This is Prog-07.gz in view mode; [Up]
From: john@gscorp.com (John C. Fox) 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: 30 Jun 1996 22:19:34 GMT Organization: Sirius Connections Message-ID: <4r6uhm$dp5@sun.sirius.com> References: <1996Jun27.230907.25091@cyantic.com> <4r39pn$l1c@soap.news.pipex.net> Cc: quite@dial.pipex.com > > - 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. > Hi There: Tailor Publishing will let you save EPS files with TIFF preview. Since Quark went cross platform, it should be able to load and display the files. However, the preview is in greyscale only. Go figure. It shouldn't upset the printing though. Hope this helps. John C. Fox
From: diego@conga.super.unam.mx (Diego Zamboni) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: gcc-2.7.2 on NS3.3-Sparcstation 5? Date: 1 Jul 1996 01:13:40 GMT Organization: Universidad Nacional Autonoma de Mexico Message-ID: <4r78o4$h63@news.mty.itesm.mx> Hi: I wanted to compile gcc-2.7.2 on my Sparcstation 5 running NS3.3, both to have a recent version of gcc and to be able to compile g77, the GNU FORTRAN compiler. However, apparently gcc doesn't yet support NEXTSTEP running on sparc processors. Does anybody have any ideas of getting around this problem? Thanks a lot, and best regards. -- Diego Martin Zamboni Jefe del Area de Seguridad en C'omputo diego@conga.super.unam.mx DGSCA, UNAM, Mexico. Tel. (52-5)622-81-69 (NeXTMail ok) Fax. (52-5)622-80-43 WWW home page: http://ds5000.super.unam.mx/~diego/ PGP key: finger diego@conga.super.unam.mx
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 Jul 1996 04:15:16 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4r7jck$744@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: "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> Newsgroups: comp.sys.next.programmer Subject: NeXT Mouse. Date: Mon, 01 Jul 1996 01:29:29 -0500 Organization: University of Wisconsin-Madison Message-ID: <31D77049.4502@students.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, Does NeXT mouse use bus type connector? If not, what type of connector does it use? Does any one know if I could use NeXT mouse for IBM PC system? Thanks in advance. Andrew.
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Re: Multi-threading in OpenStep ? Date: 1 Jul 1996 09:29:24 GMT Organization: Delft University of Technology Distribution: world Message-ID: <4r85pk$m9s@mo6.rc.tudelft.nl> References: <4r5lod$dp6@lion.embl-heidelberg.de> Georg Tuparev writes > > 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... > -- Thank you for the reply, but I'll make myself more clear:there are situations in which one would like to have multiple threads within one application, where each of this thread involves some drawing. Of course each thread can call some drawself: method but when multiple threads do this simultaneously, the windowserver gets confused: it can't deal with multiple threads. I hope you see my point. _____________________________________________________________________ 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.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Mon, 1 Jul 96 17:19:59 +0200 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9607011519.AA19988@flexus.fotek.com> Subject: Re: Multi-threading in OpenStep ? Abraham Guyt <guyt@is.twi.tudelft.nl> writes on 1996-07-01: > Thank you for the reply, but I'll make myself more > clear:there are situations in which one would like to > have multiple threads within one application, where each > of this thread involves some drawing. Of course each > thread can call some drawself: method but when multiple > threads do this simultaneously, the windowserver gets > confused: it can't deal with multiple threads. I hope > you see my point. It seems you are not in sync with the AppKit's paradigm. A program should not send drawSelf:: messages. Instead, if the result of drawSelf:: changes, it would inform the View in question about this by sending it the -update message from the main thread (you could use a proxy from a runFromAppKit connection, or use a lock that you run from AppKit). Details omitted for brevity. Note that NEXTSTEP's implementation of its own paradigm has some flaws, IMHO, so you should be prepared to apply ad-hoc remedies (see KBNS). 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. ***
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Mon, 1 Jul 96 17:53:14 +0200 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9607011553.AA20217@flexus.fotek.com> Subject: Re: Multi-threading in OpenStep ? I wrote, just a few minutes ago: > or use a lock that you run from AppKit This is confusing. I meant a lock(ing solution), which might involve shutting down the main thread. But it's easier to have an object that accepts -sendUpdateTo: messages, and is vended over a run-from-AppKit connection. -(void)sendUpdateTo:anObject{ [anObject update]; } 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: frank@this.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: OpenGl for ns3.3 ns 4.0 Date: 1 Jul 1996 20:21:44 GMT Organization: NO ORGANIZATION, INC. Message-ID: <4r9c0o$8vs@bias.ipc.uni-tuebingen.de> References: <4r3i1n$o0j@aglaia.forthnet.gr> Cc: john@openet.gr In <4r3i1n$o0j@aglaia.forthnet.gr> john sidiropoulos wrote: > Please !!!!!! > where can i find OpenGl for NeXTstep 3.3 /4.0 > else i'm loosing my mind > PLEASE !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! > thnaks in advance > > 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: Yvan Frey <yvan_frey@studio.disney.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.hardware,comp.sys.next.misc,comp.sys.next.sysadmin Subject: Novell connection dies after a while Date: Mon, 01 Jul 1996 15:31:43 -0700 Organization: Walt Disney Pictures & TV Message-ID: <31D851CF.42F8@studio.disney.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: yvan_frey@studio.disney.com Hi there! We're running NEXTSTEP on Intel and have setup Novell connectivity. (I had to edit /etc/rc and add a frame type parameter to the npsd call. Then it worked.) Unfortuately, now the connection is dropped after a while and the login panel shows up again. I've read about similar, yet quite different problems (connection dies when copying files that are more than 1M on a token ring net). Does or did anybody experience the same? Any hints? NEXTSTEP is running on a Compaq 5133, with the build in ethernet controller and I'm using the latest drivers provided by NeXT on www.next.com. Thanks in advance, Yvan Frey Walt Disney Pictures & TV
From: Luke_Howard@inter.net.au (Luke Howard) Newsgroups: comp.sys.next.programmer Subject: question regarding NeXT's purported CORBA support Date: 2 Jul 1996 07:26:59 GMT Organization: Australian Business Access Pty Ltd Message-ID: <4raj03$63@tiger.aba.net.au> Can anyone explain the following? (found on NeXT's web site in a description of what PDO/OpenStep can do) *** Integrate with CORBA services NeXT supports the CORBA 2.0 standard, allowing all your objects to interoperate easily with CORBA objects. Does this mean: a) NeXT acknowledge the fact CORBA exists, and acknowledge the possiblility that you might write ugly code using the C or C++ CORBA bindings under the odd operating system that supports PDO and CORBA (NT and Solaris come to mind). If you want to call them from Objective-C, fine, but you have to do the work yourself. or b) NeXT have integrated CORBA support into PDO, including the Objective-C bindings for IDL they submitted to the OMG, so that you can message CORBA objects as if they were native Objective-C objects. ? -- Luke
From: pete@ohm.york.ac.uk (-bat.) Newsgroups: comp.sys.next.misc,comp.sys.next.software,comp.sys.next.programmer Subject: Puzzled about OpenStep Date: 2 Jul 1996 12:42:40 GMT Organization: The University of York, UK Message-ID: <4rb5g0$qc@netty.york.ac.uk> O.K. - so I took a look at the OpenStep pages. It says that OpenStep is currently available for Mach and Windows NT. So far so good - BUT, it says that the developer stuff for Windows NT won't be available until later. Given that we need to recompile stuff to make it run under Windows NT then what use is the OpenStep for Windows without the compiler ? Surely it's totally useless or have I missed the point. Supposing I get a copy, what can I do other than install it and then stare at thhe screen wishing I had a compiler ? hmmm.... -bat.
From: aisbell@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.misc,comp.sys.next.software,comp.sys.next.programmer Subject: Re: Puzzled about OpenStep Date: 2 Jul 1996 15:12:59 GMT Organization: Netcom Distribution: world Message-ID: <4rbe9r$4h@dfw-ixnews10.ix.netcom.com> References: <4rb5g0$qc@netty.york.ac.uk> pete@ohm.york.ac.uk (-bat.) wrote: > O.K. - so I took a look at the OpenStep pages. It says that OpenStep > is currently available for Mach and Windows NT. So far so good - BUT, it > says that the developer stuff for Windows NT won't be available until > later. > Given that we need to recompile stuff to make it run under Windows NT then > what use is the OpenStep for Windows without the compiler ? Surely it's > totally useless or have I missed the point. Supposing I get a copy, what > can I do other than install it and then stare at thhe screen wishing I had > a compiler ? NeXT's Web site is offering a sneak preview (beta version) of the OPENSTEP/NT development tools which would be necessary to build an OPENSTEP/NT app. I understand that Microsoft Visual C++ must be installed as well because the NEXTSTEP linker (ld) hasn't been ported yet. -- 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: Stefan Leuker <sl@skynet.oph.rwth-aachen.de> Newsgroups: comp.sys.next.misc,comp.sys.next.software,comp.sys.next.programmer Subject: Re: Puzzled about OpenStep Date: Tue, 02 Jul 1996 15:50:51 +0200 Organization: RWTH Aachen Message-ID: <31D9293B.7B8D6549@skynet.oph.rwth-aachen.de> References: <4rb5g0$qc@netty.york.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit -bat. wrote: > > Given that we need to recompile stuff to make it run under Windows NT then > what use is the OpenStep for Windows without the compiler ? Surely it's > totally useless or have I missed the point. Supposing I get a copy, what > can I do other than install it and then stare at thhe screen wishing I had > a compiler ? > You can build your GUI and compile for NT on MachOS, given that you have the extendet OS/NT release of the dev tools. That's what I heard. Stefan
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: Profiling Foundation Kit? Message-ID: <DtwxEE.158@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4quh86$j5f@lazar.select-tech.si> Date: Tue, 2 Jul 1996 11:03:48 GMT In article <4quh86$j5f@lazar.select-tech.si> (Izidor Jerebic) writes: > > 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. Check out NeXTanswer #1898 Using_sampler_for_profiling_EOF. There is a tool called SAMPLER, which can be used to profile EOF applications. Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: "C++, Oracle, Unix, Job Opportunities" <job_op_dc@crc.com> Newsgroups: comp.sys.next.programmer Subject: C++, Unix, Oracle Positions Date: Tue, 02 Jul 1996 17:47:39 -0400 Organization: CRC HR Message-ID: <31D998FB.6EF5@crc.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Everybody wants you. Get what YOU want! Are you looking for a challenging career in information systems in a growing company with excellent benefits? Do you want the advantages associated with living in a major metropolitan area, the Southwestern sunbelt, the Northwestern ski country or the suburban Northeast? Coleman Research Corporation can offer all of these. CRC <http://www.crc.com/> is a highly diversified engineering and information services company providing services to the Federal government and private, commercial industry. CRC has over 200 openings that offer the opportunity for professional growth, exciting technical work, a stable work environment, and excellent benefits (including the full complement of insurance options, both 401(k) and pension programs, flexible paid leave programs, tuition reimbursement, and other benefits). Positions are available in the Washington, DC area (Gaithersburg, MD; Chantilly or Springfield, VA); Valley Forge, PA; Denver, CO; and Phoenix, AZ. Position educational requirements are a BS in Computer Science, Math, Engineering, Information Systems, and/or Physics. Qualified candidates will have experience in one or more of the following: software development using C/UNIX, using relational databases (Oracle), X-Windows/MOTIF, C ++ and GUI design and development, object oriented software development methodology, software through pictures, client-server technology, SQL databases, SIGINT applications and software, and digital signal processing hardware and software. Applicants selected will be subject to a government security investigation and must meet eligibility requirements for access to classified information. If you have these qualifications, CRC has opportunities for you today! Relocation assistance is available. CRC is an equal opportunity employer (EOE M/F/D/V). Interested individuals should send a current resume (with salary requirements) to: Coleman Research Corporation Attn: CTG 9302 Lee Highway, Suite 800 Fairfax, VA 22031 In addition, resumes for employment may mailto:job_op_dc@crc.com. accepted fromat for email resume: word for windows2.0 mime encoded or flat asci resume. <http://www.crc.com/>
From: mike@ali.bc.ca (Michael Cam) Newsgroups: comp.sys.next.programmer Subject: Problems using the debug panel when loading bundles Date: 3 Jul 1996 00:58:27 GMT Organization: ALI Distribution: world Message-ID: <4rcgjj$5io@cetus.ali.bc.ca> Hi, I have an app that loads a bundle and uses it to create classes etc. I was able to set breakpoints using the debug panel that comes up through the Edit.app menu when you hit the Debug button in PB. I could set breakpoints in both the app's source files as well as the bundle's source files (only after it is loaded). Now that I link Foundation libraries it seems that I cannot set breakpoints this way anymore. Instead I get this message: -------------------------------------- No source file named *.m. Error occured processing command from Edit. -------------------------------------- where * is the name of the source file I tried to set the breakpoint at. Does anyone have any ideas on how to get this to work? Thanks in advance. ..Mike. mike@ali.bc.ca --
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Imaging sizing in WebObjects Date: 3 Jul 1996 16:02:55 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4re5jf$9q9@stc06.ctd.ornl.gov> I have a series of jpeg images that will resize to the proper size in an html file using: <a href="http://www.kli.org/kli/sounds/HabQuch.au"><IMG SRC="./gifs/woo3.jpeg" ALIGN=LEFT WIDTH=90 HEIGHT=130 BORDER=1 ALT="A Great Picture of the Instructor"></a> I am trying to use those images in a WebObjects application. Where do I put the WIDTH=?? and HEIGHT=?? in order for those to be used. I've tried in the html file in various places and can't see where to put them in the wod file. Any Hints? -- - - - - - - - - - J. W. Wooten
From: mike@ali.bc.ca (Michael Cam) Newsgroups: comp.sys.next.programmer Subject: Re: Problems using the debug panel when loading bundles Date: 3 Jul 1996 16:38:35 GMT Organization: ALI Distribution: world Message-ID: <4re7mb$h2c@cetus.ali.bc.ca> References: <4rcgjj$5io@cetus.ali.bc.ca> In article <4rcgjj$5io@cetus.ali.bc.ca> mike@ali.bc.ca (Michael Cam) writes: > > Hi, > > I have an app that loads a bundle and uses it to create classes etc. > > I was able to set breakpoints using the debug panel that comes up through > the Edit.app menu when you hit the Debug button in PB. I could set > breakpoints in both the app's source files as well as the bundle's source > files (only after it is loaded). > > Now that I link Foundation libraries it seems that I cannot set > breakpoints this way anymore. Many thanks to all those that replied. The main problem is that Foundation's NSBundle strips the debug symbols by default so you must send [NSBundle stripAfterLoading:NO] before the code that loads the bundle in. ..Mike.
From: lozinski@cup.portal.com (Christopher A Lozinski) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer Subject: Nextstep Jobs Date: 3 Jul 1996 13:20:16 -0700 Organization: The Portal System (TM) Sender: pccop@unix.portal.com Distribution: world Message-ID: <152631@cup.portal.com> The NeXTSTEP job market is strong and getting stronger. It includes hardware and networking people, system administrators, developers, architects, and testers. Demand for developers is excessively strong in the Maryland, Washington, D.C., North Virginia area. It is also very strong in Texas. There is slow but steady demand in the rest of the big cities: e.g., San Francisco, Los Angeles, Seattle, Chicago, and New York. Away from the oceans there are a few other cities hiring. I know two very rural communities that are also looking for NeXTSTEP Developers. WebObjects demand is coming on strong. Really this means NeXTSTEP developers who have done a simple app on WebObjects. Telephony expertise is in strong demand. One company is willing to train in Telephony. Financial expertise is always in strong demand. One company has an 8-week training program in the financial industry for selected candidates. One Computer Systems Integrator and Internet Service Provider is willing to train developers in his industry. Computer Graphics expertise is in very strong demand all along the West Coast. One hiring company is even building Nuclear Power Plant Decommissioning Software. Two companies will let you trade your NeXTSTEP expertise for a good salary and stock. These companies are liable to go public within the year. Neither one is in California. a) WASHINGTON, D.C. METRO AREA: This area, which includes Washington, D.C., Maryland, and Western Virginia, has some of the best paying consulting positions. If you want to earn a lot of money and do not care too much what you are doing, these positions are for you. b) SEATTLE: Build Multimedia Web Applications in Seattle. You will be the primary developer on a team of multimedia experts, with the potential for very rapid group growth in this hot area. Take the initiative in figuring out what new services this company could/ should offer their customers, then go implement them. These are uncharted waters to be explored by someone with curiosity. Great for someone with an interest in cooking, gardening, and enjoying life. These guys work hard, play hard, and generally behave like a Silicon Valley Start Up, but with a sales growth curve of 100% per year. c) PITTSBURGH: This is a team of people building very complex medical radiation equipment. The employees are the design experts for the product and are treated as such. It is not worth their time to take any less than the best, and that means that you are respected and given considerable freedom and responsibility. This is not yet another database application. Rather, it is a complex scientific application with three-dimensional modelling and simulated annealing. They offer a highly competitive compensation package that includes health insurance and stock options. d) SECURITY CLEARANCES: People with TS SI/TK ISA ISSA SBI or EDI security clearances are desperately needed in the Washington, D.C. Area. You, too, can spy on foreigners. e) SAN FRANCISCO MARKETING: If you are not a developer, there is still opportunity for you. NeXTSTEP marketing professionals are needed in the San Francisco Bay Area. The company is making a major investment in NeXTSTEP technology and needs some people to help define the market requirements and help sell the product. f) HOLLAND, Sybase Administrator: We still have not found a person who can single handedly keep a production database up and running and optimize it. If we found such a person, we would take care of any visa issues, and pay such a mission-critical person well. If you do your job well, then in your free time you can develop some NeXTSTEP applications. The other interesting question is how to migrate such a production system to NT without disrupting production. g) SAN FRANCISCO BAY AREA: Become the corporate NeXTSTEP Fountainhead. Live just outside the Bay Area, close to the ocean. Wonderful location. Learn how to surf. h) SAN FRANCISCO BAY AREA: These guys are willing and able to pay the best salaries for the best people, so if you get the position, you will be working with the best possible team. Go for it. i) CHICAGO There is a strong demand for people with 6 months nextstep experience in the Chicago area. They will even train you in the financial industry. j) AUSTRALIA, Object-Oriented Database: I am very embarrassed to not yet have found a person with NeXTSTEP and object-oriented database experience for Australia. I guess most people have not visited Australia, or they would be much more eager to go. We were there just last year; give me a call if you would like me to tell you more about the country. (510) 795-6086. k) MIDWEST: Auspex System Administrator needed. Very low cost of living. Four beautiful seasons, and a very important and elegant application that needs support. l) LIFETIME EMPLOYMENT: Los Angeles Company is looking for a person who would like lifetime employment. They offer long-term job stability, and training. You get to take an existing application developed by a large consulting firm and initially maintain it and later upgrade it. m) NORTH CAROLINA: The hottest job market now is for the people building the Internet. Get out out of the corporate doldrums and into Internet building. Use your NeXTSTEP experience to land a position where you learn to be a system's integrator. You can settle down in the lovely foothills of North Carolina. Dress casually. The company is a strong NeXT advocate. You might even get to start a WebObjects consulting practice. n) LOS ANGELES: Needs graphic artists / animators with NeXTSTEP experience. Help create the hottest Hollywood blockbusters, and many of the T.V. animation movies. o) CHICAGO, Consulting Firm: Is looking for a project manager to supervise 30 people. It is a fun company filled with Smalltalk and Objective-C experts. Fun place to work, and it is a very senior position. p) MINNEAPOLIS ST. PAUL: Learn all about building the Internet. This company is building order entry systems for tracking sales of their digital services (Frame Relay, ISDN, SMDS, etc.). They are using an Oracle database to house the data, and NeXTSTEP/EOF to provide custom interfaces for the data. Duties include OOA/OOD, development, and some testing. There is both a developer and a project manager position available. q) COLORADO JOBS: For the first time, I have job openings in Colorado. This is a great place to live, to live it up if you are single, or to bring up a family if you are married. I worked there for a while. Wonderful mountains, very cheap living. Many people would do anything to find a job there, any job, just so that they could live there. You even get to do NeXTSTEP development there. r) TEXAS JOBS: This company is using NeXTSTEP to develop a Commodity Trading System capable of trading any commodity and is the only system capable of handling physical deliveries. They have found a market in the energy industry and are currently rolling out the product in two nations. They develop and customize the product for each customer, then install and support the product. Their coolest benefit is that they let you call anywhere in the world for free. Since they are willing to take foreign citizens, this works out great. I am very impressed by their project management. Every time that I call their most senior manager, he picks up his phone himself. We get our business done, and I get back to work. He must have a very smoothly running organization to be able to do that. s) WORK IN THE CITY OF YOUR CHOICE: I am now able to send your resume directly to hiring managers in the city of your choice. Whereas you may not know which companies use NeXTSTEP developers in a given city, or you may not be able to get past the HR department, I can send your resume directly to the in-box of the hiring managers of the companies you are interested in, in the cities where you would like to work. I know 104 hiring managers. If you are interested, send e-mail to <jobs@bpg.com>. Let me know what type of position you are looking for and what city you would like to work in. Even if I do not know a hiring manager in the city of your choice, I often know some NeXT developers working there, and I can forward your resume through them to the appropriate hiring manager. If you are interested in any of these jobs, please give me a call at (510) 795-6086 or better yet, send me your resume in NeXTMAIL RTF format to < jobs@bpg.com>
Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy From: benst@stuyts.nl (Ben Stuyts) Subject: Is posix file append bug fixed in 4.0? Message-ID: <DtzGB9.37w@stuyts.nl> Organization: Stuyts Engineering Haarlem BV Date: Wed, 3 Jul 1996 19:47:32 GMT Hi, Does anybody know whether the NS 3.3 file append bug (where only zeroes would be appended if a file was opened by multiple processes) that occurs in programs compiled with -posix is fixed in NS 4.0? There were also quite a few problems with the posix style tty drivers in 3.3. Are these improved? Thanks, Ben -- Ben Stuyts Stuyts Engineering BV Haarlem, The Netherlands Phone: +31 (0)23 5324609 Email: ben@stuyts.nl Fax: +31 (0)23 5421988 (NeXT/MIME Mail OK)
From: William.Clocksin@CL.cam.ac.uk Newsgroups: comp.sys.next.programmer Subject: PB reports wrong type? Date: 4 Jul 1996 14:05:01 GMT Organization: University of Cambridge, England Message-ID: <4rgj2d$6r@lyra.csx.cam.ac.uk> Has anyone else noticed on NS3.3 that when PB reports a 'multiple declarations' warning, that it cites 'char unsigned *' instead of 'unsigned char *'? For example, /NextDeveloper/Headers/appkit/NXBitmapImageRep.h is reported as defining '-(char unsigned *) data', when in fact the file (correctly) contains '-(unsigned char *) data'. (As does the line of my code that triggered the warning message). -- W F Clocksin Computer Laboratory University of Cambridge http://www.cl.cam.ac.uk/users/wfc/
From: dekorte@colt.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Control-Drag Cells Date: 4 Jul 1996 19:01:16 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4rh4ds$ksq@news.onramp.net> Is there some public source that does control-drag cells? (Like menus items can do in IB) -- Steve Dekorte "C++/Java, the COBOL of the 90s" NeXTstep consultant - Anaheim, CA (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: dekorte@colt.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4rh4jp$ksq@news.onramp.net> Control: cancel <4rh4jp$ksq@news.onramp.net> Date: 4 Jul 1996 19:12:00 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4rh520$l4j@news.onramp.net>
From: scratch@sunset.sce.carleton.ca (Craig Scratchley) Newsgroups: comp.sys.next.programmer Subject: Re: gcc for NeXT! (sort of) Date: 4 Jul 96 20:30:07 GMT Organization: Carleton University, Ottawa, Canada Message-ID: <scratch.836512207@sunset.sce.carleton.ca> References: <4qs7u0$c77@beta.datastorm.com> John A. Maier <johnam@beta.datastorm.com> writes: >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? I think the answer is to buy a version of NeXTSTEP Developer. My ld from version 3.2 seems to work fine for me. On a similar subject, does anybody know which version of gcc is used for Openstep/Mach 4.0? Craig Scratchley -- W. Craig Scratchley | internet: scratch@sce.carleton.ca Dept. of Systems and Computer Engineering | phone: (613) 520-5740 (Dept.) Carleton University | (613) 241-6952 (Home) Ottawa, ON, CANADA K1S 5B6 | fax: (613) 520-5727 (Dept.)
From: scratch@sunset.sce.carleton.ca (Craig Scratchley) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: gcc-2.7.2 on NS3.3-Sparcstation 5? Date: 4 Jul 96 20:41:26 GMT Organization: Carleton University, Ottawa, Canada Message-ID: <scratch.836512886@sunset.sce.carleton.ca> References: <4r78o4$h63@news.mty.itesm.mx> diego@conga.super.unam.mx (Diego Zamboni) writes: >Hi: >I wanted to compile gcc-2.7.2 on my Sparcstation 5 running NS3.3, both to >have a recent version of gcc and to be able to compile g77, the GNU FORTRAN >compiler. However, apparently gcc doesn't yet support NEXTSTEP running on >sparc processors. >Does anybody have any ideas of getting around this problem? Perhaps Openstep/Mach 4.0 provides a recent-enough version of gcc. Does anybody know? Craig -- W. Craig Scratchley | internet: scratch@sce.carleton.ca Dept. of Systems and Computer Engineering | phone: (613) 520-5740 (Dept.) Carleton University | (613) 241-6952 (Home) Ottawa, ON, CANADA K1S 5B6 | fax: (613) 520-5727 (Dept.)
From: kinau@zimmer.CSUFresno.EDU (Kin Hung Au) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: Mosaic 2.6 for NeXT black box Date: 4 Jul 1996 09:23:56 -0700 Organization: California State University, Fresno Message-ID: <4rgr6s$lej@zimmer.CSUFresno.EDU> Hi folks, Has anyone ever compiled Mosaic 2.6 for NeXT black box? Could anyone send me a Makefile to compile Mosaic 2.6 or give me a hint? Is there a ftp site to keep binary file for NeXT black box? Please send me an email. Thanks in advance!! =) --Kin ****************************************************************************** Kin Hung Au Internet Address: kinau@csufresno.edu Software Specialist California State University, Fresno Tel# 209-278-7509 Computing, Communications & Media Services FAX# 209-278-4660 Fresno CA 93740 http://maxwell.phys.csufresno.edu:8001/Dept/staff/kinau.html ******************************************************************************
From: tuparev@EMBL-Heidelberg.DE (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Control-Drag Cells Date: 5 Jul 1996 10:12:12 GMT Organization: EMBL Heidelberg Distribution: world Message-ID: <4ripps$g04@lion.embl-heidelberg.de> References: <4rh4ds$ksq@news.onramp.net> In article <4rh4ds$ksq@news.onramp.net> dekorte@colt.suite.com (Steve Dekorte) writes: > > Is there some public source that does control-drag cells? > (Like menus items can do in IB) NiftyMatrix (ScrollDoodScroll.app) and the new MiscDragMatrix .. on the way to be submitted... -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387305 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
From: flight@mathi.uni-heidelberg.de (Gregor Hoffleit) Newsgroups: comp.sys.next.programmer Subject: Please upload GNU sources from OS 4.0! Date: 5 Jul 1996 10:10:31 GMT Organization: Mathematisches Institut, Uni Heidelberg Message-ID: <4ripmn$f7a@sun0.urz.uni-heidelberg.de> Please, could somebody with OS/MachOS 4.0 upload the GNU sources included in the distribution to an archive ? 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: Paul Heffernan <phef@cedar.co.uk> Newsgroups: comp.sys.next.programmer Subject: Re: Control-Drag Cells Date: Fri, 5 Jul 1996 11:24:32 +0100 Organization: Cedar Systems Distribution: world Message-ID: <rUAEWAAg1O3xEwAM@cedar.co.uk> References: <4rh4ds$ksq@news.onramp.net> MIME-Version: 1.0 In article <4rh4ds$ksq@news.onramp.net>, Steve Dekorte <dekorte@colt.suite.com> writes > >Is there some public source that does control-drag cells? >(Like menus items can do in IB) > >-- >Steve Dekorte "C++/Java, the COBOL of the 90s" >NeXTstep consultant - Anaheim, CA (NeXTmail,MIME) >http://www.batech.com/~dekorte/dekorte/index.html > Yes. Get hold of Thomas Burkholder's excellent TTools from one of the ftp sites - his Ranker class should be just what you need. 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: tiggr@es.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.lang.misc,comp.sys.next.programmer,comp.lang.objective-c,comp.sys.hp.hpux,comp.lang.java Subject: ANNOUNCEMENT: TOM, version 0.01 Followup-To: comp.lang.misc Date: 05 Jul 1996 15:32:01 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.es.ele.tue.nl Message-ID: <x7afxeomge.fsf@tom.es.ele.tue.nl> The first release of `tom' is available. TOM is an object-oriented programming language. It was inspired by Objective-C, but it does not suffer a `C' heritage and its `Objective' part is richer and more consistent. For the same reason, tom is also a nicer language than Java. This first release has been tested on hppa-hpux and m68k-next machines. It also runs on i*86-linux and i*86-freebsd. Porting to a different machine is not difficult, and you are encouraged to do so. Below is a list of things addressed by tom. For more information, see "http://tom.es.ele.tue.nl:8080/". --Tiggr Extensibility In Smalltalk and Objective-C methods can be added to an already existing class, but instance variables can not. In C++, a class is defined by its one and only definition. In tom, a class need not be defined by one monolithic definition: its definition can be split into multiple parts, called extensions. An extension can add instance and class variables; it can declare new superclasses from which to inherit; and it can add, or replace, methods. Usability With tom, the semantics of the language do not change with the availability of sources. Put more strongly, the possibilities of the language are not restricted in any way if sources are not available. This in contrast to, for example, Eiffel, where the availability of only the short form of a class severly hampers the possibilities of subclassing. As another example, given a C++ library of which one wants to replace a class by a different one. In that case, the member variables must remain identical if referred to from within one of the other objects in the library. The same is true for virtual methods if invoked from the other objects. If any of the member functions of the class to be replaced is inlined in the other objects, the possibility of successfully completing this venture within the original goals is further reduced. Uniformity C++ provides signatures, Java calls them interfaces, and Objective-C refers to them as protocols. Apart from some of the more obscure problems, such as an Objective-C class object not being able to adhere to a protocol, these constructs are orthogonal to the normal class hierarchy. In Java and Objective-C, they provide a means for multiple inheritance of behaviour declaration; in C++ it is an attempt to cover some of its many deficiencies. In tom, the functionality of protocols (and signatures and interfaces) is simply provided by the class hierarchy itself; this is implied by: multiple inheritance, deferred methods, extensibility, a strict distinction between instances and classes, and the fact that instances without any instance variables can be inherited by both classes and instances. Static TOM is strongly type checked at compile time; various (builtin) basic types and the (user defined) object classes are discerned. Methods are overloaded on both the argument types and the types of the return values. The class of the receiver of a message (or call it the object of which a member function is invoked) is used to decide the availability of the method, and the types of the values accepted and returned. However, the method actually invoked is not statically bound: all methods are dynamically bound. In the context of extensibility and usability it also implies that all object references have the same type. Dynamic At runtime tom provides full typing information on objects, their instance and class variables, and arguments to and return value from methods. All methods are dynamically bound. Methods are provided to read and set instance and class variables given their name; to retrieve an object's class; to query a class for its position in the inheritance hierarchy; to send an object a message of which the selector need not be known at compile time (a bit like function pointers in C++, but being dynamically bound much more flexible; more like perform in SmallTalk and Objective-C, but without restrictions on the types and number of arguments). Storage Management All objects are allocated from the garbage collected heap. This implies that it is impossible to have problems with dangling references, previously valid pointers suddenly pointing into the void, memory leaks, etc. In tom, an object can not be a part of an entity of which the lifetime is unrelated to, and possibly shorter than, the lifetime of the object. This containment rule dictates that objects can not be allocated from the stack and they can not be part of another object. (Arrays seem a valid counterexample, but in tom object arrays hold references, never the actual objects.) Return Values In most languages, there is an arbitrary limit on the number of values a method or function can return without reverting to tricks. This limit is 1, and pointer arguments are needed to `return' more than one value, or the return values need to be packed into an aggregate entity which can then be returned as the single return value. In Common Lisp, multiple-valued returns are possible, but they are `special'. TOM imposes no limit on the number of values that can be returned from a method. Exceptions In most languages (C++, Java, NeXT's Objective-C libraries, to name but a few), an exception handler is executed in its own context instead of the context of the exception raiser, implying that it is impossible for the handler to direct the actions of the condition raiser---the stack has already been unwound and nothing can be done about that. The tom exception mechanism is modeled after Common Lisp conditions. A tom condition can be signaled; if none of the condition handlers is interested in handling this condition, execution continues in the context of the condition signaler, as if the signal were a normal method invocation. A condition handler can direct the condition signaler by setting the value to be returned from the signal invocation. It can also decide to perform a non-local exit, the equivalent of a C longjmp. It is of course possible for a method not to be able to continue execution after a condition. In this case it can force the condition to be handled, by raising it instead of signaling. Failure to handle a raised condition through a non-local exit results in an abnormal program termination.
From: William.Clocksin@CL.cam.ac.uk Newsgroups: comp.sys.next.programmer Subject: Listener/Speaker remote methods needed? Date: 5 Jul 1996 14:26:19 GMT Organization: University of Cambridge, England Message-ID: <4rj8mb$1r7@lyra.csx.cam.ac.uk> Do the Listener/Speaker remote methods - (int)msgDirectory:(char * const *)fullPath ok:(int *)flag - (int)msgVersion:(char * const *)aString ok:(int *)flag - (int)msgFile:(char * const *)fullPath ok:(int *)flag - (int)msgPrint:(const char *)fullPath ok:(int *)flag - (int)msgSelection:(char * const *)bytes length:(int *)len asType:(const char *)aType ok:(int *)flag etc still need to be defined in apps for NS3.3 and later?
From: lozinski@cup.portal.com (Christopher A Lozinski) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer Subject: Nextstep Job Market Date: 5 Jul 1996 10:40:09 -0700 Organization: The Portal System (TM) Sender: pccop@unix.portal.com Distribution: world Message-ID: <152651@cup.portal.com> The NeXTSTEP job market is strong and getting stronger. It includes hardware and networking people, system administrators, developers, architects, and testers. Demand for developers is excessively strong in the Maryland, Washington, D.C., North Virginia area. It is also very strong in Texas. There is slow but steady demand in the rest of the big cities: e.g., San Francisco, Los Angeles, Seattle, Chicago, and New York. Away from the oceans there are a few other cities hiring. I know two very rural communities that are also looking for NeXTSTEP Developers. WebObjects demand is coming on strong. Really this means NeXTSTEP developers who have done a simple app on WebObjects. Telephony expertise is in strong demand. One company is willing to train in Telephony. Financial expertise is always in strong demand. One company has an 8-week training program in the financial industry for selected candidates. One Computer Systems Integrator and Internet Service Provider is willing to train developers in his industry. Computer Graphics expertise is in very strong demand all along the West Coast. One hiring company is even building Nuclear Power Plant Decommissioning Software. Two companies will let you trade your NeXTSTEP expertise for a good salary and stock. These companies are liable to go public within the year. Neither one is in California. a) WASHINGTON, D.C. METRO AREA: This area, which includes Washington, D.C., Maryland, and Western Virginia, has some of the best paying consulting positions. If you want to earn a lot of money and do not care too much what you are doing, these positions are for you. b) SEATTLE: Build Multimedia Web Applications in Seattle. You will be the primary developer on a team of multimedia experts, with the potential for very rapid group growth in this hot area. Take the initiative in figuring out what new services this company could/ should offer their customers, then go implement them. These are uncharted waters to be explored by someone with curiosity. Great for someone with an interest in cooking, gardening, and enjoying life. These guys work hard, play hard, and generally behave like a Silicon Valley Start Up, but with a sales growth curve of 100% per year. c) PITTSBURGH: This is a team of people building very complex medical radiation equipment. The employees are the design experts for the product and are treated as such. It is not worth their time to take any less than the best, and that means that you are respected and given considerable freedom and responsibility. This is not yet another database application. Rather, it is a complex scientific application with three-dimensional modelling and simulated annealing. They offer a highly competitive compensation package that includes health insurance and stock options. d) SECURITY CLEARANCES: People with TS SI/TK ISA ISSA SBI or EDI security clearances are desperately needed in the Washington, D.C. Area. You, too, can spy on foreigners. e) SAN FRANCISCO MARKETING: If you are not a developer, there is still opportunity for you. NeXTSTEP marketing professionals are needed in the San Francisco Bay Area. The company is making a major investment in NeXTSTEP technology and needs some people to help define the market requirements and help sell the product. f) HOLLAND, Sybase Administrator: We still have not found a person who can single handedly keep a production database up and running and optimize it. If we found such a person, we would take care of any visa issues, and pay such a mission-critical person well. If you do your job well, then in your free time you can develop some NeXTSTEP applications. The other interesting question is how to migrate such a production system to NT without disrupting production. g) SAN FRANCISCO BAY AREA: Become the corporate NeXTSTEP Fountainhead. Live just outside the Bay Area, close to the ocean. Wonderful location. Learn how to surf. h) SAN FRANCISCO BAY AREA: These guys are willing and able to pay the best salaries for the best people, so if you get the position, you will be working with the best possible team. Go for it. i) CHICAGO There is a strong demand for people with 6 months nextstep experience in the Chicago area. They will even train you in the financial industry. j) AUSTRALIA, Object-Oriented Database: I am very embarrassed to not yet have found a person with NeXTSTEP and object-oriented database experience for Australia. I guess most people have not visited Australia, or they would be much more eager to go. We were there just last year; give me a call if you would like me to tell you more about the country. (510) 795-6086. k) MIDWEST: Auspex System Administrator needed. Very low cost of living. Four beautiful seasons, and a very important and elegant application that needs support. l) LIFETIME EMPLOYMENT: Los Angeles Company is looking for a person who would like lifetime employment. They offer long-term job stability, and training. You get to take an existing application developed by a large consulting firm and initially maintain it and later upgrade it. m) NORTH CAROLINA: The hottest job market now is for the people building the Internet. Get out out of the corporate doldrums and into Internet building. Use your NeXTSTEP experience to land a position where you learn to be a system's integrator. You can settle down in the lovely foothills of North Carolina. Dress casually. The company is a strong NeXT advocate. You might even get to start a WebObjects consulting practice. n) LOS ANGELES: Needs graphic artists / animators with NeXTSTEP experience. Help create the hottest Hollywood blockbusters, and many of the T.V. animation movies. o) CHICAGO, Consulting Firm: Is looking for a project manager to supervise 30 people. It is a fun company filled with Smalltalk and Objective-C experts. Fun place to work, and it is a very senior position. p) MINNEAPOLIS ST. PAUL: Learn all about building the Internet. This company is building order entry systems for tracking sales of their digital services (Frame Relay, ISDN, SMDS, etc.). They are using an Oracle database to house the data, and NeXTSTEP/EOF to provide custom interfaces for the data. Duties include OOA/OOD, development, and some testing. There is both a developer and a project manager position available. q) COLORADO JOBS: For the first time, I have job openings in Colorado. This is a great place to live, to live it up if you are single, or to bring up a family if you are married. I worked there for a while. Wonderful mountains, very cheap living. Many people would do anything to find a job there, any job, just so that they could live there. You even get to do NeXTSTEP development there. r) TEXAS JOBS: This company is using NeXTSTEP to develop a Commodity Trading System capable of trading any commodity and is the only system capable of handling physical deliveries. They have found a market in the energy industry and are currently rolling out the product in two nations. They develop and customize the product for each customer, then install and support the product. Their coolest benefit is that they let you call anywhere in the world for free. Since they are willing to take foreign citizens, this works out great. I am very impressed by their project management. Every time that I call their most senior manager, he picks up his phone himself. We get our business done, and I get back to work. He must have a very smoothly running organization to be able to do that. s) WORK IN THE CITY OF YOUR CHOICE: I am now able to send your resume directly to hiring managers in the city of your choice. Whereas you may not know which companies use NeXTSTEP developers in a given city, or you may not be able to get past the HR department, I can send your resume directly to the in-box of the hiring managers of the companies you are interested in, in the cities where you would like to work. I know 104 hiring managers. If you are interested, send e-mail to <jobs@bpg.com>. Let me know what type of position you are looking for and what city you would like to work in. Even if I do not know a hiring manager in the city of your choice, I often know some NeXT developers working there, and I can forward your resume through them to the appropriate hiring manager. If you are interested in any of these jobs, please give me a call at (510) 795-6086 or better yet, send me your resume in NeXTMAIL RTF format to < jobs@bpg.com>
From: thad@ion.rice.edu (Thad Harroun) Newsgroups: comp.sys.next.programmer Subject: NXImage question from novice Date: Fri, 05 Jul 1996 13:20:09 -0500 Organization: Rice University Message-ID: <thad-0507961320090001@news.rice.edu> I'm a novice programmer, but thanks to NeXT have written an image analysis program that works really well for what I want, and now I'm trying to optimize it because it is kind of large. One thing that I just can't understand is the flip in the image as it's displayed in the window: the origin of the window is in the lower-left corner, the data point data[0] is in the upper-left. This gets kind of confusing when I need to do polar coordinates from a new origin in the middle of the image. Why is this happening? If I NX_SOVER from the point 0,0, why then doesn't it start with data[0] at that point? Thanks for an explanation, Thad -- Thad Harroun Rice University thad@ion.rice.edu
From: monckton@arrow.mech.ubc.ca (Simon Monckton) Newsgroups: comp.sys.next.programmer Subject: Printing Renderman Date: 5 Jul 1996 19:01:28 GMT Organization: University of British Columbia, Vancouver, B.C., Canada Message-ID: <4rjoq8$g1h@nntp.ucs.ubc.ca> Hello all, I am stumped. I have successfully employed an N3DCamera object to display a wireframe scene - black shaded surfaces, matte textures on a white background. Everything works great on screen, but as soon as I try to print all I get is all black! Suggestions? Simon M.
From: Mark Strand <marks@soli.inav.net> Newsgroups: comp.sys.next.programmer Subject: Help me understand archiving (sorta long) Date: Sat, 06 Jul 1996 00:20:50 -0500 Organization: Internet Navigator, Inc. Message-ID: <31DDF7B2.6F0B@soli.inav.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Title says most of it. I have in the past used the Storage class & NXWriteTypes to save simple things like lists of names, etc. to a disk file. What I'm wondering about is archiving objects. When is this needed, and what exactly does it mean? Is it just saving the instance variables? (I can't see why you'd need to save the methods out to disk; they're part of the *.o files in your .app, and so your app is obviously aware of them.) ------------------------------- PALETTES VS. THE COMPILED APPLICATION Also I'm starting to create custom palettes with some non-UI objects, and wondering about the role of ok:, read:, and write: in the palette object vs. the "real" application. What I mean is, Say I have a simple object called "Calculator" and I want this to live on a palette so I can easily drag it into a project. All it does it add two values and print the result. So I create the palette project, and use an IBInspector to set whether it rounds numbers or not. Done correctly, this stuff works even in Test Interface mode. (See Paul Hegarty's demo at Object World to see what I'm talking about). Of course, this will also work when the program is compiled. Now my question: Where does this roundValues=YES; get built into the application? I understand that the IBInspector simply provides an interface to that object's instance variables. If I select it as YES in the inspector for that object (via a checkbox), and build the project, is it then limited to be yes? Does it get archived into the associated .nib file where the object exists, or somewhere else? And can it be changed later via some prefs editor for that app, or are values in the .nib "hard-wired" and semi-permanent? What issues do i need to be concerned about when designing an object that behaves nicely in a palette and also in a compiled application (it's kinda unclear from what I read, what isTestingInterface exactly does)? ------------------------ This leads to more of a general question... to use a simple example, let's say I want the window position of their "Tools" window to be restored each time they launch. Windows, like most classes have a build in read:/write: functionality.. how to best exploit this? Create a settings file in ~/.NeXT/MyApp, or use the defaults database with the co-ordinates? I understand most of the NeXTSTEP stuff pretty decent, but I can't seem to figure out how to correctly use read: and write: and where to actually read: from and write:to. A file? What guidelines should I follow for this file? I apologize for all the questions but I really want to get this stuff straight in my head and would be eternally grateful for any pointers.
From: Brett Allen Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: uuencode/uudecode Date: 6 Jul 1996 08:36:24 GMT Organization: Only On A Good Day Message-ID: <4rl8i8$hqk@reader1.reader.news.ozemail.net> hi, does anybody know of publicly available uuencode/uudecode source code or library? thanx --- Brett
From: Brett Allen Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: uuencode/uudecode Date: 6 Jul 1996 08:39:26 GMT Organization: Only On A Good Day Message-ID: <4rl8nu$hfh@reader1.reader.news.ozemail.net> hi, does anybody know of publicly available uuencode/uudecode function(s) or library ? thanx --- Brett
From: duboisj@mathcs.carleton.edu (Josh DuBois) Newsgroups: comp.sys.next.programmer Subject: bcp for Sybase? Date: 1 Jul 1996 21:37:03 GMT Organization: Carleton College, Northfield, MN, USA Message-ID: <4r9gdv$6ko@zorak.acns.carleton.edu> Hey, there. I am under the impression that Sybase for NeXTstep should come with a utility called bcp for dumping database info to text files. I'd think one would find this in /usr/sybase/bin, but I don't see it there. Anyone know where it's normally kept? Should I ask my sysadmin about a possible incomplete installation, or does one have to special-order this beast? Thanks, Josh -- Josh DuBois duboisj@is.com
From: jeremy@inetnebr.com (Jeremy Bettis) Newsgroups: comp.sys.next.misc,comp.sys.next.software,comp.sys.next.programmer Subject: Re: Puzzled about OpenStep Date: 6 Jul 1996 14:08:50 GMT Organization: Internet Nebraska Message-ID: <4rls1i$85j@duck.inetnebr.com> References: <4rb5g0$qc@netty.york.ac.uk> <4rbe9r$4h@dfw-ixnews10.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-User: jeremy To: aisbell@cubicsol.com (Art Isbell) In article <4rbe9r$4h@dfw-ixnews10.ix.netcom.com>, aisbell@cubicsol.com (Art Isbell) writes: > NeXT's Web site is offering a sneak preview (beta version) of the > OPENSTEP/NT development tools which would be necessary to build an > OPENSTEP/NT app. I understand that Microsoft Visual C++ must be installed as > well because the NEXTSTEP linker (ld) hasn't been ported yet. That was true in Prerelease 2 but prerelease 3 comes with the linker, no Visual C++ necessary, as far as I can tell. ( Or course I still had VC++ installed from the prior testing.) -- Jeremy Bettis -*- PGP public key available,send mail with subj "Send pgp key" Home: jeremy@tddi.inetnebr.com Work: jeremy@hksys.com NeXTMAIL/MIME/PGP accepted
From: jeremy@inetnebr.com (Jeremy Bettis) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: gcc-2.7.2 on NS3.3-Sparcstation 5? Date: 6 Jul 1996 14:11:27 GMT Organization: Internet Nebraska Message-ID: <4rls6f$85j@duck.inetnebr.com> References: <4r78o4$h63@news.mty.itesm.mx> <scratch.836512886@sunset.sce.carleton.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-User: jeremy To: scratch@sunset.sce.carleton.ca (Craig Scratchley) In article <scratch.836512886@sunset.sce.carleton.ca>, scratch@sunset.sce.carleton.ca (Craig Scratchley) writes: > diego@conga.super.unam.mx (Diego Zamboni) writes: > >>I wanted to compile gcc-2.7.2 on my Sparcstation 5 running NS3.3, both to >>have a recent version of gcc and to be able to compile g77, the GNU FORTRAN >>compiler. However, apparently gcc doesn't yet support NEXTSTEP running on >>sparc processors. > > Perhaps Openstep/Mach 4.0 provides a recent-enough version of > gcc. Does anybody know? > The prerelease does not. I suspect that the final release won't either. -- Jeremy Bettis -*- PGP public key available,send mail with subj "Send pgp key" Home: jeremy@tddi.inetnebr.com Work: jeremy@hksys.com NeXTMAIL/MIME/PGP accepted
From: ifeulner@xenon.cube.de (Ingo Feulner) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: gcc-2.7.2 on NS3.3-Sparcstation 5? Date: 7 Jul 1996 10:11:37 GMT Organization: Home, sweet home in Boeblingen, Germany Message-ID: <slrn4tv3ap.k0f.ifeulner@xenon.stgt.sub.org> References: <4r78o4$h63@news.mty.itesm.mx> <4rls6f$85j@duck.inetnebr.com> On 6 Jul 1996 14:11:27 GMT, Jeremy Bettis <jeremy@inetnebr.com> wrote: >The prerelease does not. I suspect that the final release won't either. You're right. - Ingo. -- Smail: Ingo Feulner, Wolfacher Weg 19, 71034 Boeblingen, Germany Email: ifeulner@xenon.cube.de
From: mdenchy@prolog.net (mark denchy) Newsgroups: comp.sys.next.programmer Subject: NextStep and 16Bit library Date: Sun, 07 Jul 1996 15:08:33 GMT Organization: Yuasa-Exide, Inc. Message-ID: <4rokj0$b7t@ns2.ptd.net> Will NextStep support 16Bit calls under the Intel environment? I am evaluating development platforms and I am unfortunately constrained with one development library that is still 16 bit. The company supposedly has a Mac 32 Bit version and will be porting it, but that remains to be seen. I think the answer is no, but I want to ask someone who knows. Our environment is HP3000 and Win95 clients with TCP/IP. Any help would be appreciated. Mark Denchy Yuasa-Exide, Inc.
From: gundavel@best.com (Sai Gundavelli) Newsgroups: comp.sys.next.programmer Subject: Offshore Programming Services Date: 7 Jul 1996 11:31:11 -0700 Organization: Solix Systems Incorporation Distribution: world Message-ID: <4rovpf$im7@shellx.best.com> ATTENTION: SOFTWARE ENTERPRENEURS +++++++++++++++++++++++++++++++++ Solix Systems Inc based in Sanjose, California can provide you off-shore software development services from INDIA. OUR EXPERTISE ~~~~~~~~~~~~~ Application Level Programming involving any of the tools e.g., Visual C++ Visual Basic FoxPro Dbase Clipper Borland C++ Oracle Informix Sybase Power Builder Symantec C++ Window NT Unix DOS OS/2 ....... ......... ......... .... ..... ..... .. .. .. OUR COMMUNICATION LINKAGES ~~~~~~~~~~~~~~~~~~~~~~~~~~ 64KBPS Satellite Link to INDIA, all our development staff available on Email. We have access to Ftp, Netscape, Telnet etc., OUR STRENGTH ~~~~~~~~~~~~ Marketing and Management: 6 Total # of Engineers: 75 REFERENCES ~~~~~~~~~~ Available upon request. WHY YOU SHOULD WORK WITH US ~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1. Any code written by us, will meet the ISO-9000 quality standards. 2. We guarantee on our ability. If we do not meet your expections within the first two weeks, we will gladly get away, without charging you a penny. 3. Provide you with a good economical solution. WHOM TO CONTACT ~~~~~~~~~~~~~~~ Name: Ardy Hagen, ardy@solix.com Email: sales@solix.com Phone: 408-945-9099 Fax: 408-945-9365 Web: http://www.solix.com
From: M. WILKINS <mwilkins@earthlink.net> 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,comp.sys.northstar,comp.sys.nsc.32k,comp.sys.oric,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.s Subject: I'LL TELL YOUR FUTURE, NOW!!! Date: 7 Jul 1996 22:53:52 GMT Organization: Earthlink Network, Inc. Distribution: inet Message-ID: <01bb6c56.57337d00$3f5bface@mwilkins> I'LL TELL YOUR FUTURE, NOW!!!!! JUST CALL 1-900-868-4100 EXT. 7607 $3.99per min. must be 18yrs Serv-U(619)645-8434
From: rbraver@ohww.norman.ok.us Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4rovpf$im7@shellx.best.com> Date: 7 Jul 1996 23:33:37 GMT Control: cancel <4rovpf$im7@shellx.best.com> Message-ID: <cancel.4rovpf$im7@shellx.best.com> Sender: gundavel@best.com (Sai Gundavelli) Spam cancelled. Notice ID: 19960707.24. See news.admin.net-abuse.announce or http://spam.ohww.norman.ok.us/spam_notices/19960707.24.html for complete report. Original Subject: Offshore Programming Services
Newsgroups: comp.sys.next.programmer From: Nitezki@NiDat.sub.org (Peter Nitezki) Subject: Re: uuencode/uudecode Message-ID: <Du66yE.6x9@nidat.sub.org> Sender: nitezki@nidat.sub.org (Peter Nitezki) Organization: private site of Peter Nitezki, Kraichtal, Germany References: <4rl8i8$hqk@reader1.reader.news.ozemail.net> Date: Sun, 7 Jul 1996 11:08:37 GMT In article <4rl8i8$hqk@reader1.reader.news.ozemail.net> Brett Allen writes: > hi, > > does anybody know of publicly available uuencode/uudecode source > code or library? > Oh boy, a software archeologist ;-) That should be found in the source code of BSD Unix. There are several different archives on CD/ROM now (FreeBSD by Walnut Creek, for an example). -- 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: "Patrick Ketelaar" <ketelaar@ix.netcom.com> 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,comp.sys.northstar,comp.sys.nsc.32k,comp.sys.oric,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc Subject: Re: I'LL TELL YOUR FUTURE, NOW!!! Date: Sun, 7 Jul 1996 22:36:14 -0400 Organization: Netcom Distribution: inet Message-ID: <01bb6c76.7bdde120$6589d3c6@#ketelaar> References: <01bb6c56.57337d00$3f5bface@mwilkins> Wait- I can tell my own future- I'm not gonna call, not gonna spend 3.99 a minute, gonna save for the win version of the pilot sdk. pat -- Patrick Ketelaar- Ketelaar@ix.netcom.com Disney Maps for Newton http://www.netcom.com/~ketelaar/orion.html The above is my opinion, I could be wrong, but who cares? > M. WILKINS <mwilkins@earthlink.net> wrote in article <01bb6c56.57337d00$3f5bface@mwilkins>... > I'LL TELL YOUR FUTURE, NOW!!!!! > > JUST CALL 1-900-868-4100 EXT. 7607 > > $3.99per min. > must be 18yrs > Serv-U(619)645-8434 > > >
Newsgroups: comp.sys.next.programmer From: glen@prosoft.com (Glen Biagioni) Subject: Re: bcp for Sybase? Message-ID: <Du7nFw.6Cp@prosoft.com> Sender: usenet@prosoft.com Organization: ProSoft Solutions, Inc. References: <4r9gdv$6ko@zorak.acns.carleton.edu> Date: Mon, 8 Jul 1996 06:02:20 GMT Josh DuBois writes > > Hey, there. > I am under the impression that Sybase for NeXTstep > should come with a utility called bcp for dumping database info to > text files. I'd think one would find this in /usr/sybase/bin, but I don't > see it there. Anyone know where it's normally kept? Should I ask my sysadmin > about a possible incomplete installation, or does one have to special-order > this beast? > > Thanks, > Josh > > > -- > Josh DuBois > duboisj@is.com If I recall correctly, it disappeared in one of the NextStep releases. I saved an older version which seems to work fine. Just get it from someone who had an earlier release of Sybase for NextStep. -- Glen Biagioni ProSoft Solutions Inc. glen@prosoft.com (NeXTmail Welcome) Bus:(604)324-3311 Fax:(604)324-9431
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.programmer Subject: Windows 3.1 clients and NS servers? How? Date: 8 Jul 1996 08:10:10 GMT Organization: Tech Net GmbH Message-ID: <4rqfp2$n0d@ddfservb.technet.net> Hi everybody, Following problem: Here we have a NS 3.3 Intel system running some servers. And there we have a normal PC running Windows 3.1. We would like to run some client on the Windows PC. Does anyone have some experience in this sort of thing? Expecially, what kind of communication technologie is best/available. Of cource it would be possible to just use sockets, but --> too much work. Some RPC mechanism would be much easier to use. For example ONC RPC, which was once called Sun RPC. The man page of rpcgen does not mention the word ONC, so I guess the RPC suppied with NS is some version of Sun RPC. It it compatible to ONC RPC? And then I need some ONC RPC implementaion for Win3.1... The other option is DCE RPC. It some implementaion available for NS? For Win3.1??? ??? -- Constantin Szallies, Energotec GmbH szallies@energotec.de 49211-9144012
From: s124177@student.uq.edu.au (Simon Chih-L Han) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT Mouse. Date: Mon, 08 Jul 1996 21:24:15 +1000 Organization: Centre for Research in Vascular Biology, the UQ Message-ID: <s124177-0807962124150001@pipi.anatomy.uq.oz.au> References: <31D77049.4502@students.wisc.edu> In article <31D77049.4502@students.wisc.edu>, "Andrew M. Priasmoro" <ampriasm@students.wisc.edu> wrote: > Hi, > > Does NeXT mouse use bus type connector? If not, what type of connector > does it use? Does any one know if I could use NeXT mouse for IBM PC > system? > Thanks in advance. > > > Andrew. The next adb keyboard can be used for mac without problems. I am not sure the adb mouse will. Most NeXTer think the mac mouse is better than NeXT mouse.
From: pmcneill@csed-111.csed.ida.org (Patrick McNeill) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: uuencode/uudecode Followup-To: comp.lang.objective-c,comp.sys.next.programmer Date: 8 Jul 1996 12:26:40 GMT Organization: IDA, Alexandria, Virginia Message-ID: <4rquq1$etr@news.ida.org> References: <4rl8i8$hqk@reader1.reader.news.ozemail.net> BrettAllen wrote: : does anybody know of publicly available uuencode/uudecode source code or library? What type of system? For Unix boxes, the source is available at ftp.cdrom.com. For DOS, it's probably also on that site, but I'm not entirely sure... =Patrick http://www.tjhsst.edu/people/pmcneill/ pmcneill@ida.org
From: William.Clocksin@CL.cam.ac.uk Newsgroups: comp.sys.next.programmer Subject: Examples for opening those files that look like directories? Date: 8 Jul 1996 14:27:50 GMT Organization: University of Cambridge, England Message-ID: <4rr5t6$l22@lyra.csx.cam.ac.uk> Are there any MiniExamples (or equivalent) that deal with how to open/close those files that look like directories? Or can you just say NXMapFile(".../file.ap/data.f", ...), where file.ap is what the user sees, and data.f is the file inside? -- William Clocksin http://www.CL.cam.ac.uk/users/wfc
From: thor@zems.fer.hr (Thor Legvold) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: Compile problems (pgp 2.6.3i) Date: 8 Jul 1996 17:53:55 GMT Organization: RITEH Distribution: world Message-ID: <4rrhvj$6fr@bagan.srce.hr> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Hi. I recently downloaded pgp 2.6.3i and found predefined definitions for next, next-inte, next486 and nextHP in the makefile. However, when I make I get an error: pgp.c: In function `ParseRecipients': pgp.c:2922: warning: assignment makes pointer from integer without a cast and in linking I get an exit because of undefined symbol: ld: Undefined symbols: _strdup *** Exit 1 Not being a programming genius (not even a programmer :-) I'm not sure where I should be lloking to fix this. I seem to recall a similar problem last time I compiled some generic Unix code which I managed to solve, but I don't remember how.... NeXTSTEP fIP 3.3, Developer 3.2. Regards, Thor Legvold presently in exile in Croatia -- Thor Legvold "Where I'm from is not nearly as relevant as where I've been" -me
From: kwong@morgan.ucs.mun.ca (Kai S. Wong) Newsgroups: comp.sys.next.programmer,comp.sys.next.marketplace Subject: WTB: Purple Book and Programming in EOF Date: 8 Jul 1996 18:18:10 GMT Organization: Memorial University of Newfoundland Message-ID: <4rrjd2$jv0@coranto.ucs.mun.ca> I am looking for used book "Programming The Display Postscript System withe NeXTSTEP" (the Purple book) written by Adobe and published by Addison-Wesley 1992 and books in EOF. Email to me kwong@plato.ucs.mun.ca --kai-- -- email: kwong@morgan.ucs.mun.ca url: http://web.cs.mun.ca/~kwong/
From: stuckey@mrcnext.cso.uiuc.edu (Anthony J. Stuckey) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: uuencode/uudecode Date: 8 Jul 1996 19:40:36 GMT Organization: University of Illinois at Urbana Message-ID: <4rro7k$ctv@vixen.cso.uiuc.edu> References: <4rl8i8$hqk@reader1.reader.news.ozemail.net> <4rquq1$etr@news.ida.org> pmcneill@csed-111.csed.ida.org (Patrick McNeill) writes: >BrettAllen wrote: >: does anybody know of publicly available uuencode/uudecode source code or library? >What type of system? For Unix boxes, the source is available at ftp.cdrom.com. >For DOS, it's probably also on that site, but I'm not entirely sure... Given how completely simple of a conversion uuencode/decode is, you could probably implement it as a shell script or batch file without difficulty. it's only about a hundred lines of C all told, and much of that is wrapper to the actual decoding. -- Anthony J. Stuckey stuckey@mrcnext.cso.uiuc.edu "And if you frisbee-throw a universe where does it go?" -- Steve Blunt. GCS/S -d+@ p c(++) l u+ e+(-) m+(*) s+++/-- !n h(*) f+ g+ w+ t+@ r y? KiboNumber == 1
From: Žfey@biotech.washington.edu (Curran W. Fey) Newsgroups: comp.sys.next.programmer Subject: Listener/Speaker problem Date: 8 Jul 1996 21:57:46 GMT Organization: University of Washington Message-ID: <4rs08q$n7k@nntp5.u.washington.edu> I am trying to simulate asynchronous input to an application. DigIt is my app and DigItSimulator generates the simulated data. The problem I am having is that according to all the returned values the communication link is established but in fact when the simulator sends a message it is not received by the other app. This is very difficult to explain, much easier to show. Below are the set up for both the simulator and the app please send mail if you have a clue to source of my problem. DigIt (Application that receives the simulated input): simListener = [[Listener alloc] init]; if ([simListener checkInAs:"DigItSimulatorPort"] == 0) { [simListener setDelegate:self]; [simListener setPriority:NX_MODALRESPTHRESHOLD]; [simListener addPort]; if (![[Application workspace] launchApplication:"DigItSimulator"]) { NXLogError( "%s %s %s\n", [[self class] name], sel_getName( _cmd ), "unable to launch DigItSimulator" ); } } else { NXLogError( "%s %s %s\n",[[self class] name], sel_getName( _cmd ), "unable to check in port" ); } DigItSimulator (Application that sends the simulated input) id speaker = [NXApp appSpeaker]; const char* arg = "Hello from DigItSimulator"; port_t simPort = NXPortFromName( "DigItSimulatorPort", NULL ); if (simPort == PORT_NULL) { NXLogError( "%s %s %s\n", [[self class] name], sel_getName( _cmd ), "unable to find DigItSimulatorPort" ); } else { [speaker setSendPort:simPort]; if ([speaker performRemoteMethod:"receive:" with:arg length:(strlen( arg ) + 1)] != 0) { NXLogError( "%s %s %s\n", [[self class] name], sel_getName( _cmd ), "error sending message" ); } } !!!!! What I really do not understand is that anything I put in for "receive:" returns 0 (successful) but "receive:" (or any other method name) is never executed by the receiving app DigIt. Thank you for your attention (We are running NS 3.3 user/developer)
From: scratch@tiros.sce.carleton.ca (Craig Scratchley) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: Re: gcc-2.7.2 on NS3.3-Sparcstation 5? Date: 9 Jul 96 02:13:24 GMT Organization: Carleton University, Ottawa, Canada Message-ID: <scratch.836878404@tiros.sce.carleton.ca> References: <4r78o4$h63@news.mty.itesm.mx> <4rls6f$85j@duck.inetnebr.com> <slrn4tv3ap.k0f.ifeulner@xenon.stgt.sub.org> ifeulner@xenon.cube.de (Ingo Feulner) writes: >On 6 Jul 1996 14:11:27 GMT, Jeremy Bettis <jeremy@inetnebr.com> wrote: >>The prerelease does not. I suspect that the final release won't either. >You're right. Strange, someone sent me an e-mail saying that OpenStep uses 2.7.2. Can anybody confirm this, or report which version exactly OpenStep uses. I think that at least OpenStep/NT uses a fairly recent version. Craig -- W. Craig Scratchley | internet: scratch@sce.carleton.ca Dept. of Systems and Computer Engineering | phone: (613) 520-5740 (Dept.) Carleton University | (613) 241-6952 (Home) Ottawa, ON, CANADA K1S 5B6 | fax: (613) 520-5727 (Dept.)
From: Jean-Philippe Viegas <viegas@aar.alcatel-alsthom.fr> Newsgroups: comp.sys.next.software,comp.sys.next.programmer,comp.sys.next.misc Subject: OmniWeb ? Date: Tue, 09 Jul 1996 10:50:17 +0200 Organization: Alcatel Alsthom Recherche, Marcoussis, France Message-ID: <31E21D49.41C67EA6@aar.alcatel-alsthom.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, I downloaded OmniWeb, the web browser for NextStep, but I don't know how to make it work. My Nextstep works on a pentium machine and there are the main caracteristics of the network : It's an Unix network and the access to outside is provided by a proxyhost installed on a firewall machine named...proxyhost. The communications happen through the port number XXXX. I tried to put all those informations on the proxy preferences but the answer of the browser is : temporarily unable to connect : network is unreachable. In the preferences there are the cookies preferences, what is that? I can't enter a server name... Help would be greatly appreciated THANKS -- ------------------------------------------------------ Jean Philippe VIEGAS alcatel alsthom recherche viegas@aar.alcatel-alsthom.fr ------------------------------------------------------
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Re: NXImage question from novice Date: 9 Jul 1996 15:15:29 GMT Organization: Rockwell Avionics - Collins Message-ID: <4rtt2h$rq4@castor.cca.rockwell.com> References: <thad-0507961320090001@news.rice.edu> Cc: thad@ion.rice.edu See the -setFlipped: View method. Text display and image display have different objectives. In <thad-0507961320090001@news.rice.edu> Thad Harroun wrote: > I'm a novice programmer, but thanks to NeXT have written an image > analysis program that works really well for what I want, and now I'm > trying to optimize it because it is kind of large. One thing that I just > can't understand is the flip in the image as it's displayed in the window: > the origin of the window is in the lower-left corner, the data point > data[0] is in the upper-left. This gets kind of confusing when I need to > do polar coordinates from a new origin in the middle of the image. > > Why is this happening? If I NX_SOVER from the point 0,0, why then doesn't > it start with data[0] at that point? > > Thanks for an explanation, > > Thad > >
From: bbum@friday.com Newsgroups: comp.sys.next.programmer Subject: ncurses binaries for i386/m68k Date: 9 Jul 1996 15:13:13 GMT Organization: The ThoughtPort Authority, Inc. Message-ID: <4rtsu9$5b4@chinx4.thoughtport.net> Does anyone have a [relatively modern] version of ncurses binraries available for NEXTSTEP? I downloaded Carl Edman's source packages-- but they didn't build out of the box... actually, they exhibited enough problems that I didn't have the time to fix 'em... SO-- being lazy-- I'm asking the community for a package or .tar.gz of the installed product... thanks, b.bum
From: stephen@ccc1.tamu.edu (Stephen Johnson) Newsgroups: comp.sys.next.programmer Subject: undefined type, found `NSCalendarDate' Date: 9 Jul 1996 18:18:48 GMT Organization: Texas A&M University, College Station, TX Message-ID: <4ru7q8$flk@news.tamu.edu> I'm trying to build a project in ProjectBuilder on NSFIP 3.3 and I get the following error in the ProjectBuilder detail window: cc -g -Wall -pipe -I/usr/local/Minerva/include -O -I./sym -arch i386 -I./sym -c MiniSQLAdaptor.m -o ./i386_obj/MiniSQLAdaptor.o /NextDeveloper/Headers/eoaccess/EOAdaptorChannel.h:198: undefined type, found `NSCalendarDate' *** Exit 1 Stop. *** Exit 1 Stop. NSCalendarDate is defined in NSDate.h, which is located at: /usr/include/NSDate.h and linked to the following: /usr/include/foundation/NSDate.h /NextDeveloper/Headers/foundation/NSDate.h Any ideas as to how to fix this problem? Stephen Johnson, sjohnson@myriad.net Computer Consulting Int'l, LTD PO Box 1046 College Station, TX 77841 409-696-9463
From: root@polaris.scicntr.ortn.edu (Operator) Newsgroups: comp.sys.next.programmer Subject: Time/NSDate Date: 10 Jul 1996 02:51:36 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4rv5ro$q4v@stc06.ctd.ornl.gov> I am interested in using the EOF NSDate class to measure timings as precisely as possible (milliseconds). Has anyone worked with this, or know of any examples that might be helpful in using NSDate for such things ? Thanks in advance - CB
From: moser@ifor.math.ethz.ch (Dominik Moser,CLV A4,2 40 19,720 49 89) Newsgroups: comp.sys.next.programmer Subject: Re: Compile problems (pgp 2.6.3i) Date: 10 Jul 1996 06:50:42 GMT Organization: Swiss Federal Institute of Technology (ETHZ) Distribution: world Message-ID: <4rvjs2$6oh@elna.ethz.ch> References: <4rrhvj$6fr@bagan.srce.hr> Cc: thor@zems.fer.hr In <4rrhvj$6fr@bagan.srce.hr> Thor Legvold wrote: > I recently downloaded pgp 2.6.3i and found predefined > definitions for next, next-inte, next486 and nextHP > in the makefile. However, when I make I get an > error: > pgp.c: In function `ParseRecipients': > pgp.c:2922: warning: assignment makes pointer from integer without a cast This is just a compiler warning - ignore it! > and in linking I get an exit because of undefined > symbol: > ld: Undefined symbols: > _strdup Here is a simple implementation of strdup. Put it in a file called strdup.c, compile it and link it together with pgp. #include <stdio.h> /* * Most systems don't have this (yet) */ char *strdup(str) char *str; { char *p; extern char *malloc(); extern char *strcpy(); if ((p = malloc(strlen(str)+1)) == NULL) return((char *) NULL); (void) strcpy(p, str); return(p); } Regards, Dominik --------------------------------------------------------------------- Dominik Moser Sysadmin & Webmaster Institute for Operations Research Dept. of Mathematics Phone: +41 1 632 4019 ETH Zentrum, CLV A4 Fax: +41 1 632 1025 CH - 8092 Zuerich, Switzerland E-mail: moser@ifor.math.ethz.ch Member of NiCE - NeXT User Group Switzerland
From: David Grindrod <grindrod@mailhost.NMR.EMBL-Heidelberg.DE> Newsgroups: comp.sys.next.sysadmin,comp.sys.next.software,comp.sys.next.programmer Subject: Re: Mosaic 2.6 for NeXT black box Date: Wed, 10 Jul 1996 13:05:38 +0200 Organization: EMBL Distribution: world Message-ID: <31E38E82.446B@mailhost.NMR.EMBL-Heidelberg.DE> References: <4rgr6s$lej@zimmer.CSUFresno.EDU> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Kin Hung Au <kinau@zimmer.CSUFresno.EDU> You can try the version in my ftp site: ftp://ftp.NMR.EMBL-Heidelberg.DE/pub/next/X11 I think this may be 2.6 and I can not remember if this is one that worked very well. I have compiled 2.7 beta releases with wais support but it is a painful proceedure. You need to change some source code slightly. I gave up trying to port them as Cube-X supply 2.6 with there release 5.0. Ok it does not have wais support included but this is not a problem. Contact me if you want version 2.7 and I will see if I can find the old source code I had compiled. I have compiled a few X windows programs such as mxgdb and xftp (although this has a slight problem reading local directory listings correctly). Perhaps I will make these available aswell. Dave -- --------------------------------------------------------------- David grindrod, NMR System Manager, EMBL Heidelberg. Email: mailto:grindrod@EMBL-Heidelberg.DE HTML Home Page: http://www.NMR.EMBL-Heidelberg.DE/grindrod/
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,comp.sys.northstar,comp.sys.nsc.32k,comp.sys.oric,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.s From: M. WILKINS <mwilkins@earthlink.net> Sender: M. WILKINS <mwilkins@earthlink.net> Subject: cmsg cancel <01bb6c56.57337d00$3f5bface@mwilkins> Control: cancel <01bb6c56.57337d00$3f5bface@mwilkins> Message-ID: <cancel.01bb6c56.57337d00$3f5bface@mwilkins> Followup-to: junk References: <01bb6c56.57337d00$3f5bface@mwilkins> Date: 7 Jul 1996 22:53:52 GMT Spam-cancel: "I'LL TELL YOUR FUTURE, NOW!!!"
From: "Sam R. Thangiah" <thangiah@samuel.cpsc.sru.edu> Newsgroups: comp.sys.next.programmer Subject: C++ on the NeXT Date: 10 Jul 1996 23:57:20 GMT Organization: Slippery Rock University Message-ID: <4s1g10$jc1@jake.esu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Could somebody direct me on how I could compile and run C++ code on the NeXT system. Any help would be appreciated. Thanks, .. Sam thangiah@samuel.cpsc.sru.edu
From: akira@valley.xo.com (Doug McClure) Newsgroups: comp.sys.next.programmer Subject: Crash detecting objects Date: 11 Jul 1996 05:02:12 GMT Organization: InterNex Information Services 1-800-595-3333 Message-ID: <4s21sk$5ul@masters0.news.internex.net> Quite awhile back, I remember that there was some software that you could add to an application, and when that application would die, you could get backtraces saved and/or mailed to someone. Does anything like this still exist, and where can I find it? I appreciate any help. -doug
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.next.programmer Subject: Re: Crash detecting objects Date: 11 Jul 1996 06:58:06 GMT Organization: Digital Fix Development Message-ID: <4s28lu$p0b@digifix.digifix.com> References: <4s21sk$5ul@masters0.news.internex.net> In-Reply-To: <4s21sk$5ul@masters0.news.internex.net> On 07/10/96, Doug McClure wrote: >Quite awhile back, I remember that there was some software that you >could add to an application, and when that application would die, you >could get backtraces saved and/or mailed to someone. Does anything >like this still exist, and where can I find it? > >I appreciate any help. > >-doug > > There is a free version that does this, likely on next-ftp.peak.org, although for the life of me I can't tell you the name... I think there is also a current verion of it kept on Hutchison Software's FTP archive. CrashCatcher is a commercial product that does this, although I'm not sure if its still available. Chuck? -- - 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: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: 105 key keyboards (windows keyboards) Message-ID: <1996Jul11.095835.46718@yogi.urz.unibas.ch> From: frank@ifi.unibas.ch Date: 11 Jul 96 09:58:34 MET I've finally com to installing NEXTSTEP on a PC (DEC celebris glst). Everythings fine and working (actually, NEXTSTEP was the *only* system I could install out of the box. All others, win95, winNT, and linux are causing problems in that they won't install according to the book ...) Now, the keyboard shipped with the machine has four more keys than the default keyboard suggested by NeXT (I am using the US keyboard): two of these keys have a windows logo on them, one a menu and the last is a second control key. I would like to use these keys under NEXTSTEP, but the NeXT supplied keyboard layouts and definitions won't let me do this. I've figured out how to define a new keyboard layout for use with the Keyboard.app, but I can't figure out the 'scan codes' (or whatever the key code used in the layout file stands for) of these keys. I want to map both alt keys to the NEXTSTEP alt, both controls to the NEXTSTEP control and both window keys to the NEXTSTEP command keys, and perhaps the menu key to the NEXTSTP HELP key. Can anyone help? Is this a FAQ issue I missed? -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: paul@plsys.co.uk (Paul Lynch) Newsgroups: comp.sys.next.programmer Subject: Re: Crash detecting objects Date: 11 Jul 1996 10:37:08 GMT Organization: P & L Systems, Ltd. Message-ID: <4s2lgk$btk@ironhorse.plsys.co.uk> References: <4s21sk$5ul@masters0.news.internex.net> Cc: akira@valley.xo.com In <4s21sk$5ul@masters0.news.internex.net> Doug McClure wrote: > Quite awhile back, I remember that there was some software that you > could add to an application, and when that application would die, you > could get backtraces saved and/or mailed to someone. Does anything > like this still exist, and where can I find it? There is a PD thing that does this, but CrashCatcher from ITS is a very useful commercial version of it, which we use. I don't know what ITS are doing in the NeXT market these days. Paul -- Paul Lynch (NeXTmail) paul@plsys.co.uk Tel: (01494)432422 P & L Systems Fax: (01494)432478 http://www.plsys.co.uk/~paul
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: C++ on the NeXT Date: Thu, 11 Jul 1996 11:35:26 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <DuDMv3.E7z@cam-ani.co.uk> References: <4s1g10$jc1@jake.esu.edu> In article <4s1g10$jc1@jake.esu.edu> "Sam R. Thangiah" <thangiah@samuel.cpsc.sru.edu> writes: > Could somebody direct me on how I could compile and run C++ code > on the NeXT system. Any help would be appreciated. > If you've got 3.3 then just cc it (or cc++ it - it doesn't actually matter). If the filename is xxx.C or .cc or any of the other C++ extensions it should just work. You might like to install libg++ (not strcitly necessary) which is on the distribution CD. If you're running 4.0 - don't even try (you'll only get hurt). $an
From: yannick@uranus.univ-lr.fr (Yannick Buisson) Newsgroups: comp.sys.next.programmer Subject: how can i limit the max number of records after a fetch ?? Date: 11 Jul 1996 12:54:29 GMT Organization: Universite de La Rochelle Message-ID: <4s2ti5$pls@hpuniv.univ-lr.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi, I would like to implement the same system as in EOModeler. I want to ask the user when the number of rows is more important than a limit during a fetch ?? How can do that ? any help would be nice !! Bye YANNICK -- //// (. .) ----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: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: 105 key keyboards (windows keyboards) Date: 11 Jul 1996 16:59:19 GMT Organization: ALI Message-ID: <4s3bt7$beo@cetus.ali.bc.ca> References: <1996Jul11.095835.46718@yogi.urz.unibas.ch> Cc: frank@ifi.unibas.ch In comp.sys.next.sysadmin,comp.sys.next.programmer frank@ifi.unibas.ch wrote: > Now, the keyboard shipped with the machine has four more keys than the > default keyboard suggested by NeXT (I am using the US keyboard): two of > these keys have a windows logo on them, one a menu and the last is a > second control key. > > I would like to use these keys under NEXTSTEP, but the NeXT supplied > keyboard layouts and definitions won't let me do this. > ... > Can anyone help? Is this a FAQ issue I missed? The MS-Windows keyboards came out after NS 3.3 shipped and therefore the operating system has no support for the extra keys. As far as it is concerned, these do not exist. OpenStep 4.0 for Mach OS (aka NeXTSTEP 4.0) apparently does support these extra keys and allows them to be re-mapped just like other keys. -- Allan Noordvyk, Software Artisan e-mail: allan@ali.bc.ca ALI Technologies Voice: 604.279.5422 x 317 Richmond, Canada Fax: 604.279.5468 * NeXT and MIME mail welcome * "C++ is to C as Lung Cancer is to Lung."
From: everhart@alterlife.com (Dwight Everhart) Newsgroups: comp.sys.next.programmer Subject: Re: Crash detecting objects Date: 11 Jul 1996 17:11:30 GMT Organization: Internet Partners of America Distribution: world Message-ID: <4s3ck2$m4j@katbert.ipa.net> References: <4s21sk$5ul@masters0.news.internex.net> Cc: akira@valley.xo.com In <4s21sk$5ul@masters0.news.internex.net> Doug McClure wrote: > Quite awhile back, I remember that there was some software that you > could add to an application, and when that application would die, you > could get backtraces saved and/or mailed to someone. Does anything > like this still exist, and where can I find it? HKCrashTrap will log a backtrace to the console, but it doesn't email it. You can find HKCrashTrap at: ftp://next-ftp.peak.org/pub/next/sources/classes/CrashTrap.v1.0.tar.gz I've found it to be quite useful, since it allows my beta testers to copy the backtrace out of their console and email it to me. Of course, direct emailing by the application would be better, but it wouldn't be hard to add that. -- Dwight Everhart "Without failure there is no success." Bella Vista, Arkansas -- David Geffen everhart@alterlife.com
Newsgroups: comp.sys.next.programmer Subject: Re: how can i limit the max number of records after a Message-ID: <1996Jul11.102137.638@piano.opus1.com> From: Ralph Zazula Date: 11 Jul 96 10:21:36 -0700 References: <4s2ti5$pls@hpuniv.univ-lr.fr> Organization: Running Start, Inc. yannick@uranus.univ-lr.fr (Yannick Buisson) wrote: >Hi, > >I would like to implement the same system as in EOModeler. I want to ask the >user when the number of rows is more important than a limit during a fetch ?? >How can do that ? > >any help would be nice !! > Hi - Check out this URL (from the OmniGroup EOF mailing list): http://www.omnigroup.com/MailArchive/eof/1995/0289.html Ralph -- Ralph Zazula Running Start, Inc. OpenStep Development & Consulting 520/760-4890 (voice/fax) zazula@running-start.com http://www.running-start.com
From: dekorte@colt.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: OpenStep name conflicts Date: 11 Jul 1996 20:21:23 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4s3no3$euk@news.onramp.net> 1. Foundation kit's NSException.h defines - (NSString *)name 2. Object.h defines - (const char *)name This causes hundred/thousands of compiler warnings. Why did NeXT do this? -- Steve Dekorte "C++/Java, the COBOL of the 90s" NeXTstep consultant - Anaheim, CA (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: 105 key keyboards (windows keyboards) Date: 11 Jul 1996 20:21:54 GMT Organization: J. W. Goethe-Universitaet Frankfurt/Main Message-ID: <4s3np2$1dd@tallowcross.uni-frankfurt.de> References: <1996Jul11.095835.46718@yogi.urz.unibas.ch> frank@ifi.unibas.ch wrote: > Now, the keyboard shipped with the machine has four more keys than the > default keyboard suggested by NeXT (I am using the US keyboard): two of > these keys have a windows logo on them, one a menu and the last is a > second control key. > > I would like to use these keys under NEXTSTEP, but the NeXT supplied > keyboard layouts and definitions won't let me do this. > > I've figured out how to define a new keyboard layout for use with the > Keyboard.app, but I can't figure out the 'scan codes' (or whatever the > key code used in the layout file stands for) of these keys. If you knew the scancodes it wouldn't help because the 3.3 keyboard driver simply doesn't provide them. You need to buy OS 4.0 where it's included. Bye Uli -- _____________________________________________________________________ Uli Zappe Lorscher Strasse 5 D-60489 Frankfurt Germany Phone: +49 (69) 9784 0007 E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail - Mime - ASCII) PGP: public key on request Member of WiNG (Wiesbaden NEXTSTEP Group) _____________________________________________________________________
From: silver@infomatch.com (C Silver) Newsgroups: comp.sys.next.programmer Subject: Where's the cthreads library file? Date: 11 Jul 1996 23:21:02 GMT Organization: InfoMatch Internet - Vancouver BC Message-ID: <4s428u$nsg@berlin.infomatch.com> Hi there. I'm using gnu's cc version 2.2.2 on the NeXT and when I compile a C source, it finds the library for the cthreads functions just fine. But when I compile a C++ source, it can't find the cthreads library to link with my code. If anyone knows the name of the library to explicitly include, and where it is, please email me asap. I've got a school project to do. Many thanks!! -- ---------------------------------------------------------------------- Craig Silver Computer Science at Simon Fraser University silver@sfu.ca (Internet) Burnaby, B.C., Canada
From: overa@fire.sjsu.edu (overa) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.bugs Subject: HELP! f77 compiler keeps saying "permision denied"..?? Date: Thu, 11 Jul 1996 16:32:43 +0800 Organization: Information Resources and Technology Message-ID: <overa-1107961632430001@physmq1.sjsu.edu> My compiler keeps saying "permision denied" when I want to compile a program or do anything. It is Absoft f77 fo the NeXT. Any ideas? Any help greatly appreciated. me :)
From: dekorte@colt.suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4s3no3$euk@news.onramp.net> Control: cancel <4s3no3$euk@news.onramp.net> Date: 12 Jul 1996 00:58:15 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4s47v7$jhh@news.onramp.net>
From: silver@infomatch.com (C Silver) Newsgroups: comp.sys.next.programmer Subject: Re: Where's the cthreads library file? Date: 12 Jul 1996 01:35:23 GMT Organization: InfoMatch Internet - Vancouver BC Message-ID: <4s4a4r$sb5@berlin.infomatch.com> References: <4s428u$nsg@berlin.infomatch.com> P.S. Please email replies. Thanks! -- ---------------------------------------------------------------------- Craig Silver Computer Science at Simon Fraser University silver@sfu.ca (Internet) Burnaby, B.C., Canada
From: Christof Marti <mrt@twi.ch> Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.advocacy Subject: cmsg cancel <31E71449.3089@twi.ch> Control: cancel <31E71449.3089@twi.ch> Date: Fri, 12 Jul 1996 20:14:25 -0700 Organization: Unisource Business Networks Message-ID: <31E71491.1B1E@twi.ch> References: <31E71449.3089@twi.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit This message was cancelled from within Mozilla.
From: dekorte@colt.suite.com (Steve Dekorte) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Identifying Type - NeXT's blunder Date: 12 Jul 1996 17:21:47 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4s61jb$3c8@news.onramp.net> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4rtmru$16b@saturn.genoa.com> <x7hgrhtciv.fsf@tom.es.ele.tue.nl> <4s0nnm$9q@saturn.genoa.com> <x7raqjj9jo.fsf@tom.es.ele.tue.nl> <4s3l1q$lfs@solaris.cc.vt.edu> <x7buhlhnm4.fsf@tom.es.ele.tue.nl> Cc: tiggr@es.ele.tue.nl Chuck_Esterbrook@orcacomputer.com writes: > If you test using isKindOf: then I have to inherit from that > class (or a descendant class), which may not be what I want. I agree with the point that relying on data-structure info such as what class an object is is the OO equivalent of the goto. This applies to method arguments depending on class type(instead of protocols) as well. Unfortunately OpenStep if full of these "goto"s. -- Steve Dekorte "C++/Java, the COBOL of the 90s" NeXTstep consultant - Anaheim, CA (NeXTmail,MIME) http://www.batech.com/~dekorte/dekorte/index.html
From: mcgredo@crl.com (Donald R. McGregor) Newsgroups: comp.sys.next.programmer Subject: 3.2 developer on 3.3 user Date: 12 Jul 1996 12:03:09 -0700 Organization: Miskatonic University Department of Classics Message-ID: <4s67hd$ogl@crl.crl.com> Can you run 3.2 developer on a 3.3 user system without spectacular crashes? -- Don McGregor | "That was back in the 80's huh?" mcgredo@crl.com | --student intern
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Delphi/Contract--Long term Date: 12 Jul 1996 19:10:39 GMT Organization: Norden 1 Communications Message-ID: <4s67vf$chv@tofu.alt.net> Programmer/analyst/developer NEXTSTEP-Commercial experience-2yrs+ Objective C-Commercial experience Contract-Long term Area-DC Area 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: Christof Marti <mrt@twi.ch> Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.advocacy Subject: Re: Have you gotten a Jaz drive to work with NeXTSTEP? Date: Fri, 12 Jul 1996 20:12:14 -0700 Organization: Technikum Winterthur Ingenieurschule Message-ID: <31E7140E.13CB@twi.ch> References: <31E5FFD7.75B6@ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Mario Pacheco <pacheco@ix.netcom.com> Mario Pacheco wrote: > ... > Jaz disk, problems arise when I try to format and initialize a Jaz > disk natively under NeXTSTEP. > ... > Here's the script as it transpired: > > /(2)% disk /dev/rsd1h > disk name: iomega jaz 1GB H.6204/0 > disk type: removable_rw_scsi > Disk utility > The JAZ Drive works perfectly with NEXTSTEP 3.3. The problem is, that the JAZ drive does not return its correct geometry (look at "/usr/etc/scsimodes /dev/rsd1a") or NEXTSTEP does not understand, what it delivers. You just have to add the entry below to your /etc/disktab file. (I dont know if it is totaly correct, but it works since 2 month with my Nextstation Turbo Color and on several Intels) We constructed it from the infos in the NextAnswers and some hardware infos we got from a configuration program on Windows. After rebooting the Station your JAZ is recognized and also the <Disk Initialize...> menu works. If you have problems to format an already formated DOS or Mac-Disk you have to do a lowlevel format (disk Format). Here is a working entry for /etc/disktab: (You may have to correct the entry to show your serial number.) # iomega jaz-Drive # # # Drive type: iomega jaz 1GB G.4409/2 (prototype Version) # Drive type: iomega jaz 1GB G.5411/1 (1st serie) # Drive type: iomega jaz 1GB G.5512/1 (my Drive) # Drive type: iomega jaz 1GB H.6204/0 (your Drive) # (you get this string with the disk command) # jaz|jaz1GB|jaz 1GB|iomega jaz 1GB H.6204/0:\ :ty=removable_rw_scsi:nc#1021:nt#64:ns#32:ss#512:rm#5400:\ :fp#320:bp#0:ng#0:gs#0:ga#0:ao#0:\ :os=sdmach:z0#64:z1#192:hn=localhost:rw=a:\ :pa#0:sa#2090730:ba#8192:fa#1024:ca#32:da#16384:ra#10:oa=time:\ :ia:ta=4.3BSD:aa Good luck! Christof
From: Christof Marti <mrt@twi.ch> Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.advocacy Subject: cmsg cancel <31E7140E.13CB@twi.ch> Control: cancel <31E7140E.13CB@twi.ch> Date: Fri, 12 Jul 1996 20:38:57 -0700 Organization: Unisource Business Networks Message-ID: <31E71A51.7D64@twi.ch> References: <31E7140E.13CB@twi.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit This message was cancelled from within Mozilla.
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: 12 Jul 1996 20:59:24 GMT Organization: Genoa Software Systems Message-ID: <4s6ebc$32g@saturn.genoa.com> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4rtmru$16b@saturn.genoa.com> <x7hgrhtciv.fsf@tom.es.ele.tue.nl> <4s0nnm$9q@saturn.genoa.com> <x7raqjj9jo.fsf@tom.es.ele.tue.nl> <4s3l1q$lfs@solaris.cc.vt.edu> <x7buhlhnm4.fsf@tom.es.ele.tue.nl> <4s61jb$3c8@news.onramp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: dekorte@colt.suite.com In <4s61jb$3c8@news.onramp.net> Steve Dekorte wrote: > I agree with the point that relying on data-structure info such as what class > an object is is the OO equivalent of the goto. This applies to method > arguments depending on class type(instead of protocols) as well. > Unfortunately OpenStep if full of these "goto"s. Though using protocols for every argument and return tyipe definition would allow the maximum flexibility, I don't think OPENSTEP's use of classes in this case is quite so bad (and its a big improvement over typing _everything as id) And I usually push protocols pretty heavily. Here are some reasons: a. The classes within a framework are designed to be used together, and are not really intended to be reused in another context without bringing along the rest of the framework - you can't really do that without the source code anyway. So a foundation class that expects another foundation class as an argument is not a large _additional_ constraint on future flexibility. (Note, A foundation class that depended upon an AppKit class would be a bad thing IMHO, else you couldn't use a foundation class without sucking in all of AppKit when you might want a different UI (or no UI at all)) b. The argument tests in OpenStep are compile time checks only. Thus if you wish to pass a MySpecialArray to a method that expects a NSArray, you can do so. The compiler will warn you about it, and you can silence the compiler warning if you wish with a caste. If the method only sends public API messages to its argument, and you class behaves reasonably like the one its pretending to be, this works. Here's the important point. If instead of using compile time static typing, OpenStep used isKindOf: to test argument types, not only would it be even slower, then you wouldn't be able to pass objects of classes other than those originally expected (at least, not without overriding isKindOf:) Since NeXT added the static typing, which catches lots of errors at compile time, and didn't shackle us by using isKindOf: throughout, then I won't complain too bitterly about the sparing use of protocols in OPENSTEP. Though I really really wish they had used a protocol for containers instead of expecting _everything_ to be a NSDictionary. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: wafirm@ix.netcom.com (The Washington Firm, Ltd.) Newsgroups: comp.object.corba,comp.databases.object,comp.lang.c++,ch.si.choose,comp.lang.objective-c,comp.lang.smalltalk,fj.comp.oops,comp.soft-sys.nextstep,comp.sys.next.marketplace,comp.sys.next.programmer,seattle.jobs.offered,us.jobs.offered,jobs.seattle,big.jobs.offered,misc.jobs.offered Subject: SEATTLE - SR. SOFTWARE DEVELOPER - AT&T WIRELESS Date: 12 Jul 1996 23:35:44 GMT Organization: Netcom Distribution: inet Message-ID: <4s6ngg$1jm@sjx-ixn4.ix.netcom.com> Sr. Software Developer Responsibilities: Design and development of the common object model Work with other project teams to solidify the design of the common object model through the following development cycles: requirement analysis functional design technical design construction application testing Qualifications: Application of Object Oriented design techniques adn methodologies. 3+ years C++ and / or Objective C programming experience. 1+ year UNIX Operating System experience NEXTSTEP, OpenStep, and Windows NT Operating System experience a plus Knowledge of database (Sybase preferred)
From: peron@emi.u-bordeaux.fr (Laurent Peron) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: uuencode/uudecode Followup-To: comp.lang.objective-c,comp.sys.next.programmer Date: 13 Jul 1996 12:49:04 GMT Organization: Enseignement Maths Info, Universite de Bordeaux I, France Message-ID: <4s8600$br7@news.u-bordeaux.fr> References: <4rl8i8$hqk@reader1.reader.news.ozemail.net> BrettAllen wrote: : hi, : does anybody know of publicly available uuencode/uudecode source code or library? : thanx Yes, of course. GNU has released such code. Look for them in archive named shareutils or xxx-utils. Sorry, I do not remember the exact name. Maybe fileutils. They are released with the source code, of course. -- -- MicroSoft Network may not carry this message without license to do so. License to carry this message requires a fee of $999, payable within 30 days to Laurent Peron. Appearance of this message on MicroSoft Network constitutes an agreement to terms. mailto:Laurent.Peron@emi.u-bordeaux.fr Universite Bordeaux I. France Maitrise d'Informatique Unics User, AmigaOS Fanatic ( Free Advert : Look at www.amiga.de )
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer,misc.jobs.contract Subject: NEXTSTEP/Contract-DC Area Date: 13 Jul 1996 13:11:34 GMT Organization: Norden 1 Communications Message-ID: <4s87a6$rvs@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------2YRS Commercial experience. Objective C-----------------Commercial experience Sybase or Oracle-----------A plus Contract--------------------Long term Area------------------------Greater DC Area Must Be---------------------US Citizen or Greencard holder. 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: Paul_Lynch@plsys.co.uk (Paul Lynch) Newsgroups: comp.sys.next.programmer Subject: Re: 3.2 developer on 3.3 user Date: Sat, 13 Jul 1996 06:34:34 GMT Organization: P & L Systems Sender: news@seer.demon.co.uk Message-ID: <1996Jul13.063434.17073@seer.demon.co.uk> References: <4s67hd$ogl@crl.crl.com> In comp.sys.next.programmer article <4s67hd$ogl@crl.crl.com> you wrote: > Can you run 3.2 developer on a 3.3 user system without > spectacular crashes? Yes. NeXT released 3.3 User well in advance of Developer, and this was their recommended configuration for a long time. Paul -- Paul Lynch (NeXTmail) http://www.plsys.co.uk/~paul
From: gerald@kurt.in-berlin.de (Gerald Erdmann) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.bugs Subject: Re: HELP! f77 compiler keeps saying "permision denied"..?? Date: 12 Jul 1996 20:49:55 GMT Organization: Private NEXTSTEP site, Germany Message-ID: <4s6dpj$q8v@mimi.in-berlin.de> References: <overa-1107961632430001@physmq1.sjsu.edu> Cc: overa@fire.sjsu.edu In <overa-1107961632430001@physmq1.sjsu.edu> overa wrote: > My compiler keeps saying "permision denied" when I want to compile a > program or do anything. > > It is Absoft f77 fo the NeXT. Any ideas? Sounds like file permission problems? Do you have write permissions in your project folder? Gerald -- --------------------------------------------------------------------- | GERALD ERDMANN | email: gerald @ kurt.in-berlin.de (NeXTmail welcome) | voice: +49 30 397 31 400 (Germany - Berlin) | crypt: pgp2 public key available |
From: Bob Hathaway <objcur@wwa.com> Newsgroups: comp.sys.next.programmer,comp.lang.oberon,comp.object.logic,comp.lang.scheme,comp.lang.beta Subject: OBJECT CURRENTS - Free New Monthly WWW Journal - MANY NEW ISSUES Followup-To: comp.object Date: 13 Jul 1996 21:47:42 -0500 Organization: Object Currents Sender: objcur@sashimi.wwa.com Message-ID: <4s9n4e$dbd@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> Issues: Jan, Feb, Mar, April, May, June, July - Available Now Next Issue: August 1 Publisher: SIGS: C++ Report, JOOP/ROAD, Object Magazine, Smalltalk Report, X Journal, Object Expert, Java Report, Object Buyer's Guide, ... 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. ***NEW NEWS*** The comp.object FAQ will move its official anonymous FTP and Web site to SIGSnet soon and our Code Watch Department will reflect additions or updates to the OOFAQ's Appendices in addition to any other news. The Berne site in Switzerland (the only official and up-to-date Web site) will become a mirror in Europe. I'd also like to mention a few talks I'll be giving in the next month or so, one at the TOOLS USA conference in Santa Barbara this July 29 on the Unified Modeling Language. I'll discuss the UML and give examples, alternatives, and general discussion in a half day tutorial. I'm also speaking at Object Expo in New York City this August 8 on "Advanced C++: Multiple-polymorphism/Uses/ Types", which discusses a more sophisticated object model with true types and how it relates to a more static model in terms of dependencies, metrics, and functionality from OOA thru to C++ and the available idioms and tradeoffs involved. TOOLS USA/Santa Barbara - July 29 - http://www.tools.com/tools/usa_96/td_adt.html Object Expo/New York - August 8 - http://www.sigs.com/conferences/oeny/oeny96day3.html#TH23 See you there! OCJ ARTICLES We are accepting original Feature Articles which include honorarium 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 June: Steve Mellor, Plus Jacobson (Part II) Soon: 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 Objects And Binary Relations A.Girow Practical Prototyping, Part Two Ian Mitchell, Norman Parrington, Peter Dunne, and John Moses Jul 96 O3CASE: Operational Object Oriented CASE Rakesh Agarwal, Marco Torchiano and Giorgio Bruno Practical Prototyping, Part Three Ian Mitchell, Norman Parrington, Peter Dunne, and John Moses Soon: 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 SIGS January, February, March, April, May, June, and July 1996 issues including: C++ Report, JOOP/ROAD, Object Magazine, Smalltalk Report, Object Expert, Object Buyer's Guide Thanks to our readership for patronage, praise, and feedback. Please keep visiting or give us 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
Newsgroups: comp.sys.next.programmer From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: 3.2 developer on 3.3 user Sender: news@novice.uwaterloo.ca (Mr. News) Message-ID: <DuHw9J.Aqt@novice.uwaterloo.ca> Date: Sat, 13 Jul 1996 18:48:55 GMT References: <4s67hd$ogl@crl.crl.com> Organization: University of Waterloo In article <4s67hd$ogl@crl.crl.com>, Donald R. McGregor <mcgredo@crl.com> wrote: >Can you run 3.2 developer on a 3.3 user system without >spectacular crashes? > Yes, you can. In fact, this is what you were supposed to do for a while, as 3.3 User was available before 3.3 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: Subrata_Sircar@next.com (Subrata Sircar) Newsgroups: comp.sys.next.programmer Subject: Re: Crash detecting objects Date: 13 Jul 1996 05:40:13 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4s7crt$k2o@news.next.com> References: <4s28lu$p0b@digifix.digifix.com> CC: sanguish@digifix.com, akira@valley.xo.com (Doug McClure) Scott Anguish writes: > On 07/10/96, Doug McClure wrote: > >Quite awhile back, I remember that there was some software that you > >could add to an application, and when that application would die, you > >could get backtraces saved and/or mailed to someone. Does anything > >like this still exist, and where can I find it? > There is a free version that does this, likely on next-ftp.peak.org, > although for the life of me I can't tell you the name... I think there is > also a current verion of it kept on Hutchison Software's FTP archive. There is a category on Object (written by Julie Zelenski with help from numerous other people) which uses a signal handler to attempt to dump backtraces to the console. (I've been attempting to port it to Foundation recently, with some success.) You can get it from the NeXTWorld Expo developer notes, if you have those (I think it was 1993). (I can mail you an old copy if you can't find it on the above ftp site.) Caveats: 1. It uses some stack and frame layout information that is not true on hppa and sparc. [Technical note: because these architectures don't push stuff onto the stack, choosing instead to keep things in registers which get clobbered by each successive call, the approach this code takes will not work on hppa/sparc. You'd have to try for the symbol-table-hackery approach, which is beyond my current knowledge.] 2. The compiler has a number of built-in functions which make this easier (__builtin_frame_address comes to mind) if you're rewriting the code. 3. It doesn't do functions very well. Again, as far as I know, symbol-table-hackery is needed to do better. 4. Windows does things with structure returns which invalidate the assumptions made in this code. Windows also has different signals than Mach, so it won't just compile. All that being said, just running it through the conversion scripts and a little sanity checking will give you something that works a fair bit of the time on m68k and i386, including Windows. If you need something for OpenStep-based apps and want to clean up the code, I recommend checking out NSMethodSignature (which will build a method signature from a selector). -- 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: jbf@frazer.com (James B. Frazer) Newsgroups: comp.sys.next.programmer Subject: Syscall arguments Date: 14 Jul 1996 15:28:36 GMT Organization: frazer.com Message-ID: <jbf-1407961128480001@news.tiac.net> I've been looking at proxynet with the thought of porting it to NS Motorola. There is extensive use of syscall, which takes arguments not identified in bsd/syscall.h. The only other mention of syscall seems to be in bsd/m68k/unix-traps.h which imports mach/m68k/syscall_sw.h, which isn't in my 3.3. Any suggestions? I've assumed the author had good reasons for using a direct call rather than accept et al, and I don't really want to tackle reverse engineering his package. Barney
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Crash detecting objects Date: Sun, 14 Jul 1996 15:39:35 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <IluInri00iWS82JsgS@andrew.cmu.edu> References: <4s21sk$5ul@masters0.news.internex.net> <4s28lu$p0b@digifix.digifix.com> In-Reply-To: <4s28lu$p0b@digifix.digifix.com> Excerpts from netnews.comp.sys.next.programmer: 11-Jul-96 Re: Crash detecting objects by Scott Anguish@digifix.co >> Quite awhile back, I remember that there was some software that you >> could add to an application, and when that application would die, you >> could get backtraces saved and/or mailed to someone. Does anything >> like this still exist, and where can I find it? > > There is a free version that does this, likely on next-ftp.peak.org, > although for the life of me I can't tell you the name... I think there is > also a current verion of it kept on Hutchison Software's FTP archive. > > CrashCatcher is a commercial product that does this, although I'm not > sure if its still available. > > Chuck? CrashCatcher is no longer available from ITS, since ITS is no longer in business as a NEXTSTEP software company. (ITS may or may not be going out of business entirely; I don't really know since I am not aware of everything that goes on in Chicago-- I live and work in Pittsburgh.) Personally, I would like the source for CrashCatcher to be made freely available to the NeXT community for everyone to use. Because of licensing agreements between WhiteLight and ITS, I am not sure whether this will happen, but we'll see. If any current CrashCatcher customers have problems or questions, feel free to email me and I will continue to offer what level of support I can on my own time. -Chuck PS: Email to 'chuck@its.com' should continue to work, at least for a while. Alternative email addresses include 'chuck@tertius.res.cmu.edu' (NeXTmail, MIME) and 'cs4w+@andrew.cmu.edu' (plain text only, please). Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: dave@turbocat.snafu.de (David Wetzel) Newsgroups: comp.sys.next.programmer Subject: Re: Syscall arguments Date: 14 Jul 1996 18:47:25 GMT Organization: Turbocat's Development, Germany Message-ID: <4sbfbt$225@turbocat.snafu.de> References: <jbf-1407961128480001@news.tiac.net> jbf@frazer.com (James B. Frazer) wrote: > I've been looking at proxynet with the thought of (...) What is proxynet? _ _ _(_)(_)_ 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,MIME)
From: Charles Bennett <chuck@benatong.com> Newsgroups: comp.sys.next.programmer Subject: Re: Serial port trouble (O 8 1) Date: Sun, 14 Jul 1996 16:48:01 -0400 Organization: BenaTong Message-ID: <31E95D01.5C77@benatong.com> References: <4s3cdh$5hi@news.ppp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lars Immisch wrote: > > Hi, > > I want to control a rather stupid device from my NeXT (black) that is > connected via a serial cable. This device expects odd parity, 8 data bits and > 1 stop bit and won't eat anything else. [munched] > Lars Immisch > lars@ibp.de > > (Please respond with email also) You can't get there from here. The tty device drivers don't support it. I toyed around with the idea of using a small microcontroller (8051's and 68HC11's will do that mode) for a project of mine that required 8 O 1 (sony video machine) but never got around to it. Perhaps you could take the MUX-DRIVER code and change it (If the UART you have supports that mode) Chuck -- With coarse rice to eat, with only water to drink, and my bended arm for a pillow, I am happy.
From: Bishara KHOURI <khouri@ifi.uni-klu.ac.at> Newsgroups: ch.si.choose,comp.object.corba,comp.databases.object,comp.lang.c++,ch.si.choose,comp.lang.objective-c,comp.lang.smalltalk,fj.comp.oops,comp.soft-sys.nextstep,comp.sys.next.marketplace,comp.sys.next.programmer,seattle.jobs.offered,us.jobs.offered,jobs.seattle,big.jobs.offered,misc.jobs.offered Subject: Informatik'96: Gemeinsame Jahrestagung der GI und OCG Date: Sun, 14 Jul 1996 20:23:44 +0200 Organization: Institut für Informatik Distribution: inet Message-ID: <31E93B30.3ED0@ifi.uni-klu.ac.at> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="xxx" I *************************************************** n * * f * Informatik'96 - Die 'neue' GI/OCG-Jahrestagung * o * * r * >> Beherrschung von Informationssystemen << * m * * a * Klagenfurt, 25. - 27. September 1996 * t * * i *************************************************** k * * ' * Workshops, Tutorien und Studiendendenprogramm: * 9 * Montag, Dienstag 23.- 24. September 1996 * 6 * * *************************************************** i n Es ist soweit: Das Programm der Informatik'96 steht und birgt hoechst Interessantes, fuer Wissenschaftler K und Wissenschaftlerinnen, fuer in der Praxis Taetige l und fuer Studierende. a g Informieren sie sich bitte ueber unsere home-page e n http://www.ifi.uni-klu.ac.at/Conferences/informatik'96/ f u Dort koennen Sie sich auch r ** ueber's Netz zur Tagung anmelden, t ** verguenstigte Unterkuenfte buchen, ** fuer preisguenstige Bustransfers nach und - - von Klagenfurt eintragen, ** ueber verguenstigte Fluege von Berlin, Hamburg und D Frankfurt informieren und ueber vieles andere mehr. a A propos: Wenn der Zugriff aus Deutschland auf unsere m Seiten manchmal etwas dauert, so liegt das nicht an u unserem Server sondern an den Leitungen von s Deutschland nach Oesterreich. Wir werden unsere Seiten s daher in Kuerze auch auf einen Server in Deutschland spiegeln. Das gedruckte Programm wird am 16.7. versandt. i c Wir freuen uns darauf, Sie in Klagenfurt begruessen zu h duerfen. h Heinrich C. Mayr und das Organisationsteam. i n Weitere Auskuenfte erhalten sie bei unserem Tagungssekretariat: ! ! Christine Seger ! Inst. fuer Informatik, Universitaet Klagenfurt ! Tel.: ++43-463-2700-578, Fax: ++43-463-2700-505 ! email: informatik'96@ifi.uni-klu.ac.at ************************************************************** ------- End of Forwarded Message
From: "The Washington Firm, Ltd." <wafirm@ix.netcom.com> Newsgroups: comp.object.corba,comp.databases.object,comp.lang.c++,comp.lang.objective-c,comp.lang.smalltalk,fj.comp.oops,comp.soft-sys.nextstep,comp.sys.next.marketplace,comp.sys.next.programmer Subject: SEATTLE - SR SOFTWARE DEVELOPER - AT&T WIRELESS SERVICES Date: Sun, 14 Jul 1996 18:56:45 -0700 Organization: Washington Firm Distribution: inet Message-ID: <01bb71f1.51be3ea0$97e41fcc@karen95.ix.netcom.com> RESPONSIBILITIES: Design and development of the common object model. Work with other project teams to solidify the design of the common object model through the following development cycles: Requirements Analysis Functional Design Technical Design Construction Application Testing QUALIFICATIONS: Application of Object Oriented design techniques and methodologies. 3+ years C++ and / or Objective C programming experience 1+ year UNIX Operating System experience NeXTStep, OpenStep, and Windows NT Operating System experience a plus Knowledge of database (sybase preferred) Please send response via e-mail in PLAIN TEXT to wafirm@ix.netcom.com, or via fax to 206/284.8844. Use reference #PN071296.
From: sanguish@digifix.com Newsgroups: comp.sys.next.announce,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.soft-sys.nextstep Subject: NEXTSTEP Resources on the Internet Date: 15 Jul 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4scgke$s91@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: droege@informatik.uni-koblenz.de (Detlev Droege) Newsgroups: comp.sys.next.programmer Subject: Re: 3.2 developer on 3.3 user Date: 15 Jul 1996 12:23:09 GMT Organization: University Koblenz / Germany Distribution: world Message-ID: <4sdd7d$3b1@newshost.uni-koblenz.de> References: <4s67hd$ogl@crl.crl.com> In article <4s67hd$ogl@crl.crl.com> mcgredo@crl.com (Donald R. McGregor) writes: > Can you run 3.2 developer on a 3.3 user system without > spectacular crashes? Yes. (In fact, 3.3 Dev came out later than 3.3 User, so NeXT even suggested to use 3.2 Dev with 3.3 User [until 3.3Dev came out]). Detlev -- Detlev Droege, Uni Koblenz, FB Informatik, Rheinau 1, D-56075 Koblenz, Germany Tel:+49 261 9119-421,Fax:-497,NeXT/MIME/Email:droege@informatik.uni-koblenz.de C++ is the only current language making COBOL look good. --Bertrand Meyer
From: lars@ibp.de (Lars Immisch) Newsgroups: comp.sys.next.programmer Subject: Re: Serial port trouble (O 8 1) Date: 15 Jul 1996 10:13:43 GMT Organization: Immisch, Beckert & Partner Message-ID: <4sd5kn$e7t@news.ppp.net> References: <4s3cdh$5hi@news.ppp.net> <31E95D01.5C77@benatong.com> Charles Bennett <chuck@benatong.com> wrote: >Lars Immisch wrote: >> >> Hi, >> >> I want to control a rather stupid device from my NeXT (black) that is >> connected via a serial cable. This device expects odd parity, 8 data bits and >> 1 stop bit and won't eat anything else. >[munched] Jim Redman from Ergotech (jim@ergotech.com) came up with a solution... Thanks a lot, Jim! He also remarked that this will work under 3.3 and 3.2, but not 3.1: #include <termios.h> struct termios ttymode; if (ioctl(serial, TIOCGETA, &ttymode) < 0) { perror("ioctl(TIOCGETA)"); exit(1); } ttymode.c_ispeed = B9600; ttymode.c_ospeed = B9600; ttymode.c_iflag = IGNPAR; ttymode.c_lflag = 0; ttymode.c_oflag = 0; ttymode.c_cflag = B9600 | CREAD | CS8 | CSTOPB | PARODD | PARENB; ttymode.c_cc[VMIN] = 0; ttymode.c_cc[VTIME] = 0; if (ioctl (serial, TIOCSETA, &ttymode) < 0) { perror("ioctl(TIOCSETA)"); exit(2); }
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: OpenStep design complaints Date: 15 Jul 1996 23:49:32 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4selec$fav@news.onramp.net> 1. Runtime Functions Ok so NeXT changed all the runtime functions. Now, in order to get a selector, you have to create a NSString object. NSSelectorFromString([NSString stringWithCString:astring]) NSSelectorFromString(@"addToString:") So you're taking a performance hit with creating the String, and if they're going to go through the trouble of creating a string object to make things "OO" why not have it implement a -selectorValue method? Why do: NSSelectorFromString([NSString stringWithCString:astring]) instead of: [[NSString stringWithCString:astring] selectorValue] Or why not save the allocation of unneeded NSString objects by having: [NSString selectorValueForCString:astring] Or am I missing something? 2. Inconsistent Naming Then there's the curious choice of method names that makes their naming scheme inconsistent such as -stringValue returning an object while every other -xxxxValue method returns an actual C data-type. 3. Behaviorism violations There are also type casts all over so you can't create your own classes with the same behavior to interoperate with theirs. This is anti-OO. Am I the only one finding OpenStep conversion a pain because of these things? Steve -- Remaining wild Tigers:<6K, BlackRhino:<2K GiantPanda:<1K [envirolink.org] Human population in 1900: 1.5 billion, in 1996: 5.7 billion. [census.gov]
Newsgroups: comp.sys.next.programmer From: noblenet@world.std.com (NobleNet Inc.) Subject: Re: Windows 3.1 clients and NS servers? How? Message-ID: <DuMGHt.G6G@world.std.com> Organization: NobleNet, Inc. References: <4rqfp2$n0d@ddfservb.technet.net> Date: Tue, 16 Jul 1996 05:56:16 GMT In article <4rqfp2$n0d@ddfservb.technet.net>, Constantin Szallies <szallies@energotec.de> wrote: > >Of cource it would be possible to just use sockets, but --> too much work. >Some RPC mechanism would be much easier to use. For example ONC RPC, which >was once called Sun RPC. The man page of rpcgen does not mention the word >ONC, so I guess the RPC suppied with NS is some version of Sun RPC. It it >compatible to ONC RPC? And then I need some ONC RPC implementaion for >Win3.1... > Please contact NobleNet for ONC/RPC tools and solutions for UNIX, Windows (3.1/95/NT), Macintosh, VMS, OS/2, NetWare, VxWorks, etc. NobleNet, Inc. 337 Turnpike Road Southboro, MA 01772 Voice +1 508 460 8222 Fax +1 508 460 3456 <sales@noblenet.com> <http://www.noblenet.com/>
From: rft@cg.tuwien.ac.at Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.advocacy Subject: Re: Have you gotten a Jaz drive to work with NeXTSTEP? Date: 16 Jul 1996 07:55:42 GMT Organization: Vienna University of Technology, Austria Message-ID: <4sfhtu$gfd@news.tuwien.ac.at> References: <31E5FFD7.75B6@ix.netcom.com> I think the best way to fix that problem is to do the following. At least it did work on my system (I do not guarantee that this will work on any other system). If someone in the know could confirm the procedure I'd be grateful: Login as root and type the following command: disk -q /dev/rsd1a (or rsdXa with X beeing the right number for your jaz drive) Then use the resulting name to make an entry to /etc/disktab, where the first entry in the first line is the string you got as a result of the previous command. You need to do this, since different revisions of the jaz drive will respond with a different string and the disk command uses just this string as a pointer to the correct disktab entry. My disktab entry for the jaz drive looks like this (the other strings on the first line are precaution, in case disk uses a truncated version of the string to find the correct entry). iomega jaz 1GB H.6204/0|iomega jaz 1GB H.62|iomega jaz 1GB|iomega jaz:\ :ty=removable_rw_scsi:nc#3624:nt#4:ns#144:ss#512:rm#5394:\ :fp#320:bp#0:ng#0:gs#0:ga#0:ao#0:\ :os=sdmach:z0#64:z1#192:ro=a:\ :pa#0:sa#2087104:ba#8192:fa#1024:ca#32:da#4096:ra#10:oa=time:\ :ia:ta=4.3BSD:aa: or alternatively the entry from Nextanswers: iomega jaz 1GB H.6204/0|iomega jaz 1GB H.62|iomega jaz 1GB|iomega jaz:\ :ty=removable_rw_scsi:nc#3584:nt#4:ns#72:ss#1024:rm#5400:\ :fp#160:bp#0:\ :os=sdmach:z0#32:z1#96:ro=a\ :pa#0:sa#1032192:ba#8192:fa#1024:ca#16:da#4096:ra#10:oa=time:\ :ia:ta=4.3BSD:aa: [comment: my disktab is changed from the one NeXT supplies in NextAnswers, by using more of the available space. I took whatever information was given by scsimodes on the jaz drive to build it. (The reply of the jaz drive was a bit broken, and I guessed and calculated to arrive at the above numbers). Again I'd be grateful for some confirmation by people who have more information on the layout of the jaz drive] If you have added this entry to the disktab, you can use the disk command to low-level format your drive: disk -F /dev/rsdXa and initialize your drive: disk -h yourhost -l yourdisk -i /dev/rsd1a or alternatively (after a disk -F /dev/rsdXa command) use BuildDisk on the resulting drive. (You may have to chekc on the correct installation of the Drivers before booting off the newly installed jaz cartridge). Hope that helps, ------------------------------------------------------------------------ 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: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: outer join Date: 16 Jul 1996 11:10:04 GMT Organization: Universite de La Rochelle Message-ID: <4sftac$haf@hpuniv.univ-lr.fr> I have found the source of the error but has someone a solution ? I have an entity witch contains 2 relationship : the first is an innerJoin the second an LeftOuterJoin so the thrue Query is : SELECT t0.*, t1.field1, t2.field2 FROM table0 t0, table1 t1, OUTER(table2 t2) WHERE (t0.aField = t1.theKey AND t0.aField = t2.theKey) But EOMODELER generate the folowing query : SELECT t0.*, t1.field1, t2.field2 FROM OUTER table0 t0, table1 t1, table2 t2 WHERE (t0.aField = t1.theKey AND t0.aField = t2.theKey) Wilfrid. ====================== wgaboria@iut-lr.univ-lr.fr Service Informatique IUT La Rochelle France
From: edwards@sewp.nasa.gov Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: Tue, 16 Jul 1996 13:55:28 GMT Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <960716095528.4434AADnG.edwards@vege> References: <4selec$fav@news.onramp.net> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII > From: dekorte@suite.com (Steve Dekorte) > Subject: OpenStep design complaints > Date: 15 Jul 1996 23:49:32 GMT > Newsgroups: comp.sys.next.programmer > > > 1. Runtime Functions > > Ok so NeXT changed all the runtime functions. > Now, in order to get a selector, you have to create a NSString object. > > NSSelectorFromString([NSString stringWithCString:astring]) > NSSelectorFromString(@"addToString:") > > So you're taking a performance hit with creating the String, and > if they're going to go through the trouble of creating a string object to make > things "OO" why not have it implement a -selectorValue method? > > Why do: > > NSSelectorFromString([NSString stringWithCString:astring]) > > instead of: > > [[NSString stringWithCString:astring] selectorValue] > > Or why not save the allocation of unneeded NSString objects by having: > > [NSString selectorValueForCString:astring] I don't understand how this would save you anything. Where did the NSString that you're messaging come from in the first place? I think the decision to objectify strings was good and long overdue. Since the implementation of NSStrings is hidden inside the class cluster, and the compiler was extended to recognize NSString literals, it's not clear that the performance hit in your scenario is significant in most cases. If performance is an issue, or you'd just rather use C strings in this case, I believe sel_getUid() is still there. > > Or am I missing something? > > 2. Inconsistent Naming > > Then there's the curious choice of method names that makes their > naming scheme inconsistent such as -stringValue returning an object > while every other -xxxxValue method returns an actual C data-type. This alleged inconsistency is debatable. Strings aren't data types in C, they're idiomatic uses of character arrays. Since NeXT made the strategic decision to replace all uses of C strings with NSString, it would seem more inconsistent to sporadically lapse back into using C strings. If it really bothers you, why not use a category to add a method like "CStringValue"? > > 3. Behaviorism violations > > There are also type casts all over so you can't create your own classes with > the same behavior to interoperate with theirs. This is anti-OO. Could you provide some specific examples? > > Am I the only one finding OpenStep conversion a pain because of these things? > > Steve > -- > Remaining wild Tigers:<6K, BlackRhino:<2K GiantPanda:<1K [envirolink.org] > Human population in 1900: 1.5 billion, in 1996: 5.7 billion. [census.gov] --- Stephen G. Edwards SEWP BOWL, NASA Goddard Space Flight Center Tel: 301-286-0783 Fax: 301-286-1619 edwards@sewp.nasa.gov
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: Tue, 16 Jul 1996 13:38:57 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <Dun1wy.FLB@cam-ani.co.uk> References: <4selec$fav@news.onramp.net> In article <4selec$fav@news.onramp.net> dekorte@suite.com (Steve Dekorte) writes: > > 1. Runtime Functions > NSSelectorFromString([NSString stringWithCString:astring]) > NSSelectorFromString(@"addToString:") I can live with this. Cas two is pretty trivial to enter. Case 1 is less likly to occur, as you must have got aString from somewhere - most likly an object which could return an NSString. > 2. Inconsistent Naming > > -stringValue returning an object > while every other -xxxxValue method returns an actual C data-type. Again I think this OK. > 3. Behaviorism violations > > There are also type casts all over Yep. This is a bitch. I guess it's to try and keep the C++ "strong typing is good/safe" people happy (but then again there's enough other stuff to piss them off). I haven't had many real problems as a result of it, but concpetually its a real bitch, and throws out so much good stuff that should be possible/supported. It doesn't matter what class an object is. It shoudl eb defined by it's interface. This change puts us back into the stupid C++ create a dummy superclass so that you can put things of class "dummy" in the list, but without multiple inheritance to make it possible (No I do NOT want multiple inheritance!). $an
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: Tue, 16 Jul 1996 14:45:32 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <Dun4zx.Fv0@cam-ani.co.uk> References: <Dun1wy.FLB@cam-ani.co.uk> In article <Dun1wy.FLB@cam-ani.co.uk> ians@cam-ani.co.uk (Ian Stephenson) writes: > > 3. Behaviorism violations > > > > There are also type casts all over > I haven't had many real problems as a result of it I've just remembered where/when this caused me problems, and its a real bitch: A class AGeneric has a -setB:(BGeneric*)b method. A sublcass of AGeneric, ASpecific will fail to compile if it has a method -setB:(BSpecific*)b, because the types do not match. The workaround is to simply have -setB:(BGeneric*)b in ASpecific even though it requires an object of type BSpecific. I'd prefer to define everything of type id (ie specify no class), than to have to do something like this (ie specify an object of the WRONG class). Note this produces an ERROR, not a WARNING. $an
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 16:44:00 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4sggsg$3v@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> ians@cam-ani.co.uk (Ian Stephenson) wrote: > I've just remembered where/when this caused me problems, and its a real > bitch: > > A class AGeneric has a -setB:(BGeneric*)b method. > > A sublcass of AGeneric, ASpecific will fail to compile if it has a > method -setB:(BSpecific*)b, because the types do not match. The > workaround is to simply have -setB:(BGeneric*)b in ASpecific even though > it requires an object of type BSpecific. > > I'd prefer to define everything of type id (ie specify no class), than > to have to do something like this (ie specify an object of the WRONG > class). > > Note this produces an ERROR, not a WARNING. It seems to me that the solution is to make the selector contain type information as well as the selector's name. For example, these two methods: -setB:(BSpecific*)b -setB:(BGeneric*)b right now resolve to the same selector: "setB:". Because of this, the compiler can't do proper type checking because it doesn't necessarily know the type of the receiver, and that prevents it from knowing which version of the method will be used. If you make the selector include type information, using something such as: "setB:(@BSpecific)" Then you could make the compiler support this quite nicely. Maybe the part in parenthesis should be the same as what @encode would give you... Note that this is how Java handles selectors. I like it because it allows you to create something like the above, where both messages co-exist in the same class, and which method gets called depends upon the argument type. The compiler and runtime figure out which message to send based upon the arguments as well as the method name, in effect. The downside that I see is that you end up with a method name that doesn't specifically have the type in the name, such as setIntValue: and setFloatValue:, which are nicely self-documenting. The upside is that you don't have the type in the name, with setValue: as the only method, with implementations for each supported type...which simplifies the interface quite a bit. Of course, since you could put the type in the name if you wanted to (though that is redundant) I see this more as an advantage than a disadvantage. Also note that Java will attempt to convert the variable in question if the method of the proper type isn't thee. So if you only implement a -setValue:(double)x and pass it a float, it will still work. But if you implement a -setValue:(float)x as well, then it will be used instead. I really would like to see this in Objective-C, since it would solve the problem above and it would add some syntactical sugar to the language that I would find convenient. :-) I don't see any real detriment to doing this, other than the fact that it would require some changes to the compiler and runtime. But it would be a simple enough change that it shouldn't affect the language too much... (I'm sure I'll get flooded with counter examples, so go right ahead. Post away!) At any rate, it does seem like a good idea on the surface. You don't really lose any of Obj-C's dynamism, and you get better type checking as a bonus. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 16:41:47 GMT Organization: ALI Message-ID: <4sggob$iof@cetus.ali.bc.ca> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> Cc: ians@cam-ani.co.uk In comp.sys.next.programmer Ian Stephenson wrote: > In article <Dun1wy.FLB@cam-ani.co.uk> ians@cam-ani.co.uk (Ian > Stephenson) writes: > > > There are also type casts all over > I've just remembered where/when this caused me problems, and its a real > bitch: > > A class AGeneric has a -setB:(BGeneric*)b method. > > A sublcass of AGeneric, ASpecific will fail to compile if it has a > method -setB:(BSpecific*)b, because the types do not match. The > workaround is to simply have -setB:(BGeneric*)b in ASpecific even though > it requires an object of type BSpecific. > > I'd prefer to define everything of type id (ie specify no class), than > to have to do something like this (ie specify an object of the WRONG > class). > > Note this produces an ERROR, not a WARNING. In this case I think it is better to define the object by interface (ie. protocol) not type (something I would have liked to seen more of in OpenStep but that is another issue). Rewriting your example we get: @protocol bobo ... @end @interface AGeneric:NSObject ... - (void)setB:(id <bobo>)aGenericB; @end @interface ASpecific:AGeneric ... - (void)setB:(id <bobo>)aSpecificB; @end @interface BGeneric:NSObject <bobo> @end @interface BSpecific:BGeneric @end The problem arises when the difference between genericB and specificB is the methods which are to be called in the setB: methods of the AGeneric and ASpecific classes. In this case we have the following basic options: 1. Make the bobo protocol contain all of the methods which are to be called and make genericB implment the extra methods as no-ops. (ie. QUICK AND DIRTY APPROACH) 2. As above, but have genericA's versions of the methods generate exceptions. (ie. ANAL RETENTIVE APPROACH) 3. Don't add specificB's extra methods to the bobo protocol, but put them in an extra specificBobo protocol. Then ask the argument object if it conforms to this protocol. If it does not, then abort (ie. RUN TIME BAD ARGUMENT DETECTION APPROACH). 4. As in 3 above but if it does conform to the extended protocol, then do the extra stuff, otherwise, treat it as a genericB. (ie. DYNAMIC "JUST WORKS" APPROACH). In both 3 and 4 you can silence compiler warnings with a cast assignment to a local variable of the appropriate type. Which approach you use really depends on the context of the object. For a project of any non-trivial size, I prefer the new typed argument style of OpenStep. I'd rather find my bugs sooner rather than later. There are some cases where I feel the types should have been protocols rather than classes, but that's not a big issue with me. -- 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 * "C++ is to C as Lung Cancer is to Lung."
From: adwait@uh.edu (Adwait N Nerlikar) Newsgroups: comp.sys.next.programmer Subject: Serial port communication on NeXT Date: 16 Jul 1996 17:39:05 GMT Organization: University of Houston Message-ID: <4sgk3p$kb3@masala.cc.uh.edu> I am trying to read in data over the serial port on NeXT. On the mk68 architecture machines I can read the data successfully. But when I try running the same code on an intel machine, it doesnt work! The incoming data is at 9600 baud, NO parity, 8 bits, and 1 stop bit. Can someone tell what I am doing wrong? Thanks in advance, Adwait Nerlikar Please reply to : adwait@uh.edu ----------C O D E----------- case 1 > int ttyPtr; ttyPtr = open("/dev/ttya", O_RDONLY); fcntl(ttyPtr, F_SETFL, FNDELAY); options.c_iflag = IGNBRK; options.c_oflag = 0; options.c_cflag = B9600 | CS8 | CREAD | CLOCAL; options.c_lflag = 0; options.c_iflag &= ~(IXON | IXOFF | IXANY); ioctl(ttyPtr, TCSANOW, &options); case 2 > char *command; int ttyPtr; ttyPtr = open("/dev/ttya", O_RDONLY); sprintf(command,"stty %d raw > %d", baudRate, ttyPtr); system (command);
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 17:52:41 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sgkt9$imp@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> Cc: don@misckit.com don@misckit.com wrote: > ians@cam-ani.co.uk (Ian Stephenson) wrote: > > > > A class AGeneric has a -setB:(BGeneric*)b method. > > > > A sublcass of AGeneric, ASpecific will fail to compile if it has a > > method -setB:(BSpecific*)b, because the types do not match... ... > It seems to me that the solution is to make the selector contain type > information as well as the selector's name. A bad idea, IMO. Makes all the selector related runtime functions a bitch to deal with. If NeXT had just stuck to the convention of, in the case of non-objects, putting the type followed by the word "Value" in the method name itself as they mostly did before (ex. intValue returns and int, stringValue returns a STR, etc) and leave everything else as an id, you would never have to worry about name conflicts. Name confilict are a *very* serious problem when you want to mix objects from different sources. Some manager at NeXT that doesn't understand OO (in this case, that type dependence = data structure dependence = the antithesis of OO) probably made the decision to do this. Perphaps an Taligent refugee... Steve
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 17:57:42 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sgl6m$imp@news.onramp.net> References: <4selec$fav@news.onramp.net> <960716095528.4434AADnG.edwards@vege> Cc: edwards@sewp.nasa.gov edwards@sewp.nasa.gov wrote: > dekorte@suite.com: > > 3. Behaviorism violations > > > > There are also type casts all over so you can't create your own classes with > > the same behavior to interoperate with theirs. This is anti-OO. > > Could you provide some specific examples? - (void)setStringValue:(NSString *)aString; Guess what happens when I try to pass in another object with similiar behaivior? It won't compile. Steve
From: shess@parka.winternet.com (Scott Hess) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 13:39:58 -0500 Organization: Is a sign of weakness Sender: shess@parka.winternet.com Message-ID: <acwx04ca9t.fsf@parka.winternet.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> In-reply-to: don@misckit.com's message of 16 Jul 1996 16:44:00 GMT In article <4sggsg$3v@news.xmission.com>, don@misckit.com writes: >Note that this is how Java handles selectors. I like it because >it allows you to create something like the above, where both >messages co-exist in the same class, and which method gets called >depends upon the argument type. The compiler and runtime figure >out which message to send based upon the arguments as well as the >method name, in effect. I'll weigh in with a bit veto on that one. This is one of my big complaints about C++ in general, right up there with "Why do method calls look like structure dereferences?" [Yeah, yeah, because in some sense they _are_ structure dereferences. My feeling, though, is that it is advantageous to disambiguate the OOP pieces of code from the C pieces of code, and Objective-C's message syntax does that pretty well. It calls out the message dispatches nicely.] My view on this is that Objective-C's method naming conventions tend to promote (though don't enforce) a higher level of code self-documentation. Compare: [obj init]; [obj initSize:aSize]; [obj initSize:aSize color:aColor]; ... with: obj.init(); obj.init( aSize); obj.init( aSize, aColor); Now, admittedly, it should be relatively easy to see what's up here, since I used descriptive variable names. But how can you tell if aSize and aColor are in the right order? Or change aSize to ss and aColor to cc. Add four more parameters. Now can you tell which is which without looking at the header files? [That was my bow to "A good programmer can write readable code in assembly." My point is that there are only about four good programmers out there, by that definition. For the rest of us, every little bit helps.] obj.initWithSizeAndColor( aSize, aColor); might be reasonable, but it's still not as clean as putting the parameter description right there where the parameter itself is passed. Later, -- scott hess <shess@winternet.com> (WWW to "http://www.winternet.com/~shess/") Work: 12550 Portland Avenue South #121, Burnsville, MN 55337 (612)895-1208 <I want to become so famous that people buy tapes of me reading source code>
From: shill@iphysiol.unil.ch (Sean Hill) Newsgroups: comp.sys.next.programmer Subject: 3DKit using OpenGL on Win NT? Date: 16 Jul 1996 18:48:43 GMT Organization: University of Lausanne CH (Switzerland) Message-ID: <4sgo6b$qsj@cisun2000.unil.ch> I'm wondering if anyone's looking at how to setup a OpenGL context inside a NSWindow under OpenStep NT? You can get the handle to the WINDOWS(tm) window from NSWindow. Is it possible to do more? I would imagine lot's of people would find that handy and it would be slick to have a 3DKit type API for doing it. What's the chance? Maybe John Carmack has thought about this since he's porting the Quake editor to WinNT and OpenGL... Maybe? Thanks for any ideas, info. -Sean
From: Eric_Noyau@next.com (Eric Noyau) Newsgroups: comp.sys.next.programmer Subject: Re: How to draw a shaded sphere in PSwraps? Date: 16 Jul 1996 19:01:35 GMT Organization: NeXT Software, Inc. Message-ID: <4sgouf$cqi@news.next.com> References: <4sg9du$cmv@clarknet.clark.net> In article <4sg9du$cmv@clarknet.clark.net> hacker@clark.net (Dark Hacker) writes: > > I'm interested in drawing several shaded spheres on my screen so that > they appear like planets floating in three-space. Can I just create > these spheres in Photoshop or Illustrator and export the Adobe PS into a > PSwraps file and actually expect this to work? > > Is there a way to create a sphere using the PS inline functions? That's > probably ineffecient but it may be an alternative. > Shaded spheres ? Here you are: Obfuscated version: --- Cut here --- /d{def}def/i{add d}d/Z{rand 2147483647 div mul d}d 99 -2 9{-1 1/X 500 Z/Y 700 Z/C .3 Z{C setgray/r exch d X Y r 0 360 arc fill/X X .5 i/Y Y .6 i/C C 1 C sub r div i}for}for showpage --- Cut here --- Version with comments: --- Cut here --- % <radius> <positionX> <positionY> <color> sphere /sphere { /C exch def % Color (patern of gray) /Y exch def % Position Y /X exch def % Position X /R exch def % Radius R -1 1 { % Build one shadowed sphere /r exch def % r = radius C setgray % Drawing color for the sphere X Y r 0 360 arc fill % Build an uniform color circle /X X .5 add def % Move the center to the upper right /Y Y .6 add def /C C 1 C sub r div add def % Make the color clearer } for } def % return a random number between 0 and % the number on the stack /randomize {rand 2147483647 div mul} def 99 -2 9 { % Build some spheres, largest first % the loop is done on the diameter. 500 randomize % position X 700 randomize % position Y .3 randomize % color sphere }for showpage --- Cut here --- I really like the first one... I'm sure it's possible to reduce the code a little more. -- Eric
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 19:23:55 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sgq8b$4uc@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggob$iof@cetus.ali.bc.ca> Allan Noordvyk wrote: > For a project of any non-trivial size, I prefer the new typed argument style > of OpenStep. I'd rather find my bugs sooner rather than later. I have been writing and using Objective-C libraries of 300+ classes (with my own FoundationKit and classes for all the basic C data types) without *any* Class type casting for almost four years now and I can tell you that bad message sends were never a significant source of bugs. That is, I almost never have a bad message send show up after initial debugging. Type casting creates far more problems than it solves. Steve -- Remaining wild Tigers:<6K, BlackRhino:<2K GiantPanda:<1K [envirolink.org] Human population in 1900: 1.5 billion, in 1996: 5.7 billion. [census.gov]
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: Tue, 16 Jul 1996 16:38:19 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> In-Reply-To: <acwx04ca9t.fsf@parka.winternet.com> Excerpts from netnews.comp.sys.next.programmer: 16-Jul-96 Re: OpenStep design complaints by Scott Hess@parka.wintern > I'll weigh in with a bit veto on that one. This is one of my big > complaints about C++ in general, right up there with "Why do method > calls look like structure dereferences?" [Yeah, yeah, because in > some sense they _are_ structure dereferences. C++ methods are normally invoked as functions addressed through a structure deference, because (most) C++ implementations do not have a dynamic runtime environment. Objective-C objects themeselves are implemented as structures in a manner that is very similar to C++ objects. In Obj-C, however, there is the runtime which contains metainformation about which method implementations that are associated with any particular class or instance. Obj-C method calls don't look like structure deferences because Obj-C method calls are dynamicly invoked through the Obj-C runtime instead of being implemented using structure deferencing. Generally, I slightly prefer the '[object method]' syntax to 'object.method()', but the minor difference in syntactical sugar isn't important at all relative to my other complaints about C++. [ ... ] > Now, admittedly, it should be relatively easy to see what's up here, > since I used descriptive variable names. But how can you tell if > aSize and aColor are in the right order? Or change aSize to ss and > aColor to cc. Add four more parameters. Now can you tell which > is which without looking at the header files? First, you have to remember the various parts of the Obj-C method selector-- if you get the order wrong, you'll get the same category of error that you would switching the parameters around in a C++ or Java method call. Secondly, let's pursue your example of a 6-variable method call: [obj initWithSize:aSize withColor:aColor withFoo:aFoo withBar:aBar withBaz:aBaz withBop:aBop]; obj.init(aSize, aColor, aFoo, aBar, aBaz, aBop); Call me crazy, but I prefer the second version. It's a lot more readable; it's shorter; and, if you use descriptive variable names in your code, it's as self-documenting as it needs to be. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: aisbell@ix.netcom.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 16 Jul 1996 21:01:27 GMT Organization: Netcom Distribution: world Message-ID: <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> shess@parka.winternet.com (Scott Hess) wrote: >> My view on this is that Objective-C's method naming conventions > tend to promote (though don't enforce) a higher level of code > self-documentation. Compare: > [obj init]; > [obj initSize:aSize]; > [obj initSize:aSize color:aColor]; > ... > with: > obj.init(); > obj.init( aSize); > obj.init( aSize, aColor); > Now, admittedly, it should be relatively easy to see what's up here, > since I used descriptive variable names. But how can you tell if > aSize and aColor are in the right order? Or change aSize to ss and > aColor to cc. Add four more parameters. Now can you tell which > is which without looking at the header files? > obj.initWithSizeAndColor( aSize, aColor); might be reasonable, but > it's still not as clean as putting the parameter description right > there where the parameter itself is passed. Amen! This is a REALLY NICE Objective-C (and Smalltalk, of course) feature that I wish had been adopted for Java because Java has enough other nice features that I'd like to see it REPLACE Objective-C as the OPENSTEP native language. If that were done quickly, NeXT might have its first commercially-successful product. Maybe a Java method name convention could merely substitute '_' where ':' would be in Objective-C selectors: obj.initSize_color_( aSize, aColor). Still not as nice as having the argument description immediately before the argument itself :-( -- Art Isbell NeXT/MIME Mail: aisbell@ix.netcom.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: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints (Examples) Date: 16 Jul 1996 21:47:29 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sh2lh$7vh@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggob$iof@cetus.ali.bc.ca> <4sgq8b$4uc@news.onramp.net> Cc: dekorte@suite.com > Type casting creates far more problems than it solves. As an example of problems we run into *within NeXT's own kits*: NSWindow.h defines: - (void)setContentView:(NSView *)aView; NSScrollView defines: - (void)setContentView:(NSClipView *)aView; So, among other problems, you end up with a ton of compiler warnings that you can never get rid of. Steve
From: giddings@menominee.menominee.chem.wisc.edu (Michael Giddings) Newsgroups: comp.sys.next.programmer Subject: Help with NSArchiver Date: 16 Jul 1996 22:28:55 GMT Organization: University of Wisconsin Madison Distribution: world Message-ID: <4sh537$2n20@news.doit.wisc.edu> Keywords: nsarchiver I have an app in which I'd like to implement archiving using NSArchiver. It has one main "controller" object which is the target of all of the menu commands, including "Save". This controller object stores the id's of two NSObject subclasses that need to be archived. The problem is, given the FoundationKit docs, I can't figure out how to do it cleanly. It seems the only method available is archiveRootObject:toFile:, which appears to be designed only for archiving one single root object (not two, such as I need). The docs don't say what happens if this method is called multiple times with the same file name but different objects, but I assume it is overwritten on the second call which would cause the first to be lost (if it's not, then what happens when you need to overwrite a new version of the document being saved?). I thought also of creating a file wrapper containing two files, one for each root object, but am not familiar with how to create a new directory using foundation kit. I thought about implementing encodeWithCoder/initWithCoder in the controller object, but that wouldn't work because it's initialization is done already from the .nib file. The only other thought I had was redesigning all my objects (#%$#) so there is only one that needs to be archived. Not my first choice. So I'm stuck unless I start doing some really klugey stuff. Does anyone have ideas on how this could be done the "right" way? Thanks in advance. -- ----------------------------------------------------------------------------- Michael Giddings \ For Work: Graduate Studies, Scientific Comp. UW Madison Dept. of Chemistry \ DNA sequencing Technologies Madison, Wisconsin \ DNA Fragment analysis (608) 692-2851 \ giddings@whitewater.chem.wisc.edu \ For Fun: Whitewater Kayaking UW Hoofers Water Safety Board Chair \ And anything else that gets me outdoors ------------------------------------------------------------------------------
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Java Bashing (Was: OpenStep design complaints) Date: 17 Jul 1996 00:01:08 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Distribution: world Message-ID: <4shag4$aiv@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> Cc: aisbell@ix.netcom.com Art Isbell wrote: > ..I wish had been adopted for Java because Java has enough other > nice features Such as? > that I'd like to see it REPLACE Objective-C as the OPENSTEP > native language. If that were done quickly, NeXT might have its first > commercially-successful product. Java is a static OO language(like C++ and Simula). This means it can't support Objective-C's runtime in any reasonable way. Steve
From: ampriasm@students.wisc.edu (Andrew M. Priasmoro) Newsgroups: comp.sys.next.programmer Subject: Config.app Problem on Black Hardware. Date: Wed, 17 Jul 1996 01:00:48 -0400 Organization: University of Wisconsin-Madison Message-ID: <ampriasm-1707960100480001@f180-175.net.wisc.edu> Hi, Could anyone tell me what to do if my Config.app icon is in question mark? When I tried to execute it, It said that the application is damage. Is there any way to fix this without having to reinstall the OS? Thanks. Andrew. -- Andrew Priasmoro University of Wisconsin-Madison
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 17 Jul 1996 09:22:53 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4sibdd$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > Art Isbell wrote: > > that I'd like to see it REPLACE Objective-C as the OPENSTEP > > native language. If that were done quickly, NeXT might have its first > > commercially-successful product. > > Java is a static OO language(like C++ and Simula). > This means it can't support Objective-C's runtime in any reasonable way. Ummm...that's just not true. Java has more in common with Objective-C than it does with C++. The C++ look of the language is superficial. It certainly lacks many things I consider _highly_ important (categories tops the list and it certainly doesn't end there) but it is also a lot better than C++. At least the class' type is not bound until runtime. (OK, it can be bound to a point, but if you always use the root class and casts...you effectively have Obj-C's id type.) I think the C++ syntax usage is ugly, but I also think it was a good move for Sun to make as far as market penetration goes. If someone here were enterprising enough, they could always create a new compiler that takes their favorite syntax and turn it into Java byte codes. If someone would care to start with Objective-C I'd be very appreciative. :-) I'd posit that it could almost be done (or darned close to it, at any rate), given what little I know of the Java VM. Of course I reserve the right to be proven horribly wrong. Lest anyone think I'm a raving Java fanatic, I'm not. I _much_ prefer Objective-C. But if I had to pick another language to use, it would be a close race between SmallTalk and Java...for wildly differing reasons. And Art has a point. OPENSTEP on top of Java could do very well, simply because there's a lot of hype they could ride (or be buried under--it is a dangerous game). It won't happen, though, simply because NeXT isn't into that silly "changing the world" type of thing anymore. There's no profit in it. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
Newsgroups: comp.lang.objective-c,comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Identifying Type Message-ID: <DuoLup.GpK@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <x7buhlhnm4.fsf@tom.es.ele.tue.nl> <4s61jb$3c8@news.onramp.net> <4s6ebc$32g@saturn.genoa.com> Date: Wed, 17 Jul 1996 09:47:13 GMT In article <4s6ebc$32g@saturn.genoa.com> Alex Blakemore <alex@genoa.com> writes: >quite so bad (and its a big improvement over typing _everything as id) And I >usually push protocols pretty heavily. This isn't quite correct; not "everything" was typed as "id"; only the user level objects were. I imagine that time-critical operations, like clipping, flushing window caches to screen etc. were implemented in statically bound code, not using "id" and the Objective C messenger... So I think that you had originally more of a distincition between how things were presented to the user (application programmer) and how they were internally implemented. Well, anyhow, this discussion between static typing or not seems to be eternal... although it's easy to give arguments in favor of static typing ( the compiler will do better type checking etc.) we haven't yet seen of course a product develiverd on time, that uses these concepts... I mean, the original NextStep was after all perfectly well working, and did use these "bad" techniques of typing user-level objects as "id". David.
From: bruno@gruick.univ-lr.fr (Bruno Garnier) Newsgroups: comp.sys.next.programmer Subject: EOF informix change database Date: 17 Jul 1996 10:33:05 GMT Organization: Universite de La Rochelle Message-ID: <4sifh1$g8s@hpuniv.univ-lr.fr> Hi, I work with EOF 1.1, Informix Adaptor, NextStep I work with a database named "dbname", and when using my application i want to use an other database named "dbname2" that's copy of the first (same tables,but differents data) How can i do it programatically. 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: Philippe Mougin <mougin@aar.alcatel-alsthom.fr> Newsgroups: comp.sys.next.programmer Subject: Re: Help with NSArchiver Date: Wed, 17 Jul 1996 14:59:55 +0200 Organization: Alcatel Alsthom Recherche, Marcoussis, France Message-ID: <31ECE3CB.167EB0E7@aar.alcatel-alsthom.fr> References: <4sh537$2n20@news.doit.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Michael Giddings wrote: > > I have an app in which I'd like to implement archiving using NSArchiver. It > has one main "controller" object which is the target of all of the menu > commands, including "Save". This controller object stores the id's of two > NSObject subclasses that need to be archived. > When you want to save your 2 objects, you put them in an NSArray and archive it as the root object. When you want to reload your 2 objects, you unarchive the NSArray, they will be here. Hope it's ok for your problem. --------- Phil.
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints (Examples) Date: 17 Jul 1996 13:01:22 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4sio72$ee9@btmpjg.god.bel.alcatel.be> References: <4sh2lh$7vh@news.onramp.net> Steve Dekorte writes > > > Type casting creates far more problems than it solves. > > As an example of problems we run into *within NeXT's own kits*: > > NSWindow.h defines: > > - (void)setContentView:(NSView *)aView; > > NSScrollView defines: > > - (void)setContentView:(NSClipView *)aView; > > So, among other problems, > you end up with a ton of compiler warnings that you can never get rid of. You can, but you have to use "<Class> *" instead of "id" too, as in NSScrollView * sv; [sv setContentView:aView]; or id sv; [(NSScrollView *)sv setContentView:aView]; I very like the id type, but I must admit that there are advantages in the <Class> * approach too - especially because you know from the header file which kind of object is expected. Yours, -- Ralf Suckow |------------------------ suckow@bln.sel.alcatel.de | All opinions are mine.
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,misc.jobs.contract Subject: NEXTSTEP/Contract DC Area Date: 17 Jul 1996 13:50:59 GMT Organization: Norden 1 Communications Message-ID: <4sir44$d8n@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------Commercial experience Objective C-----------------Commercial experience EOF-------------------------A Plus Sybase or Oracle------------A Plus Contract--------------------Long term Area------------------------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: edwards@sewp.nasa.gov Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: Wed, 17 Jul 1996 15:22:37 GMT Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <960717112237.5053AADnF.edwards@vege> References: <4selec$fav@news.onramp.net> <960716095528.4434AADnG.edwards@vege> <4sgl6m$imp@news.onramp.net> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII > From: dekorte@suite.com (Steve Dekorte) > Subject: Re: OpenStep design complaints > Date: 16 Jul 1996 17:57:42 GMT > Newsgroups: comp.sys.next.programmer > > edwards@sewp.nasa.gov wrote: > > dekorte@suite.com: > > > 3. Behaviorism violations > > > > > > There are also type casts all over so you can't create your own classes with > > > the same behavior to interoperate with theirs. This is anti-OO. > > > > Could you provide some specific examples? > > - (void)setStringValue:(NSString *)aString; > > Guess what happens when I try to pass in another object with similiar behaivior? > It won't compile. > > Steve Assuming this is a method on one of NeXT's classes, how would you know what behavior was required of the argument you are passing in? Since you can't assume a priori which particular messages will be sent to your NSString- impersonator, you would have to provide every method that NSString provides. And if you do that, you might as well make your class a subclass of NSString and override some or all of its methods. And if you do that, then it would be perfectly legal to pass in an instance of your class here. I find this kind of typing more liberating than restrictive. It means I can rely less on the documentation. I can just look at a method and know exactly what's expected of its arguments and return values. And subclassing provides a way to integrate customized classes. Granted this represents a significant paradigm shift for developers converting code based on the freewheeling environment of the old AppKit. But I don't think there's anything inherently wrong with the approach they've taken. Steve --- Stephen G. Edwards SEWP BOWL, NASA Goddard Space Flight Center Tel: 301-286-0783 Fax: 301-286-1619 edwards@sewp.nasa.gov
From: indy (Steve Weintz) Newsgroups: comp.sys.next.programmer Subject: Re: How to draw a shaded sphere in PSwraps? (long) Date: 17 Jul 1996 15:27:25 GMT Organization: EthnoGraphics, Ltd. Message-ID: <4sj0ot$g03@treflan.shout.net> References: <4sg9du$cmv@clarknet.clark.net> Cc: hacker@clark.net In <4sg9du$cmv@clarknet.clark.net> Dark Hacker wrote: > > I'm interested in drawing several shaded spheres on my screen so that they > appear like planets floating in three-space. Can I just create these spheres > in Photoshop or Illustrator and export the Adobe PS into a PSwraps file and > actually expect this to work? > > Is there a way to create a sphere using the PS inline functions? That's > probably ineffecient but it may be an alternative. > > How should I approach this? Mail = hacker@computation.com. > Thanks! Umm...why not use RenderMan? render the following RIB file thusly: >prman SamplePlanets.rib ------ C U T H E R E ----------- ##RenderMan RIB-Structure 1.0 ##System NeXT Release 3 ##Scene /Users/indy/SamplePlanets.rib ##Creator solidThinking Rel:2.3 Ver:171 ##CreationDate Wed ##For indy ##CameraOrientation 10.143830 8.043050 -3.111946 -1.476519 3.503783 0.123046 # Shutter 0.0 1.0 Option "shadow" "bias0" [ .05 ] "bias1" [ .1 ] Imager "clamptoalpha" ShadingRate 1 FrameBegin 1 Display "SamplePlanets.tiff" "file" "rgba" Clipping .01 1000 ScreenWindow -1 1 -.584222 .584222 Projection "perspective" "fov" [ 70.400002 ] Transform [ .268192 -.339304 -.901635 0 0 .935922 -.352207 0 .963366 .094459 .251007 0 .277451 -3.791879 12.75997 1 ] Format 469 274 1 WorldBegin Declare "intensity" "float" Declare "lightcolor" "color" LightSource "ambientlight" 9 "intensity" [ .2 ] "lightcolor" [ 0 .207846 .392163 ] Declare "intensity" "float" Declare "lightcolor" "color" Declare "from" "point" Declare "to" "point" LightSource "distantlight" 10 "intensity" [ 1 ] "lightcolor" [ 1 .877041 .709891 ] "from" [ 19.124878 22.293884 -20.171516 ] "to" [ 0 0 0 ] AttributeBegin Attribute "identifier" "name" [ "Class_M_world" ] TransformBegin ConcatTransform [ 1.57726 0 0 0 0 1.57726 0 0 0 0 1.57726 0 -2.098935 5.924119 -3.481158 1 ] AttributeBegin Attribute "identifier" "name" [ "World_1" ] TransformBegin Scale .5 .5 .5 Declare "Ka" "float" Declare "Kd" "float" Declare "spectral_exp" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "bump_scale" "float" Declare "multifractal" "float" Declare "dist_scale" "float" Declare "offset" "float" Declare "sea_level" "float" Declare "mtn_scale" "float" Declare "lat_scale" "float" Declare "nonlinear" "float" Declare "purt_scale" "float" Declare "map_exp" "float" Declare "ice_caps" "float" Declare "depth_scale" "float" Declare "depth_max" "float" Declare "mottle_limit" "float" Declare "mottle_scale" "float" Declare "mottle_dim" "float" Declare "mottle_mag" "float" Surface "terran" "Ka" [ .5 ] "Kd" [ .7 ] "spectral_exp" [ .5 ] "lacunarity" [ 10.002 ] "octaves" [ 10 ] "bump_scale" [ .14 ] "multifractal" [ .05 ] "dist_scale" [ .2 ] "offset" [ 0 ] "sea_level" [ .00325 ] "mtn_scale" [ 1 ] "lat_scale" [ .95 ] "nonlinear" [ 0 ] "purt_scale" [ .9 ] "map_exp" [ 0 ] "ice_caps" [ .9 ] "depth_scale" [ 1 ] "depth_max" [ .5 ] "mottle_limit" [ .75 ] "mottle_scale" [ 20 ] "mottle_dim" [ .25 ] "mottle_mag" [ .02 ] TransformEnd Attribute "dice" "binary" [ 1 ] TransformBegin TransformBegin Rotate 90 1 0 0 Sphere .810103 -.810103 .810103 360 TransformEnd TransformEnd AttributeEnd AttributeBegin Attribute "identifier" "name" [ "Atmosphere" ] TransformBegin Scale .21 .21 .21 Declare "Ka" "float" Declare "Kd" "float" Declare "distortionscale" "float" Declare "omega" "float" Declare "lambda" "float" Declare "octaves" "float" Declare "offset" "float" Surface "KMPlanetClouds" "Ka" [ .418 ] "Kd" [ .904 ] "distortionscale" [ 1.36 ] "omega" [ .7 ] "lambda" [ 2 ] "octaves" [ 9 ] "offset" [ 0 ] TransformEnd Attribute "dice" "binary" [ 1 ] TransformBegin TransformBegin Rotate 90 1 0 0 Sphere .82 -.82 .82 360 TransformEnd TransformEnd AttributeEnd TransformEnd AttributeEnd AttributeBegin Attribute "identifier" "name" [ "Moon_1" ] TransformBegin Scale .05 .05 .05 Declare "Ka" "float" Declare "Kd" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "H" "float" Declare "highland_color" "color" Declare "maria_basecolor" "float" Declare "maria_color" "float" Declare "arg22" "float" Declare "arg23" "float" Declare "highland_threshold" "float" Declare "highland_altitude" "float" Declare "maria_altitude" "float" Declare "peak_rad" "float" Declare "inner_rad" "float" Declare "rim_rad" "float" Declare "outer_rad" "float" Declare "peak_ht" "float" Declare "rim_ht" "float" Declare "numrays" "float" Declare "rayfade" "float" Surface "KMLuna" "Ka" [ .5 ] "Kd" [ 1 ] "lacunarity" [ 2 ] "octaves" [ 8 ] "H" [ .3 ] "highland_color" [ .700003 .700003 .700003 ] "maria_basecolor" [ 7 ] "maria_color" [ .1 ] "arg22" [ 1 ] "arg23" [ .3 ] "highland_threshold" [ -.2 ] "highland_altitude" [ .001 ] "maria_altitude" [ .0004 ] "peak_rad" [ 0075 ] "inner_rad" [ .01 ] "rim_rad" [ .02 ] "outer_rad" [ .05 ] "peak_ht" [ .005 ] "rim_ht" [ .003 ] "numrays" [ 8 ] "rayfade" [ 1 ] TransformEnd TransformBegin Scale .05 .05 .05 Declare "spectral_exp" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "bump_scale" "float" Declare "multifractal" "float" Declare "dist_scale" "float" Declare "offset" "float" Declare "sea_level" "float" Displacement "KMTerranBump" "spectral_exp" [ .5 ] "lacunarity" [ 2 ] "octaves" [ 7 ] "bump_scale" [ .0025 ] "multifractal" [ 0 ] "dist_scale" [ .2 ] "offset" [ 0 ] "sea_level" [ 0 ] TransformEnd TransformBegin ConcatTransform [ .938144 0 0 0 0 .938144 0 0 0 0 .938144 0 5.170546 7.459935 -2.866837 1 ] TransformBegin Rotate 90 1 0 0 Sphere .2 -.2 .2 360 TransformEnd TransformEnd AttributeEnd AttributeBegin Attribute "identifier" "name" [ "Class_M_world_1" ] TransformBegin ConcatTransform [ 1 0 0 0 0 1 0 0 0 0 1 0 3.532352 4.285935 0 1 ] AttributeBegin Attribute "identifier" "name" [ "World_1" ] TransformBegin Scale .5 .5 .5 Declare "Ka" "float" Declare "Kd" "float" Declare "spectral_exp" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "bump_scale" "float" Declare "multifractal" "float" Declare "dist_scale" "float" Declare "offset" "float" Declare "sea_level" "float" Declare "mtn_scale" "float" Declare "lat_scale" "float" Declare "nonlinear" "float" Declare "purt_scale" "float" Declare "map_exp" "float" Declare "ice_caps" "float" Declare "depth_scale" "float" Declare "depth_max" "float" Declare "mottle_limit" "float" Declare "mottle_scale" "float" Declare "mottle_dim" "float" Declare "mottle_mag" "float" Surface "terran" "Ka" [ .5 ] "Kd" [ .7 ] "spectral_exp" [ .5 ] "lacunarity" [ 10.002 ] "octaves" [ 10 ] "bump_scale" [ .14 ] "multifractal" [ .05 ] "dist_scale" [ .2 ] "offset" [ 0 ] "sea_level" [ .00325 ] "mtn_scale" [ 1 ] "lat_scale" [ .95 ] "nonlinear" [ 0 ] "purt_scale" [ .9 ] "map_exp" [ 0 ] "ice_caps" [ .9 ] "depth_scale" [ 1 ] "depth_max" [ .5 ] "mottle_limit" [ .75 ] "mottle_scale" [ 20 ] "mottle_dim" [ .25 ] "mottle_mag" [ .02 ] TransformEnd Attribute "dice" "binary" [ 1 ] TransformBegin TransformBegin Rotate 90 1 0 0 Sphere .810103 -.810103 .810103 360 TransformEnd TransformEnd AttributeEnd AttributeBegin Attribute "identifier" "name" [ "Atmosphere" ] TransformBegin Scale .21 .21 .21 Declare "Ka" "float" Declare "Kd" "float" Declare "distortionscale" "float" Declare "omega" "float" Declare "lambda" "float" Declare "octaves" "float" Declare "offset" "float" Surface "KMPlanetClouds" "Ka" [ .418 ] "Kd" [ .904 ] "distortionscale" [ 1.36 ] "omega" [ .7 ] "lambda" [ 2 ] "octaves" [ 9 ] "offset" [ 0 ] TransformEnd Attribute "dice" "binary" [ 1 ] TransformBegin TransformBegin Rotate 90 1 0 0 Sphere .82 -.82 .82 360 TransformEnd TransformEnd AttributeEnd TransformEnd AttributeEnd AttributeBegin Attribute "identifier" "name" [ "Moon_1_1" ] TransformBegin Scale .05 .05 .05 Declare "Ka" "float" Declare "Kd" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "H" "float" Declare "highland_color" "color" Declare "maria_basecolor" "float" Declare "maria_color" "float" Declare "arg22" "float" Declare "arg23" "float" Declare "highland_threshold" "float" Declare "highland_altitude" "float" Declare "maria_altitude" "float" Declare "peak_rad" "float" Declare "inner_rad" "float" Declare "rim_rad" "float" Declare "outer_rad" "float" Declare "peak_ht" "float" Declare "rim_ht" "float" Declare "numrays" "float" Declare "rayfade" "float" Surface "KMLuna" "Ka" [ .5 ] "Kd" [ 1 ] "lacunarity" [ 2 ] "octaves" [ 8 ] "H" [ .3 ] "highland_color" [ .700003 .700003 .700003 ] "maria_basecolor" [ 7 ] "maria_color" [ .1 ] "arg22" [ 1 ] "arg23" [ .3 ] "highland_threshold" [ -.2 ] "highland_altitude" [ .001 ] "maria_altitude" [ .0004 ] "peak_rad" [ 0075 ] "inner_rad" [ .01 ] "rim_rad" [ .02 ] "outer_rad" [ .05 ] "peak_ht" [ .005 ] "rim_ht" [ .003 ] "numrays" [ 8 ] "rayfade" [ 1 ] TransformEnd TransformBegin Scale .05 .05 .05 Declare "spectral_exp" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "bump_scale" "float" Declare "multifractal" "float" Declare "dist_scale" "float" Declare "offset" "float" Declare "sea_level" "float" Displacement "KMTerranBump" "spectral_exp" [ .5 ] "lacunarity" [ 2 ] "octaves" [ 7 ] "bump_scale" [ .0025 ] "multifractal" [ 0 ] "dist_scale" [ .2 ] "offset" [ 0 ] "sea_level" [ 0 ] TransformEnd TransformBegin ConcatTransform [ 1.668492 0 0 0 0 1.668492 0 0 0 0 1.668492 0 -1.587 -1.652513 0 1 ] TransformBegin Rotate 90 1 0 0 Sphere .2 -.2 .2 360 TransformEnd TransformEnd AttributeEnd AttributeBegin Attribute "identifier" "name" [ "Sphere" ] Color [ 1 .903409 .717658 ] TransformBegin Scale 1 1 1 Declare "Ka" "float" Declare "Kd" "float" Declare "lacunarity" "float" Declare "octaves" "float" Declare "H" "float" Declare "highland_color" "color" Declare "maria_basecolor" "float" Declare "maria_color" "float" Declare "arg22" "float" Declare "arg23" "float" Declare "highland_threshold" "float" Declare "highland_altitude" "float" Declare "maria_altitude" "float" Declare "peak_rad" "float" Declare "inner_rad" "float" Declare "rim_rad" "float" Declare "outer_rad" "float" Declare "peak_ht" "float" Declare "rim_ht" "float" Declare "numrays" "float" Declare "rayfade" "float" Surface "KMLuna" "Ka" [ .5 ] "Kd" [ 1 ] "lacunarity" [ 2 ] "octaves" [ 8 ] "H" [ .3 ] "highland_color" [ .700003 .677084 .559816 ] "maria_basecolor" [ 7 ] "maria_color" [ .1 ] "arg22" [ 1 ] "arg23" [ .3 ] "highland_threshold" [ -.2 ] "highland_altitude" [ .001 ] "maria_altitude" [ .0004 ] "peak_rad" [ 0075 ] "inner_rad" [ .01 ] "rim_rad" [ .02 ] "outer_rad" [ .05 ] "peak_ht" [ .005 ] "rim_ht" [ .003 ] "numrays" [ 8 ] "rayfade" [ 1 ] TransformEnd TransformBegin ConcatTransform [ 1 0 0 0 0 1 0 0 0 0 1 0 -4.863385 7.62939e-06 7.177093 1 ] TransformBegin Rotate 90 1 0 0 Sphere 3.993095 -3.993095 3.993095 360 TransformEnd TransformEnd AttributeEnd WorldEnd FrameEnd --------- C U T H E R E ------------- ------ Here's an URL for the RenderMan Repository, for the shaders used: http://pete.cs.caltech.edu/RMR/Shaders/KMShaders.htmld/index.html -- | Steve Weintz 217.244.3074 * indy@uiuc.edu (MIME ok) | 217.344.5303 * indy@shout.net (NeXTMail ok!) | "The anaconda is very dangerous. I'll stand over here while | Jim attaches the radio collar." -- Marlin
From: bediger@csn.net (Bruce Ediger) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Reverse ARP under NeXTStep 3.x? Date: 17 Jul 1996 12:17:54 -0600 Organization: H1DE()U5 MUT4NT5!!1!!! Distribution: world Message-ID: <4sjaoi$r1j@teal.csn.net> For very peculiar reasons, I'd like to net-boot a Sun SPARCstation with a NeXT 25MHz slab as a boot server. I can find the tftp stuff to serve the boot and kernel files with, but I think I need a Reverse ARP daemon (rarpd) for the SPARCStation to get its IP address from. Does such a beast exist for NeXTStep? I found 4.4BSD source code for a rarpd, but it uses "Berkeley Packet Filter" (bpf) to go beneath the raw IP layer. Therefore, that source is pretty useless for my purposes. What's the NeXTStep analog of "bpf" or SunOS's "nit"? Is there some way to listen for RARP packets under NeXTStep? Regards, Bruce Ediger -- Try this in a "DOS Box" for fun and profit!!! echo f 0:0 ffff 0 | debug
From: heisec@cs.tu-berlin.de (Christian Heise) Newsgroups: comp.sys.next.programmer Subject: Saving RTF-Text to GUPTA-SQLBase Date: 17 Jul 1996 20:27:44 GMT Organization: Technical University of Berlin, Germany Message-ID: <4sjic0$qfe@news.cs.tu-berlin.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Summary: Can't save RTF-Text to GUPTA using DBKit Hello, I tried to save RTF-Text from a normal textobject to the Gupta-SQLBase using the dbkit and the Gupta-adaptor, but I can't get it work. The column in the database is of type LONGVAR, in the dbmodel I set the type to NXData like it's described in the reference. But when I connect a textfield in InterfaceBuilder to that column, go to testmode, insert a text and try to save it to the database it won't work. I tried every trick I knew, but none of them worked. Has anyone experienced the same or where is my mistake. I think it also can be a bug in the adaptor, but I don't know it. Christian.
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Re: How to draw a shaded sphere in PSwraps? Date: 17 Jul 1996 22:23:27 GMT Organization: Rockwell Avionics - Collins Message-ID: <4sjp4v$v4m@castor.cca.rockwell.com> References: <4sg9du$cmv@clarknet.clark.net> Cc: hacker@clark.net Is there any reason not to use eps from any drawing program ? - Erik M. Buck
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 17 Jul 1996 22:54:26 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sjqv2$r9@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> don@misckit.com wrote: > > Java is a static OO language(like C++ and Simula). > > This means it can't support Objective-C's runtime in any reasonable way. > > Ummm...that's just not true. Java has more in common with > Objective-C than it does with C++. Perphaps you can show us how to use selectors(or -perform:) in Java? Steve
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 17 Jul 1996 22:41:29 GMT Organization: Rockwell Avionics - Collins Message-ID: <4sjq6p$v4m@castor.cca.rockwell.com> References: <4selec$fav@news.onramp.net> <960716095528.4434AADnG.edwards@vege> <4sgl6m$imp@news.onramp.net> Cc: dekorte@suite.com In <4sgl6m$imp@news.onramp.net> Steve Dekorte wrote: > edwards@sewp.nasa.gov wrote: > > - (void)setStringValue:(NSString *)aString; > > Guess what happens when I try to pass in another object with similiar behaivior? > It won't compile. > > Steve > > Probably, every class should also have a corresponding protocol. Then any class that implements the protocol could be used rather than only subclasses. - (void)setStringValue:(id <NSString>)aString; The above would be acceptable. @protocol NSString <NSObject> // NSString's interface here @end @interface NSString : NSObject <NSString> {} @end Of course, we could implement our own subclass of NSString that replaced every single method. We might even be able to reuse the NSString implementation sometimes. Oh Yah, that is what implementation inheritance is for...
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Self Date: 18 Jul 1996 01:24:42 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sk3oq$727@news.onramp.net> In responce to a post on comp.sys.next.programmer, I've gotten several requests for more info on Self. I thought this might be of general interest: > what compelling advantage does Self offer? There are a number, I'll start with the biggest: The idea behind OO is reuse by data-structure independence. In static OO languages(C++, Java, Simula, etc.) your objects are both dependent on their own structure and the structure of objects they send messages to. (This is why many don't cosider these true OO languages) In dynamic OO languages(Smalltalk, Objective-C, etc) you objects are only dependent on their own structure(until you start type casting...) (all messages are dynamically bound) In behavioristic prototype-based laguages(Self, Obliq, etc) your not even dependent on our own data-structure. In fact, objects have no static data-structures. All variables are accessed by messages. You would think this would cause poor performance but Self has compiler technology that allows it to do messaging about as fast as Objective-C. This creates new possiblilities, such as being able to dynamically add/remove variables&methods from classes and change class hierarchies. Also, new instances are created by copying objects, so you can play with a new class before propogating it's changes - this is where the "prototype" word comes in. Now, that said, imagine what sort of development environment you could create? No waiting for compiles. No class interference problems as in Smalltalk. Self also comes with a innovative graphical class editing environment. The big problems are that it's not multi-platform yet and the related problem that there's no GNU version(or free version with source). Steve
From: "The Washington Firm, Ltd." <wafirm@ix.netcom.com> Newsgroups: comp.sys.next.programmer Subject: SEATTLE - LAN Services Operations Specialist Date: Wed, 17 Jul 1996 18:13:49 -0700 Organization: Washington Firm Message-ID: <01bb7446.605a8b60$77e41fcc@karen95.ix.netcom.com> ***AT&T Wireless*** *POSITION:LAN Services Operations Specialist *LOCATION: Kirkland, WA *REFERENCE NUMBER: JO702 *TYPE : Full Time *REQUIREMENTS: BS in I.S./C.S. or equivalent experience. 2-5 yrs in Novell v3.x Networking environment. Strong working knowledge of TCP/IP and network experience. Familiarity with UNIX. *DESCRIPTION: Provide tech support to HQ PC Support group. Work with security group to ensure consistent administration of systems environments. System account management/maintenance. Responsible for system reliability. Please send your resume to: Attn: JO 2 Nickerson, Courtyard Suite; Seattle, WA 98109 Reference #: (JO702) FAX (206) 284-8844 Or, you may respond by email to "wafirm@ix.netcom.com", in PLAIN TEXT FORMAT within the body of the message.
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Java for NeXTStep Date: 18 Jul 1996 05:03:21 GMT Organization: E.S. Systems cc (Financial Systems Development) Message-ID: <4skgip$c0k@hermes.is.co.za> Is there a Java for NeXTStep available? I know that there were a port going on a while ago called "BLACKJAVA". Is this still being carried out or did they stop. I want to use WebObjects or any other product for that matter that can setup dynamic pages for the WEB that will contain Java Applets. The problem I am having is that this must be setup under NeXTStep because my Business Objects are running on this platform. -- 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.sysadmin,comp.sys.next.programmer Subject: Re: Reverse ARP under NeXTStep 3.x? Date: 18 Jul 1996 04:59:14 GMT Organization: E.S. Systems cc (Financial Systems Development) Message-ID: <4skgb2$bvp@hermes.is.co.za> References: <4sjaoi$r1j@teal.csn.net> In-Reply-To: <4sjaoi$r1j@teal.csn.net> On 07/17/96, Bruce Ediger wrote: >For very peculiar reasons, I'd like to net-boot a Sun SPARCstation with >a NeXT 25MHz slab as a boot server. I can find the tftp stuff to >serve the boot and kernel files with, but I think I need a Reverse >ARP daemon (rarpd) for the SPARCStation to get its IP address from. > >Does such a beast exist for NeXTStep? I found 4.4BSD source code for >a rarpd, but it uses "Berkeley Packet Filter" (bpf) to go beneath the >raw IP layer. Therefore, that source is pretty useless for my purposes. >What's the NeXTStep analog of "bpf" or SunOS's "nit"? Is there some >way to listen for RARP packets under NeXTStep? I don't know of a RARP for NeXTStep but if you want, you can have a look at the latest PPP on peak, there as a BPF LKS in there. It might help with the BSD 4.4 version you have -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: aisbell@ix.netcom.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 18 Jul 1996 06:42:24 GMT Organization: Netcom Distribution: world Message-ID: <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > don@misckit.com wrote: > > > Java is a static OO language(like C++ and Simula). > > > This means it can't support Objective-C's runtime in any reasonable way. > > Ummm...that's just not true. Java has more in common with > > Objective-C than it does with C++. > Perphaps you can show us how to use selectors(or -perform:) in Java? Perhaps you can show us how to use automatic garbage collection in Objective-C. My point wasn't to start a religious war over which language is "best". But rather to make a more practical observation that no matter how nice Objective-C is, it has failed as a commercially-acceptable language. Few programmers know it (and few realize how easy it is to learn). Finding good Objective-C programmers isn't easy. The commercial value of an large application suite or development effort isn't so great when Objective-C is the language used because it's not "standard". Whereas C++ is the rage, but not for any particularly good reason, as far as I can tell, other than it's a better C. But Java is the language darling of 1996 and, fortunately, it happens to be a lot better general object-oriented language than C++, so I'd like to see NeXT take advantage of its popularity and support it by making it OPENSTEP's native language. We'd lose some nice Objective-C features, but we'd gain some nice Java features, and we'd gain a lot more respectibility and credibility because we'd be more "mainstream" (gawd, I never thought I'd consider being mainstream to be a positive :-) -- Art Isbell NeXT/MIME Mail: aisbell@ix.netcom.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: Jean-Philippe VIEGAS <viegas@aar.alcatel-alsthom.fr> Newsgroups: comp.sys.next.programmer,comp.sys.next.software,fr.comp.sys.next Subject: EOF with Oracle Date: Thu, 18 Jul 1996 11:22:36 +0200 Organization: Alcatel-alsthom recherche Message-ID: <31EE025C.2781E494@aar.alcatel-alsthom.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, I'm using Enterprise Object Framework 1.1 with nextstep 3.3, with an ORACLE Database version 7.2 (first I've been using version 7.0). Can anybody explain me why, when I generate a model file, EOF doesn't recuperate the primary keys defined in the ORACLE database ? THANKS -- ------------------------------------------------------ Jean Philippe VIEGAS alcatel alsthom recherche viegas@aar.alcatel-alsthom.fr ------------------------------------------------------
From: tmb@best.com (--) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 18 Jul 1996 10:43:39 GMT Organization: home Message-ID: <TMB.96Jul18034339@best.best.com> References: <4sk3oq$727@news.onramp.net> In-reply-to: dekorte@suite.com's message of 18 Jul 1996 01:24:42 GMT FCC: /home/tmb/mail/x-nout In article <4sk3oq$727@news.onramp.net> dekorte@suite.com (Steve Dekorte) writes: In static OO languages(C++, Java, Simula, etc.) your objects are both[...] In dynamic OO languages(Smalltalk, Objective-C, etc) you objects are[...] Please stop spreading this kind of confusion. Java's and Objective-C's type systems are very similar, and the dichotomy that you are trying to create just doesn't exist in the other cases either. In behavioristic prototype-based laguages(Self, Obliq, etc) your not even dependent on our own data-structure. In fact, objects have no static data-structures. All variables are accessed by messages. You would think this would cause poor performance And one would be right: accessing instance variables through messages _is_ slow in any language. Now, that said, imagine what sort of development environment you could create? No waiting for compiles. The big problems are that it's not multi-platform yet and the related problem that there's no GNU version(or free version with source). For scripting and prototyping applications, languages like Self can be very useful. Self itself isn't widely used, but languages like Perl and Python are. For large projects and extended maintenance, however, they are less than ideal. Thomas.
From: giddings@menominee.menominee.chem.wisc.edu (Michael Giddings) Newsgroups: comp.sys.next.programmer Subject: File access in OpenStep (was: Help with NSArchiver) Date: 18 Jul 1996 13:47:39 GMT Organization: University of Wisconsin - Madison Distribution: world Message-ID: <4slf9r$2lie@news.doit.wisc.edu> Keywords: file openstep First, thanks to those who responded to my question about NSArchiver! The solution was simple: use an NSArray to encapsulate the two objects in question, then just archive the NSArray. It makes so much sense, no wonder I didn't think of doing it . . . On a related note, I'm wondering about file access in OpenStep. Though I don't have OS/4.0, having worked with foundation it appears there are no classes meant for direct file access. The only access to a filesystem seems to be given through methods as part of objects such as NSData and NSArchiver. Is there a file accessor class added to OpenStep, or is the lack of one an intentional choice to force the program to get data in & out of the program through a "standard" object such as NSData? I can see how the latter would increase portability, but it takes a bit of a mental shift to get used to the concept. I'm guessing the intention is to _not_ have the programmer use the standard C file libs. . . Thanks again -- ----------------------------------------------------------------------------- Michael Giddings \ For Work: Graduate Studies, Scientific Comp. UW Madison Dept. of Chemistry \ DNA sequencing Technologies Madison, Wisconsin \ DNA Fragment analysis (608) 692-2851 \ giddings@whitewater.chem.wisc.edu \ For Fun: Whitewater Kayaking UW Hoofers Water Safety Board Chair \ And anything else that gets me outdoors ------------------------------------------------------------------------------
From: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: accented characters Date: 18 Jul 1996 14:07:05 GMT Organization: Universite de La Rochelle Message-ID: <4slge9$b2i@hpuniv.univ-lr.fr> Hello, Have someone try to work with strings that contains accented characters ? I work with Informix. I work with EOF 1.1, but could other applications that not use EOF 1.1 (such as I-4GL or MS-Access with ODBC) read my data that contains accented characters ? Thanks for any reply. Wil wgaboria@iut-lr.univ-lr.fr
From: uhendjx@racer (Jonathan W. Hendry) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 18 Jul 1996 15:03:20 GMT Organization: Lexis-Nexis, Dayton OH Distribution: world Message-ID: <4sljno$g6b@mailgate.lexis-nexis.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> Art Isbell (aisbell@ix.netcom.com) wrote: : credibility because we'd be more "mainstream" (gawd, I never thought I'd : consider being mainstream to be a positive :-) It's amazing that anything *good* has managed to become mainstream. -- Jonathan W. Hendry Views expressed herein do Steel Driving Software, Inc. not represent those of steeldrv@ix.netcom.com Steel Driving Software, Inc. jon@exnext.com or Lexis-Nexis
From: colnet@loria.fr (Dominique Colnet) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 18 Jul 1996 16:21:52 GMT Organization: CRIN & INRIA-Lorraine - Nancy - FRANCE Message-ID: <4slob0$rmv@muller.loria.fr> References: <4sk3oq$727@news.onramp.net> <TMB.96Jul18034339@best.best.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit To: tmb@best.com (--) In article <TMB.96Jul18034339@best.best.com>, tmb@best.com (--) writes: |> In behavioristic prototype-based laguages(Self, Obliq, etc) your |> not even dependent on our own data-structure. In fact, objects |> have no static data-structures. All variables are accessed by |> messages. You would think this would cause poor performance |> |> And one would be right: accessing instance variables through messages |> _is_ slow in any language. Sorry, but the Self language use a polymorphic type inference algorithm. Accessing instance variables can be automatically inlined. Have a look at the (excellent) Phd Thesis of Ole AGESEN "Concrete Type Inference: Delivering Object Oriented Applications" Published by Sun Microsystems (SMLI TR-96-52). Ole Agesen has done a great job. IMHO, the expressivness of Self is really better than most other OOL. I also do automatic inlining in my Eiffel compiler (SmallEiffel). Program written in (Small)Eiffel are running as fast as C programs. -- ------------------------------------------------------------ Dominique COLNET -- Talk Eiffel with SmallEiffel Talk Eiffel C.R.I.N. (Centre de Recherche en Informatique de Nancy) POST: CRIN,BP 239,54506 Vandoeuvre les Nancy Cedex,FRANCE EMAIL: colnet@loria.fr VOICE:+33 83593079 FAX:+33 83413079 --------------------- Publicity for SmallEiffel. READ_ME file appended : -- Part of SmallEiffel -- Read DISCLAIMER file -- Copyright (C) -- Dominique COLNET and Suzanne COLLIN -- colnet@loria.fr -- -- SmallEiffel is not Smalltalk. D.Colnet 10/2/94 -- Is it still Eiffel ? -- 1 - What is SmallEiffel ? - SmallEiffel is a free Eiffel compiler distributed under the terms of the GNU General Public License as published by the Free Software Foundation. You can download SmallEiffel at : ftp://ftp.loria.fr/pub/loria/genielog/SmallEiffel 2 - Brief Description of SmallEiffel - SmallEiffel is the fruit of a research project done at CRIN (Centre de Recherche en Informatique de Nancy). SmallEiffel is intended to be a complete, though small and very fast, free Eiffel compiler. SmallEiffel is already used by students of the University Henri Poincare' at Nancy (FRANCE). We are using Eiffel as a first langage for teaching OOP since 1990 (SmallEiffel is used since september 1995). SmallEiffel is already portable on UNIX, Linux, BSD, HP-UX, IRIX, XENIX, DOS, OS2, WINDOWS 95, Macintosh and NeXT. Current distribution includes a compiler (command 'compile'), a pretty printer (command 'pretty') and other tools. An Eiffel interpretor will be added to the distribution in a few months (1 or 2 months). 3 - Contents of the SmallEiffel directory - 3.1 - directory misc : Miscellaneous files including the INSTALL file. 3.2 - directory man : Miscellaneous help files. 3.3 - directory bin : The binary place including the `compile' command. 3.4 - directory bin_c : C source code for tools of bin directory. 3.5 - directory lib_std : Standard Eiffel Library (string.e, integer.e, general.e, array.e, ...). 3.6 - directory lib_show : Demonstration and examples Eiffel programs. 3.7 - directory lib_test : Set of program to tests SmallEiffel (see also directory misc/self_test). 4 - Bug Report - Thank in advance for bug reports. Don't forget to have a look first in the misc/NOT_YET_IMPLEMENTED file. Please, try to isolate the bug as much as possible. The best is to be able to create a single file, as small a possible (which include the bug). D.Colnet - colnet@loria.fr
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 18 Jul 1996 19:11:27 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Distribution: world Message-ID: <4sm28v$ndf@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> Art Isbell wrote: > Steve Dekorte wrote: > > don@misckit.com wrote: > > > > Java is a static OO language(like C++ and Simula). > > > > This means it can't support Objective-C's runtime in any reasonable way. > > > > Ummm...that's just not true. > > > Perphaps you can show us how to use selectors(or -perform:) in Java? > > Perhaps you can show us how to use automatic garbage collection in > Objective-C. You can have GC in a static OO language. You can't have a decent runtime in one though. > ...[Java would make OpenStep more mainstream]... That would be fine if it were possible. Until you show me how to support an Objective-C runtime in a reasonable way(starting with selecotrs) in Java, I'm of the opinion that it's not. Steve
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 18 Jul 1996 21:08:35 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4sm94j$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sm28v$ndf@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > That would be fine if it were possible. Until you show me how to support > an Objective-C runtime in a reasonable way(starting with selecotrs) in > Java, I'm of the opinion that it's not. Well, I'm not really interested in perpetuating this thread, but I guess I'll just add a few more comments. Java, unlike C++, does have a runtime and can do late binding. Some of the stuff that they are adding in upcoming releases for doing distributed object messaging will, if used locally, give you the same thing as Obj-C selectors. The the selector is an actual object, like in OPENSTEP, and not just a C data type, as in NEXTSTEP. As it stands, right now, the runtime already uses a string based selector when you go to do native methods and Java's selectors are different from Obj-C in that they encode type as well as name into the string. It looks a lot like an @encode() if you have ever seen one. Anyway, since I'm saying that the full selector features will be in an upcoming Java, and not in today's Java, I should point out that there is no reason why you cannot do selectors today, but you have to do a little bit more work than you will have to do in the future. How? Open up Design Patterns by Gamma, et al. and read about "Command". Quite implementable in Java, and it is functionally equivalent to selectors. I personally like selectors since they are so clean syntactically, but my point here is that Java *can* do these things. It may be a bit more clunky than ObjC, but it is still way better than C++. As far as other "missing" ObjC features, the one I really miss is categories. As far as I can tell, the VM could handle them, just the language doesn't have a way to access that feature. So if you built your own Java compiler, it should be doable. Also, seeing that the VM is this powerful, I personally believe that it is possible to create an ObjC to Java bytecode compiler. Frankly, *that* would be a great thing to have, if anyone has the time to build one. I'd really like to see such a product. Imagine how much nicer it would be if you could use ObjC syntax to write your applets and other Java objects... -- 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 Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 18 Jul 1996 21:37:04 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4smaq0$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > don@misckit.com wrote: > > > Java is a static OO language(like C++ and Simula). > > > This means it can't support Objective-C's runtime in any > > > reasonable way. > > > > Ummm...that's just not true. Java has more in common with > > Objective-C than it does with C++. > > Perphaps you can show us how to use selectors(or -perform:) in Java? I another message I just posted, I mention using the "Command" design pattern from Gamma, et. al. That's how you'd do it right now until the next release (which is actually supposed to have selector objects in it). Not the best solution, since it requires a little more work than the ObjC syntax, but this will work. I'm going to add to the comments in the other post by mentioning that the compiler could easily do the work of creating the necessary Command pattern objects for you automatically. For each selector used in the app, you just have to create the appropriate command-pattern object class. You could have the compiler generate one for each of the app's methods by default . (This is in theory; I know of no such compiler in existence today--but there is NOTHING keeping you from modifying an existing compiler to do this.) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: thrall@serv.net (Dean Johnson) Newsgroups: comp.sys.next.programmer Subject: Docs for NS 4.0 Date: 19 Jul 1996 00:33:57 GMT Organization: ServNet Internet Services Message-ID: <4sml5l$sid@brockman.serv.net> Does anyone know if NeXT is going to re-release their documentation (commercial printed docs) for OPENSTEP 4.0? Thanks, dean
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Docs for NS 4.0 Date: 19 Jul 1996 02:08:44 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4smqnc$26i@news.xmission.com> References: <4sml5l$sid@brockman.serv.net> thrall@serv.net (Dean Johnson) wrote: > Does anyone know if NeXT is going to re-release their documentation > (commercial printed docs) for OPENSTEP 4.0? From what I last heard, it didn't sound likely. More like there was a lack of a publisher, not that NeXT didn't want to. But maybe they have found someone willing to do it...it sure would be nice! Of course this is all unofficial rumor and there may be some docs coming along, but my sources are usually pretty accurate, so don't hold your breath waiting for them. -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 19 Jul 1996 01:30:28 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4smofk$29j@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> Cc: don@misckit.com don@misckit.com wrote: > Steve Dekorte wrote: > > don@misckit.com wrote: > > > > Java is a static OO language(like C++ and Simula). > > > > This means it can't support Objective-C's runtime in any > > > > reasonable way. > > > > > > Ummm...that's just not true. Java has more in common with > > > Objective-C than it does with C++. > > > > Perphaps you can show us how to use selectors(or -perform:) in Java? > > I another message I just posted, I mention using the "Command" > design pattern from Gamma, et. al. That's how you'd do it right now.. Doesthis involve creating a mini Java class for the message you want to send and then having it dynamically compiled and invoke it's method? If so, that's not in the bounds of what I meant by implementing an ObjC-like runtime in a reasonable way. > ..until the next release.. With Sun things might be different, but in general: The features that would actually make a product usefull are allways in "the next release". Steve
From: bungi@gaea.titan.org (Timothy J. Wood) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 18 Jul 1996 20:39:50 -0700 Organization: Omni Development, Inc. Message-ID: <4sn026$ba3@gaea.titan.org> References: <4sk3oq$727@news.onramp.net> <TMB.96Jul18034339@best.best.com> <4slob0$rmv@muller.loria.fr> colnet@loria.fr (Dominique Colnet) writes: >In article <TMB.96Jul18034339@best.best.com>, tmb@best.com (--) writes: >|> In behavioristic prototype-based laguages(Self, Obliq, etc) your >|> not even dependent on our own data-structure. In fact, objects >|> have no static data-structures. All variables are accessed by >|> messages. You would think this would cause poor performance >|> >|> And one would be right: accessing instance variables through messages >|> _is_ slow in any language. >Sorry, but the Self language use a polymorphic type inference algorithm. >Accessing instance variables can be automatically inlined. >Have a look at the (excellent) Phd Thesis of Ole AGESEN >"Concrete Type Inference: Delivering Object Oriented Applications" >Published by Sun Microsystems (SMLI TR-96-52). Also, check out Craig Chambers' PhD thesis on the Self compiler. It contains actual benchmark data comparing Self to other languages (rather than wild speculation). You should be able to find this either on the Self home page or at his home page at the University of Washingting (somewhere under http://www.cs.washington.edu/). -- Timothy J. Wood Omni Development, Inc. http://www.omnigroup.com/
From: jeremy@inetnebr.com (Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Re: Config.app Problem on Black Hardware. Date: 19 Jul 1996 03:30:47 GMT Organization: Internet Nebraska Message-ID: <4smvh7$8eh@duck.inetnebr.com> References: <ampriasm-1707960100480001@f180-175.net.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-User: jeremy To: ampriasm@students.wisc.edu (Andrew M. Priasmoro) In article <ampriasm-1707960100480001@f180-175.net.wisc.edu>, ampriasm@students.wisc.edu (Andrew M. Priasmoro) writes: > Could anyone tell me what to do if my Config.app icon is in question mark? > When I tried to execute it, It said that the application is damage. Is there > any way to fix this without having to reinstall the OS? Thanks. There is nothing to configure, Config.app is only for i386 hardware. It shouldn't even be installed on a m68k system. (Unless you told Installer to install both m68k and i386 bins.) -- Jeremy Bettis -*- PGP public key available,send mail with subj "Send pgp key" Home: jeremy@tddi.inetnebr.com Work: jeremy@hksys.com NeXTMAIL/MIME/PGP accepted
From: lozinski@cup.portal.com (Christopher A Lozinski) Newsgroups: comp.sys.next.marketplace,comp.sys.next.programmer Subject: Nextstep Job Market Repost Date: 18 Jul 1996 22:00:16 -0700 Organization: The Portal System (TM) Sender: pccop@unix.portal.com Distribution: world Message-ID: <152778@cup.portal.com> The NeXTSTEP job market is strong and getting stronger. It includes hardware and networking people, system administrators, developers, architects, and testers. Demand for developers is excessively strong in the Maryland, Washington, D.C., North Virginia area. It is also very strong in Texas. There is slow but steady demand in the rest of the big cities: e.g., San Francisco, Los Angeles, Seattle, Chicago, and New York. Away from the oceans there are a few other cities hiring. I know two very rural communities that are also looking for NeXTSTEP Developers. WebObjects demand is coming on strong. Really this means NeXTSTEP developers who have done a simple app on WebObjects. Telephony expertise is in strong demand. One company is willing to train in Telephony. Financial expertise is always in strong demand. One company has an 8-week training program in the financial industry for selected candidates. One Computer Systems Integrator and Internet Service Provider is willing to train developers in his industry. Computer Graphics expertise is in very strong demand all along the West Coast. One hiring company is even building Nuclear Power Plant Decommissioning Software. Two companies will let you trade your NeXTSTEP expertise for a good salary and stock. These companies are liable to go public within the year. Neither one is in California. a) WASHINGTON, D.C. METRO AREA: This area, which includes Washington, D.C., Maryland, and Western Virginia, has some of the best paying consulting positions. If you want to earn a lot of money and do not care too much what you are doing, these positions are for you. b) SEATTLE: Build Multimedia Web Applications in Seattle. You will be the primary developer on a team of multimedia experts, with the potential for very rapid group growth in this hot area. Take the initiative in figuring out what new services this company could/ should offer their customers, then go implement them. These are uncharted waters to be explored by someone with curiosity. Great for someone with an interest in cooking, gardening, and enjoying life. These guys work hard, play hard, and generally behave like a Silicon Valley Start Up, but with a sales growth curve of 100% per year. c) PITTSBURGH: This is a team of people building very complex medical radiation equipment. The employees are the design experts for the product and are treated as such. It is not worth their time to take any less than the best, and that means that you are respected and given considerable freedom and responsibility. This is not yet another database application. Rather, it is a complex scientific application with three-dimensional modelling and simulated annealing. They offer a highly competitive compensation package that includes health insurance and stock options. d) SECURITY CLEARANCES: People with TS SI/TK ISA ISSA SBI or EDI security clearances are desperately needed in the Washington, D.C. Area. You, too, can spy on foreigners. e) SAN FRANCISCO MARKETING: If you are not a developer, there is still opportunity for you. NeXTSTEP marketing professionals are needed in the San Francisco Bay Area. The company is making a major investment in NeXTSTEP technology and needs some people to help define the market requirements and help sell the product. f) HOLLAND, Sybase Administrator: We still have not found a person who can single handedly keep a production database up and running and optimize it. If we found such a person, we would take care of any visa issues, and pay such a mission-critical person well. If you do your job well, then in your free time you can develop some NeXTSTEP applications. The other interesting question is how to migrate such a production system to NT without disrupting production. g) SAN FRANCISCO BAY AREA: Become the corporate NeXTSTEP Fountainhead. Live just outside the Bay Area, close to the ocean. Wonderful location. Learn how to surf. h) SAN FRANCISCO BAY AREA: These guys are willing and able to pay the best salaries for the best people, so if you get the position, you will be working with the best possible team. Go for it. i) CHICAGO There is a strong demand for people with 6 months nextstep experience in the Chicago area. They will even train you in the financial industry. j) AUSTRALIA, Object-Oriented Database: I am very embarrassed to not yet have found a person with NeXTSTEP and object-oriented database experience for Australia. I guess most people have not visited Australia, or they would be much more eager to go. We were there just last year; give me a call if you would like me to tell you more about the country. (510) 795-6086. k) MIDWEST: Auspex System Administrator needed. Very low cost of living. Four beautiful seasons, and a very important and elegant application that needs support. l) LIFETIME EMPLOYMENT: Los Angeles Company is looking for a person who would like lifetime employment. They offer long-term job stability, and training. You get to take an existing application developed by a large consulting firm and initially maintain it and later upgrade it. m) NORTH CAROLINA: The hottest job market now is for the people building the Internet. Get out out of the corporate doldrums and into Internet building. Use your NeXTSTEP experience to land a position where you learn to be a system's integrator. You can settle down in the lovely foothills of North Carolina. Dress casually. The company is a strong NeXT advocate. You might even get to start a WebObjects consulting practice. n) LOS ANGELES: Needs graphic artists / animators with NeXTSTEP experience. Help create the hottest Hollywood blockbusters, and many of the T.V. animation movies. o) CHICAGO, Consulting Firm: Is looking for a project manager to supervise 30 people. It is a fun company filled with Smalltalk and Objective-C experts. Fun place to work, and it is a very senior position. p) MINNEAPOLIS ST. PAUL: Learn all about building the Internet. This company is building order entry systems for tracking sales of their digital services (Frame Relay, ISDN, SMDS, etc.). They are using an Oracle database to house the data, and NeXTSTEP/EOF to provide custom interfaces for the data. Duties include OOA/OOD, development, and some testing. There is both a developer and a project manager position available. q) COLORADO JOBS: For the first time, I have job openings in Colorado. This is a great place to live, to live it up if you are single, or to bring up a family if you are married. I worked there for a while. Wonderful mountains, very cheap living. Many people would do anything to find a job there, any job, just so that they could live there. You even get to do NeXTSTEP development there. r) TEXAS JOBS: This company is using NeXTSTEP to develop a Commodity Trading System capable of trading any commodity and is the only system capable of handling physical deliveries. They have found a market in the energy industry and are currently rolling out the product in two nations. They develop and customize the product for each customer, then install and support the product. Their coolest benefit is that they let you call anywhere in the world for free. Since they are willing to take foreign citizens, this works out great. I am very impressed by their project management. Every time that I call their most senior manager, he picks up his phone himself. We get our business done, and I get back to work. He must have a very smoothly running organization to be able to do that. s) WORK IN THE CITY OF YOUR CHOICE: I am now able to send your resume directly to hiring managers in the city of your choice. Whereas you may not know which companies use NeXTSTEP developers in a given city, or you may not be able to get past the HR department, I can send your resume directly to the in-box of the hiring managers of the companies you are interested in, in the cities where you would like to work. I know 104 hiring managers. If you are interested, send e-mail to <jobs@bpg.com>. Let me know what type of position you are looking for and what city you would like to work in. Even if I do not know a hiring manager in the city of your choice, I often know some NeXT developers working there, and I can forward your resume through them to the appropriate hiring manager. If you are interested in any of these jobs, please give me a call at (510) 795-6086 or better yet, send me your resume in NeXTMAIL RTF format to < jobs@bpg.com>
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: Reverse ARP under NeXTStep 3.x? Date: 19 Jul 1996 08:39:11 GMT Organization: Tech Net GmbH Message-ID: <4snhjf$njq@ddfservb.technet.net> References: <4sjaoi$r1j@teal.csn.net> bediger@csn.net (Bruce Ediger) wrote: >For very peculiar reasons, I'd like to net-boot a Sun SPARCstation with >a NeXT 25MHz slab as a boot server. I can find the tftp stuff to >serve the boot and kernel files with, but I think I need a Reverse >ARP daemon (rarpd) for the SPARCStation to get its IP address from. > >Does such a beast exist for NeXTStep? I found 4.4BSD source code for >a rarpd, but it uses "Berkeley Packet Filter" (bpf) to go beneath the >raw IP layer. Therefore, that source is pretty useless for my purposes. >What's the NeXTStep analog of "bpf" or SunOS's "nit"? Is there some >way to listen for RARP packets under NeXTStep? > >Regards, >Bruce Ediger Check out man bootpd. The boot protocol daemon gives your SPARCstation it's IP address. Make sure you enter the correct hardware (ethernet) address in the configuration file. Note that the bootpd supplied with next only conforms to RFC 951 and NOT the newer protocol extensions (forgot the RFC number). If your station needs this extension, you have to get a new bootpd form somewhere. -- Constantin Szallies, Energotec GmbH szallies@energotec.de 49211-9144012
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: Reverse ARP under NeXTStep 3.x? Date: 19 Jul 1996 08:59:16 GMT Organization: Tech Net GmbH Message-ID: <4snip4$njq@ddfservb.technet.net> References: <4sjaoi$r1j@teal.csn.net> <4snhjf$njq@ddfservb.technet.net> Constantin Szallies <szallies@energotec.de> wrote: >Check out man bootpd. The boot protocol daemon gives your SPARCstation it's >IP address. Make sure you enter the correct hardware (ethernet) address in >the configuration file. Note that the bootpd supplied with next only conforms >to RFC 951 and NOT the newer protocol extensions (forgot the RFC number). If >your station needs this extension, you have to get a new bootpd form >somewhere. I forgot to mention two points: 1) Tell HostManager.app the IP and ethernet address. A correct ethernet address is mandatory. 2) On your NeXT machine in /etc/hostconfig, put NETMASTER to -YES-. Reboot the NeXT machine and check that there's a daemon process "bootpd" -- Constantin Szallies, Energotec GmbH szallies@energotec.de 49211-9144012
From: tmb@best.com (--) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 19 Jul 1996 10:19:25 GMT Organization: home Message-ID: <TMB.96Jul19031925@best.best.com> References: <4sk3oq$727@news.onramp.net> <TMB.96Jul18034339@best.best.com> <4slob0$rmv@muller.loria.fr> <4sn026$ba3@gaea.titan.org> In-reply-to: bungi@gaea.titan.org's message of 18 Jul 1996 20:39:50 -0700 FCC: /home/tmb/mail/x-nout In article <4sn026$ba3@gaea.titan.org> bungi@gaea.titan.org (Timothy J. Wood) writes: >In article <TMB.96Jul18034339@best.best.com>, tmb@best.com (--) writes: >|> In behavioristic prototype-based laguages(Self, Obliq, etc) your >|> not even dependent on our own data-structure. In fact, objects >|> have no static data-structures. All variables are accessed by >|> messages. You would think this would cause poor performance >|> >|> And one would be right: accessing instance variables through messages >|> _is_ slow in any language. > >Sorry, but the Self language use a polymorphic type inference algorithm. >Accessing instance variables can be automatically inlined. >Have a look at the (excellent) Phd Thesis of Ole AGESEN >"Concrete Type Inference: Delivering Object Oriented Applications" >Published by Sun Microsystems (SMLI TR-96-52). Also, check out Craig Chambers' PhD thesis on the Self compiler. It contains actual benchmark data comparing Self to other languages (rather than wild speculation). You should be able to find this either on the Self home page or at his home page at the University of Washingting (somewhere under http://www.cs.washington.edu/). When I'm saying "accessing variables through messages is slow", I'm talking about those situations where the program relies in an essential way on the possibility of getting multiple different object representations. When the compiler can determine statically that this isn't the case, you aren't really using messages, you are simply using the notation of messaging to express the accessing of instance variables. That does save you the trouble of having to make a lot of changes to your program if you decide to go from representing some attribute as an instance variable to representing it as a method. On the other hand, it makes it much more difficult for programmers and maintainers to figure out how costly some operation is. Which approach is better depends on your application; just having excellent compilers for highly dynamic languages is not enough to make those languages suitable for applications where speed matters. Thomas.
From: colnet@loria.fr (Dominique Colnet) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 19 Jul 1996 15:30:52 GMT Organization: CRIN & INRIA-Lorraine - Nancy - FRANCE Message-ID: <4so9nc$nuh@muller.loria.fr> References: <4sk3oq$727@news.onramp.net> <TMB.96Jul18034339@best.best.com> <TMB.96Jul19031925@best.best.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit In article <TMB.96Jul19031925@best.best.com>, tmb@best.com (--) writes: |> |> When I'm saying "accessing variables through messages is slow", I'm |> talking about those situations where the program relies in an |> essential way on the possibility of getting multiple different object |> representations. When the compiler can determine statically that this The goal of Type Inference is to remove automatically the dynamic dispatch. This is also tru for attribute (read/write). |> isn't the case, you aren't really using messages, you are simply using |> the notation of messaging to express the accessing of instance |> variables. |> |> That does save you the trouble of having to make a lot of changes to |> your program if you decide to go from representing some attribute as |> an instance variable to representing it as a method. On the other |> hand, it makes it much more difficult for programmers and maintainers |> to figure out how costly some operation is. Which approach is better Complexity of the corresponding algorithm is NOT change. |> depends on your application; just having excellent compilers for |> highly dynamic languages is not enough to make those languages |> suitable for applications where speed matters. Speed is only a question complexity. -- ------------------------------------------------------------ Dominique COLNET -- Talk Eiffel with SmallEiffel Talk Eiffel C.R.I.N. (Centre de Recherche en Informatique de Nancy) POST: CRIN,BP 239,54506 Vandoeuvre les Nancy Cedex,FRANCE EMAIL: colnet@loria.fr VOICE:+33 83593079 FAX:+33 83413079
From: wilkie@cg.tuwien.ac.at (Alexander Wilkie) Newsgroups: comp.graphics.rendering.renderman,comp.sys.next.programmer Subject: RiMakeTexture - how 2? Date: 19 Jul 1996 16:25:15 GMT Organization: Vienna University of Technology, Austria Message-ID: <4soctb$i3u@news.tuwien.ac.at> Hi newsgroups! I have tried to get the RenderMan function RtToken RiMakeTexture(RtToken image, char *map, RtToken sWrap, RtToken tWrap, RtFloatFunc filterFunction, RtFloat sWidth, RtFloat tWidth, ...); (excerpt from <ri.h>) to work on a NeXT, with no luck so far. I'd like to do this from within an application so that it can display textures that are computed at run time properly. It saves them to disk as TIFFs, which then ought to be processed by this call for later use by RiSurface(). What happens is... , well nothing. No error messages, no crashes, but no textures either. Odd. JFTR: I have made sure that this function is only called when the photorealistic version of RenderMan (as opposed to QRM) is active. Why does this function call differ from the one shown in the RM companion? I assumed that "image" refers to the source file name (why "RtToken"?) and "map" to the output file name, but I might be totally wrong here. Could anyone shed some light on this? Please e-mail me, I'll summarize. mtia Alexander Wilkie -- e-mail: wilkie@cg.tuwien.ac.at (NeXTMail preferred, MIME o.k.) www : http://www.cg.tuwien.ac.at/~wilkie/
From: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: Blinking Text Date: 19 Jul 1996 18:03:30 GMT Organization: Universite de La Rochelle Message-ID: <4soili$n9h@hpuniv.univ-lr.fr> Hello, Do you know a way to have blinking text ? Thanks for any reply. Wil. wgaboria@iut-lr.univ-lr.fr
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: GDB in OpenStep Date: 19 Jul 1996 18:39:59 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4sokpv$ikh@news.onramp.net> Does the commercial release of OpenStep have a working gdb? Steve
From: brown@bibliotech.com (Robert E. Brown) Newsgroups: comp.sys.next.programmer Subject: FP rounding modes Date: 19 Jul 1996 16:27:51 -0400 Organization: Bibliotech, Inc. Sender: brown@grettir.bibliotech.com Message-ID: <87vifk2dko.fsf@grettir.bibliotech.com> I want to set the FP rounding modes on an Intel box running NextStep. Essentially, I want the fpsetround() and fpgetround() functions of Solaris. If someone out there has code to do this, please mail me a copy. Thanks! bob
From: "Eric A. Dubiel" <ced016@email.mot.com> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Netatalk- AppleTalk for NeXT port?! Date: Fri, 19 Jul 1996 15:01:36 -0600 Organization: FPDC of Motorola, Schaumburg,IL,USA Message-ID: <31EFF7B0.5B45@email.mot.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit <a href="http://www.umich.edu/~rsug/netatalk/">http://www.umich .edu/~rsug/netatalk/</a> This package- Netatalk is an AppleTalk for UNIX implementation, like CAP (Columbia AppleTalk Protocol) perhaps someone who has the skills would be interested in completing the port to NeXT. It HAS been ported to MkLinux (Linux on MACH 3 for PowerPC Macs) There's info <a href="http://mklinux.apple.com/wip/developers/netatalk.html" >http://mklinux.apple.com/wip/developers/netatalk.html and a link at: <a href="ftp://mklinux.apple.com/pub/MkLinux_DR1/Updates/960612 .mklinux">ftp://mklinux.apple.com/pub/MkLinux_DR1/Updates/96 0612.mklinux</a> -Eric > From: "Eric A. Dubiel" <ced016@email.mot.com> > To: netatalk@umich.edu > On these web pages, it is > mentioned that a NeXT port was started, however uncompleted. I just recently talked to the person that was attempting this port (he lives in Texas, now). He said he didn't get very far, and hadn't saved his work. :wes
From: Garance A Drosehn <gad@eclipse.its.rpi.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 19 Jul 1996 08:40:07 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Message-ID: <4snhl7$g5f@usenet.rpi.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > don@misckit.com wrote: > > ..until the next release.. > > With Sun things might be different, but in general: > The features that would actually make a product usefull > are allways in "the next release". I can not remember any time when I have ever thought that about C++. If we're dealing in truisms here, it's also always true that version 1.0 of something is not really the final product. Java is at version 1.0x. It's certain that it's not the final edition. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer (MIME & NeXTmail capable) Rensselaer Polytechnic Institute; Troy NY USA
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: Help with NSArchiver Date: 20 Jul 1996 00:09:10 GMT Organization: NeXT Software, Inc. Distribution: world Message-ID: <4sp836$4v9@news.next.com> References: <4sh537$2n20@news.doit.wisc.edu> In article <4sh537$2n20@news.doit.wisc.edu> giddings@menominee.menominee.chem.wisc.edu (Michael Giddings) writes: > I have an app in which I'd like to implement archiving using > NSArchiver. It has one main "controller" object which is the > target of all of the menu commands, including "Save". This > controller object stores the id's of two NSObject subclasses > that need to be archived. > > The problem is, given the FoundationKit docs, I can't figure > out how to do it cleanly. It seems the only method available > is archiveRootObject:toFile:, which appears to be designed > only for archiving one single root object (not two, such as > I need). The docs don't say what happens if this method is > called multiple times with the same file name but different > objects, but I assume it is overwritten on the second call > which would cause the first to be lost (if it's not, then > what happens when you need to overwrite a new version of the > document being saved?). [...] See the NSArchiver methods: -initForWritingWithMutableData: -archiverData -encodeRootObject: (or the inherited -encodeObject:) and the NSData method: -writeToFile:atomically: +archiveRootObject:toFile: is just a convenience that wraps up a few of these operations for you, then saves the resulting NSData to the specified file. Don't forget the NSCoder methods that NSArchiver inherits, some of them may be handy. You want to initForWritingWithMutableData: a new NSArchiver instance, send encodeObject: to it multiple times, once for each object you wish to archive, get the archiverData out of the NSArchiver, and writeToFile:atomically: that data object. To unarchive the objects once you've created an NSUnarchiver with the data, simply call -decodeObject the appropriate number of times. Depending on how much flexibility you want, you may want to archive an integer at the start of your archive which indicates the number of objects encoded. Then on unarchival, you can first read that int and loop and appropriate number of times. And yes, calling +archiveRootObject:toFile: a multiple times will overwrite the previous file. Christopher Kane Foundation Technology
From: Samuel_Figueroa@NeXT.com (Samuel Figueroa) Newsgroups: comp.sys.next.programmer Subject: Re: gcc-2.7.2 on NS3.3-Sparcstation 5? Date: 20 Jul 1996 00:00:30 GMT Organization: NeXT Software, Inc. Message-ID: <4sp7iu$4v3@news.next.com> References: <4rls6f$85j@duck.inetnebr.com> Jeremy Bettis writes > In article <scratch.836512886@sunset.sce.carleton.ca>, > scratch@sunset.sce.carleton.ca (Craig Scratchley) writes: > > diego@conga.super.unam.mx (Diego Zamboni) writes: > > > >>I wanted to compile gcc-2.7.2 on my Sparcstation 5 running NS3.3, both to > >>have a recent version of gcc and to be able to compile g77, the GNU FORTRAN > >>compiler. However, apparently gcc doesn't yet support NEXTSTEP running on > >>sparc processors. > > > > Perhaps Openstep/Mach 4.0 provides a recent-enough version of > > gcc. Does anybody know? > > > The prerelease does not. I suspect that the final release won't either. > -- > Jeremy Bettis -*- PGP public key available,send mail with subj "Send pgp key" > Home: jeremy@tddi.inetnebr.com Work: jeremy@hksys.com > NeXTMAIL/MIME/PGP accepted If you have access to the Windows NT PR3, it might have the files you need, but I cannot make any promises, or help you to make it work. -- - Sam Figueroa
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 01:48:06 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4spdsm$q15@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> Cc: gad@eclipse.its.rpi.edu In <4snhl7$g5f@usenet.rpi.edu> Garance A Drosehn wrote: > dekorte@suite.com (Steve Dekorte) wrote: > > don@misckit.com wrote: > > > ..until the next release.. > > > > With Sun things might be different, but in general: > > The features that would actually make a product usefull > > are allways in "the next release". > > I can not remember any time when I have ever thought that > about C++. Yeah, AT&T C++ never tried to be anything it wasn't. It's the people that using that do that try to use it to do dynamic things and end up with a mess. (ie. Taligent) Steve -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdi
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 02:29:13 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4spg9q$q15@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> <4spdsm$q15@news.onramp.net> Cc: dekorte@suite.com Steve Dekorte wrote: > Garance A Drosehn wrote: > > Steve Dekorte wrote: > > > don@misckit.com wrote: > > > > ..until the next release.. > > > > > > With Sun things might be different, but in general: > > > The features that would actually make a product usefull > > > are allways in "the next release". > > > > I can not remember any time when I have ever thought that > > about C++. It's the people trying to use C++ to do dynamic things and end up with a mess (ie. Taligent) and make release promises. Steve -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdi
From: Alex Duong Nghiem <alexdn@globalobjects.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.marketplace Subject: JOB: NextStep developers needed immediately Date: Fri, 19 Jul 1996 23:12:26 -0400 Organization: Global Objects Inc. Message-ID: <31F04E9A.B42@globalobjects.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: jobs@globalobjects.com Global Objects specializes in providing quality solutions to our clients through object technology. We are currently looking for several junior and senior NextStep developers. EOF and/or WebObjects experience are a plus. For immediate consideration, please e-mail your resume to jobs@globalobjects.com (no NextMail please!) or fax it to (770) 457-7333. Thanks, - Nicki -
From: jobs@globalobjects.com (Nicki N.) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc,comp.sys.next.marketplace Subject: JOB: NextStep developers needed immediately! Date: Sat, 20 Jul 1996 03:01:16 GMT Organization: Global Objects Inc. Message-ID: <4sphvk$38hu@mule2.mindspring.com> Global Objects specializes in providing quality solutions to our clients through object technology. We are currently looking for 7 junior and senior NextStep developers. EOF and/or WebObjects experience are a plus. For immediate consideration, please e-mail your resume to jobs@globalobjects.com or fax it to (770) 457-7333. Thanks, - Nicki -
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: Reverse ARP under NeXTStep 3.x? Message-ID: <DurB7J.1F7@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4skgb2$bvp@hermes.is.co.za> Date: Thu, 18 Jul 1996 20:50:07 GMT In article <4skgb2$bvp@hermes.is.co.za> gvandyk@icon.co.za writes: > On 07/17/96, Bruce Ediger wrote: >>Does such a beast exist for NeXTStep? I found 4.4BSD source code >>for a rarpd, but it uses "Berkeley Packet Filter" (bpf) to go >>beneath the raw IP layer. Therefore, that source is pretty useless >>for my purposes. >>What's the NeXTStep analog of "bpf" or SunOS's "nit"? Is there >>some way to listen for RARP packets under NeXTStep? > I don't know of a RARP for NeXTStep but if you want, you can have a > look at the latest PPP on peak, there as a BPF LKS in there. It might > help with the BSD 4.4 version you have. BFP in PPP package can be only used to access PPP interface. At least its readme says so. Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 09:01:30 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4sq79a$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> <4spdsm$q15@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > In <4snhl7$g5f@usenet.rpi.edu> Garance A Drosehn wrote: > > dekorte@suite.com (Steve Dekorte) wrote: > > > don@misckit.com wrote: > > > I another message I just posted, I mention using the "Command" > > > design pattern from Gamma, et. al. That's how you'd do it right now.. > > Does this involve creating a mini Java class for the message you > > want to send and then having it dynamically compiled and invoke > > it's method? If so, that's not in the bounds of what I meant by > > implementing an ObjC-like runtime in a reasonable way. I just posted an in-depth explanation of what I'm proposing in the Obj-C group, where I'm sure you've seen it. Do also note Dave Griffith's post: his way is different, but emphasizes that this can be done without the programmer having to know how it is done. This makes me ask: why is that solution unacceptable to you, especially if it can be done in a way that is transparent to the programmer? Is that not the beauty of Obj-C's selector mechanism: dynamic binding of message calls done in a way that is transparent to the programmer? Obviously, with either runtime, you get that...just that the current Obj-C implementation is more efficient. Assuming Sun comes through with announced plans, that may not be true in the long haul... At any rate, right _now_ I still prefer Obj-C by a long shot. My only real interest in this is to empahasize that this _can_ be done, even if you wouldn't necessarily want to mess with the details themselves. > > > don@misckit.com wrote: > > > > ..until the next release.. > > > > > > With Sun things might be different, but in general: > > > The features that would actually make a product usefull > > > are always in "the next release". True enough. I believe that Sun will come through on this one. But I mention the other approach to show that we can do this without Sun's help if we truly care to. So it doesn't matter if Sun comes through or not, as far as my argument is concerned. It would just be more convenient for the programmer if they did. :-) Also, even though the full Command pattern is a long way to go to implement such a simple thing, technically, you COULD implement this in Java yourself by playing with the Java code for their root class and making your own native methods. (Frankly, you could do it in a subclass, too, without customizing the standard Java libs, but it wouldn't integrate into the environment as nicely. This is because Java lacks categories. I'd love to be able to use categories in Java!) Since the Java runtime has all the information about methods, the native method just has to access that already existing information and construct the method calls--just like the ObjC run time would do. Not trivial to write, but it is possible to do. The language may not support what we want (directly) but the runtime does. All that is needed is a mechanism to bridge the gap. > > I can not remember any time when I have ever thought that > > about C++. > > Yeah, AT&T C++ never tried to be anything it wasn't. > It's the people that using that do that try to use it to do dynamic > things and end up with a mess. (ie. Taligent) And I'll give no argument that C++ is awful. In order for anyone to be successful in trying to do what Taligent wanted to do, they would have to build their own runtime to layer on top of C++, since it completely lacks a runtime. The compiler and linker throw away most of the information that a runtime would need, too, which further complicates matters. Most of these attempts either fail to get the runtime right, or they miss the boat entirely and don't do the runtime. Those that try the runtime take longer before they fail, too, from what I've seen. I can think of at least three here in Utah alone that tried to do NEXTSTEPpy things with C++ and have failed or aren't doing all too well. In all cases, the runtime, if they built one, makes their C++ framework slower than what NeXT has built on top of Obj-C. And the frameworks aren't as good. For all the complaints we can make about NEXTSTEP/OPENSTEP, the fact remains: it may not be perfect, but it is d*mn good! (And if I rant any more, I'll have to move to .advocacy. But since they're ranting about Germany, freedom, and human rights right now, I'll stay here where NeXT product's can be discussed, thank you very much.) That is why I posit that Java is closer to Obj-C than it is to C++. Java _has_ a runtime. That's more than half the battle. It may have the syntax of C++, which reminds us of the "sitting on my hands" feeling that C++ gives you after having used a truly dynamic language. But it is a relief to realize that you can do the dynamic things in Java--so you aren't sitting on your hands and it isn't so bad after all. One other observation: people who program Java and came from the C++ world only really don't know how to use the language to its full potential. Those of use who come from ObjC, Smalltalk, Lisp, or any other dynamic OOP language really do have an advantage in that world. Many of the Java tutorials and books I've read give examples that show some of the most _awful_ OOD I've ever been subjected to! I suspect that many people coming into Java as their first OOP language will end up developing some really bad habits from the beginning. That is really sad, IMHO. Anyone read any _good_ books lately? :-) Well, that was longer than I meant it to be... but every once in a while I guess it is nice to vent. :-) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 20 Jul 1996 15:34:58 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4squb2$94d@solaris.cc.vt.edu> References: <4sk3oq$727@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > In behavioristic prototype-based laguages(Self, Obliq, etc) your > not even dependent on our own data-structure. In fact, objects > have no static data-structures. All variables are accessed by > messages. You would think this would cause poor performance but > Self has compiler technology that allows it to do messaging about > as fast as Objective-C. I've recently read papers on the performance of Self by both the people working on the project and people outside the project. I seem to remember numbers like 20 - 60% of the speed of optimized C code with under 50% being typical. For the paper that was independent of the Self project I believe it was under 30% the speed of the C++ version of their code. I would consider these numbers to be poor performance for the projects that I work on. I believe that Objective-C runs *much faster* than that. On the other hand, Self does offer garbage collection and a more uniform object model. Also, the paper where Self did 30% the speed of C++ used a very small application that really didn't flex the advantages of object-oriented code in a large project. The Self project has also offered runtime optimization techniques that I hadn't seen before and can be used in other languages. Maybe you can help answer a question, Steve: One thing I could never find in the Self papers that I have read is how arrays are implemented. The papers always harp on how you send a message "something" to an object which is looked up directly in the object. If not found, then you begin searching through the ancestors. Do arrays have messages "0", "1", and "2" that are hashed and read or what? How do they work? Self needs: * portability (why wasn't this done from the beginning?) * newsgroup (couldn't find one) * book (maybe there's one out there) -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 16:03:51 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4sr017$94d@solaris.cc.vt.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> aisbell@ix.netcom.com (Art Isbell) wrote: > dekorte@suite.com (Steve Dekorte) wrote: > > don@misckit.com wrote: > > > > Java is a static OO language(like C++ and Simula). > > > > This means it can't support Objective-C's runtime in any reasonable > way. > > > > Ummm...that's just not true. Java has more in common with > > > Objective-C than it does with C++. > > > Perphaps you can show us how to use selectors(or -perform:) in Java? > > Perhaps you can show us how to use automatic garbage collection in > Objective-C. > > My point wasn't to start a religious war over which language is "best". > But rather to make a more practical observation that no matter how nice I don't think Steve mentioned anything about religion. He's making a good point which I agree with. The fact is that OpenStep relies on things like dynamic binding, categories and distributed objects. That stuff just isn't found in Java right now. If you thought porting from Common/AppKit (NS3.3) to Foundation/NSAppKit (NS4.0) was a pain, wait until you take a 100,000 line project and "uncategorize" it and "command-ize" (ala design patterns) to make it work. Ack. I'd much rather see an improved Obj-C (cleanup the C part to make the language safer and simpler, throw in garbage collection, throw in some features from tom (http://tom.es.ele.tue.nl:8080/)) and then a blitzkrieg to make ObjC the development language of the next decade. (You know, a free-be compiler with source, a bytecode generator, ANSI standard, etc.) C++ programmers cry out for safety, gc, dynamic binding and simplicity. ObjC programmers cry out for a few improvements and industry acceptance. Java programmers cry out ObjC features. Even those programmers who aren't crying would benefit from these things. (Usually they're not crying because they don't know about and/or haven't experienced these other features.) It's too bad NeXT doesn't have that "we're going to change the world" attitude anymore. I think they could make a kick ass Obj-C2 and change it. -Chuck
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 20 Jul 1996 16:19:48 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4sr0v4$94d@solaris.cc.vt.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > > [obj initWithSize:aSize withColor:aColor withFoo:aFoo withBar:aBar > withBaz:aBaz withBop:aBop]; > > obj.init(aSize, aColor, aFoo, aBar, aBaz, aBop); > > Call me crazy, but I prefer the second version. It's a lot more > readable; it's shorter; and, if you use descriptive variable names in > your code, it's as self-documenting as it needs to be. I've never seen a method named "initWithSize:withColor:withFoo:withBar:withBaz:withBop:". Maybe this style is in 4.0. I haven't used it much. The programmers I know would call it "initWithSize:color:foo:bar:baz:bop:" which is much shorter and more readable. i.e., you only use one "with". -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 20 Jul 1996 16:23:13 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4sr15h$94d@solaris.cc.vt.edu> References: <4selec$fav@news.onramp.net> <960716095528.4434AADnG.edwards@vege> <4sgl6m$imp@news.onramp.net> <960717112237.5053AADnF.edwards@vege> edwards@sewp.nasa.gov wrote: > > > Could you provide some specific examples? > > > > - (void)setStringValue:(NSString *)aString; > > > > Guess what happens when I try to pass in another object with similiar > behaivior? > > It won't compile. > > > > Steve > > Assuming this is a method on one of NeXT's classes, how would you know what > behavior was required of the argument you are passing in? Since you can't > assume a priori which particular messages will be sent to your NSString- > impersonator, you would have to provide every method that NSString provides. > And if you do that, you might as well make your class a subclass of NSString > and override some or all of its methods. And if you do that, then it would > be perfectly legal to pass in an instance of your class here. > Maybe a better solution would be to use classes as protocols: @interface MyString : NSObject <NSString> { } So MyString must implement every method found in NSString or I get a warning or error. When (NSString *) is expected for a parameter, then an object of any class that conforms to NSString can be passed. Now people always say "I just proposed a change to ObjC that would be good and doesn't have far reaching implications that interefere with other features and current code." I'm afraid to do it, but I'm going to make that statement as well. I'd like to know if I'm right or not. So please, ... prove me wrong! I dare you. -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) Newsgroups: comp.sys.next.programmer Subject: perform: with: afterDelay: cancelPrevious: Date: 20 Jul 1996 17:23:31 GMT Organization: Universite de La Rochelle Message-ID: <4sr4mj$c9e@hpuniv.univ-lr.fr> Hello, Has someone implemented the method "- perform:(SEL)aSelector with:anArg afterDelay:(int)ms cancelPrevious:(BOOL)flag;" for the NSObject class ? wgaboria@iut-lr.univ-lr.fr
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Sat, 20 Jul 1996 13:55:55 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> In-Reply-To: <4sr017$94d@solaris.cc.vt.edu> Excerpts from netnews.comp.sys.next.programmer: 20-Jul-96 Re: Java Bashing (Was: Open.. by Chuck_Esterbrook@orcacom > The fact is that OpenStep relies on things like dynamic binding, categories > and distributed objects. That stuff just isn't found in Java right now. That's not very true. I agree that Java has nothing comparible to Obj-C categories. However, Java certainly has dynamic binding-- consult java.lang.Runtime and java.lang.ClassLoader. I agree that Java's dynamic binding is not as complete or as powerful as Obj-C's because methods and selectors are not first-class objects "right now". But that is something that Sun is actively working to change. Futhermore, Sun has API's and an 'early access implementation' for IDL (Interface Definition Language) and RMI (remote method invocation) available "right now". Consult http://java.sun.com/products/apiOverview.html#enterprise. > If you thought porting from Common/AppKit (NS3.3) to Foundation/NSAppKit > (NS4.0) was a pain, wait until you take a 100,000 line project and > "uncategorize" it and "command-ize" (ala design patterns) to make it work. > Ack. Sure, but then you are talking about changes to the Obj-C APIs that NeXT provides-- which have changed enough from NS 3.3 to NS 4.0 that you pretty much have to port (not just recompile) your old Obj-C code under 4.0. If you're talking about rewriting an Obj-C project in Java, you already have to rewrite all of the source code, anyway. > I'd much rather see an improved Obj-C (cleanup the C part to make the > language safer and simpler, throw in garbage collection, throw in some > features from tom (http://tom.es.ele.tue.nl:8080/)) and then a blitzkrieg to > make ObjC the development language of the next decade. (You know, a free-be > compiler with source, a bytecode generator, ANSI standard, etc.) Doesn't gcc from the Free Software Foundation qualify as a freely available Obj-C compiler that comes with source? Isn't that compiler fully-upwards compatible with ANSI-C? As for a bytecode generator, I'm not sure what you're asking for. If you'd like to see an Obj-C interpreter, then I agree with you that it would be a great thing. But the object modules (.o files) generated by a compiler are machine specific. You don't really want to define an Obj-C virtual machine and an Obj-C bytecode language spec (comparible to Java's VM and bytecode spec, etc), do you? > C++ programmers cry out for safety, gc, dynamic binding and simplicity. ObjC > programmers cry out for a few improvements and industry acceptance. Java > programmers cry out ObjC features. Even those programmers who aren't crying > would benefit from these things. (Usually they're not crying because they > don't know about and/or haven't experienced these other features.) I completely agree with this paragraph. Well said. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: Sat, 20 Jul 1996 14:04:01 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <glwFyFq00iV_A2BXcY@andrew.cmu.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> In-Reply-To: <4sr0v4$94d@solaris.cc.vt.edu> Excerpts from netnews.comp.sys.next.programmer: 20-Jul-96 Re: OpenStep design complaints by Chuck_Esterbrook@orcacom >> [obj initWithSize:aSize withColor:aColor withFoo:aFoo withBar:aBar >> withBaz:aBaz withBop:aBop]; >> >> obj.init(aSize, aColor, aFoo, aBar, aBaz, aBop); >> >> Call me crazy, but I prefer the second version. It's a lot more >> readable; it's shorter; and, if you use descriptive variable names in >> your code, it's as self-documenting as it needs to be. > > I've never seen a method named > "initWithSize:withColor:withFoo:withBar:withBaz:withBop:". Maybe this style > is in 4.0. I haven't used it much. Wait a second! :-) Someone made the argument that Obj-C was more self-documenting because you can give methods long names which describes what each argument should be. Since he suggested imagining what would happen with a 6-argument method call, I provided an example. > The programmers I know would call it "initWithSize:color:foo:bar:baz:bop:" > which is much shorter and more readable. i.e., you only use one "with". That's certainly what I would do, too. But then, that style has no particular advantage over Java's syntax in terms of describing what the various arguments should be. You simply can't make both claims at the same time.... -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: perform: with: afterDelay: cancelPrevious: Date: 20 Jul 1996 18:37:08 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4sr90k$26i@news.xmission.com> References: <4sr4mj$c9e@hpuniv.univ-lr.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) wrote: > Has someone implemented the method > "- perform:(SEL)aSelector with:anArg afterDelay:(int)ms > cancelPrevious:(BOOL)flag;" for the NSObject class ? Take a look at: http://www.misckit.com/NextOnLine/Issue.000/Frames/DelayedPerform.htmld/index .html and go to the side bar on NSObject messaging (the link is towards the end of the article). [NextOnLine isn't yet ready for primetime, but Matt and I have made a lot of good progress on it. If you want to see the full prototype, enter at http://www.misckit.com/NextOnLine/ and be prepared to WAIT, since it is graphic intensive and I have yet to tune things. The NoFrames section still needs a LOT of work, too, and many of the graphics don't quite look right--I _am_ working on it. Also, when the prototype is truly ready, I'll move it to a high bandwidth site, www.yacktman.com, which has a fiber optic link to the Internet backbone. Right now you'll have to deal with a 14.4k modem, and it is _slooooow_. If anyone wants to donate articles for the first issue, be my guest! Any comments or suggestions for improvements are also welcome. We'll be making a formal announcement to csn.announce when it is ready for primetime, so most of you will probably prefer to wait to check N.O.L. out after that announcement is made.] --- Later, -Don Yacktman don@misckit.com <a href="„http://www.misckit.com/don.html">My home page</a>
Newsgroups: comp.lang.objective-c,comp.sys.next.programmer From: chapman@grizzly.cs.washington.edu (Randy Chapman) Subject: Re: Java Bashing (Was: OpenStep design complaints) Followup-To: comp.lang.objective-c,comp.sys.next.programmer Sender: news@beaver.cs.washington.edu (USENET News System) Organization: Computer Science & Engineering, U of Washington, Seattle Message-ID: <DuuuK1.2tz@beaver.cs.washington.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> Date: Sat, 20 Jul 1996 18:40:48 GMT : As for a bytecode generator, I'm not sure what you're asking for. If : you'd like to see an Obj-C interpreter, then I agree with you that it : would be a great thing. But the object modules (.o files) generated by : a compiler are machine specific. : You don't really want to define an Obj-C virtual machine and an Obj-C : bytecode language spec (comparible to Java's VM and bytecode spec, etc), : do you? I'd like to see an ObjC -> JVM compiler, yes. I don't think the current JVM as is could likely do it well, because of the lack of -perform: However, David Hopwood has a package (relies on native code, unfortunately) to do this. And I believe Sun is adding it to the base language, so it should be doable. --randy ----------------------------------------------------------------------------- The software users top goals: * Not looking stupid * Not making any big mistakes * Getting an adequate amount of work done * Having fun (or at least not being too bored) -- Alan Cooper, "About Face"
From: don@misckit.com Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 18:56:43 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4sra5b$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > [...a bunch o' stuff I (Don) agree with...] > As for a bytecode generator, I'm not sure what you're asking for. If > you'd like to see an Obj-C interpreter, then I agree with you that it > would be a great thing. But the object modules (.o files) generated by > a compiler are machine specific. > > You don't really want to define an Obj-C virtual machine and an Obj-C > bytecode language spec (comparible to Java's VM and bytecode spec, etc), > do you? Personally, _I_ would like to see ObjC to Java bytecodes. Why bother making an ObjC VM? That's not needed. I think ObjC->Java bytecodes can be done. The runtime is there. (Just because the Java language doesn't use the full capability of the runtime engine shouldn't mean that we can't...) Not that this will ever happen, but one can dream. I'd love to be able to keep writing in ObjC but deploy to Java. By the way, I'd be happy even if it were a subset of ObjC (take out C pointers, junk like that). -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 20:30:40 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4srflg$q15@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> <4spdsm$q15@news.onramp.net> <4sq79a$26i@news.xmission.com> don@misckit.com wrote: > Steve Dekorte wrote: > > Garance A Drosehn wrote: > > > Steve Dekorte wrote: > > > > don@misckit.com wrote: > > > > I another message I just posted, I mention using the "Command" > > > > design pattern from Gamma, et. al. That's how you'd do it right now.. > > > Does this involve creating a mini Java class for the message you > > > want to send and then having it dynamically compiled and invoke > > > it's method? If so, that's not in the bounds of what I meant by > > > implementing an ObjC-like runtime in a reasonable way. > > This makes me ask: why is that solution unacceptable to you... Speed & memory. People who don't understand OO and think dynamic binding is only usefull in interfaces might see this. I use -perform: in almost every class I write, usually many times in each class and very often recursively and in tight loops. The performance in Java with the technique mentioned would be abysmal. Imagine the Java VM creating and compiling several thousand classes (one for each dynamic message) a second and you'll see what I mean. > At any rate, right _now_ I still prefer Obj-C by a long shot. My > only real interest in this is to empahasize that this _can_ be > done, even if you wouldn't necessarily want to mess with the > details themselves. It _can_ be done in C++ as well, the question was never if it could be done, but could it be done in a reasonable/practical way - that is with performance comparable to Objective-C and with a standard that actually lets you _reuse_ code without being locked-in to one companies dev tools. > > > > With Sun things might be different, but in general: > > > > The features that would actually make a product usefull > > > > are always in "the next release". > > True enough. .. > In order for anyone to be successful in trying to do what Taligent > wanted to do, they would have to build their own runtime to layer > on top of C++, since it completely lacks a runtime. ... And the same lessons will be learned and quickly forgetten once again with Java. > One other observation: people who program Java and came from the > C++ world only really don't know how to use the language to its > full potential. Those of use who come from ObjC, Smalltalk, Lisp, > or any other dynamic OOP language really do have an advantage in > that world. I agree. > Many of the Java tutorials and books I've read give > examples that show some of the most _awful_ OOD I've ever been > subjected to! This is an interesting point. Some of the worst code I've seen came from programming books, often the "Bibles" of their areas - such as the O'Reily books on X windows programming. And some of the worst Objective-C source code I've seen has come from NeXT. Steve -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdi
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: OpenStep design complaints Date: 20 Jul 1996 21:07:03 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4srhpn$g20@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> Charles William Swiger wrote: > Chuck_Esterbrook@orcacom > >> [obj initWithSize:aSize withColor:aColor withFoo:aFoo withBar:aBar > >> withBaz:aBaz withBop:aBop]; > >> > >> obj.init(aSize, aColor, aFoo, aBar, aBaz, aBop); > >> > >> Call me crazy, but I prefer the second version. You're welcome to do the same thing in ObjC syntax: - init:aSize :aColor :aFoo :aBar :aBaz :aBop; The nice thing about being able to put a note before each argument is that you can use the note to indicate how the arg will be used and the arg name to indicate what sort of object you expect: - initSize:aPoint color:aColor tailLength:aNumber otherAttributes:aList; > > The programmers I know would call it "initWithSize:color:foo:bar:baz:bop:" Personally, I don't like custom -init methods. I think -init should set the object's defaults up and any customization should be done with other messages. Aside from -init messages, I think there should be very few cases where you have to pass alot of args in one message. If you have to transport alot of data, chances are, that data is in the wrong place to begin with. -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdi
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 20 Jul 1996 21:33:27 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4srjb7$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> On the discussion re: initWithSize:withColor:withFoo:withBar:withBaz:withBop: vs. initWithSize:color:foo:bar:baz:bop: Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > That's certainly what I would do, too. > > But then, that style has no particular advantage over Java's syntax in > terms of describing what the various arguments should be. You simply > can't make both claims at the same time.... I'm with Chuck here in that in Java or ObjC you can document what is going on by using a long name--or choose not to do so. It really is a matter of style, and for readable code, you have to come up with a consistent style that includes all the information your (or your employer) deems important. I guess it is time to evangelize the book _Code_Complete_ by Steve McConnell, US ISBN 1-55615-484-4. Perhaps the best product to ever come out of Microsoft... it discusses a lot of stylistic points, and gives advantages and disadvantages of a wide variety of approaches. It doesn't have everything, despite the title, but it certainly does help to open the mind to many of the relevant issues... -- 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: Java Bashing (Was: OpenStep design complaints) Date: 20 Jul 1996 23:43:58 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4srqvu$26i@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> <4spdsm$q15@news.onramp.net> <4sq79a$26i@news.xmission.com> <4srflg$q15@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > don@misckit.com wrote: > > Steve Dekorte wrote: > > > > Does this involve creating a mini Java class for the message you > > > > want to send and then having it dynamically compiled and invoke > > > > it's method? If so, that's not in the bounds of what I meant by > > > > implementing an ObjC-like runtime in a reasonable way. > > > > This makes me ask: why is that solution unacceptable to you... > > Speed & memory. > > People who don't understand OO and think dynamic > binding is only usefull in interfaces might see this. > > I use -perform: in almost every class I write, usually many times in > each class and very often recursively and in tight loops. > The performance in Java with the technique mentioned would be abysmal. > Imagine the Java VM creating and compiling several thousand classes > (one for each dynamic message) a second and you'll see what I mean. Oh, OK. That would be unacceptable. I'm thinking rather of having a program that has, given the source files, generated the needed "mini classes" ahead of time. Once the class is loaded, it stays there. I think the performance, especially for a Java VM that does JIT compiling, will be quite acceptable. Creating the classes and compiling them on the fly would certainly cause a speed problem. Of course, if Sun comes through on their promises (and somone pointed out that betas are now available) this is all moot because the language will support what we want without this sort of tom foolery. > It _can_ be done in C++ as well, the question was never if it could > be done, but could it be done in a reasonable/practical way - > that is with performance comparable to Objective-C and with > a standard that actually lets you _reuse_ code without being > locked-in to one companies dev tools. Well, right now, Java performance for _anything_ doesn't match Obj-C. In my experience, it is darned slow. Of course I've not had a chance to play with the JIT compilers yet, and that may change my opinion of Java's speed. But everything I've seen and done to date is a lot slower than it should be, IMHO. > And the same lessons will be learned and quickly forgetten once > again with Java. Depressing thought, isn't it? I think you're probably right, though. > > Many of the Java tutorials and books I've read give > > examples that show some of the most _awful_ OOD I've ever been > > subjected to! > > This is an interesting point. Some of the worst code I've seen > came from programming books, often the "Bibles" of their areas - > such as the O'Reily books on X windows programming. > And some of the worst Objective-C source code I've seen has come > from NeXT. Yeah. A lot of the people think the MiscKit has some great code in it. IMHO, the "great" part is that (a) there's a lot there, (b) it works (mostly) and (c) it is free. As to design, there's very little there that I'm truly proud of from a design point of view. I really think many parts of the kit could be done _much_ better. (And the OPENSTEP MiscKit port will hopefully address much of that. :-) ) I think the only part I really think is fairly cool--and it still could stand some improvement--is the Merge Kit engine. And boy oh boy does it require a dynamic runtime to work its magic... Though I think Java just about has what I need to be able to port it. Maybe someday I will. It is a great example to use in job interviews with technically savvy folks, since I tell them what all it can do and how easy it is to use and they can't believe that such a thing could exist. Thanks to Obj-C, it can! :-) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: sjautomo@athena.mit.edu (John Automo) Newsgroups: comp.sys.next.programmer Subject: HELP with Printing the Display Date: 21 Jul 1996 00:03:25 GMT Organization: Massachvsetts Institvte of Technology Message-ID: <4srs4d$eks@senator-bedfellow.MIT.EDU> Hi, I need some advice. I plan to freeze an NS application anytime during the run-time, take a snapshot and get a hardcopy of its interface. Please email me if you don't mind sharing what you know. Any help will be greatly appreciated. Thanks. John sjautomo@MIT.EDU
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: OpenStep design complaints Date: 21 Jul 1996 00:03:16 GMT Organization: Genoa Software Systems Message-ID: <4srs44$b2@saturn.genoa.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> <4srhpn$g20@news.onramp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: dekorte@suite.com In <4srhpn$g20@news.onramp.net> Steve Dekorte wrote: > Personally, I don't like custom -init methods. > I think -init should set the object's defaults up and any customization should be > done with other messages. That's a reasonable position though I can think of two exceptions. 1. when the attribute is essential to the behavior of the object, and there is no reasonable default. NSView's initWithFrame: method comes to mind. 2. when the attribute must be fixed when the object is initialized and can never change. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: sjautomo@athena.mit.edu (John Automo) Newsgroups: comp.sys.next.programmer Subject: Re: HELP with Printing the Display Date: 21 Jul 1996 00:24:42 GMT Organization: Massachvsetts Institvte of Technology Message-ID: <4srtca$fdb@senator-bedfellow.MIT.EDU> References: <4srs4d$eks@senator-bedfellow.mit.edu> In article <4srs4d$eks@senator-bedfellow.mit.edu>, John Automo <sjautomo@athena.mit.edu> wrote: >Hi, I need some advice. I plan to freeze an NS application anytime >during the run-time, take a snapshot and get a hardcopy of its interface. Well, I got the answer myself finally. I found this grab.app in my NextApps. Please ignore my question. Thanks. John
From: Jon Rosen <jfrx@earthlink.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Sat, 20 Jul 1996 18:39:49 +0000 Organization: Earthlink Network, Inc. Message-ID: <31F127F5.27A2@earthlink.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Chuck_Esterbrook@orcacomputer.com wrote: > > I'd much rather see an improved Obj-C (cleanup the C part > to make the language safer and simpler, throw in garbage > collection, throw in some features from tom and then a > blitzkrieg to make ObjC the development language of the > next decade. (You know, a free-be compiler with source, > a bytecode generator, ANSI standard, etc.) Can you say "you snooze, you lose"? :-) > C++ programmers cry out for safety, gc, dynamic binding > and simplicity. ObjC programmers cry out for a few > improvements and industry acceptance. Java programmers > cry out ObjC features. Even those programmers who aren't > crying would benefit from these things. I agree with almost everything you said. The pity is that the first web stuff was done in NeXTSTEP, but NeXT didn't "get it" and now we have Netscape/Sun/Java, et al. And I don't think ANYTHING NeXT does now can change that. Can you imagine if NeXT had figured it out back in late 1994? They could have put together the COOLEST browser (using DPS as the graphics base) with ObjC/IB hooks, added the "NextScape" extension set to HTML and then "securitized" ObjC and made it the language of the future. Hell, DO might have actually supplanted CORBA then. But, NeXT has always followed it's "our way is the only way" strategy and just didn't see the writing on the walls. Sigh... I think they are too little, too late to change the direction now. > (Usually they're not crying because they don't know about and/or > haven't experienced these other features.) Nah, usually they're not crying because they are too busy figuring out how to get their blasted product to work whatever it takes. They are too busy to whine and complain. > It's too bad NeXT doesn't have that "we're going to change the > world" attitude anymore. I think they could make a kick ass > Obj-C2 and change it. I think they have to stick with a niche market today. Even if they could build the product you envision (and they probably could), the world simply won't buy a better mousetrap from Steve. Not in a million years (sadly enough). Jon Rosen
From: nurban@csugrad.cs.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 21 Jul 1996 00:21:36 -0400 Organization: Virginia Polytechnic Institute and State University Message-ID: <4ssb8g$6ng@csugrad.cs.vt.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F127F5.27A2@earthlink.net> In article <31F127F5.27A2@earthlink.net>, jfrx@earthlink.net wrote: > I agree with almost everything you said. The pity is that the first > web stuff was done in NeXTSTEP, but NeXT didn't "get it" and now we > have Netscape/Sun/Java, et al. Well, you'll have to admit that most people didn't "get it".. > And I don't think ANYTHING NeXT does > now can change that. Can you imagine if NeXT had figured it out back > in late 1994? They could have put together the COOLEST browser (using > DPS as the graphics base) with ObjC/IB hooks, added the "NextScape" > extension set to HTML and then "securitized" ObjC and made it the > language of the future. And then they could have watched it fail completely because it didn't run under Windows. -- Nathan Urban | nurban@vt.edu | Undergrad {CS,Physics,Math} | Virginia Tech
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: perform: with: afterDelay: cancelPrevious: Message-ID: <Duv380.z8@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <4sr4mj$c9e@hpuniv.univ-lr.fr> Date: Sat, 20 Jul 1996 21:48:00 GMT In article <4sr4mj$c9e@hpuniv.univ-lr.fr> wgaboria@iut-lr.univ-lr.fr (wilfrid Gaboriaud) writes: > Hello, > Has someone implemented the method > "- perform:(SEL)aSelector with:anArg afterDelay:(int)ms > cancelPrevious:(BOOL)flag;" for the NSObject class ? There is a NSTimer object in the OPENSTEP version of the Foundation classes. With NSTimer you can create timer objects. Such timer waits until a certain time interval has elapsed and then fires, sending a specified message to a specified object, which is the same thing the perform:with:afterDelay:cancelPrevious: message does. Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: tmb@best.com (--) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 21 Jul 1996 08:22:43 GMT Organization: home Message-ID: <TMB.96Jul21012243@best.best.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> In-reply-to: Chuck_Esterbrook@orcacomputer.com's message of 20 Jul 1996 16:03:51 GMT FCC: /home/tmb/mail/x-nout In article <4sr017$94d@solaris.cc.vt.edu> Chuck_Esterbrook@orcacomputer.com writes: The fact is that OpenStep relies on things like dynamic binding, categories and distributed objects. That stuff just isn't found in Java right now. Actually, Java has dynamic binding and dynamic typing (as well as protocols). And there are several implementations of distributed objects for Java out already. Structural dynamic type information and method pointers are also already available in sample implementations probably will may make it into the standard language. Categories are missing; lack of categories is a nuisance in any OOL with some form of static typing. However, there are straightforward if tedious workardounds. And I think by modifying the class loader, you could even implement them right now if you wanted to. Maybe there will be an effort to add them. C++ programmers cry out for safety, gc, dynamic binding and simplicity. ObjC programmers cry out for a few improvements and industry acceptance. Java programmers cry out ObjC features. Objective-C's big strength is interoperability with C, something Java will never parallel. But that also limits the degree to which it can achieve runtime safety, ensure link-time safety, and offer efficient precise garbage collection. They are different tools for different jobs, even if a lot of the designs and concepts carry over. Cheers, Thomas.
Newsgroups: comp.sys.next.programmer,comp.lang.objective-c From: stes@cwi.nl (David Stes) Subject: Re: Java Bashing (Was: OpenStep design complaints) Message-ID: <Duw3vz.H9L@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <TMB.96Jul21012243@best.best.com> Date: Sun, 21 Jul 1996 10:59:59 GMT In article <TMB.96Jul21012243@best.best.com> tmb@best.com (--) writes: > >Objective-C's big strength is interoperability with C, something Java >will never parallel. But that also limits the degree to which it can >achieve runtime safety, ensure link-time safety, and offer efficient >precise garbage collection. They are different tools for different >jobs, even if a lot of the designs and concepts carry over. > I agree. One the one hand, at some point, we want to benefit from the dynamic features of a language like Smalltalk, but that will not pay, of course, if somebody else has done "everything in C" and 't runs 100 times faster or so. There must be some provision to implement time critical loops in C, and than the rest of the system can be implemented at a higher level. That's not just some minor issue, to quickly be solved by some "foreign function" interface, once the language has been completely designed. It needs to be considered from the beginning. David.
From: robert@elastica.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 21 Jul 1996 13:01:22 -0400 Organization: Digital Gateway Systems Sender: robert@justine Message-ID: <fenm5v8v1.fsf@elastica.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> <4spdsm$q15@news.onramp.net> <4sq79a$26i@news.xmission.com> Don speaks about using native methods etc. David Hopwood put together a perform package which requires the JAVA source. -- You are amazed that they exist and they burn so bright whilst you can only wonder why.
From: robert@elastica.com Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 21 Jul 1996 13:05:33 -0400 Organization: Digital Gateway Systems Sender: robert@justine Message-ID: <fd91pv8o2.fsf@elastica.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4smaq0$26i@news.xmission.com> <4smofk$29j@news.onramp.net> <4snhl7$g5f@usenet.rpi.edu> <4spdsm$q15@news.onramp.net> <4sq79a$26i@news.xmission.com> Speaking of books Don ... has anybody read Copliens patterns book at all? -- You are amazed that they exist and they burn so bright whilst you can only wonder why.
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: OpenStep design complaints Date: 21 Jul 1996 19:56:03 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4su20j$g20@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> <4srhpn$g20@news.onramp.net> <4srs44$b2@saturn.genoa.com> Alex Blakemore wrote: > Steve Dekorte wrote: > > Personally, I don't like custom -init methods. > > I think -init should set the object's defaults up and any customization should > > be done with other messages. > > That's a reasonable position though I can think of two exceptions. > > 1. when the attribute is essential to the behavior of the object, and there is no > reasonable default. NSView's initWithFrame: method comes to mind. The View can have a default size. > 2. when the attribute must be fixed when the object is initialized and can never > change. I'd be interested to see an example of this. Steve -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdi
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 21 Jul 1996 19:59:03 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4su267$g20@news.onramp.net> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> <4srjb7$26i@news.xmission.com> don@misckit.com wrote: > ... a matter of style,... I guess it is time to > evangelize the book _Code_Complete_ by Steve McConnell, I recommend _Programming Smalltalk with Style_ (or something like that). An excelent book for Objective-C and OO programmers in general. Steve -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdi
From: apuleius@ix.netcom.com(William Grosso) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 22 Jul 1996 01:13:51 GMT Organization: Netcom Message-ID: <4sukkf$glo@sjx-ixn4.ix.netcom.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <fd91pv8o2.fsf@elastica.com> In <fd91pv8o2.fsf@elastica.com> robert@elastica.com writes: > >Speaking of books Don ... has anybody read Copliens patterns >book at all? > Snippets from OP: One Paragraph (book reviews for busy people). AFAIK, _Advanced C++_ is a brobdingnagian favorite amongst the static-typing crowd (as a friend put it, "Aside from Stroustrup's whiny book, it's the only C++ book worth owning"). Coplien uses the word "idioms" instead of "patterns" (and some of his "idioms" only make sense in C++ land) but, really, this is a basic-patterns book and a damn fine one. Correction. A Damn Fine One. _Pattern Languages of Program Design_ is less popular. IMO, this is because it is, roughly, a conference proceedings. Lots of papers, most of which record thoughts-in-progress. Some of the thoughts are interesting, some aren't, some are already available on the net. The material isn't fundamental (in the sense that AC++ is fundamental) and it isn't in final form. Which means PLoPD doesn't quite make it out to the hinterlands of programming. Cheers, Andy
From: Joe Freeman <jfreeman@freemansoft.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Sun, 21 Jul 1996 21:47:31 -0400 Organization: FreemanSoft Inc. Message-ID: <31F2DDB3.388C@freemansoft.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I really enjoy working with Obj-C even though the following comments may not show it. Chuck_Esterbrook@orcacomputer.com wrote: > I'd much rather see an improved Obj-C (cleanup the C part to make the > language safer and simpler, throw in garbage collection, throw in some > features from tom (http://tom.es.ele.tue.nl:8080/)) and then a blitzkrieg to > make ObjC the development language of the next decade. The "language of the next decade" train has left the station and Objective-C is not on it. NeXT went and bought out the only other Objective-C vendor while at the same time not porting the language to all platforms. There is no "next wave of computing" that NeXT can hook Objective-C onto. This means that NeXT needs to go out and sell the language since people's perception of the language is way more important than its real capabilities. (ie: the massive cleanup wouldn't be necessary if they had an partnering skills) Java is catching on because Sun got lucky and introduced at the same time the web train started to move and Sun jumped on it with both feet. Smalltalk has a superset of Jave but the single major vendor did not see the wave coming and did not want to do the appropriate marketing and positioning changes required to make a language go mainstream. So, smalltalk is sitting on the sidelines while Java gets into the game. > C++ programmers cry out for safety, gc, dynamic binding and simplicity. ObjC > programmers cry out for a few improvements and industry acceptance. Java > programmers cry out ObjC features. Even those programmers who aren't crying > would benefit from these things. (Usually they're not crying because they > don't know about and/or haven't experienced these other features.) There are plenty of successful languages without Objective-C's features. I'd love to have all of the Objective-C semnatics in Java. But I think we will see AppKit and FoundationKit and EOF put out by other folks if NeXT can't get their act together quckly both in the technology and partnering areas. > It's too bad NeXT doesn't have that "we're going to change the world" > attitude anymore. I think they could make a kick ass Obj-C2 and change it. Which track record are you using to base this prediction on? <joe '7 years in the NeXT community' freeman>
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 Jul 1996 04:15:10 GMT Organization: Digital Fix Development Distribution: inet Message-ID: <4suv8e$c8d@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: Matt Watson <Matt_Watson@NeXT.com> Newsgroups: comp.sys.next.programmer Subject: Re: - forwardInvocation: Date: Sun, 21 Jul 1996 22:28:45 -0700 Organization: NeXT Software, Inc. Message-ID: <31F3118E.6BF5@NeXT.com> References: <4spduq$q15@news.onramp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Steve Dekorte wrote: > > Anyone using this in the OpenStep Beta? > > Steve > -- Yes. You need to override methodSignatureForSelector: to make this work. I'll file a bug against the documentation if this isn't in there. matt.
From: Jon Rosen <jfrx@earthlink.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Sun, 21 Jul 1996 22:53:45 +0000 Organization: Earthlink Network, Inc. Message-ID: <31F2B4F9.5613@earthlink.net> References: <Dun1wy.FLB@cam-ani.co.uk> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F127F5.27A2@earthlink.net> <4ssb8g$6ng@csugrad.cs.vt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: nurban@vt.edu Nathan Urban wrote: > > [my moaning about NeXT "getting it" munched" > > And then they could have watched it fail > completely because it didn't run under Windows. Sorry you missed my point (or it wasn't clear). The whole point of "getting it" was that NeXT would have to have realized that this ObjC-based Browser would HAVE to run EVERYWHERE (just like Netscape). If they didn't do that, they still wouldn't have gotten it. Jon Rosen
From: Jon Rosen <jfrx@earthlink.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Sun, 21 Jul 1996 22:54:11 +0000 Organization: Earthlink Network, Inc. Message-ID: <31F2B513.242E@earthlink.net> References: <Dun1wy.FLB@cam-ani.co.uk> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F127F5.27A2@earthlink.net> <4ssb8g$6ng@csugrad.cs.vt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: nurban@vt.edu Nathan Urban wrote: > > [my moaning about NeXT "getting it" munched" > > And then they could have watched it fail > completely because it didn't run under Windows. Sorry you missed my point (or it wasn't clear). The whole point of "getting it" was that NeXT would have to have realized that this ObjC-based Browser would HAVE to run EVERYWHERE (just like Netscape). If they didn't do that, they still wouldn't have gotten it. Jon Rosen
From: Jon Rosen <jfrx@earthlink.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Sun, 21 Jul 1996 23:03:10 +0000 Organization: Earthlink Network, Inc. Message-ID: <31F2B72E.2112@earthlink.net> References: <Dun1wy.FLB@cam-ani.co.uk> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F127F5.27A2@earthlink.net> <4ssb8g$6ng@csugrad.cs.vt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: nurban@vt.edu Nathan Urban wrote: > > [my moaning about NeXT "getting it" munched" > > And then they could have watched it fail > completely because it didn't run under Windows. Sorry you missed my point (or it wasn't clear). The whole point of "getting it" was that NeXT would have to have realized that this ObjC-based Browser would HAVE to run EVERYWHERE (just like Netscape). If they didn't do that, they still wouldn't have gotten it. Jon Rosen
From: Zack Kurmas <kurmasz@river.it.gvsu.edu> Newsgroups: comp.sys.next.programmer Subject: GDB and polymorphism Date: 22 Jul 1996 07:38:31 -0000 Organization: Grand Valley State University Sender: cppmods@netlab.cs.rpi.edu Message-ID: <4svb5n$s5r@netlab.cs.rpi.edu> { This was too environment specific for comp.lang.c++.moderated -mod} I have a program which constructs a tree containing polymorphic classes. I think I have a memory leak somewhere, but when I try to look into the tree with the gdb debugger, I see stuff like this: vptr$token which points to an array with 12 pointers to __vtbl_ptr_type structures. How do I translate this into useful information so I can find the bug? Thanks in advance, Zachary Kurmas P.S. Token is the parent to such classes as operator, number, variable, etc. ------------------------------------------------------------------------------ | Zachary Kurmas | There is nothing wrong with allowing your | | kurmasz@river.it.gvsu.edu | mind to go blank; just remember to turn the | | www.csis.gvsu.edu/~kurmasz | sound off first. | |----------------------------------------------------------------------------| [ Send an empty e-mail to c++-help@netlab.cs.rpi.edu for info ] [ about comp.lang.c++.moderated. First time posters: do this! ]
From: jklein@artificial.com (Jon Klein) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Re: Netatalk- AppleTalk for NeXT port?! Followup-To: comp.sys.next.misc,comp.sys.next.programmer Date: 22 Jul 1996 14:24:29 GMT Organization: CDnow - The World's Largest Online Music Store http://cdnow.com Message-ID: <4t02ut$7b@netaxs.com> References: <31EFF7B0.5B45@email.mot.com> Eric A. Dubiel (ced016@email.mot.com) wrote: : I just recently talked to the person that was attempting : this port (he lives in Texas, now). He said he didn't get : very far, and hadn't saved his work. : : :wes So is anybody willing to give this a try? Netatalk works fine on SunOS 4.1.x, which is bsd based, and I recently heard that there is a patch which will allow it to run on freebsd... so it's certainly feasable. This is most likely a bit beyond my ability, but I think it could be done with some help. Anybody interested?
From: mlm@nectar2.info.isbiel.ch (Martin Mueller) Newsgroups: comp.sys.next.programmer Subject: IP Protocol Handlers Date: 22 Jul 1996 14:23:22 GMT Organization: Biel School of Engineering, CH-2501 Biel, Switzerland Distribution: world Message-ID: <4t02sq$g45@vega.info.isbiel.ch> I have two simple questions for which I found no answer in the on-line documentation. 1. how do you replace the default IP protocol handler in the NeXT kernel? I suppose just doing if_detach and accepting drivers that claim to be IFTYPE_ETHERNET is not enough. 2. how do you know (in a protocol handler) which netmask is being used when the interface gets a request to configure the interface (via ifconfig and IFCONTROL_SETADDR)? I want to know the netmask because the sysadmin can potentially configure the network on a non-byte boundary. If any guru(s) knows the answers, please reply to me in email as I do not read this newsgroup very often. Many thanks, -- Martin ---------------------------------------------------------------------------------------- Martin Muller mlm@info.isbiel.ch sysadmin
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 22 Jul 1996 16:11:30 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4t097i$jsk@solaris.cc.vt.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > Excerpts from netnews.comp.sys.next.programmer: 20-Jul-96 Re: Java > Bashing (Was: Open.. by Chuck_Esterbrook@orcacom > > The fact is that OpenStep relies on things like dynamic binding, categories > > and distributed objects. That stuff just isn't found in Java right now. > > That's not very true. I agree that Java has nothing comparible to Obj-C > categories. What's not very true? That OpenStep uses dynamic binding, categories, etc.? I'm unclear what you mean. > and java.lang.ClassLoader. I agree that Java's dynamic binding is not > as complete or as powerful as Obj-C's because methods and selectors are > not first-class objects "right now". But that is something that Sun is > actively working to change. All of this is very new at this stage. I wouldn't want to base a large project on it starting today. > Sure, but then you are talking about changes to the Obj-C APIs that NeXT > provides-- which have changed enough from NS 3.3 to NS 4.0 that you > pretty much have to port (not just recompile) your old Obj-C code under > 4.0. > > If you're talking about rewriting an Obj-C project in Java, you already > have to rewrite all of the source code, anyway. When you go from 3.3 source to 4.0 source you can use their scripts to save yourself an immense amount of work. Much less work that rewriting your app in a language that has less flexible runtime features. > > I'd much rather see an improved Obj-C (cleanup the C part to make the > > language safer and simpler, throw in garbage collection, throw in some > > features from tom (http://tom.es.ele.tue.nl:8080/)) and then a blitzkrieg to > > make ObjC the development language of the next decade. (You know, a free-be > > compiler with source, a bytecode generator, ANSI standard, etc.) > > Doesn't gcc from the Free Software Foundation qualify as a freely > available Obj-C compiler that comes with source? Isn't that compiler > fully-upwards compatible with ANSI-C? But it's not an ObjC2 compiler. It's an ObjC1. I don't care if Next made their own or updated GNU's. I'm just saying that if Next made ObjC2 then they should have a freebe compiler available asap. > You don't really want to define an Obj-C virtual machine and an Obj-C > bytecode language spec (comparible to Java's VM and bytecode spec, etc), > do you? No. What I'd rather see is GVM - general virtual machine. I've heard of people trying to translate languages such as Oberon to JVM and having problems because some of their language features don't map very well to JVM resulting in *very* inefficient code. It's not my impression that much needs to be added to JVM to make GVM. Probably just another addressing mode and a couple of instructions. > > C++ programmers cry out for safety, gc, dynamic binding and simplicity. ObjC > > programmers cry out for a few improvements and industry acceptance. Java > > programmers cry out ObjC features. Even those programmers who aren't crying > > would benefit from these things. (Usually they're not crying because they > > don't know about and/or haven't experienced these other features.) > > I completely agree with this paragraph. Well said. Why thank you! I thought you were going to give me a hard time throughout the whole message. :) -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 22 Jul 1996 16:23:11 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4t09tf$jsk@solaris.cc.vt.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F2DDB3.388C@freemansoft.com> Joe Freeman <jfreeman@freemansoft.com> wrote: > > It's too bad NeXT doesn't have that "we're going to change the world" > > attitude anymore. I think they could make a kick ass Obj-C2 and change it. > > Which track record are you using to base this prediction on? > <joe '7 years in the NeXT community' freeman> The wishful-thinking track record. :) I also think that a lot of people have the desire to use great technology. If they didn't, I don't think Next would even be in business anymore. It amazes me that their first machines had no floppy drive and even with one installed could not read DOS and Mac disks. A friend of mine called Next every other day for a week before he could upgrade to 4.0. Next should just concentrate on creating great software technology and hire Microsoft to do all their marketing! :) -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: indy@beckman.uiuc.edu (Steve Weintz) Newsgroups: comp.sys.next.programmer Subject: Hints on calling multiple dfaults Date: 22 Jul 1996 16:36:21 GMT Organization: University of Illinois at Urbana Message-ID: <4t0am5$56e@vixen.cso.uiuc.edu> Hi, the Subject: line is'nt very explanatory, so let me enlarge upon it: I'm trying to build a custom app that gets called by a login hook, like the old MOTD.app. Its purpose is to drasticaly simplify the process of using black hardware for complete novices (yes, I know NS is really easy to learn, but the target user base is REALLY ignorant :-). The idea is, you log in, and before the Workspace comes up you are presented with a window containing a SwapView and a button matrix. The buttons are labeled "Internet", "Games", "Word Processing", and so forth. You select a button, the SwapView loads a .nib containing a Text view and some stuff describing that category, then click a "Go" button. What I want to have happen is this: Once a selection is made and "Go" is clicked, Workspace loads with one of several defaults databases (.NeXTdefaults.inet, .NeXTdefaults.games, etc.) and only those apps and settings pertaining to the selected category appear at launch. In short -- is this possible? Any suggestions welcome. (Apart from convincing the client it's a dumb idea; I've tried that.) -- Beckman Institute | S T E V E W E I N T Z | indy@shout.net Visualization Facility |------------------------------| NeXTMail preferred indy@.uiuc.edu | Gerunding adverbly, | 217.344.5303 217.244.3074 | Noun verbed. |
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: 22 Jul 1996 14:07:13 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4t01uh$66t@netty.york.ac.uk> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4s6ebc$32g@saturn.genoa.com> On 12 Jul 1996 20:59:24 GMT, Alex Blakemore <alex@genoa.com> wrote: > b. The argument tests in OpenStep are compile time checks only. Thus if you wish > to pass a MySpecialArray to a method that expects a NSArray, you can do so. The > compiler will warn you about it, and you can silence the compiler warning if you > wish with a caste. If the method only sends public API messages to its argument, > and you class behaves reasonably like the one its pretending to be, this works. surely if you do this, then any subsequent behaviour is undefined, because the writer of the openstep class is perfectly at liberty to put in a line like : assert([obj isKindOf:[NSDictionary class]]); (as NXBrowser does for its NXBrowserCell objects, for instance) writing anything that invokes undefined behaviour is an absolute no-go for apps which are going to be around for any length of time at all, so the spec. has effectively stopped anyone from doing that (even though it might still work in the current circumstances). for this reason, i think that moving to static class typing rather than id is *not* an overall improvement. yes, it improves the ability of the compiler to statically typecheck, but IMHO it loses many of the original reasons for the existence of objective-C as a language. the real solution would be to document the interfaces well enough that you *could* plug and play classes implemented from various different frameworks. wasn't that what object-oriented programming was supposed to give us ? rog.
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 22 Jul 1996 14:23:05 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4t02s9$66t@netty.york.ac.uk> References: <4sk3oq$727@news.onramp.net> On 18 Jul 1996 01:24:42 GMT, Steve Dekorte <dekorte@suite.com> wrote: > > what compelling advantage does Self offer? > > The idea behind OO is reuse by data-structure independence. [...] > [The language Self] creates new possiblilities, such as being able to dynamically > add/remove variables&methods from classes and change class hierarchies. > Also, new instances are created by copying objects, so you can play > with a new class before propogating it's changes - this is where > the "prototype" word comes in. how does this aid data-structure independence ? it sounds like it will just contribute to the chaos where what we really want is more order (more interoperability, more compatibility leading to more software reuse) isn't Self perhaps just a dead end alley ? a sort of unstructured object oriented assembly language ? rog.
From: evan@ccpurch6.cc.vt.edu (Evan Greenberg) Newsgroups: comp.sys.next.programmer Subject: Passing Mouse Events Date: 22 Jul 1996 18:32:37 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4t0hg5$ot1@solaris.cc.vt.edu> I'm trying to write an application that will simulate performing a mouse click elsewhere on the screen. Right now I'm using - sendEvent: to place the NX_MOUSEDOWN message on the event queue, but I cannot figure out how to allow it to automatically select which view should receieve the mouse click. Thanks for any help! --Evan Greenberg --evan@ccpurch6.cc.vt.edu
From: plichta@sfu.ca (Jakub Plichta) Newsgroups: comp.sys.next.programmer Subject: ButtonCell color Date: 22 Jul 1996 18:30:38 GMT Organization: Simon Fraser University Message-ID: <4t0hce$4co@morgoth.sfu.ca> Keywords: buttonCell Is there any way to change the background and foreground (text) colors for individual buttonCells? I have a matrix of button cells and need to change their appearance according to their status, but the docs don't seem to suggest any solution. Thanks Jakub Jakub Plichta plichta@sfu.ca
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Self Date: 22 Jul 1996 19:02:19 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t0j7r$pln@news.onramp.net> References: <4sk3oq$727@news.onramp.net> <4t02s9$66t@netty.york.ac.uk> Roger Peppe wrote: > Steve Dekorte wrote: > > > what compelling advantage does Self offer? > > > > The idea behind OO is reuse by data-structure independence. > [...] > > > [The language Self] creates new possiblilities, such as being able to dynamically > > add/remove variables&methods from classes and change class hierarchies. > > Also, new instances are created by copying objects, so you can play > > with a new class before propogating it's changes - this is where > > the "prototype" word comes in. > > how does this aid data-structure independence ? The dynamic nature of self is more a result of objects only caring about their behavior(not their data-structures) than a cause. > it sounds like it will just contribute to the chaos where what we > really want is more order (more interoperability, more compatibility > leading to more software reuse) If you're worried about people changing the behavior of pre-fab libs, you can always add a "lock" option on class libs if you like. Apart from that, the only difference with being able to change things at runtime is that you can avoid compiles. > isn't Self perhaps just a dead end alley ? In itself, maybe. But the ideas & designs involved are a natural extension of OO concepts and if Self isn't the language to popularise them, another will. IMO, Prototype-based languages are the future of software development. > a sort of unstructured object oriented assembly language ? I don't know what would have given you that impression(?) Check out: http://self.sunlabs.com/ -- Steve Dekorte - OpenStep Consultant - Anahiem, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: (probably) stupid inheritence question... Date: 22 Jul 1996 19:43:23 GMT Organization: University of Illinois at Urbana Message-ID: <4t0lkr$ihm@vixen.cso.uiuc.edu> Short Version: Why does the runtime complain that I can't send a message to a subclass that is definied in its superclass? Specifically the factory defined allocation methods? I have a subclass of NSData that gets the following error when I try and use dataWithBytes:length: Jul 22 14:15:15 run[935] *** Uncaught exception: *** initialization method -initWithBytes:length:copy:freeWhenDone:bytesAreVM: cannot be sent to an abstract object of class ICDataStream: Create a concrete instance! exiting! Long Version: In order to facilitate the replacement of NXStreams with NSData objects I decided to subclass NSData to provide it with a stream-like interface. I was going to use a category but as I understand it you can not declare ivars in a category. I just added one ivar and a couple methods but the inheritence isn't working as I thought it would. I think I'm not missing something about the way objects are allocated and initialized before usage. I'm just starting out with objective-c and would appreciate any comments (even 'You shouldn't have done _that_!' :) Frankly, I'm kind of mad that I haven't been able to figure out this on my own. I've included my main and the .m and .h file for my subclass. ------ main.m ------ #import <AppKit/AppKit.h> #import <Foundation/Foundation.h> #include <stdio.h> #import "ICDataStream.h" int main(int argc, const char *argv[]) { NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; NSString* myString = @"ABCDEFG"; NSRange range = {2, 4}; ICDataStream *data1, *data2; char buf[5]; buf[4] = NULL; data1 = [ICDataStream dataWithBytes:[myString cString] length:[myString length]]; data2 = [data1 subdataWithRange:range]; [data2 getBytes:buf]; printf("The description od data2 = %s\n",buf); [pool release]; return 0; } --------- ICDataStream.h --------- #import <Foundation/NSData.h> #define IC_START 0 /* relative to start of file */ #define IC_CURRENT 1 /* relative to current position */ #define IC_END 2 /* relative to end of file */ @interface ICDataStream : NSData { long IC_position; } -(void) ICReset; -(long) ICTell; -(void) ICSeekTo: (long) offset; -(void) ICSeekTo: (long) offset from: (int) mode; -(long) ICReadCount: (long) count toBuff: (void *) buffer; @end -------- ICDataStream.m -------- #import "ICDataStream.h" #import <Foundation/NSRange.h> #include <stdio.h> @implementation ICDataStream -(void) ICReset { IC_position = 0; } -(long) ICTell { return IC_position; } -(void) ICSeekTo: (long) offset { [self ICSeekTo: offset from: IC_CURRENT]; } -(void) ICSeekTo: (long) offset from: (int) mode { switch (mode) { case IC_START: IC_position = offset; break; case IC_END: IC_position = length - offset; break; default: IC_position += offset; if (IC_position > length) { IC_position = length; } break; } } -(long) ICReadCount: (long) count toBuff: (void *) buffer { long IC_bytesRead = ((IC_position + count) > length) ? length - IC_position : count ; NSRange IC_readRange = { IC_position, IC_bytesRead }; [self getBytes:buffer range:IC_readRange]; [self ICSeekTo: IC_bytesRead + 1]; return IC_bytesRead; } @end -- Ian P. Cardenas (icardena@uiuc.edu) 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: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: (probably) stupid inheritence question... Date: 22 Jul 1996 20:55:08 GMT Organization: ALI Message-ID: <4t0prc$i8q@cetus.ali.bc.ca> References: <4t0lkr$ihm@vixen.cso.uiuc.edu> Cc: icardena@sumter.cso.uiuc.edu In comp.sys.next.programmer Ian Patrick Cardenas wrote: > Why does the runtime complain that I can't send a message to a > subclass that is definied in its superclass? Specifically the factory > defined allocation methods? I have a subclass of NSData that gets > the following error when I try and use dataWithBytes:length: > > Jul 22 14:15:15 run[935] *** Uncaught exception: *** initialization method -initWithBytes:length:copy:freeWhenDone:bytesAreVM: cannot be sent to an abstract object of class ICDataStream: Create a concrete instance! > exiting! > ... Welcome to the wonderful world of class clusters and hidden subclasses! The problem is exactly what is reported, ie. that your concrete subclass doesn't override enough methods to actually work. The NSData class is just an abstract class whose primitive methods must be over-ridden to accomplish anything. The easy way to accomplish what you have proposed (adding an instance variable which keeps track of a position within the data and then adding stream-like access methods) can be accomplished by modifying your example as follows: @interface ICDataStream : NSData { NSData *IC_realData; // actual regular data object long IC_position; // our new instance variable } ... - initWithBytes:(const void *)bytes length:(unsigned)length { [super init]; IC_realData = [[NSData dataWithBytes:bytes length:length] retain]; return self; } - (void)dealloc { [IC_realData release]; [super dealloc]; } And provide versions of NSData's primitive methods which forward the calls to IC_realData. As you can see, your ICDataStream doesn't even have to be a subclass of NSData. You could just make a subclass of NSObject which supports methods with which you would like to allow access to the bytes in the internal data object. -- 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 * "C++ is to C as Lung Cancer is to Lung."
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: Re: (probably) stupid inheritence question... Date: 22 Jul 1996 21:24:20 GMT Organization: University of Illinois at Urbana Message-ID: <4t0ri4$pkf@vixen.cso.uiuc.edu> References: <4t0lkr$ihm@vixen.cso.uiuc.edu> <4t0prc$i8q@cetus.ali.bc.ca> That'll teach me to think I know enough. I didn't understand the whole class clustering concept but after reading your article and looking at some more docs I get it now. For some good info on class clustering look at: http://www.next.com/NeXTanswers/HTMLFiles/2003.htmld/2003.html OR /NextLibrary/Frameworks/Foundation.framework/Resources/English.lproj/Documentation/Concepts/ProgrammingTopics/FoundationOverview.rtfd It's easy once you know what your problem is :) Thanks much. allan@ali.bc.ca (Allan Noordvyk) writes: > >Welcome to the wonderful world of class clusters and hidden subclasses! The >problem is exactly what is reported, ie. that your concrete subclass doesn't >override enough methods to actually work. The NSData class is just an >abstract class whose primitive methods must be over-ridden to accomplish >anything. > -- Ian P. Cardenas (icardena@uiuc.edu) 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: caljouwr@cadvision.com (Rob Caljouw) Newsgroups: comp.sys.next.programmer Subject: C++ undefined symbols Date: 23 Jul 1996 03:42:29 GMT Organization: GS/X Computing Corporation Distribution: World Message-ID: <4t1hn5$38bi@elmo.cadvision.com> Hello all. I've been trying to compile hylafax v3.0p10 on m68k rather unsuccessfully. I'm using gcc 2.7.2 and libg++. Everything compiles OK but when some of the executables link I get the following error: ls: Undefined symbols: .constructors_used .destructors_used I'm afraid I don't know much about C++, it's been a while. It appears that these symbols are created by the compiler but are not available to the linker. Any help would be greatly appreciated. Thank you. RJC
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <jim@ergotech.com> Message-ID: <199607231506.AA06491@ergotech.com> Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 3.3 v118.2) From: Jim Redman <jim@ergotech.com> Date: Tue, 23 Jul 96 09:16:52 -0600 Subject: Re: Identifying Type >for this reason, i think that moving to static class typing rather >than id is *not* an overall improvement. Most things in life are not black or white. Static typing does allow the compiler to generate "object does not respond to...." warnings, which is good. id's provide flexibility. Both have their place. Probably a good solution would be protocols with id's. With Objective-C this would take more discipline than an organization or individual normally has. Interesting that maps to java's interface concept, but I'm really trying to stay out of that thread. Jim
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Tue, 23 Jul 1996 16:20:08 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <QlxHDs600iWUA5wTUc@andrew.cmu.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> <4t097i$jsk@solaris.cc.vt.edu> In-Reply-To: <4t097i$jsk@solaris.cc.vt.edu> Excerpts from netnews.comp.sys.next.programmer: 22-Jul-96 Re: Java Bashing (Was: Open.. by Chuck_Esterbrook@orcacom >>> The fact is that OpenStep relies on things like dynamic binding, >>> categories and distributed objects. That stuff just isn't found in >>> Java right now. > >> That's not very true. I agree that Java has nothing comparible to Obj-C >> categories. > > What's not very true? That OpenStep uses dynamic binding, categories, > etc.? I'm unclear what you mean. It is not very true to claim that "things like dynamic binding, categories and distributed objects" cannot be "found in Java right now". This is because two of the three things you mention _can_ be found in Java, right now. >> I agree that Java's dynamic binding is not >> as complete or as powerful as Obj-C's because methods and selectors are >> not first-class objects "right now". But that is something that Sun is >> actively working to change. > > All of this is very new at this stage. I wouldn't want to base a large > project on it starting today. You're not refuting the statement that Java has a dynamic runtime, you're just claiming that it's new. While I agree, what was your point? [ ... ] >> Doesn't gcc from the Free Software Foundation qualify as a freely >> available Obj-C compiler that comes with source? Isn't that compiler >> fully-upwards compatible with ANSI-C? > > But it's not an ObjC2 compiler. It's an ObjC1. What is the difference between an ObjC2 compiler and an ObjC1 compiler? Did someone create a new revision of the language? [ ... ] >> I completely agree with this paragraph. Well said. > > Why thank you! I thought you were going to give me a hard time throughout > the whole message. :) Who, me? I never give people a hard time, even when I disagree with them. (Unless they deserve it, that is.... :-) -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: 23 Jul 1996 02:46:35 GMT Organization: Genoa Software Systems Message-ID: <4t1eeb$e7@saturn.genoa.com> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4s6ebc$32g@saturn.genoa.com> <4t01uh$66t@netty.york.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: rog@ohm.york.ac.uk In <4t01uh$66t@netty.york.ac.uk> Roger Peppe wrote: > i think that moving to static class typing rather > than id is *not* an overall improvement. > yes, it improves the ability of the compiler to statically typecheck, > but IMHO it loses many of the original reasons for the existence > of objective-C as a language. In what way? Adding type information to the header files merely enables the compiler to emit better warning messages. All of the dynamic features of Objective-C are still there and just as useful. There are times where you'll have to do a little extra work to turn off a warning now. There are also times where the compiler will catch an error for you. But again, none of this affects the main part of Obj-C (the runtime) at all. Why not try it for a few months before deciding whether its an improvement? -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: OpenStep design complaints Date: 23 Jul 1996 02:36:57 GMT Organization: Genoa Software Systems Message-ID: <4t1ds9$e7@saturn.genoa.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> <4srhpn$g20@news.onramp.net> <4srs44$b2@saturn.genoa.com> <4su20j$g20@news.onramp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: dekorte@suite.com Steve Dekorte wrote: >> I think -init should set the object's defaults up and any customization should >> be done with other messages. [I agree init should initialize all objects to a default usable state, usually without extra args, and that _most_ attributes should be modifiable at any time during an object's lifetime.] I wrote: > I can think of two exceptions. > 1. when the attribute is essential to the behavior of the object, and there is > no reasonable default. NSView's initWithFrame: method comes to mind. Steve: >> The View can have a default size. There is no "natural" default as there is for many attributes (zero, empty, NO etc). Requiring the size to be set during initWithFrame: ensures that there will never be an initialized view without a frame rectangle. (It establishes one of the class invariants in formal speak). This means you can write all methods assuming that there is a valid frame, simplifying both the implementation and the interface. Most importantly, a whole class of user errors is prevented by design. Clients are forced to set the frame initially and will never get some runtime exception telling them that they can't send the display message to the view until they call setFrame: first. But yes, that safety comes at a (small) cost. It forces the user to be in a position to give a frame when they init a view. But they can change it later if they wish. Alex > 2. when the attribute must be fixed when the object is initialized and can never change. Steve >> I'd be interested to see an example of this. Take a look at NSException for an extreme example. _All_ of its arguments are fixed at initialization. If you want to change one, your only choice is to init another exception object instead. Presumably, making exceptions immutable in this way simplified the exception run time. I don't suggest you wish to always prevent a user from changing some attribute after init, but that sometimes it IS useful to be able to enforce such a restriction. This allows you to ensure certain fundamental attributes won't change, except under your control. For example, you can store an object in a dictionary without dealing with the case where someone changes the key at any arbitrary moment. Or you can establish a network or database or other type of connection to some outside entity and know that the relevant connection attributes are fixed for the lifetime of the object. Being able to add these type of restrictions to the class interface is very useful at times. It is also a pain at times. The trick is knowing which circumstances call for which type of interface. After all, if all you do is replace all struct member access with set and get method invocations, then all OO has done is to make your code slower and more cluttered. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Tue, 23 Jul 1996 16:46:52 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <wlxHcwa00iWUA5wE1N@andrew.cmu.edu> References: <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <TMB.96Jul21012243@best.best.com> <Duw3vz.H9L@cwi.nl> In-Reply-To: <Duw3vz.H9L@cwi.nl> Excerpts from netnews.comp.sys.next.programmer: 21-Jul-96 Re: Java Bashing (Was: Open.. by David Stes@cwi.nl > I agree. One the one hand, at some point, we want to benefit from > the dynamic features of a language like Smalltalk, but that will not pay, > of course, if somebody else has done "everything in C" and 't runs 100 > times faster or so. Do you have any evidence to suggest every realistic application will run 100 times faster in C then in a dynamic language? Obj-C method invocations are roughly 3 to 4 times slower than C function calls, and yet you don't seem to object to Obj-C's performance. And, right now, most Java programs are being run in pure interpreters-- JIT's (Just-In-Time compilers) for Java should improve performance significantly. Even with that limitation, consider these results for a web server written in Java: http://www.w3.org/pub/WWW/Jigsaw/User/Introduction/performance.html The most anyone can claim is that Java is faster than some C implementations, and no more than a third the speed of the fastest implementation tested. That's hardly a factor of 100's. > There must be some provision to implement time critical loops in C, and > than the rest of the system can be implemented at a higher level. There was a point when people said that you'd have to have provisions to implement time critical loops in native assembly. However, with the notable exception of people writing compiler systems or operating systems (because those things need to deal with assembly directly for various reasons), very few professional developers ever write _anything_ in assembly nowadays. Do you know why? It's because modern C compilers can generate better code (in terms of speed, size, and correctness) than the average assembly programmer can, and a C compiler can do this task several orders of magnitude more quickly than even the best assembly programmer can. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: NSTableView Date: 23 Jul 1996 21:25:52 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t3g10$m7b@news.onramp.net> Anyone gotten these to work in OpenStep? When I drag in a "TableView" in IB, the inspector indicates that it's a subclass of NSScrollView(although the docs say it's a subclass of NSControl). Fine so far, but when I link it to an id and run the app it sees it as an NSScrollView only, not an NSTableView. Anyone else had this problem? -- Steve Dekorte - OpenStep Developer - Anaheim, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4t3g10$m7b@news.onramp.net> Control: cancel <4t3g10$m7b@news.onramp.net> Date: 23 Jul 1996 23:03:34 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t3lo6$oal@news.onramp.net>
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: OpenStep Exceptions Date: 23 Jul 1996 23:18:01 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t3mj9$oal@news.onramp.net> How do I avoid losinng the stack when an exception is raised? (it would make debugging alot easier) -- Steve Dekorte - OpenStep Developer - Anaheim, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep design complaints Date: 23 Jul 1996 23:45:47 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4t3o7b$8d8@solaris.cc.vt.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <Uluzqv_00iVEQ8Sq4a@andrew.cmu.edu> <4sr0v4$94d@solaris.cc.vt.edu> <glwFyFq00iV_A2BXcY@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > > I've never seen a method named > > "initWithSize:withColor:withFoo:withBar:withBaz:withBop:". Maybe this style > > is in 4.0. I haven't used it much. > > Wait a second! :-) > > Someone made the argument that Obj-C was more self-documenting because > you can give methods long names which describes what each argument > should be. Since he suggested imagining what would happen with a > 6-argument method call, I provided an example. Right. And I'm saying that your example is a bad one. No one puts "with" in their message name six names. Instead they say "initWithAttribute:ordering:" or "initWithBytes:length:". Yeah, they could put `with' everywhere, but they could do that in Java or C-- as well. > > The programmers I know would call it "initWithSize:color:foo:bar:baz:bop:" > > which is much shorter and more readable. i.e., you only use one "with". > > That's certainly what I would do, too. > > But then, that style has no particular advantage over Java's syntax in > terms of describing what the various arguments should be. You simply > can't make both claims at the same time.... >> > Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer Why doesn't it have an advantage? I like [[NSDate alloc] initWithYear:1996 month:7 day:23 hour:19 minute:0 second:0 timeZone:nil]; much better than NSDate.alloc().initYearMonthDayHourMinuteSecondTimeSize (1996,7,19,0,0,nil); Give that it's a preference I can certainly make both claims. -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Tue, 23 Jul 1996 20:20:41 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <0lxKlNC00iWTA9a0Z8@andrew.cmu.edu> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> <4sra5b$26i@news.xmission.com> In-Reply-To: <4sra5b$26i@news.xmission.com> Excerpts from netnews.comp.sys.next.programmer: 20-Jul-96 Re: Java Bashing (Was: Open.. by don@misckit.com >> You don't really want to define an Obj-C virtual machine and an Obj-C >> bytecode language spec (comparible to Java's VM and bytecode spec, etc), >> do you? > > Personally, _I_ would like to see ObjC to Java bytecodes. I'd rather see something that would aid in porting source code from Obj-C to Java. Even something as simple as rewriting Obj-C .[hm] files into .java files by keeping all of the ivars and method names the same, and leaving the original Obj-C source as Java comments would be useful (and fairly easy to write). > Why bother making an ObjC VM? That's not needed. I think ObjC->Java > bytecodes can be done. The runtime is there. (Just because the > Java language doesn't use the full capability of the runtime > engine shouldn't mean that we can't...) Not that this will ever > happen, but one can dream. I'd love to be able to keep writing > in ObjC but deploy to Java. By the way, I'd be happy even if it > were a subset of ObjC (take out C pointers, junk like that). But C arrays are implemented using pointers-- 'a[1]' is identical in meaning to '*a + 1'. You probably can find Java equivalents for the purely object-oriented Obj-C code, but a lot of pure C code-- including the actual implementations of Obj-C methods-- simply cannot be expressed in Java. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: jeremy@inetnebr.com (Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Re: (probably) stupid inheritence question... Date: 24 Jul 1996 01:26:35 GMT Organization: Internet Nebraska Message-ID: <4t3u4b$brc@duck.inetnebr.com> References: <4t0lkr$ihm@vixen.cso.uiuc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii NNTP-Posting-User: jeremy To: icardena@sumter.cso.uiuc.edu -----BEGIN PGP SIGNED MESSAGE----- In article <4t0lkr$ihm@vixen.cso.uiuc.edu>, icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) writes: > Short Version: > Why does the runtime complain that I can't send a message to a > subclass that is definied in its superclass? Specifically the factory > defined allocation methods? I have a subclass of NSData that gets > the following error when I try and use dataWithBytes:length: > > Jul 22 14:15:15 run[935] *** Uncaught exception: *** initialization > method -initWithBytes:length:copy:freeWhenDone:bytesAreVM: cannot be sent > to an abstract object of class ICDataStream: Create a concrete instance! > exiting! > This has to do with the new NS way of doing things. NSData doesn't really do anything. All the real methods are hidden away in NSConcreteData or something to prevent you from subclassing it. While this makes a nice spec (i.e. The NS version need not have the same ivars as the Solaris or GNUstep versions) it sure is a pain to make custom classes just a little different from the normal class. You will need to implement all of the methods listed as primative in the documentation. - -- Jeremy Bettis -*- PGP public key available,send mail with subj "Send pgp key" Home: jeremy@tddi.inetnebr.com Work: jeremy@hksys.com NeXTMAIL/MIME/PGP accepted -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQBVAwUBMfV7vhxzunB7hARlAQHxbwH8DPJPROjCVZc73XhIblxlT+aVCjKyrWen Gvno7xG4qx5XQvKmioeQBh5SjZ0OniFnigMRdIKmkF8JpJaZ/F7bCg== =ZTdu -----END PGP SIGNATURE-----
From: tal@cs.caltech.edu (Tal Lancaster) Newsgroups: comp.graphics.rendering.renderman,comp.sys.next.programmer Subject: Re: RiMakeTexture - how 2? Date: 24 Jul 1996 03:37:46 GMT Organization: California Institute of Technology, Pasadena Message-ID: <4t45qa$ssa@gap.cco.caltech.edu> References: <4soctb$i3u@news.tuwien.ac.at> Cc: wilkie@cg.tuwien.ac.at In <4soctb$i3u@news.tuwien.ac.at> Alexander Wilkie wrote: > Hi newsgroups! > > I have tried to get the RenderMan function > > RtToken RiMakeTexture(RtToken image, char *map, RtToken sWrap, RtToken tWrap, > RtFloatFunc filterFunction, RtFloat sWidth, RtFloat tWidth, > ...); > > (excerpt from <ri.h>) > > to work on a NeXT, with no luck so far. > > I'd like to do this from within an application so that it can display > textures that are computed at run time properly. It saves them to disk as > TIFFs, which then ought to be processed by this call for later use by > RiSurface(). What happens is... , well nothing. No error messages, no > crashes, but no textures either. Odd. > > JFTR: I have made sure that this function is only called when the > photorealistic version of RenderMan (as opposed to QRM) is active. > > Why does this function call differ from the one shown in the RM companion? I > assumed that "image" refers to the source file name (why "RtToken"?) and > "map" to the output file name, but I might be totally wrong here. > > Could anyone shed some light on this? > Please e-mail me, I'll summarize. I am not sure why isn't working. But as you are calling prman anyways, you might consider dumping out a RIB directly. The following is a Perl script from the RenderMan Repository (http://pete.cs.caltech.edu/RMR/rmDetails.htmld/index.html#mktmap) that does that and then calls prman: #!/usr/local/bin/perl #Create a prman texture-map (by calling prman + using MakeTexture) # If two arguments then the swrap,twrap, filterfunc, and swidth, twidth # will use default values. To find out more check out RC p. 256 # #NOTE: the swrap/twrap and swidth/twidth parameters are assumed to be the same as their counter part. # # Author: Tal Lancater (tal@cs.caltech.edu) # Date: 6/22/95 # # Usage: mktmap.pl src dst $len = @ARGV; if ($len != 2 && $len != 5) { print "Usage mktmap.pl sourceImage textureMap [wrap filter pixel]\n"; exit 1; } #defaults for wrap filter and pixel $Wrap="periodic"; $Filter="box"; $Pixel=2.0; @myargs = reverse(@ARGV); $src = pop(@myargs); $dst = pop(@myargs); if ($len == 5) { $wrap = pop(@myargs); $filter = pop(@myargs); $pixel = pop(@myargs); } else { # Must be just the source and destination files # use defaults $wrap = $Wrap; $filter = $Filter; $pixel=$Pixel; } $tmpName = "mk" . time . ".rib"; open(tmpfile, ">$tmpName"); #try addint 3 more params periodic box pixels #Create RIB file print (tmpfile "MakeTexture \"$src\" \"$dst\" \"$wrap\" \"$wrap\" \"$filter\" $pixel $pixel\n"); close(tmpfile); #Call prman to create the texturefile. print "Creating texture map $dst from image $src\n"; system "prman $tmpName"; unlink $tmpName; -- *********************************************************************** Tal Lancaster - tal@cs.caltech.edu (NeXTSTEP, MIME, and regular Email The RenderMan Repository -- http://pete.cs.caltech.edu/RMR/ ************************************************************************
From: suckow@bln.sel.alcatel.de (Ralf Suckow) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Exceptions Date: 24 Jul 1996 08:04:31 GMT Organization: Alcatel/Bell Distribution: world Message-ID: <4t4lef$p98@btmpjg.god.bel.alcatel.be> References: <4t3mj9$oal@news.onramp.net> Steve Dekorte writes > > How do I avoid losinng the stack when an exception is raised? > (it would make debugging alot easier) > > -- > Steve Dekorte - OpenStep Developer - Anaheim, CA > "Fundamentalism isn't about religion. It's about power." - S. Rushdie Set a breakpoint at "raise" (the method and the function). If you run on Intel, you may put the following into your .gdbinit. It allows you to get info if you are in code without symbol table. I used it alot for debugging inside IB, but it works for Intel only! Don't remember where it is from, someone sent me a while back ... -------------------------------------------------------------------- define reason po [*((id *)($fp + 8)) exceptionReason] end document reason Print the reason for an NSException. When a program is inside of the method -[NSException raise], this command prints out a textual reason for the exception. end define pself p *(id)*($fp + 8) end document pself Print self inside a method implementation end define pclass p *((id)*($fp + 8))->isa end document pclass Print the class info inside a method implementation end -------------------------------------------------------------------- Yours, ------------------------ Ralf.Suckow@bln.sel.alcatel.de | All opinions are mine.
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Exceptions Date: 24 Jul 1996 09:49:45 GMT Organization: Tech Net GmbH Message-ID: <4t4rjp$rfv@ddfservb.technet.net> References: <4t3mj9$oal@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: > >How do I avoid losing the stack when an exception is raised? >(it would make debugging alot easier) I already asked myself this question a couple of times. Yes, it's not very nice to get a message like "Uncaught exception xxxxx" and you don't know were the exception was raised. But think about it --- it's not possible to keep the stack because you don't know in advance if an exception will be caught or not. The only possibility would be to copy the whole stack when a first exception it raised and to remove that copy if the application falls of the end of some exception handling domain. ---> expensive This is true not only for Openstep exceptions. Like always --- you win some, you loose some. -- Constantin Szallies, Energotec GmbH szallies@energotec.de 49211-9144018
From: tmb@best.com (--) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: 24 Jul 1996 09:44:25 GMT Organization: home Message-ID: <TMB.96Jul24024425@best.best.com> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4s6ebc$32g@saturn.genoa.com> <4t01uh$66t@netty.york.ac.uk> In-reply-to: rog@ohm.york.ac.uk's message of 22 Jul 1996 14:07:13 GMT FCC: /home/tmb/mail/x-nout In article <4t01uh$66t@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) writes: for this reason, i think that moving to static class typing rather than id is *not* an overall improvement. yes, it improves the ability of the compiler to statically typecheck, but IMHO it loses many of the original reasons for the existence of objective-C as a language. The good thing is: both Java and Objective-C give you a choice between static and dynamic paradigms; you can pick and choose what is best for your particular application. Thomas.
From: tmb@best.com (--) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: cmsg cancel <TMB.96Jul24024425@best.best.com> Control: cancel <TMB.96Jul24024425@best.best.com> Date: 24 Jul 1996 10:01:38 GMT Organization: home Message-ID: <TMB.96Jul24030138@best.best.com>
From: mmalcolm crawford <m.crawford@shef.ac.uk> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 24 Jul 1996 11:37:20 GMT Organization: University of Sheffield, UK Message-ID: <4t51tg$bfj@bignews.shef.ac.uk> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F2DDB3.388C@freemansoft.com> In-Reply-To: <31F2DDB3.388C@freemansoft.com> On 07/22/96, Joe Freeman wrote: [...] > There are plenty of successful languages without Objective-C's > features. I'd love to have all of the Objective-C semnatics in > Java. But I think we will see AppKit and FoundationKit and EOF > put out by other folks if NeXT can't get their act together quckly > both in the technology and partnering areas. > Well, WebObjects is slated to support compiled Java later this year, and one of the news stories I saw recently (can't find it now -- anyone?) quoted a NeXT VP as saying that there was good stuff going on in Engineering with Java... so who knows? Best wishes, mmalc --
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: <9607240959.AA01986@tlpnl.tlp.nl> Date: Wed, 24 Jul 96 10:58:59 +0100 Subject: Java Bashing (Objective-C features) don@misckit.com wrote: > Charles William Swiger <cs4w+@andrew.cmu.edu> wrote: > > [...a bunch o' stuff I (Don) agree with...] > > As for a bytecode generator, I'm not sure what you're asking for. If > > you'd like to see an Obj-C interpreter, then I agree with you that it > > would be a great thing. But the object modules (.o files) generated by > > a compiler are machine specific. > > > > You don't really want to define an Obj-C virtual machine and an Obj-C > > bytecode language spec (comparible to Java's VM and bytecode spec, etc), > > do you? > > Personally, _I_ would like to see ObjC to Java bytecodes. Why > bother making an ObjC VM? That's not needed. I think ObjC->Java > bytecodes can be done. The runtime is there. (Just because the > Java language doesn't use the full capability of the runtime > engine shouldn't mean that we can't...) Not that this will ever > happen, but one can dream. I'd love to be able to keep writing > in ObjC but deploy to Java. By the way, I'd be happy even if it > were a subset of ObjC (take out C pointers, junk like that). That was my idea to. Every compiler uses byte codes before making machine specific instructions, don't they? I thought, why not make a compiler that uses an Objective-C'ish syntax and produces Java byte codes. But does de Java VM support selectors and -respondsToSelector: stuff? I personally think myTable = [ [Table alloc] initWithRows:8 columns:12 fill:NO']; is *much* clearer than myTable = new Table(8,12,false); I can't see wheter the 8 or the 12 means rows or columns, and what the 'false' means. But I am not throwing oil on the language flame. I think the language should have a new syntax, because the Objective-C syntax can use some improvements (indeed, no more *) and should be more Jave like. I even gave 'my' language the name "Pampus". Java is an Indonesian Island. Pampus is a (very small) island near Amsterdam where the Vereenigde Oostindische Compagnie (VoC) whould gather it's boats before saling to the Dutch colony of Indonesia. As usual, uttering nonsense. Maurice le Rutte.
From: NEED, CASH, FAST??<a@b.com> Newsgroups: comp.sys.next.programmer Subject: It's Here!! Date: Wed, 24 Jul 1996 09:47:50 -0700 Organization: Walt Disney Studios Message-ID: <4t5dc2$68s@phcs.phcs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello! I've got some awesome news that I think you need to take two minutes to read if you have ever thought "How could I make some serious cash in a hurry???", or been in serious debt, ready to do almost anything to get the money needed to pay off those bill collectors. So grab a snack, a warm cup of coffee, or a glass of your favorite beverage, get comfortable and listen to this interesting, exciting find! Let me start by saying that I FINALLY FOUND IT! That's right! I found it! And I HATE GET RICH QUICK SCHEMES!! I hate those schemes like multi-level marketing, mail-order schemes, envelope stuffing scams, 900 number scams... the list goes on forever. I have tried every darn get rich quick scheme out there over the past 12 years. I somehow got on mailing lists for people looking to make money (more like 'desperate stupid people who will try anything for money!'). Well, when I was a teenager, these claims to 'get me rich quick' sounded irresistible! I would shell out $14.95 here, $29.95 here, and another $49.95 there. I had maxed out my new Circuit City Card AND my Visa...I was desperate for money!! So, I gave them all a chance but failed at every one of them! Maybe they worked for some people, but not for me. Eventually, I just tossed that JUNK MAIL in the trash when I got the mail. I recognized it right away. I can smell a money scam from a mile away these days, SO I THOUGHT....I thought I could sniff out a scam easily. WAS I WRONG!!....I LOVE THE INTERNET!!! I was scanning thru a NEWSGROUP and saw an article stating to GET CASH FAST!! I though..."Here on the Internet?? Well, I'll just have to see what schemes could possible be on the internet." The article described a way to MAIL A ONE DOLLAR BILL TO ONLY FIVE PEOPLE AND MAKE $50, 000 IN CASH WITHIN 4 WEEKS! Well, the more I thought about it, the more I became very curious. Why? Because of the way it worked AND BECAUSE IT WOULD ONLY COST ME FIVE DOLLARS (AND FIVE STAMPS), THAT'S ALL I EVER PAY....EVER!! Ok, so the $50,000 in cash was maybe an tough amount to reach, but it was possible. I knew that I could at least get a return of $1,000 or so. So I did it!! As per the instructions in the article, I mailed out ('snail mail'for you e-mail fanatics) a single dollar bill to each of the five people on the list that was contained in the article. I included a small note, with the dollar, that stated "Please Add Me To Your List." I then removed the first position name of the five names listed and moved everyone up one position, and I put my name in position five of the list. This is how the money starts rolling in! I then took this revised article now with my name on the list and REPOSTED IT ON AS MANY NEWSGROUPS AND LOCAL BULLETIN BOARD MESSAGE AREAS THAT I KNEW. I then waited to watch the money come in...prepared to maybe receive about $1000 to $1500 in cash or so.... But what a welcome surprise when those envelopes kept coming in!!! I knew what they were as soon as I saw the return addresses from people all over the world-Most from the U.S., but some from Canada, even some from Australia! I tell you, THAT WAS EXCITING!! So how much did I get in total return? $1000? $5000? Not even!!! I received a total of $23,343!!! I couldn't believe it!! I now have a brand new black Acura Integra to speak for, due to this!! Now after almost 8 months, I am ready to do it again!!! So maybe it was possible to get $50,000 in cash, I don't know, but IT COMPLETELY DEPENDS ON YOU, THE INDIVIDUAL! You must follow through and repost this article everywhere you can think of! The more postings you achieve will determine how much cash will arrive in your very own mailbox!! It's just too easy to pass up!!! Let's review the reasons why you should do this: The only cost factors are for the five stamps, the 5 envelopes and the 5 one dollar bills that you send out to the listed names by snail mail (US Postal Service Mail). Then just simply repost the article (WITH YOUR NAME ADDED) to all the newsgroups and local BBS's you can. Then sit back and, (ironically), enjoy walking (you can run if you like! :o ) down your driveway to your mailbox and scoop up your rewards!! We all have five dollars to put into such an easy effortless investment with SPECTACULAR REALISTIC RETURNS OF $15,000 to $25,000 in about 3-5 weeks! So HOLD OFF ON THOSE LOTTERY NUMBERS FOR TODAY, EAT AT HOME TONIGHT INSTEAD OF TAKEOUT FROM McDONALDS AND INVEST FIVE DOLLARS IN THIS AMAZING MONEY MAKING SYSTEM NOW!!! YOU CAN'T LOSE!! So how do you do it exactly, you ask? I have carefully provided the mostdetailed, yet straightforward instructions on how to easily get this underway and get your cash on its way. SO, ARE YOU READY TO MAKE SOME CASH!!!?? HERE WE GO!!! *** THE LIST OF NAMES IS AT THE END OF THIS ARTICLE. *** OK, Read this carefully. Get a printout of this information, if you like, so you can easily refer to it as often as needed. INSTRUCTIONS: 1. Take a sheet of paper and write on it the following: "Please add my name to your list". This creates a service out of this money making system and thus making it completely legal. You are not just randomly sending a dollar to someone, you are paying one dollar for a legitimate service. Make sure you include your name and address. I assure you that, again, this is completely legal! For a neat little twist, also write what slot their name was in: "You were in slot 3", Just to add a little fun! This is all about having fun and making money at the same time! 2. Now fold this sheet of paper around a dollar bill , (no checks or money orders), and put them into an envelope and send it on its way to the five people listed. The folding of the paper around the bill will insure its arrival to its recipient. THIS STEP IS IMPORTANT!! 3. Now listen carefully, here's where you get YOUR MONEY COMING TO YOUR MAILBOX. Look at the list of five people; remove the first name from position one and move everyone on the list up slot one on the list. Position 2 name will now move to the position 1 slot , position 3 will now become position 2, 4 will be be 3, 5 wil be 4. Now put your name, address, zipcode AND COUNTRY in position 5, the bottom position on the list. 4. Now upload this updated file to as many newsgroups and local bulletin boards' message areas & file section as possible. Give a catchy description of the file so it gets noticed!! Such as: "NEED FAST CASH?, HERE IT IS!" or "NEED CASH TO PAY OFF YOUR DEBTS??", etc. And the more uploads, the more money you will make, and of course, the more money the others on the list will make too. LET'S ALL TAKE CARE OF EACH OTHER BY BEING HONEST AND BY PUTTING FORTH 120 PERCENT INTO THIS PROFITABLE & AMAZING SYSTEM!!! You'll reap the benefits, believe me!!! Set a goal for the number of total uploads you'll post, such as 15-20 postings or more! Always have a goal in mind!!! If you can UUE encode the file when uploading, that will make it easier for the poeple to receive it and have it downloaded to their hard drive. That way they get a copy of the article right on their computer without hassles of viewing and then saving the article from the File menu. Don't alter the file type, leave it as an MS-DOS Text file. The best test is to be able to view this file using Microsoft's Notepad for Windows 3.x or WordPad for Windows '95. If the margins look right without making the screen slide left or right when at the ends of the sentences, you're in business! 5. If you need help uploading, simply ask the sysop of the BBS, or "POST" a message on a newsgroup asking how to post a file, tell them who your Internet provider is and PEOPLE WILL ALWAYS BE GLAD TO HELP. I would try to describe how to do it but there are simply too many internet software packages with slightly different yet relatively simple ways to post or upload a file. Just ask for help or look in the help section for 'posting'. I do know that for GNN, you simply select 'POST' then enter a catchy description under the subject box, choose 'ATTACH', selecting 'UUE' and NOT 'TXT', then choose 'Browse' to go look for the file. Find your text file CASH.TXT and click on it and choose 'OK'. Place a one line statement in the main body section of the message post screen. Something like "Download this to read how to get cash arriving in your mailbox with no paybacks!" or whatever. Just make sure it represents its true feasibility, NOT something like..."Get one million dollars flooding in your mailbox in two days!" You'll never get ANY responses! 6. And this is the step I like. JUST SIT BACK AND ENJOY LIFE BECAUSE CASH IS ON ITS THE WAY!! Expect to see a little money start to trickle in around 2 weeks, but AT ABOUT WEEKS 3 & 4, THE MONEY STORM WILL HIT YOUR MAILBOX!! All you have to do is take it out of the mailbox and try not to scream too load (outside anyway) when you realize YOU HIT THE BIG TIME AT LAST!! 7. So go PAY OFF YOUR BILLS AND DEBTS and then get that something special you always wanted or buy that special person in your life (or the one you want in your life) a gift they'll never forget. ENJOY LIFE! 8. Now when you get low on this money supply, simply re-activate this file again; Reposting it in the old places where you originally posted and possibly some new places you now know of. Don't ever lose this file, always keep a copy at your reach for when you ever need cash. THIS IS AN INCREDIBLE TOOL THAT YOU CAN ALWAYS RE-USE TIME AND TIME AGAIN WHEN CASH IS NEEDED! 9. (This step added by Charles Reiley). Hello, This is exciting isn't it?! While I'm on the list, just add a note saying "Please include extra money tips" with your name & E-MAIL address, and I will (FOR FREE) send you some neat methods to increase the money you will receive with this plan. Why?...Why not? I'm not a selfish jerk...I like helping out others. E-mail just makes it a touch easier and cheaper, too! After I drop off the list, I can no longer offer you this advice, obviously, but maybe someone else who gets my tips will offer and simply replace my name on this step number 9. Good luck and give this plan your all, it will definitely pay off! Like Mike said, HAVE FUN WITH IT!!! ************************************************************************************* THE NAMES LIST THE NAMES LIST THE NAME LIST ************************************************************************************* * HONESTY IS WHAT MAKES THIS PROGRAM SUCCESSFUL!!! * * 1. John F. Bourgeois * 3043 New Oak Lane * Bowie, MD 20716 * * 2. David Maxwell * 1747 Stonybrook Ln. * APT 103 * Brunswick, OH 44212 * * 3. Daniel Henry * 3122-A Dagger Drive * Tyndall AFB, FL 32403-1204 * * 4. Jack Hartsfield * 1900 North Vine #404 * Los Angeles, California 90068 USA * * 5. John Ocava * 1125 N. Harvard Cir. * South Elgin, IL 60177 USA * ************************************************************************************* NOTE: Try to keep a list of everyone that sends you a dollar and always keep an eye on the local postings of this file... Just to make sure that everyone is playing the game fairly. You know where your name should be..... *** AGAIN, HONESTY IS THE BEST THING WE HAVE GOING FOR US ON THIS PLAN. -Mike Dotson, Boulder, CO *** By the way, if you try to deceive people by posting the messages with your name in the list and not sending the money to the people already included, you will not get much. I know someone who did this and only got about $150 (and that's after two months). Then he sent the 5 bills, people added him to their lists, and in 4-5 weeks he had over $10000! All the lists are re-distributed as soon as the money is received. end of article
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Exceptions Date: 24 Jul 1996 15:46:31 GMT Organization: ALI Message-ID: <4t5ggn$t3o@cetus.ali.bc.ca> References: <4t3mj9$oal@news.onramp.net> <4t4rjp$rfv@ddfservb.technet.net> Cc: szallies@energotec.de In comp.sys.next.programmer Constantin Szallies wrote: > dekorte@suite.com (Steve Dekorte) wrote: > > > >How do I avoid losing the stack when an exception is raised? > >(it would make debugging alot easier) > > I already asked myself this question a couple of times. Yes, it's not very > nice to get a message like "Uncaught exception xxxxx" and you don't know were > the exception was raised. > > But think about it --- it's not possible to keep the stack because you don't > know in advance if an exception will be caught or not. The only possibility > would be to copy the whole stack when a first exception it raised and to > remove that copy if the application falls of the end of some exception > handling domain. ---> expensive > This is true not only for Openstep exceptions. > > Like always --- you win some, you loose some. Type the following in the gdb window: b [NSException raise] this will put a break point at the beginning of the NSException class's raise method, and thus you can see the context of the upcoming exception. Note that you can't put this in your .gdbinit file since it only works after the symbols for the program have been loaded (which is done after .gdbinit is pulled in). You can put it in your PB.gdbinit file, but be prepared to replace it from time to time when ProjectBuilder rewrites the file. -- 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 * "C++ is to C as Lung Cancer is to Lung."
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: 24 Jul 1996 15:05:57 GMT Organization: Department of Electronics, University of York, UK. Message-ID: <4t5e4l$ign@netty.york.ac.uk> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4t1eeb$e7@saturn.genoa.com> On 23 Jul 1996 02:46:35 GMT, Alex Blakemore <alex@genoa.com> wrote: > In <4t01uh$66t@netty.york.ac.uk> Roger Peppe wrote: > > yes, it improves the ability of the compiler to statically typecheck, > > but IMHO it loses many of the original reasons for the existence > > of objective-C as a language. > > In what way? [...] > There are times where you'll have to do a little extra work to turn off a warning > now. warnings are not there to be ignored. in fact, when trying to build a portable application that will work indefinitely into the future, silencing type warnings like these is definitely not a good idea. it might work now, but that is absolutely no guarantee that it will work in the future (especially when openstep might have many implementations). if you do want to build portable, future-proof applications (and i do my best to) then abusing the type system is not really an option. > Why not try it for a few months before deciding whether its an improvement? i've already seen examples that made me cringe; i doubt it'll get any better. what ever happened to the mentality that looks for *good* and *appropriate* solutions rather than picking an ad hoc solution and imposing that regardless ? such a pity. rog.
From: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Java Bashing (Objective-C features) Date: Wed, 24 Jul 1996 13:12:50 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <QlxZaGK00iV984KKJ6@andrew.cmu.edu> References: <9607240959.AA01986@tlpnl.tlp.nl> In-Reply-To: <9607240959.AA01986@tlpnl.tlp.nl> Excerpts from netnews.comp.sys.next.programmer: 24-Jul-96 Java Bashing (Objective-C f.. by Maurice Le Rutte@tlpnl.t > That was my idea to. Every compiler uses byte codes before making machine > specific instructions, don't they? No. Some compilers, which are intended to be portable (or retargetable), generate what's known as 'intermediate code', which is a 'linearized representation' of the abstract syntax tree which was generated by the front end of the compiler system. Typically, this intermediate code is a variant of what's called 'three-address code', and consists of statements of the form: x := y _op_ z This intermediate code is then optimized, and the compiler then translates this intermediate representation into the target platform's assembly language (probably with additional peephole and instruction reordering optimizations happening afterwards). There are similarities between intermediate code and bytecodes, but there are some major differences, too. The biggest one is that intermediate code is intended to be converted to a real target assembly language, and is generated using assumptions that do not correspond to any (practical) virtual machine. For example, the intermediate code generator assumes that there will be an infinite number of registers available, and it is only after liveness analysis, register reuse, and if needed, register spilling into temporary areas on the stack that the code is runnable under a real environment (whether that is a virtual machine or native opcodes for the target platform). And there are other intermediate representations of the original source code that compilers use, like abstract syntax trees, postfix notation, or DAG's, which are obviously not intended for execution. Consult Chapter 8 of "Compilers: Principles, Techniques, and Tools" by Aho, Sethi, and Ullman-- it discusses intermediate code generation in detail. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: woo@polaris.scicntr.ortn.edu (J. W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Two "outlets" in a palette? Date: 24 Jul 1996 19:19:23 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4t5svr$otp@stc06.ctd.ornl.gov> I've been trying to make a palette object from a joy stick. I'd like to be able to drag the joystick into a window, then drag a connection from a text field object to the joystick and get the X value of the joystick, and another connection from another text field object to the joystick and get the Y value of the joystick. As the joystick moves, then the X and Y values show up in the textfield windows. I'd like it to be a palette object so that it functions in IB. Of course then I could connect the X and Y values directly to other objects instead of textfields. I add methods to the class JoyStick and don't see them show up as methods in IB when I reparse them and try to connect. Why? What is special about the Control class or Cell class that I can't add new methods to a subclass and see them show up? I see no way to have a palette object have more than one target for its actions, i.e. no way to have something receive the X value and another something receive the Y value. What is the trick here? John W. -- NewsGrazer, a NeXTstep(tm) news reader, posting -- M>UQR=&8P7&%N<VE[7&9O;G1T8FQ<9C!<9FUO9&5R;B!#;W5R:65R.WT*7&UA M<F=L,3(P"EQM87)G<C$R,`I<<&%R9%QT>#$Q-3)<='@R,S`T7'1X,S0U-EQT M>#0V,#A<='@U-S8P7'1X-CDQ,EQT>#@P-C1<='@Y,C$V7'1X,3`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`@ M07,@=&AE(&IO>7-T:6-K(&UO=F5S+"!T:&5N('1H92!8(&%N9"!9('9A;'5E M<R!S:&]W('5P(&EN('1H92!T97AT9FEE;&0@=VEN9&]W<RX@($DG9"!L:6ME M(&ET('1O(&)E(&$@<&%L971T92!O8FIE8W0@<V\@=&AA="!I="!F=6YC=&EO M;G,@:6X@24(N("!/9B!C;W5R<V4@=&AE;B!)(&-O=6QD(&-O;FYE8W0@=&AE M(%@@86YD(%D@=F%L=65S(&1I<F5C=&QY('1O(&]T:&5R(&]B:F5C=',@:6YS M=&5A9"!O9B!T97AT9FEE;&1S+EP*7`H*22!A9&0@;65T:&]D<R!T;R!T:&4@ M8VQA<W,@2F]Y4W1I8VL@86YD(&1O;B=T('-E92!T:&5M('-H;W<@=7`@87,@ M;65T:&]D<R!I;B!)0B!W:&5N($D@<F5P87)S92!T:&5M(&%N9"!T<GD@=&\@ M8V]N;F5C="X@(%=H>3\@(%=H870@:7,@<W!E8VEA;"!A8F]U="!T:&4@0V]N M=')O;"!C;&%S<R!O<B!#96QL(&-L87-S('1H870@22!C86XG="!A9&0@;F5W M(&UE=&AO9',@=&\@82!S=6)C;&%S<R!A;F0@<V5E('1H96T@<VAO=R!U<#\* M7`I<"DD@<V5E(&YO('=A>2!T;R!H879E(&$@<&%L971T92!O8FIE8W0@:&%V M92!M;W)E('1H86X@;VYE('1A<F=E="!F;W(@:71S(&%C=&EO;G,L(&DN92X@ M;F\@=V%Y('1O(&AA=F4@<V]M971H:6YG(')E8V5I=F4@=&AE(%@@=F%L=64@ M86YD(&%N;W1H97(@<V]M971H:6YG(')E8V5I=F4@=&AE(%D@=F%L=64N("!7 E:&%T(&ES('1H92!T<FEC:R!H97)E/UP*7`I*;VAN("!7+@I]"F%L `
From: don@misckit.com Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 24 Jul 1996 19:35:07 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4t5ttb$lpl@news.xmission.com> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <31F2DDB3.388C@freemansoft.com> <4t51tg$bfj@bignews.shef.ac.uk> mmalcolm crawford <m.crawford@shef.ac.uk> wrote: > Well, WebObjects is slated to support compiled Java later this year, > and one of the news stories I saw recently (can't find it now -- > anyone?) quoted a NeXT VP as saying that there was good stuff going > on in Engineering with Java... so who knows? Yeah. Wouldn't it be great if it was the Obj-C to Java bytecode thing? I keep hearing unsubstantiated rumors both ways...some saying they are doing it and others saying they aren't. Since I don't know which it really is, I'll guess I'll just have to wait and see. (But this wait is killing me! NeXT, tell us what's up your sleeve! :-) ) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: root@phcs.phcs.com (S00POR UZER) Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <4t5dc2$68s@phcs.phcs.com> Control: cancel <4t5dc2$68s@phcs.phcs.com> Supersedes: <4t5dc2$68s@phcs.phcs.com> Date: 24 Jul 1996 16:14:35 -0400 Organization: Private Healthcare Systems, Inc Message-ID: <4t607b$nik@phcs.phcs.com> SPAM cancelled by jq@phcs.com
Newsgroups: comp.lang.objective-c,comp.sys.next.programmer From: stes@cwi.nl (David Stes) Subject: Re: Java Bashing (Was: OpenStep design complaints) Message-ID: <Dv2C0I.JuA@cwi.nl> Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam References: <TMB.96Jul21012243@best.best.com> <Duw3vz.H9L@cwi.nl> <wlxHcwa00iWUA5wE1N@andrew.cmu.edu> Date: Wed, 24 Jul 1996 19:41:05 GMT In article <wlxHcwa00iWUA5wE1N@andrew.cmu.edu> Charles William Swiger <cs4w+@andrew.cmu.edu> writes: >Excerpts from netnews.comp.sys.next.programmer: 21-Jul-96 Re: Java >Bashing (Was: Open.. by David Stes@cwi.nl >> I agree. One the one hand, at some point, we want to benefit from >> the dynamic features of a language like Smalltalk, but that will not pay, >> of course, if somebody else has done "everything in C" and 't runs 100 >> times faster or so. > >Do you have any evidence to suggest every realistic application will run >100 times faster in C then in a dynamic language? Well, I can only quote evidence from my particular domain... Big integer division, for example, is a bottle-neck in most polynomial/symbolical computations with arbitrary precision integer coefficients. You can start with simple examples, but coefficients quickly tend to explode in size and time tends to be spend in doing operations of the type A - q B. I want to be able to write a system at a high level (you seem to insinuate at some point that I WANT to do everything in C, which is not the case !), but I do want at the same time to be able to use loops in plain C. That is why I'm wrapping up C code into Objective C objects, users don't see that internally they are implemented in C. There are hundreds of examples ! Most group theoretical stuff is best implemented in a dynamic GC language, but time will be spent in the end in computing commutators or inverses or so of permutations, large arrays of points; these operations NEED to be implemented in C : they are stupid operations like for(i=0;i<n;n++) c[i] = b[a[i]]; Another example is computing determinants of a matrix (actually that is where the number 100 that I'm quoting above comes from). If I computer a Determinant of a 100x100 over Z/p, p machine size integer, then Objective C with allocating all over the integers and messaging is going to be _slightly_ slower, no ? These are extremely important operations for symbolic computation systems. I am simply claiming that symbolic computation SHOULD be written in a dynamic, higher level language (than C), but that it is absolutely necessary to retain the ability to link in legacy C code, and C code that is necessary for time critical loops. Swiger: I'm challenging you back : show me the language that can optimize to the point , that you can implement integer arithmetic in it, performing comparably to the Integer object in my toolkit. You can't because, because if you were, the community that specializes in C optimizers would be out of business. It seems that you are claiming that programs can be implemented at a single level, and that "compilers will grow smarter". It's like saying that the whole of Unix should be implemented as a /bin/sh script ! It's evident : some filters can be best implemented in C, and other filters will rather be /bin/sh scripts that build upon the former kind... In Objective-C it's the same thing : some objects NEED to be implemented in C, and at some point you start programming in "pure Smalltalk" style. YOU will have to come up with evidence that 't isnt true. :-) David.
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: OpenStep NSRunAlertPanel Date: 24 Jul 1996 21:39:35 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t656n$gfe@news.onramp.net> Anyone else had problems getting NSRunAlertPanel() to work on OpenStep for Mach? Steve -- Steve Dekorte - OpenStep Developer - Anaheim, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: k_harbour@mail.bogo.co.uk (Karl Harbour) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Wed, 24 Jul 1996 23:06:29 GMT Organization: Orbital Computer Consultancy Message-ID: <31f6ab69.17593120@news.plsys.co.uk> References: <Dun1wy.FLB@cam-ani.co.uk> <Dun4zx.Fv0@cam-ani.co.uk> <4sggsg$3v@news.xmission.com> <acwx04ca9t.fsf@parka.winternet.com> <4sgvv7$hkn@dfw-ixnews5.ix.netcom.com> <4shag4$aiv@news.onramp.net> <4sibdd$26i@news.xmission.com> <4sjqv2$r9@news.onramp.net> <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <MlwFqfq00iV_M2BWNK@andrew.cmu.edu> <4sra5b$26i@news.xmission.com> Has no-one looked at what NetSurfer is up to? Sounds good to me. I'm sure the Java camp will argue what they're up to isn't portable, to which I would say two things: 1) The Evil Empire is about to have its wicked way with Java 2) One of NetSurfer's web pages says their stuff is available on Windows. -- Everything I've said, my opinion only. I reserve the right to be wrong.
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: 25 Jul 1996 00:05:21 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t6do1$gfe@news.onramp.net> References: <TMB.96Jul21012243@best.best.com> <Duw3vz.H9L@cwi.nl> <wlxHcwa00iWUA5wE1N@andrew.cmu.edu> <Dv2C0I.JuA@cwi.nl> (this is off the orginal topic) David Stes wrote: > In Objective-C it's the same thing : some objects NEED to be implemented > in C, and at some point you start programming in "pure Smalltalk" style. Unfortunately almost all 'Objective-C' programmers take the approach of: "I'll write everything in C and make it OO where I have to" (and usually don't even do that - they just produce alot of C functions that happen to be formatted as methods) instead of: "I'll make everthing OO and use raw C where I need the speed" (Which is what David has done a good job of in AlgebraKit) Steve -- Steve Dekorte - OpenStep Developer - Anaheim, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: mpaque@pacbell.net (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: GDB in OpenStep Date: Thu, 25 Jul 1996 00:30:23 GMT Organization: Electronics Service, Unit No. 16 Message-ID: <4t6f7n$9lk@news1.snfc21.pacbell.net> References: <4sokpv$ikh@news.onramp.net> dekorte@suite.com (Steve Dekorte) wrote: >Does the commercial release of OpenStep have a working gdb? Yes. (On Windows NT, for most values of 'working'... It's still in beta there. With all that code, and what it has to do, there's bound to be a bug or two in there that Cygnus or NeXT haven't uncovered or chased out yet. It certainly works well enough for me to do my job every day. It's certainly better than WinDbg, and more powerful than Microsoft's IDE debugger.] 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: Charles William Swiger <cs4w+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Wed, 24 Jul 1996 21:30:18 -0400 Organization: Fifth yr. senior, Computer Science, Carnegie Mellon, Pittsburgh, PA Message-ID: <glxgsee00iWR8_knJb@andrew.cmu.edu> References: <TMB.96Jul21012243@best.best.com> <Duw3vz.H9L@cwi.nl> <wlxHcwa00iWUA5wE1N@andrew.cmu.edu> <Dv2C0I.JuA@cwi.nl> In-Reply-To: <Dv2C0I.JuA@cwi.nl> Excerpts from netnews.comp.sys.next.programmer: 24-Jul-96 Re: Java Bashing (Was: Open.. by David Stes@cwi.nl >> Do you have any evidence to suggest every realistic application will run >> 100 times faster in C then in a dynamic language? > > Well, I can only quote evidence from my particular domain... Big integer > division, for example, is a bottle-neck in most polynomial/symbolical > computations with arbitrary precision integer coefficients. You can > start with simple examples, but coefficients quickly tend to explode in > size and time tends to be spend in doing operations of the type A - q B. Have you ever done, or even seen, timing comparisions between a C implementation and a truly dynamic implementation (in Java, or Smalltalk, or what have you) of one of your problems? Besides which, pure numeric computation is one of the most hardware dependant areas you could possibly have named. There is this concept of "choosing a language appropriate to a problem domain", you know. > I want to be able to write a system at a high level (you seem to insinuate > at some point that I WANT to do everything in C, which is not the case !), > but I do want at the same time to be able to use loops in plain C. Okay, fine. What's the big problem using Java's native call interface, then? [ ... ] > Swiger: I'm challenging you back : show me the language that can > optimize to the point , that you can implement integer arithmetic in it, > performing comparably to the Integer object in my toolkit. > You can't because, because if you were, the community that specializes in > C optimizers would be out of business. Nonsense! Ever heard of Fortran, Stes? You know, the language which still has the widest range of high-performance numerical and mathematical routines available? > It seems that you are claiming that programs can be implemented at a > single level, and that "compilers will grow smarter". > It's like saying that the whole of Unix should be implemented as a /bin/sh > script ! [ ... ] Of course not. You should select one or more language(s) appropriate for the task. A classic example is a large-scale system like a news server or a web server, which typically have everything from C, Perl, /bin/sh, Java, HTML, and who knows what other languages around. -Chuck Charles Swiger | cs4w@andrew.cmu.edu | standard disclaimer ----------------+---------------------+--------------------- I know you're an optimist if you think I'm a pessimist.
From: mpaque@pacbell.net (Mike Paquette) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: Thu, 25 Jul 1996 00:29:59 GMT Organization: Electronics Service, Unit No. 16 Message-ID: <4t6f72$9lk@news1.snfc21.pacbell.net> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4s6ebc$32g@saturn.genoa.com> <4t01uh$66t@netty.york.ac.uk> rog@ohm.york.ac.uk (Roger Peppe) wrote: >On 12 Jul 1996 20:59:24 GMT, Alex Blakemore <alex@genoa.com> wrote: >> b. The argument tests in OpenStep are compile time checks only. Thus if you wish >> to pass a MySpecialArray to a method that expects a NSArray, you can do so. The >> compiler will warn you about it, and you can silence the compiler warning if you >> wish with a caste. If the method only sends public API messages to its argument, >> and you class behaves reasonably like the one its pretending to be, this works. >surely if you do this, then any subsequent behaviour is undefined, >because the writer of the openstep class is perfectly at liberty >to put in a line like : > assert([obj isKindOf:[NSDictionary class]]); @interface MySpecialClass(EvilHacks_Part_2) - (BOOL) isKindOfClass:(Class)cls; @end @implementation MySpecialClass(EvilHacks_Part_2) - (BOOL) isKindOfClass:(Class)cls { if ( cls == [NSString class] ) return YES; return [super isKindOfClass:cls]; } Bogus, but if you really intend to spoof a class without subclassing it, you might as well go whole hog. >(as NXBrowser does for its NXBrowserCell objects, for instance) >writing anything that invokes undefined behaviour is an absolute >no-go for apps which are going to be around for any length of time >at all, so the spec. has effectively stopped anyone from doing that >(even though it might still work in the current circumstances). >for this reason, i think that moving to static class typing rather >than id is *not* an overall improvement. >yes, it improves the ability of the compiler to statically typecheck, >but IMHO it loses many of the original reasons for the existence >of objective-C as a language. >the real solution would be to document the interfaces >well enough that you *could* plug and play classes implemented >from various different frameworks. >wasn't that what object-oriented programming was supposed to >give us ? Yup. Note that most of the statically typed classes in the OPENSTEP API are either Foundation abstract classes or AppKit classes. In the case of most AppKit classes, details of implementation and interaction are quite important (many of these classes are rather state-heavy), and an abstract API without mechanism would be less than useful. In the case of Foundation abstract classes, these classes serve as the root of class clusters. The abstract class specifies an API, just as a protocol would, and also defines a default implementation built on top of a relatively small set of 'primitive' methods. A subclass can either override only the primitive methods, inheriting the remainder of the defaut implementation, override primitive and selected other methods, or treat the abstract class interface as a protocol, and override all methods. For a somewhat different approach, take a look at the NEXTIME server-side component API (/NextLibrary/Frameworks/NEXTIME.framework/Headers). This API relies almost entirely on protocols. All components are expected to respond to certain sets of methods, but there is very little to be shared between the implementation of (for example) a JPEG and a RLE video decompressor. Given the anticipated lack of common code among implementations, a protocol based API appears to be more appropriate. It all boils down to pragmatism, or, "What possible solution addressed the most common problems for the most programmers?" I certainly won't claim it's a panacea. It certainly seems to upset purists from both the 'static everything' and 'dynamic everything' camps. All I can say is that the OPENSTEP approach does work in the real world. 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: Mahesh Saptharishi <mahesh+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer Subject: Accessing Parallel Port Date: Thu, 25 Jul 1996 00:11:40 -0400 Organization: Sophomore, Electrical and Computer Engineering, Carnegie Mellon, Pittsburgh, PA Message-ID: <ElxjDwK00iV_4_toVD@andrew.cmu.edu> How do I communicate with the parallel port especially the data, status and control register using BSD C under NeXT on Sun SparcStation 20? The program that I am trying to port uses inb() and outb(), but those functions seem to be unavailable when you are dealing with a sparc station. Any ideas? Thanks, -Mahesh
From: David.Hinz@mci.com (David Hinz) Newsgroups: comp.sys.next.programmer Subject: pthreads api for NS 3.3 Date: 25 Jul 1996 04:37:20 GMT Organization: InternetMCI Message-ID: <4t6tm0$5e@news.internetmci.com> Does anyone know if there is a pthreads library or front-end to the cthreads available for NS 3.3? Thanks for any info, dave. David Hinz David.Hinz@mci.com
From: croehrig@cs.ubc.ca (Chris Roehrig) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Does NEXTSTEP have a /dev/tun0 tunnel interface? Date: 25 Jul 1996 04:43:25 GMT Organization: Computer Science, University of B.C., Vancouver, B.C., Canada Message-ID: <4t6u1d$8a8@nnrp.cs.ubc.ca> I need a tunnel interface to be able to use the IPIP encapsulation daemon. (I want to use this for an IP-in-IP tunnel to bridge my subnet over an ISDN connection that routes only a single IP address.) The /dev/tun0 appears to be a network interface that can be opened as a stream by a user process to read and write raw IP packets to and from the kernel. Has anyone written such a beast? I'm using the ps_if virtual interface and can route my subnet traffic to it, but I don't know how to get a user-process (the ipip daemon) to read and write raw packets from that interface to encapsulate them. Is that why I need a /dev/tun0 interface? Am I missing something? -- Chris Roehrig (croehrig@House.ORG or croehrig@sns.cs.ubc.ca) Neuroscience and Computer Science at University of British Columbia, Vancouver http://www.House.ORG/chris http://www.sns.cs.ubc.ca/chris
From: don@misckit.com Newsgroups: comp.sys.next.programmer Subject: Re: Two "outlets" in a palette? Date: 25 Jul 1996 05:16:21 GMT Organization: XMission Internet (801 539 0900) Message-ID: <4t6vv5$lpl@news.xmission.com> References: <4t5svr$otp@stc06.ctd.ornl.gov> woo@polaris.scicntr.ortn.edu (J. W. Wooten) wrote: > I see no way to have a palette object have more than one target for > its actions, i.e. no way to have something receive the X value and > another something receive the Y value. What is the trick here? I'll tackle this part of the question. The "target" outlet is treated specially by IB. However, if you are willing to kludge things a little bit, you can use the super spiffy "MiscTee" class from the MiscKit. You connect to the tee class and then connect it to the two destinations. That's probably the easiest way to send dual messages like that. All the MiscTee does is take every message sent to it and pass the message on to each of its targets, one at a time. Note that you can use it to pass any explicit message on to all it's outputs (they all get the same message) or you can send a message to it that causes a different, preset message to be sent to each "client". The really nice thing about it is that it has a pretty nice GUI inspector that allows you to harness the flexibility in IB. This little toy is pretty useful in some circumstances! By the way, there is a subclass of MiscTee that can send messages across .nib file boundaries. I don't consider it a good idea to do that, but if you *must* do it, MiscDistributor will help... (It works in a way that is analogous to the EOF rendezvous points.) -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: Jon Rosen <jfrx@earthlink.net> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Java Bashing (Was: OpenStep design complaints) Date: Thu, 25 Jul 1996 00:53:59 +0000 Organization: Earthlink Network, Inc. Message-ID: <31F6C5A7.7A4@earthlink.net> References: <4skmcg$h4t@dfw-ixnews10.ix.netcom.com> <4sr017$94d@solaris.cc.vt.edu> <TMB.96Jul21012243@best.best.com> <Duw3vz.H9L@cwi.nl> <wlxHcwa00iWUA5wE1N@andrew.cmu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Charles William Swiger wrote: > > Do you know why? > > It's because modern C compilers can generate > better code (in terms of speed, size, and > correctness) than the average assembly programmer > can, and a C compiler can do this task several > orders of magnitude more quickly than even the > best assembly programmer can. And equally important is that the advantage grows significantly as the size of the source code and resulting machine code grows. I recently had someone tell me that he can still outcode a C compiler using assembly language and showed me a 19 line program in C in which indeed he wrote better code (about 35 instructions in his assembler code vs. about 52 instructions in the compiled routine). The problem is, there is no way he could extrapolate that proof to a 19,000 line program. Whereas a C compiler could convert that 19,000 line program to machine code in minutes at worst, seconds at best, the likelihood of him writing 35,000 lines of assembler in the time it would take to write the 19,000 lines of C and have it execute properly, is almost nil, IMHO (and I was an assembler programmer for 12 years and wouldn't go back if you paid me twice what I make today). Jon Rosen
From: Constantin Szallies <szallies@energotec.de> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Exceptions Date: 25 Jul 1996 08:44:03 GMT Organization: Tech Net GmbH Message-ID: <4t7c4j$fjl@ddfservb.technet.net> References: <4t3mj9$oal@news.onramp.net> <4t4rjp$rfv@ddfservb.technet.net> <4t5ggn$t3o@cetus.ali.bc.ca> allan@ali.bc.ca (Allan Noordvyk) wrote: >In comp.sys.next.programmer Constantin Szallies wrote: >> dekorte@suite.com (Steve Dekorte) wrote: >> > >> >How do I avoid losing the stack when an exception is raised? >> >(it would make debugging alot easier) >> >> I already asked myself this question a couple of times. Yes, it's not very >> nice to get a message like "Uncaught exception xxxxx" and you don't know >were >> the exception was raised. >> >> But think about it --- it's not possible to keep the stack because you >don't >> know in advance if an exception will be caught or not. [CUT] >Type the following in the gdb window: > > b [NSException raise] > >this will put a break point at the beginning of the NSException >class's raise method, and thus you can see the context of >the upcoming exception. > [CUT] Yes, in the debugger! There were times one could debug a crashed application what was not running in the debugger by examining the core dump! And by the way, with b [NSException raise] the debugger stops whenever an exception is raised. What a programmer would like is the debugger to stop when an uncaught exception is raised!! (Although it's considered a bad programming style to have exceptions raised in the normal control flow) -- Constantin Szallies, Energotec GmbH szallies@energotec.de 49211-9144018
From: Erik Doernenburg <erik@object-factory.com> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Exceptions Date: 24 Jul 1996 12:16:04 GMT Organization: Object Factory GmbH (Germany) Message-ID: <4t5464$b1@isabella.object-factory.com> References: <4t3mj9$oal@news.onramp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit dekorte@suite.com (Steve Dekorte) wrote: > > How do I avoid losinng the stack when an exception is raised? > (it would make debugging alot easier) > (gdb) break [NSException raise] ---------------------------------------------------------------------------- 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: Holger Hoffstaette <hhoff@mvmg.de> Newsgroups: comp.sys.next.programmer Subject: Re: pthreads api for NS 3.3 Date: Thu, 25 Jul 1996 12:15:46 +0200 Organization: Modern Video Media Group Message-ID: <31F74952.3B9D@mvmg.de> References: <4t6tm0$5e@news.internetmci.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit To: David Hinz <David.Hinz@mci.com> David Hinz wrote: > Does anyone know if there is a pthreads library or front-end to the > cthreads available for NS 3.3? Thanks to Darcy Brockbank and the HASC people: http://www.hasc.ca/ I don't know the exact path, snoop around a little.. Holger -- Holger Hoffstätte (hhoff@mvmg.de) Modern Video Media Group Stuttgart, Germany
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep NSRunAlertPanel Date: Thu, 25 Jul 1996 12:14:01 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <Dv3LzE.MKq@cam-ani.co.uk> References: <4t656n$gfe@news.onramp.net> In article <4t656n$gfe@news.onramp.net> dekorte@suite.com (Steve Dekorte) writes: > > Anyone else had problems getting NSRunAlertPanel() to work on OpenStep for Mach? > No. The only issue I've had in this area is that Code kinda like NXRunAlertPanel(NULL,text,b1,b2,b3); gets converted to NSRunAlertPanel(NULL, [NSString stringWithCString:text] ,[NSString stringWithCString:b1] ,[NSString stringWithCString:b2] ,[NSString stringWithCString:b3]); which breaks if b? is NULL, because NSString dereferences the NULL pointer. If you've got this in some kinda wrapper, it's quite common to pass in NULL for some of the buttons, as that's how NXRun... specifies that buttons don't exist. You just have to use NSRunAlertPanel(NULL, [NSString stringWithCString:text] ,b1?[NSString stringWithCString:b1]:NULL ,b2?[NSString stringWithCString:b2]:NULL ,b3?[NSString stringWithCString:b3]:NULL ); b1?[NSString stringWithCString:b1]:NULL if you might need this kind of behaviour. In this particular instance it might be better if [NSString stringWithCString:NULL]==NULL but it doesn't (it rasies an exception). I'm prepared to belive that there are reasons why this would be bad for other bits of code. $an
From: amoster@cube.de (Andreas Moster) Newsgroups: comp.sys.next.programmer Subject: Problem with character set of Oracle Date: 25 Jul 1996 11:54:42 GMT Organization: Cube Informationssysteme GmbH Message-ID: <4t7na2$jg5@cubenx.cube.de> Hi! I want to store german umlauts in a Oracle 7.2 database. I use the default orcl database installed on NT with the character set WE8ISO8859P1 and WE8NEXTSTEP. Both failed. With the WE8ISO8859P1 i can insert the german characters and they are well shown with the select command in sqlplus, but the result from the sqlexecutor(demo program) are some "strange" characters. The WE8NEXTSTEP has also "strange" results. Any hints? Bye, Andreas. -- Andreas Moster, [kju:b] Cube Informationssysteme GmbH Hessbruehlstrasse 15, 70565 Stuttgart, Germany fon 0711/90669-0, fax 0711/90669-33, email: amoster@cube.de private site: Vogelsangstr. 19, 75331 Engelsbrand, Germany fon 07235/3320, fax 07235/980074, email: am@hal.ka.sub.org NeXT, MIME mail, PGP welcome!
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc,misc.jobs.offered Subject: NEXTSTEP/Career Position/ILL Date: 25 Jul 1996 14:21:37 GMT Organization: Norden 1 Communications Message-ID: <4t7vth$8jk@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------Commercial experience Objective C-----------------Commercial experience EOF-------------------------A Plus Sybase or Oracle------------A plus Career Position-------------Outstanding benefits and opportunity Must be---------------------US Citizen or Greencard holder Area------------------------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: Markus Strickler <markuss@gauss.fo.fh-koeln.de> Newsgroups: comp.sys.next.programmer Subject: PHP for NeXTStep Date: Thu, 25 Jul 1996 15:40:25 +0200 Organization: FH-Köln Message-ID: <31F77949.121F@gauss.fo.fh-koeln.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi! Has anyone out there managed to port PHP to NeXTStep? If so which changes have to be made? This is a log of the compilation errors: cc -I/usr/include/ansi -O -I. -I/www/postgres95/include -DHAVE_LIBPQ=1 -I/www/apache_1.0.5/src -DACCESS_CONTROL -DACCESS_DIR=\"/www/apache_1.0.5/cgi-da ta\" -DLOGGING=1 -DLOG_DIR=\"/www/apache_1.0.5/cgi-data\" -DFILE_UPLOAD -DAPACH E=1 -c local.c local.c: In function `strdup': local.c:87: warning: type mismatch with previous external decl /usr/include/ansi/stdlib.h:62: warning: previous external decl of `malloc' local.c:87: warning: extern declaration of `malloc' doesn't match global one cc -I/usr/include/ansi -O -I. -I/www/postgres95/include -DHAVE_LIBPQ=1 -I/www/apache_1.0.5/src -DACCESS_CONTROL -DACCESS_DIR=\"/www/apache_1.0.5/cgi-da ta\" -DLOGGING=1 -DLOG_DIR=\"/www/apache_1.0.5/cgi-data\" -DFILE_UPLOAD -DAPACH E=1 -c dns.c In file included from dns.c:29: /NextDeveloper/Headers/bsd/netdb.h:15: redefinition of `struct hostent' /NextDeveloper/Headers/bsd/netdb.h:28: redefinition of `struct netent' /NextDeveloper/Headers/bsd/netdb.h:35: redefinition of `struct servent' /NextDeveloper/Headers/bsd/netdb.h:42: redefinition of `struct protoent' /NextDeveloper/Headers/bsd/netdb.h:48: redefinition of `struct rpcent' *** Exit 1 Stop. Maybe someone has a clue? Markus -- Markus Strickler Student of Image Science FH Köln, Cologne Germany email: markuss@gauss.fo.fh-koeln.de
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: Re: Identifying Type Date: 25 Jul 1996 02:49:50 GMT Organization: Genoa Software Systems Message-ID: <4t6nce$1j8@saturn.genoa.com> References: <4rrcbd$i3i@news.ida.org> <4rsfqe$fsl@news.onramp.net> <4t1eeb$e7@saturn.genoa.com> <4t5e4l$ign@netty.york.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: rog@ohm.york.ac.uk I wrote: >>There are times where you'll have to do a little extra work to turn off a warning Roger Peppe wrote: > warnings are not there to be ignored. Absolutely. I always advocate a zero warning policy for shipping code. > in fact, when trying to build a portable application > that will work indefinitely into the future, silencing > type warnings like these is definitely not a good idea.... > abusing the type system is not really an option. When I said "do a little extra work to turn off a warning", I meant to define objects (and methods) with explicit types or protocols, rather than use typeless ids. That's the correct way to silence the warnings IMHO. And that's not abusing the type system, that's embracing it. > what ever happened to the mentality that looks for *good* and > *appropriate* solutions rather than picking an ad hoc solution and > imposing that regardless ? That's a pretty extreme view of this change. Some of us at least find that the more explicit use of static typing is an improvement. At the worst, you'll have take a little more time to use types in your definitions. More typing, and a little more thought are required. In return, you'll get more maintainable and likely more robust software, and certainly more help from the compiler in catching errors. Sounds like a good trade to me. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Exceptions Date: 25 Jul 1996 04:25:31 GMT Organization: Genoa Software Systems Message-ID: <4t6svr$1j8@saturn.genoa.com> References: <4t3mj9$oal@news.onramp.net> <4t4rjp$rfv@ddfservb.technet.net> <4t5ggn$t3o@cetus.ali.bc.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: allan@ali.bc.ca In <4t5ggn$t3o@cetus.ali.bc.ca> Allan Noordvyk wrote: > Type the following in the gdb window: > b [NSException raise] > this will put a break point at the beginning of the NSException > class's raise method, and thus you can see the context of > the upcoming exception. >... you can't put this in your .gdbinit file since it only > works after the symbols for the program have been loaded It works in .gdbinit if you use the future-break (fb) command instead. future-break [NSException raise] -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@colds.ruca.ua.ac.be> Date: Thu, 25 Jul 96 16:49:47 +0200 From: Raf Schietekat <flexus!rfschtkt@colds.ruca.ua.ac.be> Message-ID: <9607251449.AA04326@flexus> Subject: make(1) VPATH behaviour: NEXTSTEP_Dev_3.2 vs. GNU make-3.74 Dear experts, When make is called in NEXTSTEP_Dev_3.2, it apparently finds a match in the VPATH path only on demand. This makes it possible to store .o files in a directory m68k_debug_obj/, where they are subsequently found when an application is linked. However, GNU make-3.74 apparently finds a match eagerly. If an application relies on myFile.o, and it is not yet to be found anywhere in VPATH, then make fixes its mind on just myFile.o, disregarding VPATH for that file afterwards. It may then generate myFile.o somewhere on VPATH, and when it comes to the part that uses myFile.o, it will just use myFile.o as it had remembered the location. Workaround: execute make two times in a row (or more, if the generation chain length is longer, cf. .psw -> .c -> .o -> .app). Which behaviour is wrong? How (if at all) do later NeXT development environments (using GNU make) handle the separation of .o files for different architectures (m68k/i386/...) and different configurations (debug/profile/...)? If the behaviour is again like in NEXTSTEP_3.2 (don't know), is that because NeXT cured a bug, which will also be cured in the main GNU release, or because it chooses to do its own thing? I can't use NEXTSTEP_Dev_3.2's make because it crashes. 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: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: GDB in OpenStep Date: 25 Jul 1996 18:09:23 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t8d8j$gfe@news.onramp.net> References: <4sokpv$ikh@news.onramp.net> <4t6f7n$9lk@news1.snfc21.pacbell.net> Cc: mpaque@pacbell.net Mike Paquette wrote: > Steve Dekorte wrote: > > >Does the commercial release of OpenStep have a working gdb? > > Yes. > > (On Windows NT, for most values of 'working'... It's still in beta > there. .. But does it work with Edit as in NS3.3? Steve -- Steve Dekorte - OpenStep Developer - Anaheim, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: wilkie@cg.tuwien.ac.at (Alexander Wilkie) Newsgroups: comp.sys.next.programmer,comp.sys.next.hardware Subject: 3D device drivers, anyone? Date: 25 Jul 1996 21:50:02 GMT Organization: Vienna University of Technology, Austria Message-ID: <4t8q6a$6aq@news.tuwien.ac.at> Hi newsgroups! There's something that's been on my mind for quite some time now: is it possible to integrate the "polygon spitting" capabilities of graphics adapters like the MGA Millenium (or Diamond Edge, or Elsa Gloria, or...) into the NeXTStep environment? Would a NeXTStep solution work with Openstep/Mach? I'd guess that for any sensible benefit (ie one 3D viewer app working with more than one kind of 3D capable graphics card) to come from hacking in this area one would have to define something like a generic "OpenGL" or just "polygon spitter" device class that drivers linking to the specific features of the appropriate adapters can subclass. About this I've got four questions: a) Has anyone tried / is trying to do this (any Misc3DKit folk listening)? b) Is it possible for a second driver (apart from the main display) to write to the screen directly in an ordered manner, ie the 3D functionality cooperating with the windowserver with respect to window clipping etc? Judging by the fact that there are some TV cards + drivers available this might work. c) Can two drivers sit "on top" of the same graphics card (3D & normal display) or would one have to write a new dual functionality driver? d) If anything special like Interceptor developer were needed, would NeXT Inc. perhaps provide that if one could prove to be serious about a non-profit attempt to delvelop something like this? If this is at all possible, there is a good chance that I'd be able to motivate some people here at the TU to get together to do it; having something like this would turn NS/Intel PCs into (admittedly low-end-of-the-genetic-pool) graphics workstations, which is always better than no graphics workstation at all :-) Please comment! Alexander Wilkie -- e-mail: wilkie@cg.tuwien.ac.at (NeXTMail preferred, MIME o.k.) www : http://www.cg.tuwien.ac.at/~wilkie/
From: shess@subzero.winternet.com (Scott Hess) Newsgroups: comp.sys.next.programmer Subject: Projects not compiling when moved from OS/Mach to OS/NT Date: 25 Jul 1996 18:03:56 -0500 Organization: Is a sign of weakness Sender: shess@subzero.winternet.com Message-ID: <aclog86ilf.fsf@subzero.winternet.com> I've been attempting to move a project from OS/Mach to OS/NT. Since the Mach environment is nicer to work in, I want to try doing most development there, and then move the code over and rebuild. Unfortunately, when I do so, the code won't compile. make goes to work, and at some point it says "Unable to find file <classname>.m to create <classname>.o". [Sorry, that's a paraphrase, but if you've seen it you know what I mean.] Anyone know how to fix this? The release notes give no help. I ended up creating a new project and copying over each and every class manually. Then it worked. When I tar'ed the NT-created project up and moved it to Mach, changed .h and .m files, and moved it back, well, it didn't work again, same problem. ProjectBuilder seems to read everything fine, it's just make that gets wedged. Thanks, -- scott hess <shess@winternet.com> (WWW to "http://www.winternet.com/~shess/") Work: 12550 Portland Avenue South #121, Burnsville, MN 55337 (612)895-1208 <I want to become so famous that people buy tapes of me reading source code>
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc,misc.jobs.contract Subject: NEXTSTEP/Contract--Long term-Va Date: 26 Jul 1996 01:41:04 GMT Organization: Norden 1 Communications Message-ID: <4t97ng$7r4@tofu.alt.net> Programmer/analyst/developer NEXTSTEP--------------------Commercial experience Objective C-----------------Commercial experience EOF-------------------------A Plus Sybase or Oracle------------A Plus Contract--------------------Long Term Area------------------------Greater DC Area Must Be---------------------US Citizen or Greencard holder To Be Considered------------Fax resume or mail a hard copy. -- ehutch@norden1.com (419) 893-6367 [fax] Omni Search (419) 893-6334 [voice] 1310 Craig Maumee, Ohio 43537
From: dekorte@suite.com (Steve Dekorte) Newsgroups: comp.sys.next.programmer Subject: Re: Projects not compiling when moved from OS/Mach to OS/NT Date: 26 Jul 1996 02:05:27 GMT Organization: OnRamp Technologies; ISP; Dallas/Ft Worth/Houston, TX USA Message-ID: <4t9957$ig4@news.onramp.net> References: <aclog86ilf.fsf@subzero.winternet.com> Cc: shess@subzero.winternet.com Scott Hess wrote: > ProjectBuilder seems to read everything fine, it's just make that gets wedged. Have you tried doing a diff on the working-NT vs. working-Mach makefiles? Also, if PB auto creates makefiles(instead of using them as archive files) you might be able to delete all your makefiles after untarring on another platform and then it will generate good ones when you go to do a build. Steve -- Steve Dekorte - OpenStep Developer - Anaheim, CA "Fundamentalism isn't about religion. It's about power." - S. Rushdie
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: FilterServices in OS4.0 Date: 26 Jul 1996 02:30:16 GMT Organization: University of Illinois at Urbana Message-ID: <4t9ajo$31e@vixen.cso.uiuc.edu> Can anyone offer any tips on debugging/testing an OS4.0 Filter Service? I swear I can't even tell if the thing is running or not. I've tried a main funcition that uses NSConnection/NSRunLoop,another that ran as an application that registered a service, and yet a third that used the NSRegisterServiceProvier() function like the NeXT docs instruct. As far as I can tell none of it is working at all. Thanks, Ian P. Cardenas -- Ian P. Cardenas (icardena@uiuc.edu) CCSO Sites Technical Support "Are they as successful as who,Microsoft? Only drug lords from South America are as successful as Microsoft." -Tim Byars, on the success of NeXT Computer, Inc.
From: ampriasm@students.wisc.edu (Andrew M. Priasmoro) Newsgroups: comp.sys.next.programmer Subject: NeXT Machine Won't Bootup. Date: Thu, 25 Jul 1996 23:17:01 -0400 Organization: University of Wisconsin-Madison Message-ID: <ampriasm-2507962317010001@f180-171.net.wisc.edu> Hi, My NeXT machine won't bootup after I tried to install PPP-2.2.0.4.6. Does anyone have pointers how I go to the command interface so that I can restore my original hostconfig and etc. At least I could see the files in etc directory and edit them using vi editor. Thanks in advance. Andrew. -- Andrew Priasmoro University of Wisconsin-Madison
From: wilcoxb@cs.colorado.edu (Bryce) Newsgroups: comp.sys.next.programmer Subject: compiling Ecash(tm) library for Nextstep Date: 26 Jul 1996 08:18:19 GMT Organization: University of Colorado, Boulder Message-ID: <4t9v0b$d77@lace.colorado.edu> NNTP-Posting-User: wilcoxb -----BEGIN PGP SIGNED MESSAGE----- Let's say a fellow had a mostly-ANSI, mostly-POSIX compliant library known as "Ecashlib" which had already been ported to Windoze and Mac without a hitch. Let's furthermore say that this fellow got access to a Nextstep box with gcc 2.7+ and gnumake 3.73+ already installed on it. How much trouble could this hypthetical guy expect in compiling the Nextstep version of his library? More to the point, does anybody have a box like that which a person like that could borrow access to? Thanks in advance! Regards, Bryce -----BEGIN PGP SIGNATURE----- Version: 2.6.2 Comment: www.c2.net/~bryce -- 'BAP' Easy-PGP v1.1b2 iQCVAwUBMfh/TPWZSllhfG25AQFsEAQAihaHjMdjFDChaiBk4tMkzuAO4p0z2uSk F+CO77Nd3t7p+JU1Viwc60t57JhfSEDlZ/b765qXOs9J+ypzY6x+D/bL7k5byncn ZMFcZ8yAbtVi6kNGmAUOk6PvB+j4mc1Nq+mwHXMYMJcaFUjf8cSEhKpCrHXZkyJJ 2QVcHbx9IZo= =+p+0 -----END PGP SIGNATURE-----
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@colds.ruca.ua.ac.be> Date: Fri, 26 Jul 96 13:18:45 +0200 From: Raf Schietekat <flexus!rfschtkt@colds.ruca.ua.ac.be> Message-ID: <9607261118.AA11484@flexus> Subject: Problems with /usr/bin/m4(1) Hello, I was trying to process an m4 source (not written by myself), and found out that the generated result was not at all what was expected. When I obtained, built and used GNU's m4-1.4, all was well. Being blissfully ignorant of m4 syntax, I was wondering whether any of you have experienced problems with NEXTSTEP_3.2's /usr/bin/m4, i.e., should I report this as a bug, or attribute this to the use of a GNU particularity that /usr/bin/m4 does not understand? If you're an m4 power user, ask me for the example, if you wish. 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: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: Projects not compiling when moved from OS/Mach to OS/NT Date: Fri, 26 Jul 1996 13:23:15 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <Dv5Jut.4pA@cam-ani.co.uk> References: <4t9957$ig4@news.onramp.net> > Scott Hess wrote: > > ProjectBuilder seems to read everything fine, it's just make that gets wedged. I don't supose these are relevant, but I've found a number of issues wrt to makefiles: 1)remember not to set SHELL=/bin/sh!! (stupid, but I've seen it. Obviously fails on NT, but not in an help fashion) 2) $(CLASSES:=.m) doesnt work right (should tag turn class list into source file list!) (works on Mach - fails NT) 3) xargs - can be useful in makefiles (arguments are evaultated at the wrong times - looses things like " ") (works on Mach - fails NT) 4) VPATH - $* does not become full filename (Can use $^ instead) (works on Mach ($* does = filename) - fails NT ) Despite these reservations, simple PB makefiles seem to be OK. These problems were encountered building a pretty bit system with lots of quirky build scripts. $an
From: Mahesh Saptharishi <mahesh+@andrew.cmu.edu> Newsgroups: comp.sys.next.programmer Subject: Driver programming?? Date: Fri, 26 Jul 1996 12:14:07 -0400 Organization: Sophomore, Electrical and Computer Engineering, Carnegie Mellon, Pittsburgh, PA Message-ID: <AlyCvDq00YUq0BmVs0@andrew.cmu.edu> Are there any simple libraries/functions/headers available out there that will help access the status, control and data registers of a parallel port on 1) a sparc 20 2) a PC I trying to write a driver for a peripheral that hooks onto a parallel port and I am stumped. Thanks, -Mahesh
Newsgroups: comp.sys.next.programmer From: brad@instep.bc.ca Subject: How difficult to use function keys? Message-ID: <1996Jul26.172304.996@instep.bc.ca> Sender: usenet@instep.bc.ca Organization: InStep Mobile Communications Inc. Date: Fri, 26 Jul 1996 17:23:04 GMT I've scoured NextDev, FAQ, and NeXTAnswers, and I can't find anything that deals with using keyboard function keys in a NeXTSTEP application. Anyone done this? thanks. -- Brad Head <brad@instep.bc.ca> __ __o Software Developer, InStep Mobile Communications Inc. ___ `\<, Atomic Racing Club, Vancouver, BC, Canada -- O/ O
From: uhendjx@racer (Jonathan W. Hendry) Newsgroups: comp.sys.next.programmer Subject: Re: Projects not compiling when moved from OS/Mach to OS/NT Date: 26 Jul 1996 20:41:24 GMT Organization: Lexis-Nexis, Dayton OH Message-ID: <4tbahk$biq@mailgate.lexis-nexis.com> References: <4t9957$ig4@news.onramp.net> <Dv5Jut.4pA@cam-ani.co.uk> Ian Stephenson (ians@cam-ani.co.uk) wrote: : I don't supose these are relevant, but I've found a number of issues wrt : to makefiles: : 1)remember not to set SHELL=/bin/sh!! : (stupid, but I've seen it. Obviously fails on NT, but not in an : help fashion) It won't necessarily fail, if you've installed the CygWin32 package. (It include bash, which the CygWin32 installation instructions tell you to copy to \bin\sh.) Granted, this is a wee small segment of the population. ;) (The CygWin32 package is recommended, by the way. It includes a pretty complete set of Unix-environment command-line tools: everything from ls and cp to gzip and ping. Makes life a little more tolerable when you can run bash in a Win95 Dos shell.) -- Jonathan W. Hendry Views expressed herein do Steel Driving Software, Inc. not represent those of steeldrv@ix.netcom.com Steel Driving Software, Inc. jon@exnext.com or Lexis-Nexis
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: How difficult to use function keys? Date: 26 Jul 1996 20:39:19 GMT Organization: ALI Message-ID: <4tbadn$f7@cetus.ali.bc.ca> References: <1996Jul26.172304.996@instep.bc.ca> Cc: brad@instep.bc.ca In comp.sys.next.programmer brad@instep.bc.ca wrote: > > I've scoured NextDev, FAQ, and NeXTAnswers, and I can't find anything that > deals with using keyboard function keys in a NeXTSTEP application. > > Anyone done this? Assuming you are referring to the F1 through F12 keys along the top of a PC keyboard, then YES. Note that you can't use F1 since it is reserved as the equivalent to the "Help" key on the later NeXT keyboards. As well, you can't enter these into Interface Builder directly, but here are the standard hex defines for the PC keys for programatic use with the setKeyEquivalent: method. #define F1_KEY 0x3b // Note: Reserved as NeXT help-key #define F2_KEY 0x21 #define F3_KEY 0x22 #define F4_KEY 0x23 #define F5_KEY 0x24 #define F6_KEY 0x25 #define F7_KEY 0x26 #define F8_KEY 0x27 #define F9_KEY 0x28 #define F10_KEY 0x29 #define F11_KEY 0x2a #define F12_KEY 0x2b bon chance... -- 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 * "C++ is to C as Lung Cancer is to Lung."
From: allan@ali.bc.ca (Allan Noordvyk) Newsgroups: comp.sys.next.programmer Subject: Re: How difficult to use function keys? Date: 26 Jul 1996 20:51:25 GMT Organization: ALI Message-ID: <4tbb4d$14m@cetus.ali.bc.ca> References: <1996Jul26.172304.996@instep.bc.ca> <4tbadn$f7@cetus.ali.bc.ca> Cc: brad@instep.bc.ca Correcting my own reply to brad@instep's query about use of the PC function keys in NeXTSTEP applications. I forgot that there is a gotcha with working with these. The hex codes generated by the keys overlap with those from the "main" keyboard. For example, F12 and the '+' key both generate 0x2b. The way to tell the difference is to check whether the character set of the key down event. "Regular" keys have a set of zero, while "extra" keys like F1 through F12 have a non-zero character set. Here is a macro which tests this: // The macro below evaluates to TRUE if the given event pointer E // represents a key down event generated by a key outside the main // QWERTY keyboard (ie. F1 through F12 or the cursor movement keys). #define IS_FUNCTION_KEY_EVENT(E) ((E)->data.key.charSet NEQ 0) Unfortunately this is hardly something you want to do on a regular basis. In my application, I actually subclassed the Application class so that the sendEvent: method could detect such special key events and send them to a special dispatch routine rather than calling super's sendEvent: (as it did with all other types of events). This solution fit my needs very well, but may not be the best approach for every one. Remember that you can't trap the F1 key as it never gets to your application as it is mapped to the Help key. -- 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 * "C++ is to C as Lung Cancer is to Lung."
From: Holger.Hoffstaette@schwaben.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: Porting 'sp': C++, templates & POSIX lossage (solved) Date: 28 Jul 1996 09:23:31 GMT Organization: home Message-ID: <4tfbij$rki@gate.seicom.net> References: <4tdb86$7p7@gate.seicom.net> I wrote: > > [probs with C++, templates, POSIX & struct stat] > The last problem is solved: header dependencies were mixed up, and struct stat was missing some integral typedefs (u_long etc.). After adding these and some more missing #defines, the rest of sp compiled without further hiccups, and amazingly enough didn't just explode at runtime, but even completed its test suite. I'll submit the necessary diffs to the archives & the maintainer. SGML friends rejoice! Holger -- Holger Hoffstätte // Holger.Hoffstaette@schwaben.de (MIME/NeXT/Sun/PGP)
Newsgroups: comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.nsc.32k,comp.sys.oric,comp.sys.palmtops,comp.sys.pen From: dsr@lns598.lns.cornell.edu Sender: logon@my.bbs (CyberTrader) Date: 28 Jul 1996 12:16:23 EDT Control: cancel <4tdsd7$o58@ns1.autonet.net> Subject: cmsg cancel <4tdsd7$o58@ns1.autonet.net> Message-ID: <cancel.4tdsd7$o58@ns1.autonet.net> Spam/MMF cancelled by dsr@lns598.lns.cornell.edu original subject was Earn $10,000+\month Working At Home!!
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: Strange exception with NSImage Date: 28 Jul 1996 19:12:06 GMT Organization: University of Illinois at Urbana Message-ID: <4tge26$7v@vixen.cso.uiuc.edu> I've hit a bump in my program that's really annoying me. After a day and a half I'm still not sure what's wrong with this. I feel kinda silly dumping all this gdb stuff on usenet but I'm really stuck. I'd appreciate any help you can give me. I've generated an NSImage and then would like to place the image on the pasteboard. So I go to get the TIFF from the NSImage using: NSData *tiffData; tiffData = [theImage TIFFRepresentation]; However, TIFFRepresentation raises an exception on some Dictionary being used that I have no idea what it's trying to do with. What is it trying to do with on frame #4? Here's the stack: #0 0x18053ad3 in -[NSException raise] () #1 0x1805379a in +[NSException raise:format:arguments:] () #2 0x18053715 in +[NSException raise:format:] () #3 0x1802a7cd in _NSDictionaryRaiseInsertNilValueException () #4 0x1800c096 in -[NSConcreteMutableDictionary setObject:forKey:] () #5 0x60acb1c in DPSSetContext () #6 0x60cf305 in +[NSDPSContext setCurrentContext:] () #7 0x6128d4c in -[NSImage TIFFRepresentationUsingCompression:factor:] () #8 0x6129263 in -[NSImage TIFFRepresentation] () My image should be contructed fine. (only weird thing is that the print * shows the size to be 0x0 and the po shows it to 467x81). Here's the image: (gdb) print image $1 = (class NSImage *) 0x40004 (gdb) print * image $2 = { isa = 0x6268f28, _name = 0x9c120, _size = { width = 0, height = 0 }, _flags = { scalable = 0, dataRetained = 0, uniqueWindow = 0, sizeWasExplicitlySet = 0, builtIn = 0, needsToExpand = 0, useEPSOnResolutionMismatch = 0, colorMatchPreferred = 1, multipleResolutionMatching = 1, subImage = 0, archiveByName = 0, unboundedCacheDepth = 0, flipped = 0, reserved1 = 0 }, _reps = 0x221fc, _color = 0x40094 } (gdb) po image NSImage 0x40004 Name=returnImage Size={width = 467; height = 81} Reps=( NSBitmapImageRep 0x403a4 Size={width = 467; height = 81} ColorSpace=NSCalibratedRGBColorSpace BPS=8 Pixels=467x81 Alpha=YES ) -- Ian P. Cardenas (icardena@uiuc.edu) CCSO Sites Technical Support "Are they as successful as who,Microsoft? Only drug lords from South America are as successful as Microsoft." -Tim Byars, on the success of NeXT Computer, Inc.
From: levenez@worldnet.fr Newsgroups: comp.sys.next.programmer Subject: Converting mu-law sound file to a-low ? Date: 28 Jul 1996 19:57:14 GMT Organization: SCT / Worldnet - Internet Provider & Information Exchange - Paris, France Message-ID: <4tggmq$feh@news.sct.fr> Hi, I want to convert a mu-law sound file to a-low. Mu-law is used in phone ISDN in USA, and a-law is used in Europe. Giso 3.3 does not know this conversion. Can anyone help me ? -- -------------------------------------------------------------------- Eric LEVENEZ : levenez@worldnet.fr The best way to accelerate Windows is at -9.81 m/s2 --------------------------------------------------------------------
From: Kurt Deding <deding@u.washington.edu> Newsgroups: comp.sys.next.programmer Subject: PB complains about missing app icon when recompile with ... Date: Sun, 28 Jul 1996 17:32:23 -0700 Organization: University of Washington Message-ID: <Pine.A32.3.92a.960728165309.55170A-100000@homer17.u.washington.edu> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII NNTP-Posting-User: deding I had a program working fine on a NeXTstation (the slab) under 3.3 until I decided to recompile it in order to run on both NeXT and Intel machines. The PB complains that it can't find the __ICON: and __header:. The specific message reads like this cc: __ICON: No such file or directory cc: __header: No such file or directory It appears that PB fails to recognize my customized app icon which sits at the same place where it was before (when I was successful with NeXT only option). I think I am probably responsible for the still unresolved mystery. I had inadvertently dropped the app icon through the info panel first (rather than first dropped it in the app icon well in the PB, then let it automatically show up in the info panel.) It was apparently ok with NeXT only. But a nono when I try to include Intel capability. I am wondering if anybody has had such an experience or can give me some advice on how to get the thing compiled again without having to start from as a new project and rebuild the nibs that require more mouse clickings than I want to. The thing still does not compile when I disable the Intel option to get back to the original state. Kurt
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: worryAboutPortInvalidation terrible UPS utilization Date: 29 Jul 1996 14:32:16 GMT Organization: E.S. Systems cc (Financial Systems Development) Message-ID: <4tii1g$3io@hermes.is.co.za> Hi I am writing a DO that needs to run each client in a different thread. When I use [NXPort worryAboutPortInvalidation] to monitor client deaths the UPS% of ps for this task goes up to about 95. If I take it out then it is about 0.0. The problem is that I need to know about client deaths. Is there an easy answer for this one? Has anybody else come across this? -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: cnyap@dcs.shef.ac.uk (Chih Nam Yap) Newsgroups: comp.sys.next.programmer Subject: Conceptual question ! Date: 29 Jul 1996 18:49:44 GMT Organization: Department of Computer Science, University of Sheffield Message-ID: <4tj148$bvc@bignews.shef.ac.uk> Hi there, I would like to ask some conceptual questions about "instantiate". 1. Can I have more than two "ball" objects within a single .nib file Eg, I have two classes "Controller" and "List" which are both subclass of the "object" class, can I instantiate both of them within the same .nib file, Is this conceptually correct ? (I did not use another .nib file because both objects have to be loaded when the application start) 2. I did the above, and have made a connection from the controller object to the list object via an id called "root" But after I built the application and then went back to the interface builder again, strange things have happened, the connection from the controller object to the list object has changed. The controller object has now pointed to one of my panel object instead othe the list object ? (via the id "root"). Does anyone know why this can be happened ? Thank you very much. C.Yap
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: Re: Converting mu-law sound file to a-low ? Date: 29 Jul 1996 19:14:42 GMT Organization: University of Illinois at Urbana Message-ID: <4tj2j2$brq@vixen.cso.uiuc.edu> References: <4tggmq$feh@news.sct.fr> levenez@worldnet.fr writes: >I want to convert a mu-law sound file to a-low. >Mu-law is used in phone ISDN in USA, and a-law >is used in Europe. There is source code for this conversion distributed with the sox source distribution available from http://www.spies.com/Sox/ Specifically look at the file g711.c -- Ian P. Cardenas (icardena@uiuc.edu) 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: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: Re: Oracle NT->EOF Date: 29 Jul 1996 19:18:57 GMT Organization: University of Illinois at Urbana Message-ID: <4tj2r1$cd1@vixen.cso.uiuc.edu> References: <4thjpq$3t1@crl.crl.com> mcgredo@crl.com (Donald R. McGregor) writes: >Any particular trick to getting EOF/Mach talking to Oracle running on >an NT box? > It sounds like you're doing everything right...what version of SQLNet are you using? NeXTanswers specifies different procedures for connecting to v1 and v2.x. If you haven't already take a look at http://www.next.com/NeXTanswers/HTMLFiles/1896.htmld/1896.html http://www.next.com/NeXTanswers/HTMLFiles/1897.htmld/1897.html I have very limited experience with oracle and EOF but if I rember right all I did was follow the NeXTanswers. -- Ian P. Cardenas (icardena@uiuc.edu) CCSO Sites Technical Support "Are they as successful as who,Microsoft? Only drug lords from South America are as successful as Microsoft." -Tim Byars, on the success of NeXT Computer, Inc.
From: Chuck_Esterbrook@orcacomputer.com Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Pressing CDs for NEXTSTEP files Date: 29 Jul 1996 22:24:32 GMT Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <4tjdn0$nqj@solaris.cc.vt.edu> So let's say I have a product ready to ship and I've collected all my files in a subdirectory on my NEXTSTEP/Intel machine. Can anyone tell me, (or point me to a resource that can), the way to burning a CD-ROM. I have Win95 on the other partition if that helps. There is a NEXTSTEP/WinNT machine in the room on the network. Is there a way to get my NEXTSTEP files onto WinNT's file system and then press the CD from there? Thanks for any pointers in advance! -------------------------------------------------------- Chuck_Esterbrook@OrcaComputer.com NeXTmail welcome Software Engineer pho. (540) 231-3475 Orca Computer, Inc. fax. (540) 231-3480 Blacksburg, VA 24060 http://www.orcacomputer.com/~chuck
From: icardena@sumter.cso.uiuc.edu (Ian Patrick Cardenas) Newsgroups: comp.sys.next.programmer Subject: Re: Strange exception with NSImage Date: 29 Jul 1996 23:45:41 GMT Organization: University of Illinois at Urbana Message-ID: <4tjif5$snh@vixen.cso.uiuc.edu> References: <4tge26$7v@vixen.cso.uiuc.edu> Well, I still don't know why NSImage is dying when I ask for a TIFFRepresentation. But I've done some more playing around and have a theory. In order to pick which representation is displayed it trys to do some DPS stuff to pick the best one. Well, as a command-line program with no NSApplication there is no DPSContext and that's why it's dying. So. What to do? I solved (hacked) the problem so that you can at least get _a_ representation if possible by adding a method to NSImage. I'd still like to know why my original program didn't work though, if anyone has any insight. //The following goes through the representations to see if any will cooperate //and give us a TIFFRepresentation. We return the data from the first to //respond @interface NSImage (TIFFRepresentationHack) -(NSData *) firstTIFFRepresentation; @end @implementation NSImage (TIFFRepresentationHack) -(NSData *) firstTIFFRepresentation { NSArray *allReps = [self representations]; NSEnumerator *repIterator = [allReps objectEnumerator]; NSImageRep *aRepresentation; while(aRepresentation = [repIterator nextObject]) { if ([aRepresentation respondsToSelector:@selector(TIFFRepresentation)]){ return [aRepresentation performSelector:@selector(TIFFRepresentation)]; } } return NULL; } @end In case you missed my first post I came up with this because the following didn't work: #import <AppKit/AppKit.h> #import <Foundation/Foundation.h> main () { NSAutoreleasePool *thePool = [[NSAutoreleasePool alloc] init]; NSImage *theImage; NSData *tiffData; theImage = [[NSImage alloc] initWithContentsOfFile:@"input.tiff"]; //The next line dies somewhere in TIFFRepresentation tiffData = [theImage TIFFRepresentation]; [tiffData writeToFile:@"output.tiff" atomically:NO]; [theImage release]; [thePool release]; } -- Ian P. Cardenas (icardena@uiuc.edu) 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.
Newsgroups: comp.sys.next.programmer From: tom@hukatronic.cz (Tomas Hurka) Subject: Re: Driver programming?? Message-ID: <DvC07v.3KI@hurka.UUCP> Sender: tom@hurka.UUCP (Tomas Hurka) Organization: Hukatronic (H.C.C.) References: <AlyCvDq00YUq0BmVs0@andrew.cmu.edu> Date: Tue, 30 Jul 1996 01:02:19 GMT In article <AlyCvDq00YUq0BmVs0@andrew.cmu.edu> Mahesh Saptharishi <mahesh+@andrew.cmu.edu> writes: > Are there any simple libraries/functions/headers available out there > that will help access the status, control and data registers of a > parallel port on > 1) a sparc 20 > 2) a PC Check out the source code for the Parallel driver for NSFIP from Christian Starkjohann <cs@hal.kph.tuwien.ac.at>. The source is available from next-ftp.peak.org: ftp://next-ftp.peak.org/pub/next/sources/drivers/ParallelPortDriver1.5.REA DME ftp://next-ftp.peak.org/pub/next/sources/drivers/ParallelPortDriver1.5.tar .gz Best regards, -- Tomas Hurka tom@hukatronic.cz NeXTMAIL and MIME OK (international mail <50 KB accepted)
From: gvandyk@icon.co.za Newsgroups: comp.sys.next.programmer Subject: Re: worryAboutPortInvalidation terrible UPS utilization Date: 30 Jul 1996 04:21:46 GMT Organization: E.S. Systems cc (Financial Systems Development) Message-ID: <4tk2kq$mfi@hermes.is.co.za> References: <4tii1g$3io@hermes.is.co.za> In-Reply-To: <4tii1g$3io@hermes.is.co.za> Sorry for the terrible "spelling" mistake, I don't know what I was thinking of, UPS should be CPU. (dumb,dumb,dumb) -- Regards, Gerrit van Dyk email: gvandyk@icon.co.za (NeXTMail welcome) E.S. Systems cc The OBJECT is the ADVANTAGE
From: kwong@morgan.ucs.mun.ca (Kai S. Wong) Newsgroups: comp.sys.next.programmer,comp.sys.next.marketplace Subject: WTB: Purple Book Date: 30 Jul 1996 05:00:59 GMT Organization: Memorial University of Newfoundland Message-ID: <4tk4ub$oec@coranto.ucs.mun.ca> I a looking for a "Programming The Display Postscript System with NeXTSTEP" (Purple book) written by Adobe and published by Addison-Wesley 1992. --kai-- -- email: kwong@morgan.ucs.mun.ca url: http://web.cs.mun.ca/~kwong/
From: neuss@isa.informatik.th-darmstadt.de.NOSPAM (Christian Neuss) Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: Pressing CDs for NEXTSTEP files Followup-To: comp.sys.next.software,comp.sys.next.programmer Date: 30 Jul 1996 09:16:02 GMT Organization: Fachbereich Informatik, TH Darmstadt, Deutschland Message-ID: <4tkjsi$206e@rs18.hrz.th-darmstadt.de> References: <4tjdn0$nqj@solaris.cc.vt.edu> Chuck_Esterbrook@orcacomputer.com wrote: > So let's say I have a product ready to ship and I've collected all my files > in a subdirectory on my NEXTSTEP/Intel machine. Can anyone tell me, (or point > me to a resource that can), the way to burning a CD-ROM. > Is there a way to get my NEXTSTEP files onto WinNT's file system and then > press the CD from there? Yes. On the NeXT machine, use mkisofs to create an ISO image for the new CDROM. On the W95 partition, run a software that burns the ISO image to CD (can't be done safely under UNIX becausse of real time requirements, bla, bla). PD software and info can be obtained from http://www.cd-info.com/CDIC/Technology/CD-R/FAQ.html. Find mkisofs with Archie or by checking Linux archives. Compiles ok on the NeXT. Please report on your experiences here, as I have not yet completed step two (I used Astarte CD Toast on MacOS 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
From: woo@woonext.cmo.ornl.gov (John W. Wooten) Newsgroups: comp.sys.next.programmer Subject: Can services be used from inside a web app? Date: 30 Jul 1996 13:40:07 GMT Organization: Oak Ridge National Lab, Oak Ridge, TN Distribution: world Message-ID: <4tl3bn$hsn@stc06.ctd.ornl.gov> Keywords: services web speaker listener pasteboard stereoview postscript 3D I've written an app, StereoView that takes data describing a three dimensional object and displays it in a view window as a stereo-pair that can be "steered" with a joy-stick. It is available via the Services menu if you highlight data describing the 3D object. I'd like to provide a web page that allows 3D data to be entered in a textarea and upon receiving this at my server, use the services feature to have the StereoView.app process the data. How do I get the information from the web app to the StereoView.app? AND how do I get the postscript that would have gone into the view (or did) back to the web app so I can return it to the client? I'm using 3.3 (and have EOF 1.1) if that makes a difference. -- - - - - - - - - - J. W. Wooten
From: khader@katie.vnet.net (R. D. Khader) Newsgroups: comp.sys.next.programmer Subject: Q: DBTableView in c++ fails? Date: 30 Jul 1996 16:01:19 GMT Organization: Vnet Internet Access, Inc. - Charlotte, NC. (704) 374-0779 Message-ID: <4tlbkf$lib@ralph.vnet.net> I am trying to compile a dbtableview interface object in c++; As a matter of fact all the app is compiled with the C++ directive. I am getting the following error: cc -ObjC++ -g -O -Wall -I./sym -DNEXT -D__cplusplus -arch m68k -I. -c AlphaSearchController.m -o ./m68k_obj/AlphaSearchController.o In file included from /NextDeveloper/Headers/dbkit/dbkit.h:43, from AlphaSearchController.h:5, from AlphaSearchController.m:3: /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:25: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:26: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:27: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:28: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:29: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:30: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:31: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:32: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:33: invalid raises declaration /NextDeveloper/Headers/dbkit/DBEditableFormatter.h:34: invalid raises declaration *** Exit 1 Stop. *** Exit 1 Stop. I am using the extern "Objective-C" around the objc kit files. Also appkit files are comiling fine except for when I include the dbtableview, i.e., dbkit.h By the way I am using 3.1 on Black. any comments is highly appreciated, thanks, rdkhader,
From: mahesh@gussolomon.cs.cmu.edu (Mahesh Saptharishi) Newsgroups: comp.sys.next.programmer Subject: Source for NS parallel port driver for a SPARC. Date: 30 Jul 1996 16:47:49 GMT Organization: Carnegie-Mellon University, School of Computer Science Message-ID: <4tlebl$s9q@cantaloupe.srv.cs.cmu.edu> I am looking for a NS parallel port driver for a SPARC 20 (It should be the same for the 4,5 &10) with source code. I already found one for the PC, but since the parallel port architecture is completely different on a SPARC the PC version is of no help. Ideas, suggestions, solutions, comments all highly appreciated. Thanks in advance, -Mahesh
From: Holger Hoffstaette <hhoff@modern-video.de> Newsgroups: comp.sys.next.software,comp.sys.next.programmer Subject: Re: Pressing CDs for NEXTSTEP files Date: Tue, 30 Jul 1996 17:22:42 +0200 Organization: Modern Video Media Group Message-ID: <31FE28C2.1D8D@modern-video.de> References: <4tjdn0$nqj@solaris.cc.vt.edu> <4tkjsi$206e@rs18.hrz.th-darmstadt.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Christian Neuss wrote: > Chuck_Esterbrook@orcacomputer.com wrote: > > > So let's say I have a product ready to ship and I've collected all my files > > in a subdirectory on my NEXTSTEP/Intel machine. Can anyone tell me, (or point > > me to a resource that can), the way to burning a CD-ROM. > (..) > > Yes. On the NeXT machine, use mkisofs to create an ISO image for the > new CDROM. On the W95 partition, run a software that burns the ISO > image to CD (can't be done safely under UNIX becausse of real time > requirements, bla, bla). PD software and info can be obtained from Not true: Linux gets the job done just fine. Been there, seen some- one do that. I agree that the mkisofs route is a good one to take, though you really waste a lot of disk space in the process. Holger -- Holger Hoffstätte (hhoff@modern-video.de) Modern Video Media Group Stuttgart, Germany
From: kwong@morgan.ucs.mun.ca (Kai S. Wong) Newsgroups: comp.sys.next.programmer Subject: Best tool for making Icons? Date: 30 Jul 1996 19:12:13 GMT Organization: Memorial University of Newfoundland Message-ID: <4tlmqd$tan@coranto.ucs.mun.ca> Hello, Anyone know what is the best tools for making icons? (something other than Icon Builder). --kai-- -- email: kwong@morgan.ucs.mun.ca url: http://web.cs.mun.ca/~kwong/
From: ehutch@hypnos.norden1.com (E. Hutchinson) Newsgroups: comp.sys.next.programmer,misc.jobs.contract,us.jobs.contract Subject: NEXTSTEP--Contract--Long term--Greater DC Area Date: 30 Jul 1996 19:26:13 GMT Organization: Norden 1 Communications Message-ID: <4tlnkl$lfk@tofu.alt.net> Programmer/analyst/developer NEXTSTEP---------------------Commercial experience Objective C------------------Commercial experience EOF--------------------------A Plus Sybase or Oracle-------------A Plus Area-------------------------Greater DC Area Contract---------------------Long term 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
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.