This is Prog96-I.gz in view mode; [Up]
From: dale@pegasus (Dale Brisinda) Newsgroups: comp.sys.next.programmer Subject: Problem w/ Instance Drawing Flicker? Date: 31 Dec 1995 00:44:41 GMT Organization: University of Calgary CPSC Message-ID: <4c4mdp$cic@linux.cpsc.ucalgary.ca> I have a medium sized image (64x64) that I need to drag around a custom view. Problem is that using instance drawing in the standard way causes the image to flicker unbearably. Does anyone know of a way to get rid of this flicker? The code I'm using right now has the following basic structure: - mouseDown:(NXEvent *)theEvent { [ . . . ] [self lockFocus]; PSsetinstance(YES); while (nextEvent->type != NX_LMOUSEUP) { nextEvent = [NXApp getNextEvent:(NX_LMOUSEUPMASK | NX_LMOUSEDRAGGEDMASK)]; currentPoint = nextEvent->location; [self convertPoint:¤tPoint fromView:nil]; PSnewinstance(); [image composite:NX_SOVER toPoint:¤tPoint]; } PSsetinstance(NO); [self unlockFocus]; [ . . . ] } Dale --- Dale Brisinda, Grad. Student | We write in sand the benefits we Dept. of Computer Science | receive, but injuries upon marble. University of Calgary | -- Sir Thomas More (1477?-1535) brisinda@cpsc.ucalgary.ca OR dale@pegasus.cuc.ab.ca (NeXTmail) WWW: http://www.cpsc.ucalgary.ca/~brisinda/home.html
From: andrew@otter.cuug.ab.ca(Andrew Tyldesley) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer,comp.sys.next.software Subject: Re: Are you missing bcp on your black Sybase server?? Date: 31 Dec 1995 03:47:17 GMT Organization: Calgary UNIX Users' Group Message-ID: <4c5145$3vp@hp715.cuug.ab.ca> References: <DJrDF0.9CB@cuug.ab.ca> <RDL.95Dec19234317@world.std.com> <4bjaal$aq4@tempest.symnet.net> > Robert La Ferla (rdl@world.std.com) wrote: > : bcp is indeed missing. It shouldn't be hard to rewrite it. No it probably wouldn't be hard to write, but why would one want to when it used to ship with NeXTstep. Anyhow someone sent me a copy and it works fine. Art Isbell informed me that he believes it last shipped on the 3.0 CD as part of the Sybase_DB-Library.pkg package. Andrew
Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer,comp.sys.next.software From: rdl@world.std.com (Robert La Ferla) Subject: Re: Are you missing bcp on your black Sybase server?? In-Reply-To: andrew@otter.cuug.ab.ca's message of 31 Dec 1995 03:47:17 GMT Message-ID: <RDL.95Dec31031914@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <DJrDF0.9CB@cuug.ab.ca> <RDL.95Dec19234317@world.std.com> <4bjaal$aq4@tempest.symnet.net> <4c5145$3vp@hp715.cuug.ab.ca> Date: Sun, 31 Dec 1995 08:19:14 GMT Because NS 3.0 doesn't run on Intel, Sparc, or HP computers. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4c5145$3vp@hp715.cuug.ab.ca> andrew@otter.cuug.ab.ca(Andrew Tyldesley) writes: Xref: world comp.sys.next.programmer:26040 comp.sys.next.software:24262 comp.sys.next.sysadmin:27357 Path: world!news.kei.com!nntp.coast.net!howland.reston.ans.net!newsfeed.internetmci.com!newsfeed.direct.ca!sparky.insinc.net!cuugnet!usenet From: andrew@otter.cuug.ab.ca(Andrew Tyldesley) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer,comp.sys.next.software Date: 31 Dec 1995 03:47:17 GMT Organization: Calgary UNIX Users' Group Lines: 10 References: <DJrDF0.9CB@cuug.ab.ca> <RDL.95Dec19234317@world.std.com> <4bjaal$aq4@tempest.symnet.net> NNTP-Posting-Host: ppp30.cuug.ab.ca X-Newsreader: RadicalNews (TM) v0.7 Beta > Robert La Ferla (rdl@world.std.com) wrote: > : bcp is indeed missing. It shouldn't be hard to rewrite it. No it probably wouldn't be hard to write, but why would one want to when it used to ship with NeXTstep. Anyhow someone sent me a copy and it works fine. Art Isbell informed me that he believes it last shipped on the 3.0 CD as part of the Sybase_DB-Library.pkg package. Andrew
From: sanjeev@ee.umr.edu (Sanjeev Agarwal) Newsgroups: comp.sys.next.programmer Subject: Help on Objective-C++ Date: 31 Dec 1995 16:05:41 GMT Organization: UMR Missouri's Technological University Message-ID: <4c6ccl$kj7@hptemp1.cc.umr.edu> Hello there, This is my first mail to this news group. I am kind of new to using Objective-C on NextStep machines. I was tring to integrate my C++ code for Vision tasks with existing GUI developed using Next's Interface Builder in Objective-C. I created an interface file which has mixed code (both C++ and Objective C). The header file of this interface file looks something like this @interface OCEdge: Object { // some normal Objective-C instance variable ..... // C++ instance variable... class Edge *anEdge; } // the list of methods @end Now my problem is that when I include (using #import) the complier tells me undefined type: found 'class' Also how do I declare C++ class (Edge in this case) in the above header file. Thanx in advance ... Sanjeev sanjeev@isc.umr.edu
From: jbf@mitre.org (James B. Frazer) Newsgroups: comp.sys.next.programmer Subject: Multiplatform Library Support Date: Sun, 31 Dec 1995 15:24:18 -0500 Organization: The MITRE Corporation Message-ID: <jbf-3112951524180001@mbppp13.mitre.org> When applying the 3.3 patch on my NS Turbo I noticed that the distributed libraries are "fat" but that the installer lipoed them down to NeXT. I suppose this is OK for the shared libraries, since any calls can be resolved on the target machine, but do I need to reload, e.g. libposix, fat in order to build a fat application? I didn't have much luck in finding a place in the NS documentation where the building of fat applications and the library setup were clearly explained. Could someone provide pointers? Thanks, Barney
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Multiplatform Library Support Date: Sun, 31 Dec 1995 23:59:04 GMT Organization: P & L Systems Sender: news@seer. Message-ID: <1995Dec31.235904.9530@seer.demon.co.uk> References: <jbf-3112951524180001@mbppp13.mitre.org> In article <jbf-3112951524180001@mbppp13.mitre.org> jbf@mitre.org (James B. Frazer) writes: > When applying the 3.3 patch on my NS Turbo I noticed that the distributed > libraries are "fat" but that the installer lipoed them down to NeXT. I suppose > this is OK for the shared libraries, since any calls can be resolved on the > target machine, but do I need to reload, e.g. libposix, fat in order to build > a fat application? I didn't have much luck in finding a place in the NS > documentation where the building of fat applications and the library setup > were clearly explained. Could someone provide pointers? You are given the choice by the installer to install packages as fat as you like. Only the basic components of the system are forced thin. I typically install all (NeXTSTEP User/Developer) packages thin, apart from the Developer Libraries, to give me as much flexibility as possible. I know that some people prefer to install Developer Tools fat to an NFS mounted directory, and others like to install everything thin, then finish off any development projects with a remote build. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul (under construction)
From: gshaw@zeta.org.au (Greg Shaw) Newsgroups: comp.sys.next.programmer Subject: NeXTEXPO in June/July '96 Date: 1 Jan 1996 07:25:36 GMT Organization: Kralizec Dialup Internet Sydney, +61-2-837-1183 V.32bis Message-ID: <4c829g$tlu@gidora.kralizec.net.au> Hi all, Does anyone have any info on when NeXTEXPO will be held? I seem to remember that it is usually around June/July. I hope to be in California around this time and would like to attend the conference and make the trip a business deduction. Any information would be appreciated. Cheers, Greg Shaw.
From: stevem@RADium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Re: Does anybody know where sigset_t is defined? Date: 31 Dec 1995 21:09:08 GMT Organization: ftn Internet Distribution: world Message-ID: <4c6u5k$5to@master.ftn.net> References: <4bf2g1$fcc@mordred.gatech.edu> In <4bf2g1$fcc@mordred.gatech.edu> Andreas Dieberger wrote: | Hi. | | Sorry to bother you about a maybe very stupid question. I | am trying to compile the LambdaMOO system on a NeXT. Its | not my NeXT and I have access via telnet only and no | access to any documentation. [Stuff Deleted for Brevity] | >The problem is that the NeXT folks provide certain POSIX functions | >(sigemptyset, sigaddset, and sigprocmask) but don't declare the type of their | >arguments (sigset_t) in the right header file <signal.h>. If someone will tell | >me where they *do* declare that type (*if* they do), I'll try to add a case to | >the configuration script to work around NeXT's bug. | | | Can somebody of you help, please? | | Thanks a lot, | | Andreas | ----------------------------------------------------------- | Andreas Dieberger | Georgia Institute of Technology | School of Literature, Communication and Culture | Atlanta, Georgia 30332-0165 | Tel: (404) 894-8009 Fax: (404) 894-1287 | | andreas.dieberger@lcc.gatech.edu dieberger.chi@xerox.com | http://www.gatech.edu/lcc/idt/Faculty/andreas_dieberger/ | ----------------------------------------------------------- | you want to do the following: #include <sys/signal.h> which will include sigset_t (which is typedef-ed as int). This will only work if you are compiling the program as a posix program, either with a -posix flag, or a -DPOSIX_SOURCE flag. Hope this helps!! -- ==================================================== Stephen MacDougall RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: Tommy K Hwang <thwang@home.mysolution.com> Newsgroups: comp.sys.next.programmer Subject: [INN] INN1.4UNOFF3 compiled... Date: Tue, 02 Jan 1996 21:47:38 +0900 Organization: Internet Solutions Inc. Message-ID: <30E9296A.20D8@home.mysolution.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit If my inn works... This post should go out to the net... INN1.4UNOFF3 has been compiled for NextStep 3.3 on both the Intel and NeXT systems.
From: s9510225@bf.rmit.edu.au (Louis Lai) Newsgroups: comp.sys.next.programmer Subject: Programmers' behavior & needs survey Date: 2 Jan 1996 01:47:34 GMT Organization: Royal Melbourne Institute of Technology, Melbourne, Australia. Message-ID: <4ca2rm$2pa@aggedor.rmit.EDU.AU> NNTP-Posting-User: s9510225 Software development provides some of business's most lurid horror stories. Stories of projects only one-third finished but already costing double or triple their budgets are common. So are stories of projects that were outrageously late and those that never work at all. There are many views to this problem. "The major problems of our work are not so much technological as sociological in nature. Most managers are willing to concede the idea that they've got more people worries than technical worries. But they seldom manage that way." -- DeMarco & Lister: Peopleware(1987) "Many programmers and analysts have complained to me that their work holds no meaning to them... [They] stay on the job, draw the salary, and bad-mouth their employers at every safe opportunity." -- Weinberg: Understanding the Professional Programmer(1988) "While the software [programmers] develop becomes increasingly sophisticated, relations between programmer and manager remain shockingly primitive." -- Licker: The Art of Managing Software Dev. People(1985) "[Programmers are] arrogance's born in a belief of rectitude and superiority, and [are] ignorance's born in a blindness to shortcomings." -- O'Brien: Software Production Management(1992) Whatever the reasons are, there's a need to understand the programmers themselves. The aim of this research is to elicit information on the needs and behavior of software programmers. Information that will, hopefully, better managers in managing these professionals. If you are working or have worked as a professional programmer, please take 15 to 20 minutes to answer the following questionnaire. This questionnaire is in: http://www.bf.rmit.edu.au/~s9510225/ The results will be posted at this location also. I like to thank in advance to those who participated in this survey. Your contribution will surely advance the knowledge in this field. Thank you. Yours sincerely, Louis Lai, s9510225@bf.rmit.edu.au Faculty of Business, Royal Melbourne Institute of Technology, Australia.
From: robert@steffi.accessone.com (Robert Nicholson) Newsgroups: comp.sys.next.programmer Subject: Re: Multiplatform Library Support Date: 01 Jan 1996 21:32:04 -0800 Organization: x Sender: robert@steffi.accessone.com Message-ID: <ewx7byxez.fsf@steffi.accessone.com> References: <jbf-3112951524180001@mbppp13.mitre.org> <1995Dec31.235904.9530@seer.demon.co.uk> To: Paul_Lynch@plsys.co.uk <Paul_Lynch@plsys.co.uk> writes: >In article <jbf-3112951524180001@mbppp13.mitre.org> jbf@mitre.org (James >B. Frazer) writes: >>When applying the 3.3 patch on my NS Turbo I noticed that the >distributed >>libraries are "fat" but that the installer lipoed them down to NeXT. I >suppose >>this is OK for the shared libraries, since any calls can be resolved on >the >>target machine, but do I need to reload, e.g. libposix, fat in order to >build >>a fat application? I didn't have much luck in finding a place in the NS >>documentation where the building of fat applications and the library >setup >>were clearly explained. Could someone provide pointers? >You are given the choice by the installer to install packages as fat as >you like. Only the basic components of the system are forced thin. RE: Installer. Here's a question I'd like an answer on. If I install a MAB package m68k only. Do I lose the Intel etc part of that package after the Install? ie. I can only compress to re-install a previously installed package. I guess the key here is that /NextLibrary/Receipts only reflects the architecture that you originally installed. -- "For I am Costanza, lord of the Idiots" (PGP key: send email with Subject: request pgp key)
From: robert@steffi.accessone.com Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <ewx7byxez.fsf@steffi.accessone.com> Control: cancel <ewx7byxez.fsf@steffi.accessone.com> Date: 02 Jan 1996 00:18:24 -0800 Organization: x Message-ID: <eohsnyppr.fsf@steffi.accessone.com> This is a cancel message from robert@steffi.accessone.com (Robert Nicholson).
From: hwr@pilhuhn.de (Heiko W.Rupp) Newsgroups: comp.sys.next.programmer Subject: Re: [INN] INN1.4UNOFF3 compiled... Date: 2 Jan 1996 11:45:40 GMT Organization: The Home Of The Pilhuhn Message-ID: <slrn4ei6n4.jlp.hwr@pilhuhn.de> References: <30E9296A.20D8@home.mysolution.com> In <30E9296A.20D8@home.mysolution.com>, Tommy K Hwang wrote: :|If my inn works... This post should go out to the :|net... Please *.test for testing your news system. :| INN1.4UNOFF3 has been compiled for NextStep 3.3 :|on both the Intel and NeXT systems. Please send me your config.data. -- Heiko W.Rupp Gerwigstr.5 D-76131 Karlsruhe +49 721 9661524 Usenet is "the last uncensored mass medium." - Steve Crocker (aq817@cleveland.Freenet.Edu)
From: gguelden@gandalf. (Gerd Gueldenpfennig) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTEXPO in June/July '96 Date: 2 Jan 1996 14:14:57 GMT Organization: iXpoint Informationssysteme GmbH, Ettlingen, Germany Message-ID: <4cbel1$205@balu.ixpoint.de> References: <4c829g$tlu@gidora.kralizec.net.au> gshaw@zeta.org.au (Greg Shaw) wrote: > Hi all, > Does anyone have any info on when NeXTEXPO will be held? I seem to remember that it is usually around June/July. I hope to be in California around this time and would like to attend the conference and make the trip a business deduction. NeXTSTEP Expo died 1994. But there were a lot of NEXTSTEP-Infos at ObjectWorld Expo 1995. The next ObjectWorld Expo is as far as I know in the first week of august 1996 in San Jose (CA). --- ++++++++++++++++++++++++++++++++++++++++++++++++++ + Dipl.- Inform. Gerd Gueldenpfennig + + iXpoint Informationssysteme GmbH + + Daimlerstrasse 3 76275 Ettlingen Germany + + Phone ++49 7243 3775-0 Fax ++49 7243 3775-77 + + Email: gguelden@ixpoint.de (NeXTmail and MIME) + ++++++++++++++++++++++++++++++++++++++++++++++++++ + QMSprint - Network Printing for NEXTSTEP + + MERCATOR - GIS and Desktop Mapping for NEXTSTEP+ ++++++++++++++++++++++++++++++++++++++++++++++++++
From: wfc@cl.cam.ac.uk (W F Clocksin) Newsgroups: comp.sys.next.programmer Subject: Lazy Popup Lists? Date: 2 Jan 1996 16:01:09 GMT Organization: University of Cambridge, England Message-ID: <4cbks5$bct@lyra.csx.cam.ac.uk> Has anyone invented Lazy PopUp lists yet? These would be analogous to lazy browsers. I would prefer a way to do the least amount of work to a Popup list when minor modifications are made to it. For example, just before displaying a view containing a popup list of 30 items, suppose only two or three of the items are different since the last time the popup list was used, and say the number of items is different by two or three. I would like to do very minor surgery to the popup list instead of (in the worst case) removing all the items and then addItem: them back again. I would also prefer to do this "by the book", so that the various unrecommended procedures concerning menucells are not followed. Has anyone done this?
Newsgroups: comp.sys.next.programmer From: gerald@kurt.in-berlin.de (Gerald Erdmann) Subject: How to check if text object saves as rtfd Message-ID: <1996Jan1.213407.2164@kurt.in-berlin.de> Sender: news@kurt.in-berlin.de Date: Mon, 1 Jan 1996 21:34:07 GMT Hi! Does someone knows a simple way to check if a text object would save as rtfd (WITH an document folder) or as normal rtf (WITHOUT an document folder)? I don't like my current way (saveRTFDTo::: and check if my doc is an directory). Any hints are welcome. Happy new year. Gerald -- --------------------------------------------------------------------- | GERALD ERDMANN | email: gerald @ kurt.in-berlin.de (NeXTmail welcome) | voice: +49 30 372 43 10 (Germany - Berlin) | crypt: pgp2 public key available |
From: Dominik Westner <dominik@gowest.ppp.informatik.uni-muenchen.de> Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: EOF: 2nd independent fetch frees NXImage of 1st fetch Date: 2 Jan 1996 23:36:07 GMT Organization: TeDoc, Munich, Germany Message-ID: <4ccfh7$bj@cube.ppp.informatik.uni-muenchen.de> Hi, I encountered the following problem: - fetch an image from a database and display it in a NXImageView. - create a 2nd datasource+controller and fetch the same table containing that image again without displaying the image - sending the NXImageView a display message will crash the app, because the NXImage was freed by the 2nd fetch. At least this is what the stack tells me: #0 0x602baa3 in -[NXImage free] () #1 0x7835ddb in -[Object(NSCompatibility) dealloc] () #2 0x7835c29 in -[Object(NSCompatibility) release] () #3 0xe02983b in -[_EOMutableKnownKeyDictionary dealloc] () #4 0x782162b in -[NSObject release] () #5 0xe02694e in -[_EOUniqueTable recordObject:primaryKeyString:snapshot:] () #6 0xe00a554 in -[EODatabase recordObject:primaryKeyString:snapshot:] () #7 0xe02810b in -[EODatabaseChannel(Private) _fetchObject:withZone:] () #8 0xe00c90a in -[EODatabaseChannel fetchWithZone:] () #9 0xe00db6b in -[EODatabaseDataSource fetchObjects] () #10 0xf0087b6 in -[EOController _fetch:] () #11 0xf008990 in -[EOController fetch] () Can anybody enlighten me why this should happen? Is there a workaround/patch? Valid for EOF 1.1. -- Dominik --- Dominik Westner "Everything is possible in an infinite universe" _____________________________________Willi, the painter Munich, Germany (NOT :-)
From: hugues@precipice.fdn.fr (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTEXPO in June/July '96 Date: 2 Jan 1996 18:14:23 GMT Organization: Individual - France Message-ID: <4cbslv$29l@precipice.fdn.fr> References: <4c829g$tlu@gidora.kralizec.net.au> gshaw@zeta.org.au (Greg Shaw) wrote: >Hi all, >Does anyone have any info on when NeXTEXPO will be held? I seem to >remember that it is usually around June/July. ^^ was NeXTEXPO is dead. Hugues. -------------------------------------------------------------------- hugues@precipice.fdn.fr - France (small NeXTMail OK) ------------ NS3.2 ------------ NS3.0J ------------ :-) ------------
From: "Ingo Feulner" <ifeulner@xenon.stgt.sub.org> Newsgroups: comp.sys.next.programmer Subject: Re: Developing for NEXTSTEP vs OPENSTEP---a critical business decision Date: 2 Jan 1996 19:14:26 GMT Organization: Private site, Boeblingen, Germany Message-ID: <4cc06i$a5v@xenon.stgt.sub.org> References: <1995Dec27.234416.4569@heinz.com> In <1995Dec27.234416.4569@heinz.com> Federico Heinz wrote: > The technical point is the following: on one hand, NEXTSTEP > 4.0 will run NEXTSTEP 3.3 applications (at least NeXT > promised us it would), but of course it won't work the other > way around. On the other hand, NEXTSTEP 3.3 applications > will neither run nor _compile_ under OPENSTEP/Solaris > (NEXTSTEP 4.0 apps will run under OPENSTEP/Solaris, but it > requires a recompile). Now suppose you have, or are in the > process of creating an app, should you stick with NEXTSTEP > 3.3 for the time being, or just jump to the OPENSTEP > bandwagon? There are more problems: OpenStep for MachOS (aka NEXTSTEP 4.0) has some (really nice and useful classes, such as the new Text class) which are not yet in the OpenStep spec. And so these objects are not in OpenStep for Solaris (which is based/developed from NEXTSTEP 3.3). Because of this, we have to ask us one more question: Which OpenStep should I use??? (For me: I'll take the original) Questions over questions.... happy new year, Ingo. -- Smail: Ingo Feulner, Wolfacher Weg 19, 71034 Boeblingen, Germany Email: ifeulner@xenon.stgt.sub.org
From: ian_stewart@nyro.com (Ian H. Stewart) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: HTTP 1.05 MAB makefile(s) requested. Date: 3 Jan 1996 02:18:23 GMT Organization: InterNex Information Services 1-800-595-3333 Distribution: world Message-ID: <4ccp1f$55n@voyager.Internex.NET> Hi, I'm looking for someone that has modified the 1.05 HTTPD server makefiles to support ALL NS archs. NeXTmail me a copy of the files if you have them. thanks, Ian
From: Žfalkman@hegel2.cs.chalmers.se (G|ran Falkman) Newsgroups: comp.sys.next.programmer Subject: Windows with transparent background? Date: 3 Jan 1996 09:16:17 GMT Organization: Chalmers University of Technology Message-ID: <4cdhh1$dfk@nyheter.chalmers.se> Is it possible to have windows (and views) with transparent background? If this is the case, can anyone give me hint on how? Goran Falkman _____________________________________________________________________________ Goran Falkman Tel: + 46 31 772 5412 Cognition Technology Group Fax: + 46 31 16 56 55 Department of Computing Science Email: falkman@cs.chalmers.se Chalmers University of Technology URL: http://www.cs.chalmers.se/~falkman/ S-412 96 Goteborg, Sweden
Newsgroups: comp.sys.next.programmer From: martin@rat.se Subject: Re: Windows with transparent background? Message-ID: <DKLoB1.L46@rat.se> Sender: martin@rat.se (Martin Wennerberg) Organization: Research & Trade, AB. References: <4cdhh1$dfk@nyheter.chalmers.se> Date: Wed, 3 Jan 1996 10:11:25 GMT In article <4cdhh1$dfk@nyheter.chalmers.se> falkman@hegel2.cs.chalmers.se (G|ran Falkman) writes: > Is it possible to have windows (and views) with transparent background? > If this is the case, can anyone give me hint on how? It's possible, but not easy. One way to do it is to grab the background and display it in the window. That way the window will look transparent if the background is static. See the mini example "Grabber" on how to create tiffs from the screen. Martin Wennerberg ____________________________________________________ Research & Trade AB Phone: + 46 - 8 - 21 17 50 Fax: + 46 - 8 - 24 54 11 Email: martin@rat.se [MIME & NeXTmail supported] Snailmail: Box 7742, S-103 95 Stockholm, SWEDEN Visiting: Kungsgatan 33
From: Simon Lewis <simonle@bnr.ca> Newsgroups: comp.sys.next.programmer Subject: objc_loadModules And The Disappearing Variables Date: 3 Jan 1996 13:22:43 GMT Organization: Bell Northern Research Message-ID: <4cdvv3$43a@bcrkh13.bnr.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Can anybody confirm that a call to objc_loadModules or rld_load clobbers, anongst other things, NXArgv and NXArgc ? I don't know what the others are though messing about with gdb seems to indicate that there may be a number of 'hidden' variables essential to the health and well being of Appkit which likewise 'go south'. For example, calls to strftime start core-dumping some- where in the localization stuff. Thanks for any help simon lewis -- simonle@bnr.ca (but in the uk)
Newsgroups: comp.sys.next.programmer From: ian.stephenson@insignia.co.uk Subject: Smart Popup for EOF... Message-ID: <DKM2uq.CqM@isltd.insignia.com> Sender: news@isltd.insignia.com Organization: Insignia Solutions plc Date: Wed, 3 Jan 1996 15:25:37 GMT I've recently been dabling in EOF (looks good, if not as much fun as some of the old kits). The main thing I found lacking was a way to quickly convert a number into a name, and display it, then edit the name and have the number set back in the associated object. Say you've got a list of people with employee ids. You want to say that a certain action was taken by person X. The id of X is held in the action table, but when displayed you want to see X's name. Yeah, you can add a relation, but then if you edit the name (unless you're clever) you're editing the name in the employee record not the id un the action table! Enter the SmartPopupAssociation! Associate the id to the popup button, then set the target of the first item in the pop up list (not the button) to be an EOF controller for the table you want to extract the names from. Provided that the table has a "number", and a "name" attribute then it will "just work". I've used this for a few different purposes, and its been REALLY usefull, so I'm releasing it here... /* SmartPopupAssociation.h * * * Written by I Stephenson * */ #import <eointerface/EOAssociation.h> @class Window; @interface SmartPopupAssociation: EOAssociation { id titleController; id popUpList; } // Target/action method for destination to send to - controlActed:sender; - currentValue; - setUpList; - setTitleForTag:(int)val; @end /* SmartPopupAssociation.m * * * Written by I Stephenson * */ #import <appkit/appkit.h> #import <eointerface/eointerface.h> #import "SmartPopupAssociation.h" // Override the standard Button associationClass method to spot popUps @implementation Button (SmartPopupAssoc) - (Class)associationClass { if([[self target] isKindOf:[PopUpList class]]) return [SmartPopupAssociation class]; else return [EOButtonAssociation class]; } @end @implementation SmartPopupAssociation - initWithController:(EOController *)aController key: (NSString *)aKey destination: aDest { [super initWithController:aController key: aKey destination:aDest]; popUpList=[destination target]; [(Button *)destination setTarget:self]; [(Button *)destination setAction:@selector(controlActed:)]; return self; } -awakeFromNib { titleController=[[[popUpList itemList] cellAt:0:0] target]; [[[popUpList itemList] cellAt:0:0] setTarget:nil]; [self setUpList]; [self setTitleForTag:[[self currentValue] intValue]]; return self; } - setUpList { int i; NSArray *titles; id person; i=[popUpList count]; for(;i-->0;) { [popUpList removeItemAt:i]; } //This is important! //It traps the null entries... [popUpList addItem:"Unknown"]; [[[popUpList itemList] cellAt:0:0] setTag:0]; titles=[[titleController dataSource] fetchObjects]; for(i=0;i<[titles count];i++) { person=[titles objectAtIndex:i]; [popUpList addItem:[[person objectForKey:@"name"] cString]]; [[[popUpList itemList] cellAt:i+1:0] setTag:[[person objectForKey:@"number"] intValue]]; //printf("%d\t%s:%d\n",i,[[person objectForKey:@"name"] cString],[[person objectForKey:@"number"] intValue]); } return self; } - controlActed:sender { [self setUpList]; [self setTitleForTag:[[self currentValue] intValue]]; [popUpList popUp:sender]; [controller associationDidEdit: self]; return self; } - value { return [NSNumber numberWithInt: [[[popUpList itemList] selectedCell] tag]]; } - (NSString *)currentValue { NSArray *selectedObjects = [controller selectedObjects]; NSString *stringValue; if ([selectedObjects count] >= 1) { id selectedObject, aDict, val; id array = [[[NSArray alloc] initWithObjects: key, 0] autorelease]; selectedObject = [selectedObjects objectAtIndex: 0]; aDict = [selectedObject valuesForKeys: array]; val = [aDict objectForKey: key]; if (!val) stringValue = @""; else stringValue = [((NSObject *)val) description]; } else { stringValue = @""; } return stringValue; } - setTitleForTag:(int)val { id matrix; int i; matrix=[popUpList itemList]; for(i=[popUpList count];i-->0;) { if([[matrix cellAt:i:0] tag]==val) { [destination setTitle:[[matrix cellAt:i:0] title]]; return self; } } [destination setTitle:[[matrix cellAt:0:0] title]]; return nil; } - (void)selectionDidChange { [self setTitleForTag:[[self currentValue] intValue]]; return; } - (void)contentsDidChange { [self setTitleForTag:[[self currentValue] intValue]]; return; } @end However bear in mind that this is my first shot at EOF so I make no guarentees. Feel free to use this as you see fit. Send me some feedback if you do find it usefull (or infact if you don't find it usefull!). Ian
From: ian_stewart@nyro.com (Ian H. Stewart) Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: HTTP 1.5 I meant (makefiles needed) Date: 3 Jan 1996 21:28:23 GMT Organization: InterNex Information Services 1-800-595-3333 Distribution: world Message-ID: <4cesdn$s38@voyager.Internex.NET> CFLAGS=-g -arch m68k -arch i386 -arch sparc -arch hppa does not work, when it links, the binary is for a single arch, but the .o files are fat. Help appreciated. Ian
From: dale@pegasus (Dale Brisinda) Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: NeXT Demo App Source Release Policy? Date: 4 Jan 1996 00:09:44 GMT Organization: University of Calgary CPSC Message-ID: <4cf5s8$bi2@linux.cpsc.ucalgary.ca> Can anyone provide me with an e-mail contact address for someone at NeXT regarding their source release policy for a specific demonstration application? The info panel of the application I'm interested in simply says "contact NeXT for any questions on source release policy" without providing any additional information. Thanks for *any* information, Dale --- Dale Brisinda, Grad. Student | We write in sand the benefits we Dept. of Computer Science | receive, but injuries upon marble. University of Calgary | -- Sir Thomas More (1477?-1535) brisinda@cpsc.ucalgary.ca OR dale@pegasus.cuc.ab.ca (NeXTmail) WWW: http://www.cpsc.ucalgary.ca/~brisinda/home.html
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Windows with transparent background? Date: 3 Jan 1996 16:11:11 -0800 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4cf5uv$10k@nova.mdd.comm.mot.com> References: <DKLoB1.L46@rat.se> In article L46@rat.se, martin@rat.se writes: >In article <4cdhh1$dfk@nyheter.chalmers.se> falkman@hegel2.cs.chalmers.se >(G|ran Falkman) writes: >> Is it possible to have windows (and views) with transparent background? >> If this is the case, can anyone give me hint on how? > >It's possible, but not easy. One way to do it is to grab the background >and display it in the window. That way the window will look transparent if >the background is static. See the mini example "Grabber" on how to create >tiffs from the screen. I would hope that there is an easier way. Anyone care to guess how InterfaceBuilder draws the connections between windows on the screen? I always though that it slapped a big transparent window on the screen and drew in that. Can this be done? Cheers, Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: mcgredo@crl.com (Donald R. McGregor) Newsgroups: comp.sys.next.programmer Subject: Re: Windows with transparent background? Date: 3 Jan 1996 16:59:06 -0800 Organization: Universal Export Distribution: na Message-ID: <4cf8oq$mgs@crl13.crl.com> References: <DKLoB1.L46@rat.se> <4cf5uv$10k@nova.mdd.comm.mot.com> In article <4cf5uv$10k@nova.mdd.comm.mot.com>, Steve Besler <besler@mdd.comm.mot.com> wrote: :Anyone care to guess how IB draws the connections between windows :>on the screen?I always though that it slapped a big transparent window on the :>screen and drew in that. Can this be done? Close. It draws two long, skinny windows to simulate rectilinear lines. There used to be some example code of this out there somewhere. -- Don McGregor | "I did not get my Spaghetti-Os. I got spaghetti. I want mcgredo@crl.com | the press to know this."
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: US-Nationwide - TX/IL/NY/CA/NC - NeXTSTEP Developers - Top $$$$$ Date: 3 Jan 1996 20:30:47 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4cfak7$7r3@newsbf02.news.aol.com> Many of my best clients have embraced NeXTSTEP as the best OO development environment currently on the market. They have invested heavily in NeXT technology, and current projects are paying excellent dividends. Due to the my clients success with NeXTSTEP, they are aggressively hiring experienced developers for major 1996 projects. Candidates must have a development background using NeXTSTEP and Objective-C, coupled with some experience in Sybase, Informix, or Oracle. In return you will receive an excellent salary package which may include bonus and relocation. H-1 candidates may apply. Positions are available in TX, IL, NY, CA, NC, and London (England). Contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
From: woodward@onramp.net (John Woodward) Newsgroups: comp.sys.next.programmer Subject: Re: Windows with transparent background? Date: 4 Jan 1996 04:12:44 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Distribution: na Message-ID: <4cfk3s$jo9@news.onramp.net> References: <4cf5uv$10k@nova.mdd.comm.mot.com> In article <4cf5uv$10k@nova.mdd.comm.mot.com> besler@mdd.comm.mot.com (Steve Besler) writes: > Anyone care to guess how InterfaceBuilder draws the connections between > windows on the screen? I always though that it slapped a big > transparent window on the screen and drew in that. Can this be done? This was explained long ago on the net as: the lines a drawn by making very skinny windows! You draw a short & wide window, with a narraw and tall window to form the 2-part lines. john (woodward@onramp.net)
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Sybase from EOModeler Date: 4 Jan 1996 16:24:40 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4cgv08$ak2@stc06.ctd.ornl.gov> I just installed my EO 1.1 and while trying out the Enterprise Model app, I enter exactly what is stated in the documentation "People" for Database Name (new one), blank for Server, sa for user name, ( I have no idea what the passwd would be), and click okay. 1. I am running a Sybase server that works find with the Pubs Demo. 2. I don't know what the passwd and server name etc are because it never asks me. I get the message Alert sybase Server:unknown Procedure:none Line:0 Message:Login failed If i enter query or master in the server name (then it tells me this name is not found in the interfaces file). Where are these files? sybase seems to be version 1.1 (v68) but seems to work okay except for this. I also got a panic crash when I tried to open pubs from the EO Panel. -- - - - - - - - - - J. W. Wooten
From: jehu@blackbox.cl.ee.vt.edu (John Stanhope) Newsgroups: comp.sys.next.programmer Subject: Worth while to install EOF 1.1 ??? Date: 4 Jan 1996 16:44:05 GMT Organization: Virginia Tech, Blacksburg, Virginia Distribution: world Message-ID: <4ch04l$k4l@solaris.cc.vt.edu> I just upgraded to 3.3 as an academic user which means I get EOF 1.1 with NS and I was wondering if it is worth while to install this? I really don't plan on doing any database work but I would like to play with foundation a little. The real reason I ask is that in the EOF release notes it says that you need to install a patch for foundation that makes foundation stuff incompatible with non EOF foundation. How much of my foot am I shooting away if I do this? John
Newsgroups: comp.sys.next.programmer From: tom@basil.icce.rug.nl (Tom Hageman) Subject: Re: OpenStep nib conversion Message-ID: <DKMKMp.2LD@basil.icce.rug.nl> Originator: uubasil@obelix.icce.rug.nl Sender: news@basil.icce.rug.nl (NEWS pusher) Organization: Warty Wolfs References: <4bv93u$dgq@emerald.oz.net> Date: Wed, 3 Jan 1996 21:49:36 GMT art@cubicsol.com (Art Isbell) wrote: > NeXT appears to have done a VERY thorough job producing automated code > conversion scripts that will do most of the NS-to-OpenStep (OS) code > conversion. Andrew Stone, Stone Design, has produced a nice guide that > chronicles his experiences with the OS 4.0 PR1 conversion process with an > extensive problem/solution section. Is this guide avaliable somewhere? -- __/__/__/__/ Tom Hageman <tom@basil.icce.rug.nl> [NeXTmail/Mime OK] __/ __/_/ IC Group <tom@icgned.nl> (work) __/__/__/ __/ _/_/ "Unix? What's that? Is that like Linux?"
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Does anybody know where sigset_t is defined? Date: 4 Jan 1996 20:32:09 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4chdg9$kit@news.its.com> References: <4bf2g1$fcc@mordred.gatech.edu> Andreas Dieberger <andreas.dieberger@lcc.gatech.edu> wrote: >>>>my-signal.h:34: undefined type, found `sigset_t' [ ... ] >> The problem is that the NeXT folks provide certain POSIX functions >> (sigemptyset, sigaddset, and sigprocmask) but don't declare the type >> of their arguments (sigset_t) in the right header file <signal.h>. This is not correct. From /usr/include/bsd/sys/signal.h: [ ... ] #ifdef _POSIX_SOURCE /* Additional entries required by */ /* POSIX */ #include <sys/types.h> /* for pid_t */ typedef int sigset_t; [ ... ] In order to use such POSIX functions, you simply must compile and link using "cc -posix". Don't try to do anything else, such as having "-lposix" in the link line, since that will not have the executable linked correctly with the right crt0.o startup hook. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: David Spooner <spoonerd> Newsgroups: comp.sys.next.programmer Subject: digital video query Date: 4 Jan 1996 22:01:50 GMT Organization: University of Calgary CPSC Message-ID: <4chioe$4d@linux.cpsc.ucalgary.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I am developing an application for digital video processing, and am interested in using nextstep. I am concerned with the existence of the following : - image manipulation libraries (e.g. segmentation routines) - software encoders/decoders for mpeg1 or 2 - drivers for mpeg encoding/decoding cards Any information (by email) will be greatly appreciated. Also, any speculation as to whether openstep for windows will allow one to link to windows libraries? Thanks, Dave Spooner spoonerd@cpsc.ucalgary.ca
From: jcassidy@proton.genesoft.com. (James Cassidy) Newsgroups: comp.sys.next.programmer Subject: Which lib for tcgetattr? Date: 5 Jan 1996 04:35:54 GMT Organization: HookUp Communication Corporation, Waterloo, Ontario, CANADA Message-ID: <4ci9ra$to9@nic.wat.hookup.net> Hello, I'm using tcgetattr and tcsetattr calls in a program and I'm getting a message from ld that these are undefined symbols. Would anyone be able to inform me as to which library they are in? Tbanks in advance, Jim.
From: murphy@wsc.com (Paul Murphy) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: 1,600 bpi vs. 6,250 bpi on a mag tape Date: 04 Jan 1996 23:14:09 -0500 Organization: WSCIS Sender: murphy@eos Message-ID: <x4d98zqnvx.fsf@eos> Does anyone happen to know how to get a mag tape -- attached to a SCSI bus, on a NeXT machine -- to write at 6250bpi instead of the default 1600? Thanks for your help. Paul Murphy
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: objc_loadModules And The Disappearing Variables Date: 5 Jan 1996 00:41:20 GMT Organization: my own Message-ID: <4chs3g$qe5@gate.seicom.net> References: <4cdvv3$43a@bcrkh13.bnr.ca> Simon Lewis <simonle@bnr.ca> wrote: > Can anybody confirm that a call to objc_loadModules or rld_load clobbers, > anongst other things, NXArgv and NXArgc ? I don't know what the others are > though messing about with gdb seems to indicate that there may be a number > of 'hidden' variables essential to the health and well being of Appkit which > likewise 'go south'. For example, calls to strftime start core-dumping some- > where in the localization stuff. I havn't heard this yet. I recently figured out rld_load sometimes makes problems while loading the symbol table (it may got cut in the middle, so Bad Luck(tm) and SIGSEGV when you want to parse it by hand...) but got a (hackish) workaround for this. Well, I didn't use the Appkit for this application either.... - Frank --- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP & PostScript Guy "In cantonese C++ is called C ga ga"
From: dsdecasp@iiic.ethz.ch (Daniel Stefan Decasper) Newsgroups: comp.sys.next.programmer Subject: Re: Sybase from EOModeler Date: 5 Jan 1996 09:04:54 GMT Organization: Dept. Informatik, Swiss Federal Institute of Technology (ETH), Zurich, CH Message-ID: <4cipjm$rrd@neptune.ethz.ch> References: <4cgv08$ak2@stc06.ctd.ornl.gov> In article <4cgv08$ak2@stc06.ctd.ornl.gov>, [...] > >I get the message Alert >sybase Server:unknown Procedure:none Line:0 > Message:Login failed > >If i enter query or master in the server name (then it tells me this name is >not found in the interfaces file). Where are these files? > The location of the interfaces file depends on your installation, normally it's /usr/sybase/interfaces. Use SYBASE as server name (this is the default). -Dan
From: yiannis@prologos.nrl.navy.mil (John Michopoulos) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: seeking flex & bison for NEXTSTEP? Date: 5 Jan 1996 14:50:29 GMT Organization: Naval Research Laboratory Message-ID: <4cjdrl$aep@ra.nrl.navy.mil> Where can I find flex & bison compiled for NEXTSTEP? ftp.cs.orst.edu does not have them and the ones I picked up from aeneas.mit.edu gnu archive do not compile well! So, before I go into solving this problems I though I should look for compiled versions. Thanks all for your help --y ------------------------------------------------------------------- | Dr.John Michopoulos (yanni)| Tel: (202) 767-2165 or -2189 | | Naval Research Laboratory | Fax: (202) 767-9181 | | Code 6380 |e-mail: yiannis@prologos.nrl.navy.mil| | Washington DC 20375-5000 | michopoulos@anvil.nrl.navy.mil|
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: seeking flex & bison for NEXTSTEP? Date: 5 Jan 1996 19:37:31 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4cjulr$8fg@agate.berkeley.edu> References: <4cjdrl$aep@ra.nrl.navy.mil> In article <4cjdrl$aep@ra.nrl.navy.mil>, John Michopoulos <yiannis@prologos.nrl.navy.mil> wrote: >Where can I find flex & bison compiled for NEXTSTEP? ftp.cs.orst.edu does >not have them and the ones I picked up from aeneas.mit.edu gnu archive do >not compile well! So, before I go into solving this problems I though I >should look for compiled versions. I downloaded the GNU bison sources from gatekeeper.dec.com, compiled them with NeXT cc and GNU make (available at ftp.cs.orst.edu), and had no troubles. (I then used bison and GNU make to compile gcc 2.7.2.) I don't know about flex. Paul
From: dave james <djames@isl.com> Newsgroups: comp.sys.next.programmer Subject: source control software suggestion please? Date: 5 Jan 1996 19:38:58 GMT Organization: Internet Systems Ltd Message-ID: <4cjuoi$nct@bronwyn.isl.com> Hi there, we have a developement initiative under way and would like to hear any reccomendations that people have as to source control techniques or software. Thanks in advance
From: gerdes@ucla.edu Newsgroups: comp.sys.next.programmer Subject: Gnu C++ and NeXTSTEP Date: 5 Jan 1996 22:20:40 GMT Organization: University of California, Los Angeles Distribution: world Message-ID: <4ck87o$18qs@saba.info.ucla.edu> I am programming in C++ on a sun Sparc-station using the latest distribution of GNU C++ (version 2.7) and would like to use it or the distribution at ftp.cs.orst.edu (version 2.6) on my black cube running 3.3. I currently have NeXT developer 3.2 installed on my machine and the version of GNU C++ that came with it works fine. However, one drawback is that it does not support templates which keeps me from running some freeware libraries that use them. Does anyone have experience updating their GNU C++ compiler? What do I need to know? Clearly, I also want to install the latest GNU C++ library. Thanks in advance.
From: dsolis@dsp (David Solis) Newsgroups: comp.sys.next.programmer Subject: IBEditors for view class Date: 5 Jan 1996 16:10:11 GMT Organization: Datanet S.A. de C.V., Mexico City Message-ID: <4cjih4$1tn@bsdi.data.net.mx> Does somebody developed IBEditors protocol for a view class? I developed a SpreadSheet palette and I'd like to use an IBEditor similar to NXTableView's editor for editing headers -NXTableVector's title. Thanks for any help David Solis NeXT & MIME Mail:dsolis@data.net.mx LegoSoft S.C. Voice:52-5-2450477 Palacio de Versalles No. 285-A Fax:52-5-5966162 Col. Lomas Reforma Mexico, D.F. 11020
Newsgroups: comp.sys.next.programmer From: pfkeb@hpkaon.SLAC.Stanford.EDU (Paul F. Kunz) Subject: Re: source control software suggestion please? In-Reply-To: dave james's message of 5 Jan 1996 19:38:58 GMT Message-ID: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> Sender: news@unixhub.SLAC.Stanford.EDU Organization: Stanford Linear Accelerator Center References: <4cjuoi$nct@bronwyn.isl.com> Date: Sat, 6 Jan 1996 05:21:07 GMT >>>>> On 5 Jan 1996 19:38:58 GMT, dave james <djames@isl.com> said: In article <4cjuoi$nct@bronwyn.isl.com> dave james <djames@isl.com> writes: > Hi there, we have a developement initiative under way and would like > to hear any reccomendations that people have as to source control > techniques or software. Thanks in advance Native CVS works fine for NeXTSTEP development. CVS is free and available with GNU software. Fancy GUI front-ends to CVS exist for NeXTSTEP at a price; freely available front-end based in Tcl/Tk are also available. But CVS is so easy to use who needs a GUI front-end? -- Paul F. Kunz Paul_Kunz@slac.stanford.edu (NeXT mail ok) Stanford Linear Accelerator Center, Stanford University Voice: (415) 926-2884 (NeXT) Fax: (415) 926-3587
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: TeX3.14? web2c-2.61? flex-2.5? (Small world...) Date: 6 Jan 1996 07:31:58 GMT Organization: U.C. Berkeley Math. Department. Distribution: na Message-ID: <4cl8hf$qc3@agate.berkeley.edu> I need to recompile TeX with larger memory allocation parameters (the commutative diagrams in my thesis push NeXTTeX over the 5500 string limit), and I'm not programmer enough to edit the sources for flex, kpathsea, and web2c so that they compile properly on black hardware (which is all the hardware I've got!). Does anyone out there have executables or suggestions? And, considering that TeX is now old enough to drive, why hasn't someone rewritten it with dynamic memory allocation?!??! Thanks for any help. Paul
From: Tommy K Hwang <thwang@home.mysolution.com> Newsgroups: comp.sys.next.programmer Subject: cmd.h Date: Sun, 07 Jan 1996 04:39:20 +0900 Organization: Internet Solutions Inc. Message-ID: <30EECFE8.4E24@home.mysolution.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Where did cmd.h header went? Did it ever existed for NeXT? I got User Quota manager compiled except for ONE last section that need the cmd.h header. Once that part is done, I can test and release it for NS3.3.
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Which lib for tcgetattr? Date: 6 Jan 1996 08:44:12 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4clcos$hlr@news.its.com> References: <4ci9ra$to9@nic.wat.hookup.net> jcassidy@proton.genesoft.com. (James Cassidy) wrote: > Hello, I'm using tcgetattr and tcsetattr calls in a program and I'm > getting a message from ld that these are undefined symbols. Would > anyone be able to inform me as to which library they are in? Well, "man tcgetattr" will indicate that they are in section 3P, ie POSIX-specific calls. You need to compile with "cc -posix" to use those calls, which in turn means that you cannot use the AppKit. (The functions actually are in /usr/lib/libposix.a, but you cannot link from that library and get something that will work; you have to compile with "cc -posix".) -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep nib conversion Date: 6 Jan 1996 15:56:55 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4cm647$q1t@emerald.oz.net> References: <4bv93u$dgq@emerald.oz.net> <DKMKMp.2LD@basil.icce.rug.nl> tom@basil.icce.rug.nl (Tom Hageman) wrote: > art@cubicsol.com (Art Isbell) wrote: > > NeXT appears to have done a VERY thorough job producing automated code > > conversion scripts that will do most of the NS-to-OpenStep (OS) code > > conversion. Andrew Stone, Stone Design, has produced a nice guide that > > chronicles his experiences with the OS 4.0 PR1 conversion process with an > > extensive problem/solution section. > > Is this guide avaliable somewhere? I'm sorry I mentioned this :-) I asked Andrew about this guide. He prefers to wait until the OpenStep 4.0 release to release the guide he's written, probably because of the non-disclosure agreement that prerelease sites are governed by. I understand that NeXT has distributed this guide to some prerelease sites. If you are a prerelease site, then contact your NeXT representative about receiving this guide. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
Newsgroups: comp.sys.next.programmer From: belka@il.us.swissbank.com (Andy Belk) Subject: Re: source control software suggestion please? Message-ID: <1996Jan6.153717.14215@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> Date: Sat, 6 Jan 1996 15:37:17 GMT In article <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> pfkeb@hpkaon.SLAC.Stanford.EDU (Paul F. Kunz) writes: > >>>>> On 5 Jan 1996 19:38:58 GMT, dave james <djames@isl.com> said: > In article <4cjuoi$nct@bronwyn.isl.com> dave james <djames@isl.com> writes: > > > Hi there, we have a development initiative under way and would like > > to hear any recommendations that people have as to source control > > techniques or software. Thanks in advance > > Native CVS works fine for NeXTSTEP development. CVS is free and > available with GNU software. Fancy GUI front-ends to CVS exist for > NeXTSTEP at a price; freely available front-end based in Tcl/Tk are > also available. But CVS is so easy to use who needs a GUI front-end? I wouldn't want to disagree with Paul but we use DevMan here and it is excellent. Somewhat more powerful than CVS and a nice GUI, and very useful for large projects. It's by VNP Software (Luke Blanshard - luke@vnp.com - or info@vnp.com). I have no affiliation with them, etc. I just use it. Alternatively, ... Bill Bumgarner has made the mods to CVS for NeXT: here's a mail from the NeXT-Java mailing list: Quad-fat cvs/rcs binaries [with man, info, and postscript docs] are available at: ftp://ftp.thoughtport.com/pub/next/tools/ ftp://ftp.thoughtport.com/pub/next/tools/cvs_rcs.b.SHIN-README ftp://ftp.thoughtport.com/pub/next/tools/cvs_rcs.b.SHIN.tar.gz Further reading can be found at: http://ebweb.tuwien.ac.at/gnu-docs/cvs/cvs.1.html [html man page] http://ebweb.tuwien.ac.at/gnu-docs/cvs/mkmodules.1.html [html man page] http://www.loria.fr/~molli/cvs-index.html [awesome summary page] Basic CVS does need some trickery to handle .nib files and the like - I'm not sure if these changes are included in Bill's version - they are available somewhere on the net I believe. As regards techniques (as opposed to tools), you want to have a procedure by which you can freeze the source tree for a release version and continue to work on a new development branch such that if (when) release bugs are reported, you can go back to the original source to fix them (this might be obvious, so excuse me if so). Always keep time-frozen copies of any binaries (libraries, .o's) that you don't have source code for in the repository as well. Have separate repositories and build areas and a responsible individual for doing daily/weekly rebuilds. Be prepared to expend some energy explaining the source code control system to people and on a bit of training. Be paranoid about keeping copies - disk space is cheap. There's lots more but you get the idea. Cheers, Andy Belk
From: daj@ccrma.Stanford.EDU (David Jaffe) Newsgroups: comp.sys.next.programmer Subject: driver kit questions Date: 7 Jan 1996 05:27:14 GMT Organization: CCRMA, Stanford University Message-ID: <4cnlji$iq9@nntp.Stanford.EDU> 2 driver kit questions 1. I've found that when I call the following in a driver: IOGetTimestamp(&ats1); IOSleep(1); IOGetTimestamp(&ats2); the difference between ats2 and ats1 is 10 ms, not 1 ms! Is there really a minimum sleep time of 10 ms? Is there any way around that (I don't want to use IODelay() because I need for other threads to run.) (I'm a little suspicious of IOGetTimestamp(), though, or possibly the long long arithmetic I'm using to convert it to micro-seconds: long long microseconds = ats1/1000; long ms = microseconds; ) 2. I'm finding that calling vm_allocate() too many times in a driver without deallocating the buffers causes the system to freeze. Why? I'm checking the return code. I would hope that once there were no more vm available that vm_allocate() would just return an error instead of causing the system to freeze. Is there some way to tell if it's safe to call vm_allocate() in the driver? (I tried decreasing the port backlog of the reply port so that fewer allocated buffers would exist at the same time, but that only improved the situation somewhat.) Please cc: any reply to david@jaffe.com--I don't have convenient access to Usenet. thanks David Jaffe Kensington, CA david@jaffe.com
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: source control software suggestion please? In-Reply-To: dave james's message of 5 Jan 1996 19:38:58 GMT Message-ID: <RDL.96Jan7005113@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4cjuoi$nct@bronwyn.isl.com> Date: Sun, 7 Jan 1996 05:51:13 GMT DevMan is the revision control system that most sites use. The Services interface is great, the CLI is good but the 1.2 GUI is mediocre. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4cjuoi$nct@bronwyn.isl.com> dave james <djames@isl.com> writes: Xref: world comp.sys.next.programmer:26078 Path: world!news.kei.com!newsfeed.internetmci.com!tank.news.pipex.net!pipex!lade.news.pipex.net!pipex!bronwyn.isl.com!usenet From: dave james <djames@isl.com> Newsgroups: comp.sys.next.programmer Date: 5 Jan 1996 19:38:58 GMT Organization: Internet Systems Ltd Lines: 6 NNTP-Posting-Host: davejames.isl.com Hi there, we have a developement initiative under way and would like to hear any reccomendations that people have as to source control techniques or software. Thanks in advance
Newsgroups: comp.sys.next.programmer From: js@euler.hnv.icem.de (Juergen Sell) Subject: HowTo exactly One Terminal.app window for PB/gdb Debug? Message-ID: <DKo2s2.3Au@euler.hnv.icem.de> Sender: news@euler.hnv.icem.de Organization: Ink Unknown Date: Thu, 4 Jan 1996 17:19:13 GMT When running gdb from PB's Debug facility, I get a new instance of a Terminal.app window with its own gdb every single time I press the 'Debug' in PB. Over time this leaves a ton of unused windows & gdb instances around. How can I avoid that and make 'Debug' reuse the existing Terminal window & gdb? Or how can I at least setup gdb so it closes its Terminal.app window when I quit gdb? Juergen --- Fon +49 511 4406-88 NeXTMail welcome No Mime Fax +49 511 4406-17 = What time do we live in when revolution reminds us of soap powder, = when spontaneity and freedom get associated with instant coffee, = when a politician's idea of social change is changing names = when a country posing as super know-how factory cuts expenses on education?
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Developing for NEXTSTEP vs OPENSTEP---a critical business decision Date: 7 Jan 1996 16:43:39 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4cot7r$rbk@lion.embl-heidelberg.de> References: <4cc06i$a5v@xenon.stgt.sub.org> In article <4cc06i$a5v@xenon.stgt.sub.org> "Ingo Feulner" <ifeulner@xenon.stgt.sub.org> writes: > > There are more problems: OpenStep for MachOS (aka NEXTSTEP 4.0) has some > (really nice and useful classes, such as the new Text class) which are not > yet in > the OpenStep spec. And so these objects are not in OpenStep for Solaris > (which is based/developed from NEXTSTEP 3.3). > > Because of this, we have to ask us one more question: Which OpenStep should > I use??? > > (For me: I'll take the original) These NS4.0 specific extensions (the text stuff is really cool!) should be part of both MiscKit 2.x and GNUStep as additional library.... Q: When? A: God knows! :-( -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387524 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: source control software suggestion please? Date: 7 Jan 1996 17:11:58 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4cousu$rnp@lion.embl-heidelberg.de> References: <1996Jan6.153717.14215@il.us.swissbank.com> In article <1996Jan6.153717.14215@il.us.swissbank.com> belka@il.us.swissbank.com (Andy Belk) writes: > Alternatively, ... > > Bill Bumgarner has made the mods to CVS for NeXT: here's a mail from the > NeXT-Java mailing list: > > Quad-fat cvs/rcs binaries [with man, info, and postscript docs] are > available at: > [...] > Basic CVS does need some trickery to handle .nib files and the like - I'm > not sure if these changes are included in Bill's version - they are > available somewhere on the net I believe. Is it the latest CVS (GNU) version? Last time I played with modified cvs for NS there was still no support for remote repository and this is one VERY important feature. Thanks -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387524 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: source control software suggestion please? Date: 7 Jan 1996 15:42:55 GMT Organization: Genoa Software Systems Message-ID: <4copm0$6qu@saturn.genoa.com> References: <4cjuoi$nct@bronwyn.isl.com> <RDL.96Jan7005113@world.std.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <RDL.96Jan7005113@world.std.com> Robert La Ferla wrote: > DevMan is the revision control system that most sites use. The Services > interface is great, the CLI is good but the 1.2 GUI is mediocre. About the DevMan GUI It is a little different than most NeXT apps, and it could definitely be improved somewhat, but I respectfully disagree with your characterization of the GUI as mediocre. The GUI packs alot of information into a small space, and lets you see and manipulate multiple versions of files, branches, and repositories, along with their version numbers, dates etc all within a single window. It takes some getting used to, but you can work pretty quickly in a complex environment once you understand it. If it had a more "flowery" GUI like project builder, you could spend alot more time navigating and have a harder time seeing the big picture (like differences between multiple branches). Don't get me wrong, DevMan could use lots of enhancements to the GUI, to the ability to generate reports, a public API, and more. But all in all, I think they did an admirable job of building a useful CM system for a rather non-traditional environment. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: seeking flex & bison for NEXTSTEP? Date: 7 Jan 1996 23:21:59 GMT Organization: my own Message-ID: <4cpkin$4t4@gate.seicom.net> References: <4cjdrl$aep@ra.nrl.navy.mil> yiannis@prologos.nrl.navy.mil (John Michopoulos) wrote: > Where can I find flex & bison compiled for NEXTSTEP? ftp.cs.orst.edu does > not have them and the ones I picked up from aeneas.mit.edu gnu archive do > not compile well! So, before I go into solving this problems I though I > should look for compiled versions. Well, gnu flex-2.5.2 compiles without an error here on NS 3.3 / Dev. 3.2. I think bison1.24 works too (at least I have a binary in /usr/local/bin, so I must have compiled it some time ago...) Could you please specify your problem? - Frank --- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP & PostScript Guy "In cantonese C++ is called C ga ga"
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP Developer available Date: 8 Jan 1996 02:20:02 GMT Organization: Norden 1 Communications Message-ID: <4cpv0i$om0@tofu.alt.net> Omni Search has an experienced NEXTSTEP developer available for contract work. Objective C, Sybase, EOF, and design experience. Four years with NEXT, 15 years overall experience. Reasonable rates. If interested, please contact. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: marks@soli.inav.net (Mark Strand) Newsgroups: comp.sys.next.programmer Subject: Using image + text in a browser cell Date: 8 Jan 1996 02:28:20 GMT Organization: INAV Message-ID: <4cpvg4$hdn@composer.inav.net> Mime-Version: 1.0 I'm wondering how to do the following: Display in a browser, some entries with a check mark in front of them, some without. What I want is similar to what you get in ProjectBuilder's Files view. Under interfaces, your *.nib files have the little flag preceding them. Also, Headers has a space before the *.h files in the second column, as if they're allowing for a graphic of some sort. Anyhow, this is what I want, for my browser entries (it's a simple, 1-column browser) to allow the option of the check.tiff in front of the text. Is this hard to do? Does it involve setImage/setAltImage, where altImage is the check? I'm not sure where to start, and the docs didn't really go deeply into this specific scenario. Thanks in advance! - Mark (Loving IB 3.3!)
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Re: TeX3.14? web2c-2.61? flex-2.5? (Small world...) Date: 8 Jan 1996 02:55:29 GMT Organization: U.C. Berkeley Math. Department. Distribution: na Message-ID: <4cq131$oka@agate.berkeley.edu> References: <4cl8hf$qc3@agate.berkeley.edu> Thanks to all for their help. I have a working TeX3.1415 (6.1) using the kpathsea version of web2c. The problems I experienced were: flex 2.5 wouldn't build without some very minor tampering with the source vs. what configure chooses. (I guess I set up my installation of gcc strangely, but it was stock GNU.) The changes to the source can be made by just following the error messages to the definitions that cause them. sed 3.0 (necessary to run the configure scripts for the new kpathsea) looks for malloc.h in /usr/include/malloc.h; this is hardwired and should be changed to /usr/include/bsd/malloc.h. That done, it builds. kpathsea won't build unless the "hypot" extern is removed from one of the header files. (It conflicts with hypot in math.h and is never used...) Other than that, the patches were sufficient, along with the web2c.next.help (or something like that) file on ftp.cs.umb.edu. What I didn't know (but learned after the fact...) was that I *already* had the web2c files I needed, tangle and weave, in order to build TeX, so the day I spent making things work and moving files was wasted... All was not lost, though, as he new TeX and kpathsea with the associated directory heirarcy (/usr/lib/texmf/[etc]) works nicely, and the stock memory of 15000 strings (nearly triple that of the original NeXTTeX) is sufficient for more complicated TeXing. Users of LaTeX2e, amstex.sty, and P. Taylor's "diagrams.tex" package will find that the three combine to nearly fill NeXTTeX 3.1's memory all by themselves. (For anyone who is going to try to reproduce my results, be sure to save your old pk files! You can reuse them by copying them into the appropriate new directories in the kpathsea heirarchy. Look at the Make_Fonts script to find out which resolutions commonly go with which mode and then copy the files.) Thanks again. Paul Brown
From: mike@hobbs.chem.usu.edu (Michael Emmel) Newsgroups: comp.sys.next.programmer Subject: _NXPoint?? Date: 8 Jan 1996 04:12:07 GMT Organization: SouthWind Internet Access, Inc. Distribution: world Message-ID: <4cq5in$q0p@opal.southwind.net> I found this structure in dpsclient/event.h typedef struct _NXPoint { /* point */ NXCoord x, y; } NXPoint; It look suspicoulsy like the complied version of "NXPoint". The way I understand it C puts and underscore in front of sruct data and function names when compiling. So compiled NXPoint->_NXPoint->__NXPoint ??? I'm not the worlds best C programmer whay do they do this? I suspect is has something to do with calling routines written in different high level languages ? Its rather common in Nexts headers and I don't understand it. Thanks mike@hobbs.chem.usu.edu
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 8 Jan 1996 05:15:12 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4cq990$l0k@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Computer, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 188+ ISV company pages - 433+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Additionally there is a Mail Server available. You can get information on using the mail server at ns-products@stepwise.com Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Computer, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://ftp.cs.orst.edu: The main site for North American submissions ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Using image + text in a browser cell Date: Mon, 8 Jan 1996 07:26:37 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Jan8.072637.3625@seer.demon.co.uk> References: <4cpvg4$hdn@composer.inav.net> In article <4cpvg4$hdn@composer.inav.net> marks@soli.inav.net (Mark Strand) writes: > I'm wondering how to do the following: Display in a browser, some entries > with a check mark in front of them, some without. Take a look at the MiniExample ScrollDoodScroll. This implements a custom BrowserCell class with this behaviour. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul (under construction)
From: fxg@merry.imib.rwth-aachen.de (Felix H. Gatzemeier) Newsgroups: comp.sys.next.programmer Subject: messaging Previwer, like the PrintPanel Date: 08 Jan 1996 11:37:14 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Message-ID: <FXG.96Jan8123715@merry.imib.rwth-aachen.de> I want to extend TeXview.app's Services by a 'render to EPS' service that takes ASCII data, texes and dvipses ist and finally displaying it in Preview, returning naught. This Service could be used by any app that offers ASCII data, not only Apps that also take EPS. The advantage of having Preview display the data is the free scaling and small windows that Preview offers. Other alternatives are displaying the result in TeXview itself (sounds reasonable...) or a TickleService (which would have to message Preview again). The Problem is this: How do I message Preview so tat it takes EPS from my app without leaving temporary files around? Currently, I leave temp.eps and do system("open temp.eps"), but this is clearly not satisfactory. Another Question: Is there any work on TeXview being done by anyone else? Is 3.0 still the up-to-date Version? Thanks -- Felix (fxg@(Pool.Informatik|imib).RWTH-Aachen.de NeXT-Mail ok, but slow)
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: _NXPoint?? Date: 8 Jan 1996 15:47:57 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4crebd$3dr@news.its.com> References: <4cq5in$q0p@opal.southwind.net> mike@hobbs.chem.usu.edu (Michael Emmel) wrote: > > I found this structure in dpsclient/event.h > > typedef struct _NXPoint { /* point */ > NXCoord x, y; > } NXPoint; > > > It look suspicoulsy like the complied version of "NXPoint". This typedef defines "struct _NXPoint" as something with 2 NXCoords, and lets you refer to this simply as an NXPoint. "struct _NXPoint" and "NXPoint" are indentical. > The way I understand it C puts and underscore in front of sruct data and > function names when compiling. The compiler system generally prefixes symbols from user code with an underscore, true. It's completely not relevant to what was going on here, as the typedef describes the type of a variable, not a variable itself. If you have NXPoint p; ....in your code, you'll find a symbol named "_p" from that, not a symbol named "_NXPoint". (Unless you compile with debugging information, in which case the system will also describe the layout and name of the structure, but you do not need that information to run the program.) > So compiled NXPoint->_NXPoint->__NXPoint ??? No, because the typedef defines a type of a variable, but does not declare one. > I'm not the worlds best C programmer whay do they do this? They prefix all user symbols with "_" so that the system designers can freely use symbols without an underscore for system initialization stuff in the crt0.o startup code without ever worrying about a namespace collision with user symbols. By the way, completely forget about this underscore business unless you're working with symbol tables or mucking with the crt0.o initialization stuff. You'll just confuse yourself. They do "typedef struct _foo { .... } foo" because it makes it clearer what's going on. It's a very standard convention. -Chuck PS: If this isn't clear, go get the K&R and read up on the difference between a declaration and a type specification. Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: dave james <djames@isl.com> Newsgroups: comp.sys.next.programmer Subject: hp server for compiles in 25 person teams Date: 8 Jan 1996 18:29:54 GMT Organization: Internet Systems Ltd Message-ID: <4crnr2$fg@bronwyn.isl.com> Hi all, I was wondering if anyone is using a Hpserver as a compile and source straoge area. If they are can we get some guidance as to what config you are using and what you would reccomend. We are looking into supporting a 25 person environment and would expect a high amount of Openstep development activity on the server. Any advice would be helpful.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Tue, 9 Jan 96 14:06:08 +0100 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9601091306.AA07587@flexus> Subject: Re: _NXPoint?? Michael Emmel <mike@hobbs.chem.usu.edu> writes: > I suspect is has something to do with calling routines > written in different high level languages ? about NeXT's style ``typedef struct _name{ /*blah blah*/ } name;''. This is just making things difficult. I always code it as ``typedef struct name{ /*blah blah*/ } name;''. Perhaps someone didn't realise there was no danger for conflicts (because the namespace for struct is distinct)? But let's assume that they were duly cautious in the face of C++'s changes to the namespace boundaries (I'm not very familiar with C++). It has nothing to do with how the symbols are encoded in the code files, IMHO. Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be (NeXTmail), Flanders, Belgium If I don't answer: my mail relay can't handle !, % or .uucp, I think *** The year 2000 will be the last year of the 20th century. ***
From: mpaque@aol.com (Mpaque) Newsgroups: comp.sys.next.programmer Subject: Re: Making Apps that run "nice" Date: 13 Jan 1996 02:12:24 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4d7m0o$cev@newsbf02.news.aol.com> References: <4d4fbi$47n@news.duke.edu> In article <4d4fbi$47n@news.duke.edu>, altenber@acpub.duke.edu (Lee Altenberg) writes: >I wrote an app that uses the 3D kit, and is very CPU intensive. >It causes the whole workspace to slow to a crawl. Is there any >way I can configure an app so that it always runs nice, say, at >a priority equivalent to >localhost> renice +20 <app PID> Use 'thread_priority()' to drop the priority of your CPU intensive thread, or task_priority() to drop the priority of all threads in your task: task_priority( task_self(), 0, TRUE ); The second parameter is the priority, running from 0 to 31. Non-interactive processes (UNIX commands, etc) are created at priority 10 by default. Active apps run at priority 16. Setting a priority of 0 means that your app will run only when nothing else is going on. The third parameter, TRUE, means that you want the call to affect all threads that already exist in the app, not just new threads. Note that without special privileges, your app cannt raise it's priority above the base priority it was created at. This isn't a problem for your purposes. Mike Paquette -- I don't speak for NeXT Software, Inc., and they don't speak for me. Fair deal...
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: source control software suggestion please? Date: 10 Jan 1996 03:20:57 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4cvbap$72@chinx4.thoughtport.net> References: <1996Jan6.153717.14215@il.us.swissbank.com> <x7ent9brvg.fsf@tom.es.ele.tue.nl> In article <4cousu$rnp@lion.embl-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) writes: > Basic CVS does need some trickery to handle .nib files and the like - I'm > not sure if these changes are included in Bill's version - they are > available somewhere on the net I believe. Is it the latest CVS (GNU) version? Last time I played with modified cvs for NS there was still no support for remote repository and this is one VERY important feature. # CVS, current version 1.6, serves remote repositories (which is # _extremely_ useful) and has a notion of wrappers---directories # to be treated as files. --Tiggr The quad-fat CVS that is available from.... ftp://ftp.thoughtport.com/pub/next/tools contains all of the extensions necessary to handle many of the common NeXT file types including .nib, etc. As well, it is trivial to add new types. The package seems to work well [I'm using it for the port of Java], though I have not specifically tested the wrappers against NIB files. b.bum
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Suddenly: "Make: Don't know how to make project. Stop." In-Reply-To: altenber@acpub.duke.edu's message of 9 Jan 1996 19:15:20 -0500 Message-ID: <RDL.96Jan9233656@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4cv0eo$126@news.duke.edu> Date: Wed, 10 Jan 1996 04:36:56 GMT If it happens in the shell but not in ProjectBuilder, then the cause is that you are not specifying a target such as "app", "bundle", "library", "install", etc.. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4cv0eo$126@news.duke.edu> altenber@acpub.duke.edu (Lee Altenberg) writes: From: altenber@acpub.duke.edu (Lee Altenberg) Newsgroups: comp.sys.next.programmer Date: 9 Jan 1996 19:15:20 -0500 Organization: Duke University, Durham, NC, USA In the middle of working on a project with ProjectBuilder, I go to Build, and suddenly I get the error message: Make: Don't know how to make project. Stop. *** Exit 1 Stop. I can't figure out what is wrong. It happens when I run make manually in a shell with "make -f Makefile". What are the sorts of things that cause this behavior? Any tips will be GREATLY appreciated. Thanks, Lee Altenberg Maui High Performance Computing Center altenber@mhpcc.edu .
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: source control software suggestion please? In-Reply-To: vdemarco@whatnow.bou.shl.com's message of Tue, 9 Jan 1996 23:23:34 GMT Message-ID: <RDL.96Jan9234002@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> <1996Jan6.153717.14215@il.us.swissbank.com> <4cua3b$slr@shellx.best.com> <m8ivil9cll.fsf@whatnow.bou.shl.com> Date: Wed, 10 Jan 1996 04:40:02 GMT If the new CVS doesn't have a GUI or a Services interface, your developers will take a productivity hit. DevMan is worth the money. I just wish VNP would improve the GUI. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <m8ivil9cll.fsf@whatnow.bou.shl.com> vdemarco@whatnow.bou.shl.com (Vince Demarco) writes: Xref: world comp.sys.next.programmer:26115 Newsgroups: comp.sys.next.programmer Path: world!bloom-beacon.mit.edu!gatech!newsfeed.internetmci.com!in2.uu.net!shlnews!usenet From: vdemarco@whatnow.bou.shl.com (Vince Demarco) X-Nntp-Posting-Host: whatnow.bou.shl.com Lines: 25 Sender: vdemarco@whatnow.bou.shl.com Organization: SHL SystemHouse Inc. X-Newsreader: Gnus v5.1 References: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> <1996Jan6.153717.14215@il.us.swissbank.com> <4cua3b$slr@shellx.best.com> Date: Tue, 9 Jan 1996 23:23:34 GMT Denise Howard (deniseh@shellx.best.com) wrote: >Andy Belk (belka@il.us.swissbank.com) wrote: >: I wouldn't want to disagree with Paul but we use DevMan here and it is >: excellent. Somewhat more powerful than CVS and a nice GUI, and very useful >: for large projects. It's by VNP Software (Luke Blanshard - luke@vnp.com - >: or info@vnp.com). I have no affiliation with them, etc. I just use it. > >Having used both CVS and DevMan, I have to second Andy's comments. CVS >was fine as long as I was in a three-developer environment, but >for the 45-developer environment I'm in now I can't imagine working >without DevMan. It's much easier to track changes both incrementally and >across branches. Nibs can be handled without a care. It's great! I haven't used DevMan at all. But i would suggest taking a look at cvs 1.6.5 (ftp the latest snapshot from ftp://ftp.cyclic.com/pub/cvs) I have used CVS in a large group and it works fine. DevMan is as far as i understand a frontend to rcs. CVS is also a front end to rcs. Version 1.6 has wrappers support so nibs, .rtfd, .diagram etc... can all be handled without any problems vince
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Help with custom view and IB In-Reply-To: klett002@maroon.tc.umn.edu's message of Mon, 8 Jan 1996 21:59:47 GMT Message-ID: <RDL.96Jan9234206@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <DKvuBD.LA0@news.cis.umn.edu> Date: Wed, 10 Jan 1996 04:42:06 GMT When are you sending the message to your customview? If it's not in the awakeFromNib method, you can't guarantee that your outlet has been initialized. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <DKvuBD.LA0@news.cis.umn.edu> klett002@maroon.tc.umn.edu (James P. Klett) writes: Hello, I have a panel with a custom view on it, call it MyCustomTextView. I have an id variable in my controller object call userName and via IB have connected it up to MyCustomTextView. The problem is, when I try to send a message to the custom view via the id userName, I get nothing, due to the fact that userName is NULL! Why is is NULL? It is connected up in IB, I can see it on the screen when the program runs. Does it have something to do with the fact that its a custom view? If so, what do I have to do to make them connected like any other UI Object?? Thanks, JIM ___ ___....-----'---'-----....___ ========================================= ___'---..._______...---'___ (___) _|_|_|_ (___) \\____.-'_.---._'-.____// klett002@maroon.tc.umn.edu
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: source control software suggestion please? In-Reply-To: Alex Blakemore's message of 7 Jan 1996 15:42:55 GMT Message-ID: <RDL.96Jan9234528@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4cjuoi$nct@bronwyn.isl.com> <RDL.96Jan7005113@world.std.com> <4copm0$6qu@saturn.genoa.com> Date: Wed, 10 Jan 1996 04:45:28 GMT I didn't say it was a bad GUI. I said it was mediocre. In any case I like and recommend the product. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4copm0$6qu@saturn.genoa.com> Alex Blakemore <alex@genoa.com> writes: From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Date: 7 Jan 1996 15:42:55 GMT Organization: Genoa Software Systems In <RDL.96Jan7005113@world.std.com> Robert La Ferla wrote: > DevMan is the revision control system that most sites use. The Services > interface is great, the CLI is good but the 1.2 GUI is mediocre. About the DevMan GUI It is a little different than most NeXT apps, and it could definitely be improved somewhat, but I respectfully disagree with your characterization of the GUI as mediocre. The GUI packs alot of information into a small space, and lets you see and manipulate multiple versions of files, branches, and repositories, along with their version numbers, dates etc all within a single window. It takes some getting used to, but you can work pretty quickly in a complex environment once you understand it. If it had a more "flowery" GUI like project builder, you could spend alot more time navigating and have a harder time seeing the big picture (like differences between multiple branches). Don't get me wrong, DevMan could use lots of enhancements to the GUI, to the ability to generate reports, a public API, and more. But all in all, I think they did an admirable job of building a useful CM system for a rather non-traditional environment. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: cmd.h In-Reply-To: Tommy K Hwang's message of Sun, 07 Jan 1996 04:39:20 +0900 Message-ID: <RDL.96Jan9235114@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <30EECFE8.4E24@home.mysolution.com> Date: Wed, 10 Jan 1996 04:51:14 GMT When you remove that #include, what error messages does your compiler complain about? Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <30EECFE8.4E24@home.mysolution.com> Tommy K Hwang <thwang@home.mysolution.com> writes: Xref: world comp.sys.next.programmer:26083 Path: world!news.mtholyoke.edu!uhog.mit.edu!news.mathworks.com!newsfeed.internetmci.com!news.mysolution.com!news From: Tommy K Hwang <thwang@home.mysolution.com> Newsgroups: comp.sys.next.programmer Date: Sun, 07 Jan 1996 04:39:20 +0900 Organization: Internet Solutions Inc. Lines: 4 NNTP-Posting-Host: nt1.mysolution.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 2.0b4 (WinNT; I) Where did cmd.h header went? Did it ever existed for NeXT? I got User Quota manager compiled except for ONE last section that need the cmd.h header. Once that part is done, I can test and release it for NS3.3.
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: driver kit questions In-Reply-To: daj@ccrma.Stanford.EDU's message of 7 Jan 1996 05:27:14 GMT Message-ID: <RDL.96Jan9235614@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4cnlji$iq9@nntp.Stanford.EDU> Date: Wed, 10 Jan 1996 04:56:14 GMT NS is not a real-time OS. When you specify 1 ms as an argument to IOSleep(), you are only guaranteed that the delay will be at least 1 ms. i.e. 1 ms or greater. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4cnlji$iq9@nntp.Stanford.EDU> daj@ccrma.Stanford.EDU (David Jaffe) writes: 1. I've found that when I call the following in a driver: IOGetTimestamp(&ats1); IOSleep(1); IOGetTimestamp(&ats2); the difference between ats2 and ats1 is 10 ms, not 1 ms! Is there really a minimum sleep time of 10 ms? Is there any way around that (I don't want to use IODelay() because I need for other threads to run.) (I'm a little suspicious of IOGetTimestamp(), though, or possibly the long long arithmetic I'm using to convert it to micro-seconds: long long microseconds = ats1/1000; long ms = microseconds; )
Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin From: rdl@world.std.com (Robert La Ferla) Subject: Re: Root In-Reply-To: vazquezr@physics.ucla.edu's message of 9 Jan 1996 05:40:44 GMT Message-ID: <RDL.96Jan9235902@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4csv4s$1n52@saba.info.ucla.edu> Date: Wed, 10 Jan 1996 04:59:02 GMT This really belongs in comp.sys.next.sysadmin. In any case, your /.cshrc is probably set up incorrectly. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4csv4s$1n52@saba.info.ucla.edu> vazquezr@physics.ucla.edu (Rick Vazquez) writes: From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Date: 9 Jan 1996 05:40:44 GMT Organization: University of California, Los Angeles Hello, Can someone please help me. When I used to su root I once had a prompt like localhost# and I could enter commands like fsck right from the command line. Now when I su root, I get localhost> the same old promt as before, and now I have to find fsck in the directory stucure before I can execute the command. Can someone plesae tell me how to fix this??? Rick UCLA Physics
Newsgroups: comp.sys.next.programmer From: rouand@tnt.oleane.com (Jean-Michel Rouand) Subject: Foundation kit and memory management Message-ID: <1996Jan8.094545.409@tnt.oleane.com> Sender: rouand@tnt.oleane.com Organization: Terra Nova Techonologies (France) Date: Mon, 8 Jan 1996 09:45:45 GMT Hi everybody, I have to write an application in which memory allocated is very important. I 'm using for the first time foundation kit with NS 3.3 and it seems that memory is lost every time I reload a List with the same items using a NSMutableArray: [alarmList removeAllObjects]; ... insert 3 items in the list With a List object doing the same, the memory allocated is stable: [alarmList freeObjects]; ... insert 3 items in the list The items are Object subclasses with free and release methods the same. Is foundation kit memory safe ? Can anybody help me please PS: I use Process monitor to see allocated memory. Jean-Michel Terra Nova
Newsgroups: comp.sys.next.programmer From: rouand@tnt.oleane.com (Jean-Michel Rouand) Subject: Foundation kit and memory management Message-ID: <1996Jan8.095651.477@tnt.oleane.com> Sender: rouand@tnt.oleane.com Organization: Terra Nova Techonologies (France) Date: Mon, 8 Jan 1996 09:56:51 GMT Hi everybody, I have to write an application in which memory allocated is very important.I'm using for the first time foundation kit with NS 3.3 and it seems that memory is lost every time I reload a list with the same objacts using a NSMutableArray: [alarmList removeAllObjects]; ... inserting objects Doing the same with the List object from appkit, the memory allocated is stable: [alarmList freeObjects]; ... inserting objects The objects are Appkit objects with free and release methods the same. Is foundation kit memory safe ? Can anybody help me please ? PS: I use Process monitor to see allocated memory Jean-Michel Terra Nova
From: npratt@xmission.xmission.com (Nevin Pratt) Newsgroups: comp.sys.next.programmer Subject: IPT uSHARE Anybody got it? Date: 9 Jan 1996 22:13:49 -0700 Organization: XMission Internet (801 539 0900) Message-ID: <4cvhud$mr9@xmission.xmission.com> IPT used to make a uSHARE product that transformed the black NeXT into a Mac Appleshare / PrintShare server. Is that package still available? Anybody got it? You wanna sell it? Please let me know via email, as I rarely visit these groups anymore. Nevin
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: source control software suggestion please? Date: 11 Jan 1996 12:01:06 +0100 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x720p73si5.fsf@tom.es.ele.tue.nl> References: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> <1996Jan6.153717.14215@il.us.swissbank.com> <4cua3b$slr@shellx.best.com> <m8ivil9cll.fsf@whatnow.bou.shl.com> <RDL.96Jan9234002@world.std.com> In-reply-to: rdl@world.std.com's message of Wed, 10 Jan 1996 04:40:02 GMT In article <RDL.96Jan9234002@world.std.com> rdl@world.std.com (Robert La Ferla) writes: If the new CVS doesn't have a GUI or a Services interface, your developers will take a productivity hit. [DevMan plug removed.] Can you substantiate this claim? --Tiggr
From: leigh@antechinus.cs.uwa.oz.au (Leigh Smith) Newsgroups: comp.sys.next.programmer Subject: Re: Q : Common Lisp under NeXT ? Date: 11 Jan 1996 13:32:01 GMT Organization: The University of Western Australia Distribution: world Message-ID: <4d33gh$5mn@styx.uwa.edu.au> References: <4d03ag$avk@ubszh.fh.zh.ubs.com> In article <4d03ag$avk@ubszh.fh.zh.ubs.com> kjz@wsfh105 (Peter Kohut) writes: > Are there any Common Lisp Interpreters / Compilers for NeXTSTEP > ( Commercial or PD ) ? > > BTW, what's about tools for AI (Expertsystem, Knowledgebase) ? > I used to use (gcl) Gnu Common Lisp which is a descendent from Kyoto Common Lisp, which compiles to C thus producing (huge) binaries. I found it to be very buggy and in the end moved over to using clisp, which uses byte code compilation and is therefore slower but a lot less buggy and actually has a NeXTstep front end to it (not a window manager like CLX or CLIM however) and seems pretty nice. Pointers to current locations of both PD Lisps are at prep.ai.mit.edu. Clisp is at ma2s2.mathematik.uni-karlsruhe.de with a NeXTStep binary (m68k) at http://sayre.sysc.pdx.edu:8001/clisp/binaries. -- Leigh Smith Computer Science, University of Western Australia +61-9-380-1945 leigh@cs.uwa.edu.au (NeXTMail/MIME) "In a world where success means gaining time, thinking has a single but irredeemable fault: it's a waste of time" - J-F. Lyotard
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: source control software suggestion please? Date: 11 Jan 1996 16:15:39 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4d3d3b$6sm@chinx4.thoughtport.net> References: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> <RDL.96Jan9234002@world.std.com> # If the new CVS doesn't have a GUI or a Services interface, your # developers will take a productivity hit. DevMan is worth the # money. I just wish VNP would improve the GUI. I would disagree strongly with this statement; I have used DevMan's command line and GUI quite extensively. The command line blows away the UI in regards to versatility and sheer power-- and it is *extremely* well done [ie; intuitive]. As well, there are certain tasks the user interface does not handle at all well-- take map files, for example. At this point, I only use the user interface when I'm comparing different revisions of specific source files using FileMerge. [btw: i have a tutorial on how to use the devman command line. it is nearly ready to be published [via the web], but since i don't have access to devman from home, i have not been able to polish it off] The command line interface that cvs provides is equally as intuitive and *a lot* more powerful. While devman provides an excellent and intuitive interface to the powerful, but primitive, rcs environment, cvs provides both subsumes the rcs feature set (it uses rcs underneath) and implements a bunch of new features to aid in the development effort. cvs is designed for large scale development efforts involving many developers; to this end, it provides remote repositories, watch lists (ability to monitor who is editing what when and respond automatically), conflict resolution, automatic notification (send an email when a particular change is in place), and others... Since one can pretty much modify the interaction between the source repository and the working environments in any way, cvs can be made to support just about any specialized feature you might need. Just because it doesn't have a GUI [yet] does not make it inferior. BTW: Friday.COM will be releasing a set of extensions to the Make environment and ProjectBuilder that support fully revision dependent software build environments using either cvs or DevMan as the underlying revision control engine [ie; 'i want to build app x using library versions y, z, a']. b.bum
From: flight@mathi.uni-heidelberg.de (Gregor Hoffleit) Newsgroups: comp.sys.next.programmer Subject: GCC version in OpenStep 4.0 for Mach ? Date: 11 Jan 1996 16:39:42 GMT Organization: University of Heidelberg, Germany Message-ID: <4d3ege$f0v@sun0.urz.uni-heidelberg.de> Any idea which GCC version will be used in NEXTSTEP 4.0 ? Just dreaming of gcc 2.7.x for NEXTSTEP/HP and Sparc, so I can finally install gpc and g77 ;-) Gregor -- | Gregor Hoffleit admin MATHInet / contact RhiNO | | MAIL: Mathematisches Institut PHONE: (49)6221 56-5771 | | INF 288, 69120 Heidelberg / Germany FAX: 56-3812 | | EMAIL: flight@mathi.uni-heidelberg.de (NeXTmail, MIME) |
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: A way to find out which pages belongs to which task in detail? Date: 12 Jan 1996 19:46:06 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4d6dpu$r6q@news.its.com> References: <4cscbo$ioc@gate.seicom.net> frank@this.net (Frank M. Siegert) wrote: > So my questions is: Is there a (user-level) way to get on information which > task has which pages allocated in detail? Maybe someone has done such a > tool already? Take a look at '/usr/etc/vmoprint' and '/usr/bin/filemem'. Also, I believe there are published Mach function calls you can use to get that type of information; consult Digitial Librarian. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: adnan@seitz.de (Adnan Mumbasic) Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: OO development system needed Date: Thu, 11 Jan 1996 08:10:47 GMT Organization: Seitz GmbH Message-ID: <4d2ga1$726@sarajevo.seitz.de> Hello, we are searching for a object oriented development system for our new project which is funded by the German goverment. We (90% of us) have decided to use OpenStep, but I am asking you is there any alternative to OpenStep or even a much better system? For any sugesstion thnx a lot Adnan Mumbasic
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: GNU C Compiler for NeXTStep 3.3 (Black Hardware) Date: 12 Jan 1996 19:49:09 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4d6dvl$r6q@news.its.com> References: <30F3FB30.41C67EA6@cc.bellcore.com> Tom Mok <tmok@cc.bellcore.com> wrote: > Does anyone know where I can download a binary version of the > GNU C/C++ compiler for the NeXT Motorola hardware, running > NeXTStep 3.3? > > I have only seen source code versions on the net. I wonder > whether compiled versions exist, which they should. Compiled versions of GCC exist, but they won't do you any good without the header files and libraries to link from. You need to go buy NEXTSTEP/Developer if you want to compile things. (This is either a FAQ or should be in one.) -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: gary@ccrma.stanford.edu (Gary P. Scavone) Newsgroups: comp.sys.next.programmer Subject: Software License Protection without using hostid? Date: 13 Jan 1996 19:06:31 GMT Organization: Stanford University Message-ID: <4d8vrn$84d@nntp.Stanford.EDU> I'm trying to get access to a piece of software that uses a computer's hostid in the license protection string (supposedly this is NeXT's solution). Unfortunately, I have a stand-alone machine without an Ethernet card. Therefore, I have no hostid. While the software protection is really not my problem, not being able to use the software is! Does anyone know of a way that a NeXTStep software developer can incorporate software protection without using hostid?
From: perkins@cps.msu.edu (Stephen J. Perkins) Newsgroups: comp.sys.next.programmer Subject: Re: Software License Protection without using hostid? Date: 13 Jan 1996 19:42:42 GMT Organization: Michigan State University Message-ID: <4d91vi$n2v@msunews.cl.msu.edu> References: <4d8vrn$84d@nntp.Stanford.EDU> In <4d8vrn$84d@nntp.Stanford.EDU> Gary P. Scavone wrote: > I'm trying to get access to a piece of software that uses a computer's hostid > in the license protection string (supposedly this is NeXT's solution). > Unfortunately, I have a stand-alone machine without an Ethernet card. > Therefore, I have no hostid. While the software protection is really not my > problem, not being able to use the software is! > > Does anyone know of a way that a NeXTStep software developer can incorporate > software protection without using hostid? One mechanism, that is not host specific but may be better than nothing is to have a user fill in a registraion panel with information. Then have the application show that information at startup (so you know who registered the product). Futher, have the SW create a checksum from the registration information. The checksum must be provided to you from the client during registration (for instance it can be mailed with the registration info). Then, you incorporate this checksum into a licence key that is returned to the user. As long as the registration information is not changed, then the returned key will match the registration checksum and you can "enable" the program. Any change in registration information (even editing the binary) will cause the key to fail. Again, this is not a host based mechanism. Hope this helps spark some ideas... - Steve -- ============================================================== Stephen J. Perkins | mailto:perkins@cps.msu.edu Dept. of Comp. Science | NeXT, MIME, finger for PGP Michigan State University | NeXT OS 3.3 using PPP-2.3 NeXT PPP-2.2 info at http://www.thoughtport.com:8080/PPP/
Newsgroups: comp.sys.next.programmer From: gps@netcom.com (gregory scandalis) Subject: Hostid's for Intel w/o Ethernet? Message-ID: <gpsDL4wLE.52M@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Sat, 13 Jan 1996 19:24:50 GMT Sender: gps@netcom18.netcom.com We have a next software package here at CCRMA-Stanford (SynthBuilder) that is authorized to run on specific machines via an encrypted key. Part of the encrypted key is the hostid of the machine, which is derived from the ethernet card. We are starting to see examples of Intel HW without ethernet cards, thus without a proper hostid (0x0). Does anyone know if there is a way to get a "hostid" for a machine without an ethernet card, or if there is a way to programatically get to the chip id register in the pentium itself? Thanks in advance for your help, -pat +------------------------+------------------------------------------+ | Gregory Pat Scandalis | "Good taste is the enemy of creativity." | | | - Pablo Picasso | +========================+==========================================+ | | | *** All Mail forwards to the Netcom *** | | | | CCRMA gps@ccrma.stanford.edu | | Exploratorium gps@exploratorium.edu | | Netcom gps@netcom.com | +-------------------------------------------------------------------+ -- +------------------------+------------------------------------------+ | Gregory Pat Scandalis | "Good taste is the enemy of creativity." | | | - Pablo Picasso | +========================+==========================================+ | | | *** All Mail forwards to the Netcom *** | | | | CCRMA gps@ccrma.stanford.edu | | Exploratorium gps@exploratorium.edu | | Netcom gps@netcom.com | +-------------------------------------------------------------------+
From: ywu@plato.sky.bdm.com (Yibing Wu) Newsgroups: comp.sys.next.programmer,gnu.emacs.help,comp.sys.next.software Subject: Has anyone been using Emacs on NeXT? Date: Sun, 14 Jan 1996 00:44:32 GMT Organization: BDM International, Inc. Message-ID: <4d9c15$bsl@news.mcl.bdm.com> Hi, there: I have a GNU Emacs 19.28.1 on NeXT(the one with a bufflo icon), and have made it the default editor for all the .m .h files---it works well so far except that once selected, the high-light over the code NEVER goes away. For example, when the project builder tells me that line 200 at file foo.m, there is a type mismatch, and I double-click this message, emacs will be launched automatically with file "foo.m" loaded in and entire line 200 high-lighted---That is great, BUT, how can I make this "selection" go away? Since the high-light's color is darker than that of the cursor, so sometimes, I had hard time finding my cursor! I'd also appreciate if someone can tell me how to change color for the selection and cursor---fortunately, I've figured out how to do that to the foreground and background. Thanks very much! Yibing Wu
From: Carl Edman <cedman@princeton.edu> Newsgroups: comp.sys.next.programmer,gnu.emacs.help,comp.sys.next.software Subject: Re: Has anyone been using Emacs on NeXT? Date: 13 Jan 1996 23:43:31 GMT Organization: Princeton University Message-ID: <4d9g33$m56@cnn.Princeton.EDU> References: <4d9c15$bsl@news.mcl.bdm.com> In <4d9c15$bsl@news.mcl.bdm.com> Yibing Wu wrote: > I have a GNU Emacs 19.28.1 on NeXT(the one with a bufflo icon), That is a gnu, not a buffalo ! Or at least so the artist, Joe Reiss, tells me. (BTW, where are you, Joe ? Your old email address disappeared.) > and > have made it the default editor for all the .m .h files---it works > well so far except that once selected, the high-light over the code > NEVER goes away. That is a problem which was fixed in the latest source release, 4.12, which you can find on lynx.ps.uci.edu. In that version, modifying the buffer will remove the line highlighting. > I'd also appreciate if someone can tell me how to change color for the > selection and cursor---fortunately, I've figured out how to do that to > the foreground and background. The easiest way to change the color of anything emacs is to open a color panel and drag the color over the object which you want to change. That works for the cursor and all faces. To change the background color, shift-drag. Finally to make your changes permanent, use the "Info/Save Preferences" menu item. No need ever to mess with dwrites. Carl Edman
From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Subject: Help with compliling a program Date: 14 Jan 1996 01:48:27 GMT Organization: University of California, Los Angeles Message-ID: <4d9ndb$19ea@saba.info.ucla.edu> Can someone please help I am trying to compile a program under Xwindows for the NeXT and I get the following error.... localhost> xmkmf imake -DUseInstalled -I/usr/X11R6/lib/X11/config localhost> make /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc rm -f findtypes /bin/cc -o findtypes -O -L/usr/X11R6/lib findtypes.o ./findtypes Analyzing architecture, to determine types...done. /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc rm -f buildmap /bin/cc -o buildmap -O -L/usr/X11R6/lib buildmap.o debug.o ./buildmap World.ppm World.risk Countries.risk Image is [797x404]. Image reportedly has 255 colors. Compressing... 100% ...Done! There are 321988 pixels, I compressed 321988 of them, Excellent! Working on countries file. Working on country 41, [45x88] Construction of Map and Country files completed. rm -f xfrisk /bin/cc -o xfrisk -O -L/usr/X11R6/lib client.o network.o gui.o callbacks.o utils.o dice.o cards.o game.o co /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc /bin/cc -O -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DMAPFILE=\"/usc rm -f friskserver /bin/cc -o friskserver -O -L/usr/X11R6/lib server.o network.o deck.o riskgame.o debug.o server-main.o rm -f libfrisk.a ar clq libfrisk.a network.o client.o utils.o dice.o game.o riskgame.o debug.o client-main.o ranlib libfrisk.a ranlib: file: libfrisk.a(debug.o) has no symbols localhost> su root Password: # pppdown # exit localhost> cd /private/tmp/O_xconq-7.0_1/xconq-7.0.1 localhost> ls Ad2c/ Makefile.in config.guess* curses/ lib-x11/ test/ COPYING NEWS config.host* doc/ libcurses/ x11/ ChangeLog README config.sub* install.sh* mac/ INSTALL SelFile/ configure* kernel/ misc/ Imakefile config/ configure.in lib/ move-if-change* localhost> configure Configuring for a m68k-next-nextstep3 host. Created "Makefile" in /private/tmp/O_xconq-7.0_1/xconq-7.0.1 using "./config/mh-empty" localhost> xmkmf mv -f Makefile Makefile.bak imake -DUseInstalled -I/usr/X11R6/lib/X11/config localhost> make Makefiles making Makefiles in kernel... mv Makefile Makefile.bak making Makefiles in SelFile... mv Makefile Makefile.bak making Makefiles in x11... mv Makefile Makefile.bak making Makefiles in curses... mv Makefile Makefile.bak localhost> make making all in ./kernel... /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc rm -f libconq.a ar clq libconq.a init.o cmdline.o ai.o mplayer.o actions.o combat.o run.o sco ranlib libconq.a /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc /bin/cc -g -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 -DUNc rm -f libconqlow.a ar clq libconqlow.a imf.o lisp.o util.o unix.o ranlib libconqlow.a making all in ./SelFile... rm -f SelFile.o /bin/cc -c -g -I. -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 SelFilec rm -f Path.o /bin/cc -c -g -I. -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 Path.c rm -f Dir.o /bin/cc -c -g -I. -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 Dir.c rm -f Draw.o /bin/cc -c -g -I. -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTO=3 Draw.c rm -f libXgnu.a ar clq libXgnu.a SelFile.o Path.o Dir.o Draw.o ranlib libXgnu.a making all in ./x11... /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc xconq.c: In function `update_action_result_display': xconq.c:503: warning: assignment makes pointer from integer without a cast /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc /bin/cc -g -I../x11 -I../kernel -I/usr/X11R6/include -DX_LOCALE -DNeXT -DX_NOT_POSIX -DX_USEBFUNCS -DFUNCPROTc rm -f xconq /bin/cc -o xconq -g -L/usr/X11R6/lib xconq.o xinit.o xmap.o xdraw.o xcmd.o xprint.o ld: Undefined symbols: _strdup *** Exit 1 Stop. *** Exit 1 Stop. Can someone please help. Rick UCLA Physics
From: cwolf@wolfware.com (Christopher Wolf) Newsgroups: comp.sys.next.programmer Subject: Re: Software License Protection without using hostid? Date: 14 Jan 1996 05:58:29 GMT Organization: WolfWare (http://www.wolfware.com/) Message-ID: <4da625$ra9@shellx.best.com> References: <4d8vrn$84d@nntp.Stanford.EDU> In-Reply-To: <4d8vrn$84d@nntp.Stanford.EDU> On 01/13/96, Gary P. Scavone wrote: >Does anyone know of a way that a NeXTStep software developer can >incorporate software protection without using hostid? For most consumer PC hardware unqiue and unchangable machine specific host ids are not available. An alternative solution is to use a license key based protection scheme. When a user orders the application the software vendor creates a unique license key which contains information identifying the software purchaser plus a secure hash of this data. Before running, the program checks for the existence of this keyfile and checks the hash to make sure the identifying data was not altered. The identifying data is then prominently displayed either at start-up time or in the info panel or in some other way to discourage illegal distribution of key files. In this same way it's possible to have the application only function in a limitted demonstration/evaluation mode if a valid license key is not detected. Of course it's not foolproof since someone could reverse engineer the code to figure out how the secure hash is generated but if they're going to go through that much effort they could just as easily patch the binary to remove the check for any kind of hostid either. -- Christopher Wolf / WolfWare cwolf@wolfware.com (NeXTmail & MIME accepted) For information about the NewsFlash newsreader for NeXTSTEP check out WolfWare's home page: http://www.wolfware.com/
From: jreiss@discus.ise.vt.edu (Joe Reiss) Newsgroups: comp.sys.next.programmer,gnu.emacs.help,comp.sys.next.software Subject: Re: Has anyone been using Emacs on NeXT? Date: 14 Jan 1996 14:06:52 GMT Organization: Virginia Tech, Blacksburg, VA Message-ID: <4db2ls$m3g@solaris.cc.vt.edu> References: <4d9c15$bsl@news.mcl.bdm.com> <4d9g33$m56@cnn.Princeton.EDU> Carl Edman <cedman@princeton.edu> wrote... >In <4d9c15$bsl@news.mcl.bdm.com> Yibing Wu wrote: >> I have a GNU Emacs 19.28.1 on NeXT(the one with a bufflo icon), > >That is a gnu, not a buffalo ! Or at least so the artist, Joe Reiss, tells >me. There seem to be a lot of people who don't believe that the Emacs icon really is a Gnu. So I'm hereby presenting my proof. Launch Webster.app. Then define "gnu". The picture that appears is exactly the one on which I based the Emacs icon. So if it doesn't look sufficiently gnu-ish, blame the folks at Webster, not me :-) Joe -- | NeXTMail OK! | + Then we're dead. | | ________ | + I've been dead before. | | | |__) | ======================================================== | | (_|OE| \EISS | - Capt. Scott & Capt. Spock, "Star Trek VI" |
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Help Key on HP Workstations Date: 14 Jan 1996 15:02:43 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4db5uj$fga@lion.embl-heidelberg.de> References: <4ct8cd$mgu@news.sns-felb.debis.de> In article <4ct8cd$mgu@news.sns-felb.debis.de> franke@boehme.dbag.ulm.DaimlerBenz.COM (Juergen Franke) writes: > On the intel and m68 computers pushing the Help Key changes the cursur into > a ?, but it doesn't work on my hp station. How can I manage it? <F1> ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387524 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: jamesr@pennave (James Robertson) Newsgroups: comp.sys.next.programmer Subject: DO proxy uniquing problem Date: 13 Jan 1996 00:24:37 GMT Organization: WilTel Message-ID: <4d6u45$1cs@gateway.wiltel.com> References: <4ck87o$18qs@saba.info.ucla.edu> Any DO gurus out there? I'm having a problem with uniquing of proxies. I'm getting multiple proxies to the same remote object. At first, I thought my problems might be related to working in hybrid NX/NS environment. But I wrote small test applications in pure NX with same results. Here's what it looks like... Server registers itself and runs connection from appkit: connection = [NXConnection registerRoot: self withName: "Foo"]; [connection runFromAppKit]; Client gets proxy to server, gets connection for proxy and runs connection from appkit. Client also calls the server, and server stores a proxy to the client: proxyToServer = [NXConnection connectToName: "Foo"]; connection = [proxyToServer connectionForProxy]; [proxyToServer clientCalling: self]; [connection runFromAppKit]; My Server and Client run OK. They can ping each other, unsolicited, in either direction. But if the server sends the client a message: [proxyToClient heyBuddy: self]; the proxy which the client receives as the argument to -heyBuddy: will have a different address than proxyToServer which was originally created when the client connected to the server. I was expecting DO to unique the proxies, and that they would be identical. Any ideas? Thanks, - Jim Robertson
From: Jim Robertson Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 13 Jan 1996 00:31:52 GMT Organization: WilTel Message-ID: <4d6uho$1cs@gateway.wiltel.com> References: <4ck87o$18qs@saba.info.ucla.edu> <4d6u45$1cs@gateway.wiltel.com> Regarding my previous posting... My email is james.robertson@wcom.com (This is the first time I have posted using Alexandra, and I messed up the header.)
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Help with compliling a program Date: Sun, 14 Jan 1996 22:26:40 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Jan14.222640.24824@seer.demon.co.uk> References: <4d9ndb$19ea@saba.info.ucla.edu> In article <4d9ndb$19ea@saba.info.ucla.edu> vazquezr@physics.ucla.edu (Rick Vazquez) writes: > Can someone please help I am trying to compile a program under Xwindows for > the NeXT and I get the following error.... [incredibly and unnecessarily long log removed] > ld: Undefined symbols: > _strdup I find a suitable header file and insert: #define strdup(x) NXCopyStringBuffer(x) Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul (under construction)
Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer From: benst@stuyts.nl (Ben Stuyts) Subject: Anyone got mgetty+sendfax running on Black NS3.3p1? Message-ID: <DL74HJ.GG0@stuyts.nl> Organization: Stuyts Engineering Haarlem BV Date: Mon, 15 Jan 1996 00:10:31 GMT Dear net, I am trying to use mgetty+sendfax 0.98 on my Black cube with NS 3.3p1. I have the following problems: 1. I need to run it on /dev/cufa, because using /dev/ttydfa blocks the port for outgoing calls like uucico. However, using cufa causes hangups from data calls not to be recognized, which leaves the processes on that port running after a carrier loss. 2. It compiles only with -posix. Because of the file append bug in the NeXT posix support, I cannot look at the mgetty log files (with a tail -f logfile for example) without triggering this bug. Even worse, it seems that incoming data calls 'inherit' this problem. So I can't even use UUCon to check my uucp log files, because they will be filled with zeroes. 3. I tried to convert mgetty to use BSD style sgtty support. Got it running, but I ran into another problem: I cannot use /dev/ttydfa, because when mgetty opens this port, it blocks, waiting for a carrier. So it cannot talk to the modem, even not telling it to answer the phone. Apparantly this works differently under posix. If I use /dev/cufa, I have the problem of not getting SIGHUP's anymore. Does anybody know a solution to this? AFAIK there is no workaround for the append bug, right? Thanks, Ben
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 15 Jan 1996 05:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4dcnsu$m5@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Computer, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 188+ ISV company pages - 433+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Additionally there is a Mail Server available. You can get information on using the mail server at ns-products@stepwise.com Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Computer, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://ftp.cs.orst.edu: The main site for North American submissions ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: source control software suggestion please? Date: 15 Jan 1996 06:42:21 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4dct0d$845@chinx4.thoughtport.net> References: <PFKEB.96Jan5212107@hpkaon.SLAC.Stanford.EDU> <RDL.96Jan9234002@world.std.com> <RDL.96Jan12130542@world.std.com> cvs really shines in the area of multiple developer concurrent development. It takes a slightly different tack in that, by default, all developers can work on any source file simultaneously with no locking. Upon checking a source file into the repository, cvs guides the user through a conflict resolution process. Of course, no revision control system would be worth using without locking-- cvs has that as well. You can enable and disable read-only check-outs of files (just like a lock in rcs) and you can enable and disable notifications on edit, unedit, and/or commit events. As well, any of the developers can view who is currently watching a particular file... [btw: i have a bare-bones cvs UI up and running-- supports all the various services that have been mentioned as convenient for use with DevMan. i'm adding collection inspection and such now. requires the foundation kit.] b.bum
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <Schroeer.Stefan@rz.ruhr-uni-bochum.de> Message-ID: <9601150626.AA00225@localhost> Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable From: Stefan =?iso-8859-1?Q?B=F6hringer?= <Schroeer.Stefan@rz.ruhr-uni-bochum.de> Date: Thu, 11 Jan 96 18:02:17 +0100 Subject: HPDeskwriter 600 (nearly) works with NSwhite Hi all, astonishingly I got the HPDeskwriter to work with the HPDriver = (HPDJ_for3.0 I think) from the archives. I use the Mux-driver to = achive the 57600bps even over a 15450. Anything works fine except = that sometimes graphic trash results. Since reducing the = (pcl-)compresseion rate helped I assume that an internal buffer = overflow of the printer is the problem. Since the apple-serials = don't support hardware handshaking, there should be some software = handshaking. So does anybody know which kind of handshaking is involved with = Appleprinters? (If anybody is interested in the (few) changes to the source of = the driver at the moment I can mail them to him.) Thanks in advance. -stefan
From: Jeff_Sickel@sickel.com Newsgroups: comp.sys.next.programmer Subject: Scheme for NEXTSTEP anyone? Date: 15 Jan 1996 07:59:11 GMT Organization: Charm.Net Baltimore Internet Access, Hon (410) 558-3900 Message-ID: <4dd1gf$35k@canton.charm.net> Does anyone know if there is version of Scheme out there that will run under non-Motorola NEXTSTEP? Schematik sounds like it would work well, except for the Motorola lib/binary. All hints appreciated. Jeff_Sickel@sickel.com
From: jsn@audiospeech.ubc.ca (John Nicol) Newsgroups: comp.sys.next.programmer Subject: Error when linking Intel/Motorola code Date: 15 Jan 1996 17:43:07 GMT Organization: The University of British Columbia Message-ID: <4de3nb$s8p@nntp.ucs.ubc.ca> I have installed 3.3 DeveloperTools and DeveloperLibs for Motorola and Intel on a NeXTStation 25 MHz 68040. The following compile and link (for Motorola only) of a simple test project succeeds: Copied 3.3TestProject.nib cc -g -Wall -O -I./sym -arch m68k -I. -c 3.3TestProject_main.m -o ./m68k_obj/3.3TestProject_main.o cc -g -Wall -O -I./sym -arch m68k -ObjC -sectcreate __ICON __header 3.3TestProject.iconheader -segprot __ICON r r -sectcreate __ICON app /usr/lib/NextStep/Workspace.app/application.tiff -o 3.3TestProject.app/3.3TestProject m68k_obj/3.3TestProject_main.o -lMedia_s -lNeXT_s The following compile and link (for Motorola and Intel) of the same simple test project fails: Copied 3.3TestProject.nib cc -g -Wall -O -I./sym -arch m68k -arch i386 -I. -c 3.3TestProject_main.m -o ./i386_m68k_obj/3.3TestProject_main.o cc -g -Wall -O -I./sym -arch m68k -arch i386 -ObjC -L -sectcreate __ICON __header 3.3TestProject.iconheader -segprot __ICON r r -sectcreate __ICON app /usr/lib/NextStep/Workspace.app/application.tiff -o 3.3TestProject.app/3.3TestProject i386_m68k_obj/3.3TestProject_main.o -lMedia_s -lNeXT_s cc: __ICON: No such file or directory cc: __header: No such file or directory *** Exit 1 Stop. *** Exit 1 Stop. A curious consequence of the failure of the second test link is that subsequent Motorola-only links fail henceforth. Thanks very much for your suggestions as to how one might remedy this problem. John Nicol School of Audiology and Speech Sciences University of British Columbia
From: jloda@flowbee.interaccess.com (Joseph D. Loda) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Objective-C Coding Standards Date: 15 Jan 1996 14:34:49 -0600 Organization: InterAccess Co. Message-ID: <4dedp9$frm@flowbee.interaccess.com> Hi, We are trying to come up with some coding guidelines for objective-c. Has anyone developed a set that I could have a look at? In the abscence of actual objective-c coding guidelines, how about a reference to some good C or C++ coding guidelines. Thanks in advance ... Joe -- jloda@interaccess.com
From: Steve Gemma <sagemma@monolith.bellcore.com> Newsgroups: comp.sys.next.programmer Subject: Help tracking mouse Date: Mon, 15 Jan 1996 15:58:49 -0500 Organization: Bell Communications Research Message-ID: <30FAC009.167EB0E7@monolith.bellcore.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I am writing an application in which I have a matrix of buttons. As the user passes his/her mouse pointer over a button, I would like that button's title to be displayed in a separate window. I have never worked with mouse tracking before, and was wondering if anyone had any example code that could show what I would need to do. I appreciate any help anyone can give me! -Steve
Newsgroups: comp.sys.next.programmer From: brianw@sounds.wa.com (Brian Willoughby) Subject: Re: Software License Protection without using hostid? Message-ID: <DL8w6t.7q3@sounds.wa.com> Organization: Sound Consulting, Bellevue, WA, USA References: <4d8vrn$84d@nntp.stanford.edu> Date: Mon, 15 Jan 1996 23:06:29 GMT In article <4d8vrn$84d@nntp.stanford.edu>, Gary P. Scavone <gary@ccrma.stanford.edu> wrote: >I'm trying to get access to a piece of software that uses a computer's hostid >in the license protection string (supposedly this is NeXT's solution). >Unfortunately, I have a stand-alone machine without an Ethernet card. >Therefore, I have no hostid. While the software protection is really not my >problem, not being able to use the software is! > >Does anyone know of a way that a NeXTStep software developer can incorporate >software protection without using hostid? I generally prefer not to be negative, but I believe that it would be prudent here for me to recommend against using a particular copy protection scheme that has been popular in NEXTSTEP circles for some time. The README for this package identifies it as "Registration Software adapted from SGAI (Simson Garfinkle and Associates, Inc.) public use code on BANG CD-ROM." My reason for recommending against this is that I have experienced panic and other crashing behavior with every program using this copy protection scheme and on every release of NEXTSTEP so far. That includes SBook, Mesa, eXTRASCAN, eXTRAREAD, and OmniPDF, among others. I have updated with every release since 2.1 and now run release 3.3 on my NeXTdimension, and I still suffer random panics when starting these programs. I have verified that the problem is also present in the version of this copy protection software as improved by Michael McNabb, who fixed some bugs in the original code. I had the opportunity to work on a development contract with a NEXTSTEP software company, and since I had source code to the application I was able to completely cure the panics by rebuilding the application without the Registration.subproj One could argue that these panics and crashes are NeXT's fault, somewhere in the UDP broadcast port networking code in the operating system. I submit that it does not really matter to your potential customers whose fault it is, they will not respond favorably to demos or especially registered applications which cause their machine to panic. P.S. There is a note that this copy protection software uses DES encryption, which cannot be exported without following certain US Government armament export restrictions and guidelines, so you may have other reasons to choose a different solution. -- Brian Willoughby Software Design Engineer, BSEE from NCSU NeXTmail welcome Sound Consulting: Software Design and Development BrianW@SoundS.WA.com Bellevue, WA <http://nwlink.com/cyberartists/brianw/brianw.html>
From: verhagk@ix.netcom.com(Ken Verhagen ) Newsgroups: comp.sys.next.programmer Subject: EOF & Oracle Date: 15 Jan 1996 22:35:12 GMT Organization: Netcom Message-ID: <4dekr0$h4m@ixnews6.ix.netcom.com> I am trying to connect to an Oracle server running on a Sparc running Solaris 2.4 (Oracle 7.2.2.3). When I try to connect using either the EOModeler or DBModeler I can't seem to get by the login screen. (It keeps returning to the login screen without any error) I have tried running the obsolete orasrv process on the server side and adding the following line to the etc/services file orasrv 1525/tcp oracle It doesn't seem to work. Is there anyone out there who has been able to connect to this version of Oracle (or any other version) using EOModeler? I would appreciate any help. I can be reached by either replying to this posting or by email verhagk@ix.netcom.com Thanks Ken Verhagen
From: marcel@cs.tu-berlin.de (Marcel Weiher) Newsgroups: comp.sys.next.programmer Subject: Re: HPDeskwriter 600 (nearly) works with NSwhite Date: 16 Jan 1996 01:36:33 GMT Organization: Technical University of Berlin, Germany Message-ID: <4devf1$l30@news.cs.tu-berlin.de> References: <9601150626.AA00225@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Stefan =?iso-8859-1?Q?B=F6hringer?= <Schroeer.Stefan@rz.ruhr-uni-bochum.de> writes: >Hi all, >astonishingly I got the HPDeskwriter to work with the HPDriver = >(HPDJ_for3.0 I think) from the archives. I use the Mux-driver to = >achive the 57600bps even over a 15450. Anything works fine except = >that sometimes graphic trash results. Since reducing the = >(pcl-)compresseion rate helped I assume that an internal buffer = >overflow of the printer is the problem. Since the apple-serials = >don't support hardware handshaking, there should be some software = >handshaking. >So does anybody know which kind of handshaking is involved with = >Appleprinters? Actually, the Apple serial ports support hardware handshaking, but on a different pin than NeXT. If memory serves, it was DTR handshaking instead of the RTS/CTS that NeXT uses. Marcel
From: "TOMMY K. HWANG" <thwang@home.mysolution.com> Newsgroups: comp.sys.next.programmer Subject: [USER QUOTA] Compiled but... Date: Tue, 16 Jan 1996 08:38:13 -0500 Organization: INTERNET SOLUTIONS INCORPORATED Message-ID: <30FBAA45.25F0@home.mysolution.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit SVR$/BSD style User Quota Software has been successfully compiled and is operational. Unfortunately, the quota shell can not be compiled without cmd.h and the functions are not declared anywhere in and of the header files. If there are anyone with experience making shells, please contact me. Without the quota shell, the binaries are as good as useless. Thanks. (when the quotashell - qsh - has been completed and tested, the binaries will be released). -ME PS. I also contacted NeXT but do not know if they are interested in spending a little time making a quota shell for us the unfortunate.
From: altenber@acpub.duke.edu (Lee Altenberg) Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs Subject: 3D Kit bug with triangle meshes Date: 15 Jan 1996 21:46:15 -0500 Organization: Duke University, Durham, NC, USA Message-ID: <4df3hn$svg@news.duke.edu> I've discovered a bug in the 3D kit. You can plot a 3D surface using RiPointsPolygons() with a mesh of either quadrilaterals or triangles, where the triangles come from dividing each quadrilateral into 2. With triangles, the surface shading does not work using N3D_SmoothSolids. The shading no longer bears any relation to the light sources. With quadrilaterals, however, N3D_SmoothSolids works fine. Has anyone had any experience with this bug and any workaround for triangle meshes? Thanks, Lee Altenberg altenber@mhpcc.edu
Newsgroups: comp.sys.next.programmer,comp.lang.oberon,comp.object.logic,comp.lang.scheme,comp.lang.beta From: Bob Hathaway <75027.1663@compuserve.com> Subject: OBJECT CURRENTS - *FREE* MAJOR NEW ON-LINE OBJECT-ORIENTED JOURNAL!!! Message-ID: <DL9BL7.5Ko@midway.uchicago.edu> Summary: Announcing a major ***FREE*** new Object-Oriented WWW Journal. Keywords: Free WWW OO Object-Oriented Journal OBJECT CURRENTS =============== OBJECT CURRENTS ONLINE HYPERTEXT JOURNAL FREE NEW MONTHLY OBJECT-ORIENTED FORUM Sender: news@midway.uchicago.edu (News Administrator) Organization: Object Currents Date: Tue, 16 Jan 1996 04:39:07 GMT Location: http://www.sigs.com/objectcurrents/ Editor: Bob Hathaway, bob_hathaway@notesgw.compuserve.com Premiere Issue: January 1, 1996 (Available now) Next Issue: February 1, 1996 Publisher: SIGS: C++ Report, JOOP, Object Magazine, ROAD, Smalltalk Report, X Journal, Object Expert, Objeck Spectrum, Java Report, ... This is an invitation to join us at Object Currents and view, engage and participate in new object-oriented technology using the latest form of information technology. Object Currents is a complete new free journal with Feature Articles, Columns, and Departments, along with 7 to 10 *new* articles from SIGS' Journals. Object Currents' World Class Columnists: Watts Humphrey: SEI Process Director, CMM & PSP Inventor Bertrand Meyer: Eiffel, OO Design and Software Engineering Francois Bancilhon: President, O2 Technology, Leading ODBMS Expert Michael Jesse Chonoles: Chief of Methodology, Advanced Concepts Center of Lockheed Martin David Shang: OO Programming Language Designer, Motorola Labs Michael Spertus: President, Geodesic Systems, Program Automation Interviews: January: Grady Booch February: James Rumbaugh March: Ivar Jacobson Newsgroup Dialog: February: Robert Martin Week in OT: Jane Grau 1/1, 1/8, 1/15 Feature Articles: Jan 96 Understanding Objects by Understanding Variables and Functions Dr. Stefano A. Cerri A Comparison of Objects with Frames and OODBs Dr. Xindong Wu Feb 96 On the applicability of the evolving Unified Method to Rapid Functional Prototyping Ian Mitchell Others... Departments: Several more including Code Watch and Question + Answer. Best new articles from all of SIGS January and February 1996 issues including: C++ Report, JOOP, Object Magazine, ROAD, Smalltalk Report, X Journal, Object Expert, Objeck Spectrum, Java Report, ... I've concluded our premiere January issue was a resounding success. Thanks to our readership for patronage, praise, and feedback. Please keep visiting or give it a try soon. Please also feel free to inform friends and colleagues of this free new medium. Object Currents' unique hypertext media provides for advances over earlier journals - links to home pages, sites, databases and information servers, interaction, animation, graphics, code retrieval and execution, expanded pages, video, virtual reality and chat sessions. While all of these may not have appeared in the first issue, they will appear in the future. For instance, Dr. Cerri's first article's interaction and code execution didn't make it into the January issue but will be backpatched soon. It is exemplary of an improved medium, with code examples interpreted and executed in real-time, displaying results and verifying correctness. Object Currents Contents: Object Currents Articles: 2 to 4 per issue. Columns: Several columns will appear from regular columnists on important topics. SIGS articles: 7 to 10 articles from current SIGS Publications per issue (JOOP, ROAD, Obj Mag., C++ Rpt, etc.) Newsgroup Dialog: Extrapolated subjects with commentary. Fun and Games: Puzzles to be picked up from the C++ Report and the net (Obfs. C++). Questionnaire: W/compiled results from last issue. SIGS Bookshelf: Book reviews and summary. Movers & Shakers: Interview or profile of object industry VIPs. Code Watch: Source code, from both SIGS journals and elsewhere, such as on the net. Editorial Comments: By both authors and readers are encouraged. This week in OT: Prepared by SIGS News Editor. Check it out! Best Regards, Bob Hathaway Robert John Hathaway III Editor in Chief Object Currents Hypertext Journal 655 West Irving Park Road, Suite 5417 Chicago, Illinois 60613-313775 USA Ph: 312-327-2121 Fax: 312-327-2936 Email: 75027.1663@compuserve.com - Correspondence, Submissions bob_hathaway@notesgw.compuserve.com - Submissions rjh@geodesic.com - Unix, Attachments ftp://ftp.geodesic.com/objectcurrents/submissions P.S. The Questionnaire is worth filling out, so that meaningful results can be provided in the next issue. But the following aren't really from OC's Departments:-) Examples from Object Currents' Question + Answer Department: ... Q 87. What does IA stand for: A. Intelligence Amplification B. Intelligence Augmentation C. Internet Agent D. All of the above E. None of the above Q 88. Object Currents is a good Spider/Agent test(T, F)? B. This question, but not answers, can be found by a Spider/Agent(T, F)? Q 89. Lawnmower Man II's Cyber City, Cyber Network, and Cyber Chip are all conceptually valid and technologically sound(T, F)? ... Instructions from the (YTBI) Puzzle and Games... If you reach this point: Using only the virtual objects provided, enter the following Cyber City and destroy the Cyber Chip before Cyberjobe jacks into the global network.
From: df@watershed.com (Dirk Fromhein) Newsgroups: comp.sys.next.programmer Subject: DO error message : handleRequestOnPortal: id d2bd48 not available? Date: Mon, 15 Jan 1996 21:40:46 -400 Organization: Watershed Technologies, Inc. Message-ID: <4df7hm$2j8@caesar.ultra.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NS 3.3, EOF 1.1, Foundation Patch. Intel Boxes. After about 500-600 DO connections/method calls, I always get: handleRequestOnPortal: id d2bd48 not available I've looked though the docs, headers, etc. and I can't find anything. I'm using NXAutoreleaseConnections. This is for a server that has to be up 7x24, and once we get that message we are hosed for any new connections. Clearly not acceptable! Thoughts? Thanks, Dirk Fromhein df@watershed.com
Newsgroups: comp.lang.objective-c,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Objective-C Coding Standards In-Reply-To: jloda@flowbee.interaccess.com's message of 15 Jan 1996 14:34:49 -0600 Message-ID: <RDL.96Jan16000129@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4dedp9$frm@flowbee.interaccess.com> Date: Tue, 16 Jan 1996 05:01:29 GMT I have but I'll have to check to see if I can release them... Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4dedp9$frm@flowbee.interaccess.com> jloda@flowbee.interaccess.com (Joseph D. Loda) writes: From: jloda@flowbee.interaccess.com (Joseph D. Loda) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Date: 15 Jan 1996 14:34:49 -0600 Organization: InterAccess Co. Hi, We are trying to come up with some coding guidelines for objective-c. Has anyone developed a set that I could have a look at? In the abscence of actual objective-c coding guidelines, how about a reference to some good C or C++ coding guidelines. Thanks in advance ... Joe -- jloda@interaccess.com
Newsgroups: comp.sys.next.programmer Subject: [Q] writing serial program - baud, parit From: apl@kcbbs.gen.nz (Andrew Lindesay) Date: 16 Jan 96 06:58:58 GMT Message-ID: <1749615.25138.15961@kcbbs.gen.nz> Organization: Kappa Crucis Unix BBS, Auckland, New Zealand I'm writing a little serial program to set the baud rate, stop bits etc... and then dump the file given as an argument to my printer. I'm having substantial trouble getting the baud rate set, and it's very frustrating. Does somebody have a small piece of code they could mail me to show how to do this. Yes I've played w/ the man pages, the online manual etc...and it's all terminal related and I really need to output without it being a terminal. Andrew (iapl@iconz.co.nz | apl@kcbbs.gen.nz)
Newsgroups: comp.sys.next.programmer From: tom@icgned.nl (Tom Hageman) Subject: Re: Error when linking Intel/Motorola code Message-ID: <DL9vop.A8@icgned.nl> Sender: news@icgned.nl Organization: IC Group References: <4de3nb$s8p@nntp.ucs.ubc.ca> Date: Tue, 16 Jan 1996 11:53:12 GMT In article <4de3nb$s8p@nntp.ucs.ubc.ca> jsn@audiospeech.ubc.ca (John Nicol) writes: > I have installed 3.3 DeveloperTools and DeveloperLibs for Motorola and Intel > on a NeXTStation 25 MHz 68040. > > The following compile and link (for Motorola only) of a simple test project > succeeds: > > Copied 3.3TestProject.nib > cc -g -Wall -O -I./sym -arch m68k -I. -c 3.3TestProject_main.m -o > ./m68k_obj/3.3TestProject_main.o > cc -g -Wall -O -I./sym -arch m68k -ObjC -sectcreate __ICON __header > 3.3TestProject.iconheader -segprot __ICON r r -sectcreate __ICON app > /usr/lib/NextStep/Workspace.app/application.tiff -o > 3.3TestProject.app/3.3TestProject m68k_obj/3.3TestProject_main.o -lMedia_s > -lNeXT_s > > The following compile and link (for Motorola and Intel) of the same simple > test project fails: > > Copied 3.3TestProject.nib > cc -g -Wall -O -I./sym -arch m68k -arch i386 -I. -c > 3.3TestProject_main.m -o ./i386_m68k_obj/3.3TestProject_main.o > cc -g -Wall -O -I./sym -arch m68k -arch i386 -ObjC -L -sectcreate __ICON > __header 3.3TestProject.iconheader -segprot __ICON r r -sectcreate __ICON app > /usr/lib/NextStep/Workspace.app/application.tiff -o > 3.3TestProject.app/3.3TestProject i386_m68k_obj/3.3TestProject_main.o > -lMedia_s -lNeXT_s > cc: __ICON: No such file or directory > cc: __header: No such file or directory > *** Exit 1 > Stop. > *** Exit 1 > Stop. > > A curious consequence of the failure of the second test link is that > subsequent Motorola-only links fail henceforth. > > Thanks very much for your suggestions as to how one might remedy this > problem. Let me guess... You used the Options... panel in PB's Builder to select the architectures, and then clicked the Add button at the bottom. Now you have inadvertently added an empty line in the library (Search order) browser; the makefile doesn't grok this, obviously:-) Remedy: select the empty line(s) in the library browser, and Remove it. Yeah, I got bitten by this `feature' once or twice too... -- __/__/__/__/ Tom Hageman <tom@basil.icce.rug.nl> [NeXTmail/Mime OK] __/ __/_/ IC Group <tom@icgned.nl> (work) __/__/__/ "there is no difference between theory and practise, __/ _/_/ at least in theory..." -- Bruce Becker
From: jake@hep.physics.mcgill.ca (Jason) Newsgroups: comp.sys.next.programmer Subject: CLISP ---> Need Assistance Date: 16 Jan 1996 13:48:43 GMT Organization: McGill University Computing Centre Message-ID: <4dgabr$85h@sifon.cc.mcgill.ca> Good day everyone, I am wondering if anyone has a recipe for compiling CLISP on the Next under NS 3.3. Failing that, is anyone willing to try to answer questions about the errors that crop up. I am not a programming guru, obviously, but I hope that I can phrase sufficiently clear questions. The easiest solution is probably getting a binary but... Thank you all in advance. jake -- _________________________________ Why you wanna be reading them long-haired books for boy? -Foghorn Leghorn
From: bbum@friday.com Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Objective-C Coding Standards Date: 16 Jan 1996 16:00:30 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4dgi2u$hb4@chinx4.thoughtport.net> References: <4dedp9$frm@flowbee.interaccess.com> # We are trying to come up with some coding guidelines for objective-c. # Has anyone developed a set that I could have a look at? In the # abscence of actual objective-c coding guidelines, how about a # reference to some good C or C++ coding guidelines. While it is not truly an Objective-C Coding Guidline, NeXT publishes a book named +Objective-C and Object Oriented Programming+ that provides most of the information one needs to write correct Objective-C code. Not only is the book an excellent style guide, it is also one of *the best* books on true object oriented programming-- a must read for *any* NEXTSTEP or Objective-C programmer. C++ programmers shouldn't bother reading it; the elegance and dynamic nature of Obj-C will only depress the reader... :-) Seriously, last year I read +Code Complete+ and re-read the +Obj-C & OO Programming+ book and my code improved in quality immensely-- and that's after 5 years of NEXTSTEP+ObjC programming... b.bum
From: Carl Edman <cedman@princeton.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Scheme for NEXTSTEP anyone? Date: 16 Jan 1996 15:22:26 GMT Organization: Princeton University Message-ID: <4dgfri$qp8@cnn.Princeton.EDU> References: <4dd1gf$35k@canton.charm.net> In <4dd1gf$35k@canton.charm.net> Jeff_Sickel@sickel.com wrote: > Does anyone know if there is version of Scheme out there that will run > under non-Motorola NEXTSTEP? Schematik sounds like it would work > well, except for the Motorola lib/binary. Just grab any of the dozen or so Scheme implementations available from the Scheme Repository at http://www.cs.indiana.edu/scheme-repository/home.html. Almost all of them compile with few or no problems under NS. In particular, scm compiles relatively painlessly, has good performance and a decent library in the form of slib. One exception is MIT Scheme. Its maintainers distribute only an old motorola-only binary and while they make sources available as well, to me it seems impossible to create a new binary from the sources only. That is a pity as it is a fairly pleasant environment and its compiler to i386 code is said to be quite good. Carl Edman
From: charlie@why.az.stratus.com (Charles_Spitzer) Newsgroups: comp.sys.next.programmer Subject: Sound problem Date: 16 Jan 1996 16:36:03 GMT Organization: Stratus Computer Inc, Marlboro MA Message-ID: <4dgk5j$5a0@transfer.stratus.com> I am having a sound problem that I can't figure out, nor is it documented anywhere. I have an app that queries a remote app (using oneway DO calls) for data. The data comes back via other oneway DO calls. When the data is returned, a Window is displayed and Sound is played. Since Sounds are played asynchronously and the windows are displayed quickly, there is a collision. I've implented a list of sounds to be played. When a Window is displayed, the Sound is placed on a List. When didPlay: in the Sound delegate gets called, it send free: to the Sound, then picks the next one off the List and sends it a play:. This all works when up to 4 Windows are displayed. When the 5th Window gets displayed and the Sound is placed on the list, the app hangs. In order to try to solve this, I created a cleanupSound: that gets executed by a perform:afterDelay:5000.0*[sound duration] to stop the Sound, and then play the next one. Here's a stack trace of the hang when the stop is attempted: #0 0x5007c48 in msg_send_trap () #1 0x502f31b in msg_send () #2 0x60e6048 in post_message () #3 0x60e6078 in play_end () #4 0x502e00a in stop_performance () #5 0x502eebc in SNDStop () #6 0x60e5f2c in -[Sound stop] () #7 0x415bb in -[Controller cleanupSound:] (self=0x4b55b8, _cmd=0x952c8, sender=0x71ed94) at Controller.m:586 #8 0x5032fcf in -[Object perform:with:] () #9 0x603b13d in doDelayedPerform () #10 0x6021e86 in checkTEs () #11 0x601d479 in _DPSGetOrPeekEvent () #12 0x601e5f6 in NXGetOrPeekEvent () #13 0x603087d in -[Application run] () #14 0x5a59f in main (argc=1, argv=0xbffffc48) at NCSSCall_main.m:15 #15 0x2d07 in start () at crt0.c:134 What's really odd is: (gdb) p *sound $2 = { isa = 0x618b928, soundStruct = 0x73a000, soundStructSize = 17368, priority = 0, delegate = 0x4b55b8, status = 2, name = 0x0, _scratchSound = 0x0, _scratchSize = 0 } The [sound status] = 2, which is not valid. This value should be 0 or -1 from Sound.h. The Sounds are being created via a [[Sound alloc] initFromSoundfile:path], so it isn't that it's being freed somewhere that I'm not aware of. The Sound doc states in numerous places that things don't work if the main event loop goes blocked. RPC does go blocked, so this could be the problem. I tried [NXSoundDevice setUseSeparateThread:YES] and [NXSoundDevice setTimeout:10 * 1000], but that didn't solve the problem. Does this sound familiar? Does anyone have any ideas on how to solve this problem? Due to the size of the app, and the connection you need to the mainframe to run it, I don't have a small sample of the problem that I can send out. --- Charles Spitzer charlie@az.stratus.com Customer Assistance Center Stratus Computer, Inc. Phoenix, AZ 85018
From: pspriet@pegase (Paul SPRIET) Newsgroups: comp.sys.next.programmer Subject: Distributed Objects Date: 17 Jan 1996 03:12:59 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4dhpfr$fto@ai.alienor.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi, I‚m writing here for one of my friend (a pretty french woman, who develops on NeXTSTEP, called Valerie ). Here‚s what she said : << I begin working with Distributed Objetcs under NeXTSTEP 3.3. I would appreciate some examples (simple ones, and other more complete) which can explain me how does all these things work. I have to create a notification server, and use broadcast message between the client and the server (in both directions). If someone can tell me where I can find such examples (on which ftp sites), it would be great. Thanx in advance for your help and have a nice day >> It‚s me again. All reply can be sent to : pspriet@xlan.alienor.fr Thanx -- +----------------------------------------------------------+ | Paul SPRIET | Email/NeXTmail: pspriet@xlan.fdn.org | | X&LAN - FRANCE | Tel(+33)56 51 04 89 Fax(+33)56 51 05 31 | +----------------------------------------------------------+
Newsgroups: comp.lang.objective-c,comp.sys.next.programmer From: cracauer@wavehh.hanse.de (Martin Cracauer) Subject: Re: Objective-C Coding Standards Message-ID: <1996Jan16.145624.20346@wavehh.hanse.de> Organization: BSD User Group Hamburg References: <4dedp9$frm@flowbee.interaccess.com> <RDL.96Jan16000129@world.std.com> Date: Tue, 16 Jan 96 14:56:24 GMT >In article <4dedp9$frm@flowbee.interaccess.com> jloda@flowbee.interaccess.com (Joseph D. Loda) writes: > From: jloda@flowbee.interaccess.com (Joseph D. Loda) > Newsgroups: comp.lang.objective-c,comp.sys.next.programmer > Date: 15 Jan 1996 14:34:49 -0600 > Organization: InterAccess Co. > Hi, > We are trying to come up with some coding guidelines for objective-c. > Has anyone developed a set that I could have a look at? In the > abscence of actual objective-c coding guidelines, how about a reference > to some good C or C++ coding guidelines. > Thanks in advance ... Joe > -- > jloda@interaccess.com GNU libobjects-0.x.x includes some guidelines I found very reasonable (and use them). Martin -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Martin Cracauer <cracauer@wavehh.hanse.de> - Fax +49 40 522 85 36 BSD User Group Hamburg, Germany - No NeXTMail anymore, please. Copyright 1995. Redistribution via Microsoft Network is prohibited
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: Scheme for NEXTSTEP anyone? Date: 16 Jan 1996 23:25:25 GMT Organization: my own Message-ID: <4dhc55$mnb@gate.seicom.net> References: <4dd1gf$35k@canton.charm.net> Jeff_Sickel@sickel.com wrote: > Does anyone know if there is version of Scheme out there that will run under > non-Motorola NEXTSTEP? Schematik sounds like it would work well, except for > the Motorola lib/binary. > All hints appreciated. How much do you want? 8-) The following implementations should compile without much hazzle... Personally I tried elk, Siod, scsh (well ,a shell but see below...) and guile. For more information see http://www.cs.indiana.edu/scheme-repository/home.html. * Guile-iii (http://www.cygnus.com/library/ctr/guile.html) * LIBSCHEME is a C library implementing Scheme as described in R4RS. It is easily integrated into a C program as a command interpreter or extension language, and is easily extended in C with new primitive types, primitve functions and syntax. It should be portable to most machines with an ANSI C compiler. It is available by anonymous ftp from ftp.cs.indiana.edu:/pub/scheme-repository/imp/ For more information, write to Brent Benson <Brent.Benson@mail.csd.harris.com>. * Scheme->C is an R4RS compliant Scheme system that is centered around a compiler that compiles Scheme to C. Besides the base language, the system includes "expansion passing style" macros, a foreign function call capability, records, weak pointers, 3 X11 interfaces, call/cc, and a generational, conservative, copying garbage collector. The result is a system that is portable, efficient, and able to build applications that contain a mix of compiled and interpreted Scheme, and compiled code from C, C++ and other languages. The current release of Scheme->C runs on the following systems: Digital Alpha AXP systems with OSF/1, MIPS based DECstations, VAXen with ULTRIX, MIPS based SGI systems, PC's running Microsoft Windows 3.1, Apple Macintosh's running system 7.1, HP 9000/300, HP 9000/700, Sony News, Harris Nighthawk and other m88k systems, linux, Sun SPARC, and NT (Visual C++ compiler). Earlier releases also run on Sun3, DNx500, DN1000, 386's running Unix, DNx500, and DN1000 systems. The software is available for anonymous ftp from gatekeeper.dec.com:/pub/DEC/Scheme-to-C/ [16.1.0.2] There are three interfaces to X11, all written in Scheme, available from gatekeeper. The first is a complete set of stubs to Xlib included in the base system. The second is an alternative to Xlib called SCIX, found in gatekeeper.dec.com:/pub/X11/contrib/ The third, ezd, allows programs to easily produce interactive, structured graphics and is found in gatekeeper.dec.com:/pub/DEC/ezd/ Those without ftp access can also obtain Scheme->C and ezd from the Prime Time Freeware CD, Vol. 1, No. 2. Additional information is available from the author at Digital Equipment Corporation's Western Research Lab: Joel Bartlett, bartlett@decwrl.dec.com. * Scheme 48 is a Scheme implementation based on a virtual machine architecture. Scheme 48 is designed to be straightforward, flexible, reliable, and fast. It should be easily portable to 32-bit byte-addressed machines that have POSIX and ANSI C support. In addition to the usual Scheme built-in procedures and a development environment, library software includes support for hygienic macros (as described in the Revised^4 Scheme report), multitasking, records, exception handling, hash tables, arrays, weak pointers, and FORMAT. Scheme 48 implements and exploits an experimental module system loosely derived from Standard ML and Scheme Xerox. The development environment supports interactive changes to modules and interfaces. A beta release of Scheme 48 is available by anonymous ftp from ftp.cs.indiana.edu:/pub/scheme-repository/imp/scheme48-0.36.tar.Z swiss-ftp.ai.mit.edu:/archive/s48/scheme48-0.36.tar.gz ftp.cs.cmu.edu:/user/ai/lang/scheme/impl/s48/scheme48.tar.gz For more information, contact Richard Kelsey and Jonathan Rees at <scheme-48-request@martigny.ai.mit.edu>. * Scsh is a Unix shell/systems programming environment implemented on top of Scheme 48 (a portable, byte-code compiled R4RS Scheme implementation). Scsh provides - A high-level macro notation for writing typical shell-script computations: running programs, pipelines, I/O redirection, and so forth. For example, to decompress a file and mail it to someone, you might say (run (| (gzcat home.html.gz) (mail -h "Here's my home page" shivers@lcs.mit.edu))) To spell check your paper, printing out the results, you could say: (run (| (delatex (< paper.tex)) (spell) (lpr -Ppulp))) - A complete system-call interface to Unix: fork, exec, I/O, file system, time, env vars, and so forth. The I/O interface includes a *complete* interface to BSD sockets, both Unix and TCP/IP domains. I/O is completely integrated with Scheme ports. System calls return useful values, not error codes; errors are reported by raising exceptions which can be caught by handlers. - Other useful shell-programming utilities: filename globbing/pattern matching, regexp matching, macros for writing AWK-like programs, field and record parsers, and so forth. - The ability to write executable shell scripts using the Unix #! interpreter feature, with access to command-line argv values. These features are completely integrated into Scheme 48's R4RS Scheme implementation; the programming language is Scheme. The scsh release is self-contained -- it comes with its own complete Scheme 48 implementation. Scsh currently runs on the following platforms: DEC Ultrix, Harris NightHawk, HP-UX, IBM AIX, Linux, NetBSD/i386, NeXTSTEP/Intel, SGI IRIX, Solaris, and SunOS. It's not hard to port scsh to new systems. | You can get a copy of scsh via anonymous ftp, from the following: ftp://clark.lcs.mit.edu/pub/su/scsh/scsh.tar.gz ftp://swiss-ftp.ai.mit.edu/pub/su/scsh/scsh.tar.gz These tar files include a detailed manual and a paper describing the design of the system. For the lazily curious, we also have the manual separately available as ftp://clark.lcs.mit.edu/pub/su/scsh/scsh-manual.ps ftp://swiss-ftp.ai.mit.edu/pub/su/scsh/scsh-manual.ps Scsh has been implemented by the Scheme Underground http://www.ai.mit.edu/projects/su/su.html For further information, contact Olin Shivers <shivers@lcs.mit.edu>, http://clark.lcs.mit.edu/~shivers/home.html * SCM, free by anonymous ftp from swiss-ftp.ai.mit.edu:/archive/scm/ prep.ai.mit.edu:/pub/gnu/jacal/ ftp.cs.indiana.edu:/pub/scheme-repository/new/ ftp.maths.tcd.ie:/pub/bosullvn/jacal/ Current version 4e1. Runs on Amiga, Atari-ST, MacOS, MS-DOS, OS/2, NOS/VE, VMS, Unix and similar systems. SCM conforms to the Revised^4 Report on the Algorithmic Language Scheme and the IEEE P1178 specification. Scm is written in C. ASCII and EBCDIC are supported. Written by Aubrey Jaffer. To receive an IBM PC floppy disk with the source files and MSDOS and i386 executables send $99 to Aubrey Jaffer, 84 Pleasant Street, Wakefield MA 01880, <jaffer@zurich.ai.mit.edu>. SLIB (Standard Scheme Library) is a portable Scheme library which is intended to provide compatability and utility functions for all standard Scheme implementations, including SCM, Chez, Elk, Gambit, MacScheme, MITScheme, scheme->C, Scheme48, T3.1, and VSCM, and is available as the file slib2a0.tar.gz. Written by Aubrey Jaffer. JACAL is a symbolic math system written in Scheme, and is available as the file jacal1a4.tar.gz. SCMCONFIG contains additional files for the SCM distribution to build SCM on Unix machines using GNU autoconf. SLIB-PSD is a portable debugger for Scheme (requires emacs editor). TURTLSCM is a turtle graphics package which works with SCM on MSDOS or X11 machines. Written by Mkinen Sami <sjm@cc.tut.fi> and Jarkko Leppanen <jtl@cc.tut.fi>, it is available as the file turtlegr.tar.Z. (Also available from modeemi.cs.tut.fi:/pub/scm/ as turtlegr.tar.gz, along with an already-compiled MSDOS binary of scm with turtlegraphics and slib in scm4c0_b.zip.) XSCM is an X Windows interface to Xlib and the Motif and OpenLook toolkits for the SCM interpreter. It requires scm4a10 or later. It should be available at any archive of alt.sources, or on swiss-ftp, prep and indiana as the file xscm1.05.tar.Z. Contact campbell@redsox.bsw.com for more information. SMG-SCM is a package that adds VMS SMG screen management routines to SCM. It is available from swiss-ftp, prep and indiana as the file sgm-scm2a1.zip. (A VMS version of Unzip is available by anonymous FTP from ftp.spc.edu:[ANONYMOUS.MACRO32]UNZIP.EXE.) This file contains the source code, documentation, and example code. Send comments and bugs to T. Kurt Bond, <tkb@mtnet2.wvnet.edu> (preferred) or <Kurt.Bond@launchpad.unc.edu>. WB is a disk based, sorted associative array C library (database). These associative arrays consist of variable length (less that 256 bytes) keys and values. WB comes with an interface to SCM. Basic operations are creation, destruction, opening and closing of diskfiles and arrays, insertion, deletion, retrieval, successor, and predecessor (with respect to dictionary order of keys). Functional application of find-next, deletion, and modification over a range of consecutive key values is supported. Multiple associative arrays can be stored in one disk file. Simultaneous access to multiple disk files is supported. A structure checker, garbage collector are included. A repair program and ram-disk type file (for temporary structures) are in developement. The current WB implementation has a file size limit of 2^32 * block size (default 2048) = 2^43 bytes (8796 Gbytes). WB does its own memory and disk management. WB is available on swiss-ftp, prep, and indiana as wb1a1.tar.z. * SIOD (Scheme in One Defun), free by anonymous ftp from ftp.cs.indiana.edu:/pub/scheme-repository/imp/siod-v3.0-shar ftp.std.com:/pub/gjc/siod-3.0-shar or in any comp.sources.unix archive. Runs on VAX/VMS, VAX UNIX, Sun3, Sun4, Amiga, Macintosh, MIPS, Cray, Windows NT/WIN32. Small scheme implementation in C arranged as a set of subroutines that can be called from any main program for the purpose of introducing an interpreted extension language. Compiles to ~42K bytes of executable. Lisp calls C and C calls Lisp transparently. Version 3.0 includes support for manipulation of Oracle and Digital RDB relational databases (SQL interface). Written by George Carrette <gjc@paradigm.com> or <gjc@world.std.com>. * Elk (Extension Language Kit) has been designed specifically as an embeddable, reusable extension language subsystem for applications written in C or C++. Elk is also useful as a stand-alone Scheme implementation, in particular as a platform for rapid prototyping of X11-based Scheme programs. Elk was first published in 1989; the current version is Elk 3.0. The Elk distribution includes a Scheme interpreter (embeddable and stand-alone versions), several dynamically loadable extensions, run-time support (including a top-level implemented in Scheme and a debugger), and 230+ pages of documentation (troff and PostScript format). Major features of Elk are incremental, dynamic loading of compiled extensions (supported on many platforms); freezing of the interpreter or application into a new executable file; a C/C++ programmer's interface for language interoperability; Scheme bindings for X11 Xlib, Xt, Athena and Motif Widgets; a UNIX interface (not restricted to POSIX); bitstrings, records, and regular expressions; a stop-and-copy and an incremental, generational garbage collector. The Elk 3.0 distribution and more information about Elk are available in the World Wide Web at http://www.informatik.uni-bremen.de/~net/elk The distribution is also available on a number of FTP sites including ftp://ftp.x.org/contrib/devel_tools/elk-3.0.tar.gz ftp://ftp.uni-bremen.de/pub/programming/languages/scheme/elk/elk-3.0.tar.gz For more information contact Oliver Laumann <net@cs.tu-berlin.de>. (a patch to get the dynamic loading reliable on NS is available from me, frank@this.net) Hope this helps Frank --- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP & PostScript Guy "In cantonese C++ is called C ga ga"
From: espeyton@ix.netcom.com (Eric Peyton) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Objective-C Coding Standards Date: 17 Jan 1996 05:02:55 GMT Organization: Netcom Message-ID: <4dhvtv$9qn@ixnews2.ix.netcom.com> References: <4dedp9$frm@flowbee.interaccess.com> I would definitely recommend getting a copy of Code Complete by Steve McConnell. Though it is not an Objective-C standards book, the standards presented in the book are definitely applicable to ANY programming environment. Oh, and sorry, the book is published by Microsoft Press. Eric Peyton epeyton@ix.netcom.com
From: Stefan Leuker <leuker> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Objective-C Coding Standards Date: 17 Jan 1996 08:15:25 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Message-ID: <4dib6t$f3i@news.rwth-aachen.de> References: <4dedp9$frm@flowbee.interaccess.com> <4dgi2u$hb4@chinx4.thoughtport.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit bbum@friday.com wrote: >While it is not truly an Objective-C Coding Guidline, NeXT publishes a book >named +Objective-C and Object Oriented Programming+ that provides most of the >information one needs to write correct Objective-C code. > I could be wrong, but isn't this book part of the online documentation? A friend of mine bought the book, too (if it's the one I remember) and found the complete contents on his harddisk. Stefan
From: jq@papoose.quick.com (James E. Quick) Newsgroups: comp.sys.next.programmer Subject: Re: Help with compliling a program Date: 17 Jan 1996 07:32:35 -0500 Organization: PHCS Message-ID: <4diq93$h41@papoose.quick.com> References: <4d9ndb$19ea@saba.info.ucla.edu> In article <4d9ndb$19ea@saba.info.ucla.edu>, Rick Vazquez <vazquezr@physics.ucla.edu> wrote: >rm -f xconq >/bin/cc -o xconq -g -L/usr/X11R6/lib xconq.o xinit.o >xmap.o xdraw.o xcmd.o xprint.o >ld: Undefined symbols: >_strdup The strdup function is just a string copy routine. For something like this it's easiest just to roll your own and include the .o in the link phase. Here is the strdup.c that I have kept around for years. #include <stdlib.h> /* strdup ** Return a pointer to a newly allocated copy of a string. ** If the input string is NULL or malloc fails then return a NULL pointer. */ char * strdup(const char *string) { char *new = 0; if (string && (new = malloc(strlen(string) + 1))) strcpy(new, string); return new; } -- ___ ___ | James E. Quick jq@quick.com / / / | Private HealthCare Systems NeXTMail O.K. \_/ (_\/ | Systems Integration Group (617) 895-3343 ) | - My other car has a mouse.
From: mike@hobbs.chem.usu.edu (Michael Emmel) Newsgroups: comp.sys.next.programmer Subject: Subclassing hidden objects Date: 17 Jan 1996 15:02:05 GMT Organization: SouthWind Internet Access, Inc. Distribution: world Message-ID: <4dj31d$66d@opal.southwind.net> I would like to subclass some of the objects that Next decided to hide from the programmer. The Object files are in the libraries. To do this I think one must just create a accetable header file ?? Or is there more to it than that. How do I subclass or send messages to hidden objects?? Thanks Mike
From: Marcus Edwardes <75151.1557@CompuServe.COM> Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer,misc.jobs.contract Subject: NeXTstep Consulting in USA Date: 17 Jan 1996 18:07:39 GMT Organization: Computec International Resources Message-ID: <4djdtb$47t$1@mhadg.production.compuserve.com> Computec International Resources specializes in providing consultants with rare skills to a broad client base across the USA. We employ suitably qualified individuals from all over the world, and offer excellent salaries and benefits, including the correct work visa, green card sponsorship, flights and relocation (incl. family), paid vacation and paid public holidays. We are currently seeking experienced NeXTSTEP developers, with at least 18 months commercial application development experience, and preferably some exposure to EOF, to work on major projects across the USA. For further information please contact our Los Angeles office: Marcus Edwardes Computec International Resources Tel: (818) 500 3921 Fax: (818) 500 3924 Email: 75151.1557@compuserve.com
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Subclassing hidden objects Date: 17 Jan 1996 20:52:17 +0100 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7lon6zjim.fsf@tom.es.ele.tue.nl> References: <4dj31d$66d@opal.southwind.net> In-reply-to: mike@hobbs.chem.usu.edu's message of 17 Jan 1996 15:02:05 GMT In article <4dj31d$66d@opal.southwind.net> mike@hobbs.chem.usu.edu (Michael Emmel) writes: To do this I think one must just create a accetable header file ?? Create an interface declaration (or have one created by class-dump) which is accurate concerning superclass and instance variables. How do I subclass or send messages to hidden objects?? For subclassing, you need (to write) the interface of the class. To send messages to instances, you do not need anything; just refer to the object through an `id' type variable and make sure the methods you invoke have been seen---it does not matter where, as long as they have been seen. --Tiggr
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Subclassing hidden objects Date: 17 Jan 1996 20:01:30 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4djkiq$irn@news.its.com> References: <4dj31d$66d@opal.southwind.net> mike@hobbs.chem.usu.edu (Michael Emmel) wrote: > I would like to subclass some of the objects that Next decided to hide > from the programmer. The Object files are in the libraries. To do this I > think one must just create a accetable header file ?? Oh, boy. Yes, you need to create a header file; there's a tool called class-dump which may prove useful. You should be aware of some big caveats, however. First, what you are doing is reverse-engineering private API, and that API was hidden for a good reason. You can expect your code to not work between one relase of the OS and another, and possibly even not work between one architecture and another. Secondly, it is difficult (or impossible) to accurately recreate all of the instance variables used in the private class since it is not uncommon for classes to have a pointer to a structure that you can't get info about. > Or is there more to it than that. How do I subclass or send messages to > hidden objects?? The runtime system is aware of all registered classes, even ones that NeXT does not provide headers for. You can use the runtime functions (check in DL) to do various things like that, although you *really* do not want to try and subclass a class that is private and for which you do not have documentation or a complete header for. You are better off using categories to provide additional behaviors without trying to subclass. What are you trying to do, anyway? Whatever it is, you really should have a compelling reason.... -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Objective-C Coding Standards Date: 17 Jan 1996 19:41:18 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4djjcu$irn@news.its.com> References: <4dedp9$frm@flowbee.interaccess.com> <4dhvtv$9qn@ixnews2.ix.netcom.com> espeyton@ix.netcom.com (Eric Peyton) wrote: > I would definitely recommend getting a copy of Code Complete by Steve > McConnell. Though it is not an Objective-C standards book, the standards > presented in the book are definitely applicable to ANY programming > environment. Oh, and sorry, the book is published by Microsoft Press. I would strongly second that recommendation. If you develop software or manage people who do, you really should read this book. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Interface Builder won't start no more... Date: 17 Jan 1996 19:40:24 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4djjb8$4kn@news.onramp.net> References: <4d29pr$sh2@driene.student.utwente.nl> Maybe IB is trying to load a bad palette. Try changing the name of your .NeXT directory (where app preferences are stored) in your home directory and running IB again. -- Steve Dekorte Objective-C/NEXTSTEP developer http://www.marble.com/~dekorte NeXTmail, MIME welcome.
From: nivi@exhibit.media.mit.edu (Babak Nivi) Newsgroups: comp.sys.next.programmer Subject: objective c mode for emacs Date: 18 Jan 1996 01:47:34 GMT Organization: Massachvsetts Institvte of Technology Message-ID: <4dk8rm$9mr@senator-bedfellow.MIT.EDU> I know you've heard this before a million times, but I've checked the faq and the objective-c-mode that it points to on www.cs.orst.edu does not run correctly as far as I can tell. I get the following error: Symbol's value as variable is void: /* I checked the code and the version on the archive is in fact 3.03a instead of 3.02. If anyone has or knows where to get a working version of objective-c-mode, I would appreciate a hand. -- Babak Nivi
From: mark@dlib.com (Mark Middleton) Newsgroups: comp.sys.next.programmer Subject: Job Vacancy Information Date: Thu, 18 Jan 1996 08:11:59 GMT Organization: BTnet, BT Public Internet Service Message-ID: <4dkvat$1li@news.bt.net> Several Vacancies exist for the following Jobs... ================================================= SOFTWARE-AGENT DEVELOPERS WANTED Working for one of the world's leading Japanese technology companies (over 111,000 employees world-wide, and sales for 1995 exceeding US$30bn). You will be based in one of our UK offices, working within a global R&D network to implement research driven product development. This newly established organisation has been recognised as a key strategic area for the Company's business expansion into the 21st century. Project: Developing large scale, open, distributed information system using agent technology. * Innovative and challenging leading edge project * Research oriented commercial environment * Developing real industrial strength agents * Involves active participation from leading academics in agent field Qualifications: PhD in agent technology or closely related field or 3-4 years of experience Skills: a. Plan generation + execution b. Learning + adaptation c. Goal recognition + prediction d. Co-operative protocols e. Negotiation f. Strong Programmer Excellent salary, performance related bonus + comprehensive benefits. Please send resume to resume@dlib.com in ASCII or HTML. ================================================== OBJECT & WEB DEVELOPERS WANTED Working for one of the world's leading Japanese technology companies (over 111,000 employees world-wide, and sales for 1995 exceeding US$30bn). You will be based in one of our UK offices, working within a global R&D network to implement research driven product development. This newly established organisation has been recognised as a key strategic area for the Company's business expansion into the 21st century. Project: Developing large scale, open, distributed information system using agent, object and WWW technology. * Innovative and challenging leading edge project * Research oriented commercial environment * Developing real industrial strength distributed systems using distributed object and WWW technology. * Involves active participation from leading academics in the field Qualifications: Masters Degree in Computer Science or 3-4 years of experience Skills: a. Strong Object Programmer b. Framework design, patterns, systems, distributed objects c. WWW technology: HTTP (NG), HTML, PEP, Java, NSAPI, ... d. Smalltalk, Objective-C, Java e. NEXTSTEP, Solaris, Win32, OLE, NCAPI Full training and knowledge transfer will be provided. Excellent salary, performance related bonus + comprehensive benefits. Please send resume to resume@dlib.com in ASCII or HTML. +-------------------------------------------+ | Mark Middleton | mark@dlib.com | | System Architect | | +-------------------------------------------+
From: trojaner@studbox.uni-stuttgart.de (Michael Hoess) Newsgroups: comp.sys.next.programmer Subject: Problem with using Custom-Object in EO Date: 18 Jan 1996 11:07:47 GMT Organization: Comp.Center (RUS), U of Stuttgart, FRG Message-ID: <4dl9m3$2qtk@info4.rus.uni-stuttgart.de> Hi ! I'm quite a newbie with NeXT-programming and the EOF and now have a problem (-: I want to use the FixedPoint-Class by Joe Panico (which I found in the EOF-mailing-list archive on www.omnigroup.com) as class for one of the attributes of my application. I included the class in my project, set the custom-classname for the attribute in the EO-Modeler (created a new template there) and could compile everything without error. But now the application ignores everything and uses a NSString or something like that for this attribute. Methods of the FixedPoint-Class wont be called a single time during run-time. What did I forgot ? Please help me ! Thanx in advance for every answer ! Michael Hoess trojaner@studbox.uni-stuttgart.de
From: t68@nikhef.nl (Jos Vermaseren) Newsgroups: comp.sys.next.programmer Subject: Driver for Cirrus Logic GD6440 Message-ID: <3794@nikhefh.nikhef.nl> Date: 18 Jan 96 09:52:26 GMT Organization: Nikhef-H, Amsterdam (the Netherlands). Hi, Could somebody explain me what would be involved in obtaining/writing a driver for a Cirrus Logic GD6440 controler? Does anybody have the information that is needed to put into the driver (I could not get into the WWW info system of Cirrus Logic). Would it be much work? How realistic is it to do these things oneself? (I used to be quite a good systemhacker on the Atari, but since I have my NeXT --already 5 years-- I gave up on that.) I guess the debugging will be quite some fun, but I do want to see my screen in color eventually and it does not look like anybody else has made this driver. It is a bit frustrating to see that MS-DOS on my portable can make nicer screens (color) than NeXTstep. Thanks in advance Jos Vermaseren
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: NSNotificationCenter Date: 18 Jan 1996 13:22:57 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4dlhjh$4f0@lion.embl-heidelberg.de> Hi NeXTers, I'm polishing the GNU's NSNotificationCenter implementation and would like to clarify several points where the OS spec is not too verbal. 1) Should I check if an observer is registered only once with the same selector/notification-name pair or should I assume that this is a developer's responsibility? If there was no question of performance, than there is no problem this to be checked, but as I seen (writing also an WindowManager for GNU), the number of observers could be really huge! 2) If one observer is added twice with the same selector/notification-name pair (with and without object) using the following method: - (void)addObserver:(id)anObserver selector:(SEL)aSelector name:(NSString *)aName object:(id)anObject it could happen (for my current implementation) that the observer will be notified twice (Note: I'm using different NSDictionaries and trying to avoid any if statements ... in future implementation I'll will use even separate threats to post the notifications asynchronously). Is the described behavior wrong or could make some problems? I'm still waiting for NS4.0 :-( and cannot check how NeXT's implementation behaves -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387524 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: eble@ai-lab.fh-furtwangen.de (Axel Eble) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: (Off-Topic:) Re: Objective-C Coding Standards Followup-To: comp.lang.objective-c,comp.sys.next.programmer Date: 18 Jan 1996 17:03:31 GMT Organization: FH Furtwangen, AI Lab Message-ID: <4dluh3$r19@zeus.ai-lab.fh-furtwangen.de> References: <4dedp9$frm@flowbee.interaccess.com> <4dhvtv$9qn@ixnews2.ix.netcom.com> <4djjcu$irn@news.its.com> Chuck Swiger (chuck@its.com) wrote: : espeyton@ix.netcom.com (Eric Peyton) wrote: : > I would definitely recommend getting a copy of Code Complete by Steve : > McConnell. [...] : > [...] Oh, and sorry, the book is published by Microsoft Press. : : I would strongly second that recommendation. If you develop software or : manage people who do, you really should read this book. so, if it's published by MicroSoft, why don't *THEY* read it? ;o) : -Chuck Axel -- Axel Eble * GHB 02-01-03 * 78120 Furtwangen * 07723 50271 * eble@ai-lab.fh-furtwangen.de PGP Key fingerprint = 4D 6B 2B 34 1F E6 28 51 F8 67 67 C3 77 1E E9 B0 "That's the way you respond to everything: if it doesn't work, hit it!" B'ellana Torres (human) to B'ellana Torres (klingon)
From: Yibing Wu <yibwu@plato.sky.bdm.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Looking for a "Graph Display" palette Date: 18 Jan 1996 18:43:48 GMT Organization: BDM International, Inc. Message-ID: <4dm4d4$n95@news.mcl.bdm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit We are constructing a front end for a relational data source. Some of the entities either have tree/graph like relation among themselves or records that have this kind of relation. For example, a manager can manages multiple salesman and some of the salesman may share one customer, etc. I am looking for some kind of palette that would enable us to display this kind of relationship "GRAPHICALLY". Like, display a root node(manager) in a canvas-like view, when you double click on this node, all the nodes denoting salesperson will come to its right and has a line like to the root node, so on so forth. We do need the palette would be able to display network-like relation, though. Any Input? Thanks! A graph is like(roughly): b / a \ d c/ \ e Yibing Wu
From: nazari@cs.chalmers.se (Nader Nazari,1018) Newsgroups: comp.sys.next.programmer Subject: Asking for Help about RenderMan Date: 18 Jan 1996 14:32:28 GMT Organization: Chalmers University of Technology Message-ID: <4dllls$11@nyheter.chalmers.se> Hi I work on a project which needs a lot of 3D images. I start to work with the N3D library, but 3d documentation says "This documentation assume you are familiar with the RenderMan language..." If you are not a book called "The RenderMan Companion .. " is recommender. I bought the book and now I try to run the first program in the book. #include <ri/ri.h> RtPoint Square[4] = {{.5,.5,.5},{.5,-.5,.5},{-.5,.5,.5}}; main() { RiBegin(RI_NULL); RiWorldBegin(); RiSurface("constant",RI_NULL); RiPolygon(4,RI_P,(RtPointer)Square,RI_NULL); RiWorldEnd(); RiEnd(); } I succeeded to compile this program by using "cc" but how can I run it. To run the program I need a window ( it does not create it) to create a window and using RenderMan I must use N3D Kit. To use N3D kit I have to know RenderMan and to learn RenderMan I must use the book and thus run the program!! Is any body there who can save me out of this loop. Tanks Nader <nazari@cs.chalmers.se>
From: griffith@crl.com (Dave Griffith) Newsgroups: comp.sys.next.programmer Subject: NeXTStep Developer Wanted - Downtown Chicago Date: 18 Jan 1996 10:53:06 -0800 Organization: CRL Dialup Internet Access Message-ID: <4dm4ui$i61@crl.crl.com> Growing international market maker needs a talented, experienced NeXTStep developer for its Chicago office. We're looking for a strong, self-motivated individual able to deal with demanding, knowledgeable users on a daily basis. Plusses: Sybase Data Feeds: (Reuters, Telerate, Telekurs, etc.) Financial background, especially option pricing. EOF/FoundationKit Fax Resumes to: David Padgitt 312-424-3020 E-Mail Resumes: davep@fct.com !!!REPLIES TO THE ADDRESS IN THE HEADER WILL BE IGNORED!!!
From: monty@tbyte.com (Monty Zukowski) Newsgroups: comp.sys.next.programmer Subject: Emacs scroll bar Date: 18 Jan 1996 20:42:39 GMT Organization: Trilobyte, Inc. Message-ID: <4dmbbv$73c@marvin.tbyte.com> The one thing I hate about Emacs on the nExt is that when you drag the scroll bar the text doesn't move until you let go. Is there any way to change that? Monty monty@tbyte.com
From: jcs@informatik.uni-koblenz.de (Juergen Schoenke) Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 18 Jan 1996 19:56:59 GMT Organization: University Koblenz / Germany Message-ID: <4dm8mb$6ne@newshost.uni-koblenz.de> References: <4d6u45$1cs@gateway.wiltel.com> In article <4d6u45$1cs@gateway.wiltel.com> jamesr@pennave (James Robertson) writes: > Any DO gurus out there? > > I'm having a problem with uniquing of proxies. I'm getting multiple > proxies to the same remote object. At first, I thought my problems > might be related to working in hybrid NX/NS environment. But I > wrote small test applications in pure NX with same results. > Here's what it looks like... [...] > My Server and Client run OK. They can ping each other, unsolicited, > in either direction. But if the server sends the client a message: > [proxyToClient heyBuddy: self]; > the proxy which the client receives as the argument to -heyBuddy: > will have a different address than proxyToServer which was originally > created when the client connected to the server. I was expecting > DO to unique the proxies, and that they would be identical. > > Any ideas? > > Thanks, > - Jim Robertson Bug or feature? I checked up this and can confirm your results. There's nothing wrong with your code. But what you reported is rather a (undocumented) feature of DO, than a bug. The id you get when connecting to a server with "connectToName:..." is a special proxy to the root object on ther server side (whis no respect to an individual object). Normally the "name" instance variable stores the "original" id of the remote object to match it with a local proxy on the server side of the connection. When an id is recieved from the server and a proxy with that name already exists, then this Proxy is returned. The NXProxy instance you recieve when connecting to a server has no name (name is 0, you can check it with the nameForProxy instance method of the NXProxy class). I think, messages to this special proxy are routed to the object that's referenced in the "rootObject" instance variable of the corresponding (server-side) NXConnection instance. Only the first time (after connecting) you recieve the root object over the connection, a new proxy will be created. From then, it will be matched correctly with the existing proxy. I hope, this will help you JCS
From: jcs@informatik.uni-koblenz.de (Juergen Schoenke) Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 18 Jan 1996 19:59:30 GMT Organization: University Koblenz / Germany Message-ID: <4dm8r2$6or@newshost.uni-koblenz.de> References: <4d6u45$1cs@gateway.wiltel.com> In article <4d6u45$1cs@gateway.wiltel.com> jamesr@pennave (James Robertson) writes: > Any DO gurus out there? > > I'm having a problem with uniquing of proxies. I'm getting multiple > proxies to the same remote object. At first, I thought my problems > might be related to working in hybrid NX/NS environment. But I > wrote small test applications in pure NX with same results. > Here's what it looks like... [...] > My Server and Client run OK. They can ping each other, unsolicited, > in either direction. But if the server sends the client a message: > [proxyToClient heyBuddy: self]; > the proxy which the client receives as the argument to -heyBuddy: > will have a different address than proxyToServer which was originally > created when the client connected to the server. I was expecting > DO to unique the proxies, and that they would be identical. > > Any ideas? > > Thanks, > - Jim Robertson Bug or feature? I checked up this and can confirm your results. There's nothing wrong with your code. But what you reported is rather a (undocumented) feature of DO, than a bug. The id you get when connecting to a server with "connectToName:..." is a special proxy to the root object on ther server side (whis no respect to an individual object). Normally the "name" instance variable stores the "original" id of the remote object to match it with a local proxy on the server side of the connection. When an id is recieved from the server and a proxy with that name already exists, then this Proxy is returned. The NXProxy instance you recieve when connecting to a server has no name (name is 0, you can check it with the nameForProxy instance method of the NXProxy class). I think, messages to this special proxy are routed to the object that's referenced in the "rootObject" instance variable of the corresponding (server-side) NXConnection instance. Only the first time (after connecting) you recieve the root object over the connection, a new proxy will be created. From then, it will be matched correctly with the existing proxy. I hope, this will help you JCS
From: jcs@informatik.uni-koblenz.de (Juergen Schoenke) Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 18 Jan 1996 19:59:35 GMT Organization: University Koblenz / Germany Message-ID: <4dm8r7$6os@newshost.uni-koblenz.de> References: <4d6u45$1cs@gateway.wiltel.com> In article <4d6u45$1cs@gateway.wiltel.com> jamesr@pennave (James Robertson) writes: > Any DO gurus out there? > > I'm having a problem with uniquing of proxies. I'm getting multiple > proxies to the same remote object. At first, I thought my problems > might be related to working in hybrid NX/NS environment. But I > wrote small test applications in pure NX with same results. > Here's what it looks like... [...] > My Server and Client run OK. They can ping each other, unsolicited, > in either direction. But if the server sends the client a message: > [proxyToClient heyBuddy: self]; > the proxy which the client receives as the argument to -heyBuddy: > will have a different address than proxyToServer which was originally > created when the client connected to the server. I was expecting > DO to unique the proxies, and that they would be identical. > > Any ideas? > > Thanks, > - Jim Robertson Bug or feature? I checked up this and can confirm your results. There's nothing wrong with your code. But what you reported is rather a (undocumented) feature of DO, than a bug. The id you get when connecting to a server with "connectToName:..." is a special proxy to the root object on ther server side (whis no respect to an individual object). Normally the "name" instance variable stores the "original" id of the remote object to match it with a local proxy on the server side of the connection. When an id is recieved from the server and a proxy with that name already exists, then this Proxy is returned. The NXProxy instance you recieve when connecting to a server has no name (name is 0, you can check it with the nameForProxy instance method of the NXProxy class). I think, messages to this special proxy are routed to the object that's referenced in the "rootObject" instance variable of the corresponding (server-side) NXConnection instance. Only the first time (after connecting) you recieve the root object over the connection, a new proxy will be created. From then, it will be matched correctly with the existing proxy. I hope, this will help you JCS
From: jcs@informatik.uni-koblenz.de (Juergen Schoenke) Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 18 Jan 1996 19:59:57 GMT Organization: University Koblenz / Germany Message-ID: <4dm8rt$6ov@newshost.uni-koblenz.de> References: <4d6u45$1cs@gateway.wiltel.com> In article <4d6u45$1cs@gateway.wiltel.com> jamesr@pennave (James Robertson) writes: > Any DO gurus out there? > > I'm having a problem with uniquing of proxies. I'm getting multiple > proxies to the same remote object. At first, I thought my problems > might be related to working in hybrid NX/NS environment. But I > wrote small test applications in pure NX with same results. > Here's what it looks like... [...] > My Server and Client run OK. They can ping each other, unsolicited, > in either direction. But if the server sends the client a message: > [proxyToClient heyBuddy: self]; > the proxy which the client receives as the argument to -heyBuddy: > will have a different address than proxyToServer which was originally > created when the client connected to the server. I was expecting > DO to unique the proxies, and that they would be identical. > > Any ideas? > > Thanks, > - Jim Robertson Bug or feature? I checked up this and can confirm your results. There's nothing wrong with your code. But what you reported is rather a (undocumented) feature of DO, than a bug. The id you get when connecting to a server with "connectToName:..." is a special proxy to the root object on ther server side (whis no respect to an individual object). Normally the "name" instance variable stores the "original" id of the remote object to match it with a local proxy on the server side of the connection. When an id is recieved from the server and a proxy with that name already exists, then this Proxy is returned. The NXProxy instance you recieve when connecting to a server has no name (name is 0, you can check it with the nameForProxy instance method of the NXProxy class). I think, messages to this special proxy are routed to the object that's referenced in the "rootObject" instance variable of the corresponding (server-side) NXConnection instance. Only the first time (after connecting) you recieve the root object over the connection, a new proxy will be created. From then, it will be matched correctly with the existing proxy. I hope, this will help you JCS
From: jcs@informatik.uni-koblenz.de (Juergen Schoenke) Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 18 Jan 1996 20:01:23 GMT Organization: University Koblenz / Germany Message-ID: <4dm8uj$6pl@newshost.uni-koblenz.de> References: <4d6u45$1cs@gateway.wiltel.com> Bug or feature? I checked up this and can confirm your results. There's nothing wrong with your code. But what you reported is rather a (undocumented) feature of DO, than a bug. The id you get when connecting to a server with "connectToName:..." is a special proxy to the root object on ther server side (whis no respect to an individual object). Normally the "name" instance variable stores the "original" id of the remote object to match it with a local proxy on the server side of the connection. When an id is recieved from the server and a proxy with that name already exists, then this Proxy is returned. The NXProxy instance you recieve when connecting to a server has no name (name is 0, you can check it with the nameForProxy instance method of the NXProxy class). I think, messages to this special proxy are routed to the object that's referenced in the "rootObject" instance variable of the corresponding (server-side) NXConnection instance. Only the first time (after connecting) you recieve the root object over the connection, a new proxy will be created. From then, it will be matched correctly with the existing proxy. I hope, this will help you JCS
From: jcs@informatik.uni-koblenz.de (Juergen Schoenke) Newsgroups: comp.sys.next.programmer Subject: Re: DO proxy uniquing problem Date: 18 Jan 1996 20:03:12 GMT Organization: University Koblenz / Germany Message-ID: <4dm920$6q3@newshost.uni-koblenz.de> References: <4d6u45$1cs@gateway.wiltel.com> Bug or feature? I checked up this and can confirm your results. There's nothing wrong with your code. But what you reported is rather a (undocumented) feature of DO, than a bug. The id you get when connecting to a server with "connectToName:..." is a special proxy to the root object on ther server side (whis no respect to an individual object). Normally the "name" instance variable stores the "original" id of the remote object to match it with a local proxy on the server side of the connection. When an id is recieved from the server and a proxy with that name already exists, then this Proxy is returned. The NXProxy instance you recieve when connecting to a server has no name (name is 0, you can check it with the nameForProxy instance method of the NXProxy class). I think, messages to this special proxy are routed to the object that's referenced in the "rootObject" instance variable of the corresponding (server-side) NXConnection instance. Only the first time (after connecting) you recieve the root object over the connection, a new proxy will be created. From then, it will be matched correctly with the existing proxy. I hope, this will help you JCS
From: tm8025a@newssrv.soc.american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer Subject: Posix string in source Date: 19 Jan 1996 01:55:40 GMT Organization: American University, Washington DC Message-ID: <4dmtms$2op@paladin.american.edu> I would like to include in the .c files I am compiling a string that tells the compiler that this is posix code and should be compiled as so. Just like running cc with -posix. Is there a way to do this? If so, how? Thanks a bunch -- Torrey McMahon
From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Subject: Help Date: 19 Jan 1996 02:53:48 GMT Organization: University of California, Los Angeles Message-ID: <4dn13s$13au@saba.info.ucla.edu> Could someone please help me. I have X-windows and I am trying to compile some programs, but I always get an error.... localhost> cd /private/tmp/O_xboard-3.2_pl0/xboard-3.2.pl0 localhost> configure creating cache ./config.cache checking for gcc... gcc checking whether we are using GNU C... yes checking for POSIXized ISC... no checking for a BSD compatible install... /usr/bin/install -c checking for lex... lex checking for remsh... no checking for rsh... rsh checking for nroff... nroff -man checking for awk... /bin/awk checking for perl... no checking how to run the C preprocessor... gcc -E checking whether cross-compiling... no checking for ANSI C header files... yes checking whether time.h and sys/time.h may both be included... yes checking for sys/wait.h that is POSIX.1 compatible... no checking return type of signal handlers... void checking for stropts.h... no checking for sys/time.h... yes checking for string.h... yes checking for unistd.h... yes checking for sys/systeminfo.h... no checking for fcntl.h... yes checking for sys/socket.h... yes checking for stddef.h... yes checking for _getpty... no checking for grantpt... no checking for gettimeofday... yes checking for random... yes checking for gethostname... yes checking for setlocale... yes checking for -lseq... no checking for X... libraries /usr/X11R6/lib, headers /usr/X11R6/include checking for -lICE... no checking for -ldnet... no checking for -ldnet_stub... no checking for -lnsl... no checking for -lsocket... no checking host system type... m68k-next-nextstep3 checking whether ptys or pipes should be used... pipes updating cache ./config.cache creating ./config.status creating Makefile creating comment.awk creating cmail localhost> make gcc -DPRODUCT=\"xboard\" -DVERSION=\"3.2\" -DPATCHLEVEL=\"0\" -DREMOTE_SHELL=\"rsh\" -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DRETSIGTYPE=void -DHAVE_SYS_TIME_H=1 -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_SOCKET_H=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_RANDOM=1 -DHAVE_GETHOSTNAME=1 -DUSE_PTYS=0 -DATTENTION -I. -I/usr/X11R6/include -c parser.c gcc -DPRODUCT=\"xboard\" -DVERSION=\"3.2\" -DPATCHLEVEL=\"0\" -DREMOTE_SHELL=\"rsh\" -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DRETSIGTYPE=void -DHAVE_SYS_TIME_H=1 -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_SOCKET_H=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_RANDOM=1 -DHAVE_GETHOSTNAME=1 -DUSE_PTYS=0 -DATTENTION -I. -I/usr/X11R6/include -c xboard.c gcc -DPRODUCT=\"xboard\" -DVERSION=\"3.2\" -DPATCHLEVEL=\"0\" -DREMOTE_SHELL=\"rsh\" -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DRETSIGTYPE=void -DHAVE_SYS_TIME_H=1 -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_SOCKET_H=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_RANDOM=1 -DHAVE_GETHOSTNAME=1 -DUSE_PTYS=0 -DATTENTION -I. -I/usr/X11R6/include -c backend.c gcc -DPRODUCT=\"xboard\" -DVERSION=\"3.2\" -DPATCHLEVEL=\"0\" -DREMOTE_SHELL=\"rsh\" -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DRETSIGTYPE=void -DHAVE_SYS_TIME_H=1 -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_SOCKET_H=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_RANDOM=1 -DHAVE_GETHOSTNAME=1 -DUSE_PTYS=0 -DATTENTION -I. -I/usr/X11R6/include -c moves.c gcc -DPRODUCT=\"xboard\" -DVERSION=\"3.2\" -DPATCHLEVEL=\"0\" -DREMOTE_SHELL=\"rsh\" -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DRETSIGTYPE=void -DHAVE_SYS_TIME_H=1 -DHAVE_STRING_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_SOCKET_H=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_RANDOM=1 -DHAVE_GETHOSTNAME=1 -DUSE_PTYS=0 -DATTENTION -I. -I/usr/X11R6/include -c childio.c gcc -o xboard parser.o xboard.o backend.o moves.o childio.o -L/usr/X11R6/lib -lXaw -lXmu -lXt -lXext -lX11 -lm ld: Undefined symbols: _IceConnectionNumber _IceProcessMessages _SmcClientID _SmcCloseConnection _SmcDeleteProperties _SmcGetIceConnection _SmcInteractDone _SmcInteractRequest _SmcModifyCallbacks _SmcOpenConnection _SmcRequestSaveYourselfPhase2 _SmcSaveYourselfDone _SmcSetProperties *** Exit 1 Stop. localhost> I understand this is not part of the standard NeXT Developer release, but does someone have a lib or something that I could insert somewhere in the source code to fix this!! Thanks Rick UCLA Physics
From: espeyton@ix.netcom.com (Espeyton) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Default application question Date: 19 Jan 1996 00:53:26 GMT Organization: Netcom Message-ID: <4dmq26$opv@cloner2.ix.netcom.com> Mime-Version: 1.0 Hi Y'all, I want all of my .c and .m files to open with my ColorEdit application. How can I get ColorEdit.app to show it's icon on the Inspector page to be the default application? Thanks for the help, Eric Peyton espeyton@ix.netcom.com
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Help with custom view and IB Date: 18 Jan 1996 20:20:28 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4dma2c$i6l@news.onramp.net> References: <DKvuBD.LA0@news.cis.umn.edu> klett002@maroon.tc.umn.edu (James P. Klett) writes: > The problem is, when I try to send a message to the custom view > via the id userName, I get nothing, due to the fact that userName > is NULL! Why is is NULL? Make sure you don't have a method with the same name as the ivar(instance variable) you're having IB set. In this case, if you implement a method named - userName: or - setUserName: (I forget whether it needs to have a "set" in front) you will be over-riding IBs mechanism for setting that ivar. If so, try changing the name of that method and it may fix your problem. -- Steve Dekorte Objective-C/NEXTSTEP developer http://www.marble.com/~dekorte NeXTmail, MIME welcome.
Newsgroups: comp.sys.next.programmer From: brandm@sel.sub.org (Gerd Brandmeier) Subject: Uncompleted sounddriver on HP RISC 712 Message-ID: <1996Jan18.165545.1375@sel.sub.org> Sender: news@sel.sub.org Organization: ALCATEL SEL DP/SM Date: Thu, 18 Jan 1996 16:55:45 GMT Help, the sounddriver for NeXTStep 3.3 on HP RISC 712 doesn't work completely! These functions are not implemented: - separate stauncomrt/stop of line out and speaker - change source between microphone and line in - mail the transferred bytes to the DA converter - delegate messaging from the SoundStream object: - soundStreamDidOverrun: - soundStreamDidPause: - soundStreamDidResume: Who knows a solution to this problems ? Is there any new sounddriver for HP RISC ? Has someone the sourcecode from the sounddriver ? Has someone a description of the HP hardware/soundhardware ? Thanks brandm@sel.sub.org
From: lchamp@mines1.info.emn.fr (Champciaux Laurent) Newsgroups: comp.sys.next.programmer Subject: 3D kit, any bugs ? Date: 19 Jan 1996 13:28:44 GMT Organization: Ecole des Mines de Nantes Message-ID: <4do6ac$h6@wfn.emn.fr> _Hello, I am new to the 3D kit. (NeXTCube OS 3.0) I have a little and simple program that displays some cubes and spheres. Sometimes it works and sometimes not. The simple fact of changing the number of cycles in a loop can make it crash. I m very disconcerted by the strange behavior of this program. There is certainly a big bug and I must be blind. However, I d like to know if there are some bugs in the 3D kit. If so, what are they ? Also, could someone point me to some good references about the 3D kit ? Thanks in advance for your help. Laurent_ -- ======================================================= Laurent Champciaux Departement Informatique - Ecole des Mines de Nantes 4, rue A.Kastler 44070 Nantes Cedex 03 Tel: (33)51 85 82 18 (B220) email: Laurent.Champciaux@emn.fr WWW: http://mines1.info.emn.fr:8888/
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: Asking for Help about RenderMan Date: 19 Jan 1996 15:06:54 GMT Organization: Nanyang Technological University Message-ID: <4doc2e$2vd@ntuix.ntu.ac.sg> References: <4dllls$11@nyheter.chalmers.se> This question comes up more often, so I post the answer. It is in the form of a shell script. But the comment is maybe more important then the script itself. WIllem
# # name: render # # parameter: name of .c source (without extension) # assumtion is that .rib and .tiff file have # the same name as .c source # # And your C-source will look something like the following # example. This is listing 2.2 from Renderman Companion. Do # mind the RenderType and the Ribfile. To let the shell-script # work, the C-source, the RibFile and the TiffFile should all # have the same name (except extensions). # # author: Willem van Schaik, 1994-1996 # ############################################################################## # # include <ri/ri.h> # # RtPoint Square [4] = {{.6, .6, 0}, {.6, -.6, 0}, {-.6, -.6, 0}, {-.6, .6, 0}}; # RtColor SquareColor = {.2, .5, .8}; # RtColor SphereColor = {.8, .5, .2}; # # main(int argc, char *argv[]) # { # char *RenderType = "archive"; # char *RibFile = "Lst22.rib", *TiffFile = "Lst22.tiff"; # # RiBegin(RI_NULL, "renderer", &RenderType, # "filepath", &RibFile, RI_NULL); # RiDisplay (TiffFile, RI_FILE, RI_RGBA, RI_NULL); # # RiFrameBegin (1); # RiFormat (512, 512, 1); # # RiLightSource ("mylight", "distantlight", RI_NULL); # # RiProjection ("perspective", RI_NULL); # RiTranslate (0.0, 0.0, 1.2); RiRotate (45.0, -1.0, 1.0, 0.0); # # RiWorldBegin (); # # RiSurface ("matte" , RI_NULL); # RiColor (SquareColor); # RiPolygon (4, RI_P, (RtPointer) Square, RI_NULL); # # RiSurface ("matte", RI_NULL); # RiColor (SphereColor); # RiSphere (.5, -.4, .4, 360, RI_NULL); # # RiWorldEnd (); # RiFrameEnd (); # RiEnd (); # # } # ############################################################################## if ($1 == '') then echo 'syntax: render <file>'; exit endif if (-f $1'.c') then cc -o $1 $1.c -lMedia_s -lNeXT_s $1 /usr/prman/prman $1.rib /NextApps/Preview.app/Preview $1.tiff else echo 'file '$1'.c not found' endif exit
Nader Nazari,1018 (nazari@cs.chalmers.se) wrote: : Hi : I work on a project which needs a lot of 3D images. : I start to work with the N3D library, but 3d documentation says : "This documentation assume you are familiar with the RenderMan language..." : If you are not a book called "The RenderMan Companion .. " is recommender. : I bought the book and now I try to run the first program in the book. : #include <ri/ri.h> : RtPoint Square[4] = {{.5,.5,.5},{.5,-.5,.5},{-.5,.5,.5}}; : main() : { : RiBegin(RI_NULL); : RiWorldBegin(); : RiSurface("constant",RI_NULL); : RiPolygon(4,RI_P,(RtPointer)Square,RI_NULL); : RiWorldEnd(); : RiEnd(); : } : I succeeded to compile this program by using "cc" but how can I run it. : To run the program I need a window ( it does not create it) to create : a window and using RenderMan I must use N3D Kit. To use N3D kit I have to : know RenderMan and to learn RenderMan I must use the book and thus run the : program!! : Is any body there who can save me out of this loop. : Tanks : Nader <nazari@cs.chalmers.se>
Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Default application question In-Reply-To: espeyton@ix.netcom.com's message of 19 Jan 1996 00:53:26 GMT Message-ID: <RDL.96Jan19152010@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4dmq26$opv@cloner2.ix.netcom.com> Date: Fri, 19 Jan 1996 20:20:10 GMT You need to register a document extension in the Attributes section of your PB.project. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4dmq26$opv@cloner2.ix.netcom.com> espeyton@ix.netcom.com (Espeyton) writes: From: espeyton@ix.netcom.com (Espeyton) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Date: 19 Jan 1996 00:53:26 GMT Organization: Netcom Hi Y'all, I want all of my .c and .m files to open with my ColorEdit application. How can I get ColorEdit.app to show it's icon on the Inspector page to be the default application? Thanks for the help, Eric Peyton espeyton@ix.netcom.com
From: tm8025a@newssrv.soc.american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer Subject: EOF 1.1 and Oracle 7 problems Date: 20 Jan 1996 02:41:32 GMT Organization: American University, Washington DC Message-ID: <4dpkos$aib@paladin.american.edu> I am using NextStep 3.3 pl 1, EOF 1.1, ORACLE 7 ver 2 type logins and NeXT hardware. I can't get EOF to work for one second let alone get a working model. I go to Open Model and type the following in the fields: Server ID--> t:pythia:host.domain Host Machine--> "Nothing" User Name---> "My idea" Password---> "My password" The first time I tried the above it took a few seconds to get the system beep but after that it only beeps. Is that all I need or is anything else needed? Any clues would help. -- Torrey McMahon
From: bbum@friday.com Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: (Off-Topic:) Re: Objective-C Coding Standards Date: 20 Jan 1996 15:37:44 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4dr288$jd4@chinx4.thoughtport.net> References: <4dedp9$frm@flowbee.interaccess.com> <4dhvtv$9qn@ixnews2.ix.netcom.com> <4djjcu$irn@news.its.com> <4dluh3$r19@zeus.ai-lab.fh-furtwangen.de> # by Axel Eble, FH Furtwangen, AI Lab Chuck Swiger (chuck@its.com) wrote: : espeyton@ix.netcom.com (Eric Peyton) wrote: : > I would definitely recommend getting a copy of Code Complete by Steve : > McConnell. [...] : > [...] Oh, and sorry, the book is published by Microsoft Press. : : I would strongly second that recommendation. If you develop software or : manage people who do, you really should read this book. # so, if it's published by MicroSoft, why don't *THEY* read it? # ;o) Actually, a lot of the inspiration for the book came from the author's experiences working at Microsoft. [the book also has many other well-documented references to various disasters throughout the software industry] BTW: The book is basically written from the standpoint that 90% of current software projeects are flawed in some manner... and that most managers and project leaders couldn't correctly complete a software project to save their life. It is an *awesome* book-- I would reccomend it to *any* programmer and just about anyone else who is doing some kind of complex production job. : -Chuck # Axel b.bum
Newsgroups: comp.sys.next.programmer From: tomi@shinto.nbg.sub.org (Thomas Engel) Subject: Re: 3D kit, any bugs ? Message-ID: <DLHny7.I8@shinto.nbg.sub.org> Sender: tomi@shinto.nbg.sub.org (Thomas Engel) Organization: STEPeople's home (A NUGI member) References: <4do6ac$h6@wfn.emn.fr> Date: Sat, 20 Jan 1996 16:46:54 GMT In article <4do6ac$h6@wfn.emn.fr> lchamp@mines1.info.emn.fr (Champciaux Laurent) writes: > _Hello, > > I am new to the 3D kit. (NeXTCube OS 3.0) > > I have a little and simple program that displays some cubes and spheres. > Sometimes it works and sometimes not. The simple fact of changing the number of > cycles in a loop can make it crash. I m very disconcerted by the strange > behavior of this program. There is certainly a big bug and I must be blind. > However, I d like to know if there are some bugs in the 3D kit. If so, what are > they ? Well. NS 3.3 comes with a 3DKit that is quite stable. I had many crashes with big scenes under NS 3.0/3.2 but under 3.3 it works pretty stable on a NeXTdimension. > Also, could someone point me to some good references about the 3D kit ? > Besides the documenttion there are some apps that use the kit and have free source code. One of the most interesting should be WavesWorld which you should be able to locate on mit.edu (sorry forgot the real pointer :-(..but it is on the Stone CD too) You might also contact the folks at IMAG/Grenoble...they are in France and have done a lot of 3DKit work in thir labs. Aloha Tomi -- _________________________________________________________ (tomi@shinto.nbg.sub.org) Thomas Engel Neptunstr. 9 NeXTMail welcome D - 90522 Oberasbach Germany
From: woodward@onramp.net (John Woodward) Newsgroups: comp.sys.next.programmer Subject: Re: Objective-C Coding Standards Date: 20 Jan 1996 18:13:55 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4drbd3$l09@news.onramp.net> References: <4dgi2u$hb4@chinx4.thoughtport.net> In article <4dgi2u$hb4@chinx4.thoughtport.net> bbum@friday.com writes: > > C++ programmers shouldn't bother reading it; the elegance and dynamic > nature of Obj-C will only depress the reader... :-) I agree. I'm working on a really cool C++ project to provide an easy parallel computing framework for image processing. Whenever we stumble on some feature we want to add, I keep grumbling "if only we were doing this in Objective-C". > Seriously, last year I read +Code Complete+ and re-read the +Obj-C & OO > Programming+ book and my code improved in quality immensely-- and that's > after 5 years of NEXTSTEP+ObjC programming... I'll have to pick that one up and look it over. Sounds pretty good. john -- ===================================#################################################################### Path: news.informatik.uni-muenchen.de!lrz-muenchen.de!informatik.tu-muenchen.de!Germany.EU.net!howland.reston.ans.net!vixen.cso.uiuc.edu!ux1.cso.uiuc.edu!indy From: indy@ux1.cso.uiuc.edu (weintz steven cortelou) Newsgroups: comp.sys.next.programmer Subject: Re: Is NeXT Developer C++ compatible? Date: 21 Jan 1996 21:31:12 GMT Organization: University of Illinois at Urbana Lines: 45 Message-ID: <4dubb0$8qm@vixen.cso.uiuc.edu> References: <3102894E.4461@home.mysolution.com> NNTP-Posting-Host: ux1.cso.uiuc.edu "TOMMY K. HWANG" <thwang@home.mysolution.com> writes: >I am trying out C++ and am wondering if NeXT >Developer is really C++ capable? The man pages says >it is, but it does not want to compile simple program >below: [...] > The error message is below: > next1> cc finalgrade.cc > ld: Undefined symbols: > ostream::operator<<(char const *) > ostream::operator<<(ostream &(*)(ostream &)) > ostream::operator<<(float) > istream::operator>>(float &) > cin > cout > endl(ostream &) > next1> > What might my problem be? Thanks... I'm having similar problems compiling Muhummad Muquit's Homepage Access Counter. After running the cinfigure and install scripts, the initial components build fine but then I get this: gryphon> make (cd combine; make) cc -g -I. -I./combine -DALLOW_FILE_CREATION=1 -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DHAVE_STRING_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_FILE_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FLOCK=1 -DACCESS_AUTH -D__USE_FIXED_PROTOTYPES__ -o Count.cgi main.o parse.o strimage.o ./combine/libCombine.a ld: Undefined symbols: _putenv *** Exit 1 Stop. gryphon> The author is unfamliar with NS. Any suggestions? -- Steve Weintz * Partner, New Media, Ltd. P.O. Box 742 * Urbana, IL 61801 steve@dave-world.net * 217.344.5303 * 217.344.8981 fax Staff Visioneer, Beckman Institute Visualization Facility * indy@uiuc.edu
Newsgroups: comp.sys.next.programmer From: jspool@uie.com (Jared M. Spool) Subject: Drag & Drop: Usability Information Message-ID: <DLK3xp.CuL@world.std.com> Sender: news@world.std.com (Mr Usenet Himself) Organization: User Interface Engineering Date: Mon, 22 Jan 1996 00:26:07 GMT Recently, there has been some discussion on implementing Drag & Drop. I thought people would be interested in this article which recently appeared in our newsletter, Eye For Design. (See end of the article for information on getting a free issue.) DRAG & DROP HAS A LEARNING PROBLEM In testing the designs of several client's products, we've observed several issues related to how users learn drag and drop operations. While drag and drop can improve a user's productivity considerably, if they can't learn how to use it, the benefits will not be realized. Here are some of the problems we've observed. Users May Not Expect It User expectations are often shaped by other products that they've worked with. However, we've observed that drag and drop experiences always seem to translate well to new products. For example, just because users have dragged files in the Windows 3.1 File Manager doesn't mean that they look for drag and drop capability in some other application. Users don't see the File Manager as an application. (It appears as a piece of a bigger thing -- the operating environment.) Therefore, they don't seem to connect that an application's interface might be similar to the File Manager. In testing other applications, we've seen that users are often not aware of drag and drop methods, and have to learn about them in each application. Since every application that implements drag and drop seems to implement it slightly differently, it's no wonder that users have trouble transferring knowledge from one application to another. It's Only A Shortcut We've observed that new users typically learn product functions from menus or toolbars then "graduate" to faster methods such as double-clicking and drag and drop. If there is functionality that can only be accomplished by drag and drop, users may never find it at all. We saw this in the Lotus 1-2-3 dialog box for editing SmartIcons -- users didn't realize they were supposed to drag icons from one list box to another. Four Things To Learn Users must learn four concepts to use drag and drop effectively: 1) "What objects can I drag?" 2) "Where can I drop them?" 3) "What's it going to do when I let go?" 4) "If I don't like it, how do I undo it?" No Clues The first two questions arise when clues are missing. Drag and drop isn't readily visible. Users must learn that it exists through some external mechanism. One development team we worked with tried a visual cue, using a small blue square to indicate a draggable object. Users didn't get it. The team then tried another tack: illustrating the drag and drop capability on a mockup of the product box. This worked better -- users picked up on the drag and drop capability and used it successfully. Caveat: we don't know whether in real life users would have looked at the box! What's The Verb? The "what's it going to do?" issue arises because the verb is implied by a drag and drop operation often depends on the context. Consider the Windows 3.1 File manager, where dragging a file to a directory on the same disk moves the file, but dragging a file to a different disk copies it. Same drag and drop, but two different verbs are implied depending on the relationship between the source and destination. Cursor icons can provide clues, but users must still learn what the different icon mean. Microsoft Publisher tries an interesting idea: attaching words to cursor icons. For example, the "Move" icon shows a picture of a truck with the word "Move" on it. Although we have not tested drag and drop in Publisher, we do have ample evidence that an icon and word approach improves recognition for toolbar icons. How Do You Undrop? Since drag and drop can have destructive consequences, users need a way to undo drag and drop they didn't intend. In one test, we saw several users accidentally destroy spreadsheet charats as follows: * User wants to move entire chart, but positions the mouse insed the chart and drags just part of the chart * User notices error an makes an unsuccessful attempt to drag it back the way it was * User tries undo, but it only undoes the last action (the attempted correction, not the accidental drag) * After a few more attempts to correct the chart manually, user deletes the chart and re-creates it Thus, adding drag and drop capability to an interface may increase the need for verification questions and/or multiple levels of undo. ============================================= User Interface Engineering produces a bimonthly newsletter called Eye For Design, which contains the latest tips and techniques for developing excellent applications. You'll read about the latest research we've done on how to build usable applications. If you send us your postal address, we'll send you a complimentary issue. ============================================= (c) 1995, 1996 User Interface Engineering ========================================================== Jared M. Spool User Interface Engineering jspool@uie.com 800 Turnpike Street, Suite 101 (508) 975-4343 North Andover, MA 01845 fax: (508) 975-5353 USA If you send me your postal address, you'll get the next issue of our newsletter, Eye For Design. ==========================================================
Newsgroups: comp.sys.next.programmer From: dkoski@netcom.com (David Koski) Subject: Re: Is NeXT Developer C++ compatible? Message-ID: <dkoskiDLK78A.A0A@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <3102894E.4461@home.mysolution.com> Date: Mon, 22 Jan 1996 01:38:34 GMT Sender: dkoski@netcom19.netcom.com In article <3102894E.4461@home.mysolution.com>, TOMMY K. HWANG <thwang@home.mysolution.com> wrote: >I am trying out C++ and am wondering if NeXT >Developer is really C++ capable? The man pages says >it is, but it does not want to compile simple program >below: ... > next1> cc finalgrade.cc > ld: Undefined symbols: > ostream::operator<<(char const *) > ostream::operator<<(ostream &(*)(ostream &)) > ostream::operator<<(float) > istream::operator>>(float &) > cin > cout > endl(ostream &) You just need to include the proper libraries. Try this: cc finalgrade.cc -lg++ This worked for me with your example program. There is a cc++, and you might think it would include libg++ for you, but it does not, so you could also type: cc++ finalgrade.cc -lg++ And it seems to do the same thing. David dkoski@netcom.com
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 22 Jan 1996 05:15:11 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4dv6gv$7oi@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Computer, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 188+ ISV company pages - 433+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Computer, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: cnyap@dcs.shef.ac.uk (Chih Nam Yap) Newsgroups: comp.sys.next.programmer Subject: How to display an icon in a view Date: 22 Jan 1996 12:38:12 GMT Organization: Department of Computer Science, University of Sheffield Message-ID: <4e00fk$far@bignews.shef.ac.uk> Recently, I have created a simple application to perform drag-and-drop, but I faced problem in display images on the destination view, I think the reason was I do not convert the hot-spot cursor point properly, can anyone help me to rectify my code please ? What I have implemented was, I have a matric of action cells (which I called icon menu), each cell has an icon image, just like the file path in the File Viewer application. There is another custom view (which I called workingView) situated just below the matric of cells. ------------------------------------------------------------------------- | My Window | |------------------------------------------------------------------------ | | | | | ------------------------------------------------------------------ | | | ------- ------- -------- | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ------- ------- -------- | | | | icon 1 icon 2 icon 3 | | | ------------------------------------------------------------------- | | | | ------------------------------------------------------------------ | | | | | | | | | | | | | | | workingView | | | | | | This is the code of workingView @implementation VZWorking - initFrame:(const NXRect *)theFrame { NXAtom *pasteTypes = VZIDPasteTypes(); [super initFrame:theFrame]; [self registerForDraggedTypes: pasteTypes count: 1]; return self; } - (NXDragOperation) draggingEntered: sender { if ([sender draggingSource] != self) { return NX_DragOperationCopy; } return NX_DragOperationNone; } - (NXDragOperation) draggingUpdated: sender { if ([sender draggingSource] != self) { // NXPoint point declared in workingView.h file point = [sender draggingLocation]; [self convertPoint: &point fromView: nil]; return NX_DragOperationCopy; } return NX_DragOperationNone; } - draggingExited: sender { return self; } - (BOOL) prepareForDragOperation: sender { return YES; } - (BOOL) performDragOperation: sender { return YES; } - concludeDragOperation: sender { id newCopy; // read a pointer from the pasteboard dragObject = [[VZIcon class] readFromPasteboard: [sender draggingPasteboard]]; // NXImage newImage declared in workingView.h file newCopy = [dragObject copy]; newImage = [newCopy image]; [newImage composite: NX_SOVER toPoint: &point]; [self display]; return self; } @end ------------------------------------------------------------------------------ The current situation is, when I drag an icon from the icon menu, say icon1, I drag icon1 into the workingView, move around, the cursor has sucessfully changed into "copy" type. When I relieve the mouse button, the icon1 image instead of being drawn at the point where I relieve the mouse button, it appears inside the icon menu which is not what I want. If I resize the window, the new image will go, this prove that the image has drawn on the window instead of the icon menu. I have tested the readFromPasetboard, it is OK. Can anyone give suggestions to help me to solve the problem? Thank you. Cheers, C.Yap
From: jehu@jehu.async.vt.edu (john stanhope) Newsgroups: comp.sys.next.programmer Subject: Re: Is NeXT Developer C++ compatible? Date: 22 Jan 1996 14:37:29 GMT Organization: Virginia Tech, Blacksburg, Virginia Distribution: world Message-ID: <4e07f9$1oj@solaris.cc.vt.edu> References: <4dubb0$8qm@vixen.cso.uiuc.edu> In article <4dubb0$8qm@vixen.cso.uiuc.edu> indy@ux1.cso.uiuc.edu (weintz steven cortelou) writes: ->"TOMMY K. HWANG" <thwang@home.mysolution.com> writes: -> ->>I am trying out C++ and am wondering if NeXT ->>Developer is really C++ capable? The man pages says ->>it is, but it does not want to compile simple program ->>below: -> ->[...] -> ->> The error message is below: -> ->> next1> cc finalgrade.cc ->> ld: Undefined symbols: ->> ostream::operator<<(char const *) ->> ostream::operator<<(ostream &(*)(ostream &)) ->> ostream::operator<<(float) ->> istream::operator>>(float &) ->> cin ->> cout ->> endl(ostream &) ->> next1> -> ->> What might my problem be? Thanks... -> ->I'm having similar problems compiling Muhummad Muquit's Homepage ->Access Counter. After running the cinfigure and install scripts, ->the initial components build fine but then I get this: -> ->gryphon> make ->(cd combine; make) ->cc -g -I. -I./combine -DALLOW_FILE_CREATION=1 -DSTDC_HEADERS=1 -DTIME_WITH_SYS_TIME=1 -DHAVE_STRING_H=1 -DHAVE_FCNTL_H=1 -DHAVE_SYS_FILE_H=1 -DHAVE_UNISTD_H=1 -DHAVE_FLOCK=1 -DACCESS_AUTH -D__USE_FIXED_PROTOTYPES__ -o Count.cgi main.o parse.o strimage.o ./combine/libCombine.a ->ld: Undefined symbols: ->_putenv ->*** Exit 1 ->Stop. ->gryphon> -> ->The author is unfamliar with NS. Any suggestions? -> You need the putenv function, since it doesn't come in any of NS's libraries. Here's a GNU version of the function. /* Copyright (C) 1991 Free Software Foundation, Inc. This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License for more details. You should have received a copy of the GNU Library General Public License along with the GNU C Library; see the file COPYING.LIB. If not, write to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ #include <errno.h> #include <stdlib.h> #include <string.h> /* The following file does not exist in NEXSTEP 2.1. * The omission of this include does not generate an error * in NEXTSTEP versions 2.1,3.1,3.2 brd Fri Jan 21 22:54:38 PST 1994 */ #ifndef NeXT #include <unistd.h> #endif extern char **environ; #define __environ environ /* Put STRING, which is of the form "NAME=VALUE", in the environment. */ int putenv(const char *string) { char *name_end = strchr(string, '='); register size_t size; register char **ep; if (name_end == NULL) { /* Remove the variable from the environment. */ size = strlen(string); for (ep = __environ; *ep != NULL; ++ep) if (!strncmp(*ep, string, size) && (*ep)[size] == '=') { while (ep[1] != NULL) { ep[0] = ep[1]; ++ep; } *ep = NULL; return 0; } } size = 0; for (ep = __environ; *ep != NULL; ++ep) if (!strncmp(*ep, string, name_end - string) && (*ep)[name_end - string] == '=') break; else ++size; if (*ep == NULL) { static char **last_environ = NULL; char **new_environ = (char **) malloc((size + 2) * sizeof(char *)); if (new_environ == NULL) return -1; (void) memcpy((void *) new_environ, (void *) __environ, size * sizeof(char *)); new_environ[size] = (char *) string; new_environ[size + 1] = NULL; if (last_environ != NULL) free((void *) last_environ); last_environ = new_environ; __environ = new_environ; } else *ep = (char *) string; return 0; } -- John Stanhope jehu@vt.edu
From: cnyap@dcs.shef.ac.uk (Chih Nam Yap) Newsgroups: comp.sys.next.programmer Subject: Do I need "setFlipped" Date: 22 Jan 1996 17:00:17 GMT Organization: Department of Computer Science, University of Sheffield Message-ID: <4e0fr1$k64@bignews.shef.ac.uk> Hi there, Supposing I have a custom view called myView and it have been grouped by a scrollView. And the scrollView has vertical scroll bar only. Do I need to call the statement [self setFlipped: YES]; inside "myView" in order to get the y-coordinate start from the bottom left corner ? If this is the case, does it mean that if a custom view has been grouped by a scrollView with veritcal scroll bar, the custome view's origin will be "automatically" shifted to the top left corner ? Can anyone help ? I am a bit confused now. Thank you very much. Cheers, C.Yap
From: johnz@shellx.best.com (John Zollinger) Newsgroups: comp.sys.next.programmer Subject: Re: EOF 1.1 and Oracle 7 problems Date: 22 Jan 1996 11:09:01 -0800 Organization: Best Internet Communications Message-ID: <4e0ncd$iag@shellx.best.com> References: <4dpkos$aib@paladin.american.edu> Torrey McMahon (tm8025a@newssrv.soc.american.edu) wrote: : I am using NextStep 3.3 pl 1, EOF 1.1, ORACLE 7 ver 2 type logins and NeXT : hardware. I can't get EOF to work for one second let alone get a working : model. : I go to Open Model and type the following in the fields: : Server ID--> t:pythia:host.domain : Host Machine--> "Nothing" : User Name---> "My idea" : Password---> "My password" : The first time I tried the above it took a few seconds to get the system beep : but after that it only beeps. : Is that all I need or is anything else needed? Any clues would help. Well, you have to set up the connection and port information in netinfo, not to mention the orasrv stuff on the server. Check NextAnswers for more info. John Zollinger ProActive Consulting, Inc.
From: curt@lexis-nexis.com (Curt Kohler) Newsgroups: comp.sys.next.programmer Subject: Re: EOF & Oracle Date: 22 Jan 1996 20:54:03 GMT Organization: Lexis-Nexis, Dayton OH Distribution: world Message-ID: <4e0thb$1k9@meaddata.lexis-nexis.com> References: <4dekr0$h4m@ixnews6.ix.netcom.com> <4du1r3$sv3@sarajevo.seitz.de> Try checking out NeXT Answer # 1897. You may also want to look and see what version of SQLNet you are using. V2 requires another definition in /etc/services and supplying different info at the login panel... Curt In article <4du1r3$sv3@sarajevo.seitz.de>, adnan@seitz.de (Adnan Mumbasic) writes: |> verhagk@ix.netcom.com(Ken Verhagen ) wrote: |> |> >I am trying to connect to an Oracle server running on |> >a Sparc running Solaris 2.4 (Oracle 7.2.2.3). When |> >I try to connect using either the EOModeler or DBModeler |> >I can't seem to get by the login screen. (It keeps returning |> >to the login screen without any error) |> |> >I have tried running the obsolete orasrv process on the |> >server side and adding the following line to the etc/services file |> >orasrv 1525/tcp oracle |> >It doesn't seem to work. |> |> >Is there anyone out there who has been able to connect to |> >this version of Oracle (or any other version) using EOModeler? |> |> >I would appreciate any help. |> |> >I can be reached by either replying to this posting or by |> >email verhagk@ix.netcom.com |> |> >Thanks |> |> >Ken Verhagen |> |> |> Hi, |> |> I have the same problem. If you here something please let me know. |> |> |> Thnx Adnan |> |> adnan@seitz.de |> -- Curt Kohler (513) 865-1106 Lexis-Nexis Fabrication Systems - IT 9443 Springboro Pike curt@lexis-nexis.com Miamisburg, Ohio USA 45342 ...!uunet!lexis-nexis!curt
From: jim@ks.uiuc.edu (Jim Phillips) Newsgroups: comp.sys.next.programmer Subject: probing NeXT printer Date: 22 Jan 1996 21:22:41 GMT Organization: University of Illinois at Urbana Message-ID: <4e0v71$dj5@vixen.cso.uiuc.edu> Hi, Does anyone have an example of code which can probe a NeXT printer to determine if it is on? I've tried writing my own using ioctl() but I can't seem to make it work properly. What I am trying to do is create a cron job which will turn off my printer if it is inactive since I seldom print anything. I can't use "nppower off" because that first turns the printer on and then on again, which is not what I want to happend. Any ideas? Thanks. -JCP
From: tm8025a@newssrv.soc.american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer Subject: Re: EOF & Oracle Date: 23 Jan 1996 00:59:45 GMT Organization: American University, Washington DC Distribution: world Message-ID: <4e1bu1$g6b@paladin.american.edu> References: <4e0thb$1k9@meaddata.lexis-nexis.com> Just wanted to say thanks to all the people that responded with the correct information. (Its NextAnswers #1897 for those not keeping track.) -- Torrey McMahon
From: espeyton@ix.netcom.com (Eric Peyton) Newsgroups: comp.sys.next.programmer Subject: breakTable and clickTable experience anyone? Date: 23 Jan 1996 00:02:46 GMT Organization: Netcom Message-ID: <4e18j6$q7o@ixnews8.ix.netcom.com> Has anyone either used or implemented a usage of the clickTable or breakTable in Developer 3.3. I wish to use these objects in an editing application to find out when a word has been finished, can anyone help? If you need more specific information I will gladly provide it. I would love code snippets or a pointer to documentation about the usage of these object - General Reference Volume 1 is very vague about them. Thanks alot in advance, Eric Peyton
From: Howard Ship Newsgroups: comp.sys.next.programmer Subject: NeXT Tiff to gif w/ alpha Date: 22 Jan 1996 20:19:31 GMT Organization: Stratus Computer Inc, Marlboro MA Distribution: us Message-ID: <4e0rgj$qck@transfer.stratus.com> Can anyone direct me to a tool for converting NeXT TIFF images, with alpha, to GIF format (maintain alpha). OmniImage / OmniImageFilter can convert TIFF to GIF, but seems to loose alpha. A command line tool, especially one that performs other operations (especially interlacing, which looks best during incremental load of inline Web images) would be very nice. I did some hunting around the Web and found some tools for transparent GIF, but nothing specically for NeXT TIFF -> GIF. -- Let me do my TRIBUTE to FISHNET STOCKINGS... Howard Ship <hship@cac.stratus.com> `88 FZR1000 (Zed) [NeXT Mail 3.3 / MIME Mail OK]
From: Kir Royale Limited <kiroyale@kiroyale.demon.co.uk> Newsgroups: comp.sys.next.programmer Subject: NeXTSTEP developers needed for UK work Date: Tue, 23 Jan 1996 12:39:35 GMT Organization: Kir Royale Limited Distribution: world Message-ID: <838528490wnr@kiroyale.demon.co.uk> If you are interested in working in the UK on either a contract or a permanent basis then we *must* talk. I am looking for a large number of NeXTSTEP developers for four separate clients based in and around central London. All candidates must have at least 18 months NeXT experience. I am also looking for between 8 and 10 Analysts to work in the UK travelling and implementing systems throughout Europe. Contact: Duncan Campling All resumes/messages in plain text or uuencoded Word 6 formats only. Kir Royale Tel: +44 (0)181 224 6868 Fax: +44 (0)181 224 6767 E-Mail: kiroyale@kiroyale.demon.co.uk
From: griffith@crl.com (Dave Griffith) Newsgroups: comp.sys.next.programmer Subject: Re: breakTable and clickTable experience anyone? Date: 23 Jan 1996 07:04:07 -0800 Organization: The Hermetic and Esoteric Order of Clarity and Openness Message-ID: <4e2td7$bdq@crl.crl.com> References: <4e18j6$q7o@ixnews8.ix.netcom.com> In article <4e18j6$q7o@ixnews8.ix.netcom.com>, Eric Peyton <espeyton@ix.netcom.com> wrote: >Has anyone either used or implemented a usage of the clickTable or breakTable >in Developer 3.3. I wish to use these objects in an editing application to >find out when a word has been finished, can anyone help? If you need more >specific information I will gladly provide it. I would love code snippets or >a pointer to documentation about the usage of these object - General >Reference Volume 1 is very vague about them. > After chasing down all the clickTable and breakTable information I could find (very little), and much painful debugger work, I finally came to the conclusion that clickTable and breakTable are simply not usable by programmers outside of NeXT. The workaround I came up with was to subclass Text and override the mouseDown: method to do the selection manipulation I needed. (It looks like you'd need to override keyDown: instead). Not pretty, but it works, and it's documented. -- --Dave Griffith, griffith@crl.com
From: david@pfi.ibk.baum.ethz.ch (David EKCHIAN) Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: Help: Malformed Mach-o file. Date: 23 Jan 1996 16:26:38 GMT Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4e327u$3bt@elna.ethz.ch> Hi all, I use NEXTSTEP 3.3 on a SPARC Station 20 and my code is a mixture of Objective-C (for the interface) and C++ (for the back-end). I often get the message "Malformed Mach-o file" after linking my App. If I link again (always using Project Builder), it works fine! Any explanation for this? Many thanks, David. -- o _ /-;c __________________________________________________________________(@)#\(@)___ _/_/_/_/ _/_/_/_/ _/ _/ _/_/_/_/ David C. EKCHIAN _/ _/ _/ _/ _/ _/_/_/ _/ _/_/_/_/ _/ IBK-Informatik(HIL) _/ _/ _/ _/ _/ Phone: ++41-1-633 31 10 _/_/_/_/ _/ _/ _/ _/_/_/_/ Email: david@pfi.ibk.baum.ethz.ch _____________________________________________________________________________ Swiss Federal Institute of Technology Zurich - Inst. Structural Engineering _____________________________________________________________________________
From: yibwu@napoleon.plato.sky.bdm.com (Yibing Wu) Newsgroups: comp.sys.next.programmer Subject: Change DataSource for EOController, Can I? Date: 23 Jan 1996 21:36:10 GMT Organization: BDM International, Inc. Message-ID: <4e3kca$jru@news.mcl.bdm.com> Hi, there: I have 2 EOModels built on 2 database instances, some entities in both have same names. What I was trying to do, at first, was to change the EODatabaseDataSource for EOController dynamically(on the fly). So, I generated 2 EOModels from EOModeler, created templates for them and saved them. In my code, once a button is clicked, I will [alloc] a EODatabaseDataSource object and init with the message [initWithModel: entityName]. Then set the EOController object's DataSource to this new instance by sending [setDataSource]. However, it give me some unexpected errors: since the UI objects are linked to EOModel 1 by default, so when I call [initWithModel: entityName] and initialize it with EOModel 2, it give me an error: cannot Incorporate EOModel2, because conflicting entity named "blah blah blah". As some of you may be able to tell, what I am trying to achieve is certain level of data source transparency---maybe I've taken the word too literally, but as a new comer to the grandios NeXT, I really don't know what was going on. Help, please. Thanks, Yibing Wu
Newsgroups: comp.sys.next.programmer From: klett002@maroon.tc.umn.edu (James P. Klett) Subject: Transparent Icons, Again... Message-ID: <DLns4G.B6J@news.cis.umn.edu> Sender: news@news.cis.umn.edu (Usenet News Administration) Organization: UofM Alumni Date: Wed, 24 Jan 1996 00:04:37 GMT I know, I've seen this one many times, I even saved the last article on this subject, but I couldn't get it to work. There is a check box in the color panel called Paint in Overlay mode, but it is grey'd out for my Icon. So, I can't use this function. Once again, How do you create an Icon with transparent parts to it? Thanks, JIM ___ ___....-----'---'-----....___ ========================================= ___'---..._______...---'___ (___) _|_|_|_ (___) \\____.-'_.---._'-.____// klett002@maroon.tc.umn.edu
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: probing NeXT printer Date: 24 Jan 1996 04:09:10 GMT Organization: Nanyang Technological University Message-ID: <4e4bd6$hg6@ntuix.ntu.ac.sg> References: <4e0v71$dj5@vixen.cso.uiuc.edu> Jim Phillips (jim@ks.uiuc.edu) wrote: : Does anyone have an example of code which can probe a NeXT printer to : determine if it is on? I've tried writing my own using ioctl() but I can't : seem to make it work properly. Problem has been solved before :-) Willem --------------------------------------------- Hi printers, Lately there were some messages about how to switch off the NeXTprinter, without having to switch it on manually. The method I picked up from the net, one or two years ago is to start a deamon from rc.local. In my case (I think a rather typical one) this method works flawless. Once again, thanx to the original authors. Willem van Schaik ---8X====---------------------------------------------------nppowerd.README--- Compile nppowerd.c, copy it to /usr/local/bin, take care that root is the owner and call it from /etc/rc.local. The following works fine: --------- (echo -n 'Starting local daemons:') >/dev/console # # Run your own commands here # # Start NextPrinter power-off deamon if [ -f /usr/local/bin/nppowerd ]; then /usr/local/bin/nppowerd 10 & (echo -n ' nppowerd') >/dev/console fi # (echo '.') >/dev/console --------- created by: Peter Nitezki (Nitezki@NiDat.sub.org) modified by: Piotr Twarecki, M.Eng. (B7NS@MUSICB.McGill.CA) ---8X====---------------------------------------------------nppowerd.README--- ---8X====--------------------------------------------------------nppowerd.c--- /* * nppowerd.c rjesse 19 oct 92 * * This daemon turns off a NeXT 400 dpi printer (/dev/np0) * that has been idle for more than "interval" minutes. * * Usage: nppowerd [interval minutes] */ #define INTERVAL (60*60) /* 60 minute default */ #include <stdio.h> #include <bsd/sys/types.h> #include <mach/boolean.h> #include <bsd/sys/ioctl.h> #include <bsd/sys/file.h> #include <bsd/sys/stat.h> #include <bsd/dev/m68k/npio.h> char printer[] = "/dev/np0"; time_t pmtime() { struct stat pstat; if (stat(printer, &pstat) < 0) { fputs("nppowerd: ", stderr); perror(printer); exit(1); } return pstat.st_mtime; } turnpoff() { struct npop op; register pfd; extern errno; errno = 0; if ((pfd=open(printer, O_WRONLY, 0)) < 0) return errno; op.np_op = NPSETPOWER; op.np_power = 0; ioctl(pfd, NPIOCPOP, &op); close(pfd); return errno; } main(argc, argv) int argc; char **argv; { time_t offtime, mtime, idle; register interval = INTERVAL; if (argc > 2) { fputs("Usage: nppowerd [minutes]\n", stderr); return 1; } switch (fork()) { default: return 0; case -1: return 1; case 0: ; /* child */ } if (argc > 1) interval = atoi(argv[1]) * 60; if (interval<60 || interval>(24*60*60) ) interval = INTERVAL; /* * it seems (see NP(4)) to be impossible to test whether the * printer is on or off without first turning it on - opening * the device, even without O_WRONLY, turns it on. so we assume * that upon nppowerd invocation (probably from /etc/rc) that * the printer is on (it gets turned on during the boot sequence). */ offtime = 0; mtime = pmtime(); /* * go to sleep during startup to allow the printer daemon * to start up */ sleep(15); while (1) { if (mtime != offtime) { /* there has been some printer activity */ while ((idle = time() - mtime) < interval) { sleep(interval-idle+1); mtime = pmtime(); } if (turnpoff() == 0) offtime = mtime = pmtime(); } sleep(interval); mtime = pmtime(); } /*NOTREACHED*/ } ---8X====--------------------------------------------------------nppowerd.c--- W i l l e m v a n S c h a i k ------------------------------------------------------------------------------- Gintic - Singapore gwillem@ntuvax.ntu.ac.sg
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: Transparent Icons, Again... Date: 24 Jan 1996 04:12:16 GMT Organization: Nanyang Technological University Message-ID: <4e4bj0$hg6@ntuix.ntu.ac.sg> References: <DLns4G.B6J@news.cis.umn.edu> James P. Klett (klett002@maroon.tc.umn.edu) wrote: : I know, I've seen this one many times, I even saved the last article on this : subject, but I couldn't get it to work. There is a check box in the color : panel called Paint in Overlay mode, but it is grey'd out for my Icon. : So, I can't use this function. : Once again, How do you create an Icon with transparent parts to it? Probably what you did wrong is that in the Format menu (out of my head, so naming will not be 100%), you should select an format with alpha. Willem
Newsgroups: comp.sys.next.programmer From: embuck@hogan.cca.rockwell.com (Erik M. Buck) Subject: How to use portrait display Message-ID: <DLo4tw.Dn6@lazrus.cca.rockwell.com> Sender: news@lazrus.cca.rockwell.com Organization: Rockwell Avionics - Collins Date: Wed, 24 Jan 1996 04:37:08 GMT For a variety of reasons, I want to display the workspace as if on a portrait display. The display will actually be landscape (scan along the long axis) but I want the images rotated 90 deg. In other words, I want my workspace to be taller than it is wide and I want it to look upright if I turn the monitor on its side. I hacked Workspace.ps (sp?) and made some progress but I suspect that is not the best way. Should I try to write a device driver that fools the system. Any suggestions ? Thanks in anvance. - Erik M. Buck Do any of the commercial workspace extenders support this ? I may be willing to purchase this capability.
Newsgroups: de.comp.sys.next,comp.sys.next.software,comp.sys.next.programmer From: ergo@merlin.stuttgart.netsurf.de (Olaf Foellinger) Subject: S: class 'Serial.h' Message-ID: <DLHMzE.M5o@merlin.stuttgart.netsurf.de> Followup-To: comp.sys.next.programmer Sender: news@merlin.stuttgart.netsurf.de Organization: none Date: Sat, 20 Jan 1996 16:26:02 GMT NOTE: FOLLOWUPS TO COMP.SYS.NEXT.PROGRAMMER A few weeks ago I've downloaded from CompuServe a package called NXCim. A part of this package (which contains sources for a browser for CompuServe) is a class 'Serial'. Unfortunately, the sources for this class are only the header-file and the object-code for Next-hardware. I use 'white' hardware, so it's of no use for me. The header-file contains copyright-restrictions of the following form: "Copyright (C) BenaTong 1991 All rights reserved" Can someone point me in any direction, for instance where I can find BenaTong or an object-file for NsfIP or the Sources or so. Thanks Olaf -- Olaf Foellinger homepage: http://www1.stuttgart.netsurf.de/~ergo/ NeXTMail & MIME welcome!
From: jacques.garbi@colombus.ch (Jacques Garbi) Newsgroups: comp.sys.next.programmer Subject: Problems with date in precomp Date: 23 Jan 1996 18:43:14 GMT Organization: Colombus Inc. Distribution: world Message-ID: <4e3a82$gns@news.petrel.ch> Hi, I'm sorry to have this problem, but since I'm no programmer, I can't find the reason why the compiler keeps complaining about different dates between a header and precomp. Here is one instance of what I get by the hundreds : /NextDeveloper/Headers/appkit/nextstd.h:14: warning: 'sys/wait.h' has different date than in precomp Now, can anyone help me to remove this ? I don't know how it came here. Probably I did something wrong, installed a new package, whatever. Please, help me to get rid of that ! Thanks --- Jacques GARBI Colombus Inc. Av. de France 33 1004 Lausanne Switzerland Phone : 011 41 21 661 16 76 Fax : 011 41 21 661 16 73 NeXTMail or MIME : jacques.garbi@colombus.ch
From: stanj@cs.stanford.edu (Stan Jirman) Newsgroups: comp.sys.next.programmer Subject: Writing SCSI device drivers Date: 24 Jan 1996 07:42:02 GMT Organization: Stanford University Message-ID: <4e4nsa$kd1@nntp.Stanford.EDU> Hello, I have to write a couple of device drivers, one of them for a DAT drive, something similar to that one used in SafetyNet. As you may know, once you quit SafetyNet the device /dev/nrst0 is still operational and can be used for access to the tape. This leads to the conclusion that a device driver has been installed as a kernel server. However, the documentation says: > Currently, third parties can't write device drivers for non-NeXTbus > interfaces, such as SCSI and serial ports. Third parties can > write three types of loadable kernel servers: Also, the SCSITape example is all based on the DriverKit and hence does not compile for m68k. Has anyone any idea how SafetyNet does the job, and can possibly point me at an example? Thanks a ton, - Stan --- +-------------------------------------------+-------------------------+ | Stan Jirman -- The Swiss CS Guy | | | stanj@cs.stanford.edu NeXTmail / MIME | When you find yourself | | http://www-leland.stanford.edu/~stanj/ | in a hole, stop digging | | Box 2642, Stanford, CA 94309, USA | | +-------------------------------------------+-------------------------+
From: me@localhost.wesleyan.edu (My Account) Newsgroups: comp.sys.next.programmer Subject: Re: Scheme for NEXTSTEP anyone? Date: 24 Jan 1996 10:22:02 GMT Organization: Wesleyan University Message-ID: <4e518a$leh@news.wesleyan.edu> References: <4dhc55$mnb@gate.seicom.net> In article <4dhc55$mnb@gate.seicom.net> frank@this.net (Frank M. Siegert) writes: > Jeff_Sickel@sickel.com wrote: > > Does anyone know if there is version of Scheme out there that will run under > > non-Motorola NEXTSTEP? Schematik sounds like it would work well, except for > > the Motorola lib/binary. > > > All hints appreciated. > > How much do you want? 8-) > > The following implementations should compile without much hazzle... Personally I tried > elk, Siod, scsh (well ,a shell but see below...) and guile. For more information see > http://www.cs.indiana.edu/scheme-repository/home.html. > > * Guile-iii (http://www.cygnus.com/library/ctr/guile.html) > > * LIBSCHEME is a C library implementing Scheme as described in R4RS. It > is easily integrated into a C program as a command interpreter or > extension language, and is easily extended in C with new primitive > types, primitve functions and syntax. It should be portable to most > machines with an ANSI C compiler. It is available by anonymous ftp > from > ftp.cs.indiana.edu:/pub/scheme-repository/imp/ > For more information, write to Brent Benson > <Brent.Benson@mail.csd.harris.com>. > > * Scheme->C is an R4RS compliant Scheme system that is centered around > a compiler that compiles Scheme to C. Besides the base language, > the system includes "expansion passing style" macros, a foreign function > call capability, records, weak pointers, 3 X11 interfaces, call/cc, and a > generational, conservative, copying garbage collector. The result is a > system that is portable, efficient, and able to build applications that > contain a mix of compiled and interpreted Scheme, and compiled code > from C, C++ and other languages. The current release of Scheme->C runs > on the following systems: Digital Alpha AXP systems with OSF/1, MIPS > based DECstations, VAXen with ULTRIX, MIPS based SGI systems, PC's > running Microsoft Windows 3.1, Apple Macintosh's running system 7.1, > HP 9000/300, HP 9000/700, Sony News, Harris Nighthawk and other m88k > systems, linux, Sun SPARC, and NT (Visual C++ compiler). Earlier > releases also run on Sun3, DNx500, DN1000, 386's running Unix, > DNx500, and DN1000 systems. The software is available for > anonymous ftp from > gatekeeper.dec.com:/pub/DEC/Scheme-to-C/ [16.1.0.2] > There are three interfaces to X11, all written in Scheme, available > from gatekeeper. The first is a complete set of stubs to Xlib included > in the base system. The second is an alternative to Xlib called SCIX, > found in > gatekeeper.dec.com:/pub/X11/contrib/ > The third, ezd, allows programs to easily > produce interactive, structured graphics and is found in > gatekeeper.dec.com:/pub/DEC/ezd/ > Those without ftp access can also obtain Scheme->C and ezd from the > Prime Time Freeware CD, Vol. 1, No. 2. Additional information is > available from the author at Digital Equipment Corporation's Western > Research Lab: Joel Bartlett, bartlett@decwrl.dec.com. > > * Scheme 48 is a Scheme implementation based on a virtual machine > architecture. Scheme 48 is designed to be straightforward, flexible, > reliable, and fast. It should be easily portable to 32-bit > byte-addressed machines that have POSIX and ANSI C support. > In addition to the usual Scheme built-in procedures and a development > environment, library software includes support for hygienic macros (as > described in the Revised^4 Scheme report), multitasking, records, > exception handling, hash tables, arrays, weak pointers, and FORMAT. > Scheme 48 implements and exploits an experimental module system > loosely derived from Standard ML and Scheme Xerox. The development > environment supports interactive changes to modules and interfaces. > A beta release of Scheme 48 is available by anonymous ftp from > ftp.cs.indiana.edu:/pub/scheme-repository/imp/scheme48-0.36.tar.Z > swiss-ftp.ai.mit.edu:/archive/s48/scheme48-0.36.tar.gz > ftp.cs.cmu.edu:/user/ai/lang/scheme/impl/s48/scheme48.tar.gz > For more information, contact Richard Kelsey and Jonathan Rees > at <scheme-48-request@martigny.ai.mit.edu>. > > * Scsh is a Unix shell/systems programming environment implemented on top of > Scheme 48 (a portable, byte-code compiled R4RS Scheme implementation). Scsh > provides > > - A high-level macro notation for writing typical shell-script > computations: running programs, pipelines, I/O redirection, and so > forth. For example, to decompress a file and mail it to someone, > you might say > (run (| (gzcat home.html.gz) > (mail -h "Here's my home page" shivers@lcs.mit.edu))) > > To spell check your paper, printing out the results, you could say: > (run (| (delatex (< paper.tex)) > (spell) > (lpr -Ppulp))) > > - A complete system-call interface to Unix: fork, exec, I/O, file > system, time, env vars, and so forth. The I/O interface includes > a *complete* interface to BSD sockets, both Unix and TCP/IP domains. > I/O is completely integrated with Scheme ports. System calls return > useful values, not error codes; errors are reported by raising > exceptions which can be caught by handlers. > > - Other useful shell-programming utilities: filename globbing/pattern > matching, regexp matching, macros for writing AWK-like programs, field > and record parsers, and so forth. > > - The ability to write executable shell scripts using the Unix #! > interpreter feature, with access to command-line argv values. > > These features are completely integrated into Scheme 48's R4RS Scheme > implementation; the programming language is Scheme. The scsh release > is self-contained -- it comes with its own complete Scheme 48 > implementation. Scsh currently runs on the following platforms: DEC Ultrix, > Harris NightHawk, HP-UX, IBM AIX, Linux, NetBSD/i386, NeXTSTEP/Intel, > SGI IRIX, Solaris, and SunOS. It's not hard to port scsh to new systems. > | > You can get a copy of scsh via anonymous ftp, from the following: > ftp://clark.lcs.mit.edu/pub/su/scsh/scsh.tar.gz > ftp://swiss-ftp.ai.mit.edu/pub/su/scsh/scsh.tar.gz > These tar files include a detailed manual and a paper describing > the design of the system. For the lazily curious, we also have the > manual separately available as > ftp://clark.lcs.mit.edu/pub/su/scsh/scsh-manual.ps > ftp://swiss-ftp.ai.mit.edu/pub/su/scsh/scsh-manual.ps > Scsh has been implemented by the Scheme Underground > http://www.ai.mit.edu/projects/su/su.html > For further information, contact Olin Shivers <shivers@lcs.mit.edu>, > http://clark.lcs.mit.edu/~shivers/home.html > > > * SCM, free by anonymous ftp from > swiss-ftp.ai.mit.edu:/archive/scm/ > prep.ai.mit.edu:/pub/gnu/jacal/ > ftp.cs.indiana.edu:/pub/scheme-repository/new/ > ftp.maths.tcd.ie:/pub/bosullvn/jacal/ > Current version 4e1. Runs on Amiga, Atari-ST, MacOS, MS-DOS, OS/2, > NOS/VE, VMS, Unix and similar systems. SCM conforms to the Revised^4 > Report on the Algorithmic Language Scheme and the IEEE P1178 > specification. Scm is written in C. ASCII and EBCDIC are supported. > Written by Aubrey Jaffer. > To receive an IBM PC floppy disk with the source files and MSDOS > and i386 executables send $99 to Aubrey Jaffer, 84 Pleasant Street, > Wakefield MA 01880, <jaffer@zurich.ai.mit.edu>. > SLIB (Standard Scheme Library) is a portable Scheme library > which is intended to provide compatability and utility functions for > all standard Scheme implementations, including SCM, Chez, Elk, > Gambit, MacScheme, MITScheme, scheme->C, Scheme48, T3.1, and VSCM, and is > available as the file slib2a0.tar.gz. Written by Aubrey Jaffer. > JACAL is a symbolic math system written in Scheme, and is > available as the file jacal1a4.tar.gz. > SCMCONFIG contains additional files for the SCM distribution to build > SCM on Unix machines using GNU autoconf. > SLIB-PSD is a portable debugger for Scheme (requires emacs editor). > TURTLSCM is a turtle graphics package which works with SCM on MSDOS > or X11 machines. Written by Mkinen Sami <sjm@cc.tut.fi> and Jarkko > Leppanen <jtl@cc.tut.fi>, it is available as the file turtlegr.tar.Z. > (Also available from modeemi.cs.tut.fi:/pub/scm/ as turtlegr.tar.gz, > along with an already-compiled MSDOS binary of scm with turtlegraphics > and slib in scm4c0_b.zip.) > XSCM is an X Windows interface to Xlib and the Motif and > OpenLook toolkits for the SCM interpreter. It requires scm4a10 or > later. It should be available at any archive of alt.sources, or on > swiss-ftp, prep and indiana as the file xscm1.05.tar.Z. > Contact campbell@redsox.bsw.com for more information. > SMG-SCM is a package that adds VMS SMG screen management routines > to SCM. It is available from swiss-ftp, prep and indiana as the file > sgm-scm2a1.zip. (A VMS version of Unzip is available by anonymous > FTP from ftp.spc.edu:[ANONYMOUS.MACRO32]UNZIP.EXE.) This file > contains the source code, documentation, and example code. Send > comments and bugs to T. Kurt Bond, <tkb@mtnet2.wvnet.edu> (preferred) > or <Kurt.Bond@launchpad.unc.edu>. > WB is a disk based, sorted associative array C library (database). These > associative arrays consist of variable length (less that 256 bytes) > keys and values. WB comes with an interface to SCM. Basic > operations are creation, destruction, opening and closing of > diskfiles and arrays, insertion, deletion, retrieval, successor, and > predecessor (with respect to dictionary order of keys). Functional > application of find-next, deletion, and modification over a range of > consecutive key values is supported. Multiple associative arrays > can be stored in one disk file. Simultaneous access to multiple > disk files is supported. A structure checker, garbage collector > are included. A repair program and ram-disk type file (for > temporary structures) are in developement. The current WB > implementation has a file size limit of 2^32 * block size (default > 2048) = 2^43 bytes (8796 Gbytes). WB does its own memory and disk > management. WB is available on swiss-ftp, prep, and indiana as wb1a1.tar.z. > > * SIOD (Scheme in One Defun), free by anonymous ftp from > ftp.cs.indiana.edu:/pub/scheme-repository/imp/siod-v3.0-shar > ftp.std.com:/pub/gjc/siod-3.0-shar > or in any comp.sources.unix archive. Runs on VAX/VMS, VAX UNIX, Sun3, > Sun4, Amiga, Macintosh, MIPS, Cray, Windows NT/WIN32. Small scheme > implementation in C arranged as a set of subroutines that can be > called from any main program for the purpose of introducing an > interpreted extension language. Compiles to ~42K bytes of > executable. Lisp calls C and C calls Lisp transparently. Version > 3.0 includes support for manipulation of Oracle and Digital RDB > relational databases (SQL interface). > Written by George Carrette <gjc@paradigm.com> or <gjc@world.std.com>. > > > * Elk (Extension Language Kit) has been designed specifically as an > embeddable, reusable extension language subsystem for applications written > in C or C++. Elk is also useful as a stand-alone Scheme implementation, > in particular as a platform for rapid prototyping of X11-based Scheme > programs. Elk was first published in 1989; the current version is Elk 3.0. > The Elk distribution includes a Scheme interpreter (embeddable and > stand-alone versions), several dynamically loadable extensions, run-time > support (including a top-level implemented in Scheme and a debugger), > and 230+ pages of documentation (troff and PostScript format). > Major features of Elk are incremental, dynamic loading of compiled > extensions (supported on many platforms); freezing of the interpreter or > application into a new executable file; a C/C++ programmer's interface > for language interoperability; Scheme bindings for X11 Xlib, Xt, Athena > and Motif Widgets; a UNIX interface (not restricted to POSIX); bitstrings, > records, and regular expressions; a stop-and-copy and an incremental, > generational garbage collector. > The Elk 3.0 distribution and more information about Elk are available > in the World Wide Web at > http://www.informatik.uni-bremen.de/~net/elk > The distribution is also available on a number of FTP sites including > ftp://ftp.x.org/contrib/devel_tools/elk-3.0.tar.gz > ftp://ftp.uni-bremen.de/pub/programming/languages/scheme/elk/elk-3.0.tar.g z > For more information contact Oliver Laumann <net@cs.tu-berlin.de>. > (a patch to get the dynamic loading reliable on NS is available from me, frank@this.net) > > Hope this helps > > Frank > > --- > Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank > NeXTSTEP & PostScript Guy "In cantonese C++ is called C ga ga" -- I had no trouble at all making Aubrey Jaffer's _scm_ on NS 3.3 (Intel). I've used it quite a bit. I'm impressed with the interpreter's speed. ------------------------------------------------------------------------- Jim Lipton | Math Dept. Wesleyan University, Middletown CT 06459-0128 | e-mail: lipton@wesleyan.edu =========================================================================
From: me@localhost.wesleyan.edu (My Account) Newsgroups: comp.sys.next.programmer Subject: Re: Is NeXT Developer C++ compatible? Date: 24 Jan 1996 10:27:14 GMT Organization: Wesleyan University Message-ID: <4e51i2$lek@news.wesleyan.edu> References: <3102894E.4461@home.mysolution.com> In article <3102894E.4461@home.mysolution.com> "TOMMY K. HWANG" <thwang@home.mysolution.com> writes: > I am trying out C++ and am wondering if NeXT > Developer is really C++ capable? The man pages says > it is, but it does not want to compile simple program > below: > > #include <iostream.h> > > int main() > > { > float test1 = 0.0; > > float test2 = 0.0; > > float finalExam = 0.0; > > float courseGrade = 0.0; > > > cout << " Test1: "; > > cin >> test1; > > cout << " Test2: "; > > cin >> test2; > > cout << " Final Exam: "; > cin >> finalExam; > > courseGrade = 0.25 * test1 + 0.25 * test2 + 0.50 * finalExam; > > cout << endl; > > cout << "Course Grade: " << courseGrade << "%" << endl; > > return 0; > } > The error message is below: > > next1> cc finalgrade.cc > ld: Undefined symbols: > ostream::operator<<(char const *) > ostream::operator<<(ostream &(*)(ostream &)) > ostream::operator<<(float) > istream::operator>>(float &) > cin > cout > endl(ostream &) > next1> > > What might my problem be? Thanks... -- I got the same results with cc. Try /NextDeveloper/Source/GNU/cc/cc/g++ ------------------------------------------------------------------------- Jim Lipton | Math Dept. Wesleyan University, Middletown CT 06459-0128 | e-mail: lipton@wesleyan.edu =========================================================================
Newsgroups: comp.databases.sybase,comp.sys.next.programmer,comp.sys.next.software From: Fabien_Roy@free.fdn.org Subject: (EOF) How to tell EOF that a trigger had rollback a transaction Message-ID: <1996Jan24.072702.14580@free.fdn.org> Sender: news@free.fdn.org Organization: Fabien Roy Consultant, Paris, France Date: Wed, 24 Jan 1996 07:27:02 GMT Hi, I got a problem with the handling of transaction with EOF when a trigger rolls back the current transaction. EOF seems out of sync, the only way to cure it is to quit the application and restart. Here are the messages: Sybase Server:unknown Procedure:none Line:1 Message:Attempt to issue 'SAVE TRANsaction' when there is no active transaction Sybase Server:unknown Procedure:none Line:1 Message:The commit transaction request has no corresponding BEGIN TRANSACTION. I know about the delegate method: - (BOOL)sybaseChannel:(SybaseChannel *)channel willReportMessageWithNumber:(int)msgNo msgState:(int)msgState severity:(int)severity msgText:(char *)msgText srvname:(char *)srvname procname:(char *)procname lineno:(int)lineno and tried to send to the EODatabaseContext the message transactionDidRollback with no luck :-( Thanks Fabien -- --------------------------------------------------------------------- Fabien_Roy@free.fdn.org (NextMail/MIME accepted) Fabien Roy Consultant NEXTSTEP/OPENSTEP/EOF Consultant, SYBASE DBA 10 rue de la DEFENSE 93100 MONTREUIL, France Tel: 33 1 45 28 32 23 Fax: 33 1 48 55 09 90
From: "Christoph Widmer (by ubsswop)" <zhwit> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: FrameMaker to HTML conversion Date: 24 Jan 1996 14:59:48 GMT Organization: Union Bank of Switzerland Message-ID: <4e5hh4-qas@svstch.ubs.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi all I'm looking for a FrameMaker (mif) to HTML converter, which can handle different pargraph formats, table of contents, ... Any suggestions ? Thanks for your help. Christoph Widmer Christoph.s.z.h.w.i.t.Widmer@ubs.ch
From: franke@boehme.dbag.ulm.DaimlerBenz.COM (Juergen Franke) Newsgroups: comp.sys.next.programmer Subject: Dependencies problems with make Date: 24 Jan 1996 15:51:56 GMT Organization: Daimler-Benz AG Message-ID: <4e5kis$988@news.sns-felb.debis.de> Question concerning evaluation of dependencies by the project builder via linked headers. In my project I have the following instruction in my Makefile.preamble OTHER_CFLAGS = -I/LocalDeveloper/ClassLab/Headers I have links to all header files in this directory (/LocalDeveloper/ClassLab/Headers) which I need in my project. Additionally I have references to header files in my makefile.dependencies to compile my .m and .c files. But if one of these headers is changed, the project builder does not compile the dependend .m and .c files. It seems that make ignores the links in the Headers directory. Even when we create all link again, it has no effect. Only when I start depend in the project builder again, the project builder compiles the relevant files. Jörgen Franke Text Understanding Department F3M/T Daimler-Benz AG, Research Center Ulm P.O. Box 2360 89013 Ulm Germany Telephone Germany 731 505 2355 Fax Germany 731 505 4113 e-mail: franke@dbag.ulm.DaimlerBenz.COM
From: bettsc@wcpost.rfweston.com Newsgroups: comp.sys.next.programmer Subject: y Date: 24 Jan 1996 16:25:40 GMT Organization: InterServ News Service Message-ID: <4e5mi4$8r5@lal.interserv.net> test
From: Yibing Wu <ywu@lan.sky.bdm.com> Newsgroups: comp.sys.next.programmer Subject: Problem Solved Date: 24 Jan 1996 23:30:39 GMT Organization: CACI, Inc. Message-ID: <4e6fev$op4@news.mcl.bdm.com> References: <4e3kca$jru@news.mcl.bdm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit When I init the EODatabaseDataSource object, I should have init it with new context and channel and database. Thanks.
From: david@pfi.ibk.baum.ethz.ch (David EKCHIAN) Newsgroups: comp.sys.next.programmer Subject: Re: 3D kit, any bugs ? Date: 25 Jan 1996 10:32:21 GMT Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4e7m7l$flv@elna.ethz.ch> References: <4do6ac$h6@wfn.emn.fr> <1996Jan21.193209.8856@media.mit.edu> In <1996Jan21.193209.8856@media.mit.edu> Michael B. Johnson wrote: > To begin playing with the 3DKit, take a look at the "Simple" example > in NextDeveloper/Examples/3DKit. YOu could also take a look at my > WavesWorld tools, available off of my home page. I extend the 3DKit > in a number of ways. > --> Michael B. Johnson, SMVS, Ph.D. -- wave@media.mit.edu|wave@pixar.com > --> http://wave.www.media.mit.edu/people/wave/ IMHO, the www address is : "http://www.media.mit.edu/people/wave/" The address Michael gave (as Steve Weintz) doesn't work. -- o _ /-;c ___ David C. EKCHIAN ______________________________________(@)#\(@)___
From: escribe@dial.pipex.com (Alan Barclay) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 25 Jan 1996 14:33:37 GMT Organization: The Electric Scribe Co. Ltd. Sender: barclay@frodo.ntc.nokia.com (Alan Barclay) Message-ID: <4e84c1$m0n@axl02it.ntc.nokia.com> Keywords: NeXT, Plan9 Some ideas for you to think about ... 1. *** 4.4 BSD port to NeXT Black Hardware *** 2. *** OPENSTEP 4.0 binaries support (m68k, FIP, SP, HP) under 4.4 BSD port *** The following are extracts from NeXT newsgroup articles. Many other NeXT users have also expressed concern about the elderly 4.3BSD (on Mach) which NeXT sells as its OS. Mark Crispin <mrc@CAC.Washington.EDU> wrote: > Black hardware is, for the most part, still reasonably capable > Unix hardware. I have reason to believe that either Linux or > BSD Lite would perform on black hardware as a much better UNIX > workstation than NEXTSTEP does. Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: > One would wonder how much information about NeXT hardware people > could figure out by looking at Plan 9. There's some students on > campus here interested in Plan 9, and I've been musing about > putting one of my NeXT's towards that. louie@va.pubnix.com (Louis A. Mamakos) wrote > I'd love to run FreeBSD or NetBSD on my black hardware. > Perhaps starting with the Plan 9 release is an approach? Has there > ever been any released documentation on the hardware specs, registers, > etc? Who knows, maybe the SCSI bus could do synchronous tranfers? paul@nagshead.plsys.co.uk (Paul Lynch) wrote: > Personally, it seems to me that the BSD 4.4 Lite upgrade was the > potentially most valuable to NeXTSTEP users. It will be a great shame if > that is lost. 1. Back in Oct/Nov 95 I made some noises myself in some groups about porting 4.4BSD to NeXT black hardware. I thought then and still think that the effort to do this would not be very great as 4.4 BSD, in the form of NetBSD, is already compatible with m68k architecture and Plan 9 sources, from AT&T, contain the specific hardware information required to implement low level drivers. At the time I heard some noises from a few others who were interested but no real action from anybody interested in doing a port. At the time, I was in no position to give time to such a task, but I now am more available and have a couple of mono stations sitting idle for such a task. If NetBSD (4.4BSD) were ported to NeXTstation (and NeXTstation colour) hardware - How many of you would be interested in using it? How many of you would be interested in helping? 2. The recent unhappiness with NeXTs elderly 4.3BSD and their lack of further development for it has prompted another idea by current NEXTSTEP 3.x users. Forget NeXTs 4.3BSD and run OPENSTEP on another UNIX instead. At first I thought this idea was ridiculous, but now I am much less sure and would welcome input from others. If we remember that 4.3BSD binaries for one machine will execute on other machines 4.3BSD (with the same processor architecture) then perhaps we can simply take the OPENSTEP libraries and binaries and run them under our own 4.3BSD compatible OS, say 4.4BSD NetBSD. Infact, this would not work immediately because of the fact that NeXTs 4.3BSD is really MACH. However there is a version of 4.4BSD for MACH called 4.4BSD LITES (produced at Helsink Uni) - (Slightly strange coincidence is that Linux comes out of Helsink too!). Perhaps these sources offer the possibility of running native OPENSTEP libraries and binaries on an upto date UNIX (for those people who really want OPENSTEP on UNIX rather than on NT/W95). I accept that there are hurdles to overcome, not least of which might be the fact that LITES is based on a version of MACH later than that which NeXT used or that new code to support the NeXT MAB executable file formats would be required. I am interested in your comments on this. Alan Barclay The Electric Scribe Co. Ltd. escribe@dial.pipex.com alan@escribe.co.uk (Down!)
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: US- CA - Senior NeXT Architect & NeXT Developers - to 90k + Date: 25 Jan 1996 13:46:06 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4e8j5e$ffk@newsbf02.news.aol.com> My California based client currently relies very heavily on NeXT as their platform of choice for mission critical financial systems. Due to the companies continued growth my client currently has several positions open in their top development group. *~ NeXT Architect~* Candidate must have 4/5 years experience in NeXTSTEP, with experience in architecting, designing, and developing large NeXT based applications. Must have good Objective-C and Sybase or Oracle. B.S or M.S required. Must be Citizen or Permanent Resident Salary - $70k to $90k (possibly more) + benefits + 96' Bonus *~NeXT Developers~* Candidates must have 2+ years NeXT design and development experience with good Objective-C skills coupled with Sybase or Oracle. B.S required U.S Citizen or Permanent Resident Salary - $50 to $75+ , + benefits + 1996 Bonus Contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
Newsgroups: comp.sys.next.programmer From: gemoe@proximus.north.de (Gerhard =?iso-8859-1?Q?M=F6ller?=) Subject: Re: S: class 'Serial.h' References: <DLHMzE.M5o@merlin.stuttgart.netsurf.de> Organization: German NeXT User Group, Oldenburg. Date: Thu, 25 Jan 1996 09:13:52 GMT Message-ID: <1996Jan25.091352.11829@proximus.north.de> In article <DLHMzE.M5o@merlin.stuttgart.netsurf.de>, Olaf Foellinger <ergo@merlin.stuttgart.netsurf.de> wrote: Olaf> A few weeks ago I've downloaded from CompuServe a package called NXCim. A part of this package (which contains sources for a browser for CompuServe) is a class 'Serial'. Unfortunately, the sources for this class are only the header-file and the object-code for Next-hardware. I use 'white' hardware, so it's of no use for me. The header-file contains copyright-restrictions of the following form: Olaf> "Copyright (C) BenaTong 1991 All rights reserved" Olaf> Can someone point me in any direction, for instance where I can find BenaTong or an object-file for NsfIP or the Sources or so. BenaTong email: info@bentong.com Tel.: (614) 276-7859 Fax: (614) 276-7859 1057 Racine Ave. Columbus, Ohio 43204 USA Gerhard. -- Private: gemoe@proximus.north.de NeXTmail & MIME Uni: Gerhard.Moeller@Informatik.Uni-Oldenburg.DE "Personally, I don't think NT runs worth diddly on a PeeCee, either, but it does boot (many times a day...)" --Mike Paquette
From: indy@ux1.cso.uiuc.edu (weintz steven cortelou) Newsgroups: comp.sys.next.programmer Subject: Re: 3D kit, any bugs ? Date: 26 Jan 1996 19:58:20 GMT Organization: University of Illinois at Urbana Message-ID: <4ebbos$baa@vixen.cso.uiuc.edu> References: <4do6ac$h6@wfn.emn.fr> <1996Jan21.193209.8856@media.mit.edu> <4e7m7l$flv@elna.ethz.ch> david@pfi.ibk.baum.ethz.ch (David EKCHIAN) writes: >In <1996Jan21.193209.8856@media.mit.edu> Michael B. Johnson wrote: >> To begin playing with the 3DKit, take a look at the "Simple" example >> in NextDeveloper/Examples/3DKit. YOu could also take a look at my >> WavesWorld tools, available off of my home page. I extend the 3DKit >> in a number of ways. >> --> Michael B. Johnson, SMVS, Ph.D. -- wave@media.mit.edu|wave@pixar.com >> --> http://wave.www.media.mit.edu/people/wave/ >IMHO, the www address is : >"http://www.media.mit.edu/people/wave/" >The address Michael gave (as Steve Weintz) doesn't work. Oops; my goof, not wave's. -- Steve Weintz * Partner, New Media, Ltd. P.O. Box 742 * Urbana, IL 61801 steve@dave-world.net * 217.344.5303 * 217.344.8981 fax Staff Visioneer, Beckman Institute Visualization Facility * indy@uiuc.edu
Newsgroups: comp.sys.next.programmer From: gerald@kurt.in-berlin.de (Gerald Erdmann) Subject: How to run a FontPanel in a modal loop Message-ID: <1996Jan26.114936.1097@kurt.in-berlin.de> Sender: news@kurt.in-berlin.de Date: Fri, 26 Jan 1996 11:49:36 GMT Hi! Does somebody knows a way to run a FontPanel in a modal loop like an OpenPanel oder a SavePanel? Thanx and have a nice weekend, Gerald -- --------------------------------------------------------------------- | GERALD ERDMANN | email: gerald @ kurt.in-berlin.de (NeXTmail welcome) | voice: +49 30 372 43 10 (Germany - Berlin) | crypt: pgp2 public key available |
From: Tal Lancaster <tlan> Newsgroups: comp.sys.next.programmer Subject: Re: 3D kit, any bugs ? Date: 26 Jan 1996 21:38:46 GMT Organization: Walt Disney Studios Message-ID: <4ebhl7$bb0@cabinboy.studio.disney.com> References: <4do6ac$h6@wfn.emn.fr> <1996Jan21.193209.8856@media.mit.edu> <4e7m7l$flv@elna.ethz.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit david@pfi.ibk.baum.ethz.ch (David EKCHIAN) wrote: >In <1996Jan21.193209.8856@media.mit.edu> Michael B. Johnson wrote: [Sniped] >> YOu could also take a look at my >> WavesWorld tools, available off of my home page. I extend the 3DKit >> in a number of ways. >> --> Michael B. Johnson, SMVS, Ph.D. -- wave@media.mit.edu|wave@pixar.com >> --> http://wave.www.media.mit.edu/people/wave/ > >IMHO, the www address is : > >"http://www.media.mit.edu/people/wave/" > >The address Michael gave (as Steve Weintz) doesn't work. Actually both links work. It is just for a while MIT was changing things around and for a while neither would work. But all if fine now. -- Tal Lancaster Technical Director, Disney Feature Animation email: tlan@fa.disney.com WEB: The RenderMan Repository (http://pete.cs.caltech.edu/RMR/)
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Flakey black patched NS 3.3/EOF 1.1 debugging environment Date: 27 Jan 1996 00:21:10 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4ebr5m$17f@emerald.oz.net> I haven't seen much discussion of the debugging environment of patched NS 3.3 black with EOF 1.1, but I'm finding it to be very frustrating to use, especially relative to NS 3.2. Either gdb or Edit crashes very frequently severely limiting debugging capabilities. Has anyone else experienced this? If so, have you solved the problem, or at least adopted measures that minimize the problem. I have found that if I quit the debugged app, restarting it in the same gdb session will almost guarantee that gdb will crash during the next launch of the app. But if I merely rerun the app without quitting it first, gdb *usually* doesn't crash. Any help or advice would be appreciated. This problem is really taking a serious toll on my debugging proficiency. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: DO error message : handleRequestOnPortal: id d2bd48 not available? Message-ID: <DLu797.2G4@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4df7hm$2j8@caesar.ultra.net> Date: Sat, 27 Jan 1996 11:15:07 GMT Hi Dirk, In article <4df7hm$2j8@caesar.ultra.net> df@watershed.com (Dirk Fromhein) writes: > > NS 3.3, EOF 1.1, Foundation Patch. Intel Boxes. > After about 500-600 DO connections/method calls, I always get: > handleRequestOnPortal: id d2bd48 not available > > I've looked though the docs, headers, etc. and I can't find anything. > I'm using NXAutoreleaseConnections. If you have multi-client environment, check how you call the [NSAutoreleasePool release] on the server side. I and Ondra Cada <ocs@ms.mff.cuni.cz> had the similar problem. Here is the summary we posted about this problem some time ago: --------------------------------------------------------------------- Newsgroups: comp.sys.next.programmer Path: hurka!tom From: tom@hukatronic.cz (Tomas Hurka) Subject: DO and NSAutoreleasePool (was: Re: DO nice trick) Message-ID: <DJt03v.2G5@hurka.UUCP> Keywords: DO NSAutoreleasePool Sender: tom@hurka.UUCP (Tomas Hurka) Reply-To: tom@hukatronic.cz (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <199512181046.AA03169@ns.ms.mff.cuni.cz> Date: Mon, 18 Dec 1995 22:35:55 GMT Hi ! In article <199512181046.AA03169@ns.ms.mff.cuni.cz> writes: > Are you aware that no method in multi client server is guaranted to > be processed mutually exclusive with any other method? [...] > In fact, this very strange behavior is indirectly documented: the > connection (used to communicate with the proxy) can serve other > client's DO request, which in turm might trigger any other server > method. Sorry to bore you if you had known this, but I guess almost > nobody understand till his/her program start crashing irreguralry > due to some critical section violation. This is especially very dangerous when combining DO with autoreleasing used in EOF. [NSAutoreleasePool release] on the server side must be called only from the server main event loop. Otherwise, due to the DO behavior mentioned above, [NSAutoreleasePool release] can be called inside any server method, which communicates with the client. --------------------------------------------------------------------- I hope it helps you, and if you have any questions, please let us know. Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: fox@jeans.fokus.gmd.de (Oliver Fox) Newsgroups: comp.sys.next.programmer Subject: insert .sound in SybaseDB with EOF.app -needHelp Date: 27 Jan 1996 14:32:25 GMT Organization: GMD-FOKUS Message-ID: <4edd1p$8n@stern.fokus.gmd.de> Hi everybody, Iam going to write an application which store all kinds of Data-Files in a Sybase DataBase. Are there any examples to insert sound ps mpeg with a NeXT EOF application ? Thank You Oliver
From: btipping@ix.netcom.com (roybatty) Newsgroups: comp.sys.next.programmer Subject: NeXTSTEP and Gateway laptop Date: Sat, 27 Jan 1996 19:10:14 GMT Organization: Daugherty Systems, Inc. Message-ID: <4ediov$2aj@speedy.daugherty.com> I have a Gateway S120 (120 Pentium, 24MB RAM, 1.2 GB, 256 PB-cache). Does anyone have any experience loading NeXTSTEP 3.3 on a Gateway? If so, what problems did you encounter? Thanks in advance... btipping@ix.netcom.com
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs Subject: Is NSString's compare:options:range: broken? Date: 27 Jan 1996 18:33:08 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4edr54$ao1@emerald.oz.net> I have been unable to get compare:options:range: to work in the manner I expect from the documentation. I expect the comparison to occur on only those characters in the specified range, but my experience suggests that the length of the range is ignored; i.e., the comparison result is based on characters past the end of the range. For example, if the receiver is "p" and the string to compare with is "pr" with a range of {0, 1} and options of 0, I expect the comparison result to be NSOrderedSame. But the comparison result actually returned is NSOrderedAscending. This is with the EOF 1.1 Foundation Kit. Has anyone else observed this? Or am I just interpreting the documentation incorrectly? From my perspective, this is either a functionality or documentation bug, so I'm filing a bug report with NeXT. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: marcos@imana.com (Marcos Javier Polanco) Newsgroups: comp.sys.next.programmer Subject: DO: NX_objectNotAvailableException & NX_receiveTimedOut Date: 27 Jan 1996 19:45:37 GMT Organization: TLGnet Inc., (formerly The Little Garden) Message-ID: <4edvd1$p5@kadath.zeitgeist.net> Hello, I am vending an NSObject over a normal NXConnection (I had difficulties using the NXAutoreleaseConnection): + startServer:(NSRoot *)root named:(NSString *)serverName; { NXConnection* server = [NXConnection registerRoot: root withName:[serverName cString]]; [NXConnection setDefaultTimeout:6000000]; [server runFromAppKit]; return root; } Clients connect to the server like this: myServer = [[NXConnection connectToName: [serverName cString]] addReference]; . . . [myServer free]; which I understand will return the server's proxy; I dutifully addReference and free the proxy. This setup usually works very well. ocassionally, however, I two kinds of exceptions, both defined for NXProxy: NX_receiveTimedOut (#11013) NX_objectNotAvailableException (#11007) I don't understand how either exception is possible. I've set the timeout to an almost eternal number when vending the server. The 11007 error I just don't understand. I know there's a leak when vending an NSObject over an NXConnection, but that should not make the object become *unavailable*. Your help is appreciated! I don't usually read the newsgroups, so an email response is much appreciated. Thanks! -- ========================================================== Marcos J. Polanco The House of Imana 415.255.1102 (voice) 1011A Fell Street 415.255.1102 (fax) San Francisco, CA 94117 marcos@imana.com http://www.imana.com/ ==========================================================
From: black@winternet.com (Ben Black) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 27 Jan 1996 21:36:05 GMT Organization: StarNet Communications, Inc Message-ID: <4ee5s5$sh2@blackice.winternet.com> References: <4e84c1$m0n@axl02it.ntc.nokia.com> In article <4e84c1$m0n@axl02it.ntc.nokia.com> escribe@dial.pipex.com (Alan Barclay) writes: > > I accept that there are hurdles to overcome, not least of which might be the > fact that LITES is based on a version of MACH later than that which NeXT > used or that new code to support the NeXT MAB executable file formats > would be required. I am interested in your comments on this. > not only is LITES based on mach 3.0 rather than mach 2.x, but the nextstep kernel is a mangled melange of mach 2.x, 3.0, and BSD 4.2, 4.3 and, i think, some 4.4. it would likely be easier to make an emulation library for 4.4BSD than to try to get the (very buggy) 3.0 Mach + LITES to work. of course, then you have to figure out exactly what to emulate and how. not a pleasant thought. whatever happens, i hope the end result doesn't involve using xwindoze. --- Benjamin Black black@winternet.com Network Engineer http://www.winternet.com/~black Minnesota Regional Network OC-3 Or Bust
From: nicholas@infosciences.com (Nicholas Winterhalter) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.sysadmin Subject: FS: NEXTSTEP DEVELOPER CD-ROM Date: Sat, 27 Jan 1996 17:37:01 -0700 Organization: Rt66.COM, Public Internet Access in New Mexico Message-ID: <nicholas-2701961737010001@pmc25.rt66.com> For sale used: NEXTSTEP DEVELOPER CD-ROM o Full Developer's Release 3.1 o For Intel Processors and NeXT Computers Asking $900 --- Nicholas Winterhalter nicholas@infosciences.com
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: DO error message : handleRequestOnPortal: id d2bd48 not available? Date: 28 Jan 1996 01:56:25 +0100 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x768dxyw5i.fsf@tom.es.ele.tue.nl> References: <4df7hm$2j8@caesar.ultra.net> In-reply-to: df@watershed.com's message of Mon, 15 Jan 1996 21:40:46 -400 In article <4df7hm$2j8@caesar.ultra.net> df@watershed.com (Dirk Fromhein) writes: handleRequestOnPortal: id d2bd48 not available This is for a server that has to be up 7x24, and once we get that message we are hosed for any new connections. Clearly not acceptable! I once saw a similar message, which I tracked down to a problem with oneway void messages not being all that oneway void: the client would invoke a oneway void method and exit immediately. In the server, DO seemed to check for the originator being valid, issueing this error, before it checked that the message was oneway, thus needing no reply. --Tiggr
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs Subject: Re: Is NSString's compare:options:range: broken? Followup-To: comp.sys.next.programmer Date: 28 Jan 1996 01:59:49 +0100 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x74tthyvzu.fsf@tom.es.ele.tue.nl> References: <4edr54$ao1@emerald.oz.net> In-reply-to: art@cubicsol.com's message of 27 Jan 1996 18:33:08 GMT In article <4edr54$ao1@emerald.oz.net> art@cubicsol.com (Art Isbell) writes: For example, if the receiver is "p" and the string to compare with is "pr" with a range of {0, 1} and options of 0, I expect the comparison result to be NSOrderedSame. But the comparison result actually returned is NSOrderedAscending. My interpretation is that you compare the string with the receiver limited to the range. In this case, `p' (the string limited to the range) is shorter than `pr' and the NSOrderedAscending result is what I'd expect. --Tiggr
From: sl@skynet.oph.rwth-aachen.de (Stefan Leuker) Newsgroups: comp.sys.next.programmer Subject: Re: How to run a FontPanel in a modal loop Date: 28 Jan 1996 16:43:23 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Distribution: world Message-ID: <4eg93b$bkq@news.rwth-aachen.de> References: <1996Jan26.114936.1097@kurt.in-berlin.de> In article <1996Jan26.114936.1097@kurt.in-berlin.de> gerald@kurt.in-berlin.de (Gerald Erdmann) writes: > Hi! > > Does somebody knows a way to run a FontPanel in a modal loop like an > OpenPanel oder a SavePanel? > > Thanx and have a nice weekend, Gerald > Why would you like to do this? But there is a method like runModalFor:(Window*) in the Application class. Hope this helps. Stefan
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs Subject: Re: Is NSString's compare:options:range: broken? Date: 28 Jan 1996 19:45:18 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4egjoe$p6i@emerald.oz.net> References: <4edr54$ao1@emerald.oz.net> <x74tthyvzu.fsf@tom.es.ele.tue.nl> tiggr@es.ele.tue.nl (Pieter Schoenmakers) wrote: > In article <4edr54$ao1@emerald.oz.net> art@cubicsol.com (Art Isbell) writes: > For example, if the receiver is "p" and the string to compare with > is "pr" with a range of {0, 1} and options of 0, I expect the > comparison result to be NSOrderedSame. But the comparison result > actually returned is NSOrderedAscending. > My interpretation is that you compare the string with the receiver limited > to the range. In this case, `p' (the string limited to the range) is > shorter than `pr' and the NSOrderedAscending result is what I'd expect. Ah, that makes sense. So maybe there's no way to use compare:... to compare just the characters in a range within a string as I was trying to do. I could instead use rangeOfString:... and examine the length of the returned range to determine whether a matching prefix exists. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: rragner@panther.vm.iastate.edu (Rod Ragner) Newsgroups: comp.sys.next.programmer Subject: saveRTFDTo and openRTFDFrom: for Text Object Date: 29 Jan 1996 00:24:27 GMT Organization: Iowa State University, Ames, Iowa Message-ID: <4eh43r$jrr@news.iastate.edu> I can not seem to get my Text object to open the same thing that I previously saved. I have a ScrollView in my main window, with a corresponding Text (the dault one) object within it. The user fills this Text object with text and graphics and slects the Save button. The application saves the Text object to a ".rtfd" document using the saveRTFDTo: method. Careful inspection notes that it is saves as a RTFD folder. The application then opens the file just saved into the same Text object using the openRTFDFrom: method. Thankfully, the graphics are still present. The application saves the contents of the Text object to the same file again using the same saveRTFDTo: method (careful inspections shows, however, that it is saved as a RTF file, not a RTFD folder), then opens the same file again using the openRTFDFrom: method, but this time the graphics dissappear and appear as NeXT "placeholder" icons! It appears that the saveRTFDTo: method works when the user has dragged the graphics into the Text object. The openRTFDFrom: method may not be including everything necessary to permit re-saving it as a RTFD however! This could explain why the second and subsequent saveRTFDTo: methods to not work properly (they leave the links to the graphics in the document, but no icons are written, since there is no folder in which to put them. Is there a bug with the openRTFDFrom: or saveRTFDTo: methods (NEXTSTEP 3.3 User and Developer on a NeXTstation Turbo) or have I done something wrong or left something out? I appreciate any suggestions! Thank you! -- Rod Ragner, UNIX Systems Administrator/NEXTSTEP Application Developer Veterinary Diagnostic Laboratory, College of Veterinary Medicine, Iowa State University, 2630 Vet. Med. Bldg., Ames, Iowa 50011 Voice: (515) 294-4751, FAX: (515) 294-6961 or 3564, (NeXT Mail accepted) Email: rragner@stallion.vm.iastate.edu or stryder@iastate.edu
From: trombino@wendy.ucsd.edu (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Drag and drop Date: 29 Jan 1996 01:51:02 GMT Organization: The University of California at San Diego Message-ID: <4eh966$djc@news1.ucsd.edu> Hello, I'm having a bit of a problem implementing drag and drop in an application, and I was hoping that someone out there could please help he out... I've used code from the developer's examples and from Solitaire.app, but what seems to be happening is that I'm not getting any messages sent to my sending view (or the view that initiates the drag session), other than - (NXDragOperation)draggingSourceOperationMaskForLocal: No other messages are sent to the initiating view. The receiving view is getting every message that its supposed to, and the draggingPBoard is functioning as expected. What gives? Am I missing something really obvious here? Please help!! I've looked endlessly over source code examples and can't find out what I'm doing wrong! Thanks in advance! -- ================================================================== Mark Trombino trombino@wendy.ucsd.edu Center For Research in Computing the Arts, U.C.S.D. ==================================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical SplitView Date: 29 Jan 1996 01:29:57 GMT Organization: ftn Internet Message-ID: <4eh7ul$kfe@master.ftn.net> Does anyone know how to create an equivalent to NXSplitView, except with the split bar being vertical instead of horizontal? If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical SplitView Date: 29 Jan 1996 01:30:04 GMT Organization: ftn Internet Message-ID: <4eh7us$kfe@master.ftn.net> Does anyone know how to create an equivalent to NXSplitView, except with the split bar being vertical instead of horizontal? If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical SplitView Date: 29 Jan 1996 01:31:28 GMT Organization: ftn Internet Message-ID: <4eh81g$kfe@master.ftn.net> Does anyone know how to create an equivalent to NXSplitView, except with the split bar being vertical instead of horizontal? If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical SplitView Date: 29 Jan 1996 01:33:08 GMT Organization: ftn Internet Message-ID: <4eh84k$kfe@master.ftn.net> Does anyone know how to create an equivalent to NXSplitView, except with the split bar being vertical instead of horizontal? If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical SplitView Date: 29 Jan 1996 01:54:50 GMT Organization: ftn Internet Message-ID: <4eh9da$kfe@master.ftn.net> Does anyone know how to create an equivalent to NXSplitView, except with the split bar being vertical instead of horizontal? If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical SplitView Date: 29 Jan 1996 01:57:54 GMT Organization: ftn Internet Message-ID: <4eh9j2$kfe@master.ftn.net> Does anyone know how to create an equivalent to NXSplitView, except with the split bar being vertical instead of horizontal? If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Sorry for the multiple posts!! Date: 29 Jan 1996 02:01:38 GMT Organization: ftn Internet Message-ID: <4eh9q2$l0u@master.ftn.net> Sorry for the multiple posts!! I was having a slight technical difficulty with my news reader. Mea culpa, mea culpa, mea maxima culpa!! -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: Dennis Glatting <dennis.glatting@plaintalk.bellevue.wa.us> Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: A problem with NeXTSTEP 3.3 m68k, sendmail 8.7.3, and POSIX Date: 29 Jan 1996 03:47:49 GMT Organization: AccessOne Distribution: world Message-ID: <4ehg15$ock@news.accessone.com> I am having a problem with sendmail 8.7.3 when compiled and linked with the -posix option under NeXTSTEP 3.3 m68k, patch level 1. I defined NEWDB and previously compiled and installed Berkeley db 1.85 using the -posix option, though that appears not to be relevant. The problem I am having is related to the readdir() function used by the orderq() function in queue.c. The following lines of code are the problem focal point: register struct dirent *d; ... while ((d = readdir(f)) != NULL) The compiler generates a warning message indicated d and readdir() are different types. I preprocessed (cc -posix -E) queue.c and determined d and readdir() are the same type. When the code executes it appears that d is the type "struct direct" (indicated by the debugger) but readdir() returns struct dirent, as it should. Consequently, indexes into d are using the wrong offsets. I tried a trick. I preprocessed queue.c (cc -posix -E) then compiled the result without the -posix flag and linked it with the rest of the code specifying -posix. It worked. It appears the the flag -posix has an effect after preprocessing, possibly redefining "dirent" to "direct". Have any of you come across a problem like this? If so, how did you solve it? -- Dennis P. Glatting
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 29 Jan 1996 05:15:19 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4ehl57$e8t@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: saveRTFDTo and openRTFDFrom: for Text Object Date: Mon, 29 Jan 1996 12:05:50 +0000 Organization: Cedar Systems Distribution: world Message-ID: <WJbRvCAegLDxEwWP@cedar.co.uk> References: <4eh43r$jrr@news.iastate.edu> MIME-Version: 1.0 Rod, We have done a lot with the Text class, so thought I'd try to help. To replicate your problem I took the following steps: First, created a new ProjectBuilder project and opened the nib file. Dragged a window off the palette and set to be visible at launch time. Dragged Text ScrollView off palette into window. Positioned to top left of window and resize to fill width of window and height of window leaving space for couple of buttons at bottom. Dragged two buttons into window below ScrollView - set titles to Read and Save. Created subclass of Object in ProjectBuilder and included two action methods: doRead: and doSave:, and one instance: theText. Instantiated an instance of the myObject subclass, and unparsed into project. Connected buttons to myObject instance to invoke doRead: and doSave: as appropriate and connected theText instance to the Text within the ScrollView. Saved and quitted InterfaceBuilder. Edited the myObject.m to contain the following: - doRead:sender { NXRTFDError retval; retval = [theText openRTFDFrom:"/tmp/rtfdtest.rtfd"]; return self; } - doSave:sender { NXRTFDError retval; retval = [theText saveRTFDTo:"/tmp/rtfdtest.rtfd" removeBackup:YES errorHandler:nil]; return self; } Saved and run the project. I keyed in some text and copied and pasted in a graphic from another rtfd file. I clicked the save button and verified the rtfdtest.rtfd file was correctly written away. I changed the text and clicked the read button - the text was correctly redisplayed. Next I changed the text and clicked the save button. Again, I verified that the file was correctly saved as rtfd. Having changed the text and clicking the read button, the text was redisplayed exactly as when it was last saved. I have not been able to reproduce your problem. Try following the steps above to determine whether your problem is related to your installation of NEXTSTEP or perhaps something else in your project. If the above steps work correctly on your installation and you can't get any further with the problem, feel free to send me your minimal source that causes the problem and I can replicate here and see if I can get any further with it. Regards, Paul. email: phef@cedar.co.uk ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
Newsgroups: comp.sys.next.programmer From: tini@gurke.ping.de (Constantin Szallies) Subject: Sorting an array and other stuff: Doing it the right way? Message-ID: <DLy7A1.7H@gurke.ping.de> Sender: usenet@gurke.ping.de Organization: Legalize THC Date: Mon, 29 Jan 1996 15:06:00 GMT I recently wondered how to support sorting an array object. I saw two solutions. There first one is used by the MiscSortedList class which is included in the MiscKit distribution. There second solution is used by NSArray and NSMutableArray in the foundation kit. First I will present both solutions, then I will show you how I would do it. 1) MiscKit Every object which is included in a MiscSortedList array must conform to the MiscCompare protocol which means it must implement a method called "compare:". MiscSortedList compares two objects "anObjectA" and "anObjectB" by calling [anObjectA compare:anObjectB] 2) Foundation Kit An NSArray object can return a sorted array using one of these methods: - (NSArray *)sortedArrayUsingSelector:(SEL)comparator - (NSArray *)sortedArrayUsingFunction:(int(*)(id, id, void *))comparator context:(void *)context Note that the sorting context is supplied in the method call. In the MiscKit solution, an object knows how to compare itself with some other object. I used the MiscSortedList in a newsreader project. The user can sort articles by subject, by sequence number, by date or by the authors name. The class "Article" has a class method +setSortType:(int)type to set how the list should be sorted. The "compare:" method looks like this: - (int)compare:someObject { switch(sortType){ case SORT_BY_DATE: .... break; case SORT_BY_SUBJECT: .... break; case SORT_BY_NUMBER: .... break; } } This solution is not very oo-like. The case statement should be replaced by polymorphism (an early version of Eiffel even did not have such thing --- guess why) So I came up with my solution (I guess other oo kits do exactly the same thing? This is an obvious solution.): A sortable array has the method - sortArrayUsingComparator:id<Comparator>aComparator A comparator object is used to compare two objects by calling [aComparator compareObject:anObjectA withObject:anObjectB]. The method compareObject:withObject: is defined in the Comparator protocol. This solution is quite similar to the foundation solution, but it uses objects not C functions. A design rule for reusable oo design is: "Encapsulate the context that varies" In this case of sorting an array, 1) the sorting algorithm is fixed 2) the way comparing is done is variable Now one can sort an array differently without modifying the sorted objects class, just be writing a new Comparator class. Some other interesting things I will use in the new version of my newsreader are iterators and filters. Iterators allready are a well known concept (NSEnumerator protocol). Filters just implement a method -(BOOL)accept:anObject which accepts or rejects anObject. If you are using NewsGrazer or Alexandra, you will see a matrix in a scroller with a list of newsgroups in the upper left corner. If this matrix defines a method: -(BOOL)selectCellUsingEnumerator:anEnumerator andFilter:aFilter which uses the iterator to iterate through the cell list and uses the filter to find the right cell to select, one can do lots of things just by feeding a different iterator and filter object to the method. For example - Select the next/previous article/newsgroup. - Select the next/previous unread article. - Select the next/previous article with a certain subject, author or .... - Select the next/previous newsgroup, which name includes some substring. Filter objects can be used in other contexts as well, for example to decide which newsgroups/articles to display in the matrix. A single method - showGroupsUsingFilter:aFilter can be used to 1) Show all groups 2) Show all subscribed groups 3) Show all groups with at least one unread article 4) Show all groups which name includes some substring etc Have a nice day -- <--------------------------------------------------------------> | Constantin Szallies ||| tini@gurke.ping.de | | Emil Figge 7 44227 Dortmund (o o) Tel: 0231/7519681 | <------------------------------oOo--(_)--oOo------------------->
From: yibwu@napoleon.plato.sky.bdm.com (Yibing Wu) Newsgroups: comp.sys.next.programmer Subject: How can I use gdb in Emacs? Date: 29 Jan 1996 19:37:25 GMT Organization: BDM International, Inc. Message-ID: <4ej7ll$h77@news.mcl.bdm.com> I am using GNU Emacs for NeXT as my development tool. However, the GDB interface is different, I guess it is because the executables and .m .h files are not in the same directory, the gdb always give me ".m" file not found error. What should I do? Thanks, Yibing Wu
From: ivo.welch@agsm.ucla.edu (Ivo Welch) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Newer than gcc 2.7.0 ? Followup-To: comp.sys.next.sysadmin Date: 29 Jan 1996 19:05:00 GMT Organization: The Anderson School at UCLA Message-ID: <4ej5os$24n@risc.agsm.ucla.edu> My gcc 2.7.0 compiler for NS regularly croaks with the message that there is an internal compiler bug and asks me to submit a full bug report. Because 2.7.0 is no longer the last version, and my input file is quite complex, I presume this is not worthwhile. Is there a binary version of a later version of gcc for NS/FIP available somewhere? /ivo welch -- Ivo Welch ivo.welch@anderson.ucla.edu Assoc Prof of Finance Anderson GSM at UCLA 110 Westwood Plaza, Box 951481, LA CA 90095-1481
From: xander@interlog.com (A Dobrowolski) Newsgroups: comp.sys.next.programmer Subject: PROGRAMMER WANTED!!! Date: Mon, 29 Jan 1996 15:49:15 -0500 Organization: InterLog Internet Services Message-ID: <xander-2901961549150001@xander.interlog.com> CONTRACT WORK POSITION: Programmer. REQUIREMENTS: Excellent knowledge of clipper (XBASE) Experience in Data Communications knowledge of Windows Knowledge of Network Environments PROJECT: To assist in the development of a costomized Data Entry program for a specific proffesional clientele. The application involves file transfers to and from various platforms. PROJECT: To design form output for a laser printer. Please send replies to: xander@interlog.com
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical NXSplitView Date: 29 Jan 1996 20:51:01 GMT Organization: ftn Internet Message-ID: <4ejbvl$bfo@master.ftn.net> I'd like to thank everyone who responded to my enquiry!! Since the list was rather long, I thought that I would respond publicly and thank everyone in this forum! For those who were also wondering what the answer was, check out the ZooView MiniExample on ftp.next.com. This gives a very clear and concise example of how to accomplish this. Again, thanks to all who responded! Regards, -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: stevem@radium.ca (Stephen MacDougall) Newsgroups: comp.sys.next.programmer Subject: Vertical NXSplitView Date: 29 Jan 1996 20:51:02 GMT Organization: ftn Internet Message-ID: <4ejbvm$bfo@master.ftn.net> I'd like to thank everyone who responded to my enquiry!! Since the list was rather long, I thought that I would respond publicly and thank everyone in this forum! For those who were also wondering what the answer was, check out the ZooView MiniExample on ftp.next.com. This gives a very clear and concise example of how to accomplish this. Again, thanks to all who responded! Regards, -- ==================================================== Stephen MacDougall Director of Research and Development RADium Technology Centre (Canada) stevem@RADium.ca NeXT and MIME mail gladly accepted ====================================================
From: David Hinz <dhinz@dna406.dna.mci.com> Newsgroups: comp.sys.next.programmer Subject: Q: Sybase and EOF Date: Mon, 29 Jan 1996 16:31:13 -0700 Organization: MCI Message-ID: <310D58C1.167E@dna406.dna.mci.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I tried using EOmodeler for the first time to connect to a remote Sybase server. After entering the database name, server, name, and password I got an error that the server name was not in the interfaces file. A few questions. 1) Do I need any local Sybase software on my system to connect to a remote Sybase server? Or does EOF handle all of the local stuff. 2) Is there documentation somewhere on how to setup a interfaces file for Sybase? Anything documented in the Bookshelves? 3) Will the Sybase adaptor that is shipped with EOF work with Sybase 10.x? Thanks for any assistance. dave. -- David Hinz --- MCI Telecommunications Corporation E-mail: dhinz@dna406.dna.mci.com Pager: 1-800-5Page50 PIN: 101-6795 Phone: (303) 291-6108 VNET: 636-6108 Fax: (303) 291-6365
From: yibwu@napoleon.plato.sky.bdm.com (Yibing Wu) Newsgroups: comp.sys.next.programmer Subject: Can I put EOModel in a bundle?! Date: 30 Jan 1996 02:34:21 GMT Organization: BDM International, Inc. Message-ID: <4ek03d$q3k@news.mcl.bdm.com> I created a bundle in which there is a nib file that has a EOController in it. The EOController was first instantiated from a EOModel created in EOModeler. It works fine as long as I have the EOModel in the top-most "Other Resources" folder in that project. In other words, if I remove the EOModel (i.e.*.eomodel file) from the top-most "Other Resources" folder, and put it to that of the bundle's (that is where it is only needed anyway), I had problem creating the EODatabaseSource for the EOController (error message: Cannot create DataSource for Model xxx Entity yyy). I tried to disconnect the EODatabaseSource from the EOController, however, it does not prevent the DatabaseSource from initializing itself during "NibLoading". Either there is way allows me to move the EOModel (*.eomodel file) to the "Other Resources" folder in the Bundle directory, or somehow I can stop the EOController in my code from initializing its datasource, either way helps. Thanks, Bing P.S. A little background intro of why I am doing this: We had a project that all the sub-projects are sharing one big *.eomodel file in the top-most "Other Resources" folder. Now, we are in a second phase dev. which require us to regroup some of the modules in different projects.
From: altenber@acpub.duke.edu (Lee Altenberg) Newsgroups: comp.sys.next.programmer Subject: PopUpList woes- changes class Date: 29 Jan 1996 20:56:53 -0500 Organization: Duke University, Durham, NC, USA Message-ID: <4ejtt5$dfo@news.duke.edu> I am trying to get PopUpLists to work, and getting peculiar behavior. It changes its object class from a PopUpList (when application is launched) to a Matrix (when pop-up list is clicked). This is NS/I 3.3. Here is what happens: I pulled a pop-up list from the IB pallete to a window, and connect the cells to a method "whatAmI" in my Controller.m file: - whatAmI:sender { printf("This object is a %s\n", [sender name]); } When the application is launched, in my console I get: This object is a PopUpList When I click on the PopUpList, however, I now get: This object is a Matrix What the hell is going on here? I don't want the first value, since in my real program I want the pop-up list to call : [[sender selectedCell] title]; and this crashes the program when launched because PopUpList doesn't have selectedCell as a method. As a model, I looked at the BusyBox example application, but when it launches, the pop-up list that is connected to the method "changeAnimal" in the AnimalView.m object does not get called. The first time you click it, it will say it's a Matrix. So I guess the question is, why in my application does the pop-up list get called upon launching the application, and why is it a PopUpList instead of the Matrix I want it to be? Thanks for any tips, Lee Altenberg altenber@mhpcc.edu
From: miz@jsg.co.jp (Ryuichi MIZUNO) Newsgroups: comp.sys.next.programmer Subject: Q. How can I connect to the remote server on other domain via (P)DO ? Date: 30 Jan 1996 06:35:03 GMT Organization: Internet Initiative Japan, Inc. Message-ID: <4eke6n$28a@news3.iij.ad.jp> Keywords: DO,distributed Objects,lookupd,NXConnection Hi, this is Ryuichi Mizuno, a programming staff of NeXus,Tokyo,Japan. ---- Q. How can I connect to the remote server on other domain via (P)DO ? ---- My DO application on NEXTSTEP 3.2J/3.3J works well under LAN environment . When a client checks in the server, I usually write below. remoteHub = [NXConnection connectToName:"FooServer" onHost:LOCAL_DOMAIN_SERVER_NAME]; if (remoteHub){ ourServer = [remoteHub connectionForProxy]; [remoteHub setProtocolForProxy:@protocol(SomePrivateProtocols)]; } LOCAL_SERVER_NAME is like graf,seles,chang,agassi,... From netInfo, they can be seen in the same domain on which the client host locates. Every host can be found as a server if only the DO server process is running. That is very nice. Now I will try to connect to the server running on other domain. remoteHub = [NXConnection connectToName:"FooServer" onHost:OTHER_DOMAIN_SERVER_NAME]; ... OTHER_DOMAIN_SERVER_NAME is like dolphin.ocean.linc.or.jp,...,which can't be seen from our netinfo, but telnet/rlogin is available by my account and it has no firewall. Unfortunately, My DO client app can't connect to the OTHER_DOMAIN_SERVER yet. How should I setup the network configuration for the DO server host and client hosts ? Or my implementation has anything wrong ? I think that while host name is resolved by DNS or by lookupd, -connectToName:onHost: must work good. Why NOT ? I'm very happy if you let me know any suggestions. Best Regards, ---- Ryuichi Mizuno (miz@jsg.co.jp NeXTmail welcome) Jidouseigyo Institute Inc. (voice)+81-3-3776-1213 (fax)+81-3-3775-1142 ---- NeXus (NeXT User's society),Tokyo,Japan ----
From: gshaw@zeta.org.au (Greg Shaw) Newsgroups: comp.sys.next.programmer Subject: mixing Objective-C and C++ Date: 30 Jan 1996 09:14:13 GMT Organization: Kralizec Dialup Internet Sydney, +61-2-837-1183 V.32bis Message-ID: <4eknh5$lqb@gidora.kralizec.net.au> Keywords: C++, Objective-C Has anyone had experience using both Objective-C and C++ in the one application. Even more so, using both under NeXTSTEP InterfaceBuilder. Basically, I have a client with tools in C++ wanting to use OpenStep as their development environment. I either have to combine the two in some manner or get them to move to Objective-C. Although I would prefer the latter the first would be the most likely to succeed in the short term. Any help would be appreciated. Thanks, Greg Shaw.
From: gshaw@zeta.org.au (Greg Shaw) Newsgroups: comp.sys.next.programmer Subject: mixing Objective-C and C++ Date: 30 Jan 1996 09:14:43 GMT Organization: Kralizec Dialup Internet Sydney, +61-2-837-1183 V.32bis Message-ID: <4ekni3$lqk@gidora.kralizec.net.au> Keywords: C++, Objective-C Has anyone had experience using both Objective-C and C++ in the one application. Even more so, using both under NeXTSTEP InterfaceBuilder. Basically, I have a client with tools in C++ wanting to use OpenStep as their development environment. I either have to combine the two in some manner or get them to move to Objective-C. Although I would prefer the latter the first would be the most likely to succeed in the short term. Any help would be appreciated. Thanks, Greg Shaw.
From: gshaw@zeta.org.au (Greg Shaw) Newsgroups: comp.sys.next.programmer Subject: mixing Objective-C and C++ Date: 30 Jan 1996 20:16:46 +1100 Organization: Kralizec Dialup Unix Sydney, +61-2-837-1183 V.32bis Message-ID: <4eknlu$ar@godzilla.zeta.org.au> Keywords: C++, Objective-C Has anyone had experience using both Objective-C and C++ in the one application. Even more so, using both under NeXTSTEP InterfaceBuilder. Basically, I have a client with tools in C++ wanting to use OpenStep as their development environment. I either have to combine the two in some manner or get them to move to Objective-C. Although I would prefer the latter the first would be the most likely to succeed in the short term. Any help would be appreciated. Thanks, Greg Shaw.
Newsgroups: comp.sys.next.programmer From: ian.stephenson@insignia.co.uk Subject: Re: Q: Sybase and EOF Message-ID: <DM02pH.sL@isltd.insignia.com> Sender: news@isltd.insignia.com Organization: Insignia Solutions plc References: <310D58C1.167E@dna406.dna.mci.com> Date: Tue, 30 Jan 1996 15:22:28 GMT In article <310D58C1.167E@dna406.dna.mci.com> David Hinz <dhinz@dna406.dna.mci.com> writes: > I tried using EOmodeler for the first time to connect to a remote Sybase > server. After entering the database name, server, name, and password I > got an error that the server name was not in the interfaces file. A few > questions. > > 1) Do I need any local Sybase software on my system to connect to a > remote Sybase server? Or does EOF handle all of the local stuff. No more than you should have if you've done a standard install of NeXTStep (ie none are part of the sybase distribution). Get /usr/sybase/bin/isql going first, before trying the EOF stuff. > 2) Is there documentation somewhere on how to setup a interfaces file > for Sybase? Anything documented in the Bookshelves? add to /usr/sybase/interfaces: MYSERVERNAME query tcp ether host.domain.etc.com 2500 Just drop in the name you want to call your server, the full ,machine name of your server, and the final number is the port number you've configed sybase to use on the remote machine. If your server is set up OK then you should run /usr/sybase/bin/isql -SMYSERVERNAME and you should be connected. After that, try EOF, and it should just work... > 3) Will the Sybase adaptor that is shipped with EOF work with Sybase > 10.x? Thats what I'm using - works fine. Ian
From: sl@skynet.oph.rwth-aachen.de (Stefan Leuker) Newsgroups: comp.sys.next.programmer Subject: Q: How to get the complete host name? Date: 30 Jan 1996 16:47:57 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Distribution: world Message-ID: <4eli3t$dl8@news.rwth-aachen.de> Hi there! How do I get the full host name programmatically? E.g., my system is called skynet.oph.rwth-aachen.de, but a call to gethostname returns only skynet. How do I get the domain name, or the full name including the domain name? Any hint is wellcome Stefan
Newsgroups: comp.sys.next.programmer From: ktwu@alw.nih.gov (Kuo-Tsung Wu) Subject: Help: building gcc-2.7.2 Message-ID: <1996Jan30.230611.29244@alw.nih.gov> Sender: postman@alw.nih.gov (AMDS Postmaster) Organization: NIH/NIDCD/ESDSB Date: Wed, 31 Jan 1996 00:05:06 GMT I'm pretty new on C programming and hope to build gcc-2.7.2 on my NeXTstation running 3.2OS. There are lot of packages on GNU that related to gcc, such as glibc, libg++, gdb, gdbm g77 and binutils which include as and ld. I've tried on most of them and have no luck on some of the packages. I would be indebeted if someone can answer a few elementary questions. If this is a FAQ, my apology. Please point me to the available source. 1). I was unable to compile as and ld included in binutils. Would gcc-2.7.2 work with as and ld that come with 3.2OS? If yes, what should I do during compilation to make it work? 2). With some help, I was able to compile bison. Could someone tell me what is the most complete packages of gcc that can be build on NeXTstation with 3.2OS? Much thanks. KT -- Kuo-Tsung Wu ktwu@alw.nih.gov NIH/NIDCD/ESDSB voice: 301/402-1843 Rockville, MD fax: 301/402-0390
From: wombat@smartt.com Newsgroups: comp.sys.next.programmer Subject: Possible to use Developer 3.2 with NS FIPS 3.3? Date: Tue, 30 Jan 1996 17:05:55 -0800 Organization: KTK Commmunications Ltd. (SmarttNet) Message-ID: <310EC073.DB5@smartt.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, Does anyone know for sure whether it is possible to use NEXSTEP Developer on an Intel system running NEXSTEP 3.3? The documentation for Devloper says something like "make sure you are using the same version of NEXTSTEP as the version of Developer you are going to install..." I wonder if there are any serious problems? Thanks, Shea Phillips. wombat@smartt.com
From: hkhenson@cup.portal.com (H Keith Henson) Newsgroups: comp.sys.next.programmer Subject: Odd memory problem Date: 30 Jan 1996 19:00:46 -0800 Organization: The Portal System (TM) Sender: pccop@unix.portal.com Distribution: world Message-ID: <150162@cup.portal.com> We are having an odd problem with some code. This is an app which eats a lot of memory space, hundreds of meg at the moment. Our problem is that after a considerable amount of activity (typically tens of minutes of near 100% processer time) the app breaks on trying to access memory which the OS does not think belongs to it. By tracing on the virtual memory release, and logging this data we can tell that our app has *not* returned the memory block which is causing the problem to the system. Somehow the OS has lost track of this memory belonging to the process. (Or such is my speculation.) While the problem is sort of reproduceable, the place where it breaks is not consistent--it is random over a wide range, with something on the order of a 0.5% chance of failing on a cycle. Your thoughts on this problem would be appreciated. We are using GDB. I have not yet tried to figure out how to look at the memory at some other level to see if the contents got stomped on. I know somewhat more about the way unix memory management works. It was a little surprising that mach permits "allocating" much larger amount of memory space than the physical and/or available disk. Any suggestions would be much appreciated. Keith Henson
From: hwr@pilhuhn.de (Heiko W.Rupp) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.mis Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 31 Jan 1996 06:05:33 GMT Organization: The Home Of The Pilhuhn Message-ID: <slrn4gu1ld.egn.hwr@snert.pilhuhn.de> References: <4e84c1$m0n@axl02it.ntc.nokia.com> In article <4e84c1$m0n@axl02it.ntc.nokia.com>, Alan Barclay wrote: >is already compatible with m68k architecture and Plan 9 sources, from AT&T, >contain the specific hardware information required to implement low level Plan 9 only supports Cube and MonoStations, no color. I guess der Mouse who wrote Mouse X and who hangs around the NetBSD mailinglists, could perhaps help for the display. -- Heiko W.Rupp hwr@pilhhuhn.de INN FAQ can be found in ftp://ftp.xlink.net/pub/news/docs/
From: cwolf@wolfware.com (Christopher A. Wolf) Newsgroups: comp.sys.next.programmer Subject: Compresshelp error messages during build? Date: 31 Jan 1996 05:34:02 GMT Organization: WolfWare (http://www.wolfware.com) Message-ID: <4emv0a$5iu@shellx.best.com> I used HelpBuilder 1.3 to generate some menu documentation but when I go to build the application I get the following warnings from compresshelp: compresshelp warning: Files Newsgroups/Unsubscribe.cmd.rtfd/menupath1.eps and Kill/Kill_Panel.cmd.rtfd/menupath1.eps have the same sum but their contents differ. compresshelp warning: Files Kill/Kill_Subject.cmd.rtfd/menupath1.eps and Main/Articles.mn.rtfd/menupath1.eps have the same sum but their contents differ. compresshelp warning: Files Main/Articles.mn.rtfd/menupath1.eps and Articles/Mark_Unread.cmd.rtfd/menupath1.eps have the same sum but their contents differ. compresshelp warning: Files Thread/Select.cmd.rtfd/menupath1.eps and Save/Extract_Binary.cmd.rtfd/menupath4.eps have the same sum but their contents differ. ... and it goes on like that for -quite- a while. Everything seems to build and work OK but I was wondering if anyone could enlighten me about what these warnings mean and how to avoid them. Thanks! -- Christopher Wolf / WolfWare cwolf@wolfware.com (NeXTmail & MIME accepted) For information about the NewsFlash newsreader for NeXTSTEP check out WolfWare's home page: http://www.wolfware.com/
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: lockf Date: 30 Jan 1996 21:52:15 GMT Organization: Turbocat's Development, Germany Message-ID: <4em3uf$ru@turbocat.snafu.de> Hello! Does lockf also work on NFS-mounted volumes? Thanks. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Possible to use Developer 3.2 with NS FIPS 3.3? Date: 31 Jan 1996 02:51:22 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4emlfa$f7i@usenet.rpi.edu> References: <310EC073.DB5@smartt.com> wombat@smartt.com wrote: > Does anyone know for sure whether it is possible to use NEXSTEP > Developer on an Intel system running NEXSTEP 3.3? The documentation > for Devloper says something like "make sure you are using the > same version of NEXTSTEP as the version of Developer you are > going to install..." It is absolutely certain that you can run NS-3.2 developer on a NS-3.3 User system. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu ITS Systems Programmer (handles NeXT-type mail) Rensselaer Polytechnic Institute; Troy NY USA
From: kloppen@zorro.gmd.de (Jelske Kloppenburg) Newsgroups: comp.sys.next.programmer Subject: Re: Drag and drop Date: 31 Jan 1996 09:32:28 GMT Organization: GMD, Sankt Augustin, Germany Message-ID: <4encvc$5nr@omega.gmd.de> References: <4eh966$djc@news1.ucsd.edu> trombino@wendy.ucsd.edu (Mark Trombino) wrote: > Hello, > I'm having a bit of a problem implementing drag and drop in an application, > and I was hoping that someone out there could please help he out... > I've used code from the developer's examples and from Solitaire.app, but what > seems to be happening is that > I'm not getting any messages sent to my sending view (or the view that > initiates the drag session), other than > - (NXDragOperation)draggingSourceOperationMaskForLocal: > No other messages are sent to the initiating view. The receiving view is > getting every message that its supposed to, and the draggingPBoard is > functioning as expected. What gives? Am I missing something really obvious > here? I've got the message - draggedImage:(NXImage *)image endedAt:(NXPoint *)screenPoint deposited:(BOOL)didDeposit too. Is it that what You want? I can't remember doing anything extra for that. > Please help!! I've looked endlessly over source code examples and can't find > out what I'm doing wrong! If You want another source code example, I've made me a new AltDock from the sources of the old AltDock (1989, for NS 1.0). -- Jelske Kloppenburg, +49 2241 14-2433, <Jelske.Kloppenburg@gmd.de> GMD - German National Research Center for Information Technology "Don't kill the Winners!" Dennis Tsichritzis > Thanks in advance! > -- > ================================================================== > Mark Trombino > trombino@wendy.ucsd.edu > Center For Research in Computing the Arts, U.C.S.D. > ================================================================== -- Jelske Kloppenburg, +49 2241 14-2433, <Jelske.Kloppenburg@gmd.de> GMD - German National Research Center for Information Technology "Don't kill the Winners!" Dennis Tsichritzis
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: First WebObjects question Date: 31 Jan 1996 09:34:25 GMT Organization: University of Heidelberg, Germany Message-ID: <4end31$39k@sun0.urz.uni-heidelberg.de> Well, I've downloaded WOF, installed it and checked that all the examples work (TimeOff is even cooler than Dodge-Lite, but S_L_O_W). Now, I want to try my own, a simple example connecting to a miniSQL database (just a simple address-book type web-app). How? There's nothing at all in the WOF documentation on how to do this. ( see http://www.next.com/Pubs/Documents/WebObjects/ ) Though the docs are not complete, there's not even a hint of _how_to_ connect to an external datasource... There's no typing, and no (documented) way to import headers: how can I connect to a custom object (even one that only takes & returns NSArrays , NSDicts & so on...)? I tried using DO, but that didn't work: NXConnection: does not recognize selector +methodSignatureForSelector: I'm getting the impression that this free beta version is "WOF-Lite" and that without WOF-Pro or WOF-Enterprise, what I want to do won't be possible. your Web Manic-depressive, Dave ------------------------------------------------------------------ David A. Coyle Father of Annealer.app. Fission-track research. Ask for my PGP public key. ------------------------------------------------------------------ Max-Planck-Institut f r Kernphysik Heidelberg, Germany ------------------------------------------------------------------ dcoyle@goanna.mpi-hd.mpg.de dcoyle@weizen.rt.schwaben.de
From: /usr/lib/NextStep/Resources/German.lproj/AlertPanel.nib/data.nib@ Newsgroups: comp.sys.next.programmer Subject: Re: PopUpList woes- changes class Date: 31 Jan 1996 09:59:47 GMT Organization: GMD, Sankt Augustin, Germany Message-ID: <4eneij$5nr@omega.gmd.de> References: <4ejtt5$dfo@news.duke.edu> altenber@acpub.duke.edu (Lee Altenberg) wrote: > I am trying to get PopUpLists to work, ... I connected the trigger Button in IB to my outlet midiChanPopUp and did the rest in my class: - awakeFromNib { . . . midiChanList = [[midiChanPopUp target] itemList]; [[midiChanList setTarget:self] setAction:@selector(takeMidiChan:)]; . . . return self; } - takeMidiChan:sender { int midiChannel = [sender selectedTag]; . . . return self; } see the class description of PopUpList. -- Jelske Kloppenburg, +49 2241 14-2433, <Jelske.Kloppenburg@gmd.de> GMD - German National Research Center for Information Technology "Don't kill the Winners!" Dennis Tsichritzis > Lee Altenberg > altenber@mhpcc.edu -- Jelske Kloppenburg, +49 2241 14-2433, <Jelske.Kloppenburg@gmd.de> GMD - German National Research Center for Information Technology "Don't kill the Winners!" Dennis Tsichritzis
From: Jelske Kloppenburg <Jelske.Kloppenburg@gmd.de> Newsgroups: comp.sys.next.programmer Subject: Re: PopUpList woes- changes class Date: 31 Jan 1996 10:45:05 GMT Organization: GMD, Sankt Augustin, Germany Message-ID: <4enh7h$57o@omega.gmd.de> References: <4ejtt5$dfo@news.duke.edu> Please excuse the wrong posting from "/usr/lib/.... " Now I'm using NewsGrazer. j.k. -- Jelske Kloppenburg, +49 2241 14-2433, <Jelske.Kloppenburg@gmd.de> GMD - German National Research Center for Information Technology "Don't kill the Winners!" Dennis Tsichritzis
From: lin@lorien.umd.edu (Lin) Newsgroups: comp.sys.next.programmer Subject: open and install package on tty Date: 31 Jan 1996 10:59:04 GMT Organization: University of Maryland, College Park Message-ID: <4eni1o$e89@hecate.umd.edu> Is there any way to install a package from a remote site ( UNIX, not next ) instead of on the console ?
From: hugues@precipice.fdn.fr (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4eh7us$kfe@master.ftn.net> Control: cancel <4eh7us$kfe@master.ftn.net> Date: 30 Jan 1996 19:29:47 GMT Organization: Individual - France Message-ID: <4elrjb$17f@precipice.fdn.fr> cancel
From: hugues@precipice.fdn.fr (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4eh9da$kfe@master.ftn.net> Control: cancel <4eh9da$kfe@master.ftn.net> Date: 30 Jan 1996 19:29:56 GMT Organization: Individual - France Message-ID: <4elrjk$17f@precipice.fdn.fr> cancel
From: hugues@precipice.fdn.fr (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4eh9j2$kfe@master.ftn.net> Control: cancel <4eh9j2$kfe@master.ftn.net> Date: 30 Jan 1996 19:30:00 GMT Organization: Individual - France Message-ID: <4elrjo$17f@precipice.fdn.fr> cancel
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 31 Jan 1996 12:43:59 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4eno6f$m1j@lion.embl-heidelberg.de> References: <4eli3t$dl8@news.rwth-aachen.de> In article <4eli3t$dl8@news.rwth-aachen.de> sl@skynet.oph.rwth-aachen.de (Stefan Leuker) writes: > How do I get the full host name programmatically? > > E.g., my system is called skynet.oph.rwth-aachen.de, > but a call to gethostname returns only skynet. > > How do I get the domain name, or the full name including the domain name? gethostname() + getdomainname(); I think they are even posix compatible, but cannot guarantee... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387524 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: david@pfi.ibk.baum.ethz.ch (David EKCHIAN) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Friends in Objective-C and C++. Date: 31 Jan 1996 13:54:48 GMT Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4ensb8$pgh@elna.ethz.ch> Hi all, I develop a computer program for bridge construction simulation. The back-end is in C++, the GUI in Objective-C. I use the Controller-Model-View technique and want to declare my view objects (in Objective-C) as "friends" for the model objects (in C++). Have you ever tried to do this? I have some problems to declare this friendship. Many thanks in advance, David. PS: I have another question since one week without answer under the title "Help: Malformed Mach-o file." -- o _ /-;c ___ David C. EKCHIAN ______________________________________(@)#\(@)___
Newsgroups: comp.sys.next.programmer From: ian.stephenson@insignia.co.uk Subject: Re: First WebObjects question Message-ID: <DM1x45.Arq@isltd.insignia.com> Sender: news@isltd.insignia.com Organization: Insignia Solutions plc References: <4end31$39k@sun0.urz.uni-heidelberg.de> Date: Wed, 31 Jan 1996 15:16:53 GMT In article <4end31$39k@sun0.urz.uni-heidelberg.de> dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) writes: > Well, I've downloaded WOF, installed it and checked that all the examples > work (TimeOff is even cooler than Dodge-Lite, but S_L_O_W). > > Now, I want to try my own, a simple example connecting to a miniSQL > database (just a simple address-book type web-app). > > I'm getting the impression that this free beta version is "WOF-Lite" and > that without WOF-Pro or WOF-Enterprise, what I want to do won't be > possible. > > your Web Manic-depressive, > http://www.next.com/WebObjects/Products.html pretty much confirms this. talking to Obj-C will cost you $3000, while talking to EOF will cost you $25000. Running the demo's is free! To be fair, WOF-Lite gives you access to a few things that might be cool, and $3000 for WOF-Pro is probably a fair price (though remember that 5 years ago $3000 would by you a NeXTStation with full developer!). Unfortunatly to do anything usefull is going to require selling my house! WOF is just not an option for most of the people who have been waiting for it. I've got a protoype app that we'd love to deploy cross platform, and we were prepared to spend sensible money to do it (say WOF-Pro prices), but $25000 is just a bit steep to service a dozen users who need access prehaps 1 hour a week. For that money they'll just have to walk to the Cube in the corner and work there. I can see that for big corporate sites (ie the target market) writing the corporate web page that this stuff is valuable, but I can't see many companies justifying this for in-house deployment - despite the fact that they're desprate for such a solution. Prehaps what is needed is some price scaling based on the number of users, usage, and access time - a cube running in house, is very different to a SparcServer with megaStream links to the net. One is worth spending $25000 on - the other is more difficult to justify. Ian [I suspect that this could be the start of a long running discussion] [Forget "Developer is expensive!" ]
Newsgroups: comp.sys.next.programmer Subject: GNU Fortran on NeXTSTEP ? Message-ID: <1996Jan30.210014.20210@roper.uwyo.edu> From: norbert@uwyo.edu (Norbert Pirzkal) Date: 30 Jan 96 21:00:14 MST Distribution: world Organization: University of Wyoming MIME-Version: 1.0 Content-Type: Text/Plain; charset=US-ASCII Has anyone gotten GNU FORTRAN to compile/work under NeXTSTEP 3.3 (intel) ? Is there a copy of working binaries available for ftp anywhere?? Nor
Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.mis From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Sender: news@novice.uwaterloo.ca (system PRIVILEGED account) Message-ID: <DM1y1q.H9p@novice.uwaterloo.ca> Date: Wed, 31 Jan 1996 15:37:02 GMT References: <4e84c1$m0n@axl02it.ntc.nokia.com> <slrn4gu1ld.egn.hwr@snert.pilhuhn.de> Organization: University of Waterloo In article <slrn4gu1ld.egn.hwr@snert.pilhuhn.de>, Heiko W.Rupp <hwr@pilhuhn.de> wrote: >In article <4e84c1$m0n@axl02it.ntc.nokia.com>, Alan Barclay wrote: > >>is already compatible with m68k architecture and Plan 9 sources, from AT&T, >>contain the specific hardware information required to implement low level > >Plan 9 only supports Cube and MonoStations, no color. > >I guess der Mouse who wrote Mouse X and who hangs around the NetBSD >mailinglists, could perhaps help for the display. > He likely could. Course, us Dimension people are once again out of luck (unless someone really likes hacking i860 binaries with otool. :-)) -- David Evans (NeXTMail OK) dfevans@bbcr.uwaterloo.ca Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/ University of Waterloo "Default is the value selected by the composer Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: sl@skynet.oph.rwth-aachen.de (Stefan Leuker) Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 31 Jan 1996 18:33:27 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Distribution: world Message-ID: <4eocln$ea6@news.rwth-aachen.de> References: <4eno6f$m1j@lion.embl-heidelberg.de> In article <4eno6f$m1j@lion.embl-heidelberg.de> tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) writes: > > How do I get the full host name programmatically? > > gethostname() + getdomainname(); getdomainname() returns an empty string. What is wrong with my config? Stefan
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: Help: building gcc-2.7.2 Date: 31 Jan 1996 20:18:02 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4eoipq$e8n@agate.berkeley.edu> References: <1996Jan30.230611.29244@alw.nih.gov> In article <1996Jan30.230611.29244@alw.nih.gov>, Kuo-Tsung Wu <ktwu@alw.nih.gov> wrote: >1). I was unable to compile as and ld included in binutils. > Would gcc-2.7.2 work with as and ld that come with 3.2OS? > If yes, what should I do during compilation to make it work? gcc-2.7.2 builds and runs fine under 3.2OS with the standard assembler and linker. I don't remember having any troubles building it, or if I did have some troubles, they were minor. >2). With some help, I was able to compile bison. Could someone > tell me what is the most complete packages of gcc that can be > build on NeXTstation with 3.2OS? Can't help you there. I compiled gcc so that I could compile and run some standard X-windows and UNIX software with minimal modifications. If I were planning on programming C applications for general use on other UNIX machines, I wouldn't use (GNU or other) C++ or objective C. If I were writing for NeXT machines, I would use NeXT's native compiler. Best, Paul
From: Joe Freeman <Joe@FreemanSoft.COM> Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: Wed, 31 Jan 1996 18:16:23 -0500 Organization: FreemanSoft Inc. Message-ID: <310FF847.7F5@FreemanSoft.COM> References: <4end31$39k@sun0.urz.uni-heidelberg.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: "David A. Coyle" <dcoyle@goanna.mpi-hd.mpg.de> David A. Coyle wrote: > Now, I want to try my own, a simple example connecting to a miniSQL > database (just a simple address-book type web-app). You may be able to load in a bundle using [[NSBundle xxx] principalClass] under NEXTSTEP. There may be an issue of whether it was linked against all of the shlibs. WOF "too-lite" appears to be a script only release for folks that don't do custom objects. > I'm getting the impression that this free beta version is "WOF-Lite" and > that without WOF-Pro or WOF-Enterprise, what I want to do won't be > possible. You may also be able to DO out to a remote process to do the database access but I'm not sure DO is there either. It will be if they are using the DO version of the CGI helper app but .... -- The governer of MD wants to spend $300M on a rent free pro football stadium which is only used 10 times a year. For whom is he really doing this deal?
From: yibwu@napoleon.plato.sky.bdm.com (Yibing Wu) Newsgroups: comp.sys.next.programmer Subject: Oracle OCI library for NeXT? Date: 1 Feb 1996 02:49:04 GMT Organization: BDM International, Inc. Message-ID: <4ep9n0$em@news.mcl.bdm.com> We are trying to use PL/SQL to fully utilize the Oracle database; However, I don't know if there is a NeXT compatible OCI library or ProC compiler available? The assumption I have here is: in NeXT, you can do no more than Insert, delete, update and select. EOAdaptorChannel supports methods to form the sql expression and send it internally to the physical database server, however, if you want to get a unique key or something, you cannot use any existing EOF objects. Please correct me if I am wrong. Bing
From: tm8025a@mailhost.soc.american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 1 Feb 1996 02:26:55 GMT Organization: American University, Washington DC Message-ID: <4ep8df$i3@paladin.american.edu> References: <4end31$39k@sun0.urz.uni-heidelberg.de> In article <4end31$39k@sun0.urz.uni-heidelberg.de> dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) writes: :Well, I've downloaded WOF, installed it and checked that all the examples :work (TimeOff is even cooler than Dodge-Lite, but S_L_O_W). : :Now, I want to try my own, a simple example connecting to a miniSQL :database (just a simple address-book type web-app). : :How? : You wait until March and shell out the $25,000 for WebObjects Enterprise. I think the beta is only for simple scripting and to teach people the basics of the product. -- Torrey McMahon
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 1 Feb 1996 04:54:57 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4eph31$4j6@emerald.oz.net> References: <4end31$39k@sun0.urz.uni-heidelberg.de> <4ep8df$i3@paladin.american.edu> tm8025a@mailhost.soc.american.edu (Torrey McMahon) wrote: > dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) writes: > :Well, I've downloaded WOF, installed it and checked that all the examples > :work (TimeOff is even cooler than Dodge-Lite, but S_L_O_W). > :Now, I want to try my own, a simple example connecting to a miniSQL > :database (just a simple address-book type web-app). > :How? > You wait until March and shell out the $25,000 for WebObjects Enterprise. I > think the beta is only for simple scripting and to teach people the basics > of the product. NeXT needs to get a lot of feedback about this pricing which effectively prevents reuse of existing EOF code from within WebObjects unless you are a large, wealthy site that can afford WebObjects Enterprise. Of course, you could use WebObjects Pro (*only* $2,500) and use the miniSQL access library in a compiled WebObjects app. Or if miniSQL supports access from Perl or another scripting language, you might be able to use the free version of WebObjects if speed isn't an issue. Interpreted script languages are going to be a lot slower than compiled WebObjects apps. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: Carl Edman <cedman@princeton.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 31 Jan 1996 15:38:46 GMT Organization: Princeton University Distribution: world Message-ID: <4eo2e6$c4e@cnn.Princeton.EDU> References: <4eli3t$dl8@news.rwth-aachen.de> In <4eli3t$dl8@news.rwth-aachen.de> Stefan Leuker wrote: > How do I get the full host name programmatically? On a properly configured machine, you get the FQDN this way: #include <netdb.h> { char s[256]; struct hostent *h; gethostname(s,sizeof(s)); h=gethostbyname(s); puts(h->h_name); } Do _NOT_ use getdomainname(). getdomainname() is a YP/NIS-ism and will only return any value at all on YP/NIS machines (which does not include most NeXT machines). Even if it does return a value there is no guarantee whatsoever that the value it returns will have any relationship to the fully qualified host name you are looking for. Please help stamp out that bug in our life times. Carl Edman
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9602010920.AA09327@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Thu, 1 Feb 96 12:20:28 +0300 Subject: Re: First WebObjects question Joe Freeman wrote: > David A. Coyle wrote: > > Now, I want to try my own, a simple example connecting to a miniSQL > > database (just a simple address-book type web-app). > > You may be able to load in a bundle using [[NSBundle xxx] principalClass] > under NEXTSTEP. There may be an issue of whether it was linked against all > of the shlibs. WOF "too-lite" appears to be a script only release for folks > that don't do custom objects. At least one could use Foundation classes. Check out TimeOf/Main.wo/Main.wos. I like this discussion to continue. Is there a way to relink against all of shlibs? Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: wfc@cl.cam.ac.uk (W F Clocksin) Newsgroups: comp.sys.next.programmer Subject: Removing lockFocus'ed Views Date: 1 Feb 1996 12:34:53 GMT Organization: University of Cambridge, England Message-ID: <4eqc1d$shr@lyra.csx.cam.ac.uk> What is the usual cause of getting the console message Feb 1 12:34:20 ...: Assertion failed: You removed a View from the View hierarchy that had been lockFocus'ed I am not aware of removing views from the view hierarchy, although it might be happening behind my back. What might it be? Thanks. William Clocksin Computer Laboratory Cambridge
Newsgroups: comp.sys.next.programmer From: benoit@quest.fdn.org (Benoit Marchant) Subject: W3Kit Message-ID: <1996Feb1.104441.367@quest.fdn.org> Keywords: W3Kit Sender: news@quest.fdn.org Organization: Quest International / Unilever - Neuilly, France Date: Thu, 1 Feb 1996 10:44:41 GMT As everyone talk about WebObjects, does anyone found the W3Kit ? It seems very well done, exists for 2 years, written in Objective-C and comes with the source as it is free ! You can look at it on http://www.geom.umn.edu/software/w3kit/ Benoit Marchant benoit@quest.fdn.org
From: sl@skynet.oph.rwth-aachen.de (Stefan Leuker) Newsgroups: comp.sys.next.programmer Subject: Re: Removing lockFocus'ed Views Date: 1 Feb 1996 13:33:50 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Distribution: world Message-ID: <4eqffu$8l0@news.rwth-aachen.de> References: <4eqc1d$shr@lyra.csx.cam.ac.uk> In article <4eqc1d$shr@lyra.csx.cam.ac.uk> wfc@cl.cam.ac.uk (W F Clocksin) writes: > What is the usual cause of getting the console message > Feb 1 12:34:20 ...: Assertion failed: You removed a View from the View > hierarchy that had been lockFocus'ed I think the usual situation is you are removing a view in IB.app. This doen't mean anything, actually. If you do that programmatically, you remove e.g. a button while it is pressed. That happened once in my NetTalk.app and was a lot of work to work around. sl -- Reply to sl@skynet.oph.rwth-aachen.de NeXTmail/MIME welcome. ----------------------------------------------------------------- Mia: Is that a fact? Vince: No, it's just what I heard; just what I heard. (pf)
From: Carl Edman <cedman@princeton.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 1 Feb 1996 00:27:53 GMT Organization: Princeton University Distribution: world Message-ID: <4ep1e9$mu5@cnn.Princeton.EDU> References: <4eno6f$m1j@lion.embl-heidelberg.de> <4eocln$ea6@news.rwth-aachen.de> In <4eocln$ea6@news.rwth-aachen.de> Stefan Leuker wrote: > In article <4eno6f$m1j@lion.embl-heidelberg.de> > tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) writes: > > > How do I get the full host name programmatically? > > > > gethostname() + getdomainname(); > > getdomainname() returns an empty string. > What is wrong with my config? Nothing. It will do the same for most NeXTstep machines. It is getdomainname() which is wrong for this purpose. Carl Edman
From: Robert Mueller <rmueller@goofy.zdv.Uni-Mainz.DE> Newsgroups: comp.sys.next.programmer Subject: Interfacing SoftPC programmatically Date: Thu, 1 Feb 1996 16:36:57 +0100 Organization: Johannes Gutenberg-Universität Mainz, Germany Message-ID: <Pine.OSF.3.91.960201163520.577C-100000@goofy.zdv.Uni-Mainz.DE> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII I have a question concerning SoftPC: From a AppKit-program, I tried out to start SoftPC and let it execute a list of dos-commands with /LocalApps/SoftPC.app/SoftPC -NXOpen /softpc.hdf -c "some dos commands", and it worked fine. However, is there a possibility to tell a RUNNING SoftPC-application to execute a dos-command (as starting SoftPC takes quite a long time)? Thank you very much in advance ... Robert ___________________________________________________________ Robert Mueller Tel. (+49) (0) 6131 17 68 52 (or 17 20 22) Medical Informatics Fax. (+49) (0) 6131 17 29 68 Institut fuer Medizinische Statistik und Dokumentation Universitaetskliniken Mainz 55101 Mainz, Germany email: rmueller@duchesse.zdv.uni-mainz.de homepage: http://www.Uni-Mainz.DE/FB/Medizin/IMSD/TheMPO/mueller.html
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: Removing lockFocus'ed Views Date: 1 Feb 1996 15:51:38 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4eqnia$rrs@chinx4.thoughtport.net> References: <4eqc1d$shr@lyra.csx.cam.ac.uk> <4eqffu$8l0@news.rwth-aachen.de> #Re: Removing lockFocus'ed Views #by Stefan Leuker, RWTH -Aachen / Rechnerbetrieb Informatik In article <4eqc1d$shr@lyra.csx.cam.ac.uk> wfc@cl.cam.ac.uk (W F Clocksin) writes: > What is the usual cause of getting the console message > Feb 1 12:34:20 ...: Assertion failed: You removed a View from the View > hierarchy that had been lockFocus'ed # I think the usual situation is you are removing a view in IB.app. # This doen't mean anything, actually. Nope-- not specific to IB... but in IB, it doesn't really mean much [though it is definitely a bug]. # If you do that programmatically, you remove e.g. a button while # it is pressed. That happened once in my NetTalk.app and was a # lot of work to work around. It can also happen as a side-affect of other actions. For example, it seems to commonly occur when using NXBrowsers -- if the browser is tracking an event in one of its panes and the processing of that event causes the browser to move its panes around, you'll receive that message. Diagnosing exactly WHY it is ocurring can be difficult Once diagnosed, a fix is not as difficult; you just need to make sure that any code that causes views to shuffle around is NOT being executed while a view is tracking the mouse or actively drawing-- ie; push the 'respond to event by shuffling view hierarchy code' to after processing/drawing associated with the current event is done. -perform:with:afterDelay:cancelPrevious: is an excellent candidate for these kinds of problems. If I remember correctly, invoking -perform:with:afterDelay:cancelPrevious: with a delay of 0 will cause the target method to be invoked at the top of the Main Event Loop (MEL)-- this ensures that it will be invoked after the currently lockFocus'd view is done and unlockFocus'ed. Of course, you still have to deal with ensuring that your UI doesn't appear clunky because of goofy event handling... but that is a case-by-case sort of thing. b.bum
Newsgroups: comp.sys.next.programmer From: jpanico@netcom.com (Joe Panico) Subject: Re: Removing lockFocus'ed Views Message-ID: <jpanicoDM3rEz.GEG@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4eqc1d$shr@lyra.csx.cam.ac.uk> Date: Thu, 1 Feb 1996 15:08:58 GMT Sender: jpanico@netcom.netcom.com W F Clocksin (wfc@cl.cam.ac.uk) wrote: : What is the usual cause of getting the console message : Feb 1 12:34:20 ...: Assertion failed: You removed a View from the View : hierarchy that had been lockFocus'ed : I am not aware of removing views from the view hierarchy, although it might be : happening behind my back. What might it be? Thanks. DPS locks the focus on a view that is in the process of drawing. This includes drawing a button that is in the pressed state. If, from a button press, you call a method that results in swapping out the view that held the button, a la MiscSwapView, you will get the error message you are seeing. The common work-around is to use: - perform:(SEL)aSelector with:anObject afterDelay:(int)ms cancelPrevious:(BOOL)flag; in the call that invokes the view swap. That way, the button will popup and finish drawing, unlocking the focus, before the view is removed. There is a Nextanswer that deals with this problem (I don't know which one). : William Clocksin : Computer Laboratory : Cambridge -- Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc. jpanico@netcom.com /* Please no NeXTMail, I can't read it at this address */
From: Sebastian Niesen <sniesen@imib.rwth-aachen.de> Newsgroups: comp.sys.next.programmer Subject: Re: GNU Fortran on NeXTSTEP ? Date: Thu, 01 Feb 1996 18:28:10 +0100 Organization: NaixT - The NEXTSTEP User Group Aachen Message-ID: <3110F82A.15F9@imib.rwth-aachen.de> References: <1996Jan30.210014.20210@roper.uwyo.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Norbert Pirzkal wrote: > > Has anyone gotten GNU FORTRAN to compile/work under NeXTSTEP 3.3 (intel) ? > Is there a copy of working binaries available for ftp anywhere?? > > Nor Get yourself the source of gcc 2.7.2 and of g77 compile gcc and then g77 using gcc, and then it should work. Sebastian _______________________________________________________________________________ Sebastian Niesen sniesen@pool.informatik.rwth-aachen.de Student Of Computer Science sniesen@imib.rwth-aachen.de RWTH Aachen http://www-users.informatik.rwth-aachen.de/~sniesen
Newsgroups: comp.sys.next.programmer From: beaver@il.us.swissbank.com (Jason Beaver) Subject: Q: Putting images in libraries Message-ID: <1996Feb1.170815.11376@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Thu, 1 Feb 1996 17:08:15 GMT I am trying to put an image in a library so that I don't have to include it in each app which links against the library. I've managed to do this by using ld to put the image in the __TIFF segment for each architecture... ld -r -arch m68k -o libFoo.a.m68k -sectcreate __TIFF MyImage MyImage.tiff i386_m68k_obj/*.o ld -r -arch i386 -o libFoo.a. i386 -sectcreate __TIFF MyImage MyImage.tiff i386_m68k_obj/*.o then using libtool to put them together... libtool -o libFoo.a libFoo.a.m68k libFoo.a.i386 and everything works as expected...apps of either architecture can access the image... However, the resulting library is about 200K smaller than a normal library build using: libtool -o libFoo.a i386_m68k_obj/*.o otool -S claims the library (with the image) has a table of contents, but all the object offsets in the TOC are the same... Does anyone have any idea why the library is smaller? If it doesn't contain a real TOC, will it cause any problems (other than a slower link time)? jason __________________ Jason Beaver Swiss Bank Corp. Jason_Beaver@swissbank.com (NeXTMail/MIME preferred) #import <std/disclaimer.h>
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Removing lockFocus'ed Views Date: 1 Feb 1996 17:27:54 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4eqt6q$duv@news.xmission.com> References: <4eqc1d$shr@lyra.csx.cam.ac.uk> <4eqffu$8l0@news.rwth-aachen.de> sl@skynet.oph.rwth-aachen.de (Stefan Leuker) wrote: > In article <4eqc1d$shr@lyra.csx.cam.ac.uk> wfc@cl.cam.ac.uk (W F Clocksin) > writes: > > What is the usual cause of getting the console message > > > Feb 1 12:34:20 ...: Assertion failed: You removed a View from the View > > hierarchy that had been lockFocus'ed > > [...]. > If you do that programmatically, you remove e.g. a button while it is > pressed. That happened once in my NetTalk.app and was a lot of work to > work around. It is actually very easy; just use the Object method -perform:with:afterDelay:cancelPrevious: to delay the message until the app returns to its event loop (use a delay time of zero). That way the View will be removed after the focus is unlocked. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: michael@rum.cs.wwu.edu (Michael F. DeMan) Newsgroups: comp.sys.next.programmer Subject: Re: GNU Fortran on NeXTSTEP ? Date: 1 Feb 1996 20:16:54 GMT Organization: Western Washington University Distribution: world Message-ID: <4er73m$1g4@ra.cc.wwu.edu> References: <1996Jan30.210014.20210@roper.uwyo.edu> In article <1996Jan30.210014.20210@roper.uwyo.edu> writes: > Has anyone gotten GNU FORTRAN to compile/work under NeXTSTEP 3.3 (intel) ? > Is there a copy of working binaries available for ftp anywhere?? > > Nor Hey, I could use this information as well. We've had a couple projects on the back burner for lack of a cheap (free) fortran compiler. - michael@rum.cs.wwu.edu
Newsgroups: comp.sys.next.programmer From: dan@pointbreak.com (Daniel S. Decasper) Subject: Freeing Enterprise Objects Message-ID: <DM45LD.JnA@eunet.ch> Sender: usenet@eunet.ch Organization: A customer of EUnet AG, Switzerland Date: Thu, 1 Feb 1996 20:15:12 GMT Could somebody please give me a pointer to how I can free Enterprise Objects after fetching them from the database? I have a huge loop which fetches almost 100MB of data out of a SYBASE database in about 40 steps to write the data to different flat files. The steps are not related to each other (from a data point of view), so I would like to free the EO's after each step. I couldn't find anything in the documentation. If I don't free the memory, I'm running into some nasty allocation problem. Thanks, -Dan --- ---------------------------------------------------------------------------- = Dan S. Decasper email: dan@pointbreak.com 09/24/69 = = ch de chandieu 5 voice: 41-21-6010777 NRD# RD272 = = CH-1006 Lausanne Switzerland fax: 41-21-6010774 [NeXTMail ok] = = SDTF@Pointbreak Software & CS-student@ETH -> fight against racism = ----------------------------------------------------------------------------
From: "Benjamin G. Benson" <benson@spiderline.com> Newsgroups: comp.sys.next.programmer Subject: NEXT Database w/ web interface? Date: Thu, 1 Feb 1996 15:52:29 -0500 Organization: Shore.Net/Eco Software, Inc; (info@shore.net) Message-ID: <Pine.SUN.3.91.960201154908.20412D-100000@northshore.shore.net> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII I'm looking to start up a NEXT server on the web and would like to know what is available in the means of web-accessable databases. I've heard there are might be an easy solution to building a web searcheable (through perl?) database on a NEXT machine... is this true? Thanks! ------------------------------------------------------------------------------- | S P I D E R L I N E The Web Source for Web Consultants! | | Consulting Services info@spiderline.com http://www.spiderline.com/ | ------------------------------------------------------------------------------- | Benjamin G. Benson behaas@spiderline.com POBox 466 Newburyport MA 01950 | | President Cellular: 508-837-1612 Page/Voice Mail: 508-859-6143 | | Alpha-Numeric Paging Service: 800-225-0256, #016-9162 | -------------------------------------------------------------------------------
From: dse@dse.ho.att.com (att.general) Newsgroups: comp.sys.next.programmer Subject: Coloring Rows in NXTableViews Date: 1 Feb 1996 22:21:59 GMT Organization: AT&T Bell Laboratories, Columbus, Ohio Message-ID: <4eree7$6qv@nntpb.cb.att.com> Does anyone know of a way to color individual rows in an NXTableView. Thanks, Don
From: tm8025a@mailhost.soc.american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer Subject: Re: A problem with NeXTSTEP 3.3 m68k, sendmail 8.7.3, and POSIX Date: 2 Feb 1996 01:01:37 GMT Organization: American University, Washington DC Distribution: world Message-ID: <4ernph$h9i@paladin.american.edu> References: <4ehg15$ock@news.accessone.com> In article <4ehg15$ock@news.accessone.com> Dennis Glatting <dennis.glatting@plaintalk.bellevue.wa.us> writes: : : :Have any of you come across a problem like this? If so, how :did you solve it? : I used the sendmail package from the ftp sites. Works like a charm. Kudo's to Robert once again. But on a serious note I compiled it once myself and never saw that error. Hmmm....strange. -- Torrey McMahon
Newsgroups: comp.sys.next.programmer From: beaver@il.us.swissbank.com (Jason Beaver) Subject: Re: Freeing Enterprise Objects Message-ID: <1996Feb2.005613.15627@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <DM45LD.JnA@eunet.ch> Date: Fri, 2 Feb 1996 00:56:13 GMT In article <DM45LD.JnA@eunet.ch> dan@pointbreak.com (Daniel S. Decasper) writes: > > Could somebody please give me a pointer to how I can free Enterprise Objects > after fetching them from the database? > > I have a huge loop which fetches almost 100MB of data out of a SYBASE > database in about 40 steps to write the data to different flat files. > The steps are not related to each other (from a data > point of view), so I would like to free the EO's after each step. > > I couldn't find anything in the documentation. If I don't free the memory, > I'm running into some nasty allocation problem. > > Just set up your own autorelease pool for each step... i.e. NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; // Some stuff which allocs LOTS of objects... [pool release]; When you allocate an autorelease pool, it becomes the current pool...Hence, all objects you autorelease in the middle will be placed in your autorelease pool... Since autorelease pools are stacked, you don't have to worry about exceptions generated in the middle (which are caught higher up the stack frame) causing all the objects in your pool to be leaked... jason __________________ Jason Beaver Swiss Bank Corp. Jason_Beaver@swissbank.com (NeXTMail/MIME preferred) #import <std/disclaimer.h>
From: u9113582@muss.cis.McMaster.CA (S. Sharma) Newsgroups: comp.sys.next.programmer Subject: Glibc question Date: 1 Feb 1996 01:33:15 -0500 Organization: McMaster University, Hamilton, Ontario, Canada. Sender: u9113582@muss.mcmaster.ca Distribution: World Message-ID: <4epmrb$k5s@muss.CIS.McMaster.CA> Hello! A simple question really: I have gcc V 2.3.3 with the glibc header files and lib's that came with that t level of the release. Now I am trying to compile things that require functions like strdup() in string.h. The old version of the lib doesn't have this function. I got the glibc.tar.gz but it is unsupported for NeXT running NeXTStep 3.0. I was wondering if there is a site out there where I can find a glibc with port already done? I am on a tight time budget. Or will I just have to compile in the parts of code from the glibc version that I need to get the job done? Any suggestions would be greatly appreciated! Thx!
From: marcel@star.ali.bc.ca (Marcel Sutanto) Newsgroups: comp.sys.next.programmer Subject: How to reserve 2Meg of physically contiguous memory? Date: 2 Feb 1996 03:44:39 GMT Organization: ALI Message-ID: <4es1b7$ar4@cetus.ali.bc.ca> Keywords: PCI driver IOMalloc IOMallocLow I am writing a NeXT driver for a PCI video frame grabber. The board requires that the driver allocates 2 Mbytes of physically contiguous HOST memory for DMA transfer of live video to host memory. Unfortunately, I cannot use IOMalloc () because IOMalloc does not guarantee physically contiguous memory. Using IOMallocLow (2Meg) always failed probably because the driver was asking a large chuck of memory. The driver was never able to allocate 2Meg of contiguous memory. Does anyone know how to reserve 2Meg of physically contiguous memory on NS 3.3? Thanks... -- Marcel Sutanto, marcel@ali.bc.ca -- Marcel Sutanto, Software Manager marcel@ali.bc.ca A.L.I. Technologies Ltd., (NeXT Mail accepted here) 95-10551 Shellbridge Way, Richmond, BC, Canada V6X 2W9 Phone: (604) 279-5422 Fax: (604) 279-5468
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 2 Feb 1996 03:10:04 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4ervac$ebv@news.onramp.net> References: <DM1x45.Arq@isltd.insignia.com> ian.stephenson@insignia.co.uk writes: > talking to Obj-C will cost you $3000, while talking to EOF will cost you $25000. .. > WOF is just not an option for most of the people.. What WOF does (Taking web form output, translating to messages and outputing the results in html) is pretty simple. With the exception of WebScript, it would not be difficult to implement a GNU version. -- Steve Dekorte Objective-C/NEXTSTEP consultant mailto:dekorte@suite.com (NeXTmail, MIME welcome) http://www.marble.com/~dekorte
Newsgroups: comp.sys.next.programmer From: ian.stephenson@insignia.co.uk Subject: Re: Interfacing SoftPC programmatically Message-ID: <DM5A4I.2u5@isltd.insignia.com> Sender: news@isltd.insignia.com Organization: Insignia Solutions plc References: <Pine.OSF.3.91.960201163520.577C-100000@goofy.zdv.Uni-Mainz.DE> Date: Fri, 2 Feb 1996 10:50:42 GMT In article <Pine.OSF.3.91.960201163520.577C-100000@goofy.zdv.Uni-Mainz.DE> Robert Mueller <rmueller@goofy.zdv.Uni-Mainz.DE> writes: > > I have a question concerning SoftPC: > > From a AppKit-program, I tried out to start SoftPC and let it execute a > list of dos-commands > with /LocalApps/SoftPC.app/SoftPC -NXOpen /softpc.hdf -c "some dos > commands", and it worked fine. However, is there a possibility to > tell a RUNNING SoftPC-application to execute a dos-command (as starting > SoftPC takes quite a long time)? There's no trivial way of doing this. However you could try getting data into SoftPC by a couple of routes... Prehaps the easiest would be to set up an fsa drive, as a "spool" type directory, and copy batch files (or exe's) into it. Then write a DOS program (or prehaps a batch file) which loops, looking for files in the directory, executing them when it finds them. You'd have to be carefull that you get "whole" files - mv'ing or ln'ing the files in rather than cp'ing the should deal with this. something like (excuse the mix of C, shell and bat syntax) net use g: $FSADRIVE forever { while !(testforfile g:runme.bat) { sleep 5 } system runme.bat rm runme.bat } on the DOS side. Then to execute a command just ln ToBeExecuted $FSADRIVE/runme.bat from unix. You might like to consider the command EXITSPC from the insignia directory - you can then start everything up with SoftPC -c "EXECUTE.BAT" and shutdown with ln exitspc.bat $FSADRIVE/runme.bat where exitspc.bat just contains the command exitspc. An alternative strategy would be to try using the serial drivers, but thats probably more difficult. Ian
Newsgroups: comp.sys.next.programmer From: jpanico@netcom.com (Joe Panico) Subject: Re: Coloring Rows in NXTableViews Message-ID: <jpanicoDM5Dr4.E0p@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4eree7$6qv@nntpb.cb.att.com> Date: Fri, 2 Feb 1996 12:09:04 GMT Sender: jpanico@netcom16.netcom.com att.general (dse@dse.ho.att.com) wrote: : Does anyone know of a way to color individual rows in an : NXTableView. You could override the: drawFieldAt::inside:inView:withAttributes::usePositions:: of the NXFormatter to handle the drawing of the cell yourself. The formatter gets the row number as an arg, so you could put in some logic to decide which rows to color. Changing the backgroundGrey of Cells in the NXTableView will break the NXTableViews built in selection highlighting method. The built in selection highlighting apparently uses NXHighlightRect() to achieve its effect, and NXHighlightRect() uses the compositerect operator to toggle white to light gray and vice versa. The highlighting only recognizes these values, so if you change the background gray of some cells, the highlighting doesn't work. : Thanks, : Don -- Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc. jpanico@netcom.com /* Please no NeXTMail, I can't read it at this address */
From: rabbit@tygra.Michigan.COM (Roger Rabbit) Newsgroups: comp.sys.next.programmer Subject: How to get The MORTIMER Story Message-ID: <14755@tygra.Michigan.COM> Date: 2 Feb 96 12:04:45 GMT Organization: Shut Up Blair This posting was made automatically by machine. It will appear once every minute or so. Direct any questions to rabbit@Buster.Michigan.COM ------- The MORTIMER story can be retreived from a special archive server (the Mort-Server) To request a part of the MORTIMER story, send mail to one of the addresses below. Put your request on the Subject: line of your message. Requests are can be the in the following form: ALL - (in upper case) will cause the entire story to be sent to you. n - (where n is an integer) will cause part one to be sent n,m... - (where n and m, etc are integers) will cause parts n, m and so forth, to be sent. You may send your request to any of the following addresses: MORTIMER@Buster.Michigan.COM MORTIMER@Babs.Michigan.COM MORTIMER@Plucky.Michigan.COM MORTIMER@Hamton.Michigan.COM MORTIMER@McLoon.Michigan.COM MORTIMER@Elmyra.Michigan.COM If you send off a request and don't get a response within 5 days, send e-mail to the sysop here (NOT ME!!). His address is: ARCHIVER@Michigan.COM -- >>> BAN: Nuclear Power, US Intervention in The Gulf, Toxic Waste, >>> rdc, carasso, Trash Incinerators, Nuclear Weapons, Poverty, KiBoIsM >>> Racism, Kent Paul Dolan, Specieism, etc... Write: Rabbits for a Better >>> Hutch, Roscommon, MI 48653 E-MAIL: rabbit@Buster.Michigan.COM
From: jeremy@tddi.hksys.com (Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Re: How can I use gdb in Emacs? Date: 2 Feb 1996 17:20:55 GMT Organization: University of Nebraska--Lincoln Message-ID: <4eth5o$b9l@crcnis3.unl.edu> References: <4ej7ll$h77@news.mcl.bdm.com> In-Reply-To: <4ej7ll$h77@news.mcl.bdm.com> To: yibwu@napoleon.plato.sky.bdm.com (Yibing Wu) In article <4ej7ll$h77@news.mcl.bdm.com>, yibwu@napoleon.plato.sky.bdm.com (Yibing Wu) writes: >I am using GNU Emacs for NeXT as my development tool. However, the GDB >interface is different, I guess it is because the executables and .m .h >files are not in the same directory, the gdb always give me ".m" file not >found error. Create a .gdbinit file in that directory, and put: dir /path/to/the/source/files in it. -- Jeremy Bettis -*- PGP public key available,send mail with subj "Send pgp key" INET: jbettis@cse.unl.edu UUCP: jeremy@tddi.UUCP NeXTmail: jeremy@thistle.hksys.com NeXTMAIL/MIME/PGP accepted
From: majka@next.com (Marc Majka) Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 2 Feb 1996 18:59:14 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4etmu2$qi7@news.next.com> References: <4ep1e9$mu5@cnn.Princeton.EDU> To be certain that you get the fully qualified domain name of a host, you need to fetch the host name and the domain name, and append the domain name if required. The semantics of gethostname(), gethostbyname(), and gethostbyaddr() don't guarentee that you get a fully qualified name. If depends on how your system administator has things set up. Here's a sample program that looks up the host name and the domain name, and appends them if necessary. I haven't tested this program in all circumstances, so caveat developer :-) -- Marc Majka /* * fqdn.c - get a host's fully qualified domain name * * Copyright 1996, NeXT Software Inc. * You may freely copy, distribute and reuse the code in this example. * NeXT disclaims any warranty of any kind, expressed or implied, as to * its fitness for any particular use. */ #include <stdio.h> #include <sys/types.h> #include <libc.h> #include <arpa/nameser.h> #include <resolv.h> #include <sys/param.h> char *fqdn(void) { static char host[MAXHOSTNAMELEN]; int status, hlen, dlen; /* Get the host name */ status = gethostname(host, MAXHOSTNAMELEN); if (status < 0) return(NULL); hlen = strlen(host); /* res_init() will look up the domain name */ res_init(); /* If there's no domain name set, _res.defdname is zero length */ dlen = strlen(_res.defdname); if (dlen == 0) return(host); /* Append the domain name to the host name if necessary */ if ((hlen < dlen) || strcmp(host + (hlen - dlen), _res.defdname)) { strcat(host, "."); strcat(host, _res.defdname); } return(host); } int main(int argc, char *argv[]) { printf("Fully qualified domain name = %s\n", fqdn()); exit(0); }
Date: 2 Feb 1996 18:58:53 GMT From: rabbit@tygra.Michigan.COM (Roger Rabbit) Message-ID: <cancel.14755@tygra.Michigan.COM> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <14755@tygra.Michigan.COM> Control: cancel <14755@tygra.Michigan.COM> RABBIT spam cancelled by clewis@ferret.ocunix.on.ca
From: trey@hsv.tybrin.com (Trey McClendon) Newsgroups: comp.sys.next.programmer Subject: Question about Driver Dev. Date: 2 Feb 1996 11:03:12 -0600 Organization: TYBRIN Corporation Message-ID: <4etg4g$cnv@tybrin1.hsv.tybrin.com> I'm trying to write a driver for a FASTCOM/HSCX IO card. This card is rated at up to 4Mbit/sec when externally clocked. My target speed is 2Mbit/sec. This is my first C program in a year and my first driver ever, so I'm a bit slow coming to grips with the whole thing. For example, it took a while just to figure that the write buffer passed to the driver gets automagically converted into a series of io vectors. Anyway, my question involves the write command. After the driver is finished, and I issue a 'write(dev, buf, ..)' to the device, will it wait until the driver completes writing all of buf to the hardware, or will (should) it return control to the user program immediately while the hardware pumps the data out? I'll be using DMA, and the card has a 4k DMA buffer space. I plan on sending 4 or 8k chunks to the driver. The size is small since I have to stuff some time tag information into the data stream. If the write command returns control immediately then I'll have 4 to 8 ms to prepare another buffer. If not... eek. It may be that I should punt doing it in NS and do it under DOS [sorry]. Any info/insight/wisdom on this is appreciated. --Trey -- Trey McClendon TYBRIN Corporation trey@hsv.tybrin.com Huntsville, AL NeXT / MIME Mail Accepted Fax: 205-837-3472
From: friendly@hotspur.psych.yorku.ca (Michael Friendly) Newsgroups: comp.sys.next.programmer Subject: unistd.h needed to build Movie.app Date: 2 Feb 1996 22:40:46 GMT Organization: York University, Ontario, Canada Message-ID: <4eu3te$mjf@sunburst.ccs.yorku.ca> |Movie 3.0 has support for some Quicktime types ('.mov', it uses xanim as a backend) as |far as I know. The package should be available on peanuts |(ftp.informatik.uni-muenchen.de) somewhere. I tried to build Movie.app on black NS 3.0 from the source on the Munich ftp site. It failed in xanim with Header file 'unistd.h' not found Could someone send me this header file, or suggest a workaround? [I'm not a C++ sort of guy; I tried sending email to the developer, Windemuth, but it was returned, user unknown.] -- Michael Friendly Internet: friendly@hotspur.psych.yorku.ca (NeXTmail OK) Psychology Department York University Voice: 416 736-5118 4700 Keele Street http://www.math.yorku.ca/SCS/friendly.html Toronto, ONT M3J 1P3 CANADA
From: mathiasp@infotell.isar.de (Mathias Picker) Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 1 Feb 1996 12:41:09 GMT Organization: InfoTell Message-ID: <4eqcd5$1mb@mp.infotell.isar.de> References: <4end31$39k@sun0.urz.uni-heidelberg.de> In <4end31$39k@sun0.urz.uni-heidelberg.de> David A. Coyle wrote: > Well, I've downloaded WOF, installed it and checked that all the examples > work (TimeOff is even cooler than Dodge-Lite, but S_L_O_W). > > Now, I want to try my own, a simple example connecting to a miniSQL > database (just a simple address-book type web-app). > > How? > [snip] > > I'm getting the impression that this free beta version is "WOF-Lite" and > that without WOF-Pro or WOF-Enterprise, what I want to do won't be > possible. > > your Web Manic-depressive, > > Dave > > ------------------------------------------------------------------ > David A. Coyle Father of Annealer.app. > Fission-track research. Ask for my PGP public key. > ------------------------------------------------------------------ > Max-Planck-Institut f r Kernphysik Heidelberg, Germany > ------------------------------------------------------------------ > dcoyle@goanna.mpi-hd.mpg.de dcoyle@weizen.rt.schwaben.de > Right, about what I made out. We probably need WO pro. I first thought the Enterprise edition was it and got really depressed (who can pay that), but it seems that includes tools for legacy application access. Hmmm, interesting idea to try DO, didn*t thing that far. Sad it didn't work out. I have a question, too: Any way to have dynamic components. Oh that doesn't describe it: I want to have something like a content view (the real text of the page), one page wich describes the layout, header, footer, navigational tools, and this should import the real content as (probably) component, depentent on the state of the navigational elements in the main page. I.e instead of making something like /header.wo /footer.wo /page1.wo /page2.wo /pageX.wo I want /page.wo /page.wo/content1.wo /page.wo/content2.wo /page.wo/contentX.wo Or, even better (much better): store all of the content in a database instead of in components. The problem: I want to have WOHyperlinks in them, so they have to be components if I understand the WO design right. -- Mathias Picker mathiasp@infotell.isar.de Hypermedia Phone: +49 89 54070199 Knowledge re/pre sentation Fax: +49 89 54070198
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 3 Feb 1996 02:04:48 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Distribution: world Message-ID: <4eufs0$jnl@news.onramp.net> References: <4eli3t$dl8@news.rwth-aachen.de> In article <4eli3t$dl8@news.rwth-aachen.de> sl@skynet.oph.rwth-aachen.de (Stefan Leuker) writes: > Hi there! > > How do I get the full host name programmatically? Check out the UNIX gethostname(name, namelen) function. Man pages are a good source of info like this - See below. ------------------ > man -k hostname L.aliases (5) - UUCP hostname alias file ethers (5) - Ethernet address to hostname database or YP domain gethostname, sethostname (2) - get/set name of current host hostname (1) - set or print name of current host system hostname (7) - host name resolution description > man gethostname Reformatting page. Wait... done GETHOSTNAME(2) UNIX Programmer's Manual GETHOSTNAME(2) NAME gethostname, sethostname - get/set name of current host SYNOPSIS gethostname(name, namelen) char *name; int namelen; .
Newsgroups: comp.sys.next.programmer From: lorgb@netcom.com (LOR/Geske Bock Associates) Subject: Problem Loading Sybase Adaptor Message-ID: <lorgbDM6K5u.M2s@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Sat, 3 Feb 1996 03:25:06 GMT Sender: lorgb@netcom5.netcom.com Hello. I am getting "Cannot load adaptor: SybaseAdaptor" msg. The adaptor is in place in /NextLibrary/Adaptors/SybaseAdaptor.adaptor. In gdb, I also get the msg: Error loading /NextLibrary/Adaptors/SybaseAdaptor.adaptor/SybaseAdaptor rld(): Undefined symbols: _send _setsockopt _sys_errlist _sys_nerr I am running 3.3 on HP 712, attempting to connect to Sybase 10. Help! Thanks in advance. Joseph Kim LOR/Geske Bock Investments lorgb@netcom.com jkim@la.lorgb.com phone 310.789.2000 x4012 fax 310.789.2010
Newsgroups: comp.sys.next.programmer From: beatty@netcom.com (Derek Lee Beatty) Subject: Re: Help: building gcc-2.7.2 Message-ID: <4eteuf$4l5@beatty.slip.netcom.com> Sender: netnews@mork.netcom.com Organization: none References: <1996Jan30.230611.29244@alw.nih.gov> <4eoipq$e8n@agate.berkeley.edu> Date: Fri, 2 Feb 1996 16:42:55 GMT Yesterday I built gcc-2.7.2 under 3.2 on black hardware. It was very straightforward and I had no problems. This morning I built libg++-2.7.1 using this gcc. I had one problem: I had to comment out the line //extern "C" double hypot (double, double) __attribute__ ((const)); from the file libstdc++/std/complext.h. Not all of "make check" passes due to sillyness with the index() macro, but otherwise so far it seems to work well. -- Derek Lee Beatty _ beatty@netcom.com _| ~-, Austin, Texas \, * ) \_(
From: Carl Edman <cedman@princeton.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to get the complete host name? Date: 2 Feb 1996 22:40:20 GMT Organization: Princeton University Distribution: world Message-ID: <4eu3sk$k07@cnn.Princeton.EDU> References: <4ep1e9$mu5@cnn.Princeton.EDU> <4etmu2$qi7@news.next.com> In <4etmu2$qi7@news.next.com> Marc Majka wrote: > To be certain that you get the fully qualified domain name of a host, you > need to fetch the host name and the domain name, and append the domain > name if required. The semantics of gethostname(), gethostbyname(), and > gethostbyaddr() don't guarentee that you get a fully qualified name. If > depends on how your system administator has things set up. The only case in which this will fail is if the sysadmin has listed a truncated hostname before the fully qualified name in the hosts file (or the netinfo equivalent), but from reading the man pages that is clearly an error. Carl Edman
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: ObjectPak Date: 3 Feb 1996 13:02:25 GMT Organization: University of Ghent, Belgium Message-ID: <4evmd1$jcr@infoserv.rug.ac.be> OBJECTPAK --------- "ObjectPak" is not just "another" collection class library. ObjectPak is a rewrite from scratch of the objects described in Brad Cox' book. ObjectPak is largely compatible in interface with the collection classes of Stepstone's ICpak101 (but not with the ICpak101 foundation classes). The library comes with source code, header files, RTF and HTML documentation. Runtime dependencies are left out of ObjectPak. You can therefore work with the same objects on any current or future implementation of Objective C, yet take advantage of each particular environment : ObjectPak has been tested on NextStep, Stepstone and GNU Objective C, and has been ported to several Unix platforms. Classes in ObjectPak -------------------- ObjPak : root class of ObjectPak ObjStr : string objects ObjCltn : collections of objects, ordered ObjSet : sets of objects, no duplicate entries ObjDic : dictionaries, key-value tables ObjSort : instances keep their contents sorted in a tree ObjSeq : sequencing over groups of objects ObjectPak is an ideal starting-point for writing all sorts of text filters or processors in Objective C : as a demonstration of what can be done, the library comes with a small program to print a collection, a set without duplicates, a dictionary or a sorted list of the words in a text file. ObjectPak is public domain software. Get a copy from "http://www.can.nl/~stes". David Stes stes@can.nl
From: Joe Freeman <Joe@FreemanSoft.COM> Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: Sat, 03 Feb 1996 09:51:49 -0500 Organization: FreemanSoft Inc. Message-ID: <31137685.658C@FreemanSoft.COM> References: <4end31$39k@sun0.urz.uni-heidelberg.de> <4eqcd5$1mb@mp.infotell.isar.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Mathias Picker <mathiasp@infotell.isar.de> I hope I understand your question accurately Mathias Picker wrote: > > Any way to have dynamic components. Oh that doesn't describe it: I want to > have something like a content view (the real text of the page), one page > wich describes the layout, header, footer, navigational tools, and this > should import the real content as (probably) component, depentent on the > state of the navigational elements in the main page. > You can use a WOConditional construct to hold the reference to the component you want. The conditional will only include the component if the value of a variable in your script is set to something other than nil. (Nothing like breaking the "don't use nil" rule in OPENSTEMP eh?) So, it looks something like this. foo.html -- <WEBOBJECT name=some_conditional> <WEBOBJECT name=some_component> </WEBOBJECT> </WEBOBJECT> foo.wod -- some_conditional = WOConditional ... -- some_component = YourComponent {} foo.wos -- id some_flag Component.html -- whatever you want Component.wod -- whatever you want Component.wos -- whatever you want This isn't exact since it is off the top of my head but I make heavy use of conditional code in an application I built under the alpha. As an aside I have a long list of pointers, hints and issues that I came up with based on the alpha but I haven't been able to port my app to the beta yet to update the list. <joe freeman> FreemanSoft.com in MD -- The governer of MD wants to spend $300M on a rent free pro football stadium which is only used 10 times a year. For whom is he really doing this deal?
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: Help: building gcc-2.7.2 Date: 3 Feb 1996 21:32:41 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4f0k9p$d0c@agate.berkeley.edu> References: <1996Jan30.230611.29244@alw.nih.gov> <4eoipq$e8n@agate.berkeley.edu> <4eteuf$4l5@beatty.slip.netcom.com> In article <4eteuf$4l5@beatty.slip.netcom.com>, Derek Lee Beatty <beatty@netcom.com> wrote: >Yesterday I built gcc-2.7.2 under 3.2 on black hardware. It was very >straightforward and I had no problems. This morning I built libg++-2.7.1 >using this gcc. I had one problem: I had to comment out the line > //extern "C" double hypot (double, double) __attribute__ ((const)); >from the file libstdc++/std/complext.h. This is because hypot is defined in the standard library math.h, probably. It creates other silly problems, too. Paul
From: Wei Chen <weic@msn.com> Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 4 Feb 1996 00:38:46 GMT Organization: Internet Online Services Message-ID: <4f0v6m$d0l@news.ios.com> References: <4end31$39k@sun0.urz.uni-heidelberg.de> <4eqcd5$1mb@mp.infotell.isar.de> <31137685.658C@FreemanSoft.COM> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Joe Freeman <Joe@FreemanSoft.COM> wrote: > >foo.html -- <WEBOBJECT name=some_conditional> > <WEBOBJECT name=some_component> > </WEBOBJECT> > </WEBOBJECT> >foo.wod -- some_conditional = WOConditional ... > -- some_component = YourComponent {} >foo.wos -- id some_flag > > >Component.html -- whatever you want >Component.wod -- whatever you want >Component.wos -- whatever you want > What's rational behind NeXT's design to have a separate declaration file? It seems to me quite problematic from a design point of view, because it's very hard to tell what attributes can appear in a WEBOBJECTS clause just from context. For example, an attribute like <INPUT TYPE = "SUBMIT"> should only be allowed to appear in a form. I would like to see the type info is part of WEBOBJECT clause. Something like <WEBOBJECT type=WOFFORM name=A_FORM> ... </WEBOBJECT> Besides, I wish that there would be a USE clause to locate where WEBOBJECT clause is, rather than putting WEBOBJECT clause directly into my html template. Advantages? For one, then I can reuse those clauses. Fro example: ... <USE type=WEBOBJECT location=my_form_dir/this_form>... </USE> sometime later,... <USE type=WEBOBJECT location=my_form_dir/this_form>... </USE> so I use the same form object twice!! I think that NeXT is going in a right direction, but to make an abstract model for HTML really elegant may take some more time.
From: nwc@wsc.com (Nick Christopher) Newsgroups: comp.sys.next.programmer Subject: Anybody done any ODBC stuff on NEXTSTEP? Date: 2 Feb 1996 20:02:11 GMT Organization: WSC Investment Services, Inc. Message-ID: <4etqk3$1m8@cerberus.wsc.com> I'd assume it would be level one but thats ok....just need to know if even that can be done and what was used. \n
From: tlm@ameslab.gov(Dr. T. L. Marchioro II) Newsgroups: comp.sys.next.programmer Subject: Use of NXPerformService Date: 4 Feb 1996 05:35:23 GMT Organization: Iowa State University, Ames, Iowa Message-ID: <4f1gir$2cn@news.iastate.edu> After so many years of owning and actively using a NeXT I'm finally writing my own Obj-C application... eventually hope to announce a nice little program that makes some of the fancy WEB software I work on in my "day job" work with NeXTstep. Anyway, the as with most program I write, it seems to work perfectly but for the last 5%. The program needs to load some information on the Pasteboard and then perform utilize a service already available under NeXTstep. I'm successfully loading the information onto the pasteboard, but the function documented in NeXTDeveloper references, NXPerformService does not seem to be working. Has anyone out there used this function? And if so, can you give an example? A 5 line program that causes "cat" to be piped to digital Webster would be enough I think. Thanks in advance --- Tom PS: Can anyone point me to documentation on Terminal Services over and above the rather scanty stuff in the release notes? -- Dr. Thomas L. Marchioro II Two-wheeled theoretical physicist Applied Mathematical Sciences 515-294-9779 Ames Laboratory 515-432-9142 (home) Ames, Iowa 50011 tlm@ameslab.gov Project Coordinator: Undergraduate Computational Engineering and Sciences http://uces.ameslab.gov/
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Sorting an array and other stuff: Doing it the right way? Date: 4 Feb 1996 11:15:26 GMT Organization: University of Ghent, Belgium Message-ID: <4f24ge$nlo@infoserv.rug.ac.be> References: <DLy7A1.7H@gurke.ping.de> Constantin Szallies (tini@gurke.ping.de) wrote: : I recently wondered how to support sorting an array object. I saw two : solutions. There first one is used by the MiscSortedList class which is : included in the MiscKit distribution. There second solution is used by : NSArray and NSMutableArray in the foundation kit. First I will present both : solutions, then I will show you how I would do it. It might also be feasible to temporarily convert the array of objects into another structure, for sorting, and then, when they are sorted, convert them back into an array.
From: Joe Freeman <Joe@FreemanSoft.COM> Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: Sun, 04 Feb 1996 08:56:59 -0500 Organization: FreemanSoft Inc. Message-ID: <3114BB2B.3D04@FreemanSoft.COM> References: <4end31$39k@sun0.urz.uni-heidelberg.de> <4eqcd5$1mb@mp.infotell.isar.de> <31137685.658C@FreemanSoft.COM> <4f0v6m$d0l@news.ios.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Wei Chen wrote: > > What's rational behind NeXT's design to have a separate declaration file? > It seems to me quite problematic from a design point of view, because > it's very hard to tell what attributes can appear in a WEBOBJECTS clause > just from context. This was an attempt to split out the user interface description from the binding of data objects to the interface. One of the problems with nib files is that all of the control, data binding and user interface is in one file. In theory, if the binding file was done correctly, you could put a meta description of the data object binding into the wod file and you could bind different user interfaces to it. The removal of WebObjects code from the HTML , in theory, would make it easier to use standard HTML editors to do layout. The last point doesn't work out quite as well as one would hope since much of the HTML is replaced with <WEBOBJECT> constructs anyway. > For example, an attribute like <INPUT TYPE = "SUBMIT"> should only be > allowed to appear in a form. Actually, I think you can put any parameter you want in the HTML and it will get passed to the the wod file and then onto the Web Object that represents the HTML block. > I would like to see the type info is part of WEBOBJECT clause. Something > like > > <WEBOBJECT type=WOFFORM name=A_FORM> ... </WEBOBJECT> Do you really want to specify the class to be used on the backend from inside the HTML file. There is a partitioning issue here where people are going to take different positions. Actually, a builder would fix this problem for most people since it would read in this stuff and present it in a user interface. There is no excuse for the beta coming out without a builder application. > Besides, I wish that there would be a USE clause to locate where WEBOBJECT > clause is, rather than putting WEBOBJECT clause directly into my html > template. Advantages? For one, then I can reuse those clauses. Fro > example: > > ... <USE type=WEBOBJECT location=my_form_dir/this_form>... </USE> > sometime later,... <USE type=WEBOBJECT location=my_form_dir/this_form>... > </USE> so I use the same form object twice!! You can do this in the wod file by using "reusable" .wo bundles. > I think that NeXT is going in a right direction, but to make an abstract > model for HTML really elegant may take some more time. Agreed. But, I think they big win of stuff from NeXT is that they do a good job at abstraction and paradigm mapping. -- The governer of MD wants to spend $300M on a rent free pro football stadium which is only used 10 times a year. For whom is he really doing this deal?
From: louie@va.pubnix.com (Louis A. Mamakos) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 4 Feb 1996 09:49:18 -0500 Organization: Pubnix Access Systems (Virginia) Message-ID: <4f2h1e$jo9@pub02.va.pubnix.com> References: <4e84c1$m0n@axl02it.ntc.nokia.com> Keywords: NeXT, Plan9 In article <4e84c1$m0n@axl02it.ntc.nokia.com>, Alan Barclay <escribe@dial.pipex.com> wrote: >Infact, this would not work immediately because of the fact that NeXTs >4.3BSD is really MACH. However there is a version of 4.4BSD for MACH called >4.4BSD LITES (produced at Helsink Uni) - (Slightly strange coincidence is that >Linux comes out of Helsink too!). Perhaps these sources offer the possibility >of running native OPENSTEP libraries and binaries on an upto date UNIX (for >those people who really want OPENSTEP on UNIX rather than on NT/W95). > >I accept that there are hurdles to overcome, not least of which might be the >fact that LITES is based on a version of MACH later than that which NeXT >used or that new code to support the NeXT MAB executable file formats >would be required. I am interested in your comments on this. Again, I'd reiterate that I'd love to run a modern OS on my black next hardware. There is no requirement, for me at least, that it run existing NeXTSTEP applications, or host OPENSTEP. All I'm looking for is a stock 4.4BSD based OS, like NetBSD. louie
From: stefanos@Vir.com (Stefanos Kiakas) Newsgroups: comp.sys.next.programmer Subject: Help with the indexing kit. Date: 4 Feb 1996 09:33:46 -0500 Organization: Communications Vir, Internet Access Montreal. Message-ID: <4f2g4a$ako@Vir.com> Hello, I'm using the indexing kit to maintain some data files. I'm trying to use IXAttributeQuery to search for objects in an IXRecordManager. I'm trying to find the objects with an attribute year for which year is greater or equal to some value. Here is the query string I'm passing to the method initQueryString:andAttributeParser: and the initialization of an instance of IXAttributeQuery. char *queryStr = "ge( year 1993 )" foo = [[IXAttributeQuery alloc] initQueryString: queryStr andAttributeParser: nil ]; When the above method returns foo is nil, which means there was an error in compiling the query. The attribute year is of type int and has been defined in the IXRecordManager instance object as an attribute, as follows: [adRecMan addAttributeNamed:"year" forSelector:@selector(year)]; [adRecMan setComparator:&IXCompareLong andContext:NULL forAttributeNamed:"year" ]; I read the documentation, but I can't figure out what I'm doing wrong. Any help would be appreciated. stef
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 4 Feb 1996 17:24:54 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4f2q56$n9c@emerald.oz.net> References: <4end31$39k@sun0.urz.uni-heidelberg.de> dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) wrote: > Now, I want to try my own, a simple example connecting to a miniSQL > database (just a simple address-book type web-app). > > How? > > There's nothing at all in the WOF documentation on how to do this. > ( see http://www.next.com/Pubs/Documents/WebObjects/ ) > Though the docs are not complete, there's not even a hint of _how_to_ > connect to an external datasource... > > There's no typing, and no (documented) way to import headers: how can I > connect to a custom object (even one that only takes & returns NSArrays , > NSDicts & so on...)? > I tried using DO, but that didn't work: > NXConnection: does not recognize selector +methodSignatureForSelector: > > I'm getting the impression that this free beta version is "WOF-Lite" and > that without WOF-Pro or WOF-Enterprise, what I want to do won't be > possible. I have no experience with WebObjects, but at Webmania, NeXT stated that the free version of WebObjects doesn't support the use of custom objects, DO, or the use of anything compiled. It supports the use of WebScript only, but Foundation is included so you should be able to create NSArrays, NSDicts, etc. using WebScript. Apparently Perl can be used with WebObjects and a series of Perl scripts called sybperl can be used to access Sybase data. So I assume that something similar could be done to access MiniSQL. A couple of things to keep in mind: WebObjects is beta software and it costs nothing, not even the final release. So there's probably a lot of functionality there for free. However, NeXT has to be profitable to survive, so it must make money somewhere. WebObjects Pro supports the use of compiled custom objects, probably DO, etc. - it's more like what we're all accustomed to in building a NS app. Whether $2,500 is an appropriate licensing fee for WebObjects Pro, I'm not sure. I presume most organizations would need only a single license, so the volumes won't be nearly as large as for popular shrink-wrap apps. It doesn't take long to pay a Web programmer $2,500 in compensation, so if WebObjects Pro really does reduce the Web page design time significantly and produce a more dynamic Web site, its licensing fee is probably reasonable. However, I have a real problem with the $25,000 licensing fee for WebObjects Enterprise. All it appears to add to WebObjects Pro is EOF - i.e., powerful database access. This licensing fee seems so high that it will limit the use of WebObjects Enterprise to only large companies. We have customers who could use WebObjects Enterprise to access their data using our existing EOF code, but they'll never pay $25,000 for the privilege :-( I suggest that NeXT tie WebObjects Enterprise licensing fees to Web server activity levels. The low end license would limit a site to X hits per day. If that limit is reached frequently, the customer could license the next level which would increase the maximum number of hits per day. This might allow smaller sites to take advantage of WebObjects Enterprise at an affordable price. Then once they see how well it works, if they need a higher license level, they'd be more likely to pay for it. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: David M Holscher <holscher@mixi.net> Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: Sun, 04 Feb 1996 14:13:20 -0500 Organization: Midwest Internet Exchange, Inc. Message-ID: <31150550.58C@mixi.net> References: <DM1x45.Arq@isltd.insignia.com> <4ervac$ebv@news.onramp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Steve Dekorte wrote: > > ian.stephenson@insignia.co.uk writes: > > talking to Obj-C will cost you $3000, while talking to EOF will cost you $25000. > .. > > WOF is just not an option for most of the people.. > > What WOF does (Taking web form output, translating to messages > and outputing the results in html) is pretty simple. > > With the exception of WebScript, it would not be difficult to implement a GNU version. Agreed. Actually I am working on an internal Web based project for my company. Orignally we did everything in CGI and Perl. These were great for whipping up simple form interfaces, but are impossible to maintain and scale horribly. We started using the foundation classes with a few additions (regular expression objects and a stream based object) to write our CGI programs. It didn't take long for the idea of an object oriented server to begin forming so I wrote one. It was easy to do using Foundation and some HTTP classes to represent HTTP messages and URIs. The server is very configurable and works by mapping a URI to a call to a certain class. Classes are loaded dynamically as bundles and each request gets its own thread. Except for the WOS interpreter I'm sure we could add a class to emulate Web objects in a few weeks. The next major evolutionary step in our server was going to be a seperation of data bearing objects from interface objects (which is always a good design but in the real world we have time constraints). Note that our Web server will be written to compile under all OpenStep implementations. I doubt our server will ever be released but I would be happy to help work on a free version. -- David Holscher CAE Systems Architect, OpenStep Developer Navistar International
From: Mark Crispin <mrc@CAC.Washington.EDU> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: Sun, 4 Feb 1996 13:00:51 -0800 Organization: Networks & Distributed Computing Message-ID: <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII NNTP-Posting-User: dma In-Reply-To: <4f2h1e$jo9@pub02.va.pubnix.com> On 4 Feb 1996, Louis A. Mamakos wrote: > Again, I'd reiterate that I'd love to run a modern OS on my black next > hardware. There is no requirement, for me at least, that it run > existing NeXTSTEP applications, or host OPENSTEP. All I'm looking for > is a stock 4.4BSD based OS, like NetBSD. I second this. It'd be awfully nice if there was /dev/audio support to make the DSP do its thing, but it isn't really necessary. NeXT's software offerings are completely uninteresting. It would also open up the possibility of running real Web browsers, instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does a better job on the Web than an "insanely great" NEXTSTEP system, that says something very sad. -- Mark -- DoD #0105, R90/6 pilot, FAX: (206) 685-4045 ICBM: N 47 39'35" W 122 18'39" Science does not emerge from voting, party politics, or public debate.
From: cfleming@alleg.edu (Chuck Fleming) Newsgroups: comp.sys.next.programmer Subject: Re: Use of NXPerformService Date: 5 Feb 1996 01:52:39 GMT Organization: Best Internet Communications Message-ID: <4f3nt7$pmr@shellx.best.com> On 02/03/96, Dr. T. L. Marchioro II wrote: > >NXPerformService >does not seem to be working. Has anyone out there used this function? >And if so, can you give an example? > Assuming you are using InterfaceBuilder, did you drag a Services menu into your application's menu? If so, the following should work. - retrieveWebPage:sender { Pasteboard *privatePasteboard = [Pasteboard newName:"URLPboard"]; const char *types[] = {NXAsciiPboardType, NULL}; [privatePasteboard declareTypes:types num:1 owner:nil]; [privatePasteboard writeType:NXAsciiPboardType data:"http://www.next.com/" length:20]; // NXPerformService("Netsurfer/Open URL", privatePasteboard); // or // NXPermformService("OmniWeb/Open URL", privatePasteboard); return self; }
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 5 Feb 1996 05:15:09 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4f43ot$t3j@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: black@winternet.com (Ben Black) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 5 Feb 1996 06:46:10 GMT Organization: StarNet Communications, Inc Distribution: world Message-ID: <4f493i$t2n@blackice.winternet.com> References: <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> In article <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> Mark Crispin <mrc@CAC.Washington.EDU> writes: > On 4 Feb 1996, Louis A. Mamakos wrote: > > Again, I'd reiterate that I'd love to run a modern OS on my black next > > hardware. There is no requirement, for me at least, that it run > > existing NeXTSTEP applications, or host OPENSTEP. All I'm looking for > > is a stock 4.4BSD based OS, like NetBSD. > > I second this. It'd be awfully nice if there was /dev/audio support to > make the DSP do its thing, but it isn't really necessary. NeXT's software > offerings are completely uninteresting. > > It would also open up the possibility of running real Web browsers, > instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does > a better job on the Web than an "insanely great" NEXTSTEP system, that > says something very sad. > Trash like OmniWeb? Have you actually used the 2.0 beta, or are you speaking about 1.0? I am perfectly happy using OmniWeb instead of the Netscape trash I have to use at work. I'd like to see Java support, but your precious Powerbook won't allow that even with Netscape. Ben black@winternet.com
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: WOTextField: doesn't like NSDictionary? Date: 5 Feb 1996 12:35:57 GMT Organization: University of Heidelberg, Germany Message-ID: <4f4tjd$l7t@sun0.urz.uni-heidelberg.de> I want to load a series of values into WOTextFields from an NSDictionary. The values display in anything _other_than_ a WOTextField from the .wod: FAXFIELD:WOTextField { value = currentRecord.fax; WIDTH = 40; }; // value *does not* display // but if I change it to WOString, it does! PGPFIELD:WOText { value = currentRecord.publicKey; rows="10"; cols="67"; // make the key block look pretty }; // value *does* display // if I change it to WOTextField it *doesn't* Is this aberrant behaviour? (ie: a bug) Dave dcoyle@goanna.mpi-hd.mpg.de
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Functionality of Perl's CGI.pm in WebObjects? Date: 5 Feb 1996 17:50:03 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4f5g0b$199@stc06.ctd.ornl.gov> I have written a Web application that provides part of our company with a form fill out, signature, and routing capability. It allows one to archive a filled out form and then notify the next "signor" that it is ready of their signature. I've written it in Perl using CGI.pm, but wanted to convert it to WebObjects. How can I save (write the name/variable pairs) to disk from WebObjects? or read them in later. Is there some WebObject construct that would provide the equivalent? If I can't use WebObjects to do something that is quite easy in CGI.pm, then it appears to be limited in its present form! Does anyone know where I could find a comparison of CGI.pm with WebObjects? What do I get for $2500. -- - - - - - - - - - J. W. Wooten
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: US - CA - NeXT Architect - Develop New OO Systems - to 90k + Bonus Date: 5 Feb 1996 13:39:14 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4f5isi$h76@newsbf02.news.aol.com> California based position My client a well established Fortune 500 company, is currently looking for a senior developer/architect for a position in their top development group. Candidates must have a strong NeXT development background, with a succesful track record of developing large scale OO systems in NeXTSTEP. The succesful applicant will receive an excellent starting salary, benefits, and a '96 bonus. Contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: Re: Functionality of Perl's CGI.pm in WebObjects? Date: 5 Feb 1996 19:18:17 GMT Organization: University of Illinois at Urbana Message-ID: <4f5l5p$jic@vixen.cso.uiuc.edu> References: <4f5g0b$199@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. Wooten) writes: >How can I save (write the name/variable pairs) to disk from WebObjects? or read >them in later. Is there some WebObject construct that would provide the >equivalent? > I don't believe there's a need to save the parameters to disk as WO will maintain it's state across pages. If you believe you must save the parmeters to disk it's not hard to mimic the save and load functions of CGI.pm. Take a look at the source for pointers. -- Ian P. Cardenas (icardena@uiuc.edu) CCSO NeXT System Administrator, CCSO Sites Technical Support "Are they as successful as who,Microsoft? Only drug lords from South America are as successful as Microsoft." -Tim Byars, on the success of NeXT Computer, Inc.
From: mikecam@unixg.ubc.ca (Michael C. Cam) Newsgroups: comp.sys.next.programmer Subject: Where is fmod() Date: 5 Feb 1996 19:14:06 GMT Organization: The University of British Columbia Message-ID: <4f5ktu$sdo@nntp.ucs.ubc.ca> Reply-To: mikecam@unixg.ubc.ca (Michael C. Cam) Distribution: world Hi, Does anyone know if fmod() is part of the math library in the Next release of the GNU C compiler? I have been trying to dynamically load some bundles that utilize it and it comes up with an error stating it is an undefined symbol. It appears as its own man page but it is absent from the math library man page. It is, however, defined in the header file math.h inside ansi/ Any ideas? ..Mike. -- ___________________________________________________________________ | | | ___ ^ ... /\ BEAUTIFUL | | _|_::| ___o '|`^ .. o_ . .. /\ / \ BRITISH | | |:::|:| \ \, ^ '|`|` (`_|/____') / / /\ COLUMBIA | | |:::|:| (o)/ (o) '|`'|`|`` ,,/ . ... . .. / \ |
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Re: Functionality of Perl's CGI.pm in WebObjects? Date: 5 Feb 1996 20:43:35 GMT Organization: University of Heidelberg, Germany Message-ID: <4f5q5n$2o8@sun0.urz.uni-heidelberg.de> References: <4f5l5p$jic@vixen.cso.uiuc.edu> In article <4f5l5p$jic@vixen.cso.uiuc.edu> icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) writes: > woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > >How can I save (write the name/variable pairs) to disk from WebObjects? or read > >them in later. Is there some WebObject construct that would provide the > >equivalent? > > > I don't believe there's a need to save the parameters to disk as WO will > maintain it's state across pages. If you believe you must save the parmeters > to disk it's not hard to mimic the save and load functions of CGI.pm. Take > a look at the source for pointers. Why not write NSDictionaries or NSArrays of NSStrings to disk? See the EOF example FlatFileDatabase. Then one can also write an EOF app to manage the data, if one wishes. Dave
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Ivar link to view in ScrollView Date: 5 Feb 1996 20:49:34 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4f5qgu$o7j@news.onramp.net> I'd like to set an ivar of a view within IB the usual way - using a link to some other object. But I found that there is a problem doing this for a view within a scrollview. Are there any tricks around this? (Of course, I write some extra code to set the ivar, but that's ugly - I want to do this in IB) -- Steve Dekorte Objective-C/NEXTSTEP consultant mailto:dekorte@suite.com (NeXTmail, MIME welcome) http://www.marble.com/~dekorte
From: weic@msn.com (Wei Chen) Subject: Re: First WebObjects question Date: 5 Feb 96 23:40:44 -0800 References: <4end31$39k@sun0.urz.uni-heidelberg.de> <4eqcd5$1mb@mp.infotell.isar.de> <31137685.658C@FreemanSoft.COM> <4f0v6m$d0l@news.ios.com> <3114BB2B.3D04@FreemanSoft.COM> Message-ID: <00001c42+00000a77@msn.com> Newsgroups: comp.sys.next.programmer Organization: The Microsoft Network (msn.com) Joe Freeman wrote: >I wrote: >> Besides, I wish that there would be a USE clause to locate where WEBOBJECT >> clause is, rather than putting WEBOBJECT clause directly into my html > >template. Advantages? For one, then I can reuse those clauses. Fro >> example: >> >> ... <USE type=WEBOBJECT location=my_form_dir/this_form>... </USE> >> sometime later,... <USE type=WEBOBJECT location=my_form_dir/this_form>... >> </USE> so I use the same form object twice!! >You can do this in the wod file by using "reusable" .wo bundles. I am completely against this paradigm. Why should there be a rule that dictates which components can be reusable or not!? Not to mention a reusable component is determined by a component name, and that webobject clause has to be repeated. Whatever I write should be reused by other people without any need of revision of its structure. Sun, I think, has made a big mistake in that Java applet has to be revised into a Java application.
From: weic@msn.com (Wei Chen) Subject: RE: First WebObjects question Date: 5 Feb 96 23:48:17 -0800 References: <4end31$39k@sun0.urz.uni-heidelberg.de> Message-ID: <00001c42+00000a78@msn.com> Newsgroups: comp.sys.next.programmer Organization: The Microsoft Network (msn.com) Does anybody know the url format to launch a WebScript program, in general? Right now I have to put all my programs under /NextLibrary/WebObjects/Examples, then use http://localhost/cgi-bin/WebObjects/Examples/myApp. Thanks.
From: gcasa@wam.umd.edu (Gregory John Casamento) Newsgroups: comp.sys.next.programmer Subject: Problem w/ Mouse-X X11R4 Server Date: 6 Feb 1996 02:53:19 GMT Organization: University of Maryland College Park Message-ID: <4f6fqv$hl8@cville-srv.wam.umd.edu> Hi, I am having trouble w/ Mouse-X. Basically the problem is that after I switch back to NeXTSTEP, using the Command-Command-Delete sequence, when I go back into X the xterm window will not accept anything and the mouse doesn't seem to work. Does anybody know what is going on here?? Thanks, -- Gregory John Casamento -- gcasa@wam.umd.edu (c) G. Casamento -- Permission to distribute on MS network denied!!
From: caljouwr@cadvision.com Newsgroups: comp.sys.next.programmer Subject: NXTableView Date: 6 Feb 1996 03:54:32 GMT Organization: CADVision Message-ID: <4f6jdo$sre@huey.cadvision.com> Does anyone know how to get a value from a NXTableView? I have set up the setActions and selectors and can get the row, column and dataSource but am unable to retrieve the value in the selected row and column. I think the difficulty lies in declaring an object to be a NXTableValue for the into parameter but I am not sure. Any help would be greatly appreciated! Best regards, Rob
From: otto@olcs.com (Otto Lind) Newsgroups: comp.sys.next.programmer Subject: Status of Java port to NeXT? Date: 6 Feb 1996 05:43:44 GMT Organization: Otto Lind Consulting Services Message-ID: <4f6pqg$pd1@olcs.olcs.com> Hi, I know there is a mailing list, but I don't want to sign up on (yet another) list just to find out what the status is. If anyone sees this post and is a member of the list (or on the porting team), a brief summary posted as a response would be appreciated. Thanks is advance, Otto -- Otto Lind Otto Lind Consulting Services otto@olcs.com 4890 Ashley Lane #311, Inver Grove Hts, MN 55077 skypoint!olcs!otto voice:(612)457-1080 fax:(612)457-0761
From: cph@next.com (Charles d'Harcourt) Newsgroups: comp.sys.next.programmer Subject: Re: WOTextField: doesn't like NSDictionary? Date: 6 Feb 1996 10:49:41 GMT Organization: NeXT Software, Inc. Message-ID: <4f7bo5$jo1@news.next.com> References: <4f4tjd$l7t@sun0.urz.uni-heidelberg.de> On 5 Feb 1996, dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) wrote: > I want to load a series of values into WOTextFields from an > NSDictionary. > > The values display in anything _other_than_ a WOTextField > > from the .wod: > > FAXFIELD:WOTextField { > value = currentRecord.fax; > WIDTH = 40; > }; // value *does not* display > // but if I change it to WOString, it does! > > PGPFIELD:WOText { > value = currentRecord.publicKey; > rows="10"; > cols="67"; // make the key block look pretty > }; // value *does* display > // if I change it to WOTextField it *doesn't* > > Is this aberrant behaviour? (ie: a bug) I have tested using a WOTextField as you described and it worked OK for me. However, when I neglected to put a <FORM>...</FORM> around the HTML where the WOTextField WebObject was, I did observe the "missing value" you describe, not only for WOTextField, but also for WOText. Perhaps you omited to put your WOTextField in a form? Since WOTextField (and WOText) generate <INPUT ...> HTML elements, they need to be in a form to function properly. WOString, on the other hand, functions just as well in as out of a form. I have entered a suggestion that an error be generated if WOTextField (and WOText) WebObjects are encountered outside of a form in HTML. Here's the app I tested (as "http:/cgi-bin/WebObjects/Tests/TextField"): File <DOCROOT>/WebObjects/Tests/TextField/Main.wo/Main.html: <HTML> <HEAD> <TITLE>TextField Tester</TITLE> </HEAD> <BODY> <FORM> FAXFIELD:<WEBOBJECT NAME = "FAXFIELD"></WEBOBJECT><BR> PGPFIELD:<WEBOBJECT NAME = "PGPFIELD"></WEBOBJECT><BR> </FORM> </BODY> </HTML> File <DOCROOT>/WebObjects/Tests/TextField/Main.wo/Main.wod: FAXFIELD:WOTextField { value = currentRecord.fax; WIDTH = 40; }; PGPFIELD:WOText { value = currentRecord.publicKey; ROWS = "10"; COLS = "67"; }; File <DOCROOT>/WebObjects/Tests/TextField/Main.wo/Main.wos: id currentRecord; - awake { currentRecord = [NSMutableDictionary dictionary]; [currentRecord setObject:@"FOO" forKey:@"fax"]; [currentRecord setObject:@"BAR" forKey:@"publicKey"]; } I hope this helps, C= ________________________________________________________ Charles d'Harcourt, WebObjects Group, NeXT Software, Inc. mailto:cph@next.com http://www.next.com/~cph Not speaking on behalf of NeXT or anyone in particular
From: fox@jeans.fokus.gmd.de (Oliver Fox) Newsgroups: comp.sys.next.programmer Subject: Q:Examples for insert .sound with EOF ? Date: 6 Feb 1996 16:12:39 GMT Organization: GMD-FOKUS Message-ID: <4f7uln$dbi@stern.fokus.gmd.de> Hi everybody, Iam going to write an application which store all kinds of Data-Files in a Sybase DataBase. Are there any examples to insert sound or ps or mpeg with a NeXT EOF application ? Thank You Oliver Oliver Fox Tel.: +49 30 25 499 123 Fax: +49 30 25 499 202 E-Mail: fox@fokus.gmd.de Hardenbergplatz 2 D-10623 Berlin Germany GMD-FOKUS GMD - German National Research Center for Information Technology
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Re: WOTextField: doesn't like NSDictionary? Date: 6 Feb 1996 15:18:59 GMT Organization: University of Heidelberg, Germany Message-ID: <4f7rh3$qmo@sun0.urz.uni-heidelberg.de> References: <4f7bo5$jo1@news.next.com> In article <4f7bo5$jo1@news.next.com> cph@next.com (Charles d'Harcourt) writes: > On 5 Feb 1996, dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) wrote: > > I want to load a series of values into WOTextFields from an > > NSDictionary. > > I have tested using a WOTextField as you described and it worked OK for > me. However, when I neglected to put a <FORM>...</FORM> around the HTML > where the WOTextField WebObject was, I did observe the "missing value" Ah! How stupid of me! Just goes to show that making WOF work does in fact presuppose a good knowledge of HTML. :-) Now the next question: was it an error for OmniWeb to have displayed the WOTextFields at all??? ;-)) ( I discovered that NetScape didn't, which may have put me on the trail sooner ) thanks, Dave
From: seanl@carmi.cs.umd.edu (Sean Luke) Newsgroups: comp.sys.next.programmer Subject: Re: Removing lockFocus'ed Views Date: 6 Feb 1996 15:59:48 GMT Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742 Message-ID: <4f7ttk$2a2@mimsy.cs.umd.edu> References: <4eqc1d$shr@lyra.csx.cam.ac.uk> <4eqffu$8l0@news.rwth-aachen.de> <4eqt6q$duv@news.xmission.com> don@misckit.com wrote: >> > What is the usual cause of getting the console message >> >> > Feb 1 12:34:20 ...: Assertion failed: You removed a View from the View >> > hierarchy that had been lockFocus'ed [good explanations follow] Oh, YEAH? Well how about this one: When assembling Soundviews in Resound, I've been unable to avoid the following assertion popping up in the Console: Feb 6 10:56:05 carmi Resound[798]: Assertion failed: lockFocus sent to a view which is not in a window Trouble is: I have no views that aren't in windows. My hierarchy is thus: Console Window -> Console Buttons -> MiscVolumeMeter -> TextField Sound Windows (one is already open and another is about to be created after recording a sound) -> ScrollView -> ClipView -> MiscSoundView The MiscSoundView is just an ordinary SoundView with some enhancements. Any ideas? _____________________________________________________________________________ Sean Luke "I've discovered that P==NP, but the proof is too U Maryland at College Park small to fit in the margins of this signature." seanl@cs.umd.edu URL: http://www.cs.umd.edu/~seanl/
From: griffith@crl.com (Dave Griffith) Newsgroups: comp.sys.next.programmer Subject: Re: Removing lockFocus'ed Views Date: 6 Feb 1996 09:49:37 -0800 Organization: The Hermetic and Esoteric Order of Clarity and Openness Message-ID: <4f84bh$kf@crl.crl.com> References: <4eqc1d$shr@lyra.csx.cam.ac.uk> <4eqffu$8l0@news.rwth-aachen.de> <4eqt6q$duv@news.xmission.com> <4f7ttk$2a2@mimsy.cs.umd.edu> In article <4f7ttk$2a2@mimsy.cs.umd.edu>, Sean Luke <seanl@carmi.cs.umd.edu> wrote: > >Oh, YEAH? Well how about this one: > >When assembling Soundviews in Resound, I've been unable to avoid the >following assertion popping up in the Console: > >Feb 6 10:56:05 carmi Resound[798]: Assertion failed: lockFocus sent to a >view which is not in a window > >Trouble is: I have no views that aren't in windows. My hierarchy is thus: I get this problem as well, when unarchiving windows with DBTableViews in them. It looks as though part of the unarchiving process tries to draw the view, even though at that point in the unarchiving process the view is not part of a window. It doesn't seem to have any other effect other than a console message, so I've stopped worrying about it. -- --Dave Griffith, griffith@crl.com
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 5 Feb 1996 09:29:45 GMT Organization: Academic Computing Services, Sheffield University Message-ID: <4f4im9$9jt@bignews.shef.ac.uk> References: <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4e84c1$m0n@axl02it.ntc.nokia.com>?<4f2h1e$jo9@pub02.va.pubnix.com>?0 In-Reply-To: <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> On 02/04/96, Mark Crispin wrote: > It would also open up the possibility of running real Web browsers, > instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does > a better job on the Web than an "insanely great" NEXTSTEP system, that > says something very sad. > Umm, Mark, which version of OmniWeb are you running? IMHO version 2 (albeit in beta) blows away NetScape for elegance, robustness and general ease of use -- the chaps from Lighthouse/Omni even seemed to have frames working properly before NetScape. Best wishes, mmalc. --
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: Where is fmod() Message-ID: <DMCHEs.F6@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4f5ktu$sdo@nntp.ucs.ubc.ca> Date: Tue, 6 Feb 1996 08:11:15 GMT In article <4f5ktu$sdo@nntp.ucs.ubc.ca> mikecam@unixg.ubc.ca (Michael C. Cam) writes: > Does anyone know if fmod() is part of the math library in the > Next release of the GNU C compiler? I have been trying to > dynamically load some bundles that utilize it and it comes up > with an error stating it is an undefined symbol. > Any ideas? I guess that you forgot to use -all_load flag, when working with bundles. You need to force the loader to load all the symbols that your bundle might refer to; this is done by editing your project's Makefile.preamble file (located in Supporting Files in your project) to include this line: OTHER_LDFLAGS = -all_load Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: kennedy@b0ru01.fnal.gov (Robert Daniel Kennedy) Newsgroups: comp.sys.next.programmer Subject: GCC configure memory faults - hardware? Date: 6 Feb 1996 19:55:06 GMT Organization: Fermi National Accelerator Laboratory, Batavia IL, USA Distribution: world Message-ID: <4f8bmq$ep1@fnnews.fnal.gov> Hello, While installing GCC 2.7.2, I came across a problem which is not caused by GCC, and I do not know what is actually causing it. I have already successfully installed GCC 2.7.1 in the past on my m68k-next-nextstep3, running NEXTSTEP 3.3.1. Now, when I execute './configure m68k-next-nextstep3', things roll along until the end of a big 'sh' case-switch when I get a message saying something like: ./configure - memory fault (some #'s) (core dump). I think the core dump is of 'sh', but 'sh' is stripped so the core is useless. I have tried to use 'bash' instead, and to carefully remove any non-standard programs from the PATH seen by the script, and this prevents the core-dumps but produces zero-length Makefiles (ie, configure does not seem to operate quite right under bash). After much thrashing about looking for a cause, I finally tried to rebuild GCC-2.7.1 just to see what was different, since I have already built it in the past... and this memory fault occurs again! So either hardware or software changed since I last built GCC-2.7.1. Since the PATH avoids any /usr/local stuff, only patches to NEXTSTEP could be involved in a software problem... I think. Suggestions as to what may be causing this, or similar experiences would be much appreciated. Thanks, Rob Kennedy kennedy@fnal.gov Rutgers/CDF/Fermilab The Top Quark Folks
From: rsww@quanta.com (Ross Walker) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Followup-To: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Date: 6 Feb 1996 21:10:53 GMT Organization: Quanta Communications, Inc. Distribution: inet Message-ID: <4f8g4t$2dh@news.quanta.com> References: <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4e84c1$m0n@axl02it.ntc.nokia.com>?<4f2h1e$jo9@pub02.va.pubnix.com>?0 <4f4im9$9jt@bignews.shef.ac.uk> mmalcolm crawford (m.crawford@shef.ac.uk) wrote: : On 02/04/96, Mark Crispin wrote: : > It would also open up the possibility of running real Web browsers, : > instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does : > a better job on the Web than an "insanely great" NEXTSTEP system, that : > says something very sad. : > : Umm, Mark, which version of OmniWeb are you running? : IMHO version 2 (albeit in beta) blows away NetScape for elegance, robustness : and general ease of use -- the chaps from Lighthouse/Omni even seemed to : have frames working properly before NetScape. It would just be nice to have 64bit filesystems. But a new NFS, and Kerberos would also be neat (nee-t-o). Most extras are sugar candy. Cheers, Ross Walker
From: Eric Peyton Newsgroups: comp.sys.next.programmer Subject: Question about MABs Date: 7 Feb 1996 02:06:39 GMT Organization: Netcom Message-ID: <4f91ff$qvt@cloner2.ix.netcom.com> Hi Y'all, Got a question. I just installed the libs for NIHS (I just had I installed before), and compiled my program MAB. It is now 10 times larger than the origianl executable, it went from ~500K to ~5M. What happened, will all of my executables be 10 times larger? Is there a way that I can reduce this size?? Any help or pointers would be appreciated. Eric Peyton espeyton@ix.netcom.com
From: wjs@omnigroup.com (William Shipley) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 7 Feb 1996 04:42:59 GMT Organization: Omni Development, Inc. Message-ID: <4f9akj$2co@gaea.titan.org> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> Mark Crispin <mrc@CAC.Washington.EDU> wrote: >>All I'm looking for is a stock 4.4BSD based OS, like NetBSD. >It would also open up the possibility of running real Web browsers, >instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does >a better job on the Web than an "insanely great" NEXTSTEP system, that >says something very sad. I'd find it easier to accept your criticism of OmniWeb if (a) it were more specific in its complaints, (b) you weren't self-admittedly someone who wants to leave NEXTSTEP for BSD, or (c) you didn't have an old vendetta against me and my company dating from back when we were stuck at the UW as well. Calling OmniWeb "trash" gives me nothing I can fix to please you, and thus makes it clear that all you really want is to be heard whining. If you don't like a piece of free software, and feel you must complain, couldn't you at least be specific? Most people love OmniWeb, many customers prefer it to Netscape. OmniWeb 2.0 is the ONLY browser in existence on any platform that implements all the functionality that Netscape 2.0 has (well, besides Netscape), and we have a search panel and bookmarks that no other browser has. If you can find another browser that supports frames, tables, in-line JPEGs, true multi-threading, loadable bundles, side-aligned images, and everything else we do, I invite you to compare it to OmniWeb and let us know how we come out. -Wil
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Removing lockFocus'ed Views Date: 7 Feb 1996 06:19:33 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4f9g9l$lh1@news.xmission.com> References: <4eqc1d$shr@lyra.csx.cam.ac.uk> <4eqffu$8l0@news.rwth-aachen.de> <4eqt6q$duv@news.xmission.com> <4f7ttk$2a2@mimsy.cs.umd.edu> seanl@carmi.cs.umd.edu (Sean Luke) wrote: [...] > When assembling Soundviews in Resound, I've been unable to avoid the > following assertion popping up in the Console: > > Feb 6 10:56:05 carmi Resound[798]: Assertion failed: lockFocus sent to a > view which is not in a window > > Trouble is: I have no views that aren't in windows. My hierarchy is thus: [...] > Any ideas? I think Dave Griffiths is on the right track. However, have you tried setting a GDB breakpoint on -lockFocus and doing a backtrace to see which View is generating the exception? If it is one of your views, you can make it not attempt to draw until after the View is in a Window. If it is a NeXT class, then it is a bug in NS. Also, are your windows deferred? If they are, that can sometimes cause minor console messages, too... -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: rankin@ewl.uky.edu (David W. Rankin Jr.) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 7 Feb 1996 06:29:29 GMT Organization: UK Microlabs -- Engineering Workstation Laboratory Message-ID: <4f9gs9$n0d@t3.mscf.uky.edu> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> contents of this message do not represent the opinions of the University of Kentucky, UK Microlabs, UKMSCF, or UKCC. In article <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU>, Mark Crispin <mrc@CAC.Washington.EDU> wrote: }On 4 Feb 1996, Louis A. Mamakos wrote: }> Again, I'd reiterate that I'd love to run a modern OS on my black next }> hardware. There is no requirement, for me at least, that it run }> existing NeXTSTEP applications, or host OPENSTEP. All I'm looking for }> is a stock 4.4BSD based OS, like NetBSD. }I second this. It'd be awfully nice if there was /dev/audio support to }make the DSP do its thing, but it isn't really necessary. NeXT's software }offerings are completely uninteresting. }It would also open up the possibility of running real Web browsers, }instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does }a better job on the Web than an "insanely great" NEXTSTEP system, that }says something very sad. It's also mean that people with 030 boxen could actually get REAL use out of their hardware, and those of us who like the other M68k architectures (like myself, a mac68k user) could have another binary-compatable alternative. Also, non-gui binary support for all of the m68k machines would be nothing to sneeze at. ;) In fact, I've seen this thread going on just about enough to suspect a critical mass of people to start a porting effort. If you are willing to put some amount of time and/or resources into a NetBSD/next68k port, Email me at rankin@ewl.uky.edu, and let me know. If enough people who can code will, we might just get the thing off the ground. :) Also, if anyone is already working on a port, please let me know. David -- David W. Rankin, Jr. Technical Support, Engineering Workstation Laboratory rankin@ewl.uky.edu PGP Key via http://www.ewl.uky.edu/rankin/key.asc Alphanumeric Pager: rankin-pager@ewl.uky.edu or (606) 330-4922 Seen on a California license plate on a VW Beetle recently: "FEATURE"
From: rankin@ewl.uky.edu (David W. Rankin Jr.) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: cmsg cancel <4f9gs9$n0d@t3.mscf.uky.edu> Control: cancel <4f9gs9$n0d@t3.mscf.uky.edu> Date: 7 Feb 1996 07:01:21 GMT Organization: University of Kentucky, Dept. of Math Sciences Message-ID: <4f9io1$nhi@t3.mscf.uky.edu> <4f9gs9$n0d@t3.mscf.uky.edu> was cancelled from within trn. -- David W. Rankin, Jr. Technical Support, Engineering Workstation Laboratory rankin@ewl.uky.edu PGP Key via http://www.ewl.uky.edu/rankin/key.asc Alphanumeric Pager: rankin-pager@ewl.uky.edu or (606) 330-4922 Seen on a California license plate on a VW Beetle recently: "FEATURE"
From: rankin@ewl.uky.edu (David W. Rankin Jr.) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: cmsg cancel <4f9gs9$n0d@t3.mscf.uky.edu> Control: cancel <4f9gs9$n0d@t3.mscf.uky.edu> Date: 7 Feb 1996 07:02:38 GMT Organization: University of Kentucky, Dept. of Math Sciences Message-ID: <4f9iqe$nhr@t3.mscf.uky.edu> <4f9gs9$n0d@t3.mscf.uky.edu> was cancelled from within trn. -- David W. Rankin, Jr. Technical Support, Engineering Workstation Laboratory rankin@ewl.uky.edu PGP Key via http://www.ewl.uky.edu/rankin/key.asc Alphanumeric Pager: rankin-pager@ewl.uky.edu or (606) 330-4922 Seen on a California license plate on a VW Beetle recently: "FEATURE"
From: rankin@ewl.uky.edu (David W. Rankin Jr.) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: 7 Feb 1996 07:19:17 GMT Organization: UK Microlabs -- Engineering Workstation Laboratory Message-ID: <4f9jpl$nnh@t3.mscf.uky.edu> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> contents of this message do not represent the opinions of the University of Kentucky, UK Microlabs, UKMSCF, or UKCC. In article <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU>, Mark Crispin <mrc@CAC.Washington.EDU> wrote: }On 4 Feb 1996, Louis A. Mamakos wrote: }> Again, I'd reiterate that I'd love to run a modern OS on my black next }> hardware. There is no requirement, for me at least, that it run }> existing NeXTSTEP applications, or host OPENSTEP. All I'm looking for }> is a stock 4.4BSD based OS, like NetBSD. }I second this. It'd be awfully nice if there was /dev/audio support to }make the DSP do its thing, but it isn't really necessary. NeXT's software }offerings are completely uninteresting. }It would also open up the possibility of running real Web browsers, }instead of trash like OmniWeb. When a dinky 68030 Mac PowerBook 165c does }a better job on the Web than an "insanely great" NEXTSTEP system, that }says something very sad. I'm a NetBSD bigot myself, so I can't really say TOO much about OmniWeb, except that ANYTHING is better than Nutscape. :) Even so, NetBSD/next68k would be a win for a lot of people, esp. those who have another NetBSD/m68k box in their clutches. I for one like the idea of a NetBSD/next68k port because it'd allow me to use the same binaries as my Mac IIsi uses. Plus, black box hardware can't be easily beaten. :) In fact, I've seen this thread going on enough in NetBSD areas to suspect a critical mass of people exists for starting a porting effort. If you are willing to put some amount of time and/or resources into a NetBSD/next68k port, Email me at rankin@ewl.uky.edu, and let me know. If enough people who can code will, we might just get the thing off the ground. :) Also, if anyone is already working on a port, please let me know. David -- David W. Rankin, Jr. Technical Support, Engineering Workstation Laboratory rankin@ewl.uky.edu PGP Key via http://www.ewl.uky.edu/rankin/key.asc Alphanumeric Pager: rankin-pager@ewl.uky.edu or (606) 330-4922 Seen on a California license plate on a VW Beetle recently: "FEATURE"
From: rsmallwo@islandnet.com Newsgroups: comp.sys.next.programmer Subject: Problem installing Server provided with WebObjects Beta Date: Wed, 7 Feb 1996 06:06:07 GMT Organization: Island Net in Victoria, B.C. Canada Message-ID: <960206220607.293AAD6E.rsmallwo@canada> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII To install the Apache server provided with the WebObjects beta, I have followed Next's instructions posted on their web site for starting the Next installer program from a terminal session as follows: /NextAdmin/Installer.app/Installer &</pre> Executing this line gives me a 'missing name for redirect' error message. Obviously my knowledge of the native unix side of Next is scant. Can anyone suggest how to correct this error? Please respond to 'rsmallwo@islandnet.com' Thanks in advance. Rhys
From: Paul_Lynch@plsys.co.uk (Paul Lynch) Newsgroups: comp.sys.next.programmer Subject: Re: NXTableView Date: Wed, 7 Feb 1996 08:44:15 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Feb7.084415.6004@seer.demon.co.uk> References: <4f6jdo$sre@huey.cadvision.com> In article <4f6jdo$sre@huey.cadvision.com> caljouwr@cadvision.com writes: > Does anyone know how to get a value from a NXTableView? > > I have set up the setActions and selectors and can get the > row, column and dataSource but am unable to retrieve the > value in the selected row and column. > > I think the difficulty lies in declaring an object to be a NXTableValue > for the into parameter but I am not sure. Unfortunately, this isn't possible to do in a well documented way. I use: [(id)[table dataSource] getValueFor:[[table columnAt:j] identifier] at:i into:tmpValue]; where tmpValue is declared as: id <NXTableValue> tmpValue = [[[[[table formatterAt:0 :0] value] class] alloc] init]; I use a category of NXValueFormatter that looks like: #import <eointerface/NXFormatter.h> @interface NXFormatter(NXValueFormatter) - value; @end @implementation NXFormatter(NXValueFormatter) - value { return value; } @end Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: Gerrit van Dyk <gvandyk@icon.co.za> Newsgroups: comp.sys.next.programmer Subject: Forwarding Date: Wed, 07 Feb 1996 07:09:24 +0200 Organization: E.S. Systems Message-ID: <31183404.754E@icon.co.za> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I want to implement an object that will forward all messages that it does not understand to a delegate object. I implemented the foward:: method but the problem i am facing now is that i can only forward messages that will return either a pointer or an id, which is also a pointer. Is there any way around this? There should be, because NeXT is doing the same type of thing in a NXProxy. I know that this mechanism is much easier in using the foundation classes, but my problem right know is that the object that needs to forward the messages is a view. In the current implementation of NeXTStep these classes inherit of Object and the foundation behaviour is not implemented in the Object hierarchy. I need to deploy the application using the current release of NeXTStep as we cannot wait until the middle of this year when version 4.0 should be released. Any Help would be appreciated. Thanks Gerrit
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.next.programmer Subject: Re: Problem w/ Mouse-X X11R4 Server Date: 7 Feb 1996 09:48:23 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <4f9sh7$e2d@news.lth.se> References: <4f6fqv$hl8@cville-srv.wam.umd.edu> In article <4f6fqv$hl8@cville-srv.wam.umd.edu>, Gregory John Casamento <gcasa@wam.umd.edu> wrote: >Hi, I am having trouble w/ Mouse-X. Basically the problem is that >after I switch back to NeXTSTEP, using the Command-Command-Delete >sequence, when I go back into X the xterm window will not accept >anything and the mouse doesn't seem to work. > >Does anybody know what is going on here?? If you are talking abot the same thing that happen to me, then I have a pretty good guess: When you press command-command-delete, the X server receives keydown events for the command keys as you press them -- but when you press the delete key, you switch to NeXTSTEP. Switching back to X, the X server still thinks the command keys are pressed, since it never got the corresponding key-up events. Pressing the command key(s) usually does the trick for me (I usually press Command, Shift, and Alt,one at a time, in quick succession,and after that, everything works just fine). > >Thanks, > >-- >Gregory John Casamento -- gcasa@wam.umd.edu >(c) G. Casamento -- Permission to distribute on MS network denied!! // Christian Brunschen #import <http://www.efd.lth.se/~d89cb/index,html>
From: jcr@idiom.com (John C. Randolph) Newsgroups: comp.sys.next.programmer Subject: Dictionary interface to Ivars. Date: 7 Feb 1996 06:05:10 -0800 Organization: A poorly-installed InterNetNews site Message-ID: <jcr.823701682@idiom.com> I'm writing an NSDictionary subclass that gives a dictionary-style interface to the ivars of any object. If anyone would like a copy of this when I'm done, drop me a note at <jcr@idiom.com>. I'm not creating a Dictionary, and then filling it in with the values of the ivars, the way that it's done in EOF. My NSDictionary subclass doesn't duplicate storage, it just attaches to the object in question, and looks like a dict where the keys are the names of the ivars, and the values are whatever's in those ivars. Anyway, if you want a copy of it, drop me a note. This might be a misckit submission soon. -jcr
From: fa@emf.emf.net (Farhad Afrahi) Newsgroups: comp.sys.next.programmer Subject: GCC binaries? Date: 7 Feb 1996 14:40:52 GMT Organization: Objekt Asset Management Message-ID: <4fadlk$tm@emf.emf.net> I am looking for binaries for gcc (actually, any ANSI-C compiler would do), so that I can build gcc 2.7.2 for NEXTSTEP 3.3 on Intel. I obviously don't have NEXTSTEP Developer. I would greatly appreciate any pointers to ftp site which carry such a beast. --- Farhad Afrahi <fa@emf.net> NeXTmail + MIME accepted
From: Gerrit van Dyk <gvandyk@icon.co.za> Newsgroups: comp.sys.next.programmer Subject: Archive List of Newsgroup Date: Wed, 07 Feb 1996 13:08:25 +0200 Organization: E.S. Systems Message-ID: <31188829.3E21@icon.co.za> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Is there an Archive of this newsgroup lying around somewhere that i can ftp. Thanks
From: cph@next.com (Charles d'Harcourt) Newsgroups: comp.sys.next.programmer Subject: Re: Problem installing Server provided with WebObjects Beta Date: 7 Feb 1996 16:18:32 GMT Organization: NeXT Software, Inc. Message-ID: <4fajco$r61@news.next.com> References: <960206220607.293AAD6E.rsmallwo@canada> On Wed, 7 Feb 1996 rsmallwo@islandnet.com writes: > To install the Apache server provided with the WebObjects beta, I have > followed Next's instructions posted on their web site for starting the > Next installer program from a terminal session as follows: > > /NextAdmin/Installer.app/Installer &</pre> > > Executing this line gives me a 'missing name for redirect' error > message. There was a typo (which should now be fixed) in the instructions for downloading WebObjects for NEXTSTEP: The </pre> should not be visible, and so the command line to type in a terminal window should be: /NextAdmin/Installer.app/Installer & Best of luck with installing and using WebObjects! C= __________________________________________________________ Charles d'Harcourt, WebObjects Group, NeXT Software, Inc. mailto:cph@next.com http://www.next.com/~cph Not speaking on behalf of NeXT or postmodern punctuations
From: Joe Freeman <Joe@FreemanSoft.COM> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: CORBA IDL Date: Wed, 07 Feb 1996 11:32:42 -0500 Organization: FreemanSoft Inc. Message-ID: <3118D42A.2FB0@FreemanSoft.COM> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I've been looking at the NeXT web site for more information on when their "CORBA xxxx" kits will actually be useable in a CORBA environment. One of the documents says they will have an IDL specification and compiler "this year" but the document looks like it was written in 1995. Has anyone seen more information on when a CORBA IDL layer will be available for Objective-C? I know that Duh-OLE is supposed to have this Objective-C to CORBA transfer mechanism , as was show two years a go in SJs demonstration, but how do you specify the language binding so it knows how to do the translation. <joe> -- The governer of MD wants to spend $300M on a rent free pro football stadium which is only used 10 times a year. For whom is he really doing this deal?
Newsgroups: comp.sys.next.programmer From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: GCC binaries? Sender: news@novice.uwaterloo.ca (system PRIVILEGED account) Message-ID: <DMEyJr.2Mr@novice.uwaterloo.ca> Date: Wed, 7 Feb 1996 16:16:38 GMT References: <4fadlk$tm@emf.emf.net> Organization: University of Waterloo In article <4fadlk$tm@emf.emf.net>, Farhad Afrahi <fa@emf.emf.net> wrote: >I am looking for binaries for gcc (actually, any ANSI-C compiler would >do), so that I can build gcc 2.7.2 for NEXTSTEP 3.3 on Intel. I >obviously don't have NEXTSTEP Developer. I would greatly appreciate >any pointers to ftp site which carry such a beast. > The gcc binaries won't help you, since you won't have the necessary linker libraries or include files. NS Dev is pretty much your only solution. -- David Evans (NeXTMail OK) dfevans@bbcr.uwaterloo.ca Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/ University of Waterloo "Default is the value selected by the composer Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: Matt Watson <Matt_Watson@NeXT.com> Newsgroups: comp.sys.next.programmer Subject: Re: Dictionary interface to Ivars. Date: Wed, 07 Feb 1996 10:55:35 -0800 Organization: NeXT Software, Inc. Message-ID: <3118F5A7.5308@NeXT.com> References: <jcr.823701682@idiom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit John C. Randolph wrote: > > I'm writing an NSDictionary subclass that gives a dictionary-style > interface to the ivars of any object. [...] > My NSDictionary subclass doesn't duplicate storage, it just attaches to the > object in question, and looks like a dict where the keys are the names of the > ivars, and the values are whatever's in those ivars. > Is it mutable? How do you represent scalar and aggregate types? matt.
From: Matt Watson <Matt_Watson@NeXT.com> Newsgroups: comp.sys.next.programmer Subject: Re: Dictionary interface to Ivars. Date: Wed, 07 Feb 1996 10:57:51 -0800 Organization: NeXT Software, Inc. Message-ID: <3118F62F.3AC9@NeXT.com> References: <jcr.823701682@idiom.com> <3118F5A7.5308@NeXT.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Matt Watson wrote: > > John C. Randolph wrote: > > > > I'm writing an NSDictionary subclass that gives a dictionary-style > > interface to the ivars of any object. [...] > > My NSDictionary subclass doesn't duplicate storage, it just attaches to the > > object in question, and looks like a dict where the keys are the names of the > > ivars, and the values are whatever's in those ivars. (Silly Netscape. I meant not to quote myself...) Is it mutable? How do you represent scalar and aggregate types? matt.
From: mmuni@plato.sky.bdm.com (Murali Muni,,,) Newsgroups: comp.sys.next.programmer Subject: Adaptor Object for File I/O Date: 7 Feb 1996 23:14:59 GMT Organization: BDM International, Inc. Message-ID: <4fbbpj$76n@news.mcl.bdm.com> I have the following requirement on my project: "To Read/Log Errors [Database and Logical(functional)], we need to read and write to a resource file on the system. We are using the EOF environment for database operations, But as per the requirement, the application should log logical errors to a flat file on the system. Is there an Adaptor Object which could do File based Operations, either supplied by NeXT or any of the Internet sites ? Thanks Murali Muni ------------ mmuni@plato.sky.bdm.com
From: "Dennis G. Norrod" <dnorrod@visuallink.com> Newsgroups: comp.sys.next.programmer Subject: Need to clear window counterPart info Date: 7 Feb 1996 23:49:56 GMT Organization: MRJ, inc Message-ID: <4fbdr4$m1j@visuallink.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi All, When the Window class is archived, the miniwindow and counterPart info is saved. When that window is read:, the miniwindow goes to the same location that it did before, instead of the first open slot. If an icon is already in this slot, it is overlayed (covered). Is there any way to set counterPart to nil? I think this would make the miniwindow go where it's suppose to go. Thanks. Dennis
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: CONTRACT POSITION/NEXTSTEP Developer Date: 8 Feb 1996 01:11:57 GMT Organization: Norden 1 Communications Message-ID: <4fbikt$rgj@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF---A plus Sybase---A plus Contract--long, long term Greater DC Area To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: Developer/Perm/ILL Date: 8 Feb 1996 01:14:52 GMT Organization: Norden 1 Communications Message-ID: <4fbiqc$rgj@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF---a plus Sybase---A plus Career position Full benefits Relocation assistance To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: peckover@atlsci.com (Kevin D. Peckover, B.Sc.(M.E.)) Newsgroups: comp.sys.next.programmer Subject: GNU gcc and Objective-C on Win95 Date: Tue, 06 Feb 1996 15:59:26 GMT Organization: Entrenet Systems Message-ID: <4f80nh$dun@entrenet.entrenet.com> The GNU documentation that comes with GCC-2.7.2 briefly mensions how to compile GCC for WIN NT. I am told that this target will probably also work for WIN 95. The ADA package has a successfully built GNU-C (only) compiler for Win NT/95. I have looked at making the entire package compile on my Win 95 box, but, I would rather find somone else who has already done this :-). So, I ask: Does anybody have or know of where One can get the GNU 2.7.2 compiler in binary form with the C,C++, & Obj-C compilers and libraries ? Where the binaries will work on WIN95/NT ? (ie: where native NT calls are made rather then having executables dependent on GO32). it has been compiled for Win NT and I have gotten the binary form of the compiler from the ADA package with only C, and C++ compiled in the binary. But, I would like to programm with Obj-C. Thanx Much. kevin.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9602072356.AA21834@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Thu, 8 Feb 96 02:56:44 +0300 Subject: Re: First WebObjects question In article <00001c42+00000a78@msn.com>, Wei Chen wrote: > Does anybody know the url format to launch a WebScript program, in general? > Right now I have to put all my programs under > /NextLibrary/WebObjects/Examples, > then use http://localhost/cgi-bin/WebObjects/Examples/myApp. So have anybody already figured out how to start programs from some other directory? I have encountered the following strange behavior 1. When I copied my test WebObjects program Foo originally located at /Users/zander/Library/WebObjects to /NextLibrary/WebObjects/Examples everything works just fine. 2.When I created a link /NextLibrary/WebObjects/Examples/Foo.wo -> /Users/zander/Library/WebObjects/Foo/@, DefaultApp fails to find reusable components and I've got the following message Application Name = 'Examples/Foo' Exception Name = 'NSGenericException' Exception Reason = 'Error while parsing Declaration file: Main.wod. Path: /Users/zander/Library/WebObjects/Foo/Main.wo. Cannot find HTML template for component named Footer. Component may not exist.' 3. When I created a test directory /NextLibrary/WebObjects/Test, copied and linked Foo and pointed my browser to http://localhost/cgi-bin/WebObjects/Test/Foo WebObjects adaptor a.k.a. WebServer/cgi-bin/WebObjects fails to autostart my programs at all and as a result I've got "Did not receive any response from application.". Could someone please explain this behavior and give me a solution. I don't like copying components into Examples directory each time I modify them. And the last but not least. It will be cool to be able to configure which directories WebObjects recognize as directories contained WebObjects components, i.e. something like srm.conf's ScriptAlias for cgi scripts. I want each user to have a directory for WebObjects' components and ability to autostart these components by opening URL: http://localhost/cgi-bin/WebObjects/~username/Foo. Regards, Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9602080103.AA22279@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Thu, 8 Feb 96 04:03:02 +0300 Subject: WebObjects, NSArray, FoundationKit and how it works on Solaris and NT? Hello, While checking WebObjects documentation I have found NSArray methods I haven't seen before. The methods are -arrayWithContentsOfFile: and -writeToFile: atomically:. Neither EOF's FoundationKit documentation, nor OpenStep specification doesn't define such methods. So am I missing something or these methods are WebObjects extensions? I wonder how WebObjects beta works on non NeXTSTEP platforms. Does it require FoundationKit or they hardlinked FoundationKit into DefaultApp? Have anybody successfully run any WebObjects's application on Solaris or NT? Regards, Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Newsgroups: comp.sys.next.programmer From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Window delegate doesn't get keyDown: Message-ID: <DMFsIp.94@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College Date: Thu, 8 Feb 1996 03:04:01 GMT Why doesn't the Window delegate receive any keyDown: events that are passed up the responder chain? Is this a bug? Did I misinterpret the docs? This is what it says in the Window doc: setDelegate: - setDelegate:anObject Makes anObject the Window's delegate, and returns self. A Window's delegate is given a chance to respond to action messages that work their way up the responder chain to the Window (through Application's sendAction:to:from: method). It can also respond to notification messages sent by the Window. -- Chris Osborn, Network Administrator Voice: 707 253 3130 Napa Valley College Fax: 707 253 3063 2277 Napa-Vallejo Hwy., Napa, CA, 94558 <fozztexx@nvc.cc.ca.us> MIME ok, NeXTMail tolerated
Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin From: cjones@netcom.com (Carl Jones) Subject: too many "beeps" Message-ID: <cjonesDMFzFu.Kpq@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Thu, 8 Feb 1996 05:33:30 GMT Sender: cjones@netcom21.netcom.com I have NeXT runing on a Gateway Intelmachine; verion 3.3 On the occasions when I do something that would give me a system beep, I get eight (8!) of them! each and every time! for instance, if I open a terminal window and, at the system name prompt (hostname>) I type a back arrow, I get a [D printed and 8 "you goofed" sounds! This did not used to happen (before I re-installed 3.3 and installed ppp; that is the only thing I can think of that I did). Anyway, anyone have any suggestions? It is driving me nuts! thanks, carl jones
Newsgroups: comp.sys.next.programmer From: Fabien_Roy@free.fdn.org Subject: Re: Adaptor Object for File I/O Message-ID: <1996Feb8.072511.7472@free.fdn.org> Sender: news@free.fdn.org Organization: Fabien Roy Consultant, Paris, France References: <4fbbpj$76n@news.mcl.bdm.com> Date: Thu, 8 Feb 1996 07:25:11 GMT See the FlatFileDataSource in /NextDeveloper/Examples/EnterpriseObjects. Hope that helps. Fabien -- --------------------------------------------------------------------- Fabien_Roy@free.fdn.org (NextMail/MIME accepted) Fabien Roy Consultant NEXTSTEP/OPENSTEP/EOF Consultant, SYBASE DBA 10 rue de la DEFENSE 93100 MONTREUIL, France Tel: 33 1 45 28 32 23 Fax: 33 1 48 55 09 90
From: Mark Crispin <mrc@CAC.Washington.EDU> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: NeXTBSD - 4.4BSD For NeXTstep machines ? Date: Thu, 8 Feb 1996 00:44:03 -0800 Organization: Networks & Distributed Computing Message-ID: <Pine.NXT.3.92.960207232551.25734A-100000@Tomobiki-Cho.CAC.Washington.EDU> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII NNTP-Posting-User: den To: William Shipley <wjs@omnigroup.com> In-Reply-To: <4f9akj$2co@gaea.titan.org> On 7 Feb 1996, William Shipley wrote: > (c) you didn't have an old vendetta against me and my company dating from > back when we were stuck at the UW as well. I have no idea what you are talking about. The best that I can guess is that maybe you were some student who flamed me because we didn't support Mail.app or some such. If you honestly believe that I would remember you for that, you have an excessively high opinion of yourself. > couldn't you at least be specific? I only used Omniweb once, and the experience was such an incredible exercise in frustration that I gave up. Here are some specifics for you: I tried to paste a URL from an email message into the "Open URL" panel. It beeped at me. I had to type it in manually. [No, I can't reproduce it now, but it was solid then.] When it got to the URL, I was hit by a barrage of model error panels complaining about the GIF images. It took a while before I realized that for some reason it decided to call an external program to resolve the GIF images and it was that program that was blatting out these panels. I didn't even know that program did GIF (it's a special purpose image resolution tool for a completely different application). I could find no way to tell Omniweb not to use that program, other than removing it. Netscape lets you configure this. But I plodded through, until I got to filling in a form and hit the submit button. It pressed, but did nothing. It was at that point ther I gave up and ran Netscape. I just tried it now, with a URL that displays perfectly well with Netscape, Mosaic, and eWorld's browser. I was rewarded with: "Fatal Error (#500) Can't Access Document http://icons/ftp.gif Reason: Can't locate remote host icons." The maintainer of that URL said, and I quote, "Your browser doesn't handle relative pointers right, common bug in old code". He also pointed out that <LI> entries weren't displaying bullets, and that lines which are delimited by <BR> had bogus doublespaces. "Fatal Error (#404)" when trying to open bookmarks, well golly gee, couldn't it have looked to find the Mosaic bookmarks in the well-known filename? Why does every dinky Web browser think that it has to have its own private bookmarks file? The fact that others get this wrong too doesn't excuse Omniweb. Plus, it should have offered to make a new file instead of saying "Fatal Error (#404)". "Open URL..." sometimes forces you to click in the TextCell for no apparent reason. To reproduce, first up OmniWeb and immediately type Command-Shift-O (which, by the way, is a terrible choice since Command-o is also used -- consider Command-l which is more-or-less standard). At other times, the cursor appears in the TextCell without requiring the mouse. "Open URL..." opens a new window, even though the Preference to do this is unchecked. I tried the URL that gave me the frustration before and all the same things happened, including the submit button doing nothing at all. As far as I could tell, it first collects the images, then invokes the external program. It seems to do this in lockstep, with no attempt to thread the gathering from the net with image resolution. This makes image display much slower than Netscape (especially Netscape 2.0 which just flies), even when the image resolver is the right program. This slowness is quite frustrating, particularly with URLs with lots of images; it reminds me of the way Mosaic used to perform. Try as I might, I was unable to find any international character support other than for Western Europe (ISO-8859-1). So much for using any URLs in East Asia. Nor did there seem to be any way to set fonts. I have no idea what version of Omniweb it is; I didn't install it. Maybe it is an old buggy version. The version number in the Info Panel is unreadable; don't you test things in monochrome (or is it your attitude that monochrome users shouldn't be using your software)? The latest file date seems to be February 3, 1995, if that helps. -- Mark -- DoD #0105, R90/6 pilot, FAX: (206) 685-4045 ICBM: N 47 39'35" W 122 18'39" Science does not emerge from voting, party politics, or public debate.
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Adaptor Object for File I/O Date: 8 Feb 1996 08:49:42 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4fcdf6$mhg@lion.embl-heidelberg.de> References: <4fbbpj$76n@news.mcl.bdm.com> In article <4fbbpj$76n@news.mcl.bdm.com> mmuni@plato.sky.bdm.com (Murali Muni,,,) writes: > I have the following requirement on my project: > > "To Read/Log Errors [Database and Logical(functional)], we > need to read and write to a resource file on the system. > > We are using the EOF environment for database operations, But > as per the requirement, the application should log logical > errors to a flat file on the system. > > Is there an Adaptor Object which could do File based > Operations, either supplied by NeXT or any of the > Internet sites ? I don't think so, but you can write it yourself. An example is the MiscLogFile from MiscKit. If you write your loging procedure, do not forget to lock the log file. The easiest way for file locking is creating and deleting a directory. You can also look closer the MiskLockFile class later -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: cph@next.com (Charles d'Harcourt) Newsgroups: comp.sys.next.programmer Subject: Re: WebObjects, NSArray, FoundationKit Date: 8 Feb 1996 09:18:25 GMT Organization: NeXT Software, Inc. Message-ID: <4fcf51$2kd@news.next.com> References: <9602080103.AA22279@cnext.crec.mipt.ru> On Thu, 8 Feb 96, Aleksey Sudakov <zander@cnext.crec.mipt.ru> writes: > While checking WebObjects documentation I have found NSArray > methods I haven't seen before. The methods are > -arrayWithContentsOfFile: and -writeToFile: atomically:. > Neither EOF's FoundationKit documentation, nor OpenStep > specification doesn't define such methods. So am I missing something > or these methods are WebObjects extensions? These are indeed WebObjects extensions to foundation. They are meant to simplify the foundation API for those accessing foundation through WebScript and not familiar with all of foundation and concepts like autoreleasing (we're serious when we say we want to make this product simple). The additional methods are: NSString: + string; + stringWithString:(NSString *)aString; + (NSString *)stringWithContentsOfFile:(NSString *)path; - (BOOL)writeToFile:(NSString *)path atomically:(BOOL)flag; - (NSComparisonResult)caseInsensitiveCompare:(NSString *)aString; NSArray: + arrayWithArray:(NSArray *)anArray; + (NSArray *)arrayWithContentsOfFile:(NSString *)path; - (BOOL)writeToFile:(NSString *)path atomically:(BOOL)flag; NSDictionary: + dictionaryWithDictionary:(NSDictionary *)aDictionary; + (NSDictionary *)dictionaryWithContentsOfFile:(NSString *)path; - (NSArray *)keysSortedByValueUsingSelector:(SEL)comparator; These are documented in the WebObjects documentation, and should be incorporated into OpenStep 4.0 for Mach and NT. Enjoy, C= ________________________________________________________ Charles d'Harcourt, WebObjects Group, NeXT Software, Inc. mailto:cph@next.com http://www.next.com/~cph Not speaking on behalf of NeXT or contemporary archives
From: kiwi@buran.fb10.tu-berlin.de (Axel Habermann) Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: Re: too many "beeps" Followup-To: comp.sys.next.programmer,comp.sys.next.sysadmin Date: 8 Feb 1996 09:25:31 GMT Organization: Technical University Berlin, Germany Message-ID: <4fcfib$16b@brachio.zrz.TU-Berlin.DE> References: <cjonesDMFzFu.Kpq@netcom.com> Carl Jones (cjones@netcom.com) wrote: : I have NeXT runing on a Gateway Intelmachine; verion 3.3 : On the occasions when I do something that would give me a system beep, : I get eight (8!) of them! each and every time! Repeating sounds are mostly due to a misconfigured (that is: used twice) irq line for the soundcard. -- Axel Habermann kiwi@buran.fb10.tu-berlin.de Fon:+49 30 45478986 Fax:4542296 Die Dateien, in denen die Programmdokumentation enthalten ist, haben normalerweise die Endung ".c", -- Kristian Koehntopp
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Dictionary interface to Ivars. Date: 8 Feb 1996 09:51:54 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4fch3q$nbr@lion.embl-heidelberg.de> References: <jcr.823701682@idiom.com> In article <jcr.823701682@idiom.com> jcr@idiom.com (John C. Randolph) writes: > I'm writing an NSDictionary subclass that gives a dictionary-style > interface to the ivars of any object. If anyone would like a copy of > this when I'm done, drop me a note at <jcr@idiom.com>. I wrote the same thing for class variables. Very nifty :-( It's part of the SciTools project (e.g. free ;-) later -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: cph@next.com (Charles d'Harcourt) Newsgroups: comp.sys.next.programmer Subject: Re: First WebObjects question Date: 8 Feb 1996 10:02:34 GMT Organization: NeXT Software, Inc. Message-ID: <4fchnq$2r8@news.next.com> References: <9602072356.AA21834@cnext.crec.mipt.ru> On Thu, 8 Feb 96, Aleksey Sudakov <zander@cnext.crec.mipt.ru> writes: > In article <00001c42+00000a78@msn.com>, Wei Chen wrote: > > Does anybody know the url format to launch a WebScript program, > > in general? Right now I have to put all my programs under > > /NextLibrary/WebObjects/Examples, then use > > http://localhost/cgi-bin/WebObjects/Examples/myApp. > > So have anybody already figured out how to start programs from some > other directory? For autostarting (I won't get into the more complex case of "manually" started apps, which will be covered by documentation shipping with the product in March), web apps need to be somewhere in the "WebObjects Document Root". The "WebObjects Document Root" is the directory called "WebObjects" in your web server's document root. So if you're using NeXT's version of apache with the default configuration, you WebObjects Document Root will be /NextLibrary/WebServer/htdocs/WebObjects. if you're using NCSA with the default configuration, your WebObjects Document Root will be /usr/local/etc/httpd/htdocs/WebObjects. As part of installing WebObjects, a directory called Examples will be created in your WebObjects Document Root. You can create other directories there (for example I have a Tests subdirectory), or simply put your web apps directly under the WebObjects Document Root. Indeed, it will make upgrading WebObjects easier if you do your tests and work in a directory other than Examples. The general format (simple form) of a WebObjects URL is: http://YourHost/cgi-bin/WebObjects/YourApp Where YourHost is your web server's host name (or localhost if you're accessing this from the web server itself), and YourApp is you "web app name", which is actually the path of the web app relative to your WebObjects Document Root (eg. "Examples/HelloWorld", "Tests/Frames", "Foo"). I hope this helps, C= ________________________________________________________________ Charles d'Harcourt, WebObjects Group, NeXT Software, Inc. mailto:cph@next.com http://www.next.com/~cph Not speaking on behalf of NeXT or the grand duchy of Luxenbourg
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: RE: WebObjects, NSArray, FoundationKit and how it works on Solaris and NT? Date: 8 Feb 1996 10:11:15 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4fci83$nje@lion.embl-heidelberg.de> References: <9602080103.AA22279@cnext.crec.mipt.ru> In article <9602080103.AA22279@cnext.crec.mipt.ru> writes: > Hello, > > While checking WebObjects documentation I have found NSArray > methods I haven't seen before. The methods are > -arrayWithContentsOfFile: and -writeToFile: atomically:. > Neither EOF's FoundationKit documentation, nor OpenStep > specification doesn't define such methods. So am I missing something > or these methods are WebObjects extensions? Hmmmm. As you know, OpenStep is still version 1.0 (e.g. hack ;-) There is a NSDictionary method called + (id)dictionaryWithContentsOfFile:(NSString *)path; ... very useful one... I hope the NeXT crew have seen that they are missing analogues methods in several foundation classes and decide to correct their mistake. Anyway, this was still not done in the NS4.0 alpha :-( BTW: This is my wish list for OS extensions: 1) The promised nib file format; 2) The not yet promised coding format; 3) ASCII coding format; 4) The Text class extensions; .. and a lot of other cools stuff ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: ukr8@rzstud2.rz.uni-karlsruhe.de (Jochen Quante) Newsgroups: comp.sys.next.programmer Subject: Sometimes field editor does not become first responder?! Date: 8 Feb 1996 16:30:59 GMT Organization: University of Karlsruhe, Germany Message-ID: <4fd8g3$mm@nz12.rz.uni-karlsruhe.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit NNTP-Posting-User: ukr8 Hi, I have a problem concerning the TextField/TextFieldCell/Text mechanism. Usually, when the user mouses down into a TextField (or TextFieldCell?), the TextFin? May I change text already typed in? 3. Why did NeXT implement the filtering of text via filter functions? ^^^^^^^^^ There's no nice way to figure out the current text filter within a character filter function. Any pointers would be appreciated. Thanx so much in advance. Ivo B"ohme 24hours@work: i@abm.de 0hours@home: <>
From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Subject: Help connecting two modems Date: 8 Feb 1996 20:20:08 GMT Organization: University of California, Los Angeles Message-ID: <4fdlto$iu8@saba.info.ucla.edu> Can someone please tell me how to directly connect two modems, so that I may transfer huge amoundt of data from one computer to the other. Rick UCLA Physics I think all I need is a cable and some command on each modem.
From: rasmussn@swallow.eecs.umich.edu (Craig Rasmussen) Newsgroups: comp.sys.next.programmer Subject: How to kill a DO connection gracefully? Date: 8 Feb 1996 20:58:27 GMT Organization: University of Michigan, Ann Arbor, Michigan, USA Message-ID: <4fdo5j$mk9@news.eecs.umich.edu> I thought that I had this problem solved and now I'm beginning to wonder. Suppose I get a connection to a server: server = [NXConnection connectToName:"aName" onHost:"aHost"]; Then I'm vended an object by the server: remote = [server returnSomeObjectTo:self]; Now I do some work and want to gracefully free the server and remote objects and the NXConnection used to communicate with the server. I've tried the following: connection = [server connectionForProxy]; [remote freeProxy]; [server freeProxy]; [connection free]; But the [connection free] leads to chaos and without it, I can't kill the connection and have the death reported on the server side. Thanks to all, Craig Rasmussen
From: ivo.welch@agsm.ucla.edu (Ivo Welch) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: OmniWeb Date: 8 Feb 1996 21:49:28 GMT Organization: The Anderson School at UCLA Message-ID: <4fdr58$ojc@risc.agsm.ucla.edu> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> In-Reply-To: <4f9akj$2co@gaea.titan.org> For the record: OmniWeb is a terrific program. Works as advertised. Much better than netscape/win in my opinion. And I have no connection with Lighthouse Design. The price is terrific, too, on such a low distribution platform as NS. /ivo --- Ivo Welch ivo.welch@anderson.ucla.edu Assoc Prof of Finance Anderson GSM at UCLA 110 Westwood Plaza, Box 951481, LA CA 90095-1481 -- Ivo Welch ivo.welch@anderson.ucla.edu Assoc Prof of Finance Anderson GSM at UCLA 110 Westwood Plaza, Box 951481, LA CA 90095-1481
From: klett002@maroon.tc.umn.edu (James P. Klett) Newsgroups: comp.sys.next.programmer Subject: Inheritance and EOF Question Date: 8 Feb 1996 22:25:58 GMT Organization: UofM Alumni Message-ID: <4fdt9m$juk@epx.cis.umn.edu> Suppose you want to model a simple real world relationship using Objects and inheritance, and then using EOF, EOModeler, and A relational DB like SQL Server or QuickBase etc... want to put it in motion, how would it work? There are two examples I can think of that both work differently, 1. (Two Classes, 1 DB Table?) Suppose you have a Superclass called Pets. That class contains the name and age of the pet, and a subclass for each type of pet, like Cat, Dog... which might have data about the food it eats etc... Using the above tools how would this work so that when a Cat or Dog is accessed from the DB, the information is correctly placed in both the parent and child classes. 2. (Two Classes, 2 Tables?) The classic Order Entry, You have an order table/class and an order_detail table class. The order detail is a child class of order since it requires access to all the information in order, but it also has its own data as to the item ordered, qty's etc... Using the tools, you want to read data into the parent class, and write it back, if modified, when the child class is accessed. Shouldn't this "Just Work"? or is there more to it?? I have not seen any EOF Examples handling inheritance like this, therefore, I ask the net :) Thanks, JIM ___ ___....-----'---'-----....___ ========================================= ___'---..._______...---'___ (___) _|_|_|_ (___) \\____.-'_.---._'-.____// klett002@maroon.tc.umn.edu
From: klett002@maroon.tc.umn.edu (James P. Klett) Newsgroups: comp.sys.next.programmer Subject: Garbage collection in NeXTStep? Date: 8 Feb 1996 22:31:32 GMT Organization: UofM Alumni Message-ID: <4fdtk4$juk@epx.cis.umn.edu> After reading the article in this months Object magazine, I am wondering if there are any Garbage collecting librairies available for NeXTStep. For those who are wondering what the hech I mean, this is a bit of code that links in with an application and keeps track of an allocated memory that is no longer in use (incorrectly deallocated by the program) and it frees it for you. It is supposed to increase debugging time, run-time and provide for a more robust application. Thanks, JIM ___ ___....-----'---'-----....___ ========================================= ___'---..._______...---'___ (___) _|_|_|_ (___) \\____.-'_.---._'-.____// klett002@maroon.tc.umn.edu
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: How to kill a DO connection gracefully? Date: 9 Feb 1996 09:02:29 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4ff2j5$bme@lion.embl-heidelberg.de> References: <4fdo5j$mk9@news.eecs.umich.edu> In article <4fdo5j$mk9@news.eecs.umich.edu> rasmussn@swallow.eecs.umich.edu (Craig Rasmussen) writes: > I thought that I had this problem solved and now I'm beginning to wonder. > Suppose I get a connection to a server: > > server = [NXConnection connectToName:"aName" onHost:"aHost"]; > > Then I'm vended an object by the server: > > remote = [server returnSomeObjectTo:self]; > > Now I do some work and want to gracefully free the server and remote objects > and the NXConnection used to communicate with the server. > > I've tried the following: > > connection = [server connectionForProxy]; > [remote freeProxy]; > [server freeProxy]; > [connection free]; > > But the [connection free] leads to chaos and without it, I can't kill the > connection and have the death reported on the server side. If you only want prevent new clients from connecting to the server you first should remove the server's name from the network message server. To do this write a category like this one: @implementation NXConnection(UnregisteringServerName) + unregisterServerWithName:(char *)aName { return [NXNetNameServer checkOutPortWithName:aName]; } @end The following line then will terminate connections with existing clients: port_deallocate(task_self(),[[connection inPort] machPort]); All clients will receive senderIsInvalide: and you can happily free the connection. BTW: It is really good idea to use NXAutoreleaseConnection instead of NXConnection! -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Garbage collection in NeXTStep? Date: 09 Feb 1996 00:29:47 +0100 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7ivhhtmz8.fsf@tom.es.ele.tue.nl> References: <4fdtk4$juk@epx.cis.umn.edu> In-reply-to: klett002@maroon.tc.umn.edu's message of 8 Feb 1996 22:31:32 GMT In article <4fdtk4$juk@epx.cis.umn.edu> klett002@maroon.tc.umn.edu (James P. Klett) writes: It [garbage collection] is supposed to increase debugging time, run-time and provide for a more robust application. This is a joke, right? --Tiggr
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Setting PopUp's value? Date: 9 Feb 1996 09:37:30 GMT Organization: Turbocat's Development, Germany Message-ID: <4ff4kr$bo@turbocat.snafu.de> Hi! How do I set the initial Value of a PopUp? I know how to get the actual value and save it in the prefs. If the app restarts -- the wrong item will be visible. How do I change the visible Item? Can I ftp an demo-app somewhere that does this? Thanks in advance. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: jes@rednsi.com (Josep Egea i Sanchez) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: fflush not working Date: 9 Feb 1996 12:32:39 GMT Organization: Medusa Message-ID: <4ffet7$2c4@news.medusa.es> Keywords: posix fflush Hi all, Compiling third party software which needs the -posix flag to work properly, I've found that fflush() doesn't work as stated. In the manual pages can be read that if the argument is a NULL pointer, all output and update streams are flushed but in my case it causes a bus error. Have you experienced this before? Any comments welcome! Thanks and regards. -- Josep Egea - jes@rednsi.com - NeXTMail & MIME OK Nexus Servicios de Informacion - Barcelona (Spain) Telf: + 34 3 285 00 70 - Fax: + 34 3 285 00 70
From: mark@dlib.com (Mark Middleton) Newsgroups: comp.lang.java,comp.lang.smalltalk,comp.object,comp.sys.next.programmer Subject: Agents/Web/Objects Developers and Systems Integrators Wanted Date: Fri, 09 Feb 1996 12:05:06 GMT Organization: BTnet, BT Public Internet Service Message-ID: <4ffd7a$43e@news.bt.net> Please check out http://www.dlib.com/ for job vacancy information in the Mitsubishi Electric Digital Library Co. +-------------------------------------------+ | Mark Middleton | mark@dlib.com | | System Architect | | +-------------------------------------------+
From: Mark Middleton <mark@dlib.com> Newsgroups: comp.sys.next.programmer Subject: Agents/Web/Objects Developers and Systems Integrators Wanted Date: Fri, 09 Feb 1996 13:10:31 +0000 Organization: Mitsubishi Electric Digital Library Group Message-ID: <311B47C7.373B@dlib.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Please check out http://www.dlib.com/ for job vacancy information in the Mitsubishi Electric Digital Library Co. Mark -- +-------------------------------------------+ | Mitsubishi Electric Digital Library Group | |-------------------------------------------| | Mark Middleton | mark@dlib.com | | System Architect | http://www.dlib.com/ | +-------------------------------------------+
From: Steve Edwards - SEWP <edwards@sewp.nasa.gov> Newsgroups: comp.sys.next.programmer Subject: Re: Inheritance and EOF Question Date: Fri, 9 Feb 1996 14:58:44 GMT Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <960209095844.205AADnE.edwards@noodle> References: <4fdt9m$juk@epx.cis.umn.edu> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII >Suppose you want to model a simple real world relationship using Objects >and inheritance, and then using EOF, EOModeler, and A relational DB like >SQL Server or QuickBase etc... want to put it in motion, how would it work? > >There are two examples I can think of that both work differently, > >1. (Two Classes, 1 DB Table?) Suppose you have a Superclass called Pets. >That class contains the name and age of the pet, and a subclass for each >type of pet, like Cat, Dog... which might have data about the food it eats >etc... > >Using the above tools how would this work so that when a Cat or Dog is >accessed from the DB, the information is correctly placed in both the >parent and child classes. EOF has a hook that addresses this problem directly. You have to include a type identifier as one of the attributes of the primary key for the Pet table (i.e., a column that says, "this is a dog/cat/hamster row"). Then in your Pet class, implement the method initWithPrimaryKey:entity. This will give you the chance to substitute an instance of a specific child class of Pet. The resulting object will then be initialized with takeValuesFromDictionary: as usual. See the section "Multiple Classes for an Entity" in the "Designing Enterprise Objects" chapter of the EOF documentation. > >2. (Two Classes, 2 Tables?) The classic Order Entry, You have an order >table/class and an order_detail table class. The order detail is a child >class of order since it requires access to all the information in order, >but it also has its own data as to the item ordered, qty's etc... > >Using the tools, you want to read data into the parent class, and write >it back, if modified, when the child class is accessed. Shouldn't this >"Just Work"? or is there more to it?? If you really wanted to structure this as an inheritance relationship, then it seems the above approach would work in this case also. However, I would treat this as a containment relationship. Instead of Order and OrderDetail classes, have an Order and OrderedItem class. The OrderedItem describes a single item on the order (quantity, price, etc.). The Order, in addition to its own information (date, customer, etc.), contains a list of OrderedItem objects. You can then use EOModeler to specify a to-many relationship between the Order and OrderedItem entities. > >I have not seen any EOF Examples handling inheritance like this, therefore, >I ask the net :) > >Thanks, JIM > > > ___ > ___....-----'---'-----....___ > ========================================= > ___'---..._______...---'___ > (___) _|_|_|_ (___) > \\____.-'_.---._'-.____// klett002@maroon.tc.umn.edu --- Stephen G. Edwards NASA Goddard Space Flight Center Tel: 301-286-0631 Fax: 301-286-1619 edwards@sewp.nasa.gov
From: windemut@cumbnd.bioc.columbia.edu (Andreas Windemuth) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: 9 Feb 1996 16:21:13 GMT Organization: Columbia University Message-ID: <4ffs9p$m5f@apakabar.cc.columbia.edu> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> ivo.welch@agsm.ucla.edu (Ivo Welch) wrote: > >For the record: OmniWeb is a terrific program. Works as advertised. Much >better than netscape/win in my opinion. And I have no connection with >Lighthouse Design. The price is terrific, too, on such a low distribution >platform as NS. > I second that. OmniWeb 2.0 blows Netscape out of the water. --- Andreas Windemuth +-------------------------------------------------------------------- |Columbia University, Department of Biochemistry and Biophysics |630 West 168th St. BB-221 | tel: (212)-305-6884, fax: 6926, NeXTmail |New York, NY 10032 | email: windemut@cumbnd.bioc.columbia.edu +--------------------------------------------------------------------
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9602091441.AA06783@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Fri, 9 Feb 96 17:41:05 +0300 Subject: Re: Agents/Web/Objects Developers and Systems Integrators Wanted Cc: Mark Middleton <mark@dlib.com> In article <311B47C7.373B@dlib.com>, you wrote: > Please check out http://www.dlib.com/ for job vacancy information in the > Mitsubishi Electric Digital Library Co. cnext> ping www.dlib.com ping: unknown host www.dlib.com Regards, Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re:Re: Functionality of Perl's CGI.pm in WebObjects? Date: 9 Feb 1996 18:08:13 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4fg2id$94e@stc06.ctd.ornl.gov> References: <4f5q5n$2o8@sun0.urz.uni-heidelberg.de> In article <4f5q5n$2o8@sun0.urz.uni-heidelberg.de> dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) writes: > In article <4f5l5p$jic@vixen.cso.uiuc.edu> icardena@sumter.cso.uiuc.edu > (Ian Patrick Cardenas) writes: > > woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > > >How can I save (write the name/variable pairs) to disk from WebObjects? > or read > > >them in later. Is there some WebObject construct that would provide > the > > >equivalent? > > > > > I don't believe there's a need to save the parameters to disk as > WO will > > maintain it's state across pages. The reason for saving parameters to disk is that we are building a web application that creates a form, allows signatures to be added, and notifies the next-to-sign that a form awaits a signature. The actual document is retained on disk. I wouldn't want to loose the document (they must be kept for several years) because I shut down the system or rebooted. From WebObjects (just the script version) can I write to disk somehow?: - - - - - - - - - J. W. Wooten
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: Help connecting two modems Date: 9 Feb 1996 19:49:19 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4fg8fv$9dr@agate.berkeley.edu> References: <4fdlto$iu8@saba.info.ucla.edu> In article <4fdlto$iu8@saba.info.ucla.edu>, Rick Vazquez <vazquezr@physics.ucla.edu> wrote: >Can someone please tell me how to directly connect two modems, so that I may >transfer huge amoundt of data from one computer to the other. If the two computers have BNC-type ethernet connectors, then you don't need a modem cable, you need a BNC ethernet cable with some terminators. BNC cabling can be used to create a standalone network; I have two computers in a very small network, one NeXT on each end of a BNC cable. It works great, and you'll get a lot better through-put than on a serial port... Paul
From: john@luke.az.stratus.com (John Acuna) Newsgroups: comp.sys.next.programmer Subject: Problem loading .nib from bundle Date: 9 Feb 1996 23:28:48 GMT Organization: Stratus Computer Inc, Marlboro MA Message-ID: <4fglbh$ori@transfer.stratus.com> Problem: Attempting to load a .nib file from within a .bundle but failing in loadNibFile:owner: Console Error: Unknown error code 100003 in NXReportError Scenario: (below) TestApp (application delegate) loads a .bundle with... myBundle = [NSBundle bundleForClass:[self class]]; bundlePath = [myBundle pathForResource:@"QueryBuilder" ofType:@"bundle"]; aBundle = [NSBundle bundleWithPath:bundlePath]; queryBuilder = [[[aBundle classNamed:@"QueryBuilder"] alloc] init]; The .bundle contains a QueryBuilder (NSObject subclass) that is the target of 3 MenuCells from the main menu. All 3 target methods do basically the same thing which is to allocate an object (NSObject subclass) and request that they display their panel. All 3 objects load .nib files, 2 work correctly and the third returns the console error described above. The QueryBuilder class allocates the problem object with... [[[QBDocument alloc] init] display:mainWindowDelegate]; and the init of QBDocument fails in loadNibFile:owner: here... myBundle = [NSBundle bundleForClass:[self class]]; nibPath = [myBundle pathForResource:@"QBDocument" ofType:@"nib"]; . . . if (![NXApp loadNibFile:[nibPath cString] owner:self]) { . . . ...I can see from gdb that nibPath contains a valid path to the .nib file and that self is QBDocument. I have completely stripped QBDocument.nib to nothing but an empty .nib file with a File's Owner of a QBDocument that has no methods or instance variables and the problem will not go away. I have stripped all code out of QBDocument.m but the init routine and still cannot load the .nib file. Being a novice with .bundles, I am not sure where to go from here. Any ideas of what I am doing wrong here? --- John Acuna Services Systems Engineer IV VOS Mail: John_Acuna@vos.stratus.com Stratus Computer, Inc. NeXT Mail: john@luke.az.stratus.com Phoenix Customer Assistance Center:(800)828-8513 4455 E.Camelback Rd. -- Suite #115A -- Phoenix, AZ 85018
From: yliang@leland.Stanford.EDU (Ying Liang) Newsgroups: comp.sys.next.programmer Subject: Problem using addRetrieveOrder in DBKit Date: 10 Feb 1996 01:47:56 GMT Organization: Stanford University Message-ID: <4fgtgc$8ej@nntp.Stanford.EDU> I have a problem using DBRecordList method addRetrieveOrder:(DBRetrieveOrder)anOrder for:(id< DBProperties> )aProperty; in DBkit. For one search, I can get about 11 results back by just using didFetch = ([recordList fetchUsingQualifier:qualifier] but if I put [recordList addRetrieveOrder:DB_AscendingOrder for:sortProp]; right before the fetch above, I get 0 result back for exactly the same search. The sortProp is in the recordList before adding the RetrieveOrder, and the whole program has nothing to do with the IB, so it doesn't have the problem mentioned in the NextAnwers for addRetrieveOrder. Can anybody shed some light on this ? Any help would be greatly appreciated! Ying Liang yliang@leland.stanford.edu
From: "C. Knight" <knight@arsenal.com> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: Sat, 10 Feb 1996 02:28:18 -0600 Organization: Knight Productions, Inc. Message-ID: <311C5722.76AB@arsenal.com> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> <4ffs9p$m5f@apakabar.cc.columbia.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit windemut@cumbnd.bioc.columbia.edu (Andreas Windemuth) writes: >I second that. OmniWeb 2.0 blows Netscape out of the water. Who are you kidding? Netscape 2.0 b6 clobbers OmniWeb. Later, - C
From: devan2m@imap2.asu.edu Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: NEXTSTEP'er in Hong Kong? Date: 10 Feb 1996 09:22:56 GMT Organization: Arizona State University Message-ID: <4fho5g$38u@news.asu.edu> Keywords: NEXTSTEP, Advocates, Hong Kong Hello All, First, I was wondering whether there are any NEXTSTEP/OPENSTEP advocates in Hong Kong and who they might be? I know that there used to be a listing for a user group there, but I don't know how current this is. Secondly, would anybody also know about Internet access in Hong Kong? Any and all assistance would be dearly appreciated. Regards-
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9602100718.AA13215@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Sat, 10 Feb 96 10:18:44 +0300 Subject: Any way to force WOFrame reloading? Hi, I have a two frame component (smth like in a WOFrame example) with submit button in the top frame and WORepetition in the bottom one. I want bottom frame to be reloaded (i.e. WORepetition to be refilled) when I press submit button. Is it possible at all? Any ideas how to do it? Thanks in advance, Aleksey. _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9602101320.AA15820@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Sat, 10 Feb 96 16:20:28 +0300 Subject: How to set reusable components' attributes in WORepetition? Hi, I have WORepetition in one of my component with a list of reusable components Foo{ value1 = val1; value2 = val2; }; I want to fill this WORepetition with contents of NSArray of same (in sence of keys (val1, val2) NSDictionaries. Any ideas how could I do that? Here are component files <!-- html --> <HTML> <BODY> <WEBOBJECT NAME="REP"> <WEBOBJECT NAME="FOO"></WEBOBJECT><HR> </WEBOBJECT> </BODY> </HTML> // // wod // REP:WORepetition { list = aList; item = dict; }; FOO:Foo { dictionary = dict; } Thanks in advance, Aleksey. _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: kkwan@cs.hku.hk (Kelvin Kwan) Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: Re: NEXTSTEP'er in Hong Kong? Followup-To: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Date: 10 Feb 1996 15:35:37 GMT Organization: Department of Computer Science, The University of Hong Kong Message-ID: <4fie09$96q@ns.cs.hku.hk> References: <4fho5g$38u@news.asu.edu> devan2m@imap2.asu.edu wrote: : Hello All, : First, I was wondering whether there are any NEXTSTEP/OPENSTEP : advocates in Hong Kong and who they might be? I know that there used to be a : listing for a user group there, but I don't know how current this is. : Secondly, would anybody also know about Internet access in Hong : Kong? Any and all assistance would be dearly appreciated. I live in Hong Kong. I live on NEXTSTEP. Email me. -Kelvin
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Question about MABs Date: 10 Feb 1996 00:39:22 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4fgpfq$dgh@usenet.rpi.edu> References: <4f91ff$qvt@cloner2.ix.netcom.com> Eric Peyton wrote: > Hi Y'all, > > Got a question. I just installed the libs for NIHS (I just had > I installed before), and compiled my program MAB. It is now 10 > times larger than the origianl executable, it went from ~500K to > ~5M. What happened, will all of my executables be 10 times > larger? Is there a way that I can reduce this size?? Any help > or pointers would be appreciated. I wouldn't expect an application to get ten times larger by compiling it quad-fat. However, it could get ten times larger (easily) if you're comparing an "installed" version of a program compared to a "non-installed" version. The install target strips out unnecessary symbols in an executable, and that can trim a lot of data off. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu ITS Systems Programmer (handles NeXT-type mail) Rensselaer Polytechnic Institute; Troy NY USA
From: cph@next.com (Charles d'Harcourt) Newsgroups: comp.sys.next.programmer Subject: Re: Functionality of Perl's CGI.pm in WebObjects? Date: 10 Feb 1996 21:59:05 GMT Organization: NeXT Software, Inc. Message-ID: <4fj4f9$ht1@news.next.com> References: <4fg2id$94e@stc06.ctd.ornl.gov> On 9 Feb 1996 woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > > (Ian Patrick Cardenas) writes: > > > woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > > > > How can I save (write the name/variable pairs) to disk from > > > > WebObjects? or read them in later. Is there some WebObject > > > > construct that would provide the equivalent? > > > > > > > I don't believe there's a need to save the parameters to disk as > > > WO will maintain it's state across pages. > > The reason for saving parameters to disk is that we are building a web > application that creates a form, allows signatures to be added, and > notifies the next-to-sign that a form awaits a signature. The actual > document is retained on disk. I wouldn't want to loose the document > (they must be kept for several years) because I shut down the system or > rebooted. > > From WebObjects (just the script version) can I write to disk somehow?: You could do: id stringToWrite; is stringRead; ... stringToWrite = @"Foo"; [stringToWrite writeToFile:@"/A/B/C.txt" atomically:YES]; ... later ... stringRead = [NSString stringWithContentsOfFile:@"/A/B/C.txt"]; Or, if more structure is you cup of tea id dictionaryToWrite; is dictionaryRead; ... dictionaryToWrite = [NSMutableDictionary dictionary]; [dictionaryToWrite setObject:@"one" forKey:@"1"]; [dictionaryToWrite setObject:@"two" forKey:@"2"]; [dictionaryToWrite setObject:@"three" forKey:@"3"]; [dictionaryToWrite writeToFile:@"/A/B/C.dict" atomically:YES]; ... later ... dictionaryRead = [NSDictionary dictionaryWithContentsOfFile:@""/A/B/C.dict"]; ... There are other options for handling data in different formats. Check out http://www.next.com/Pubs/Documents/WebObjects/Foundation/Foundation.book.h tml for more info. C= __________________ Charles d'Harcourt WebObjects Group, NeXT Software, Inc. cph@next.com, http://www.next.com/~cph Obviously, I'm not speaking on behalf of NeXT.
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: 10 Feb 1996 22:05:08 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4fj4qk$ebp@emerald.oz.net> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> <4ffs9p$m5f@apakabar.cc.columbia.edu> <311C5722.76AB@arsenal.com> "C. Knight" <knight@arsenal.com> wrote: > windemut@cumbnd.bioc.columbia.edu (Andreas Windemuth) writes: > >I second that. OmniWeb 2.0 blows Netscape out of the water. > > Who are you kidding? Netscape 2.0 b6 clobbers OmniWeb. Is that the version of Netscape that crashed Windows 95 during Steve Jobs' Webmania demo, apparently because he ignored some sort of Netscape alert panel? Pretty serious penalty for ignoring an alert panel :-) OmniWeb has never crashed NEXTSTEP since I've been using it. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: apandit@acs.uswest.com (Ajay Pandit) Newsgroups: comp.sys.next.programmer Subject: Problem downloading WebObjects Date: 9 Feb 1996 00:29:30 GMT Organization: U S WEST Marketing Resources Message-ID: <4fe4ha$2v8@engineer.mrg.uswest.com> While trying to download Web server and WebObjects, I get the following: WebServer.pkg/ 755 4205 13 0 6103205510 6513 WebServer.pkg/WebServer.bom 400 4205 13 40000 6103205504 11206 ` BI allo @ p @ p p p p p p p p p @ - z _ G - & B & BOM Next Fake Inode69 $ BOM Format Version1 BOM Content Size824550 dBOM Archs Info ` @ ? p p . N ; ? . l N . ? a WebObjects.pkg/ 755 4205 13 0 6103205436 6645 WebObjects.pkg/WebObjects.bom 400 4205 13 60000 6103205423 11456 ` BI allo ` @ | @ D ` @ | Is there any problem here ? I am using OminWeb. Thanks, - Ajay
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Help connecting two modems In-Reply-To: vazquezr@physics.ucla.edu's message of 8 Feb 1996 20:20:08 GMT Message-ID: <RDL.96Feb11001841@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4fdlto$iu8@saba.info.ucla.edu> Date: Sun, 11 Feb 1996 05:18:41 GMT First the chastise: You should have posted this to comp.sys.next.sysadmin or misc. This group is for programming issues. Now for the answer: If the two computers are in the same room, you can use two serial cables and a null modem. You can purchase a null modem (which is just a little convertor that joins the two cables and flips the pins) from Radio Shack for under $10. On the software side, you should pick up a copy of kermit. This is available from ftp://ftp-next.peak.org. If you need to transfer between two computers in different locations, you just need kermit. Set one modem to auto answer on one ring with "AT SO=1" and dial that modem up with the other modem using "AT DT thephonenumber" Switch to the kermit prompt on the receiving end and type "rec" and type "send filename" on the sending end. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4fdlto$iu8@saba.info.ucla.edu> vazquezr@physics.ucla.edu (Rick Vazquez) writes: Xref: world comp.sys.next.programmer:26457 Path: world!blanket.mitre.org!sed.psrw.com!psinntp!psinntp!psinntp!howland.reston.ans.net!gatech!newsfeed.internetmci.com!info.ucla.edu!nnrp.info.ucla.edu!usenet From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Date: 8 Feb 1996 20:20:08 GMT Organization: University of California, Los Angeles Lines: 8 NNTP-Posting-Host: ts30-7.wla.ts.ucla.edu X-Newsreader: Alexandra.app (Version 0.81) Can someone please tell me how to directly connect two modems, so that I may transfer huge amoundt of data from one computer to the other. Rick UCLA Physics I think all I need is a cable and some command on each modem.
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Setting PopUp's value? In-Reply-To: dave@turbocat.snafu.de's message of 9 Feb 1996 09:37:30 GMT Message-ID: <RDL.96Feb11010529@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4ff4kr$bo@turbocat.snafu.de> Date: Sun, 11 Feb 1996 06:05:29 GMT Assuming that what you're asking is: "How do I programmatically set a PopUpList from a default?", then here's one way to do it. Note that I haven't compiled or tested this so bear with me... In the controller that has an outlet to the PopUpList cover button (popUpListCoverButton), you can add the following methods: - (MenuCell *)findPopUpMenuCellNamed:(const char *)aString { MenuCell * popUpMenuCell = nil; PopUpList * popUpList = nil; Matrix * popUpListMatrix = nil; unsigned int row = 0; unsigned int column = 0; int numberOfRows = 0; int numberOfColumns = 0; unsigned int i = 0; popUpList = [popUpListCoverButton target]; popUpListMatrix = [popUpList itemList]; [popUpListMatrix getNumRows:&numberOfRows numCols:&numberOfColumns]; row = 0; column = 0; for (i = 0; (i < [popUpListMatrix cellCount]) && (popUpMenuCell == nil); i++) { id theCell; theCell = [popUpListMatrix cellAt:row:column]; if ([theCell isEnabled] == YES) { if (strcmp([theCell title], aString) == 0) { popUpMenuCell = theCell; } } row++; if (row >= numberOfRows) { row = 0; column++; } } return popUpMenuCell; } - (void)selectPopUpMenuCellNamed:(const char *)aString { Matrix *popUpListMatrix = [[popUpListCoverButton target] itemList]; MenuCell *popUpMenuCell = nil; popUpMenuCell = [self findPopUpMenuCellNamed:aString]; if (popUpMenuCell != nil) { [popUpListMatrix selectCell:popUpMenuCell]; [popUpListCoverButton setTitleNoCopy:[[popUpListMatrix selectedCell] title]]; } } - (void)selectPopUpMenuCellTagged:(int)aTag { Matrix *popUpListMatrix = [[popUpListCoverButton target] itemList]; [popUpListMatrix selectCellWithTag:aTag]; [popUpListCoverButton setTitleNoCopy:[[popUpListMatrix selectedCell] title]]; } After you retrieve the default using NXGetDefaultValue(), you will need to call selectPopUpMenuCellNamed: with that string OR map it to an integer tag and call selectPopUpMenuCellTagged:. From the NS reference manual: If you want to change the title of a pop-up list's trigger Button, be aware that this title represents the selected item to the user, so your code will have to also change the selected Cell in the PopUpList's Matrix. It can do this either by scanning for the title, or by changing the selected Cell by position first, then getting that Cell's title to use as the title of the trigger Button. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 In article <4ff4kr$bo@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) writes: How do I set the initial Value of a PopUp? I know how to get the actual value and save it in the prefs. If the app restarts -- the wrong item will be visible. How do I change the visible Item? Can I ftp an demo-app somewhere that does this?
From: michaelv@MindBender.HeadCandy.com (Michael L. VanLoon) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: 11 Feb 1996 10:06:01 GMT Organization: HeadCandy Associates... Sweets for the lobes. Distribution: world Message-ID: <MICHAELV.96Feb11020601@MindBender.HeadCandy.com> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> <4ffs9p$m5f@apakabar.cc.columbia.edu> <311C5722.76AB@arsenal.com> <4fj4qk$ebp@emerald.oz.net> In-reply-to: art@cubicsol.com's message of 10 Feb 1996 22:05:08 GMT In article <4fj4qk$ebp@emerald.oz.net> art@cubicsol.com (Art Isbell) writes: "C. Knight" <knight@arsenal.com> wrote: > windemut@cumbnd.bioc.columbia.edu (Andreas Windemuth) writes: > >I second that. OmniWeb 2.0 blows Netscape out of the water. > Who are you kidding? Netscape 2.0 b6 clobbers OmniWeb. Is that the version of Netscape that crashed Windows 95 during Steve Jobs' Webmania demo, apparently because he ignored some sort of Netscape alert panel? Pretty serious penalty for ignoring an alert panel :-) OmniWeb has never crashed NEXTSTEP since I've been using it. I'm not a proponent of either browser, but you might note that the "b" up there means "beta"... Believe it or not, beta software has bugs in it... And, don't blame all Win95 crashes on the software running under Win95. ;-) -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Michael L. VanLoon michaelv@HeadCandy.com --< Free your mind and your machine -- NetBSD free un*x >-- NetBSD working ports: 386+PC, Mac 68k, Amiga, HP300, Sun3, Sun4, DEC PMAX (MIPS), DEC Alpha, PC532 NetBSD ports in progress: VAX, Atari 68k, others... - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
From: wongj@alum01.its.rpi.edu (Jasper Y. Wong) Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: Re: NEXTSTEP'er in Hong Kong? Date: 11 Feb 1996 08:10:08 GMT Organization: Rensselaer Polytechnic Institute Message-ID: <4fk890$as5@usenet.rpi.edu> References: <4fho5g$38u@news.asu.edu> Keywords: NEXTSTEP, Advocates, Hong Kong In article <4fho5g$38u@news.asu.edu>, <devan2m@imap2.asu.edu> wrote: >Hello All, > > First, I was wondering whether there are any NEXTSTEP/OPENSTEP >advocates in Hong Kong and who they might be? I know that there used to be a >listing for a user group there, but I don't know how current this is. > Secondly, would anybody also know about Internet access in Hong >Kong? Any and all assistance would be dearly appreciated. > >Regards- > > I'm not sure if there still exists a user group in Hong Kong or not, even though I've heard one of the newsfeed provider in HK runs his operation with NeXT stations....I can't confirm it though. Anyways, I'm now working in Hong Kong with my two Cubes....so I would like to join a local users group here as well if one exists.... Jasper -- --< NeXT >--< BeBox >--< Amiga >---< SGI >--< Sparc >--< PowerMac >-- Homepage forever under construction... IRC's nick (mizzle@#amiga) Penhold/Globe999/Friendship755 wongj@rpi.edu jyw@jasper.hko.net --< Table Tennis >--< MIDI >--< Volleyball >---| NeXTMail Welcome |--
From: thrall@serv.net (Dean Johnson) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: 11 Feb 1996 23:28:24 GMT Organization: ServNet Internet Services Message-ID: <4flu2o$q5q@brockman.serv.net> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> <4ffs9p$m5f@apakabar.cc.columbia.edu> <311C5722.76AB@arsenal.com> <4fj4qk$ebp@emerald.oz.net> <MICHAELV.96Feb11020601@MindBender.HeadCandy.com> In-Reply-To: <MICHAELV.96Feb11020601@MindBender.HeadCandy.com> On 02/10/96, Michael L. VanLoon wrote: >In article <4fj4qk$ebp@emerald.oz.net> art@cubicsol.com (Art Isbell) writes: > > "C. Knight" <knight@arsenal.com> wrote: > > windemut@cumbnd.bioc.columbia.edu (Andreas Windemuth) writes: > > > >I second that. OmniWeb 2.0 blows Netscape out of the water. > > > Who are you kidding? Netscape 2.0 b6 clobbers OmniWeb. > > Is that the version of Netscape that crashed Windows 95 during Steve > Jobs' Webmania demo, apparently because he ignored some sort of Netscape > alert panel? Pretty serious penalty for ignoring an alert panel > :-) OmniWeb has never crashed NEXTSTEP since I've been using it. The alert panel is not a Netscape problem; it is Windows NT. If you ignore an alert panel for longer than NT thinks you should, it locks up your next action and you have to reboot. I wish it was a Netscape problem, since OPENSTEP is going to be on NT. Kinda scary to me :) -- NeXTmail Please | More computer stuff! | thrall@serv.net | - My wife |
From: df@watershed.com (Dirk Fromhein) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Job Posting- Technical QA Positions Date: Sun, 11 Feb 1996 21:30:16 -400 Organization: Watershed Technologies, Inc. Message-ID: <4fm9hg$dug@caesar.ultra.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit A large financial corporation in the center of Boston is looking for QA people, both contract and full time. Extensive knowlege of SQL and Unix scripting (sh/ksh shell) required. These are high load positions in a dynamic and rapidly changing environment. Work with some of the best people in the industry. Unix, NeXTSTEP, Windows NT, and SQL experience a plus. Send me your resume, and I will pass it on. (NeXTMail, MIME, (Postscript or Word), or just ASCII) Dirk Fromhein df@watershed.com Watershed Technologies, Inc.
From: "TOMMY K. HWANG" <thwang@home.mysolution.com> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: Sun, 11 Feb 1996 22:34:38 -0500 Organization: INTERNET SOLUTIONS INC. Message-ID: <311EB54E.104F@home.mysolution.com> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> <4ffs9p$m5f@apakabar.cc.columbia.edu> <311C5722.76AB@arsenal.com> <4fj4qk$ebp@emerald.oz.net> <MICHAELV.96Feb11020601@MindBender.HeadCandy.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > > >I second that. OmniWeb 2.0 blows Netscape out of the water. > > > Who are you kidding? Netscape 2.0 b6 clobbers OmniWeb. OmniWeb 2.0b4h has problems that has been acknowledged by the company (like the bug associated with fixed frames). Also, the lack of JAVA support makes my home page look funny... http://www.mysolution.com/ The UNIX and the Win32 version of NetScape, or NetScape Gold, currently are better, IMHO, than OmniWeb. > :-) OmniWeb has never crashed NEXTSTEP since I've been using it. You are lucky. But then, it is Beta 4.0h > And, don't blame all Win95 crashes on the software running under > Win95. ;-) I only use NT (in terms of modern MS OS), so no comment here.
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 12 Feb 1996 05:15:17 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4fmid5$2m3@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: ukr8@rzstud2.rz.uni-karlsruhe.de (Jochen Quante) Newsgroups: comp.sys.next.programmer Subject: Sometimes field editor does not become first responder?! Date: 12 Feb 1996 13:27:51 GMT Organization: University of Karlsruhe, Germany Message-ID: <4fnf8n$tu@nz12.rz.uni-karlsruhe.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit NNTP-Posting-User: ukr8 Hi, Well, last time, something went completly wrong. So I am doing it again now. I have a problem concerning the TextField/TextFieldCell/Text mechanism. Usually, when the user mouses down into a TextField (or TextFieldCell?), the TextField will be made temporary the first responder and afterwards the field editor yields first responder status. When the user exits the TextField, -endEditing: is sent to the TextFieldCell and -textDidEnd:endChar: is sent to TextField. So far, so good. Sometimes, if one clicks near the border of a bezeled TextField, the TextField becomes the first responder BUT NOT THE TEXT OBJECT SERVING AS A FIELD EDITOR!!! This hides the input cursor (I-Beam) as well. Unneccessary to mention, when the user leaves the TextField, -endEditing: is not called!!! My app depends on this messages to set up specific text filters. Does anyone has an idea why: 1. this thing happens. 2. in which way I should change the field editors contents inside a text filter function? May I change text already typed in? 3. Why did NeXT implement the filtering of text via filter functions? ^^^^^^^^^ There's no nice way to figure out the current text filter within a character filter function. Any pointers would be appreciated. Thanx so much in advance. Jochen -- =============================================================== | Jochen Quante Computer Science Student | | University of Karlsruhe, Germany | | E-Mail: ukr8@rz.uni-karlsruhe.de S_Quante@ira.uka.de | | WWW: http://www.uni-karlsruhe.de/~ukr8 Jeckel@IRC | ===============================================================
From: gene@fw.bluestone.com (Dr. Gene De Lisa) Newsgroups: comp.sys.next.programmer Subject: character metrics? Date: 12 Feb 1996 10:05:38 -0500 Organization: bluestone.com Message-ID: <4fnl02$bhc@fw.bluestone.com> Keywords: metrics,font I've been trying to retrieve character metrics from a Font object (black hardware running 3.0). I keep getting a null pointer an the numCharMetrics member is always 0. I've tried different fonts etc. The font loads and other members print (displayName e.g.) The font metrics are ok. I've also tried [font metrics] besides [font readMetrics:...] Does anyone see anything wrong with this code? If I can't do this, how do I get the char's bounding box? NXFontMetrics *metrics; NXCharMetrics *cmetrics; Font * theFont = [Font newFont:"Times-Roman" size:18.0 matrix:NX_IDENTITYMATRIX] ; int j; float *bbox; printf("TEST name: %s\n", [theFont displayName] ); if( (metrics = [theFont readMetrics:NX_FONTHEADER|NX_FONTMETRICS|NX_FONTWIDTHS]) == NULL) { printf("TEST font metrics are null!\n"); return; } printf("TEST version: %s\n", metrics->formatVersion ); printf("TEST n charmetrics: %d\n", metrics->numCharMetrics ); if( (cmetrics = metrics->charMetrics) == NULL) { printf("TEST char metrics are null!\n"); return; } // the bbox for space bbox = metrics->charMetrics[ metrics->encoding[32] ].bbox; for(j=0; j< 4; j++) { printf("bb %f \n", bbox[j]); } for(j=0; j< metrics->numCharMetrics; j++) { printf("Name: %s xwidth: %f\n", cmetrics[ metrics->encoding[j] ].name, metrics->charMetrics[ metrics->encoding[j] ].xWidth ); } -- Dr. Gene De Lisa Bluestone Inc. 1000 Briggs Rd. Mt. Laurel, NJ 08054-4101 gene@bluestone.com Vox Humana: (609) 727-4600 x261 Fax: (609) 778-4925 <a href "http://www.bluestone.com">Bluestone Home Page</a>
From: trombino@wendy.ucsd.edu (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Please Help!!! Drag and Drop Date: 12 Feb 1996 18:53:01 GMT Organization: The University of California at San Diego Message-ID: <4fo2ad$b5p@news1.ucsd.edu> Hello... I've posted this question several times here, but haven't had any responses yet. I guess this is my last-ditch effort to BEG someone to help me out. I'm working on a game and I want to incorporate 3.x drag and drop protocol, but I'm having a bit of trouble. In short, the sending view is not receiving all the messages it should be, but the receiving view is receiving all its messages. The sending view gets the operationMaskForLocal message, but that's it. And, probably because of this, the drag session returns unsuccessful, even when in a lot of ways it is successful (the receiving view gets the dragged object from the pasteboard). Is there a kind soul out there who has a bit of time to look over my code? It wouldn't take long, I'm sure that I've just overlooked something... Please? Pretty please? PL-EEEEEASE???? Desparate, -- ================================================================== Mark Trombino trombino@wendy.ucsd.edu Center For Research in Computing the Arts, U.C.S.D. NEXTMail and MIME okay ==================================================================
From: rasmussn@swallow.eecs.umich.edu (Craig Rasmussen) Newsgroups: comp.sys.next.programmer Subject: DO question, [NXConnection Run]? Date: 12 Feb 1996 20:30:06 GMT Organization: University of Michigan, Ann Arbor, Michigan, USA Message-ID: <4fo80e$ceo@news.eecs.umich.edu> If you think you've seen everything, you really ought to try programming distributed objects. Every day is a new day, full of surprises. I will sometimes get the following error on the client end, [NXConnection run] - tossing received reply msg which I believe is due to the server returning a value, when none is expected by the client. Anyone care to verify this? The latest strangeness is this. The clients sends the server a message, self_of_server = [server returnSelf]; which is defined in the server protocol and just returns self on the server end. Now this often (not always) works the first time called, but not the second time. The server happily returns self, but the client fails to read the returned value and times out. Sometimes when this fails, the client end stderr gets the message [NXConnection run] - tossing received reply msg which to me indicates, actually I'm not sure what this indicates. Any suggestions? Craig Rasmussen rasmussn@umich.edu
From: john@gscorp.com (John C. Fox) Newsgroups: comp.sys.next.programmer Subject: Re: Problem installing Server provided with WebObjects Beta Date: 13 Feb 1996 01:13:54 GMT Organization: North Bay Network, Inc. news server - not responsible for content Message-ID: <4fooki$gad@miwok.nbn.com> References: <960206220607.293AAD6E.rsmallwo@canada> In <960206220607.293AAD6E.rsmallwo@canada> rsmallwo@islandnet.com wrote: > To install the Apache server provided with the WebObjects beta, I have > followed Next's instructions posted on their web site for starting the Next > installer program from a terminal session as follows: > > /NextAdmin/Installer.app/Installer &</pre> > > Executing this line gives me a 'missing name for redirect' error message. > Obviously my knowledge of the native unix side of Next is scant. Can anyone > suggest how to correct this error? > > Please respond to 'rsmallwo@islandnet.com' > > Thanks in advance. > > Rhys > Hi There: Never believe the insructions! ;-) I got caught up on this bug too. The rc.httpd file should _really_ say this: /NextLibrary/WebServer/httpd -d /NextLibrary/WebServer >>/dev/console 2>&1 Hope this helps, John http://www.gscorp.com
From: zazula@netcom.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: Please Help!!! Drag and Drop Date: 13 Feb 1996 01:54:30 GMT Organization: Running Start, Inc. Message-ID: <4for0m$92s@shellx.best.com> References: <4fo2ad$b5p@news1.ucsd.edu> In-Reply-To: <4fo2ad$b5p@news1.ucsd.edu> On 02/12/96, Mark Trombino wrote: > Hello... > > I've posted this question several times here, but haven't had > any responses yet. I guess this is my last-ditch effort to > BEG someone to help me out. > > I'm working on a game and I want to incorporate 3.x drag and > drop protocol, but I'm having a bit of trouble. In short, > the sending view is not receiving all the messages it should > be, but the receiving view is receiving all its messages. > The sending view gets the operationMaskForLocal message, but > that's it. And, probably because of this, the drag session > returns unsuccessful, even when in a lot of ways it is > successful (the receiving view gets the dragged object from > the pasteboard). > Hi - How big is the source in question? Can it be posted? There are a bunch of methods you need to implement. On the source side, there is: - draggingSourceOperationMaskForLocal: and, of course, a -mouseDown: or something to kick off the dragging. On the destination side, there is: - draggingEntered: - prepareForDragOperation: - performDragOperation: and, of course, a call to -registerForDraggedTypes:count: to enable dropping. There are some slides and accompanying code to a talk I did at the 1994 NEXTSTEP EXPO that give examples of window-to-window dragging via a class called DragDropControl (an object-well type object). I think these are in NEXTANSWERS or at the FTP sites. Good Luck! Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@netcom.com
From: prog@.coss.de (Juergen Gnoss) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.hardware,comp.sys.next.programmer Subject: disktab entry for panasonic OD Date: 13 Feb 1996 10:42:16 GMT Organization: Nacamar Data Communications Message-ID: <4fppu8$cqq@peanuts.nacamar.de> Hi all, what entry in disktab for a Panasonic LF-7300 and cardridge LM-D702W anybody now please post me or send mail to < prog@coss.de > thanks all
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <ocs@ms.mff.cuni.cz> Message-ID: <199602130936.AA08404@ns.ms.mff.cuni.cz> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Ondra Cada <ocs@ms.mff.cuni.cz> Date: Mon, 12 Feb 96 21:01:39 +0100 Subject: public/private headers in PB??? Organization: OCSoftware Hi all of you, does somebody happen to know what are the "Files/Make public" and "Files/Make private" items in the PB good for? They can be used to make headers 'public' or 'private', but I haven't noticed any difference between the former and the latter (let alone the nice small flag in the browser)... --- Ondra Cada ocs@earn.cvut.cz NeXTMAIL and MIME OK
From: rasmussn@swallow.eecs.umich.edu (Craig Rasmussen) Newsgroups: comp.sys.next.programmer Subject: Re: DO question, [NXConnection Run]? Date: 13 Feb 1996 14:53:11 GMT Organization: University of Michigan, Ann Arbor, Michigan, USA Message-ID: <4fq8kn$i3e@news.eecs.umich.edu> References: <4fo80e$ceo@news.eecs.umich.edu> In article <4fo80e$ceo@news.eecs.umich.edu>, Craig Rasmussen <rasmussn@swallow.eecs.umich.edu> wrote: >The latest strangeness is this. The clients sends the server a message, > > self_of_server = [server returnSelf]; > >which is defined in the server protocol and just returns self on the >server end. Now this often (not always) works the first time called, but >not the second time. The server happily returns self, but the client fails >to read the returned value and times out. Sometimes when this fails, the >client end stderr gets the message > > [NXConnection run] - tossing received reply msg > >which to me indicates, actually I'm not sure what this indicates. Any >suggestions? > >Craig Rasmussen >rasmussn@umich.edu > I'm just carrying on a conversation with myself but I've discovered part of problem of with the client timing out. The client is actually another server and opens a root connection itself after the first self_of_server = [server returnSelf]; is sent. It appears that the return value from [server returnSelf] comes in on the wrong connection. Very strange. I've moved to asynchronous communication which cures this particular problem. Craig Rasmussen rasmussn@umich.edu
From: rasmussn@swallow.eecs.umich.edu (Craig Rasmussen) Newsgroups: comp.sys.next.programmer Subject: Re: How to kill a DO connection gracefully? Date: 13 Feb 1996 14:58:55 GMT Organization: University of Michigan, Ann Arbor, Michigan, USA Message-ID: <4fq8vf$idc@news.eecs.umich.edu> References: <4fdo5j$mk9@news.eecs.umich.edu> <4ff2j5$bme@lion.embl-heidelberg.de> In article <4ff2j5$bme@lion.embl-heidelberg.de>, Georg Tuparev <tuparev@mailserver.EMBL-Heidelberg.DE> wrote: >In article <4fdo5j$mk9@news.eecs.umich.edu> rasmussn@swallow.eecs.umich.edu (Craig >Rasmussen) writes: >> I thought that I had this problem solved and now I'm beginning to wonder. >> Suppose I get a connection to a server: >> >> server = [NXConnection connectToName:"aName" onHost:"aHost"]; >> >> Then I'm vended an object by the server: >> >> remote = [server returnSomeObjectTo:self]; >> >> Now I do some work and want to gracefully free the server and remote objects >> and the NXConnection used to communicate with the server. >> > > >The following line then will terminate connections with existing clients: > >port_deallocate(task_self(),[[connection inPort] machPort]); > >All clients will receive senderIsInvalide: and you can happily free the connection. Is this a safe practice to kill a port out from under a connection? This is most certainly not an object orientation. It looks to me like the DO API is incomplete. > >BTW: It is really good idea to use NXAutoreleaseConnection instead of NXConnection! > I've looked all over the documentation and can't find anything on NXAutoreleaseConnection. The name alone looks promising, where do I find the documentation? >-- >Georg Tuparev >EMBL / Protein Design Phone: +49 - 6221 - 387437 >Meyerhofstr. 1 FAX: +49 - 6221 - 387517 >D-69117 Heidelberg >Germany >Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: "Timothy J. Luoma" <luomat@capitalist.princeton.edu> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.unix.bsd.netbsd.misc,comp.sys.next.advocacy Subject: Re: OmniWeb Date: Tue, 13 Feb 1996 12:14:12 -0500 Organization: Princeton University Message-ID: <Pine.NXT.3.91.960213121117.1161J-100000@capitalist.princeton.edu> References: <4e84c1$m0n@axl02it.ntc.nokia.com> <4f2h1e$jo9@pub02.va.pubnix.com> <Pine.NXT.3.92.960204125419.17765B-100000@Tomobiki-Cho.CAC.Washington.EDU> <4f9akj$2co@gaea.titan.org> <4fdr58$ojc@risc.agsm.ucla.edu> <4ffs9p$m5f@apakabar.cc.columbia.edu> <311C5722.76AB@arsenal.com> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <311C5722.76AB@arsenal.com> On Sat, 10 Feb 1996, C. Knight wrote: > windemut@cumbnd.bioc.columbia.edu (Andreas Windemuth) writes: > >I second that. OmniWeb 2.0 blows Netscape out of the water. > > Who are you kidding? Netscape 2.0 b6 clobbers OmniWeb. Oh I beg you not to start this again.... not "you" specifically, but "you" as in anyone who thinks they are going to change anyone's mind when it comes to GUI/workability of one app versus another... Can't we just agree that some people prefer one and some people prefer others? It'd save a lot of bandwidth. Anyway, if you do decide to followup on this, please remove csn-misc from the list, and keep it in *advocacy. TjL -- Timothy J. Luoma <luomat@capitalist.princeton.edu> NeXTStep Swapfile and Swapdisk FAQ: send mail with subject 'send swapfaq' Other info-via-email available, send mail with subject 'send info' For information on PPP under NeXTStep <nextppp@listproc.thoughtport.com>
From: gladish@suite.com (Brian Gladish) Newsgroups: comp.sys.next.programmer Subject: Sybase 10.0 db-library Date: 13 Feb 1996 23:35:26 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4fr77u$pgl@news.onramp.net> Is there a Sybase 10.0 db-library available for NeXT? We're developing in Objective-C for multiple platforms and would like to have a full-featured 10.0 client capability (without EOF). TIA, Brian gladish@suite.com
From: american@internetMCI.COM (Charles C. Hocker) Newsgroups: comp.sys.next.programmer Subject: mgetty patches--how do I apply them? Date: 14 Feb 1996 04:47:25 GMT Organization: InternetMCI Message-ID: <4frpgt$pdq@news.internetmci.com> Hello, I am running NSFIP and I want to upgrade getty to mgetty so I can run 57,600 on my 28.8 modem. I download mgetty-0.22 and the patches for mgetty. Now my question is--how do I apply the patches? Thanks, Charles -- ‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘ Charles C. Hocker War is Peace american@internetMCI.COM Freedom is Slavery american@aztec.asu.edu Ignorance is Strength ASCII, MIME & NeXTmail (preferred) George Orwell's 1984 ‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <ocs@ms.mff.cuni.cz> Message-ID: <199602132054.AA18246@ns.ms.mff.cuni.cz> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Ondra Cada <ocs@ms.mff.cuni.cz> Date: Tue, 13 Feb 96 20:23:35 +0100 Subject: Protocols Organization: OCSoftware Hi all, does somebody know what is the proper way of using protocols? Here's the problem - my first try looked like: @protocol AProtocol -aMethod; @end ... id <AProtocol> obj; ... [obj retain]; // COMPILER WARNING!!!!!! - 'retain' isn't in protocol Well, then I've tried this: @protocol AProtocol <NSObject> -aMethod; @end ... @interface AClass:NSObject <AProtocol> {...} @end @implementation AClass -aMethod {...} @end // COMPILER WARNING!!!!!! - the methods inherited from NSObject are "undefined" So, what is _proper_ way of using protocols with inherited classes, so as not to get a handful of warnings? If any? (Don't, please, write me "use plain id" or "cast the receiver to id" - I know and use these workarounds, but frankly don't believe they are _intended_ to be hugely used with protocols...) BTW, in case I use plain ids (which is not so bad, after all), do you know some workaround for the following IMNSHO compiler bug? @protocol One -(void)x; @end @protocol Two -(int)x; @end id q=...; [(id <One>)q x]; // WARNING!!! despite the cast: "multiple declarations..." The only - ant pretty sucking - workaround I know is to declare new variable: ... the same as above ... {id <One> qq=q; [qq x];} // no warning, but :-S Looking 4ward your answers... --- Ondra Cada ocs@earn.cvut.cz NeXTMAIL and MIME OK
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: character metrics? Date: Wed, 14 Feb 1996 09:00:05 +0000 Organization: Cedar Systems Distribution: world Message-ID: <89wN3AAVSaIxEwZw@cedar.co.uk> References: <4fnl02$bhc@fw.bluestone.com> MIME-Version: 1.0 In article <4fnl02$bhc@fw.bluestone.com>, "Dr. Gene De Lisa" <gene@fw.bluestone.com> writes >I've been trying to retrieve character metrics from a Font object >(black hardware running 3.0). > >I keep getting a null pointer an the numCharMetrics member is always 0. >I've tried different fonts etc. >The font loads and other members print (displayName e.g.) >The font metrics are ok. >I've also tried [font metrics] besides [font readMetrics:...] > >Does anyone see anything wrong with this code? If I can't do this, >how do I get the char's bounding box? > > > NXFontMetrics *metrics; > NXCharMetrics *cmetrics; > Font * theFont = [Font newFont:"Times-Roman" size:18.0 > matrix:NX_IDENTITYMATRIX] ; > int j; > float *bbox; > > printf("TEST name: %s\n", [theFont displayName] ); > if( (metrics = [theFont > readMetrics:NX_FONTHEADER|NX_FONTMETRICS|NX_FONTWIDTHS]) > == NULL) { (Rest of code removed) Gene, I think your problem here is that your flags do not include NX_FONTCHARDATA - including that in your OR should return the additional data that you require. Regards, Paul. ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: How to kill a DO connection gracefully? Date: 14 Feb 1996 14:15:28 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4fsqq0$gdu@lion.embl-heidelberg.de> References: <4fq8vf$idc@news.eecs.umich.edu> In article <4fq8vf$idc@news.eecs.umich.edu> rasmussn@swallow.eecs.umich.edu (Craig Rasmussen) writes: > >The following line then will terminate connections with existing clients: > > > >port_deallocate(task_self(),[[connection inPort] machPort]); > > > >All clients will receive senderIsInvalide: and you can happily free the connection. > > Is this a safe practice to kill a port out from under a connection? Safe - yes; elegant - no ;-) > >This > is most certainly not an object orientation. It looks to me like the DO > API is incomplete. Yes they definitely missing a disconnect method > > > >BTW: It is really good idea to use NXAutoreleaseConnection instead of NXConnection! > I've looked all over the documentation and can't find anything on > NXAutoreleaseConnection. The name alone looks promising, where do I find > the documentation? EnterpriseObjects ... UsingDistributedObjects The class does not introduce any new methods. Its just a bugfix saving a lot of trouble... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Protocols Date: 14 Feb 1996 14:20:13 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4fsr2t$ghh@lion.embl-heidelberg.de> References: <199602132054.AA18246@ns.ms.mff.cuni.cz> In article <199602132054.AA18246@ns.ms.mff.cuni.cz> writes: > Hi all, > > does somebody know what is the proper way of using protocols? > Here's the problem - my first try looked like: > > @protocol AProtocol > -aMethod; > @end > ... > id <AProtocol> obj; > ... > [obj retain]; // COMPILER WARNING!!!!!! - 'retain' isn't in protocol > use something like: id <Protocol,NSObject> obj; or even better: Aclass *obj; -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: thain@ (Doug Thain) Newsgroups: comp.sys.next.programmer Subject: How to pass app options from the command line? Date: 14 Feb 1996 20:00:48 GMT Organization: University of Minnesota, Society of Physics Students Sender: thain@calhoun.spa.umn.edu Distribution: usa Message-ID: <4ftf1g$n2n@epx.cis.umn.edu> Where can one find what options may be passed to NeXT apps from the shell? Specifically, I want to start an application hidden, much as the Workspace starts Preferences hidden. I am successfully running NS apps from /etc/rc as root, but I wish to bring up Preferences hidden and without the Workspace. Any help would be appreciated. NeXTCube, NS 3.1 Regards, -- Douglas Thain University of Minnesota thai0009@gold.tc.umn.edu http://cedar.spa.umn.edu/~thain/
From: Luis Arias <luis@club-internet.fr> Newsgroups: comp.sys.next.programmer Subject: WebObjects Warning: WOCheckbox checked attribute Date: Tue, 13 Feb 1996 14:09:09 +0100 Organization: Elysia, SARL Message-ID: <31208D75.43AD@club-internet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, I had been having some trouble getting a form with a WORepetition object containing a WOCheckbox object to work. The documentation states that the checked attribute will be set to YES if the checkbox is checked and NO otherwise. I have not found this to be true. In fact, if the checkbox is checked, the checked attribute is set to a NSCString object, if it is unchecked, it is set to a WONull object. I now test the class of the checked attribute to know if my checkbox is checked or not. -- Luis Arias Elysia, SARL 13, avenue Morane Saulnier 78140 VELIZY [33] (1) 30 70 63 42 [33] (1) 34 65 36 79 fax info@elysia.com
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Hiding Redraw Date: 15 Feb 1996 00:12:38 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4fttpm$q8m@news.onramp.net> What's the best way to avoid seeing a view being redrawn? Using a buffered window? Steve
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Hiding Redraw Date: 15 Feb 1996 00:23:04 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4ftud8$r8d@news.onramp.net> Nevermind, this did the trick; .. [window disableFlushWindow]; [self display]; [window reenableFlushWindow]; [window flushWindowIfNeeded]; .. -- Steve Dekorte NEXTSTEP/OPENSTEP consultant - Anaheim, CA mailto:dekorte@suite.com (NeXTmail, MIME welcome) http://www.marble.com/~dekorte
From: marco@sentenext1.epfl.ch (Marco Scheurer) Newsgroups: comp.sys.next.programmer Subject: Re: Protocols Date: 15 Feb 1996 10:36:21 GMT Organization: Ecole Polytechnique Federale de Lausanne Message-ID: <4fv2b5$nms@info.epfl.ch> References: <199602132054.AA18246@ns.ms.mff.cuni.cz> <4fsr2t$ghh@lion.embl-heidelberg.de> tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) wrote: > In article <199602132054.AA18246@ns.ms.mff.cuni.cz> writes: > > Hi all, > > > > does somebody know what is the proper way of using protocols? > > Here's the problem - my first try looked like: > > > > @protocol AProtocol > > -aMethod; > > @end > > ... > > id <AProtocol> obj; > > ... > > [obj retain]; // COMPILER WARNING!!!!!! - 'retain' isn't in protocol > > > use something like: > id <Protocol,NSObject> obj; > or even better: > Aclass *obj; Why better? It would, in essence, means it is better not to use protocols. I believe the reason for the confusion about these warnings is this: From Concepts/ObjectiveC/2_ObjC: "In Objective C, objects are identified by a distinct data type, id. This type is defined as a pointer to an object [...] The id type is completely nonrestrictive. By itself, it yields no information about an object, except that it is an object." In the old days, all objects were descendant from Object, so it was possible (and easy) to think that an id variable could respond to all the methods of Object. That is not the case anymore, because there is not a single inheritance root, but many: NSObject, Object, NSProxy, ... So 'id' can not imply 'inherits from NSObject'. One could argue that the compiler should not issue a warning with a declaration such as "id <AProtocol> obj;", because in that case what's wanted is the ability to check conformance to the protocol at runtime, using the conformsToProtocol method, and not compile time type checking. I don't think so. Some type checking cannot hurt. If you declare obj as NSObject <AProt> *obj; or, as Georg suggested: id <Protocol,NSObject> obj; the compiler warning will go away. --- Marco Scheurer Sen:te Parc Scientifique EPFL -- CH-1015 Lausanne -- Switzerland marco@sente.epfl.ch (NeXT, MIME mail welcome)
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: NIB:Window Inspector:Free when closed? Date: 14 Feb 1996 14:32:29 GMT Organization: Turbocat's Development, Germany Message-ID: <4fsrpu$mc@turbocat.snafu.de> Hi! When I enable the Checkbox "Free when closed" when inspecting my infopanel it only loads once. The second try crashes the app. Here is my code: #import "InfoController.h" @implementation InfoController - showInfoPanel:sender { if (myInfoPanel==NULL) { [NXApp loadNibSection:"info.nib" owner:self withNames:NO]; } [myInfoPanel makeKeyAndOrderFront:sender]; return self; } @end I think, myInfoPanel is not set to NULL after freeing. Is that true? Regards _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: Stefan Boehringer <boehrsbu@rz.ruhr-uni-bochum.de> Newsgroups: comp.sys.next.programmer Subject: Re: Sometimes field editor does not become first responder?! Date: Thu, 15 Feb 1996 17:49:07 +0100 Organization: Ruhr-Uni-Bochum Message-ID: <31236403.41C6@rz.ruhr-uni-bochum.de> References: <4fnf8n$tu@nz12.rz.uni-karlsruhe.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Jochen Quante wrote: ... > 3. Why did NeXT implement the filtering of text via filter functions? > ^^^^^^^^^ > There's no nice way to figure out the current text filter within > a character filter function. > > Any pointers would be appreciated. Thanx so much in advance. > > Jochen clue: try [[NXApp firstResponder] textFilterFunc] inside the character filter func. Daniel
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 19 Feb 1996 05:15:12 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4g9110$6nk@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: rworne@primenet.com (Robert Worne) Newsgroups: comp.sys.next.programmer Subject: How to control (replace) the mouse... Date: 17 Feb 1996 22:28:01 -0700 Organization: My Cube Sender: root@primenet.com Message-ID: <4g6dd1$3sb@nnrp1.news.primenet.com> I have resurrected several NeXT's running 3.0 and am working with them to use alternate input devices. Does anyone know a convenient way to access the mouse cursor in Workspace and use another device to control it? How to change the size/shape of the cursor? I've seen this done with other apps, unfortunately, they don't supply any source to comb through. Any suggestions would be welcome... -- Robert Worne NeXT-OS/2-MacOS Starving CS Undergrad...Sorry, I don't *do* Windows! I'd rather starve... -----BEGIN GEEK CODE BLOCK----- Version: 3.1 GCS/P/S d-?>pu s+:+> a- C++$>++++ UX++++>$ P+>+++ L+ !E !W++ N+++ !o-- !K w--- O++$ M+ V PS>--- !PE+ Y+ !PGP- t@ 5++ X+++ R- tv b+>++ DI !D G e>+++ h--- r++ y+++** ------END GEEK CODE BLOCK------
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: US - CA - NeXT Developers - to 80k + Bonus Date: 19 Feb 1996 17:31:32 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4gato4$4cs@newsbf02.news.aol.com> Location : California Position : Full - Time Salary : $50k to $80k Project : Develop major new NeXT based systems using Obj-C, EOF, Sybase. Contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
From: patricia@cco.caltech.edu (Patricia M. Schwarz) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: HELP - init exited with 2, can't boot the NS partition Date: 20 Feb 1996 03:56:16 GMT Organization: California Institute of Technology, Pasadena Message-ID: <4gbgp0$o76@gap.cco.caltech.edu> I actually had to set up the HyperTerminal on Windows 95 to post this because the NeXTstep side of my P133 is no longer starting up. It starts to boot then stops after the kmDevice is registered and gives me init exited with 2 and hangs. THIS IS THE FIRST TIME I AM NEXTLESS IN SIX YEARS :( :( :( :( Waahhhhh. I'm not ashamed to cry. Please help. :-) -a loyal community member PS I have no NeXTmail now, for obvious reasons. :-{ I'm not on any networks. I do have the Dial-Up Networking installed in the WinDoze side. -- there must be something i can dream tonight -patti smith
From: Luke Howard <lukeh@vurt.schnet.edu.au> Newsgroups: comp.sys.next.programmer Subject: Re: Problem with libposix -- fixed! Date: Tue, 20 Feb 1996 02:25:45 +1000 Organization: SchoolsNET Pty Ltd, Melbourne, Australia. Message-ID: <3128A489.275@vurt.schnet.edu.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > The version is bootp-DD2.4.3, which is bootpd-2.4.3 with DHCP patches. > I'm compiling with -posix and -traditional-cpp (to suppress the warning > about not being able to use the precompiled headers) and linking with > -lposix. All the executables built in this crash in NXCreateZone() with > a floating point error. Attached is some output from gdb. Sorry to follow up my own post. It needs to link against posixcrt0.o, not crt0.o, and that appears to be done by using -posix as a *LINKER* flag, not a compiler flag. Because the compiling/linking were been done seperately in the two programs I tried, I didn't notice. So, I now have a working (well, untested yet) DHCP server and a DCE IDL compiler under NS. Now to port the rest of the free DCE code.. :) cheers, luke
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: problems running on HPPA Date: 20 Feb 1996 06:10:48 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4gbol8$pra@news.its.com> References: <4g4j3t$hnd@netty.york.ac.uk> pete@ohm.york.ac.uk (-bat.) wrote: > can anyone tell me of significant problems in getting code to > run under HPPA 3.3 ? Sure. The stack grows the opposite way, and the argument frames and calling convention and so forth are completely different. The HP requires more stringent alignment than most other platforms, such as doubles being aligned on an 8-byte boundary. Arguments may be passed in registers and not on the stack if you're running code that's optimized (including code from the system libraries), which makes debugging difficult. GDB sort of works, but it has some important limitations and problems that also make debugging difficult. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Tue, 20 Feb 96 09:43:24 +0100 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9602200843.AA00533@flexus> Subject: never-shrinking swapfile Hello, I'm running NEXTSTEP_3.2 (yes...) on a NeXTstation, and I was examining a file by printing diagnostics about it to a terminal, which was set to retain an infinite amount of lines of text. This ran to completion, and then a file needed to be written, but by that time the swapfile had filled the disk. So I did cmd-k on the shell involved, then closed Terminal, then closed other applications, then logged out, then did exit on the login prompt (which restarts the Window Server). At some stage I expected (having read previous discussions about the exact nature of the problem) to have cleaned up that last allocated memory in the swapfile, to allow it to shrink back. But no. I logged in using the console, and there was still no disk space free. So I finally rebooted, and here I am again. To all those who are saying that I should just buy a larger disk to accommodate my swapfile and then some, I can now safely say that this is, just maybe, a bit exaggerated. I thought that perhaps with an external disk of 1 GB raw space (in addition to my internal 400-MB disk), with 200 MB available, I should be free from this problem. LET THAT SWAPFILE SHRINK BACK AGAIN!!! (Pardon my shouting. You're right, of course. NeXT knows best. What's the emoticon for irony again, if you've missed that bit?) Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be (NeXTmail), Flanders, Belgium If I don't answer: my mail relay can't handle !, % or .uucp, I think *** The year 2000 will be the last year of the 20th century. ***
From: ronald@trace.com.tw (Ronald Wiplinger) Message-ID: <cancel.728708Au4640to52158@news.trace.com.tw> Control: cancel <728708Au4640to52158@news.trace.com.tw> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <728708Au4640to52158@news.trace.com.tw> Date: 20 Feb 1996 11:21:58 GMT Cancelled by jem@xpat.com. 824815318 NEWYEAR Original Subject was: *+* Happy New Lunar Year *+*
From: otto@olcs.com (Otto Lind) Newsgroups: comp.sys.next.programmer Subject: Re: Problem with libposix while compiling DHCP server Date: 20 Feb 1996 13:25:06 GMT Organization: Otto Lind Consulting Services Message-ID: <4gci3i$2mk@olcs.olcs.com> References: <3127A446.596B@vurt.schnet.edu.au> Luke Howard (lukeh@schools.net.au) wrote: : : vurt[18:46:21]/usr/local/src/dhcp/bootp-DD2.4.3$ make : make SYSDEFS="-posix" SYSLIBS="-lposix" : cc -o bootpd bootpd.o dovend.o readfile.o hash.o dumptab.o lookup.o : getif.o hwaddr.o tzone.o report.o -lposix Use "cc -posix" instead of "-lposix". This will use posixcrt0.o in addition to the posix library (use cc -v to find out exactly what's being linked). While this may fix your particular problem, I've found the NeXT posix library so buggy that I've stopped using it for any development. Otto -- Otto Lind Otto Lind Consulting Services otto@olcs.com 4890 Ashley Lane #311, Inver Grove Hts, MN 55077 skypoint!olcs!otto voice:(612)457-1080 fax:(612)457-0761
From: marcelor@acs.bu.edu Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Fortran M (Anyone got it working on White ?) Date: 20 Feb 1996 18:24:11 GMT Organization: Boston University Message-ID: <4gd3kb$ec9@news.bu.edu> Hello, I am interested on Fortran M (http://www.mcs.anl.gov/fortran-m .) The installation instructions mention that it can be compiled for White but the makefiles and such do not seem to ( at least not the sources from the site that I got it from. ) On another vein, has anyone gotten G77 to work on that same platfom ? Also would like to hear any positive/negative comments on Absoft's FORTRAN compiler. Thanks, Marcelo
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: NEXTSTEP vs. Delphi: An Informed comparison Date: 20 Feb 1996 20:46:27 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4gdbv3$s5c@shelob.afs.com> Several weeks ago, I promised the readers on c.s.n.advocacy that I would post an informed comparison of the differences our firm has discovered between programming in NEXTSTEP/OPENSTEP and Delphi. This article is being cross-posted to c.s.n.programmer and comp.lang.pascal.delphi.misc, because I thought it might be of interest there too. Much of this article talks about techicial issues, but I do raise business issues at the end. A short paragraph of background, for those unfamiliar with AFS: We write custom trading systems for Wall Street firms. We have 13 years of total experience, 5 of which are with NEXTSTEP. We started using Delphi about 6 months ago, at the request of a customer. We have maintained essentially the same award-winning OOP design through that transition. (Yes, it truly is award-winning. We were a finalist in the OMG/ComputerWorld competition at ObjectWorld two years ago.) Without further ado, here are our observations. Feedback is welcomed. Language/runtime issues ----------------------- Delphi is based on Object Pascal, the latest iteration of almost a decade of Pascal technology from Borland. NEXTSTEP uses Objective C, which reads like traditional C with the Smalltalk '[]' messaging syntax. Like ObjC, ObjPascal offers runtime type information. One thing I like better about ObjPascal is that you can control the private/protected/public visibility of methods, not just ivars as in ObjC. In ObjPascal, you can also decide whether to dispatch methods statically (resolved by the linker, similar to function calls), virtually (resolved at runtime with VMTs, like C++), or dynamically (resolved at runtime by selector, like ObjC method calls). ObjPascal is more strongly typed than ObjC, but you can typecast all objects as necessary. The 'IS' and 'AS' operators determine class type information at runtime, plus there are numerous methods to query the RTTI information. Message passing: You can vary messages dynamically through the Windows messaging system using Dispatch(Message). That's not quite the same as passing @selectors, but it covers most cases in which you would pass selectors. You can look up method addresses dynamically in ObjPascal, but you must refer to them in conjunction with an object. Remember, the programmer decides the visibility of its methods. As a result, not all messages are resolvable at runtime, even if the methods exist deep in the class information. This is a good thing. Any method declared as 'public' and 'dynamic' can be dispatched a-la ObjC. A disadvantage of Delphi is that it doesn't do categories, but it is capable of building and using DLLs, which gives you dynamic loading. DLLs do have some limitations -- for one thing, you can't pass OUT a pointer to memory allocated within the DLL -- but those limitations exist for all Windows programs, not just Delphi. Development environment ----------------------- I like Delphi's better. There, I've said it, and I wasn't struck down by lightning. One of the two other AFS developers working in Delphi strongly agrees with this sentiment. The other finds them about equal, based on different strengths and weaknesses in each product. I find Delphi's pieces to be more strongly integrated, although it is still missing a few built-in things that NEXTSTEP provides. Form Design: Both visual designers use a drag-from-palette approach to instantiate new objects. Delphi's designer is more flexible with event handling; you're not limited to a single target/action. Delphi does not provide as much direct manipulation and in-place editing, but I like its Object Inspector better. NeXT's InterfaceBuilder is better at manipulating groups of objects visually, but its Inspector cannot set attributes in bulk when the group contains objects of different classes. The one major Delphi 1.0 designer deficiency -- that all Forms must descend directly from the TForm class -- has been corrected in the 2.0 release. Editing: Delphi's built-in editor is a bit quirky, but you get used to it. It's missing some things I like in Edit.app, like top-level compression (Cmd-0). On the other hand: it supports the full set of cursor keys, text can be selected using only the keyboard, breakpoints can be set directly in the margins, and when you click on objects on the visual form, it takes you straight to the right place in the source code. A tossup. Compiling: The Delphi compiler/linker is blazingly fast. You have to see it to believe it. You can 'Build All' in Delphi from scratch in the time it takes NEXTSTEP to rebuild one module with a normal number of include files. The 1.0 version was limited to linking Delphi object files, but 2.0 is willing to deal with conforming .obj files from other products. Debugging: NeXT's gdb is more powerful than Delphi's integrated debugger. Delphi's integration is better, but its features are weaker. Still, I have yet to run into a problem I couldn't solve in Delphi. Another deficiency: Delphi does not include a profiler. I expect to miss that when it comes time for final tuning. Components: Components and palettes are easier to build and maintain in Delphi. Delphi's VCL (Visual Component Library, equivalent to AppKit/ FoundationKit) comes with source. This has proven invaluable in learning how to design new components without hours of hit-or-miss testing. I've been harping about this deficiency in NEXTSTEP since day one, and I still think it's a mistake not to provide source. 'Nuff said. Project Manager: Project-level defaults are easier to maintain in Delphi, and inter-module dependencies are monitored automatically. However, when it comes time to make major changes (like names of things in multiple modules), Delphi does not have built-in tools to grep and diff. You can perform those tasks with other tools, but it's not as convenient. Database access --------------- Delphi ships with a powerful set of data-aware components, and it's easy to add data-aware features to 'plain' components as you need them. There is good support for tables, queries, stored procedures, and batch moves. Tables can be joined programmatically in the designer, and you can work with a live data set in design mode (even open-ended queries!). Delphi data access uses a three-layer model, with intermediate data-set objects bridging raw tables to user interface objects. Our AfsModel wrappers inherit from the intermediate DataSet objects, and that has proven an excellent place to define business rules without UI reliance. Delphi database access is generally accomplished through the well-proven BDE interface. Even the baseline version of the product comes with a copy of the Local Interbase server and the Database Desktop (for schema maintenance). For C/S projects, you can either buy a copy of the high-end product -- which includes native 'adapters' for all popular databases -- or you can use any ODBC driver in your possession. On balance, I would say that Delphi is well beyond the original NEXTSTEP DBKit, but not yet on par with EOKit. However, for anyone enterprising enough to tackle such a project, I don't see any technical limitations in Delphi that would stand in the way of building it. Other factors ------------- Delphi is essentially a tight and understandable wrapper around the Windows API. In this respect, it reminds me of the way NEXTSTEP wraps around PostScript and Mach: You can accomplish an amazing amount of useful work almost immediately, without learning too much about the more complicated layers. But when you are ready to delve deeper, the tools are there to do it. In this respect, Delphi blows the doors off all other Windows RAD tools. PowerBuilder and Visual BASIC, by comparison, just don't keep up. For deployment purposes, Delphi's approach of building self-standing applications that use the standard Windows DLLs makes it superior to OPENSTEP/Win in two key respects: no runtime royalties, and minimal RAM footprint. Delphi requires no special DLLs or extra drawing layers to distribute or charge extra money for. Personally, I think this will be OPENSTEP/Win's achilles heel. I know what NeXT wants to charge third- parties to distribute its runtime libraries with applications, and it's ludicrous. Even for MCCA developers, given the choice between no runtime royalties and $x00 per seat, OPENSTEP will be a tough sell. Especially since the Delphi development environment costs only half as much as OPENSTEP. Throw in $200 for an extra 8MB of RAM with OPENSTEP, and it all adds up. Depending on needs, one of Delphi's disadvantages is lack of support for multiple operating systems. My customers generally don't care, because at this point they are committed to variants of Win32. For developers that need to support the same source tree for both Windows and Unix (or OS/2), Delphi is not the right tool. Conclusions ----------- As skilled and knowledgeable OOP practitioners, we are extremely pleased with Delphi. We will have a native Windows trading application deployed to a very sophisticated client before OPENSTEP even ships. That application is based on the same hierarchy of objects that comprises our NEXTSTEP release, with very few design changes to accommodate differences between the products. If a customer demands OPENSTEP for Win32 deployment, we will use it. But for general purposes, OPENSTEP's runtime liabilities are too significant to ignore. Delphi seems destined to be our first choice for future work on Win32 projects. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: richard@brainstorm.co.uk. (Richard Frith-Macdonald) Newsgroups: comp.sys.next.programmer Subject: Fax server for non-next machines? Date: 21 Feb 1996 08:13:36 GMT Organization: EUnet-GB distributed news service, +44 227 266466 Message-ID: <4gek7g$3jg@bsdi201.britain.eu.net> I want to use my NeXTstep (intel) system as a fax server (send only) for the other unix systems on my network. I tried writing a small daemon to wait for requests to send (plain text) documents, put the data into a Text object, and send the fax using - [text faxPSCode:self toList:names numberList:numbers sendAt:0 wantsCover:YES wantsNotify:NO wantsHires:NO faxName:title]; This seemed to work when I ran it in a test mode (program reads from stdin rather than waiting for incoming network connections), by double-clicking on the program icon. When I ran it from the command line, it was unable to get a postscript context for imaging. I got it to work by setting my preferences to have a public postscript display. This is not a solution for when it is running as a daemon, and I am not logged in. Q1. How do I get it to have a valid postscript context when running as a daemon? Q2. How do I get my program to select different fax cover sheets? Q3. How do I set the field values to be substituted into the cover sheet? Any help greatly appreciated. Richard Frith-Macdonald
From: rking@rmode.com (Richard L. King) Newsgroups: comp.sys.next.programmer Subject: Creating boot program from assembler output Date: Wed, 21 Feb 1996 9:55:21 GMT Organization: R-Mode Corp. Message-ID: <Dn0001.1Dr@news.rmode.com> Keywords: networking, Sun 3/60, NeXT, ethernet, help I'm looking for guidance or a working example of how to use NeXT's assembler (perhaps with other programs) to produce a binary image suitable for writing to a diskette as (1) a boot program; and (2) a mini-OS. Thanks in advance for any assistance, Richard
Newsgroups: comp.sys.next.programmer From: ploeger@aplki.toppoint.de (Andreas Ploeger) Subject: Access to the window server while no one is logged in Message-ID: <1996Feb21.111105.14592@aplki.toppoint.de> Sender: ploeger@aplki.toppoint.de (Andreas Ploeger) Organization: Andreas Ploeger Date: Wed, 21 Feb 96 11:11:05 GMT Hi, I need to run an application even when nobody is logged in (it is controlled by remote messages - in case you wonder why). I could rewrite it to not use the AppKit but I would prefer not do so. There must be a way to use the WindowServer while the login panel is showing. NeXT does it for alert panels like "no paper in printer". Can you tell me how? Thanks, A. Ploeger -- Andreas Ploeger E-Mail: ploeger@tpki.toppoint.de Kiel University Phone: (49) 431 597 1757 Clinic for Pediatric Cardiology FAX: (49) 431 597 1828 Schwanenweg 20, 24105 Kiel, Germany *** NeXT Mail welcome ***
From: Robert Lutwak <robert@amo.mit.edu> Newsgroups: comp.sys.next.programmer Subject: Determining if app supports NeXTstep Help Date: 21 Feb 1996 13:11:52 GMT Organization: Massachvsetts Institvte of Technology Message-ID: <4gf5mo$5lm@senator-bedfellow.MIT.EDU> Hi. My BackSpace module (MailWatchView) supports NeXTstep help, which is quite slick with BackSpace.app. Unfortunately, Fiend.app does not support NeXTstep help, so Fiend users get an annoying alert panel the first time MailWatch tries to access the help system. Unfortunately, + newForDirectory:(const char *)helpDirectory Doesn't tell you whether or not the NXHelpPanel already existed when it was called. Has anyone out there solved this? Thanks, Robert --- Robert Lutwak robert@amo.mit.edu MIT Atomic Resonance and Spectroscopy Laboratory ---> NeXTmail always welcome <--- -- Robert Lutwak robert@amo.mit.edu MIT Atomic Resonance and Spectroscopy Laboratory ---> NeXTmail always welcome <---
From: Luke Howard <lukeh@auswired.net> Newsgroups: comp.sys.next.programmer Subject: Re: GNU gcc and Objective-C on Win95 Date: 21 Feb 1996 11:17:15 GMT Organization: AusWired, Melbourne, Australia. Message-ID: <4geuvr$pkm@news.mel.aone.net.au> References: <4f80nh$dun@entrenet.entrenet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit peckover@atlsci.com (Kevin D. Peckover, B.Sc.(M.E.)) wrote: >it has been compiled for Win NT and I have gotten the >binary form of the compiler from the ADA package with only C, and C++ >compiled in the binary. But, I would like to programm with Obj-C. There is a gnu-win32 mailing list at cygnus.com. There is gcc, bash and an assortment of GNU utilities. As far as I recall, Objective-C is included, but I did have some problems linking ObjC code under NT with it. I suspect it's possible though. cheers, luke.
From: izumi@pinoko.berkeley.edu (Izumi Ohzawa) Newsgroups: comp.sys.next.programmer Subject: Re: Access to the window server while no one is logged in Date: 21 Feb 1996 18:39:34 GMT Organization: University of California, Berkeley Distribution: world Message-ID: <4gfot6$gt9@agate.berkeley.edu> References: <1996Feb21.111105.14592@aplki.toppoint.de> In article <1996Feb21.111105.14592@aplki.toppoint.de> ploeger@aplki.toppoint.de (Andreas Ploeger) writes: >I need to run an application even when nobody is logged in (it is controlled by >remote messages - in case you wonder why). I don't know if this is still true, but... I recall that the solution is to run your application at boot time from /usr/lib/NextPrinter/lpd.children The script is run at boot time. -- Izumi Ohzawa <izumi@pinoko.berkeley.edu> [ $@Bg_78^=;(J ] USMail: Univ. of California, 360 Minor Hall, Berkeley, CA 94720-2020 Tel: 510-642-6440, Fax: -3323, Web: http://totoro.berkeley.edu/~izumi/
From: paulus@nextdown.pe.utexas.edu (Paulus Adisoemarta) Newsgroups: comp.sys.next.programmer Subject: Compiling IP-IP Encap daemon ? Date: 22 Feb 1996 23:36:26 GMT Organization: Petroleum and Geosystem Engineering, U of Texas at Austin, Austin TX Distribution: world Message-ID: <4giulq$loi@brazos.pe.utexas.edu> I need to know if somebody has already compile the IP-IP encap daemon on NS. thanks, Paulus -- Paulus Suryono Adisoemarta, N5SNN / YG1QN yono@parokinet.org n5snn@mail.utexas.edu paulus@nextdown.pe.utexas.edu
From: juberEmail=juber@aloha.net (Jack E. Uber II) Newsgroups: comp.sys.next.programmer Subject: Study of Object-Oriented Methodologies Date: Fri, 23 Feb 96 06:31:04 GMT Organization: Hawaii OnLine - Honolulu, HI Distribution: world Message-ID: <4gjn1n$mfr@nuhou.aloha.net> 20 February 1996 Greetings, I am a graduate student at Hawaii Pacific University in Honolulu, Hawaii, U.S. For my graduate thesis, I am conducting a study of object-oriented methodology and its implementation in current software projects using current techniques and theory. To this end I am submitting the accompanying questionnaire to individuals who have taken part in a software project that employed object-oriented methodologies to any extent. If you have been a part of a software development team that employed object-oriented methodologies of any kind to any extent, please take a moment of your time and answer this questionnaire. You do not need to identify yourself. The only identifying information gathered by this survey pertains to general information about your organization. The deadline for my submission of the findings of my study is 1 April 1996. Please submit your responses before this date. Please submit your response to me through Internet e-mail using the attached text file questionnaire to indicate your answers. If you would like to receive a copy of my completed study paper, please indicate so when you respond to my questionnaire. Thank you very much for taking time out to assist me. Please, only individuals who personally took part in a software project that employed object-oriented methodologies to any extent need reply. Thank you Jack E. Uber II juber@aloha.net Object-Oriented Methodology Questionnaire ----------------------------------------------------------+------------------- | The following questions are used to identify the specific | Answers characteristics of your software engineering project. | | ----------------------------------------------------------+------------------- | 1. What was the size of your project? | (Some suggested measures are: lines of code (LOC), | number of modules (NOM), number of classes (NOC), | methods per class (MPC), etc.) | | Please rate your project using a 1-to-5 scale where | 1 2 3 4 5 a "1" indicates a "small" or "trivial" project and |[ ] [ ] [ ] [ ] [ ] a "5" indicates a "very large" project. | | 2. What was the scope of your project? | (i.e. the size of the target user group supported by | your project) | | Please select the number that corresponds to the | closest match from the following choices: | (1) Personal, supporting a single person. | (2) Working Group, supporting an organizational | 1 2 3 4 5 branch from 1 to 20 people. |[ ] [ ] [ ] [ ] [ ] (3) Departmental, supporting an organizational | department from 21 to 100 people. | (4) Organizational, supporting all personnel within | a single organization. | (5) Enterprise Wide, supporting multiple | organizations. | | 3. What was the duration of your project? | (i.e. the amount of time from start to delivery) | | Please enter the number in months. | ____________ | 4. What was the actual, ending budget in U.S. dollars | for your project? | | Please enter the amount in U.S. dollars. | ____________ | 5. What was the size of your project team at the time | of your projects completion? | | Please enter the number people on your team. | ____________ | 6. What is category or type of organization you and | your team members work for? | | Please select the number that corresponds to the | closest match from the following choices: | (1) Educational or research oriented facility. | 1 2 3 4 5 (2) Private company. |[ ] [ ] [ ] [ ] [ ] (3) Government (local, state, or federal). | (4) Military. | (5) Other. (Please write your organizational | type in space provided.) | ______________________________________________ | | ----------------------------------------------------------+------------------- ----------------------------------------------------------+------------------- | The following questions ask what object-oriented | Answers methodologies (OOMs) you employed in your project. | Please indicate the level that each listed OOM was | employed in your project according to the following | criteria: | | NU - not used | SC - some concepts used | BC - basic concepts used | MC - most concepts used | EU - extensive use of the methodologys concepts | | For each methodology your project employed (those | you did not mark "NU"), please indicate the effect | the methodology had on your project using the | following criteria: | | GH - the OOM greatly hindered the development process | SH - the OOM somewhat hindered the development process | NE - the OOM had no effect on the development process | SE - the OOM somewhat enhanced the development process | GE - the OOM greatly greatly enhanced the process | | ----------------------------------------------------------+------------------- | 7. The Booch method. | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 8. Object Modeling Technique (OMT). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] 9. The Slaer/Mellor method (Object-Oriented Systems | Analysis [OOSA] and Object-Oriented Domain | Analysis [OODA]). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 10. The Jacobson method (Object-Oriented Software | Engineering [OOSE]). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 11. The Coad and Yurdon method (Object-Oriented | Analysis [OOA] and Object-Oriented Design [OOD]). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 12. Hierarchical Object-Oriented Design (HOOD). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 13. The Firesmith method (Object-Oriented Requirements | Analysis and Logical Design [OORALD]). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 14. The Cox method. | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 15. The Berard method (Object-Oriented Requirements | Analysis [OORA]). | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | 16. Other. (Please write the name in the space provided.)| ____________________________________________________ | | Extent of use? | NU SC BC MC EU |[ ] [ ] [ ] [ ] [ ] Project impact? | GH SH NE SE GE |[ ] [ ] [ ] [ ] [ ] | ----------------------------------------------------------+------------------- ----------------------------------------------------------+------------------- | The following questions ask whether your software project | Answers encountered specific types of problems. Please use the | following criteria to evaluate each statement as it | pertains to your project: | | SD - strongly disagree | D - disagree | N - no opinion | A - agree | SA - strongly agree. | | ----------------------------------------------------------+------------------- | 17. Your software project missed deadlines. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 18. Your project suffered from unmet requirements. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 19. Your projects cost exceeded its budget. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 20. Your delivered product suffered from an | unacceptable number of errors. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 21. Your delivered product suffered from a poor design | (e.g. unmaintainable code, inefficient or difficult | operation, lack of conceptual integrity, lack of | adequate comments, or poor coding style). | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 22. The code generated during the construction or coding | phase of your project was not reusable. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 23. The classes developed during your project were not | reusable. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 24. Nonobject-oriented methods were used in your project, | such as structured analysis, structured design, | procedural languages, and the like. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 25. Your software projects final design was overly | complex. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | ----------------------------------------------------------+------------------- ----------------------------------------------------------+------------------- | The following questions list some possible causes for | Answers the problems listed in the previous set of questions. | Please indicate whether any of these issues played a part | in the level of success of your software project using | the following criteria: | | SD - strongly disagree | D - disagree | N - no opinion | A - agree | SA - strongly agree | | ----------------------------------------------------------+------------------- | 26. The goals of your project were NOT clearly defined. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 27. The goals of your project were NOT reasonable. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 28. Adequate time was NOT scheduled for analysis, design, | coding, and/or testing. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 29. The project team was NOT provided reasonable | resourcing to acquire the tools or training | necessary to accomplish the project. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 30. Reasonable and meaningful metrics were NOT | established for verification and validation of | project deliverables. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 31. The software development team suffered from | interpersonal problems (e.g. bickering, rivalries, | lack of consensus, etc.). | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 32. There was a general lack of formal policy for | management of software development teams. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 33. Your software development team was too large. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 34. Your software development team was too small. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 35. There was a general lack of communication between | project leadership and development team members. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 36. Project management was too controlling. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 37. Project management did NOT provide enough control. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 38. Your project teams leaders were inexperienced with | the selected object-oriented methodologies (OOMs). | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 39. Your project teams analysts, designers, coders, | and/or testers were inexperienced with the selected | OOMs. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 40. Your project teams leaders, analysts, designers, | and/or testers received ineffective or inadequate | training on the selected OOMs. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 41. The analysis, design, coding, and/or testing tools | and methods chosen to support the selected OOMs for | your projectwere inappropriate. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 42. The OOMs selected for your project were NOT scaleable.| SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 43. The OOMs selected for your project did NOT adequately | support the analysis, design, coding, and/or testing | phase of your project. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 44. The OOMs selected for your project were limited to a | specific object-oriented programming language (OOPL). | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 45. There was NO training available for the OOMs selected | for your project. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 46. The available training for the OOMs selected for your | project was inadequate. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 47. The OOMs selected for your project were overly | complex. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 48. The OOMs selected for your project did NOT adequately | provide for smooth integration of the project phases. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 49. The OOMs selected for your project did NOT adequately | provide for smooth translation to the selected OOPL. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 50. The OOMs selected for your project were NOT | appropriate to the size of the project. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 51. The OOMs selected for your project were NOT | appropriate to the scope of the software engineering | project. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 52. The OOMs selected for your project were NOT | adequately supported by any comprehensive and | robust set of object-oriented computer-assisted | software engineering (OOCASE) tools. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 53. Structured techniques were used for the analysis, | design, and/or coding phase of your project. | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | 54. There were violations of object integrity in your | project (e.g. use of global data structures, use | of "friend" functions, direct accessing of an objects | data without using that objects methods, etc.). | SD D N A SA Please mark your answer to the right. |[ ] [ ] [ ] [ ] [ ] | ----------------------------------------------------------+------------------- Thank you for completing this questionnaire. ------------------------------------------------------------------------------
From: Steve Spicklemire <steve@estel.uindy.edu> Newsgroups: comp.sys.next.programmer Subject: clock() timing and all that..... Date: Fri, 23 Feb 1996 04:23:32 -0500 Organization: University of Indianapolis Message-ID: <312D8794.13CC@estel.uindy.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I'm trying to make sense out of values returned by clock(). There is no CLOCKS_PER_SEC defined in time.h. The numbers look like they *might* be consistent with 60 clocks per second... but I'd like to be more certain. Does anyone have a clear source that says something like: 'the correct interpretation of times returned by clock() is ____'??? thanks! -steve
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: PB: changing font of NXBrowserCells Date: 23 Feb 1996 11:12:23 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4gk7en$k1k@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Is it really impossible to change the font of a browser cell or am i missing something? I am using the standard NXBrowser with the delegate method "browser:fillMatrix:inColumn" in which i tried to "setFont" to a userFixedPitchFontOfSize:10.0. I tried several approches, with setFont applied to the prototype cell, to the matrix and to the cell. NO WAY! I get the right font in the cell, but the cell isn't resized and i can see only the upper part of the cell text, the lower being cut off. Any cues? ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <maurice@tlpnl.tlp.nl> From: Maurice Le Rutte <maurice@tlpnl.tlp.nl> Message-ID: <9602231428.AA09033@tlpnl.tlp.nl> Date: Fri, 23 Feb 96 15:24:49 +0100 Subject: EOF: Missing charset & Intel runs, Sparc crashes I am making a program that uses the Foundation and the EOAccess libraries, but it does not use the appkit. Although I am not new to the NeXTSTEP enviroment I have been able to keep my nose out of Enterprise objects as long as possible, but not anymore... :-( In my project I only need the access layer, and for the project the EOAdaptorXXXX classes are better than the EODatabaseXXXX classes. In our small office we have two stations, one Sun SparcStation 4 and a Pentium, both with NextStep 3.3 and EOF 1.1. On the Pentium my small program works just fine, doing what I thought it should be doing (that's always a suprise, isn't it?). On my sun it does not work, both using exactly the same code... My sun reports: Feb 23 14:56:11 [7830] Unable to load extended character set(s); using builtin. opening channel Feb 23 14:56:17 [7830] *** [<SybaseChannel: 0x10b968> executing SQL: SELECT t0.c_fakt_desc, t0.active, t0.dt_active, t0.nr_user, t0.dt_processing_date FROM foo t0] Feb 23 14:56:17 [7830] *** [<SybaseChannel: 0x10b968> SQL execution succeeded] Bus error The intel does not report the "Unable to load extended character set(s)..." message (it didn't report it even when my program did not crash) and does not cause a bus error. Does anybody know: - Why do I get the "Unable to load..." message, what can I do about it? - Why does the code on the Sun crash and not on the Intel? Maurice. Appendix: 1. Gdb Stack 2. Program text 3. Makefile ***** GDB STACK ***** (gdb) where #0 0x5004828 in objc_msgSend () #1 0x78240d4 in objectPrintfProc () #2 0x500e17c in NXVPrintf () #3 0x7833494 in +[NSInlineCString newWithFormat:arguments:zone:] () #4 0x78374d4 in -[NSTemporaryPlaceholderString initWithFormat:arguments:] () #5 0x7818e70 in +[NSException(NSExceptionRaisingConveniences) raise:format:arguments:] () #6 0x7818e3c in +[NSException(NSExceptionRaisingConveniences) raise:format:] () #7 0x782c100 in NSRequestConcreteImplementation () #8 0x783c7d4 in +[NSNumber allocWithZone:] () #9 0x2f2cfc in ?? () #10 0x2f70bc in ?? () #11 0x2f71fc in ?? () #12 0x2f75f0 in ?? () #13 0x2f0e4c in ?? () #14 0x3db4 in main () at test.m:49 #15 0x3990 in start (kfp=0xeffffc90) at crt0.c:161 ***** PROGRAM TEXT ***** #import <eoaccess/eoaccess.h> #import <foundation/foundation.h> #include <stdio.h> int main(void) { EOAdaptor *myAdaptor; EOAdaptorContext *myContext; EOAdaptorChannel *myChannel; EOModel *myModel; EOEntity *myEntity ; EOAttribute *myAttribute; EOQualifier *myQualifier ; NSMutableArray *attributes = [NSMutableArray arrayWithCapacity:10],*result = [NSMutableArray arrayWithCapacity:10]; BOOL ok; myModel = [ [EOModel alloc] initWithContentsOfFile:@"/Users/maurice/Test/foo.eomodel"]; myAdaptor = [EOAdaptor adaptorWithModel:myModel]; myContext = [myAdaptor createAdaptorContext]; myChannel = [myContext createAdaptorChannel]; [myChannel openChannel]; myEntity = [ myModel entityNamed:@"Foo"]; myAttribute = [myEntity attributeNamed:@"invoiceDescription"]; [attributes addObject:myAttribute]; [attributes addObject:[myEntity attributeNamed:@"isActive"]]; [attributes addObject:[myEntity attributeNamed:@"dateActive"]]; [attributes addObject:[myEntity attributeNamed:@"userId"]]; [attributes addObject:[myEntity attributeNamed:@"processingDate"]]; myQualifier = [myEntity qualifier]; [[myChannel adaptorContext] beginTransaction]; [myChannel setDebugEnabled:YES]; ok = [myChannel selectAttributes:attributes describedByQualifier:myQualifier fetchOrder:nil lock:NO]; if(ok) { id o; while((o = [myChannel fetchAttributes:attributes withZone:NULL]) != nil) [result addObject:o]; } [[myChannel adaptorContext] commitTransaction]; [myChannel closeChannel]; NSLog(@"result:%@",result); return 0; } ***** MAKEFILE ***** LIBS = -lNeXT_s -lEOAccess_s -lFoundation_s OBJ = test.o CC = cc LINK = cc CFLAGS = -c -ObjC -ggdb -Wno-precomp LDFLAGS = -all_load temple: $(OBJ) $(LINK) $(OBJ) $(LDFLAGS) $(LIBS) -o temple test.o: test.[mh] $(CC) $(CFLAGS) test.m
From: Steve Edwards - SEWP <edwards@sewp.nasa.gov> Newsgroups: comp.sys.next.programmer Subject: EOF Updating Relationships Date: Fri, 23 Feb 1996 15:06:13 GMT Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <960223100613.11473AADnF.edwards@noodle> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII I'm hoping someone might be able to shed some light on the problem of adding/removing detail items in a master-detail relationship using EOControllers. I'm having trouble coming up with a good approach. Suppose I have an Order class. Orders, in addition to having various other attributes, contain Item objects, which can be added and removed dynamically. A natural way to define such an object would be: @interface Order : NSObject { ... NSMutableArray* items; ... } ... - (BOOL)addItem:(Item*)anItem; - (void)removeItem:(Item*)anItem; ... @end The reason addItem: returns BOOL is because there is business logic associated with the adding of an item. It might fail, for example, if adding the item would exceed the amount of money allocated to the order. Now suppose I want to create a user interface for modifying the contents of an Order. EOF suggests that you define a to-many relationship from the Order entity to the Item entity and then set up master and detail EOControllers for them. This works just fine for displaying the Order and its contents, but when it comes time to add a new Item to the Order, a problem arises. To add a new item, I want to use Order's addItem: method, because it has the business logic to perform that operation. But if I do that, the detail controller won't know anything about this new Item, and when I save to the data source, the new Item won't be saved. On the other hand, if I instead insert Items directly into the detail controller, I'm bypassing the business logic which resides in the Order EO. I've tried both calling the EO's addItem: method as well as doing an insert into the detail controller, but this seems to have the effect of saving duplicate Items, and even if it did work it really negates the value of having a detail controller. I've come up with a number of solutions to this problem, and none are very pretty. My design objectives are to keep the business logic completely within the EOs, but also to have the EOs know as little as possible about EOF beyond the KeyValueCoding protocol. These objectives appear to be mutually exclusive in this case. Am I missing something basic? --- Stephen G. Edwards NASA Goddard Space Flight Center Tel: 301-286-0783 Fax: 301-286-1619 edwards@sewp.nasa.gov
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 23 Feb 1996 17:44:45 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4gkued$3q6@lion.embl-heidelberg.de> References: <4gi5cq$5vi@shelob.afs.com> In article <4gi5cq$5vi@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > Delphi will be the better choice even after OPENSTEP shows up, I think. > Once you suss out the Delphi RTTI system (which admittedly needs better > documentation, but we did it), there is almost nothing you can't do in > the way of dynamic object programming ala NEXTSTEP. Greg! This is a bit to much for a guy spending his free hours working on GNUstep! ;-) Well, I wanted to post a followup to your firs article, but now I _have_ to do it! Although I almost agree with your comparative analysis, what you told is only the half of the story. What you _forgot_ to mentioned is that NEXTSTEP is a complete OO environment. Of course this is well known to all c.s.n.* groups, but you are crossposting to other newsgroups too. But let me start with more general comments: - You are comparing Delphi from today with NeXTSTEP from the past but this is not very fair! Yes, it is true that AppKit and IB are not unique any more, but nowadays nobody claims it! The main focus now is not the AppKit, but the (P)DO. The times where we had to write huge monolithic programs belong to the past. The reality today is "distributed". And here NS/OS is light years ahead of anybody else. Just to give an example from what I'm testing now: I wrote a small molecular dynamics app witch consumes hell a lot of CPU time. Well, the main calculator object looks for the box with the lowest load (it could be in US, Japan or even on Mars ;-), and starts working there, then moves to another host, and then again, and again ... one can follow it how it migrates around the globe. When the calculation is ready, the result is waiting in my mailbox. The source for implementing this behavior is about 500 lines (including pgp security and exception handling). Now try to do this with Delphi or any other tool! Try to do it even with CORBA & NEO! I wish you good luck! And then all this nifty Foundation kit classes! You was talking about target/action. And why on earth not mention the Notification Center? What about Object Archiving, The autorelease pool, the exception handling, the Run/Event loops, the events, the Unicode strings, threads, .... well I can continue... You may claim, "well, Win has no multitasking, threads, sockets, ...", but what about D'OLE? - Now to return to the GUI classes. None of them is unique, but because they are building an entire environment the net result is something very special! Just several keywords: Services, drag and drop, pasteboard, and yes of course, FILTERS! You drag one document (say a jpeg image) from an ImageViewer and drop it in your Mail.app The filter converts it to tiff and here we are! the image is in your mail. Drag the folder from the fileviewer and drop it in a terminal, <Alt> drag the document (h-file) from Edit.app and drop it in IB ... and Oh wonder! , IB parses the file! Every single bit of NEXTSTEP is the money worth! It makes your life much easier. What you are talking about are details, but take NEXTSTEP as an entire developer environment and compare it again! - The NEXTSTEP is uniform. Thsi is something very, very, very, important, but you forgot to mentioned it. This uniformity has one huge implication to _all_ developers: one can write very easily reusable general purpose classes or even entire frameworks, and these frameworks can work together without any problem! Have you ever seen something like the MiscKit for any other developer environment? Only NEXTSTEP makes it possible for over 30 people from I don't know how many countries to work together so nicely! Wake up Greg! very nextish... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: heberlei@cs.ucdavis.edu (Louis Todd Heberlein) Newsgroups: comp.sys.next.programmer Subject: Re: Simple PS question: string width Date: Tue, 20 Feb 1996 10:17:04 -0800 Organization: UC Davis, Computer Science Dept. Message-ID: <heberlei-2002961017040001@dissertation.cs.ucdavis.edu> References: <4g3nen$4ng@mimsy.cs.umd.edu> <DMxrvI.4uJ@basil.icce.rug.nl> > Check out: > -[Font getWidthOf:(const char *)string] > -(NXFontMetrics *)[Font metrics] > > Here's a snippet of code that demonstrates its use: > > Font *f; > NXSize size; > char *string; > ... > // Determine rectangle occupied by annotation. > { > NXFontMetrics *metrics = [f metrics]; > > size.width = [f getWidthOf:string]; > size.height = metrics->ascender - metrics->descender + 1;//fudge > } Ahh. The "fudge" factor. I've run into this problem a couple of times. In particular, I've tried doing my own optimization with scrolling the bits around a terminal screen instead of doing a complete redisplay. The problem appears to be that the font sizes (the width in particular) come out to fractional sizes (e.g., 7.2), which causes problems over time. To solve this problem, I am using courier font at either 10 or 15 point, both of which generate nice whole numbers for the width. This allows the optimizations to remain consistent with what the user sees with a true drawSelf call (which happens when I create an EPS file of the terminal window). I realize this is not a very clear posting, but I am hoping that someone who has run into similar problems will recognize what I am trying to say. Anybody have any suggestions? Thanks, Todd PS. The program really works well at 10 and 15 point fonts, but I would prefer to give the user a greater selection of font sizes.
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 23 Feb 1996 19:18:52 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <4gl3us$alj@shelob.afs.com> References: <4gkued$3q6@lion.embl-heidelberg.de> Georg Tuparev writes > Wake up Greg! Oh, I am quite awake, thank you. NEXTSTEP _was_ a complete OO environment (including a really good unified user environment), but OPENSTEP will not be. In fact, many of the complaints I have heard about OPENSTEP/Win from a functional point of view is how much of that wonderful NEXTSTEP we take for granted has been "dumbed down" for the sake of being a "first class NT citizen." In other words, NEXTSTEP loses some of its magic on the way to OPENSTEP/NT. Are you sure about those Services? I've heard otherwise. You described a lot of cool stuff that we currently take for granted which I'm not sure will be included in OPENSTEP/NT. I share your rapture for that environment, but it's on its way out. Mail.app? NeXT has already stated they won't be shipping it. Furthermore, NEXTSTEP is not uniform once it becomes OPENSTEP. It will use the traditional UI in the Mach version, but Windows UI in the NT version. I think Pete Clark posted that you will have a choice between traditional and Motif in the Solaris version. And I hate to burst your bubble, but I have extremely good information that the Sun and NeXT versions _already_ differ in some important ways. GNUstep will differ from both of them. Even if you maintain 100% fidelity to the spec, Sun and NeXT will not. Face it, OPENSTEP will never be as source-compatible and MAB-capable as NEXTSTEP already is. As to PDO, I have never found a "must have" use for it in my applications. You may prefer to focus there instead of AppKit, but PDO is NOT where the bulk of the corporate applications market is. Yes, I know about options and derivatives firms who need to run massive Monte Carlo simulations on Alpha chips, but you'll have to work awfully hard to convince me that straightforward forms-based apps -- the bread and butter corporate apps -- benefit from PDO. NeXT made this argument to the customer who is paying us to move from NEXTSTEP to Delphi, and the customer didn't buy it. Besides, if you really need PDO/D'OLE, my understanding is that it will not be limited to OPENSTEP apps. I hope not, or it will find a limited market. In which case, there is no reason a Delphi app couldn't leverage it too. Georg, it doesn't sound like you have actually worked much on Delphi or even Windows 95/NT. For example, the integration of Delphi's editor and form designer goes WAY beyond IB's limited parsing of header files. The statement that "Only NEXTSTEP makes it possible for over 30 people from I don't know how many countries to work together so nicely [on MiscKit]!" is patently wrong. More is possible in other environments than you think. Before you spend much more time on your GNUstep labor of love, you should take a step back, draw a deep breath, survey the market, and ask who you're building this for and how much it is worth the effort. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: kboyes@noc.tor.hookup.net (Kevin Boyes) Newsgroups: comp.sys.next.programmer Subject: sccs programs for next Date: 23 Feb 1996 16:01:35 -0500 Organization: HookUp Communication Corporation, Oakville, Ontario, CANADA Message-ID: <4gl9vf$npt@noc.tor.hookup.net> NNTP-Posting-User: kboyes Is sccs available for NeXT? I found the front end, sccs proper, from a 4.4BSD distribution but the supporting proograms (admin, get, etc.) were not there. Free programs would be best but I am open to for-pay solutions. Thanks Kevin -- Kevin Boyes kboyes@multipath.com Multipath Business Systems Inc. (416)865-9757x239
From: MaRK_BeSSeY@next.com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: clock() timing and all that..... Date: 24 Feb 1996 00:08:39 GMT Organization: NeXT Software, Inc. Message-ID: <4glku7$7s2@news.next.com> References: <312D8794.13CC@estel.uindy.edu> Steve Spicklemire <steve@estel.uindy.edu> writes > Hi, > > I'm trying to make sense out of values returned by > clock(). There is no CLOCKS_PER_SEC defined in time.h. The > numbers look like they *might* be consistent with 60 clocks > per second... but I'd like to be more certain. Does anyone > have a clear source that says something like: 'the correct > interpretation of times returned by clock() is ____'??? > > thanks! > -steve Well, the *source code* for libc implies that you might want to use CLK_TCK, which is #defined in time.h Yes, there really should be a CLOCKS_PER_SEC value #defined in time.h, in order to be ANSI-compliant. Hope this helps, -Mark -- Mark Bessey NeXT Software, Inc Software Quality Assurance -->I DON'T SPEAK FOR NeXT <--
From: stanj@cs.stanford.edu (Stan Jirman) Newsgroups: comp.sys.next.programmer Subject: Filter Services for directories? Date: 24 Feb 1996 05:25:01 GMT Organization: Stanford University Message-ID: <4gm7fd$k34@nntp.Stanford.EDU> Hello, I just tried to write a filter service to filter an image document type that is a directory (like rtfd) to the native TIFF format. After quite a while I noticed that the service does not get called at all. I put in several debugging points and realized that if I take a plain file and rename it so that it matches the file type, the service gets called; if I call it to open the "real" thing, it is not even fired up. Can anyone help me? Anyone successfully filtered e.g. rtfd? Thanks a lot! - Stan --- +-------------------------------------------+-------------------------+ | Stan Jirman -- The Swiss CS Guy | | | stanj@cs.stanford.edu NeXTmail / MIME | When you find yourself | | http://www-leland.stanford.edu/~stanj/ | in a hole, stop digging | | Box 2642, Stanford, CA 94309, USA | | +-------------------------------------------+-------------------------+
Date: 24 Feb 1996 11:17:23 GMT From: left@www4.clever.net (EXPERT SPORTS) Newsgroups: comp.sys.next.programmer Message-ID: <cancel.4gmj0o$19a5@hearst.cac.psu.edu> Control: cancel <4gmj0o$19a5@hearst.cac.psu.edu> Subject: cmsg cancel <4gmj0o$19a5@hearst.cac.psu.edu> EMP/ECP (aka SPAM) cancelled by jem@xpat.com. See news.admin.net-abuse.announce, report 19960224.04 for further details
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.software Subject: WODocumentation Date: 24 Feb 1996 19:01:00 GMT Organization: Academic Computing Services, Sheffield University Message-ID: <4gnn9c$4b9@bignews.shef.ac.uk> Some people seem to be getting a bit, umm, frustrated, with trying to retrieve the WebObjects documentation over the Web... I've bundled up as much of it as I can (whilst still remaining sane!) in rtf and rtfd format and put it in: ftp://ftp.dcs.shef.ac.uk/home/malc/NeXT/WebObjects/WODocumentation.compressed (note that this is an automount point, so may not be accessible through a browser -- if an access attempt fails, please try again using command-line!) Documents include: Getting Started DynamicElements A Foundation for WebScript Programmers Using WebScript Maybe it might be a good idea if NeXT could create an official version of this...?! :-) Best wishes, mmalc. posn. research facilitator where institute for language speech and hearing sheffield university west court 2 mappin street sheffield s1 4dt england vox (+44) 114 282 5269 fax (+44) 114 278 0972 email m.crawford@dcs.shef.ac.uk NeXTMail, SunMail, MIME welcome PGP key available on request http://www.dcs.shef.ac.uk/research/ilash/ --
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 24 Feb 1996 19:13:54 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4gno1i$jvu@news.xmission.com> References: <4gkued$3q6@lion.embl-heidelberg.de> <4gl3us$alj@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > [...] The statement that "Only NEXTSTEP makes it possible for over > 30 people from I don't know how many countries to work together so > nicely [on MiscKit]!" is patently wrong. More is possible in other > environments than you think. As the administrator of the MiscKit, I ought to at least state that I agree with Greg's statement here. The is no technical reason that I am aware of that would rule out a Java-based MiscKit, or even a Delphi MiscKit. That said, the NEXTSTEP _community_ is what bred the MiscKit, which is more a political/social thing than anything to do with the NEXTSTEP environment per se. I don't know of anything like the MiscKit for Java or Delphi, but it could happen someday if it hasn't already. Perhaps enough NEXTSTEP developers will cross over that the MiscKit will make the same transition. Heck, if some NEXTSTEP-based operation (or anyone for that matter) decided that they wanted a Java or Delphi MiscKit and wanted to offer me a research grant to work on it, I'd cross over right now. I've actually discussed the idea with a few folks... :-) For those not in the NeXT groups who may be wondering what this MiscKit thing is, it is basically a huge library of Objective-C objects--foundation classes, GUI widgets, and all kinds of wrappers around UNIX functionality. More information on the kit is available from http://www.misckit.com/ including listings of exactly what the kit contains and a more in-depth description of license, charter, etc. Feel free to check it out if you're curious. (Warning: this web site is slow since it is still in experimental stages and therefore located on one of my home machines across a 14.4k link to the Internet.) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: Dominik Westner <dominik@gowest.ppp.informatik.uni-muenchen.de> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.software Subject: Re: WODocumentation Date: 26 Feb 1996 11:16:51 GMT Organization: TeDoc, Munich, Germany Message-ID: <4gs4r3$8e@cube.ppp.informatik.uni-muenchen.de> References: <4gnn9c$4b9@bignews.shef.ac.uk> mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > Some people seem to be getting a bit, umm, frustrated, with trying to > retrieve the WebObjects documentation over the Web... > > I've bundled up as much of it as I can (whilst still remaining sane!) in rtf > and rtfd format and put it in: > > ftp://ftp.dcs.shef.ac.uk/home/malc/NeXT/WebObjects/WODocumentation.co mpressed > > (note that this is an automount point, so may not be accessible through a > browser -- if an access attempt fails, please try again using command-line!) > > Documents include: > Getting Started > DynamicElements > A Foundation for WebScript Programmers > Using WebScript > > Maybe it might be a good idea if NeXT could create an official version of > this...?! :-) They do ... maybe you should try: http://www.next.com:80/Pubs/Documents/WebObjects/Beta1/WODoc.v2.tar.Z so long -- Dominik --- Dominik Westner "Everything is possible in an infinite universe" _____________________________________Willi, the painter Munich, Germany (NOT :-)
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc From: Fabien_Roy@free.fdn.org Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Content-Type: text/plain; charset=iso-8859-1 Message-ID: <1996Feb26.182506.12657@free.fdn.org> Sender: news@free.fdn.org Content-Transfer-Encoding: 8bit Organization: Fabien Roy Consultant, Paris, France References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4grtuf$h78@info.epfl.ch> Mime-Version: 1.0 Date: Mon, 26 Feb 1996 18:25:06 GMT FYI. I just received a promotional offer from BORLAND France for Delphi. This offer is only for Delphi 1.0 customer (Which I am not! :-) ) Here are the price: Delphi Desktop 2.0: 1200 FF or $234, reg. price: 2405 FF or $481 Delphi Developper 2.0: 1561 FF or 312, reg. price: 3612 FF or $722 Delphi Client/Server 2.0 which i think could be compared to NS+EOF 18900 FF or $3780 (For this one I don‚t have the regular price and will post it later). Delphi Client/Server 2.0 also includes Interbase NT 2 users licence and ReportSmith 3.0. ReportSmith 3.0 might be of some interest, but not Interbase NT. Conclusion: The discounted price Delphi Client/Server 2.0 is not so far from NS+NSDev+EOF, but I think that I get more for my money with the NEXSTEP stuff. Cheers Fabien (Prices quoted above include french sales taxes (20,6%) 8^( ) -- --------------------------------------------------------------------- Fabien_Roy@free.fdn.org (NextMail/MIME accepted) Fabien Roy Consultant NEXTSTEP/OPENSTEP/EOF Consultant, SYBASE DBA 10 rue de la DEFENSE 93100 MONTREUIL, France Tel: 33 1 45 28 32 23 Fax: 33 1 48 55 09 90
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 26 Feb 1996 23:41:55 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4gtgg3$m4q@shelob.afs.com> References: <1996Feb26.182506.12657@free.fdn.org> Fabien_Roy@free.fdn.org writes > [French pricing information deleted] In the US, the price differential between Delphi and NS is considerable. Remember, if you don't need Borland's custom C/S drivers, the basic "developer" package is sufficient for use with local or ODBC drivers. Even the C/S version is only $1995. Compare that to NS, which costs $795 for the User version, plus $4,995 for the Developer edition, plus $495 for EOF, plus... And your description completely ignores runtime fees. NS has them. Delphi doesn't. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 19:57:57 -0500 Organization: Steel Driving Software, Inc. Message-ID: <31325715.70F8@ix.netcom.com> References: <4gkued$3q6@lion.embl-heidelberg.de> <4gl3us$alj@shelob.afs.com> <3132418e.12598508@news.bogo.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Karl Harbour wrote: > > Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > <snip> > > > >As to PDO, I have never found a "must have" use for it in my applications. > >You may prefer to focus there instead of AppKit, but PDO is NOT where the > >bulk of the corporate applications market is. Yes, I know about options > >and derivatives firms who need to run massive Monte Carlo simulations on > >Alpha chips, but you'll have to work awfully hard to convince me that > >straightforward forms-based apps -- the bread and butter corporate apps -- > But I thought your argument was based on a nearly-award-winning OO > app which you translated remarkably easily from Next to Delphi, not > some straightforward triviality? Straightforward != trivial Straightforward means 'minimal chrome' not 'minimal complexity'. Greg's design was lauded not because it had every whiz-bang feature under the sun, but because it's a solid, finely crafted design which has shown real benefits. > >benefit from PDO. NeXT made this argument to the customer who is paying us > >to move from NEXTSTEP to Delphi, and the customer didn't buy it. Besides, > >if you really need PDO/D'OLE, my understanding is that it will not be > >limited to OPENSTEP apps. I hope not, or it will find a limited market. > >In which case, there is no reason a Delphi app couldn't leverage it too. > > > Surely, the point with (P)DO is that you hardly know you're doing it, That's all well and good. But most applications simply don't need it. Many applications of DO are little more than prettied-up RPC calls. For this, DDE would work as well. And, it doesn't cost $500/seat to deploy. > >Georg, it doesn't sound like you have actually worked much on Delphi or > >even Windows 95/NT. For example, the integration of Delphi's editor and > >form designer goes WAY beyond IB's limited parsing of header files. > I would think IB doesn't need to go further than header file parsing, > due to its object archiving nature, a subject which I know has been > discussed here before. (IB does need improving, but not in this > direction.) Alas, all the objects you use in IB are not palettized. In that case, you do have to parse headers. Unfortunately, IB doesn't notice if you change the headers in the editor.
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 20:08:51 -0500 Organization: Steel Driving Software, Inc. Message-ID: <313259A3.5433@ix.netcom.com> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4gi8ei$ekv@cheyenne.iac.net> <31324223.12747679@news.bogo.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Karl Harbour wrote: > > jon@heinlein (Jonathan Hendry) wrote: > > > >Everyone who's written an app where all the functionality is implemented > >in 'AppController.m' or such, raise your hand. > Maybe so, but at least it would be a controller object, and not a > subclass of Window or View. All this example shows is that novices put > too much functionality in one class, but at least a NextStep novice > separates Controller logic from View logic right from the word go, and > they don't do bizarre things like subclassing Window because it > contains some Controls that the designer of the Window class didn't > anticipate. NextStep novices don't create a Controller because of anything inherent in NeXTSTEP. They generally do it because that's what Garfinkel & Mahoney did, or what was in NeXTDeveloper/Examples. Actually, the Form-reliance makes some sense considering that Windows is based on Windows, and not Applications. If I'm not mistaken, if you want to get a handle to another process in Windows, you have to go through the on-screen windows. Yeah, it sucks. But that's Bill Gates for ya. > > > >NeXTSTEP certainly doesn't enforce good design. > But it does encourage it, and that's what matters. Oh, yeah. I forgot about the NeXT OOA/D Wizard that raps you on the knuckles if you access an ivar without going through the accessor. > Something "which in many cases is adequate" is hardly encouraging good > OO design. A person who hasn't a clue about OOD will botch just as well on NeXTSTEP as on Delphi. Good OOD doesn't come from your tools.
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 20:23:19 -0500 Organization: Steel Driving Software, Inc. Message-ID: <31325D07.5DED@ix.netcom.com> References: <4gl3us$alj@shelob.afs.com> <4gq6ok$9hl@lion.embl-heidelberg.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Georg Tuparev wrote: > Now I'm really lost! I was expecting to hear such an argument form a newcomer, but > it surprise me to get it from you! Each (even small) company needs tools like: > distributed time/appointment scheduler; documentation/ordering/accounting systems; > database ... just to mention few. By using PDO you can develop such applications an > order of magnitude faster! I do not start talking about large corporate > environments, like Hospital management system ... there the benefits of using PDO > are gigantic! Georg, NeXT is moving to the Windows world. Most of this software already exists - there's no need to write it. Sure, PDO is nice. If the benefits (and more importantly, cost-benefits) of using PDO were that gigantic, it wouldn't be the DO bit-player that it is. > You know better then I that the NS community is not very large. So one could expect > that for larger communities one can find much more universally applicable > frameworks. But this is just not true! Every now and then I have to work on a X11 + > C++ project. There are no two different class frameworks (written by different > software vendors) with the size of MiscKit which are capable to work together. Think > about it! I even played with the Talignet ... no wonder that they bit the dust! > Believe it or not, NS/OS was and still is the only complete development environment. > If I found a better one I will perhaps change the train, but there is nothing else > out! C++ wasn't exactly made for easy exchange of classes. Bad example. :) Delphi was made for that. The Free/Shareware Delphi component 'market' is booming. New components are coming out faster than NeXTSTEP palettes *ever* did, even when NeXT had an academic presence. They just plug right into the Delphi palette, and 'just work'. Check http://sunsite.icm.edu.pl/archive/delphi
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 20:35:59 -0500 Organization: Steel Driving Software, Inc. Message-ID: <31325FFF.D68@ix.netcom.com> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4grtuf$h78@info.epfl.ch> <1996Feb26.182506.12657@free.fdn.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Fabien_Roy@free.fdn.org wrote: > Conclusion: The discounted price Delphi Client/Server 2.0 is not so far from > NS+NSDev+EOF, but I think that I get more for my money with the NEXSTEP > stuff. That's not quite right. For one, you can do database development without Delphi C/S. Second, a large part of the cost of Delphi C/S is due to database driver licenses which allow you to redistribute the drivers with your application. It's more akin to NSDev+EOF *with no runtime fee*. Since the EOF runtime is $500/seat it doesn't take many deployment seats to get a major price advantage from Delphi.
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 27 Feb 1996 16:10:19 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4gvadb$op2@shelob.afs.com> References: <31323e5e.11782536@news.bogo.co.uk> Karl Harbour writes > Obviously I haven't seen your AfsController class, but I find it hard > to accept that you can achieve 100% MVC with it. I accept that it > separates out the business logic from the UI, but I think there is > more to MVC than that, especially considering your claim of complete > purity. I need to be more careful with my wording; everyone takes things too literally in this group. By "100% MVC," I mean that we use MVC as a fundamental design principal in all of our work. I do NOT mean that we adhere to slavish execution of a particular object organizational model. Pardon my smugness, but in a methodology as new as OOP, I do not accept that Booch's or Rumbaugh's or anyone else's ideas are necessarily better than mine. My point was that we DO divide our apps into Models, Views, and Controllers in a way that works for us. I would never claim that we are "pure" to anyone else's ideas. > Surely it's a requirement of MVC that the model objects ARE objects. > A TDataSet operates on a SET of rows in a table. How is it possible to > subclass something which represents a set, and yet have instances of > it represent a single model object instance? A TDataSet CAN operate on a set, it doesn't have to. What if it's hooked up to a TQuery? You're thinking if TTable mode. In any event, this is one of the compromises we made, given budget and delivery schedules. Yes, we could have built AfsModel as a direct subclass of TObject, but why? We found an existing class that seemed appropriate and started from there. That's what REUSE is all about. > In any case, how can it be good OO design for a Currency class, say, > to inherit methods such as > > procedure CursorPosChanged; > procedure Delete; > procedure DisableControls; > procedure Edit; > procedure EnableControls; In what universe do models not have to be deleted? Or prepare for editing? And just because Enable|DisableControls methods exist in the superclass, it doesn't mean we actually use them. > If my model classes have non-persistent attributes, how is this > modelled in your scheme? Calculated fields are no good, as I may want > the attribute to be settable by another object through public API. How > do I do this with a TDataSet subclass, because clearly any instance > variables apply to a set of model objects, not just one. Even if a TDataSet happens to contain multiple records at a given time, there is only one "cursor" position. The attributes of that record are expressed through a series of TField-derived objects, which may or may not be hooked up to UI. You only operate on one record/object at a time. Each has access to all properties declared in the Model subclass. > Also, I'd be interested in how you solved the classic OO to > relational-database mapping problem? When you've got a Model class > hierarchy, do subclasses have their own table, go in their parent > class's table, or spread across both? If this hasn't been an issue for > you, is it perhaps because your model is computationally complex but > structurally fairly simple? Actually, this problem solves itself with multiple tables. An AfsModel subclass defines the properties it needs. Those properties are then attached to whatever TFields from whatever TTables/TQuerys are needed to provide the necessary data. The model literally does not care how the parent/child aspect of database tables are organized structurally. However, it does support a parent/child concept (analogous to visible parents and children) for more complex situations where multiple MODELs need to get a particular job done, like cross-model validation. > Does your framework provide an object-uniqueing facility like EOF > does? If you've got an OO API, surely it's important that object > identity is preserved in this way? No, but the OOPness of the database is not what matters in our app. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 27 Feb 1996 16:42:32 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4gvc9o$osh@shelob.afs.com> References: <31324223.12747679@news.bogo.co.uk> Karl Harbour writes > jon@heinlein (Jonathan Hendry) wrote: > > Everyone who's written an app where all the functionality is > > implemented in 'AppController.m' or such, raise your hand. > Maybe so, but at least it would be a controller object, and not a > subclass of Window or View. All this example shows is that novices > put too much functionality in one class, but at least a NextStep > novice separates Controller logic from View logic right from the > word go, and they don't do bizarre things like subclassing Window > because it contains some Controls that the designer of the Window > class didn't anticipate. The first iteration of our now-properly-organized TradeKit was based entirely on Window subclasses. It turned out to be a sub-optimal design, but it didn't suck. Just don't try to tell me that NS naturally forces people to use controllers. We paid for that mistake for several months, when we finally decided to make it right. > > NeXTSTEP certainly doesn't enforce good design. > But it does encourage it, and that's what matters. Perhaps you would like to cite an example of how that oh-so-famous two-layer database access library, DBKit, separated Controller logic and encouraged novices not to do bizarre things. I'm not beating on NeXT, I'm just pointing out that at a time when most NeXT programmers were cutting their teeth on database access, NeXT's solution was poor from several perspectives, notably overall design. I am proud to say that AFS stuck with its homegrown solution through all that, and has never regretted it for a moment. In fact, when Felix visited with his DBKit II (now EOF) slide show, my first comment was "That looks great. Would you like to buy it right now?" Delphi's database access layer, mercifully, began its life as a 3-layer hierarchy, which is why we were able to interpose our AfsModel class between the raw tables and the UI expression of the data. Delphi newbies will never know how much pain they have been spared by that decision. The bottom line is, only clueful managers and skilled programmers can encourage and support good design. Unfortunately, both are in dreadfully short supply. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 22:58:23 GMT Organization: Orbital Computer Consultancy Message-ID: <31338a40.12358521@news.plsys.co.uk> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4gi8ei$ekv@cheyenne.iac.net> <31324223.12747679@news.bogo.co.uk> <313259A3.5433@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> wrote: >Karl Harbour wrote: >> >> jon@heinlein (Jonathan Hendry) wrote: >> > >> >Everyone who's written an app where all the functionality is implemented >> >in 'AppController.m' or such, raise your hand. > >> Maybe so, but at least it would be a controller object, and not a >> subclass of Window or View. All this example shows is that novices put >> too much functionality in one class, but at least a NextStep novice >> separates Controller logic from View logic right from the word go, and >> they don't do bizarre things like subclassing Window because it >> contains some Controls that the designer of the Window class didn't >> anticipate. > >NextStep novices don't create a Controller because of anything inherent in >NeXTSTEP. They generally do it because that's what Garfinkel & Mahoney did, >or what was in NeXTDeveloper/Examples. > And why did Garfinkel & Mahoney and the people who wrote the NeXTDeveloper/Examples do it that way? Because of target/action and the design of the AppKit generally separating UI from application logic. I'm sorry, but your argument is completely bogus. I suppose you don't think the AppKit provides any help to the developer for writing consistent apps in terms of the look and feel of cut & paste, spell checking, fonts, colours, printPSCode, faxPSCode and so forth? If two apps handle these things in the same way, is it down to the application developer or what's provided by Next's AppKit? >Actually, the Form-reliance makes some sense considering that Windows is >based on Windows, and not Applications. If I'm not mistaken, if you want to >get a handle to another process in Windows, you have to go through the >on-screen windows. Yeah, it sucks. But that's Bill Gates for ya. > >> > >> >NeXTSTEP certainly doesn't enforce good design. >> But it does encourage it, and that's what matters. > >Oh, yeah. I forgot about the NeXT OOA/D Wizard that raps you on the knuckles >if you access an ivar without going through the accessor. > >> Something "which in many cases is adequate" is hardly encouraging good >> OO design. > >A person who hasn't a clue about OOD will botch just as well on NeXTSTEP >as on Delphi. Good OOD doesn't come from your tools. I beg to differ. If a newbie follows the Delphi User's Guide, or any of the numerous books now on the market, they're certainly going to end up with a sub-optimal design. If they had you or Greg as a tutor, of course they'd do better, but that's not the point. A newbie who follows the Next Developer documentation (or reads G&M) is much more likely to develop a reasonably well-partitioned app.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 22:58:10 GMT Organization: Orbital Computer Consultancy Message-ID: <313388c9.11983625@news.plsys.co.uk> References: <31323e5e.11782536@news.bogo.co.uk> <4gvadb$op2@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >Karl Harbour writes >> Obviously I haven't seen your AfsController class, but I find it hard >> to accept that you can achieve 100% MVC with it. I accept that it >> separates out the business logic from the UI, but I think there is >> more to MVC than that, especially considering your claim of complete >> purity. > >I need to be more careful with my wording; everyone takes things too >literally in this group. By "100% MVC," I mean that we use MVC as a >fundamental design principal in all of our work. I do NOT mean that >we adhere to slavish execution of a particular object organizational >model. Pardon my smugness, but in a methodology as new as OOP, I do >not accept that Booch's or Rumbaugh's or anyone else's ideas are >necessarily better than mine. My point was that we DO divide our apps >into Models, Views, and Controllers in a way that works for us. I >would never claim that we are "pure" to anyone else's ideas. > Fine, you've got your own definition of MVC. I believe there's a commonly accepted definition of MVC, in which a reasonable definition of a model object is an object that has an analogue in the business problem domain, such as Currency, Customer, Account etc. When comparing different development tools, it surely makes sense to use the commonly accepted definition of terms, not personal definitions. Your Delphi app seems to have no model objects as such by what I believe is the common definition. >> Surely it's a requirement of MVC that the model objects ARE objects. >> A TDataSet operates on a SET of rows in a table. How is it possible to >> subclass something which represents a set, and yet have instances of >> it represent a single model object instance? > >A TDataSet CAN operate on a set, it doesn't have to. What if it's hooked >up to a TQuery? You're thinking if TTable mode. Surely it's TDataSources that are hooked up to TDataSets? TQuery is a wrapper round a SQL query is it not? I was under the impression SQL SELECT statements returned a _set_ of results. Perhaps you thought I meant TTable when I said "SET of rows in a table" - I meant table in the database. TTable and TQuery both work on sets. >In any event, this is one >of the compromises we made, given budget and delivery schedules. Yes, we >could have built AfsModel as a direct subclass of TObject, but why? We >found an existing class that seemed appropriate and started from there. >That's what REUSE is all about. > >> In any case, how can it be good OO design for a Currency class, say, >> to inherit methods such as >> >> procedure CursorPosChanged; >> procedure Delete; >> procedure DisableControls; >> procedure Edit; >> procedure EnableControls; > >In what universe do models not have to be deleted? Or prepare for editing? I agree CursorPosChanged, DisableControls and EnableControls are the most bogus. I wanted to show these 3 methods whilst keeping authenticity, so I cut and pasted the whole block from DB.INT >And just because Enable|DisableControls methods exist in the superclass, >it doesn't mean we actually use them. > >> If my model classes have non-persistent attributes, how is this >> modelled in your scheme? Calculated fields are no good, as I may want >> the attribute to be settable by another object through public API. How >> do I do this with a TDataSet subclass, because clearly any instance >> variables apply to a set of model objects, not just one. > >Even if a TDataSet happens to contain multiple records at a given time, >there is only one "cursor" position. The attributes of that record are >expressed through a series of TField-derived objects, which may or may >not be hooked up to UI. You only operate on one record/object at a time. I think you have to concede that you are working with database records, and not model objects. This one record at a time, may / may not be connected to UI aspect seems to introduce a nasty coupling into your framework. For example, if one part of the program wants to access the nth record in the dataset, it could ruin the cursor position for another part of the program also accessing the dataset, and also undesirably cause the UI to display another record. Okay, you could solve the first problem by having separate datasets, but this could be very inefficient and gives you a UI synchronisation problem. You can disableControls to solve the second problem, but this only proves that the bad coupling is there. These problems simply don't exist with EOF. >Each has access to all properties declared in the Model subclass. > >> Also, I'd be interested in how you solved the classic OO to >> relational-database mapping problem? When you've got a Model class >> hierarchy, do subclasses have their own table, go in their parent >> class's table, or spread across both? If this hasn't been an issue for >> you, is it perhaps because your model is computationally complex but >> structurally fairly simple? > >Actually, this problem solves itself with multiple tables. An AfsModel >subclass defines the properties it needs. Those properties are then >attached to whatever TFields from whatever TTables/TQuerys are needed >to provide the necessary data. The model literally does not care how >the parent/child aspect of database tables are organized structurally. >However, it does support a parent/child concept (analogous to visible >parents and children) for more complex situations where multiple MODELs >need to get a particular job done, like cross-model validation. > As far as I can see, you didn't solve the problem, you avoided it. Instead of model objects, you have an object which manages a set of database records. As far as I'm concerned, this is not really OO business programming, and it's one reason why NextStep/EOF is fundamentally superior to Delphi. >> Does your framework provide an object-uniqueing facility like EOF >> does? If you've got an OO API, surely it's important that object >> identity is preserved in this way? > >No, but the OOPness of the database is not what matters in our app. Surely it's desirable to have an API where the objects you create in your design, which are derived from the business domain analysis, can be treated as real objects, for example by passing them around as arguments? Using a TDataSet subclass and expecting the current cursor position to transiently define an object for you does not cut the mustard for me.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 22:58:20 GMT Organization: Orbital Computer Consultancy Message-ID: <313389d7.12254284@news.plsys.co.uk> References: <31324223.12747679@news.bogo.co.uk> <4gvc9o$osh@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >Karl Harbour writes >> jon@heinlein (Jonathan Hendry) wrote: >> > Everyone who's written an app where all the functionality is >> > implemented in 'AppController.m' or such, raise your hand. > >> Maybe so, but at least it would be a controller object, and not a >> subclass of Window or View. All this example shows is that novices >> put too much functionality in one class, but at least a NextStep >> novice separates Controller logic from View logic right from the >> word go, and they don't do bizarre things like subclassing Window >> because it contains some Controls that the designer of the Window >> class didn't anticipate. > >The first iteration of our now-properly-organized TradeKit was based >entirely on Window subclasses. It turned out to be a sub-optimal design, >but it didn't suck. Just don't try to tell me that NS naturally forces >people to use controllers. We paid for that mistake for several months, >when we finally decided to make it right. > >> > NeXTSTEP certainly doesn't enforce good design. > >> But it does encourage it, and that's what matters. > >Perhaps you would like to cite an example of how that oh-so-famous >two-layer database access library, DBKit, separated Controller logic >and encouraged novices not to do bizarre things. Well, this is an unexpected change in the direction of the debate. It seems a rather weak argument to criticise a kit which has been defunct for over 18 months. What relevance does DBKit have to Nextstep and EOF now? How well did Turbo Pascal 5.0 separate Controller logic? >I'm not beating on NeXT, Not much! >I'm just pointing out that at a time when most NeXT programmers >were cutting their teeth on database access, NeXT's solution was poor >from several perspectives, notably overall design. I am proud to say >that AFS stuck with its homegrown solution through all that, and has >never regretted it for a moment. In fact, when Felix visited with his >DBKit II (now EOF) slide show, my first comment was "That looks great. >Would you like to buy it right now?" > Next's past is irrelevant as far as I'm concerned. >Delphi's database access layer, mercifully, began its life as a 3-layer >hierarchy, which is why we were able to interpose our AfsModel class >between the raw tables and the UI expression of the data. Delphi newbies >will never know how much pain they have been spared by that decision. > No Next newbie needs to go through any of this pain either. DBKit has been completely superceded (some time ago) by EOF. EOF allows you to do real business objects, and also, for example, provides an edit undo mechanism - how does Delphi support undo again? >The bottom line is, only clueful managers and skilled programmers can >encourage and support good design. Unfortunately, both are in dreadfully >short supply. >-- I think the quality of the tools is relevant as well. (Clueful?)
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 22:58:28 GMT Organization: Orbital Computer Consultancy Message-ID: <3133840e.10772188@news.plsys.co.uk> References: <1996Feb26.182506.12657@free.fdn.org> <4gtgg3$m4q@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >Fabien_Roy@free.fdn.org writes >> [French pricing information deleted] > >In the US, the price differential between Delphi and NS is considerable. >Remember, if you don't need Borland's custom C/S drivers, the basic >"developer" package is sufficient for use with local or ODBC drivers. >Even the C/S version is only $1995. > >Compare that to NS, which costs $795 for the User version, plus $4,995 >for the Developer edition, plus $495 for EOF, plus... And your description >completely ignores runtime fees. NS has them. Delphi doesn't. > >-- >Gregory H. Anderson | "Honey, there're few programming >Gaffer/Best Boy/Key Grip | problems that can't be solved >Anderson Financial Systems | with duct tape." -- 'Father' Duke >greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95 For once, can we not make a comparison on purely technical grounds? There's enough Next pricing policy bashing as it is.
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.lang.pascal.delphi.misc,comp.sys.next.advocacy Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 19:14:43 -0500 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <olAttnq00iV9IAr4M2@andrew.cmu.edu> References: <1996Feb26.182506.12657@free.fdn.org> <4gtgg3$m4q@shelob.afs.com> <3133840e.10772188@news.plsys.co.uk> In-Reply-To: <3133840e.10772188@news.plsys.co.uk> Excerpts from netnews.comp.sys.next.advocacy: 27-Feb-96 Re: NEXTSTEP vs. Delphi: An.. by Karl Harbour@mail.bogo.c > For once, can we not make a comparison on purely technical grounds? > There's enough Next pricing policy bashing as it is. Greg Anderson did perform a rather detailed comparision on purely technical grounds, without ever mentioning price tags. The article was posted on 2/20/96 with a message-ID of <4gdbv3$s5c@shelob.afs.com>. I didn't see any "bashing" of NeXT's pricing policies in Greg's comments, which I'll quote in full at the end of this message. For that matter, I can't think of any other comments in this thread that I'd call "bashing" either, although certainly several people have been trying to figure out the economics of NeXT's pricing. Perhaps Karl is fortunate enough to have the luxury to make decisions without considering financial matters, but that isn't something that I'm free do to. The results I've gotten lead me to criticise NeXT's proposed runtime licensing fees because they appear to make it unviable for an ISV to develop many products using OPENSTEP, at least those which do not sell for thousands of dollars per seat. -Chuck >>In the US, the price differential between Delphi and NS is considerable. >>Remember, if you don't need Borland's custom C/S drivers, the basic >>"developer" package is sufficient for use with local or ODBC drivers. >>Even the C/S version is only $1995. >> >>Compare that to NS, which costs $795 for the User version, plus $4,995 >>for the Developer edition, plus $495 for EOF, plus... And your description >>completely ignores runtime fees. NS has them. Delphi doesn't. >> >>-- >>Gregory H. Anderson | "Honey, there're few programming >>Gaffer/Best Boy/Key Grip | problems that can't be solved >>Anderson Financial Systems | with duct tape." -- 'Father' Duke >>greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95 Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: Joe Freeman <Joe@FreemanSoft.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 22:23:46 -0500 Organization: FreemanSoft Inc. Message-ID: <3133CAC2.7ED4@FreemanSoft.com> References: <31324223.12747679@news.bogo.co.uk> <4gvc9o$osh@shelob.afs.com> <313389d7.12254284@news.plsys.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit It appears that I must agree with Gregg on this one!? Karl Harbour wrote: > > >Perhaps you would like to cite an example of how that oh-so-famous > >two-layer database access library, DBKit, separated Controller logic > >and encouraged novices not to do bizarre things. > > Well, this is an unexpected change in the direction of the debate. It > seems a rather weak argument to criticise a kit which has been defunct > for over 18 months. What relevance does DBKit have to Nextstep and EOF > now? How well did Turbo Pascal 5.0 separate Controller logic? OK, we can look at some other kits and see the same lack of separation. The Appkit has a very poor separation between the application and UI control logic. Applications which follow the 3.x appkit generally have a very poor partition between the program control and the UI control layer. EOF 1.x has similar issues where the best "Controller" class is the EOController class which has user interface features rolled into it also. (EOF 2.x should fix this with the breakup of the controller class) 3.x DO is a great product except that it has no split between the DO protocol and the underlying transport layer. The controllers responsible for user object managment and notifications are the same objects which go down to the wire. (Hopefully the new DO will fix this). > Next's past is irrelevant as far as I'm concerned. Only if the progress is across the board. Progress in one group, EOF, does not translate to understanding across all projects. > No Next newbie needs to go through any of this pain either. DBKit has > been completely superceded (some time ago) by EOF. EOF allows you to > do real business objects, and also, for example, provides an edit undo > mechanism - how does Delphi support undo again? The current, EOF 1.x, undo architecture is nice except that it is only really supported in the UI layer. I dont see anything UI centric about object attribute editing rollbacks. -- FreemanSoft Inc. NEXTSTEP and WEBOBJECTS software in the DC area. Governer Glendenning of MD wants to spend $300M of the people's money on a rent free pro football stadium which is only used 10 times a year. For whom is he really doing this deal?
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 28 Feb 1996 16:16:21 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4h1v4l$lb@shelob.afs.com> References: <313388c9.11983625@news.plsys.co.uk> Karl Harbour writes >[reasonable stuff about the use of TDataSet vs TDataSource] > Surely it's TDataSources that are hooked up to TDataSets? > I think you have to concede that you are working with database > records, and not model objects. This one record at a time, may / may > not be connected to UI aspect seems to introduce a nasty coupling into > your framework. Yikes. I need to wait until I'm fully awake before posting. I think this whole argument could have been avoided if I hadn't typed 'TDataSet' where I meant 'TDataSource'. I've been talking about the three-layer DB access hierarchy, and how we use the middle part, but when it came to naming it, I used the wrong one. AfsModel derives from Source, NOT Set. And yes, our models do fully encapsulate the business rules and logic. Sorry for the confusion. I think we're actually on the same page. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 28 Feb 1996 16:23:58 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4h1viu$m1@shelob.afs.com> References: <313389d7.12254284@news.plsys.co.uk> Karl Harbour writes > Greg Anderson wrote: > >Perhaps you would like to cite an example of how that oh-so-famous > >two-layer database access library, DBKit, separated Controller logic > >and encouraged novices not to do bizarre things. > > Well, this is an unexpected change in the direction of the debate. It > seems a rather weak argument to criticise a kit which has been defunct > for over 18 months. What relevance does DBKit have to Nextstep and EOF > now? How well did Turbo Pascal 5.0 separate Controller logic? As I said, I raised this point only to demonstrate that even NeXT, whose programmers I respect enormously, botched this first time out, and a lot of early programmers either learned a bad thing or spent too much time working around the deficiencies. Delphi at least got the DB part right, but I agree they botched the Form-dependence part. I guess I'm saying, cut Borland some slack. Delphi is (was) a 1.0 tool, and a better 1.0 tool than NS was when it was 1.0. They already fixed some of the worst TForm buggery. And don't forget, some of this window-centric stuff comes from the Win API. Messaging is intimately tied to visual objects. > Next's past is irrelevant as far as I'm concerned. To you, perhaps. But it affected a lot of programmers. I know, I have the DBKit-workaround class library to prove it. And it's ugly. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: sumner@tiac.net (Sumner Paine) Newsgroups: comp.sys.mac.misc,comp.sys.mac.system,comp.sys.mac.wanted,comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: Steve Jobs' email address? Date: 26 Feb 1996 02:47:56 GMT Organization: The Internet Access Company Message-ID: <sumner-2502962151550001@sumner.tiac.net> Does anyone know Steve Jobs' email address? (I want to quote him in a research paper). If you know, please email me at sumner@tiac.net. Thanks, Sumner
From: teddy@hubert.fukt.hk-r.se (Teddy Hogeborn) Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: How to update the resolver in NS 3.0? Followup-To: comp.sys.next.programmer Date: 29 Feb 1996 05:49:04 +0100 Organization: FUKT Computer Society Sender: teddy@hubert.fukt.hk-r.se Message-ID: <kkhgwa4tz3.fsf@hubert.fukt.hk-r.se> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Keywords: bind, shared libraries, resolver I use some features in our local nameserver setup that requires resolvers to understand CNAMEs in reverse lookups. I have therefore compiled the newest version of bind (4.9.3-REL-Patch1), and it works fine. But naturally the system binaries don't use the libresolv.a, so *how* do I fix it? I think I that all of /lib/libsys_p.a, /lib/libsys_s.a and /usr/shlib/libsys_s.B.shlib have the resolver routines in them, (NS 3.0, black hw) but I don't dare to muck with a shared library libc for fear of creating a forever unbootable system. So how do I replace the resolver routines with the new versions? What part of the online manual should I read? I found nothing simple in it about creating shared libraries. Please reply by email, and I'll summarize, as I don't read this group. /Teddy
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 27 Feb 1996 15:52:07 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4gv9b7$ont@shelob.afs.com> References: <rfenney-260296204937@rfenney.slip.netcom.com> Robert J. Fenney writes > Just because you have not had the need for PDO does not mean that the > corporate environment is not going in that direction! When you start > talking about sales force automation, distributed processing centers and > branch office automation you are talking about a PDO functionallity. Please don't overstate what I said. I have nothing against [P]DO; in fact, I pointed out that the WriteUp API relies on it. I do understand that some enterprise apps need [P]DO functionality, I just pointed out that my company's fairly sophisticated Wall Street MCCA does NOT. As a result, I do not accept this is a must-have, make-or-break feature. Georg seemed to be saying NEXTSTEP couldn't take a leak without it. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 25 Feb 1996 17:37:24 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4gq6ok$9hl@lion.embl-heidelberg.de> References: <4gl3us$alj@shelob.afs.com> In article <4gl3us$alj@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > Georg Tuparev writes > > Wake up Greg! > [...] > Furthermore, NEXTSTEP is not uniform once it becomes OPENSTEP. It will use > the traditional UI in the Mach version, but Windows UI in the NT version. > I think Pete Clark posted that you will have a choice between traditional > and Motif in the Solaris version. And I hate to burst your bubble, but I > have extremely good information that the Sun and NeXT versions _already_ > differ in some important ways. GNUstep will differ from both of them. Even > if you maintain 100% fidelity to the spec, Sun and NeXT will not. Face it, > OPENSTEP will never be as source-compatible and MAB-capable as NEXTSTEP > already is. Again this is not very true. OpenStep specifies stuff like services, pasteboard, communication to the workspace, user default database etc. What it is not specified is if the window close button is on the left or the right side, if the title bar is blue or black, which mouse button is doing what , etc.pp. This is IMHO very wise decision, because having the same source and even the same nib files, you can get different look & feels. Of course stuff like drag & drop will work only between OpenStep apps, but if I have Tailor.app, Stone Design, Lighthouse, and AFS ;-) apps on top of W$ I will be completely happy and never will think on using any other Win crap! > As to PDO, I have never found a "must have" use for it in my applications. > You may prefer to focus there instead of AppKit, but PDO is NOT where the > bulk of the corporate applications market is. Oh really? And how your apps are talking to the workspace manager? How is your license server pinged? How could you implement pasteboard daemon? How to implement a API in a heterogeneous environment? And further: everybody is talking about client/server systems, accessing remote databases, EOF, WOF,... I'm really very curious how you can implement it without PDO. Well, you can use sockets on UNIX! Have you used them? It's great fun indeed ;-) And what about W$? Don't tell me OLE please! I prefer to go to work in a sheep farm but not using OLE! > Yes, I know about options > and derivatives firms who need to run massive Monte Carlo simulations on > Alpha chips, but you'll have to work awfully hard to convince me that > straightforward forms-based apps -- the bread and butter corporate apps -- > benefit from PDO. Now I'm really lost! I was expecting to hear such an argument form a newcomer, but it surprise me to get it from you! Each (even small) company needs tools like: distributed time/appointment scheduler; documentation/ordering/accounting systems; database ... just to mention few. By using PDO you can develop such applications an order of magnitude faster! I do not start talking about large corporate environments, like Hospital management system ... there the benefits of using PDO are gigantic! > Georg, it doesn't sound like you have actually worked much on Delphi or > even Windows 95/NT. For example, the integration of Delphi's editor and > form designer goes WAY beyond IB's limited parsing of header files. Yes, it's true that I'm not a Delphi guru, and yes it is true that "the integration of Delphi's editor and form designer goes WAY beyond IB's limited parsing of header files", but I'm talking about overall integration! What if I do not like the Delphi's editor? With NS I can use any! With Delphi I have only one option... > The > statement that "Only NEXTSTEP makes it possible for over 30 people from > I don't know how many countries to work together so nicely [on MiscKit]!" > is patently wrong. More is possible in other environments than you think. You know better then I that the NS community is not very large. So one could expect that for larger communities one can find much more universally applicable frameworks. But this is just not true! Every now and then I have to work on a X11 + C++ project. There are no two different class frameworks (written by different software vendors) with the size of MiscKit which are capable to work together. Think about it! I even played with the Talignet ... no wonder that they bit the dust! Believe it or not, NS/OS was and still is the only complete development environment. If I found a better one I will perhaps change the train, but there is nothing else out! > Before you spend much more time on your GNUstep labor of love, you should > take a step back, draw a deep breath, survey the market, and ask who > you're building this for and how much it is worth the effort. Or perhaps you should spend a bit more time with GNUstep ;-) It will work on virtually any platform. Not to mentioned that some of the most intelligent and talented folks have gravitated to GNUstep. Amen ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: charlie@why.az.stratus.com (Charles_Spitzer) Newsgroups: comp.sys.next.programmer Subject: ObjectError Date: 28 Feb 1996 16:26:39 GMT Organization: Stratus Computer Inc, Marlboro MA Message-ID: <4h1vnv$85c@transfer.stratus.com> Anyone have this rewritten as a poseAs: for NSObject? I tried doing so, but found that NSObject doesn't have Object's +error: or -error methods. Can one just eliminate these methods? -- Charles Spitzer charlie@az.stratus.com Stratus Computer, Inc.
From: wpmoyne@xingu.mit.edu (William Moyne) Newsgroups: comp.sys.next.programmer Subject: NeXT display on SGI Date: 28 Feb 1996 17:05:45 GMT Organization: Massachvsetts Institvte of Technology Message-ID: <4h2219$1li@senator-bedfellow.MIT.EDU> Is there a way to display NeXT Step programs on an SGI? I know they both use display postscript. I would like to remotely log into my NeXT box (intel 3.3) and send the display of a program to my sgi. Is this possible?? Please reply via Email --William -- -------- Teaching Assistant by day, Research Assistant by night MSEECS 2-'95 Massachusetts Institute of Technology Bld 36, Rm 279 e-mail: wpmoyne@mit.edu WWW: http://www-mtl.mit.edu/~wpmoyne/
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: The Enema Within Date: 26 Feb 1996 16:57:31 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4gsopr$l9n@shelob.afs.com> References: <4grhhd$1s1@prometheus.localnet.com> levman@localnet.com writes: [blah blah blah] Geez, people call _me_ long-winded! I think it was Will Rogers who once said, "The problem with that man is, he has a diarrhea of words, but a constipation of ideas." That seemed an appropriate commentary for a post titled "The Enema Within." I hope it's not long-term exposure to NeXT machines that causes this condition, or we're all in deep trouble... -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: ngervae@sirius.com Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 27 Feb 1996 18:33:51 GMT Organization: NeXT Software, Inc. Message-ID: <4gviqf$dh@news.next.com> References: <4guqt5$5nd@ixnews3.ix.netcom.com> In article <4guqt5$5nd@ixnews3.ix.netcom.com> far@ix.netcom.net(Felipe A. Rodriguez) writes: > I've been trying to implement a lightweight database engine using > the Indexing Kit's Query Language with little success. > > Unfortunately when I try to use the Query Language's relational > operators (i.e. eq, gt) things do not work out. The selector within > my records for the registered attribute is simply not called. > > I've heard there are bugs in the Indexing Kit, is this one of them? Indeed. About half of the documented features in the query language were never really implemented. The engineer assured me they would be, so I documented them as if they were (silly, naive new employee that I was). I thought this got mentioned in a release note at some point, but it's been a while and I haven't even looked at IXKit since I wrote the doc for 3.0. I really must recommend that you avoid the Indexing Kit. It was a decent idea, but the engineer had more work than he could do, and it hasn't been properly supported since it was released. It isn't going to evolve, improve, or be fixed, so best keep away.... -- Nik Gervae <ngervae@next.com> Technical Writer NeXT Software, Inc.
From: far@ix.netcom.net(Felipe A. Rodriguez) Newsgroups: comp.sys.next.programmer Subject: Indexing Kit Query Language Date: 27 Feb 1996 11:45:41 GMT Organization: Netcom Message-ID: <4guqt5$5nd@ixnews3.ix.netcom.com> I've been trying to implement a lightweight database engine using the Indexing Kit's Query Language with little success. I started out with the RecordManager example and with a few mods can use the Query Language's Search operators (i.e. whole, prefix) to find strings based on a registered attribute. Unfortunately when I try to use the Query Language's relational operators (i.e. eq, gt) things do not work out. The selector within my records for the registered attribute is simply not called. Does anyone know of example code which uses the Indexing Kit Query Language? I've heard there are bugs in the Indexing Kit, is this one of them? Before I go and implement things in the C/UNIX way does anyone have any ideas I might try? -- Felipe A. Rodriguez # ...it cannot be called ingenuity to Agoura Hills, CA # kill one's fellow citizens, to betray # friends, to be without faith, without far@ix.netcom.com # mercy, without religion; by these means (NeXTmail preferred) # one can acquire power but not glory. (MIMEmail welcome) # --Nicolo Machiavelli
Newsgroups: comp.sys.next.programmer From: rfenney@netcom.com (Robert J. Fenney) Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Message-ID: <rfenney-260296204937@rfenney.slip.netcom.com> Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Sender: netnews@mork.netcom.com Organization: FenTek References: <4gkued$3q6@lion.embl-heidelberg.de> <4gl3us$alj@shelob.afs.com> Date: Tue, 27 Feb 1996 04:51:59 GMT In article <4gl3us$alj@shelob.afs.com>, Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >> Stuff Cut << Greg, Just because you have not had the need for PDO does not mean that the corporate environment is not going in that direction! When you start talking about sales force automation, distributed processing centers and branch office automation you are talking about a PDO functionallity. My company has spent the last 8 years building these kinds of applications and they are just plain hard without something like PDO. PDO like most of NeXTSTEP is a least 5 years a head of the market. The problem is that Next just is not able to capitolize to a greater extent on it's advantage and the rest of the market catches up. But like WebObjects they just keep moving ahead. Robert
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 26 Feb 1996 20:43:56 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <4gt62c$lph@shelob.afs.com> References: <4gq6ok$9hl@lion.embl-heidelberg.de> [I took Delphi off the distribution, since this isn't really about my original comparison anymore] Georg Tuparev writes > In article <4gl3us$alj@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) > > I have extremely good information that the Sun and NeXT versions > > _already_ differ in some important ways. GNUstep will differ from both > > of them. Even if you maintain 100% fidelity to the spec, Sun and NeXT > > will not. Face it, OPENSTEP will never be as source-compatible and > > MAB-capable as NEXTSTEP already is. > Again this is not very true. OpenStep specifies stuff like services, > pasteboard, communication to the workspace, user default database etc. > What it is not specified is if the window close button is on the left or > the right side, if the title bar is blue or black, which mouse button is > doing what , etc. I should have been more clear about what I meant. I meant the PROGRAMMING API is starting to diverge, not the look-and-feel. To quote from one of my sources who is familiar with the work Sun is doing: "[DPS on top of X is] causing no end of problems with incompatible imaging models, incompatible input models, and incompatible everything else. Not to mention performance. OpenStep on Solaris on DPS on X makes my old Mac-IIci look like a Ferrari. Nexts' OpenStep and Suns' OpenStep are already incompatible [with respect to] ... the new text subsystem ... NSTableView ... locale related stuff ... help (Sun wants to keep using AnswerBook format, can't do RTF) ... and debugger (instead of gdb, they hacked dbx to make it understand Objective-C; performance is miserable)." > Of course stuff like drag & drop will work only between OpenStep apps, > but if I have Tailor.app, Stone Design, Lighthouse, and AFS ;-) apps > on top of W$ I will be completely happy and never will think on using > any other Win crap! Assuming NeXT makes it possible for ISVs to rewrite and distribute their apps profitably under OpenStep. The current discussion about OS/Win runtime pricing in c.s.n.advocacy does not give much hope of that occurring. Then what will you do? Run "Win crap", exactly. > > As to PDO, I have never found a "must have" use for it in my > > applications. You may prefer to focus there instead of AppKit, > > but PDO is NOT where the bulk of the corporate applications market is. > Oh really? And how your apps are talking to the workspace manager? How > is your license server pinged? How could you implement pasteboard > daemon? How to implement a API in a heterogeneous environment? I don't know, and I don't care. As an application programmer, the implementation of these features is irrelevant. I DO know that Win95 doesn't seem to have any trouble launching apps by double-clicking files, and there's a workable pasteboard. Do I care if it's DO? No. > And further: everybody is talking about client/server systems, accessing > remote databases, EOF, WOF,... I'm really very curious how you can > implement it without PDO. Well, you can use sockets on UNIX! C/S and remote database provide their own client libraries, and I use them. As an application programmer, I don't know or care how they're implemented. But I can understand why _you_ do, as a systems programmer. > And what about W$? Don't tell me OLE please! I prefer to go to work in > a sheep farm but not using OLE! God gives you lemons, you make lemonade, I guess. This is where the world is moving. Be sure to send me a sweater when you get where you're going. > Each (even small) company needs tools like: distributed time/appointment > scheduler; documentation/ordering/accounting systems; database ... just > to mention few. By using PDO you can develop such applications an order > of magnitude faster! Most people don't develop these apps, they buy them. They don't care how they're programmed, as long as they do the job. > What if I do not like the Delphi's editor? With NS I can use any! With > Delphi I have only one option... Did NeXT give you the keys to the car? Last I heard, the magic PB/gdb interface to Edit.app was not formally or publically documented. It's not in my help system, at any rate. Delphi 2.0, OTOH, further expands upon the open tools interface that began in 1.0. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: fox@jeans.fokus.gmd.de (Oliver Fox) Newsgroups: comp.sys.next.programmer Subject: need help to chnage sybase sizes Date: 28 Feb 1996 19:46:35 GMT Organization: GMD-FOKUS Message-ID: <4h2ber$d14@stern.fokus.gmd.de> Keywords: sybase size problem Hi need help with my sybase database when I installed the sybase database i agree to the size of 16 Mb - now it is to small how I can change the overall size ? I new about ALTER DATABASE but with this I only can make some of my training Databases bigger till the summ of them all reaches 16Mb and than -all its over OK here are the two questions: How I can make the overall size for the database bigger? How I can reduce the size of a example database? thank you for help Oliver Tel.: +49 30 25 499 123 Fax: +49 30 25 499 202 E-Mail: fox@fokus.gmd.de (NeXT mail welcome)
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 29 Feb 1996 10:47:11 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4h407f$gpe@lion.embl-heidelberg.de> References: <4h1viu$m1@shelob.afs.com> In article <4h1viu$m1@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: [...] > As I said, I raised this point only to demonstrate that even NeXT, whose > programmers I respect enormously, botched this first time out, and a lot > of early programmers either learned a bad thing or spent too much time > working around the deficiencies. Delphi at least got the DB part right, [...] At the time when DBKit was out, it was the coolest R/DB access framework around. It's not a shame when other people write something better later. But it's a shame if other people write _worse_ stuff later, and that is what is continually happening. NS is already 8 years old, and even NS0.8 was better than most OO environments are now. And another point: if you are saying that it's very horrible to change/rewrite sources you better work something else. But I think we should stop this discussion. I'll just summarize: - You made a comparison between Delphi and NeXTSTEP mentioning details only; - I answered you that you should compare the OpenStep & Delphi, and that the NeXT. Inc's focus changed drastically last 3 years and that NeXTSTEP != OpenStep; - You continue to bringing arguments from the NeXT's past completely ignoring OS (with an exception of arguments from one of your NEO sources ... but these arguments ware just false or misleading). You never compared the new features of OS and never mentioned the new stuff NeXT is bringing to the market. We are talking two different languages. I presume your inaccurate statements according the present products of NeXT Inc. are because of negative emotions. In the past you ware one of the most accurate and open minded people in this news group, so I'll recommend you to join the GNUstep community and when our OpenStep book we are writing now is ready I'll promise to send you one free copy with the hope to bring you again where I'm sure you (as a developer) will feel best. Regards -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
Newsgroups: comp.sys.next.programmer From: markus g <tm@burrow.muc.de> Subject: [q] perl5, where to put it? Message-ID: <7xravdydkz.fsf@burrow.muc.de> Sender: tm@burrow.muc.de Organization: hardly any... Date: Thu, 29 Feb 1996 22:25:00 GMT hi! i have perl4 installed on my system. where do i have to put perl5 stuff in order to have both perls coexist peacefully? (what paths are hard-coded into b.bum's (bbum@friday.com) compilation for NS?) currently i have: /usr/local/bin/perl4 and /usr/local/lib/perl (with lots of subdirectories used by perl4) and the man pages of course (but they pose a lesser problem). any suggestions? AtDhVaAnNkCsE, markus g a/k/a the mole -- / . . . . . . . . . . . . . . . . . . . . . . . . . . __ __/ __ __ . . . . . . . . . . . . . . . . . . . . . the mole / / / / . . . . . . . . . . . . . . . . . tm@burrow.muc.de __/ __/ __/ __/ . . . . . . . . . . . . http://www.muc.de/~mgloede/
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 29 Feb 1996 20:51:33 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4h53kl$lsa@news.its.com> References: <4gq6ok$9hl@lion.embl-heidelberg.de> <4gt62c$lph@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >> What if I do not like the Delphi's editor? With NS I can use any! With >> Delphi I have only one option... > > Did NeXT give you the keys to the car? Last I heard, the magic PB/gdb > interface to Edit.app was not formally or publically documented. It's > not in my help system, at any rate. [ ... ] You can look through the source code to NeXT's version of GDB which NeXT provides on the developer CD_ROM, and take a look at the debugger side of the interface. That'd probably be a good starting point, although you're right that the API is not "documented" at all. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: How to create example.eomodel? Date: 28 Feb 1996 22:06:57 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4h2jm1$28m@stc06.ctd.ornl.gov> In the Example FlatFileDataSource, how is the example.eomodel created? The eomodeler seems to always want either Oracle or Sybase as an adapter. -- - - - - - - - - - J. W. Wooten
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: latex2html Date: 29 Feb 1996 13:58:54 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4h4beu$l3t@stc06.ctd.ornl.gov> Has anyone gotten latex2html to work on NeXT? I'd like to know the tricks. -- - - - - - - - - - J. W. Wooten
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 29 Feb 1996 00:08:07 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4h2qp7$2ll@news.xmission.com> References: <4guqt5$5nd@ixnews3.ix.netcom.com> <4gviqf$dh@news.next.com> ngervae@sirius.com wrote: > I really must recommend that you avoid the Indexing Kit. It was a > decent idea, but the engineer had more work than he could do, and > it hasn't been properly supported since it was released. It isn't > going to evolve, improve, or be fixed, so best keep away.... I disagree a little bit. I, too, wouldn't recommend relying too heavily upon the IXKit right now, but that recommendation is also subject to change. The latter statement I disagree with, though. I can think of two cases--both of which could happen--where the IXKit could become a viable resource to developers: 1) The GNUStep people create a GNU IXKit from scratch that both implements the old API and works right. 2) NeXT "gives" the kit to a third party and the third party supports the kit properly, fixing and finishing it. Either way, we could yet end up with something worth considering. So, rather that avoiding the kit entirely, I'd say to wait a few months and see what happens. In my mind, either of the above scenarios (or both) could quite easily happen. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9602290012.AA25049@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Wed, 28 Feb 96 18:14:49 -0600 Subject: Problems with NSBundle Greetings, all. I'm developing an EOApplication which (hopefully) will make good use of the NSBundle class. Let me start off with what may be a sordid confession: I'm subclassing the EOApplication class to add some central/misc. functionality, including the determination of which bundle to load when called for. When I trained (about 2 years ago), it seems I vaguely remember the instructor (Ralph Z.) expressing distaste for subclassing Application, but I can't remember why. At any rate, this is the relevant portion of the bundle-loading method out of the app object. Right now, I have one bundle hardwired in to load every time, but ultimately, the app will swap different bundles in, depending on one of the fields returned in an EOGenericObject. @implementation EODealSheetApp (TargetActions ) - fetch: sender { id fetchArray; EOQualifier *dealNumberQualifier; char buf[ 55 ]; NSString *tradeID; id rootEObject; NSString *dealType; NSString *bundlePath; NSBundle *theBundle; id bundlePrincipalClass; { snip... } // This call works OK... bundlePath = [[ NSBundle mainBundle ] pathForResource: @"NOTFOUND" ofType: @".bundle" ]; // This one works OK too...sees the bundle theBundle = [ NSBundle bundleWithPath: bundlePath ]; // ******* We die right here... bundlePrincipalClass = [ theBundle principalClass ]; { snip... } return self; } @end As soon as I make the <principalClass> call, the console burps out: Feb 28 18:02:30 geek DealSheet[2756]: Error loading /private/Net/devhost/users5/tappd/Projects/DealSheet/DealSheet.app/NOTFOUND.bundle/NOTFOUND Feb 28 18:02:30 geek DealSheet[2756]: rld(): Undefined symbols: Feb 28 18:02:30 geek DealSheet[2756]: The bundle I'm loading only has one class in it, so there's not a problem hooking the wrong class. Also, I tried replacing the <principalClass> method with a <classNamed:> call, and got exactly the same results. I have the OTHER_LDFLAGS = -all_load set in the main Makefile.preamble, and BUNDLELDFLAGS = -all_load set in the bundle's preamble (which I'm not sure is appropriate, but the compiler doesn't complain). Can anybody spot what's up? Am I being punished for subclassing an Application object? (gdb doesn't seem to like what I'm doing at all, and reports all kinds of protected access exceptions, which appear to be spurious...) Thanks & regards to all, - Dan Tapp
From: levman@localnet.com Newsgroups: comp.sys.next.programmer Subject: The Enema Within Date: 26 Feb 1996 05:47:25 GMT Organization: LocalNet Corporation Message-ID: <4grhhd$1s1@prometheus.localnet.com> To all my netizen friends: I have a NeXT. Mine is black hardware. I have a 21' monitor and 400 meg internal drive. I believe in my machine and read this news group. Though Jobs never saw it, I free lanced a great marketing strategy for NeXT. I named the b&w model, the Orca Kriller Baby. The color model, (like mine), I named, the Colorful Cachalot, after the sperm whale. The three main slogans were, "The Rolls Royce Machines with the Model T price", "The Platform for Individual and World Wide Designs", and, "Today's Machine for Tomorrow's Technology". The main metaphor, in all the ad copy was, ›In a sea of information these machines crunch through data the way whales eat herring. . . . . by the ton". All the ads had a sperm whale on the monitor, swishing by a bunch of open windows. How I planned the distribution would have made Jobs into a billionaire, but then you wouldn't be reading this with vanilla hardware-evryone would be using black. OK. I'm jealous of vanilla, sort of. I could get into my clickedy NeXT keyboard with the Alpha chip secretly running the show. So onward. This post is here because I plan on having my NeXT in the oval office. Find out why they killed my friend Danny Casalaro, at the end of this letter to you. I'm a candidate for President of USA. When I post my drug platform you are no doubt going to want to join my campaign, whether you have ever used any drugs or not, because I have worked out a world solution to the prob limb. I have analyzed the election laws and have devised a way to put ten thousand dollars in your pocket, walkin around money, as one of my reps, and leave me wiith enough bucks to expose the fascists on TV, but I might not live that long as I am targeted by Stassi-USA, my name for the fascist domestic intelligence op, that brought us Waco and Ruby Ridge. So read, whether you could use an additional ten grand or not, because I want to live. I'm entitled to that.. You saw me a week ago, on Charles Osgood, and again, last tuesday, my 15 seconds of fame was on the ABC News. You might haave seen me getting ripped off, on America's talking, onJanuary 16. For every prob limb facing our nation I have an innovative non -partisan solution. I will post my platform, The Lev Deal, soon. Something else first. Your government, those volks from Stassi-USA, the dudes connected with those triggers at Waco and Ruby Ridge, are at the stage of talking, just talking now, about some more extreme prejudice, against an American citizen, one of your not quite known candidates for President, and, the same as in the Danny Casalaro matter, the Stassi-USA has begun to harass me, a bona fide candidate for President, with cryptic phone calls-messages left on my voice mail box -just when I step out the house for a few minutes-suddenly it's the FBI announcing they are FBI and they will be getting back to me later. The agent's name-that left in my voice mail box -is not associated with the Buffalo office of FBI, as the man announces in his message-it is a cryptic, hidden message they were leaving me-that I should drop out of politics altogether, or we, the Stassi-USA are going to smear your character. These fascist scummies are very liberal. Just like with Danny Casalaro, they are letting me know, in advance, that whatever happens, I will not be allowed to speak during the campaign for the presidency, and as long as I am going to keep pushing on my First Amendment rights, only harm will befall me. My response is to them is to push that much harder, thus my first posting. I have provided the access codes to my voice mail box one person on the net, known as, John Q. Public, who I have read around in all the political news groups, and who I actually accidentaly met up with, in the flesh, and I have personally challenged him to listen to the FBI harrasment on my voice mail box, for himself. That FBI message is followed by one, not directly connected to FBI, from Robert Baker, of the FCC. The slimy bureaucrat is telling me that the PBS station's counter offer of five minutes to my proposed New Hampshire two hour speech, that followed by a second request for one hour of access, was on the table.....in other words, an affirmative right to access, exercised by the other candidates, is not my right. Though I am a United States citizen, born here, so qualified to run for President, my first amendment rights to political speech are/have been stripped away by this fascist bureaucracy. Even a write-in candidate is entitled to access, though I am the only citizen ever to have actually crossed the FCC threshold of being certified a true bona fide write-in candidate. Now they are busy rewriting the rules to insure my broadcast rights will continue to be denied. What the FCDC jerk is saying, on my voice mail, is that the states where I got on the ballot aren't enough-you can smell the fascist in his voice-now there are additional requirements-and the Internet, as a campaign vehicle that can level the political field, does not count. Got that. The new threshold is that I have to physically travel to your state and .....then what? Give a speech on a street corner. For who.? How do I prove to the fascists that I was there. Can't. Now you can understand why I want to send you my recorded 58 minute speech. What you will find from that is I am an inspired man who could, given the chance, show you the way to a 6 hour day, 4 day week, 7 month year, with enough for a spouse and two kids. Or explain how to cut your rent in half and begin the slow disinflation of the world's economy so that in 40 years beer cost a nickle and space travel is free, with every line a derlicate sensible rhyme. And give you this info in a mull tie lingual format, running and punning in every spoken tongue, as above, with every line a delicate sensible rhyme. If Bob Dole wanted an hour or two from PBS, to make a speech, the door is open and the time is his bacause that is the law. But not Michael Stephen Levinson. I don't have any right to make any speech on TV. Only the others have that affirmative right to access which they mostly use for 30 sec. ads, though the law was written and enacted for substantive speech. I also submitted my one hour studio taped speech to PBS, in Virginia, and they are running their own fascist confusion of reality, as we speak, in order to continue the denial of my rights, until something happens to me . The PBS Virginia-Langely attorney writes, ›Without reaching the question of whether you are a nationally qualified candidate, you have not yet complied with our proceedures for scheduling access requests, and we cannot distribute any programming on your behalf until you do so.— This is in writing. The FCC jerk on my phone claims he talked with the PBS volks and Jews and Dogs need not alpply for First Amendment rights to speech. Very tricky. Everything is delay delay delay ubntil the primary is past and then-tough luck sucker. Without recognizing my bona fides, as a legally qualified candidate, any satellite distribution of my speech merely goes over the air without even being downloaded onto tape, for broadcasting anywhere. My speech would be down loaded for FBI domesrtic Stassi-USA purposes only. So how do I get them to write down their fascistic crap on paper so I can begin the thankless task of challenging them. And how long before they raise the ante on their veiled telephone threats and wipe me out altogether, because they aren't playing. I would love to make a copy of my one hour speech available to all of the readers of this post. You could then, upon viewing, submit it to your local access channel. I can't. The only people with access to the community channels are those people in the community who subscribe. It's all very tricky. These access channels were all spun off from the Cable stations, to be ultimately under funded and mismanaged into oblivion. The fascist FBI-Stassi hand in this, though hidden, was/is-to limit access by individuals for political sprechens. Were you able to see fresh faces in our pallah tics, you might get a feel for some other candidate for an important federal office, like President, or congress, and vote for them or give them a hand in their campaigns. Originally local access meant local, professionally operated studios, with programs originating in the communities. Spinning them off was a mistake, obviously, from my point of view, a fascist plot. The access law, on which the must carry element of all those political commercials is based, was written for speech. Telecommunications 312(a)(7) states the station can have its license revoked for the willful and repeated failure to allow access. PBS is under the same access law. So you don't have to compromise your integrity, begging money to run, and the PBS doors are supposed to be open to political speech. Could it be that I am the only citizen in America who understands this. Michael Stephen Levinson, 87 FCC2d, 433, 1980; (there are two versions of this case-one they sent me in the mail, and the other one they printed in law books that was fatally flawed-this to discourage anyone from ever examining this First Amendment issue my original complaint shows. The most recent, Michael Steven Levinson (note they willfully misspell my middle name-to discourage access via electronic data base) is from 1992. When Tom Brokaw had the debate with the six democrats, at the end Gerry Brown was going 1-800-me me me....I wrote a request for access, to NBC, and PBS, pointing out I was entitled to parity, and as long as they were not on the ballot in ten states (the arbitrary FCC threshold) then why should I have to be on the ballot in ten states. Get The New York Slimes, from January 16, 1992. Maureen Dowd had a front page article: Immersing Himself in Nitty-Gritty, Bush Barnstorms New Hampshire, that line appeared beneath a picture of Bush petting a cow in some New Hampshire barn. Dowd writes: Swinging Wildly: "Mr. Bush started out apologetic but he ended up swinging wildly at opponents on the left and right, using his high pitched campaign voice and what he has called his "red meat" vocabulary. "He lashed out at "mournful pundits," "egg head academicians," "smart aleck columnists" (and) "jacklegs jumping up demanding equal time with some screwy scheme." Who put my request for access on the President's desk back in '92-is that a scary compliment? (would that have set you back on your heels) I sought my own time-not e quill time as Bush mischaracterized my request, as there isn't any e quill time because everything the networks do is exempt. All of Perot's speeches could have been delivered for free on PBS, but Ross' media advisors forgot to tell him that. I was entitled to make a nationwide speech and this was suppressed and I waited two and a half years for FCC to finally rule with their fascistic slime, and I immediately appealed in the Dist. of Columbia Circuit-and now to the Supreme "Coat, where I hang barely by a thread." Three days after I filed in the Court of Appeals, a government person came to Belleville, Wisconsin, where I was hanging out, and photographed me, and then he photographed my car. Three weeks later the same dude showed up where I was working, just after I left for breakfast (I was semi-independent), claiming he'd heard from one of my friends, who was a friend of his, that I was selling my car, and that was why he was there (sure; my 1985 Dodge Charger, with 187,000 miles on it at the time was a hot car to be looking to purchase, and he had a cellular phone so he got a call letting him know when I was finishing eating, and he was gone. I didn't have any friends in Wisconsin. I was merely living there, without roots put down. This is your government's way of letting me, the citizen know that they, the FBI, for whatever their reasons, don't like me, (or you), and that they know who you are (I am) and where and when you go to and fro. As far as the highest court goes, I reached the Conference level, but it is veddy skeddy. They don't hear any case they don't read about first, in The Wall Street Gurgle, or The New York Slimes. I said to a clerk: This is scary. The clerk blurted out, Yes it is. But I digress all over the place, and why not , when FBI is getting ready- at least talking about -another Casalaro to cover their slime. I had a website. Recently I had a dispute with my provider who accused me of spamming. Spamming, from my point of view, is a million emails-not a couple hundred queries-so he erased my pages or removed them. As soon as I reestablish my web site somewhere else I will let all know where you can see pages from a prophetic work of art that was written down in multi-lingual form by a man who walked and talked with [ the lan lord uh pin heaven ] (don't fret-such an experience had to happen to some buddy who would then be into re-ledgend-a retelling of all the ancient stories) that was written down in design to perform on world wide television, like old blind Homer, for all the world's peoples to see, listen to and, interactively be a part of all at once, via the AT&T, ITT natch a rill rig up. Heavy duty. World pizza. every buddy getz a slice. The recitration of a prophetic work of art where every major event of the century was foretold in advance becomes DEUTERONOMY, a whole new bookie, the laying down of the law-you lay down-relax-sit back, like a God, get set-get yer television set in yer living room and the inspired Homeric one tells his own vision-it don't cost nuthin to listen-and delivers a dusk until dawn twelve our video trans crypt world thriller. What's wrong with a world vid cultural event that qualifies, politically, as the first peaceful night in five thousand years of recorded history. That would be the first peaceful night in five thousand years of recorded history because everyone would be doing the same thing at the same time-watchin TV. Even the most alienated person would tune in to see what everyone else is doing. Start with the creation-Adman and Even in the Gar Den ov Edum. You saw me on ABC News last Tuesday night, in New Hampshire-I'm watchable. With giant concepts like the one haphazardly expressed above, of course the Washington control freaks have someone surreptiously watching me like a hawk. Stay with me because I am going to expose the whole Casalaro business rye-cheer, including the pipeline from The Washington Post, where Casalaro brought his synopsis and signed his own death warrant. But , getting back, could I ask for your vote based on one five minute sprechen. Of course not. I requested two hours for a major address. There are giant issues facing our nation. You are entitled to spend an evening with the person who could send your children to war, the day after election. Before the primary opened in New Hampshire, my phone was ringing from the Secretary of State's office in New Hamp. Was I going to run for president, and would I be coming up on the first day because lots of local reporters wanted to interview me. So I did it. I'm on the docket in the United States Supreme Court, hanging by a thread: No 95-5876. The case is against the FCC, and behind them WGBH, in Boston, and the New Hampshire NHPTV. It won't be heard so your rights are surely being disolved. Ok Danny Casalaro. Here is why they murdered my friend. In the mid eighties, the govt. stopped printing a limited giant book that contained the names and addresses of every person getting a check from the Federal government. This book was available to Members of the House and Senate, and a few of the agencies. The info is out there, but with the rise of computers, they stopped snail printing that particular volume. Does the federal gov use electronic depositing. It used to be, that direct deposit of a social security check meant that a physical check was delivered to the actual bank for processing. Are you ready: With INSLAW software the direct deposits can be quickly separated from the other 30-50,000,000 checks going out every month. Then, using INSLAW, that super powerful search software, access the nationwide white pages, which are now on CD rom. Guess what dummies? There are 57,000 checks going out, every month, being directly deposited where there isn't anyone in the local telephone book, or even unlisted, to match up with the bank account. Whoops. For that the scum bags murdered poor Danny. To protect the 57,000 little fascists making a living giving FALSE WITNESS about ordinary people, to justify their domestic above-the-law intelligence bank account. Family. 57,000 fake bank accounts, into which are directly deposited your tax dollars. Danny Casalaro was on the edge of it-just starting to see why the governmemnt tried to kill the INSLAW company by stealing their search software and playing hardball-when he signed his own death warrant by bringing his synopsis up to the Washington Post, a noose paper famous from Watergate. He handed his synopsis over to one Joanne Aramao, who is or was The Washington Post City Desk Editor. She copied it and gave itr to her husband, Larry Mcniece, who works in a print shop-they tell me-outside of Washington. The two of them are a Stassi team and have been connected to domestic intelligence since they were students at the State University of Buffalo where the both of them were working for the student newspaper, The Spectrum, as undergraduates. McNiece is easily exposed. You come to Buffalo and I will show you his FBI photographic handiwork. Then, under Freedom of Information Act you check out the files of people whose pics he took and see the Joanne Aramao's husband's picturesas thewy were taken to be placed in that person's files. As soon as my name appeared in Time Mag, on December 18, 1995,as a poet, from Buffalo, N.Y. with a jobs program Larry, Jo-anne Aramao's husband was on the phone to his superiors, telling them I could be a prob limb in a big way, because I could expose and identify both him and his wife. As an undergraduate, she became editor-in-chief of The Spectrum, and HOWIE KURTZ, the Wash Toast medja critic, was her protogé. In the early eighties, after I left buffalo, N.Y., coincidenrtally, so did Jo-anne Aramao and Larry McNiece. She called in the chits, and Howie got her a job at the Post. Kurtz is not involved except that he does not like Michael Stephen Levinson because one of the deals, years ago, when he became the editor of The Spectrum, was that he was NOT to print any thing written by your humble poet. That was the word that came from the back room of the paper, where Larry McNiece stayed on working, after graduating. Anyone Post connected now begins to understand why I am badmouthed in the Post newsroom-not that I am the main topic for anything there except for someone to say that I am not worth covering. This is the good ship mother earth. Whatever deck you live on, the cards are dealt out evenly. When it comes time to change the course of human history, on the good ship mother earth, all the world cries out for, is a spokesman, a spokes person, to turn the wheel. You can quote me. I don't have any ax to grind with these fascist scummies.: Joanne Aramao and her little Hitler husband, Larry. They have to live with themselves, knowing forever that were it not for them Casalaro might still be a father to his son. I have spent fifteen years simply trying to give a speech. Where ever I have ever gone there is always someo9ne there to say, "Not you, Mr. take a walk. you are not allowed to speak: Nightwatch on the ship's prow The stars are out in disorder. Everything ever been seen By the naked eye Is out tonight. Michael Stephen Levinson My next post will have planks in my platform, the innovative solutions I bring to our highest office. Peace.
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 27 Feb 1996 16:25:16 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <4gvb9c$oql@shelob.afs.com> References: <4guk39$a4i@lion.embl-heidelberg.de> Georg Tuparev writes > You better not believe too much to your NEO sources: > - Sun's implementation _has_ RTF; Fair enough, they've added that. But that was a subsidiary point. Does NEO use NeXTHelp or AnswerBook format? That matters a lot if you're a harried CEO trying to determine how much effort it would take to port your app. WriteUp has 5MB of help that I don't feel like reformatting. > - OS on Sun & next are _not_ incompatible. OS is a standard. How can you say this.... > But everybody can add his own classes. So if you like the new Text > features, you should use NeXTSTEP 4.0. If you like flickering windows, > ugly colors, but fast ObjectiveC-C++ compiler, then you should use NEO. .. and follow it with THIS? You just admitted that it will be entirely poossible to write apps with dependencies on fundamental classes that will not exist in all implementations of OPENSTEP. For that matter, what about EOF? Yes, NeXT has promised to provide an NEO version, but think about it: Where is their incentive to hustle it out the door? They will be making their money on Windows now. Why feed the competition? > But, if your source is OS compliant, it will work on all three > implementation, and even produce the same results Assuming source is available. NeXT has never made its source available. What if I REALLY REALLY need that Text framework on NEO, and NeXT won't provide it? I'm stuck. In Delphi, at least, I get the source code. > when your customer comes to you says:"Dear Greg, we decided to change > our DB engine" ... you are in the happy position to rewrite your > sources. Of coarse, you will charge your customer ... First of all, this happens very rarely. But I don't know why you assume this requires a source rewrite. Delphi has a plug-in adapter mechanism similar to NEXTSTEP. All DB access uses the BDE API. It's even fully ODBC-compliant. Is EOF? > please open the preferences of PB. Oh wonder! You will see, that you may > well change the, editor and debugger, and even the builder. Uh huh. How does the developer of those tools find out how to integrate as fully as the NeXT-supplied tools? Maybe it's finally documented in 4.0, but it never has been before. Without that level of integration, it's pretty pointless to use anything else. > - The IB format (it's promised to be part of OS spec, but I've never \ > seen it); This seems pretty important to me. Especially because ... > - The object coding format! (it's neither promised nor I've seen it). > But this is important, because otherwise different OS implementation > can not share objects, and we are where CORBA is now ... nowhere... .. I forgot to include this in my original list. The NIB typedstream has NEVER been formally documented (although someone has cracked it), and my understanding is that NeXT and Sun are definitely NOT going to have the same NIB/archiving/whatever format. That means whenever I change a label or add a field, I will have a multiple maintenance chore. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: matthias@amg.de (Matthias Schuerhoff) Newsgroups: comp.sys.next.programmer Subject: Re: PB: changing font of NXBrowserCells Date: 26 Feb 1996 15:28:34 GMT Organization: AMG Industrieconsulting GmbH Message-ID: <4gsjj2$ccc@hagen.amg.de> References: <4gk7en$k1k@naiad.grenet.fr> Matija wrote: > Is it really impossible to change the font of a browser cell or am i missing something? > I am using the standard NXBrowser with the delegate method "browser:fillMatrix:inColumn" in which i tried to "setFont" to a userFixedPitchFontOfSize:10.0. > I tried several approches, with setFont applied to the prototype cell, to the matrix and to the cell. NO WAY! I get the right font in the cell, but the cell isn't resized and i can see only the upper part of the cell text, the lower being cut off. As far as I know the TextFontInfo is important. You should do something like the following: in MyCell.h; NXCoord ascender, descender, lineHeight, offset; - setFont:fontObj; in MyCell.m: - setFont:fontObj { [super setFont:fontObj]; NXTextFontInfo(support, &ascender, &descender, &lineHeight); return self; } Perhaps that helps... Matthias ===================================== Matthias Schuerhoff AMG Industrie Consulting GmbH Joseph-von-Fraunhofer-Str. 27 44227 Dortmund / Germany Telefon 231 / 97 53 54 - 0 Telefax 231 / 97 53 54 - 55 E-Mail matthias@amg.de (NeXTmail ok) =====================================
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 27 Feb 1996 09:49:29 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4guk39$a4i@lion.embl-heidelberg.de> References: <4gt62c$lph@shelob.afs.com> In article <4gt62c$lph@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > I should have been more clear about what I meant. I meant the PROGRAMMING > API is starting to diverge, not the look-and-feel. To quote from one of > my sources who is familiar with the work Sun is doing: > > "[DPS on top of X is] causing no end of problems with incompatible > imaging models, incompatible input models, and incompatible everything > else. Not to mention performance. OpenStep on Solaris on DPS on X makes my > old Mac-IIci look like a Ferrari. Nexts' OpenStep and Suns' OpenStep are > already incompatible [with respect to] ... the new text subsystem ... > NSTableView ... locale related stuff ... help (Sun wants to keep using > AnswerBook format, can't do RTF) ... and debugger (instead of gdb, they > hacked dbx to make it understand Objective-C; performance is miserable)." You better not believe too much to your NEO sources: - Sun's implementation _has_ RTF; - I was playing with the Sun's (_alpha_) implementation on a Voyager: well, it's not a Ferrari ... but just a Porche ;-) ... and you perhaps know, that no alpha or even beta is claiming to be fast! Think about NS3.0 beta ;-) - The problems are caused by X limited color map, not by DPS; - OS on Sun & next are _not_ incompatible. OS is a standard. But everybody can add his own classes. So if you like the new Text features, you should use NeXTSTEP 4.0. If you like flickering windows, ugly colors, ... but fast ObjectiveC-C++ compiler, than you should use NEO. And if you do not like paying money, than wait couple of months and start using GNUstep. The choice is yours. But, if your source is OS compliant, it will work on all three implementation, and even produce the same results ... something not very natural in M$ world ;-) > I don't know, and I don't care. As an application programmer, the > implementation of these features is irrelevant. I DO know that Win95 > doesn't seem to have any trouble launching apps by double-clicking > files, and there's a workable pasteboard. Do I care if it's DO? No. A workable pasteboard that does not work in Dialog box ;-) Happy working! BTW, you should start making destination between user and application programmer. As an application programmer, the implementation of these features is _very_ relevant! > C/S and remote database provide their own client libraries, and I use > them. As an application programmer, I don't know or care how they're > implemented. But I can understand why _you_ do, as a systems programmer. Yep! And when your customer comes to you says:"Dear Greg, we decided to change our DB engine" ... you are in the happy position to rewrite your sources. Of coarse, you will charge your customer ... > Most people don't develop these apps, they buy them. They don't care how > they're programmed, as long as they do the job. But I thought AFS _is_ developing apps? Correct me if I'm wrong ... may be I'm missing the announcement, where you said, that you are changing your focus from developing apps to a WinWord consultant company... > Did NeXT give you the keys to the car? Last I heard, the magic PB/gdb > interface to Edit.app was not formally or publically documented. It's > not in my help system, at any rate. Delphi 2.0, OTOH, further expands > upon the open tools interface that began in 1.0. So, please open the preferences of PB. Oh wonder! You will see, that you may well change the, editor and debugger, and even the builder. In the early days of PB, when there was no library model, I wrote my own library model (changing a lot of stuff in PB ... even the interface). But at one point you are right: NeXT should publish several things: - The API to dev tools (I think it is already done with NS4.0) - The IB format (it's promised to be part of OS spec, but I've never seen it); - The object coding format! (it's neither promised nor I've seen it). But this is important, because otherwise different OS implementation can not share objects, and we are where CORBA is now ... nowhere... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 26 Feb 1996 05:15:09 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4grfkt$iv@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
Newsgroups: comp.sys.next.programmer Subject: Re: 3.0 Developer on 3.3, possible? Message-ID: <DnAnHo.nM@udcf.gla.ac.uk> From: dmurray@music.glas.ac.uk (Duncan Murray(EMU92)) Date: Sat, 24 Feb 1996 19:00:58 GMT Sender: news@udcf.gla.ac.uk (News) Organization: Glasgow University Computing Service In article <4flbj6$2hl@gypsy.iisys.com> dguthrie@iisys.com writes: > Is it possible to install the User 3.3 version of NeXTStep and > still run the 3.0 version of the development system. I only > require the libraries and tools that 3.0 thus I though I could > run the latest OS (3.3) but still use my existing development > tools? > > Is this possible or is mentioned somewhere in a FAQ? > You can run 3.3 User and 3.2 Developer (it works on an Intel system anyway!). Seemingly (according to the bloke who supplied me) that was the intended way to run them unti 3.3 Dev came out. Duncan Murray dmurray@music.gla.ac.uk
From: far@ix.netcom.net(Felipe A. Rodriguez) Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 29 Feb 1996 06:54:46 GMT Organization: Netcom Message-ID: <4h3ijm$8b9@ixnews3.ix.netcom.com> References: <4h2qp7$2ll@news.xmission.com> In article <4h2qp7$2ll@news.xmission.com> don@misckit.com writes: >ngervae@sirius.com wrote: >> I really must recommend that you avoid the Indexing Kit. It was a >> decent idea, but the engineer had more work than he could do, and >> it hasn't been properly supported since it was released. It isn't >> going to evolve, improve, or be fixed, so best keep away.... > >I disagree a little bit. I, too, wouldn't recommend relying too >heavily upon the IXKit right now, but that recommendation is also >subject to change. The latter statement I disagree with, though. >I can think of two cases--both of which could happen--where the >IXKit could become a viable resource to developers: > >1) The GNUStep people create a GNU IXKit from scratch that both > implements the old API and works right. > >2) NeXT "gives" the kit to a third party and the third party > supports the kit properly, fixing and finishing it. > >Either way, we could yet end up with something worth considering. >So, rather that avoiding the kit entirely, I'd say to wait a few >months and see what happens. In my mind, either of the above >scenarios (or both) could quite easily happen. > >-Don Yacktman >don@misckit.com ><a href="http://www.misckit.com/don.html">My home page</a> I hope one of these two possibilities comes to fruition. IMO, IXKit is truly a remarkable piece of work. The Query Language would be the "coup de grace". I can't think of an easier way to implement a lightweight database engine under any platform. Had the Query Language been usable it would have saved me from the C/S solution I am now leaning towards :-( -- Felipe A. Rodriguez # ...it cannot be called ingenuity to Agoura Hills, CA # kill one's fellow citizens, to betray # friends, to be without faith, without far@ix.netcom.com # mercy, without religion; by these means (NeXTmail preferred) # one can acquire power but not glory. (MIMEmail welcome) # --Nicolo Machiavelli
Newsgroups: comp.sys.next.programmer From: mapdjb@bath.ac.uk (D J Batey) Subject: condition timed wait, in cthreads? Organization: School of Mathematical Sciences, University of Bath, UK Message-ID: <Dn8G6F.17w.B.mary@bath.ac.uk> Date: Fri, 23 Feb 1996 14:27:52 GMT Hi, Has anyone written, or does anyone know of, an implementation of a condition timed wait for cthreads? That is, a function that will wait upon a condition or until a timeout, whichever happens first. Duncan Batey, School of Mathematics, University of Bath, UK.
Newsgroups: comp.sys.next.programmer From: newcube!pascal (pascal) Subject: problem oracle v6 on NeXTStation and EOF Message-ID: <1996Feb26.193038.5192@tnt.oleane.com> Sender: pascal@tnt.oleane.com Organization: Terra Nova Techonologies (France) Date: Mon, 26 Feb 1996 19:30:38 GMT Hi all, I'm having a bit of trouble accessing our Oracle server from NEXTSTEP. Some info: Oracle v6.0.30.3.1 on a NeXTStation NextStep 3.2 with EOF 1.1 When I start Oracle with dbstart with a TWO_TASK configuration where the user variable is equal to T:oracle:oracle Also EOModeler does not seem to connect to the server,I think it will be normal. ORACLE_HOME=/Users/oracle export ORACLE_HOME ORACLE_SID=oracle export ORACLE_SID ORACLE_OWNER=oracle export ORACLE_OWNER TWO_TASK=T:oracle:oracle export TWO_TASK If I unset TWO_TASK my database works fine but I can not use the EOModeler. I have the orasrv service entered in NetInfo. orasrv 1525/tcp here my oratab file oracle:/Users/oracle:Y here my rc.local file fbshow -B -I "Starting Oracle Server" -z 100 su - oracle -c '/Users/oracle/bin/dbstart' fbshow -B -I "Starting SQLNET Process" -z 100 su - oracle -c '/Users/oracle/bin/orasrv' It seems that i 've got problem with my SQL*NET server wich is v1.2.7.1.3 version . I would like to know if someone 've got a new version of SQL*NET and if he can help me . Does anyone out there have any similar experiences? Any answers? Any thing? Regards, Jean Louis ------------------------------------------------------------------- Jean Louis PASCAL Terra Nova Technologies TOULON, France Internet : (NeXTMail welcome) pascal@tnt.oleane.com
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Thu, 29 Feb 1996 21:53:21 -0500 Organization: Steel Driving Software, Inc. Message-ID: <313666A1.F2B@ix.netcom.com> References: <4gq6ok$9hl@lion.embl-heidelberg.de> <4gt62c$lph@shelob.afs.com> <4h53kl$lsa@news.its.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Chuck Swiger wrote: > > Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > >> What if I do not like the Delphi's editor? With NS I can use any! With > >> Delphi I have only one option... > > > > Did NeXT give you the keys to the car? Last I heard, the magic PB/gdb > > interface to Edit.app was not formally or publically documented. It's > > not in my help system, at any rate. [ ... ] > > You can look through the source code to NeXT's version of GDB which NeXT > provides on the developer CD_ROM, and take a look at the debugger side of the > interface. That'd probably be a good starting point, although you're right > that the API is not "documented" at all. Which still leaves you in the position of having to graft PB/gdb compatability onto your editor of choice. Which is difficult if you don't have source for said editor. Thus the options are Edit.app, emacs, and any other PD or copyleft editors with source. If nobody takes the time to modify those editors, though, you get the same level of editor interchangeability as you get with Delphi. (Actually, the developers of MultiEdit have added features for Delphi interoperability.)
From: root@maverick.jboller.dialup.access.net (Operator) Newsgroups: comp.sys.next.programmer Subject: GBB for GCC 2.7.2 on NeXTStation m68k-next-nextstep3.3 Date: 25 Feb 1996 22:36:10 GMT Organization: PANIX Public Access Internet and Unix, NYC Message-ID: <4gqo8q$635@news1.panix.com> Keywords: GNU GDB GCC NeXT m68k Hi! I am looking for GBB for GCC 2.7.2 on NeXTStation m68k-next-nextstep3.3. Currently GDB 4.7 (NeXT 3.1) bus errors/segmentation faults with ObjC executables generated by GCC 2.7.2. (er..Mach executables?). gdb-4.15.1 does not build for the NeXT Motorola 68040 (m68k-next-nextstep3.3) I would greatly appreciate and hints or pointers to help! Thanks, John Boller jboller@panix.com
From: Wei Chen <wchen@webquill.com> Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Thu, 29 Feb 1996 23:20:52 -0500 Organization: Internet Online Services Message-ID: <31367B24.4AD2@webquill.com> References: <4gi5cq$5vi@shelob.afs.com> <4gru6c$5as@btmpjg.god.bel.alcatel.be> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Gregory H. Anderson writes > > > Now this is a legitimate point that I meant to mention in my original > > article. The Delphi and NEXTSTEP form designers have different strengths > > in the area of "event attachment," but both are deficient. In NEXTSTEP, > > InterfaceBuilder allows any target, but can only assign one action. In > > Delphi, the form designer can attach multiple actions, but only one target > > (the form itself). You can work around both deficiencies, but it requires > > programmatic runtime assignment. This is basically false. For example, Button in NS/OS can be assigned to one out of ANY number of actions reponded by ANY number of targets, while Dephi's button allows one and only one action to assign to a fixed target viz. Form. I am curious to know in what systematic way your form responds to multiple buttons on the form in your program in order to make your code more reusable. I hope that you will not tell me that you have a big case statement to achieve this. IMHO, using a form as a target is a past technology which started from 4GL and later used in VisualBasic, Powerbuilder and even Forte ... Too bad, Delphi could not get over this, even though it came out years later than NS. However, Smalltalk is similar to NS. This might be telling how a TRUE OO development environment should be like. :-)
From: dbora@tezcat.com (Donald F. Bora) Newsgroups: comp.sys.next.programmer Subject: rendering problems. Renderman Date: 24 Feb 1996 22:53:18 -0600 Organization: Breakfast at Tezcat. Message-ID: <4gopvu$mp6@huitzilo.tezcat.com> Keywords: renderman problem I recently upgraded my Next Cube to NS3.3. I have written an application which renders a scene with the photo-realistic renderer. For some reason the program is not finding the photo-realistic renderer. I didn't do anything before to help my application find this service.. it sorta just worked. I have checked for the existence of /usr/prman and it exists. rpc.renderd is there as well but when I try and run it under root I get the message: cannot create upd service. I didn't think the upgrade messed with any of the services files in /etc. If anyone has a clue as to why this would be having a problem I would greatly appreciiate your assistance. much thanks! Don Bora (dbora@tezcat.com)
From: wgaboria@iut-lr.univ-lr.fr Newsgroups: comp.sys.next.programmer Subject: Default UpperShift Char in textField ? Date: 26 Feb 1996 10:06:16 GMT Organization: Service Informatique IUT La Rochelle Message-ID: <4gs0mo$nmb@hpuniv.univ-lr.fr> Hello, Here is a problemwich seems simple : I'am trying to have only upper shift char in a textfield or text object. How to force the user to enter upper shift char ? Thanks to reply to : /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/ NeXT Mail are Welcome !
From: sundial@primenet.com (Sundial Services) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 24 Feb 1996 10:09:01 -0700 Organization: Primenet Sender: root@primenet.com Distribution: world Message-ID: <sundial.2403.003F7335@primenet.com> References: <4gkued$3q6@lion.embl-heidelberg.de> <4gl3us$alj@shelob.afs.com> In article <4gl3us$alj@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: >As to PDO, I have never found a "must have" use for it in my applications. >You may prefer to focus there instead of AppKit, but PDO is NOT where the >bulk of the corporate applications market is. Yes, I know about options >and derivatives firms who need to run massive Monte Carlo simulations on >Alpha chips, but you'll have to work awfully hard to convince me that >straightforward forms-based apps -- the bread and butter corporate apps -- >benefit from PDO. NeXT made this argument to the customer who is paying us >to move from NEXTSTEP to Delphi, and the customer didn't buy it. Besides, >if you really need PDO/D'OLE, my understanding is that it will not be >limited to OPENSTEP apps. I hope not, or it will find a limited market. >In which case, there is no reason a Delphi app couldn't leverage it too. >Georg, it doesn't sound like you have actually worked much on Delphi or >even Windows 95/NT. For example, the integration of Delphi's editor and >form designer goes WAY beyond IB's limited parsing of header files. The >statement that "Only NEXTSTEP makes it possible for over 30 people from >I don't know how many countries to work together so nicely [on MiscKit]!" >is patently wrong. More is possible in other environments than you think. >Before you spend much more time on your GNUstep labor of love, you should >take a step back, draw a deep breath, survey the market, and ask who >you're building this for and how much it is worth the effort. >-- >Gregory H. Anderson | "Honey, there're few programming >Gaffer/Best Boy/Key Grip | problems that can't be solved >Anderson Financial Systems | with duct tape." -- 'Father' Duke >greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95 Gentlemen, I found your collective postings to be among the most interesting ones I've read on the net in recent memory. But before the conversation degrades too much from the central points that have been so well-made, I think you are truly comparing apples and oranges here. NextSTEP does some things magnificently well, such as the things Georg is doing; Delphi does not and it was never designed to do so. For a "bread and butter business application" it might be overkill -- but that's not an issue because as far as I can see, NeXT never really penetrated that market at all (because they were overkill). Perhaps the strategy for OpenSTEP is indeed a deliberate compromise -- to give up some features and/or perfect compatibility, but in so doing bring down the cost and increase the market penetration of NeXT products into the areas where, as Gregory points out, less sophisticated functionality is adequate. Delphi is a dramatically-better development product partly because it introduces some advanced concepts into the Windows world, and indeed it is comparable to NeXT products, but it's not the same and does not need to be. It bears not being compared to NeXT too closely. /mr/
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Thu, 29 Feb 1996 22:28:00 -0500 Organization: Steel Driving Software, Inc. Message-ID: <31366EC0.3F87@ix.netcom.com> References: <4h1viu$m1@shelob.afs.com> <4h407f$gpe@lion.embl-heidelberg.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I'll let Greg speak for himself, but I feel a need to respond... Georg Tuparev wrote: > > In article <4h1viu$m1@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) > writes: > [...] > > As I said, I raised this point only to demonstrate that even NeXT, whose > > programmers I respect enormously, botched this first time out, and a lot > > of early programmers either learned a bad thing or spent too much time > > working around the deficiencies. Delphi at least got the DB part right, > [...] > At the time when DBKit was out, it was the coolest R/DB access framework around. > It's not a shame when other people write something better later. But it's a shame if > other people write _worse_ stuff later, and that is what is continually happening. > NS is already 8 years old, and even NS0.8 was better than most OO environments are > now. > And another point: if you are saying that it's very horrible to change/rewrite > sources you better work something else. The point (I think) was this: Mr. Harbour claimed that NeXTSTEP encourages good design. DBKit points out the fallacy in this: NeXT themselves, with the advantage of having massive brain power, source code, and the designers of NeXTSTEP itself, STILL produced a piece of software which is widely recognized as a suboptimal design. If *NeXT* doesn't 'automagically' produce good designs with NeXTSTEP, how can one claim that NeXTSTEP somehow automagically encourages good design? Perhaps I'm wrong, but I don't think that EOF includes the spirit of Grady Booch. > But I think we should stop this discussion. > I'll just summarize: > - You made a comparison between Delphi and NeXTSTEP mentioning details only; Details? As opposed to what? Opinion? Fantasy? > - I answered you that you should compare the OpenStep & Delphi, and that the NeXT. > Inc's focus changed drastically last 3 years and that NeXTSTEP != OpenStep; Seemed to me that Greg compared OpenStep & Delphi. OpenStep/Windows - NeXT's main product now - is competing against Delphi (in some problem domains). > - You continue to bringing arguments from the NeXT's past completely ignoring OS > (with an exception of arguments from one of your NEO sources ... but these arguments > ware just false or misleading). You never compared the new features of OS and never > mentioned the new stuff NeXT is bringing to the market. Are those new features really enough to justify OpenStep to many customers? > We are talking two different languages. I presume your inaccurate statements > according the present products of NeXT Inc. are because of negative emotions. > In the past you ware one of the most accurate and open minded people in this > news group, soI'll recommend you to join the GNUstep community and when our > OpenStep book we are writing now is ready I'll promise to send you one free > copy with the hope to bring you again where I'm sure you (as a developer) will > feel best. Translation: I prefer not to read wrong-thinking statements critical of NeXT and/or OpenStep. Those who would make critical or negative statements must be feverish and should lay down for a bit, until they regain their senses. A little warm milk and a nap should have Greg back to normal in no time. Georg, GnuStep may be fabulous. Your book may be wonderful. But if OpenStep tanks out, GnuStep isn't likely to keep someone's family fed, his mortgage paid, his investors happy. No business owner can afford to go through life with their head firmly planted in the sand. Such business owners don't own businesses for long. Ignoring the realities of the software marketplace, and only believing NeXT's PR announcements is an excellent strategy for rapidly achieving bankruptcy. This may not be an issue for you. Many people have more at stake than their spare time.
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 29 Feb 1996 20:05:28 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4h50u9$m6m@usenet.rpi.edu> References: <4h1viu$m1@shelob.afs.com> <4h407f$gpe@lion.embl-heidelberg.de> tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) wrote: > Greg_Anderson@afs.com (Gregory H. Anderson) writes: > [...] > > As I said, I raised this point only to demonstrate that even > > NeXT, whose programmers I respect enormously, botched this > > first time out, and a lot of early programmers either learned > > a bad thing or spent too much time working around the deficiencies. > > Delphi at least got the DB part right, > [...] > At the time when DBKit was out, it was the coolest R/DB access > framework around. It's not a shame when other people write > something better later. But it's a shame if other people write > _worse_ stuff later, and that is what is continually happening. > NS is already 8 years old, and even NS0.8 was better than most > OO environments are now. And another point: if you are saying > that it's very horrible to change/rewrite sources you better work > something else. > > But I think we should stop this discussion. > I'll just summarize: > - You made a comparison between Delphi and NeXTSTEP mentioning > details only; > - I answered you that you should compare the OpenStep & Delphi, > and that the NeXT Inc's focus changed drastically last 3 years > and that NeXTSTEP != OpenStep; > - You continue to bringing arguments from the NeXT's past > completely ignoring OS (with an exception of arguments from > one of your NEO sources ... but these arguments were just false > or misleading). You never compared the new features of OS and > never mentioned the new stuff NeXT is bringing to the market. I think you've completely misunderstood the comments Greg has been making. He isn't bringing up examples of NeXT's past to say how bad NeXT was, he's doing it to say how quickly improvements can happen. The fact that Delphi is where it is now indicates that they at least "get it". Now that they "get it", the fact that they aren't as far along as NeXT does not prove that they will never be as far along as NeXT. Thus, it is very reasonable to assume that some of the deficiencies that have been noted in Delphi (by both you and Greg) may be addressed soon. From my reading of Greg's comments, it seems fairly clear to me that he is interested in the long-term success of OpenStep. He is just commenting on the issues which he considers a barrier to that success. You're reading his comments as if he's got some vendetta against NeXTSTEP, and I don't see that at all. The reasoned comparision of NeXTSTEP to Delphi was not the work of a raving lunatic who wants to tear down NeXT. It was the analysis of someone who has worked a lot with NeXTSTEP, and has the experience to do an intelligent comparison of the two environments. > We are talking two different languages. I presume your inaccurate > statements according the present products of NeXT Inc. are because > of negative emotions. In the past you ware one of the most accurate > and open minded people in this news group, In my mind, he still is. > so I'll recommend you to join the GNUstep community and when our > OpenStep book we are writing now is ready I'll promise to send > you one free copy with the hope to bring you again where I'm sure > you (as a developer) will feel best. I think that GNUStep goes a long way to addressing some of the concerns that Greg has mentioned, even though it will not be a viable solution for every potential customer. I think you're reading his remarks as an enemy attack, when in fact he's still quite much on the side of NeXTSTEP and OpenStep. The fact that he's on the side of these products does not mean he has to overlook the areas which he thinks might cause trouble for those products. I am totally baffled by the people who are responding to Greg as if he's got some grudge against NeXT. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu ITS Systems Programmer (handles NeXT-type mail) Rensselaer Polytechnic Institute; Troy NY USA
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.software Subject: Re: WODocumentation Date: Thu, 29 Feb 1996 22:08:27 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Feb29.220827.354@seer.demon.co.uk> References: <4gs4r3$8e@cube.ppp.informatik.uni-muenchen.de> In article <4gs4r3$8e@cube.ppp.informatik.uni-muenchen.de> Dominik Westner <dominik@gowest.ppp.informatik.uni-muenchen.de> writes: > mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > > Maybe it might be a good idea if NeXT could create an official > version of > > this...?! :-) > > They do ... maybe you should try: > > http://www.next.com:80/Pubs/Documents/WebObjects/Beta1/WODoc.v2.tar.Z As a matter of curiousity, are there any discernable links to this from the WebObjects or Pubs pages? Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 23:49:14 GMT Organization: Orbital Computer Consultancy Message-ID: <31323e5e.11782536@news.bogo.co.uk> References: <312BE9EC.4B0A@gramercy.ios.com> <4gi5cq$5vi@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >Wei Chen <wchen@gramercy.ios.com> writes >> Although I essentially agree with your assesments on Delphi. I would >> like to point out that Delphi is not truly object-oriented development >> platform as far as reusability is concerned. > >> Without taking additional steps, Delphi developers would put all the >> logic (be business or control) on the form object( NeXT's equivalent >> Window object .) It's quite well-known that when programming in >> NeXTstep, developers should put this logic on a custom controller object >> rather than a Window or its subclass object in order to achive greatest >> reusability. NeXT also advised developers not to subclass Application >> for exactly the same reason. > >Whoa, whoa, whoa, hold the phone, Chester! You can do bad design with ANY >development tool. Just because most Delphi developers are newer to OOP and >make some mistakes, you can't blame that on the tool. If you know what >you're doing, Delphi is every bit as capable as NEXTSTEP of promoting >reusability. Our trading app framework is 100% M-V-C. Our AfsController >component is dragged to a form, all business models descend from AfsModel >(which inherits from TDataSet), and TForms are very minimally subclassed. > Obviously I haven't seen your AfsController class, but I find it hard to accept that you can achieve 100% MVC with it. I accept that it separates out the business logic from the UI, but I think there is more to MVC than that, especially considering your claim of complete purity. Surely it's a requirement of MVC that the model objects ARE objects. A TDataSet operates on a SET of rows in a table. How is it possible to subclass something which represents a set, and yet have instances of it represent a single model object instance? In any case, how can it be good OO design for a Currency class, say, to inherit methods such as procedure CursorPosChanged; procedure Delete; procedure DisableControls; procedure Edit; procedure EnableControls; If my model classes have non-persistent attributes, how is this modelled in your scheme? Calculated fields are no good, as I may want the attribute to be settable by another object through public API. How do I do this with a TDataSet subclass, because clearly any instance variables apply to a set of model objects, not just one. If such attributes were stored somewhere else, it wouldn't be a very OO design. That's fine if you're putting a simple bread-and-butter business app together, but you have made bold statements about awards for OO design and 100% MVC purity. Also, I'd be interested in how you solved the classic OO to relational-database mapping problem? When you've got a Model class hierarchy, do subclasses have their own table, go in their parent class's table, or spread across both? If this hasn't been an issue for you, is it perhaps because your model is computationally complex but structurally fairly simple? Does your framework provide an object-uniqueing facility like EOF does? If you've got an OO API, surely it's important that object identity is preserved in this way? <-- snip -->
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 23:49:17 GMT Organization: Orbital Computer Consultancy Message-ID: <3132418e.12598508@news.bogo.co.uk> References: <4gkued$3q6@lion.embl-heidelberg.de> <4gl3us$alj@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >Georg Tuparev writes >> Wake up Greg! > >Oh, I am quite awake, thank you. NEXTSTEP _was_ a complete OO environment Still is, last time I looked. >(including a really good unified user environment), but OPENSTEP will not >be. In fact, many of the complaints I have heard about OPENSTEP/Win from >a functional point of view is how much of that wonderful NEXTSTEP we take >for granted has been "dumbed down" for the sake of being a "first class NT >citizen." In other words, NEXTSTEP loses some of its magic on the way to >OPENSTEP/NT. I think you confused the issue by using the word NextStep in the thread title. <-- snip --> > >As to PDO, I have never found a "must have" use for it in my applications. >You may prefer to focus there instead of AppKit, but PDO is NOT where the >bulk of the corporate applications market is. Yes, I know about options >and derivatives firms who need to run massive Monte Carlo simulations on >Alpha chips, but you'll have to work awfully hard to convince me that >straightforward forms-based apps -- the bread and butter corporate apps -- But I thought your argument was based on a nearly-award-winning OO app which you translated remarkably easily from Next to Delphi, not some straightforward triviality? >benefit from PDO. NeXT made this argument to the customer who is paying us >to move from NEXTSTEP to Delphi, and the customer didn't buy it. Besides, >if you really need PDO/D'OLE, my understanding is that it will not be >limited to OPENSTEP apps. I hope not, or it will find a limited market. >In which case, there is no reason a Delphi app couldn't leverage it too. > Surely, the point with (P)DO is that you hardly know you're doing it, i.e. local and remote messaging have the same syntax. With Distributed OLE as your ORB, surely OLE is gonna come and smack you in the face like a really badly smelling fish? BTW, for a really good and impartial treatment of CORBA/OLE etc, I can recommend "The Essential Distributed Objects Survival Guide" (ISBN 0471-12993-3). >Georg, it doesn't sound like you have actually worked much on Delphi or >even Windows 95/NT. For example, the integration of Delphi's editor and >form designer goes WAY beyond IB's limited parsing of header files. I would think IB doesn't need to go further than header file parsing, due to its object archiving nature, a subject which I know has been discussed here before. (IB does need improving, but not in this direction.) <-- SNIP -->
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 26 Feb 1996 23:49:19 GMT Organization: Orbital Computer Consultancy Message-ID: <31324223.12747679@news.bogo.co.uk> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4gi8ei$ekv@cheyenne.iac.net> jon@heinlein (Jonathan Hendry) wrote: >Wei Chen (wchen@gramercy.ios.com) wrote: >: Although I essentially agree with your assesments on Delphi. I would >: like to point out that Delphi is not truly object-oriented development >: platform as far as reusability is concerned. > >: Without taking additional steps, Delphi developers would put all the >: logic (be business or control) on the form object( NeXT's equivalent >: Window object .) It's quite well-known that when programming in > >This isn't because of an inherent shortcoming in Delphi. This is >a result of being new to objects, as most Delphi users are. A >similar phenomenon can be found among people new to NeXTSTEP. > >Everyone who's written an app where all the functionality is implemented >in 'AppController.m' or such, raise your hand. Maybe so, but at least it would be a controller object, and not a subclass of Window or View. All this example shows is that novices put too much functionality in one class, but at least a NextStep novice separates Controller logic from View logic right from the word go, and they don't do bizarre things like subclassing Window because it contains some Controls that the designer of the Window class didn't anticipate. > >: NeXTstep, developers should put this logic on a custom controller object >: rather than a Window or its subclass object in order to achive greatest >: reusability. NeXT also advised developers not to subclass Application >: for exactly the same reason. NeXT achieved this by providing, e.g., >: target/action configurable at the design time via IB. However, Dephi's >: target is hard-coded to be the form object. In terms of M-V-C paradigm, >: Delphi seems mixing the C component with the V. > >That's not entirely true. There is a default, which you get by double-clicking >on an event. Which in many cases is adequate. You can also specify different >methods. > >NeXTSTEP certainly doesn't enforce good design. But it does encourage it, and that's what matters. Something "which in many cases is adequate" is hardly encouraging good OO design. >
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 26 Feb 1996 09:23:24 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4gru6c$5as@btmpjg.god.bel.alcatel.be> References: <4gi5cq$5vi@shelob.afs.com> Gregory H. Anderson writes > Now this is a legitimate point that I meant to mention in my original > article. The Delphi and NEXTSTEP form designers have different strengths > in the area of "event attachment," but both are deficient. In NEXTSTEP, > InterfaceBuilder allows any target, but can only assign one action. In > Delphi, the form designer can attach multiple actions, but only one target > (the form itself). You can work around both deficiencies, but it requires > programmatic runtime assignment. For NEXTSTEP IB you can write your own connection classes (IBConnection and the appropriate connection inspector) and then, for each object, have as many targets/actions as you want (depending on object state or single-double-triple click, whatever). Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine.
From: filip@filtronix.eunet.be (Filip Lingier) Newsgroups: comp.sys.next.programmer Subject: Re: Getting EOF Adaptor to work with Microsoft SQL 6.0 server Date: 25 Feb 1996 19:56:35 GMT Organization: Filtronix Inc. Distribution: world Message-ID: <4gqetj$da@andromeda.filtronix.eunet.be> References: <4gii2e$3bv@cetus.ali.bc.ca> In article <4gii2e$3bv@cetus.ali.bc.ca> marcel@star.ali.bc.ca (Marcel Sutanto) writes: > Did anyone ever setup an EOF connection to Microsoft SQL 6.0 server > running on Windows-NT? Yes, I'm running the Microsoft SQL Server 6.0. > My understanding is that the SQL 6.0 is really a Sybase SQL server > and the EOF adaptor for Sybase should work with SQL 6.0. The development of the Sybase and the Microsoft SQL server have went their own paths after v4.2. But MS SQL Server 6.0 is backwards compatible with the older version and still works with the Sybase EOF adaptor. Filip -- ---------------------------- FILTRONIX ----------------------------- |-- --- \ / Software Development - OpenStep|Windows|X11 |- | X Web Design & Development - HTML|CGI|JAVA|WebObjects | | / \ ---> info@filtronix.eunet.be
From: apuleius@ix.netcom.com(William Grosso) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 1 Mar 1996 07:08:28 GMT Organization: Netcom Message-ID: <4h67pc$kdl@reader2.ix.netcom.com> References: <4h1viu$m1@shelob.afs.com> <4h407f$gpe@lion.embl-heidelberg.de> <31366EC0.3F87@ix.netcom.com> In <31366EC0.3F87@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> writes: > >If *NeXT* doesn't 'automagically' produce good designs with >NeXTSTEP, how can one claim that NeXTSTEP somehow automagically >encourages good design? Perhaps I'm wrong, but I don't think that >EOF includes the spirit of Grady Booch. > I've been staring at this for a few minutes now and I still don't know: Is this last statement pro or con EOF ? If EOF somehow included the spirit of GB, would that make it a better product ? Personally, I'd rather have my apps channel Albert Camus. For one thing, he'd probably write much better alert panels. Aujourd'hui, le programmation est morte... Andy
From: nurban@csugrad.cs.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer Subject: Re: glibc or libg++ for NEXTSTEP? Date: 1 Mar 1996 03:08:01 -0500 Organization: Virginia Polytechnic Institute and State University Message-ID: <4h6b91$q0@csugrad.cs.vt.edu> References: <4h6b4r$13g@csugrad.cs.vt.edu> In article <4h6b4r$13g@csugrad.cs.vt.edu>, nurban@vt.edu wrote: > Does anyone have patched source or x86 binaries of glibc or libg++ for > NEXTSTEP? I built gcc and then tried compiling glibc, but 'configure' By the way, I've already found them on the standard FTP archives, but I'm looking for the latest versions. The ones I found were pitifully old. I think the latest versions are 1.09 for glibc and 2.7.1 for libg++. -- -------------------------------------------------------------------------- Nathan Urban | Undergrad {CS,Physics,Math} | Virginia Tech nurban@vt.edu | {NeXT,MIME} mail welcome | http://nurban.campus.vt.edu/ --------------------------------------------------------------------------
From: marco@sentenext1.epfl.ch (Marco Scheurer) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 26 Feb 1996 09:19:11 GMT Organization: Ecole Polytechnique Federale de Lausanne Message-ID: <4grtuf$h78@info.epfl.ch> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> Wei Chen <wchen@gramercy.ios.com> wrote: [...] > In terms of M-V-C paradigm, > Delphi seems mixing the C component with the V. Isn't it also also the case with NEXTSTEP, where the View class implements many things that are in the Controller component of the MVC triad, like mouse event handling? NEXTSTEP "Controllers" are different from the original MVC Controllers. Not a bad thing IMHO. --- Marco Scheurer Sen:te Parc Scientifique EPFL -- CH-1015 Lausanne -- Switzerland marco@sente.epfl.ch (NeXT, MIME mail welcome)
From: stefan.boehringer@rz.ruhr-uni-bochum.de (Stefan Boehringer) Newsgroups: comp.sys.next.programmer Subject: Does EOF coerce values properly? Date: 1 Mar 1996 06:47:19 GMT Organization: Ruhr-Universitaet Bochum, Rechenzentrum Message-ID: <4h66hn$5j1@sun168.rz.ruhr-uni-bochum.de> I've defined my own convenience class for an entity. However I find a date property initialized with a NSString which the documentation says should not happen (the EOController should coerceValue:forKey: to get proper value objects). This seems to apply to EOGenericRecord as well. Has anybody experienced this as well? Am I missing something about this behaviour? Thanks. - stefan boehringer
From: stefan.boehringer@rz.ruhr-uni-bochum.de (Stefan Boehringer) Newsgroups: comp.sys.next.programmer Subject: Does EOF coerce values properly? Date: 1 Mar 1996 06:49:23 GMT Organization: Ruhr-Universitaet Bochum, Rechenzentrum Message-ID: <4h66lj$5j9@sun168.rz.ruhr-uni-bochum.de> I've defined my own convenience class for an entity. However I find a date property initialized with a NSString which the documentation says should not happen (the EOController should coerceValue:forKey: to get proper value objects). This seems to apply to EOGenericRecord as well. Has anybody experienced this as well? Am I missing something about this behaviour? Thanks. - stefan boehringer
From: hÛö’ pa Newsgroups: comp.sys.next.programmer Subject: Help On Linking EOF to a Bundle Date: 27 Feb 1996 21:29:56 GMT Organization: ALI Message-ID: <4gvt4k$43n@cetus.ali.bc.ca> Hello, We are writing a kit to make use of EOF. It works fine on application level by including necessary libraries in Project Builder. However, it fails to run when put in a feature bundle. The compiling and linking process are ok but at run time we would receive error loading message if try to invoke the bundle. Are we missing something here or EOF has its limit to be integrated within a bundle? The bundle we are using is a subclass of NXBundle but not NSBundle, does it matter? Below is part of our Makefile.preamble : ... ... # Flags passed to compiler (in addition to -g, -O, etc) OTHER_CFLAGS = -all_load -I$(MYHDRLIB) -I$(MYCLASSLIB) -I$(MYPROTOLIB) \ -I$(HDRLIB) -I$(CLASSLIB) -I$(PROTOLIB) -I$(EOHDRLIB) \ -DANSI # Flags passed to ld (in addition to -ObjC, etc.) OTHER_LDFLAGS = -L$(MYHDRLIB) -L$(MYCLASSLIB) \ -L$(HDRLIB) -L$(CLASSLIB) -L /usr/lib/ ## Configure what is linked in at each level here. Libraries are only used in ## the final 'app' linking step. Final 'app' linking is only done via the ## 'app', 'debug', and 'profile' targets when they are invoked for ## the top-level app. # Additional relocatables to be linked in at this level OTHER_OFILES = $(CLASSLIB)/libKit.a \ /usr/lib/libFoundation_s.a /usr/lib/libEOAccess_s.a \ /usr/lib/libEOInterface_s.a ... ... We would very appreciate if anybody can give some ideas. Thank your very much for the attention. - Jim -- Yi-Ming (Jim) Sun jimsun@ali.bc.ca Software Developer (NeXT Mail accepted here) A.L.I. Tenologies Inc. Richmond, BC, Canada (604)279-5422 ext.361
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.software Subject: Re: WODocumentation Date: 1 Mar 1996 12:14:27 GMT Organization: Academic Computing Services, Sheffield University Message-ID: <4h6pn3$2mh@bignews.shef.ac.uk> References: <4gnn9c$4b9@bignews.shef.ac.uk> <4gs4r3$8e@cube.ppp.informatik.uni-muenchen.de> In-Reply-To: <4gs4r3$8e@cube.ppp.informatik.uni-muenchen.de> On 02/26/96, Dominik Westner wrote: > mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > > Maybe it might be a good idea if NeXT could create an official > > version of this...?! :-) > > They do ... maybe you should try: > > http://www.next.com:80/Pubs/Documents/WebObjects/Beta1/WODoc.v2.tar.Z > Hmm, so where's the pointer from the WWW pages? I hope that my efforts will nevertheless be useful as documentation which can be easily incorporated into DL, which seems to make it easier to do wider searches on the rest of the developer docs at the same time. Best wishes, mmalc. --
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.lang.pascal.delphi.misc,comp.sys.next.advocacy Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Fri, 1 Mar 1996 10:35:29 -0500 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <klBlZ1_00iV5A16cAD@andrew.cmu.edu> References: <4h1viu$m1@shelob.afs.com> <4h407f$gpe@lion.embl-heidelberg.de> <4h50u9$m6m@usenet.rpi.edu> In-Reply-To: <4h50u9$m6m@usenet.rpi.edu> Excerpts from netnews.comp.sys.next.advocacy: 29-Feb-96 Re: NEXTSTEP vs. Delphi: An.. by Garance A Drosehn: > I am totally baffled by the people who are responding to > Greg as if he's got some grudge against NeXT. I find it baffling too, although I believe I understand why it's happening. You see, Greg has been performing a detailed analysis of NeXT's technology in public. Some of his comments actually criticise NeXT. Some people equate even the most honest, technically sound, and justified criticism of NeXT as an attack. Along these lines, recently someone described a problem with the Display PostScript interpreter used by the windowing system of NEXTSTEP. Although DPS is multithreaded, it is still possible for a sufficiently complex PostScript request to effectively lock up the entire system. This is because DPS also handles keyboard events, so it is impossible to kill the offending task from the console-- you have to telnet in from another machine. I responded, and pointed out that round-robin scheduling of DPS contexts would probably work well (assuming it's properly implemented, of course). The gentleman at NeXT who had written the DPS scheduler regarded my comments as a personal attack. Since a private apology apparantly did not suffice, I will repeat it publicly: ----------------- > [ ... ] Someone described a problem with DPS and I provided a technical criticism. If you choose to take that as a personal attack then there is nothing I can do except say that it was not meant to be one, and I apologize. You know, I really don't understand you NeXT people. NeXT has had some of the brightest and most creative people in the computer industry working for you, and NeXT has created some brilliant software. You guys also seem to be completely unable to deal with constructive criticism. NeXT has repeatly ignored serious problems with your software for years. Your typical response is to either pretend that the problems don't exist, or else to say "we're NeXT and we've got our way of doing things; if it wasn't invented here, it obviously can't matter very much." What's the big deal about acknowledging that problems exist? > [ ...The person in question provided me with a few general comments about > the design of DPS interpreter, and why it sometimes has problems... ] The problem in question appears to result from DPS being unable to switch to working on another context if the interpreter is busy trying to perform some complex operation and is locking the critical sections; presumably, no other thread is runnable because they're blocked waiting for the currently active thread to release it's locks. The most obvious thing that comes to mind is to schedule a SIGALRM interrupt for a second or so in the future every time DPS switches to a new context. If the alarm goes off, this means that the interpreter is spending too much time on the current context and you should forcibly perform a context switch by saving current state and yielding. ----------------- Sigh. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 1 Mar 1996 16:28:13 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4h78it$877@shelob.afs.com> References: <31367B24.4AD2@webquill.com> Wei Chen <wchen@webquill.com> writes > Gregory H. Anderson writes > > > > > The Delphi and NEXTSTEP form designers have different strengths > > > in the area of "event attachment," but both are deficient. In > > > NEXTSTEP, InterfaceBuilder allows any target, but can only assign > > > one action. In Delphi, the form designer can attach multiple > > > actions, but only one target (the form itself). You can work around > > > both deficiencies, but it requires programmatic runtime assignment. > > This is basically false. For example, Button in NS/OS can be assigned > to one out of ANY number of actions reponded by ANY number of targets, > while Dephi's button allows one and only one action to assign to a fixed > target viz. Form. I am curious to know in what systematic way your form > responds to multiple buttons on the form in your program in order to > make your code more reusable. I hope that you will not tell me that you > have a big case statement to achieve this. Heavens no. Here's a further clarification of what I see as the event attachment differences: In NEXTSTEP, you can attach a Button to any "sender" method of any other object. But the only Button message you can attach is its 'click', the primary action of the Button. That's because the target->action paradigm is hard-wired, non-extensibly, into InterfaceBuilder. This assumes, of course, that you are using the standard classes. It is possible, with some effort, to build custom IBConnectors for custom classes. In Delphi, the Designer can only attach Button events to methods of the Form on which the Button resides. This is a deficiency, and I acknowledged that. HOWEVER, Delphi Buttons can send many events, not just clicks. In Delphi, the Designer can attach all of the drag/drop, enter/exit, and keypress events to custom methods. This part of Delphi's scheme is better than NEXTSTEP, because multiple attachments are possible. The downside is that they all point to the same, fixed responder. For standard data entry forms, we have an AfsController object that points to all the primary buttons, like 'New', 'Edit', 'Delete', etc. (It needs those attachments primarily so it can enable/disable the buttons, depending on edit state.) When the DFM file is loaded into memory, the AfsController object checks to see if the OnClick events of these buttons are attached to anything. If so, it assumes the programmer has something special in mind. Otherwise, they are attached to the standard processing methods of the controller. That is what I meant by automatic programmatic attachment. We have to do the same thing in NEXTSTEP, where the validation and field-help targets must be attached programmatically, since only one target->action is assignable in IB. > IMHO, using a form as a target is a past technology which started from > 4GL and later used in VisualBasic, Powerbuilder and even Forte ... Too > bad, Delphi could not get over this, even though it came out years later > than NS. However, Smalltalk is similar to NS. This might be telling how > a TRUE OO development environment should be like. :-) My understanding is that Delphi is 'Form-centric' with respect to events, because Windows itself is so window-centric about dispatching events. That's not an excuse, just an observation. As to "TRUE" OOP environments, clearly you can make any Delphi object send a message to any other receptive Delphi object. You just can't use the Designer tool to do it. Frankly, I can't think of any reason why you couldn't write a custom component editor for Delphi to accomplish this. I've seen how the existing event editor does its work, and the RTTI is certainly available to do it. It shouldn't be hard to extend that functionality for any random receiver. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Delphi vs. NEXTSTEP: A summary of what we've learned Date: 1 Mar 1996 16:33:41 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4h78t5$87n@shelob.afs.com> This has been fascinating discussion. I'd like to wrap up a few points before the current set of arguments turns circular (as they inevitably will, this being Usenet). First, for those of you who read my initial comparison as an emotional attack on NEXTSTEP, I only have one thing to say: your prejudices are showing. Read it again. I brought up plenty of places where each tool has advantages and disadvantages relative to the other. I'm sorry if I burst any bubbles about how fast the competition is catching up to NeXT, but we have to live in the real world. If Win32 is your target, Delphi is real. OPENSTEP/NT will have to compete against it. No, it does not possess every bell and whistle NeXT offers, but for what my company does -- fairly sophisticated MCCA Wall Street trading apps -- it is more than sufficient. As it relates to AFS, this is an unarguable fact. Please don't try to change my mind by citing a list of features I can prove I don't need. Second, I _do_ have a prejudice. In an ideal world, NeXT would have done a better job of marketing a great development environment, and I wouldn't have had to consider working on Windows. Nothing would have made me happier. Alas, things didn't turn out that way. So we had to ask ourselves what was the best way to get to Windows: wait for OS/NT, or find an existing native OOP tool? With the direction of a customer, we chose the latter, because NeXT could not provide timely answers for things like runtime fees, minimum hardware/RAM configuration, and deployment dates. I have never regretted we chose Delphi. We are going to get to deployment, we are going to do it in approximately the same timeframe, and we did not have to compromise our existing OOP design. My 'NEXTSTEP-spoiled' programmers have not mutinied; in fact, they are quite happy. The customer will be running the app before OS/NT even makes it to beta testing. It will run comfortably on a 16MB Pentium machine running NT 3.51, talking to a Sybase server. It will not require a runtime fee. Call me crazy, but I don't see the downside. Georg keeps pestering about OPENSTEP. I have referred to it only minimally for a simple reason: OPENSTEP is not real yet. You cannot buy it. It's an interesting spec, and I know that at least three organizations (including the GNU gang) are working on implementations, but it is not deployable at this time. I have opinions about if and when each of those parties is likely to ship a product, but I will not share them, because they are based on information I am not entitled to disclose. They'll get here when they get here. In the meantime, Delphi 2.0 is already in customer hands, so it is fair to talk about it. A troubling aspect of OPENSTEP is the amount of work it will take to get there from NEXTSTEP. Yes, I know there will be automated tools to do much of the grunt work, but for big apps, there will still be a lot of cleanup and retesting. Furthermore, the name space and class hierarchies are different in significant ways. When we were making our original decision, one point that did NOT work in OPENSTEP's favor was training time. That may sound surprising, given how many experienced NEXTSTEP programmers we employ. But when you look at the OPENSTEP spec, there's a lot of new stuff to learn about classes where our programming reflexes now run on auto-pilot. Five years into it, I can recite almost all of the AppKit methods in my sleep. I almost never refer to help while I'm coding. But in OPENSTEP, it seems like everything changed, sometimes arbitrarily (and always with longer method names to type 8^). So even if the conversion tool does most of the grunt work, it will take a while before I can read the results without referring constantly to online help. In practice, the three of us who switched to Delphi were programming fluently within a month, using good OOP techniques. (And I had never used Pascal before, not even in school.) The user manuals are not great, but the context-sensitive online help is decent. We had figured that relearning OPENSTEP would take approximately the same amount of time. So that aspect of the transition was a wash. Ultimately, what counts in this business are results. Are you delivering on time, are you staying close to budget, are you proud of what you're writing, and will it be easy to maintain after deployment? I can honestly answer all of these questions with a "yes" for both NEXTSTEP and Delphi. As I mentioned the other day, I think the determining factor in our success is that the app is being written with objects either way. In that context, the differences between the development environments seem incidental to our basic rate of progress. If I'm right, then the NEXTSTEP advantage we have all come to know and love in the comp.sys.next hierarchy is about to face its most substantial challenge yet. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: soren@zeus.datashopper.dk (Soren Mathiasen) Newsgroups: comp.sys.next.programmer Subject: NeXT programing ABC Date: 1 Mar 1996 16:37:34 +0100 Organization: DataShopper Danmark, Copenhagen, Denmark Message-ID: <MFHNxcm0fpic089yn@zeus.datashopper.dk> Hi... Is there a good book for a Newbie in NS programming. I'm experienced i C. Thanks, Soren
From: soren@zeus.datashopper.dk (Soren Mathiasen) Newsgroups: comp.sys.next.programmer Subject: COMM probs Date: 1 Mar 1996 16:37:38 +0100 Organization: DataShopper Danmark, Copenhagen, Denmark Message-ID: <XMHNxcm0fRed089yn@zeus.datashopper.dk> Hi. How do I write to my com2 using C source. I have tried to open the port as a file, and then used fprintf to write down to it. didn't work. Please Help.... -- Soren Mathiasen soren@datashopper.dk Student at Roskilde EDB skolde, DK
From: crath@bnr.ca (Christopher Rath) Newsgroups: comp.sys.next.programmer Subject: Re: [q] perl5, where to put it? Date: 01 Mar 1996 15:33:34 GMT Organization: Bell Northern Research, Ottawa, Canada Message-ID: <CRATH.96Mar1103334@bmerhe83.bnr.ca> References: <7xravdydkz.fsf@burrow.muc.de> In-reply-to: markus g's message of Thu, 29 Feb 1996 22:25:00 GMT One of perl's problems, that it uses hard-coded paths to find its files, is also one of its strengths. We have overcome this perl4/perl5 problem here within my company by installing perl4 in its own directory tree (/opt/corp/local/perl4) and perl5 in its own tree (/opt/corp/local/perl5). We then install a couple of symlinks to actually put perl into place, ln -s /opt/corp/local/perl4/bin/perl /usr/bin/perl4 ln -s /opt/corp/local/perl5/bin/perl /usr/bin/perl Both perls now reside peacefully together on the system, and perl5 is the default perl used by perl scripts. We fell upon this idea because this is basically how perl installs itself if you give an install location different than /usr/local. Christopher -- === Christopher Rath ===== crath@bnr.ca ===== (613) 765-3141 === Northern Telecom Ltd. | Box 3511, Station `C' | ``Hydrogen is a colourless, odourless Ottawa, ON K1Y 4H7 | gas which, given enough time, turns FAX: (613) 763-4101 | into people.'' --- Henry Hiebert
From: Bob Lunney - Imonics Development <bob.lunney@imonics.com> Newsgroups: comp.sys.next.programmer Subject: RE: NeXT programming ABC Date: Fri, 01 Mar 1996 13:21:10 -0500 Organization: Imonics Corporation Message-ID: <31374016.39EF@imonics.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Check out http://www.stepwse.com/Resources/Books/index.html. For beginners, I recommend Alex Nghiem's book - it explains everything in readily understood terms. Good luck and welcome to the fold! -- Bob Lunney | Building tomorrow's legacy Imonics, Inc. | systems today. (tm) bob.lunney@imonics.com | #include <std_disclaimer.h>
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: precomp differences? Date: 1 Mar 1996 20:25:23 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4h7mfj$hk3@stc06.ctd.ornl.gov> How does one fix it so that we don't get these messages and use the precompiled headers? /NextDeveloper/Headers/appkit/nextstd.h:14: warning: could not use precompiled header '/NextDeveloper/Headers/bsd/libc.p', because: /NextDeveloper/Headers/appkit/nextstd.h:14: warning: 'stdio.h' has different date than in precomp J. W. Wooten
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: Re: How to update the resolver in NS 3.0? Date: 1 Mar 1996 23:11:13 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4h806h$sm6@news.its.com> References: <kkhgwa4tz3.fsf@hubert.fukt.hk-r.se> teddy@hubert.fukt.hk-r.se (Teddy Hogeborn) wrote: > I use some features in our local nameserver setup that requires > resolvers to understand CNAMEs in reverse lookups. [ ... ] > So how do I replace the resolver routines with the new versions? What > part of the online manual should I read? I found nothing simple in it > about creating shared libraries. Unfortunately, you are simply out of luck. You can't replace the resolver routines in the shared libraries, because NeXT does not provide any means of creating or changing their shared libraries. You can try linking with the "-m" option to allow multiply defined symbols, but I've tried and never gotten the resolver routines to work correctly under such circumstances. If you had a newer version of NEXTSTEP than 3.0, you could compile with "cc -posix" and the libresolv.a from the BIND distribution; this combination works. However, you can't use things like the AppKit with "cc -posix", and NeXT's POSIX routines have some severe bugs, so that doesn't get you very far either. Sorry. Go complain to NeXT. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
Newsgroups: comp.sys.next.programmer From: rfenney@netcom.com (Robert J. Fenney) Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Message-ID: <rfenney-010396153944@rfenney.slip.netcom.com> Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Sender: netnews@mork.netcom.com Organization: FenTek References: <31324223.12747679@news.bogo.co.uk> <4gvc9o$osh@shelob.afs.com> Date: Fri, 1 Mar 1996 23:38:10 GMT In article <4gvc9o$osh@shelob.afs.com>, Greg_Anderson@afs.com (Gregory H. Anderson) wrote: Greg, I have to save that you are doing a great job in presenting your case! The stuff in this thread is awsome! Thanks for taking the time. Robert
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Sat, 02 Mar 1996 00:12:56 GMT Organization: Orbital Computer Consultancy Message-ID: <3137870c.13652991@news.plsys.co.uk> References: <4h1viu$m1@shelob.afs.com> <4h407f$gpe@lion.embl-heidelberg.de> <31366EC0.3F87@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> wrote: >I'll let Greg speak for himself, but I feel a need to respond... > >Georg Tuparev wrote: >> >> In article <4h1viu$m1@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) >> writes: >> [...] >> > As I said, I raised this point only to demonstrate that even NeXT, whose >> > programmers I respect enormously, botched this first time out, and a lot >> > of early programmers either learned a bad thing or spent too much time >> > working around the deficiencies. Delphi at least got the DB part right, >> [...] >> At the time when DBKit was out, it was the coolest R/DB access framework around. >> It's not a shame when other people write something better later. But it's a shame if >> other people write _worse_ stuff later, and that is what is continually happening. >> NS is already 8 years old, and even NS0.8 was better than most OO environments are >> now. >> And another point: if you are saying that it's very horrible to change/rewrite >> sources you better work something else. > > >The point (I think) was this: Mr. Harbour claimed that NeXTSTEP encourages good >design. > >DBKit points out the fallacy in this: Actually, it doesn't. If you encourage someone to do something, and they don't do it, it may be because they were unwilling or unable to do it. Failure on someone's behalf to do something is certainly not proof that you didn't encourage them. >NeXT themselves, with the advantage >of having massive brain power, source code, and the designers of NeXTSTEP >itself, STILL produced a piece of software which is widely recognized as >a suboptimal design. > DBKit was a poor design, as Next themselves have admitted. But what does DBKit have to do with the AppKit or any other part of NextStep? It was designed from first principles, and, yes, Next got it wrong. At a guess, I'd say DBKit and the AppKit were designed by different teams in Next. My point is that the design principles of the AppKit and EOF encourage / enable an application developer to create well-designed applications. Perhaps enable is a better word. <-- snip -->
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: Sat, 02 Mar 1996 00:13:01 GMT Organization: Orbital Computer Consultancy Message-ID: <3137873e.13702266@news.plsys.co.uk> References: <4h78t5$87n@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >This has been fascinating discussion. I'd like to wrap up a few points >before the current set of arguments turns circular (as they inevitably >will, this being Usenet). > >First, for those of you who read my initial comparison as an emotional >attack on NEXTSTEP, I only have one thing to say: your prejudices are >showing. Read it again. I brought up plenty of places where each tool has >advantages and disadvantages relative to the other. I'm sorry if I burst >any bubbles about how fast the competition is catching up to NeXT, but we >have to live in the real world. If Win32 is your target, Delphi is real. >OPENSTEP/NT will have to compete against it. No, it does not possess every >bell and whistle NeXT offers, but for what my company does -- fairly >sophisticated MCCA Wall Street trading apps -- it is more than sufficient. >As it relates to AFS, this is an unarguable fact. Please don't try to >change my mind by citing a list of features I can prove I don't need. > This is completely fair Greg, you're saying Delphi was good enough for your app. But where I differ with you is in the following statement: >Delphi will be the better choice (for PC development) even after OPENSTEP shows up, I think. >Once you suss out the Delphi RTTI system (which admittedly needs better >documentation, but we did it), there is almost nothing you can't do in >the way of dynamic object programming ala NEXTSTEP. Here, you seem to have removed the qualification about suitability for _your_ requirements and made a completely general statement. There are many areas in which OpenStep has a clear technical lead over Delphi. True enterprise objects, a proven DO system, pseudo-garbage collection, delegation, notification, single imaging model to name a few. Okay, you don't need these features, but the mileage of others may vary. Even Wall Street is not the entire world. I also feel that Delphi's SQL transaction-handling is a kludge, and may be inadequate in certain application scenarios. The one advantage Delphi does have is the upfront cost being lower. But this can often be a short-sighted metric. >Second, I _do_ have a prejudice. In an ideal world, NeXT would have done a >better job of marketing a great development environment, and I wouldn't >have had to consider working on Windows. Nothing would have made me >happier. Amen to that! >Alas, things didn't turn out that way. So we had to ask ourselves >what was the best way to get to Windows: wait for OS/NT, or find an >existing native OOP tool? With the direction of a customer, we chose the >latter, because NeXT could not provide timely answers for things like >runtime fees, minimum hardware/RAM configuration, and deployment dates. > >I have never regretted we chose Delphi. We are going to get to deployment, >we are going to do it in approximately the same timeframe, and we did not >have to compromise our existing OOP design. My 'NEXTSTEP-spoiled' >programmers have not mutinied; in fact, they are quite happy. The customer >will be running the app before OS/NT even makes it to beta testing. It >will run comfortably on a 16MB Pentium machine running NT 3.51, talking to >a Sybase server. It will not require a runtime fee. Call me crazy, but I >don't see the downside. > >Georg keeps pestering about OPENSTEP. I have referred to it only minimally >for a simple reason: OPENSTEP is not real yet. You cannot buy it. It's an >interesting spec, and I know that at least three organizations (including >the GNU gang) are working on implementations, but it is not deployable at >this time. I have opinions about if and when each of those parties is >likely to ship a product, but I will not share them, because they are >based on information I am not entitled to disclose. They'll get here when >they get here. In the meantime, Delphi 2.0 is already in customer hands, >so it is fair to talk about it. > >A troubling aspect of OPENSTEP is the amount of work it will take to get >there from NEXTSTEP. Yes, I know there will be automated tools to do much >of the grunt work, but for big apps, there will still be a lot of cleanup >and retesting. Furthermore, the name space and class hierarchies are >different in significant ways. > Well, Next do like to go with the best stuff they've got. >When we were making our original decision, one point that did NOT work in >OPENSTEP's favor was training time. That may sound surprising, given how >many experienced NEXTSTEP programmers we employ. But when you look at the >OPENSTEP spec, there's a lot of new stuff to learn about classes where our >programming reflexes now run on auto-pilot. Five years into it, I can >recite almost all of the AppKit methods in my sleep. I almost never refer >to help while I'm coding. But in OPENSTEP, it seems like everything >changed, sometimes arbitrarily (and always with longer method names to >type 8^). So even if the conversion tool does most of the grunt work, it >will take a while before I can read the results without referring >constantly to online help. > After a year of using the great stuff in the Foundation Kit, I can't imagine having to go back to the old 3.2 API. NSMutableArray vs List, NSMutableDictionary vs HashTable, NSString vs char * , autorelease vs free - Foundation wins hands down every time for me. >In practice, the three of us who switched to Delphi were programming >fluently within a month, using good OOP techniques. (And I had never used >Pascal before, not even in school.) What about all those begin / ends all over the shop? Don't you miss C's terseness ;-) <-- snip -->
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Sat, 02 Mar 1996 00:12:53 GMT Organization: Orbital Computer Consultancy Message-ID: <31379259.16545756@news.plsys.co.uk> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4gi8ei$ekv@cheyenne.iac.net> <31324223.12747679@news.bogo.co.uk> <313259A3.5433@ix.netcom.com> <31338a40.12358521@news.plsys.co.uk> <3133CC9B.59BF@FreemanSoft.com> Joe Freeman <Joe@FreemanSoft.com> wrote: >Karl Harbour wrote: >> And why did Garfinkel & Mahoney and the people who wrote the >> NeXTDeveloper/Examples do it that way? Because of target/action and >> the design of the AppKit generally separating UI from application >> logic. > >This is definately not true. The target/action paradigm has done >nothing to aid in separating out the application logic. > Applications have several different types of process, presentation, application, >and model to name the most common. I haven't the faintest idea what this last sentence means. Are there some words missing? > The insertion of controllers into >nib files where those controllers implement UI control logic and >application logic is a "bad thing" IMHO. It makes it very difficult >to deploy an application across presentation layer architectures.. > Couldn't find "presentation layer architecture" in my buzzword dictionary. What are you on about? <-- snip --> >Almost all >of the subsystems you mention are really abstractions for system resources >and are basicly consumable objects. What? > (printPSCode and faxPSCode are actually interface protocols) Thanks for the quick lesson on the AppKit API. I always thought printPSCode was a method in the View class. > So, NeXT did a nice job wrapping their >internally developed resources which is good. It is not clear what >portion of the development time that these resources represent What is not clear is what the hell you are on about. developer" package is sufficient for use with local or ODBC drivers. > >>Even the C/S version is only $1995. > >> > >>Compare that to NS, which costs $795 for the User version, plus $4,995 > >>for the Developer edition, plus $495 for EOF, plus... And your description > >>completely ignores runtime fees. NS has them. Delphi doesn't. > >> > >>-- > >>Gregory H. Anderson | "Honey, there're few programming > >>Gaffer/Best Boy/Key Grip | problems that can't be solved > >>Anderson Financial Systems | with duct tape." -- 'Father' Duke > >>greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95 > > > Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer > ----------------+---------------------+--------------------- > I know you're an optimist if you think I'm a pessimist.
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: sccs programs for next Date: 25 Feb 1996 14:19:17 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4gpr55$295@news.its.com> References: <4gl9vf$npt@noc.tor.hookup.net> kboyes@noc.tor.hookup.net (Kevin Boyes) wrote: > Is sccs available for NeXT? I found the front end, sccs proper, > from a 4.4BSD distribution but the supporting proograms (admin, get, etc.) > were not there. > > Free programs would be best but I am open to for-pay solutions. SCCS is a commercial product of SMI. RCS and CVS from ftp://prep.ai.mit.edu/pub/gnu/ are freely available, upwards compatible version management software packages. Or you could get DevMan from VNP, which is an RCS plus a native NEXTSTEP GUI interface. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: espeyton@ix.netcom.com (Eric Peyton) Newsgroups: comp.sys.next.programmer Subject: Why does NXColorWell hang on a drag Date: 2 Mar 1996 01:26:27 GMT Organization: Netcom Message-ID: <4h8843$lrn@dfw-ixnews1.ix.netcom.com> I recently moved my system from 8bit to 16bit color. Everyhting seems to work well (and look better). However, I have an application that whenever I drag a color from the NXColor dialog onto an NXColorWell hangs the system for about 2 minutes. The color is never transferred to the color well. If I click on the border of the color well and change the color that way everything is fine. This problem did not occur in 8 bit color mode. What gives??? Why can't I drag colors in my app??? Any suggestions would be appreciated. Thanks in advance, Eric Peyton espeyton@ix.netcom.com
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 2 Mar 1996 13:11:00 GMT Organization: University of Ghent, Belgium Message-ID: <4h9hd4$qg4@infoserv.rug.ac.be> References: <4guqt5$5nd@ixnews3.ix.netcom.com> <4gviqf$dh@news.next.com> ngervae@sirius.com wrote: : I really must recommend that you avoid the Indexing Kit. It was a : decent idea, but the engineer had more work than he could do, and : it hasn't been properly supported since it was released. It isn't : going to evolve, improve, or be fixed, so best keep away.... I agree. I did a few years ago some tests with adding Computer Algebra Kit objects to Indexing Kit objects. Actually was trying to make them compatible because that's what the idea is about, no ? I think it would be extremely "cool" if you could use big integers etc. as keys and so on for small databases, in the same way that you can use Sound or Image objects. But I gave up after a while. The IXKit API heavily uses protocols, which is an unproven concept after all, instead of consisting of just a few, clean, powerful Objective C classes... And this "serialization of instance variables", that probably lies at the root of how archiving is done in the Foundation Kit, is really over-optimized. You can write a "non-efficient" strong public interface instead, and then do the optimizing _internally_ instead. By effectively converting into some other data structures, instead of making all these specialized data structures public. But these things take time of course... If this would happen, then IXkit would be a great tool, since I have a feeling that there's a lot of theory hidden in it; the "C" side is OK, it's just the Smalltalk side that is lacking... Could anyone give me a hint on where to look for a book or so on this "Boyer-Moore" compiling, matching etc. ?? David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 2 Mar 1996 13:24:36 GMT Organization: University of Ghent, Belgium Message-ID: <4h9i6k$qg4@infoserv.rug.ac.be> References: <4guqt5$5nd@ixnews3.ix.netcom.com> <4gviqf$dh@news.next.com> <4h2qp7$2ll@news.xmission.com> don@misckit.com wrote: : 1) The GNUStep people create a GNU IXKit from scratch that both : implements the old API and works right. We're not talking about bugs here; it's the API that matters. If it implements the old API,well... Algorithms you can improve later on, you can make things faster etc. What you can't do, is change, once it's released, since then people will depend upon it etc. So the IXkit shows how important it is to do good design beforehand... But I agree with you that it would be nice if someone would come up with a set of objects to _do_ exactly what IXkit was meant to do. But it would really be important to take a critical look at the API, to make half or more of the stuff private, and to retain only a few public classes. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: ObjectError Date: 2 Mar 1996 13:38:09 GMT Organization: University of Ghent, Belgium Message-ID: <4h9j01$qg4@infoserv.rug.ac.be> References: <4h1vnv$85c@transfer.stratus.com> Charles_Spitzer (charlie@why.az.stratus.com) wrote: : Anyone have this rewritten as a poseAs: for NSObject? I tried doing so, but : found that NSObject doesn't have Object's +error: or -error methods. Can one : just eliminate these methods? Is this true ? I wouldn't be surprised. Wouldn't it be simpler to just make NSObject a subclass of Object, and inherit these methods ? I really don't understand NeXT on the issue. It would save work ! I don't see why Object would be less portable than NSObject. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Date: 2 Mar 1996 13:59:24 GMT Organization: University of Ghent, Belgium Message-ID: <4h9k7s$qg4@infoserv.rug.ac.be> References: <4h78t5$87n@shelob.afs.com> Gregory H. Anderson (Greg_Anderson@afs.com) wrote: : Georg keeps pestering about OPENSTEP. I have referred to it only minimally : for a simple reason: OPENSTEP is not real yet. You cannot buy it. It's an Never mind Georg, Greg. When I contacted him, to ask to place a link towards the Computer Algebra Kit site on the eduStep server, which I thought basically to be a good initiative, the answer was that my use of : #ifdef __NeXT__ - read:(NXTypedStream *)aStream - write:(NXTypedStream *)aStream #endif ,to be able to port my software to NeXT ObjC, Stepstone ObjC and GNU ObjC, was not OpenStep compliant... Hence no link, I guess. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: COMM probs Date: 2 Mar 1996 14:18:19 GMT Organization: Nanyang Technological University Message-ID: <4h9lbb$u7s@ntuix.ntu.ac.sg> References: <XMHNxcm0fRed089yn@zeus.datashopper.dk> Soren Mathiasen (soren@zeus.datashopper.dk) wrote: : How do I write to my com2 using C source. I have tried to open the port : as a file, and then used fprintf to write down to it. didn't work. THis is a C-source to access the port in raw mode. ---------------------- /* * include files */ #include <stdio.h> #include <stdlib.h> #include <strings.h> #include <sys/file.h> #include <sys/ioctl.h> /* * defines */ #define TRUE 1 #define FALSE 0 #define TTY_DEVICE "/dev/ttyb" #define STR_LEN 255 /* * main */ extern int main () { int ttyfd; struct sgttyb ttypars; int ttyrd; int ttywr; char ttystr[STR_LEN]; long avail; if ((ttyfd = open (TTY_DEVICE, O_RDWR)) <= 0) { printf ("Tty232: can't open %s\n", TTY_DEVICE); exit (-1); } else { if (gtty (ttyfd, &ttypars) < 0) { printf ("Tty232: can't get port parameters\n"); close (ttyfd); exit (-1); } printf ("ttypars.ispeed = %d\n", ttypars.sg_ispeed); printf ("ttypars.ospeed = %d\n", ttypars.sg_ospeed); printf ("ttypars.erase = %02X\n", ttypars.sg_erase); printf ("ttypars.kill = %02X\n", ttypars.sg_kill); printf ("ttypars.flags = %02X\n", ttypars.sg_flags); ttypars.sg_flags |= (RAW); ttypars.sg_flags &= ~(ECHO); if (stty (ttyfd, &ttypars) < 0) { printf ("Tty232: can't set port parameters\n"); close (ttyfd); exit (-1); } while (TRUE) { ioctl (ttyfd, FIONREAD, &avail); if (avail > 0) { ttyrd = read (ttyfd, ttystr, STR_LEN); ttywr = write (ttyfd, ttystr, ttyrd); putc ('.', stdout); printf ("%d", avail); fflush (stdout); } else { putc ('.', stdout); fflush (stdout); } sleep(1); } close (ttyfd); } return; } ------------------------------ Hope this helps, WIllem
From: pedja@mina.TipTop.COM (Pedja Bogdanovich) Newsgroups: comp.sys.next.programmer Subject: WebObjects: request header fields Followup-To: comp.sys.next.programmer Date: 02 Mar 1996 15:30:54 GMT Organization: TipTop Software Distribution: world Message-ID: <PEDJA.96Mar2103054@mina.TipTop.COM> Currently, WOF doesn't provide any info from the HTTP request headers. What do you think how useful it is to have this info available? For example, since WOF is used to generate "dynamic HTML", it makes sense to be able to generate HTML based on what the client browser is (e.g., to take advantage of some specific browser features). Cheers. Pedja -- ===================================================================== | Spoken: Pedja Bogdanovich Phone: +1-301-656-3837 | | TipTop Software USPS: 20824-0681, USA Fax: +1-301-656-8432 | | URL: http://www.tiptop.com/ Email: pedja@tiptop.com | =====================================================================
From: markdo@myriad.net (Mark A. Doucet) Newsgroups: comp.sys.next.programmer Subject: Panasonic Power Drive Date: 1 Mar 1996 22:39:53 GMT Organization: DigiPhone Corporation, Bryan/College Station Texas 409-693-8885 Message-ID: <4h7ubp$aao@news.myriad.net> I am trying to install a Panasonic PowerDrive onto my black hardware running 3.2. The drive is a SCSI multi-function drive that operates as a Quad speed CDROM and as a read/write optical drive. The CDROM portion seems to work but not without the error below: Target 6: BUSY; retry 1 other retries Target 6: BUSY; retry 49 sd2 (6,0): ERROR op:0x0 sd_state:4 scsi status:0x8 Target 6: BUSY; retry 1 probing for DOS probing for CDROM CDROM File System: LOADED CDROM File System: Initialized Mar 1 16:10:49 Workspace: Mounted scsi disk However, the read/write portion chokes up after trying to initialize the media. I am assuming that this is because of the lack of a properly formatted disktab entry. I have most of the specifications on the drive but am not sure I have all of what I need to create something that is workable. I have also gone through the next answers but they say nothing about optical drives other than the old Canon drive. If anyone has had any experience in making this drive work on black or white hardware I would greatly appreciate any info on the subject. Also if anyone thinks they have the wherewithall to configure this thing so that it works properly I would be glad to pay them for their time. Thanks -- Mark A. Doucet Dominion Technologies, LTD. "Solutions for the Next Generation" markdo@mail.myriad.net 409.778.3615
From: izumi@pinoko.berkeley.edu (Izumi Ohzawa) Newsgroups: comp.sys.next.programmer Subject: Re: Panasonic Power Drive Date: 2 Mar 1996 22:08:02 GMT Organization: University of California, Berkeley Distribution: world Message-ID: <4hags2$ien@agate.berkeley.edu> References: <4h7ubp$aao@news.myriad.net> In article <4h7ubp$aao@news.myriad.net> markdo@myriad.net (Mark A. Doucet) writes: >I am trying to install a Panasonic PowerDrive onto my black hardware >running 3.2. The drive is a SCSI multi-function drive that operates as a >Quad speed CDROM and as a read/write optical drive. The CDROM portion >seems to work but not without the error below: > >Target 6: BUSY; retry 1 >other retries >Target 6: BUSY; retry 49 >sd2 (6,0): ERROR op:0x0 sd_state:4 scsi status:0x8 >Target 6: BUSY; retry 1 >probing for DOS >probing for CDROM >CDROM File System: LOADED >CDROM File System: Initialized >Mar 1 16:10:49 Workspace: Mounted scsi disk > >However, the read/write portion chokes up after trying to initialize the >media. I am assuming that this is because of the lack of a properly >formatted disktab entry. I have most of the specifications on the drive >but am not sure I have all of what I need to create something that is >workable. I have also gone through the next answers but they say nothing >about optical drives other than the old Canon drive. If anyone has had >any experience in making this drive work on black or white hardware I >would greatly appreciate any info on the subject. Also if anyone thinks >they have the wherewithall to configure this thing so that it works >properly I would be glad to pay them for their time. > >Thanks >-- >Mark A. Doucet >Dominion Technologies, LTD. >"Solutions for the Next Generation" >markdo@mail.myriad.net >409.778.3615 I have no experience with this drive, but... There was a thread on Panasonic/Matsushita PD in a Japanese NeXT newsgroup fj.sys.next which you probably cannot read. But, I include the article below anyway, because it contains a disktab for PD that is attributed to kiura@narc.affrc.go.jp (Takuji Kiura). Briefly, the first line of disktab includes "MATSHITA PD-1 LF-1000 A" which is noted in the alert logged in Console from Workspace when it fails to format from menu. That is also displayed at boot time. Since Matsushita is Panasonic here, you may get a different message in the Console. Find that out and change that portion of the first line in the disktab. With that disktab, it is reported that you can format PD from Workspace. Good luck, and if this works, please post the correct distab for the US version of Panasonic PD. -- Izumi Ohzawa <izumi@pinoko.berkeley.edu> [ $@Bg_78^=;(J ] USMail: Univ. of California, 360 Minor Hall, Berkeley, CA 94720-2020 Tel: 510-642-6440, Fax: -3323, Web: http://totoro.berkeley.edu/~izumi/ Forwarded Article from fj.sys.next follows: ====================================================================== Newsgroups: fj.sys.next From: hrst@papageno.rim.or.jp (Hiroshi TOMARU) Subject: Re: Booting from PD Message-ID: <Dn5H7w.38A@papageno.rim.or.jp> Organization: Independent researcher, Takasaki, Gunma, JAPAN References: <DMx072.HvK@vega.aichi-u.ac.jp> Distribution: fj Date: Wed, 21 Feb 1996 23:57:31 GMT Lines: 57 In article <DMx072.HvK@vega.aichi-u.ac.jp> arisawa@vega.aichi-u.ac.jp (Kenji Arisawa) writes: : $B$7$+$7Ha$7$$$+$J!";d$N(JSCSI$B%I%i%$%P$O(JAIC6X60$B$G!"$3$l$,2?$H(J300KB$B$b$"$k$N$G$9!# (J : $B$;$a$F!"(Jboot$B;~$K(J(installer$B$,9T$C$F$$$kMM$K(J)$B$"$H$+$i(JSCSI $B%I%i%$%P$rDI2C$G$-$l(J : $B$P2?$H$+$J$k$N$G$9$,!"$d$j$+$?$,J,$+$j$^$;$s!#(J $B$*$=$i$/!"4{$K;n$5$l$?$+$H$O;W$$$^$9$,(J... FD$B$N(J/usr/Devices/System.config/Instance0.table $B$G(J ======================================================================== "Install Mode" = "Yes"; "Installation Driver Families" = "Disk SCSI"; ======================================================================== $B$H$9$l$P!"%$%s%9%H!<%k%b!<%I$K$J$C$F!"(J $BNc$N<ALd$r$7$F$/$k$+$b$7$l$^$;$s!#(J $B$"$H!"(J ======================================================================== "Boot Drivers" = "PS2Keyboard Adaptec1542B DPT2012 IDE"; ======================================================================== $B$N%U%#!<%k%I$r(J K/B$B$H(JAIC6X60$B%I%i%$%P$@$1$K$7$F!"(JFD$B>e$N%I%i%$%P(J $B$b$3$N(J2$B$D$K9J$k$3$H$b2DG=$+$b$7$l$^$;$s!#(J $B85$N5-;v$KLa$j$^$9$,!"(Jdisktab$B$N$H$3$m$G!"(J1$B9TL\$rJQ99$7$F!"(J ------------------------------------------------------------- disktab # # pd634: 5.25" SCSI PD disk for NEXTSTEP # pd634|PD634|MATSHITA PD-1 LF-1000 A:\ :ty=removable_rw_scsi:ns#64:nt#64:nc#634:ss#512:\ :fp#256:bp#256\ :os=sdmach:z0#80:z1#168:ro=a:\ :pa#0:sa#1298176:ba#8192:fa#1024:ca#3:da#4096:ra#5:oa=space:\ :ta=4.3BSD:ia:aa: --------------------------------------------------------------------- $B$N$h$&$K$9$k$H!"(JWorkSpace$B$N%a%K%e!<$+$i%U%)!<%^%C%H$G$-$k$H$$(J $B$&>pJs$r!"(J kiura@narc.affrc.go.jp($BLZ1:Bn<#$5$s(J)$B$+$i!"$$$?$@$-(J $B$^$7$?!#$"$j$,$H$&$4$6$$$^$9!#(J"MATSHITA PD-1 LF-1000 A"$B$H$$$&(J $B$N$O!"(JWorkSpace$B$,%U%)!<%^%C%H$K<:GT$9$k;~$K%3%s%=!<%k$KI=<($5(J $B$l$k%a%C%;!<%8$@$=$&$G$9!#3N$+!"%9%?!<%H%"%C%W;~$K$b!"I=<($5$l(J $B$F$$$^$9$M!#(JWorkSpace$B$N%a%K%e!<$+$i%U%)!<%^%C%H$G$-$k$H$9$l$P!"(J BuildDisk$B$b2DG=$+$b$7$l$^$;$s!#6u$N%a%G%#%"$rGc$&5!2q$,$G$-$?(J $B$i!"(JBuildDisk.app$B$r(JCD$B$+$i(JHD$B$K%3%T!<$7$F!"(J BuildDisk.app/BuildDisk$B$r(Jsetuid (# chmod u+s BuildDisk.app/BuildDisk) $B$7$F!";n$7$F$_$?$$$H;W$$$^$9!#(J In article <DMsHo7.1DF@papageno.rim.or.jp> I wrote: : $B$b$7!"(JHD$B$N(Jroot device$B$,(J630MB$B$h$j$bBg$-$$>l9g$O(J....Uhmmm... tar : $B$d(Jfind|cpio$B$J$I$r;H$C$F%3%T!<$7$F2<$5$$!#(Jgnutar$B$O!"(Jdevice : special file$B$b%3%T!<$G$-$k$N$+$J(J? hyodo@youkoso.wincnet.or.jp($BJ<F#$5$s(J)$B$+$i!"%a!<%k$r$$$?$@$-$^(J $B$7$F!"(Jgnutar$B$,%P%s%I%k$5$l$F$$$k$3$H!"$=$7$F!"(J-p$B%*%W%7%g%s$G(J device special file$B$,%3%T!<$G$-$k$3$H$r65$($F$$$?$@$-$^$7$?!#(J $B$"$j$,$H$&$4$6$$$^$9!#(J --- Hiroshi TOMARU ($BEP4]9@(J) <- Japanese Kanji hrst@papageno.rim.or.jp (accepts NeXT-type mail < 100KB)
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: ObjectError Date: 2 Mar 1996 23:15:26 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4hakqe$5d@news.its.com> References: <4h1vnv$85c@transfer.stratus.com> charlie@why.az.stratus.com (Charles_Spitzer) wrote: > Anyone have this rewritten as a poseAs: for NSObject? I tried doing so, but > found that NSObject doesn't have Object's +error: or -error methods. > Can one just eliminate these methods? @Begin(SHAMELESS_PLUG) { CrashCatcher 2 understands NSObject and the Foundation kit. You are free to work on ObjectError or CrashTrap if you prefer those, of course, but CrashCatcher is a lot more powerful and has a lot more capabilities. }@End -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: Ken Foust <fousto@neta.com> Newsgroups: comp.sys.next.programmer Subject: sis 6205 video chip HELP HELP HELP Date: 3 Mar 1996 01:10:00 GMT Organization: Internet Access, Chandler Arizona Message-ID: <4harh8$gob@trojan.neta.com> Maybe you can help? I recently purchased a asus motherboard with the on board video chip sis6205. I can get it to boot in mono but it won't boot in color. I have disable the boot graphics as was the problem with the triton. I have tried using several drivers. I can get it to detect the trios and set the resolution. when boot goes to the login window the screen goes blank. some drivers will give a color strip at the bottom of the screen. It seems so close that just maybe some tweak in the instance O file would work. It works very well with win 95 and the color is absoutley great. I miss my nextstep. If you have any ideas I would love to hear. I have the technical manual from sis but I am not competent enough to write a driver. If someone does this will be a great motherboard for nextstep. It has the best color I have seen under win 95. my system is as follows asus 55sp3av 100 mz pentium 32 mg ram ncr pci scsi ns 3.3 ns devel 3.2 Thanks a lot Ken Foust Phoenix Arizona
From: name Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Java & NEXTSTEP Date: 3 Mar 1996 05:13:19 GMT Organization: org Message-ID: <4hb9pf$q34@thales.nmia.com> I have to do some Java development and I'd like to know if there are any Java-aware browsers running under NEXTSTEP. Thanks.
From: droux@nmia.com (Nicolas Droux) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Java & NEXTSTEP Date: 3 Mar 1996 05:18:28 GMT Organization: New Mexico Internet Access Message-ID: <4hba34$q34@thales.nmia.com> [I just realized that the headers of my previous message were not initialized correctly. Here we go again...] I have to do some Java development and I'd like to know if there are any Java-aware browsers running under NEXTSTEP. Thanks.
From: reich@cip.mathematik.uni-wuerzburg.de (Stefan Reich) Newsgroups: alt.unix.wizards,comp.sys.next.programmer,comp.unix.internals,comp.unix.misc,comp.unix.programmer,comp.unix.questions Subject: Q: Problem redirecting stdout to a socket Date: 3 Mar 1996 12:33:09 GMT Organization: University of Wuerzburg, Germany Message-ID: <4hc3i5$omn@winx03.informatik.uni-wuerzburg.de> Hello world! I have a problem connecting two programs via UNIX stream sockets using Nextbsd 4.3. Program A opens two sockets, binds() and listens(). It sets up the asynchronous IO facility that selects() the sockets and finally accepts() the incoming messages. It then does a SUN RPC call to program B that: - forks a child which * connects to the supplied portnumbers of program A (close on exec Flag is clear) * substitutes stdout/stderr with the two socketdescriptors * execs the desired program So far, so good. But: the program started at last cannot write to stdout/stderr and receives errno=49=EADDRNOTAVAIL <=> "Can't assign requested address " Why? Thank you for any help! Please reply by Email. Stefan ------------------------------------------------------------------------- Email: 1reich@informatik.uni-hamburg.de (Plain, MIME) or reich@cip.mathematik.uni-wuerzburg.de (NeXTMail)
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: precomp differences? Date: 3 Mar 1996 19:34:08 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4hcs7g$gh8@news.its.com> References: <4h7mfj$hk3@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. Wooten) wrote: > How does one fix it so that we don't get these messages and use the > precompiled headers? Run 'cc -precomp' as root on the precompiled header files to rebuild them. By the way, it's generally not a good idea to change the system header files.... -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Fri, 01 Mar 1996 23:37:57 -0500 Organization: Steel Driving Software, Inc. Message-ID: <3137D0A5.5962@ix.netcom.com> References: <4h1viu$m1@shelob.afs.com> <4h407f$gpe@lion.embl-heidelberg.de> <31366EC0.3F87@ix.netcom.com> <3137870c.13652991@news.plsys.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Karl Harbour wrote: > > "Jonathan W. Hendry" <steeldrv@ix.netcom.com> wrote: > >DBKit points out the fallacy in this: > > Actually, it doesn't. If you encourage someone to do something, and > they don't do it, it may be because they were unwilling or unable to > do it. Failure on someone's behalf to do something is certainly not > proof that you didn't encourage them. This might make some sense if DBKit were written outside NeXT. NeXTSTEP cannot provide more 'encouragement' than NeXT itself is capable of. Which is essentially my point: The tool itself won't magically create good design. Using NeXTSTEP does not ensure good design, regardless of how much encouragement is found in the docs. The developers of NeXTSTEP have far more knowledge of OO design that could ever be captured in documentation, but they still managed to put out a suboptimal design. > > >NeXT themselves, with the advantage > >of having massive brain power, source code, and the designers of NeXTSTEP > >itself, STILL produced a piece of software which is widely recognized as > >a suboptimal design. > > > > DBKit was a poor design, as Next themselves have admitted. But what > does DBKit have to do with the AppKit or any other part of NextStep? > It was designed from first principles, and, yes, Next got it wrong. > > At a guess, I'd say DBKit and the AppKit were designed by different > teams in Next. > > My point is that the design principles of the AppKit and EOF encourage > / enable an application developer to create well-designed > applications. Perhaps enable is a better word. So, with those design principles handy, and the source, and easy access to the architects of the 'good parts', how did the DBKit happen? Surely, the designers of DBKit had access to far more expertise in NeXTSTEP software design than any of us plebes could ever hope to have. Yet they still managed to botch it somewhat. Face it, the tool does not enable good design. You can do good design with gcc on a linux box, or with java, or even with Basic. If you know what you're doing. No matter what tool you use, if you haven't got a clue about design, you're going to produce crap. NeXTSTEP may allow you to produce crap '10 times faster'. But it's still crap.
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 4 Mar 1996 05:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4hdu8u$2c0@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: awe@inhb.co.nz (Mark Derricutt) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Mon, 04 Mar 1996 12:12:42 GMT Organization: Chalice of Blood Message-ID: <313ad031.16464043@news.inhb.co.nz> References: <4gl3us$alj@shelob.afs.com> <4gq6ok$9hl@lion.embl-heidelberg.de> On 25 Feb 1996 17:37:24 GMT, tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) wrote: >curious how you can implement it without PDO. Well, you can use sockets on UNIX! >Have you used them? It's great fun indeed ;-) And what about W$? Don't tell me OLE >please! I prefer to go to work in a sheep farm but not using OLE! Call me a dill, or a New Zealander (lot of sheep here), but what is PDO??? >of Delphi's editor and form designer goes WAY beyond IB's limited parsing of header >files", but I'm talking about overall integration! What if I do not like the >Delphi's editor? With NS I can use any! With Delphi I have only one option... There are other 'programmers' editors available, which integrate directly with delphi, and work quite well. Also, I have successfully subclassed and made my own glyph browser, instead of the standard one. So I assume it should/could be possible to sub-class the editor and write your own... Not exactly what you mean, but the option is there.. Also, the ability to write your own experts which integrate with the IDE etc. etc. is very nice. Catch you on the reply, Mark "Alone, at last just you and me - Fight"
From: fox@jeans.fokus.gmd.de (Oliver Fox) Newsgroups: comp.sys.next.programmer Subject: color PostScrip to tiff problem Date: 4 Mar 1996 15:28:31 GMT Organization: GMD-FOKUS Message-ID: <4hf26v$aiq@stern.fokus.gmd.de> Keywords: PostScript tiff color Does anyone know how to convert a color postScript to a tiff ? i wrote a application what do this quit good for black/white "pictures" but with collor PostScripts there is a propblem: from the color information in the postScript for example black cian magenta yellow only the last color is convertet and the tiff only have the "yellow"information If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! Oliver Fox Tel.: +49 30 25 499 123 Fax: +49 30 25 499 202 E-Mail: fox@fokus.gmd.de (NeXT mail welcome)
From: fox@jeans.fokus.gmd.de (Oliver Fox) Newsgroups: comp.sys.next.programmer Subject: 2 bit tiff to 1 bit tiff Date: 4 Mar 1996 15:33:30 GMT Organization: GMD-FOKUS Message-ID: <4hf2ga$anr@stern.fokus.gmd.de> Keywords: tiff 1 bit Does anyone know how to convert a 2 bit tiff to a 1 bit tiff a silly archiving system from a client only accepts 1 bit tiffs If anyone could inform me of how this is done, or point me to a piece of code somewhere on the net that demonstrates this, I'd be grateful!! Oliver Fox Tel.: +49 30 25 499 123 Fax: +49 30 25 499 202 E-Mail: fox@fokus.gmd.de (NeXT mail welcome)
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 4 Mar 1996 15:57:39 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4hf3tj$2m1@news.its.com> References: <4guqt5$5nd@ixnews3.ix.netcom.com> <4gviqf$dh@news.next.com> <4h9hd4$qg4@infoserv.rug.ac.be> stes@cage.rug.ac.be (David Stes) wrote: > Could anyone give me a hint on where to look for a book or so on this > "Boyer-Moore" compiling, matching etc. ?? A decent bookstore that carries computer literature? A university bookstore (if the campus offers a computer science program, anyway) is a likely place. You're looking for books on machine languages, finite automata, and searching/matching algorithms. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: 4 Mar 1996 16:26:55 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4hf5kf$ti@shelob.afs.com> References: <31378720.13672300@news.plsys.co.uk> Karl Harbour writes > Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > > In NEXTSTEP, you can attach a Button to any "sender" method of any > > other object. But the only Button message you can attach is its > > 'click', the primary action of the Button. That's because the > > target->action paradigm is hard-wired, non-extensibly, into > > InterfaceBuilder. This assumes, of course, that you are using the > > standard classes. It is possible, with some effort, to build custom. > > IBConnectors for custom classes > > Surely an important point is that Delphi event handlers correspond to > two different paradigms in Nextstep, namely target/action and > delegation. Anything the Delphi event handlers can do can be handled > by one of these two paradigms as appropriate. Buttons have delegates? Please explain how, WITHOUT SUBCLASSING, I can get notified and then do something special about mouse in/out and drag enter/exit events in NEXTSTEP buttons. I can in Delphi. > Granted, there are some classes in the AppKit which are missing > desirable delegate methods, but this is a minor problem for an > experienced developer. By which you mean, anyone willing to write a subclass, add a 'delegate' ivar, create a delegate messaging protocol, and add default processing methods for these events. Which is already done for you in Delphi. > As I'm sure you're aware, the OpenStep version of Interface Builder > allows any Control to be easily subclassed, so you'll be able to roll > your own delegate methods. I think this is easier said than done. But to each his own. > Nevertheless, I agree IB could be better in this area in some ways. > For example, double-click actions should be assignable in IB. > But the key thing for me is that workrounds to these limitations > only require a few lines of code. Once you know what you're doing, have a firm sense of how NeXT does things, and trial-and-error your way around brain damaged classes, yes. I can do it. You can do it. But unless someone really knows what they're doing as an object architect, I don't recommend it. I have seen many poor examples of apparently simple override stuff. > Plus of course delegation gives you the unique "Can I do this?" / > denial functionality, which is absent from Delphi. This is a lie. OnCloseQuery, OnChange, and all BeforeXxxx methods of Tables -- just to name three that spring to mind -- provide exellent structured exception handling (or Boolean function returns) to deny continuation of a proposed operation. > And NSNotificationCenter (available now) gives you "To whoever wants > to know, I did this...", which is also absent from Delphi. Yeah, it's a new class in FoundationKit. Our apps have had such a class for _five_years_. If anyone in Delphi-land needs such a component, I'll be happy to sell it to them. It's just not fair to compare the native environments object-for-object. I can cite numerous counter-examples. So what? They are not 100% orthogonal, nor do they need to be. > Oh, and what's the equivalent of the Responder chain in Delphi? There isn't one. Again, so what? This happens to be how NeXT implemented a part of its control hierarchy. Explain how its absence cripples Delphi in some way for MCCA apps, because I haven't missed it. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Format and Documentation of eomodel files? Date: 4 Mar 1996 18:03:39 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4hfb9r$5av@stc06.ctd.ornl.gov> I'm creating a WebObject example that uses an NSDictionary for storage of the information. I've tested the NSDictionary with a piece of regular Objective C code and it works fine. I need to be able to create the eomodel file describing the NSDictionary data, but don't have Sybase or Oracle that will connect to the EOModeler. Where can I find the documentation on what to put in the eomodel file so that I can use it in an NextStep interface (i.e. with EOBrowser, and table view, etc.). Also when I try to add a method to my Application.wos file, I get the message that the WO'Application.wos does not have a selector for my method. I thought I added a selector by simply putting the definition of that "method" into the Application.wos script. -- - - - - - - - - - J. W. Wooten
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: NSScanner problems? Date: 4 Mar 1996 17:39:59 GMT Organization: University of Heidelberg, Germany Message-ID: <4hf9tf$rpc@sun0.urz.uni-heidelberg.de> I'm using an NSScanner to parse a text-only data file. It appears to work. Up to a point. After correctly scanning 32 lines into the text ( while ([myScanner isAtEnd]==NO) ), it stops. Interestingly, always at the same place. Curiouser, when I scan the original Mac file (other line terminating characters), it stops one line earlier. Always at the same place. This file should not be too big to choke NSString or NSScanner... My inclination is to trust NeXT's implementation, and blame the Mac file (I've had these problems before, Mathematica also chokes on NPCs). While there are no *visible* problems in the file, this doesn't rule out non-printing characters. Is there a simple editor that will let me see all the ASCII codes in my file, or at least show the NPCs with some kind of glyph? Please don't say emacs! (but if you do, provide detailed instructions ;-) Dave
From: edx@cc.usu.edu Newsgroups: comp.sys.next.programmer Subject: GNUStep install fails Message-ID: <1996Mar4.101421.75753@cc.usu.edu> Date: 4 Mar 96 10:14:21 MDT Organization: Utah State University I'm trying to install GNUStep on an SGI Indigo. When I attempt to perform a "configure" in the libobjects directory, it's giving an error "configure: Couldn't find GCC Objective C runtime library" Now, I've installed the runtime library libobjc.a in the usual place. Configure doesn't say where it's looking for it, or where I should place it, or what I should do to tell it where this library is. Any help appreciated. - HRC - hcole@spanky.idec.sdl.usu.edu
From: johnz@PROBLEM_WITH_INEWS_GATEWAY_FILE (John Zollinger) Newsgroups: comp.sys.next.programmer Subject: Re: NSScanner problems? Date: 4 Mar 1996 19:55:22 GMT Organization: Best Internet Communications Message-ID: <4hfhra$nuo@nntp1.best.com> References: <4hf9tf$rpc@sun0.urz.uni-heidelberg.de> David A. Coyle (dcoyle@goanna.mpi-hd.mpg.de) wrote: : [...] Is there a simple editor that will let me see : all the ASCII codes in my file, or at least show the NPCs with some kind : of glyph? Well... you could try the Unix command 'od'. This command dumps out the passed file in various methods showing the characters used in a bunch of useful ways. Do a man on od for the syntax. Good luck, John Zollinger ProActive Consulting, Inc. johnz@best.com
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Bounds checking for NS apps Date: 4 Mar 1996 20:28:04 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4hfjok$o3l@solaris.cc.vt.edu> Hi, I'm looking for a bounds checking product for NEXTSTEP applications on Intel platforms. The product would be similar to the Purify product, which according to the Purify web page, is not available for NEXTSTEP. The most important feature required is the checking of reading and writing past the bounds of arrays. If you could point me to any products, URLs etc I would appreciate it. Thanks! -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer & Analyst pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
Newsgroups: comp.sys.sun.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.next.misc From: David Fulton <dfulton> Subject: System Development Survey Message-ID: <1996Mar4.204245.54412@ucl.ac.uk> Date: Mon, 4 Mar 1996 20:42:45 GMT Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Mime-Version: 1.0 Organization: University College London System Development Survey Dear Reader, This survey is part of a PhD research study looking at the links between the organisational context in which a system is developed, the problem being tackled and the system development strategy utilised. Currently, knowledge of the type of pressures and problems faced by developers is fragmented across case studies particular to individual disciplines (HCI, Safety-related etc). I am looking for input from system development professionals from a range of disciplines in order that parallels (and contrasting interpretations) across different areas of system development can be drawn. Those of you with form-compatible WWW browsers are cordially invited to contribute to this research by accessing a survey web-page at: http://boom.cs.ucl.ac.uk/staff/dfulton/survey.html For those who are interested in the research but do not have form-compatible web browsers, a text version follows which can be returned to: dfulton@cs.ucl.ac.uk Many thanks! David Fulton -------------------------------------------------- Please fill in all sections in as much detail as possible. The form should take around 10-15 minutes to complete. Personal details are only recorded so that we can get back to you - if required. Otherwise, individuals remain anonymous and your input is kept confidential. Your name: Organisation: Email Addr: 1: The Project You are asked to complete this questionnaire in relation to a recent (or current) project with which you are familiar. 1.1) What was the title of the project? 1.2) Please give a brief description of the project? (if possible): 1.3) What system characteristics did the product incorporate? (more than one of the options can be ticked by marking them with an 'X'): ___ Interactive ___ Safety-Related * ___ Real-Time ___ Embedded * ___ Defence * Safety-related projects being developments where safety (in whatever form) is an important non-functional requirement. Embedded projects being software systems that are used to run hardware 2: Focus of Analysis/Initial Statement of Objectives 2.1) Please mark the development paradigm used by the project: ___ In-house development- for another part of the company ___ Contract-based development- for an outside client ___ Generic/Shrink-Wrap development- for general use 2.2) Which of the following most closely describe the starting point for the project? ___ The project started out without a clear view of the system to be developed, or of the underlying problem to be tackled ___ The project started out with an explicit problem to tackle but without a clear definition of the expected application ___ The project started with an explicit problem to ameliorate and a clear definition of the expected application ___ The project started with an explicit problem to solve and a clear definition of the expected application. In addition, the client has set the acceptance criteria for the project, and penalty provisions if criteria were not met 3: Development Pressures 3.1) Was the system to be developed designed to replace an existing manual or automated organisational system? ___ Yes? ___ No? 3.2) Which of the following describes the development pressures (in generating and using requirements) faced by the project? NB: Only make a choice here if the system was to be used by a user/operator and had an identifiable user population. ___ Users/Representatives had little or no experience with a system of this kind. As a result, they were unable to offer tangible requirements or ideas until a concrete representation (prototype) was available ___ Users/Representatives had some experience with similar systems in the past and they were able to put forward ideas and requirements for future systems The next section looks at the skills of developers and the degree to which their familiarity with the application being developed contributed to their knowledge of a suitable definition. 3.3) Which of the following, in your opinion, describes the degree to which developers (in general) were experienced in this type of development? ___ Developers had limited experience in developing a system of this kind and some experimentation was required to evaluate technical options ___ Developers had some experience of developing similar systems in the past and had a limited ability to predict changes or other types of uncertainty ___ Developers were experts in a particular problem domain and had a clearer understanding of technical options and possible solutions 4: The Organisational Context In considering the organisational context, we are really looking at the context within which the project is based, and are therefore looking at the structures and processes (specific to the project) within the developer organisation, the client organisation (if the two differ), and developer/client communication. 4.1: Structure of the project 4.1.1) Which of the following most closely describe the job roles within the project? ___ Responsibilities and job roles were explicitly defined. No-one on the project could carry out tasks for which they did not have authorisation ___ While there were notional job roles and responsibilities, project personnel tended to carry out a range of tasks when they were required to do so 4.1.2) Which of the following most closely describes the manner in which changes or requests for verification were addressed? ___ All requests for changes had to be routed via the project manager or to personnel at a higher level in the project structure. ___ Possible changes would generally be discussed by the appropriate parties before action was decided (on an informal basis). 4.2: Processes within the project 4.2.1) Which of the following describe the degree of stability in project-related processes? ___ Processes were largely stable. Interaction between stakeholders in the project occured at pre-determined times. ___ Processes were subject to change. Interaction between stakeholders in the project was random - occured as and when needed. 5: System Development Strategy 5.1) In your opinion, which of the following points most closely summarise the system development strategy that was taken? ___ The emphasis was on a constantly changing specification, adaptive code and minimal use of design documentation. ___ The emphasis was on the use of experimental methods, including evolutionary or iterative prototyping. ___ The emphasis was on producing a robust design and implementation, but allowing limited flexibility and ability to react to change. ___ The emphasis was on the production of error-free code and an accurate transformation of the specification into design material. 5.2) Which of the following most closely match the model of development used? ___ Stage-based models of development: A variation on the waterfall software life-cycle model is used ___ Iterative models of development: Spiral or iterative build models of development are utilised 5.3) Which of the following most closely match the emphasis of the development strategy? ___ Emphasis on high-level design: Design activities are closely regulated and monitored ___ Emphasis on a mixture of high and low-level design: Although some procedural restrictions may be in place, low-level design tasks are largely unregulated ___ Emphasis on low-level design: There was little regulation of day to day practice 6: Outstanding issues 6.1) As far as you can recollect, what forms of changing requirements had a particular impact on the project and what was the extent of that impact? Answers on the bi-polar scale refer to how confident you are with options at each end of the scale. If you feel that there is a 'major impact' choose the option on the far right. If you felt there was some impact, but you don't feel you could define it using either of the opposing labels, then choose one of the options in-between 6.1.1) Mutable Requirements: Changes brought about by changing organisational goals and environmental turbulence eg. change because of organisational restructuring mid-way through the project Negligable impact 1 2 3 4 5 Major impact 6.1.2) Consequential Requirements: Changes brought about as a consequence of particular design decisions, or through the testing of prototypes eg. change when users discover new ways of working etc, change when technical staff discover a new way of solving a technical problem Negligable impact 1 2 3 4 5 Major impact 6.1.3) Migration Requirements: Changes when there are difficulties in moving from the current state to the desired state eg, change when considerations have to be made for the technical platforms the working version will have to work on, data management etc Negligable impact 1 2 3 4 5 Major impact 6.1.4) Emergent Requirements: Changes when participants slowly develop a better understanding of what they really want eg, Users clarifying what they really want and negating previous requirements Negligable impact 1 2 3 4 5 Major impact 6.2) Was a method or methodology used in order to aid the analysis and design of the system? ___ Yes? ___ No? 6.3) If the answer to the last question was 'Yes', which method(ology) was used? ___ Information Engineering ___ SSADM ___ Yourdon (or SA/SD) ___ A Systems approach (SSM, ETHICS etc) ___ An Object-Oriented Method ___ Jackson Structured Design or other (please specify) 6.4) if the answer to question 6.2 was 'yes'- Was the method used in its entirety? ___ Yes? ___ No? If not, what aspects of the method were not used, and why? 6.5) Which of the following tools or techniques were used on the project? (A number of selections can be made) ___ Structure Charts ___ Data Flow Diagrams ___ Entity-Life Histories ___ Brainstorming ___ JAD Workshops ___ Prototyping Reviews ___ Entity-Relationship Models ___ Rich Pictures ___ Flow-Charts ___ Class diagrams ___ State Transition Diagrams ___ Storyboarding or other (please specify) 7.1.1) New requirements or changing requirements were coming in at too fast a rate for us to cope effectively Not Applicable 1 2 3 4 5 Very Applicable 7.1.2) It would have been nice if we could get together with the user representatives on a more regular basis, as it was, we were just getting too little feedback Not Applicable 1 2 3 4 5 Very Applicable 7.1.3) We had a number of difficulties in fitting the approach we took with the quality assurance procedures that we had to follow Not Applicable 1 2 3 4 5 Very Applicable 7.1.4) Turnover of staff (with connections to the project) was a major problem and distrupted the project. Not Applicable 1 2 3 4 5 Very Applicable 7.1.5) Too much attention was paid to user interface issues and not enough to the underlying functionality of the system Not Applicable 1 2 3 4 5 Very Applicable 7.1.6) Getting agreement on changes was a time-consuming process, even when the change seemed to be very minor. Not Applicable 1 2 3 4 5 Very Applicable 8: General Comments Are there any other aspects of development practice that you feel are particularly important and haven't been covered within this survey? Are there any other points that you would like to make? 9: On completion of this survey would you like a summary of results? ___ Yes? ___ No? Many thanks for taking the time to fill in this form! David Fulton Department of Computer Science University College London
From: ivo.welch@agsm.ucla.edu (Ivo Welch) Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Re: Bounds checking for NS apps Date: 5 Mar 1996 00:48:19 GMT Organization: The Anderson School at UCLA Message-ID: <4hg30j$jh4@risc.agsm.ucla.edu> References: <4hfjok$o3l@solaris.cc.vt.edu> In-Reply-To: <4hfjok$o3l@solaris.cc.vt.edu> There are some *free* products under LINUX that do this sort of thing, but none under NeXTStep to the best of my knowledge. /ivo welch On 03/04/96, Chuck_Esterbrook@orcacomputer.com wrote: >Hi, > >I'm looking for a bounds checking product for NEXTSTEP applications on Intel >platforms. The product would be similar to the Purify product, which >according to the Purify web page, is not available for NEXTSTEP. The most >important feature required is the checking of reading and writing past the >bounds of arrays. > >If you could point me to any products, URLs etc I would appreciate it. > >Thanks! > >-------------------------------------------------------- >Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome >Software Engineer & Analyst pho. (540) 231-3475 >Orca Computer, Inc. fax. (540) 231-3480 >Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck > -- Ivo Welch ivo.welch@anderson.ucla.edu Assoc Prof of Finance Anderson GSM at UCLA 110 Westwood Plaza, Box 951481, LA CA 90095-1481 UCLA AGSM Finance Faculty Homepage: HTTP://next.agsm.ucla.edu/
From: Joe Freeman <Joe@FreemanSoft.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Tue, 27 Feb 1996 22:31:39 -0500 Organization: FreemanSoft Inc. Message-ID: <3133CC9B.59BF@FreemanSoft.com> References: <4gdbv3$s5c@shelob.afs.com> <312BE9EC.4B0A@gramercy.ios.com> <4gi8ei$ekv@cheyenne.iac.net> <31324223.12747679@news.bogo.co.uk> <313259A3.5433@ix.netcom.com> <31338a40.12358521@news.plsys.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Karl Harbour wrote: > And why did Garfinkel & Mahoney and the people who wrote the > NeXTDeveloper/Examples do it that way? Because of target/action and > the design of the AppKit generally separating UI from application > logic. This is definately not true. The target/action paradigm has done nothing to aid in separating out the application logic. Applications have several different types of process, presentation, application, and model to name the most common. The insertion of controllers into nib files where those controllers implement UI control logic and application logic is a "bad thing" IMHO. It makes it very difficult to deploy an application across presentation layer architectures.. > I'm sorry, but your argument is completely bogus. I think that completely should be used about as often as "never" and "always". > I suppose you don't think the AppKit provides any help to the > developer for writing consistent apps in terms of the look and feel of > cut & paste, spell checking, fonts, colours, printPSCode, faxPSCode > and so forth? If two apps handle these things in the same way, is it > down to the application developer or what's provided by Next's AppKit? These are nice features but I'm not as convinced any more that they are as important as an application frame work, a standard doc architecture, application and document controllers and some policy aides. Almost all of the subsystems you mention are really abstractions for system resources and are basicly consumable objects. (printPSCode and faxPSCode are actually interface protocols) So, NeXT did a nice job wrapping their internally developed resources which is good. It is not clear what portion of the development time that these resources represent -- FreemanSoft Inc. NEXTSTEP and WEBOBJECTS software in the DC area. Governer Glendenning of MD wants to spend $300M of the people's money on a rent free pro football stadium which is only used 10 times a year. For whom is he really doing this deal?
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9603050213.AA07982@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Mon, 4 Mar 96 20:15:16 -0600 Subject: Releasing elements within NSMutableArray Hello, all: I'm sending the [removeAllObjects] method to an instance of class NSMutableArray, and then following it up with a [release] method to the array object itself. I had subclassed each element in the array to release some local variables, and then to send an [autorelease] command to its superclass (Hmm...wonder if it should've sent [autorelease] to itself, since I hadn't overridden the call...) Anyway, the above sequence of calls resulted in each object in the array also getting passed the release command from the array which had held it, even though the array had already been requested to remove the object. (In other words, each element was being over-released.) I got rid of the problem by replacing the [super autorelease] call in each element with a [super release], which of course released each element immediately and prevented it from receiving the extra [release] call from the array. My concern is why the array was able to pass the [release] call along in the first place, to elements which had already been "removed". Apparently the slots in the array aren't nilled out when all the elements are removed??? Instead, I guess some type of nextAvailableSlot counter is getting reset...I wonder what happens if I release an object at a specific index... Regards, - Dan
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Ivar link to view in ScrollView Date: 5 Mar 1996 02:32:57 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4hg94p$fac@news.onramp.net> References: <RDL.96Feb18225038@world.std.com> In article <RDL.96Feb18225038@world.std.com> rdl@world.std.com (Robert La Ferla) writes: > You should be able to double-click on the View inside the ScrollView and > Ctrl-drag out a connection. And you can just Ctrl-drag a connection to the > View inside the ScrollView as well. This at least works in IB (v460) which is > the version that I believe came with EOF 1.1. I have that version of IB. But I made a mistake in my post - it's a Browser that I'm trying to do this with - and with a Browser, it doesn't work for me. Anyone know how to get it to work with a Browser?? Steve
From: wgaboria@iut-lr.univ-lr.fr Newsgroups: comp.sys.next.programmer Subject: only upper case char in a textField object Date: 28 Feb 1996 10:01:45 GMT Organization: Universite de La Rochelle Message-ID: <4h1969$d32@hpuniv.univ-lr.fr> Hello, Here is a problemwich seems simple : I'am trying to have only upper case char in a textfield or text object. How to force the user to enter upper shift char ? Thanks to reply to : /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/ NeXT Mail are Welcome !
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: WebScript spinoff...? Date: 28 Feb 1996 11:59:43 GMT Organization: Academic Computing Services, Sheffield University Message-ID: <4h1g3f$9nr@bignews.shef.ac.uk> I haven't noticed any comment here on what seems to me as if it might be an important spin-off from WebScript (or maybe it was this that made WebScript possible?) -- namely an Objective-C interpreter... I might be misunderstanding things, but WebScript to me looks like it's pretty much Obj-C, it's not compiled, and it has (or should have -- I haven't been able to get it to work, anyone else?) a couple of neat features for tracing, printing out logs etc. I wouldn't have thought it would take too much effort to build this into the "main" (IB/PB) development environment, which would seem to be what a number of people have asked for over the years (and an improvement on old Eval.app). Any other takes on this, or have I just not been sleeping enough?! Best wishes, mmalc. --
From: Carl Edman <cedman@princeton.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Re: Bounds checking for NS apps Date: 5 Mar 1996 02:16:18 GMT Organization: Princeton University Message-ID: <4hg85i$pft@cnn.Princeton.EDU> References: <4hfjok$o3l@solaris.cc.vt.edu> <4hg30j$jh4@risc.agsm.ucla.edu> On 03/04/96, Chuck_Esterbrook@orcacomputer.com wrote: > >I'm looking for a bounds checking product for NEXTSTEP applications on > >Intel platforms. The product would be similar to the Purify product, which > >according to the Purify web page, is not available for NEXTSTEP. The most > >important feature required is the checking of reading and writing past the > >bounds of arrays. In <4hg30j$jh4@risc.agsm.ucla.edu> Ivo Welch wrote: > There are some *free* products under LINUX that do this sort of thing, but > none under NeXTStep to the best of my knowledge. There are also a free set of patches against recent versions of gcc available which add bounds checking on the compiler level. The only question is whether the current version of gcc compiles under NS. Anybody tried ? Carl Edman
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Re: NSScanner problems? Date: 5 Mar 1996 10:09:59 GMT Organization: University of Heidelberg, Germany Message-ID: <4hh3tn$kqt@sun0.urz.uni-heidelberg.de> References: <4hfhra$nuo@nntp1.best.com> In article <4hfhra$nuo@nntp1.best.com> johnz@PROBLEM_WITH_INEWS_GATEWAY_FILE (John Zollinger) writes: > David A. Coyle (dcoyle@goanna.mpi-hd.mpg.de) wrote: > : [...] Is there a simple editor that will let me see > : all the ASCII codes in my file, or at least show the NPCs with some > > Well... you could try the Unix command 'od'. This command dumps out the > passed file in various methods showing the characters used in a bunch of > useful ways. Do a man on od for the syntax. Thanks John, I tried this. Now it appears that I'll have to lay the blame on NSScanner. (I've appended the offending portion of the file at the end of this posting). I estimate that the Scanner has scanned just over 200 elements by the time it loses track of whatever it's doing. There are no specified limits to the size of the string you can process, and I can't imagine that this file is too much for it to chew on, but... Has anyone else much experience with using this object? Is my file just too big for it? Dave ------------------------------------ verbatim portion of the text being scanned (you can see that it's pretty monotonous): (...) 27 14305 20930 C 12.7570 67.0927 28 14305 20930 C 15.6601 77.0180 29 24226 -25515 C 15.4542 78.0100 30 23901 -25387 C 16.0759 30.1988 31 24308 -24972 C 12.2331 80.3413 32 24366 -24773 C 11.8189 62.0549 33 24503 -25050 C 14.4868 37.1141 34 24568 -24792 C 11.0830 40.1090 35 24264 -24447 C 11.1345 58.9242 (...) It dies after processing item 33 (Mac file) or after item 32 (Unix). Always at the same place (these aren't the actual line numbers, there's header stuff that the NSScanner scans past, correctly...). It's reasonable to expect that if it gets 32 lines like this correct, it can manage the next 70...
Newsgroups: comp.sys.next.programmer From: ian.stephenson@insignia.co.uk Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Message-ID: <DnsM47.AEJ@isltd.insignia.com> Sender: news@isltd.insignia.com Organization: Insignia Solutions plc References: <3137D0A5.5962@ix.netcom.com> Date: Tue, 5 Mar 1996 11:48:06 GMT In article <3137D0A5.5962@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> writes: > Which is essentially my point: The tool itself won't magically create good > design. the REAL programmer can write FORTRAN in any language. >Face it, the tool does not enable good design. You can > do good design with gcc on a linux box, or with java, >or even with Basic. I tried this a few years ago! I took everythin gI'd learnt about OO, and NeXTStep and design etc and tried to build a really cool app in BASIC on an 8bit micro. It was actually pretty sucsessful! What I produced was of a VERY high standard compared to what was being shipped commercially a few years previously on that same platform (which was at its peak in the mid 80's). However I ran into problems trying to reuse that code, and it didn't scale well. You can only go so far with poor tools. > NeXTSTEP may allow you > to produce crap '10 times faster'. But it's still crap. Now this is the real trick! Under NeXTStep it ISN'T 10 times faster to produce crap. To produce crap takes just as long as on any other platform (say X). It's only faster if you're doing the right things. If you want menu's along the top of your window like MSWindows then thats going to be a bitch to do (I've seen apps try it!), and it'll take you forever. However if you want a NeXT menu with the standard look and feel it's very fast. Programming X, it takes hugh amounts of code to build an App which looks and feels right (check out the weight of the Motif style guides). Under NeXT its hard not to look and feel right (ever bothered to read the NeXT style guide - Yes there is one, you just didn't need it). Hence most NeXT apps feel right, and most X apps feel wrong (even to X people!). Sure NeXTStep isn't perfect, and everyone has made mistakes building OO apps, but the tools DO encourage you to do the right thing. I've worked on bad NeXTStep code that was just generic code with a NeXT front end tagged on (sometimes this is the right solution), and I've worked on real NeXTStep code. A well built system of objects models the problem, and flows into a solution in a way that traditional code never could. "It is truly alive withing the tao" Ian To teach is an immposible task you can only hope that you assist others to learn.
Newsgroups: comp.sys.next.programmer From: jpanico@netcom.com (Joe Panico) Subject: Re: Format and Documentation of eomodel files? Message-ID: <jpanicoDnsptw.AxM@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4hfb9r$5av@stc06.ctd.ornl.gov> Date: Tue, 5 Mar 1996 13:08:19 GMT Sender: jpanico@netcom19.netcom.com John W. Wooten (woo@woonext.cmo.ornl.gov) wrote: : the eomodel file so that I can use it in an NextStep interface (i.e. with : EOBrowser, and table view, etc.). The eomodel file is simply an ASCII representation of a "property list". Property lists are nowhere defined in the documentatioin (as far as I can tell), but basically a property list is a way to represent certain information (in the case of EOModel, the state of the EOModel, EOEntity and EOAttribute instances) using a collection of heirarchically arranged NSDictioanrys, NSArray and NSStrings (and NSData, but these are not used for eomodel). On the concretet level: pList = [anEOModel modelAsPropertyList] return an NSDictionary. The dictioanry contains an object/value pair for many of its important ivars, and an array of NSDictionarys, each of which represents the state of one of the models EOEntitys. The same holds true for EOEntitys and their EOAttributes. asciiList = [[pList description] cString] This is an ascii representation of the property list, where {...} denotes an NSDictionary, (...) denotes an NSArray and key = value denotes an NSDictioanry entry where the value is an NSString. Now write asciiList to a file and you have an .eommodel file! : -- : - - - - - - - - - : J. W. Wooten -- Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc. jpanico@netcom.com /* Please no NeXTMail, I can't read it at this address */
From: stes@maze.ruca.ua.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 5 Mar 1996 13:58:10 GMT Organization: RUCA , Universiteit Antwerpen Message-ID: <4hhh9i$mmv@maze.ruca.ua.ac.be> References: <4guqt5$5nd@ixnews3.ix.netcom.com> <4gviqf$dh@news.next.com> <4h9hd4$qg4@infoserv.rug.ac.be> <4hf3tj$2m1@news.its.com> Chuck Swiger (chuck@its.com) wrote: : You're looking for books on machine languages, finite automata, and : searching/matching algorithms. I was asking because of IXBMTable and IXBMCompile() in the "ToDoList" mini-example. Seems to me that these things would better have been kept internal, so that you don't have to know about all that to _use_ the kit. If the Indexing Kit were to be reworked some day, I hope that efforts will be focused on the interface instead of the implementation. What does it help that the objects are zonified, thread safe, use advanced Boyer Moore and finite automata theory, if you have to learn a total of 17 classes and 18 protocols !? That's really what I liked about the (original) Application Kit : just a few strong concepts - Application, Window, View. And it's exactly the interface of _these_ objects that is being reworked for OpenStep, I don't get it. In Indexing Kit, yeah , _there_ you have deficiencies. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re: Format and Documentation of eomodel files? Selector Ref solved Date: 5 Mar 1996 14:19:13 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4hhih1$c2v@stc06.ctd.ornl.gov> References: <4hfb9r$5av@stc06.ctd.ornl.gov> In article <4hfb9r$5av@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > I'm creating a WebObject example that uses an NSDictionary for storage of the .. > > Also when I try to add a method to my Application.wos file, I get the message > that the WO'Application.wos does not have a selector for my method. I thought > I added a selector by simply putting the definition of that "method" into the > Application.wos script. > Found out something here of possible interest to others! If you've ever referenced a WebObject that has an Application.wos, then you must kill the process running that WebObject Application before it will allow you to add or change "selectors" in the Application.wos. Once I did that, then my selector was recogized and worked fine. - - - - - - - - - J. W. Wooten
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP Developers/Experienced/$50hr+ Date: 5 Mar 1996 14:43:20 GMT Organization: Norden 1 Communications Message-ID: <4hhju8$ne2@tofu.alt.net> OMNI SEARCH currently has several well experienced NEXTSTEP developers available for long term assignments. These developers have years of experience that includes NEXTSTEP, Objective C, EOF, and Sybase or Oracle. Rates can be arranged for as low as $50hr for long term assignments. If interested, please contact me as soon as possible. Thank You Tom Gugger -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: ObjectError Date: 5 Mar 1996 17:09:12 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4hhsfo$1n2@news.its.com> References: <4h1vnv$85c@transfer.stratus.com> <4h9j01$qg4@infoserv.rug.ac.be> stes@cage.rug.ac.be (David Stes) wrote: > Charles_Spitzer (charlie@why.az.stratus.com) wrote: >: Anyone have this rewritten as a poseAs: for NSObject? I tried doing so, >: but found that NSObject doesn't have Object's +error: or -error methods. >: Can one just eliminate these methods? > > Is this true ? I wouldn't be surprised. > > Wouldn't it be simpler to just make NSObject a subclass of Object, and > inherit these methods ? I really don't understand NeXT on the issue. No can do. Object and NSObject both need to be "root classes" (ie, one can't be a subclass of the other), since they need to do some common fundamental tasks in an incompatible way from each other. There is a category on Object to make it more compatible with the FoundationKit world, however.... -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: frederic@pegase (Frederic GALOT) Newsgroups: comp.sys.next.programmer Subject: EOApplication? Date: 6 Mar 1996 01:52:43 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4hir5b$g2v@ai.alienor.fr> Project Builder does'nt keep the change in appName_main.m. The fact is that when I save my project I've got a message : "header file EOApplication.h not found". So, I modify my appName_main.n : #import "EOApplication.h" becomes #import "eointerface/EOApplication.h". But I'll have to do it again next save. What to do? -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: frederic@pegase (Frederic GALOT) Newsgroups: comp.sys.next.programmer Subject: Model Browser does not browse! Date: 6 Mar 1996 01:55:22 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4hiraa$g2v@ai.alienor.fr> My problem : I uses a Custom class for some EOModeler attributes, subclass of NSString, but while doing that I can't load the concerned attributes in my Model Browser. Description: * to perform character set conversion between my database and NEXTSTEP I've made a NSString subclass. * the character conversion is OK in my application. * but if I'm using EOModeler.app's Model Browser on that attribute the message is : "Unable to find value class ConvertToNextStepNSString for attribute myAttribute". I've tried to modify the .model file and many other things but no way! -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: frederic@pegase (Frederic GALOT) Newsgroups: comp.sys.next.programmer Subject: Multiple declaration of method... Date: 6 Mar 1996 01:56:18 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4hirc2$g2v@ai.alienor.fr> I'm using EOF and I've got a compilation problem : The messages are : * "multiple declarations for method -foo" * "appkit/toto.h using -foo" * "eointerface/toto2.h also found -foo" What can I do to avoid this? -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: frederic@pegase (Frederic GALOT) Newsgroups: comp.sys.next.programmer Subject: NXLineDesc? Date: 6 Mar 1996 01:57:17 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4hirdt$g2v@ai.alienor.fr> Has anybody ever use NXLineDesc? I've managed doing it but not in a standard way. My goal is to count the number of chars in each line of a Text object. Perhaps it exists another way to do that? -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: eugene@raddi.uah.ualberta.ca (Eugene Mah) Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Re: Bounds checking for NS apps Date: 5 Mar 1996 17:01:43 GMT Organization: University of Alberta, Edmonton, Canada Message-ID: <4hhs1n$s04@pulp.ucs.ualberta.ca> References: <4hfjok$o3l@solaris.cc.vt.edu> <4hg30j$jh4@risc.agsm.ucla.edu> <4hg85i$pft@cnn.Princeton.EDU> In <4hg85i$pft@cnn.Princeton.EDU> Carl Edman wrote: > On 03/04/96, Chuck_Esterbrook@orcacomputer.com wrote: > > >I'm looking for a bounds checking product for NEXTSTEP applications on > > >Intel platforms. The product would be similar to the Purify product, which > > >according to the Purify web page, is not available for NEXTSTEP. The most > > >important feature required is the checking of reading and writing past the > > >bounds of arrays. > > In <4hg30j$jh4@risc.agsm.ucla.edu> Ivo Welch wrote: > > There are some *free* products under LINUX that do this sort of thing, but > > none under NeXTStep to the best of my knowledge. > > There are also a free set of patches against recent versions of gcc available > which add bounds checking on the compiler level. The only question is > whether the current version of gcc compiles under NS. Anybody tried ? > I haven't tried the bounds checking patches, but both gcc 2.7.2 and g77 (gnu fortran) compile and work pretty nicely on my slab (NS 3.2). I use gcc all the time now instead of the NeXT cc. Don't know about Intel yet. Eugene -- ---------------------------------------------------------------------- Eugene Mah eugene@raddi.uah.ualberta.ca (NeXTmail) Systems Administrator "For I am a Bear of Very Little Department of Radiology Brain, and long words Bother University of Alberta Hospitals me." Winnie the Pooh Edmonton, Alberta, Canada http://raddi.uah.ualberta.ca/~eugene/ ----------------------------------------------------------------------
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <doug@thoughtful.com> Date: Tue, 5 Mar 96 09:50:10 -0700 From: doug@thoughtful.com (Douglas Simons) Message-ID: <9603051650.AA04717@thoughtful.com> Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > In practice, the three of us who switched to Delphi were > programming fluently within a month, using good OOP > techniques. (And I had never used Pascal before, not even > in school.) The user manuals are not great, but the > context-sensitive online help is decent. We had figured > that relearning OPENSTEP would take approximately the > same amount of time. So that aspect of the transition > was a wash. That sounds all very well and good, for undertaking a new project. But what about converting existing code? Surely the job of taking thousands of lines of Objective C code (based on AppKit etc. classes) and rewriting them in Delphi's version of Pascal (using their class library) is far from being a trivial task. I'm curious to know -- how much code did you actually port, and what was your experience with that? Doug Simons Thoughtful Software
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Indexing Kit Query Language Date: 6 Mar 1996 00:26:07 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4him2v$e7@news.xmission.com> References: <4guqt5$5nd@ixnews3.ix.netcom.com> <4gviqf$dh@news.next.com> <4h9hd4$qg4@infoserv.rug.ac.be> <4hf3tj$2m1@news.its.com> <4hhh9i$mmv@maze.ruca.ua.ac.be> stes@maze.ruca.ua.ac.be (David Stes) wrote: > If the Indexing Kit were to be reworked some day, I hope that efforts will > be focused on the interface instead of the implementation. What does it > help that the objects are zonified, thread safe, use advanced Boyer > Moore and finite automata theory, if you have to learn a total of 17 > classes and 18 protocols !? If anyone has some ideas on ways to improve the IXKit design, I'd love to hear about them. Ie., I'd like to hear ideas about how to fix it, not just that it is "broken". > In Indexing Kit, yeah , _there_ you have deficiencies. I agree, but I don't yet have enough of a handle on all of the design decisions behind the IXKit to be able to propose anything that would be simpler and still remain at least as powerful as the current design. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.lang.pascal.delphi.misc Subject: Re: NEXTSTEP vs. Delphi: An Informed comparison Date: Wed, 06 Mar 1996 00:26:25 GMT Organization: Orbital Computer Consultancy Message-ID: <313cda97.3272539@news.plsys.co.uk> References: <31378720.13672300@news.plsys.co.uk> <4hf5kf$ti@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >Karl Harbour writes >> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: >> > In NEXTSTEP, you can attach a Button to any "sender" method of any >> > other object. But the only Button message you can attach is its >> > 'click', the primary action of the Button. That's because the >> > target->action paradigm is hard-wired, non-extensibly, into >> > InterfaceBuilder. This assumes, of course, that you are using the >> > standard classes. It is possible, with some effort, to build custom. >> > IBConnectors for custom classes >> >> Surely an important point is that Delphi event handlers correspond to >> two different paradigms in Nextstep, namely target/action and >> delegation. Anything the Delphi event handlers can do can be handled >> by one of these two paradigms as appropriate. > >Buttons have delegates? Please explain how, WITHOUT SUBCLASSING, I can >get notified and then do something special about mouse in/out and drag >enter/exit events in NEXTSTEP buttons. I can in Delphi. > I agree that it's easier in Delphi, especially with buttons. I guess I was trying to say the difference is less in the case of TForm vs Window, since Window has a good set of delegate methods. <-- snip ---> >> Plus of course delegation gives you the unique "Can I do this?" / >> denial functionality, which is absent from Delphi. > >This is a lie. OnCloseQuery, OnChange, and all BeforeXxxx methods of >Tables -- just to name three that spring to mind -- provide exellent >structured exception handling (or Boolean function returns) to deny >continuation of a proposed operation. > Well, a mistake actually, rather than a lie. I didn't look closely enough at TForm's events. Sorry if I misled anyone. <-- snip -->
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP Developers/Experienced/$50hr+ Date: Tue, 05 Mar 1996 20:23:40 -0500 Organization: Steel Driving Software, Inc. Message-ID: <313CE91C.4CB3@ix.netcom.com> References: <4hhju8$ne2@tofu.alt.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit E. Hutchinson wrote: > > OMNI SEARCH currently has several well experienced > NEXTSTEP developers available for long term assignments. > These developers have years of experience that includes > NEXTSTEP, Objective C, EOF, and Sybase or Oracle. > Rates can be arranged for as low as $50hr for long > term assignments. If interested, please contact me > as soon as possible. Wow! As low as $50/hr! I don't think I've ever seen a phrase like that in a recruiting ad before.
Newsgroups: comp.sys.next.programmer Subject: Intimidation of shareware author Message-ID: <19960305.204723.622475.NETNEWS@WVNVM.WVNET.EDU> From: "J. Ross" <oma00085@wvnvm.wvnet.edu> Date: Tue, 5 Mar 1996 20:47:23 -0500 Life as a shareware author is hard enough. A majority of users mistake "shareware" for "freeware" and use the products without sending in the registration fees. This is to be expected. People are forgetful and most don't stop to think that someone's livelihood depends on the generally modest registration fee. What is not expected, however, is when a giant corporation with high-priced lawyers threatens you with a lawsuit for a copyright violation on a product you didn't even know existed. This is what happened to me, and I'm sure it's a fairly common practice in the industry, however, it's certainly disconcerting. I spent months developing what I thought was a unique product-- a Spy application for debugging user applications utilizing the standard Winsock API. After putting the finishing touches on the application and posting it to my Web Page for distribution & evaluation, I receive a threatening e-mail letter from a corporation clear across the country claiming a copyright violation on a competing product. WHAT GIVES HERE????? Do they think they have exclusive rights to a product simply because they thought of it first? After all, the Winsock API is public information, and I've been writing Winsock applications for the past three years. Or are they simply trying to force me out of the market because I'm asking a shareware registration fee which is less than 1/3 of their product price? Do these intimidation tactics work? While respecting any company's right, (including my own), to defend its copyright, there should be room in the marketplace for competing products. In this case, there is absolutely no legal basis for their claim of copyright infringement, however, they have more money to spend on lawyers than I do. If forced to withdraw my SocktSpy applications from the market, (http://www.win-tech.com), I lose the time spent in honest development of the product, but the real loser is the programming community which no longer has the option of choosing between competing products. Have others experienced this kind of strong-armed tactics & if so, what was the outcome? J.Ross
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: 2 bit tiff to 1 bit tiff Date: 6 Mar 1996 04:42:03 GMT Organization: Nanyang Technological University Message-ID: <4hj52r$8kc@ntuix.ntu.ac.sg> References: <4hf2ga$anr@stern.fokus.gmd.de> Oliver Fox (fox@jeans.fokus.gmd.de) wrote: : Does anyone know how to convert a 2 bit tiff to a 1 bit tiff : a silly archiving system from a client only accepts 1 bit tiffs : I would try the following: tifftopnm your-file.tiff | pgmtopbm | pnmtotiff > your-new-file or similarly tifftopnm your-file.tiff | pnmdepth 1 | pnmtotiff > your-new-file PBMplus and/or netpbm is available on many sites. Willem
From: lionel@cyberlab.ch (Lionel Tinguely) Newsgroups: comp.sys.next.programmer Subject: NeXTSTEP and JAVA Date: Tue, 05 Mar 1996 23:46:09 GMT Organization: CYBERLaB NeTWORK Message-ID: <4hinam$sjp@news.worldcom.ch> Hello ! Does any one know if someone is porting JAVA on NeXTSTEP ? Thanks *********************** THE SWISS JAVA PAGES *********************** Lionel Tinguely E-mail: lionel@cyberlab.ch (Mime and NeXTMAIL OK) WWW: http://www.cyberlab.ch/
From: wgaboria@ainfo3.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: Re: Multiple declaration of method... Date: 6 Mar 1996 08:54:09 GMT Organization: Universite de La Rochelle Message-ID: <4hjjrh$h3v@hpuniv.univ-lr.fr> References: <4hirc2$g2v@ai.alienor.fr> frederic@pegase (Frederic GALOT) wrote: > I'm using EOF and I've got a compilation problem : > The messages are : > * "multiple declarations for method -foo" > * "appkit/toto.h using -foo" > * "eointerface/toto2.h also found -foo" > What can I do to avoid this? In this case you must indicate the class you want to use for this method : 1-bad example : ================ In myController.m : myMethod [myTextField stringValue]; at compilation : ---------------- myController.m: In function `-[myController myMethod:]': myController.m:100: warning: multiple declarations for method `stringValue' /NextDeveloper/Headers/appkit/Cell.h: In function `stringValue': /NextDeveloper/Headers/appkit/Cell.h:191: warning: using `-(const char *)stringValue' /NextDeveloper/Headers/foundation/NSValue.h: In function `stringValue': /NextDeveloper/Headers/foundation/NSValue.h:53: warning: also found `-(NSString *)stringValue' 2-good example : ================ [(TextField *)myTextField stringValue]; /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/
From: wgaboria@ainfo3.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: Re: EOApplication? Date: 6 Mar 1996 09:02:05 GMT Organization: Universite de La Rochelle Message-ID: <4hjkad$h3v@hpuniv.univ-lr.fr> References: <4hir5b$g2v@ai.alienor.fr> frederic@pegase (Frederic GALOT) wrote: > Project Builder does'nt keep the change in > appName_main.m. > The fact is that when I save my project I've got a > message : "header file EOApplication.h not found". > So, I modify my appName_main.n : > #import "EOApplication.h" becomes > #import "eointerface/EOApplication.h". > > But I'll have to do it again next save. > What to do? You must set "generate main file on save" at false in ProjectBuilder.app Attributes . /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: NXLineDesc? Date: Wed, 6 Mar 1996 09:31:19 +0000 Organization: Cedar Systems Distribution: world Message-ID: <juJyMCAntVPxEwye@cedar.co.uk> References: <4hirdt$g2v@ai.alienor.fr> MIME-Version: 1.0 In article <4hirdt$g2v@ai.alienor.fr>, Frederic GALOT <frederic@pegase> writes > Has anybody ever use NXLineDesc? I've managed >doing it but not in a standard way. My goal is to count the >number of chars in each line of a Text object. > Perhaps it exists another way to do that? Yes, we use NXLineDesc a lot in the internal workings of our extended Text classes - I don't know of a more efficient way of getting the number of characters on each line. When working with NXBreakArray, be sure to check for negative values that indicate NXHeightChange. If you have any concerns about your code, email me a copy and I can check you are handling NXLineDesc correctly. Regards, Paul. ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <maurice@tlpnl.tlp.nl> From: Maurice Le Rutte <maurice@tlpnl.tlp.nl> Message-ID: <9603060937.AA08033@tlpnl.tlp.nl> Date: Wed, 6 Mar 96 10:37:17 +0100 Subject: Re: Releasing elements within NSMutableArray Dan Tapp <tappd@copernicus.sccsi.com> asked: > I'm sending the [removeAllObjects] method to an instance of class > NSMutableArray, and then following it up with a [release] method to > the array object itself. I had subclassed each element in the array > to release some local variables, and then to send an [autorelease] > command to its superclass <snip> I am not sure if I understand you correctly, but if you are doing something like the code example below I'd advice you to read: * /NextLibrary/Documentation/NextDev/Foundation/IntroFoundation.rtfd ('Object Ownership and Automatic Disposal') * /NextLibrary/Documentation/NextDev/Foundation/Protocols/NSObject.rtf If you did not mean this, please forgive my stupidity. > My concern is why the array was able > to pass the [release] call along in the first place, to elements > which had already been "removed". Apparently the slots in the array > aren't nilled out when all the elements are removed??? You can check this using the GNU debugger (gdb), using the commands p <plain-style-c-address-expression> or po <an-objects-address>. For details on using the gdb please check your local reseller, ehh, I mean your manual. Maurice. <code> - foo { ... [yourArray removeAllObjects]; [yourArray release]; ... } @implementation YourSubClass - (void)release { [myRelation1 release]; ... [myRelationN release]; [super autorelease]; } @end </code> This is wrong, your object registers itself one more time for releasing, but the super class's 'autorelease' does not call it's own release, it merely puts itself into the current autoreleasepool (NSAutoreleasePool). When the pool gets emptied it calls for all the objects 'release'. In the implementation I have illustrated above the retaincount never gets decreased, so never a real deallocation is made and each time the object gets released by some other object it wants to release itself one more time, but that has no effect 'cause... (ad infinitum) Should be: <code> @implementation YourSubClass - (void)dealloc { [myRelation1 release]; ... [myRelationN release]; [super dealloc]; } @end </code>
Date: 6 Mar 1996 09:38:01 EST Newsgroups: comp.sys.next.programmer Message-ID: <cancel.19960305.204723.622475.NETNEWS@WVNVM.WVNET.EDU> Control: cancel <19960305.204723.622475.NETNEWS@WVNVM.WVNET.EDU> From: "J. Ross" <oma00085@wvnvm.wvnet.edu> Subject: cmsg cancel <19960305.204723.622475.NETNEWS@WVNVM.WVNET.EDU> EMP/ECP (aka SPAM) cancelled by clewis@ferret.ocunix.on.ca. See news.admin.net-abuse.announce, report 19960306.02 for further details
From: 93925730@plink.cityu.edu.hk (David P. FOK) Newsgroups: comp.sys.next.programmer Subject: inb(), outb(), How? Date: Wed, 06 Mar 1996 23:18:04 GMT Organization: CityU of HK Message-ID: <4hka40$ndm@hpg30a.csc.cuhk.hk> Dear Programmers, I have a big problem on using these 2 functions. All I need to do is a simple 8-bit I/O but the codes keep on refusing to work. The Builder reports no problem and the app can be built successfully but whenever it runs, it quits immediately. Here below are the codes to test the outb() functions, #import <appkit/appkit.h> #import <driverkit/i386/ioPorts.h> #import <driverkit/generalFuncs.h> #define sourcePort 0x220 #define measurePort 0x221 @interface Calculator:Object { id inputField; id outputField; } - doConversion:sender; @end ================================= #import "Calculator.h" @implementation Calculator - doConversion:sender { int i; float degreeF; void IOInitGeneralFuncs(); [inputField selectText:self]; degreeF=((9.0*[inputField floatValue])/5.0)+32.0; [outputField setFloatValue:degreeF]; //below are the lines for the outb() testing, which is inserted to the //NeXT calculator example. outb((unsigned int) sourcePort,(unsigned char) 0x80); for(i=0;i<1000;i++); return self; } @end I did try to remove the void IOInitGeneralFuncs(); as it's required for user drivers only and casting the "port" to IOEISAAddress(?) also. No magic happens then. Could anyone give me some hints? The hardware I built works okay under dos by using the inportb() and outport() functions and it requires no interrupt or DMA. In addition, I did really confused by the DriverKit. It is documented that the IODevice class is a subclass of Object but, I cannot subclass from the IODevice by any means, why? Sorry for wasting your time if this is an FAQ but I did consult all possible source of answers. The SMC16 driver example seems just so straight forward but... Thanks in advance for any suggestions. I am using NeXTSTEP 3.3 user and developer under Intel. -david
From: jbettis@cse.unl.edu (Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Re: 2 bit tiff to 1 bit tiff Date: 6 Mar 1996 13:58:19 GMT Organization: University of Nebraska--Lincoln Message-ID: <4hk5lr$5uo@crcnis3.unl.edu> References: <4hf2ga$anr@stern.fokus.gmd.de> <4hj52r$8kc@ntuix.ntu.ac.sg> GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) writes: >I would try the following: >tifftopnm your-file.tiff | pgmtopbm | pnmtotiff > your-new-file >or similarly >tifftopnm your-file.tiff | pnmdepth 1 | pnmtotiff > your-new-file >PBMplus and/or netpbm is available on many sites. I can't get tifftopnm to work right. It usually puts a bunch of garbage on the right side of the image. -- Jeremy Bettis -*- PGP Public key available -*- University of Nebraska INET: jbettis@cse.unl.edu "Those who stand in the middle of the UUCP: jeremy@tddi.inetnebr.com road are often hit by passing cars." <a href="http://cse.unl.edu/~jbettis/">Click Here</a>
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: Re: Bounds checking for NS apps Date: 6 Mar 1996 14:32:38 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4hk7m7$1f0@news.its.com> References: <4hfjok$o3l@solaris.cc.vt.edu> Chuck_Esterbrook@orcacomputer.com wrote: > I'm looking for a bounds checking product for NEXTSTEP applications on > Intel platforms. The product would be similar to the Purify product, which > according to the Purify web page, is not available for NEXTSTEP. Yup. I talked with the people from Purify a couple of years back about porting Purify or writing a similar tool from scratch. The project never got out of the planning stage since (a) Purify wasn't interested, and (b) it would have required a great deal of information about things such as how the system loader and shared libraries work that NeXT wasn't willing or able to document adequately. So there aren't any post-compilation tools available. Try including runtime checks into your code (as in Carl's suggestion to patch the compiler). > The most important feature required is the checking of reading and writing > past the bounds of arrays. As a purely practical experiment, have you tried running your code after declaring automatic local arrays to be static? This is helpful for preventing array overruns from blowing away the stack and rendering the program undebuggable. Or try using the Foundation kit; it lets you do things at a higher level.... -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: 6 Mar 1996 16:20:57 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4hke19$870@shelob.afs.com> References: <9603051650.AA04717@thoughtful.com> Douglas Simons writes > Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > > In practice, the three of us who switched to Delphi were > > programming fluently within a month, using good OOP techniques. > > That sounds all very well and good, for undertaking a new project. > But what about converting existing code? Surely the job of taking > thousands of lines of Objective C code (based on AppKit etc. classes) > and rewriting them in Delphi's version of Pascal (using their class > library) is far from being a trivial task. We don't have much source code to begin with. This is OOP, right? 8^) Seriously, our libraries are divided into four sections: generic 'C' functions, AppKit extensions, TradeKit core classes, and afs:TRADE customer projects. The calculation library stayed in C and was compiled into a DLL. The remaining C functions without Pascal equivalents were ported very quickly. Most of the AppKit extensions were not needed, because they were only there to address AppKit deficiencies. AfsModel and AfsController did have to be ported, but it wasn't that hard. The differences between C and Pascal are syntactical, not functional. I suppose we could have written a 'sed' script (if we knew how) to change '{' to 'begin' and so on. The TradeKit will of course take additional time, but again, the logic and business rules are invariant. The point is, the basic methodology of building classes is the same in both environment. Since you don't have to learn and apply new paradigms, it's less work than you'd think. In my original article, I was contasting this effort to the time it would take to (1) build and run NS->OS conversion scripts; (2) walk through the resulting code and hand-convert whatever the script didn't; (3) understand the new code, since the name of almost every class and method seems to have changed; (4) completely recertify the new code. If NeXT had not decided to so thoroughly change everything, this factor would definitely come down more on its side. But I have to be honest: every time I read the documentation on converting NS apps to OS, my mind starts spinning and I get a headache. I think the first few souls brave enough to convert a serious-sized app are in for a surprise. -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: jcr@idiom.com (John C. Randolph) Newsgroups: comp.sys.next.programmer Subject: Re: NEXTSTEP Developers/Experienced/$50hr+ Date: 6 Mar 1996 05:09:56 -0800 Organization: A poorly-installed InterNetNews site Message-ID: <jcr.826117686@idiom.com> References: <4hhju8$ne2@tofu.alt.net> <313CE91C.4CB3@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> writes: >E. Hutchinson wrote: >> >> OMNI SEARCH currently has several well experienced >> NEXTSTEP developers available for long term assignments. >> These developers have years of experience that includes >> NEXTSTEP, Objective C, EOF, and Sybase or Oracle. >> Rates can be arranged for as low as $50hr for long >> term assignments. If interested, please contact me >> as soon as possible. >Wow! As low as $50/hr! I don't think I've ever seen a >phrase like that in a recruiting ad before. You've got to admire their candor. Gosh, I'd just *love* to work at Fannie Mae, in a coat and tie, for less than half what I'm getting now, working at home! ;-) -jcr.
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: Releasing elements within NSMutableArray Date: 6 Mar 1996 17:12:01 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4hkh11$kd0@news.next.com> References: <9603050213.AA07982@copernicus> In article <9603050213.AA07982@copernicus> Dan Tapp writes: > I'm sending the [removeAllObjects] method to an instance of class > NSMutableArray, and then following it up with a [release] method to > the array object itself. I had subclassed each element in the array > to release some local variables, and then to send an [autorelease] > command to its superclass (Hmm...wonder if it should've sent > [autorelease] to itself, since I hadn't overridden the call...) > Anyway, the above sequence of calls resulted in each object in the > array also getting passed the release command from the array which > had held it, even though the array had already been requested to > remove the object. (In other words, each element was being > over-released.) When NSMutableArray receives -removeAllObjects, it releases its previous retain on each of the objects. If the array had the only retained reference to each of the objects, the objects are released and deallocated at that point. NSMutableArray's -dealloc method will remove all objects, if any, in the array first (and release its reference to them) before deallocating itself. In your case, NSMutableArray's -dealloc won't have any objects to remove, and will just free the array. Your custom objects should be implementing -(void)dealloc to free themselves (not -release, but it's unclear how you are going about that). Implement -retain and -release only if the object is maintaining its own reference count internally; release should not be used, normally, to invalidate or deallocate objects. Your -dealloc method should release instance variables that are objects, if you previously retained the objects in those ivars, and as the last thing, call [super dealloc];. Objects should rarely (if ever) be autoreleasing or releasing or retaining themselves. They also, in practice, rarely modify their behavior based on the retain count of some other object... > I got rid of the problem by replacing the [super autorelease] > call in each element with a [super release], which of course > released each element immediately and prevented it from > receiving the extra [release] call from the array. If NSMutableArray was sending a -release message to objects after they were removed from the array, replacing [super autorelease] with [super release] would not prevent the objects from receiving that message, and its likely that you'd just end up sending a message to a freed object sooner in your program. Nothing prevents you from sending a message to a freed object. Chris Kane NeXT Software, Inc.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9603061659.AA03506@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Wed, 6 Mar 96 19:59:34 +0300 Subject: Re: NeXTSTEP and JAVA Cc: lionel@cyberlab.ch In article <4hinam$sjp@news.worldcom.ch>, you wrote: > Hello ! > > Does any one know if someone is porting JAVA on NeXTSTEP ? As far as I know there are two different ports BlackJava and set of patches to original Sun's environment a.k.a. NEXTSTEP Java Port. Both could be found on peanuts (munich). I don't know how far they are differ for I haven't look into BlackJava (to unpack it one should license Java from Sun and get crypt key) and while building the other port I've got the following cnext> gmake >>>Recursively making java all... gmake[1]: Entering directory `/Users/zander/Incoming/Java/build/next/java' >>>Recursively making javah all... gmake[2]: Entering directory `/Users/zander/Incoming/Java/build/next/java/javah' gmake ../../../../build/bin/javah_g VARIANT=DBG gmake[3]: Entering directory `/Users/zander/Incoming/Java/build/next/java/javah' single architecture bin directories.... target: {i386} native: {i386} if [ -s .classes.list ] ; \ then ../../../../build/bin/java_g -ms8m -classpath ../../../../build/classes sun.tools.javac.Main -classpath ../../../../build/classes:../../../../src/next/java:../../../../src/share/java:../../../../build/classes-JDK -d ../../../../build/classes \ \ ; \ fi gmake[3]: *** No rule to make target `obj_g/i386/classloader.o', needed by `../../../../build/bin/i386/javah_g'. Stop. gmake[3]: Leaving directory `/Users/zander/Incoming/Java/build/next/java/javah' gmake[2]: *** [debug] Error 2 gmake[2]: Leaving directory `/Users/zander/Incoming/Java/build/next/java/javah' gmake[1]: *** [all] Error 1 gmake[1]: Leaving directory `/Users/zander/Incoming/Java/build/next/java' gmake: *** [all] Error 1 I went thru "One time preparations" 3 times so I guess all the instruction have been followed exactly and cnext>find build/classes -type f -exec touch {} \; didn't help, so could somebody tell me how could I patch original Sun's Java code? Thanks, Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Indexing Kit Design Date: 6 Mar 1996 18:48:27 GMT Organization: University of Ghent, Belgium Message-ID: <4hkmlr$rvd@infoserv.rug.ac.be> don@misckit.com wrote: : If anyone has some ideas on ways to improve the IXKit design, I'd love : to hear about them. Ie., I'd like to hear ideas about how to fix it, : not just that it is "broken". To improve things, there must of course be a consensus that they need improvement... Nothing helps, for making a kit easier to use, better than reducing the number of objects, in my opinion. I really think there should be a "Table" object in the Indexing Kit, a two-dimensional collection, like you have in relational databases, but where in this case, each field would be an Objective C object. That seems to me the natural data-structure for representing, for example, the output of a file-search of a filesystem (level 4 use of IXKit). You could have auxiliary FilePath objects, and these would be the fields in the table that would be the output of the query. But the Table object could cover other levels of use of the IXKit as well. For example, given a text file, I understand that the Indexing Kit is able to compute the number of occurences of words in that file. A Table seems again the natural _user_level_ data-structure for containing the result of such an operation (the rows of the table would be pairs of string and count). I hope you don't immediately brush that proposal away, saying that it would compromise the "power" of the current interface. You're talking about computing "power", not user-friendlyness "power". And with ObjC, you can always internally use non-object representations (for speed). But fixing an API is not easy, so there I would not optimize. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: 6 Mar 1996 18:15:50 GMT Organization: University of Heidelberg, Germany Message-ID: <4hkkom$dnh@sun0.urz.uni-heidelberg.de> References: <4hke19$870@shelob.afs.com> In article <4hke19$870@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > > If NeXT had not decided to so thoroughly change everything, this factor > would definitely come down more on its side. But I have to be honest: > every time I read the documentation on converting NS apps to OS, my > mind starts spinning and I get a headache. I think the first few souls > brave enough to convert a serious-sized app are in for a surprise. Well, there must be someone out there (Omni? Lighthouse? Stone? NeXT?) who _has_ done the conversion, and could illuminate us with *hard* *facts*. Personally, I don't see how a conversion from Common Classes/AppKit to Foundation Kit/Application Kit could be more difficult than Greg's switch to Delphi. sed script or not... Complaining about the size of the method names is a pretty bogus argument, because the new names are unambiguous, and they've removed a lot of discrepancies. "initWithContentsOfFile:" (to pick one off the top of my head) is pretty damned easy to remember, and what's more important, easy for a beginner to pick up. Once I started using Foundation, I was amazed at how naturally it all started coming to me. I find it much easier than the old Common classes. I expect the same to hold true for the Application Kit. Thank God they've tidied up the View object! Except I'll never know, because I've got HP-PA. But let's not open *that* wound (where, I'll say it now, I'll agree with whatever you're going to say, Greg!). ciao, Dave ------------------------------------------------------------------ David A. Coyle Father of Annealer.app. Fission-track research. Ask for my PGP public key. ------------------------------------------------------------------ Max-Planck-Institut f r Kernphysik Heidelberg, Germany ------------------------------------------------------------------ dcoyle@goanna.mpi-hd.mpg.de dcoyle@weizen.rt.schwaben.de ------------------------------------------------------------------
Newsgroups: comp.sys.next.programmer From: rburns@netcom.com (Randall J. Burns) Subject: NextStep Programmers sought by Palo Alto Startup Message-ID: <rburnsDnv1xw.DLE@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Wed, 6 Mar 1996 19:25:08 GMT Sender: rburns@netcom4.netcom.com Filoli Information system seeks additional experienced NextStep programmers. Experience with Sybase or other database systems is helpful. We pay special attention to examples of your previous programming projects and offer competitive compensation and a chance to work with people who are as good as you are. Filoli Information Systems' goal is to become the premier provider of software to insurance companies. We have already fielded two packages to a major customer: Portions of our Underwriting System and our Customer Information Sharing System. We are now completing our Claims Management System. CMS is a document management and decision support system that uses Sybase running on SparcServer 1000's to track a large repository of documents stored on CD-R jukeboxes. CMS will enable our customers to access all documents relating to insurance claims using their computer workstations. CMS will be the first insurance claims management system which will _really_ make good on the promise of of _completely_ paperless operation. We feel that CMS has the potential to revolutionize an important segment of the financial industry-with dramatic improvements in productivity-enabling many insurance companies to switch from maintaining their own extensive staffs of in-house programmers to greater reliance on a turnkey, of-the-shelf solution. Filoli intends to bridge the gap between Silicon Valley and one of the most important segments of the financial industry. If you want to join a company destined to become one of Silicon Valleys greatest success stories, send your resume to jobs@filoli.com.
From: Subrata_Sircar@next.com (Subrata Sircar) Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: 6 Mar 1996 22:24:27 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4hl3ar$ln3@news.next.com> References: <4hke19$870@shelob.afs.com> Gregory H. Anderson writes > If NeXT had not decided to so thoroughly change everything, this factor > would definitely come down more on its side. But I have to be honest: > every time I read the documentation on converting NS apps to OS, my > mind starts spinning and I get a headache. I think the first few souls > brave enough to convert a serious-sized app are in for a surprise. As someone who has converted a serious-sized app to OpenStep from 3.3 NEXTSTEP/EOF 1.1, I'd like to offer a summary of my experience. In general, the conversion scripts did an excellent job. Yes, there were mistakes. Yes, I did have to re-read every source file (skim some, peruse others in detail) - wouldn't you anyway? Yes, you will have to retest the entire app. (Again, wouldn't you anyway?) No, I didn't have to understand lots of new meaning. Method names have changed, and a lot of the code looks different, but it was suprisingly easy to figure out what had happened (and I can't even recite the AppKit names and methods in my sleep :<). Mostly, I re-read the code and referred to the spec sheet for any methods I couldn't figure out from the new names. A lot of the changes simplified my code to the point that I could understand it again :<) I'm not going to pretend that it was trivial. Porting never is. It wasn't as easy as porting to Intel from black hardware, but it was a lot easier than porting this app to any other environment would have been. In general, isn't re-reading it all a lot easier than re-writing it all? -- Subrata Sircar | ssircar@next.com | Prophet & SPAMIT Charter Member I don't speak for NeXT, and they | http://www.next.com/~ssircar don't speak for me. I suspect we | Relationships are complex because both like that. | they're part real & part imaginary. - Martin Terman
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Ivar link to view in ScrollView In-Reply-To: dekorte@suite.com's message of 5 Mar 1996 02:32:57 GMT Message-ID: <RDL.96Mar6212240@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <RDL.96Feb18225038@world.std.com> <4hg94p$fac@news.onramp.net> Date: Thu, 7 Mar 1996 02:22:40 GMT NXBrowser is a kind of View so you can easily make connections to and from it with IB. Please be a bit more specific as to what you are trying to accomplish. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4hg94p$fac@news.onramp.net> dekorte@suite.com (Steve Dekorte) writes: From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Date: 5 Mar 1996 02:32:57 GMT In article <RDL.96Feb18225038@world.std.com> rdl@world.std.com (Robert La Ferla) writes: > You should be able to double-click on the View inside the ScrollView and > Ctrl-drag out a connection. And you can just Ctrl-drag a connection to the > View inside the ScrollView as well. This at least works in IB (v460) which is > the version that I believe came with EOF 1.1. I have that version of IB. But I made a mistake in my post - it's a Browser that I'm trying to do this with - and with a Browser, it doesn't work for me. Anyone know how to get it to work with a Browser?? Steve
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Re: Format and Documentation of eomodel files? Date: 7 Mar 1996 12:28:01 GMT Organization: No organisation supplied Message-ID: <4hmkoh$hr8@hermes.is.co.za> References: <4hfb9r$5av@stc06.ctd.ornl.gov> In-Reply-To: <4hfb9r$5av@stc06.ctd.ornl.gov> On 03/04/96, John W. Wooten wrote: >I'm creating a WebObject example that uses an NSDictionary for storage of the >information. I've tested the NSDictionary with a piece of regular Objective C >code and it works fine. I need to be able to create the eomodel file >describing the NSDictionary data, but don't have Sybase or Oracle that will >connect to the EOModeler. Where can I find the documentation on what to put in >the eomodel file so that I can use it in an NextStep interface (i.e. with >EOBrowser, and table view, etc.). > > > >-- >- - - - - - - - - >J. W. Wooten > The format of an eomodel file is simply a dictionary that was written out to a file. "writeToFile:" method in NSDictionary. The format of the file is quite simple and I suggest that you open an existing model file with Edit.app to see what it looks like. You can always copy some of the stuff out and paste it into your own model file. On the other hand get yourself msql and the related EOF adaptor, then you will have a free database that you can experiment with. I can't exactly recall know where these are but if you want I can email the addresses to you. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and JAVA Date: 7 Mar 1996 12:31:53 GMT Organization: No organisation supplied Message-ID: <4hmkvp$hso@hermes.is.co.za> References: <4hinam$sjp@news.worldcom.ch> In-Reply-To: <4hinam$sjp@news.worldcom.ch> On 03/06/96, Lionel Tinguely wrote: >Hello ! > >Does any one know if someone is porting JAVA on NeXTSTEP ? > Yes, there is a port going on. The site where you can find some of the files is "ftp://ftp.thoughtport.com/pub/next/Java" Hope this helps -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: kiwi@ariane.fb10.tu-berlin.de (Axel Habermann) Newsgroups: comp.sys.next.programmer Subject: Re: inb(), outb(), How? Date: 7 Mar 1996 14:35:15 GMT Organization: Technical University Berlin, Germany Message-ID: <4hms73$rc3@brachio.zrz.TU-Berlin.DE> References: <4hka40$ndm@hpg30a.csc.cuhk.hk> David P. FOK (93925730@plink.cityu.edu.hk) wrote: : Dear Programmers, : I have a big problem on using these 2 functions. All I need to do is a : simple 8-bit I/O but the codes keep on refusing to work. The Builder : reports no problem and the app can be built successfully but whenever : it runs, it quits immediately. inb and outb can only be used from within the kernel, e.g. a loadable kernel server. The kernel of any unix like operating system prevents user programs from accessing the hardware directly. Read 'Writing a loadable kernel server' from the Developer documentation. Write a LKS. Connect your program via mach messages or via the device switch tables (and a node in /dev/) with your driver. -- Axel Habermann kiwi@buran.fb10.tu-berlin.de Fon:+49 30 45478986 Fax:4542296 Die Dateien, in denen die Programmdokumentation enthalten ist, haben normalerweise die Endung ".c", -- Kristian Koehntopp
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: 7 Mar 1996 18:21:14 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <4hn9eq$c6l@shelob.afs.com> References: <4hl3ar$ln3@news.next.com> Subrata Sircar writes > In general, isn't re-reading it all a lot easier than re-writing it all? Probably so. My point was that it _seems_ like so much work, the decision to rewrite seems doesn't seem as onerous by comparison. The potential trapdoors are certainly smaller for an AppKit-derived application, like our trading system, than for applications like WriteUp and PasteUp. The latter make heavy use of NXTypedStreams, -read:, -write:, and a lot of classes that have gone away in FoundationKit. I'd be interested in hearing experiences from early adopters who had to move from DBKit to EOKit. How much effort was it, and did you effectively rewrite those portions of your apps anyway? -- Gregory H. Anderson | "Honey, there're few programming Gaffer/Best Boy/Key Grip | problems that can't be solved Anderson Financial Systems | with duct tape." -- 'Father' Duke greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: eric@skatter.USask.Ca Newsgroups: comp.sys.next.programmer Subject: Re: inb(), outb(), How? Date: 7 Mar 1996 19:53:12 GMT Organization: University of Saskatchewan Message-ID: <4hner8$he0@tribune.usask.ca> References: <4hka40$ndm@hpg30a.csc.cuhk.hk> 93925730@plink.cityu.edu.hk (David P. FOK) wrote: >Dear Programmers, > >I have a big problem on using these 2 functions. All I need to do is a >simple 8-bit I/O but the codes keep on refusing to work. The Builder >reports no problem and the app can be built successfully but whenever >it runs, it quits immediately. > > >Could anyone give me some hints? The hardware I built works okay under >dos by using the inportb() and outport() functions and it requires no >interrupt or DMA. > >In addition, I did really confused by the DriverKit. It is documented >that the IODevice class is a subclass of Object but, I cannot subclass >from the IODevice by any means, why? > You are trying to use kernel-level functions in a user-level program. Your program works in DOS because DOS isn't a *real* operating system :-) A *real* operating system (NEXTSTEP, VMS, UNIX, etc) usually doesn't allow user-level programs to fiddle directly with I/O devices. Only the kernel is permitted to do that. I'm not sure what I can tell you that will help. It sounds like you're going to have to start right with the basics. Some books on the design of the UNIX operating system might be a good place to start. Then look the code for an existing NEXTSTEP device driver. At the very least you're going to have to learn about: Operating system design Loadable device drivers Driver Kit -- Eric Norum eric@skatter.usask.ca Saskatchewan Accelerator Laboratory Phone: (306) 966-6308 University of Saskatchewan FAX: (306) 966-6058 Saskatoon, Canada. NeXTMail accepted.
From: dekorte@intrepid.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Objective-C -> JAVA byte code Date: 7 Mar 1996 19:46:34 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4hneeq$pgg@news.onramp.net> Is anyone working on an Objective-C -> JAVA byte code compiler? Is it possible? (problems with byte code security checks? GC system? AWT?) Having a VM on web clients would be alot more usefull with a dynamic OO language. GNUstep would also have interesting new possibilities... Steve
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <doug@thoughtful.com> Date: Thu, 7 Mar 96 12:10:32 -0700 From: doug@thoughtful.com (Douglas Simons) Message-ID: <9603071910.AA10537@thoughtful.com> Subject: protocol problems We're having a problem with Protocols, in which we haven't found a clean way of avoiding warnings from the compiler. Here's a summary of the situation: We have two protocols, X and Y, which both incorporate a third protocol, Z. Class A adopts protocol X (including Z's methods). Class B is a subclass of A. C is a category of B, and adopts protocol Y. A<X> | <Z> | (subclass) | | B<X>---(category)---C<Y> <Z> <Z> <==== ???? ==-== Class A completely conforms to both protocols X and Z, and so does B by inheritence. Nothing special has to be done in B to make it conform to X and Z. Category C implements all of the methods in protocol Y, except for those that are part of protocol Z, since those are already inherited from class A. Everything works just fine, except the compiler complains that category C doesn't fully implement protocol Y, citing the methods of protocol Z. This seems like a "bug" in the compiler -- it's very annoying that it isn't smart enough to figure out that class B does indeed implement all of the needed methods. Does anyone know a good, clean workaround for this type of problem. We've run into other similar problems and been forced to rearrange the way we work in order to satisfy the compiler. In this current case, we've added the methods of protocol Z into category C, with implementations which simply pass the message to super, but this is a really ugly way to avoid compiler warnings. Any thoughts or suggestions? Thanks, Doug Simons Thoughtful Software
From: ralf@reswi.ruhr.de (Ralf E. Stranzenbach) Newsgroups: comp.sys.next.programmer Subject: DO Serevr Problem Date: 07 Mar 1996 21:28:46 +0100 Organization: News Server fuer en.open.de Sender: ralf@jodokus.ruhr.de Message-ID: <eg2bkmyv5.fsf@jodokus.ruhr.de> Hi, i have some problems implementing a server using Distributed Objects. I've registered my server object (Server) as the delegate of the NXConnection and eacht time a new connection starts up, this method is called: - connection: (NXConnection *) conn didConnect: (NXConnection *) new { fprintf (stderr, "neue Connection %d\n", clients++); [new registerForInvalidationNotification: self]; return new; } After the client is done with the connection, my Server object gets notified about this death using this senderIsInvalid: method. - senderIsInvalid: sender { fprintf (stderr, "connection is lost\n"); // return [sender free]; return self; } It works the first time round, and some times afterwards. But after 128 tries (one after another, not all at the same time) my server drops the following note: handleRequestOnPortal: id 0 not available What's wrong with my server ? - ralf -- Ralf E. Stranzenbach <ralf@reswi.ruhr.de> at Home: +49 2302 / 96200-3 at Work: +49 231 / 975354-31 Die Grossen sind nicht durch sich selbst gross, sondern durch die andern, durch alle die, denen es ein Entzuecken bereitet, sie als gross zu erklaeren. Durch vieler Leute Wuerdelosigkeit entsteht diese eine ueberragende Ehre und Wuerde. Durch vieler Leute Kleinheit und Feigheit entsteht diese auf einem Punkt aufgehaeufte Summe von Groesse und durch vieler Leute Verzicht auf Macht diese gewaltige Macht. Ohne Gehorsam ist der Befehlshaber und ohne Diener ist der Herr nicht moeglich. - Robert Walser
Newsgroups: comp.sys.next.programmer,comp.sys.os.ms-windows.programmer.ole From: sargent@shuriken (Jeff Sargent) Subject: is the D'OLE nextorb timeout configurable? Message-ID: <1996Mar7.212624.11747@il.us.swissbank.com> Keywords: D'OLE, OLE, NEXTSTEP Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Thu, 7 Mar 1996 21:26:24 GMT I have a visual basic for applications program (in Excel) that is synchronously messaging (thru D'OLE) a server application running on a NeXT. D'OLE allows the NeXT process to appear as an OLE automation server. If the server takes longer than 60 seconds to process a method, the VBA program halts with the error: Macro Error Run-time error '1000': Objective-C unhandled exception during invocation of clearAll. It appears the nextorb (or maybe just the OLE hook in VBA/Excel) is timing out. Is there a way to change the value of the timeout? Thanks -- - Jeff sargent@il.us.swissbank.com
From: droux@nmia.com (Nicolas Droux) Newsgroups: comp.sys.next.programmer Subject: Re: inb(), outb(), How? Date: 7 Mar 1996 18:44:51 GMT Organization: New Mexico Internet Access Message-ID: <4hnar3$s8@thales.nmia.com> References: <4hka40$ndm@hpg30a.csc.cuhk.hk> 93925730@plink.cityu.edu.hk (David P. FOK) wrote: >I have a big problem on using these 2 functions. All I need to do is a >simple 8-bit I/O but the codes keep on refusing to work. The Builder >reports no problem and the app can be built successfully but whenever >it runs, it quits immediately. Sure, user processes cannot access IO ports directly. You have to write a kernel-level driver to do that. There are a few drivers with sources on the NeXT ftp sites, you may take a look at them for some examples. Drivers based solely on IO on ports are generally easy to write. You may further create a /dev entry for your driver and use UNIX entry points (see addToCdevswFromDescription) to allow user processes to easily communicate with the driver. Nicolas droux@cs.sandia.gov
From: df@watershed.com (Dirk Fromhein) Newsgroups: comp.sys.next.programmer Subject: Re: DO Serevr Problem Date: Fri, 08 Mar 96 01:17:20 GMT Organization: Watershed Technologies, Inc. Message-ID: <4ho2kj$ro6@caesar.ultra.net> References: <eg2bkmyv5.fsf@jodokus.ruhr.de> If you find an answer to this I'd love to know... this has been troubling us to no end. I don't remember this behavior three years ago! I believe that this bug was introduced with NeXTSTEP 3.3. Also check out Randy Tidd's article on NeXT's web site... From : http://www.next.com/NeXTanswers/HTMLFiles/2040.htmld/2040.html >However, the reference-counting scheme has bugs that still exist in NEXTSTEP >3.3. The reference count of the objects is incorrectly increased and >decreased, resulting in objects being leaked or prematurely freed. These >problems are prevalent and unfortunately there is no known workaround. ^^^^^^^^^^ >Note that in the Foundation Kit implementation of the OpenStep specification, >the autorelease strategy fixes these problems and makes the process of memory >management over (P)DO much cleaner. This is not available with the current >NEXTSTEP 3.3 and PDO implementations, but it is something to look forward to >in NEXTSTEP 4.0. Now what would make me think that there will not be equally deadly bugs in 4.0? How totally like them; give us more money, then we will think about fixing your problem. My trust has been shaken, we are looking at DCE for our transport layer now. Dirk Fromhein df@watershed.com In article <eg2bkmyv5.fsf@jodokus.ruhr.de>, ralf@reswi.ruhr.de (Ralf E. Stranzenbach) wrote: > >Hi, > >i have some problems implementing a server using Distributed Objects. > >I've registered my server object (Server) as the delegate of the >NXConnection and eacht time a new connection starts up, this method is >called: > [munch] >It works the first time round, and some times afterwards. But after >128 tries (one after another, not all at the same time) my server >drops the following note: > handleRequestOnPortal: id 0 not available > >What's wrong with my server ? >
From: ukr8@rzstud2.rz.uni-karlsruhe.de (Jochen Quante) Newsgroups: comp.sys.next.programmer Subject: drawing b/w PS & writing to FAX3-tiff Date: 7 Mar 1996 17:44:47 GMT Organization: University of Karlsruhe, Germany Message-ID: <4hn7af$mrd@nz12.rz.uni-karlsruhe.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit NNTP-Posting-User: ukr8 Hi there, I want to draw an image using postscript and save the resulting ONE BIT bitmap as a CCITTFAX3-compressed TIFF-file. I tried to use an NXBitmapImageRep and draw into that, but it didn't work at all. Either it used a 2 bit depth cache window, or it still used 2 bits per pixel (and 1 bit per sample, but it still didn't work). Then I tried to read an NXBitmapImageRep back from the NXCacheImageRep's cache window, but it still was 2 bits... Nothing worked. :( I thought this must be easy! Any hints? Best regards, Jochen -- =============================================================== | Jochen Quante Computer Science Student | | University of Karlsruhe, Germany | | E-Mail: ukr8@rz.uni-karlsruhe.de S_Quante@ira.uka.de | | WWW: http://www.uni-karlsruhe.de/~ukr8 Jeckel@IRC | ===============================================================
Newsgroups: comp.sys.next.programmer,comp.lang.oberon,comp.object.logic,comp.lang.scheme,comp.lang.beta From: Bob Hathaway <75027.1663@compuserve.com> Subject: Object Currents - Free New WWW Journal - Announcement/Call For Papers Message-ID: <DnxGww.3F8@midway.uchicago.edu> Followup-To: comp.object Summary: Object Currents - Free New Journal - Announcement/Call For Papers Keywords: Free WWW OO Object-Oriented Journal Sender: news@midway.uchicago.edu (News Administrator) Organization: Object Currents Date: Fri, 8 Mar 1996 02:43:44 GMT OBJECT CURRENTS =============== OBJECT CURRENTS ONLINE HYPERTEXT JOURNAL FREE NEW MONTHLY OBJECT-ORIENTED FORUM Location: http://www.sigs.com/objectcurrents/ Editor: Bob Hathaway, bob_hathaway@notesgw.compuserve.com Issues: Jan, Feb, Mar 1996 (All Available now) Publisher: SIGS: C++ Report, JOOP, Object Magazine, ROAD, Smalltalk Report, X Journal, Object Expert, Objeck Spectrum, Java Report, ... This is an invitation to join us at Object Currents and view, engage and participate in new object-oriented technology using the latest form of information technology. Object Currents is a complete new free journal with Feature Articles, Columns, and Departments, along with 7 to 10 *new* articles from SIGS' Journals. We are also in need of contributors for feature articles, which include stipends and the opportunity to publish. Please see our URL for Authors' Guidelines. Object Currents' World Class Columnists: Watts Humphrey: SEI Process Director, CMM & PSP Inventor Bertrand Meyer: Eiffel, OO Design and Software Engineering Francois Bancilhon: President, O2 Technology, Leading ODBMS Expert Michael Jesse Chonoles: Chief of Methodology, Advanced Concepts Center of Lockheed Martin David Shang: OO Programming Language Designer, Motorola Labs Michael Spertus: President, Geodesic Systems, Program Automation Interviews: January: Grady Booch February: James Rumbaugh March: Ivar Jacobson (Part I) - Get the latest on the UML Newsgroup Dialog: Robert Martin Week in OT: Jane Grau 4 times/month Feature Articles: Jan 96 Understanding Objects by Understanding Variables and Functions Dr. Stefano A. Cerri A Comparison of Objects with Frames and OODBs Dr. Xindong Wu Feb 96 On the applicability of the evolving Unified Method to Rapid Functional Prototyping Ian Mitchell Extending the Object Oriented Languages Type Notion Hugues Lecoeuche Mar 96 Back To Basics - Greater Programming Power Multiple Polymorphism and True Type Systems, Recursive Types and Pointer Semantics, Meta Systems and Dynamic Inheritance: Part I Bob Hathaway - Editor Simple Shared Objects Robert Ott Departments: Several, including Code Watch, Question + Answer, and C++ Puzzle. Best new articles from all of SIGS January and February 1996 issues including: C++ Report, JOOP, Object Magazine, ROAD, Smalltalk Report, X Journal, Object Expert, Objeck Spectrum, Java Report, ... Thanks to our readership for patronage, praise, and feedback. Please keep visiting or give it a try soon. Please also feel free to inform friends and colleagues of this free new medium. From the Guidelines: Object Currents' unique hypertext media provides for advances over earlier journals - links to home pages, sites, databases and information servers, interaction, animation, graphics, code retrieval and execution, expanded pages, video, virtual reality and chat sessions. While all of these may not have appeared in these first issues, they will appear in the future. Check it out! Best Regards, Bob Hathaway Robert John Hathaway III Editor in Chief Object Currents Hypertext Journal 655 West Irving Park Road, Suite 5417 Chicago, Illinois 60613-313775 USA Ph: 312-327-2121 Fax: 312-327-2936 Email: 75027.1663@compuserve.com - Correspondence, Submissions bob_hathaway@notesgw.compuserve.com - Submissions rjh@geodesic.com - Unix, Attachments ftp://ftp.geodesic.com/objectcurrents/submissions
From: df@watershed.com (Dirk Fromhein) Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: Fri, 08 Mar 96 01:06:59 GMT Organization: Watershed Technologies, Inc. Distribution: world Message-ID: <4ho218$ro6@caesar.ultra.net> References: <4hl3ar$ln3@news.next.com> <4hn9eq$c6l@shelob.afs.com> Not quite what you asked for (we started from scratch with EOF), but I am also troubled by NeXT's arbitrary changes to method names. We had a hard time with that and EOF. And now it looks like we are in for some of the same changes with EOF 2.0. In some cases change is bad, very bad. If the "porting" effort is too great then I have to agree with Greg, why not go with native code (NO runtime, and standard hardware/software)... and Delphi looks like a real killer. I don't know, the competition was decades behind five years ago. They are now months behind. Dirk Fromhein df@watershed.com In article <4hn9eq$c6l@shelob.afs.com>, Greg_Anderson@afs.com wrote: >Subrata Sircar writes >> In general, isn't re-reading it all a lot easier than re-writing it all? > >Probably so. My point was that it _seems_ like so much work, the decision >to rewrite seems doesn't seem as onerous by comparison. The potential >trapdoors are certainly smaller for an AppKit-derived application, like >our trading system, than for applications like WriteUp and PasteUp. The >latter make heavy use of NXTypedStreams, -read:, -write:, and a lot of >classes that have gone away in FoundationKit. > >I'd be interested in hearing experiences from early adopters who had to >move from DBKit to EOKit. How much effort was it, and did you effectively >rewrite those portions of your apps anyway? >-- >Gregory H. Anderson | "Honey, there're few programming >Gaffer/Best Boy/Key Grip | problems that can't be solved >Anderson Financial Systems | with duct tape." -- 'Father' Duke >greg@afs.com (NeXTmail OK) | (paraphrased), Doonesbury, 2/17/95
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: Fri, 8 Mar 1996 08:02:59 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Mar8.080259.3191@seer.demon.co.uk> References: <4hn9eq$c6l@shelob.afs.com> In article <4hn9eq$c6l@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > Subrata Sircar writes > > In general, isn't re-reading it all a lot easier than re-writing it all? > > Probably so. My point was that it _seems_ like so much work, the decision > to rewrite seems doesn't seem as onerous by comparison. The BIG problem with this question is that anyone who can answer this question appropriately is under non-disclosure. I seem to remember from the classes at ObjectWorld on conversion, that the scripts in question had quite a few areas where they would insert comments into your source, rather than convert the code. The NXTypedStreams that Greg mentioned are probably one example of this. I also remember Steve Jobs saying that it would take one man-month to convert a 50,000 lines of code project. This could include or omit lots of things, but it still points to there being a whole lot more to do than just re-reading code. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Message-ID: <DnyAuE.43D@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4hl3ar$ln3@news.next.com> <4hn9eq$c6l@shelob.afs.com> Date: Fri, 8 Mar 1996 13:30:13 GMT In article <4hn9eq$c6l@shelob.afs.com> Greg_Anderson@afs.com writes: >Subrata Sircar writes >> In general, isn't re-reading it all a lot easier than re-writing it all? Nonsense. Suppose the authors of the Linux kernel adopted an incompatible interface, and then said "yeah, but we're going to provide tools to convert the rest of BSD Unix to our interface". Besides, the conversion tools don't work. Here's what a developer at Omni told me : >> Do they convert or not ? > Directly ? For the vast majority of apps, of course not. Note the words "of course". David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Message-ID: <DnyBp7.4p7@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4hn9eq$c6l@shelob.afs.com> <1996Mar8.080259.3191@seer.demon.co.uk> Date: Fri, 8 Mar 1996 13:48:43 GMT In article <1996Mar8.080259.3191@seer.demon.co.uk> Paul_Lynch@plsys.co.uk writes: >The BIG problem with this question is that anyone who can answer this >question appropriately is under non-disclosure. That is not true. It is known in general that "changes in the large" on a software system are the perfect recipee for disaster. It's not NeXT specific. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 8 Mar 1996 14:53:08 GMT Organization: Norden 1 Communications Message-ID: <4hphkk$7ev@tofu.alt.net> Programmer/analyst NEXTSTEP Objective C EOF Sybase ILL To be considered--fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: DO Serevr Problem Date: 8 Mar 1996 15:14:26 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4hpisi$hqd@lion.embl-heidelberg.de> References: <eg2bkmyv5.fsf@jodokus.ruhr.de> In article <eg2bkmyv5.fsf@jodokus.ruhr.de> ralf@reswi.ruhr.de (Ralf E. Stranzenbach) writes: > It works the first time round, and some times afterwards. But after > 128 tries (one after another, not all at the same time) my server > drops the following note: > handleRequestOnPortal: id 0 not available > > What's wrong with my server ? 1) You should *never* use NXConnection but NXAuroreleaseConnection; 2) -senderIsInvalid has also to free the connection ... this should be already a FAQ. I just checked the experimental SciTools server (it runs last 40 days) and till now it processed successfully over 20k connections. The number 128 should tell you that it's not a random bug ;-) The max number of all connections in a given moment is indeed 128 (and 32 of them are really active ... think so). Happy connecting -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: Bruce Montegani <bmont@rconnect.com> Newsgroups: comp.sys.next.programmer Subject: Precompiled headers Date: Fri, 8 Mar 1996 15:30:42 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <960308093042.245AADyG.bruce@theborg> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII Does anyone know how to read a precompiled header from an application. I'm experimenting on creating my own Class Browser similar HeaderViewer. I noticed that HeaderViewer uses precompiled headers. Thanks Bruce Montegani Objects Alive Custom Software, Inc.
From: ralf@reswi.ruhr.de (Ralf E. Stranzenbach) Newsgroups: comp.sys.next.programmer Subject: Re: DO Serevr Problem Date: 08 Mar 1996 22:10:18 +0100 Organization: News Server fuer en.open.de Sender: ralf@jodokus.ruhr.de Message-ID: <en35r1ebp.fsf@jodokus.ruhr.de> References: <eg2bkmyv5.fsf@jodokus.ruhr.de> In-reply-to: ralf@reswi.ruhr.de's message of 07 Mar 1996 21:28:46 +0100 >>>>> "Ralf" == Ralf E Stranzenbach <ralf@reswi.ruhr.de> writes: In article <eg2bkmyv5.fsf@jodokus.ruhr.de> ralf@reswi.ruhr.de (Ralf E. Stranzenbach) writes: Hi all, Ralf> i have some problems implementing a server using Distributed Ralf> Objects. Ralf> It works the first time round, and some times Ralf> afterwards. But after 128 tries (one after another, not all Ralf> at the same time) my server drops the following note: Ralf> handleRequestOnPortal: id 0 not available Further investigation unsing MallocDebug reveals that the NXConnection objects are leaking. I've got the senderIsInvalid: message from the connection but i would not go away... - ralf -- Ralf E. Stranzenbach <ralf@reswi.ruhr.de> at Home: +49 2302 / 96200-3 at Work: +49 231 / 975354-31 Ich weiss nicht, welches Bild die Welt sich von mir macht, aber mir selbst scheint es so, als sei ich nur ein Junge gewesen, der am Strand spielte und sich damit unterhielt, ab und zu einen glatteren Kiesel oder eine huebschere Muschel als gewoehnlich zu finden, waehrend der weite Ozean der Wahrheit unerkannt vor mir lag. - Sir Isaac Newton
From: MaRK_BeSSeY@NeXT.CoM (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: OPENSTEP conversion (Re: Delphi vs. NEXTSTEP: A summary of what we've learned) Date: 9 Mar 1996 00:14:03 GMT Organization: NeXT Software, Inc. Message-ID: <4hqigb$4nl@news.next.com> References: <DnyAuE.43D@cwi.nl> David Stes writes > Subrata Sircar writes > > In general, isn't re-reading it all a lot easier than re-writing it > > all? > > Nonsense. Suppose the authors of the Linux kernel adopted an > incompatible interface, and then said "yeah, but we're going to > provide tools to convert the rest of BSD Unix to our interface". I don't see the point of this comment. *If* they did change the interface, would you rather they *didn't* provide tools to ease the transition? I think everyone reading this group (or comp.lang.objective-c) knows exactly how you feel about NeXT's FoundationKit, but the discussion here was about whether it was easier to migrate from NEXTSTEP->OPENSTEP than from NEXTSTEP->Something Else. > Besides, the conversion tools don't work. Here's what a developer at > Omni told me : > >> Do they convert or not ? > > Directly ? For the vast majority of apps, of course not. > > Note the words "of course". Well, it's unrealistic (IMO) to expect a perfect, 100% accurate conversion from one paradigm to the other. In practice, most of the code I've written converts quite easily to OPENSTEP. I find that the scripts do a fairly good job of: (1) converting most of the code directly (2) flagging potential trouble spots with the #warning or #error directives The problems I run into with the conversion scripts usually involve: The scripts being too conservative - For instance, almost every use of variables of type char * in my programs could be replaced with the equivalent NSString constructs. The conversion scripts don't do that for me, so I end up with a lot of [someString CString] type constructs in the code where they don't really belong. Fortunately, removing those is fairly easy... Not trying hard enough - In a few cases, the conversion scripts stopped just short of understanding what *should* have been a fairly common construct. This causes some problems when the behavior of the object changed significantly between NEXTSTEP and OPENSTEP (see the OpenPanel for an example). The only *really bad bug* I got from the scripts was where I was doing something wrong already, that didn't happen to cause any problems under 3.3. The converted code ended up releasing an object a little too early, which caused the dreaded [message sent to freed object] exception. However, once I understood what I was doing wrong, even that bug was trivially easy to fix. One last point: The conversion scripts are based on tops, a syntax-aware search and replace engine. This allows you to write custom scripts to handle converting your own API (if you used your own String class, for instance). -Mark -- Mark Bessey NeXT Software, Inc Software Quality Assurance -->I DON'T SPEAK FOR NeXT <--
Newsgroups: comp.sys.next.programmer From: paul@cyantic.com (Paul Guiness) Subject: Pagination and inserting page headers/footers in the Text object. Organization: CYANTIC Systems Date: Fri, 08 Mar 1996 21:08:23 GMT Message-ID: <1996Mar08.210823.20777@cyantic.com> I am trying to insert page headers and footers into a Text object containing RTF text. I have tried two different methods: 1. Inserting a PS header into the PS stream when the Text object prints. 2. Adding header and footer text into the Text object and then printing it. I haven't been able to get either approach to work properly. Can anyone point me in the direction of sample code that can do this? Commercial NeXT Apps seem to offer this capability, but I suspect they probably produce formatted page PS from scratch. I wonder if there is anyway to do formatted pages with headers and footer in PS using the Text object. Perhaps there is a better approach to pagination/header/footers with regard to using the Text object. Any hints or suggestions would be appreciated. Does anyone know if the NS4.0 Text object will support better pagination features? Paul Guinness paul@cyantic.com
From: 93925730@plink.cityu.edu.hk (David P. FOK) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4hka40$ndm@hpg30a.csc.cuhk.hk> Control: cancel <4hka40$ndm@hpg30a.csc.cuhk.hk> Date: Sat, 09 Mar 1996 09:09:30 GMT Organization: CityU of HK Message-ID: <4hqlgr$fpp@hpg30a.csc.cuhk.hk>
From: mermiramar_a@ppp.com(merMiramar Associates) Newsgroups: comp.sys.next.programmer Subject: mer...Renters .... Buy a Home with No $$ Downmer Date: 9 Mar 1996 04:30:42 GMT Organization: adnc.com Message-ID: <4hr1hi$5tl@taurus.adnc.com> WOULD YOU CONVERT YOUR CURRENT RENTAL PAYMENTS INTO MORTGAGE PAYMENTS ON A NEW HOME IF THE DOWN PAYMENT AND CLOSING COSTS WERE TAKEN CARE OF FOR YOU?................................................. IF YOUR EMPLOYER GAVE YOU AN INCREASE IN YOUR TAKE HOME PAY TO COVER THE LEASE PAYMENTS ON A NEW CAR WOULD YOU TRADE YOUR OLD CAR AND LEASE A NEW ONE?............................................ If you answered YES to both questions I have a program which can make this a reality for you...................NOW................ NEW HOME - What may be the most fantastic government give-away began in March 1994 when the Clinton Administration expanded the FHA 203K loan program. This loan permits investors, like the company I represent, to buy a home, that you select, and acquire a government backed loan greater than the cost of the property. This loan can be assumed by qualified buyer with no down payment required. Yes, that's right, the investor can buy the home, add a modest profit and sell it to you with no out of pocket costs and sell it to you at the bank appraised value. NEW CAR - This company has added a unique twist which allows you the opportunity to lease a new car and also receive an increase in your take home pay to cover the payments. The increase does not cost your employer a dime. EARN MORE MONEY - You can become a representative, like myself, and communicate this incredible program to others and earn a small fortune. MORE INFORMATION - I am just one of several thousand independent contractors through out the United States and I would like to send you this nationally recognized company's promotional package on how you can buy a home, trade your old car and lease a new one, increase your take home pay and have the opportunity to make more money. To cover my costs for the package with shipping and handling, please send $5.00 in cash, personal check or money order to I apologize if you feel that this post is not in the proper place, but, we're looking for you! People that are renting, and don't know how to buy a new home are the people that need us the most. We're real....we close residential real estate transactions across the United States daily. The next one can be yours. MIRAMAR ROAD ASSOCIATES 6920 Miramar Road Ste. 207 San Diego, CA 92121.2641 Name ________________________________ Address ________________________________ City ________________________________ State ______________ Zip ___________ Telephone (opt) ______________________________ No salesperson will call. Your information will be held strictly confidential.
From: robert@steffi.dgsys.com (Robert Nicholson) Newsgroups: comp.sys.next.programmer Subject: BNF for plist format? Date: 08 Mar 1996 02:48:10 -0500 Organization: x Sender: robert@steffi.dgsys.com Message-ID: <eka0wyqis.fsf@steffi.dgsys.com> Does anybody have the BNF for NeXT's plist format supported by... NSString's propertyList
From: Subrata_Sircar@next.com (Subrata Sircar) Newsgroups: comp.sys.next.programmer Subject: Re: Delphi vs. NEXTSTEP: A summary of what we've learned Date: 9 Mar 1996 05:14:32 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4hr43o$606@news.next.com> References: <DnyAuE.43D@cwi.nl> David Stes writes >Subrata Sircar writes >> In general, isn't re-reading it all a lot easier than re-writing it all? > Nonsense. Suppose the authors of the Linux kernel adopted an incompatible > interface, and then said "yeah, but we're going to provide tools to convert > the rest of BSD Unix to our interface". Isn't that what moving between BSD and System V is like? :<) In all seriousness, I did say "In general". First, the interfaces aren't *that* different. Second, EOF apps will have an easier time of it, since most EOF apps I've seen use Foundation internally, which makes the transition much less of a production. If, on the other hand, you're talking about porting palettes which archived stuff via NXTypedStreams, for example, that's another problem. > Besides, the conversion tools don't work. Here's what a developer at Omni > told me : > >> Do they convert or not ? > > Directly ? For the vast majority of apps, of course not. (It's a big leap from "Directly? ... Of course not." to "They don't work", btw. How'd you get over that chasm? :<) Note the word "Directly". If you expect that the scripts will turn your 3.3 code to error-free, does-what-I-meant-it-to-do 4.0 code, you probably won't get what you expect. (If you expect that the scripts will turn your 3.3 code to compiling 4.0 code that needs to be re-reread and re-tested before you deploy it, then you'll get more than you expect, but not much.) Here's a developer at NeXT telling you, "Directly? For the app(s) I work on, yes." just so you have another data point :<) (For the record, I work on BugNeXT and internal tools. BugNeXT converted without a hitch, compiled and had 95% the same behavior as before. It turned out that most of it was actually much much easier to write using Foundation and the new AppKit, and so for the sake of maintaining it I rewrote that part. That's not a knock on the conversion scripts, but a merit for the new libraries.) That's not to say that your developer at Omni is wrong. To pull up a quote of Greg Anderson's from the article you quoted: > The potential trapdoors are certainly smaller for an > AppKit-derived application, like our trading system, than > for applications like WriteUp and PasteUp. The latter make > heavy use of NXTypedStreams, -read:, -write:, and a lot of > classes that have gone away in FoundationKit. If the Omni people were using a lot of that (or, in their experience, the vast majority of apps use that functionality), I can certainly see where they're coming from. My experience is that most apps don't, but I've leaned as heavily on the AppKit as I could from the moment I found out it existed. (That kind of tool is what pulled me to NeXT.) -- Subrata Sircar | ssircar@next.com | Prophet & SPAMIT Charter Member I don't speak for NeXT, and they | http://www.next.com/~ssircar don't speak for me. I suspect we | Relationships are complex because both like that. | they're part real & part imaginary. - Martin Terman
From: Subrata_Sircar@next.com (Subrata Sircar) Newsgroups: comp.sys.next.programmer Subject: Re: OPENSTEP conversion (Re: Delphi vs. NEXTSTEP: A summary of what we've learned) Date: 9 Mar 1996 05:20:47 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4hr4ff$61g@news.next.com> References: <4hqigb$4nl@news.next.com> Mark Bessey writes > For instance, almost every use of variables of type char * in my > programs could be replaced with the equivalent NSString constructs. The > conversion scripts don't do that for me, so I end up with a lot of > [someString CString] type constructs in the code where they don't really > belong. Fortunately, removing those is fairly easy... There's an optional set of scripts that you can run which will do deep conversion of the code, which can eridicate the char * references and leave nothing but NSStrings behind. The reason they're optional is that some API (like libc) doesn't use NSString ... > One last point: The conversion scripts are based on tops, a syntax-aware > search and replace engine. This allows you to write custom scripts to > handle converting your own API (if you used your own String class, for > instance). This is how the above functionality works: it scans your code, builds a set of optional replacements (which you can then edit), and then runs that conversion. Very handy for EOF/3.3-based apps, which often have the schizophrenic Foundation-internals/Appkit-UI NSString/char * wars ... -- Subrata Sircar | ssircar@next.com | Prophet & SPAMIT Charter Member I don't speak for NeXT, and they | http://www.next.com/~ssircar don't speak for me. I suspect we | Relationships are complex because both like that. | they're part real & part imaginary. - Martin Terman
From: mermiramar_a@ppp.com(merMiramar Associates) Newsgroups: comp.sys.next.programmer Date: 09 Mar 1996 10:37:31 EST Control: cancel <4hr1hi$5tl@taurus.adnc.com> Subject: cmsg cancel <4hr1hi$5tl@taurus.adnc.com> Message-ID: <cancel.4hr1hi$5tl@taurus.adnc.com> Spam cancelled by dsr@lns598.lns.cornell.edu
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: NS encoding <-> Latin 1? Date: 9 Mar 1996 20:38:37 GMT Organization: Turbocat's Development, Germany Message-ID: <4hsq8d$uf@turbocat.snafu.de> Hi! Is there a simple way to convert ISO Latin 1 to NEXTSTEP encoded strings? (and the other way) Thanks in advance. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: OPENSTEP conversion (Re: Delphi vs. NEXTSTEP: A summary of what we've learned) Date: 10 Mar 1996 16:16:00 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4huv80$oic@lion.embl-heidelberg.de> References: <4hqigb$4nl@news.next.com> In article <4hqigb$4nl@news.next.com> MaRK_BeSSeY@NeXT.CoM (Mark Bessey) writes: > David Stes writes > > Subrata Sircar writes > > > In general, isn't re-reading it all a lot easier than re-writing it > > > all? > > > > Nonsense. Suppose the authors of the Linux kernel adopted an > > incompatible interface, and then said "yeah, but we're going to > > provide tools to convert the rest of BSD Unix to our interface". > > I don't see the point of this comment. *If* they did change the > interface, would you rather they *didn't* provide tools to ease the > transition? I think everyone reading this group (or > comp.lang.objective-c) knows exactly how you feel about NeXT's > FoundationKit, but the discussion here was about whether it was easier > to migrate from NEXTSTEP->OPENSTEP than from NEXTSTEP->Something Else. ;-) ;-) -) ;-) ;-) -) ;-) ;-) -) ;-) ;-) -) You wrote it Mark ;-) I was trying to keep my word producing device silent because I was afraid that David will start claiming that I'm the devil himself ;-). A year ago I was stupid enough telling him that the new Foundation's archiving is better then the old one. Since that time I'm on his black list! Of course it doesn't stop me producing OS sources ... but there are people around who really love backwards compatibility a la M$'s DOS. God bless the NeXT developers not to be so stupid and for giving them enough courage to change the old interface! What surprise me again and again is that there are still people claiming that converting from NS to OS is more difficult than porting NS sources to Delphi or any other environment. This is just absurd! I can find just 4 possible explanation about these claims: - These people have strange sense of humor; - They are members of the SSIII club (a group of people blaming NeXT for their personal misssuccess. There is only one rule to get a membership card for this club: never say a positive word about NeXT Inc. - easy! ;-) - They never understood the concepts behind OOP; - Their old sources are so horrible, that the new conversion tools (and the authors themselves) have hard time tp understand them; > > Besides, the conversion tools don't work. Here's what a developer at > > Omni told me : > > >> Do they convert or not ? > > > Directly ? For the vast majority of apps, of course not. > > > > Note the words "of course". Hmmm. Onece I've seen parts of the OmniWeb sources (very early version). If they keep the same spaghetti, well, no wonder having problems "of course"... > The scripts being too conservative - > For instance, almost every use of variables of type char * in my > programs could be replaced with the equivalent NSString constructs. The > conversion scripts don't do that for me, so I end up with a lot of > [someString CString] type constructs in the code where they don't really > belong. Fortunately, removing those is fairly easy... Hmm. I was told there is a flag ... but I still have no beta, so I may be wrong... > -->I DON'T SPEAK FOR NeXT <-- -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 11 Mar 1996 05:15:09 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4i0cst$fs8@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Pagination and inserting page headers/footers in the Text object. Date: Mon, 11 Mar 1996 08:56:06 +0000 Organization: Cedar Systems Distribution: world Message-ID: <NamDyFAmq+QxEwdJ@cedar.co.uk> References: <1996Mar08.210823.20777@cyantic.com> MIME-Version: 1.0 In article <1996Mar08.210823.20777@cyantic.com>, Paul Guiness <paul@cyantic.com> writes >I am trying to insert page headers and footers into a Text object containing RTF >text. I have tried two different methods: > 1. Inserting a PS header into the PS stream when the Text object prints. > 2. Adding header and footer text into the Text object and then printing >it. > ... >Perhaps there is a better approach to pagination/header/footers with regard to >using the Text object. Any hints or suggestions would be appreciated. > There certainly is! Cedar Systems have developed extensive extensions to the Text classes which include, amongst many other features, pagination with headers and footers. If you would like further information, please email me with your requirements. Regards, Paul. ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: mikem@afs.com (Mike Matlack) Newsgroups: comp.sys.next.programmer Subject: rtf2ps Date: 11 Mar 1996 16:40:18 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4i1l1i$r5t@shelob.afs.com> Anyone have the source for rtf2ps, a utility that converts rtf files to postscript? The version I have does not deal with landscape pages correctly. Thanks! -- Michael J. Matlack Anderson Financial Systems, +1 215 653 0911 Mike_Matlack@afs.com (MIME & NEXT mail OK)
Newsgroups: comp.sys.next.programmer From: martin@rat.se Subject: Re: Pagination and inserting page headers/footers in the Text object. Message-ID: <Do3I0r.8tI@rat.se> Keywords: headers footers drawPageBorder Sender: martin@rat.se (Martin Wennerberg) Organization: Research & Trade, AB. References: <1996Mar08.210823.20777@cyantic.com> Date: Mon, 11 Mar 1996 08:53:08 GMT The general way to print headers and footers for a view is to override the View method -drawPageBorder::. One way to do this for a text object is to implement a subclass of Text that asks the delegate for a background view (containing the header/footer). - drawPageBorder:(float)width :(float)height { View *paperView = [delegate paperView]; if (paperView) { [[NXApp focusView] addSubview:paperView]; [paperView display]; [paperView removeFromSuperview]; } return self; } One can then do a +poseAsClass to make all text instances do this. /Martin -- Martin Wennerberg ____________________________________________________ Research & Trade AB Phone: + 46 - 8 - 21 17 50
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: quad fat 3/11 build of cvs available Date: 11 Mar 1996 18:48:29 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4i1sht$9cb@chinx4.thoughtport.net> A Quad-fat build of cvs version 1.7 is now available from ftp.thoughtport.com (URL below). The [rather large] package includes all resources necessary to run cvs under NEXTSTEP (including example repository configuration files extended for use with NEXTSTEP). The package should be installed as /usr/local/cvs -- I suggest you untar it to /usr/local/ and create a symbolic link from the directory that is created to '/usr/local/cvs'. Once installed, execute 'cvsinit' to create or update a repository. [Note that there appears to be a bug in Emacs that requires a cvs binary to be present in /usr/local/bin -- I recommend that you create a symbolic link from /usr/local/cvs/bin/cvs to /usr/local/bin/cvs] If anyone needs source, send me mail. b.bum
From: hugues@precipice.fdn.org (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: Re: NS encoding <-> Latin 1? Date: 11 Mar 1996 18:48:47 GMT Organization: Individual - France Distribution: world Message-ID: <4i1sif$15f@precipice.fdn.fr> References: <4hsq8d$uf@turbocat.snafu.de> In article <4hsq8d$uf@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) writes: > Hi! > > Is there a simple way to convert ISO Latin 1 to NEXTSTEP encoded strings? > (and the other way) You should find that in the Alexandra news reader sources... See the peanut's or Muenchen archives... It may also be in the MiscKit. if not wrote it and give them :-) Hugues. -------------------------------------------------------------------- hugues@precipice.fdn.fr - France (small NeXTMail OK) ------------ NS3.2 ------------ NS3.0J ------------ :-) ------------
From: Dmichael@rumah.rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: latest gnu c++ Date: 12 Mar 1996 13:30:48 GMT Organization: Unconfigured Message-ID: <4i3ua8$gs6@jaring.my> Does the latest gnu c++ have templates and exception handling supported? -- Dr. Michael Olan Email: michael@rumah.pc.my (NeXT Mail OK) Lecturer - Computer Science michael@ppp.itm.my American Degree Program Institut Teknologi MARA Section 17, Shah Alam, Malaysia ----------------------------------------------------------- "To see the universe as it is, you must step beyond the net. It is not hard to do so, for the net is full of holes." -- Sri Nisargadatta -----------------------------------------------------------
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Contract/DC Area Date: 12 Mar 1996 18:55:08 GMT Organization: Norden 1 Communications Message-ID: <4i4hac$h9k@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF Sybase Contract--long, long term DC Area To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: MaRK_BeSSeY@NeXT.CoM (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: OPENSTEP conversion (Re: Delphi vs. NEXTSTEP: A summary of what we've learned) Date: 12 Mar 1996 19:03:15 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4i4hpj$snq@news.next.com> References: <4huv80$oic@lion.embl-heidelberg.de> Georg Tuparev writes > Mark Bessey writes: > > The scripts being too conservative - > > For instance, almost every use of variables of type char * in my > > programs could be replaced with the equivalent NSString constructs. > > The conversion scripts don't do that for me, so I end up with a lot > > of [someString CString] type constructs in the code where they don't > > really belong. Fortunately, removing those is fairly easy... > Hmm. I was told there is a flag ... but I still have no beta, so I may > be wrong... No, it's been pointed out to me that this does work (now). I really should try to keep up with the changes in the newer versions... Thanks, all. -- Mark Bessey NeXT Software, Inc Software Quality Assurance -->I DON'T SPEAK FOR NeXT <--
From: Joe Pompei <pompei@audia.music.nwu.edu> Newsgroups: comp.sys.next.programmer Subject: help! NXBrowser usage Date: Tue, 12 Mar 1996 16:29:43 -0600 Organization: Northwestern University, Evanston, IL, US Message-ID: <3145FAD7.41C6@audia.music.nwu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I'm developing a NeXT application and would like to use a browser to keep track a list of active files (much like SoundRecorder's). I need to be able to add, insert, and remove items from the list, and do something with a selected item (basically just know what it is). It seems like it's a lot harder than it should be (much like the popuplist).. and I'm quite confused about how the process works - the librarian's info was very incomplete. From what I can tell, to add list items I need a delegate to do it.. now, what messages does this delegate send to the browser, and how does it link to my controller object? To get info about what is selected, some kind of message is coming out of the 'target' outlet.. but what is the message, and how do I get it? they could have made this a lot easier... Any help is appreciated... Joe
From: efr611@shell01.ozemail.com.au (PAUL ARMSTRONG) Newsgroups: comp.sys.next.programmer Subject: barcode algorithm Date: 13 Mar 1996 15:45:22 GMT Organization: OzEmail Pty Ltd - Australia Message-ID: <4i6qii$b4a@oznet07.ozemail.com.au> any one knows where can i find an algorithms to print bar codes??? any help would be appreciate. thanks
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: Re: DO Serevr Problem Date: 13 Mar 1996 15:06:30 -0600 Organization: MCSNet Services Distribution: world Message-ID: <4i7dcm$fls@Mercury.mcs.com> References: <eg2bkmyv5.fsf@jodokus.ruhr.de> <4hpisi$hqd@lion.embl-heidelberg.de> Georg Tuparev (tuparev@mailserver.EMBL-Heidelberg.DE) wrote: : In article <eg2bkmyv5.fsf@jodokus.ruhr.de> ralf@reswi.ruhr.de (Ralf E. Stranzenbach) : writes: : > It works the first time round, and some times afterwards. But after : > 128 tries (one after another, not all at the same time) my server : > drops the following note: : > handleRequestOnPortal: id 0 not available : > : > What's wrong with my server ? : 1) You should *never* use NXConnection but NXAuroreleaseConnection; Why should you always use NXAutoReleaseConnection? What if the program in question has no interaction with the DPS system? When do the autoreleased connections get freed? : 2) -senderIsInvalid has also to free the connection ... this should be already a : FAQ. I just checked the experimental SciTools server (it runs last 40 days) and till : now it processed successfully over 20k connections. Thanks for the information Peter Richardson
From: katzlbt@vuse.vanderbilt.edu (Thomas Katzlberger) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: _G_config.h in gcc-2.7.[12] missing ? Date: 14 Mar 1996 00:22:43 GMT Organization: a white NeXT Message-ID: <4i7osj$g56@news.vanderbilt.edu> Hi all, Hmmm, I just wanted to use iostreams with gcc and ran into compiler errors with this missing _G_config.h file. It is inclued in NeXT's g++ library but doesn't create itself in the gcc distribution ... hmmm ? (And the old one doesn't work) Is anyone using iostreams sucessfully on NeXT with gcc-2.7.2 ? I really like printf better but there is no choice ... Thanks, Cat. -- _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/_/ _/_/ _/_/ Thomas Katzlberger _/_/ _/_/ katzlbt@vuse.vanderbilt.edu _/_/ _/_/ @aWhiteNeXT.called.garfield _/_/ _/_/ http://www.vuse.vanderbilt.edu/~katzlbt/ _/_/ _/_/ _/_/ _/_/ "You can tune a file system, but you can't tune a fish." _/_/ _/_/ UNIX man page for tunefs. _/_/ _/_/ _/_/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
Newsgroups: comp.sys.next.programmer From: kjell@bart (Kjell Nilsson) Subject: appkitServer API Message-ID: <Do98Bs.8q6@oops.se> Sender: usenet@oops.se Organization: OOPS art, HB Date: Thu, 14 Mar 1996 11:09:27 GMT By coincident I stumbled on this error message in the console. Mar 13 16:21:17 bart Diagram[7364]: Assertion failed: _registerDocAt: appkitServer didn't recognize our doc Diagram seems to comunicate with the appkitServer. I have tried to find some doc about the appkitServer process but with no luck. Is there a hidden API to this process? What does this process do? It seems to register documents if I understand the errormessage right, but why? -- -- Kjell Nilsson - Member of Swedish Object Guild OOPS art HB, Phone +46-31-499713, Fax +46-31-474594 kjell@oops.se - NeXTMail, Mime welcome
From: katzlbt@vuse.vanderbilt.edu (Thomas Katzlberger) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: cmsg cancel <4i7osj$g56@news.vanderbilt.edu> Control: cancel <4i7osj$g56@news.vanderbilt.edu> Date: 14 Mar 1996 17:13:51 GMT Organization: a black NeXT Message-ID: <4i9k4f$fd7@news.vanderbilt.edu> Article cancelled from within tin [v1.2 PL2]
From: Puchau Jara <cur3n83> Newsgroups: comp.sys.next.programmer Subject: how can I get software to my new NeXTStation Date: 14 Mar 1996 17:33:40 GMT Organization: Universidad Autonoma de Madrid, Spain Message-ID: <4i9l9k$qu@acebo.sdi.uam.es> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit I started to work with my NeXTStation few days ago. I have an old version of the Operative System, and i´m looking for better software. While i do so, i´m trying to write programs in c++ for mi NeXT. I found sites with GNU software, but i no have pkunzip ( I don´t have gcc to compile it :( ). Any Idea? Info and MIME are welcome. Alejandro.Puchau@ii.uam.es
From: zazula@running-start.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: help! NXBrowser usage Date: 14 Mar 1996 15:04:44 GMT Organization: Running Start, Inc. Message-ID: <4i9cic$s4q@nntp1.best.com> References: <3145FAD7.41C6@audia.music.nwu.edu> In-Reply-To: <3145FAD7.41C6@audia.music.nwu.edu> Hi - On 03/12/96, Joe Pompei wrote: >I'm developing a NeXT application and would like to use a browser to >keep track a list of active files (much like SoundRecorder's). > > I need to be able to add, insert, and remove items from the list, and >do something with a selected item (basically just know what it is). > > It seems like it's a lot harder than it should be (much like the >popuplist).. and I'm quite confused about how the process works - the >librarian's info was very incomplete. > You can examples using the NXBrowser class in the NeXT mini-example "NormalBrowser" (http://www.next.com/NeXTanswers/) and more extensive use in my "RZToDoList" program (/http://www.running-start.com/downloads/sources.html). >From what I can tell, to add list items I need a delegate to do it.. >now, what messages does this delegate send to the browser, and how does >it link to my controller object? > The basic method you need to implement as the NXBrowser's delegate is: - (int)browser:sender fillMatrix:matrix inColumn:(int)column; You will be sent this message when the browser wants to display itself. Your responsibility is to add or remove cells from "matrix" and set the titles of these cells. The return value indicates the number of cells you have to display. Another method you'll want to know about is NXBrowser's -loadColumnZero. You send this to an NXBrowser to get it to reload its contents (via your delegate method) and then redisplay. >To get info about what is selected, some kind of message is coming out >of the 'target' outlet.. but what is the message, and how do I get it? > There are two ways to get the selection. As you mention, you can become the target of the browser and keep some state around every time the user selects a cell. However, this will not work if you allow multiple selections in the browser. A better way might be to query the NXBrowser when you need to know the selection. There are two methods in NXBrowser that you can use, depending on its configuration. The first is -selectedCell. This will return the NXBrowserCell that is currently selected. You should use this when your browser is NOT set to allow multiple selection. The second method is -getSelectedCells:. This method will either fill a list you provide with the selected cells, OR, return a list if the argument is 'nil'. I tend to use the latter, e.g.,: List *selecteCells = [browser getSelectedCells:nil]; NOTE: there appears to be a BUG in NXBrowser that makes the -getSelectedCells: method ALWAYS return 'nil' if the NXBrowser does not allow multiple selection. This is really nasty and can drive you crazy if you don't know about it... Also, -selectedCell is pretty meaningless for a NXBrowser with multiple selection enabled. Here's a general method you can use to make this easier: - (List *)selectedCells:(NXBrowser *)browser { List *list = nil; if(browser) { if([browser isMultipleSelectionEnabled]) { list = [[browser matrixInColumn: [browser selectedColumn]] getSelectedCells:nil]; } else if([browser selectedCell]) { List *list = [[List alloc] init]; [list addObject:[browser selectedCell]]; } } return list; } Of course, the above could be reworked to use NSArray's if you are using Foundation. >they could have made this a lot easier... > Well, NXBrowser's aren't too bad once you figure them out. It's much simpler than the Matrix in a ScrollView stuff... Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@running-start.com http://www.running-start.com http://www.running-start.com/zazula/index.html
From: Luke Howard <lukeh@auswired.net> Newsgroups: comp.sys.next.programmer Subject: NSBundle and shared library problem Date: 14 Mar 1996 10:41:59 GMT Organization: AusWired, Melbourne, Australia. Message-ID: <4i8t5n$59n@news.mel.aone.net.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I'm trying to dynamically load some code into an application for which I don't have the source. The bundle being dynamically loaded has some dependencies on shared libraries whose stubs aren't linked by the application which is doing the dynamic loading. I can link the libXXX_s.a libraries against the bundle's object file using ld -r, and I get an object file suitable for further linking. Unfortunately, the application which is trying to NSBundle-load this code complains about not being able to dynamically load fixed VM shared library. If I don't link in libXXX_s.a, I get undersolved symbols (as would be expected). Is there a way around this, or does a bundle have to have all code that isn't in the loading object file statically linked into it? Any help would be appreciated; otherwise, I'm going to have to use a DO server just so I can access the library, and that's hardly the most efficient means of deploying this particular application. -- Luke.
From: dcoyle@goanna.mpi-hd.mpg.de (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Re: how can I get software to my new NeXTStation Date: 14 Mar 1996 19:11:08 GMT Organization: University of Heidelberg, Germany Message-ID: <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> References: <4i9l9k$qu@acebo.sdi.uam.es> In article <4i9l9k$qu@acebo.sdi.uam.es> Puchau Jara <cur3n83> writes: > I started to work with my NeXTStation few days ago. I have an old > version of the Operative System, and i m looking for better > software. While i do so, i m trying to write programs in c++ for > mi NeXT. I found sites with GNU software, but i no have pkunzip > ( I don t have gcc to compile it :( ). Any Idea? Info and MIME > are welcome. > Alejandro.Puchau@ii.uam.es Alejandro: 1) your NeXTstation isn't new 2) you DO have gcc 3) if you really have a NeXTstation, you will be programming in Objective-C 4) You can get almost all the software you need and more at: ftp:peanuts.leo.org/pub/comp/platforms/next/ 5) You should upgrade to NextStep 3.3 as soon as your budget/disk space/ RAM allow it. Hope this helps, Dave
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: different DO problem Date: 14 Mar 1996 14:05:16 -0600 Organization: MCSNet Services Message-ID: <4i9u5s$mkc@Mars.mcs.com> I have a problem with DO that I can't find the answer to. The "server" often takes an hour or more to process a request. This leads me to want to use a "oneway" call so that the client can continue other operations. I want to pass something to the server so that it can send back status information periodically. The question is, what has to be done in the client to allow the callback once the method that called the server completes? If I don't make this a oneway call and make the client wait, it works fine. Thanks for any information. Peter Richardson
Newsgroups: comp.sys.next.programmer From: markus g <tm@burrow.muc.de> Subject: Re: [q] perl5, where to put it? In-Reply-To: crath@bnr.ca's message of 01 Mar 1996 15:33:34 GMT Message-ID: <7xg2bcii37.fsf@burrow.muc.de> To: crath@bnr.ca (Christopher Rath) Sender: tm@burrow.muc.de Organization: hardly any... References: <7xravdydkz.fsf@burrow.muc.de> <CRATH.96Mar1103334@bmerhe83.bnr.ca> Date: Thu, 14 Mar 1996 01:22:52 GMT >>>>> "CR" == Christopher Rath <crath@bnr.ca> writes: CR> One of perl's problems, that it uses hard-coded paths to find CR> its files, is also one of its strengths. We have overcome CR> this perl4/perl5 problem here within my company by installing CR> perl4 in its own directory tree (/opt/corp/local/perl4) and CR> perl5 in its own tree (/opt/corp/local/perl5). We then CR> install a couple of symlinks [...] thanks for your reply. but i'm not concerned about the binaries it's rather the libraries where i fear conflicts. markus g -- / . . . . . . . . . . . . . . . . . . . . . . . . . . __ __/ __ __ . . . . . . . . . . . . . . . . . . . . . the mole / / / / . . . . . . . . . . . . . . . . . tm@burrow.muc.de __/ __/ __/ __/ . . . . . . . . . . . . http://www.muc.de/~mgloede/
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer Subject: Re: how can I get software to my new NeXTStation Date: Thu, 14 Mar 1996 18:27:46 -0500 Organization: Steel Driving Software, Inc. Message-ID: <3148AB72.1B3F@ix.netcom.com> References: <4i9l9k$qu@acebo.sdi.uam.es> <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit David A. Coyle wrote: > > In article <4i9l9k$qu@acebo.sdi.uam.es> Puchau Jara <cur3n83> writes: > > I started to work with my NeXTStation few days ago. I have an old > > version of the Operative System, and i m looking for better > > software. While i do so, i m trying to write programs in c++ for > > mi NeXT. I found sites with GNU software, but i no have pkunzip > > ( I don t have gcc to compile it :( ). Any Idea? Info and MIME > > are welcome. > > Alejandro.Puchau@ii.uam.es > Alejandro: > 1) your NeXTstation isn't new > 2) you DO have gcc Is gcc installed if it's not a Developer machine? > 3) if you really have a NeXTstation, you will be programming in > Objective-C Or C, or C++. > 4) You can get almost all the software you need and more at: > ftp:peanuts.leo.org/pub/comp/platforms/next/ > 5) You should upgrade to NextStep 3.3 as soon as your budget/disk space/ > RAM allow it. What for?
From: purdys7@aol.com (Purdys7) Newsgroups: comp.sys.next.programmer Subject: NextStep Developer Wanted - NYC area Date: 15 Mar 1996 00:56:15 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4ib0pv$5ds@newsbf02.news.aol.com> NextStep developer/sys admin wanted: NY metro area. 2-4 month assignment with opportunity for full-time position. Experience with C, Objective C, EOF, and QuickBase required. Exciting applications. Reply to us, the employer, directly at purdys7@aol.com (we are not a recruiter or agency).
From: cnyap@next (Chih Nam Yap) Newsgroups: comp.sys.next.programmer Subject: "Not Applicable" Date: 15 Mar 1996 13:00:10 GMT Organization: University of Sheffield, UK Message-ID: <4ibpkq$eqf@bignews.shef.ac.uk> Hi there Can anyone tell me what is the simple way to show the word "Not Applicable" on an inspector window ? Thank you. c.Yap
From: dse@dse.ho.att.com (att.general) Newsgroups: comp.sys.next.programmer Subject: Help. Setting TextField colors Date: 15 Mar 1996 13:54:27 GMT Organization: AT&T Bell Laboratories, Columbus, Ohio Message-ID: <4ibsqj$9ke@nntpb.cb.att.com> I'm trying to use setBackgroundColor: to set a the color of TextFields. The color changes OK, but the selection indication is lost. That is double clicking on text selects the text, but there is no visual indication. In fact it appears as if the cursor is lost, until one single clicks to bring the I-beam cursor back. Any hints on how to fix this problem? Thanks, Don
From: woo@polaris.scicntr.ortn.edu (J. W. Wooten) Newsgroups: comp.sys.next.programmer Subject: missing "flag" on nib file in ProjectBuilder Date: 15 Mar 1996 13:57:04 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4ibsvg$7t9@stc06.ctd.ornl.gov> I've been converting an older palette over to newer ProjectBuilder and IB. Most everything is going fine, BUT I have two nib files in the ProjectBuilder, JoyStickInspector.nib and JoyStickPalette.nib. The second one has a little flag in front of it but the first one doesn't. In other cases, all the nib files have the little flag. What would cause this one not to have the flag? Is this a "bad thing"? thanks, John Wooten
From: caljouwr@cadvision.com Newsgroups: comp.sys.next.programmer Subject: Debugging bundles Date: 15 Mar 1996 14:51:37 GMT Organization: CADVision Distribution: world Message-ID: <4ic05p$2t1m@huey.cadvision.com> Does anyone know if there is a way to use a source level debugger when debugging bundles. According to NeXTAsnwers if you make the call [NSBundle stripAfterLoading:NO]; you will be able to use the debugger but this call must be made before any bundles are loaded. I am working on a bundle for a NeXT application so I don't have access to the main calling routines, only my own source for the bundle. If any one has any information it would be greatly appreciated! Debugging using printing and logging of information becomes rather tedious and time consuming. Thanks in advance, Rob Caljouw
From: gene@fw.bluestone.com (Dr. Gene De Lisa) Newsgroups: comp.sys.next.programmer Subject: Re: help! NXBrowser usage Date: 14 Mar 1996 13:42:00 -0500 Organization: bluestone.com Message-ID: <4i9p9o$eve@fw.bluestone.com> References: <3145FAD7.41C6@audia.music.nwu.edu> In article <3145FAD7.41C6@audia.music.nwu.edu>, Joe Pompei <pompei@audia.music.nwu.edu> wrote: >I'm developing a NeXT application and would like to use a browser to >keep track a list of active files (much like SoundRecorder's). > I need to be able to add, insert, and remove items from the list, and >do something with a selected item (basically just know what it is). If you don't need to show a heirarchy then why not just use a matrix? It is pretty easy to add cells to a matrix and it can be configured to look like a "list". -- Dr. Gene De Lisa Bluestone Inc. 1000 Briggs Rd. Mt. Laurel, NJ 08054-4101 gene@bluestone.com Vox Humana: (609) 727-4600 x261 Fax: (609) 778-4925 <a href "http://www.bluestone.com">Bluestone Home Page</a>
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: Re: different DO problem Date: 15 Mar 1996 13:13:14 -0600 Organization: MCSNet Services Message-ID: <4icfga$pu7@Venus.mcs.com> References: <4i9u5s$mkc@Mars.mcs.com> Peter Richardson (par@MCS.COM) wrote: : I have a problem with DO that I can't find the answer to. : The "server" often takes an hour or more to process a request. This : leads me to want to use a "oneway" call so that the client can : continue other operations. I want to pass something to the server : so that it can send back status information periodically. : The question is, what has to be done in the client to allow the : callback once the method that called the server completes? If I : don't make this a oneway call and make the client wait, it works fine. Never mind. I found the answer on the NeXTAnswers page. I simply needed to send a runFromAppKit message to the connection on the client end. Peter Richardson
From: bff@icarus.pvh.org (Brendan Forsyth) Newsgroups: comp.sys.next.programmer Subject: selectedCell Stuff Date: 15 Mar 1996 21:19:10 GMT Organization: SuperNet Inc. +1.303.296.8202 Denver Colorado Message-ID: <4icmse$klm@news-2.csn.net> Is there a way to tell which cell was --> not <-- selected from a matrix? Thanks Brendan
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: How to add takeFloatValueFrom to subclass of Control? Date: 15 Mar 1996 20:26:34 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4icjpq$4eq@stc06.ctd.ornl.gov> I am building an object for a palette that should be just like a slider BUT has two degrees of freedom. I need to have a method takeXFloatValueFrom and a takeYFloatValueFrom somehow to allow someone using the palette object to obtain the X and Y values of the object. What do the takeFloatValueFrom methods do in Control and what else do I have to have to make something like this work? -- - - - - - - - - - J. W. Wooten
Newsgroups: comp.sys.next.programmer,comp.mail.uucp From: westes@netcom.com (Will Estes) Subject: Does NS 3.2 FIP have TERMIO or TERMIOS? Message-ID: <westesDoCKu4.Lv3@netcom.com> Organization: U.S. Computer Date: Sat, 16 Mar 1996 06:32:28 GMT Sender: westes@netcom4.netcom.com I am compiling Taylor UUCP 1.06 under NeXTSTEP FIP 3.2. Does NS 3.2 have either SYSV TERMIO or POSIX TERMIOS? Are these as reliable as BSD TTY under NS? -- Will Estes U.S. Computer Internet: westes@usc.com POB 3150 Saratoga, CA 95070 FAX: 408-446-1013
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:16 GMT Organization: Unconfigured Message-ID: <4idomg$9u6@jaring.my> cancel
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:19 GMT Organization: Unconfigured Message-ID: <4idomj$9u6@jaring.my> cancel
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:22 GMT Organization: Unconfigured Message-ID: <4idomm$9u6@jaring.my> cancel
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:25 GMT Organization: Unconfigured Message-ID: <4idomp$9u6@jaring.my> cancel
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:28 GMT Organization: Unconfigured Message-ID: <4idoms$9u6@jaring.my> cancel
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:31 GMT Organization: Unconfigured Message-ID: <4idomv$9u6@jaring.my> cancel
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Control: cancel <4i9r0c$rp9@sun0.urz.uni-heidelberg.de> Date: 16 Mar 1996 06:56:41 GMT Organization: Unconfigured Message-ID: <4idon9$9u6@jaring.my> cancel
Newsgroups: comp.sys.next.programmer,comp.mail.uucp From: westes@netcom.com (Will Estes) Subject: Taylor UUCP 1.06 Build Failing Under NeXTSTEP 3.2 FIP Message-ID: <westesDoCnDw.EqE@netcom.com> Organization: U.S. Computer Date: Sat, 16 Mar 1996 07:27:32 GMT Sender: westes@netcom18.netcom.com I am compiling Taylor UUCP 1.06.1 under NeXTSTEP FIP 3.2. Everything runs fine until we get to linking, and then: rm -f libuucp.a ar rc libuucp.a buffer.o crc.o debug.o escape.o getopt.o getop1.o parse.o spool.o status.o xfree.o xmall.o xreall.o getlin.o strdup.o ranlib libuucp.a cc -g -O -o uucico uucico.o trans.o send.o rec.o xcmd.o prot.o protg.o protf.o prott.o prote.o proti.o protj.o proty.o protz.o time.o log.o chat.o conn.o util.o copy.o unix/libunix.a uuconf/libuuconf.a lib/libuucp.a ld: Undefined symbols: _tcgetattr _cfgetospeed _cfsetospeed _cfsetispeed _tcflush _tcsetattr _tcsendbreak *** Exit 1 Stop. Any ideas on what I am doing wrong here? -- Will Estes U.S. Computer Internet: westes@usc.com POB 3150 Saratoga, CA 95070 FAX: 408-446-1013
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: barcode algorithm Date: 16 Mar 1996 13:52:33 GMT Organization: Nanyang Technological University Message-ID: <4ieh31$ft7@ntuix.ntu.ac.sg> References: <4i6qii$b4a@oznet07.ozemail.com.au> PAUL ARMSTRONG (efr611@shell01.ozemail.com.au) wrote: : any one knows where can i find an algorithms to print bar codes??? : any help would be appreciate. I created (years ago) a Code39 PostScript font. E-mail me when that would help you. WIllem gwillem@ntuvax.ntu.ac.sg
From: mbecker@uml.edu (Mark Becker) Newsgroups: comp.sys.next.programmer,gnu.gcc.help Subject: I/O on long long int under cc/gcc? Date: 16 Mar 1996 14:59:38 GMT Organization: A poorly-installed InterNetNews site Message-ID: <4iel0q$itj@ulowell.uml.edu> Hello * Pardon the cross-post; looked like the best places to ask for help. Was doing some test coding for a project and found cc (NS 2.1 stock) and gcc 2.7.2 both recognize "long long int" as a valid data type with size 8-bytes. This size data type just saved me from having to simulate a 64-bit value. However, in looking through man-pages, I don't see the format letter used in printf()/scanf() (or other conversion) to convert a variable of this size. Q: What is the format-letter for ASCII conversion of this size variable? Thanks - Mark mbecker@caehbl.uml.edu
From: <> Newsgroups: comp.sys.next.programmer Subject: Re: NextStep Developer Wanted - NYC area Date: 16 Mar 1996 16:27:23 GMT Organization: C.I.S.M. Universite de Lyon 1 / INSA de Lyon Message-ID: <4ieq5b$r9r@tempo.univ-lyon1.fr> purdys7@aol.com (Purdys7) s'écrit : > NextStep developer/sys admin wanted: NY metro area. 2-4 month assignment > with opportunity for full-time position. Experience with C, Objective C, > EOF, and QuickBase required. Exciting applications. Reply to us, the > employer, directly at purdys7@aol.com (we are not a recruiter or agency). Hello, I write you in case of ... So i work on Nextstep since 1991. I've written a cad software that you can download at : "ftp://ftp.asi.fr/pub/next" the programm is "Jessycad", try it to evaluate my capability. I live in france, but i'm going in a near time to live in US or canada. I come to U.S in may. Cordialy, Olivier Chapiteau olivchap@net.asi.fr
From: tm8025a@american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer,comp.mail.uucp Subject: Re: Taylor UUCP 1.06 Build Failing Under NeXTSTEP 3.2 FIP Date: 16 Mar 1996 19:08:52 GMT Organization: American University, Washington DC Message-ID: <4if3k4$sq5@paladin.american.edu> References: <westesDoCnDw.EqE@netcom.com> In <westesDoCnDw.EqE@netcom.com> Will Estes wrote: > cc -g -O -o uucico uucico.o trans.o send.o rec.o xcmd.o prot.o protg.o > protf.o prott.o prote.o proti.o protj.o proty.o protz.o time.o > log.o chat.o conn.o util.o copy.o unix/libunix.a uuconf/libuuconf.a > lib/libuucp.a > ld: Undefined symbols: > _tcgetattr > _cfgetospeed > _cfsetospeed > _cfsetispeed > _tcflush > _tcsetattr > _tcsendbreak > *** Exit 1 > Stop. > > Any ideas on what I am doing wrong here? > > You need to include POSIX flags. Those are all POSIX symbols. Try adding -posix to your cc flags. -- Torrey McMahon American University School of Communication System Administrator, WebMaster, and Security Consultant NeXT, MIME, and ASCII PGP and CyptorBundle PGP Key available via finger and home page. Check the headers for both.
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: barcode algorithm In-Reply-To: GWILLEM@alpha.ntu.ac.sg's message of 16 Mar 1996 13:52:33 GMT Message-ID: <RDL.96Mar16143050@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4i6qii$b4a@oznet07.ozemail.com.au> <4ieh31$ft7@ntuix.ntu.ac.sg> Date: Sat, 16 Mar 1996 19:30:49 GMT Also, my company HTI develops and markets the BarCodeKit which includes a wide range of bar code palettes. BarCodeKit was nominated for Best of Breed ObjectWare. Demonstration palettes can be found in every copy of NEXTSTEP 3.3 on the Third Party/ObjectWare CD-ROM. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP /Developer/Career Position Date: 16 Mar 1996 21:41:42 GMT Organization: Norden 1 Communications Message-ID: <4ifcim$ei0@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF--A plus Sybase or Oracle--A plus Career Position Exceptional company Relocation assistance To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: "Ben Bernhard" <bbernhar@cs.indiana.edu> Newsgroups: comp.sys.next.programmer Subject: Smalltalk for my old black box Date: Sat, 16 Mar 1996 19:04:55 -0500 (EST) Organization: Computer Science, Indiana University Message-ID: <24212@827021099> Hi, I'm looking to compile gnu smalltalk on my next maching running 3.2. Has anybody done this? Thanks for any info, Ben -- __________________________________________________________________________ Ben Bernhard "Nothing that results from human progress 205/322-1604 (fax) is achieved with unanimous consent." bbernhar@cs.indiana.edu ---Christopher Columbus
From: miguel@sphinx.nuclecu.unam.mx (Miguel de Icaza) Newsgroups: comp.sys.next.programmer Subject: Binary emulation of NeXTStep. Date: 16 Mar 1996 20:49:55 -0600 Organization: Universidad Nacional Autonoma de Mexico Message-ID: <s83f78l9gs.fsf@sphinx.nuclecu.unam.mx> Hello, I'm looking for technical information on how the WindowServer works and how applications talk to the WindowServer under Next's Mach. This is just to examine the possibility of adding a personality for Linux (at least for the SPARC port, since I have a Sparc running NextStep) that would let us run the NeXTStep environement on top of the Linux kernel. If you know of any tool that would let me trace system calls under NeXTStep I would appreciate that (trace or truss or strace, or any other tool). Emulating Mach under Linux will be easy, I just need the information on what does the WindowServer expect from the operating system: it requires a frame buffer interface to the video display? How are applications supposed to talk to the WindowServer? By means of a mach_message or through a socket connection to the server to send Display Postscript code? Information on the WindowServer mechanics will be greately appreciated. Thanks in advance, Miguel. -- miguel@roxanne.nuclecu.unam.mx The GNU Midnight Commander: http://stekt.oulu.fi/~jtklehto/mc Linux/SPARC project: http://www.geog.ubc.ca/sparclinux.html
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: How to add takeFloatValueFrom to subclass of Control? In-Reply-To: woo@woonext.cmo.ornl.gov's message of 15 Mar 1996 20:26:34 GMT Message-ID: <RDL.96Mar17013147@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4icjpq$4eq@stc06.ctd.ornl.gov> Date: Sun, 17 Mar 1996 06:31:47 GMT Here's an example of what you need: - takeFloatValueFrom:sender { return [self setFloatValue:[sender floatValue]]; } Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4icjpq$4eq@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. Wooten) writes: I am building an object for a palette that should be just like a slider BUT has two degrees of freedom. I need to have a method takeXFloatValueFrom and a takeYFloatValueFrom somehow to allow someone using the palette object to obtain the X and Y values of the object. What do the takeFloatValueFrom methods do in Control and what else do I have to have to make something like this work?
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Debugging bundles In-Reply-To: caljouwr@cadvision.com's message of 15 Mar 1996 14:51:37 GMT Message-ID: <RDL.96Mar17013800@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4ic05p$2t1m@huey.cadvision.com> Date: Sun, 17 Mar 1996 06:38:00 GMT Go into gdb and set a breakpoint on initialize for the Application class (or whatever your app uses) before you run it, run the app, it will break, type [NSBundle stripAfterLoading:0] in gdb and then continue. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4ic05p$2t1m@huey.cadvision.com> caljouwr@cadvision.com writes: Xref: world comp.sys.next.programmer:26795 Path: world!uunet!in2.uu.net!huey.cadvision.com!usenet From: caljouwr@cadvision.com Newsgroups: comp.sys.next.programmer Date: 15 Mar 1996 14:51:37 GMT Organization: CADVision Lines: 23 Distribution: world Reply-To: caljouwr@cadvision.com NNTP-Posting-Host: cadb142.cadvision.com X-Newsreader: RadicalNews (TM) v0.8.1 Beta Does anyone know if there is a way to use a source level debugger when debugging bundles. According to NeXTAsnwers if you make the call [NSBundle stripAfterLoading:NO]; you will be able to use the debugger but this call must be made before any bundles are loaded. I am working on a bundle for a NeXT application so I don't have access to the main calling routines, only my own source for the bundle. If any one has any information it would be greatly appreciated! Debugging using printing and logging of information becomes rather tedious and time consuming. Thanks in advance, Rob Caljouw
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: selectedCell Stuff In-Reply-To: bff@icarus.pvh.org's message of 15 Mar 1996 21:19:10 GMT Message-ID: <RDL.96Mar17014327@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4icmse$klm@news-2.csn.net> Date: Sun, 17 Mar 1996 06:43:27 GMT Using cellList, you can get all the cells in the Matrix and getSelectedCells: gives you all the selected ones. Thus, it's easy to write a method that will give you all the unselected ones. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4icmse$klm@news-2.csn.net> bff@icarus.pvh.org (Brendan Forsyth) writes: Xref: world comp.sys.next.programmer:26799 Path: world!uunet!in1.uu.net!solaris.cc.vt.edu!news.mathworks.com!gatech!swrinde!howland.reston.ans.net!math.ohio-state.edu!magnus.acs.ohio-state.edu!csn!news-1.csn.net!news-2.csn.net!usenet From: bff@icarus.pvh.org (Brendan Forsyth) Newsgroups: comp.sys.next.programmer Date: 15 Mar 1996 21:19:10 GMT Organization: SuperNet Inc. +1.303.296.8202 Denver Colorado Lines: 4 NNTP-Posting-Host: 204.131.215.253 Is there a way to tell which cell was --> not <-- selected from a matrix? Thanks Brendan
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: Joystick drivers for NextStep? Date: 17 Mar 1996 19:31:34 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4ihpam$1ah5@core.bard.edu> I need a joystick driver for nextstep for a projext I am doing right now. Has anyone heard of one? Otherwise I will have to write my own, which I though about briefly. The first question I would have is: is there somekind of standard to follow in assigning event numbers to suchj things as joystick events? -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
Newsgroups: comp.sys.next.programmer,comp.mail.uucp From: rdl@world.std.com (Robert La Ferla) Subject: Re: Taylor UUCP 1.06 Build Failing Under NeXTSTEP 3.2 FIP In-Reply-To: tm8025a@american.edu's message of 16 Mar 1996 19:08:52 GMT Message-ID: <RDL.96Mar17223539@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <westesDoCnDw.EqE@netcom.com> <4if3k4$sq5@paladin.american.edu> Date: Mon, 18 Mar 1996 03:35:39 GMT Actually, don't do that. Instead, edit your policy.h: #define HAVE_BSD_TTY 1 #define HAVE_SYSV_TERMIO 0 #define HAVE_POSIX_TERMIOS 0 Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4if3k4$sq5@paladin.american.edu> tm8025a@american.edu (Torrey McMahon) writes: In <westesDoCnDw.EqE@netcom.com> Will Estes wrote: > cc -g -O -o uucico uucico.o trans.o send.o rec.o xcmd.o prot.o protg.o > protf.o prott.o prote.o proti.o protj.o proty.o protz.o time.o > log.o chat.o conn.o util.o copy.o unix/libunix.a uuconf/libuuconf.a > lib/libuucp.a > ld: Undefined symbols: > _tcgetattr > _cfgetospeed > _cfsetospeed > _cfsetispeed > _tcflush > _tcsetattr > _tcsendbreak > *** Exit 1 > Stop. > > Any ideas on what I am doing wrong here? > > You need to include POSIX flags. Those are all POSIX symbols. Try adding -posix to your cc flags. -- Torrey McMahon American University School of Communication System Administrator, WebMaster, and Security Consultant NeXT, MIME, and ASCII PGP and CyptorBundle PGP Key available via finger and home page. Check the headers for both.
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 18 Mar 1996 05:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4iirgu$bqa@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: jacob@dannug.dk (Jacob Nielsen) Newsgroups: comp.sys.next.programmer Subject: Re: NS encoding <-> Latin 1? Date: 13 Mar 1996 22:15:43 GMT Organization: Danish NeXT User Group Distribution: world Message-ID: <4i7hef$73@jnext.dannug.dk> References: <4hsq8d$uf@turbocat.snafu.de> <4i1sif$15f@precipice.fdn.fr> In <4i1sif$15f@precipice.fdn.fr> Hugues RICHARD wrote: > In article <4hsq8d$uf@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) > writes: > > Hi! > > > > Is there a simple way to convert ISO Latin 1 to NEXTSTEP encoded strings? > > (and the other way) > > You should find that in the Alexandra news reader sources... > See the peanut's or Muenchen archives... Or get (and compile) recode. Availible from your nearest GNU archive. Given the right command line options it will output a C function that performs the translation. Regards, Jacob -- Jacob Nielsen Maintainer of NEXTSTEP Software Reviews http://www.dannug.dk/jacob & My own home page :-) http://www.dannug.dk/~jacob NeXTMail, MIMEMail and SUNMail jacob@dannug.dk
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9603162151.AA19619@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Sat, 16 Mar 96 15:51:42 -0600 Subject: Scanning Hello, all... I was wondering what experience anyone had with scanner control under NS...I'm not looking for color desktop so much as for control of some brand of grayscale batch-oriented multisheet scanner. Regards, - Dan
From: james.robertson@wcom.com Newsgroups: comp.sys.next.programmer Subject: Help. WebObjects. Creating new components Date: 18 Mar 1996 17:47:33 GMT Organization: WilTel Message-ID: <4ik7jl$mkv@gateway.wiltel.com> I'm having trouble creating simple Web Objects application. I've installed Web Objects on a NeXTSTEP 3.3 platform. The example apps (e.g. HelloWorld, Visitors, etc.) work OK. I copied the HelloWorld app into my own directory under the doc root, and that works OK. I can make trivial changes to the html in the HelloWorld app, and it still works OK. But I can't add any new pages/components. Even changing the name of the existing "Hello" component from "Hello" to "Goodbye" produces an error. If I change the names of directories/files from Hello.wo, Hello.html, Hello.wod, Hello.wos to Goodbye.wo, Goodbye.html, Goodbye.wod, Goodbye.wos and change the page-creation message in Main.wos to nextPage = [WOApp pageWithName:@"Goodbye"]; the web browser can't bring up the "Goodbye" page. It reports... Exception occurred while handling request: Application Name = 'JimsExamples/HelloWorld' Exception Name = 'NSGenericException' Exception Reason = 'Method or procedure: 'sayHello' Evaluation error in script Erroneous statement: 'nextPage = [WOApp pageWithName: @"Goodbye"]' Exception while evaluating Objective-C expression:[WOApp pageWithName:@"Goodbye"] Cannot find HTML template for component named Goodbye. Component may not exist. ' I couldn't find anywhere in the docs that describes how to make Web Objects aware of the components in an application, beyond placing .html, .wod and .wos files in a .wo subdirectory. Any suggestions or pointers to docs would be appreciated. Thanks. Jim Robertson LDDS WorldCom jim.robertson@wcom.com (713) 364-4133 voice (713) 367-9036 fax
Newsgroups: comp.sys.next.programmer From: zollingj@ripley (John Zollinger) Subject: Re: barcode algorithm Sender: news@docws001.shl.com Message-ID: <DoH0H6.C3M@docws001.shl.com> Date: Mon, 18 Mar 1996 16:00:42 GMT References: <4i6qii$b4a@oznet07.ozemail.com.au> <4ieh31$ft7@ntuix.ntu.ac.sg> <RDL.96Mar16143050@world.std.com> Organization: SHL Systemhouse Inc. Robert La Ferla (rdl@world.std.com) wrote: : Also, my company HTI develops and markets the BarCodeKit which includes : a wide range of bar code palettes. BarCodeKit was nominated for Best of : Breed ObjectWare. Demonstration palettes can be found in every copy : of NEXTSTEP 3.3 on the Third Party/ObjectWare CD-ROM. And an excellent kit it is. A must for anyone who needs to do barcodes. If you have purchased a pre-paid calling card from AT&T (and some others) you have probably seen them in action. Good thing Robert doesn't charge per barcode generated. :-) Just a happy user... John Zollinger ProActive Consulting, Inc. johnz@netropolis.net
From: paulrs@lgs-systems.buf.servtech.com (Paul R. Summermatter) Newsgroups: comp.sys.next.programmer Subject: piping information... Date: 18 Mar 1996 19:55:34 GMT Organization: LGS Systems Message-ID: <paulrs-1803961551310001@lgs-systems.buf.servtech.com> Hello all, I'm working on a few projects that require for me to use the terminal... My question is this... How can I make a link up with my app the GUI and the Unix back-end? Are there any examples of this anywhere? Any responces sould go to: peter@lgs-systems.com Thank you for the help. Peter J
From: james.robertson@wcom.com Newsgroups: comp.sys.next.programmer Subject: Re: Help. WebObjects. Creating new components Date: 18 Mar 1996 21:17:41 GMT Organization: WilTel Message-ID: <4ikjtm$ncp@gateway.wiltel.com> References: <4ik7jl$mkv@gateway.wiltel.com> Followup to my own posting: I've made some progress since posting below question regarding adding components to a WebObjects application. I've found that if I add a new component (i.e. .wo directory, and the .html, .wod, and .wos files contained therein) to an application, it breaks. I can't seem to incrementally add pages to an application. But if I copy my entire application, with the additional component(s), to a new directory, then it works fine. This workaround will at least keep me going for a while, though I'm going to accumulate a lot of applications. Does anybody know how you're really supposed to incrementally develop a WebObjects application? - Jim Robertson Jim Robertson LDDS WorldCom jim.robertson@wcom.com (713) 364-4133 voice (713) 367-9036 fax james.robertson@wcom.com wrote: >I'm having trouble creating simple Web Objects application. >I've installed Web Objects on a NeXTSTEP 3.3 platform. >The example apps (e.g. HelloWorld, Visitors, etc.) work OK. >I copied the HelloWorld app into my own directory under the >doc root, and that works OK. >I can make trivial changes to the html in the >HelloWorld app, and it still works OK. >But I can't add any new pages/components. Even changing the name >of the existing "Hello" component from "Hello" to "Goodbye" >produces an error. If I change the names of directories/files from > Hello.wo, Hello.html, Hello.wod, Hello.wos >to > Goodbye.wo, Goodbye.html, Goodbye.wod, Goodbye.wos >and change the page-creation message in Main.wos to > nextPage = [WOApp pageWithName:@"Goodbye"]; >the web browser can't bring up the "Goodbye" page. It reports... >Exception occurred while handling request: >Application Name = 'JimsExamples/HelloWorld' >Exception Name = 'NSGenericException' >Exception Reason = 'Method or procedure: 'sayHello' >Evaluation error in script >Erroneous statement: 'nextPage = [WOApp pageWithName: >@"Goodbye"]' >Exception while evaluating Objective-C expression:[WOApp >pageWithName:@"Goodbye"] >Cannot find HTML template for component named Goodbye. >Component may not exist. >' >I couldn't find anywhere in the docs that describes how to make >Web Objects aware of the components in an application, beyond placing >.html, .wod and .wos files in a .wo subdirectory. >Any suggestions or pointers to docs would be appreciated. Thanks. >Jim Robertson
From: vbragin@uranus.calstatela.edu (Victoria Bragin) Newsgroups: comp.sys.next.programmer Subject: View on laptop monitor deffers from that on desktop monitor - why? Date: 19 Mar 1996 03:18:17 GMT Organization: Information Resources and Technology Message-ID: <4il91p$8dt@nuke.csu.net> Keywords: I have developed an application which my students run on Intel machines. Why is it that when I run the application on a laptop, what I see on the screen is not automatically resized to fill the laptop screen. That is, I am able to view only the top 3/4 of what I would ordinarily see when I run the application on the desktop Intels. The problem is I cannot even resize a window as I cannot even get to the resize bar at the bottom - it is completely hidden from view. Thanks. Vicki Bragin
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Re: Help. WebObjects. Creating new components Date: 19 Mar 1996 04:05:19 GMT Organization: Digital Fix Development Message-ID: <4ilbpv$itk@digifix.digifix.com> References: <4ik7jl$mkv@gateway.wiltel.com> <4ikjtm$ncp@gateway.wiltel.com> In-Reply-To: <4ikjtm$ncp@gateway.wiltel.com> I responded to Jim privately the first time, but the information might be useful. Remember that there is a great WebObjects development mailing list run by Omnigroup.. To subscribe to the WebObjects mailing list, send mail to _listproc@omnigroup.com_ saying (in the body of the message): subscribe WebObjects Your Name To unsubscribe, send mail to _listproc@omnigroup.com_ saying: unsubscribe WebObjects If this automated procedure fails, send mail to _netsec@omnigroup.com_ and a human will help you. -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: Adding GNU 2.6.2.diff to cc on Next Date: 19 Mar 1996 03:32:52 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4il9t4$f50@core.bard.edu> I need gcc 2.6.2 (i beleive) to compile gcl-2.2 and I was wondering whether I could use the gcc .diff files, which are the difference between the old and the new versions, to add on NeXT's cc, since it is based on gcc. And by the way, since I think gcc supports obj-c, why doesn't next work with GNU and produce one product? -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: Debugging bundles Date: 16 Mar 1996 03:17:10 GMT Organization: Genoa Software Systems Distribution: world Message-ID: <4idbrm$4kd@saturn.genoa.com> References: <4ic05p$2t1m@huey.cadvision.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <4ic05p$2t1m@huey.cadvision.com> caljouwr@cadvision.com wrote: > if you make the call [NSBundle stripAfterLoading:NO]; > you will be able to use the debugger but this call must be made before > any bundles are loaded. ... > I don't have access to the main calling routines you can still set a breakpoint on main and send a message to NSBundle using the p command not having symbols, just means you can't view the stack conveniently. You can still break, and examine the stack for instance, it just is a little more work. Julie Z's Zen of Debugging talk shows some how. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <maurice@tlpnl.tlp.nl> From: Maurice Le Rutte <maurice@tlpnl.tlp.nl> Message-ID: <9603191312.AA01019@tlpnl.tlp.nl> Date: Tue, 19 Mar 96 14:12:08 +0100 Subject: RTF documentation needed. For a project I'm doing I need to generate Rich Text, without using the Appkit (so no Text class). I am looking for the definition of the Rich Text Format. I vaguely recall having read some documentation about RTF, but I can't find it nowhere. We do not have the documentation avaliable here and ordering books will take too long. Who can mail me some documents about RTF? As I only have mail access I cannot browse the web. :-( Thanks in advance, Maurice le Rutte _____________________________________________________________ | | |OpenStep Software Engineering | |NeXT/MIME/ASCII mail welcome Suo sibi gladio hunc iugulo* | |____________________________________________________________| * I will beat him with his own weapons.
From: zmahmood@site.gmu.edu (Zakir Mahmood (ELEN)) Newsgroups: comp.sys.next.programmer Subject: AVATAR help, please Date: 19 Mar 1996 15:01:50 GMT Organization: George Mason University, Fairfax, Virginia, USA Message-ID: <4imi8u$sgn@portal.gmu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Hello: I was wondering if some out there has used the product called AVATAR. It is by a company called International Data Systems. I is supposedly "automated software that can convert/re-engineer application code from any source to any new target environment." I would like to know if anyone has used this product and what they think of it. I would like to know what types of files it can convert and with what ease.Does it work with smalltalk, C++,cobol,vm,mvs? what type of operating system(s) does it run under? I would like to know what you think of it. Any help in this regard would be greatly appreciated. If you can share some information, please e-mail me at: zmahmood@gmu.edu or at mahmood@usairln.usair.com Thanks Zakir
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: US - Chicago - NeXT Developers - Financial Systems - $50 to $80k+ Date: 19 Mar 1996 11:50:53 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4imold$pra@newsbf02.news.aol.com> My client, a major North American Corporation, has opened a new office in Chicago, and is currently searching for NeXT Developers. You will be responsible for developing NeXTSTEP based financial applications in Objective-C and Sybase. Candidates will receive an excellent starting salary + Bonus. My client will sponsor for H1. Contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: US - TX - NeXt Developers - Career Opprtunities - to $75k Date: 19 Mar 1996 12:00:49 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4imp81$q1t@newsbf02.news.aol.com> My client a major Fortune 50 company is currently establishing a new development group that will build systems using NeXT technology. Candidates must have a proven track record in NeXT development using Objective-C and Sybase/Oracle. Excellent salary or daily rates. Contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
From: lisap23@aol.com (LisaP23) Newsgroups: comp.sys.next.programmer Subject: Cap Gemini America Has Immediate Opportunities for NeXTSTEP PA Date: 19 Mar 1996 12:05:52 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4imphg$q4s@newsbf02.news.aol.com> We are Cap Gemini America, a transnational computer consulting company with over 22,000 consultants globally. Our company focus is to resolve our clients' information technology issues. Presently, we have immediate opportunities for NeXTSTEP application developers in Florida. If you possess at least two years of NeXT development experience and would like to discover the options for your career with a progressive consulting firm, please contact Lisa Peterson at lisap23@aol.com or (305) 776-0621. We offer excellent salaries, benefits, a career enhancement program and the opportunity for you to develop a stellar career!
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer,comp.mail.uucp Subject: Re: Does NS 3.2 FIP have TERMIO or TERMIOS? Date: 19 Mar 1996 17:05:38 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4imph2$822@news.its.com> References: <westesDoCKu4.Lv3@netcom.com> westes@netcom.com (Will Estes) wrote: > I am compiling Taylor UUCP 1.06 under NeXTSTEP FIP 3.2. Does NS 3.2 have > either SYSV TERMIO or POSIX TERMIOS? Compile with "cc -posix". > Are these as reliable as BSD TTY under NS? Not even close. Unfortunately, NeXT's POSIX support contains some severe bugs-- the most likely one to bite you is the "file append bug", where files get overwritten by 0x0. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9603191648.AA05998@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Tue, 19 Mar 96 10:48:07 -0600 Subject: Variation in Behavior of Initialized NXImages Hello, all (again...:-) I dinked together a little app to load a designated TIFF or JPEG image into a view. I had never really worked with images much, and I wanted to try to implement a drag-and-drop. I cobbled together some of the code from the G&M book with an example out of Backlin's chapters on image management and drag-and-drop. Things went along swimmingly until I decided I wanted to scale the image at 200% along both axes, preparatory to implementing a 'zoom' feature. I started getting inconsistent behavior between those images loaded from a dialog and those dropped into the view. To back up a little bit, when the app initializes (and when the "load another image" menu item is selected), this method chain runs for the ImageView object, to get the first image: - getANewImage: sender { id myOpenPanel = [ OpenPanel new ]; if ( [ myOpenPanel runModalForTypes: [NXImage imageFileTypes ] ] ) { const char *theFileName = [ myOpenPanel filename ]; [ self setImage: [[ NXImage alloc ] initFromFile: theFileName ] ]; // See below [ window setTitle: theFileName ]; } return self; } - setImage: newImage { NXSize theSize; if (displayImage ) [ displayImage free ]; displayImage = newImage; [ displayImage setScalable: YES ]; [[ displayImage bestRepresentation ] getSize: &theSize ]; theSize.width *= 2.0; theSize.height *= 2.0; [[ displayImage bestRepresentation ] setSize: &theSize ]; // Look here... [ self setSize: &theSize ]; // see below [ self display ]; [[ self window ] makeKeyAndOrderFront: self ]; return self; } - setSize: (NXSize *)aSize { NXRect obounds = bounds; [ self convertRectToSuperview: &obounds ]; [ self sizeTo: (aSize->width) : (aSize->height) ]; [ self convertRectFromSuperview: &obounds ]; [ self displayFromOpaqueAncestor: NXUnionRect( &obounds, &bounds ) : 1 : NO ]; return self; } When an image is dropped into the view, the same chain gets kicked off from here: - (BOOL) performDragOperation: sender { int length; int i; Pasteboard *pboard; char *data = '\0'; id imageRepClass = nil; pboard = [ sender draggingPasteboard ]; [ pboard readType: NXFilenamePboardType data: &data length: &length ]; for ( i = strlen( data ); data[i] != '.'; i-- ) ; // I know there's a stringlib call; I just can't remember it... imageRepClass = [ NXImage imageRepForFileType: (const char *)&data[ ++i ] ]; fprintf( stderr, "Asked to accept %s...", data ); if (imageRepClass == nil ) { fprintf( stderr, "declined.\n" ); return NO; } fprintf( stderr, "Accepted.\n" ); [ self setImage: [[ NXImage alloc ] initFromPasteboard: pboard ] ]; // Different init... [ window setTitle: data ]; [ pboard deallocatePasteboardData: data length: length ]; return YES; } Now, things work as anticipated until a window resize event comes along: - windowDidResize: sender { NXSize theSize; [[ displayImage bestRepresentation ] getSize: &theSize ]; fprintf( stderr, "The WindowDidResize function found an imageRep size of %7.2fw by %7.2fh\n", theSize.width, theSize.height ); [ self setSize: &theSize ]; [ self display ]; return self; } Up in the setImage: method, you'll note that I doubled the size for the image's best representation...that operation "takes" for images initialized from the pasteboard, but it seems to be temporary for images initialized from a file! So, the first time I resize the window for an image loaded by dialog, I lose 50% of the width and 50% of the height (actually, the remaining part of the image is properly scaled to twice the original size; it's just that the top and right are clipped away). If I drop an image into the view via the pasteboard service, the image representation resize operation 'sticks' and the double-sized image responds properly to all window resizing. The fprintf() statement in windowDidResize: shows the original sizing for dialog-loaded images, and a properly doubled size for dropped-in images. Any takers? Thanks & regards (as always), - Dan
From: dekorte@intrepid.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: column titles of DBTableView Date: 19 Mar 1996 21:34:44 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4in99k$41l@news.onramp.net> Anyone have a code snippet that gets at them? It would be nice if DBTableView had a method like: - columnTitleAt:(int)column. Thanks for any info, Steve
From: yibwu@napoleon.plato.sky.bdm.com(Yibing Wu) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Any news group for OpenStep? Date: 19 Mar 1996 23:51:09 GMT Organization: BDM International, Inc. Message-ID: <4inh9d$qem@news.mcl.bdm.com> Is there a news group (or groups) for OpenStep only? e.g. the compatibility issues, the conversions and, of course, bugs. An urgent question is, does OpenStep on Mach work with SQL+Net V2 from Oracle? I know it works with V1 and NeXTSTEP 3.3 works with SQL+Net V2. However, after I created the /etc/tnsnames.ora file, modified the .eomodels, the EOModeler tells me: "Cannot connect to server!". SIgh, I am not expecting any response... I found NeXT community less responsive than many other groups. Thanks Bing
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 19 Mar 1996 23:11:54 GMT Organization: Norden 1 Communications Message-ID: <4inevq$2tu@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF--a plus Sybase or Oracle--a plus Career Position Exceptional Company Full Benefits Relocation assistance To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: Mehdi Bousaidi <mbousaidi@cen.com> Newsgroups: comp.sys.next.programmer Subject: JDK by NeXT, Date: Tue, 19 Mar 1996 15:44:40 -0500 Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <314F1CB8.6415@cen.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Does any of you know whether NeXT is planning on building a Java Developer Kit . ? Imagine, an IB for Java !! Thanks. -- ------------------------------------------------------------------- Mehdi Bousaidi | ____/* | email: mbousaidi@cen.com Century Computing | |___| | web: http://www.cen.com 8101 Sandy Spring | (~ ~) | phone: (301) 953-3330 ext. 129 Laurel, MD 20707 | @ V @ | Fax: (301) 953-2368 | \W/ | -------------------------------------------------------------------
Newsgroups: comp.sys.next.programmer From: wave@ml.media.mit.edu (Michael B. Johnson) Subject: compiling the latest gcc Message-ID: <1996Mar19.072412.406@media.mit.edu> Sender: news@media.mit.edu (USENET News System) Organization: Pixar Animation Studios (East Coast Office) Date: Tue, 19 Mar 1996 07:24:12 GMT Hi folks. Well, for the last several months, I've been moving more and more towards the dark side, writing and compiling C++ code. I'm interested in compiling and installing the latest gcc/libg++ for NEXTSTEP 3.3, and was wondering if others had any words of wisdom. I'll be happy to post back if I get any good tidbits... -- --> Michael B. Johnson, SMVS, Ph.D. -- wave@media.mit.edu|wave@pixar.com --> http://wave.www.media.mit.edu/people/wave/ --> alumnus, MIT Media Lab, Computer Graphics & Animation Group --> Media Arts Technologist, Pixar Animation Studios (East Coast Office)
From: reichman@scf.usc.edu (Matthew N. Reichman) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: Any news group for OpenStep? Date: 20 Mar 1996 06:10:24 GMT Organization: University of Southern California, Los Angeles, CA Sender: reichman@comserv-b-43.usc.edu Message-ID: <4io7gg$cmj@usc.edu> References: <4inh9d$qem@news.mcl.bdm.com> In <4inh9d$qem@news.mcl.bdm.com> Yibing Wu wrote: > Is there a news group (or groups) for OpenStep only? e.g. the compatibility > issues, the conversions and, of course, bugs. > > An urgent question is, does OpenStep on Mach work with SQL+Net V2 from > Oracle? I know it works with V1 and NeXTSTEP 3.3 works with SQL+Net V2. > However, after I created the /etc/tnsnames.ora file, modified the > .eomodels, the EOModeler tells me: "Cannot connect to server!". > > SIgh, I am not expecting any response... I found NeXT community less > responsive than many other groups. I certainly don't know the answer to your question. But that last comment I can say from experience is patently untrue! -- Be well, Matthew Reichman reichman@scf.usc.edu USC-CNTV NeXTStep v.3.3 m68k NeXTMAIL & MIME welcome =============================================================== PGP key --> email w/ subject "request_PGP" --------------------------------------------------------------- Computer Privacy Information --> http://www.eskimo.com/~joelm/
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: View on laptop monitor deffers from that on desktop monitor - why? Date: Wed, 20 Mar 1996 08:16:12 +0000 Organization: Cedar Systems Distribution: world Message-ID: <l3vXrDAM77TxEwya@cedar.co.uk> References: <4il91p$8dt@nuke.csu.net> MIME-Version: 1.0 The desktop configuration is not only saved for each user, but also for each screen resolution for each user. This enables you to have less frequently used applications on the dock for larger screens and your most frequently used ones for smaller screens. In addition, users will probably require different window sizes and positioning when using larger and smaller screen resolutions. Paul.
From: lchamp@mines1.info.emn.fr (Champciaux Laurent) Newsgroups: comp.sys.next.programmer Subject: 3D Kit: How to use RiObjectInstance() ? Date: 20 Mar 1996 13:20:40 GMT Organization: Ecole des Mines de Nantes Message-ID: <4ip0n8$1en@wfn.emn.fr> Hello, I am new to the 3D kit and I have a problem with renderman object instancing: In ri.h, RiObjectBegin(), RiObjectEnd() and RiObjectInstance() use RtToken in place of RtObjectHandle (cf. The Renderman Companion). Is there something that I missed ? The object I want to instance is a simple cube (6 polygons) and my program crashes when using the statements above. Here is my code, a method that is called in the renderSelf method of my N3DShape subclass: - drawTheire { int x, y, z; RtToken cube; RiAttributeBegin(); cube = RiObjectBegin(cube); [self unitCube]; // 6 polygons, nothing else RiObjectEnd(); for (x = boundingBox[0]; x <= boundingBox[1]; x=x+inc) for (y = boundingBox[2]; y <= boundingBox[3]; y=y+inc) for (z = boundingBox[4]; z <= boundingBox[5]; z=z+inc) { if (mat[x+25][y+25][z+25]) { RiTransformBegin(); RiTranslate(x, y, z); RiObjectInstance(cube); RiTransformEnd(); } } RiAttributeEnd(); return self; } All is Ok when I don t instance the cube and so use [self unitCube] in place of RiObjectInstance(cube). I have NeXTStep 3.0 on a black cube. Any clue ? -- ======================================================= Laurent Champciaux Departement Informatique - Ecole des Mines de Nantes 4, rue A.Kastler 44070 Nantes Cedex 03 Tel: (33)51 85 82 18 (B220) email: Laurent.Champciaux@emn.fr WWW: http://www.emn.fr/dept_info/perso/laurent/
From: tuparev@asterix.geog.uni-heidelberg.de (Georg Tuparev) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: SPARCprinter Date: 20 Mar 1996 15:26:35 GMT Organization: University of Heidelberg, Germany Message-ID: <4ip83b$hcd@sun0.urz.uni-heidelberg.de> Hi, I have here NS3.3Us/Dev running on SPARC10. Now I installed a SPARCprinter (connected to the serial port) ... but how can I access it? It does not applier in the list of printers, and the serial port cannot be chosen from the PrintManager? Anybody has an idea? Thanks -- georg --
From: paulrs@lgs-systems.buf.servtech.com (Paul R. Summermatter) Newsgroups: comp.sys.next.programmer Subject: checking passwords... Date: 20 Mar 1996 18:07:53 GMT Organization: LGS Systems Message-ID: <paulrs-2003961403490001@lgs-systems.buf.servtech.com> Hello all, If I wanted to ask a user to verify that they are on the system at that point in time by asking them to pype in their own login name and password... how would the app check if that was true? Are their any examples? Thanks for the help.... Peter J peter@lgs-systems.com
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Adding a NXImageView to FlatFile Example Date: 20 Mar 1996 19:38:08 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4ipmr0$ktd@stc06.ctd.ornl.gov> I'm examining FlatFileDataSource Example and wanted to experiment with adding an NXImageView so that when an item in Product was selected, an image of it would be shown. I put a new entry in the eomodel file for each product item called image and as its value put the address of a file containing an image. How do I get that image to appear in the NXImageView? I connected the image entry in the eomodel to the NXImageView using IB, but nothing happens. Can someone point me to an example of using a NXImageView? - - - - - - - - - J. W. Wooten
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: dbadaptor for mSQL? Date: 20 Mar 1996 19:38:52 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4ipmsc$ktq@stc06.ctd.ornl.gov> I've heard that there is a dbadaptor for mSQL available, but have been unable to locate it. Any clues? -- - - - - - - - - - J. W. Wooten
From: mpaque@pbinet.com (Mike Paquette) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: Any news group for OpenStep? Date: Wed, 20 Mar 1996 20:48:16 GMT Organization: Pacific Bell Internet Services Message-ID: <4ipr3s$7hv@SNFC21_SRVR_WWW.PBI.net> References: <4inh9d$qem@news.mcl.bdm.com> yibwu@napoleon.plato.sky.bdm.com(Yibing Wu) wrote: >Is there a news group (or groups) for OpenStep only? e.g. the compatibility >issues, the conversions and, of course, bugs. I suggest looking at comp.sys.next.software and comp.sys.next.programmer for user and programmer discussions, respectively. Note that the OPENSTEP implementations currently out are beta releases, and those who have them are under nondisclosure agreements, as (suprise!) the beta releases aren't quite the same as the anticipated final releases. There's not much discusson activity yet. Once the final releases of OPENSTEP for Mach, Windows, and Solaris are out, I expect to see plenty of discussion on these groups. >An urgent question is, does OpenStep on Mach work with SQL+Net V2 from >Oracle? I know it works with V1 and NeXTSTEP 3.3 works with SQL+Net V2. >However, after I created the /etc/tnsnames.ora file, modified the .eomodels, >the EOModeler tells me: "Cannot connect to server!". Don't know about EOF/Database stuff. Perhaps one of the EOF gurus will share that information with us. >SIgh, I am not expecting any response... I found NeXT community less >responsive than many other groups. Some folks get tired of being flamed, mail-bombed, and generally abused when they try to answer questions. (Some of the USENET equivalent of stalkers lurk in comp.sys.next.*) Me, I just beef up the mail filters and killfiles and go on my way. Also, don't forget that USENET connectivity and reliability isn't what it used to be. About 20% of articles appearing on major East Coast (USA) sites never make it to many sites where I am (San Francisco Bay Area). Sometimes I'd swear that backbone sites are expiring based on posting dates, or giving priority to alt.binaries.nekkid..swedish.chefs.bork.bork.bork... Mike Paquette I don't speak for my employer, and they don't speak for me. mpaque@pbinet.com Personal E-mail mpaque@next.com NeXT business mail only, please
Newsgroups: comp.sys.next.programmer From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: checking passwords... Sender: news@novice.uwaterloo.ca (Mr. News) Message-ID: <DoL389.nsM@novice.uwaterloo.ca> Date: Wed, 20 Mar 1996 20:50:33 GMT References: <paulrs-2003961403490001@lgs-systems.buf.servtech.com> Organization: University of Waterloo In article <paulrs-2003961403490001@lgs-systems.buf.servtech.com>, Paul R. Summermatter <paulrs@lgs-systems.buf.servtech.com> wrote: >Hello all, > >If I wanted to ask a user to verify that they are on the system at that >point in time by asking them to pype in their own login name and >password... how would the app check if that was true? Are their any >examples? > Check out the NetInofKit. From GeneralRef/11_NetInfoKit/IntroNetInfoKit.rtf, Panels The NILoginPanel class provides a mechanism for authenticating a user for access to a NetInfo domain. The panel includes text fields for the user account name and password along with Login and Cancel buttons. This panel can be used to allow the user to authenticate as a user with permission to run an application, as with UserManager and HostManager, or to authenticate as a user with permission to modify a domain, as with NetInfoManager and NFSManager. Hope this helps. -- David Evans (NeXTMail OK) dfevans@bbcr.uwaterloo.ca Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/ University of Waterloo "Default is the value selected by the composer Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: proux@enstb.enst-bretagne.fr (PROUX Guillaume) Newsgroups: comp.sys.next.programmer Subject: Re: dbadaptor for mSQL? Date: 21 Mar 1996 08:40:03 GMT Organization: ENSTBR, Brest, France Distribution: world Message-ID: <4ir4l3$hnc@melimelo.enst-bretagne.fr> References: <4ipmsc$ktq@stc06.ctd.ornl.gov> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit It should be easy to retrieve it... It is called MsqlAdaptor Why not try to search in the msql home page (i don't have the time myself) But I already compiled it in a quad fat manner.. After a few hours of fight against the compiler :) Guillaume.
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: Any news group for OpenStep? Date: 21 Mar 1996 04:12:35 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4iqkvj$8ii@usenet.rpi.edu> References: <4inh9d$qem@news.mcl.bdm.com> <4ipr3s$7hv@SNFC21_SRVR_WWW.PBI.net> mpaque@pbinet.com (Mike Paquette) wrote: > Once the final releases of OPENSTEP for Mach, Windows, and Solaris > are out, I expect to see plenty of discussion on these groups. Will there be a new set of Addison-Wesley books? > > SIgh, I am not expecting any response... I found NeXT community > > less responsive than many other groups. > > Also, don't forget that USENET connectivity and reliability isn't > what it used to be. About 20% of articles appearing on major > East Coast (USA) sites never make it to many sites where I am > (San Francisco Bay Area). Sometimes I'd swear that backbone > sites are expiring based on posting dates, ... This is definitely true here at RPI. One of the reasons I'm on usenet so much is that RPI expires articles based on the date posted. Of course, we don't *get* them on the date posted, and I *know* that there are articles which only exist on our server for a few hours. It's also clear that I'm not seeing all the articles posted, because I see replies to things that I never see the original to (and it's not likely that the original was expired). I've got to check into alternate sources for reading news sometime. Of course, it's long been my contention that all of usenet was bound to collapse at some point. I think we're nearing that point. There are many newsgroups I'm interested in which I simply can't follow anymore, because there isn't enough time to do a good job at it. Of course, I wouldn't mind if the NeXT newsgroups were a bit busier... --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9603211408.AA14826@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Thu, 21 Mar 96 17:08:57 +0300 Subject: Re: Help. WebObjects. Creating new components Hi, In article <4ik7jl$mkv@gateway.wiltel.com>, Jim Robertson wrote about his troubles with his 1st WO application: > But I can't add any new pages/components. Even changing the name > of the existing "Hello" component from "Hello" to "Goodbye" > produces an error. If I change the names of directories/files from > Hello.wo, Hello.html, Hello.wod, Hello.wos > to > Goodbye.wo, Goodbye.html, Goodbye.wod, Goodbye.wos > and change the page-creation message in Main.wos to > nextPage = [WOApp pageWithName:@"Goodbye"]; > the web browser can't bring up the "Goodbye" page. It reports... > > Exception occurred while handling request: > Application Name = 'JimsExamples/HelloWorld' > Exception Name = 'NSGenericException' > Exception Reason = 'Method or procedure: 'sayHello' > Evaluation error in script > Erroneous statement: 'nextPage = [WOApp pageWithName: > @"Goodbye"]' > Exception while evaluating Objective-C expression:[WOApp > pageWithName:@"Goodbye"] > Cannot find HTML template for component named Goodbye. > Component may not exist. > ' You should restart your WebObjects application (i.e. kill it) after such changes. Flushing cache from time to time won't hurt either. Hope that will help. Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9603211417.AA14834@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Thu, 21 Mar 96 17:17:10 +0300 Subject: Re: JDK by NeXT, In article <314F1CB8.6415@cen.com>, Mehdi Bousaidi wrote: > Does any of you know whether NeXT is planning on building a Java > Developer Kit . ? Imagine, an IB for Java !! There already is an IB palette for Java written by Ralf Suckow. It's called aCupOfCoffee and available on all major next related ftp. As for JDK you might want to have a look at BlackJava and set of patches to original Sun's environment a.k.a. NEXTSTEP Java Port. You could find these ports at thoughtport. Here is the URL ftp://ftp.thoughtport.com/pub/next/Java I don't know how far they are differ for I haven't look into BlackJava (to unpack it one should license Java from Sun and get crypt key) and while building the other port I've got the several errors. Could somebody tell me how could I patch original Sun's Java code? Regards. Aleksey. _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Date: 21 Mar 1996 15:00:15 GMT From: zmahmood@site.gmu.edu (Zakir Mahmood (ELEN)) Newsgroups: comp.sys.next.programmer Message-ID: <cancel.4imi8u$sgn@portal.gmu.edu> Control: cancel <4imi8u$sgn@portal.gmu.edu> Subject: cmsg cancel <4imi8u$sgn@portal.gmu.edu> EMP/ECP (aka SPAM) cancelled by jem@xpat.com. See news.admin.net-abuse.announce, report 19960321.02 for further details
From: silvio@jabberwock.swarthmore.edu (Silvio P Eberhardt) Newsgroups: comp.sys.next.programmer Subject: Wanted: parallel port driver Date: 21 Mar 1996 15:03:47 GMT Organization: Swarthmore College, Swarthmore, PA, USA Message-ID: <4irr4j$re8@larch.cc.swarthmore.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Hi- I would like to connect a response box for psychophysical experiments to the parallel port of a NS intel system (I do not need the port for my printer). Has anyone out there used the parallel port bidirectionally? For that matter, how does one use it unidirctionally? Since I only need two input and two output lines, is it possible to use the error and on-line printer bits as two data bits? I can't even find documentation on ioctl'ing the parallel port. Thanks- Silvio Eberhardt Department of Engineering Swarthmore College
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: RTF documentation needed. Date: 21 Mar 1996 18:22:58 GMT Organization: ALI Message-ID: <4is6q2$ksr@cetus.ali.bc.ca> References: <9603191312.AA01019@tlpnl.tlp.nl> In comp.sys.next.programmer Maurice Le Rutte <maurice@tlpnl.tlp.nl> wrote: > For a project I'm doing I need to generate Rich Text, without > using the Appkit (so no Text class). I am looking for the definition > of the Rich Text Format. I vaguely recall having read some > documentation about RTF, but I can't find it nowhere. We do not > have the documentation avaliable here and ordering books will > take too long. > > Who can mail me some documents about RTF? As I only have mail > access I cannot browse the web. :-( Those with web access can find some RTF docs and tools at: http://night.primate.wisc.edu/software/RTF/ I'll contact Maurice via e-mail about sending him the spec. -- Allan Noordvyk, Software Artisan e-mail: allan@ali.bc.ca ALI Technologies Voice: 604.279.5422 x 317 Richmond, Canada Fax: 604.279.5468 * NeXT and MIME mail welcome *
From: jacob@dannug.dk (Jacob Nielsen) Newsgroups: comp.sys.next.programmer,comp.mail.uucp Subject: Re: Taylor UUCP 1.06 Build Failing Under NeXTSTEP 3.2 FIP Date: 19 Mar 1996 21:03:37 GMT Organization: Danish NeXT User Group Message-ID: <4in7f9$10o@jnext.dannug.dk> References: <westesDoCnDw.EqE@netcom.com> <4if3k4$sq5@paladin.american.edu> In <4if3k4$sq5@paladin.american.edu> Torrey McMahon wrote: > In <westesDoCnDw.EqE@netcom.com> Will Estes wrote: > > [Why the build of Taylor UUCP 1.06 fails] > > > You need to include POSIX flags. Those are all POSIX symbols. Try adding > -posix to your cc flags. > Since I've built Taylor 1.06.1 successfully using NEXTSTEP Developer 3.2, I can say that in _this_ case the easiest route is just to do a: #define HAVE_BSD_TTY 1 in 'policy.h'. This avoids the posix functions alltogether. I can send you my notes on how to compile Taylor. You could also get precompiled versions for NeXTs (m68k) and Intel from 'ftp://ftp.dannug.dk/' in some subdirectory I don't remember (most likely pub/next/). Regards, Jacob -- Jacob Nielsen Maintainer of NEXTSTEP Software Reviews http://www.dannug.dk/jacob & My own home page :-) http://www.dannug.dk/~jacob NeXTMail, MIMEMail and SUNMail jacob@dannug.dk
From: tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: dbadaptor for mSQL? Date: 22 Mar 1996 12:42:42 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4iu782$4l3@lion.embl-heidelberg.de> References: <4ipmsc$ktq@stc06.ctd.ornl.gov> In article <4ipmsc$ktq@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > I've heard that there is a dbadaptor for mSQL available, but have been unable > to locate it. Any clues? Check: ftp://ftp.nmr.embl-heidelberg.de/pub/next/Database/Adaptors -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387437 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: martin@gecko.fb6.fhtw-berlin.de (Martin Bachmayer) Newsgroups: comp.sys.next.programmer Subject: PS to TIFF Date: 22 Mar 1996 13:26:31 GMT Organization: Fachhochschule fuer Technik und Wirtschaft Berlin, FB1 Message-ID: <4iu9q7$ps4@commserv.fb1.fhtw-berlin.de> Hi, there! At the moment it is not immediatly evident how to convert a PS to TIFF. There are three points making trouble. 1. Do you have a clue, how you can get an 1-Bit-Tiff (black and white)? By now I use NX_TwoBit... or NX_EightBit... Tiff. 2. How can you adjust the dpi value of a Tiff? For example, not 72 but 144 dpi. 3. The PS can consist of colors too, so its not only black, but CMYK. Converting this PS by now, brings out only the last color to Tiff (f.e. yellow). But I do need all colors to be converted. This is the piece of code I use by now: /* start*/ inputEPSImage = [[NXImage alloc] initFromFile: sourceName]; [inputEPSImage getSize: &aSize]; aRect.size = aSize; outputTIFFImage = [[NXImage alloc] initSize: &aSize]; [outputTIFFImage setBackgroundColor: NX_COLORWHITE]; [outputTIFFImage useCacheWithDepth: NX_TwoBitGrayDepth]; stream = NXOpenMemory(NULL, 0, NX_WRITEONLY); if ([outputTIFFImage lockFocus]) { [[inputEPSImage lastRepresentation] draw]; [outputTIFFImage unlockFocus]; } [outputTIFFImage writeTIFF: stream allRepresentations: YES usingCompression: NX_TIFF_COMPRESSION_NONE andFactor: 0.0]; /* end */ Any hints appreciated! I would be grateful for a piece of code! Thanks, Martin ------ Martin Bachmayer martin@gecko.fb6.fhtw-berlin.de (NeXTMail welcome) FHTW Berlin
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re: dbadaptor for mSQL? Date: 22 Mar 1996 14:14:14 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4iucjm$ikg@stc06.ctd.ornl.gov> References: <4iu782$4l3@lion.embl-heidelberg.de> In article <4iu782$4l3@lion.embl-heidelberg.de> tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) writes: > In article <4ipmsc$ktq@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. > Wooten) writes: > > I've heard that there is a dbadaptor for mSQL available, but have been > unable > > to locate it. Any clues? > Check: > > ftp://ftp.nmr.embl-heidelberg.de/pub/next/Database/Adaptors > Thanks, Located that, then discovered that I needed the "patched" version 1.0.7 of msql for the Next. Well, located that, downloaded it. Followed installation instructions. Guess What! msql.o next_tmpnam.o libmsql.a ../lib/libmisc.a -o msql -L../lib ld: warning next_tmpnam.o cputype (7, architecture i386) does not match cputype (6 architecture m68k) of objects files previously loaded (file not loaded) ld: Undefined symbols: _NeXT_tmpnam *** Exit 1 Stop. *** Exit 1 Stop. Somehow, the "patched" version is only for intel? Or else how do I fix the above problem. thanks again.
From: gblock@homer.alpha.net (Greg Block) Newsgroups: comp.sys.next.programmer Subject: MiscKit gains 3DKit, but what does that *mean*??? Date: 22 Mar 1996 15:05:59 GMT Organization: Alpha.net -- Milwaukee, WI Message-ID: <4iufkn$867@homer.alpha.net> Okay, so 3DKit is being added to the MiscKit. But will it be compiled for new versions of NS/OpenStep? Will it remain backwards compatible? Will new features be added, or will it stagnate and die? Is this the death of 3DKit, or a way of passing it to the vendors that rely on it? What does this mean for apps based on the 3DKit? Should I be writing a 3DKit-replacement for my own use to allow 3DKit to die? If I wanted to work on adding stuff to 3DKit, would that be possible? Boy, I'm just a bottle of questions, aren't I. Greg -- ): ): ): ): Have you overdosed on frownies today? Why NOT!?! :( :( :( :( ): "One if by land, two if by sea" - Paul Revere, encryption 1775 :( ): Fight government censorship. Greg.Block@perspex.com :( ): ): ): ): ): ): ): ): ): ): ): ): :( :( :( :( :( :( :( :( :( Wubba :(
From: ton@tiac.net (Terence C. O'Neill) Newsgroups: comp.sys.next.programmer Subject: Re: dbadaptor for mSQL? Date: 22 Mar 1996 15:03:24 GMT Organization: The Internet Access Company Message-ID: <4iuffs$djn@sundog.tiac.net> References: <4ipmsc$ktq@stc06.ctd.ornl.gov> <4iu782$4l3@lion.embl-heidelberg.de> In-Reply-To: <4iu782$4l3@lion.embl-heidelberg.de> On 03/22/96, Georg Tuparev wrote: >In article <4ipmsc$ktq@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. >Wooten) writes: > >> I've heard that there is a dbadaptor for mSQL available, but > >> have been unable to locate it. Any clues? > >Check: > > > >ftp://ftp.nmr.embl-heidelberg.de/pub/next/Database/Adaptors > > The distribution refers to the MiniSQL server distribution version 1.0.7.NS from: ftp://ftp.stepwise.com/pub/Vendors/M_Onyschuk_and_Associates_Inc. Unfortunately, I can't get access to the server. Anyone know where to find a copy? Thanks, Terry -- Terry O'Neill......................................................ton@tiac.net Open Systems Consulting, Inc......................................NeXTMail/MIME
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: sybase kernel panic with EOModeler Date: 22 Mar 1996 18:40:12 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4ius6c$s2d@stc06.ctd.ornl.gov> I have an original version of sybase that came with my cube. I've been keeping it alive and using it with the DBModeler and other projects. It still works fine except when I try to create a new EOModel and choose Sybase as the adaptor. Then I get a kernel panic when I log into the SYBASE server (even though if I do this from DBModeler or isql things work fine). Is there a cure? Is this just the incompatibility of the older version of sybase with EOModeler? -- - - - - - - - - - J. W. Wooten
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re: dbadaptor for mSQL? Date: 22 Mar 1996 18:37:29 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4ius19$rvq@stc06.ctd.ornl.gov> References: <4iucjm$ikg@stc06.ctd.ornl.gov> In article <4iucjm$ikg@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. Wooten) writes: > In article <4iu782$4l3@lion.embl-heidelberg.de> > tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) writes: > > In article <4ipmsc$ktq@stc06.ctd.ornl.gov> woo@woonext.cmo.ornl.gov (John W. > > Wooten) writes: > > > I've heard that there is a dbadaptor for mSQL available, but have been > > unable > > > to locate it. Any clues? > > Check: > > > > ftp://ftp.nmr.embl-heidelberg.de/pub/next/Database/Adaptors > > > Thanks, > > Located that, then discovered that I needed the "patched" version 1.0.7 of msql > for the Next. Well, located that, downloaded it. Followed installation > instructions. Guess What! > > msql.o next_tmpnam.o libmsql.a ../lib/libmisc.a -o msql -L../lib > ld: warning next_tmpnam.o cputype (7, architecture i386) does not match cputype > (6 architecture m68k) of objects files previously loaded (file not loaded) > ld: Undefined symbols: > _NeXT_tmpnam > *** Exit 1 > Stop. > *** Exit 1 > Stop. > > > Somehow, the "patched" version is only for intel? Or else how do I fix the > above problem. Well, found that if I changed directory into the Next-3.3 directory and carefully went through each subdirectory removing all *.o files, then redid the make install, things worked better. I now have a working msql and a EOFAdaptor for it. thanks to all who helped.
From: andrew_abernathy@omnigroup.com Newsgroups: comp.sys.next.programmer Subject: cd audio control Date: 22 Mar 1996 20:05:21 GMT Organization: Omni Development, Inc. Message-ID: <4iv161$5jl@gaea.titan.org> Can anyone help me programmatically control the audio functions on a cdrom drive? I can play, stop, eject, and I can pause... but I can't resume after pausing - I get scsi i/o errors when I send the resume command. (I also can't get subchannel data to determine current disc position, but I haven't knocked on that one so much yet.) Suggestions or pointers gratefully accepted. -- andrew_abernathy@omnigroup.com - NeXTmail & MIME ok I probably don't speak for Omni, but I'm going to keep you guessing.
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: Help with event loops Date: 22 Mar 1996 11:10:42 -0600 Organization: MCSNet Services Message-ID: <4iumui$7u6@Mercury.mcs.com> [ Article crossposted from comp.sys.next.software ] [ Author was Peter Richardson ] [ Posted on 22 Mar 1996 11:08:12 -0600 ] I have a background task process that I have written. It worked ok until I wanted to receive back unsolicited DO messages. Then I followed the example in NeXT.Answers to add an event loop to allow this to work. And it works!! The problem is that it uses every available CPU cycle to keep checking the status of the event queue. Is there a better way of doing this? Thanks for any help Peter Richardson
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Contract/DC Area Date: 22 Mar 1996 20:52:24 GMT Organization: Norden 1 Communications Message-ID: <4iv3u8$6c2@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF--A Plus Sybase or Oracle Contract--long term Greater DC Area To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: miniSQLEOFAdaptor problem Date: 22 Mar 1996 22:31:17 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4iv9nl$ddn@stc06.ctd.ornl.gov> I have loaded the MiniSQLEOFAdaptor for EOF and then converted several example Sybase database loads to load under mSQL without error. Whenever I try some of the examples, (after editing the eomodel file), and then trying them, I get a message. Adaptor channel could not evaluate expression "SELECT t0.ADDRESS, t0.CITY, t0.DEPT_ID, t0.EMP_ID, t0.FIRST_NAME, t1.TITLE, t0.LAST_NAME, t0.SALARY, t0.TITLE_ID FROM JOB_TITLE t1, EMPLOYEE t0 WHERE t0.DEPT_ID = '100' AND t0.TITLE_ID = t1.TITLE_ID". I thought that msql could handle expressions like this. - - - - - - - - - J. W. Wooten
From: osg@onramp.net (Glenda ) Newsgroups: comp.sys.next.programmer Subject: Good work/Great company Date: Sat, 23 Mar 1996 01:20:14 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4ivci7$juk@news.onramp.net> NeXT DEVELOPERS OSG (Object Systems Group) is a technology-based consultancy that provides assistance to Global 1000 corporations in new and emerging technologies. We currently have multiple opening in the N.E. for someone with the following experience: NeXT Developer Minimum 2+ years total experience with extensive NeXTStep, AppKit, DBKit and Objective C. Send resume to osg@onramp.net , US mail your resume to Object Systems Group, Inc., Attn: Glenda Maddox, 3048 Infomart, 1950 Stemmons Freeway, Dallas, Tx. 75207, or fax to 214 742-5847.
From: systemx@val.net (Dave Ian) Newsgroups: comp.sys.next.programmer Subject: Looking for work? Understand PERL & TCL? Date: Sat, 23 Mar 1996 05:24:39 GMT Organization: ValNet, Inc. Message-ID: <4j01fi$6ua@ooze.val.net> I am looking for a cgi, perl, tcl expert to work part-time from the convienence of their home. If you are intrested please contact me at Intelek Internet Services, 818.752.8252 Dave Ian systemx@val.net
From: flight@mathi.uni-heidelberg.de (Gregor Hoffleit) Newsgroups: comp.sys.next.programmer Subject: Re: dbadaptor for mSQL? Date: 23 Mar 1996 17:24:36 GMT Organization: University of Heidelberg, Germany Distribution: world Message-ID: <4j1c4k$112@sun0.urz.uni-heidelberg.de> References: <4iucjm$ikg@stc06.ctd.ornl.gov> <4ius19$rvq@stc06.ctd.ornl.gov> AFAIK, all the patches of the NEXTSTEP msql-1.0.7 have been integrated into the `mainstream' mSQL distribution. Therefore, you should be able to fetch the source of the most recent mSQL version (1.0.13) and make and install it for NEXTSTEP ("make target" in top directory, "cd targets/NEXTSTEP-XXX", "setup" and "make"). BTW, MiniSQL's home page is http://hughes.com.au/product/msql, and it's NO PUBLIC DOMAIN software any more!!! Gregor -- | Gregor Hoffleit admin MATHInet / contact RhiNO | | MAIL: Mathematisches Institut PHONE: (49)6221 54-5771 | | INF 288, 69120 Heidelberg / Germany FAX: 54-8312 | | EMAIL: flight@mathi.uni-heidelberg.de (NeXTmail, MIME) |
From: trombino@wendy.ucsd.edu (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Text Fields & First Responder Date: 24 Mar 1996 18:05:29 GMT Organization: The University of California at San Diego Message-ID: <4j42t9$d06@news1.ucsd.edu> Hi all, I have a simple question that I hope someone will help me with... I have a few text fields on a window. The window is set up to notice some keydown events, which it does fine. However, when you select the text or modify one of the text fields, it becomes the first responder and intercepts the keydown events... My question is, how can I either: 1) unselect text in a text field or 2) how can I force the text field to resign the first responder? 3) whatever else I could do to get keydown events to the window when the user is finished editing the text field. The documentation doesn't describe this directly. I hoping that someone out there can help.... Thanks in advance! -- ================================================================== Mark Trombino trombino@wendy.ucsd.edu Center For Research in Computing the Arts, U.C.S.D. ==================================================================
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: MiscKit gains 3DKit, but what does that *mean*??? In-Reply-To: gblock@homer.alpha.net's message of 22 Mar 1996 15:05:59 GMT Message-ID: <RDL.96Mar24151707@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4iufkn$867@homer.alpha.net> Date: Sun, 24 Mar 1996 20:17:07 GMT This generous move by NeXT means that we get the SOURCE CODE to both 3DKit and IndexingKit. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: dbora@tezcat.com (Donald F. Bora) Newsgroups: comp.sys.next.programmer Subject: Calling C Function from Objective C Date: 24 Mar 1996 16:45:16 -0600 Organization: Breakfast at Tezcat. Message-ID: <4j4j9s$peu@huitzilo.tezcat.com> I have a C function in a file and that file is the the "Other sources" section of the Project. I have an Objective C file which contains a method which calls that C routine. I am getting linking errors. I have even tried to compile the C routine into it's own lib file but the linker is still claiming that the C Routine is undefined. Any clues?
Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: PRESS RELEASE: MiscKit Announces the Transfer of the IndexingKit and 3DKit In-Reply-To: Van_Simmons@vnp.com's message of Fri, 22 Mar 96 13:29:02 GMT Message-ID: <RDL.96Mar24221230@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <1996Mar22.132902.12833@vnp.com> Date: Mon, 25 Mar 1996 03:12:30 GMT When I first read the press release, I somehow managed to skip the fine print about these kits being binary-only distribution. That's bad. It should be distributed like MusicKit. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: lozinski@cup.portal.com (Christopher A Lozinski) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer Subject: NEXTSTEP JOBS Date: 24 Mar 1996 21:20:04 -0800 Organization: The Portal System (TM) Sender: pccop@unix.portal.com Distribution: world Message-ID: <151155@cup.portal.com> NEXTSTEP DEVELOPERS NEEDED a. NEXTSTEP DEVELOPERS URGENTLY NEEDED b. VACATION SPECIAL $90 K c. INTRODUCTORY NEXTSTEP POSITIONS d. FOREIGN NEXT DEVELOPERS e. GERMAN SPEAKER NEEDED $80K f.CHICAGO FINANCIAL FIRMS g.FLORIDA POSITION h.NORTHERN VIRGINIA POSITION NEXTSTEP DEVELOPERS NEEDED There are companies all over the United States, even in Europe asking me to help them find people. In addition to the jobs listed, companies in Boston, Holland, and the San Francisco Bay Area are accepting additional resumes, but are not actively recruiting. Additional jobs arrive on my desktop all the time. I have one or two companies a week asking me to help them find someone. I maintain a list of job seekers, and I post announcements to that list as the jobs become available. If you are thinking of switching jobs, let me know. I probably have just the job for you. I can give you advice on your best job strategy, and I can tell you a lot about the companies that you are considering. If you are interested in any of the following positions, please give me a call at (510) 795-6086, or send email to lozinski@bpg.com with your resume. NeXTMAIL RTF preferred. NEXTSTEP DEVELOPERS URGENTLY NEEDED. This San Francisco area company needs NEXTSTEP developers urgently. They are interviewing, and flying people in like crazy. 22 days paid vacations. 10 days holidays. 6 Sick Days. Almost as much vacation as in Europe. Project Bonuses. $1500- $5,000. Stock Options. Excellent benefits. They pay all the health premiums for the employees and their dependents. The project is large enough to have a variety of positions, so they should have something just right for you. They are looking for a number of people, and are particularly interested in someone with extensive GUI experience. VACATION SPECIAL $85 K Warm Sunny Climate. Help a resort develop their business applications, and enjoy the resort while you are at it. Cheap Housing, World Class Entertainment, International City, strong education systems. You need experience developing hotel or resort reservation applications. You also need1 Year NEXTSTEP, 2 Years Objective-C, 4 Years Unix , 2 Years RDBMS. The company would prefer if you had experience with HP Unix , Informix, and EOF. US Citizens Only. Just don't tell the company that I called it a vacation special. INTRODUCTORY NEXTSTEP POSITIONS I am now able to find positions for developers with not much experience. I have companies in Los Angeles, Michigan, the San Francisco Bay Area, Germany, and Virginia looking for beginning NEXTSTEP developers. I probably know companies in the city of your choice looking for beginning NEXTSTEP developers. This generally means people with up to six months NEXTSTEP development experience, although sometimes I can place people who are active in their user group, but have not developed professionally on NEXTSTEP. FOREIGN NEXT DEVELOPERS I am now able to find positions for foreign developers with NEXTSTEP experience. I have at least one US consulting company interested and willing to hire foreign developers for projects opening up around the US. They will process your H1 visa. I also have one german company looking for NEXTSTEP developers. EEC Citizenship preferred. GERMAN SPEAKER NEEDED $80K The company needs a german speaking NEXTSTEP developer. They are willing to pay $80,000 = Deutsche Marks 118,400 per Year. This is considered a desirable location, close to the Alps and skiing. 28 days vacation and 11 days holiday per year. CHICAGO FINANCIAL FIRMS. Continue to need NeXT developers. They filled the last openings. One company seeks two more strong NeXT developers with experience in relational databases and EOF. The other company seeks one experienced, and one beginner NEXTSTEP developer. Financial Industry experience is a strong positive. Full Benefits, top pay, bonuses, tuition assistance. FLORIDA POSITION This is an interesting and large project. They are running multiple databases and the Tuxedo Encina transaction Processor ported to NEXTSTEP. The applications are being hosted on the Web. Great place to live. You have the choice of working for the company directly as an employee, or, if you have enough experience, working through a consulting company. NORTHERN VIRGINIA POSITION Beautiful place to live one hour to mountains. 1 Hour to Chesapeake Bay. 2 Hours to the Atlantic. 1-5 acre plots average homes. Affluent, Educated neighborhood. It is the best time of year. Almost Spring. You have a choice of two consulting companies to work for. Each one reports to the same end client. 6 Month extensible contracts. Your ability to sell yourself to the client is very important. Oracle Database. EOF. Multi-Company Distributed Databases. 4 Years of work Experience needed. 1 Year of NEXTSTEP Experience preferred.
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Calling C Function from Objective C Date: Mon, 25 Mar 1996 08:23:56 +0000 Organization: Cedar Systems Distribution: world Message-ID: <GJexMDAcglVxEwE4@cedar.co.uk> References: <4j4j9s$peu@huitzilo.tezcat.com> MIME-Version: 1.0 In article <4j4j9s$peu@huitzilo.tezcat.com>, "Donald F. Bora" <dbora@tezcat.com> writes > >I have a C function in a file and that file is the the "Other sources" >section of the Project. I have an Objective C file which contains a >method which calls that C routine. I am getting linking errors. I have >even tried to compile the C routine into it's own lib file but the >linker is still claiming that the C Routine is undefined. > > > Any clues? > > > Include an external declaration for the method in the file containing the function call: extern int foo(); Paul. ------------------------------------------------------------------ Paul Heffernan Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: PJ@pros.com (Patrick (PJ) Caraher) Newsgroups: comp.sys.next.programmer Subject: Mail App and sendmail Date: 25 Mar 1996 16:47:46 GMT Organization: Argonne National Laboratory Message-ID: <4j6inj$e2d@milo.mcs.anl.gov> I have a NeXTStation connected via PPP to a local internet provider and I have some questions about sending mail. My incoming mail gets sent to the internet provider and I download it to my machine via POP. I use the PopOver App and this places incoming mail directly into my Mail App (very nice indeed!) I would now like to send mail from my Mail App. My Mail App is currently configured to call sendmail. How do I go about configuring sendmail to forward my mail out to my PPP-connected internet provider? (I assume that it has something to do with the sendmail.subsidiary.cf file but I am not certain.) Any help or pointers would be greatly appreciated. PJ Caraher pj@pros.com
From: lph@sei.cmu.edu (Larry Howard) Newsgroups: comp.sys.next.programmer Subject: linking posix apps Date: 25 Mar 1996 17:00:22 GMT Organization: The Software Engineering Institute Distribution: world Message-ID: <4j6jf6$g89@news.sei.cmu.edu> Could someone give me a few pointers on linking (NS3.3) applications compiled with -posix. If there's something to read somewhere, just let me know. Thanks. lph -- Larry Howard Software Engineering Institute, Carnegie Mellon University lph@sei.cmu.edu (NeXT Mail OK) (412) 268-6397
From: lutzray@9bit.qc.ca. (Raymond Lutz) Newsgroups: comp.sys.next.programmer Subject: Re: NS encoding <-> Latin 1? Date: 13 Mar 1996 06:08:56 GMT Organization: Bell Global Solutions Message-ID: <4i5opo$acj@news.bellglobal.com> References: <4hsq8d$uf@turbocat.snafu.de> <4i1sif$15f@precipice.fdn.fr> In-Reply-To: <4i1sif$15f@precipice.fdn.fr> In article <4hsq8d$uf@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) writes: > Hi! > > Is there a simple way to convert ISO Latin 1 to NEXTSTEP encoded strings? > (and the other way) following a pointer given by Charles William Swiger in csn.misc (sept 94), I use the GNU recode utility, available to any good ftp site near you... and at ftp://prep.ai.mit.edu/pub/gnu/recode-3.4.tar.gz (361k) As Chuck said, "It understands NeXT's ISOLatin1 along with the more common formats...." Ray
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: NXImage in EOF Date: 25 Mar 1996 21:51:57 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4j74ht$kms@stc06.ctd.ornl.gov> I've implemented an EOF app that has Departments listed in the top browser. When you select a department, you get a list of employees in that department in the lower browser. I want an image of the employee to appear in a NXImageView when I select an employee. I'm using an NSDictionary with the image key having the value of a path to a tiff image. How do I get the retrieved value (that path) to get imaged in the NXImageView? -- - - - - - - - - - J. W. Wooten
From: jbn@mystery-train.cu-online.com (J.B. Nicholson-Owens) Newsgroups: comp.sys.next.programmer Subject: Re: MiscKit gains 3DKit, but what does that *mean*??? Date: 26 Mar 1996 02:25:46 GMT Organization: Organizing Organisms & Organs Message-ID: <slrn4lehnm.6qt.jbn@mystery-train.cu-online.com> References: <4iufkn$867@homer.alpha.net> <RDL.96Mar24151707@world.std.com> On Sun, 24 Mar 1996 20:17:07 GMT, Robert La Ferla <rdl@world.std.com> wrote: > This generous move by NeXT means that we get the SOURCE CODE to both 3DKit > and IndexingKit. According to the FAQ (http://www.misckit.com/faq/NXKits.html) and announcement on comp.sys.next.announce, this is untrue. The MiscKit public does not get access to the source code. As the announcement stated, binaries will be distributed. Those chosen to maintain the source, get access to the source: (from the FAQ) Q: What about the source code? How can I get it? A: The terms of the agreement do not allow the source code for either kit to be publically distributed. However, a select few MiscKit workers will be able to help in maintaining the code for the kits. At this time it has not been decided who these people will be nor how many people will be put on each team. However, we do know that the working team for each kit will be fairly small. If you want to be a part of the development team for either or both of the kits, contact Don Yacktman at don@misckit.com if you meet these criteria: * You have enough time to actually be able to contribute. * You aren't legally bound by your current employer in such a way that you cannot return your fixes and enhancements. * You already have a solid foundation in the kit(s) you wish to participate in and use them yourself. * You aren't expecting to get paid for any of your work. The above criteria are not absolute requirements, but they do encapsulate some of the things that Don will be looking for as he chooses the teams for maintaining the kits. The reason for being selective is that only a limited number of people will be able to participate, so we want to assure that the people working on the project are the best we can get, so that all those who use these kits will achieve the maximum benefit.
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: missing "flag" on nib file in ProjectBuilder In-Reply-To: woo@polaris.scicntr.ortn.edu's message of 15 Mar 1996 13:57:04 GMT Message-ID: <RDL.96Mar25203017@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4ibsvg$7t9@stc06.ctd.ornl.gov> Date: Tue, 26 Mar 1996 01:30:17 GMT The flag just means that it is localized. i.e. not global and specific to a (.lproj) language subproject. You can change it by going to the Files menu and choosing Make Localizable / Make Global. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: jsamson@agt.net (Jean-Paul Samson) Newsgroups: comp.sys.next.programmer Subject: Re: Mail App and sendmail Date: 26 Mar 1996 00:43:05 GMT Organization: Public Live Access Network (PLAnet) Message-ID: <4j7eip$dbr@tigger.planet.eon.net> References: <4j6inj$e2d@milo.mcs.anl.gov> In-Reply-To: <4j6inj$e2d@milo.mcs.anl.gov> On 03/25/96, Patrick (PJ wrote: >I would now like to send mail from my Mail App. My Mail App is >currently configured to call sendmail. How do I go about >configuring sendmail to forward my mail out to my PPP-connected >internet provider? You've got the right idea--you will have to make some minor but trivial modifications to "/etc/sendmail/sendmail.subsidiary.cf". First, make a backup copy of the file for safe keeping. Near the top of sendmail.subsidiary.cf, you will see a couple of lines like this: DRmailhost CRmailhost Replace "mailhost" with the host name of your Internet service provider's mail server. Typically, the mail server is called "mail", in which case these lines should be altered thus: DRmail CRmail Just below this, there is a setting for the domain name of your Internet service provider: DmPodunk.EDU Change this to be the domain name of your provider. For example, my provider uses the domain name "agt.net", so the ammendment is: Dmagt.net This should get everything working properly for sending mail using the Mail application. -- -= Jean-Paul (Sean) Samson =------= St. Albert, Alberta, Canada =- -= jsamson@agt.net =-------------------------= NeXTmail welcome =-
From: gad@eclipse.its.rpi.edu (Garance A. Drosehn) Newsgroups: comp.sys.next.programmer Subject: Re: cd audio control Date: 26 Mar 1996 03:59:00 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Distribution: world Message-ID: <4j7q24$jnv@usenet.rpi.edu> References: <4iv161$5jl@gaea.titan.org> andrew_abernathy@omnigroup.com writes: > > Can anyone help me programmatically control the audio functions > on a cdrom drive? I can play, stop, eject, and I can pause... > but I can't resume after pausing - I get scsi i/o errors when I > send the resume command. (I also can't get subchannel data to > determine current disc position, but I haven't knocked on that > one so much yet.) Suggestions or pointers gratefully accepted. Look for something called mCD.app on many of the standard archives. One location (and the one that's the most up-to-date) is: ftp://eclipse.its.rpi.edu/NeXT/apps/mCD_f/ That directory includes four single-architecture binaries, one quad-fat binary, and another archive which is all of the source code. I do not claim the source is pretty, but it generally seems to work. Note that some of the routines in the scsi_cd.subproj have not been updated to work on NS/Intel (or other architectures), but all of those routines which are called by the main mCD program should work OK. -- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: moellney@michi.bota.uni-bonn.de (Michael Moellney) Newsgroups: comp.sys.next.programmer Subject: NeXTSTEP and CORBA 2.0 Date: 26 Mar 1996 13:52:33 GMT Organization: University of Bonn, Germany Message-ID: <4j8sr1$9h5@news.rhrz.uni-bonn.de> Hi! On CeBIT'96 I visited SUN and NEXT. Sun told me about their NEO Project that it now uses CORBA 1.2 and will use CORBA 2.0 at the end of the year. Someone at NEXT told me, that PDO 3.0 (and so DO of NS ??) uses the NEXT ORB, but PDO 4.0 will use CORBA 2.0 also. My Question is, will OPENSTEP for MACH 4.0 and OPENSTEP for W/NT contain DO using CORBA 2.0 or will it still use NEXT ORB (or both???) Thanks for any hints. Bye, Michael
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: linking posix apps Date: 26 Mar 1996 17:44:43 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4j9aeb$7ml@news.its.com> References: <4j6jf6$g89@news.sei.cmu.edu> lph@sei.cmu.edu (Larry Howard) wrote: > Could someone give me a few pointers on linking (NS3.3) > applications compiled with -posix. If there's something > to read somewhere, just let me know. If you search Digital Librarian for posix, you'll find some comments in the compiler docs and release notes. Anyway, just use "cc -posix -o executable file1.o file2.o...." to link. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: NextDeveloper/EOF/Examples in msql Date: 26 Mar 1996 18:13:00 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4j9c3c$q2b@stc06.ctd.ornl.gov> I've converted the installPEOPLE.sybase loaddataset so that it builds a msql database and then converted the eomodel files for UniqueKey, Validation, Projects, and PeopleDemo so that they use the miniSQL Adaptor and work with the examples. I thought I'd pass on a few things I've learned. 1) The miniSQL adaptor checks for the type assuming the eomodel file has the type listed in caps. Thus if you have type = int; It will get passed in the SQL call as a string rather than as an integer. I had to edit the eomodel for all the variable types to be in caps. 2) Special types like varchar need to be made CHAR, float becomes FLOAT. 3) derived values that use combined keys have to be removed. 4) Some of the nib files have to have derived columns removed and replaced by something else. One question for msql experts: Sybase can handle the expression coming from qualifierString = [NSString stringWithFormat: @"EmpId not in (SELECT z1.EMP_ID FROM EMPLOYEE z1, EMP_PROJECT z2 " @"SELECT z1.EMP_ID FROM EMPLOYEE z1, EMP_PROJECT z2 " @"WHERE z1.EMP_ID = z2.EMP_ID AND z2.PROJECT_ID = %@",projectId]; (note the not in and the nested selects). I implemented this as qualifierString = [NSString stringWithFormat:@""]; and thus don't remove the present projects that the employee is in from the list to be displayed for adding the employee to. How could I do this in EOF? I.E. after otherEmployees = [[[EOQualifier alloc] initWithEntity:employeeEntity qualifierFormat:qualifierString] autorelease]; What could I do to remove the projects that the employee is already in? -- - - - - - - - - - J. W. Wooten
From: rthrford@Glue.umd.edu (Reginald Rutherford) Newsgroups: comp.mail.mh,comp.sys.next.sysadmin,comp.sys.next.programmer,comp.sys.next.misc Subject: NeXT mh installation Date: 26 Mar 1996 16:54:36 GMT Organization: Project GLUE, University of Maryland, College Park, MD Message-ID: <4j97gc$2ts@mojo.eng.umd.edu> Hiya I'm trying to install mh 6.8.3 on a NeXT Cube running NeXTStep 3.2 and I'm running into trouble. Any help at all would be appreciated! The makefile will make with no errors (but a bunch of warnings), but when I try to run "inc" I get the error message: inc: unable to lock and fopen /usr/spool/mail/rthrford Here's the conf/MH file: # NeXTStep 3.2 bin: /usr/local/src/mh/bin chown: /etc curses: -lcurses -ltermlib editor: /usr/ucb/vi etc: /usr/local/src/mh/lib/mh ldoptions: -s mail: /usr/spool/mail mandir: /usr/local/src/mh/man manuals: local mts: sendmail options ATTVIBUG BIND BSD42 BSD43 LOCKF options NORUSERPASS NTOHLSWAP RENAME UNISTD VSPRINTF ZONEINFO options MHRC MIME ranlib: on remove: mv -f Note: I have tried using both the "LOCKF" and "FLOCK" options, both with and without "UNISTD." I have used all four of these combinations with both the "BDS42" option and with "BSD42 BDS43", and I always get the same error with "inc". Thank you very much! -- Rick Rutherford rthrford@sirius.umd.edu The above opinions are mine. "It seems to me that the nearer painting approaches sculpture the better it is, and that sculpture is the worse the nearer it approaches painting." -- Michelangelo
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 25 Mar 1996 05:15:09 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4j5a4t$cj0@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: Joan ordinas <jordinas@xtec.es> Newsgroups: comp.sys.next.programmer Subject: Missing waitpid and setpgid Date: Wed, 27 Mar 1996 15:27:25 +0100 Organization: PIE Message-ID: <3159504D.49EA@xtec.es> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I've problems porting an utility, Posix 1 compliant, to the NeXT. The problem consists in the absence of waitpid and setpgid functions. The functions are declared in the correspondig headers, but the loader say ld: Undefined symbols: _waitpid *** Exit 1 Could somebody help me? Please, answer directly to my email address. Thanks in advance. +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Joan Josep Ordinas Rosa | | Email: jordinas@pie.xtec.es Home page: http://phonos.upf.es/~jordinas/ | | "Comunication is too important to leave it in the specialists' hands" | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
From: paulrs@lgs-systems.buf.servtech.com (Paul R. Summermatter) Newsgroups: comp.sys.next.programmer Subject: looking, comparing, and duplicating folders... Date: 27 Mar 1996 19:48:24 GMT Organization: LGS Systems Message-ID: <paulrs-2703961544330001@lgs-systems.buf.servtech.com> Hello All, Due to some security issues that are poping up I wanted to write a few applications on montering folders and doing an invetory on them. Does anyone have any examples that I can look at? What should I look at? Any help would be very nice. Thanks peter@lgs-systems.com NeXT mail ok.
From: cbrooks@futon.SFSU.EDU () Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: Help with Sybase on NS 3.3 (black) Date: 27 Mar 1996 20:55:03 GMT Organization: San Francisco State University Distribution: world Message-ID: <4jc9v7$28d@nuke.csu.net> Hi. We're working on an online quiz program here at San Francisco State using Sybase as the back-end DB. We're running this on a Black NeXTstation running NS 3.3 with 32 MB of RAM. Here's the problem - we're unable to open more than a few connections to the SQL server at one time (we'd like to have about 15) - the docs say that this can be changed by reconfiguring the database, but that the maximum number of connections is bounded by the variable @@max_connections. All they say about this is that it's OS dependent. On our system this is 8, which is much lower than we'd like it to be. (The examples in the documentation show @@max_connections = 30, and the SQL Server can supposedly support 250 connections.) So, my questions are, how can we raise this number, and/or, what exactly is it dependent on? Is it file descriptors? Semaphores? Available memory? If anyone out there has any help or insight, it'd be greatly appreciated. Thanks, Chris Brooks cbrooks@futon.sfsu.edu
From: "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> Newsgroups: comp.sys.next.programmer Subject: OPENSTEP/NEXTSTEP Coursework in Illinois Date: Wed, 27 Mar 1996 22:17:24 -0600 Organization: Illinois State University, Instructional Technology Services Message-ID: <315A12D4.5F12@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Can anyone tell me where a good place is in Illinois to take OPENSTEP/NEXTSTEP Coursework? Thanks! --------------------------------------------------------------------------------- Eric A. Dubiel; http://138.87.201.11 Instructional mailto:eadubie@rs6000.cmp.ilstu.edu Technology PEACE LOVE Services UNITY RESPECT Illinois State University "Understanding is best learned via experience." "Follow our instinct, not a trend. Go against the grain until the end." Unknown .... VIEWS EXPRESSED ARE ENTIRELY MY OWN
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Storage? Date: 27 Mar 1996 20:48:43 GMT Organization: Turbocat's Development, Germany Message-ID: <4jc9jb$cr@turbocat.snafu.de> Hi! The Storage class behaves like an array. Are the elements stored in a list? If the index changes (when I call removeElementAt) are all other elements copied to another position or is just one element in a double linked list changed? _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: Mehdi Bousaidi <mbousaidi@cen.com> Newsgroups: comp.sys.next.programmer Subject: Sun and OPENSTEP Date: Thu, 28 Mar 1996 10:54:59 -0500 Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <315AB653.133C@cen.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Would Sun's commitment to OPENSTEP be slipping away, I hope not, but check out the artcile about Sun's JOE and what it means for OPENSTEP in this week's InfoWeek magazine online at the following URL: http://techweb.cmp.com/iw/current/ -- ------------------------------------------------------------------ Mehdi Bousaidi | ____/* | email: mbousaidi@cen.com Century Computing | |___| | web: http://www.cen.com 8101 Sandy Spring | (~ ~) | phone: (301) 953-3330 ext. 129 Laurel, MD 20707 | @ V @ | Fax: (301) 953-2368 | \W/ | -------------------------------------------------------------------
From: allenlee@innet.be Newsgroups: comp.sys.next.programmer Subject: Leaves to nodes in NXBrowser Date: 28 Mar 1996 18:52:18 GMT Organization: INnet NV (post doesn't reflect views of INnet NV) Message-ID: <4jen52$5lt@news.be.innet.net> Hello. I have a NXBrowser in which leaves change to nodes. I implemented a normal delegate method that creates the columns from a tree data structure. A method adds the element to the structure and redisplays the column. My problem is that I can't get the new element to be displayed because its column is created by the delegate which is only called when I click on the new node (even the data structure is up-to-date). I'd like the new element of the node to be displayed automatically. I would appreciate any help or advice in solving the problem. Thanks in advance, allen
Newsgroups: comp.sys.next.programmer From: pelletk@il.us.swissbank.com (Ken Pelletier) Subject: Re: Help with Sybase on NS 3.3 (black) Message-ID: <1996Mar28.174600.29673@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <4jc9v7$28d@nuke.csu.net> Date: Thu, 28 Mar 1996 17:46:00 GMT Chris, You can't increase the number of connections beyond 5 - the version of Sybase sold (bundled in the good ol' days) by NeXT had a hardwired limit of 5 client connections. - Ken writes > > Hi. > > We're working on an online quiz program here at San Francisco State > using Sybase as the back-end DB. We're running this on a Black NeXTstation > running NS 3.3 with 32 MB of RAM. > Here's the problem - we're unable to open more than a few connections > to the SQL server at one time (we'd like to have about 15) - the docs say > that this can be changed by reconfiguring the database, but that the maximum > number of connections is bounded by the variable @@max_connections. All > they say about this is that it's OS dependent. On our system this is 8, > which is much lower than we'd like it to be. (The examples in the > documentation show @@max_connections = 30, and the SQL Server can > supposedly support 250 connections.) > So, my questions are, how can we raise this number, and/or, what > exactly is it dependent on? Is it file descriptors? Semaphores? > Available memory? > If anyone out there has any help or insight, it'd be greatly > appreciated. > > Thanks, > Chris Brooks > cbrooks@futon.sfsu.edu -- ________________o__________________ | \ | | Ken Pelletier | | NiKA Software | | 1207 W. Newport Ave. __| | Chicago, IL 60657 | /
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Leaves to nodes in NXBrowser Date: 28 Mar 1996 19:25:32 GMT Organization: University of Ghent, Belgium Message-ID: <4jep3c$n37@infoserv.rug.ac.be> References: <4jen52$5lt@news.be.innet.net> allenlee@innet.be wrote: : I have a NXBrowser in which leaves change to nodes. I implemented a : normal delegate method that creates the columns from a tree data : structure. A method adds the element to the structure and redisplays : the column. Although the name is somewhat frightening, I find it easier to work with a "very lazy" delegate. (or a "lazy" one). David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: nagendra@cs.bu.edu (nagendra mishr) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer Subject: Re: NEXTSTEP JOBS Followup-To: comp.sys.next.marketplace,comp.sys.next.programmer Date: 28 Mar 1996 22:03:30 GMT Organization: Computer Science Department, Boston University, Boston, MA, USA Distribution: world Message-ID: <4jf2bi$5h5@news.bu.edu> References: <151155@cup.portal.com> Christopher A Lozinski (lozinski@cup.portal.com) wrote: : NEXTSTEP DEVELOPERS NEEDED : a. NEXTSTEP DEVELOPERS URGENTLY NEEDED : b. VACATION SPECIAL $90 K : c. INTRODUCTORY NEXTSTEP POSITIONS : d. FOREIGN NEXT DEVELOPERS : e. GERMAN SPEAKER NEEDED $80K : f.CHICAGO FINANCIAL FIRMS : g.FLORIDA POSITION : h.NORTHERN VIRGINIA POSITION : NEXTSTEP DEVELOPERS NEEDED : There are companies all over the United States, even in Europe asking : me to help them find people. In addition to the jobs listed, : companies in Boston, Holland, and the San Francisco Bay Area are : accepting additional resumes, but are not actively recruiting. : Additional jobs arrive on my desktop all the time. I have one or : two companies a week asking me to help them find someone. I maintain : a list of job seekers, and I post announcements to that list as the : jobs become available. If you are thinking of switching jobs, let me : know. I probably have just the job for you. I can give you advice : on your best job strategy, and I can tell you a lot about the : companies that you are considering. If you are interested in any of : the following positions, please give me a call at (510) 795-6086, or : send email to lozinski@bpg.com with your resume. NeXTMAIL RTF : preferred. : NEXTSTEP DEVELOPERS URGENTLY NEEDED. : This San Francisco area company needs NEXTSTEP developers urgently. : They are interviewing, and flying people in like crazy. 22 days : paid vacations. 10 days holidays. 6 Sick Days. Almost as much : vacation as in Europe. Project Bonuses. $1500- $5,000. Stock : Options. Excellent benefits. They pay all the health premiums for : the employees and their dependents. The project is large enough to : have a variety of positions, so they should have something just right : for you. They are looking for a number of people, and are : particularly interested in someone with extensive GUI experience. : VACATION SPECIAL $85 K : Warm Sunny Climate. Help a resort develop their business : applications, and enjoy the resort while you are at it. Cheap : Housing, World Class Entertainment, International City, strong : education systems. You need experience developing hotel or resort : reservation applications. You also need1 Year NEXTSTEP, 2 Years : Objective-C, 4 Years Unix , 2 Years RDBMS. The company would prefer : if you had experience with HP Unix , Informix, and EOF. US Citizens : Only. Just don't tell the company that I called it a vacation : special. : INTRODUCTORY NEXTSTEP POSITIONS : I am now able to find positions for developers with not much : experience. I have companies in Los Angeles, Michigan, the San : Francisco Bay Area, Germany, and Virginia looking for beginning : NEXTSTEP developers. I probably know companies in the city of your : choice looking for beginning NEXTSTEP developers. This generally : means people with up to six months NEXTSTEP development experience, : although sometimes I can place people who are active in their user : group, but have not developed professionally on NEXTSTEP. : FOREIGN NEXT DEVELOPERS : I am now able to find positions for foreign developers with NEXTSTEP : experience. I have at least one US consulting company interested and : willing to hire foreign developers for projects opening up around the : US. They will process your H1 visa. I also have one german company : looking for NEXTSTEP developers. EEC Citizenship preferred. : GERMAN SPEAKER NEEDED $80K : The company needs a german speaking NEXTSTEP developer. They are : willing to pay $80,000 = Deutsche Marks 118,400 per Year. This is : considered a desirable location, close to the Alps and skiing. 28 : days vacation and 11 days holiday per year. : CHICAGO FINANCIAL FIRMS. : Continue to need NeXT developers. They filled the last openings. One : company seeks two more strong NeXT developers with experience in : relational databases and EOF. The other company seeks one : experienced, and one beginner NEXTSTEP developer. Financial : Industry experience is a strong positive. Full Benefits, top pay, : bonuses, tuition assistance. : : FLORIDA POSITION : This is an interesting and large project. They are running multiple : databases and the Tuxedo Encina transaction Processor ported to : NEXTSTEP. The applications are being hosted on the Web. Great : place to live. You have the choice of working for the company : directly as an employee, or, if you have enough experience, working : through a consulting company. : NORTHERN VIRGINIA POSITION : Beautiful place to live one hour to mountains. 1 Hour to Chesapeake : Bay. 2 Hours to the Atlantic. 1-5 acre plots average homes. : Affluent, Educated neighborhood. It is the best time of year. Almost : Spring. You have a choice of two consulting companies to work for. : Each one reports to the same end client. 6 Month extensible : contracts. Your ability to sell yourself to the client is very : important. Oracle Database. EOF. Multi-Company Distributed : Databases. 4 Years of work Experience needed. 1 Year of NEXTSTEP : Experience preferred.
From: gshaw@zeta.org.au (Greg Shaw) Newsgroups: comp.sys.next.programmer Subject: ORB palette or some such Date: 28 Mar 1996 23:00:43 GMT Organization: Kralizec Dialup Internet Sydney, +61-2-837-1183 V.32bis Message-ID: <4jf5mr$jq2@gidora.kralizec.net.au> Keywords: ORB orbject request broker A client of mine heard about a mechanism under NEXTSTEP for SPARC whereby InterfaceBuilder could talk to an Object Request Broker (ORB) object. I figured that this would have to be a palette of some description. Has anyone heard about a palette that does this sort of thing on any platform? Or another mechanism which would allow communication with an Object Request Broker. Thanks for any assistance, Greg Shaw.
From: gshaw@zeta.org.au (Greg Shaw) Newsgroups: comp.sys.next.programmer Subject: Linking Visual C++ libraries under OPENSTEP for Windoze Date: 28 Mar 1996 22:56:01 GMT Organization: Kralizec Dialup Internet Sydney, +61-2-837-1183 V.32bis Message-ID: <4jf5e1$jks@gidora.kralizec.net.au> Keywords: Windows Visual C++ gcc Does anyone know if it is possible to link Visual C++ libraries (third party) to an application written under OPENSTEP for windoze? My client wants to use IB and PorjectBuilder to do the development but they must link to a third party library of Visual C++ routines to complete their database interaction. Any hints? Thanks, Greg Shaw.
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Storage? Message-ID: <DozwpE.3HB@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4jc9jb$cr@turbocat.snafu.de> Date: Thu, 28 Mar 1996 20:53:38 GMT In article <4jc9jb$cr@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) writes: >The Storage class behaves like an array. Are the elements stored in a list? No, not in a linked list. >If the index changes (when I call removeElementAt) are all other elements >copied to another position or is just one element in a double linked list >changed ? It's the former. I find Storage objects very useful, for example, to store NXRect struct's for cellframes in a View. David Stes RIACA / Computer Algebra Objects e-mail : stes@can.nl www : http://www.can.nl/~stes
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: OPENSTEP/NEXTSTEP Coursework in Illinois In-Reply-To: "Eric A. Dubiel"'s message of Wed, 27 Mar 1996 22:17:24 -0600 Message-ID: <RDL.96Mar28224229@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <315A12D4.5F12@rs6000.cmp.ilstu.edu> Date: Fri, 29 Mar 1996 03:42:29 GMT NeXT offers classes in their downtown Chicago office. Call (800)-TRY-NEXT for details. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: OPENSTEP/NEXTSTEP Coursework in Illinois In-Reply-To: rdl@world.std.com's message of Fri, 29 Mar 1996 03:42:29 GMT Message-ID: <RDL.96Mar28224542@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <315A12D4.5F12@rs6000.cmp.ilstu.edu> <RDL.96Mar28224229@world.std.com> Date: Fri, 29 Mar 1996 03:45:42 GMT Oh yeah... The NeXT office is in the NBC Tower building. What's interesting about the building is that they film the NBC talk shows in the basement. e.g. Jenny Jones. Robert
From: kohler@lithnext (Vincent Kohler) Newsgroups: comp.sys.next.programmer Subject: Re: Leaves to nodes in NXBrowser Date: 29 Mar 1996 05:50:18 GMT Organization: Ecole Polytechnique Federale de Lausanne Message-ID: <4jftmq$eva@disunms.epfl.ch> References: <4jen52$5lt@news.be.innet.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit allenlee@innet.be wrote: : My problem is that I can't get the new element to be displayed because : its column is created by the delegate which is only called when I : click on the new node (even the data structure is up-to-date). I'd : like the new element of the node to be displayed automatically. Hi Allen, You should use the reloadColumn: method to force the browser to reload all cells within the specified column. Then, your delegate will update all cells according to your data structure. Regards Vincent --------------------------------------------------------------- Vincent Kohler Sen:te PSE 1010 Lausanne-Ecublens +41 21 693 8383 kohler@sente.epfl.ch +41 21 693 8380 (fax) http://sente.epfl.ch/
Newsgroups: comp.sys.next.programmer From: pemmerik@solair1.inter.NL.net (P.J.L. van Emmerik) Subject: syslog:-[MethodSignature engodeMethodParams... ????? Message-ID: <Dp0q00.AD1@solair1.inter.NL.net> Sender: news@solair1.inter.NL.net (News at news) Organization: Holec Projects Date: Fri, 29 Mar 1996 08:21:24 GMT From ObjectError i get a bactrace reporting a segmentation violation. In the backtrase it gives: syslog: -[NXPortal encodeData:ofType: :(ptr)0x1ff8d4:(char *)n^{FEPIOMessage=SS[500S]}16] syslog: -[MethodSignature encodeMethodParams:onto: :(ptr)0x1ff8cc :(id)0x0da038] I can not find any information on encodeMethodParams. I do not directly call this myself, so it must be generated by something i did call. In order to be able to solve the crasch i would like to know what the metod does and in what case it is called. Alse the "n^" and "16" part in the encodeData:ofType are not clear to me. I do use a structure: home FEPIOMessage { unsigned short startreg; unsigned short regCount; unsigned short regValue[500]; } This does explain the {FEPIOMessage=SS[500S]} part but not the rest. Thanks in advance Pieter. -- P.J.L. van Emmerik Holec Projects B.V. Email: pemmerik@solair1.inter.nl.net PO.BOX 565, 7550 AN Hengelo Phone: +31 74 2558 688 The Netherlands
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: Help with event loops Date: 29 Mar 1996 08:48:54 -0600 Organization: MCSNet Services Message-ID: <4jgt8m$fi8@Venus.mcs.com> Is there anybody out here who understands what happens in the Application run method? I am writing a process that requires by directional messages in the DO system, but does not have an application. I followed the example for creating an event loop in the main program that I got from the NeXT Answers page. The problem is that it is a tight loop that uses the entire CPU even when idle. Is there some way to get the correct idle behavior? Can you have an application that has no workspace(ie started from rc.local) that can then use the run method? Any help would be greatly appreciated. Peter Richardson
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9603291454.AA01774@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Fri, 29 Mar 96 08:53:58 -0600 Subject: NEXT bundled/Sold Sybase References: <199603291247.EAA03512@shellx.best.com> > Chris, > You can't increase the number of connections beyond 5 - the version of Sybase > sold (bundled in the good ol' days) by NeXT had a hardwired limit of 5 client > connections. > - Ken I'm glad to hear that something like this may be available. I'm considering putting a NEXT box out at the house, on which I would need to do EOF work in order to justify the cost...a bit problematic, since I don't have a Sybase instance sitting around the living room. Ultimately, if the idea pans out, the office will get single-channel ISDN for me, but I need to do some proof-of-concept for the whole idea of working at home, and I wasn't sure doing 28.8 kbaud dial-in to the Sybase servers here was going to be bearable, even in the short-term. - Dan
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: MiscKit gains 3DKit, but what does that *mean*??? Date: 29 Mar 1996 18:04:22 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4jh8n6$3vl@news.xmission.com> References: <4iufkn$867@homer.alpha.net> My news feed is slow at times, so I didn't get this until just today. Here goes... gblock@homer.alpha.net (Greg Block) wrote: > Okay, so 3DKit is being added to the MiscKit. But will it be compiled for > new versions of NS/OpenStep? That's the plan. > Will it remain backwards compatible? Yes. > Will new features be added, or will it stagnate and die? I'd prefer the former. That's the goal. > Is this the death of 3DKit, > or a way of passing it to the vendors that rely on it? > What does this mean for apps based on the 3DKit? The latter is, again, our goal. This way, apps using the 3DKit should be able to keep on living. > Should I be writing a > 3DKit-replacement for my own use to allow 3DKit to die? No. GNU is working on a GNUSTEP alternative, though, so you might want to contribute to that. Someday it might be good enough to be a replacement to the NeXT kit and be better. If that ever happens, then we'd probably have everyone move to that kit and let the NeXT kit drop. But it would have to be a compatabile replacement for that to happen. Of course, you may wish to contribute to GNU's work on this... > If I wanted to work > on adding stuff to 3DKit, would that be possible? Yes. But you can't work on the MiscKit _and_ the GNU project. You'd have to pick one or the other for legal reasons. > Boy, I'm just a bottle of questions, aren't I. That seems to be normal, most people have plenty of questions. :-) I've got a FAQ on http://www.misckit.com/faq/NXKits.html that gives more detailed info on the answers above, so I suggest that any interested parties give it a look-see. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: Re: NEXT bundled/Sold Sybase Date: 29 Mar 1996 14:24:20 -0600 Organization: MCSNet Services Message-ID: <4jhgtk$9si@Mars.mcs.com> References: <199603291247.EAA03512@shellx.best.com> <9603291454.AA01774@copernicus> Dan Tapp (tappd@copernicus.sccsi.com) wrote: : I'm glad to hear that something like this may be available. I'm : considering putting a NEXT box out at the house, on which I would : need to do EOF work in order to justify the cost...a bit : problematic, since I don't have a Sybase instance sitting around the : living room. Ultimately, if the idea pans out, the office will get : single-channel ISDN for me, but I need to do some proof-of-concept : for the whole idea of working at home, and I wasn't sure doing 28.8 : kbaud dial-in to the Sybase servers here was going to be bearable, : even in the short-term. I don't think it should be that bad for a short term test. It all depends on how much data is needed back and forth. We have a client that has been working with a long distance 28.8 connection to our database for most of a year. It is not a system that I would want to run data intesive apps over for long, but for simple retrieval and updates or for a short test, it works fine. Peter Richardson
From: frank@cs.ucla.edu (Frank Fabbrocino) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXT and CORBA? Date: 29 Mar 1996 21:48:40 GMT Organization: UCLA Computer Science Dept. Message-ID: <4jhlro$6gp@delphi.cs.ucla.edu> References: <1996Mar29.191900.4514@il.us.swissbank.com> Keywords: CORBA, PDO, DO, DOLE, OLE sargent@ch1d305nwk (Jeff Sargent) writes: > Are any CORBA implementations available for NeXTSTEP? (Besides D'OLE, > which has an ORB, but isn't really based on CORBA). > How about just CORBA IDL compilers for Objective-C? > I've heard rumors here that NeXT is working on a CORBA compliant product. > Anyone know when such a thing might be available? The only thing I ever heard was that NeXT was going to provide a gateway between PDO and CORBA. I personally haven't heard anything about a native CORBA ORB for NeXTSTEP, unless you want to count OpenStep for Solaris which will have NEO. Later, Frank --------------------------------------------------------------------------- Frank Fabbrocino | OK, I admit it: My girlfriend's just an object University of California | to me. Unfortunately there is some information Los Angeles | hiding, but thankfully, she's fairly Data Mining Laboratory | encapsulated, nicely modular, and has a very frank@cs.ucla.edu | well defined interface! ;) ---------------------------------------------------------------------------
From: apuleius@ix.netcom.com(William Grosso) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXT and CORBA? Date: 30 Mar 1996 01:15:48 GMT Organization: Netcom Message-ID: <4ji204$frj@dfw-ixnews2.ix.netcom.com> References: <1996Mar29.191900.4514@il.us.swissbank.com> In <1996Mar29.191900.4514@il.us.swissbank.com> sargent@ch1d305nwk (Jeff Sargent) writes: > >Are any CORBA implementations available for NeXTSTEP? > No. There are two major free ORBs which you could probably get to compile, though. Electra is available from somebody at Cornell (whose name I have completely forgotten) and ILU is available from Xerox Parc. Cheers, Andy
Newsgroups: comp.sys.next.programmer From: shayman@Objectario.com (Steve Hayman) Subject: Re: OPENSTEP/NEXTSTEP Coursework in Illinois Message-ID: <1996Mar30.023655.6468@objectario.com> Sender: news@objectario.com Organization: Steve Hayman + Associates / NeXTSTEP Consulting / Toronto References: <RDL.96Mar28224542@world.std.com> Date: Sat, 30 Mar 1996 02:36:55 GMT In article <RDL.96Mar28224542@world.std.com> rdl@world.std.com (Robert La Ferla) writes: > Oh yeah... The NeXT office is in the NBC Tower building. Not any more - the Chicago office (and classroom) has moved to 311 S. Wacker, Suite 2250 (across the street from the Sears Tower.) For those of you keeping score, the New York City office has moved recently too. It's now at 1 Park Avenue. --- Steve Hayman Systems Engineer, NeXT Software Canada, Toronto Steve_Hayman@NeXT.com || (416) 769-8995
From: robert@justine.elastica.com (Robert Nicholson) Newsgroups: comp.sys.next.programmer Subject: NeXT's Sybase EOF adaptor against an MSQL server? Date: 23 Mar 1996 10:54:04 -0500 Organization: x Message-ID: <elokriz4z.fsf@justine.elastica.com> CC: eof@omnigroup.com Anybody have any experience doing the above? with Microsoft SQL V 6.0? -- "Under the circumstances I will sit down." (PGP key: send email with Subject: request pgp key)
From: james.robertson@wcom.com Newsgroups: comp.sys.next.programmer Subject: help: dual boot, swapping Win95 for DOS/Win3.1 Date: 30 Mar 1996 21:05:00 GMT Organization: WilTel Message-ID: <4jk7ls$b18@gateway.wiltel.com> I have a dual-boot Intel with NeXT in one partition and DOS/Win3.1 in the other. I want to swap in Win95 in place of DOS/Win3.1. (Actually, my NeXT partition is OpenStep/Mach/Intel PR2.) A few weeks back, I saw a posting (in CSNA?) on how to do this. As I recall, there were some tricks/pitfalls involved. I didn't save the posting, and can't find it in news archives on stepwise. NextAnswers says... "DOS 6.2 and Windows95 kindly removes our booter. You can replace the boot sector by following the instructions under "Replacing the NEXTSTEP Boot Sector" above. The instructions on replacing the boot sector say to log in as root in a terminal window. How can you do that if you can't boot NeXT? Do I boot from the installation floppy? I'd appreciate it if someone could repost or email me. - Jim Robertson LDDS WorldCom (WilTel) (713) 364-4133 james.robertson@wcom.com
From: nextjet@ids.net Newsgroups: comp.sys.next.programmer Subject: looking for new next color printer driver. Date: Sat, 30 Mar 96 18:12:55 +500 Organization: IDS World Network Internet Access Service, (800)IDS-1680 Message-ID: <4jkf5q$37t@paperboy.ids.net> does anyone remember were the site is that contains this driver.
From: Luke Howard <lukeh@auswired.net> Newsgroups: comp.sys.next.programmer Subject: Re: ORB palette or some such Date: 31 Mar 1996 03:41:45 GMT Organization: AusWired, Melbourne, Australia. Message-ID: <4jkutp$5v3@news.mel.aone.net.au> References: <4jf5mr$jq2@gidora.kralizec.net.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit gshaw@zeta.org.au (Greg Shaw) wrote: >A client of mine heard about a mechanism under NEXTSTEP for SPARC whereby >InterfaceBuilder could talk to an Object Request Broker (ORB) object. I >figured that this would have to be a palette of some description. It's an intriuging idea, but (a) what kind of ORB are you talking to? (under PDO the language is the ORB, save for locating an object by a string reference with nmserver) and (b) what kind of objects do you want to connect? >Has anyone heard about a palette that does this sort of thing on any >platform? Or another mechanism which would allow communication with an BTW -- what Nextstep development are you doing? It seems to be pretty rare in this country :-) -- Luke.
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: help: dual boot, swapping Win95 for DOS/Win3.1 Message-ID: <Dp4GxH.Dq@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4jk7ls$b18@gateway.wiltel.com> Date: Sun, 31 Mar 1996 08:00:52 GMT In article <4jk7ls$b18@gateway.wiltel.com> james.robertson@wcom.com writes: > The instructions on replacing the boot sector say to log in as root > in a terminal window. How can you do that if you can't boot NeXT? > Do I boot from the installation floppy? You don't need to. Use fdisk under DOS to set your NEXTSTEP partition as active. Now, when you reboot computer, you will boot into NEXTSTEP and you can restore NEXTSTEP boot sector as mentioned in NeXTanswers. Good luck, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: bbum@friday.com Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Netinfo support broken in Sendmail 8.7.x; patch enclosed Date: 31 Mar 1996 21:02:27 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4jmrt3$nqd@chinx4.thoughtport.net> The Netinfo support in sendmail 8.7.5 was *extremely* broken; - it *always* returned whatever value was found at the top-most domain. For example, if one used the /locations/sendmail netinfo directory to define the configuration file to be used by sendmail and one had a local-to-the-mailhost entry that specified /etc/sendmail/sendmail.mailhost.cf AND one had a generic, root-level entry that specified /etc/sendmail/sendmail.sharedsubsidiary.cf, the code would return the value encountered at the root level (as a result, there was no way to correctly configure netinfo such that sendmail would grok the correct configuration file from the machine's local domain). - The code leaked memory like a sieve; anytime information was found in more than one domain, the code would allocate seperate hunks of memory for the values contained in each domain. But it returns the pointer to the memory containing the value returned by the TOP most domain and would effectively 'forget' about all memory associated with the values found in the sub-domains. The enclosed patch fixes both problems; the value returned will always be the first encountered. Since the ni_propval() function starts it's search at the bottom-most (the local machine's) netinfo domain and searches up the hierarchy, the search stops at the first value encountered (which is returned), no memory is lost and the correct behaviour of being able to override global netinfo information with local netinfo configuration is preserved. The patch has been submitted to UCB. I will be making a quad-fat build of the patched sendmail available on ThoughtPort Authority's ftp site later tonight. The delay in posting to the ftp site is because I'm in the midst of adding a really useful feature; the ability to define a class (sendmail class, not object class) from the results of a netinfo lookup. The main purpose is to define the 'w' class (all of the hostnames for which the local machine is masquerading) from the contents of netinfo and/or the contents of a file. Be forewarned, the lookup semantics now supported by the ni_propval() class are the result of my beliefs derived from the exhibited behaviour, applied logic, and discussions with a NeXT employee. I honestly believe it is correct-- if your mileage comes up short, let me know so we can come to concensus, fix the problem, and make a new version available to the community. b.bum <bbum@friday.com> BTW: you should be able to apply this patch by uudecoding it and then 'patch -p < patchfile' from within the sendmail/src directory. The latest version of patch is available from ftp://ftp.thoughtport.net/pub/next/tools/. The patched sendmail binary will show up there later tonight [i'll make a posting here when it does; please be patient. As always, thanks to ThoughtPort Authority for providing the space and resources for my toolbox. :-) http://www.thoughtport.net/ ---- patch enclosed --- begin 644 sendmail-map-netinfo.patch M26YD97@Z(&UA<"YC"CT]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T] M/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T]/3T*4D-3(&9I;&4Z M("],;V-A;$1E=F5L;W!E<B]C=G,O;F5T+W-E;F1M86EL+W-R8R]M87`N8RQV M"G)E=')I979I;F<@<F5V:7-I;VX@,2XQ+C$N,0ID:69F("UR,2XQ+C$N,2!M M87`N8PHQ-3DR8S$U.3(*/"`)"BTM+0H^(`HQ-C0U8S$V-#4*/"`)"BTM+0H^ M(`HQ-S$W8S$W,3<*/"`)"0HM+2T*/B`*,3DQ.&,Q.3$X"CP@"0HM+2T*/B`* M,C`R-V,R,#(W"CP@"0HM+2T*/B`*,C$Y,6$R,3DR+#(Q.3,*/B`C(&1E9FEN M92!35%))3D=/4DY53$PH7W-?*0DH7W-?(#\@7W-?(#H@(CQ.54Q,/B(I"CX@ M"C(R,3!A,C(Q,RPR,C$W"CX@"6EF("AT5&0H-S$L(#$I*0H^(`D)<')I;G1F M*"(M+3X@;FE?<')O<'9A;"AD:7(Z("5S+"!P<F]P.B`E<RP@=F%L.B`E<RP@ M=F%L<')O<#H@)7,L('-E<&-H87(Z(#!X)7@I7&XB+`H^(`D)"2`@(%-44DE. M1T]23E5,3"AK97ED:7(I+"!35%))3D=/4DY53$PH:V5Y<')O<"DL(%-44DE. M1T]23E5,3"AK97EV86PI+`H^(`D)"2`@(%-44DE.1T]23E5,3"AV86QP<F]P M*2P@<V5P8VAA<BD["CX@"C(R-CAC,C(W-2PR,C<W"CP@"BTM+0H^(`D):68@ M*'149"@W,2P@-2DI"CX@"0D)<')I;G1F*"(M+7P@;FE?<')O<'9A;#H@:6YV M;VMI;F<@;FE?<&%T:'-E87)C:"@I('=I=&@@:V5Y(&]F('LE<WU<;B(L"CX@ M"0D)"2`@(&ME>6)U9BD["C(R.#)D,C(Y,`H\(`HR,C@V8S(R.30L,C,P-`H\ M(`D):68@*'-E<&-H87(@/3T@)UPP)R`F)B!N:6YL+FYI7VYA;65L:7-T7VQE M;B`^(#$I"BTM+0H^("`@("`@("`@+RH*/B`@("`@("`@("HJ($EF('1H92!N M86UE;&ES="!D;V5S;B=T(&-O;G1A:6X@86YY(&5N=')I97,L(&]R(&EF(&UU M;'1I<&QE"CX@("`@("`@("`J*B!E;G1R:65S('=E<F4@96YC;W5N=&5R960@ M=VAE;B!T:&4@<V5P97)A=&]R(&-H87)A8W1E<B!W87,@)UPP)RP*/B`@("`@ M("`@("HJ('1H96X@8VQE86X@=7`@86YD(&-O;G1I;G5E+@H^("`@("`@("`@ M*BH*/B`@("`@("`@("HJ($YO=&4Z("!3:&]U;&0@=V4@4D5!3$Q9(&-O;G1I M;G5E(&%S('1H92!R97-U;'0@;V8@82!S97`@;V8@)UPP)PH^("`@("`@("`@ M*BH@=VET:"!M=6QT:7!L92!E;G1R:65S.R`@;W(@<VAO=6QD('=E(&IU<W0@ M=&%K92!T:&4@9FER<W0@96YT<GD*/B`@("`@("`@("HJ(&%N9"!R971U<FX@ M:70_/S\_"CX@("`@("`@("`J*@H^("`@("`@("`@*B\*/B`)"6EF("@H;FEN M;"YN:5]N86UE;&ES=%]L96X@/3T@,"D@?'P@("@H<V5P8VAA<B`]/2`G7#`G M*2`F)B`H;FEN;"YN:5]N86UE;&ES=%]L96X@/B`Q*2DI"C(R.35D,C,Q,@H\ M(`HR,CDY83(S,3<L,C,Q.`H^(`D):68@*'149"@W,2P@.2DI"CX@"0D)<')I M;G1F*"(M+7P@;FE?<')O<'9A;#H@86QL;V-A=&EN9R!B=69F97(@;V8@>R5D M?2!B>71E<UQN(BP@86QE;BD["C(S,#AA,C,R."PR,S,T"CX@"CX@("`@("`@ M("`O*@H^("`@("`@("`@*BH@268@<')O<'9A;"!I<R!D969I;F5D("AI9B!T M:&4@;&]O:W5P('-U8V-E<W-F=6QL>2!F;W5N9"!A('9A;'5E*0H^("`@("`@ M("`@*BH@8G)E86L@86YD(')E='5R;BX*/B`@("`@("`@("HO"CX@"0EI9B`H M<')O<'9A;"D*/B`)"0EB<F5A:SL*,C,Q.&$R,S0U+#(S-#@*/B`*/B`):68@ M*'149"@W,2P@,2DI"CX@"0EP<FEN=&8H(CPM+2!N:5]P<F]P=F%L*"D@<F5T M=7)N:6YG("5S7&XB+`H^(`D)"2`@(%-44DE.1T]23E5,3"AP<F]P=F%L*2D[ !"G)N ` end
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Fab: Objective C Fabricator Message-ID: <Dp5Ip7.IAu@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam Date: Sun, 31 Mar 1996 21:36:42 GMT O B J E C T I V E - C F A B R I C A T O R David Stes Computer Algebra Objects E-mail: stes@can.nl WWW: http://www.can.nl/~stes Fab is a Unix filter for producing Objective C class libraries; the source code of the Computer Algebra Kit, a set of Objective C objects for arbitrary precision arithmetic and computer algebra, is written in "fab". Fab extracts private and public header, source and documentation files from files with a ".fab" extension. Given a file named "MyObject.fab", fab deposits, depending on the options being used, files called "MyObject.h", "MyObject.m", "MyObject.rtf", "My Object.3", "MyObject.html" etc. in the specified directories. The most important feature of using "fab" is the ability to produce public and private header files. Private header files can be used by the developers of a class library; users include the public headers in their source files. This encourages hiding implementation details from the user, and exporting a "strong" interface. Availability ------------ Fab is written in Objective C, and compiles on NextStep, on Stepstone and GNU Objective C, and has been ported to several Unix platforms. FeedBack -------- "Fab" is public domain software. If you find bugs, add features, port to new platforms, improve in any other way or have suggestions, please send e-mail and the necessary changes can perhaps be incorporated into future versions of "Fab", available from the Computer Algebra Objects web site, currently at : "http://www.can.nl/~stes". Disclaimer ---------- I'll try to fix bugs in "fab", and to continue to maintain the software, but you agree to use "fab" on an "as-is" basis.
Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Netinfo support broken in Sendmail 8.7.x; patch enclosed In-Reply-To: bbum@friday.com's message of 31 Mar 1996 21:02:27 GMT Message-ID: <RDL.96Mar31215404@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4jmrt3$nqd@chinx4.thoughtport.net> Date: Mon, 1 Apr 1996 02:54:04 GMT Before you go through all that trouble and create a second version, let me say that my sendmail 8.7.5 distribution DOES NOT contain the first bug that you mentioned. I have to check to see if there are memory leaks but it's not as alarming as you've made it out to be. I believe you are referring to the Berkeley source distribution which does not contain the NetInfo fixes that my binary distribution has. Eric Allman has been aware of it for some time but won't be fixing it until sendmail 8.8. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 1 Apr 1996 05:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4jnoou$2or@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Netinfo support broken in Sendmail 8.7.x; patch enclosed In-Reply-To: rdl@world.std.com's message of Mon, 1 Apr 1996 02:54:04 GMT Message-ID: <RDL.96Mar31231821@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4jmrt3$nqd@chinx4.thoughtport.net> <RDL.96Mar31215404@world.std.com> Date: Mon, 1 Apr 1996 04:18:21 GMT I think I've found another source of confusion. Berkeley sendmail 8.7.5 treats aliases with MULTIPLE members differently than NeXT's. With NeXT's implementation, you can specify each name as a different value whereas with Berkeley, it's one value which is a comma separated list. e.g. NeXT: key=name value=mygroup key=members value=myself value=another value=athird Berkeley: key=name value=mygroup key=members value=myself,another,athird For compatibility, we should have both so that needs to be changed. I will make a release this week that contains this change and patches to any memory leaks. Bill, you should mail me your modifications and I'll incorporate them (and credit you) for them. What I don't want to happen is to have multiple versions maintained by multiple people. It's bad enough that I have to have a version that is slightly different from Berkeleys. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: sargent@ch1d305nwk (Jeff Sargent) Subject: NeXT and CORBA? Message-ID: <1996Mar29.191900.4514@il.us.swissbank.com> Keywords: CORBA, PDO, DO, DOLE, OLE Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Fri, 29 Mar 1996 19:19:00 GMT Are any CORBA implementations available for NeXTSTEP? (Besides D'OLE, which has an ORB, but isn't really based on CORBA). How about just CORBA IDL compilers for Objective-C? I've heard rumors here that NeXT is working on a CORBA compliant product. Anyone know when such a thing might be available? Thanks - - Jeff
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: NSCharacterSet problem (long) Message-ID: <Dp6JA1.10B@hurka.UUCP> Keywords: NSCharacterSet foundation Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) Date: Mon, 1 Apr 1996 10:46:49 GMT Hi All, does anybody know, why the following code does not work? I suspect that this is a bug in the Foundation library, but before I post it to the bug_next, I want to know your opinion. Here is the code: ---------------------------------------------------------------------- // cc -Wall -g smazat.m -lFoundation_s ; a.out #import <foundation/foundation.h> #import <stdio.h> #define SHOW printf("o=%x,%d i=%x,%d\n",(int)operators,[operators retainCount],(int)illegal,[illegal retainCount]); void main() { NSCharacterSet *operators=nil; NSMutableCharacterSet *illegal=nil; NSAutoreleasePool *pool=[[NSAutoreleasePool alloc] init]; [NSAutoreleasePool enableDoubleReleaseCheck:YES]; [NSAutoreleasePool _enableClassNameLogging:YES]; operators=[NSCharacterSet characterSetWithCharactersInString:@"+-*/<>=,|&"]; SHOW [operators retain]; SHOW illegal=(NSMutableCharacterSet*)[NSCharacterSet alphanumericCharacterSet]; SHOW illegal=[illegal mutableCopy]; SHOW illegal=[illegal retain]; SHOW illegal=[illegal retain]; SHOW illegal=[illegal retain]; SHOW [illegal addCharactersInString:@"."]; SHOW [illegal formUnionWithCharacterSet:operators]; SHOW [pool release]; SHOW } ---------------------------------------------------------------------- And here is the result: ---------------------------------------------------------------------- /Users/tom/test> a.out Apr 1 12:38:14 [1283] Double free check enabled; break on _NSDoubleFreeError Apr 1 12:38:14 [1283] autorelease called for 45384 of NSImmutableStringCharacterSet o=45384,1 i=0,0 o=45384,2 i=0,0 o=45384,2 i=447c8,-1 Apr 1 12:38:14 [1283] autorelease called for 4496c of NSCStringWithGap Apr 1 12:38:14 [1283] autorelease called for 4496c of NSCStringWithGap Apr 1 12:38:14 [1283] autorelease called for 4495c of NSAutoreleasedMallocMemory Apr 1 12:38:14 [1283] autorelease called for 47c54 of NSConcreteMutableData Apr 1 12:38:14 [1283] autorelease called for 47c54 of NSConcreteMutableData Apr 1 12:38:14 [1283] autorelease called for 47bb8 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 45434 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4dd28 of NSConcreteMutableArray o=45384,2 i=454f8,-1 Apr 1 12:38:15 [1283] autorelease called for 4bcd4 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4bce4 of NSConcreteMutableArray Apr 1 12:38:15 [1283] autorelease called for 4ddb4 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4ddd0 of NSConcreteMutableArray o=45384,2 i=454f8,-1 Apr 1 12:38:15 [1283] autorelease called for 4ddec of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4ddfc of NSConcreteMutableArray Apr 1 12:38:15 [1283] autorelease called for 4de10 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4de20 of NSConcreteMutableArray o=45384,2 i=454f8,-1 Apr 1 12:38:15 [1283] autorelease called for 4dec0 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4ded0 of NSConcreteMutableArray Apr 1 12:38:15 [1283] autorelease called for 4dee4 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4def4 of NSConcreteMutableArray o=45384,2 i=454f8,-1 Apr 1 12:38:15 [1283] autorelease called for 4df10 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4df20 of NSConcreteMutableArray o=45384,2 i=454f8,-1 Apr 1 12:38:15 [1283] autorelease called for 4df74 of NSUnicodeString Apr 1 12:38:15 [1283] autorelease called for 4df88 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4dd70 of NSConcreteMutableCharacterSet Apr 1 12:38:15 [1283] autorelease called for 4dcfc of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4dd0c of NSConcreteData Apr 1 12:38:15 [1283] autorelease called for 44590 of NSConcreteData Apr 1 12:38:15 [1283] autorelease called for 475e8 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 4dd70 of NSConcreteMutableCharacterSet Apr 1 12:38:15 [1283] autorelease called for 52058 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 54218 of NSCStringWithGap Apr 1 12:38:15 [1283] *** _NSDoubleFreeError: autorelease called 2 times, but object's retainCount is only 1, for object: <NSConcreteMutableCharacterSet: 0x4dd70> Apr 1 12:38:15 [1283] autorelease called for 52068 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 541d8 of NSConcreteMutableArray o=45384,2 i=454f8,-1 Apr 1 12:38:15 [1283] autorelease called for 44a30 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 44a30 of _NSDictionaryEnumerator Apr 1 12:38:15 [1283] autorelease called for 47850 of NSConcreteMutableArray o=45384,1 i=454f8,-1 /Users/tom/test> ---------------------------------------------------------------------- What surprise me is the retainCount == -1. I can tolerate it in [NSCharacterSet alphanumericCharacterSet], but it stays there after mutableCopy, which is definitely not correct (IMHO). Moreover retain does not change the retainCount at all. Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.soft-sys.nextstep,comp.sys.mac.advocacy,comp.sys.mac.programmer.tools,comp.sys.mac.programmer.misc,comp.sys.mac.oop.misc,comp.unix.advocacy,comp.sys.powerpc,comp.software-eng,comp.programming,comp.os.linux.advocacy,comp.sys.linux.development.system,isu.isunix Subject: OpenDoc with Corba and OPENSTEP/NEXTSTEP licensing... Date: Mon, 01 Apr 1996 00:29:23 -0600 Organization: Illinois State University, Instructional Technology Services Distribution: inet Message-ID: <315F77C3.4AC9@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Well now that OMG is integrating Corba with OpenDoc, wouldn't it make sense for NeXT to integrate OpenDoc? Also Sun has OpenDoc as being the #1 most requested feature for Java... NeXT is supposed to be integrating Corba connectivity, so perhaps OpenDoc is inevitable. Then all we need is OPENSTEP on MAC OS Power PC, but that will probably come from GNU rather than NeXT. (It will probably be ported to PowerPC Linux too) It seems the name of the game is cooperation and interoperability, not dominance with a single OS or set of tools. Companies should realize this (NeXT are you listening?) The more people can interoperate (make connections) with software they have, they will be more prone to using those tools. I'm trying to get my two object-oriented professors at my university (Dr. Bee Lee Lim mailto:bllim@rs6000.cmp.ilstu.edu and Dr. Tibor Gyires mailto:tbgyires@rs6000.cmp.ilstu.edu) to take a look at what OPENSTEP/NEXTSTEP can do for the software world, and I'm having a dificult time convincing them that this is worthwhile as course curiculum. Dr. Gyires is interested in SmallTalk (which I understand is VERY similar to Objective-C) and has watched 2 NeXT videos (one on NS and another on WebObjects and D'OLE.), and I found that Dr. Lim is only interested in OPENSTEP and its relation to Java after talking to him, his concern is that much of the software industry and companies DON'T USEOPENSTEP/NEXTSTEP. I need assistance in getting this institution (Illinois State University) to adopt some OPENSTEP curiculum because I feel these development tools are great, and they get me excited about software. If anyone has any valid arguments I can make, please email me or the two professors. Thanks all ------------------------------------------------ Eric A. Dubiel; http://www.ilstu.edu/~eadubie Instructional mailto:eadubie@rs6000.cmp.ilstu.edu Technology PEACE LOVE Services UNITY RESPECT Illinois State University "Understanding is best learned via experience." "Follow our instinct, not a trend. Go against the grain until the end." VIEWS EXPRESSED ARE ENTIRELY MY OWN
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 1 Apr 1996 12:26:12 GMT Organization: Norden 1 Communications Message-ID: <4joi14$ku2@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF--A Plus Sybase or Oracle exceptional opportunity Full benefits US Citizen or Greencard To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: msm@pswtech.com (Mark Mertel) Newsgroups: comp.sys.next.programmer Subject: Re: NEXT bundled/Sold Sybase Date: 1 Apr 1996 12:55:07 GMT Organization: PSW Technologies Message-ID: <4jojnb$ern@digdug.pswtech.com> References: <199603291247.EAA03512@shellx.best.com> <9603291454.AA01774@copernicus> Dan Tapp <tappd@copernicus.sccsi.com> wrote: > >I'm glad to hear that something like this may be available. I'm >considering putting a NEXT box out at the house, on which I would >need to do EOF work in order to justify the cost...a bit >problematic, since I don't have a Sybase instance sitting around the >living room. > >- Dan i've been doing EOF development at home for over a year. i've got a slab with sybase installed, and access the system 10 server over 1 14.4 or 28.8 modem when i need it no problem. -- Mark Mertel Email: msm@pswtech.com PSW Technologies Phone: (512) 343-6666 9050 Capital of Texas Hwy. N. FAX: (512) 343-9650 Austin, TX 78759 Home: (512) 847-2061
From: rseymour@rseymour.com (Robert J Seymour) Newsgroups: comp.sys.next.programmer,comp.lang.perl.misc Subject: Re: [Q] Fat perl5.002 for NextStep3.3 Patch1? Date: Mon, 01 Apr 1996 14:57:23 GMT Organization: MCSNet Internet Services Message-ID: <4joqug$bom@Nntp1.mcs.net> References: <wp68bk45vm.fsf@frsx01.fr.bosch.de> r_bock@frsx01.fr.bosch.de (Rene Bock) writes in comp.lang.perl.misc: > Doses is exists somewhere? Or can somone give a hint on how to build it? The simplest mechanism is to conjoin the binaries for the individual architectures using the lipo tool. I also have a patch (the basis of which was generously provided by another perl porter whose name escapes me at the moment) that I'll send along to you when I'm next on my NeXT. I've been meaning to fix this up and submit it to the porters ... hopefully this will be set up for the next release (though there are a few issues that need to be addressed first). Regards, Robert -- Robert J Seymour rseymour@rseymour.com <URL:http://rseymour.com/>
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT's Sybase EOF adaptor against an MSQL server? Date: 1 Apr 1996 19:00:53 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4jp955$4fp@stc06.ctd.ornl.gov> References: <elokriz4z.fsf@justine.elastica.com> In article <elokriz4z.fsf@justine.elastica.com> robert@justine.elastica.com (Robert Nicholson) writes: > Anybody have any experience doing the above? with Microsoft SQL V 6.0? There is a MiniSQLEOFAdaptor available on the web. I've installed both msql and the adaptor, changed sybase .eomodel files and redone the examples using Sybase using msql. It seems to work fine and is only limited by the limitations of msql. -- - - - - - - - - - J. W. Wooten
From: hhoff@cube.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer,comp.lang.perl.misc Subject: Re: [Q] Fat perl5.002 for NextStep3.3 Patch1? Date: 1 Apr 1996 17:48:48 GMT Organization: Cube Informationssysteme GmbH Message-ID: <4jp4u0$lcp@cubenx.cube.de> References: <wp68bk45vm.fsf@frsx01.fr.bosch.de> <4joqug$bom@Nntp1.mcs.net> Robert J Seymour wrote: > r_bock@frsx01.fr.bosch.de (Rene Bock) writes in comp.lang.perl.misc: > > Doses is exists somewhere? Or can somone give a hint on how to build it? > > The simplest mechanism is to conjoin the binaries for the individual > architectures using the lipo tool. I also have a patch (the basis of > which was generously provided by another perl porter whose name > escapes me at the moment) that I'll send along to you when I'm next on > my NeXT. I've been meaning to fix this up and submit it to the > porters ... hopefully this will be set up for the next release (though > there are a few issues that need to be addressed first). An even simpler mechanism is a simple download of the the precompiled quad-fat installer package, done by Pedja B. of TipTop Software. :-) I've seen (and downloaded) this pgk on next-ftp.peak.org, and it should be on the mirrors too. Thanks Pedja! Holger -- ___ /\__\ Holger Hoffstätte net: hhoff@cube.de (MIME, NeXT, PGP) \/__/ Cube Informationssysteme GmbH vox: (+49)0711-90669-0 (or -11) [kju:b] Stuttgart, Germany fax: (+49)0711-90669-16
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT's Sybase EOF adaptor against an MSQL server? Date: 1 Apr 1996 22:00:48 GMT Organization: Digital Fix Development Message-ID: <4jpjmg$967@digifix.digifix.com> References: <elokriz4z.fsf@justine.elastica.com> <4jp955$4fp@stc06.ctd.ornl.gov> In-Reply-To: <4jp955$4fp@stc06.ctd.ornl.gov> On 04/01/96, John W. Wooten wrote: >In article <elokriz4z.fsf@justine.elastica.com> robert@justine.elastica.com >(Robert Nicholson) writes: >> Anybody have any experience doing the above? with Microsoft SQL V 6.0? >There is a MiniSQLEOFAdaptor available on the web. I've installed both msql >and the adaptor, changed sybase .eomodel files and redone the examples using >Sybase using msql. It seems to work fine and is only limited by the >limitations of msql. There needs to be some clarification here... Robert was asking about Microsoft SQL, but put MSQL in the subject. The msql adaptor is for the msql database is an entirely different beast. -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
From: franke@boehme.dbag.ulm.DaimlerBenz.COM (Juergen Franke) Newsgroups: comp.sys.next.programmer Subject: Arabic Character Code Palette Date: 1 Apr 1996 09:40:13 GMT Organization: debis Network Services GmbH Message-ID: <4jo89t$ff7@news.sns-felb.debis.de> How and where do I get an arabic charcter code palette for Keyboard.app? -- Jörgen Franke Text Understanding Department F3M/T Daimler-Benz AG, Research Center Ulm P.O. Box 2360 89013 Ulm Germany Telephone Germany 731 505 2355 Fax Germany 731 505 4113 e-mail: franke@dbag.ulm.DaimlerBenz.COM
From: mo@spielberg.unice.fr (Mathieu Olivier) Newsgroups: comp.sys.next.programmer Subject: renderman on nextstep Date: 2 Apr 1996 13:39:14 GMT Organization: University of Nice Sophia-Antipolis Message-ID: <4jram3$b18@malibu.unice.fr> Keywords: 3DKit Hi All, does anybody know, where i can find "The RenderMan Companion" ??? it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) Best regards,
From: jessica@swamp.chem.wisc.edu (Jessica Severin) Newsgroups: comp.sys.next.programmer Subject: Problem linking in posix library Date: 2 Apr 1996 19:18:44 GMT Organization: University of Wisconsin, Madison Message-ID: <4jruik$1oiu@news.doit.wisc.edu> I'm trying to write a Nextstep 3.3 application using the Mathematica 2.2 MathLink libraries. The problem I'm running into is that the MathLink libraries use bsd sockets to communicate with the Mathematica application, and these functions are not part of the NeXT_s library. They are however part of libposix.a. But when I include the posix library into my project I get a series of link errors like: ld: multiple definitions of symbol _strrchr /usr/lib/libposix.a(strrchr.o) definition of _strrchr in section(__TEXT,__text) /lib/libsys_s.a(strrchr.o) definition of absolute _strrchr (value 0x5002dab) I even ftped the new libposix.a from next and it produced the same errors. I need the following functions (all of which can be found in libposix.a and nowhere else): _accept _bsd_printf _bzero _dup _freopen _htonl _htons _isatty _kill _listen _memcpy _ntohl _ntohs _recv _send _setsockopt If you do a > otool -L /usr/lib/lib*.a | grep setsocketopt the response is /usr/lib/libposix.a(setsockopt.o): How do I link in both NeXT_s and posix? thanks --- Jessica Severin-Hayden jessica@swamp.chem.wisc.edu (608) 262-0296 Programmer / Analyst Dr. Lloyd Smith's Lab Dept Chemistry University of Wisconsin - Madison
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Tue, 02 Apr 1996 20:00:00 GMT Organization: Orbital Computer Consultancy Message-ID: <316186b9.4262367@news.plsys.co.uk> References: <Dp6nnF.2xB@cwi.nl> <4jqv4t$d20@sun0.urz.uni-heidelberg.de> tuparev@mailserver.EMBL-Heidelberg.DE (Georg Tuparev) wrote: >In article <Dp6nnF.2xB@cwi.nl> stes@cwi.nl (David Stes) writes: >> In article <Dp5nt2.6vA@novice.uwaterloo.ca> dfevans@bbcr.uwaterloo.ca (David >Evans) writes: >> It's natural that in such circumstances NeXT's interest, and the interests >> of the supporters, partners, customers etc. diverge. > >You mean *your* and NeXT's interest diverge. There are a lot of people here >happy with the new development, and most of us do not need backwards >compatibility a la M$. But you already know that I'm sick of listening your >anti-Foundation complains... > I agree that the FoundationKit is nice to develop with, but can you, or anyone, give me a really good technical reason why its functionality could not have been fitted within the old Object hierarchy? When I consider the possible reasons, I find myself able to discount them. 1) The memory management? Release, retain and autorelease could easily have been added to Object; after all, the retain counts are externally maintained. Dealloc could have remained free ; so what that its name pairs better with alloc? 2) The more consistent method names? The new methods could surely have been added to Object (or other classes as appropriate), and obsolete _methods_ phased out, instead of a whole class hierarchy 3) The NS class name prefix? An advantage surely outweighed by the dual hierarchy / need for conversion scripts. I'd rather have Objective-C enhanced to support namespaces. Mind you, proper support for private methods would be a nice start. What exactly _have_ Next done with the language itself since they acquired the rights? 4) NSDictionary / NSArray / NSEnumerator ? Again, surely HashTable and List could simply have been enhanced? And NSDictionary requires all keys to be NSStrings, unlike the more general HashTable. 5) The mutability concepts? Okay, NSImmutableList:List would violate the substitutability principle, but again I think this is probably worth it to prevent disposing of an entire existing class hierarchy. And anyway, it still doesn't explain why you couldn't have NSMutableArray:NSArray:Object 6) EOF? EOFault is a root class anyway, and Object already supports EOKeyValueCoding via a category. 7) DO? Don't see why. Again, NXProxy is a root class anyway. 8) The "new" encoding scheme? Yet again, why not just add it to Object. 9) Something else......? Just to make it clear what I'm saying, I think Next's decision to clean the API up for OpenStep was a good one in itself, but this could have been done in a more incremental fashion. Perhaps I've been "Living in a Hybrid World" for too long..... ;-)
From: Paul R. Brown <pbrown@math.berkeley.edu> Newsgroups: comp.sys.next.programmer Subject: Re: renderman on nextstep Date: 2 Apr 1996 20:20:13 GMT Organization: University of California, Berkeley Message-ID: <4js25t$3ee@agate.berkeley.edu> References: <4jram3$b18@malibu.unice.fr> mo@spielberg.unice.fr (Mathieu Olivier) wrote: >does anybody know, where i can find "The RenderMan Companion" ??? >it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) I bought a copy from Powell's Technical Books in Portland, OR. If you can't find it in France, I'm sure that they could probably send you a copy. (Just so you know, Powell's is one of the largest (single) bookstores in the U.S; they have most everything. Last time I was there, they had full sets of NeXT technical docs, for example.) Paul
From: steel456@aol.com (Steel456) Newsgroups: comp.sys.next.programmer Subject: NeXT Developers and Architects Date: 2 Apr 1996 15:00:12 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4js10c$oqm@newsbf02.news.aol.com> We are the Premier search and recruiting firm for NeXT Professionals in the USA. We have a number of exclusive clients who have retained our services to search for the best NeXSTEP, Objective-C and strong OO Architects on the market.We are now recruiting for positions from coast to coast from Developers to Architects. If you have experience developing within the NeXT environment and are considering looking to leverage your career, now is the time and we are the firm to talk to. Please call us at: Corporate Resource Associates Bruce Rennert 800-662-9797 Fax 800-814-3100
Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) In-Reply-To: k_harbour@mail.bogo.co.uk's message of Tue, 02 Apr 1996 20:00:00 GMT Message-ID: <RDL.96Apr2221809@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <Dp6nnF.2xB@cwi.nl> <4jqv4t$d20@sun0.urz.uni-heidelberg.de> <316186b9.4262367@news.plsys.co.uk> Date: Wed, 3 Apr 1996 03:18:09 GMT How about binary compatibility? What about archiving? Making major changes to base classes is not exactly a low-risk proposition. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Netinfo support broken in Sendmail 8.7.x; patch enclosed In-Reply-To: bbum@friday.com's message of 31 Mar 1996 21:02:27 GMT Message-ID: <RDL.96Apr2222247@world.std.com> Sender: rdl@world.std.com (Robert La Ferla) Organization: The World Public Access UNIX, Brookline, MA References: <4jmrt3$nqd@chinx4.thoughtport.net> Date: Wed, 3 Apr 1996 03:22:47 GMT Looks like my followup to this article didn't get crossposted here. Not that I think this is the place for these discussions anyhow. But in a nutshell, my binary distribution of sendmail does NOT contain these problems. He is referring to the Berkeley source code distribution which does NOT contain a critical NetInfo bug fix. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: schultz@darkwing.uoregon.edu (Stewart Schultz) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: How to change finger display? Date: 2 Apr 1996 14:42:00 -0800 Organization: University of Oregon Message-ID: <4jsafo$7tm@darkwing.uoregon.edu> Is there a simple way to change the NeXT finger command so that it returns only the .plan contents, and no login information? If a fingerd is available, perhaps one could simply delete the portion that looks in a system file for login info? Thanks, -SS
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: renderman on nextstep Date: 3 Apr 1996 05:49:18 GMT Organization: Nanyang Technological University Message-ID: <4jt3gu$gdr@ntuix.ntu.ac.sg> References: <4jram3$b18@malibu.unice.fr> <4js25t$3ee@agate.berkeley.edu> Paul R. Brown (pbrown@math.berkeley.edu) wrote: : mo@spielberg.unice.fr (Mathieu Olivier) wrote: : >does anybody know, where i can find "The RenderMan Companion" ??? : >it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) : I bought a copy from Powell's Technical Books in Portland, OR. If you can't : find it in France, I'm sure that they could probably send you a copy. (Just : so you know, Powell's is one of the largest (single) bookstores in the U.S; : they have most everything. Last time I was there, they had full sets of NeXT : technical docs, for example.) Any bookshop should be able to order it for you. You just need the ISBN. Which I don't have at hand now, (it is at home), but if nobody else posts/mails it, give me an e-mail and I'll look it up. Willem gwillem@ntuvax.ntu.ac.sg
Newsgroups: comp.sys.next.programmer,comp.lang.oberon,comp.object.logic,comp.lang.scheme,comp.lang.beta From: Bob Hathaway <rjh@geodesic.com> Subject: Comp.Object FAQ Version 1.0.9 (04-02) Announcement Message-ID: <Dp9r86.C4p@midway.uchicago.edu> Followup-To: comp.object Summary: Frequently Asked Questions (FAQ) List and Available Systems For Object-Oriented Technology Sender: news@midway.uchicago.edu (News Administrator) Organization: Geodesic Systems, Inc. Date: Wed, 3 Apr 1996 04:31:18 GMT [This is a copy of an archive messsage to your newsgroup. Well, my new 10 minute "cut" program seems to work, so take a look! The OOFAQ is posted to the comp.object Newsgroup and is on the news and comp.answer groups too.] This announces the April version 1.0.9 of the Comp.Object FAQ! It has many new questions, entries, and updates and has several very up-to-date appendices on object-oriented methodologies and systems, with a new Appendix on commercial Object-Oriented Libraries and Systems. This document may very well comprise the latest and most up-to-date collection of object-oriented systems and terms in the world today! I'd also like to pay attention to a related new free hypertext journal on Object Technology called Object Currents. There are many references to the comp.object FAQ and it provides a great opportunity to keep up with Object Technology and contribute as authors! There are many world-class columnists and authors with unique contributions *as well as* many of the best articles from the very best of the OO Journals placed on the Web. So please take a look at: http://www.sigs.com/objectcurrents The first three interviews are with Grady Booch, James Rumbaugh, and Ivar Jacobson and a Steve Mellor interview is in the works. Robert Martin has been providing a Newsgroup Dialog and columnists include Watts Humphrey, Bertrand Meyer, Michael Chonoles, Francois Bancilhon, David Shang, and Michael Spertus along with an editorial by yours truly. This last month's Editor's Corner is on Metasystems and Reflection, so check it out! The April Issue went up on 4/1. >I will try to update the FAQ monthly with system entries and information from >the net. I have a new TQM system in place which should proceed from the >old Level 1 to Level 5 in no time - not hard for a one man job. *Many* >thanks to the patience of those sending system updates and corrections, >my new system should provide fast acknowledgement, inclusion (within a few >weeks) and tracking, and 6 sigma quality should provide essentially defect >free performance for the FAQ. Perhaps I'll even go for ISO 9000 certification! Well, now that that system has been shelved for 1 year its about time to update it! A great example of a good OOA/D system with no time for implementation. But that should change very soon! It's the volunteer efforts that suffer the most when in a (prolonged) time crunch, but with a new book coming out that should change. And with a new 4GL hypertext system in place it will be ***much*** easier. Sending comments, suggestions, additions, updates, corrections, and new systems and site entries is strongly encouraged and should be directed to the author at rjh@geodesic.com. The FAQ is posted to the comp.object, comp.answers and news.answers newsgroups and is available via anonymous ftp from zaphod.uchicago.edu and rtfm.mit.edu, although new versions may take a short period of time to be installed. Anonymous FTP Sites and Hypertext Server: anonymous@zaphod.uchicago.edu:/pub/CompObj8.faq(.Z) (128.135.72.61) anonymous@rtfm.mit.edu:/pub/usenet/comp.object/*_Part_* (18.181.0.24 Tmp) http://iamwww.unibe.ch/~scg/OOinfo/FAQ/index.html (new IAM location) Mail Server: (See also section 1.24) mail mail-server@rtfm.mit.edu Subject: send usenet/comp.object/* Zaphod is preferred over rtfm for anonymous ftp retrieval, as it provides a single file. Rtfm contains the FAQ as posted. To use the hypertext system, see APPENDIX E, entries 27. Comp.Object Archive: A workable comp.object archive is now available on the www, with much thanks to Markus A. Beckmann, beckmann@informatik.mathematik.uni-mainz.de. http://aaimzb.mathematik.uni-mainz.de/Personal/Mitarbeiter/comp.object.idx.html Again, a short period of time may be required to retrieve the latest version from anonymous ftp, allowing the posted version to propagate and get out as quickly as possible. Thank you to the many people who have contributed their time and effort to help this document spread the word about object-oriented technology and available systems! It is hoped it will be most useful in that endeavor. Best Regards! Bob
From: dirk@kalium.physik.TU-Berlin.DE (Dirk Schwarzhans) Newsgroups: comp.sys.next.programmer Subject: Re: renderman on nextstep Date: 3 Apr 1996 08:19:40 GMT Organization: Technical University Berlin, Germany Message-ID: <4jtcas$t53@brachio.zrz.TU-Berlin.DE> References: <4jram3$b18@malibu.unice.fr> In-Reply-To: <4jram3$b18@malibu.unice.fr> Hello, On 04/02/96, Mathieu Olivier wrote: > >Hi All, > >does anybody know, where i can find "The RenderMan Companion" ??? >it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) NeXT's front end to RenderMan, the 3D Kit, is not the topic of "The RenderMan Companion"! It "only" covers the low level RenderMan language (as far as I know). Dirk Schwarzhans
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: gcc 2.7.x for hppa or sparc Date: 30 Mar 1996 18:57:23 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4jk06j$nqd@chinx4.thoughtport.net> Has anyone done the work necessary to support gcc under hppa or sparc? If someone is kind enough to send me the patches necessary to make hppa and/or sparc support possible [in any form], I would be happy to clean them up and both submit them back to the GNU folks for [possible] integration as well as releasing a canonical patch for the NeXT community. thanks, b.bum
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: Pb while loading a nib? Date: 3 Apr 1996 17:39:42 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4jud4u$jld@ai.alienor.fr> Pb while loading a nib? I'm using MiscKit1.2.8 and I'd like to load the Unix-HatersTool's nib. But the result is : "Unable to load palette . Use /LocalDeveloper/Palettes instead?" I say OK (why not?) and : "Unable to load palette /LocalDeveloper/Palettes." Finally : "Cannot open file /LocalDeveloper/Examples/Apps/UNIX-HatersTool.app/Engl ish.lproj/UNIX-HatersTool.nib." -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: pbrown@asparagus.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: renderman on nextstep Date: 3 Apr 1996 07:20:13 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4jt8rd$lt5@agate.berkeley.edu> References: <4jram3$b18@malibu.unice.fr> <4js25t$3ee@agate.berkeley.edu> <4jt3gu$gdr@ntuix.ntu.ac.sg> In article <4jt3gu$gdr@ntuix.ntu.ac.sg>, Van Schaik Willem Anthon Johan <GWILLEM@alpha.ntu.ac.sg> wrote: >Paul R. Brown (pbrown@math.berkeley.edu) wrote: >: mo@spielberg.unice.fr (Mathieu Olivier) wrote: >: >does anybody know, where i can find "The RenderMan Companion" ??? >: >it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) >Any bookshop should be able to order it for you. You just need the ISBN. And right Willem is. Here you go: ISBN 0-201-50868-0 The Renderman Companion by Steve Upstill Addison-Wesley Publishing Company Sorry for being unhelpful the first time around. Paul
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: How to change finger display? Date: Wed, 3 Apr 1996 07:13:33 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Apr3.071333.3042@seer.demon.co.uk> References: <4jsafo$7tm@darkwing.uoregon.edu> In article <4jsafo$7tm@darkwing.uoregon.edu> schultz@darkwing.uoregon.edu (Stewart Schultz) writes: > Is there a simple way to change the NeXT finger command so that it > returns only the .plan contents, and no login information? > > If a fingerd is available, perhaps one could simply delete the > portion that looks in a system file for login info? If you can't find a suitable fingerd recplacement, then tcp-wrappers lets you write a script to run whena finger is called, which could easily be set up to do what you suggest. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 3 Apr 1996 10:11:45 GMT Organization: University of Heidelberg, Germany Message-ID: <4jtit1$am5@sun0.urz.uni-heidelberg.de> References: <316186b9.4262367@news.plsys.co.uk> In article <316186b9.4262367@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: > I agree that the FoundationKit is nice to develop with, but can you, > or anyone, give me a really good technical reason why its > functionality could not have been fitted within the old Object > hierarchy? Firs one general point: Foundation is a *new* kit! It is not just an extension to previous AppKit hacks like List and HashTable. Before we had just a collection of nifty useful classes helping us to do our GUI work more easily. Now we have a kit of cooperating classes, giving us strong and powerful design patterns. But being a *new* kit, it has the *rights* to be really new! I've seen a lot of different implementations of array, hash table, string, etc. pp classes. Should Foundation try to be backwards compatible with all of them? No! And because NeXTSTEP was using similar classes in the past do not make them more privileged ;-) After all, I was told that the conversion scripts do very proper job... > When I consider the possible reasons, I find myself able to discount > them. > > 1) The memory management? > > Release, retain and autorelease could easily have been added to > Object; after all, the retain counts are externally maintained. > > Dealloc could have remained free ; so what that its name pairs better > with alloc? Oh no! If you allocating an object, you should DEALLOCating it, not free-ing it. Free is really a silly name ;-) > 2) The more consistent method names? > > The new methods could surely have been added to Object (or other > classes as appropriate), and obsolete _methods_ phased out, instead of > a whole class hierarchy Why? Why on earth should we carry tons of garbage? One could find always conservative thinking people in any community. Does it means that we should stop the progress, just because this handful of people do not feel very comfortable in a rapidly changing environment? > 3) The NS class name prefix? > > An advantage surely outweighed by the dual hierarchy / need for > conversion scripts. I can really not see the point here! What a "dual hierarchy" you are talking about? > I'd rather have Objective-C enhanced to support namespaces. Huh!?! > Mind you, proper support for private methods would be a nice start. > What exactly _have_ Next done with the language itself since they > acquired the rights? One of the biggest advantages of ObjC over C++ is simplicity. I'm perfectly happy with what I have now. Every experienced programmer / dev team has his own style for private methods. This is mine: If the (public) class name is called say Blah, I'm writing a category in the Blah.m just before @implementation Blah It looks like this: @interface Blah(__local_Blah) .. priv-methods .. @end and than implementing priv-methods in arbitrary order in the implementation part of Blah.m And yes, NeXT done a lot for ObjC! The PROTOCOLS! They are really cool! > 4) NSDictionary / NSArray / NSEnumerator ? > > Again, surely HashTable and List could simply have been enhanced? And > NSDictionary requires all keys to be NSStrings, unlike the more > general HashTable. What's wrong with strings as *only* key type for dictionaries? You can always produce a NSSting from the Object ID. But strings give you even more power and flexibility. I'm using them for unique IDs in SciTools project where we have really distributed system. Really cool! BTW, I was told that the NSDictionary in 4.0 can eat id's as dictionary keys either. > 5) The mutability concepts? > > Okay, NSImmutableList:List would violate the substitutability > principle, but again I think this is probably worth it to prevent > disposing of an entire existing class hierarchy. And anyway, it still > doesn't explain why you couldn't have NSMutableArray:NSArray:Object RTFM ... class clusters! And, see comments like "this method is fast..." in various places in the docs. All this is possible because of the class clusters and is *not* possible by simple subclassing. BTW, NSNumber is great example where one can add his/her own private classes! Really great! > 6) EOF? > > EOFault is a root class anyway, and Object already supports > EOKeyValueCoding via a category. Could you explain your problem? > 7) DO? > > Don't see why. Again, NXProxy is a root class anyway. So what? Pre-OS DO was a hack. Now it is a concept. Being such, it has it's own rights of existence, beauty and elegance. See my general comments. > 8) The "new" encoding scheme? > > Yet again, why not just add it to Object. Because is *new*. It's not a replacement! This should be really a FAQ! > 9) Something else......? > > Just to make it clear what I'm saying, I think Next's decision to > clean the API up for OpenStep was a good one in itself, but this could > have been done in a more incremental fashion. Perhaps I've been > "Living in a Hybrid World" for too long..... ;-) Yes! Thank you NeXT Dev team for saving us of a la M$ backwards compatibility stupidity. -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 3 Apr 1996 12:38:56 GMT Organization: Norden 1 Communications Message-ID: <4jtrh0$oqo@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF-A Plus Sybase or Oracle--A Plus Excellent opportunity Relocation assistance Greater Chicago Area To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 3 Apr 1996 18:22:23 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4jufkv$jbc@shelob.afs.com> References: <4jtit1$am5@sun0.urz.uni-heidelberg.de> Georg Tuparev writes > [logic that gets stranger with each passing message] Let me spell out what's wrong with NeXT strategy. Five years ago, AFS starting building a pretty nice object-oriented trading system. It works. All things considered, we're not looking for excuses to change it. We just want to keep reselling it as frequently as possible, with as few changes as possible, and make some money in the process. Now OPENSTEP comes along and we are basically FORCED to change everything. Give me ONE good reason why List could not have survived as an available class in OPENSTEP. You can't. There isn't one. I don't give a sh*t whether NSArrays are better or more elegant than Lists. Lists work OK for me, and I do not feel like doing the extra work. But I don't have a choice. I have to run a script, and learn a whole bunch of new classes, and hand-fix all the things the script can't do... WHERE'S THE REUSE IN THAT?! All of the supposed advantages of OOP disappear when your underlying class library vendor tells you, in so many words, "convert or die." I agree with Karl: A LOT of this stuff could have been handled through new methods in existing objects, and there's no reason I can see why the Object class couldn't have offered BOTH forms of archive encoding (read/ write and encode/decode), depending on which you wanted. I understand that for certain classes, there were Mach dependencies, but the changes in OS far exceed what was mandated by those considerations. -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: izumi@pinoko.berkeley.edu (Izumi Ohzawa) Newsgroups: comp.sys.next.programmer Subject: Re: How to change finger display? Date: 3 Apr 1996 19:04:21 GMT Organization: University of California, Berkeley Distribution: world Message-ID: <4jui3l$e3d@agate.berkeley.edu> References: <4jsafo$7tm@darkwing.uoregon.edu> In article <4jsafo$7tm@darkwing.uoregon.edu> schultz@darkwing.uoregon.edu (Stewart Schultz) writes: >Is there a simple way to change the NeXT finger command so that it >returns only the .plan contents, and no login information? >If a fingerd is available, perhaps one could simply delete the >portion that looks in a system file for login info? Actually, you should look at the source code for "finger" to change what is returned to remote finger requests. Fingerd merely passes on requests to finger. Changes you mention are easy to do, and sources for finger are very short. -- Izumi Ohzawa <izumi@pinoko.berkeley.edu> [ $@Bg_78^=;(J ] USMail: Univ. of California, 360 Minor Hall, Berkeley, CA 94720-2020 Tel: 510-642-6440, Fax: -3323, Web: http://totoro.berkeley.edu/~izumi/
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Building a palette that has Scroller-like behavior (but 2-D) Date: 3 Apr 1996 21:10:05 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4jupfd$rfb@stc06.ctd.ornl.gov> I've been trying to create a palette object that behaves like a scroller, except has two dimensional behavior. My problems are: 1) I've tried subclassing Scroller, but find that I must override the takeFloatValueFrom:sender method and reparse to get my method to show up when I try to connect a TextField to the object. Then it doesn't work (i.e. doesn't put the floatValue into the TextField). 2) If I don't subclass Scroller, I don't seem to be able to introduce methods takeXFloatValueFrom:sender and takeYFloatValueFrom:sender into my object in such a way that TextField can use them. Apparently there is something special about these takeFloatValueFrom:sender methods that I can't locate in the documentation. The connection from a TextField to a Scroller allows me to select takeFloatValueFrom: and then if I move the scroller, the value tracks in the TextField. But apparently, TextField uses the connection to set itself as the target and setFloatValue as the action of the takeFloatValueFrom: method in Scroller. What do I put into my object that that will cause this behavior with TextField? i.e. I want to have two textFields in a window, connected to my palette object so that as I move the control around, the X value tracks in one textField, and the Y value tracks in the other. Is there an example somewhere that shows this? Thanks, -- - - - - - - - - - J. W. Wooten
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 3 Apr 1996 14:50:20 -0800 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4juvbc$2i4@nova.mdd.comm.mot.com> References: <4jufkv$jbc@shelob.afs.com> In article jbc@shelob.afs.com, Greg_Anderson@afs.com (Gregory H. Anderson) writes: [ranting about all his legacy code] > All of the >supposed advantages of OOP disappear when your underlying class library >vendor tells you, in so many words, "convert or die." OpenStep for Mach 4.0 will run NS 3.x binaries won't it? So just keep one box around with an old version of NEXTSTEP and compile with that (for any necessary bugfixes/changes to your existing code). Test it on whatever NS versions your customers will run it on. Am I missing something here? [beginning of my own rant] If you want to have your NS app run on other operating systems, then port it to OpenStep, or rewrite the whole thing from scratch (Greg could use Delphi :-)). Look at that! In order to get some *benefit*, you have to do some *work*. What was that little ditty about a "free lunch"? It seems to me that the more legacy code people have laying around, the more they want to stick with "the old way" of doing things. Not because the old way is "better" but because of a mixture of laziness and greed... >Let me spell out what's wrong with NeXT strategy. Five years ago, AFS >starting building a pretty nice object-oriented trading system. It works. >All things considered, we're not looking for excuses to change it. We >just want to keep reselling it as frequently as possible, with as few >changes as possible, and make some money in the process. In other words, you want to write an app once, then milk that app until the end of time. Perhaps there is something wrong with AFS strategy, rather than NeXT's strategy? I guess its just one more example of NeXT "stomping on" ISV's... It seems the same as when NEXTSTEP went cross-platform. Would I be justified in whining because my apps using inline m68k assembly wouldn't run on Intel/Sun/HP hardware? It amazes me to no end when sysadmins and programmers ( denizens of the so-called rapidly changing computer industry) whine when they have to learn something new! [ending of my own rant] OK, OK, that was a bit unfair :-). It's just that it seems that a lot of people agree that Foundation has many benefits. If NeXT had just made some sort of arbitrary changes all over the place with no real benefits, I could understand the complaints. And I think the idea is that since OpenStep is a specification/standard of some form, they won't be able to just change it willy-nilly in the future. This is good news. Comments are welcome! -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: Matt_Watson@next.com (Matt Watson) Newsgroups: comp.sys.next.programmer Subject: Re: Problem linking in posix library Date: 3 Apr 1996 18:26:12 GMT Organization: NeXT Software, Inc. Distribution: usa Message-ID: <4jufs4$9fg@news.next.com> References: <4jruik$1oiu@news.doit.wisc.edu> In article <4jruik$1oiu@news.doit.wisc.edu> jessica@swamp.chem.wisc.edu (Jessica Severin) writes: > I'm trying to write a Nextstep 3.3 application using the Mathematica 2.2 > MathLink libraries. The problem I'm running into is that the MathLink > libraries use bsd sockets to communicate with the Mathematica application, and > these functions are not part of the NeXT_s library. They are however part of > libposix.a. [...] > How do I link in both NeXT_s and posix? > You don't. libsys has support for bsd sockets. You should find the equivalent API in libsys to do what you need (including setsockopt). /lib/libsys_s.a is linked implicitly, so you don't need to specify it on the link line. matt. -- Matt Watson Sherwood Forestry Division NeXT Software, Inc.
From: mateev@ifor.math.ethz.ch Newsgroups: comp.sys.next.programmer Subject: Q: NXConnection & Hosts Date: 4 Apr 1996 11:41:48 GMT Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4k0chs$5hb@elna.ethz.ch> Hallo, I use in my program the following message to establish a connection: server=[NXConnection connectToName:name onHost:"*"]; ---------- I.e. query the machines on the subnet until an object registered under name is found. If the connection was established and a proxy is returned, how can I find out on which host was established the connection ? Thanks in advance, Bojidar
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 4 Apr 1996 12:02:03 GMT Organization: University of Heidelberg, Germany Message-ID: <4k0dnr$pue@sun0.urz.uni-heidelberg.de> References: <4jufkv$jbc@shelob.afs.com> In article <4jufkv$jbc@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > Georg Tuparev writes > > [logic that gets stranger with each passing message] [logic that gets even stranger with each passing message] Greg: Now I can see that you have three basic problems: 1) You are not willing to understand what's behind Foundation (and OS). As I mentioned it in my "strange logic" posting, this is something *new* (neu, nouveau, novuj ... ). If you don't want to learn it, than it's your problem, not NeXT's 2) There is a general acceptance (even in any intro comp. sci. textbook) that most efforts during the lifetime and most of the money are spent in the maintenance of a software project. If you believe that this is not true, you are really on a wrong track. 3) You are not free of anti-NeXT emotions ... and this is not the best way to be objective... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
Newsgroups: comp.sys.next.programmer From: malc@daneel (Malcolm Crawford) Subject: Re: looking for new next color printer driver. Message-ID: <DpC02C.4Ms@exeter.ac.uk> Sender: news@exeter.ac.uk (news admin) Organization: University of Exeter, UK References: <4jkf5q$37t@paperboy.ids.net> Date: Thu, 4 Apr 1996 09:37:24 GMT nextjet@ids.net wrote: : does anyone remember were the site is that contains this driver. Look at the GSCorp WWW pages: http://www.gscorp.com/ It's listed under eXTRAPRINT. It's outstanding -- you wouldn't believe it's the same printer. Best wishes, mmalc.
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 4 Apr 1996 16:55:52 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4k0uuo$mub@shelob.afs.com> References: <4juvbc$2i4@nova.mdd.comm.mot.com> Steve Besler writes > OpenStep for Mach 4.0 will run NS 3.x binaries won't it? > So just keep one box around with an old version of NEXTSTEP > and compile with that (for any necessary bugfixes/changes to your > existing code). Test it on whatever NS versions your customers > will run it on. Am I missing something here? Yes, 3.3 binaries will be runnable on 4.0 MACH versions of OPENSTEP, but that doesn't solve the Windows case, which is where my customers were already headed. My complaint is that there is NO EARTHLY REASON why classes like List and Storage had to be discarded in OS/Windows, other than NeXT's arbitrary decision to scrap them and force developers to use the latest Kits Of The Month. Some developers make the hybrid solution work in 3.3 now. But as currently constructed for OS/Win, this becomes an all-or-nothing decision. > If you want to have your NS app run on other operating systems, > then port it to OpenStep, or rewrite the whole thing from > scratch (Greg could use Delphi :-)). Look at that! In order > to get some *benefit*, you have to do some *work*. What was > that little ditty about a "free lunch"? But my NS code already runs fine, dammit. Why fix what wasn't broken? > It seems to me that the more legacy code people have laying > around, the more they want to stick with "the old way" of doing > things. Not because the old way is "better" but because of a > mixture of laziness and greed... And a desire to minimize maintenance, which has now been forced upon us. > In other words, you want to write an app once, then milk that app > until the end of time. Perhaps there is something wrong with AFS > strategy, rather than NeXT's strategy? I guess its just one more > example of NeXT "stomping on" ISV's... I'm not looking to milk the same old code forever, but I am looking for enough stability to allow some return on my investment. For the record, I have 12-year-old DOS-based systems that still run fine for users who even today are in no hurry to upgrade to GUI-based alternatives. By comparison, the average age of our NS systems is 2-3 years, and none of them can survive the "new" era without serious rework. > It seems the same as when NEXTSTEP went cross-platform. Would I be > justified in whining because my apps using inline m68k assembly > wouldn't run on Intel/Sun/HP hardware? No, because NeXT didn't encourage you to use assemply. If they provide a Kit and say "this is the official way to do things," then you should be able to rely upon its continuing availability. NeXT has a very bad history of dropping whole kits that it no longer deems interesting enough to keep supporting, and that has left many people in the lurch over time. My personal opinion is that the risk of "unscheduled turbulence" has gotten so great, you would be foolish to use NeXT-supplied classes as a starting point for any serious project with a 5-10 year deployment lifecycle unless they reverse their "no source code" policy. Better to write your own and be guaranteed they will survive the whole period. > It amazes me to no end when sysadmins and programmers ( denizens of > the so-called rapidly changing computer industry) whine when they have > to learn something new! I love to learn new things. I resent being forced to learn them in order to maintain a path I have already paid for. Sun learned a major lesson when they tried to force a similar transition from SunOS to Solaris. Many people never switched or dragged their feet for years, others used it as their excuse to abandon the platform. The same thing is happening here in NeXT-land, best as I can determine. > I think the idea is that since OpenStep is a specification/standard > of some form, they won't be able just change it willy-nilly in the > to future. This is good news. Did you just fall off the turnip truck? 8^) -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Problem linking in posix library Date: 4 Apr 1996 09:46:55 -0800 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4k11uf$2jj@nova.mdd.comm.mot.com> References: <4jufs4$9fg@news.next.com> In article 9fg@news.next.com, Matt_Watson@next.com (Matt Watson) writes: >In article <4jruik$1oiu@news.doit.wisc.edu> jessica@swamp.chem.wisc.edu >Matt Watson >Sherwood Forestry Division ^^^^^^^^^^^^^^^^^^ >NeXT Software, Inc. Does anyone care to speculate what he means by this? :-) 1) Matt steals from the rich and gives to the poor. 2) Matt is one of Jobinhood's merry men. 3) Matt cuts down the trees in the Sherwood Forest. 4) All of the above. Comments anyone? ;-) Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Thu, 04 Apr 1996 21:37:09 GMT Organization: Orbital Computer Consultancy Message-ID: <316440fc.1103470@news.plsys.co.uk> References: <Dp6nnF.2xB@cwi.nl> <4jqv4t$d20@sun0.urz.uni-heidelberg.de> <316186b9.4262367@news.plsys.co.uk> <RDL.96Apr2221809@world.std.com> rdl@world.std.com (Robert La Ferla) wrote: >How about binary compatibility? What about archiving? That's what I meant by the new encoding scheme. Why not just have the new classes, such as NSString, adopt the NSCoding protocol instead of the root class? > Making major changes >to base classes is not exactly a low-risk proposition. > High risk has never stopped Next before. I think most MCCA shops would rather see controlled incremental delivery of enhancements than the Big Bang approach of a replacement hierarchy, which is itself not a low-risk proposition. I mean, not being able to use NSObjects in Interface Builder until the NSObjectIBFixes patch was hardly a great win was it?
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Thu, 04 Apr 1996 21:37:26 GMT Organization: Orbital Computer Consultancy Message-ID: <3164410a.1117449@news.plsys.co.uk> References: <316186b9.4262367@news.plsys.co.uk> <4jtit1$am5@sun0.urz.uni-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: >In article <316186b9.4262367@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl >Harbour) writes: > >> I agree that the FoundationKit is nice to develop with, but can you, >> or anyone, give me a really good technical reason why its >> functionality could not have been fitted within the old Object >> hierarchy? > >Firs one general point: Foundation is a *new* kit! It is not just an extension >to previous AppKit hacks like List and HashTable. In what sense is List a hack? It's a perfectly respectable class as far as I'm concerned. Which isn't to say it couldn't have been improved. And, if memory serves, a post last year, which I accept may have been bogus, said NSMutableArray is 3 (?) times slower than List. > Before we had just a >collection of nifty useful classes helping us to do our GUI work more easily. >Now we have a kit of cooperating classes, giving us strong and powerful design >patterns. But being a *new* kit, it has the *rights* to be really new! It still didn't have to introduce a new root object. Should every new kit have a new root? Let's forget about re-use then. > I've >seen a lot of different implementations of array, hash table, string, etc. pp > classes. Should Foundation try to be backwards compatible with all of them? >No! And because NeXTSTEP was using similar classes in the past do not make >them more privileged ;-) > Well excuse me, but I thought one of the fundamental principles of OO was to be able to extend, enhance, and change the internal implementation, of your old classes without completely replacing them. By not being able to do this with their _own_ classes, what message is Next sending to the OO community? How, as a consultant, can you sell the maintenance benefits of OO (re-use, extensibility, ability to change implementation without changing interface etc) with Nextstep, when Next themselves have had to re-code an entire class hierarchy? >After all, I was told that the conversion scripts do very proper job... > I was told that Digital, HP and Sun would ship OpenStep implementations by mid-1995. I'm still waiting... >> When I consider the possible reasons, I find myself able to discount >> them. >> >> 1) The memory management? >> >> Release, retain and autorelease could easily have been added to >> Object; after all, the retain counts are externally maintained. >> >> Dealloc could have remained free ; so what that its name pairs better >> with alloc? > >Oh no! If you allocating an object, you should DEALLOCating it, not free-ing >it. Free is really a silly name ;-) > A rose is a rose by any other name. >> 2) The more consistent method names? >> >> The new methods could surely have been added to Object (or other >> classes as appropriate), and obsolete _methods_ phased out, instead of >> a whole class hierarchy > >Why? Why on earth should we carry tons of garbage? One could find always >conservative thinking people in any community. Does it means that we should >stop the progress, just because this handful of people do not feel very >comfortable in a rapidly changing environment? > I think in the Next community one finds people who work in / come from a rarefied academic background, and people that work in the real world of commercial applications. >> 3) The NS class name prefix? >> >> An advantage surely outweighed by the dual hierarchy / need for >> conversion scripts. > >I can really not see the point here! What a "dual hierarchy" you are talking >about? > Er, hello??? You know, that dual hierarchy that is introduced by a new root object. OpenStep isn't shipping yet; perhaps you're too deep into Gnu OpenStep or a PR2 to realise that. >> I'd rather have Objective-C enhanced to support namespaces. >Huh!?! > So that Next could call their classes what the hell they like, so could I, and there would be no fear of conflict. >> Mind you, proper support for private methods would be a nice start. >> What exactly _have_ Next done with the language itself since they >> acquired the rights? > >One of the biggest advantages of ObjC over C++ is simplicity. I'm perfectly >happy with what I have now. One could find always conservative thinking people in any community. Does it means that we should stop the progress, just because this handful of people do not feel very comfortable in a rapidly changing environment? > Every experienced programmer / dev team has his >own style for private methods. In the commercial world, it's a pain in the ass that every developer that joins a team has his own way of doing things, especially things that the language should support directly anyway. And Next could get rid of the crap "Your code should not invoke this method" phrase which is all over the AppKit documentation; you'd know that if the language allowed methods to be declared private. <-- snipped Georg's way of doing private methods --> Thanks for the lesson. You're not the only experienced programmer in the world you know. > >And yes, NeXT done a lot for ObjC! The PROTOCOLS! They are really cool! > Er, excuse me? Why don't you read what I said. Next did introduce protocols, but _before_ they got the rights to ObjectiveC from StepStone. >> 4) NSDictionary / NSArray / NSEnumerator ? >> >> Again, surely HashTable and List could simply have been enhanced? And >> NSDictionary requires all keys to be NSStrings, unlike the more >> general HashTable. > >What's wrong with strings as *only* key type for dictionaries? You can always >produce a NSSting from the Object ID. But strings give you even more power and >flexibility. I'm using them for unique IDs in SciTools project where we have >really distributed system. Really cool! > >BTW, I was told that the NSDictionary in 4.0 can eat id's as dictionary keys >either. > >> 5) The mutability concepts? >> >> Okay, NSImmutableList:List would violate the substitutability >> principle, but again I think this is probably worth it to prevent >> disposing of an entire existing class hierarchy. And anyway, it still >> doesn't explain why you couldn't have NSMutableArray:NSArray:Object > >RTFM ... class clusters! And, see comments like "this method is fast..." in >various places in the docs. All this is possible because of the class clusters >and is *not* possible by simple subclassing. BTW, NSNumber is great example >where one can add his/her own private classes! Really great! > Don't tell me to RTFM. I'm very familiar with _all_ the classes in Foundation. Explain why NSNumber could not be a subclass of Object and still have private subclasses? >> 6) EOF? >> >> EOFault is a root class anyway, and Object already supports >> EOKeyValueCoding via a category. >Could you explain your problem? I don't have a problem. My _point_ is that some of NSObject's methods, such as in the EOKeyValueCoding protocol, are there especially to support EOF, so there could conceivably be some reason EOF couldn't have used an enhanced Object. As I'm something of an EOF expert, I don't think so. Enterprise Objects should have their own base class IMHO, not least because EOKeyValueCoding allows you to read and set _any_ NSObject's ivars, hardly a behaviour which is globally desirable. > >> 7) DO? >> >> Don't see why. Again, NXProxy is a root class anyway. > >So what? Pre-OS DO was a hack. Now it is a concept. Being such, it has it's >own rights of existence, beauty and elegance. See my general comments. > ROTFL. You really do swallow everything Next says. Is it a shipping concept, or just a regular concept? >> 8) The "new" encoding scheme? >> >> Yet again, why not just add it to Object. >Because is *new*. It's not a replacement! This should be really a FAQ! So object didn't have -read: and -write: methods? > >> 9) Something else......? >> >> Just to make it clear what I'm saying, I think Next's decision to >> clean the API up for OpenStep was a good one in itself, but this could >> have been done in a more incremental fashion. Perhaps I've been >> "Living in a Hybrid World" for too long..... ;-) >Yes! Thank you NeXT Dev team for saving us of a la M$ backwards compatibility >stupidity. Microsoft's backward compatability is part of the reason why they have so completely dominated commercial computing over the last 10 years, and are likely to continue to do so. Cheapness and awareness of shrinkwrap are others. Show me an IT director / manager who's happy about converting their existing Next apps to the new class hierarchy. Personally, I'd be far happier with some new method's in the old root class and some new subclasses. I maintain my assertion that all of the facilities of Foundation (and OpenStep by implication) could have been provided without NSObject. Then Next would have provided the best of both worlds - fantastic new features, _and_ backward compatability. That's what OO is supposed to be about, for Christ's sake. BTW, I _know_ the autorelease behaviour can be done without a new root class, because I've _done_ it.
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 4 Apr 1996 13:23:50 -0800 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4k1el6$2k4@nova.mdd.comm.mot.com> References: <4k0uuo$mub@shelob.afs.com> In article mub@shelob.afs.com, Greg_Anderson@afs.com (Gregory H. Anderson) writes: >I'm not looking to milk the same old code forever, but I am looking for >enough stability to allow some return on my investment. For the record, >I have 12-year-old DOS-based systems that still run fine for users who >even today are in no hurry to upgrade to GUI-based alternatives. Just curious, can the newest version of the development tools you used still compile the original source, completely unchanged? Are you running them on the same DOS version, or a newer one? >My personal opinion is that the risk of "unscheduled turbulence" has >gotten so great, you would be foolish to use NeXT-supplied classes as a >starting point for any serious project with a 5-10 year deployment >lifecycle unless they reverse their "no source code" policy. Better to >write your own and be guaranteed they will survive the whole period. What is your take on NeXT giving IXKit and 3DKit to the MiscKit folks? >Did you just fall off the turnip truck? 8^) I don't think so... What does that mean? :-) Cheers, Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 4 Apr 1996 22:31:51 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4k1ikn$nrk@shelob.afs.com> References: <3164410a.1117449@news.plsys.co.uk> Much as I generally despise "me too" posts... Karl Harbour writes > Well excuse me, but I thought one of the fundamental principles of OO > was to be able to extend, enhance, and change the internal > implementation, of your old classes without completely replacing them. > By not being able to do this with their _own_ classes, what message is > Next sending to the OO community? Amen, brother. > How, as a consultant, can you sell the maintenance benefits of OO > (re-use, extensibility, ability to change implementation without > changing interface etc) with Nextstep, when Next themselves have had > to re-code an entire class hierarchy? Personally, I'd be far happier > with some new method's in the old root class and some new subclasses. I said, let me hear you say "Amen!" > I maintain my assertion that all of the facilities of Foundation (and > OpenStep by implication) could have been provided without NSObject. > Then Next would have provided the best of both worlds - fantastic new > features, _and_ backward compatability. That's what OO is supposed to > be about, for Christ's sake. Everything Karl said goes double for me. I STILL have not read a valid technical reason why this "great renaming" was considered necessary. I have heard stories that this was done to placate Sun by removing all of the NEXTSTEPish vestiges from a supposedly "open" spec. But as you said, one of the greater ironies in the short history of OOP is that one of its most skilled practitioners couldn't pull off a few iterations of reuse on this scale. I've wondered about the impact this will have on skeptics. By comparison, the bump-up from Delphi 1.0 to 2.0 has been pretty smooth, considering how many improvements were made (including support for the entire Win/32 API and new Win95 standard controls). -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.lang.pascal.delphi.misc,comp.sys.next.programmer,phl.jobs.offered,comp.object,misc.jobs.offered Subject: JOB: Delphi in Philadelphia Date: 4 Apr 1996 23:11:24 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4k1kus$nuf@shelob.afs.com> It's spring, and everything seems reborn except your sagging career. We have the solution! Join a leading provider of Wall Street object technology as we apply our NEXTSTEP/Unix knowledge to Win/NT and Borland's amazing Delphi. Pascal, Win API, RDBMS, NEXTSTEP/OPENSTEP a plus, but mostly we're looking for the right total person to fit the team. Most convincing cover letter on the topic "OOP will rule the early 21st century" gets the prize. Anderson Financial Systems 909 Sumneytown Pike, Suite 106 Springhouse, PA 19477 jobs@afs.com (MIME and NeXTmail OK) This is a full-time employment (NOT consulting) opportunity which requires daily on-site attendance at our offices in suburban Philadelphia. Sorry, consultants and off-site employees are not appropriate for this position. -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: Matt Watson <Matt_Watson@NeXT.com> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Thu, 04 Apr 1996 16:44:26 -0800 Organization: NeXT Software, Inc. Message-ID: <31646CEA.7994@NeXT.com> References: <4k0dnr$pue@sun0.urz.uni-heidelberg.de> <4k0u7t$mtc@shelob.afs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Gregory H. Anderson wrote: > > Georg Tuparev writes > > Now I can see that you have three basic problems: > > 1) You are not willing to understand what's behind Foundation (and OS). [...] > I resent being FORCED to use them when the old ones > were not broken for my purposes... You shouldn't be talking about pre-release software, but since you did... All the common classes still exist in OPENSTEP for Windows: $ cat > foo.m #import <stdio.h> #import <objc/Storage.h> void main(void) { Storage *foo = [Storage new]; printf("Hello from %s\n", [[foo class] name]); } ^Z $ make foo $ ./foo Hello from Storage $ echo $OS Windows_NT $ matt. --- Matt Watson Thaumaturge NeXT Software, Inc. Matt_Watson@NeXT.com
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Why does Foundation Kit introduce a "n Message-ID: <DpCzD1.F7n@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4juvbc$2i4@nova.mdd.comm.mot.com> <4k0uuo$mub@shelob.afs.com> Date: Thu, 4 Apr 1996 22:19:49 GMT Come on, Steve, Georg, the point is not to remove NSObject from the system or so; the point is to retain Object in order to be able to continue to maintain NextStep software, and to ease porting NextStep software to OpenStep. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: osg@onramp.net (Glenda ) Newsgroups: comp.sys.next.programmer Subject: NeXT AND WWW Date: Fri, 05 Apr 1996 00:00:39 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4k1gol$qv1@news.onramp.net> NeXT AND WWW DEVELOPER Object Systems Group is a OO technology-based consultancy that provides assistance to Global 1000 corporations. These clients have already made the commitment to move to objects. That means that we can offer you the opportunity to build robust infrastructures, develop good designs, and direct state of the art implementations for large scale OO projects. Because OSG has a proven successful OO process and a reputation for excellence, we can keep you progressing in OO technology while you are making contractors wages. We currently have an assignment in the NE for the following area: NeXT AND WWW Minimum 2+ years total experience with extensive Objective C on a NeXT platform. Individual will be working in a team of developers for a application which is NeXT based, and will be responsible for the development of the WWW portion of that project. If you or someone you know might be interested in cutting edge work, please contact me either by e-mail or sending your resume to osg@onramp.net, US mail your resume to Object Systems Group, Inc., Attn: Glenda Maddox, 3048 Infomart, 1950 Stemmons Freeway, Dallas, Tx. 75207, or fax it to 214-742-5847.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9604051417.AA00703@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Fri, 5 Apr 96 18:16:56 +0400 Subject: [Q:] EOQualifier inner format Hi, Could someone please tell me the format of inner data in EOQualifier and how - initWithEntity:(EOEntity *)entity qualifierFormat:(NSString *)qualifierFormat, ...; method works. Neither headers nor [myQualifier description] say much. The problem is that - initWithEntity: qualifierFormat: doesn't work with NSCalendarDates (reference: 49659 for more details) and in my flat file database project I can't EOAdaptor - formatValue: forAttribute: in suggested by NeXT workaround, i.e. EOAdaptor *adaptor = [[[[ds databaseChannel] adaptorChannel] adaptorContext] adaptor]; qualifier = [[EOQualifier alloc] initWithEntity:[ds entity] qualifierFormat:@"%A < %@", @"NAME", [adaptor formatValue:startDate forAttribute:dateAttr]]; for I've got no EOAdaptor. So I decided to reimplement EOQualifier and want my implementation to be compatible with NeXT's one. Well, any information regarding EOQualifier, NSCalendarDate and flat file datasource EOF example will be greatly appreciated. Thanks. Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: uhendjx@meaddata.com (John Hendry) Newsgroups: comp.sys.next.programmer Subject: NSCalendarDate prob? Or our bug? Date: 5 Apr 1996 15:56:30 GMT Organization: Steel Driving Software, Inc. Cincinnati, OH Message-ID: <4k3fre$2dt@mailgate.lexis-nexis.com> Has anyone run into a problem with NSCalendarDates, when trying to create a new date which has no spaces in the format? When I try to create a new date using [NSCalendarDate dateWithString:@"19960101" calendarFormat:@"%Y%m%d"] an exception is raised during an NSString method. If there are spaces in the date & format (eg: @"1996 01 01" + @"%Y %m %d") there's no problem. Has anyone run into this before? Or shall I go cruft spelunking? Thanks, Jon
Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: NSCalendarDate prob? Or our bug? Message-ID: <DpFyAD.I2@hurka.UUCP> Keywords: NSCalendarDate bug Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4k3fre$2dt@mailgate.lexis-nexis.com> Date: Sat, 6 Apr 1996 12:49:24 GMT In article <4k3fre$2dt@mailgate.lexis-nexis.com> uhendjx@meaddata.com (John Hendry) writes: > Has anyone run into a problem with NSCalendarDates, when trying > to create a new date which has no spaces in the format? > > When I try to create a new date using > > [NSCalendarDate dateWithString:@"19960101" calendarFormat:@"%Y%m%d"] > > an exception is raised during an NSString method. > > If there are spaces in the date & format > (eg: @"1996 01 01" + @"%Y %m %d") there's no problem. > > Has anyone run into this before? Or shall I go cruft spelunking? I ran into this just a few weeks ago. It looks to me that implementation of %Y does not read four-characters-integer (%4d) as it should, but whole number (%d). Here is example I prepared: #import <stdio.h> #import <foundation/foundation.h> int main() { NSCalendarDate *date; date=[NSCalendarDate dateWithString:@"951205" calendarFormat:@"%y%m%d"]; NSLog(@"%@\n",date); date=[NSCalendarDate dateWithString:@"1995 1205" calendarFormat:@"%Y%m%d"]; NSLog(@"%@\n",date); date=[NSCalendarDate dateWithString:@"11111 1205" calendarFormat:@"%Y%m%d"]; NSLog(@"%@\n",date); date=[NSCalendarDate dateWithString:@"19951205" calendarFormat:@"%Y%m%d"]; return 0; } I just had not time to report it to bug_next@next.com nor KBNS. :-( Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: Sat, 06 Apr 1996 15:22:11 GMT Organization: Orbital Computer Consultancy Message-ID: <31668a63.4169258@news.plsys.co.uk> References: <4juvbc$2i4@nova.mdd.comm.mot.com> <4k0uuo$mub@shelob.afs.com> <4k2gfs$9j4@usenet.rpi.edu> Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: <-- snip --> >The whole **-POINT-** of object-oriented technology is code reuse. > >You should not have to change the code for some object until the >real-world equivalent of that object changes. The issue at the >heart of this discussion is not "Does NeXT do nasty things to people >it should be working with?", the issue is "Does Object-Oriented >Technology live up to it's claims?". Notice that the real issue >is not necessarily related to NeXT, it's a much more important >issue. If you have to rewrite all your objects for every major >release of NeXTSTEP/OpenStep, then what has object-oriented technology >really bought you? > No, I think the debate is specifically about what Next have done, since we are not making the a priori assumption that Next _had_ to replace Object with NSObject. It is however important to consider what message Next is sending having made this decision. >The only valid argument for the changes must be based on the merit >of those changes. It is utterly unacceptable to defend the changes >by implying that someone is "greedy" to want to reduce maintaince >costs. There are enough programs left to write in this world that >we (as an industry) do not need to keep going back and rewriting >the code we've already written. > >I've read the ideas behind some of the changes in OpenStep, and >they do seem worthwhile. The issue is more _how_ the changes have been implemented, as opposed to the changes themselves. > I would like to see them defended on that >basis, or else one is left to assume they were not worthwhile. I >look forward to the final release, so we can have an informed feel >for how well these ideas will play out in the real world. > All the FK stuff has been available since EOF 1.0, which is, what, 2 years now? The final release of OpenStep won't prove anything, since there will be no control experiment to compare it with. You'd need two products to compare, an OpenStep based on Object and an OpenStep based on NSObject.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Sat, 06 Apr 1996 15:22:14 GMT Organization: Orbital Computer Consultancy Message-ID: <31668ad2.4279932@news.plsys.co.uk> References: <4jufkv$jbc@shelob.afs.com> <4k0dnr$pue@sun0.urz.uni-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: <-- snip --> >2) There is a general acceptance (even in any intro comp. sci. textbook) that >most efforts during the lifetime and most of the money are spent in the >maintenance of a software project. This is exactly the problem that OO technology is supposed to address!
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 6 Apr 1996 07:59:10 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4k588e$7fv@news.its.com> References: <4jufkv$jbc@shelob.afs.com> <4k0dnr$pue@sun0.urz.uni-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: [ ... ] > Greg: > Now I can see that you have three basic problems: > 1) You are not willing to understand what's behind Foundation (and OS). > As I mentioned it in my "strange logic" posting, this is something *new* > (neu, nouveau, novuj ... ). If you don't want to learn it, than it's your > problem, not NeXT's I have seen no evidence to support the assertion that Greg does not understand and is not willing to learn what's behind Foundation and OPENSTEP. > 2) There is a general acceptance (even in any intro comp. sci. textbook) > that most efforts during the lifetime and most of the money are spent in > the maintenance of a software project. That depends entirely upon the situation. And let's remember that this is a matter of perspective-- if my company does some MCCA consulting for a big customer, we're generally happy to perform maintanance work and so forth on the project later on. > 3) You are not free of anti-NeXT emotions ... and > this is not the best way to be objective... Nonsense. I have seen no evidence to support this assertion, either. In fact, the only person I've seen with an obvious bias is *you*, Georg-- whether you choose to admit it or not. -Chuck, who probably will be added to Georg's list of suspects who harbor "anti-NeXT emotions".... Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
Newsgroups: comp.sys.next.programmer From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Subject: Re: Why does Foundation Kit introduce a "n Message-ID: <DpGt2n.KDs@news2.new-york.net> References: <4juvbc$2i4@nova.mdd.comm.mot.com> <4k0uuo$mub@shelob.afs.com> <4k2gfs$9j4@usenet.rpi.edu> <31668a63.4169258@news.plsys.co.uk> Date: Sat, 6 Apr 1996 23:54:23 GMT k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: > > The only valid argument for the changes must be based on the > > merit of those changes. It is utterly unacceptable to defend > > the changes by implying that someone is "greedy" to want to > > reduce maintaince costs. There are enough programs left to > > write in this world that we (as an industry) do not need to > > keep going back and rewriting the code we've already written. > > > > I've read the ideas behind some of the changes in OpenStep, > > and they do seem worthwhile. > > The issue is more _how_ the changes have been implemented, as > opposed to the changes themselves. This is true. I should have said something a bit different. I remember reading some magazine articles why NeXT decided to go with a new NSObject heirarchy, and those reasons did seem reasonable at the time I was reading it. I wasn't *happy* about it, but I did think some good reasons were presented, and that they had thought about things before deciding to make that major of a change. I think that article (or articles?) are at work though, and I'm at home right now (I've finally given up on the usenet hub at RPI, and am switching my news reading to home for a little while). > > I would like to see them defended on that basis, or else one > > is left to assume they were not worthwhile. I look forward to > > the final release, so we can have an informed feel for how well > > these ideas will play out in the real world. > > All the FK stuff has been available since EOF 1.0, which is, > what, 2 years now? > > The final release of OpenStep won't prove anything, since there > will be no control experiment to compare it with. You'd need two > products to compare, an OpenStep based on Object and an OpenStep > based on NSObject. Well, I was thinking that the final release was important, because it'll be the first time we see all of these ideas in action on multiple operating systems. But you're right, it's not going to be absolute proof because we'll only have the NSObject version. I guess the question is whether they could have put together all the Foundation Kit improvements into the Object heirarchy in the same amount of time, given similar resources. I don't know the answer, of course. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
Newsgroups: comp.sys.next.programmer From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Message-ID: <DpGtow.L87@news2.new-york.net> References: <4k0dnr$pue@sun0.urz.uni-heidelberg.de> <4k0u7t$mtc@shelob.afs.com> <31646CEA.7994@NeXT.com> Date: Sun, 7 Apr 1996 00:07:43 GMT Matt Watson <Matt_Watson@NeXT.com> wrote: > Gregory H. Anderson wrote: > > I resent being FORCED to use them when the old ones > > were not broken for my purposes... > You shouldn't be talking about pre-release software, but since > you did... He wasn't talking about pre-release software, he was talking about the published description of OpenStep. That published description is all that one can expect to be available on all OpenStep platforms. Nothing about pre-release in that. > All the common classes still exist in OPENSTEP for Windows: This is good news. I, who definitely have no pre-release software, would have assumed that NeXT would have only implemented the official OpenStep specs, and not worked on getting the older methods up and running on WindowsNT. I'm glad to hear that the extra effort has been made. This should make it a bit easier to move code to the WindowsNT version, although I'm not sure what that does for any of the other OpenStep implementations (such as solaris). Also, it's not clear what that means as far as IB support. In any case, thanks for the info. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: pbrown@palms.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: How to build libg++-2.7.1 on black Date: 7 Apr 1996 00:33:19 GMT Organization: U.C. Berkeley Math. Department. Distribution: na Message-ID: <4k72gf$91t@agate.berkeley.edu> I just installed the g++-2.7.1 libraries using gcc-2.7.2; no problems except that you have to comment out the definition of hypot in one of the files. The compiler will stop and complain. you fix it, and the build runs to completion. The newer g++ will not work with NeXT's g++ libraries, so this is a necessity. Paul
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Developer/DC Area Date: 6 Apr 1996 16:52:40 GMT Organization: Norden 1 Communications Message-ID: <4k67go$t1o@tofu.alt.net> Developer NEXTSTEP Objective C EOF---A Plus Sybase or Oracle---A Plus DC Area Contract--long long term To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
Newsgroups: comp.sys.next.programmer From: jpanico@netcom.com (Joe Panico) Subject: Re: NSCalendarDate prob? Or our bug? Message-ID: <jpanicoDpFvK4.C0n@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4k3fre$2dt@mailgate.lexis-nexis.com> Date: Sat, 6 Apr 1996 11:50:28 GMT Sender: jpanico@netcom11.netcom.com John Hendry (uhendjx@meaddata.com) wrote: : Has anyone run into a problem with NSCalendarDates, when trying : to create a new date which has no spaces in the format? : When I try to create a new date using : [NSCalendarDate dateWithString:@"19960101" calendarFormat:@"%Y%m%d"] I use dates of the form @"%m-%d-%y" without any problems. -- jpanico@netcom.com Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc.
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 7 Apr 1996 14:26:39 GMT Organization: University of Heidelberg, Germany Message-ID: <4k8jav$3fh@sun0.urz.uni-heidelberg.de> References: <4k3rli$qv7@shelob.afs.com> In article <4k3rli$qv7@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > Steve Besler writes > > Just curious, can the newest version of the development tools you > > used still compile the original source, completely unchanged? > > Are you running them on the same DOS version, or a newer one? > > As a matter of fact, yes. Microsoft's BASIC compiler has always maintained > full backward compatibility. We started with version 2, and version 7 > still compiles all that code. Back at version 4, WE decided to change our > dynamic library management scheme, but that was not a requirement of the > tool upgrade.. Aaaaaaaaaaaaaaah! Now I can predict the next AFS job offer! It's for M$ BASIC programmer! Welcome to the world where objects are fiction ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 7 Apr 1996 14:37:36 GMT Organization: University of Heidelberg, Germany Message-ID: <4k8jvg$3j9@sun0.urz.uni-heidelberg.de> References: <4k0u7t$mtc@shelob.afs.com> In article <4k0u7t$mtc@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > NeXT has forced me into DOUBLE maintenance mode if I choose to stay with > them for the long term. Old customers would not dare to take my word that > "the conversion went great" and switch to a new OS-based production system > without extensive retesting. So that's a major piece of work. For that > matter, so is the unscheduled (and UNDESIRED) maintenance required for the > conversion process itself. The bottom line is, to move to Windows I have > to adopt a new paradigm. Since I'm stuck with a bunch of new work and > two source trees in either event, it becomes easier to consider moving > to a different platform altogether. Hmmm. Have you taken the decision if you are going to write your stuff for DOS, Win3.x or Win95? Perhaps you will even enter into a TRIPLE mode? > > > 3) You are not free of anti-NeXT emotions ... and this is not the best > > way to be objective... > > Wake up and smell the toast burning, Georg. I have been a consistent > critic of NeXT [Computer|Software] Inc. for many years. That has not > interfered with my advocacy for its PRODUCTS, nor the significant volume > of resources my company has pledged to those products. I'm sure you cannot > name five other ISVs who have devoted more time, money, personnel, and > products to the NEXTSTEP market than AFS, sticking through thick and thin > over a period of five years.If that doesn't give me the right to complain, > I don't know what does. That's true. That's why I'm wondering if you are really Greg or just SSIII is posting with a falsified signature -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 7 Apr 1996 16:02:43 GMT Organization: University of Heidelberg, Germany Message-ID: <4k8ov3$4j3@sun0.urz.uni-heidelberg.de> References: <3164410a.1117449@news.plsys.co.uk> In article <3164410a.1117449@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: > >Firs one general point: Foundation is a *new* kit! It is not just an extension > >to previous AppKit hacks like List and HashTable. > > In what sense is List a hack? It's a perfectly respectable class as > far as I'm concerned. Which isn't to say it couldn't have been > improved. It's true that List was one of the most useful classes. But it does not fit in a framework. It has no Enumerator, no mutable version, no Autorelease stuff (are you really willing to use List and then send to any of the objects it contains retain and release methods? ... I'm eager to see if you can read your own source). > Well excuse me, but I thought one of the fundamental principles of OO > was to be able to extend, enhance, and change the internal > implementation, of your old classes without completely replacing them. > By not being able to do this with their _own_ classes, what message is > Next sending to the OO community? The messages are very simple: - Nobody is perfect - Don't be backwards compatible with your imperfectness; - Never write helper classes without any concept or OO design. Adam was probably the first peson who discovered the firs two "truths". The last one is a NeXT discovery. Even though, they did not called OS "New technology" ;-) > How, as a consultant, can you sell the maintenance benefits of OO > (re-use, extensibility, ability to change implementation without > changing interface etc) with Nextstep, when Next themselves have had > to re-code an entire class hierarchy? Huh? AppKit is basically the same. Just polished names, several new classes, and simplified display strategy. FK is *new*. So where is the "re-coding"? > >After all, I was told that the conversion scripts do very proper job... > > > > I was told that Digital, HP and Sun would ship OpenStep > implementations by mid-1995. I'm still waiting... Well, I've seen it ... but I tought I was not allowed to talk about it ;-) > >> When I consider the possible reasons, I find myself able to discount > >> them. > >> > >> 1) The memory management? > >> > >> Release, retain and autorelease could easily have been added to > >> Object; after all, the retain counts are externally maintained. > >> > >> Dealloc could have remained free ; so what that its name pairs better > >> with alloc? > > > >Oh no! If you allocating an object, you should DEALLOCating it, not free-ing > >it. Free is really a silly name ;-) > > > > A rose is a rose by any other name. Are you really sure? Once I had to write a Win program. I found something like the NXRect structure. Two of the fields ware called top and bottom. Well, I used them ... and the program was not doing what I was expecting. After a day of debugging, I figured out, that top is bottom and vs. It's even documented "feature", but who on earth is reading a doc about Top's and Bottom's? So if you call your rose a tulip, you can even buy a PC hardware thinking it's a RDBMS ;-) > >Why? Why on earth should we carry tons of garbage? One could find always > >conservative thinking people in any community. Does it means that we should > >stop the progress, just because this handful of people do not feel very > >comfortable in a rapidly changing environment? > > > > I think in the Next community one finds people who work in / come from > a rarefied academic background, and people that work in the real world > of commercial applications. Even COM programmers have not infinite amount of biopolymer based RAM. > Er, hello??? You know, that dual hierarchy that is introduced by a new > root object. OpenStep isn't shipping yet; perhaps you're too deep into > Gnu OpenStep or a PR2 to realise that. Aha! And where is the trouble there? I'm including in any of my projects a library called libdual_world ;-) and don't care if I'm using NSO or O classes. It just works... that's BTW all I need. > >> Mind you, proper support for private methods would be a nice start. > >> What exactly _have_ Next done with the language itself since they > >> acquired the rights? > > > >One of the biggest advantages of ObjC over C++ is simplicity. I'm perfectly > >happy with what I have now. > > One could find always conservative thinking people in any community. > Does it means that we should stop the progress, just because this > handful of people do not feel very comfortable in a rapidly changing > environment? Ha ha! What a cool pun! But there is a difference between keeping something simple and conservatism?!?! BTW, the new FK classes having an similar helper class included in the old AppKit are much easier! Think about it! > >And yes, NeXT done a lot for ObjC! The PROTOCOLS! They are really cool! > > > > Er, excuse me? Why don't you read what I said. Next did introduce > protocols, but _before_ they got the rights to ObjectiveC from > StepStone. It makes a big difference ... for the history of the computer science ;-) > Don't tell me to RTFM. I'm very familiar with _all_ the classes in > Foundation. Explain why NSNumber could not be a subclass of Object and > still have private subclasses? Explain me why you are not willing to pass your source trough a perl, csh, or NeXT provided script which replaces all occurrences of Object with NSObject? Perhaps we should stop using cpp either? > >> 6) EOF? > >> > >> EOFault is a root class anyway, and Object already supports > >> EOKeyValueCoding via a category. > >Could you explain your problem? > > I don't have a problem. My _point_ is that some of NSObject's methods, > such as in the EOKeyValueCoding protocol, are there especially to > support EOF, so there could conceivably be some reason EOF couldn't > have used an enhanced Object. As I'm something of an EOF expert, I > don't think so. Could you give examples? > Enterprise Objects should have their own base class IMHO, not least > because EOKeyValueCoding allows you to read and set _any_ NSObject's > ivars, hardly a behaviour which is globally desirable. Hmmm. It's is philosophical question. In general I do not care of _any_ of the ivars ... just few of them. But you are right that it will be nice if there was a wrapper around a NSDictionary which only purpose is to provide more general way for setting/getting properties of any type. It may reduce the class explosion. > >> 8) The "new" encoding scheme? > >> > >> Yet again, why not just add it to Object. > >Because is *new*. It's not a replacement! This should be really a FAQ! > > So object didn't have -read: and -write: methods? ..and they ware very useful for the DO indeed! > > > >> 9) Something else......? This time some wishes: NeXT should make the nib and the archiving formats part of OS .. and a question: Why on earth is the NSScanner class called so? What aboout calling it NSStringParser? -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 7 Apr 1996 17:38:37 GMT Organization: University of Heidelberg, Germany Message-ID: <4k8uit$5fl@sun0.urz.uni-heidelberg.de> References: <4k2ftr$9j4@usenet.rpi.edu> In article <4k2ftr$9j4@usenet.rpi.edu> Garance A Drosehn <gad@eclipse.its.rpi.edu> writes: > tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: > This is a true statement, but what's your point? That NeXT has > every right to change around everything every release just to make > sure that all even more of your costs are in maintanence? If > that is your best defense, then it's not very convincing. No! Not after every release. But after 7 years? Everything? Hey guys! What's the big fuzz of parsing your sources through several scripts compared to the benefits of having all these new cool concepts? I thought I don't like unix and all these ugly shells, but it seams you are even more militant ;-) > > 3) You are not free of anti-NeXT emotions ... and this is not > > the best way to be objective... > But I am increasingly irritated of you dismissing Greg's remarks > as being "anti-NeXT". If you have something to say about his > statements, then say it. Don't pretend that you're contributing > anything useful by acting as if you have some right to ignore his > remarks. I told it. I was expecting a fair comparison between OS! and delphi. I asked questions and I never got answers (if you don't count complains like why all class names start now with NS, or DO is not important). > > I'd like to maintain my gung-ho enthusiasm for NeXT's technology, > but if that means I find myself on the same side as someone who is > so utterly brainwashed Thanks! I know brainwashed people here, but that's not me. I started posting pro-NeXT mails when I've seen that everybody is explaining his/her own nonsuccess with NeXT's decisions, no matter what they are. But I denny to believe, that it is a huge catastrophe for all of as to move from NS to OS. I denny to believe that programmers with almost a decade of NS experience have *real* problem learning the new OS docs (e.g. it's easier for Greg to learn Delphi de novo, then to get the new OS docs), I denny to believe that it's easier for somebody to write a new app from scratch then to convert one old NS app to OS... as a matter of a fact, this list could be very long. The single explanation I found is that experience programmers posting such claims are full of anti-NeXT emotions. If you have better, than I will be happy to agree with you. > then I have to start worrying that maybe > I'm on the wrong side of the issue. Please stop the smart-aleck > dismissive attacks, pretending as if you're the only one here who's > written any significant amount of code. The only person you're > fooling is yourself, and even then only if you're particularly > easy to fool. Long time ago as this discussion flamed for the fist time (just after OS spec came out) I made a proposal to move it to a group alt.next.names. Yes! it is true I'm fooling myself, but at least a have the fun seeing respected people fooling themselves too. It's really a great fun! Isn't it? But at least I didn't start this stupid drama! And yes, it is a drama, because I almost never got answers to my FK related questions, and I was asked in private mails by *very* respected people to give them examples for use of DO, Scanner, Not.Center.... All that tells me that this discussion was introduced and continued by people who have no idea of what's going on with OS. And all these people are CEO's, presidents, heads, and so on of companies buiding the bridge between NeXT and all other NS users. It is sad to know, that an app will be not more supported because it is very hard to convert it to OS. And I already got such an explanation from one very "respected" guy! If I was a greenhorn, I could buy this explanation, but without calming to be the coolest expert around, I could say that this is not NeXT fault! I had to rewrite almost all of my sources several times, but I never blamed other people for my own stupidity. It is sometimes hard to say "oh stupid me", and it is always hard to see, that if you had better design, your project could be done faster and maintained easier, but why claim that other people are responsible for that? And it seams to me, that one is added to a virtual black sheep list, just because is saying to somebody else "thank you for giving us this product! It's fun to use it". But that is what gives most motivation, and the excitement of thinking on new ways and finding solutions to complicated and yet unsolved problems. You should think if it itn't something wrong with you if you really believe, that every "thank-you-saying" person is brainwashed. > You're doing more to discourage me about NeXTSTEP than anything > Greg has ever said or done. Yes! There are a lot of people loosing the interest on or the attraction of a software product when they are told that it's easy to use it ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Why does Foundation Kit introduce a "n Message-ID: <DpI6uw.42r@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4k2gfs$9j4@usenet.rpi.edu> <31668a63.4169258@news.plsys.co.uk> <DpGt2n.KDs@news2.new-york.net> Date: Sun, 7 Apr 1996 17:49:44 GMT In article <DpGt2n.KDs@news2.new-york.net> Garance A Drosehn <gad@eclipse.its.rpi.edu> writes: >multiple operating systems. But you're right, it's not going to >be absolute proof because we'll only have the NSObject version. But fortunately there is now the semi-official assurance of Matt Watson that OpenStep will "still" contain the Object hierarchy; finally some good news. It's a matter of pragmatism : some people have large amounts of Object subclasses (in my case even portable to other operating systems than NextStep, nota bene), so it would be stupid from NeXT's side to remove Object. The word "still" implies however that they reserve the right to remove the Object subclasses at some future point, of course :-( But anyhow, the fact that Object is now part of OpenStep is extremely good news; you can now use *single* sources to build software that will run on both NextStep (all versions!) and OpenStep. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Message-ID: <DpI79r.496@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <3164410a.1117449@news.plsys.co.uk> <4k8ov3$4j3@sun0.urz.uni-heidelberg.de> Date: Sun, 7 Apr 1996 17:58:38 GMT In article <4k8ov3$4j3@sun0.urz.uni-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) writes: > >Explain me why you are not willing to pass your source trough a perl, csh, or >NeXT provided script which replaces all occurrences of Object with NSObject? > Because 1+1 might very well give something else than 2, after such a change on my system. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: Marc@AccessOne.com(Marc Salvatori) Newsgroups: comp.sys.next.programmer Subject: PB & IB: Forgetting Window Placements? Date: 7 Apr 1996 17:49:49 GMT Organization: AccessOne Message-ID: <4k8v7t$5ll@news.accessone.com> Keywords: PB Project Builder IB Interface I'm new to these tools, and am dismayed to find that they fail to remember the most recent placement of browsers, windows, and panels irregardless of saving projects and interfaces. -- >< Marc J. Salvatori | >< >< mailto:salvo@eskimo.com | MIME & NeXTMail are accepted ><
Newsgroups: comp.sys.next.programmer Subject: Sorting a Storage Object? Message-ID: <1996Apr7.114620.77792@cc.usu.edu> From: hcole@spanky.idec.sdl.usu.edu (Howard R. Cole) Date: 7 Apr 96 11:46:20 MDT Distribution: world Is a an instance of a Storage Class allocated contiguous memory so that I can use qsort with it? Is there a quick and easy way to sort a Storage object? Thanks - - HRC - hcole@spanky.idec.sdl.usu.edu
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: Anyone have a joystick driver? Date: 7 Apr 1996 20:46:42 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4k99ji$188n@core.bard.edu> -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: ambi@world.std.com Newsgroups: comp.sys.next.programmer Subject: Salaries for NEXTSTEP programmers. Date: 8 Apr 1996 01:16:51 GMT Organization: Alternate Access Incorporated Message-ID: <4k9pe3$c3u@Holly.aa.net> Can anyone help me get some ballpark figures for salaries for NEXTSTEP programmers? I realize there may be lots of factors involved such as location, other benefits, etc., but I'm just trying to get some basic idea (within 5-10K) for the following individuals: NEXTSTEP Programmer : 1 year experience. NEXTSTEP Programmer : 3 years experience. NEXTSTEP Programmer : 5 years experience. Any help would be greatly appreciated!
From: stanj@cs.stanford.edu (Stan Jirman) Newsgroups: comp.sys.next.programmer Subject: Seeks with /dev/rst0 Date: 8 Apr 1996 02:31:57 GMT Organization: Stanford University Message-ID: <4k9tqt$gb6@nntp.Stanford.EDU> Hello, I just found out that one can't perform a seek with the /dev/rst0 device. Now that's what I really need, and SafteyNet does it, too. So I tried the SCSI commands like in the 3.3 examples; for no avail. If I want to pass my class, I need to make it work, fast. So, do you have any idea how to execute a seek to the n-th block? I tried the code below, but nada. Thanks a lot, - Stan #import <stdio.h> #import <stdlib.h> #import <libc.h> #import <sys/file.h> #import <bsd/dev/scsireg.h> #import "scsi_commands.h" #define DEVICE "/dev/rst0" main () { int fh, pos; char r[100]; struct esense_reply e; struct timeval time; fh = open (DEVICE, O_RDONLY); if (fh > 1) { ioctl (fh, MTIOCVARBLK, &pos); pos = do_seek (fh, 7, &time, &e); // from examples printf ("%d\n", pos); // says -1 pos = read (fh, r, sizeof (r)); // reads the 1st block printf ("%d %s:\n", pos, r); close (fh); } else printf ("%d\n", -1); } -- +-------------------------------------------+-------------------------+ | Stan Jirman -- The Swiss CS Guy | | | stanj@cs.stanford.edu NeXTmail / MIME | When you find yourself | | http://www-leland.stanford.edu/~stanj/ | in a hole, stop digging | | PO Box 2642, Stanford, CA 94309, USA | | +-------------------------------------------+-------------------------+
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 8 Apr 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4ka3se$mta@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: skeezics@teleport.com (Skeezics Boondoggle) Newsgroups: comp.sys.next.programmer Subject: Re: renderman on nextstep Date: 8 Apr 1996 03:00:07 -0700 Organization: Teleport - Portland's Public Access (503) 220-1016 Message-ID: <4kao37$hfa@claudia.teleport.com> References: <4jram3$b18@malibu.unice.fr> <4js25t$3ee@agate.berkeley.edu> In <4js25t$3ee@agate.berkeley.edu> Paul R. Brown <pbrown@math.berkeley.edu> writes: >mo@spielberg.unice.fr (Mathieu Olivier) wrote: >>does anybody know, where i can find "The RenderMan Companion" ??? >>it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) >I bought a copy from Powell's Technical Books in Portland, OR. If you can't >find it in France, I'm sure that they could probably send you a copy. (Just >so you know, Powell's is one of the largest (single) bookstores in the U.S; >they have most everything. Last time I was there, they had full sets of NeXT >technical docs, for example.) Powell's "city of books" is *awesome*. You can wander for _days_ in there - it's grown from one full city block to multiple locations. They're even on the web, with a search engine: http://www.powells.portland.or.us Great people, great store! -- Chris Civic pride, nuthin' - we've got more microbreweries and espresso shops than any city in the nation! Powell's is just icing on the cake. :-) -- skeezics@teleport.com it's a pretty safe bet that my opinions have little bearing on reality, especially teleport's version of it.
From: jbettis@cse.unl.edu (Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Linking OpenStep/NT programs with WinNT libs Date: 8 Apr 1996 01:59:02 GMT Organization: University of Nebraska--Lincoln Message-ID: <4k9rt6$7pb@crcnis3.unl.edu> I need to link a OpenStep App with a windows NT based database server library. Namely Borland Interbase. But when I do so, I get undefined symbols for every function that I use which is supposed to be in the library. I am using Prerelease 1, with Interbase 4.0 and I compiled some test programs with Microsoft C++ and interbase so I know the library is fine, and I compiled OpenStep Yap from the Next Examples, so I know that works too. How can I mix those? Is there some special way to prototype a WinNT compiled function or something? Please help. -- Jeremy Bettis -*- PGP Public key available -*- University of Nebraska INET: jbettis@cse.unl.edu "Those who stand in the middle of the UUCP: jeremy@tddi.inetnebr.com road are often hit by passing cars." <a href="http://cse.unl.edu/~jbettis/">Click Here</a>
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Mon, 8 Apr 1996 07:13:49 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Apr8.071349.18229@seer.demon.co.uk> References: <4k8uit$5fl@sun0.urz.uni-heidelberg.de> In article <4k8uit$5fl@sun0.urz.uni-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) writes: > What's > the big fuzz of parsing your sources through several scripts compared to the > benefits of having all these new cool concepts? I thought I don't like unix > and all these ugly shells, but it seams you are even more militant ;-) Firstly, we have Steve Jobs personal asssurance that the conversion scripts won't be enough to convert your code; about one man month per 50,000 lines of code will be required, he said. This is a lot of effort (particularly if it takes longer than that :-). Secondly, such a major change will force very extensive regression testing. With multiple architectures supported by NeXTSTEP, we have been very fortunate in that for the most part a single set of source will work on all architectures, and testing across different versions hasn't needed to be extensive. With OpenStep, I expect that will no longer be the case, and testing requirements for OpenStep/NT, OpenStep/Solaris and NeXTSTEP will be much more gruelling. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: Benhur.Stein@imag.fr (Benhur Stein) Newsgroups: comp.sys.next.programmer Subject: Dragging Matrix's Cells Date: 8 Apr 1996 12:42:27 GMT Organization: IMAG, Grenoble, France Message-ID: <4kb1jj$p0d@imag.imag.fr> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit Hi all, I'd like to change the positions of cells in a matrix by dragging them, just like one can do in the IB. If this is a feature of Matrix itself, how do I enable it? And if it is not, does anybody know if this is implemented in any example or public domain software? Thanks for any help, benhur
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 5 Apr 1996 06:51:39 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4k2ftr$9j4@usenet.rpi.edu> References: <4jufkv$jbc@shelob.afs.com> <4k0dnr$pue@sun0.urz.uni-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: > Greg_Anderson@afs.com (Gregory H. Anderson) writes: > > Georg Tuparev writes > > > [logic that gets stranger with each passing message] > [logic that gets even stranger with each passing message] > > Greg: > Now I can see that you have three basic problems: No he does not. > 1) You are not willing to understand what's behind Foundation (and > OS). As I mentioned it in my "strange logic" posting, this > is something *new* (neu, nouveau, novuj ... ). If you don't > want to learn it, than it's your problem, not NeXT's It would just as easy to say that you're so enamoured with the promise of new technology, that you have no memory for the promises of last year's "new technology". > 2) There is a general acceptance (even in any intro comp. sci. > textbook) that most efforts during the lifetime and most of > the money are spent in the maintenance of a software project. > If you believe that this is not true, you are really on a > wrong track. This is a true statement, but what's your point? That NeXT has every right to change around everything every release just to make sure that all even more of your costs are in maintanence? If that is your best defense, then it's not very convincing. > 3) You are not free of anti-NeXT emotions ... and this is not > the best way to be objective... I'm as gung-ho for NeXT as just about anyone, and I've managed to maintain that enthusiasm for about five years now. Most of the people who were gung-ho about NeXT at that time are now gone. I'm still enamoured with many aspects of the technology. But I am increasingly irritated of you dismissing Greg's remarks as being "anti-NeXT". If you have something to say about his statements, then say it. Don't pretend that you're contributing anything useful by acting as if you have some right to ignore his remarks. I'd like to maintain my gung-ho enthusiasm for NeXT's technology, but if that means I find myself on the same side as someone who is so utterly brainwashed then I have to start worrying that maybe I'm on the wrong side of the issue. Please stop the smart-aleck dismissive attacks, pretending as if you're the only one here who's written any significant amount of code. The only person you're fooling is yourself, and even then only if you're particularly easy to fool. You're doing more to discourage me about NeXTSTEP than anything Greg has ever said or done. Please stop, unless of course your goal is to help make NeXT look stupid. You're doing a fine job of that. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 5 Apr 1996 07:01:16 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4k2gfs$9j4@usenet.rpi.edu> References: <4juvbc$2i4@nova.mdd.comm.mot.com> <4k0uuo$mub@shelob.afs.com> Steve Besler apparently has written: > It seems to me that the more legacy code people have laying > around, the more they want to stick with "the old way" of doing > things. Not because the old way is "better" but because of a > mixture of laziness and greed... > In other words, you want to write an app once, then milk that > app until the end of time. Perhaps there is something wrong > with AFS strategy, rather than NeXT's strategy? I guess its > just one more example of NeXT "stomping on" ISV's... No, it's a much more serious issue than that. Now there may be perfectly fine reasons for all the changes NeXT has made on the move to OpenStep, but the above attitude *must* not be used in defense of those changes. The whole **-POINT-** of object-oriented technology is code reuse. You should not have to change the code for some object until the real-world equivalent of that object changes. The issue at the heart of this discussion is not "Does NeXT do nasty things to people it should be working with?", the issue is "Does Object-Oriented Technology live up to it's claims?". Notice that the real issue is not necessarily related to NeXT, it's a much more important issue. If you have to rewrite all your objects for every major release of NeXTSTEP/OpenStep, then what has object-oriented technology really bought you? The only valid argument for the changes must be based on the merit of those changes. It is utterly unacceptable to defend the changes by implying that someone is "greedy" to want to reduce maintaince costs. There are enough programs left to write in this world that we (as an industry) do not need to keep going back and rewriting the code we've already written. I've read the ideas behind some of the changes in OpenStep, and they do seem worthwhile. I would like to see them defended on that basis, or else one is left to assume they were not worthwhile. I look forward to the final release, so we can have an informed feel for how well these ideas will play out in the real world. Of course, it's also possible that I'm just expecting too much from any discussion held in usenet. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: randy@is.com (Randy Marchessault) Newsgroups: comp.sys.next.programmer Subject: Employment Opportunity Date: 5 Apr 1996 21:12:36 GMT Organization: Integrity Solutions, Inc. Message-ID: <4k42c4$jmi@medusa.is.com> Keywords: NEXTSTEP, Objective-C, Oracle, Sybase Object-Oriented Developer Young dynamic company with strong growth opportunities seeks Object-Oriented Developer to work in creative high-tech environment. Candidates should have excellent written and verbal communication skills and knowledge in one or more of the following languages: Objective-C (NEXTSTEP), Smalltalk, and Java; and familiarity with OO methodologies. Oracle and/or Sybase development experience a definite plus. Project management experience also a plus. Must have strong motivation and self management skills. Excellent compensation. Please send email to randy@is.com. -- Randy Marchessault V.P. Integrity Solutions, Inc. 23 Empire Drive St. Paul, MN 55103 Voice: (612)-223-8474 X207 Fax: (612)-223-8481
From: xela@acm.org (Matt Bezark) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Problem connecting to Oracle7 with EOF Date: 8 Apr 1996 14:53:29 GMT Organization: Rocky Mountain Internet Inc. Message-ID: <4kb999$jp@natasha.rmii.com> I have an Oracle7 server running on an HP 712 with HPUX 10. SQL*NET is running on the server. When I try to connect with EOModeler, I get a beep, and no message in the Console. Do I need any Oracle or NeXT software on the client besides the Oracle adaptor that is Oracle specific? Do I need an entry in /etc/services on the client? Thanks, Matt
From: Alexis Rzewski <arzewski@mcimail.com> Newsgroups: comp.sys.next.programmer Subject: when an expression evaluates to nil Date: Mon, 08 Apr 1996 08:56:04 -0600 Organization: InternetMCI Message-ID: <31692904.28AD@mcimail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, in smalltalk, if in the following expression aPerson dateOfBirth year messaging dateOfBirth to aPerson evaluates to nil, a runtime error occurs and a "walkback" window is opened. In Objective-C, in the expression [[aPerson dateOfBirth] year] any messages following the one that evaluated to nil are ignored. Is this correct? - Alexis
From: Matt Watson <Matt_Watson@NeXT.com> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: Mon, 08 Apr 1996 08:05:52 -0700 Organization: NeXT Software, Inc. Message-ID: <31692B50.731@NeXT.com> References: <4k0dnr$pue@sun0.urz.uni-heidelberg.de> <4k0u7t$mtc@shelob.afs.com> <31646CEA.7994@NeXT.com> <DpGtow.L87@news2.new-york.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Garance A Drosehn wrote: > He wasn't talking about pre-release software... Well, actually, he was. I believe his exact words were: "The bottom line is, to move to Windows I have to adopt a new paradigm." If you're talking about OPENSTEP for Windows in the present tense, you're talking about pre-release software, not just the spec. I was was following up on his concern about a specific OPENSTEP implementation, just trying to address the "bottom line." matt.
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Fast scrolling in View? Date: 7 Apr 1996 10:58:36 GMT Organization: Turbocat's Development, Germany Message-ID: <4k874s$js@turbocat.snafu.de> Hi! In my app, I draw in the lower line of a view. The lines should scroll really fast to the top of the view. Now I redraw the whole view when a new line comes in my view. That's not very fast. - drawTemp; { int i,line=0, pixel; float r,g,b; float mybuffer[2048]; NXRect myRect; [self getBounds:&myRect]; [self lockFocus]; for (line=0;line < 22; line++){ pixel=[myTempMap getLine:mybuffer:2048:line]; for (i=0;i < pixel+1; i++){ [self temp2rgb:mybuffer[i]:100:350:&r:&g:&b]; PSWdrawLine(i*2.0,myRect.size.height-(10*(line+1)), i*2.0,(myRect.size.height-20*(line+1)),r,g,b, 2.0); } } [window flushWindow]; [self unlockFocus]; return self; } defineps PSWdrawLine (float x; float y;float xx; float yy; float r; float g; float b; float width) x y moveto xx yy lineto width setlinewidth r g b setrgbcolor stroke endps How can I speed that up? Thanks in advance for any help. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 8 Apr 1996 15:58:00 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4kbd28$7vp@shelob.afs.com> References: <31692B50.731@NeXT.com> Matt Watson <Matt_Watson@NeXT.com> writes > Garance A Drosehn wrote: > > > He wasn't talking about pre-release software... > > Well, actually, he was. I believe his exact words were: > > "The bottom line is, to move to Windows I have > to adopt a new paradigm." > If I was, it was only accidental. If a pre-release version of OS/Windows exists, I promise I am not programming with it. So I really was speaking only from my reading of the published OPENSTEP spec. That's why I said I was glad to have the information. And "new paradigm" was inclusive of more than just NeXT products. -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 8 Apr 1996 16:12:12 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4kbdss$819@shelob.afs.com> References: <4k8uit$5fl@sun0.urz.uni-heidelberg.de> Georg Tuparev writes > What's the big fuzz of parsing your sources through several scripts > compared to the benefits of having all these new cool concepts? For the benefit of those who still don't get it: a) I don't NEED "all these new cool concepts". They're nice, but not necessary to get my work done. It's needless, costly work for which I receive insufficient benefit. b) They create double maintenance, because if a bug is found, I have to go fix it in two source trees. (Unless you think I should fix it in one place and rerun the script.) I admit the same problem exists as a result of switching to Delphi. The point is, this is no longer an issue that can be automatically decided in NeXT's favor. > I told it. I was expecting a fair comparison between OS! and delphi. > I asked questions and I never got answers (if you don't count complains > like why all class names start now with NS, or DO is not important). I am not aware of any questions about Delphi that I did not attempt to answer back during that discussion. > I denny to believe, that it is a huge catastrophe for all of as to move > from NS to OS. I denny to believe that programmers with almost a decade > of NS experience have *real* problem learning the new OS docs (e.g. it's > easier for Greg to learn Delphi de novo, then to get the new OS docs), > I denny to believe that it's easier for somebody to write a new app from > scratch then to convert one old NS app to OS. You are free to believe and deny whatever you want, as long as you stop short of denying the actual experience others are willing to relate. And in my case, that's what you are doing. I can prove the counterpoint to your statements above, AS IT RELATES TO MY COMPANY. Stop telling me that I NEED all this stuff I don't, and stop telling me that my real-world experience with other products is wrong and misguided, and we won't have anything left to argue about. I will say one last time: Georg, in an objective sense, I agree about the technical advantages of OPENSTEP. As a complete business proposition, I don't find it nearly so compelling. -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: Matt Watson <Matt_Watson@NeXT.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: Mon, 08 Apr 1996 09:46:08 -0700 Organization: NeXT Software, Inc. Message-ID: <316942D0.6038@NeXT.com> References: <31692904.28AD@mcimail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Alexis Rzewski wrote: > > Hi, > > in smalltalk, if in the following expression > > aPerson dateOfBirth year > > messaging dateOfBirth to aPerson evaluates to nil, a runtime error > occurs and a "walkback" window is opened. > > In Objective-C, in the expression > > [[aPerson dateOfBirth] year] > > any messages following the one that evaluated to nil are ignored. > > Is this correct? > > - Alexis Not quite. Messages to nil return nil from objc_msgSend(). In the runtime, this means the register that holds object values upon return is set to 0. This is very different from "zero cast to the return value I expected." The runtime has no idea what the return type of a message to nil is supposed to be. A consequence of this is that if you sent a message to nil expecting to get a float, double, or struct back, you'll get garbage in your float, double or struct. It is always faster and often more correct to check for nil before sending a message to an object which is possibly nil. It's not that hard, happens less often than you'd expect, plus makes your program faster and your code more maintainable. matt.
From: bmont@rconnect.com (Bruce Montegani) Newsgroups: comp.sys.next.programmer Subject: NeXT/Windows95 clock discrepencies Date: 9 Apr 1996 15:21:31 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <4kdv9r$k9d@news.mr.net> I have noticed that when I boot into Window95 on my machine that the clock is 5 hours later than the clock when I boot under NeXT. When I fix either one the other one is out. Why can I not syncronize these two clocks? Don't I have just one clock on my system? Thanks Bruce Montegani
From: lbo@sma.ch (-Laurent Bourqui) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 9 Apr 1996 16:03:55 GMT Organization: Swiss Meteorological Institute Distribution: world Message-ID: <4ke1pbINN451@maz4.sma.ch> References: <4kdv9r$k9d@news.mr.net> In article k9d@news.mr.net, bmont@rconnect.com (Bruce Montegani) writes: > I have noticed that when I boot into Window95 on my machine that the clock is > 5 hours later than the clock when I boot under NeXT. When I fix either one > the other one is out. > I have exactly the same problem... > Why can I not syncronize these two clocks? > > Don't I have just one clock on my system? > I think that the answer to this problem is the following: NEXTSTEP set the system clock (in bios) to the GMT time and then calculate itself the difference with your time zone. At the opposite, Windows stores the local time to the bios!!! The only way I found to overcome this problem is to set the bios time to the local time and then set the time zone in NEXTSTEP and WinNT to GMT+00 The problem is that the daylight saving time isn't automatically done, but it's not a problem for me. Hope this help, Laurent --- -------------------------------------------------------------------------------- Laurent Bourqui | Phone (private) +41(0)1 451 62 92 SMI (Swiss Meteorological Institute) | Phone (office) +41(0)1 256 95 73 Kraebuehlstr. 58 | Fax (office) +41(0)1 256 92 78 | email lbo@sma.ch CH-8044 Zuerich | Switzerland | --------------------------------------------------------------------------------
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 9 Apr 1996 15:54:13 GMT Organization: Anderson Financial Systems Inc. Distribution: na Message-ID: <4ke175$2oo@shelob.afs.com> References: <4kcfuv$2qm@nova.mdd.comm.mot.com> Steve Besler writes > Just out of interest, what about Borland? Can I compile my Borland > Turbo Pascal 3.0 source using Delphi with no problems? We have no direct experience, but from what I have read in the Delphi groups, the answer is "yes" (except with respect to new keywords that might overlap old variable names). Pascal is still Pascal from the standpoint of procedures and functions not associated with classes. > Didn't NeXT at one point state that they would submit > the OpenStep standard to a body such as the OMG? Yes. I don't know whether that has been done, but the fact remains that if only one vendor shipping a product, they constitute the "de facto" standards committee. Viz M$. -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 9 Apr 1996 13:29:17 GMT Organization: Genoa Software Systems Message-ID: <4kdond$25b@saturn.genoa.com> References: <31692904.28AD@mcimail.com> <316942D0.6038@NeXT.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <316942D0.6038@NeXT.com> Matt Watson wrote: > Messages to nil return nil from objc_msgSend(). In the runtime, this means > the register that holds object values upon return is set to 0. This is > very different from "zero cast to the return value I expected." .... > A consequence of this is that if you sent a message to nil expecting to get a > float, double, or struct back, you'll get garbage in your float, double or struct. All true. You have to be aware of this possibility when the return value is float, double or struct. > It is always faster and often more correct to check for nil before > sending a message to an object which is possibly nil. What does "more correct" mean? > It's not that hard, happens less often than you'd expect, plus makes your program faster > and your code more maintainable. I doubt the speed difference is noticeable in most cases. Personally, I think relying upon this feature makes Objective-C code much _more_ maintainable. You can remove alot of clutter from the code so that the essential logic is unobscured. Just remember to a. make sure your methods handle nil arguments gracefully b. handle those methods that return floats, doubles and structs explicitly. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9604091842.AA07754@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Tue, 9 Apr 96 22:42:08 +0400 Subject: Re: Problem connecting to Oracle7 with EOF Cc: comp-sys-next-programmer@antigone.com Hi, > I have an Oracle7 server running on an HP 712 > with HPUX 10. SQL*NET is running on the > server. When I try to connect with EOModeler, > I get a beep, and no message in the Console. > Do I need any Oracle or NeXT software on the > client besides the Oracle adaptor that is > Oracle specific? Do I need an entry in > /etc/services on the client? Check out NeXTanswer 1897. It has tips on hooking up to Oracle. Briefly, you need the following entries in /etc/services orasrv 1525/tcp # Oracle SQL+Net V1 port tnslsnr 1521/tcp # Oracle SQL+Net V2 port or there equivalent into NetInfo and in case of SQL*Net V2 copy of server tnsnames.ora in /etc directory. Hope that will help. Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: NeXT Developers - London, Houston, Chicago, CA, and New York Date: 9 Apr 1996 14:34:02 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4keaiq$skf@newsbf02.news.aol.com> Datacom Technology Group Inc has been steadily building up an excellent reputation with our clients as a major source for full-time NeXT developers. We currently have several excellent opportunities for experienced NeXT professionals in London England, Houston, Chicago, Southern California, and New York. Candidates must have 2 to 5 years NeXT development experience, including Objective-C, AppKit, and any RDBMS. Knowledge and experience with EOF and/or PDO/DO are a big+. Salaries range from $45 to $95 + Bonus. For more information, contact: Brian Mitchell Datacom Technology Group Inc 212-629-5720 212-629-3374(FAX)
From: MaRK_BeSSeY@NeXT.CoM (Mark Bessey) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 9 Apr 1996 19:22:07 GMT Organization: NeXT Software, Inc. Message-ID: <4kedd0$1e7@news.next.com> References: <4kdond$25b@saturn.genoa.com> Alex Blakemore <alex@genoa.com> writes > In <316942D0.6038@NeXT.com> Matt Watson wrote: > > > It is always faster and often more correct to check for nil before > > sending a message to an object which is possibly nil. > > What does "more correct" mean? It means that the output of your program is more likely to be correct.(what did you think it meant? :-)) Returning nil on messages to nil objects is *not* a benign feature of Objective-C. It's a definite rough edge. > > It's not that hard, happens less often than you'd expect, plus makes > > your program faster > > and your code more maintainable. > > I doubt the speed difference is noticeable in most cases. True enough, but since the check needs to be there in most cases anyway, it's nice to know that it makes your program faster, too. > Personally, I think relying upon this feature makes Objective-C code > much _more_ maintainable. > You can remove a lot of clutter from the code so that the essential > logic is unobscured. > > Just remember to > a. make sure your methods handle nil arguments gracefully > b. handle those methods that return floats, doubles and structs > explicitly. The big problem is this: Messaging a nil object causes a chain of useless results to propagate through your program. By the time you get to somewhere where ignoring the problem is no longer acceptable, you have no idea where the failure occurred. The "right thing" to do on messaging a nil object is throw an exception. This would preserve the advantage of streamlining the application logic, but provide a better, more structured way of dealing with the occasional nil message. Since neither of the two most popular Objective-C compilers do this, you're forced to explicitly check for invalid receivers. How does a nil object end up in your program? Usually it's from a method returning nil to indicate a FAILURE of one sort or another. You ignore these failures at your peril. This isn't much different in principle from checking the return value from C library functions - sure, that fprintf() will usually succeed, but wouldn't you rather know if something went wrong? Here's another way to look at it: What if I changed the behavior of the compiler and runtime so that messages to freed objects returned nil, instead of producing a run-time error? How about other arbitrary invalid addresses? Would programs written using that compiler be more or less likely to contain unknown errors? Would you want to use one to prepare your taxes? It basically comes down to the "do it right" versus "bull ahead regardless" schools of software design. One requires more effort to write programs that give the right result more often. The other makes it easy to create programs that keep running, but occasionally give bogus results... -- Mark Bessey NeXT Software, Inc Software Quality Assurance -->I DON'T SPEAK FOR NeXT!<--
From: dekorte@intrepid.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Salaries for NEXTSTEP programmers. Date: 9 Apr 1996 20:26:34 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4keh5q$4oq@news.onramp.net> References: <4k9pe3$c3u@Holly.aa.net> <4kchse$cqj@newsbf02.news.aol.com> In my experience, the rates are a little higher: In <4kchse$cqj@newsbf02.news.aol.com> BriMan101 wrote: NEXT Developer only academic experience $50k+/yr consulting, $40k+/yr perm > NEXT Developer with 1 year of experience - $30k to $45k. $60k-90k consulting, $50-$80k/yr perm > NEXT Developer with 3 years of experience - $50 to $75k. $100k+/yr consulting, ? permanent > NEXT Developer with 5 years experience - $75 to $90k. I've heard of as much as $150+/hr for a good consultant, ? permanent There also seems to be a great variation with skill, which tends to be more dependent on natural talent than experience after a year or so. And as far as experience goes, it's value seems to depend on what areas of NeXTstep you have experience with - for example, EOF experience is in high demand right now. -- Steve Dekorte NEXTSTEP/OPENSTEP consultant - Anaheim, CA mailto:dekorte@suite.com (NeXTmail, MIME welcome) http://www.batech.com/~dekorte/dekorte/index.html
From: synper1@aol.com (Syn Per 1) Newsgroups: comp.sys.next.programmer Subject: $105k per year plus stocks and benifits/ and live anywhere in the USA Date: 9 Apr 1996 19:06:40 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4keqi0$5rh@newsbf02.news.aol.com> Hello, My name is Matt Greedy, I am a technical recruiter who is looking of post sales consultants for the fastest growing system software company to date. They need people who have 4 solid years building client/server applications on relational databases, and who has used a 4gl; and who know Unix; and who have RDBMS knowlegde, and who know Data modelling, CASE, and Entity Relationship Diagrams. It would be helpful if you understand life cycle project planing, and have exposure to object oriented concepts and methodologies, but it is not mandatory. If you interested let me know, If you may know somebody pass the message on, I pay referal fees!!! Best Regards, M. Greedy
From: hugues@precipice.fdn.org (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 9 Apr 1996 17:33:02 GMT Organization: Individual - France Distribution: world Message-ID: <4ke70e$2bg@precipice.fdn.fr> References: <31696cb0.27573286@news.plsys.co.uk> In article <31696cb0.27573286@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: > But that still leaves us with the "hybrid world", dual hierarchy > problem. Some bits of your code will support autorelease, some won't, > for example. Could still be a bit of a mess I reckon, like it is at > the moment with EOF 1.1 projects. Same problem as with win/16 and win/32 :-) Welcome to the MS world... Bill will eat you all :-9 Hugues. -------------------------------------------------------------------- hugues@precipice.fdn.fr - France (small NeXTMail OK) ------------ NS3.2 ------------ NS3.0J ------------ :-) ------------
From: esky@marathon.cs.ucla.edu (Eskandar Ensafi) Newsgroups: comp.sys.next.programmer Subject: Using -unmounting:ok: (I can't get any results!) Date: 10 Apr 1996 04:20:37 GMT Organization: University of California, Los Angeles Message-ID: <4kfcul$bb5@delphi.cs.ucla.edu> Hello, I have been unable to implement -unmounting:ok: in any of my programs. I have written a few test programs under NEXTSTEP 3.3, using the floppy disk on an Intel system and the optical disk on a Motorola NeXTcube. In both cases, the Application's delegate can receive -app:mounted: and -app:unmounted: messages, but the -unmounting:ok: message is ignored. I tried both documented versions: - (int)app:sender unmounting:(const char *)fullPath; - (int)unmounting:(const char *)fullPath ok:(int *)flag; Why are there two different methods? In any case, -unmounting:ok: (or app:unmounting:) is never invoked! Is this a bug or am I overlooking something? These are the steps taken in my test programs: 1. I have a class MyObject which implements the following: - appDidInit:sender; - app:sender mounted:(const char *)fullPath; - app:sender unmounted:(const char *)fullPath; - (int)app:sender unmounting:(const char *)fullPath; - (int)unmounting:(const char *)fullPath ok:(int *)flag; 2. I instantiate MyObject in InterfaceBuilder and make it my Application's delegate by draggging a connection to the file's owner. I have only one NIB file, and the Application object is the file's owner. 3. I implement -appDidInit: in MyObject. It performs the following as part of initialization: [[Application workspace] beginListeningForDeviceStatusChanges]; 4. As expected, -app:mounted: and -app:unmounted work. However, -unmounting:ok: (or app:unmounting:) are never invoked. Why? Any assistance would be appreciated. Please e-mail me a copy of your response if possible. Thank you! - Eskandar ------------------------------------------------------------------------------ Eskandar Ensafi Object-Oriented Software Engineer University of California, Los Angeles Department of Biomathematics esky@cs.ucla.edu (MIME, NeXT) School of Medicine http://www.ph.ucla.edu/~esky ------------------------------------------------------------------------------ -- "Was ich kann und was ich konnte Weiss ich gar nicht mehr Gib mir wieder etwas schones Zieh mich aus dem Meer" -- Marian (The Sisters Of Mercy)
From: Luis Arias <luis@club-internet.fr> Newsgroups: comp.sys.next.programmer Subject: Calling Stored Procedures from EOF Date: Wed, 10 Apr 1996 10:10:03 +0200 Organization: Elysia, SARL Message-ID: <316B6CDB.2A2A@club-internet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, Does anyone have an example handy showing me how to call stored PL/SQL procedures from EOF (or even DBKIT) ? Thanks, -- Luis Arias Elysia, SARL 13, avenue Morane Saulnier 78140 VELIZY FRANCE [33] (1) 30 70 63 42 [33] (1) 34 65 36 79 fax info@elysia.com
From: martin@gecko.fb6.fhtw-berlin.de (Martin Bachmayer) Newsgroups: comp.sys.next.programmer Subject: Preview in EPS Date: 10 Apr 1996 10:23:17 GMT Organization: Fachhochschule fuer Technik und Wirtschaft Berlin, FB1 Message-ID: <4kg26l$4oh@commserv.fb1.fhtw-berlin.de> Hi, there. Does anybody have a clue how I can put a preview into an EPS file? Any hints appreciated! A piece of code would be heaven. Ha. Thanks. Bye. Martin
From: lchamp@mines1.info.emn.fr (Champciaux Laurent) Newsgroups: comp.sys.next.programmer Subject: 3DKit: problem with renderman object instancing Date: 10 Apr 1996 11:27:04 GMT Organization: Ecole des Mines de Nantes Distribution: world Message-ID: <4kg5u9$3ga@wfn.emn.fr> Hello, I am new to the 3D kit and I have a problem with renderman object instancing: In ri.h, RiObjectBegin(), RiObjectEnd() and RiObjectInstance() use RtToken in place of RtObjectHandle (cf. The Renderman Companion). What about that ? The object I want to instance is a simple cube (6 polygons) and my program crashes when using the statements above. Here is the method called in the renderSelf method of my N3DShape subclass: - drawTheire { int x, y, z; RtToken cube; RiAttributeBegin(); cube = RiObjectBegin(cube); // This call is bad, how to do it ? [self unitCube]; // 6 polygons, nothing else RiObjectEnd(); for (x = boundingBox[0]; x <= boundingBox[1]; x=x+inc) for (y = boundingBox[2]; y <= boundingBox[3]; y=y+inc) for (z = boundingBox[4]; z <= boundingBox[5]; z=z+inc) { if (mat[x+25][y+25][z+25]) { RiTransformBegin(); RiTranslate(x, y, z); RiObjectInstance(cube); RiTransformEnd(); } } RiAttributeEnd(); return self; } I use NeXTSTEP 3.0 on a black hardware. Any clue ? -- ======================================================= Laurent Champciaux Departement Informatique - Ecole des Mines de Nantes 4, rue A.Kastler 44070 Nantes Cedex 03 Tel: (33)51 85 82 18 (B220) email: Laurent.Champciaux@emn.fr WWW: http://www.emn.fr/dept_info/perso/laurent/
Newsgroups: comp.sys.next.programmer From: gery@ares.fdn.org(Gery_Divry) Subject: Re: Preview in EPS Message-ID: <1996Apr10.112928.3513@ares.fdn.org> Sender: news@ares.fdn.org Organization: ARES - Lyon, France. References: <4kg26l$4oh@commserv.fb1.fhtw-berlin.de> Date: Wed, 10 Apr 1996 11:29:28 GMT In article <4kg26l$4oh@commserv.fb1.fhtw-berlin.de> martin@gecko.fb6.fhtw-berlin.de (Martin Bachmayer) writes: > > Hi, there. > > Does anybody have a clue how I can put a preview into an EPS file? > Any hints appreciated! > A piece of code would be heaven. Ha. > > Thanks. Bye. > > Martin send to your view a message copyPSCodeInside:to: example: -saveEPS : (NXStream *)stream { [myView copyPSCodeInside:NULL to:stream]; return self; } Gery DIVRY ( ZZVolume Daddy ) ARES Publisher 8, rue Victor Lagrange Phone: (+33) 72 80 16 30 69007 LYON Fax: (+33) 72 80 16 32 France Email: gery@ares.fdn.org Earth, Solar System, Galaxy MW1 NeXT Mail accepted
Newsgroups: comp.sys.next.programmer From: gery@ares.fdn.org(Gery_Divry) Subject: Re: Preview in EPS Message-ID: <1996Apr10.134042.6360@ares.fdn.org> Sender: news@ares.fdn.org Organization: ARES - Lyon, France. References: <1996Apr10.112928.3513@ares.fdn.org> Date: Wed, 10 Apr 1996 13:40:42 GMT In article <1996Apr10.112928.3513@ares.fdn.org> gery@ares.fdn.org(Gery_Divry) writes: > In article <4kg26l$4oh@commserv.fb1.fhtw-berlin.de> > martin@gecko.fb6.fhtw-berlin.de (Martin Bachmayer) writes: > > > > Hi, there. > > > > Does anybody have a clue how I can put a preview into an EPS file? > > Any hints appreciated! > > A piece of code would be heaven. Ha. > > > > Thanks. Bye. > > > > Martin > > send to your view a message copyPSCodeInside:to: > > example: > > -saveEPS : (NXStream *)stream > { > [myView copyPSCodeInside:NULL to:stream]; > return self; > } > > Take care .. it will change in openstep Gery
Newsgroups: comp.sys.next.programmer From: gery@ares.fdn.org(Gery_Divry) Subject: Re: Sorting a Storage Object? Message-ID: <1996Apr10.061159.2789@ares.fdn.org> Sender: news@ares.fdn.org Organization: ARES - Lyon, France. References: <1996Apr7.114620.77792@cc.usu.edu> Date: Wed, 10 Apr 1996 06:11:59 GMT In article <1996Apr7.114620.77792@cc.usu.edu> writes: > Is a an instance of a Storage Class allocated contiguous memory > so that I can use qsort with it? > > Is there a quick and easy way to sort a Storage object? > > Thanks - > > - HRC - > hcole@spanky.idec.sdl.usu.edu the storage Object will disappear in OpenStep .. so it is a good opportunity to replace it by something else ... instead I used a lot and the best way to sort it is to index it ... otherwise you can use a subclass of Storage and then use the internal variable 'dataPtr'to get the beginning of the array in memory but be careful to get the value of this variable each time after an operation on the storage object .. the array can move at any moment. I used it to do a shellsort on the storage but ... it is not object programming ... it was a very poor solution Follow a storageObject subclasse who sort the elements when they are inserted .. (sorry the comments are in french) StorageOrdonne.h ----------------- // Gery DIVRY Fri Jun 12 16:56:29 GMT+0100 1992 #import <objc/Storage.h> @interface StorageOrdonne:Storage { @public BOOL unique; } - setUnique:(BOOL) etat; // methodes a sousclasser pour d finir la methode de comparaison // you should subclass this method to sort the storage - (int) compare: (void *) el1 : (void*) el2; // -1 si inferieur // 0 si egal // 1 si superieur // surd finir la methode insertAt pour interdir son emploi - insert:(void*)anElement at:(unsigned) index; // surdefinir la fonction addElement pour interdir son emploi - addElement:(void*)anElement ; // creer une fonction d ajout d un element // si le flag unique est YES alors l element n est pas ins r si un egal existe // et l index de l existant est retourne // sinon l element est ins r et son index retourn -(unsigned) ajouteElement:(void*)anElement; -(unsigned) dicotomeTranche:(unsigned) debut : (unsigned) fin : (void *) anElement; @end ------------------- StorageOrdonne.m ------------------- // Gery DIVRY Fri Jun 12 17:15:57 GMT+0100 1992 #import "StorageOrdonne.h" @implementation StorageOrdonne -setUnique:(BOOL) etat { unique=etat; return self; } // methode a sousclasser pour d finir la methode de comparaison - (int) compare: (void *) el1 : (void*) el2 // -1 si inferieur // 0 si egal // 1 si superieur { if((int)el1<(int)el2) return -1; else if((int)el1>(int)el2) return 1; else return 0; } // surd finir la methode insertAt pour interdir son emploi - insert:(void*)anElement at:(unsigned) index { return nil; } // surdefinir la fonction addElement pour interdir son emploi - addElement:(void*)anElement { return nil; } // creer une fonction d ajout d un element // si le flag unique est YES alors l element n est pas ins r si un egal existe // et l index de l existant est retourne // sinon l element est ins r et son index retourn -(unsigned) ajouteElement:(void*)anElement { // deux cas, le storage est vide ou pas int taille= [self count]; if(!taille) { // on ajoute de suite [super addElement: anElement]; return 0; } else { return [self dicotomeTranche: 0 :taille : anElement]; } } -(unsigned) dicotomeTranche:(unsigned) debut : (unsigned) fin : (void *) anElement { unsigned taille=fin-debut, newDebut=0,newFin=0; unsigned pivot=debut + (taille)/2; void *elPivot; int test; elPivot=[super elementAt:pivot]; // comparaison avec l element a inserer test=[self compare : anElement : elPivot]; switch (test) { case -1: newDebut=debut; newFin=pivot; if((newFin - newDebut)<1) { // alors on est sur l element juste inferieur //et on insere juste derriere le debut [super insertElement:anElement at:(newFin)]; return (newFin); } else return ([self dicotomeTranche:newDebut : newFin :anElement]); break; case 1: newDebut=pivot; newFin=fin; if((newFin - newDebut)<=1) { // alors on est sur l element juste inferieur //et on insere juste derriere le debut [super insertElement:anElement at:(newDebut+1)]; return (newDebut+1); } else return ([self dicotomeTranche:newDebut : newFin :anElement]); break; case 0: if(unique) { return pivot; } else { [super insertElement:anElement at:pivot]; return pivot; } break; } return pivot; } @end --------- Gery DIVRY ( ZZVolume Daddy ) ARES Publisher 8, rue Victor Lagrange Phone: (+33) 72 80 16 30 69007 LYON Fax: (+33) 72 80 16 32 France Email: gery@ares.fdn.org Earth, Solar System, Galaxy MW1 NeXT Mail accepted
From: wilkens@linkwood.linguistics.ruhr-uni-bochum.de (Rolf Wilkens) Newsgroups: comp.sys.next.programmer Subject: How to write from right to left Date: 10 Apr 1996 15:44:52 GMT Organization: Ruhr-Universitaet Bochum, Rechenzentrum Message-ID: <4kgl1k$gb@sun168.rz.ruhr-uni-bochum.de> Hi, I'm looking for some kind of Text Object where I can write from right to left in a Hebrew / Yiddish font (with NS 3.3). I've tried to subclass Text, but I have problems with the line breaks. All I know is that NXScanALine() is probably used to calculate the line breaks and that it is possible to replace this function. Where can I find documentation / examples of this function ? Thanks rolf wilkens wilkens@linguistics.ruhr-uni-bochum.de
From: rragner@panther.vm.iastate.edu (Rod Ragner) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 10 Apr 1996 16:47:52 GMT Organization: Iowa State University, Ames, Iowa Message-ID: <4kgono$772@news.iastate.edu> References: <4kdv9r$k9d@news.mr.net> In article <4kdv9r$k9d@news.mr.net> bmont@rconnect.com (Bruce Montegani) writes: > I have noticed that when I boot into Window95 on my machine that the clock is > 5 hours later than the clock when I boot under NeXT. When I fix either one > the other one is out. NEXTSTEP runs the clock at GMT and adjusts the time shown for the current time zone specified in Preferences. If you move to a different time zone and change preferences, then your documents will appear to have changed when they were last saved, but they really have not, since their time is based upon GMT. > Why can I not syncronize these two clocks? Because MS software is dumb. DOS/Windows knows nothing about time zones. In the MS world, you are everything and the rest of the world does not exist (at least it is not relative to MS). > Don't I have just one clock on my system? Yes, see above. It all depends upon how the software uses the clock (actually the chronomator). > Bruce Montegani There does exist a DOS program which someone wrote which will adjust the time in DOS to reflect the time zone that you specify, but I do not know what it is called or where it can be found. I run DOS/Windows applications, when I am forced to use them, in SoftPC... I actually have a DOS/WIndows partition, but I have not booted with it for more than a year. -- Rod Ragner, UNIX Systems Administrator/NEXTSTEP Application Developer Veterinary Diagnostic Laboratory, College of Veterinary Medicine, Iowa State University, 2630 Vet. Med. Bldg., Ames, Iowa 50011 Voice: (515) 294-4751, FAX: (515) 294-6961 or 3564, (NeXT Mail accepted) Email: rragner@stallion.vm.iastate.edu or stryder@iastate.edu
From: franke@boehme.dbag.ulm.DaimlerBenz.COM (Juergen Franke) Newsgroups: comp.sys.next.programmer Subject: Keyboard.app How to change the Character Code Palette? Date: 10 Apr 1996 06:39:19 GMT Organization: debis Network Services GmbH Message-ID: <4kfl2n$s1u@news.sns-felb.debis.de> -- Jörgen Franke Text Understanding Department F3M/T Daimler-Benz AG, Research Center Ulm P.O. Box 2360 89013 Ulm Germany Telephone Germany 731 505 2355 Fax Germany 731 505 4113 e-mail: franke@dbag.ulm.DaimlerBenz.COM
From: Haibo Chen <haiboc@bnr.ca> Newsgroups: comp.sys.next.programmer Subject: windows 95 Date: Wed, 10 Apr 1996 12:43:05 -0400 Organization: Hi-Wave Research (HWR) Message-ID: <316BE519.144B@bnr.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: haiboc@bnr.ca Hi! I would like to seek the help from all of you: I'd like to build a LAN between WINDOWS 95 (PC) to NeXTCube for sharing the NeXT printer. I am facing the workgroup configuration problem in windows 95 network configuration, which workgroup I should specify for NeXT? Any more information and help would appreciated. Thanks a lot. Haibo Chen haiboc@bnr.ca
From: Markus Pilzecker <markus.pilzecker@rhein-neckar.netsurf.de> Newsgroups: comp.sys.next.programmer Subject: Re: Dragging Matrix's Cells Date: Wed, 10 Apr 1996 20:18:37 +0000 Organization: GNI -- Internet fuer Odenwald, Rhein-Neckar, Karlsruhe, Kaiserslautern, Heilbronn und Pfaffenhofen Message-ID: <316C179D.677085FA@rhein-neckar.netsurf.de> References: <4kb1jj$p0d@imag.imag.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: markus.pilzecker@rhein-neckar.netsurf.de Benhur Stein wrote: > > Hi all, > I'd like to change the positions of cells in a matrix by dragging them, > just like one can do in the IB. If this is a feature of Matrix > itself, how do I enable it? And if it is not, does anybody know if > this is implemented in any example or public domain software? > > Thanks for any help, > benhur Hallo Benhur, I have exactly the same problem. But as far is I understand Matrix, it is not possible to override its decisions about the positions of its Cells. It would be a fine thing, if NEXT would put position computation of Matrix into a method, so that we can override it, if we want. I don't yet know of a library that does that job (so if You get personal mail, concerning this problem, it would be nice to hear of You), but my idea is, to subclass the desired Cell class with an overridden method drawSelf:inView that does not take the position from the given frame parameter, but from an additional local position variable, which could be set by methods that we define for that purpose. Ciao, Markus ------------------------------------------------------------------------------- Sackpost: Markus Pilzecker voice: +49 6251 39415 Am Wechsel 9a V34,V17: +49 6251 39575 D-64625 Bensheim email: markus.pilzecker@rhein-neckar.netsurf.de -------------------------------------------------------------------------------
From: pbrown@palms.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: How to write from right to left Date: 10 Apr 1996 21:00:51 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4kh7i3$2k5@agate.berkeley.edu> References: <4kgl1k$gb@sun168.rz.ruhr-uni-bochum.de> In article <4kgl1k$gb@sun168.rz.ruhr-uni-bochum.de>, Rolf Wilkens <wilkens@linkwood.linguistics.ruhr-uni-bochum.de> wrote: >I'm looking for some kind of Text Object where I can write >from right to left in a Hebrew / Yiddish font (with NS 3.3). I can't speak to the programming aspects of your question, but if you're just looking for a L->R typesetting package or functionality, you might consider building TeX-XeT (a dyslexic TeX) for your machine. Paul
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer, Subject: Re: Anyone want a SLIP up/down app? Date: 10 Apr 1996 21:19:36 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4kh8l8$1gf1@core.bard.edu> References: <DpK90w.Gzp@waldo.com> I was wondering what SLIP package you are using. I have a Next/intel machine that I would like to run SLIP on, but Mamakos' SLIP_904??? package is for black hardware. Know of any quad-fat, or at least intel SLIP's for Next? Thanks in advance...HC -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.misc Subject: NEXTSTEP Java Porting Issues-mailing list! Date: Wed, 10 Apr 1996 17:19:41 -0500 Organization: Illinois State University, Instructional Technology Services Message-ID: <316C33FD.308C@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NEXTSTEP Java Porting Issues Sponsor: Bill Bumgarner (bbum@friday.com) next-java@friday.com An open mailing list for for discussion of porting and integration esoteria that are unique to the NeXT platform. To subscribe, send the word "subscribe" in the BODY of a message to next-java-request@friday.com. ------------------------------------------------ Eric A. Dubiel; http://www.ilstu.edu/~eadubie Instructional mailto:eadubie@rs6000.cmp.ilstu.edu Technology PEACE LOVE Services UNITY RESPECT Illinois State University "Understanding is best learned via experience." "Follow our instinct, not a trend. Go against the grain until the end." VIEWS EXPRESSED ARE ENTIRELY MY OWN
From: everhart@alterlife.com (Dwight Everhart) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 11 Apr 1996 00:25:45 GMT Organization: IntelliNet-ISP, LLC. 1-800-290-7677 Message-ID: <4khji9$mt4@zeus.intellinet.com> References: <4kdv9r$k9d@news.mr.net> In <4kdv9r$k9d@news.mr.net> Bruce Montegani wrote: > I have noticed that when I boot into Window95 on my machine that the clock is > 5 hours later than the clock when I boot under NeXT. When I fix either one > the other one is out. > > Why can I not syncronize these two clocks? > > Don't I have just one clock on my system? NEXTSTEP keeps its time internally in GMT. I think Windows 95 keeps it in local time. -- Dwight Everhart "Spring will come for IBM when people Ex Nicole Bobek Coach start buying big computers again." Bella Vista, Arkansas -- John Akers, ex-CEO of IBM everhart@alterlife.com
From: everhart@alterlife.com (Dwight Everhart) Newsgroups: comp.sys.next.programmer Subject: Re: Sorting a Storage Object? Date: 11 Apr 1996 00:30:34 GMT Organization: IntelliNet-ISP, LLC. 1-800-290-7677 Distribution: world Message-ID: <4khjra$mt4@zeus.intellinet.com> References: <1996Apr7.114620.77792@cc.usu.edu> In <1996Apr7.114620.77792@cc.usu.edu> Howard R. Cole wrote: > Is a an instance of a Storage Class allocated contiguous memory > so that I can use qsort with it? > > Is there a quick and easy way to sort a Storage object? The MiscKit has a subclass of Storage called MiscSortedStorage. It may do what you need. -- Dwight Everhart "Spring will come for IBM when people Ex Nicole Bobek Coach start buying big computers again." Bella Vista, Arkansas -- John Akers, ex-CEO of IBM everhart@alterlife.com
From: Ravi Mendis <lady0098@sable.ox.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: WebObjects' transaction integrity? Date: Wed, 10 Apr 1996 17:59:57 +0100 Organization: Oxford University Message-ID: <Pine.OSF.3.91.960410174843.10337B-100000@sable.ox.ac.uk> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII What mechanism is in place that ensures information (transactions) aren't lost over the Web? (We all know that sometime our Web pages don't download properly: sometimes with errors; images in particular tend to 'get lost') Isn't this a commonly asked question by NeXT customers? They wouldn't want to implement business critical applications otherwise. On the plus side, WebObjects' client/SERVER (Server intense) approach has an edge over CLIENT/SERVER strategies like Sun's JOE, looking at things from the perspective of better transaction integrity. But still, i have found little information on NeXT's WebSite regarding this issue. And i'm not refering to transaction security! ravi
From: pete@ohm.york.ac.uk (-bat.) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 10 Apr 1996 15:07:53 GMT Organization: The University of York, UK Message-ID: <4kgis9$ltk@netty.york.ac.uk> References: <4kdv9r$k9d@news.mr.net> bmont@rconnect.com (Bruce Montegani) writes: > I have noticed that when I boot into Window95 on my machine that the clock is > 5 hours later than the clock when I boot under NeXT. When I fix either one > the other one is out. Probably because UNIX stores the time in GMT and then adds the local adjustment whereas windown95 stores the local time directly (though ti does seem to do summertime correctly) Workaround: come and live in the UK :-) -bat.
From: pete@ohm.york.ac.uk (-bat.) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 10 Apr 1996 15:11:39 GMT Organization: The University of York, UK Message-ID: <4kgj3b$lu2@netty.york.ac.uk> References: <1996Apr8.071349.18229@seer.demon.co.uk> Paul Lynch <Paul_Lynch@plsys.co.uk> writes: > scripts won't be enough to convert your code; about one man month per > 50,000 lines of code will be required, he said. This is a lot of effort > (particularly if it takes longer than that :-). Ouch. I hadn't seen that before. Preseumably this is just for the frontend code as well. Those of us who have a significant amount of UNIXy code underneath all the Objective C will have an even longer job I guess :-( > to be extensive. With OpenStep, I expect that will no longer be the case, > and testing requirements for OpenStep/NT, OpenStep/Solaris and NeXTSTEP > will be much more gruelling. I really really hope not - it would make things a lot harder and maybe make me think twice about actually moving to OpenStep anyway. Mind you, the single source tree hasn't always worked fine in the past either. I never managed to get an App which runs fine on intel and 68000 to work under HP no matter what I did. Sigh... -bat.
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 11 Apr 1996 03:54:32 GMT Organization: ALI Message-ID: <4khvpp$6ce@cetus.ali.bc.ca> References: <4kdv9r$k9d@news.mr.net> In comp.sys.next.programmer Bruce Montegani wrote: > I have noticed that when I boot into Window95 on my machine that the clock is > 5 hours later than the clock when I boot under NeXT. When I fix either one > the other one is out. > > Why can I not syncronize these two clocks? > > Don't I have just one clock on my system? The problem is that NeXTSTEP (like all Unices) wants to store Greenwich Mean Time in the system clock and use the assigned time zone to calculate the local time as needed. Windows and DOS on the otherhand, want to store the local time in the system clock and use the assigned time zone to calculate Greenwich Mean Time (and others) as needed. The advantage of the Windows and DOS approach is that you save a few computing cycles when returning the local time. The disadvantage is that whenever you switch in and out of daylight savings time or relocate the computer to a different time zone, it has to change the value stored in the system clock. This in turn can cause problems for programs which depend on the system clock as an absolute time base which doesn't jump around. Thus, to solve your problem, set the time correctly for whichever OS you run most of the time, and then change the time zone of the other to compensate. Expect to handle some weirdness around daylight savings switchovers. -- Allan Noordvyk, Software Artisan e-mail: allan@ali.bc.ca ALI Technologies Voice: 604.279.5422 x 317 Richmond, Canada Fax: 604.279.5468 * NeXT and MIME mail welcome *
Newsgroups: comp.sys.next.programmer From: gery@ares.fdn.org(Gery_Divry) Subject: Re: Dragging Matrix's Cells Message-ID: <1996Apr11.054916.9482@ares.fdn.org> Sender: news@ares.fdn.org Organization: ARES - Lyon, France. References: <316C179D.677085FA@rhein-neckar.netsurf.de> Date: Thu, 11 Apr 1996 05:49:16 GMT In article <316C179D.677085FA@rhein-neckar.netsurf.de> Markus Pilzecker <markus.pilzecker@rhein-neckar.netsurf.de> writes: > Benhur Stein wrote: > > > > Hi all, > > I'd like to change the positions of cells in a matrix by dragging them, > > just like one can do in the IB. If this is a feature of Matrix > > itself, how do I enable it? And if it is not, does anybody know if > > this is implemented in any example or public domain software? > > > > Thanks for any help, > > benhur > > Hallo Benhur, > > I have exactly the same problem. But as far is I understand Matrix, it is > not possible to override its decisions about the positions of its Cells. > It would be a fine thing, if NEXT would put position computation of Matrix > into a method, so that we can override it, if we want. > > I don't yet know of a library that does that job (so if You get personal mail, > concerning this problem, it would be nice to hear of You), but my idea is, to > subclass the desired Cell class with an overridden method drawSelf:inView that > does not take the position from the given frame parameter, but from an additional > local position variable, which could be set by methods that we define for that > purpose. > > > Ciao, Markus > > Hello ... you should use the niftyMatrix By Jayson Adams which do the move of the cells when the ctrl key is depressed Gery DIVRY ( ZZVolume Daddy ) ARES Publisher 8, rue Victor Lagrange Phone: (+33) 72 80 16 30 69007 LYON Fax: (+33) 72 80 16 32 France Email: gery@ares.fdn.org Earth, Solar System, Galaxy MW1 NeXT Mail accepted
From: "Timothy J. Luoma" <luomat@capitalist.princeton.edu> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Re: Anyone want a SLIP up/down app? Date: Tue, 9 Apr 1996 12:20:16 -0400 Organization: Princeton University Message-ID: <Pine.NXT.3.92.960409121121.12749C-100000@capitalist.princeton.edu> References: <DpK90w.Gzp@waldo.com> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII To: Matthew Hocker <hocker@waldo.com> In-Reply-To: <DpK90w.Gzp@waldo.com> I don't use SLIP, so I don't know what exactly you have to go through to turn it on once you have it all setup. For PPP it is as easy as typing 'pppup' in a Shell. If SLIP is the same way, then I see this as a time for TickleServices. Say you have to do 'slipup' or 'slipdown'. In TickleServices, you can write a Service which will do this as easily as: exec /usr/local/bin/slipup name it 'SLIP connect' (or whatever) save, and check the 'services' part of the menus. Click on 'SLIP connect' and it will execute 'slipup'. The same thing is possible with bringing the connection down. I don't say this to mean "Geez, you should have done this before making an entire app". If you can program (which you obviously can) and you have the developer tools, that's great. TickleServices is available to those of us who can't program, and to those who do not have the developer/compiling tools needed. It also allows you to do these little tasks without needing a separate application running (the main TickleServer.daemon runs on login). I'd encourage people to check it out. See the most excellent Scott Hess' Home Page http://www.winternet.com/~shess and TickleServices page http://www.winternet.com/~shess/TickleServices.html TjL -- Timothy J. Luoma <luomat@capitalist.princeton.edu> USENET: Please CC this address on replies to my posts, things vanish fast. NeXT Info available via email! Send message with SUBJECT 'send info' >>>> Please: No NeXTMail, use MIME for attachments <<<<
From: rputtkam@iiic.ethz.ch (Roman Puttkammer) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 11 Apr 1996 06:32:25 GMT Organization: Dept. Informatik, Swiss Federal Institute of Technology Message-ID: <4ki91p$d9u@neptune.ethz.ch> References: <4kdv9r$k9d@news.mr.net> <4kgono$772@news.iastate.edu> Rod Ragner (rragner@panther.vm.iastate.edu) wrote: : ... I actually have : a DOS/WIndows partition, but I have not booted with it for more than a : year. Congratulations! I wish, I could say the same ;-) roman
From: pbrown@palms.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 11 Apr 1996 09:15:30 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4kiiji$mhi@agate.berkeley.edu> References: <4kdv9r$k9d@news.mr.net> <4kgono$772@news.iastate.edu> <4ki91p$d9u@neptune.ethz.ch> In article <4ki91p$d9u@neptune.ethz.ch>, Roman Puttkammer <rputtkam@iiic.ethz.ch> wrote: >Rod Ragner (rragner@panther.vm.iastate.edu) wrote: >: ... I actually have >: a DOS/WIndows partition, but I have not booted with it for more than a >: year. >Congratulations! I wish, I could say the same ;-) Windows '95 is the cause of "mad cow" disease in humans... you should be careful... pb
From: tal@cs.caltech.edu (Tal Lancaster) Newsgroups: comp.sys.next.programmer Subject: Re: renderman on nextstep Date: 11 Apr 1996 04:57:32 GMT Organization: California Institute of Technology, Pasadena Distribution: world Message-ID: <4ki3fs$4u9@gap.cco.caltech.edu> References: <4jram3$b18@malibu.unice.fr> <4jtcas$t53@brachio.zrz.TU-Berlin.DE> In <4jtcas$t53@brachio.zrz.TU-Berlin.DE> Dirk Schwarzhans wrote: > Hello, > > On 04/02/96, Mathieu Olivier wrote: > > > >Hi All, > > > >does anybody know, where i can find "The RenderMan Companion" ??? > >it is a book who deals with the 3DKit on NEXTSTEP. ( i hope !!) > > NeXT's front end to RenderMan, the 3D Kit, is not the topic of "The > RenderMan Companion"! > It "only" covers the low level RenderMan language (as far as I know). This is true the RenderMan Companion only covers RenderMan. Nothing on 3D Kit. To the best of my knowledge there isn't any books dealing with 3D Kit. But if you really want to get down and get dirty with 3D Kit, take a look at WavesWorld: (http://www.media.mit.edu/people/wave/) The distribution contains the full source. And there are a lot of 3D Kit usage in there. This is in my belief shows off really what NeXTSTEP and 3D Kit can really do. Tal -- *********************************************************************** Tal Lancaster - tal@cs.caltech.edu (NeXTSTEP, MIME, and regular Email The RenderMan Repository -- http://pete.cs.caltech.edu/RMR/ ************************************************************************
From: matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) Newsgroups: comp.sys.next.programmer Subject: pswrap programming: How to transfer large image data? Date: 11 Apr 1996 13:11:03 GMT Organization: debis Network Services GmbH Message-ID: <4kj0d7$4to@news.sns-felb.debis.de> With my very limited knowledge of PostScript I'm trying to write a pswrap for imaging data using the Indexed feature (I think the AppKit itself does not support it). This is the solution I finally came up with: defineps drawIndexedImage( char colors[nc]; int nc; char pic[np]; int np, w, h) /ncolors nc 3 idiv def /picstr w string def /infile (pic) 0 () /SubFileDecode filter def [/Indexed /DeviceRGB ncolors 1 sub (colors) ] setcolorspace <</ImageType 1 /Width w /Height h /BitsPerComponent 8 /Decode [0 ncolors] /ImageMatrix [w 0 0 h neg 0 h] /DataSource {infile picstr readstring pop} >> image endps The central point is that I convert my string variable 'pic' which holds the image data into a file (see definition of 'infile'). The whole code works in principal. But, alas, if the image is quite large, the program encounters a limit check error, probably because PostScript string variables cannot hold arbitrary many bytes. I really spent some time on this before I finally gave up. So, hey, PostScript wizards, what would be the real way for solving my task? Thanks in advance! Matthias -- ____ __ _/_/ . ( / / ( / / / / ===================================================================== Matthias Oberlaender, Daimler-Benz AG, Research Center Ulm F3M/T (Information Technology / Text Understanding) Wilhelm-Runge-Str. 11, P.O. Box 2360, 89013 Ulm, Germany Phone: +49 731 505 2354 Fax: +49 731 505 4113 Email: oberlaender@dbag.ulm.DaimlerBenz.COM =====================================================================
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.hardware Subject: Full Duplex Sound Card Capability Under NS? Date: 11 Apr 1996 16:16:23 GMT Organization: monoChrome, Inc. Message-ID: <4kjb8n$q2k@news4.digex.net> Other than on NeXT hardware, is there NEXTSTEP compatible platform that can handle full duplex sound? That is, to be able to play and record sound simultaneously. This ability is critical to be able to implement certain tools, like a full duplex internet phone. The old NeXT hardware could handle this no problem...however, NS on intel processors doesn't seem to be able to do this. Well, at least not with either a SoundBlaster 16 or a PAS16. So, is there some way to do this under NS? Has anyone tried any of the following: a) SoundBlaster 32 (and higher, I believe have full duplex capabilities). The question is, do the full duplex features work w/ the regular NS driver, and if not, is a driver available that does support those features? b) Is it possible to put in multiple sound blaster 16's to get this feature and just load up multiple instances of the SB16 driver? If so, how can I be sure which card I'm record and which card I'm playing from, if for example, I run one copy of SoundEditor.app for playing and another for recording? c) Are any drivers or cards currently, or will become available, that can do this? Thanks very much for any help :) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Dragging Matrix's Cells Date: 11 Apr 1996 18:03:20 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4kjhh8$avl@lion.embl-heidelberg.de> References: <4kbu4j$2n7@digifix.digifix.com> In article <4kbu4j$2n7@digifix.digifix.com> sanguish@digifix.com (Scott Anguish) writes: > On 04/08/96, Benhur Stein wrote: > >Hi all, > >I'd like to change the positions of cells in a matrix by dragging > them, > >just like one can do in the IB. If this is a feature of Matrix > >itself, how do I enable it? And if it is not, does anybody know if > >this is implemented in any example or public domain software? > > > > Its not a feature of the default Matrix class. > > There is a MiniExample from NeXT called ScrollDoodScroll that > implements this. > > Its on ftp.next.com or available via nextanswers I wrote one for the new version of MiscKit but it will work only with NS4.0 It contains also a scroll view in a pallet. It's based on NiftyMatrix ... of course ... but one can rearrange columns either. -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: How to write from right to left Date: 11 Apr 1996 18:07:44 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4kjhpg$b07@lion.embl-heidelberg.de> References: <4kgl1k$gb@sun168.rz.ruhr-uni-bochum.de> In article <4kgl1k$gb@sun168.rz.ruhr-uni-bochum.de> wilkens@linkwood.linguistics.ruhr-uni-bochum.de (Rolf Wilkens) writes: > Hi, > I'm looking for some kind of Text Object where I can write > from right to left in a Hebrew / Yiddish font (with NS 3.3). > > I've tried to subclass Text, but I have problems with > the line breaks. All I know is that NXScanALine() is > probably used to calculate the line breaks and that it is > possible to replace this function. Where can I find > documentation / examples of this function ? I wrote my own semitic Text object ;-) If you give me hebrew fonts, I'll fix and submit my (free) semitic editor.... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: Robert Worne <rworne@primenet.com> Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 11 Apr 1996 12:13:01 -0700 Organization: Primenet Services for the Internet Sender: root@primenet.com Message-ID: <4kjljt$f4k@nnrp1.news.primenet.com> References: <4kdv9r$k9d@news.mr.net> bmont@rconnect.com (Bruce Montegani) wrote: >I have noticed that when I boot into Window95 on my machine that the clock is >5 hours later than the clock when I boot under NeXT. When I fix either one >the other one is out. > >Why can I not syncronize these two clocks? > >Don't I have just one clock on my system? This is due to a discrepancy on what you system clock is perceived as by 'doze 95 and NeXTSTEP. A utility exists on the archives that will synch the clock when NeXTSTEP is booted, and switch it back when it's shut down. I don't know the name of it specifically, but someone here must use/have it. My black HW makes all this unnecessary :) -- Robert Worne NeXT-OS/2-MacOS Starving CS Undergrad...Sorry, I don't *do* Windows! I'd rather starve... -----BEGIN GEEK CODE BLOCK----- Version: 3.1 GCS/P/S d-?>pu s+:+> a- C++$>++++ UX++++>$ P+>+++ L+ !E !W++ N+++ !o-- !K w--- O++$ M+ V PS>--- !PE+ Y+ !PGP- t@ 5++ X+++ R- tv b+>++ DI !D G e>+++ h--- r++ y+++** ------END GEEK CODE BLOCK------
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer,comp.sys.next.hardware Subject: Re: Full Duplex Sound Card Capability Under NS? Date: 11 Apr 1996 19:55:08 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4kjo2s$sis@core.bard.edu> References: <4kjb8n$q2k@news4.digex.net> >a) SoundBlaster 32 (and higher, I believe have full duplex capabilities). >The question is, do the full duplex features work w/ the regular NS driver, >and if not, is a driver available that does support those features? We have a SB AWE which I believe is 32 bit. But the only available drivers is for SB16 and it does not support anything more. >c) Are any drivers or cards currently, or will become available, that can >do this? You should look into the i*link i56 card. It is made a German company, i*link, whose main product is mix, a telecommunications package for NextStep. They made the i56 card so that their software could be used on intel machines. It is designed very similarly to NeXT DSP setup, from what I hear, except for the DAC/ADC. -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
Newsgroups: comp.sys.next.programmer,comp.lang.objective-c From: jpanico@netcom.com (Joe Panico) Subject: Re: when an expression evaluates to nil Message-ID: <jpanicoDppG9I.KAy@netcom.com> Followup-To: comp.sys.next.programmer,comp.lang.objective-c Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> Date: Thu, 11 Apr 1996 15:56:06 GMT Sender: jpanico@netcom.netcom.com Mark Bessey (MaRK_BeSSeY@NeXT.CoM) wrote: : Alex Blakemore <alex@genoa.com> writes : The big problem is this: Messaging a nil object causes a chain of : useless results to propagate through your program. By the time you get : to somewhere where ignoring the problem is no longer acceptable, you : have no idea where the failure occurred. : The "right thing" to do on messaging a nil object is throw an exception. : This would preserve the advantage of streamlining the application logic, : but provide a better, more structured way of dealing with the occasional : nil message. uhmm, if this is the case, then why do NeXT software engineers themselves message nil objects? For instance, if you ask EOModeler to produce a template EO, the accessor methods will use constructs like: - (void)setTYPE_CODE:(NSString *)value { [TYPE_CODE autorelease]; TYPE_CODE = [value retain]; } Please clarify. -- jpanico@netcom.com Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc.
From: rickg@sun.com (Richard Goldstein) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 11 Apr 1996 13:02:41 -0700 Organization: SunSoft, Inc. Sender: rickg@lamancha Message-ID: <ku7u3yqedge.fsf@lamancha> References: <4kj0d7$4to@news.sns-felb.debis.de> In-reply-to: matti@dbag.ulm.DaimlerBenz.COM's message of 11 Apr 1996 13:11:03 GMT From: matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) Newsgroups: comp.sys.next.programmer Date: 11 Apr 1996 13:11:03 GMT Organization: debis Network Services GmbH With my very limited knowledge of PostScript I'm trying to write a pswrap for imaging data using the Indexed feature (I think the AppKit itself does not support it). This is the solution I finally came up with: defineps drawIndexedImage( char colors[nc]; int nc; char pic[np]; int np, w, h) /ncolors nc 3 idiv def /picstr w string def /infile (pic) 0 () /SubFileDecode filter def [/Indexed /DeviceRGB ncolors 1 sub (colors) ] setcolorspace <</ImageType 1 /Width w /Height h /BitsPerComponent 8 /Decode [0 ncolors] /ImageMatrix [w 0 0 h neg 0 h] /DataSource {infile picstr readstring pop} >> image endps The central point is that I convert my string variable 'pic' which holds the image data into a file (see definition of 'infile'). The whole code works in principal. But, alas, if the image is quite large, the program encounters a limit check error, probably because PostScript string variables cannot hold arbitrary many bytes. I really spent some time on this before I finally gave up. So, hey, PostScript wizards, what would be the real way for solving my task? Correct, the given max for a PS string object on a 32bit architecture is 64k and you'd actually run a tad short of that given the BOS encoding. The answer is that you invoke the image operator, similar to above, and follow it up with the data via DPSWriteData. The "Modello" library was just released for X/DPS and C++, which does this kind of stuff for you, without having to resort to pswrap. It would be an easy port, for the most part, to NeXT DPS assuming a decent C++ compiler. If interested, its on ftp.x.org /contrib/devel_tools/DPS/. rick -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "... push the button, Frank." - Dr. Clayton Stonewall Forresster Richard M. Goldstein rickg@Eng.Sun.COM Desktop Performance Group SunSoft, Inc.
From: bmont@rconnect.com (Bruce Montegani) Newsgroups: comp.sys.next.programmer Subject: Adjusting text field size after changing font Date: 11 Apr 1996 20:43:33 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <4kjqtl$gjq@news.mr.net> I would like to know programmatically how to adjust the height of a text field after I have changed the font size. I have noticed that in Interface Builder that when you change the font from a smaller size to a larger size the field height adjusts accordingly. By the way the field will be empty when I change the font size. Thanks Bruce Montegani
From: wrb@biostr.washington.edu (William Barker) Newsgroups: comp.sys.next.programmer Subject: DBKit-Foundation problem? Date: 11 Apr 1996 20:53:21 GMT Organization: University of Washington Message-ID: <4kjrg1$ha5@nntp5.u.washington.edu> I've got a DBKit program that refuses to start up since installing the FoundationKit patch on 3.2 and 3.3 NeXT (black) systems. The app's behavior is the same on 3.2 and 3.3 systems--the program begins to initialize, then dies. The only message I get on the console is: KeyDB[3713]: objc: FREED(id): message propertyNamed: sent to freed object=0xf5284 KeyDB is the name of the app. The code that seems to be generating the problem is: [[KeyDBModule entity] propertyNamed:"lname"]; where "KeyDBModule" is a DBModule object. I do not have this problem on 3.2 systems that have not been upgraded with the FoundationKit patch. Are there any know problems with the DBKit and the FoundationKit patch? I've checked NextAnswers, but so far no joy. Thanks in advance. bb -- Bill Barker Biological Structure, Box 357420 University of Washington Seattle WA 98195-7420 (206) 543-7315 "In Wine there is Wisdom, In Beer there is Strength; In Water is Bacteria." --Old German Saying.
From: "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: Canon NEXTSTEP Developer Support-Japan Date: Thu, 11 Apr 1996 17:27:39 -0500 Organization: Illinois State University, Instructional Technology Services Message-ID: <316D875B.6E2C@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Has anyone seen the Canon NEXTSTEP Developer Support web pages in Japan? http://www.cnds.canon.co.jp/Japanese_EUC/WelcomeJ.html I can't understand the attitude of American Canon VS. Japanese Canon. I wish the two would truly work together and also fully translate the web pages to English and provide object.station driver downloads/updates! I am suprised that a company like Canon would drop a product that was so expensive and leave users to fend for themselves...It's so hard to put faith in companies- people too for that matter since they make up companies. Although, I have MUCH MORE RESPECT for the Japanese divisions due to their greater attention to quality and support (my opinions). side note: Has anyone also noticed that a lot of new activity/apps are coming out of Germany lately? There seems to be a lot of interest there... I would really appreciate feedback from NeXT users and Canon (although they've been completely unresponsive) ------------------------------------------------ Eric A. Dubiel; http://www.ilstu.edu/~eadubie Instructional mailto:eadubie@rs6000.cmp.ilstu.edu Technology PEACE LOVE Services UNITY RESPECT Illinois State University "Understanding is best learned via experience." "Follow our instinct, not a trend. Go against the grain until the end." VIEWS EXPRESSED ARE ENTIRELY MY OWN
From: craigh@pswtech.com Newsgroups: comp.sys.next.programmer Subject: Re: Dragging Matrix's Cells Date: 12 Apr 1996 05:15:46 GMT Organization: PSW Technologies Message-ID: <4kkou2$kon@digdug.pswtech.com> References: <4kb1jj$p0d@imag.imag.fr> <316C179D.677085FA@rhein-neckar.netsurf.de> >Benhur Stein wrote: > > Hi all, > I'd like to change the positions of cells in a matrix by dragging them, > just like one can do in the IB. If this is a feature of Matrix > itself, how do I enable it? And if it is not, does anybody know if > this is implemented in any example or public domain software? > > Thanks for any help, > benhur This functionality is demonstrated in the ScrollDoodScroll example that is distributed with NEXTSTEP 3.3. Check out the NiftyMatrix class (implemented by Jayson Adams I think) - it should show you how to do this. If you can't find it, let me know and I'll mail the code to you... Craig craigh@pswtech.com
From: "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: NS VIDEO CONFERENCING SERVER Date: Fri, 12 Apr 1996 01:24:26 -0500 Organization: Illinois State University, Instructional Technology Services Message-ID: <316DF71A.417D@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Use your NEXTSTEP box to serve as a videoconferencing server for MAC and WIN, two options I know of are: QuickTime Conferencing Internet Reflector http://www.devtools.apple.com/qtc/ The Reflector is a one-to-many program which distributes QTC audio and video streams over the Internet. It is a temporary solution to the lack of wide MBone deployment. In the absence of a multicast protocol, the Reflector is the best way to alleviate the network load on a QTC media source. Currently, qtcd runs on the following dialects of Unix: Solaris, SunOS 4.3, Irix, AIX, Linux, FreeBSD, and NeXTStep. It is little-endian friendly. If you have a platform you'd like to see the reflector run on, send mail. Also: White Pine CU-SEEME Video Conferencing Reflectors http://goliath.wpine.com/reflectdemo.htm and/or http://cu-seeme.cornell.edu/#Reflector Download a demo version of White Pine's Enhanced CU-SeeMe Reflector software. White Pine's unique Reflector technology accepts multiple CU-SeeMe connections and reflects the video, audio, and additional data to multiple participants concurrently -- providing an environment for group videoconferencing, document sharing, large audience broadcasts or just plain chat. Available soon for Windows NT. * NEXT OS Version 3.2 (reflector-4.0-b3-next.tar.Z) ftp://ftp.wpine.com/pub/product/demo/reflector/4.0B3/reflector-4.0-b3-next.tar.Z or ftp://gated.cornell.edu/pub/video/Reflector/4.00b3.dist/reflector-4.0-b3-next.tar.Z ------------------------------------------------ Eric A. Dubiel; http://www.ilstu.edu/~eadubie Instructional mailto:eadubie@rs6000.cmp.ilstu.edu Technology PEACE LOVE Services UNITY RESPECT Illinois State University "Understanding is best learned via experience." "Follow our instinct, not a trend. Go against the grain until the end." VIEWS EXPRESSED ARE ENTIRELY MY OWN
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 11 Apr 1996 22:26:05 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4kk0tt$68d@usenet.rpi.edu> References: <4juvbc$2i4@nova.mdd.comm.mot.com> <4k0uuo$mub@shelob.afs.com> <4k2gfs$9j4@usenet.rpi.edu> <31668a63.4169258@news.plsys.co.uk> <DpGt2n.KDs@news2.new-york.net> <31696ca4.27561865@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: > > I remember reading some magazine articles why NeXT decided to > > go with a new NSObject heirarchy, and those reasons did seem > > reasonable at the time I was reading it. I wasn't *happy* > > about it, but I did think some good reasons were presented, > > and that they had thought about things before deciding to make > > that major of a change. > > > > I think that article (or articles?) are at work though, and > > I'm at home right now (I've finally given up on the usenet hub > > at RPI, and am switching my news reading to home for a little > > while). > Do dig it out and post them if you can. Well, there's one in the OpenStep Journal for Spring 1995. Looks like it's NeXTanswer #1989 I've only had time to look over it very quickly, but it seems to me there might be another article around somewhere with some other information on the new heirarchy. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 11 Apr 1996 23:02:33 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4kk329$68d@usenet.rpi.edu> References: <4k2gfs$9j4@usenet.rpi.edu> <4kcict$2qo@nova.mdd.comm.mot.com> besler@mdd.comm.mot.com (Steve Besler) wrote: > Garance A Drosehn <gad@eclipse.its.rpi.edu> writes: > > >The whole **-POINT-** of object-oriented technology is code reuse. > > While reuse is a large benefit of OO, that is certainly not > the whole point. Hmm. It seems to me that most of the reasons for going with OO programming (as opposed to competent programming via some other methodolgy) would come under the heading of reliable code reuse. Shrug. At the very least, code reuse shouldn't be listed as some kind of evil thing to want to do, not in world of OOP. > > The issue at the heart of this discussion is not "Does NeXT do > > nasty things to people it should be working with?", the issue > > is "Does Object-Oriented Technology live up to it's claims?". > > No, I don't think that is what this thread is about. It's really > about whether NeXT should have maintained backward compatability > with NEXTSTEP in OpenStep. One way they could have done this > was to build all the new functionality to into the existing Object > hierarchy. I think you're jumping the gun there. First we have to determine if perhaps they really did need to have a new heirarchy -- and if so, then "Why?". Some people believe that they could have had all the benefits of the new technology without the new heirarchy, but I don't think that's really been determined yet. Yes, you can implement autorelease without a new heirarchy, but OpenStep has changed a lot more than just that one pittly little issue. Maybe the change to Unicode is enough to require a whole new heirarchy, maybe it's other things. I have yet to see anyone go thru all the openstep changes and prove that *all* of them could have been done in the older heirarchy. Maybe the ones which require the new heirarchy are ones that many would rather do without (if for no other reason than to avoid the new heirarchy). I'm still quite open to the suggestion that the new heirarchy is worth it, I've just been disappointed with some of the tactics that have been used to defend it. I took the question (the one in the subject of this thread) to mean "Why did NeXT feel it had to use a new heirarchy?", and not "Moses has come down the mountain with proof from G-d that NeXT did not have to use a new heirarchy, so why is NeXT such a bunch of bastards?". Maybe I just misunderstood the intent of the post. In any case, it occurs to me that this actually is in csn.programmer and not csn.advocacy, so I don't mean to sound quite as free-wheeling and inflammatory as I would be prone to in advocacy. My apologies for not paying attention to the group. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: droege@informatik.uni-koblenz.de (Detlev Droege) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 12 Apr 1996 08:58:55 GMT Organization: University Koblenz / Germany Distribution: world Message-ID: <4kl60f$62r@newshost.uni-koblenz.de> References: <4kj0d7$4to@news.sns-felb.debis.de> In article <4kj0d7$4to@news.sns-felb.debis.de> matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) writes: .. > /ncolors nc 3 idiv def > /picstr w string def > /infile (pic) 0 () /SubFileDecode filter def > > [/Indexed /DeviceRGB ncolors 1 sub > (colors) > ] setcolorspace > > <</ImageType 1 > /Width w /Height h > /BitsPerComponent 8 > /Decode [0 ncolors] > /ImageMatrix [w 0 0 h neg 0 h] > /DataSource {infile picstr readstring pop} > >> > image > endps Why do you take this long way via "readstring" ? You can directly use a (filtered or not) file for /DataSource, not only procedures or strings. So just use /DataSource infile or if pic holds the filename anyway /DataSource (pic) (r) file Detlev -- Detlev Droege, Uni Koblenz, FB Informatik, Rheinau 1, D-56075 Koblenz, Germany Tel:+49 261 9119-421,Fax:-497,NeXT/MIME/Email:droege@informatik.uni-koblenz.de C++ is the only current language making COBOL look good. --Bertrand Meyer
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer,comp.sys.next.hardware Subject: Re: Full Duplex Sound Card Capability Under NS? Date: Fri, 12 Apr 1996 12:24:33 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960412121647.17856A-100000@hphalle0.informatik.tu-muenchen.de> References: <4kjb8n$q2k@news4.digex.net> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4kjb8n$q2k@news4.digex.net> On 11 Apr 1996, John Kheit wrote: > a) SoundBlaster 32 (and higher, I believe have full duplex capabilities). > The question is, do the full duplex features work w/ the regular NS driver, > and if not, is a driver available that does support those features? > > b) Is it possible to put in multiple sound blaster 16's to get this feature > and just load up multiple instances of the SB16 driver? If so, how can I > be sure which card I'm record and which card I'm playing from, if for > example, I run one copy of SoundEditor.app for playing and another for > recording? 1) It doesn't matter wether you use a SB16 or SB32. Both cards are 16Bit. The difference is the wavetable support in the SB32. 2) Currently you can't record from one input and do output from another input connectore because: there is no app which supports this! 3) The SB 16 driver supports recording from differnt input lines and therefore might be able to support differen recording streams. I know for sure, that you can do record the currently played sound, so it should be possible to playback a sound and record from a different stream. I'm currently investigating in a new Sound application which should do this (I put it on my ToDo list) 4) There is one solution which works today: Use Sound.app twice, once on your local machine and once on a host machine. Best regards, Boerny -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> Newsgroups: comp.sys.next.programmer Subject: TransSys PNI 1.11 problem on NeXTStep 3.3. Date: Fri, 12 Apr 1996 05:14:19 -0500 Organization: University of Wisconsin-Madison Message-ID: <316E2CFB.53A6@students.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: ampriasm@students.wisc.edu I tried to run TransSys PNI SLIP 1.11 on my NeXTStep 3.3 Intel architecture system. However, when I typed "/etc/pni/pnirun pni0 &" at the terminal's prompt, it gave me a message saying that "kernel device was not loaded". This caused my modem was not able to dial the assigned phone number. Does anyone know how to take care of this problem? I am using localhost with IP 127.0.0.1. Thanks. Andrew.
From: jq@papoose.quick.com (James E. Quick) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT/Windows95 clock discrepencies Date: 12 Apr 1996 07:03:35 -0400 Organization: PHCS Message-ID: <4klda7$ek6@papoose.quick.com> References: <4kdv9r$k9d@news.mr.net> <4kjljt$f4k@nnrp1.news.primenet.com> In article <4kjljt$f4k@nnrp1.news.primenet.com>, Robert Worne <rworne@primenet.com> wrote: >A utility exists on the archives that will synch the clock >when NeXTSTEP is booted, and switch it back when it's shut down. >I don't know the name of it specifically, but someone here must >use/have it. My black HW makes all this unnecessary :) The utility that fixes the time at shoutdown and startup is TimeShift. It can be downloaded from the archives. ftp://next-ftp.peak.org/pub/next/binaries/util/TimeShift.I.b.tar.gz ftp://next-ftp.peak.org/pub/next/binaries/util/TimeShift.README Once installed it will add and subtract the difference between localtime and GMT from the system hardware clock each time the system enters and leaves NEXTSTEP. You won't have to keep resetting the time, and you also won't have to fake out the system by using the wrong timezone. -- ___ ___ | James E. Quick jq@quick.com / / / | Private HealthCare Systems NeXTMail O.K. \_/ (_\/ | Systems Integration Group (617) 895-3343 ) | - My other car has a mouse.
From: matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 12 Apr 1996 12:08:28 GMT Organization: debis Network Services GmbH Distribution: world Message-ID: <4klh3s$9tv@news.sns-felb.debis.de> References: <4kj0d7$4to@news.sns-felb.debis.de> <4kl60f$62r@newshost.uni-koblenz.de> In <ku7u3yqedge.fsf@lamancha> Richard Goldstein wrote: > The answer is that you invoke the image operator, similar to above, and > follow it up with the data via DPSWriteData. Got that! In <4kl60f$62r@newshost.uni-koblenz.de> Detlev Droege wrote: > Why do you take this long way via "readstring" ? You can directly use a > (filtered or not) file for /DataSource, not only procedures or strings. So > just use DataSource infile or if pic holds the filename anyway DataSource > (pic) (r) file Ok! I incorporated both suggestions. Thanks to both of you. Now the code looks like this: defineps drawIndexedImage( char colors[nc]; int nc; int w, h) /ncolors nc 3 idiv def [/Indexed /DeviceRGB ncolors 1 sub (colors) ] setcolorspace <</ImageType 1 /Width w /Height h /BitsPerComponent 8 /Decode [0 ncolors] /ImageMatrix [w 0 0 h neg 0 h] /DataSource currentfile >> image endps And the calling Obj-C programm should read like this: drawIndexedImage(colorList, 768, cols, rows); DPSWriteData(DPSGetCurrentContext(), image_buffer, cols*rows); It almost works, however, not really! The problem that remains possibly has to do with padding data sent by the pswrap, preceding the real image data and thus being used for imaging. Here's a look into the C code generated by the pswrap compiler: void drawIndexedImage(const char colors[], int nc, int w, int h) { ... (stuff eliminated) DPSBinObjSeqWrite(_dpsCurCtxt,(char *) &_dpsF,479); DPSWriteStringChars(_dpsCurCtxt, (char *)colors, nc); DPSWriteStringChars(_dpsCurCtxt, (char *)pad, ~(nc + 3) & 3); if (0) *pad = 0; /* quiets compiler warnings */ } How can I reliably avoid padding (provided this is the cause of my problem)? -- Matthias
From: marcel@cs.tu-berlin.de (Marcel Weiher) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 13 Apr 1996 06:12:42 GMT Organization: Technical University of Berlin, Germany Message-ID: <4kngkq$ev7@news.cs.tu-berlin.de> References: <4kj0d7$4to@news.sns-felb.debis.de> <4kl60f$62r@newshost.uni-koblenz.de> <4klh3s$9tv@news.sns-felb.debis.de> <ku7spe9e9oe.fsf@lamancha> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit rickg@sun.com (Richard Goldstein) writes: > From: matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) > Newsgroups: comp.sys.next.programmer > Date: 12 Apr 1996 12:08:28 GMT > Organization: debis Network Services GmbH >...... > It almost works, however, not really! The problem that remains possibly has > to do with padding data sent by the pswrap, preceding the real image data and > thus being used for imaging. Here's a look into the C code generated by the > pswrap compiler: > void drawIndexedImage(const char colors[], int nc, int w, int h) > { > ... (stuff eliminated) > DPSBinObjSeqWrite(_dpsCurCtxt,(char *) &_dpsF,479); > DPSWriteStringChars(_dpsCurCtxt, (char *)colors, nc); > DPSWriteStringChars(_dpsCurCtxt, (char *)pad, ~(nc + 3) & 3); > if (0) *pad = 0; /* quiets compiler warnings */ > } > How can I reliably avoid padding (provided this is the cause of my problem)? >Hmm, I've never seen the pswrap with DPS/X generate this, >maybe you're using the -p option or something? This is all >the more reason to avoid pswrap -- doing this with something >like the Modello library would look something like this: > TransformMatrix mtx; > mtx.scale(width, -height); > mtx.translate(0, height); > ImageDict dict(ctxt, width, height, ImageDict::Type1, mtx); > /* default source uses current file, else > could pass in a PSArray */ > > ctxt->image(dict); > ctxt->write_data(data, size); > ctxt->wait_context(); >rick While Modello looks like a nice idea (I have to say that, b/c I am implementing something similar as part of my NS projects:-) the AppKit provides a pretty convenient way of doing this, while also taking advantage of the fast Mach based image data transfer mechanisms. For example, I have a method for drawing images in calibrated instead of device colorspaces: -(BOOL)drawCalibrated:(const char*)colorSpace { NXColorSpace old=_colorSpace; BOOL retVal; _colorSpace = NX_CustomColorSpace; DPSPrintf( DPSGetCurrentContext(), "\n %s setcolorspace \n",colorSpace); retVal = [self draw]; _colorSpace = old; return retVal; } (It is in a category of NXBitmapImageRep) In a drawSelf:: method. I then call it as follows: if ( image ) [image drawCalibrated:" [ /Indexed /DeviceGray 255 { 255 div 1 exch sub } ] "]; Works like a charm. Marcel
From: csaldanh@mae.carleton.ca (Chris Saldanha) Newsgroups: comp.sys.next.programmer Subject: Pascal compiler? Date: 12 Apr 1996 11:37:46 GMT Organization: computerActive Inc Message-ID: <4klfaa$ppg@bertrand.ccs.carleton.ca> Does anyone know of an actual Pascal _compiler_ for NeXTSTEP/UNIX systems? I have p2c, but I'd really like a real compiler. I heard once that there was a plug-in for the GNU compiler that would allow it to compile pascal, but I've never actually seen such a package. --Chris Chris Saldanha, Software Analyst -------------------------------------- computerActive, Inc |"The telephone was not invented by | chris@computerActive.on.ca (NeXTMail) | Alexander Graham Unitel" -Bell Ad | http://www.mae.carleton.ca/~csaldanh --------------------------------------
From: samurai@random.hasc.ca (Darcy Brockbank) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 12 Apr 1996 14:47:19 GMT Organization: disorganized Message-ID: <SAMURAI.96Apr12104719@random.hasc.ca> References: <316186b9.4262367@news.plsys.co.uk> <4jtit1$am5@sun0.urz.uni-heidelberg.de> <3164410a.1117449@news.plsys.co.uk> In-reply-to: k_harbour@mail.bogo.co.uk's message of Thu, 04 Apr 1996 21:37:26 GMT <k_harbour@mail.bogo.co.uk> writes: >tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: >>In article <316186b9.4262367@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl >>Harbour) writes: >> >>>I agree that the FoundationKit is nice to develop with, but can you, >>>or anyone, give me a really good technical reason why its >>>functionality could not have been fitted within the old Object >>>hierarchy? >> >>Firs one general point: Foundation is a *new* kit! It is not just an extension >>to previous AppKit hacks like List and HashTable. >In what sense is List a hack? It's a perfectly respectable class as >far as I'm concerned. Which isn't to say it couldn't have been >improved. >And, if memory serves, a post last year, which I accept may have been >bogus, said NSMutableArray is 3 (?) times slower than List. Not quite... here's the data you were referring to: 1) NSMutableArray: Adding 15000 items at the end. 0.11 2) NSMutableArray: Adding 15000 items in the centre. 24.19 3) NSMutableArray: Accessing first 15000 items. 0.01 4) NSMutableArray: Removing 15000 items from the beginning. 46.80 5) NSMutableArray: Removing 15000 items from the end. 0.10 Total time:71.21 1) NKQueue: Adding 15000 items to the end. 0.17 2) NKQueue: Adding 15000 items in the centre. 9.65 3) NKQueue: Accessing first 15000 items. 0.01 4) NKQueue: Removing 15000 items from the beginning. 0.15 5) NKQueue: Removing 15000 items from the end. 0.16 Total time: 10.14 1) NKList: Adding 15000 items to the end. 0.14 2) NKList: Adding 15000 items in the centre. 14.32 3) NKList: Accessing first 15000 items. 0.01 4) NKList: Removing 15000 items from the beginning. 27.09 5) NKList: Removing 15000 items from the end. 0.14 Total time: 41.70 1) List: Adding 15000 items at the end. 0.02 2) List: Adding 15000 items in the centre. 16.00 3) List: Accessing first 15000 items. 0.01 4) List: Removing 15000 items from the beginning. 33.18 5) List: Removing 15000 items from the end. 0.02 Total time:49.22 Notes: List has a slight advantage of not having to retain and release the elements that go into it. However, when you compare, it's running something like 40% faster than the NSMutableArray. The two NK objects are both classes I wrote. They have identical interfaces, but the NKList uses a simple linear array algorithm as do List and NSMutableArray, while the NKQueue uses a more complicated circular array. The benefits of the circular array are quite obvious, no? Both are NSObject subclasses, and retain/release their contents. Anyway, since my NKList beats both List and NSMutableArray hands down, and is really a replacement for NSMutableArray, I'd conclude that: (i) The FoundationKit, and NeXT classes in general are fat and slow due to poor implementation. Since my NKQueue totally blows them all away, I'd conclude that: (ii) NeXT chose non-optimal algorithms for implementing base classes. Since the interfaces of the Foundation and OpenStep classes REALLY SUCK, and retain/release cycles bite me in the rear end every day, making my life miserable, and NeXT took things like copyWithZone: out of the root class (which is another total ass biter, just think about it a while) I'd also conclude that: (iii) The designers of OpenStep and Foundation did nowhere near as good a job as the original AppKit designers. Let's face it... OpenStep as an API sucks. It's all we have for the next little bit, but the writing is finally absolutely on the wall now. With Sun backing off OpenStep, and moving into Java, NeXT's technology becomes less meaningful every day... >>Before we had just a >>collection of nifty useful classes helping us to do our GUI work more easily. >>Now we have a kit of cooperating classes, giving us strong and powerful design >>patterns. But being a *new* kit, it has the *rights* to be really new! >It still didn't have to introduce a new root object. Should every new >kit have a new root? Let's forget about re-use then. I know that we're not even going to bother with a whole heap of our applications. Porting them to OpenStep presents such a painful alternative, we're just going to move people into NT and use shareware or $100 apps where we used to write our own. OpenStep is not a kit of cooperating classes, giving us strong and powerful design. It's a dramatic rewrite of something that worked pretty damn well for a long time. It breaks everything you currently have. Does it work well? Who can say for sure until you use it for a year, or try to implement something difficult. That's the true test. >>I've >>seen a lot of different implementations of array, hash table, string, etc. pp >>classes. Should Foundation try to be backwards compatible with all of them? >>No! And because NeXTSTEP was using similar classes in the past do not make >>them more privileged ;-) >> >Well excuse me, but I thought one of the fundamental principles of OO >was to be able to extend, enhance, and change the internal >implementation, of your old classes without completely replacing them. >By not being able to do this with their _own_ classes, what message is >Next sending to the OO community? NeXT is saying: you can't reuse all your code because we just yanked the carpet out from under you. Your entire base of hundreds of thousands of lines of painstakenly proven and tested code is now a big black hole containing many ass-biting bugs which will crop up in many wierd places over the next year or so, continually causing you expensive repairs, and embarassment. >How, as a consultant, can you sell the maintenance benefits of OO >(re-use, extensibility, ability to change implementation without >changing interface etc) with Nextstep, when Next themselves have had >to re-code an entire class hierarchy? >>After all, I was told that the conversion scripts do very proper job... >I was told that Digital, HP and Sun would ship OpenStep >implementations by mid-1995. I'm still waiting... I can't comment on the conversion scripts because of beta. The tests I wrote up above were with Foundation that shipped with EOF. All I'll say is that I hope the shipping conversion scripts are easy to use, intelligent, and don't insert tons bugs into my code. I also hope that they convert things properly, and don't flag everything for me to repair manually. I also hope that they don't force me to re-read every single method that I ever wrote which returns (id). >>One of the biggest advantages of ObjC over C++ is simplicity. I'm perfectly >>happy with what I have now. Simplicity... OK, say I register myself for an update of realtime data. Gotta wait for the data to come in. So, a periodic check is registered, I set my object up as a target to be called later. This object that's going to call me, retains me to make sure I'm still around for its update method call. Fine. The update comes in, and the object passes itself in as a parameter. The recipient cannot retain the object because that will create a cycle. It *should* retain the object, to make sure it still exists, or perhaps the recipient wants to keep the data around for a while... How does the implementor of the recipient know to do this? He reads the documentation. Very simple. Every programmer always reads all documentation, never misinterprets what they read, or forgets what they read. This is why today's compilers don't bother checking types anymore. Because everyone reads and knows all the documentation, and writes all their code correctly, there's no need to implement features in syntax. Just tell people how they're supposed to use things, and it'll be all right. >>>5) The mutability concepts? >>> >>>Okay, NSImmutableList:List would violate the substitutability >>>principle, but again I think this is probably worth it to prevent >>>disposing of an entire existing class hierarchy. And anyway, it still >>>doesn't explain why you couldn't have NSMutableArray:NSArray:Object >> >>RTFM ... class clusters! And, see comments like "this method is fast..." in >>various places in the docs. All this is possible because of the class clusters >>and is *not* possible by simple subclassing. BTW, NSNumber is great example >>where one can add his/her own private classes! Really great! >Don't tell me to RTFM. I'm very familiar with _all_ the classes in >Foundation. Explain why NSNumber could not be a subclass of Object and >still have private subclasses? I always thought they should have introduced the idea of a const object. - (const HashTable *)dada; The HashTable you get is essentially read-only, just like you'd expect in C. Mutable/Immutable is silly. Methods are then declared as methods that modify the object, or just access the data (ie. mutable and immutable). This would have been much much nicer. Anyway, this just goes to show you that no matter the shit that someone shovels at a group of people, some people will eat it and smile, some people will eat it and gag, and some people will refuse to eat. - db -- I have a heart condition... if you hit me, it's murder. -- Withnail -- We are millionares! -- and I -- To you, we're not deep -- Housemartins -- A bit small for a share certificate -- Cleese -- Smoke me a kipper, skipper, I'll be back for breakfast. -- Ace Rimmer -- He's the tallest man by a dam site. -- Crow --
From: rickg@sun.com (Richard Goldstein) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 12 Apr 1996 08:36:33 -0700 Organization: SunSoft, Inc. Sender: rickg@lamancha Message-ID: <ku7spe9e9oe.fsf@lamancha> References: <4kj0d7$4to@news.sns-felb.debis.de> <4kl60f$62r@newshost.uni-koblenz.de> <4klh3s$9tv@news.sns-felb.debis.de> In-reply-to: matti@dbag.ulm.DaimlerBenz.COM's message of 12 Apr 1996 12:08:28 GMT From: matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) Newsgroups: comp.sys.next.programmer Date: 12 Apr 1996 12:08:28 GMT Organization: debis Network Services GmbH ...... It almost works, however, not really! The problem that remains possibly has to do with padding data sent by the pswrap, preceding the real image data and thus being used for imaging. Here's a look into the C code generated by the pswrap compiler: void drawIndexedImage(const char colors[], int nc, int w, int h) { ... (stuff eliminated) DPSBinObjSeqWrite(_dpsCurCtxt,(char *) &_dpsF,479); DPSWriteStringChars(_dpsCurCtxt, (char *)colors, nc); DPSWriteStringChars(_dpsCurCtxt, (char *)pad, ~(nc + 3) & 3); if (0) *pad = 0; /* quiets compiler warnings */ } How can I reliably avoid padding (provided this is the cause of my problem)? Hmm, I've never seen the pswrap with DPS/X generate this, maybe you're using the -p option or something? This is all the more reason to avoid pswrap -- doing this with something like the Modello library would look something like this: TransformMatrix mtx; mtx.scale(width, -height); mtx.translate(0, height); ImageDict dict(ctxt, width, height, ImageDict::Type1, mtx); /* default source uses current file, else could pass in a PSArray */ ctxt->image(dict); ctxt->write_data(data, size); ctxt->wait_context(); rick -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "... push the button, Frank." - Dr. Clayton Stonewall Forresster Richard M. Goldstein rickg@Eng.Sun.COM Desktop Performance Group SunSoft, Inc.
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.hardware Subject: Re: Full Duplex Sound Card Capability Under NS? Date: 12 Apr 1996 17:17:44 GMT Organization: monoChrome, Inc. Message-ID: <4km37o$8a@news4.digex.net> References: <4kjb8n$q2k@news4.digex.net> <Pine.HPP.3.91.960412121647.17856A-100000@hphalle0.informatik.tu-muenchen.de> Hi Bernhard :) Bernhard Scholz <scholz@informatik.tu-muenchen.de> wrote: > On 11 Apr 1996, John Kheit wrote: > 1) It doesn't matter wether you use a SB16 or SB32. Both cards are 16Bit. > The difference is the wavetable support in the SB32. As I understand it, the newer SB's have support for full duplex where the older ones did not...But as always, I reserve my right to be 100% wrong :) > 2) Currently you can't record from one input and do output from another > input connectore > because: there is no app which supports this! This is not the problem... If you go to any NeXT computer and run two copies of the Sound.app at the same time...then in one of the apps play a long song, then go to the second running app and hit record...On the NeXT computer both seperate Programs/processes will work. However, doing the same on a NEXTSTEPintel machine will not work. I think it is a function of the sound hardware and driver not being able to handle full duplex.... > 3) The SB 16 driver supports recording from differnt input lines and > therefore might be able to support differen recording streams. > I know for sure, that you can do record the currently played sound, so > it should be possible to playback a sound and record from a different > stream. I'm currently investigating in a new Sound application which > should do this (I put it on my ToDo list) Well, please let me know if you come up with anything, as it has completely stumped me (not a difficult thing to do ;). But again, you might try my little experement above (re running multiple applications and doing plays and records through the different channels)...I would think that if the intel hardware were capable, then that experiment would have the same results on both NeXT and Intel hardware... > 4) There is one solution which works today: Use Sound.app twice, once on > your local machine and once on a host machine. Yes, the ultimate second sound card...a whole new machine, networking, and another copy of NEXTSTEP...maybe it is a marketing device :) Yes...I actually do know about that, but unfortunatly, I need it to work on one machine (sigh). Anyway, thanks for the help. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: CVS Date: 12 Apr 1996 18:08:49 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4km67h$28h@lion.embl-heidelberg.de> Which version of CVS can handle nib files properly and use remote repository? I know it's a FAQ, but I lost my dev mailbox :-( -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: How to start a daemon on remote host? Date: 12 Apr 1996 18:14:17 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4km6hp$2ca@lion.embl-heidelberg.de> I want to start a daemon on remote host from an app running locally. system("rsh hostname nohup daemon parameters"); works somehow, but I feel it isn't the right way.... tnx -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: root@ecs.soton.ac.uk (Eren Kotan) Newsgroups: comp.sys.next.programmer Subject: Loading files and parsing RenderMAn Date: 12 Apr 1996 18:23:19 GMT Organization: I need to put my ORGANIZATION here. Message-ID: <4km72n$mt9@wapping.ecs.soton.ac.uk> Hi, I am writing a mini app for my final year university project, which needs to parse an ASCII file and render a number of 3D general cylinders depending on the numbers in the file. I am using NEXTSTEP 3.3 and the 3DKit for this, and despite my lack of programming experience with the 3DKit, I made some small progress so far. There is one major problem though: I cannot get more than one object displayed at a time. Can anyone help? I am guessing I need to construct some sort of a list and then render the objects in it, but I cannot figure out the right syntax and I cannot seem to find any examples of this. The input file is of the form: 0 3.0 3.0 4.0 2.0 4.5 6.0 2.0 4.1 ... Where the first field is an ID, and the rest give the x, y, z coords and the radii of the two ends of the cylinders. This is just an academic exercise for my course, and I cannot find any example code which does this, and noone at my university has experienced NEXTSTEP. I'd appreciate any help, please email eek93el@soton.ac.uk if you wish to respond. Regards, Eren
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Canon NEXTSTEP Developer Support-Japan Date: 12 Apr 1996 11:19:40 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4km6rs$35u@nova.mdd.comm.mot.com> References: <316D875B.6E2C@rs6000.cmp.ilstu.edu> In article 6E2C@rs6000.cmp.ilstu.edu, "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> writes: >Has anyone seen the Canon NEXTSTEP Developer Support web pages in Japan? Yes. >I can't understand the attitude of American Canon VS. Japanese Canon. I wish the two >would truly work together and also fully translate the web pages to English ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [munch] God Bless Americer! If English is good enough for Steve Jobs, its good enough for me. Yeeeehaaaaaaw! Cheers from Canada, Steve. PS: Why the heck should they? I don't see NeXT translating everything on www.next.com to Japanese. -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: mpaque@pbinet.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: Fri, 12 Apr 1996 19:29:40 GMT Organization: Pacific Bell Internet Services Message-ID: <4kmb0q$k5q@nntp.snfc21.pbi.net> References: <4kj0d7$4to@news.sns-felb.debis.de> matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) wrote: >With my very limited knowledge of PostScript I'm trying to write a pswrap for >imaging data using the Indexed feature (I think the AppKit itself does not >support it). This is the solution I finally came up with: [munch] >The central point is that I convert my string variable 'pic' which holds the >image data into a file (see definition of 'infile'). The whole code works in >principal. But, alas, if the image is quite large, the program encounters a >limit check error, probably because PostScript string variables cannot hold >arbitrary many bytes. >I really spent some time on this before I finally gave up. So, hey, >PostScript wizards, what would be the real way for solving my task? As you found, PostScript has some definite limits on things like the length of strings, stack contents, etc. Even if one finds a way past these limits for the Window Server's DPS implementation, one may still hit these limits on other PS devices (e.g. PS printers, DPS for X, etc). What I would suggest is that you look into converting your indexed data into RGB data yourself, and write that data into an NSBitmapImageRep (or NX...). This can be imaged very efficiently using the AppKit and NeXT DPS mechanisms. If you must do this through the DPS interpreter, you might want to try band rendering, dicing your input stream into smaller chunks on scanline boundries, and imaging into a series of horizontal strips. If Scott Byer is lurking about, he might have some other more clever ideas. I'm afraid I can't give you a better answer without understanding more of the problem (e.g, the source and format of the input data, expected size range, and what you would ultimately like to do with it). Mike Paquette I don't speak for my employer, and they don't speak for me. mpaque@pbinet.com Personal E-mail mpaque@next.com NeXT business mail only, please
From: Mark_Bessey@next.com (Mark Bessey) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil (really long!) Date: 12 Apr 1996 19:58:37 GMT Organization: NeXT Software, Inc. Message-ID: <4kmcld$q91@news.next.com> References: <jpanicoDppG9I.KAy@netcom.com> Joe Panico writes > Mark Bessey (MaRK_BeSSeY@NeXT.CoM) wrote: > > The big problem is this: Messaging a nil object causes a chain of > useless results to propagate through your program. By the time you get > to somewhere where ignoring the problem is no longer acceptable, you > have no idea where the failure occurred. > > The "right thing" to do on messaging a nil object is throw an > exception. This would preserve the advantage of streamlining the > application logic, but provide a better, more structured way of > dealing with the occasional nil message. > > uhmm, if this is the case, then why do NeXT software engineers > themselves message nil objects? Carelessness? Blindness to the "one true way" to software enlightenment? Differences in opinion about what constitutes "proper" behavior? There are many reasons why. Ok, let's take a look at your example: > For instance, if you ask EOModeler to produce a template EO, the > accessor methods will use constructs like: > > - (void)setTYPE_CODE:(NSString *)value > { > [TYPE_CODE autorelease]; > TYPE_CODE = [value retain]; > } > > Please clarify. There are two message expressions here. Let's take a look at the consequences if either receiver is nil: [TYPE_CODE autorelease]; if TYPE_CODE if nil, autorelease won't do anything, and returns a nil value (which is discarded) TYPE_CODE = [value retain]; if value is nil, retain does nothing, then returns nil, which is assigned to TYPE_CODE, which is probably what the programmer intended. Based on that analysis, a software engineer could easily conclude (and someone obviously did) that both cases are benign, and the code's just fine as it stands. HOWEVER, I DISAGREE. In the case of the autorelease message, I have to admit that the risk of unintended behavior is very small. It's certainly easier to do it this way. However, it's actually faster to explicitly check for a nil object before sending the message (thanks, Matt!), so we'd still win here. In the case of the second message, there's obviously a problem. What is the *meaning* of (NSString *)nil? Is it an empty string, or the absence of a value? The two are not the same. Here's my attempt at a "better behaved" version: - (void)setTYPE_CODE:(NSString *)value { if (nil != TYPE_CODE) [TYPE_CODE autorelease]; if (nil != value) { TYPE_CODE = [value retain]; } else { value = nil; [NSException raise: NSinvalidArgumentException fromat: @"setTYPE_CODE: value cannot == nil"]; } } OK, so it's a little longer than the original version. But it's faster, and emits a (somewhat) helpful diagnostic message if you pass in nil for the value... Of course, this is a really simple method, so the overhead of trapping the error looks pretty huge. With any other type of method the proportion of exception handler lines to method lines would be smaller... But since this is automatically generated code, why not just do things the long way anyway? Maybe I should go file a bug against EOModeler's template generation... <<DANGER!!! Speculation ahead!!!>> Now, in a more ideal world, the compiler would generate the if statement and exception raise for you, which would eliminate about half that code... An additional optimization might be to not raise an exception if there are no side effects (the return value of the message isn't used). I'm not totally convinced this is necessary or wanted, though it would simplify the code for the first case. Then we'd have something like this: - (void)setTYPE_CODE:(NSString *)value { [TYPE_CODE autorelease]; NS_DURING TYPE_CODE = [value retain]; NS_HANDLER if ([[localException name] isEqualToString: NSNilObjectException]) { value = nil; } [localException raise]; NS_ENDHANDLER } -Mark -- Mark Bessey NeXT Software, Inc Software Quality Assurance -->I DON'T SPEAK FOR NeXT <--
Date: 13 Apr 1996 15:30:58 GMT From: David@Cetex.zynet.co.uk (David De la Haye) Newsgroups: comp.sys.next.programmer Message-ID: <cancel.4ko1g8$ngq@sol.zynet.net> Control: cancel <4ko1g8$ngq@sol.zynet.net> Subject: cmsg cancel <4ko1g8$ngq@sol.zynet.net> EMP/ECP (aka SPAM) cancelled by jem@xpat.com. See news.admin.net-abuse.announce, report 19960413.23 for further details
Newsgroups: comp.sys.next.programmer From: embuck@hogan.cca.rockwell.com (Erik M. Buck) Subject: Re: Why does Foundation Kit introduce a "n Message-ID: <DpnJwK.Ax9@lazrus.cca.rockwell.com> Sender: news@lazrus.cca.rockwell.com Organization: Rockwell Avionics - Collins References: <4juvbc$2i4@nova.mdd.comm.mot.com> <4k0uuo$mub@shelob.afs.com> <4k2gfs$9j4@usenet.rpi.edu> Distribution: world Date: Wed, 10 Apr 1996 15:19:31 GMT I agree that all or almost all of the foundation kit enhancements could have been made within the existing Object hierarchy. Many of the method name changes could have been implemented as additional methods rather than replacements. It would have been confusing to new comers, but they would get the idea fast enough. Next could have insisted that the legacy classes such as List remain in the Openstep standard that Sun et all implement. All of the renaming having to do with pre-pending NS or even NX was probably needed for name space reasons. A very simple script could have added NS to all class names where appropriate. That would have broken very little code and probably would not necessitate complete re-testing. So the question remains - why did NeXT do it ? There are some advantages to a "clean" break with the past. Given the choice of being confined to only NeXTstep systems with perfect Object re-use or having a clear but slightly painful transition to Openstep from multiple vendors, I choose the latter. NeXT may have felt that they were doing us developers a favor by expanding the market. They may have felt that we should be grateful for a chance to move the bulk of our code to a wider audience. They may have thought that their last chance to "get it right" was upon them. It is conceivable that Sun pushed NeXT into pissing off customers. It may be that studies indicated new developers would think the legacy support was all a big cludge. I don't know. With regard to Delphi... Saying that Delphi transitioned from 1.0 to 2.0 without a lot of pain is not fair. That is similar to saying that C was transitioned to Objective-C without pain. If you limited yourself to the C sub-set, that is true but valueless. Of course old Turbo-Pascal code works with Delphi. Old C works with Openstep. - Erik M. Buck I do not speak for NeXT or Rockwell. I speak for myself only and the opinions I have are my own property. I am grateful to Rockwell for letting me post this note with the assistance of their facilities.
From: rickg@sun.com (Richard Goldstein) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 12 Apr 1996 13:28:09 -0700 Organization: SunSoft, Inc. Sender: rickg@lamancha Message-ID: <ku7rattdw6e.fsf@lamancha> References: <4kj0d7$4to@news.sns-felb.debis.de> <4kmb0q$k5q@nntp.snfc21.pbi.net> In-reply-to: mpaque@pbinet.com's message of Fri, 12 Apr 1996 19:29:40 GMT From: mpaque@pbinet.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Date: Fri, 12 Apr 1996 19:29:40 GMT Organization: Pacific Bell Internet Services matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) wrote: >With my very limited knowledge of PostScript I'm trying to write a pswrap for >imaging data using the Indexed feature (I think the AppKit itself does not >support it). This is the solution I finally came up with: [munch] >The central point is that I convert my string variable 'pic' which holds the >image data into a file (see definition of 'infile'). The whole code works in >principal. But, alas, if the image is quite large, the program encounters a >limit check error, probably because PostScript string variables cannot hold >arbitrary many bytes. >I really spent some time on this before I finally gave up. So, hey, >PostScript wizards, what would be the real way for solving my task? As you found, PostScript has some definite limits on things like the length of strings, stack contents, etc. Even if one finds a way past these limits for the Window Server's DPS implementation, one may still hit these limits on other PS devices (e.g. PS printers, DPS for X, etc). Wait a minute, the limit problem here is on the use of a descrete PS type, i.e. the string. Using the currentfile as a source effectively eliminates that problem irreguardless of the interpreter. Just be careful that the scratch string you read into is within the limit. Usually things read a scan line at a time off of the currentfile, as in: /scratch image-span def ... /DataSource {currentfile scratch readstring pop} ... What I would suggest is that you look into converting your indexed data into RGB data yourself, and write that data into an NSBitmapImageRep (or NX...). This can be imaged very efficiently using the AppKit and NeXT DPS mechanisms. May be sound advice for other reasons. rick -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "... push the button, Frank." - Dr. Clayton Stonewall Forresster Richard M. Goldstein rickg@Eng.Sun.COM Desktop Performance Group SunSoft, Inc.
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: sending a (const char *) to a (char *) Date: 13 Apr 1996 16:26:10 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4kokj2$gpk@core.bard.edu> I am writing a program using the MusicKit in which I would like to be able to change parameters with Buttons and PopUpLists. Here's my problem: [theNote setPar: [Note parTagForName: [aButton title] ] .... parTagForName: expects to recieve a (char *) and [aButton title] sends a (const char *). This caused the program to crash, which I believe corrupted my system also. I had to reinstall the OS after this program crashed. I also tried: char *parName; strcpy(parName, [aButton title]) but the compiler gave the same warning as before and the program crashed. Please help...this is rather crucial to my program. Thanks in advance...HC -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer Subject: Re: Canon NEXTSTEP Developer Support-Japan Date: 12 Apr 1996 22:26:59 GMT Organization: monoChrome, Inc. Message-ID: <4kmlbj$28u@news4.digex.net> References: <316D875B.6E2C@rs6000.cmp.ilstu.edu> <4km6rs$35u@nova.mdd.comm.mot.com> besler@mdd.comm.mot.com (Steve Besler) wrote: > God Bless Americer! If English is good enough for Steve Jobs, its good enough for me. Yeeeehaaaaaaw! That's right...God bless it. Wow, one more slag at Americans, how brave and unique...well I'll try and supress my urge to share my opinions about canadians... > PS: Why the heck should they? I don't see NeXT translating everything > on www.next.com to Japanese. Perhaps b/c english (which you may recall is from ENGLAND, not the US) has become the international go-between language. Europe uses it to talk between itself... Kinda annoying to learn 5 or more languages just to do business. So it never hurts to put things in english, b/c the majority of the planet is versed in it. That may be all the more true for a country that kinda relies on exports... Jeez, get the American battery off your shoulder already. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 13 Apr 1996 18:06:12 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4koqek$kjl@emerald.oz.net> References: <4k2gfs$9j4@usenet.rpi.edu> <4kcict$2qo@nova.mdd.comm.mot.com> <4kk329$68d@usenet.rpi.edu> > besler@mdd.comm.mot.com (Steve Besler) wrote: > > Garance A Drosehn <gad@eclipse.its.rpi.edu> writes: > > > The issue at the heart of this discussion is not "Does NeXT do > > > nasty things to people it should be working with?", the issue > > > is "Does Object-Oriented Technology live up to it's claims?". > > > > No, I don't think that is what this thread is about. It's really > > about whether NeXT should have maintained backward compatability > > with NEXTSTEP in OpenStep. One way they could have done this > > was to build all the new functionality to into the existing Object > > hierarchy. A few thoughts... Dealing with a major API change is a pain for all developers. But NeXT is the OpenStep developer with the most NS code to convert, so any API changes will affect NeXT more than anyone else. Of course, NeXT has resources that the rest of us don't have, but they're not likely to do something that's going to cause them a great deal of grief unless the rewards are substantial. I have no inside info, but I can imagine NeXT determining that NS just wasn't going to make it as a proprietary OS. NeXT began using the Mac model of proprietary software on proprietary hardware that all worked great together, but that was relatively expensive. Apple's experience is showing that this model is flawed and has been for some time. NeXT realized this when it abandoned its proprietary hardware first. I can imagine its proprietary operating system being the next to go with only its proprietary development and run-time environment remaining. But there are many examples of successful proprietary development environments, so the OpenStep model seems valid. But moving NS to become an OS-independent development environment requires lots of API enhancements. Probably much of the existing NS API is just sub-optimal for an OS-independent development environment. And there are always portions of an API that just didn't work out as intended and could be improved. So faced with the need to make major changes to become OS-independent, NeXT had an opportunity to "fix" other problems as well. But this had to be done with a minimum of hassle for developers, so NeXT has spent a considerable effort developing automated conversion tools. Yes, NeXT could probably have done most of what needed to be done within the Object hierarchy thus preserving object reuse, but the mixing of both worlds would have been very messy as those of you familiar with programming EOF under NS can attest. I would much rather spend a little effort now and make a clean break with the past than drag around a messy API forever and ever as some others have chosen to do. I have heard that Windows 95 includes 1,000,000 lines of code just to implement backward compatibility. No wonder it was so late in being released! No wonder it hasn't been as wonderful as had been expected. I don't want to have to deal with such a monstrosity, and that's what OpenStep under the Object hierarchy would likely have been. So we're going to grin and bear the inconvenience of the OpenStep conversion while continuing to ship our NS apps. Once we've made the conversion, we hope to be much better off with apps that will run under NT and thus will be much more commercially viable than our NS apps would ever have been. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: Robert Worne <rworne@primenet.com> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: Re: Canon NEXTSTEP Developer Support-Japan Date: 13 Apr 1996 13:00:03 -0700 Organization: Primenet Services for the Internet Sender: root@primenet.com Message-ID: <4kp143$dfv@nnrp1.news.primenet.com> References: <316D875B.6E2C@rs6000.cmp.ilstu.edu> "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> wrote: >Has anyone seen the Canon NEXTSTEP Developer Support web pages in Japan? > >http://www.cnds.canon.co.jp/Japanese_EUC/WelcomeJ.html > >I can't understand the attitude of American Canon VS. Japanese Canon. I wish the two >would truly work together and also fully translate the web pages to English and provide >object.station driver downloads/updates! That's nothing... did you see they supply drivers for the Canon Laser/Bubblejets for the Japanese version of NS? For free? These drivers supposedly are for the Japanese versions of these printers, and probably will not work with the American/European versions, but the mind boggles. >I am suprised that a company like Canon would drop a product that was so expensive and >leave users to fend for themselves...It's so hard to put faith in companies- people too for >that matter since they make up companies. Although, I have MUCH MORE RESPECT for the >Japanese divisions due to their greater attention to quality and support (my opinions). The Japanese seem quite fond of computers/OS's that support their native language. OS/2 does quite well over there, for instance. -- Robert Worne NeXT-OS/2-MacOS Starving CS Undergrad...Sorry, I don't *do* Windows! I'd rather starve... -----BEGIN GEEK CODE BLOCK----- Version: 3.1 GCS/P/S d-?>pu s+:+> a- C++$>++++ UX++++>$ P+>+++ L+ !E !W++ N+++ !o-- !K w--- O++$ M+ V PS>--- !PE+ Y+ !PGP- t@ 5++ X+++ R- tv b+>++ DI !D G e>+++ h--- r++ y+++** ------END GEEK CODE BLOCK------
From: dekorte@intrepid.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil (really long!) Date: 13 Apr 1996 20:44:02 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4kp3mi$94l@news.onramp.net> References: <jpanicoDppG9I.KAy@netcom.com> <4kmcld$q91@news.next.com> If there was a nil object, then one could choose to have it ignore messages or to report a warning/error or pop up a debug panel or whatever you wanted. So how could you go wrong by adding this feature to Objective-C? -- Steve Dekorte mailto:dekorte@suite.com (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: csaldanh@mae.carleton.ca (Chris Saldanha) Newsgroups: comp.sys.next.programmer Subject: Re: sending a (const char *) to a (char *) Date: 13 Apr 1996 20:24:48 GMT Organization: computerActive Inc Message-ID: <4kp2ig$29a@bertrand.ccs.carleton.ca> References: <4kokj2$gpk@core.bard.edu> Hans Steiner (hs283@bard.edu) wrote: : parTagForName: expects to recieve a (char *) and [aButton title] sends a : (const char *). This caused the program to crash, which I believe : corrupted my system also. I had to reinstall the OS after this program : crashed. I find that quite hard to believe! Are you sure that was the reason for corrupting your system? I also tried: : char *parName; : strcpy(parName, [aButton title]) You need to allocate the space for the pointer passed in to strcpy() in advance. Do: char *parName; parName = calloc(strlen([aButton title])+1, sizeof(char)); strcpy(parName, [aButton title]); --Chris Chris Saldanha, Software Analyst -------------------------------------- computerActive, Inc |"The telephone was not invented by | chris@computerActive.on.ca (NeXTMail) | Alexander Graham Unitel" -Bell Ad | http://www.mae.carleton.ca/~csaldanh --------------------------------------
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: [[sender selectedCell] title] returns nil Date: 13 Apr 1996 22:39:24 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4kpaes$1oll@core.bard.edu> I am trying to use a PopUpList to cchange MusicKit parameters, but in the action method that the PopUpList's MenuCells call, setXPar: sender, [[sender selectedCell] title] only gives me what I assume is a pointer to nil. From gdb: (gdb) p parName $1 = 0x0 Could there be some problem trying to copy a (const char *) to a (char *)? I have used three different methods of doing so: char *parName parName = malloc(strlen([xButton title]) + 1); strcpy (parName, [xButton title]); char parName[15]; strncpy(parName,[xButton title], 15); char *parName parName = malloc(strlen([[sender selectedCell] title]) + 1); strcpy(parName, [[sender selectedCell] title]); and each time the program crashed saying: Memory access exception on address0x0(protectionfailure) This is crucial to my project, so any help would be appreciated. -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: Lee Willis <lee@multigen.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Why does Foundation Kit introduce a "n Date: 12 Apr 1996 20:59:02 GMT Organization: MultiGen Message-ID: <4kmg6m$f90@SantaClara01.News.Internex.NET> References: <4jufkv$jbc@shelob.afs.com> <4juvbc$2i4@nova.mdd.comm.mot.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit besler@mdd.comm.mot.com (Steve Besler) wrote: [snippage] >It seems to me that the more legacy code people have laying >around, the more they want to stick with "the old way" of doing >things. Not because the old way is "better" but because of a >mixture of laziness and greed... And golly, what an unreasonable and contempible attitude that is, too! I got news for you, Steve: "Laziness and greed" is the whole point. The goal of this game called business is to get the most return (greed) from the least investment (laziness). That's how you stay in business. That's how you make money. That's how you pay your employees. That's how you buy new equipment. That's how you have capital to invest in R&D. ... Anything that allows you to get more return, or put in less investment is a *GOOD* thing. Anything that decreases your return, or increases your investment, is a *BAD* thing. >-- >Steven Besler, Coop Student "I do not speak for Motorola." >Motorola Wireless Data Group besler@mdd.comm.mot.com -- Lee Willis MultiGen Inc. 550 S Winchester Blvd. Suite 500 San Jose, CA 95128 (408) 261-4100 "Experience is what you get when you don't get what you want."
From: tm8025a@american.edu (Torrey McMahon) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 13 Apr 1996 23:18:29 GMT Organization: American University, Washington DC Distribution: world Message-ID: <4kpco5$ps5@paladin.american.edu> References: <4k2gfs$9j4@usenet.rpi.edu> <4kcict$2qo@nova.mdd.comm.mot.com> <4kk329$68d@usenet.rpi.edu> <4koqek$kjl@emerald.oz.net> In <4koqek$kjl@emerald.oz.net> Art Isbell wrote: > I have no inside info, but I can imagine NeXT determining that NS just > wasn't going to make it as a proprietary OS. NeXT began using the Mac model > of proprietary software on proprietary hardware that all worked great > together, but that was relatively expensive. Apple's experience is showing > that this model is flawed and has been for some time. NeXT realized this > when it abandoned its proprietary hardware first. I can imagine its > proprietary operating system being the next to go with only its proprietary > development and run-time environment remaining. But there are many examples > of successful proprietary development environments, so the OpenStep model > seems valid. > Just out of curiosity, what is a non-propriatary OS? I agree with almost all of your posts but NeXTSTEP....er...NEXTSTEP is not a propriatary OS is it? It runs, on more hardware platforms then DOS, Win95, SunOS...etc. etc. etc. -- Torrey McMahon American University School of Communication System Administrator, WebMaster, Security Consultant NeXT, MIME, SUN, and ASCII --- PGP Key available via finger and home page. GCM d- s:- a-- C++ UX++++ P+ L+ E---- W+++ N+++ K- w-- !O M V-- PS+ PE- Y+ PGP+ t+ 5++ X+ !R tv- b+ DI++ D++ G e++ h* r++ y+
Newsgroups: comp.sys.next.programmer,comp.lang.objective-c From: jpanico@netcom.com (Joe Panico) Subject: Re: when an expression evaluates to nil (really long!) Message-ID: <jpanicoDptpF1.JJF@netcom.com> Followup-To: comp.sys.next.programmer,comp.lang.objective-c Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <jpanicoDppG9I.KAy@netcom.com> <4kmcld$q91@news.next.com> <4kp3mi$94l@news.onramp.net> Date: Sat, 13 Apr 1996 23:04:13 GMT Sender: jpanico@netcom15.netcom.com Steve Dekorte (dekorte@intrepid.suite.com) wrote: : If there was a nil object, then one could choose to have it ignore messages : or : to report a warning/error or pop up a debug panel or whatever you wanted. : So how could you go wrong by adding this feature to Objective-C? This is the obvious solution, and I assume that most custom shops roll their own "null" object for many different purposes. For instance, the collection classes like NSArrays and NSDictionaries cannot hold nil values, but there are times when it is quite useful to have some sort of "null" placeholder in the collection. Also, a universal null object could replace EONull in EOF. So I agree that a null object should be added to the language. It would be nice to be able to specify that ivars and loca variables are guaranteed to be init'ed to the null object. I'm not sure what kind of scheme NeXT could cook up to minimize the overhead assocaited with this scheme. : -- : Steve Dekorte : mailto:dekorte@suite.com (NeXTmail,MIME) : http://www.batech.com/~dekorte/dekorte/index.html -- jpanico@netcom.com Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc.
From: pbrown@palms.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Suggestions for memory allocation technique Date: 14 Apr 1996 01:45:55 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4kplcj$d4m@agate.berkeley.edu> I'm not a novice programmer, but I'm just warming up to C and objective C on the NeXTs. I need to write a program that will allocate and manage, say, 10 million long integers worth of storage, maybe more. I may not actually need that much, but it's likely that some computations will need even more. (This is for some computations related to my thesis work.) What is the preferred method to deal with such a large block of memory? Will just realloc()'ing over and over again be a suitable solution? (I would probably allocate the storage 100,000 longs at a time.) Also, I know this is messy on a 680x0 machine; would I be better off to run this on a different arch? Thanks in advance. Paul Brown
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: CVS Date: 13 Apr 1996 13:48:49 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7zq8gmjj2.fsf@tom.es.ele.tue.nl> References: <4km67h$28h@lion.embl-heidelberg.de> In-reply-to: tuparev@EMBL-Heidelberg.DE's message of 12 Apr 1996 18:08:49 GMT In article <4km67h$28h@lion.embl-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) writes: Which version of CVS can handle nib files properly and use remote repository? As far as I know, the combination of the two (i.e. wrappers to a remote repository) does not work and never has worked. --Tiggr
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 11 Apr 1996 02:18:21 GMT Organization: Genoa Software Systems Message-ID: <4khq5d$6sv@saturn.genoa.com> References: <4k8uit$5fl@sun0.urz.uni-heidelberg.de> <4kbdss$819@shelob.afs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <4kbdss$819@shelob.afs.com> Gregory H. Anderson wrote: > a) I don't NEED "all these new cool concepts". They're nice, but not > necessary to get my work done. It's needless, costly work for which > I receive insufficient benefit. You receive the benefit of being able to run your apps under NT, and hopefully Solaris. I agree the conversion is a pain, but you can view it as an _opportunity_ to make an improvement pass over your own class design/implementation. (every class library can benefit from hindsight) If AFSTradeKit or whatever is so important to your company than surely you can bundle some other valuable improvements with the conversion. From my experience, things are alot simpler and more uniform in the OPENSTEP world, so you should end up with a better more maintainable class library in the end. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 11 Apr 1996 05:00:43 GMT Organization: Genoa Software Systems Message-ID: <4ki3lr$6sv@saturn.genoa.com> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <4kedd0$1e7@news.next.com> Mark Bessey wrote: > The big problem is this: Messaging a nil object causes a chain of > useless results to propagate through your program. By the time you get > to somewhere where ignoring the problem is no longer acceptable, you > have no idea where the failure occurred. 1. The fact is that messages sent to nil silently do nothing is part of the language definition. It isn't going to change. That fact has both good and bad ramifications. 2. The "bad" ramification you seem concerned with is that one class of programming error may be detected later in the execution flow than when they are first caused. This is true, and there are several ways to guard against it, such as to use assertions and explicit tests for nil. So if you have some code where a nil argument would be an error, you would suggest that one should test for nil and detect/handle the error. Fair enough. You will be happy to know that the new foundation classes from NeXT raise exceptions when passed nil arguments in many situations, rather than just silently return. 3. There are also situations where a nil argument or variable would be acceptable as a degenerate case. A common situation is the free or dealloc methods when you are freeing a set of ivars that may or may not have been lazily allocated. Other situations involve iterating over lazily allocated containers. Or set methods. In this case, (when you decide in advance that a nil receiver is acceptable), the fact that messages to nil are ignored is beneficial, in reducing some amount of code clutter. Given the fact that the language definition is not going to change in this regard, why should you not take advantage of the feature in those cases it helps matters? I see no real benefit of the if part of the following line: if (someObject) [someObject free]; -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: yibwu@plato.sky.bdm.com(Yibing Wu) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Use NeXT cc to compile C++ code---can I? Date: 13 Apr 1996 15:09:48 GMT Organization: BDM International, Inc. Message-ID: <4kog3s$2r8@news.mcl.bdm.com> Well, to be specific, the configuration is NeXT 3.3/Mach/Intel with NeXT Developer installed. Now I have some code that has been compiled by g++ for Ultrix, uses iostream and company. Can I recompile it with the cc compiler? or I have to install gcc again? Thanks, Bing
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Why does Foundation Kit introduce a "n Message-ID: <DpuG1w.HrD@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4kcict$2qo@nova.mdd.comm.mot.com> <4kk329$68d@usenet.rpi.edu> <4koqek$kjl@emerald.oz.net> Date: Sun, 14 Apr 1996 08:39:31 GMT In article <4koqek$kjl@emerald.oz.net> art@cubicsol.com (Art Isbell) writes: > So we're going to grin and bear the inconvenience of the OpenStep >conversion while continuing to ship our NS apps. Once we've made the >conversion, we hope to be much better off with apps that will run under NT >and thus will be much more commercially viable than our NS apps would ever >have been. It's true that you touch the essence of the issue here; one group of people (among which myself) is saying "it isn't how it's suppposed to be according to the rules of good software engineering and object-orientation", while the other group simply says "heck, the hell to object-orientation, if it runs on WindowsNT as soon as possible". But the association "pure Object reuse = nonportable NextStep", while "little inconvenient NSObject conversion = portable OpenStep" is false. David Stes RIACA / Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil (really long!) Date: 14 Apr 1996 13:14:56 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7ratrrr9r.fsf@tom.es.ele.tue.nl> References: <jpanicoDppG9I.KAy@netcom.com> <4kmcld$q91@news.next.com> <4kp3mi$94l@news.onramp.net> <jpanicoDptpF1.JJF@netcom.com> In-reply-to: jpanico@netcom.com's message of Sat, 13 Apr 1996 23:04:13 GMT In article <jpanicoDptpF1.JJF@netcom.com> jpanico@netcom.com (Joe Panico) writes: I'm not sure what kind of scheme NeXT could cook up to minimize the overhead assocaited with this scheme. Either `+alloc' would need to know all object references in an object to be able to initialize those ivars to point to the null object (time consuming) or the null object should be all-bits-zero, i.e. nil. Early versions of my libtl (http://www.es.ele.tue.nl/tiggr/sw/#tl) provided (with a patch to the gnu runtime) a true nil object, but in the end it turned out not to be all that useful. Also, the fact that it was incompatible with nextstep further limited its usefulness. Anyway, in my opinion, the only right way to handle a message sent to nil is to dump core. --Tiggr
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 14 Apr 1996 13:19:29 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7pw9brr26.fsf@tom.es.ele.tue.nl> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> In-reply-to: Alex Blakemore's message of 11 Apr 1996 05:00:43 GMT In article <4ki3lr$6sv@saturn.genoa.com> Alex Blakemore <alex@genoa.com> writes: You will be happy to know that the new foundation classes from NeXT raise exceptions when passed nil arguments in many situations, rather than just silently return. Apart from the fact that exceptions suck deep, the fact that one can not put a nil object in an NSArray very much decreases their usefulness. An NSArray is an abstraction of a `id a[]'. In the latter, I can put nil if I want to, so why not in an NSArray? What if one wants to add 100 objects to an array, but not at increasing indices, i.e. what if the index of an object matters? --Tiggr
From: yoda@tatooine.ppp.cis.uni-muenchen.de (Marc Guenther) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 14 Apr 1996 11:57:32 GMT Organization: Institut fuer Informatik der Universitaet Muenchen Message-ID: <4kqp7c$90h@arcadia.informatik.uni-muenchen.de> References: <316186b9.4262367@news.plsys.co.uk> <4jtit1$am5@sun0.urz.uni-heidelberg.de> <3164410a.1117449@news.plsys.co.uk> <SAMURAI.96Apr12104719@random.hasc.ca> In <SAMURAI.96Apr12104719@random.hasc.ca> Darcy Brockbank wrote: In comp.sys.next.programmer,comp.sys.next.advocacy you wrote: > Not quite... here's the data you were referring to: [...] > (i) The FoundationKit, and NeXT classes in general are fat and slow due > to poor implementation. [...] > (ii) NeXT chose non-optimal algorithms for implementing base classes. [...] > (iii) The designers of OpenStep and Foundation did nowhere near as good > a job as the original AppKit designers. Perhaps SUN had something to do with it ? :-) > Let's face it... OpenStep as an API sucks. It's all we have for the next > little bit, but the writing is finally absolutely on the wall now. With > Sun backing off OpenStep, and moving into Java, NeXT's technology > becomes less meaningful every day... [munch, munch] Well, all I can say, I'm glad that there are at least some sane people still around. I quite liked the new Foundation, and I still prefer it when starting a new program, but you raised some really good points, and all I can do is to agree with you ! Now, if only there would be anything which comes even remotely close to Nextstep, I would switch yesterday. Hmm. BTW, did you ever try the Array classes from the GNU foundation classes ? Would be very interesting indeed. Linux with Gnustep would be an alternative, but it's almost none existant as of now, and I hate PC's, I hate X-Windows, and I begin to hate Unix as well. Unfortunatley I don't see any other alternative. Marc -- If you have problems with my reply-address, please use yoda@cis.uni-muenchen.de Thanks
From: yoda@tatooine.ppp.cis.uni-muenchen.de (Marc Guenther) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Why does Foundation Kit introduce a "n Date: 14 Apr 1996 11:59:08 GMT Organization: Institut fuer Informatik der Universitaet Muenchen Message-ID: <4kqpac$90h@arcadia.informatik.uni-muenchen.de> References: <4jufkv$jbc@shelob.afs.com> <4juvbc$2i4@nova.mdd.comm.mot.com> <4kmg6m$f90@SantaClara01.News.Internex.NET> In <4kmg6m$f90@SantaClara01.News.Internex.NET> Lee Willis wrote: > besler@mdd.comm.mot.com (Steve Besler) wrote: > > [snippage] > > >It seems to me that the more legacy code people have laying > >around, the more they want to stick with "the old way" of doing > >things. Not because the old way is "better" but because of a > >mixture of laziness and greed... > > And golly, what an unreasonable and contempible attitude that is, too! > > I got news for you, Steve: "Laziness and greed" is the whole point. > The goal of this game called business is to get the most return > (greed) from the least investment (laziness). That's how you > stay in business. That's how you make money. That's how you > pay your employees. That's how you buy new equipment. That's > how you have capital to invest in R&D. ... > > Anything that allows you to get more return, or put in less > investment is a *GOOD* thing. Anything that decreases > your return, or increases your investment, is a *BAD* thing. And that's exactly why crap like Windows 95, Intel PC's, X-Windows, WWW and Java succeeds, and interesting things like NeWS, Nextstep, Amoeba, Hyper-G,... don't. To do some really cool stuff, you have to invent a lot of new things. Most probably you can't build on anything already there. Anyway, it will cost you a lot of money, more than you would have to spent when you just extend existing bad technology. So that's what happens all the time. A simple truth when doing business in computer science is: "If there is a good product, and a bad product, more people will buy the bad product". I wonder if this is true for other marketplaces too. Some people say it's not enough to have a technollogically better product than the competition. That's not even true, it doesn't matter at all. This, I think, is really sad. Obviously, not all people think that way. NeXT seems to have learned that lession. (I mean, in the beginning they thought, a good product is ALL you need to be successful. Now really !) That's why they are only doing nonsense these days, WebObjects, Openstep on Windows (the joke that kills), OLE, Corba..., instead of doing the really important and interesting (at least to me) stuff like objectoriented filesystem, incremental compiler/linker, language aware graphical source/interface editor, really distributed environment, etc, and most important finally getting rid of Unix. But well, toys like these are not very interesting nowadays. I mean, we have much cooler stuff, like webpages with little animations in them, and distributed full text indexing (Alta Vista), errr... Marc --- If you have problems with my reply-address, please use yoda@cis.uni-muenchen.de Thanks
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer Subject: NXAutoreleaseConnection problem: timeouts Date: Sat, 13 Apr 1996 18:16:13 -0400 Organization: Steel Driving Software, Inc. Message-ID: <317027AD.64C6@ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Folks, I've got a system which uses DO and Foundation. I'm using a multithreaded server (based on the MTS example). If I use NXConnections, it all works, but I get warnings about there being no autorelease pool. (Just as I'd expect) When I use NXAutoreleaseConnections, I start getting timeout exceptions, bringing things to a screeching halt. Any tips on what it is about NXAutoreleaseConnection that could be hanging me? Are there any things that it's sensitive about? (BTW, I'm using NS 3.3 with EOF 1.2, and with the Foundation User Patch installed) Thanks, JH
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Suggestions for memory allocation technique Date: 14 Apr 1996 13:29:02 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4kquiu$6c4@lion.embl-heidelberg.de> References: <4kplcj$d4m@agate.berkeley.edu> In article <4kplcj$d4m@agate.berkeley.edu> pbrown@palms.berkeley.edu (Paul Robert Brown) writes: > I'm not a novice programmer, but I'm just warming up to C and objective C > on the NeXTs. > > I need to write a program that will allocate and manage, say, 10 million > long integers worth of storage, maybe more. I may not actually need that > much, but it's likely that some computations will need even more. (This > is for some computations related to my thesis work.) What is the > preferred method to deal with such a large block of memory? Will just > realloc()'ing over and over again be a suitable solution? (I would > probably allocate the storage 100,000 longs at a time.) It depends strongly on what you are going to do with all these integers. If you are going to access them sequentially (typical case for statistics), use a memory stream. If you are going to do some complex (say matrix) aurithmetix, it's better to allocates different junks of memory each in a separate zone. I think it's more algorithmic problem then allocation.... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: NXAutoreleaseConnection problem: timeouts Date: 14 Apr 1996 13:38:37 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4kqv4t$6gf@lion.embl-heidelberg.de> References: <317027AD.64C6@ix.netcom.com> In article <317027AD.64C6@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> writes: > Folks, > > I've got a system which uses DO and Foundation. I'm using a > multithreaded server (based on the MTS example). > > If I use NXConnections, it all works, but I get warnings about > there being no autorelease pool. (Just as I'd expect) > > When I use NXAutoreleaseConnections, I start getting timeout > exceptions, bringing things to a screeching halt. > > Any tips on what it is about NXAutoreleaseConnection that > could be hanging me? Are there any things that it's sensitive > about? > > > (BTW, I'm using NS 3.3 with EOF 1.2, and with the Foundation > User Patch installed) I'm using EOF 1.1 and have no problem. I have no idea why you got these timeouts, but here couple of hints how you may find the bug: - Run the connection with infinite timeout. Perhaps it's just messaging of a freed object... - Use the debug method of NXAutoreleaseConnection: Add these 3 lines somewhere: @interface NXConnection (Debug) + (void)debug:(const char *)header; @end and to start debugging: #ifdef DEBUG [NXAutoreleaseConnection debug:"[Process name]"]; #endif DEBUG before any othor message. -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: nelson@santafe.edu (Nelson Minar) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 14 Apr 1996 11:59:17 -0600 Organization: Self-organized Sender: nelson@nelson.santafe.edu Message-ID: <ozd95au1oq.fsf@nelson.santafe.edu> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> In-reply-to: Alex Blakemore's message of 11 Apr 1996 05:00:43 GMT Alex Blakemore <alex@genoa.com> writes: >1. The fact is that messages sent to nil silently do nothing is part of the >language definition. It isn't going to change. It could easily change with gcc. A lot of us have been asking for some option in the runtime so we could override the behaviour of messages to nil. Have it default to return silently, but at least let us change it if we want. >That fact has both good and bad ramifications. Right, and they boil down to "you need to put in explicit tests to catch bad code, but do nothing for good code to work". I think the other way around: you shouldn't need to do anything for bad code to show up, but you may have to put in checks for good code to work. For instance, you say that freeing nil is a common thing: [someObjectThatIsNil free] and that checking it is a lot of trouble: >I see no real benefit of the if part of the following line: >if (someObject) [someObject free]; Whereas, as a defensive programmer, I work the opposite way. I want messages to nil to be errors. Then, if you know that something could be nil sometimes, you should put in the check. And when you blow a piece of code and messages go to nil, it automatically fails - the programmer doesn't have to preantcipate errors by putting in checks. It's a serious quality-of-code issue for my project. I'm writing a set of tools for building simulations - other people, many of them not experienced programmers, take our Objective C libraries and build simulations with them. I can think of four cases already where a person's bugs would have been caught if they could have messages sent to nil fail. I'd like to make that the default for our users. -- __ nelson@santafe.edu \/ http://www.santafe.edu/~nelson/ PGP key 9D719FAD Fingerprint 3B 9B 8E 58 1C 90 57 3E B7 99 ED 13 65 2E 0B 24
Newsgroups: comp.sys.next.bugs,comp.sys.next.hardware,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,de.alt.binaries.next From: martin@jeramy.westfalen.de Subject: this "interleaced-mode" is nearly killing me !!! Message-ID: <DpuE49.1K7@jeramy.westfalen.de> Sender: news@jeramy.westfalen.de (news) Organization: JerySoft Date: Sun, 14 Apr 1996 07:57:45 GMT Hi you out there, well I have a Hercules Dynamite VL Pro VGA-grphic card installed in my PC ... I am using NSi 3.3 and the only driver Ive got by now is that on, that gives you exactly 1024 x 768 RGB by 45 Hz INTERLEACED !!! I think its revision b by a modified NS 3.2 driver (from Ricardo J .Parada and John Ruggentaler) So, pleas, tell me that there is a newer on without this interleace in a high color standard !!! Where can I get it from ?!? -- ... M.Roggon (martin@jeramy.westfalen.de)
From: nisse@lysator.liu.se (Niels Möller) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 15 Apr 1996 00:22:42 +0200 Organization: Lysator Computer Society, Link|ping University, Sweden Sender: nisse@lysistrate.lysator.liu.se Message-ID: <nnhgumxx71.fsf@lysistrate.lysator.liu.se> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <x7pw9brr26.fsf@tom.es.ele.tue.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit In-reply-to: tiggr@es.ele.tue.nl's message of 14 Apr 1996 13:19:29 +0200 tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: > Apart from the fact that exceptions suck deep, What's bad with exceptions? Do you dislike the concept in general, or just the NeXT style of exceptions? Personally, I haven't decided yet if I like exceptions or not. But I've been programming in python lately, and have gotten used to them. Also, I wrote up a new implementation of exceptions for C and Objective-C some months ago (hopefully to be included into libobjects soon), that should be much more flexible than the NeXT exceptions. Look at http://www.lysator.liu.se/~nisse/archive/exceptions-0.5.[readme|tar.gz] if you're interested (this is not quite the final version, I made some more changes while integrating it into libobjects). /Niels
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 12 Apr 1996 03:36:42 GMT Organization: Genoa Software Systems Message-ID: <4kkj4a$8f8@saturn.genoa.com> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <jpanicoDppG9I.KAy@netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <jpanicoDppG9I.KAy@netcom.com> Joe Panico wrote: > : Alex Blakemore <alex@genoa.com> [you claim] writes > : ... The "right thing" to do on messaging a nil object is throw an exception. ... >> if this is the case, then why do NeXT software engineers themselves >> message nil objects? Wait a minute Joe. You've switched the attributions. Mark Bessey wrote those words. I send messages to nil all the time without guilt. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 15 Apr 1996 04:15:12 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4ksigg$hur@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: NSData in NSArray causing Malloc errors (code=5) Date: Mon, 15 Apr 1996 09:33:58 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960415091813.9001A-100000@hphalle0.informatik.tu-muenchen.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hello, I discovered a problem with the use of NSData xNoCopy method. I'm doing something similar like this: (running in a seperate thread) -------------- char *buffer = vm_alloc(...); // alloc length bytes NSData *data; data=[NSData dataWithBytesNoCopy:buffer length:length]; [myArray addObject:data]; a seperate thread would dealloc the data in the different NSData objects of the array and release the NSData objects (after removeing form array). compare this to: ... data = [NSData dataWithBytes:buffer length:length]; [myArray addObject:data]; vm_deallocate(task,buffer,length); a seperate thread would again dealloc the data in the NSData objects and release the NSData objects (after remove from the array). -------------- I hope the mothed is clear, even if the code isn't 100% correct (writing from memory). When using the NoCopy method, I'll free the data in the NSData object later if it's no longer used. If I use copy method I do free the original data at once. I don't use 'autorelease' because it doesn't suit me well in several threads. The first runs ok, but gives Malloc error (code=5) on the console. The second runs without any problems, causing no memory leaks. Adding [data autorelease] to any code would cause memory leaks, because the thread running the code will only exit on programm exit and the producer thread in my case _doesn't_ release NSData objects. Any hints? BTW: how is the correct mechanism of handling autorelease objects in a multi-threaded environment? Best regards, Boerny. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: Mon, 15 Apr 1996 10:52:51 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960415105203.9702F-100000@hphalle0.informatik.tu-muenchen.de> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <x7pw9brr26.fsf@tom.es.ele.tue.nl> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <x7pw9brr26.fsf@tom.es.ele.tue.nl> On 14 Apr 1996, Pieter Schoenmakers wrote: > I want to, so why not in an NSArray? What if one wants to add 100 objects > to an array, but not at increasing indices, i.e. what if the index of an > object matters? --Tiggr Use the supplied hash value? Greetings, Boerny. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: matti@dbag.ulm.DaimlerBenz.COM (Matthias Oberlaender) Newsgroups: comp.sys.next.programmer Subject: Re: pswrap programming: How to transfer large image data? Date: 15 Apr 1996 08:26:34 GMT Organization: debis Network Services GmbH Message-ID: <4kt17q$bi5@news.sns-felb.debis.de> References: <4kj0d7$4to@news.sns-felb.debis.de> <4kl60f$62r@newshost.uni-koblenz.de> <4klh3s$9tv@news.sns-felb.debis.de> <ku7spe9e9oe.fsf@lamancha> <4kngkq$ev7@news.cs.tu-berlin.de> In <4kngkq$ev7@news.cs.tu-berlin.de> Marcel Weiher wrote: | the AppKit provides a pretty convenient way of doing this, | while also taking advantage of the fast Mach based image | data transfer mechanisms. For example, I have a method | for drawing images in calibrated instead of device | colorspaces: | | -(BOOL)drawCalibrated:(const char*)colorSpace | { | NXColorSpace old=_colorSpace; | BOOL retVal; | | _colorSpace = NX_CustomColorSpace; | DPSPrintf( DPSGetCurrentContext(), "\n %s setcolorspace \n",colorSpace); | retVal = [self draw]; | _colorSpace = old; | return retVal; | } | | (It is in a category of NXBitmapImageRep) | | In a drawSelf:: method. I then call it as follows: | | if ( image ) | [image drawCalibrated:" [ /Indexed /DeviceGray 255 { 255 div 1 exch sub } ] "]; | | Works like a charm. Yeah, it really works like a charm. It also seems to be the shortest and cleanest way of doing what I originally had in mind, namely forcing the appkit to use a user-supplied color look-up table. Thanks a lot ! Matthias -- ____ __ _/_/ . ( / / ( / / / / ===================================================================== Matthias Oberlaender, Daimler-Benz AG, Research Center Ulm F3M/T (Information Technology / Text Understanding) Wilhelm-Runge-Str. 11, P.O. Box 2360, 89013 Ulm, Germany Phone: +49 731 505 2354 Fax: +49 731 505 4113 Email: oberlaender@dbag.ulm.DaimlerBenz.COM =====================================================================
From: paul@plsys.co.uk (Paul Lynch) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Why does Foundation Kit introduce a "n Date: 15 Apr 1996 09:03:31 GMT Organization: P & L Systems, Ltd. Message-ID: <4kt3d3$cf@ironhorse.plsys.co.uk> References: <4jufkv$jbc@shelob.afs.com> <4juvbc$2i4@nova.mdd.comm.mot.com> <4kmg6m$f90@SantaClara01.News.Internex.NET> <4kqpac$90h@arcadia.informatik.uni-muenchen.de> In <4kqpac$90h@arcadia.informatik.uni-muenchen.de> Marc Guenther wrote: > In <4kmg6m$f90@SantaClara01.News.Internex.NET> Lee Willis wrote: > I got news for you, Steve: "Laziness and greed" is the whole point. > > The goal of this game called business is to get the most return > > (greed) from the least investment (laziness). That's how you > > stay in business. That's how you make money. That's how you > > pay your employees. That's how you buy new equipment. That's > > how you have capital to invest in R&D. ... > > > > Anything that allows you to get more return, or put in less > > investment is a *GOOD* thing. Anything that decreases > > your return, or increases your investment, is a *BAD* thing. The timescale of your expections matters a lot to this sort of business model. If you need a short term return, then you are right. If you are prepared to consider a longer time span, then you might choose to invest in the short term to receive a greater long term return. Otherwise you will never train staff, for instance. > And that's exactly why crap like Windows 95, Intel PC's, X-Windows, WWW and > Java succeeds, and interesting things like NeWS, Nextstep, Amoeba, > Hyper-G,... don't. The difference between a strictly short term view on profits and a longer term one is called "Vision" (in the US and UK, at least). Vision is what lets people sweat for years in a start-up; it has nothing to do with any kind of mystical anti-business thing, contrary to impressions. Rather it is all about having a better understanding of the long term development of a business than typical big company investors and stock market analysts, who are almost exclusively interested in short term dividends. > NeXT seems to have learned that lession. (I mean, in the beginning they > thought, a good product is ALL you need to be successful. Now really !) > That's why they are only doing nonsense these days, WebObjects, Openstep on > Windows (the joke that kills), OLE, Corba..., instead of doing the really > important and interesting (at least to me) stuff like objectoriented > filesystem, incremental compiler/linker, language aware graphical > source/interface editor, really distributed environment, etc, and most > important finally getting rid of Unix. Applause!! If you look at NeXT's shifting emphasis in recent years, and consider the Wired interview with Steve Jobs at the end of last year, you will see that the big difference is a shortening of acceptable time scales for returning profits. Paul -- Paul Lynch (NeXTmail) paul@plsys.co.uk Tel: (01494)432422 P & L Systems Fax: (01494)432478 http://www.plsys.co.uk/~paul
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Exceptions (was: Re: when an expression evaluates to nil) Date: 15 Apr 1996 12:12:36 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7g2a5yewb.fsf_-_@tom.es.ele.tue.nl> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <x7pw9brr26.fsf@tom.es.ele.tue.nl> <nnhgumxx71.fsf@lysistrate.lysator.liu.se> In-reply-to: nisse@lysator.liu.se's message of 15 Apr 1996 00:22:42 +0200 In article <nnhgumxx71.fsf@lysistrate.lysator.liu.se> nisse@lysator.liu.se (Niels Möller) writes: What's bad with exceptions? Do you dislike the concept in general, or just the NeXT style of exceptions? I like the functionality of and possibilities offered by exceptions; in some circumstances they're invaluable and to be preferred very much over setjmp and longjmp. The problem I have with exceptions is that, >99% of the times they are misused. For instance, an NSArray raising an invalid whatever exception because an index is out of bounds is a nonsensical application of exceptions, as such an error is an error and not some occasion which one might want to handle. The programmer is better off fixing the reason of the error than putting handlers anywhere to handle it. In this case, the array should preferrably output some message and, certainly, call abort. (I won't mention the bizar behaviour of putting a nil object reference in an NSArray causing an exception to be raised. I can come up with reasons why inserting nil into an NSArray would be a valid (and valuable) action; I can't think of any reason why nil should not be allowed to be put in an NSArray; and I'm pretty sure I can refute any argument put forward in favour of it.) Another example is the exception which is raised by NeXT's distributed objects if, after invocation of a (local) method which was invoked as the result of a DO call, the return values of the invocation can not be delivered back because the connection was invalidated. If the connection were invalidated one instance after the results had been delivered (i.e. scheduled to be sent out over the net to the other machine), no exception would have been raised and some invalidation notification would be posted, and, moreover, the result would be exactly the same: the calling party did not get a chance to digest the results. Things are even worse than this: the same exception can also raised for _oneway_ void methods. In this case, obviously, the code raising the exception should be deleted. Anyway, there is one more serious problems with exceptions, if one were to actually use them: When an exception is not caught properly, postconditions along the method invocation path from the handler to the exception raiser are not necessarily met. Basically this means that every method invocation <EM>should</EM> be encapsulated by a handler when an exception during the invocation would cause a violation of the postcondition of the current method, statement or compound statement. Failing to install a handler since `that method will not raise an exception' implies knowledge of the internal workings of a method implementation, violating the concepts of object oriented programming since in general one can not know the inner workings of a method based on the fact that one knows the method's name and thinks to know the class (and thus implementation) of the receiver. A solution to the unexpected exception raising is to write methods in such a way that any (implicit) postconditions it has are always met, even if any method being invoked raises an exception. But writing methods that way is cumbersome and not worth the effort. When are exceptions OK? Emacs' C-g is a good example. Most other exceptions in Emacs are present because Emacs provides a programming environment and the user, when s/he has become the programmer, can make programming mistakes which Emacs must handle. --Tiggr
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 15 Apr 1996 06:06:26 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4ksp13$bo8@news.its.com> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> MaRK_BeSSeY@NeXT.CoM (Mark Bessey) wrote: > Alex Blakemore <alex@genoa.com> writes >> In <316942D0.6038@NeXT.com> Matt Watson wrote: >>> It is always faster and often more correct to check for nil before >>> sending a message to an object which is possibly nil. Why is it always faster? Checking for nil is generally going to a single instruction (BNE on a register or stack offset). If you can have the compiler inline objc_msgSend or have it available as macro, you can implement the Check-If-Nil test as the first thing you do. So it should be the same speed, right? If the size of objc_msgSend is excessive for inlining it everywhere, you could still macroize or inline just the Check-If-Nil part and do a function call to the actual objc_msgSend functionality if the receiver is not nil. >> What does "more correct" mean? > > It means that the output of your program is more likely to be > correct.(what did you think it meant? :-)) Returning nil on messages to > nil objects is *not* a benign feature of Objective-C. It's a definite > rough edge. It would be nice if there was an option to control whether the Obj-C runtime raised an exception upon messaging a nil object or whether it simply ignores it. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 15 Apr 1996 12:45:16 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7enppyddv.fsf@tom.es.ele.tue.nl> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <x7pw9brr26.fsf@tom.es.ele.tue.nl> <Pine.HPP.3.91.960415105203.9702F-100000@hphalle0.informatik.tu-muenchen.de> In-reply-to: Bernhard Scholz's message of Mon, 15 Apr 1996 10:52:51 +0200 In article <Pine.HPP.3.91.960415105203.9702F-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: On 14 Apr 1996, Pieter Schoenmakers wrote: > I want to, so why not in an NSArray? What if one wants to add 100 objects > to an array, but not at increasing indices, i.e. what if the index of an > object matters? --Tiggr Use the supplied hash value? No way. Not general. Nonsensical detour. Much to expensive. --Tiggr
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Known float bugs? Date: 13 Apr 1996 18:11:27 GMT Organization: Turbocat's Development, Germany Message-ID: <4koqof$4ci@turbocat.snafu.de> Hi! Are there known float bugs? If I do a "myfloat=(myint/100.0)" then myfloat seems to be incorrect in some cases. NXSwapHostFloatToBig and NXSwapBigFloatToHost seem to have problems as well. Any hints? Tia. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <alex@conextions.com> Message-ID: <9604151338.AA00613@conextions.com> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Alexander Elkin <alex@conextions.com> Date: Mon, 15 Apr 96 09:38:05 -0400 Subject: Re: Why does Foundation Kit introduce a "n Www: http://www.conextions.com Dear developers, do you know any development environment, combining the short development circle with backward compatibility for 5-7 years? If yes, let's just switch to it. Unfortunately, I don't know such. It seems for me the NeXT approach is not too bad. Yes, they do major changes from time to time, that requires (let's be realistic and not rely too much on the conversion scripts) to rewrite the custom software. But it pays off. Even if we put aside the benefits from the cleaning up the code and getting the better environment, the overall development circle is still shorter. I was using C++ under DOS and Windows a while ago. To develop an 50,000 lines application I had to spend at least 6-9 months and the maintainance problems were very big. Doing the same with NeXT, you may get a pretty good result in 3 months. Even if you'll have to rewrite it every 2-3 years, it saves you time (and company's money). Once again, I'm not NeXT's advocate. We all know a lot of problems with support, etc. But making the choice between backward compatibility and better environment, my vote is for the last one. Alex Elkin. Conextions, Inc alex@conextions.com (508)689-3570
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil Date: 15 Apr 1996 15:04:51 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4ktoij$k3d@emerald.oz.net> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <ozd95au1oq.fsf@nelson.santafe.edu> nelson@santafe.edu (Nelson Minar) wrote: > It could easily change with gcc. A lot of us have been asking for some > option in the runtime so we could override the behaviour of messages > to nil. Have it default to return silently, but at least let us change > it if we want. Seems that you could simply redefine objc_msgSend() to get the behavior that you want (not tested!): id objc_msgSendOrig(id theReceiver, SEL theSelector, ...); id objc_msgSendNilError(id theReceiver, SEL theSelector, ...) { if (theReceiver != nil) { return objc_msgSendOrig(theReceiver, theSelector, dealWithVarargs); } else { return someSortOfError(...); } } Then early in your app's execution: objc_msgSendOrig = objc_msgSend; objc_msgSend = objc_msgSendNilError; -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: dirk@kalium.physik.TU-Berlin.DE (Dirk Schwarzhans) Newsgroups: comp.sys.next.programmer Subject: Better timing than 10ms possible???? Date: 15 Apr 1996 16:09:46 GMT Organization: Technical University Berlin, Germany Message-ID: <4ktsca$n3f@brachio.zrz.TU-Berlin.DE> Hello, I want to execute parts of my program in regular intervals of less than 10ms. On NSfIP systems the timing seems to be not better than 10ms (starting with release 3.1!). Even the function ns_timeout() in the kernel (which was renamed from us_timeout to ns_timeout (!)) doesn't provide more accuracy than 10ms. Any idea how to get more accurate timing? Dirk Schwarzhans
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil (really long!) Date: 15 Apr 1996 18:42:43 GMT Organization: ALI Message-ID: <4ku5b3$3db@cetus.ali.bc.ca> References: <jpanicoDppG9I.KAy@netcom.com> <4kmcld$q91@news.next.com> In comp.sys.next.programmer,comp.lang.objective-c Mark Bessey wrote: > Here's my attempt at a "better behaved" version: > - (void)setTYPE_CODE:(NSString *)value > { > if (nil != TYPE_CODE) > [TYPE_CODE autorelease]; > if (nil != value) > { > TYPE_CODE = [value retain]; > } > else > { > value = nil; > [NSException raise: NSinvalidArgumentException > fromat: @"setTYPE_CODE: value cannot == nil"]; > } > } This brings up another point that has bugged me about a lot of NSObject code examples I have seen. In his "improved" version, you will notice that Mark has switched to using autorelease rather than release (in addition to checking for nil - the original point of the thread). I can't count the number of times I've seen people give example code like: - (void)setFoo:newFoo { [myFoo release]; myFoo = [newFoo retain]; } while I write either: - (void)setFoo:newFoo { id oldFoo = myFoo; myFoo = [newFoo retain]; [oldFoo release]; } or, more clearly, - (void)setFoo:newFoo { // If the new value is the same as the old... if (newFool == myFoo) { return; // do nothing } [myFoo release]; myFoo = [newFoo retain]; } I do this because I have used my own equivalent to retain and release for about a year before NeXT introduced EOF, and in my implementation, release would immediately deallocate myFoo if it was not otherwise retained, thus the call [self setFoo:myFoo]; would cause a problem when the first version of setFoo: tried to retain the object which it had just released. Mark's use of autorelease would seem to be a third (perhaps less memory efficient) way of dealing with the same issue. Nothing in the NSObject documentation seems to assure that NeXT's or another vendor's OpenStep implementation would be required different from my kit. So, my question is: shouldn't we be watching for set...: arguments which are the same object we are currently retaining? -- Allan Noordvyk, Software Artisan e-mail: allan@ali.bc.ca ALI Technologies Voice: 604.279.5422 x 317 Richmond, Canada Fax: 604.279.5468 * NeXT and MIME mail welcome *
From: esky@marathon.cs.ucla.edu (Eskandar Ensafi) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: Re: Using -unmounting:ok: (I can't get any results!) Date: 15 Apr 1996 19:35:02 GMT Organization: University of California, Los Angeles Message-ID: <4ku8d6$3a3@delphi.cs.ucla.edu> From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> To: Eskandar Ensafi <esky@cs.ucla.edu> [Reposted with corrections and RTF to ASCII conversion] Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be (NeXTmail), Flanders, Belgium *** The year 2000 will be the last year of the 20th century. *** Bug Report: KBNS.33.3.007_o3.213.3o ---------- -[Application unmounting:ok:] (and hence -app:unmounting:) never invoked Description: ----------- There are three Application instance methods to monitor mounting and unmounting volumes: -mounted: and -unmounted:, which are called after [[Application workspace] beginListeningForDeviceStatusChanges]; and -unmounting:ok:, which is supposedly called if and only if the application is launched from the Workspace Manager. Apparently the former two work as described, but the latter method is never called. (There are three corresponding delegate methods called from the main method, see the documentation; the former two are called correctly, the latter was not tested.) How to Verify: ------------- Use the following in an Application subclass. Don't declare these methods, then you'll have an indication that the AppKit declares them. #ifdef TEST_KBNS_33_3_007 -app:sender mounted:(const char *)fullPath{ // from mounted: NXRunAlertPanel(NULL,"app:mounted:``%s''",NULL,NULL,NULL,fullPath); return 0; // should actually test whether super understands, then relay // note: this is not achievable using just -respondsTo:! } -app:sender unmounted:(const char *)fullPath{ // from app:unmounted NXRunAlertPanel(NULL,"app:unmounted:``%s''",NULL,NULL,NULL,fullPath); return 0; // should actually test whether super understands, then relay // note: this is not achievable using just -respondsTo:! } -(int)app:sender unmounting:(const char *)fullPath{ // from unmounting:ok: NXRunAlertPanel(NULL,"app:unmounting:``%s''",NULL,NULL,NULL,fullPath); return [super app:sender unmounting:fullPath]; } -(int)mounted:(const char *)fullPath{ // after beginListeningForDeviceStatusChanges NXRunAlertPanel(NULL,"mounted:``%s''",NULL,NULL,NULL,fullPath); return [super mounted:fullPath]; } -(int)unmounted:(const char *)fullPath{ // after beginListeningForDeviceStatusChanges NXRunAlertPanel(NULL,"unmounted:``%s''",NULL,NULL,NULL,fullPath); return [super unmounted:fullPath]; } -(int)unmounting:(const char *)fullPath ok:(int *)flag{ // if and only if application launched from Workspace Manager NXRunAlertPanel(NULL,"unmounting:``%s'' ok:",NULL,NULL,NULL,fullPath); return [super unmounting:fullPath ok:flag]; } #endif Urgency: ------- Does no immediate damage, not immediately obvious, partial workaround available. Confirmed: --------- Eskandar_Ensafi (o3.3o, reported), Raf_Schietekat (o3.2o) Details: ------- Eskandar_Ensafi tested with floppy disk on Intel and optical disk on NeXTcube. Raf_Schietekat tested with floppy disk and CD-ROM on NeXTstation. Workaround: ---------- Don't use -unmounting:ok:/-app:unmounting:, use -unmounted:/-app:unmounted: instead (that's like using a <Did< method instead of a <Will< method). You can prevent a file system from being unmounted by keeping a file descriptor open to something on that file system as long as required. But then of course you also can't present a panel to the user to ask permission to unmount, and the user may be confused by what application it is that is preventing him/her from unmounting that file system, so this is not very friendly. Cure: (for NeXT) ---- Cure -unmounting:ok:/-app:unmounting:. Don't assume it is not required anymore. Report History: -------------- Eskandar_Ensafi reported this to c.s.n.programmer on 1996-04-10. Comment: ------- The AppKit Release Notes specific to 3.0 (as viewed on 3.2) say ``The following Speaker/Listener messages have been obsoleted and replaced with improved API described below under "Interapplication Image Dragging" and "Workspace Protocol". '', mentioning -unmounting:ok:. But there is no indication elsewhere about this. So maybe NeXT just stopped supporting this method in 3.0? But it is certainly not obsoleted by the new ones (as described in the KBNS Workaround for this item)! ------------------------------------------------------------------------------ Eskandar Ensafi Object-Oriented Software Engineer University of California, Los Angeles Department of Biomathematics esky@cs.ucla.edu (ASCII, MIME, NeXT) School of Medicine http://www.cs.ucla.edu/csd-lanai/fweb/esky ------------------------------------------------------------------------------ -- "Was ich kann und was ich konnte Weiss ich gar nicht mehr Gib mir wieder etwas schones Zieh mich aus dem Meer" -- Marian (The Sisters Of Mercy)
From: drbob@milquetoast (Robert E Beaty) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: Problem connecting to Oracle7 with EOF Followup-To: comp.sys.next.sysadmin,comp.sys.next.programmer Date: 15 Apr 1996 18:03:59 GMT Organization: First Chicago NBD Corporation, Chicago IL, USA Message-ID: <4ku32f$ail@news.fnbc.com> References: <4kb999$jp@natasha.rmii.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Matt Bezark (xela@acm.org) wrote: : I have an Oracle7 server running on an HP 712 ... : Do I need any Oracle or NeXT software on the : client besides the Oracle adaptor that is : Oracle specific? Nope. : Do I need an entry in : /etc/services on the client? I believe so. When I installed my Oracle Enterprise Server for NT, I needed to give the listener a tcp port - 1525 I believe. In any case, I could connect to it from my NT box, but not from a NeXTStep 3.3 or OPENSTEP/Mach 4.0 PR2 box until I put the line for: orasrv 1525 tcp in the services. Acutally, I did it via nidump/niload, but it's the same thing. : Thanks, : Matt No problem! Let me know if I can help more... We're trying EOF 2.0 to Sybase and Oracle, and it looks very interesting. -- Cheers, Bob (drbob@cmg.fcnbd.com) Architecture Nazi "No tools for You! Come back - One Year!" Seinfeld First Chicago - NBD - Technical Architecture Group Capital Markets Systems - (312)732-3087
From: Mark_Bessey@next.com (Mark Bessey) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Exceptions (was: Re: when an expression evaluates to nil) Date: 15 Apr 1996 20:46:33 GMT Organization: NeXT Software, Inc. Message-ID: <4kucj9$i7u@news.next.com> References: <x7g2a5yewb.fsf_-_@tom.es.ele.tue.nl> Pieter Schoenmakers writes > In article <nnhgumxx71.fsf@lysistrate.lysator.liu.se> > nisse@lysator.liu.se (Niels M ller) writes: > > What's bad with exceptions? Do you dislike the concept in general, or > just the NeXT style of exceptions? > > I like the functionality of and possibilities offered by exceptions; in > some circumstances they're invaluable and to be preferred very much over > setjmp and longjmp. The problem I have with exceptions is that, >99% of > the times they are misused. > > For instance, an NSArray raising an invalid whatever exception because > an index is out of bounds is a nonsensical application of exceptions, as > such an error is an error and not some occasion which one might want to > handle. The programmer is better off fixing the reason of the error than > putting handlers anywhere to handle it. > > In this case, the array should preferrably output some message and, > certainly, call abort. Wow. That's a little harsh, isn't it? At least with an exception you can catch it and try to recover. What if the NSArray being accessed was only used for some minor display function? In that case, aborting the program might be an overreaction. > (I won't mention the bizarre behaviour of putting a nil object > reference in an NSArray causing an exception to be raised. I can come > up with reasons why inserting nil into an NSArray would be a valid (and > valuable) action; I can't think of any reason why nil should not be > allowed to be put in an NSArray; and I'm pretty sure I can refute any > argument put forward in favour of it.) Well, you did mention it, so...There is no such thing as a *nil object*, according to Foundation. Since the NSArray is supposed to be a collection of objects, you can't put nil references in there. You can get the functional equivalent in many cases by using a null object, like the EONull class from Enterprise Objects. > Another example is the exception which is raised by NeXT's distributed > objects if, after invocation of a (local) method which was invoked as > the result of a DO call, the return values of the invocation can not be > delivered back because the connection was invalidated. Seems reasonable so far - you might want to attempt some kind of error recovery at that point... > If the connection were invalidated one instance after the results had > been delivered (i.e. scheduled to be sent out over the net to the other > machine), no exception would have been raised and some invalidation > notification would be posted, and, moreover, the result would be exactly > the same: the calling party did not get a chance to digest the results. Ok, now I get it. The behavior is different depending on when the connection becomes invalid? Maybe someone who actually understands DO would like to comment on this, but I would bet it was easier (or faster) to do things this way. Probably *not* the ideal situation, though. > Things are even worse than this: the same exception can also raised for > _oneway_ void methods. > In this case, obviously, the code raising the exception should be > deleted. Really? That *does* sound like a bug. > Anyway, there is one more serious problems with exceptions, if one were > to actually use them: > When an exception is not caught properly, postconditions along the > method invocation path from the handler to the exception raiser are not > necessarily met. Basically this means that every method invocation > <EM>should</EM> be encapsulated by a handler when an exception during > the invocation would cause a violation of the postcondition of the > current method, statement or compound statement. Well, that is a problem, but a real exception handling system would at least have the ability to restart whatever statement caused the exception, or proceed to the next statement, or exit the local method. Implementing such complex non-linear program flow is left as an exercise for the reader. Here's a hint - #define macros aren't going to cut it. > Failing to install a handler since `that method will not raise an > exception' implies knowledge of the internal workings of a method > implementation, violating the concepts of object oriented programming > since in general one can not know the inner workings of a method based > on the fact that one knows the method's name and thinks to know the > class (and thus implementation) of the receiver. Either that, or REALLY, REALLY, good documentation...Of course, that won't really work, since upstream classes could change anytime. > A solution to the unexpected exception raising is to write methods > in such a way that any (implicit) postconditions it has are always met, > even if any method being invoked raises an exception. But writing > methods that way is cumbersome and not worth the effort. Yeah, it's tough to write software that does the right thing. But plain ANSI C code that always checks for proper return codes ends up looking like this: if (foo()) { if (bar()) { bar's postconditions... } else { error recovery... } foo's postconditions... } else { error recovery... } So, maybe it's something that plagues C-style languages in general..Anybody ever used a C-derived language with exceptions that were: 1. Easy to use? 2. Comprehensive? 3. Low run-time overhead? 4. All of the above? >[stuff about Emacs... I don't use Emacs, so I won't comment] -Mark -- Mark Bessey NeXT Software, Inc Software Quality Assurance -->I DON'T SPEAK FOR NeXT <--
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: How to start a daemon on remote host? Date: 15 Apr 1996 14:38:57 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4kuflh$3be@nova.mdd.comm.mot.com> References: <4km6hp$2ca@lion.embl-heidelberg.de> In article 2ca@lion.embl-heidelberg.de, tuparev@EMBL-Heidelberg.DE (Georg Tuparev) writes: >I want to start a daemon on remote host from an app running locally. > >system("rsh hostname nohup daemon parameters"); > >works somehow, but I feel it isn't the right way.... Another way that I can think of is to have it spawned by inetd. Then again, that is kind of icky as well. Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "n Date: 15 Apr 1996 14:55:43 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4kugl0$3bg@nova.mdd.comm.mot.com> References: <4koqek$kjl@emerald.oz.net> In article kjl@emerald.oz.net, art@cubicsol.com (Art Isbell) writes: > I can imagine its >proprietary operating system being the next to go with only its proprietary >development and run-time environment remaining. But there are many examples >of successful proprietary development environments, so the OpenStep model >seems valid. I don't understand how you equate the "OpenStep model" with "proprietary development environments". To me, the idea of the "OpenStep model" is to have multiple vendors implement development environments which would all be able to produce executables from OpenStep compliant source code. Now, whether this model ends up existing in reality is a separate (very significant) issue. The model itself seems (to me) to be open rather than proprietary. Please explain, Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: Canon NEXTSTEP Developer Support-Japan Date: 15 Apr 1996 15:27:11 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4kuifv$3bk@nova.mdd.comm.mot.com> References: <4kmlbj$28u@news4.digex.net> In article 28u@news4.digex.net, John Kheit <jkheit@cnj.digex.net> writes: >besler@mdd.comm.mot.com (Steve Besler) wrote: >> God Bless Americer! If English is good enough for Steve Jobs, its good >> enough for me. Yeeeehaaaaaaw! >That's right...God bless it. Wow, one more slag at Americans, how brave >and unique.. My Goodness, now I've been branded as an unoriginal coward! Hmmm, I must have left out the smileys, I seem to have hit a nerve here. > So it never hurts to put things in english, b/c the majority of >the planet is versed in it. Now *that* is funny. :-) >Jeez, get the American battery off your shoulder already. Sorry for being a bully. I should have known better than picking on the poor down-trodden Americans. :-) Lighten up, it was supposed to be a *joke*! Cheers, Steve PS: Sorry for the wasted bandwidth, this clearly has very little to do with c.s.n.p -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: Fabien_Roy@free.fdn.org Subject: Re: Why does Foundation Kit introduce a "n Message-ID: <1996Apr15.211230.8328@free.fdn.org> Sender: news@free.fdn.org Organization: Fabien Roy Consultant, Paris, France References: <4jufkv$jbc@shelob.afs.com> <4juvbc$2i4@nova.mdd.comm.mot.com> <4kmg6m$f90@SantaClara01.News.Internex.NET> Date: Mon, 15 Apr 1996 21:12:30 GMT Lee Willis <lee@multigen.com> wrote: > besler@mdd.comm.mot.com (Steve Besler) wrote: > > [snippage] > > >It seems to me that the more legacy code people have laying > >around, the more they want to stick with "the old way" of doing > >things. Not because the old way is "better" but because of a > >mixture of laziness and greed... > > And golly, what an unreasonable and contempible attitude that is, too! > > I got news for you, Steve: "Laziness and greed" is the whole point. > The goal of this game called business is to get the most return > (greed) from the least investment (laziness). That's how you > stay in business. That's how you make money. That's how you > pay your employees. That's how you buy new equipment. That's > how you have capital to invest in R&D. ... > > Anything that allows you to get more return, or put in less > investment is a *GOOD* thing. Anything that decreases > your return, or increases your investment, is a *BAD* thing. So why I have so much trouble convincing some of my customers to use NEXSTEP/OPENSTEP/EOF (on MCCA apps)! laziness: less time to develop and maintain! greed: cheaper reusable ojects! > > >-- > >Steven Besler, Coop Student "I do not speak for Motorola." > >Motorola Wireless Data Group besler@mdd.comm.mot.com > > -- > Lee Willis MultiGen Inc. 550 S Winchester Blvd. Suite 500 > San Jose, CA 95128 (408) 261-4100 > "Experience is what you get when you don't get what you want." > -- Fabien Roy --------------------------------------------------------------------- Fabien_Roy@free.fdn.org (NextMail/MIME accepted) Fabien Roy Consultant NEXTSTEP/OPENSTEP/EOF Consultant, SYBASE DBA 10 rue de la DEFENSE 93100 MONTREUIL, France Tel: 33 1 45 28 32 23 Fax: 33 1 48 55 09 90
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer Subject: Re: Canon NEXTSTEP Developer Support-Japan Date: 16 Apr 1996 01:13:07 GMT Organization: monoChrome, Inc. Message-ID: <4kus73$ghk@news4.digex.net> References: <4kmlbj$28u@news4.digex.net> <4kuifv$3bk@nova.mdd.comm.mot.com> besler@mdd.comm.mot.com (Steve Besler) wrote: > In article 28u@news4.digex.net, John Kheit <jkheit@cnj.digex.net> writes: > >besler@mdd.comm.mot.com (Steve Besler) wrote: > > So it never hurts to put things in english, b/c the majority of > >the planet is versed in it. > Now *that* is funny. :-) The truth ofttimes is. Anyway, I apologize for my over protective fervor when it comes to anything I sense as an anti-US sentiment (I'm sure the US can likely get on, likely even better, w/o my help :) I don't know why I get touchy about it, normally I'm a jerk in private and spare people that experience :) But it always just busts one of my fuses... So my apologies to the group for wasting space, and to you Steve for not getting your sense of levity. Anyway, truce :) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: mouseDragged event not being sent Date: 16 Apr 1996 02:56:23 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4kv28n$j6o@core.bard.edu> I am writing an app to control MusicKit synthInstruments with the mouse but the app is not receiving the mouseDragged event that I have asked for. I determined this by setting a breakpoint there in gdb. The strange thing is that I am getting rightMouseDragged and all the rest of the events I am asking for. I am stumped. Any help or suggestions will be greatly appreciated. Here's the code: #import <appkit/appkit.h> #import <musickit/musickit.h> #import <musickit/synthpatches/Pluck.h> #import <musickit/pitches.h> #import "PitchWindow.h" #import "ButtonIns.h" static ButtonIns *leftButton, *rightButton; static Orchestra *theOrch; @implementation PitchWindow static void handleMKError(char *msg) { if (!NXRunAlertPanel("M2",msg,"OK","Quit",NULL,NULL)) [NXApp terminate:NXApp]; } - initContent: (const NXRect *) contentRect style: (int) aStyle backing: (int) bufferingType buttonMask: (int) mask defer: (BOOL) flag { [self makeFirstResponder: self]; // [self addToEventMask: NX_MOUSEDRAGGEDMASK]; //tried this [self addToEventMask: NX_LMOUSEDRAGGEDMASK]; [self addToEventMask: NX_RMOUSEDRAGGEDMASK]; [self removeFromEventMask: NX_KEYUP]; [self removeFromEventMask: NX_KEYDOWN]; leftButton = [ [ButtonIns alloc] init]; rightButton = [ [ButtonIns alloc] init]; // Music Kit initialization MKSetErrorProc(handleMKError); /* Intercept music kit errors. */ theOrch = [Orchestra new]; #ifdef DEBUG [theOrch setOnChipMemoryConfigDebug:YES patchPoints:0]; DSPEnableErrorFile("/dev/tty"); // extern int _DSPVerbose; _DSPVerbose = 1; #endif MKSetDeltaT(0.01); [Orchestra setFastResponse:YES]; if (![theOrch open]) { fprintf(stderr,"Can't open DSP. Perhaps some other process has it.\n"); exit(1); } [Conductor setFinishWhenEmpty:NO]; [Conductor useSeparateThread:YES]; [Conductor setThreadPriority:1.0]; [theOrch run]; [Conductor startPerformance]; [self becomeFirstResponder]; return [super initContent: contentRect style: aStyle backing: bufferingType buttonMask: mask defer: flag]; } - mouseDown: (NXEvent *) theEvent { [leftButton playNoteWithX: theEvent->location.x andY: theEvent->location.y]; return self; } - mouseUp: (NXEvent *) theEvent { [leftButton playNoteOff]; return self; } - mouseDragged: (NXEvent *) theEvent { [leftButton playUpdateWithX: theEvent->location.x andY: theEvent->location.y]; return self; } - rightMouseDown: (NXEvent *) theEvent { [rightButton playNoteWithX: theEvent->location.x andY: theEvent->location.y]; return self; } - rightMouseDragged: (NXEvent *) theEvent { [rightButton playUpdateWithX: theEvent->location.x andY: theEvent->location.y]; return self; } - appWillTerminate: sender { [Conductor lockPerformance]; [Conductor finishPerformance]; [Conductor unlockPerformance]; [theOrch close]; /* Free Orchestra resources and release the DSP. */ return self; } @end -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: RLTHOMPSON@gnn.com (Roy Thompson) Newsgroups: comp.sys.next.programmer Subject: System Engineer Opportunity in Interactive TV Date: Tue, 16 Apr 1996 00:14:05 Organization: GNN Message-ID: <4kv6sq$t7e@news-e2b.gnn.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" LOOKING FOR SYSTEM ENGINEERS/ANALYSTS A unique team of System Engineers/Analysts is currently being assembled within our "global village". Are you a System Engineer or Analyst with design or development skills in one or more of the following functional areas? Telco Network Systems, Cable or Telco Billing Systems, Signaling Systems (SS7,ISDN for example), Cable or Telco Systems Integration, Telco Intelligent Networks, Network Management, Video-On-Demand and other Media Servers, Device Drivers, Multimedia Authoring (IconAuthor, Authorware), ATM Switching Systems, SONET/SDH, Cable Head-end Systems, Cable Set-top Operating Systems (Microware), Client/Server Systems, and Service Creation and Management Systems. What are we trying to accomplish? We want to design and develop an end-to-end signaling solution for Interactive Television. This is a large-scale effort that requires many hours of analysis, design, development, testing, integration and support. This type of effort usually requires significant capitalization to implement (millions of dollars in US terms). However, we intend to produce this large-scale class of product in the same manner that thousands of small-scale products have been produced in the past. Through the efforts of individuals, or small teams in some cases, many products have been produced in a ‘garage’ environment. No budgets or funding, just good ideas and the strong will and desire to succeed on the part of the developer. Think about the number of products developed at near-zero cost. Most recently we have significant Internet-related products that are produced on a low-budget or no-budget basis. And many of the developers have invested significant hours, with the risk of little or no financial return, but with the satisfaction that only comes with such creations. Well, can it work? Can the concept of low-budget, garage-environment product development work for large-scale projects? Even in the recent past, it probably wasn’t feasible. But now, with Internet, it is possible to create an environment, even on a global scale, that is similar to what is possible locally when two or three colleagues or friends get together to co-develop some product. With proper communication and project management tools, facilitated by Internet, I believe most any software development project, large or small, is possible. We hope you are interested in working with us, as a global team, to accomplish what has never been achieved before. The model we establish with this project has the potential of making the business community rethink "traditional" approaches and methodologies to product development. Hopefully, we will be able to chronicle our experiences and lay the foundation for those that follow. What we create will be ours, collectively. Any value the asset achieves will likewise be ours, collectively. Based on a method we all agree to, we will be able to invest in our ‘garage’ project and be able to proportionately reap any future benefits. AN OPPORTUNITY One of the hottest emerging markets today is Interactive Television, or ITV. There are major ITV initiatives and trials going on around the world. Yet, as with most emerging markets, implementations vary widely. Much of the variation is dependent on the core network structure. Many telephone, cable-television, utility, wireless, and satellite-television companies have plans for ITV implementation, but the infrastructure differences dictate different approaches. In addition, their are very few standard components involved in ITV, complicating plans for service roll-out. There are multiple set-top boxes, with multiple set-top operating systems offered for each of those. The number of video and other media server products are as varied as ice cream flavors. And when everyone figures out that they need service creation or authoring components, they’ll find an exhaustive source of potential solutions. ITV has suffered significant trial defeats over the past two years or so, due in part to cost of implementation and partly by technical failures and unexpected disappointments in performance. During initial planning, several trial scenarios idealistically assumed that certain partnerships would increase the chances of success. But in implementation, it became painfully obvious that integration was a bigger problem than ever expected. Vaporware - promised compression products, inexpensive set-tops and digital video servers - has been a cause of much grief, especially when the well-funded partners came to the party only to find half-baked components from some of the less-funded and staffed partners. Many of these early alliances, where the big players jumped quickly in the pants of aggressive startups, have ended abruptly when reputations were put on the line and losses inevitable. To date, ITV has not been heralded as a major success. So where are we then with this emerging ITV market, given the poor turnout of many of the trials? Is it in fact still emerging? Frankly, I believe we are in a similar situation to what happened just prior to Microsoft Windows being developed for the PC. Needless to say, expectations for PC’s at that time were very high, although acceptance was slow. One of the contributing factors to slow acceptance, besides the fact that it was new and high-tech, was the vast array of interfaces and peripherals and the issue of compatibility. As an application designer/developer, you may be aware of the challenges that existed for PC developers in the early days before Windows. Two limiting problems existed, one related to device support and the other related to application development environment. If your software needed to interface to a peripheral of any sort, support for specific device drivers where required, often times causing significantly longer development cycles and/or minimized peripheral support. And without a common development environment, everyone made independent, non-standard interfacing decisions. Then came Microsoft Windows and almost single-handedly removed the device support and interface problems from the application developer. Today when a developer writes software for Windows and it’s common interface, applications are immediately compatible with hundreds of devices, reducing time to market and overall support. How does this situation relate to the ITV market? Seamless integration of end-to-end ITV solutions has been elusive. Of all the trials I’ve been involved in, the integration effort has been significantly underestimated. Deployable implementations, with a broadband path for data, a signaling path for service management, and a mix of set-top and media devices all integrated, are yet to be successful. Most trials and small deployments have resorted to very limited support for devices, similar to early PC application implementations. There are no real standards for interfacing to users, although several make the claim. See the similarities? The scope of the ITV problem is certainly larger than the PC problem in many respects. For example, there’s only two ways really to execute an application in Windows, either as a standard Window’s application, or as a native DOS application. With ITV however, an application can be deployed and executed in a number of different ways. Part of an application may reside in the set-top device, while a smaller piece is deployed to the network, and a third piece is deployed to the media server. With these three basic deployment targets, with primary application execution occurring at any one of those three points, application development, deployment and management becomes a uniquely challenging task. Should someone introduce a real solution for ITV integration, which properly addresses application interface and device support, one would do for ITV what Windows did for PC’s. Providing a true end-to-end solution is the answer for this emerging market, and because we know this fact, our opportunity. The solution would be based principally on the signaling path and would address application development and deployment, common human interfacing, application and network management, and standard and non-standard device support. THE PROJECT The common requirements for any type of ITV system, regardless of the network differences, are: o Broadband transport; content/media server (or feed) to content subscriber o Transaction Management o Service Creation; content packaging o Service Management; content package or application installation and updates o Network Management; broadband and signaling path o Content Servers/Feeds o Content Subscriber Devices; set-tops, PC’s, other Our team will focus only on the following areas: o Transaction Management; "level one and two gateways" o Connection and Session Management associated with service delivery o Service Creation; repackaging authored applications and integration into end-to-end o Service Management o Network Management for signaling path o Service Compatibility for set-top devices and content server/feeds We will not address: o Broadband Transport o Network Management for transport o Content Servers/Feeds o Content Subscriber Devices; including set-top operating systems The initial phase of the project plan specifies the development of a series of requirements documents, all of which collectively define the requirements for the end-to-end ITV solution. Specific methodologies have not been pre-determined, although recommendations will be presented to the team initially as a starting point. The planned documents are: Level 1 Gateway Level 2 Gateway Connection Management; Cable-television specific Connection Management; Telco/ATM specific Connection Management; Telco/Non-ATM specific Connection Management; Satellite-television specific Session Management Account Management; Cable-television specific Account Management; Telco specific Network Management; Cable-television specific Network Management; Telco specific Fault Management Resource Management Configuration Management Service Creation Integration Device Driver; access devices Device Driver; content server/feeds Quality of Service/Bandwidth Management Database Management End-to-End Architecture After the requirements are inspected and finalized, a phase for High-Level Design and Detailed Design will proceed. Then follows the two most significant challenges we will face as a team, development and integration. We will need to understand many external systems and devices to be successful, requiring communications to a number of different companies as well as other team members. Besides leveraging the Internet as much as possible, certain alliances may need to be negotiated. During the integration of our solution, we will require a physical environment with an accumulation of several devices or simulators. Creating unique relationships will be necessary, including the possibility of limited exclusives. The higher quality our Requirements and Design documents, our initial assets, the more likely teaming and sponsoring relationships will happen. A well defined project plan will be established after the team is assembled. Change control and issues management will be implemented after requirements are defined. Communications of all status, change, issues, and corporate/team activities will occur on a weekly basis through Internet email. Other resources will be experimented with, including IRC/chat and telephone service over Internet. It is likely that for certain significant events, we may be able to find sponsorship from interested companies, covering any associated costs. Again, the likelihood of this may hinge on the perceived value of our initial assets, the Requirements and Design documentation. What happens after we complete the development project is yet to be determined. But I can assure you that if we develop a workable, affordable solution, someone will use it. How valuable it becomes depends on the quality of the product and the team behind it. OTHER TEAM MEMBERS In addition to the core-team of designer/developers, we will also team with resources related to the functions of marketing, project management, and testing. As much as possible, we will leverage corporate relationships for assistance in certain areas. It may be that we will be able to work out an arrangement with a specific computer system vendor (such as Stratus or Tandem) for a test lab environment and testers, in exchange for a short-term exclusive. This type of leveraging will be scrutinized over the duration of the project, in order to maintain a near-zero cost budget. We intend to package our assets with the highest degree of professionalism, whether it’s the cover of our System Requirements Specification, or a screen on our management software. We will find team players that will ensure that professionalism. A system is only as good as it is tested. If we can’t shake our system down ourselves, our customers will, leaving us with the stigma of poor quality. A reputation for quality product will be of highest priority. We will find team members within the testing and quality assurance functions that will meet that objective. JOIN THE TEAM! If you have the experience, or are an engineer or analyst willing to learn quickly and be an asset to our team, please contact me via email soon. My address is RLTHOMPSON@gnn.com. My plan at this point is to assemble the initial "Analyst Team" as follows (Functional area, number of members): Set-top, 2 Signaling, 2 Network Management, 2 Content Server, 1 Device Drivers, 2 ATM/SONET/SDH, 2 Cable/Head-end, 1 Client/Server, 1 Telco/CO, 2 Billing Systems, 2 ITV Integration, 2 Multimedia Authoring, 2 This structure would produce a team of 21 analysts. Because I expect most to work part-time hours, this is probably a small team. We’ll re-evaluate the needs of this team as we understand the available working hours and scope of work. It is hoped that the information provided in this post will be sufficient for you to make an initial decision to send your resume and make further inquiry. If you do need more information, please contact me as soon as you can. We are trying to move things along quickly. Thanks for reading and hope you can contribute to our project! Roy
Newsgroups: comp.sys.next.programmer From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: How to start a daemon on remote host? Sender: news@novice.uwaterloo.ca (Mr. News) Message-ID: <DpxnFI.1A5@novice.uwaterloo.ca> Date: Tue, 16 Apr 1996 02:11:42 GMT Distribution: na References: <4km6hp$2ca@lion.embl-heidelberg.de> <4kuflh$3be@nova.mdd.comm.mot.com> Organization: University of Waterloo In article <4kuflh$3be@nova.mdd.comm.mot.com>, Steve Besler <besler@mdd.comm.mot.com> wrote: >In article 2ca@lion.embl-heidelberg.de, tuparev@EMBL-Heidelberg.DE (Georg Tuparev) writes: >>I want to start a daemon on remote host from an app running locally. >> >>system("rsh hostname nohup daemon parameters"); >> >>works somehow, but I feel it isn't the right way.... > >Another way that I can think of is to have it spawned by >inetd. Then again, that is kind of icky as well. > rexec(3) works very well--I've used that successfully. -- David Evans (NeXTMail OK) dfevans@bbcr.uwaterloo.ca Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/ University of Waterloo "Default is the value selected by the composer Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: hoffmann@ehmgs1.et.tu-dresden.de (Thomas Hoffmann) Newsgroups: comp.sys.next.programmer Subject: Info about dbkit available? Date: 16 Apr 1996 10:39:34 +0200 Organization: Dresden University of Technology Sender: hoffmann@ehmgs1.et.tu-dresden.de Message-ID: <u87mvgzho9.fsf@ehmgs1.et.tu-dresden.de> I try to do some ObjC programming with gnu objc on my OS/2 box. Because I want to interface some of my programs to a SQL database, I would like to know if any information about NeXT's DBKit is publicly available (design philosophy, APIs, ..). Any help is highly appreciated. Regards, Thomas. -- ============================================================================= Thomas Hoffmann E-mail: hoffmann@ehms1.et.tu-dresden.de TU Dresden Phone : +49 351 463 5399 Institut fuer Halbleiter- Fax: : +49 351 463 7172 und Mikrosystemtechnik 01062 Dresden, Germany.
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: when an expression evaluates to nil (really long!) Date: 16 Apr 1996 08:40:38 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4kvme6$n0e@btmpjg.god.bel.alcatel.be> References: <4ku5b3$3db@cetus.ali.bc.ca> Allan Noordvyk writes > I can't count the number of times I've seen people give example code like: > > - (void)setFoo:newFoo > { > [myFoo release]; > myFoo = [newFoo retain]; > } > > while I write either: > > - (void)setFoo:newFoo > { > id oldFoo = myFoo; > myFoo = [newFoo retain]; > [oldFoo release]; > } > I think this should not be necessary since release should never really destroy an object. All objects, even those with reference count 0, should survive to the end of the event loop cycle. This is the feature that allows for interfaces to not specify, who allocates the object and who releases it. I.e. you don't know the living time of an object you get through an interface and nobody needs to know about the living time of objects you have. So, if somebody passes an object to you, which is autoreleased only, and you retain it, then release (in the same event loop cycle), than its state should be again autoreleased only, not destroyed, For example, consider calling setFoo: as follows: id o1, o2; o1 = [[O alloc] init]; // you get it autoreleased o2 = [[O alloc] init]; // you get it autoreleased [otherO setFoo:o1]; ... [otherO setFoo:o2]; // releases o1 [o1 doSomething]; // if release would destroy, this fails, but it should not > or, more clearly, > > - (void)setFoo:newFoo > { > // If the new value is the same as the old... > if (newFool == myFoo) > { > return; // do nothing > } > [myFoo release]; > myFoo = [newFoo retain]; > } > In performance-critical parts, if often newFool == myFoo, you save the release/retain overhead, of course. So this is an idea, nevertheless. Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine.
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Exceptions (was: Re: when an expression evaluates to nil) Date: 16 Apr 1996 12:15:38 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7u3yk5vat.fsf@tom.es.ele.tue.nl> References: <x7g2a5yewb.fsf_-_@tom.es.ele.tue.nl> <4kucj9$i7u@news.next.com> In-reply-to: Mark_Bessey@next.com's message of 15 Apr 1996 20:46:33 GMT In article <4kucj9$i7u@news.next.com> Mark_Bessey@next.com (Mark Bessey) writes: Wow. That's a little harsh, isn't it? At least with an exception you can catch it and try to recover. What if the NSArray being accessed was only used for some minor display function? In that case, aborting the program might be an overreaction. As I said, the programmer is better off fixing the reason of the error than putting handlers anywhere to handle it. One exception to this is abort, which could be an exception which can only be handled globally. The idea behind this is that `the program should continue to run, even though some very serious programming error has occurred'. There is no such thing as a *nil object*, according to Foundation. Since the NSArray is supposed to be a collection of objects, you can't put nil references in there. You can get the functional equivalent in many cases by using a null object, like the EONull class from Enterprise Objects. NSArray does not store objects (go to C++ if you want that and all the problems and garbage caused by that). NSArray stores object references. And nil is a very nice object reference. Ok, now I get it. The behavior is different depending on when the connection becomes invalid? Maybe someone who actually understands DO would like to comment on this, but I would bet it was easier (or faster) to do things this way. Probably *not* the ideal situation, though. The behaviour is different, i.e. inconsistent. One of them is wrong, and it is not the one not raising the exception. The exception raising is _extra_ code so I don't think raising the exception is any easier. And I actually understand DO, trust me :-) > Things are even worse than this: the same exception can also raised for > _oneway_ void methods. Really? That *does* sound like a bug. It is the same bug worse, since it raises before the check for oneway (instead of after). Well, that is a problem, but a real exception handling system would at least have the ability to restart whatever statement caused the exception, I don't think exceptions, in any language, look that much like interrupts to a CPU. or proceed to the next statement, or exit the local method. Implementing such complex non-linear program flow is left as an exercise for the reader. Here's a hint - #define macros aren't going to cut it. Such complex flow is not advantageous in any way. `goto' has been considered harmful since ages. So, maybe it's something that plagues C-style languages in general..Anybody ever used a C-derived language with exceptions that were: 1. Easy to use? 2. Comprehensive? 3. Low run-time overhead? 4. All of the above? Take a look at how exceptions are used in Lisp (and Emacs). Most of the time, they're never caught by the program; they are caught by the top-level user interface, which can be helpful and let you enter the debugger or *beep* and put you back in the top-level event loop. Exceptions are not meant to be handled locally. What would be better: opening a file returns a result or opening a file causes an exception? I go for the first; it saves a lot of time setting up the error handler; if I do not wish to check the object returned, then I'll generate an abort the first time I'll try to write something to it. Or should I say, `then the NSNilReceiverException exception will be raised...'? --Tiggr
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: How to safely exit from NSThread Date: Tue, 16 Apr 1996 13:13:25 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960416130558.29309B-100000@hphalle0.informatik.tu-muenchen.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hello, I was just playing around a little bit with Foundations's NSThread class. There are a lot of things which worry me about this class, and it would be nice to hear your comments: - The exit method stops the thread at once. You don't get a chance to cleanup a thread. There is no Exceptions raised for thread exiting, right? So what is the correct method to exit a thread which works in a loop: Send it a [thread exit]: propably not, the thread might to somethin important. Setting a condition either via NSLock or an instance variable. Not nice, too, because you need another instance variable in the controller of your thread. - How do you change the sheduling policy of a thread? Foundation declares A thread structures and claims that you can get access to this structure by using class methods, but this isn't true, or is it? The only way changing sheduling policy would be by calling a Mach function with a cthread as argument. But the cthread is hidden by NSThread. Thanks in advance for any advise. Bernhard. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: when an expression evaluates to nil (really long!) Date: 16 Apr 1996 13:07:35 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7rato5sw8.fsf@tom.es.ele.tue.nl> References: <jpanicoDppG9I.KAy@netcom.com> <4kmcld$q91@news.next.com> <4ku5b3$3db@cetus.ali.bc.ca> In-reply-to: allan@ali.bc.ca's message of 15 Apr 1996 18:42:43 GMT In article <4ku5b3$3db@cetus.ali.bc.ca> allan@ali.bc.ca (Allan Noordvyk) writes: This brings up another point that has bugged me about a lot of NSObject code examples I have seen. In his "improved" version, you will notice that Mark has switched to using autorelease rather than release (in addition to checking for nil - the original point of the thread). As you point out, it is totally unnecessary. We use the following macro for every retaining assignment to a variable: #define RETAIN(VAR,VAL) \ do \ { \ id _val; \ \ _val = [(id) (VAL) retain]; \ if (VAR) \ [(id) (VAR) release]; \ (VAR) = _val; \ } while (0) (This does not perform the check for the VAL already being the current value of VAR, which could be added if desirable.) Anyway, this macro is correct and fast. There's no reason to do an autorelease. There's no reason for any code to have an explicit retain anymore (any retain is considered an error, except in methods named `init...'); with the accompanying RELEASE (which sets the VAR to nil) explicit invocations of release are also no longer needed; bugs due to dangling reference counts because of a reference being overwritten before a release do no longer exist; bugs due to a reference being released twice no longer exist. And all this retain/[auto]release crap shows that reference counting is not good enough as you're basically still managing storage by hand. --Tiggr check the newsgroups. Thanks. -- Dino Bagdadi ex nihilo, inc. dino@ex-nihilo.com (ASCII, NeXTmail and MIME) Public PGP key available via `finger -l dbagdadi@hyper.shadow.net'
From: bjanzen@rainbow.rmii.com (Barry Janzen) Newsgroups: comp.sys.next.programmer Subject: Wiring down virtual memory? Date: 16 Apr 1996 17:14:06 GMT Organization: Rocky Mountain Internet Inc. Message-ID: <4l0kgu$nrn@natasha.rmii.com> I read some time back about support for user-level wiring of virtual memory. Was this ever implemented, and if so, what are the Mach calls to do this? I have a need to lock down large raster images for printing. Barry Janzen bjanzen@objectmfg.com
From: kander@cac.stratus.com Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 16 Apr 1996 19:51:21 GMT Organization: Stratus Computer Inc, Marlboro MA Distribution: world Message-ID: <4l0tnp$et1@transfer.stratus.com> References: <4kgj3b$lu2@netty.york.ac.uk> <1996Apr8.071349.18229@seer.demon.co.uk> In Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) comp.sys.next.programmer pete@ohm.york.ac.uk (-bat.) writes, > Paul Lynch <Paul_Lynch@plsys.co.uk> writes: >> scripts won't be enough to convert your code; about one man month per >> 50,000 lines of code will be required, he said. This is a lot of effort >> (particularly if it takes longer than that :-). > > Ouch. I hadn't seen that before. Preseumably this is just for the frontend > code as well. Those of us who have a significant amount of UNIXy code > underneath all the Objective C will have an even longer job I guess :-( > >> to be extensive. With OpenStep, I expect that will no longer be the case, >> and testing requirements for OpenStep/NT, OpenStep/Solaris and NeXTSTEP >> will be much more gruelling. > > I really really hope not - it would make things a lot harder and maybe make > me think twice about actually moving to OpenStep anyway. Mind you, the > single source tree hasn't always worked fine in the past either. I never > managed to get an App which runs fine on intel and 68000 to work under HP > no matter what I did. > > Sigh... > > -bat. I'm surprised by your comment regarding running well under Intel & 68K, but not HP. I moved a system with hundreds of thousands of lines of code to the HP, with ZERO source changes. Porting to Sparc was a different matter entirely - although the HP prefers aligned data, the Sparc requires it. There were a number of places where we were directly manipulating memory, and the sparc wouldn't dereference it. Ken Anderson anderson@interramp.com Advanced Information Solutions, Inc. Piper Turbo Arrow III
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Odd Crasher.. perhaps a compiler problem? Date: 16 Apr 1996 21:23:01 GMT Organization: Digital Fix Development Message-ID: <4l133l$8he@digifix.digifix.com> I've got this odd problem... If that conditional before the NX_DURING is false, then things work just fine If however that conditional before the NX_DURING is true and I remove the NXLogError line, then the app dies! It shouldn't be reaching that NXLogError anyways, but if it isn't there... BOOM I'm communicating with this bundle using DO if that makes a difference.. if (![self _imageFormatAvailable:filetype]) { NXLogError("We can't convert %s.\n",filetype); // If we can't do the conversion, then lets get out now return nil; } NXLogError("Hello Everybody!\nHi Doctor Nick!"); NX_DURING { // Create a pasteboard object (using [Pasteboard newUnique]). originalFilePb = [Pasteboard newUnique]; a=[originalFilePb declareTypes:typeList num:1 owner:nil]; -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Why does Foundation Kit introduce a "new" hierarchy? (was Re: Betatron fool NeXT) Date: 16 Apr 1996 22:04:30 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <4l15he$105@shelob.afs.com> References: <4l0tnp$et1@transfer.stratus.com> kander@cac.stratus.com writes > Porting to Sparc was a different matter entirely - although the HP > prefers aligned data, the Sparc requires it. There were a number of > places where we were directly manipulating memory, and the sparc > wouldn't dereference it. Under emergency conditions, it is possible to make a SPARC do what you want by booting with 'fat=1'. However, my understanding is that this slows down all memory referencing, since it is difficult/impossible to predetermine which accesses may be misaligned. The fat flag fixes the problem, but you'd want to patch your code as a long-term solution that does not sacrifice speed. -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman NSData dataWithBytesNoCopy:buffer length:length]; > [myArray addObject:data]; > > a seperate thread would dealloc the data in the different NSData objects > of the array and release the NSData objects (after removeing form array). > compare this to: > ... > data = [NSData dataWithBytes:buffer length:length]; > [myArray addObject:data]; > vm_deallocate(task,buffer,length); > > a seperate thread would again dealloc the data in the NSData objects and > release the NSData objects (after remove from the array). > > [...] > The first runs ok, but gives Malloc error (code=5) on the console. The > second runs without any problems, causing no memory leaks. Adding [data > autorelease] to any code would cause memory leaks, because the thread > running the code will only exit on programm exit and the producer thread > in my case _doesn't_ release NSData objects. NSData isn't doing what you expect with the NoCopy: method, but it is correctly doing what it was, for whatever reason, designed to do: an NSData ALWAYS takes ownership of the bytes you give it. When the NSData is deallocated, it free()s the bytes. The NoCopy: just changes what happens at initialization; either the data creates its own copy of the bytes ("copy"), or it uses yours ("no-copy"). The always-taking-ownership has the interesting additional effect that you may only give (in the NoCopy: case) a chunk of memory to NSData which it can free with free() (or NSZoneFree()). You cannot give a vm_allocate'd chunk of memory to the NSData, because when it is deallocated it will attempt to free() it, generating the Malloc error (code=5) message you've seen. There is no public way to enable free() to work on a vm_allocate'd block of memory, nor to tell an NSData to use vm_deallocate instead of free(). If you use the "copy" flavor (dataWithBytes:length:), your bytes will be copied, but on architectures that support it, it will be copied-on-write (say, with vm_copy() on MACH). > BTW: how is the correct mechanism of handling autorelease objects in a > multi-threaded environment? There shouldn't be much different in a multi-threaded environment, as long as you are retaining and releasing as you should (even in the absense of multiple threads). Each NSThread has its own stack of autorelease pools -- an object autoreleased in one thread of execution will get released by whatever pool was topmost whenever that thread of execution releases that pool. What other threads do with autorelease pools has no effect. Christopher Kane Foundation Technology
From: Tim Triemstra <empath@a.crl.com> Newsgroups: comp.sys.next.programmer Subject: "conversion to non-scalar type requested" Date: Fri, 19 Apr 1996 15:30:50 -0400 Organization: CRL Dialup Internet Access Message-ID: <3177E9EA.6E2D@a.crl.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit When converting a project from NS Developer 3.2 to 3.3 I've encountered a number of lines of code that will no longer compile. Most of it can be attributed to ANSI C++ spec changes which are fairly easy to track down, and easy to understand the rationalle behind. However, I have encountered a new error that wasn't even a warning before and I can't track down how to make it work. The error is the subject line, "conversion to non-scalar type requested" - I took this to mean it was a pointer conversion problem of some sort. Essentially what the line of code that is croaking is doing is passing a pointer to an method of a class. The method is typed like this: typedef int (PtCallback::*PtFuncCallback)(ParamType *) Ie: the function is a method of a class derived from "PtCallback" returning "int" with a parameter of "ParamType*" However, when I to pass: (PtFuncCallback)&PtCallbackDerived::MyNewFunc I get the non-scalar conversion error? Any ideas? I'm still not possitive what the term non-scalar means (I used to think scalar refered to a "real" variable vs. a pointer.) This code used to work without error or warning. Thanks. -- Tim Triemstra Go Wings!!!! ---------------------------------------------- There are parts of me that don't get nervous Not the parts that shake... empath@a.cr.com
From: hhoff@cube.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: varibles declared outside of @implementation? Date: 19 Apr 1996 21:49:45 GMT Organization: Cube Informationssysteme GmbH Distribution: world Message-ID: <4l91pp$fj@cubenx.cube.de> References: <4l5rvg$194b@core.bard.edu> <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz wrote: > ... > To hide methods completely out of the public interface file, > I prefer the following method of programming: > > .m file: > > @interface MyClass (private) > > - private_method; > @end > > @implementation MyClass:... > ... > @end > > @implementation MyClass (private) > ... > @end > > Note that this only hides private methods by using the category technique > on MyClass. > > To hide instance variables out of the public interface I currently don't > know a way; maybe putting them inside @implementation @end will give this > effect, but I haven't tried it yet. > > The ObjC grammar didn't help me with this issue, but writing a test > object should be easy. I think I tried this once, and it didn't work - you can optionally list all your ivars in the @implementation{..} section, but you cannot define new or different ones than in the @interface. Makes the whole point of private ivars moot - but then again, you're supposed to use accessors anyway. This won't stop classdump, though. =:) Holger -- ___ /\__\ Holger Hoffstätte net: hhoff@cube.de (MIME, NeXT, PGP) \/__/ Cube Informationssysteme GmbH vox: (+49)0711-90669-0 (or -11) [kju:b] Stuttgart, Germany fax: (+49)0711-90669-16
Newsgroups: comp.sys.next.programmer Subject: Making a Window always on top. Message-ID: <Dq4IzG.33u@udcf.gla.ac.uk> From: dmurray@music.glas.ac.uk (Duncan Murray(EMU92)) Date: Fri, 19 Apr 1996 19:18:50 GMT Sender: news@udcf.gla.ac.uk (News) Organization: Glasgow University Computing Service Hi there, anyone know how to make a window always on top a la NXAlert panel? Is is possible to do this? Im not sure if I've seen it in a NeXT app (except the aformentioned NXAlertPanel & Dock). Any info greatly appriciated. Cheers, Duncan. dmurray@music.gla.ac.uk
Newsgroups: comp.sys.next.programmer From: al@atd.rochester.ny.us (Al Davis) Subject: Re: C compiler for NeXT? Message-ID: <1996Apr19.223848.4511@atd.rochester.ny.us> References: <4l2sue$208@transfer.stratus.com> <Dq118u.J2z@novice.uwaterloo.ca> <4l4gd6$cph@bertrand.ccs.carleton.ca> <4l4v6i$3go@news.rwth-aachen.de> Date: Fri, 19 Apr 1996 22:38:48 GMT Stefan Leuker (sl@skynet.oph.rwth-aachen.de) wrote: : You still wouldn't have a linker that produces machO output. What about "atom"? (do "man atom" for more info.) I have converted Sun-3 binaries with it. --
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Making a Window always on top. Date: Sat, 20 Apr 1996 11:08:23 +0100 Organization: Cedar Systems Distribution: world Message-ID: <QY1ebBAXeLexEw+G@cedar.co.uk> References: <Dq4IzG.33u@udcf.gla.ac.uk> MIME-Version: 1.0 In article <Dq4IzG.33u@udcf.gla.ac.uk>, "Duncan Murray(EMU92)" <dmurray@music.glas.ac.uk> writes >Hi there, anyone know how to make a window always on top a la NXAlert >panel? Is is possible to do this? Im not sure if I've seen it in a NeXT >app (except the aformentioned NXAlertPanel & Dock). > >Any info greatly appriciated. > If you use a Panel rather than a Window, invoking [thePanel setFloatingPanel:YES]; will do this. Paul ------------------------------------------------------------------ Paul Heffernan Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Better timing than 10ms possible???? Date: 19 Apr 1996 01:02:28 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4l6on4$1bn@news.its.com> References: <4ktsca$n3f@brachio.zrz.TU-Berlin.DE> dirk@kalium.physik.TU-Berlin.DE (Dirk Schwarzhans) wrote: > I want to execute parts of my program in regular intervals of less > than 10ms. On NSfIP systems the timing seems to be not better than > 10ms (starting with release 3.1!). [ ... ] > Any idea how to get more accurate timing? You can't depend on the system's scheduler to switch to your process at that small an interval-- the minimum quantum (time slice per process) is somewhere around 5 - 20 ms on many kernels. You're going to have to busy-wait between intervals by creating a calibrated do-nothing loop: record the time, iterate the loop for a couple of thousand times, determine the actual time interval, and from that and the number of loops you did, you can determine how many iterations would correspond to a 10 ms interval. Of course, nothing prevents the scheduler from taking control away from your process. You're probably better off using a real-time operating system if you really need to perform activities reliably at such a small interval. What are you trying to do, anyway? -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Exceptions (was: Re: when an expression evaluates to nil) Date: 19 Apr 1996 01:12:35 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4l6pa3$1bn@news.its.com> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <x7pw9brr26.fsf@tom.es.ele.tue.nl> <nnhgumxx71.fsf@lysistrate.lysator.liu.se> <x7g2a5yewb.fsf_-_@tom.es.ele.tue.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit tiggr@es.ele.tue.nl (Pieter Schoenmakers) wrote: > In article <nnhgumxx71.fsf@lysistrate.lysator.liu.se> > nisse@lysator.liu.se (Niels Möller) writes: > > What's bad with exceptions? Do you dislike the concept in general, or > just the NeXT style of exceptions? > > I like the functionality of and possibilities offered by exceptions; in > some circumstances they're invaluable and to be preferred very much over > setjmp and longjmp. That's ironic-- NXExceptions and NSExceptions are implemented as an OO wrapper over setjmp and longjmp. [ ... ] > For instance, an NSArray raising an invalid whatever exception because an > index is out of bounds is a nonsensical application of exceptions, as such > an error is an error and not some occasion which one might want to handle. I disagree. Having an exception raised for accessing an array with on out-of-bounds index is much cleaner than having to augment loop over a data structure with some kind of 'sentinel' value. Consult Jon Bentley's "Programming Pearls", Knuth's "Algorithms", or some AI applications (eg, computer chess programs) for examples of searching/sorting algorithms where this is useful. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: varibles declared outside of @implementation? Date: 20 Apr 1996 16:10:45 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4lb2a5$j3j@lion.embl-heidelberg.de> References: <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> In article <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: > > I was looking at MusicKit demo programs and I noticed that they have a > > strange way of declaring variables: > > > > ... > > #import "ExampAmp.m" > > > > #define DEFAULTFREQ 440 > > > > static Note *theNote, *theNoteUpdate; > > > > @implementation > > ... > Defines are legal anywhere. True > Concerning the static declared variables the intention was probably the > following: having instance variables which need _not_ to be declared in > the public interface file (.h). No. static vars are used (visible) only in the c/m file where they are declared and have absolutely nothing to do with Ivars. > The result is different. The compiler > treats these variables like objcet variables. They are not visible > outisde the class, but all instances share the same variables. It's like > putting these declarations in the @public part of your interface file. In other OO languages like C++ (hmm, is it really OO) the concept of class vars vs. object vars could be found. This is one of the things people are missing in ObjC. But if you want to stay OO, then you can write your own Class-Ivar system. Below you can find such a system. I wrote it for the SciTools base kit. It's possible to add it as a category to NSObject class (note: the source might be buggy because is being converted to the new version of the kit). /************************************************************************* *** Accessing class variables *************************************************************************/ static NSMutableDictionary *classVariables; + (NSMutableDictionary *)_getVariablesForClass:(Class)aClass /* * The method returns a private to class aClass mutable dictionary of * class variables. If aClass does not respond to +className * a NSInvalidArgumentException is raised. */ { NS_DURING if (![aClass respondsToSelector:@selector(className)]) { [NSException raise:NSInvalidArgumentException format:@"Class does not respond to selector + className"]; } NS_HANDLER NSLog( @">>>STObject$_getVariablesForClass:\n%@\n%@", [exception exceptionName], [exception exceptionReason]); [exception raise]; NS_ENDHANDLER return [classVariables objectForKey:[aClass className]]; } + (NSDictionary *)getVariablesForClass:(Class)aClass /* * Note: calls _getVariablesForClass (for exeptions) */ { NSDictionary *classDictionary = nil; classDictionary = [self _getVariablesForClass:aClass]; if (classDictionary) { id newDivtionary; newDivtionary = [classDictionary copyWithZone:[classDictionary zone]]; return [newDivtionary autorelease]; } return nil; } + (id <STObjectProtocol>)getVariableWithName:(NSString *)aName forClass:(Class)aClass /* * Note: calls _getVariablesForClass (for exeptions) */ { NSDictionary *classDictionary = nil; id anObject = nil; classDictionary = [self _getVariablesForClass:aClass]; if (classDictionary) { anObject = [classDictionary objectForKey:aName]; } return anObject; } + (void)setVariable:(id)anObject withName:(NSString *)aName forClass:(Class)aClass /* * Note: calls _getVariablesForClass (for exeptions) */ { NSMutableDictionary *classDictionary = nil; classDictionary = [self _getVariablesForClass:aClass]; if (!classDictionary) { classDictionary = [NSMutableDictionary dictionaryWithCapacity:1]; [classVariables setObject:classDictionary forKey:[aClass className]]; } [classDictionary setObject:anObject forKey:aName]; return; } [...] > To hide instance variables out of the public interface I currently don't > know a way; maybe putting them inside @implementation @end will give this > effect, but I haven't tried it yet. Hmmm. Use @private -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: jbettis@cse.unl.edu (Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Using NXImage to make tiffs smaller Date: 20 Apr 1996 17:08:40 GMT Organization: University of Nebraska--Lincoln Message-ID: <4lb5mp$3il@crcnis3.unl.edu> I am trying to make a thumbnail of a TIFF with this code: tmpImage = [image copy]; [tmpImage setScalable:YES]; [tmpImage getSize:&size]; if (size.height * size.width > 2304.0) { float change; change = 48.0 / sqrt(size.height * size.width); size.height = size.height * change; size.width = size.width * change; [tmpImage setSize:&size]; [[tmpImage bestRepresentation] setSize:&size]; } It gives a 48x48 tiff just like I want, but it stiff takes up the same abount of space on disk. But if I use setPixelsWide: and setPixelsHigh: it crops the image instead of scaling it, but it is smaller. What can I do? -- Jeremy Bettis -*- PGP Public key available -*- University of Nebraska INET: jbettis@cse.unl.edu "Those who stand in the middle of the UUCP: jeremy@tddi.inetnebr.com road are often hit by passing cars." <a href="http://cse.unl.edu/~jbettis/">Click Here</a>
From: pbrown@palms.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 18 Apr 1996 08:12:40 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4l4tho$c6d@agate.berkeley.edu> References: <4l3qnm$hcj@news.cyberlink.ch> In article <4l3qnm$hcj@news.cyberlink.ch>, Lionel Tinguely <lionel@cyberlab.ch> wrote: >I made a very little "C" program. Whern I compile it unde NS 3.3 it >is about 50K....! WhenI compile on DOSWindows compilers it get 3K !!! >Is there any way to get it smaller under NS 3.3 ? You can strip the program with the command "strip a.out" or whatever your file is called. Once you're totally happy with the program (i.e., think you're done debugging), strip should cut the size by quite a bit. Paul
From: csaldanh@mae.carleton.ca (Chris Saldanha) Newsgroups: comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: Does a NEXTSTEP aware version of EMACS exist? Followup-To: comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Date: 20 Apr 1996 18:25:38 GMT Organization: computerActive Inc Message-ID: <4lba72$igc@bertrand.ccs.carleton.ca> References: <4l66mc$38p@mozo.cc.purdue.edu> <3177EDF7.41C6@fa.disney.com> Tal Lancaster (tlan@fa.disney.com) wrote: : > Would anybody please tell me where may I find a NEXTSTEP aware : > version of EMACS for my NEXTSTEP v3.2 (INTEL)? I got used to using a version : Regular emacs can be found on your NeXT-CD. Actually, I prefer using Not in NS 3.2. Emacs was part of the Developer Tools package, and it was only the terminal version. Emacs for NeXTSTEP 4.1 can be found on the archives, which is an excellent port of Emacs 19.28 to the NeXTSTEP window system. --Chris Chris Saldanha, Software Analyst -------------------------------------- computerActive, Inc |"The telephone was not invented by | chris@computerActive.on.ca (NeXTMail) | Alexander Graham Unitel" -Bell Ad | http://www.mae.carleton.ca/~csaldanh --------------------------------------
From: sl@skynet.oph.rwth-aachen.de (Stefan Leuker) Newsgroups: comp.sys.next.programmer Subject: Re: C compiler for NeXT? Date: 18 Apr 1996 08:40:50 GMT Organization: RWTH -Aachen / Rechnerbetrieb Informatik Message-ID: <4l4v6i$3go@news.rwth-aachen.de> References: <4l2sue$208@transfer.stratus.com> <Dq118u.J2z@novice.uwaterloo.ca> <4l4gd6$cph@bertrand.ccs.carleton.ca> Copyright: This message is copyrighted (c) by Stefan Leuker csaldanh@mae.carleton.ca (Chris Saldanha) wrote: > Does GNU libc build properly under NeXTSTEP? If it did, a binary > distribution of that and gcc could make a basic C environment... You still wouldn't have a linker that produces machO output. Stefan ----------------------------------------------------------------- Mia: Is that a fact? Vince: No, it's just what I heard; just what I heard. (pf)
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <9604181103.AA02421@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Thu, 18 Apr 96 15:03:25 +0400 Subject: Re: WO does not recognize new components ? BUG ? Cc: comp-sys-next-programmer@antigone.com Hi, In article <Dq0638.Hn@shinto.nbg.sub.org>, Tomi wrote: > Since it is definitly there you start to restart the browser, change names, > variables, methods etc. pp.. until on the next day everything turns out to > work magically...just like the problem never existed. I guess you've rebooted your machine. After adding new component one gotta restart the Defaultapp from a shell. I guess WOApplication object records all application components (i.e wo files) when it instantated (NeXT guys please point out if I'm wrong) and since new components are unknown to running WOApplication object it generates "HTML template may not exist" error which is gone if one restart the application. > Now to me this definitly is a BUG since "wo's" are bundles to me which are > searched for when requested. > > I new checked with the WODocumentation and I could not find a pointer in the > most obivous sections which talk about the componentes. > IF it really is a feature then it should be documented in <bold> face > somewhere in the examples or the docu. You might want to join WebObject mailing list or at least browse the archive at URL: http://www.omnigroup.com/MailArchive/WebObjects/ This feature as well as many others is "documented" there. Regards. Aleksey. _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Exceptions (was: Re: when an expression evaluates to nil) Date: 20 Apr 1996 23:57:15 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7spdy1rus.fsf@tom.es.ele.tue.nl> References: <4kdond$25b@saturn.genoa.com> <4kedd0$1e7@news.next.com> <4ki3lr$6sv@saturn.genoa.com> <x7pw9brr26.fsf@tom.es.ele.tue.nl> <nnhgumxx71.fsf@lysistrate.lysator.liu.se> <x7g2a5yewb.fsf_-_@tom.es.ele.tue.nl> <4l6pa3$1bn@news.its.com> In-reply-to: chuck@its.com's message of 19 Apr 1996 01:12:35 GMT In article <4l6pa3$1bn@news.its.com> chuck@its.com (Chuck Swiger) writes: > I like the functionality of and possibilities offered by exceptions; in > some circumstances they're invaluable and to be preferred very much over > setjmp and longjmp. That's ironic-- NXExceptions and NSExceptions are implemented as an OO wrapper over setjmp and longjmp. I know now that I referred to exceptions, not NSExceptions (thanks to Rainer who explained Common Lisp conditions to me). > For instance, an NSArray raising an invalid whatever exception because an > index is out of bounds is a nonsensical application of exceptions, as such > an error is an error and not some occasion which one might want to handle. I disagree. I agree with your disagreement (thanks to Andrew who explained that a DO client should never crash the server and thus an invalid index should raise an exception (to be passed back to the client, which reminds me of the question where exceptions in a oneway void DO method invocation are returned to; returning them to the client certainly being the wrong thing to do), and not call abort). The example should have been the exception raised at an attempt to add nil to an NSArray. --Tiggr
From: zazula@running-start.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: NSBundle, trouble..... Date: 18 Apr 1996 16:10:43 GMT Organization: Running Start, Inc. Message-ID: <4l5pi3$9eb@nntp1.best.com> References: <4l34tb$bj1@srvr1.engin.umich.edu> In-Reply-To: <4l34tb$bj1@srvr1.engin.umich.edu> On 04/17/96, Sean Willson wrote: >I am having trouble getting a listing of all the bundles in an >applications directory. I am able to ask the workspace for the >application path name, but I can't figure out how to get a file >listing of the bundles. It is not in the NXWorkspaceProtocol >definition to ask the workspace, and I don't want to kludge a system >call. I am using Foundation, is there an easier way? Any help would be >great! > Hi - If you know all the names of the bundles you are looking for, you can use NSBundle itself to do this. You can create an NSBundle for any directory and then use it to get full path names. For example: id bundle = [[NSBundle alloc] initWithPath:@"/tmp"]; id path = [bundle pathForResource:@"console" ofType:@"log"]; should return @"/tmp/console.log". If you *don't* know the names of the bundles, you'll probably have to use those ugly system calls. Here's the code from RZBundle that does this: + (void)addBundlesInPath:(NSString *)path withExtension:(NSString *)extension /* * Creates RZBundle instances for all files *."extension" * in the folder "path". */ { DIR *thedir; struct direct *entry; char *ext; thedir = opendir([path cString]); if(thedir) { while ((entry = readdir(thedir))) { if (*entry->d_name != '.') { ext = strrchr(entry->d_name, '.'); if(ext) { if(!strncmp(++ext, [extension cString], [extension length])) { [[self alloc] initWithPath: [NSString stringWithFormat:@"%@/%s", path, entry->d_name]]; } } } } closedir(thedir); } } Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@running-start.com http://www.running-start.com http://www.running-start.com/zazula/index.html
Newsgroups: comp.sys.next.programmer From: beaver@il.us.swissbank.com (Jason Beaver) Subject: Re: Printing "invisible" Text Message-ID: <1996Apr18.164941.13773@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <4l2v3u$3dt@turbocat.snafu.de> Date: Thu, 18 Apr 1996 16:49:41 GMT In article <4l2v3u$3dt@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) writes: > Hi! > > I have a Text object in my app. I load a rtf into that text and modify it. > Then I save the modified rtf to /tmp/test.rtf. Works great. > Now I want to print the text (not the saved one). How do I do that? Preview > gives me 2 empty pages. The saved rtf can be loaded in Edit.app and printed > but that's not really comfortable. > > Here is my source: > > <source deleted> You need to put the Text object in a window before it will generate PostScript. jason __________________ Jason Beaver Swiss Bank Corp. Jason_Beaver@swissbank.com (NeXTMail/MIME preferred) #import <std/disclaimer.h>
Newsgroups: comp.sys.next.programmer From: beaver@il.us.swissbank.com (Jason Beaver) Subject: Re: when an expression evaluates to nil (really long!) Message-ID: <1996Apr18.165606.13893@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <4l4maa$6jr@news.next.com> Date: Thu, 18 Apr 1996 16:56:06 GMT In article <4l4maa$6jr@news.next.com> cph@next.com (Charles d'Harcourt) writes: > ... > > If you wanted to, you could see this as a problem of the "currentOoo" > method, which could be rewritten (but usually isn't, for performance > reasons): > > ... An easier solution is to replace the release with autorelease in the foo: method... i.e. - (void)foo:newOoo { [myOoo autorelease]; myOoo = [newOoo retain]; } This way if myOoo and newOoo are the same object, the only thing which happens is the the object is retained one additional time and later released one additional time. A slight waste of time, but I'd rather have smaller code... jason __________________ Jason Beaver Swiss Bank Corp. Jason_Beaver@swissbank.com (NeXTMail/MIME preferred) #import <std/disclaimer.h>
From: andrew@hydra.cfm.brown.edu (Andrew Jones) Newsgroups: comp.sys.next.programmer Subject: Absoft's Fx Debugger Date: 21 Apr 1996 01:06:35 GMT Organization: Brown University Center for Fluid Mechanics Distribution: world Message-ID: <4lc1mr$fir@cocoa.brown.edu> Do you know where I can get the grapical FORTRAN debugger by Absoft. If so please contact me. Thanks, Andrew andrew@cfm.brown.edu (401) 861-MATH
Newsgroups: comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc From: rdl@world.std.com (Robert La Ferla) Subject: Re: Does a NEXTSTEP aware version of EMACS exist? In-Reply-To: csaldanh@mae.carleton.ca's message of 20 Apr 1996 18:25:38 GMT Message-ID: <u41n346gjys.fsf@world.std.com> Sender: rdl@world.std.com Organization: The World Public Access UNIX, Brookline, MA References: <4l66mc$38p@mozo.cc.purdue.edu> <3177EDF7.41C6@fa.disney.com> <4lba72$igc@bertrand.ccs.carleton.ca> Date: Sun, 21 Apr 1996 06:38:35 GMT Speaking of emacs, GNU's latest version 19.30 totally rocks! Don't know if you can easily compile it quad-fat yet... Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Making a Window always on top. Date: Sun, 21 Apr 1996 08:57:28 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Apr21.085728.3923@seer.demon.co.uk> References: <Dq4IzG.33u@udcf.gla.ac.uk> In article <Dq4IzG.33u@udcf.gla.ac.uk> writes: > Hi there, anyone know how to make a window always on top a la NXAlert > panel? Is is possible to do this? Im not sure if I've seen it in a NeXT > app (except the aformentioned NXAlertPanel & Dock). Alert panels are on top because they are called using runModal. "Floating" windows work in another way. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: Using NXImage to make tiffs smaller Date: 21 Apr 1996 13:20:21 GMT Organization: Nanyang Technological University Message-ID: <4ldcml$usv@ntuix.ntu.ac.sg> References: <4lb5mp$3il@crcnis3.unl.edu> Jeremy Bettis (jbettis@cse.unl.edu) wrote: : tmpImage = [image copy]; : [tmpImage setScalable:YES]; : [tmpImage getSize:&size]; : if (size.height * size.width > 2304.0) { : float change; : change = 48.0 / sqrt(size.height * size.width); : size.height = size.height * change; : size.width = size.width * change; : [tmpImage setSize:&size]; : [[tmpImage bestRepresentation] setSize:&size]; : } : It gives a 48x48 tiff just like I want, but it stiff takes up the same abount : of space on disk. : But if I use setPixelsWide: and setPixelsHigh: it crops the image instead of : scaling it, but it is smaller. What happens, when you first use setSize to scale it and then setPixels... to crop it? Just an idea. Willem
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: Re: Canon NEXTSTEP Developer Support-Japan Date: 21 Apr 1996 13:26:28 GMT Organization: University of Sheffield, UK Message-ID: <4ldd24$e3f@bignews.shef.ac.uk> References: <316D875B.6E2C@rs6000.cmp.ilstu.edu> In-Reply-To: <316D875B.6E2C@rs6000.cmp.ilstu.edu> On 04/12/96, "Eric A. Dubiel" wrote: > Has anyone seen the Canon NEXTSTEP Developer Support web pages in Japan? > > http://www.cnds.canon.co.jp/Japanese_EUC/WelcomeJ.html > > I can't understand the attitude of American Canon VS. Japanese > Canon. I wish the two would truly work together and also fully > translate the web pages to English and provide object.station driver > downloads/updates! > Just as the pages was loading, I thought to myself, I wonder if this is Toru-san's work... ... and sure enough: WWW Server Maintained by Toru Sato I suspect that Toru-san is putting in bit more effort on this than his employers pay for. He's an excellent chap, and totally dedicated. If Canon ATO had had him in charge, I'm sure the Oregon shop would still be open, and would be making big profits. If you're listening, Toru, it was a pleasure to meet you (heavens, 18 months ago now), and I hope maybe I will again one day. Maybe you could come over for the NeXT Expo Europe?! ;-) Best wishes, mmalc. --
Newsgroups: comp.sys.next.programmer Subject: Text class & writing to an instance ? From: apl@kcbbs.gen.nz (Andrew Lindesay) Date: 21 Apr 96 13:18:28 GMT Message-ID: <17496111.47908.12620@kcbbs.gen.nz> Organization: Kappa Crucis Unix BBS, Auckland, New Zealand Hi, I'm very much enjoying writing under NeXTSTEP apart from this little area - the Text handling classes are *VERY* unintuative to the novice. I'm writing an application in which I want to occasionally write results out to a text field; simple problem and you'd think the solution was equally as simple! I was much surprised to find that the Text class has generally really obtuse text handling methods. OK so am I stupid nad there really is an easy way or is it tricky and can somebody please shown me a few lines of code so I can make headway? I'd appreciate any help you can give me on this. I have MiscKit installed (which is bloody marvellous) so if that helps....??? ---------------------------------------------------------------------- A N D R E W P. L I N D E S A Y My Home Page is at http://www.iconz.co.nz/~iapl 9 St. Benedicts Street,Newton,Auckland,New Zealand. Fx 3581797. eMail apl@kcbbs.gen.nz (NeXTmail/MIME OK)
From: howlandr@cadvision.com (Royce Howland) Newsgroups: comp.sys.next.programmer Subject: Re: when an expression evaluates to nil (really long!) Date: 21 Apr 1996 13:46:41 GMT Organization: CADVision Message-ID: <4lde81$195g@rex.cadvision.com> References: <4l4maa$6jr@news.next.com> <1996Apr18.165606.13893@il.us.swissbank.com> beaver@il.us.swissbank.com (Jason Beaver) wrote: > An easier solution is to replace the release with autorelease in the > foo: method... > > i.e. > > - (void)foo:newOoo > { > [myOoo autorelease]; > myOoo = [newOoo retain]; > } > > This way if myOoo and newOoo are the same object, the only thing which > happens is the the object is retained one additional time and later > released one additional time. A slight waste of time, but I'd rather > have smaller code... Curious trade-off. Personally speaking, I have only found, so far, one type of situation where autorelease is necessary. (There may be others, of course.) I'm going to unconditionally shoot down any other use of it I find at our site unless I can be convinced it's solidly justifiable. (Don't think I would accept smaller code as a justification. :~) Indiscriminate use of autorelease leads to memory-bloated apps suffering performance degradation due to autorelease pool processing, and memory management errors that are difficult to track down and fix. (This is experience, not opinion, based on over a year's experience with our own reference counting and object caching implementation here at PCP.) Explicit use of retain/release will always be faster, easier to debug and more maintainable, except in cases where autorelease is the only way to preserve encapsulation. The only place I've found this to be necessary so far is in situations where an object has to release something and also return it to its caller. There, use autorelease to avoid both forcing the caller to clean up the callee's internals, as well as the possibility of returning a freed object. I'd be curious to know if there are scenarios besides this one where people have found autorelease to be solidly justified. Let's move past "101 ways to code set methods" and see what else of interest is around. -- Royce Howland, Object Systems Group howlandr@cadvision.com Howland_Royce@pcp.ca (PanCanadian Petroleum) I speak, but not for OSG or PCP.
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: varibles declared outside of @implementation? Date: 21 Apr 1996 18:11:02 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4ldtnm$p0@emerald.oz.net> References: <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> <4lb2a5$j3j@lion.embl-heidelberg.de> tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: > In article > <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> > Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: > > To hide instance variables out of the public interface I currently don't > > know a way; maybe putting them inside @implementation @end will give this > > effect, but I haven't tried it yet. > > Hmmm. Use @private That doesn't eliminate the need to declare them in the interface file used during compilation. One solution is to distribute an interface file that's different from the one actually used for class compilation. NeXT seems to take this approach. Private ivars are merely omitted from the public interface file. -- Art Isbell NeXT & MIME Mail: art@cubicsol.com NeXT Registered Consultant Voice/Fax: +1 408 335 2515 Trego Systems Voice Mail: +1 408 335 1154 CaseServ: NEXTSTEP managed care US Mail: Felton, CA 95018-9442 contract and case management solutions
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: when an expression evaluates to nil (really long!) Date: 21 Apr 1996 21:47:46 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7zq85e4v1.fsf@tom.es.ele.tue.nl> References: <4l4maa$6jr@news.next.com> <1996Apr18.165606.13893@il.us.swissbank.com> <4lde81$195g@rex.cadvision.com> In-reply-to: howlandr@cadvision.com's message of 21 Apr 1996 13:46:41 GMT In article <4lde81$195g@rex.cadvision.com> howlandr@cadvision.com (Royce Howland) writes: I'd be curious to know if there are scenarios besides this one where people have found autorelease to be solidly justified. In methods which clean up an object because what it stands for has vanished, such as, for example, an NSPort's `-invalidate' when its mach port has ceased to exist, it can be necessary to have the object retain and autorelease itself to avoid it being deallocated while still in the invalidation method. Possible scenario: -(void) cleanup { [delegate notifyEverybodyAboutOurCleanup: self]; [delegate release]; delegate = nil; } In this situation, it is possible that everybody previously holding a retain to this object does a release, implying that accessing the DELEGATE in the second line is invalid. There are two possible ways to solve this: The first is to set a local variable to the delegate and release the delegate through the local variable (but this is not always possible in situations more difficult than this trivial example); the second solution is to do a `[[self retain] autorelease]' at the start of this method. --Tiggr
From: lutzray@9bit.qc.ca (Raymond Lutz) Newsgroups: comp.sys.next.programmer Subject: Darcy's view on NeXT classes undisputed? (was Why does FK introduce a new..) Date: 22 Apr 1996 03:52:30 GMT Organization: SPC Message-ID: <4levpu$lpu@wagner.spc.videotron.ca> Well, I'm not a programmer so I can't comment but his view seems backed with hard facts... I was a little deceived to not read even a small reply to his post... Is he right? samurai@random.hasc.ca (Darcy Brockbank) wrote: > [timings about Darcy's objects VS FK ones snipped] > >Anyway, since my NKList beats both List and NSMutableArray hands >down, and is really a replacement for NSMutableArray, I'd conclude >that: > >(i) The FoundationKit, and NeXT classes in general are fat and >slow due to poor implementation. >Since my NKQueue totally blows them all away, I'd conclude that: >(ii) NeXT chose non-optimal algorithms for implementing base classes. >Since the interfaces of the Foundation and OpenStep classes REALLY >SUCK, and retain/release cycles bite me in the rear end every day, >making my life miserable, and NeXT took things like copyWithZone: >out of the root class (which is another total ass biter, just >think about it a while) I'd also conclude that: >(iii) The designers of OpenStep and Foundation did nowhere near >as good a job as the original AppKit designers. > >Let's face it... OpenStep as an API sucks. It's all we have for >the next little bit, but the writing is finally absolutely on the >wall now. With Sun backing off OpenStep, and moving into Java, >NeXT's technology becomes less meaningful every day... [other thread replies deleted] >Anyway, this just goes to show you that no matter the shit that >someone shovels at a group of people, some people will eat it and >smile, some people will eat it and gag, and some people will refuse >to eat. In which group are you? 8^) RL, lutzray@9bit.qc.ca
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 22 Apr 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4lf14e$mva@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: jungr@uni-muenster.de (Reinhard Jung) Newsgroups: comp.sys.acorn.programmer,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.unix.sco.programmer,comp.sys.atari.programmer,comp.sys.next.programmer,z-netz.rechner.apple.programmieren Subject: Questionnaire about Distributed Software Development (YAQ :) Date: Mon, 22 Apr 1996 06:40:02 GMT Organization: Institut fuer Wirtschaftsinformatik, WWU Muenster Message-ID: <4lf9fo$fd2@majestix.uni-muenster.de> Dear Internet participant, IMPORTANT NOTE: Although this subject might seem to be out of place in this newsgroup, I don't want to annoy its members! I'm interested in your opinion concerning worldwide-distributed software development and in your readiness to take part in such projects. The questionnaire is designed to investigate the opportunities only. It is part of a research project at the University of Muenster (Germany). There are no real projects or commercial goals! In order to save bandwidth, this post only contains the announcement of the questionnaire. Interested persons are kindly asked to visit our WWW form at the following address. If you use a WWW browser for reading netnews you can simply click here: <a href="http://www.uni-muenster.de/WiWi/quest.html">Questionnaire</a> <p> If not, please use the following address to locate the form: http://www.uni-muenster.de/WiWi/quest.html Greetings from Germany, Reinhard Jung, University of Muenster, Institute of Business Informatics, Grevener Strasse 91, 48159 Muenster (GERMANY) Tel. +49 251 927551, Fax +49 251 839754, email: jungr@uni-muenster.de
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Text class & writing to an instance ? Date: Mon, 22 Apr 1996 09:11:27 +0100 Organization: Cedar Systems Distribution: world Message-ID: <SUaW1CAv8zexEw$9@cedar.co.uk> References: <17496111.47908.12620@kcbbs.gen.nz> MIME-Version: 1.0 In article <17496111.47908.12620@kcbbs.gen.nz>, Andrew Lindesay <apl@kcbbs.gen.nz> writes >Hi, > >I'm very much enjoying writing under NeXTSTEP apart from this little >area - the Text handling classes are *VERY* unintuative to the novice. > >I'm writing an application in which I want to occasionally write results >out to a text field; simple problem and you'd think the solution was >equally as simple! I was much surprised to find that the Text class has >generally really obtuse text handling methods. > >OK so am I stupid nad there really is an easy way or is it tricky and >can somebody please shown me a few lines of code so I can make headway? >I'd appreciate any help you can give me on this. I have MiscKit >installed (which is bloody marvellous) so if that helps....??? > A common problem with the Text class is correctly instantiating a Text instance, this can be addressed by looking at /NextLibrary/Documentation/NextDev/Concepts/Pre3.0_Concepts/09_UIObjects .rtfd or by dragging a Scrolling Text instance from the palette in IB into a Window and loading that into your app. Once you have a correctly configured Text instance, to write results to it should be straightforward; the following method will append a string to a Text instance - it assumes that scrollText is an instance variable that has been connected to the ScrollView part of a Scrolling Text instance in IB. - (void)appendString:(NSString *)aString { Text *theText = [scrollText docView]; int endOfText = [theText textLength]; [theText setSel:endOfText :endOfText]; [theText replaceSel:[aString cString]]; [theText scrollSelToVisible]; } Hope this helps get started with Text. Paul. ------------------------------------------------------------------ Paul Heffernan Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: jq@papoose.quick.com (James E. Quick) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 22 Apr 1996 07:24:56 -0400 Organization: PHCS Message-ID: <4lfqa8$4i0@papoose.quick.com> References: <4l3qnm$hcj@news.cyberlink.ch> <4l4tho$c6d@agate.berkeley.edu> <4l7l36$cf@ironhorse.plsys.co.uk> In article <4l7l36$cf@ironhorse.plsys.co.uk>, Paul Lynch <paul@plsys.co.uk> wrote: >In <4l4tho$c6d@agate.berkeley.edu> Paul Robert Brown wrote: >> In article <4l3qnm$hcj@news.cyberlink.ch>, >> Lionel Tinguely <lionel@cyberlab.ch> wrote: >> >I made a very little "C" program. Whern I compile it unde NS 3.3 it >> >is about 50K....! WhenI compile on DOSWindows compilers it get 3K !!! >> >Is there any way to get it smaller under NS 3.3 ? >> >> You can strip the program with the command "strip a.out" or whatever your >> file is called. Once you're totally happy with the program (i.e., think >> you're done debugging), strip should cut the size by quite a bit. > >I think the minimum size is about 16K under NeXTSTEP. For very small programs you can trade a bit of time for a small footprint by using the -object flag. This puts all of the sections in a single unnamed segment with no segment padding performed. The smallest executable I have created using this was 1300 bytes (after running strip). Note that 'extremely small' executables should be careful to use library functions only from shared libraries. This executable type is documented in the ld(1) man page. A fragment from the ld(1) man page says this about -object: . . . . . . . . . In this format all sections are placed in one unnamed segment with all protections (read, write, execute) allowed on that segment. This is intended for extremely small programs that would otherwise be large due to segment padding. . . . . -- ___ ___ | James E. Quick jq@quick.com / / / | Private HealthCare Systems NeXTMail O.K. \_/ (_\/ | Systems Integration Group (617) 895-3343 ) | - My other car has a mouse.
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: varibles declared outside of @implementation? Date: 22 Apr 1996 17:36:43 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4lgg3b$448@chinx4.thoughtport.net> References: <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> <4lb2a5$j3j@lion.embl-heidelberg.de> <4ldtnm$p0@emerald.oz.net> #Re: varibles declared outside of @implementation? #by Art Isbell, Sense Networking Seattle (www.oz.net) tuparev@EMBL-Heidelberg.DE (Georg Tuparev) wrote: > In article > <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> > Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: > > To hide instance variables out of the public interface I currently don't > > know a way; maybe putting them inside @implementation @end will give this > > effect, but I haven't tried it yet. > > Hmmm. Use @private # That doesn't eliminate the need to declare them in the # interface file used during compilation. # # One solution is to distribute an interface file that's # different from the one actually used for class compilation. # NeXT seems to take this approach. Private ivars are merely # omitted from the public interface file. Can't do that; it will lead to some HORRENDOUSLY DIFFICULT to debug problems. Specifically, if you simply omit variables from the header, the compiler will no longer know how to look the ivar up in an instance. Basically, an instance is just a ANSI C structure that happens to have methods associated with it [seriously, that's all it is]. For example, View: Object Responder View Literally, an instance of view's ivar structure looks like: view { { { ... object's ivars ... } ... responders ivars ... } ... view's ivars ... } Ie; if you 'print *view' in gdb, you see a long list of ivars that include all the ivars for view and its superclasses... So, when compiled, the compiler treats an access to an instance variable as if you said self->ivar. It must compute the offset from the beginning of the structure by using the declaration of *every single* ivar in self and super and super's super, etc... If you REMOVE an ivar from anywhere in the inheritance chain, the calculation will be wrong and your ivar will suddenly have values you don't expect. If you don't believe me, try this: 1) create a set of files like those at the end of this message. 2) compile and run [see the main() function]: 31-wellington % ./ivar ivar1: 0xCAFEBEEF ivar2: 0xFACECAFE 3) Comment out ivar1 in Garbage.h and touch MoreGarbage.m (but not Garbage.m) 4) compile and run: 32-wellington % ./ivar ivar1: 0xCAFEBEEF ivar2: 0xCAFEBEEF The compiler computed the offset for ivar2 with the assumption that ivar1 no longer existed... ie; the compiler bases the offset calculation PURELY on the information provided in the headers. If NeXT releases headers with ivars removed, it is only with those missing ivars replaced with ivars of the same physical size.... ie; they cannot simply be removed. BTW: This is the source of a common and very confusing bug... if you *ever* see ivars that have values that are *totally* unexpected, ask yourself whether or not the instance variables of any of the superclasses have changed and, if they have, whether or not *all* .m files that inherit information from that class have been *recompiled*. If not, recompile and all will likely be happy. b.bum --- Garbage.h: @interface Garbage : Object { unsigned ivar1; } - test; @end --- Garbage.m: #import <objc/Object.h> #import <stdio.h> #import "Garbage.h" @implementation Garbage - init { [super init]; ivar1 = 0xCAFEBEEF; return self; } - test { printf("ivar1:\t0x%X\n", ivar1); return self; } @end --- MoreGarbage.h: @interface MoreGarbage:Garbage { unsigned ivar2; } + test; - test; @end --- MoreGarbage.m: #import <objc/Object.h> #import <stdio.h> #import "Garbage.h" #import "MoreGarbage.h" @implementation MoreGarbage + test { [[[[MoreGarbage alloc] init] test] free]; return self; } - init { ivar2 = 0xfacecafe; // init BEFORE calling super [super init]; return self; } - test { [super test]; printf("ivar2:\t0x%X\n", ivar2); return self; } @end -- main.m: #import <stdio.h> #import <objc/Object.h> #import "Garbage.h" #import "MoreGarbage.h" void main() { [MoreGarbage test]; exit(0);}
From: kc@criamon (Ken Case) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 22 Apr 1996 17:22:57 GMT Organization: Omni Development, Inc. Message-ID: <4lgf9h$2ob@gaea.titan.org> References: <4l3qnm$hcj@news.cyberlink.ch> <4l4tho$c6d@agate.berkeley.edu> <4l7l36$cf@ironhorse.plsys.co.uk> Paul Lynch (paul@plsys.co.uk) wrote: : In <4l4tho$c6d@agate.berkeley.edu> Paul Robert Brown wrote: : > In article <4l3qnm$hcj@news.cyberlink.ch>, : > Lionel Tinguely <lionel@cyberlab.ch> wrote: : > >I made a very little "C" program. Whern I compile it unde NS 3.3 it : > >is about 50K....! WhenI compile on DOSWindows compilers it get 3K !!! : > >Is there any way to get it smaller under NS 3.3 ? : > : > You can strip the program with the command "strip a.out" or whatever your : > file is called. Once you're totally happy with the program (i.e., think : > you're done debugging), strip should cut the size by quite a bit. : I think the minimum size is about 16K under NeXTSTEP. You can get programs smaller by linking with the "-object" flag. Compiling "Hello world" (for my Intel) with the command cc -object -s hello.c -o hello results in a binary which is just over 1K. (For those who didn't know, the "-s" flag strips the debugging symbols immediately, just like running the strip command later would.) Ken -- Ken Case Omni Development, Inc.
From: charlie@why.az.stratus.com (Charles_Spitzer) Newsgroups: comp.sys.next.programmer Subject: Re: Making a Window always on top. Date: 22 Apr 1996 19:14:32 GMT Organization: Stratus Computer Inc, Marlboro MA Message-ID: <4lglqq$b14@transfer.stratus.com> References: <Dq4IzG.33u@udcf.gla.ac.uk> In article <Dq4IzG.33u@udcf.gla.ac.uk> writes: > Hi there, anyone know how to make a window always on top a la NXAlert > panel? Is is possible to do this? Im not sure if I've seen it in a NeXT > app (except the aformentioned NXAlertPanel & Dock). > > Any info greatly appriciated. > > Cheers, > Duncan. > dmurray@music.gla.ac.uk place the follwing function in a .psw and compile it. ---cut here--- defineps setWindowLevel(userobject windowNum; int level) % sets a windows tier to level level windowNum setwindowlevel endps ---cut here--- you can then call it with the window number and a level. see the doc about window numbers and max levels. -- Charles Spitzer charlie@az.stratus.com Customer Assistance Center Stratus Computer, Inc. Phoenix, AZ 85018
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <jim@ergotech.com> Message-ID: <199604231519.AA00245@ergotech.com> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Jim Redman <jim@ergotech.com> Date: Mon, 22 Apr 96 10:32:58 -0600 Subject: RE: "conversion to non-scalar type requested" Tim Triemstra <empath@a.crl.com> writes: >The method is typed like this: >typedef int (PtCallback::*PtFuncCallback)(ParamType *) >Ie: the function is a method of a class derived from >"PtCallback" returning "int" with a parameter of "ParamType*" That's a bug in gcc 2.5.8 (or 2.5.x). It's not in 2.4.x or 2.6.x or later. If you are just using C++ you might consider a later version of gcc but I don't know how those would work for Objc-C++. Jim
From: Žruss@physical27.chem.ufl.edu (russ bowers) Newsgroups: comp.sys.next.programmer Subject: making g77 on HPPA Date: 23 Apr 1996 03:18:34 GMT Organization: University of Florida Message-ID: <4lhi6a$eso@no-names.nerdc.ufl.edu> Has anyone succeeded in compiling g77 (gnu fortran) under NS 3.3 running on HPPA? If so, what ./configure command did you use? I have heard that its been done on NS3.3 Intel, but the HPPA-next combination doesn't seem to be supported. Thanks. Russ Bowers
From: Erik Doernenburg <erik@object-factory.com> Newsgroups: comp.sys.next.programmer Subject: Re: Darcy's view on NeXT classes undisputed? (was Why does FK introduce a new..) Date: 23 Apr 1996 08:04:35 GMT Organization: Object Factory GmbH (Germany) Message-ID: <4li2uj$610@isabella.object-factory.com> References: <4levpu$lpu@wagner.spc.videotron.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit You got me! Normally, I'm just too lazy to get involved into discussions like this but since nobody commented on Darcy's statements... The problem is that you can't really be right or wrong. There are no "hard facts." It all depends on how you *interpret* the results you get. For example, Darcy is right to say that his implementation of an array based on a queue "totally blew them all away." However, this is only true for the tests he did. Of course it is very tempting to generalise and assume that under all circumstances his NKQueue would blow them all away. But that is dangerous and, as I will explain, in this case plain wrong! As you might remember, it was part of the test to delete lots of elements directly from the beginning of the array. This is something that can be done very efficiently in a queue based algorithm and only awkwardly in the way NeXT implemented their arrays. And (oh wonder!) Darcy's array beats NeXT's. If you change the test to always delete the, say, 23rd element, Darcy's algorithm will run into the same problems and wont outperform NeXT by much. Of course I could tell you all this and then implement an algorithm that would allow for efficient delete operations of arbitrary elements and say "Hey folks, my array is much better than Darcy's." The whole point is that this doesn't proof what Darcy originally implied. It simply shows that it is possible to outperform NeXT's implementations in some special cases. Big deal! From a computer scientist's point of view it is not too surprising to find that it is possible to write special algoritms that solve special problems more efficiently than standard algorithms that do allright in most cases. To give you an example: The standard C library has a sort function built in. This is implemented using a quicksort algorithm which *usually* has good performance, but has a really bad worst-case that brings it close to bubblesort if the array is almost sorted before. Okay, now I write a little shellsort that has a different worst-case condition, compare it to the standard implementation using pre-sorted arrays and conclude: "The standard implementation sucks!" (right?) After using them extensively for the last year I can say that in my opinion the implementation of the FK classes is okay (Of course, there are some things which I would like to see changed and there's some general overhead.) Most importantly, though, I found that that software development is generally much faster with them. This gives me time to identify problematic areas in my programs and implement efficient algorithms for these. erik -- -------------------------------------------------------------------------- Erik Dörnenburg OBJECT FACTORY Gesellschaft für Informatik und Datenverarbeitung mbH Lohbachstraße 12, 58239 Schwerte, Germany erik@object-factory.com or erik@sultan.free.de (priv)
From: Erik Doernenburg <erik@object-factory.com> Newsgroups: comp.sys.next.programmer Subject: Re: NSBundle, trouble..... Date: 23 Apr 1996 08:27:06 GMT Organization: Object Factory GmbH (Germany) Message-ID: <4li48q$610@isabella.object-factory.com> References: <4l34tb$bj1@srvr1.engin.umich.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit premise@engin.umich.edu (Sean Willson) wrote: >I am having trouble getting a listing of all the bundles in an >applications directory. I am able to ask the workspace for the >application path name, but I can't figure out how to get a file >listing of the bundles. It is not in the NXWorkspaceProtocol >definition to ask the workspace, and I don't want to kludge a system >call. I am using Foundation, is there an easier way? Any help would be >great! There is an easy solution using the Foundation Kit: Use the method 'completePathIntoString:caseSensitive:matchesIntoArray:filterTypes:' in the NSString class. An elegant way is to add the following method in a category to NSArray: + (NSArray *)arrayWithFilesOfTypes:(NSArray *)types inPath:(NSString *)aPath; { NSString *firstName; NSArray *allNames = nil; unsigned int count; count = [aPath completePathIntoString:&firstName caseSensitive:YES matchesIntoArray:&allNames filterTypes:types]; return allNames; } good luck erik ---------------------------------------------------------------------------- Erik Dörnenburg OBJECT FACTORY Gesellschaft für Informatik und Datenverarbeitung mbH Lohbachstraße 12, 58239 Schwerte, Germany erik@object-factory.com or erik@sultan.free.de (priv)
Newsgroups: comp.sys.next.programmer From: gerald@kurt.in-berlin.de (Gerald Erdmann) Subject: Re: Printing "invisible" Text Message-ID: <1996Apr22.214217.2819@kurt.in-berlin.de> Sender: news@kurt.in-berlin.de References: <4l2v3u$3dt@turbocat.snafu.de> <1996Apr18.164941.13773@il.us.swissbank.com> Date: Mon, 22 Apr 1996 21:42:17 GMT In <1996Apr18.164941.13773@il.us.swissbank.com> Jason Beaver wrote: > In article <4l2v3u$3dt@turbocat.snafu.de> dave@turbocat.snafu.de > (David Wetzel) writes: > > Hi! > > > > I have a Text object in my app. I load a rtf into that text and > modify it. > > Then I save the modified rtf to /tmp/test.rtf. Works great. > > Now I want to print the text (not the saved one). How do I do that? > Preview > > gives me 2 empty pages. The saved rtf can be loaded in Edit.app and > printed > > but that's not really comfortable. > > > > Here is my source: > > > > <source deleted> > > You need to put the Text object in a window before it will generate > PostScript. OK, but is this a bug or a feature? Does someone has a good reason why a single text object can't generate PS?! Gerald -- --------------------------------------------------------------------- | GERALD ERDMANN | email: gerald @ kurt.in-berlin.de (NeXTmail welcome) | voice: +49 30 397 31 400 (Germany - Berlin) | crypt: pgp2 public key available |
Newsgroups: comp.sys.next.programmer From: js@euler.han.de (Juergen Sell) Subject: Re: Where are "tcgetattr" and "tcsetattr"? Message-ID: <DqB4FJ.zI@euler.han.de> Sender: js@euler.han.de (Juergen Sell) Organization: Ink Unknown References: <4l444f$e7j@pollux.usc.edu> Date: Tue, 23 Apr 1996 08:47:43 GMT Sungdo Moon writes > I'm a user of NeXTSTEP 3.3 for Intel and also have Developer 3.3. > I compiled a program which uses "tcgetattr" and "tcsetattr". > Those are defined in <termios.h> and exists man pages for them. > But I got link errors. Does any body know what library need to be linked? > POSIX ? Juergen --- Fon +49 511 4406-88 NeXTMail welcome No Mime Fax +49 511 4406-17 = What time do we live in when revolution reminds us of soap powder, = when spontaneity and freedom get associated with instant coffee, = when a politician's idea of social change is changing names = when a country posing as super know-how factory cuts expenses on education?
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: varibles declared outside of @implementation? Date: 23 Apr 1996 15:12:16 GMT Organization: Sense Networking Seattle (www.oz.net) Distribution: world Message-ID: <4lis0g$8vb@emerald.oz.net> References: <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> <4lb2a5$j3j@lion.embl-heidelberg.de> <4ldtnm$p0@emerald.oz.net> <4lgg3b$448@chinx4.thoughtport.net> bbum@friday.com wrote: > #Re: varibles declared outside of @implementation? > #by Art Isbell, Sense Networking Seattle (www.oz.net) > # One solution is to distribute an interface file that's > # different from the one actually used for class compilation. > # NeXT seems to take this approach. Private ivars are merely > # omitted from the public interface file. > > Can't do that; it will lead to some HORRENDOUSLY DIFFICULT to debug > problems. Specifically, if you simply omit variables from the header, the > compiler will no longer know how to look the ivar up in an instance. > > [thoughtful comments omitted] Well, that's what can happen when one posts without thinking. Duh!! Bill is correct - ivars cannot merely be omitted from headers. Anyone who hasn't kept Makefile.dependencies updated after ivar changes has experienced what would happen. What NeXT appears to do is replace some ivars with void * ivars which are opaque pointers to unknown types of data, thus hiding some ivar info. But what probably caused my foolish post was NeXT's increasing use of class clusters with private subclasses with ivars and abstract superclasses with no ivars. The only public headers are for the abstract superclasses, so subclass ivar information is hidden. Hmmm. Wonder how I $%@&-up this posting ?-) -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: wolfgang@mime.amg.de (Wolfgang Fehmer) Newsgroups: comp.sys.next.programmer Subject: DO-Question Date: 23 Apr 1996 13:25:23 GMT Organization: AMG Industrieconsulting GmbH Message-ID: <4lilo3$ont@hagen.amg.de> Hi, I have the following situation: Client ---- ISDN Router ----- Server I form a connection between them like that: myServer = [NXConnection connectToName: "TESTNAME" onHost:"TESTHOST"] ; [[myServer connectionForProxy] registerForInvalidationNotification:self]; [[myServer connectionForProxy] runFromAppKit]; I try to end the connection like that: [myServer free] ; Its all right as long as I do this in a LAN. The Server gets a senderIsInvalid and free's the ports. But if there is a router inbetween, the senderIsInvalid message is lost and there is still traffic on the line. So how can I solve this problem ? Any hints are welcome. Wolfgang email: wolfgang@amg.de as a justification. :~) Indiscriminate use of > autorelease leads to memory-bloated apps suffering performance degradation > due to autorelease pool processing, and memory management errors that are > difficult to track down and fix. (This is experience, not opinion, based on > over a year's experience with our own reference counting and object caching > implementation here at PCP.) Does this imply that you avoid using all object allocation methods that return autoreleased instances (e.g., NSArray's array method)? If so, you've returned to the previous approach of requiring that temporary objects (e.g., those local to a single method) be manually released when no longer needed, a source of memory leaks when programmers forget to release an object later. Creation of an autoreleased object eliminates this common source of memory leaks. -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: csaldanh@mae.carleton.ca (Chris Saldanha) Newsgroups: comp.sys.next.programmer Subject: Re: Where are "tcgetattr" and "tcsetattr"? Date: 23 Apr 1996 15:13:25 GMT Organization: computerActive Inc Distribution: world Message-ID: <4lis2l$nbs@bertrand.ccs.carleton.ca> References: <4l444f$e7j@pollux.usc.edu> Sungdo Moon (sungdomo@pollux.usc.edu) wrote: : I'm a user of NeXTSTEP 3.3 for Intel and also have Developer 3.3. : I compiled a program which uses "tcgetattr" and "tcsetattr". : Those are defined in <termios.h> and exists man pages for them. : But I got link errors. Does any body know what library need to be linked? Link with -lposix (the POSIX library), and those symbols will be resolved. It's possible that the object files that compose your program may need to be compiled with -posix, the POSIX source flag. --Chris Chris Saldanha, Software Analyst -------------------------------------- computerActive, Inc |"The telephone was not invented by | chris@computerActive.on.ca (NeXTMail) | Alexander Graham Unitel" -Bell Ad | http://www.mae.carleton.ca/~csaldanh --------------------------------------
From: Erik Doernenburg <erik@object-factory.com> Newsgroups: comp.sys.next.programmer Subject: methodSignatureForSelector broken? Date: 23 Apr 1996 15:21:23 GMT Organization: Object Factory GmbH (Germany) Message-ID: <4lishk$cnt@isabella.object-factory.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Is it me or does the method 'methodSignatureForSelector' always produce a memory access exception on address 0x0 when used with the Foundation Kit that comes with EOF 1.1? erik ---------------------------------------------------------------------------- Erik Dörnenburg OBJECT FACTORY Gesellschaft für Informatik und Datenverarbeitung mbH Lohbachstraße 12, 58239 Schwerte, Germany erik@object-factory.com or erik@sultan.free.de (priv)
Newsgroups: comp.sys.next.programmer From: beaver@il.us.swissbank.com (Jason Beaver) Subject: Re: Making a Window always on top. Message-ID: <1996Apr23.160740.11729@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <4lglqq$b14@transfer.stratus.com> Date: Tue, 23 Apr 1996 16:07:40 GMT In article <4lglqq$b14@transfer.stratus.com> charlie@why.az.stratus.com (Charles_Spitzer) writes: > > place the follwing function in a .psw and compile it. > ---cut here--- > defineps setWindowLevel(userobject windowNum; int level) > > % sets a windows tier to level > level windowNum setwindowlevel > endps > ---cut here--- > you can then call it with the window number and a level. see the doc about > window numbers and max levels. > There's already a PostScript function to do this, PSsetwindowlevel. i.e. @implementation Window(LevelControl) - setWindowLevel:(int)level { PSsetwindowlevel(level,[self windowNum]); return self; } @end jason __________________ Jason Beaver Swiss Bank Corp. Jason_Beaver@swissbank.com (NeXTMail/MIME preferred) #import <std/disclaimer.h>
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 23 Apr 1996 21:23:59 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7g29uoib4.fsf_-_@tom.es.ele.tue.nl> In-reply-to: aisbell@cubicsol.com's message of 23 Apr 1996 15:12:16 GMT In article <4lis0g$8vb@emerald.oz.net> aisbell@cubicsol.com (Art Isbell) writes: But what probably caused my foolish post was NeXT's increasing use of class clusters with private subclasses with ivars and abstract superclasses with no ivars. The only public headers are for the abstract superclasses, so subclass ivar information is hidden. With the added disadvantage that NeXT's gdb actually uses the debugging information it can find on a struct Foo to be able to print the contents of an object of the Foo class. Since there is no debugging information, you won't be able to print the contents of instances of private classes. And I do not want to hear `they are private so you're not supposed to look at them', since I do want to be able to `print foo->delegate' if I know `foo' has a delegate. In this context, I do not care about privacy. --Tiggr
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <bruce@trwlasd.com> From: Bruce McKenzie <bruce@trwlasd.com> Message-ID: <199604232316.AA24241@trwlasd.com> Date: Tue, 23 Apr 96 16:16:06 -0700 Subject: Ever growing swapfile...Any suggestions? I'm working on an app that allocates and frees a lot of data. The swap file grows significantly during the app's lifespan. After the app is quit/terminated, the swapfile does not decrease. If the app is rerun, the swapfile starts growing again. I'm inferring that swapfile space is "lost" when a process exits. Has anyone solved this yet? NeXT acknowledges that it is a problem, and refuses to fix it themselves. Please email any responses; I'll summarize significant ones to the group. Thanks in advance... (Developing under NS 3.2) Bruce McKenzie bruce@trwlasd.com
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Darcy's view on NeXT classes undisputed? (was Why does FK introduce a new..) Date: 24 Apr 1996 00:42:43 GMT Organization: University of Ghent, Belgium Message-ID: <4ljte3$b7e@infoserv.rug.ac.be> References: <4levpu$lpu@wagner.spc.videotron.ca> <4li2uj$610@isabella.object-factory.com> Erik Doernenburg (erik@object-factory.com) wrote: : You got me! Normally, I'm just too lazy to get involved into discussions like : this but since nobody commented on Darcy's statements... I'd also greatly prefer to just continue with the "real" (?) development issues of my NextStep/OpenStep software instead of this Object-NSObject thing, but I think that in this case "this kind of discussion" was definitely not unsuccesful : there is now the confirmation of Matt Watson that Object will be there in all versions of OpenStep; the alternative, not speaking up, would have been worse. What does OpenStep win by excluding some "faithful" NextStep developers, and by labeling them as "anti-NeXT", simply because of a genuine concern about what's happening with all these Object-NSObject changes and by the rewriting in the large of things that were working perfectly fine before ? David Stes e-mail: stes@can.nl www: http://www.can.nl/~stes
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: DO-Question Date: 24 Apr 1996 08:13:44 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4lknro$qf0@lion.embl-heidelberg.de> References: <4lilo3$ont@hagen.amg.de> In article <4lilo3$ont@hagen.amg.de> wolfgang@mime.amg.de (Wolfgang Fehmer) writes: > Hi, > > I have the following situation: > > Client ---- ISDN Router ----- Server Hmmm. The ISDN Router should not make any difference! > I form a connection between them like that: > > > I try to end the connection like that: > > [myServer free] ; I think this is the problem. Your mach port is not closed. > Its all right as long as I do this in a LAN. The Server gets a > senderIsInvalid and free's the ports. > But if there is a router inbetween, the senderIsInvalid message is lost and > there is still traffic on the line. > So how can I solve this problem ? > This line will terminate connections with existing clients: port_deallocate(task_self(), [[serverConnection inPort] machPort]); In this case, clients will receive the senderIsInvalid: notification in response to the port deallocation. Also you have to check out the port name from NI database: @implementation NXConnection(unregister) +unregisterRootWithName: (const char *)name { return [NXNetNameServer checkOutPortWithName: name]; } @end .. and last suggestion: use NXAutoreleaseConnection instead of NSConnection. This will save you a lot of headache. I hope this will help ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: EOCustomvalues??? Date: 24 Apr 1996 09:10:03 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4lkr5b$gl0@ai.alienor.fr> I'm working with EOF, NexTStep 3.3 and Oracle 7. I had an accent problem in the communications between Oracle and NexT. So I made a custom data type (NSString's subclass) for the concerned attributes. The accent conversion works, but I had an other problem while doing a select and a fetch. I'm sure the problem come from the custom data type because if I change it back to NSString it works(without the accents). The message is : [273] *** Freeing freed object [0x47fc70] ignored; [NSAutoreleasePool enableDoubleReleaseCheck:YES] to debug! I think I need to implement some others methods in my custom class, but which ones? Or What? Thanks for help. -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: when an expression evaluates to nil (really long!) Date: 24 Apr 1996 10:49:43 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4ll107$38b@zdi.informatik.uni-stuttgart.de> References: <4l4maa$6jr@news.next.com> <1996Apr18.165606.13893@il.us.swissbank.com> beaver@il.us.swissbank.com (Jason Beaver) writes: >An easier solution is to replace the release with autorelease in the >foo: method... >i.e. >- (void)foo:newOoo >{ > [myOoo autorelease]; > myOoo = [newOoo retain]; >} And an even easier solution is: - (void)foo:newOoo { [newOoo retain]; [myOoo release]; myOoo = newOoo; } Using -autorelease is really not necessary here. Klaus Brouwer <NO SIG>
From: howlandr@cadvision.com (Royce Howland) Newsgroups: comp.sys.next.programmer Subject: Re: when an expression evaluates to nil (really long!) Date: 24 Apr 1996 12:45:51 GMT Organization: CADVision Message-ID: <4ll7q0$icm@rex.cadvision.com> References: <4l4maa$6jr@news.next.com> <1996Apr18.165606.13893@il.us.swissbank.com> <4lde81$195g@rex.cadvision.com> <4liseh$8vb@emerald.oz.net> aisbell@cubicsol.com (Art Isbell) wrote: > Does this imply that you avoid using all object allocation methods that > return autoreleased instances (e.g., NSArray's array method)? If so, you've > returned to the previous approach of requiring that temporary objects (e.g., > those local to a single method) be manually released when no longer needed, a > source of memory leaks when programmers forget to release an object later. > Creation of an autoreleased object eliminates this common source of memory > leaks. No, it doesn't imply that at all. Further down in my original post, I said the only valid use for autorelease that I had so far encountered was the scenario where object A has to release its reference to object B, but return B to the caller. Your example above is a specific case of this scenario. Encapsulation would dictate that the caller should not have to know if the returned object it received needs to be cleaned up or not; rather the caller should only have to know that if it retains its own reference, then it is responsible for releasing it later. PCP has had a "he who allocates, frees" policy long before we had a retain/release/autorelease mechanism that actually enables developers to more easily maintain strict encapsulation of memory management policies. (I personally would like the option [<-- note I said *option*] of GC, so I didn't have any memory management to encapsulate. :~) The huge number of memory leaks and memory-related crashes and data corruption bugs that I continually find, through testing and code inspections, take a substantial chunk of time that would be well worth the performance hit of GC to avoid, in most cases.) Now, many "factory" methods break the memory management encapsulation rule. Convention is that +alloc (and any others like -copyWithZone:, and all of the +newWith... type of methods) return something that I don't have to retain, but that I do have to release. This is an inconsistency that we're stuck with. Fortunately, most such methods are clearly recognizable as factory methods, whereas most methods that return autoreleased objects are also recognizable as *not* being factory methods... -- Royce Howland, Object Systems Group howlandr@cadvision.com Howland_Royce@pcp.ca (PanCanadian Petroleum) I speak, but not for OSG or PCP.
From: walter@biw-ag.de (Sabine Walter) Newsgroups: biz.jobs,biz.jobs.offered,biz.next,comp.lang.objective-c,comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,ct.jobs,de.markt.jobs,de.markt.jobs.d,misc.jobs.misc,misc.jobs.offered,misc.jobs.offered.entry,stgt.next,stgt.uni-s.general Subject: Anwendungsentwickler(innen) NEXTSTEP gesucht Date: 24 Apr 1996 13:16:14 GMT Organization: BIW Systemhaus Distribution: inet Message-ID: <4ll9iu$1id@next01.biw-ag.de> Als Systemhaus setzen wir seit öber 20 Jahren MaûstÙbe bei der Beratung, Entwicklung und Einföhrung modernster Business- und PPS-Software. Mit öber 280 Mitarbeitern und Niederlassungen in der Schweiz, –sterreich, Portugal und Ruûland haben wir eine föhrende Stellung am Markt und verzeichnen weiterhin ein starkes Wachstum. Unser Erfolg beruht im wesentlichen auf dem Engagement und der Leistung unserer Mitarbeiter. Dies wissen wir entsprechend zu honorieren. Unsere positive Unternehmensentwicklung bietet auch Ihnen ausgezeichnete Perspektiven. Wir suchen mehrere Anwendungsentwickler(innen) för den Bereich NEXTSTEP/OpenStep Voraussetzungen sind: Kenntnisse der objektorientierten Softwareentwicklung, C-Programmiererfahrung, Studium mit starker Softwareorientierung und Freude an der Arbeit im Team Von Vorteil wÙren: Erfahrungen in NEXTSTEP/OpenStep und Objective C, Datenbanken, Unix etc. Bewerbungen bitte an: BIW GmbH z.H. Herrn Faûbinder Werkstr. 24 71384 Weinstadt Infos unter: e-mail: walter@biw-ag.de http://www.biw-ag.de -- Sabine Walter BIW Beratung und Informationssysteme GmbH
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: varibles declared outside of @implementation? Date: 20 Apr 1996 18:15:13 GMT Organization: Genoa Software Systems Distribution: world Message-ID: <4lb9jh$dv@saturn.genoa.com> References: <4l5rvg$194b@core.bard.edu> <Pine.HPP.3.91.960419091903.24783B-100000@hphalle0.informatik.tu-muenchen.de> <4l91pp$fj@cubenx.cube.de> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In <4l91pp$fj@cubenx.cube.de> Holger Hoffstaette wrote: > I think I tried this once, and it didn't work - you can optionally list > all your ivars in the @implementation{..} section, but you cannot define > new or different ones than in the @interface. Makes the whole point > of private ivars moot - but then again, you're supposed to use accessors > anyway. This won't stop classdump, though. =:) just define your ivars after the @private keyword (ivars can be private, public or protected) @interface X:Y { @public ... // some ivars visible to all by direct access "->", no method call needed @protected ...// some ivars visible directly by the implementation, and subclasses @private ..// some ivars visible only by the class implementation, not even subclasses } Don't define "private" methods in the interface file, so people at least get a compiler warning if they use them, as Bernhard Scholz illustrated. Finally, you can declare static variables outside of a class just as in traditional C -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: dirk@kalium.physik.TU-Berlin.DE (Dirk Schwarzhans) Newsgroups: comp.sys.next.programmer Subject: How to register interrupt handlers without using DriverKit? Date: 24 Apr 1996 17:04:04 GMT Organization: Technical University Berlin, Germany Message-ID: <4llmu4$366@brachio.zrz.TU-Berlin.DE> Hello, I am writing a loadable kernel server of standard MACH style without using the DriverKit. How do I register a function to be called if a specific hardware interrupt occured. All refrences I found in the documentation refered to the "NeXTBUS". How to do it on Intel Hardware, if I want to receive interrupts from the printer port p.e.? Thanks, Dirk Schwarzhans
From: paul@plsys.co.uk (Paul Lynch) Newsgroups: comp.sys.next.programmer Subject: Re: Darcy's view on NeXT classes undisputed? (was Why does FK introduce a new..) Date: 24 Apr 1996 17:00:52 GMT Organization: P & L Systems, Ltd. Message-ID: <4llmo4$lbo@ironhorse.plsys.co.uk> References: <4levpu$lpu@wagner.spc.videotron.ca> <4li2uj$610@isabella.object-factory.com> <4ljte3$b7e@infoserv.rug.ac.be> In <4ljte3$b7e@infoserv.rug.ac.be> David Stes wrote: > I'd also greatly prefer to just continue with the "real" (?) development > issues of my NextStep/OpenStep software instead of this Object-NSObject > thing, but I think that in this case "this kind of discussion" was > definitely not unsuccesful : there is now the confirmation of Matt Watson > that Object will be there in all versions of OpenStep; the alternative, > not speaking up, would have been worse. No, he just said that it was in NeXT's current versions of OpenStep. Paul -- Paul Lynch (NeXTmail) paul@plsys.co.uk Tel: (01494)432422 P & L Systems Fax: (01494)432478 http://www.plsys.co.uk/~paul
From: premise@engin.umich.edu (Sean Willson) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 24 Apr 1996 19:12:27 GMT Organization: University of Michigan Engineering, Ann Arbor Message-ID: <4lluer$c2m@srvr1.engin.umich.edu> References: <4l3qnm$hcj@news.cyberlink.ch> <4l4tho$c6d@agate.berkeley.edu> <4l7l36$cf@ironhorse.plsys.co.uk> <4lgf9h$2ob@gaea.titan.org> In-Reply-To: <4lgf9h$2ob@gaea.titan.org> I ran the strip operator on a NeXTStep application I am writing and it shrunk it down from 2M to 100K, this was huge, BUT it failed to load bundles and more. I defined a protocol the bundles respond to, and it appears that strip removed it. Any way to selectively strip? NeXTApps? Sean Willson premise@engin.umich.edu -- _____________________________________________________________________ Sean M. Willson. University of Michigan +1 313 761 8003 premise@engin.umich.edu College of Engineering +1 313 669 0565 MIME and NeXTmail http://www-personal.engin.umich.edu/~premise/ 1800 Washtenaw Ann Arbor MI, 48104 USA
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Darcy's view on NeXT classes undisputed? (was Why does FK introduce a new..) Date: 24 Apr 1996 23:06:21 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4lmc5d$a18@usenet.rpi.edu> References: <4levpu$lpu@wagner.spc.videotron.ca> <4li2uj$610@isabella.object-factory.com> <4ljte3$b7e@infoserv.rug.ac.be> stes@cage.rug.ac.be (David Stes) wrote: > I'd also greatly prefer to just continue with the "real" (?) > development issues of my NextStep/OpenStep software instead of > this Object-NSObject thing, but I think that in this case "this > kind of discussion" was definitely not unsuccesful : there is > now the confirmation of Matt Watson that Object will be there in > all versions of OpenStep; the alternative, not speaking up, would > have been worse. Hmm. The only thing I saw was the proof that "OpenStep for WindowsNT" had all the object classes. That is not the same thing as "all versions of OpenStep". On the other hand, our usenet hub is as flakey as usual so it is very believable that I'm missing some articles. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: droege@informatik.uni-koblenz.de (Detlev Droege) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 25 Apr 1996 08:02:58 GMT Organization: University Koblenz / Germany Distribution: world Message-ID: <4lnbji$73n@newshost.uni-koblenz.de> References: <4lluer$c2m@srvr1.engin.umich.edu> In article <4lluer$c2m@srvr1.engin.umich.edu> premise@engin.umich.edu (Sean Willson) writes: > > I ran the strip operator on a NeXTStep application I am writing and it > shrunk it down from 2M to 100K, this was huge, BUT it failed to load > bundles and more. I defined a protocol the bundles respond to, and it > appears that strip removed it. Any way to selectively strip? NeXTApps? For applications loading bundles and for bundles themselves add the options "-x -u" to the call of strip, e.g. strip -x -u DynaLoad.app/DynaLoad strip -x -u Some.bundle/Some Detlev -- Detlev Droege, Uni Koblenz, FB Informatik, Rheinau 1, D-56075 Koblenz, Germany Tel:+49 261 9119-421,Fax:-497,NeXT/MIME/Email:droege@informatik.uni-koblenz.de C++ is the only current language making COBOL look good. --Bertrand Meyer
From: braun@athen.biw-ag.de (Pirmin Braun) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 25 Apr 1996 09:56:45 GMT Organization: BIW Systemhaus Distribution: world Message-ID: <4lni8t$4u5@next01.biw-ag.de> References: <4lluer$c2m@srvr1.engin.umich.edu> <4lnbji$73n@newshost.uni-koblenz.de> In <4lnbji$73n@newshost.uni-koblenz.de> Detlev Droege wrote: > For applications loading bundles and for bundles themselves > add the options "-x -u" to the call of strip, e.g. > > strip -x -u DynaLoad.app/DynaLoad > strip -x -u Some.bundle/Some > You may also try "strip -S ...."; It works on EOF Apps. -- Pirmin Braun; BIW GmbH; Werkstr. 24; 71384 Weinstadt; Germany Phone: 0049 7151 602 484; Fax: 0049 7151 602 341 #include <StandardDisclaimers.h>
From: gmecchia@cc.colorado.edu Newsgroups: comp.sys.next.programmer Subject: using +new versus +alloc, -init Date: 25 Apr 1996 13:50:10 GMT Organization: The Colorado College, Colorado Springs, CO Message-ID: <4lnvui$3gf@lace.colorado.edu> Hi, what were the reasons for NeXT to stop using +new and instead use +alloc, -init? Translating an example from Smalltalk to Objective-C, in the smalltalk implementation, -init of class B gets called twice resulting with age being equal to 30. Is the result the same in an Objective-C implementation? /* class A - subclass of Object */ + new { return [[super new] init]; } - init { [self age: 7]; return self; } - age: anInteger { age = age + anInteger; } /* class B - subclass of A */ + new { return [[super new] init]; } - init { [super init]; [self age: 8]; } ------------------------------ Because in Smalltalk code it's common to have the factory method #new implemented as ^super new initialize if a subclass overrides #new and implements it the same as above, then the message #initialize is sent twice. That is why in "style guidelines" among the smalltalk literature there is a rule such as "if you need to create and initialize an instance using the #new and #initialize methods, check if the superclass already implements the same methods. If it does, then do not override the #new method." Any comments on the above are much appreciated. - Alexis gmecchia@cc.colorado.edu
From: wam@jaguar. (Wayne Morrison) Newsgroups: comp.sys.next.programmer Subject: Re: Printing "invisible" Text Date: 25 Apr 1996 17:17:01 GMT Organization: BT Labs, Martlesham Heath, Ipswich, UK Message-ID: <4loc2d$7s3@pheidippides.axion.bt.co.uk> References: <4l2v3u$3dt@turbocat.snafu.de> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit > > I have a Text object in my app. I load a rtf into that text and > modify it. > > Then I save the modified rtf to /tmp/test.rtf. Works great. > > Now I want to print the text (not the saved one). How do I do that? > Preview > > gives me 2 empty pages. The saved rtf can be loaded in Edit.app and > printed > > but that's not really comfortable. > Source code to create window and text in code deleted. The only way I managed to get it to work was by creating a Window containing a scrolled text in IB (but never displaying it). Holding an outlet to Text, I then shove my text inside it and send printPSCode:. I too couldn't do this in code and I also believe it's something to do with the sizing of the views. My problem was that it duplicated lines at the end of every page. I'd be interested if anyone else knows anything. Snippet: //thePrintingView is my outlet to the Text inside the ScrollView in the NIB. NSMutableString *printText = nil; //Somewhere to store the stuff for printing. while (obj = [theIssuesLogEnumerator nextObject]) { [printText appendFormat:@"Name%@\n", [obj name]]; } [[NXApp printInfo] setVertCentered:NO]; //Cosmetic touches. [thePrintingView setText:[printText cString]]; //Set our hidden view with printText. return [thePrintingView printPSCode:sender]; //Eventually print the bloody thing.
From: premise@news-server.engin.umich.edu (sean michael willson) Newsgroups: comp.sys.next.programmer Subject: View duplication... Date: 25 Apr 1996 17:47:18 GMT Organization: University of Michigan Engineering, Ann Arbor Message-ID: <4lodr6$q5o@srvr1.engin.umich.edu> I have an application I am developing that has an inspector and a preference panel. In the preference panel I have a view that updates the preferences for say a Name. This state is changed in the database. In the inspector I have the same view, but it changes the state of the name only in the current session. I use the basic multivew from the sams NeXT programming book, but when I have a preference panel, and an inspector panel up, one of the views freezes, since I guess it is swapped out. Is there any way to copy a view, yet maintain the logical connections, or do I have to create a seperate preference and inspector view? I can also entertain some other type of view swapping if anyone has any suggestions. Should I reswap everytime the window becomes key, this seems like alot of overhead? Just for reference, I load these inspector and preference views out of bundles. Sean Willson _________________________________________________________________________ Sean M. Willson premise@engin.umich.edu Foundation Software Solutions MIME, ASCII, NeXTMail Object Oriented Application Development Specializing in NeXT/OpenStep Net Home Page http://www-personal.engin.umich.edu/~premise/foundation/ 1800 Washtenaw Ave. Ann Arbor Mi, 48104 (313) 669-0565 _________________________________________________________________________
From: randyj@lubra.sbs.ohio-state.edu (Randy Jackson) Newsgroups: comp.sys.next.programmer Subject: f2c question Date: 25 Apr 1996 20:47:34 GMT Organization: The Ohio State University Message-ID: <4lood6$24t@charm.magnus.acs.ohio-state.edu> What would be the simplest way for me to install f2c? Help appreciated. Regards, Randy -- Randy Jackson, Associate Professor ,_ o __o Geography, The Ohio State University / //\, _`\<,_ 1036 Derby Hall, 154 North Oval Mall \>> | (*)/ (*) Columbus OH 43210-1361 \\, FAX (614) 292 6213 randyj@lubra.sbs.ohio-state.edu
From: Eric_Noyau@next.com (Eric Noyau) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 25 Apr 1996 21:21:46 GMT Organization: NeXT Software, Inc. Message-ID: <4loqda$bpf@news.next.com> References: <x7g29uoib4.fsf_-_@tom.es.ele.tue.nl> In article <x7g29uoib4.fsf_-_@tom.es.ele.tue.nl> tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: > In article <4lis0g$8vb@emerald.oz.net> aisbell@cubicsol.com (Art Isbell) writes: > > But what probably caused my foolish post was NeXT's increasing use > of class clusters with private subclasses with ivars and abstract > superclasses with no ivars. The only public headers are for the > abstract superclasses, so subclass ivar information is hidden. > > With the added disadvantage that NeXT's gdb actually uses the debugging > information it can find on a struct Foo to be able to print the contents > of an object of the Foo class. Since there is no debugging information, > you won't be able to print the contents of instances of private classes. > And I do not want to hear `they are private so you're not supposed to > look at them', since I do want to be able to `print foo->delegate' if I > know `foo' has a delegate. In this context, I do not care about > privacy. > --Tiggr What's wrong with 'print [foo delegate]' ? -- Eric
From: "James R. Pooton" <james@digisys.net> Newsgroups: comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.programmer Subject: POPPASSD compilation Date: Thu, 25 Apr 1996 16:25:43 -0600 Organization: DigiSys Incorporated Message-ID: <317FFBE7.6980@digisys.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, If anyone has gotten POPPASSD to compile under NS could you please drop me a line. If you'd like to give it a shot I can send you the source (only one source file). It's a password changing daemon for use with POP clients like Eudora. -James Pooton james@digisys.net
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Setting TextField's Backround to Transparent Date: 26 Apr 1996 03:46:55 GMT Organization: No organisation supplied Message-ID: <4lpgvf$99u@hermes.is.co.za> Hi, I have a TextField in one of my Apps. It's background needs to be in one of 2 states either transparent or it must be Red. The problem I am experiencing is that when I have changed the background color to red I cannot make it transparent again. My code looks something like this: - makeRed { [aTextField setBordered:YES]; [aTextField setBackgroundColor:NX_COLORRED]; return self; } - makeTransparent { [aTextField setBordered:NO]; [aTextField setBackgroundTransparent:YES]; return self; } I call makeRed to set the field red, and then makeTransparent to make it transparent. What happens is that the border does disappear but the background is still red. Any help in this regard would be appreciated. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Drawing lines in color? Date: 25 Apr 1996 23:43:59 GMT Organization: Turbocat's Development, Germany Message-ID: <4lp2nv$2ob@turbocat.snafu.de> Hi! How do I draw fast lines in color? DPSDoUserPath works only with one color, I think. I use this now: defineps PSWdrawLine (float x; float y;float xx; float yy; float r; float g; float b; float width) x y moveto xx yy lineto width setlinewidth r g b setrgbcolor stroke endps Thanks in advance. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: Optimize ??? Message-ID: <DqEv43.EH@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4lluer$c2m@srvr1.engin.umich.edu> Date: Thu, 25 Apr 1996 09:16:50 GMT Sean, In article <4lluer$c2m@srvr1.engin.umich.edu> premise@engin.umich.edu (Sean Willson) writes: > > I ran the strip operator on a NeXTStep application I am writing and it > shrunk it down from 2M to 100K, this was huge, BUT it failed to load > bundles and more. I defined a protocol the bundles respond to, and it > appears that strip removed it. Any way to selectively strip? NeXTApps? If you use bundles or you use any database adaptor (which is bundle too) you must use 'strip -x' to strip your executable. Bye, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 26 Apr 1996 11:37:14 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7afzzmilx.fsf@tom.es.ele.tue.nl> References: <x7g29uoib4.fsf_-_@tom.es.ele.tue.nl> <4loqda$bpf@news.next.com> In-reply-to: Eric_Noyau@next.com's message of 25 Apr 1996 21:21:46 GMT In article <4loqda$bpf@news.next.com> Eric_Noyau@next.com (Eric Noyau) writes: What's wrong with 'print [foo delegate]' ? The `foo' might not respond to `delegate'. --Tiggr
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: Setting TextField's Backround to Transparent Date: 26 Apr 1996 10:30:59 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4lq8l3$bkm@zdi.informatik.uni-stuttgart.de> References: <4lpgvf$99u@hermes.is.co.za> In <4lpgvf$99u@hermes.is.co.za> gvandyk@icon.co.za writes: >Hi, >I have a TextField in one of my Apps. It's background needs to be in >one of 2 states either transparent or it must be Red. >The problem I am experiencing is that when I have changed the >background color to red I cannot make it transparent again. [code deleted] >I call makeRed to set the field red, and then makeTransparent to make >it transparent. What happens is that the border does disappear but the >background is still red. Transparent display means "drawing nothing", i.e. any time you display your TextField, you also have to display the view behind it. If this view isn't opaque either, the view behind that view has to be displayed and so on. The View class has a method called -displayFromOpaqueAnchestor::: which searches up the view hierarchy for an opaque view and displays it and all of its subviews. Unfortunately this method isn't called automatically by the -display, -dipsplay: and -display:: methods, even if the view is set to be not opaque. Bug or Feature? Have a nice day, Klaus Brouwer
From: flight@mathi.uni-heidelberg.de (Gregor Hoffleit) Newsgroups: comp.lang.python,comp.sys.next.programmer Subject: python with NEXTSTEP NetInfo ? Date: 26 Apr 1996 11:44:02 GMT Organization: University of Heidelberg, Germany Message-ID: <4lqcu3$hr@sun0.urz.uni-heidelberg.de> Summary: python with NetInfo: wishes, bugs ? Keywords: python,NetInfo Has anybody written code to access NEXTSTEP's NetInfo database with python ? I'd like to retrieve additional fields in the NetInfo user passwd database like phone, room and default email address, fields that aren't accessible with getpwent. Until now, I'm using a hack like popen("niutil -read / /users/name="+name+"), but it might be nice to have a general python interface to NetInfo, not ? ;-) Gregor -- | Gregor Hoffleit admin MATHInet / contact RhiNO | | MAIL: Mathematisches Institut PHONE: (49)6221 54-5771 | | INF 288, 69120 Heidelberg / Germany FAX: 54-8312 | | EMAIL: flight@mathi.uni-heidelberg.de (NeXTmail, MIME) |
From: Ravi Mendis <lady0098@sable.ox.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.next.misc Subject: Does WebObjects have Transaction Integrity? Date: Fri, 26 Apr 1996 12:59:34 +0100 Organization: Oxford University Message-ID: <Pine.OSF.3.91.960426122740.23197A-100000@sable.ox.ac.uk> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII The last NEXTSTEP newsletter cited the lack of transaction integrity of NeXT's WebObjects product. In particular it mentioned an Internet application which crashed, dissing the order for a bunch of flowers made over the Web, when demonstrated at a NeXT event. This kind of thing is exactly what NeXT's potential corporate clients are weary of (using Web for business-critical applications). WebObjects' design is well enough to provide transaction processing/monitoring at 2 levels at least: 1. TP HTTP protocol: Like the SSL standard, Web browsers and servers may implement HTTP over a (distributed) transaction processing protocol. This is the cleanest/easiest solution? WebObjects, being independent of both Web servers and browsers, would instantaneously gain transaction integrity. 2. Integrate TP into WebObjects: at the WebAdaptor level it may be possible to transaction monitor WebObject transactions. NeXT will be able to use any (non-distributed) Transaction API available: CICS, MQI, Tuxedo. This is by far the most elegant solution. Further, the transaction-sensitive WebAdaptors could be purchased by those who really require it with the Pro or Enterprise edition of WebObjects. Any third party may be able to supply a transaction secure WebAdaptor for different WebServers?!!? (Hint!) But really it will be to NeXT's tremendous advantage if it were bundled with the WebObjects Pro/Enterprise products. Any other suggestions for integrating transaction integrity into WebObjects? ravi
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: f2c question Date: 26 Apr 1996 14:35:09 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4lqmut$8vr@lion.embl-heidelberg.de> References: <4lood6$24t@charm.magnus.acs.ohio-state.edu> In article <4lood6$24t@charm.magnus.acs.ohio-state.edu> randyj@lubra.sbs.ohio-state.edu (Randy Jackson) writes: > What would be the simplest way for me to install f2c? > > Help appreciated. Get it from: ftp://ftp.nmr.embl-heidelberg.de/pub/next/Developer/Languages. I compiled it for all 4 architectures. -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: Norbert Heger <bertl@hal.kph.tuwien.ac.at> Newsgroups: comp.sys.next.programmer Subject: Re: Making a Window always on top. Date: 26 Apr 1996 14:16:22 GMT Organization: Vienna University of Technology, Austria Message-ID: <4lqlrm$5as@news.tuwien.ac.at> References: <1996Apr23.160740.11729@il.us.swissbank.com> Originator: bertl@black13 Jason Beaver <beaver@il.us.swissbank.com> wrote: > There's already a PostScript function to do this, > PSsetwindowlevel. And there's already an undocumented window method available: @interface Window (Undocumented) - _setWindowLevel:(int)level; @end _______________________________________________________________ Norbert Heger <bertl@hal.kph.tuwien.ac.at> NeXTmail preferred, MIME is welcome Please finger for PGP public key PGP fingerprint = E0 10 3C BB 67 4C 7E 82 76 99 E9 6D B1 8C 32 C2 -- CIAO, Bertl _______________________________________________________________ Norbert Heger <bertl@hal.kph.tuwien.ac.at> NeXTmail preferred, MIME is welcome
Newsgroups: comp.sys.next.programmer From: beaver@il.us.swissbank.com (Jason Beaver) Subject: Re: Making a Window always on top. Message-ID: <1996Apr26.163426.4933@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <4lqlrm$5as@news.tuwien.ac.at> Date: Fri, 26 Apr 1996 16:34:26 GMT In article <4lqlrm$5as@news.tuwien.ac.at> Norbert Heger <bertl@hal.kph.tuwien.ac.at> writes: > And there's already an undocumented window method available: > > @interface Window (Undocumented) > - _setWindowLevel:(int)level; > @end True, but as trivial as this method is, I'd rather not rely on undocumented code... jason __________________ Jason Beaver Swiss Bank Corp. Jason_Beaver@swissbank.com (NeXTMail/MIME preferred) #import <std/disclaimer.h>
Newsgroups: comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.sys.powerpc.advocacy,comp.sys.powerpc.misc,comp.sys.powerpc.tech,comp.sys.psion From: house@direct.ca (HOUSE) Subject: FREE ESSAYS!!! ESSAY EXCHANGE!!! Message-ID: <41a7cc$12f11.1a9@intrepid.cia.com> Date: Fri, 26 Apr 1996 08:47:17 GMT Organization: Direct.ca MIME-Version: 1.0 Content-Type: Text/Plain; charset=US-ASCII I AM NOT AFFLIATED WITH THIS SITE!!!!!! CHECK OUT THE EVIL HOUSE OF CHEAT!!!!!! http://www.hvision.nl/~lovkraft LOTS OF FREE ESSAYS IN ALL TOPICS AND LANGUAGES! CHECK IT OUT!
From: jlandwehr@next.com Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: Re: Does WebObjects have Transaction Integrity? Date: 26 Apr 1996 18:58:12 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4lr6c4$gsc@news.next.com> References: <Pine.OSF.3.91.960426122740.23197A-100000@sable.ox.ac.uk> In article <Pine.OSF.3.91.960426122740.23197A-100000@sable.ox.ac.uk> Ravi Mendis <lady0098@sable.ox.ac.uk> writes: > > > The last NEXTSTEP newsletter cited the lack of transaction integrity of > NeXT's WebObjects product. In particular it mentioned an Internet > application which crashed, dissing the order for a bunch of flowers made > over the Web, when demonstrated at a NeXT event. > > This kind of thing is exactly what NeXT's potential corporate clients are > weary of (using Web for business-critical applications). > FYI, these events have been misrepresented. The Internet application that crashed was at 800FLOWERS website (http://www.800flowers.com). This particular story originated when I was trying to send flowers via the web. They *DO NOT* run WebObjects - they use the Netscape Merchant System. As it turns out, orders placed on their website are only updated to their main order management system (that the telephone reps use) *3 times a day*. You cannot call them to check on the status of your order until that update has been performed. I order flowers at 8:30pm, their server hung, and I couldn't check the status of the order until 7am the next morning! Steve Jobs tells this story because WebObjects would vastly improve their site. WebObjects would allow them to integrate their website directly into their customer order management system and not have to worry about updates happening only three times a day! -- John Landwehr Product Manager NeXT Software Inc. john_landwehr@next.com
From: Eric_Noyau@next.com (Eric Noyau) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 26 Apr 1996 21:24:45 GMT Organization: NeXT Software, Inc. Message-ID: <4lreut$hf8@news.next.com> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> In article <x7afzzmilx.fsf@tom.es.ele.tue.nl> tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: > In article <4loqda$bpf@news.next.com> Eric_Noyau@next.com (Eric Noyau) writes: > > What's wrong with 'print [foo delegate]' ? > > The `foo' might not respond to `delegate'. --Tiggr > Please do not remove the context of my answer !! We where talking about looking into the instances variables of a private class, subclass of a class cluster, for debugging purpose. My point is the only state you may need from these objects is the public state, available thru method calls on the abstract superclass. In another words, if there is no API for to access an instance variable on such a class, there is no need for you to access it. And this is specially important while debugging, because it's easy for you to misinterpret... (Aka: "Yeah, right, this is a NSConcreteMutableArray, so the first 32 bits after the isa pointer should contains the length of the array...". Wrong! Guess again. Or use [foo count].) In addition, I'd like to point that as a user of Foundation (I'm working on EOF), I have to use gdb sometimes, and, believe it or not, I do not have access to the instance variables of any NSArray, NSDictionary or NSString instance. Like any of you. And I have no need to access them, because the state of these objects is always available thru method calls. I'm sure you've heard about Object Oriented development. This is a part of it: you don't want to know how an object work to be able to use it. And please: I DO NOT SPEAK FOR NeXT! -- Eric PS: Okay, stop guessing ;^) The length of an NSConcreteMutableArray is stored on the first 30 Bits after the isa pointer...
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 27 Apr 1996 01:07:36 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4lrs0o$ido@news.next.com> References: <4lreut$hf8@news.next.com> In article <4lreut$hf8@news.next.com> Eric_Noyau@next.com (Eric Noyau) writes: > PS: Okay, stop guessing ;^) The length of an NSConcreteMutableArray is > stored on the first 30 Bits after the isa pointer... In 4.0, that is. In 4.1, it will be different. Chris Kane Foundation Technology
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Where are "tcgetattr" and "tcsetattr"? Date: 25 Apr 1996 15:45:21 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4lo6mh$f7v@news.its.com> References: <4l444f$e7j@pollux.usc.edu> <4lis2l$nbs@bertrand.ccs.carleton.ca> csaldanh@mae.carleton.ca (Chris Saldanha) wrote: > Sungdo Moon (sungdomo@pollux.usc.edu) wrote: > : I'm a user of NeXTSTEP 3.3 for Intel and also have Developer 3.3. > : I compiled a program which uses "tcgetattr" and "tcsetattr". > : Those are defined in <termios.h> and exists man pages for them. > : But I got link errors. Does any body know what library need to be linked? > > Link with -lposix (the POSIX library), and those symbols will be resolved. > It's possible that the object files that compose your program may need to > be compiled with -posix, the POSIX source flag. Don't try to link with -lposix directly. You will almost always not get a correctly functioning executable-- about the only time that works is if you're using some trivial POSIX function that does not depend on the special initialization done in /lib/posixcrt0.a. If you need to use POSIX terminal handling, which tcgetattr() and tcsetattr() qualify as, you must compile and link everything using "cc -posix". -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: house@direct.ca (HOUSE) Newsgroups: comp.sys.mac.programmer.codewarrior,comp.sys.mac.programmer.games,comp.sys.mac.programmer.help,comp.sys.mac.programmer.misc,comp.sys.mac.scitech,comp.sys.mac.system,comp.sys.mac.wanted,comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: FREE ESSAYS!!!!! EVIL HOUSE OF CHEAT!!!! Date: 27 Apr 1996 07:25:52 GMT Organization: Direct.ca Message-ID: <4lsi60$ki5@spark.law.vill.edu> Mime-Version: 1.0 Content-Type: Text/Plain; charset=US-ASCII I AM NOT ASSOSIATED WITH THIS PAGE! JUST A FREQUENT USER!!! CHECK OUT THE EVIL HOUSE OF CHEAT!!!! http://www.hvision.nl/~lovkraft LOTS OF FREE ESSAYS IN ALL TOPICS AND LANGUAGES!! CHECK IT OUT!!!!
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Darcy's view on NeXT classes undisputed? (was Why does FK introduce a new..) Date: 26 Apr 1996 19:50:22 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4lr9du$nam@news.its.com> References: <4levpu$lpu@wagner.spc.videotron.ca> <4li2uj$610@isabella.object-factory.com> <4ljte3$b7e@infoserv.rug.ac.be> stes@cage.rug.ac.be (David Stes) wrote: > What does OpenStep win by excluding some "faithful" NextStep developers, > and by labeling them as "anti-NeXT", simply because of a genuine concern > about what's happening with all these Object-NSObject changes and by the > rewriting in the large of things that were working perfectly fine before ? Good question! Surely no one could believe that NeXT's products have always been so perfect that they cannot possibly be criticized. Given this, why do some people believe that you are "anti-NeXT" if you voice a criticism? Hmm-- I wonder if Georg Tuparev has any insight about this...? -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Optimize ??? Date: 26 Apr 1996 20:56:43 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4lrdab$nam@news.its.com> References: <4l3qnm$hcj@news.cyberlink.ch> <4l4tho$c6d@agate.berkeley.edu> <4l7l36$cf@ironhorse.plsys.co.uk> <4lgf9h$2ob@gaea.titan.org> <4lluer$c2m@srvr1.engin.umich.edu> premise@engin.umich.edu (Sean Willson) wrote: > I ran the strip operator on a NeXTStep application I am writing and it > shrunk it down from 2M to 100K, this was huge, BUT it failed to load > bundles and more. I defined a protocol the bundles respond to, and it > appears that strip removed it. Any way to selectively strip? NeXTApps? Read the documentation. From 'man strip': -A Save all global absolute symbols except those with a value of zero, and save Objective C class symbols. This is intended for use of programs that use rld(3) and want the loaded code to use symbols from the shared libraries. -n Save all N_SECT global symbols. This is intended for use with executable programs in combination with -A to remove the symbols needed for correct static link edit- ing which are not needed for use with rld(3) where using the -s filename would be too much trouble. -u Save all undefined symbols. This is intended for use with relocatable objects to save symbols referred to by external relocation entries. Note that common symbols are also referred to by external relocation entries and this flag does not save those symbols. These options specify symbols to be removed from the result- ing output file. -S Remove the debugging symbol table entries (those created by the -g option to cc(1) and other compilers). -X Remove the local symbols whose names begin with `L'. -x Remove all local symbols (saving only global symbols). And from Makefile.postamble: # Options to strip for bundles, apps with bundles, and apps without bundles, # respectively. RELOCATABLE_STRIP_OPTS = -x -u DYLD_APP_STRIP_OPTS = -A -n APP_STRIP_OPTS = $(DYLD_APP_STRIP_OPTS) TOOL_STRIP_OPTS = -X # (Note: APP_STRIP_OPTS and TOOL_STRIP_OPTS default to empty, but # developers doing their own dynamic loading should set this to # $(DYLD_APP_STRIP_OPTS)). Of course, doing a "build install" using ProjectBuilder.app should do the right things to the various components of your application... -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: headi@now.ch (Daniel Scheidegger) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Strange inetd problem Date: Fri, 26 Apr 1996 10:00:16 GMT Organization: NOW GmbH, Baar, Switzerland Sender: headi@now.ch (Daniel Scheidegger) Message-ID: <DqGrsG.Cz.0.astra@now.ch> Folks, this is driving me insane... I want to write a little daemon which should be called by inetd. This is what my /etc/inetd.conf entry looks like: mptn stream tcp nowait root /Users/headi/Apps/MPTN.app/MPTN /Users/headi/Apps/MPTN.app/MPTN As you may notice, it is a real .app which should act as a daemon. The service mptn is correctly registered in netinfo. If I do a 'telnet localhost mptn', the app gets started, but crashes when it tries to load it's nib, at the following statement in main(): if ([NXApp loadNibSection:"MPTN.nib" owner:NXApp withNames:NO]) [NXApp run]; However, if I manually kill inetd and restart it (as root of course, and a simple kill -HUP does not work) and try again, it works!. Why do I have to kill the inetd strated in etc/rc and start a new one first?? I have to do this everytime I reboot my machine (which is really not often, but I don't like this behaviour anyway...) Any help would be greatly appreciated Daniel -- Daniel Scheidegger Software Engineer, System Administrator NOW GmbH, Zugerstrasse 11a, CH-6340 Baar ++41-1-2208223 / dscheide@now.ch
From: taweil@skat.usc.edu (Ta-Wei Li) Newsgroups: comp.sys.next.programmer,comp.sys.next.bug Subject: rld limitation on the length of the symbol name?! Date: 27 Apr 1996 11:14:45 -0700 Organization: University of Southern California, Los Angeles, CA Sender: taweil@skat.usc.edu Message-ID: <c4aivelsfe5.fsf@skat.usc.edu> I ran into a strange problem when I tried to use rld functions. There seems to be a limit on the length of symbol length for these functions but they are are documented anywhere. The following two programs show that. The foo works and load bar.o with no problem. However, if I uncomment the Kaff... function, the rld_load fails. Is this a bug? Any way to work arround it? ------ bar.c #include <stdio.h> bar () { printf ("This is bar\n"); } /* Kaffe_java_io_FileDescriptor_initSystemFD_stub () { print ("This is Kaffe_java_io_FileDescriptor_initSystemFD_stub\n"); }*/ ------ foo.c #include <stdio.h> #include <mach-o/rld.h> #include <streams/streams.h> main () { char *files[] = {"bar.o", NULL}; void (*f)(); NXStream *s = NXOpenFile(fileno(stdout), NX_WRITEONLY); if (!rld_load (s, NULL, (const char * const *)files, NULL)) printf ("can not load\n"); if(!rld_lookup(s, "_bar", (long *)&f)) printf ("can not lookup\n"); (*f)(); } -------- -- Ta-Wei Li
From: stefanos@Vir.com (Stefanos Kiakas) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: NeXTSTEP driver information Web site. Date: 27 Apr 1996 13:13:16 -0400 Organization: Communications Vir, Internet Access Montreal. Message-ID: <4ltkjc$iq1@Vir.com> Hello, There is a new Web site available, which contains information about creating device drivers NeXTSTEP and OpenStep. The Web site is under construction and there is a need for submissions and people to review the contents to create a truly useful Web site. URL http://www.uniscape.com/NSDrivers You need a browser which can view tables (Netscape or OmniWeb2.0) If you have any information to submit please, email me at stefanos@uniscape.com I'm maintaining the Web site on my spare time so I don't always have the time to quickly respond to your comments or any requests, but I will try to get to them as quickly as possible. Sincerely, Stefanos Kiakas stefanos@uniscape.com
From: Ravi Mendis <lady0098@sable.ox.ac.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: Re: Does WebObjects have Transaction Integrity? Date: Sat, 27 Apr 1996 19:33:11 +0100 Organization: Oxford University Message-ID: <Pine.OSF.3.91.960427191111.16883B-100000@sable.ox.ac.uk> References: <Pine.OSF.3.91.960426122740.23197A-100000@sable.ox.ac.uk> <4lr6c4$gsc@news.next.com> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4lr6c4$gsc@news.next.com> On 26 Apr 1996 jlandwehr@next.com wrote: [edit] > ...The Internet application that > crashed was at 800FLOWERS website (http://www.800flowers.com). This > particular story originated when I was trying to send flowers via the web. > They *DO NOT* run WebObjects - they use the Netscape Merchant System. As it > turns out, orders placed on their website are only updated to their main order > management system (that the telephone reps use) *3 times a day*. You cannot > call them to check on the status of your order until that update has been > performed. I order flowers at 8:30pm, their server hung, and I couldn't check > the status of the order until 7am the next morning! > > Steve Jobs tells this story because WebObjects would vastly improve their > site. WebObjects would allow them to integrate their website directly into > their customer order management system and not have to worry about updates > happening only three times a day! > Thanks for straightening out the facts. However, i'm still clueless as to the transaction integrity of applications based on WebObjects? The fact that the 800FLOWERS application did not use WebObjects doesn't mean that WebObjects possess transaction integrity does it? Explicitly: the level of transaction integrity of corporate applications based on CICS, MQI or other (distributed) TP APIs. It could. Would it not be a point in favor of WebObjects' greater success in the corporate world? Would it not be 'straight-forward' to implement within the current WebObject Framework? (e.g: via TP monitored WebAdaptors?) ravi
From: house@direct.ca (HOUSE) Message-ID: <cancel.4lsi60$ki5@spark.law.vill.edu> Control: cancel <4lsi60$ki5@spark.law.vill.edu> Newsgroups: comp.sys.mac.programmer.codewarrior,comp.sys.mac.programmer.games,comp.sys.mac.programmer.help,comp.sys.mac.programmer.misc,comp.sys.mac.scitech,comp.sys.mac.system,comp.sys.mac.wanted,comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: cmsg cancel <4lsi60$ki5@spark.law.vill.edu> Date: 27 Apr 1996 20:18:41 GMT Cancelled by jem@xpat.com. 830636321 LovKraft Original Subject was: FREE ESSAYS!!!!! EVIL HOUSE OF CHEAT!!!!
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 27 Apr 1996 23:59:21 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7pw8tjpl2.fsf@tom.es.ele.tue.nl> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> <4lreut$hf8@news.next.com> In-reply-to: Eric_Noyau@next.com's message of 26 Apr 1996 21:24:45 GMT In article <4lreut$hf8@news.next.com> Eric_Noyau@next.com (Eric Noyau) writes: Please do not remove the context of my answer !! Why not? We where We what? talking about looking into the instances variables of a private class, subclass of a class cluster, for debugging purpose. And only through debugger commands; not from within a program. My point is the only state you may need from these objects is the public state, available thru method calls on the abstract superclass. Non-source-distributed libraries can have bugs. Tracking down these bugs should not be harmed by any state protection or whatever other Object Oriented Programming paradigm. Debugging is not programming. Object Oriented Debugging does not exist. In another words, if there is no API for to access an instance variable on such a class, there is no need for you to access it. I am not a program. A program must adhere to an API; that's all it can do; it has no choice. And this is specially important while debugging, because it's easy for you to misinterpret... I do not want to be protected against misinterpreting something. I want access to information which is available in the system. If the system refuses to output something useful if I type `p *$', I can always type `x/10x $' (which I never do). (Aka: "Yeah, right, this is a NSConcreteMutableArray, so the first 32 bits after the isa pointer should contains the length of the array...". Wrong! Guess again. Or use [foo count].) There is no need to do any guessing, as every class defines the state contained in its instances. I'm not interested in how many bits at which offset `count' is stored; I'm only interested in its value. Given the object information already present, the debugger can present me this value. And if I type `p *$' I think I can interpret the meaning of the variable named `___my_private_length'. Besides, the whole issue doesn't concern single simple ivar printing; it concerns (1) presenting multiple instance variables of the same object. Typing `p *$' is much nicer than typing `p [$23 foo] RET p [$23 bar]'. Another point is (2) the speed of the whole thing: a method invocation in the child process costs much more time than a simple child-address-space pointer dereference. And this is especially important if you're interested in values at a breakpoint which is hit often. In addition, I'd like to point that as a user of Foundation (I'm working on EOF), I have to use gdb sometimes, and, believe it or not, I do not have access to the instance variables of any NSArray, NSDictionary or NSString instance. Like any of you. And I have no need to access them, because the state of these objects is always available thru method calls. The objects you mention are trivial examples. Their state is interesting at a much higher level, available through `p [[$ description] cString]'. I'm sure you've heard about Object Oriented development. This is a part of it: you don't want to know how an object work to be able to use it. Debugging is not programming. Object Oriented Debugging does not exist. And please: I DO NOT SPEAK FOR NeXT! Relax. --Tiggr
From: stanj@cs.stanford.edu (Stan Jirman) Newsgroups: comp.sys.next.programmer Subject: Color NSBrowserCells (NS4.0) Date: 28 Apr 1996 01:21:58 GMT Organization: Stanford University Message-ID: <4luh7m$ilp@nntp.Stanford.EDU> This is about a problem with creating colored cells in a browser on 4.0 PR2. If you have any insight, please let me know... My described approach worked just fine at Sun with their OpenStep, and should work according to the docs in PR2 as well. In 3.3, when I wanted a browser with colorful cells, the thing was to set the NXBrowser to use a cell subclass, and then make the subclass of the NXBrowserCell take care of the changing of the color, like this: - setTextAttributes:textObject { [super setTextAttributes:textObject]; [textObject setTextGray:NX_BLACK]; return self; } Now, this method has been renamed. My DirEntryCell:NSBrowserCell has this: - (NSText *) setUpFieldEditorAttributes:(NSText *)textObject { [super setUpFieldEditorAttributes:textObject]; [textObject setTextColor:[NSColor redColor]]; return textObject; } But it doesn't work. I even set a breakpoint in [NSCell setUpFieldEditorAttributes:], but gdb never halted... What's today's secret? It's just about recoloring the text, so I think to overload the whole drawing function is rather an overkill... Especially when the class is NSBrowserCell, writing the whole proper displaying of the arrow etc. is a mess. Any help is appreciated. - Stan --- +-------------------------------------------+-------------------------+ | Stan Jirman -- The Swiss CS Guy | | | stanj@cs.stanford.edu NeXTmail / MIME | When you find yourself | | http://www-leland.stanford.edu/~stanj/ | in a hole, stop digging | | PO Box 2642, Stanford, CA 94309, USA | | +-------------------------------------------+-------------------------+
From: John McPherson <mcphersj@sojourn.com> Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc Subject: FOR SALE ($2000) : NSFIP 3.3 User, Developer, and EOF commercial versions Date: Sat, 27 Apr 1996 23:59:12 -0400 Organization: Sojourn Systems, Lansing, MI (USA) Message-ID: <3182ED10.2C6B@sojourn.com> References: <317C0D17.7719@ftp.test.com> <317C0DAA.55C@sojourn.com> <317EE46B.426E@sojourn.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit John McPherson wrote: > > FOR SALE: > > NeXTstep 3.3 User (Commercial version) $799.00 > NeXTstep 3.3 Developer (Commercial version) $4999.00 > NeXTstep EOF (Commercial version) $499.00 > -------- > (Total) $6297.00 value (Obviously this is not the asking > price...) > > These are being sold as on package. I am also including for free the book NEXTSTEP PROGRAMMING >Step One: Object Oriented Programming by Garfinkel and Mahoney. In addition, I will be including 3 >OPENSTEP video tapes, a NeXT Vs. Sun video tape, and The NeXT Advantage video tape. I have >installed this only once. All the manuals are 'like new!' > > Please send mailto://mcphersj@sojourn.com > > I intend to sell this quickly! > > Thanks!
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside of @implementation?) Date: 27 Apr 1996 23:04:36 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4lu964$l1@news.its.com> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> <4lreut$hf8@news.next.com> Eric_Noyau@next.com (Eric Noyau) wrote: [ ... ] > Please do not remove the context of my answer !! Very well, I'll include the context as you've defined it: > We where talking about looking into the instances variables of a private > class, subclass of a class cluster, for debugging purpose. My point is the > only state you may need from these objects is the public state, available > thru method calls on the abstract superclass. [ ... ] > In addition, I'd like to point that as a user of Foundation (I'm > working on EOF), I have to use gdb sometimes, and, believe it or not, > I do not have access to the instance variables of any NSArray, > NSDictionary or NSString instance. Like any of you. And I have no > need to access them, because the state of these objects is always > available thru method calls. I disagree with the assertion that a developer never has a need to be able to see the implementation (instance variables, et cetera). The public interface may not be sufficient due to: a) bugs in NeXT classes b) insufficient or just plain incorrect documentation on the effects of certain method calls. Since NeXT does not provide source for your classes, unlike many other OO environments I'm familiar with, using the debugger to see instance variables is the only available mechanism for finding out what's actually going on inside the black box of an object. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 29 Apr 1996 04:15:12 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4m1fog$kcg@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <xaf@saem-ales.saem-ales.fr> Message-ID: <9604290827.AA00279@saem-ales.saem-ales.fr> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: fabre xavier <xaf@saem-ales.saem-ales.fr> Date: Mon, 29 Apr 96 10:27:18 +0200 Subject: Card with multi serial port under NeXTStep ? I need, for an application, 5 to 7 serials ports. The big question is, if this card exists, how can I drive it under NeXTStep ?? Xavier Fabre. ********************************************************************* * Do you want to work under NeXTStep only with your eyes ? * * http://www.saem-ales.fr/~chet/eyeaccess.html * *********************************************************************
From: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: EORowAssociation class ? Date: 29 Apr 1996 10:13:31 GMT Organization: Universite de La Rochelle Message-ID: <4m24ob$a1v@hpuniv.univ-lr.fr> I create an NXTableView and EOColumnAssociations. I connect them : it works ! But I want to display EObject in column instead of row ! so I replace (NXTableView *) addColumn: by (NXTableView *) addRow: in my code but there is not an EORowAssociation corresponding to EOColumnAssociation!!! Has someone allready develop an EORowAssociation class to display EObjects from an EOController into an NXTableView ? here is my Code : (statTV is an outlet connected to a real NXTableView in IB.) (statWindow is an outlet connected to a real Window in IB.) - createStatGUI { NXTableView *myTV; NXTableVector *myTVc; NXRect myRect; NXTextFormatter *oldF = [statTV formatterAt:0:0]; int i, numrows=[statTV columnCount]; NSArray *attributesIdentifiers=[NSArray alloc]; NSString *modelPath=[EOModel findPathForModelNamed:@"myEOModel"]; EOModel *model=[[EOModel alloc] initWithContentsOfFile:modelPath]; EODatabaseDataSource *datasource; EOController *controller; EOAdaptor *adaptor = [[EOAdaptor adaptorWithModel: model] retain]; EOEntity *entity = [[adaptor model] entityNamed:@"MYENTITY"]; EODatabase *database = [[EODatabase allocWithZone: [self zone]] initWithAdaptor: adaptor]; EODatabaseContext *dbContext = [[EODatabaseContext alloc] initWithDatabase: database]; EODatabaseChannel *dbChannel = [[EODatabaseChannel alloc] initWithDatabaseContext: dbContext]; if (! [dbChannel openChannel]) NSLog(@"database channel could not be opened."); [[dbChannel adaptorChannel] setDebugEnabled:YES]; datasource = [[EODatabaseDataSource allocWithZone:[self zone]] initWithDatabaseChannel:dbChannel entityNamed:[entity name]]; controller=[[EOController allocWithZone: [self zone]] initWithDataSource:datasource]; attributesIdentifiers=[entity attributes]; [dbChannel setDelegate:self]; [controller setDelegate:self]; [statWindow disableDisplay]; [statTV getFrame:&myRect]; myTV = [[NXTableView alloc]initFrame:&myRect]; [myTV setAutosizing:[statTV autosizing]]; [[statTV superview] addSubview:myTV]; [statTV removeFromSuperview]; statTV = myTV; [statTV setFrame:&myRect]; [statTV setVertScrollerRequired:YES]; [statTV setHorizScrollerRequired:YES]; // [statTV setRowHeadingVisible:YES ]; // [statTV setColumnHeadingVisible:NO ]; [statTV setRowHeadingVisible:NO]; [statTV setColumnHeadingVisible:YES]; [statTV setMode:NX_RADIOMODE]; for(i=0;i<numrows;i++) id association; // [statTV addRow: [attributesIdentifiers objectAtIndex:i] [statTV addColumn:[attributesIdentifiers objectAtIndex:i] withFormatter:oldF andTitle:[[(EOAttribute *)[attributesIdentifiers objectAtIndex:i] name] cString] at:i]; // myTVc = [statTV rowAt :i]; myTVc = [statTV columnAt:i]; [myTVc setResizable:NO]; [myTVc setContentAlignment:NX_RIGHTALIGNED]; association=[[ EOColumnAssociation allocWithZone:[self zone]] initWithController:controller key:[(EOAttribute *)[attributesIdentifiers objectAtIndex:i] name] destination:myTVc]; [association setTableView :statTV]; [controller addAssociation:association]; [association autorelease]; [ controller fetch]; [statWindow reenableDisplay]; [controller redisplay]; [statWindow makeKeyAndOrderFront:self]; return self; } /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Drawing lines in color? Date: 29 Apr 1996 12:38:29 GMT Organization: University of Sheffield, UK Message-ID: <4m2d85$b7g@bignews.shef.ac.uk> References: <4lp2nv$2ob@turbocat.snafu.de> In-Reply-To: <4lp2nv$2ob@turbocat.snafu.de> On 04/26/96, David Wetzel wrote: > How do I draw fast lines in color? DPSDoUserPath works only with one color, > I think. > > I use this now: > > defineps PSWdrawLine (float x; float y;float xx; float yy; float r; float g; float b; float width) > x y moveto > xx yy lineto > width setlinewidth > r g b setrgbcolor > stroke > endps > You will get a small improvement if you _bind_ the definition: e.g. defineps PSWDefs() /PDL { % x y xx yy width r g b moveto lineto setlinewidth setrgbcolor stroke } bind def endps then refer to it it thus: defineps PSWdrawLine (float x; float y;float xx; float yy; float width float r; float g; float b) x y xx yy width r g b PDL endps You have to set up the PSWDefs in the initFrame method: - initFrame(cnst NXRect *)frameRect { [super initFrame:frameRect]; ... PSWDefs(); return self; } This is documented in Adobe's excellent Purple book ("Programming the Display PostScript System with NeXTstep", pp 54-55): the book also gives timings for the various methods: for 1000 lines of random width and colour wraps take 2639ms, whereas wraps with binding takes 2584ms. Not a lot in it, but for the extra effort worthwhile, particularly if you ever save the image as the PostScript file may well be much smaller. Best wishes, mmalc. posn. research facilitator where institute for language speech and hearing sheffield university west court 2 mappin street sheffield s1 4dt england vox (+44) 114 282 5269 fax (+44) 114 278 0972 email m.crawford@dcs.shef.ac.uk NeXTMail, SunMail, MIME welcome PGP key available on request http://www.dcs.shef.ac.uk/research/ilash/ --
Date: 29 Apr 1996 10:35:29 EST Newsgroups: comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.sys.powerpc.advocacy,comp.sys.powerpc.misc,comp.sys.powerpc.tech,comp.sys.psion Message-ID: <cancel.41a7cc$12f11.1a9@intrepid.cia.com> Control: cancel <41a7cc$12f11.1a9@intrepid.cia.com> From: clewis@ferret.ocunix.on.ca Sender: house@direct.ca (HOUSE) Subject: cmsg cancel <41a7cc$12f11.1a9@intrepid.cia.com> EMP/ECP (aka SPAM) cancelled by clewis@ferret.ocunix.on.ca. See news.admin.net-abuse.announce, report 19960429.01 for further details
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Contract/DC Area Date: 29 Apr 1996 14:50:35 GMT Organization: Norden 1 Communications Message-ID: <4m2kvr$l0l@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF---A plus Sybase or Oracle---A plus Contract-----Long term DC Area----Nice sub Must Be US Citizen or Greencard To be considered---Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: EORowAssociation class ? Date: Mon, 29 Apr 1996 20:53:28 GMT Organization: Orbital Computer Consultancy Message-ID: <31852b58.815244@news.plsys.co.uk> References: <4m24ob$a1v@hpuniv.univ-lr.fr> wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) wrote: > >I create an NXTableView and EOColumnAssociations. I connect them : it works ! >But I want to display EObject in column instead of row ! >so I replace (NXTableView *) addColumn: by (NXTableView *) addRow: in my code >but there is not an EORowAssociation corresponding to EOColumnAssociation!!! >Has someone allready develop an EORowAssociation class to display EObjects from an EOController into an NXTableView ? > > >here is my Code : >(statTV is an outlet connected to a real NXTableView in IB.) >(statWindow is an outlet connected to a real Window in IB.) <-- code deleted --> You'll be pleased to know there's a simple solution to this. Simply paste the following into your code: @implementation EOColumnAssociation(RowHack) - (unsigned int)columnCount { return [[[self controller] allObjects] count]; } @end Why Next didn't do this anyway, and call the class EOVectorAssociation, is anyone's guess. BTW, turn row and column headings on for an interesting effect. -- If I wrote more, I'd probably make more sense. But time is the fire in which we burn.
From: Žruss@physical27.chem.ufl.edu (russ bowers) Newsgroups: comp.sys.next.programmer Subject: gnu fortran, g77 Date: 30 Apr 1996 01:37:31 GMT Organization: University of Florida Message-ID: <4m3qsr$4eq@no-names.nerdc.ufl.edu> Hi, A few weeks ago I posted a request for information on how to compile g77 running on NS3.3 HPPA. Although I learned that this is possible on NS Intel or Motorola, I did not get any replies on how to run ./configure and it appears that g77 is not supported with my particular CPU/Hardware configuration. An alternative that I did find very useful is the fortran-to-C converter, called f2c. You may already know that it exists, but I just recently discovered it. I have used it to convert a numerical differential equations solver from NETLIB (netlib.att.com:/netlib/f2c) and it works great. Just thought this might be useful information to someone out there. -------------------- Russ Bowers Chemistry Department University of Florida
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Berkeley libdb available? In-Reply-To: chinay@cig.mot.com's message of 28 Apr 1996 16:47:13 GMT Message-ID: <u41pw8q9w45.fsf@world.std.com> Sender: rdl@world.std.com Organization: The World Public Access UNIX, Brookline, MA References: <4m07eh$68l@trotsky.cig.mot.com> Date: Tue, 30 Apr 1996 04:25:14 GMT I have. E-mail me. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com In article <4m07eh$68l@trotsky.cig.mot.com> chinay@cig.mot.com (Albert Chin-A-Young) writes: From: chinay@cig.mot.com (Albert Chin-A-Young) Newsgroups: comp.sys.next.programmer Date: 28 Apr 1996 16:47:13 GMT Organization: Motorola Cellular Infrastructure Group Xref: world comp.sys.next.programmer:27300 Path: world!blanket.mitre.org!news.mathworks.com!zombie.ncsc.mil!nntp.coast.net!lll-winken.llnl.gov!uwm.edu!math.ohio-state.edu!newsfeed.acns.nwu.edu!ftpbox!mothost!news-in.cig.mot.com!booboo!chinay Lines: 5 NNTP-Posting-Host: booboo.cig.mot.com X-Newsreader: TIN [version 1.2 PL2] Has anyone ported Berkeley's libdb to NS 3.3? I've tried but get to a point where I can't resolve munmap. -- albert chin (ITS Service Team 2)
Newsgroups: comp.sys.next.programmer From: rdl@world.std.com (Robert La Ferla) Subject: Re: Card with multi serial port under NeXTStep ? In-Reply-To: fabre xavier's message of Mon, 29 Apr 96 10:27:18 +0200 Message-ID: <u41ohoa9w2d.fsf@world.std.com> Sender: rdl@world.std.com Organization: The World Public Access UNIX, Brookline, MA References: <9604290827.AA00279@saem-ales.saem-ales.fr> Date: Tue, 30 Apr 1996 04:26:18 GMT The Mux driver will drive it. The ISASerialPort/PortServer drivers will probably do so too. DigiBoard is a good hardware candidate. Robert La Ferla Registered OPENSTEP / NEXTSTEP Consultant Boston, MA Tel: + 1 (617) 252-0088 Fax: + 1 (617) 252-0004 E-mail: Robert_La_Ferla@hot.com
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: Problem with key equivalents in buttons Date: 30 Apr 1996 09:40:21 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4m4n65$7bu@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hello, i have problems with key equivalents in buttons: i use buttons with \r (return) key equivalents within Panel windows: in some of these panels, the buttons' equivalents work while in others they do not work. I could not see the difference between the 2 kinds: both kinds of panels were made with IB (i am working under version 3.3) in the most usual way. I checked at run time the key equivalents of these buttons (with Button's keyEquivalent method, which returns 13 in all cases) but have found nothing wrong. ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
From: egeze@stage1.univ-lr.fr (Emmanuel Geze) Newsgroups: comp.databases.oracle,comp.sys.next.software,comp.sys.next.programmer Subject: Using SchemaView of SchemaResearch Date: 30 Apr 1996 09:03:49 GMT Organization: Universite de La Rochelle Message-ID: <4m4l1l$k4m@hpuniv.univ-lr.fr> Keywords: Oracle, SchemaView,SchemaResearch,NextStep,EOF,Adaptor Hi all, I work with NeXTSTEP 3.3, EOF 1.1 and Oracle v7.0 or v7.2. I would like use SchemaView product, but I'm problems with it. Then I'd like know if there are people who use this application. What do you do to install this product ? I've already installed the Oracle 6 compatibility views but my problem is always the same. When I launch SchemaView, I've this error message: Table ACCESSIBLE_COLUMNS was not found. Do you have the Oracle 6 compatibility views installed? See file rdbms/admin/catalog6.sql in the Oracle distribution. But with sqlplus, I can see the table ACCESSIBLE_COLUMNS and I can fetch it !!! Now, I don't know what I can do. Do you know another product which are the same functions ? Any hints are welcome. Emmanuel email: egeze@cri.univ-lr.fr (NeXTMail, MIME mail)
From: wolfgang@wi.WHU-Koblenz.de (Wolfgang Roeckelein) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: compiling pgp263is on intel 3.3 w. patch Date: 30 Apr 1996 10:16:22 GMT Organization: University of Regensburg, Germany Message-ID: <4m4p9m$30r@rrzs3.uni-regensburg.de> Hi, has anybody succeeded in compiling pgp263is on NS intel 3.3 with the patch. The linker keeps complaining about ld: Undefined symbols: _strdup Please help. Thank you very much in advance, Wolfgang -- Dipl.-Wirtsch.Inf. Voice: +49 941 943 2998 Wolfgang Roeckelein Fax: +49 941 943 4986 Uni Regensburg E-Mail: roeckelein@wi.whu-koblenz.de Universitaetsstr. 31 Wolfgang.Roeckelein@wiwi.uni-regensburg.de D-93053 Regensburg (MIME and NeXTmail ok) Germany WWW: http://www.whu-koblenz.de/~wolfgang/
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: gnu fortran, g77 Date: 30 Apr 1996 12:00:30 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4m4vcu$l8a@lion.embl-heidelberg.de> References: <4m3qsr$4eq@no-names.nerdc.ufl.edu> In article <4m3qsr$4eq@no-names.nerdc.ufl.edu> russ@physical27.chem.ufl.edu (russ bowers) writes: > > Hi, > > A few weeks ago I posted a request for information on how to compile > g77 running on NS3.3 HPPA. Although I learned that this is possible > on NS Intel or Motorola, I did not get any replies on how to run > ./configure and it appears that g77 is not supported with my > particular CPU/Hardware configuration. The reason is very simple. To compile g77 on HP/NS you need gcc.2.7.x but this version of gcc cannot be compiled on HP/NS and it seams that nobody is working on it although Paul Kunz was spreading rumors a while ago :-( -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: crath@bnr.ca (Christopher Rath) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: compiling pgp263is on intel 3.3 w. patch Date: 30 Apr 1996 13:51:07 GMT Organization: Bell Northern Research, Ottawa, Canada Message-ID: <CRATH.96Apr30095107@bmerhe83.bnr.ca> References: <4m4p9m$30r@rrzs3.uni-regensburg.de> In-reply-to: wolfgang@wi.WHU-Koblenz.de's message of 30 Apr 1996 10:16:22 GMT In article <4m4p9m$30r@rrzs3.uni-regensburg.de> wolfgang@wi.WHU-Koblenz.de (Wolfgang Roeckelein) writes: > The linker keeps complaining about ld: Undefined symbols: _strdup Here's some code I keep around for just this type of situation: #if defined(NeXT) /** * strdup() - Duplicate a string. * * Returns a pointer to a new string which is a duplicate of the * string to which s1 points. The space for the new string is * obtained using the calloc() function (see calloc(3C)). * * Parameters: * [1] inStr - String to duplicate. * * Returns: * Pointer to duplicated string; or NULL if an error occurred. **/ char *strdup(const char *inStr) { char *outStr; if (NULL == inStr) { outStr = NULL; } else { outStr = calloc((strlen(inStr) + 1), sizeof(char)); if (NULL != outStr) { strcpy(outStr, inStr); } } return outStr; } #endif -- === Christopher Rath ===== crath@bnr.ca ===== (613) 765-3141 === Northern Telecom Ltd. | Box 3511, Station `C' | ``Hydrogen is a colourless, odourless Ottawa, ON K1Y 4H7 | gas which, given enough time, turns FAX: (613) 763-4101 | into people.'' --- Henry Hiebert
From: matthias@amg.de (Matthias Schuerhoff) Newsgroups: comp.sys.next.programmer Subject: Re: Problem with key equivalents in buttons Date: 30 Apr 1996 12:20:31 GMT Organization: AMG Industrieconsulting GmbH Message-ID: <4m50if$sje@hagen.amg.de> References: <4m4n65$7bu@naiad.grenet.fr> Matija Exel <exel@lag.ensieg.fr> wrote: > > hello, > > i have problems with key equivalents in buttons: i use > buttons with \r (return) key equivalents within Panel > windows: in some of these panels, the buttons' equivalents > work while in others they do not work. I could not see the > difference between the 2 kinds: both kinds of panels were > made with IB (i am working under version 3.3) in the most > usual way. > I checked at run time the key equivalents of these buttons > (with Button's keyEquivalent method, which returns 13 in all > cases) but have found nothing wrong. > > > > ----------------------------------------------------------------- > Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., > Tel : 76 82 71 12 Fax: 76 82 63 88 > e-mail : exel@lag.ensieg.fr (NO NextMail please) > ----------------------------------------------------------------- > > Hello, only an idea: Are you sure that all of your panels are set to the class "Panel" in IB inspector? With then class "Window" the key equivalents don't work. Perhaps that helps... Matthias ===================================== Matthias Schuerhoff AMG Industrie Consulting GmbH Joseph-von-Fraunhofer-Str. 27 44227 Dortmund / Germany Telefon: 0231 / 97 53 54 - 0 Telefax: 0231 / 97 53 54 - 55 E-Mail matthias@amg.de (NeXTmail ok) =====================================
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: Re: Problem with key equivalents in buttons Date: 30 Apr 1996 15:51:49 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4m5cul$e7e@naiad.grenet.fr> References: <4m4n65$7bu@naiad.grenet.fr> <4m50if$sje@hagen.amg.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit matthias@amg.de (Matthias Schuerhoff) wrote: >> hello, >> >> i have problems with key equivalents in buttons >Hello, > >only an idea: Are you sure that all of your panels are set to the class >"Panel" in IB inspector? Yes, they are Panels all right (I know about this one!). ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
From: monty@tbyte.com (Monty Zukowski) Newsgroups: comp.sys.next.programmer,gnu.emacs.help Subject: compiling emacs on NEXTSTEP 3.3 HPPA? Date: 30 Apr 1996 15:24:45 GMT Organization: Trilobyte, Inc. Message-ID: <4m5bbt$sb6@marvin.tbyte.com> I'm trying to compile emacs 19.30 on NEXTSTEP 3.3 HPPA with X windows support from the beta Xnext shareware. Anyway, at the end when it does the ./temacs thing it runs for 10 minutes and then has a segmentation fault. The PROBLEMS doc says it could be alloca not working right. I'm pretty sure that I didn't run out of swap space because my swapfile didn't get too huge (180 Meg). Has anybody solved this problem? I configured with m68k-next-nextstep since it didn't know what to do with the default machine it came up with (hppa-next-nextstep3). Thanks for any help, Monty monty@tbyte.com
From: eugene@raddi.uah.ualberta.ca (Eugene Mah) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: compiling pgp263is on intel 3.3 w. patch Date: 30 Apr 1996 15:56:17 GMT Organization: University of Alberta, Edmonton, Canada Message-ID: <4m5d71$o0u@pulp.ucs.ualberta.ca> References: <4m4p9m$30r@rrzs3.uni-regensburg.de> In <4m4p9m$30r@rrzs3.uni-regensburg.de> Wolfgang Roeckelein wrote: > Hi, > > has anybody succeeded in compiling pgp263is on NS intel 3.3 with the > patch. > > The linker keeps complaining about ld: Undefined symbols: _strdup > > Please help. > > Thank you very much in advance, > > Wolfgang > There's a patch for this that should eliminate this problem. It should be available from wherever pgp 2.6.3i is at. The patched version calls itself pgp 2.6.3ia. -- ---------------------------------------------------------------------- Eugene Mah eugene@raddi.uah.ualberta.ca (NeXTmail) Systems Administrator "For I am a Bear of Very Little Department of Radiology Brain, and long words Bother University of Alberta Hospitals me." Winnie the Pooh Edmonton, Alberta, Canada http://raddi.uah.ualberta.ca/~eugene/ PGP KeyID = 1F9779FD or E37A1591 O- ----------------------------------------------------------------------
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Setting TextField's Backround to Transparent Date: Tue, 30 Apr 1996 11:26:41 +0100 Organization: Cedar Systems Distribution: world Message-ID: <danD2BAhrehxEw3Q@cedar.co.uk> References: <4lpgvf$99u@hermes.is.co.za> MIME-Version: 1.0 In article <4lpgvf$99u@hermes.is.co.za>, gvandyk@icon.co.za writes >Hi, > >I have a TextField in one of my Apps. It's background needs to be in >one of 2 states either transparent or it must be Red. > >The problem I am experiencing is that when I have changed the >background color to red I cannot make it transparent again. > >My code looks something like this: > >- makeRed >{ > [aTextField setBordered:YES]; > [aTextField setBackgroundColor:NX_COLORRED]; > return self; >} > >- makeTransparent >{ > [aTextField setBordered:NO]; > [aTextField setBackgroundTransparent:YES]; > return self; >} > >I call makeRed to set the field red, and then makeTransparent to make >it transparent. What happens is that the border does disappear but the >background is still red. > >Any help in this regard would be appreciated. > Have you tried using [aTextField setBackgroundColor:NX_COLORCLEAR]; ? Paul ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: Ravi Mendis <lady0098@sable.ox.ac.uk> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer Subject: Re: Does WebObjects have Transaction Integrity? Date: Tue, 30 Apr 1996 14:43:08 +0100 Organization: Oxford University Message-ID: <Pine.OSF.3.91.960430133338.31260B-100000@sable.ox.ac.uk> References: <Pine.OSF.3.91.960426122740.23197A-100000@sable.ox.ac.uk> <4lr6c4$gsc@news.next.com> <Pine.OSF.3.91.960427191111.16883B-100000@sable.ox.ac.uk> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <Pine.OSF.3.91.960427191111.16883B-100000@sable.ox.ac.uk> On Sat, 27 Apr 1996, Ravi Mendis wrote: > Would [transaction integrity] not be 'straight-forward' to implement > within the current WebObject Framework? > (e.g: via TP monitored WebAdaptors?) Actually, monitoring a WebObject application's [http] transactions via a TP-monitored WebAdaptor, is equivalent to "logging" those http requests *received* (and those responded to), without having a clue as to which requests went astray over the 'net. [For database Web applications, this is roughly equivalent to TP-monitoring database transactions via a TP EOFAdaptor (like a CICS/DB2 EOF adaptor) : a far more efficient use of TP monitoring] What it *does not do* is guarantee that [http request/response] transactions are always delivered intact. So although it may monitor [http] transactions, it does not provide transaction integrity for WebObject applications. If you really want 'bullet-proof' transaction messaging over the Web, the link [in the chain of communication] to be tightened is its weakest: the 'Weblink'. It would be simplest to deploy your WebObject applications over Web servers and browsers implementing http on a TP API like CICS or MQI: it could be as simple as using something like the "shttp" protocol. As this solution is *external* to the WOFramework, NeXT need not provide it themselves: they would be infringing on Web browser/server territory. It is really up to one of the Web browser/server providers to implement!? ravi
Newsgroups: comp.sys.next.programmer,comp.sys.next.misc From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: How can I add undo to my program? Message-ID: <Dqovtn.8E6@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College Date: Tue, 30 Apr 1996 19:08:10 GMT How can I add undo to my program? Even better, how about a multi-level undo? It seems like a nearly impossible task of getting undo to work when there are so many different things to keep track of that can be undone. Especially in a text object! I remember a few years ago seeing something that someone had done which had some sort of undo manager, and all you had to do was add some kine of undo method to your objects to get it to work. I'm looking for that, or something similar. Thanks! -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: jonas@eunet.si (Jonas Znidarsic) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: compiling pgp263is on intel 3.3 w. patch Date: 30 Apr 1996 21:59:47 GMT Organization: Academic and Research Network of Slovenia Message-ID: <4m62gk$5qg@cmir.arnes.si> References: <4m4p9m$30r@rrzs3.uni-regensburg.de> <4m5d71$o0u@pulp.ucs.ualberta.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: eugene@raddi.uah.ualberta.ca In <4m5d71$o0u@pulp.ucs.ualberta.ca> Eugene Mah wrote: > There's a patch for this that should eliminate this problem. It should be > available from wherever pgp 2.6.3i is at. The patched version > calls itself pgp 2.6.3ia. I'm kinda embarrased to admitt... I don't have a compiler :( Could somebody mail me binary PGP? I would love to have it... Next mail accepted. -- *** Who is General Failure and why is he reading my disk? *** Jonas Znidarsic http://www.jonas.eunet.si
Newsgroups: comp.sys.next.misc,comp.sys.next.programmer From: jrhoads@netcom.com (Jason Rhoads) Subject: FS:QuickBase SQL Database for NeXTSTEP Message-ID: <jrhoadsDqp2un.JM7@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Tue, 30 Apr 1996 21:39:59 GMT Sender: jrhoads@netcom3.netcom.com I purchased a copy of QuickBase for a project I was working on. I no longer need the software, so its for sale! QuickBase 4.1h - 3 User License $695 OBO (Retail $895) Supports EOF. Includes disks and manual. I will pay shipping to anywhere in the continental US. For information about QuickBase check out http://www.quickbase.com Please send all inquiries to me at jrhoads@netcom.com. Thanks!
From: adachi@xnet.com (Kris Kasalo) Newsgroups: comp.sys.next.programmer Subject: HAIR LOSS?..JAPAN'S LEADING TREATMENT MAKES MINOXIDIL OBSOLETE... Date: 1 May 1996 09:05:10 GMT Organization: XNet - A Full Service Internet Provider - (708) 983-6064 Message-ID: <4m79g6$12j@flood.xnet.com> Now available in the U.S. Pentadecanoic Acid Glyceride and FA-BAO. World's leading treatments for thinning hair from Japan. 2.5X more effective than minoxidil. Featured on CNN, NEWSWEEK, New York Times. Doctor recommended. For free info, please call 1-800-555-8655
From: randy@is.com (Randy Marchessault) Newsgroups: comp.sys.next.programmer Subject: Employment opportunity Date: 1 May 1996 14:25:50 GMT Organization: Integrity Solutions, Inc. Distribution: world Message-ID: <4m7s9e$qnm@medusa.is.com> Keywords: NEXTSTEP, OpenStep, RDBMS, Oracle Object-Oriented Developer Young dynamic company with strong growth opportunities seeks Object-Oriented Developer to work in creative high-tech environment. Candidates should have excellent written and verbal communication skills and knowledge in one or more of the following languages: Smalltalk, Objective-C, and Java; and familiarity with OO methodologies. Oracle and/or Sybase development experience a definite plus. Project management experience also a plus. Must have strong motivation and self management skills. Excellent compensation. Please send email to randy@is.com. -- Randy Marchessault V.P., Integrity Solutions, Inc. 23 Empire Drive St. Paul, MN 55103 Voice: (612)-223-8474 X207 Fax: (612)-223-8481
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: How can I add undo to my program? Date: 1 May 1996 15:33:09 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4m807l$cn4@lion.embl-heidelberg.de> References: <Dqovtn.8E6@nvc.cc.ca.us> In article <Dqovtn.8E6@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) writes: > > How can I add undo to my program? Even better, how about a > multi-level undo? > > It seems like a nearly impossible task of getting undo to work when > there are so many different things to keep track of that can be > undone. Especially in a text object! > > I remember a few years ago seeing something that someone had done > which had some sort of undo manager, and all you had to do was add > some kine of undo method to your objects to get it to work. Perhaps it will be useful to read the "command pattern" in E.Gamma's et al Design Patterns. It's not exactly what you need, but it can help you to see how such a system has to be designed and used. -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
Date: 1 May 1996 12:12:46 EST Newsgroups: comp.sys.next.programmer Message-ID: <cancel.4m79g6$12j@flood.xnet.com> Control: cancel <4m79g6$12j@flood.xnet.com> From: clewis@ferret.ocunix.on.ca Sender: adachi@xnet.com (Kris Kasalo) Subject: cmsg cancel <4m79g6$12j@flood.xnet.com> EMP/ECP (aka SPAM) cancelled by clewis@ferret.ocunix.on.ca. See news.admin.net-abuse.announce, report 19960501.08 for further details
From: Paul Desrochers <paul@planon.qc.ca> Newsgroups: qc.jobs,qc.jobs.wanted,comp.sys.next.programmer,fr.comp.sys.next Subject: Offre d'emploi - Job offer <Programmeur analyste> Date: Wed, 01 May 1996 09:37:35 -0400 Organization: Planon Telexpertise, Inc. Boucherville QC Canada Message-ID: <3187691F.4056@planon.qc.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Planon Telexpertise developpe des systemes informatiques de controle en temps reel pour le domaine de la television. Les systemes concus par l'entreprise utilisent des technologies par objets. Nous sommes actuellement a la recherche de plusieurs analystes-programmeurs. Le defi: Le programmeur aura la responsabilite de developper et tester des applications sous NEXTSTEP (Objective C). Le candidat possede: DEC ou Bacc. en informatique ou experience pertinente Minimum 3 ans d'experience en programmation Minimum d'un an d'experience en langage C Connaissance des technologies orientees objets. S.v.p. envoyer votre c.v. a : paul@planon.qc.ca Planon Telexperise Inc. 1370 Joliot Curie Suite 708 Boucherville, Quebec Tel: (514) 449-6481 Fax: (514) 449-6484 Planon Telexpertise develops real-time computer systems for broadcast television market. Our systems are based on object-oriented technologies (NextStep). We are in the process of hiring several analyst-programmers. The candidat must at least: have a College or Graduate degree in computer science or strong experience; 3 years experiences in programming; 1 year of C language experience; Being familiar with object-oriented technologies. Please send your resume to: paul@planon.qc.ca Planon Telexperise Inc. 1370 Joliot Curie Suite 708 Boucherville, Quebec Tel: (514) 449-6481 Fax: (514) 449-6484
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside o Date: 1 May 1996 18:31:22 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4m8alq$p6m@netty.york.ac.uk> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> <4lreut$hf8@news.next.com> On 26 Apr 1996 21:24:45 GMT, Eric Noyau <Eric_Noyau@next.com> wrote: > We where talking about looking into the instances variables of a private > class, subclass of a class cluster, for debugging purpose. My point is the > only state you may need from these objects is the public state, available > thru method calls on the abstract superclass. this is not true when debugging from a core dump. rog.
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside o Date: 02 May 1996 01:53:18 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7u3y0kl1t.fsf@tom.es.ele.tue.nl> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> <4lreut$hf8@news.next.com> <4m8alq$p6m@netty.york.ac.uk> In-reply-to: rog@ohm.york.ac.uk's message of 1 May 1996 18:31:22 GMT In article <4m8alq$p6m@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) writes: this is not true when debugging from a core dump. But that is only because the kernel is too stupid to (re-) create a process from a core dump :-) --Tiggr
Newsgroups: comp.sys.next.misc,comp.sys.next.programmer From: "Eric K. Ringger" <ringger@cs.rochester.edu> Subject: Re: How can I add undo to my program? In-Reply-To: Your message of "Tue, 30 Apr 1996 19:08:10 GMT." <Dqovtn.8E6@nvc.cc.ca.us> Message-ID: <199605020233.WAA09449@slate.cs.rochester.edu> Sender: ringger@cs.rochester.edu (Eric K. Ringger) Cc: comp.sys.next.misc, comp.sys.next.programmer Organization: University of Rochester Computer Science Dept Date: Wed, 01 May 1996 22:33:46 -0400 Hi. Chris Osborn wrote: >How can I add undo to my program? Even better, how about a >multi-level undo? > >It seems like a nearly impossible task of getting undo to work when >there are so many different things to keep track of that can be >undone. Especially in a text object! > >I remember a few years ago seeing something that someone had done >which had some sort of undo manager, and all you had to do was add >some kine of undo method to your objects to get it to work. 'Sounds like Draw.app ! >I'm looking for that, or something similar. [...] Check out the following files in the NEXTSTEP Developer distribution: /NextDeveloper/Examples/AppKit/Draw/undo.subproj /NextDeveloper/Examples/AppKit/Draw/UndoDoc.rtf /NextDeveloper/Examples/AppKit/Draw/UndoREADME.rtf /NextDeveloper/Examples/AppKit/Draw/textUndo.subproj /NextDeveloper/Examples/AppKit/Draw/graphicsUndo.subproj Good luck. --Eric --- Eric K. Ringger mailto:ringger@cs.rochester.edu Dept. of Computer Science Office: +1-716-275-0922; Lab: +1-716-275-5377 University of Rochester Fax: +1-716-461-2018 Rochester NY 14627-0226 http://www.cs.rochester.edu/u/ringger/ ||||| | | | | | | | | | | | | | | | | |||||
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Re: How can I add undo to my program? Date: 2 May 1996 05:03:38 GMT Organization: Digital Fix Development Message-ID: <4m9fna$n3@digifix.digifix.com> References: <199605020233.WAA09449@slate.cs.rochester.edu> In-Reply-To: <199605020233.WAA09449@slate.cs.rochester.edu> On 05/01/96, "Eric K. Ringger" wrote: >Hi. > >Chris Osborn wrote: >>How can I add undo to my program? Even better, how about a >>multi-level undo? >> >>It seems like a nearly impossible task of getting undo to work when >>there are so many different things to keep track of that can be >>undone. Especially in a text object! >> >>I remember a few years ago seeing something that someone had done >>which had some sort of undo manager, and all you had to do was add >>some kine of undo method to your objects to get it to work. > >'Sounds like Draw.app ! > >>I'm looking for that, or something similar. >[...] > >Check out the following files in the NEXTSTEP Developer distribution: > I sent this to Chris earlier as well, but it might be of general interest too. In the Summer 1994 issue of NXApp, Jeff Martin wrote an article on message based Undo. Its on ftp.next.com or via next answers.. 1998_Message_Based_Undo_1994Summer.rtfd -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati From: aq004@torfree.net (Steve Guidi) Subject: Need Help!! Tricky problem involving user input! Message-ID: <DqrHw8.GsC.0.sheppard@torfree.net> Organization: Toronto Free-Net Date: Thu, 2 May 1996 05:00:08 GMT on,fras.text.programmieren,mcnc.programmers,maus.sys.mac.*,no.slekt.programmer,relcom.comp.lang.*,pl.comp.programming.umn.cs.lang.*,ucd.cs.programming,uw.mac-users,york.comp.*,z-netz.rechner.amiga.programmieren,z-netz.rechner.ibm.programmieren,z-netz.rech ner.apple.programmieren,z-netz.rechner.promgrammieren,york.cs.*,uw.cs.*,a.bsu.programming,adass.iraf.programming,comp.programming.*,zer.lanex.programmieren X-Newsreader: TIN [version 1.2 PL2] [ Article crossposted from comp.programming ] [ Author was Steve Guidi ] [ Posted on Thu, 2 May 1996 03:58:56 GMT ] Please, please somebody help me with this... I am working on a school project which basically asks a user to enter a mathematical formula and then the program will give the solution to that formula. For example: if the formula is (R*S)/(R+S) where R and S are predefined values, the comptuer will calculate that using the values and give a solution. The problem I have encountered is that the USER has to define this formula via an input/output routine. From my knowledge of datatypes, there is no type that will allow the computer to simply use the input as if it were a mathematical formula in the source code itself. Taking the above formula used, the computer has to realize that R and S are variables in the formula and that the brackets and operateors are operators and not text. Also, I have to develop this algorithm to use predefined functions such as SIN, SQRT, ABS, LOG, etc.. The formula can take any level of complexity also (there could be hundreds of brackets in there!) I have consulted with a teacher at my school and he mentioned that the trick must lie with the way spreadsheet programs work. Say I were to enter a formula = A3 * A5 / ABS (SUM E2:E7), the spreadsheet program would work with that forumala with the values in the respective cells. This is VERY similar to what I have to do except I am lost and don't know how to do it! I plan to use Think Pascal for Macintosh as I am pretty familiar with the software. Any help on the matter would be greatly appreciated. My deadline is coming up real soon so please email me with some advice ASAP! Thanks for all your help in advance.. Email me ... aq004@torfree.net PS: I know they teach this stuff in University.. I'm still a highschool student =) -- | hacker@wbb.com aq004@torfree.net steve.guidi@canada.tor250.org | | Steve Guidi - 1:250/620 (Fido) 1/Steve Guidi (PCPN) | -- | hacker@wbb.com aq004@torfree.net steve.guidi@canada.tor250.org | | Steve Guidi - 1:250/620 (Fido) 1/Steve Guidi (PCPN) |
From: barnhoorn@nlev00 () Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati Subject: Re: Need Help!! Tricky problem involving user input! Date: 2 May 1996 07:53:26 GMT Organization: Alcatel Bell Distribution: world Message-ID: <4m9plm$s6a@btmpjg.god.bel.alcatel.be> References: <DqrHw8.GsC.0.sheppard@torfree.net> In article <DqrHw8.GsC.0.sheppard@torfree.net>, aq004@torfree.net (Steve Guidi) writes: > >Please, please somebody help me with this... I am working on a school >project which basically asks a user to enter a mathematical formula and >then the program will give the solution to that formula. > >For example: if the formula is (R*S)/(R+S) where R and S are predefined >values, the comptuer will calculate that using the values and give a >solution. > Don't know anything about the way spreadsheets work, but your problem is easy to solve when you are going to use a recursive function. Scan from left to right - as soon as an opening bracket pops up, call your function recursively. I guess that is enough information for you to solve this tiny little problem. > >Email me ... aq004@torfree.net > >PS: I know they teach this stuff in University.. I'm still a highschool >student =) > yeah, sure. You don't need a university degree to solve your problem!! -- --------------------------------------------------------------------------- Jaco Barnhoorn barnie@xs4all.nl Software Test Engineer barnhoorn%nlev00@btmv56.se.bel.alcatel.be Alcatel Telecom Systems Rijswijk, The Netherlands ---------------------------------------------------------------------------
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4m79g6$12j@flood.xnet.com> Control: cancel <4m79g6$12j@flood.xnet.com> Date: 1 May 1996 17:21:48 GMT Organization: Turbocat's Development, Germany Message-ID: <4m86jc$2ei@turbocat.snafu.de> cancel
From: mnb20@cam.ac.uk (Mark Baker) Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati Subject: Re: Need Help!! Tricky problem involving user input! Date: 2 May 1996 09:30:30 GMT Organization: University of Cambridge, England Message-ID: <4m9vbm$3cn@lyra.csx.cam.ac.uk> References: <DqrHw8.GsC.0.sheppard@torfree.net> In article <DqrHw8.GsC.0.sheppard@torfree.net>, aq004@torfree.net (Steve Guidi) writes: >For example: if the formula is (R*S)/(R+S) where R and S are predefined >values, the comptuer will calculate that using the values and give a >solution. > >The problem I have encountered is that the USER has to define this >formula via an input/output routine. From my knowledge of datatypes, >there is no type that will allow the computer to simply use the input as >if it were a mathematical formula in the source code itself. No. In a compiled language, any expressions such as that in your source code are read at compile time and converted into an assembly language program to evaluate the expression. There is no expression parser in the output of the compiler, so no simple way to pass a user entered expression to it. You'll have to do the parsing yourself. What you need to do is to scan the line looking for terms to add up. Each term is then passed to a routine that scans through it looking for terms to multiply or divide, then each term of that is passed to a routine that scans through it looking for terms to raise to a power. Each of those terms is then either a number, or a function such as sin or cos, or an expression in brackets, in which case you pass the contents of the brackets to the first routine. It sounds complicated, but it's fairly easy, particularly if you've got good string functions. >I have consulted with a teacher at my school and he mentioned that the >trick must lie with the way spreadsheet programs work. Say I were to >enter a formula = A3 * A5 / ABS (SUM E2:E7), the spreadsheet program >would work with that forumala with the values in the respective cells. >This is VERY similar to what I have to do except I am lost and don't know >how to do it! Well yes, spreadsheets also have to have an expression evaluator of course. But the difficult bits about writing a spreadsheet are storing the data in an efficient way and doing natural order evaluation, neither of which do you have to worry about.
From: marcel@cs.tu-berlin.de (Marcel Weiher) Newsgroups: comp.sys.next.programmer Subject: Nested Functions? Date: 2 May 1996 10:41:01 GMT Organization: Technical University of Berlin, Germany Distribution: world Message-ID: <4ma3ft$gp6@news.cs.tu-berlin.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Summary: How do I get nested functions to work using NeXT CC? Keywords: compiler,next,cc,gcc,gnu cc I can't seem to get nested functions to work with NeXT's CC, even though the doc says that they are supported. This is on a 3.3 with patch installed and: NeXT Computer, Inc. version cc-437.2.6, gcc version 2.5.8 I tried the 'nested.c' sample from the libobjects package, as well as some other tests, which all fail with syntax errors. Please tell me where I'm being dense. Thanks, Marcel
From: marco@sentenext1.epfl.ch (Marco Scheurer) Newsgroups: comp.sys.next.programmer Subject: Re: How can I add undo to my program? Date: 2 May 1996 11:30:38 GMT Organization: Ecole Polytechnique Federale de Lausanne Message-ID: <4ma6cu$nqg@info.epfl.ch> References: <199605020233.WAA09449@slate.cs.rochester.edu> <4m9fna$n3@digifix.digifix.com> sanguish@digifix.com (Scott Anguish) wrote: > On 05/01/96, "Eric K. Ringger" wrote: > >Hi. > > > >Chris Osborn wrote: > >>How can I add undo to my program? Even better, how about a > >>multi-level undo? > Its on ftp.next.com or via next answers.. > 1998_Message_Based_Undo_1994Summer.rtfd This example is great, IMHO the way to go. Try to do this in C++ ! --- Marco Scheurer Sen:te Parc Scientifique EPFL -- CH-1015 Lausanne -- Switzerland marco@sente.epfl.ch (NeXT, MIME mail welcome)
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 2 May 1996 14:12:29 GMT Organization: Norden 1 Communications Message-ID: <4mafsd$gih@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF-----A plus Sybase or Oracle Career Position Exceptional Opportunity Relocation Assistance To Be Considered----Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: How can I add undo to my program? Date: 2 May 1996 14:31:36 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4mah08$15u@lion.embl-heidelberg.de> References: <4ma6cu$nqg@info.epfl.ch> In article <4ma6cu$nqg@info.epfl.ch> marco@sentenext1.epfl.ch (Marco Scheurer) writes: > sanguish@digifix.com (Scott Anguish) wrote: > > On 05/01/96, "Eric K. Ringger" wrote: > > >Hi. > > > > > >Chris Osborn wrote: > > >>How can I add undo to my program? Even better, how about a > > >>multi-level undo? > > > Its on ftp.next.com or via next answers.. > > > 1998_Message_Based_Undo_1994Summer.rtfd > > This example is great, IMHO the way to go. Try to do this in C++ ! ;-))) Anyway, this is not all you can get. Using slightly more general approach it is possible (although not very trivially) to do much more: - a real macro recorder; - brunches of undo/redo; - log file for events (so you can restore from crash); - if you put everything in a shared undo manager (accessible with DO) you can also remotely teach somebody how to use an app (temporary he can connect his sgared undo manager to your manager) and both application are executed in sync; Happy hacking -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: wolfgang@wi.WHU-Koblenz.de (Wolfgang Roeckelein) Newsgroups: comp.sys.next.programmer Subject: Re: compiling pgp263is on intel 3.3 w. patch Date: 2 May 1996 14:47:12 GMT Organization: University of Regensburg, Germany Message-ID: <4mahtg$feh@rrzs3.uni-regensburg.de> References: <4m5d71$o0u@pulp.ucs.ualberta.ca> In article <4m5d71$o0u@pulp.ucs.ualberta.ca> eugene@raddi.uah.ualberta.ca (Eugene Mah) writes: > In <4m4p9m$30r@rrzs3.uni-regensburg.de> Wolfgang Roeckelein wrote: > > The linker keeps complaining about ld: Undefined symbols: _strdup > > There's a patch for this that should eliminate this problem. It should be > available from wherever pgp 2.6.3i is at. The patched version > calls itself pgp 2.6.3ia. Where is this version available? I found only a special atari version called 263ia. Wolfgang -- Dipl.-Wirtsch.Inf. Voice: +49 941 943 2998 Wolfgang Roeckelein Fax: +49 941 943 4986 Uni Regensburg E-Mail: roeckelein@wi.whu-koblenz.de Universitaetsstr. 31 Wolfgang.Roeckelein@wiwi.uni-regensburg.de D-93053 Regensburg (MIME and NeXTmail ok) Germany WWW: http://www.whu-koblenz.de/~wolfgang/
From: eric@skatter.USask.Ca Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati Subject: Re: Need Help!! Tricky problem involving user input! Date: 2 May 1996 17:21:34 GMT Organization: University of Saskatchewan Message-ID: <4maquu$40j@tribune.usask.ca> References: <DqrHw8.GsC.0.sheppard@torfree.net> <4m9vbm$3cn@lyra.csx.cam.ac.uk> mnb20@cam.ac.uk (Mark Baker) wrote: >In article <DqrHw8.GsC.0.sheppard@torfree.net>, > aq004@torfree.net (Steve Guidi) writes: > >>For example: if the formula is (R*S)/(R+S) where R and S are predefined >>values, the comptuer will calculate that using the values and give a >>solution. >> >>The problem I have encountered is that the USER has to define this >>formula via an input/output routine. From my knowledge of datatypes, >>there is no type that will allow the computer to simply use the input as >>if it were a mathematical formula in the source code itself. > >No. In a compiled language, any expressions such as that in your source >code are read at compile time and converted into an assembly language >program to evaluate the expression. There is no expression parser in >the output of the compiler, so no simple way to pass a user entered >expression to it. You'll have to do the parsing yourself. > >What you need to do is to scan the line looking for terms to add up. >Each term is then passed to a routine that scans through it looking >for terms to multiply or divide, then each term of that is passed to >a routine that scans through it looking for terms to raise to a power. >Each of those terms is then either a number, or a function such as >sin or cos, or an expression in brackets, in which case you pass the >contents of the brackets to the first routine. > >It sounds complicated, but it's fairly easy, particularly if you've >got good string functions. And even easier if you use LEX and YACC (or better, the more modern FLEX and BISON) to do the lexical analysis and parsing. Setting up a simple expression interpreter using these tools is very easy. The YACC manual even used to come with a little example that did just that. -- Eric Norum eric@skatter.usask.ca Saskatchewan Accelerator Laboratory Phone: (306) 966-6308 University of Saskatchewan FAX: (306) 966-6058 Saskatoon, Canada. NeXTMail accepted.
From: "Ingo Feulner" <ifeulner@cube.de> Newsgroups: comp.sys.next.programmer Subject: Subclass of NXBrowserCell works with normal delegate but not with (very) lazy Date: 2 May 1996 16:50:00 GMT Organization: Cube Informationssysteme GmbH Message-ID: <4map3o$ddk@cubenx.cube.de> Can anybody imagine what kind of problem this is? The only methods which are overriden are - setFont:fontObj; - calcCellSize:(NXSize *)theSize inRect:(const NXRect *)aRect; - drawInside:(const NXRect *)cellFrame inView:controlView; - highlight:(const NXRect *)cellFrame inView:controlView lit:(BOOL)flag; so nothing special. but the delegate method browser:loadCell:atRow:inColumn:column gets the first time called if the user clicks on the browser cell (better to say: into the space where the Cell should be...) So whats wrong? (It works perfect with the standard NXBrowserCell, but not with my subclass). - Ingo. -- Ingo Feulner <ifeulner@cube.de>, Cube Informationssysteme GmbH Stuttgart.
From: Douglas Joshua Behnke <behnkedj@cs.rose-hulman.edu> Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati Subject: Re: Need Help!! Tricky problem involving user input! Date: Thu, 02 May 1996 12:03:09 -0700 Organization: Trusted Information Systems, Inc. Message-ID: <318906ED.41C67EA6@cs.rose-hulman.edu> References: <DqrHw8.GsC.0.sheppard@torfree.net> <4m9vbm$3cn@lyra.csx.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Mark Baker wrote: > > In article <DqrHw8.GsC.0.sheppard@torfree.net>, > aq004@torfree.net (Steve Guidi) writes: > > >For example: if the formula is (R*S)/(R+S) where R and S are predefined > >values, the comptuer will calculate that using the values and give a > >solution. > > > >The problem I have encountered is that the USER has to define this > >formula via an input/output routine. From my knowledge of datatypes, > >there is no type that will allow the computer to simply use the input as > >if it were a mathematical formula in the source code itself. > > No. In a compiled language, any expressions such as that in your source > code are read at compile time and converted into an assembly language > program to evaluate the expression. There is no expression parser in > the output of the compiler, so no simple way to pass a user entered > expression to it. You'll have to do the parsing yourself. > You could always write the function out to a text file, wrapped in a suitable function declaration like: double userExpression() { return <insert_user_function_here>; } Then all you have to do is run the compiler, dynamically link the object file your executable, and call the function. It might be tricky to get the variables passed in correctly, though. They may need to be globals. ;-) Cheers, -Josh Behnke behnkedj@cs.rose-hulman.edu
From: "Timothy J. Luoma" <luomat@capitalist.princeton.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: How can I add undo to my program? Date: Wed, 1 May 1996 15:57:01 -0400 Organization: Princeton University Message-ID: <Pine.NXT.3.93.960501155643.266D-100000@capitalist.princeton.edu> References: <Dqovtn.8E6@nvc.cc.ca.us> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII To: Chris Osborn <fozztexx@nvc.cc.ca.us> In-Reply-To: <Dqovtn.8E6@nvc.cc.ca.us> http://www.next.com/NeXTanswers/HTMLFiles/1998.htmld/1998.html might be what you are looking for..... might be... TjL -- Timothy J. Luoma <luomat@capitalist.princeton.edu> USENET: Please CC this address on replies to my posts, things vanish fast. NeXT Info available via email! Send message with SUBJECT 'send info' >>>> Please: No NeXTMail, use MIME for attachments <<<<
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside o Date: 2 May 1996 19:36:13 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4mb2rd$kjc@netty.york.ac.uk> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> <4lreut$hf8@news.next.com> <x7u3y0kl1t.fsf@tom.es.ele.tue.nl> On 02 May 1996 01:53:18 +0200, Pieter Schoenmakers <tiggr@es.ele.tue.nl> wrote: > In article <4m8alq$p6m@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) writes: > > this is not true when debugging from a core dump. > > But that is only because the kernel is too stupid to (re-) create a > process from a core dump :-) --Tiggr which doesn't invalidate the point. i'm sure there are other occasions when it is not possible to invoke method calls in a process being debugged (e.g. stack pointer corruption, corruption of other objects, etc) as was said earlier - this isn't programming, it's debugging. there's a big difference. cheers, rog.
From: alanf@izzy.net Newsgroups: comp.sys.next.programmer Subject: mmap or shmget Date: 3 May 1996 02:29:30 GMT Organization: Isthmus Corporation, Inc. Message-ID: <4mbr2a$2vq@izzy4.izzy.net> Keywords: backup, shared memory, amanda I'm trying to compile the University of Maryland AMANDA backup software on NEXTSTEP. The client will compile, but the server requires additional libraries, and fails. The error I get is ld: Undefined symbols: _munmap which I interpret to mean it can't find the mmap() or munmap() calls for shared memory management. I can't find any reference to mmap (Berkeley) or shmget (Sys V) in Digital Librarian... Any thoughts? Thanks, Alan Frabutt mailto:alanf@izzy.net
From: pspriet@pegase (Paul SPRIET) Newsgroups: comp.sys.next.programmer Subject: Pb with NXTableView and EOF Date: 3 May 1996 09:27:40 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4mcjid$nq8@ai.alienor.fr> Hi all, I am using an NXTableview with EOF and I would like to modify a row of a colomn. I 've tried : if([[anEOController allObjects]count]){ for(i=0;i<[[anEOController allObjects] count];i++){ unEssai = [[anEOController allObjects]objectAtIndex:i]; if([[unEssai ETAT]intValue] == 0){ id <NXTableVectors > aVector; NSString *aValue ; aVector = [nxTableEssais columnAt:0]; aValue = [NSString setStringValue:"A"]; [[nxTableEssais dataSource] setValueFor:[aVector identifier] at:i from:aValue]; } } } But I did not succeded. Thanks to help me. ANY REPLY IS TO SENT TO : Valerie Moukdarath email : vmoukdar@x-lan.alienor.fr -- +-------------------------------------------------------------+ | Paul SPRIET | Email/NeXTmail: pspriet@x-lan.alienor.fr | | X&LAN - FRANCE | Tel(+33)56 51 04 89 Fax(+33)56 51 05 31 | +-------------------------------------------------------------+
From: Jyakobi@visinet.ca Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.novell Subject: CyberChat // CyberChat // CyberChat // CyberChat // CyberChat Date: 2 May 1996 22:24:51 GMT Organization: VisiNet Internet Services Distribution: inet Message-ID: <4mbcnj$ol@kenya.visinet.ca> Tired experiencing untimely delays when chatting on the Internet? Is your curiosty getting the better of you? Here's your chance to do something about it! No Keyboards, No Monitors, and No Modems! Call CyberChat your telephone connection to people you meet on the net! 1 - 809 - 474 - 4437 Only long distance charges apply for this call. Must be 18 years of age or older. This service brought to you by Top Communications. Call CyberChat now!
From: Lars Konieczny Newsgroups: comp.sys.next.programmer Subject: Q: How to use foundation Date: 3 May 1996 15:18:12 GMT Organization: Rechenzentrum der Universitaet Freiburg, Germany Message-ID: <4md83k$onr@n.ruf.uni-freiburg.de> Hi, sorry if this is a stupid question, but how can I make the compiler be aware of the foundationKit classes? When the correct header is imported in the definition file (#import <foundation/NSObject.h>), the compiler exits with: ld: Undefined symbols: .objc_class_name_NSObject *** Exit 1 Another question: can I make IB unparse subclasses of NSObjects correctly? After subclassing NSObject, IB puts #import <appkit/appkit.h> #import "NSObject.h" into the class definition file obviously ignoring that NSObject is not part of the appkit. Thanx, Lars
From: nathan@cco.caltech.edu (Nathan Mates) Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati Subject: Re: Need Help!! Tricky problem involving user input! Date: 3 May 1996 17:20:27 GMT Organization: California Institute of Technology, Pasadena Message-ID: <4mdf8r$4ti@gap.cco.caltech.edu> References: <DqrHw8.GsC.0.sheppard@torfree.net> In article <DqrHw8.GsC.0.sheppard@torfree.net>, >Please, please somebody help me with this... I am working on a school >project which basically asks a user to enter a mathematical formula and >then the program will give the solution to that formula. Read a computer science book, but don't expect us to do your homework for you. Find a good book on recursive parsing-- the source code for that sort of thing is a page of code at most. Nathan Mates -- <*> Nathan Mates http://www.ugcs.caltech.edu/~nathan/ <*> # And he said to them, "To you has been given the secret of the # Kingdom of God, but for those outside, everything comes in # parables" Mark 4:11, NRSV
Date: 3 May 1996 17:22:04 GMT From: jem@xpat.com Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.novell Message-ID: <cancel.4mbcnj$ol@kenya.visinet.ca> Control: cancel <4mbcnj$ol@kenya.visinet.ca> Sender: Jyakobi@visinet.ca Subject: cmsg cancel <4mbcnj$ol@kenya.visinet.ca> EMP/ECP (aka SPAM) cancelled by jem@xpat.com. See news.admin.net-abuse.announce, report 19960503.21 for further details
From: bkmoore@uci.edu (Brian Kevin Moore) Newsgroups: comp.sys.next.programmer Subject: Learning Objective C Date: 4 May 1996 02:28:06 GMT Organization: University of California, Irvine Message-ID: <4mefbm$931@news.service.uci.edu> Hello All: I've been trying to learn the intricacies of NeXT STEP programming for some time now. I am an experienced C programmer, but moving to Objective C has been a little more difficult than I had at first anticipated. It has required me to move from linear-type code to a more parallel object oriented model. This transition is very difficult for me to make. I'm looking for any good resources on Objective C programming. I do not need a book on developing for the NeXT GUI. At least not yet anyway. I'm just looking for a good introduction to Objective C with a few simple examples. Once I master small text-based programs, I'll try to move on to using the GUI. I would like to build a strong foundation on Objective C so that I can write better code when I eventually start writing full-blown NeXT STEP apps. Thanks for any help, Brian K. Moore
From: svail@regression.next.com (Scott Vail) Newsgroups: comp.sys.next.programmer Subject: Re: Card with multi serial port under NeXTStep Date: 4 May 1996 04:18:44 GMT Organization: NeXT Software, Inc. Message-ID: <4melr4$2fi@news.next.com> References: <9604290827.AA00279@saem-ales.saem-ales.fr> In article <9604290827.AA00279@saem-ales.saem-ales.fr> writes: > > I need, for an application, 5 to 7 serials ports. The big question > is, if this card exists, how can I drive it under NeXTStep ?? > > Xavier Fabre. > > Assuming you are talking about 3.3 with the new ISASerialPort driver, you can drive as many serial ports are there are IRQ's and port-addresses available for each port. There are some limitations though, the serial ports must not share interrupts and you cannot set the serial ports to speeds greater than 57600 baud. Also, since you are planning on using more than two serial ports, it is a definite must that your serial ports use 16550 (or better) UARTs and be sure to get modem cables with hardware flow-control lines. Using a Quickpath Port-Folio I/O card, I once configured 4 serial ports and had them each transferring data at 57600 baud for several days without too many problems. Unfortunately, there were hardware bugs in this card, so I can't recommend it. Unfortunately, I haven't seen or used any other serial I/O cards, so I can't really make any recommendations. Sorry. --Scott
From: hnalgae@soback.kornet.nm.kr (SungJin Kang (kornet)) Newsgroups: comp.sys.next.programmer Subject: Reference about Input Manager? Date: 4 May 1996 12:24:45 GMT Organization: KORNET (Korea Telecom) Message-ID: <4mfiad$g6d@usenet.kornet.nm.kr> In NeXT Developer Korean Edition, there are reference about Input Manager. But it was writted in Japanese, so I can't read it. Do you know where can I get the reference about IM in English? Thank you for reading this. Hayan Nalgae hnalgae@soback.kornet.nm.kr
From: Jim_Brownfield@Radical.Com (Jim Brownfield) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Anybody using Objective-C with MS Visual C++? Date: 4 May 1996 14:03:08 GMT Organization: Radical System Solutions, Inc. Message-ID: <4mfo2s$m6m@radical1.radical.com> Does anyone have any experience mixing the gnu Objective-C (or Stepstone, etc.) with Microsoft's s Visual C++ environment? Does it even work, and are there any "gotchas?" -- Jim Brownfield (Jim_Brownfield@Radical.Com) NeXTmail/MIME accepted Radical System Solutions, Inc. (http://www.radical.com/) System/Network/Database Design, Development, Consulting rad i cal \'rad-i-kel\ n -- a basic principle: FOUNDATION
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Anybody using Objective-C with MS Visual C++? Followup-To: comp.lang.objective-c,comp.sys.next.programmer Date: 4 May 1996 16:13:08 GMT Organization: University of Ghent, Belgium Message-ID: <4mfvmk$5co@infoserv.rug.ac.be> References: <4mfo2s$m6m@radical1.radical.com> Jim Brownfield (Jim_Brownfield@Radical.Com) wrote: : Does anyone have any experience mixing the gnu Objective-C (or Stepstone, : etc.) with Microsoft's s Visual C++ environment? Does it even work, and : are there any "gotchas?" I have no experience with their non-UNIX versions myself, but Stepstone suggests Visual C++ for their Windows (16bit) and WindowsNT (32bit) Objective C precompilers. David Stes Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: stes@cage.rug.ac.be (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: The (questioned) privacy of private classes (was: Re: varibles declared outside o Date: 4 May 1996 16:29:40 GMT Organization: University of Ghent, Belgium Message-ID: <4mg0lk$5co@infoserv.rug.ac.be> References: <x7afzzmilx.fsf@tom.es.ele.tue.nl> <4lreut$hf8@news.next.com> <4m8alq$p6m@netty.york.ac.uk> On 26 Apr 1996 21:24:45 GMT, Eric Noyau <Eric_Noyau@next.com> wrote: > We where talking about looking into the instances variables of a private > class, subclass of a class cluster, for debugging purpose. My point is the > only state you may need from these objects is the public state, available > thru method calls on the abstract superclass. OK, but if these classes were really private, then you would not need to access them either by sending a message or by looking into their instance variables... Actually you would't even have to know that they exist, and they would be there merely as implementation of a single public class. So calling these subclasses of an abstract class private, merely because the header file was removed, is rather dubious, because the first thing that is being told in IntroFoundationKit is how to deal with them; if they were private, we wouldn't have to deal with them. David Stes Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer Subject: Re: Card with multi serial port under NeXTStep Date: 4 May 1996 15:59:50 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4mfutm$3ag@tallowcross.uni-frankfurt.de> References: <9604290827.AA00279@saem-ales.saem-ales.fr> <4melr4$2fi@news.next.com> svail@regression.next.com (Scott Vail) wrote: > There are some limitations though, the serial > ports must not share interrupts and you cannot set the serial ports to > speeds greater than 57600 baud. Well, you can. There are several serial cards available which multiply the speed set by the software with a factor of 2 or 4 by means of hardware. Thus, a (software) setting of 57600 will actually result in 230400 baud. It works perfectly with an ISDN ZyXEL for me. Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: gfa@komsys-pc-gf.ethz.ch (George Fankhauser) Newsgroups: comp.sys.next.programmer Subject: Re: Nested Functions? Date: 4 May 1996 10:55:12 -0700 Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4mg5m0$ju@komsys-pc-gf.ethz.ch> References: <4ma3ft$gp6@news.cs.tu-berlin.de> Keywords: compiler,next,cc,gcc,gnu cc > NeXT Computer, Inc. version cc-437.2.6, gcc version 2.5.8 > >I tried the 'nested.c' sample from the libobjects package, as well >as some other tests, which all fail with syntax errors. > >Please tell me where I'm being dense. the magic word is -traditional-cpp which enables nested procedures. I don't why this isn't the default behaviour... George -- George Fankhauser Swiss Federal Institute of Technology Phone/Fax +41 1 632 7017/1035 Computer Engineering and Networks http://www.tik.ee.ethz.ch/~gfa Laboratory, CH-8092 Zurich mailto:gfa@tik.ee.ethz.ch "Happyness is a journey, not a destination"
From: Johan R=?iso-8859-1?Q?=F6nnblom <JRo0605@ba.ssdn.skellefte.se>?= Newsgroups: alt.lang,alt.alt.lang.*,comp.sys.amiga.programmer,comp.sys.apple2.programmer,comp.sys.atari.programmer,comp.sys.mac.programmer.*,comp.sys.next.programmer,de.comp.lang.*,fjnet.programming,fj.sys.mac.programming,maus.lang.*,franscomm.programmati Subject: Re: Need Help!! Tricky problem involving user input! Date: Fri, 03 May 1996 13:52:49 -0700 Organization: Balderskolan Skelleftea Message-ID: <318A7221.18BA@ba.ssdn.skellefte.se> References: <DqrHw8.GsC.0.sheppard@torfree.net> <4m9vbm$3cn@lyra.csx.cam.ac.uk> <318906ED.41C67EA6@cs.rose-hulman.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable > > >For example: if the formula is (R*S)/(R+S) where R and S are = > > >predefined > > >values, the comptuer will calculate that using the values and give a > > >solution. > > > > > >The problem I have encountered is that the USER has to define this > > >formula via an input/output routine. From my knowledge of datatypes, > > >there is no type that will allow the computer to simply use the = > > >input as > > >if it were a mathematical formula in the source code itself. > You could always write the function out to a text file, wrapped in a = > suitable > function declaration like: > double userExpression() > { > return <insert_user_function_here>; > } > = > Then all you have to do is run the compiler, dynamically link the = > object file > your executable, and call the function. It might be tricky to get the = > variables > passed in correctly, though. They may need to be globals. That's an interesting idea.. but a bit clumsy to start the compiler just = for that. Or else you could write it as a scriptfile(batchfile) like = this: eval <userinputstring> then you could simply execute this file, and then use the returned value = as the aswer. Note that <eval> is the Amiga command for this kind of = operation, replace this with whatever appropriate for your system. Could = be a problem with a Mac though.. it must be possible in some way. /Johan R=F6nnblom, Team Amiga
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: Window not taking title Date: 4 May 1996 21:42:06 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4mgive$1ats@core.bard.edu> I am trying to set the title of a Window with setTitle:. When I check with [Window title] I get a null pointer. Here's how I'm doing it: - setName: (const char *)name { moduleName = malloc(strlen(name) + !); strcpy(moduleName,name); [myWindow setTitle:name] // [myWindow setTitle:moduleName]; //tried this return self; } name and moduleName are both getting the correct value. here's the rest: - initWithController:(StickMusic *)controller { name=[controller getName]; [self setName:name]; [nameField setStringValue:name]; return self; } from StickMusic: - (char *)getName { char *name; name = malloc(3); sprintf(name,"%i",++moduleCount); return name; } -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: mail@carecom.com (Bjoern B. Dorra) Newsgroups: comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer Subject: looking for a wiz Date: Sat, 04 May 1996 13:53:44 -0700 Organization: CareCom Message-ID: <mail-0405961353440001@carebrain.carecom.com> Howdy, I am looking for a wiz , Yes a real object knowledgeable person to lead my new companys engineering group. I¹ve got the customer but not the technical visionary person that can take this companys product line from the prototype to the finish line. Yes this is a startup, and we are based in San Francisco . Have you got what it takes? Send me your info by email to: mail@carecom.com Snail Mail to: CARECOM 1482 Page Street San Francisco, CA. 94117
Sender: jagooch@alf.UCCS.edu (John A. Gooch) From: jem@xpat.com Message-ID: <cancel.4mh07b$a7u@harpo.uccs.edu> Control: cancel <4mh07b$a7u@harpo.uccs.edu> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4mh07b$a7u@harpo.uccs.edu> Date: 5 May 1996 03:47:07 GMT Cancelled by jem@xpat.com. 831268027 MMFscan This posting is spam - thousands of these have been posted. This posting is also a chain-letter. Chain-letters via Usenet may illegal. See: http://www.usps.gov/websites/depart/inspect/chainlet.htm Original Subject was: $$$Need Money? Look Here!!$$$
From: paul@softland.demon.co.uk (Paul Taylor) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: Sun, 05 May 1996 10:15:23 GMT Message-ID: <318c5f5d.4530409@news.demon.co.uk> References: <4mefbm$931@news.service.uci.edu> Brian Kevin Moore wrote: >I'm looking for any good resources on Objective C programming. Check out NeXT's web site (http://www.next.com/<somwehere or other>) for quite an in-depth discussion of Objective C the language and how it is implemented under NextStep. It is well worth looking at. -- Paul Taylor (paul@softland.demon.co.uk)
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: Card with multi serial port under NeXTStep Date: 5 May 1996 05:25:07 GMT Organization: Nanyang Technological University Message-ID: <4mhe3j$6oh@ntuix.ntu.ac.sg> References: <9604290827.AA00279@saem-ales.saem-ales.fr> <4melr4$2fi@news.next.com> Scott Vail (svail@regression.next.com) wrote: : In article <9604290827.AA00279@saem-ales.saem-ales.fr> writes: : > : > I need, for an application, 5 to 7 serials ports. The big question : > is, if this card exists, how can I drive it under NeXTStep ?? : > : > Xavier Fabre. : > : > Have you considered using a TCP/IP terminal server, connected to your ethernet port? Willem
From: dekorte@intrepid.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 5 May 1996 19:53:59 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4mj10n$8g3@nntp.onramp.net> References: <4mefbm$931@news.service.uci.edu> Cc: bkmoore@uci.edu In <4mefbm$931@news.service.uci.edu> Brian Kevin Moore wrote: > .. I'm just looking for a good introduction to Objective C with a few simple examples. http://www.batech.com/~dekorte/Objective-C/objc.html may have what you're looking for. -- Steve Dekorte "C++, the COBOL of the 90s"- UHH NeXTstep consultant - Anaheim, CA mailto:dekorte@suite.com (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Disabling menu trees? Date: 4 May 1996 10:30:25 GMT Organization: Turbocat's Development, Germany Message-ID: <4mfbk1$c9@turbocat.snafu.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi! I have the following menue in my app: . . . |Auftrag ->||neuer Auftrag a | . |Auftrag klonen -> | . |Auftrag ändern -> || über Aufträge 3| . || über Kunden 4| How do I disable/enable the whole tree? [auftragMenu setEnabled:NO] disables only one cell and cmd-3 still works. There must me a nicer way than to set all cells by hand. Thanks in advance. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: Marc@AccessOne.com(Marc Salvatori) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 6 May 1996 02:32:57 GMT Organization: AccessOne Message-ID: <4mjocp$m1l@news.accessone.com> References: <4mefbm$931@news.service.uci.edu> Cc: bkmoore@uci.edu In <4mefbm$931@news.service.uci.edu> Brian Kevin Moore wrote: > I'm looking for any good resources on Objective C programming. I do not > need a book on developing for the NeXT GUI. At least not yet anyway. I'm > just looking for a good introduction to Objective C with a few simple > examples. Once I master small text-based programs, I'll try to move on to > using the GUI. I would like to build a strong foundation on Objective C so > that I can write better code when I eventually start writing full-blown NeXT > STEP apps. I recognize that everyone has a different approach to learning something new, so I won't bother asking why you wish to initially divorce yourself from the interface. While my background in C is limited, I share your perception that the learning curve is initially very steep. I'm still reading through the stack of books I acquired last month to prepare myself for NS development. One I have yet to read may appeal to you. It's "Objective-C: Object-Oriented Programming Techniques" by Lewis J. Pinson and Richard S. Wiener. It does make some inescapable references to the interface, but as its title suggests, focuses mostly on the language. Another book I plan to read after that is "Programming Under Mach" by Joseph Boykin, David Kirschen, Alan Langerman & Susan LoVerso. It largely focuses on threads and interprocess communications. -- >< Marc J. Salvatori | >< >< mailto:salvo@accessone.com | MIME & NeXTMail are accepted ><
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 6 May 1996 04:15:13 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4mjuch$6k8@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: Using modal loops Date: 6 May 1996 05:45:12 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4mk3l8$cob@core.bard.edu> I am running a joystick calibrator at the beginning of the program. I am using a while loop now. It works but it seems needlessly wasteful of system resources. I would like to use a modal loop, but I couldn't figure it out from the general reference. A simple example would be great. Thanks in advance... -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
From: trey@castle.wizards.com (Trey Valenta) Newsgroups: comp.sys.next.programmer Subject: Missing 'strip' Date: 1 May 1996 07:44:19 -0700 Organization: Testing site Message-ID: <trey.830961790@castle> Someone before me did some funky things to a NeXT station I am maintaining. And well, 'strip' is missing from the compiler. Does anyone know where I can get a copy? I'm running NS3.3 on a NeXT station color. Any help appreciated, trey -- Trey Valenta trey@wizards.com <standard legal stuff here> Wizards of the Coast Renton, WA 206.204.7372
From: rft@cg.tuwien.ac.at Newsgroups: comp.sys.next.programmer Subject: HELP: Problems with variable argument methods Date: 6 May 1996 11:58:10 GMT Organization: Vienna University of Technology, Austria Message-ID: <4mkpgi$jj@news.tuwien.ac.at> For a specific reason I need to use a new/init pair with variable number of arguments. The Objective C manual in /NextLibrary/Documentation/NextDev/Concepts/ObjectiveC/2_ObjC/CoreObjC.rtfd states that this can be done exactly as with standard C functions. I have problems calling a method with a variable number of arguments from within another method with a variable number of arguments, in order to just pass along all arguments. (I.e. I have problems calling -init:number,...; from within +new:number,...; and thereby passing along all arguments.) Here are the relevant classes for stating my problem: @interface NaryNode : Node // Node is the baseclass of my hierarchy { // NaryNode should have n subnodes long subNodeNumber; Node ** subNodeTable; } - init : (long) newNumber, ... ; @end @interface GroupNode : NaryNode // a specialized version of NaryNode {} + (GroupNode *) new : (long) newNumber, ... ; @end @implementation NaryNode - init : (long) newNumber, ... // this works { long i; Node * newSubNode; va_list argPtr; subNodeNumber = newNumber; subNodeTable = malloc(sizeof(Node *) * newNumber); va_start(argPtr, newNumber); for (i = 0; i < newNumber; i++) subNodeTable[i] = va_arg(argPtr, Node *); va_end(argPtr); return self; } @end @implementation GroupNode + (GroupNode *) new : (long) newNumber, ... // DOESN'T WORK !!! { va_list argPtr; GroupNode * newGroup = [GroupNode alloc]; va_start(argPtr, newNumber); [newGroup init:newNumber,argPtr]; // Here seems to be the // problem !!! va_end(argPtr); return newGroup; } @end At the indicated point, the NaryNode - init should be called, with all the arguments that have been passed to GroupNode + new, but this doesn't seem to happen. It seems that at the indicated point, the method call is not compiled with the variable number of arguments that are subsumed in argPtr. Have I overlooked something? Does anybody have an idea what's happening? I'd appriciate any help! Thanks in advance for any effort! Here is some additional info: Just to test my implementation of variable arguments, I used the following implementation of GroupNode + new, and found that it worked: + (GroupNode *) new : (long) newNumber, ... // this works! { long i; GroupNode * newGroup = [GroupNode alloc]; Node * newSubNode; va_list argPtr; newGroup->subNodeNumber = newNumber; newGroup->subNodeTable = malloc(sizeof(Node *) * newNumber); va_start(argPtr, newNumber); for (i = 0; i < newNumber; i++) newGroup->subNodeTable[i] = va_arg(argPtr, Node *); va_end(argPtr); return newGroup; } ------------------------------------------------------------------------ Robert F. Tobler - tel:+43(1)58801-4585,fax:5874932 Institute of Computer Graphics - mailto:rft@cg.tuwien.ac.at Vienna University of Technology - http://www.cg.tuwien.ac.at/~rft/ ------------------------------------------------------------------------
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Q: How to use foundation Date: 6 May 1996 14:00:19 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4ml0lj$hei@lion.embl-heidelberg.de> References: <4md83k$onr@n.ruf.uni-freiburg.de> In article <4md83k$onr@n.ruf.uni-freiburg.de> Lars Konieczny writes: > Hi, > > sorry if this is a stupid question, but how can I make the compiler be > aware of the foundationKit classes? When the correct header is > imported in the definition file (#import <foundation/NSObject.h>), the > compiler exits with: > > ld: Undefined symbols: > .objc_class_name_NSObject > *** Exit 1 Well, you should add the Foundation library to your project ... > > Another question: can I make IB unparse subclasses of NSObjects > correctly? After subclassing NSObject, IB puts No > #import <appkit/appkit.h> > > #import "NSObject.h" > > into the class definition file obviously ignoring that NSObject is not > part of the appkit. .. and you will have trouble using NSObject subclasses in IB without hacks (read a NeXT answer ... I cannot remember the number). -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Window not taking title Date: 6 May 1996 14:08:25 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4ml14p$hf5@lion.embl-heidelberg.de> References: <4mgive$1ats@core.bard.edu> In article <4mgive$1ats@core.bard.edu> hs283@bard.edu (Hans Steiner) writes: > I am trying to set the title of a Window with setTitle:. When I check > with [Window title] I get a null pointer. Here's how I'm doing it: > > - setName: (const char *)name > { > moduleName = malloc(strlen(name) + !); > strcpy(moduleName,name); > [myWindow setTitle:name] > // [myWindow setTitle:moduleName]; //tried this > > return self; > } > > name and moduleName are both getting the correct value. > > here's the rest: > > - initWithController:(StickMusic *)controller > { > name=[controller getName]; > [self setName:name]; > [nameField setStringValue:name]; > > return self; > } > > from StickMusic: > - (char *)getName > { > char *name; > > name = malloc(3); > sprintf(name,"%i",++moduleCount); > > return name; > } ;-) ... ;-) Well, a while ago I had exactly the same problem. The answer (as usual) was RTFM. To be more informative: you should change the name of the method setName to setWhateverOtherName:... IB has strange (documented) behavior: it sets the name (the ivar) to nil and executes the setName: method before any awake method. I think this is because of historical reasons and hope should disappear with NS 4.0. .. so as a general rule: if you have an ivar called anObjects, do not call a method setAnObject: if you do not want to lose your IB connections ;-) hope this helps.... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Using bowman? Date: 6 May 1996 16:57:18 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4mlb1e$bcs@stc06.ctd.ornl.gov> I'm interested in trying to have my NextStep app available to people running X-Windows somewhere on the network. I've seen information about gnu's libXopen that may provide this and also ran into a reference to bowman, a BOGusWindowMANager that makes XWindows look like Next's interface. I've downloaded it and find that it requires xmkmf. I don't have a full X distribution, just CubXWindow. Also I don't have Imake. Anyone have the Makefiles for BOWMAN already made? Or else how to translate Imakefiles into Makefiles. -- - - - - - - - - - J. W. Wooten
From: esky@marathon.cs.ucla.edu (Eskandar Ensafi) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: Using -unmounting:ok: <<< BUG REPORT Date: 6 May 1996 17:36:01 GMT Organization: University of California, Los Angeles Message-ID: <4mlda1$c5k@delphi.cs.ucla.edu> <<< Date: Wed, 24 Apr 96 23:56:02 +0200 <<< From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> <<< To: Eskandar Ensafi <esky@CS.UCLA.EDU> <<< Subject: Re: Using -unmounting:ok: (I can't get any results!) KBNS.33.3.007_o3.2-3.3o Application -unmounting:ok:, -app:unmounting: Description: ----------- Three Application instance methods monitor mounting and unmounting volumes: -mounted: and -unmounted:, which are called after [[Application workspace] beginListeningForDeviceStatusChanges];, and -unmounting:ok:, which is supposedly called in ``all active'' applications and ``only if the application was launched from the Workspace Manager'' without the need for the message to [Application workspace]; three corresponding delegate methods are (seemingly correctly) called from the main method, see the documentation. Problems: -------- 1. Apparently the former two work as described, but the latter pair is called only in applications holding a reference to the volume, even if they are not active (probably the author meant ``running'' applications), and even if they were started from a shell. 2. Even though the documentation says that the default implementation of -unmounting:ok: sets flag to the return value of -app:unmounting: and NO objects to unmounting, and the return value to -unmounting:ok: is ``an arbitrary integer; your application defines and interprets it.'', unmounting is blocked either by * returning 1 from -app:unmounting:, if the default implementation of -unmounting:ok: is used, * both setting flag to 0 and returning 0 in -unmounting:ok: (both are required). How to verify: ------------- Use this in an Application subclass. Don't declare these methods, otherwise the compiler can't check them against the declarations in the AppKit. Don't forget to begin listening for device status changes, or not, depending on what you want to test. #ifdef TEST_KBNS_33_3_007 -app:sender mounted:(const char *)fullPath{ // from mounted: NXRunAlertPanel(NULL,"app:mounted:``%s''",NULL,NULL,NULL,fullPath); return [[self superclass] instancesRespondTo:_cmd] ? [super app:sender mounted:fullPath] : 0 /*value ignored by AppKit*/ ; } -app:sender unmounted:(const char *)fullPath{ // from app:unmounted NXRunAlertPanel(NULL,"app:unmounted:``%s''",NULL,NULL,NULL,fullPath); return [[self superclass] instancesRespondTo:_cmd] ? [super app:sender unmounted:fullPath] : 0 /*value ignored by AppKit*/ ; } -(int)app:sender unmounting:(const char *)fullPath{ // from unmounting:ok: NXRunAlertPanel(NULL,"app:unmounting:``%s''",NULL,NULL,NULL,fullPath); return [[self superclass] instancesRespondTo:_cmd] ? [super app:sender unmounting:fullPath] : 0 /*objects to unmounting*/ ; } -(int)mounted:(const char *)fullPath{ // after beginListeningForDeviceStatusChanges NXRunAlertPanel(NULL,"mounted:``%s''",NULL,NULL,NULL,fullPath); chdir(fullPath); return [super mounted:fullPath]; } -(int)unmounted:(const char *)fullPath{ // after beginListeningForDeviceStatusChanges NXRunAlertPanel(NULL,"unmounted:``%s''",NULL,NULL,NULL,fullPath); return [super unmounted:fullPath]; } -(int)unmounting:(const char *)fullPath ok:(int *)flag{ // if and only if application launched from Workspace Manager NXRunAlertPanel(NULL,"unmounting:``%s'' ok:&%i", NULL,NULL,NULL,fullPath,*flag); chdir("/"); switch(NXRunAlertPanel(NULL,"set flag to:","0","1",NULL)){ case NX_ALERTDEFAULT : *flag=0; break;case NX_ALERTALTERNATE: *flag=1; } switch(NXRunAlertPanel(NULL,"return:","0","1","from super")){ case NX_ALERTDEFAULT : return 0; break;case NX_ALERTALTERNATE: return 1; break;default : return [super unmounting:fullPath ok:flag]; } } #endif Urgency: ------- Does no damage, not immediately obvious, fairly easy workarounds. Confirmed: --------- Eskandar_Ensafi (o3.3o, reported), Raf_Schietekat (o3.2o) Details: ------- * Eskandar_Ensafe tested with floppy disk on Intel and optical disk on NeXTcube. * Raf_Schietekat tested with floppy disk and CD-ROM on NeXTstation. Workaround: ---------- 1. Open a reference in -mounted: or its delegate method, if this is what you want. 2. Always return 0 in mounting:ok: if you implement this, and return 1 to object in app:mounting:, otherwise. If your application blocks a volume because you followed the documentation, quit the application first, or do something else to stop it referring to that volume, and try again. Cure (for NeXT): --------------- Make documentation match reality (specify which applications get this message). Also don't think this method is obsolete: a ...Did...-like method does not obsolete a ...Will...-like method (see Comment). Report history: -------------- * Eskandar_Ensafi reported this to c.s.n.programmer on 1996-04-10. * On 1996-04-15 a first draft of this KBNS entry was posted to programmer and bugs on the mistaken assumption that the message was never sent. * On <todo specify> a corrected draft was posted to both newsgroups, with the stuff about 2. added and some other changes. Comment: ------- The AppKit Release Notes specific to 3.0 (as viewed on 3.2) say ``The following Speaker/Listener messages have been obsoleted and replaced with improved API described below under "Interapplication Image Dragging" and "Workspace Protocol". '', mentioning -unmounting:ok:. Well, it still works for those applications holding references to the volume. Also, there is functionality that you can't get from -unmounted:. Without -unmounting:ok:, a user gets a meaningless panel saying the volume could not be unmounted because some application is using it. With the method, that application gets the chance to take appropriate action, like asking the user for what should be done. Or will just app:unmounting: be kept? It is not clear what is meant, to say the least. <<< -- In a long black car And the lights flash blue | Vision With the prettiest shit My vision thing | Thing From Panama Come | THE SISTERS When the sirens wail Slamming through | OF MERCY
From: jsickel@charm.net (Jeffrey Sickel) Newsgroups: comp.sys.next.programmer Subject: GDB and NSRange problems Date: 6 May 1996 18:30:44 GMT Organization: Charm.Net : Baltimore Local Internet Access, Hon Message-ID: <4mlggk$mjd@canton.charm.net> Summary: bad NSRanges via GDB Keywords: gdb, NSRange Does anyone know of a way to get gdb to provide valid NSRange information? Other than NSLog other coded printing... jas
From: tony@gst1.gestel.it (Antonio Flores) Newsgroups: comp.sys.next.programmer Subject: Adaptec 2940 and command queueing Date: 7 May 1996 12:06:53 GMT Organization: Italia Online Message-ID: <4mnect$69s@mikasa.iol.it> Is there a way to enable the "command queueing" feature when one or more scsi devices can t handle it? Antonio Flores tony@gestel.it
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: HowTo: shlib and/or dylib Date: 8 May 1996 04:45:27 GMT Organization: No organisation supplied Message-ID: <4mp8t7$61@hermes.is.co.za> Hi, How can I make a .shlib or .dylib under NeXTStep. I need this urgently as we have a library of all the business classes we use in our project that we want to cenvert to a .shlib. Our framework is almost in place and we are starting to turn out a lot of applications in a very short period of time. The problem we are facing however is that we need to link our applications with the -all_load flag. This causes our library to be linked into every single application in full. Our current space requirements just to place our Apps on a user machine is about 200MB for our current apps. I think that this will grow still very substantially in the next 2 - 3 months. There is also a lot of other problems with static linking approach. I know that OpenStep makes it relatively easy to this, but we can't wait for it , as we need to roll out our system's first phase at the end of June. Any help will be greatly appreciated. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: zazula@running-start.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: HowTo: shlib and/or dylib Date: 8 May 1996 16:31:13 GMT Organization: Running Start, Inc. Message-ID: <4mqi8h$d7o@nntp1.best.com> References: <4mp8t7$61@hermes.is.co.za> In-Reply-To: <4mp8t7$61@hermes.is.co.za> On 05/07/96, gvandyk@icon.co.za wrote: >Hi, > >How can I make a .shlib or .dylib under NeXTStep. I need this urgently >as we have a library of all the business classes we use in our project >that we want to cenvert to a .shlib. Our framework is almost in place >and we are starting to turn out a lot of applications in a very short >period of time. > >The problem we are facing however is that we need to link our >applications with the -all_load flag. This causes our library to be >linked into every single application in full. Our current space >requirements just to place our Apps on a user machine is about 200MB >for our current apps. I think that this will grow still very >substantially in the next 2 - 3 months. There is also a lot of other >problems with static linking approach. > Hi - You can avoid using -all_load if you can predict the symbols that will be needed at run-time. Instead of -all_load, add "-u" options to the link flags (OTHER_LDFLAGS in Makefile.preamble). This will force only the symbols you list to be linked in. You can get the list of symbols by running under gdb and watching for rld() errors (running from the Workspace never seems to print out the actual symbols). Also, make sure you strip all your stuff well (you probably do already but I thought I'd mention in just in case...). Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@running-start.com http://www.running-start.com http://www.running-start.com/zazula/index.html
From: vrobi@futon (Robert Vasvari) Newsgroups: comp.sys.next.programmer Subject: sybperl for NextStep Date: 8 May 1996 18:53:49 GMT Organization: California State University Sacramento Message-ID: <4mqqjt$nnp@news.csus.edu> an anybody tell me where I can get a copy of sybperl compiled for NextStep on Intel? The one that lives at next-ftp.peak.org is broken and it is probably m68k only. thanx... =[vrobi]=
From: kris@siva.captech.com (Kris Skrinak) Newsgroups: comp.os.ms-windows.programmer.networks,comp.os.msdos.programmer,comp.programming,comp.sys.next.programmer,comp.unix.sco.programmer Subject: TCP/IP Server Bug Date: 8 May 1996 12:30:13 -0700 Organization: Capital Technologies, 604 Mission, Ste. 830, SF, CA 94105 800.sys.admin Message-ID: <4mqso5$nut@siva.captech.com> I have a tcp/ip programming roadblock that I need help getting over. Let's name the two hosts in this scenario S and C. S starts up, gets a socket, bind()s it and listen()s. C gets a socket, bind()s it and connect()s to S. S accept()s the connection, and begins servicing requests from C. So far, so good. Some important details: When the connection is established, an transitory port is created on C's end to talk to the assigned port on S's side. netstat -an | grep <port> shows both ends of the connections as ESTABLISHED, and the transitory port number can be observed. Now the fun begins: If C gets shot in the head, S can easily detect it, and close the connection. BUT, C's end of the connection ends up in TIME_WAIT mode, waiting for either an ACK or a timeout (5-minutes). In this case, there's not really a problem, since the port being tied up is transitory, and restarting C will just cause a new one to be generated. However, if S dies, C can close its side of the connection, but the assigned port is tied up until the timeout occurs, which prevents bringing S back up for 5-minutes. (The scenario is the same, but this time, the port being held captive is the advertised one.) So, what I need is some way to either bypass that timeout, or send some sort of ACK to the other end before dying, or something. The client code detects that the server went away, and starts a thread to reconnect to it. The server code detects when a client went away, and removes it from its list. All the pieces are there to create this interface *EXCEPT* this timeout.
Newsgroups: comp.sys.next.misc,comp.sys.next.programmer From: tom@basil.icce.rug.nl (Tom Hageman) Subject: Re: How can I add undo to my program? Message-ID: <Dqsu13.41G@basil.icce.rug.nl> Originator: uubasil@obelix.icce.rug.nl Sender: news@basil.icce.rug.nl (NEWS pusher) Organization: Warty Wolfs References: <199605020233.WAA09449@slate.cs.rochester.edu> Date: Thu, 2 May 1996 22:19:51 GMT "Eric K. Ringger" <ringger@cs.rochester.edu> wrote: > Chris Osborn wrote: > >How can I add undo to my program? Even better, how about a > >multi-level undo? > > > >It seems like a nearly impossible task of getting undo to work when > >there are so many different things to keep track of that can be > >undone. Especially in a text object! > > > >I remember a few years ago seeing something that someone had done > >which had some sort of undo manager, and all you had to do was add > >some kine of undo method to your objects to get it to work. > > 'Sounds like Draw.app ! > > >I'm looking for that, or something similar. > [...] > > Check out the following files in the NEXTSTEP Developer distribution: [...] Also check out the UndoManager MiniExample. It is (was? I haven't been at ftp.next.com recently...) available as NeXTanswer 1582. Hope this helps, -- __/__/__/__/ Tom Hageman <tom@basil.icce.rug.nl> [NeXTmail/Mime OK] __/ __/_/ IC Group <tom@icgned.nl> (work) __/__/__/ "Ed is the standard text editor" __/ _/_/ -- Unix Programmer's Manual
From: alanf@izzy.net Newsgroups: comp.sys.next.programmer Subject: embedded graphic Date: 9 May 1996 03:15:37 GMT Organization: Isthmus Corporation, Inc. Message-ID: <4mro0p$t8e@izzy4.izzy.net> I'm working on an app in which I'd like to use either a Text in a ScrollView, or an NXBrowser, with an embedded graphic. A couple examples of where I've seen this are FaxReader, and the Processes window of OmniWeb. It seems common enough; I just can't seem to figure out how to do it from the General Reference and HeaderViewer documentation. Is there an example or a code fragment available that might enlighten me? Thanks, Alan Frabutt alanf@izzy.net
From: svail@regression.next.com (Scott Vail) Newsgroups: comp.sys.next.programmer Subject: Re: Card with multi serial port under NeXTStep Date: 9 May 1996 06:43:37 GMT Organization: NeXT Software, Inc. Message-ID: <4ms46p$gu1@news.next.com> References: <4mfutm$3ag@tallowcross.uni-frankfurt.de> In article <4mfutm$3ag@tallowcross.uni-frankfurt.de> uli@tallowcross.uni-frankfurt.de (Uli Zappe) writes: > svail@regression.next.com (Scott Vail) wrote: > > There are some limitations though, the serial > > ports must not share interrupts and you cannot set the serial ports to > > speeds greater than 57600 baud. > > Well, you can. There are several serial cards available which multiply the > speed set by the software with a factor of 2 or 4 by means of hardware. Thus, > a (software) setting of 57600 will actually result in 230400 baud. It works > perfectly with an ISDN ZyXEL for me. > > Bye > Uli > Nextstep does not support port speeds above 57600 with standard hardware. If you are using unsupported or non-standard hardware, and it works without error, then by all means use it, but as an employee I can't encourage it. --Scott ----- This is not official NeXT stuff, so ignore me if you want. http://www.next.com/~svail/
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: embedded graphic Date: Thu, 9 May 1996 08:51:57 +0100 Organization: Cedar Systems Distribution: world Message-ID: <GscNXDAdQakxEw+w@cedar.co.uk> References: <4mro0p$t8e@izzy4.izzy.net> MIME-Version: 1.0 In article <4mro0p$t8e@izzy4.izzy.net>, alanf@izzy.net writes >I'm working on an app in which I'd like to use either a Text in a >ScrollView, or an NXBrowser, with an embedded graphic. A couple examples >of where I've seen this are FaxReader, and the Processes window of >OmniWeb. It seems common enough; I just can't seem to figure out how to >do it from the General Reference and HeaderViewer documentation. Is there >an example or a code fragment available that might enlighten me? > >Thanks, >Alan Frabutt >alanf@izzy.net Use the Text method replaceSelWithCell: to insert graphics into a Text object. If you look at the Text documentation in HeaderViewer, there is a section entitled 'Adding Graphics To The Text' that describes the methods that an embedded graphic 'cell' must implement. Hope this is enough to get you started. Paul. ------------------------------------------------------------------ Paul Heffernan. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: andrewg@microlise.co.uk (Andrew Gierth) Newsgroups: comp.os.ms-windows.programmer.networks,comp.os.msdos.programmer,comp.programming,comp.sys.next.programmer,comp.unix.sco.programmer Subject: Re: TCP/IP Server Bug Followup-To: comp.unix.programmer Date: 9 May 1996 12:58:34 GMT Organization: Microlise Engineering Ltd. Message-ID: <4msq5q$8ip@microl4.microlise.UUCP> References: <4mqso5$nut@siva.captech.com> In <4mqso5$nut@siva.captech.com>, kris@siva.captech.com (Kris Skrinak) writes: > >I have a tcp/ip programming roadblock that I need help getting over. [description of usual TIME-WAIT scenario snipped] I have answered this same post in comp.unix.programmer. Please don't post multiple copies of the same question; find the most relevent group(s) and crosspost ONE copy. -- Andrew (andrewg@microlise.co.uk) "How is this place run - is it an anarchy?" "No, I wouldn't say so; it is not that well organised..."
From: lopatic@malaka.dbs.informatik.uni-muenchen.de (Thomas Lopatic) Newsgroups: comp.os.ms-windows.programmer.networks,comp.os.msdos.programmer,comp.programming,comp.sys.next.programmer,comp.unix.sco.programmer Subject: Re: TCP/IP Server Bug Date: 9 May 96 14:04:23 GMT Organization: Institut fuer Informatik der Universitaet Muenchen Message-ID: <lopatic.831650663@malaka> References: <4mqso5$nut@siva.captech.com> kris@siva.captech.com (Kris Skrinak) writes: [lines deleted, we have a server S and a client C] >However, if S dies, C can close its side of the connection, but the >assigned port is tied up until the timeout occurs, which prevents >bringing S back up for 5-minutes. (The scenario is the same, but this >time, the port being held captive is the advertised one.) [lines deleted] Setting the socket option SO_REUSEADDR should cure this problem. Hope this helps. -Thomas -- Thomas Lopatic lopatic@informatik.uni-muenchen.de
Newsgroups: comp.sys.next.programmer From: pascal@burgond.remcomp.fr (Pascal Bourguignon) Subject: Re: HELP: Problems with variable argument methods Message-ID: <Dr0But.1po@burgond.remcomp.fr> Sender: pascal@burgond.remcomp.fr (Pascal Bourguignon) Organization: P. Bourguignon Informatique References: <4mkpgi$jj@news.tuwien.ac.at> Date: Mon, 6 May 1996 23:28:05 GMT In article <4mkpgi$jj@news.tuwien.ac.at> rft@cg.tuwien.ac.at writes: > I have problems calling a method with a variable number of arguments from > within another method with a variable number of arguments, in order to just > pass along all arguments. (I.e. I have problems calling -init:number,...; > from within +new:number,...; and thereby passing along all arguments.) You have to do exactly like with C: declare one more method with only a va_list parameter, and pass your argPtr to it. Inside this methode, use va_arg() to access the arguments in the list. __Pascal Bourguignon__
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: More on: Problem with key equivalents in buttons Date: 9 May 1996 15:03:56 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4mt1gs$a84@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hello, i posted recently the folowing problem: i have problems with key equivalents in buttons: i use buttons with \r (return) key equivalents within Panel windows: in some of these panels, the buttons' equivalents work while in others they do not. I could not see the difference between the 2 kinds: both kinds of panels were made with IB (i am working under version 3.3) in the most usual way. I got a couple of suggestions which did not solve the problem. Then i made further tests to check how the firstResponder is set when the Panels are displayed: I got the following results: 1) the 2 Panels where the buttons with Return key equivalents DID work, the firstResponder was of class Panel; 2) the 2 Panels where these buttons did NOT work had first responders of class Text (in one case, the Text seems to correspond to a TextField (selectable, not editable), in the other the Text is within a scroll view (selectable, not editable)). Note that no cursor was active when displaying the Panels; - in all cases the Panels are displayed with makeKeyAndOrderFront. I can see no illuminating pattern here! I finally solved the problem with explicitly assigning the first responder status (with makeFirstResponder) to the buttons in the offending Panels. However, the following questions remain: - HOW the first responder is assigned when a Panel becomes key: to WHICH object within the Panel is it assigned? - WHY, in 1) above, the buttons worked? As far as i understand the next responder chaining correctly, how can the button WITHIN the Panel get the first responder status from that Panel? ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
From: Jonathan.Rice@tcd.ie (Jonathan Rice) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Printing Alpha-Channel Images? DPS vs. PS Date: 9 May 1996 17:27:51 GMT Organization: TCD, Computer Science Message-ID: <4mt9un$q2e@news.cs.tcd.ie> Hi. I'd like to do some diagrams in Create using its transparency feature, so that I can do cut-away-style effects. However, since such alpha-channel features are only part of NeXT's Display Postscript implementation, I would appear to be rather stymied as regards printing such diagrams. Or am I? One general solution would be to use the DPS system to create a TIFF of a suitably-high dpi (say 300dpi?) and then print this image. So it's rather inefficient, but at least it should work. Has anyone figured out a neat way to do this sort of thing? A key question, I suppose, is whether I can manage to convince Create to export an EPS file with the alpha information intact. I can't specifically test this out yet, as I'm only using a demo version of the app. I have done a little test with Draw.app, however, in which I draw two overlapping red ovals and then "Save To" an EPS. I then edit the EPS and comment out its header redefinition (neutralisation) of setalpha and currentalpha. Down at the end of the EPS file, where it defines the two ovals, I add two setalpha calls, giving something like this: ... %%BeginSetup /oval { translate scale newpath 0.5 0.5 0.5 0 360 arc closepath } def ... %%EndSetup gsave 125 294 234 291 rectclip 0 setgray 0.333333 setgray 0 0 0 setup gsave 1 0 0 nxsetrgbcolor 0.5 setalpha % added this line 173 209 130 371 oval fill grestore gsave 1 0 0 nxsetrgbcolor 0.2 setalpha % added this line too 160 173 194 299 oval fill grestore ... Dragging this back into Draw (or whatever) produces two intersecting ovals, one with 50% opacity, and the second overlapping it with 20% opacity! Almost there! However, the intersection region between the two ovals is *also* at 20% opacity, rather than 60%, as would be hoped. What's going on here? I guess that the transfer mode is the problem, that the second "fill" is just wiping out the coverage info of the first one. How can I fix this? And presuming this can be worked out, how do I then convince DPS to render everything at 300dpi, etc.? This must have been done before... --- Jonathan Rice Jonathan.Rice@tcd.ie (NeXTMail OK) http://isg.cs.tcd.ie/jorice/jorice.html Dept. Comp. Sci., Trinity College, Dublin 2, Ireland.
From: dkoski@running-start.com (David Koski) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Re: Printing Alpha-Channel Images? DPS vs. PS Date: 10 May 1996 15:10:49 GMT Organization: Best Internet Communications Message-ID: <4mvm9p$as1@nntp1.best.com> References: <4mt9un$q2e@news.cs.tcd.ie> In-Reply-To: <4mt9un$q2e@news.cs.tcd.ie> On 05/09/96, Jonathan Rice wrote: ... > Dragging this back into Draw (or whatever) produces two > intersecting ovals, one with 50% opacity, and the second > overlapping it with 20% opacity! Almost there! However, the > intersection region between the two ovals is *also* at 20% > opacity, rather than 60%, as would be hoped. What's going on > here? I guess that the transfer mode is the problem, that the > second "fill" is just wiping out the coverage info of the > first one. How can I fix this? Transparency is really only good for compositing. When you draw the ovals, you are laying down paint with a certain opacity, not compositing in SOVER mode. There are three ways that I can think of to get the desired effect: 1. using a bitmap editor, draw some transparent ovals and composite them over one another 2. you might be able to do this by hand: try doing it in a paint program to see what the colors will look like, then do it in your draw program with three shapes: the two ovals, and a piece where they overlap (some drawing programs can create this intersection for you). 3. use a drawing program that can simulate transparency effects. -- David Koski Running Start, Inc. OpenStep Development & Consulting dkoski@running-start.com http://www.running-start.com
Newsgroups: comp.sys.next.programmer From: chane@ln1d274nwk (Eddie Chan) Subject: Problem with dragging rows in NXTableView Message-ID: <1996May10.160817.4782@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Fri, 10 May 1996 16:08:17 GMT Hi all, Dragging a row in a NXTableView locks up the process. However, if I drag a column frist then everything works. Any work arounds? many thanks, eddie -- Edward C.P Chan Work: eddie_chan@swissbank.com Home: eddie@boston.demon.co.uk
From: agm@dataphone.se (Axel G. Merk) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.hardware,comp.sys.next.programmer Subject: Digital HiNote Ultra II Display Driver Date: 10 May 1996 14:41:33 GMT Organization: Dataphone Communication Networks Distribution: world Message-ID: <4mvkit$j3@nic.dataphone.se> Keywords: notebook, display driver The fairly new Ultra II notebook by Digital seems like one of the first useful notebooks on the market with few compromises, with the following available: * 133 MHz Pentium * 3.9lbs * 10.4" SVGA TFT Active Matrix color * 1.35GB HD * 40MB RAM * the usual extras possible (incl 'multi-media', various types of docks) * price is competitive (call Digital at 800.642 4532) (among the few compromises: no PCI bus to keep the size/weight low) I'm not affiliated with DEC and don't really want to have many others order the notebook as I don't want to increase its lead time. To run NS on the system, one will need a display driver. My question is whether anyone is currently developing a driver for the notebook or knows of someone who is interested in developing that driver? I would appreciate direct email as reply. I will summarize if there is an interest. Thank you and best regards, axel Axel Merk agm@cubiclum.com agm@doc.ic.ac.uk agm@daphone.se
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer Subject: Q:Sound[Kit] & OPENSTEP? Date: 10 May 1996 21:19:24 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4n0bss$s58@news4.digex.net> Hi, I'm wondering what is the future of sound under OPENSTEP? I'm guessing that the soundKit will continue to be available and work under OPENSTEP for Mach. However, I wonder if it will be available, or it's functionality, under other OPENSTEP implementations? In other words, if I port my application to OPENSTEP for Mach, and then later wish to port it to OPENSTEP for NT, will it work under NT with sound? (I'm hoping that somekind of sound facility is available through the Postscript server layer of OPENSTEP). I realize this is probably not a big deal for most people but it is an important issue for me... Thanks for any info. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: ericw@cmg.fcnbd.com (Eric Wespestad) Newsgroups: comp.sys.next.programmer Subject: code to return ip address machine Date: 10 May 1996 17:29:44 GMT Organization: First Chicago NBD Corporation, Chicago IL, USA Message-ID: <4mvue8$snk@news.fnbc.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Howdy, Does anyone have a routine that just returns the ip address of the current machine (and hides the system call details), either as an unsigned long, or struct in_addr, either will do. Please share this if you have it, so I don't have to write it :-) Thanks!
From: Luke Howard <Luke_Howard@inter.net.au> Newsgroups: comp.sys.next.programmer Subject: Re: Card with multi serial port under NeXTStep Date: Sun, 12 May 1996 13:24:08 +1100 Organization: Internet Australia, Melbourne. Message-ID: <31954BC8.6FE2@inter.net.au> References: <9604290827.AA00279@saem-ales.saem-ales.fr> <4melr4$2fi@news.next.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Scott Vail wrote: > Assuming you are talking about 3.3 with the new ISASerialPort driver, you > can drive as many serial ports are there are IRQ's and port-addresses > available for each port. There are some limitations though, the serial > ports must not share interrupts and you cannot set the serial ports to We have a couple of Digiboards running under 3.3 with the Mux driver, using the new ISASerialPort driver for the internal serial ports. We did have some problems with the ttydf* devices - as I recall, we ended up using ttyd*. But we did get a login prompt (had some problems with PPP though). The other approach is just to setup a FreeBSD box as a terminal server :-) - we're using one with two 16-port Cyclades boards, and it's relatvely reliable/inexpensive. -- Luke
From: synper5@aol.com (Syn Per5) Newsgroups: comp.sys.next.programmer Subject: EMPLOYMENT AT FORTE SOFTWARE Date: 13 May 1996 04:14:05 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4n6r0d$r8f@newsbf02.news.aol.com> FORTE SOFTWARE HAS SEVERAL NATIONAL CAREER OPPORTUNITIES AND WOULD BE INTERESTED IN DISCUSSING THESE OPPORTUNITIES WITH YOU OR A COLLEAGUE OF YOURS. PLEASE CONTACT: MATTHEW GREEDY AT FORTE SOFTWARE 510-869-2116
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.programmer Subject: why does this program core dump ? Date: 13 May 1996 15:16:08 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4n7jno$e00@netty.york.ac.uk> the following program gives me a bus error (null argument to strtod() from +[FrameView initialise]) has anyone any idea why ? cheers, rog. #include <appkit/appkit.h> main() { id win; NXRect fr = {0, 0, 100, 100}; win = [[Window alloc] initContent:&fr style:NX_PLAINSTYLE backing:NX_RETAINED buttonMask:0 defer:NO]; return 0; }
From: hs283@bard.edu (Hans Steiner) Newsgroups: comp.sys.next.programmer Subject: any way to stop loadNibSection: from initing all vars? Date: 13 May 1996 16:08:05 GMT Organization: Bard College, Annandale-on-Hudson, New York 12504 Message-ID: <4n7mp5$bi8@core.bard.edu> I was wondering if there is any way that anyone knows of that will stopp loadNibSection from initializing all instance variables to zero. I ran into some annoying complications because awakeFromNib was being sent twice to my object being loading from a .nib file so I would like to be able to do thesew things in init without them being zeroed by the NIB -- (C)Copyright Hans-Christoph Steiner. Permission for use of this material is freely granted to all except Microsoft. Microsoft can secure distribution rights for US$1000. Use of this material without permission constitutes an agreement to these terms. Report violations to me and postmaster@microsoft.com
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <jim@ergotech.com> Message-ID: <199605141447.AA03074@ergotech.com> Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable From: Jim Redman <jim@ergotech.com> Date: Mon, 13 May 96 10:02:40 -0600 Subject: Re: Card with multi serial port under NeXTStep I don't know whether anyone mentioned this. Central Data (800) = 482 0316 - fax (217) 359 6904 have SCSI->serial solutions with NS = support. Jim=
From: zazula@running-start.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: why does this program core dump ? Date: 13 May 1996 21:38:47 GMT Organization: Running Start, Inc. Message-ID: <4n8a57$ch0@nntp1.best.com> References: <4n7jno$e00@netty.york.ac.uk> In-Reply-To: <4n7jno$e00@netty.york.ac.uk> On 05/13/96, Roger Peppe wrote: >the following program gives me a bus error (null >argument to strtod() from +[FrameView initialise]) > >has anyone any idea why ? > >#include <appkit/appkit.h> > >main() >{ > id win; > NXRect fr = {0, 0, 100, 100}; > > win = [[Window alloc] initContent:&fr style:NX_PLAINSTYLE > backing:NX_RETAINED buttonMask:0 > defer:NO]; > return 0; >} > > Hi - The problem is that you are trying to create a window without a connection to the WindowServer. The Window object is the AppKit front-end to the DisplayPostscript window operators but it does not establish this connection for you. The Applicaton object does when it is run. Here is a modified version of the program that will get your window to display. Unless you have a particular reason for creating Window objects by-hand, you should check out /NextDeveloper/Apps/InterfaceBuilder.app for creating UI's graphically. Also, check out the docs on becoming the Application's delegate for a better way to interact with the Application event-loop (in particular, -appDidInit:). #include <appkit/appkit.h> main() { id win; NXRect fr = {0, 0, 100, 100}; NXApp = [Application new]; win = [[Window alloc] initContent:&fr style:NX_PLAINSTYLE backing:NX_RETAINED buttonMask:0 defer:NO]; [win makeKeyAndOrderFront:nil]; [NXApp run]; return 0; } P.S. You have to kill this off since there is no main-menu to quit with. Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@running-start.com http://www.running-start.com http://www.running-start.com/zazula/index.html
From: i455@stio1 (Bergmann Winfried) Newsgroups: comp.sys.next.programmer Subject: Tahiti Sound Driver Date: 14 May 1996 00:17:57 GMT Organization: University of Wuerzburg, Germany Message-ID: <4n8jfl$khb@winx03.informatik.uni-wuerzburg.de> Hello, I'd like to write a SoundDriver for the Turtle Beach Tahiti (at least I'll try). Does anybody know, where I can get more information to the soundcard ? (I don't want to spend money for it!) thanx Winfried -- ========================================================================== Winfried Bergmann | Germany - 91478 Ulsenheim 14 | I read it in the paper i455@stio1.sari.fh-wuerzburg.de | There's death on every page bergmann@rz-ppp-12.fh-wuerzburg.de|
From: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: duplicate value in a UNIQUE INDEX column Date: 14 May 1996 12:46:42 GMT Organization: Universite de La Rochelle Message-ID: <4n9vbi$s25@hpuniv.univ-lr.fr> I have an EOController (in Interface Builder) and an NXTableView. I have : - (EODataSourceFailureResponse)controller:(EOController *)controller failedToInsertObject:anEO inDataSource:aDataSource { return EORollbackDataSourceFailureResponse; } My problem: When I insert a new row which causes duplicate value in a UNIQUE INDEX column, Here is what EOF does: 1 --controller willInsertObject: (I send beginTransaction in this method) 2 --ALERTE: SQL: -239, Could not insert new row - duplicate value in a UNIQUE INDEX column. INSERT INTO myTable (myKey, myFieldA, myFieldB) VALUES ( ? , ? , ? <<< ) 3 ---controller willRollbackDataSource 4 ---controller didRollbackDataSource 5 --controller didInsertObject: (I send commitTransaction in this method) ------------------------------ "controller didInsertObject" should not be called!!! and My NXTableView Display the new row !!! (the new row is not inserted in my INFORMIX Table) The RollBack operation don't seems to work!!! /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: why does this program core dump ? Date: 14 May 1996 13:10:26 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4na0o2$hlv@lion.embl-heidelberg.de> References: <4n7jno$e00@netty.york.ac.uk> In article <4n7jno$e00@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) writes: > the following program gives me a bus error (null > argument to strtod() from +[FrameView initialise]) > > has anyone any idea why ? > > cheers, > rog. > > #include <appkit/appkit.h> > > main() > { > id win; > NXRect fr = {0, 0, 100, 100}; > > win = [[Window alloc] initContent:&fr style:NX_PLAINSTYLE > backing:NX_RETAINED buttonMask:0 > defer:NO]; > return 0; > } You have not created a new context for the window! If you want to see any effect, you should surround the window allocation with: NXApp = [Application new]; .. // Window stuff here [NXApp run]; Have fun -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.programmer Subject: Re: why does this program core dump ? Date: 14 May 1996 12:51:43 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4n9vkv$5ms@netty.york.ac.uk> References: <4n7jno$e00@netty.york.ac.uk> <4n8a57$ch0@nntp1.best.com> On 13 May 1996 21:38:47 GMT, Ralph Zazula <zazula@running-start.com> wrote: > The problem is that you are trying to create a window without a > connection to the WindowServer. The Window object is the AppKit > front-end to the DisplayPostscript window operators but it does not > establish this connection for you. The Applicaton object does when it is > run. Here is a modified version of the program that will get your > window to display. thanks; i should have realised that Window would use Application directly. i guess i expected NeXT objects to be written in a robust manner (i.e. to check that external objects are not nil before using them) has anyone found a decent method of intercepting method calls to nil yet ? > Unless you have a particular reason for creating Window objects by-hand, i did, in this case. (i was writing a minimal test program, to check some of AppKit's behaviour) i often find it's easier to write a small single file test program than going through the laborious stages of creating a new project, creating a new interface, creating and parsing new source files, etc, etc, required of a full blown new application. cheers, rog.
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 14 May 1996 09:28:26 GMT Organization: University of Sheffield, UK Message-ID: <4n9jnq$gm3@bignews.shef.ac.uk> References: <4mefbm$931@news.service.uci.edu> <4mjocp$m1l@news.accessone.com> In-Reply-To: <4mjocp$m1l@news.accessone.com> On 05/06/96, Marc Salvatori wrote: > While my background in C is limited, I share your perception that > the learning curve is initially very steep. > Hmm, that wasn't my experience. I started in NEXTSTEP as a novice C programmer, and I think I found it *easier* to pick up Objectice-C as a result. Really all it boils down to is: [anObject aMessage:withArgument]; Tell someone to perform an action with something. Simple, really. > One I have yet to read may appeal to you. It's "Objective-C: > Object-Oriented Programming Techniques" by Lewis J. Pinson and > Richard S. Wiener. It does make some inescapable references to > the interface, but as its title suggests, focuses mostly on the > language. > > Yup, I found that moderatley useful. I suspect NeXT's own book, and indeed the online docs, would now be of greater utility. Best wishes, mmalc. --
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Comparison of JYACC's JAM with NEXTSTEP? Date: Tue, 14 May 1996 14:57:19 +0100 Organization: Cedar Systems Distribution: world Message-ID: <ipQ47BA$EJmxEwak@cedar.co.uk> MIME-Version: 1.0 If anyone has experience of JAM, could they let me know how it compares with NEXTSTEP. Many thanks, John.
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: NSScanner problemette Date: 14 May 1996 16:38:13 GMT Organization: University of Sheffield, UK Message-ID: <4nactl$qg4@bignews.shef.ac.uk> I'm using an NSScanner to read in some data -- no surprises there... What is confusing (to me at least) is the fact that it's skipping over whitespace when I don't want it to. I have a function which looks like this: BOOL scanInt(NSScanner *scanner, int *buf) { NSString *tmp; if (![scanner scanUpToString:@"\t" intoString:&tmp]) { *buf = 0; return NO; } [scanner scanString:@"\t" intoString:NULL]; if ([tmp isEqualToString:space]) { *buf = 0; return YES; } *buf = [tmp intValue]; return YES; } and data which looks like this: 1 2 3 4 5 but occasionally like this 6 7 9 10 ^ note space character What I was expecting was that the space character would be "read in" and, having detected it, I could simply set the buf value to 0 (which, although inelegant, in this context is correct). Unfortunately the space character seems to be skipped: I even put in an explicit [scanner setCharactersToBeSkipped:newLineSet]; and it's still skipped. I've tried this a different way too (originally using scanUpToCharactersFromSet:tabSet intoString:&tmp -- I haven't tried scanCharactersFromSet:numbersAndSpaceSet intoString:&tmp but I guess I should?), but with no success. This seems really simple, and should be straightforward, so there's probably something blindingly obvious I'm missing? Anyone care to enlighten me? Best wishes, mmalc. --
From: yoda@cis.uni-muenchen.de (Marc Guenther) Newsgroups: comp.sys.next.programmer Subject: Re: any way to stop loadNibSection: from initing all vars? Date: 14 May 1996 17:16:38 GMT Organization: Institut fuer Informatik der Universitaet Muenchen Message-ID: <4naf5m$h9f@arcadia.informatik.uni-muenchen.de> References: <4n7mp5$bi8@core.bard.edu> Cc: hs283@bard.edu In <4n7mp5$bi8@core.bard.edu> Hans Steiner wrote: > I was wondering if there is any way that anyone knows of that will stopp > loadNibSection from initializing all instance variables to zero. I ran > into some annoying complications because awakeFromNib was being sent > twice to my object being loading from a .nib file so I would like to be > able to do thesew things in init without them being zeroed by the NIB It doesn't do that. If I'm not mistaken, what happens is, that objects created by loadNibSection are unarchived, not initialised. This means that +alloc, -read and -awake are called, but not -init. So you have to implement -awake, too. LoadNibSection never manipulates any instance variables, which have no connection. Those with connections are set either by calling -setXXX (another possibly, but documented, trap) if it exists, or by manipulating the variable directly. Your other problem with awakeFromNib being called twice is probably caused by some undocumented peculiarity (I call it a bug) of the loadNibSection code. awakeFromNib should only be sent to the objects that were newly created, but its also sent to the File's Owner of the NIB. Now this File's Owner already existed, it wasn't created in the process, and got no awake or read recently, but it gets an awakeFromNib. Rather strange. If your File's Owner was created by some other nibfile, it gets that message twice. So you have to check for that. have fun, Marc -- Marc Guenther ---------------------------------------------------- Centrum fuer Informations und Sprachverarbeitung University of Munich, Germany NEW ADDRESS AND PHONE: Oettingenstr.67 | phone: +49 89 2178 2724 / fax 2701 80538 Muenchen | email: yoda@cis.uni-muenchen.de ----------------------------------------------------
From: yoda@cis.uni-muenchen.de (Marc Guenther) Newsgroups: comp.sys.next.programmer Subject: Re: code to return ip address machine Date: 14 May 1996 17:36:50 GMT Organization: Institut fuer Informatik der Universitaet Muenchen Message-ID: <4nagbi$h9f@arcadia.informatik.uni-muenchen.de> References: <4mvue8$snk@news.fnbc.com> Cc: ericw@cmg.fcnbd.com In <4mvue8$snk@news.fnbc.com> Eric Wespestad wrote: > Howdy, > > Does anyone have a routine that just returns the ip address of the current > machine (and hides the system call details), either as an unsigned long, or > struct in_addr, either will do. Please share this if you have it, so I don't > have to write it :-) First of all, machines DON'T have ip addresses ! Only network interfaces do. All NeXT machines already have two ip-adresses, 127.0.0.1 (the loopback if) and the address of the ethernet if. Some Intel machines don't even have ethernet. You probably want the address of the ethernet device. I can't halp you with that, but I can tell you how to find the address of the device, through which an existing connection is routed. Assume you create a socket sock = socket(PF_INET,SOCK_STREAM,0); and connect it to some given address if( (connect(sock, (struct sockaddr *)&addr, sizeof(addr)) < 0)) error... you can get the address of the local interface (and it's ip number) len = sizeof (dataaddr); if (getsockname(sock, (struct sockaddr *)&dataaddr, &len) < 0) error... Well, now that I think of it, you could also do a gethostbyname() of your own hostname (available via hostname() ), and you get some ip number in the returned struct hostent. Marc -- Marc Guenther ---------------------------------------------------- Centrum fuer Informations und Sprachverarbeitung University of Munich, Germany NEW ADDRESS AND PHONE: Oettingenstr.67 | phone: +49 89 2178 2724 / fax 2701 80538 Muenchen | email: yoda@cis.uni-muenchen.de ----------------------------------------------------
From: haverber@winternet.com (Bill Haverberg) Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc Subject: Friend on a tight budget needs help with black broken keyboard - he can't post either Date: Tue, 14 May 1996 00:27:02 GMT Organization: StarNet Communications, Inc Message-ID: <4nb81m$mdl@blackice.winternet.com> A friend of mine has a black-hardware next and has a rather serious problem with his mouse (doesn't work) and his keyboard (all y keystrokes map to the number 7). He needs 1) some software that will let him remap his keyboard, or 2) a REAL CHEAP replacement for the mouse and/or keyboard. The keyboard problem is preventing him from accessing the newsgroups so I promised to do so for him...I know absolutely zilch about next except that its supposed to be one of those really cool power systems or something like that (Bill Gates didn't write it so that is one point in its favor :) Anyways, apparently this is the kind of hardware where he can't just walk into a goodwill and grab one off a junked PC or apple. Oh, and if there is software that lets him map his mouse to the keyboard that would be really neat as well. I am going to make an effort to keep up with this newsgroup, but since its not my area of expertise or interest I probably won't be keeping up with it everyday. I would appreciate it if any respondants would EMail me at haverber@winternet.com and I will forward whatever I get. Thanks :)
From: marcos@imana.com (Marcos Javier Polanco) Newsgroups: comp.sys.next.programmer,gnu.gcc.help Subject: Compiling without the -g flag; C++ template code Date: 15 May 1996 00:58:34 GMT Organization: TLGnet Inc., (formerly The Little Garden) Message-ID: <4nba7q$o8@kadath.zeitgeist.net> Hello, Note: I'm using gcc within Nextstep's ProjectBuilder. I'm compiling C++ code with templates, and it is just killing gcc. This is what the ProjectBuilder output looks like: ------------------------------------------------ cc -g -Wall -ObjC++ -I/usr/include/g++ -DDEBUG -I./sym -arch i386 -I. -c Bond.m -o ./i386_debug_obj/Bond.o tnPrimitives.h: In function `int tnSizeAsString(long unsigned int)': In file included from tnServices.h:66, from Bond.m:5: tnPrimitives.h:8: warning: return-type defaults to `int' tnPrimitives.h: In function `int tnSizeAsString(long int)': tnPrimitives.h:11: warning: return-type defaults to `int' /usr/tmp/cc654094-011042.s:unknown:Undefined local symbol LBB160 /usr/tmp/cc654094-011042.s:unknown:Undefined local symbol LBE160 *** Exit 1 Stop. *** Exit 1 Stop. ------------------------------------------------ The problem goes away if I compile without the -g flag. So I have two requests: one is help with getting rid of the compiler problem in the first. If short of that, then I'd like to know how to teach ProjectBuilder to compile without the -g flag. Thanks! -- ========================================================== Marcos J. Polanco Imana, LLC 415.436.9700 (voice) 480 Potrero Ave. 415.436.9793 (fax) San Francisco, CA 94110 marcos@imana.com http://www.imana.com/ ==========================================================
From: trecom1@adams.patriot.net (Trecom Business Systems) Newsgroups: comp.sys.next.programmer Subject: US-VA/DC(metro area)/NC-MULTI SOFTWARE,TELECOM,DBA,more... Date: 15 May 1996 03:04:58 GMT Organization: PatriotNet, (703) 277-7737 Message-ID: <4nbhkq$k4g@washington.patriot.net> IMMEDIATE MULTIPLE FULL-TIME OFFERINGS: In Northern Virginia and Research Triangle, North Carolina with: TRECOM Business Systems Washington, DC Office 8521 Leesburg Pike, Suite 405 Vienna, VA 22182 http://www.trecomdc.com FAX (703) 883-2019 PHONE (703) 883-2012 ----------------------------------------------------------------- Northern Virginia Washington,DC-Metro Area All Northern Virginia positions are in the Tysons Corner/Dulles Airport corridor. *OOA/D Architects(Enterprise wide project) *Telecommunications Specialists with extensive ISDN experience. *Network/Telecommunications Analysts with strong communication and writing skills *Systems Analysts *SYBASE *UNIX *C/C++ *NeXT *COBOL/IMS ---------------- Cary, North Carolina Research Triangle Area *Network Manager *Network Architects *Telecommunications Analysts with TELCO switch experience *System Architects *System Analysts *ISDN Analysts *ORACLE related positions *UNIX/C/C++ development and programming positions We are looking for information systems professionals with expertise in a wide variety of technologies. You may send us your resume as an attachment via e-mail to jobs@trecomdc.com, fax (703) 883-2019 or send your resume to: Director of Recruiting TRECOM Business Systems, Inc. 8521 Leesburg Pike, Suite 405 Vienna, VA 22182 FAX (703) 883-2019 PHONE (703) 883-2012 For more information about TRECOM Business Systems, take a look at a our Web page at http://www.trecomdc.com -- --- ____________________________________________________________________________ TRECOM Business Systems, Inc. http://www.trecomdc.com 8521 Leesburg Pike, Suite 405 mailto:info@trecomdc.com Vienna, VA 22182 ____________________________________________________________________________
From: rhayden@next.com (Ronald Hayden) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 15 May 1996 04:13:45 GMT Organization: NeXT Software, Inc. Message-ID: <4nbllp$rud@news.next.com> References: <4mefbm$931@news.service.uci.edu> In article <4mefbm$931@news.service.uci.edu> bkmoore@uci.edu (Brian Kevin Moore) writes: > I'm looking for any good resources on Objective C programming. You might check out NeXT's well-received "Object-Oriented Programming and the Objective-C Language", available online with the product or at http://www.next.com/Pubs/Documents/OPENSTEP/ObjectiveC/objctoc.htm. This is NOT a book about NeXT's GUI -- it is purely about programming in Objective-C, and understanding object-oriented programming. -- Ron Hayden, Manager of Technical Publications
From: zazula@running-start.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 15 May 1996 05:05:34 GMT Organization: Running Start, Inc. Message-ID: <4nbomu$oge@nntp1.best.com> References: <4mefbm$931@news.service.uci.edu> <4mjocp$m1l@news.accessone.com> <4n9jnq$gm3@bignews.shef.ac.uk> In-Reply-To: <4n9jnq$gm3@bignews.shef.ac.uk> On 05/13/96, mmalcolm crawford wrote: >Yup, I found that moderatley useful. I suspect NeXT's own book, and indeed >the online docs, would now be of greater utility. > NeXT's Objective-C reference can be found on-line at: http://www.next.com/Pubs/Documents/OPENSTEP/ObjectiveC/objctoc.htm This is what much of the training materials were based on - very good reading. Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@running-start.com http://www.running-start.com http://www.running-start.com/zazula/index.html
From: hnalgae@soback.kornet.nm.kr (SungJin Kang (kornet)) Newsgroups: comp.sys.next.programmer Subject: What is .lepf format? (NEXTSTEP font) Date: 15 May 1996 06:12:48 GMT Organization: KORNET (Korea Telecom) Message-ID: <4nbsl0$mrb@usenet.kornet.nm.kr> I have some question about NEXTSTEP font. 1. Can I use the font which was made in only bitmap font? If I want to add my new font, should I add bitmap font with outline font? I wanted to add my new bitmap font in .bepf(BDF format). So I added it, and I could see that it was registered in Font Panel, but I could not use that. Of course, I added AFM file and did some unix command. 2. What format is .lepf? I searched many kinds of books about PostScript and many web sites, but I couldn't find about .lepf. What is that? Thanks for reding this. Hayan Nalgae hnalgae@soback.kornet.nm.kr
From: privacy@interlink-bbs.com Newsgroups: comp.sys.next.programmer Subject: New Anonymous Remailer Date: Wed, 15 May 1996 07:05:16 GMT Message-ID: <9605141436.D7532wg@interlink-bbs.com> You may be familiar with anon.penet.fi, which give you an anonymous account. Our service allows YOU to choose what the return address will be! Please write for more info.
Sender: privacy@interlink-bbs.com From: jem@xpat.com Message-ID: <cancel.9605141436.D7532wg@interlink-bbs.com> Control: cancel <9605141436.D7532wg@interlink-bbs.com> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <9605141436.D7532wg@interlink-bbs.com> Date: 15 May 1996 07:39:37 GMT Cancelled by jem@xpat.com. 832145977 IneterlinkBBS Original Subject was: New Anonymous Remailer
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Remote Execution Date: 15 May 1996 04:46:16 GMT Organization: No organisation supplied Message-ID: <4nbnio$7kj@hermes.is.co.za> Hi, I want to execute a program remotely. I use the rexec function call and it works fine provided I pass the user name and password in the call. The problem I am facing is that I can't use the logged on users name and password as this is most probably going to be root. I want to execute the remote program as a specific user for example "remoteuser". In order to do this I need to hard code the users password into my call which I don't want to do as the password might change. I did a setuid on the program to be executed and changed its owner to "remoteuser", the problem with this though is that the program does not use the "remoteuser"'s environment. Could anybody give me some pointers to what I should do? Basically what I need to do is execute a program remotely with a specific userid so that this program can pick up its environment correctly. I don't know what this user's password is as it might have changed since I wrote the program. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: Alex Kolesov Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: HELP!!! Sybase/MS SQL DB-Library headers Date: 15 May 1996 07:43:16 GMT Organization: TverUniversal Bank Distribution: world Message-ID: <4nc1uk$k9d@hobbit.tubank.msk.su> Keywords: SQL Hi ! I apologize for using these newsgroups for my post, but folks from comp.databases.<blah-blah-blah> seems not to be concern to help anyone who needs help. Nothing strange - it is _not_ a NEXTSTEP community ... Another world, another rules ... Ok, my problem is: I wrote a lot of code using DB-Library for Sybase on my NeXT and now I have to port it to Windoze somehow to work with MS SQL Server (I don't want to, but ..). Using MS SQL DB-Library seems to be a very attractive solution. Unfortunately my distribution disk for MS SQL does not contain it (and says that it does). I found DLL with DBLib API, but there's no C headers available for that stuff. I've tried to use my sybase ones - it didn't work. So if somebody has these .h files for MS or Borland C compilers (sqlfront.h and sqldb.h), please, send'em to me via e-mail. Thanks a lot in advance for any help. Be well ... ------------------------------------------------------------ Alex Kolesov Moscow, RUSSIA Tver Universal Bank Nostro Accounts Department Phone : 7+(095)-432-9483 NeXT, MIME & Microsoft E-mail: <alex@next.tubank.msk.su> mail accepted ------------------------------------------------------------
Date: 15 May 1996 09:11:37 GMT From: jem@xpat.com Newsgroups: comp.sys.next.programmer Message-ID: <cancel.4nbhkq$k4g@washington.patriot.net> Control: cancel <4nbhkq$k4g@washington.patriot.net> Sender: trecom1@adams.patriot.net (Trecom Business Systems) Subject: cmsg cancel <4nbhkq$k4g@washington.patriot.net> EMP/ECP (aka SPAM) cancelled by jem@xpat.com. See news.admin.net-abuse.announce, report 19960515.09 for further details
From: IHKONG@chollian.dacom.co.kr (õ¸®¾È NEWS GROUP ÀÌ¿ëÀÚ) Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc Subject: Re: Friend on a tight budget needs help with black broken keyboard - he can't post either Followup-To: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc Date: 15 May 1996 10:28:22 GMT Organization: DACOM Internet Message-ID: <4ncbk6$16o@nis.dacom.co.kr> References: <4nb81m$mdl@blackice.winternet.com>
From: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: NSCalendarDate Date: 15 May 1996 12:58:17 GMT Organization: Universite de La Rochelle Message-ID: <4nckd9$6gu@hpuniv.univ-lr.fr> Hello, I have a little problem with a Time field: I create a column in my Informix database which holds only a time value. the type is DATETIME HOUR TO MINUTE (a time like 14:23) In EOModeler I set the NSDate format to %H:%M but When I fetch, the adaptor tell me It can't convert the data to that format. When I try to do [NSCalendarDate dateWithString:@"8:00" calendarFormat:@"%H:%M"] There is an error : Uncaught exception: NSInvalidArgumentException. return code = 100003. incomplete date specification; missing year, month, or day How Can I use time field ? thanks for our help. /*************************************** * Wilfrid Gaboriaud * Service Informatique * IUT La Rochelle * 15 rue de Vaux de Foletier * 17026 La Rochelle cedex * France * Phone. (33) 46 51 39 24 * Fax. (33) 46 51 39 39 * e-mail: wgaboria@iut-lr.univ-lr.fr ***************************************/
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Learning Objective C Message-ID: <DrG7DF.Br1@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4mefbm$931@news.service.uci.edu> <4nbllp$rud@news.next.com> Date: Wed, 15 May 1996 13:12:50 GMT In article <4nbllp$rud@news.next.com> rhayden@next.com writes: >In article <4mefbm$931@news.service.uci.edu> bkmoore@uci.edu (Brian Kevin >Moore) writes: >> I'm looking for any good resources on Objective C programming. > >You might check out NeXT's well-received "Object-Oriented Programming and >the Objective-C Language", available online with the product or at >http://www.next.com/Pubs/Documents/OPENSTEP/ObjectiveC/objctoc.htm. > In addition to this NeXT specific book, it's also a good idea to take up the book of the author of Objective C, or to read a book on Smalltalk, to keep the design e.g. of the original Collection classes in mind. There's also a web-site on Objective C, maintained by Steve Dekorte, whcih can provide pointers to all this. (http://www.batech.com/~dekorte, I believe) David Stes Computer Algebra Objects e-mail: stes@can.nl www: http://www.can.nl/~stes
From: mpaque@pbinet.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Wed, 15 May 1996 15:07:23 GMT Organization: Pacific Bell Internet Services Message-ID: <4ncs66$k2j@nntp.snfc21.pbi.net> References: <4n0bss$s58@news4.digex.net> John Kheit <jkheit@cnj.digex.net> wrote: > I'm wondering what is the future of sound under OPENSTEP? I'm >guessing that the soundKit will continue to be available and work under >OPENSTEP for Mach. However, I wonder if it will be available, or it's >functionality, under other OPENSTEP implementations? The SoundKit is not part of the OPENSTEP specification. > In other words, if I port my application to OPENSTEP for Mach, and >then later wish to port it to OPENSTEP for NT, will it work under NT with >sound? (I'm hoping that somekind of sound facility is available through >the Postscript server layer of OPENSTEP). Well, the NeXT playsound DPS extension makes the speaker beep. This may not be sufficient for your needs, however... Mike Paquette I don't speak for my employer, and they don't speak for me. mpaque@pbinet.com Personal E-mail mpaque@next.com NeXT business mail only, please
From: ash@now.ch (Michael Aeschbach) Newsgroups: comp.sys.next.programmer Subject: Custom Data Types for EOModeler Date: Tue, 14 May 1996 10:41:38 GMT Organization: NOW GmbH, Baar, Switzerland Sender: news@now.ch Message-ID: <DrE5pE.DF7.0.astra@now.ch> Hello, could anyone tell me how to use a Custom Data Type for EOModeler? I created a subclass of NSObject (EOCustomValues informal protocol). But how can I use this in EOModeler with the Attribute Inspector? Many thanks for your help. Michael
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9605151834.AA12937@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Wed, 15 May 96 13:34:33 -0500 Subject: Simple Animation Timing Hi all, I want to place a Netscape-style-gleaming-company-logo-doohickey on a panel. To that end, I've been fiddling with a raytracer (not Renderman, sorry...don't have adequate docs, so I'm using a NEXT compile of the excellent POVray instead). I've had good success in generating an appropriate series of 50-100 small (64-pixel-square) TIFFs of a 3-D logo with an incrementally moving light source. I'm calling a DPSTimedEntry to simply composite the frames, one after the other. I'm reasonably pleased with the results, but I'd like to smooth things out a little. Given that DSPTimedEntry frequency may vary from the requested interval, my animation gets noticeably jerky at times. Right now, I have the DPSTimedEntry set to run every 1/30 of a second. I'm trying to figure a strategy to ensure that a new frame does happen that often. I recall reading somewhere that my sampling rate needs to be at least twice the expected event rate for best results, so I take it I need to set the timed entry to fire sixty times a second, and determine for myself when I've entered a new thirtieth-of-a-second interval. All of that was to ask this: how do I interpret the "now" argument in the timed entry? I take it that's the key to figuring out if it's really time to composite a new frame, but the docs only indicate that the argument is "the number of seconds since...some arbitrary time in the past", or somesuch. Many thanks, and if anyone has any other hints on polishing an animation, I'd certainly welcome those, too. Regards, - Dan
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9605151919.AA13027@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Wed, 15 May 96 14:19:23 -0500 Subject: Speaking of animation... ....does anyone know of a good TARGA-to-TIFF converter that can be called from the command line? I'm having to hand-convert the TARGA output from POVray frame by frame, which I'd rather not. - Dan
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 15 May 1996 22:19:55 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4ndlab$325@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> mpaque@pbinet.com (Mike Paquette) wrote: > Well, the NeXT playsound DPS extension makes the speaker beep. This may not be sufficient for your needs, however... Aye! I guess it's clear that NeXT does not want to provide development tools that might help the game makers develop games. This is sad, b/c it would be a definate market, and a high profile one at that... -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
Sender: IHKONG@chollian.dacom.co.kr (õ¸®¾È NEWS GROUP ÀÌ¿ëÀÚ) From: jem@xpat.com Message-ID: <cancel.4ncbk6$16o@nis.dacom.co.kr> Control: cancel <4ncbk6$16o@nis.dacom.co.kr> Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc Subject: cmsg cancel <4ncbk6$16o@nis.dacom.co.kr> Date: 16 May 1996 1:33:7 GMT Spam, local newbie mistake, or other rubbish removed by jem@xpat.com
From: ronnie@lanl.gov (Ronnie Mainieri) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Re: How can I add undo to my program? Date: 16 May 1996 03:12:21 GMT Organization: Complex Systems Group, LANL Message-ID: <4ne6el$a34@newshost.lanl.gov> References: <199605020233.WAA09449@slate.cs.rochester.edu> In the text editor sam by Rob Pike a very elegant approach is taken to undo. He looked at the text file as database of characters. There are two basic operations that affect the file: insert and remove. All other modification commands are implemented in terms of these two operations. These two commands are the transactions issued to the database. As the file is modified, another file is kept with all the transactions. To undo one just has to replay the transactions backwards with a few shifts of the addresses used. Ronnie Mainieri ronnie@cnls.lanl.gov
From: yannick@uranus.univ-lr.fr (Yannick Buisson) Newsgroups: comp.sys.next.programmer Subject: mail and bug report Date: 16 May 1996 09:36:58 GMT Organization: Universite de La Rochelle Message-ID: <4nesvq$eb2@hpuniv.univ-lr.fr> Hi all, I want to implement a big report in my applications and i'm looking for the best way to do it !!! So can anyone tell me the best way to it ? Do some classes exist to do it? How can i prepare and deliver automatically a mail ? Do i use the unix mail and senmail commands ? Any help would be appreciate thanx for your answers yannick
From: tony@gst1.gestel.it (Antonio Flores) Newsgroups: comp.sys.next.programmer Subject: Wacom tablet driver Date: 15 May 1996 11:37:46 GMT Organization: Italia Online Message-ID: <4ncfma$b66@mikasa.iol.it> Is there a Wacom tablet driver available? Thanks for help. -------------- Antonio Flores tony@gestel.it
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: Career Position/NEXT/Developer/ILL Date: 16 May 1996 14:28:30 GMT Organization: Norden 1 Communications Message-ID: <4nfe2e$qdt@tofu.alt.net> Programmer/analyst/developer NEXTSTEP---------------NEXTSTEP------------------NEXTSTEP Programmer/analyst/developer------------Commercial experience NEXTSTEP--------------------Commercial experience Objective C-----------------Commercial experience EOF-------------------------A Plus Sybase or Oracle------------A Plus Career Position-----Outstanding benefits & working conditions. Relocation----------Company assistance Must Be-------------US Citizen or Greencard To be considered----Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <copernicus.sccsi.com!tappd@nuchat.sccsi.com> Message-ID: <9605161408.AA15181@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@copernicus.sccsi.com> Date: Thu, 16 May 96 09:08:36 -0500 Subject: Re: Simple Animation Timing Cc: copernicus.sccsi.com!antigone.com!comp-sys-next-programmer@nuchat.sccsi.com References: <199605161227.HAA21606@subzero> Mr. Hess, I appreciate the reply! Next had sent me a snippet which used one of the techniques you describe (i.e. placing all the frames in a single rectangle and compositing subranges from that). I set that aside for a bit because I wasn't sure I was smart enough about images to prep that type of image file. Instead, my little animation controller makes its first iteration through a list of images in an off-screen view, and then calls a delegate when the animation is ready to move onscreen. Maybe I need to revisit your way; it sounds simpler, if I'll put together a tool to composite a series of images into a rectangle. I know it would definitely please Project Builder more to handle a single TIFF file...it tends to drop things off its background command-line processing when I hand it over 30 images at a whack. One comment you made compels me to (again) share my inexperience in NEXT...you said I should keep the CPU utilization below about 30%. I'm not sure what I can do to control that, other than to fiddle with the frequency of the animation. Oh, and thanks much for the sample code on reconciling "now" with the animation start time...it's a help! Regards, - Dan
From: griffith@crl.com (Dave Griffith) Newsgroups: comp.sys.next.programmer Subject: Looking for Phone Interface Software/Hardware Date: 16 May 1996 09:24:08 -0700 Organization: The Hermetic and Esoteric Order of Clarity and Openness Message-ID: <4nfkr8$9r7@crl.crl.com> We're looking to extend our trading applications with automatic paging or calling, i.e. if july eurodollars hit 94.25 according to Reuters, page number 555-1234. Now there are any number of turnkey DOS products that would do this for us, but being a heavy NeXT shop with a big investment in onsite development, we were looking for something a bit more flexible. At best, it would be something like the late lamented PhoneKit was supposed to be, allowing our programmers to quickly write or extend custom apps that would do paging or calling. Does anyone know of any such kits, either free or commercial, which work with readily available hardware. -- --Dave Griffith, griffith@crl.com
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Passing ... To another method Date: 16 May 1996 17:26:02 GMT Organization: No organisation supplied Message-ID: <4nfofa$3it@hermes.is.co.za> Hi, How can I achieve this, I have a method defined in one object as follows: - (void)forwardVarArg:(NSString *)format, ...; In this method I want to forward the vararg list to another object that has a method of the following form: - (void)execute:(NSString *)format, ...; How can a forward the variable argument list to another method that also accepts a variable argument list? Any help will be appreciated. --- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Passing ... To another method Date: 16 May 1996 20:32:04 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7loiseafv.fsf@tom.es.ele.tue.nl> References: <4nfofa$3it@hermes.is.co.za> In-reply-to: gvandyk@icon.co.za's message of 16 May 1996 17:26:02 GMT In article <4nfofa$3it@hermes.is.co.za> gvandyk@icon.co.za writes: - (void)execute:(NSString *)format, ...; How can a forward the variable argument list to another method that also accepts a variable argument list? Your best shot is to fire the programmer responsible for writing the `-execute:' method without also writing -(void) execute: (NSString *) format : (va_list) ap; --Tiggr
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: Wacom tablet driver Date: 16 May 1996 22:15:22 GMT Organization: ipc, U of Tuebingen, Germany Message-ID: <4ng9dq$adt@bias.ipc.uni-tuebingen.de> References: <4ncfma$b66@mikasa.iol.it> Cc: tony@gst1.gestel.it In <4ncfma$b66@mikasa.iol.it> Antonio Flores wrote: > > Is there a Wacom tablet driver available? > Yes, Graeme Barnes <graeme@cam-ani.co.uk> did a driver for NS 3.3. It is available on ftp.cs.tu.berlin.de/pub/NeXT/new/hardware/driver/intel/WacomTablet3.36h.b.I.tar.gz. -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
Newsgroups: comp.sys.next.software,comp.sys.next.misc,comp.sys.next.programmer From: Sean Willson <sean_willson@il.us.swissbank.com> Subject: OmniWeb, Netscape, and MIME types Message-ID: <1996May16.214513.19600@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Thu, 16 May 1996 21:45:13 GMT Fellow NeXT Users, I have the following problem, I have several Microsoft Word documents sitting on a web server. I create pages with links to those docs and open those pages in both Netscape and OmniWeb. When I select the link in either of the two applications they both display the file in the browser. In Netscape a new mime-type was defined to open Word, although I am not sure if it was done properly. In OmniWeb, the .doc file has a default application of OpenWrite.app yet in both cases the browser ALWAYS displays the output (mangled obviously) in the browser window instead of spawning an external viewer. Is there some reason for this. It is imperative that doc's sit on the web server and can be served up. Can someone make some suggestions? This seems so basic and straight forward, yet persists!!!! I am sure this problem has been dealt with before.....thanks. Sean Willson -- ____________________________________________________________________ Sean M. Willson sean_willson@il.us.swissbank.com Swiss Bank Corp. NeXT, MIME, and ascii 141 West Jackson Boulevard Chicago, Illinois 60604 (312) 554-5399
From: allanmac@blueprint.com (Allan MacKinnon) Newsgroups: comp.sys.next.programmer Subject: Example source for an EOF adaptor? Date: 16 May 1996 21:19:20 GMT Organization: Channel 1 Communications Message-ID: <4ng64o$jsd@news1.channel1.com> I'm looking for example source code for an EOF _adaptor_. Any pointers? Thanks, Allan -- Allan MacKinnon allanmac@blueprint.com Boston, MA NeXT / MIME Mail Welcome (617) 424-0615
From: Roland Telfeyan <roland@telf.com> Newsgroups: comp.sys.next.programmer Subject: Re: Is there an MS-Windows-like "combo-box" for NEXTSTEP? Date: 17 May 1996 01:51:27 GMT Organization: Vnet Internet Access, Charlotte, NC - info@char.vnet.net Distribution: world Message-ID: <4ngm2v$dkn@ralph.vnet.net> References: <4ne848$o2j@ralph.vnet.net> I know this is was detestable question. Sorry! So how about if I ask it this way: Is there any pop-up list object or subclass that stays popped up the way menus can? -- Roland Telfeyan roland@telf.com
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re: Example source for an EOF adaptor? Date: 17 May 1996 12:53:51 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4nhssv$nq8@stc06.ctd.ornl.gov> References: <4ng64o$jsd@news1.channel1.com> In article <4ng64o$jsd@news1.channel1.com> allanmac@blueprint.com (Allan MacKinnon) writes: > > I'm looking for example source code for an EOF _adaptor_. > Any pointers? > Do a web search for msqlAdaptor, it comes with source. -- - - - - - - - - - J. W. Wooten
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <199605171341.RAA02322@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Fri, 17 May 96 17:41:05 +0400 Subject: mSQL and EOF? Hi, Anybody here use mSQL database? I guess there are a lot of you since mSQL is almost free. :-) Could you please tell me about your experience with it. I want to use it for WebObjects project and have got several questions. I'll greatly appreciate your answers. 1. Since mSQL doesn't have any daytime type how does NSDates are handled by mSQLAdaptor, or NSDates are beyond its scope? 2. Could I have varchar field to be a primary key? 3. What about relationship? Does mSQL support them at all? 4. From mSQL faq I know that it doesn't support multiple primary keys. How could I live with it? I have tables which rows could be uniquely identified only by values of 2-3 field. Thanks. Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Re:mSQL and EOF? Date: 17 May 1996 17:49:00 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4nie6c$dcq@stc06.ctd.ornl.gov> References: <199605171341.RAA02322@cnext.crec.mipt.ru> In article <199605171341.RAA02322@cnext.crec.mipt.ru> writes: > Hi, > > Anybody here use mSQL database? I guess there are a lot of you > since mSQL is almost free. :-) Can't answer your questions yet, but I did convert the Demos of EOF over to using msql. There were a couple of things I had to change slightly because I couldn't perform some operations in msql that could be done in Sybase. > Could you please tell me about your experience with it. I want to > use it for WebObjects project and have got several questions. I'll > greatly appreciate your answers. > > 1. Since mSQL doesn't have any daytime type how does NSDates are > handled by mSQLAdaptor, or NSDates are beyond its scope? > > 2. Could I have varchar field to be a primary key? > > 3. What about relationship? Does mSQL support them at all? > > 4. From mSQL faq I know that it doesn't support multiple primary > keys. How could I live with it? I have tables which rows could be > uniquely identified only by values of 2-3 field. > > Thanks. > Aleksey > _________________________________________________________________ > > Aleksey Sudakov > 1A-1 Kerchenskaya Street > Moscow 113303,Russia. > ________________________________________________________________ > Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 > EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) > WWW: http://www.crec.mipt.ru/~zander/ -- - - - - - - - - - J. W. Wooten
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: mSQL and EOF? Date: 17 May 1996 17:59:17 GMT Organization: University of Sheffield, UK Message-ID: <4niepl$t2s@bignews.shef.ac.uk> References: <199605171341.RAA02322@cnext.crec.mipt.ru> In-Reply-To: <199605171341.RAA02322@cnext.crec.mipt.ru> On 05/17/96, Aleksey Sudakov wrote: > Anybody here use mSQL database? I guess there are a lot of you > since mSQL is almost free. :-) > I've played with it a *little*. > Could you please tell me about your experience with it. I want to > use it for WebObjects project and have got several questions. I'll > greatly appreciate your answers. > It's been vaguely useful as a learning tool. > 1. Since mSQL doesn't have any daytime type how does NSDates are > handled by mSQLAdaptor, or NSDates are beyond its scope? > Same as I use dates in the free WebObjects -- simply store and retrieve as strings in a given format, e.g.: set up date with: formatDate = [NSString stringWithFormat:@"%@ %@ %@", theDay, theMonth, theYear]; [aConference setObject:formatDate forKey:@"date"]; retrieve with: conferenceDate = [NSCalendarDate dateWithString: [aConference objectForKey: @"date"] calendarFormat:@"%d %B %Y"]; I'd be interested to hear others' answers to your other questions. Best wishes, mmalc. --
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 17 May 1996 18:12:21 GMT Organization: University of Sheffield, UK Message-ID: <4nifi5$2p@bignews.shef.ac.uk> References: <4n0bss$s58@news4.digex.net> In-Reply-To: <4n0bss$s58@news4.digex.net> On 05/10/96, John Kheit wrote: > I'm wondering what is the future of sound under OPENSTEP? I'm > guessing that the soundKit will continue to be available and work under > OPENSTEP for Mach. However, I wonder if it will be available, or it's > functionality, under other OPENSTEP implementations? > I guess an interesting question is what's going to happen to MusicKit...? Will there (can there?) be ports to other OSes? If so, then it might offer the most robust solution (probably more robust than ordinary sound on Intel) -- but at the cost of a DSP card. Best wishes, mmalc. --
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <199605171842.WAA11525@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Fri, 17 May 96 22:42:38 +0400 Subject: OpenBase Hi, Anybody here use OpenBase v5.0 database?I have got several questions. I'll greatly appreciate any help. 1. Could I have varchar field to be a primary key? If I could, how could I remove _rowid - default field using OpenBaseManager? If I couldn't how could I make OpenBase autogenerate unique values for _rowid when I insert objects using EOF? 2. How could I set the primary key other than _rowid using OBM? Is it possible at all? 3. How could I backup relationships (contents of _SYS_RELATIONSHIP table) and users (contents of _SYS_USER table) using OpenBaseManager? The problem is that neither Tools->Backup nor Edit Database Schema window's Save button dump contents of system tables. Thanks. Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <199605172026.AAA11830@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Sat, 18 May 96 00:26:31 +0400 Subject: Re: mSQL and EOF? In article <4niepl$t2s@bignews.shef.ac.uk>, mmalc wrote: > On 05/17/96, Aleksey Sudakov wrote: > > 1. Since mSQL doesn't have any daytime type how does NSDates are > > handled by mSQLAdaptor, or NSDates are beyond its scope? > > > Same as I use dates in the free WebObjects -- simply store and retrieve as > strings in a given format, e.g.: > > set up date with: > > formatDate = [NSString stringWithFormat:@"%@ %@ %@", > theDay, theMonth, theYear]; > [aConference setObject:formatDate forKey:@"date"]; > > retrieve with: > > conferenceDate = [NSCalendarDate dateWithString: > [aConference objectForKey: @"date"] > calendarFormat:@"%d %B %Y"]; But ain't it just what msql EOF Adaptor supposed to do? Regards, Aleksey. _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303,Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
Newsgroups: comp.sys.next.programmer From: schalle@smiles.cen.encompass.com (Dirk Schalle) Subject: Re: why does this program core dump ? Message-ID: <1996May17.182203.19855@glv.cen.encompass.com> Sender: news@glv.cen.encompass.com Organization: Encompass References: <4n7jno$e00@netty.york.ac.uk> Date: Fri, 17 May 1996 18:22:03 GMT In article <4n7jno$e00@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) writes: > the following program gives me a bus error (null > argument to strtod() from +[FrameView initialise]) > > has anyone any idea why ? > > cheers, > rog. > > #include <appkit/appkit.h> > > main() > { > id win; > NXRect fr = {0, 0, 100, 100}; > > win = [[Window alloc] initContent:&fr style:NX_PLAINSTYLE > backing:NX_RETAINED buttonMask:0 > defer:NO]; > return 0; > } .. there is a bunch of things missing ... .. .. basically it's the Display PostScript context ... .. but also an event loop ... .. .. what you might want to do is ... .. create an application before creating the window ... .. and run the application after that ... .. that's the easiest way ... .. if you don't want to deal with Display PostScript directly ... .. so your code should look like ... #include <appkit/appkit.h> main() { id win; NXRect fr = {0, 0, 100, 100}; // implicitly create a DPS context ... [Application new]; win = [[Window alloc] initContent:&fr style:NX_PLAINSTYLE backing:NX_RETAINED buttonMask:0 defer:NO]; // run the event loop ... [NXApp run]; return 0; } .. --- ... Dirk
From: murphy@wsc.com (Paul Murphy) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Help wanted Date: 17 May 1996 18:07:51 -0400 Organization: WSC Sender: murphy@eos Message-ID: <x4u3xfhs20.fsf@eos> We are looking for a programmer to work on an OpenStep project. WSC is a NYC-based financial services and technology company. Anyone applying for this job should be interested in product development. Having lived through a product lifecycle is a plus. Knowledge of NEXTSTEP or Objective-C is required. Also required: strong basic CS skills, knowledge of Unix, intelligence, and willingness to live in NYC. If interested, please email resume and salary requirement to murphy@wsc.com.
From: Marc@AccessOne.com(Marc Salvatori) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 18 May 1996 02:13:53 GMT Organization: AccessOne Message-ID: <4njbp1$hfi@news.accessone.com> References: <4mefbm$931@news.service.uci.edu> <4mjocp$m1l@news.accessone.com> <4n9jnq$gm3@bignews.shef.ac.uk> Cc: m.crawford@shef.ac.uk In <4n9jnq$gm3@bignews.shef.ac.uk> mmalcolm crawford wrote: > Hmm, that wasn't my experience. I started in NEXTSTEP as a novice C > programmer, and I think I found it *easier* to pick up Objectice-C as a > result. Really all it boils down to is: > [anObject aMessage:withArgument]; > > Tell someone to perform an action with something. True indeed if all you wish to do is focus on programming. As a beginner, Objective-C comes across to me more as a message passer than a language. I find it easier to consider the design approach if I set aside all past paradigms. -- >< Marc J. Salvatori | >< >< mailto:salvo@accessone.com | MIME & NeXTMail are accepted ><
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 17 May 1996 23:57:08 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4nj3ok$6ds@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4nifi5$2p@bignews.shef.ac.uk> mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > On 05/10/96, John Kheit wrote: > > I'm wondering what is the future of sound under OPENSTEP? I'm > > guessing that the soundKit will continue to be available and work under > > OPENSTEP for Mach. However, I wonder if it will be available, or it's > > functionality, under other OPENSTEP implementations? > > > I guess an interesting question is what's going to happen to MusicKit...? > Will there (can there?) be ports to other OSes? If so, then it might offer the most robust solution (probably more robust than ordinary sound on Intel) > -- but at the cost of a DSP card. That's a cheap price to pay... But my gut tells me there is little hope :( -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: lutzray@9bit.qc.ca (Raymond Lutz) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: WebCopy perl script -> anyone ran it on NS? Date: 18 May 1996 19:13:07 GMT Organization: SPC Message-ID: <4nl7g3$sr9@wagner.spc.videotron.ca> Hi all Hmmm... I know nothing about perl (neither camels) and I can't run this script that recursively slurps html pages.. This my first attempt at using perl... I installed it successfully (that's what Installer.app told me 8^) but it chocks on an include statement when running webcopy script: Can't locate sys/socket.ph in @INC (did you run h2ph?) In the FAQs they talk about h2xs, but I can't find it in the TipTop binary distribution... Should I fetch it somewhere? And what does this script do, anyway? Is this the relevent group? How OpenStep runtime for NT will be priced? Where can I buy a hardware compression board for my ND? Can I use a NeXT Laser printer with a PC? CDPlayer doesn't work, what should I do? And what about dimming NeXT Mono megapixel monitors? Thanx for your answers... 8^) RL
From: Robert La Ferla <Robert_La_Ferla@hot.com> Newsgroups: comp.sys.next.programmer Subject: Java Date: 18 May 1996 22:49:28 GMT Organization: HTI Message-ID: <4nlk5o$lie@news.cais.com> Is there a Java compiler/interpreter/JDK for NEXTSTEP? Robert La Ferla Registered OPENSTEP Consultant Boston, MA + 1 (617) 252-0088
From: GWILLEM@alpha.ntu.ac.sg (Van Schaik Willem Anthon Johan ) Newsgroups: comp.sys.next.programmer Subject: Re: Speaking of animation... Date: 19 May 1996 05:33:18 GMT Organization: Nanyang Technological University Message-ID: <4nmbqu$gtk@ntuix.ntu.ac.sg> References: <9605151919.AA13027@copernicus> Dan Tapp (tappd@copernicus.sccsi.com) wrote: : ....does anyone know of a good TARGA-to-TIFF converter that can be : called from the command line? I'm having to hand-convert the TARGA : output from POVray frame by frame, which I'd rather not. : - Dan Have a look at the pbmplus/netpbm toolset. willem gwillem@ntuvax.ntu.ac.sg
From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Subject: ld: can't locate file for: -lX11 Date: 20 May 1996 05:41:48 GMT Organization: University of California, Los Angeles Message-ID: <4np0ms$1cfo@uni.library.ucla.edu> Hello everyone, could someone please help me with the following error. ld: can't locate file for: -lX11 I get this error alot when trying to compile unix or X11 windows programs under my NeXT (Black). Can someone please help me. Rick vazquezr@physics.ucla.edu
From: pbrown@palms.berkeley.edu (Paul Robert Brown) Newsgroups: comp.sys.next.programmer Subject: grrr... _strdup Date: 20 May 1996 07:29:11 GMT Organization: U.C. Berkeley Math. Department. Message-ID: <4np707$cht@agate.berkeley.edu> Anyone have a couple of lines of code for strdup() sitting around? I found it once ("Handbook of C Functions" or something like that), but I can't find it now... I can't figure why NeXT'd go to the trouble of defining hypot() in math.h but leave out strdup... Many thanks. Paul
From: droege@informatik.uni-koblenz.de (Detlev Droege) Newsgroups: comp.sys.next.programmer Subject: Re: grrr... _strdup Date: 20 May 1996 09:24:33 GMT Organization: University Koblenz / Germany Distribution: world Message-ID: <4npdoh$h4i@newshost.uni-koblenz.de> References: <4np707$cht@agate.berkeley.edu> In article <4np707$cht@agate.berkeley.edu> pbrown@palms.berkeley.edu (Paul Robert Brown) writes: > Anyone have a couple of lines of code for strdup() sitting around? I > found it once ("Handbook of C Functions" or something like that), but I > can't find it now... I can't figure why NeXT'd go to the trouble of > defining hypot() in math.h but leave out strdup... Try something like #define strdup(s) NXCopyStringBuffer(s) (won't work for hypot() :-) ) Detlev -- Detlev Droege, Uni Koblenz, FB Informatik, Rheinau 1, D-56075 Koblenz, Germany Tel:+49 261 9119-421,Fax:-497,NeXT/MIME/Email:droege@informatik.uni-koblenz.de C++ is the only current language making COBOL look good. --Bertrand Meyer
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: ld: can't locate file for: -lX11 Date: 20 May 1996 12:09:58 GMT Organization: ipc, U of Tuebingen, Germany Message-ID: <4npnem$rsh@bias.ipc.uni-tuebingen.de> References: <4np0ms$1cfo@uni.library.ucla.edu> Cc: vazquezr@physics.ucla.edu In <4np0ms$1cfo@uni.library.ucla.edu> Rick Vazquez wrote: > Hello everyone, could someone please help me with the following error. > > ld: can't locate file for: -lX11 > > I get this error alot when trying to compile unix or X11 windows programs > under my NeXT (Black). > Well, NeXTSTEP is not X, so this lib is not there. Install the free MouseX (or the shareware Xnext) or a commercial X package and it should work. But be prepared: you cannot run X11 programs direct under NS, you have to start a XServer before. Have a nice day Frank -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
From: bruno@gruick.univ-lr.fr (Bruno Garnier) Newsgroups: comp.sys.next.programmer Subject: editing textField Date: 20 May 1996 13:11:29 GMT Organization: Universite de La Rochelle Message-ID: <4npr21$7k9@hpuniv.univ-lr.fr> Hi all, I have a problem with editing textfield. My EOController is in BufferEdits mode. After successfully building my project, I run it. When editing a textfield, I modify the display value, valid with enter, remodify the same textfield wih a different value, and another time again. (so 3 times). When I send savetoObject to my EOController, only the second value is saved !!! I don't understand why the third value (that I want to save) is ignored. Please help .... Thanks -- Bruno Garnier Centre de Ressources Informatique Universite de La Rochelle Avenue Marillac 17042 LA ROCHELLE CEDEX 01 Tel : 46 45 82 14 Fax : 46 45 82 45 http://www.univ-lr.fr/ bruno@cri.univ-lr.fr (NeXTMail, MIME Mail) |~~~| |~~~|D Guinness C|. :|D is |: .|D good for you ! |___| \\|// || |^ _| || ( -_+ ) // /_\ O / / __ __ __ || \_/ / | | | | | | ||____/ | | | | | | ---`----- \-- ---| |-| |--| |-- v |==| |==| |==|
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: mail and bug report Date: 20 May 1996 15:16:43 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4nq2cr$dpa@lion.embl-heidelberg.de> References: <4nesvq$eb2@hpuniv.univ-lr.fr> In article <4nesvq$eb2@hpuniv.univ-lr.fr> yannick@uranus.univ-lr.fr (Yannick Buisson) writes: > > Hi all, > > I want to implement a big report in my applications and i'm looking for the > best way to do it !!! > So can anyone tell me the best way to it ? > Do some classes exist to do it? How can i prepare and deliver automatically a > mail ? > Do i use the unix mail and senmail commands ? > > Any help would be appreciate Use MiscKit.... I think the class is called MiscMailApp -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: mSQL and EOF? Date: 20 May 1996 15:22:41 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4nq2o1$dt1@lion.embl-heidelberg.de> References: <199605172026.AAA11830@cnext.crec.mipt.ru> In article <199605172026.AAA11830@cnext.crec.mipt.ru> writes: > In article <4niepl$t2s@bignews.shef.ac.uk>, mmalc wrote: > > On 05/17/96, Aleksey Sudakov wrote: > > > 1. Since mSQL doesn't have any daytime type how does NSDates are > > > handled by mSQLAdaptor, or NSDates are beyond its scope? > > > > > Same as I use dates in the free WebObjects -- simply store and > retrieve as > > strings in a given format, e.g.: > > > > set up date with: > > > > formatDate = [NSString stringWithFormat:@"%@ %@ %@", > > theDay, theMonth, theYear]; > > [aConference setObject:formatDate forKey:@"date"]; > > > > retrieve with: > > > > conferenceDate = [NSCalendarDate dateWithString: > > [aConference objectForKey: @"date"] > > calendarFormat:@"%d %B %Y"]; > > But ain't it just what msql EOF Adaptor supposed to do? > You're right. As far as I can remember, NSDate works just fine. Well, I could imagine that they are storing the date either as a string or as some kind of number. But why you care about it? It works, and that's enough in most of the cases ;-) -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: IB editor for PopUpList's trigger button Date: 20 May 1996 17:06:31 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4nq8qn$f9b@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hello, i am trying to write an IB Palette for a subclass "PopupsView" of View containing trigger buttons for PopUpLists; programmatically all works fine. However, when i select (within IB) a popup's trigger within a "PopupsView" object, i get of course only the Button's inspector corresponding to the popup's trigger. What would be nice however, is to get hold of the real IB's PopUpList editor - the one you get when you pick the standard PopUpList trigger from the standard IB palette's window: with the possibility of directly editing the underlying PopUpList. Any hints? ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: NSUserDefaultsChanged notification Date: 20 May 1996 17:23:40 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4nq9qs$f9b@lion.embl-heidelberg.de> When is the NSUserDefaultsChanged notification sent? The OS spec is really very sparse! Well, I think it's even wrong... or at least not very logical. I'm finishing the GNUstep's default system, and seen that the doc is completely misleading. I think it's logical if the notification is sent every time when the defaults (at least the persistent defaults) are changed (either by the application itself or by any other application). Any (even ND ;-) idea? -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Learning Objective C Date: 20 May 1996 17:39:34 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4nqaom$9p4@nntp.onramp.net> References: <DrG7DF.Br1@cwi.nl> stes@cwi.nl (David Stes) writes: > There's also a web-site on Objective C, maintained by Steve Dekorte, > whcih can provide pointers to all this. (http://www.batech.com/~dekorte, > I believe) It's at: http://www.batech.com/~dekorte/Objective-C/objc.html Steve
From: ronp@sun5.cs.wisc.edu (Ronp) Newsgroups: comp.sys.next.programmer Subject: Re: What is .lepf format? (NEXTSTEP font) Date: 20 May 1996 17:36:42 GMT Organization: U of Wisconsin CS Dept Message-ID: <4nqaja$174@spool.cs.wisc.edu> References: <4nbsl0$mrb@usenet.kornet.nm.kr> In article <4nbsl0$mrb@usenet.kornet.nm.kr>, SungJin Kang (kornet) <hnalgae@soback.kornet.nm.kr> wrote: >I have some question about NEXTSTEP font. >2. What format is .lepf? .lepf is for little-endian machines, where .bepf is for big-endian machines. I haven't been able to find any further documentation. Ron
From: marcos@imana.com (Marcos Javier Polanco) Newsgroups: comp.sys.next.programmer Subject: Compiling GCC 2.7.2 Date: 20 May 1996 21:07:41 GMT Organization: TLGnet Inc., (formerly The Little Garden) Message-ID: <4nqmut$98@kadath.zeitgeist.net> Hello, I am dowloading (as we speak) gcc 2.7.2, presumably the latest release of the GNU compiler. If any of you have compiled it before, any pointers would be most welcome. Thanks. -- ========================================================== Marcos J. Polanco Imana, LLC 415.436.9700 (voice) 480 Potrero Ave. 415.436.9793 (fax) San Francisco, CA 94110 marcos@imana.com http://www.imana.com/ ==========================================================
From: Robert La Ferla <Robert_La_Ferla@hot.com> Newsgroups: comp.sys.next.programmer Subject: Re: grrr... _strdup Date: 20 May 1996 23:45:37 GMT Organization: HTI Message-ID: <4nr071$fg6@news.cais.com> References: <4np707$cht@agate.berkeley.edu> pbrown@palms.berkeley.edu (Paul Robert Brown) wrote: > Anyone have a couple of lines of code for strdup() sitting around? I > found it once ("Handbook of C Functions" or something like that), but I > can't find it now... I can't figure why NeXT'd go to the trouble of > defining hypot() in math.h but leave out strdup... Use NXCopyStringBuffer() from <objc/hashtable.h> Robert La Ferla Registered OPENSTEP Consultant Boston, MA + 1 (617) 252-0088
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: mSQL and EOF? Date: 21 May 1996 09:17:28 GMT Organization: University of Sheffield, UK Message-ID: <4ns1n8$e7@bignews.shef.ac.uk> References: <199605172026.AAA11830@cnext.crec.mipt.ru> <4nq2o1$dt1@lion.embl-heidelberg.de> In-Reply-To: <4nq2o1$dt1@lion.embl-heidelberg.de> > > But ain't it just what msql EOF Adaptor supposed to do? > > <sigh> Bad day.... Sorry. Best wishes, mmalc. --
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <ttakeo@amtec.co.jp> From: ttakeo@amtec.co.jp (tetuya takeo) Message-ID: <9605210615.AA01069@amtec001.amtec.co.jp> Date: Tue, 21 May 96 15:15:57 +0900 Subject: Infrating OmniImageFilter Cc: ttakeo@tcp-ip.or.jp Hi, Steppers. I wrote little code to convert TIFF file into GIF file using pasteboard. The code exploits OmniImageFilter which already installed on my system. Here's my code fragment. ------------------------------------------------------------------ const char *ImageFormatGifType = "image format gif"; void convertFileToGif( const char *filename ) { id pb; NXAtom *ap; BOOL found = NO; NXStream *gifStream; char newname[MAXPATHLEN], tempname[MAXPATHLEN]; char *cp; strcpy( newname, filename ); cp = rindex( newname, '.' ); if ( cp != NULL ) *cp = '\0'; strcat( newname, ".gif" ); pb = [Pasteboard newByFilteringFile:filename]; for ( ap = [pb types]; *ap; ap ++ ) { if ( strcmp( *ap, ImageFormatGifType ) ) { found = YES; break; } } if ( !found ) { NXLogError( "can't convert" ); return; } gifStream = [pb readTypeToStream:ImageFormatGifType]; sprintf( tempname, "%s.tmp", newname ); NXSaveToFile( gifStream, tempname ); NXCloseMemory( gifStream, NX_FREEBUFFER ); remove( newname ); rename( tempname, newname ); [pb free]; } ------------------------------------------------------------------ I'm success to convert files, but after time, my system goes down by shortage of memory. At last I found that memory usage of OmniImageFilter is infrating by each calling this code. Please let me know bad point of above code. Or, should I kill OIF each time? Thanks in advance. --- ttakeo@tcp-ip.or.jp (NeXTmail welcomed!) ttakeo@amtec.co.jp (NeXTmail welcomed!) phone +81-532-63-1253(office), +81-532-55-5517(home)
From: gfa@komsys-pc-gf.ethz.ch (George Fankhauser) Newsgroups: comp.sys.next.programmer Subject: Re: What is .lepf format? (NEXTSTEP font) Date: 21 May 1996 12:15:23 +0200 Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4ns53r$22q@komsys-pc-gf.ethz.ch> References: <4nbsl0$mrb@usenet.kornet.nm.kr> <4nqaja$174@spool.cs.wisc.edu> >In article <4nbsl0$mrb@usenet.kornet.nm.kr>, >SungJin Kang (kornet) <hnalgae@soback.kornet.nm.kr> wrote: >>I have some question about NEXTSTEP font. >>2. What format is .lepf? > >.lepf is for little-endian machines, where .bepf is for big-endian >machines. I haven't been able to find any further documentation. see /usr/bin/prebuild. lepf/bepf is just a binary bitmap font representation, hence the le/be distinction George -- George Fankhauser Swiss Federal Institute of Technology Phone/Fax +41 1 632 7017/1035 Computer Engineering and Networks http://www.tik.ee.ethz.ch/~gfa Laboratory, CH-8092 Zurich mailto:gfa@tik.ee.ethz.ch "Happyness is a journey, not a destination"
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <zander@cnext.crec.mipt.ru> Message-ID: <199605211245.QAA04941@cnext.crec.mipt.ru> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Aleksey Sudakov <zander@cnext.crec.mipt.ru> Date: Tue, 21 May 96 16:45:54 +0400 Subject: EOF and primary keys Hi EOFers, I've got simple question. Could I have EOEntity primary key other than database unique key? I checked the documentation one more time and it's not clear from there. If I could have primary key other than database unique key would EOF ensure that value for that key should be unique or I gotta do it myself? Thanks. Aleksey _________________________________________________________________ Aleksey Sudakov 1A-1 Kerchenskaya Street Moscow 113303, Russia. ________________________________________________________________ Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) WWW: http://www.crec.mipt.ru/~zander/
From: creditweb@neont.com (CreditWeb) Newsgroups: comp.sys.misc,comp.sys.msx,comp.sys.ncr,comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.bugs,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.novell,comp.sys.nsc.32k,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.sys.powerpc.misc,comp.sys.powerpc.tech,comp.sys.prime,comp.sys.proteon,comp.sys.psion,comp.sys.psion.misc,comp.sys.pyramid,comp.sys.ridge Subject: FIX YOUR OWN CREDIT... Date: 21 May 1996 18:44:23 GMT Organization: CreditWeb Distribution: inet Message-ID: <creditweb-2105961449370001@s193.neont.com> DON'T PAY ANYONE TO FIX YOUR CREDIT. FIX IT YOURSELF. No one can do anything to fix your credit that you can't do yourself. That is a fact! The so called "credit clinics" charge consumers hundreds, even thousands of dollars to "fix" credit files. DON'T PAY them... You can repair your own credit. According to most members of the banking industry, negative information can not be removed from a consumers credit file. THEY ARE LYING. We will provide you with the knowledge and tools needed to repair your own credit for a one time LOW price. You will receive a comprehensive step by step instruction set detailing proven, and successful methods for removing negative information from credit reports. Also included are sample letters and tips for nearly every credit situation. These are EXACTLY the same methods that the so called "credit clinics" use. You will be surprised how easy this is, and how little time it takes. Take Action TODAY! Send a check or money order for $6.95, plus $1 shipping and handling to: CreditWEB PO Box 2364 Youngstown, OH 44509-0364
From: dhunt@fnbc.com (David Hunt) Newsgroups: comp.sys.next.programmer Subject: Re: NSScanner problemette Date: 21 May 1996 17:22:23 GMT Organization: First Chicago NBD Corporation, Chicago IL, USA Message-ID: <4nsu4f$ebd@news.fnbc.com> References: <4nactl$qg4@bignews.shef.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit In article <4nactl$qg4@bignews.shef.ac.uk> mmalcolm crawford <m.crawford@shef.ac.uk> writes: > I'm using an NSScanner to read in some data -- no surprises there... > What is confusing (to me at least) is the fact that it's skipping over > whitespace when I don't want it to. According to the EOF 1.1 Foundation documentation on NSScanner Class Cluster, the charactersToBeSkipped method mentions "The default set is the whitespace and newline character set." The whitespace and newline character set includes ' ', '\t' and '\n'. Thus, by default spaces, tabs and newlines are skipped in the NSScanner input string. -- David G. Hunt One First National Plaza Email: dhunt@cmg.FCNBD.com Mail Suite 0581 Phone: (312) 732-5151 Chicago, IL 60670 Fax: (312) 732-1657
Date: 21 May 1996 19:41:59 GMT From: clewis@ferret.ocunix.on.ca (Chris Lewis) Sender: creditweb@neont.com (CreditWeb) Message-ID: <cancel.creditweb-2105961449370001@s193.neont.com> Newsgroups: comp.sys.misc,comp.sys.msx,comp.sys.ncr,comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.bugs,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.novell,comp.sys.nsc.32k,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.sys.powerpc.misc,comp.sys.powerpc.tech,comp.sys.prime,comp.sys.proteon,comp.sys.psion,comp.sys.psion.misc,comp.sys.pyramid,comp.sys.ridge Subject: cmsg cancel <creditweb-2105961449370001@s193.neont.com> Control: cancel <creditweb-2105961449370001@s193.neont.com> CREDITWEB spam cancelled by clewis@ferret.ocunix.on.ca Original Subject: FIX YOUR OWN CREDIT...
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: EOF and primary keys Date: Tue, 21 May 1996 19:34:06 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996May21.193406.18987@seer.demon.co.uk> References: <199605211245.QAA04941@cnext.crec.mipt.ru> In article <199605211245.QAA04941@cnext.crec.mipt.ru> writes: > Could I have EOEntity primary key other than database unique key? I > checked the documentation one more time and it's not clear from > there. You can mnake whatever you like a part of the key. A "real" SQL database won't have a unique key of its own, anyhow :-). > If I could have primary key other than database unique key would > EOF ensure that value for that key should be unique or I gotta do it > myself? EOF 2.0 will be different, but you have to arrange that your EO has a key and that it is unique. Mysterious things start to happen if you don't do this. See the UniqueKey example that comes with EOF for a solution. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: zazula@running-start.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: SUBMISSION: Foundationalized RZBundle Date: 21 May 1996 21:19:53 GMT Organization: Running Start, Inc. Message-ID: <4ntc1p$aj1@nntp1.best.com> Hi - I have placed a foundation compatible version of the RZBundle class on our WWW site: http://www.running-start.com Click on the "Sources" link under "Free Software". RZBundle provides on-demand dynamic loading of bundles and automatic dependancy resolution. This means you can place objects in bundles and have RZBundle load them when you refer to the classes in code. We place virtually all of our classes in bundles during development and statically link them in when shipping. This gives us very short link times and speeds debugging/development. There is a document included with the code that describes the motivation behind this class and the API it provides. Enjoy, Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting zazula@running-start.com http://www.running-start.com http://www.running-start.com/zazula/index.html
Newsgroups: comp.sys.next.programmer From: pascal@burgond.remcomp.fr (Pascal Bourguignon) Subject: Re: Is there an MS-Windows-like "combo-box" for NEXTSTEP? Message-ID: <DrMLrv.1o6@burgond.remcomp.fr> Sender: pascal@burgond.remcomp.fr (Pascal Bourguignon) Organization: P. Bourguignon Informatique References: <4ne848$o2j@ralph.vnet.net> Date: Sun, 19 May 1996 00:09:30 GMT In article <4ne848$o2j@ralph.vnet.net> Roland Telfeyan <roland@telf.com> writes: > Is there an MS-Windows-like "combo-box" GUI object for NEXTSTEP? > > Thanks, > > Roland You may want to use the PopUpList menus.
Newsgroups: comp.sys.next.programmer From: ernst@fritz.snafu.de (Ernst Kloecker) Subject: Re: Comparison of JYACC's JAM with NEXTSTEP? Message-ID: <DrI25F.3oG@fritz.snafu.de> Sender: news@fritz.snafu.de Cc: phef@cedar.co.uk Organization: dasburo Berlin References: <ipQ47BA$EJmxEwak@cedar.co.uk> Date: Thu, 16 May 1996 13:15:15 GMT In <ipQ47BA$EJmxEwak@cedar.co.uk> Paul Heffernan wrote: > If anyone has experience of JAM, could they let me know how it compares > with NEXTSTEP. > I was doing a bit of JAM programming about 2.5 years ago and at that time it was a mega-kludge. Not the least bit object-oriented at the time, very limited etc. I thought it was just plain horrible. I heard it improved a bit since, but for me it is hard to believe that it got any way close to one of its competitors, e.g. PowerBuilder, Delphi or SqlWindows on the Windows side or EOF/InterfaceBuilder on NEXTSTEP. Its only big advantage at the time was the support for a broad range of GUIs, from VT100 over Windows3.1 to X. But meanwhile you can do OPENSTEP on Unix/Windows or PowerBuilder on Windows/Solaris/Mac. So what is the point of using JAM these days ? Cheers, Ernst. -- -------------------------------------------- Ernst Kloecker ernst@fritz.snafu.de --------------------------------------------
From: George Johnston <orak@earthlink.net> Newsgroups: comp.sys.next.programmer Subject: Re: EOF and primary keys Date: Tue, 21 May 1996 23:27:37 -0700 Organization: Earthlink Network, Inc. Message-ID: <31A2B3D9.593B@earthlink.net> References: <199605211245.QAA04941@cnext.crec.mipt.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: zander@cnext.crec.mipt.ru Aleksey Sudakov wrote: > > Hi EOFers, > > I've got simple question. > > Could I have EOEntity primary key other than database unique key? I > checked the documentation one more time and it's not clear from > there. > Yes, you are able to, whether you want to is another question. The "keys" in the EOModel are used for determining uniqueness so that if you get five rows back that all have the same value in a column that has been marked as the unique key then you will only see one of those five rows. > If I could have primary key other than database unique key would > EOF ensure that value for that key should be unique or I gotta do it > myself? > Another possibility, if you want a unique column or combination of columns that is unique and different from the primary key, is to create an index on the column(s) and specify it as unique. If you try to add values to it that are not unique you will get an error and the transaction will be rolled back. If there is an overwhelming reason to do this you might also configure the EOController handling that table to not buffer edits and commands so that you will know when you try to add a row that is not unique at the time you add it rather than after it is bunched up with a bunch of other changes. > Thanks. > Aleksey > _________________________________________________________________ > > Aleksey Sudakov > 1A-1 Kerchenskaya Street > Moscow 113303, Russia. > ________________________________________________________________ > Home: +7(095)121-6100 + 3-45, +7(095)121-0561 + 3-45 > EMail: zander@cnext.crec.mipt.ru (NeXTMail & MIME are OK) > WWW: http://www.crec.mipt.ru/~zander/ I am curious, why do you need unique column(s) other than the primary key? George Johnston
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Contracts/DC & ILL Date: 22 May 1996 16:31:02 GMT Organization: Norden 1 Communications Message-ID: <4nvfg6$oil@tofu.alt.net> Programmer/analyst/developers NEXTSTEP---------------------Commercial experience Objective C------------------Commercial experience EOF--------------------------A strong plus Sybase or Oracle-------------A strong plus Contract positions-----------DC area & ILL Start Date-------------------June 1996 Must Be US Citizen or Greencard To Be Considered-------------Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: hugues@precipice.fdn.fr (Hugues RICHARD) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4ntc1p$aj1@nntp1.best.com> Control: cancel <4ntc1p$aj1@nntp1.best.com> Date: 22 May 1996 20:21:10 GMT Organization: Individual - France Message-ID: <4nvsvm$17e@precipice.fdn.fr> cancel
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: hideous bug when compositing on 8 bit displays under intel Date: 22 May 1996 20:13:48 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4nvshs$kbm@netty.york.ac.uk> aaargh the frustration! after two days, i've finally found my problem, and it seems to be a genuine bug. compositing a transparent white image from a bitmap containing interleaved R, G, B and alpha channels fails to work correctly. it overwrites the destination, rather than transferring transparently as it should. this *only* happens on 8 bit displays on intel platforms. i currently have an extremely pressing deadline, and would appreciate it if anyone has a decent fix for this problem. (there are users out there using 8 bit displays and telling them all to change their configuration is just not on!) any ideas ? cheers, rog. (rog@ohm.york.ac.uk) P.S. here is some code that reproduces the problem. try running it from the shell. it should produce a window containing a blocky "OK" when all is working; if the window is blank, you have seen the bug. #import <appkit/appkit.h> #import <libc.h> #define BMWIDTH 10 char bmspec[][BMWIDTH] = { "oooo k kk", "o o k kk ", "o o kkk ", "o o k k ", "o o k kk ", "oooo k kk" }; #define BMHEIGHT (sizeof(bmspec) / sizeof(bmspec[0])) #define SPP 4 /* samples per pixel */ id new_win(NXCoord width, NXCoord height); #define WHITE 255 #define BLACK 0 #define TRANSPARENT 0 #define OPAQUE 255 int main() { NXSize winsize = {BMWIDTH * 32.0, BMHEIGHT * 32.0}; unsigned char *bm; id bmrep, image; int x, y, i; [Application new]; /* * create bitmap * all pixels are white; letter pixels are transparent */ bm = (unsigned char *)malloc(BMWIDTH * BMHEIGHT * SPP); for (y = 0; y < BMHEIGHT; y++) for (x = 0; x < BMWIDTH; x++) { BOOL isletter = (bmspec[y][x] != ' '); unsigned char *place = &bm[y * BMWIDTH * SPP + x * SPP]; for (i = 0; i < SPP - 1; i++) place[i] = WHITE; place[i] = (isletter ? TRANSPARENT : OPAQUE); } bmrep = [[NXBitmapImageRep alloc] initData:bm pixelsWide:BMWIDTH pixelsHigh:BMHEIGHT bitsPerSample:8 samplesPerPixel:SPP hasAlpha:YES isPlanar:NO colorSpace:NX_RGBColorSpace bytesPerRow:0 bitsPerPixel:0]; [bmrep setSize:&winsize]; image = [[NXImage alloc] initSize:&winsize]; [image useRepresentation:bmrep]; { id win = new_win(winsize.width, winsize.height); NXPoint org = {0, 0}; [[win contentView] lockFocus]; PSsetgray(NX_BLACK); PSrectfill(0.0, 0.0, winsize.width, winsize.height); [image composite:NX_SOVER toPoint:&org]; DPSFlush(); [[win contentView] unlockFocus]; } printf("press return..."); getchar(); return 0; } id new_win(NXCoord width, NXCoord height) { NXRect fr; NXSize scrsz; id ret; [NXApp getScreenSize:&scrsz]; fr.origin.x = (scrsz.width - width) / 2; fr.origin.y = (scrsz.height - height) / 2; fr.size.width = width; fr.size.height = height; ret = [[[Window alloc] initContent:&fr style:NX_PLAINSTYLE backing:NX_RETAINED buttonMask:0 defer:NO] reenableDisplay]; [ret orderFront:nil]; return ret; }
From: Bob Hathaway <objcur@wwa.com> Newsgroups: comp.sys.next.programmer,comp.lang.oberon,comp.object.logic,comp.lang.scheme,comp.lang.beta Subject: OBJECT CURRENTS - Free New Monthly WWW Journal - Many New Issues Followup-To: comp.object Date: 23 May 1996 02:22:24 -0500 Organization: Object Currents Sender: objcur@sashimi.wwa.com Message-ID: <4o13ng$5i7@shoga.wwa.com> Summary: Object Currents - Free New Journal - Announcement/Call For Papers Keywords: Free WWW OO Object-Oriented Journal OBJECT CURRENTS =============== OBJECT CURRENTS ONLINE HYPERTEXT JOURNAL FREE NEW MONTHLY OBJECT-ORIENTED FORUM Location: http://www.sigs.com/objectcurrents/ Editor: Bob Hathaway (objcur@wwa.com, bhathaway@sigs.com) Issues: Jan, Feb, Mar, April, May 1996 (All Available now) Next Issue: June 1 Publisher: SIGS: C++ Report, JOOP, Object Magazine, ROAD, Smalltalk Report, X Journal, Object Expert, Objeck Spectrum, Java Report, ... This is an invitation to join us at Object Currents and view, engage in, and participate in the latest in object-oriented technology using the newest in information technology, the WWW. Object Currents is a complete new free monthly journal with original Feature Articles, Columns, and Departments along with 7 to 10 *new* articles from SIGS' Journals. Object Currents' World Class Columnists: Watts Humphrey: SEI Process Director, CMM & PSP Inventor Bertrand Meyer: Eiffel, OO Design and Software Engineering Francois Bancilhon: President, O2 Technology, Leading ODBMS Expert Michael Jesse Chonoles: Chief of Methodology, Advanced Concepts Center of Lockheed Martin David Shang: OO Programming Language Designer, Motorola Labs Michael Spertus: President, Geodesic Systems, Program Automation Interviews: January: Grady Booch February: James Rumbaugh March: Ivar Jacobson (Part I) - Get the latest on the UML New: June: Steve Mellor, Jacobson (Part II) July: Sally Shlaer Newsgroup Dialog: Robert Martin Week in OT: Jane Grau - Late breaking news on object technology 4 times/month Feature Articles: Jan 96 Understanding Objects by Understanding Variables and Functions Dr. Stefano A. Cerri A Comparison of Objects with Frames and OODBs Dr. Xindong Wu Feb 96 On the applicability of the evolving Unified Method to Rapid Functional Prototyping Ian Mitchell Extending the Object Oriented Languages Type Notion Hugues Lecoeuche Mar 96 Back To Basics - Greater Programming Power Multiple Polymorphism and True Type Systems, Recursive Types and Pointer Semantics, Meta Systems and Reflection, Prototyping and Dynamic Inheritance Part I (MP & TTS, RT & PS) - Editor Simple Shared Objects Robert Ott Apr 96 On a Semantic Net for Object Spotting Ian Mitchell, Norman Parrington, Peter Dunne, and John Moses Measuring Coupling in Object-Oriented Systems Martin Hitz and Behzad Montazeri Special: An Updated *Running* Version Of Understanding Objects Dr. Stefano A. Cerri Editorial: Back To Basics - Greater Programming Power Part II - Meta Systems and Reflection May 96 Practical Prototyping, Part One Ian Mitchell, Norman Parrington, Peter Dunne, and John Moses UniSQL's Next-Generation Object-Relational Database Management System F. Kim New: Newsgroup Dialog - Exploring a C++ Metamorphism Proposal Jun 96 (will appear June 1) Objects And Binary Relations A.Girow Practical Prototyping, Part Two Ian Mitchell, Norman Parrington, Peter Dunne, and John Moses Editorial: Back To Basics - Greater Programming Power Part III - Prototyping and Dynamic Inheritance Departments: Several including Newsgroup Dialog, Editorial, C++ Puzzle, Code Watch, Question + Answer. Best new articles from all of SIGS January, February, March, April, May, and June 1996 issues including: C++ Report, JOOP, Object Magazine, ROAD, Smalltalk Report, Object Expert, Object Buyer's Guide We are accepting original Feature Articles which include honorarium and the opportunity to publish. Please see our URL for Authors' Guidelines. Thanks to our readership for patronage, praise, and feedback. Please keep visiting or give it a try soon. Please also feel free to inform friends and colleagues of this free new medium. From the Guidelines: Object Currents' unique hypertext media provides for advances over earlier journals - links to home pages, sites, databases and information servers, interaction, animation, graphics, code retrieval and execution, expanded pages, video, virtual reality and chat sessions. While all of these may not have appeared in these first issues, they will appear in the future. Check it out! Best Regards, Bob Hathaway Robert John Hathaway III Editor in Chief Object Currents Hypertext Journal 655 West Irving Park Road, Suite 5417 Chicago, Illinois 60613-313775 USA Ph: 312-327-2121 Fax: 312-327-2936 Email: objcur@wwa.com - Correspondence, Submissions 75027.1663@compuserve.com - Cc bhathaway@sigs.com - Cc
From: stanj@cs.stanford.edu (Stan Jirman) Newsgroups: comp.sys.next.programmer Subject: DAT tape capacity reading Date: 23 May 1996 07:55:47 GMT Organization: Stanford University Message-ID: <4o15m3$89g@nntp.Stanford.EDU> Hello, does someone know how I can get the remaining capacity from a DAT tape, once I write onto it through /dev/nrst0 ? SafetyNet does it, somehow... Thanks a lot! - Stan --- Nature photography: http://www-leland.stanford.edu/~stanj NeXTmail and MIME: stanj@cs.stanford.edu
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Is there an MS-Windows-like "combo-box" for NEXTSTEP? Date: 23 May 1996 13:52:50 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4o1qji$8dj@lion.embl-heidelberg.de> References: <DrMLrv.1o6@burgond.remcomp.fr> In article <DrMLrv.1o6@burgond.remcomp.fr> pascal@burgond.remcomp.fr (Pascal Bourguignon) writes: > In article <4ne848$o2j@ralph.vnet.net> Roland Telfeyan <roland@telf.com> > writes: > > Is there an MS-Windows-like "combo-box" GUI object for NEXTSTEP? > > > > Thanks, > > > > Roland > > You may want to use the PopUpList menus. .. and have hell a lot of fun ;-) -- georg --
From: yannick@uranus.univ-lr.fr (Yannick Buisson) Newsgroups: comp.sys.next.programmer Subject: emacs and textfield !! Date: 23 May 1996 14:37:56 GMT Organization: Universite de La Rochelle Message-ID: <4o1t84$qnc@hpuniv.univ-lr.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Firm: University of La Rochelle Hi all, i'm looking for a subclass of textfield which allow users to use the emacs'keys ?? Does such a class exist ? Where can i find it ? thanx for your help -- //// (. .) ----oOO--(_)--OOo-------------------------------------------- Yannick BUISSON Centre de Ressources Informatiques Université de La Rochelle tel prof. : 46 45 82 14. fax prof. : 46 45 82 45. Email (NeXTMail , MIME) : -> yannick@cri.univ-lr.fr
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: Re: hideous bug when compositing on 8 bit displays under intel Date: 23 May 1996 14:51:28 GMT Organization: Netcom Distribution: world Message-ID: <4o1u1g$k05@dfw-ixnews2.ix.netcom.com> References: <4nvshs$kbm@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) wrote: > compositing a transparent white image from a bitmap containing > interleaved R, G, B and alpha channels fails to work correctly. > > it overwrites the destination, rather than transferring transparently > as it should. > > this *only* happens on 8 bit displays on intel platforms. I have never found any documentation about how 8-bit NS color allocates bits among R, G, B, and alpha. Does 8-bit color even have an alpha channel? How are bits allocated? -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: Mark Anenberg <marka@Eng.Sun.COM> Newsgroups: comp.sys.next.programmer Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Thu, 23 May 1996 10:53:18 -0700 Organization: Sun Microsystems Inc. Message-ID: <31A4A60E.2B41@Eng.Sun.COM> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Mike Paquette wrote: > > John Kheit <jkheit@cnj.digex.net> wrote: > > > I'm wondering what is the future of sound under OPENSTEP? I'm > >guessing that the soundKit will continue to be available and work under > >OPENSTEP for Mach. However, I wonder if it will be available, or it's > >functionality, under other OPENSTEP implementations? > > The SoundKit is not part of the OPENSTEP specification. > > > In other words, if I port my application to OPENSTEP for Mach, and > >then later wish to port it to OPENSTEP for NT, will it work under NT with > >sound? (I'm hoping that somekind of sound facility is available through > >the Postscript server layer of OPENSTEP). > > Well, the NeXT playsound DPS extension makes the speaker beep. This > may not be sufficient for your needs, however... > > Mike Paquette > > I don't speak for my employer, and they don't speak for me. > mpaque@pbinet.com Personal E-mail > mpaque@next.com NeXT business mail only, please True, the SoundKit is not part of the OPENSTEP spec; However Solaris OpenStep from Sun will include an implementation of the NX-sound apis; while not the complete implementation, it does suffice to support everything that Mail.app does with Sound including the soundview stuff. -- Mark Anenberg , OpenStep Development Team, SunSoft, Inc. [E,NeXT,Mime]Mail: mark@velodrome.com, http://www.velodrome.com/ Disclaimer: The opinions expressed above are my own and in no way represent those of Sun Microsystems, Inc.
From: Roland Telfeyan <roland@telf.com> Newsgroups: comp.sys.next.programmer Subject: Re: DAT tape capacity reading Date: 23 May 1996 22:45:45 GMT Organization: Vnet Internet Access, Charlotte, NC - info@char.vnet.net Distribution: world Message-ID: <4o2pqp$2j1@ralph.vnet.net> References: <4o15m3$89g@nntp.Stanford.EDU> In article <4o15m3$89g@nntp.Stanford.EDU> stanj@cs.stanford.edu (Stan Jirman) writes: > Hello, > > does someone know how I can get the remaining capacity from a DAT tape, > once I write onto it through /dev/nrst0 ? SafetyNet does it, somehow... Put a filter right in front of the >/dev/nrst0 that counts the bytes that fly by. Roland
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 24 May 1996 04:30:11 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4o3e0j$9qt@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> Mark Anenberg <marka@Eng.Sun.COM> wrote: > True, the SoundKit is not part of the OPENSTEP spec; > However Solaris OpenStep from Sun will include an implementation of the NX-sound apis; while not the complete implementation, it does suffice to support everything that Mail.app does with Sound including the soundview stuff. Seems like a good time to have a further re-specification of OPENSTEP to include at least these basic sound functions...All we need now is for that to diverge (note sarcasm :) Thanks for the note, it's good to know. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: stanj@cs.stanford.edu (Stan Jirman) Newsgroups: comp.sys.next.programmer Subject: Re: DAT tape capacity reading Date: 24 May 1996 04:45:01 GMT Organization: Stanford University Distribution: world Message-ID: <4o3esd$qna@nntp.Stanford.EDU> References: <4o2pqp$2j1@ralph.vnet.net> Roland Telfeyan <roland@telf.com> writes > In article <4o15m3$89g@nntp.Stanford.EDU> stanj@cs.stanford.edu > (Stan Jirman) writes: > > Hello, > > > > does someone know how I can get the remaining capacity from a > > DAT tape, once I write onto it through /dev/nrst0 ? SafetyNet > > does it, somehow... > > Put a filter right in front of the >/dev/nrst0 that counts the > bytes that fly by. > > Roland Well, that is nice, as this will tell me how many bytes I have written to the tape. But I don't know how well the data was then hardware compressed, nor do I know what length the tape is (i.e., *I* know, the program which I am writing does not). SafetyNet properly recognizes compressed data from uncompressed (when I force the drive to compress or not to; the "available capacity" is always reasonable). Hence, the question remains. I just spent the day trying... Someone there must know, right? :-) Thanks! - Stan --- Nature photography: http://www-leland.stanford.edu/~stanj NeXTmail and MIME: stanj@cs.stanford.edu
From: vazquezr@physics.ucla.edu (Rick Vazquez) Newsgroups: comp.sys.next.programmer Subject: Re: ld: can't locate file for: -lX11 Date: 24 May 1996 05:17:45 GMT Organization: University of California, Los Angeles Message-ID: <4o3gpp$sma@uni.library.ucla.edu> References: <4np0ms$1cfo@uni.library.ucla.edu> <4npnem$rsh@bias.ipc.uni-tuebingen.de> frank@this.net (Frank M. Siegert) wrote: >In <4np0ms$1cfo@uni.library.ucla.edu> Rick Vazquez wrote: >> Hello everyone, could someone please help me with the following error. >> >> ld: can't locate file for: -lX11 >> >> I get this error alot when trying to compile unix or X11 windows programs >> under my NeXT (Black). >> > >Well, NeXTSTEP is not X, so this lib is not there. > >Install the free MouseX (or the shareware Xnext) or a commercial X package >and it should work. But be prepared: you cannot run X11 programs direct under >NS, you have to start a XServer before. > >Have a nice day > > Frank >-- >Frank M. Siegert [frank@this.net] -- Home Page >http://hades.tue.schwaben.de/~frank >NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C >ga ga" > Hello, I have Cub'x installed. I have all the Developer Libs installed. I just need to know command to replace. Rick
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Re: emacs and textfield !! Date: 24 May 1996 05:16:11 GMT Organization: No organisation supplied Message-ID: <4o3gmr$nf8@hermes.is.co.za> References: <4o1t84$qnc@hpuniv.univ-lr.fr> In-Reply-To: <4o1t84$qnc@hpuniv.univ-lr.fr> On 05/23/96, Yannick Buisson wrote: > >Hi all, >i'm looking for a subclass of textfield which allow users to use the >emacs'keys ?? >Does such a class exist ? >Where can i find it ? > >thanx for your help > Yes there is such a TextField or at least of an example of how to do something like this in NeXTAnswers. Document: 1259. or the TextORama example in the MiniExamples -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: Jay <s2139473@cse.unsw.EDU.AU> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep Subject: Why NeXT? Date: Fri, 24 May 1996 16:49:09 +1000 Organization: University of New South Wales Distribution: inet Message-ID: <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> References: <4norkd$97v@digifix.digifix.com> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4norkd$97v@digifix.digifix.com> After hearing some great things about NeXTSTEP-OS, I'd like to know something: 1. Why do id software develop under it ? Whats so great about it (ie that you can't do under NT or OS/2)? 2. Why should _I_ get it? (I'm a wannabe Game Programmer currently finishing my CompSci degree at University) Would glad to hear convincing arguments... (emailed if possible) Thanks in advance. Justin s2139473@cse.unsw.edu.au http://www.cse.unsw.edu.au/~s2139473 <-"The Quake Bestiary"
Newsgroups: comp.sys.next.programmer From: yves (Yves Akakpo) Subject: Taylor_UUCP config Message-ID: <1996May24.140340.938@yves.fdn.fr> Keywords: Taylor UUCP, Sitename different Hostname Sender: news@yves.fdn.fr Organization: Individual Date: Fri, 24 May 1996 14:03:40 GMT Hello Has anyone out there had trouble with a slab wich have sitename different of hostname in the way to config Taylor_UUCP? -NeXTStation Turbo with NS3.3 User + Developer -Taylor_UUCP_1.05 -Sitename: alpha -Hostname: beta -Loginname: Ualpha -Domain(s): alpha.machin.fr I think that an modification of my SENDMAIL in their rulesetes can solve the conflict, the disharmonize. But I don t know how? Any and all comments welcome. Thanks in advance! Yves.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <@vm.cnuce.cnr.it,@ISS.IT:ZANITTI@ISS.IT> Message-ID: <199605240752.AAA17335@isp.net> Date: Fri, 24 May 96 09:51:24 ISS From: ZANITTI@ISS.IT Subject: Compile problems I have two users (root and me) USER GROUP root wheel me hother (NeXTDeveloper is installed by root) Can the me user compiles an application? If the me user compiles an application it obtains an Error message on directory tmp I don't want develop an application as super user (root) because this can crash the system and the security (Program errors and so) Thanks Zanitti Leo
From: bruno@gruick.univ-lr.fr (Bruno Garnier) Newsgroups: comp.sys.next.programmer Subject: insert Null value Date: 24 May 1996 14:28:57 GMT Organization: Universite de La Rochelle Message-ID: <4o4h39$3db@hpuniv.univ-lr.fr> Hi all, I work with InformixAdaptor. When I want to insert an [EONull null] value in a field in float format of my EOController, the inserted value is 0.00. I Have no default value in my table. How can I insert a null instead of 0.00 ?? Please help ... -- Bruno Garnier Centre de Ressources Informatique Universite de La Rochelle Avenue Marillac 17042 LA ROCHELLE CEDEX 01 Tel : 46 45 82 14 Fax : 46 45 82 45 http://www.univ-lr.fr/ bruno@cri.univ-lr.fr (NeXTMail, MIME Mail) |~~~| |~~~|D Guinness C|. :|D is |: .|D good for you ! |___| \\|// || |^ _| || ( -_+ ) // /_\ O / / __ __ __ || \_/ / | | | | | | ||____/ | | | | | | ---`----- \-- ---| |-| |--| |-- v |==| |==| |==|
From: lbo@sma.ch (-Laurent Bourqui) Newsgroups: comp.sys.next.programmer Subject: SoundBlaster Driver problem Date: 24 May 1996 17:46:55 GMT Organization: Swiss Meteorological Institute Distribution: world Message-ID: <4o4smfINNk4b@maz4.sma.ch> Hi, I have trouble with my soundblaster board under NEXTSTEP. Sometime the dirver doesn't regognize the soundblaster and when it regognize it most of the time I have no sound output. Any idea about this problem ? Is there a new driver ? The system I'm running is the following: Intet Pentium 133 Asus Motherboard AHA 2940 SCSI controller Sounblaster16 (with IDE controller, not PnP) NEXTSTEP 3.3 Sound driver : Soundblaster16 v. 3.32 Thank you for your help. Laurent --- -------------------------------------------------------------------------------- Laurent Bourqui | Phone (private) +41(0)1 451 62 92 SMI (Swiss Meteorological Institute) | Phone (office) +41(0)1 256 95 73 Kraebuehlstr. 58 | Fax (office) +41(0)1 256 92 78 | email lbo@sma.ch CH-8044 Zuerich | Switzerland | --------------------------------------------------------------------------------
Newsgroups: comp.sys.next.bugs,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin From: fuguen@paris.fdn.fr (Francois UGUEN) Subject: CraftMan Engine+ ETI Birds of Europe Message-ID: <DrvsE9.8Bx@paris.fdn.fr> Sender: news@paris.fdn.fr Organization: Individual - PARIS - Francois UGUEN Date: Thu, 23 May 1996 23:10:57 GMT Hello Birds of Europe for NEXTSTEP, a Multimedia Software I purchased in 1994, failed to function when I recently switched to NEXTSTEP 3.3 Martine van der Eerden, from ETI Biodiversity Center told me that the problem is due to the fact that the CraftMan Engine they used to build their software is failing to properly open color windows under NS 3.3 As this is very inconvenient and make Birds of Europe almost unusable, I'd like to know when a bug-fix release of the CraftMan Engine is planned to be released. Best regards, Francois UGUEN fuguen@paris.fdn.fr -- MM*: ?ìP8$
From: waldek@bg.univ.gda.pl (Waldemar Chrzan) Newsgroups: comp.sys.next.hardware,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: Q: HP DeskJet 1600CM (suplement) Date: 24 May 1996 15:11:03 GMT Organization: Politechnika Gdanska Message-ID: <4o4ji7$l29@sunrise.pg.gda.pl> Hi ! I send 3 weeks above mail with question about drivers for HP DeskJet 1600 printer for NeXt Step 3.3 (intel). I send error in my question. My printer is a DeskJet 1600CM. (This is a PostScript network printer) I want to use high quality color printing mode. Where I found drivers form my printer or haw to resolve my problem ? Help mi
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: hideous bug apology Date: 24 May 1996 14:45:28 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4o4i28$16@netty.york.ac.uk> i wrote: > after two days, i've finally found my problem, and it seems to be > a genuine bug. > > compositing a transparent white image from a bitmap containing > interleaved R, G, B and alpha channels fails to work correctly. [...] > this *only* happens on 8 bit displays on intel platforms. sorry everyone, of course it was my fault, and it was not a bug at all, merely a feature. as documented in the second last sentence of the 6th paragraph of the "Premultiplication" section about two thirds of the way through the document: /NextLibrary/Documentation/NextDev/Concepts/Pre3.0_Concepts/04_Drawing.rtfd the pixel value cannot be greater than the coverage value. the means that the pixel value : r=255,g=255,b=255,alpha=0 (i.e. transparent white) produces an undefined result, leading to different behaviours on 8-bit and 16-bit displays. the correct way to do it is something like: void set_pixel(unsigned char *pix, int r, int g, int b, int alpha) { float mult = (float)alpha / 255; pix[0] = r * mult; pix[1] = g * mult; pix[2] = b * mult; pix[3] = alpha; } sorry about the shouting. i have no excuse. cheers, rog. p.s. all thanks to Ken Case for pointing this out.
From: Jay <s2139473@cse.unsw.EDU.AU> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep Subject: Re: Why NeXT? Date: Sat, 25 May 1996 14:14:52 +1000 Organization: University of New South Wales Distribution: inet Message-ID: <Pine.ULT.3.91.960525140625.313F-100000@flute07.orchestra.cse.unsw.EDU.AU> References: <4norkd$97v@digifix.digifix.com> <199605241950.MAA00686@math.berkeley.edu> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <199605241950.MAA00686@math.berkeley.edu> On Fri, 24 May 1996, Paul Robert Brown wrote: > Justin- > > You should just sit down at a NeXTstep machine and check it out for > yourself. I don't think it's the ideal game programming environment, but > that's coming from someone who wrote Apple II games using a macro > assmebler, page-flipping animation, and other low-level techniques... Use > of a high-level language and a compiler seems a little excessive and > slows things down... For anything that involves a GUI, NeXTstep beats > everything else available, hands-down. In addition, it's UNIX. I've loved > UNIX for a long time, since the late '80's, say, and I can't stand DOS... > OS/2 is a darn good platform, kicks ass on Winblowz .095. Mmmmmm Unix. Yuck, saccarine sweet Win95, all sugar no substance. I agree 100% Linux is a little _too_ meaty. Recompiling kernels isn't my idea of a good time. NeXTstep seems like a good compromise. > NeXTstep has good capabilities to allow many people to work on one > project; it was designed with that in mind. It has the power and > flexibility of UNIX. It is the premier object-oriented development > platform out there, period. If you're not familiar with OO, you might not > see what the hubbub is about; try it, you'll never go back. (Even though > the last program I wrote was coded directly in assembly code for a > Pentium machine...) Yep. Just discovered OO at College, and I like it. I still like to assembly optimize my graphics inner-loops though :). Thanks a lot... Justin s2139473@cse.unsw.edu.au http://www.cse.unsw.edu.au/~s2139473 <-"The Quake Bestiary" "A bird in hand is better than one overhead..."
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: ld: can't locate file for: -lX11 Date: 24 May 1996 13:54:49 GMT Organization: ipc, U of Tuebingen, Germany Message-ID: <4o4f39$fq4@bias.ipc.uni-tuebingen.de> References: <4np0ms$1cfo@uni.library.ucla.edu> <4npnem$rsh@bias.ipc.uni-tuebingen.de> <4o3gpp$sma@uni.library.ucla.edu> Cc: vazquezr@physics.ucla.edu In <4o3gpp$sma@uni.library.ucla.edu> Rick Vazquez wrote: > frank@this.net (Frank M. Siegert) wrote: > >In <4np0ms$1cfo@uni.library.ucla.edu> Rick Vazquez wrote: > >> Hello everyone, could someone please help me with the following error. > >> > >> ld: can't locate file for: -lX11 > >> > >> I get this error alot when trying to compile unix or X11 windows programs > >> under my NeXT (Black). > >> > > > >Well, NeXTSTEP is not X, so this lib is not there. > > > >Install the free MouseX (or the shareware Xnext) or a commercial X package > >and it should work. But be prepared: you cannot run X11 programs direct > under > >NS, you have to start a XServer before. > > > >Have a nice day > > > > Frank > >-- > >Frank M. Siegert [frank@this.net] -- Home Page > >http://hades.tue.schwaben.de/~frank > >NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C > >ga ga" > > > > > Hello, I have Cub'x installed. I have all the Developer Libs installed. I > just need to know command to replace. > > Rick > [I am not sure if I parsed the above sentence correctly... 8-)] Then the lib search path is not set correctly, you may have to add some symlinks in /usr/lib/... For a quick test you can add a -Ldir Add dir to the list of directories in which to search for libraries. Directories specified with -L are searched before the standard directories. (where dir should point to your libX11.a containing directory) to the link options in the Makefile... Bye Frank -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.advoocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 24 May 1996 15:20:02 GMT Organization: University of Sheffield, UK Message-ID: <4o4k32$lv0@bignews.shef.ac.uk> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> In-Reply-To: <31A4A60E.2B41@Eng.Sun.COM> On 05/23/96, Mark Anenberg wrote: > However Solaris OpenStep from Sun will include an implementation of the > NX-sound apis; while not the complete implementation, it does suffice > to support everything that Mail.app does with Sound including the soundview > stuff. > > -- > Mark Anenberg , OpenStep Development Team, SunSoft, Inc. > And here were some of us thinking that Mail.app wouldn't be part of Sun's OpenStep toolkit... ;-) Best wishes, mmalc. --
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Foundation memory management blues (longish ramble) Date: 24 May 1996 19:01:09 GMT Organization: Netcom Distribution: world Message-ID: <4o511l$cl4@dfw-ixnews10.ix.netcom.com> Periodically, I post because I still don't get it when it comes to using Foundation's memory management scheme correctly. I'm still looking for the magic key to understanding. Not having this key is very frustrating. My problem lies in knowing when to retain object ivars. NeXT's documentation seems to suggest that object ivars that are "intrinsic" should be retained or copied, and those that are "extrinsic", shouldn't. However, an exception occurs when retaining an intrinsic ivar forms a retain cycle, a basic flaw in any reference-counting memory management system that results in memory leaks. To partially deal with this, NeXT advises that objects shouldn't retain their delegates and that children shouldn't retain their parents in hierarchical architectures. Most of my object architectures are graphs, not hierarchies. Peer objects maintain references to each other so they can serve each other. How does one decide which retains which? I haven't discovered any general guidelines, so designing class graphs seems anarchic and error-prone. NeXT has also stated that retain cycles haven't been a major problem in their apps. But I find many possibilities for retain cycles in my object graphs, usually involving more than just 2 objects which makes detecting these potential retain cycles difficult and an unfortunate additional cognitive load that programmers just shouldn't have to deal with. Maybe I maintain too many ivars or design flawed architectures. This makes me want to move toward minimal retaining - retain just those objects that an object creates and don't retain any object received from another object. But that defeats one of the goals of the Foundation memory management system: objects are responsible for the existence of their intrinsic object ivars, so they must retain or copy them to prevent their release due to the actions of other objects. But how does one decide what's "intrinsic"? NeXT seems to suggest that children are intrinsic and should be retained, but parents are extrinsic and shouldn't be retained. Maybe NeXT is just reflecting U.S. culture, but that's another story :-) For example, if a class, MenuManager, has ivars pointing to various menu cells, should MenuManager retain these menu cells? MenuManager didn't create these menu cells, but if they are freed because some other object releases them, MenuManager will send messages to freed objects. But this is true for any object ivar, be it "intrinsic" or "extrinsic". But working with menu cells seems central to MenuManager, so maybe menu cell ivars are intrinsic. But if a menu cell is removed from a menu by another object, it's not available to the user, so who cares that MenuManager has retained the removed menu cell so it can still operate on it? Maybe a menu cell is thus extrinsic and shouldn't be retained. Then there's the issue of object ivars assigned when an object is unarchived from a nib. Objects unarchived from a nib have a retain count of 1 before they're assigned to an object's ivar as a result of an IB outlet connection, so some object must step forward to assume responsibility for releasing this "extra" retain to prevent memory leaks. Which object should take on this responsibility? I presume it should be an object that considers this unarchived object to be intrinsic - it would have retained an object assigned to the same ivar had it been done programmatically rather than as a result of nib loading. But more than one object might consider this object intrinsic, so which one of these objects should release the "extra" retain? If more than one object releases this "extra" retain, a message to a freed object results. But the programmatic setting of an intrinsic ivar may have been done using a "setIvarName:" method in which the ivar is retained. This method would be invoked during nib loading as well, so an object assigned this way might have a retain count 1 greater than an object assigned programmatically. How does one get rid of this extra retain count? I guess it could be done in awakeFromNib. This all feels pretty kludgy and definitely error-prone :-( Am I just totally missing the boat? Does everyone else feel comfortable with Foundation's memory management? Am I too old to learn new tricks? -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <doug@thoughtful.com> Message-ID: <9605241948.AA18304@thoughtful.com> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Douglas Simons <doug@thoughtful.com> Date: Fri, 24 May 96 13:48:52 -0600 Subject: A Solution to the Problem of Cyclical References Among Objects Hello, Paul Summermatter's recent post to the EOF list on the problem of how to reliably release objects when there are cyclical references among them got me thinking about a possible solution which was hinted at in that post. I discussed my idea with Nathan Schroeder here at Thoughtful Software, and together we designed a HyperSense document which implemented a simulation of objects and allowed us to test the algorithm. I thought folks on this list might be interested also. Other people have probably invented similar schemes, but since we created the simulator, we thought we'd share it (and our approach) with everyone, so you can play with it and get an actual look at how it all works. OVERVIEW The basic approach involves two additional methods, retainFor: and releaseFor: which allow an object to maintain a list of objects which have retained it. We called this the refList. By keeping such a list, when an object is released, it can consult with the known objects which have retained it to see if any of them are being retained by any objects "outside the group". If not, then the object can safely be freed. By implementing the approach in an interpreted mock-up environment we were able to quickly test a variety of simple and more complex inter-relationships among objects. Some of them proved to have tricky interactions which uncovered weak points in the basic approach. As difficult cases were discovered, we refined the algorithm to handle them, and have arrived at what I believe is a robust and general-purpose solution to the problem. PLEASE TRY IT! We have posted the simulator to the archives and invite everyone to play around with it, using cases you are familiar with, to see if there are any other tricky situations we may have overlooked (details on obtaining and using the simulator follow this article). We've tested it with simple binary cycles (two objects retaining each other) and more complex scenarios involving up to 5 objects with various combinations of interconnections. If nothing else, the simulator is fun to watch! :-) Now, some more details, for those who are interested... PROS AND CONS One immediately obvious drawback to this approach is that it requires additional overhead, both in memory (the refList of objects in each participating object) and in cpu time (for messaging related objects when an object is released). In actual practice, I believe both of these will prove to be minimal, except in situations where there is a great deal of interdependency between objects. Another consideration is the necessity for the programmer to choose when to use retainFor: and releaseFor: rather than retain and release. There is a fairly simple answer, though: if your objects are being "leaked" because of cyclic references, then they should use these new methods which deal with that problem! The good news is that the standard retain and release methods can still be used on any of these objects. It's perfectly reasonable for objectA to be retained by objectB using retainFor: and by objectX using the standard retain method. In fact, it's a good idea, as long as there's no chance that objectA (or its friends) know anything about objectX. The really big plus, of course, is that this approach solves the problem of cyclic references in a fairly general way, and is not too difficult to implement. HOW IT WORKS When an object is released, it decrements its refCount and then checks whether the refCount is greater than the number of objects in its refList. If so, that means there are still "outside" references to the object (from unknown retainers) so we simply return. If the refCount goes to zero, there are no references to the object and it may be freed. On the other hand, if the refCount is equal to the number of objects in the refList, that means that we know all of the objects which are retaining us. So we send each of them an isRetainedByOutsiderFor: message to see if any of them are retained by "outsider" objects (those which are not part of our interdependency graph). The implementation of that method checks the receiving object's refCount against the number of objects in its refList, and then in turn calls isRetainedByOutsider: for those objects. In this way, all related objects are checked to see whether they are retained by any object which is outside the group. If none of the objects which are retaining an object are themselves being retained by any unknown "outsider" objects, and all of their known retainers are in the same situation, then it is safe to free the original object, and to release any objects which it retains. Rather than freeing the object directly, it is added to a special pool of objects which will be freed upon return to the main run loop. This is necessary because there are situations in which other objects will send messages to the object after it has determined that it will free itself, as part of the joint release process. So it can't actually be freed until later. Infinite recursion is a problem which had to be dealt with in a couple of situations. In the case of the isRetainedByOutsiderFor: message, we chose to pass a list of objects as an argument, adding each sender to the list as it was forwarded. This list is consulted before forwarding the isRetainedByOutsiderFor: message to other objects. In this way, no object is ever asked to respond to the same request twice. In some other cases, we set a flag in the object indicating that it is already checking whether it can be freed, or will be freed, or is being freed. In the simulator, this is represented by a notation on the refCount line. In a real implementation, a single recursion-prevention flag would suffice for all three cases. UNFINISHED BUSINESS It should be noted that, so far, we have implemented this algorithm only in the HyperSense simulator, not in actual Objective C code. One consequence is that we've totally ignored any implementation for an autoreleaseFor: method, although one will most likely be needed in the real world. A potentially more serious concern will be relevant if the application can in any situation ever re-access an object after it has been released, but before it is freed from the freePool. If this is possible in your application, you should take a very hard look at the implications. It may be possible to add code to the retain method to deal with this situation, but I haven't explored it. Also, I'm not an EOF expert, so there may be other ramifications besides the one mentioned above having to do with the way EOF caches and fetches objects. OBTAINING AND USING THE SIMULATOR The simulator has been posted to the archives at next-ftp.peak.org, in the submissions directory: ftp://next-ftp.peak.org/pub/next/submissions/RetainSimulator.sense.compressed It should migrate to the HyperSense directory and possibly other locations. To use the simulator, you will need a copy of either HyperSense.app or HyperSensePlayer.app. Both of these are available in the HyperSense directory at peak. In order to be able to examine (and experiment with changes to) the SenseTalk scripts which implement the algorithm, you should use the full developer version of HyperSense rather than the Player. Running HyperSense in demo mode will give you full access to the scripts and let you modify them if you like, although changes will not be saved to disk. HyperSensePlayer (including FREE single-user license and sample documents): ftp://next-ftp.peak.org/pub/next/HyperSense/HyperSenseREADME1.00B14a.rtfd.compressed ftp://next-ftp.peak.org/pub/next/HyperSense/HyperSensePlayer.pkg.1.00B14a.NIHS.b.tar HyperSense (full development version and documentation): ftp://next-ftp.peak.org/pub/next/HyperSense/HyperSenseApp.pkg.1.00B14a.NIHS.b.tar ftp://next-ftp.peak.org/pub/next/HyperSense/HyperSenseDocs.pkg.1.00B14a.tar After everything is installed on your disk, launch HyperSense and then open the RetainSimulator document. Click the button labelled "?" for some instructions on using the simulator. You can easily change the inter-dependency relationships between the simulated objects (in either HyperSense or the Player) and play around with referencing and releasing objects in any order. Most of the handlers which implement the functionality are contained in the page script. To edit this script, in HyperSense (not the Player), open the Message Box (from the Tools menu), and type "edit the script of this page". Have fun! We'd be happy to hear any questions or comments (you can reach us at support@thoughtful.com). Using HyperSense to simulate the objects and behavior proved to be a tremendous help in discovering and correcting problems with the algorithm. I highly recommend it as a rapid prototyping tool!!! (note: this should be considered a shameless plug, since I am the chief designer of HyperSense, and I'm not ashamed of it. Nevertheless, everything I said about HyperSense's utility is absolutely true! :-) I hope this is helpful to some of you. Best Regards, Doug Simons Thoughtful Software doug@thoughtful.com
Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep From: nitezki@nidat.sub.org (Peter Nitezki) Subject: Re: Why NeXT? Message-ID: <DryK56.G0B@nidat.sub.org> Sender: news@nidat.sub.org Cc: s2139473@cse.unsw.EDU.AU Organization: private site of Peter Nitezki, Kraichtal, Germany References: <4norkd$97v@digifix.digifix.com> <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> Distribution: inet Date: Sat, 25 May 1996 11:05:30 GMT In <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> Jay wrote: > > After hearing some great things about NeXTSTEP-OS, I'd like to know something: > > 1. Why do id software develop under it ? Whats so great about it (ie that > you can't do under NT or OS/2)? > This question can only be posed by some one who's never done development both under Unix and one of the other systems mentioned (If he had, he'd not pose such a stupid question!) And if he'd know the Unix way and seeing how it's done on NEXTSTEP he'd want to have one. Period! > 2. Why should _I_ get it? (I'm a wannabe Game Programmer currently > finishing my CompSci degree at University) > If you're only used to DOS you'd probably better leave it that way, since you'd spend a considerable bit of time worrying why things must be so complicated before you'd catch the point. And afterwards you probably could never be a DOS programmer again ;-) As a summary: NEXTSTEPpers are bored by the request to talk someboy into it since they can't understand why people don't just come and see. Or let it put me that way: it's the same as with oysters. If you don't like'em half a dozen is too much, and if you like'em half a dozen is not enough. -- Peter Nitezki | Nitezki@NiDat.sub.org # Blessed art thou who knoweth Staarenbergstr. 44 | Tel.: +49 7251 62495 # not about the pleasure and D-76703 Kraichtal | Fax : +49 7251 69215 # delight of being hooked GERMANY | E-mail defunct, sorry # up to the Net. Peter 1,3-5
Newsgroups: comp.sys.next.programmer From: nitezki@nidat.sub.org (Peter Nitezki) Subject: Re: DAT tape capacity reading Message-ID: <DryKt7.G1n@nidat.sub.org> Sender: news@nidat.sub.org Cc: stanj@cs.stanford.edu Organization: private site of Peter Nitezki, Kraichtal, Germany References: <4o2pqp$2j1@ralph.vnet.net> <4o3esd$qna@nntp.Stanford.EDU> Date: Sat, 25 May 1996 11:19:55 GMT In <4o3esd$qna@nntp.Stanford.EDU> Stan Jirman wrote: > Roland Telfeyan <roland@telf.com> writes > > In article <4o15m3$89g@nntp.Stanford.EDU> stanj@cs.stanford.edu > > (Stan Jirman) writes: > > > Hello, > > > > > > does someone know how I can get the remaining capacity from a > > > DAT tape, once I write onto it through /dev/nrst0 ? SafetyNet > > > does it, somehow... > > > > Put a filter right in front of the >/dev/nrst0 that counts the > > bytes that fly by. > > > > Roland > > Well, that is nice, as this will tell me how many bytes I have written > to the tape. But I don't know how well the data was then hardware > compressed, nor do I know what length the tape is (i.e., *I* know, the > program which I am writing does not). SafetyNet properly recognizes > compressed data from uncompressed (when I force the drive to compress or > not to; the "available capacity" is always reasonable). > > Hence, the question remains. I just spent the day trying... Someone > there must know, right? :-) > My guess: they're just guessing. They have heuristics on the tape capacity and they know how to read the media parameters from the drive (tape lenght is encoded in the DAT cassette body, for instance). Then they have the archive directory on file and just substract the written amount from the estimate capacity. If the archive directory should be inaccessible they'd be as blind as all of us :-) But, remember! It's just a guess! -- Peter Nitezki | Nitezki@NiDat.sub.org # Blessed art thou who knoweth Staarenbergstr. 44 | Tel.: +49 7251 62495 # not about the pleasure and D-76703 Kraichtal | Fax : +49 7251 69215 # delight of being hooked GERMANY | E-mail defunct, sorry # up to the Net. Peter 1,3-5
From: reichman@usc.edu (Matthew N. Reichman) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep Subject: Re: Why NeXT? Date: 25 May 1996 18:03:30 GMT Organization: Como me Gusta productions Sender: reichman@comserv-j-22.usc.edu Distribution: inet Message-ID: <4o7i1i$7ku@usc.edu> References: <4norkd$97v@digifix.digifix.com> <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> <DryK56.G0B@nidat.sub.org> Cc: nitezki@nidat.sub.org In <DryK56.G0B@nidat.sub.org> Peter Nitezki wrote: > As a summary: NEXTSTEPpers are bored by the request to talk someboy into it > since they can't understand why people don't just come and see. Or let it > put me that way: it's the same as with oysters. If you don't like'em half a > dozen is too much, and if you like'em half a dozen is not enough. You are definitely a riot, Peter! 8-) -- Be well, Matthew Reichman reichman@usc.edu USC-CNTV NeXTStep v.3.3 m68k NeXTMAIL & MIME welcome =============================================================== PGP key --> email w/ subject "request_PGP" --------------------------------------------------------------- Computer Privacy Information --> http://www.eskimo.com/~joelm/
From: Harpreet S Sandhu <harpreet@sandhu.demon.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Comparison of JYACC's JAM with NEXTSTEP? Date: Sat, 25 May 1996 18:38:46 +0100 Organization: LSE Distribution: world Message-ID: <9oikZCAmW0pxEwwm@sandhu.demon.co.uk> References: <ipQ47BA$EJmxEwak@cedar.co.uk> <DrI25F.3oG@fritz.snafu.de> MIME-Version: 1.0 In article <DrI25F.3oG@fritz.snafu.de>, Ernst Kloecker <ernst@fritz.snafu.de> writes >In <ipQ47BA$EJmxEwak@cedar.co.uk> Paul Heffernan wrote: >> If anyone has experience of JAM, could they let me know how it compares >> with NEXTSTEP. >> > >I was doing a bit of JAM programming about 2.5 years ago and at that time it >was a mega-kludge. Not the least bit object-oriented at the time, very >limited etc. I thought it was just plain horrible. > >I heard it improved a bit since, but for me it is hard to believe that it got >any way close to one of its competitors, e.g. PowerBuilder, Delphi or >SqlWindows on the Windows side or EOF/InterfaceBuilder on NEXTSTEP. > >Its only big advantage at the time was the support for a broad range of GUIs, >from VT100 over Windows3.1 to X. > >But meanwhile you can do OPENSTEP on Unix/Windows or PowerBuilder on >Windows/Solaris/Mac. > >So what is the point of using JAM these days ? > > >Cheers, Ernst. > >-- >-------------------------------------------- >Ernst Kloecker ernst@fritz.snafu.de >-------------------------------------------- > > I totally agree - I also used JAM v5.3 for Windows a couple of years ago. One really cannot compare JAM with NEXTSTEP: NEXTSTEP can be used to produce robust, mission-critical applications; JAM includes a 4GL with is loosely typed and very slow. We had to re-engineer a human resources client/server application which used the 4GL to use C++ to improve its speed! Moreover, since it is portable across different GUIs, it cannot - by definition - provide all of the functionality of any one GUI: the GUI functionality is a subset of the functionality of all the GUIs it supports. NEXTSTEP is a dream come true after having used JAM. -- Harpreet S Sandhu hsandhu@stockex.co.uk harpreet@sandhu.demon.co.uk
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advoocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 25 May 1996 21:33:39 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4o7ubj$ud@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > And here were some of us thinking that Mail.app wouldn't be part of Sun's > OpenStep toolkit... ;-) Well, they showed it at CeBIT. Problem is, it's Mail.app, but without NeXTMail :-((( (SUN: PLEASE tell me that's not true!) Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep Subject: Re: Why NeXT? Followup-To: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep Date: 26 May 1996 09:01:02 -0500 Organization: MCSNet Services Distribution: inet Message-ID: <4o9o6u$3og@Venus.mcs.com> References: <4norkd$97v@digifix.digifix.com> <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> <DryK56.G0B@nidat.sub.org> Peter Nitezki (nitezki@nidat.sub.org) wrote: : In : <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> : Jay wrote: : > 1. Why do id software develop under it ? Whats so great about it (ie that : > you can't do under NT or OS/2)? : > : This question can only be posed by some one who's never done development both : under Unix and one of the other systems mentioned (If he had, he'd not pose : such a stupid question!) : And if he'd know the Unix way and seeing how it's done on NEXTSTEP he'd want : to have one. Period! This has to be the most narrow minded NON answer every seen. I have programmed under many UNIX OSs (Solaris, HPUX, AIX), DOS and Windows 3.X. I now am doing most of my work under NeXTStep. I wouldn't trade it back to the Windows 3.x for anything. But I know nothing about programming OS/2 or NT or what might be there. AND even programming under NeXTStep I fail to see what would make it such a great environment of developing games like ID does. So if you know what makes this environment so GREAT get off you HIGH HORSE and say something useful. PERIOD! : > 2. Why should _I_ get it? (I'm a wannabe Game Programmer currently : > finishing my CompSci degree at University) : > : If you're only used to DOS you'd probably better leave it that way, since : you'd spend a considerable bit of time worrying why things must be so : complicated before you'd catch the point. And afterwards you probably could : never be a DOS programmer again ;-) I feel sorry for you that you can only program on 1 system. That must seriously limit you. I hope that you have a very stable job. Peter Richardson
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: How to write an own filesystem module? Date: 26 May 1996 17:41:00 GMT Organization: ipc, U of Tuebingen, Germany Message-ID: <4oa53c$9rs@bias.ipc.uni-tuebingen.de> Howdy! Does there exist any documentation on how to write an own filesystem module (like the DOS.fs only for Linux ext2fs or similar?) I have found nothing so far... Have a nice day Frank -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
From: apuleius@ix.netcom.com(William Grosso) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep Subject: Re: Why NeXT? Date: 26 May 1996 17:42:30 GMT Organization: Netcom Distribution: inet Message-ID: <4oa566$du8@sjx-ixn2.ix.netcom.com> References: <4norkd$97v@digifix.digifix.com> <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> <DryK56.G0B@nidat.sub.org> <4o9o6u$3og@Venus.mcs.com> In <4o9o6u$3og@Venus.mcs.com> par@MCS.COM (Peter Richardson) writes: > >Peter Nitezki (nitezki@nidat.sub.org) wrote: > >> If you're only used to DOS you'd probably better leave it that >>way, since you'd spend a considerable bit of time worrying why >>things must be so complicated before you'd catch the point. >>And afterwards you probably could never be a DOS programmer again ;-) > >I feel sorry for you that you can only program on 1 system. That must >seriously limit you. I hope that you have a very stable job. > Well, he was pretty adept with the horseshit, wasn't he ? Cheers, Andy
From: "Jonathan W. Hendry" <steeldrv@ix.netcom.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advoocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Sun, 26 May 1996 17:32:04 -0400 Organization: Steel Driving Software, Inc. Message-ID: <31A8CDD4.7D3C@ix.netcom.com> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Uli Zappe wrote: > > mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > > And here were some of us thinking that Mail.app wouldn't be part of Sun's > > OpenStep toolkit... ;-) > > Well, they showed it at CeBIT. Problem is, it's Mail.app, but without > NeXTMail :-((( > > (SUN: PLEASE tell me that's not true!) Does it do MIME? If it does MIME well, who cares if it does NeXTMail?
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.bugs,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: Re: CraftMan Engine+ ETI Birds of Europe Date: 26 May 1996 17:51:31 GMT Organization: University of Sheffield, UK Message-ID: <4oa5n3$roq@bignews.shef.ac.uk> References: <DrvsE9.8Bx@paris.fdn.fr> In-Reply-To: <DrvsE9.8Bx@paris.fdn.fr> On 05/24/96, Francois UGUEN wrote: > As this is very inconvenient and make Birds of Europe almost > unusable, I'd like to know when a bug-fix release of the CraftMan > Engine is planned to be released. > Umm, sorry -- CraftMan was developed by Xanthus, who sadly went out of business a while back. Most of their applications were taken over by Lighthouse, but I've seen no indication that a re-release is in the offing. Best wishes, mmalc. --
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 27 May 1996 04:15:13 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4oba8h$g04@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Where GNU Source Date: 27 May 1996 15:02:06 GMT Organization: No organisation supplied Message-ID: <4ocg5e$ip3@hermes.is.co.za> Where can I find the GNU sources for gcc and gdb that NeXT are using? I looked on ftp.next.com but only found EOF_1.1_GNUSOURCE and PDO_3.0_GNUSOURCE which is not what I want. I know that the source is on the CD that the NeXT Developer software is shipped on, but this release hasn't got the fixes in that came with the Patches for 3.3. These sources should be publicly available as they are part of the GNU project. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: How to use (Im)mutability Date: Mon, 27 May 1996 21:24:01 GMT Organization: Orbital Computer Consultancy Message-ID: <31aa1cdb.31202737@news.plsys.co.uk> Kind of following on from Art Isbell's interesting Foundation Kit question, I'd also like to pose a question that's been bothering me about FK. The mutable / immutable distinction seems very useful at first sight. However, I think it suffers from the following problem. Suppose a class has a collection of children, and we want it to provide API for getting the children and adding a new child. For example, - (NSArray *)children; - (void)addChild:aChild; The children method should return an immutable array (or to be precise it should not return the exact same mutable array that the parent stores its children in, otherwise the parent's children could be changed without its knowledge). Obviously, the class needs an instance variable for the children, but the question is should this ivar (call it theChildren) be an NSArray or NSMutableArray? If we choose NSMutableArray, the children method must copy the contents of theChildren into an immutable NSArray, using [[NSArray alloc] initWithArray:theChildren]. The addChild: method can just directly add the child. If we choose NSArray, the children method can directly return theChildren. However, the addChild method needs to mess about doing array= [theArray mutableCopy]; [array addObject:aChild]; [self setChildren:[[NSArray alloc] initWithArray:array]] ; i.e. two array copies! (Incidentally, I think it's wrong that copy copies the contents of the array but mutableCopy doesn't, since the concept of mutability in copying the array should apply to the array itself and not the array's contents, IMHO - but that's another issue.) Doing array copies is fine with a small number of children, but could be an unacceptable performance hit with a large (definition dependent on performance needs) number of children. It seems to me that the decision on whether to make the ivar mutable should be based on whether the class will be queried for its children more often than having children added / removed. But the problem is, that means your class is _internally_ implemented based on _external_ criteria, which seems contrary to OO principles. Any comments? -- If I wrote more, I'd probably make more sense. But time is limited.
Newsgroups: comp.sys.next.programmer From: gosman@netcom.com (www.servsol.com) Subject: http://www.servsol.com Message-ID: <gosmanDs2y7v.F8z@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Mon, 27 May 1996 19:59:55 GMT Sender: gosman@netcom11.netcom.com Please evaluate our home page www.servsol.com Thanks
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 28 May 1996 00:37:58 GMT Organization: Netcom Distribution: world Message-ID: <4odht6$14q@sjx-ixn6.ix.netcom.com> References: <31aa1cdb.31202737@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > - (NSArray *)children; > - (void)addChild:aChild; > The children method should return an immutable array (or to be precise > it should not return the exact same mutable array that the parent > stores its children in, otherwise the parent's children could be > changed without its knowledge). > Obviously, the class needs an instance variable for the children, but > the question is should this ivar (call it theChildren) be an NSArray > or NSMutableArray? > If we choose NSMutableArray, the children method must copy the > contents of theChildren into an immutable NSArray, using > [[NSArray alloc] initWithArray:theChildren]. The addChild: method can > just directly add the child. I would do the following in hopes that NSArray implements lazy copying: NSMutableArray *theChildren; - (NSArray *)children { return [[theChildren copy] autorelease]; } > If we choose NSArray, the children method can directly return > theChildren. However, the addChild method needs to mess about doing > array= [theArray mutableCopy]; > [array addObject:aChild]; > [self setChildren:[[NSArray alloc] initWithArray:array]] ; > i.e. two array copies! > Doing array copies is fine with a small number of children, but could > be an unacceptable performance hit with a large (definition dependent > on performance needs) number of children. This doesn't seem like the better approach as you've noted. -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 28 May 1996 09:33:24 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> References: <31aa1cdb.31202737@news.plsys.co.uk> <4odht6$14q@sjx-ixn6.ix.netcom.com> In-reply-to: aisbell@cubicsol.com's message of 28 May 1996 00:37:58 GMT In article <4odht6$14q@sjx-ixn6.ix.netcom.com> aisbell@cubicsol.com (Art Isbell) writes: I would do the following in hopes that NSArray implements lazy copying: NSMutableArray *theChildren; - (NSArray *)children { return [[theChildren copy] autorelease]; } I would do: NSMutableArray kids; -(NSArray *) children { return kids; } This is the simplest and fastest implementation, and anybody talking mutable methods to an NSArray should be fired. --Tiggr
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 28 May 1996 09:02:41 GMT Organization: Tech Net GmbH Message-ID: <4oeffh$6m5@ddfservb.technet.net> References: <31aa1cdb.31202737@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: >Kind of following on from Art Isbell's interesting Foundation Kit >question, I'd also like to pose a question that's been bothering me >about FK. > >The mutable / immutable distinction seems very useful at first sight. >However, I think it suffers from the following problem. > >Suppose a class has a collection of children, and we want it to >provide API for getting the children and adding a new child. > >For example, > >- (NSArray *)children; >- (void)addChild:aChild; > >The children method should return an immutable array (or to be precise >it should not return the exact same mutable array that the parent >stores its children in, otherwise the parent's children could be >changed without its knowledge). > >Obviously, the class needs an instance variable for the children, but >the question is should this ivar (call it theChildren) be an NSArray >or NSMutableArray? @interface SomeClass : Object { NSMutableArray *children; } - (NSArray *)children; - (void)addChild:aChild; @end @implementation SomeClass - (NSArray *)children { return chirdren; } - (void)addChild:aChild { [children addObject:aChild]; } Just store the children in a mutable array and return this array as an immutable array. This is 100% correct, because NSMutableArray is a subclass of NSArray. When your client code tries this: someObject=[[someClass alloc] initWithZone:[self zone]]; [[someObject children] addObject:aNewChild]; This will result in a compiler warning. This code is 100% incorrect, because the SomeClass class MAY return an object of type NSArray. But it's perfectly legal for you class to return an object of ANY subclass of NSArray. Greetings Constantin
From: "James C. Phillips" <jcp@viking.lanl.gov> Newsgroups: comp.sys.next.programmer Subject: source for mkfifo? Date: Wed, 22 May 1996 18:52:30 -0600 Organization: Los Alamos National Lab Message-ID: <31A3B6CE.7545@viking.lanl.gov> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I'd like to have the mkfifo command found on other unices. I tried writing my own with the following results: jim@monaco(164) cat mkfifo.c #define _POSIX_SOURCE #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> int main(int argc, char** argv) { mode_t mode = 0666; if ( argc != 2 ) { fprintf(stderr,"usage: %s file\n",argv[0]); return -1; } else return mkfifo(argv[1],mode); } jim@monaco(165) cc -o mkfifo mkfifo.c -lposix jim@monaco(166) mkfifo test Floating exception This works under solaris, so I don't know what's up? Does someone have a working mkfifo source they can send me? -Jim
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <gosmanDs2y7v.F8z@netcom.com> Control: cancel <gosmanDs2y7v.F8z@netcom.com> Date: 28 May 1996 08:51:11 GMT Organization: Turbocat's Development, Germany Message-ID: <4oeepv$cd@turbocat.snafu.de> cancel
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Contract/DC & ILL Date: 28 May 1996 12:16:34 GMT Organization: Norden 1 Communications Message-ID: <4oeqr2$85f@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------Commercial experience Objective C-----------------Commercial experience EOF-------------------------A plus Sybase or Oracle------------A plus Contracts-------------------Long & Mid term Areas-----------------------DC & ILL To Be Considered------------Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
Date: 28 May 1996 13:28:41 GMT From: clewis@ferret.ocunix.on.ca (Chris Lewis) Sender: news@tip.nl (The News User) Message-ID: <cancel.Ds1GDL.GM8@tip.nl> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <Ds1GDL.GM8@tip.nl> Control: cancel <Ds1GDL.GM8@tip.nl> MMF spam cancelled by clewis@ferret.ocunix.on.ca Original Subject: YOU SHOULD READ THIS
From: far@ix.netcom.net(Felipe A. Rodriguez) Newsgroups: comp.sys.next.programmer Subject: Re: source for mkfifo? Date: 28 May 1996 10:39:03 GMT Organization: Netcom Message-ID: <4oel47$f29@dfw-ixnews3.ix.netcom.com> References: <31A3B6CE.7545@viking.lanl.gov> In article <31A3B6CE.7545@viking.lanl.gov> "James C. Phillips" <jcp@viking.lanl.gov> writes: >Hi, > >I'd like to have the mkfifo command found on other unices. I tried >writing my own with the following results: > [snip] > >This works under solaris, so I don't know what's up? Does someone >have a working mkfifo source they can send me? > >-Jim The following source was sent to me when I faced a similar obstacle. The FIFO code in GateKeeper is based on this. /* Copyright (C) 1991 Free Software Foundation, Inc. This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with the GNU C Library; see the file COPYING. If not, write to the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ // #include <ansidecl.h> #include <stdio.h> #include <errno.h> #include <stddef.h> #include <sys/stat.h> #include <sys/types.h> typedef unsigned short mode_t; /* Create a named pipe (FIFO) named PATH with protections MODE. */ //#ifdef NEXT_MKFIFO_MAIN main(int argc, char **argv) { if (argc != 2) usage(argv); mkfifo(argv[1],(mode_t) 644); } usage(char **argv) { printf("Usage: %s PATH\n",argv[0]); exit(1); } //#endif int mkfifo (const char *path, mode_t mode) { return mknod (path, mode | S_IFIFO, 0); } -- Felipe A. Rodriguez # Francesco Sforza became Duke of Milan from Agoura Hills, CA # being a private citizen because he was # armed; his successors, since they avoided far@ix.netcom.com # the inconveniences of arms, became private (NeXTmail preferred) # citizens after having been dukes. (MIMEmail welcome) # --Nicolo Machiavelli
From: jfranco@montgomery.plato.sky.bdm.com (Jeff Franco,,,) Newsgroups: comp.sys.next.programmer Subject: Emacs and GDB Date: 28 May 1996 12:27:53 GMT Organization: BDM International, Inc. Message-ID: <4oerg9$g0u@bdmserver.mcl.bdm.com> I making the move from Editor to emacs and am having a bit of trouble using gdb. When I set a break point in a library I am using (that has been compiled fro debug) I get the following... (gdb) b JMValidator.m:789 Reading in symbols for JMValidator.m...done. Breakpoint 1 at 0x49ed8: JMValidator.m:789. (gdb) r Starting program: /jfranco/Projects/PDSS/Online/PDSSUser/PDSSUser.debug/PDSSUser Type emacs unknown TERM = (unknown) To all of you emacs fans out there, what does this mean? -- Thanks, Jeff Franco (703) 824-4301 Price Waterhouse LLP jfranco@plato.sky.bdm.com (NeXTMail, MIME)
Date: 28 May 1996 12:22:24 EST Newsgroups: comp.sys.next.programmer Message-ID: <cancel.gosmanDs2y7v.F8z@netcom.com> Control: cancel <gosmanDs2y7v.F8z@netcom.com> From: clewis@ferret.ocunix.on.ca Sender: gosman@netcom11.netcom.com Subject: cmsg cancel <gosmanDs2y7v.F8z@netcom.com> EMP/ECP (aka SPAM) cancelled by clewis@ferret.ocunix.on.ca. See news.admin.net-abuse.announce, report 19960528.10 for further details
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Tue, 28 May 1996 21:06:27 GMT Organization: Orbital Computer Consultancy Message-ID: <31ab666d.754158@news.plsys.co.uk> References: <31aa1cdb.31202737@news.plsys.co.uk> <4odht6$14q@sjx-ixn6.ix.netcom.com> aisbell@cubicsol.com (Art Isbell) wrote: >k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > >> - (NSArray *)children; >> - (void)addChild:aChild; > >> The children method should return an immutable array (or to be precise >> it should not return the exact same mutable array that the parent >> stores its children in, otherwise the parent's children could be >> changed without its knowledge). > >> Obviously, the class needs an instance variable for the children, but >> the question is should this ivar (call it theChildren) be an NSArray >> or NSMutableArray? > >> If we choose NSMutableArray, the children method must copy the >> contents of theChildren into an immutable NSArray, using >> [[NSArray alloc] initWithArray:theChildren]. The addChild: method can >> just directly add the child. > > I would do the following in hopes that NSArray implements lazy copying: > >NSMutableArray *theChildren; > >- (NSArray *)children >{ > return [[theChildren copy] autorelease]; >} > Interesting idea, but I would have thought that pretty soon after requesting the children, it's a safe bet the client code would actually access the array.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Tue, 28 May 1996 21:06:33 GMT Organization: Orbital Computer Consultancy Message-ID: <31ab6a4c.201639@news.plsys.co.uk> References: <31aa1cdb.31202737@news.plsys.co.uk> <4oeffh$6m5@ddfservb.technet.net> Constantin Szallies <szallies@energotec.de> wrote: >k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: <-- snip --> >someObject=[[someClass alloc] initWithZone:[self zone]]; >[[someObject children] addObject:aNewChild]; > id children = [someObject children]; [children addObject:aNewChild]; won't result in a compiler warning.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Tue, 28 May 1996 21:06:30 GMT Organization: Orbital Computer Consultancy Message-ID: <31ab69d0.78081@news.plsys.co.uk> References: <31aa1cdb.31202737@news.plsys.co.uk> <4odht6$14q@sjx-ixn6.ix.netcom.com> <x7ohn98d6j.fsf@tom.es.ele.tue.nl> tiggr@es.ele.tue.nl (Pieter Schoenmakers) wrote: >In article <4odht6$14q@sjx-ixn6.ix.netcom.com> aisbell@cubicsol.com (Art Isbell) writes: > > I would do the following in hopes that NSArray implements lazy copying: > > NSMutableArray *theChildren; > > - (NSArray *)children > { > return [[theChildren copy] autorelease]; > } > >I would do: > >NSMutableArray kids; > >-(NSArray *) children >{ > return kids; >} > >This is the simplest and fastest implementation, and anybody talking >mutable methods to an NSArray should be fired. --Tiggr Sorry, don't agree. A reusable class should protect itself against misuse, not rely on well-behaved client code. Any client code of your class could do: id children = [someObject children]; [children addObject:aChild]; without even a compiler warning.
From: chongl%sc@mhs.elan.af.mil Newsgroups: comp.sys.next.programmer Subject: I need help with netware Date: Tue, 28 May 1996 23:19:05 GMT Organization: United States Air Force Message-ID: <4og1of$cmr@rcp6.elan.af.mil> Hi, I have a netware connection from my Next to an email server and I'm writing this program that periodically fetches a datafile from the server. The problem is, the server periodically crashes causing me to have to reconnect using the executable nwlogin. So, my dumb solution is to nwlogin into the server just before fetching the data and then nwlogout afterwards all using the "system" command. In order for this to work, I had to setup an account on my email server so that I don't need to enter a password. So my code looks like this: system ("nwlogin SC-NGM2/stats"); NXApp = [NGMApp new]; [NXApp run: argv[1]]; system ("nwlogout SC-NGM2"); 1. Are there any doco on nwlogin? 2. Even without a password, nwlogin will prompt for one. Is there a way to echo a carriage return in C right after the command? 3. Is there a netware C library so that I don't have to use nwlogin to connect to my server? 4. I have Mach Rel 3.0 with 1.0 manuals. The Air Force had these sitting around when I got here. Do I have the right to obtain Rel 3.0 manuals from Next? Is Next still around? I'm running Next 68040/Rel 3.0/40.0MB ram/1041MB hd. Thanks! Leo Chong, 2Lt, USAF 95CS/SCSY chongl%sc@mhs.elan.af.mil
From: bmw@tesla.visgen.com (Bruce Walker) Newsgroups: comp.sys.next.programmer Subject: Re: grrr... _strdup Date: 28 May 1996 13:12:33 -0400 Organization: Visible Genetics Inc. Message-ID: <4ofc61$hvd@tesla.visgen.com> References: <4np707$cht@agate.berkeley.edu> In article <4np707$cht@agate.berkeley.edu>, Paul Robert Brown <pbrown@palms.berkeley.edu> wrote: >Anyone have a couple of lines of code for strdup() sitting around? I >found it once ("Handbook of C Functions" or something like that), but I >can't find it now... I can't figure why NeXT'd go to the trouble of >defining hypot() in math.h but leave out strdup... Yeah, it's a pain in the asymptote. This worked for me the other day when I was porting something... ----------------------------- #include <strings.h> char * strdup(const char *str) { char *new = (char *) malloc(strlen(str) + 1); if (new == NULL) return new; return strcpy(new, str); } ----------------------------- -- Bruce M. Walker | Visible Genetics Inc. | bmw@visgen.com | Suite 1000, Box 333 | Phone: 416-813-3240 | 700 Bay Street | FAX: 416-813-3250 | Toronto M5G 1Z6 | Pager: 416-609-6941
From: afi@d-tech.de (Alfred Fiedler) Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.soft-sys.nextstep Subject: Moving Subviews in Superviews Date: Wed, 29 May 1996 07:10:04 GMT Organization: Customer of EUnet Germany; Info: info@Germany.EU.net Distribution: inet Message-ID: <4ogsie$frd@Dortmund.Germany.EU.net> Hi all,, I do not have much practice in programming with Obj-C, but have a problem:. Assume there is a View (e.g. called background) containing some subviews. Now I want to move the subviews around in an arbitrary way.. I know I can do this by e.g. [subView1 moveBy:xLoc:yLoc];; Unfortunately by doing this "subView1" is presented twice: At the old and at the new position.. Ok, ok one solution to this problem is to erase the whole contents of superview and build the whole scenery from the very first beginning but with the new positions of subviews. This seems to be an ugly and slow solution. Please e-mail me, if there is a better or faster or more elegant solution for the described problem. Thanks in advance, Alfred.
Newsgroups: comp.sys.next.programmer From: pascal@burgond.remcomp.fr (Pascal Bourguignon) Subject: Re: Foundation memory management blues (longish ramble) Message-ID: <Ds04F8.1L2@burgond.remcomp.fr> Sender: pascal@burgond.remcomp.fr (Pascal Bourguignon) Organization: P. Bourguignon Informatique References: <4o511l$cl4@dfw-ixnews10.ix.netcom.com> Date: Sun, 26 May 1996 07:21:07 GMT In article <4o511l$cl4@dfw-ixnews10.ix.netcom.com> aisbell@cubicsol.com (Art Isbell) writes: > Most of my object architectures are graphs, not hierarchies. Peer > objects maintain references to each other so they can serve each other. How > does one decide which retains which? I haven't discovered any general > guidelines, so designing class graphs seems anarchic and error-prone. In the case of graphs, I would suggest to have an object (Graph) that keep the list of the nodes and retains them. Inside the graph, nodes can relate without retain. If a node needs a reference to another node from another graph, it should retain this other graph. > For example, if a class, MenuManager, has ivars pointing to various menu > cells, should MenuManager retain these menu cells? MenuManager didn't create > these menu cells, but if they are freed because some other object releases > them, MenuManager will send messages to freed objects. But this is true for > any object ivar, be it "intrinsic" or "extrinsic". But working with menu > cells seems central to MenuManager, so maybe menu cell ivars are intrinsic. In some cases, I just keep a list of dependants (a kind of delegate). Dependants don't retain the dependee. When some other object releases the dependee, it can message the dependants that it will be deleted. __Pascal Bourguignon__
From: mike@ali.bc.ca (Michael Cam) Newsgroups: comp.sys.next.programmer Subject: instance drawing Date: 28 May 1996 22:46:14 GMT Organization: ALI Distribution: world Message-ID: <4ofvnm$bo8@cetus.ali.bc.ca> Hi, I'm having a problem getting instance drawing to work. Can anyone tell me what the sequence of calls to PSsetinstance and PSnewinstance for this to work in a view class inside or outside of drawSelf::? I am using buffered windows. Thanks in advance. ..Mike. mike@ali.bc.ca -- ___________________________________________________________________ | | | ___ ^ ... /\ BEAUTIFUL | | _|_::| ___o '|`^ .. o_ . .. /\ / \ BRITISH | | |:::|:| \ \, ^ '|`|` (`_|/____') / / /\ COLUMBIA | | |:::|:| (o)/ (o) '|`'|`|`` ,,/ . ... . .. / \ |
From: kwong@morgan.ucs.mun.ca (Kai S. Wong) Newsgroups: comp.sys.next.programmer Subject: How to use the MiscSwapView and MiscTabMatrix? Date: 29 May 1996 13:02:36 GMT Organization: Memorial University of Newfoundland Message-ID: <4ohhtc$i37@coranto.ucs.mun.ca> Hello, I am looking for a sample code for using MiscSwapView and MiscTabMatrix. I have tried to open the nib file in Unix for Hater which uses SwapView and TabMatrix but I couldn't? Help needed? --KAI-- -- email: kwong@morgan.ucs.mun.ca url: http://web.cs.mun.ca/~kwong/
From: stes@kokomo.wri.com (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: Foundation memory management blues (longish ramble) Date: 29 May 1996 14:56:53 GMT Organization: Wolfram Research, Inc. Distribution: world Message-ID: <4ohojl$alo@dragonfly.wolfram.com> References: <Ds58u4.v7@lazrus.cca.rockwell.com> Erik M. Buck writes > > I do not think unarchived objects have a retain count of1. I always > have to retain > In the CAObject class of the Computer Algebra Kit, I write the reference count of the object being-archived to the stream. And restore it when being read in, again. This moves responsibility for setting the count to "1" from the unarchiver to the archiver. id myPolynomial = [aPolynomial deepCopy]; NXWriteObject(aTypedStream,myPolynomial); That's because of compatibility with other code, so that it can do a simple id anObject = NXReadObject(aTypedStream); It doesn't have to know that Computer Algebra Kit objects are actually reference counted. That's important for using e.g. Integer objects together with the Indexing Kit. David Stes
Newsgroups: comp.sys.next.programmer From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Subject: Re: Foundation memory management blues (longish ramble) Message-ID: <Ds58u4.v7@lazrus.cca.rockwell.com> Sender: news@lazrus.cca.rockwell.com Cc: aisbell@cubicsol.com Organization: Rockwell Avionics - Collins References: <4o511l$cl4@dfw-ixnews10.ix.netcom.com> Date: Wed, 29 May 1996 01:44:27 GMT Re: The retain count of unarchived objects I do not think unarchived objects have a retain count of1. I always have to retain them in - initWithCoder:. Re: The specific menu example The menu class must have a collection of menu cells. As a cell is added to the collection (by whatever means) it is retained. When the cell is removed from the collection, it is released. By making all cell add/remove operations go through the interface (encapsulation) of the collection, the issues of menus accessing freed cells are avoided. I would argue that the menu exclusively owns the menu cells and that no other object should retain them. (I regard the collection as an implementation detail of the menu) Even if another object allocates the menu cells and adds them to the menu, the menu should have exclusive right to manage the cells and/or reject them. ---------------- The design difficulties I experience with retain cycles are very uncommon and usually have to do with situations when I want to delete objects in mass and I can not be bothered to find all dependencies. For example, imagine a class that encapsulates connections between objects (like IBConnector). When any of the objects that are connected are released, the connection itself should be release also. It does not work the other way. If the connection is released, it should not matter to the objects that were connected. I hope I am clear... Anyway, connected objects can not/should not have any knowledge of connections yet they must notify connections when they are released so that the connection can be released. Further, if the connection retains the connected objects, then the connected objects can not be dealloced due to the connection. Just try to implement IBConnector using foundation! The solution is to make sure that the connectors do NOT retain the object they connect and use a notification system to inform connections that one or more connected objects has been released (dealloced!). - Erik M. Buck
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 29 May 1996 20:20:11 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x74tozpcis.fsf@tom.es.ele.tue.nl> References: <31aa1cdb.31202737@news.plsys.co.uk> <4odht6$14q@sjx-ixn6.ix.netcom.com> <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <31ab69d0.78081@news.plsys.co.uk> In-reply-to: k_harbour@mail.bogo.co.uk's message of Tue, 28 May 1996 21:06:30 GMT In article <31ab69d0.78081@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: Sorry, don't agree. A reusable class should protect itself against misuse, not rely on well-behaved client code. Any client code of your class could do: id children = [someObject children]; [children addObject:aChild]; without even a compiler warning. If the user of the class choses to ignore the fact that an array is returned, and not a mutable array, it is his/her problem. If you return a non-mutable array created upon request from the mutable one, the same user will knock on your door asking why your class is so slow. --Tiggr
From: drmac19@mail.idt.net (Dylan Blankenship) Newsgroups: comp.sys.next.programmer Subject: Nextstation For Sale Date: 29 May 1996 19:14:59 GMT Organization: ViViD Message-ID: <drmac19-0805961522200001@pmdyn73.vivid.net> Here we go!!!!!!!!! -------------------------------------------------------------- |NeXT Color Workstation =) | |-----------------------------| |megabytes of RAM |***32 |megabyte hard disk |***420 |monitor size(inches) |17" |monitor(b&w or color) |color |Sound box |yes |printer |laser |Keyboard |Full NeXT black keys in good condition |mouse |Black NeXT |Case |Black Slab NeXT pizza box |Disk Drive |1.44 DDHD 3.5" |OS |NextSTEP 3.0 ---------------------------------------------------------------- This configuration is for a NeXT color Workstation . I'm willing to trade for a macintosh powerbook or make an offer. ******TRADE OR FOR SALE********** _/\/\/\/-Visit Static Realm-/\/\/\/\/_ -- drmac19@mail.idt.net _/\/\/-Static_Realm_/\/\/\-
From: jfranco@montgomery.plato.sky.bdm.com (Jeff Franco,,,) Newsgroups: comp.sys.next.programmer Subject: Re: insert Null value Date: 29 May 1996 19:41:17 GMT Organization: BDM International, Inc. Message-ID: <4oi98t$g5f@bdmserver.mcl.bdm.com> References: <4o4h39$3db@hpuniv.univ-lr.fr> In-Reply-To: <4o4h39$3db@hpuniv.univ-lr.fr> On 05/24/96, Bruno Garnier wrote: >Hi all, > >I work with InformixAdaptor. >When I want to insert an [EONull null] value in a field in float format of my >EOController, the inserted value is 0.00. >I Have no default value in my table. >How can I insert a null instead of 0.00 ?? > >Please help ... > I've never worked with the Informix adaptor/database, but you might want to verify that Informix accepts nulls for scalars. If your database does accept nulls, then look at the class you controller is returning. If it is a custom EO and not EOGenericRecord you may have to override the unableToSetNullForKey:(NSString *)key method in EOKeyValueCoding.h. Or yet another solution would be do do away with ints, floats and doubles all together and use NSNumbers. Our database accepts nulls for scalars and we have gone with the last option. Hope this helps. -- Thanks, Jeff Franco (703) 824-4301 Price Waterhouse LLP jfranco@plato.sky.bdm.com (NeXTMail, MIME)
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 29 May 1996 19:37:44 GMT Organization: Tech Net GmbH Message-ID: <4oi928$hec@ddfservb.technet.net> References: <31aa1cdb.31202737@news.plsys.co.uk> <4oeffh$6m5@ddfservb.technet.net> <31ab6a4c.201639@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: >Constantin Szallies <szallies@energotec.de> wrote: > >>k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: ><-- snip --> >>someObject=[[someClass alloc] initWithZone:[self zone]]; >>[[someObject children] addObject:aNewChild]; >> > id children = [someObject children]; > > [children addObject:aNewChild]; > >won't result in a compiler warning. This is not the point. This code is incorrect. NXAssray doesn't implement addObject:. So you don't know if your app crashes here or not. Greeting Constantin Szallies
From: lbo@sma.ch (-Laurent Bourqui) Newsgroups: comp.sys.next.programmer Subject: How to write a printer driver Date: 29 May 1996 12:50:32 GMT Organization: Swiss Meteorological Institute Distribution: world Message-ID: <4ohh6oINN3ck@maz4.sma.ch> Hi all, I would like to try to write a driver for my printer. Where should I begin ? I already looked trough the driverkit but it seems that there is no support for printer driver. Did someone already do something like that or has some experience in this direction ? Any help will be appreciate. Laurent --- -------------------------------------------------------------------------------- Laurent Bourqui | Phone (private) +41(0)1 451 62 92 SMI (Swiss Meteorological Institute) | Phone (office) +41(0)1 256 95 73 Kraebuehlstr. 58 | Fax (office) +41(0)1 256 92 78 | email lbo@sma.ch CH-8044 Zuerich | Switzerland | --------------------------------------------------------------------------------
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Wed, 29 May 1996 21:47:39 GMT Organization: Orbital Computer Consultancy Message-ID: <31acc41e.1614449@news.plsys.co.uk> References: <31aa1cdb.31202737@news.plsys.co.uk> <4odht6$14q@sjx-ixn6.ix.netcom.com> <31ab666d.754158@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: >aisbell@cubicsol.com (Art Isbell) wrote: > >>k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: >> >>> - (NSArray *)children; >>> - (void)addChild:aChild; >> >>> The children method should return an immutable array (or to be precise >>> it should not return the exact same mutable array that the parent >>> stores its children in, otherwise the parent's children could be >>> changed without its knowledge). >> >>> Obviously, the class needs an instance variable for the children, but >>> the question is should this ivar (call it theChildren) be an NSArray >>> or NSMutableArray? >> >>> If we choose NSMutableArray, the children method must copy the >>> contents of theChildren into an immutable NSArray, using >>> [[NSArray alloc] initWithArray:theChildren]. The addChild: method can >>> just directly add the child. >> >> I would do the following in hopes that NSArray implements lazy copying: >> >>NSMutableArray *theChildren; >> >>- (NSArray *)children >>{ >> return [[theChildren copy] autorelease]; >>} >> > >Interesting idea, but I would have thought that pretty soon after >requesting the children, it's a safe bet the client code would >actually access the array. I knew I replied too quickly here! The whole reason I did the initWithArray: stuff in my example code, rather than copy, is because if you send copy to an NSMutableArray the objects within the array are all sent copy messages as well - which of course is very unfortunate if these objects don't implement copy. -- If I wrote more, I'd probably make more sense. But time is limited.
Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.soft-sys.nextstep From: nitezki@nidat.sub.org (Peter Nitezki) Subject: Re: Why NeXT? Message-ID: <Ds6IuL.LLv@nidat.sub.org> Sender: news@nidat.sub.org Cc: reichman@usc.edu Organization: private site of Peter Nitezki, Kraichtal, Germany References: <4norkd$97v@digifix.digifix.com> <Pine.ULT.3.91.960524163016.12494A-100000@cello13.orchestra.cse.unsw.EDU.AU> <DryK56.G0B@nidat.sub.org> <4o7i1i$7ku@usc.edu> Distribution: inet Date: Wed, 29 May 1996 18:18:18 GMT In <4o7i1i$7ku@usc.edu> Matthew N. Reichman wrote: > In <DryK56.G0B@nidat.sub.org> Peter Nitezki wrote: > > As a summary: NEXTSTEPpers are bored by the request to talk someboy into it > > since they can't understand why people don't just come and see. Or let it > > put me that way: it's the same as with oysters. If you don't like'em half a > > dozen is too much, and if you like'em half a dozen is not enough. > > You are definitely a riot, Peter! 8-) > > Ok, ok. I plead guilty of not always applying the amount humble forbearance I've learned to devote towards c.s.n.a postings for postings outside said newsgroup ;-) But I must insist that I was not geniunely out for a flame war. Sometimes I just can't stand NEXTSTEP being mentioned with Microshit in one breath. >:-| But yes, I know this won't help with my karma :-P -- Peter Nitezki | Nitezki@NiDat.sub.org # Blessed art thou who knoweth Staarenbergstr. 44 | Tel.: +49 7251 62495 # not about the pleasure and D-76703 Kraichtal | Fax : +49 7251 69215 # delight of being hooked GERMANY | E-mail defunct, sorry # up to the Net. Peter 1,3-5
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 30 May 1996 02:36:42 GMT Organization: Netcom Distribution: world Message-ID: <4oj1jq$329@sjx-ixn6.ix.netcom.com> References: <31aa1cdb.31202737@news.plsys.co.uk> <4oeffh$6m5@ddfservb.technet.net> <31ab6a4c.201639@news.plsys.co.uk> <4oi928$hec@ddfservb.technet.net> Constantin Szallies <szallies@energotec.de> wrote: > k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > >Constantin Szallies <szallies@energotec.de> wrote: > > > >>k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > ><-- snip --> > >>someObject=[[someClass alloc] initWithZone:[self zone]]; > >>[[someObject children] addObject:aNewChild]; > >> > > id children = [someObject children]; > > > > [children addObject:aNewChild]; > > > >won't result in a compiler warning. > This is not the point. > > This code is incorrect. NXAssray doesn't implement addObject:. So you don't > know if your app crashes here or not. But children is a NSMutableArray, so it does implement addObject: and the app won't crash. Strong typing of children to NSArray to match the declaration of the children method would have produced a compiler warning, but not an error, so the code would be built and would run fine allowing children to be modified. So do we want to prevent this from happening, or do we want to merely lay the blame on the programmer? -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: ggarrabe@terre.univ-lr.fr (Gilles Garrabella) Newsgroups: comp.sys.next.programmer Subject: How to write and read an NSArray in a File ? Date: 30 May 1996 09:07:05 GMT Organization: Universite de La Rochelle Message-ID: <4ojofp$8nj@hpuniv.univ-lr.fr> Keywords: foundation , nextStep , file I'm a novice in objective-C programmation with Interface Builder and i don't now how to write and read an NSArray in a file. I try an NXOpenTypedStreamForFile() to open the file and a NXWriteNSObject() to write the NSArray in but when i try to read the file and print the NSArray i have this error : NSInvalidArgumentException: *** -[NSMutableDictionary setObject:forKey:]: nil object for key 'entity' Maybe, a NSCoding is necessary but how to make it ? Thanks to reply in NewsGrazer or by mail to ggarrabe@cri.univ-lr.fr
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: I need help with netware Date: Wed, 29 May 1996 22:50:07 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996May29.225007.20949@seer.demon.co.uk> References: <4og1of$cmr@rcp6.elan.af.mil> In article <4og1of$cmr@rcp6.elan.af.mil> chongl%sc@mhs.elan.af.mil writes: > 1. Are there any doco on nwlogin? man nwlogin > 2. Even without a password, nwlogin will prompt for one. Is there a > way to echo a carriage return in C right after the command? The password stuff is quite likely trying to read from the tty device, so you may have a problem hiding this. > 3. Is there a netware C library so that I don't have to use nwlogin > to connect to my server? Look in /NextLibrary/Headers/netware; I think they go with libnw. Not very easy to read, I am afraid. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: Norbert Heger <bertl@hal.kph.tuwien.ac.at> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 30 May 1996 11:57:36 GMT Organization: Vienna University of Technology, Austria Message-ID: <4ok2fh$bln@news.tuwien.ac.at> References: <31ab6a4c.201639@news.plsys.co.uk> Originator: bertl@black13 Karl Harbour wrote: > id children = [someObject children]; > [children addObject:aNewChild]; > > won't result in a compiler warning. It's a question of programming style. Casting typed object pointers to typeless id's is always a dangerous thing that may lead to runtime errors. id children = [[NSObject alloc] init]; [children addObject:aNewChild]; won't result in a compiler warning either... _________________________________________________________________ Norbert Heger <bertl@hal.kph.tuwien.ac.at> NEXTSTEP / OpenStep Software Development NeXTmail preferred, MIME is welcome Please finger for PGP public key PGP fingerprint = E0 10 3C BB 67 4C 7E 82 76 99 E9 6D B1 8C 32 C2
From: Matthias Klose <doko@cs.tu-berlin.de> Newsgroups: comp.lang.objective-c,comp.lang.tcl,comp.sys.next.programmer Subject: dynamic loading problem with Tcl7.5 and ObjC Date: 30 May 1996 19:21:40 +0200 Organization: Technical University of Berlin, Germany Message-ID: <78ivde3wm3.fsf@cs.tu-berlin.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit I have a problem with loading a Tcl module (via the tcl load command), which contains ObjC code. The first method call in the dynamically loaded module core dumps in the objc_msg_lookup function with a segmentation fault. The same method call from the main function succeeds. The parameters passed to this function are different in both calls (but should be the same). In the loaded module no classes and categories are defined. Is there anything of the objc stuff to initialize in the loaded module? I tried this an Solaris 2.5, Tcl7.5, gcc-2.7.2 and the gnu-objc runtime with the thread patch applied. Appended is a short shar file with demonstrates the behaviour. #!/bin/sh # This is a shell archive (produced by shar 3.49) # To extract the files from this archive, save it to a file, remove # everything above the "!/bin/sh" line above, and type "sh file_name". # # made 05/30/1996 17:18 UTC by doko@gargleblaster # Source directory /home/gargleblaster/doko/uf4/development/runtime/to # # existing files will NOT be overwritten unless -c is specified # # This shar contains: # length mode name # ------ ---------- ------------------------------------------ # 907 -rw-r--r-- to.m # 459 -rw-r--r-- bar.m # 382 -rw-r--r-- Makefile # # ============= to.m ============== if test -f 'to.m' -a X"$1" != X"-c"; then echo 'x - skipping to.m (File already exists)' else echo 'x - extracting to.m (Text)' sed 's/^X//' << 'SHAR_EOF' > 'to.m' && #include <tcl.h> //#include <Foundation/NSString.h> #include <objc/Object.h> X @interface Foo: Object { X int x; } - (Foo *) initX: (int) anX; - (int) getX; @end X @implementation Foo - (Foo *) initX: (int) anX { X x = anX; X return self; } X - (int) getX; { X return x; } @end X static Tcl_Interp *interp; X Foo* aFoo; X int main (int argc, char *argv[], char *envp[]) { X interp = Tcl_CreateInterp(); X if (Tcl_Init(interp) == TCL_ERROR) { X fprintf(stderr, "%s\n", interp->result); X return 1; X } X X aFoo = [[Foo alloc] initX: 12345]; X printf("x = %d\n", [aFoo getX]); X X if (Tcl_Eval(interp, X "load /home/swt/doko/uf4/development/runtime/to/libbar.so") X != TCL_OK) { X fprintf(stderr, "%s\n", interp->result); X return 1; X } X X if (Tcl_GlobalEval(interp, "bar arg1 arg2") != TCL_OK) { X fprintf(stderr, "%s\n", interp->result); X return 1; X } X X printf("done\n"); X X return 0; } SHAR_EOF chmod 0644 to.m || echo 'restore of to.m failed' Wc_c="`wc -c < 'to.m'`" test 907 -eq "$Wc_c" || echo 'to.m: original size 907, current size' "$Wc_c" fi # ============= bar.m ============== if test -f 'bar.m' -a X"$1" != X"-c"; then echo 'x - skipping bar.m (File already exists)' else echo 'x - extracting bar.m (Text)' sed 's/^X//' << 'SHAR_EOF' > 'bar.m' && #include <tcl.h> X @class Foo; X extern Foo *aFoo; X static int barCmd(ClientData clientData, X Tcl_Interp *interp, X int argc, char *argv[]) { X int x; X X printf("bar called\n"); X X x = [aFoo getX]; X printf("x = %d\n", x); X X return TCL_OK; } X int Bar_Init(Tcl_Interp *interp) { X printf("bar inited\n"); X Tcl_CreateCommand(interp, "bar", barCmd, (ClientData) NULL, X (Tcl_CmdDeleteProc *) NULL); X X return TCL_OK; } SHAR_EOF chmod 0644 bar.m || echo 'restore of bar.m failed' Wc_c="`wc -c < 'bar.m'`" test 459 -eq "$Wc_c" || echo 'bar.m: original size 459, current size' "$Wc_c" fi # ============= Makefile ============== if test -f 'Makefile' -a X"$1" != X"-c"; then echo 'x - skipping Makefile (File already exists)' else echo 'x - extracting Makefile (Text)' sed 's/^X//' << 'SHAR_EOF' > 'Makefile' && TCL=/usr/tcl/4.1 X all: to libbar.so X to: to.o X gcc -g -o to to.o -L$(TCL)/lib -R$(TCL)/lib -ltcl7.5 \ X /usr/gnu/src/gcc-2.7.2/sol2-stabs/objc/libobjc.a -lm X to.o: to.m X gcc -g -c -I$(TCL)/include -I$(OLIB)/include to.m X bar.o: bar.m X gcc -g -c -fpic -DPIC -I$(TCL)/include -I$(OLIB)/include bar.m X libbar.so: bar.o X ld -G -o libbar.so bar.o \ X -L$(TCL)/lib -R$(TCL)/lib -ltcl7.5 SHAR_EOF chmod 0644 Makefile || echo 'restore of Makefile failed' Wc_c="`wc -c < 'Makefile'`" test 382 -eq "$Wc_c" || echo 'Makefile: original size 382, current size' "$Wc_c" fi exit 0
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: How to use the MiscSwapView and MiscTabMatrix? Date: 30 May 1996 18:29:23 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4okpe3$hj3@nntp.onramp.net> References: <4ohhtc$i37@coranto.ucs.mun.ca> > I am looking for a sample code for using MiscSwapView and > MiscTabMatrix. I have tried to open the nib file in Unix for Hater > which uses SwapView and TabMatrix but I couldn't? Help needed? > > --KAI-- You may need to load those palettes into Interface Builder first. Double-clicking on the .palette files for them should do the trick. -- Steve Dekorte "C++, the COBOL of the 90s"- UHH NeXTstep consultant - Anaheim, CA mailto:dekorte@suite.com (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 30 May 1996 02:48:38 GMT Organization: Netcom Distribution: world Message-ID: <4oj2a6$329@sjx-ixn6.ix.netcom.com> References: <31aa1cdb.31202737@news.plsys.co.uk> <4odht6$14q@sjx-ixn6.ix.netcom.com> <31ab666d.754158@news.plsys.co.uk> <31acc41e.1614449@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > > >aisbell@cubicsol.com (Art Isbell) wrote: > > > >>k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > >> > >>> - (NSArray *)children; > >>> - (void)addChild:aChild; > >> > >>> The children method should return an immutable array (or to be precise > >>> it should not return the exact same mutable array that the parent > >>> stores its children in, otherwise the parent's children could be > >>> changed without its knowledge). > >> > >>> Obviously, the class needs an instance variable for the children, but > >>> the question is should this ivar (call it theChildren) be an NSArray > >>> or NSMutableArray? > >> > >>> If we choose NSMutableArray, the children method must copy the > >>> contents of theChildren into an immutable NSArray, using > >>> [[NSArray alloc] initWithArray:theChildren]. The addChild: method can > >>> just directly add the child. > >> > >> I would do the following in hopes that NSArray implements lazy copying: > >> > >>NSMutableArray *theChildren; > >> > >>- (NSArray *)children > >>{ > >> return [[theChildren copy] autorelease]; > >>} > >> > > > >Interesting idea, but I would have thought that pretty soon after > >requesting the children, it's a safe bet the client code would > >actually access the array. > > I knew I replied too quickly here! The whole reason I did the > initWithArray: stuff in my example code, rather than copy, is because > if you send copy to an NSMutableArray the objects within the array are > all sent copy messages as well - which of course is very unfortunate > if these objects don't implement copy. > -- > > If I wrote more, I'd probably make more sense. I'll write more, but it probably won't make much more sense than my original suggestion :-) Karl is correct. Sending a mutable array a copy message will perform a deep copy without any apparent copy-on-write efficiency. So even if the object receiving the copied array doesn't modify the array, a needless copy will have been made. So much for copy-on-write :-( Furthermore, the deep copy of a mutable array results in every element being sent a copy message. copy is inherited from NSObject and implemented using copyWithZone: which raises if any array element doesn't override it :-( So copy a mutable array only if you're certain that every element has overridden copyWithZone: or you're prepared to catch and deal with an exception. This seems likely to crash a few Foundation apps... Karl's suggestion is an intermediate solution that will create an independent immutable array, but that shares all array elements with the original mutable array. Whether this is acceptable probably depends on the situation. But at least the app won't crash. Foundation is feeling more like a mine field the deeper I venture :-( -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: tmb@best.com (.) Newsgroups: comp.sys.next.programmer Subject: Re: Foundation memory management blues (longish ramble) Date: 30 May 1996 05:19:48 GMT Organization: home Distribution: world Message-ID: <TMB.96May29221948@best.best.com> References: <4o511l$cl4@dfw-ixnews10.ix.netcom.com> In-reply-to: aisbell@cubicsol.com's message of 24 May 1996 19:01:09 GMT FCC: /home/tmb/mail/x-nout In article <4o511l$cl4@dfw-ixnews10.ix.netcom.com> aisbell@cubicsol.com (Art Isbell) writes: Periodically, I post because I still don't get it when it comes to using Foundation's memory management scheme correctly. I'm still looking for the magic key to understanding. Not having this key is very frustrating. [...] Most of my object architectures are graphs, not hierarchies. Peer objects maintain references to each other so they can serve each other. How does one decide which retains which? I haven't discovered any general guidelines, so designing class graphs seems anarchic and error-prone. You may want to see whether you can use Boehm's GC (parcftp.xerox.com:pub/gc). I have used it with Objective-C and threaded programs on other platforms, and it works like a charm. Cheers, Thomas.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <eott.com!tappd@nuchat.sccsi.com> Message-ID: <9605301401.AA22481@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@eott.com> Date: Thu, 30 May 96 09:01:53 -0500 Subject: Simple Animation Timing - Epilogue I wanted to thank everybody who provided advice (and sample code, and entire class definitions!) for my animation problem early last week. You can probably visualize the results...a textured corporate logo, spinning smoothly in the corner of a data-entry panel...but I'm not sure you can imagine the look on my Systems Director's face when he saw it. I wish I had a snapshot I could post. His reaction followed an entirely predictable sequence: 1. I can't believe you actually spent time doing that totally superfluous, ridiculous thing, and 2. we need a couple of changes to make it **really** cool, before we roll this app out. Quick note to POV users...If you'll do this in your scene file: #declare Next_Gray = color red 170/255 green 170/255 blue 170/255 background { color Next_Gray } you should get a scene that merges pretty-nigh perfectly with the gray on your panels. ('Course, you'll probably want to render the frames on hardware that's as close as possible to the stuff that'll be running your app, just to be certain.) Thanks, all! - Dan
From: dboyce@albert (Douglas W Boyce) Newsgroups: comp.sys.next.programmer Subject: Bad gdb Date: 30 May 1996 19:36:22 GMT Organization: a Digital Internet AlphaServer Site Message-ID: <4oktbm$i3n@ns1.moran.com> I have a project with 3 bundle sub-projects. Gdb suddenly refuses to recognize any sources in the sub-projects in order to set breakpoints. The standard error is: No source file named FileName.m. Error occured processing command from Edit. Anyone ever have this problem with a valid PB.gdbinit file. NS3.3 Intel, with user and developer patches applied.
From: tmb@best.com (.) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 30 May 1996 19:30:56 GMT Organization: home Message-ID: <TMB.96May30123056@best.best.com> References: <31ab6a4c.201639@news.plsys.co.uk> <4ok2fh$bln@news.tuwien.ac.at> In-reply-to: Norbert Heger's message of 30 May 1996 11:57:36 GMT FCC: /home/tmb/mail/x-nout In article <4ok2fh$bln@news.tuwien.ac.at> Norbert Heger <bertl@hal.kph.tuwien.ac.at> writes: It's a question of programming style. Casting typed object pointers to typeless id's is always a dangerous thing that may lead to runtime errors. id children = [[NSObject alloc] init]; [children addObject:aNewChild]; won't result in a compiler warning either... Well, I'm sure we agree on the technical issues, but I would put this differently. Casting typed object pointers to "typeless id's" is not dangerous, but then applying methods to "typeless id's" can result in runtime errors; those runtime errors are not all that dangerous, though, since they are reliably detected. That's a choice and tradeoff people make when using dynamic typing. Without some form of dynamic typing, UI programming, distributed programming, and dynamic loading become headaches. What is more dangerous is casting id to typed pointers (downcasts) and then using those pointers, in particular to access instance variables. You will, in general, not get notified of runtime errors and just get crashes or incorrect results if you make a mistake. Cheers, Thomas.
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advoocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 30 May 1996 17:28:58 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4oklsq$3ge@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> "Jonathan W. Hendry" <steeldrv@ix.netcom.com> wrote: > Uli Zappe wrote: > > Well, they showed it at CeBIT. Problem is, it's Mail.app, but without > > NeXTMail :-((( > > Does it do MIME? If it does MIME well, who cares if it does NeXTMail? Me. I really prefer NeXTMail to MIME because the latter for my purposes has severe shortcomings in layout possibilities. One of the great advantages of NEXTSTEP is that things work all the same. NeXTMail works like Edit.app; MIME doesn't. I hate that. :-( Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: Foundation memory management blues (longish ramble) Date: 30 May 1996 02:30:01 GMT Organization: Netcom Distribution: world Message-ID: <4oj179$329@sjx-ixn6.ix.netcom.com> References: <4o511l$cl4@dfw-ixnews10.ix.netcom.com> <Ds58u4.v7@lazrus.cca.rockwell.com> embuck@palmer.cca.rockwell.com (Erik M. Buck) wrote: > Re: The retain count of unarchived objects > > I do not think unarchived objects have a retain count of1. I always have to > retain > them in - initWithCoder:. Objects unarchived programmatically are not retained if decodeObject is used, but are already retained if decodeValuesOfObjCTypes: is used. This unfortunate inconsistency occurs to maintain backward compatibility. Seems that in this period of massive API changes, this backward compatibility could have been sacrificed for future clarity. Apparently, the various loadNib... methods use decodeValuesOfObjCTypes: so that objects unarchived from nibs are already retained. Just another minor detail to deal with... -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 30 May 1996 22:17:07 GMT Organization: ALI Distribution: world Message-ID: <4ol6p3$bv4@cetus.ali.bc.ca> References: <31aa1cdb.31202737@news.plsys.co.uk> <4oeffh$6m5@ddfservb.technet.net> <31ab6a4c.201639@news.plsys.co.uk> <4oi928$hec@ddfservb.technet.net> <4oj1jq$329@sjx-ixn6.ix.netcom.com> Cc: aisbell@cubicsol.com In comp.sys.next.programmer Art Isbell wrote: > Constantin Szallies <szallies@energotec.de> wrote: > > k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > > >Constantin Szallies <szallies@energotec.de> wrote: > > > > > >>k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: > > ><-- snip --> > > >>someObject=[[someClass alloc] initWithZone:[self zone]]; > > >>[[someObject children] addObject:aNewChild]; > > >> > > > id children = [someObject children]; > > > > > > [children addObject:aNewChild]; > > > > > >won't result in a compiler warning. > > This is not the point. > > > > This code is incorrect. NXAssray doesn't implement addObject:. So you don't > > know if your app crashes here or not. > > But children is a NSMutableArray, so it does implement addObject: and the > app won't crash. Strong typing of children to NSArray to match the > declaration of the children method would have produced a compiler warning, > but not an error, so the code would be built and would run fine allowing > children to be modified. > > So do we want to prevent this from happening, or do we want to merely lay > the blame on the programmer? > If you are really worried about this, it would be a trivial excercise to create a subclass of NSArray (ie. immutable) which simply contains a reference to another array (ie. mutable) which actually contains the data. This would be a lot cheaper than creating a copy of the original array and would still prevent addObject: et. al. from being called by the unwashed masses. It would get tedious however to have to do this with every mutable/immutable pair of classes, however. So this isn't a great solution. Also, the programmer could still make assumptions about the constancy of the number of elements and other attributes of the stand-in array which wouldn't be true since objects which are allowed to reference the "real" array could change its characteristics at will. The stand-in would reflect these changes the next time it was asked for information (since it just forwards the methods on to the mutable array) thus blowing the programmers assumptions out of the water. ciao -- Allan Noordvyk, Software Artisan e-mail: allan@ali.bc.ca ALI Technologies Voice: 604.279.5422 x 317 Richmond, Canada Fax: 604.279.5468 * NeXT and MIME mail welcome *
Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer From: Sean Willson <sean_willson@il.us.swissbank.com> Subject: Reality Check....NeXTSTEP and OpenStep Message-ID: <1996May30.222007.21443@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Thu, 30 May 1996 22:20:07 GMT Fellow NeXT Users, This could be a rehash of other posting's, but I have a few questions regarding NeXT Software's direction on their development tools. With the recent talk about Tsunami and the fact that it is a NT only product, what is in store for Solaris and NeXTSTEP besides as deployment environments? How well will NeXT be able to compete in the NT market while maintaining ties with the MACH and Solaris installed base? I myself love NeXTStep and the technologies used in it, but am beginning to see the light. Its dimming.... As a developer and advocate of NeXTSTEP, what can I expect to embrace in the future, other than Java and the supposed Obj-Pascal stuff? I am not impress by either, and would like to stay in Obj-C. Will the dev environment that NeXT is going to deploy on NT be worth the investment in time and money? Do other developers think that it will compete with the likes of Borland, Microsoft, and Java in the NT marketplace? Is NeXT focusing just on the web, as it appears? Sorry for the collage of questions, but coming from a soon to graduate student, I am lacking a bit of focus.... Sean Willson -- ____________________________________________________________________ Sean M. Willson sean_willson@il.us.swissbank.com Swiss Bank Corp. NeXT, MIME, and ascii 141 West Jackson Boulevard Chicago, Illinois 60604 (312) 554-5399
From: pascal@Galileo.slip.u-bordeaux.fr (Pascal Thibaudeau) Newsgroups: comp.sys.next.programmer Subject: dlopen and dlsym Date: 30 May 1996 19:48:01 GMT Organization: CRIBX1 , Universite de Bordeaux I , France Message-ID: <4oku1h$bfh@news.u-bordeaux.fr> -- -Pascal Thibaudeau ********************************************************************** E-Mail: pthibaud@frbdx11.cribx1.u-bordeaux.fr ( NO NEXTMAIL ) pascal@galileo.slip.u-bordeaux.fr ( Small NEXTMAIL OK )
From: pascal@Galileo.slip.u-bordeaux.fr (Pascal Thibaudeau) Newsgroups: comp.sys.next.programmer Subject: dlopen and dlsym Date: 30 May 1996 19:49:23 GMT Organization: CRIBX1 , Universite de Bordeaux I , France Message-ID: <4oku43$bfi@news.u-bordeaux.fr> Hello, What are the NS equivalents of dlopen(s,m) and dlsym(l,s) ? -- -Pascal Thibaudeau ********************************************************************** E-Mail: pthibaud@frbdx11.cribx1.u-bordeaux.fr ( NO NEXTMAIL ) pascal@galileo.slip.u-bordeaux.fr ( Small NEXTMAIL OK )
From: cdl@proxima.ucsd.edu (Carl Lowenstein) Newsgroups: comp.sys.next.programmer Subject: shell script as .app Date: 30 May 1996 23:22:34 GMT Organization: The Avant-Garde of the Now, Ltd. Message-ID: <4olajq$r9i@news1.ucsd.edu> I would like to do something which should be really simple. Namely, use a shell script as a genuine NeXT point-and-click .app. Do I have to write an Objective-C wrapper for the script, and hook it all up with InterfaceBuilder? Or is there something more obvious that I am overlooking? -- carl lowenstein marine physical lab u.c. san diego clowenstein@ucsd.edu
From: pascal@galileo.slip.u-bordeaux.fr (Pascal Thibaudeau) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer,fr.comp.sys.next Subject: Help: Preferences -> Bus error Date: 31 May 1996 07:49:08 GMT Organization: CRIBX1 , Universite de Bordeaux I , France Message-ID: <4om89k$g0h@news.u-bordeaux.fr> Hello When I lauch Preferences.app or Configure.app I have a Bus error and as Root, the dread -l returns: #dread -l dread: Can't open defaults database #./Preferences.app/Preferences Bus error #./Configure.app/Configure Bus error # How can I do now ? Thanks. -- -Pascal Thibaudeau ********************************************************************** E-Mail: pthibaud@frbdx11.cribx1.u-bordeaux.fr ( NO NEXTMAIL ) pascal@galileo.slip.u-bordeaux.fr ( Small NEXTMAIL OK )
From: Mark Anenberg <marka@Eng.Sun.COM> Newsgroups: comp.sys.next.programmer,comp.sys.next.advoocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Fri, 31 May 1996 09:19:00 -0700 Organization: Sun Microsystems Inc. Message-ID: <31AF1BF3.4ECA@Eng.Sun.COM> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Jonathan W. Hendry wrote: > > Uli Zappe wrote: > > > > mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > > > And here were some of us thinking that Mail.app wouldn't be part of Sun's > > > OpenStep toolkit... ;-) > > > > Well, they showed it at CeBIT. Problem is, it's Mail.app, but without > > NeXTMail :-((( > > > > (SUN: PLEASE tell me that's not true!) > > Does it do MIME? If it does MIME well, who cares if it does NeXTMail? It still reads NeXTMail messages just fine. But you can send only PlainText or Mime. If you forward a NeXTMail message it can become Mime and look just fine within Mail.app. The way NeXT has done Mime you still get all your fonts, images and attachments, just like in NeXTMail format; If you read it with another Mime compliant mail reader, you may get the non-text versions as attachments (depending on your other reader and whether you have mime-alternatives enabled). -- Mark Anenberg , OpenStep Development Team, SunSoft, Inc. [E,NeXT,Mime]Mail: mark@velodrome.com, http://www.velodrome.com/ Disclaimer: The opinions expressed above are my own and in no way represent those of Sun Microsystems, Inc.
From: Charles F Waltrip <waltrip@zephyr.jhuapl.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: Reality Check....NeXTSTEP and OpenStep Date: Fri, 31 May 1996 14:23:18 -0400 Organization: The Johns Hopkins University Applied Physics Lab Message-ID: <31AF3916.794B@zephyr.jhuapl.edu> References: <1996May30.222007.21443@il.us.swissbank.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sean Willson wrote: > > Fellow NeXT Users, > > This could be a rehash of other posting's, but I have a few > questions regarding NeXT Software's direction on their development > tools. With the recent talk about Tsunami and the fact that it is a > NT only product, what is in store for Solaris and NeXTSTEP besides > as deployment environments? I wouldn't make too much of this for now. Tsunami is an OpenStep app and, whatever NeXT's plans or decisions turn out to be, it would be premature to announce it's availability for anything other than NT at this point. It's good to express your concern, however, which I second. > How well will NeXT be able to compete > in the NT market while maintaining ties with the MACH and Solaris > installed base? I myself love NeXTStep and the technologies used in > it, but am beginning to see the light. Its dimming.... As it stands now, OpenStep on Mach/Solaris/Digital UNIX(??)/HPUX(??) represents a very stable development environment for NT. Many would argue that it's the ONLY stable development environment for NT. Unless/until NT itself becomes more stable, OpenStep on other OSs will likely continue to have a growing following. > > As a developer and advocate of NeXTSTEP, what can I expect > to embrace in the future, other than Java and the supposed Obj-Pascal > stuff? I am not impress by either, and would like to stay in Obj-C. > Will the dev environment that NeXT is going to deploy on NT be > worth the investment in time and money? Do other developers think that > it will compete with the likes of Borland, Microsoft, and Java in the > NT marketplace? Is NeXT focusing just on the web, as it appears? Hardly. The "Object Solutions on Tour" consists of half-a-day on WebObjects and half-a-day on other object technologies (mostly EOF and integration of legacy apps with EOF). WebObjects simply leverages off the existing object technologies and represents a new market that NeXT can enter nearly for free. Conversely, WebObjects will present new opportunities for NeXT to pitch their MCCA advantage to new audiences with old applications to nurse along. Since the Web is generating a lot of noise at the moment, NeXT has to generate a lot of noise just to be heard. The noise level is probably very disproportionate to the actual level of resources behind the scenes. I just attended "Object Solutions on Tour" and these are my *impressions* only...I speak with no authority. I would encourage you to hang in there, Sean. Good luck. Chuck ____________ Charles F. Waltrip The Johns Hopkins University Applied Physics Laboratory Opinions expressed are my own. email: waltrip@zephyr.jhuapl.edu phone: (410) 792-6596 fax: (410) 792-5597
From: Mark Anenberg <marka@Eng.Sun.COM> Newsgroups: comp.sys.next.programmer Subject: Re: Where GNU Source Date: Fri, 31 May 1996 13:47:40 -0700 Organization: Sun Microsystems Inc. Message-ID: <31AF5AEC.4485@Eng.Sun.COM> References: <4ocg5e$ip3@hermes.is.co.za> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit gvandyk@icon.co.za wrote: > > Where can I find the GNU sources for gcc and gdb that NeXT are using? > > I looked on ftp.next.com but only found EOF_1.1_GNUSOURCE and > PDO_3.0_GNUSOURCE which is not what I want. > > I know that the source is on the CD that the NeXT Developer software > is shipped on, but this release hasn't got the fixes in that came with > the Patches for 3.3. > > These sources should be publicly available as they are part of the GNU > project. > -- > > Regards, > Gerrit van Dyk > email: gvandyk@icon.co.za (NeXTMail welcome) > E.S. Systems cc > The OBJECT is the ADVANTAGE The PDO_3.0_GNUSOURCE source archive you mention contains all the gcc, gdb, gnumake, etc source that was used with PDO3.0 ---- Mark Anenberg , OpenStep Development Team, SunSoft, Inc. [E,NeXT,Mime]Mail: mark@velodrome.com, http://www.velodrome.com/ Disclaimer: The opinions expressed above are my own and in no way represent those of Sun Microsystems, Inc.
From: dev_dp@mfa.com (Dave Plonka) Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Subject: Re: How can I query where does a symbolic link points to? Followup-To: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Date: 31 May 1996 22:10:21 GMT Organization: McHugh Freeman Message-ID: <4onqod$l8c@homer.alpha.net> References: <4omku4$el4@news.sns-felb.debis.de> Torsten Caesar (caesar@dbag.ulm.DaimlerBenz.COM) wrote: > I want to query the path to which a symbolic link points to. The symbolic > link itself is a file which contains this information and "ls -l" manages it > to show this information but how can I get this info from my program? The readlink(2) system call is what you're looking for. Dave -- Dave Plonka ARS:N9HZF mailto:Dave.Plonka@mfa.com Waukesha, Wisconsin
Newsgroups: comp.sys.next.programmer From: Sean Willson <sean_willson@il.us.swissbank.com> Subject: Re: Reality Check....NeXTSTEP and OpenStep Message-ID: <1996May31.210411.8945@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <31AF3916.794B@zephyr.jhuapl.edu> Date: Fri, 31 May 1996 21:04:11 GMT In article <31AF3916.794B@zephyr.jhuapl.edu> Charles F Waltrip <waltrip@zephyr.jhuapl.edu> writes: > Sean Willson wrote: > > > > Fellow NeXT Users, > > > > [ MUNCH ]...... > > > I wouldn't make too much of this for now. Tsunami is an OpenStep app > and, whatever NeXT's plans or decisions turn out to be, it would be > premature to announce it's availability for anything other than NT at > this point. It's good to express your concern, however, which I second. It is an application only being released on NT, with no apparent talk of Solaris let alone MACH release.....it is as if we have fallen off the face of the earth......Why is is premature to announce on any other platform than NT? You would think it would instill some faith in the NeXTSTEP community to release it for MACH on HP/Sun/Intel/Moto....... > > [ MUNCH ]...... > > As it stands now, OpenStep on Mach/Solaris/Digital UNIX(??)/HPUX(??) > represents a very stable development environment for NT. Many would > argue that it's the ONLY stable development environment for NT. > Unless/until NT itself becomes more stable, OpenStep on other OSs will > likely continue to have a growing following. I totally agree with this point, but again NeXT is saying NT is their development platform of choice, so unless I am reading this wrong, they are saying NT is where they are going for development as well as products. It could mean that they are developing on NeXTSTEP and releasing to NT, but this is not how I am seeing it, anyone else? > > [ MUNCH ]...... > > Hardly. The "Object Solutions on Tour" consists of half-a-day on > WebObjects and half-a-day on other object technologies (mostly EOF and > integration of legacy apps with EOF). WebObjects simply leverages off > the existing object technologies and represents a new market that NeXT > can enter nearly for free. Conversely, WebObjects will present new > opportunities for NeXT to pitch their MCCA advantage to new audiences > with old applications to nurse along. From what I heard, the solutions tour was a bunch of advertisments to business folk, not about development. I understand that WebObjects leverages off of OpenStep technologies, but with the apparent future end of NeXTSTEP due to pulling programmer off projects, the demise of OpenStep on Solaris (Face it Sun has other things in mind with Java and the Java-OS), and nothing but Web talk from NeXT, how can anyone think NeXT has not changed focus completely yet again. DON'T GET ME WRONG, I hope that I am sitting here in 4-8 months saying, Guys and Gals, I got it all wrong....being young and about to enter the market officially just got me scared. NeXT proved OpenStep was really where it was at and Web talk was just leveraging off OpenStep technologies.....but right now, the signs are pointing otherwise.... > Since the Web is generating a lot of noise at the moment, NeXT has to > generate a lot of noise just to be heard. The noise level is probably > very disproportionate to the actual level of resources behind the > scenes. I'll give you this and pray...... > I would encourage you to hang in there, Sean. Good luck. Thanks.....I just want to make the right decision in a company out of the gate. I would like to start in OpenStep and go from there, but it seems that few companies are going that direction. They are using this bend in the rode to jump ship to Delphi and the likes of Java. Can anyone suggest a company or group of them doing otherwise? One that is hiring? I would greatly appreciate it since this is what I really love, and would like to do. Thank for the help. Sean Willson -- ____________________________________________________________________ Sean M. Willson sean_willson@il.us.swissbank.com Swiss Bank Corp. NeXT, MIME, and ascii 141 West Jackson Boulevard Chicago, Illinois 60604 (312) 554-5399
Newsgroups: comp.sys.next.programmer From: paul@cyantic.com (Paul Guiness) Subject: RTF to PS Filter Organization: CYANTIC Systems Date: Thu, 30 May 1996 19:01:23 GMT Message-ID: <1996May30.190123.23765@cyantic.com> Keywords: RTF, PS, Text Hi, I am looking for ways to convert an RTF file to PostScript without using an application built with IB. My options seem to be: 1. Create a tool that programatically takes care of allocating a Text object within a View and sending a printPSCode message to it. 2. Finding a filter that converts RTF (NeXT RTF) to PS without using the ApplicationKit at all. Option 1 means trying to make use of the ApplicationKit but avoiding ever displaying a view (and I don't mean just having a hidden window). I remember seeing a posting recently (I think) that explained how to startup the WindowServer manually. I'm wondering though if it is really possible to separate Application objects from the display. Some test code I've written crashes when I try allocating a Window object. Option 2 would be very helpful knowing that I can print RTF files without depending on the ApplicationKit at all. This would be used for running on Solaris where I have some NeXT libraries, but not (as far as I know) the appkit lib. I am looking for sample code or products that can do either option 1 or 2. It would be surprising to find that the only way to print NeXT RTF files is through the Text object. Thanks, Paul Guinness paul@cyantic.com
From: ts110@pmms.cam.ac.uk (Tomaz Slivnik) Newsgroups: comp.sys.next.programmer Subject: Checking if slip interface is up Date: 1 Jun 1996 11:46:37 GMT Organization: Cambridge University, Pure Mathematics and Mathematical Statistics Message-ID: <4opait$eqc@lyra.csx.cam.ac.uk> How do I check from a C program if a SLIP interface is up? Thank you for your replies. Tomaz Slivnik
From: james@jraynard.demon.co.uk (James Raynard) Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Subject: Re: How can I query where does a symbolic link points to? Date: 31 May 1996 23:37:13 -0000 Organization: A FreeBSD Box Message-ID: <4onvr9$393@jraynard.demon.co.uk> References: <4omku4$el4@news.sns-felb.debis.de> In article <4omku4$el4@news.sns-felb.debis.de>, Torsten Caesar <caesar@dbag.ulm.DaimlerBenz.COM> wrote: >I want to query the path to which a symbolic link points to. The symbolic >link itself is a file which contains this information and "ls -l" manages it >to show this information but how can I get this info from my program? The >problem is that fopen() (and I guess open() too) resolves the soft link and >opens the file where the symbolic link points to. > >I am working with Nextstep3.3 which is BSD4.3 API compatible. readlink(2) -- James Raynard, Edinburgh, Scotland jraynard@freebsd.org jraynard@dial.pipex.com james@jraynard.demon.co.uk
From: Roland Telfeyan <roland@telf.com> Newsgroups: comp.sys.next.programmer Subject: Re: Bad gdb Date: 1 Jun 1996 00:53:38 GMT Organization: Telf Design Corp. Distribution: world Message-ID: <4oo4ai$eh0@ralph.vnet.net> References: <4oktbm$i3n@ns1.moran.com> In article <4oktbm$i3n@ns1.moran.com> dboyce@albert (Douglas W Boyce) writes: > I have a project with 3 bundle sub-projects. Gdb suddenly refuses to > recognize any sources in the sub-projects in order to set breakpoints. > The standard error is: > > No source file named FileName.m. > Error occured processing command from Edit. > > Anyone ever have this problem with a valid PB.gdbinit file. > > > NS3.3 Intel, with user and developer patches applied. If your PB.gdbinit or .gdbinit file doesn't already have it, add the necessary 'dir' commands to include the subprojects' directories. If you are invoking gdb from the command line yourself, don't forget to say gdb yourProgram -x PB.gdbinit. If you're invoking gdb from the command line a lot, you could soft link PB.gdbinit -> .gdbinit and then forget about the -x deal. Gdb always looks for ~/.gdbinit and ./.gdbinit automatically. Roland -- Roland Telfeyan roland@telf.com
From: Roland Telfeyan <roland@telf.com> Newsgroups: comp.sys.next.programmer Subject: Re: Emacs and GDB Date: 1 Jun 1996 01:02:48 GMT Organization: Telf Design Corp. Distribution: world Message-ID: <4oo4ro$elv@ralph.vnet.net> References: <4oerg9$g0u@bdmserver.mcl.bdm.com> The shell environment that emacs provides by default is a scrolling text, not a screen-oriented vt100-type visual text terminal. That's the problem I think. When emacs forks a shell for you, your .login script is trying to set the terminal type and is complaining that it doesn't know about emacs's default terminal type. I'd like to know how to solve it as well. When trying to debug full-screen vt100 apps inside emacs, I get the vt100 codes flying down the screen. Is there a way to configure emacs to spoof a vt100? Roland -- Roland Telfeyan roland@telf.com
From: Roland Telfeyan <roland@telf.com> Newsgroups: comp.sys.next.programmer Subject: Re: Emacs and GDB Date: 1 Jun 1996 01:02:56 GMT Organization: Telf Design Corp. Distribution: world Message-ID: <4oo4s0$em0@ralph.vnet.net> References: <4oerg9$g0u@bdmserver.mcl.bdm.com> The shell environment that emacs provides by default is a scrolling text, not a screen-oriented vt100-type visual text terminal. That's the problem I think. When emacs forks a shell for you, your .login script is trying to set the terminal type and is complaining that it doesn't know about emacs's default terminal type. I'd like to know how to solve it as well. When trying to debug full-screen vt100 apps inside emacs, I get the vt100 codes flying down the screen. Is there a way to configure emacs to spoof a vt100? Roland -- Roland Telfeyan roland@telf.com
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 1 Jun 1996 17:17:40 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4optvk$19t@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> Mark Anenberg <marka@Eng.Sun.COM> wrote: > It still reads NeXTMail messages just fine. Well, at least something... > But you can send only PlainText or Mime. If you forward a NeXTMail > message it can become Mime and look just fine within Mail.app. > The way NeXT has done Mime you still get all your fonts, images > and attachments, just like in NeXTMail format; Well, you get different font sizes and attributes (bold, italic) You DON'T get different fonts and text colour. To me, these are VERY severe limitations. If you implement reception of NeXTMail, why stop halfways? It can't be that much effort to implement sending of NeXTMail, too. Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: crath@nortel.ca (Christopher Rath) Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Subject: Re: How can I query where does a symbolic link points to? Date: 31 May 1996 13:06:24 -0400 Organization: Northern Telecom Ltd., Ottawa, Canada Message-ID: <iwkivdcydpr.fsf@bmerhe83.nortel.ca> References: <4omku4$el4@news.sns-felb.debis.de> In-reply-to: caesar@dbag.ulm.DaimlerBenz.COM's message of 31 May 1996 11:24:52 GMT Here's a bit of code I hacked up to solve a softlink resolution problem I was having. I'm not sure what you're trying to do; but, this might help you. /** * ResolveSymlinks() - Resolve symlink, if it exists. * * Examine the basename of filename. If it is a symlink, then rescurse on it * until the real file is found. Stop recursing when rLevel is no longer * greater than zero. * * Parameters: * [1] filename - The filename to resolve. * [2] rLevel - Recursion level. * * Returns: * Nothing. **/ static void ResolveSymlinks(char *filename, const int rLevel) { struct stat fileBuf; /* Used in stat(). */ /* * If this if() is true then we have found a symlink. We will simply call * ourselves to handle nested symbolic links. */ if ((0 < rLevel) && (! lstat(filename, &fileBuf)) && (S_ISLNK(fileBuf.st_mode))) { char buf[BUFSIZ + 1]; int readcnt = readlink(filename, buf, BUFSIZ); if ((-1 == readcnt) || (BUFSIZ == readcnt)) { error(RLINKERR); } else { /* * Readlink() doesn't put a c-string in buf, instead is puts a * bunch of characters there, and it is our responsibility to * ensure we don't copy too many out. */ buf[readcnt] = '\0'; if ('/' == buf[0]) { /* * If the link was to an absolute path, then fully replace * filename with the retrieved link. */ strcpy(filename, buf); } else { int i; /* * If the link was to a relative file, then replace the * basename with the retrieved link. We haven't used strncat() * because we ensured earlier in this block that buf[] contained * a properly terminated c-string. */ for (i = BUFSIZ ; (i >= 0) && ('/' != filename[i]) ; i--) { filename[i] = '\0'; } if (BUFSIZ < strlen(filename) + strlen(buf)) { error(NMLENERR); } else { strcat(filename, buf); /* * Call ourself again just in case the newly formed basename is * also a softlink. */ ResolveSymlinks(filename, (rLevel - 1)); } } } } } -- === Christopher Rath ===== crath@bnr.ca ===== (613) 765-3141 === Northern Telecom Ltd. | Box 3511, Station `C' | ``Hydrogen is a colourless, odourless Ottawa, ON K1Y 4H7 | gas which, given enough time, turns FAX: (613) 763-4101 | into people.'' --- Henry Hiebert
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Re: RTF to PS Filter Date: 2 Jun 1996 04:01:46 GMT Organization: Digital Fix Development Message-ID: <4or3na$k9n@digifix.digifix.com> References: <1996May30.190123.23765@cyantic.com> In-Reply-To: <1996May30.190123.23765@cyantic.com> On 05/30/96, Paul Guiness wrote: >Hi, > >I am looking for ways to convert an RTF file to PostScript without using an application built with IB. My options seem to be: > 1. Create a tool that programatically takes care of allocating a Text object within a View and sending a printPSCode message to it. > 2. Finding a filter that converts RTF (NeXT RTF) to PS without using the ApplicationKit at all. Eric P. Scott wrote something that did this... and for the longest time it sat in my Unix folder after he posted it, and I would email it to those who asked... Unfortunately, it appears to have been lost... It was a single .m file... If someone has this code, please let me know, I'd like to have a copy of it again, and I think its what Paul is looking for too. > >It would be surprising to find that the only way to print NeXT RTF files is through the Text object. > >Thanks, > >Paul Guinness >paul@cyantic.com > -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
From: pjoisha@usa.pipeline.com(Prashanth G. Joisha) Newsgroups: comp.sys.next.programmer Subject: Problems with EOCOntroller's saveToObjects Date: 2 Jun 1996 15:11:07 GMT Organization: Pipeline Message-ID: <4osaub$oht@news2.h1.usa.pipeline.com> This question must have been answered number of times. However, I would be grateful if someone could point me to the right direction. I'm having problems with following logic (modifications to MainController in NeXT's PeopleDemo example) .. .. [deptController savesToDataSourceAutomatically]; [deptController savesToObjectsAutomatically]; if(/*Some logic*/) [dept setDepartmentName:@ "Marketing"]; else [dept setDepartmentName:@"Customer Relations"]; [deptController saveToObjects:self]; //Fails to save to database [deptController saveToDataSource:self]; //Though not //required, included to see if it helps The problem I think is for some reason the objects are not marked "dirty" and hence the EOController does not save them. I use the EOController's save methods, instead of dataSource's [[deptController dataSource] updateObject:dept]; [[deptController dataSource] saveObjects]; , because I can use it's delegate method controller:didUpdateObject:inDataSource: to refetch the object (the other way is to use DatabaseChannel's delegate method, I guess) So my question is: Is this a bug in EOController or am I missing something? Thanks a lot GPJ
From: chinay@cig.mot.com (Albert Chin-A-Young) Newsgroups: comp.sys.next.programmer Subject: Re: Emacs and GDB Date: 2 Jun 1996 17:08:48 GMT Organization: Motorola Cellular Infrastructure Group Distribution: world Message-ID: <4oshr0$noh@trotsky.cig.mot.com> References: <4oerg9$g0u@bdmserver.mcl.bdm.com> <4oo4s0$em0@ralph.vnet.net> Roland Telfeyan (roland@telf.com) wrote: : The shell environment that emacs provides by default is a scrolling text, : not a screen-oriented vt100-type visual text terminal. That's the problem I : think. When emacs forks a shell for you, your .login script is trying to : set the terminal type and is complaining that it doesn't know about emacs's : default terminal type. : : I'd like to know how to solve it as well. When trying to debug full-screen : vt100 apps inside emacs, I get the vt100 codes flying down the screen. Is : there a way to configure emacs to spoof a vt100? I think there's a discussion going on in comp.emacs about this. Supposedly there's a terminal.el that might get you started. : Roland -- albert chin
From: erbalch@adnc.com (Eric R. Balch) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Accessing the Parallel Port (QuickCam Software Development) Date: 3 Jun 1996 02:30:53 GMT Organization: adnc.com Message-ID: <4otiot$gev@taurus.adnc.com> Hi there... I have a few questions for anyone who wants to take a shot at em: First of all I should mention that I am in the early stages of trying to develop some software to use the Connectix(C) QuickCam(TM) under NSFIP... Basicly the idea is to port the portable QuickCam(TM) interface that already runs on several other platforms and give is a slick NS interface. By the way, if anyone knows the I am duplicating something that has already been done, please let me know where I can pick up a copy ;) The cam hooks up through the parallel port so what I am looking for is some detailed information on sending data to and receiving data from the port in the NS environment. NeXT states in their documentation that the parallel port driver under NSFIP 3.2 does not support "bidirectional" operation. I am not sure exactly what they mean here, maybe someone could elaborate. Also, I don't know if this is still the case under 3.3. Any information on or Help in this endeavor would be greatly appreciated! I plan to release the resulting driver, application(s), tech info, source code, etc. as freeware. Thankx, Eric R. Balch
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 3 Jun 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4otose$9p@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: brian@devnull.saturn.net (Brian Mitchell) Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Subject: Re: How can I query where does a symbolic link points to? Date: 3 Jun 1996 08:57:21 GMT Organization: Saturn Internet - ISP Message-ID: <slrn4r5a7l.n7.brian@tcpip.geek.net> References: <4omku4$el4@news.sns-felb.debis.de> On 31 May 1996 11:24:52 GMT, Torsten Caesar <caesar@dbag.ulm.DaimlerBenz.COM> wrote: >I want to query the path to which a symbolic link points to. The symbolic >link itself is a file which contains this information and "ls -l" manages it >to show this information but how can I get this info from my program? The >problem is that fopen() (and I guess open() too) resolves the soft link and >opens the file where the symbolic link points to. > >I am working with Nextstep3.3 which is BSD4.3 API compatible. "man lstat" is what you are looking for, I believe. -- Brian Mitchell brian@saturn.net "I never give them hell. I just tell the truth and they think it's hell" - H. Truman
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: Object does not respond to mouseDown!!! Date: 3 Jun 1996 09:36:04 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4oubm4$kn4@ai.alienor.fr> I've got a Text Object in which I add a matrix during the initFrame. .... [aMatrix setEnabled:NO]; [aMatrix setEmptySelectionEnabled:YES]; [Textobj addSubview:aMatrix]; [Textobj setNextResponder:nil]; During the drawSelf: I put the cells I my matrix and say they are not enabled .... [cellField setEnabled:NO]; [cellField setSelectable:NO]; [cellField setEditable:NO]; ... What I want is that my Text Object responds to the mouseDown: I wrote for this subClass. -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 3 Jun 1996 11:58:27 GMT Organization: Norden 1 Communications Message-ID: <4ouk13$g60@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------Commercial experience Objective C------------------Commercial experience EOF---------------------------A Plus Sybase or Oracle---------------A plus Career Position-----------------Full benefits & exceptional opportunity Relocation-----------------------Company assistance Must Be---------------------------US Citizen or Greencard To Be Considered-----Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: Kevin Walsh <kevin@cursor.demon.co.uk> Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Subject: Re: How can I query where does a symbolic link points to? Date: Mon, 3 Jun 1996 14:50:31 GMT Message-ID: <DsFIK7.GBz@cursor.demon.co.uk> References: <4omku4$el4@news.sns-felb.debis.de> <slrn4r5a7l.n7.brian@tcpip.geek.net> Organisation: Cursor Software Limited, Littlehampton, West Sussex, England In article <slrn4r5a7l.n7.brian@tcpip.geek.net> brian@saturn.net writes: > On 31 May 1996 11:24:52 GMT, Torsten Caesar <caesar@dbag.ulm.DaimlerBenz.COM> wrote: > > I want to query the path to which a symbolic link points to. The symbolic > > link itself is a file which contains this information and "ls -l" manages it > > to show this information but how can I get this info from my program? The > > problem is that fopen() (and I guess open() too) resolves the soft link and > > opens the file where the symbolic link points to. > > > > I am working with Nextstep3.3 which is BSD4.3 API compatible. > > > "man lstat" is what you are looking for, I believe. > It depends on what information you want about the symlink. If you are after the stat(2) information then lstat(2) will work ok. If you only really want to know the pathname of the file/directory pointed to by the symlink then try readlink(2). -- _/ _/ _/_/_/_/ _/ _/ _/_/_/ _/ _/ _/_/_/ _/_/ _/ _/ _/ _/_/ _/ Professor Kevin Walsh _/ _/ _/ _/ _/ _/ _/ _/_/ kevin@cursor.demon.co.uk _/ _/ _/_/_/_/ _/ _/_/_/ _/ _/
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <@vm.cnuce.cnr.it,@ISS.IT:ZANITTI@ISS.IT> Message-ID: <199606031225.FAA21116@isp.net> Date: Mon, 03 Jun 96 14:12:35 ISS From: ZANITTI@ISS.IT Subject: tty In my application I must set the tty in this mode: Baud Rate = 2400 Data Bits = 8 Stop Bits = 1 Parity = None This code is correct? - myMetod:(const char *)tty ( struct sgttyb modes = (B2400,B2400,0,0,RAW); file *myVar; if ((myVar = fopen(tty,"r+")) == NULL) ( syslog(FACILITY | LOG_CRIT,"message"); /* this (|) is or (keyboard limits) */ exit(1); ) setbuf(myVar,NULL); ioctl(fileno(myVar),TIOSECTP,&modes); fcntl(fileno(myVar),F_SETFL,FNDELAY); ...... ...... Thanks Zanitti Leo
From: bgomes@neptune.univ-lr.fr (Bruno Gomes) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software Subject: Trouble with gettytab configuration Date: 3 Jun 1996 16:03:40 GMT Organization: Universite de La Rochelle Message-ID: <4ov2cs$4t9@hpuniv.univ-lr.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hello, I try to connect a Next station and a PC, using the Next system, on dialin. But when I‚m connecting to the remote device, the stream format isn‚t good. I think, it‚s a gettytab file configuration problem. I had the same problem with an hp station using Next system. Trouble was resolved by gettytab file modification. Anybody knows gettytab configuration for Next station and PC using NextStep ? Thanks for help. -- Bruno Gomes E-mail : bgomes@cri.univ-lr.fr Universite de La Rochelle Tel : 46-45-82-14 Centre de Ressources Informatiques Fax : 46-45-82-45
From: Mark Anenberg <marka@Eng.Sun.COM> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Mon, 03 Jun 1996 10:15:07 -0700 Organization: Sun Microsystems Inc. Message-ID: <31B31D9B.2BAE@Eng.Sun.COM> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Uli Zappe wrote: > > Mark Anenberg <marka@Eng.Sun.COM> wrote: > > It still reads NeXTMail messages just fine. > > Well, at least something... > > > But you can send only PlainText or Mime. If you forward a NeXTMail > > message it can become Mime and look just fine within Mail.app. > > The way NeXT has done Mime you still get all your fonts, images > > and attachments, just like in NeXTMail format; > > Well, you get different font sizes and attributes (bold, italic) > > You DON'T get different fonts and text colour. > > To me, these are VERY severe limitations. > > If you implement reception of NeXTMail, why stop halfways? It can't be that > much effort to implement sending of NeXTMail, too. > > Uli Zappe well I don't speak officially for Sun, but... Its not a matter of stopping short; its a matter of a conscious limiting effort; It doesn't make sense to send email in a proprietary format when the rest of the world speaks Mime/X.400/X.500. The whole point is information exchange. -- Mark Anenberg , OpenStep Development Team, SunSoft, Inc. [E,NeXT,Mime]Mail: mark@velodrome.com Disclaimer: The opinions expressed above are my own and in no way represent those of Sun Microsystems, Inc.
From: yannick@uranus.univ-lr.fr (Yannick Buisson) Newsgroups: comp.sys.next.programmer Subject: NXTableView and editFieldAt method Date: 3 Jun 1996 19:37:59 GMT Organization: Universite de La Rochelle Message-ID: <4oveun$5nd@hpuniv.univ-lr.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi all, i have some problems with a NXTableview and the way to edit some cells in it !! To edit a cell in a NXTableview, i'm using the editFieldAt method, but if i call this method for a cell and for another, and so on, i have some displaying pbs !!!! some cells are white and we can't read the text written in it ??? can someone help me ?? Shall i use the endEditing method, or disabling the display from the window or something like that !!! thanks for your help -- //// (. .) ----oOO--(_)--OOo-------------------------------------------- Yannick BUISSON Centre de Ressources Informatiques Université de La Rochelle tel prof. : 46 45 82 14. fax prof. : 46 45 82 45. ARPAE . Appt 128. av. jean Monnet. Les minimes 17042 La Rochelle - cedex 1 tel : 46 45 95 87. Email (NeXTMail , MIME) : -> yannick@cri.univ-lr.fr
Newsgroups: comp.sys.next.programmer From: jpanico@netcom.com (Joe Panico) Subject: Re: Problems with EOCOntroller's saveToObjects Message-ID: <jpanicoDsF8CF.GKv@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4osaub$oht@news2.h1.usa.pipeline.com> Date: Mon, 3 Jun 1996 11:09:50 GMT Sender: jpanico@netcom19.netcom.com Prashanth G. Joisha (pjoisha@usa.pipeline.com) wrote: : in NeXT's PeopleDemo example) : .. : .. : [deptController savesToDataSourceAutomatically]; : [deptController savesToObjectsAutomatically]; : if(/*Some logic*/) : [dept setDepartmentName:@ "Marketing"]; : else : [dept setDepartmentName:@"Customer Relations"]; : : [deptController saveToObjects:self]; //Fails to save to database : [deptController saveToDataSource:self]; //Though not //required, : included to see if it helps : : The problem I think is for some reason the objects are : not marked "dirty" and hence the EOController does not : save them. Yes, you are absolutely correct here-- the EOController does not see the change, and therefore does not save anything. EOControllers are designed to coordinate information flow to and from a bunch of EOAssociations. They assume that changes are coming from the UI. If you directly modify an eo through its accessor methods, the EOController has no way of getting notified about the change. If you want to programmatically modify your eos you can use: - (void)setValues:(NSDictionary *)newValues forObject:anEO which simulates the changes coming through EOAssociations. If there is no UI involved in your task, then you should probably skip the COController and just use the EODataSource, since the EODataSource can detect changes made directly to eos. Good luck. -- jpanico@netcom.com Joe Panico NeXTStep/OpenStep Developer BLaCKSMITH Inc.
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: NXTableView and editFieldAt method Date: 3 Jun 1996 20:59:55 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4ovjob$8j5@news.onramp.net> References: <4oveun$5nd@hpuniv.univ-lr.fr> (Yannick Buisson) writes: > > i have some problems with a NXTableview and the way to edit some cells in it !! > To edit a cell in a NXTableview, i'm using the editFieldAt method, but if i call this method > for a cell and for another, and so on, i have some displaying pbs !!!! some cells are white > and we can't read the text written in it ??? > > can someone help me ?? > Shall i use the endEditing method, or disabling the display from the window or something like > that !!! I've had this problem too. When the user is editing a cell and hits return/tab, I'd like the selection to go to the cell to the right, instead on the cell below, but editFieldAt has problems... Steve
From: jfranco@montgomery.plato.sky.bdm.com (Jeff Franco,,,) Newsgroups: comp.sys.next.programmer Subject: Re: Emacs and GDB Date: 3 Jun 1996 21:34:41 GMT Organization: BDM International, Inc. Message-ID: <4ovlph$reg@bdmserver.mcl.bdm.com> References: <4oerg9$g0u@bdmserver.mcl.bdm.com> <4oo4ro$elv@ralph.vnet.net> In-Reply-To: <4oo4ro$elv@ralph.vnet.net> On 05/31/96, Roland Telfeyan wrote: >The shell environment that emacs provides by default is a scrolling text, >not a screen-oriented vt100-type visual text terminal. That's the problem I >think. When emacs forks a shell for you, your .login script is trying to >set the terminal type and is complaining that it doesn't know about emacs's >default terminal type. > >I'd like to know how to solve it as well. When trying to debug full-screen >vt100 apps inside emacs, I get the vt100 codes flying down the screen. Is >there a way to configure emacs to spoof a vt100? > >Roland > >-- I have no idea, but this unknown term thing would be great to get rid of. I'm chalking it up to .profile/.login/heterogeneous unix environment magic until I get more information. Please let me know if you exorcise this beast. -- Thanks, Jeff Franco (703) 824-4301 Price Waterhouse LLP jfranco@plato.sky.bdm.com (NeXTMail, MIME)
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: Accessing the Parallel Port (QuickCam Software Development) Date: 3 Jun 1996 23:13:32 GMT Organization: NO ORGANIZATION, INC. Message-ID: <4ovris$kun@bias.ipc.uni-tuebingen.de> References: <4otiot$gev@taurus.adnc.com> Cc: erbalch@adnc.com In <4otiot$gev@taurus.adnc.com> Eric R. Balch wrote: > Hi there... > > I have a few questions for anyone who wants to take a shot at em: > > First of all I should mention that I am in the early stages of trying to > develop some software to use the Connectix(C) QuickCam(TM) under NSFIP... > Basicly the idea is to port the portable QuickCam(TM) interface that already > runs on several other platforms and give is a slick NS interface. > > By the way, if anyone knows the I am duplicating something that has already > been done, please let me know where I can pick up a copy ;) > > The cam hooks up through the parallel port so what I am looking for is some > detailed information on sending data to and receiving data from the port in > the NS environment. > > NeXT states in their documentation that the parallel port driver under NSFIP > 3.2 does not support "bidirectional" operation. I am not sure exactly what > they mean here, maybe someone could elaborate. Also, I don't know if this is > still the case under 3.3. > > Any information on or Help in this endeavor would be greatly appreciated! I > plan to release the resulting driver, application(s), tech info, source code, > etc. as freeware. > As far as I know both the NeXT Driver and the PD ParallelPortDriver 1.5 as available from Peanuts support only printing (are not bidirectional). But you can ask Christian Starkjohann <cs@hal.kph.tuwien.ac.at> or <cs@ds1.kph.tuwien.ac.at> who wrote the PD driver for help. Bye Frank -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Copy/Paste Date: 4 Jun 1996 00:04:14 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4ovuhu$bs5@news.onramp.net> Is there a simple PD example of copy/paste on something besides text. Draw.app is a mess. Thanks for any info, Steve
From: Christian Gruber <cgruber@virtualcity.com> Newsgroups: comp.sys.next.programmer Subject: EOF inserts and pre-fab key-index fields. Date: Mon, 03 Jun 1996 23:26:04 -0600 Organization: Quintessence Software Foundry Message-ID: <31B3C8EC.7C50@virtualcity.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I talked to a friend at NeXT regarding this, and he agreed that this was probably the way to go, but I just wanted to check with you'se guys to see if I can't find a more elegant solution... I'm running a sybase SQL server, and I have MY_TABLE, with MY_ID set to auto-increment (IDENTITY property for you Sybasers). I go to insert a record from an EOController object, and of course it fails, because Sybase fails when you try to assign a value to an IDENTITY field. I tried setting the ID attribute to read only (in my model) and that gave me worse headaches. I don't want to implement a destributed ID-number broker for my app, so I thought I'd delegate my EOAdaptorChannel, create an object with a dummy key-field, trap the - (EODelegateResponse)adaptor:(blah..)blah willInsertRow:(NSMutableDictionary *) aRow forEntity:(blah..)blah (or whatever), strip out the MY_ID key=object pair, and send it on it's merry way with the modified row. This should (right?) generate SQL less the reference to the MY_ID field, and that is what sybase expects for IDENTITY fields (I checked this) Any comments? Any improvements. I'm having some hell of a time getting the stupid thing to work, so I'm looking for both concept criticism, as well as warnings, and perhaps clues as to how EOF2.0 might handle such a thing. Christian Gruber Quintessence Software Foundry mailto:cgruber@q-soft.com http://www.gruber.org/ (403) 229-4883 (voice) (403) 244-9777 (fax) P.S. Rave in Saskatoon. http://www.gruber.org/Ascendance/rave.html
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer Subject: Text and textDidChange: Date: 4 Jun 1996 01:04:21 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4p022l$49q@solaris.cc.vt.edu> Hi, I have an app that uses a text editing window where the contents of the text are constantly changed as the user goes clicking through the browser in the same window. I'd like to capture the text whenever the user clicks a browser cell (or the main window is resigned) *and* the text was changed. If the text wasn't changed, no point in capturing it as this makes the app sluggish. It's also important to know when it has changed so that I can "setDocEdited:YES". I implemented -textDidChange: but the "text" in this message seems to refer to the ASCII characters and not the text object. If the user changes fonts, drops colors, etc. the message is not sent. I want something equivalent to "-theContentsOfTheTextObjectDidChange:" which would be sent whenever the characters *or* formatting changed. Has anyone faced this problem and worked around it? Perhaps there is a custom subclass of Text on the net that fixes this design flaw? Thanks, Chuck
From: michael@rumah.pc.my (Michael Olan) Newsgroups: comp.sys.next.programmer Subject: libg++ Date: 4 Jun 1996 10:26:01 GMT Organization: Unconfigured Message-ID: <4p12vp$gg2@jaring.my> I recently installed gcc-2.7.2 and libg++-2.7.1 on a NEXTSTEP 3.2 machine. When I compiled an old program which #includes builtin.h, I got the following ld error: ld: Undefined symbols: _start_timer _return_elapsed_time When I searched in libg++.a the symbols indicated were in there. I also manually told gcc where to look & which library to use, but no change. It worked with the old gcc. Any suggestions. Several other #includes I tried in g++-include worked ok. algobase.h which #includes builtin.h has the same problem. Email responses appreciated. -- --------------------------------------------------------------------- Dr. Michael Olan Email: michael@rumah.pc.my (NeXT Mail OK) Senior Lecturer - Computer Science michael@ppp.itm.my American Degree Program Fax: 6-03-5482329 Institut Teknologi MARA Section 17, Shah Alam, Malaysia ---------------------------------------------------------------------
Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next From: Bill Zissimopoulos <B.Zissimopoulos@cs.ucl.ac.uk> Subject: Re: How can I query where does a symbolic link points to? Sender: news@ucl.ac.uk (Usenet News System) Message-ID: <31B43AEB.5FDB@cs.ucl.ac.uk> Date: Tue, 4 Jun 1996 13:32:27 GMT Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii References: <4omku4$el4@news.sns-felb.debis.de> <iwkivdcydpr.fsf@bmerhe83.nortel.ca> Mime-Version: 1.0 Organization: University College London Christopher Rath wrote: > > Here's a bit of code I hacked up to solve a softlink resolution problem I was > having. I'm not sure what you're trying to do; but, this might help you. > > /** > * ResolveSymlinks() - Resolve symlink, if it exists. > * > * Examine the basename of filename. If it is a symlink, then rescurse on it > * until the real file is found. Stop recursing when rLevel is no longer > * greater than zero. > * > * Parameters: > * [1] filename - The filename to resolve. > * [2] rLevel - Recursion level. > * > * Returns: > * Nothing. > **/[code snipped] After seeing your code I really wondered if you have learnt programming in a functional language. You have a recursive call, where a for loop would do nicely. Not that I have any problem with that, it's just not very common (and probably not very efficient, since this was tail recursion and can be trivially eliminated). The problem with your code is that it resolves symlinks for the last path component only. Since I don't have the original poster's article I don't know if that's what he wanted. If you want the real path of a file you could just use realpath() (resolves all symlinks). Below I also enclose a function/program of my own which I call hardpath(). I think that it is better than realpath() for the following reasons: 1. realpath() may fail if the cwd is not readable 2. I prefer to keep the ".." entries, instead of going up one level in the path (it could be that a directory has more than one hard link... very rare I admit it) 3. I believe that my implementation is more elegant (of course I am wrong :-) The code for tha hardpath function/program is included at the end of the post. I think that it is bugfree. Then again there is always one more bug... Bill -- Bill Zissimopoulos <mailto:B.Zissimopoulos@cs.ucl.ac.uk> <http://www-dept.cs.ucl.ac.uk/students/B.Zissimopoulos/> /* $Id: hardpath.c,v 1.3 1996/05/24 02:05:49 bill Exp bill $ */ /* ** hardpath.c ** Translate path to hard link path. ** ** Written by Bill Zissimopoulos, 1996. */ /* ** $Log: hardpath.c,v $ ** Revision 1.3 1996/05/24 02:05:49 bill ** Preconditions done first in hardpath(). ** ** Revision 1.2 1996/05/24 00:25:28 bill ** Added stat() call at the beginning of hardpath(). ** ** Revision 1.1 1996/05/23 23:58:34 bill ** Initial revision ** */ static char rcsid[] = "$Id: hardpath.c,v 1.3 1996/05/24 02:05:49 bill Exp bill $"; #include <stdlib.h> #include <stdio.h> #include <string.h> #include <errno.h> #include <limits.h> #include <sys/param.h> #include <sys/types.h> #include <sys/stat.h> #ifndef PATH_MAX #ifndef MAXPATHLEN #define PATH_MAX 1024 #else #define PATH_MAX MAXPATHLEN #endif #endif #define MAX_READLINKS 64 char *hardpath(const char *path, char *buf, size_t bufsiz) { char gapbuf[PATH_MAX + 1 + 1], *gapsrc, *gapdst, *savdst; char lnkbuf[PATH_MAX]; int readlinks = 0; int siz; struct stat statbuf; /* preconditions */ if (bufsiz == 0) { if (buf != NULL) { errno = EINVAL; return NULL; } bufsiz = PATH_MAX + 1; } /* do we need to go into the trouble? */ if (stat(path, &statbuf) == -1) return NULL; /* init gap buffer [a gap (even 1 char only) should always exist!] */ if ((siz = strlen(path) + 1) > PATH_MAX + 1) { errno = ENAMETOOLONG; return NULL; } gapdst = gapbuf; gapsrc = gapbuf + sizeof gapbuf - siz; memcpy(gapsrc, path, siz); /* go over all path components */ while (*gapsrc) { savdst = gapdst; /* remember this */ if (*gapsrc == '.') /* .* entry */ { *gapdst++ = '.', gapsrc++; if (*gapsrc == '.') /* ..* entry */ *gapdst++ = '.', gapsrc++; if (*gapsrc == '/' || *gapsrc == '\0') goto slash; /* fall through */ } while (*gapsrc && *gapsrc != '/') *gapdst++ = *gapsrc++; *gapdst = '\0'; /* check if what we have here is a symlink */ if (gapdst > savdst) { lnkbuf[0] = '\0'; if ((siz = readlink(gapbuf, lnkbuf, sizeof lnkbuf)) >= 0) { if (++readlinks >= MAX_READLINKS) /* paranoid: alrdy stat'ed file */ { errno = ELOOP; return NULL; } gapdst = lnkbuf[0] == '/' ? gapbuf : savdst; if (gapsrc - gapdst <= siz) /* always keep at least 1 char gap */ { errno = ENAMETOOLONG; return NULL; } gapsrc -= siz; memcpy(gapsrc, lnkbuf, siz); } else if (errno != EINVAL) /* EINVAL: file not a symlink (but exists) */ return NULL; } slash: if (*gapsrc == '/') { *gapdst++ = '/'; while (*++gapsrc == '/'); /* eat extra slashes */ } } *gapdst = '\0'; /* copy hard path to user supplied buffer */ if ((siz = strlen(gapbuf) + 1) > bufsiz) { errno = ERANGE; return NULL; } if (buf == NULL) if ((buf = malloc(bufsiz)) == NULL) return NULL; memcpy(buf, gapbuf, siz); return buf; } #define PROGNAME "hardpath" void warn(const char *s) { fprintf(stderr, PROGNAME ": %s: %s\n", s, strerror(errno)); } int main(int argc, char *argv[]) { char buf[PATH_MAX + 1]; struct stat statbuf; dev_t dev; ino_t ino; while (*++argv) { if (hardpath(*argv, buf, sizeof buf)) { puts(buf); if (stat(*argv, &statbuf) == -1) warn(*argv); dev = statbuf.st_dev, ino = statbuf.st_ino; if (stat(buf, &statbuf) == -1) warn(buf); if (dev != statbuf.st_dev || ino != statbuf.st_ino) fprintf(stderr, PROGNAME ": internal error: hardpath() failed\n" " origpath=%s, (dev=%u ino=%u)\n" " hardpath=%s, (dev=%u ino=%u)\n", *argv, dev, ino, buf, statbuf.st_dev, statbuf.st_ino); } else warn(*argv); } return 0; }
From: neuss@isa.informatik.th-darmstadt.de.NOSPAM (Christian Neuss) Newsgroups: comp.sys.next.programmer Subject: Re: RTF to PS Filter Date: 4 Jun 1996 15:08:42 GMT Organization: Fachbereich Informatik, TH Darmstadt, Deutschland Message-ID: <4p1jhq$1q5@rs18.hrz.th-darmstadt.de> References: <1996May30.190123.23765@cyantic.com> <4or3na$k9n@digifix.digifix.com> Scott Anguish (sanguish@digifix.com) wrote: : On 05/30/96, Paul Guiness wrote: : >I am looking for ways to convert an RTF file to PostScript without : using an application built with IB. My options seem to be: : It was a single .m file... : If someone has this code, please let me know, I'd like to have : a copy of it again, and I think its what Paul is looking for too. I've found it. I've placed it in the Web. Please retrieve it from http://www.informatik.th-darmstadt.de/~neuss/next-printrtf.m If somebody doesn't have Web access, let me know. I'll mail it to you. Hope this helps. Chris -- // Christian Neuss "I ride tandem with a random.." // http://www.informatik.th-darmstadt.de/~neuss/ // +49 6151 16-3414 fax: -5472
From: Steve Spicklemire <steve@estel.uindy.edu> Newsgroups: comp.sys.next.programmer Subject: Problems with "cc -posix ..." Date: Tue, 04 Jun 1996 09:40:02 +0000 Organization: Silicon Prairie Ventures Message-ID: <31B4046E.4D28@estel.uindy.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi folks, Somehow it isn't quite working for me! This has happened to me twice now and I still haven't found a good resolution (other than some blind hacks I'm too embarrased to admit!) I'm trying to build some code that requires posix functions (e.g., the bsddb library, and freeWAIS-0.5,...). Anyway... if I put '-posix' on the cc command line the .c files build OK, but I get unresolved refs with -lposix (like floor, ceil, and log) or -lm (then it's getcwd), but if I use -lm *and* -lposix I get a boatload of multiply defined symbols. Has anyone figured out how to handle this? I looked in the man page for a compiler flag that told the linker to ignore multiple defs of functions (i.e., just take the first found def) but I didn't see it in there. BTW I'm on NS Developer 3.2. thanks! -steve
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: How to use the MiscSwapView and MiscTabMatrix? Date: 4 Jun 1996 16:23:12 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4p1ntg$jog@lion.embl-heidelberg.de> References: <4okpe3$hj3@nntp.onramp.net> In article <4okpe3$hj3@nntp.onramp.net> dekorte@suite.com (Steve Dekorte) writes: > > I am looking for a sample code for using MiscSwapView and > > MiscTabMatrix. I have tried to open the nib file in Unix for Hater > > which uses SwapView and TabMatrix but I couldn't? Help needed? > > > > --KAI-- > > You may need to load those palettes into Interface Builder first. > Double-clicking on the .palette files for them should do the trick. Well, I hope Kai knows it ;-) I sent him an example of a nib file. I'll put this example on my www page next weekend. http://www.nmr.embl-heidelberg.de/tuparev/ Look for my developer's cookbook -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: crath@nortel.ca (Christopher Rath) Newsgroups: comp.unix.programmer,comp.sys.next.programmer,de.comp.sys.next Subject: Re: How can I query where does a symbolic link points to? Date: 04 Jun 1996 12:54:32 -0400 Organization: Northern Telecom Ltd., Ottawa, Canada Message-ID: <iwkloi3ldbr.fsf@bmerhe83.nortel.ca> References: <4omku4$el4@news.sns-felb.debis.de> <iwkivdcydpr.fsf@bmerhe83.nortel.ca> <31B43AEB.5FDB@cs.ucl.ac.uk> In-reply-to: Bill Zissimopoulos's message of Tue, 4 Jun 1996 13:32:27 GMT In article <31B43AEB.5FDB@cs.ucl.ac.uk> Bill Zissimopoulos <B.Zissimopoulos@cs.ucl.ac.uk> writes: > ... > After seeing your code I really wondered if you have learnt > programming in a functional language. You have a recursive call, > where a for loop would do nicely. Not that I have any problem with > that, it's just not very common (and probably not very efficient, > since this was tail recursion and can be trivially eliminated). I am familiar with the conversion you refer to (tail recursion elimination); however, in some cases I prefer to leave the recursion in place. This code came from an application where performance was not a concern, and the recursion simply reflects how I visualized the algorithm I was coding. > The problem with your code is that it resolves symlinks for the last > path component only. Since I don't have the original poster's > article I don't know if that's what he wanted. If you want the real > path of a file you could just use realpath() (resolves all > symlinks). Below I also >... * realpath() doesn't exist on all of the systems my code has to run on (HP-UX 9.x, specifically) * my code exists only so that when I perform an fopen() on the filename, I am guarenteed that the basename I'm using is not a symbolic link. Symbolic links within the directory list preceding the basename do not affect this condition. * I'm not sure whether or not my code completely answered his question, but I wanted to give him an example usage of the readlink() function along with its name. You'll notice that other posters to this thread only gave the function name in their replies. Christopher -- === Christopher Rath ===== crath@bnr.ca ===== (613) 765-3141 === Northern Telecom Ltd. | Box 3511, Station `C' | ``Hydrogen is a colourless, odourless Ottawa, ON K1Y 4H7 | gas which, given enough time, turns FAX: (613) 763-4101 | into people.'' --- Henry Hiebert
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 4 Jun 1996 18:11:57 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4p1u9d$1ip@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> Mark Anenberg <marka@Eng.Sun.COM> wrote: > Uli Zappe wrote: > > If you implement reception of NeXTMail, why stop halfways? It can't be > > that much effort to implement sending of NeXTMail, too. > > well I don't speak officially for Sun, but... > > Its not a matter of stopping short; its a matter of a conscious > limiting effort; It doesn't make sense to send email in a proprietary > format when the rest of the world speaks Mime/X.400/X.500. The whole > point is information exchange. I'm well aware that basically it comes down to this old "the technologically excellent vs. the most widespread" thing, and we all know who apparently always has to suffer :-((( In this case, though, it's not only different because NeXTMail would be just another *option*, allowing better quality in communication with a selected group of people (just like colour faxes to those who own colour fax machines). More than that, SUN could argue that with NEO on SUN machines NeXTMail isn't something "proprietary" anymore ;-))) (And don't forget GNUStep!) I'd understand if a whole lot of effort were necessary for the implementation, but in this case - why not leave the choice to the user??? Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Re: RTF to PS Filter Date: 4 Jun 1996 20:59:40 GMT Organization: Digital Fix Development Message-ID: <4p283s$j5j@digifix.digifix.com> References: <1996May30.190123.23765@cyantic.com> <4or3na$k9n@digifix.digifix.com> <4p1jhq$1q5@rs18.hrz.th-darmstadt.de> In-Reply-To: <4p1jhq$1q5@rs18.hrz.th-darmstadt.de> On 06/04/96, Christian Neuss wrote: >Scott Anguish (sanguish@digifix.com) wrote: >: On 05/30/96, Paul Guiness wrote: >: >I am looking for ways to convert an RTF file to PostScript without >: using an application built with IB. My options seem to be: > >: It was a single .m file... > >: If someone has this code, please let me know, I'd like to have >: a copy of it again, and I think its what Paul is looking for too. > >I've found it. I've placed it in the Web. Please retrieve it >from http://www.informatik.th-darmstadt.de/~neuss/next-printrtf.m > Thanks BTW to everyone who emailed me a copy... it wasn't what Paul wanted in the end, but it was what I was looking for! Scott -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
From: louie@va.pubnix.com (Louis A. Mamakos) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 4 Jun 1996 23:18:12 -0400 Organization: Pubnix Access Systems (Virginia) Message-ID: <4p2u9k$oal@pub02.va.pubnix.com> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> >I'm well aware that basically it comes down to this old "the technologically >excellent vs. the most widespread" thing, and we all know who apparently >always has to suffer :-((( NeXTMail format isn't any technologially "better" than MIME is; it's just different and proprietary and not significantly better. NeXT missed a golden opportuntity to have a showcase MIME mail user agent by having their NextMail.app program support MIME rather than their own format. They didn't, and no one cared how whizzy NextMail was because it wasn't interoperable with the rest of the world. >In this case, though, it's not only different because NeXTMail would be just >another *option*, allowing better quality in communication with a selected >group of people (just like colour faxes to those who own colour fax >machines). Not better. Different. >More than that, SUN could argue that with NEO on SUN machines NeXTMail isn't >something "proprietary" anymore ;-))) (And don't forget GNUStep!) I'd love to run GNUStep. Is it hosted on FreeBSD? louie
From: baudouin@firmin.labri.u-bordeaux.fr (Thierry Baudouin [Paries]) Newsgroups: comp.sys.next.programmer Subject: RunTime Date: 4 Jun 1996 15:16:39 GMT Organization: Laboratoire Bordelais de Recherche en Informatique Message-ID: <4p1k0n$nq9@news.u-bordeaux.fr> I am working with runtime Functions and i have a problem : I want to call dynamically methods for objects using the method "performv" : -performv:SEL :void * First argument is the selector and second argument is for arguments of method. The return value of "performv" is an id ,object, and if the method called dynamically return an other type a problem appear. However, there is a good conversion with int,short,long,char and any pointer but a bad conversion with float and double.How can i get with this two types. Thanks to answer if you have a idea.
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Re: How to write and read an NSArray in a File ? Date: 5 Jun 1996 03:55:52 GMT Organization: No organisation supplied Message-ID: <4p30g8$5gq@hermes.is.co.za> References: <4ojofp$8nj@hpuniv.univ-lr.fr> In-Reply-To: <4ojofp$8nj@hpuniv.univ-lr.fr> On 05/30/96, Gilles Garrabella wrote: >I'm a novice in objective-C programmation with Interface Builder and i >don't now how to write and read an NSArray in a file. > >I try an NXOpenTypedStreamForFile() to open the file and a >NXWriteNSObject() to write the NSArray in but when i try to read the file >and print the NSArray i have this error : >NSInvalidArgumentException: *** -[NSMutableDictionary setObject:forKey:]: >nil object for key 'entity' > >Maybe, a NSCoding is necessary but how to make it ? > >Thanks to reply in NewsGrazer or by mail to ggarrabe@cri.univ-lr.fr > Well the way I did it was to archive the array to a NSData class and then I write out the NSData. I then read back the NSData and unarchive it using NSUnArchiver which produces the "original" object. I am not saying that this is the best solution but it seems to work OK. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: OmniWeb Hierarchical List Date: 4 Jun 1996 21:16:40 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4p293o$ra1@news.onramp.net> Is there a PD palette (or NeXT cell view) for a hierarchical list like in OmniWebs hot-links window or InterfaceBuilder's "outline" view of classes and instances? Thanks for any info, Steve
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: How can I query where does a symbolic Date: 4 Jun 1996 14:57:20 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4p2bg0$pc@nova.mdd.comm.mot.com> References: <31B43AEB.5FDB@cs.ucl.ac.uk> In article 5FDB@cs.ucl.ac.uk, Bill Zissimopoulos <B.Zissimopoulos@cs.ucl.ac.uk> writes: >You have a recursive call, where a for loop would do nicely. Not that I have >any problem with that, it's just not very common (and probably not very >efficient, since this was tail recursion and can be trivially eliminated). Its so trivial that I believe today's compilers (e.g. gcc) can optimize it out for you. Functional languages often teach you to write tail recursion, because they can also optimize out tail recursion. Cheers, Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: maag@now.ch Newsgroups: comp.sys.next.programmer Subject: White Next and x25? Date: Wed, 5 Jun 1996 11:55:02 GMT Organization: NOW GmbH, Baar, Switzerland Sender: news@now.ch Message-ID: <DsIzrq.ByK.0.astra@now.ch> hi there i consider writing an application using x25 on a white next with nextstep 3.3 (later 4.0 / openstep). what i need at the moment are informations about x25-cards and/or drivers for those cards: what cards are suppported by nextstep? are there existing drivers? are there pieces of code/documentation concerning those cards/drivers? any piece of information is highly appreciated! thanks a lot Rolf -- sounds.rtfremoteSpotMerge.0406
From: laurent@planon.qc.ca (Laurent Daudelin) Newsgroups: comp.sys.next.programmer Subject: Re: Text and textDidChange: Date: 5 Jun 1996 14:57:59 GMT Organization: Communications Accessibles Montreal, Quebec Canada Distribution: world Message-ID: <4p479n$b28@tandem.CAM.ORG> References: <4p022l$49q@solaris.cc.vt.edu> In article <4p022l$49q@solaris.cc.vt.edu> Chuck_Esterbrook@orcacomputer.com writes: > Hi, > > I have an app that uses a text editing window where the contents of the text > are constantly changed as the user goes clicking through the browser in the > same window. > > I'd like to capture the text whenever the user clicks a browser cell (or the > main window is resigned) *and* the text was changed. If the text wasn't > changed, no point in capturing it as this makes the app sluggish. It's also > important to know when it has changed so that I can "setDocEdited:YES". > > I implemented -textDidChange: but the "text" in this message seems to refer > to the ASCII characters and not the text object. If the user changes fonts, > drops colors, etc. the message is not sent. > > I want something equivalent to "-theContentsOfTheTextObjectDidChange:" which > would be sent whenever the characters *or* formatting changed. > > Has anyone faced this problem and worked around it? Perhaps there is a custom > subclass of Text on the net that fixes this design flaw? > > Thanks, > > Chuck Chuck, There is the less known Text delegate method 'text:textObject isEmpty:(BOOL)empty' that gets called whenever the content changes. As you recall, whenever a TextField is selected for edition, the window text field editor object is set with the attributes of the text field, then editing can begin. The text field that's get represented by the text object becomes the text object delegate. If an object is made the text field delegate, any delegate messages sent by the text object to the text field will be fowarded to the text field delegate. This is how I can know whenever the text field content changes. Maybe a bit difficult to understand, but give it a try, who knows... - Laurent -- ****************************************************************** Laurent Daudelin, Lead Software Engineer- Planon TELEXPERTISE Inc. laurent@planon.qc.ca <-- NeXTMail welcome! (MIME Mail welcome too!) Join "EvangeList," Guy Kawasaki's (un)official Apple listserver of good news about Apple, Macintosh, and third-party developers. To subscribe to
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 5 Jun 1996 20:32:13 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4p4qsd$74p@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> louie@va.pubnix.com (Louis A. Mamakos) wrote: > >I'm well aware that basically it comes down to this old "the technologically > >excellent vs. the most widespread" thing, and we all know who apparently > >always has to suffer :-((( > NeXTMail format isn't any technologially "better" than MIME is; it's just different and proprietary and not significantly better. NeXT missed a golden opportuntity to have a showcase MIME mail user agent by having their NextMail.app program support MIME rather than their own format. They didn't, and no one cared how whizzy NextMail was because it wasn't interoperable with the rest of the world. > Not better. Different. Hi Louis :) Well, I think it is both different and better, especially when you have the choice to down grade to MIME. Quite honestly, the RTFD used for NeXTmail is canvas allowing richer expression than the MIME pallatte. Noth that NeXTmail couldn't be improved... -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: "BEST" <markb@bestnet.com> Newsgroups: comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,pgh.next-users,sdnet.next,slac.users.next,su.computers.next Subject: IMMEDIATE "NeXTstep Developers" CONSULTANT OPPORTUNITIES!!! Date: Wed, 5 Jun 1996 16:08:18 -0700 Organization: BEST CONSULTING Message-ID: <01bb5333.e2460de0$3b3240ce@kwaa02w059.bestnet.com> BEST Consulting is an Information Systems Consulting firm headquartered in the Seattle area with 11 offices throughout the western U.S. The majority of our Associate Consultants are salaried however, we do have hourly contract Associates as well. All BEST Associates are proficient in a variety of technical disciplines and are strong in professional demeanor, attitude and other non-technical attributes. BEST Consulting has built a solid reputation by providing QUALITY SERVICE to our clients and team building and career oriented employment to our Associates. Our clients keep coming back and referring others and our Associate employee base increases because we are committed to personal pride and the delivery of which we never compromise. Our idea is "total quality", quality in our offering itself and in all the services that come with it. BEST has competitive pay rates, excellent cafeteria style benefits, and an employee oriented history. BEST is constantly looking for people that want to build a relationship and career with a great company. BEST's focus is to "make a difference" JOB TITLE: NeXTstep Developers (#470) JOB LOCATION: Seattle Area NR. OF OPENINGS: Our client has immediate requirements for 10 developers PART/FULL TIME: Full Time START DATE: Immediately SALARY: DOE/DOQ JOB QUALIFICATIONS: NeXTstep, Objective C JOB DESCRIPTION: To provide development porting our clients application from NeXTstep to open step on NT. If you are interested in these opportunities please send/fax your resume to: Mark Barton Sourcing Manager BEST Consulting 12910 Totem Lake Blvd., #270 Kirkland, WA 98034 (206) 814-8104 ext. 185 (206) 814-8108 Fax markb@bestnet.com http://www.bestnet.com
From: cph@next.com (Charles d'Harcourt) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 6 Jun 1996 08:58:59 GMT Organization: NeXT Software, Inc. Message-ID: <4p66kj$dei@news.next.com> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: > aisbell@cubicsol.com (Art Isbell) writes: > > I would do the following in hopes that NSArray implements lazy > > copying: > > > > NSMutableArray *theChildren; > > > > - (NSArray *)children > > { > > return [[theChildren copy] autorelease]; > > } > > I would do: > > NSMutableArray kids; > > -(NSArray *)children > { > return kids; > } > > This is the simplest and fastest implementation, and anybody talking > mutable methods to an NSArray should be fired. --Tiggr I also think this is the only resonable way to go. When object X gets (either as a return value or as a passed in argument) an object Y that's typed as an instance of an immutable class, this doesn't mean that Y will never change, just that X is not allowed to change it. Note that the argument that with the "returning an NSMutableArray" implementation one could do something like: id children = [someObject children]; [children addObject:aChild]; is pointless, since even with the "returning an NSArray" implementation an unreasonable programmer could do the similar: id children = [someObject children]; ((void **)children)[2+(((int *)children)[1])++] = (void *)aChild; This snippet of code is wrong for exactly the same reason the first snippet is wrong (because it makes unwarranted assumptions about how the children method and the object it returns is implemented). It is not even more wrong that the first one, just a bit more complex. A more interesting characteristings of how immutability works is this: Consider an object X knowing about a non-mutable object Y. Since the fact that Y is non-mutable doesn't guarantee it won't change, but just that X can't legally change it, X will needs to create a copy of Y if it wants to keep it intact over time. Thus the following code is incorrect, even though it is "legal": NSArray *children = [someObject children]; [someObject addChild:aChild]; NSLog(@"The children were: %@", children); The correct way to do this would be: NSArray *children = [[someObject copy] autorelease]; [someObject addChild:aChild]; NSLog(@"The children were: %@", children); (initWithArray: could have been used instead of copy if a shallow copy was needed). C= _________________________________ Charles d'Harcourt, cph@next.com, http://www.next.com/~cph NOT SPEAKING ON BEHALF OF NeXT!!!
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Problems with "cc -posix ..." Date: 4 Jun 1996 21:27:52 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4p29oo$gvg@news.its.com> References: <31B4046E.4D28@estel.uindy.edu> Steve Spicklemire <steve@estel.uindy.edu> wrote: [ ... ] > I'm trying to build some code > that requires posix functions (e.g., the bsddb library, > and freeWAIS-0.5,...). Anyway... if I put '-posix' on > the cc command line the .c files build OK, but I get > unresolved refs with -lposix (like floor, ceil, and log) > or -lm (then it's getcwd), but if I use -lm *and* -lposix > I get a boatload of multiply defined symbols. Has anyone > figured out how to handle this? Read the compiler documentation-- you will not get the correct linking semantics for POSIX executables by using "-lposix". You must compile everything using "cc -posix" and you must link using "cc -posix". Add libraries like "-lm" at the end of the link line. > I looked in the man page for a compiler flag that told the linker > to ignore multiple defs of functions (i.e., just take the first found > def) but I didn't see it in there. BTW I'm on NS Developer 3.2. Look at 'man ld' for the -m flag. (You should not have to use it to link a POSIX app, though....) -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 6 Jun 1996 04:34:13 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4p5n45$5lm@usenet.rpi.edu> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> louie@va.pubnix.com (Louis A. Mamakos) wrote: > >I'm well aware that basically it comes down to this old "the > >technologically excellent vs. the most widespread" thing, and > >we all know who apparently always has to suffer :-((( > > NeXTMail format isn't any technologially "better" than MIME is; > it's just different and proprietary and not significantly better. > NeXT missed a golden opportuntity to have a showcase MIME mail > user agent by having their NextMail.app program support MIME > rather than their own format. Didn't they have Mail.app shipping before MIME was finalized? I don't see why they couldn't pick some format and go with it, to have an example of a format, instead of waiting around until the MIME group figured out what they were going to do. I don't think they need to use MIME *rather* than their own format, which would have meant delaying any multimedia email support, but it would have been nice if they *added* the MIME formatting option much sooner than they did. > They didn't, and no one cared how whizzy NextMail was because it > wasn't interoperable with the rest of the world. No one cared now whizzy NeXTmail was because they didn't know. It had nothing much to do with interoperating with the rest of the world. I was doing NeXTmail quite nicely when there was no one else on campus doing *any* kind of multi-media email. It wouldn't have mattered if it had been MIME format, because users on the other platforms didn't have a decent mail application (by "decent" I mean "one usable by normal people"), so they were basically all using plain-text readers anyway. My first MIME message came from Mark Crispin, probably a year or two after I was setup with my first NeXT. And he was only doing it to say "See, you can't read this message, so NeXT sucks". Just about no one else on campus could have read it at the time either, but as I remember he didn't consider that point particularly relevent. > > In this case, though, it's not only different because NeXTMail > > would be just another *option*, allowing better quality in > > communication with a selected group of people (just like colour > > faxes to those who own colour fax machines). > > Not better. Different. I think there are some (few) areas where NeXTmail really is better, but I agree that there's no point in anyone (such as Sun) working on supporting it for other platforms. Not even for OpenStep platforms. There's many projects which would be have a better payoff. If Apple or Microsoft wanted to do NeXTmail for their OS's, then that might be significant. But getting NeXTmail supported on other Unix platforms at this point is irrelevant (given that all platforms are supporting MIME). It'd be nice for nostalgia of NeXTSTEP fans, but nothing else. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 6 Jun 1996 15:32:06 GMT Organization: Netcom Distribution: world Message-ID: <4p6tlm$2tj@dfw-ixnews4.ix.netcom.com> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> cph@next.com (Charles d'Harcourt) wrote: > tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: > > NSMutableArray kids; > > > > -(NSArray *)children > > { > > return kids; > > } > > I also think this is the only resonable way to go. When object X gets > (either as a return value or as a passed in argument) an object Y that's > typed as an instance of an immutable class, this doesn't mean that Y will > never change, just that X is not allowed to change it. > Thanks for responding on what seems to be a simple situation but really needs careful consideration. This exact situation must occur repeatedly with various classes supplied by NeXT, so can we assume that you *know* that NeXT uses this approach? > A more interesting characteristings of how immutability works is this: > Consider an object X knowing about a non-mutable object Y. Since the fact > that Y is non-mutable doesn't guarantee it won't change, but just that X > can't legally change it, X will needs to create a copy of Y if it wants to > keep it intact over time. Thus the following code is incorrect, even > though it is "legal": > NSArray *children = [someObject children]; > [someObject addChild:aChild]; > NSLog(@"The children were: %@", children); > The correct way to do this would be: > NSArray *children = [[someObject copy] autorelease]; > [someObject addChild:aChild]; > NSLog(@"The children were: %@", children); > (initWithArray: could have been used instead of copy if a shallow copy was > needed). But this seems dangerous as well because copying an array, a deep copy, will send each element a copy message which, although inherited from NSObject, is implemented by copyWithZone: which raises if not overridden by NSObject subclasses. Why wouldn't defaulting to a shallow copy if copyWithZone: hasn't been overridden be a better approach? Placing each copy within an exception handling domain to deal gracefully with this potential feels like a rather onerous solution for programmers. -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 6 Jun 1996 19:52:15 GMT Organization: Norden 1 Communications Message-ID: <4p7ctf$582@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------Commercial experience Objective C-----------------Commercial experience EOF-------------------------A plus Sybase or Oracle------------A plus Career Position-------------Full benefits & exceptional opportunity. Relocation------------------Company assistance Area------------------------ILL To Be Considered------------Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: brian@cadaver.acm.ndsu.NoDak.edu (Brian Glaeske) Newsgroups: comp.sys.next.programmer Subject: Comments requested on NXBrowser clip cell code Date: 6 Jun 1996 19:55:21 GMT Organization: NDSU ACM Message-ID: <4p7d39$8rs@daily-planet.nodak.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit I am looking on comments on the following code which implements a NXBrowserCell which clips the text contained in the cell before displaying it like the Workspace browser behaves. Comments I am looking for include but are not limited to: - Is this the right way to accomplish this task - Usage of malloc and free - How to not use the FUDGE factor in the width calculation - Should it ignore whitespace - Will this work correctly for all lazy and very lazy delegates? I only tested against a rather simple delegate. Thanks, -- Code -- #import <appkit/NXBrowserCell.h> @interface NXBrowserCellClip : NXBrowserCell { } - drawInside:(const NXRect *)cellFrame inView:aView; @end #import <appkit/Font.h> #import <appkit/Matrix.h> #import <appkit/NXImage.h> #import <appkit/NXBrowser.h> #include <string.h> #include <stdlib.h> #include <math.h> #define CLIPTEXT "..." #define FUDGE 4.0 #ifndef strdupn #define strdupn(x, n) strcpy((char *)malloc(sizeof(char) * (strlen(x) + n + 1)), x) #endif #ifndef strdup #define strdup(x) strcpy((char *)malloc(sizeof(char) * (strlen(x) + 1)),x) #endif @implementation NXBrowserCellClip - drawInside:(const NXRect *)cellFrame inView:aView { Font *mainFont = [self font]; Font *scrnFont = [mainFont screenFont]; Font *theFont = (scrnFont ? scrnFont : mainFont); // try to use screen font NXSize iconSize; int row, cellCol; id myBrowser; char *contentCopy, *tempCopy; float contentWidth, elipseWidth; if ([self isLoaded] == NO) { [aView getRow:&row andCol:&cellCol ofCell:self]; // It's a long way to get to my browser myBrowser = [[[aView superview] superview] superview]; /* This needs to be done because in the case of a lazy or very lazy browser, I will not be loaded yet, but since I am going to display I need this information. */ if([[myBrowser delegate] respondsTo:@selector(browser:loadCell:atRow:inColumn:)]) { [[myBrowser delegate] browser:myBrowser loadCell:self atRow:row inColumn:[myBrowser lastColumn]]; } } theFont = [self font]; contentWidth = ceil([theFont getWidthOf:contents]); [[[self class] branchIcon] getSize:&iconSize]; if (contentWidth < cellFrame->size.width - (iconSize.width + FUDGE)) { return [super drawInside:cellFrame inView:aView]; } else { elipseWidth = ceil([theFont getWidthOf:CLIPTEXT]); contentCopy = strdupn(contents, strlen(CLIPTEXT)); while(contentWidth >= cellFrame->size.width - (iconSize.width + FUDGE)) { contentCopy[strlen(contentCopy) - 1] = '\0' ; contentWidth = ceil([theFont getWidthOf:contentCopy]) + elipseWidth; } strcat(contentCopy, CLIPTEXT); tempCopy = contents; contents = contentCopy; [super drawInside:cellFrame inView:aView]; contents = tempCopy; free(contentCopy); } return self; } @end -- Brian Glaeske brian@cadaver.acm.ndsu.NoDak.edu ---------------------------------------------------------------------------- "Men Rule. Rap Sucks. Eat Fatty Food." The PIPC, Outland
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Thu, 06 Jun 1996 21:29:58 GMT Organization: Orbital Computer Consultancy Message-ID: <31b74471.157975@news.plsys.co.uk> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> cph@next.com (Charles d'Harcourt) wrote: >tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: >> aisbell@cubicsol.com (Art Isbell) writes: >> > I would do the following in hopes that NSArray implements lazy >> > copying: >> > >> > NSMutableArray *theChildren; >> > >> > - (NSArray *)children >> > { >> > return [[theChildren copy] autorelease]; >> > } >> >> I would do: >> >> NSMutableArray kids; >> >> -(NSArray *)children >> { >> return kids; >> } >> >> This is the simplest and fastest implementation, and anybody talking >> mutable methods to an NSArray should be fired. --Tiggr > >I also think this is the only resonable way to go. When object X gets >(either as a return value or as a passed in argument) an object Y that's >typed as an instance of an immutable class, this doesn't mean that Y will >never change, just that X is not allowed to change it. > I read what you say carefully and notice that you say "typed as an instance", not "is an instance of". You were careful to say this, but of course the fact is that typing an object has no bearing on what class it really is at run-time, much less whether it's immutable. What seems sensible to me is that if a method states it returns an NSArray, i.e. an immutable class, it _will_ return an immutable array. Syntatically it's valid to return NSArray's mutable subclass; but I think methods which return classes which recognise the mutability concept should be treated as special cases, i.e. the immutability will be honoured as well as the object really being an instance of the stated class. Next's production code seems to vindicate this to my mind; to the best of my knowledge, all the EOF methods ([EOEntity attributes], for example) really do return an immutable array. Like you said, you don't speak for Next. >Note that the argument that with the "returning an NSMutableArray" >implementation one could do something like: > id children = [someObject children]; > [children addObject:aChild]; >is pointless, since even with the "returning an NSArray" implementation an >unreasonable programmer could do the similar: > id children = [someObject children]; > ((void **)children)[2+(((int *)children)[1])++] = (void *)aChild; Ok, you win this month's cryptic code award. >This snippet of code is wrong for exactly the same reason the first >snippet is wrong (because it makes unwarranted assumptions about how the >children method and the object it returns is implemented). It is not even >more wrong that the first one, just a bit more complex. > And a whole lot less likely to be written, I'd wager. Of course you shouldn't build classes which protect against _any_ possible misuse, but you should build in reasonable protection. Protection against id typecasting is reasonable and necessary I'd say. >A more interesting characteristings of how immutability works is this: >Consider an object X knowing about a non-mutable object Y. Since the fact >that Y is non-mutable doesn't guarantee it won't change, Double-speak. Instances of immutable classes should be immutable, period. Exactly what is the point of FK introducing the immutability concept if it's not guaranteed that what you think is an immutable object won't change? (Pardon the triple negative!) > but just that X >can't legally change it, X will needs to create a copy of Y if it wants to >keep it intact over time. Thus the following code is incorrect, even >though it is "legal": > NSArray *children = [someObject children]; > [someObject addChild:aChild]; > NSLog(@"The children were: %@", children); >The correct way to do this would be: > NSArray *children = [[someObject copy] autorelease]; Did you mean [[[someObject children] copy] autorelease] ? > [someObject addChild:aChild]; > NSLog(@"The children were: %@", children); >(initWithArray: could have been used instead of copy if a shallow copy was >needed). > > C= >_________________________________ >Charles d'Harcourt, cph@next.com, http://www.next.com/~cph >NOT SPEAKING ON BEHALF OF NeXT!!!
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: RunTime Date: Thu, 06 Jun 1996 21:37:10 GMT Organization: Orbital Computer Consultancy Message-ID: <31b74f38.2917053@news.plsys.co.uk> References: <4p1k0n$nq9@news.u-bordeaux.fr> baudouin@firmin.labri.u-bordeaux.fr (Thierry Baudouin [Paries]) wrote: >I am working with runtime Functions and i have a problem : I want to >call dynamically methods for objects using the method "performv" : > > -performv:SEL :void * >First argument is the selector and second argument is for arguments of >method. >The return value of "performv" is an id ,object, and if the method >called >dynamically return an other type a problem appear. However, there is a >good conversion with int,short,long,char and any pointer but a bad >conversion >with float and double.How can i get with this two types. > >Thanks to answer if you have a idea. Take a look at /NextDeveloper/Examples/IndexingKit/Ledger/JFTableVectorConfiguration.m
From: jhills@unix.infoserve.net Newsgroups: comp.sys.next.programmer Subject: perl Date: 7 Jun 1996 04:15:09 GMT Message-ID: <4p8acd$atg@news.infoserve.net> Where can I find an implementation of perl that will compile on my NeXT? Thanks.
From: BCOG Webmaster <webmaster@bcog.org> Newsgroups: comp.sys.next.programmer Subject: Re: perl Date: 7 Jun 1996 05:48:33 GMT Organization: British Columbia Openstep Group Message-ID: <4p8frh$t6j@scipio.cyberstore.ca> References: <4p8acd$atg@news.infoserve.net> jhills@unix.infoserve.net wrote: >Where can I find an implementation of perl that will compile on my NeXT? >Thanks. ftp://ftp.bcog.org/pub/unix/perl5.001m.tar.gz Enjoy! -- Webmaster for the British Columbia Openstep Group <webmaster@bcog.org> + <URL:http://www.bcog.org> * MIME/NeXTmail welcome * PGP key available *
Newsgroups: comp.sys.next.programmer From: Stefan Kruger <stefan@cs.bris.ac.uk> Subject: how compile c++ under nextstep ? Content-Type: text/plain; charset=us-ascii Message-ID: <31B801BA.167E@cs.bris.ac.uk> Sender: usenet@uns.bris.ac.uk (Usenet news owner) Content-Transfer-Encoding: 7bit Organization: University of Bristol, England Mime-Version: 1.0 Date: Fri, 7 Jun 1996 10:17:30 GMT Hello, I need to be able to compile some c++ code under nextstep. I was under the impression that this is possible, but I must be missing something central here: tranquillity% cat hello.cc #include <iostream.h> int main(int argc, char **argv) { cout << "hello world\n"; } tranquillity% c++ -g -o hello hello.cc ld: Undefined symbols: ostream::operator<<(char const *) cout tranquillity% Right, either I'm required to explicitly link in some library here, or something is not installed correctly ? Greatful for any clarification, and sorry if this is a faq. (NSi86 3.3) cheers, stef -- Stefan Kruger (stefan@cs.bris.ac.uk) Finger me for pgp public key. University of Bristol (http://www.cs.cf.ac.uk/User/Stefan.Kruger)
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: RCS with interface? Date: 7 Jun 1996 11:00:18 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4p9243$b1b@ai.alienor.fr> Hi, I'm looking for an interfaced version of RCS working with HP-PA. Can you direct me the best place to get it please. Thanks Fred. -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 7 Jun 1996 11:16:45 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4p932t$473@zdi.informatik.uni-stuttgart.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> In <31b74471.157975@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: >cph@next.com (Charles d'Harcourt) wrote: >What seems sensible to me is that if a method states it returns an >NSArray, i.e. an immutable class, it _will_ return an immutable array. >Syntatically it's valid to return NSArray's mutable subclass; but I >think methods which return classes which recognise the mutability >concept should be treated as special cases, i.e. the immutability will >be honoured as well as the object really being an instance of the >stated class. Sorry, but then NSMutableArray must not be a subclass of NSArray - but it is. Something like (NSArray *) just says something about the view on an object, not about the concrete instance of an object. This is polymorphism and what OOP is all about! Anyway the class is named NSArray and not NSImmutableArray, so I don't see any assurences about immutability here. Creating an NSImmutableArray subclass and use it as an return type would make the whole thing clearer. >Protection against id typecasting is reasonable and necessary I'd say. And I'd say, strong typing was invented to prevent the programmer from making hard-to-find type errors, and not to hinder every malicious hacker from doing illegal things. >Double-speak. Instances of immutable classes should be immutable, >period. Exactly what is the point of FK introducing the immutability >concept if it's not guaranteed that what you think is an immutable >object won't change? (Pardon the triple negative!) Because it does not GUARANTEE immutability. (Pardon accepted.) By the way: const char *foofunction(.....); ... char *str = (char *)foofunction(.....); *str = '?'; Would you say this is silly code? Yes? Would you, as an implementor of foofunction(), create a copy of the string before returning it? Just to make the above silly code "legal" or just not harmfull? Would any sensible programmer waste cpu time and main memory just for the case that some prgrammer got wiered? DID YOU EVER DO THAT? I think the answer is no. And it's the same problem with NSArray/ NSMutableArray. Typing ensures that no type errors will occur - provided everyone adheres to the rules implied. Just my $2. Bye, Klaus Brouwer
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: perl Date: 7 Jun 1996 11:48:22 GMT Organization: University of Sheffield, UK Message-ID: <4p94u6$2dg@bignews.shef.ac.uk> References: <4p8acd$atg@news.infoserve.net> <4p8frh$t6j@scipio.cyberstore.ca> In-Reply-To: <4p8frh$t6j@scipio.cyberstore.ca> On 06/07/96, BCOG Webmaster wrote: > jhills@unix.infoserve.net wrote: > > >Where can I find an implementation of perl that will compile on my NeXT? > >Thanks. > > ftp://ftp.bcog.org/pub/unix/perl5.001m.tar.gz > 5.002 is precompiled on Peanuts: ftp://www2.informatik.uni-muenchen.de/pub/comp/platforms/next/Unix/script/ perl.5002.NIHS.b.tar.gz Best wishes, mmalc. --
From: flight@mathi.uni-heidelberg.de (Gregor Hoffleit) Newsgroups: comp.sys.next.programmer Subject: Source for flex-PB and bison-PB ? Date: 7 Jun 1996 12:59:49 GMT Organization: Mathematisches Institut, Uni Heidelberg Message-ID: <4p9945$rs4@sun0.urz.uni-heidelberg.de> A kind, anonymous soul put `PB-aware' binary packages of bison and flex on the archives. Would it be possible to post the sources, too, or - even better - a diff file against the original source tree ? Many thanks in advance, Gregor -- | Gregor Hoffleit admin MATHInet / contact RhiNO | | MAIL: Mathematisches Institut PHONE: (49)6221 54-5771 | | INF 288, 69120 Heidelberg / Germany FAX: 54-8312 | | EMAIL: flight@mathi.uni-heidelberg.de (NeXTmail, MIME) |
From: flight@mathi.uni-heidelberg.de Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4p98mo$rmj@sun0.urz.uni-heidelberg.de> Control: cancel <4p98mo$rmj@sun0.urz.uni-heidelberg.de> Date: 7 Jun 1996 12:54:26 GMT Organization: University of Heidelberg, Germany Message-ID: <4p98q2$rmj@sun0.urz.uni-heidelberg.de> ignore Article canceled by slrn 0.8.8.2
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: OmniWeb Hierarchical List Date: 7 Jun 1996 13:33:23 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4p9b33$f76@lion.embl-heidelberg.de> References: <4p293o$ra1@news.onramp.net> In article <4p293o$ra1@news.onramp.net> dekorte@suite.com (Steve Dekorte) writes: > > Is there a PD palette (or NeXT cell view) for a hierarchical list like in > OmniWebs hot-links window or InterfaceBuilder's "outline" view of classes > and instances? I wrote such a class, but is not completely finished. It will be part of the new MiscKit. I will finish & submit it whet I get NS4.0. Although officially I was even on the NS4.--alpha tester list [flame on !&*)(&*(^%^&$%$%^ flame off], I'm still waiting for a copy. If you want to see it sooner, call, or mail the NeXT office in Germany and tell them "!@!^&%%#%@^%" See my www page for more: http://www.nmr.embl-heidelberg.de/tuparev ... Projects .... MiscKit -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: how compile c++ under nextstep ? Date: 7 Jun 1996 13:17:05 GMT Organization: NO ORGANIZATION, INC. Message-ID: <4p9a4h$5uo@bias.ipc.uni-tuebingen.de> References: <31B801BA.167E@cs.bris.ac.uk> Cc: stefan@cs.bris.ac.uk In <31B801BA.167E@cs.bris.ac.uk> Stefan Kruger wrote: > Hello, > > I need to be able to compile some c++ code under nextstep. > I was under the impression that this is possible, but I > must be missing something central here: > > tranquillity% cat hello.cc > #include <iostream.h> > > int main(int argc, char **argv) > { > cout << "hello world\n"; > } > > tranquillity% c++ -g -o hello hello.cc > ld: Undefined symbols: > ostream::operator<<(char const *) > cout > tranquillity% > > Right, either I'm required to explicitly link in some > library here, or something is not installed correctly ? > > Greatful for any clarification, and sorry if this is a faq. > (NSi86 3.3) > % cc++ hello.cc -lg++ % a.out hello world be sure to install libg++, it should be available on the archives. Bye Frank -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 7 Jun 1996 14:06:47 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4p9d1n$lo@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> louie@va.pubnix.com (Louis A. Mamakos) wrote: > NeXTMail format isn't any technologially "better" than MIME is; it's > just different and proprietary and not significantly better. I cannot but repeat it: It IS better in a (to me) significant way because it can use text colour and different fonts. Me, I can't live without that, it's absolutely critical for the way I use E-Mail. Is it so hard to accept that? Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: stes@kokomo.wri.com (David Stes) Newsgroups: comp.sys.next.programmer Subject: Re: OmniWeb Hierarchical List Date: 7 Jun 1996 14:48:25 GMT Organization: Wolfram Research, Inc. Distribution: world Message-ID: <4p9ffp$r1u@dragonfly.wolfram.com> References: <4p9b33$f76@lion.embl-heidelberg.de> Georg Tuparev writes > > I will finish & submit it whet I get NS4.0. Although officially I was > even on the NS4.--alpha tester list [flame on !&*)(&*(^%^&$%$%^ flame > off], I'm still waiting for a copy. If you want to see it sooner, call, > or mail the NeXT office in Germany and tell them "!@!^&%%#%@^%" Here we now have a nice picture of the average NSObject supporter.
From: michal@gortel.phys.ualberta.ca (Michal Jaegermann) Newsgroups: comp.sys.next.programmer Subject: Re: perl Date: 7 Jun 1996 16:56:10 GMT Organization: Disorganized Bits Message-ID: <4p9mva$14dg@pulp.ucs.ualberta.ca> References: <4p8acd$atg@news.infoserve.net> <4p8frh$t6j@scipio.cyberstore.ca> BCOG Webmaster (webmaster@bcog.org) wrote: : jhills@unix.infoserve.net wrote: : : >Where can I find an implementation of perl that will compile on my NeXT? : >Thanks. : : ftp://ftp.bcog.org/pub/unix/perl5.001m.tar.gz : Which is obsolete. The current release is 5.002 + patches. It can be found on any CPAN archive and compiles on NeXT just out-of-the-box. --mj p.s. Just for a reference - here is a list of CPAN sites (likely incomplete) Africa South Africa ftp://ftp.is.co.za/programming/perl/CPAN/ Asia Japan ftp://ftp.lab.kdd.co.jp/lang/perl/CPAN/ Taiwan ftp://dongpo.math.ncu.edu.tw/perl/CPAN/ Australasia Australia ftp://coombs.anu.edu.au/pub/perl/ ftp://ftp.mame.mu.oz.au/pub/perl/CPAN/ New Zealand ftp://ftp.tekotago.ac.nz/pub/perl/CPAN/ Canada British Columbia ftp://mango.pinc.com/pub/perl/CPAN/ Europe Czech Republic ftp://sunsite.mff.cuni.cz/Languages/Perl/CPAN/ Finland ftp://ftp.funet.fi/pub/languages/perl/CPAN/ France ftp://ftp.ibp.fr/pub/perl/CPAN/ ftp://ftp.pasteur.fr/pub/computing/unix/perl/CPAN/ Germany ftp://ftp.leo.org/pub/comp/programming/languages/perl/CPAN/ ftp://ftp.rz.ruhr-uni-bochum.de/pub/CPAN/ Poland ftp://ftp.pk.edu.pl/pub/lang/perl/CPAN/ Portugal ftp://ftp.ci.uminho.pt/pub/lang/perl/ Slovenia ftp://ftp.arnes.si/software/perl/CPAN/ Sweden ftp://ftp.sunet.se/pub/lang/perl/CPAN/ Switzerland ftp://ftp.switch.ch/mirror/CPAN/ the Netherlands ftp://ftp.cs.ruu.nl/pub/PERL/CPAN/ UK ftp://ftp.demon.co.uk/pub/mirrors/perl/CPAN/ ftp://unix.hensa.ac.uk/mirrors/perl-CPAN/ USA Florida ftp://ftp.cis.ufl.edu/pub/perl/CPAN/ Illinois ftp://uiarchive.cso.uiuc.edu/pub/lang/perl/CPAN/ Massachusetts ftp://ftp.delphi.com/pub/mirrors/packages/perl/CPAN/ Oklahoma ftp://ftp.uoknor.edu/mirrors/CPAN/ Texas ftp://ftp.metronet.com/pub/perl/ ftp://ftp.sedl.org/pub/mirrors/CPAN/ ftp://ftp.sterling.com/CPAN/
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: NeXTMail is just different Message-ID: <Dsn8zF.3xs@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p4qsd$74p@news4.digex.net> Date: Fri, 7 Jun 1996 19:04:26 GMT In article <4p4qsd$74p@news4.digex.net>, John Kheit <jkheit@cnj.digex.net> wrote: >Well, I think it is both different and better, especially when you have the >choice to down grade to MIME. Quite honestly, the RTFD used for NeXTmail >is canvas allowing richer expression than the MIME pallatte. Wrong. NeXTMail is just different. Mynah uses the same Text Class that NeXTMail does, except it breaks it down and encodes it as MIME, instead of as RTF. I can have different fonts/sizes/colors, files, and pictures in MIME messages. They look to the user just like NeXTMail messages. No difference in capabilities of the message format. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
Newsgroups: comp.sys.next.programmer From: dhoman@il.us.swissbank.com (David Homan) Subject: EOF : is there an equivalent for dbsetifile() ? Message-ID: <1996Jun7.191946.7165@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division Date: Fri, 7 Jun 1996 19:19:46 GMT I noticed that NeXT's EO Sybase adaptor will not connect unless the interface file can be located in the ~sybase directory. Unfortunately, I need to store the interface file in another directory. Does anyone know of a way to achieve the affect of a dbsetifile() call within the framework of the EO kit? I can hack around this problem by linking ~sybase/interfaces to my interface file - but this solution is far from optimal. Please respond directly to me, since I am not a frequent reader of news. Thank you, - David --- David Homan SBC Warburg 222 Broadway New York, NY 10038 All opinions are mine - not my employer's.
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 7 Jun 1996 20:22:25 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pa321$852@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > In article <4p4qsd$74p@news4.digex.net>, > John Kheit <jkheit@cnj.digex.net> wrote: > >Well, I think it is both different and better, especially when you have the > >choice to down grade to MIME. Quite honestly, the RTFD used for NeXTmail > >is canvas allowing richer expression than the MIME pallatte. > Wrong. NeXTMail is just different. Very confident aren't you...Maybe I'm behind the times...Let's find out... > Mynah uses the same Text Class that NeXTMail does, except it breaks it down and encodes it as MIME, instead of as RTF. I can have different fonts/sizes/colors, files, and pictures in MIME messages. They look to the user just like NeXTMail messages. No difference in capabilities of the message format. That is the point... MIME<RTF(D) as a format. Doesn't matter that the MIME format is in a sense converted back into and RTF text object... As far as MIME being as capable...Is Mynah a NS app? Regardless, is it using some custom extensions to MIME like the Netscape folks custom extensions of html? If so, that is not MIME... Anyway, the MIME apps I've seen, and the one I'm using do not support any postscript the fonts I use, nor doe they remember color and formatting details anywhere nearly as accurately as RTF(D). Fractional spaces, line sizes, various tab formats, etc...etc... I'd appreciate you sending me a MIME mail that has all of the capabilities of RTF(D)...like your résumé or some such decently heavily laid out document...mine goes through just dandy in NeXTmail, and just terribly in MIME... Also, I'd appreciate getting info as to what version of MIM supports the following: Accurate memory of fonts used (not merely an HTML-like font substitution for common serif, heading, and mono spaced fonts (e.g.--the original could have had Garmond, Avante Garde and Ohlfs, and the MIME product I'll get will be Times, Helvetica, and courier even though I have all the fonts used in the original), color text, variable line spacing, varying margins and tabs... -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
Newsgroups: comp.sys.next.programmer From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Subject: Re: How to use (Im)mutability Message-ID: <Dsn0Kt.FIo@lazrus.cca.rockwell.com> Sender: news@lazrus.cca.rockwell.com Cc: k_harbour@mail.bogo.co.uk Organization: Rockwell Avionics - Collins References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> Date: Fri, 7 Jun 1996 16:02:52 GMT >Double-speak. Instances of immutable classes should be immutable, >period. Exactly what is the point of FK introducing the immutability >concept if it's not guaranteed that what you think is an immutable >object won't change? (Pardon the triple negative!) This is not a reasonable criticism. Maybe we just disagree about the definition of immutable. I think it means that "I" can not change it. You think it means the "nobody" can change it. Why can't one instance of a class maintain a reference to a mutable container as a mutable container and every one else reference it immutably ? I use this all the time. One class maintains and modifies a collection and every one else references (read only). - Erik M. Buck
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <eott.com!tappd@nuchat.sccsi.com> Message-ID: <9606072207.AA13942@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@eott.com> Date: Fri, 7 Jun 96 17:07:42 -0500 Subject: Fishing for NextMeisters Ahem...any Next-experienced folks in the Houston area looking for full-time work in the near future? Pls. reply privately to: tappd@copernicus.sccsi.com Regards, - Dan
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 7 Jun 1996 23:38:47 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4paei7$j0@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: > I think there are some (few) areas where NeXTmail really is better, > [...] > It'd be nice for nostalgia of NeXTSTEP fans, > but nothing else. What with those who need it as a communication system? Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: edwintam@webhk.com (Edwin TAM) Newsgroups: comp.sys.next.programmer Subject: HELP: EOF without UI, code example? Date: 8 Jun 1996 07:21:06 GMT Message-ID: <4pb9l2$hpi@hkt001.hkt.net> Hi, I am developing a application to insert records into an Oracle database, the application has to be run without having somebody to login to the Workspace, ie, it is sort of a daemon, pulling data from a serial port and dump the results into the database. I had written the basic stuffs using EOF and no UI elements but it has to be lauched in Workspace, what is missing ? Is it possible to develop applications using EOF without interfacing to the UI layer ? Any code example would be very helpful. Please help, thanks a lot. Edwin TAM
From: Jacob Nielsen <jacob@dannug.dk> Newsgroups: comp.sys.next.programmer Subject: Re: RTF to PS Filter Date: 3 Jun 1996 22:09:57 GMT Organization: Danish NeXT User Group Message-ID: <4ovnrl$bm@jnext.dannug.dk> References: <1996May30.190123.23765@cyantic.com> <4or3na$k9n@digifix.digifix.com> sanguish@digifix.com (Scott Anguish) wrote: > On 05/30/96, Paul Guiness wrote: > >Hi, > > > >I am looking for ways to convert an RTF file to PostScript without > using an application built with IB. My options seem to be: > > 1. Create a tool that programatically takes care of allocating > a Text object within a View and sending a printPSCode message to it. > > 2. Finding a filter that converts RTF (NeXT RTF) to PS without > using the ApplicationKit at all. > > Eric P. Scott wrote something that did this... and for the > longest time it sat in my Unix folder after he posted it, and I would > email it to those who asked... > > Unfortunately, it appears to have been lost... Here it is: Regards, Jacob // Print RTF files from the command line // (without a connection to the Window Server!) // Eric P. Scott, San Francisco State University, November 1993 // + for NS 2.1 and 3.1 // cc -o printrtf -s -O printrtf.m -lNeXT_s #ifdef NX_COMPILER_RELEASE_3_0 #include <ansi/stdio.h> #import <defaults/defaults.h> #else #include <stdio.h> #import <appkit/defaults.h> #endif #import <streams/streams.h> #import <objc/Object.h> #import <objc/Storage.h> #import <appkit/PrintInfo.h> #import <appkit/Text.h> #import <dpsclient/dpsclient.h> #import <dpsclient/psops.h> @interface FakeApp:Object { id _focusStack; id printInfo; } + initialize; - init; - printInfo; - (BOOL)shouldRunPrintPanel:aView; - (const char *)appName; - (const char *const *)systemLanguages; - _focusStack; - (DPSContext)context; #ifdef NX_COMPILER_RELEASE_3_0 - (BOOL)_calibratedColorOK; #endif @end @implementation FakeApp + initialize { static NXDefaultsVector PRDefaults = { { "NXFont", "Helvetica" }, { "NXFontSize", "12" }, { "NXMargins", "72 72 90 90" }, { "NXPaperType", "Letter" }, #ifdef NX_COMPILER_RELEASE_3_0 { "NXBoldSystemFonts", "Helvetica" }, { "NXSystemFonts", "Helvetica-Bold" }, #endif { (char *)NULL, (char *)NULL } }; NXRegisterDefaults("printrtf", PRDefaults); return self; } - init { [super init]; _focusStack=[[Storage allocFromZone:[self zone]] initCount:0 elementSize:10 description:"@*c"]; return self; } - printInfo { if (!printInfo) printInfo=[[PrintInfo allocFromZone:[self zone]] init]; return printInfo; } - (BOOL)shouldRunPrintPanel:aView { return NO; } - (const char *)appName { return "printrtf"; } - (const char *const *)systemLanguages { return (const char *const *)NULL; } - _focusStack { return _focusStack; } - (DPSContext)context { return (DPSContext)NULL; } #ifdef NX_COMPILER_RELEASE_3_0 - (BOOL)_calibratedColorOK { return YES; } #endif @end @interface Text(printrtf) - (BOOL)lockFocus; - unlockFocus; - (BOOL)getVisibleRect:(NXRect *)theRect; #ifndef NX_COMPILER_RELEASE_3_0 - beginPSOutput; #endif - display:(const NXRect *)rects :(int)rectCount :(BOOL)clipFlag; - spoolFile:(const char *)filename; @end @implementation Text(printrtf) - (BOOL)lockFocus { return (NXDrawingStatus==NX_PRINTING) ? NO : [super lockFocus]; } - unlockFocus { return (NXDrawingStatus==NX_PRINTING) ? self : [super unlockFocus]; } - (BOOL)getVisibleRect:(NXRect *)theRect { if (NXDrawingStatus!=NX_PRINTING) return [super getVisibleRect:theRect]; [self getBounds:theRect]; return YES; } #ifndef NX_COMPILER_RELEASE_3_0 - beginPSOutput { extern id NXApp; [[NXApp printInfo] setPageOrder:NX_ASCENDINGORDER]; return [super beginPSOutput]; } #endif - display:(const NXRect *)rects :(int)rectCount :(BOOL)clipFlag { if (NXDrawingStatus!=NX_PRINTING) return [super display:rects:rectCount:clipFlag]; PSgsave(); PStranslate(0.0, -frame.size.height+ (rects[0].size.height+rects[0].origin.y)*2.0); PSscale(1.0, -1.0); [self drawSelf:rects:rectCount]; PSgrestore(); return self; } - spoolFile:(const char *)filename { extern id NXApp; id pi; const char *pn; int cp; char cmd[256]; pi=[NXApp printInfo]; (void)strcpy(cmd, "exec lpr"); #ifdef NX_COMPILER_RELEASE_3_0 if ((pn=[[pi printer] name])&&*pn) { #else if ((pn=[pi printerName])&&*pn) { #endif (void)strcat(cmd, " -P"); (void)strcat(cmd, pn); } if ((cp=[pi copies])>1) (void)sprintf(&cmd[strlen(cmd)], " -#%d", cp); (void)strcat(cmd, " -r -s "); (void)strcat(cmd, filename); (void)system(cmd); /* really should check status here! */ return self; } @end main(int argc, char *argv[]) { extern id NXApp; register int i; NXStream *f; Text *t; int errs; NXRect fr; char *streambuf; int len, maxlen; id pi; char path[24]; errs=0; [Text excludeFromServicesMenu:YES]; NXApp=[[FakeApp alloc] init]; pi=[NXApp printInfo]; [pi setHorizCentered:NO]; [pi setVertCentered:NO]; #ifdef NX_COMPILER_RELEASE_3_0 [pi setReversePageOrder:YES]; #endif for (i=1;i<argc;i++) { const NXSize large={ 504.0, 1.0e38 }; if (!(f=NXMapFile(argv[i], NX_READONLY))) { perror(argv[i]); errs++; continue; } NXGetMemoryBuffer(f, &streambuf, &len, &maxlen); NXSetRect(&fr, 0.0, 0.0, 504.0, 21.0); t=[[Text alloc] initFrame:&fr]; [t setMaxSize:&large]; [t setVertResizable:YES]; if (streambuf&&!strncmp(streambuf, "{\\rtf0\\", 7)) { [t setMonoFont:NO]; [t readRichText:f]; } else [t readText:f]; [t sizeToFit]; (void)strcpy(path, "/tmp/.printrtfXXXXXX.ps"); (void)NXGetTempFilename(path, 14); [pi setOutputFile:path]; [t printPSCode:NXApp]; NXCloseMemory(f, NX_FREEBUFFER); } exit(errs); } -- Jacob Nielsen Email: jacob@dannug.dk WWW: http://www.dannug.dk/~jacob/
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: how to issue a modem command at computer shutdown? Date: 8 Jun 1996 11:30:06 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4pbo7u$hk@tallowcross.uni-frankfurt.de> Hi, I've just added DRAM to my ZyXEL 2864 thereby enabling it to receive faxes while the computer is turned off. Unfortunately, it only does so if issued two special commands (at+fclass=z and then at+fzf=3). Of course it would be a drag to issue these commands manually each time you turn off the computer so I'm looking for a way to do that automatically. This arises two problems for me: 1. How do I issue these two commands over the serial port not using tip or the like but a little script or C program? This may be an easy one for a professional programmer, but I'm none :-( 2. (I'm afraid that makes much more problems): How do I execute this little script/program automatically when the computer is turned off? It must be started AFTER all apps have already been terminated to make sure no other program locks the serial device. Basically, I'm looking for something like /etc/rc to make an entry in, but for the other way round; but I couldn't find anything :-( Any help would be much appreciated! Thanks! Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
Newsgroups: comp.sys.next.programmer From: alex@gauss.franken.de (Alexander Much) Subject: Re: how compile c++ under nextstep ? In-Reply-To: Stefan Kruger's message of Fri, 7 Jun 1996 10:17:30 GMT Message-ID: <x77mtieibu.fsf@gauss.franken.de> Sender: alex@gauss.franken.de Organization: KNF (KommunikationsNetz Franken e.V.) References: <31B801BA.167E@cs.bris.ac.uk> Date: Sat, 8 Jun 1996 09:49:41 GMT I need to be able to compile some c++ code under nextstep. I was under the impression that this is possible, but I must be missing something central here: tranquillity% cat hello.cc #include <iostream.h> int main(int argc, char **argv) { cout << "hello world\n"; } tranquillity% c++ -g -o hello hello.cc ld: Undefined symbols: ostream::operator<<(char const *) cout tranquillity% Right, either I'm required to explicitly link in some library here, or something is not installed correctly ? Yes, libg++. You can use the libg++ from next in /usr/lib. But if you want to program more in C++ install gcc-2.7.2 and the GNU-libg++. -- Alex +----------------------------------------------------------------+ | Alexander Much Voice: +49-9131-207850Q | | alex@gauss.franken.de Public-Key on request | | <A HREF="http://www.franken.de/users/gauss/alexander/"> | | PGP: 1024/9F7C4F05 BADC 2E8A 0931 8967 646E 6C8D A2C9 31F2 | +----------------------------------------------------------------+
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Sat, 8 Jun 1996 12:31:32 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <sliOfYC00iWV01cK4O@andrew.cmu.edu> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> In-Reply-To: <4paei7$j0@tallowcross.uni-frankfurt.de> Excerpts from netnews.comp.sys.next.advocacy: 7-Jun-96 Re: Q:Sound[Kit] & OPENSTEP? by Uli Zappe@tallowcross.un > Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: >> I think there are some (few) areas where NeXTmail really is better, >> [...] >> It'd be nice for nostalgia of NeXTSTEP fans, >> but nothing else. > > What with those who need it as a communication system? A communications system is something that allows people to exchange information. The way multimedia mail is exchanged on the Internet is called MIME (for 'Multipurpose Internet Mail Extensions') and is well-documented via RFCs. NeXTmail is not a publicly-documented format, and it's likely that the only people who will ever use NeXTmail are NEXTSTEP users. Judging by their utility as 'communications systems', MIME is superior to NeXTmail because there are a couple of orders of magnitude more people with mail readers that understand MIME messages than understand NeXTmail. NeXTmail is completely unintelligible to non-NeXTmail mail readers, whereas MIME works pretty well even in non-MIME readers. Uli, you seem to be convinced that you either have to use NeXTmail or you cannot communicate at all. Obviously, that is nonsense (people have managed to deal with 7-bit ASCII email for decades!), but if you really want the precise formatting capabilities of NeXTmail, why can't you simply send an RTF(D) document as an enclosure in a MIME message? -Chuck PS: I like NeXTmail more than MIME too. The fact that MIME is the accepted standard and NeXTmail is going to become extinct is unfortunate, but you can place the resposibility for this squarely on NeXT. NeXTmail was something that NeXT regarded as proprietary and as a 'competitive advantage' of NEXTSTEP as an operating system. NeXT deliberately avoided supporting MIME natively for several years (since MIME was the alternative in competition to NeXTmail) and NeXT's refusal to publicly document the NeXTmail format as a proposed Internet standard via the RFC process ensured that no one else could write their own NeXTmail-compliant reader. NeXT's strategy succeeded, in a way. Read Mark Anenberg's comments again: "Its not a matter of stopping short; its a matter of a conscious limiting effort; It doesn't make sense to send email in a proprietary format when the rest of the world speaks Mime/X.400/X.500. The whole point is information exchange." Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 8 Jun 1996 17:20:40 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4pccp8$hk@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > Mynah uses the same Text Class that NeXTMail does, except it breaks it > down and encodes it as MIME, instead of as RTF. I can have different > fonts/sizes/colors, files, and pictures in MIME messages. They look to > the user just like NeXTMail messages. No difference in capabilities of > the message format. Sounds interesting. What is Mynah? Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: bmont@rconnect.com (Bruce Montegani) Newsgroups: comp.sys.next.programmer Subject: WebObjects:Can't start DefaultApp from command line Date: 7 Jun 1996 22:36:27 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <4paatb$gq5@news.mr.net> I'm using NextStep 3.3 with WebObjects 1.0 installed. The instructions in the documentation say to start the HelloWorld application manually do the following: 1. Change to /NextLibrary/WebObjects/Executables directory. 2. Enter DefaultApp Examples/HelloWorld I did this and the message came back that DefaultApp: command not found. However when I entered the command ./DefaultApp Examples/HelloWorld the application started. Why is this? Also, even though the application started the terminal window did not return to the prompt. It just kind of hanged. When I closed the terminal window the DefaultApp also ended. How come the command did not put the DefaultApp in the background?? Thanks Bruce Montegani
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer Subject: Re: WebObjects:Can't start DefaultApp from command line Date: 8 Jun 1996 22:45:01 GMT Organization: University of Sheffield, UK Message-ID: <4pcvpd$kga@bignews.shef.ac.uk> References: <4paatb$gq5@news.mr.net> In-Reply-To: <4paatb$gq5@news.mr.net> On 06/08/96, Bruce Montegani wrote: > The instructions in the documentation say to start the HelloWorld > application manually do the following: > > 1. Change to /NextLibrary/WebObjects/Executables directory. > > 2. Enter DefaultApp Examples/HelloWorld > > I did this and the message came back that DefaultApp: command not found. > > However when I entered the command ./DefaultApp Examples/HelloWorld the > application started. > > Why is this? > For some reason the current working directory (.) isn't part of your PATH environment variable. What to do about this will depend on what shell you're using. If you're using the c-shell (csh) you could add the following to the .cshrc file in your home directory: set path=(. $path) I wonder why it's not in your path already, though? Are you logged in as root? > Also, even though the application started the terminal window did not > return to the prompt. It just kind of hanged. When I closed the terminal > window the DefaultApp also ended. How come the command did not put the > DefaultApp in the background?? > Because you haven't told it to run in the background! :-) To do this, use the & parameter: ./DefaultApp Examples/HelloWorld & It will then return. If you're debugging the programme, however, I'd advise you to let it run in the foreground, as any logging will be output to this shell, and it'll be easier to make sense of if your have the app in the foreground. If you need another shell, just use Command-n in Terminal. If you want to background the app later, use Command-z, and then bg . You may find it useful to find a good introductory book to Unix -- if you do find a good one, please let me know! :-) Best wishes, mmalc. --
From: Paul Lynch <Paul_Lynch@plsys.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: WebObjects: Can't start DefaultApp from command line Date: Sat, 8 Jun 1996 22:25:33 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Jun8.222533.27649@seer.demon.co.uk> References: <4paatb$gq5@news.mr.net> In article <4paatb$gq5@news.mr.net> bmont@rconnect.com (Bruce Montegani) writes: > 2. Enter DefaultApp Examples/HelloWorld > > I did this and the message came back that DefaultApp: command not found. > > However when I entered the command ./DefaultApp Examples/HelloWorld the > application started. > > Why is this? Because . isn't in your path (and it shouldn't be in the path of root). > Also, even though the application started the terminal window did not return > to the prompt. It just kind of hanged. When I closed the terminal window the > DefaultApp also ended. How come the command did not put the DefaultApp in the > background?? Because you didn't tell it to run in the background! Try putting a & at the end of the command line. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: nettech@crl.com (Joseph W. Stroup) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Web (Netscape Questions) Date: 8 Jun 1996 19:55:28 -0700 Organization: CRL Network Services (415) 705-6060 [Login: guest] Message-ID: <4pdef0$2cp@crl.crl.com> Does the Netscape server & browser work under NExt 3.2 O/S ? Whats the deal witht their web software ? -- Joseph Stroup " I don't work for the Cable Company.
From: Håkan Jonsson <hajons@fy.chalmers.se> Newsgroups: comp.sys.next.programmer Subject: Q:Huge object file. Date: Fri, 07 Jun 1996 09:16:46 +0200 Organization: Chalmers University of Technology Message-ID: <31B7D75E.6D6@fy.chalmers.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi! I have developed a program using Sun's cc compiler, which have worked fine. When I tried to compile it on my NS machine a problem appeared: On the Sun the size of the objectfile was 43kb, but on NS it turned out to be 41MB. Does anyone have any idea how this happens? Yours, Hakan Jonsson
From: nazari@cs.chalmers.se (Nader Nazari,1018) Newsgroups: comp.sys.next.programmer Subject: HelP Text and Renderman together Date: 7 Jun 1996 14:13:52 GMT Organization: Chalmers University of Technology Message-ID: <4p9df0$ibp@nyheter.chalmers.se> Hi I am working on a project and most of the time I work with 3D visualization. I have some " shapes" on my " shape hierarchy" and I move the light and change the distance and rotate the figure and ... as you can see in the famous Renderman example which can be found here: /NextDeveloper/Examples/3Dkit/Simple Now I need to write some text on the different part of the figure and also draw some line and curve on it . I can do it with postscript but the problem is that I want to "move" the whole figure all together. You can use all Renderman facility in Next with (Shape Class) but the postscript code must be in (View Class (Camera class) ). How can I force the postscript figures/Text to Follow the (Shape). Tanks Nader nazari@cs.chalmers.se
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Re: WebObjects:Can't start DefaultApp from command line Date: 9 Jun 1996 03:19:32 GMT Organization: Digital Fix Development Message-ID: <4pdfs4$d9n@digifix.digifix.com> References: <4paatb$gq5@news.mr.net> In-Reply-To: <4paatb$gq5@news.mr.net> On 06/07/96, Bruce Montegani wrote: >I'm using NextStep 3.3 with WebObjects 1.0 installed. > >The instructions in the documentation say to start the HelloWorld application >manually do the following: > >1. Change to /NextLibrary/WebObjects/Executables directory. > >2. Enter DefaultApp Examples/HelloWorld > >I did this and the message came back that DefaultApp: command not found. > >However when I entered the command ./DefaultApp Examples/HelloWorld the >application started. > >Why is this? > This isn't a WO question, but a basic Unix one. DefaultApp searches the directories you have specified in your shell startup (likely the default) for the app with that name. Since its a potential security risk the directory '.' (the current directory) is not included in the default directories to check for the app. ./DefaultApp tells the shell to look for DefaultApp in the current directory. >Also, even though the application started the terminal window did not return >to the prompt. It just kind of hanged. This is what its supposed to do. You do this for debugging, not necessarily for running an actual WebObjects application in a real environment (although I have!) >When I closed the terminal window the >DefaultApp also ended. How come the command did not put the DefaultApp in the >background?? > Again, this is because Terminal is killing the processes. You'd have to put it in the background when you start it (and even then I'm not sure if it would still run, or why you would want to do it like that) WebObjects will 'just start' if you use the right path.. in the case of the above, it would be http://YOURSERVERHERE/cgi-bin/WebObjects/Examples/HelloWorld -- - Scott Anguish - sanguish@digifix.com (NextMail) next-announce@digifix.com (comp.sys.next.announce submissions) http://www.stepwise.com/ (Stepwise NEXTSTEP/OpenStep Information Server)
From: lin@lorien.umd.edu (Lin) Newsgroups: comp.sys.next.programmer Subject: Did Developer 3.3 forgot something ? Date: 9 Jun 1996 09:14:26 GMT Organization: University of Maryland, College Park Message-ID: <4pe4li$jl7@hecate.umd.edu> After upgrade to NeXTStep 3.3 and Developer 3.3 on a black NeXT box, there seems to be more problem in posix than it has in Developer 3.2. I was trying to compile Lynx 2-5. If I used 'make next', those emphasis words on a web page will not be highlighted. So I have to use 'make next-ncurses'. But this cause the code to use several library routines that only exist in postscript library ( according to the man page of those routines not found when linking ). So I added -posix to the compiler and link flags. This solve most problem but then failed at link time. The routines that's missing is isascii. To test the problem, I wrote this following short C program and compiled it with -posix. Sure enough, it can't find isascii. When compiled without -posix, it has no problem. I guess I can put a macro definition in Lynx code to deal with this. But can someone with the same developer version test this short program out and let me know the result, please ? Thanks. #include <stdio.h> #include <ctype.h> main(int argc, char *argv[]) { int i; for (i=0; i < 128; i++) if (isascii(i) != 0) printf(" %d : ascii\n", i); else printf(" %d : ascii\n", i); }
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: Q:Sound[Kit] & OPENSTEP? Message-ID: <DsqsEK.JqD@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p9d1n$lo@tallowcross.uni-frankfurt.de> Date: Sun, 9 Jun 1996 16:56:44 GMT In article <4p9d1n$lo@tallowcross.uni-frankfurt.de>, Uli Zappe <uli@tallowcross.uni-frankfurt.de> wrote: >It IS better in a (to me) significant way because it can use text colour and >different fonts. Can do that in Mynah, and it sends *only* MIME. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: pascal@localhost (Pascal Thibaudeau) Newsgroups: comp.sys.next.programmer Subject: HELP!!Unable to load FontPanel Date: 9 Jun 1996 20:52:08 GMT Organization: CRIBX1 , Universite de Bordeaux I , France Message-ID: <4pfdho$d78@news.u-bordeaux.fr> Hello, When I want to start the FontPanel, the WM.app crashes. What's wrong ? It seems to me that the FontPanel object which uses the defaults database to remember its size and so one, can't be invoqued. What's wrong with the defaults database and how to rebuild it ? Thanks. -- -Pascal Thibaudeau ********************************************************************** E-Mail: pthibaud@frbdx11.cribx1.u-bordeaux.fr ( NO NEXTMAIL ) pascal@galileo.slip.u-bordeaux.fr ( Small NEXTMAIL OK ) Powered by NEXTSTEP **********************************************************************
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 9 Jun 1996 21:57:47 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pfhcr$e63@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p9d1n$lo@tallowcross.uni-frankfurt.de> <DsqsEK.JqD@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > In article <4p9d1n$lo@tallowcross.uni-frankfurt.de>, Uli Zappe <uli@tallowcross.uni-frankfurt.de> wrote: > >It IS better in a (to me) significant way because it can use text colour and > >different fonts. > Can do that in Mynah, and it sends *only* MIME. Will it be represented properly (all color, line spacing info) when opened in another mail reader other than Mynah? Like Mail.app, or some other MIME capable reader? Or just in Mynah? -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: brian@homer.systemix.com (Brian Cuthie) Newsgroups: comp.sys.next.programmer Subject: Getting the remaining capacity from a DAT drive Date: 9 Jun 1996 19:25:59 -0400 Organization: Systemix Software, Inc. Message-ID: <4pfmi7$c9g@homer.systemix.com> NNTP-Posting-User: unknown Hello. A while ago someone asked about calculating the remaining capacity of a DAT tape once it had been written. I missed the original article and therefore can't post a followup. But, I'll try to answer the question here. Determining the remaining capacity of a DAT tape is done by issuing a special MODE SENSE command to the tape drive. Most modern drives support a "Tape Capacity" sense page. The page number is 0x31 (HEX) and gives the number of remaining blocks on the tape. Since these are tape blocks they represent the remaining capacity of the tape with no compression applied. You can figure the compression ratio of the date already written to the tape and then multiply that value by the remaining uncompressed capacity to get a feel for how much real space is left. SafetyNet always resturns the value for uncompressed results. Hope this helps. Cheers, Brian
Newsgroups: comp.sys.next.programmer From: js@euler.han.de (Juergen Sell) Subject: Re: Checking if slip interface is up Message-ID: <DsqxCs.6Ap@euler.han.de> Sender: js@euler.han.de (Juergen Sell) Organization: Ink Unknown References: <4opait$eqc@lyra.csx.cam.ac.uk> Date: Sun, 9 Jun 1996 18:43:40 GMT Tomaz Slivnik writes > How do I check from a C program if a SLIP interface is up? > Thank you for your replies. Have a look at man pages for ifconfig, if for starters. Hope that helps. Juergen --- Fon +49 511 92455-51 NeXTMail welcome No Mime Fax +49 511 92455-52 = What time do we live in when revolution reminds us of soap powder, = when spontaneity and freedom get associated with instant coffee, = when a politician's idea of social change is changing names = when a country posing as super know-how factory cuts expenses on education?
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 10 Jun 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4pg7ge$nms@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: What is KBNS Date: 10 Jun 1996 06:10:17 GMT Organization: No organisation supplied Message-ID: <4pge89$ms5@hermes.is.co.za> I recently downloaded the Projectbuilder aware Flex and Bison. In the description of the package it states the following: This is the binary distribution of flex, GNU's lex replacement. This version is made ``ProjectBuilder aware'', using an undocumented API as described in KBNS.32.2.029. What is KBNS and where can I find it? -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Mon, 10 Jun 1996 11:07:11 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <31b74471.157975@news.plsys.co.uk> On Thu, 6 Jun 1996, Karl Harbour wrote: Sorry I had to quote a lot, but I would quote out of context otherwise: > cph@next.com (Charles d'Harcourt) wrote: > > >tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: > >> I would do: > >> > >> NSMutableArray kids; > >> > >> -(NSArray *)children > >> { > >> return kids; > >> } > >> > >> This is the simplest and fastest implementation, and anybody talking > >> mutable methods to an NSArray should be fired. --Tiggr > >> > I read what you say carefully and notice that you say "typed as an > instance", not "is an instance of". You were careful to say this, but > of course the fact is that typing an object has no bearing on what > class it really is at run-time, much less whether it's immutable. > > What seems sensible to me is that if a method states it returns an > NSArray, i.e. an immutable class, it _will_ return an immutable array. > Syntatically it's valid to return NSArray's mutable subclass; but I > think methods which return classes which recognise the mutability > concept should be treated as special cases, i.e. the immutability will > be honoured as well as the object really being an instance of the > stated class. > Absolutely correct! In this thread people are thinking about the 'immutable' concept of the FK as a different implementation of the 'const' keyword, which is wrong. Somebody stated already the following string code example: const char * foobar(); char *str = (char *)foobar(); *str='?'; A nice example but completely wrong for the FK implementation: NSString *foobar(); NSMutableString *str = (NSMutableString *)foobar(); [str ...]; // changing content will and _has_to_ fail. Both examples are based on the same programmer's goal, but with different approach. It should be clear: The mutable/immutable concept _is_not_ a replacement for the 'const' keyword. It extends programmers protections. Tiggr's example is wrong! Yes, it is. Although it might be syntactical legal it is wrong in the concept of FK and OpenStep. It should be: NSMutableArray kids; -(const NSMutableArray *)children { return kids; } Best regards Boerny. P.S.: It is nice, but I rarely see people using the 'const' keyword. But with the introduction of mutable/immutable objects they think that an immutable object is the same as a 'const' mutable object --- which is wrong. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: louie@va.pubnix.com (Louis A. Mamakos) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 10 Jun 1996 07:36:38 -0400 Organization: Pubnix Access Systems (Virginia) Message-ID: <4ph1c6$amu@pub02.va.pubnix.com> References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p9d1n$lo@tallowcross.uni-frankfurt.de> In article <4p9d1n$lo@tallowcross.uni-frankfurt.de>, Uli Zappe <uli@tallowcross.uni-frankfurt.de> wrote: >louie@va.pubnix.com (Louis A. Mamakos) wrote: >> NeXTMail format isn't any technologially "better" than MIME is; it's >> just different and proprietary and not significantly better. > >I cannot but repeat it: > >It IS better in a (to me) significant way because it can use text colour and >different fonts. > >Me, I can't live without that, it's absolutely critical for the way I use >E-Mail. Is it so hard to accept that? You could certainly carry wretched NeXT RTF as an alternative content-type in a MIME message. Hopefully, you'd include it as part of a 'multipart/alternative' content type, so that people with other platforms could still read your message. While we can have a discussion about how lame or not the NextMail.app user interface is, the point of this discussion, I thought, was the encoding on the wire and across the network. The reason that MIME is "better" is because it's standard and portable and is on a bunch of different platforms. Even NeXT finally figured this out, as NeXTMail in release 3.3 (I think) can use a MIME wrapper to carry the NeXTMail propretary content. NeXT blew it again here. If you look at the Mail application that they shipped, the "hard part" was the user interface, not the encoding on the wire. They could have had a really wonderful enterprise-wide mail platform solution, but they tossed it all away. For instance, while I was at the University of Maryland, we had a project to get a UNIX platform on the desk of every College dean, President and VP of this and that. The NeXT would have been an ideal solution for this given the quality of it's user interface, but for the fact that these users would have inevitably sent unreadable NeXTMail to users reading mail on IBM MVS, other UNIX, VMS, etc. systems. And they would have not been able to handle MIME messages from those same people containing spreadsheets, etc. So, instead a bunch of DECstation 3100 systems went out there instead, and that's what those people saw *EVERY* day, and not a NeXT. And NeXT could have been the leader in the WWW browser business. Oh well. louie
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: What is KBNS Date: Mon, 10 Jun 1996 17:51:22 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960610174748.4458D-100000@hphalle0.informatik.tu-muenchen.de> References: <4pge89$ms5@hermes.is.co.za> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4pge89$ms5@hermes.is.co.za> On 10 Jun 1996 gvandyk@icon.co.za wrote: > I recently downloaded the Projectbuilder aware Flex and Bison. In the > description of the package it states the following: > > This is the binary distribution of flex, GNU's lex replacement. This > version is made ``ProjectBuilder aware'', using an undocumented API as > described in KBNS.32.2.029. > > What is KBNS and where can I find it? It's always a good choice to search a good ftp site for this: ftp://peanuts.leo.org/pub/comp/platforms/next/Documents/software/KBNS.33.2.* Have fun, Boerny. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: OmniWeb Hierarchical List Date: 10 Jun 1996 15:50:06 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4phg7e$dn1@lion.embl-heidelberg.de> References: <4p9ffp$r1u@dragonfly.wolfram.com> In article <4p9ffp$r1u@dragonfly.wolfram.com> stes@kokomo.wri.com (David Stes) writes: > Georg Tuparev writes > > > > I will finish & submit it whet I get NS4.0. Although officially I was > > even on the NS4.--alpha tester list [flame on !&*)(&*(^%^&$%$%^ flame > > off], I'm still waiting for a copy. If you want to see it sooner, call, > > or mail the NeXT office in Germany and tell them "!@!^&%%#%@^%" > > Here we now have a nice picture of the average NSObject supporter. Just for your info: There is no connection between Object, NSObject and PR/Marketing mismanagement .. -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: NeXTMail is just different Message-ID: <DssKxy.I68@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pccp8$hk@tallowcross.uni-frankfurt.de> Date: Mon, 10 Jun 1996 16:10:45 GMT In article <4pccp8$hk@tallowcross.uni-frankfurt.de>, Uli Zappe <uli@tallowcross.uni-frankfurt.de> wrote: >Sounds interesting. What is Mynah? A MIME Mailer for NeXTSTEP. More info can be found at <http://www.nvc.cc.ca.us/Mynah>. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: NeXTMail is just different Message-ID: <DssL2p.I8u@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> Date: Mon, 10 Jun 1996 16:13:37 GMT In article <4pa321$852@news3.digex.net>, John Kheit <jkheit@cnj.digex.net> wrote: >Anyway, the MIME apps I've seen, and the one I'm using do not support any >postscript the fonts I use, nor doe they remember color and formatting >details anywhere nearly as accurately as RTF(D). It's not my fault the MIME apps you've been using are so limited. >variable line spacing, varying margins and tabs... Since when does the Text class support variable line spacing and variable margins? NeXTMail's capabilities are only the ones the Text class has (and man is the Text class a mess). -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 10 Jun 1996 15:49:45 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4phg6p$nc@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > A communications system is something that allows people to exchange > information. > [...] > Judging by their utility as 'communications systems', MIME is superior > to NeXTmail because there are a couple of orders of magnitude more > people with mail readers that understand MIME messages than understand > NeXTmail. True for internet purposes, wrong for intranet purposes. > NeXTmail is completely unintelligible to non-NeXTmail mail > readers, whereas MIME works pretty well even in non-MIME readers. Hmmm... in former days, when I had no NEXTSTEP computer, I now and then got a NeXTMail. I just uudecoded and untared/unzipped it, and there it was, perfectly readable on my ATARI computer's RTF word processor. I never was able to do that with MIME. Admittedly, this is for tech freaks only. > Uli, you seem to be convinced that you either have to use NeXTmail or > you cannot communicate at all. Obviously, that is nonsense (people have > managed to deal with 7-bit ASCII email for decades!) No, it is plain truth in the context of the project I work for. The people who use NeXTMail there (all non-technicians) would have never used ASCII mail; they'd have faxed or written letters instead because they consciously refuse to adapt to the logic of a machine. > but if you really > want the precise formatting capabilities of NeXTmail, why can't you > simply send an RTF(D) document as an enclosure in a MIME message? Because they'd regard even this as too "technical"; technology does not totally vanish behind the use, and they refuse that. Only because of NeXTMail I was able to install a email system for them. All I want to point out is that use of NeXTMail may well be because of other reasons than nostalgia, and that I can't understand SUN if they don't offer this as an *alternative* (that nobody is forced to use), given that fact that recompilation probably wouldn't make that much of an effort for them. Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: neuss@isa.informatik.th-darmstadt.de.NOSPAM (Christian Neuss) Newsgroups: comp.sys.next.programmer Subject: Re: WebObjects:Can't start DefaultApp from command line Date: 10 Jun 1996 16:09:28 GMT Organization: Fachbereich Informatik, TH Darmstadt, Deutschland Message-ID: <4phhbo$1omm@rs18.hrz.th-darmstadt.de> References: <4paatb$gq5@news.mr.net> <4pcvpd$kga@bignews.shef.ac.uk> mmalcolm crawford (m.crawford@shef.ac.uk) wrote: > On 06/08/96, Bruce Montegani wrote: > > > > However when I entered the command ./DefaultApp Examples/HelloWorld the > > application started. > > > For some reason the current working directory (.) isn't part of your PATH > environment variable. He probably strated DefaultApp as root. When you're root, "." should not be in your PATH for security reasons. Simple use ./DefaultApp instead. Best wishes, Chris -- // Christian Neuss "I ride tandem with a random.." // http://www.informatik.th-darmstadt.de/~neuss/ // +49 6151 16-3414 fax: -5472
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: NeXTMail is just different Message-ID: <DssLqC.IGB@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> Date: Mon, 10 Jun 1996 16:27:48 GMT In article <4pa321$852@news3.digex.net>, John Kheit <jkheit@cnj.digex.net> wrote: > That is the point... MIME<RTF(D) as a format. Doesn't matter > that the MIME format is in a sense converted back into and RTF > text object... As far as MIME being as capable...Is Mynah a NS > app? Regardless, is it using some custom extensions to MIME like > the Netscape folks custom extensions of html? If so, that is not > MIME... Mynah is a NeXTSTEP app. Info can be found at <http://www.nvc.cc.ca.us/Mynah>. Mynah adheres strictly to the MIME related RFCs. > Anyway, the MIME apps I've seen, and the one I'm using do not > support any postscript the fonts I use, nor doe they remember > color and formatting details anywhere nearly as accurately as > RTF(D). Fractional spaces, line sizes, various tab formats, > etc...etc... The Text class is able to encode and decode RTF, but the Text class != RTF. There are a *lot* of things that RTF can do that the Text class cannot. NeXTMail relies on the Text class to do all of its work. A lot of those things you cite as being part of RTF are *not* part of the Text class. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) Newsgroups: comp.sys.next.programmer Subject: mk[s]tmp Bus errors Date: 10 Jun 1996 18:32:18 +0200 Organization: Institut fuer medizinische Informatik und Biometrie Sender: fxg@merry.imib.rwth-aachen.de Message-ID: <s4ywx1f7h7x.fsf@merry.imib.rwth-aachen.de> When running a trivial program like this #include<libc.h> void main(void) { int f = mkstmp("tstXXXXXX"); } mkstmp Bus errors on me. That is, in NeXTstep 3.2 on a NeXTstation, NeXT's compiler. Can anyone verify this? And, since I've noticed that mkstmp is not commonplace in unix programming, what would be the most convenient and portable way to create a temporary file? Thanks -- Felix (fxg@(Pool.Informatik|imib).RWTH-Aachen.de NeXT-Mail ok, but slow)
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 10 Jun 1996 16:45:16 GMT Organization: University of Sheffield, UK Message-ID: <4phjes$i98@bignews.shef.ac.uk> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> In-Reply-To: <DssL2p.I8u@nvc.cc.ca.us> On 06/10/96, Chris Osborn wrote: > Since when does the Text class support variable line spacing and > variable margins? NeXTMail's capabilities are only the ones the Text > class has (and man is the Text class a mess). > Umm, you can do this easily enough in NeXTMail... Use Command-r to show the ruler, and you can set margins how you wish. Use superscript and subscript to alter line spacing. As I just mentioned on the UK list, you can also use the Escape key in the To: (and CC:) field in Mail.app to complete any addresses in your Address list (or on the local domain) -- a number of us hadn't noticed this, even after using the app for years! Best wishes, mmalc. --
From: David Hinz <dhinz@dna406.dna.mci.com> Newsgroups: comp.sys.next.programmer Subject: Q:Am I using NSException correctly? Date: Fri, 07 Jun 1996 15:16:32 -0600 Organization: MCI Message-ID: <31B89C30.52BF@dna406.dna.mci.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I wrote the following test program to learn exceptions but I'm getting an error which says: "Attempt to remove unrecognized exception handler." From the exception handling system I guess. Is this the correct way to use exceptions? Does someone have a good short example of a way to write an exception handler? - (NSNumber *)createDirName:(NSString *)dirName mode:(NSNumber *)dirMode { NSNumber *createResult; NSException *mkdirException; int mkdirResult; const char *directory = [dirName cString]; int mode = [dirMode intValue]; if ([self debug]) { NSLog (@"Input directory is: %s", directory); NSLog (@"Mode is: %o", mode); } NS_DURING mkdirResult = mkdir (directory, mode); /* Successful? */ if (mkdirResult == -1) { // Create the exception info from errno mkdirException = [NSException exceptionWithName:@"mkdir" reason:[NSString stringWithCString: strerror (errno)] userInfo:nil]; // Raise the exception [mkdirException raise]; } if ([self debug]) { NSLog (@"Result is %d", mkdirResult); } createResult = [NSNumber numberWithInt:mkdirResult]; NS_VALRETURN (createResult); // If mkdir failed NS_HANDLER NSLog (@"Exception: %s Reason: %s\n", [[mkdirException exceptionName] cString], [[mkdirException exceptionReason] cString]); createResult = [NSNumber numberWithInt:-1]; NS_VALRETURN (createResult); NS_ENDHANDLER } /* end createDir */ Thanks dave. -- David Hinz E-mail: dhinz@dna406.dna.mci.com
From: Misha Neverov <misha@sccs.swarthmore.edu> Newsgroups: comp.sys.next.programmer Subject: [Q] Updating cursor position Date: Mon, 10 Jun 1996 13:42:51 -0400 Organization: Swarthmore College Message-ID: <31BC5E9B.4B2F@sccs.swarthmore.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit We are looking for a way to update cursor position either from within an application or by a driver that would receive data from a parallel port. Any helpful information would be appreciated. Misha
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 10 Jun 1996 17:52:00 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4phnc0$ltv@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > In article <4pa321$852@news3.digex.net>, > John Kheit <jkheit@cnj.digex.net> wrote: > >Anyway, the MIME apps I've seen, and the one I'm using do not support any > >postscript the fonts I use, nor doe they remember color and formatting > >details anywhere nearly as accurately as RTF(D). > It's not my fault the MIME apps you've been using are so limited. Blame has nothing to do w/ it... If the number of people using MIME capable readers that can do what Mynah can is less than the number of people using NeXTmail, then NeXTmail is more usefull to me. The bottom line is, not enough mail readers can do what NeXTmail can, even if technically MIME can do as much or more than NeXTmail, that standard is not followed. Moreover...I am assured a certain minimum, which is quite high, under NeXTmail in the way of formatting, and the same is not true of MIME readers... Some like Mynah do everything, others do substantially less, I have no guarantee as to how things will be seen on the other end. And that is what makes MIME less usefull; it is a capable standard that is employed more slackly than even html. Technically capable or not, the standard does not widely do what NeXTmail does. > >variable line spacing, varying margins and tabs... > Since when does the Text class support variable line spacing and variable margins? NeXTMail's capabilities are only the ones the Text class has (and man is the Text class a mess). Variable margins...always as near as I can tell...Varible spacing...Well I can't comment on what can or can't be done in 4.0 ;) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 10 Jun 1996 17:55:21 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4phni9$ltv@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssLqC.IGB@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > The Text class is able to encode and decode RTF, but the Text class != RTF. There are a *lot* of things that RTF can do that the Text class cannot. NeXTMail relies on the Text class to do all of its work. True enough... > A lot of those things you cite as being part of RTF are *not* part of the Text class. We obviosly have disparite definitions as to what accounts for alot...regardless, it depends on what version of the text class we are talking about... -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: silvio@jabberwock.swarthmore.edu (Silvio P Eberhardt) Newsgroups: comp.sys.next.programmer Subject: mouse driver source needed Date: 10 Jun 1996 18:02:18 GMT Organization: Swarthmore College, Swarthmore, PA, USA Message-ID: <4phnva$2up@larch.cc.swarthmore.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Keywords: mouse driver Hi- We have connected a Phantom force feedback robotic mechanism to a NS486 system. We would like to use the Phantom as a high-tech mouse. Does anyone out there have the source for a NS mouse driver that we could use as a prototype for writing our own? Thanks. Silvio Eberhardt Department of Engineering Swarthmore College
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Mon, 10 Jun 1996 14:20:30 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> In-Reply-To: <4phg6p$nc@tallowcross.uni-frankfurt.de> Excerpts from netnews.comp.sys.next.advocacy: 10-Jun-96 Re: Q:Sound[Kit] & OPENSTEP? by Uli Zappe@tallowcross.un >> Judging by their utility as 'communications systems', MIME is superior >> to NeXTmail because there are a couple of orders of magnitude more >> people with mail readers that understand MIME messages than understand >> NeXTmail. > > True for internet purposes, wrong for intranet purposes. Er? A factual claim falls into one of three catagories: true, false, or unanswered/unknown; it cannot be both true and false at the same time. Even if you want to consider just intranet usage, there are still orders of magnitude more people using MIME than are using NeXTmail. >> NeXTmail is completely unintelligible to non-NeXTmail mail >> readers, whereas MIME works pretty well even in non-MIME readers. > > Hmmm... in former days, when I had no NEXTSTEP computer, I now and then got > a NeXTMail. I just uudecoded and untared/unzipped it, and there it was, > perfectly readable on my ATARI computer's RTF word processor. I never was > able to do that with MIME. Admittedly, this is for tech freaks only. The NeXTmail message _was_ unintelligible to your mail reader. You had to save the message somewhere (or use the pasteboard), and then manually run a series of programs over the message, so that then you could look through the tar archive and be able to examine the individual pieces of an RTFD document. Even assuming that you don't have a MIME mail reader available, you can still read a compound MIME message composed of text sections and enclosures and be able to read the text, without leaving your ASCII mail reader. And you can still manually decode a MIME message just like you can NeXTmail. >> Uli, you seem to be convinced that you either have to use NeXTmail or >> you cannot communicate at all. Obviously, that is nonsense (people have >> managed to deal with 7-bit ASCII email for decades!) > > No, it is plain truth in the context of the project I work for. The people > who use NeXTMail there (all non-technicians) would have never used ASCII > mail; they'd have faxed or written letters instead because they consciously > refuse to adapt to the logic of a machine. Electronic mail is a means of communicating information, and the precise way that any message is displayed must be under the control (and preferences) of the reader, in much the same way that HTML is a markup language that defines "what things should look like", but not the precise fonts, whitespace, and so forth. If you want to exchange precisely formatted documents like RTFD, PostScript, DVI, TeX, and so forth, then send them as enclosures. [ ... ] > All I want to point out is that use of NeXTMail may well be because of other > reasons than nostalgia, and that I can't understand SUN if they don't offer > this as an *alternative* (that nobody is forced to use), given that fact that > recompilation probably wouldn't make that much of an effort for them. Sure, Sun (presumably) has the source code to NeXT's Mail.app available, so they could certainly provide support for sending NeXTmail without much work at all, _if_ they chose to. However, a decision obviously was made at Sun that MIME mail was the multimedia email format that their users will use. This decision is hardly surprising for a lot of reasons. First, MIME is the accepted communication standard for electronic mail, and it is superior to NeXTmail (which has never been, and probably never will be, a "standard") in every technical aspect that I can think of. Furthermore, MIME provides multimedia layout capabilities that are only slightly inferior to the layout capabilities of NeXTmail. If NeXT had any interest at all in trying to get NeXTmail commonly accepted, they should have documented it thoroughly and written NeXTmail mailreaders for MS-Windows, the Mac, and various Unix/X11 environments. And they should have done this several years ago-- because now, Sun isn't willing to support sending NeXTmail messages even with NeXT helping out with OPENSTEP.... -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: bmont@rconnect.com (Bruce Montegani) Newsgroups: comp.sys.next.programmer Subject: Reading Precompiled Headers Date: 10 Jun 1996 18:38:45 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <4phq3l$bmk@news.mr.net> Does anyone know how to read a precompiled header file? ie. that is the structure? Thanks Bruce
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: NeXTMail is just different Date: Mon, 10 Jun 1996 15:07:49 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <wlj7_5K00iVG44jJ96@andrew.cmu.edu> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> In-Reply-To: <4phnc0$ltv@news3.digex.net> Excerpts from netnews.comp.sys.next.advocacy: 10-Jun-96 Re: NeXTMail is just different by John Kheit@cnj.digex.net > Moreover...I am assured a certain minimum, which is quite high, > under NeXTmail in the way of formatting, and the same is not true of MIME > readers... Some like Mynah do everything, others do substantially less, I > have no guarantee as to how things will be seen on the other end. That's correct. You should not, repeat, should not, expect to know exactly how email you send will look like in the recipient's mail reader. For example: The user might decide to read mail in a big window or a small window. The user will decide which mail headers should be hidden. The user might decide to read their mail on any of a number of different kinds of machines-- a NEXTSTEP box, a Sun running Solaris, in Netscape on a Mac or Windows box, in some native Mac/Windows mail reader, or even remotely logged in somewhere with only the display capabilities of a VT100 terminal. > And that is what makes MIME less usefull; it is a capable standard that is > employed more slackly than even html. Technically capable or not, the > standard does not widely do what NeXTmail does. Those exact reasons are why I believe MIME is considerably _more_ useful than NeXTmail. Any MIME-compliant mail reader is required to comply with the minimum specifications, and is expected to try to comply with with rest of the MIME specification as far as is possible and reasonable given the limitations of the system the mail reader is targetted for. It's perfectly reasonable to deal with MIME email or Usenet news posts in a VT100 terminal because of this, whereas reading NeXTmail under a VT100 terminal is impossible for anyone except an expert, and is a painful and tedious process even for someone who does know how to uudecode, uncompress, untar, and rtf2ascii the message. Someone running a decent MIME reader on a color X-Windows display should be able to get multiple fonts and colors, foreign language support, inline rendering of images, drag-n-drop enclosures, and so forth-- a mail environment that is completely comparable to that provided by NeXTmail and Mail.app, except that MIME does not pretend to substitute for a page-layout system (it does, however, let you send such precisely formatted documents rather easily as enclosures). If you're going to spend the time to create such a complex document that you could meaningfully utilize the difference between MIME's formatting capabilities and NeXTmail's, then I cannot believe that the time spent to double-click a MIME enclosure is significant. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: uhendjx@meaddata.com (Jonathan W. Hendry) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy Date: 10 Jun 1996 19:25:01 GMT Organization: Lexis-Nexis, Dayton OH Message-ID: <4phsqd$l60@mailgate.lexis-nexis.com> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> John Kheit (jkheit@cnj.digex.net) wrote: : fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: : > In article <4pa321$852@news3.digex.net>, : > John Kheit <jkheit@cnj.digex.net> wrote: : > >Anyway, the MIME apps I've seen, and the one I'm using do not support : any : > >postscript the fonts I use, nor doe they remember color and formatting : > >details anywhere nearly as accurately as RTF(D). : > It's not my fault the MIME apps you've been using are so limited. : Blame has nothing to do w/ it... If the number of people using MIME : capable readers that can do what Mynah can is less than the number of : people using NeXTmail, then NeXTmail is more usefull to me. The bottom : line is, not enough mail readers can do what NeXTmail can, even if : technically MIME can do as much or more than NeXTmail, that standard is not : followed. Moreover...I am assured a certain minimum, which is quite high, : under NeXTmail in the way of formatting, and the same is not true of MIME : readers... Some like Mynah do everything, others do substantially less, I : have no guarantee as to how things will be seen on the other end. And that : is what makes MIME less usefull; it is a capable standard that is employed : more slackly than even html. Technically capable or not, the standard does : not widely do what NeXTmail does. Seems like the 'certain minimum' which NeXTMail provides is a garbled, uuencoded message if read by anyone not using NeXTMail (most people). If I use Netscape to MIME-mail a web page to someone with Netscape, that person gets the full web page in their message, including Java applets if present. If the receipient is using something else, like Mail.app, or even elm, they still get something fairly intelligible. The HTML tags would clutter the message, but there'd be useful content in there, unconcealed. Somehow that seems more useful than a bunch of uuencoded data. -- Jonathan W. Hendry Views expressed herein do Steel Driving Software, Inc. not represent those of steeldrv@ix.netcom.com Steel Driving Software, Inc. jon@exnext.com or Lexis-Nexis
From: Mark Anenberg <marka@Eng.Sun.COM> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Mon, 10 Jun 1996 13:37:23 -0700 Organization: Sun Microsystems Inc. Message-ID: <31BC8783.5811@Eng.Sun.COM> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Charles William Swiger wrote: > > Excerpts from netnews.comp.sys.next.advocacy: 7-Jun-96 Re: Q:Sound[Kit] > & OPENSTEP? by Uli Zappe@tallowcross.un > > Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: > >> I think there are some (few) areas where NeXTmail really is better, > >> [...] > >> It'd be nice for nostalgia of NeXTSTEP fans, > >> but nothing else. > > > > What with those who need it as a communication system? > > A communications system is something that allows people to exchange > information. The way multimedia mail is exchanged on the Internet is > called MIME (for 'Multipurpose Internet Mail Extensions') and is > well-documented via RFCs. NeXTmail is not a publicly-documented format, > and it's likely that the only people who will ever use NeXTmail are > NEXTSTEP users. > > Judging by their utility as 'communications systems', MIME is superior > to NeXTmail because there are a couple of orders of magnitude more > people with mail readers that understand MIME messages than understand > NeXTmail. NeXTmail is completely unintelligible to non-NeXTmail mail > readers, whereas MIME works pretty well even in non-MIME readers. > > Uli, you seem to be convinced that you either have to use NeXTmail or > you cannot communicate at all. Obviously, that is nonsense (people have > managed to deal with 7-bit ASCII email for decades!), but if you really > want the precise formatting capabilities of NeXTmail, why can't you > simply send an RTF(D) document as an enclosure in a MIME message? > > -Chuck > > PS: I like NeXTmail more than MIME too. The fact that MIME is the > accepted standard and NeXTmail is going to become extinct is > unfortunate, but you can place the resposibility for this squarely on > NeXT. > > NeXTmail was something that NeXT regarded as proprietary and as a > 'competitive advantage' of NEXTSTEP as an operating system. NeXT > deliberately avoided supporting MIME natively for several years (since > MIME was the alternative in competition to NeXTmail) and NeXT's refusal > to publicly document the NeXTmail format as a proposed Internet standard > via the RFC process ensured that no one else could write their own > NeXTmail-compliant reader. > > NeXT's strategy succeeded, in a way. Read Mark Anenberg's comments again: > "Its not a matter of stopping short; its a matter of a conscious > limiting effort; It doesn't make sense to send email in a proprietary > format when the rest of the world speaks Mime/X.400/X.500. The whole > point is information exchange." > > Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer > ----------------+---------------------+--------------------- > I know you're an optimist if you think I'm a pessimist. thanks for a little sanity Charles. It never ceases to amaze how hard it is for some people to understand to OpenStep is an API. And an implementation of OpenStep doesn't necessarily include a desktop; And a desktop written to the OpenStep API's doesn't have to look or act anything like NEXTSTEP did. It just so happens that Sun's does because we licensed the 3.3 sources. Its a fine balance, wanting to preserve the good things in the Nextstep desktop and change the things that do not make sense in an open system. You don't want to change too much because of the retraining cost. I was involved with the IBM port of Nextstep 1.0/2.0 and now I am here trying make it succeed on Solaris. I have a strong interest in seeing it succeed in the market, any market, and I believe that part of that includes changing the things that have kept it part of small proprietary group. Mail format was one thing; Moving from a the NeXT windowserver notion of windows and remote display via NXHost to X11 based protocol and its notion of remote display is another; moving from netinfo to yp/nis/nis+ ..... my opinions only... -- Mark Anenberg , OpenStep Development Team, SunSoft, Inc. [E,NeXT,Mime]Mail: mark@velodrome.com, http://www.velodrome.com/ Disclaimer: The opinions expressed above are my own and in no way represent those of Sun Microsystems, Inc.
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 10 Jun 1996 21:28:11 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4pi41b$odv@news.xmission.com> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > Er? A factual claim falls into one of three catagories: true, false, or > unanswered/unknown; it cannot be both true and false at the same time. How about this statement: "It is raining." At this very moment it will be true for somebody and not true for somebody else. You're right on the NeXTMail issue, though. I think that MIME is better for Internet and intranet usage for a variety of reasons that have already been discussed. But, nevertheless, most statements are not black and white true or false only. While there are absolute truths, the truth of many things depends upon the individual circumstances. For some people in some situations, NeXTMail is far superior to MIME. There are fewer and fewer cases where that is true, but for some it IS true. For most it is not. > Even if you want to consider just intranet usage, there are still orders > of magnitude more people using MIME than are using NeXTmail. I think the only intranet usage that makes sense for NeXTMail is if everyone in the company is running NEXTSTEP. How many of those are there? I doubt there are very many, if any. For all the other cases, MIME makes a lot more sense. > If you want to exchange precisely formatted documents like RTFD, > PostScript, DVI, TeX, and so forth, then send them as enclosures. And that is the solution. Even with NeXTMail, I often send rtf enclosures. Why? I can then set the width of the document. Can't do that with NeXTMail, MIME, or any other rich mail format... One problem that I had, though, was the result of someone dropping a folder into a MIME message. Each file ended up being encoded separately. If they had tarred up and gzipped the directory first, there would have been no problem. But, because they didn't, the 14M of a very large project tree caused my machine to use up all 150M of swap space and crash when I tried to display the message. I would place the blame, though, on NeXT and not on MIME. Analyzing the dust left behind by it gave me the impression that MIME attachments are handled in a very sub optimal way under NEXTSTEP; analogous to the moronic implementation of WorkSpace chunking. (That one I have hard data: a file chunked across 40 disks requires over 400M swap+/tmp to reassemble.) > If NeXT had any interest at all in trying to get NeXTmail commonly > accepted, they should have documented it thoroughly and written NeXTmail > mailreaders for MS-Windows, the Mac, and various Unix/X11 environments. > And they should have done this several years ago-- because now, Sun > isn't willing to support sending NeXTmail messages even with NeXT > helping out with OPENSTEP.... I agree with this fully. NeXT kept many things closed up for too long, and as a result the industry has rejected them. Now it is far too late. Let's hope that some of the upcoming products will fare better. IMHO there is a lot of technical excellence, but a marked lack of business sense. But, complain as I might, I certainly couldn't do any better in their shoes... so I think it is time that I shut up for now. :-) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 00:18:15 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pie07$dr7@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> Hi Chuck :) Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > Excerpts from netnews.comp.sys.next.advocacy: 10-Jun-96 Re: Q:Sound[Kit] & OPENSTEP? by Uli Zappe@tallowcross.un > >> Judging by their utility as 'communications systems', MIME is superior > >> to NeXTmail because there are a couple of orders of magnitude more > >> people with mail readers that understand MIME messages than understand > >> NeXTmail. > > > > True for internet purposes, wrong for intranet purposes. > Er? A factual claim falls into one of three catagories: true, false, or unanswered/unknown; it cannot be both true and false at the same time. I won't get into this too in depth...but there are other states as well... Regardless, a statement like 'his dog is mean' can be both true and false at the same time as the dog can be mean to some and nice to others... Many 'facts' depend upon the person and perspective of the person doing the factual interpretation. And here we have the same fact having different 'values' based upon one's perspective. i.e. whether we view it from a micro intranet viewpoint or a macro internet view point. Makes sense to me (but that's not saying much ;) > If you want to exchange precisely formatted documents like RTFD, PostScript, DVI, TeX, and so forth, then send them as enclosures. Not as cool...and considering that there is NO technical reason why we shouldn't be able to view things directly in our mail programs, I shouldn't have to do this...and I imagine that within a year or two all the MIME apps will catch up and raise the bar of mail viewing to that of NeXTmail...but that hasn't happened yet (certainly not widely). > This decision is hardly surprising for a lot of reasons. First, MIME is the accepted communication standard for electronic mail, and it is superior to NeXTmail (which has never been, and probably never will be, a "standard") in every technical aspect that I can think of. Furthermore, MIME provides multimedia layout capabilities that are only slightly inferior to the layout capabilities of NeXTmail. Again, just because it is technically superior via spec, doesn't mean that spec is widely used... It's like this... We have HTML1,2,3... If your browser claims to support HTML3 it must support ALL those tags...otherwise it will only be, say, an HTML2 capable browser. With MIME we have definitions to, say, the level of HTML3, lets call it MIME3 for ease. So we have MIME3 levels of technical layout prowess available in the MIME format. Unfortunatly, there are no discrete MIME 1, 2, or 3 _levels_ that mailers can say they comply w/, so the result is each mailer author just junks in whatever they feel is enough from any of the MIME3 feature set (i.e. one or a zillion MIME3 features implemented...either case will allow you to deem your mailer a MIME mailer even though you will be able to see either perfectly or perfectly poorly formatted MIME mail)...And the same MIME message mailed out to different programs is NOT guaranteed to look the same, and currently is quite _likely_ too look MUCH worse. NeXTmail, on the otherhand, has been the same, and is identicle on all NeXTmail apps... Which is pretty easy since there are not that many NeXTmail apps, and all of them use the same Text object :) Regardless...until more MIME mail programs _standardize_ on some minimum level of formatting ability, like say RTF(D) levels or higher...NeXTmail will remain a superior format, in practice though not technically. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 11 Jun 1996 00:26:21 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4piefd$dr7@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> uhendjx@meaddata.com (Jonathan W. Hendry) wrote: > Seems like the 'certain minimum' which NeXTMail provides is a garbled, uuencoded message if read by anyone not using NeXTMail (most people). OK, seems like you didn't read my premise... > If I use Netscape to MIME-mail a web page to someone with Netscape, that person gets the full web page in their message, including Java applets if present. If the receipient is using something else, like Mail.app, or even elm, they still get something fairly intelligible. The HTML tags would clutter the message, but there'd be useful content in there, unconcealed. Somehow that seems more useful than a bunch of uuencoded data. Do you claim that, currently, there are more MIME mail readers that can do the above _AND_ format text as accurately as RTF(D) then there are NeXTmail readers? If that is so, then obviously MIME has now become more useful to the majority of people on this earth... However, the only app I've heard of that does do formatting this good using MIME as its format vehicle is Myhnah (sp?) and that is likely to be a smaller number than NeXTmail app mailers... When this number increases then it will widely be more useful...no doubt...and if the number has already gotten so great, I was caught unaware. Could you please tell me what mail apps can do MIME at an RTF(D) formatting quality other than Myhnah (sp?) ? -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: Q:Sound[Kit] & OPENSTEP? Message-ID: <Dst85p.sB@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4p9d1n$lo@tallowcross.uni-frankfurt.de> <DsqsEK.JqD@nvc.cc.ca.us> <4pfhcr$e63@news3.digex.net> Date: Tue, 11 Jun 1996 00:32:12 GMT In article <4pfhcr$e63@news3.digex.net>, John Kheit <jkheit@cnj.digex.net> wrote: >Will it be represented properly (all color, line spacing info) when opened >in another mail reader other than Mynah? Like Mail.app, or some other MIME >capable reader? Or just in Mynah? Unfortunately, I don't know of any other MIME mail readers that truly support the Text/Enriched content type. Some will filter it back into ASCII, but that's all. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: NeXTMail is just different Date: 11 Jun 1996 00:41:54 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pifci$dr7@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <wlj7_5K00iVG44jJ96@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > Excerpts from netnews.comp.sys.next.advocacy: 10-Jun-96 Re: NeXTMail is just different by John Kheit@cnj.digex.net > > Moreover...I am assured a certain minimum, which is quite high, > > under NeXTmail in the way of formatting, and the same is not true of MIME > > readers... Some like Mynah do everything, others do substantially less, I > > have no guarantee as to how things will be seen on the other end. > That's correct. You should not, repeat, should not, expect to know exactly how email you send will look like in the recipient's mail reader. Although that maybe correct...I _should_, _repeat,_yes_I_should_be_assured that if a person really WANTS to see the document fully formatted, they in fact WILL be able to do so...but right now there is absolutely no way to guarantee this unless the two people are using the exact same mailer, namely Myhnah (sp?). There are no discrete feature sets in MIME ala HTML1,2,or3 to guarantee me that that when both the sender and receiver wish to see very richly formatted text that they indeed will see the same very richly formatted text. Conversly, NeXTmail does give this guarantee. > For example: The user might decide to read mail in a big window or a small window. The user will decide which mail headers should be hidden. > The user might decide to read their mail on any of a number of different kinds of machines-- a NEXTSTEP box, a Sun running Solaris, in Netscape on a Mac or Windows box, in some native Mac/Windows mail reader, or even remotely logged in somewhere with only the display capabilities of a VT100 terminal. True, and for this MIME is great. I love it for that. For that it is superior. But you'll find the reason people, or at least I, love NeXTmail is that when people don't wan't the receiver to see the document any way except as originally presented/formatted, then NeXTmail is the way to go. Sometimes I don't want you to be able to view the thing I'm sending in vt100...and even if I do, I want you to be able to get back to your decent mailer and see it in full blazen formatted glory, guaranteed to be like the original. MIME, currently, generally, and widely doesn't provide such a guarantee. > Those exact reasons are why I believe MIME is considerably _more_ useful than NeXTmail. Any MIME-compliant mail reader is required to comply with the minimum specifications, and is expected to try to comply with with rest of the MIME specification as far as is possible and reasonable given the limitations of the system the mail reader is targetted for. Considerably more useful is a WHOLE DIFFERENT issue, and not one that was originally brought up. If I had to choose one or the other, then I'd agree with you. But we are talking about NeXTmail being very useful and good and people still have a use for it b/c it does provide this very high base line guarantee of formatting that MIME doesn't. > Someone running a decent MIME reader on a color X-Windows display should be able to get multiple fonts and colors, foreign language support, inline rendering of images, drag-n-drop enclosures, and so forth-- a mail environment that is completely comparable to that provided by NeXTmail and Mail.app, except that MIME does not pretend to substitute for a page-layout system (it does, however, let you send such precisely formatted documents rather easily as enclosures). Name the apps other than Myhnah (sp? which is also a NS app...)... That's the problem, no guarantee, and not widely implemented to a high rtf(d)-like quality... > If you're going to spend the time to create such a complex document that you could meaningfully utilize the difference between MIME's formatting capabilities and NeXTmail's, then I cannot believe that the time spent to double-click a MIME enclosure is significant. Listen, the time to do many things may not be significant (although I think it is much quicker to copy/paste a segment from one app to the mailer than it is to copy/paste a segment, launch an editor, open new rtf(d) doc, paste in segment, save doc, drag/drop doc into mailer...)...but when you add up all the little pain in the ass little time penalties, you get the difference akin to the difference between using the windows UI and NeXTSTEP's... Considering the might of today's hardware, and the development tools availble, there is NO GOOD reason why I shouldn't be able to see my mail come in richly formatted and guaranteed to be formatted to x (or rtf(d)) quality... -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 00:46:55 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4piflv$dr7@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4p9d1n$lo@tallowcross.uni-frankfurt.de> <DsqsEK.JqD@nvc.cc.ca.us> <4pfhcr$e63@news3.digex.net> <Dst85p.sB@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > In article <4pfhcr$e63@news3.digex.net>, > John Kheit <jkheit@cnj.digex.net> wrote: > >Will it be represented properly (all color, line spacing info) when opened > >in another mail reader other than Mynah? Like Mail.app, or some other MIME > >capable reader? Or just in Mynah? > Unfortunately, I don't know of any other MIME mail readers that truly support the Text/Enriched content type. Some will filter it back into ASCII, but that's all. Exactly the point... So although MIME on paper maybe superior to NeXTmail...currently NeXTmail is more widely capable of distributing rtf(d) quality of formatting than any MIME mailer, and is, for now, for that reason, a superior format to people that need to send mail guaranteeing an rtf(d) level of formatting will be received and seen by the receiver. I'm guessing that w/in a year or so, that all other MIME readers will catch up to or surpass the rtf(d) level of formatting...but for now, that is not the case. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 9 Jun 1996 16:09:43 GMT Organization: Genoa Software Systems Message-ID: <4pet07$2bq@saturn.genoa.com> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: uli@tallowcross.uni-frankfurt.de In <4p1u9d$1ip@tallowcross.uni-frankfurt.de> Uli Zappe wrote: > ... NeXTMail would be just another *option*, allowing better quality > in communication [compared to MIME] with a selected group of people > (just like colour faxes to those who own colour fax machines). Why do you feel NeXTMAIL is superior to MIME as a transport format? I send and receive both formats using the NEXT mail client, and haven't noticed any difference, except that with MIME I can communicate with people who don't use NEXTSTEP. I agree that MIME capable mail clients vary in their capability, but I don't see how that hurts people who use other mail clients. In particular, I haven't noticed any degrading in quality when using MIME to transport mail between two NEXTSTEP hosts. If the user can't tell the difference, why should the vendor bother to support a non-standard format? You seem to think the user can tell the difference. How? -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: wilkie@cg.tuwien.ac.at (Alexander Wilkie) Newsgroups: comp.sys.next.programmer Subject: Re: Q:Huge object file. Date: 11 Jun 1996 06:38:39 GMT Organization: Vienna University of Technology, Austria Message-ID: <4pj49f$s80@news.tuwien.ac.at> References: <31B7D75E.6D6@fy.chalmers.se> Cc: hajons@fy.chalmers.se Hi newsgroup, hi Hakan! > I have developed a program using Sun's cc compiler, which have worked > fine. When I tried to compile it on my NS machine a problem appeared: > On the Sun the size of the objectfile was 43kb, but on NS it turned out > to be 41MB. > > Does anyone have any idea how this happens? NS objectfiles are normally quite reasonable in size, so it sounds like something is screwed in your setup. Would you mind telling us more about the program you are writing? It's hard to make an educated guess without any data available. ys Alexander Wilkie -- e-mail: wilkie@cg.tuwien.ac.at (NeXTMail preferred, MIME o.k.) www : http://www.cg.tuwien.ac.at/~wilkie/
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: What is KBNS Date: 11 Jun 1996 02:11:41 GMT Organization: Genoa Software Systems Message-ID: <4pikkt$i9@saturn.genoa.com> References: <4pge89$ms5@hermes.is.co.za> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: gvandyk@icon.co.za In <4pge89$ms5@hermes.is.co.za> gvandyk@icon.co.za wrote: > What is KBNS and where can I find it? Known Bugs in NEXTSTEP (I think) It was on the main NEXT related ftp archives last I checked. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 13:38:15 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4pjss7$274@tallowcross.uni-frankfurt.de> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4pet07$2bq@saturn.genoa.com> Alex Blakemore <alex@genoa.com> wrote: > You seem to think the user can tell the difference. How? As now already mentioned several times: Just try to send a precisely formatted mail (including tabs and the like), use different font families, font sizes and different text colours, send the mail to yourself via MIME and look at the result... Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: pete@ohm.york.ac.uk (-bat.) Newsgroups: comp.sys.next.programmer Subject: Re: Q:Huge object file. Date: 11 Jun 1996 13:13:09 GMT Organization: The University of York, UK Message-ID: <4pjrd5$d0j@netty.york.ac.uk> References: <4pj49f$s80@news.tuwien.ac.at> > I have developed a program using Sun's cc compiler, which have worked > fine. When I tried to compile it on my NS machine a problem appeared: > On the Sun the size of the objectfile was 43kb, but on NS it turned out > to be 41MB. I'd make a shrewd guess that you've got loads of very big globally declared uninitialised arrays of floats. These used to go into the object file (for reasons best known to the compiler) I can't reproduce this bug under 3.3 though so it's probably been fixed. Are you using NS3.2 by any chance ? -bat.
From: fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) Newsgroups: comp.sys.next.programmer Subject: Re: mk[s]tmp Bus errors Date: 11 Jun 1996 16:51:57 +0200 Organization: Institut fuer medizinische Informatik und Biometrie Sender: fxg@merry.imib.rwth-aachen.de Message-ID: <s4ypw76tmur.fsf@merry.imib.rwth-aachen.de> References: <s4ywx1f7h7x.fsf@merry.imib.rwth-aachen.de> In-reply-to: fxg@imib.rwth-aachen.de's message of 10 Jun 1996 18:32:18 +0200 To: fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) [Bus error Solved] It has been pointed out to me via email (Silly me to delete it too fast to be able to give the due credit here) that I was trying to write over constant data. In article <s4ywx1f7h7x.fsf@merry.imib.rwth-aachen.de> fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) writes: From: fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) Newsgroups: comp.sys.next.programmer Date: 10 Jun 1996 18:32:18 +0200 #include<libc.h> void main(void) { int f = mkstmp("tstXXXXXX"); "tstXXXXXX" is a constant string, a pointer to it is passed to mkstmp which tries to overwrite it => bust. Better: char myTmpName[] = "tstXXXXXX"; This creates an auto char array of length 10 (computed at compile time), initialized with the value ['t','s',...,'X','\0']. This copy is writable. int f = mkstmp(myTmpName); A pointer to this writable structure is passed => bingo. } And, since I've noticed that mkstmp is not commonplace in unix programming, what would be the most convenient and portable way to create a temporary file? still remains unsolved -- Felix (fxg@(Pool.Informatik|imib).RWTH-Aachen.de NeXT-Mail ok, but slow)
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer Subject: Re: Text and textDidChange: Date: 11 Jun 1996 16:09:31 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4pk5nr$k0i@solaris.cc.vt.edu> References: <4p022l$49q@solaris.cc.vt.edu> laurent@planon.qc.ca (Laurent Daudelin) wrote: > > Chuck, > > There is the less known Text delegate method 'text:textObject > isEmpty:(BOOL)empty' that gets called whenever the content changes. As you > recall, whenever a TextField is selected for edition, the window text > field editor object is set with the attributes of the text field, then > editing can begin. The text field that's get represented by the text > object becomes the text object delegate. If an object is made the text > field delegate, any delegate messages sent by the text object to the text > field will be fowarded to the text field delegate. This is how I can know > whenever the text field content changes. > > Maybe a bit difficult to understand, but give it a try, who knows... > > - Laurent No, I've already tried that. If I double click a word, and hit Command b for bold that delegation message doesn't get sent. What I finally had to do was subclass Text and override those messages that change the formatting. Then in my main.m I do a poseAs:. By trial and error I'll discover if I forgot to override a particular method or overrode one that shouldn't have been. It would have been nice if all formatting changes went through a small set of well defined methods. This might be the case, but the docs don't seem to say so explicitly. I guess I'll find out the old fashion way -- printf()! -Chuck
From: neuss@isa.informatik.th-darmstadt.de.NOSPAM (Christian Neuss) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy Date: 11 Jun 1996 16:35:53 GMT Organization: Fachbereich Informatik, TH Darmstadt, Deutschland Message-ID: <4pk799$hsb@rs18.hrz.th-darmstadt.de> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> Chris Osborn (fozztexx@nvc.cc.ca.us) wrote: > Since when does the Text class support variable line spacing and > variable margins? NeXTMail's capabilities are only the ones the Text > class has (and man is the Text class a mess). Amen. (sorry, couldn't resist ;-) Chris -- // Christian Neuss "I ride tandem with a random.." // http://www.informatik.th-darmstadt.de/~neuss/ // +49 6151 16-3414 fax: -5472
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: NeXTMail is just different Message-ID: <DsuGsq.I2A@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> Date: Tue, 11 Jun 1996 16:36:26 GMT In article <4phnc0$ltv@news3.digex.net>, John Kheit <jkheit@cnj.digex.net> wrote: >Blame has nothing to do w/ it... If the number of people using MIME >capable readers that can do what Mynah can is less than the number of >people using NeXTmail, then NeXTmail is more usefull to me. The bottom >line is, not enough mail readers can do what NeXTmail can, even if >technically MIME can do as much or more than NeXTmail, that standard is not >followed. By that same argument though, you can use Mynah. NeXTMail is restricted to the NeXTSTEP platform, as is Mynah. The people who you are sending NeXTMail to could get Mynah instead, and they would then be able to see the same things you have been doing with NeXTMail. The difference would be that the people who *don't* have Mynah would still be able to view your messages. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: Tue, 11 Jun 1996 14:15:22 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <EljPSuG00iUxQ1gOx9@andrew.cmu.edu> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> <4pi41b$odv@news.xmission.com> In-Reply-To: <4pi41b$odv@news.xmission.com> Excerpts from netnews.comp.sys.next.advocacy: 10-Jun-96 Re: Q:Sound[Kit] & OPENSTEP? by don@misckit.com > Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: >>>> Judging by their utility as 'communications systems', MIME is superior >>>> to NeXTmail because there are a couple of orders of magnitude more >>>> people with mail readers that understand MIME messages than understand >>>> NeXTmail. >>> >>> True for internet purposes, wrong for intranet purposes. >> >> Er? A factual claim falls into one of three catagories: true, false, or >> unanswered/unknown; it cannot be both true and false at the same time. > > How about this statement: > > "It is raining." > > At this very moment it will be true for somebody and not true for > somebody else. ...and John Kheit <jkheit@cnj.digex.net> wrote: > I won't get into this too in depth...but there are other states as well... > Regardless, a statement like 'his dog is mean' can be both true and false > at the same time as the dog can be mean to some and nice to others... Many > 'facts' depend upon the person and perspective of the person doing the > factual interpretation. And here we have the same fact having different > 'values' based upon one's perspective. i.e. whether we view it from a > micro intranet viewpoint or a macro internet view point. Makes sense to me > (but that's not saying much ;) A fully-qualified proposition (one where everyone agrees what all of the terms refer to) can be tested and determined to be true or false; if it isn't possible to test the proposition for some reason, it is unanswered. For example, the statement "it is raining" I know to be true, because it's raining here in Pittsburgh-- off and on, anyway. But it would still be true if the sky was clear where I am but it was raining in London or anywhere else. (*) If you want to consider, "is it raining here?", than you've got a relative proposition (because 'here' is a relative term, one whose meaning depends on the context), and the answer would be 'unknown' until everyone who's considering the question agrees on what 'here' means. Most people would interpret 'here' differently, if they are in different places, and the single proposition would turn into many: "is it raining in Pittsburgh?", "is it raining in London?", and so forth-- each of which can be answered definitively because each is fully-qualified. Likewise, John's statement "his dog is mean" is a relative statement because several terms are unqualified ('his', 'dog', and 'mean') and the truth of that statement is 'unknown'. But, if you assign a specific definition to each term, you can then try to determine whether the now fully-qualified statement is true or false. My statement about MIME and NeXTmail above is fully-qualified (I believe), and is pretty obviously true. It remains a true statement regardless of what other propositions you consider. If you consider the issue of MIME, NeXTmail and Intranet usage, that is a different proposition which has it's own answer that is (mostly) independent of the statement that I made. ----------- (*) There is quite a difference between the colloquial meaning of "is it raining?" and the logical meaning-- most people actually mean, "is it raining here?" when they ask you the question "is it raining?", but, obviously, those two questions are not the same. [ ... ] > But, nevertheless, most statements are not black and white true or > false only. While there are absolute truths, the truth of many > things depends upon the individual circumstances. I surely don't have any relation to "absolute truth"! :-) But I will say that you can look at specific statements or propositions and resolve them for a particular set of circumstances. That's good enough.... > For some people in some situations, NeXTMail is far superior to MIME. > There are fewer and fewer cases where that is true, but for some it IS true. > For most it is not. Certainly, at least if you are judging 'superior' according to the perspective of the users. But the users don't care whether their email goes through sendmail and the network, or gets moved from machine to machine by gremlins carrying floppy disks. If you discount the display aspect (in which I agree that NeXTmail is more capable than MIME of precise formatting), and consider technical issues like the encoding mechanisms used (uuencode vs base-64), or the size of the messages (every try to send a compressed/gziped file via NeXTmail?), MIME is superior to NeXTmail. [ ... ] > One problem that I had, though, was the result of someone dropping > a folder into a MIME message. Each file ended up being encoded > separately. Oooo! That's not right! > If they had tarred up and gzipped the directory first, > there would have been no problem. But, because they didn't, the > 14M of a very large project tree caused my machine to use up all > 150M of swap space and crash when I tried to display the message. > I would place the blame, though, on NeXT and not on MIME. I would too-- NeXT's implementation of MIME, at least in this aspect, leaves much to be desired. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: Q:Am I using NSException correctly? Date: 11 Jun 1996 18:35:31 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4pke9j$13f@news.next.com> References: <31B89C30.52BF@dna406.dna.mci.com> In article <31B89C30.52BF@dna406.dna.mci.com> David Hinz <dhinz@dna406.dna.mci.com> writes: > I wrote the following test program to learn exceptions but I'm getting > an error which says: "Attempt to remove unrecognized exception handler." > [...] > // If mkdir failed > NS_HANDLER > NSLog (@"Exception: %s Reason: %s\n", > [[mkdirException exceptionName] cString], > [[mkdirException exceptionReason] cString]); > > createResult = [NSNumber numberWithInt:-1]; > NS_VALRETURN (createResult); > > NS_ENDHANDLER Don't use NS_VALRETURN() from within the handler (or handling) region, use it only within the handled region (which you do correctly). Just use "return" in the handler. NS_HANDLER, NS_VALRETURN(), NS_VOIDRETURN and raising an exception all terminate a handled region, only one of which should be executed to do so. When the handling region is entered after raising an exception, the handled region has been exited, so using NS_VALRETURN() in the handling region is attempting to terminate an exception handled region which has already been terminated (and hence the message you get). Christopher Kane Software Engineer, NeXT Software, Inc.
From: uhendjx@meaddata.com (Jonathan W. Hendry) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy Date: 11 Jun 1996 19:09:50 GMT Organization: Lexis-Nexis, Dayton OH Message-ID: <4pkg9u$9l5@mailgate.lexis-nexis.com> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> John Kheit (jkheit@cnj.digex.net) wrote: : uhendjx@meaddata.com (Jonathan W. Hendry) wrote: : > Seems like the 'certain minimum' which NeXTMail provides is a garbled, : uuencoded message if read by anyone not using NeXTMail (most people). : OK, seems like you didn't read my premise... : > If I use Netscape to MIME-mail a web page to someone with Netscape, that : person gets the full web page in their message, including Java applets if : present. If the receipient is using something else, like Mail.app, or even : elm, they still get something fairly intelligible. The HTML tags would : clutter the message, but there'd be useful content in there, unconcealed. : Somehow that seems more useful than a bunch of uuencoded data. : Do you claim that, currently, there are more MIME mail readers that can do : the above _AND_ format text as accurately as RTF(D) then there are NeXTmail : readers? If that is so, then obviously MIME has now become more useful to : the majority of people on this earth... However, the only app I've heard : of that does do formatting this good using MIME as its format vehicle is : Myhnah (sp?) and that is likely to be a smaller number than NeXTmail app : mailers... When this number increases then it will widely be more : useful...no doubt...and if the number has already gotten so great, I was : caught unaware. Could you please tell me what mail apps can do MIME at an : RTF(D) formatting quality other than Myhnah (sp?) ? I've no idea, frankly, but that's not the point. The point is that a NeXTMail message is nearly useless if sent to someone without NeXTMail. It's *completely* useless if the receipient has no access to uudecode and compress software (most Windows and Mac users). At least with a MIME message, useful content is available on platforms that do not support every nifty feature. Even if an attachment is not usable, text in the rest of the body should still be readable. (Unlike NeXTMail, which garbles the entire message.) If your message's appearance is so important, you ought to send it as an attached Acrobat file. After all, there's no guarantee the receiver will have your fonts, which would mess up the layout anyway. -- Jonathan W. Hendry Views expressed herein do Steel Driving Software, Inc. not represent those of steeldrv@ix.netcom.com Steel Driving Software, Inc. jon@exnext.com or Lexis-Nexis
From: bmont@rconnect.com (Bruce Montegani) Newsgroups: comp.sys.next.programmer Subject: How to trace ObjC messages Date: 11 Jun 1996 20:53:23 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <4pkmc3$p47@news.mr.net> Does anyone know how to trace all messages that are executed when an application is running including API kit messages. i think this would be very helpful in determining the message sequencing of an application especially with the NeXT kits. Thanks Bruce Montegani
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 21:26:57 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pkob1$83t@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4pet07$2bq@saturn.genoa.com> Alex Blakemore <alex@genoa.com> wrote: > In <4p1u9d$1ip@tallowcross.uni-frankfurt.de> Uli Zappe wrote: > how that hurts people who use other mail clients. In particular, I haven't noticed > any degrading in quality when using MIME to transport mail between two NEXTSTEP hosts. > If the user can't tell the difference, why should the vendor bother to support a non-standard format? > You seem to think the user can tell the difference. How? Then you are not making documents with varying margins, tabs, fonts, and colors... Trust me, there is a degredation. For the avarge email it might not matter, but when you need the email to be relatively heavily formatted, MIME pukes on most mailers... -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 21:34:27 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pkop3$83t@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> <4pi41b$odv@news.xmission.com> don@misckit.com wrote: > At this very moment it will be true for somebody and not true for somebody else. You're right on the NeXTMail issue, though. I think that MIME is better for Internet and intranet usage for a variety of reasons that have already been discussed. But, nevertheless, most statements are not black and white true or false only. While there are absolute truths, the truth of many things depends upon the individual circumstances. For some people in some situations, NeXTMail is far superior to MIME. There are fewer and fewer cases where that is true, but for some it IS true. For most it is not. Wow, would you please tell me any/all the absolute truths you know. I will be VERY VERY VERY grateful! (BTW, agree w/ your description above..for some there just is no replacement of NeXTmail...for now until MIME mailers move up in ability...but generally speaking, MIME is just fine [wow a rhyme :]) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 11 Jun 1996 21:43:12 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pkp9g$83t@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <DsuGsq.I2A@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > In article <4phnc0$ltv@news3.digex.net>, > John Kheit <jkheit@cnj.digex.net> wrote: > >Blame has nothing to do w/ it... If the number of people using MIME > >capable readers that can do what Mynah can is less than the number of > >people using NeXTmail, then NeXTmail is more usefull to me. The bottom > >line is, not enough mail readers can do what NeXTmail can, even if > >technically MIME can do as much or more than NeXTmail, that standard is not > >followed. > By that same argument though, you can use Mynah. NeXTMail is restricted to the NeXTSTEP platform, as is Mynah. The people who you are sending NeXTMail to could get Mynah instead, and they would then be able to see the same things you have been doing with NeXTMail. > The difference would be that the people who *don't* have Mynah would still be able to view your messages. Not quite...then I'm working w/ an even smaller universe than those mail readers that can read NeXTmail (mainly Mail.app, and what a small universe that is!!!). The only people that would be able to see my text properly formatted as I _wish_ them to see, would be those with Mynah...and even smaller group. So again, until more apps raise the bar and can view/send MIME significantly richly like Mynah, and IF I need for the receiver to view the mail richly (notwithstanding the many great reasons, all of which I agree completely with, that is cool that MIME can downgrade when needed to the viewer used), NeXTmail is the only format that guarantees me that the receiver will be able to view the document as intended...That is currently its advantage over MIME--and that advantage will fade as more Mynah like MIME viewers become common. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Tue, 11 Jun 1996 21:40:38 GMT Organization: Orbital Computer Consultancy Message-ID: <31bde709.1476496@news.plsys.co.uk> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) wrote: <-- snip --> > >Next's production code seems to vindicate this to my mind; to the best >of my knowledge, all the EOF methods ([EOEntity attributes], for >example) really do return an immutable array. Actually, having checked this now, it isn't true. Methods like (NSArray *)[EOEntity attributes] actually return an NSMutableArray subclass. Apologies. Whether it's "right" for such methods to work this way is a moot point, at least to me.
From: james@ali.bc.ca (James Nakashima) Newsgroups: comp.sys.next.programmer Subject: Help with CVS on NeXT please Date: 11 Jun 1996 22:00:17 GMT Organization: ALI Message-ID: <4pkq9h$fjh@cetus.ali.bc.ca> Hi, I'm trying to use CVS on NeXT and I'm having a lot of problems with wrappers. I have version 1.8.1 as well as a changed version 1.3 for NeXTSTEP. With version 1.3, and the example scripts gunzipuntar and gtarzip I run into the problem of losing some files out of .rtfd files. No small problem. With version 1.8.1 I can input .rtfd files and pull them out no problem, but when I try to change them and put them back in again I get a message saying that a diff failed or something and it doesn't work. Please, any help would be most appreciated. Thanks, Jim -- ________________________________________________________ Jim Nakashima Voice: (604) 279-5422 Ext 369 ALI Technologies Inc. Fax: (604) 279-5468 Richmond, BC Email: james@ali.bc.ca
From: james@ali.bc.ca (James Nakashima) Newsgroups: comp.sys.next.programmer Subject: Help with CVS on NeXTSTEP please. Date: 11 Jun 1996 22:01:00 GMT Organization: ALI Message-ID: <4pkqas$fji@cetus.ali.bc.ca> Hi, I'm trying to use CVS on NeXT and I'm having a lot of problems with wrappers. I have version 1.8.1 as well as a changed version 1.3 for NeXTSTEP. With version 1.3, and the example scripts gunzipuntar and gtarzip I run into the problem of losing some files out of .rtfd files. No small problem. With version 1.8.1 I can input .rtfd files and pull them out no problem, but when I try to change them and put them back in again I get a message saying that a diff failed or something and it doesn't work. Please, any help would be most appreciated. Thanks, Jim -- ________________________________________________________ Jim Nakashima Voice: (604) 279-5422 Ext 369 ALI Technologies Inc. Fax: (604) 279-5468 Richmond, BC Email: james@ali.bc.ca
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 22:17:53 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pkrah$83t@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> <4pi41b$odv@news.xmission.com> <EljPSuG00iUxQ1gOx9@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > A fully-qualified proposition (one where everyone agrees what all of the terms refer to) can be tested and determined to be true or false; if it isn't possible to test the proposition for some reason, it is unanswered. The above is likely to be true only if and when you quantify things to be stochastic and discrete, or better yet mutually exclusive. i.e. if something is not good, it must be bad. This is nice an academic modeling tool, but things in life and reality are not so nicely discrete, and so the above tends not to be so cut and dried. There is truth, there is falsity, there is non truth, non falsity and many gradations in between; non falsity is not necessarily truth, and non truth is not necessarily falsity... > For example, the statement "it is raining" I know to be true, because it's raining here in Pittsburgh-- off and on, anyway. But it would still be true if the sky was clear where I am but it was raining in London or anywhere else. (*) Ok, now If I were in Pittsburgh, and I was the size of an atom or smaller, perhaps where I stood no rain would fall on me, and to me it wouldn't be raining... Or if I were in a plane above Pittsburgh it wouldn't be raining..._FOR ME_ that (it raining) would be false...for _you_ it would be true... So when we discuss things, the viewpoint of the speaker makes a difference, it is difficult to say things are absolutely either true or false for everyone...and so, the question of MIME or NeXTmail depends upon what functions/features you find necessary when describing which is superior... > If you want to consider, "is it raining here?", than you've got a relative proposition (because 'here' is a relative term, one whose meaning depends on the context), and the answer would be 'unknown' until everyone who's considering the question agrees on what 'here' means. Most people would interpret 'here' differently, if they are in different places, and the single proposition would turn into many: "is it raining in Pittsburgh?", "is it raining in London?", and so forth-- each of which can be answered definitively because each is fully-qualified. Again, no matter how specific you get, there is always a counter example...the counter examples become less and less likely, but always remain possible (this is from a pure knowledge/ontological/truth type perspective)... So no matter how discrete you try and make things, I will always show you a counter example as I did above re the Pittsburgh rain hypothetical you posed... Now, when you do get more and more specific, my counters will get more and more insane and less probable counters, though not completely impossible...ergo, you won't be able to say something is absolutely true... Anyway, this gets deeper into philosophy than I'm sure most care to get :) And I'm certain I've bored everyone into coma already %) > Likewise, John's statement "his dog is mean" is a relative statement because several terms are unqualified ('his', 'dog', and 'mean') and the truth of that statement is 'unknown'. But, if you assign a specific definition to each term, you can then try to determine whether the now fully-qualified statement is true or false. You cannot fully make any example absolutely discrete...I ask you to try and I'll still provide you w/ an example to make the opposite result. Lets try some more specificity...John's dog is mean to Harry Anderson of night court. True, when Harry comes over, the dog can bark and hump his leg...Something Harry dislikes and finds mean, and yet something the dog thoroughly enjoys :) So where's the truth to the meaness, in Harry's or the dogs interpretation :) > My statement about MIME and NeXTmail above is fully-qualified (I believe), and is pretty obviously true. It remains a true statement regardless of what other propositions you consider. If you consider the issue of MIME, NeXTmail and Intranet usage, that is a different proposition which has it's own answer that is (mostly) independent of the statement that I made. No it doesn't (remain a true statement regardless...). If I send a highly formatted document I wish you the receiver to view EXACTLY as I sent it, and do it via MIME, it will be worse than if I sent you it via NeXTmail (assuming you can read either MIME or NeXTmail, and both you and I don't have Myahn). So in that situation MIME is inferior...and for now, that situation exists for some folks... So your statement is merely true for you, and likely a majority of others that don't think a guarantee of the receiver being able to see a highly formatted document exactly as sent as being valuable, yet false for those of us who do. > > But, nevertheless, most statements are not black and white true or > > false only. While there are absolute truths, the truth of many > > things depends upon the individual circumstances. > I surely don't have any relation to "absolute truth"! :-) But I will say that you can look at specific statements or propositions and resolve them for a particular set of circumstances. No, what you are doing here is resolving what is true and specific for _you_ for a particular set of circumstances...which is exactly what we all do, but that leads oftimes to disparate results...which is fine and dandy... But something that is true for you, while the opposite is true for me, can hardly unequivocally be stated to be the truth for all. > That's good enough.... Right, good enough for you, but not a panacea for all... > But the users don't care whether their email goes through sendmail and the network, or gets moved from machine to machine by gremlins carrying floppy disks. If you discount the display aspect (in which I agree that NeXTmail is more capable than MIME of precise formatting), and consider technical issues like the encoding mechanisms used (uuencode vs base-64), or the size of the messages (every try to send a compressed/gziped file via NeXTmail?), MIME is superior to NeXTmail. I completely agree...NeXTmail's superiority is not being judged, at least not by me, by its technical transfer mechanisms, which I agree w/ you MIME is superior. But rather for some, the biggest criterion is the display quality of the received mail...for those folks NeXTmail is still a superior format... That's all. How nice and rare when everyone has a nice chat and all come to agreement in the end; especially in an advocacy group !-) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 11 Jun 1996 22:29:50 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pks0u$83t@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> <4pkg9u$9l5@mailgate.lexis-nexis.com> uhendjx@meaddata.com (Jonathan W. Hendry) wrote: > I've no idea, frankly, but that's not the point. The point is that a NeXTMail message is nearly useless if sent to someone without NeXTMail. It's *completely* useless if the receipient has no access to uudecode and compress software (most Windows and Mac users). I disagree, that is the point for me judging it's usefulness. I agree with you 100% that NeXTmail is all but useless unless viewed on a mailer supporting it... But if what _I_ _NEED_ is for the receiver to get mail that is highly formatted and view it accurately, MIME that degrades when received is also all but useless to me... Is it just me, is this point I'm making so stupid that noone sees how NeXTmail's higher standard guarantee can be very important for some people and tasks? > At least with a MIME message, useful content is available on platforms that do not support every nifty feature. Even if an attachment is not usable, text in the rest of the body should still be readable. (Unlike NeXTMail, which garbles the entire message.) Gosh, I've said that is great a zillion times to this feature, but if that is NOT the feature I need if what I do need IS a guarantee that the receiver will NOT see a less formatted version... I.E. when I would prefer the receiver to either see nothing, or highly formatted text accurately, MIME's downgrade feature is actually a handicap...When might this happen??? When for example sending my resume... When viewed properly under NeXTmail it looks great and professional, if viewed in MIME degraded format, or anything less than the NeXTmail standard it looks like I'm an unprofessional idiot. So in that circumstance, MIME don't do me any good. > If your message's appearance is so important, you ought to send it as an attached Acrobat file. After all, there's no guarantee the receiver will have your fonts, which would mess up the layout anyway. Why should I do what you think I ought to do? I think mailers ought to be able to do a simple thing like properly display a simple rtf(d) document when that is what I want. Sometimes I do send attachments, but other times what I WANT is for the message to be seen in the mailer itself. Why is this so strange a thing to people? Am I the only one that finds this useful???? So am I to guess, none of you would ever care for that ability to be available in your MIME mailers? -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
Newsgroups: comp.sys.next.programmer From: paul@cyantic.com (Paul Guiness) Subject: re: HELP: EOF without UI, code example? Organization: CYANTIC Systems Date: Mon, 10 Jun 1996 13:05:34 GMT Message-ID: <1996Jun10.130534.25080@cyantic.com> Although the EOF examples which come with EOF do contain examples on how to do programatic attachments to tables, the following code snippet might give you some ideas. Note that this was extracted from a class method, so I hope I included all the relevant information. - attachToTable: (NSString *) name { /* A similiar example found in SHLExamples */ EODatabaseDataSource* datasource; EOController * myController; EOEntity *entity; NSString *modelPath; EOModel *oldModel; // Get the path of a known model. modelPath = [EOModel findPathForModelNamed:@"test"]; // Create the known model. oldModel = [[EOModel alloc] initWithContentsOfFile:modelPath]; // Get adaptor information from the known model. connectionDictionary = [oldModel connectionDictionary]; // Creating the oldModel entity for the new table. entity = [[oldModel entityNamed:name] retain]; // Create a datasource. datasource = [[[EODatabaseDataSource alloc] initWithModelName:(NSString *)[oldModel name] entityName:(NSString *)[entity name]] retain]; // Create a myController. myController = [[[EOController alloc] initWithDataSource: datasource] retain]; ... } Hope this helps. Paul Guinness paul@cyantic.com
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 20:57:18 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4pkmje$dru@usenet.rpi.edu> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4pet07$2bq@saturn.genoa.com> Alex Blakemore <alex@genoa.com> wrote: > If the user can't tell the difference, why should the vendor > bother to support a non-standard format? > > You seem to think the user can tell the difference. How? I can almost always tell the difference between messages that were sent in NeXTmail and those sent in MIME. The NeXTmail ones I get generally look like the sender intended, as long as I have the fonts the sender used (most users stick with the standard fonts, and I have a lot of other fonts in addition to those so I'm usually OK). MIME messages often look like a mess. Why? Probably because the default fonts I have for Mail.app are different than the normal defaults. Perhaps that's the fault of some MIME implementation details in Mail.app. In any case, it really does come out a mess. The font is much larger than the sender probably intended, and the line-breaks are all wrong for the font size (making it all the more obvious that the font size is wrong). I'm not saying that this proves that Sun should include NeXTmail support in their product, but I am saying that I do think NeXTmail does a better job at formatting than MIME does (mainly because it can depend on postscript font definitions). I'd also say that I think we've beaten this dead horse long enough. I can't imagine that a half-dozen people vocally supporting NeXTmail in the NeXT newsgroups are going to influence much of anyone. It doesn't even influence me that Sun should bother with the NeXTmail format, and I agree that NeXTmail does better than MIME for some things. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: mcgredo@crl.com (Donald R. McGregor) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 11 Jun 1996 16:49:59 -0700 Organization: Miskatonic University Department of Classics Message-ID: <4pl0n7$n51@crl.crl.com> References: <4n0bss$s58@news4.digex.net> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4pet07$2bq@saturn.genoa.com> <4pjss7$274@tallowcross.uni-frankfurt.de> In article <4pjss7$274@tallowcross.uni-frankfurt.de>, Uli Zappe <uli@tallowcross.uni-frankfurt.de> wrote: :>> You seem to think the user can tell the difference. How? :> :>As now already mentioned several times: :> :>Just try to send a precisely formatted mail (including tabs and the like), :>use different font families, font sizes and different text colours, send the :>mail to yourself via MIME and look at the result... That's an argument for a MIME type for rtf, or perhaps sending out the document in adobe acrobat, not an argument for NeXTMail. The NeXTMail format is just some simple tweaks (tar, compress, uuencode) on a collection arbitary file types. There happens to be a built-in rtf reader in the NeXTMail app called the Text Object. If you wanted to send out precisely formatted mail in an interoperable format, you have to decide on what filetype you're going to use. Then you have to write a reader for that filetype on every platform. The NeXTMail format does nothing towards getting you a reader for whatever filetype you decide upon. You can just as easily (in fact probably more reliably) send the file in MIME format, then read it in your rtf or pdf viewer. So you can: 1) write a viewer for rtf, or similar capability, files; or 2) write the above PLUS, try to change the rest of the world's choice for mail transport formats. (2) is both not bloody likely and a Bad Idea. -- Don McGregor | The wages of sin is death--but after taxes it's mcgredo@crl.com | only a tired feeling at the end of the day.
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: How to trace ObjC messages Date: 12 Jun 1996 00:48:04 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4pl444$3v@news.onramp.net> References: <4pkmc3$p47@news.mr.net> (Bruce Montegani) writes: > Does anyone know how to trace all messages that are executed when an > application is running including API kit messages. > > i think this would be very helpful in determining the message sequencing of > an application especially with the NeXT kits. When you compile, there's a debug option you can use and you can mark which classes you want to see message invocations in. Check out the gdb docs. -- Steve Dekorte "C++/Java, the COBOL of the 90s" NeXTstep consultant - Anaheim, CA mailto:dekorte@suite.com (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: OpenStep and EOF Date: 12 Jun 1996 04:33:12 GMT Organization: E.S. Systems cc (Financial Systems Development) Message-ID: <4plha8$177@hermes.is.co.za> I recently visited the gnustep web site to see what the progress was with the project, and I was amazed about the amount of work they have already done. The reason I visited the site was to see how far they were, as we have a potential client that wants to run on OS/2 and the initial goal of gnustep, that I saw, was to get OpenStep on OS/2. My problem is however with the whole Openstep thing is that next made available the foundation and appkit specs but never the EOF specs. This means that if you use EOF, you are stuck to using NeXT's Openstep and why the effort of even developing Openstep/EOF Apps, if they will only run on NeXT's Openstep. The only solution I could find to this problem was to write your own Database layer on top of EOF. Then you could change this layer on other environments where EOF is not available and the interface to your Databases should still be the same. By doing my solution to the problem you might as well not use EOF at all as you are going to develop the layer in any case for the other environments you want to run on. Am I missing something somewhere, or is this the way I will have to do it? If this is the case then we might have the same problem with any other potential NeXT products that might be released. -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 12 Jun 1996 06:56:12 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4plpmc$68g@news.xmission.com> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> <4pkg9u$9l5@mailgate.lexis-nexis.com> uhendjx@meaddata.com (Jonathan W. Hendry) wrote: > [...] > a NeXTMail message is nearly useless if sent to someone without > NeXTMail. It's *completely* useless if the receipient has no access > to uudecode and compress software (most Windows and Mac users). > > At least with a MIME message, useful content is available on > platforms that do not support every nifty feature. Even if an > attachment is not usable, text in the rest of the body should > still be readable. (Unlike NeXTMail, which garbles the entire > message.) I've often posited that NeXTMail would have been a lot better if it had taken an approach similar to NewsGrazer with its messages. Put a plain ASCII text version of the message before the uuencoded data. On a mailer that handles the rich data, display it that way. On any other system, at least they can read the message easily. I think that this could pretty much be accomplished right now with MIME anyway, if NeXT had done more than athe existing quick and dirty MIME wrapper around NeXTMail. In other words, it *would* be possible to include the "NeXTMail" message in a MIME mail message as an attachment. Mail.app could have a little intelligence added to recognize the "NeXTMail" attachment in the message and throw away the MIME message and only display the RTF part. IMHO, this has a *lot* of advantages over the current NeXTMail implementation: it solves John's concerns with MIME, makes NeXTMail more MIME compatible, and would require very little work to be done to Mail.app itself. All in all, it could be a big win. And it will never happen, I'll bet. I wanted this back in 1991...it hasn't happened yet. Back then MIME wasn't really a player. I would have been happy with the actual Newsgrazer solution that was used by Jayson. But the MIME enhancement I describe here is quite an improvement. Especially since someone on Windows with a good MIME app could extract the RTF and actually view it if they wanted to. The current NeXTMail would utterly baffle the typical Windows user! By the way, don't think from this that I'm bashing NeXTMail. I like it a lot. But I think it could be better...and wish it were. In the current computing climate, there are fewer and fewer cases where NeXTMail is the optimal solution. I think John's point still stands: there are a few cases where it may be the best solution. However, MIME is the most sensible solution in nearly every case... -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 12 Jun 1996 07:05:48 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4plq8c$68g@news.xmission.com> References: <4n0bss$s58@news4.digex.net> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4p2u9k$oal@pub02.va.pubnix.com> <4p5n45$5lm@usenet.rpi.edu> <4paei7$j0@tallowcross.uni-frankfurt.de> <sliOfYC00iWV01cK4O@andrew.cmu.edu> <4phg6p$nc@tallowcross.uni-frankfurt.de> <Ilj6Rim00iVGI4jHgu@andrew.cmu.edu> <4pi41b$odv@news.xmission.com> <EljPSuG00iUxQ1gOx9@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > A fully-qualified proposition (one where everyone agrees what all of the > terms refer to) can be tested and determined to be true or false; if it > isn't possible to test the proposition for some reason, it is unanswered. OK. I agree with that. > My statement about MIME and NeXTmail above is fully-qualified (I > believe), and is pretty obviously true. It remains a true statement > regardless of what other propositions you consider. If you consider the > issue of MIME, NeXTmail and Intranet usage, that is a different > proposition which has it's own answer that is (mostly) independent of > the statement that I made. I agree with you on this, and did when I made my post. > > For some people in some situations, NeXTMail is far superior to MIME. > > There are fewer and fewer cases where that is true, but for some it > > IS true. For most it is not. > > Certainly, at least if you are judging 'superior' according to the > perspective of the users. > [...]. If you discount the display aspect (in which I agree that > NeXTmail is more capable than MIME of precise formatting), and consider > technical issues like the encoding mechanisms used (uuencode vs > base-64), or the size of the messages (every try to send a > compressed/gziped file via NeXTmail?), MIME is superior to NeXTmail. And I agree with this by far. Note my other post suggesting that NeXTMail's functionality could be made a heck of a lot more friendly in this respect by simply sending NeXTMail as MIME with the RTF as a MIME attachment. Makes a lot of sense to me. It is still proprietary, but at least it is accessible to non-NeXT users and still retains the functionality that the USER of NeXTMail wants to have in there. > > [... The big folder problem ...] > > I would place the blame, though, on NeXT and not on MIME. > I would too-- NeXT's implementation of MIME, at least in this aspect, > leaves much to be desired. No kidding! :-) By the way, WorkSpace's chunking is just as broken. With all the good things they've done, I find it annoying that stupid little things like these don't get fixed. I guess I'm becoming a cranky old curmudgeon like all the other long time NeXT users that are still here...how sad. ( :-) ) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 12 Jun 1996 07:13:31 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4plqmr$68g@news.xmission.com> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> <4pkg9u$9l5@mailgate.lexis-nexis.com> <4plpmc$68g@news.xmission.com> don@misckit.com wrote: > In other words, it *would* be possible to include the "NeXTMail" > message in a MIME mail message as an attachment. Mail.app could > have a little intelligence added to recognize the "NeXTMail" > attachment in the message and throw away the MIME message and only > display the RTF part. IMHO, this has a *lot* of advantages over > the current NeXTMail implementation: it solves John's concerns > with MIME, makes NeXTMail more MIME compatible, and would require > very little work to be done to Mail.app itself. All in all, it > could be a big win. A quick addendum as I think of this. If they were willing to go to just a little tiny bit more trouble, change "NeXTMail" to be based upon PDF instead of RTF. PDF is richer than NeXT's RTF and would therefore provide a superset of the formatting options that would make a user appreciate NeXTMail. And it even has the added advantage of being more universal. The only problem with this is that PDF and the Text object aren't very close friends right now. I wish they were, though! :-) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: Laurent.Champciaux@emn.fr Newsgroups: comp.sys.next.programmer Subject: How to write a threadsafe Objective-C class ? Date: 12 Jun 1996 14:05:39 GMT Organization: Ecole des Mines de Nantes Distribution: world Message-ID: <4pmirj$3hb@wfn.emn.fr> Keywords: thread, Objc Hi, How to write a threadsafe Objective-C class such as List or HashTable ? I want to use the MiscThreadedObject class of the MiscKit. My threadedObject will have to send messages to other objects of my own classes that are not threadsafe. Also, I am looking for any information about multi-threaded applications under NEXTSTEP. Thanks in adance for your help. Laurent. -- ======================================================= Laurent Champciaux Departement Informatique - Ecole des Mines de Nantes 4, rue A.Kastler 44070 Nantes Cedex 03 Tel: (33)51 85 82 18 (B220) email: Laurent.Champciaux@emn.fr WWW: http://www.emn.fr/dept_info/perso/laurent/
Newsgroups: comp.sys.next.programmer From: olaf@megatel.de (Olaf Seifert) Subject: where can i get a javainterpreter for next ? Message-ID: <1996Jun12.145250.9180@megatel.de> Sender: news@megatel.de Organization: megatel GmbH, Bremen, Germany Date: Wed, 12 Jun 96 14:52:50 GMT I want develop on a save and useful platform, so I'm interrested for a jdk for NeXT !!!!!
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 12 Jun 1996 23:08:25 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pnil9$eln@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4p1u9d$1ip@tallowcross.uni-frankfurt.de> <4pet07$2bq@saturn.genoa.com> <4pkmje$dru@usenet.rpi.edu> Hi Garance :) Garance A Drosehn <gad@eclipse.its.rpi.edu> wrote: > I'd also say that I think we've beaten this dead horse long enough. I can't imagine that a half-dozen people vocally supporting NeXTmail in the NeXT newsgroups are going to influence much of anyone. It doesn't even influence me that Sun should bother with the NeXTmail format, and I agree that NeXTmail does better than MIME for some things. I agree, but what good is an advacocay group if we cant beat a dead horse every once in a while...who cares if SUN or NeXT ever listens...this place is my therapy :) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 12 Jun 1996 23:14:12 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4pnj04$eln@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4p4qsd$74p@news4.digex.net> <Dsn8zF.3xs@nvc.cc.ca.us> <4pa321$852@news3.digex.net> <DssL2p.I8u@nvc.cc.ca.us> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> <4pkg9u$9l5@mailgate.lexis-nexis.com> <4plpmc$68g@news.xmission.com> Hi Don :) don@misckit.com wrote: > In other words, it *would* be possible to include the "NeXTMail" message in a MIME mail message as an attachment. Mail.app could have a little intelligence added to recognize the "NeXTMail" attachment in the message and throw away the MIME message and only display the RTF part. IMHO, this has a *lot* of advantages over the current NeXTMail implementation: it solves John's concerns with MIME, makes NeXTMail more MIME compatible, and would require very little work to be done to Mail.app itself. All in all, it could be a big win. This is such a good and simple and brilliant idea, it's a wonder why it's not done! Perhaps NeXT will hear and do just that...I can dream can't I :) -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: nazari@cs.chalmers.se (Nader Nazari,1018) Newsgroups: comp.sys.next.programmer Subject: Help!! Renderman .. Text... Date: 11 Jun 1996 14:18:37 GMT Organization: Chalmers University of Technology Message-ID: <4pjv7t$qc5@nyheter.chalmers.se> Hi I am working on a project and most of the time I work with 3D visualization. ( Renderman) I have some " shapes" on my " shape hierarchy" and I move the light and change the distance and rotate the figure and ... as you can see in the famous Renderman example which can be found here: /NextDeveloper/Examples/3Dkit/Simple Now I need to write some text on the different part of the figure and also draw some line and curve on it . I can do it with postscript but the problem is that I want to "move" the whole figure all together. You can use all Renderman facility in Next with (Shape Class) but the postscript code must be in (View Class (Camera class) ). How can I force the postscript figures/Text to Follow the (Shape). Tanks Nader nazari@cs.chalmers.se
From: mcgredo@crl.com (Donald R. McGregor) Newsgroups: comp.sys.next.programmer Subject: loading dbadaptor as a bundle Date: 12 Jun 1996 23:13:17 -0700 Organization: Miskatonic University Department of Classics Message-ID: <4pobht$o1b@crl.crl.com> I'm assuming that the Oracle7.dbadaptor directory, the adaptor for Oracle under EOF, is really a bundle. It seems that one ought to be able to do an aBundle = [[NXBundle alloc] initFromDirectory:"/NextLibrary/Adaptors/Oracle7.dbadaptor"];; [aBundle principleClass]; and load all the code present in the bundle into the application. But when I try this, I get an error: memory address access exception [...] in _doprnt Eh? Is it possible to do this? -- Don McGregor | The wages of sin is death--but after taxes it's mcgredo@crl.com | only a tired feeling at the end of the day.
From: martin@rat.se (Martin Wennerberg) Newsgroups: comp.sys.next.programmer Subject: Re: methodSignatureForSelector broken? References: <4lishk$cnt@isabella.object-factory.com> Message-ID: <31bfb3be.0@news-sto1.tfi.net> Date: 13 Jun 96 06:22:54 GMT In <4lishk$cnt@isabella.object-factory.com> Erik Doernenburg wrote: > > Is it me or does the method 'methodSignatureForSelector' always produce a > memory access exception on address 0x0 when used with the Foundation Kit that > comes with EOF 1.1? I use it a lot and it "just works" :~) It's slow though, so some cashing is needed if performance is important. Martin Wennerberg ____________________________________________________ Research & Trade AB Phone: + 46 - 8 - 21 17 50 Fax: + 46 - 8 - 24 54 11 Email: martin@rat.se [MIME & NeXTmail supported] Snailmail: Box 7742, S-103 95 Stockholm, SWEDEN Visiting: Kungsgatan 33
From: hldpub@gramercy.ios.com (Hector Davila) Newsgroups: comp.sys.next.programmer Subject: Is your computer being bugged??? Date: 13 Jun 1996 09:28:43 GMT Organization: Internet Online Services Message-ID: <4pon0b$dog@news2.ios.com> Is your computer being monitored by someone else? Is someone using your computer without your knowledge? Is your mate chatting online with someone else? Are your children chatting online with the wrong crowd? Now , you can monitor your computer with my private collection of keyboard recoders from around the world. Also known as:Keyboard Grabber, Keyboard Key Logger, Keyboard Monitor. PURPOSE: Captures keystrokes and sends & saves them to a hidden file. Now you can keep a record of any keyboard activity on your computer. Monitor your computer at home or office. My private collection of keyboard recorders is yours for only $9.95. You will receive 19 different programs on a 3 1/2 disk. For Dos,Windows,and Mac's.(some come with actual source codes) You'll get:Keycopy,Keyfake,Keyread,Keytrap,Keyrec,Keylogwn(Windows), Hackkey,Bagkeys,Getit,Playback,Robokey,Record,Encore, Kcap10,Ptm229N,Qwertman,GKG,Depl,Maclife(Mac). Just send $9.95 plus $1.00 for shipping and handling to: HLD PLUBLISHING COMPANY 1680 N. VINE ST. #1103 LOS ANGELES, CA. 90028 *All orders shipped within 48hrs. *100% Satisfaction unconditional money back guarantee. *Foreign orders add $2.00 for shipping and handling.
Sender: hldpub@gramercy.ios.com (Hector Davila) From: jem@xpat.com Message-ID: <cancel.4pon0b$dog@news2.ios.com> Control: cancel <4pon0b$dog@news2.ios.com> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4pon0b$dog@news2.ios.com> Date: 13 Jun 1996 10:17:55 GMT Cancelled by jem@xpat.com. 834661075 HLDPLUB Original Subject was: Is your computer being bugged???
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep and EOF Date: 13 Jun 1996 12:44:46 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4pp2fu$6on@lion.embl-heidelberg.de> References: <4plha8$177@hermes.is.co.za> In article <4plha8$177@hermes.is.co.za> gvandyk@icon.co.za writes: > Am I missing something somewhere, or is this the way I will have to do > it? If this is the case then we might have the same problem with any > other potential NeXT products that might be released. Although EOF is in principle nice package, the object persistent could be solved in even more elegant and flexible way. So if GNU guys one day decide to write something like EOF, I'll vote for something very different. BTW, Gerrit, if you read this posting, please resubscribe to the gnustep list. I had to take your e-mail address out, because it makes a lot of trouble (user unknown, etc). -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 13 Jun 1996 15:14:48 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7g27zq20n.fsf@tom.es.ele.tue.nl> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> In-reply-to: Bernhard Scholz's message of Mon, 10 Jun 1996 11:07:11 +0200 In article <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: Tiggr's example is wrong! Yes, it is. No it isn't :-) Although it might be syntactical legal it is wrong in the concept of FK and OpenStep. The argument has nothing to do with syntax. It should be: NSMutableArray *kids; -(const NSMutableArray *)children { return kids; } If you do not want other objects to modify your kids, you certainly should not return it as a MutableArray. Let's change the example a bit to avoid getting into a yes/no war. Suppose there's this class Foo, and a class MutableFoo, the one not necessarily a subclass of the other. Suppose both adhere to the Bar protocol. Then, the same example could become: MutableFoo *kids; -(id <Bar>) children { return kids; } Should I now not be allowed to return this MutableFoo object, only because the caller might think it a good idea to cast the `id <Bar>' object returned by the callee to a MutableFoo (what it is) instead of a Foo (what I want it to be). The answer is NO. The class indication of an object is something very useful, as it allows the compiler to do static checking. There is no need to force something to be immutable at runtime if the sources say it is immutable (at compile time). Any cast from a class to a subclass, or from a protocol to a class implementing that protocol, or from anything to id (everything), is performed under the responsibility of the person typing that cast. Any bugs in code which seemed legal to the compiler as a result of such a cast is, again, the responsibility of the person that typed the cast. P.S.: It is nice, but I rarely see people using the 'const' keyword. The `const' keyword is a hack which is rather unusable. As soon as some function forgets to promise not to modify a string, you'll get compiler warnings all over the place. And since most code is not under your own control, you're stuck with it. const does not buy you anything. But with the introduction of mutable/immutable objects they think that an immutable object is the same as a 'const' mutable object --- which is wrong. An immutable object is an object which does not provide methods which affect the object's state. If the compiler tells you you're dealing with an immutable object, then it is immutable. The fact that, at runtime, it actually might be something different, is irrelevant. As said, casting is the responsibility of the person at the keybord when the cast appeared in the source code. Anybody casting through mutability should better have a very good reason for doing so. If not he should be fired :-) --Tiggr
From: "Timothy J. Luoma" <luomat@capitalist.princeton.edu> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: ~/.AppInfo Date: Wed, 12 Jun 1996 09:13:33 -0400 Organization: Princeton University Message-ID: <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi there. I'm having a bad day, and so I'm posting about something that has been very much bothering me of late... Somewhere along the way, someone came up with the idea to have the app files which would be placed in ~/ placed in a centralized folder called ~/.AppInfo. I saw this as yet another very smart idea. It seems like a good way to keep my $HOME uncluttered.... I've seen a few apps lately which have ignored this idea, and that bothers me because, well, it's there, why not use it? I don't remember what apps they were right now, but anyway, I just wanted to mention it. Thank you for helping to lower my blood pressure... TjL -- Timothy J. Luoma (Formerly known as: luomat@capitalist.princeton.edu) New email address: 476tjl@ptsmail.ptsem.edu (No NeXTmail) *** Please note: the new address cannot read the 'reply-to' line ***
From: "Timothy J. Luoma" <luomat@capitalist.princeton.edu> Newsgroups: comp.sys.next.programmer,comp.sys.next.sysadmin Subject: Re: how to issue a modem command at computer shutdown? Date: Wed, 12 Jun 1996 08:44:48 -0400 Organization: Princeton University Message-ID: <Pine.NXT.3.93.960610095154.2381B-100000@capitalist.princeton.edu> References: <4pbo7u$hk@tallowcross.uni-frankfurt.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII To: Uli Zappe <uli@tallowcross.uni-frankfurt.de> In-Reply-To: <4pbo7u$hk@tallowcross.uni-frankfurt.de> > 2. (I'm afraid that makes much more problems): How do I execute this little > script/program automatically when the computer is turned off? It must be > started AFTER all apps have already been terminated to make sure no other > program locks the serial device. Basically, I'm looking for something like > /etc/rc to make an entry in, but for the other way round; but I couldn't find > anything :-( Well, if you are willing to give up a little GUI-nicetie.... What I'd try is making a little script, called 'powerdown' something like this: ---cut here--- #!/bin/sh echo "Powering down `hostname`" # put your commands here # this must be the last line: /usr/etc/shutdown -h -p +5 # which tells the NeXT to 'halt' (-h) and then 'power down' (-p) in 5 # minutes (+5) exit 0 The 5 minutes is so you can quit/save anything you still might be working on. You could change it to '+1' or 'now' if you preferred. Good luck TjL -- Timothy J. Luoma (Formerly known as: luomat@capitalist.princeton.edu) New email address: 476tjl@ptsmail.ptsem.edu (No NeXTmail) *** Please note: the new address cannot read the 'reply-to' line ***
From: fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 13 Jun 1996 15:18:23 +0200 Organization: Institut fuer medizinische Informatik und Biometrie Sender: fxg@merry.imib.rwth-aachen.de Message-ID: <s4y20jjdeqq.fsf@merry.imib.rwth-aachen.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <4p932t$473@zdi.informatik.uni-stuttgart.de> In-reply-to: brouwer@minnie.informatik.uni-stuttgart.de's message of 7 Jun 1996 11:16:45 GMT To: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) In article <4p932t$473@zdi.informatik.uni-stuttgart.de> brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) writes: From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Date: 7 Jun 1996 11:16:45 GMT In <31b74471.157975@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: >cph@next.com (Charles d'Harcourt) wrote: >What seems sensible to me is that if a method states it returns an >NSArray, i.e. an immutable class, it _will_ return an immutable array. It's not called NSImmutableArray for a reason... >Syntatically it's valid to return NSArray's mutable subclass; but I >think methods which return classes which recognise the mutability >concept should be treated as special cases, i.e. the immutability will >be honoured as well as the object really being an instance of the >stated class. Sorry, but then NSMutableArray must not be a subclass of NSArray - but it is. Something like (NSArray *) just says something about the view on an object, not about the concrete instance of an object. This is polymorphism and what OOP is all about! Not so fast! This is a design decision, not an OO intrinsic. Actually, it violates one of the best OO guidelines I've ever heard. I agree with your conclusion, but I want to make the difficulties plain: OO is about inheritance. Inheritance can be used as generalisation or specialization. The root class is the least general one becaus it is least assumption you can make about an object. It is the most special class (in a weird sense) in that it accepts the fewest messages. Any derived class (well, many of them) is more general in that it implements a broader protocol. It is more special in that it may accept a limited range of arguments (Child : Person [age]). The former would be A-is-a-B-plus-C, the latter A-is-a-special-B. Which one of these contradictory views you take, is a design decision that should be consistent in the entire code. Specialisation: An Array is writeable, an immutable array is a special array. Generelisation: An Array is not writeable, a mutable array is. NeXT took the first way here, which looks rather odd at first sight, violating the 'programming by contract' paradigm which is another view on all that OO is about. ("Any subclass instance should accept whatever the superclass instance would", or at least (a) deny politely or (b) tell the static type system about it.) This means that as the 'customer', you get an array you might be allowed to change. This is rather awkward and illustrated why 'programming by contract' is such a tough aim. Anyway the class is named NSArray and not NSImmutableArray, so I don't see any assurences about immutability here. Creating an NSImmutableArray subclass and use it as an return type would make the whole thing clearer. Yes. Consider Array : Storage MutableArray : Array : Storage ImmutableArray : Array : Storage Wouldn't we all feel more comfortable with this? Array defines the interface, the class design puts emphasis on not guaranteeing mutability and you can be sure you can write the array if it's mutable. P.S.: I offer this point from behind a cloud of theory that has kept me away from NS/OS for a while. All I know about NS/OS design in this area is what's in this thread. There may be a pragmatic solution that has not been pointed out here but included by NeXT to alleviate this. -- Felix (fxg@(Pool.Informatik|imib).RWTH-Aachen.de NeXT-Mail ok, but slow)
From: Adam Fedor <fedor@boulder.colorado.edu> Newsgroups: comp.sys.next.programmer Subject: Re: What can be used for streams/streams.h in GNUstep? Date: Thu, 13 Jun 1996 10:16:01 -0600 Organization: Optoelectronic Computing Systems Center Message-ID: <31C03EC1.7CB9@boulder.colorado.edu> References: <4po0g6$dfp@Vir.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Stefanos Kiakas wrote: > > I'm porting some software from NeXTSTEP to GNUstep, and I'm wondering > what header file I should use for streams? > > I want to initialize an NXStringTable from memory in GNUstep. > GNUstep supports steams via an Objective-C class (*Stream), but not through some generic C interface. I guess you have two options, then: (1)Write a C interface for streams and update NXStringTable to handle these. That's not something in my plans, but it might be interesting to do. (2) Use the OpenStep interface and read the table into a NSDictionary through the use of property lists. Unfortuanetly this isn't completely implemented in the current release of GNUstep. --- Adam Fedor. CU, Boulder | Fudd's Law of Opposition: fedor@colorado.edu (MIME) | "If force doesn't work, adam@rmnug.org (NeXTMail) | your not using enough."
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: fozztexx@nvc.cc.ca.us (Chris Osborn) Subject: Re: NeXTMail is just different Message-ID: <Dsx9C2.1o8@nvc.cc.ca.us> Sender: news@nvc.cc.ca.us Organization: Napa Valley College References: <4n0bss$s58@news4.digex.net> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> Date: Thu, 13 Jun 1996 04:48:01 GMT In article <4piefd$dr7@news4.digex.net>, John Kheit <jkheit@cnj.digex.net> wrote: >uhendjx@meaddata.com (Jonathan W. Hendry) wrote: >However, the only app I've heard >of that does do formatting this good using MIME as its format vehicle is >Myhnah (sp?) and that is likely to be a smaller number than NeXTmail app >mailers. I only know of *one* NeXTMail mailer, Mail.app. I also only know of *one* MIME mailer that has equivalent capabilites, Mynah. I don't think *one* is a smaller number than *one*. It seems to me, since you insist on restricting your mail to NeXTSTEP *only* platforms, and Mynah is a NeXTSTEP app, you could switch to Mynah. Then if you *did* send mail to someone on a non-NeXTSTEP platform that was using a MIME mailer, they'd at least be able to make *some* sense of your message. And everyone else using Mynah would still see your beautifully formatted message. -- Chris Osborn, Network Administrator Napa Valley College 707 253 3130 - Voice 2277 Napa-Vallejo Hwy. 707 253 3063 - Fax Napa, CA 94558 <fozztexx@nvc.cc.ca.us> <http://www.nvc.cc.ca.us/~fozztexx>
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: 13 Jun 1996 18:30:30 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4ppmo6$mca@news4.digex.net> References: <4n0bss$s58@news4.digex.net> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> <Dsx9C2.1o8@nvc.cc.ca.us> fozztexx@nvc.cc.ca.us (Chris Osborn) wrote: > In article <4piefd$dr7@news4.digex.net>, > John Kheit <jkheit@cnj.digex.net> wrote: > >uhendjx@meaddata.com (Jonathan W. Hendry) wrote: > >However, the only app I've heard > >of that does do formatting this good using MIME as its format vehicle is > >Myhnah (sp?) and that is likely to be a smaller number than NeXTmail app > >mailers. > I only know of *one* NeXTMail mailer, Mail.app. There are several out their... Eloquence comes immediatly to mind...and a couple of others... Of course they are all NS applications... I know of no app that can do NeXTmail that is not a NS app... > I don't think *one* is a smaller number than *one*. I'll spare you the metaphysics :) See above for ez response... > It seems to me, since you insist on restricting your mail to NeXTSTEP *only* platforms, and Mynah is a NeXTSTEP app, you could switch to Mynah. Then if you *did* send mail to someone on a non-NeXTSTEP platform that was using a MIME mailer, they'd at least be able to make *some* sense of your message. And everyone else using Mynah would still see your beautifully formatted message. More over...there are LESS people using Mynah than even just Mail.app (even w/o the other two or three NeXTmail capable mailers). AGAIN!!!! THE POINT IS I DON'T WANT THEM TO SEE THE DOCUMENT AT _ALL_ UNLESS THEY CAN SEE IT PERFECTLY. IF THE ONLY PEOPLE THAT CAN SEE IT PERFECTLY USE MYNAH, AND THERE ARE LESS MYNAH USERS THAN MAIL.APP USERS, WHY WOULD I USE MYNAH? I must be miserably unclear on this... Ok, let me try and restate it better...I need the guarantee...perfection or they see NOTHING AT ALL. I don't want them to make any sense of my document at all unless they see it perfectly. MIME doesn't offer that. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 13 Jun 1996 18:05:52 GMT Organization: Rockwell Avionics - Collins Message-ID: <4ppla0$qkt@castor.cca.rockwell.com> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <4p932t$473@zdi.informatik.uni-stuttgart.de> <s4y20jjdeqq.fsf@merry.imib.rwth-aachen.de> Cc: fxg@imib.rwth-aachen.de > Which one of these contradictory views you take, is a design decision that > should be consistent in the entire code. > > Specialisation: An Array is writeable, an immutable array is a special array. > > Generelisation: An Array is not writeable, a mutable array is. > > NeXT took the first way here, which looks rather odd at first sight, violating > the 'programming by contract' paradigm which is another view on all that OO is > about. ("Any subclass instance should accept whatever the superclass instance > would", or at least (a) deny politely or (b) tell the static type system about > it.) > This is exactly wrong! NeXT chose the second way (Generalization) saying (An Array is mutable, an immutable array is a special array) requires implementations of the methods that change the array in the immutable class. > This means that as the 'customer', you get an array you might be allowed to > change. This is rather awkward and illustrated why 'programming by contract' is > such a tough aim. > > Anyway the class is named NSArray and not NSImmutableArray, so I don't see > any assurences about immutability here. Creating an NSImmutableArray subclass > and use it as an return type would make the whole thing clearer. > > Yes. Consider > > Array : Storage > MutableArray : Array : Storage > ImmutableArray : Array : Storage > What is the difference between the Array interface and the Immutable Array interface in this decomposition ? MutableArray presumably adds methods to modify the Array. - Erik M. Buck
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Southern CA NeXT users Date: 13 Jun 1996 21:25:26 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4pq106$6ik@news.onramp.net> If you're in the Southern CA area and interested in getting together with other users for bi-monthly user group meeting please send me email. Steve Dekorte dekorte@suite.com
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Adding popUp to Scrollview Slider Date: 14 Jun 1996 01:39:50 GMT Organization: On-Ramp; Individual Internet Connections; Dallas/Ft Worth/Houston, TX USA Message-ID: <4pqft6$aag@news.onramp.net> Anyone know the easiest way to ad a popup menu to the send of a Slider controling a scrollview. I can imagine ugly ways to do this, but I'm sure NeXT has made this easy to do although I can't find any docs on it. Thanks for any info, Steve
From: cpayne@optical.fiber.net (Carl Payne) Newsgroups: comp.sys.next.programmer Subject: munmap, INN, NS3.3: undefined symbol Date: 12 Jun 1996 17:28:26 GMT Organization: Fibernet! Satisfy the need for speed Message-ID: <4pmunq$qm2@optical.fiber.net> INN unoff4 on NS 3.3, x86, built single architecture. See, funny thing is, it had to work at some time because I'm running INN unoff1 right now. Anyhow... Here's where it dies in the compilation process: cd innd ; make all ; cd .. cc -g -o innd art.o cc.o chan.o his.o icd.o innd.o lc.o nc.o newsfeeds.o ng.o proc.o rc.o site.o ../libinn.a ld: Undefined symbols: _munmap *** Exit 1 Stop. *** Exit 1 Stop. Hmmm... OK, so I did a little digging, and thought maybe sys/types.h and sys/mmap.h weren't found, so I added an extra path: -I/usr/include/bsd and that still didn't work. References to mmunmap are only scarce, but _mmunmap is downright extinct. Does anyone have a clue I can borrow? -- Carl Payne cpayne@fiber.net <--NeXTMAIL, MIME, ASCII all OK Cogito Ergo Amicrosoft
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: ~/.AppInfo Date: 13 Jun 1996 22:03:27 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4pq37f$7f6@usenet.rpi.edu> References: <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> "Timothy J. Luoma" <luomat@capitalist.princeton.edu> wrote: > Somewhere along the way, someone came up with the idea to have > the app files which would be placed in ~/ placed in a centralized > folder called ~/.AppInfo. I saw this as yet another very smart > idea. It seems like a good way to keep my $HOME uncluttered.... > > I've seen a few apps lately which have ignored this idea, and > that bothers me because, well, it's there, why not use it? I > don't remember what apps they were right now, but anyway, I just > wanted to mention it. Yeah, I like this idea too. It seemed like a good place to tuck away configuration files for individual applications. I'd like to see more NeXTSTEP apps take advantage of it too. (well, actually I'd like to see unix utilities in general do that too, but that's probably too much to hope for) --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: Re: Adding popUp to Scrollview Slider Date: 14 Jun 1996 07:09:52 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4pr380$h5f@ai.alienor.fr> References: <4pqft6$aag@news.onramp.net> If I understand what who want... You should look at an example named scrollDoodScroll which performs the thing. I've got it if you don't. -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: synper5@aol.com (Syn Per5) Newsgroups: comp.sys.next.programmer Subject: Tell me what you want for a change! Date: 14 Jun 1996 04:06:13 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4pr6hl$9eb@newsbf02.news.aol.com> Rather than wasting your time looking though all my job openings just tell me what YOU want in you next career move, and when you will be ready to make this move. I will have one of my staff hunt down that opportunity for you. Just give us all the specifications: location; compensation range; titles and descriptions of the type of work; the type of company etc... the more details the better. Do not send us your resume. We will contact you when we found what you want and come up with a game plan for landing you the opportunity, but until then lets hold off on sending a resume. However, please included all relevant contact information, so we can get a hold of you. If your a contractor, you've come to the right place, we are "the experts" in 1099. If you are a company person moving up the latter or just looking for a position that will give you peace of mind. Again we have been doing this for 12 years and have great contacts and the know how. Even if you are not interested in making a move for sometime, we would like to plan ahead for you. Also we give great bonus for the right referrals so let your friends know, and make sure they us your name! Best Regards, Matthew Greedy, Technical Placement Supervisor Synergy Inc.
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Text and textDidChange: Date: Fri, 14 Jun 1996 09:31:32 +0100 Organization: Cedar Systems Distribution: world Message-ID: <jJAKxBAkNSwxEw4P@cedar.co.uk> References: <4p022l$49q@solaris.cc.vt.edu> MIME-Version: 1.0 In article <4p022l$49q@solaris.cc.vt.edu>, Chuck_Esterbrook@orcacomputer .com writes >Hi, > >I have an app that uses a text editing window where the contents of the text >are constantly changed as the user goes clicking through the browser in the >same window. > >I'd like to capture the text whenever the user clicks a browser cell (or the >main window is resigned) *and* the text was changed. If the text wasn't >changed, no point in capturing it as this makes the app sluggish. It's also >important to know when it has changed so that I can "setDocEdited:YES". > >I implemented -textDidChange: but the "text" in this message seems to refer >to the ASCII characters and not the text object. If the user changes fonts, >drops colors, etc. the message is not sent. > >I want something equivalent to "-theContentsOfTheTextObjectDidChange:" which >would be sent whenever the characters *or* formatting changed. > >Has anyone faced this problem and worked around it? Perhaps there is a custom >subclass of Text on the net that fixes this design flaw? > >Thanks, > >Chuck There is a private Text method _replaceSel::::::: that gets invoked to update the text contents. You can override it in a subclass to handle your requirements: - replaceSel:(int)p1 :(wchar *)p2 :(int)p3 :(NXRunArray *)p4 :(int)p5 :(int)p6 :(int)p7 { id retval = [super replaceSel:p1 :p2 :p3 :p4 :p5 :p6 :p7]; [self theContentsOfTheTextObjectDidChange:self]; return retval; } Hope this is of some use to you. Regards, Paul. P.S. I am currently looking for Beta testers for our budget NEXTSTEP word processor - please email me: phef@cedar.co.uk if you would like to help. ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: ~/.AppInfo Date: 14 Jun 1996 10:29:18 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4pretu$je8@zdi.informatik.uni-stuttgart.de> References: <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> In <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> "Timothy J. Luoma" <luomat@capitalist.princeton.edu> writes: >Somewhere along the way, someone came up with the idea to have the app >files which would be placed in ~/ placed in a centralized folder called >~/.AppInfo. I saw this as yet another very smart idea. It seems like a >good way to keep my $HOME uncluttered.... Well, it stands in the docs, but I can't tell where at the moment. >I've seen a few apps lately which have ignored this idea, and that bothers >me because, well, it's there, why not use it? I don't remember what apps >they were right now, but anyway, I just wanted to mention it. What do you mean with "lately"? Not even NeXT's own apps (eg. Edit and IB) use .AppInfo. And with such a paragon....no wonder...... >Thank you for helping to lower my blood pressure... Thank you for highering my blood preasure now. >TjL Klaus Brouwer
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 14 Jun 1996 10:43:21 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4prfo9$lco@zdi.informatik.uni-stuttgart.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> <x7g27zq20n.fsf@tom.es.ele.tue.nl> In <x7g27zq20n.fsf@tom.es.ele.tue.nl> tiggr@es.ele.tue.nl (Pieter Schoenmakers) writes: >In article <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: > Tiggr's example is wrong! Yes, it is. >No it isn't :-) Right. >Suppose there's this class Foo, and a class MutableFoo, the one not >necessarily a subclass of the other. Suppose both adhere to the Bar >protocol. Then, the same example could become: > MutableFoo *kids; > -(id <Bar>) children > { > return kids; > } >Should I now not be allowed to return this MutableFoo object, only because >the caller might think it a good idea to cast the `id <Bar>' object >returned by the callee to a MutableFoo (what it is) instead of a Foo (what >I want it to be). The answer is NO. The class indication of an object is >something very useful, as it allows the compiler to do static checking. >There is no need to force something to be immutable at runtime if the >sources say it is immutable (at compile time). Any cast from a class to a >subclass, or from a protocol to a class implementing that protocol, or >from anything to id (everything), is performed under the responsibility of >the person typing that cast. Any bugs in code which seemed legal to the >compiler as a result of such a cast is, again, the responsibility of the >person that typed the cast. RIGHT, RIGHT, RIGHT! I hope you post will put this discussion to a reasonable end. > P.S.: It is nice, but I rarely see people using the 'const' keyword. Anyway: if you have a fooFunction(const char *str); and fooFunction caches str in some static variable, just NOBODY guarantess you that the creator of str doesn't modify it in some way, since it might be declared as char *str. >The `const' keyword is a hack which is rather unusable. As soon as some >function forgets to promise not to modify a string, you'll get compiler >warnings all over the place. And since most code is not under your own >control, you're stuck with it. const does not buy you anything. Because C does not support the concept of actual constants. [munch] Have a nice day, Klaus Brouwer
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 14 Jun 1996 11:22:22 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4pri1e$qre@zdi.informatik.uni-stuttgart.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <4p932t$473@zdi.informatik.uni-stuttgart.de> <s4y20jjdeqq.fsf@merry.imib.rwth-aachen.de> In <s4y20jjdeqq.fsf@merry.imib.rwth-aachen.de> fxg@imib.rwth-aachen.de (Felix H. Gatzemeier) writes: > >Syntatically it's valid to return NSArray's mutable subclass; but I > >think methods which return classes which recognise the mutability > >concept should be treated as special cases, i.e. the immutability will > >be honoured as well as the object really being an instance of the > >stated class. > > Sorry, but then NSMutableArray must not be a subclass of NSArray - but it > is. Something like (NSArray *) just says something about the view on an > object, not about the concrete instance of an object. This is polymorphism > and what OOP is all about! >Not so fast! This is a design decision, not an OO intrinsic. Actually, it >violates one of the best OO guidelines I've ever heard. I agree with your >conclusion, but I want to make the difficulties plain: OO is about >inheritance. NO! This IS an OO intrinsic! If there's a parameter typed (NSArray *) I can pass instances of all subclasses of NSArray to it. This is guaranteed by the language! If this leads to rubbish than the class design is wrong, which is nothing else then a more complicated form of a programming error. This polymorphism is the reason for dynamic binding, and dynamic binding has to do much with OOP, right? If there always has to be a real NSArray instance when there is and (NSArray *) parameter or result type, we have no OOP any more but pure boring procedural programming (well Ada does it's job a little better) >Inheritance can be used as generalisation or specialization. The root class is >the least general one because it is least assumption you can make about an >object. It is the most special class (in a weird sense) in that it accepts the >fewest messages. Any derived class (well, many of them) is more general in >that it implements a broader protocol. It is more special in that it may >accept a limited range of arguments (Child : Person [age]). The former would >be A-is-a-B-plus-C, the latter A-is-a-special-B. >Which one of these contradictory views you take, is a design decision that >should be consistent in the entire code. >Specialisation: An Array is writeable, an immutable array is a special array. >Generelisation: An Array is not writeable, a mutable array is. The classical "is an ellipse a subclass of circle or vice-versa" problem. This is addressed in modern OO programming languages (eg. Sather) by separation of the type hierarchy and the inheritance hierarchy: a Type in this sense is quite the same as an Objective-C protocol and it determines the possible usages of an object. On the other hand the inheritance hierarchy provides code reuse. This seems odd at first sight, but all the confusion about immuatbility results from the mixture of both hierarchies in Objective-C and the consequences for class design (I really like protocols!). >NeXT took the first way here, which looks rather odd at first sight, violating >the 'programming by contract' paradigm which is another view on all that OO is >about. ("Any subclass instance should accept whatever the superclass instance >would", or at least (a) deny politely or (b) tell the static type system about >it.) NeXT took the second approach. An NSMutableArray behaves just as NSArray, but if you store it somewhere it might of course be changed by its creator. This is just the same as a (const char *str) where str might in fact be declared as (char *) and so be modified by its creator. If you want be sure that this doesn't happen, you (YOU! Not the real owner of str.) have to make a copy of it and become yourself an owner. Have a nice day, Klaus Brouwer
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 14 Jun 1996 11:29:21 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4prieh$rr6@zdi.informatik.uni-stuttgart.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <4p932t$473@zdi.informatik.uni-stuttgart.de> <s4y20jjdeqq.fsf@merry.imib.rwth-aachen.de> <4ppla0$qkt@castor.cca.rockwell.com> In <4ppla0$qkt@castor.cca.rockwell.com> embuck@palmer.cca.rockwell.com (Erik M. Buck) writes: >> Anyway the class is named NSArray and not NSImmutableArray, so I don't >> see any assurences about immutability here. Creating an NSImmutableArray >> subclass and use it as an return type would make the whole thing clearer. >> >> Yes. Consider >> >> Array : Storage >> MutableArray : Array : Storage >> ImmutableArray : Array : Storage >> >What is the difference between the Array interface and the Immutable Array >interface in >this decomposition ? >MutableArray presumably adds methods to modify the Array. It's not a question of adding methods. The point here is that MutableArray and ImmutableArray are in different branches of the class tree and so (ImmuatbleArray *)someArray; will never return a MutableArray instance. Have a nice day, Klaus Brouwer
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <eott.com!tappd@nuchat.sccsi.com> Message-ID: <9606141419.AA02784@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@eott.com> Date: Fri, 14 Jun 96 09:19:39 -0500 Subject: Re: Adding popUp to Scrollview Slider Steve Dekorte wrote: > Anyone know the easiest way to ad a popup menu to > the send of aSlider controling a scrollview. I can > imagine ugly ways to do this, but > I'm sure NeXT has made this easy to do although > I can't find any docs on it. If I understand what you're after, Garfinkel & Mahoney outline just such a technique in Chapter 19 of their text. If you don't have the book handy, let me know and I'll send you the code for their "ZoomScrollView" class. Caveat: I have found occasional minor problems with their code as written. I haven't worked through Chapter 19, so while my confidence is high that their code fires up right out of the box, it isn't absolute. Regards, - Dan
From: eric@skatter.USask.Ca Newsgroups: comp.sys.next.programmer Subject: Re: Adding popUp to Scrollview Slider Date: 14 Jun 1996 16:04:58 GMT Organization: University of Saskatchewan Message-ID: <4ps2ja$dtv@tribune.usask.ca> References: <4pqft6$aag@news.onramp.net> <4pr380$h5f@ai.alienor.fr> The book NeXTSTEP PROGRAMMING by Garfinkel and Mahoney has a good example of how to do this. -- Eric Norum eric@skatter.usask.ca Saskatchewan Accelerator Laboratory Phone: (306) 966-6308 University of Saskatchewan FAX: (306) 966-6058 Saskatoon, Canada. NeXTMail accepted.
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: ~/.AppInfo Date: 14 Jun 1996 16:28:44 GMT Organization: University of Sheffield, UK Message-ID: <4ps3vs$5op@bignews.shef.ac.uk> References: <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> <4pretu$je8@zdi.informatik.uni-stuttgart.de> In-Reply-To: <4pretu$je8@zdi.informatik.uni-stuttgart.de> On 06/14/96, Klaus Brouwer wrote: > In <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> "Timothy J. Luoma" <luomat@capitalist.princeton.edu> writes: > > >Somewhere along the way, someone came up with the idea to have the app > >files which would be placed in ~/ placed in a centralized folder called > >~/.AppInfo. I saw this as yet another very smart idea. It seems like a > >good way to keep my $HOME uncluttered.... > Well, it stands in the docs, but I can't tell where at the moment. > > >I've seen a few apps lately which have ignored this idea, and that bothers > >me because, well, it's there, why not use it? I don't remember what apps > >they were right now, but anyway, I just wanted to mention it. > What do you mean with "lately"? Not even NeXT's own apps (eg. Edit and IB) > use .AppInfo. And with such a paragon....no wonder...... > I knew this sort of thing had been discussed before... Best wishes, mmalc. --- Date: Sat, 13 Jun 92 13:09:46 BST ^^ From: M.Crawford To: Martin_Reed@com.next Subject: Re: British spelling and A4 Preview Cc: UK-NeXT-USERS@uk.ac.york.ohm Comments: UK NeXT Users Mailing List Hi Martin, > As for "dot" files, I certainly sympathise - at last count > I had 35 of this little perishers. However, I would really > recommend people avoid adding to the .NeXT directory as > we reserve the right to change it in unpredictable ways; > in essence it is a reserved space. > Thanks for pointing this out, I didn't notice it mentioned during the previous discussion. Perhaps, then, NeXT could try to encourage developers to put their applications' initialisation files/directories into a .NeXTApps directory, just to make things a bit tidier? Have fun, Malcolm. --
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: How to trace ObjC messages Date: 14 Jun 1996 10:25:37 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4ps7ah$q0@df13h.mdd.comm.mot.com> References: <4pkmc3$p47@news.mr.net> In article p47@news.mr.net, bmont@rconnect.com (Bruce Montegani) writes: >Does anyone know how to trace all messages that are executed when an >application is running including API kit messages. Try linking your application with /usr/lib/libPeep.a (using "-lPeep") Then, run your app, and look at it with AppInspector with the Peep option turned on. Unfortunately, I don't know how to link libPeep into an existing application... Cheers, Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: mpaque@pacbell.net (Mike Paquette) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: NeXTMail is just different Date: Fri, 14 Jun 1996 19:08:30 GMT Organization: Electronics Service, Unit No. 16 Message-ID: <4psde9$8f3@news.snfc21.pacbell.net> References: <4n0bss$s58@news4.digex.net> <4phnc0$ltv@news3.digex.net> <4phsqd$l60@mailgate.lexis-nexis.com> <4piefd$dr7@news4.digex.net> <Dsx9C2.1o8@nvc.cc.ca.us> <4ppmo6$mca@news4.digex.net> Fortunately, as of Release 3.3 Mail.app supports both Mime and NeXTMail formats. Here's what I do: Bring up Mail's Preferences panel. Select 'Compose' on the popup. Set the Compose Format to be MIME. In the Reply box, check the Reply in Same Format option. Now when sending mail to J Random User I'll default to the lossy but standard MIME format. When replying to mail, I'll be using a format that the sender can understand (ASCII, MIME, or NeXTMail). If I just gotta have NeXTMail, it's only a mouse click away on the Compose panel. The two formats were designed for slightly different purposes. MIME is intended to be a standard for communicating between diverse machines of unknown capability. A MIME mail reader can be written to work reasonably well on a VT-52, a DOS box,, or a nice graphical display. The information presented to the user should be a 'best effort' for the capabilities of the display system. NeXTMail was developed to support collaborative efforts among groups of people working on a local area network, all using the same type of machine, along with supporting the 7 bit ASCII mail format then in common use over wide area networks. (In fact, Bryan wrote it as his National Small Programs Week project!) It wasn't intended to be an Internet standard. It's biggest advantage over the standard ASCII format that dominated in 1988 was that it existed and worked well for it's intended use. Mike Paquette I don't speak for my employer, and they don't speak for me. mpaque@pacbell.net Personal E-mail mpaque@next.com NeXT business mail only, please
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Fax Cover? Date: 14 Jun 1996 20:45:52 GMT Organization: Turbocat's Development, Germany Message-ID: <4psj20$26r@turbocat.snafu.de> Hi! I a fax cover, you can insert fields for date etc. When you enable the faxing of a cover sheet these fields are filled with data. I want to do that too. But I want to print the "cover" as the document and fill it with custom data. How is this done? Which online manuals explain that? I found nothing. Thanks in advance. _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: johnam@beta.datastorm.com (John A. Maier) Newsgroups: comp.sys.next.programmer Subject: gcc exiting because it can't fork Date: 14 Jun 1996 22:16:06 GMT Organization: Datastorm Technologies Inc. Message-ID: <4psob6$7s9@beta.datastorm.com> Here is what I am doing... wilbur:/root>make cc -O2 assign1.c -o assign1 /usr/local/lib/i386/as: can't fork a new process to execute: /usr/local/lib/i386/as (No more processes) *** Exit 1 Stop. I have a 486DX2-66 with 20Megs of RAM, 65 Megs free (400 Meg Drive), Cirrus Logic GD5428 running at 1024x768 (2 bit grey), everything shutdown that I can think of (I even killed nfs processes to free additional resources). What do I need to do to get a compile to work? jam
From: Matt_Watson@NeXT.com (Matt Watson) Newsgroups: comp.sys.next.programmer Subject: Re: How to trace ObjC messages Date: 15 Jun 1996 04:33:00 GMT Organization: NeXT Software, Inc. Message-ID: <4pteds$nsm@news.next.com> References: <4pkmc3$p47@news.mr.net> In article <4pkmc3$p47@news.mr.net> bmont@rconnect.com (Bruce Montegani) writes: > Does anyone know how to trace all messages that are executed when an > application is running including API kit messages. > (gdb) b objc_msgSend Breakpoint 1 at 0x60044f0 (gdb) comm 1 Type commands for when breakpoint 1 is hit, one per line. End with a line saying just "end". >silent >p *(char **)($sp+8) >up >cont >end (gdb) r (Getting this to work on HP-PA is left as an exercise to the reader...) matt. -- Matt Watson Thaumaturgist NeXT Software, Inc.
From: "Timothy J. Luoma" <luomat@capitalist.princeton.edu> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: ~/.AppInfo Date: Fri, 14 Jun 1996 09:41:07 -0400 Organization: Princeton University Message-ID: <Pine.NXT.3.93.960614093720.2754B-100000@capitalist.princeton.edu> References: <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> <4pretu$je8@zdi.informatik.uni-stuttgart.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII To: Klaus Brouwer <brouwer@minnie.informatik.uni-stuttgart.de> In-Reply-To: <4pretu$je8@zdi.informatik.uni-stuttgart.de> On 14 Jun 1996, Klaus Brouwer wrote: > >I've seen a few apps lately which have ignored this idea, and that bothers > >me because, well, it's there, why not use it? I don't remember what apps > >they were right now, but anyway, I just wanted to mention it. > > What do you mean with "lately"? Not even NeXT's own apps (eg. Edit and > IB) use .AppInfo. And with such a paragon....no wonder...... I don't know much about IB, but as far as Edit.app and other NeXT apps, they all use the defaults database for their info. ~/.AppInfo is a place to store information which is not appropriate for the defaults database, such as Fiend's dock and shelf information... If your app needs to drop extra files someplace, why not ~/.AppInfo/Appname? > >Thank you for helping to lower my blood pressure... > Thank you for highering my blood preasure now. I do not hope to raise it now, but I don't think you really understood my original post, which I take 50% blame for.... TjL -- Timothy J. Luoma (Formerly known as: luomat@capitalist.princeton.edu) New email address: 476tjl@ptsmail.ptsem.edu (No NeXTmail) *** Please note: the new address cannot read the 'reply-to' line ***
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: Reading Precompiled Headers Date: 12 Jun 1996 21:24:25 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4pnci9$hgl@news.its.com> References: <4phq3l$bmk@news.mr.net> bmont@rconnect.com (Bruce Montegani) wrote: > Does anyone know how to read a precompiled header file? ie. that is the > structure? The precompiled header files basicly contain a portable dump of the internal data structures of the compiler after all of the header files have been processed normally. Therefore, you could determine what the format is by examining the source to NeXT's version of gcc.... -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator Immutable means "not capable of or susceptible to change"-- an immutable object is an object that cannot be changed (by anything). > Why can't one instance of a class maintain a reference to a mutable > container as a mutable container and every one else reference it > immutably ? I use this all the time. One class maintains and modifies > a collection and every one else references (read only). That's nothing more than using casting to effectively enforce "call by value" argument semantics under Objective-C, instead of objects being passed around via "call-by-reference" as is normal. Casting a mutable object to an immutable object does not mean that the original mutable object actually becomes immutable somehow; it just means that other methods have been told to treat it that way. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: gfa@komsys-pc-gf.ethz.ch (George Fankhauser) Newsgroups: comp.sys.next.programmer Subject: Re: munmap, INN, NS3.3: undefined symbol Date: 15 Jun 1996 23:25:05 +0200 Organization: Swiss Federal Institute of Technology (ETHZ) Message-ID: <4pv9nh$2nb@komsys-pc-gf.ethz.ch> References: <4pmunq$qm2@optical.fiber.net> In article <4pmunq$qm2@optical.fiber.net>, Carl Payne <cpayne@optical.fiber.net> wrote: >ld: Undefined symbols: >_munmap > >References to mmunmap are only scarce, but _mmunmap is downright extinct. > >Does anyone have a clue I can borrow? > > As far as I know, mmap/munmap are system V vm functions. You should be able to replace them with map_fd/vm_deallocate, they're the corresponding mach vm functions... and, be careful with mmap: next's system library has an mmap for the driver kit but it does something completely different. last time I link this one I got an illegal instruction in my i386 box :-) Hope this helps, George -- George Fankhauser Swiss Federal Institute of Technology Phone/Fax +41 1 632 7017/1035 Computer Engineering and Networks http://www.tik.ee.ethz.ch/~gfa Laboratory, CH-8092 Zurich mailto:gfa@tik.ee.ethz.ch "Happyness is a journey, not a destination"
From: shultzb@cvn.net Newsgroups: comp.sys.next.programmer Subject: HELP declaring variables Date: 14 Jun 1996 17:54:17 -0700 Organization: Zippo Message-ID: <4pt1jp$khn@clark.zippo.com> Greetings, most brilliant programmers. I am finding myself in front of a NeXTstation color. I find that I can not declare variables in the shell on this NeXT. I am accustomed to Sperry and Intel Unix machines. I would be eternally greatful, if you would please pass the secret of declaring variables along to me. Many thanks, shultzb@cvn.net
From: Eric_Noyau@next.com (Eric Noyau) Newsgroups: comp.sys.next.programmer Subject: Re: How to trace ObjC messages Date: 16 Jun 1996 00:53:03 GMT Organization: NeXT Software, Inc. Message-ID: <4pvltf$t25@news.next.com> References: <4pteds$nsm@news.next.com> In article <4pteds$nsm@news.next.com> Matt_Watson@NeXT.com (Matt Watson) writes: > In article <4pkmc3$p47@news.mr.net> bmont@rconnect.com (Bruce Montegani) > writes: > > Does anyone know how to trace all messages that are executed when an > > application is running including API kit messages. > > > > (gdb) b objc_msgSend > Breakpoint 1 at 0x60044f0 > (gdb) comm 1 > Type commands for when breakpoint 1 is hit, one per line. > End with a line saying just "end". > >silent > >p *(char **)($sp+8) > >up > >cont > >end > (gdb) r > > (Getting this to work on HP-PA is left as an exercise to the reader...) > Matt, you are giving away corporate secrets ! I have another solution to the same problem: (enoyau@pelane) gdb test GDB is free software and you are welcome to distribute copies of it under certain conditions; type "show copying" to see the conditions. There is absolutely no warranty for GDB; type "show warranty" for details. GDB 4.14 (NEXTSTEP 4.0 --target i386), Copyright 1995 Free Software Foundation, Inc... (gdb) b main Reading in symbols for test_main.m...done. Breakpoint 1 at 0x3b47: test_main.m:9. (gdb) r [...] Breakpoint 1, main (argc=1, argv=0xbffff6f8) at test_main.m:9 test_main.m:9: No such file or directory. (gdb) b objc_msgSend Breakpoint 2 at 0x5002230 (gdb) comm 2 Type commands for when breakpoint 1 is hit, one per line. End with a line saying just "end". >silent >printf "%c[%s %s]\n", (((*((**(int ***)($sp+4)) + 4)) & 2) ? '+' : '-'), (char *)(*((**(int ***)($sp+4)) + 2)), (char *)sel_getName(*(int *)($sp+8)) >cont >end (gdb)cont Continuing. +[NSAutoreleasePool alloc] +[NSAutoreleasePool initialize] +[NSAutoreleasePool allocWithZone:] +[NSThread allocWithZone:] +[NSThread initialize] -[NSThread init] -[NSAutoreleasePool init] +[NSUserDefaults standardUserDefaults] +[NSUserDefaults initialize] +[NSAutoreleasePool allocWithZone:] -[NSAutoreleasePool init] -[NSConstantString getCString:maxLength:] -[NSConstantString length] -[NSConstantString getCString:maxLength:range:remainingRange:] [... etc...] (Getting this to work on HP-PA is still left as an exercise to the reader... ;^) ) -- Eric
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: How to trace ObjC messages Date: 16 Jun 1996 01:26:54 GMT Organization: Anderson Financial Systems Inc. Message-ID: <4pvnsu$90c@shelob.afs.com> References: <4pvltf$t25@news.next.com> Eric Noyau writes > I have another solution to the same problem: Wait, Eric, you forgot something: ----------------------- BEGIN GEEK CODE BLOCK ----------------------- > printf "%c[%s %s]\n",(((*((**(int***)($sp+4))+4))&2)?'+':'-'), > (char*)(*((**(int ***)($sp+4))+2)),(char*)sel_getName(*(int*)($sp+8)) ------------------------ END GEEK CODE BLOCK ------------------------ There. Now it is properly formatted for USENET. Man, I loves dem triple '*'s... -- Gregory H. Anderson | "I wander'd off by myself, In the Crystal Ball/Star Gazer | mystical moist night-air, and from Anderson Financial Systems | time to time, Look'd up in perfect greg@afs.com (NeXTmail OK) | silence at the stars." Walt Whitman
From: lbell@ucsd.edu Newsgroups: comp.sys.next.programmer Subject: Re: Q:Am I using NSException correctly? Date: 16 Jun 1996 05:04:41 GMT Organization: The Avant-Garde of the Now, Ltd. Message-ID: <4q04l9$2fk@news1.ucsd.edu> References: <31B89C30.52BF@dna406.dna.mci.com> Do not use NS_VALRETURN in the HANDLER part. If you want you can just use a plain return createResult; Here is how the exception handling works NS_DURING installs an exception handler and NS_HANDLER removes it. So in order to return anything inbetween NS_DURING and NS_HANDLER you have to use NS_VALRETURN as it also removes the exception handler and then returns a value. The block of code inbetween NS_HANDLER and NS_ENDHANDLER is only jumped to when an exception occurrs. The variable exception is declared in that bloc, and represents the exception that occurred. Using mkdirException is incorrect, as if different exception could occurr and mkdirException wouldn't reflect it. > > NS_ENDHANDLER >} /* end createDir */ > Ideally you can do without any NS_VALRETURN's. - (NSNumber *)createDirName:(NSString *)dirName mode:(NSNumber *)dirMode { NSNumber *createResult; NSException *mkdirException; int mkdirResult; const char *directory = [dirName cString]; int mode = [dirMode intValue]; if ([self debug]) { NSLog (@"Input directory is: %s", directory); NSLog (@"Mode is: %o", mode); } NS_DURING mkdirResult = mkdir (directory, mode); /* Successful? */ if (mkdirResult == -1) { // Create the exception info from errno mkdirException = [NSException exceptionWithName:@"mkdir" reason:[NSString stringWithCString: strerror (errno)] userInfo:nil]; // Raise the exception [mkdirException raise]; } if ([self debug]) { NSLog (@"Result is %d", mkdirResult); } createResult = [NSNumber numberWithInt:mkdirResult]; // If mkdir failed NS_HANDLER NSLog (@"Exception: %s Reason: %s\n", [[exception exceptionName] cString], [[exception exceptionReason] cString]); createResult = [NSNumber numberWithInt:-1]; NS_ENDHANDLER return createResult; } /* end createDir */
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: How to trace ObjC messages Date: 16 Jun 1996 04:58:52 GMT Organization: Netcom Distribution: world Message-ID: <4q04ac$hj3@dfw-ixnews5.ix.netcom.com> References: <4pteds$nsm@news.next.com> <4pvltf$t25@news.next.com> Eric_Noyau@next.com (Eric Noyau) wrote: > >printf "%c[%s %s]\n", (((*((**(int ***)($sp+4)) + 4)) & 2) ? '+' : '-'), > (char *)(*((**(int ***)($sp+4)) + 2)), (char *)sel_getName(*(int > *)($sp+8)) Oh my gawd!! -- Art Isbell NeXT/MIME Mail: aisbell@cubicsol.com Trego Systems Voice/Fax: +1 408 335 2515 CaseServ: NEXTSTEP/OpenStep Voice Mail: +1 408 335 1154 managed care solutions US Mail: Felton, CA 95018-9442
Newsgroups: comp.sys.next.programmer From: paul@cyantic.com (Paul Guiness) Subject: Running EOF under Solaris Organization: CYANTIC Systems Date: Fri, 14 Jun 1996 12:46:33 GMT Message-ID: <1996Jun14.124633.10163@cyantic.com> I have written a small test program that compiles without warnings under Solaris. All the program does is to connect to Sybase through EOF. The program errors with a message about not finding an adaptor. Here is a code fragment and the sample output: // Code fragment. // Get the path of a known model. modelPath = [EOModel findPathForModelNamed:modelName]; // Create the known model. myModel = [[EOModel alloc] initWithContentsOfFile:modelPath]; if (!myModel) { NSLog(@"Unable to find the model %@.\n", modelName); return nil; } NSLog(@"Model name = %@.\n", [myModel name]); NSLog(@"Adaptor Name = %@.\n", [myModel adaptorName]); NSLog(@"Connection DIct. = %@.\n", [[myModel connectionDictionary] description]); myDatabase = [[EODatabase alloc] initWithModel:myModel]; The ouput is: Jun 12 16:58:22 [22238] Model name = A_Model_Name. Jun 12 16:58:22 [22238] Adaptor Name = Sybase. Jun 12 16:58:22 [22238] Connection DIct. = { databaseName = a_database; hostName = A_HOSTNAME; password = a_password; userName = a_user; }. Jun 12 16:58:22 [22238] NSInvalidArgumentException: initWithAdaptor:: invalid nil value for adaptor The program errors on the line myDatabase = [[EODatabase alloc] initWithModel:myModel] A suggestion offered by NeXT online help was to statically link in the adaptor library, so the compile command is: gcc -g TestDataSource.m /usr/lib/libnsl.so.1 /NextLibrary/Adaptors/Sybase.dbadaptor/Sybase -lEOAccess -lFoundation However, this doesn't make a difference. I know this test program works under NeXT OS, but not Solaris. Has anybody experience this problem or have any suggestions? Please respond directly. Thanks. G. Paul Guinness paul@cyantic.com
From: stes@kokomo.wri.com (David Stes) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: How to trace ObjC messages Date: 16 Jun 1996 15:50:53 GMT Organization: Wolfram Research, Inc. Distribution: world Message-ID: <4q1agt$qo5@dragonfly.wolfram.com> References: <4pvltf$t25@news.next.com> > Does anyone know how to trace all messages that are executed when an > application is running including API kit messages. On the NeXT 2.1, I found libPeep.a & AppInspector.app very useful for this (but is this still working ? I haven't been able to use this app since 3.0 for anything else than "inspecting" the objects in the application - not for actually tracing messages. Maybe I'm missing something here...) If you have a Solaris or WindowsNT machine at hand, you could also try using the Stepstone objcc which has an excellent feature for tracing messages. You can turn it on or off from within your program or from within the debugger by the "msgFlag" boolean. (The Stepstone compiler comes with runtime source... here's the relevant code at the start of the C messenger) IMP _imp(Object *recvr, SEL sel) { SHR shr; REGISTER int index; #if defined(NO_TRACEBACK) if (msgFlag)_prnFrame(msgIOD, recvr, sel, NO); #else if (msgFlag)_prnFrame(msgIOD, GETFRAME(recvr), ARGPTR(recvr), NO); #endif if (recvr == (Object *)nil) return (IMP)_nilHandler; if ((shr=recvr->isa) == Nil) return (IMP)_freedHandler;
From: cwolf@filoli.com (Christopher Wolf) Newsgroups: comp.sys.next.programmer,comp.sys.next.marketplace Subject: Seeking NEXTSTEP Device Driver expert for short term contract Date: 16 Jun 1996 19:33:00 GMT Organization: Best Internet Communications Message-ID: <4q1nhc$o89@nntp1.best.com> The company I work for is seeking someone very familiar with writing NeXTSTEP device drivers to perform some short-term (a few weeks probably) one-time contract work in the SF Bay Area. I don't speak for my employer but if you know someone with the necessary experience who may be interested I can put you in contact with the appropriate people. Chris Wolf --
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 17 Jun 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4q2m4e$bv4@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Postgres95 EOAdapter? Date: 15 Jun 1996 13:38:57 GMT Organization: Turbocat's Development, Germany Message-ID: <4puedh$s8@turbocat.snafu.de> Hi! Is there a Postgres95 EOAdapter available? _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: ~/.AppInfo Date: 17 Jun 1996 10:48:58 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4q3d6q$7jh@zdi.informatik.uni-stuttgart.de> References: <Pine.NXT.3.93.960612090909.2125B-100000@capitalist.princeton.edu> <4pretu$je8@zdi.informatik.uni-stuttgart.de> <Pine.NXT.3.93.960614093720.2754B-100000@capitalist.princeton.edu> In <Pine.NXT.3.93.960614093720.2754B-100000@capitalist.princeton.edu> "Timothy J. Luoma" <luomat@capitalist.princeton.edu> writes: >On 14 Jun 1996, Klaus Brouwer wrote: >> >I've seen a few apps lately which have ignored this idea, and that bothers >> >me because, well, it's there, why not use it? I don't remember what apps >> >they were right now, but anyway, I just wanted to mention it. >> >> What do you mean with "lately"? Not even NeXT's own apps (eg. Edit and >> IB) use .AppInfo. And with such a paragon....no wonder...... >I don't know much about IB, but as far as Edit.app and other NeXT apps, >they all use the defaults database for their info. ~/.AppInfo is a place >to store information which is not appropriate for the defaults database, >such as Fiend's dock and shelf information... Edit.app stores files like ".commanddict" and ".pipedict" and the expansion dictionary file (name?) in the users home directory. Hm. Thinking twice about it these are all developer-mode items.... >If your app needs to drop extra files someplace, why not >~/.AppInfo/Appname? I totally agree. But unfortunately the reality is different. Not even all apps sold by Lighthouse use it, though Diagram! does. > >> >Thank you for helping to lower my blood pressure... >> Thank you for highering my blood preasure now. >I do not hope to raise it now, but I don't think you really understood my >original post, which I take 50% blame for.... I think I did :-) I came around the topic a while ago and got angry about NeXT not sticking to their own advices. Reading your post everything came up again. Clear Statement: I think it should be done! Klaus Brouwer, hoping not to cause an other 50% of confusion.
From: 1kyriaki@swt1.informatik.uni-hamburg.de (Valentino Kyriakides) Newsgroups: comp.sys.next.programmer Subject: Searching for NS 2.1 gdb binary... Date: 17 Jun 1996 13:15:34 GMT Organization: University of Hamburg -- Germany Message-ID: <4q3lpm$3c3@rzsun02.rrz.uni-hamburg.de> I need an old NS 2.1 gdb binary for some stuff I want to do. (NS 3.x gdb binaries won't do the job and needless to say, I have these by my own. As I don't have an old NS 2.1 backup, I need some help from the NS black hardware comunity.) So my question is if somebody could send me a gzipped NS 2.1 binary via email? Thanks in advance, Valentino -- Valentino Kyriakides Email: 1kyriaki@informatik.uni-hamburg.de Rugenbergener Str. 27a 25474 Ellerbek Tel.: 04101-35319
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Mon, 17 Jun 1996 17:24:21 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960617164511.7023A-100000@hphalle0.informatik.tu-muenchen.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> <x7g27zq20n.fsf@tom.es.ele.tue.nl> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <x7g27zq20n.fsf@tom.es.ele.tue.nl> On 13 Jun 1996, Pieter Schoenmakers wrote: After resthinking a lot about the problem, I finally have to admit to your code. It is the best approach currently available within the language definition of Objective-C. > In article <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: > P.S.: It is nice, but I rarely see people using the 'const' keyword. > > The `const' keyword is a hack which is rather unusable. As soon as some > function forgets to promise not to modify a string, you'll get compiler > warnings all over the place. And since most code is not under your own > control, you're stuck with it. const does not buy you anything. > I still don't admit here. It's correct, that most libraries accecpt non const parameters, even if they don't change a pointer or it's contents, but that's a problem of the library designers and history. The const keyword was intruduced after the standard libraries were already written. Using const within your own functions is a very good way to force better rethinking about functions etc. We use it all over our projects. After a hard phase on how to deal with the standard libraries we finally cleaned up things in a way which isn't perfect, but makes most things much more clearer. > But with the introduction of mutable/immutable objects they think that > an immutable object is the same as a 'const' mutable object --- which > is wrong. > > An immutable object is an object which does not provide methods which > affect the object's state. If the compiler tells you you're dealing with > an immutable object, then it is immutable. The fact that, at runtime, it > actually might be something different, is irrelevant. As said, casting is > the responsibility of the person at the keybord when the cast appeared in > the source code. Anybody casting through mutability should better have a > very good reason for doing so. If not he should be fired :-) --Tiggr > I was wrong about my thinking of the keyword 'const'. You should consider the inspiration, why it was introduced: to prevent the compiler from changing a constant variable. In the OO world we would need some kind of a const keyword, but on object base. The C 'const' keyword can't provide this mechanism, because the compiler can't figure out, which methods of a class would change and object and should be rejected when used with the const keyword. Interesting idea, isn't it? Distributed Objects intruduced new keywords, how about the following approach (abstract class description) of a new keyword (just an idea): NSString: - initWithCString:(char *)string; - change (NSString)toUpper; One could imagine the followin now: - (ooconst NSString) foo { ooconst NSString *myString=[NSString initWithCString:"BlaBla"]); return return myString; } - bar { ooconst NSString *myString=[self foo]; NSString mySecondString=[NSString initWithCString:"Blubber"]); [myString toUpper]; // fails, myString is ooconstant [mySecondString toUpper]; // succeeds, not constant. } A lot of blabla you might think, and what are the benefits. Ok, although I'm not too good in explaining theoretics in english, I'll try it: It's a more clear approach. NSMutableString and NSString are two different classes and _should_not_ get casted. E.g. NSMutableString _should_not be returned by some method as an NSString, because it simply isn't. Using ooconst as a new keyword would instructs the compiler that the user of an ooconst class is not allowed to use the methods which might change the state of an object. What's the difference to the Mutable/Immutable concept? Returning a casted value always triggers side effects. One might think: "Hey, I've got an immutable object, so it's contents will never change!". In contrast if he receives an ooconst object he will think "Hey, I've got an object which I'm not allowed to change, but it will probably change over the time when I'm finished using it". Different isn't it? By returning an immutable object you simple indicate an constant object, which it isn't. Using this concept you are responsible by very carefully documenting and designing your methods. Ok, I admit, this was a lot of blabla. I got your point, and maybe you got mine. I was wrong, Tiggr was right. Anyway I still believe there should be better concepts available. The introduction of NeXT's mutable classes indicates that there is a need for better security in access methods. But on the other hand, we are the programmers and get paid for our thinking not to do worse things with our classes and class access method. Best regards, Boerny. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 17 Jun 1996 15:20:48 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4q3t4g$np9@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> Mark Anenberg <marka@Eng.Sun.COM> wrote: > well I don't speak officially for Sun, but... > Its not a matter of stopping short; its a matter of a conscious limiting effort; It doesn't make sense to send email in a proprietary format when the rest of the world speaks Mime/X.400/X.500. The whole point is information exchange. That doesn't exactly make sense. If the person can choose to send MIME anyway, nothing is limited except the users ability to send in a richer format? This sounds like something a bit more political than anything else. -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Mon, 17 Jun 1996 17:33:27 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> <x7g27zq20n.fsf@tom.es.ele.tue.nl> <4prfo9$lco@zdi.informatik.uni-stuttgart.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4prfo9$lco@zdi.informatik.uni-stuttgart.de> On 14 Jun 1996, Klaus Brouwer wrote: > >from anything to id (everything), is performed under the responsibility of > >the person typing that cast. Any bugs in code which seemed legal to the > >compiler as a result of such a cast is, again, the responsibility of the > >person that typed the cast. > RIGHT, RIGHT, RIGHT! I hope you post will put this discussion to a reasonable > end. > I admit. > > P.S.: It is nice, but I rarely see people using the 'const' keyword. > Anyway: if you have a fooFunction(const char *str); and fooFunction caches > str in some static variable, just NOBODY guarantess you that the creator of > str doesn't modify it in some way, since it might be declared as char *str. > As you said: YES, YES, YES, that's exactly the problem returning an casted NSMutableArray as an NSArray. NOBODY guarantees you that the creator of the NSArray doesn't modify it in some way. That's why you shouldn't use casting. See also my other posting. The const keyword, just indicates you that _you_ are not allowed to change it. Returning an immutable object suggests that you and others (even the creator) can't change it. Have a nice day, Boerny. -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: window partitioning widget Date: 17 Jun 1996 17:54:24 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4q464g$jci@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hello, is there any way to program the small spherical widget used to dynamically resize the workspace window in two horizontal parts? I am certain the question was already made, but can get no FAQ reference about it... ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: programming help links Date: 17 Jun 1996 18:26:46 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4q4816$jci@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hello, I am looking for documentation and examples for programming help (hypertext) links within Text objects. The documentation on NXHelpPanel says the Text class supports this functionality, but I don't see anything explicit within the Text class documentation ... I must be missing something... Any hints? ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
From: uhendjx@meaddata.com (Jonathan W. Hendry) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Followup-To: comp.sys.next.programmer,comp.sys.next.advocacy Date: 17 Jun 1996 18:32:07 GMT Organization: Lexis-Nexis, Dayton OH Message-ID: <4q48b7$q7g@mailgate.lexis-nexis.com> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4q3t4g$np9@news3.digex.net> John Kheit (jkheit@cnj.digex.net) wrote: : Mark Anenberg <marka@Eng.Sun.COM> wrote: : > well I don't speak officially for Sun, but... : > Its not a matter of stopping short; its a matter of a conscious limiting : effort; It doesn't make sense to send email in a proprietary format when : the rest of the world speaks Mime/X.400/X.500. The whole point is : information exchange. : That doesn't exactly make sense. If the person can choose to send MIME : anyway, nothing is limited except the users ability to send in a richer : format? This sounds like something a bit more political than anything : else. No more political than a decision to not support, say, the Apple Double file format for file attachments. Sure, some people might find it useful, and not supporting it may be considered 'limiting the users ability to send in a richer format'. But it's not a political decision. It's 1996. The number of NeXTMail users is dwindling fast. Why bother including NeXTMail? If NeXT was going to sell NeXTMail on Windows, there'd be a better reason for including NeXTMail. If someone wrote a Eudora-esque mail app for Windows that could send NeXTMail, there might be a reason for it. As things stand? Nope. -- Jonathan W. Hendry Views expressed herein do Steel Driving Software, Inc. not represent those of steeldrv@ix.netcom.com Steel Driving Software, Inc. jon@exnext.com or Lexis-Nexis
From: John Kheit <jkheit@cnj.digex.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Q:Sound[Kit] & OPENSTEP? Date: 17 Jun 1996 18:46:58 GMT Organization: monoChrome, Inc., NJ, USA Message-ID: <4q4972$p0b@news3.digex.net> References: <4n0bss$s58@news4.digex.net> <4ncs66$k2j@nntp.snfc21.pbi.net> <31A4A60E.2B41@Eng.Sun.COM> <4o4k32$lv0@bignews.shef.ac.uk> <4o7ubj$ud@tallowcross.uni-frankfurt.de> <31A8CDD4.7D3C@ix.netcom.com> <31AF1BF3.4ECA@Eng.Sun.COM> <4optvk$19t@tallowcross.uni-frankfurt.de> <31B31D9B.2BAE@Eng.Sun.COM> <4q3t4g$np9@news3.digex.net> <4q48b7$q7g@mailgate.lexis-nexis.com> uhendjx@meaddata.com (Jonathan W. Hendry) wrote: > : That doesn't exactly make sense. If the person can choose to send MIME > : anyway, nothing is limited except the users ability to send in a richer > : format? This sounds like something a bit more political than anything > : else. > No more political than a decision to not support, say, the Apple Double file format for file attachments. Sure, some people might find it useful, and not supporting it may be considered 'limiting the users ability to send in a richer format'. But it's not a political decision. It's 1996. The number of NeXTMail users is dwindling fast. Why bother including NeXTMail? Because it is NO bother. And that makes no sense. No one is in essance 'cloning' MacOS the way NeXTSTEP3.3 is practically being clone by SUN. If you want to make everyone comfortable, and it takes no extra effort to enable sending, then there is no reason left but politics...They say so, so it must be THE right way. Where have I heard that 'not invented here' philosophy before? -- Thanks, be well, take care, later, John Kheit )^> %^) =^) monoChrome, Inc. | New York Law School NEXTSTEP Developer | Opinions expressed represent me only MIME, SUN, & NeXTmail OK | http://cnj.digex.net/~jkheit mailto:jkheit@cnj.digex.net | Telepathy...It's coming...
From: ngervae@sirius.com Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 17 Jun 1996 21:34:17 GMT Organization: NeXT Software, Inc. Message-ID: <4q4j0p$b0n@news.next.com> References: <31b74471.157975@news.plsys.co.uk> In article <31b74471.157975@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: > What seems sensible to me is that if a method states it returns an > NSArray, i.e. an immutable class, it _will_ return an immutable array. > Syntatically it's valid to return NSArray's mutable subclass; but I > think methods which return classes which recognise the mutability > concept should be treated as special cases, i.e. the immutability will > be honoured as well as the object really being an instance of the > stated class. > Double-speak. Instances of immutable classes should be immutable, > period. Exactly what is the point of FK introducing the immutability > concept if it's not guaranteed that what you think is an immutable > object won't change? (Pardon the triple negative!) This sounds to me like a C++ programmer speaking. :-) I think the present confusion has more to do with the documentation than with the Foundation Framework's design. Being one of the principal documentors of the Framework, I can only claim my share of guilt. Maybe these notes of the moment will help a bit. In ObjC, declaring an object with a particular class means that the object can be an instance of that class or of any subclass. There are NO guarantees that this is not the case, with regard to mutability or other characteristics, unless you actually created the object yourself using alloc/init or copy. Ownership is a more significant indicator than class declaration in Foundation. Use of the term "immutable" is a documentation construct, not a formal part of the API. We should have clarified this concept much more than we did, as it now turns out. The distintion was made primarily to assist in compile-time type checking; run-time checking, to my mind (NOT necessarily the designer's), is a nice side benefit, but not one to be relied on. Again: There is NO formal guarantee that an object declared as NSArray won't change, unless YOU created that object as an NSArray. A non-mutable-by-the-recipient typing (how's that for obfuscation?) merely means that the recipient shouldn't change the object. It makes no claims about the owner's ability to change that object. If you aren't the owner, and you want to be sure an object you get from somewhere else is in fact immutable, it falls upon you, not the provider, to make it immutable, whether by shallow or deep copy. Only you, the recipient, know what you want. Consider these two cases: You want a continual up-to-date picture of a received object's state, vs. you want a static snapshot guaranteed not to change. Which must be the default case to enable *you* to have it either way? If the provider gives you an immutable copy of its mutable object, you're stuck with that or with additional API on the provider to get a separate, mutable-but-not-by-you copy. On the other hand, if the provider gives you its own (possibly mutable) object, you can copy it if you need to (which is commonly necessary only for simple value classes: NSString, NSDate, NSNumber). A fair amount of the confusion here also comes from our documentation on How to Copy Objects being hidden in the framework directory. You might want to check the NSCopying protocol for some nitty-gritty on implementing copy methods. As far as getting the results you want: - For a really shallow, no-guarantees approach, just retain the object. - For a shallow, truly immutable copy, use alloc/initWith... rather than relying on the provider or using the copy method. (This is something we should've made much more prominent in the documentation.) - For a deep, truly immutable copy, use copy. If the receiving class doesn't give you a deep, truly immutable copy, that's an error in the class implementation. (And remember, since immutable values can implement copy by retaining, this isn't always as expensive as it might seem.) The above applies only to classes that make the nonmutable/mutable distinction. Copying an NSView doesn't produce a nonmutable NSView, for example, nor does it necessarily produce a completely deep copy. > Next's production code seems to vindicate this to my mind; to the best > of my knowledge, all the EOF methods ([EOEntity attributes], for > example) really do return an immutable array. Like you said, you don't > speak for Next. I wouldn't count on this. Again, if you want true immutability in an object you don't own, you have to make your own copy. I'm entering a bug report on this confusing issue, and will see what I can do to clarify it in the documentation for the next release. Until then, save this article. :-) -- Nik Gervae <ngervae@sirius.com> Technical Writer NeXT Software, Inc.
From: besler@mdd.comm.mot.com (Steve Besler) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 17 Jun 1996 16:40:51 -0700 Organization: Motorola - Wireless Data Group Distribution: na Message-ID: <4q4qe3$ve@df13h.mdd.comm.mot.com> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> In article 100000@hphalle0.informatik.tu-muenchen.de, Bernhard Scholz >As you said: YES, YES, YES, that's exactly the problem returning an >casted NSMutableArray as an NSArray. NOBODY guarantees you that the >creator of the NSArray doesn't modify it in some way. That's why you >shouldn't use casting. See also my other posting. The const keyword, just >indicates you that _you_ are not allowed to change it. Returning an >immutable object suggests that you and others (even the creator) can't >change it. But that's Just it! Notice that NSArray is *NOT* named NSImmutableArrary or NSConstantArray! This is because the NSArray *interface* makes no claim of being "immutable", it is simply the interface that all the members of the NSArray class cluster (mutable and immutable) implement. I don't really get why people don't understand this. Maybe I'm missing the point, but I doubt it... Cheers, Steve -- Steven Besler, Coop Student "I do not speak for Motorola." Motorola Wireless Data Group besler@mdd.comm.mot.com
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 18 Jun 1996 15:46:54 GMT Organization: Norden 1 Communications Message-ID: <4q6j1e$ral@tofu.alt.net> Programmer/analyst NEXTSTEP----------Commercial experience Objective C-------Commercial experience EOF---------------A plus Sybase or Oracle--A plus Career Position---Full Benefits++ Opportunity-------Exceptional Relocation--------Company assistance To Be Considered--Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: tremblay_michel@tandem.com (Pierre-Michel Tremblay) Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc Subject: Help: Screaming cc compiler using Lex&Yacc. Date: 18 Jun 1996 17:57:31 GMT Organization: Tandem Computers Message-ID: <4q6qmb$af@news.mpd.tandem.com> Bonjour all, I'm a novice interested in using Lex and Yacc. I went on the tutorial located at URL: http://www.ee.ic.ac.uk/course/advanced/lex/lex.html. I've tried to use their tutorial without success. cc screams at me and I don't really know what's necessary to get it to run properly. Here are the very simple steps I followed: ********** Step 1. ********** I've modified the makefile to support the cc compiler; not acc. I also took care of replacing the 8 spaces by a tab. Hence makefile at http://www.ee.ic.ac.uk/course/advanced/lex/what.html: setup % cat > makefile CC= cc CFlags= -w LdFlags= -ll DEMO: $(Project).o $(CC) $(CFlags) -o DEMO $(Project).o $(LdFlags); make last $(Project).o: $(Project).c $(CC) $(CFlags) -c $(Project).c; make last $(Project).c: $(Project).l lex -t $(Project).l > $(Project).c; make last last: # displays most recent file ls -lt | head -2 | tail -1 clean: # removes intermediate files rm *.o *.c .SILENT: ********** Step 2. ********** Then, I've wrote this small example and replaced the spaces with tabs: setup % cat > tr2.l %% [A-Z] printf("%c", yytext[0]+'a'-'A'); . ECHO; ********** Step 3. ********** Now. Could anyone tell me what's wrong with this? What are the flags I should use? % make Project=tr2 -rw-r--r-- 1 pmt 6853 Jun 18 13:47 tr2.c /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: illegal external declaration, found `.' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: undefined type, found `W' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: illegal external declaration, missing `;' after `h' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: undefined type, found `W' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: illegal external declaration, missing `;' after `h' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:2: illegal method selector, found ` diablo 12 pitch .ds L' /NextDeveloper/Headers/ansi/i386/stdtypes.h:3: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:4: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:5: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:6: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:7: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:8: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:9: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:11: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:12: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:13: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:14: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:15: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:16: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:17: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:17: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:18: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:19: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:20: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:21: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:22: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:23: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:24: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:25: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:26: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:27: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:27: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:28: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:29: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:30: warning: ANSI C forbids newline in character constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:32: warning: ANSI C forbids newline in character constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:33: warning: ANSI C forbids newline in character constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: warning: numeric constant contains digits beyond the radix /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: illegal method definition, missing `{' after `SMI' /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: undefined type, found `from' /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: illegal external declaration, missing `;' after `UCB' /NextDeveloper/Headers/ansi/i386/stdtypes.h:37: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:37: illegal method definition, found `mark' /NextDeveloper/Headers/ansi/i386/stdtypes.h:39: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:40: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:42: warning: nondigits in number and not hexadecimal /NextDeveloper/Headers/ansi/i386/stdtypes.h:43: warning: nondigits in number and not hexadecimal /NextDeveloper/Headers/ansi/i386/stdtypes.h:44: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:45: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:46: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: warning: nondigits in number and not hexadecimal /NextDeveloper/Headers/ansi/i386/stdtypes.h:11: illegal method definition, missing `{' after `ptrdiff_t' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: illegal external declaration, found `.' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: undefined type, found `W' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: illegal external declaration, missing `;' after `h' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: undefined type, found `W' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: illegal external declaration, missing `;' after `h' /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:1: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:2: illegal method selector, found ` diablo 12 pitch .ds L' /NextDeveloper/Headers/ansi/i386/stdtypes.h:3: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:4: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:5: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:6: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:7: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:8: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:9: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:11: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:12: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:13: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:14: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:15: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:16: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:17: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:17: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:18: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:19: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:20: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:21: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:22: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:23: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:24: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:25: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:26: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:27: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:27: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:28: warning: ANSI C forbids newline in string constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:29: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:30: warning: ANSI C forbids newline in character constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:31: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:32: warning: ANSI C forbids newline in character constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:33: warning: ANSI C forbids newline in character constant /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: warning: numeric constant contains digits beyond the radix /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: illegal method definition, missing `{' after `SMI' /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: undefined type, found `from' /NextDeveloper/Headers/ansi/i386/stdtypes.h:36: illegal external declaration, missing `;' after `UCB' /NextDeveloper/Headers/ansi/i386/stdtypes.h:37: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:37: illegal method definition, found `mark' /NextDeveloper/Headers/ansi/i386/stdtypes.h:39: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:40: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:42: warning: nondigits in number and not hexadecimal /NextDeveloper/Headers/ansi/i386/stdtypes.h:43: warning: nondigits in number and not hexadecimal /NextDeveloper/Headers/ansi/i386/stdtypes.h:44: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:45: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:46: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: stray '\' in program /NextDeveloper/Headers/ansi/i386/stdtypes.h:47: warning: nondigits in number and not hexadecimal /NextDeveloper/Headers/ansi/i386/stdtypes.h:22: illegal method definition, missing `{' after `)' *** Exit 1 Stop.
From: steel456@aol.com (Steel456) Newsgroups: comp.sys.next.programmer Subject: NeXT Developers needed Date: 18 Jun 1996 14:38:29 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4q6t35$9l6@newsbf02.news.aol.com> We are the Premier search and recruiting firm for NeXT Professionals in the USA. We have a number of exclusive clients who have retained our services to search for the best permanent NeXTSTEP, Objective-C and strong OO Architects on the market.We are now recruiting for positions from coast to coast from Developers to Architects.We have critical needs for 2 Architects and 2 Senior Software Engineers If you have experience developing within the NeXT environment and are considering looking to leverage your career, now is the time and we are the firm to talk to. Please call us at: Corporate Resource Associates Bruce Rennert 800-662-9797 Fax 800-814-3100
From: bmont@rconnect.com (Bruce Montegani) Newsgroups: comp.sys.next.programmer Subject: Re: Text and textDidChange: Date: 18 Jun 1996 20:46:10 GMT Organization: Minnesota Regional Network (MRNet) Message-ID: <4q74ii$f7d@news.mr.net> References: <4p022l$49q@solaris.cc.vt.edu> <jJAKxBAkNSwxEw4P@cedar.co.uk> >There is a private Text method _replaceSel::::::: that gets invoked to >update the text contents. >You can override it in a subclass to handle your requirements: I'd be very careful subclassing a private method. Their is no way of knowing whether NeXT will change this method in the future. Bruce Montegani
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Text and textDidChange: Message-ID: <Dt7wv4.50o@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4p022l$49q@solaris.cc.vt.edu> <jJAKxBAkNSwxEw4P@cedar.co.uk> <4q74ii$f7d@news.mr.net> Date: Tue, 18 Jun 1996 22:52:16 GMT In article <4q74ii$f7d@news.mr.net> bmont@rconnect.com (Bruce Montegani) writes: > >I'd be very careful subclassing a private method. Their is no way of knowing >whether NeXT will change this method in the future. > I agree that it's risky to do that, but on the other hand, in these days using public methods only, is hardly a guarantuee that you will not have to change your software either...
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: stripping EOF-app? Date: 18 Jun 1996 22:46:36 GMT Organization: Turbocat's Development, Germany Message-ID: <4q7bkc$71r@turbocat.snafu.de> Hello! The app, I wrote works fine but I wanted to strip it to make it smaller. That worked for me fine on non-eof apps. But when I strip the EOF app: dave@turbocat> ./OPPlan Jun 19 00:40:06 [7107] Unable to load extended character set(s); using builtin. Error loading /NextLibrary/Adaptors/Sybase.dbadaptor/Sybase rld(): Undefined symbols: .objc_class_name_EOAdaptor .objc_class_name_EOAdaptorChannel .objc_class_name_EOAdaptorContext .objc_class_name_EOAttribute .objc_class_name_EOEntity .objc_class_name_EOJoin .objc_class_name_EOModel .objc_class_name_EONull .objc_class_name_EOQualifier .objc_class_name_EOQuotedExpression .objc_class_name_EORelationship .objc_class_name_EOSQLExpression .objc_class_name_NSArray .objc_class_name_NSAutoreleasePool .objc_class_name_NSCalendarDate .objc_class_name_NSData .objc_class_name_NSException .objc_class_name_NSMutableArray .objc_class_name_NSMutableDictionary .objc_class_name_NSMutableString .objc_class_name_NSNumber .objc_class_name_NSObject .objc_class_name_NSString .objc_class_name_NSUnarchiver .objc_class_name_NSValue .objc_class_name_NXConstantString .objc_class_name_Object .objc_class_name_Panel _NSDefaultMallocZone _NSInternalInconsistencyException _NSLog _NSZoneMalloc _NXApp _NXCloseMemory _NXCloseTypedStream _NXGetDefaultValue _NXGetMemoryBuffer _NXOpenMemory _NXOpenTypedStream _NXReadObject _NXRunAlertPanel _NXWriteRootObject __NSAddHandler __NSRemoveHandler __ctype_ __filbuf __flsbuf __iob __setjmp _abort _access _atoi _close _connect _ctime _errno _exit _fclose _fcntl _fflush _fgets _fopen _fprintf _fread _free _fscanf _fseek _ftell _fwrite _getdtablesize _getenv _geteuid _gethostbyname _gethostname _getpid _getpwnam _getpwuid _gettimeofday _lseek _malloc _memchr _memcmp _memcpy _memset _objc_lookUpClass _objc_msgSend _objc_msgSendSuper _object_getClassName _open _printf _read _sel_getName _select _send _setsockopt _sigblock _signal _sigsetmask _sleep _socket _sprintf _stat _strcat _strchr _strcmp _strcpy _strlen _strncmp _strncpy _strrchr _strtod _strtol _sys_errlist _sys_nerr _vfprintf _write Jun 19 00:40:11 [7107] *** Uncaught exception: _newDataBaseWithModel:databaseName:: EOModel file:"/Users/dave/projects/op/OPPlan/OPPlan.app/GUSTAV.eomodel" specifies nonexistent adaptor with name "Sybase" exiting! The adaper is there! (NS 3.3 PL 1 EOF 1.1) ??????? _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Tue, 18 Jun 1996 23:39:11 GMT Organization: Orbital Computer Consultancy Message-ID: <31c73d30.6172470@news.plsys.co.uk> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> besler@mdd.comm.mot.com (Steve Besler) wrote: >In article 100000@hphalle0.informatik.tu-muenchen.de, Bernhard Scholz > >>As you said: YES, YES, YES, that's exactly the problem returning an >>casted NSMutableArray as an NSArray. NOBODY guarantees you that the >>creator of the NSArray doesn't modify it in some way. That's why you >>shouldn't use casting. See also my other posting. The const keyword, just >>indicates you that _you_ are not allowed to change it. Returning an >>immutable object suggests that you and others (even the creator) can't >>change it. > >But that's Just it! Notice that NSArray is *NOT* named NSImmutableArrary >or NSConstantArray! This is because the NSArray *interface* makes no >claim of being "immutable", it is simply the interface that all the members >of the NSArray class cluster (mutable and immutable) implement. > >I don't really get why people don't understand this. Maybe I'm missing the >point, but I doubt it... > Here's a quote from the Next docs: "NSArray objects manage arrays of objects. The cluster's two public classes, NSArray and NSMutableArray, declare the programmatic interface for static and dynamic arrays, respectively. The objects you create using these classes are referred to as array objects. Because of the nature of class clusters, array objects are not actual instances of the NSArray or NSMutableArray classes but of one of their private subclasses. Although an array object's class is private, its interface is public, as declared by these abstract superclasses, NSArray and NSMutableArray" In other words, the NSArray and MSMutableArray classes _define_ the interface, one immutable the other mutable. Here's an example gdb session, which shows that arrays created with NSArray don't respond to addObject:, and so are immutable: (gdb) set $obj=[NSNumber numberWithInt:1] (gdb) set $mutableArray = [NSMutableArray array] (gdb) set $immutableArray = [NSArray array] (gdb) po [[$mutableArray class] description] NSConcreteMutableArray (gdb) set [$mutableArray addObject:$obj] (gdb) po [[$immutableArray class] description] NSConcreteArray (gdb) set [$immutableArray addObject:$obj] Target does not respond to method.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Tue, 18 Jun 1996 23:39:15 GMT Organization: Orbital Computer Consultancy Message-ID: <31c73d49.6197959@news.plsys.co.uk> References: <31b74471.157975@news.plsys.co.uk> <4q4j0p$b0n@news.next.com> ngervae@sirius.com wrote: >In article <31b74471.157975@news.plsys.co.uk> k_harbour@mail.bogo.co.uk >(Karl Harbour) writes: >> What seems sensible to me is that if a method states it returns an >> NSArray, i.e. an immutable class, it _will_ return an immutable array. >> Syntatically it's valid to return NSArray's mutable subclass; but I >> think methods which return classes which recognise the mutability >> concept should be treated as special cases, i.e. the immutability will >> be honoured as well as the object really being an instance of the >> stated class. > >> Instances of immutable classes should be immutable, >> period. Exactly what is the point of FK introducing the immutability >> concept if it's not guaranteed that what you think is an immutable >> object won't change? (Pardon the triple negative!) > >This sounds to me like a C++ programmer speaking. :-) I think the >present confusion has more to do with the documentation than with >the Foundation Framework's design. Being one of the principal >documentors of the Framework, I can only claim my share of guilt. > Well, I think the NSArray documentation is fine. For me, there are two significant errors in the Foundation documentation: 1) NSObject.rtf talking about NSObject's non-existent copy method 2) IntroToFoundation's Object Ownership and Automatic Disposal section, which I think is totally misleading. For example, the statement: "The Foundation Kit therefore sets this policy: If you create an object you alone are responsible for releasing it." Clearly, the objects responsible for releasing an object are those that have retained it. Creating an object merely implicitly retains it, as the documentation says elsewhere, so the creator's responsibility with regards to releasing the created object is no different to any other object that retains the created object. >Maybe these notes of the moment will help a bit. > >In ObjC, declaring an object with a particular class means that >the object can be an instance of that class or of any subclass. >There are NO guarantees that this is not the case, with regard to >mutability or other characteristics, unless you actually created >the object yourself using alloc/init or copy. Ownership is a more >significant indicator than class declaration in Foundation. > Isn't the point of Foundation's (auto)release scheme to do away with the concept of ownership? >Use of the term "immutable" is a documentation construct, not a >formal part of the API. Isn't the fact NSArray does not provide an addObject: etc method enough to classify it as having an API which makes it immutable? >We should have clarified this concept much >more than we did, as it now turns out. The distintion was made >primarily to assist in compile-time type checking; run-time >checking, to my mind (NOT necessarily the designer's), is a nice >side benefit, but not one to be relied on. Again: There is NO >formal guarantee that an object declared as NSArray won't change, >unless YOU created that object as an NSArray. > In that case, [NSArray copy] should only be implemented as [NSArray retain] if the caller is the same as the object that created the array. But doesn't it in fact disregard who the caller is? >A non-mutable-by-the-recipient typing (how's that for obfuscation?) >merely means that the recipient shouldn't change the object. It >makes no claims about the owner's ability to change that object. >If you aren't the owner, and you want to be sure an object you get >from somewhere else is in fact immutable, it falls upon you, not >the provider, to make it immutable, whether by shallow or deep copy. >Only you, the recipient, know what you want. > I was considering the provider's requirements, not the recipient's. It was your colleague, Charles D'Harcourt, that brought the recipient into it. <-- snip --> > >A fair amount of the confusion here also comes from our documentation >on How to Copy Objects being hidden in the framework directory. You >might want to check the NSCopying protocol for some nitty-gritty on >implementing copy methods. As far as getting the results you want: I have checked the NSCopying protocol. > >- For a really shallow, no-guarantees approach, just retain > the object. >- For a shallow, truly immutable copy, use alloc/initWith... > rather than relying on the provider or using the copy method. > (This is something we should've made much more prominent in the > documentation.) >- For a deep, truly immutable copy, use copy. If the receiving > class doesn't give you a deep, truly immutable copy, that's > an error in the class implementation. (And remember, since > immutable values can implement copy by retaining, this isn't > always as expensive as it might seem.) > One must be wary of NSArrays which contain objects which don't conform to NSCopying. >The above applies only to classes that make the nonmutable/mutable >distinction. Copying an NSView doesn't produce a nonmutable NSView, >for example, nor does it necessarily produce a completely deep copy. > So is the nonmutable/mutable distinction just a documentation construct or not? > >> Next's production code seems to vindicate this to my mind; to the best >> of my knowledge, all the EOF methods ([EOEntity attributes], for >> example) really do return an immutable array. Like you said, you don't >> speak for Next. > >I wouldn't count on this. Again, if you want true immutability in >an object you don't own, you have to make your own copy. > I retracted this statement in another post. However, given everything that's been said, I would like to know why EOModeler's create class template function creates ivars for to-many relationships which are NSArrays, and not NSMutableArrays. I haven't checked whether EODatabaseChannel actually uses NSArrays for relationships, but presumably it does. >I'm entering a bug report on this confusing issue, and will see >what I can do to clarify it in the documentation for the next >release. Until then, save this article. :-) > Do you mean the release after 4.0? Surely you're going to have to be quick to get new documentation onto the 4.0 CD-ROMs? Or do you know something we don't ;-)
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 19 Jun 1996 10:51:49 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4q8m45$qr9@zdi.informatik.uni-stuttgart.de> References: <31b74471.157975@news.plsys.co.uk> <4q4j0p$b0n@news.next.com> In <4q4j0p$b0n@news.next.com> ngervae@sirius.com writes: >In ObjC, declaring an object with a particular class means that >the object can be an instance of that class or of any subclass. >There are NO guarantees that this is not the case, with regard to >mutability or other characteristics, unless you actually created >the object yourself using alloc/init or copy. Ownership is a more >significant indicator than class declaration in Foundation. >Use of the term "immutable" is a documentation construct, not a >formal part of the API. We should have clarified this concept much >more than we did, as it now turns out. The distintion was made >primarily to assist in compile-time type checking; run-time >checking, to my mind (NOT necessarily the designer's), is a nice >side benefit, but not one to be relied on. Again: There is NO >formal guarantee that an object declared as NSArray won't change, >unless YOU created that object as an NSArray. >A non-mutable-by-the-recipient typing (how's that for obfuscation?) >merely means that the recipient shouldn't change the object. It >makes no claims about the owner's ability to change that object. >If you aren't the owner, and you want to be sure an object you get >from somewhere else is in fact immutable, it falls upon you, not >the provider, to make it immutable, whether by shallow or deep copy. >Only you, the recipient, know what you want. I agree to every sentence :-) ....with a little exception: I think the people got confused about the topic because they WANT some mechanism that guarantess immutability and thought it would be provided by FK. But this is not a FK problem and even an Objective-C problem, it's a problem of the underlying C language in general. Intuitivly the following should work: static const int M = 10; static const int MM = M*M; Unfortunately the compilers opinion is different.... Again, just nobody guarantees you constants or immutable buffers or objects or whatever. >Nik Gervae <ngervae@sirius.com> >Technical Writer >NeXT Software, Inc. Have a nice day, Klaus Brouwer her NSArray is immutable since it's subclasses need not to be! A subclass might overwrite every method and inherit quite nothing but it is still a legal subclass. There is just no way of preventing somebody to write a mutable subclass of an immutable superclass - and this means every programmer has to anticipate such a case. >Have a nice day, Hey, I wanted to say that.... Have an even nicer day, Klaus Brouwer :-)
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 19 Jun 1996 10:53:20 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> In <31c73d30.6172470@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: >besler@mdd.comm.mot.com (Steve Besler) wrote: >Here's a quote from the Next docs: [munch] >In other words, the NSArray and MSMutableArray classes _define_ the >interface, one immutable the other mutable. >Here's an example gdb session, which shows that arrays created with >NSArray don't respond to addObject:, and so are immutable: [example deleted] This is correct but it doesn't matter. What matters here is that you can pass an NSMutableArray instance to an NSArray parameter - this is not only legal but essential to OOP as I've explained already in two previous posts. You, as the receiver on a NSArray, may not change the object! But of course the real owner of object may. If you want a really immutable instance you have to make a copy or the like and become yourself owner of an instance. Once again: an objects declared as NSArray just states that it is *at least* a NSArray but might in fact be an instance of any subclass. Have a nice day, Klaus Brouwer
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Wed, 19 Jun 1996 15:57:53 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> On 19 Jun 1996, Klaus Brouwer wrote: > This is correct but it doesn't matter. What matters here is that you can pass > an NSMutableArray instance to an NSArray parameter - this is not only legal > but essential to OOP as I've explained already in two previous posts. > > You, as the receiver on a NSArray, may not change the object! But of course > the real owner of object may. If you want a really immutable instance you > have to make a copy or the like and become yourself owner of an instance. > > Once again: an objects declared as NSArray just states that it is *at least* a > NSArray but might in fact be an instance of any subclass. > In your previous postings, you did 3 statements which I'd like to rewrite 1) 'const' gives you nothing 2) the people need constant objects which FK can't provide 3) casting upwards the inheritance tree is _always_ save. For point 1) see 2) *grin* 2) That's the point. We want/need objects which are immutable. I already suggested the introduction of an 'ooconst' keyword, which could instruct the compiler to restrict an mutable object at runtime to become an immutable object. The logic for doing this has to be implemented in the compiler by using new keywords. I have no hope to see this in the near future. 3) This isn't true. Imagine a method which collects mutable and immutable objects in a single array and after collecting all objects it thinks: "Let's change all mutable objects I own by performing mutable methods. The immutable objects I collected won't get affected". That's a big mistake, and it's hard, and even very time consuming to find such an error. This was my assumption when stating 'casting is always dangerous'. Have a sunny day, Boerny :-) -- Bernhard Scholz (IRC: Boerny) scholz@informatik.tu-muenchen.de http://peanuts.leo.org/ scholz@ve1.rm.op.dlr.de http://www.leo.org/~scholz/
From: mpaque@pacbell.net (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: stripping EOF-app? Date: Wed, 19 Jun 1996 19:43:48 GMT Organization: Electronics Service, Unit No. 16 Message-ID: <4q9ld9$mpp@news.snfc21.pacbell.net> References: <4q7bkc$71r@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) wrote: >Hello! >The app, I wrote works fine but I wanted to strip it to make it smaller. That >worked for me fine on non-eof apps. But when I strip the EOF app: >dave@turbocat> ./OPPlan >Jun 19 00:40:06 [7107] Unable to load extended character set(s); using >builtin. >Error loading /NextLibrary/Adaptors/Sybase.dbadaptor/Sybase >rld(): Undefined symbols: >.objc_class_name_EOAdaptor > etc.... What's changed on you is that your EOF app wants to load a bit of code at runtime, the database adaptor. For this to work properly, the runtime linker rld() needs to be able to find the symbols for items used by the loadable code module. You can partially strip your application, removing debugging, relocation, and local symbols, but leaving globals. Use the strip options -S -x. In ProjectBuilder driven projects, yoc can set these in the Makefile.postamble. Look for the makefile variable APP_STRIP_OPTS, uncomment, and set as appropriate. Mike Paquette I don't speak for my employer, and they don't speak for me. mpaque@pacbell.net Personal E-mail mpaque@next.com NeXT business mail only, please
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Wed, 19 Jun 96 21:56:50 +0200 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9606191956.AA12255@flexus> Subject: What is _NEXT_SOURCE? Dear experts, I'm building a project using gcc-2.7.2 and libg++-2.7.1 to have template support (and also make-3.74 because /bin/make consistently crashed on my NeXTstation running NEXTSTEP_3.2), but it stopped because it did not see the declaration of struct stat in <sys/stat.h>. I #define'd _NEXT_SOURCE in the program source, because struct stat would only be visible #if defined(_POSIX_SOURCE) || defined(_NEXT_SOURCE), and I had read enough about the troubles that can be caused if _POSIX_SOURCE is defined. What's the deal here? The build is now continuing, but can I trust that all is well? Why haven't I had any problem with this until now (I have compiled using struct stat before)? Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be (NeXTmail), Flanders, Belgium If I don't answer: my mail relay can't handle !, % or .uucp, I think *** The year 2000 will be the last year of the 20th century. ***
From: Victor Solakhian <vicsol@eecs.umich.edu> Newsgroups: comp.sys.next.programmer Subject: Re: stripping EOF-app? Date: 19 Jun 1996 21:14:42 GMT Organization: University of Michigan EECS Distribution: world Message-ID: <4q9qk2$37j@news.eecs.umich.edu> References: <4q7bkc$71r@turbocat.snafu.de> In article <4q7bkc$71r@turbocat.snafu.de> dave@turbocat.snafu.de (David Wetzel) writes: > Hello! > > The app, I wrote works fine but I wanted to strip it to make it smaller. That > worked for me fine on non-eof apps. But when I strip the EOF app: > > dave@turbocat> ./OPPlan > Jun 19 00:40:06 [7107] Unable to load extended character set(s); using > builtin. > Error loading /NextLibrary/Adaptors/Sybase.dbadaptor/Sybase > rld(): Undefined symbols: > .objc_class_name_EOAdaptor > .objc_class_name_EOAdaptorChannel > .objc_class_name_EOAdaptorContext > .objc_class_name_EOAttribute > .objc_class_name_EOEntity > .objc_class_name_EOJoin > .objc_class_name_EOModel > .objc_class_name_EONull > .objc_class_name_EOQualifier > .objc_class_name_EOQuotedExpression > .objc_class_name_EORelationship > .objc_class_name_EOSQLExpression > .objc_class_name_NSArray > .objc_class_name_NSAutoreleasePool > .objc_class_name_NSCalendarDate > .objc_class_name_NSData > .objc_class_name_NSException > .objc_class_name_NSMutableArray > .objc_class_name_NSMutableDictionary > .objc_class_name_NSMutableString > .objc_class_name_NSNumber > .objc_class_name_NSObject > .objc_class_name_NSString > .objc_class_name_NSUnarchiver > .objc_class_name_NSValue > .objc_class_name_NXConstantString > .objc_class_name_Object > .objc_class_name_Panel > _NSDefaultMallocZone > _NSInternalInconsistencyException > _NSLog > _NSZoneMalloc > _NXApp > _NXCloseMemory > _NXCloseTypedStream > _NXGetDefaultValue > _NXGetMemoryBuffer > _NXOpenMemory > _NXOpenTypedStream > _NXReadObject > _NXRunAlertPanel > _NXWriteRootObject > __NSAddHandler > __NSRemoveHandler > __ctype_ > __filbuf > __flsbuf > __iob > __setjmp > _abort > _access > _atoi > _close > _connect > _ctime > _errno > _exit > _fclose > _fcntl > _fflush > _fgets > _fopen > _fprintf > _fread > _free > _fscanf > _fseek > _ftell > _fwrite > _getdtablesize > _getenv > _geteuid > _gethostbyname > _gethostname > _getpid > _getpwnam > _getpwuid > _gettimeofday > _lseek > _malloc > _memchr > _memcmp > _memcpy > _memset > _objc_lookUpClass > _objc_msgSend > _objc_msgSendSuper > _object_getClassName > _open > _printf > _read > _sel_getName > _select > _send > _setsockopt > _sigblock > _signal > _sigsetmask > _sleep > _socket > _sprintf > _stat > _strcat > _strchr > _strcmp > _strcpy > _strlen > _strncmp > _strncpy > _strrchr > _strtod > _strtol > _sys_errlist > _sys_nerr > _vfprintf > _write > > Jun 19 00:40:11 [7107] *** Uncaught exception: > _newDataBaseWithModel:databaseName:: EOModel > file:"/Users/dave/projects/op/OPPlan/OPPlan.app/GUSTAV.eomodel" specifies > nonexistent adaptor with name "Sybase" > > exiting! > > The adaper is there! (NS 3.3 PL 1 EOF 1.1) > > ??????? > _ _ > _(_)(_)_ David Wetzel, Turbocat's Development, > (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, > _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 > (______) dave@turbocat.snafu.de (NeXTMail) -- Did you edit your Makefile.preamble and Makefile.postamble files as it is described in chapter 6 of Enterprise Objects Framework Guide (/NextLibrary/Documentation/NextDev/EnterpriseObjects/Guide/Part2_AppDesig n/06_CreatingAProject.rtfd): ----- Loading Unreferenced Symbols The Enterprise Objects Framework dynamically loads the adaptor or adaptors your application needs. The adaptor code may refer to symbols from the Enterprise Objects Framework library, but since not all of these references are resolved when your application is being compiled, they won't necessarily be loaded into your application's executable image. To rectify this, you need to force the loader to load all the symbols that the adaptor might refer to; this is done by editing your project's Makefile.preamble file (located in Supporting Files in your project) to include this line: OTHER_LDFLAGS = -all_load This tells the loader to load all the symbols from the named libraries even if they aren't referred to when the application is being compiled. When, in the course of your application's execution, an adaptor is dynamically loaded and used, the symbols that it needs are thus guaranteed to be present. To keep these symbols from being stripped when you build your application with the install target (that is, when you make install your application), edit your Makefile.postamble file (also located in Supporting Files) to include this line: APP_STRIP_OPTS = $(DYLD_APP_STRIP_OPTS) ------ ________________ Victor Solakhian Electrical Engineering and Computer Science Dept. Center for Display Technology and Manufacturing University of Michigan vicsol@eecs.umich.edu (NeXT, MIME) _________________________________________________
From: derek@slab.isdn.uiuc.edu (Derek Taubert) Newsgroups: comp.sys.next.programmer Subject: NXPlayStream and Mulaw8 data Date: 20 Jun 1996 04:04:37 GMT Organization: University of Illinois at Urbana Message-ID: <4qaikl$6be@vixen.cso.uiuc.edu> In short, what I need to do is feed Mulaw8 data at 8KHz as it being generated by my program to a NXSoundOut object using an NXPlayStream. However, the NXPlayStream only supports Linear16 at 22KHz and 44KHz. I have managed to work up a sample application using Linear16 22KHz, but ultimately I need to use Mulaw8 8KHz. Am I stuck converting my Mulaw8 data to Linear16 at 22KHz on the fly before feeding it to the NXPlayStream with playBuffer: size: tag:? How does the Sound object play back Mulaw8 8KHz data through the (black next) DAC? Does it convert it on the fly? Use the DSP to convert it on the fly? Why isn't the CODEC data/speed supported by NXPlayStream? -- Derek Taubert - http://slab.isdn.uiuc.edu/~derek taubert@uiuc.edu - NeXTMail welcome
From: stefan.boehringer@rz.ruhr-uni-bochum.de (Stefan Boehringer) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: Indexing Kit Date: 20 Jun 1996 07:13:32 GMT Organization: Ruhr-Universitaet Bochum, Rechenzentrum Message-ID: <4qatms$sor@sun168.rz.ruhr-uni-bochum.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Ladies and gentlemen, I recently started a mission critical app using the indexing kit. I ran into serious problems jeopardizing the whole project: the application frequently crashes during replaceRecord:with: methods sent to IXRecordManager. This seems to happen unless all objects in this recordManager are *direct* subclasses of Object. The default typedStream archiving-methods (read:/write:) are used for serialization. Even if all objects are are *direct* subclasses of Object the 9001 - exception is occasionally raised from inside -replaceRecord:with: with the result that the database is rendered unusable. I could mail a small PB.project that reproduces a crash. We have to eruate the feasability of our IXKit project fast. So please email us wheter there is some hope to get IXKit do its work in a reliable fashion. Thanks in advance. - Stefan Böhringer
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Deleting a line in Text object? Date: 20 Jun 1996 08:56:40 GMT Organization: Turbocat's Development, Germany Message-ID: <4qb3o8$gq@turbocat.snafu.de> Hello! Let's say I have a Text that countains this: 12 12.06.96 1 150,00 20,00 2546,98 <RGNR2> <RGDATUM2> <MS2> <ANZAHLUNG2> <ZINSEN2> <OFFEN2> <RGNR3> <RGDATUM3> <MS3> <ANZAHLUNG3> <ZINSEN3> <OFFEN3> How do I delete the last two lines? Thanks in advance _ _ _(_)(_)_ David Wetzel, Turbocat's Development, (_) __ (_) Buchhorster Strasse, D-16567 Muehlenbeck/Berlin, FRG, _/ \_ Phone +49 33056 82151, Fax +49 33056 82152 (______) dave@turbocat.snafu.de (NeXTMail)
From: don@misckit.com Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer Subject: Re: Indexing Kit Date: 20 Jun 1996 13:58:56 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4qblf0$tg@news.xmission.com> References: <4qatms$sor@sun168.rz.ruhr-uni-bochum.de> stefan.boehringer@rz.ruhr-uni-bochum.de (Stefan Boehringer) wrote: > [... IXKit crasher ...]. > So please email > us wheter there is some hope to get IXKit do its work in a > reliable fashion. I would like to think that the MiscKit folks will be a resounding success. I see no reason at this point to expect otherwise. :-) Although I've not set up our bug tracking software yet, the following addresses do exist: ixbugs@misckit.com for IXKit bug reports 3dbugs@misckit.com for 3DKit bug reports When the bug tracking stuff is in place, we will put out an announcement in csn.announce. In the meantime, though, you can still use the addresses; you just won't get a tracking number or response sent back automatically until we get the bug tracker up and running. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Deleting a line in Text object? Date: Thu, 20 Jun 1996 20:34:40 +0100 Organization: Cedar Systems Distribution: world Message-ID: <B5rODCAQfayxEwuM@cedar.co.uk> References: <4qb3o8$gq@turbocat.snafu.de> MIME-Version: 1.0 In article <4qb3o8$gq@turbocat.snafu.de>, David Wetzel <dave@turbocat.snafu.de> writes >Hello! > >Let's say I have a Text that countains this: > >12 12.06.96 1 150,00 20,00 2546,98 ><RGNR2> <RGDATUM2> <MS2> <ANZAHLUNG2> <ZINSEN2> <OFFEN2> ><RGNR3> <RGDATUM3> <MS3> <ANZAHLUNG3> <ZINSEN3> <OFFEN3> > >How do I delete the last two lines? > To delete text, you can select the text to be removed using [aText setSel:sPos :ePos] (where sPos is the character position of the start of the text to be deleted and ePos is the position after the last character to be deleted), and then remove the text with [aText replaceSel:""] To identify the positions of the start and end of the text you wish to delete, you could use the -stream or -firstTextBlock methods to inspect the contents of the text. Hope this helps, Paul. ------------------------------------------------------------------ Paul Heffernan, Product Development Manager. Cedar Systems, 2440 The Quadrant, Aztec West, Bristol BS12 4AQ, UK Phone: +44 (0) 1454 878708 Fax : +44 (0) 1454 878608
From: klingler@mack.rt66.com (Dave Klingler) Newsgroups: comp.sys.next.programmer Subject: Where can I find some Interceptor docs? Date: 20 Jun 1996 14:51:14 -0600 Organization: Engineering International, Inc. Message-ID: <4qcdk2$r6r@mack.rt66.com> Quite a long time ago I thought I remembered a discussion on c.s.n.p. about programming games under NS, Interceptor, etc.. I didn't pay much attention then because I was thinking about another project, but now I could really use some information. I've been combing through various archives but I can't seem to find the exchange. Basically I'd like to be able to use the equivalent of 1.0's NXDirectBitmap to do some animation. I've just finished poring through Omni's archive, but no dice. Can anyone help? I seem to have version 20 of Interceptor installed (think it came with NEXTTIME, but I have no docs. Thanksthanksthanks, Dave
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Thu, 20 Jun 1996 21:34:52 GMT Organization: Orbital Computer Consultancy Message-ID: <31c9c25f.9285985@news.plsys.co.uk> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) wrote: > <-- snip --> >Once again: an objects declared as NSArray just states that it is *at least* a >NSArray but might in fact be an instance of any subclass. > What the string "NSArray" means in Objective-C source-code depends on context. In a method declaration, yes it means the class NSArray or any subclass. In a method invocation, such as [NSArray array], it refers explicitly and distinctly to the NSArray class object.
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Thu, 20 Jun 1996 22:52:11 GMT Organization: Orbital Computer Consultancy Message-ID: <31c9d102.13033739@news.plsys.co.uk> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> besler@mdd.comm.mot.com (Steve Besler) wrote: >In article 100000@hphalle0.informatik.tu-muenchen.de, Bernhard Scholz > >>As you said: YES, YES, YES, that's exactly the problem returning an >>casted NSMutableArray as an NSArray. NOBODY guarantees you that the >>creator of the NSArray doesn't modify it in some way. That's why you >>shouldn't use casting. See also my other posting. The const keyword, just >>indicates you that _you_ are not allowed to change it. Returning an >>immutable object suggests that you and others (even the creator) can't >>change it. > >But that's Just it! Notice that NSArray is *NOT* named NSImmutableArrary >or NSConstantArray! This is because the NSArray *interface* makes no >claim of being "immutable", it is simply the interface that all the members >of the NSArray class cluster (mutable and immutable) implement. > >I don't really get why people don't understand this. Maybe I'm missing the >point, but I doubt it... > Let's put it another way, which is hopefully less bizarre and wierd than my other posts. If the NSArray interface, as you put it, makes no claim of being "immutable", what benefit is there in casting NSMutableArray to NSArray? If what you said were true and NSArray is the interface for the whole class cluster, mutable and all, then "the Tiggr agreement" wouldn't work, since the receiver couldn't know it was being told not to change the returned array. It's precisely because NSArray does signify immutability to the caller that, provided all callers abide by "the Tiggr agreement", everything's ok. Whether you, as the implementer of the method, wish to assume all callers will abide "by the Tiggr agreement" is a design issue for you, especially if you're designing a class to be used outside your project. As stated by another poster in this thread, "the Tiggr agreement", from the caller's point of view, requires them to copy the returned array if you want to be sure that the contents won't change. Far be it from me to suggest what the Foundation designers were thinking of when they named the NSArray and NSMutableArray classes, since I obviously have absolutely no idea, but I'd guess they thought it would be a little silly to have a class hierarchy such as NSMutableArray:NSImmutableArray. Well, what I've said makes sense to me, but no doubt I'm going to get some more lectures on basic OO now.
From: ngervae@next.com Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 21 Jun 1996 00:57:42 GMT Organization: NeXT Software, Inc. Message-ID: <4qcs26$2vf@news.next.com> References: <31c73d49.6197959@news.plsys.co.uk> In article <31c73d49.6197959@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: > For me, there are two significant errors in the Foundation > documentation: > > 2) IntroToFoundation's Object Ownership and Automatic Disposal > section, which I think is totally misleading. For example, the > statement: "The Foundation Kit therefore sets this policy: If you > create an object you alone are responsible for releasing it." Well, since I wrote that section I should address this. That sentence appears in a context before reference-counting has been introduced. I was trying to keep things clear from the context that's been established so far, and to deconstruct that as the reader proceeds. I guess it didn't work so well, but at least the summary is clearer about which specific methods are involved. Believe me, I struggled with a lot of people about how best to present the whole area. We knew it was going to be a sticky area.... > >In ObjC, declaring an object with a particular class means that > >the object can be an instance of that class or of any subclass. > >There are NO guarantees that this is not the case, with regard to > >mutability or other characteristics, unless you actually created > >the object yourself using alloc/init or copy. Ownership is a more > >significant indicator than class declaration in Foundation. > > > > Isn't the point of Foundation's (auto)release scheme to do away with > the concept of ownership? Depends on the context. With regard to who releases the object, yes, that's true. With regard to who can modify an object, no, ownership is definitely still a significant indicator. That's why many classes define ivars as NSMutable* privately, but pass out NS(Nonmutable)*'s publicly. > >Use of the term "immutable" is a documentation construct, not a > >formal part of the API. > > Isn't the fact NSArray does not provide an addObject: etc method > enough to classify it as having an API which makes it immutable? Yes and no. Sure, NSArray is immutable, but an object typed as NSArray may well be an NSMutableArray in fact. You *can* send addObject: to an object typed as NSArray--you'll get a compiler warning, but at run time, if the array responds to the method, it will modify its contents. In Objective-C, immutability is not enforced at run time by type declaration. This would be quite different in C++.... > >We should have clarified this concept much > >more than we did, as it now turns out. The distintion was made > >primarily to assist in compile-time type checking; run-time > >checking, to my mind (NOT necessarily the designer's), is a nice > >side benefit, but not one to be relied on. Again: There is NO > >formal guarantee that an object declared as NSArray won't change, > >unless YOU created that object as an NSArray. > > In that case, [NSArray copy] should only be implemented as > [NSArray retain] if the caller is the same as the object that created > the array. But doesn't it in fact disregard who the caller is? It can't regard who the sender is, since there's no means for it to know, so yes, it does disregard the sender. Still, NSArray *could* implement copy by retaining, since the class itself is immutable, regardless of the sender. But NSMutableArray would have to implement it by creating a separate copy. Who the sender is doesn't matter in this case, since the array itself "knows" if it's truly immutable or mutable. > >A non-mutable-by-the-recipient typing (how's that for obfuscation?) > >merely means that the recipient shouldn't change the object. It > >makes no claims about the owner's ability to change that object. > >If you aren't the owner, and you want to be sure an object you get > >from somewhere else is in fact immutable, it falls upon you, not > >the provider, to make it immutable, whether by shallow or deep copy. > >Only you, the recipient, know what you want. > > I was considering the provider's requirements, not the recipient's. It > was your colleague, Charles D'Harcourt, that brought the recipient > into it. His article has expired here so I can't check that. But I think the recipient is the logical bearer of this burden. The provider's requirements are quite simple to my mind: 1. Declare the return type appropriate to the recipient's use of the returned object. 2. Return an object that is a subtype of that type (or the type itself). Providers that want to be careful are free to return truly immutable copies of private ivars, or to trust the recipient to abide by the return type. Recipients that want private copies of received objects must copy them explicitly regardless, according to the default policy. You can document individual cases to be explicit about what the provider does, but in the default case the recipient *cannot*, by the principle of inheritance, assume that a nonmutable type is in fact absolutely IMmutable. > One must be wary of NSArrays which contain objects which don't conform > to NSCopying. One must indeed. Copying objects is actually a really hard problem, as Bruce Webster covers in his book, Pitfalls of Object-Oriented Programming. Wish I had a copy with me right now to pull a quote or two.... Note that copy does produce shallow copies of collections, so this is actually safe for contents that don't implement a copy method. It's mutableCopy that you have to worry about. There things can get ugly, and might force you to do a shallow copy yourself using alloc/initWithObjects:. > So is the nonmutable/mutable distinction just a documentation > construct or not? Although it's backed by an actual implementation difference, the distinction is just a documentation construct as far as the *consumers* of the classes are concerned. Like I said earlier, the distinction for the consumer is based on type declaration, which cannot be enforced at run time (without change to the language). And changing the language isn't something done lightly.... > I retracted this statement in another post. However, given everything > that's been said, I would like to know why EOModeler's create class > template function creates ivars for to-many relationships which are > NSArrays, and not NSMutableArrays. I haven't checked whether > EODatabaseChannel actually uses NSArrays for relationships, but > presumably it does. My guess is that it's a matter of laziness. :-) EOModeler only generates the accessor methods needed for key-value coding, and because of that it doesn't "need" mutable arrays. Even though you very well might want them. You can edit the template to make the arrays mutable, but EOModeler--and EOF in and of itself--just don't need them to be mutable. I hope all that clarifies what I'm saying. These replies aren't going through an edit/review cycle, after all. :-) > Do you mean the release after 4.0? Surely you're going to have to > be quick to get new documentation onto the 4.0 CD-ROMs? Or do you > know something we don't ;-) I mean the release after 4.0, which is out the door as far as documentation is concerned. -- Nik Gervae <ngervae@next.com> Technical Writer NeXT Software, Inc.
From: "Elwood K.F. Chu" <c3801721@comp.polyu.edu.hk> Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: OpenGL support for NEXTSTEP? Date: Thu, 20 Jun 1996 10:34:20 +0800 Organization: Hong Kong Polytechnic University Message-ID: <01bb5e50.fc5f01a0$0e0a40ca@alexander> I would like to know whether OpenGL support as well as development kit for NEXTSTEP? --------------- Elwood K.F. Chu BA (Hons) Computing Year 3 Hong Kong Polytechnic University
From: vadim@uccon.kiev.ua (Vadim Timchenko) Newsgroups: comp.sys.next.programmer Subject: [Q] How ps command work ? Date: 16 Jun 1996 20:01:31 GMT Organization: UCCON Message-ID: <4q1p6s$24r@verylong.uccon.kiev.ua> Hi, Can anybody say me how I may get list of all processes and its attributes (command line, environment) without root access right? Device /dev/kmem has only owner and group read rights but ps(1) works nice. Is there other way to do this? -- Best regards Vadim ( vadim@uccon.kiev.ua ) || ( 2:463/59.10@fidonet )
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 21 Jun 1996 06:13:58 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4qdej6$hup@chinx4.thoughtport.net> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> <x7g27zq20n.fsf@tom.es.ele.tue.nl> <4prfo9$lco@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q8m1j$qpg@zdi.informatik.uni-stuttgart.de> <4q92uq$92s@news.its.com> If only the Obj-C world were so simple... Regardless of how an immutable class is implemented, if it accepts and stores references to *any* objects created by an outside source, it is vulnerable to the fact that the Obj-C runtime *does not and cannot gurantee* immutable. For example, within the Foundation Kit itself, examine NSArray Vs. NSMutableArray. There are *numerous* methods that take an (NSArray *) as an argument-- almost all of which mean "take either a mutable or non-mutable array object as an argument". As well, throughout the foundation kit, the inheritance hierarchy is basically Mutable : Immutable -- that is, the Immutable Class acts as root class for a particular cluster and the Mutable implementation acts as the subclass. It makes a lot of sense (from the traditional, pointer-based, developer-is-effectively-god viewpoint) because Mutability is generally some simple additions to the API of a non-mutable class. Unfortunately, this means that the developer must go to great lengths to ensure that a mutable version of an object is *never* passed in where a mutable version is required. There are no warnings associated with casting, say, an NXMutableArray to NXArray. Because the concrete classes are supposed to be totally hidden from the developer AND the mutable concrete classes are generally subclasses of the non-mutable concreate classes, simple isKindOf: tests don't work. isMemberOf: only works if one is willing to acknowledge the existance of hidden, undocumented, and possibly transient classes. Then categories are added to the mix; with categories, one basically would have to validate the entire API of any given object to see if it is similar to a mutable or non-mutable class. Of course, even THEN, it is not impossible for awry developer to explode carefully-- nay, totally anal-- code. How many people have ever run into a piece of code that type-casts a (char *) to (id) and then added the (char *) to a List object? One Goes Ka-Boom without a CrashCatcher-like validation of the object... "Bullet-proof" programming in Obj-C is little more than agreeing with others that a particular API will act a particular way. The runtime/compiler does LITTLE to actually prevent the developer from breaking the rules. b.bum brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) wrote: >> See also my other posting. The const keyword, just >> indicates you that _you_ are not allowed to change it. Returning an >> immutable object suggests that you and others (even the creator) can't >> change it. > > Sorry, but this is just wrong. The language does not guarantee > immutability, so what mechanism does it guarantee then? The documenation? > The class name? Forget it. Thy really just "suggest" something, but > nothing else. Chuck Swiger replied: # The immutable class itself guarantees that its instances cannot # be changed by not providing setFoo accessor methods and by not # making the ivars public. > It is important that it really doesn't matters whether NSArray is immutable > since it's subclasses need not to be! A subclass might overwrite every > method and inherit quite nothing but it is still a legal subclass. # So what? That hypothetical subclass cannot override a setFoo # accessor method if it's immutable superclass does not implement # setFoo. And, if the immutable class declares it's ivars to be # "@private", the subclass cannot see the ivars to change them. > There is just no way of preventing somebody to write a mutable subclass > of an immutable superclass - and this means every programmer has to > anticipate such a case. # Disagree. I most certainly can write an immutable class and be # confident that nobody is going to change my instances. It's # interesting to note that Java provides the "final" keyword-- # which is used to indicate that a particular class cannot be # subclassed and to prevent precisely the situation you describe.
From: chuck@its.com (Chuck Swiger) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 19 Jun 1996 14:30:50 GMT Organization: Information Technology Solutions, Inc. Message-ID: <4q92uq$92s@news.its.com> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> <x7g27zq20n.fsf@tom.es.ele.tue.nl> <4prfo9$lco@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q8m1j$qpg@zdi.informatik.uni-stuttgart.de> brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) wrote: >> See also my other posting. The const keyword, just >> indicates you that _you_ are not allowed to change it. Returning an >> immutable object suggests that you and others (even the creator) can't >> change it. > > Sorry, but this is just wrong. The language does not guarantee > immutability, so what mechanism does it guarantee then? The documenation? > The class name? Forget it. Thy really just "suggest" something, but > nothing else. The immutable class itself guarantees that its instances cannot be changed by not providing setFoo accessor methods and by not making the ivars public. > It is important that it really doesn't matters whether NSArray is immutable > since it's subclasses need not to be! A subclass might overwrite every > method and inherit quite nothing but it is still a legal subclass. So what? That hypothetical subclass cannot override a setFoo accessor method if it's immutable superclass does not implement setFoo. And, if the immutable class declares it's ivars to be "@private", the subclass cannot see the ivars to change them. > There is just no way of preventing somebody to write a mutable subclass > of an immutable superclass - and this means every programmer has to > anticipate such a case. Disagree. I most certainly can write an immutable class and be confident that nobody is going to change my instances. It's interesting to note that Java provides the "final" keyword-- which is used to indicate that a particular class cannot be subclassed and to prevent precisely the situation you describe. -Chuck Charles Swiger -- chuck@its.com | Information Technology Solutions, Inc. --------------------------------+--------------------------------------- CrashCatcher Development, Systems and Networking Administrator
From: "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> Newsgroups: comp.sys.next.programmer Subject: GateKeeper 1.0 Problem. Date: Fri, 21 Jun 1996 05:21:54 -0500 Organization: University of Wisconsin-Madison Message-ID: <31CA77C2.3F22@students.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I tried to use GateKeeper together with ppp2.2.0.1.9. My problem was when I set up the ppd path in the GateKeeper preference menu as "/usr/local/ppp/bin" I got a message "error execu'ing pppd:Permission denied" right after connected to my ISP and clicking the pppd button on the GateKeeper dial-up menu. Does anyone know what is the solution to fix this problem? The thing I don't understand is I checked every related directories and link-directories content ppp files, and they were all group, others, user executable, but why the GateKeeper said that permission to execute pppd was denied? I really appreciate if anyone can help me to solve this problem. Thanks. Andrew.
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 21 Jun 1996 10:51:26 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4qdure$rte@zdi.informatik.uni-stuttgart.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> In <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: >In your previous postings, you did 3 statements which I'd like to >rewrite [del] >3) casting upwards the inheritance tree is _always_ save. >3) This isn't true. Imagine a method which collects mutable and immutable >objects in a single array and after collecting all objects it thinks: >"Let's change all mutable objects I own by performing mutable methods. >The immutable objects I collected won't get affected". That's a big mistake, >and it's hard, and even very time consuming to find such an error. What are your assumptions here? Let's make some: 1.) The mutable class is a subclass of the immutable one (it won't make much sense the other way round). 2.) a) The array is declared as an array of the immutable class. Then just nothing can go wrong, because sending an add: message or the like will lead to a type error (which is unfortunately just worth a warning for gcc) since such methods are not defined by the mutable class. b) The array is declared as an array of the mutable class. Then it is impossible to collect instances of the immutable class, based on assumption 1) So as far as I can see, nothing can go wrong. You might of course use the 2a) case and then determine dynamically which object is mutable an which not, but then your are bypassing the type system and its of course your responsibility to ensure that the semantics are reasonable. >This was my assumption when stating 'casting is always dangerous'. Please make your assumptions more explicit, mayby I just didn't understand your example. >Have a sunny day, > Boerny :-) Have a nice and sunny day, Klaus Brouwer P.S.: of course it is always possible to write crashing programs - in any language and with any paradigma.
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 21 Jun 1996 11:18:12 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4qe0dk$2e3@zdi.informatik.uni-stuttgart.de> References: <x7ohn98d6j.fsf@tom.es.ele.tue.nl> <4p66kj$dei@news.next.com> <31b74471.157975@news.plsys.co.uk> <Pine.HPP.3.91.960610105207.12925A-100000@hphalle0.informatik.tu-muenchen.de> <x7g27zq20n.fsf@tom.es.ele.tue.nl> <4prfo9$lco@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q8m1j$qpg@zdi.informatik.uni-stuttgart.de> <4q92uq$92s@news.its.com> In <4q92uq$92s@news.its.com> chuck@its.com (Chuck Swiger) writes: >The immutable class itself guarantees that its instances cannot be changed by >not providing setFoo accessor methods and by not making the ivars public. >> It is important that it really doesn't matters whether NSArray is immutable >> since it's subclasses need not to be! A subclass might overwrite every >> method and inherit quite nothing but it is still a legal subclass. >So what? That hypothetical subclass cannot override a setFoo accessor method >if it's immutable superclass does not implement setFoo. And, if the >immutable class declares it's ivars to be "@private", the subclass cannot see >the ivars to change them. >> There is just no way of preventing somebody to write a mutable subclass >> of an immutable superclass - and this means every programmer has to >> anticipate such a case. >Disagree. I most certainly can write an immutable class and be confident >that nobody is going to change my instances. I think my English is too bad, but we have a big misunderstanding here. Of course somebody (and most certainly you :-) can write classes which instances are truely immutable. But the problem leading to this discussion was different: Assume there is a class "Imm" with one private ivar, say "num"; the user must provide nums value on initialization, other methods for modifing "num" don't exist. Now somebody writes a subclass "Mut" with another ivar "mnum"; it overwrites the -(int)num method to return now the value of "mnum" and provides an additional method -setNum:(int)val modify it. It doesn't resuse a single line of code here, but "Mut" is still a legal subclass of "Imm". Ok so far? The original problem was: if I receive an object typed as (Imm *) then of course I cannot change it! But it might be changed by somebody else since in fact it might be an instance of "Mut" and not of "Imm"; an object declared as (Imm *) can be an instance of ANY subclass of "Imm". My conclusion was that you cannot rely on immutability of an object until you are the one who created it. If you want it to be immutable you'll have to make a copy! This no different to the problem of so called constant string buffers in traditional C programming. >It's interesting to note that Java provides the "final" keyword-- which is >used to indicate that a particular class cannot be subclassed and to prevent >precisely the situation you describe. Correct. Would be nice to have in fact. Have a nice day, Klaus Brouwer
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 21 Jun 1996 11:27:18 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4qe0um$3la@zdi.informatik.uni-stuttgart.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> <31c9c25f.9285985@news.plsys.co.uk> In <31c9c25f.9285985@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: >brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) wrote: >> ><-- snip --> >>Once again: an objects declared as NSArray just states that it is *at least* a >>NSArray but might in fact be an instance of any subclass. >> >What the string "NSArray" means in Objective-C source-code depends on >context. In a method declaration, yes it means the class NSArray or >any subclass. In a method invocation, such as [NSArray array], it >refers explicitly and distinctly to the NSArray class object. I know. Did I say something different? Maybe I should learn English now :-) Even more precise: a class name is ALWAYS a type name EXCEPT as an receiver of a message. But it is NEVER some sort of variable pointing to the class object! To get the class object you must say [ClassName class] instead of using ClassName directly. Now I hope you know that I know what you wanted to tell me :-) Have a nice day, Klaus Brouwer
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: Where can I find some Interceptor docs? Date: Fri, 21 Jun 1996 11:11:56 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <DtCKFx.DqM@cam-ani.co.uk> References: <4qcdk2$r6r@mack.rt66.com> In article <4qcdk2$r6r@mack.rt66.com> klingler@mack.rt66.com (Dave Klingler) writes: > Quite a long time ago I thought I remembered a discussion on c.s.n.p. > about programming games under NS, Interceptor, etc.. I didn't pay much > attention then because I was thinking about another project, but now I > could really use some information. Don't do it... I've worked with interceptor officially, and with support from NeXT. It's a good system, but it's more trouble than it's worth. Before you even think about trying it, I assume that you do have a mono NeXT, NeXTStation Colour, Dimmension, Intel 2bit, Intel 8bit gray, Intel 8 Colour, Intel 12444 Colour, Intel 12555 Colour, Intel 24 Colour, Sparc 8 Colour, Sparc 12 Colour, Sparc 24 Color and any other sparc's I've forgotten. I guess I we can let you off with HP, though really you should get one of each of those as well. Once you've bought all those machines you can start writing reliable interceptor code. Seriously you can code to be portable accross several of these machines, but you do need to test on most of these platforms. For this work you will get about 20-30% performance improvement over WELL WRITTEN dps code (OK you mgith be lucky and get more, but don't count on it). There's nothing wrong with Interceptor but it talks to the low level bitmap which is held on different formats on different platforms. > I seem to have version 20 of Interceptor installed (think it came with > NEXTTIME, but I have no docs. Thats the run time. The developer is not publically released. It has been reversed engineered, and additionally has been released to key developers. Ever used a Mac/Amiga/X? (PC's aren't too bad at this actually!) You find a great app you want to run, only to find your graphics card isn't supported. It's one of the best things about NeXTStep - you NEVER get told to get lost because you don't have the right hardware (MusicKit excepted!). Unless you've got the resources to support an Interceptor based app, then don't even think about doing it. Because NeXT withhold interceptor, everyone uses it as an excuse, and claims NeXT are keeping the goodiesd to themselves. Interceptor is NOT a panacea. It helps, but not much. Use NXBitmapImageRep, BUT MAKE IT MATCH THE SCREEN DEPTH! Detect the screen depth, and write a driver for each depth (only four required - 8bit colour doesn't work, but apparently 4.0 has some tricks here). Moderate sized bitmaps can be blitted very fast PROVIDED THEY'RE THE RIGHT DEPTH (remember to switch off buffering - use retained). If this still isn't fast enough, then the work you've done isn't wasted, as it is half way to writing an intercrptor driver (Looks much like an NXBitmapImageRep). $an
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 21 Jun 1996 11:52:04 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4qe2d4$7al@zdi.informatik.uni-stuttgart.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c9d102.13033739@news.plsys.co.uk> In <31c9d102.13033739@news.plsys.co.uk> k_harbour@mail.bogo.co.uk (Karl Harbour) writes: >besler@mdd.comm.mot.com (Steve Besler) wrote: >>In article 100000@hphalle0.informatik.tu-muenchen.de, Bernhard Scholz >> >>>As you said: YES, YES, YES, that's exactly the problem returning an >>>casted NSMutableArray as an NSArray. NOBODY guarantees you that the >>>creator of the NSArray doesn't modify it in some way. That's why you >>>shouldn't use casting. See also my other posting. The const keyword, just >>>indicates you that _you_ are not allowed to change it. Returning an >>>immutable object suggests that you and others (even the creator) can't >>>change it. >> >>But that's Just it! Notice that NSArray is *NOT* named NSImmutableArrary >>or NSConstantArray! This is because the NSArray *interface* makes no >>claim of being "immutable", it is simply the interface that all the members >>of the NSArray class cluster (mutable and immutable) implement. >> >>I don't really get why people don't understand this. Maybe I'm missing the >>point, but I doubt it... >> >Let's put it another way, which is hopefully less bizarre and wierd >than my other posts. >If the NSArray interface, as you put it, makes no claim of being >"immutable", what benefit is there in casting NSMutableArray to >NSArray? If what you said were true and NSArray is the interface for >the whole class cluster, mutable and all, then "the Tiggr agreement" >wouldn't work, since the receiver couldn't know it was being told not >to change the returned array. It's precisely because NSArray does >signify immutability to the caller that, provided all callers abide by >"the Tiggr agreement", everything's ok. The RECEIVER must not modify the returned array, and it it must not make a cast to NSMutableArray, which is away from the root of the class tree and almost always illegal, even it might work after some runtime typechecking. But the CREATOR/OWNER of the array might modify it, since (NSArray *) does not guarantee that it is really an instance of NSArray. Agreed? >Whether you, as the implementer of the method, wish to assume all callers will >abide "by the Tiggr agreement" is a design issue for you, especially if you're >designing a class to be used outside your project. If I'm designing a class (library) using strong typing I don't feel responsible for programmers who bypass the type system. This is of course just an opinion but since it is always possible for an programmer to crash a system I must ask you where to stop: copying returned ivars, run time check for parameters, consistency checks for dependend ivars (somebody might have bypassed your access methodss by adding his own through a category), ... what else? This all costs time: development time and run time. And all for some bad (!) programmers not sticking to standards and conventions - I say NO. >As stated by another poster in this thread, "the Tiggr agreement", >from the caller's point of view, requires them to copy the returned >array if you want to be sure that the contents won't change. Correct. But now a copy is made only if the caller really need immutability and not in any case. Isn't this an advantage? >Far be it from me to suggest what the Foundation designers were >thinking of when they named the NSArray and NSMutableArray classes, >since I obviously have absolutely no idea, but I'd guess they thought >it would be a little silly to have a class hierarchy such as >NSMutableArray:NSImmutableArray. Correct. >Well, what I've said makes sense to me, but no doubt I'm going to get >some more lectures on basic OO now. Correct, but less then last time :-) Have a nice day, Klaus Brouwer
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: Re: OpenGL support for NEXTSTEP? Date: 21 Jun 1996 12:10:05 GMT Organization: NO ORGANIZATION, INC. Message-ID: <4qe3et$bec@bias.ipc.uni-tuebingen.de> References: <01bb5e50.fc5f01a0$0e0a40ca@alexander> Cc: c3801721@comp.polyu.edu.hk In <01bb5e50.fc5f01a0$0e0a40ca@alexander> "Elwood K.F. Chu" wrote: > I would like to know whether OpenGL support as well as development kit for > NEXTSTEP? > > --------------- > Elwood K.F. Chu > BA (Hons) Computing Year 3 > Hong Kong Polytechnic University > Mesa (the OpenGL library, not the spreadsheet 8-)) has some support for NeXTSTEP: http://www.ssec.wisc.edu/~brianp/Mesa.html -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "In cantonese C++ is called C ga ga"
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: bug in cursor management? Date: 21 Jun 1996 14:20:59 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4qeb4b$5pb@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hello, I built an interface (using IB, v3.3) with a Matrix of switches having as background a ScrollView: the problem (a BUG?) is that i do not have the usual NXArrow cursor over the Matrix but the NXIBeam cursor (which is normal for ScrollViews). It seems the background ScrollView is "forcing" its beam cursor upon the Matrix view (which is however on top of it and completely comprised within the scroll view's rectangle). I tried to solve the problem with the following piece of code: [myMatrix getBounds:&bounds]; [myMatrix removeCursorRect:&bounds cursor:NXIBeam]; // -- this "remove" doesn't make any difference... [myMatrix addCursorRect:&bounds cursor:NXArrow]; [myWindow makeKeyAndOrderFront:..]; I get the most unusual result: this works, but ONLY on the SECOND (and later) display of the window!! Any hints? ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer,comp.sys.next.misc From: klaus@cruella.interpc.de (Klaus Bscheid) Subject: Printing portrait on a paper that is wider then high Message-ID: <DtD4z9.JH9@interpc.de> Sender: usenet@interpc.de Organization: interpersonal-computing GmbH Date: Fri, 21 Jun 1996 18:35:32 GMT Hi everyone I couldnÁt believe it but I wasn't able to set a PrintInfo where the width is greater than the height, without flipping automatically to landscape. After that I have tried to set the orientation back to portrait but this doesnÁt work. It's the same behavior if you try to do such a thing with the Format->PageLayout Panel. There is no way to enter a portrait pagesize witch is wider then high. Now you can say why not turning the paper to landscape orientation, but if you tell me a way to feed continuous paper in landscape orientation... But if anyone has a solution for this really strange behavior please let me know. have a nice day klaus
Newsgroups: comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: How to use (Im)mutability Message-ID: <DtD64n.G2F@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4q8m1j$qpg@zdi.informatik.uni-stuttgart.de> <4q92uq$92s@news.its.com> <4qdej6$hup@chinx4.thoughtport.net> Date: Fri, 21 Jun 1996 19:00:23 GMT In article <4qdej6$hup@chinx4.thoughtport.net> bbum@friday.com writes: > >Regardless of how an immutable class is implemented, if it accepts and stores >references to *any* objects created by an outside source, it is vulnerable to >the fact that the Obj-C runtime *does not and cannot gurantee* immutable. > I think of it in opposite terms... and so did the engineers who wrote NextStep and choose for Objective C. Thanks to the fact that the runtime doesn't make assumptions about what kind of objects you are actually working with, it's possible to design e..g "container" classes that can contain an arbitrary mix of "mutable" or "nonmutable" (whatever that means) objects.
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Sat, 22 Jun 96 22:44:48 +0200 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9606222044.AA17375@flexus> Subject: Re: How to use (Im)mutability Are we getting anywhere with this? Nik Gervae has already clarified that the purpose of (im)mutable is essentially the same one as ``const'' (though not in those exact terms): an access privilege/protection. And yet some people still deal with the subject in other terms. Here's my conclusion. I've omitted my (longhanded) reasoning because, sadly enough, people then tend to read nothing at all (sent to ``nobody'', ask for a copy if you want it). I've also omitted a summary of the statements about ``const'' (same remark). Conclusion: Instead of flattening everything into an inheritance tree (which is what the FoundationKit has tried to do), we need orthogonal: -protocols for conformance, -inheritance for implementation, -const for access control. Then we need a static checking mechanism without holes (for const at least, modelled after C++ with methods that can be declared const), ProjectBuilder support to remember relevant warnings, templates. If access protection needs to be dynamic instead of static (FK's (im)mutability is also static), then proxies should be provided. (Now don't assume I can't think of other nice things to have. Try me if you dare. :-) ) Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be (NeXTmail), Flanders, Belgium If I don't answer: my mail relay can't handle !, % or .uucp, I think *** The year 2000 will be the last year of the 20th century. ***
From: "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> Newsgroups: comp.sys.next.programmer Subject: GateKeeper Connection Date: Sat, 22 Jun 1996 20:33:43 -0500 Organization: University of Wisconsin-Madison Message-ID: <31CC9EF7.1F15@students.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I am trying to use GateKeeper-0.9.NI and PPP-2.2.0.4.6 for my NeXTStep Intel machine. Unfortunately, I have encountered problems with the "link" in GateKeeper menu. I created my scripts according to the exam- ples given by the GateKeeper. My modem dialed up to my ISP properly and obtained the local and remote IPs as well. However, after it ob- tained the local and remotes IPs, the connection disconnected. Here is the part of the messages I checked from the PPP-2.2.log file: . . . . . ########################## #### local IP address is: 204.120.5.128 ########################## #### remote IP address is: 204.120.4.7 ########################## June 22 18:41:31 silicon pppd[1504]: Setting interface mask to 255.255.255.0 June 22 18:41:33 silicon pppd[1504]: rcvd[IPCP ConfReq id=0x3 <compress VJ 0f 00><addr 204.120.4.7>] June 22 18:41:33 silicon pppd[1504]: ipcp:down June 22 18:41:33 silicon pppd[1504]: sent[IPCP ConfReq id=0x3 <addr 0.0.0.0><compress VJ 0f 01>] June 22 18:41:33 silicon pppd[1504]: ipcp:received COMPRESSTYPE June 22 18:41:33 silicon pppd[1504]: (45) June 22 18:41:33 silicon pppd[1504]: (ACK) June 22 18:41:33 silicon pppd[1504]: ipcp:received ADDR June 22 18:41:33 silicon pppd[1504]: (204.120.4.7) June 22 18:41:33 silicon pppd[1504]: (ACK) June 22 18:41:33 silicon pppd[1504]: ipcp:returning Configure-ACK June 22 18:41:33 silicon pppd[1504]: sent [IPCP confAck id=0x3 <compress VJ 0f 00> <addr 204.120.4.7>] June 22 18:41:33 silicon pppd[1504]: Terminating on signal 2. June 22 18:41:33 silicon pppd[1504]: sent [LCP TermReq id=0x2] June 22 18:41:34 silicon pppd[1504]: rcvd [IPCP ConfNak id=0x3 <addr 204.120.5.128>] June 22 18:41:34 silicon pppd[1504]: Hangup (SIGHUP) June 22 18:41:34 silicon pppd[1504]: rcvd [LCP TermAck id=0x4] June 22 18:41:34 silicon pppd[1504]: Connection terminated. June 22 18:41:35 silicon pppd[1504]: Exit. Does anyone have any pointers, why after pppd sucessfully negotiated the local and remote IPs, it went to "ipcp:down" and eventually terminated the connection of my modem to my ISP? By the way I am using inexpress.net as my ISP, if anyone knows about this ISP, what kinds of scripts (pppup, options, ipup, ipdown, etc) do I need to write or modify? And also I would be real appreciate if anyone could give me some scripts that will sustain the connection between my modem and my ISP. Until now, I don't even have enough time to execute the OMNIWEB, the connection has already timed out. Once again, I want to thank you in advance for your helps. I hope, I will hear the answers soon from you. Andrew.
From: "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> Newsgroups: comp.sys.next.programmer Subject: Resolving Problem. Date: Sun, 23 Jun 1996 04:03:34 -0500 Organization: University of Wisconsin-Madison Message-ID: <31CD0866.39B2@students.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, Does anyone know the solution how to resolve nameserver? My problem was after connection through PPP-2.2 I could not telnet or use browser by just typing the machine names, instead I had to type full IP addresses. I have my resolv.conf as follows: domain wisc.edu # WISC name servers # nameserver x.x.x.x where x.x.x.x is the DNS number. Could anyone tell me what was wrong with my resolv.conf file? Andrew.
From: rsmallwo@islandnet.com Newsgroups: comp.sys.next.programmer Subject: Re: NS V2.0 <table.h> file Date: Sun, 23 Jun 1996 13:11:23 GMT Organization: Island Net in Victoria, B.C. Canada Message-ID: <960623061123.263AAD6E.rsmallwo@canada> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII Does anyone have access to the definition of <table.h> in Nextstep 2.0. I need this for a development project I am involved with. If so, please send the file via 'nextmail' to 'rsmallwo@islandnet.com' Thanks in advance.
From: Mark Strand <marks@soli.inav.net> Newsgroups: comp.sys.next.programmer Subject: Help me make custom palettes Date: Sun, 23 Jun 1996 20:21:51 -0500 Organization: Internet Navigator, Inc. Message-ID: <31CDEDAF.5CCE@soli.inav.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I've been starting to toy with creating custom palettes with IB. The dynamic ones are easy, of course, and static palettes are also easy: if you're only using views, apparently (this is what my IB book & some NeXTanswers I've read suggest). I can follow the simple example in the online docs to make a progress bar (view) that has outlets and responds to methods. But I want to build a non-UI object such as Paul Hegarty did in that OPENSTEP video demo of IB. He drags an object (temperature converter) off the palette into the nib window (NOT an application window) and uses it just like he'd gone through and created a custom class, outlets, methods, etc. This is what I want to do. Maybe make a simple "calculator" object which reads 2 form rows/textFields for its values, and prints the sum to another field (a simple "sum" method). I'm having no luck in representing the object on the palette, or getting this thing loadable. I understand part of it entails using "associateObject:type:with:" to set the target (menu/window/nib window). As for making the icon, it sounds like you put a view on the PaletteView in the palette project (TTools calls it a "cover" and that is what you drag around. If someone could show me a simple demo from start to finish, I'd be eternally grateful (just a simple one, to get the basics down). The TTools had lots of neat stuff, but too complex to sit down and write a _simple_ palette.
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 24 Jun 1996 04:15:09 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4ql4od$a16@digifix.digifix.com> Topics include: Stepwise NEXTSTEP/OpenStep Information WWW site eduSTEP WWW site NeXT Software, Inc. WWW site comp.sys.next newsgroups related newsgroups comp.sys.next newsgroups mailing list ftp sites NeXTanswers Stepwise NEXTSTEP/OpenStep Information WWW site =============================================== This online community resource includes - 150+ ISV company pages - 350+ ISV product descriptions - NEXTSTEP Developer Directory - NEXTSTEP Community WhitePages - NEXTSTEP User Group Directory - comp.sys.next archives - User Group information - Mailing List archives and information You can connect via the world wide web at: http://www.stepwise.com/ Suggestions or comments can be directed to me at sanguish@digifix.com If you would like to get your company and product information on Stepwise, please contact me at sanguish@digifix.com. eduSTEP WWW site ================ http://www.nmr.embl-heidelberg.de/eduStep/ eduStep aims to provide up-to-date information on: - NextStep tools and projects for scientists. - Third-party products interesting for the educational and scientific community (with educational discounts noted, where they exist). - A listing of resellers and shops interested in working with customers in the educational community. - Conferences, meetings, workshops - Major projects, such as SciTools, EMBL's project to develop a NextStep scientific work environment - Status reports on GNUStep, a freely-available implementation of OpenStep now being developed NeXT Software, Inc. WWW site ============================ http://www.next.com comp.sys.next.* newsgroups ========================== news:comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. news:comp.sys.next.announce Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Archives are available by ftp at ftp://ftp.stepwise.com/pub/Next_Announce_Archives Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. news:comp.sys.next.bugs A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT- specific groups as well. news:comp.sys.next.hardware Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. news:comp.sys.next.marketplace NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. news:comp.sys.next.misc For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! news:comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. news:comp.sys.next.software This is a place to talk about [third party] software products that run on NEXTSTEP systems. news:comp.sys.next.sysadmin Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. Related Newsgroups ================== news:comp.soft-sys.nextstep Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. news:comp.lang.objective-c Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. news:comp.object Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep-d next-advocacy-d next-announce-d next-bugs-d next-hardware-d next-marketplace-d next-misc-d next-programmer-d next-software-d next-sysadmin-d (For a full description, send mail saying LISTS to <digestif@antigone.com>). The subscription syntax is essentially the same as LISTSERV's. To subscribe, send a message to <digestif@antigone.com> saying: SUB Listname YourName Example: SUB next-hardware-d John Doe The ftp sites ============= ftp://next-ftp.peak.org: The main site for North American submissions (formerly ftp.cs.orst.edu) ftp://ftp.informatik.uni-muenchen.de: (Peanuts) Located in Germany. ftp://ftp.dn.net/pub/next Peanuts mirror in the US ftp://terra.stack.urc.tue.nl (Dutch NEXTSTEP User Group) and ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next/ eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO <your-address> If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! Written by: Eric P. Scott (mailto:eps@toaster.SFSU.EDU) and Scott Anguish (mailto:sanguish@digifix.com) Additions from: Greg Anderson (mailto:Greg_Anderson@afs.com) Michael Pizolato (mailto:alf@epix.net) Dan Grillo (mailto:dan_grillo@next.com)
From: pascal@localhost (Pascal Thibaudeau) Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: Re: OpenGL support for NEXTSTEP? Followup-To: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Date: 24 Jun 1996 05:18:22 GMT Organization: CRIBX1 , Universite de Bordeaux I , France Message-ID: <4ql8eu$l7b@news.u-bordeaux.fr> References: <01bb5e50.fc5f01a0$0e0a40ca@alexander> Elwood K.F. Chu (c3801721@comp.polyu.edu.hk) wrote: : I would like to know whether OpenGL support as well as development kit for : NEXTSTEP? : --------------- : Elwood K.F. Chu : BA (Hons) Computing Year 3 : Hong Kong Polytechnic University Try Mesa-OpenGL for NeXTSTEP Getting the software ==================== The primary Mesa ftp site is iris.ssec.wisc.edu in the pub/Mesa directory. Mesa is also mirrored on sunsite in the directory pub/packages/development /graphics/mesa. The archive file Mesa-1.2.8.tar.Z can be unpacked with: zcat Mesa-1.2.8.tar.Z | tar xf - The archive file Mesa-1.2.8.tar.gz can be unpacked with: gzcat Mesa-1.2.8.tar.gz | tar xf - After you unpacking you should have the following files in the Mesa-1.2.8 directory: README - this file README.AMIWIN - instructions for using Mesa on Amigas with AmiWin (X11) LICENSE - the GNU library license IAFA-PACKAGE - description file Makefile - top-level Makefile Make-config - system configurations used by the Makefiles mklib.* - scripts for making shared libraries for some systems include/ - application include files lib/ - application libraries, created during installation src/ - source code for core library src-glu/ - source code for utility library src-tk/ - source code for tk library src-aux/ - source code for aux library demos/ - demo programs samples/ - sample OpenGL programs from SGI book/ - example programs from the OpenGL Programming Guide widgets/ - the Mesa widgets windows/ - Microsoft Windows driver stuff NeXT/ - NeXT demo programs mondello/ - Cirrus Logic Mondello 3-D accelerator code -- -Pascal Thibaudeau ********************************************************************** E-Mail: pthibaud@frbdx11.cribx1.u-bordeaux.fr ( NO NEXTMAIL ) pascal@galileo.slip.u-bordeaux.fr ( Small NEXTMAIL OK ) Powered by NEXTSTEP **********************************************************************
From: mcgredo@crl.com (Donald R. McGregor) Newsgroups: comp.sys.next.programmer Subject: Porting to Foundation Date: 23 Jun 1996 22:45:15 -0700 Organization: Miskatonic University Department of Classics Message-ID: <4qla1b$o8f@crl.crl.com> I'm modifying a large application to include a little bit of Foundation. I include the appropriate libraries and header files, recompile, and fire it up under gdb. Shortly thereafter I get an error in NXGetOrPeekEvent, from [Application run]. This happens in the appDidInit methd of the delegate. I'm assuming I'm getting bit by some sort of foundation kit memory management change; but what sort? What are the rules for mixing and matching old-style appkit programming and enough of Foundation to get EOF running? -- Don McGregor | Cthulhu '96: The choice for a new millenium. mcgredo@crl.com |
Newsgroups: comp.sys.next.programmer From: "Scott W. Bradley" <scottwb@grizzly.cs.washington.edu> Subject: libposix.a problems Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: news@beaver.cs.washington.edu (USENET News System) Organization: Computer Science & Engineering, U of Washington, Seattle Message-ID: <Pine.ULT.3.91.960624002520.21524H-100000@grizzly.cs.washington.edu> Mime-Version: 1.0 Date: Mon, 24 Jun 1996 07:28:56 GMT it seems to me that compiling with the flag -lposix ruins everything. however libposix.a is the only one i can find that defines setsid(). i have a program: void main() { printf("hello world\n"); exit(0); } when compiled with 'cc hello.c -lposix' and run, i get: Floating exception and the thing crashes. if anybody has any insight to this problem, i'd greatly appreciate hearing it. this has been tormenting me for some time now. thanks in advance -scott +---------------------------------+---------------------------------------+ | Scott W. Bradley | Home: 206.485.4142 | | 14128 NE 181st PL | Internet: scottwb@cs.washington.edu | | Apt. # K205 | | | Woodinville, WA 98072 | | +---------------------------------+---------------------------------------+
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: Printing portrait on a paper that is wider then high Date: 23 Jun 1996 21:44:28 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4qkdrs$b32@usenet.rpi.edu> References: <DtD4z9.JH9@interpc.de> klaus@cruella.interpc.de (Klaus Bscheid) wrote: > Hi everyone > > I couldn't believe it but I wasn't able to set a PrintInfo where > the width is greater than the height, without flipping automatically > to landscape. After that I have tried to set the orientation > back to portrait but this doesn't work. > > It's the same behavior if you try to do such a thing with the > Format->PageLayout Panel. There is no way to enter a portrait > pagesize which is wider then high. I've run into this too, trying to use an HP DesignJet plotter. The maximum size of a plot is 36 inches wide by 60 inches long. The width is fixed at 36 inches, but the height depends on the size of the image that you send the plotter. If I want to print something that is (say) 34 inches by 8 inches, I'd like that image to be printed 34 inches *wide* by 8 inches *long*. This uses 8 inches of paper. NeXTSTEP printing insists on printing it 8 inches wide by 34 inches high, which uses 34 inches of paper and gives me a lot of wasted space... We probably should BugNeXT about this, but I haven't had the time to yet. For what it's worth, it seems many (but not all) programs on the Mac exhibit the same behavior. They are determined to print the longest dimension of a document along the direction of the longest dimension of the maximum-size plot. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Mon, 24 Jun 1996 14:24:16 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> <4qdure$rte@zdi.informatik.uni-stuttgart.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <4qdure$rte@zdi.informatik.uni-stuttgart.de> On 21 Jun 1996, Klaus Brouwer wrote: > >3) casting upwards the inheritance tree is _always_ save. > > >3) This isn't true. Imagine a method which collects mutable and immutable > >objects in a single array and after collecting all objects it thinks: > >"Let's change all mutable objects I own by performing mutable methods. > >The immutable objects I collected won't get affected". That's a big mistake, > >and it's hard, and even very time consuming to find such an error. > What are your assumptions here? Let's make some: > 1.) The mutable class is a subclass of the immutable one (it won't make much > sense the other way round). > 2.) a) The array is declared as an array of the immutable class. Then just > nothing can go wrong, because sending an add: message or the like will > lead to a type error (which is unfortunately just worth a warning for > gcc) since such methods are not defined by the mutable class. > b) The array is declared as an array of the mutable class. Then it is > impossible to collect instances of the immutable class, based on > assumption 1) > > P.S.: of course it is always possible to write crashing programs - in any > language and with any paradigma. Maybe I got something wrong. Imagine the following: @implementation Foo + (NSString *)getString { // Foo uses NSMutableString for private usage, but returns NSString // to indicate that nobody is allowed to change it, without using // Foo methods // So in fact, the returned string is an instance of NSMutableString NSMutableString *myPrivateMutableString= [[NSMutableString alloc] initWithCString:"string of class FOO"]; return (NSString *)[myPrivateMutableString autorelease]; } @end @implementation Bar NSMutableArray *myArray; - initMyArray { if ((self=[super init])==nil) return nil; if (myArray == nil) myArray=[NSMutableArray new]; [myArray addObject:@"an immutable string, isn't it?"]; [myArray addObject:[[NSString alloc] initWithCString:"another immutable string]]; [myArray addObject:[[NSMutableString alloc] initWithCString:"mutable string"]]; [myArray addObject:[Foo getString]]; // now let's change our array // the following selector is only defined for mutable strings! [myArray makeObjectsPerformSelector:@selector(appendString:) withObject:[[NSString alloc] initWithCString:" *changed*"]]; // print out strings .. iterate myArray and print all strings ... return self; } @end I didn't test the code, but I'll bet the output is for any instance of Bar if called with [bar initMyArray]: an immutable string, isn't it? another immutable string mutable string *changed* string of class Foo *changed* The last line isn't what the author of this code was probably expecting when using the class Foo. I hope this clears up my previous posting, and shows the typing problem I wanted to point out. Best wishes, Boerny. -- Bernhard Scholz (IRC: Boerny) scholzb@pst.informatik.uni-muenchen.de http://peanuts.leo.org/ http://www.leo.org/~scholz/
From: Bernhard Scholz <scholz@informatik.tu-muenchen.de> Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: Mon, 24 Jun 1996 14:46:16 +0200 Organization: Technische Universitaet Muenchen, Germany Distribution: world Message-ID: <Pine.HPP.3.91.960624143547.3431B-100000@hphalle0.informatik.tu-muenchen.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> <4qdure$rte@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> On Mon, 24 Jun 1996, Bernhard Scholz wrote: Geez, reading this thread made me confused and I wrote wrong code. I shouldn't have written it. It uses the runtime system and is a big hack. On the other hand it gives a bad example, and doesn't obey the law "don't use methods on objects not responding to them". As Tiggr pointed out: these people should get fired :-) I only wanted to give a constructed example on how thinsg can get bad by hiding different objects in a seperate container class. In fact the container class can't be a simple NSMutableArray; one needs to use a container class which checks it's objects to perform to a specific selector before sending it to its objects. I'm not sure NSArray implements this behaviour, and I doub't it does. Best regards, Boerny. -- Bernhard Scholz (IRC: Boerny) scholzb@pst.informatik.uni-muenchen.de http://peanuts.leo.org/ http://www.leo.org/~scholz/
From: shill@iphysiol.unil.ch (Sean Hill) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> Control: cancel <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> Date: 24 Jun 1996 13:50:23 GMT Organization: University of Lausanne CH (Switzerland) Message-ID: <4qm6ev$4rk@cisun2000.unil.ch> cancel
From: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: libposix.a problems Date: 24 Jun 1996 14:14:48 GMT Organization: NO ORGANIZATION, INC. Message-ID: <4qm7so$g8l@bias.ipc.uni-tuebingen.de> References: <Pine.ULT.3.91.960624002520.21524H-100000@grizzly.cs.washington.edu> Cc: scottwb@grizzly.cs.washington.edu In <Pine.ULT.3.91.960624002520.21524H-100000@grizzly.cs.washington.edu> "Scott W. Bradley" wrote: > it seems to me that compiling with the flag -lposix ruins everything. > however libposix.a is the only one i can find that defines setsid(). > > i have a program: > > void main() { > printf("hello world\n"); > exit(0); > } > > when compiled with 'cc hello.c -lposix' > and run, i get: > Floating exception > > and the thing crashes. > > if anybody has any insight to this problem, i'd greatly appreciate > hearing it. this has been tormenting me for some time now. > Try 'cc hello.c -posix' instead of 'cc hello.c -lposix'. It's only a little char for the programmer, but a gigant leap for the compiler 8-). -- Frank M. Siegert [frank@this.net] -- Home Page http://hades.tue.schwaben.de/~frank NeXTSTEP, Linux & PostScript Guy "You'll never know what's in your lib until you look"
From: Matija Exel <exel@lag.ensieg.fr> Newsgroups: comp.sys.next.programmer Subject: Re: bug in cursor management? Date: 24 Jun 1996 16:14:06 GMT Organization: E.N.S.I.E.G./Service Informatique et L.A.G. Message-ID: <4qmese$7ui@naiad.grenet.fr> References: <4qeb4b$5pb@naiad.grenet.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Matija Exel <exel@lag.ensieg.fr> wrote: >hello, > >I built an interface (using IB, v3.3) with a Matrix of switches >having as background a ScrollView: the problem (a BUG?) is that >i do not have the usual NXArrow cursor over the Matrix but the >NXIBeam cursor (which is normal for ScrollViews). .. >Any hints? this is a post to myself...: well, no real solutions so far, but i got a hint, which suggests the following workaround (which i tested OK): --> group, with IB, the switch matrix within a second dummy ScrollView and use it to reassign a cursor for the group with [dummyScrollView setDocCursor:NXArrow]; this gives me back the arrow cursor over the matrix control! Sounds a bit clumsy, but it works... >----------------------------------------------------------------- >Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., >Tel : 76 82 71 12 Fax: 76 82 63 88 >e-mail : exel@lag.ensieg.fr (NO NextMail please) >----------------------------------------------------------------- > > ----------------------------------------------------------------- Matija Exel, E.N.S.I.E.G., Service Informatique et L.A.G., Tel : 76 82 71 12 Fax: 76 82 63 88 e-mail : exel@lag.ensieg.fr (NO NextMail please) -----------------------------------------------------------------
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <eott.com!tappd@nuchat.sccsi.com> Message-ID: <9606222138.AA24475@copernicus> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Dan Tapp <tappd@eott.com> Date: Sat, 22 Jun 96 16:38:50 -0500 Subject: Suggestions: Beyond RTF? Hello, all, again. I've rolled out the executive recap tool I've been working on (and pestering everybody about) for the past few months. Everyone likes the concept, but there's significant dissatisfaction with the recap format. Turns out the users who saw our mockups gave a great deal of lip service but limited thought to what they actually wanted, until confronted with the working product. I'm sure that's never happened to anyone else. Enough whining: the report uses RTF rulers and tabs to provide a nicely-formatted narrative document on the retrieved item's history and status. Turns out, the users want the detail-level stuff in tabular form...tabular to the point that we may be talking about actual tables, with boxes around rows and columns and stuff like that. While I'm waiting for revised specs, I thought it would be clever of me to figure out just what exactly I, as a journeyman NEXTer, am supposed to do about it if that **IS** what they want. I'm not really experienced with UNIX environments (NEXT is the first UNIX environment I've ever used), so I've been noodling around trying to figure out what my options are. I'm not sure which ones are even workable, let alone optimal. I might add that I'm going to need to turn this around reasonably quickly (i.e. 30-45 days) when I get the specs back, so the obvious answer (which is to learn enough about PostScript to directly build, control, and display a document with multiple embedded tables) is probably not the best answer, unless it's the **only** answer, in which case I suspect I'll want help figuring out how to tell management they need to wait for a year. Herewith, the apparent options. Your comments and criticisms (and especially your revelatory impartations) are welcome. Option the First: Shut up, learn more PostScript, and deal with it. Please see above. Option the Second: Prep an input file for another tool (like tex, maybe? ptroff?), spawn an appropriate background task using G&M's Process class, and display the resulting PostScript. Sounds vaguely promising, but I have no experience with these other tools, so I don't know if this is worthy of pursuit, or if I'm talking nonsense. Option the Third: We have a developer's kit for Complete Access, which another team uses extensively. Complete Access wasn't smart enough to handle the decision tree for this recap, but it might be smart enough to handle formatting. Also seems promising; I just need a closer look at the report formatting options and the API. Option the Fourth: Do the retrieval as a CGI process, format the output as HTML, and let folks use a Web browser to look at it. I think I like this option. Bit of a learning curve for me, but it seems decidedly less nasty than trying to turn myself into a PostScriptMeister in two months. However, intranet stuff is kind of low on the management priority list right now...we're just starting to dink with an Apache server, etc. and nobody here really understands http server mechanics, so if this is the best solution, I'll have to convince management that it's worth shifting other people's time and priorities. (I might add that management would be the only party I'd need to convince; the staffers are looking for an excuse, or at least a break in the workload, to swan dive into this stuff.) That's it. It's about 101 degrees in Houston right now, so I'm going to go to my apartment and sit at the bottom of the pool for awhile, while all you bright, non-acerbic people send me mail, tactfully pointing out the incredibly obvious and highly trivial solution to this. Regards, - Dan
From: xkueno@isid.co.jp (UENO Kotaro) Newsgroups: comp.sys.next.programmer Subject: Help: yppasswd() function on NeXTSTEP Date: 24 Jun 1996 08:08:50 GMT Organization: Information Services International - DENTSU, ltd. TOKYO, JAPAN. Message-ID: <4qliei$595@c0033.isid.co.jp> Hi, I would like to make a simple security program on NIS environment by using yppasswd() function. But yppasswd() function of NeXTSTEP occurs "Bus error". However, when I compiled the same program on HP-UX[9.05], that behavior was good. Then, do anybody have an experience about yppasswd() function on NeXTSTEP? If you know correct usage, please teach me. My using NeXTSTEP version is 3.3J. === main.c #include <pwd.h> #include <rpcsvc/yppasswd.h> main(argc, argv) int argc; char **argv; { struct passwd newpw; memset((char *)&newpw, 0x00, sizeof(struct passwd)); newpw.pw_name = "usr3"; newpw.pw_passwd = (char *)crypt(argv[2], argv[3]); newpw.pw_uid = 11043; newpw.pw_gid = 0; newpw.pw_dir = "/home/usr3"; newpw.pw_shell = "/bin/csh"; if (yppasswd(argv[1], &newpw)) printf("ERROR\n"); else printf("OK\n"); } === cc $ cc main.c -lrpcsvc === a.out $ a.out abcd00 abcd01 99 A^^^^^ B^^^^^ C^ A : old password B : new password C : cryptic key Thanks, --------------------------------------------------------------------- Jun/24/96[Mon] Kotaro Ueno/CNS,Ltd.
From: murphyx@earth.carleton.ca (Mitch Murphy) Newsgroups: comp.sys.next.programmer Subject: [Q] NXImage caching Date: 24 Jun 1996 20:45:03 GMT Organization: Carleton University Message-ID: <4qmuof$qfq@bertrand.ccs.carleton.ca> NNTP-Posting-User: murphyx Hello All. I am currently writing a NextStep 3.3 app that has to display EPS files from disk. By default, the EPS file is cached using 24 bit. Since the EPS files are in b/w, I would like to make the NXImage object cache itself using a 2 bit (b/w) cache. I tried using - useCacheWithDepth: NX_TwoBitGrayDepth, but the image does not appear in the view when I use composite:toPoint: , unless I use NX_TwentyFourBitDepth instead of NX_TwoBitGrayDepth. Any ideas ? Mitch ====================================================================== WELCOME TO THE PSYCHIATRIC HOTLINE. + If you are obsessive-compulsive, please press 1 repeatedly. + If you are co-dependent, please ask someone to press 2. + If you have multiple personalities, please press 3, 4, 5, and 6. + If you are paranoid-delusional, we know who you are and what you want. . . . Just stay on the line so we can trace the call. + If you are schizophrenic, listen carefully and a little voice will tell you which number to press. + If you are manic-depressive, it doesn't matter which number you press. No one will answer! =======================================================================
From: <todd@montrose.net> Newsgroups: comp.sys.next.programmer Subject: PowerGlove Library Sources and PaintLab package Date: 24 Jun 96 22:41:28 GMT Organization: Wonderful World of Dizzy Message-ID: <31cf1998.0@news1.montrose.net> Does anyone have the Joe Freeman package for the power glove. Also, what about the library source code... If you do, please email it to me via next or MIME. I thank you a bunch! TIA, \t -- Todd Nathan <todd@montrose.net> (all mail) + 1.800.556.5538 (all voice)
From: coty@netcom.com (Coty Rosenblath) Newsgroups: comp.sys.next.programmer Subject: NSImageRep for GIF format? Date: Mon, 24 Jun 1996 22:51:09 -0500 Organization: Emory University Message-ID: <coty-2406962251090001@news.cc.emory.edu> I hope I am asking the right question--I am new to NEXTSTEP. I'd like to do some work with GIF files, and it would seem that having a concrete subclass of NSImageRep that knew about GIF files would help. Does anyone out there know of such a thing? Thanks in advance, Coty
Newsgroups: comp.sys.next.programmer From: paul@cyantic.com (Paul Guiness) Subject: RE: Porting to Foundation Organization: CYANTIC Systems Date: Mon, 24 Jun 1996 12:37:21 GMT Message-ID: <1996Jun24.123721.11992@cyantic.com> You wrote: > I'm assuming I'm getting bit by some sort of foundation kit > memory management change; but what sort? What are the rules > for mixing and matching old-style appkit programming and > enough of Foundation to get EOF running? There are several small changes to make to your project so that it runs successfully under EOF. Chapter 6: Creating an Enterprise Objects Framework Project covers most of them. Two important parts not mentioned are that PB incorrectly writes out the import "EOApplication.h" line in the _main file, and that you may need to include a small .m called NSObject-IBFixes.m as well (can be found in the EOF examples). The above changes give you autorelease memory management as well. If you are not using EOF at all, then Foundation classes can be mixed with old Object classes by creating autorelease pools manually. Good luck, Paul Guinness paul@cyantic.com
From: IAMJYJ@chollian.dacom.co.kr (õ¸®¾È NEWS GROUP ÀÌ¿ëÀÚ) Newsgroups: comp.sys.next.bugs,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: Printing portrait on a paper that is wider then high Followup-To: comp.sys.next.bugs,comp.sys.next.programmer,comp.sys.next.misc Date: 25 Jun 1996 09:09:50 GMT Organization: DACOM Internet Message-ID: <4qoacu$8ct@nis.dacom.co.kr> References: <DtD4z9.JH9@interpc.de> <4qkdrs$b32@usenet.rpi.edu> Garance A Drosehn (gad@eclipse.its.rpi.edu) wrote: : klaus@cruella.interpc.de (Klaus Bscheid) wrote: : > Hi everyone : > : > I couldn't believe it but I wasn't able to set a PrintInfo where : > the width is greater than the height, without flipping automatically : > to landscape. After that I have tried to set the orientation : > back to portrait but this doesn't work. : > : > It's the same behavior if you try to do such a thing with the : > Format->PageLayout Panel. There is no way to enter a portrait : > pagesize which is wider then high. : I've run into this too, trying to use an HP DesignJet plotter. : The maximum size of a plot is 36 inches wide by 60 inches long. : The width is fixed at 36 inches, but the height depends on the : size of the image that you send the plotter. : If I want to print something that is (say) 34 inches by 8 : inches, I'd like that image to be printed 34 inches *wide* : by 8 inches *long*. This uses 8 inches of paper. NeXTSTEP : printing insists on printing it 8 inches wide by 34 inches : high, which uses 34 inches of paper and gives me a lot of : wasted space... : We probably should BugNeXT about this, but I haven't had : the time to yet. : For what it's worth, it seems many (but not all) programs : on the Mac exhibit the same behavior. They are determined : to print the longest dimension of a document along the : direction of the longest dimension of the maximum-size plot. : --- : Garance Alistair Drosehn = gad@eclipse.its.rpi.edu : Senior Systems Programmer (MIME & NeXTmail capable) : Rensselaer Polytechnic Institute; Troy NY USA
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <@vm.cnuce.cnr.it,@ISS.IT:ZANITTI@ISS.IT> Message-ID: <199606251203.FAA25780@isp.net> Date: Tue, 25 Jun 96 13:59:27 ISS From: ZANITTI@ISS.IT Subject: animate icon I would like to animate an application's icon, (by changing the application's icon). Could you help me please. The only information I found in Release Notes (NXAppTile) is very short. Thanking in advance Zanitti Leo zanitti@iss.it
From: briman101@aol.com (BriMan101) Newsgroups: comp.sys.next.programmer Subject: JOB: London & Paris - NeXT/UNIX/NT - OLE/PDO Date: 25 Jun 1996 10:15:58 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <4qosau$a4g@newsbf02.news.aol.com> European Investment Powerhouse currently seeks O-O architects and developers for a major new systems development project based in London and Paris. Candidates will be involved in developing a new architecture and front-end for several trading systems in the banks Fixed Income Group. You must have a strong background in Object-Oriented analysis and design, preferably from a Unix, NeXT, or Windows NT background. The project will be broken up into numerous small teams, each responsible for key elements of the trading environment. Architects need to have experience in analysis, design, and development of medium to large scale environments, with a strong hands-on knowledge of a wide range of development tools and operating systems, such as OMT, UNIX, Solaris, C++, X-Motif, and/or NeXTSTEP/OpenStep, Obj-C, EOF, and/or WinNT, VC++. Experience with Corba, Orbix, PDO, or OLE would be a strong advantage. Developers need to have experience in design and development of classic client/server systems. You must have strong hands-on experience in C++ or Objective-C development, coupled with one or more of the following skills: X-Motif, OpenStep, WinNT, VC++, OLE, Visual Basic, EOF, PDO. Salary range 30k to 65k + Bonus + Car + Benefits. Contact: Brian Mitchell Datacom Technology Group Inc New York Office 212-629-5720 212-629-3374(FAX) brianm@datacom1.com
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: FAT compiles Date: 25 Jun 1996 11:04:42 -0500 Organization: MCSNet Services Message-ID: <4qp2mq$29r@Mars.mcs.com> I know this is stupid, but I looked through the man pages and the on-line doc and could not find an answer. When I compile something with "cc -arch m68k -arch i386" it actually compiles each platform seperately. Is there some Define that can be checked to see what platform it is building to change headers or defines? Thanks. Peter Richardson
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: FAT compiles Date: Wed, 26 Jun 1996 09:58:03 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <DtLqCs.Ezv@cam-ani.co.uk> References: <4qp2mq$29r@Mars.mcs.com> In article <4qp2mq$29r@Mars.mcs.com> par@MCS.COM (Peter Richardson) writes: > I know this is stupid, but I looked through the man pages and the > on-line doc and could not find an answer. > > When I compile something with "cc -arch m68k -arch i386" it actually > compiles each platform seperately. Is there some Define that can be > checked to see what platform it is building to change headers or defines? I've never found doc's on this, but doing strings on the appropriate cpp's gives: -Dmc68000 -Dm68k -DNeXT -Dunix -D__MACH__ -D__BIG_ENDIAN__ -D__ARCHITECTURE__="m68k" and -Di386 -DNeXT -Dunix -D__MACH__ -D__LITTLE_ENDIAN__ -D__ARCHITECTURE__="i386" if you're just worried about endianness then: N[XS]SwapHostLongToBig, and other (little documented) related functions should be used. $an
From: pete@ohm.york.ac.uk (-bat.) Newsgroups: comp.sys.next.programmer,alt.3d Subject: Advice required for 3D graphics problem. Date: 26 Jun 1996 12:31:55 GMT Organization: The University of York, UK Message-ID: <4qrajr$fig@netty.york.ac.uk> Hi, I was wondering if someone could give me some advice for a way of doing some 3D programming under NeXTStep. I have what I consider to be a fairly simple problem - a matrix of values which I would like to be represented as a surface. Each value has a height and an associated colour. We did a few experiments a few years ago using NextStep's 3D kit and found that a single colour surface was fairly slow. We did not find a way to generate the multi-coloured surface which we require. The problem was then dropped for a bit and a crude implementation doe using straight PostScript and a fixed projection. Now we would like somethign which looks a lot better. Can anyone recommend a method of doing this ? Would Mesa help here - or is there a way of doing this in the 3D kit that I am unaware of ? Any advice would be appreciated. -bat.
Newsgroups: comp.sys.next.programmer Subject: Re: FAT compiles Message-ID: <1996Jun26.151228.46663@yogi.urz.unibas.ch> From: frank@ifi.unibas.ch Date: 26 Jun 96 15:12:28 MET References: <4qp2mq$29r@Mars.mcs.com> par@MCS.COM (Peter Richardson) wrote: > I know this is stupid, but I looked through the man pages and the > on-line doc and could not find an answer. > > When I compile something with "cc -arch m68k -arch i386" it actually > compiles each platform seperately. Is there some Define that can be > checked to see what platform it is building to change headers or defines? > Not stupid! Theoretically, you should be able to get the predefined macros with cc -E [ -dM | -dN | -dD ] file.c, but this just doesn't work. However, the compilers do have a configuration file, which is found in /lib/arch/specs, where arch is one of m68k, i386, hppa, or sparc. The files are a bit difficult to read, but under *cpp: you'll find the macros which are defined depending on what compiler flags you used. The notation is: %{compiler_flag:-Ddefine} for a define to be set if the compiler flags was given or %{!compiler_flag:-Ddefine} for a define to be set if the compiler flag was not given. For you case, the relevant defines are (you'd never have guessed!): m68k, i386, hppa, and sparc and __ARCHITECTURE__ these are listed under *predefines: -Robert -- Institut fuer Informatik tel +41 (0)61 321 99 67 Universitaet Basel fax. +41 (0)61 321 99 15 Robert Frank Mittlere Strasse 142 rfc822: frank@ifi.unibas.ch (NeXT,MIME mail ok) CH-4056 Basel X400: S=frank;OU=ifi;O=unibas;P=switch;A=arcom;C=ch Switzerland
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Syntax of the wod file in WebObjects Date: 26 Jun 1996 19:07:04 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4qs1oo$7g3@stc06.ctd.ornl.gov> I'm trying to connect an NSDictionary object to the Web page as in the DodgeLite example. Looking at the following in the Main.wod file of DodgeLite: MODEL.NAME:WOBrowser { list = models; item = model; value = model.name; selections = selectedModels; }; In the Main.wos there are the lines: id models, model, selectedModels; .. - awake { models = [[WOApp modelsDict] allValues]; .. } in the Application.wos there is - modelsDict { return [dodgeData objectForKey:@"models"]; } and of course in Main.html there is: <WEBOBJECT NAME="MODEL.NAME"><SELECT SIZE=5 MULTIPLE></SELECT></WEBOBJECT> NOW, Obviously in When the Main.html page is to be vended, then MODEL.NAME causes the main.wod file to be searched for MODEL.NAME where it finds that this represents an object of class WOBrowser. Since the awake method has already been done, the variable models has the list of all information with the key "models" from the NSDictionary. The items models, model, and selectedModels all appear as id's in the Main.wos. What does the item model.name mean? How is it associated with the NSDictionary data that is referenced by the variable model? Could I have substituted aModel for model and aModel.name for model.name? I assume "name" refers to the actual name value pair that is located in the model variable. The documentation doesn't make it clear to me the "mapping" here between the wod file and the wos file. Since the example has the word model in the NSDictionary in so many different contexts, it isn't clear what it is in the wod file. Clarification appreciated, -- - - - - - - - - - J. W. Wooten
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <randyn@crt.com> Message-ID: <9606261949.AA05837@nx_chi> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Randy Nelson <randyn@crt.com> Date: Wed, 26 Jun 96 14:48:54 -0500 Subject: Connecting to the window server with no one logged in Our apps have some parts that are tied to the interface and therefore we must have workspace running in order to connect to it. However, we want to run these jobs from cron during the night. When the apps are finished all the generated data is dumped directly to a file so there is not actual interaction with the interface. The problem is that we have to leave machines logged in in order to do this and this is becoming a problem of its own. I have tried using the work around posted by Eric P. Scott for connecting to lpr's window server port, but it appears that the Apps would have to be set-UID to whatever the lpr process run as. Q. Is there anyway to connect to the lpr's window server without being setUID and without anyone being logged into the console? or Q. Is it possible to create a window server that can run as anyone (or allow anyone to connect to) if no one is logged in? What would even be better is a kind of Off Screen window server where the display would never really show up on the screen similar to if you draw in an off screen window or drawing to /dev/null. Is this feasible at all? and if so, is there any example code to start from? Thanks for any ideas. -Randy Nelson __ Randy Nelson Software Engineer (NEXTSTEP Developer) NationsBanc--CRT (312) 234-3032 Fax: (312) 234-2186 randyn@crt.com (NeXT mail accepted) Drive thy Business, or it will drive thee. -Ben Franklin, Poor Richard's Almanack
From: John A. Maier <johnam@beta.datastorm.com> Newsgroups: comp.sys.next.programmer Subject: gcc for NeXT! (sort of) Date: 26 Jun 1996 20:52:16 GMT Organization: Datastorm Message-ID: <4qs7u0$c77@beta.datastorm.com> I download the following files off of peak and installed them: as-1.38.NIH.b.tar.gz cc-NeXT.2.5.8.NIH.ba.tar.gz gdb-4.7.NIH.b.tar.gz gnumake-3.74.b.SHIN.tar.gz libg++_for_3.2.NIH.ba.tar.gz I can start a compile, cc launchs as and then as wants ld (i386 3.3 version). Boom! So how do I get ld? jam *************************************************************************** * John A. Maier * Research and Development * * E-mail: johnam@datastorm.com * Datastorm Technologies Inc. * * * Procomm Plus for Windows and DOS * * * (314) 443-3283 * ***************************************************************************
From: monty@tbyte.com (Monty Zukowski) Newsgroups: comp.sys.next.programmer Subject: anti-aliased font rendering? Date: Wed, 26 Jun 96 22:17:57 GMT Organization: Trilobyte, Inc. Message-ID: <4qsdt2$3ga@marvin.tbyte.com> I wrote a little app to throw some text in a box (text object), justified, and shrink it until it fits. Then I composite that into a tiff and save the file for our CD-ROM game. Now the artists want anti-aliased text. How can I do this? I'd even be willing to generate and modify EPS if I can do it through that. Thanks for any help! Monty monty@tbyte.com
From: alanf@izzy.net (Alan Frabutt) Newsgroups: comp.sys.next.programmer Subject: MiscKit problem Date: 26 Jun 1996 18:59:31 -0400 Organization: Isthmus Corporation, Inc. Message-ID: <4qsfcj$ml@izzy4.izzy.net> I've installed MiscKit on a NeXTStation running the 3.2 User and Developer environments. I believe I've followed the instructions in the "Using the MiscKit" document, however I get the following run-time error: ----- Jun 26 18:44:52 manta Phone[4382]: objc: class `MiscTabActionCell' not linked into application Jun 26 18:44:52 manta Phone[4382]: An uncaught exception was raised Jun 26 18:44:52 manta Phone[4382]: Typed streams library error: class error for 'MiscTabActionCell': class not loaded ------ The -ObjC and -L/LocalDeveloper/Libraries are in place: --------- cc -g -O -Wall -I./sym -arch m68k -I. -c PhoneController.m -o ./m68k_obj/PhoneController.o cc -g -O -Wall -I./sym -arch m68k -ObjC -L/LocalDeveloper/Libraries -sectcreate __ICON __header Phone.iconheader -segprot __ICON r r -sectcreate __ICON app YourCall.tiff -o Phone.app/Phone m68k_obj/PhoneController.o m68k_obj/Phone_main.o -lMedia_s -lNeXT_s --------- Any help is greatly appreciated. -Alan (alanf@izzy.net)
From: chongl%sc@mhs.elan.af.mil Newsgroups: comp.sys.next.programmer Subject: Help! Please help with porting NeXT ObjC code to GCC Date: Wed, 26 Jun 1996 21:43:00 GMT Organization: United States Air Force Message-ID: <4qsb25$jo@rcp6.elan.af.mil> Hi, I'm trying to port my Objc code from my NeXT to a Unixware system running gcc version 2.7.2. The only include file I use is <objc/Object.h> and in my program, I only use the class Object. This is what I get: $ make gcc -O -c ngm.m -o ngm.o gcc -O main.m ngm.o -o ngmstat Undefined first referenced symbol in file __objc_exec_class /var/tmp/cca000Bx1.o __objc_class_name_Object ngm.o objc_msg_lookup /var/tmp/cca000Bx1.o objc_msg_lookup_super ngm.o objc_get_class /var/tmp/cca000Bx1.o ld: ngmstat: fatal error: Symbol referencing errors. No output written to ngms tat *** Error code 1 (bu21) make: fatal error. I would truly appreciate it if someone out there can help me because I have not clue on what the above errors mean. Thanks. --Leo chongl%sc@mhs.elan.af.mil
From: hnalgae@soback.kornet.nm.kr (SungJin Kang (kornet)) Newsgroups: comp.sys.next.programmer Subject: CID Keyed font in NextStep? Date: 27 Jun 1996 08:35:30 GMT Organization: KORNET (Korea Telecom) Message-ID: <4qth4i$2a1@usenet.kornet.nm.kr> Hi, I am Hayan Nalgae in Korea. I want to use CID Keyed font for Korean Language. Is it possible to use CID Keyed font in NextStep? Thanks for reading this. Hayan Nalgae hnalgae@soback.kornet.nm.kr http://soback.kornet.nm.kr/~hnalgae/
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 27 Jun 1996 09:00:45 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4qtijt$1nj@zdi.informatik.uni-stuttgart.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> <4qdure$rte@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> <Pine.HPP.3.91.960624143547.3431B-100000@hphalle0.informatik.tu-muenchen.de> In <Pine.HPP.3.91.960624143547.3431B-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: >On Mon, 24 Jun 1996, Bernhard Scholz wrote: >Geez, reading this thread made me confused and I wrote wrong code. I >shouldn't have written it. It uses the runtime system and is a big hack. >On the other hand it gives a bad example, and doesn't obey the law "don't >use methods on objects not responding to them". As Tiggr pointed out: >these people should get fired :-) I've read this just after my follow up to your code example - sorry for my unnecessary article. >I only wanted to give a constructed example on how thinsg can get bad by >hiding different objects in a seperate container class. >In fact the container class can't be a simple NSMutableArray; >one needs to use a container class which >checks it's objects to perform to a specific selector before sending it >to its objects. I'm not sure NSArray implements this behaviour, and I >doub't it does. Yes, but even then the semantics would be wrong! It would be really nice to have the generic mechanism from Ada. Then there would be full type checking even in container classes. *Sigh* Have a nice day, Klaus Brouwer
From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 27 Jun 1996 08:52:15 GMT Organization: Informatik, Uni Stuttgart, Germany Message-ID: <4qti3v$ft@zdi.informatik.uni-stuttgart.de> References: <Pine.HPP.3.91.960617172807.7023B-100000@hphalle0.informatik.tu-muenchen.de> <4q4qe3$ve@df13h.mdd.comm.mot.com> <31c73d30.6172470@news.plsys.co.uk> <4q8m70$r0m@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960619154911.265A-100000@hphalle0.informatik.tu-muenchen.de> <4qdure$rte@zdi.informatik.uni-stuttgart.de> <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> In <Pine.HPP.3.91.960624135817.28174A-100000@hphalle0.informatik.tu-muenchen.de> Bernhard Scholz <scholz@informatik.tu-muenchen.de> writes: >On 21 Jun 1996, Klaus Brouwer wrote: >> 2.) a) The array is declared as an array of the immutable class. Then just >> nothing can go wrong, because sending an add: message or the like >> will lead to a type error (which is unfortunately just worth a >> warning for gcc) since such methods are not defined by the mutable >> class. >> b) The array is declared as an array of the mutable class. Then it is >> impossible to collect instances of the immutable class, based on >> assumption 1) >> So as far as I can see, nothing can go wrong. >> You might of course use the 2a) case and then determine dynamically which >> object is mutable an which not, but then your are bypassing the type system >> and its of course your responsibility to ensure that the semantics are >> reasonable. [much deleted; important: myArray contains mutable and immutable strings] > // now let's change our array > // the following selector is only defined for mutable strings! > [myArray makeObjectsPerformSelector:@selector(appendString:) > withObject:[[NSString alloc] initWithCString:" *changed*"]]; I don't have any FK docu, but if -makeObjectsPerformSelector:withObject: behaves like -perform:with: then app will crash because NSString don't respond to -appendString: messages as you mentioned yourself. If it does check whether the receiving objects can respond then its just case 2a) together with dynamic type checking as I've mentioned in my previous post. If you assign an object to a variable of type id gives up on type checking - the responsibility is up to the programmer; and adding an object to a NSArray is just same. So the above code is legal from the compilers point of view. The error here is that objects of different classes are stored in the same array but on the other hand methods are sent to them which aren't implemented by all - this is a type error. Again: the code above is legal as far as Objective-C is concerend; but its up to programmer to ensure that it makes any sense. And - as your example showed - it doesn't. Have a nice day, Klaus Brouwer
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: Syntax of the wod file in WebObjects Date: 27 Jun 1996 10:22:35 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4qtndb$kan@btmpjg.god.bel.alcatel.be> References: <4qs1oo$7g3@stc06.ctd.ornl.gov> John W. Wooten writes .. -> MODEL.NAME:WOBrowser { -> list = models; -> item = model; -> value = model.name; -> selections = selectedModels; -> }; -> -> In the Main.wos there are the lines: -> -> id models, model, selectedModels; -> ... ... -> What does the item model.name mean? It assumes that model (the value of the variable 'model') is a dictionary Then model.name is the value associated to the key "name" in that dictionary. -> ... Could I have substituted aModel for -> model and aModel.name for model.name? Yes, if you declare it in Main.wos too. Hope that helps (and hope its true :-). Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine. e.
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: Syntax of the wod file in WebObjects Date: 27 Jun 1996 10:21:59 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4qtnc7$kak@btmpjg.god.bel.alcatel.be> References: <4qs1oo$7g3@stc06.ctd.ornl.gov> John W. Wooten writes .. -> MODEL.NAME:WOBrowser { -> list = models; -> item = model; -> value = model.name; -> selections = selectedModels; -> }; -> -> In the Main.wos there are the lines: -> -> id models, model, selectedModels; -> ... ... -> What does the item model.name mean? It assumes that model (the value of the variable 'model') is a dictionary Then model.name is the value associated to the key "name" in that dictionary. -> ... Could I have substituted aModel for -> model and aModel.name for model.name? Yes, if you declare it in Main.wos too. Hope that helps (and hope its true :-). Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine.
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: Syntax of the wod file in WebObjects Date: 27 Jun 1996 10:21:12 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4qtnao$kaf@btmpjg.god.bel.alcatel.be> References: <4qs1oo$7g3@stc06.ctd.ornl.gov> John W. Wooten writes .. | MODEL.NAME:WOBrowser { | list = models; | item = model; | value = model.name; | selections = selectedModels; | }; | | In the Main.wos there are the lines: | | id models, model, selectedModels; | ... ... | What does the item model.name mean? It assumes that model (the value of the variable 'model') is a dictionary Then model.name is the value associated to the key "name" in that dictionary. | ... Could I have substituted aModel for | model and aModel.name for model.name? Yes, if you declare it in Main.wos too. Hope that helps (and hope its true :-). Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine.
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: Syntax of the wod file in WebObjects Date: 27 Jun 1996 10:20:45 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4qtn9t$kac@btmpjg.god.bel.alcatel.be> References: <4qs1oo$7g3@stc06.ctd.ornl.gov> John W. Wooten writes .. | MODEL.NAME:WOBrowser { | list = models; | item = model; | value = model.name; | selections = selectedModels; | }; | | In the Main.wos there are the lines: | | id models, model, selectedModels; | ... .. | What does the item model.name mean? It assumes that model (the value of the variable 'model') is a dictionary Then model.name is the value associated to the key "name" in that dictionary. | ... Could I have substituted aModel for | model and aModel.name for model.name? Yes, if you declare it in Main.wos too. Hope that helps (and hope its true :-). Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine.
From: Frederic Vernier <Frederic.Vernier@imag.fr> Newsgroups: comp.sys.next.programmer Subject: editor & RiPolygon Date: Thu, 27 Jun 1996 15:38:34 +0000 Organization: IMAG, Grenoble, France Message-ID: <31D2AAFA.794B@imag.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi! (excuse my poor language, I'm just french) I would like to know which text editor you use ? Is there a better one than "edit" to write objective c ? I also have a second question : Where can i find the prototype for RiPolygon (3d kit) because the manual often use it but it never define it (RiPolygon is the constructor for polygon in 3dkit) . I found examples but none are in objective-c code. thanks for help (a last thing : does someone get a tcsh for NS 3.3 because csh make me crazy) _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Frederic VERNIER : CLIPS-IMAG, Bat. B Bureau 204 Grenoble FRANCE _/ _/ Email : Frederic.Vernier@imag.fr : Accept NeXT mails _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
From: griffith@crl.com (Dave Griffith) Newsgroups: comp.sys.next.programmer Subject: Adding a bundle to Preferences.app Date: 27 Jun 1996 08:05:51 -0700 Organization: The Hermetic and Esoteric Order of Clarity and Openness Message-ID: <4qu80f$lf5@crl.crl.com> References: <4qsb25$jo@rcp6.elan.af.mil> We've got a need at our organization to add a simple selection window to Preferences.app. According to the documentation, nothing could be easier. Just take the handy template provided in /NeXTApps/Preferences.app/Template.bproj, modify it to taste, and install. Piece of cake. The only problem is that said template doesn't exist on our system. What's the deal? A pointer to said template (or even better just dropping it in the mail to griffith@fct.com) would be greatly appreciated. -- --Dave Griffith, griffith@crl.com
From: par@MCS.COM (Peter Richardson) Newsgroups: comp.sys.next.programmer Subject: Extracting FAT objects Date: 27 Jun 1996 10:17:12 -0500 Organization: MCSNet Services Message-ID: <4qu8lo$i1b@Venus.mcs.com> I need to extract an object from an existing library. This comes as part of the distribution of a library package. The make file uses "ar x ..." to remove an object from an existing library and add it to the new library. This does not work with FAT libraries and I want to create the new library as FAT. Is there some way to extract the FAT object from a FAT library? Thanks Peter Richardson
From: chongl%sc@mhs.elan.af.mil Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4qsb25$jo@rcp6.elan.af.mil> Control: cancel <4qsb25$jo@rcp6.elan.af.mil> Date: Thu, 27 Jun 1996 16:19:14 GMT Organization: United States Air Force Message-ID: <4qucf6$8u7@rcp6.elan.af.mil>
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: Re: How to use (Im)mutability Date: 27 Jun 1996 17:42:52 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4quh6s$rs2@chinx4.thoughtport.net> References: <4q8m1j$qpg@zdi.informatik.uni-stuttgart.de> <4q92uq$92s@news.its.com> <4qdej6$hup@chinx4.thoughtport.net> <DtD64n.G2F@cwi.nl> Yes, when one views the Objective-C environment for what it is-- an extension on top of ANSI C... the design thinking of the Foundation is feature. But-- what if your custom container's purpose is to cache some expensive-to-generate description of the encapsulated objects. Unless one develops and forces everyone and everything to adhere to a 'Hey, this datum changed' notification system-- which may not be appropriate-- there is no way to gurantee that the datum's passed to your container are immutable (short of what I discussed in a previous message)! Specific example, say your caching container performs some summary calculation across a set of NSArray's-- and it REALLY MEANS a set of immutable NSArray instances. Because of the design of the Foundation [well, more because of Objective-C], the developer could easily [as is done in the Foundation all the time] pass a MUTABLE array to your container... If that array were to mutate while within your container-- but you don't have a notification set up because the specific design of your container says 'non-mutable arrays only'-- your container's cached calculation will now be out of sync with the real contents. The objective-c environment and design of the Foundation Kit do not provide what is necessary to prevent such things from happening. It could, but it doesn't-- it is yet another responsibility and point of failure laid in the hands of the developer! So-- yes, it CAN be feature. Similarly, it CAN be an extreme liability. b.bum #Re: How to use (Im)mutability #by David Stes, CWI, Amsterdam In article <4qdej6$hup@chinx4.thoughtport.net> bbum@friday.com writes: > >Regardless of how an immutable class is implemented, if it accepts and stores >references to *any* objects created by an outside source, it is vulnerable to >the fact that the Obj-C runtime *does not and cannot gurantee* immutable. > # I think of it in opposite terms... and so did the engineers who # wrote NextStep and choose for Objective C. Thanks to the fact # that the runtime doesn't make assumptions about what kind of # objects you are actually working with, it's possible to design # e..g "container" classes that can contain an arbitrary mix of # "mutable" or "nonmutable" (whatever that means) objects.
Newsgroups: comp.sys.next.programmer From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: gcc for NeXT! (sort of) Sender: news@novice.uwaterloo.ca (Mr. News) Message-ID: <Dto29v.C9I@novice.uwaterloo.ca> Date: Thu, 27 Jun 1996 16:10:43 GMT References: <4qs7u0$c77@beta.datastorm.com> Organization: University of Waterloo In article <4qs7u0$c77@beta.datastorm.com>, John A. Maier <johnam@beta.datastorm.com> wrote: >I download the following files off of peak and installed them: > >as-1.38.NIH.b.tar.gz >cc-NeXT.2.5.8.NIH.ba.tar.gz >gdb-4.7.NIH.b.tar.gz >gnumake-3.74.b.SHIN.tar.gz >libg++_for_3.2.NIH.ba.tar.gz > >I can start a compile, cc launchs as and then as wants ld (i386 3.3 version). Boom! > >So how do I get ld? > As far as I know, buy NS Developer. -- David Evans (NeXTMail OK) dfevans@bbcr.uwaterloo.ca Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/ University of Waterloo "Default is the value selected by the composer Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: anti-aliased font rendering? Date: 27 Jun 1996 18:35:06 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4quk8q$ahe@news.xmission.com> References: <4qsdt2$3ga@marvin.tbyte.com> monty@tbyte.com (Monty Zukowski) wrote: > I wrote a little app to throw some text in a box (text object), > justified, and shrink it until it fits. Then I composite that > into a tiff and save the file for our CD-ROM game. Now the artists > want anti-aliased text. How can I do this? I'd even be willing to > generate and modify EPS if I can do it through that. Have you looked at the source code to Sam Streeper's "ShipBuilder" application, which creates anti-aliased sprites in multiple rotations for Xox? You could get your code from there, for free. The basic idea is to create your text in an area 2X larger (or some other integral size larger) than desired for the final output. Then, you average the pixels together to determine what the pixels of the 1:1 scale image will be. For a 2X image, you will have four pixels to average to get one output pixel. The exact math used can slightly vary your results, so you may want to experiment some. (ie, taking averages vs. medians can change things a bit...etc.) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: sams@best.com (Samuel G. Streeper) Newsgroups: comp.sys.next.programmer Subject: Re: anti-aliased font rendering? Date: 27 Jun 1996 14:23:01 -0700 Organization: BEST Internet Communications Message-ID: <sams.835909545@shellx> References: <4qsdt2$3ga@marvin.tbyte.com> <4quk8q$ahe@news.xmission.com> don@misckit.com writes: >monty@tbyte.com (Monty Zukowski) wrote: >> (I need anti-aliased fonts...) >Have you looked at the source code to Sam Streeper's "ShipBuilder" >application, which creates anti-aliased sprites in multiple >rotations for Xox? For the record, I didn't write this; Linus Upson did. I mentioned to him that something like this would be useful to me and a couple days later he was done with it... It used to be on the orst ftp site, so it's probably now on next-ftp.peak.org, though I don't know where. > You could get your code from there, for free. Indeed this app would do the trick as-is, though it's not really intended for this. Any decent image editing app should work as well. I seem to recall that Keith made Pixelist available on his site (www.pixelsight.com) and this app is excellent if you use it in 24 bit mode (in the past it was buggy in 12 bit modes) It's a bit tricky to figure out though. >The basic idea is to create your text in an area 2X larger (or >some other integral size larger) than desired for the final >output. This is not necessary either for anti-aliasing in general or for ship-builder in particular, though it can make filtering conceptually a bit simpler. cheers, -sam
From: "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> Newsgroups: comp.sys.next.programmer Subject: Sound Blaster 16 Problem. Date: Thu, 27 Jun 1996 15:49:33 -0500 Organization: University of Wisconsin-Madison Message-ID: <31D2F3DD.79A0@students.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, I have a sound blaster 16 (not plug and play), it did not give problems when I played my CD ROM music. But, when it played sound files (au, snd, etc) sometimes it played the sound and sometimes it did not. Does anyone have any pointers or solutions to fix this problem? Is this a bug from the sound blaster 16 driver? I use the driver I obtained from NeXT Answer and I believe it's the recent driver released by NeXT. Andrew.
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: anti-aliased font rendering? Date: 27 Jun 1996 23:33:31 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4qv5ob$ahe@news.xmission.com> References: <4qsdt2$3ga@marvin.tbyte.com> <4quk8q$ahe@news.xmission.com> <sams.835909545@shellx> sams@best.com (Samuel G. Streeper) wrote: > don@misckit.com writes: > >monty@tbyte.com (Monty Zukowski) wrote: > >> (I need anti-aliased fonts...) > >Have you looked at the source code to Sam Streeper's "ShipBuilder" > >application, which creates anti-aliased sprites in multiple > >rotations for Xox? > > For the record, I didn't write this; Linus Upson did. > I mentioned to him that something like this would be > useful to me and a couple days later he was done with it... > It used to be on the orst ftp site, so it's probably now > on next-ftp.peak.org, though I don't know where. Thanks for the clarification! :-) > > You could get your code from there, for free. > > Indeed this app would do the trick as-is, though it's > not really intended for this. Any decent > image editing app should work as well. That is true; WetPaint or TIFFany II can do it easily. However, I got the impression that he wanted some custom code to do the job. The ShipBuilder code is the only NEXTSTEP based code that is freely available of which I am aware. I'm assuimg that he would rip that algorithm/code out of ShipBuilder and use it in his own custom app. It should be fairly easy to do that, and why not do the custom stuff? Then you can batch the image creation process and let it chug away...just feed it more text when its hungry. :-) > I seem to recall > that Keith made Pixelist available on his site > (www.pixelsight.com) and this app is excellent if > you use it in 24 bit mode (in the past it was buggy > in 12 bit modes) It's a bit tricky to figure out though. I don't think it is free, though...I remember seeing a hefty price tag. TIFFany II is a good app if you're going to be spending cash, and it seems to be better supported and also easier to use, from what I can tell. > >The basic idea is to create your text in an area 2X larger (or > >some other integral size larger) than desired for the final > >output. > > This is not necessary either for anti-aliasing in general > or for ship-builder in particular, though it can make > filtering conceptually a bit simpler. And probably a bit faster...since integral scale factors allow the coder to make some assumptions that can add dramatically to the speed of conversion. (Of course, an optimized algorithm would lose generality, but in this application, that may be just fine to do.) Anyway, anti-aliasing is a lot like scaling images down, with the addition of a mean instead of just copying pixels directly from the original image. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: (Izidor Jerebic) Newsgroups: comp.sys.next.programmer Subject: Profiling Foundation Kit? Date: 27 Jun 1996 17:43:34 GMT Organization: Select Technology Message-ID: <4quh86$j5f@lazar.select-tech.si> Keywords: foundation kit, performance, profile I have an application which spends (according to gprof) 90% of its time somewhere where profiler doesn't see it. I suspect Foundation Kit. But there is no libFoundation_p.a (profile library) on our systems. This version of library should probably enable profiling of Foundation Kit. Anybody knows where to obtain it? Izidor Jerebic izidor@select-tech.si
From: lbo@unconfigured.xvnews.domain (Laurent Bourqui) Newsgroups: comp.sys.next.programmer Subject: Re: Sound Blaster 16 Problem. Date: 28 Jun 1996 05:44:39 GMT Organization: The unconfigured xvnews people Message-ID: <4qvrg7INN2n6@maz4.sma.ch> References: <31D2F3DD.79A0@students.wisc.edu> In article <31D2F3DD.79A0@students.wisc.edu>, "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> writes: >Hello, > >I have a sound blaster 16 (not plug and play), it did not give problems >when I played my CD ROM music. But, when it played sound files (au, >snd, etc) sometimes it played the sound and sometimes it did not. >Does anyone have any pointers or solutions to fix this problem? Is this >a bug from the sound blaster 16 driver? I use the driver I obtained >from NeXT Answer and I believe it's the recent driver released by NeXT. > I have exactly the same problem. I suspect a bug in the sound blaster 16 driver. I just installed the version 3.33 (beta) from NeXT Answer and it seems to work a little bit better, but it's not perfect. If anybody know a workaround, It would help me. Thanks, Laurent -- -------------------------------------------------------------------------------- Laurent Bourqui | Phone (private) +41(0)1 272 62 92 SMI (Swiss Meteorological Institute) | Phone (office) +41(0)1 256 95 73 Kraebuehlstr. 58 | Fax (office) +41(0)1 256 92 78 | email lbo@sma.ch CH-8044 Zuerich | Switzerland | --------------------------------------------------------------------------------
From: yannick@uranus.univ-lr.fr (Yannick Buisson) Newsgroups: comp.sys.next.programmer Subject: textfield and floatValue Date: 28 Jun 1996 08:01:03 GMT Organization: Universite de La Rochelle Message-ID: <4r03fv$d3c@hpuniv.univ-lr.fr> Hi all, I have some probs with textfield ans the floatValue method !! When i send the floatValue method to a textfield, it return me a wrong value !! For example : if the user enter 15.71, the floatValue method will return 15.70099 or if the user enter 0.14 the floatValue method will return 0.130095 (but sometimes it return the right value !!) It's the same with the doubleValue method !! Any help would be appreciated :)) thanks for your answers yannick -- //// (. .) ----oOO--(_)--OOo-------------------------------------------- Yannick BUISSON Centre de Ressources Informatiques Uni. of La Rochelle (FRANCE) tel prof. : 46 45 82 14. fax prof. : 46 45 82 45. Email (NeXTMail , MIME) : -> yannick@cri.univ-lr.fr
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: NXImage always white!!! Date: 28 Jun 1996 13:27:20 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4r0mjo$osi@ai.alienor.fr> This is comming from the Next documentation about the NXImage object: [self getBounds:&theRect]; [self getFrame:&theRectFrame]; aNXImage=[[NXImage alloc] initSize:&theRectFrame.size]; if([aNXImage useCacheWithDepth:NX_EightBitGrayDepth]) { printf("bng bng\n"); if([aNXImage lockFocus]) { printf("bingo bingo\n"); [self drawSelf:&theRect :1]; [aNXImage unlockFocus]; } } But I've got "bng bng" and "bingo bingo" and it's not bingo : the NXImage is always WHITE... WHY??? Thanks for help -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: fgalot@x-lan.alienor.fr Newsgroups: comp.sys.next.programmer Subject: How seeing pasteboard's content? Date: 28 Jun 1996 13:43:21 GMT Organization: Alienor Informatique, Bordeaux, France Message-ID: <4r0nhp$osi@ai.alienor.fr> How seeing pasteboard's content? NXGetMemoryBuffer(aStream, &data, &length, &maxlen); [pboard writeType:NXPostScriptPboardType data:data length:length]; And After?? Today is the day where everything seems to be wrong sorry for all this questions and thanks for help. -- --------------------------------------- ® ® | ® O_O ® ® | O_O -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Fred Galot fgalot@x-lan.alienor.fr
From: Jean-Philippe Viegas <viegas@aar.alcatel-alsthom.fr> Newsgroups: comp.sys.next.programmer Subject: Re: textfield and floatValue Date: Fri, 28 Jun 1996 17:38:13 +0200 Organization: Alcatel Alsthom Recherche, Marcoussis, France Message-ID: <31D3FC65.167EB0E7@aar.alcatel-alsthom.fr> References: <4r03fv$d3c@hpuniv.univ-lr.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Salut, I tried the floatValue method to a textfield with the values you gave, but the method returned me the right values. Hem...Are you sure you're not using an old version of pentium processor... I continue looking at the problem while keeping your adress. bonne chance Jean Philippe
Newsgroups: comp.object,comp.software-eng,comp.sys.next.programmer,comp.os.ms-windows.programmer.tools From: madison@atlanta.twr.com (Madison Cloutier) Subject: TowerEiffel 2.0 available Message-ID: <MADISON.96Jun28115706@twratl.atlanta.twr.com> Sender: madison@atlanta.twr.com (Madison Cloutier) Organization: Tower Technology Date: Fri, 28 Jun 1996 16:57:06 GMT TowerEiffel Release 2.0 is Available An expanded version of this announcement was posted to comp.lang.eiffel. Full details are available on our Web site: http://www.twr.com. Tower Technology Corp. is now shipping TowerEiffel Release 2.0. This release bundles an integrated development environment, tools, compiler and over 300 classes including GUI (Motif and Windows), ODBC, direct persistence, distributed processing, international language support and the Eiffel Booch Components. Madison Cloutier madison@twr.com VP Marketing tower@twr.com Tower Technology Corp. http://www.twr.com 1501 W. Koenig Lane Tel (512) 452-9455 Austin, TX 78756 USA Fax (512) 452-1721
From: gclem@dannug.dk Newsgroups: comp.sys.next.programmer Subject: Re: anti-aliased font rendering? Date: 28 Jun 1996 07:11:32 GMT Organization: Danish NeXT User Group Distribution: world Message-ID: <4r00j4$b6@snaps.dannug.dk> References: <4qsdt2$3ga@marvin.tbyte.com> Appsoft Image can produce anti-aliased text and save it as TIFF while you are at it. Maybe other apps, e.g. TIFFany II, can do the same. Appsoft Image is a discontinued product, but maybe you can get a copy somewhere. Geert Monty Zukowski writes > I wrote a little app to throw some text in a box (text object), justified, and > shrink it until it fits. Then I composite that into a tiff and save the file > for our CD-ROM game. Now the artists want anti-aliased text. How can I do > this? I'd even be willing to generate and modify EPS if I can do it through > that. > > Thanks for any help! > > Monty > monty@tbyte.com
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Multi-threading in OpenStep ? Date: 28 Jun 1996 14:51:20 GMT Organization: Delft University of Technology Message-ID: <4r0rh8$t5e@mo6.rc.tudelft.nl> Hello, Openstep for Windows NT will support multiple threads, since the specification contains an NSThread class. However, the NS 3.3 documentation tells me that the Applicatin Kit is NOT thread safe, so that any drawing should be done only by the main thread. This is rather annoying; because this would mean you can't have two windows with each window being served by a seperate thread. Does someone know more than I ? _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Technical Mathematics & Informatics The Netherlands Delft University of Technology tel: +31 15 78 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
Newsgroups: comp.lang.postscript,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc From: mark@cyantic.com (Mark T. Dornfeld) Subject: Next EPS to Macintosh (Quark Express) Conversion Organization: CYANTIC Systems Date: Thu, 27 Jun 1996 23:09:07 GMT Message-ID: <1996Jun27.230907.25091@cyantic.com> We need to convert Next EPS to Macintosh EPS format including colour attributes. This must be done programmatically, not at the command line. Has anyone done this before (of course) and is there a utility (class or code snippet) available out there? -- Mark T. Dornfeld, Cyantic Systems Corporation Voice: (416) 621-6166 1 Eva Road Suite 301 Facsimile: (416) 621-6212 Etobicoke, Ontario, M9C 4Z5 CANADA Email: mark@cyantic.com "Vasily, give me one ping; one ping only" -- -- Mark T. Dornfeld, Cyantic Systems Corporation Voice: (416) 621-6166 1 Eva Road Suite 301 Facsimile: (416) 621-6212 Etobicoke, Ontario, M9C 4Z5 CANADA Email: mark@cyantic.com
From: quite@dial.pipex.com (Aandi Inston) Newsgroups: comp.lang.postscript,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: Next EPS to Macintosh (Quark Express) Conversion Date: Sat, 29 Jun 1996 13:07:37 GMT Organization: UnipalmPIPEX server (post doesn't reflect views of UnipalmPIPEX) Message-ID: <4r39pn$l1c@soap.news.pipex.net> References: <1996Jun27.230907.25091@cyantic.com> mark@cyantic.com (Mark T. Dornfeld) wrote: >We need to convert Next EPS to Macintosh EPS format including colour >attributes. This must be done programmatically, not at the command line. >Has anyone done this before (of course) and is there a utility (class or >code snippet) available out there? There are only three formats of EPS-with-preview that I've ever heard of: - PC-format, which includes a TIFF or WMF preview inside the EPS file - Mac-format, which includes a PICT preview accompanying the file - EPSI-format, which includes a greyscale preview as PostScript comments in the file. So, what is NeXT format EPS. Actually, I rather suspect it is EPS without preview, with NeXT using its display PostScript facilities to show you the preview. No answer, I'm afraid, just trying to clarify the question... ---------------------------------------- Aandi Inston quite@dial.pipex.com Visit http://ds.dial.pipex.com/quite for info on PostScript, PSAlter, psalters, tea, and small furry animals. And stuff.
From: john@openet.gr (john sidiropoulos) Newsgroups: comp.sys.next.programmer Subject: OpenGl for ns3.3 ns 4.0 Date: Sat, 29 Jun 1996 12:28:38 GMT Organization: Foundation for Research and Technology Hellas, FORTHnet Message-ID: <4r3i1n$o0j@aglaia.forthnet.gr> Please !!!!!! where can i find OpenGl for NeXTstep 3.3 /4.0 else i'm loosing my mind PLEASE !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! thnaks in advance
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: MiscKit problem Date: 29 Jun 1996 19:31:26 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4r40ae$g9g@news.xmission.com> References: <4qsfcj$ml@izzy4.izzy.net> alanf@izzy.net (Alan Frabutt) wrote: > I've installed MiscKit on a NeXTStation running the 3.2 User and > Developer environments. I believe I've followed the instructions in the > "Using the MiscKit" document, however I get the following run-time error: > [...] > cc -g -O -Wall -I./sym -arch m68k -ObjC -L/LocalDeveloper/Libraries > -sectcreate __ICON __header Phone.iconheader -segprot __ICON r r > -sectcreate __ICON app YourCall.tiff -o Phone.app/Phone > m68k_obj/PhoneController.o m68k_obj/Phone_main.o -lMedia_s -lNeXT_s > --------- > Any help is greatly appreciated. If you look at the end of the linking line, you'll see that your app is being linked against only two libraries: Media_s NeXT_s. In order to use the MiscTabActionCell you need to link against the proper library, which in this case is libMiscTabMatrix.a. You add it using the Libraries suitcase in ProjectBuiler's files inspector. Once you add the library, MiscTabMatrix will be listed along with Media_s and NeXT_s. Rebuild the app and that should fix your problem! -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Multi-threading in OpenStep ? Date: 30 Jun 1996 10:43:25 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4r5lod$dp6@lion.embl-heidelberg.de> References: <4r0rh8$t5e@mo6.rc.tudelft.nl> In article <4r0rh8$t5e@mo6.rc.tudelft.nl> guyt@is.twi.tudelft.nl (A. Guyt) writes: > Hello, > > Openstep for Windows NT will support multiple threads, since the > specification contains an NSThread class. However, the NS 3.3 > documentation tells me that the Applicatin Kit is NOT thread safe, so that > any drawing should be done only by the main thread. > > This is rather annoying; because this would mean you can't have two > windows with each window being served by a seperate thread. I really cannot see the point. Each view has its own drawSelf:... so you do not need to create a separate thread for each window. But to make you happy, with OS you can use NSConnection to send safely messages between threads... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer Subject: NEXTSTEP/Career Position/ILL Date: 30 Jun 1996 15:35:46 GMT Organization: Norden 1 Communications Message-ID: <4r66sj$d5@tofu.alt.net> Programmer/analyst/developer NEXTSTEP Objective C EOF-------------------------A Plus Sybase or Oracle------------A Plus Career Position-------------Exceptional opportunity Relocation -----------------Company Assistance To Be Considered------------Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.