This is Prog94-II.gz in view mode; [Up]
From: wwright@shell.portal.com (Bradly William Wright) Newsgroups: comp.sys.next.programmer Subject: IODirectDevice Memory Mapping Bug? Date: 23 Jun 1994 16:42:08 GMT Organization: Portal Communications Company Message-ID: <2uce10$fq6@news1.svc.portal.com> Summary: IODirectDevice seems unable to do memory mapping. Keywords: IODirectDevice, memory mapping I'm trying to get an IODirectDevice (DriverKit) driver to use memory mapped IO. I believe that I set the Default.table up correctly. I created a line that specifies the valid memory maps for the hardware. The problem is that when I try to call the function to map memory it always returns an invalid arguement error message. If I assume that the initFromDeviceDescription routine has already mapped the memory, I get a kernel panic when trying to write to the address of the shared memory. Has anyone implemented an IODirectDevice driver with memory mapping before? I need to know the correct way of setting up memory mapped IO. Brad Wright WrightSystems wwright@shell.portal.com
From: chris@iastate.edu (Chris Wong) Newsgroups: comp.sys.next.programmer Subject: Can't change button's icon image. Weird behavior. Date: 23 Jun 1994 17:32:28 GMT Organization: Iowa State University, Ames, Iowa (USA) Message-ID: <2ucgvc$ir7@news.iastate.edu> Very weird. I use [[myMatrix selectedCell] setImage:someImage] to set the icon image of one of the button at runtime. However, the image doesn't appear when I launch the app. The weirdest thing is the image appear if I launch the app in gdb, i.e., in debug mode. Here is the method that get executed when one of the button in the Matrix is click. - mark:sender { if (someCondition) { [[mineField selectedCell] setImage:flag]; } return self; } Please share some insight here. I am totally clueless. Thank you very very very much for any kind of information... Chris -- desperate -- NeXTMail super welcomed!!| Chris Wong | "Hardware is supposed to serve Software." chris@iastate.edu | Computer Engineering & Computer Science <:)>:)<:)>:)<:)>:)<:) | Iowa State University of Science and Technology
From: mow@marsu.tynet.sub.org (Markus Wenzel) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP 3.2 implementation of POSIX Date: 23 Jun 1994 09:59:11 +0100 Organization: Palumbian Research Labs Message-ID: <2ubisv$138@marsu.tynet.sub.org> References: <CrrGFK.n2J@fwrdc.rtsg.mot.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Keywords: POSIX signals sysconf gmarler@sunmail1.fwrdc.rtsg.mot.com (Gordon Marler x6624) writes: >To what extent does NeXTSTEP 3.2 support the POSIX C interface >and system calls? Does it have the sysconf() call, and POSIX >signals? I'm using 3.0 on black hardware and just want to know >if it 3.2 has enough of these POSIX features to make it worth my >while to upgrade. If someone who has 3.2 could look for the >constants _POSIX_VERSION and _XOPEN_VERSION (if they exist) and >tell me what they are, I'd really appreciate it. The POSIX compatibility is not very convincing. Try to compile and link the following short C prog. It compiles fine, but it fails linking - and I just don't know why: ---------------------------------------------------- #include <stdio.h> #include <time.h> int main(int argc, char **argv) { time_t clock = time((time_t *) NULL); static char date[9]; (void) strftime(date, (size_t)sizeof(date), "%y/%m/%d", localtime(&clock)); printf("Today is : %s\n",date); } ---------------------------------------------------- The result is: marsu [Misc] % cc -posix posix_test1.c ld: Undefined symbols: _objc_msgSend _NXGetDefaultValue _NXRegisterDefaults .objc_class_name_NXStringTable .objc_class_name_NXBundle Any suggestions? Markus. -- /dev/ Markus Wenzel /usr/spool/mail/ mow@marsu.tynet.sub.org /etc/zoneinfo/ University of Stuttgart /bin/ps System administration, Consulting, Networking
From: michaelr@pengyou.Stanford.EDU (J. Michael Richards) Newsgroups: comp.sys.next.programmer Subject: using custom text field with DBKit Date: 24 Jun 1994 01:13:18 GMT Organization: Stanford University Message-ID: <2udbve$ljb@nntp2.Stanford.EDU> Here's my problem: I want to have a subclass of TextField parses the value provided to it by a DBModule via the method association:setValue: I created such a subclass, providing an association:setValue:, but it seems that this method never gets called because DBKit already defines a setValue: for Control objects which gets called instead. The only way I could find to get my code called was to make a subclass of Box, which has no border or title and serves as a wrapper for a TextField. Since Box is not a Control subclass, association:setValue: gets called. (This is how I understand it anyway.) Now clearly this is a kludge, since what I really want is an object which inherits all the functionality of a TextField and just does things a bit differently when interfacing with a DBModule. Can anyone tell me what I'm doing wrong? I really believe this should be possible, and maybe I'm misunderstanding something about the kit or about objective-C. thanks, mike
From: he@engin.swarthmore.edu (Ye He) Newsgroups: comp.sys.next.programmer Subject: help: writing a DirectDriver Date: 24 Jun 1994 03:15:58 GMT Organization: Swarthmore College Engineering, Swarthmore PA Message-ID: <2udj5e$1g5@larch.cc.swarthmore.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Hello Folks, I need to write a driver for a timer board for one of our NS 486's. I am currently reading the DriverKit docs. Does anyone out there have the experience of writing an NS driver and if so could you give me some advice? It would be even better if I can have some driver templates to use. Please reply to my email address. Thanks. -- Ye he@engin.swarthmore.edu
From: jmack@skye.phys.ualberta.ca Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP 3.2 implementation of POSIX Date: 24 Jun 1994 07:02:51 GMT Organization: Computer and Network Services, U of Alberta, Edmonton, Canada Message-ID: <2ue0er$l3k@quartz.ucs.ualberta.ca> References: <2ubisv$138@marsu.tynet.sub.org> In article <2ubisv$138@marsu.tynet.sub.org> mow@marsu.tynet.sub.org (Markus Wenzel) writes: > gmarler@sunmail1.fwrdc.rtsg.mot.com (Gordon Marler x6624) writes: > The POSIX compatibility is not very convincing. Try to compile > and link the following short C prog. It compiles fine, but it fails > linking - and I just don't know why: .. > The result is: > > marsu [Misc] % cc -posix posix_test1.c > ld: Undefined symbols: > _objc_msgSend > _NXGetDefaultValue > _NXRegisterDefaults > ..objc_class_name_NXStringTable > ..objc_class_name_NXBundle > > Any suggestions? > Try: cc -posix -c posix_test1.c cc -o posix_test1 posix_test1.o /tmp> posix_test1 Today is : 94/06/24 -- James S. MacKinnon Office: P-139 Avahd-Bhatia Physics Lab Computing/Networking Phone : (403) 492-8226 Department of Physics email : jmack@phys.ualberta.ca University of Alberta uucp : uofaphys!jmack iskye!jmack Edmonton, Canada T6G 2N5 bitnet: jmack@triumfcl jsm1@ualtamts
From: jmack@skye.phys.ualberta.ca Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP 3.2 implementation of POSIX Date: 24 Jun 1994 07:03:53 GMT Organization: Computer and Network Services, U of Alberta, Edmonton, Canada Message-ID: <2ue0gp$pfm@quartz.ucs.ualberta.ca> References: <2ubisv$138@marsu.tynet.sub.org> In article <2ubisv$138@marsu.tynet.sub.org> mow@marsu.tynet.sub.org (Markus Wenzel) writes: > Try to compile > and link the following short C prog. It compiles fine, but it fails > linking - and I just don't know why: .. > The result is: > > marsu [Misc] % cc -posix posix_test1.c > ld: Undefined symbols: > _objc_msgSend > _NXGetDefaultValue > _NXRegisterDefaults > ..objc_class_name_NXStringTable > ..objc_class_name_NXBundle > > Any suggestions? > Try: cc -posix -c posix_test1.c cc -o posix_test1 posix_test1.o /tmp> posix_test1 Today is : 94/06/24 -- James S. MacKinnon Office: P-139 Avahd-Bhatia Physics Lab Computing/Networking Phone : (403) 492-8226 Department of Physics email : jmack@phys.ualberta.ca University of Alberta uucp : uofaphys!jmack iskye!jmack Edmonton, Canada T6G 2N5 bitnet: jmack@triumfcl jsm1@ualtamts
Newsgroups: comp.sys.next.programmer From: esprit@netcom.com (Alan F. Perry) Subject: Is there a SCSIKit or anything similar in NeXTSTEP? Message-ID: <espritCrw2v9.4xF@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Fri, 24 Jun 1994 06:47:33 GMT Are there any objects in NeXTSTEP that allow a program to deal with SCSI devices in a uniform way, that is, the program can just send SCSI commands to the object and the object will deal with figuring out how and where to send it? -- ----------------------------+----------------------------------------------- Alan F. Perry | Life is short, but by achieving greater speeds alanp@eng.sun.com (work) | a man can make his life a little longer and esprit@netcom.com (home) | more affluent - Soichiro Honda
Newsgroups: comp.sys.next.programmer From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Subject: Magic links in Edit Message-ID: <Crw7vz.Ho8@news.informatik.uni-stuttgart.de> Summary: Edit automtically updates pictures in RTFD documents Sender: news@informatik.uni-stuttgart.de Organization: Informatik, Uni Stuttgart, Germany Date: Fri, 24 Jun 1994 08:35:58 GMT Hello all! If you drag&drop a picture into a RTFD document you can create a copy by holding down the <Alternate> key (or doing nothing) or create a link using the <Control> key. Unfortunately the copy doesn't seem to be a real copy: - drag&drop an image into an RTFD document - close the document - change the image file on disk using Icon (or something like that) - open the RTFD document again => the changed image shows up Is this a kind of "magic link" (found no docu)? What is the underlying mechanism? Is it a feature or a bug? What if I want to have a real copy? Many questions, Klaus...
Newsgroups: comp.sys.next.programmer From: Atze (Alexander Spohr) Subject: Re: using custom text field with DBKit Message-ID: <1994Jun24.090358.1747@dart.de> Sender: aspohr@dart.de Organization: d'ART Software GmbH References: <2udbve$ljb@nntp2.Stanford.EDU> Date: Fri, 24 Jun 94 09:03:58 GMT In article <2udbve$ljb@nntp2.Stanford.EDU> michaelr@pengyou.Stanford.EDU (J. Michael Richards) writes: > Here's my problem: > > I want to have a subclass of TextField parses the value provided to it by > a DBModule via the method association:setValue: > > I created such a subclass, providing an association:setValue:, but it > seems that this method never gets called because DBKit already defines a > setValue: for Control objects which gets called instead. The only way I > could find to get my code called was to make a subclass of Box, which has > no border or title and serves as a wrapper for a TextField. Since Box is > not a Control subclass, association:setValue: gets called. (This is how I > understand it anyway.) > > Now clearly this is a kludge, since what I really want is an object which > inherits all the functionality of a TextField and just does things a bit > differently when interfacing with a DBModule. Can anyone tell me what I'm > doing wrong? I really believe this should be possible, and maybe I'm > misunderstanding something about the kit or about objective-C. > Mike, this is wrong. We have lots of Control subclasses (eg. Button) and they all work with - association:association setValue:(DBValue *)value without problems. Maybe you should recheck you code :-) Atze -- Alexander Spohr, Hamburg, Germany Alexander_Spohr@DART.de (Mail ) Faces and faces (+49) 40 / 380 23-0 (Sound) See them and complain not (+49) 40 / 380 23-290 (Fax ) And am content with all
Newsgroups: comp.sys.next.programmer From: bbernhar@silver.ucs.indiana.edu (benjamin bernhard) Subject: DBAssociation (subclass) for NXBrowser? Message-ID: <CrwpuA.F3u@usenet.ucs.indiana.edu> Sender: news@usenet.ucs.indiana.edu (USENET News System) Organization: Indiana University Date: Fri, 24 Jun 1994 15:03:46 GMT Hello all, I'm currently moving my 0.8 product from QuickBase's Obj-c interface to DBKit compliance. It looks like I'm going to be writing a DBAssociation to link by NXBrowser to the appropriate DBFetchgroups. My questions: Has anybody done this? Are there any solid protocols I should follow? This is what I have so far: /* from the DBFetchgroup */ - contentsDidChange; - currentRecordDidDelete; - endEditing;/* no editing: this won't do much*/ - (unsigned int)selectedRowAfter:(unsigned int)previousRow; - validateEditing; /* no editing: this won't do much*/ /* from the browser*/ - rowCount; - getValueFor:rowIdentifier into:aValue; - getValueFor:identifier at:(unsigned int)aPosition into:aValue; I'm planning to have one DBAssociation and one DBFetchgroup per browser column. Any suggestions, pointers much appreciated! thanks, ben ben@spot.bloomington.in.us -- __________________________________________________________________________ Ben Bernhard "Nothing that results from human progress 812/339-5304 (fax) is achieved with unanimous consent." bbernhar@cs.indiana.edu ---Christopher Columbus
From: elitman@proxima.com (Eric A. Litman) Newsgroups: comp.sys.next.programmer Subject: doing the right thing with AppleSingle'd files Date: 24 Jun 1994 10:48:29 -0500 Organization: Proxima, Inc. Sender: nobody@cs.utexas.edu Message-ID: <9406241520.AA10079@proxima.com> I've come across some code written by StarNine to do AppleSingle encoding and decoding, but it's really weak. If you have some first-hand experience dealing with this format, I'd be grateful (read: will pay) to hear about your experiences. -- Eric Litman Proxima, Inc. vox: (703) 506.1661 Director, Network Services McLean, VA elitman+@proxima.com
From: pcu@umich.edu Newsgroups: comp.sys.next.programmer Subject: problem with qsort'ing Date: 24 Jun 1994 15:54:14 GMT Organization: University of Michigan Distribution: world Message-ID: <2uevj6$c44@lastactionhero.rs.itd.umich.edu> Well, this problem has me stumped. I am attempting to use qsort to sort an array of POINT, by looking at the first float variable of the two variable structure. Unfortunately, qsort not only sorts the array, but it modifies the contents of the array. What am I doing wrong? Any help would be gratefully accepted. I include the following in this message: 1. the POINT structure. 2. the comparison function 3. an edited version of the method that calls qsort 4. a gdb trace showing the values of the first point, the pointer point, and the value returned by qsort, before and after qsort is called 5. the man page for qsort. typedef struct { float energy, intensity; } POINT; int comparePoints(const POINT *a, const POINT *b); { /* kludgy but explicit function */ if (a->energy > b->energy) return 1; else if (a->energy == b->energy) return 0; else // (a->energy < b->energy) return -1; } - objectmethod:sender { POINT *points; void *returnVal; ..... number = {sum of integers}; /* number is around 2500 */ NX_MALLOC(points, POINT, number); /* number may change, points is NX_REALLOC'd if so */ ..... *points are calculated and assigned values ..... number = {sum of integers}; /* number is still somewhere around 2500 */ returnVal = qsort(points, number, sizeof(POINT), (int (*)(const void *, const void *))comparePoints); /* only difference between casting or not is compiler warning */ ..... return self; } (gdb) next /* just before calling qsort */ 5: *points = { energy = 0.099070914, intensity = 0.51011169 } 4: points = (POINT *) 0x120d7c 1: returnVal = (void *) 0xdf5c8 (gdb) next /* just after calling qsort */ 5: *points = { energy = -2.58569021e+38, intensity = -1.58836229e-19 } 4: points = (POINT *) 0x120d7c 1: returnVal = (void *) 0x0 man 3 qsort NAME qsort - quicker sort SYNOPSIS #include <stdlib.h> void *qsort(void *base, size_t nmemb, size_t size, int (*compar)(const void *, const void *)) DESCRIPTION Qsort is an implementation of the quicker-sort algorithm. The first argument is a pointer to the base of the data; the second is the number of elements; the third is the size of an element in bytes; the last is the name of the comparison routine to be called with two arguments which are pointers to the elements being compared. The routine must return an integer less than, equal to, or greater than 0 according as the first argument is to be considered less than, equal to, or greater than the second. SEE ALSO sort(1) -- Peter Urka <pcu@umich.edu> Dept. of Chemistry, Univ. of Michigan Anything to me is sweeter, Than to see Shock-headed Peter. - H. Hoffmann
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: DBAssociation (subclass) for NXBrowser? Date: 24 Jun 1994 17:25:15 GMT Organization: University of California, Santa Cruz Distribution: world Message-ID: <2uf4tr$2lo@darkstar.UCSC.EDU> References: <CrwpuA.F3u@usenet.ucs.indiana.edu> In article <CrwpuA.F3u@usenet.ucs.indiana.edu> bbernhar@silver.ucs.indiana.edu (benjamin bernhard) writes: > I'm currently moving my 0.8 product from QuickBase's Obj-c interface > to DBKit compliance. Just a general word of advice: Next is replacing DBKit with Enterprise Objects Framework (EOF), so if you can hold off until the October release or get a copy of the beta given to all who attended the Expo developer conference, do so! EOF seems to solve the architectural problems of DBKit so will be much nicer and more powerful. Those who have used the alpha version of EOF report that it's very stable and that the beta is in some respects more stable and less buggy than the current DBKit. EOF is also built upon the new Foundation classes upon which OpenStep will be based, so building OpenStep-compliant apps will mean not using DBKit (however, neither DBKit nor EOF are included in the initial OpenStep spec). There doesn't appear to be any clear path for the conversion from DBKit to EOF :-( So any DBKit implementation that you DON'T do today will be less conversion work that you'll have to do in the future. --- Art Isbell Cubic Solutions NeXT Registered Consultant NEXTSTEP software development and consulting NeXTmail: art@cubicsol.com Voice: +1 408 335 1154 USmail: 95018-9442 Fax: +1 408 335 2515
From: tyler@vsc.com (Tyler Gingrich) Newsgroups: comp.sys.next.programmer Subject: Re: using custom text field with DBKit Date: 24 Jun 1994 17:32:42 GMT Organization: OARnet Message-ID: <2uf5bq$5qd@ns.oar.net> References: <2udbve$ljb@nntp2.Stanford.EDU> In article <2udbve$ljb@nntp2.Stanford.EDU> michaelr@pengyou.Stanford.EDU (J. Michael Richards) writes: > Here's my problem: > > I want to have a subclass of TextField parses the value provided to it > by a DBModule via the method association:setValue: > > I created such a subclass, providing an association:setValue:, but it > seems that this method never gets called because DBKit already defines a > setValue: for Control objects which gets called instead. The only way I > could find to get my code called was to make a subclass of Box, which > has no border or title and serves as a wrapper for a TextField. Since > Box is not a Control subclass, association:setValue: gets called. (This > is how I understand it anyway.) > > Now clearly this is a kludge, since what I really want is an object > which inherits all the functionality of a TextField and just does things > a bit differently when interfacing with a DBModule. Can anyone tell me > what I'm doing wrong? I really believe this should be possible, and > maybe I'm misunderstanding something about the kit or about > objective-C. > You need to create a custom association for your class. The easiest way to do this is to subclass DBAssociation overriding setValue: and getValue: in your new class. You also need to add a category to your custom TextField so it can tell the kit which Class to use for it's association. Sort of like this... #import <appkit/appkit.h> #import <dbkit/dbkit.h> @interface MyTextField (DBAssociationStuff) - (const char *) _dbAssociationClass; @end @interface MyTextFieldAssociation:DBAssociation - setValue:value; - getValue:value; @end #import "MyTextFieldAssociation.h" @implemenation MyTextField (DBAssociationStuff) - (const char *) _dbAssociationClass { return "MyTextFieldAssociation"; } @end @implementation MyTextFieldAssociation - setValue:value { [[self destination] setStringValue:[value stringValue]]; return self; } - getValue:value { [value setStringValue:[[self destination] stringValue]; return self; } @end --- Tyler Gingrich tyler@vsc.com Vanguard Software, Corp. PS: There is an official/approved way to return the class name, but I never bother to use it, choosing instead to directly override _dbAssociationClass.
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and C++ Date: 24 Jun 1994 17:48:05 GMT Organization: University of California, Santa Cruz Distribution: world Message-ID: <2uf68l$2sl@darkstar.UCSC.EDU> References: <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> In article <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> pfkeb@kaon.SLAC.Stanford.EDU (Paul F. Kunz) writes: > What is really annoying is that NeXTSTEP for HP PA-RISC will come > out in July and is already in beta. Guess what? The compiler for > NeXTSTEP-For-HP is based on gcc 2.5.8! So if you have an HP version > of NeXTSTEP, you can have templates in your C++ code. You can even > produce Motorola and Intel code on the HP with a gcc 2.5.8 based > compiler. But if you only have Motorola or Intel machines, then you > have to wait another year before NeXT gives you a update to the C++ > compiler. > Not only this, but NeXT's developer tools team (the guys responsible for the compiler tools as well as the UI developer tools like IB, PB, HV, etc.) has done an incredible job on improving the development environment, some of which was demoed at the Expo and looks very stable. But this, too, will have to wait until NS 4.0, another year down the road. I think I sensed the frustration of the team leader about this situation, and he recommended talking to Rick_Jackson@next.com who apparently made the marketing decision to not ship a 3.3 developer version. I'd like to see all supported platforms have the same compiler and have those developer tools that are ready to ship by October, the 3.3 release date. So I encourage everyone who agrees to make Rick Jackson aware of your desires. Enterprise Objects Framework will be available in October as a separate product in both user and developer versions. So a developer release of the compiler, ProjectBuilder, etc. could be scheduled for the same release. The new IB is included with the EOF release, so maybe all the developer tools should be made part of the EOF developer release. EOF user could be a separate, less expensive product. --- Art Isbell Cubic Solutions NeXT Registered Consultant NEXTSTEP software development and consulting NeXTmail: art@cubicsol.com Voice: +1 408 335 1154 USmail: 95018-9442 Fax: +1 408 335 2515
From: art@cubicsol.com (Art Isbell) Newsgroups: comp.sys.next.programmer Subject: Re: What new features will be in 3.3? Date: 24 Jun 1994 18:17:20 GMT Organization: University of California, Santa Cruz Distribution: world Message-ID: <2uf7vg$38q@darkstar.UCSC.EDU> References: <1994Jun23.000853.7943@millennium.com> In article <1994Jun23.000853.7943@millennium.com> Jayson Adams writes: > In article <ROBERT.94Jun19074704@steffi.demon.co.uk>, Robert Nicholson writes: > > <rgc@cs.umd.edu> writes: > > > > >What new utilities will be in NS 3.3? Specifically, will 3.3 provide > > >version controll? An improved Edit (maybe will color syntax > > highlighting).... > > > > If you want _automatic_ color syntax highlighting then go no further > > than Emacs for NS.. with font-lock.e. or hilit19.e. > > That sounds great. How do I get it within Edit? Wait for NS 4.0, unfortunately, even though it appears that the new PB editor demoed at the Expo will support this feature (although it hadn't yet been implemented). Tell Rick Jackson at NeXT that you want the new developer tools including PB as part of a 3.3 developer release. --- Art Isbell Cubic Solutions NeXT Registered Consultant NEXTSTEP software development and consulting NeXTmail: art@cubicsol.com Voice: +1 408 335 1154 USmail: 95018-9442 Fax: +1 408 335 2515
From: dekelver@cowley.uwlax.edu (Jeff J. DeKelver S92) Newsgroups: comp.sys.next.programmer Subject: Manuals, Books, Information, etc... Date: 24 Jun 1994 18:12:22 GMT Organization: University of Wisconsin - La Crosse Message-ID: <2uf7m6$jsn@alfred.acs.uwlax.edu> Keywords: Books, Help, programming, objective C I was wondering what books, or manuals, etc. are out there for someone who is not all that experienced in programming in objective C, using IB on NeXTSTEP. Titles, and authors would be appreciated. Please E-Mail to dekelver@cslab.uwlax.edu Thank you in advance for any help.
From: hoff@weschnitz.darmstadt.gmd.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and C++ Date: 24 Jun 1994 18:31:42 GMT Organization: German National Research Centre for Information Technology Sender: hoff@weschnitz (Holger Hoffstaette) Distribution: world Message-ID: <2uf8qe$lnu@omega.gmd.de> References: <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> <2uf68l$2sl@darkstar.UCSC.EDU> In article <2uf68l$2sl@darkstar.UCSC.EDU>, art@cubicsol.com (Art Isbell) writes: >[Developer Tools not coming in 3.3] WHAT ON EARTH is the reason for this ? Does the person responsible for this decision even _remotely_ know what kind of damage he is doing ? It would be good if someone 'who knows' could shed a bit more light on this. The complete lack of a usable C++ compiler with NeXTSTEP is a _serious_ shortcoming, no matter what NeXT think about C++. Same goes for the many limitations in IB/PB/HV/whatnot. The Geckos could have a huge amount of PD or shareware apps in virtually no time, unlike any other UNIX port before. Or is it some kind of 'clever move' to make people switch to a Gecko (and 'encourage' everybody to junk their recently bought Intel system) ? I just don't get it. Angry, H.Hoffstaette -- Holger Hoffstaette, GMD-IPSI/PaVE (hoff@darmstadt.gmd.de) [no NeXTmail] Smalltalker, NeXTstepper, Amiga veteran, Net citizen. "I'm gonna get myself connected.." - Stereo MC's
Newsgroups: comp.sys.next.programmer From: eike@ilink.de (Eike Dierks) Subject: Re: Stopping mach-threads from switching Message-ID: <Crx3M0.C6B@ilink.de> Sender: usenet@ilink.de Organization: i.link Kommunikationssysteme GmbH, Berlin References: <2u736f$593@bilbo.suite.com> Date: Fri, 24 Jun 1994 20:01:11 GMT thread_suspend() & thread_resume() should help. You can access the mach thread for a cthread via cthread_t->real_thread. --Eike
From: buddha@samsara.circus.com (Adam Deishu Beeman) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and C++ Date: 24 Jun 1994 13:46:27 -0700 Organization: The Marshmallow Peanut Circus Message-ID: <2ufgn3$omr@samsara.circus.com> References: <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> <2uf68l$2sl@darkstar.UCSC.EDU> <2uf8qe$lnu@omega.gmd.de> In article <2uf8qe$lnu@omega.gmd.de>, Holger Hoffstaette <hoff@darmstadt.gmd.de> wrote: > > >WHAT ON EARTH is the reason for this ? Does the person responsible for this >decision even _remotely_ know what kind of damage he is doing ? >It would be good if someone 'who knows' could shed a bit more light on this. >The complete lack of a usable C++ compiler with NeXTSTEP is a _serious_ >shortcoming, no matter what NeXT think about C++. geez... ftp to prep.ai.mit.edu and get /pub/gnu/gcc-2.5.8.tar.gz, and build it yourself. Same goes for the latest libc/libg++, if you don't want to wait for NeXT to carefully inspect the GNU source, make their changes, test, and release the absolute latest compiler. Of course, this won't give you fat c++ binaries, and for some of us this will make a stock g++ unusable. Depends on your needs. IMHO the future lies in making compiler upgrades readily available for those who need them. >Same goes for the many >limitations in IB/PB/HV/whatnot. The Geckos could have a huge amount of PD >or shareware apps in virtually no time, unlike any other UNIX port before. >Or is it some kind of 'clever move' to make people switch to a Gecko (and >'encourage' everybody to junk their recently bought Intel system) ? > >I just don't get it. A shot in the dark, perhaps, but I would surmise that they're working on some substantial improvements to the developer environment which are not completely ready to be released. That would certainly explain things. [Adam soapboxMode: 1]; Be careful what you ask for. Do you want NeXT to fix everything absolutely and make it totally ideal, or release what they have now? Do you want them to make frequent releases and charge for every little upgrade, or infrequent landmark upgrades? Should they give their software away and go broke or target people with money? People seem to expect NeXT to do everything at once. It's not a huge company. They do seem to be working quite hard at improving their software. I for one think they're doing fairly well at juggling the tasks they have. [Adam soapboxMode: 0]; >Angry, > > H.Hoffstaette >-- >Holger Hoffstaette, GMD-IPSI/PaVE (hoff@darmstadt.gmd.de) [no NeXTmail] >Smalltalker, NeXTstepper, Amiga veteran, Net citizen. > "I'm gonna get myself connected.." - Stereo MC's I do sympathize, though, and hopefully NeXT will find the happy medium here. -Adam -- //#*=--=*#*=--=*#*=--=*#*=--=*#*=--=*#*=--=*#*=--=*#*=--=*#*=--=*#*=--=*#// // Adam Beeman \\ Standard Disclaimers Apply! // // Home = buddha@circus.com \\ work = 0; // // http://samsara.circus.com/~buddha/ \\ I don't speak for anyone //
Newsgroups: comp.sys.next.programmer From: rlabelle@lamar.colostate.edu Subject: Looking for an example of using rld Sender: news@yuma.ACNS.ColoState.EDU (News Account) Message-ID: <Crx7uG.210u@yuma.ACNS.ColoState.EDU> Date: Fri, 24 Jun 1994 21:32:40 GMT Organization: Colorado State University, Fort Collins, CO 80523 Keywords: rld Hello, I'm looking for an example of using the rld functions. I'd like to use dynamic loading in straight c-code. Thanks for any info..... Rob LaBelle rlabelle@lamar.colostate.edu
Newsgroups: comp.sys.next.programmer From: robert@ictv.com (Robert Patrick Thille) Subject: Printing out objects. Message-ID: <1994Jun24.224951.29318@ictv.com> Sender: usenet@ictv.com Organization: ICTV, Santa Clara, CA (408) 562-9200 Distribution: usa Date: Fri, 24 Jun 1994 22:49:51 GMT I'd like to print out some objects, like gdb does. I was hoping that printForDebugger: would do what I wanted, but it just prints the class and the address. Does anyone have any code which will query the run-time and print out an object like GDB does? Also, I remember the thread about ASCII archiving here awhile ago. Did anyone come up with a good solution? What I'm trying to do is use IB to generate some screen layouts, for use in another environment, so I want an ASCII representation of all the objects and their connections. I've got the code that talks with IB's active document, and gets all the objects, and I can get their connections. Now I just need a nice way to print them out. Would a category to Object work? Thanks, Robert -- Robert Thille NeXTMail robert@ictv.com OK 660 Bair Island Rd #40 Redwood City CA 94063 PGP Public Key via finger
From: hoff@darmstadt.gmd.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and C++ Message-ID: <8129@darmstadt.gmd.de> Date: 24 Jun 94 23:37:24 GMT References: <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> <2uf68l$2sl@darkstar.UCSC.EDU> <2uf8qe$lnu@omega.gmd.de> <2ufgn3$omr@samsara.circus.com> Sender: hoff@darmstadt.gmd.de Organization: German National Research Centre for Information Technology Adam Deishu Beeman (buddha@samsara.circus.com) wrote: >In article <2uf8qe$lnu@omega.gmd.de>, >Holger Hoffstaette <hoff@darmstadt.gmd.de> wrote: >>[no improved development tools in 3.3] >geez... ftp to prep.ai.mit.edu and get /pub/gnu/gcc-2.5.8.tar.gz, and build >it yourself. Same goes for the latest libc/libg++, if you don't want to >wait for NeXT to carefully inspect the GNU source, make their changes, test, >and release the absolute latest compiler. I know this. Actually, this was the first thing I did when I got NeXTSTEP. My point was not only about C++ support (since especially templates are still broken in 2.5.8), but rather about something much more fundamental, namely _consistency_ with all the releases. GCC 2.5.8 doesn't give me ObjC++, and as to debugging C++ code, support is 'less' than optimal there, too. Also, imagine the mess once people start building programs with their own GCC's 2.5.8, built with slightly different options, generating slightly different bugs which don't show up with the supplied 2.2.2 - should be lots of fun to hunt ! I'm sure everybody has better ways to spend their time. If we all have to use a compiler with bugs (and GCC _has_ bugs) let's at least have consistent bugs, if possible on all the three platforms. >Of course, this won't give you fat c++ binaries, and for some of us this will >make a stock g++ unusable. Depends on your needs. IMHO the future lies >in making compiler upgrades readily available for those who need them. Right. No FAT support -> no multiple platform support, one of the greatest strenghts of NeXTSTEP. >A shot in the dark, perhaps, but I would surmise that they're working on >some substantial improvements to the developer environment which are not >completely ready to be released. That would certainly explain things. Then why don't they just say so ? If the official reason would be 'it's not completely ready yet', I'll gladly stick my foot in my mouth, stand in the corner and shut up; on the other hand, simply leaving people in the dark is just _so_ much worse, especially since there was this EXPO tutorial where the new features were supposed to be introduced (has anybody been there ? a short summary would be nice). They can be _honest_ to their customers - surely not too much to ask for. >[Adam soapboxMode: 1]; >Be careful what you ask for. Do you want NeXT to fix everything absolutely >and make it totally ideal, or release what they have now? Do you want them >to make frequent releases and charge for every little upgrade, or infrequent >landmark upgrades? Should they give their software away and go broke or >target people with money? >[..] See above - I perfectly understand the situation they're in, given their current status and the way they seem to handle dancing at all the various parties at once. I think they're doing a great job, too. But it just doesn't make any sense to save ressources on development tools. How are we supposed to travel to MECCA ? Walk ? (naa..get a high ride on a Camel. 8-) >[Adam soapboxMode: 0]; >I do sympathize, though, and hopefully NeXT will find the happy medium here. [self agree]; Holger -- Holger Hoffstaette, GMD-IPSI/PaVE (hoff@darmstadt.gmd.de) [no NeXTmail] Smalltalker, NeXTstepper, Amiga veteran, Net citizen. "I'm gonna get myself connected.." - Stereo MC's
From: chris@iastate.edu (Chris Wong) Newsgroups: comp.sys.next.programmer Subject: SUMMARY: Can't change button's icon image. Weird behavior. Date: 25 Jun 1994 04:52:07 GMT Organization: Iowa State University, Ames, Iowa (USA) Message-ID: <2ugd5n$r9b@news.iastate.edu> References: <2ucgvc$ir7@news.iastate.edu> I've got a few messages from a lot of kind people in the net. Here is a summary for what I got. The reason why the image didn't appear when [aButton setImage:someImage] was sent is because NXImage can't locate the .tiff file. Reading carefully from the doc, I realize that - initFromFile:filename in which the filename can be either relative or absolute. Therefore, e.g., in my case, [[NXImage alloc] initFromFile:"flag.tiff"] would search for the file, flag.tiff, in the current directory. But, flag.tiff is only located in the app wrapper. Therefore, no instance of NXImageRepresentation is created. Why running in gdb without problem? Because Project Builder change the current directory to the .app wrapper before launching gdb. So the followup question is: How on the earth can I tell the NeXTSTEP to search the filename under my app wrapper? Thanks for any help. I'll post a summary after I get an answer. Chris Learning hard... >:) -- NeXTMail super welcomed!!| Chris Wong | "Hardware is supposed to serve Software." chris@iastate.edu | Computer Engineering & Computer Science <:)>:)<:)>:)<:)>:)<:) | Iowa State University of Science and Technology
From: starksm@genesis.mcs.com (Scott M. Stark) Newsgroups: comp.sys.next.programmer Subject: Re: Looking for an example of using rld Date: 25 Jun 1994 05:05:33 GMT Organization: MCSNet Services Distribution: world Message-ID: <2ugdut$8qs@News1.mcs.com> References: <Crx7uG.210u@yuma.ACNS.ColoState.EDU> Keywords: rld, dynamic load In article <Crx7uG.210u@yuma.ACNS.ColoState.EDU> rlabelle@lamar.colostate.edu writes: > Hello, > I'm looking for an example of using the rld functions. > I'd like to use dynamic loading in straight c-code. > > Thanks for any info..... > Rob LaBelle > rlabelle@lamar.colostate.edu Rob, Here are two files that demonstrate the use of rld_load() and rld_lookup(): --- main.c --- #include <mach-o/rld.h> #include <stdio.h> main() { NXStream *errors; struct mach_header *obj_header; const char *files[] = {"to_load.o", 0}; void (*dyn_func)(void); errors = NXOpenMemory(0, 0, NX_READWRITE); if( !rld_load(errors, &obj_header, files, 0) ) { char *buffer; int len, maxlen; perror("rld_load():: "); NXGetMemoryBuffer(errors, &buffer, &len, &maxlen); fprintf(stderr, "%s\n", buffer); exit(1); } /* Get the func address */ if( !rld_lookup(errors, "_dyn_func", (unsigned long *)&dyn_func) ) { char *buffer; int len, maxlen; perror("rld_load():: "); NXGetMemoryBuffer(errors, &buffer, &len, &maxlen); fprintf(stderr, "%s\n", buffer); exit(2); } /* Call the new function */ dyn_func(); } --- to_load.c --- #include <stdio.h> void dyn_func(void) { fprintf(stderr, "Loaded dyn_func() called\n"); } The result of running the program is: Loaded dyn_func() called Scott -- Scott Stark Stark International Software 4950 N. Marine Dr. #102 Chicago, IL 60640 starksm@genesis.mcs.com (NeXT mail accepted)
From: sbender@harmony.digex.net (Scott Bender) Newsgroups: comp.sys.next.programmer Subject: Re: SUMMARY: Can't change button's icon image. Weird behavior. Date: 25 Jun 1994 06:51:28 GMT Organization: Harmony Data Systems Message-ID: <2ugk5g$5kc@news1.digex.net> References: <2ugd5n$r9b@news.iastate.edu> In article <2ugd5n$r9b@news.iastate.edu> > So the followup question is: How on the earth can I tell the NeXTSTEP to > search the filename under my app wrapper? > Use: [[NXBundle mainBundle] getPath:path forResource:"fileName" ofType:"tiff"] - Scott
From: bungi@shark.omnigroup.com (Timothy J. Wood) Newsgroups: comp.sys.next.programmer Subject: Re: DBAssociation (subclass) for NXBrowser? Date: 25 Jun 1994 06:25:24 -0700 Organization: Omni Development, Inc. Message-ID: <2uhb84$123@shark.omnigroup.com> References: <CrwpuA.F3u@usenet.ucs.indiana.edu> <2uf4tr$2lo@darkstar.UCSC.EDU> art@cubicsol.com (Art Isbell) writes: [...] >EOF is also >built upon the new Foundation classes upon which OpenStep will be based, >so building OpenStep-compliant apps will mean not using DBKit (however, >neither DBKit nor EOF are included in the initial OpenStep spec). From the OpenStep overview given by Jean-Marie Hullot, DBKit will not be part of OpenStep in any form, but EOF will (and is, supposedly) OpenStep compliant. I presume this means that EOF will be an add-on available on any OpenStep system. -- Timothy J. Wood Omni Development, Inc. http://www.omnigroup.com -- Timothy J. Wood Omni Development, Inc. http://www.omnigroup.com
Newsgroups: comp.sys.next.programmer From: jgrace@netcom.com (Joseph Grace) Subject: Q: How do I {get,put} accents into a font? Message-ID: <jgraceCrynu1.FJ6@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Sat, 25 Jun 1994 16:15:37 GMT Hi, I am trying to put together a text editor to edit classical greek. I have the Symbol font, but it doesn't have the accents (macron, acute, grave, circumflex, soft breathing, and hard breathing) that I need. I would like some easy way (not necessarily "the right way" --- since I'm the only one using the app) to add the accents. Several ways I can think of (any (or all :-) of which I would be happy with): Buy another Postscript font which has the accents (I note that Symbol has a macron symbol which actually superimposes over its previous character! I need the other accents also, though.). Make a custom Symbol font which includes the characters. Other ways??? If anyone knows how to build such a font easily, or has pointers to the correct accents (a la the macron, above), that would help. Also, the accents don't have to be pretty, just legible. I also need to know how to get these superimposed characters to show up in an NS text object (I need this to be easy), i.e., are there any tricks. Quick and dirty is OK! Any and all help appreciated. Thanks, = Joe =
Newsgroups: comp.sys.next.programmer From: Manabu Tokunaga <pec@netcom.com> Subject: Questions on Monitor Support Message-ID: <netnewsCryt67.Mpn@netcom.com> Sender: netnews@netcom.com (USENET Administration) Organization: Pacific Electronic Connection (PEC) Date: Sat, 25 Jun 1994 18:10:55 GMT Hello, I am very new to NeXTSTEP environment, but I would like to know; 1. If there is an design limitations in NeXTSTEP on the number of display monitors it can support simultaneously. I was not able to get a clear answer from people I visited with at NeXTSTEP Expo. 2. We are also interested in desigining a system much higher pixel counts (in the order of 2000 x 2500 pixels) per monitor. Again, I am wondering if there is any design limitation in NeXTSTEP that may make support of such display difficult. Of course, this all assumes that the underlying hardware supports it. Any input is welcome on this topic either via E-mail or posts here. Thank you! Manabu Tokunaga <via pec@netcom.com>
Newsgroups: comp.sys.next.programmer Subject: HELP for app: applicationWillLaunch: needed Message-ID: <1994Jun26.101930.171362@eros.embl-heidelberg.de> From: tuparev@miles.NMR.EMBL-Heidelberg.DE (Georg Tuparev) Date: 26 Jun 94 10:19:29 +0100 Good morning C.S.N.P Gurus, Yesterday night I decided to implement app:sender applicationWillLaunch: (for the first time in my life ) but somehow S.P.J. and his crew make me feeling completely drunk :^( That's the source I wrote: ------------ Controller.h ... What Else ;-) #import <appkit/appkit.h> @interface Controller:Object { } @end @interface Controller(ApplicationDelegate) - appDidInit:sender; - app:sender applicationWillLaunch:(const char *)appName; @end ------------- ApplicationDelegate.m #import "Controller.h" @implementation Controller(ApplicationDelegate) - appDidInit:sender { // Start talking with Workspace Manager [[Application workspace] beginListeningForDeviceStatusChanges]; NXRunAlertPanel("","AppDidInit",0,0,0); return self; } - app:sender applicationWillLaunch:(const char *)appName { NXRunAlertPanel("","Launch:%s",0,0,0,appName); return self; } @end The Controller is the NXApp's delegate and the appDidInit alert panel comes up (as expected). The problem is, that nothing happens when I launch another application. Am I really completely drunk or I'm missing something? Thanks for the help -- Georg Tuparev EMBL / Protein Design Phone: +49 - 6221 - 387534 Meyerhofstr. 1 FAX: +49 - 6221 - 387517 D-69117 Heidelberg Germany Tuparev@EMBL-Heidelberg.de (NeXT-mail)
Newsgroups: comp.sys.next.programmer From: pfkeb@kaon.SLAC.Stanford.EDU (Paul F. Kunz) Subject: Re: NeXTSTEP and C++ In-Reply-To: buddha@samsara.circus.com's message of 24 Jun 1994 13:46:27 -0700 Message-ID: <PFKEB.94Jun26081049@kaon.SLAC.Stanford.EDU> Sender: news@unixhub.SLAC.Stanford.EDU Organization: Stanford Linear Accelerator Center References: <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> <2uf68l$2sl@darkstar.UCSC.EDU> <2uf8qe$lnu@omega.gmd.de> <2ufgn3$omr@samsara.circus.com> Date: Sun, 26 Jun 1994 15:10:49 GMT >>>>> On 24 Jun 1994 13:46:27 -0700, buddha@samsara.circus.com (Adam Deishu Beeman) said: > geez... ftp to prep.ai.mit.edu and get /pub/gnu/gcc-2.5.8.tar.gz, > and build it yourself. Same goes for the latest libc/libg++, if you > don't want to wait for NeXT to carefully inspect the GNU source, > make their changes, test, and release the absolute latest compiler. Unfortunately, that wouldn't do it. If I had a pure C++ program it would work, but I want to integrate my C++ code with the GUI in Objective-C. For this I need Objective-C++. Gcc from prep doesn't have Objective-C++ yet. I also can't compile part of the code with NeXT's Objective-C++ and the rest with gcc 2.5.8. GNU changed the symbol name mangling between gcc 2.2 and 2.3, so the two parts wouldn't link. So on Motorola or Intel, we are stuck with gcc 2.2.2 if any of the C++ code is going to be part of a NeXTSTEP Application. However, for HPPA, NeXT is shipping NS 3.2 with a compiler based on gcc 2.5.8 and NS/fHP was compiled with it (about 1M lines of code). It is certainly well tested. I'd like to see this version available for Motorola and Intel before next summer. I made the suggestion to NeXT that it be put up for grabs via anonymous ftp as a unsupported product, such as libg++ is distributed. -- Paul F. Kunz Paul_Kunz@slac.stanford.edu (NeXT mail ok) Stanford Linear Accelerator Center, Stanford University Voice: (415) 926-2884 (NeXT) Fax: (415) 926-3587
From: bakker@cyclon.aoml.erl.gov (Cary Bakker) Newsgroups: comp.sys.next.programmer Subject: How to use vm_deallocate() ? and a Q abt. PBoard... Date: 26 Jun 1994 19:01:34 GMT Organization: U.S. Department of Commerce, NOAA/AOML Message-ID: <2ukjae$sou@wave.aoml.erl.gov> This may be a stupic question, and I'm sorry if it is a FAQ. The ref manual says to use NXCloseMemory(), and then use NXGetMemoryBuffer() to get the internal buffer, and then vm_deallocate() to free it. I can't seem to find out how to use vm_deallocate anywhere... Object uses it: vm_deallocate(task_self(), sharedMemory, memorySize); but what is task_self()? Use it for all vm_deallocate calls? And is memorySize the maxLen returned from the NXGetMemoryBuffer()? Also, the ref manual uses the following: NXCloseMemory(theStream, NX_TRUNCATEBUFFER); which is equivalent to NXClose(theStream); I understand how that works, but what is the option: NX_FREEBUFFER for? Does that actually free the extra allocated pages rather than mark them to be deleted (I assume from the vm_deallocate)? Another simple question: I assume that when using ReadTypeToStream: handles the -deallocatePasteboardData:length: call, and that is why we should use NXCloseMemory(theStream, NX_FREEBUFFER) when we are done instead of calling deallocatePasteboardData:length: ourselves, right? How come we don't have to use vm_deallocate() with this stream? Thanks a lot for your help, Cary -- ------------------------------------------------------------------------- Cary A. Bakker | Music is the pleasure the human soul C.S. Major | experiences from counting without FLorida International University | being aware that it is counting... NeXT Programmer | NOAA/AOML | -- Gottfried W. Leibniz
From: eleeb08@iris.egr.uh.edu (Zhong ) Newsgroups: comp.sys.next.programmer Subject: How to archive cross linked objects? Date: 26 Jun 1994 20:11:23 GMT Organization: University of Houston Message-ID: <2ukndb$iuq@masala.cc.uh.edu> Is there any easy way to archive cross linked objects? For example, object1 has a (id *) pointing to object2 and at the same time object2 has a (id *) pointing to object1. Will NXWriteObjectReference() automatically resolve the problem, if I use it in the -write method of both objects? Any answer is appreciated. Please email to eleeb08@tree.egr.uh.edu, if possible.
From: tyler@vsc.com (Tyler Gingrich) Newsgroups: comp.sys.next.programmer Subject: Re: HELP for app: applicationWillLaunch: needed Date: 26 Jun 1994 23:13:03 GMT Organization: OARnet Message-ID: <2ul21v$kad@ns.oar.net> References: <1994Jun26.101930.171362@eros.embl-heidelberg.de> In article <1994Jun26.101930.171362@eros.embl-heidelberg.de> writes: > Yesterday night I decided to implement app:sender applicationWillLaunch: > (for the first time in my life ) but somehow S.P.J. and his crew make me > feeling completely drunk :^( > > // Start talking with Workspace Manager > [[Application workspace] beginListeningForDeviceStatusChanges]; ^^^^^^ > The Controller is the NXApp's delegate and the appDidInit alert panel > comes up (as expected). The problem is, that nothing happens when I > launch another application. Am I really completely drunk or I'm missing > something? > You should stop drinking while you code NS. :-) You should listen for _APPLICATION_ status changes by sending the beginListeningForApplicationStatusChanges message rather than the beginListeningForDeviceStatusChanges message that you actually sent. Hope this helps... --- Tyler Gingrich tyler@vsc.com Vanguard Software, Corp.
Newsgroups: comp.sys.next.programmer From: mikef@xnet.com (Mike Fleming) Subject: Sound Drivers: dealing with stereo Message-ID: <Cs1G9t.GLs@amiserv.chi.il.us> Keywords: sound drivers gravis ultrasound Sender: news@amiserv.chi.il.us Organization: XNet Public Access Internet, Naperville, IL (708-983-6435) Date: Mon, 27 Jun 1994 04:25:04 GMT I'm working on a sound driver for the Gravis Ultrasound board, and have run into a mite bit of a wall. The GUS will play stereo, however it does not deal well with sound streams that interleave the stereo data. Unfortunately, I can't figure out a way for the driver to de-interleave the data. (of course, this is more complicated: the data must be de-interleaved, and both datasets must be DMAed seperately onto the card...) Is there a way? (also, anyone who might have any tips on audio drivers: your input is welcome! :) -- Mike Fleming/mikef@sonata.cc.purdue.edu/Undergrad, EE, Purdue University "Where is the knowledge we have lost in information?"--T.S.Eliott "...shaped together in real-time: my television and my subconscious mind..."--Machines of Loving Grace
From: zeno@zebu.abstractsoft.com (Sean T. Lamont) Newsgroups: comp.sys.next.programmer Subject: EOF IB Date: 27 Jun 1994 03:46:52 -0700 Organization: Abstract Software Distribution: world Message-ID: <2umams$2c3@zebu.abstractsoft.com> Has anyone else found that using the EOF version of IB causes non-NeXT palettes to load flakily, if at all? I've had extremely limited success with all of my own palettes, plus all of the MiscKit ones. They occasionally work with the right circumstances, but usually they just say 'unable to load' -- Sean T. Lamont | Ask me about the WSI-Fonts Abstract Software | Professional collection for NeXT lamont@abstractsoft.com |____________________________________
From: kampp@informatik.uni-koblenz.de (Ralf Kampp) Newsgroups: comp.sys.next.programmer Subject: nxzonefreenolock()????? Date: 27 Jun 1994 11:40:50 GMT Organization: University Koblenz / Germany Message-ID: <2umds2$i25@newshost.uni-koblenz.de> Hi folks !! Following problem: When I try to free a certain object ([<object> free]) my application CRASHES !!! 1) the object is a subclass of an own subclass of Object 2) the object's code is loaded dynamically via objc_loadModules() 3) when I send a free-message to the object, the debugger says: "Memory access exception at address 0x..... in function nxzonefreenolock()." Does anyone have any experience with errors involving the functions nxzonefreenolock() or also nxzonemallocnolock() (which are called indirectly by the alloc- and free-messages) ???? Please help !!!! Ralf
From: otto@coactive.com (Otto Lind) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP 3.2 implementation of POSIX Date: 27 Jun 1994 07:23:44 GMT Organization: Coactive Aesthetics Inc. Message-ID: <2uluq0$lqj@disc.coactive.com> References: <2ubisv$138@marsu.tynet.sub.org> <2ue0er$l3k@quartz.ucs.ualberta.ca> jmack@skye.phys.ualberta.ca wrote: : In article <2ubisv$138@marsu.tynet.sub.org> mow@marsu.tynet.sub.org (Markus : Wenzel) writes: : > gmarler@sunmail1.fwrdc.rtsg.mot.com (Gordon Marler x6624) writes: : > The POSIX compatibility is not very convincing. Try to compile : > and link the following short C prog. It compiles fine, but it fails : > linking - and I just don't know why: [munch] : > Any suggestions? : > : : Try: : cc -posix -c posix_test1.c : cc -o posix_test1 posix_test1.o : : /tmp> posix_test1 : Today is : 94/06/24 Wrong answer. While it may work with the example given, it will not work in the general case of using POSIX functions. All you did was to access POSIX headers with the compile, and ignore them in the link. This will have _really_ bad effects when the data structures or function calls don't match. In looking at the nm output of /usr/lib/libposix.a library, it appears that NeXT botched the strftime.o module by using functions which are only defined in /lib/libsys_s.a. Makes you kinda wonder what they use for regression testing of their release... My pet peave with the POSIX support is that NeXT "forgot" to include the profile libraries /lib/gposixcrt0.o and /usr/lib/libposix_p.a. This means that any POSIX compiled source cannot be profiled. Otto -- Otto Lind Coactive Aesthetics otto@coactive.com P.O. Box 425967, San Francisco, CA 94142 netcom!coactive!otto voice:(415)626-5152 fax:(415)626-6320
From: misha@raptor.swarthmore.edu (Omar "not Irene or Misha" Haneef) Newsgroups: comp.sys.next.programmer Subject: Q: updating windows Followup-To: comp.sys.next.programmer Date: 27 Jun 1994 14:49:39 GMT Organization: Swarthmore College Message-ID: <misha-270694103020@mac01.hicks.swarthmore.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Howdy! Here a routine that I need to do and don't know how to do: 1. Disable the window update. 2. Change the contents of two subviews of the window in the off-screen buffer. 3. Enable window update (flush the window), so the two changed subviews will appear on the screen updated *simulaneously*. What I need to _avoid_ is updating the two subviews one at a time, one earlier than the other. Just in case you are curious why I need the simultaneuity, I explain. One of the window contentview's subviews is located in the lower left-hand corner of the screen and turns white (from black) whenever the other subview is updated. 'Turning white' is sensed by a photocell, which starts an external timer. For timing to be correct it is essential that both subviews (the one being timed and the one sending siganls) are updated simulataneously. Thanks very much to everyone who responds -- I appreciate your help! Misha -- Misha Neverov E-mail: misha@raptor.sccs.swarthmore.edu NeXT-mail: misha@jabberwock.swarthmore.edu Voice-mail: (610)690-3912 (until Aug '94)
Newsgroups: comp.sys.next.programmer From: wave@media.mit.edu (Michael B. Johnson) Subject: Re: EOF IB Message-ID: <1994Jun27.161458.26477@news.media.mit.edu> Sender: news@news.media.mit.edu (USENET News System) Organization: MIT Media Laboratory References: <2umams$2c3@zebu.abstractsoft.com> Date: Mon, 27 Jun 1994 16:14:58 GMT In article <2umams$2c3@zebu.abstractsoft.com> zeno@zebu.abstractsoft.com (Sean T. Lamont) writes: >> >>Has anyone else found that using the EOF version of IB causes non-NeXT >>palettes to load flakily, if at all? I've had extremely limited >>success with all of my own palettes, plus all of the MiscKit ones. >>They occasionally work with the right circumstances, but usually they >>just say 'unable to load' >> The only thing I had to do to the current version of the WavesWorld palettes was comment out the line in palette.table that had "Export Images". This is a documented (in the paper release notes) bug. After that, they both loaded fine, and everything seems fine, including all my weird stuff I do in test interface mode with loading nibs, etc. I love the new IB. -- --> Michael B. Johnson -- wave@media.mit.edu --> MIT Media Lab -- Computer Graphics & Animation Group --> 20 Ames St. E15-023G -- (617) 547-0563 (day office) --> Cambridge, MA 02139 -- (617) 253-0663 (night office)
From: hship@zen.cac.stratus.com (Howard Ship) Newsgroups: comp.sys.next.programmer Subject: EOF beta / Headerviewer Date: 27 Jun 1994 17:04:59 GMT Organization: Stratus Computer Inc, Marlboro MA Message-ID: <2un0rr$2bo@transfer.stratus.com> I'm having trouble convincing HeaderViewer to search for the new EOF and Foundation classes. I've added eoaccess.h and eointerface.h to the headers list, but they don't work too well, I get an error about 'missing introduction'. I've added the /NextLibrary/Documentation/NextDev/EnterpriseObjects/Reference/Access Layer and .../InterfaceLayer to the documentation directories, but I still sometimes get the headers, not the documentation. There doesn't seem to be a .h file for the Foundation classes, either, which prevents me from viewing them. Did NeXT not ship something, or am I missing the boat on some of this? -- This MUST be a good party -- My RIB CAGE is being painfully pressed up against someone's MARTINI!! Howard Ship hship@zen.cac.stratus.com [NeXT Mail OK] `84 GPz 750 "I lurk not, neither do I flame." -- NewsGrazer, a NeXTstep(tm) news reader, posting -- M>UQR=&8P7&%N<VE[7&9O;G1T8FQ<9C!<9G-W:7-S($AE;'9E=&EC83M<9C%< M9FUO9&5R;B!/:&QF<SM]"EQM87)G;#$R,`I<;6%R9W(Q,C`*7'!A<F1<='@Q M,34R7'1X,C,P-%QT>#,T-39<='@T-C`X7'1X-3<V,%QT>#8Y,3)<='@X,#8T M7'1X.3(Q-EQT>#$P,S8X7'1X,3$U,C!<9C!<8C!<:3!<=6QN;VYE7&9S,C1< M9F,P7&-F,"!))VT@:&%V:6YG('1R;W5B;&4@8V]N=FEN8VEN9R`*7&(@2&5A M9&5R5FEE=V5R"EQB,"`@=&\@<V5A<F-H(&9O<B!T:&4@;F5W(`I<8B!%3T8* M7&(P("!A;F0@"EQB($9O=6YD871I;VX*7&(P("!C;&%S<V5S+B`@22=V92!A M9&1E9"`*7&8Q(&5O86-C97-S+F@@"EQF,"!A;F0@"EQF,2!E;VEN=&5R9F%C M92YH"EQF,"`@=&\@=&AE(&AE861E<G,@;&ES="P@8G5T('1H97D@9&]N)W0@ M=V]R:R!T;V\@=V5L;"P@22!G970@86X@97)R;W(@86)O=70@)VUI<W-I;F<@ M:6YT<F]D=6-T:6]N)RX@($DG=F4@861D960@=&AE(`I<9C$@+TYE>'1,:6)R M87)Y+T1O8W5M96YT871I;VXO3F5X=$1E=B]%;G1E<G!R:7-E3V)J96-T<R]2 M969E<F5N8V4O06-C97-S3&%Y97(@"EQF,"!A;F0@"EQF,2`N+BXO26YT97)F M86-E3&%Y97(@"EQF,"!T;R!T:&4@9&]C=6UE;G1A=&EO;B!D:7)E8W1O<FEE M<RP@8G5T($D@<W1I;&P@<V]M971I;65S(&=E="!T:&4@:&5A9&5R<RP@;F]T M('1H92!D;V-U;65N=&%T:6]N+EP*7`I4:&5R92!D;V5S;B=T('-E96T@=&\@ M8F4@82`N:"!F:6QE(&9O<B!T:&4@1F]U;F1A=&EO;B!C;&%S<V5S+"!E:71H M97(L('=H:6-H('!R979E;G1S(&UE(&9R;VT@=FEE=VEN9R!T:&5M+EP*7`I$ M:60@3F585"!N;W0@<VAI<"!S;VUE=&AI;F<L(&]R(&%M($D@;6ES<VEN9R!T M:&4@8F]A="!O;B!S;VUE(&]F('1H:7,_7`I<"@I<<&%R9%QT>#4R,%QT>#$P M-C!<='@Q-C`P7'1X,C$R,%QT>#(V-C!<='@S,C`P7'1X,S<R,%QT>#0R-C!< M='@T.#`P7'1X-3,R,%QF8S!<8V8P("TM7`I4:&ES($U54U0@8F4@82!G;V]D M('!A<G1Y("TM($UY(%))0B!#04=%(&ES(&)E:6YG('!A:6YF=6QL>5P*('!R M97-S960@=7`@86=A:6YS="!S;VUE;VYE)W,@34%25$E.22$A7`H@7`I(;W=A M<F0@4VAI<"`@:'-H:7!`>F5N+F-A8RYS=')A='5S+F-O;2`@6TYE6%0@36%I M;"!/2UT@(&`X-"!'4'H@-S4P7`H)(DD@;'5R:R!N;W0L(&YE:71H97(@9&\@ 22!F;&%M92XB7`H*?0H@ `
Newsgroups: comp.sys.next.programmer From: eike@ilink.de (Eike Dierks) Subject: Re: [?] Project Builder with C++ (make depend fix) Message-ID: <Cs2K8F.IHD@ilink.de> Sender: usenet@ilink.de Organization: i.link Kommunikationssysteme GmbH, Berlin References: <1994Jun20.152624.3019@dart.de> Date: Mon, 27 Jun 1994 18:48:14 GMT make depend is also broken in PB/NeXT makefiles regarding C++ To fix this: Put all C++ files in the "Other Sources" category of PB. If you use extension other then .c (i.e. C,cc,M,...) then add the following to your Makefile.preamble: # all will stay the default: all depend debug:: # this is to fix deficies in NeXT Makefiles, in 3.2 PB there is no way to # specify *.cc (c++) files except as $(OTHERLINKED) but these files are # not included in a make depend depend:: /bin/rm -rf Makefile.dependencies $(MAKE) Makefile.dependencies 'CCFILES=$(OTHERLINKED)' project:: @if [ ! -f Makefile.dependencies ] ;then \ $(MAKE) Makefile.dependencies 'CCFILES=$(OTHERLINKED)' ;\ fi KNOWN BUGS: This only works if all files in "Other Sources" are C/C++. In article <1994Jun20.152624.3019@dart.de> Atze (Alexander Spohr) writes: > In article <2tv7d3$pdf@umd5.umd.edu> rprice@reunion.umd.edu (Rodney Price) > writes: > > > > I've been writing C++ code on my NeXT (NS 3.2) for a while, using > > [...] > > I'd like to use Project Builder, since my projects are getting > > Just use New and switch to Tool instead of Application in the New Project > panel. However subprojects are not supported for tools :-( >
From: tkylling@magnus.acs.ohio-state.edu (Tobin J Kyllingstad) Newsgroups: comp.sys.next.hardware,comp.sys.next.programmer Subject: Orchid Kelvin 64 support for NeXT? Date: 27 Jun 1994 20:43:35 GMT Organization: The Ohio State University Message-ID: <2undln$8lc@charm.magnus.acs.ohio-state.edu> Keywords: orchid kelvin vga driver I am very interested in Orchid's Kelvin 64 vga controller. I am also more interested in moving to NeXTSTEP. I called Orchid; they do not support NeXTSTEP. Is there any third party supporting the Kelvin 64 for NeXTSTEP? (according to the CL-GD543X data book, the CL-GD543X is register and software compatible with the CL-GD542X) -- Tobin Kyllingstad Applied Innovation Inc. 614-798-2000 tobin@aiinet.com
Newsgroups: comp.sys.next.programmer Organization: Antigone Press gateway, San Francisco Return-Path: <flexus!rfschtkt@maze.ruca.ua.ac.be> Date: Mon, 27 Jun 94 20:48:08 +0200 From: Raf Schietekat <flexus!rfschtkt@maze.ruca.ua.ac.be> Message-ID: <9406271848.AA03864@flexus> Subject: Re: How to use vm_deallocate() ? and a Q abt. PBoard... LS, This is a reply to Cary Bakker's message of 1994-06-26. Sorry, it's long for a post, but I think it is of general interest (NEXTSTEP's self-compatibility and documentation). > This may be a stupic question, and I'm sorry if it is a > FAQ. I am having a deja-vu feeling here... but that was about mapped files, mainly on c.s.n.b, around 1994-06-05. > The ref manual says to use NXCloseMemory(), and then use > NXGetMemoryBuffer() to get the internal buffer, and then > vm_deallocate() to free it. No need to do that. Use NXCloseMemory() with NX_FREEBUFFER. Warning: that used to be the prescribed procedure, but NeXT changed the docs, for no apparent reason. It should continue to work, however, and the documentation should be restored to support both methods, or else these guys have a lot of explaining/apologising to do (breaking our code and all that, well, not yet anyway). > I can't seem to find out how to use vm_deallocate > anywhere... Object uses it: > vm_deallocate(task_self(), sharedMemory, > memorySize); > but what is task_self()? Use it for all vm_deallocate > calls? And is memorySize the maxLen returned from the > NXGetMemoryBuffer()? You didn't look very well. To find out about vm_deallocate, task_self, use Librarian. The hit you want to inspect is OperatingSystem/Part1_Mach/04_MachFunctions/MachFunctions.rtf. The size argument is there because this function will actually deallocate all *pages* ``touched'' by the specified ``segment'' (this is not like the ``segments'' that malloc() works with and whose sizes it does remember all by itself). > Also, the ref manual uses the following: > NXCloseMemory(theStream, NX_TRUNCATEBUFFER); > which is equivalent to NXClose(theStream); I understand > how that works, but what is the option: NX_FREEBUFFER > for? > Does that actually free the extra allocated pages rather > than mark them to be deleted (I assume from the > vm_deallocate)? Actually, it's the other way around. NXClose on a memory stream is equivalent to NXCloseMemory(theStream,NX_TRUNCATEBUFFER). This makes sense if you have previously obtained the buffer (NXGetMemoryBuffer), and you want to deallocate any memory you may have used before, but not the present contents. NX_FREEBUFFER is for actually freeing everything, and you may choose between vm_deallocate on NXGetMemoryBuffer after NXClose, or NX_FREEBUFFER on the stream itself. > Another simple question: > I assume that when using ReadTypeToStream: handles the > -deallocatePasteboardData:length: call, and that is why > we should use > NXCloseMemory(theStream, NX_FREEBUFFER) > when we are done instead of calling > deallocatePasteboardData:length: ourselves, right? > How come we don't have to use vm_deallocate() with this > stream? I don't think your sentence is syntactically correct, but I think it is about something I mentioned in the above-mentioned message to c.s.n.b: >>>>>> Well, with regard to Tom's question, there's a similar problem with how -[pasteboard readType:data:length:] data should be released. This used to be with vm_deallocate, but now ``you should *not* attempt to free the returned memory using vm_deallocate() or free()''. I suppose it's OK for old code (even recompiled under 3.X) to continue to use vm_deallocate as long as it doesn't use the new readTypeToStream: as well? Such things should be documented (it's not in the ReleaseNotes), if only to reassure us that NeXT knows what it is doing, in this case. <<<<< Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be, Flanders, Belgium *** The designation BANRUC60.BITNET for this host will disappear by July 1 *** (real, i.e., with triangle in the Deliver button) NeXTmail preferred Addressing limitations: no !, % or .uucp (? I should test this again)
From: Mark_Wagner@NeXT.COM (Mark Wagner) Newsgroups: comp.sys.next.programmer Subject: Re: EOF beta / Headerviewer Date: 27 Jun 1994 22:43:56 GMT Organization: NeXT, Inc. Message-ID: <2unknc$lia@rosie.next.com> References: <2un0rr$2bo@transfer.stratus.com> The error message about missing introductions comes up because there aren't any introductions in the beta documentation for EOF. The EOF header files should include the foundation headers so you probably shouldn't need to include them directly. -- Mark In article <2un0rr$2bo@transfer.stratus.com> hship@zen.cac.stratus.com (Howard Ship) writes: I'm having trouble convincing HeaderViewer to search for the new EOF and Foundation classes. I've added eoaccess.h and eointerface.h to the headers list, but they don't work too well, I get an error about 'missing introduction'. I've added the /NextLibrary/Documentation/NextDev/EnterpriseObjects/Reference/AccessLayer and .../InterfaceLayer to the documentation directories, but I still sometimes get the headers, not the documentation. There doesn't seem to be a .h file for the Foundation classes, either, which prevents me from viewing them. Did NeXT not ship something, or am I missing the boat on some of this? -- This MUST be a good party -- My RIB CAGE is being painfully pressed up against someone's MARTINI!! Howard Ship hship@zen.cac.stratus.com [NeXT Mail OK] `84 GPz 750 "I lurk not, neither do H [YK B 750
Newsgroups: comp.sys.next.programmer From: pelletk@il.us.swissbank.com (Ken Pelletier) Subject: Re: sybase pubs database? Message-ID: <1994Jun27.223840.15600@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <2tr6ii$lu4@jake.esu.edu> Date: Mon, 27 Jun 1994 22:38:40 GMT Howard B Ott writes > Could some one please mail me the pubs (version 1, not pubs2) database > create for sybase? Or direct me to an ftp site? > > Thanks in advance, Howard, You don't need to get the pubs db from someone else, you can build it yourself by running a script that comes with Sybase. If you look in /usr/sybase/scripts, you'll find a script called installpubs (or something like that - I don't have access to a Sybase installation at the moment). Take a look that the script - it just uses ISQL to create and populate pubs. - Ken (ken@nika.com)
From: nils@daisy.syd.de Newsgroups: comp.sys.next.hardware,comp.sys.next.programmer Subject: Problems with writing Graphics Driver for ATI MACH 64 Date: 27 Jun 1994 15:33:39 GMT Organization: Sydlink e.V., 24937 Flensburg, Germany Message-ID: <2umrgj$9na@flens.syd.de> Hi! I'am currently working on a Graphics Driver for the ATI Mach 64 PCI Card. The good news: -8 Bit gray works -15 Bit color works The bad news: -I got trouble with the 32 Bit color mode ! -There is a kind of "noise" on the screen ! Okay, to be more specific: The ATI has an IOPixelEncoding of "RRRRRRRRGGGGGGGGBBBBBBBB--------" When I set the IOPixelEncoding field to this value instead of the Default "--------RRRRRRRRGGGGGGGGBBBBBBBB", there is a strange behaviour of the Window Server ( or of something else ??#@%$ ): after booting, the Background is set, but the login-panel is invisible ! after "blind" login, only the Workspace-Menu appears, but NOTHING else ! (The Menu looks correct, but there are some missing, black, rectangular Sections) ! Opening Sub Menus works fine, all functions seem to work, but ALL other graphic items except the Application Menus are invisible ! Is this a bug in the Window Server with handling the RGBa IOPixelEncoding instead of the standart aRGB IOPixelEncoding ????????? Now the second bug: Noise !?! In both 8bpp and 15bpp (aRGB) Modes, there are "some" pixels, which don't display there correct color/gray values, but some kind of color/gray noise! I Know, that this is a problem of my hardware programming, not of NEXTSTEP, but maybe some of you did some programming stuff on the MACH 64 already !? The behaviour changes a bit if i modify the CRTC Timing, but i haven't been able to eliminate all noisy pixels (even if i use standart VESA timing rules) The behaviour looks a bit like the behaviour of the Revision 3 ATI MACH 32 Chipset in 1132x820 Modes ! (before a new ATI 32 driver arrived) So: PLEASE HELP ! Thank you for any hints, E-Mail Adresses of "WISE GUYS" and any other kinds of Help: Nils Heidorn -- -- Nils Heidorn nils@daisy.syd.de Muehlenstrasse 21b 24937 Flensburg 0461 / 55049 ( am Wochenende 04674 / 1477 )
From: alxs@cs.tu-berlin.de (Alexander Sparkowsky) Newsgroups: comp.sys.next.programmer Subject: DBTableView's floating-point format Date: 28 Jun 1994 07:18:16 GMT Organization: Technical University of Berlin, Germany Message-ID: <2uoiro$inc@news.cs.tu-berlin.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit High, can anybody tell me, how I can set the floating-point format for a field/column in a DBTableView and for a Form/FormCell/TextField which displays data from this DBTableView. I tried the setFloatingPointFormat:left:right method but without success. -- so long alxs Alxeander Sparkowsky, 10707 Berlin
From: arnouilj@oahu.esiee.fr (Jean-Claude Arnouil) Newsgroups: comp.sys.next.programmer Subject: Searching YourCall.app sources from NeXTSTEP Advantage Date: 28 Jun 1994 10:12:07 GMT Organization: Groupe ESIEE - PARIS Message-ID: <2uot1n$hdk@srv5.esiee.fr> Hello NeXTSTEP fellows, Does someone knows where I can find the sources for the YourCall example application used in the Step-by-step chapter of the last NeXTSTEP Advantage document. The sources are in the table-of-contents but NOT included in the postscript files I got. I've explored the NeXT archives and archie without success. Thank you in advance for any help -- Jean-Claude Arnouil Unix Systems Administrator Ecole Superieure d'Electrotechnique et d'Electronique email: arnouilj@esiee.fr phone:(33 1)45.92.67.71 BP 99 - 93162 Noisy-le-Grand CEDEX - FRANCE
Newsgroups: comp.sys.next.programmer From: jerald@mrk.com (Jerald Dawson) Subject: Re: EOF IB Message-ID: <Cs2GLM.96H@mrk.com> Sender: jerald@mrk.com (Jerald Dawson) Organization: Investor's Advantage References: <1994Jun27.161458.26477@news.media.mit.edu> Distribution: usa Date: Mon, 27 Jun 1994 17:29:45 GMT In article <1994Jun27.161458.26477@news.media.mit.edu> wave@media.mit.edu (Michael B. Johnson) writes: > In article <2umams$2c3@zebu.abstractsoft.com> zeno@zebu.abstractsoft.com (Sean T. Lamont) writes: > >> > >>Has anyone else found that using the EOF version of IB causes non-NeXT > >>palettes to load flakily, if at all? I've had extremely limited > >>success with all of my own palettes, plus all of the MiscKit ones. > >>They occasionally work with the right circumstances, but usually they > >>just say 'unable to load' > >> > > The only thing I had to do to the current version of the WavesWorld palettes > was comment out the line in palette.table that had "Export Images". This > is a documented (in the paper release notes) bug. After that, they both loaded > fine, and everything seems fine, including all my weird stuff I do in test > interface mode with loading nibs, etc. > > I love the new IB. > > > -- > --> Michael B. Johnson -- wave@media.mit.edu > --> MIT Media Lab -- Computer Graphics & Animation Group > --> 20 Ames St. E15-023G -- (617) 547-0563 (day office) > --> Cambridge, MA 02139 -- (617) 253-0663 (night office) How does one go about getting the prerelease of the new IB and EOF stuff. Was it just something that was available to expo attendees or can anyone get it? -- jerald dawson * Systems Manager/Head Programmer * Investor's Advantage NeXTmail - jerald@mrk.com | phone - (708) 487-3221 | Don't blame me. I voted for Bush.
Newsgroups: comp.sys.next.programmer From: gordie@cyclesoft.com (Gordie Freedman) Subject: Re: How to archive cross linked objects? Message-ID: <1994Jun28.081031.15652@netcom.com> Sender: gordie@netcom.com Organization: Cyclesoft Media Works References: <2ukndb$iuq@masala.cc.uh.edu> Date: Tue, 28 Jun 1994 08:10:31 GMT In article <2ukndb$iuq@masala.cc.uh.edu> eleeb08@iris.egr.uh.edu (Zhong ) writes: > Is there any easy way to archive cross linked objects? For example, > object1 has a (id *) pointing to object2 and at the same time object2 has > a (id *) pointing to object1. Will NXWriteObjectReference() automatically > resolve the problem, if I use it in the -write method of both objects? > > Any answer is appreciated. Please email to eleeb08@tree.egr.uh.edu, if > possible. First, I think you mean "has an (id) pointing ..." rather than "has an (id *) pointing ...". An id is a pointer (it's an Object*). For the answer straight out of the horse's mouth, Look in the documentation in Concepts/Pre3.0_Concepts/Chapter 10 "Support Objects and Functions" via Digi Librarian - but here's my interpretation (disclaimer - I haven't actually tried this, so no warranty expressed or implied): In the set of objects being archived, all of them must be owned by one *and only one* object, but any object can be referenced by many objects. in your example, that means either object1 can own object2 (and something else must call NXWriteRootObject on object1), or object2 can own object1 (and see prev comment). Both, however, cannot own each other. When I say object1 owns object2, that means that object1's write: method will archive object2 with NXWriteObject or NXWriteRootObject. If object1 references (but does not own) object2, than object1 "writes" object2 via NXWriteObjectReference. Note that if object1 will ever call NXWriteObjectReference on a subobject, NXWriteRootObject (and not NXWriteObject) must be called on object1, else you get an exception. So one way to do this in your example is: 1) Call NXWriteRootObject(typedStream, object1). 2) In object1's [write:] method, do NXWriteRootObject(typedStream, object2). Remember you have to call NXWriteRootObject in both cases, else you will get an exception. 3) In object2's [write:] method call NXWriteObjectReference(typedStream, object1). You do not say what class each belongs to - if they are of the same class, you will have to do some fancy footwork in the [write:] method, so that one of the objects NXWriteRootObjects the other, and the NXWriteObjectReferences the former. I'll leave this as an exercise for the reader! -- >>> Gordie Freedman -> gordie@cyclesoft.com NeXTMail Yes! >>> Thou shalt not inline functions more complicated than 20
Newsgroups: comp.sys.next.programmer From: gordie@cyclesoft.com (Gordie Freedman) Subject: Re: JPG in NEXT Application Message-ID: <1994Jun28.081957.15717@netcom.com> Sender: gordie@netcom.com Organization: Cyclesoft Media Works References: <Crt0ID.H19@wst.edvz.sbg.ac.at> Date: Tue, 28 Jun 1994 08:19:57 GMT In article <Crt0ID.H19@wst.edvz.sbg.ac.at> cuebl@hai (Christoph Uebleis) writes: > How can I display a jpg-file on the screen in my NeXT-Application??? > > I have tried the following: > > System("/LocalApps/OmniImage.app/OmniImage -NXOpen <path of picture>"); > > it works, but the picture is displayed at the background. > Are there any routines or functions with which I can solve that problem > better. > I believe you can just open the jpeg file in your app treating it as a tiff - OmniImage (or other filters/apps such as ImageViewer) should filter it into a tiff from the jpeg automagically passing that into the NXImage (actually the workspace will know that OmniImage does this and fire it up to do it behind the scenes). I don't have the exact details on this, but I know that both Mail.app and Preview.app suddenly acquired the ability to display jpegs after I installed ImageViewer.app on my system - so I assume this is why. If this isn't clear or doesn't work send me mail and I'll hack around and send you a code example. -- >>> Gordie Freedman -> gordie@cyclesoft.com NeXTMail Yes! >>> Thou shalt not inline functions more complicated than 20
Newsgroups: comp.sys.next.programmer From: gordie@cyclesoft.com (Gordie Freedman) Subject: Re: JPG in NEXT Application Message-ID: <1994Jun28.100408.16421@netcom.com> Sender: gordie@netcom.com Organization: Cyclesoft Media Works References: <1994Jun28.081957.15717@netcom.com> Date: Tue, 28 Jun 1994 10:04:08 GMT In article <1994Jun28.081957.15717@netcom.com> gordie@cyclesoft.com (Gordie Freedman) writes: > In article <Crt0ID.H19@wst.edvz.sbg.ac.at> cuebl@hai (Christoph > Uebleis) writes: > > How can I display a jpg-file on the screen in my NeXT-Application??? > > > > I have tried the following: > > > > System("/LocalApps/OmniImage.app/OmniImage -NXOpen <path of > > picture>"); > > > > it works, but the picture is displayed at the background. Are > > there any routines or functions with which I can solve that > > problem better. > > > > I believe you can just open the jpeg file in your app treating > it as a tiff - OmniImage (or other filters/apps such as > ImageViewer) should filter it into a tiff from the jpeg > automagically passing that into the NXImage (actually the > workspace will know that OmniImage does this and fire it up to > do it behind the scenes). > > I don't have the exact details on this, but I know that both > Mail.app and Preview.app suddenly acquired the ability to > display jpegs after I installed ImageViewer.app on my system > - so I assume this is why. If this isn't clear or doesn't work > send me mail and I'll hack around and send you a code example. Following up on my own answer - I just hacked around a bit, and here's an example. Let's say your app delegate is of class MyAppDelegateClass and has two instance vars, "myWindow" and "myView". Also, assume that myView is your own class MyViewClass, which inherits directly from View, and has one instance variable "myCell". Wire things up in IB such that an instance of MyAppDelegateClass (call it appDelegate) is the Application object's Delegate. Also, the appDelegate's myWindow outlet is wired to a Window, which has one view inside it. The view is a custom view, of class MyViewClass. The appDelegate's myView instance variable is wired to the custom view. One other thing - the JPEG. Using Project Builder, add a jpg file to "Other Resources". In this example, I added the file "Picture.jpg" - note that when you build the project, the jpg file will get copied into the app wrapper. Of course, you could skip the bundle stuff, and use the file dialog or whatever to get the path to the jpg. (Sorry if I am stating the obvious with some of these explanations - just trying to be complete.) So in MyAppDelegateClass you implement: @implementation MyAppDelegateClass - appDidInit:sender { char buf[MAXPATHLEN + 1]; id bundle; id jpegImage; // Set up a bundle object and get the path to Picture.jpg bundle = [NXBundle bundleForClass:[self class]]; rc = [bundle getPath:buf forResource:"Picture" ofType:"jpg"]; // Make a jpg image, give it a name, and than // allocate a Cell object using that image // Note that the buf in initFromFile: is an absolute // path to the jpg file ... jpegImage = [[NXImage alloc] initFromFile:buf]; [jpegImage setName:"MyPicture"]; [myView setCell:[[Cell alloc] initIconCell:"MyPicture"]]; // Display the window [myWindow display]; return self; } @end And in the MyView class, you implement these two methods: @implementation MyViewClass // This is just called out of appDidInit to set // the Cell object we are using. - setCell:theCell { myCell = theCell; return self; } // This is called when the View needs to draw itself for // some reason - it will get called when the appDidInit // code invokes [myWindow display] for example ... - drawSelf:(const NXRect *)rects :(int)rectCount { NXRect rect; // There's probably a better way to do this, // such as getting the size from the image // in the cell, and resizing [self window] // and self (the view) and getting the rect // after that ... [self getFrame:&rect]; rect.origin.x = 0; rect.origin.y = 0; [myCell drawSelf:&rect inView:self]; return self; } @end Now - assuming you have something installed in your system that will filter a jpg into a tiff, when you do the [[NXImage alloc] initFromFile:buf] in appDidInit, the Workspace or some special NeXT magic somewhere will convert the jpg into something the NXImage can use, and you're all set. -- >>> Gordie Freedman -> gordie@cyclesoft.com NeXTMail Yes! >>> Thou shalt not inline functions more complicated than 20
Newsgroups: comp.sys.next.programmer From: gordie@cyclesoft.com (Gordie Freedman) Subject: Re: JPG in NEXT Application - one more thing Message-ID: <1994Jun28.102005.16629@netcom.com> Sender: gordie@netcom.com Organization: Cyclesoft Media Works References: <1994Jun28.100408.16421@netcom.com> Date: Tue, 28 Jun 1994 10:20:05 GMT In article <1994Jun28.100408.16421@netcom.com> gordie@cyclesoft.com (Gordie Freedman) writes: > @implementation MyAppDelegateClass > - appDidInit:sender > { > ... > return self; > } > @end Damn it's too late - I don't know why I'm messing with a custom view object. Scratch that - make myView a Button (instead of MyViewClass) dragged off of Good old IB's palette - and scratch the whole silly MyViewClass code. The only difference is in appDidInit. Again, you can get the size from the NXImage, and set myView and myWindow to the size of the Image before displaying it. If you don't, it will display in the center of the Button. Replace: jpegImage = [[NXImage alloc] initFromFile:buf]; [jpegImage setName:"MyPicture"]; [myView setCell:[[Cell alloc] initIconCell:"MyPicture"]]; With: jpegImage = [[NXImage alloc] initFromFile:buf]; [jpegImage setName:"MyPicture"]; [myView setIcon:"MyPicture"]; -- >>> Gordie Freedman -> gordie@cyclesoft.com NeXTMail Yes! >>> Thou shalt not inline functions more complicated than 20
From: pcu@umich.edu Newsgroups: comp.sys.next.programmer Subject: Re: Q: updating windows Date: 28 Jun 1994 14:07:57 GMT Organization: University of Michigan Distribution: world Message-ID: <2upart$4pi@lastactionhero.rs.itd.umich.edu> References: <misha-270694103020@mac01.hicks.swarthmore.edu> In article <misha-270694103020@mac01.hicks.swarthmore.edu> misha@raptor.swarthmore.edu (Omar "not Irene or Misha" Haneef) writes: > 1. Disable the window update. > 2. Change the contents of two subviews of the window in the off-screen > buffer. > 3. Enable window update (flush the window), so the two changed subviews > will appear on the screen updated *simulaneously*. /* window MUST be of type NX_BUFFERED */ /[window disableFlushWindow]; /* call display (Window, View) as you see fit */ [window reenableFlushWindow]; [window flushWindow]; /* or flushWindowIfNeeded */ > Just in case you are curious why I need the simultaneuity, I explain. One > of the window contentview's subviews is located in the lower left-hand > corner of the screen and turns white (from black) whenever the other > subview is updated. 'Turning white' is sensed by a photocell, which starts > an external timer. For timing to be correct it is essential that both > subviews (the one being timed and the one sending siganls) are updated > simulataneously. You, of course, must make sure that your photocell detector takes into account the refresh rate of the the actual hardware, 'cause there's no way, no how, you'll get true simultaneity. -- Peter Urka <pcu@umich.edu> Dept. of Chemistry, Univ. of Michigan Anything to me is sweeter, Than to see Shock-headed Peter. - H. Hoffmann
From: sean@zapotec.math.byu.edu (Sean O. Luke) Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and C++ Date: 28 Jun 1994 15:22:16 GMT Organization: Brigham Young University Message-ID: <2upf78$shc@hamblin.math.byu.edu> References: <PFKEB.94Jun21214759@kaon.SLAC.Stanford.EDU> <2uf68l$2sl@darkstar.UCSC.EDU> <PFKEB.94Jun26081049@kaon.SLAC.Stanford.EDU> Paul F. Kunz (pfkeb@kaon.SLAC.Stanford.EDU) wrote: : Unfortunately, that wouldn't do it. If I had a pure C++ program it : would work, but I want to integrate my C++ code with the GUI in : Objective-C. For this I need Objective-C++. Gcc from prep doesn't : have Objective-C++ yet. : I also can't compile part of the code with NeXT's Objective-C++ and : the rest with gcc 2.5.8. GNU changed the symbol name mangling : between gcc 2.2 and 2.3, so the two parts wouldn't link. : So on Motorola or Intel, we are stuck with gcc 2.2.2 if any of the : C++ code is going to be part of a NeXTSTEP Application. I think the answer is simple. Lobby FSF to force NeXT to release their compiler as source. Then recompile. ..."yeah, right.." :-) +--------------------------------------------------------------+ | Sean Luke This signature no verb | | sean@digaudio.byu.edu ,,, <- finger for PGP key | | sean@zapotec.math.byu.edu (o o) | +------------------------oOO--(_)--OOo-------------------------+
Newsgroups: comp.sys.next.programmer From: jhall@madeline.umhc.umn.edu (Jeff Hallgren) Subject: Need help compiling "straight" C. Message-ID: <Cs46L3.196@news.cis.umn.edu> Sender: news@news.cis.umn.edu (Usenet News Administration) Organization: University of Minnesota, Twin Cities Date: Tue, 28 Jun 1994 15:45:55 GMT I need to compile some C code I got off an ftp server. It has large numbers of locally declared and defined functions, they look like this... void aLargeThing(myStruct *aStruct) { int i, j; if (something) { ---> int isrelop(char c){ return strchr("<>", c) != NULL; } int issign(char c){ return strchr("-+", c) != NULL; } char *item0 = item; (...) } } I put the code together in a PB tool project. When I try to compile I get syntax errors: thing.c:224: syntax error, missing `;' after `)' for the above combined function declaration and implementation. I tried adding -traditional to CFLAGS in makefile.preamble but that gives me errors like: In file included from /NextDeveloper/Headers/bsd/sys/types.h:91, from /NextDeveloper/Headers/bsd/sys/dir.h:45, from ./thing1.h:6, from thing2.c:9: /NextDeveloper/Headers/bsd/machine/label_t.h:11: `#include' expects "fname" or <fname> Does anyone know what options are necessary to compile C code like this? Thanks, Jeff -- "You are Gunther Heindrickson from Appenzell, you moved to Grindelwald to drive the cog train to Muren." jeff.hallgren@tahiti.umhc.umn.edu (NeXTMail OK)
From: reed@dedekind.math.luc.edu (dale f reed; DH307; x8-3567) Newsgroups: comp.sys.next.programmer Subject: Re: Can't change button's icon image. Weird behavior. Date: 28 Jun 1994 15:40:12 GMT Organization: Loyola University of Chicago Message-ID: <2upg8s$q34@apollo.it.luc.edu> References: <2ucgvc$ir7@news.iastate.edu> Chris Wong (chris@iastate.edu) wrote: : Very weird. I use [[myMatrix selectedCell] setImage:someImage] to set the : icon image of one of the button at runtime. : However, the image doesn't appear when I launch the app. The weirdest thing : is the image appear if I launch the app in gdb, i.e., in debug mode. : Here is the method that get executed when one of the button in the Matrix is : click. : - mark:sender : { : if (someCondition) : { : [[mineField selectedCell] setImage:flag]; : } : return self; : } : Please share some insight here. I am totally clueless. : Thank you very very very much for any kind of information... : Chris -- desperate What seems to me might be the problem is that the icon is not found in the directory being searched by the App. I have an App that works great from within SuperDebugger, but has directory problems when launched from the Workspace. (It does however work correctly when I specify the path explicitly from the command line) The author of SuperDebugger, Bing-Qing Chen wrote to me: "While debuging, the current directory will be the Project directory. If the app is launched from Workspace, its current directory will be your home directory. You should use chdir() to change to your Project directory if you want a consistent behavior." This doesn't completely solve the problem, however, if you want to be able to install your App ANYWHERE and have it run correctly. To do this you must have code that uses a relative path name. I haven't been able to get this to work (yet), but am interested in the solution. Dale Reed reed@math.luc.edu
From: stufduff@albion.unmc.edu (Sean W. Duffy) Newsgroups: comp.sys.next.programmer Subject: HELP:IB DBModule & Interbase Date: 28 Jun 1994 16:23:01 GMT Organization: University of Nebraska Medical Center, Omaha, NE, USA Message-ID: <2upip5$3hl@netserv.unmc.edu> Keywords: IB DBMODULE INTERBASE Help! I'm using Interbase & a netinfo network. I use dll to create a database on the network server. - OK I use qli to enter some test data & commit it via rsh to the server. - OK Then I use DBModeler to create a model on my local machine and save it in ~/Library/Databases. - OK I examine the model - it looks correct. I can even use DBEdit from hksys and see the data! I Open Project Builder, add libdbkit_s.a, and the model file. Then open the project's nib in interfaces, select the DBModel drag it off - it shows up fine and has my model in it. No Problem. So I slap on a DBTableView following all the instructions in chapter 4 'Creating a Database Project' and wire up a 'Fetch' button. The following page sez test the interface; so I do - I NEVER GET ANY DATA! What in heck am I doing wrong? please reply to stufduff@albion.unmc.edu
Newsgroups: comp.sys.next.programmer From: chuston@nugget.rmNUG.ORG (Chris Huston) Subject: Foundation kit and autorelease Message-ID: <1994Jun28.163109.7366@nugget.rmNUG.ORG> Organization: Rocky Mountain NeXT Users' Group Date: Tue, 28 Jun 1994 16:31:09 GMT I just finished reading the Foundation kit concepts... The file describes the new autorelease, retain, release paradigm... > When your code completely finishes executing and control > returns to the application object (that is, at the end > of the event loop), the application object releases the > object. Am I misunderstanding or does it say the NXApp will take care of freeing objects? I've got more than a few backend programs that have no NXApp... they don't have an interface. My concern is what happens when a NXApp-less program uses objects that assume that the Application object will take care of things? Does anyone have more light to shine on this subject? - Chris
Newsgroups: comp.sys.next.programmer From: wave@media.mit.edu (Michael B. Johnson) Subject: Re: Foundation kit and autorelease Message-ID: <1994Jun28.180319.23485@news.media.mit.edu> Sender: news@news.media.mit.edu (USENET News System) Organization: MIT Media Laboratory References: <1994Jun28.163109.7366@nugget.rmNUG.ORG> Date: Tue, 28 Jun 1994 18:03:19 GMT In article <1994Jun28.163109.7366@nugget.rmNUG.ORG> chuston@nugget.rmNUG.ORG (Chris Huston) writes: >>I just finished reading the Foundation kit concepts... The file describes the new autorelease, retain, release paradigm... >> >>> When your code completely finishes executing and control >>> returns to the application object (that is, at the end >>> of the event loop), the application object releases the >>> object. >> >>Am I misunderstanding or does it say the NXApp will take care of freeing objects? >> >>I've got more than a few backend programs that have no NXApp... they don't have an interface. >> >>My concern is what happens when a NXApp-less program uses objects that assume that the Application object will take care of things? >> >>Does anyone have more light to shine on this subject? >> Well, if you're handling your own main loop, you have to provide your own autorelease pools. Should be pretty straightforward... -- --> Michael B. Johnson -- wave@media.mit.edu --> MIT Media Lab -- Computer Graphics & Animation Group --> 20 Ames St. E15-023G -- (617) 547-0563 (day office) --> Cambridge, MA 02139 -- (617) 253-0663 (night office)
From: sean@zapotec.math.byu.edu (Sean O. Luke) Newsgroups: comp.sys.next.programmer Subject: NeXT's own IB Inspectors (Question) Date: 28 Jun 1994 18:32:54 GMT Organization: Brigham Young University Message-ID: <2upqcm$31p@hamblin.math.byu.edu> Does anyone know the names of the Attributes inspectors for each of NeXT's own palette objects (Button, Matrix, Slider, Text-in-ScrollView, etc.)? I need to know the names of these inspectors badly but can't find them anywhere. +--------------------------------------------------------------+ | Sean Luke This signature no verb | | sean@digaudio.byu.edu ,,, <- finger for PGP key | | sean@zapotec.math.byu.edu (o o) | +------------------------oOO--(_)--OOo-------------------------+
From: sean@zapotec.math.byu.edu (Sean O. Luke) Newsgroups: comp.sys.next.programmer Subject: Categories Date: 28 Jun 1994 18:33:39 GMT Organization: Brigham Young University Message-ID: <2upqe3$31p@hamblin.math.byu.edu> Does anyone know of some evil hack which allows me to access the original method of an object, even if that method's been overridden by a category method of the same name? Thanks! +--------------------------------------------------------------+ | Sean Luke This signature no verb | | sean@digaudio.byu.edu ,,, <- finger for PGP key | | sean@zapotec.math.byu.edu (o o) | +------------------------oOO--(_)--OOo-------------------------+
From: hauert@itp.unibe.ch (Christoph Hauert) Newsgroups: comp.sys.next.programmer Subject: Scrolling of a View by dragging the mouse Date: 28 Jun 1994 19:27:41 GMT Message-ID: <2uptjd$lg5@aragorn.unibe.ch> Keywords: scrolling, dragging Hi, a friend of mine and i are writting a nice tool concerning fonts - more details will be given (hopefully soon) in comp.sys.next.announce... one problem we are dealing with at the moment is scrolling or shifting a view controlled by the mouse. actually we are looking for hints how scrolling is done in apps like Grab.app with the "hand cursor" active. Another example would be PasteUp. any hints and tips on shifting a view directly using the mouse are appreciated! thanks chris -- Christoph Hauert Institute for Theoretical Physics Berne, Switzerland hauert@butp.unibe.ch (NeXTMail)
From: vanguard@oar.net (Tyler Gingrich) Newsgroups: comp.sys.next.programmer Subject: Re: Foundation kit and autorelease Date: 28 Jun 1994 16:44:04 -0400 Organization: OARnet, Columbus, OH. Message-ID: <2uq22k$6ah@dialup.oar.net> References: <1994Jun28.163109.7366@nugget.rmNUG.ORG> In article <1994Jun28.163109.7366@nugget.rmNUG.ORG> chuston@nugget.rmNUG.ORG (Chris Huston) writes: >I just finished reading the Foundation kit concepts... The file describes the new autorelease, retain, release paradigm... > >I've got more than a few backend programs that have no NXApp... they don't have an interface. > >My concern is what happens when a NXApp-less program uses objects that assume that the Application object will take care of things? > While the Application class (EOApplication during the Beta) does maintain the main autorelease pool, you are able to create your own AutoReleasePool instances and handle releasing the arp yourself. The application object just does this release whenever it processes a new event. --- Tyler Gingrich tyler@vsc.com Vanguard Software, Corp.
Newsgroups: comp.sys.next.programmer From: drew@fnbc.com (Drew Davidson) Subject: Re: Magic links in Edit Message-ID: <1994Jun28.192019.26882@fnbc.com> Sender: news@fnbc.com Organization: First National Bank Of Chicago, Chicago IL, USA References: <Crw7vz.Ho8@news.informatik.uni-stuttgart.de> Date: Tue, 28 Jun 94 19:20:19 GMT In article <Crw7vz.Ho8@news.informatik.uni-stuttgart.de> brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) writes: > Hello all! > > If you drag&drop a picture into a RTFD document you can create a copy by > holding down the <Alternate> key (or doing nothing) or create a link using the > <Control> key. Unfortunately the copy doesn't seem to be a real copy: > - drag&drop an image into an RTFD document > - close the document > - change the image file on disk using Icon (or something like that) > - open the RTFD document again > => the changed image shows up > The TXT.rtf contains the link through a NeXT extension to RTF (new tag) of the following form: {\NeXTGraphic<sequence #> <graphic file name here> \width<width * 20> \height<height * 20> } where anything between angle brackets <> is a variable. The width and height are in units that RTF requires (can't remember the terminology though), which is why the actual dimensions are multiplied by 20 in this example. The sequence number must be unique in the file, however and Edit just keeps incrementing it. > Is this a kind of "magic link" (found no docu)? What is the underlying > mechanism? Is it a feature or a bug? What if I want to have a real copy? > > Many questions, Klaus... No magic, just sufficiently advanced technology. :-) -- +--------------------------------+---------------------------------------+ | Drew Davidson - Software Guy | "Computers are useless. They can | | First National Bank of Chicago | only give you answers." | | drew@fnbc.com (NeXTmail) | - Pablo Picasso | +--------< All opinions expressed here are mine and mine alone >---------+
Newsgroups: comp.sys.next.programmer From: alex@cs.umd.edu (Alex Blakemore) Subject: Re: DBTableView's floating-point format Message-ID: <Cs447D.21K@genoa.com> Sender: alex@genoa.com (Alex Blakemore) Organization: Genoa Software Systems References: <2uoiro$inc@news.cs.tu-berlin.de> Date: Tue, 28 Jun 1994 14:57:12 GMT Alexander Sparkowsky writes > how I can set the floating-point format for a > field/column in a DBTableView subclass DBFormatter [works for me] > ... a Form send setFloatingPointFormat:left:right to each cell? [untested] > ... a FormCell send it setFloatingPointFormat:left:right [untested, should work] > [a] TextField. send it setFloatingPointFormat:left:right [tested] -- Alex Blakemore alex@cs.umd.edu NeXT mail accepted
Newsgroups: comp.sys.next.programmer From: joe@FreemanSoft.com (Joe Freeman) Subject: Re: Foundation kit and autorelease Message-ID: <1994Jun28.182514.1497@FreemanSoft.com> Sender: jfreeman@FreemanSoft.com Organization: FreemanSoft Inc. References: <1994Jun28.163109.7366@nugget.rmNUG.ORG> Distribution: usa Date: Tue, 28 Jun 1994 18:25:14 GMT In article <1994Jun28.163109.7366@nugget.rmNUG.ORG> chuston@nugget.rmNUG.ORG (Chris Huston) writes: > I just finished reading the Foundation kit concepts... The file describes the new autorelease, retain, release paradigm... > > > When your code completely finishes executing and control > > returns to the application object (that is, at the end > > of the event loop), the application object releases the > > object. > > Am I misunderstanding or does it say the NXApp will take care of freeing objects? > > I've got more than a few backend programs that have no NXApp... they don't have an interface. > > My concern is what happens when a NXApp-less program uses objects that assume that the Application object will take care of things? I believe the goal is to include a non-UI event loop object similar to the DOEventLoop object that is included with PDO. The event loop object acts as an event dispatcher like Application without making the connection to the window server and all that other good UI stuff. -- Joe Freeman Home of FSPreferences and FSGreyBoard. NEXTSTEP mentoring and consulting services also available. email: Joe@FreemanSoft.COM voice: 919.783.7033
From: borzilda@NeXTwork.Rose-Hulman.Edu (David A Borzillo) Newsgroups: comp.sys.next.programmer Subject: Setting permissions with PB.... Date: 28 Jun 1994 21:00:39 GMT Organization: Rose-Hulman Institute of Technology Message-ID: <2uq31n$o8q@master.cs.rose-hulman.edu> Is there any arguments or anything I can add to Project Builder so that when I build my app everyone will have permission to run it? Also, whereever I move it it will have the same permissions? We're on the AFS network here and changing the attributes in the Workspace inspector doesn't help. I have to use a series of 'chmod' commands to change the permission. -- David A. Borzillo (borzilda@nextwork.rose-hulman.edu) Software Developer, Fluid Science Learning Center Mech. Engineering Senior, Rose-Hulman Inst of Tech, Terre Haute,IN +=================================================================+ Disclaimer: Any views and opinions expressed in this message are my own and not of Rose-Hulman or the FSLC Project.
From: Charles William Swiger <infidel+@CMU.EDU> Newsgroups: comp.sys.next.programmer Subject: Re: NeXTSTEP and C++ Date: Tue, 28 Jun 1994 19:33:50 -0400 Organization: Carnegie Mellon, Pittsburgh, PA Message-ID: <gi4=BSS00WBL85oDJ9@andrew.cmu.edu> In-Reply-To: <2upf78$shc@hamblin.math.byu.edu> Excerpts from netnews.comp.sys.next.programmer: 28-Jun-94 Re: NeXTSTEP and C++ by Sean O. Luke: > : So on Motorola or Intel, we are stuck with gcc 2.2.2 if any of the > : C++ code is going to be part of a NeXTSTEP Application. > > I think the answer is simple. Lobby FSF to force NeXT to release their > compiler as source. Then recompile. > > ..."yeah, right.." :-) Presumably the NeXT-modified source to gcc 2.5.8 used with NEXTSTEP/HP will be made available on the NS/HP Dev Disk as has been the case with other NEXTSTEP 3.2 releases. Failing that, everyone has the legal right to get the modifications NeXT made to gcc; all you have to do is ask for them and they will be made available for no cost (beyond media and a "reasonable" service charge). I've had friends do this to get the source for other FSF-based NeXT programs in the past and they had no problems, but your mileage may vary. -Chuck Charles William Swiger - WhiteLight Systems | "All the world's a stage, and" --------------------------------------------+ "we are merely players...." AMS & normal mail: infidel@cmu.edu | NeXTmail: chuck@cswiger.slip.andrew.cmu.edu | "Semper ubi sub ubi."
From: mai@next.com (Mai Nguyen) Newsgroups: comp.sys.next.programmer Subject: MiniExamples for Beta EOF Date: 28 Jun 1994 23:34:38 GMT Organization: NeXT, Inc. Message-ID: <2uqc2e$7u@rosie.next.com> Keywords: MiniExamples Hi Folks: For those of you who have received the Beta EOF at NEXTSTEP EXPO and who wonder about example code, we have just released 4 miniexamples: ftp site: ftp.next.com ftp path: pub/NeXTanswers/Files/Software_and_Tools/MiniExamples/EOFramework If you want to get them by numbers, they are: 1628.compressed --> UniqueKey_oracle 1629.compressed --> PeopleDemo_sybase 1630.compressed --> PeopleDemo_oracle 1631.compressed --> UniqueKey_sybase (Note: The EOF supplemental release notes explains how to use ftp) PeopleDemo_sybase Sybase version of PeopleDemo example. It shows how to write a simple enterprise object, and manipulate records with the EOController class. See the README file for more details. PeopleDemo_oracle Oracle version of PeopleDemo example. It is an exact copy of the sybase example apart from the oracle eomodel and the oracle sql script. UniqueKey_sybase Sybase version of UniqueKey example. It shows how to automatically generate unique keys when inserting a new record. It also shows how to manipulate records without using any custom EO. UniqueKey_oracle Oracle version of UniqueKey example. All the above examples are based on a People database provided in the example folders. You only need to create the database once either for Sybase or Oracle, even though all 4 examples contain the corresponding install SQL script. Your comments about the usefulness of the examples are welcome. Have fun, Mai Nguyen
From: lupson@next.com (Linus Upson) Newsgroups: comp.sys.next.programmer Subject: Re: Foundation kit and autorelease Date: 28 Jun 1994 23:39:34 GMT Organization: NeXT, Inc. Message-ID: <2uqcbm$87@rosie.next.com> References: <1994Jun28.163109.7366@nugget.rmNUG.ORG> In article <1994Jun28.163109.7366@nugget.rmNUG.ORG> chuston@nugget.rmNUG.ORG (Chris Huston) writes: > Am I misunderstanding or does it say the NXApp will take care of freeing > objects? > > I've got more than a few backend programs that have no NXApp... they > don't have an interface. > > My concern is what happens when a NXApp-less program uses objects that > assume that the Application object will take care of things? > > Does anyone have more light to shine on this subject? Take a look at the documentation for NSAutoreleasePool in: /NextLibrary/Documentation/NextDev/Foundation/Classes/NSAutoreleasePool.rt f And also look at the section in the release notes on enabling autorelease pools in 3.2 applications: /NextLibrary/Documentation/NextDev/ReleaseNotes/EnterpriseObjects.rtfd Hopefully everything you need to know is there. (There are code examples in there too.) Have Fun, Linus Upson
From: peter@bert.psyc.upei.ca (Beaker) Newsgroups: comp.sys.next.programmer Subject: Re: Need help compiling "straight" C. Date: 29 Jun 1994 00:29:18 GMT Organization: University of Prince Edward Island, Charlottetown, PEI Canada Distribution: world Message-ID: <2uqf8u$2g2@atlas.cs.upei.ca> References: <Cs46L3.196@news.cis.umn.edu> In article <Cs46L3.196@news.cis.umn.edu> jhall@madeline.umhc.umn.edu (Jeff Hallgren) writes: > I need to compile some C code I got off an ftp server. It has large numbers > of locally declared and defined functions, they look like this... > void aLargeThing(myStruct *aStruct) > { > int i, j; > if (something) > { > ---> int isrelop(char c){ return strchr("<>", c) != NULL; } > int issign(char c){ return strchr("-+", c) != NULL; } > char *item0 = item; > (...) > } > } > > I put the code together in a PB tool project. When I try to compile I get syntax > errors: > thing.c:224: syntax error, missing `;' after `)' > for the above combined function declaration and implementation. > . . . > Does anyone know what options are necessary to compile C code like this? > Thanks, > Jeff > I don't know the right flag to use, but what you're looking at is C++ code. C++, unlike C or Objective-C, allows you to declare variables anywhere. Since NeXT's compiler allows C++, it should compile if the right flag is set. -- Peter 'Beaker' Burka |finger peter@bert.psyc.upei.ca U.P.E.I., Prince Edward Island, Canada |for PGP pburka@upei.ca | GCS d--- p--- c++ l u+++ e+(*) m--- s+ n- h* f- g+ w+ t+ r-- y-(*)
Newsgroups: comp.sys.next.programmer From: Robert La Ferla <Robert_La_Ferla@hot.com> Subject: Re: ** STONE DESIGN RAVE ** Message-ID: <1994Jun28.043110.2640@hot.com> Sender: robertl@hot.com (Robert La Ferla) Organization: Hot Technologies References: <1994Jun14.145356.692@stone.com> Date: Tue, 28 Jun 1994 04:31:10 GMT Here's a bit of wit from the rave. I was schmoozing with Dwight Koop of Swiss Bank and mentioned that you (Andrew Stone) have corrupted the souls and minds of the Fortune 500. Dwight replied: "No, you've got it all wrong. Corporate America has subverted him." Robert La Ferla Hot Technologies Registered NEXTSTEP Developer and Consultant In article <1994Jun14.145356.692@stone.com> andrew@stone.com (Andrew Stone) writes: > Since my mailbox has been flooded with requests for info, I thought it > would be useful to make a general announcement about this year's party... > > The Stone Design Rave will be held on the Solstice this year, Wednesday > night at 9pm until dawn. Come celebrate the shortest night of the year! > Spin by our booth to get your invitation - and a CD perhaps. > > ======================================================================= == > > T H E C O S M I C D A N C E > > > Dance is the manifestation of the primal vibrations of the > universe. It is the means by which the Absolute brings creation into > existence, sustains and maintains it, and eventually destroys it. > God/dess, the supreme dancer, has no purpose other than the enjoyment of > her/his own dynamic movements. Life's continual prodigal transformation of > energy IS the dance. > > -- > ||<<->>||<<==>>|S<<++>>|T<<?>O<+>>N|<<-->>E|<<==>>||<<+>>|| > !! Andrew Stone | (505) 345-4800 !! > !! andrew@stone.com | Have Modem Why Travel !! > ||<<->>|D<<==>>|E<<++>>|S<<?>|<+>>G|<<-->>N|<<==>>!|<<+>>||
Newsgroups: comp.sys.next.programmer From: gelato@netcom.com (Steve James) Subject: Uncompressing Help.store? Message-ID: <gelatoCs4vKH.CKu@netcom.com> Organization: Me, organized? Date: Wed, 29 Jun 1994 00:48:16 GMT Does anybody have a tool to find out what's in a Help.store compressed help directory? I couldn't find anything in Librarian. Is compresshelp even documented anywhere? If it is, I couldn't find it. Alternately, can anybody point me to a ftp site that archives the c.s.n hierarchy; if memory serves there was a thread on Help.store last year.
Newsgroups: comp.sys.next.programmer From: jay@charlie (Jay Liew) Subject: Help .... Message-ID: <1994Jun28.194856.7949@cs.uno.edu> Sender: news@cs.uno.edu Organization: University of New Orleans Date: Tue, 28 Jun 1994 19:48:56 GMT Hi, I am writing an editor and have some problems getting the right NXCoord the tabsize given certain fonts. Here is part of my code. int NextTab = 0; double OldTabWidth = (float)[OldFont getWidthOf : ' '] * (float) _CurrentTabSize; double NewTabWidth = (float)[NewFont getWidthOf : ' '] * (float) _CurrentTabSize; [_Text setAutodisplay : NO]; for (NextTab = 1; NextTab != NumberOfTabsInText; ++ NextTab){ [_Text changeTabStopAt : (float) NextTab * OldTabWidth to : (float) NextTab * NewTabWidth]; } [_Text setAutodisplay : YES]; [_Text display]; What I am thinking here is that I can get the size of a blank, multiply with the Current Tab Size (indentation when writing C code), and I should get the right X coordinate. Unfortunately, this is not working. Can someone write to me, if you know the solution or find something wrong with my code above? Thank you. Jay (jay@charlie.smobject.com)
From: reed@dedekind.math.luc.edu (dale f reed; DH307; x8-3567) Newsgroups: comp.sys.next.programmer Subject: Arrays as method args: how to? Date: 28 Jun 1994 23:26:13 GMT Organization: Loyola University of Chicago Message-ID: <2uqbil$add@apollo.it.luc.edu> I want to pass a 2 dimensional array of int's as a method argument. It works fine with a 1 dimensional array, but I'm having trouble declaring the 2 dimensional array. Here's what I have. The method is defined as: - (void) getNumbersFrom:(FILE*)ifile andStoreIn: (int*) twoDimensionalArray; It is called elsewhere using: int twoDimensionalArray[ 2][ 10]; [self getNumbersFrom: theFilePointer andStoreIn: twoDimensionalArray]; Any ideas on how to declare this? When I compile, I get a message saying "Incompatible pointer type". I appreciate any leads you might give me. Dale Reed reed@math.luc.edu
From: chris@iastate.edu (Chris Wong) Newsgroups: comp.sys.next.programmer Subject: Re: Can't change button's icon image. Weird behavior. Date: 29 Jun 1994 05:25:40 GMT Organization: Iowa State University, Ames, Iowa (USA) Message-ID: <2ur0kk$1v3@news.iastate.edu> References: <2ucgvc$ir7@news.iastate.edu> <2upg8s$q34@apollo.it.luc.edu> In article <2upg8s$q34@apollo.it.luc.edu>, dale f reed; DH307; x8-3567 <reed@dedekind.math.luc.edu> wrote: > >: - mark:sender >: { >: if (someCondition) >: { >: [[mineField selectedCell] setImage:flag]; >: } >: return self; >: } > >What seems to me might be the problem is that the icon is not found in the >directory being searched by the App. I have an App that works great from Yeah. This is the problem. I've got a few respone from those kind NS souls on the net. Some suggest that I query the mainBundle to get the absolute filename. But, I've got a better idea. I used: [[NXImage alloc] initFromSection:"afilename.tiff"]; which would search for the file in the app wrapper. That might not the best/most general way to do thing. But, it do what I need to do cleanly. Thanks to all who help me solve the problem. Chris -- NeXTMail super welcomed!!| Chris Wong | "Hardware is supposed to serve Software." chris@iastate.edu | Computer Engineering & Computer Science <:)>:)<:)>:)<:)>:)<:) | Iowa State University of Science and Technology
Newsgroups: comp.sys.next.programmer From: jfr@trwlasd.com Subject: Re: How to archive cross linked objects? Message-ID: <Cs55w0.905@trwlasd.com> Sender: jfr@trwlasd.com (Jon Rosen) Organization: TRW References: <1994Jun28.081031.15652@netcom.com> Date: Wed, 29 Jun 1994 04:31:11 GMT Gordie Freedman writes > > So one way to do this in your example is: > > 1) Call NXWriteRootObject(typedStream, object1). > > 2) In object1's [write:] method, do NXWriteRootObject(typedStream, > object2). > Remember you have to call NXWriteRootObject in both cases, else you will > get an exception. NO NO NO NO NO!!!!!!!!!!!! Don't EVER do this! I quote from the "OOP and Obj-C Language" manual from NeXT, page 140 (my emphasis): "NXWriteRootObject() can be used ONLY to initiate archiving, NEVER inside a write: method." The whole point of NXWriteRootObject is that it initiates the two pass scan of ALL the objects archived underneath it. You don't want to (and can't) start another two-pass scan inside the one that has already been started. In object1, use an NXWriteObject for object2 and in object2, use an NXWriteObject for object1 and the thing will be sorted out properly. In fact, it doesn't matter WHICH object you start with at the top, it will still work. Archiving is VERY smart and as long as all the objects are being archived together, you don't even need NXWriteObjectReference (that is only for archiving objects that MAY not be part of the archive tree, so-called extrinsic references). Circular references are handled automatically by NXWriteRootObject() and each and every object will be archived ONLY once. Jon Rosen
Newsgroups: comp.sys.next.programmer From: alex@cs.umd.edu (Alex Blakemore) Subject: Re: How to archive cross linked objects? Message-ID: <Cs536z.2JK@genoa.com> Sender: alex@genoa.com (Alex Blakemore) Organization: Genoa Software Systems References: <1994Jun28.081031.15652@netcom.com> Date: Wed, 29 Jun 1994 03:32:58 GMT Gordie Freedman writes > An id is a pointer (it's an Object*). almost, its a pointer to any objective C object. the only real criteria is that the first ivar be a pointer to a class object. from objc.h. typedef struct objc_class *Class; typedef struct objc_object { Class isa; } *id; There are other root objects besides Object (e.g. NSObject, NXProxy and EOFault) -- Alex Blakemore alex@cs.umd.edu NeXT mail accepted
Newsgroups: comp.sys.next.programmer From: alex@cs.umd.edu (Alex Blakemore) Subject: Re: Can't change button's icon image. Weird behavior. Message-ID: <Cs543G.2LK@genoa.com> Sender: alex@genoa.com (Alex Blakemore) Organization: Genoa Software Systems References: <2upg8s$q34@apollo.it.luc.edu> Date: Wed, 29 Jun 1994 03:52:27 GMT dale f reed; DH307; x8-3567 writes > I have an App that works great from within SuperDebugger, but > has directory problems when launched from the Workspace. > if you want to be able to install your App ANYWHERE and have it > run correctly. To do this you must have code that uses a relative > path name. I haven't been able to get this to work (yet), but am > interested in the solution. the solution is to use the NXBundle class (or now NSBundle) read the docs. to get a path to the app wrapper or debug wrapper directories use [[NXBundle mainBundle] directory] to find a tiff file named bozo.tiff (looking in appropriat .lproj dirs) [[NXBundle mainBundle] getPath:buff forResource:"bozo" ofType:"tiff"]; -- Alex Blakemore alex@cs.umd.edu NeXT mail accepted
Newsgroups: comp.sys.next.programmer,comp.sys.next.software From: alex@cs.umd.edu (Alex Blakemore) Subject: Looking for a COTS or shareware Bug Tracker Message-ID: <Cs54Dp.2MB@genoa.com> Sender: alex@genoa.com (Alex Blakemore) Organization: Genoa Software Systems Date: Wed, 29 Jun 1994 03:58:37 GMT Does anyone know of a commercial or shareware package suitable for tracking bug reports during the development, testing and deployment of a software product? Ideally, it would have simple NEXTSTEP interface for entering and viewing bug reports, facilities for getting reports broken down by area, priority, status etc, extra credit if able to provide info in a form that can be easily plotted. I guess BugNext.app is half of what I'm looking for. I know it wouldn't be hard to build, but there must be something like this out there. somewhere. P.S. I'm aware of GNU's gnats, but am really interested in a NEXTSTEP app. -- Alex Blakemore alex@cs.umd.edu NeXT mail accepted
From: ivo@hasc.ca Newsgroups: comp.sys.next.programmer Subject: Re: returning void instead of self Date: 28 Jun 1994 21:28:04 GMT Organization: Hutchison Avenue Software Message-ID: <2uq4l4$eil@giuliani.gun.com> References: <1994Jun22.233601.7850@millennium.com> In article <1994Jun22.233601.7850@millennium.com> Jayson Adams writes: > In article <1994Jun16.105911.705@prim.demon.co.uk>, Dave Griffiths writes: > > Don't forget the original promise of DO's was that we could use them just > > as though they were local objects, completely transparently. Of course > > anyone who's ever written anything with DO's knows this to be untrue! > > It's pretty damn close Dave. Not really. And I don't think it will be without garbage collection. -ivo ("saver")
Newsgroups: comp.sys.next.programmer From: Atze (Alexander Spohr) Subject: Re: Need help compiling "straight" C. Message-ID: <1994Jun29.075931.355@dart.de> Sender: aspohr@dart.de Organization: d'ART Software GmbH References: <2uqf8u$2g2@atlas.cs.upei.ca> Date: Wed, 29 Jun 94 07:59:31 GMT In article <2uqf8u$2g2@atlas.cs.upei.ca> peter@bert.psyc.upei.ca (Beaker) writes: > In article <Cs46L3.196@news.cis.umn.edu> jhall@madeline.umhc.umn.edu (Jeff > Hallgren) writes: > > I need to compile some C code I got off an ftp server. It has > large numbers > > of locally declared and defined functions, they look like this... > > void aLargeThing(myStruct *aStruct) > > { > > int i, j; > > if (something) > > { > > ---> int isrelop(char c){ return strchr("<>", c) != NULL; } > > int issign(char c){ return strchr("-+", c) != NULL; } > > char *item0 = item; > > (...) > > } > > } > > > > I put the code together in a PB tool project. When I try to compile I > get syntax > > errors: > > thing.c:224: syntax error, missing `;' after `)' > > for the above combined function declaration and implementation. > > . . . > > Does anyone know what options are necessary to compile C code like this? > > Thanks, > > Jeff > > > > I don't know the right flag to use, but what you're looking at > is C++ code. C++, unlike C or Objective-C, allows you to declare > variables anywhere. Since NeXT's compiler allows C++, it should > compile if the right flag is set. > Sorry, these are not variables, these are local functions. I do not know if gcc supports them. For now and if it is not too much work, redefine them by using #define. <#define isrelop(c) (strchr("<>", c) != NULL)> Atze PS. You can declare variables at the start of any block in C, so this would not be a problem. -- Alexander Spohr, Hamburg, Germany Alexander_Spohr@DART.de (Mail ) Faces and faces (+49) 40 / 380 23-0 (Sound) See them and complain not (+49) 40 / 380 23-290 (Fax ) And am content with all
Newsgroups: comp.sys.next.programmer From: brouwer@minnie.informatik.uni-stuttgart.de (Klaus Brouwer) Subject: Re: Magic links in Edit Message-ID: <Cs5G8G.M7w@news.informatik.uni-stuttgart.de> Sender: news@informatik.uni-stuttgart.de Organization: Informatik, Uni Stuttgart, Germany References: <Crw7vz.Ho8@news.informatik.uni-stuttgart.de> <1994Jun28.192019.26882@fnbc.com> Date: Wed, 29 Jun 1994 08:14:39 GMT In <1994Jun28.192019.26882@fnbc.com> drew@fnbc.com (Drew Davidson) writes: >> Is this a kind of "magic link" (found no docu)? What is the underlying >> mechanism? Is it a feature or a bug? What if I want to have a real copy? >> >> Many questions, Klaus... >No magic, just sufficiently advanced technology. :-) Seems not to be so. I also posted the article to comp.sys.next.bugs and here is the reaction: ------------ Beginning of Article -------------- From comp.sys.next.bugs Tue Jun 28 10:50:10 1994 From: flexus!rfschtkt@maze.ruca.ua.ac.be (Raf Schietekat) Date: 27 Jun 1994 16:04:07 -0500 Newsgroups: comp.sys.next.bugs Subject: Re: Magic links in Edit? LS, Here's a new entry in KBNS. Version 32.0 is available on cs.orst.edu. >>>>> KBNS.32.1.036_o3.2o Edit.app does not handle dropped images correctly Description: Whether you alternate-drop (copy) or control-drop (link) images, Edit always uses a hard link to include them (even though it differentiates the changed cursor). How to verify: Copy some image to a scratch directory, duplicate it. Do both operations on a new RTFD document, close that. Change each image using IconBuilder.app. Open the RTFD document again. The changes are visible. Evidence of the hard links is obtained using ls -lga (link count 2). Urgency: A document may be damaged because this is unexpected behaviour. Confirmed: Klaus_Brouwer (reported, unknown version), Raf_Schietekat (o3.2o) Workaround: Duplicate the included message, delete the original, rename the copy to whatever is appropriate (yes, this is tedious). Note that merely renaming the included image does not work (didn't verify this though). Cure (for NeXT): See KBNS.00.0.136: the authors apparently don't know that normally one overwrites a file rather than replacing it, so a hard link is not a suitable decision. Until there are copy-on-write semantics in the file system, copy-dropping an image should use a real file copy operation. Link-dropping should probably use a symbolic link instead. Report history: Reported by Klaus_Brouwer to c.s.n.b on 1994-06-24. <<<<< Raf Schietekat, RfSchtkt@maze.ruca.ua.ac.be, Flanders, Belgium *** The designation BANRUC60.BITNET for this host will disappear by July 1 *** (real, i.e., with triangle in the Deliver button) NeXTmail preferred Addressing limitations: no !, % or .uucp (? I should test this again) ------------ End of Article -------------- Seems to be a design failure....or just a bug :-( KLaus Brouwer
Newsgroups: comp.sys.next.programmer From: Dave THOMAS <dave@softpac.com.au> Subject: **looking for work in Sunny Aussie?*** Message-ID: <Cs5H58.479@softpac.com.au> Keywords: job australia Sender: dave@softpac.com.au (Dave THOMAS) Organization: Softpac Pty. Ltd., Sydney, AUSTRALIA Date: Wed, 29 Jun 1994 08:34:20 GMT How would you like to spend a few years in the sun? Enjoy real beaches? Satisfy your pioneering spirit? Seriously, we do not have much currently on the go with NeXTSTEP, but this will grow slowly, but we do have stacks of contracting work out here.. IBM, VM,REXX,SQL IMS DB/DC COBOL ADBAS, NATURALII, CICS ORACLE 6/7 & CASE DEC VAX PATHWORKS TANDEM, BASE 24 TAL SAP R2/3 ABAP 4 PC EXPRESS VISUAL BASIC, C, C++ CITIDESK,SIR,HOLOS,HURON DIDOL,KAPITI, UNIFACE If any of these are your bag, and you want to have a go, mail me a resume and I'll get back to you. Most contracts are 3 or 6 months, and mostly renewable. The earnings are comfortable to good. Some of the above may end up longer term, so give us an idea of your preference as to time frames.. look forward to haering from you.. dave t, downunder I_Io "Here's to being awarded Object Honors '94!"
From: Ingo Feulner <ifeulner@cube.de> Newsgroups: comp.sys.next.programmer Subject: Re: EOF IB Date: 29 Jun 1994 08:47:40 GMT Organization: Cube Informationssysteme GmbH, Stuttgart, FRG Distribution: world Message-ID: <2urcfc$2hc@cubenx.cube.de> References: <1994Jun27.161458.26477@news.media.mit.edu> In article <1994Jun27.161458.26477@news.media.mit.edu> wave@media.mit.edu (Michael B. Johnson) writes: > In article <2umams$2c3@zebu.abstractsoft.com> > zeno@zebu.abstractsoft.com (Sean T. Lamont) writes: > The only thing I had to do to the current version of the WavesWorld palettes > was comment out the line in palette.table that had "Export Images". This > is a documented (in the paper release notes) bug. After that, > they both loaded fine, and everything seems fine, including all my > weird stuff I do in test interface mode with loading nibs, etc. I'm having problems with the new IB, too. Some of my NIBs created with the "old" IB couldn't be loaded by the new... it crashes with a memory exception. (But once I was able to load the NIB... but only once...) Has anybody mentioned similar problems? Thanks for any help, Ingo. > I love the new IB. Yes, it looks really nice and the EO and Foundation classes are great... but I wanna load my old Nibs!
From: shtmar04@sonnenberg.uct.ac.za (Mark Richard Shuttleworth) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.next.software Subject: MDI - Multiple Document Interface onNextStep? Date: Wed, 29 Jun 1994 14:46:10 GMT Organization: University of Cape Town Message-ID: <shtmar04.37.2E118932@sonnenberg.uct.ac.za> Hiya, I've never seen NEXTSTEP, so I may be way out of my depth here. Does it have any sort of standard, de facto or official, to handle MDI approaches to document work? Or has it already gone past that, to a more document centered OLE or OpenDoc approach? Does this not apply to NEXT? Did MDI really evolve from MS? Thanks for any info! Cheers, Mark
From: shtmar04@sonnenberg.uct.ac.za (Mark Richard Shuttleworth) Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.advocacy Subject: Distributed Object Frameworks: Next OpenStep vs SOM vs OLE/COM Date: Wed, 29 Jun 1994 14:50:30 GMT Organization: University of Cape Town Message-ID: <shtmar04.38.2E118A36@sonnenberg.uct.ac.za> Hiya, How does OpenStep's persistent distributedd object model compare with SOM, or COM as implemented by Microsoft and/or Digital. I understand that SOM is infinitely superior to COM, but how does Next's fit in to the industry? SOM, combined with Bento from Apple, will give birth to OpenDoc. Does Next have any big guns to shoot down that effort? What are the chances of getting SOM and OpenStep objects to talk to one another? Are they both CORBA based? It seems as if IBM has placed itself extremely well with SOM soon to be available on OS/2, Windows, AIX and even the mainfram OS's. Cheers, Mark
From: schalle@nextstn2.igd.fhg.de (Dirk Schalle (K.Vaananen)) Newsgroups: comp.sys.next.programmer Subject: NXTextCache functions Date: 29 Jun 1994 14:56:33 GMT Organization: IGD Darmstadt Message-ID: <2us231$lk9@korfu.igd.fhg.de> Keywords: Text NXTextCache Hi there, has anybody out there experiences or information about the three functions declared in /usr/include/appkit/Text.h: extern void NXSetTextCache(id self, NXTextCache *cache, int pos); extern int NXAdjustTextCache(id self, NXTextCache *cache, int pos); extern void NXFlushTextCache(id self, NXTextCache *cache); furthermore I'm trying hard to prevent Cells in a Text object from being deleted by backspace; is it possible to filter backspace from case to case (probably depending on the moon :-) thanx Dirk
Newsgroups: comp.sys.next.programmer From: dmurray@music.glas.ac.uk (Duncan Murray(EMU92)) Subject: NXHelpPanel. Message-ID: <Cs603p.pp@udcf.glasgow.ac.uk> Sender: news@udcf.glasgow.ac.uk (News) Organization: Glasgow University Computing Service Date: Wed, 29 Jun 1994 15:23:48 GMT Does anyone know how to avoid using the default NXHelpPanel when compiling an app? I want to have a 'coloured' panel with some custom graphics rather than the boring old black and white help panel. Can I just change a path in the Makefile.postamble/.preamble? I want to avoid changing the nib in lib/usr/NeXTStep/etc..... Cheers. Duncan Murray. dmurray@music.glas.ac.uk NeXTMail accepted.
From: reed@dedekind.math.luc.edu (dale f reed; DH307; x8-3567) Newsgroups: comp.sys.next.programmer Subject: Re: Arrays as method args: how to? Date: 29 Jun 1994 15:10:16 GMT Organization: Loyola University of Chicago Message-ID: <2us2so$r5c@apollo.it.luc.edu> References: <2uqbil$add@apollo.it.luc.edu> dale f reed; DH307; x8-3567 (reed@dedekind.math.luc.edu) wrote: : I want to pass a 2 dimensional array of int's as a method argument. It : works fine with a 1 dimensional array, but I'm having trouble declaring the : 2 dimensional array. : Here's what I have. The method is defined as: : - (void) getNumbersFrom:(FILE*)ifile andStoreIn: (int*) twoDimensionalArray; : It is called elsewhere using: : int twoDimensionalArray[ 2][ 10]; : [self getNumbersFrom: theFilePointer andStoreIn: twoDimensionalArray]; : Any ideas on how to declare this? When I compile, I get a message saying : "Incompatible pointer type". I appreciate any leads you might give me. : Dale Reed : reed@math.luc.edu Apparently this question belonged in misc.c.programmers.beginners rather than here. My appologies. The problem was that I was declaring (int) twoDimensionalArray [][ 10] rather than (int [][ 10]) twoDimensionalArray Thanks to all for their help. Dale reed@math.luc.edu
From: Ralph_Zazula@next.com (Ralph Zazula) Newsgroups: comp.sys.next.programmer Subject: Re: Categories Date: 29 Jun 1994 17:13:25 GMT Organization: NeXT, Inc. Distribution: world Message-ID: <2usa3l$1c6@rosie.next.com> References: <2upqe3$31p@hamblin.math.byu.edu> Sean O. Luke writes > Does anyone know of some evil hack which allows me to access the original > method of an object, even if that method's been overridden by a category > method of the same name? > The basic idea is to use the "methods" entry in the objc_class structure. This entry is a linked-list of methods known by the class. Methods added by categories should appear first (toward the head of the linked-list). If you traverse the entire list, you will find the replaced objc_method structure further down. From there, just get the IMP and call it as a function... Take a look at /usr/inlcude/objc/objc-class.h (or the run-time data types docs) for info on the objc_class and objc_method structures. Ralph -- Ralph Zazula Developer Trainer NeXT, Inc. Ralph_Zazula@next.com (415) 780-2893
Newsgroups: comp.sys.next.programmer From: karl@trapac.com (Karl Kraft) Subject: Porting Curses based programs Message-ID: <Cs65zB.uG@trapac.com> Sender: karl@trapac.com (Karl Kraft) Organization: Trans Pacific Container Service Corporation Date: Wed, 29 Jun 1994 17:30:47 GMT I've ported two curses based programs to NeXTSTEP, and both share the same problem, they insist on ^J to end a line instead of ^M. Not being a curses kinda guy, what do I need to do to get these programs to use ^M? Thanks. -- Karl Kraft Karl_Kraft@trapac.com Karl_Kraft@ensuing.com [My opinions are my own]
Newsgroups: comp.sys.next.programmer From: pelletk@il.us.swissbank.com (Ken Pelletier) Subject: Re: Can't change button's icon image. Weird behavior. Message-ID: <1994Jun29.184758.7664@il.us.swissbank.com> Sender: root@il.us.swissbank.com (Operator) Organization: Swiss Bank Corporation CM&T Division References: <2upg8s$q34@apollo.it.luc.edu> Date: Wed, 29 Jun 1994 18:47:58 GMT dale f reed; DH307; x8-3567 writes > Chris Wong (chris@iastate.edu) wrote: > : Very weird. I use [[myMatrix selectedCell] setImage:someImage] to set the > : icon image of one of the button at runtime. > > : However, the image doesn't appear when I launch the app. The weirdest thing > : is the image appear if I launch the app in gdb, i.e., in debug mode. > > : Here is the method that get executed when one of the button in the Matrix is > : click. [...] > > : Please share some insight here. I am totally clueless. > > : Thank you very very very much for any kind of information... > > : Chris -- desperate > > What seems to me might be the problem is that the icon is not found in the > directory being searched by the App. I have an App that works great from > within SuperDebugger, but has directory problems when launched from the > Workspace. (It does however work correctly when I specify the path > explicitly from the command line) > > The author of SuperDebugger, Bing-Qing Chen wrote to me: "While debuging, > the current directory will be the Project directory. If the app is launched > from Workspace, its current directory will be your home directory. You > should use chdir() to change to your Project directory if you want a > consistent behavior." > > This doesn't completely solve the problem, however, if you want to be able > to install your App ANYWHERE and have it run correctly. To do this you > must have code that uses a relative path name. I haven't been able to get > this to work (yet), but am interested in the solution. > > Dale Reed > reed@math.luc.edu Unless I'm missing something (not unusual), you can just put the image into your app wrapper by plopping it into the images suitcase in PB. Then it follows your app wherever it gets installed. To get the image from the wrapper, you can use: myImage = [NXImage findImageNamed:"myImageName"]; - Ken
From: cfleming@alleg.edu (Chuck Fleming) Newsgroups: comp.sys.next.programmer Subject: mmap help Date: 29 Jun 1994 19:35:08 GMT Organization: Allegheny College Distribution: world Message-ID: <2usidc$aqo@mustang.alleg.edu> Keywords: mmap, kernel, driver Some pre-3.0 documentation indicates that one can get a pointer to the EvGlobals structure by calling mmap (or _mmap). Can anyone pass along information on using mmap or alternative ways of manipulating EvGlobals from within the kernel? This is needed for work on black hardware. Thanks, Chuck -- ########################################### Chuck Fleming (cfleming@alleg.edu) Educational Computing Services Allegheny College NeXTMail welcome ###########################################
From: root@net23.com (Operator) Newsgroups: alt.2600,alt.tv.max-headroom,comp.sys.next.sysadmin,comp.sys.next.programmer,comp.sys.next.software,comp.sys.infosystems.www Subject: Max Web Server Message-ID: <Cs6DLD.MKx@spcuna.spc.edu> Date: 29 Jun 94 20:15:12 GMT Sender: news@spcuna.spc.edu (Network News) Organization: Network 23, Inc. Explore http://www.net23.com with your favorite client. Razor -- Nick Jarecki | Network 23 - InterNet Services Provider razor@net23.com | (shell,FTP,WWW) in the New York/Metro area Voice: [917-424-8806] | Email "info@net23.com" Ask me about our Hamburgers | Telnet to net23.com, login:info To access our REVOLUTIONARY WWW server, point your client to http://net23.com
Newsgroups: comp.sys.next.programmer From: jhall@barbados.umhc.umn.edu (Jeff Hallgren) Subject: Re: Need help compiling "straight" C. (Summary) Message-ID: <Cs6I0y.HrA@news.cis.umn.edu> Sender: news@news.cis.umn.edu (Usenet News Administration) Organization: University of Minnesota, Twin Cities References: <1994Jun29.075931.355@dart.de> Date: Wed, 29 Jun 1994 21:53:25 GMT Thanks to everyone that responded. These are "inline nested functions" and can be compiled using gcc 2.5 or later which explains why I can't compile them using NeXT's compiler. Apparently, nested functions are non-ansi and are only compilable using recent gcc compilers. void aLargeThing(myStruct *aStruct) { int i, j; if (something) { ---> int isrelop(char c){ return strchr("<>", c) != NULL; } int issign(char c){ return strchr("-+", c) != NULL; } char *item0 = item; (...) } }
Newsgroups: comp.sys.next.programmer,comp.lang.objective-c From: monty@hounddog.tbyte.com (Monty) Subject: Anyone using python on Nextstep? Message-ID: <1994Jun29.211217.28796@newsserver.rrzn.uni-hannover.de> Sender: news@newsserver.rrzn.uni-hannover.de (News Service) Organization: RRZN Date: Wed, 29 Jun 1994 21:12:17 GMT Looks like I will be. A Python <-> Objective-C binding would be most useful. Thanks, Monty monty@tbyte.com
Newsgroups: comp.sys.next.programmer From: andrew@stone.com (Andrew Stone) Subject: OpenStep issues Message-ID: <1994Jun28.211212.909@stone.com> Sender: andrew@stone.com Organization: Stone Design Corp Date: Tue, 28 Jun 1994 21:12:12 GMT As I pore over the new docs & specs of Enterprise Objects and Foundation classes, several questions come to mind: 1] what were the design motivations for altering the hierarchies? 2] How will Apps with docs that were archived in one hierarchy be resuscitated in the brave new world? 3] Is there sufficient interest in these topics to warrant a mailing list? andrew -- ||<<->>||<<==>>|S<<++>>|T<<?>O<+>>N|<<-->>E|<<==>>||<<+>>|| !! Andrew Stone | (505) 345-4800 !! !! andrew@stone.com | Have Modem Why Travel !! ||<<->>|D<<==>>|E<<++>>|S<<?>|<+>>G|<<-->>N|<<==>>!|<<+>>||
From: Charles William Swiger <infidel+@CMU.EDU> Newsgroups: comp.sys.next.programmer Subject: Re: Porting Curses based programs Date: Wed, 29 Jun 1994 19:21:59 -0400 Organization: Carnegie Mellon, Pittsburgh, PA Message-ID: <Ui4U8LS00iV706gaZb@andrew.cmu.edu> In-Reply-To: <Cs65zB.uG@trapac.com> Excerpts from netnews.comp.sys.next.programmer: 29-Jun-94 Porting Curses based programs by Karl Kraft@trapac.com > I've ported two curses based programs to NeXTSTEP, and both share the > same problem, they insist on ^J to end a line instead of ^M. > > Not being a curses kinda guy, what do I need to do to get these > programs to use ^M? Also watch out for Delete (DEL or '\177') versus BackSpace (^H or '\010'). Unless you have a pressing need to have every possible key combination available (as perhaps is needed for an editor such as Emacs), it's a very good idea to make your code work across such platform-specific assumptions by having the above key alternates (^M and ^J, DEL and ^H) perform the same functionality. -Chuck Charles William Swiger - WhiteLight Systems | "All the world's a stage, and" --------------------------------------------+ "we are merely players...." AMS & normal mail: infidel@cmu.edu | NeXTmail: chuck@cswiger.slip.andrew.cmu.edu | "Semper ubi sub ubi."
From: Greg.Titus@mccaw.com (Greg Titus) Newsgroups: comp.sys.next.programmer Subject: Re: SUMMARY: Can't change button's icon image. Weird behavior. Date: 30 Jun 1994 00:00:07 GMT Organization: McCaw Cellular Communications, Inc. Message-ID: <2ut1u7$edf@ftp-p.mccaw.com> References: <2ugk5g$5kc@news1.digex.net> In article <2ugk5g$5kc@news1.digex.net> sbender@harmony.digex.net (Scott Bender) writes: > In article <2ugd5n$r9b@news.iastate.edu> > > So the followup question is: How on the earth can I tell the NeXTSTEP to > > search the filename under my app wrapper? > > > > Use: > > [[NXBundle mainBundle] getPath:path forResource:"fileName" ofType:"tiff"] > > - Scott Unfortunately, if you do any rloading there is a bug that hoses the main bundle, and this won't work any more. Besides, what you really want is +findImageNamed:, which caches images by name. From the NXImage documentation: When you ask for an NXImage object by name (with the findImageNamed: method), the class object returns the one from its database, which also includes all the system bitmaps provided by the Application Kit. If there's no object in the database for the specified name, the class object tries to create one by looking (for historical reasons) in the _ _ICON, _ _EPS, and _ _TIFF segments of the application's executable file, and then in the application bundle. If a section or file matches the name, an NXImage is created from the data stored there. You can therefore create NXImage objects simply by including EPS or TIFF data for them within the executable file, or in files inside the application's file package. -Greg ------------- Greg Titus Omni Development Inc. toon@omnigroup.com
Newsgroups: comp.sys.next.programmer From: cedman@princeton.edu (Carl Edman) Subject: Re: Need help compiling "straight" C. (Summary) In-Reply-To: jhall@barbados.umhc.umn.edu's message of Wed, 29 Jun 1994 21:53:25 GMT To: jhall@barbados.umhc.umn.edu (Jeff Hallgren) Message-ID: <CEDMAN.94Jun29205621@capitalist.princeton.edu> Originator: news@nimaster Sender: news@Princeton.EDU (USENET News System) Organization: Princeton University References: <1994Jun29.075931.355@dart.de> <Cs6I0y.HrA@news.cis.umn.edu> Date: Thu, 30 Jun 1994 00:56:21 GMT In article <Cs6I0y.HrA@news.cis.umn.edu> jhall@barbados.umhc.umn.edu (Jeff Hallgren) writes: These are "inline nested functions" and can be compiled using gcc 2.5 or later which explains why I can't compile them using NeXT's compiler. Apparently, nested functions are non-ansi and are only compilable using recent gcc compilers. That is correct. void aLargeThing(myStruct *aStruct) { int i, j; if (something) { ---> int isrelop(char c){ return strchr("<>", c) != NULL; } int issign(char c){ return strchr("-+", c) != NULL; } char *item0 = item; (...) } } I wonder how these local functions work anyway. What if isrelop made a reference to 'i' and aLargeThing returned a pointer to isrelop ? If the surrounding program ever made use of that pointer to call isrelop, to what 'i' does isrelop refer ? Adding closures to C seems like a large hassle for a minor convenience, but I don't see how you could consistently have static scope and local functions otherwise. Carl Edman
Newsgroups: comp.sys.next.programmer From: bruce@trwlasd.com (Bruce McKenzie) Subject: Re: inspector for sereral file-types? Message-ID: <Cs70D7.5qs@trwlasd.com> Sender: bruce@trwlasd.com (Bruce McKenzie) Organization: TRW References: <940622222502.432AACUI.malc@jeeves> Date: Thu, 30 Jun 1994 04:27:05 GMT In article <940622222502.432AACUI.malc@jeeves> "mmalcolm Crawford" <malc@dcs.shef.ac.uk> writes: > Is there a way to specify that a Workspace Inspector will inspect several > file-types...? > Could you please tell me how...! put multiple lines in the bundle.registry file: {type=InspectorCommand; mode=contents; extension=app; selp=selectionOneOnly; class=AppInfoInspector} {type=InspectorCommand; mode=contents; extension=debug; selp=selectionOneOnly; class=AppInfoInspector} {type=InspectorCommand; mode=contents; extension=palette; selp=selectionOneOnly; class=AppInfoInspector} {type=InspectorCommand; mode=contents; extension=bundle; selp=selectionOneOnly; class=AppInfoInspector} Good luck... -- Bruce McKenzie (spuds@netcom.com, NeXTMail welcome) Atlas Software Ventures, Inc. PO Box 1299, Santa Clara, CA 95052-1299 800/278-9909 (ASV-9909)
Newsgroups: comp.sys.next.programmer From: bruce@trwlasd.com (Bruce McKenzie) Subject: Re: Scrolling of a View by dragging the mouse Message-ID: <Cs7112.68y@trwlasd.com> Sender: bruce@trwlasd.com (Bruce McKenzie) Organization: TRW References: <2uptjd$lg5@aragorn.unibe.ch> Date: Thu, 30 Jun 1994 04:41:25 GMT In article <2uptjd$lg5@aragorn.unibe.ch> hauert@itp.unibe.ch (Christoph Hauert) writes: > one problem we are dealing with at the moment is scrolling or shifting a > view controlled by the mouse. actually we are looking for hints how > scrolling is done in apps like Grab.app with the "hand cursor" active. > Another example would be PasteUp. Check out TTools/ScrollDoodScroll MiniExamples. These keep getting events until a MouseUp is seen. For each event you can move a speed/amount (based on distance outside window rectangle). Basically, you've got a tight little event loop going on in this special circumstance, in mouseDown. -- Bruce McKenzie (spuds@netcom.com, NeXTMail welcome) Atlas Software Ventures, Inc. PO Box 1299, Santa Clara, CA 95052-1299 800/278-9909 (ASV-9909)
Newsgroups: comp.sys.next.programmer From: bruce@trwlasd.com (Bruce McKenzie) Subject: Re: SUMMARY: Can't change button's icon image. Weird behavior. Message-ID: <Cs71x0.6qE@trwlasd.com> Sender: bruce@trwlasd.com (Bruce McKenzie) Organization: TRW References: <2ugk5g$5kc@news1.digex.net> Date: Thu, 30 Jun 1994 05:00:33 GMT In article <2ugk5g$5kc@news1.digex.net> sbender@harmony.digex.net (Scott Bender) writes: > In article <2ugd5n$r9b@news.iastate.edu> > > So the followup question is: How on the earth can I tell the NeXTSTEP to > > search the filename under my app wrapper? > > Use: > [[NXBundle mainBundle] getPath:path forResource:"fileName" ofType:"tiff"] (I know that this will be changed in 4.0 bitch, bitch, bitch...) I recommend "[NXBundle bundleForClass:[self class]]" instead of [NXBundle mainBundle]. It just that little bit more portable.... -- Bruce McKenzie (spuds@netcom.com, NeXTMail welcome) Atlas Software Ventures, Inc. PO Box 1299, Santa Clara, CA 95052-1299 800/278-9909 (ASV-9909) *** Software Development, specializing in NEXTSTEP *** --- -- Bruce McKenzie (spuds@netcom.com, NeXTMail welcome) Atlas Software Ventures, Inc. PO Box 1299, Santa Clara, CA 95052-1299 800/278-9909 (ASV-9909)
Newsgroups: comp.sys.next.programmer From: bruce@trwlasd.com (Bruce McKenzie) Subject: Re: What new features will be in 3.3? Message-ID: <Cs7235.6uD@trwlasd.com> Sender: bruce@trwlasd.com (Bruce McKenzie) Organization: TRW References: <1994Jun23.000853.7943@millennium.com> Date: Thu, 30 Jun 1994 05:04:16 GMT In article <1994Jun23.000853.7943@millennium.com> Jayson Adams writes: > In article <ROBERT.94Jun19074704@steffi.demon.co.uk>, Robert Nicholson writes: > > <rgc@cs.umd.edu> writes: > > > > >What new utilities will be in NS 3.3? Specifically, will 3.3 provide > > >version controll? An improved Edit (maybe will color syntax > > highlighting).... > > > > If you want _automatic_ color syntax highlighting then go no further > > than Emacs for NS.. with font-lock.e. or hilit19.e. > > That sounds great. How do I get it within Edit? (I'm not sure if you're being facetious...) Do you know about the RTFSyntax app? on cs.orst.edu... Now the question that I didn't get a chance to ask at Expo... Does the 3.3/EOF IB support RTF source? I haven't loaded it up yet... Anyone know? -- Bruce McKenzie (spuds@netcom.com, NeXTMail welcome) Atlas Software Ventures, Inc. PO Box 1299, Santa Clara, CA 95052-1299 800/278-9909 (ASV-9909) *** Software Development, specializing in NEXTSTEP *** --- -- Bruce McKenzie (spuds@netcom.com, NeXTMail welcome) Atlas Software Ventures, Inc. PO Box 1299, Santa Clara, CA 95052-1299 800/278-9909 (ASV-9909)
From: "Scott A. McIntyre" <S.A.McIntyre@durham.ac.uk> Newsgroups: comp.sys.next.programmer Subject: udb, munmap() and sendmail Date: 30 Jun 1994 09:26:53 GMT Organization: University of Durham, Durham, UK. Message-ID: <2uu34u$ajg@mercury.dur.ac.uk> I'm tring to get berkeley's udb (db 1.79) to work with the latest version of sendmail (8.6.9) but found that udb needs snprintf() and munmap() neither of which the NeXT seems to have; is there a suitable replacement that someone can point me in the general direction of? thanks -- GAT d-- -p+ c++++ !l u++ e+++@ m++(*) s+/+@ n+ h-- f+@(?) g+ w+++ t++ r- y**(+) EMAIL: scott@shrug.org (NeXTmail accepted) OR S.A.McIntyre@durham.ac.uk WWW: http://shrug.org/~scott/ On the Internet, no one knows if you have a crap haircut.
Newsgroups: comp.sys.next.programmer From: brouwer@mickey.informatik.uni-stuttgart.de (Klaus Brouwer) Subject: Foundation Classes Message-ID: <Cs7IKw.M66@news.informatik.uni-stuttgart.de> Sender: news@informatik.uni-stuttgart.de Organization: Informatik, Uni Stuttgart, Germany Date: Thu, 30 Jun 1994 11:00:31 GMT Hello! How could a non Expo visitior and no database programmer get the new Foundation Classes? Is at least the docu available via ftp? Someone said that there will be no 3.3 Developer and I don't want to wait until 4.0. Klaus Brouwer
From: Charles William Swiger <infidel+@CMU.EDU> Newsgroups: comp.sys.next.programmer Subject: Re: Need help compiling "straight" C. (Summary) Date: Thu, 30 Jun 1994 09:42:14 -0400 Organization: Carnegie Mellon, Pittsburgh, PA Message-ID: <ki4giqa00iV481BrF8@andrew.cmu.edu> In-Reply-To: <CEDMAN.94Jun29205621@capitalist.princeton.edu> Excerpts from netnews.comp.sys.next.programmer: 30-Jun-94 Re: Need help compiling "st.. by Carl Edman@princeton.edu RE: static inline functions > I wonder how these local functions work anyway. What if isrelop made > a reference to 'i' and aLargeThing returned a pointer to isrelop ? If > the surrounding program ever made use of that pointer to call isrelop, > to what 'i' does isrelop refer ? Adding closures to C seems like a > large hassle for a minor convenience, but I don't see how you could > consistently have static scope and local functions otherwise. The reason for adding this syntax is probably based on compiler design issues, with a nod towards encapsulation. The apparent purpose is to have these functions inlined for maximum optimization purposes without (a) static inline functions (which run into namespace collisions, particularly if you want to have the same mini-function which behaves differently in different "main" functions), and (b) macros (although I don't see many problems with them, but the designers of gcc seem to have a mild dislike of them). The compiler can certainly optimize these locally faster then it could if it had to mess with the global tables for a real static function. I am quite surprised to see the gcc team produce such a kludge considering their stance on ANSI compatibility with respect to trigraphs. I may well be doing the designers an injustice, but I simply don't see what they are trying to provide that can't be done otherwise without changing the C language. ----------- In case it isn't obvious, this rather dubious syntax is not going to be portable outside of modern versions of gcc, and is also using code that has not been tested nearly as thoroughly as the rest of gcc. I'd suggest avoiding using this syntax until and unless you understand where gcc is going with this idea. -Chuck Charles William Swiger - WhiteLight Systems | "All the world's a stage, and" --------------------------------------------+ "we are merely players...." AMS & normal mail: infidel@cmu.edu | NeXTmail: chuck@cswiger.slip.andrew.cmu.edu | "Semper ubi sub ubi."
Newsgroups: comp.sys.next.programmer From: rpomeroy@aunext1.amd.com (Ron Pomeroy x(Coop)) Subject: EOF with Sybase Limited Server Message-ID: <Cs7rop.FCE@txnews.amd.com> Keywords: eof sybase limited server Sender: news@txnews.amd.com Organization: Advanced Micro Devices, Austin TX Distribution: usa Date: Thu, 30 Jun 1994 14:17:11 GMT Hey there NextHeads, I've got EOF and the Sybase Limited Server (from the 2.x days) and would like to start making some of those Mission Critical Objects everyone has been going on about. A few questions persist (if only my object could do the same :-) ). 1) In the 2.x days there was a client libs package for Sybase in /NextLibrary/Packages/3rd_Party/Sybase. Do I still need this package for things to work ? 2) I heard that there are problems running the graphical Sybase Administrator application under 3.x. Does anyone know exactly what the problems are and the workarounds ? In addition, any general comments from those who have already traveled this path are *extremely* welcome. This is all on black hardware - one machine - no network etc. Thanks in Advance -- Ronald Pomeroy Advanced Micro Devices CIM Applications Group rpomeroy@aunext1.amd.com
Newsgroups: comp.sys.next.programmer From: rpomeroy@aunext1.amd.com (Ron Pomeroy x(Coop)) Subject: Re: OpenStep issues Message-ID: <Cs7ru8.FE2@txnews.amd.com> Sender: news@txnews.amd.com Organization: Advanced Micro Devices, Austin TX References: <1994Jun28.211212.909@stone.com> Distribution: usa Date: Thu, 30 Jun 1994 14:20:30 GMT In article <1994Jun28.211212.909@stone.com> andrew@stone.com (Andrew Stone) writes: >>As I pore over the new docs & specs of Enterprise Objects and Foundation >>classes, several questions come to mind: >> >>1] what were the design motivations for altering the hierarchies? >>2] How will Apps with docs that were archived in one hierarchy be >> resuscitated in the brave new world? >>3] Is there sufficient interest in these topics to warrant a mailing list? >> I'm not sure I understand 2) (killed quite a few brain cells at that damn rave!). Could you elaborate ? -- Ronald Pomeroy Advanced Micro Devices CIM Applications Group rpomeroy@aunext1.amd.com
Newsgroups: comp.sys.next.programmer From: Atze (Alexander Spohr) Subject: Duplicate key is not an error? Message-ID: <1994Jun30.142304.5171@dart.de> Keywords: SYBASE DBModule saveChanges Sender: aspohr@dart.de Organization: d'ART Software GmbH Date: Thu, 30 Jun 94 14:23:04 GMT Hi out there! If I save my DBModule containing a duplicated record, SYBASE gives me "Duplicate key was ignored. (0 rows affected)". I get a panel telling me the same stuff. This is exactly what I want: No doublettes of records in my db. The problem I have is that the DBModule does not tell me that the save failed. Question: How do I get the information about the duplicate key into my program? I can bring up panels myself. But I have to know WHEN. :-{ All ideas are wellcome! Atze -- Alexander Spohr, Hamburg, Germany Alexander_Spohr@DART.de (Mail ) Faces and faces (+49) 40 / 380 23-0 (Sound) See them and complain not (+49) 40 / 380 23-290 (Fax ) And am content with all
From: dekorte@ibm19.scri.fsu.edu (Stephen L. DeKorte) Newsgroups: comp.sys.next.programmer Subject: DPS problem Date: 30 Jun 1994 08:11:18 GMT Organization: Supercomputer Computations Research Institute Message-ID: <2utun6$n4l@news.scri.fsu.edu> I'm trying to open a DPS window without appkit and draw some animations on it. In order to avoid seeing the new frames be drawn, I thought I would use an NX_BUFFERED window, but I can't figure out how to swap the buffers - I see no DPSswapbuffer or the like in the dpsclient includes, and the other buffer related DPS&PS calls don't seem to work. Does anyone know what I should do? NeXT has no real docs on the DPS stuff. And I couldn't find anything usefull from ftp.adobe.com. Any help would be greatly appreciated, Steve Dekorte
Newsgroups: comp.sys.next.programmer From: glen@instep.wimsey.bc.ca Subject: Re: Setting permissions with PB.... Message-ID: <1994Jun30.155157.10355@instep.wimsey.bc.ca> Sender: usenet@instep.wimsey.bc.ca Organization: InStep Mobile Communications Inc. References: <2uq31n$o8q@master.cs.rose-hulman.edu> Date: Thu, 30 Jun 1994 15:51:57 GMT In article <2uq31n$o8q@master.cs.rose-hulman.edu> borzilda@NeXTwork.Rose-Hulman.Edu (David A Borzillo) writes: > Is there any arguments or anything I can add to Project Builder so that > when I build my app everyone will have permission to run it? Also, > whereever I move it it will have the same permissions? We're on the AFS > network here and changing the attributes in the Workspace inspector > doesn't help. I have to use a series of 'chmod' commands to change the > permission. > Permissions after make install will normally be affected by the making user's umask in the normal Unix way. We fix it at the end of the makefile postamble; eg. after_install:: -$(CHMOD) -Rf a+r $(DSTROOT)$(INSTALLDIR)/$(NAME).app Replace a+r with whatever you need for your developers. Moving an app shouldn't change the permissions (or ownership). Copying it will affect the permissions with the umask in the normal Unix way. > -- > David A. Borzillo (borzilda@nextwork.rose-hulman.edu) > Software Developer, Fluid Science Learning Center > Mech. Engineering Senior, Rose-Hulman Inst of Tech, Terre Haute,IN > +=================================================================+ > Disclaimer: Any views and opinions expressed in this > message are my own and not of Rose-Hulman or the FSLC Project.
Newsgroups: comp.sys.next.programmer,comp.lang.objective-c From: gene@netcom.com (gene m. stover) Subject: Re: Anyone using python on Nextstep? Message-ID: <geneCs80Br.F0C@netcom.com> Followup-To: comp.sys.next.programmer,comp.lang.objective-c Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <1994Jun29.211217.28796@newsserver.rrzn.uni-hannover.de> Date: Thu, 30 Jun 1994 17:23:45 GMT I'll bite: What is Python? Monty (monty@hounddog.tbyte.com) wrote: : Looks like I will be. A Python <-> Objective-C binding would be most : useful. : Thanks, : Monty : monty@tbyte.com -- ------------------------------------------------------------ gene m. stover "Making the world safe from democracy." Internet: gene@CyberTiggyr.com NeXT Mail spoken here For PGP public key, finger gene@netcom.com
From: sweeney@seahawk.Eng.Sun.COM (William Sweeney) Newsgroups: comp.sys.next.programmer Subject: Re: DPS problem Date: 30 Jun 1994 18:06:09 GMT Organization: Sun Microsystems, Inc. Distribution: world Message-ID: <2uv1ih$81a@engnews1.Eng.Sun.COM> References: <2utun6$n4l@news.scri.fsu.edu> > I thought I would use an NX_BUFFERED window, but I > can't figure out how to swap the buffers - PSflushgraphics(void) > NeXT has no real docs on the DPS stuff General Reference Volume 2 Chapter 5.
Newsgroups: comp.sys.next.programmer From: pasqua@mv.us.adobe.com Subject: Re: DPS problem Message-ID: <1994Jun30.180504.7477@adobe.com> Sender: usenet@adobe.com (USENET NEWS) Organization: Adobe Systems Incorporated References: <2utun6$n4l@news.scri.fsu.edu> Date: Thu, 30 Jun 1994 18:05:04 GMT In article <2utun6$n4l@news.scri.fsu.edu> dekorte@ibm19.scri.fsu.edu (Stephen L. DeKorte) writes: >I'm trying to open a DPS window without appkit and draw some >animations on it. In order to avoid seeing the new frames be >drawn, I thought I would use an NX_BUFFERED window, but I >can't figure out how to swap the buffers - I see no DPSswapbuffer >or the like in the dpsclient includes, and the other buffer >related DPS&PS calls don't seem to work. Does anyone know >what I should do? NeXT has no real docs on the DPS stuff. >And I couldn't find anything usefull from ftp.adobe.com. > >Any help would be greatly appreciated, >Steve Dekorte In this context buffered does not mean double buffered. A buffered window in NEXTSTEP is one that has an off-screen window that receives all output from DPS. Under programmer control, this output is flushed to the corresponding window on the screen. When you begin to draw again, you draw into the same buffer as before and it contains the same contents as before (unless you explicitly clear it). You can achieve what you want by allocating your own buffer(s) and compositing. There are a number of programs on the archives that do this. You can also take a look at BackSpace (in /NextDeveloper/Examples/AppKit/Backspace) to see an example of animation in the NS environment. Unless you have some overriding reason to do so, you probably should use the AppKit rather than going directly to low level window operators. Joe Pasqua Adobe Systems Incorporated
Newsgroups: comp.sys.next.programmer From: Robert La Ferla <Robert_La_Ferla@hot.com> Subject: Re: EOF IB Message-ID: <1994Jun30.185237.382@hot.com> Sender: robertl@hot.com (Robert La Ferla) Organization: Hot Technologies References: <Cs2GLM.96H@mrk.com> Distribution: usa Date: Thu, 30 Jun 1994 18:52:37 GMT It was given to everyone who attended the Developer's conference. Robert La Ferla Hot Technologies Registered NEXTSTEP Developer and Consultant In article <Cs2GLM.96H@mrk.com> jerald@mrk.com (Jerald Dawson) writes: > In article <1994Jun27.161458.26477@news.media.mit.edu> wave@media.mit.edu > (Michael B. Johnson) writes: > > In article <2umams$2c3@zebu.abstractsoft.com> zeno@zebu.abstractsoft.com > (Sean T. Lamont) writes: > > >> > > >>Has anyone else found that using the EOF version of IB causes non-NeXT > > >>palettes to load flakily, if at all? I've had extremely limited > > >>success with all of my own palettes, plus all of the MiscKit ones. > > >>They occasionally work with the right circumstances, but usually they > > >>just say 'unable to load' > > >> > > > > The only thing I had to do to the current version of the WavesWorld palettes > > was comment out the line in palette.table that had "Export Images". This > > is a documented (in the paper release notes) bug. After that, they both > loaded > > fine, and everything seems fine, including all my weird stuff I do in test > > interface mode with loading nibs, etc. > > > > I love the new IB. > > > > > > -- > > --> Michael B. Johnson -- wave@media.mit.edu > > --> MIT Media Lab -- Computer Graphics & Animation Group > > --> 20 Ames St. E15-023G -- (617) 547-0563 (day office) > > --> Cambridge, MA 02139 -- (617) 253-0663 (night office) > > How does one go about getting the prerelease of the new IB and EOF stuff. Was > it just something that was available to expo attendees or can anyone get it? > > -- > jerald dawson * Systems Manager/Head Programmer * Investor's Advantage > NeXTmail - jerald@mrk.com | > phone - (708) 487-3221 | Don't blame me. I voted for Bush.
Newsgroups: comp.sys.next.programmer From: Jayson Adams Subject: Re: What new features will be in 3.3? Message-ID: <1994Jun30.200621.2461@millennium.com> Keywords: If I wanted to use Emacs, I would use Emacs Sender: jayson@millennium.com Organization: Millennium Software Labs, Inc. References: <1994Jun23.000853.7943@millennium.com> <Cs7235.6uD@trwlasd.com> Date: Thu, 30 Jun 1994 20:06:21 GMT In article <Cs7235.6uD@trwlasd.com>, Bruce McKenzie writes: > In article <1994Jun23.000853.7943@millennium.com> Jayson Adams writes: > > In article <ROBERT.94Jun19074704@steffi.demon.co.uk>, Robert Nicholson > writes: > > > <rgc@cs.umd.edu> writes: > > > > > > >What new utilities will be in NS 3.3? Specifically, will 3.3 provide > > > >version controll? An improved Edit (maybe will color syntax > > > highlighting).... > > > > > > If you want _automatic_ color syntax highlighting then go no further > > > than Emacs for NS.. with font-lock.e. or hilit19.e. > > > > That sounds great. How do I get it within Edit? > > (I'm not sure if you're being facetious...) That's almost always a safe bet. __jayson
Newsgroups: comp.sys.next.programmer From: brianw@sounds.wa.com (Brian Willoughby) Subject: Re: Setting permissions with PB.... Message-ID: <Cs8Bo7.J3H@sounds.wa.com> Organization: Sound Consulting, Bellevue, WA, USA References: <2uq31n$o8q@master.cs.rose-hulman.edu> <1994Jun30.155157.10355@instep.wimsey.bc.ca> Date: Thu, 30 Jun 1994 21:28:54 GMT <glen@instep.wimsey.bc.ca> wrote: >borzilda@NeXTwork.Rose-Hulman.Edu (David A Borzillo) writes: >> Is there any arguments or anything I can add to Project Builder so that >> when I build my app everyone will have permission to run it? Also, >> whereever I move it it will have the same permissions? We're on the AFS >> network here and changing the attributes in the Workspace inspector >> doesn't help. I have to use a series of 'chmod' commands to change the >> permission. > >Permissions after make install will normally be affected by the making >user's umask in the normal Unix way. We fix it at the end of the makefile >postamble; eg. > >after_install:: > -$(CHMOD) -Rf a+r $(DSTROOT)$(INSTALLDIR)/$(NAME).app > >Replace a+r with whatever you need for your developers. > >Moving an app shouldn't change the permissions (or ownership). Copying it >will affect the permissions with the umask in the normal Unix way. You may want to use the support which is already designed into the NEXTSTEP Makefiles for this. Here is an excerpt from my Makefile.postamble to illustrate: # # Makefile.postamble # # Ownership and permissions of files installed by 'install' target #INSTALL_AS_USER = root # User to chown app to INSTALL_AS_GROUP = staff # Group to chgrp app to INSTALL_PERMISSIONS = ug+s # If set, 'install' chmod's executable to this e.g. in Glen's example above, her would only need to add the following line to Makefile.postamble: INSTALL_PERMISSIONS = a+r -- Brian Willoughby Software Design Engineer, BSEE from NCSU NeXTmail welcome Sound Consulting: Software Design and Development BrianW@SoundS.WA.com Bellevue, WA
From: magus@shell.portal.com (Magus Company) Newsgroups: comp.lang.postscript,comp.os.os2.programmer.misc,comp.os.ms-windows.programmer.misc,comp.sys.mac.programmer,comp.sys.next.programmer,comp.text.tex,comp.text.sgml Subject: A question about the uses of PostScript Date: 30 Jun 1994 22:30:56 GMT Organization: Portal Communications Company -- 408/973-9111 (voice) 408/973-8091 (data) Message-ID: <2uvh30$1ua@news1.svc.portal.com> My company (Magus) sells a commercial PostScript file viewer*. We are interested in the different ways people use PostScript files, and would like to know more so as to focus our efforts along useful lines. I am hoping that some Usenet experts can help me find answers to the following questions: 1) What are the largest repositories of PostScript documents on the Internet, and for what purposes are they used? 2) What organizations currently distribute documents in PostScript form to customers, clients, or collaborators; how do they do it; and for what purposes? (I know that PostScript documents appear on CD ROMs from numerous companies, but am trying to get more specific information.) Contact information is desirable here; i.e. how can I acquire some of these documents? Please reply by email. Thanks in advance for your help. Kevin W. Thompson 415-940-1109 President 800-848-8037 thompson@magus.com 415-940-1238 Fax Magus * P.O. Box 390965 * Mountain View CA 94039-0965 * USA * Currently for OS/2; MS Windows and Macintosh versions will follow.
From: gad@eclipse.its.rpi.edu (Garance A. Drosehn) Newsgroups: comp.sys.next.programmer Subject: Re: Uncompressing Help.store? Date: 30 Jun 1994 22:28:54 GMT Organization: Rensselaer Polytechnic Institute, Troy NY, USA Distribution: world Message-ID: <2uvgv7$7g8@usenet.rpi.edu> References: <gelatoCs4vKH.CKu@netcom.com> gelato@netcom.com (Steve James) writes: > Does anybody have a tool to find out what's in a Help.store > compressed help directory? I couldn't find anything in Librarian. > Is compresshelp even documented anywhere? If it is, I couldn't > find it. On ftp.cs.orst.edu, check directory pub/next/binaries/util for the files uncompresshelp.README and uncompresshelp.m.Z I haven't used this utility, but I believe it does what you want. -- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu ITS Systems Programmer (handles NeXT-type mail) Rensselaer Polytechnic Institute; Troy NY USA
From: rp9@unix.york.ac.uk (R Peppe) Newsgroups: comp.sys.next.programmer Subject: using the right mouse button in a Text object Date: 30 Jun 1994 17:13:32 GMT Organization: University of York, Computing Service, UK Distribution: world Message-ID: <2uuufs$mac@castle.york.ac.uk> Keywords: text subclassing mouse I have an application for which I need to make the right mouse button have a function different from that normally given it under NeXTStep. More particularly, I would like to subclass the Text object so that I can trap usage of the right mouse button. (this includes 'chording' - i.e. touching the right mouse button while the left button is already depressed.) So far, it seems very possible, as the right-button mouse event passes through a Text object just like all other events. The problem is that all the functionality to do with the left mouse button is in the mouseDown method. If the right button gets pressed when the left button is down, the Text object is in a modal loop inside mouseDown, so the 'rightMouseDown' method is never called. So it seems that the mouseDown method has to be replaced by a subclass implementation. Which leads me to a real question : How can I tell what character in the Text object has been clicked on from a coordinate in its view ? e.g. - mouseDown:(NXEvent *)e { int chrpos; [self convertPoint:&e->location fromView:nil]; /* how can I do this ?! */ chrpos = [self getChrPosForPoint:&e->location]; ... } If anyone has any suggestions, or any better ideas on how to go about accomplishing this task, I'd be most grateful to them. (A definite statement like "the only way to do this is to delve into those horribly complicated Text instance variables" would also be most useful - that's my only other option at the moment, and I don't like the prospect!) Please email replies, as our news feed is extremely shaky at the moment. cheers, rog. p.s. is it a known bug that the mouseMoved method in a View is never called despite setting its window's event mask appropriately. (code on request) p.p.s. all comments apply to nextstep 3.0 and 3.2 (black hardware and intel) roger peppe +44 904 432379 rog@ohm.york.ac.uk adaptive systems engineering university of york england
Newsgroups: comp.sys.next.programmer,comp.sys.next.software From: sdavis@laforge.ksc.nasa.gov (Steve Davis) Subject: Two-column Landscape Printing Utility? Message-ID: <1994Jun30.191658.28908@dale.ksc.nasa.gov> Sender: news@dale.ksc.nasa.gov Organization: NASA Date: Thu, 30 Jun 1994 19:16:58 GMT Greetings! I'm looking for a utility that will allow me to print long source code files in two-column landscape format-- with a reduced font, obviously. Does anyone know of such a utility for the NeXT? (I'm using a NeXTstation Turbo Color-- NeXTmail OK.) Thanks, Steve
Newsgroups: comp.sys.next.programmer From: macrae@pandora.geo.ucalgary.ca (Andrew MacRae) Subject: Re: EOF with Sybase Limited Server Message-ID: <Jul1.021219.51105@acs.ucalgary.ca> Date: Fri, 1 Jul 1994 02:12:19 GMT References: <Cs863H.J3s@txnews.amd.com> Organization: The University of Calgary, Alberta, Canada In article <Cs863H.J3s@txnews.amd.com> rpomeroy@aunext1.amd.com (Ron Pomeroy x(Coop)) writes: > In article <Cs7rop.FCE@txnews.amd.com> rpomeroy@aunext1.amd.com (Ron Pomeroy > x(Coop)) writes: > >>Hey there NextHeads, > >> > >>I've got EOF and the Sybase Limited Server (from the 2.x days) and > >> .. > >>2) I heard that there are problems running the graphical Sybase > >>Administrator application under 3.x. Does anyone know exactly what the > >>problems are and the workarounds ? It brings up two terminal windows during the installation stage, and runs two installation scripts simultaneously - a bad thing. However, if you close one of the windows as soon as it opens, the other will work ok. Even better, get the updated version of Sybase Administrator. It is available from NeXT. > > With respect to the second question: No conclusion, but I found out how > to perform the pertinent steps via the command line. > -Andrew macrae@pandora.geo.ucalgary.ca or: macrae@geo.ucalgary.ca
Newsgroups: comp.sys.next.programmer,mot.comp.unix.admin,comp.unix.programmer From: mschwage@next3.corp.mot.com (Mike Schwager) Subject: Help! Generic Unix programming problem... Organization: MOTOROLA Date: 30 Jun 94 21:07:59 GMT Message-ID: <mschwage.773010479@cssrjs> Sender: news@schbbs.mot.com (Net News) Hi, This isn't so NeXTStep-specific, but our NeXT's are the only ones with the problem (ie, they use BSD-style sockets only, no SysV named pipes): I have an application that tails a file, much like tail -f, but if the file it's looking at disappears out from under it, no big whoop. It simply tries to find it again. It's called taild. Now I need to take the output from taild and send it to an arbitrary number of other processes. These programs need to run on AIX, HP-UX, SunOS, and NeXTStep. On most Unices, no problem. I just: make_named_pipe npipe taild file > npipe & process1 < npipe & process2 < npipe & process3 < npipe & But Mach doesn't have named pipes! How do I duplicate this functionality under Mach? Can I do it with sockets? Or, what about Dup'ing file descriptors in the shell? (I'm still fuzzy wrt understanding dup(2)). Thanks for the help. Just another, -Motorola Mike ================================================ Mike Schwager schwager@mot.com (soon to be: schwager@wwa.com ) Working for, not speaking for Motorola, Inc., in Beautiful Schaumburg, IL "Our doubts are traitors," said Lucio, "and make us lose the good we might often win, by fearing to attempt it." -Wm. Shakespeare
From: zeno@zebu.abstractsoft.com (Sean T. Lamont) Newsgroups: comp.sys.next.programmer,mot.comp.unix.admin,comp.unix.programmer Subject: Re: Help! Generic Unix programming problem... Date: 30 Jun 1994 21:32:51 -0700 Organization: Abstract Software Message-ID: <2v069j$4sb@zebu.abstractsoft.com> References: <mschwage.773010479@cssrjs> >But Mach doesn't have named pipes! How do I duplicate this functionality >under Mach? Can I do it with sockets? Or, what about Dup'ing file >descriptors in the shell? (I'm still fuzzy wrt understanding dup(2)). BZZT. Wrong. look at the call to mkfifo(), or get gnu mknod, either method will create a named pipe. -- Sean T. Lamont | Ask me about the WSI-Fonts Abstract Software | Professional collection for NeXT lamont@abstractsoft.com |____________________________________
Newsgroups: comp.sys.next.programmer From: rpomeroy@aunext1.amd.com (Ron Pomeroy x(Coop)) Subject: Re: EOF with Sybase Limited Server Message-ID: <Cs863H.J3s@txnews.amd.com> Sender: news@txnews.amd.com Organization: Advanced Micro Devices, Austin TX References: <Cs7rop.FCE@txnews.amd.com> Distribution: usa Date: Thu, 30 Jun 1994 19:28:28 GMT In article <Cs7rop.FCE@txnews.amd.com> rpomeroy@aunext1.amd.com (Ron Pomeroy x(Coop)) writes: >>Hey there NextHeads, >> >>I've got EOF and the Sybase Limited Server (from the 2.x days) and would like >>to start making some of those Mission Critical Objects everyone has been going >>on about. >> >>A few questions persist (if only my object could do the same :-) ). >> >>1) In the 2.x days there was a client libs package for Sybase in >>/NextLibrary/Packages/3rd_Party/Sybase. Do I still need this package for things >>to work ? >> >>2) I heard that there are problems running the graphical Sybase Administrator >>application under 3.x. Does anyone know exactly what the problems are and the >>workarounds ? >> >>In addition, any general comments from those who have already traveled this >>path are *extremely* welcome. >> >>This is all on black hardware - one machine - no network etc. >> >>Thanks in Advance >> They say talking to yourself is a sign of intelligence....or schizophrenia. Anyway, I received some excellent help on this from Tyler Gingrich (hope I spelled your name correct). Let's see if I can summarize... With respect to the first question: That package has been folded into NEXTSTEP3.2. /usr/sybase contains all the necessary goodies for a client to talk to a sybase database. With respect to the second question: No conclusion, but I found out how to perform the pertinent steps via the command line. -- Ronald Pomeroy Advanced Micro Devices CIM Applications Group rpomeroy@aunext1.amd.com
Newsgroups: comp.sys.next.programmer From: rouand@tnt.oleane.com (Jean-Michel Rouand) Subject: Wanted: DriverKit example of mouse/Tablet serial driver Message-ID: <1994Jun23.155418.21420@tnt.oleane.com> Sender: rouand@tnt.oleane.com Organization: Terra Nova Techonologies (France) Date: Thu, 23 Jun 1994 15:54:18 GMT Keywords: DriverKit Reply-To: rouand@tnt.oleane.com Hello NeXT people, We are currently trying to design a TouchScreen serial driver. The first thing we have done was to design an object which was receiving data from the touch screen serial line (ttya) building an NXEvent with these data and sending this NXEvent structure to the application event list by the DPSPostEvent() call. This is partially working since drag sessions are not correctly interpreted : events generated by this way cannot make Sliders or even ThumbWheel move. That kind of driver was therefore sufficient but cannot satisfy our needs now. So we decided to design a real driver based on DriverKit such as the Serial Mouse driver. But there, we encountered many problems in understanding how is working this DriverKit. We tried to understand the driver examples located in /NextDeveloper/Examples/DriverKit but there is no driver similar to the one we want. We then had a look to the EventProtocols.h header but that was not so simple. The problem is that we have not enough documentation on this kind of driver since classes like IOEventSource are not documented. Has someone an example more close to our problem ? How can we use the IOEventSource class which is not in the libDriver library ? How and to which object can we send the absolute pointer method ? How to write a line discipline for the tty driver ? Is there a way to build an event source object that would send events directly to the window server or the event driver, not based on DriverKit ? Thanks in advance Jean Michel ROUAND (NeXTMail welcome)
Newsgroups: comp.sys.next.programmer From: rouand@tnt.oleane.com (Jean-Michel Rouand) Subject: Recording 16 bits samples sounds Message-ID: <1994Jun27.104857.1079@tnt.oleane.com> Keywords: sound Sender: rouand@tnt.oleane.com Organization: Terra Nova Techonologies (France) Date: Mon, 27 Jun 1994 10:48:57 GMT We have some questions about processing sound with NEXTSTEP/FIP. Our audio device is a PRO-AUDIO SPECTRUM 16. We have to record and playback sounds with 16-bit samples. For this purpose we tried to use a NXSoundDevice object (and its subclass NXSoundIn) connected to a NXRecordStream object. Playback is OK. But unfortunately, NeXT Documentation (NXRecordStream.rtf) says : "The sound data in the buffers that are returned to the delegate is a single channel of 8-bit mu-law samples at the CODEC sanpling rates." Q1: is this a NeXT software limitation ? Obviouly, the PRO-AUDIO card is able to do 16-bit samples recording as it does under DOS O.S. Q2: if we can't find Objects in the sound kit to do our work (now or in further release), can we use driver functions ? wich and how ? do you have any examples ? Q3: if the PRO-AUDIO is limited, can we find any other sound device to do our work ? Q4: any other suggestions ... ? Thanks for your precious help. J-M Rouand
From: p9026363@deneb.cslab.tuwien.ac.at (Harald Brun) Newsgroups: comp.sys.next.programmer Subject: Macro: Record and Execute Date: 1 Jul 1994 11:44:56 GMT Organization: Technical University Vienna, Austria Message-ID: <2v0vjo$kuu@email.tuwien.ac.at> Hallo ! I'm looking for a code for the window - commands "Record Macro", "Execute Macro". Do you know existing code or how to get informations about my problem? Please answer, thank's !
From: akostyrk@track.cslab.tuwien.ac.at (Andreas Kostyrka) Newsgroups: comp.sys.next.programmer Subject: [Q] Objective C++ ? Date: 1 Jul 1994 11:51:11 GMT Organization: Technical University Vienna, Austria Message-ID: <2v0vvf$jib@email.tuwien.ac.at> Summary: Question: How to ObjC++? Keywords: Objective-C Objective-C++ compiler-options Hi folks, I've read the NeXT docs about Objective C++ and it sounds nice. I've got only one problem: I just overlooked the information how to make gcc compile ObjC++ sources? Thanx in advance, Andreas Kostyrka P.S.: please email, I do not follow this newsgroup often. -- \|/ \|/ (o o) (o o) -------------oOo--(_)--oOo-----------------oOo--(_)--oOo------ ( Email: * andreas@siegfried.smc.univie.ac.at [preferred] * akostyrk@cslab.tuwien.ac.at * * FTP: * siegfried.smc.univie.ac.at:/incoming/andreas * (if you put something there, then please email me * also a short note) * [preferred] (compared to UUENCODE) )
Newsgroups: comp.sys.next.programmer From: ian_stewart@nyro.com (Ian Stewart) Subject: Need help with FAT arch - cc options Message-ID: <Cs9AtE.1GF@nyro.com> Sender: ian@nyro.com (Ian Stewart) Organization: NYRO Technix, Inc. - makers of Faxcess(tm), UUCP-EZ(tm) and other NEXTSTEP software. Date: Fri, 1 Jul 1994 10:08:02 GMT I have noticed that more developers are using the CPU_SUBTYPE_<arch>. I know I can add to my makefile.preamble something like: OTHER_CFLAGS = -arch i486 -arch m68040 But how do I get NEXTSTEP to stop adding the default -arch on its own. i.e. on a NeXT it automatically adds -arch m68k if you uncheck both arch types. Also if you select Intel only from the options button in PB, but then it creates a -arch i386 -arch i486 -arch m68040 version. (-arch i386 is autoadded by PB? ) If I: lipo -remove <unwanted_extra_auto_added_dam_arch> -output /something /MyApp.app/MyApp and then do a bunch of renaming etc everything is fine. This is a pain. Basically I want to turn the automatic -arch option off and only deal with the ones I specifically ask for! Ian --- NYRO Technix, Inc. 236 W. Portal Ave Suite 341 San Francisco CA 94127 415 664-1170 voice 415 664-5530 fax NYRO Technix, Inc. markets VirtSpace, Faxcess(tm), UUCP-EZ(tm), On-Vacation(tm), READIT-EZ(tm) and other custom software.
From: hauert@itp.unibe.ch (Christoph Hauert) Newsgroups: comp.sys.next.programmer Subject: Changing cursor from deactivated app Date: 1 Jul 1994 12:55:01 GMT Message-ID: <2v13n5$6k4@aragorn.unibe.ch> Keywords: cursor hi, when the mouse enters a window of our app the cursor changes its appeareance according to the pressed modifier keys (shift, ctrl, etc.). Now comes the problem: all this works fine as long as our app is the active application, but we would like to have the same mouse behaviour even when our app is deactivated! So, does anybody know out there how one can change the appeareance of the cursor when entering a window of a deactivated application? Any hints and tips appreciated! Thanks Chris -- Christoph Hauert Institute for Theoretical Physics Berne, Switzerland hauert@butp.unibe.ch (NeXTMail)
Newsgroups: comp.sys.next.programmer From: js@euler.hnv.icem.de(Juergen Sell) Subject: Do you like developing multiple thread apps? What else is there? Message-ID: <Cs8JF3.rG@euler.hnv.icem.de> Sender: js@euler.hnv.icem.de (Juergen Sell) Organization: Ink Unknown Date: Fri, 1 Jul 1994 00:16:15 GMT Well, I just tried to write my first multiple cthreaded app and it doesn't work and it's difficult to debug for me. Also I am surprised whether you really use cthreads within your apps for time consuming stuff. I do not like it so far as the appkit is not threadsafe and making it threadsafe comes with a stated factor 3 performance penalty. Apart from the debugging pain, you lose the nice objective touch and have to access instance vars via the object's pointer (if you do not use the threadsafe runtime nor the mutex locking facilities), argh. So my big question is: What technique would you recommend to implement an app with "background" processing where this processing needs to update the GUI and access lots of objects while the frontend remains ever ready for user input? Is using oneway messages via a NXConnection to a different (forked) process an alternative? But then you have to get data back somehow. Leaves me confused, so please enlighten my soul. Juergen --- Fon ++49-511-440688 NeXTMail welcome Fax ++49-511-440617 == What time do we live in when laying people off gets called {right,down}sizing, == when spontaneity and freedom gets associated with instant coffee?
From: apang@mindlink.bc.ca (Anthon Pang) Newsgroups: comp.sys.next.programmer Subject: Wanted: sbrk() like function Date: Fri, 01 Jul 94 07:37:55 -0700 (PDT) Organization: MIND LINK! - British Columbia, Canada Distribution: world Message-ID: <48018@mindlink.bc.ca> Could someone email me an sbrk() function that demonstrates using vm_allocate()? Thanks.
From: stufduff@albion.unmc.edu (Sean W. Duffy) Newsgroups: comp.sys.next.programmer Subject: Interbase dbkit & IB Examples Wanted Date: 1 Jul 1994 13:26:44 GMT Organization: University of Nebraska Medical Center, Omaha, NE, USA Message-ID: <2v15ik$75q@netserv.unmc.edu> Keywords: INTERBASE DBKIT IB EXAMPLES Hi, I'm working on a little project involving DBKit & Interbase. I'm looking for any source code examples using Interbase. I can find plenty of Oracle & Sybase stuff so please don't send those. What I'm looking for are things like connecting to the database, inserting new records, expanding & compressing bit mapped flags, ddl triggers and just about anything else that might be useful. TIA Sean (I did WHAT?) Duffy stufduff@albion.unmc.edu
From: Charles William Swiger <infidel+@CMU.EDU> Newsgroups: comp.sys.next.programmer,comp.unix.programmer Subject: Re: Help! Generic Unix programming problem... Date: Fri, 1 Jul 1994 10:58:05 -0400 Organization: Carnegie Mellon, Pittsburgh, PA Message-ID: <wi52vxm00WBO82Jw01@andrew.cmu.edu> In-Reply-To: <2v069j$4sb@zebu.abstractsoft.com> Excerpts from netnews.comp.sys.next.programmer: 30-Jun-94 Re: Help! Generic Unix pro.. by Sean T. Lamont: > >But Mach doesn't have named pipes! How do I duplicate this functionality > >under Mach? Can I do it with sockets? Or, what about Dup'ing file > >descriptors in the shell? (I'm still fuzzy wrt understanding dup(2)). > > BZZT. Wrong. look at the call to mkfifo(), or get gnu mknod, either method > will create a named pipe. Umm, Sean: You'll get a kernel panic under earlier versions of NEXTSTEP (pre 3.0?) if you actually use the named pipe, but the mkfifo() call does exist (!). I would imagine this wasn't what the person wanted. :-) Getting GNU mknod is a very good idea. -Chuck Charles William Swiger - WhiteLight Systems | "All the world's a stage, and" --------------------------------------------+ "we are merely players...." AMS & normal mail: infidel@cmu.edu | NeXTmail: chuck@cswiger.slip.andrew.cmu.edu | "Semper ubi sub ubi."
Newsgroups: comp.sys.next.programmer From: rpomeroy@aunext1.amd.com (Ron Pomeroy x(Coop)) Subject: mouseDown: in the Application icon Message-ID: <Cs9oI5.AK6@txnews.amd.com> Sender: news@txnews.amd.com Organization: Advanced Micro Devices, Austin TX Distribution: usa Date: Fri, 1 Jul 1994 15:03:39 GMT Dear NEXTSTEP Gurus. I want my Application's icon in the dock to respond to a single mouseDown:. I've tried something along the lines of [[NXApp appIcon] setDelegate: self]; in an appDidInit method in a Controller object (which is a delegate of the NXApp). In my Controller I implement a mouseDown:sender method, but I never gets invoked. It occured to me that the event mask for the Window (appIcon) might prevent single mouse clicks from making it through. Am I on the right track or am I in the woods on this one ? I also want to change the appIcon each time the user songle clicks it (it's like a power switch that toggles). Any hints there are also appreciated. -- Ronald Pomeroy Advanced Micro Devices CIM Applications Group rpomeroy@aunext1.amd.com
Newsgroups: comp.sys.next.programmer From: jay@charlie (Jay Liew) Subject: Font Class - getwidth of ... Bug ???? Message-ID: <1994Jul1.145344.18342@cs.uno.edu> Sender: news@cs.uno.edu Organization: University of New Orleans Date: Fri, 1 Jul 1994 14:53:44 GMT Why is the -(float)getWidthOf:(const char *)string function in the Font Class yields the same width for both Courier and Ohlfs (given the same point Size) Is there a bug in this function or am I misusing this method? Jay (jay@charlie.smobject.com)
Newsgroups: comp.sys.next.programmer,comp.sys.next.software From: sdavis@laforge.ksc.nasa.gov (Steve Davis) Subject: Re: Two-column Landscape Printing Utility? Message-ID: <1994Jul1.131244.5643@dale.ksc.nasa.gov> Sender: news@dale.ksc.nasa.gov Organization: NASA References: <1994Jun30.191658.28908@dale.ksc.nasa.gov> Date: Fri, 1 Jul 1994 13:12:44 GMT Everyone: Thanks for the suggestions; enscript works great! Steve
From: <JTRQC@CUNYVM.CUNY.EDU> Newsgroups: comp.sys.next.programmer Subject: NXBundle..how come..... Date: Fri, 1 Jul 1994 15:23:23 EDT Organization: City University of New York/University Computer Center Message-ID: <94182.152323JTRQC@CUNYVM.CUNY.EDU> Disclaimer: Author bears full responsibility for this post Hello.. How come when I load code in from an xxx.bundle file that contains a nib the nib stuff e.g. windows etc.. don't work???? How come [NXApp loadNibSection:......] doesn't work but [NXApp loadNibFile....] does work but I can't orderFront: etc.. my windows..????? Feedback please!! James T. Romano romano@emc-sole.com
Newsgroups: comp.sys.next.programmer From: jay@charlie (Jay Liew) Subject: Font Class - getwidth of ... Bug ???? Message-ID: <1994Jul1.191135.3541@cs.uno.edu> Sender: news@cs.uno.edu Organization: University of New Orleans Date: Fri, 1 Jul 1994 19:11:35 GMT Chuck from carnegie melon replied me with this : Ohlfs and Courier are really the same font (ie, same font metrics, mathematical glyph descriptions, etc); they simply have different bitmap representations so they look different on your screen. If you print them out to a printer, they'll print the same, which is why -getWidthOf returns the same size. No bug. ---------- Since that is the case how can i get the bitmap representations of the fonts I am working on and editor i am not getting the my desired tab stop widths using the getWidthOf method. Jay (jay@charlie.smobject.com)
From: sam_s@NeXT.com (Sam Streeper) Newsgroups: comp.sys.next.programmer Subject: Re: Font Class - getwidth of ... Bug ???? Date: 1 Jul 1994 21:14:43 GMT Organization: NeXT, Inc. Message-ID: <2v2103$4g6@rosie.next.com> References: <1994Jul1.145344.18342@cs.uno.edu> jay@charlie (Jay Liew) writes: > Why is the -(float)getWidthOf:(const char *)string function in the Font Class > yields the same width for both Courier and Ohlfs (given the same point Size) > Is there a bug in this function or am I misusing this method? Ohlfs was derived from courier, and the font itself is rather broken. I think you get the right result in any point size for which a screen font exists, but otherwise all bets are off. -sam -- Opinions expressed herein are not those of my employer. They're not even mine. They're probably wrong besides. How did they get in here, anyway?
Newsgroups: comp.sys.next.programmer From: Gregory_Mutzel@afs.com Subject: Re: NeXT's own IB Inspectors (Question) Message-ID: <1994Jul1.141411.3333@afs.com> Sender: mutz@afs.com References: <2upqcm$31p@hamblin.math.byu.edu> Date: Fri, 1 Jul 1994 14:14:11 GMT In comp.sys.next.programmer article <2upqcm$31p@hamblin.math.byu.edu> you wrote: > Does anyone know the names of the Attributes inspectors for each > of NeXT's own palette objects (Button, Matrix, Slider, > Text-in-ScrollView, etc.)? > > I need to know the names of these inspectors badly but can't > find them anywhere. > > +--------------------------------------------------------------+ > | Sean Luke This signature no verb | > | sean@digaudio.byu.edu ,,, <- finger for PGP key | > | sean@zapotec.math.byu.edu (o o) | > +------------------------oOO--(_)--OOo-------------------------+ Take a peek in ../InterfaceBuilder.app/English.lproj and you should find what you want. -Greg -- Gregory L. Mutzel, 215-653-0911 Anderson Financial Systems, Inc. email: Gregory_Mutzel@afs.com [NeXTmail Appreciated]
Newsgroups: comp.sys.next.programmer From: dmurray@music.glas.ac.uk (Duncan Murray(EMU92)) Subject: Project Builder won't compile edited files!!! Help!!!! Message-ID: <Cs9LqH.9AB@udcf.glasgow.ac.uk> Sender: news@udcf.glasgow.ac.uk (News) Organization: Glasgow University Computing Service Date: Fri, 1 Jul 1994 14:03:53 GMT Hi everyone, Project Builder on our network, refuses to compile files that it has previously compiled properly, even if the files have been changed. After building an app and changing the .m/.h files associated with it and then clicking Build/run, Project Builder just jumps straight to the Linking whatever.app/whatever message completely ignoring the changes to the files. Someone suggested that this is an NFS server problem and that resetting the time would do the trick, but no joy!!! Any ideas? Answers to me directly if possible. Duncan Murray, dmurray@music.glasgow.ac.uk NeXTMail accepted.
Newsgroups: comp.sys.next.programmer From: fred@tof.fdn.org (Frederic Pralong) Subject: NXImageView with EOF Message-ID: <1994Jul1.165415.5110@object_factory.fr> Sender: news@object_factory.fr Organization: THE OBJECT FACTORY - Paris, France Date: Fri, 1 Jul 1994 16:54:15 GMT Has someone try the NXImageView in EOF without problem ? If you success, can you explain how you do that ? Thanks FReD -- Frederic Pralong Paris France The Object Factory 5 rue de la cour des Noues 75020 Paris France e-mail : frederic_pralong@object-factory.fr
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.