From kc at omnigroup.com Wed Sep 1 00:58:49 1999 From: kc at omnigroup.com (Ken Case) Date: Thu Nov 3 14:42:15 2005 Subject: NSCStringText questions Message-ID: <199909010758.AAA21380@scyther.omnigroup.com> > I need to convert an app from NEXTSTEP to OPENSTEP/YellowBox. This > app makes heavy use of NSCStringText's find features. How do I solve > this problem whithout using NSCStringText anymore? Is there some > kind of a find controller somewhere in the public, or ... ?! There are at least two: You'll find the source code to Omni's find controller (OAFindController, the one used in OmniWeb and OmniPDF) in the source code to our OmniAppKit framework--download it from http://www.omnigroup.com/community/developer/sourcecode/. This isn't really specific to the Text system, though: we also use it to search PDF documents, bookmarks, etc. What might be a closer fit for your purpose is the source code to TextEdit's Find/Replace class, TextFinder, which you'll find in /System/Developer/Examples/Java/AppKit/TextEdit/. (It's actually an Objective C class, despite being part of the TextEdit Java example.) This TextFinder class supports full undo using the new undo mechanism introduced in Mac OS X Server. Hope this helps! Ken From jerome at openbase.com Wed Sep 1 11:18:35 1999 From: jerome at openbase.com (Jerome Lebel) Date: Thu Nov 3 14:42:16 2005 Subject: Convert color to a RGB Color Message-ID: <9909011826.AA00277@mozart.openbase.com> I would to know RGB component of some system colors (like controlColor). How can I convert a color from one space to another ? I found nothing about that on the doc. Thanks, -- Jerome Lebel jerome@openbase.com +1-603-547-8404 ext. 207 From andrew at stone.com Wed Sep 1 14:22:48 1999 From: andrew at stone.com (Andrew Stone) Date: Thu Nov 3 14:42:16 2005 Subject: Project Builder emacs keys? Message-ID: <199909012123.PAA05634@floyd.stone.com> >Is there a bundle that activates emacs-style keys in PB? This activates emacs keys in EVERY NSTextView and NSTextField, save this to the following path: /* ~/Library/KeyBindings/DefaultKeyBinding.dict */ { /* Additional Emacs bindings */ "~f" = "moveWordForward:"; "~b" = "moveWordBackward:"; "~<" = "moveToBeginningOfDocument:"; "~>" = "moveToEndOfDocument:"; "~v" = "pageUp:"; "~d" = "deleteWordForward:"; "~^h" = "deleteWordBackward:"; "~\010" = "deleteWordBackward:"; /* Alt-backspace */ "~\177" = "deleteWordBackward:"; /* Alt-delete */ /* Escape should really be complete: */ "\033" = "complete:"; /* Escape */ } === Andrew Stone, QuixMaster of Stone Design Corp mailto:andrew@stone.com (505) 345-4800 http://www.stone.com - Create(TM) From andrew at stone.com Wed Sep 1 14:28:22 1999 From: andrew at stone.com (Andrew Stone) Date: Thu Nov 3 14:42:16 2005 Subject: Convert color to a RGB Color Message-ID: <199909012128.PAA05642@floyd.stone.com> > I would to know RGB component of some system colors (like > controlColor). How can I convert a color from one space to another ? It's so easy you're going to laugh: NSColor *rgbColor = [color colorUsingColorSpaceName:NSCalibratedRGBColorSpace]; > Jerome Lebel > jerome@openbase.com > +1-603-547-8404 ext. 207 > > === Andrew Stone, QuixMaster of Stone Design Corp mailto:andrew@stone.com (505) 345-4800 http://www.stone.com - Create(TM) From bungi at omnigroup.com Wed Sep 1 16:01:16 1999 From: bungi at omnigroup.com (Timothy J. Wood) Date: Thu Nov 3 14:42:16 2005 Subject: EGCS on MacOS X Server (was Re: Using STL in Cocoa apps) Message-ID: <199909012301.QAA01161@electabuzz.omnigroup.com> >p.s.: Did anyone else notice the AltiVec descriptions in the ppc >configuration subdirectories? Yeah, there is a whole ChangeLog.AltiVec that has stuff about this. I haven't looked at it yet since I hadn't looked for it yet. Yeah! -tim From pshum at alitech.com Thu Sep 2 12:56:49 1999 From: pshum at alitech.com (Paul Shum) Date: Thu Nov 3 14:42:16 2005 Subject: Installation of packages through a telnet session Message-ID: <013901bef57d$4ba9f820$0202010a@ali.bc.ca> Hi, Apple has released NS3.3 User Patch 2 and OS4.2 Mach User Patch 4. We need to install these Y2K packages to our customers. As our customers are all over the States and Canada, we would like to do the installation of the packages through a telnet session (i.e. we cannot use the Installer to do the installation). Does anyone have any idea how to do the job? I have tried to uncompress the OS42MachUserPatch4.tar.Z and use installer_bigtar xpvfiT to extract the files. After the extraction has completed, I run the OS42MachUserPatch4.post_install script. However, I do not feel very comfortable because the files installed in this way have different file sizes than the ones installed using the Installer. In addition, there is no receipt generated under /NextLibrary/Receipts. Could anyone show me some guidelines what I should do? Thanks! _____________ Paul S. L. Shum Christianity is not a way of life but is the truth of life. From karl at nfox.com Thu Sep 2 16:24:32 1999 From: karl at nfox.com (Karl Kraft) Date: Thu Nov 3 14:42:16 2005 Subject: Installation of packages through a telnet session References: <013901bef57d$4ba9f820$0202010a@ali.bc.ca> Message-ID: <004d01bef59a$5138d3c0$0329040a@nfox.com> You might get a better response that what follows if you directed this to the admin list of osx, or even one of the next newsgroups. This group is for discussion of OSX. > However, I do not feel very comfortable because the files installed in this > way have different file sizes than the ones installed using the Installer. I don't have any nextstep machines anymore, so I haven't looked at this patch. My guess would be that the files you are installing this way are larger becuase you have installed them FAT instead of for just the specific architecture. If the files are smaller, then you have a real problem. The CDIS stuff has actual usable telnet tools for installing packages, and stripping them down to a specific or set of architectures. > In addition, there is no receipt generated under /NextLibrary/Receipts Yeah, neither tar nor uncompress will do that magic for you. You could do it by hand, or you could use those CDIS tools. You will find them on your original 3.x CDs. They are used by the initial install process. Why do you feel you need the receipt? Are you planning on unistalling the Y2K patch later :-) ? From cybergrog at bigpond.com Thu Sep 2 21:26:45 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:16 2005 Subject: Getting the paper source from NSPrinter. Message-ID: <04225871006352@domain0.bigpond.com> I am trying to get the paper sources available from the printer chosen in a popup menu. The code looks something like this: (NSArray *) getPaperTrays:(NSString *)printerName { NSPrinter *selectedPrinter = [NSPrinter printerWithType:printerName]; NSArray *printTrays = [selectedPrinter stringListForKey:@"InputSources" inTable:@"PPD"]; return printTrays; } I call this function when the user changes the printer from one to another so it updates the correct paper sources available. I believe that thing I am doing wrong is passing the wrong strings into stringListForKey as it returns nil. I could not find how to get the correct list when reading the NSPrinter information. Any help is greatly appreciated. Greg. From cybergrog at bigpond.com Fri Sep 3 07:26:59 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:16 2005 Subject: Reading a file into an NSTextView Message-ID: <14243809719617@domain6.bigpond.com> How does one read in a file into an NSTextView. I have found this documentation quite confusing - but I am sure that it is not. Any help is greatly appreciated. Greg From waynez at erols.com Fri Sep 3 07:46:08 1999 From: waynez at erols.com (Wayne) Date: Thu Nov 3 14:42:16 2005 Subject: Reading a file into an NSTextView References: <14243809719617@domain6.bigpond.com> Message-ID: <010e01bef61b$110ee010$c959453f@rover> Greg, Check out: http://www.stepwise.com/Articles/Technical/HTMLEditor/index.html This example was provided by Scott Anguish and in addition to answering your question deals with some other issues your likely to encounter. Wayne ----- Original Message ----- From: Greg Hulands To: Multiple recipients of list Sent: Friday, September 03, 1999 10:42 AM Subject: Reading a file into an NSTextView > How does one read in a file into an NSTextView. I have found this > documentation quite confusing - but I am sure that it is not. > > Any help is greatly appreciated. > > Greg > > From cybergrog at bigpond.com Fri Sep 3 21:03:40 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:16 2005 Subject: Listening for an NSNotification Message-ID: <04013481034483@domain2.bigpond.com> I need to listen for an NSNotification for when the window is about to be close, so I can save the state of the objects that are used in the controller. What message do I listen for. Greg From jjfeiler at relief.com Fri Sep 3 21:40:06 1999 From: jjfeiler at relief.com (John Jay Feiler) Date: Thu Nov 3 14:42:16 2005 Subject: Listening for an NSNotification Message-ID: <199909030844.BAA23046@haque.relief.com> > From: "Greg Hulands" > I need to listen for an NSNotification for when the window is about to be > close, so I can save the state of the objects that are used in the > controller. What message do I listen for. APPKIT_EXTERN NSString *NSWindowWillCloseNotification; -- John Feiler jjfeiler@relief.com From cybergrog at bigpond.com Fri Sep 3 23:19:40 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:16 2005 Subject: DocDisplayer Message-ID: <06172077334397@domain4.bigpond.com> I have made an object that allows you to display any rtf or rtfd files in a window and title it. I am using it in the app I am currently writing. The rtf files need to be in the application wrapper. Hope someone finds it useful. Greg -------------- next part -------------- An HTML attachment was scrubbed... URL: /mailman/archive/macosx-dev/attachments/19990904/b8de0fd3/attachment.html From cybergrog at bigpond.com Sun Sep 5 07:33:30 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:16 2005 Subject: NSTableView Message-ID: <14312677350297@domain7.bigpond.com> I have an NSTableView which I want to be able to have an empty record available to fill in (like MS Access's table view). When I enter something in any of the fields in the new row, it will then create another empty row. I thought NSTableView would implement this cause I have told it the data source. What is the best implementation to use? Any pointers *greatly appreciated! Greg. From mark at jiiva.com Sun Sep 5 14:02:15 1999 From: mark at jiiva.com (Mark Ericksen) Date: Thu Nov 3 14:42:16 2005 Subject: Choosing paper-tray for printing (NT) Message-ID: <199909052055.PAA23955@dfw-ix6.ix.netcom.com> I have this very issue. According to Apple technical support there is a bug in the AppKit when parsing PPD files which prevents any other paper tray from being selected and used. They told me that the bug was scheduled to be fixed for MacOS X (Personal Edition). No patch will be provided. I have a brand new HP 4050TN printer that will only print from tray 3... Mark Ericksen Jiiva, Inc. Markus Ullius Wrote: > Hi > > I'm developing an application which should also print some created documents. > Now there is a problem in choosing the tray. Independant of the printer > settings under NT (where tray 2 is selected), the printer always takes its > paper from tray 1? > > What can I do? > > Thanks > > Markus > ------------------------------------------------------------------ > Markus Ullius > Institute of Transportation, Traffic, Highway & Railway Engineering (IVT) > ETH Hoenggerberg > 8092 Zurich > Switzerland > > Fax: +41 1 633 10 57 > Phone: +41 1 633 31 05 > Direct: +41 1 633 24 15 > > http://www.ivt.baum.ethz.ch > http://www.ivt.baum.ethz.ch/opentimetable.html > > From ahoesch at smartsoft.de Mon Sep 6 08:55:57 1999 From: ahoesch at smartsoft.de (Andreas Hoschler) Date: Thu Nov 3 14:42:16 2005 Subject: Autoresizing NSTextView Message-ID: <9909061555.AA04514@smartsoft.de> Hi, I would like to autoresize an NSTextView if the (entered) text reaches the right edge of the textview so that the text can still be laid out in one line. Is there any notification or delegate method or whatever that can be used to determine when and how much the textview width has to be increased (automatically or manually)? I've played a bit with - setHorizontallyResizable: and - sizeToFit, unfortunately without success yet. How are these methods intended to be used? Thanks a lot, Andreas From fischer at fokus.gmd.de Mon Sep 6 07:57:08 1999 From: fischer at fokus.gmd.de (Robert Fischer) Date: Thu Nov 3 14:42:16 2005 Subject: NSPerformService ruins NSPasteboard? Message-ID: <37D3D644.1D008983@fokus.gmd.de> Hi, I am trying to invoke a service programmatically with NSPerformService(NSString*, NSPasteboard*). At the first time everything works as expected, but then I got a 'Bad return value'. In the following ProjectBuilder from which I started my app will do no more pasteboard operations until being restarted. I am using the pasteboard [NSPasteboard generalPasteboard], which should be the same, whenever i call it, or? Has anyone a clue what's going on? Thanks in advance, Robert -- --- - .-. -- -- --- / \ ---- Robert Fischer .-. / \ --- .-. __o .-. @ / \ / \ / \ _`\<,_ / \ GMD-Fokus / \ / \ / \ (*)/ (*) / `-------------- / `---' `-' `-----------' From paulrs at lgs-systems.com Mon Sep 6 07:08:50 1999 From: paulrs at lgs-systems.com (Paul R. Summermatter) Date: Thu Nov 3 14:42:16 2005 Subject: Context Menu & selections Message-ID: <9909061408.AA01291@slab> Hey guys, I recently posted on a problem regarding context menus (as generated by menuForEvent:) and mouseDown: not being called. In an attempt to resolve this issue, I took matters into my own hands in menuForEvent: and set the selection based on the mouse location. In one instance, I have a context menu popping up in a custom subclass of NSMatrix. I want the selection in the matrix to change based on the location of the right (or control) click in the matrix. I added code to find the appropriate cell for the click and select it in the matrix. On MOSXS, this worked fine and the selection changed before the context menu appeared. However, on NT, the selection does not cause a redisplay until after the context menu is dismissed. I assumed that I could force the issue by calling an explicit 'display' on the matrix just after changing the selection, but this had no impact. Is this a bug on NT? It appears that normal event processing does not go on while a context menu is raised on NT (ie, views are not being asked to redisplay themselves each time through the loop). Does anyone know how else I might work around this problem? Regards, Paul --- Paul Summermatter LGS Systems, Inc. Medical Computing Division 15 TJ Gamester Ave Portsmouth, NH 03801-5871 (603) 433-9822 voice (603) 433-9818 fax (603) 498-6935 voice mail paulrs@lgs-systems.com (NeXT or MIME Mail Welcome) http://www.lgs-systems.com From buck.erik at mcleod.net Mon Sep 6 20:03:29 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:16 2005 Subject: Context Menu & selections Message-ID: <008501bef8dd$b6747960$b52a10d0@elcry> I do not know how to work around your redisplay problem when a menu is up. I can suggest that perhaps you shouldn't work around it. Windows Apps do not select with a right click. Windows File Explorer is the perfect example. Right clicking an icon does not change the selection. I honestly do not know which behavior is better. I can only guess at which behavior is expected. From jojo at thinkapple.com Tue Sep 7 07:53:09 1999 From: jojo at thinkapple.com (Adrian) Date: Thu Nov 3 14:42:16 2005 Subject: Offscreen graphics Message-ID: I was wondering if there is a better strategy than using NSBitmapImageRep for directly setting pixels in its offscreen bitmap and then pushing it out to screen as you normally do? Is there anything faster/efficient? -- Kind Regards Adrian From bbum at codefab.com Tue Sep 7 09:38:39 1999 From: bbum at codefab.com (Bill Bumgarner) Date: Thu Nov 3 14:42:16 2005 Subject: New version of Mssh released (and mailing lists) Message-ID: <199909071638.MAA15005@bjork.codefab.com> Mssh version 0.01 is now available. Lots of new features, bug fixes, and other hacquery in this version-- release notes at end of message. I have also created a mailing list for the discussion of mssh; it will very likely also include discussions of runtime hacquing, framework hacquing, and otehr relatively undocumented fun stuff that one can do to the Apple supplied frameworks. See the following for access to the archives and to subscribe (or manage your subscription): http://darwin.codefab.com/mailman/listinfo/mssh You can also send email to mssh-request@darwin.codefab.com with 'help' in the subject or body and the system will reply with documentation as to how to subscribe/unsubscribe via email. (To subscribe, simply send a message with "subscribe" in the subject/body). I will be making various random extensions, scripts, and hacques available shortly and, once (if) there is a reasonable sized community on the mailing list, will be asking for volunteers to script up some solutions to various nagging problems and to finish off partial solutions that have already been implemented. enjoy! b.bum Release 0.01: - Formalized Scripted Module support. Scripted modules have: - search paths and search bundles - cache of already loaded modules - Added defaults module; when used, it adds the name of the script to the defaults domains that are searched. This allows per-standalone-script defaults. - Improved command line modules ability to find executables. - differentiates between absolute paths and relative paths (i.e. won't search if executable "name" passed to -which: is already a full path to an executable - now has "-launchedTaskWithCommand:arguments:" convenience API that will create and execute an NSTask given a command name (including findind the command's executable by search the path environment variable). - now has "-stringFromExecutingCommand:arguments:" API that returns the string based result of executing a command. - Added a Java based scripted module. This module makes it trivial to wrap a java package for execution from the command line. It takes care of searching for the necessary packages, setting up various environment variables, and launching the 'java' virtual machine with the appropriate arguments. (see javatest.wos example in Scripts directory) - Added a proxy object that makes it possible to use a scripted object as the target of distributed notifications, invocations, and the various performSelector: methods. (see example(s) in Scripts that contain references to "ScriptedPerformProxy") - exposed and normalized more of the NSScanner, NSPAthUtilities, NSString and NSFileManager APIs that were not previously available in scripted code (because they either used ranges or had arguments like (NSString **)). - exposed many more "standard" variables as globals. Invoke the showglobals.wos example in Scripts. - added a number of new examples; includes darray.wos -- a command line tool for easily manipulating the contents of default values that contain arrays. - added API for retrieving full set of command line arguments Release 0.00: - execute webscript files from shell - write standalone shell scripts by starting script with: #!/usr/local/bin/mssh (Path should be adjusted to your local installation) - define most common global variables that are normally found in ObjC; string encodings, exception names, etc... - beginnings of scripted module support (includes a command line module that has notion of executable paths and 'which' command) From mark at jiiva.com Tue Sep 7 11:19:28 1999 From: mark at jiiva.com (Mark Ericksen) Date: Thu Nov 3 14:42:16 2005 Subject: distant object invalidation Message-ID: <199909071813.NAA03408@dfw-ix6.ix.netcom.com> Greetings, I have a distributed objects question. I have a client and server connected via distributed objects. The server stores objects from the client and passes them on to other clients. The server also stores objects in a cache in case a new client registers and needs to catch up with the others. All is working great except for one thing. When I client dies and is unable to unregister properly, I need to clean up the server so it doesn't contain stale object proxies. So far my attempt is to do the following: I register for the NSConnectionDidDieNotification and try to ask each NSDistantObject if this is its connection so I can then remove the object from the cache. However when I try to access the NSDistantObject that is from the dead connection I get an expection about accessing the NSDistantObject from another thread... NSThread says that I am not multi-threaded, however. What am I doing wrong? What is the proper way to handle dead connections? Is there a hidden NSDistantObjectInvalidatingNotification? That would be perfect. Thanks, Mark Ericksen Jiiva, Inc. From alexnet at gestalt.com Tue Sep 7 13:03:17 1999 From: alexnet at gestalt.com (Alex Molochnikov) Date: Thu Nov 3 14:42:16 2005 Subject: distant object invalidation Message-ID: <9909072003.AA00527@gestalt.com> Here is how we do it in STEP FORWARD which is also a Client-Server system: 1. Register an object for NSConnectionDidDieNotification notification with senderIsInvalid: method to be called. Allocate an NSArray in the Server to hold ClientRecord objects (say, clientArray). These objects are intended to hold identifying information about the remote Client, e.g. the proxy to the Client's object. 2. When a client registers in the Server, have the Client pass one of its objects back to the Server (as a callback). On the Server side, implement the following method that the Client should call as a part of registering: - (void) registerClient: (id) proxy { id client, connection; connection = [proxy connectionForProxy]; client = [[ClientRecord alloc] initWithConnection:connection]; [clientArray addObject:client]; [[NSNotificationCenter defaultCenter] addObserver:self selector:@selector(senderIsInvalid:) name:NSConnectionDidDieNotification object:connection]; } This is a bare bones method, add any other info about the remote client you want to keep in the ClientRecord object instantiation or in separate methods. 3. Implement the notification handling method: - (void) senderIsInvalid: (NSNotification *) sender { int i, count; id client, deadConnection; NSString *eName; deadConnection = [sender object]; [[NSNotificationCenter defaultCenter] removeObserver:self name:NSConnectionDidDieNotification object:deadConnection]; count = [clientArray count]; for (i = 0; i < count; i++) { client = [clientArray objectAtIndex:i]; if ([client connection] == deadConnection) { /* ** Do your cleanup here. If you want to be on ** the safe side, add the following error handling code: */ NS_DURING /* ** Do something... */ NS_HANDLER eName = [localException name]; if (![eName isEqualToString:NSExceptionYouWantToSuppress]) [localException raise]; NS_ENDHANDLER break; } } } BTW, the NSConnectionDidDieNotification will NOT work on WO/NT if the Server runs as an NT Service. This is a known and very old bug that neither NeXT nor Apple ever bothered to fix. Hope this helps. Alex Molochnikov Gestalt Corporation From gclem at frontline-software.dk Tue Sep 7 13:33:09 1999 From: gclem at frontline-software.dk (Geert B. Clemmensen) Date: Thu Nov 3 14:42:16 2005 Subject: Timers and notifications in Java Message-ID: <199909072033.WAA28108@snaps.frontline-software.dk> Hello there, Any good ideas on how to achieve -scheduledTimerWithTimeInterval: and NSNotification (and associated methods) functionality in in pure Java? Thx. for your time, Geert B. Clemmensen Frontline Software From xavierb at proxyware.fr Wed Sep 8 02:51:18 1999 From: xavierb at proxyware.fr (Xavier Barrier) Date: Thu Nov 3 14:42:16 2005 Subject: distant object invalidation In-Reply-To: <199909071813.NAA03408@dfw-ix6.ix.netcom.com> References: <199909071813.NAA03408@dfw-ix6.ix.netcom.com> Message-ID: <199909080951.AA13904@proxyware.fr> You wrote: > Greetings, > > I have a distributed objects question. > > I have a client and server connected via distributed objects. The server > stores objects from the client and passes them on to other clients. The > server also stores objects in a cache in case a new client registers and > needs to catch up with the others. All is working great except for one > thing. When I client dies and is unable to unregister properly, I need to > clean up the server so it doesn't contain stale object proxies. > > So far my attempt is to do the following: > I register for the NSConnectionDidDieNotification and try to ask each > NSDistantObject if this is its connection so I can then remove the object > from the cache. However when I try to access the NSDistantObject that is > from the dead connection I get an expection about accessing the > NSDistantObject from another thread... NSThread says that I am not > multi-threaded, however. > > What am I doing wrong? What is the proper way to handle dead connections? > Is there a hidden NSDistantObjectInvalidatingNotification? That would be > perfect. > > Thanks, > > Mark Ericksen > Jiiva, Inc. > > It is normal that accessing the distant object via a dead NSConection fails, how could it be possible to acces it since the NSConnection is no more present ;-) the best way to handle it, would be to store the proxies linked to each NSConnection. When receiving NSConnectionDidDieNotification you are then able to free evry proxy using this NSConnection. BTW we are handling the same kind of pb in our architecture, and solved it like it is described above! Good luck --- Xavier BARRIER --------------------------------------------------------- ProxyWare SARL 147, rue Saint Martin Paris France Financial Software Company Tel : (+33) 1 42 71 97 66 Fax : (+33) 1 42 71 97 06 e-mail : xavierb@proxyware.fr From rvamerongen at zeelandnet.nl Wed Sep 8 05:44:49 1999 From: rvamerongen at zeelandnet.nl (R.H van Amerongen) Date: Thu Nov 3 14:42:16 2005 Subject: Localization Message-ID: Hi All, Maybe someone can use it. Do you want to program or Localize a language who isn't available? Let say you want to localize you App. to the Dutch Speaking People. There a lot of them in Belgium, Suriname, Curacao, Aruba, St. Maarten, South of Africa, America, Canada Australia and in Holland (The Netherlands) the country with the wooden shoes. Ok enough serious talk. First. change inside the Preferences.framework ( /System/Library/PrivateFrameworks ) the file global.defaults - NSLanguages = (Dutch, English, French, German, Japanese, Spanish, Italian, Swedish); AvailableLanguages = (Dutch, English, French, German, Japanese, Spanish, Italian, Swedish); - change the file language.strings in /System/Library/CoreServices/Resources - /* in native languages. */ "Netherlands" = "Nederland"; "Dutch" = "Nederlands"; "England" = "England"; - Second. To change PB to get this working. Change MainInfo.table The file exists in System/Library/Frameworks/ProjectBuilder.framework/Versions/Current/Re sources/ change MainInfo.table to - Architectures = (C64, ppc, i386); HumanLanguages = ("Dutch", "English", "French", "German", "Spanish", "Portuguese", "Italian", "Swedish", "Japanese"); CompilerLanguages = ("ObjectiveC++", "Java"); CompilerLanguagesPrefixes = (OBJCPLUS_, JAVA_); - So far as I see it is not really needed to sort them alphabetic LET OP!!! <== Dutch for 'Be Aware'. You see it works already. You did make a backup of those files before you....... Greetings Ren? From cybergrog at bigpond.com Wed Sep 8 04:20:17 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:16 2005 Subject: oddities in my for loop Message-ID: <13044293120068@domain2.bigpond.com> I am having a strange thing happen to me when using a for loop and an array. The code: for (i=0; i<[exercises count]-1; i++) //exercises is the array { newExercise = [[Exercise alloc] init]; [newExercise setName: [exercises objectAtIndex:i]]; [myWorkout addExercise:newExercise]; [newExercise autorelease]; } I put 2 items in the array (exercises), so the [exercises count] message should return 2, which it does, the only thing wrong is that it assigns it to i (that is i = 2). That is the strange thing - an assignment taking place in the condition of the loop. Another strange thing about it is that if i=2 then it doesn't satisfy the condition but allows it to go through the loop. This causes an object out of bounds exception in the [newExercise setName:[exercises objectAtIndex:i]], which would be expected because i = 2. Any Suggestions or explanations? Greg From paulrs at lgs-systems.com Wed Sep 8 05:41:17 1999 From: paulrs at lgs-systems.com (Paul R. Summermatter) Date: Thu Nov 3 14:42:16 2005 Subject: Context Menu & selections In-Reply-To: <008501bef8dd$b6747960$b52a10d0@elcry> References: <008501bef8dd$b6747960$b52a10d0@elcry> Message-ID: <9909081241.AA02715@slab> You wrote: [ I do not know how to work around your redisplay problem when a menu is [ up. I can suggest that perhaps you shouldn't work around it. Windows [ Apps do not select with a right click. Windows File Explorer is the [ perfect example. Right clicking an icon does not change the selection. I [ honestly do not know which behavior is better. I can only guess at which [ behavior is expected. [ Eric, Based on my tests (NT 4.0 sp 4), right clicking a file or folder in the Explorer window DOES select the item you have clicked prior to bringing up the context menu. I tried a couple of other applications such as Outlook and found the same behaviour. I believe this all relates to the same problems that we have with being unable to redraw views while dragging and dropping on NT. If any of the AppKit guys could comment on this, I would love to hear it. Regards, Paul --- Paul Summermatter LGS Systems, Inc. Medical Computing Division 15 TJ Gamester Ave Portsmouth, NH 03801-5871 (603) 433-9822 voice (603) 433-9818 fax (603) 498-6935 voice mail paulrs@lgs-systems.com (NeXT or MIME Mail Welcome) http://www.lgs-systems.com From mike at lorax.com Wed Sep 8 07:50:51 1999 From: mike at lorax.com (mike) Date: Thu Nov 3 14:42:16 2005 Subject: Autoresizing NSTextView Message-ID: <199909081544.IAA09425@boom.lorax.com> See the TextSizing example. Basically you want a very wide (say 1.0e6) text container that does not track the text view width, and a horizontally resizable text view. The view will then size itself to the usage of the container and be as wide as necessary to display the text. The text will never wrap (until it gets a million pixels wide :-). Mike Ferris Begin forwarded message: > From: Andreas Hoschler > Date: 1999-09-06 08:05:50 -0700 > To: Multiple recipients of list > Subject: Autoresizing NSTextView > X-Nextstep-Mailer: Mail 4.2mach (Enhance 2.1) > X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas > Originator: macosx-dev@omnigroup.com > X-Comment: To unsubscribe, follow directions at > http://www.omnigroup.com/MailArchive/ > > Hi, > > I would like to autoresize an NSTextView if the (entered) text reaches the > right edge of the textview so that the text can still be laid out in one > line. Is there any notification or delegate method or whatever that can be > used to determine when and how much the textview width has to be increased > (automatically or manually)? > > I've played a bit with - setHorizontallyResizable: and - sizeToFit, > unfortunately without success yet. How are these methods intended to be used? > > Thanks a lot, > > Andreas > From hordur at mail.lausn.is Wed Sep 8 09:31:51 1999 From: hordur at mail.lausn.is (hordur) Date: Thu Nov 3 14:42:16 2005 Subject: Character encoding woes - help needed ! Message-ID: <199909081631.QAA29741@gremlin.skyrr.is> Hi, I am having an awful problem with my WO app. I am using an OpenBase 6.0.4 database (local to my machine) and when I run my WO app on my machine and insert a record with foreign characters in some strings (like a name), the value that goes into the database is all messed up. On another machine which is also running MacOSXServer and OpenBase 6.0.4 and the same WO app, everything works fine, all characters that go into the database are fine. I have EOAdaptorDebugEnabled turned on for my app and I can see exactly what the INSERT statement looks like on both machines. When I insert a record with some foreign (special Icelandic characters, the same INSERT statement is being generated on both machines. The only difference is that when I go into OpenBaseManager and examine what was inserted, it is messed up on one machine and not on the other one. By the way, both databases are set to ASCII encoding and all the webpages/webbrowsers being used are set to Windows-Latin1 encoding. It does not matter what webbrowser is used (I've tried Netscape 4.5 and Omniweb 3.0). Any light anyone could possibly shed on this would be ever so much appreciated since I am totally clueless as to why this is happening. I've allready re-installed OpenBase on the machine where things are not OK and that did not help (I also deleted and recreated the database). One other thing I tried was replicating the database in question from the machine where things are OK, to the database on the machine where things are NOT OK, and the replicated records are identical in both databases. Sincerely, Hordur Thordarson Iceland From mtarbell at tarbell.jpl.nasa.gov Wed Sep 8 09:33:58 1999 From: mtarbell at tarbell.jpl.nasa.gov (Mark Tarbell) Date: Thu Nov 3 14:42:16 2005 Subject: Localization In-Reply-To: References: Message-ID: <199909081633.JAA02507@tarbell.jpl.nasa.gov> > Do you want to program or Localize a language who isn't available? 'Localization' is fairly slick, but the downside is that you'll still need to manually create & arrange a duplicate set of .nibs for each language you wish to support. For a large project with many supported languages, it can really be a pain to have to go into each .nib, for each language, for every change. (It would be nice if there were an alternative to this based on a single set of .nib files. You can programmatically update text fields, buttons, etc., on a .nib (for instance, using a single English.lproj with TextFields (no hard-coded text), and use NSLocalizedString() with -setStringValue: to populate the fields), but this doesn't work out very well due to the differing lengths of text strings and fonts. And, having to populate each .nib as it is brought up is kludgey.) Mark -- Mark Tarbell (Mark.A.Tarbell@jpl.nasa.gov) Applications Development Section Information Systems Development and Operations Division Engineering and Science Directorate Jet Propulsion Laboratory 4800 Oak Grove Drive MS 502-500 Pasadena, CA 91109 Opinions expressed herein do not reflect those of JPL. From mark at jiiva.com Wed Sep 8 09:58:06 1999 From: mark at jiiva.com (Mark Ericksen) Date: Thu Nov 3 14:42:16 2005 Subject: Conclusion - distant object invalidation Message-ID: <199909081652.LAA08479@dfw-ix4.ix.netcom.com> Thanks to all those who replied to my post. I appreciate the quick responses. The error I received regarding the access of the distant object from another thread was resolved by calling [connection runInNewThread] after I setup the connection and registered it with the name service. I admit that I do not know the technical reason that this solves the issue, but I'm glad it works. Additionally, a few replies I received were concerned about accessing an NSDistantObject when its connection is dead. This is okay to do as long as you're not causing the NSDistantObject to forward its message to the real object. NSDistantObject has a method named, connectionForProxy, which returns the NSConnection object that I used to compare. What does make a difference is how the NSDistantObjects are stored in your cache. I have a dictionary with array values. The array contains more dictionaries and the distant objects are in these dictionaries. When I wanted to invalidate the connection and remove it from the cache, I tried to remove the dictionary from the array. This, to my surprise, caused a copy operation and therefore the distant object tried to access its real object causing another exception. Instead I removed the invalid distant object from its dictionary first and then remove the dictionary from the array. This worked for me and now my cache is being cleaned properly. Thanks again. Mark Ericksen Jiiva, Inc. Greetings, I have a distributed objects question. I have a client and server connected via distributed objects. The server stores objects from the client and passes them on to other clients. The server also stores objects in a cache in case a new client registers and needs to catch up with the others. All is working great except for one thing. When I client dies and is unable to unregister properly, I need to clean up the server so it doesn't contain stale object proxies. So far my attempt is to do the following: I register for the NSConnectionDidDieNotification and try to ask each NSDistantObject if this is its connection so I can then remove the object from the cache. However when I try to access the NSDistantObject that is from the dead connection I get an expection about accessing the NSDistantObject from another thread... NSThread says that I am not multi-threaded, however. What am I doing wrong? What is the proper way to handle dead connections? Is there a hidden NSDistantObjectInvalidatingNotification? That would be perfect. Thanks, Mark Ericksen Jiiva, Inc. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 4069 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990908/c089cb7d/attachment.bin From dyoung at vviuh221.vvi.com Wed Sep 8 10:26:20 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:16 2005 Subject: Interesting quotes from Sun Message-ID: <9909081726.AA02473@vviuh221.vvi.com> Excuse the interruption but I thought these quotes from Sun regarding their Sun Ray computer are interesting. John Loiacono, VP marketing at Sun says when JavaStation floundered: "... We've learned that users don't want to just use Java, ..." Sounds like Apple is the only one promoting Sun's Java technology on the desktop nowadays. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com Full article: http://www.mercurycenter.com/svtech/news/indepth/docs/sun090899.htm From rthomas at uiuc.edu Wed Sep 8 19:57:34 1999 From: rthomas at uiuc.edu (Ryan Thomas) Date: Thu Nov 3 14:42:16 2005 Subject: EGCS on MacOS X Server (was Re: Using STL in Cocoa apps) In-Reply-To: <199909010407.VAA17644@electabuzz.omnigroup.com> Message-ID: <4.1.19990908215531.00b08100@ntx1.cso.uiuc.edu> At 09:30 PM 8/31/99 -0700, Timothy J. Wood wrote: [...] > As far as installing the results, I haven't any idea. For my work, I = >only need the actual compiler executable. Which is what? All I want to do is use egcs to try to compile htdig and python since I can't get them to compile with existing compiler. How do I do it with the provided egcs-2? -Ryan >-tim > > > >Begin forwarded message: > > >From: Ryan Thomas = >< > >Date: 1999-08-31 20:13:02 -0700 > >To: Multiple recipients of list < > >Subject: Re: EGCS on MacOS X Server (was Re: Using STL in = >Cocoa apps) > > >Thanks, Tim. I'm now subscribed to PublicSource. Please bear with me. > > >I downloaded said egcs package and ran ./configure and make without any=20= > > >parameters and it seemed to compile just fine. But, what's the next = >step?=20 > >Everything went into subdirs of <, and the makefile seems to = >suggest=20 > >they were supposed to. (?) Why do these Darwin packages never have any=20= > > >INSTALL or README docs? > > >Where can I read about whatever next/rhapsody/mosxs -specific stuff it = >is=20 > >I'm lacking in order to know how to install these Darwin packages? For=20= > > >example, how the heck did you come up with this? > > >RC_OS=3D3Dnextstep RC_ARCHS=3D3Dppc RC_CFLAGS=3D3D'-arch ppc' make = >TARGETS=3D3Dppc=3D > > HOSTS=3D3Dppc build > > >Also, I've not seen DSTROOT, SYMROOT, OBJROOT anywhere but in=20 > >next/rhapsody/mosxs -related installs (often needing to be defined on = >the=20 > >command line, I assume). > > >Where can I learn about these idiosyncracies? > > >Thanks. > > >-Ryan > > >At 03:07 PM 8/27/99 -0700, Timothy J. Wood wrote: > >0000,0000,DEB7> If you read the = >PublicSource@public.lists.apple.com, you might have > >>noticed the thread about this. > >> > >> > >> Wilfredo just put up the real egcs package at =3D > >= >>0000,0000,FFFFhttp://www.publicsource.ap= >ple.com/projects/darwin/source/egcs-2.tar.gz > >> > >> > >> This isn't 2.95.1, but it is close enought that I can open it in =3D > >>FileMerge and not just get an uncomprehensible mess. It's close enough = >=3D > >>that it looks like someone could actually do the merge with 2.95.1 with = >=3D > >>a non-zero chance of success (although hopefully Apple will do this =3D > >>before MacOS X Client ships). > >> > >> > >> I've downloaded it and built it and compiled some code with it = >already =3D > >>(No C++, just ObjC). The command I used to build it was: > >> > >> > >>cd <<< > >> > >>RC_OS=3D3Dnextstep RC_ARCHS=3D3Dppc RC_CFLAGS=3D3D'-arch ppc' make = >TARGETS=3D3Dppc=3D > >> HOSTS=3D3Dppc build > >> > >> > >> If you define RC_OS=3D3Dmacosx, it will end up wanting the pthreads = >=3D > >>stuff that isn't in MacOS X Server. Partway through the build, = >nextstep =3D > >>somehow gets magically changed to macosx_server for the configure =3D > >>scripts. Dunno what is up with still having 'nextstep' in the source =3D= > > >>anywhere, but it seems to work this way. > >> > >> > >>-tim > > > From fischer at fokus.gmd.de Thu Sep 9 01:32:32 1999 From: fischer at fokus.gmd.de (Robert Fischer) Date: Thu Nov 3 14:42:16 2005 Subject: .keyboard file for B&W keyboard Message-ID: <37D770A2.452C02F3@fokus.gmd.de> Hi, am I blind or is there no keyboard layout description file available with MacOSX for configuring the 'standard' USB keyboard? Cheers, Robert -- --- - .-. -- -- --- / \ ---- Robert Fischer .-. / \ --- .-. __o .-. @ / \ / \ / \ _`\<,_ / \ GMD-Fokus / \ / \ / \ (*)/ (*) / `-------------- / `---' `-' `-----------' From Peter.Ochmann at ifs-gmbh.de Thu Sep 9 03:13:49 1999 From: Peter.Ochmann at ifs-gmbh.de (Peter.Ochmann@ifs-gmbh.de) Date: Thu Nov 3 14:42:16 2005 Subject: IB-style window Message-ID: <813D90EEB732D211878800A0244B6D54017449E4@A0200030.ess.ifs-gmbh.de> Hi there, I would like to have a view similar to the IB instances view, i.e. a view with icons and context sensitive menus. How is this done or is there any example application how to do this ? Peter Ochmann From garrison at talosiv.cit.cornell.edu Thu Sep 9 07:55:51 1999 From: garrison at talosiv.cit.cornell.edu (William Garrison) Date: Thu Nov 3 14:42:16 2005 Subject: oddities in my for loop Message-ID: <199909091453.KAA00353@postoffice.mail.cornell.edu> > From: "Greg Hulands" > Date: 1999-09-08 09:19:58 -0400 > To: Multiple recipients of list > Subject: oddities in my for loop > > I am having a strange thing happen to me when using a for loop and an array. > > The code: > > for (i=0; i<[exercises count]-1; i++) //exercises is the array > { > newExercise = [[Exercise alloc] init]; > [newExercise setName: [exercises objectAtIndex:i]]; > [myWorkout addExercise:newExercise]; > [newExercise autorelease]; > } Greg, Have you tried putting [exercises count] in a local variable to see if the behavior changes? int cnt = [exercises count]; for (i=0; i < cnt -1; i++) //exercises is the array { newExercise = [[Exercise alloc] init]; [newExercise setName: [exercises objectAtIndex:i]]; [myWorkout addExercise:newExercise]; [newExercise autorelease]; } Bill -- ---------------------------------- William Garrison, wvg2@cornell.edu CIT/ATS, Cornell University From Paul.Hoehne at ps.net Thu Sep 9 08:23:55 1999 From: Paul.Hoehne at ps.net (Hoehne, Paul) Date: Thu Nov 3 14:42:16 2005 Subject: Using the XML parser Message-ID: <8F31BE142C54D21194BD0008C756AD6C012CA2C3@resexch01.rmf.ps.net> The P-Lists are stored as XML documents so therefore there must be an XML parser at some level which reads those p-lists. Is the parser that reads those XML docs accessible as a Yellow Box class or set of function calls? If it is, what am I missing? I keep trying to track it down with precious little success. perotsystems Paul C. Hoehne (703) 295 1279 Paul.Hoehne@ps.net From infinity at apple.com Thu Sep 9 08:36:10 1999 From: infinity at apple.com (Sam Krishna) Date: Thu Nov 3 14:42:16 2005 Subject: [Fwd: Inaugural DaWG (Darwin and WebObjects Group) Meeting] Message-ID: <37D7D3EA.DFCB7F75@apple.com> All, Apologies for the spam, but I figured if anyone who's in the Triange this weekend and who's interested in (or developing for): EOF WebObjects Mac OS X (or X Server) Darwin Cocoa (formerly known as Yellow Box and OPENSTEP) NEXTSTEP would be interested in this email. Take care, Sam -------------- next part -------------- An embedded message was scrubbed... From: "Michelle P. Malach" Subject: Inaugural DaWG (Darwin and WebObjects Group) Meeting Date: Thu, 09 Sep 1999 11:05:40 -0400 Size: 2569 Url: /mailman/archive/macosx-dev/attachments/19990909/2d67ea89/attachment.mht From ckane at apple.com Thu Sep 9 09:15:29 1999 From: ckane at apple.com (Chris Kane) Date: Thu Nov 3 14:42:16 2005 Subject: Using the XML parser Message-ID: <001701befade$88e5d540$9d0eca11@verbosity.apple.com> The XML parser is in the CoreFoundation framework in Mac OS X DP1, but the header is not public. It is public in DP2, and is a general XML parser. Chris Kane Apple Computer, Inc. -----Original Message----- From: Hoehne, Paul To: Multiple recipients of list Date: Thursday, September 09, 1999 8:31 AM Subject: Using the XML parser >The P-Lists are stored as XML documents so therefore there must be an XML >parser at some level which reads those p-lists. Is the parser that reads >those XML docs accessible as a Yellow Box class or set of function calls? >If it is, what am I missing? I keep trying to track it down with precious >little success. > >perotsystems >Paul C. Hoehne >(703) 295 1279 >Paul.Hoehne@ps.net From infinity at apple.com Thu Sep 9 11:37:52 1999 From: infinity at apple.com (Sam Krishna) Date: Thu Nov 3 14:42:16 2005 Subject: Inaugural DaWG (Darwin and WebObjects Group) Meeting Message-ID: <37D7FE80.C6BD9F68@apple.com> All, It seems like my original posting didn't go through. Here's the original note: (Please direct all replies to Mike Laster - mailto:mlaster@apple.com ) Take care, Sam -------------------------------------------------------------------------- All, Apologies for the spam, but I figured if anyone who's in the Triange this weekend and who's interested in (or developing for): EOF WebObjects Mac OS X (or X Server) Darwin Cocoa (formerly known as Yellow Box and OPENSTEP) NEXTSTEP would be interested in this email. Subject: Inaugural DaWG (Darwin and WebObjects Group) Meeting Date: Thu, 09 Sep 1999 11:05:40 -0400 All, There will be an inaugural meeting for all people interested in participating in the Darwin and WebObjects Group (DaWG) this Saturday, September 11, 3pm at 2nd City Grill, 506 Daniels Street, Raleigh, NC 27605 (919) 856-0336 http://yp.yahoo.com/py/ypMap.py?Pyt=Typ&YY=6505&city=Raleigh&state=NC&country=US&slt=35.7719&sln=-78.6389&cs=5&stx=7737232&stp=y&ad=2500&ycat=&l=7&tuid=2784491&tq=8&btype=default The meeting will address future meetings and future meeting agenda. Please RSVP Mike Laster (mlaster@apple.com) or reply to this mailing if you are interested in attending. If you cannot make this meeting, but are interested in participating in future meetings please reply as well. Michelle Malach Mike Laster David Coyle William Wei From rcfa at cubiculum.com Thu Sep 9 12:10:13 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:16 2005 Subject: Using the XML parser In-Reply-To: <001701befade$88e5d540$9d0eca11@verbosity.apple.com> References: <001701befade$88e5d540$9d0eca11@verbosity.apple.com> Message-ID: <199909091910.PAA05220@wenix.cubiculum.com> you wrote: > The XML parser is in the CoreFoundation framework in Mac OS X DP1, but the > header is not public. It is public in DP2, and is a general XML parser. Will there be ObjC bindings, or will this remain a plain C-based thing? Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From ckane at apple.com Thu Sep 9 12:59:59 1999 From: ckane at apple.com (Chris Kane) Date: Thu Nov 3 14:42:16 2005 Subject: Using the XML parser Message-ID: <003e01befafd$e5aad2c0$9d0eca11@verbosity.apple.com> Ronald C.F. Antony >> The XML parser is in the CoreFoundation framework in Mac OS X DP1, but the >> header is not public. It is public in DP2, and is a general XML parser. > >Will there be ObjC bindings, or will this remain a plain C-based thing? There will be no ObjC binding in DP2, and likely never will be. Chris Kane Apple Computer, Inc. From marcel at system.de Thu Sep 9 14:59:31 1999 From: marcel at system.de (Marcel Weiher) Date: Thu Nov 3 14:42:16 2005 Subject: Using the XML parser Message-ID: > From: "Chris Kane" > > Ronald C.F. Antony > >> The XML parser is in the CoreFoundation framework in Mac OS X DP1, but > the > >> header is not public. It is public in DP2, and is a general XML parser. > > > >Will there be ObjC bindings, or will this remain a plain C-based thing? > > There will be no ObjC binding in DP2, and likely never will be. Anyone else sickened by this? Apple really seems to be going through with their crazy plan. Oh well, I've got an Objective-C XML-parser in development here, currently used for object-archiving and XML-property-list parsing. If anyone is interested in helping out, I could make it available as OpenSource. Marcel From rcfa at cubiculum.com Thu Sep 9 15:55:10 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:16 2005 Subject: Using the XML parser In-Reply-To: <003e01befafd$e5aad2c0$9d0eca11@verbosity.apple.com> References: <003e01befafd$e5aad2c0$9d0eca11@verbosity.apple.com> Message-ID: <199909092255.SAA05745@wenix.cubiculum.com> you wrote: > Ronald C.F. Antony > >> The XML parser is in the CoreFoundation framework in Mac OS X DP1, but > >> the header is not public. It is public in DP2, and is a general XML > >> parser. > > > >Will there be ObjC bindings, or will this remain a plain C-based thing? > > There will be no ObjC binding in DP2, and likely never will be. No personal offense, but this *severely sucks*, particularly the "likely never will be" part. Does Apple really think we lust for Java, C++ and plain C S&M sessions? I'd prefer the cozier relationship I have with ObjC, Scheme and SmallTalk. Apple better stop the talk about "object oriented" and "advanced" in connection with their software, if all that's being done is a regression. Nice hardware really can't cover up the fact that NeXT's direction was much more modern in the software development area than anything Apple has so far come up with. It's pathetic. Please forward this to the people in charge. Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From rcfa at cubiculum.com Thu Sep 9 16:01:04 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909092301.TAA05760@wenix.cubiculum.com> > > > >The XML parser is in the CoreFoundation framework in Mac OS X DP1, but > > > >the header is not public. It is public in DP2, and is a general XML > > > >parser. > > > > > >Will there be ObjC bindings, or will this remain a plain C-based > > >thing? > > > >There will be no ObjC binding in DP2, and likely never will be. > > Anyone else sickened by this? Yes, very much so. > Apple really seems to be going through with their crazy plan. I hope that it will come back to bite them. > Oh well, I've got an Objective-C XML-parser in development here, > currently used for object-archiving and XML-property-list parsing. If > anyone is interested in helping out, I could make it available as > OpenSource. How about working with the GNUStep people on this? It seems that there are people there looking into the same issue, in particular in reference to OSX XML based NIB files. They need such a parser, and there are people who probably can help out, too. Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From mmarkowitz at ceiss.org Thu Sep 9 16:32:06 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: On Thursday, September 09, 1999 3:05 PM, Marcel Weiher [SMTP:marcel@system.de] wrote: > Anyone else sickened by this? Yes and no. My take on this is that it's a little bit of slight of hand. Ask yourself this, is Apple trying to kill AppKit by putting stuff only in CF, or is Apple making a completely new "OS" (API) called CF? I believe it's the later. Apple's in a bit of a tough place. It's clear that many MacOS people are not going to go to Obj-C no matter how good it is (and before people start telling me they're luddites, I've see the same in reverse just as many times), yet they've got all this great code in Foundation and such that would be of tremendous use to all Mac apps, and even Apple itself. So we separate out all the cool code, stick it into CF and now both the Obj-C people and the Mac people can use it. We've already seen the results, new API's coming out of Apple are using CF types. The Keychain uses them for instance, as does the MacOS side of the URL handling stuff. As time goes on I think we'll see this more and more, we'll get language neutral API's using CF types - maybe even newer versions of older API's that use them (please Apple, make QT CF friendly!). Sooner or later CF becomes rich and complete. Foundation and AppKit are wrappers over CF. At that point the language debate goes away. Want to use C? Be our guest! Want to use Obj-C? Hey, great? Java? Fine! I can't see any downside to this. Sure it means smaller numbers of direct updates to the Obj-C world (the NS things), but at the same time it means more and more API's that we can get easy access to via CF. I think the later is more important to us. Of course one problem is that the "original language" colours the API, no matter how hard you try. Consider for example that even a CF based AppKit-like layer is likley to use target-action, making it look weird to Java developers who are used to (uggg) the throw/catch system. I don't know if the Obj-C flavouring is bad when you see it in CF though, as Obj-C specific things like protocols are invisible. Another reason why I like the concept at least is that Obj-C is getting very long in the tooth. Apple is clearly not interested in upgrading Obj-C, so essentially we can expect the libs to remain largely the same. Yet you can't help but wonder how much nicer AppKit would be if it was multiple-inherited rather than protocoled (for instance) and included garbage collection - and CF allows me to move to TOM in theory. > Apple really seems to be going through with their crazy plan. I don't think it's that crazy, layering is good after all. It might not be great for us, but overall I think it's a greater good. Maury From mmarkowitz at ceiss.org Thu Sep 9 16:34:03 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: On Thursday, September 09, 1999 4:16 PM, Ronald C.F. Antony [SMTP:rcfa@cubiculum.com] wrote: > How about working with the GNUStep people on this? Noooooo! Please do NOT put a GPL on anything you write! Maury From marcel at system.de Thu Sep 9 16:42:33 1999 From: marcel at system.de (Marcel Weiher) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: > From: David Young > > > Apple really seems to be going through with their crazy plan. > > And what plan is that? The "let's not give developers a choice but force Java down their throats now matter how loud they scream or what the technical merits of the case are, or just how ridiculously badly it works"-plan. Anyway, another question: I have a patch to the egcs version recently released that makes certain types of meta-programming a lot easier. Anybody know where/how I can submit the patch? Thanks, Marcel From cmh at greendragon.com Thu Sep 9 16:51:25 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: At 3:05 PM -0700 9/9/99, Marcel Weiher wrote: >Anyone else sickened by this? > >Apple really seems to be going through with their crazy plan. Oh, for God's sake. Just GIVE IT UP. Objective-C is nice, but the sky is not going to fall if Apple doesn't implement an Objective-C binding for everything under the sun. Put together an LGPL'd Objective-C binding to Apple's XML parser and I'll bet Apple adopts it for inclusion. They just don't want to spend the effort to maintain such a thing when the primary API to it -- thanks to the vast majority of Mac OS X applications being based on Carbon for a long time to come (probably forever) -- will be C. From cmh at greendragon.com Thu Sep 9 17:02:36 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: At 4:52 PM -0700 9/9/99, Maury Markowitz wrote: > Noooooo! Please do NOT put a GPL on anything you write! Just to clear things up, GNUstep is under the Library GPL. This means it's okay for proprietary software to dynamically-link against it. From marcel at system.de Thu Sep 9 17:18:24 1999 From: marcel at system.de (Marcel Weiher) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: > Sooner or later CF becomes rich and complete. Foundation and AppKit are > wrappers over CF. At that point the language debate goes away. Want to use > C? Be our guest! Want to use Obj-C? Hey, great? Java? Fine! It would be great if that were the case, but the statement that made me sick was that there wouldn't be any Objective-C bindings, meaning either plain C or Java. Neither of these is good enough. Heck, Objective-C isn't *really* good enough, but just barely good enough to get by on. > Another reason why I like the concept at least is that Obj-C is getting > very long in the tooth. In what way, except for Apple not supporting it any longer? > Apple is clearly not interested in upgrading Obj-C, Obj-C doesn't really need much upgrading, and hasn't really needed much in the last ~15 years, because it is a small and sensible design with enough flexibility ( e.g. meta-programmability) that you can get both flexible, high-level interfaces and lightning fast implementations. C++ gives you fast implementation and brittle interfaces. Smalltalk gives you even more flexibility at a slight speed penalty. Java gives you both brittle interfaces and an additional speed penalty. For example, the XML-parser I referred to earlier uses a small plain-C routine at the lowest level, character scanning and progressively more abstract techniques in the higher levels. Since this is Objective-C, the layers are freely mixable, so you can, for example, create very fast partial parsers that only incur the parsing overhead for the parts they are actually interested in. The only problem with Objective-C right now is that Apple is actively trying to kill it, by willfully not making licenses availabe and removing interfaces from the system, despite strong developer opposition (from those that have actually used it). In my feeble mind, that's not a problem with Objective-C, but with Apple, and the result is that if Apple removes Objective-C without providing a better replacement (neither C nor Java fit that bill), it will be Apple being dumped, not Objective-C. Heck, there are now what appear to be very usable Smalltalk implementations for Windows + cross platform versions. If I'm going to pay the VM-speed penalty, I might as well get something for it. That said, I also like the layering aspect of CoreFoundation. It sure can be useful having access to the bare metal at times. Marcel From mmarkowitz at ceiss.org Thu Sep 9 17:59:09 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: On Thursday, September 09, 1999 5:18 PM, Marcel Weiher [SMTP:marcel@system.de] wrote: > It would be great if that were the case, but the statement that made > me sick was that there wouldn't be any Objective-C bindings, meaning > either plain C or Java. No, I think that means plain C. However let us not confuse "won't be" with "can't be". Writing Obj-C wrappers for CF based API is trivial. I don't think this is an issue, this should take you a few minutes to wrap. > Objective-C isn't *really* good enough, but just barely good enough > to get by on. TOM on the other hand is *really* good. In fact it's so good that I think this is where the debate should be turning. TOM includes modules for namespace issues (similar to Java), multiple inheritance (yes, you can do this safely if you have modules), garbage collection, reflection, class vars, etc. It also has the Obj-C syntax, but cleaned up to a degree (some things I don't like). > In what way, except for Apple not supporting it any longer? When Obj-C was being built, a lot of the questions out there weren't yet answered. For instance there was a lot of confusion as to whether or not multiple inheritance could be done safely, and whether or not auto-GB could be done realistically. In the time since then we've learned that both of these can be made to work, and to work well. We've also learned a lot about namespaces, Eiffel introduced programming by contract, etc. In fact most of the languages in use today are old in this same way. C++ isn't even worth talking about, but Java is also fairly old in most ways. Dylan, TOM and Cecil are some of the newer ones. Of all of these, TOM has the advantage of looking and working like Obj-C, and even expanding some of the concepts, like it's version of catagorization, which allows you to even add ivars. You get all sorts of things for free, for instance you get NSCoding without a single line of code (no void *) > Obj-C doesn't really need much upgrading, and hasn't really needed > much in the last ~15 years, because it is a small and sensible design > with enough flexibility ( e.g. meta-programmability) that you can > get both flexible, high-level interfaces and lightning fast > implementations. I'm sorry, but I simply don't agree. Every second you spend tracing down a NSZombie is one wasted second. TOM is better, more powerful, and easier to use. > interfaces. Smalltalk gives you even more flexibility at a slight > speed penalty. Java gives you both brittle interfaces and an > additional speed penalty. You are naming a bunch of old languages. You need to do a little research here I think. > In my feeble mind, that's not a problem with Objective-C, but with > Apple No, it's the market. Apple simply has zero chance of being able to support Obj-C in the Java and Obj-C world. If YOU want to use something better fine, but you can't expect Apple to do it. Maury From rcfa at cubiculum.com Thu Sep 9 19:10:40 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909100210.WAA06234@wenix.cubiculum.com> > TOM on the other hand is *really* good. In fact it's so good that I think > this is where the debate should be turning. TOM includes modules for > namespace issues (similar to Java), multiple inheritance (yes, you can do > this safely if you have modules), garbage collection, reflection, class > vars, etc. It also has the Obj-C syntax, but cleaned up to a degree (some > things I don't like). Sounds interesting. I'm not married to ObjC, I just don't want to regress back to the past. (Links to information on TOM appreciated...) However what chances do you think a language like TOM stands? ObjC is at least 100% ANSI-C compatible, and while the step from procedural to OOP-thinking is large, the step of learning ObjC as a C programmer is small, and equally small to non-existent are the issues involved with linking in legacy code with ObjC projects. Now unless TOM is an ANSI-C extension like ObjC, TOM is going to fight the same up-hill battle as Scheme, T, Dylan, SmallTalk, etc. Out of the barely-useful and better languages, ObjC has the best chances to succeed, because it has both a proven track record, is compatible with legacy code, and it has a dedicated following. Apple is only shooting itself in the foot. Apple offering Java bindings and plain-C based CF, is a great thing, because it offers choice and calms down the programers stuck in the past who get a panic attack when they hear about OOP. But as much as I'd like to see a language better than ObjC, I can't believe that any of the other languages has even a small chance, given the amount of resistance even ObjC runs into. But before this goes completely off topic, a little remark about your comment that "Writing Obj-C wrappers for CF based API is trivial. I don't think this is an issue, this should take you a few minutes to wrap." I'd refute that. Ask the Omni people, if it was just a matter of minutes to write wrappers around AIAT. "a wrapper" and "a well designed, subclassable wrapper with an API that's consistent with the rest of the OpenStep/Cocoa API" are two totally different things. The former may be an issue of "a few minutes", the latter is an engineering effort. Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From wsanchez at apple.com Thu Sep 9 19:59:01 1999 From: wsanchez at apple.com (Wilfredo Sanchez) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909100258.TAA04742@scv3.apple.com> | Anyway, another question: I have a patch to the egcs version | recently released that makes certain types of meta-programming a lot | easier. Anybody know where/how I can submit the patch? To the FSF or on http://www.publicsource.apple.com/modifications.html. I suggest you try the FSF first. -Fred -- Wilfredo Sanchez, wsanchez@apple.com Apple Computer, Inc., Core Operating Systems / BSD Technical Lead, Darwin Project 1 Infinite Loop, 302-4K, Cupertino, CA 95014 From rcfa at cubiculum.com Thu Sep 9 17:55:18 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909100055.UAA06039@wenix.cubiculum.com> you wrote: > On Thursday, September 09, 1999 4:16 PM, Ronald C.F. Antony > [SMTP:rcfa@cubiculum.com] wrote: > > How about working with the GNUStep people on this? > > Noooooo! Please do NOT put a GPL on anything you write! That's another inane comment. First GNUStep is under LGPL and not under GPL, and that's a very relevant distinction. LGPL can be freely used in commercial projects, but it doesn't allow freeloaders to take it, and make it their private property. Either you don't understand the GNU licensing, or you're trying to benefit more from someone else's work than you should be entitled to, given your response... Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From rcfa at cubiculum.com Thu Sep 9 17:52:50 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909100052.UAA06027@wenix.cubiculum.com> > Sooner or later CF becomes rich and complete. Foundation and AppKit are > wrappers over CF. At that point the language debate goes away. Want to use > C? Be our guest! Want to use Obj-C? Hey, great? Java? Fine! There are a few problems with this: CF is only as useful as the abstractions it makes. If it ends up just being yet another cleaned up procedural API, then the improvement over Toolbox/Carbon is marginal to non-existent. Wrappers around procedural code work, as long as the procedural stuff is so basic, that it constitutes the building blocks of everything else. If things become more complex, there will be a clash of software models and/or a problem with such things as subclassability, etc. Worst of all, I could deal with the fact that CF provides building blocks and a procedural API for those eternally stuck in the past, and that higher level OO abstractions are provided by proper ObjC wrappers around that base. We have had similar things in the past e.g. ixKit, EOF adaptors, etc. No problem. *BUT* that requires that these wrappers be made. While Chris Cane's statements may not be "offical Apple words", they are at least an honest assessment as to how he sees the situation as someone who works at that joint. So if he thinks it's not likely that ever there will be ObjC wrappers around these things, then your theory breaks down. If every developer has to spend months wrapping CF API on their own, and if everyone rolls their own more or less well designed wrappers, etc. then all the benefits of a consistent, standardized, well integrated and rich set of OO APIs are right out the window, and we end up with the "job security" mess we have on all other platforms: a plethora of so called object oriented toolkits and APIs in a wild mixture, none designed to work with the other, each employing a different set of conventions, etc. In short, we're back in the programming nightmare I tried to escape by moving to SmallTalk and then NeXT. Very f*cking pleasant to feel trown back in time by about 20 or so years. I start understanding the agony of the people who used to work on Lisp-machines, and the original SmallTalk-80 environment about 20 years ago, who still bitch and moan that in all this time no commercial platform has ever come close to providing an environment of similar consistency, integration and programming efficiency. Rather pathetic for an industry that prides itself to have a product life cycle of only a few months. It might behoove some people here to revisit history for a while. It's also pathetic that an entire industry is dominated by a mass of brainless lemings, and that nobody has any guts anymore to show leadership. If we add up the inefficiency, wasted resources, etc. that are the result of an intentionally delayed development in the software field, then we start to understand the saying "Behind each large fortune hides a large crime" in a new dimension. I don't think giving to charity can offset the damage done in the first place. Sort of like burning down a forest, and then offering to plant a few hundred trees as a tax deduction... > I can't see any downside to this. Sure it means smaller numbers of direct > updates to the Obj-C world (the NS things), but at the same time it means > more and more API's that we can get easy access to via CF. I think the > later is more important to us. I can't say I agree. First, there's no intention to ever create the relevant wrappers, and second, productivity is the main advantage Apple can offer. If I have to use a procedural API, then I can go with Linux or Win* and have a bigger market and a more portable product. NS allows me to achieve my goals in a fraction of the time, it allows me as a single individual to compete with entire teams when it comes to creating custom software. Enter the world of CF, that is history. Besides, what irony: while people were ditching OOP because it's supposedly a performance hog, NeXT succeeded creating an OOP environment and OS with acceptable performance, on hardware, that by modern standards can only be called slow. Now Apple has the fastest boxes around, which could easily handle the little bit of potential OOP overhead, and guess what: we're going back to procedural APIs. I suggest, while Apple is at it, they may want to reconsider their support of m68k Macs... They should be able to handle CF just fine. > Apple is clearly not interested in upgrading Obj-C, > so essentially we can expect the libs to remain largely the same. ObjC doesn't need much upgrading. Other than name spaces, which by convention already exist anyway, the language is fine. GC, the other much requested item, is not a matter of the language, but of the runtime/libraries, as GNUStep versions can do GC quite nicely. > Yet you > can't help but wonder how much nicer AppKit would be if it was > multiple-inherited rather than protocoled That is flame bait. Multiple inheritance is not in ObjC *by design*, not by omission. There is *a lot* of critizism against multiple inheritance, e.g. that it's a bad idea to mix implementation sharing with conceptual relationships, the issue of undefined results on subclasses, categories, etc. Personally, I'd be the first person to oppose adding multiple-inheritance to ObjC. (Note: AFAIK, Java doesn't support that either, for good reason.) > I don't think it's that crazy, layering is good after all. It might not > be great for us, but overall I think it's a greater good. Again: layering requires however that the layers be built, which is exactly what has been denied. Nobody here minds or opposes that the XML parser is a CF framework, but what causes a riot is that there neither are nor won't be any ObjC wrappers. Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From wsanchez at apple.com Thu Sep 9 20:28:47 1999 From: wsanchez at apple.com (Wilfredo Sanchez) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909100328.UAA06655@scv3.apple.com> | That's another inane comment. First GNUStep is under LGPL and not | under GPL, and that's a very relevant distinction. LGPL can be freely | used in commercial projects, but it doesn't allow freeloaders to | take it, and make it their private property. | Either you don't understand the GNU licensing, or you're trying to | benefit more from someone else's work than you should be entitled to, | given your response... Um, I think the GPL is a pain. And I think I do know what I'm talking about, because I deal with licenses and lawyers quite a bit. The LGPL is OK, except when one uses the option to switch to the GPL. If you have a choice, and aren't married to the GPL, I suggest you go with a BSD license, the Artisitc license (Perl), or something a little friendlier than the GPL. -Fred -- Wilfredo Sanchez, wsanchez@apple.com Apple Computer, Inc., Core Operating Systems / BSD Technical Lead, Darwin Project 1 Infinite Loop, 302-4K, Cupertino, CA 95014 From sanguish at digifix.com Thu Sep 9 21:34:11 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909100434.AAA04669@digifix.com> Maury Markowitz wrote: > On Thursday, September 09, 1999 3:05 PM, Marcel Weiher > [SMTP:marcel@system.de] wrote: > > Anyone else sickened by this? > > Yes and no. > > My take on this is that it's a little bit of slight of hand. Ask yourself > this, is Apple trying to kill AppKit by putting stuff only in CF, or is > Apple making a completely new "OS" (API) called CF? I believe it's the > later. > Considering CF is entirely function call based, I hope not. > > So we > separate out all the cool code, stick it into CF and now both the Obj-C > people and the Mac people can use it. > But Apple won't be providing a standard Objective-C API for these calls... which means that passing them around is going to be less than clear, and EVERYONE gets to roll their own. > We've already seen the results, new API's coming out of Apple are using CF > types. The Keychain uses them for instance, as does the MacOS side of the > URL handling stuff. As time goes on I think we'll see this more and more, > we'll get language neutral API's using CF types - maybe even newer versions > of older API's that use them (please Apple, make QT CF friendly!). > > Sooner or later CF becomes rich and complete. Foundation and AppKit are > wrappers over CF. At that point the language debate goes away. Want to use > C? Be our guest! Want to use Obj-C? Hey, great? Java? Fine! > So we can all write in whatever language we want, but in order to provide Object Interfaces to these C-APIs, we have to all roll our own. That is so much the Mac OS Toolbox approach, and 10 years out of date that it sickens me. From sanguish at digifix.com Thu Sep 9 21:37:07 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909100437.AAA04684@digifix.com> Maury Markowitz wrote: > On Thursday, September 09, 1999 5:18 PM, Marcel Weiher > [SMTP:marcel@system.de] wrote: > > It would be great if that were the case, but the statement that made > > me sick was that there wouldn't be any Objective-C bindings, meaning > > either plain C or Java. > > No, I think that means plain C. However let us not confuse "won't be" > with "can't be". Writing Obj-C wrappers for CF based API is trivial. I > don't think this is an issue, this should take you a few minutes to wrap. > All the more reason that we should have OFFICIALLY SUPPORTED APIs for these things. It IS trivial, and yet by Apple providing them, they are ubiquitous.. From theisen at akaMail.com Thu Sep 9 23:17:37 1999 From: theisen at akaMail.com (Dirk Theisen) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <199909100434.AAA04669@digifix.com> Message-ID: <1dxwpcv.w62gri1b8se4gM@[192.168.0.2]> Hi, Scott! > > Sooner or later CF becomes rich and complete. Foundation and AppKit are > > wrappers over CF. At that point the language debate goes away. Want to use > > C? Be our guest! Want to use Obj-C? Hey, great? Java? Fine! > > > So we can all write in whatever language we want, but in order to > provide Object Interfaces to these C-APIs, we have to all roll our own. I agree! But if Apple really doesn't wrap the CF stuff for us (I hope for the opposite), we have to make sure there is an early public release that can be integrated with GNUstep, so we all have access to this and few developers see any reason to roll their own. Regards, Dirk From dyoung at vviuh221.vvi.com Fri Sep 10 04:51:12 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:17 2005 Subject: patch to the egcs version References: Message-ID: <9909101151.AA06404@vviuh221.vvi.com> Marcel, > Anyway, another question: I have a patch to the egcs version > recently released that makes certain types of meta-programming a lot > easier. Anybody know where/how I can submit the patch? Here are some ideas: Sign up to the mail list: gcc@gcc.gnu.org gcc is actually egcs from what I know. post the patch there, see what people say. Also, go to: http://egcs.cygnus.com See the section titled: "Bugs", "Known Bugs" for bugs See the section: "Mailing Lists" to sign up. http://egcs.cygnus.com/lists.html See: http://egcs.cygnus.com/onlinedocs/objc-features_toc.html to see the egcs compiler obj-c features. Cygnus seems like a nice compiler-oriented tools company. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From marcel at system.de Fri Sep 10 04:50:12 1999 From: marcel at system.de (Marcel Weiher) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: > Another reason why I like the concept at least is that Obj-C is getting > very long in the tooth. Apple is clearly not interested in upgrading Obj-C, > so essentially we can expect the libs to remain largely the same. Yet you > can't help but wonder how much nicer AppKit would be if it was > multiple-inherited rather than protocoled (for instance) and included > garbage collection - and CF allows me to move to TOM in theory. Of course, the dynamic nature of Objective-C makes such interfaces *much* easier. For example, an interface from Squeak-Smalltalk to Objective-C took about a day to write and minimal code ( 2-3 C functions, 3 Squeak classes with a total of ~ 15 methods). Tcl was very similar, FScript (a very nice scripting language with Smalltalk-like syntax and some other interesting ideas) is also similar etc. Wrapping C-procedural APIs on the other hand is much more painful because there is no run-time information available (Lawson's GX-wrapper, anyone? :-) Marcel From helge.hess at mdlink.de Fri Sep 10 06:14:18 1999 From: helge.hess at mdlink.de (Helge Hess) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser References: Message-ID: <37D9042A.EBC5CA5F@mdlink.de> Marcel Weiher wrote: > > Another reason why I like the concept at least is that Obj-C is > getting > > very long in the tooth. Apple is clearly not interested in > upgrading Obj-C, > > so essentially we can expect the libs to remain largely the same. > Yet you > > can't help but wonder how much nicer AppKit would be if it was > > multiple-inherited rather than protocoled (for instance) and included > > garbage collection - and CF allows me to move to TOM in theory. Can you give an example what benefits MI would bring in this context ? Regarding GC: GNU Objective-C provides this as well. Regarding NameSpaces: could be easily added to Objective-C as well if Apple would have interest in this. > Wrapping C-procedural APIs on the other hand is much more painful > because there is no run-time information available (Lawson's > GX-wrapper, anyone? :-) I don't know how the CF API looks like, but is it really procedural ? What people seem to miss in this discussion is that Objective-C is after all done in plain C as well. Actually one can use *any* features of Objective-C from plain-C, it just looks different, eg instead of: [a blah] one can use the 'Objective-C C API' and write objc_msgSend(a, sel_getName('blah')) Of course the C API requires more work, but there is actually no functionality loss. What is gained is that this code can be used with any language without problem (eg from C++ without an ObjC++ compiler). So, personally I can imagine that the MacOSX C API is in fact an object system which might be based on the Objective-C one, therefore providing all it's features, including classes, categories and so on. If this is the case I wouldn't have any problem with a C based API. Greetings Helge -- MDlink online service center GmbH http://www.mdlink.de/ From mmarkowitz at ceiss.org Fri Sep 10 09:32:27 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: On Thursday, September 09, 1999 7:17 PM, Ronald C.F. Antony [SMTP:rcfa@cubiculum.com] wrote: > However what chances do you think a language like TOM stands? That all depends on who supports it. If everyone here were to join the TOM mailing list and steer it in the direction of an Obj-C replacement, then I'd say it would have a big chance. > ObjC is at least 100% ANSI-C compatible, and while the step from > procedural to OOP-thinking is large, the step of learning ObjC as a > C programmer is small, and equally small to non-existent are the > issues involved with linking in legacy code with ObjC projects. TOM is essentially a modern Obj-C. It bootstrapped just like Obj-C, albiet on egcs, but now also has it's own native compiler as well. Unlike Obj-C, TOM defines it's own basic types, so full reflection is possible (ie, no void *). Yet because we all work in a c world, TOM allows pure ANSI-C blocks inside it's methods. This is a little bit different than Obj-C where the mix is purely inline, in TOM the c code is blocked in it's own {}'s. The difference is tiny, the long and short of it is that it works the same. If you're at all interested go to http://www.gerbil.org/tom/doc/why.shtml and read the portion on Extensibility. Under TOM you can categorize in new ivars, cvars, add new superclasses to existing classes, REMOVE methods, add methods, and even replace existing methods! This kicks Obj-C's butt. Apple faces a tough fight with OC, essentially it's power is hidden in non-obvious places. Unless you really use the language, it's hard to explain why it's better - and after many years in tech sales on the road, I can say that the basic rule is that if it takes more than 2 minutes to explain, you lose. Obj-C takes more than 2 minutes to explain. Comparing it to Java is even harder, because they both have the same set of buzzwords. Not so with TOM, TOM wins every buzzword war with any language you care to toss at it. Yet again, I don't expect Apple can support it directly. > But before this goes completely off topic, a little remark about > your comment that "Writing Obj-C wrappers for CF based API is trivial. > I don't think this is an issue, this should take you a few minutes to wrap." > I'd refute that. Ask the Omni people, if it was just a matter of minutes > to write wrappers around AIAT. AIAT is not a CF based API, your argument is a non-sequitur. Maury From mmarkowitz at ceiss.org Fri Sep 10 09:03:29 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: On Thursday, September 09, 1999 9:41 PM, Scott Anguish [SMTP:sanguish@digifix.com] wrote: > Maury Markowitz wrote: > > On Thursday, September 09, 1999 3:05 PM, Marcel Weiher > > [SMTP:marcel@system.de] wrote: > > > Anyone else sickened by this? > > > > Yes and no. > > > > My take on this is that it's a little bit of slight of hand. Ask yourself > > this, is Apple trying to kill AppKit by putting stuff only in CF, or is > > Apple making a completely new "OS" (API) called CF? I believe it's the > > later. > > > Considering CF is entirely function call based, I hope not. > > > > > So we > > separate out all the cool code, stick it into CF and now both the Obj-C > > people and the Mac people can use it. > > > But Apple won't be providing a standard Objective-C API for these > calls... which means that passing them around is going to be less than clear, > > and EVERYONE gets to roll their own. > > > We've already seen the results, new API's coming out of Apple are using CF > > types. The Keychain uses them for instance, as does the MacOS side of the > So we can all write in whatever language we want, but in order to > provide Object Interfaces to these C-APIs, we have to all roll our own. > > That is so much the Mac OS Toolbox approach, and 10 years out of > date that it sickens me. Perhaps so, but one thing that has happened in the last 10 years is the open source movement. This proves, conclusively, that a community can come together and provide standardized API - even good ones some times - if there's interest in doing so. So enough bellyaching, we never expected Apple to do all the work for us in the past, and we shouldn't now. This time around we have widespread internet use to help us though. Again, I'll happy act as coordinator on this effort if you all think it's important. But if all you're going to do (not you Scott) is complain that Apple isn't doing it that's not getting anyone anywhere - indeed it will turn into exactly the problem you're talking about. Also, while I'm here, I've got a lot of e-mail asking me about TOM. Go here...http://www.gerbil.org/tom/ And please join the mailing list and make it clear that we're interested in a modern Obj-C replacement, as opposed to an entirely new system (I for instance am not terribly interested in using TOM's X stuff, I'm far more interested in using TOM against Foundation). Maury From cybergrog at bigpond.com Fri Sep 10 06:31:22 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:17 2005 Subject: Setting Up A CVS Server Message-ID: <13282362720369@domain2.bigpond.com> Does anyone know of any good websites or books which talk about setting up a CVS server. Greg From mark at jiiva.com Fri Sep 10 09:52:24 1999 From: mark at jiiva.com (Mark Ericksen) Date: Thu Nov 3 14:42:17 2005 Subject: Future of Apple Technology (Re: Using the XML parser) Message-ID: <199909101646.LAA01150@dfw-ix10.ix.netcom.com> As developers, the future of Apple's technology is of the upmost importance. Sometimes I work myself up into a small rage with Apple and their lack of communication to us about what is *really* going on with Objective-C et al. In reading the many emails on this subject I've been thinking about a "what if" scenario. What if Objective-C was to go away (or at least stop being supported). Where does that leave us developers who have been depending upon it for the last 10 years? Apple came up with the idea of Yellow Box and then Cocoa to replace OPENSTEP. If there is no more Objective-C what language to we program in? According to Apple marketing it will be Java. For now Java sits on top of Objective-C via the bridge. Even if Apple ports all the Objective-C APIs to Java, Java will still be sitting on top of the Quartz graphics library. Now, I could be wrong, but doesn't this smell like a law suit waiting to happen? Didn't Sun sue Microsoft for trying to create a Java that worked best on Windows? Isn't Apple using the Java language and providing the APIs that are written in C? Doesn't this create a platform specific advantage to those who use Java-Cocoa? So if Sun gets upset and sues Apple, where does that leave Cocoa? Will Apple just kill off Cocoa? Apple did say at the Apple Developer's conference in May that the "crown jewel" of Cocoa is the API. So what language would they implement the API in? C++? Go back to Objective-C after burning everyone for years? Or go after a new language? It seems that Apple has another "crown jewel" that they fail to recognize. That is Objective-C. After all when they purchased NeXT they purchased the rights to the language. Perhaps their marketing group can come up with another name for the language and Apple could tout it as the better, faster, natively cross-platform environment. Apple could make a few modifications to the language to give it new life. Look at the marketing success of Java! As people already suggested, Objective-C has a better migration path for legacy code since it is ANSI based, all it needs is a successful marketing campaign! Imagine: Steve Jobs enters the stage and tells the audience, "Java was a great idea. Building truly cross-platform software would have revolutionized the industry. However these goals and dreams were never realized. That is why Apple is announcing today a new development language that is built from the ground up to bring your legacy application into the next millennium and letting new applications take advantage of all the features of modern languages we've come to know and love. Today we give you ! Build on top of ANSI C it allows you to create blazingly fast code while providing a messaging syntax rivaling Smalltalk. bla bla bla..." You get my point. Alright I'm done, back to coding and wondering what I'll be doing next year! Mark Ericksen Jiiva, Inc. From ckane at apple.com Fri Sep 10 09:19:21 1999 From: ckane at apple.com (Chris Kane) Date: Thu Nov 3 14:42:17 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <009d01befba8$3fa04c60$9d0eca11@verbosity.apple.com> Normally when the hornets get stirred up*, I don't bother getting back into the fray, because it's unclear how relevant to actual development continuing the discussion would be, even when it's my fault (as here). But some comments on the role and goal of CoreFoundation may puts its intended use into perspective which may help when deciding whether or not to use it. I may also throw in some comments about the internal development perspective here wrt Cocoa. Let the rambling begin! [* merely descriptive imagery, not meant to be derogatory] CoreFoundation in a C-based API (and implemented in C) with a lot of the "look and feel" of facilities found previously only in the Foundation framework in Cocoa. It was and is currently developed by the same group which maintains the two principal Cocoa frameworks, Foundation and AppKit, and some other Cocoa-related technologies. The primary intent is to broaden the availability of Foundation-like APIs and data types to a wider audience: Mac OS developers and internal providers of other C-based APIs. It uses as many of the principals of OO design as practical: encapsulation/information hiding is used heavily, there is some polymorphism (CF has its own tiny simple runtime), but no subclassing generally. We could have brought the concept of abstract superclasses down to CF to get a sort of subclassibility, and a few modules do allow for something like that, but decided that that would interfere with another primary architectural goal: highest performance. If you were at the CoreFoundation talk at WWDC you heard about this. When you look at the API too, though, you definately see the influence of Mac OS APIs as well; for example, since Mac developers are (we are told) used to not dealing with seeing pointers, the pointers are often embedded in the types (you deal with CFArrayRef rather than CFArray *); or, constants are named kCFBlah, rather than CFBlah. I'm somewhat dubious of all that, but when the alternative was near-zero adoption, we compromised. The CF APIs are very consistent at least, which was another goal. And you get used to it, though it seems strange at first. [Java APIs end up being in a similar position.] CoreFoundation is what I call a "sharp tool"; it isn't safe and friendly like one might consider the Foundation APIs to be, but rather very precise and literal, a tool for intermediate to advanced programmers. For example, it tends to simply do what you tell it to do: if you ask a CFArray to remove the value at index -17, it believes that you know what you are doing, and will remove that element (even though valid indices are always non-negative) and shift all the other elements to close the gap. This may have no observable affect on the program, may corrupt memory causing a crash possibly days later, or may cause immediate program termination. What you've asked for is undefined behavior, so in a sense CF gives it to you. This is a cruel API to thrust on developers who don't type in perfectly correct code (and I myself cannot cast the first stone), so there is also a debug version of the library, which has assertions turned on and assertions all over (though even more would be useful, as is always the case). CoreFoundation also tends to not have all of the convenience APIs of the Foundation in some cases, but in other cases has some that Foundation doesn't have. So, when deciding whether or not to use CoreFoundation, you need to be aware of its sharp nature and decide if you really need the performance (if that's why you think you need to go to it) or use caution if you have to use it because it is the only source of the functionality you need, and use the debug library. Since we also develop Cocoa libraries, we went to some effort to make some types of interaction with Foundation/AppKit APIs convenient. We call this toll-free bridging, where a CFFoo can be used (eg, messaged) like an NSFoo, and an NSFoo can be used where parameters of type CFFoo are expected. They're completely interchangeable. You can even subclass NSFoo and use it like a CFFoo. Not all CF/NS types have a correspondence in the other, and not even all of those are toll-free bridged (sometimes the NSBar contains a CFBar to which messages are forwarded). It's a fair bit of hassle to do this toll-free bridging, but we think it's worth it. > [On not providing an ObjC interface to the XML parser etc.] I should have said "possibly" there will never be an ObjC version of the CFXMLParser API, rather than "likely never". Now I love Objective C more than life itself, and know it better than the back of my hand. But I'm also realistic. There are only a small number of us tasked to maintain and develop CoreFoundation, Foundation, AppKit, etc. (I won't quote a number, but it's small, trust me), so we begin to see additional APIs as additional liability; we have to maintain backwards compatibility with new APIs after all. If we change something in the CF API, then we need to go to another project and change the ObjC version, and another project and change the Java version (possibly). And because of language differences we get differences from the base API, or different constraints (for example "You can't make that change because somebody might have subclassed in ObjC and be doing XYZ." Sometimes we decide to break that anyway, because that is the least of evils, and then the people who are broken are up in arms.) [It be nice if we could "write once, deploy multiple versions" of an API, but autogenerated APIs are usually not great or don't make proper use of the target language or constrain the APIs to least-common-denominators of capabilities. But we haven't figured out how to simplify the maintainence of 3 or more APIs at once yet. We have some tools but they are complex and cumbersome.] I know what the state of the open bugs assigned to me is and how much time I have to devote to bug fixing and feature enhancement in existing APIs and how long bugs tend to wait before being fixed, and what's similarly happens with the other engineers. And I still spend most of my time having to write new code and repair the new code rather than work in the 150K lines I already am responsible for. Providing a cover API is fine as long as you don't expect bugs in it to be fixed for a release or two; but people do want bugs fixed, so providing the new API is not so appealing. Eventually we get around to things (eventually the Foundation came along; eventually we added a document architecture in the AppKit), but it takes a while, which is why I say an ObjC wrapper for CFXMLParser may not happen (or may not happen in a useful timeframe). There are a number of ancillary things that would need to be wrapped, in this case, not just the parser itself, too. We know and hear the "Apple should do it so everybody doesn't have to do it themselves" argument. And XML parsing is a big enough ticket item that an ObjC API may appear. We also hear a lot of request for (and no offense here intended to anyone) dumb or low-bang conveniences which are tied to this argument, that it colors our opinion of that argument (like, "I need a function to sort only every other object in an array and it should take a parameter to allow choosing between the even or the odd indices, and Apple should add this to the frameworks they ship so everybody doesn't have to invent it themselves.") At NeXT we never had any intent of covering all operating system or lower-level functionality with ObjC APIs (though the requests did/do come in). The same still applies. We try to do what will be potentially useful to 80%+ of apps, but we can't do everything nor are we inclined to do so. This gives us, we hope, the biggest bang for the resources we have. Perhaps though it would be a better strategy to tell people what they want to hear rather than honest assesments, and then not be able to deliver. That sort of strategy doesn't seem to have made people any happier [eg. Yellow Box for Windows licensing], though. I've gotten off of the development angle here with some self-pitying raving, for which I apologies to the list. Please decide if replies are appropriate to the discussion of Mac OS X development. Chris Kane Apple Computer, Inc. My opinions are just and merely that. From bentley at crenelle.com Fri Sep 10 11:28:02 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <37D9042A.EBC5CA5F@mdlink.de> References: <37D9042A.EBC5CA5F@mdlink.de> Message-ID: I concur with Ron Antony regarding the choice of direction Apple is apparently taking with regard to non-OO wrappers. Apple has always been driven by the lowest common denominator in their development APIs, with the occasional homage to objects (Open Transport API defines a few C++ classes). The advantage NeXT had that OSX doesn't have is mobility and freedom! NeXT was able to use the experiences of the early 80's to build a late 80's development environment. Apple has a large legacy community that's literally steering the company. It's like a few thousand developers compared to a couple hundred. And there's likely significant top secret work being done for the Java side. The overt Apple Java work seems sound and occasionally cool, but it doesn't get my heart started in the morning. Yet. Apple tried to do something quite clever in OpenDoc by using SOM. The results were quite, hm, indefinite when someone pulled the plug on OD. The whole of OpenDoc seemed ponderous. C++ thing. Is there a commercial need for support for Objective C? Wrappers, libraries, and so on? Can you think of dev products that use ObjC that would sell to legacy Mac OS developers? I asked on the OSX talk list in the same thread (by name), but I'll ask here too, perhaps in a more appropriate form of question: how does one address the same needs, but in Objective C, that C++ addresses with the Standard Template Library (STL)? The STL appears to have become a vital component for current C++ development projects everywhere. The biggest reason, I think, that ObjC is not going anywhere on the legacy Mac OS side is because there's nothing for the legacy Mac developer to grab onto and make work with their stuff. You'd think the upgrade path would be obvious, but for some reason it isn't. Nobody provides cool C enhancements for Metrowerks via Objective C to whet people's appetite. There's no reasonable OO cross library available for legacy Mac OS developers to use to get to Windows. YB could create critical mass by itself, Apple isn't letting that happen. The legacy Mac OS community has to be able to use it too for their stuff. I've been caught short recently, Metrowerks took out the Pascal compiler from CW Pro 5. Lots of the legacy Mac OS code I've been maintaining is at least partly written in Pascal, including the majority of the client app. My new code tends to be entirely C++, but clients still use twelve year old code. What a _mess_. -m From hordur at lausn.is Fri Sep 10 07:10:15 1999 From: hordur at lausn.is (Hordur Thordarson) Date: Thu Nov 3 14:42:17 2005 Subject: Foreign chars in ProjectBuilder Message-ID: <199909101409.OAA20872@gremlin.skyrr.is> Hi, I can?t write some foreign characters in the text editor in Project Builder - PB offers to convert them to ASCII and if so instructed, the chars turn into underscore chars. Of the 20 or so special Icelandic characters (10 upper case and 10 lower), this problem occurs with 3 characters in both upper and lower case. Can anything be done about this ? Thanks, Hordur Thordarson From kcd at jumpgate.com Fri Sep 10 12:00:31 1999 From: kcd at jumpgate.com (Kenneth C. Dyke) Date: Thu Nov 3 14:42:17 2005 Subject: [Off-Topic] Fun C++ quote In-Reply-To: References: Message-ID: <199909101900.MAA09732@babylon5.jumpgate.com> I just thought this might lighten someone's day. "If you think C++ is not overly complicated, just what is a abstract virtual base pure virtual private destructor, and when was the last time you needed one?" -- Tom Cargil in the C++ Journal -Ken --- Kenneth Dyke, kcd@jumpgate.com (personal), kcd@3dfx.com (work) Sr. Software Engineer, Cross Platform Development, 3dfx Interactive C++: The power, elegance and simplicity of a hand grenade. From phr at projectcenter.ch Fri Sep 10 12:02:08 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909101902.AA00448@projectcenter.ch> You wrote: > On Thursday, September 09, 1999 4:16 PM, Ronald C.F. Antony > [SMTP:rcfa@cubiculum.com] wrote: > > How about working with the GNUStep people on this? > Noooooo! Please do NOT put a GPL on anything you write! Please be careful about what you say, GNUstep is LGPL... sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From phr at projectcenter.ch Fri Sep 10 12:02:12 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909101902.AA00453@projectcenter.ch> You wrote: > > There will be no ObjC binding in DP2, and likely never will be. > Anyone else sickened by this? yup...;-( > Apple really seems to be going through with their crazy plan. > > Oh well, I've got an Objective-C XML-parser in development here, > currently used for object-archiving and XML-property-list parsing. If > anyone is interested in helping out, I could make it available as > OpenSource. This would be VERY cool. If you are interested, what about releasing it with GNUstep? sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From sanguish at digifix.com Fri Sep 10 21:06:52 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <199909110406.AAA08458@digifix.com> Maury Markowitz wrote: > > That is so much the Mac OS Toolbox approach, and 10 years out of > > date that it sickens me. > > Perhaps so, but one thing that has happened in the last 10 years is the > open source movement. This proves, conclusively, that a community can come > together and provide standardized API - even good ones some times - if > there's interest in doing so. > Have you noticed just how little of this 'community' is left? Look at the lack of forward momentum the MiscKit has had in the last 2 years. For the longest time we had NO Foundation. MiscKit came along and filled that gap (as best as possible). Then Apple decided to make that a priority, and we got Foundation. But we also had much, much pain along the way to Foundation because of this lack of leadership on (then NeXT's) part. Ask yourself this.. will CF's XML parser have a Java interface? > So enough bellyaching, we never expected Apple to do all the work for us > in the past, and we shouldn't now. Low level objects like this, I think it IS Apple's responsibility to provide. Especially when it consists of virtually no effort on their part aside from sitting down and hashing out the API. The development effort is minimal. Getting a third party solution widely adopted on the other hand, is not. I don't hold any grudge against Chris or Ali for their choice.. just I'd like to see it corrected. :-) From don at misckit.com Fri Sep 10 22:20:35 1999 From: don at misckit.com (Donald A. Yacktman) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <199909110406.AAA08458@digifix.com> References: <199909110406.AAA08458@digifix.com> Message-ID: <9909110520.AA05333@misckit.com> You wrote: > Have you noticed just how little of this 'community' is left? Look > at the lack of forward momentum the MiscKit has had in the last 2 years. That's not _just_ the community. It is partly my fault... ...after a string of several disasters (various key hard drives dying and worse), plus being swamped at work, it has been hard to get the next (very long overdue) releases out. I *do* have enough submitted material to justify both a new NEXTSTEP and a new OPENSTEP/MOSXS release. Now if I hadn't had all these delays, excitement over the kit might not have waned as much as it has. Hopefully someday I'll get my act together...and all that will change. However, there's also truth in what Scott is saying: a lot of the key people who comprised the original set of MiscKit authors have moved on to other pastures and are no longer paying one whit of attention to Apple. Any "revival" of the MiscKit will happen with (for the most part) new blood. (I'd personally like to have the next release include the opening up of a public CVS repository and a bug tracking database. I'm not quite done setting all this up yet...but it is coming, and hopefully with the next release...) -- Later, Don Yacktman don@misckit.com From sanguish at digifix.com Fri Sep 10 23:13:02 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:17 2005 Subject: Adding file extensions to an app on the fly Message-ID: <199909110614.CAA09321@digifix.com> I'm working on an app that has an extendable architecture... adding new bundles would allow you to deal with more file types... I need to be able to add these new file types to the list that the Workspace allows to be opened automatically by the application (which will load the correct bundle at runtime). I'm thinking that maybe the only way to do this that won't require modifying the data in the app wrapper directly is by creating filter services... BUT... some of the types that I'm going to need to deal with can have 'compound' file-extensions and would need to be recognized as being distinct from a file that has only the single trailing extension. As a practical example... I'd want a different bundle to be activated if it was fed a .tar.gz than if it was just a .gz The other concern is how Workspace will interact with the compound files... I suppose I could have just one massive services entry for all the file types, which I could build on the fly and save in the ~/Library/Services when a user launches the app for the first time.. then I could determine the appropriate action to take after my app gets launched.. that way I can put all the granular stuff there... Any comments? Please? From bbum at codefab.com Sat Sep 11 00:13:05 1999 From: bbum at codefab.com (Bill Bumgarner) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909110616.BAA15759@wellington.friday.com> It isn't just that the community fragmented, but also that we haven't had a consistent set of APIs or ***publically accessible*** vision for the future of the OpenStep/Macintosh platform. Note the emphasis there. Consider Darwin; for all intents and purposes, any effort at the kernel/driver layer down is wasted-- Apple has openly discussed that the kernel/driver layers are being replaced in OS X Client and that that work will migrate to Darwin and OS X Server. The same can be said for pretty much the entire line of NeXTSTEP/OpenStep technologies-- they are all in flux right now. It is very hard to motivate oneself to undertake any kind of community project at this point in time-- any effort put forth now will have to be largely revisited in the very near future. This isn't necessarily a bad thing. Actually, all things considered, I would rather be in the situation we are in now than where we were, say, three years ago in the NeXT community OR the Macintosh community. In the NeXT community, we had insanely great programming tools, but absolutely nobody in the marketplace cared.... In the Apple community, we had an insanely great UI and user community, but the development community had been promised Major Changes since '89-- but NOTHING ever came to fruition. Though there have been growing pains... though some of the technologies that are resulting are not exactly what some of us idealists would like to see... the path being taken will very likely result in Apple releasing what could be considered to be a pretty damned impressive achievement: The first commercial operating system with a large suite of productivity applications that is actually stable and quite potentially secure (when set up correctly). For the NeXT community, this means that folks like Andy Stone will be able to ship great applications into a third party market that actually exists. Stone Design -- a 2.5 person software house -- will be able to leverage the OpenStep/Yellow/Cocoa/Quartz technologies to effectively be able to build products that compete directly with Adobe. That's cool! For the Apple community, it means that Macintosh developers will actually be able to develop software on an API and operating system that isn't 15 years old... that they will finally be able to take advantage of some truly killer modern computing technology in an environment that won't crash/freeze/require-a-reboot every time you look at it funny. For the rest of us, it means that we will have a platform that: - runs a plethora of excellent apps, tools, games, etc... - is a consumer system that is *way* more stable than any other commercial platform out there - is incredibly flexible and is running on some very exciting hardware And, most importantly, the technology has a very bright future. OS X Server/Client is a really well designed system from the top to the bottom-- from the user to the system bus. It is modular; that means that bits and pieces can be easily augmented to take advantage of new technologies that haven't even been dreamed of yet! It is flexible; it can play a server role... it can play a client role... it can be a multi-headed monster of a desktop machine... or it can serve as a great environment for lightweight portables. Finally-- and very important-- it leverages standards without being mired in standards; it will leverage XML, PDF, HTML, HTTP, QuickTime, etc... but it is built in a fashion that *none* of these technologies are truly inseparable from the system. And when it comes, I suspect that the developer communities will be back in full force... that once we have the APIs in a form where they are on our primary desktops, there will be a lot of developers quite willing to contribute to various community projects. With the openness of the system, we are going to see a rich array of very interesting projects-- who wouldn't want to build out their cool idea/experiment/research project/thesis system in an environment where all of those standards come for free? b.bum (BTW: No, I didn't forget Linux in the above.... Linux is wonderful-- we use Linux in production systems on a regular basis. I have used and-- will use it again-- as a desktop OS many times. However, Linux is *not* a consumer operating system-- it has come a long, long way, but the human interaction experience is just not there yet. Nor does it have the applications. It'll happen... but it is going to take time. Same with Be-- very cool technologies there, as well. Lots of compelling reasons to use it for certain niche markets. As it stands, it is *not* a consumer desktop environment by any stretch of the imagination) From mark at jiiva.com Sat Sep 11 09:29:39 1999 From: mark at jiiva.com (Mark Ericksen) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909111623.LAA01942@dfw-ix13.ix.netcom.com> > Stone Design -- a 2.5 person software house -- will be able to leverage the > OpenStep/Yellow/Cocoa/Quartz technologies to effectively be able to build > products that compete directly with Adobe. That's cool! So can you explain what "Cocoa" will be in five years from now? This is my biggist hangup. CoreFoundation is written in C. AIAT is written in C++ I believe. Foundation is written in Objective-C with Java wrappers. TextEdit is written in Java (using AppKit) The Finder will be written for Carbon. Will I be able to access an Objective-C API for the Finder or will I have to know C++? Is Objective-C worth programming in if it may go away? There is no clear standard for what technologies we as developers will need to know in order to write a Cocoa application. Yes, for now there is Objective-C and I hope it stays (only because I love it and know it so well). But other Mac programmers may not know Objective-C or care about it so they are screaming to keep their C or C++ libraries and all of Carbon. I worry about the coherency of the various APIs and the ability to write a fully integrated application using all of Apple's technologies. This mybrid of languages in the system may be of concern with those wanting to write applications. Some may say this is a good thing to have the choice. But it is also a larger learning curve for someone not familiar with a particular language, let alone the APIs and design patterns that go along with them. Especially if there are three to four languages, APIs, and patterns to learn just to write one app. I'm not suggesting Apple drop all other languages and go with one, but I've been waiting to start writing a few desktop apps, just wondering what the right language (etc.) will be and how I'll be able to access other system functions and features before I start spending time and money. I bet you that in a few years after MacOS X Personal Edition is out Apple will start dropping one of the languages to simplify things. Any guesses as to which one will be first? Cheers, Mark Ericksen BTW, I really do love the new OS, just need continued communication from Apple leadership... From bentley at crenelle.com Sat Sep 11 10:39:12 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <199909110616.BAA15759@wellington.friday.com> References: <199909110616.BAA15759@wellington.friday.com> Message-ID: To me, a living implementation language has really good libraries, and the language undergoes change. A good implementation language doesn't change that much., but it does appear to change some because requirements change and our knowledge of how things are done improve with time. Objective C doesn't appear to have changed much in the last ten years, which is both good and bad, and perhaps unnatural. It also doesn't seem to have many users, and the only libraries I can find and use appear to belong to Apple. Right now there is no strong case for learning Objective C stated anywhere. One of the things I'd have thought would make sense is to provide some easy migration for C developers to use objects. Remember when C++ was supposed to be an easy migration path for C developers to employ objects? There's no such migration path exploited on the legacy Mac OS side, yet there's a HUGE amount of C code out there. Apple is steering a very overloaded bus filled with legacy Mac OS developers. The burden of the pre-Carbon code base is immense. You can get these folks' attention with working compilers, working libraries (an easier-to-fathom equivalent of STL would go a lonnnng way, by this I mean the classes, not the template mechanism), and some PR. And maybe it is time for Objective C to add a new feature or two; it may not take much at all to get people's attention. I'd have to say that Objective C gathers zero interest because there are no libraries for it where the developers are, but there is tremendous library support for C and C++, some of it nicely cross-platform. Nothing out there takes advantage of the fact that adding way cool stuff to C code is very easy, turns it into Objective C--which is beneficial! (Somehow.) -m From phr at projectcenter.ch Sat Sep 11 11:42:12 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909111842.AA01009@projectcenter.ch> You wrote: > years, which is both good and bad, and perhaps unnatural. It also > doesn't seem to have many users, and the only libraries I can find > and use appear to belong to Apple. Not quite true, GNUstep doesn't belong to Apple and neither the ObjC bindings for gtk. I just mention this to encourage people to have a look at GNUstep. Version 0.6.0 "Dawn" has been released these days! sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From dyoung at vviuh221.vvi.com Sat Sep 11 12:56:25 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser References: Message-ID: <9909111956.AA08245@vviuh221.vvi.com> > I'd have to say that Objective C gathers zero interest because there > are no libraries for it where the developers are, but there is > tremendous library support for C and C++ Many of the c++ libraries out there are redundant and/or superficial. Also, there are obj-c based libraries. Apparently you aren't aware of any of them. Also, you can link any c or c++ lib. into an obj-c based app. So those are suitable consumables for obj-c apps. > So can you explain what "Cocoa" will be in five years from now? Yes, indeed I can: "Five years from now Cocoa will be whatever Apple thinks it wants at the time and what developers will be porting to". You can quote me on that, although the name Cocoa may change. > Objective C doesn't appear to have changed much in the last ten > years Do you use distributed objects with in/out/bycopy argument extensions? Categories? Don't try to rewrite history! Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From cmh at greendragon.com Sat Sep 11 15:20:54 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:17 2005 Subject: Localization In-Reply-To: <199909081633.JAA02507@tarbell.jpl.nasa.gov> References: <199909081633.JAA02507@tarbell.jpl.nasa.gov> Message-ID: At 9:56 AM -0700 9/8/99, Mark Tarbell wrote: >(It would be nice if there were an alternative to this based on a single set >of .nib files. You can programmatically update text fields, buttons, etc., If nibs go XML as Apple's said they will, this should be really easy (as long as they're *real* UTF-8 or UTF-16 XML). Just build some sort of processing tool to auto-generate localized nibs from a string table and a reference nib. From bentley at crenelle.com Sat Sep 11 16:38:56 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <9909111956.AA08245@vviuh221.vvi.com> References: <9909111956.AA08245@vviuh221.vvi.com> Message-ID: >Many of the c++ libraries out there are redundant and/or superficial. Also, >there are obj-c based libraries. Apparently you aren't aware of any of them. I know about GNUStep, which appears to work only with Linux. I think something is needed to help get Mac OS developers interested in using Objective C. Perhaps someone can name some Objective C libraries that can be used to build Mac OS 8 apps? This is part of what I'm asking for. The GNU Objective C Class Library, for example, isn't certified on Mac OS, but aside from that there doesn't seem to be that much in it. >Also, you can link any c or c++ lib. into an obj-c based app. So those are >suitable consumables for obj-c apps. Why would a Mac OS developer do that? > > Objective C doesn't appear to have changed much in the last ten > > years > >Do you use distributed objects with in/out/bycopy argument extensions? >Categories? Don't try to rewrite history! I don't use ObjC for much of anything, much less legacy Mac OS projects. There's little out there for the old OS to work with. Also, even though as you say there have been changes, as far as I can tell Objective C hasn't changed much since 1986, according to the documents I've been able to find so far, among the many orphaned web links. ObjC is not in the legacy MacOS literature... I personally have a desire to be able to work with the language on substantive projects, and I can do that right now on OSXS, but I think that unless the rank and file of Mac OS developerdom actively express an interest, Apple at least seems to emphasize Java, and Metro of course emphasizes C, C++ and Java. Maybe the name Objective C should be changed to Messenger; objective languages are "a dime a dozen" but a message-based language is more unique and cool and helps to explain that "wierd non-C++ like syntax." Also, that way, nobody shoots the messenger. Ah-heh. I just called a Mac developer on the phone and asked him all about Objective C. He didn't know Metrowerks supported it (.m files). He told me the conniptions he had to go through to build his C++ game class library, and we compared notes about the sorts of things he ran into and what the Objective C code would look like. The short of it was that he'd have run into none of the problems. He has an open mind, but he's not going to switch to an unsupported language, he has enough trouble with a supported one. Clearly there's a bit of an education gap here... >Thanks A Bunch! David Young; VVI-DCS >dyoung@vvi.com -m From phr at projectcenter.ch Sat Sep 11 17:49:40 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909120049.AA00330@projectcenter.ch> You wrote: > >Many of the c++ libraries out there are redundant and/or superficial. Also, > >there are obj-c based libraries. Apparently you aren't aware of any of them. > > I know about GNUStep, which appears to work only with Linux. I think > something is needed to help get Mac OS developers interested in using > Objective C. Nope, GNUstep runs on most Unix flavours, eg. FreeBSD, Solaris, Linux, ... Windows will follow, the foundation already works on it. sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From richard at brainstorm.co.uk Sat Sep 11 22:43:54 1999 From: richard at brainstorm.co.uk (Richard Frith-Macdonald) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909120543.AA01103@tiptree.brainstorm.co.uk> On Sat, 11 Sep 1999 16:42:19 -0700 (PDT), bentley@crenelle.com wrote: > >Many of the c++ libraries out there are redundant and/or superficial. Also, > >there are obj-c based libraries. Apparently you aren't aware of any of them. > > I know about GNUStep, which appears to work only with Linux. Wrong - not just GNU/Linux. Also BSD, SunOS/Solaris, Unixware ... Hardware architecture is more of an issue than operating system - it only works fully on ix86, sparc, powerpc and alpha processors as far as I know. I use it on various flavours of GNU/Linux, Unixware-2.1.3, and used it on SunOS until our sparcstation was stolen. From bentley at crenelle.com Sat Sep 11 23:39:31 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <9909120543.AA01103@tiptree.brainstorm.co.uk> References: <9909120543.AA01103@tiptree.brainstorm.co.uk> Message-ID: >On Sat, 11 Sep 1999 16:42:19 -0700 (PDT), bentley@crenelle.com wrote: > > >Many of the c++ libraries out there are redundant and/or >superficial. Also, > > >there are obj-c based libraries. Apparently you aren't aware of >any of them. > > > > I know about GNUStep, which appears to work only with Linux. > >Wrong - not just GNU/Linux. >Also BSD, SunOS/Solaris, Unixware ... > >Hardware architecture is more of an issue than operating system - it only >works fully on ix86, sparc, powerpc and alpha processors as far as I know. > >I use it on various flavours of GNU/Linux, Unixware-2.1.3, and used it on >SunOS until our sparcstation was stolen. How much of the support classes will work under Mac OS 8 or IX (heh)? -m From richard at brainstorm.co.uk Sun Sep 12 01:26:52 1999 From: richard at brainstorm.co.uk (Richard Frith-Macdonald) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909120826.AA01180@tiptree.brainstorm.co.uk> On Sat, 11 Sep 1999 23:41:39 -0700 (PDT), bentley@crenelle.com wrote: > >On Sat, 11 Sep 1999 16:42:19 -0700 (PDT), bentley@crenelle.com wrote: > > > >Many of the c++ libraries out there are redundant and/or > >superficial. Also, > > > >there are obj-c based libraries. Apparently you aren't aware of > >any of them. > > > > > > I know about GNUStep, which appears to work only with Linux. > > > >Wrong - not just GNU/Linux. > >Also BSD, SunOS/Solaris, Unixware ... > > > >Hardware architecture is more of an issue than operating system - it only > >works fully on ix86, sparc, powerpc and alpha processors as far as I know. > > > >I use it on various flavours of GNU/Linux, Unixware-2.1.3, and used it on > >SunOS until our sparcstation was stolen. > > How much of the support classes will work under Mac OS 8 or IX (heh)? I'd image that most of the basic foundation stuff would work with minimal or no porting effort - NSArray, NSDictionary, NSData, NSSet, NSString, NSAttributedString, NSCoder, NSArchiver, NSUnarchiver, NSValue, NSNumber, NSNotification... etc. NSInvocation and distributed objects would not work. Forget the gui library - it's currently X-windows based. It's true though, that most of the stuff that actually interfaces with the O/S would need substantial porting work for MacOS (pre MacOS-X). When I said that O/S wasn't the major issue - I guess I meant that in the context of a 'real' O/S :-) ... anything that is something like POSIX compliant. That rules out early MacOS and all Windoze (though windoze-nt is close enough that the GNUstep base library will mostly run on it, and could probably work quite nicely if anyone wanted to spend a few weeks work on it). From phr at projectcenter.ch Sun Sep 12 03:08:06 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: References: Message-ID: <9909121008.AA00323@projectcenter.ch> You wrote: > How much of the support classes will work under Mac OS 8 or IX (heh)? I don't want to sound rude at all, but Mac OS 8.x and also 9 is 'dead technology' like Windows9x. Yes I know, those systems will still be used for some time and some parts will be included in X, but when we talk about the future, then they are not important anymore. Just my $0.02... sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From tdwoodward at radiodigest.com Sun Sep 12 10:27:28 1999 From: tdwoodward at radiodigest.com (Todd Daniel Woodward) Date: Thu Nov 3 14:42:17 2005 Subject: Compiling 2.2S4 on Mac OS X Server? illegal cast error in ident.c In-Reply-To: <199909062149.OAA26375@ircache.net> Message-ID: Back in April, Mike Laster was so kind to modify Squid 2.1P2 to work with Mac OS X Server , and it works better and much much faster than 1.1.20. (I can't find any contact info for Mr. Laster however.) Here's his diff: Only in squid-2.1.PATCH2-rhapsody/auth_modules: dummy diff -cr squid-2.1.PATCH2/cfgaux/config.guess squid-2.1.PATCH2-rhapsody/cfgaux/config.guess *** squid-2.1.PATCH2/cfgaux/config.guess Mon Aug 3 18:24:47 1998 --- squid-2.1.PATCH2-rhapsody/cfgaux/config.guess Sat Apr 17 12:36:55 1999 *************** *** 555,560 **** --- 555,563 ---- news*:NEWS-OS:[56].*:*) echo mips-sony-newsos${UNAME_RELEASE} exit 0 ;; + *:Rhapsody:*:*) + echo `arch`-apple-rhapsody${UNAME_RELEASE} + exit 0;; esac #echo '(No uname command or uname output not recognized.)' 1>&2 diff -cr squid-2.1.PATCH2/cfgaux/config.sub squid-2.1.PATCH2-rhapsody/cfgaux/config.sub *** squid-2.1.PATCH2/cfgaux/config.sub Wed Apr 8 01:11:33 1998 --- squid-2.1.PATCH2-rhapsody/cfgaux/config.sub Sat Apr 17 12:38:08 1999 *************** *** 753,758 **** --- 753,761 ---- -xenix) os=-xenix ;; + -rhapsody*) + os=-rhapsody + ;; -none) ;; *) diff -cr squid-2.1.PATCH2/include/config.h.in squid-2.1.PATCH2-rhapsody/include/config.h.in *** squid-2.1.PATCH2/include/config.h.in Mon Aug 3 17:00:08 1998 --- squid-2.1.PATCH2-rhapsody/include/config.h.in Sat Apr 17 13:02:23 1999 *************** *** 94,99 **** --- 94,102 ---- #elif defined(__CYGWIN32__) || defined(WIN32) || defined(WINNT) || defined(__WIN32__) || defined(__WIN32) #define _SQUID_MSWIN_ + #elif defined(__APPLE__) + #define _SQUID_APPLE_ + #endif #if !defined(CACHE_HTTP_PORT) diff -cr squid-2.1.PATCH2/src/ftp.c squid-2.1.PATCH2-rhapsody/src/ftp.c *** squid-2.1.PATCH2/src/ftp.c Mon Nov 16 12:22:46 1998 --- squid-2.1.PATCH2-rhapsody/src/ftp.c Sat Apr 17 13:05:52 1999 *************** *** 1787,1793 **** --- 1787,1797 ---- comm_close(ftpState->data.fd); debug(9, 3) ("ftpAcceptDataConnection: Connected data socket on FD %d\n", fd); ftpState->data.fd = fd; + #ifdef _SQUID_APPLE_ + ftpState->data.port = ntohs((u_short)peer.sin_port); + #else ftpState->data.port = ntohs(peer.sin_port); + #endif ftpState->data.host = xstrdup(inet_ntoa(peer.sin_addr)); commSetTimeout(ftpState->data.fd, Config.Timeout.read, ftpTimeout, ftpState); diff -cr squid-2.1.PATCH2/src/squid.h squid-2.1.PATCH2-rhapsody/src/squid.h *** squid-2.1.PATCH2/src/squid.h Fri Oct 30 14:48:16 1998 --- squid-2.1.PATCH2-rhapsody/src/squid.h Sat Apr 17 13:33:59 1999 *************** *** 51,56 **** --- 51,63 ---- #define CHANGE_FD_SETSIZE 0 #endif + /* Trying to redefine CHANGE_FD_SETSIZE causes a slew of warnings on + Mac OS X Server */ + #if defined(_SQUID_APPLE_) + #undef CHANGE_FD_SETSIZE + #define CHANGE_FD_SETSIZE 0 + #endif + /* Cannot increase FD_SETSIZE on FreeBSD before 2.2.0, causes select(2) * to return EINVAL. */ /* Marian Durkovic */ *************** *** 64,72 **** --- 71,81 ---- #endif /* Increase FD_SETSIZE if SQUID_MAXFD is bigger */ + #ifndef _SQUID_APPLE_ #if CHANGE_FD_SETSIZE && SQUID_MAXFD > DEFAULT_FD_SETSIZE #define FD_SETSIZE SQUID_MAXFD #endif + #endif #if PURIFY /* disable assert() under purify */ *************** *** 346,352 **** --- 355,365 ---- #include "globals.h" #include "util.h" + /* Mac OS X Server already has radix.h as a standard header, so this causes + conflicts */ + #ifndef _SQUID_APPLE_ #include "radix.h" + #endif #if !HAVE_TEMPNAM #include "tempnam.h" I tried to implement most of these changes in 2.2S4, but unfortunately I get the following error during make all: cc -g -O2 -Wall -I. -I../include -I../include -c ident.c ident.c:200: illegal cast, missing `)' after `peer' make[1]: *** [ident.o] Error 1 make: *** [all] Error 1 Anyone have any insight into tweaking Squid 2.2S4 enough to compile correctly under Mac OS X Server? Thanks, Todd Daniel Woodward Technical Consultant RadioDigest.com Inc. http://radiodigest.com "Radio you can READ!" From wessels at ircache.net Sun Sep 12 10:28:05 1999 From: wessels at ircache.net (Duane Wessels) Date: Thu Nov 3 14:42:17 2005 Subject: Compiling 2.2S4 on Mac OS X Server? illegal cast error in ident.c In-Reply-To: Message-ID: On Tue, 7 Sep 1999, Todd Daniel Woodward wrote: > + #ifdef _SQUID_APPLE_ > + ftpState->data.port = ntohs((u_short)peer.sin_port); > + #else > ftpState->data.port = ntohs(peer.sin_port); > + #endif How is 'sockaddr_in' defined on your system? On Unixes it's like: struct sockaddr_in { u_char sin_len; u_char sin_family; u_short sin_port; struct in_addr sin_addr; char sin_zero[8]; }; So, sin_port is already of type u_short and the cast is not needed. Hopefully your system also defines sin_port as short. It would really suck to have to insert casts for Apple everywhere sin_port is used. > I tried to implement most of these changes in 2.2S4, but > unfortunately I get the following error during make all: > > cc -g -O2 -Wall -I. -I../include -I../include -c ident.c > ident.c:200: illegal cast, missing `)' after `peer' > make[1]: *** [ident.o] Error 1 > make: *** [all] Error 1 Weird. Sounds like maybe the preprocessor is getting confused? Can you run 'cpp' by hand? % cpp -I. -I../include ident.c > foo Duane W. From mark at jiiva.com Sun Sep 12 13:20:20 1999 From: mark at jiiva.com (Mark Ericksen) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909122014.PAA21420@dfw-ix15.ix.netcom.com> >From a business point of view, if it took six months to get your technology running on either Mac OS 8 or Win 9x it would be worth it. You instantly have millions of potential customers while the MacOS X market is in its infancy. Users of 'dead technology' are still paying customers and should not be ignored. Your profit from these customers should pay for your effort, support, AND the development of products used in the latest (but smaller user base) technologies. Mark Ericksen Jiiva, Inc. > You wrote: > > How much of the support classes will work under Mac OS 8 or IX (heh)? > > I don't want to sound rude at all, but Mac OS 8.x and also 9 is 'dead technology' like Windows9x. > Yes I know, those systems will still be used for some time and some parts will be included in X, > but when we talk about the future, then they are not important anymore. > > Just my $0.02... > > sweet dreams, Phil > --- > Philippe C.D. Robert > FreeBSD/Linux Hacker > http://www.nice.ch/~phip > > From ericchan at mindless.com Sun Sep 12 13:49:17 1999 From: ericchan at mindless.com (Eric Chan) Date: Thu Nov 3 14:42:17 2005 Subject: Setting Up A CVS Server Message-ID: <1274978240-26396926@Istar.Stanford.EDU> The following is a good place to start: http://www.gnu.org/manual/cvs/html_mono/cvs.html ec ---------- >From: "Greg Hulands" >To: Multiple recipients of list >Subject: Setting Up A CVS Server >Date: Fri, Sep 10, 1999, 8:58 PM > > Does anyone know of any good websites or books which talk about setting up a > CVS server. > Greg > From phr at projectcenter.ch Sun Sep 12 14:27:20 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <199909122014.PAA21420@dfw-ix15.ix.netcom.com> References: <199909122014.PAA21420@dfw-ix15.ix.netcom.com> Message-ID: <9909122127.AA00575@projectcenter.ch> You wrote: > >From a business point of view, if it took six months to get your technology > running on either Mac OS 8 or Win 9x it would be worth it. You instantly have > millions of potential customers while the MacOS X market is in its infancy. > Users of 'dead technology' are still paying customers and should not be > ignored. Your profit from these customers should pay for your effort, support, > AND the development of products used in the latest (but smaller user base) > technologies. You are completely right, that's why I have said: > > but when we talk about the future, then they are not important anymore. Here I thought we talk about API questions concerning the next few years, not the past! To use/maintain elder Mac OS resources there is Carbon, at least this is what Apple told everyone, I am not talking about Carbon, and I never will have a look at it (I hope). If Apple chooses not to push ObjC and Cocoa (and Java...sigh) as their modern OO API, what will they offer us? A mix of C, C++, Embedded C++ and Java? This can not be their intention... There was once a system, completely object oriented, well designed with a great UI, it was called OPENSTEP... ...and now we are back at the roots: Carbon everywhere, CF as a plain C API, the IOKit using Embedded C++, Quartz instead of DPS.... IMHO Apple and its insanely great PR staff COULD have made everyone believe that ObjC is the great new language of the Millenium (why not renaming it to iAva...:), Steve Jobs COULD have made everyone want to develop for Cocoa, they just did not want to see this happen!!! Instead they tell use to use Java (in conjunction with Cocoa) and feed us with pure C/Embedded C++ APIs... BTW how many *new* native Cocoa/Java developers are really out there?! In this point IMHO they just failed. The point of Java is to be crossplattform and not to be dependant of any vendor extensions or hardware - did someone say crossplattform here....?! And now I am going to get some sleep...grin sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From bentley at crenelle.com Sun Sep 12 20:18:15 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <9909122127.AA00575@projectcenter.ch> References: <9909122127.AA00575@projectcenter.ch> Message-ID: > >IMHO Apple and its insanely great PR staff COULD have made everyone >believe that ObjC is the >great new language of the Millenium (why not renaming it to >iAva...:), Steve Jobs COULD have made >everyone want to develop for Cocoa, they just did not want to see >this happen!!! Instead they tell >use to use Java (in conjunction with Cocoa) and feed us with pure >C/Embedded C++ APIs... >BTW how many *new* native Cocoa/Java developers are really out >there?! In this point IMHO they >just failed. The point of Java is to be crossplattform and not to be >dependant of any vendor >extensions or hardware - did someone say crossplattform here....?! > > >And now I am going to get some sleep...grin > >sweet dreams, Phil >--- >Philippe C.D. Robert From my conversations with rank and file legacy Mac OS developers, the jury hasn't convened for Mac OS X or Server. There's some bright folks out there that dabble in languages and libraries, and so they're rarin' to go, but they appear to be a minority. Most of the rest are roll-up-the sleeves folks who have a goal and a means, and OSX isn't ready yet, tho Carbon beckons. My take is that the propaganda machine knows that doing the hard sell right now would just plain bounce, for the reasons I am alluding to in this thread, nothing for Mac OS legacy developers to play with much less develop commercial product with, that can somehow dock with the ObjC world when they drop the OSX gangplanks. The propaganda machine is focused on selling IX in large quantities. The allure of the web is much stronger than Mac OS X. There are some takers of Mac OS X Server, but there are many ways to host web sites, and WebObjects appears to be very pricey. Yesterday, one developer said, "50 hits a minute is "free", but a 100 hits a minute is $1500. What the heck is that? And how about that unlimited license for $25K? So while they might be drawn toward the Java WO and maybe a little toward ObjC, the price tag seems a little wacky for a CGI. I could convince people of the efficacy of the system, but I have to do my own development and don't have time for that kind of stuff. While they'll be attracted to Cocoa, they're going to be trying to attack it from the legacy angle, using Carbon. They'll anticipate a large hump between OS IX and OS X. It sort of looks like ObjC is going to come up very short in the Mac community without some sort of outreach, and I don't think that's going to substantially happen without real working dev versions of OS X. OSXS is much too flaky. It is too late to do anything about ObjC libraries for Metrowerks. -m From owolf at advancenet.net Sun Sep 12 22:19:30 1999 From: owolf at advancenet.net (Scott Johnson) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909130520.AAA14088@babba.advancenet.net> >I don't want to sound rude at all, but Mac OS 8.x and also 9 is 'dead >technology' like Windows9x. >Yes I know, those systems will still be used for some time and some parts >will be included in X, >but when we talk about the future, then they are not important anymore. > >Just my $0.02... > >sweet dreams, Phil BFD, to paraphrase Galbraith, in the long run *all* operating systems are 'dead technology'. Advocacy belongs elsewhere. Scott Johnson From hordur at lausn.is Mon Sep 13 06:40:14 1999 From: hordur at lausn.is (Hordur Thordarson) Date: Thu Nov 3 14:42:17 2005 Subject: Problem with foreign characters in PB Message-ID: <199909131339.NAA24550@gremlin.skyrr.is> A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 483 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990913/b71fc4f2/attachment.bin From mmarkowitz at ceiss.org Mon Sep 13 09:20:09 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: On Saturday, September 11, 1999 12:19 AM, Bill Bumgarner [SMTP:bbum@codefab.com] wrote: > It isn't just that the community fragmented, but also that we haven't had a > consistent set of APIs or ***publically accessible*** vision for the future > of the OpenStep/Macintosh platform. Indeed, I couldn't agree more. Most of the teeth knashing going on is because no one really has an idea what Apple's plans are, so we're all left guessing based on current moves. I tend to be pretty happy-go-lucky (ask my girlfriend) in most things in life and I tend to interpret Apple's moves as overall positive although it's going to take a lot longer to get where I want the OS to be. However the majority of people in this forum appear to have been hit about the ears one too many times by NeXT/Apple and form different opinions than I do - vive la differance. That said the number of assumptions that "apple is stupid" makes my stomach ache. What really makes my head spin is when the argument is that Apple's C libs (ie, CF) will only be good if they're well designed - as well designed as Foundation. How it is that the exact people who wrote Foundation in the first place could apparently lose so many IQ points is a mystery. The long and short of it is that Apple has done a tremendously poor job of telling us what the plan is. Thus many are led to believe there is none. That's a bad thing, and Apple needs to address it. Having some people who don't like your plan is far far better than having a lot of people who don't know what the plan is. > The first commercial operating system with a large suite of productivity > applications that is actually stable and quite potentially secure (when set > up correctly). The issue from an OS standpoint is a toughy though, Unix simply isn't very good for users, and Apple really does have a LOT of work to do in terms of wrapping some of the knobbly bits before this thing is good for consumers. For instance, the current split in the NetInfo tools into two _really_ needs to be addressed. Frankly these matters are much more important than Cocoa base development in the short term, and anyone that suggests otherwise is dreaming. Maury From mmarkowitz at ceiss.org Mon Sep 13 09:22:03 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:17 2005 Subject: Adding file extensions to an app on the fly Message-ID: On Friday, September 10, 1999 11:20 PM, Scott Anguish [SMTP:sanguish@digifix.com] wrote: > I'm working on an app that has an extendable architecture... adding > new bundles would allow you to deal with more file types... > > I need to be able to add these new file types to the list that the > Workspace allows to be opened automatically by the application (which > will load the correct bundle at runtime). I'm having the _exact_ same problem. I plan on offloading G2's filters out from the base app and into bundles, but as you discovered the NSDocument system simply doesn't deal with this case - at least not in any public API I've seen. Maury From leigh at tomandandy.com Mon Sep 13 09:45:44 1999 From: leigh at tomandandy.com (Leigh Smith) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909131645.MAA12351@ernst.tomandandy.com> As my .sig attests, I'm certainly not advocating a Microsoft approach, but their approach of pushing a dynamic object system (COM) as the means for programmers to interact with different libraries independent of development languages seems an appropriate approach, effectively scaling the issue from one of language adoption to protocol adoption. Obviously Apple's distributed objects runs rings around COM (especially with CORBA support). So to deal with the real world of programmer inertia (coding in antiquated languages like C++), DO support across *all* languages (that Apple currently supports: C, C++, ObjC, Java) would mean people aren't forced to adopt any particular language, but that the growing body of evidence would lead to Mac developers eventually adopting more dynamic languages as it would be simply easier to interact with the DOs. That would weigh in favour of ObjC, but not rule out other languages which are appropriate (including programmer retraining costs) for coding particular objects. -- Leigh Smith leigh@tomandandy.com (MIME) tomandandy +1-212-334-0421 (W) +1-212-334-0422 (F) 89 Greene St. New York, NY 10012, USA http://www.cs.uwa.edu.au/~leigh Microsoft - What do you want to re-install today? From karl at nfox.com Mon Sep 13 10:09:24 1999 From: karl at nfox.com (Karl Kraft) Date: Thu Nov 3 14:42:17 2005 Subject: Setting Up A CVS Server References: <13282362720369@domain2.bigpond.com> Message-ID: <00d801befe0a$bb3826e0$0329040a@nfox.com> www.codefab.com - OSX specific items. CVSWRAPPERS is there. Also, a mailing list working on a CVS replacement. www.omnigroup.com - has part of the online docs www.sente.ch - Has a GUI for CVS on OSX. Has several links to useful web sites about cvs. There are also a couple of newsgroups, comp.software.config-mgmt comes to mind ----- Original Message ----- From: Greg Hulands To: Multiple recipients of list Sent: Friday, September 10, 1999 8:56 PM Subject: Setting Up A CVS Server Does anyone know of any good websites or books which talk about setting up a CVS server. Greg From mpelzsherman at yahoo.com Mon Sep 13 13:41:10 1999 From: mpelzsherman at yahoo.com (Michael Pelz Sherman) Date: Thu Nov 3 14:42:17 2005 Subject: VSS/ProjectBuilder Woes Message-ID: <19990913204110.19895.rocketmail@web601.yahoomail.com> Folks: Among the many joys of consulting is the opportunity to learn inferior products that you might never chose to use otherwise. :-) My current client has run into some problems with Microsoft's Visual Source Safe (surprise!), some of which may be related to the need to upgrade to Visual Studio service pack 3, others which are probably "pilot error", although we've found that VSS makes it very easy to commit such errors, especially in combination with our favorite IDE, ProjectBuilder. :-) The situation that seems most common is the following: ProjectBuilder allows editing of files marked "Read Only" on NT. (It does warn you first, but will let you save changes to the files without changing their read-only status.) So developer (let's call him "Homer") does a "Get Latest Version" in VSS and start making changes to read-only files using ProjectBuilder. OK - two things VSS probably doesn't like here: Homey is editing read-only files, which VSS assumes he can't do, and he hasn't checked out those files, so VSS assumes he shouldn't be changing them! Now Homer decides to check in his changes. First VSS says he has to check OUT the files he's modified. Homer checks out the files, but forgets to check the "Don't get local copy" box. "D'OH!" Without asking or warning, VSS will now replace the files Homer has been editing with the repository versions. It doesn't make a backup copy of the files either; all of Homer's hard work is now completely toasted. This problem has cost our team several hours of lost work! Questions: 1) Is this behavior normal for VSS? 2) If so, is there any way of setting up VSS to always create backups of files it replaced in your work area (like CVS does)? 3) If not, do you think the upgrade to SP3 will do anything to fix this problem? Thanks much! Michael Pelz-Sherman __________________________________________________ Do You Yahoo!? Bid and sell for free at http://auctions.yahoo.com From mdj at guru.localdomain Mon Sep 13 20:24:07 1999 From: mdj at guru.localdomain (Matt(hew) Jenkins) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser In-Reply-To: <199909092301.TAA05760@wenix.cubiculum.com>; from Ronald C.F. Antony on Thu, Sep 09, 1999 at 07:01:04PM -0400 References: <199909092301.TAA05760@wenix.cubiculum.com> Message-ID: <19990914132407.A22604@one.net.au> On Thu, Sep 09, 1999 at 07:01:04PM -0400, Ronald C.F. Antony wrote: > > >There will be no ObjC binding in DP2, and likely never will be. > > > > Anyone else sickened by this? > > Yes, very much so. Probably not too bad since it's likely to be a minimal parser intended just for serialization stuff, and as such needs to be nice and fast. Other uses of XML will probably require a 'full' parser. > > Oh well, I've got an Objective-C XML-parser in development here, > > currently used for object-archiving and XML-property-list parsing. If > > anyone is interested in helping out, I could make it available as > > OpenSource. > > How about working with the GNUStep people on this? It seems that there > are people there looking into the same issue, in particular in reference > to OSX XML based NIB files. They need such a parser, and there are people > who probably can help out, too. Well I've been working on a XML 'kit' for GNUstep for a while now.. I did an initial implementation a while ago which built straight to a tree and made excessive use of Foundation Objects to do it's job. As such, it was a bit on the slow side. I've since been rewriting my parser to not use NSStrings until the last minute, and also to be used via the SAX interface that many Java based parsers use, although I've had to change the SAX interface somewhat since Foundation doesn't include any general purpose stream classes. If anyone wants to collaborate with me on this, that's fine, although keep in mind I intend to donate the entire work to the GNUstep project (presuming the GNUstep project wants it :) ) and as such the copyright will be assigned to the Free Software Foundation.. Cheers, Matt From fischer at fokus.gmd.de Tue Sep 14 01:49:49 1999 From: fischer at fokus.gmd.de (Robert Fischer) Date: Thu Nov 3 14:42:17 2005 Subject: Pasteboard communications failure ...? Message-ID: <199909140849.KAA26464@mailhub.fokus.gmd.de> Hi, I still have problems with my pasteboard in conjunction with NSPerformService(). Maybe anyone has encountered this error before and may give me a hint, what I could try eventually. What I do is the following: pb = [NSPasteboard generalPasteboard]; [pb declareTypes:[NSArray arrayWithObject:NSStringPboardType] owner:nil]; fileContent = [NSString stringWithContentsOfFile:filename]; [pb setString:fileContent forType:NSStringPboardType]; NSPerformService( @"TextEdit/Open Selection", pb ); This works fine for ONE time. After that I got a 'Bad return value' or finally a 'Pasteboard communications failure during dataShouldBeSet:, code -3'. I could'nt even work with PB afterwards, because Copy/Paste is corrupted. Please help! Thanks, Robert -- -- --- - .-. -- -- --- / \ ---- Robert Fischer .-. / \ --- .-. __o .-. @ / \ / \ / \ _`\<,_ / \ GMD-Fokus / \ / \ / \ (*)/ (*) / `-------------- / `---' `-' `-----------' From cybergrog at bigpond.com Tue Sep 14 04:13:10 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:17 2005 Subject: CVS and CVL Message-ID: <11484626115752@domain0.bigpond.com> I am new to cvs and would like to know some things about it. I created a cvsroot directory at /usr/cvsroot. I then imported an already created project called GymBuddy with the following command: cvs import -m "Imported Source for GymBuddy" Development/Personal/GymBuddy greg start I then checked out the modules file and edited it by adding a line: GymBuddy -a Development/Personal/GymBuddy I then used CVL.app to then checkout the project, it listed it in the modules by name dialog but when trying to checkout the project CVL returns an error : cvs checkout: cannot find module 'Development/Personal/GymBuddy' - ignored. I then removed the alias (-a) in the modules file and tried again but a different error occured: cvs [checkout aborted]: there is no repository /usr/cvsroot/Development/Personal/GymBuddy Questions: 1. What is the relationship between the import command and how it associates itself with modules? 2. How do I fix the problem. Any help is greatly appreciated. Thanks in advance, Greg Hulands. From thomas.roehl at cai.com Tue Sep 14 08:20:02 1999 From: thomas.roehl at cai.com (Roehl, Thomas) Date: Thu Nov 3 14:42:17 2005 Subject: Mac OS X Client on G4 Message-ID: <4C094419B0E6D2119FF100805F85CABE012DDD52@usilmse1.cai.com> Has anyone tried to install the MacOS X Client preview on a G4 machine yet? Tom Roehl From andreas at mevis.de Tue Sep 14 08:29:47 1999 From: andreas at mevis.de (Andreas Bohne) Date: Thu Nov 3 14:42:17 2005 Subject: Additional connection inspector Message-ID: <9909141729.AA470041@matisse> Hi all, some time ago I wrote an additional (to the usual) connection inspector for notification handling. I made it a separate entry in the inspector window popup (just after the "old" connection inspector). This works, but gives you some odd behaviour (like the need to switch between the inspectors every time you draw a connection). Now I would like to behave it like the EOAssociation inspector - with a popup inside the inspector for selecting the old inspector(s) or my own ones. Unfortunately, searching is disabled in the Omnigroup dev mailing lists now, so I couldn't do a search there - sure this have been asked earlier. If anyone has done this or has some hints how to do this I would appreciate the help. TIA, Andreas --------------------------------------------------------------------- Andreas Bohne MeVis - Center for Medical Diagnostic Systems and Visualization e-mail: andreas@mevis.de http://www.mevis.de/ --------------------------------------------------------------------- From tdwoodward at radiodigest.com Tue Sep 14 10:39:46 1999 From: tdwoodward at radiodigest.com (Todd Daniel Woodward) Date: Thu Nov 3 14:42:17 2005 Subject: Compiling EGCS from Apple? Message-ID: Anyone have luck compiling EGCS from Apple's Publicsource? http://www.publicsource.apple.com/projects/darwin/source/egcs-2.tar.gz ./configure isn't complete and complains that there's no config-guess or config-in If you just do make and install, it compiles without a problem, but doesn't install outside of the source directory. Anyone have any tips to share with a compiling newbie? Thanks, Todd Daniel Woodward Technical Consultant RadioDigest.com Inc. http://radiodigest.com "Radio you can READ!" From kgb at yale.graduate.net Tue Sep 14 13:27:44 1999 From: kgb at yale.graduate.net (John Kuszewski) Date: Thu Nov 3 14:42:17 2005 Subject: EGCS, Java, and Altivec Message-ID: <199909142026.PAA00729@yale.graduate.net> Hi, I remember hearing at WWDC that Apple would be moving to EGCS for the MacOS X compiler. That has me wondering about a few issues: 1. EGCS includes a native Java compiler, GCJ (from Cygnus). Will Apple support using GCJ to compile Cocoa apps written in Java? 2. How will Altivec optimization be handled? Will there be an automatic vectorizer, or will we have to scatter pragmas throughout our source? I'm curious, because I do some heavy-duty scientific computation, and I'd like to write my next program in Java and have it run screamingly fast on a G4. Thanks, --John Kuszewski -- John Kuszewski, Genius cleric@yale.graduate.net From bentley at crenelle.com Tue Sep 14 14:44:35 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: EGCS, Java, and Altivec In-Reply-To: <199909142026.PAA00729@yale.graduate.net> References: <199909142026.PAA00729@yale.graduate.net> Message-ID: >Hi, > >I remember hearing at WWDC that Apple would be moving to EGCS for >the MacOS X compiler. That has me wondering about a few issues: > >1. EGCS includes a native Java compiler, GCJ (from Cygnus). Will >Apple support using GCJ >to compile Cocoa apps written in Java? > >2. How will Altivec optimization be handled? Will there be an >automatic vectorizer, or will we >have to scatter pragmas throughout our source? > >I'm curious, because I do some heavy-duty scientific computation, >and I'd like to write my >next program in Java and have it run screamingly fast on a G4. > >Thanks, > >--John Kuszewski What's the current policy regarding so-called fast floating point in Java? It was my last understanding that Java numerics must follow the exact same number system, that the results on one machine must exactly match the results on another machine, down to the last decimal place for float computations, otherwise you don't get to that lofty goal of write once run everywhere. -m From pdetina at comnetix.com Tue Sep 14 14:52:39 1999 From: pdetina at comnetix.com (Paolo DeTina) Date: Thu Nov 3 14:42:17 2005 Subject: OLE and Openstep Message-ID: <3.0.5.32.19990914175239.00902db0@comnetix.com> Hello all, I am trying to use OLE automation. My client is an OPenstep application. The OLE automation server I am trying to work with wants the client to register a callback with it providing a window handle (WHND) and a message id. My question is how do I grab/trap the event the OLE server will eventually send back. Being relatively new to OPenstep I do not know how to make the connection between the eventual event and a method in one of my classes. Paolo --------------------------------- Paolo De Tina ComnetiX Computer Systems Inc. pdetina@comnetix.com (905)829-9988 x272 http://www.comnetix.com From pierce at twinforces.com Tue Sep 14 14:52:32 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:17 2005 Subject: EGCS, Java, and Altivec In-Reply-To: <199909142026.PAA00729@yale.graduate.net> References: <199909142026.PAA00729@yale.graduate.net> Message-ID: At 2:22 PM -0700 9/14/99, John Kuszewski wrote: >Hi, > >I remember hearing at WWDC that Apple would be moving to EGCS for >the MacOS X compiler. That has me wondering about a few issues: > >1. EGCS includes a native Java compiler, GCJ (from Cygnus). Will >Apple support using GCJ >to compile Cocoa apps written in Java? That would be cool. > >2. How will Altivec optimization be handled? Will there be an >automatic vectorizer, or will we >have to scatter pragmas throughout our source? Generally, the Altivec stuff is accessed through things that look like function calls, but aren't. > >I'm curious, because I do some heavy-duty scientific computation, >and I'd like to write my >next program in Java and have it run screamingly fast on a G4. You'll probably have to do it in C, or at least do the computation in C. PIerce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From toon at omnigroup.com Tue Sep 14 15:07:04 1999 From: toon at omnigroup.com (Greg Titus) Date: Thu Nov 3 14:42:17 2005 Subject: Pasteboard communications failure ...? Message-ID: <199909142207.PAA25525@scyther.omnigroup.com> Hi Robert, I don't have any idea why this is, but apparently reusing pasteboards that you use for services is a bad idea. Just use +[NSPasteboard pasteboardWithUniqueName] instead of the general pasteboard, and you should be fine. Hope this helps, --Greg From troy.stephens at page44.com Tue Sep 14 15:37:40 1999 From: troy.stephens at page44.com (Troy Stephens) Date: Thu Nov 3 14:42:17 2005 Subject: VSS/ProjectBuilder Woes Message-ID: Michael Pelz-Sherman wrote: > 1) Is this behavior normal for VSS? Unfortunately, yes. > 2) If so, is there any way of setting up VSS to always create backups of > files it replaced in your work area (like CVS does)? Not as far as I know. > 3) If not, do you think the upgrade to SP3 will do anything to fix this > problem? Again, not AFAIK... In my experience, the most reliable policy is to get in the habit of *always* checking out _any_ file before you make changes to it. (Even if the changes you're about to make are temporary and you don't intend to check them in, you may care enough about them to turn red when you absent-mindedly do a "Get" and lose them!) You can always Undo Checkout later, and this way you get a warning/reminder when you do a Get, that prompts you for what to do for each file that's checked out to you (usually you want to Merge their changes in with yours). If you have multiple checkout disabled at your site (we have it enabled here), this advice may not work as well for you, since your checking something out as a simple precaution will block someone else from working on it. (My suggestion is: Enable multiple checkout -- just be cautious when you do a merged Get, especially when some time has passed and the files in SourceSafe have changed significantly out from under you. I always back up my local source files before doing a merged Get, if my changes are non-trivial. SourceSafe's merge isn't always what you want, especially when things get complicated...) Incidentally, I've noticed that SourceSafe's diff's syntax coloring has some problems with my source files -- probably because ProjectBuilder uses the UNIX newline convention when saving. If, for instance, I add a message to a class @interface declaration, SourceSafe's diff will recognize that something changed and bring up the side-by-side difference window, but displays the added line without highlighting it appropriately. I wish I could set SS up to OS/E's FileMerge.app! It's so much better! Anyway, hope this helps, and good luck! Troy Stephens Page 44 Studios troy.stephens@page44.com From rasmussn at soren.acl.lanl.gov Tue Sep 14 14:43:15 1999 From: rasmussn at soren.acl.lanl.gov (Craig E Rasmussen) Date: Thu Nov 3 14:42:17 2005 Subject: EGCS, Java, and Altivec References: Message-ID: <199909142143.PAA02687@soren.acl.lanl.gov> >I'm curious, because I do some heavy-duty scientific computation, >and I'd like to write my >next program in Java and have it run screamingly fast on a G4. Those interested in scientific computation using Java should look into the Java Grande Forum (www.javagrande.org). There you will find a discussion on what it takes to make numerically intensive applications run fast. The issues have more to do with the language itself, rather than the specific processor (i.e., G4/Altivec) that the Java VM is running on. Craig Rasmussen From dyoung at vviuh221.vvi.com Tue Sep 14 16:21:13 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:17 2005 Subject: EGCS, Java, and Altivec References: <199909142026.PAA00729@yale.graduate.net> Message-ID: <9909142321.AA13136@vviuh221.vvi.com> John Kuszewski, Genius, > Will there be an automatic vectorizer You mean like a compiler for the CRAY -- a real super computer? I don't think egcs has an automatic vectorizer for any language. > How will Altivec optimization be handled? Essentially function calls like a DSP chip embedded programming from what I know. > because I do some heavy-duty scientific computation, and I'd like to > write my next program in Java Hum a mix of automatic vectorizer feature and Java features? This seems bizarre to nauseating. I think you will end up programming in ANSI C and wrapping in your object language of choice. > and have it run screamingly fast on a G4 If you know how to unroll loops correctly and all that. Fast but not screamingly. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From bungi at omnigroup.com Tue Sep 14 16:47:13 1999 From: bungi at omnigroup.com (Timothy J. Wood) Date: Thu Nov 3 14:42:17 2005 Subject: Using the XML parser Message-ID: <199909142347.QAA25362@electabuzz.omnigroup.com> >Well I've been working on a XML 'kit' for GNUstep for a while now.. I did an >initial implementation a while ago which built straight to a tree and made >excessive use of Foundation Objects to do it's job. As such, it was a bit on >the slow side. I've since been rewriting my parser to not use NSStrings until >the last minute, and also to be used via the SAX interface that many Java based >parsers use, although I've had to change the SAX interface somewhat since >Foundation doesn't include any general purpose stream classes. >If anyone wants to collaborate with me on this, that's fine, although keep in >mind I intend to donate the entire work to the GNUstep project (presuming the >GNUstep project wants it :) ) and as such the copyright will be assigned to the >Free Software Foundation.. I wrote a flex scanner for XML and a bison parser. The part that is sort of interesting here is that the flex scanner handles UTF-8 (and correctly rejects any invalid byte sequences). It is VERY fast at scanning and parsing the XML that I've fed it (I gave it this big blob of Shakespeare plays in XML -- 250,000 lines I think and it scanned/parsed in under 2 seconds on a 400Mhz G3... this is from memory so I may be off a bit). It doesn't handle some of the stuff that it would need to (like entity replacement), but if you are interesting in seeing what I have, just let me know. -tim From niemann at mvp.net Tue Sep 14 19:09:38 1999 From: niemann at mvp.net (Jim Niemann) Date: Thu Nov 3 14:42:17 2005 Subject: Mac OS X Client on G4 In-Reply-To: <4C094419B0E6D2119FF100805F85CABE012DDD52@usilmse1.cai.com> References: <4C094419B0E6D2119FF100805F85CABE012DDD52@usilmse1.cai.com> Message-ID: >Has anyone tried to install the MacOS X Client preview on a G4 machine yet? > >Tom Roehl I was unsuccessful attempting to install either Mac OS X Client and Mac OS X Server on my new G4/400. Does anyone know how soon Apple will release versions that work on a G4? Jim Niemann From fischer at fokus.gmd.de Tue Sep 14 23:43:42 1999 From: fischer at fokus.gmd.de (Robert Fischer) Date: Thu Nov 3 14:42:17 2005 Subject: Pasteboard communications failure ...? References: <199909142207.PAA25525@scyther.omnigroup.com> Message-ID: <37DF401E.65BFB707@fokus.gmd.de> Greg Titus wrote: > > Hi Robert, > > I don't have any idea why this is, but apparently reusing > pasteboards that you use for services is a bad idea. Just use > +[NSPasteboard pasteboardWithUniqueName] instead of the general > pasteboard, and you should be fine. Thank you! Maybe I was a little bit too avaricious! Now I create a new unique pasteboard every time I need one and everything works fine. Robert -- --- - .-. -- -- --- / \ ---- Robert Fischer .-. / \ --- .-. __o .-. @ / \ / \ / \ _`\<,_ / \ GMD-Fokus / \ / \ / \ (*)/ (*) / `-------------- / `---' `-' `-----------' From bentley at crenelle.com Wed Sep 15 08:05:07 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:17 2005 Subject: EGCS, Java, and Altivec In-Reply-To: <9909142321.AA13136@vviuh221.vvi.com> References: <9909142321.AA13136@vviuh221.vvi.com> Message-ID: On the legacy side, the Metrowerks C/C++ compilers support Velocity Engine. -m From mailagent at clickto.com Wed Sep 15 09:24:17 1999 From: mailagent at clickto.com (Clickto Network) Date: Thu Nov 3 14:42:17 2005 Subject: Mac OS X Client on G4 In-Reply-To: Message-ID: I think Nov 1. > -----Original Message----- > From: macosx-dev@omnigroup.com [mailto:macosx-dev@omnigroup.com]On > Behalf Of Jim Niemann > Sent: Tuesday, September 14, 1999 10:17 PM > To: Multiple recipients of list > Subject: Re: Mac OS X Client on G4 > > > >Has anyone tried to install the MacOS X Client preview on a G4 > machine yet? > > > >Tom Roehl > > I was unsuccessful attempting to install either Mac OS X Client and > Mac OS X Server on my new G4/400. Does anyone know how soon Apple > will release versions that work on a G4? > > Jim Niemann > From etienne at univ-lr.fr Wed Sep 15 10:56:15 1999 From: etienne at univ-lr.fr (Etienne Gourdon) Date: Thu Nov 3 14:42:17 2005 Subject: cross compilation Windows on Mac OS Message-ID: <199909151756.TAA15653@cri.univ-lr.fr> A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 3115 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990915/103c1c5a/attachment.bin From bbum at codefab.com Wed Sep 15 11:09:39 1999 From: bbum at codefab.com (Bill Bumgarner) Date: Thu Nov 3 14:42:17 2005 Subject: cross compilation Windows on Mac OS Message-ID: <199909151809.OAA12262@bjork.codefab.com> A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 4475 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990915/cd4e80fa/attachment.bin From bbum at codefab.com Wed Sep 15 12:24:25 1999 From: bbum at codefab.com (Bill Bumgarner) Date: Thu Nov 3 14:42:17 2005 Subject: cross compilation Windows on Mac OS Message-ID: <199909151924.PAA12703@bjork.codefab.com> Mac OS X Server == Rhapsody. Rhapsody DR1 (DR2, too? Can't remember) was shipped on Intel.... And, again, whether or not Apple has OS X Server Intel-- I don't know, haven't been disclosed on anything, and have no idea-- it seems awfully fishy that so many third party developers are actually expending effort on producing Intel binaries for various applications for a platform that doesn't exist. That the targets still remain in PB or the shlibs or anything else Intel still ships with OS X Server-- all things that could be easily removed-- seems kinda fishy, as well. Of course, this is all complete conjecture based on observations-- none on any kind of official information-- and is likely so far away from reality that those in the know are laughing at us about now. It is of interest to developers to note that it really is very trivial to port Yellow applications between Yellow NT, Cocoa PPC and-- if you had a machine to run it on-- Cocoa Intel. "It just works". So, whatever Apple might decide to do with future systems, at least getting there will be relatively painless. And, as you very succintly put it, the marketing and support from Apple behind OS X Server and WinNT Yellow Box were very, very different.... b.bum From mnowak at umich.edu Wed Sep 15 13:04:15 1999 From: mnowak at umich.edu (Mike Nowak) Date: Thu Nov 3 14:42:17 2005 Subject: Running Java applications under Mac OS X Server? Message-ID: <199909152004.QAA05186@purgatory.healthmedia.umich.edu.> I am having trouble running Java programs that I have created with the Project Builder. I create a new project as a Java Package and do a build and I end up with .zip file. When I double click it, OpenUp conveniently unpacks it. If I use the inspector, I can change the default to JavaRun but when I double-click it, JavaRun runs but there is no app. If I try to run "java SwingApplication" (the sample in the Swing tutorial) in Terminal, I get "class SwingApplication" not found. The CLASSPATH variable was not set so I did "setenv CLASSPATH .:`javaconfig DefaultClasspath`" as suggested in the WebObjects documentation. "." is included in the CLASSPATH as is ~/Library/Java but having the .zip file in either of those directories doesn't seem to matter. If I generate .class files, either by decompressing the .zip file or generating the .class files with javac or using the .class files in the .build directory, then double-clicking on the .class file in the Workspace Manager or running "java SwingApplication" works. I apologize if this is obvious, but I'm used to the Metrowerks environment. How can I run the Java application from the .zip file? Thanks in advance for your help. -- Mike Nowak Instant Messenger: mykkro The University of Michigan mnowak@umich.edu Health Media Research Lab http://www-personal.umich.edu/~mnowak/ ............................................................................ "Every gun that is made, every warship launched, every rocket fired signifies a theft from those who hunger and are not fed, those who are cold and are not clothed. The world is in arms not spending money alone. It is spending the sweat of its laborers, the genius of its scientists, the hopes of its children." -- President Eisenhower, 1953. From pierce at twinforces.com Wed Sep 15 13:30:47 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:17 2005 Subject: cross compilation Windows on Mac OS In-Reply-To: <199909151924.PAA12703@bjork.codefab.com> References: <199909151924.PAA12703@bjork.codefab.com> Message-ID: At 12:31 PM -0700 9/15/99, Bill Bumgarner wrote: >Mac OS X Server == Rhapsody. Rhapsody DR1 (DR2, too? Can't remember) was >shipped on Intel.... And, again, whether or not Apple has OS X >Server Intel-- I >don't know, haven't been disclosed on anything, and have no idea-- it seems >awfully fishy that so many third party developers are actually >expending effort >on producing Intel binaries for various applications for a platform that >doesn't exist. That the targets still remain in PB or the shlibs >or anything >else Intel still ships with OS X Server-- all things that could be easily >removed-- seems kinda fishy, as well. > >Of course, this is all complete conjecture based on observations-- >none on any >kind of official information-- and is likely so far away from reality that >those in the know are laughing at us about now. > >It is of interest to developers to note that it really is very >trivial to port >Yellow applications between Yellow NT, Cocoa PPC and-- if you had a >machine to >run it on-- Cocoa Intel. "It just works". So, whatever Apple might decide >to do with future systems, at least getting there will be relatively painless. > >And, as you very succintly put it, the marketing and support from >Apple behind >OS X Server and WinNT Yellow Box were very, very different.... > >b.bum Apple has basically stated here and there that part of the problem with YellowBox for Intel were licensing issues with _other_ companies. I.e. there's stuff they don't own, and therefore can't release, and the other companies don't choose to be reasonable. For instance, perhaps Adobe wants $150/pop for display postscript... I suspect that somewhere in the Darwin 0.5 timeframe, we'll see some sort of release that has "stuff missing for Intel goes " along with a document listing what needs to be done. At that point, it will be up to the open source community. In other words, if Apple COULD release Yellow Box for Intel, they would! That's why there's all that stuff left in ProjectBuilder and such. In fact, wasn't there a YellowBox SDK for Intel on my MOSXS CD? Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From pierce at twinforces.com Wed Sep 15 13:35:36 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:17 2005 Subject: Running Java applications under Mac OS X Server? In-Reply-To: <199909152004.QAA05186@purgatory.healthmedia.umich.edu.> References: <199909152004.QAA05186@purgatory.healthmedia.umich.edu.> Message-ID: At 1:11 PM -0700 9/15/99, Mike Nowak wrote: >I am having trouble running Java programs that I have created with the >Project Builder. I create a new project as a Java Package and do a build and >I end up with .zip file. When I double click it, OpenUp conveniently unpacks >it. If I use the inspector, I can change the default to JavaRun but when I >double-click it, JavaRun runs but there is no app. > >If I try to run "java SwingApplication" (the sample in the Swing tutorial) >in Terminal, I get "class SwingApplication" not found. The CLASSPATH >variable was not set so I did "setenv CLASSPATH .:`javaconfig >DefaultClasspath`" as suggested in the WebObjects documentation. "." is >included in the CLASSPATH as is ~/Library/Java but having the .zip file in >either of those directories doesn't seem to matter. > >If I generate .class files, either by decompressing the .zip file or >generating the .class files with javac or using the .class files in the >.build directory, then double-clicking on the .class file in the Workspace >Manager or running "java SwingApplication" works. > >I apologize if this is obvious, but I'm used to the Metrowerks environment. >How can I run the Java application from the .zip file? Here is the script I wrote to launch jcvs: setenv CLASSPATH ~/Library/Java:/Local/Library/Java:/System/Library/Java:/Network/Libra ry/Java:/System/Library/Frameworks/JavaVM.framework/Classes/classes.ja r:/System/Library/Frameworks/JavaVM.framework/Classes/awt.jar:/System/ Library/Frameworks/JavaVM.framework/Classes/swingall.jar java -classpath "./jcvsii.jar:$CLASSPATH" com.ice.jcvsii.JCVS --------- the setenv is necessary because CLASSPATH never seems to be set. the -classpath option tells java to look in jcvsii.jar in the current directory, and then to run com.ice.jcvsii.JCVS So, all you need is probably: java -classpath "./myapp.zip:$CLASSPATH" yourMainClass where myapp.zip is your zip file, and "yourMainClass" is the name of the class that holds your main procedure. Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From mnowak at umich.edu Wed Sep 15 13:48:51 1999 From: mnowak at umich.edu (Mike Nowak) Date: Thu Nov 3 14:42:17 2005 Subject: Running Java applications under Mac OS X Server? Message-ID: <199909152049.QAA05319@purgatory.healthmedia.umich.edu.> > So, all you need is probably: > > java -classpath "./myapp.zip:$CLASSPATH" yourMainClass > > where myapp.zip is your zip file, and "yourMainClass" is the name of > the class that holds your main procedure. > > Pierce Thanks, that worked. I guess I don't understand why I need to include the .zip file explicitly if the directory is in the class path. But thanks again. -- Mike Nowak Instant Messenger: mykkro The University of Michigan mnowak@umich.edu Health Media Research Lab http://www-personal.umich.edu/~mnowak/ ............................................................................ "Every gun that is made, every warship launched, every rocket fired signifies a theft from those who hunger and are not fed, those who are cold and are not clothed. The world is in arms not spending money alone. It is spending the sweat of its laborers, the genius of its scientists, the hopes of its children." -- President Eisenhower, 1953. From pierce at twinforces.com Wed Sep 15 13:59:43 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:17 2005 Subject: Running Java applications under Mac OS X Server? In-Reply-To: <199909152049.QAA05319@purgatory.healthmedia.umich.edu.> References: <199909152049.QAA05319@purgatory.healthmedia.umich.edu.> Message-ID: At 4:48 PM -0400 9/15/99, Mike Nowak wrote: > > So, all you need is probably: > > > > java -classpath "./myapp.zip:$CLASSPATH" yourMainClass > > > > where myapp.zip is your zip file, and "yourMainClass" is the name of > > the class that holds your main procedure. > > > > Pierce > >Thanks, that worked. I guess I don't understand why I need to include the >.zip file explicitly if the directory is in the class path. But thanks >again. Because you can name your zip ANYTHING, so you have to tell Java what you named it... Plus its supposed to be a .jar file anyways... Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From pcoskren at bbn.com Wed Sep 15 14:04:31 1999 From: pcoskren at bbn.com (Patrick Coskren) Date: Thu Nov 3 14:42:17 2005 Subject: Running Java applications under Mac OS X Server? Message-ID: <199909152105.RAA05455@po2.bbn.com> > Thanks, that worked. I guess I don't understand why I need to include the > .zip file explicitly if the directory is in the class path. But thanks > again. Basically because the CLASSPATH is a list of your class archives. Those are either directories that contain the whole com/mystuff/classname nonsense, or they're zip files: the two archive types that Sun supports. CLASSPATH is _not_ a list of directories that could contain such archives. The 'PATH' in the name might be a little misleading, I'll admit. Maybe a legacy from before ZIP files? But it wouldn't be the first time Sun gave something a lousy name. I'm still trying to figure out why a "GridBagLayout" is called that. :-) -Patrick From pdetina at comnetix.com Wed Sep 15 14:20:59 1999 From: pdetina at comnetix.com (Paolo DeTina) Date: Thu Nov 3 14:42:17 2005 Subject: OLE and Openstep p2 Message-ID: <3.0.5.32.19990915172059.0093e270@comnetix.com> Hello, I should have mentioned that I am dealing with Openstep on NT, but I guess most of you figured that out. > >I am trying to use OLE automation. My client is an OPenstep application. >The OLE automation server I am trying to work with wants the client to >register a callback with it providing a window handle (WHND) and a message >id. My question is how do I grab/trap the event the OLE server will >eventually send back. Being relatively new to OPenstep I do not know how to >make the connection between the eventual event and a method in one of my >classes. Any ideas/suggestions would be appreciated. --------------------------------- Paolo De Tina ComnetiX Computer Systems Inc. pdetina@comnetix.com (905)829-9988 x272 http://www.comnetix.com From oloft at cs.chalmers.se Wed Sep 15 17:17:10 1999 From: oloft at cs.chalmers.se (Olof Torgersson) Date: Thu Nov 3 14:42:17 2005 Subject: Windows and Frameworks problem Message-ID: <9909152217.AA102613@metis> I have Windows/NT SP5 and WebObjects 4.0 When I try to run an application that uses frameworks I have written I get errors like the following (gdb) Starting program: D:/YBBuilds/TestFrameworks/TestFrameworks.app/TestFrameworks.exe 34020000:d:\Apple\Library\Executables\AppKit.dll 32010000:d:\Apple\Library\Executables\Foundation.dll 31000000:d:\Apple\Library\Executables\nextpdo.dll [failed reading symbols from DLL] "C:\WINNT\system32\MSVCRT.dll": can't read symbols: File format not recognized. 77f00000:C:\WINNT\system32\KERNEL32.dll 77e70000:C:\WINNT\system32\USER32.dll 77ed0000:C:\WINNT\system32\GDI32.dll 77dc0000:C:\WINNT\system32\ADVAPI32.dll 77e10000:C:\WINNT\system32\RPCRT4.dll 77c40000:C:\WINNT\system32\SHELL32.dll [failed reading symbols from DLL] "C:\WINNT\system32\COMCTL32.dll": can't read symbols: File format not recognized. 776d0000:C:\WINNT\system32\WSOCK32.dll 776b0000:C:\WINNT\system32\WS2_32.dll 776a0000:C:\WINNT\system32\WS2HELP.dll 77fd0000:C:\WINNT\System32\WINMM.dll 77c00000:C:\WINNT\System32\WINSPOOL.DRV 77d80000:C:\WINNT\system32\comdlg32.dll 77b20000:C:\WINNT\system32\ole32.dll Program received signal SIGSEGV, Segmentation fault. If I compile and run applications not using any extra frameworks I have no problems. Any clues? Olof Torgersson -- Olof Torgersson oloft@cs.chalmers.se Department of Computing Science +46 31 772 54 06 G?teborg University & Chalmers University of Technology S-412 96 G?TEBORG, SWEDEN http://www.cs.chalmers.se/~oloft/ From moy at parc.xerox.com Wed Sep 15 14:41:00 1999 From: moy at parc.xerox.com (Edward Moy) Date: Thu Nov 3 14:42:18 2005 Subject: Running Java applications under Mac OS X Server? In-Reply-To: <99Sep15.131242pdt."358605"@louise.parc.xerox.com> Message-ID: This is one of the confusing things about Java when one is first starting. The Java compiler, javac, takes file names as arguments, but the runtime environment, java, takes a class name. So "java foo" looks for a class "foo" in a foo.class file somewhere in the path. If '.' is in the CLASSPATH, it looks in the current directory for foo.class. And "java foo.boo.who" looks for who.class in the "foo.boo" package, somewhere in the path. To execute a class contained in a zip or jar file, you have to have that zip or jar file in the CLASSPATH explicitly. This seemed counter-intuitive to me, since I also would expect if I put a zip or jar file in a directory specified in my CLASSPATH, it should find it. Nope. So I wrote this little shell script that I source (very important to "source" the file, not just execute it). The file is named ~/Java/classpath, and I have the following line in my .cshrc (or compatible) file: source ~/Java/classpath The ~/Java/classpath script is: set myclasspath = ( /boo/hoo /some/where /a/cookie.jar ) set c = '.' foreach i ( $myclasspath ) set c = $c\:$i set suf = $i:e if($suf != 'zip' && $suf != 'jar') then foreach j ( `echo $i/*.jar $i/*.zip |& egrep 'jar|zip'` ) set c = $c\:$j end endif end setenv CLASSPATH $c So the script runs through the $myclasspath, looking for directories (those things not ending in .zip or .jar) and then looks for .zip and .jar files in those directories and adds them to the CLASSPATH. I was tempted to edit the java* scripts to put this in, so that this inclusion would be automatic, but that would make them non-standard... ------------------------------------------------------------------------ Edward Moy Xerox Palo Alto Research Center P.S. The classpath script doesn't deal with paths containing whitespace charaters, so beware. Also, if you add a new zip or jar file, you need to re-source the classpath script. On Wed, 15 Sep 1999, Mike Nowak wrote: > I am having trouble running Java programs that I have created with the > Project Builder. I create a new project as a Java Package and do a build and > I end up with .zip file. When I double click it, OpenUp conveniently unpacks > it. If I use the inspector, I can change the default to JavaRun but when I > double-click it, JavaRun runs but there is no app. > > If I try to run "java SwingApplication" (the sample in the Swing tutorial) > in Terminal, I get "class SwingApplication" not found. The CLASSPATH > variable was not set so I did "setenv CLASSPATH .:`javaconfig > DefaultClasspath`" as suggested in the WebObjects documentation. "." is > included in the CLASSPATH as is ~/Library/Java but having the .zip file in > either of those directories doesn't seem to matter. > > If I generate .class files, either by decompressing the .zip file or > generating the .class files with javac or using the .class files in the > .build directory, then double-clicking on the .class file in the Workspace > Manager or running "java SwingApplication" works. > > I apologize if this is obvious, but I'm used to the Metrowerks environment. > How can I run the Java application from the .zip file? > > Thanks in advance for your help. > > -- > Mike Nowak Instant Messenger: mykkro > The University of Michigan mnowak@umich.edu > Health Media Research Lab http://www-personal.umich.edu/~mnowak/ > ............................................................................ > "Every gun that is made, every warship launched, every rocket fired > signifies a theft from those who hunger and are not fed, those who are cold > and are not clothed. The world is in arms not spending money alone. It is > spending the sweat of its laborers, the genius of its scientists, the hopes > of its children." -- President Eisenhower, 1953. > > From theisen at akaMail.com Wed Sep 15 15:56:49 1999 From: theisen at akaMail.com (Dirk Theisen) Date: Thu Nov 3 14:42:18 2005 Subject: Running Java applications under Mac OS X Server? In-Reply-To: <199909152049.QAA05319@purgatory.healthmedia.umich.edu.> Message-ID: <1dy77zo.1o3y9a1195rplM@[192.168.0.2]> Hi, Mike! > > So, all you need is probably: > > > > java -classpath "./myapp.zip:$CLASSPATH" yourMainClass > > > > where myapp.zip is your zip file, and "yourMainClass" is the name of > > the class that holds your main procedure. > > > > Pierce > > Thanks, that worked. I guess I don't understand why I need to include the > .zip file explicitly if the directory is in the class path. But thanks > again. This was invented by a genious :-) at Apple (unix guys could never think of something that simple/comfortable) and implemented in MRJ. That's why you are probably used to it. Sun realized that and meanwhile provides this feature in the /current/ JDK, AFAIK. They even build a global hashtable which includes all files in all zip files in the path speeding up the class loader. Regards, Dirk -- http://theisen.home.pages.de/ From theisen at akaMail.com Wed Sep 15 15:56:58 1999 From: theisen at akaMail.com (Dirk Theisen) Date: Thu Nov 3 14:42:18 2005 Subject: Setting Up A CVS Server In-Reply-To: <00d801befe0a$bb3826e0$0329040a@nfox.com> Message-ID: <1dy78pg.16zoyc47eof0gM@[192.168.0.2]> Hi, everybody! After going through all the troubles of setting up CVS with wrapper support (and discovering that CVS has enormous problems with the colons in path names on NT -- what crap) I'm now trying to make the pserver mode work. After following the instructions on http://durak.org:81/cvswebsites/howto-cvs/node37.html and http://www.loria.fr/%7Emolli/cvs/doc/cvs_2.html#SEC20 and after discovering that /etc/services is ignored on MXS (look for it in netinfo instead -- why the f*** doesn't the services file state that - as with the other flat files!?) it is now invoked correctly by inetd. Can anyone tell how to make authentification work? The docs state that $CVSROOT/CVSROOT/passwd is consultet (how do I crypt a password on MXS?), it uses the system user lookup, but I cannot login using my system login/password combo (authentication error). Any help welcome!! Regards Dirk P.S. Please Bill, etc. -- make your replacement a good one. CVS is a PAIN in almost every part of my body. From buck.erik at mcleod.net Wed Sep 15 15:59:59 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <001501beffce$0a89c720$062410d0@default> > In other words, if Apple COULD release Yellow Box for Intel, they >would! That's why there's all that stuff left in ProjectBuilder and >such. In fact, wasn't there a YellowBox SDK for Intel on my MOSXS CD? > Sorry to be so blunt, but BULL SHIT. You are being far too generous to Apple. They stated publicly that outstanding licensing issues had been resolved. Apple is just screwing us developers because they can. They asked publicly what the business case for shipping YellowBox for Windows would be. The fact that they do not know already is frightening. (I will not repeat the long discussions of the topic here) Apple released Rhapsody DR2 for Intel. Apple released YellowBox DR2 for Windows WebObjects 4.0 includes YellowBox for Windows. Apple is a shifty unreliable dishonest purveyor of mediocrity. Do not be deceived. They ARE out to get you in this case. PS I Suspect the reason ProjectBuilder etc. have not been changed is because they have less than 12 people working on YellowBox and 1200 on Carbon. These things do not get changed because nobody is working on them. This is just speculation. From pierce at twinforces.com Wed Sep 15 15:44:29 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:18 2005 Subject: Compiling EGCS from Apple? In-Reply-To: References: Message-ID: At 2:08 PM -0700 9/14/99, Todd Daniel Woodward wrote: >Anyone have luck compiling EGCS from Apple's Publicsource? > >http://www.publicsource.apple.com/projects/darwin/source/egcs-2.tar.gz > >./configure isn't complete and complains that there's no >config-guess or config-in > >If you just do make and install, it compiles without a >problem, but doesn't install outside of the source >directory. > >Anyone have any tips to share with a compiling newbie? make install will put it in ./dst, which you can then move to / You could probably change DSTROOT in the Makefile to point to /, but this sounds dangerous to me, personally, as if it doesn't work it could be bad. I'll put a tarchive on darwinlinux when I get done building my version. Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From buck.erik at mcleod.net Wed Sep 15 16:04:53 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: Windows and Frameworks problem Message-ID: <003001beffce$b94b0580$062410d0@default> Normally the "failed reading symbols" stuff just indicates that there was no debugging information in the DLL. The DLL still works. I have no idea why you get SEG FAULTS. From phil at ns2.tcp.net.au Wed Sep 15 16:09:53 1999 From: phil at ns2.tcp.net.au (Phil Blake) Date: Thu Nov 3 14:42:18 2005 Subject: Project Builder: java.lang.OutOfMemoryError Message-ID: I've just added a framework of java classes (about 300) to a project and get an out of memory error when I build. The error is below. I'm running a G3/350 with 448Mb of RAM so I'm assuming that the problem lies with "sun.tools.javac.BatchEnvironment.parseFile" and whatever BatchEnvironment is doing for me. I don't know how to go about resolving this. Is there a way to get ProjectBuilder to not "Batch" so much? Thanks, Phil java.lang.OutOfMemoryError at sun.tools.javac.BatchEnvironment.parseFile(Compiled Code) at sun.tools.javac.Main.compile(Main.java:266) at sun.tools.javac.Main.main(Main.java:473) error: An error has occurred in the compiler; please file a bug report (http://java.sun.com/cgi-bin/bugreport.cgi). ------------------------------ Phil Blake spark tcp Level 12, 323 Castlereagh Street Sydney NSW 2000 Tel 9519 4235 Fax 9519 9445 From rcfa at cubiculum.com Wed Sep 15 16:33:53 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:18 2005 Subject: [OT] Re: cross compilation Windows on Mac OS In-Reply-To: <001501beffce$0a89c720$062410d0@default> References: <001501beffce$0a89c720$062410d0@default> Message-ID: <199909152333.TAA25184@wenix.cubiculum.com> > PS I Suspect the reason ProjectBuilder etc. have not been changed is > because they have less than 12 people working on YellowBox and 1200 on > Carbon. These things do not get changed because nobody is working on them. > This is just speculation. Plus who knows: maybe in three years they do see a business case, and who'd like to put it back in. Plus it hurts a lot more seeing what could be but isn't... ;-) Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From theisen at akaMail.com Wed Sep 15 16:32:52 1999 From: theisen at akaMail.com (Dirk Theisen) Date: Thu Nov 3 14:42:18 2005 Subject: Running Java applications under Mac OS X Server? Message-ID: <1dy7axj.1o4hkvn1yqyr31M@[192.168.0.2]> Hi, Mike! > > So, all you need is probably: > > > > java -classpath "./myapp.zip:$CLASSPATH" yourMainClass > > > > where myapp.zip is your zip file, and "yourMainClass" is the name of > > the class that holds your main procedure. > > > > Pierce > > Thanks, that worked. I guess I don't understand why I need to include the > .zip file explicitly if the directory is in the class path. But thanks > again. This was invented by a genious :-) at Apple (unix guys could never think of something that simple/comfortable) and implemented in MRJ. That's why you are probably used to it. Sun realized that and meanwhile provides this feature in the /current/ JDK, AFAIK. They even build a global hashtable which includes all files in all zip files in the path speeding up the class loader. Regards, Dirk From pierce at twinforces.com Wed Sep 15 16:41:05 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS In-Reply-To: <001501beffce$0a89c720$062410d0@default> References: <001501beffce$0a89c720$062410d0@default> Message-ID: At 5:59 PM -0500 9/15/99, Michelle L. Buck wrote: > > In other words, if Apple COULD release Yellow Box for Intel, they > >would! That's why there's all that stuff left in ProjectBuilder and > >such. In fact, wasn't there a YellowBox SDK for Intel on my MOSXS CD? > > > >Sorry to be so blunt, but BULL SHIT. > >You are being far too generous to Apple. They stated publicly that >outstanding licensing issues had been resolved. Apple is just screwing us >developers because they can. They asked publicly what the business case for >shipping YellowBox for Windows would be. The fact that they do not know >already is frightening. (I will not repeat the long discussions of the >topic here) Actually, they've stated publicly that outstanding licensing issues HAVEN'T been resolved, or resolved in such a way that they could make YellowBox licensing nominal or free, but they've pushed as far as they can. Translation: Someone won't budge on price. The question they asked at WWDC was: Are you interested in YellowBox for Windows at: Free, $50-$100, $1000. And they got the answer: 100%, 50%, 3 people. You can read that multiple ways. I have come to read it to mean that they have to pay a per-seat royalty or a large fee on some of the YellowBox for Intel components. Lets say they have to pay $10. After Apple adds overhead and such, that becomes $20. Since its no longer "Free" Apple now has to figure out how to make money on it, which means they need to retail it for at least $50. That's not a very good solution for a development library, and I'm sure when they told developers: "Look, we're getting bu-fu'ed by Company-X, and we have to charge at least $50, the developers said 'not interested'." And if less developers are interested, then the per-seat price has to go up to support more of the NRE, which means less developers are interested... However, that's YB as it exists today. There's nothing to say that down the road, it might be possible to put a YB solution together, but Apple has to get Cocoa running on PPC first. And perhaps when all thats ready, who knows where the Wintel market will be? Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From buck.erik at mcleod.net Wed Sep 15 18:30:09 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <006801beffe3$04583020$022810d0@elcry> > In other words, if Apple COULD release Yellow Box for Intel, they >would! That's why there's all that stuff left in ProjectBuilder and >such. In fact, wasn't there a YellowBox SDK for Intel on my MOSXS CD? > Sorry to be so blunt, but BULL SHIT. You are being far too generous to Apple. They stated publicly that outstanding licensing issues had been resolved. Apple is just screwing us developers because they can. They asked publicly what the business case for shipping YellowBox for Windows would be. The fact that they do not know already is frightening. (I will not repeat the long discussions of the topic here) Apple released Rhapsody DR2 for Intel. Apple released YellowBox DR2 for Windows WebObjects 4.0 includes YellowBox for Windows. Apple is a shifty unreliable dishonest purveyor of mediocrity. Do not be deceived. They ARE out to get you in this case. PS I Suspect the reason ProjectBuilder etc. have not been changed is because they have less than 12 people working on YellowBox and 1200 on Carbon. These things do not get changed because nobody is working on them. This is just speculation. From phil at ns2.tcp.net.au Wed Sep 15 18:39:53 1999 From: phil at ns2.tcp.net.au (Phil Blake) Date: Thu Nov 3 14:42:18 2005 Subject: ProjectBuilder: OutOfMemory with Java Message-ID: I've just added a framework of java classes (about 300) to a project and get an out of memory error when I build. The error is below. I'm running a G3/350 with 448Mb of RAM so I'm assuming that the problem lies with "sun.tools.javac.BatchEnvironment.parseFile" and whatever BatchEnvironment is doing for me. I don't know how to go about resolving this. Is there a way to get ProjectBuilder to not "Batch" so much? Thanks, Phil java.lang.OutOfMemoryError at sun.tools.javac.BatchEnvironment.parseFile(Compiled Code) at sun.tools.javac.Main.compile(Main.java:266) at sun.tools.javac.Main.main(Main.java:473) error: An error has occurred in the compiler; please file a bug report (http://java.sun.com/cgi-bin/bugreport.cgi). ------------------------------ Phil Blake spark tcp Level 12, 323 Castlereagh Street Sydney NSW 2000 Tel 9519 4235 Fax 9519 9445 From cybergrog at bigpond.com Wed Sep 15 19:49:18 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <02464316750973@domain8.bigpond.com> > Of course, this is all complete conjecture based on observations-- none on any > kind of official information-- and is likely so far away from reality that > those in the know are laughing at us about now. > It is these laughing people who are the ones holding back with the official information. It really pisses me off that a company which used to pride itself on innovation and "insanely great" software and hardware now has just taken the first 6 letters of it and are insane for constricting the developers' target markets. Just an opinion. Greg From mtrent at msn.fullfeed.com Wed Sep 15 22:39:46 1999 From: mtrent at msn.fullfeed.com (Mike Trent) Date: Thu Nov 3 14:42:18 2005 Subject: FreeSpace Central's Fall 1999 Module Contest Message-ID: <199909160539.AAA06190@fullfeed.msn.fullfeed.com> FreeSpace Central is pleased to announce our Fall 1999 Module Contest! BackSpace and FreeSpace module authors can win Apple-branded merchandise (well, t-shirts and such) by releasing new screen saver modules to our eager, internet-savvy, Mac OS X Server wielding public. To enter: 1) Write a new screen saver module for BackSpace or FreeSpace. 2) Make it available for public download via the internet. 3) Let FreeSpace Central know about the new module. Hop to FreeSpace Central for more information: http://www.abunai.org/freespace/ Have fun! Mike Trent FreeSpace Central From andreas at mevis.de Thu Sep 16 03:32:27 1999 From: andreas at mevis.de (Andreas Bohne) Date: Thu Nov 3 14:42:18 2005 Subject: Additional connection inspector (2) Message-ID: <9909161232.AA274708@matisse> I'm not sure if this got through to the list (I got some strange errors from the omnigroup mailer), so I try to post it once more. Sorry for the repost, if the first one was sent anyway. Hi all, some time ago I wrote an additional (to the usual) connection inspector for notification handling, that had to appear for all objects. I made it a separate entry in the inspector window popup (just after the "old" connection inspector). This works, but gives you some odd behaviour (like the need to switch between the two connection inspectors every time you draw a connection). Now I would like to behave it like the EOAssociation inspector - with a popup inside the inspector for selecting the old inspector(s) or my own ones. Unfortunately, searching is disabled in the Omnigroup dev mailing lists now, so I couldn't do a search there. If anyone has done this or has some hints how to do this I would appreciate the help. TIA, Andreas --------------------------------------------------------------------- Andreas Bohne MeVis - Center for Medical Diagnostic Systems and Visualization e-mail: andreas@mevis.de http://www.mevis.de/ --------------------------------------------------------------------- From cybergrog at bigpond.com Thu Sep 16 05:30:45 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:18 2005 Subject: Thoughts for the Stepwise Challenge Message-ID: <12280301613459@domain2.bigpond.com> Has anyone thought of writing a commanche equivalent for OSXS and making it freeware. Greg. From paulrs at lgs-systems.com Thu Sep 16 09:03:00 1999 From: paulrs at lgs-systems.com (Paul R. Summermatter) Date: Thu Nov 3 14:42:18 2005 Subject: Pop Up Menus in table views Continued Message-ID: <51325BD9D61DD31186A50040F6340DB811BF@NAPOLI> Hey guys, It seems quite a number of people are interested in this issue, so I will just post my general findings. Firstly, from a programmatic standpoint, it's pretty much a piece of cake to create an NSPopUpButtonCell and make it the data cell of an NSTableColumn. According to the doc, the NSPopUpButtonCell API is under construction, but I feel confident that the basic methods for creating a pop up button cell and adding items to the cell are going to remain in their current form. The real issue is the association. I have discovered (but I cannot find the documentation on this) that if an object calls setObjectValue: on an NSPopUpButtonCell, the cell tries to take the intValue of what you hand it and pick the item in the list with the tag corresponding to the intValue. This means that if you have a static list of choices which correspond to integer or enumerated values, EOColumnAssociation just works. However, this is not satisfactory for most situations and, I believe, requires a subclass of EOColumnAssociation which implements much of the functionality that EOPopUpAssociation does, but in the context of a table view. I have not written this object yet, but I think it should be straightforward. Since I am thinking about this, I will probably write this object soon. I will report back to the list when I do. Regards, Paul --- Paul Summermatter LGS Systems, Inc. Medical Computing Division 15 TJ Gamester Ave. Portsmouth, NH 03801-5871 (603) 433-9822 voice (603) 433-9818 fax (603) 498-6935 voice mail From jjfeiler at relief.com Thu Sep 16 09:10:56 1999 From: jjfeiler at relief.com (John Jay Feiler) Date: Thu Nov 3 14:42:18 2005 Subject: Pop Up Menus in table views Continued Message-ID: <199909152012.NAA02630@haque.relief.com> If you have a single editing context where the destination objects exist, you can set the tag of the NSPopUpButtonCell to be the address of the destination object..... J Begin forwarded message: > From: "Paul R. Summermatter" > Date: 1999-09-15 13:06:39 -0700 > To: Multiple recipients of list > Subject: Pop Up Menus in table views Continued > Originator: macosx-dev@omnigroup.com > X-Comment: To unsubscribe, follow directions at > http://www.omnigroup.com/MailArchive/ > > Hey guys, > > It seems quite a number of people are interested in this issue, so I > will just post my general findings. > > Firstly, from a programmatic standpoint, it's pretty much a piece of > cake to create an NSPopUpButtonCell and make it the data cell of an > NSTableColumn. According to the doc, the NSPopUpButtonCell API is under > construction, but I feel confident that the basic methods for creating a pop > up button cell and adding items to the cell are going to remain in their > current form. > > The real issue is the association. I have discovered (but I cannot > find the documentation on this) that if an object calls setObjectValue: on > an NSPopUpButtonCell, the cell tries to take the intValue of what you hand > it and pick the item in the list with the tag corresponding to the intValue. > This means that if you have a static list of choices which correspond to > integer or enumerated values, EOColumnAssociation just works. > > However, this is not satisfactory for most situations and, I > believe, requires a subclass of EOColumnAssociation which implements much of > the functionality that EOPopUpAssociation does, but in the context of a > table view. I have not written this object yet, but I think it should be > straightforward. Since I am thinking about this, I will probably write this > object soon. I will report back to the list when I do. > > Regards, > Paul > > --- > > Paul Summermatter > > LGS Systems, Inc. > Medical Computing Division > > 15 TJ Gamester Ave. > Portsmouth, NH 03801-5871 > (603) 433-9822 voice > (603) 433-9818 fax > (603) 498-6935 voice mail > > -- John Feiler jjfeiler@relief.com From wsanchez at apple.com Thu Sep 16 12:01:36 1999 From: wsanchez at apple.com (Wilfredo Sanchez) Date: Thu Nov 3 14:42:18 2005 Subject: Thoughts for the Stepwise Challenge Message-ID: <199909161901.MAA27105@scv3.apple.com> | Has anyone thought of writing a commanche equivalent for OSXS and making it | freeware. One might try porting Tk to AppKit and then use Commanche, if Commanche is good, plus we get Tk, which would be nice. -Fred -- Wilfredo Sanchez, wsanchez@apple.com Apple Computer, Inc., Core Operating Systems / BSD Technical Lead, Darwin Project 1 Infinite Loop, 302-4K, Cupertino, CA 95014 From sanguish at digifix.com Thu Sep 16 14:17:28 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: Thoughts for the Stepwise Challenge In-Reply-To: <12280301613459@domain2.bigpond.com> References: <12280301613459@domain2.bigpond.com> Message-ID: <199909162117.RAA00409@digifix.com> Greg Hulands wrote: > Has anyone thought of writing a commanche equivalent for OSXS and making it > freeware. Let me expand a we bit on the 'Challenge' (a much better name that what I had.. :-) We've got a programming contest that I've been trying to get announced for weeks now. We've got some great prizes (licenses for apps from three different ISVs, and for the best app with source provided, a 3Com LANOffice Modem)... I've just not gotten the verbage out as far as rules wise.. I've gotten some good suggestions from Mike Trent today on wording and from David 'YB/NT poster boy' Herren. So I'll try and get that up on the site tonight. From sanguish at digifix.com Thu Sep 16 14:43:08 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: Thoughts for the Stepwise Challenge Message-ID: <199909162143.RAA00487@digifix.com> Greg Hulands wrote: > Has anyone thought of writing a commanche equivalent for OSXS and making it > freeware. Let me expand a we bit on the 'Challenge' (a much better name that what I had.. :-) We've got a programming contest that I've been trying to get announced for weeks now. We've got some great prizes (licenses for apps from three different ISVs, and for the best app with source provided, a 3Com LANOffice Modem)... I've just not gotten the verbage out as far as rules wise.. I've gotten some good suggestions from Mike Trent today on wording and from David 'YB/NT poster boy' Herren. So I'll try and get that up on the site tonight. From nhoj at cd.chalmers.se Thu Sep 16 15:58:05 1999 From: nhoj at cd.chalmers.se (nhoj@cd.chalmers.se) Date: Thu Nov 3 14:42:18 2005 Subject: GraphicKit beta 5 available. In-Reply-To: <199909162143.RAA00487@digifix.com> References: <199909162143.RAA00487@digifix.com> Message-ID: <19990916225806.7869.qmail@glenn> Beta 5 of the GraphicKit is now available. You can download it from http://www.cd.chalmers.se/~nhoj/GraphicKit/index.html. New in this release is support for animation. The framework documentation has been updated, and saving is now fully implemented, although mostly untested. Regards, John Hornkvist From phr at projectcenter.ch Thu Sep 16 16:01:43 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:18 2005 Subject: Mac OS X Client on G4 In-Reply-To: <4C094419B0E6D2119FF100805F85CABE012DDD52@usilmse1.cai.com> References: <4C094419B0E6D2119FF100805F85CABE012DDD52@usilmse1.cai.com> Message-ID: <9909162301.AA00353@projectcenter.ch> You wrote: > Has anyone tried to install the MacOS X Client preview on a G4 machine yet? I was told @ the Apple Expo in Paris yesterday that there will be a new MOSXS Version in November that runs on the G4s...;-) sweet dreams, Phil --- Philippe C.D. Robert http://www.nice.ch/~phip From bbum at codefab.com Thu Sep 16 17:45:43 1999 From: bbum at codefab.com (Bill Bumgarner) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <199909170045.TAA23178@pathos.codefab.com> As a publically owned, company, Apple's single most important business goal is to increase the value of its stock. Considering that the stock has gone from a low of 13 to a 7 year high of 77 in the last two years, Steve & Friends are doing a remarkably good job of this by introducing some pretty damned cool hardware products. And that is the key, Apple is fundamentally a hardware company. Hardware provides the profit and the hardware is sold with a margin way beyond that of software. All things considered WebObjects is a bit of an anomaly-- while, in and of itself, it is a very successful product (despite the complete lack of marketing), WebObjects really doesn't contribute a noticeable amount to the overall value of the company. Widening the developer's market to include platforms that do not sell Apple hardware is not necessarily in Apple's best interest. Third party developers will argue that things like OS X Server Intel and Yellow/NT would contribute to Apple value by increasing the number of developers on the platform-- but is that really the case? Would the additional value provided by a handful of additional products balance the value of the lost hardware sales to folks that decide to take advantage of insanely great software running on a lowest common denominator system (Intel running Windows)? The press and some developers are constantly stating that Apple can't survive as a hardware company... that Apple must take on a Microsoft business model to succeed. Yet, Apple continues to grow hardware sales, continues to lead the industry in terms of minimizing on hand inventory, and continues to release hardware products that are revolutionary (not in their individual features, but in the packaging of a complete set of features that simply aren't found in one place at one time on other systems for the price). Considering the stock price, the continued quarter after quarter profit gains, the buy ratings across all of the investment houses, the clear direction for growth of the announced product lines, and the vast potential in the unnannounced, but rumored, products, Apple is both extremely successful and, painful though it has been to the development community, has been making the correct business decisions. Unfortunately, the developers recieve the short end of the stick. It is unfortunate that Apple promised Yellow/NT licensing and then canceled. It is less unfortunate-- but still unfortunate-- that OS X Server Intel never shipped. But, again, would either of those products have directly contributed to the stock being of a greater value today than it already is? It is also interesting to note that Apple's story on one front has remained clear and consistent throughout the entire NeXT takeover; The company must stop the bleeding as a top priority. Steve and others said this from the start of the merger and they have said that there would be some very dififcult decisions to be made that were going to piss off a lot of people (I still distinctly remember the OpenDoc discussion at WWDC '97). That is exactly what happened and the bleeding has mostly stopped. The hardware future is well mapped and everything indicates that next year will be the year of Software for Apple. It is going to be interesting and, personally, I doubt that technologies like Yellow or WebObjects aren't going to make a very strong showing in the coming years. What form and which markets remains to be seen. b.bum From: "Greg Hulands" Date: 1999-09-15 19:51:17 -0700 To: Multiple recipients of list Subject: Re: cross compilation Windows on Mac OS > Of course, this is all complete conjecture based on observations-- none on any > kind of official information-- and is likely so far away from reality that > those in the know are laughing at us about now. > It is these laughing people who are the ones holding back with the official information. It really pisses me off that a company which used to pride itself on innovation and "insanely great" software and hardware now has just taken the first 6 letters of it and are insane for constricting the developers' target markets. Just an opinion. Greg From cybergrog at bigpond.com Thu Sep 16 18:25:29 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:18 2005 Subject: Mac OS X Client on G4 Message-ID: <01231043639988@domain8.bigpond.com> Will this be OSXS version 2.0 ? Greg ---------- >From: "Philippe C.D. Robert" >To: Multiple recipients of list >Subject: Re: Mac OS X Client on G4 >Date: Fri, 17 Sep 1999 9:08 AM > > You wrote: >> Has anyone tried to install the MacOS X Client preview on a G4 machine yet? > > I was told @ the Apple Expo in Paris yesterday that there will be a new > MOSXS Version in November > that runs on the G4s...;-) > > sweet dreams, Phil > --- > Philippe C.D. Robert > http://www.nice.ch/~phip > From rthomas at uiuc.edu Thu Sep 16 18:36:06 1999 From: rthomas at uiuc.edu (Ryan Thomas) Date: Thu Nov 3 14:42:18 2005 Subject: Compiling EGCS from Apple? In-Reply-To: References: Message-ID: But the only thing in the ./dst dir after install are the c++ headers (and copy of source). What about all of the objects, symbols, libraries etc in ./sym, ./obj, ./gcc dirs? How are these installed? -Ryan At 4:33 PM -0700 9/15/99, Pierce T. Wetter III wrote: >At 2:08 PM -0700 9/14/99, Todd Daniel Woodward wrote: >>Anyone have luck compiling EGCS from Apple's Publicsource? >> >>http://www.publicsource.apple.com/projects/darwin/source/egcs-2.tar.gz >> >>./configure isn't complete and complains that there's no >>config-guess or config-in >> >>If you just do make and install, it compiles without a >>problem, but doesn't install outside of the source >>directory. >> >>Anyone have any tips to share with a compiling newbie? > > make install will put it in ./dst, which you can then move to / > > You could probably change DSTROOT in the Makefile to point to /, >but this sounds dangerous to me, personally, as if it doesn't work >it could be bad. > > I'll put a tarchive on darwinlinux when I get done building my version. > >Pierce From buck.erik at mcleod.net Thu Sep 16 18:42:04 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <004a01bf00ad$d8cfc320$332910d0@elcry> People keep stating that Apple is primarily a hardware company. Margins on hardware are at best less than 30 % and often less than 10 %. Margins on software are routinely more than 90 %. Even if software sales bit into Apple's hardware sales, why would they care if they make more money from software anyway ? If they profit $200 every time they sell a G4 and they profit $250 every time they sell Rhapsody Intel, they win either way. Even better, not every Intel sale costs a G4 sale. There is an opportunity to radically increase market share and revenue and profit. Regarding YellowBox for Windows, there is absolutely NO reason not to ship. Sell the dev tools for $500. That is at least 90% margin considering all development costs are sunk. So what if software becomes available for Mac and Wintel both. There is hardly any Mac only software anyway. They have nothing to loose and more software to gain. Face it...they are already playing catch-up. YellowBox for Windows is really important to small developers. The big players already have cross platform libraries. They already sell for both platforms or Windows only. Allowing small developers to maintain one code base across platforms is a big win for them. Perhaps it is the big player's influence on Apple that has scuttled things. As a small developer, I am first interested in deployment on Wintel. If it costs little to release for Mac too, we might. As it is there is no chance. From rthomas at uiuc.edu Thu Sep 16 21:07:58 1999 From: rthomas at uiuc.edu (Ryan Thomas) Date: Thu Nov 3 14:42:18 2005 Subject: Compiling EGCS from Apple? Message-ID: But the only thing in the ./dst dir after install are the c++ headers (and copy of source). What about all of the objects, symbols, libraries etc in ./sym, ./obj, ./gcc dirs? How are these installed? -Ryan At 4:33 PM -0700 9/15/99, Pierce T. Wetter III wrote: >At 2:08 PM -0700 9/14/99, Todd Daniel Woodward wrote: >>Anyone have luck compiling EGCS from Apple's Publicsource? >> >>http://www.publicsource.apple.com/projects/darwin/source/egcs-2.tar.gz >> >>./configure isn't complete and complains that there's no >>config-guess or config-in >> >>If you just do make and install, it compiles without a >>problem, but doesn't install outside of the source >>directory. >> >>Anyone have any tips to share with a compiling newbie? > > make install will put it in ./dst, which you can then move to / > > You could probably change DSTROOT in the Makefile to point to /, >but this sounds dangerous to me, personally, as if it doesn't work >it could be bad. > > I'll put a tarchive on darwinlinux when I get done building my version. > >Pierce From sanguish at digifix.com Thu Sep 16 22:14:37 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS In-Reply-To: <199909170045.TAA23178@pathos.codefab.com> References: <199909170045.TAA23178@pathos.codefab.com> Message-ID: <199909170514.BAA02115@digifix.com> Bill Bumgarner wrote: > > Unfortunately, the developers recieve the short end of the stick. It is > unfortunate that Apple promised Yellow/NT licensing and then canceled. It > is less unfortunate-- but still unfortunate-- that OS X Server Intel never > shipped. But, again, would either of those products have directly > contributed to the stock being of a greater value today than it already is? > Credibility with developers is important. All the hardware sales in the world will disappear if Apple looses developers again. Apple had a chance to prove that they had stopped treating all developers as they had, and they chose to continue to screw some of them over. Whatever excuses that people decide to except as far as YB licensing goes, there are those who are very bitter because they __know__ for a fact that the issues are resolved. Paper in hand, so to speak. From sanguish at digifix.com Thu Sep 16 22:15:39 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: Mac OS X Client on G4 In-Reply-To: <01231043639988@domain8.bigpond.com> References: <01231043639988@domain8.bigpond.com> Message-ID: <199909170515.BAA02128@digifix.com> Greg Hulands wrote: > Will this be OSXS version 2.0 ? No... 1.2 2.0, assuming we get a 2.0, would likely be with the new underpinnings (Mac OS X Client). From wchin at acm.org Thu Sep 16 23:21:56 1999 From: wchin at acm.org (Bill Chin) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <199909170621.CAA22879@mail.richmond.net> I truly don't understand what Apple is afraid of here. There is no way Apple can get the Yellow Box runtime license fee down to $0 with the current release that is part of WebObjects on NT. Plus, it doesn't perform all that well under Windows NT and there are plenty of problems using it under NT, from user, sysadmin, and developer standpoints. There are plenty of packaging issues too... if I buy GlyphiX and Create, will the end user end up paying for two copies of the Yellow Box runtime? Would end users buy it separately? Therefore, the only people that want to buy it _really_ want to buy it and know full well what limitations exist. They want to buy it anyways! The truth is, in its current incarnation, it's not going to take over the Windows development toolkit world. It's not going to supplant VC++ or VB. Developers as a whole hate runtime licenses. Why not slap it into a box and take money for it (with appropriate disclaimers attached)? Apple has to maintain the code base for WebObjects on NT anyways. Don't put any money into marketing it. Don't make a big splash about it. Do sell it at the lowest price point Apple can sell it for to developers (if you make $10 on a $30 item, that's still a great profit margin) and if users end up paying for two or three or four copies of the runtime, so much the better! That way, developers that are _already_ committed to MacOS X Server and MacOS X can deliver to customer desktops now ... it's not like there exists a way to deliver Yellow Box apps to MacOS 8.x now nor are companies deploying MacOS X Server as a desktop in mass. Developers that are targeting MacOS 8.x or even 9.x now aren't using Yellow Box anyways... and can't. Besides, they've got that thing that all life forms are based on. A current Mac user isn't going to choose a Wintel box over this... after all, they're not using any Yellow Box apps anyways! A current Wintel user isn't going to give up a Power Macintosh purchase of this. How many Power Macintoshes are running MacOS X Server as a desktop? So the only market are those that are committed to Windows on the desktop and have crazy developers that want to deploy an already written Yellow Box application on the desktop. Or, developers want to write to Cocoa for a future product. These developers then will have or do have products that are pretty much ready to ship for MacOS X when that arrives. The alternative is that they write a Windows app w/o Yellow Box and has ties directly to Windows (I would bet most Java apps to be deployed on the desktop have JNI calls that bind them to Windows specifically). Thus these apps won't be ready for the move to MacOS X Client. In summary, it's not going to make the difference between a Power Macintosh purchase or a WinTel machine purchase in the grand scale of things _now_. Thus it doesn't affect the question of whether Apple is a hardware or software company. Now, once Apple completes MacOS X, and hopefully ports all that work to Windows and has a possibility of a $0 runtime fee for Cocoa, then it becomes a different world. But Apple can choose when to release that product, and it might be significantly after MacOS X client ships for the Power Macintosh. Again, we're talking the middle of next year. I certainly hope that by that point in time, one would be willing to choose to buy a Power Macintosh because its value as hardware alone, regardless of the operating system that it runs. Apple as a hardware company only has to fear if its hardware cannot stand on it own. ..Bill Chin BTW, I really do like Yellow Box, even on NT. -- Bill Chin (NeXTmail/MIME accepted) From sanguish at digifix.com Fri Sep 17 00:39:13 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: Stepwise Mac OS X Server/Cocoa Application Challenge Message-ID: <199909170740.DAA02592@digifix.com> [if you need some ideas, there will be an article on Stepwise shortly to help give you some ideas. If you want to add prizes, drop me some email - Scott] We thought that perhaps one good method of generating some new applications for Mac OS X Server would be to hold a contest and offer prizes. Any newly released, non-commercial Cocoa API application can be entered by the author. The contest will be divided up into two stages, the Editors' Choice awards, and the Readers' Choice awards. Editors' Choice awards Our panel of judges will pick their favorite Cocoa based application based on suitability for the task, user interface, useability and practicality. That author of that app will receive our Editor's Choice award consisting of a FrontBase Web Deployment license (donated by Frontline Software) and an Enterprise level Communigate Pro license (courtesy Stalker Software, Inc.). The winner and four runners-up will receive a stylish Stepwise "Think Wisely" polo shirt. If the winning application includes source code, the author will also receive an additional prize, a 3Com LanOffice 56K Modem. Using this modem, you can get your OS X Server based PowerMacintosh connected to the net! The hope is that this award will entice developers to provide source so that others can learn from their apps. Readers' Choice Award After the contest has closed, and all entries have been judged for the Editors' Choice Award, we'll open up the voting to you, the readers. You will be able to download the apps, try them, and then vote on-line for your favorite. The author of the application with the most votes will receive a license for the complete Stone Web Suite (courtesy Stone Design Corporation). To get the rules, or find out more about the contest hit the site at http://www.stepwise.com/Contests/CocoaChallenge From olive-tl at ee.uwa.edu.au Fri Sep 17 03:22:51 1999 From: olive-tl at ee.uwa.edu.au (Timothy Lee Oliver) Date: Thu Nov 3 14:42:18 2005 Subject: Cant connect to registered connections from a remote machine. Message-ID: <99091718360800.23518@zeus> Morning people, (at least, I assume it is for the majority of you). I'm having trouble getting DO going between machines. The source code I'm using is the ProxyEx source that came through this list a month or two ago. Remote MachIPC is on. (Anyone know if changing that setting in the control panel takes immediate effect/ requires signalling something/ requires reboot?) The MachIPC setting alone wasn't enough though. Looking through the archives for this list, I found that the portname needs to be in the nmserver.access (sp?) file. Now, ProxyEx registers a connection with the name @"ProxyEx", so I added the line: ProxyEx:+ to nmserver and restarted the machine just in case (what does it take to notify nmserver of such a change anyway)? The client on a remote machine still couldn't connect, it would just timeout after a few seconds. The two machines aren't connected through netinfo in any way (that is, one isnt a child of the others domain). Is that whats stoping it? Or is it something else I've forgotten. Anyone got some ideas? From Philippe.Robert at uptime.ch Fri Sep 17 03:42:38 1999 From: Philippe.Robert at uptime.ch (Philippe Robert) Date: Thu Nov 3 14:42:18 2005 Subject: Mac OS X Client on G4 In-Reply-To: <01231043639988@domain8.bigpond.com> References: <01231043639988@domain8.bigpond.com> Message-ID: <9909171044.AA27769@california.uptime.ch> You wrote: > Will this be OSXS version 2.0 ? 1.2 AFAIK, just a 'minor upgrade' I think. sweet dreams, Phil -- Philippe C.D. Robert | Uptime ObjectFactory Inc Unix/OpenStep Software Engineer | http://www.uptime.ch http://www.nice.ch/~phip | info@uptime.ch From maury at oaai.com Fri Sep 17 11:30:10 1999 From: maury at oaai.com (Maury Markowitz) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS References: <199909170514.BAA02115@digifix.com> Message-ID: <37E288B2.685C7FA5@oaai.com> Scott Anguish wrote: > Credibility with developers is important. All the hardware sales in > the world will disappear if Apple looses developers again. Sure, but from their standpoint, how many will they lose really? 1% or so? The market realities are pretty jumbled these days. Heck, they could likely trash WO on OS-X and not _really_ effect that many people, or WO. Sad but true. > Whatever excuses that people decide to except as far as YB licensing > goes, there are those who are very bitter because they __know__ for a fact > that the issues are resolved. Paper in hand, so to speak. But that's only one aspect of the problem. And before anyone (Eric) accuses me of being some brain dead Apple appologist, let me remind you all that I actually have a _shipping_commercial_product_ on YB/NT and thus Apple's decision has _actually_ affected our bottom line - unlike many of the more "theoretical" complaints I've seen here. While the licensing is the single biggest issue involved, it's by no means the only one. Another is that the system is about to undergo a massive upgrade, and it's unclear whether or not they are interested in bringing the libs along for the ride. Personally I think they are pretty much forced to, but that's my guess. If they _don't_ bring the libs forward, then all of us would be in a much worse place if they _did_ release the current stuff. Then we'd have real shipping products being yanked out of the channel. Another issue involves the reality that YB/NT may indeed be the world's best cross-platform development system, but simply put it's a crappy way to ship NT apps. Don't fool yourself on this. YB/NT apps don't really look like NT apps (it needs real NT-style toolbar support for one). They can't really share data with other NT apps - even cut and paste doesn't work in many cases, and Apple's EPS doesn't provide a file that will be viewable under NT apps that read them (it's missing the preview - this is a problem on Mac apps too). Text file differences are ignored making even text editor data exchange a problem. And then there's the whole "Apple" folder issue. I can't even find an installer on Win that can install my app. The long and short of it is that this is NOT a "productized" system. It may indeed be fine for WO building and such, because people _expect_ to install all sorts of crap all over the place. But do you _really_ think that customers will accept you installing 338 megs of over 18000 files to run your application? Dream on. Even a deployment version is unlikely to cut that to less that 500-1000 files. Some have compared this to QT, but QT installs about 7.5 megs in 10 files, and installs them in the right place. YB/NT is, and will be for the forseeable future, a custom application development system. It is not, nor is likely to be, a generalized application deployment system. I have to live with my market being those people running WO, but simply put that was the majority of our market anyway. Unless your product is custom and not developer related (like David's) then the real effect on your sales is likely to be pretty much zero. I can't say I'm terribly happy about the broken promises too, but everyone involved seems to suggest this is some simple issue of releasing the libs. It's not. Maury From alexnet at gestalt.com Fri Sep 17 11:24:17 1999 From: alexnet at gestalt.com (Alex Molochnikov) Date: Thu Nov 3 14:42:18 2005 Subject: Cant connect to registered connections from a remote machine. Message-ID: <9909171824.AA00436@gestalt.com> Timothy Lee Oliver wrote: > I'm having trouble getting DO going between machines. >From your description it appears that you have not one but two problems: 1. Mapping of the host name to the IP address. NSConnection takes the name of the remote host for getting a proxy to a remote object, but it must be able to internally translate this host name into the IP address. If you do not have NetInfo running (or did not register your hosts with it) then you may have to add host/IP address entries in your /etc/hosts file (that is, the client machine needs these mappings in its own /etc/hosts file that will point to the remote server). 2. The host machine that runs the server (i.e. the one that accepts the incoming NSConnections) must be configured to allow such connections. Setting Remote MachIPC on is not enough, you have to open the /etc/hostconfig file and change MACHIPC entry in it to -UNRESTRICTED-. Refer to /System/Documentation/ReadMe/Networking&Services/Security.html for detailed explanation of this paranoia. Alex Molochnikov Phoenix Data Trend -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 1088 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990917/11da41a5/attachment.bin From dyoung at vviuh221.vvi.com Fri Sep 17 12:49:30 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS References: <37E288B2.685C7FA5@oaai.com> Message-ID: <9909171949.AA19150@vviuh221.vvi.com> Maury Markowitz said ... > > excuses ... excuses... and more excuses ... > ... ditto ... The fact of the matter is that we have already SHIPPED applications on Win NT and our customers love it! All the reasons you gave are mere excuses for not forging ahead. I prefer to forge ahead and solve problems. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From alexnet at gestalt.com Fri Sep 17 14:38:13 1999 From: alexnet at gestalt.com (Alex Molochnikov) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS Message-ID: <9909172138.AA00232@gestalt.com> Maury Markowitz wrote: > do you _really_ think that customers will accept you installing 338 megs of over 18000 > files to run your application? Dream on. Even a deployment version is unlikely to cut that > to less that 500-1000 files. BS, pure and simple. All that is needed for the deployment license are a few DLL libraries and executable files. Namely: 93 files in Apple/Library/Executables directory totalling about 27.5 Mb. M$ Service Pack 4 for Win NT is about 32 Mb, and nobody complains. Why somebody would try so hard to defend the unethical behavior of Apple is beyond my comprehension. Alex Molochnikov Phoenix Data Trend -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 647 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990917/2b494fe7/attachment.bin From buck.erik at mcleod.net Fri Sep 17 16:33:27 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <003101bf0165$0b7b0700$252410d0@default> >CoreFoundation in a C-based API (and implemented in C) with a lot of the >"look and feel" of facilities found previously only in the Foundation >framework in Cocoa. It was and is currently developed by the same group >which maintains the two principal Cocoa frameworks, Foundation and AppKit, >and some other Cocoa-related technologies. The primary intent is to >broaden the availability of Foundation-like APIs and data types to a wider >audience: Mac OS developers and internal providers of other C-based APIs. Given that any C program can be compiled with Objective-C, what is the point of new functional APIs ? In order to use new core foundation APIs, developers will have to rewrite portions of their existing C based applications. Since they are rewriting anyway, why stick them with half assed half OO APIs when they can use full OO with the same compiler and the same amount of effort ? Is it really that painful for developers to type myArray = [[NSMutableArray alloc] init] than to type myArray = CFCreateNewEmptyMutableArray() ? As you claim they result in the same utility. The former just has light years more power and flexibility via categories, posing, subclassing, ecapsulation, etc. (I do not know the CF API so I apologize for getting the function wrong, but I think my point is clear.) From cdouty at netcom.com Fri Sep 17 17:26:08 1999 From: cdouty at netcom.com (Chris Douty) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <003101bf0165$0b7b0700$252410d0@default> Message-ID: On Fri, 17 Sep 1999, Michelle L. Buck wrote: > >CoreFoundation in a C-based API (and implemented in C) with a lot of the > >"look and feel" of facilities found previously only in the Foundation > >framework in Cocoa. It was and is currently developed by the same group > >which maintains the two principal Cocoa frameworks, Foundation and AppKit, > >and some other Cocoa-related technologies. The primary intent is to > >broaden the availability of Foundation-like APIs and data types to a wider > >audience: Mac OS developers and internal providers of other C-based APIs. > > > Given that any C program can be compiled with Objective-C, what is the point > of new functional APIs ? In order to use new core foundation APIs, > developers will have to rewrite portions of their existing C based > applications. Since they are rewriting anyway, why stick them with half > assed half OO APIs when they can use full OO with the same compiler and the > same amount of effort ? Is it really that painful for developers to type > myArray = [[NSMutableArray alloc] init] than to type myArray = > CFCreateNewEmptyMutableArray() ? As you claim they result in the same > utility. The former just has light years more power and flexibility via > categories, posing, subclassing, ecapsulation, etc. (I do not know the CF > API so I apologize for getting the function wrong, but I think my point is > clear.) The stated rationale for Core Foundation is Carbon (other C-based APIs). Both Carbon and Cocoa use CF to some degree so that low level functionality need only be implemented once. I don't know how successful this will really be; I can see the CF API being extended in one way to support Carbon and another to support Cocoa, but that's really Apple's problem. The real implication of this is that Carbon is intended to stay for the long haul. Anyone still clinging to a fantasy where Carbon is a transitional API ought to give it up now. MOSX and CF will give the Mac toolbox life support long into the next decade. Only the adventurous will sample Cocoa, probably through some silly Java too. -Chris -- Christopher Douty - Rogue Engineer trapped in a land of software cdouty@netcom.com "Frequently the messages have meaning; that is they refer to or are correlated according to some system with physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem." -Shannon From cmh at greendragon.com Fri Sep 17 17:47:13 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <003101bf0165$0b7b0700$252410d0@default> References: <003101bf0165$0b7b0700$252410d0@default> Message-ID: At 4:40 PM -0700 9/17/99, Michelle L. Buck wrote: >same amount of effort ? Is it really that painful for developers to type >myArray = [[NSMutableArray alloc] init] than to type myArray = >CFCreateNewEmptyMutableArray() ? When the former will never be supported on the several million non-G3/G4 PowerMacs that are out there, YES. I'm sure if Apple had Cocoa running on both MacOS 9 and Mac OS X that there would be (almost) no need for a C-based CoreFoundation. But since developers will have to deal with a MacOS 8/9 user base of *millions* for at least a couple more years, Apple had to either provide a library with Foundation's functionality that works without any Objective-C runtime, provide a standard MacOS Objective-C runtime, get compiler vendors to adopt it, and then port Foundation to it, or face near-zero adoption and start writing a version of the Toolbox that interoperates well with STL. (My assumption with that last bit is that Carbon will be using CoreFoundation types. STL is the only reasonable alternative; I'll take CoreFoundation any day, thanks.) From buck.erik at mcleod.net Fri Sep 17 19:11:43 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <002601bf017b$27cfd3c0$552b10d0@default> Seriously, Why can't OS 9 have the portions of the FoundationKit that do not require real operating system feature. Is Carbon with CF going to be available for OS 9 ? Legacy Mac developers can not use any OS "feature" that is not supported in OS 9 ? (Isn't that what is being claimed ?) What about features that are not in OS8 ? I'll stick to my guns on this. Legacy C code must be changed to take advantage of CF. Why change it to a half OO procedural API when for the same effort it could be changed to the real thing ? Hell, the Foundation classes could be wrapped in silly simple C APIs with typedefs to hide the class names. What would be the point ? CF seems to be creating something similar to function wrapped classes by writing everything from scratch and introducing new bugs and closing the door of future use of OO and a dynamic runtime. It should be possible to compile any Carbon application with Apple's Objective-C compiler without any code modification. If you do not use object Objective-C features, you have C. If you are going to modify the previously mentioned Carbon application in order to use CF, why not modify it to use Foundation instead ? I suspect the answer is because Apple is dropping Foundation and all other YellowBox stuff in the near future. From sanguish at digifix.com Fri Sep 17 20:55:16 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: cross compilation Windows on Mac OS In-Reply-To: <37E288B2.685C7FA5@oaai.com> References: <37E288B2.685C7FA5@oaai.com> Message-ID: <199909180355.XAA06276@digifix.com> Maury Markowitz wrote: > Scott Anguish wrote: > > Credibility with developers is important. All the hardware sales in > > the world will disappear if Apple looses developers again. > > Sure, but from their standpoint, how many will they lose really? 1% > or so? The market realities are pretty jumbled these days. Heck, they > could likely trash WO on OS-X and not _really_ effect that many people, > or WO. Sad but true. > You tell me.. how much credibility is lost when a major company dumps ALL Apple hardware because of this.. or ALL NeXT solutions because of this? How long until it effects WebObjects sales (a major money maker at Apple at this point)... Apple needs to understand that EVERY developer counts, not just the Adobes. Apple needs to understand that EVERY sale counts. Apple doesn't understand that. > > Whatever excuses that people decide to except as far as YB licensing > > goes, there are those who are very bitter because they __know__ for a > > fact that the issues are resolved. Paper in hand, so to speak. > > But that's only one aspect of the problem. And before anyone (Eric) > accuses me of being some brain dead Apple appologist, let me remind you > all that I actually have a _shipping_commercial_product_ on YB/NT and > thus Apple's decision has _actually_ affected our bottom line - unlike > many of the more "theoretical" complaints I've seen here. I've not seen ANYONE here make theoretical complaints about not being able to ship.. On the other hand, I've seen much theorizing on why Apple shouldn't do it, or can't do it. > Then we'd have real shipping products > being yanked out of the channel. > Yes, because we all know that our apps would immediately stop working. To hell with the fact that customers would have a useable, deployable situation that would run for the next few years, that they could recoup their investment in during the meantime.. > And then > there's the whole "Apple" folder issue. I can't even find an installer > on Win that can install my app. > So, a Windows app, which litters files from here to beyond can be installed, but a YB app that copies files to specific places can't? Come on Maury... Apple is able to install WebObjects using InstallShield. There is no reason that you can't install Glyphix similarly. > Some have compared > this to QT, but QT installs about 7.5 megs in 10 files, and installs > them in the right place. > Others have already pointed out that a minimum of files are absolutely necessary. And regardless, if OUR customers are willing to do this, what difference does it make if yours aren't? > YB/NT is, and will be for the forseeable future, a custom application > development system. It is not, nor is likely to be, a generalized > application deployment system. Reality check.. Even the custom application people are screwed, since we can't deploy our apps either. From afleming at andrew.cmu.edu Fri Sep 17 21:50:05 1999 From: afleming at andrew.cmu.edu (Andrew M Fleming) Date: Thu Nov 3 14:42:18 2005 Subject: WaveLAN Message-ID: Ok, I know it may be crazy, but is there any possibility that someone has or knows how to get a driver for a PCMCIA waveLAN card for MOSXS? --Andy Fleming Carnegie Mellon University Carnegie Institute of Technology Electrical and Computer Engineering From rcfa at cubiculum.com Sat Sep 18 02:43:01 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: References: Message-ID: <199909180943.FAA03228@wenix.cubiculum.com> > The real implication of this is that Carbon is intended to stay for the > long haul. Anyone still clinging to a fantasy where Carbon is a > transitional API ought to give it up now. MOSX and CF will give the Mac > toolbox life support long into the next decade. Only the adventurous will > sample Cocoa, probably through some silly Java too. That may be the case, but then the Cocoa-Java strategy makes even less sense: if Carbon-CoreFoundation is the bread&butter API even in the future, and the entire purpose of Java is to be cross-platform and to be free of encumbrances through proprietary APIs such as Cocoa, then Java should remain pure, Carbon should be the mass market API and Cocoa remains esoteric. In that case however, Cocoa and Apple won't be hurt if Cocoa stays linked to an appropriately esoteric language like ObjC. If there's little point in making Cocoa a mass market API, then there's no point either in shoehorning it into a mass market hype languate like Java that cripples Cocoa's potential. Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From sebestyen_g at accentcomm.com Sat Sep 18 06:34:12 1999 From: sebestyen_g at accentcomm.com (Gabor Sebestyen) Date: Thu Nov 3 14:42:18 2005 Subject: __ NE2K for MacOSX Server __ Message-ID: <37E394D4.74B96A56@accentcomm.com> Hi! I just wanted to tell you I'm going to release the first beta of NE2K (v0.9.2) driver for MacOSX Server ASAP. It's a really hot stuff therefore it's not yet fully tested (but working!). I intend to release NE2K for test purposes so dear testers welcome! :)) The binary actually is for MacOSX Server only (not planned to compile on OPENSTEP/Mach). It would be good to have it tested on Darwin ppc/i386 ... The source code will also be included. The driver will be available on our server at: ftp://ftp.accentcomm.com/pub/downloads/Rhapsody/ Best regards, Gabor Sebestyen -------------------------------------------------------------- A C C E N T C O M M U N I C A T I O N S - We Make Your Web Groove - Tel: (+36 1) 315-0701 http://www.accentcomm.com (+36 1) 343-4440 e-mail: info@accentcomm.com GSM: (+36 30) 951-4334 Fax: (+36 30) 980-0344 -------------------------------------------------------------- From ckane at apple.com Sat Sep 18 10:33:16 1999 From: ckane at apple.com (Chris Kane) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <011101bf01fb$e4572a40$9d0eca11@verbosity.apple.com> Michelle L. Buck writes: >Seriously, Why can't OS 9 have the portions of the FoundationKit that do not >require real operating system feature. Is Carbon with CF going to be >available for OS 9 ? Legacy Mac developers can not use any OS "feature" >that is not supported in OS 9 ? (Isn't that what is being claimed ?) What >about features that are not in OS8 ? Most of CF is part of CarbonLib, and is available on Mac OS 9 and whatever other older versions of Mac OS CarbonLib will ship on. Some new APIs in Mac OS 9/CarbonLib make use of CF in API and implementation. >I'll stick to my guns on this. Legacy C code must be changed to take >advantage of CF. Why change your legacy C code at all? If it works, don't bother. If you want to enhance it with new features in CF, then yes, you have to use CF. If you're enhancing it, you can equally choose to use Foundation instead and make it (or parts of it) ObjC code. You can do both. You can change it to use old-Mac OS-style Carbon APIs if you like, or use the new APIs in Carbon. Whatever. Not everyone wants to use ObjC however, particularly since it won't be available from Apple on Mac OS Frankly, I think the other issue that Apple is ignoring is that GnuStep will in a short amount of time run on Windows...there are enough people with enough of a commercial business plan to make that happen. When that does, and developers go through the effort of porting from MOXS Cocoa to GS/Linux & Win32 (expected to be comparatively small), they break through the insanely great reality distortion field and realise the other 87% of the computing population doesn't buy Apple machines regardless of what colour they are. In anybody's market, that is a real win for small developers, the public domain being the means to counteract the cartel behaviour of the large developers who are happy to see cross-platform solutions remain unavailable to the small developers. If this situation becomes reality, Apple will have once again squandered opportunities to not only make a product, but to keep responding to their developers in order to continue to advance the cutting edge of their software. Apple needs to realise that by responding to Cocoa developers they are investing in R&D and have the opportunity to leverage that work towards other new products in such a way that produced WebObjects. If they are responding to a bigger market, (MOXS + YB/NT) they advance their software further. If they leave the NT market, then they lose the innovative small developers who will always prefer an open source solution to a closed shop. I think Apple's expectation that developers will always follow whereever they may lead (excessive advocacy of Java, expectation that NT deployment requirements will just go away) is misplaced. Maury's comments of the limitations of YB/NT are well taken, but none of these are problems that couldn't be fixed with an extra developer on the YB/NT team addressing such typical developer bug reports. -- Leigh Smith leigh@tomandandy.com (MIME) tomandandy +1-212-334-0421 (W) +1-212-334-0422 (F) 89 Greene St. New York, NY 10012, USA http://www.cs.uwa.edu.au/~leigh Microsoft - What do you want to re-install today? From cmh at greendragon.com Sat Sep 18 12:55:47 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:18 2005 Subject: EOF licensing question Message-ID: I know that deployment licensing for WebObjects can cost quite a bit. But what if you're not planning on using the WebObjects part, just EOF or Foundation? I'm planning the development of a three-tier system involving a large database, some form of application server talking to the database and implementing business logic, and lots of clients that communicate with the application server via a custom TCP-based protocol. (Sorry, I can't go into any more specifics on the clients, except to say that using web browsers or HTTP -- and therefore WebObjects -- really isn't a solution in my case.) What I'm curious about is the possibility of developing the application server with Foundation and EOF on Mac OS X Server and then deploying on either Solaris or HP-UX using the WebObjects deployment system. Does it sound like a reasonable plan? What kind of issues might I run into? Is anyone who's done something like this before willing to share some wisdom? I can take responses off-list and summarize. From mail2agp at zeelandnet.nl Sat Sep 18 13:52:29 1999 From: mail2agp at zeelandnet.nl (R.H van Amerongen) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <011101bf01fb$e4572a40$9d0eca11@verbosity.apple.com> References: <011101bf01fb$e4572a40$9d0eca11@verbosity.apple.com> Message-ID: > >Not everyone wants to use ObjC however, True, but a lot of them are using it. You can also turn it around. Not everyone wants to use Java, CF, carbon or whatever however. The rest of us will use ObjC. > particularly since it won't be >available from Apple on Mac OS [and before this statement excites >everybody, that's just a strategic decision that Apple has made, Do they have a reason for this, or is this another..... >which I >have nothing to do with]. Happy to read this. (Private/) Are you not tired off this strategic decision stuff from Apple (/Private) >Some people are upset with lack of Pascal on Mac >OS X, too. if this is the excuse what Apple is using, I find it very cheap. > >Chris Kane >Apple Computer, Inc. Ren? van Amerongen A still believing in -ObjC Cocoa- Developer. BUT when Apple.... I won't use Java with carbon, takes to much time to get something finished. From cmh at greendragon.com Sat Sep 18 14:08:37 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: References: Message-ID: At 1:59 PM -0700 9/18/99, R.H van Amerongen wrote: >>particularly since it won't be >>available from Apple on Mac OS >Can you explain more about this. What are the options. Where can we >find the ( again changed) strategies? Will it still be possible to >build Cocoa Apps? Chris means Objective-C won't be available from Apple on any Mac OS less than Mac OS X (i.e. on MacOS 9.0, 9.1, 9.5, 9.6, etc.). From buck.erik at mcleod.net Sat Sep 18 14:45:23 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: EOF licensing question Message-ID: <002301bf021f$1d5ae0c0$7e2310d0@elcry> > >What I'm curious about is the possibility of developing the >application server with Foundation and EOF on Mac OS X Server and >then deploying on either Solaris or HP-UX using the WebObjects >deployment system. Does it sound like a reasonable plan? What kind >of issues might I run into? Is anyone who's done something like this >before willing to share some wisdom? > My company has done this and I strongly caution you against it. I know the lure of Foundation/AppKit/EOF is very strong for just the type of application you describe. Resist it. Regardless of technical merit, Apple is determined to prevent you from succeeding. The only viable deployment option, as you suggested, is WebObjects licensing. The pricing, availability, unreliability of the vendor, and the vendor's stated hostility to the idea, should be more than enough reason to tighten your belt and look for another solution. If Apple came right out and stated that there is a 50% chance that one year from now there will be no deployment option at any price, would you continue the project ? Well, I estimate the chances are more like 80 % against. Furthermore, licensing WebObjects for deployment is not reasonable now. Run away as fast as you can. I could have saved millions and a lot of pain, hardship, and reputation if I had fled 3 years ago. From ckane at apple.com Sat Sep 18 14:45:23 1999 From: ckane at apple.com (Chris Kane) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <001301bf021f$1c6cd740$9d0eca11@verbosity.apple.com> R.H van Amerongen writes: >> particularly since it won't be >>available from Apple on Mac OS > Can you explain more about this. What are the options. Where > can we find the ( again changed) strategies? Will it still be possible > to build Cocoa Apps? That's "Mac OS less than 10" in case anybody missed that. Metrowerks may have an ObjC compiler and runtime on Mac OS now, I don't know. It will still be possible to build Cocoa apps on Mac OS X. Apple has not stated any release in which this will no longer be the case, if any. Developer-oriented strategies are communicated at WWDC primarily. Sometimes in the past this information is available afterwards at developer.apple.com, but I've forgotten what my Apple ADR ID is, and searching w/o logging in didn't turn up anything obvious quickly. >>Some people are upset with lack of Pascal on Mac >>OS X, too. > if this is the excuse what Apple is using, I find it very cheap. It's not an excuse, it's just an observation on my part. Chris Kane Apple Computer, Inc. From bwebster at owlnet.rice.edu Sat Sep 18 15:43:05 1999 From: bwebster at owlnet.rice.edu (Brian Webster) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation Message-ID: <199909182244.RAA07692@marsh.owlnet.rice.edu> Michelle Buck wrote: > Seriously, Why can't OS 9 have the portions of the FoundationKit that do not > require real operating system feature. Is Carbon with CF going to be > available for OS 9 ? Legacy Mac developers can not use any OS "feature" > that is not supported in OS 9 ? (Isn't that what is being claimed ?) What > about features that are not in OS8 ? All of Carbon, including CF, will be able to run on MacOS 8.x with CarbonLib installed and MacOS 9 built-in. CarbonLib will include all the stuff in Carbon that's not in the MacOS 8 API, like CF, Navigation Services, etc. I really don't see any incentive to implement a whole ObjC runtime on MacOS 9 just for a few Foundation classes which probably won't be used by anyone since basically all of developers for MacOS 9 will be coming from the Toolbox, not Yellow Box. > > I'll stick to my guns on this. Legacy C code must be changed to take > advantage of CF. Why change it to a half OO procedural API when for the > same effort it could be changed to the real thing ? Hell, the Foundation > classes could be wrapped in silly simple C APIs with typedefs to hide the > class names. What would be the point ? CF seems to be creating something > similar to function wrapped classes by writing everything from scratch and > introducing new bugs and closing the door of future use of OO and a dynamic > runtime. One thing to consider about CoreFoundation is the "toll free bridging" that Apple pointed out at WWDC, how you could call CF functions on NS objects and send NS messages to CF objects. Perhaps the underlying CF data structures are ObjC objects after all. That way, if the data structure pointed to by a CF reference is passed to objc_msgSend, it can do the method lookup as normal via the isa pointer. Then whatever Foundation class is being used can call the appropriate CF implementation. CFxxx() calls would just ignore the isa pointer entirely and just do their thing. > > It should be possible to compile any Carbon application with Apple's > Objective-C compiler without any code modification. If you do not use > object Objective-C features, you have C. If you are going to modify the > previously mentioned Carbon application in order to use CF, why not modify > it to use Foundation instead ? I suspect the answer is because Apple is > dropping Foundation and all other YellowBox stuff in the near future. If Apple made Carbon developers use Foundation classes in order to get CF's functionality, they would be forced into using a compiler that supports ObjC, which is not a wise move on Apple's part since they want to make it as easy as possible for current developers to use Carbon. As for wrapping Foundation in C APIs, this would probably be significantly harder to do well than wrapping Foundation around CF, which is what they're doing. And I _really_ don't see how Apple could possibly just "drop" Foundation. For one, according to the Apple slides from WWDC, PDO is being used as the core for all interapplication communication, including AppleEvents. There a nice little illustration showing abstraction layers between the Classic, Carbon, and Cocoa AppleEvent mechanisms and PDO, which handles the details of passing info between all three. I'm sure there are other examples supporting this too. -- Brian Webster bwebster@owlnet.rice.edu http://www.owlnet.rice.edu/~bwebster/ From phr at projectcenter.ch Sat Sep 18 16:44:26 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <001301bf021f$1c6cd740$9d0eca11@verbosity.apple.com> References: <001301bf021f$1c6cd740$9d0eca11@verbosity.apple.com> Message-ID: <9909182344.AA01257@projectcenter.ch> You wrote: > Metrowerks may have an ObjC compiler and runtime on Mac OS > now, I don't know. It will still be possible to build Cocoa apps on > Mac OS X. Apple has not stated any release in which this will no > longer be the case, if any. I am not a Mac OS guy, so please forgive my stupdi question: which compiler does Apple use for Mac OS 9? Is this gcc(egcs) as well?! sweet dreams, Phil --- Philippe C.D. Robert http://www.nice.ch/~phip From rcfa at cubiculum.com Sat Sep 18 17:30:20 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <011101bf01fb$e4572a40$9d0eca11@verbosity.apple.com> References: <011101bf01fb$e4572a40$9d0eca11@verbosity.apple.com> Message-ID: <199909190030.UAA05122@wenix.cubiculum.com> > Why change your legacy C code at all? If it works, don't bother. If you > want to enhance it with new features in CF, then yes, you have to use CF. > If you're enhancing it, you can equally choose to use Foundation instead > and make it (or parts of it) ObjC code. You can do both. Well, that's where the cat bites its tail: we started this whole discussion with the premise that certain APIs will *not* be available in ObjC, and thus we will be forced to deal with a suboptimal pseudo-OOP API called CF. If we really could do either, and had CF equivalency in FoundationKit then this entire discussion would be largely moot. So, no, we can't do both. We are forced to use CF, XML parser being one example. Only because of this situation came up the question as to why not push OOP over procedural APIs, rather than vice versa. Apple can only survive in its niche and expand on it by being a leader, not a follower. CF as good as it may be, doesn't show much leadership, it's at best a refinement of old-style programming, it is better, but not fundamentally different from Win32. FoundationKit and AppKit are, which is why there should be an emphasis on that if technology leadership is a goal. Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From bierman at apple.com Sat Sep 18 17:50:39 1999 From: bierman at apple.com (Peter Bierman) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <9909182344.AA01257@projectcenter.ch> Message-ID: At 4:50 PM -0700 9/18/99, Philippe C.D. Robert wrote: >You wrote: >> Metrowerks may have an ObjC compiler and runtime on Mac OS >> now, I don't know. It will still be possible to build Cocoa apps on >> Mac OS X. Apple has not stated any release in which this will no >> longer be the case, if any. > >I am not a Mac OS guy, so please forgive my stupdi question: which compiler >does Apple use for Mac OS 9? Is this gcc(egcs) as well?! Apple uses Mr.C, available for free with the MPW tools on apple's webpages. Apple also uses Metrowerks' compilers for various subprojects. Mac OS X is built almost entirely with egcs. -pmb -- "UNIX shells in Mac OS X should be unneeded but functional... and have the same installed base as MPW." From buck.erik at mcleod.net Sat Sep 18 18:54:30 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> Wait wait wait! 1) The stated reason for the existence of CF was to bring some portion of Foundation to a much broader audience including teams within Apple. 2) Every team within Apple has access to Objective-C and Java 3) Every developer for MacOSX client will have access to Objective-C and Java 4) Porting part of Foundation to OS 8/9 can not be harder than writing CF as new C code from scratch 5) developers with legacy code including Apple MUST modify the code to use CF Why not modify the code to use Foundation instead of CF ? The only reason I can think of is that one of the premises above is false. Assuming Apple is not irrational, which of the premises above are false ? I do not think numbers 4 or 5 are false. I will grant number 1 on face value. If any of numbers 2 or 3 are false, Apple is clearly dumping Cocoa completely. I do not object to C APIs. We have always had C APIs in NeXTstep/Openstep. The part that scares the shit out of me is that the only explanation I can think of for CF is an intention to drop Cocoa. Will there be ANY Cocoa dependencies in OSX Client ? I can not think of any. Given the 100% rate of promise breaking regarding Openstep technology to date from Apple, why does anyone (even Apple employees) think Cocoa will ship or be supported by Apple for any platform ? If they can kill it for Windows where people already depended on it, they can kill it for Mac where nobody does. Why won't they ? Because they promised ? Ha Ha ha ha ha ha Because they are dependent on it ? They are working as fast as possible to remove any such dependency. Because they think it is better...the future ? Then why aren't they using it themselves ? I can think of lots of reasons completely consistent with Apple's behavior of the past few years for killing Cocoa. Adobe doesn't want it. Mac developers don't want it or don't know what it is. It is written in Objective-C which Apple has decided they don't like. They don't have the resources to develop it. I am sure we can all think of more reasons. I defy anyone to give any plausible reason for Apple to ship Cocoa libraries on OSX other than that they promised to. From rcfa at cubiculum.com Sat Sep 18 20:36:09 1999 From: rcfa at cubiculum.com (Ronald C.F. Antony) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> References: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> Message-ID: <199909190336.XAA05535@wenix.cubiculum.com> > I defy anyone to give any plausible reason for Apple to ship Cocoa libraries > on OSX other than that they promised to. A plausible reason, in the face of rational thought, would be it's better. However given the Java vs. ObjC outcome, we already know that technical merit is not relevant. Remains essentially WebObjects and EOF which depend on it. So the question is, does Apple consider these products fundamental enough or are they considered peripheral in regards to the "Apple is a hardware company" strategy? If WO is considered dispensable, now that Apple is making a healthy profit with hardware again, then they might just drop/sell WO/Cocoa. Otherwise it's here to stay in one form or another. Of course if OpenStep users still will be able to recognize it, is a different matter. Since the mass of Mac programmers never used it, and since these are the only people that seem to matter to Apple, they might as well use CF w/ Java to do an entirely Java-centric thing, replacing most of Foundation/AppKit with native Java stuff like JFC, Swing, etc. and doing some EOF mock-up CF and providing a Java interface to it. That would of course be thoroughly disgusting. The only good thing is, as Apple deconstructs OpenStep, GNUStep is becoming useable. So by the time Apple is done unraveling OpenStep, GNUStep is ready for prime time. Makes me wonder why Apple doesn't just simply donate the entire code to Darwin... Ronald ============================================================================== "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." G.B. Shaw | rcfa@cubiculum.com | NeXT-mail welcome From infinity at ntwrks.com Sat Sep 18 21:02:55 1999 From: infinity at ntwrks.com (Sam Krishna) Date: Thu Nov 3 14:42:18 2005 Subject: Broken HTML mail digest Message-ID: <199909190353.XAA06943@netcube.ntwrks.com> Guys, It seems as if the HTML digest is broken. I tried to find some posts later than 8/27/1999 and macosx-dev wasn't showing them. I also verified this problem on webobjects as well. Thanks, Sam From sanguish at digifix.com Sat Sep 18 21:13:12 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> References: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> Message-ID: <199909190413.AAA09287@digifix.com> Erik M. Buck wrote: > Wait wait wait! > > 5) developers with legacy code including Apple MUST modify the code to use > CF > > Why not modify the code to use Foundation instead of CF ? > Is it more straight forward to re-implement the Foundation functionality as CF (with the increases in speed that Chris had alluded to) and then have Foundation use that... or is it more straight forward to move foundation, with all its dependcies on Objective-C, and Objective-C, and then try and convince Mac developers (the entrenched that is) to use it? If you go back and read what Chris said about formatting certain passed vars using Blah instead of Blah *, you'd see the problems that they're up against on the Mac side. Its not about providing the best APIs (YB), or the most common sense method of going from task A to completion.. but writing what these developers will USE. I've seen ALOT of reticence about even using CF from Mac developers.. they don't understand why Apple just isn't adopting STL. Re-implementing the Foundation functionality as C, and then wrapping it is probably the best way to go given the varied circumstances that have to be dealt with (little of which appears to be technological).. From marcel at metaobject.com Sun Sep 19 00:21:58 1999 From: marcel at metaobject.com (Marcel Weiher) Date: Thu Nov 3 14:42:18 2005 Subject: EOF licensing question Message-ID: > From: Chris Hanson > > What I'm curious about is the possibility of developing the > application server with Foundation and EOF on Mac OS X Server and > then deploying on either Solaris or HP-UX using the WebObjects > deployment system. Does it sound like a reasonable plan? What kind > of issues might I run into? Is anyone who's done something like this > before willing to share some wisdom? At my previous company, we built our product using gcc + libFoundation with great success on Solaris. The real bummer is EOF, but for many, many apps an almost trivial DB-layer ( Objects <-> Dicts <-> DB ) does the trick just (well, almost) as well, sometimes even better. Marcel From richard at brainstorm.co.uk Sun Sep 19 01:00:14 1999 From: richard at brainstorm.co.uk (Richard Frith-Macdonald) Date: Thu Nov 3 14:42:18 2005 Subject: EOF licensing question In-Reply-To: References: Message-ID: <9909190800.AA01062@tiptree.brainstorm.co.uk> On Sun, 19 Sep 1999 00:23:34 -0700 (PDT), marcel@metaobject.com wrote: > > From: Chris Hanson > > > > What I'm curious about is the possibility of developing the > > application server with Foundation and EOF on Mac OS X Server and > > then deploying on either Solaris or HP-UX using the WebObjects > > deployment system. Does it sound like a reasonable plan? What kind > > of issues might I run into? Is anyone who's done something like this > > before willing to share some wisdom? > > At my previous company, we built our product using gcc + > libFoundation with great success on Solaris. The real bummer is EOF, > but for many, many apps an almost trivial DB-layer ( Objects <-> > Dicts <-> DB ) does the trick just (well, almost) as well, sometimes > even better. I can second this - my company runs distributed objects based systems supporting tens of thousands of users using the gstep-base library on GNU/Linux, and Unixware (we used to use Solaris/sparc, but GNU/Linux/intel eventually began to win on price, and reliability is getting close). We use a simple wrapping of a lightweight C database library for simple database work. However - if you really need the functionality of EOF, I don't think you have an alternative to the Apple software (unless you want to rewrite/upgrade the existing GNUstep EOF-1 clone yourself). Unless you can get a legally binding comittment from Apple to provide EOF for deployment of your product - perhaps you'd better look at something different. From marcel at metaobject.com Sun Sep 19 01:28:48 1999 From: marcel at metaobject.com (Marcel Weiher) Date: Thu Nov 3 14:42:18 2005 Subject: EOF licensing question In-Reply-To: Message-ID: > From: Richard Frith-Macdonald > > I can second this - my company runs distributed objects based systems > supporting tens of thousands of users using the gstep-base library on > GNU/Linux, and Unixware (we used to use Solaris/sparc, but GNU/Linux/intel > eventually began to win on price, and reliability is getting close). Actually, we also started out using gstep-base, but then switched to libFoundation, which we found much cleaner. YMMV. Marcel From richard at brainstorm.co.uk Sun Sep 19 01:48:46 1999 From: richard at brainstorm.co.uk (Richard Frith-Macdonald) Date: Thu Nov 3 14:42:18 2005 Subject: EOF licensing question In-Reply-To: References: Message-ID: <9909190848.AA01123@tiptree.brainstorm.co.uk> On Sun, 19 Sep 1999 10:28:48 +0200, marcel@metaobject.com wrote: > > From: Richard Frith-Macdonald > > > > I can second this - my company runs distributed objects based systems > > supporting tens of thousands of users using the gstep-base library on > > GNU/Linux, and Unixware (we used to use Solaris/sparc, but > GNU/Linux/intel > > eventually began to win on price, and reliability is getting close). > > Actually, we also started out using gstep-base, but then switched to > libFoundation, which we found much cleaner. YMMV. I guess this was a some years back then :-) But I'm biased - having rewritten/cleaned-up a lot of the gstep-base code. >> However - if you really need the functionality of EOF, I don't >think you have >> an alternative to the Apple software (unless you want to >rewrite/upgrade the >> existing GNUstep EOF-1 clone yourself). > >This is starting to be a real option, though I'd want to concentrate >on the stuff I really need. There seems to be quite a bit in EOF >that we don't really need. Yes - I didn't mean to imply that rewriting/upgrading the GNUstep database stuff was an impossible option. In fact - a sensible business plan for anyone needing EOF functionality would probably be to develop for Apple EOF, while devoting a couple of developer-days per week to the GNUstep database stuff. Then, if Apple drops EOF in a years time, you can deploy using GNUstep, and even if they don't, your time wouldn't be wasted since you would be able to deploy on GNUstep as well as Apple. Certainly, for non-gui apps, nobody really needs to fear that Apple will drop the technology any more - since only a little investment is needed to ensure that the free alternatives are viable in the areas you need them. All the fear and uncertainly about the direction Apple is taking with YB seems to me rather out of proportion - even the GNUstep gui stuff could be made usable for commercial apps within a year if a few companies put a couple of people to work on them. From bruce at iterative.com Sun Sep 19 01:52:48 1999 From: bruce at iterative.com (Bruce Fancher) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> Message-ID: <000201bf027c$5979fbb0$02e9e90a@deadlord> I don't think that Apple is planning on dropping Cocoa completely for the simple reason that why would they bother to rename YB and announce it if they were planning to drop it? That would just be stupid. I don't think that it'll ever ship on any platform other than the Mac, but I don't think they'll kill it completely. From buck.erik at mcleod.net Sun Sep 19 01:55:18 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <003801bf027c$b34afea0$672a10d0@default> > If you go back and read what Chris said about formatting certain >passed vars using Blah instead of Blah *, you'd see the problems that they're >up against on the Mac side. Its not about providing the best APIs (YB), or >the most common sense method of going from task A to completion.. but writing >what these developers will USE. > > I've seen ALOT of reticence about even using CF from Mac >developers.. they don't understand why Apple just isn't adopting STL. > > Re-implementing the Foundation functionality as C, and then wrapping >it is probably the best way to go given the varied circumstances that have >to be dealt with (little of which appears to be technological).. > > Wow. Both Scott Anguish and Chris Hanson have given very good explanations for why CF must exist and why everything is being rewritten. New features that Apple wants developers to use must also be C APIs. So I reiterate my challenge. Why will Apple ship Cocoa libraries for OSX client ? The one answer I will not accept is because they promised to. Doesn't this look like the end of Openstep technology, dynamic OO, and Objective-C in particular ? From phr at projectcenter.ch Sun Sep 19 03:05:44 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> References: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> Message-ID: <9909191005.AA00367@projectcenter.ch> You wrote: > If any of numbers 2 or 3 are false, Apple is clearly dumping Cocoa > completely. Wait a minute, you seem to be a little bit in rage (well, I do understand the reasons for it), but how can Apple dump Cocoa whithout dumping WOF? This isn't possible, so I don't think your assumptions are correct, since WOF *is* good money for them! sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From rvamerongen at zeelandnet.nl Sun Sep 19 04:38:50 1999 From: rvamerongen at zeelandnet.nl (R.H van Amerongen) Date: Thu Nov 3 14:42:18 2005 Subject: NSString, Cstring and string Message-ID: Hi All, When I read in the description about cString (NSString) I read that the returned cString is automically freed. But it doesn't say when ( when automatic deallocation takes place). Is the used, NSString his private temp subclass, the owner and will it released when the owner is released or....? Ren? From dave.pl at ping.at Sun Sep 19 06:05:19 1999 From: dave.pl at ping.at (Dietmar Planitzer) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <199909191301.PAA13828@endjinn.Austria.EU.net> ---------- >From: Scott Anguish > Is it more straight forward to re-implement the Foundation >functionality as CF (with the increases in speed that Chris had alluded to) >and then have Foundation use that... > > or is it more straight forward to move foundation, with all its >dependcies on Objective-C, and Objective-C, and then try and convince Mac >developers (the entrenched that is) to use it? > > If you go back and read what Chris said about formatting certain >passed vars using Blah instead of Blah *, you'd see the problems that >they're >up against on the Mac side. Its not about providing the best APIs (YB), or >the most common sense method of going from task A to completion.. but >writing what these developers will USE. Sure, we MacOS developers - I've been one for more than 6 years now - are such idiots that we aren't able to deal with pointers; constants if they don't come with an inital 'k' or OOP in general. And, dynamic languages with a not so usual syntax like ObjC are clearly invented by the devil himself and thus naturally bad for us. It really seems that some Apple employees have done a very good job in presenting us MacOS developers to the ex-NeXTs in a way that makes it easy for Apple to misuse us as a cheap pretext whenever Apple changes their plans regarding Cocoa technology (MacOS developers don't want ObjC... they don't want Foundation... blah, blah, blah...). Look, the situation for a MacOS developer is actually quite easy to understand. His customers - the ones who pay his bills - have MacOS 7, 8 and soon 9 installed - they don't run MacOS X Server or MacOS X and also not WinXX or OPENSTEP. Thus, if a MacOS developer should be able to use Cocoa technologies in his programs, then Apple has FIRST to make sure that all this stuff runs on the OLD MacOS versions - a thing called "back-porting new technology". Now, while Apple has obviously no problems with back-porting things like Carbon, QT or Navigation services - it seems that at the same time it is sheer impossible for Apple to do the same with Cocoa. But if Cocoa is not available for the old MacOS line and a MacOS developer wants to use this stuff in his apps, then he is faced with a dilemma (well he's been facing that for two years now): Either he switches to Cocoa and abondens his current customer base - what a ridiculous thought, or he would have to maintain at least two build targets just for the MacOS market (this would also required him to juggle around with two different MacOS systems) - another ridiculous thought. While I see good reasons why porting of the whole Cocoa stuff to the old MacOS is not reasonably possible, I can not see any reason why it should be impossible to port at least the ObjC runtime, egcs compiler (would fit perfectly into the CLI based MPW environment) and the Foundation framework. If Apple would have done just this 1 or 2 years ago, the situation would be completly different today, because it would have enabled MacOS developer to actually USE this stuff in their apps and SHIP such apps to their customers, rather than listing to theoric discussions of how much better the OpenStep stuff would be than the MacOS Toolbox. But there is another problem that would prevent MacOS developers from adopting Cocoa even if it would ship on the old MacOS line: It's not complete enough. Sure the OpenStep frameworks are one of the best object frameworks that I've ever seen, but they are not perfect and simply not complete. It's not only about QT functionality that is missing (I find it very sad that Apple hasn't carried NEXTTime and RenderMan over from OPENSTEP), it's also about severe weaknesses in things like the AppKit text system. It has big problems with big documents (did you ever open an 11MB disassembly listing with TextEdit ? Or try to open a 4100 line document in PB - it takes for ever), it needs much to much memory or it doesn't support things like footers or headers. This is problematic because adding these functionality as an afterthough is hard to do - to say it in a friendly way. Another weak area is Unicode support in Foundation. No way to find out the combining class of a character, or it's script code, numeric value, bidirectional behavior or mirrored attribute - again very sad because the exact definition of these attributes is one of the most important advantages of using Unicode rather than another encoding scheme. I don't even want to speak about missing features like normalization (K, KC, D, KD) or localized casing (Greek and Turkish require some special handling). These are just a few examples - I could come up with a lot more if someone is really interested... > I've seen ALOT of reticence about even using CF from Mac >developers.. they don't understand why Apple just isn't adopting STL. Yes, and I fully agree with them. The last thing that the world was waiting for is yet another container library which doesn't even support inheritence or at least some parameter validation. > Re-implementing the Foundation functionality as C, and then wrapping >it is probably the best way to go given the varied circumstances that have >to be dealt with (little of which appears to be technological).. Yeah, but this solution has some big disadvantages: 1) It occupies development and debugging time that could have been spent on improving Foundation & AppKit itself. 2) As a new piece of code, it also introduces new chances for bugs - even in Foundation which had years of testing behind it - now thrown away. 3) It adds overhead to Foundation, makes it slower and thus even more unlikely that it will get adopted by MacOS developers in the future. In any way, it's good to know that there will be at least two ways out of the Apple Cocoa chaos :) Regards, Dietmar Planitzer PS: Just a thought - but wouldn't this discussen be more appropriate for the (seemingly dead) MacOS X Talk list ? From holger at wizards.de Sun Sep 19 07:53:20 1999 From: holger at wizards.de (Holger Hoffstaette) Date: Thu Nov 3 14:42:18 2005 Subject: NSString, Cstring and string In-Reply-To: References: Message-ID: <199909191453.QAA14686@ragnarok.en.uunet.de> >When I read in the description about cString (NSString) I read that >the returned cString is automically freed. But it doesn't say when ( Right. >when automatic deallocation takes place). Is the used, NSString his >private temp subclass, the owner and will it released when the owner >is released or....? Not quite - this method (IMHO) combines the worst aspects of Foundation with the worst aspects of C and is one good example why many people are so turned off of Foundation and YB. It's also a terrible potential performance hole without saying so in the documentation. What happens when you call -cString is that the receiver gives you a (const char*) that points to a block of memory which contains the NSString's contents in the +defaultCStringEncoding. Obviously, there are two cases to consider - what happens when the string already is in the default encoding (e.g. simple ASCII) and what happens when it is not just plain ASCII and must be converted. In the first case, NString will simply return the pointer to its internal buffer (hence the 'const': you wouldn't want to poke around in an object's guts). This buffer will obviously go away when the instance is deallocated. In the second case, the NSString's Unicode contents have to be converted into a temporary buffer. The regular way to create an object with a temporarily floating ownership is the autorelease mechanism - unfortunately one cannot autorelease a char*, so what you really get is a pointer into some piece of autoreleased memory that will go away at an (seemingly) unspecified point of time in the future. Great, eh? So, you'll have to malloc another buffer yourself and then strcpy() to it if you need the string for a longer time. Or just hope that this invisible buffer won't be deallocated while you're diddling around with the pointer; chances are that it'll be valid until the end of the current event loop. However, the 'documentation' states that this is not guaranteed. Recommendations: 1) Don't use -cString. 2) Use getCString: to get the characters into a buffer which you have allocated and under lifetime control. Either on the heap or on the stack by using alloca([theString -cStringLength]) for a buffer that is needed only for your method's scope. Since this will always copy stuff, it will be slightly slower when you're only dealing with simple strings, but provides a more consistent usage pattern when you come across the occasional non-ASCII string. In that case you only pay for two cheap method calls (-cStringLength and -getCString:, which does a memcpy()) instead of 'sometimes' autoreleased, hidden buffer creation and content copying. IMHO, of course. Regards, Holger From dyoung at vviuh221.vvi.com Sun Sep 19 08:14:44 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) References: <199909191301.PAA13828@endjinn.Austria.EU.net> Message-ID: <9909191514.AA21485@vviuh221.vvi.com> Dietmar, > Or try to open a 4100 line document in PB - it takes for ever) Don't know exactly what you are doing, but ... You may want to try turning indexing off. Indexing has nothing to do with the NSText class. > big problems with big documents (did you ever open an 11MB .. To tune NSText objects to a particular problem please consult the documentation. TextEdit (a tool written in Java depending on what version you are using) is not necessarily the best tool choice for the problem at hand. > footers or headers Embed the NSTextView in a view hierarchy, define some page-oriented parameters and add footer/header view classes. > ... are one of the best object frameworks ... but they are not perfect > and simply not complete. Good grief, when did you ever get perfection in anything? > ...even more lame excuses, excuses ... Yes, everything can be better and have more options, be simpler to use, be more complex, be better supported, cost less or be free, etc. etc. So what. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From dyoung at vviuh221.vvi.com Sun Sep 19 08:33:49 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:18 2005 Subject: NSString, Cstring and string References: <199909191453.QAA14686@ragnarok.en.uunet.de> Message-ID: <9909191533.AA21521@vviuh221.vvi.com> > Not quite - this method (IMHO) combines the worst aspects of Foundation > with the worst aspects of C and is one good example why many people > are so turned off of Foundation and YB. It's also a terrible > potential performance hole without saying so in the documentation. Funny, I think the idea of -cString is intuitive and a very nice way to mix traditional ansi c temps with obj-c autoreleased temps. The only thing really lacking is the exact description of the implementation and a little better description of the different implementation cases and what they mean regarding "released at a future date". Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From sebestyen_g at accentcomm.com Sun Sep 19 09:24:15 1999 From: sebestyen_g at accentcomm.com (Gabor Sebestyen) Date: Thu Nov 3 14:42:18 2005 Subject: ANN: NE2K 0.92 beta available now Message-ID: <37E50E2E.F27DBD58@accentcomm.com> As I promised yesterday, NE2K 0.92 beta, the one-and-only popular NE2000 compatible ethernet adapter driver is now available at: ftp://ftp.accentcomm.com/pub/downloads/Rhapsody/ With NE2K you can use cheap PCI based NE2000 based PC Ethernet card in your Macintosh, MacOS X Server will accept it! For more information please read the info page at ftp://ftp.accentcomm.com/pub/downloads/Rhapsody/NE2K-0.92b.html WARNING! This product in its current state is unfinished and may contain bugs therefore suggested mainly for test purposes. USE AT YOUR OWN RISK! Best regards, Gabor Sebestyen -------------------------------------------------------------- A C C E N T C O M M U N I C A T I O N S - We Make Your Web Groove - Tel: (+36 1) 315-0701 http://www.accentcomm.com (+36 1) 343-4440 e-mail: info@accentcomm.com GSM: (+36 30) 951-4334 Fax: (+36 30) 980-0344 -------------------------------------------------------------- From dave.pl at ping.at Sun Sep 19 10:20:12 1999 From: dave.pl at ping.at (Dietmar Planitzer) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <199909191716.TAA26952@endjinn.Austria.EU.net> ---------- >From: David Young >Dietmar, > >> Or try to open a 4100 line document in PB - it takes for ever) > >Don't know exactly what you are doing, but ... I work a lot with tools that automatically generate source code, and as you can see the generated code can get quite large... >You may want to try turning indexing off. Indexing has nothing to do with >the NSText class. Turning off indexing doesn't help much. It still takes at least 5 seconds for PB to open and layout my file on a 8600/200. >> big problems with big documents (did you ever open an 11MB .. > >To tune NSText objects to a particular problem please consult the >documentation. What's that ? Should Apple tell their MacOS X users, whenever they have problems with large text documents and TextEdit, that they should consult the docs and individually tune their TextEdit copy ? Surely not a way to win friends among your customer base. >TextEdit (a tool written in Java depending on what version >you are using) is not necessarily the best tool choice for the problem at hand. I known that the Java TextEdit sucks and honestly, I don't have any idea why someone inside Apple thought that it would be a good idea to switch from the ObjC version to the Java version. However, Java is not responsible for the performance of the text layout machinery largely hidden behind NSLayoutManager. It is also not responsible for the huge cache that NSLayoutManager deploys and it is not responsible for the display performance in general and the storage requirements of the text system in particular. All these problems reside in the ObjC world. Java, on the other side, is responsible for the infinitely long start up time of TextEdit. >> footers or headers > >Embed the NSTextView in a view hierarchy, define some page-oriented >parameters and add footer/header view classes. The display task is the least problem. The bigger one is to get the footer/header info out of a RTF file and to get it back in there without duplicating what's already part of the AppKit. The problem - as I see it - is that though the text system has a RTF parser and generator built in, it doesn't give me a way to overload it's functionality. There is simply no way to set a delegate that would be called whenever the RTF parser runs across a keyword it doesn't understand. The same applies to the RTF generator. >> ... are one of the best object frameworks ... but they are not perfect >> and simply not complete. > >Good grief, when did you ever get perfection in anything? Why don't you comment on the things that I have listed as missing ? This would be more interesting than to start a discussion about the philosphical meaning of the word "perfect". >> ...even more lame excuses, excuses ... > >Yes, everything can be better and have more options, be simpler to use, be >more complex, be better supported, cost less or be free, etc. etc. So what. Yeah, right. Now we're back to square one. NeXT Dev: Why the hell do this lame MacOS developers not instantly switch to Cocoa ? Mac OS Dev: It doesn't run where my customers are. Also, there are some defficiencies that need to be addressed before I can replace my code with the Cocoa provided objects. Regarding the language issue: Well, we've never worked with ObjC and thus it naturally looks strange to us. Though, the syntax isn't the show-stopper here, but the question if Apple clearly and undoubtly embraces it, is (we were burnt once to often to trust on the future of any technology that Apple doesn't use for core components of it's own mainstream OS). NeXT Dev: Lame excuses, all I here from them is lame excusses. Okay, this is a bit of an oversimplification of the situation, but maybe it's easier to understand :) Regards, Dietmar Planitzer From bigdog at bulldogbeach.com Sun Sep 19 12:07:51 1999 From: bigdog at bulldogbeach.com (Gary Teter) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: > From: "Dietmar Planitzer" > >From: David Young > >> big problems with big documents (did you ever open an 11MB .. > > > >To tune NSText objects to a particular problem please consult the > >documentation. > > > What's that ? Should Apple tell their MacOS X users, whenever they > have problems with large text documents and TextEdit, that they should > consult the docs and individually tune their TextEdit copy ? > > Surely not a way to win friends among your customer base. I'm confused as to why this is a problem (coming from the Mac side of things) -- I always thought of TextEdit as a demo app which happened to be useful in many circumstances, like SimpleText. But I would never use SimpleText as a code editor, or a word processor, or anything else important. I have BBEdit and *cough* Word for the big stuff. > >TextEdit (a tool written in Java depending on what version > >you are using) is not necessarily the best tool choice for the > problem at hand. Exactly. Or should I be expecting more from it? Just because the classes are included with the system doesn't mean I have to use them, or that they're perfect for every situation. I can't remember offhand what the equiv to NSTextView is called on the MacOS, but I don't think there are many developers that would use it as the basis for a programmers editor or a word processor. They'd roll their own. (A "third party opportunity" :-) Granted, the trend in operating systems is to subsume more and more functionality as time goes on that would otherwise have been in third-party libraries or home-grown code, but naturally that functionality will be least-common-denominator. For example, while I'm sure Adobe Premiere makes an awful lot of QuickTime calls, I doubt that's all the application consists of. All my dev on OS X has been WO and EOF, so I guess I have a different perspective than the old NeXT-ers. WO and EOF are remarkably feature-complete* for building narrowly-defined classes of applications; they're frameworks built to do just that. I have no such expectations that operating system-supplied classes would (or even should) do as much for me in building an arbitrary desktop application. Or am I wrong about what I should expect to find when I finally get around to using Interface Builder for a real project? (Written in Java, of course. :-) *(I feel obligated to add that WO is "feature-complete" except for, of course, the ability to manipulate generated HTML or the component object graph.... one can only hope for that in 4.5 or 5.0 or whatever.) From sanguish at digifix.com Sun Sep 19 12:21:32 1999 From: sanguish at digifix.com (sanguish) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <199909191918.PAA10770@digifix.com> > > Wow. Both Scott Anguish and Chris Hanson have given very good explanations > for why CF must exist and why everything is being rewritten. New features > that Apple wants developers to use must also be C APIs. So I reiterate my > challenge. Why will Apple ship Cocoa libraries for OSX client ? The one > answer I will not accept is because they promised to. > With all due respect Erik, you're not acting in any sort of a rational manner. Regardless of WHAT is said, you'll believe what you have already chosen to believe. Apple has large numbers of Apps that are part of OS X that rely on Cocoa and Objective-C. All the admin stuff, all the development tools that they themselves are pushing for Carbon development. Finally, Java on its own, using the Java UI libraries isn't enough to produce useable apps. > Doesn't this look like the end of Openstep technology, dynamic OO, and > Objective-C in particular ? > > The sky is falling.. the sky is falling... From sanguish at digifix.com Sun Sep 19 12:28:13 1999 From: sanguish at digifix.com (sanguish) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <199909191925.PAA10794@digifix.com> > > Sure, we MacOS developers - I've been one for more than 6 years now - > are such idiots that we aren't able to deal with pointers; constants > if they don't come with an inital 'k' or OOP in general. And, dynamic > languages with a not so usual syntax like ObjC are clearly invented by > the devil himself and thus naturally bad for us. > Finally a MacOS developer admits it..:-) From dyoung at vviuh221.vvi.com Sun Sep 19 13:31:37 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) References: <19990919200949.21749.qmail@hotmail.com> Message-ID: <9909192031.AA21899@vviuh221.vvi.com> Gary Teter said: > of things) -- I always thought of TextEdit as a demo app which > happened to be useful in many circumstances, like SimpleText. But I Yes, TextEdit demonstrates where the NSText[View] class is applicable and where it is not applicable for the attribute values TextEdit has set for the NSText[View] object. NSText class is great for small to medium text problems like about panels, text entry formatting, and online manuals where the manual is sensibly structured into separate chapters (files). It is not a replacement for Framemaker or the likes; BUT it does go a long ways to solving a lot of problems. However, TextEdit is not the all inclusive demonstration program for the NSText[View] class. For example, You can set a few parameters of NSText class and get a lot more mileage out of it than you see in TextEdit. And subclassing goes even further. "Dietmar Planitzer" said: > Mac OS Dev v.s. NeXT Dev. Do me a favor, cut out that me-you nonsense. I'm a whatever it takes developer. I program in many things including MFC. > Mac OS Dev: It doesn't run where my customers are. Yes, that is a problem. Hopefully OS X, and Apple, will give more incentive. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From buck.erik at mcleod.net Sun Sep 19 13:49:39 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <003101bf02e0$7e25ce80$612a10d0@elcry> It is kind of ironic that I used to impress people by opening an 80 Mb text file with good old Edit.app on a machine with only 32Mb memory. The document would open instantly and display the first page. Scrolling was smooth unless you jumped into the middle of the document. The magic was mapped files (one of the innovations of Mach). Apple is giving people a bad impression by downgrading on every release since NeXTstep 1.0. It was not that long ago that the Mac text facilities were limited to 32K and after that would crash rather than fail nicely. From buck.erik at mcleod.net Sun Sep 19 14:07:43 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <006401bf02e3$04c1ba60$612a10d0@elcry> > With all due respect Erik, you're not acting in any sort of >a rational manner. Regardless of WHAT is said, you'll believe what >you have already chosen to believe. > Fair enough. I suspect that analysis applies generally to any kind of opinion. However, do we know of any application essential to MacOSX relies on Carbon? I have been led to believe that everything essential is being rewritten in C or Java. I should think that if we can agree on anything it is that Apple is NOT pushing Cocoa technology. > The sky is falling.. the sky is falling... > Folks, we have all been through this before. Every time someone said Rhapsody was dead, eventually the sarcastic "sky is falling" would come up. Then again with Display Postscript. Then again with cross platform. Then again with OSX Server general availability and features. How many times do we have to be burned ? Everyone argues that Apple is making decisions in their own best interests and developers be damned. People have even suggested that we look to Apple's interests as a guide to the future. After all, every past Apple decision has been justified that way. So someone please show me how Apple's interests as clearly defined by them benefit from ever releasing Cocoa libraries as a standard part of OSX client. We even have a Mac developer telling us that Objective-C can not be used because Apple is not publicly committed to it and they have been burned too many times in the past. How can we Openstep developers not be just as wary or more ? >From Dietmar Planitzer: "... Regarding the language issue: Well, we've never worked with ObjC and thus it naturally looks strange to us. Though, the syntax isn't the show-stopper here, but the question if Apple clearly and undoubtly embraces it, is (we were burnt once to often to trust on the future of any technology that Apple doesn't use for core components of it's own mainstream OS)." From dyoung at vviuh221.vvi.com Sun Sep 19 15:44:55 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) References: <199909191716.TAA26952@endjinn.Austria.EU.net> Message-ID: <9909192244.AA22050@vviuh221.vvi.com> "Dietmar Planitzer" wrote: > it doesn't give me a way to overload > it's functionality. There is simply no way to set a delegate that > would be called whenever the RTF parser runs across a keyword it > doesn't understand. The same applies to the RTF generator. You may consider parsing in RTF code, embedding custom dictionary tags into the runtime representation of text in the NSText object and then storing the archive of the runtime. That is the ultimate in flexibility since you can store your own classes in the archive stream (serialization) as well as tags and embedded parameters, etc. in the text serialization. Of course, you are then beholden to the curse of Apple's proprietary archiving format. The curse is why we wrote our own (much faster) archiver/serializer that runs and parses on multiple platforms. Thus our documents aren't tied into Apple's proprietary systems. Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From bierman at apple.com Sun Sep 19 15:35:42 1999 From: bierman at apple.com (Peter Bierman) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <01d301bf0241$ea83cfe0$5e2b10d0@elcry> Message-ID: >I defy anyone to give any plausible reason for Apple to ship Cocoa libraries >on OSX other than that they promised to. I am writing Cocoa apps for Mac OS X client. Other people at Apple are writing Cocoa apps. You have been suitably defied. (Other people at Apple are writing Carbon apps too, so don't take this as a "Apple's moving 100% to Yellow!") IMHO, the "Yellow Paranoia" is borne from a small group of developers who are concerned that they will always remain a small group. The best way to prevent this is to evangelize your advantages to new developers... WHEN those new developers have painless access to your toys of choice. Frankly the shouts of "Apple is killing Yellow!" from the Yellow dev community will probably do more harm right now than good. You'll scare the hundreds to thousands of developers who will be picking up Mac OS X client for the first time in the next 12 months. Not everyone jumps into the pool before it's full. Apple's made it clear that both Carbon and Cocoa will be available on Mac OS X Client. The future of Yellow will be determined by adoption in *the future* not the developer base today. If Yellow is so good, then people will use it. I'm a longtime Mac OS developer, and there's a lot about Yellow to like, once you find the time to start learning Yet Another Toolbox. >2) Every team within Apple has access to Objective-C and Java Depends on what you mean by "access". Apple's a big company. If I was an engineer working on Mac OS 9 components, I might have "access" to ObjC, but it might make no sense at all for me to use it for a Mac OS 9 component. -pmb -- "UNIX shells in Mac OS X should be unneeded but functional... and have the same installed base as MPW." From dyoung at vviuh221.vvi.com Sun Sep 19 16:28:59 1999 From: dyoung at vviuh221.vvi.com (David Young) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) References: Message-ID: <9909192329.AA22100@vviuh221.vvi.com> Peter, > IMHO, the "Yellow Paranoia" is borne from a small group of developers who > are concerned that they will always remain a small group. Count me out of that, but ... Uh, Peter need I bring up the YB/Windows debacle? or the FACT that Mac OS (the present Mac OS) ---> IS <--- small by any standard? It would help YOUR credibility if YOUR company actually learned how to BE credible. > The best way to > prevent this is to evangelize your advantages to new developers... WHEN > those new developers have painless access to your toys of choice. Well Peter, it is up to YOUR company to make it painless. We're just on the sidelines waiting for YOU to make it that way. > Frankly the shouts of "Apple is killing Yellow!" Frankly, keep me out of that, I'm having too much fun with Yellow. On the other hand YOUR company could have prevented that by NOT killing (for all practical purposes) Yellow on NT after YEARS of WRITTEN statements to the contrary. > will probably do more harm right now than good. So true! so bring back Yellow on NT now! PROVE YOUR credibility. > If Yellow is so good, then people will use it. Most of the lure of Yellow was its cross platform capability. Without that it has lure of about 5% of the marketplace. If YOU want lure then YOU (Apple) have to prove it, not third party developers. YOU have to go the distance. > You have been suitably defied. With all respect, I think not. If I get e-mail from Steve Jobs with a follow-up legally binding affidavit stating exactly what you mean by that then maybe YOUR statement will have some credibility. > Apple's made it clear that both Carbon and Cocoa will be available on > Mac OS X Client. Make my day! Thanks A Bunch! David Young; VVI-DCS dyoung@vvi.com From bierman at apple.com Sun Sep 19 16:53:25 1999 From: bierman at apple.com (Peter Bierman) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <9909192329.AA22100@vviuh221.vvi.com> Message-ID: First-- I AM NOT APPLE. I work for Apple. I write code for Apple. I do not interact with marketting, developer relations, or legal on any regular basis. I don't speak FOR Apple, I speak ABOUT Apple. And I try to pick my words carefully, but one can never be careful enough, and I know I am walking on thin ice by contributing to this at all, but I feel it's important to convey some perspective to a commuinity that appears to feel underappreciated. >Uh, Peter need I bring up the YB/Windows debacle? or the FACT that Mac OS >(the present Mac OS) >---> IS <--- small by any standard? It would help YOUR credibility if YOUR >company actually learned how to BE credible. Since I've been following Apple, Steve's return to the throne marks the first time I've ever seen Apple following a coherant marketing plan. Keep that in mind, and a lot of the things you see comming from Apple make a lot more sense. Now if I were to GUESS at this plan, with all of the same info available to all of you, I might notice that a commonality is "focus on the stuff we do well, make sure we're doing it well, and send the PROFITS (when we make 'em) to the projects that show the best chances. That means no more QuickDrawGX BS. Or overambitious AOCE. Actually, AOCE is a good template for looking at Yellow. Dumping a fully hatched, *massive*, new set of APIs on your developers and then wondering why no one adopted the whole mess. Yellow's plan looks like someone learned. Make it so you can adopt it bit by bit. Then more people will be adopting parts of it, and a FEW people will go the 9 yards and show the real power. It's about EVOLUTION, not transition. In that sense, Carbon is brilliant. >> The best way to >> prevent this is to evangelize your advantages to new developers... WHEN >> those new developers have painless access to your toys of choice. > >Well Peter, it is up to YOUR company to make it painless. We're just on the >sidelines waiting for YOU to make it that way. And here I am at work at 4:30 on a Sunday afternoon, working on the Mac OS X installer, as I was yesterday, and the day before, and Tuesday from noon until 7am. My point isn't that I'm working like a madman, it's that there's SO MUCH WORK to be done. All of you are developers. You all know how much work actually goes into writing, debugging, polishing, and shipping an App. An OS is all of the same, but an order of magnitude bigger. So there's no one at Apple that wants Yellow development (OR ANY DEVELOPMENT) to be painful! And the solution for that is TIME (god I hate gdb). And Mac OS X Client 1.0 will ship sometime in the future, and *then* a whole bunch of people will start playing with these technologies. People who have heard about them, mostly in the context that the Yellow Community is now talking about it And I tell you, as an outsider to the "Yellow Community" that it is the community who is sending off the "Yellow Is Dead" message, not Apple. You should be getting other developers on board any way you can. Not with big projects, but with shareware, and freeware. Stuff that longtime Mac OS developers can do in their miniscule spare time, that gets them excited about Yellow. Because Client 1.0 will ship, and then they'll want to play with this stuff! And then you'll have that many more people who want Yellow on NT. Or not. Who knows? >> Frankly the shouts of "Apple is killing Yellow!" > >Frankly, keep me out of that, I'm having too much fun with Yellow. On the >other hand YOUR company could have prevented that by NOT killing (for all >practical purposes) Yellow on NT after YEARS of WRITTEN statements to the >contrary. I know ZERO about licensing. I don't know what's holding it up, but I believe that we keep saying we're trying. There's a million explanations, but some of them are more believable than others. You have to decide for yourself based on your info and the source. -pmb -- "UNIX shells in Mac OS X should be unneeded but functional... and have the same installed base as MPW." From sanguish at digifix.com Sun Sep 19 17:30:45 1999 From: sanguish at digifix.com (sanguish) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <199909200027.UAA11764@digifix.com> > Fair enough. I suspect that analysis applies generally to any kind of > opinion. However, do we know of any application essential to MacOSX relies > on Carbon? I have been led to believe that everything essential is being > rewritten in C or Java. > > ProjectBuilder, InterfaceBuilder (both pushed for Carbon development at WWDC) WebObjects MailViewer (pushed by Steve at WWDC as THE mail client for OSX) The network management stuff is all written in Cocoa Infact,I don't think Apple has commited publically to writing ANYTHING in Carbon aside from the Finder. Writing it in JAVA does NOT mean writing it to something other than COCOA. They aren't using SWING for this. These are YB apps. Why do you ignore this so stubornly and repeatedly?? > Folks, we have all been through this before. Every time someone said > Rhapsody was dead, eventually the sarcastic "sky is falling" would come up. You're making insane claims here that things are dead because of what you've been told by others, someone you've already admitted (privately) has a potential grudge. > Then again with Display Postscript. The loss of DPS is NOT a major problem. Seeing what Quartz can do makes DPS look pale, and the ADOBE monkey is off our backs. PS printers are not as common as they were, even in Desktop Publishing.. You'll be able to get $2000 printer quality out of a $200 printer. > Then again with cross platform. Then > again with OSX Server general availability and features. Hmmm. I'm confused.. those who claimed that we weren't going to get OSXS definately were crying wolf. We got it, and at a reasonable price point. I'm curious why you can't get your product in use at your clients site Erik... Apple has in place a license that you can use for Enterprise deployment of the current YB/NT implementation. The price sucks somewhat, but it is doable. Its the commercial deployment that is not in place. From sanguish at digifix.com Sun Sep 19 21:16:41 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: It isn't his fault - Re: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: References: Message-ID: <199909200416.AAA12102@digifix.com> Peter Bierman wrote: > And I try to pick my words carefully, but one can never be careful enough, > and I know I am walking on thin ice by contributing to this at all, but I > feel it's important to convey some perspective to a commuinity that appears > to feel underappreciated. > Yes, we feel underappreciated.. I'm a little concerned that it isn't obvious to everyone at Apple that we're feeling this way. I know I feel this way (as I've said privately to you). [to the group] Peter isn't one to attack here especially on the licensing issue. Nor is Chris Kane... This is a leadership issue. It is the LEADERSHIP that has decided to hang us out to dry on this. It isn't the people who are doing the actual work... The fact that Peter was (foolish) enough to tread here, knowing full well that there is a mine under one foot, and dog-poop under the other is admirable. Foolish.. but admirable.. :-) Infact, licencing discussion really doesn't belong here at all. From phr at projectcenter.ch Sun Sep 19 23:50:19 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: References: Message-ID: <9909200650.AA00352@projectcenter.ch> You wrote: > >I defy anyone to give any plausible reason for Apple to ship Cocoa libraries > >on OSX other than that they promised to. > > I am writing Cocoa apps for Mac OS X client. Other people at Apple are > writing Cocoa apps. You have been suitably defied. (Other people at Apple > are writing Carbon apps too, so don't take this as a "Apple's moving 100% > to Yellow!") Thanks Peter, thanks for your comment! > IMHO, the "Yellow Paranoia" is borne from a small group of developers who > are concerned that they will always remain a small group. The best way to > prevent this is to evangelize your advantages to new developers... WHEN > those new developers have painless access to your toys of choice. The problem is - and remains - that Apple doesn't communicate its longterm or midterm plans, and thus it is not very easy to plan... > Apple's made it clear that both Carbon and Cocoa will be available on Mac > OS X Client. The future of Yellow will be determined by adoption in *the > future* not the developer base today. If Yellow is so good, then people > will use it. I'm a longtime Mac OS developer, and there's a lot about > Yellow to like, once you find the time to start learning Yet Another > Toolbox. Thanks, again...;-) sweet dreams, Phil --- Philippe C.D. Robert FreeBSD/Linux Hacker http://www.nice.ch/~phip From phr at projectcenter.ch Sun Sep 19 23:57:45 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <199909200027.UAA11764@digifix.com> References: <199909200027.UAA11764@digifix.com> Message-ID: <9909200657.AA00357@projectcenter.ch> You wrote: > > Then again with Display Postscript. > The loss of DPS is NOT a major problem. Seeing what Quartz > can do makes DPS look pale, and the ADOBE monkey is off our backs. In fact in some cases it is just better not to have DPS...;-) > Its the commercial deployment that is not in place. We should just get the point that Apple is not NeXT anymore, and thus the customers of the company are different - it is a pitty, but it's up to us to realise and react on this. Apple is not an big Enterprise company and never won't be. Just my $0.02... sweet dreams, Phil --- Philippe C.D. Robert http://www.nice.ch/~phip From sanguish at digifix.com Mon Sep 20 00:04:25 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: <199909200027.UAA11764@digifix.com> Message-ID: <199909200706.DAA12719@digifix.com> > > Its the commercial deployment that is not in place. > > We should just get the point that Apple is not NeXT anymore, and thus the customers > of the > company are different - it is a pitty, but it's up to us to realise and react on this. > Apple is > not an big Enterprise company and never won't be. > All the more reason to put a commercial deployment option into place. It would help us (6400 educational seats for our free app) a bunch. The enterprise option is covered off... it is the non-enterprise situation that has been left hanging.. From Paul_Lynch at plsys.co.uk Mon Sep 20 04:17:40 1999 From: Paul_Lynch at plsys.co.uk (Paul Lynch) Date: Thu Nov 3 14:42:18 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) In-Reply-To: References: Message-ID: <199909201117.MAA02339@srabon.plsys.co.uk> Gary Teter wrote: > I'm confused as to why this is a problem (coming from the Mac side > of things) -- I always thought of TextEdit as a demo app which > happened to be useful in many circumstances, like SimpleText. But I > would never use SimpleText as a code editor, or a word processor, or > anything else important. I have BBEdit and *cough* Word for the big > stuff. And right there is the difference, or perhaps the *advantage* of Cocoa. The applications and classes are robust enough for use in this way. Compare the lines of code count of TextEdit and the equivalent MacOS 8 application. This is why old NeXT people (including those at Apple) are confident that this is a better way of developing. Paul --- Paul Lynch (NeXTmail) paul@plsys.co.uk Tel: (01494)432422 P & L Systems Fax: (01494)432478 http://www.plsys.co.uk/~paul From cybergrog at bigpond.com Mon Sep 20 06:17:16 1999 From: cybergrog at bigpond.com (Greg Hulands) Date: Thu Nov 3 14:42:19 2005 Subject: Should there be a new Mailing List? Message-ID: <13145369517900@domain5.bigpond.com> With the amount of discussion on the topic of Apple's decisions, it would be nice if Omni could create a new list like pissedOfAtApple@omnigroup.com or something similar. Any suggestions? Greg From bentley at crenelle.com Mon Sep 20 09:16:40 1999 From: bentley at crenelle.com (Michael Brian Bentley) Date: Thu Nov 3 14:42:19 2005 Subject: Should there be a new Mailing List? In-Reply-To: <13145369517900@domain5.bigpond.com> References: <13145369517900@domain5.bigpond.com> Message-ID: >With the amount of discussion on the topic of Apple's decisions, it would be >nice if Omni could create a new list like pissedOfAtApple@omnigroup.com or >something similar. > >Any suggestions? > >Greg MacOSX-talk is where this is done, to keep -dev and -admin traffic focused on tech issues. These conversations erupt here now because lots of the participants don't/no longer subscribe to -talk because such traffic was quite high and distracting. So they effectively bring it with 'em. I suggest they move it on over to -talk whenever I see this sort of thing being done on -dev, and someone else hasn't beaten me to it. -m From cmh at greendragon.com Mon Sep 20 09:48:42 1999 From: cmh at greendragon.com (Chris Hanson) Date: Thu Nov 3 14:42:19 2005 Subject: post-install script Message-ID: I'm building an Installer.app package for my application, and I've got a resource that I'd like to put in a special place. What do I need to do to create a post-install script that Installer.app will run? Will its working directory be the .pkg folder? This doesn't seem to be documented with the rest of the package-building docs. From buck.erik at mcleod.net Mon Sep 20 10:01:11 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:19 2005 Subject: The intent behind CoreFoundation (was Using the XML parser) Message-ID: <001601bf0389$be7e14c0$a12910d0@elcry> >>I defy anyone to give any plausible reason for Apple to ship Cocoa libraries >>on OSX other than that they promised to. > >I am writing Cocoa apps for Mac OS X client. Other people at Apple are >writing Cocoa apps. You have been suitably defied. (Other people at Apple >are writing Carbon apps too, so don't take this as a "Apple's moving 100% >to Yellow!") > Thank you. I am defied. I was under the impression that there was no Cocoa work in progress. You have corrected me. From marcel at metaobject.com Mon Sep 20 11:26:51 1999 From: marcel at metaobject.com (Marcel Weiher) Date: Thu Nov 3 14:42:19 2005 Subject: NSString, Cstring and string Message-ID: > Recommendations: > > 1) Don't use -cString. > 2) Use getCString: to get the characters into a buffer which you > have allocated and under lifetime control. Either on the heap or > on the stack by using alloca([theString -cStringLength]) for a > buffer that is needed only for your method's scope. > Since this will always copy stuff, it will be slightly slower > when you're only dealing with simple strings, but provides a more > consistent usage pattern when you come across the occasional > non-ASCII string. In that case you only pay for two cheap method > calls (-cStringLength and -getCString:, which does a memcpy()) > instead of 'sometimes' autoreleased, hidden buffer creation and > content copying. 3. Don't use NSString. Use NSData. At least, don't use it for stuff where you want to byte-munge the contents or act on them with C-string semantics. Use NSData instead, it has the proper semantics for storing bags of bytes. Add easy-to-use compatibility/conversion methods. I personally have also added a 'sub-data' object that (a) refers to a range in a NSData object (b) has both NSData and NSString interfaces. Not perfect, but pretty nice. Marcel From alexnet at gestalt.com Mon Sep 20 15:15:17 1999 From: alexnet at gestalt.com (Alex Molochnikov) Date: Thu Nov 3 14:42:19 2005 Subject: Java bridge for GNUstep Message-ID: <9909202215.AA07808@gestalt.com> A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 468 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990920/f936f99a/attachment.bin From mail2agp at zeelandnet.nl Mon Sep 20 15:24:01 1999 From: mail2agp at zeelandnet.nl (R.H van Amerongen) Date: Thu Nov 3 14:42:19 2005 Subject: NSString, Cstring and string In-Reply-To: <199909191453.QAA14686@ragnarok.en.uunet.de> References: <199909191453.QAA14686@ragnarok.en.uunet.de> Message-ID: Thank you for the suggestions. Ren? From sanguish at digifix.com Mon Sep 20 21:01:57 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:19 2005 Subject: post-install script In-Reply-To: References: Message-ID: <199909210401.AAA15941@digifix.com> Chris Hanson wrote: > I'm building an Installer.app package for my application, and I've > got a resource that I'd like to put in a special place. What do I > need to do to create a post-install script that Installer.app will > run? Will its working directory be the .pkg folder? This doesn't > seem to be documented with the rest of the package-building docs. Depending on your package, you could put this file in its 'special place' using the standard decompression... thats probably the best way.. Where are you trying to put it? I think (last I recall) the DHCP package I created for Mac OS X Server (check on Stepwise or in Softrak) had an post-install script... I also tend to look at an existing package receipt for these kinds of things if they're not documented.. they do get stuck in the Receipts folder... From mmalc at stepwise.com Mon Sep 20 22:33:08 1999 From: mmalc at stepwise.com (mmalcolm crawford) Date: Thu Nov 3 14:42:19 2005 Subject: EOF licensing question Message-ID: <199909210532.XAA14450@www3.stepwise.com> Erik wrote: >>What I'm curious about is the possibility of developing the >>application server with Foundation and EOF on Mac OS X Server and >>then deploying on either Solaris or HP-UX using the WebObjects >>deployment system. Does it sound like a reasonable plan? What kind >>of issues might I run into? Is anyone who's done something like this >>before willing to share some wisdom? >> > My company has done this and I strongly caution you against it. I know the > lure of Foundation/AppKit/EOF is very strong for just the type of > application you describe. Resist it. Regardless of technical merit, Apple > is determined to prevent you from succeeding. The only viable deployment > option, as you suggested, is WebObjects licensing. The pricing, > availability, unreliability of the vendor, and the vendor's stated hostility > to the idea, should be more than enough reason to tighten your belt and look > for another solution. > Given that just about all the major WebObjects deployments are on Solaris, I do not consider it a realistic possibility that the runtime for that OS will be discontinued. As for licensing: I would suggest that you talk with your Apple sales rep; they are known to be flexible, and if you present a reasonable case you will probably be listened to sympathetically. I haven't noticed any "stated hostility". Apple is most definitely *not* "determined to prevent you from succeeding": Apple is, as behoves the company, trying to ensure *it* succeeds. Its path to success may or may not intersect yours. mmalc. From paulrs at lgs-systems.com Tue Sep 21 09:17:30 1999 From: paulrs at lgs-systems.com (Paul R. Summermatter) Date: Thu Nov 3 14:42:19 2005 Subject: Compile Problems on NT Message-ID: <51325BD9D61DD31186A50040F6340DB811D6@NAPOLI> Hey guys, I've been having the weidest problem on NT for the past month or so. We have a couple of machines which are set up with NT and WOF 4.0, but one of them is much faster than the others, so we prefer to use the faster machine for full compiles. However, for some unknown reason, there is one framework which fails to compile on this machine. This problem appeared recently and has been very consistent. I have been unable to figure out why it fails to compile. If I tar up the framework and move it onto another machine, it compiles just fine. My obvious guess as to the problem is that there is a difference in the environment, but I'm not sure what it is. Here is the error that I am getting. It comes at the very end of the compile and appears to be at the stage of stripping the install product. When this happens, sh.exe Dr. Watsons on me: ... libtool: warning: -install_name ignored error in [C:\Apple\Library\Exeuctables/sh.exe] There were 0 arguments begin environment ... I believe what was supposed to happen at this stage was that the .dll was to be copied into the framework's directory in the install dir. The compile actually fails a bit later on when it attempts to move the .dll from the framework's install directory into the Executables directory. The reason I am guessing that this is all happening at the stripping stage is that, even though the full compile fails, the dll is built but is about 10 times larger than the dll which I build on another machine. Again, just this one framework out of about 10 fails to compile on this machine. If anyone has any suggestions as to what might be going on, I would love to hear it. Regards, Paul --- Paul Summermatter LGS Systems, Inc. Medical Computing Division 15 TJ Gamester Ave. Portsmouth, NH 03801-5871 (603) 433-9822 voice (603) 433-9818 fax (603) 498-6935 voice mail From cordero at pros.com Tue Sep 21 09:58:30 1999 From: cordero at pros.com (Manuel Cordero) Date: Thu Nov 3 14:42:19 2005 Subject: Compile Problems on NT Message-ID: <199909211658.AA18041@pros.com> Paul, Do you have the exact same absolute path for the framework tree in both machines? If you don't then the explanation below might solve your problem: I've seen something similar when trying to build frameworks with lots of files in it. As far as I can tell, the command lines get too long to be processed in one shot. Adding a single character to the pathname where you framework is (from C:\tree to C:\tree1) will increase the length of the command line by (at least) 1*number_of_files_in_framework. To get around it, we broke down the large frameworks into several subprojects so you don't run into the command line limitation. Hope this helps. Manuel From server@omnigroup.com Tue Sep 21 11:19:51 1999 Date: Tue, 21 Sep 1999 09:19:39 -0700 (PDT) Errors-To: netsec@omnigroup.com Reply-To: paulrs@lgs-systems.com Originator: macosx-dev@omnigroup.com Sender: macosx-dev@omnigroup.com Precedence: bulk From: "Paul R. Summermatter" To: Multiple recipients of list Subject: Compile Problems on NT X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas X-Comment: To unsubscribe, follow directions at http://www.omnigroup.com/MailArchive/ Content-Type: text/plain; X-Mailer: Internet Mail Service (5.5.2448.0) Mime-Version: 1.0 Hey guys, I've been having the weidest problem on NT for the past month or so. We have a couple of machines which are set up with NT and WOF 4.0, but one of them is much faster than the others, so we prefer to use the faster machine for full compiles. However, for some unknown reason, there is one framework which fails to compile on this machine. This problem appeared recently and has been very consistent. I have been unable to figure out why it fails to compile. If I tar up the framework and move it onto another machine, it compiles just fine. My obvious guess as to the problem is that there is a difference in the environment, but I'm not sure what it is. Here is the error that I am getting. It comes at the very end of the compile and appears to be at the stage of stripping the install product. When this happens, sh.exe Dr. Watsons on me: ... libtool: warning: -install_name ignored error in [C:\Apple\Library\Exeuctables/sh.exe] There were 0 arguments begin environment ... I believe what was supposed to happen at this stage was that the .dll was to be copied into the framework's directory in the install dir. The compile actually fails a bit later on when it attempts to move the .dll from the framework's install directory into the Executables directory. The reason I am guessing that this is all happening at the stripping stage is that, even though the full compile fails, the dll is built but is about 10 times larger than the dll which I build on another machine. Again, just this one framework out of about 10 fails to compile on this machine. If anyone has any suggestions as to what might be going on, I would love to hear it. Regards, Paul --- Paul Summermatter LGS Systems, Inc. Medical Computing Division 15 TJ Gamester Ave. Portsmouth, NH 03801-5871 (603) 433-9822 voice (603) 433-9818 fax (603) 498-6935 voice mail From mmarkowitz at ceiss.org Tue Sep 21 11:08:27 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:19 2005 Subject: Compile Problems on NT Message-ID: Manuel Cordero wrote: > I've seen something similar when trying to build frameworks with lots > of files in it. As far as I can tell, the command lines get too > long to be processed in one shot. We ran into this one too. We moved the dev folder closer to the root of the drive to fix it in our case. The message also had one other thing I've not seen before... error in [C:\Apple\Library\Exeuctables/sh.exe] Shouldn't that /sh.exe be \sh.exe? Or is this something that's happening on my machine too and I just never noticed it? Maury From nhoj at cd.chalmers.se Tue Sep 21 11:36:30 1999 From: nhoj at cd.chalmers.se (nhoj@cd.chalmers.se) Date: Thu Nov 3 14:42:19 2005 Subject: Protocol inheritance -- ConcreteProtocol framework Message-ID: <19990921183631.10858.qmail@glenn> I've just finished a small framework that lets you create protocols with implementations. The framework can be found on my homepage; http://www.cd.chalmers.se/~nhoj. The interface is a bit rugged since there is no matching compiler support, but judging from what limited testing I've had time for, it works well enough. Consider this an alpha/beta release... Regards, John Hornkvist From vinet at ina.fr Wed Sep 22 06:39:03 1999 From: vinet at ina.fr (Laurent Vinet) Date: Thu Nov 3 14:42:19 2005 Subject: PB Linking in WO 4.0 Message-ID: <199909221338.PAA01160@gaston.ina.fr> Hi, I have problems linking a framework, maybe someone can help me out. This problem is new (before I can link without pb but before what ?) I want to build the framework, the link fails like this: PROGRAM : fatal error LNK1181: cannot open input file "Files/Microsoft.obj" Regards, Laurent == Making IVKit for i386-nextpdo-winnt3.5 == Pre-build setup... C:/Apple/Developer/Executables/Utilities/rm -f c:/Users/laurent/IVBuild/IVKit.build/derived_src/TrustedPrecomps.txt Building... C:/Apple/Developer/Executables/libtool -Lc:/Users/laurent/IVBuild/IVKit.build/objects-optimized -def c:/Users/laurent/IVBuild/IVKit.build/objects-optimized/IVKit.def -dynamic -compatibility_version 1 -current_version 1 -install_name c:/laurentInstall/Frameworks/IVKit.framework/IVKit.dll -arch_only i386-nextpdo-winnt3.5 -F/System/Library/PrivateFrameworks -Fc:/laurentInstall/Frameworks -FC:/Apple/Library/Frameworks -o c:/Users/laurent/IVBuild/IVKit.framework/IVKit.dll c:/Users/laurent/IVBuild/IVKit.build/objects-optimized/IVCell.o c:/Users/laurent/IVBuild/IVKit.build/objects-optimized/IVControl.o c:/Users/laurent/IVBuild/IVKit.build/objects-optimized/NSFramework_IVKit.o c:/Users/laurent/IVBuild/IVKit.build/objects-optimized/NSFrameworkForceLoad_ IVKit.o -framework AppKit -framework Foundation libtool: warning: -compatibility_version ignored libtool: warning: -current_version ignored libtool: warning: -install_name ignored PROGRAM : warning LNK4044: unrecognized option "LIB"; ignored PROGRAM : fatal error LNK1181: cannot open input file "Files/Microsoft.obj" C:/Apple/Developer/Libraries/gcc-lib/i386-nextpdo-winnt3.5\2.7.2.1\ld.exe: warning: -F: directory name (/System/Library/PrivateFrameworks) does not exist make: *** [c:/Users/laurent/IVBuild/IVKit.framework/IVKit.dll] Error 1 ___________________________________________________________ Laurent Vinet | tel: 33 (0)1 49 83 22 63 INA - Direction de la Recherche | fax: 33 (0)1 49 83 25 82 4, avenue de l'Europe | Email: vinet@ina.fr 94366 Bry sur Marne Cedex France | http://www.ina.fr/ ___________________________________________________________ From ahoesch at smartsoft.de Thu Sep 23 08:58:04 1999 From: ahoesch at smartsoft.de (Andreas Hoschler) Date: Thu Nov 3 14:42:19 2005 Subject: makeKey - mouseDown event question Message-ID: <9909231558.AA06886@smartsoft.de> Hi, I've a window with a textview that displays some search results. A click on a line in this textview causes the corresponding document to be opened (mouseDown: event). This works great, but I have to make the search window key, before the click on the search result (requires two clicks versus only one like in the PB's Find Reference window). I read somewhere in the docu how to change this behaviour, but I can't find it anymore. A corresponding hint would be appreciated! Thanks, Andreas From buck.erik at mcleod.net Thu Sep 23 09:32:13 1999 From: buck.erik at mcleod.net (Erik M. Buck) Date: Thu Nov 3 14:42:19 2005 Subject: makeKey - mouseDown event question Message-ID: <000a01bf05e1$3188e9a0$4f2b10d0@elcry> >I read somewhere in the docu how to change this behaviour, but I can't find >it anymore. A corresponding hint would be appreciated! > -acceptsFirstMouse; From nhoj at cd.chalmers.se Thu Sep 23 09:33:09 1999 From: nhoj at cd.chalmers.se (nhoj@cd.chalmers.se) Date: Thu Nov 3 14:42:19 2005 Subject: makeKey - mouseDown event question In-Reply-To: <9909231558.AA06886@smartsoft.de> References: <9909231558.AA06886@smartsoft.de> Message-ID: <19990923163310.20569.qmail@glenn> You wrote: > I read somewhere in the docu how to change this behaviour, but I can't find > it anymore. A corresponding hint would be appreciated! I believe you're looking for NSView's acceptsFirstMouse method: - (BOOL)acceptsFirstMouse:(NSEvent *)theEvent Overridden by subclasses to return YES if the receiver should be sent a mouseDown: message for theEvent, an initial mouse-down event over the receiver in its window, NO if not. The receiver can either return a value unconditionally, or use theEvent's location to determine whether or not it wants the event. NSView's implementation ignores theEvent and returns NO. Override this method in a subclass to allow instances to respond to initial mouse-down events. For example, most view objects refuse an initial mouse-down event, so that the event simply activates the window. Many control objects, however, such as NSButton and NSSlider, do accept them, so that the user can immediately manipulate the control without having to release the mouse button. See also: - hitTest: Regards, John Hornkvist From ahoesch at smartsoft.de Thu Sep 23 10:55:23 1999 From: ahoesch at smartsoft.de (Andreas Hoschler) Date: Thu Nov 3 14:42:19 2005 Subject: Forcing all relationships (faults) to be fired Message-ID: <9909231755.AA08348@smartsoft.de> Hello, I've just encountered an EOF multi thread problem on OPENSTEP. I know, that EOF 2.1 is not multithread capable, but I hoped I could get it to work in my special case anyway (read only access to already fetched eo's). Unfortunately the application crashed, most likely while firing a fault in the subthread. Is there any method that could be send to an eo (in the main thread before passing it to the subthread) that fires all faults (up to a specfied depth), so that no fault firing is needed in the subthread anymore? Thanks, Andreas From tbumgarner at itools.symantec.com Thu Sep 23 10:48:14 1999 From: tbumgarner at itools.symantec.com (Tim Bumgarner) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder Message-ID: <199909231755.KAA12094@itools.symantec.com> This is what I would like to be able to do in Interface Builder but haven't been able to figure out a way to do so: 1. Drag a view based object from a palette onto a window. 2. Drag another view based object (such as a button) into the first view such that the first view contains it. -View - -button I now there is a group into box command, but that isn't what I need here. I've tried all the keyboard shortcuts that I could think of during the drag and have searched through all of the documentation available, but no such luck. This seems like such a common thing to do. (Setting up view hierarchies that is). Anyone have any ideas? Thanks! twb From tbumgarner at itools.symantec.com Thu Sep 23 11:08:55 1999 From: tbumgarner at itools.symantec.com (Tim Bumgarner) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder In-Reply-To: <199909231755.KAA12094@itools.symantec.com> Message-ID: <199909231816.LAA12228@itools.symantec.com> Something like a draggable view subclass that could contain a row of buttons such that when the user drags the view all of it's subviews (buttons) are moved as well. The dragable view would not be subclassable from NSBox. twb > From: Eric Peyton > Date: 1999-09-23 14:09:15 -0400 > To: Tim Bumgarner > Subject: Re: Nesting views in Interface Builder > In-Reply-To: <199909231755.KAA12094@itools.symantec.com> > X-Sender: epeyton@shell-3.enteract.com > > What kind of view based object (like an NSImageView I assume) do you want > to hold the NSButton? > > NSBox is generally the way that most things are nested like this ... Or > NSScrollView, or NSSplitView. What other kind of stuff are you trying? > > Eric > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > This is what I would like to be able to do in Interface Builder but > > haven't been able to figure out a way to do so: > > > > 1. Drag a view based object from a palette onto a window. > > 2. Drag another view based object (such as a button) into the first > > view such that the first view contains it. > > -View > > - -button > > > > I now there is a group into box command, but that isn't what I need > > here. I've tried all the keyboard shortcuts that I could think of > > during the drag and have searched through all of the documentation > > available, but no such luck. This seems like such a common thing to > > do. (Setting up view hierarchies that is). > > > > Anyone have any ideas? > > > > Thanks! > > > > > > twb > > > > > > Eric Peyton > epeyton@epicware.com > > Software and Source for Mac OS X Server > > From epeyton at epicware.com Thu Sep 23 11:20:02 1999 From: epeyton at epicware.com (Eric Peyton) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder In-Reply-To: <199909231816.LAA12228@itools.symantec.com> Message-ID: Say, like a toolbar? Why don't you create you're own custom subclass of NSView, NSToolBarView which is just a Custom View in interface builder. Then create an NSBox which holds the buttons. When you load the nib set the custom NSToolbarView to have a content view of the contents of the NSBox. Would this work? Eric On Thu, 23 Sep 1999, Tim Bumgarner wrote: > Something like a draggable view subclass that could contain a row of > buttons such that when the user drags the view all of it's subviews > (buttons) are moved as well. The dragable view would not be > subclassable from NSBox. > > twb > > > From: Eric Peyton > > Date: 1999-09-23 14:09:15 -0400 > > To: Tim Bumgarner > > Subject: Re: Nesting views in Interface Builder > > In-Reply-To: <199909231755.KAA12094@itools.symantec.com> > > X-Sender: epeyton@shell-3.enteract.com > > > > What kind of view based object (like an NSImageView I assume) do > you want > > to hold the NSButton? > > > > NSBox is generally the way that most things are nested like this ... Or > > NSScrollView, or NSSplitView. What other kind of stuff are you trying? > > > > Eric > > > > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > > > This is what I would like to be able to do in Interface Builder but > > > haven't been able to figure out a way to do so: > > > > > > 1. Drag a view based object from a palette onto a window. > > > 2. Drag another view based object (such as a button) into the first > > > view such that the first view contains it. > > > -View > > > - -button > > > > > > I now there is a group into box command, but that isn't what I need > > > here. I've tried all the keyboard shortcuts that I could think of > > > during the drag and have searched through all of the documentation > > > available, but no such luck. This seems like such a common thing to > > > do. (Setting up view hierarchies that is). > > > > > > Anyone have any ideas? > > > > > > Thanks! > > > > > > > > > twb > > > > > > > > > > Eric Peyton > > epeyton@epicware.com > > > > Software and Source for Mac OS X Server > > > > > > Eric Peyton epeyton@epicware.com Software and Source for Mac OS X Server From tbumgarner at itools.symantec.com Thu Sep 23 11:33:29 1999 From: tbumgarner at itools.symantec.com (Tim Bumgarner) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder In-Reply-To: <199909231816.LAA12228@itools.symantec.com> Message-ID: <199909231840.LAA12352@itools.symantec.com> Yup, exactly like a toolbar. But what I'm trying to do is build a set of frameworks along with palettes for Interface Builder that makes it very simple for a developer to create toolbars for their apps without having to jump through a bunch of hoops to do so. It appears that IB wasn't built with this in mind, since they tried to capture the most common usage of nesting (for scroll views, and nested controls in dialogs, etc.) in the Group into commands. I already have all of the classes designed and implemented (well mostly), and when I use them from source code standpoint I create something like a 'tool window' and then create a 'dock' that is the content view of the 'tool window'. I then create one or more 'tool bars' that can be added to the 'dock' view of the 'tool window'. These toolbars are draggable. I'm sure that I can do something like let the user drag out a 'tool window' add any number of 'tool bar' views and then have them add the buttons over the top of the 'tool bar' that will contain those buttons and then when I 'awakeFromNib' go through and figure out where the toolbars were placed and take any views that were positioned above them and nest them accordingly, but this seems like a big hack to me. Basically, what I'm shooting for is the ability to construct these toolbars within IB and not require a lot of hard wired code by the developer to get them to work. twb > From: Eric Peyton > Date: 1999-09-23 14:25:55 -0400 > To: Tim Bumgarner > Subject: Re: Nesting views in Interface Builder > cc: Multiple recipients of list > In-Reply-To: <199909231816.LAA12228@itools.symantec.com> > X-Sender: epeyton@shell-3.enteract.com > > Say, like a toolbar? > > Why don't you create you're own custom subclass of NSView, NSToolBarView > which is just a Custom View in interface builder. Then create an NSBox > which holds the buttons. When you load the nib set the custom > NSToolbarView to have a content view of the contents of the NSBox. > > Would this work? > > Eric > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > Something like a draggable view subclass that could contain a row of > > buttons such that when the user drags the view all of it's subviews > > (buttons) are moved as well. The dragable view would not be > > subclassable from NSBox. > > > > twb > > > > > From: Eric Peyton > > > Date: 1999-09-23 14:09:15 -0400 > > > To: Tim Bumgarner > > > Subject: Re: Nesting views in Interface Builder > > > In-Reply-To: <199909231755.KAA12094@itools.symantec.com> > > > X-Sender: epeyton@shell-3.enteract.com > > > > > > What kind of view based object (like an NSImageView I assume) do > > you want > > > to hold the NSButton? > > > > > > NSBox is generally the way that most things are nested like this ... Or > > > NSScrollView, or NSSplitView. What other kind of stuff are you trying? > > > > > > Eric > > > > > > > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > > > > > This is what I would like to be able to do in Interface Builder but > > > > haven't been able to figure out a way to do so: > > > > > > > > 1. Drag a view based object from a palette onto a window. > > > > 2. Drag another view based object (such as a button) into the first > > > > view such that the first view contains it. > > > > -View > > > > - -button > > > > > > > > I now there is a group into box command, but that isn't what I need > > > > here. I've tried all the keyboard shortcuts that I could think of > > > > during the drag and have searched through all of the documentation > > > > available, but no such luck. This seems like such a common thing to > > > > do. (Setting up view hierarchies that is). > > > > > > > > Anyone have any ideas? > > > > > > > > Thanks! > > > > > > > > > > > > twb > > > > > > > > > > > > > > Eric Peyton > > > epeyton@epicware.com > > > > > > Software and Source for Mac OS X Server > > > > > > > > > > > > Eric Peyton > epeyton@epicware.com > > Software and Source for Mac OS X Server > > From tbumgarner at itools.symantec.com Thu Sep 23 12:08:44 1999 From: tbumgarner at itools.symantec.com (Tim Bumgarner) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder In-Reply-To: <199909231840.LAA12352@itools.symantec.com> Message-ID: <199909231915.MAA12574@itools.symantec.com> Well, I'm sure that I'm not doing the best job at describing the problem here. I have created a static palette for IB that contains all of the interesting items such as a tool window, a dock window, a tool bar (which is a subclass of NSView), rollover button, and a toolbar spacer. In the test app using IB (after adding my static palette to IB) I can drag out a tool window (which is a floating utility window). Then what I would like to do next (from a users perspective) is drag a toolbar into the tool window (since I can have more than one tool bar per tool window) and then drag in some of my rollover buttons, or spacers, (or any other views for that matter) into the toolbar such that that tool bar is the parent of these buttons, etc. I guess the bottom line is that there isn't a way to drag a view into another view and establish a hiearchy of views within IB. Thanks, Eric for your responses! twb > From: Eric Peyton > Date: 1999-09-23 15:02:39 -0400 > To: Tim Bumgarner > Subject: Re: Nesting views in Interface Builder > In-Reply-To: <199909231840.LAA12352@itools.symantec.com> > X-Sender: epeyton@shell-3.enteract.com > > Why don't you create your own palette and put your toolbar onto the > palette? > > Then they just drag a toolbar off of the palette and use it? > > Shouldn't be impossible to do - I know NeXT at one time had an NSToolbar > class, but no palette for it - I would assume it may be re-introduced at > some point. > > Eric > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > Yup, exactly like a toolbar. But what I'm trying to do is build a > > set of frameworks along with palettes for Interface Builder that > > makes it very simple for a developer to create toolbars for their > > apps without having to jump through a bunch of hoops to do so. It > > appears that IB wasn't built with this in mind, since they tried to > > capture the most common usage of nesting (for scroll views, and > > nested controls in dialogs, etc.) in the Group into commands. > > > > I already have all of the classes designed and implemented (well > > mostly), and when I use them from source code standpoint I create > > something like a 'tool window' and then create a 'dock' that is the > > content view of the 'tool window'. I then create one or more 'tool > > bars' that can be added to the 'dock' view of the 'tool window'. > > These toolbars are draggable. > > > > I'm sure that I can do something like let the user drag out a 'tool > > window' add any number of 'tool bar' views and then have them add the > > buttons over the top of the 'tool bar' that will contain those > > buttons and then when I 'awakeFromNib' go through and figure out > > where the toolbars were placed and take any views that were > > positioned above them and nest them accordingly, but this seems like > > a big hack to me. > > > > Basically, what I'm shooting for is the ability to construct these > > toolbars within IB and not require a lot of hard wired code by the > > developer to get them to work. > > > > twb > > > > > From: Eric Peyton > > > Date: 1999-09-23 14:25:55 -0400 > > > To: Tim Bumgarner > > > Subject: Re: Nesting views in Interface Builder > > > cc: Multiple recipients of list > > > In-Reply-To: <199909231816.LAA12228@itools.symantec.com> > > > X-Sender: epeyton@shell-3.enteract.com > > > > > > Say, like a toolbar? > > > > > > Why don't you create you're own custom subclass of NSView, > > NSToolBarView > > > which is just a Custom View in interface builder. Then create an NSBox > > > which holds the buttons. When you load the nib set the custom > > > NSToolbarView to have a content view of the contents of the NSBox. > > > > > > Would this work? > > > > > > Eric > > > > > > > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > > > > > Something like a draggable view subclass that could contain a > > row of > > > > buttons such that when the user drags the view all of it's subviews > > > > (buttons) are moved as well. The dragable view would not be > > > > subclassable from NSBox. > > > > > > > > twb > > > > > > > > > From: Eric Peyton > > > > > Date: 1999-09-23 14:09:15 -0400 > > > > > To: Tim Bumgarner > > > > > Subject: Re: Nesting views in Interface Builder > > > > > In-Reply-To: <199909231755.KAA12094@itools.symantec.com> > > > > > X-Sender: epeyton@shell-3.enteract.com > > > > > > > > > > What kind of view based object (like an NSImageView I assume) do > > > > you want > > > > > to hold the NSButton? > > > > > > > > > > NSBox is generally the way that most things are nested like > > this ... Or > > > > > NSScrollView, or NSSplitView. What other kind of stuff are > > you trying? > > > > > > > > > > Eric > > > > > > > > > > > > > > > On Thu, 23 Sep 1999, Tim Bumgarner wrote: > > > > > > > > > > > This is what I would like to be able to do in Interface > > Builder but > > > > > > haven't been able to figure out a way to do so: > > > > > > > > > > > > 1. Drag a view based object from a palette onto a window. > > > > > > 2. Drag another view based object (such as a button) into > > the first > > > > > > view such that the first view contains it. > > > > > > -View > > > > > > - -button > > > > > > > > > > > > I now there is a group into box command, but that isn't what > > I need > > > > > > here. I've tried all the keyboard shortcuts that I could > > think of > > > > > > during the drag and have searched through all of the > > documentation > > > > > > available, but no such luck. This seems like such a common > > thing to > > > > > > do. (Setting up view hierarchies that is). > > > > > > > > > > > > Anyone have any ideas? > > > > > > > > > > > > Thanks! > > > > > > > > > > > > > > > > > > twb > > > > > > > > > > > > > > > > > > > > > > Eric Peyton > > > > > epeyton@epicware.com > > > > > > > > > > Software and Source for Mac OS X Server > > > > > > > > > > > > > > > > > > > > > > > > Eric Peyton > > > epeyton@epicware.com > > > > > > Software and Source for Mac OS X Server > > > > > > > > > > > > Eric Peyton > epeyton@epicware.com > > Software and Source for Mac OS X Server > > From toon at omnigroup.com Thu Sep 23 13:18:14 1999 From: toon at omnigroup.com (Greg Titus) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder Message-ID: <199909232018.NAA13381@scyther.omnigroup.com> Tim Bumgarner wrote: > I guess the bottom line is that there isn't a way to drag a view > into another view and establish a hiearchy of views within IB. This is incorrect. It is possible to drag subviews into a vew and establish a hierarchy of views in IB, it is just more work, and it isn't documented very well. The most obvious example is nesting views inside NSBoxes - just double click on the box, and drag views in and you can nest boxes and other views arbitrarily deeply. What you need to do is to look at the IBEditors protocol. This protocol is how IB adds all of its editing behavior which is unique to IB, and not part of the view classes themselves. For instance, when you double-click on a button, you can edit its title. How does that work? The NSButton class doesn't let you do that in your own applications. What happens is that the InterfaceBuilder palette for the views includes another view class called IBButtonEditor that gets put on to the IB window in the place of the NSButton, and it responds to your mouse-clicks and keyboard events to edit the underlying NSButton which is what is actually saved into the nib. All of the unique editing behavior for the various views is implemented this way, mainly as subclasses (so there is an IBTableViewEditor which is a subclass of NSTableView, and adds the extra column cutting/pasting/etc. stuff that normal NSTableViews do not do.) So, your tool bar view needs to implement the -editorClassName method, and you need to have a second view class to be the toolbar editor. The toolbar editor just implements the dragging protocols to allow other views to be dropped upon it. So when you double-click on your toolbar (after dragging it into a window in IB), your editor class will be invoked, and replace the "real" toolbar view with your IB-only toolbar editor view. After that, just make your toolbar editor view accept IB view drags, and when you get a drag dropped on you, add that view as a subview to the "real" toolbar. (By the way, this info is all from class-dumping the IB-framework and IB itself. I have no first-hand knowledge of how IB works internally besides that. I highly recomend class-dump to anyone who wants to try to figure out the right way to hook into and extend all of Apple's developer applications.) Hope this helps, ---Greg From sanguish at digifix.com Thu Sep 23 13:56:42 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:19 2005 Subject: makeKey - mouseDown event question In-Reply-To: <19990923163310.20569.qmail@glenn> References: <19990923163310.20569.qmail@glenn> Message-ID: <199909232056.QAA28731@digifix.com> nhoj@cd.chalmers.se wrote: > You wrote: > > I read somewhere in the docu how to change this behaviour, but I can't > > find it anymore. A corresponding hint would be appreciated! > > I believe you're looking for NSView's acceptsFirstMouse method: > > - (BOOL)acceptsFirstMouse:(NSEvent *)theEvent > > Overridden by subclasses to return YES if the receiver should be sent a > mouseDown: message for theEvent, an initial mouse-down event over the > receiver in its window, Can anyone give a good explanation of why this isn't a delegate method? It seems like a perfect application for it... and would eliminate the need to subclass if that is the only change you need to make ... From sanguish at digifix.com Thu Sep 23 13:58:48 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder In-Reply-To: References: Message-ID: <199909232058.QAA28757@digifix.com> Eric Peyton wrote: > Say, like a toolbar? > > Why don't you create you're own custom subclass of NSView, NSToolBarView > which is just a Custom View in interface builder. Apple seems to have their own private NSToolBar class in the private frameworks... I suppose making a request for this to become a public class would be futile (and tax Chris and teams already overstuffed workload). Maybe we could ask for some OpenSource on these bits.. :-) From jmiser at loudoun.com Thu Sep 23 19:43:31 1999 From: jmiser at loudoun.com (Serial # 0) Date: Thu Nov 3 14:42:19 2005 Subject: Removal from list Message-ID: <37eae4e7.69319923@mail.loudoun.com> Hello, could you please remove me from this list. I did not sign up for it, and i believe my address was entered in error. Thank you. Jason Miser From buck.erik at mcleod.net Thu Sep 23 22:54:16 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder Message-ID: <001d01bf0651$3d3a1ec0$4b2910d0@default> First, everyone can do as they please, but I hate tool bars. That is what detachable menus and context menus are for. They use no screen space. They have text rather than incomprehensible Icons. I can put them wherever I want. All that said... IB does need a key modifier that adds views as children of views upon drop. I assume that you do not want to use an NSMatrix of cells. My advice would be to subclass NSBox (even though you do not want to) OR place the "grouped" views in their own window and than programmatically add the content view of that window (including all child views) as a subview of you special view. That way you do not have to figure out placements or add lots of subviews. You can get away with adding just one subview: the former content view of the grouping window. This is of course how inspectors are implemented. One hint: You MUST retain the content view before you take it out of the window. The following is from memory: NSView *temp = [window contentView]; [temp retain]; [window setContentView:nil]; [myView addSubview:temp]; [temp release]; Of course, I have to ask why you spend time on YellowBox given Apples history and propensity for breaking promises. What makes you think you will ever be able to use YB widely other than Apple's promise ? How much faith are you putting in that promise ? How much time and money are you wagering ? From mmalc at stepwise.com Fri Sep 24 00:15:35 1999 From: mmalc at stepwise.com (mmalcolm crawford) Date: Thu Nov 3 14:42:19 2005 Subject: Nesting views in Interface Builder Message-ID: <199909240715.BAA07954@www3.stepwise.com> Erik wrote: > Of course, I have to ask why you spend time on YellowBox given Apples > history and propensity for breaking promises. What makes you think you will > ever be able to use YB widely other than Apple's promise ? How much faith > are you putting in that promise ? How much time and money are you wagering > ? > No, you don't have to ask, at least not here; this is a topic more suitable for the -talk list. Having spent this week in Cupertino, though, I put a lot of faith in Apple's promises. How much faith do I have that you're not simply waging a jihad against the company? mmalc. ... followups to -talk, please. From troy.stephens at page44.com Fri Sep 24 12:01:51 1999 From: troy.stephens at page44.com (Troy Stephens) Date: Thu Nov 3 14:42:19 2005 Subject: [OSE 4.2] NSScrollView and centering after zoom? Message-ID: The OpenStep Enterprise 4.2 app I'm working on has a custom view enclosed in an NSScrollView, and lets the user zoom in/out on the custom view by factors of 2. So far, everything works fine. When the user asks to zoom in by 2x, I enlarge the custom view's frameSize by 2x, while keeping its boundsRect the same. As a result, the visible content is drawn with the appropriate magnification, while the scrollbars' range still covers exactly the extents of the stuff I'm drawing. Yay, AppKit! :) Now, the one improvement I'd like to make is to keep the view centered on the same point in my "document" as the user zooms in by successive steps. (More generally, it might be nice to zoom in centered on any point the user chooses -- say, with a magnifying glass tool, as you would zoom in on part of an image in a paint program -- but for now I'm keeping it simple.) Has anyone here succeeded in implementing such a feature? I've tried it various ways, and am just not getting it right. My latest attempt: @implementation MyCustomView ... - (void) zoomIn:sender { NSRect oldVisibleRect; unsigned oldMagnification; NSPoint oldCenterPoint; NSPoint newCenterPoint; NSPoint newScrollToPoint; oldVisibleRect = [scrollView documentVisibleRect]; // (-magnification and -setMagnification: are methods I added) oldMagnification = [customView magnification]; [customView setMagnification:(oldMagnification * 2)]; oldCenterPoint.x = oldVisibleRect.origin.x + oldVisibleRect.size.width / 2.0f; oldCenterPoint.y = oldVisibleRect.origin.y + oldVisibleRect.size.height / 2.0f; newCenterPoint.x = oldCenterPoint.x * 2.0f; newCenterPoint.y = oldCenterPoint.y * 2.0f; newScrollToPoint.x = newCenterPoint.x - oldVisibleRect.size.width / 2.0f; newScrollToPoint.y = newCenterPoint.y - oldVisibleRect.size.height / 2.0f; [[scrollView contentView] scrollToPoint:newScrollToPoint]; [scrollView reflectScrolledClipView:[scrollView contentView]]; } Am I getting warm? :) Thanks! Troy Troy Stephens Page 44 Studios troy.stephens@page44.com ---------------------------------------------------------------------- From andrew at stone.com Fri Sep 24 13:12:38 1999 From: andrew at stone.com (Andrew Stone) Date: Thu Nov 3 14:42:19 2005 Subject: [OSE 4.2] NSScrollView and centering after zoom? Message-ID: <199909242011.OAA05410@floyd.stone.com> > From: Troy Stephens > [[scrollView contentView] scrollToPoint:newScrollToPoint]; > [scrollView reflectScrolledClipView:[scrollView contentView]]; > } > I usually do this simple trick: [[scrollView contentView] scrollRectToVisible:RectAroundPoint(point,4.0)]; with this handy function which you could make inline if you wanted: NSRect RectAroundPoint(NSPoint pt, float amount) { return NSMakeRect(pt.x - amount,pt.y - amount, 2 * amount, 2 * amount); } > Am I getting warm?? > > Thanks! > Troy > > > Troy Stephens > Page 44 Studios > troy.stephens@page44.com > > ---------------------------------------------------------------------- > > === Andrew Stone, QuixMaster Stone Design Corp mailto:andrew@stone.com (505) 345-4800 http://www.stone.com - Create(TM) From erik at free.de Fri Sep 24 15:35:57 1999 From: erik at free.de (Erik Doernenburg) Date: Thu Nov 3 14:42:19 2005 Subject: makeKey - mouseDown event question Message-ID: > From: Scott Anguish > > [...] > > I believe you're looking for NSView's acceptsFirstMouse method: > > > > - (BOOL)acceptsFirstMouse:(NSEvent *)theEvent > > > > Overridden by subclasses to return YES if the receiver should be sent a > > mouseDown: message for theEvent, an initial mouse-down event over the > > receiver in its window, > > Can anyone give a good explanation of why this isn't a delegate method? Consistency? Users get used to what happens when the click on various user interface elements. It would certainly be surprising/annoying if the same user interface elements has different behaviour in different applications. (We all love EOModeler for accepting first mouses on the attribute icons...) > It seems like a perfect application for it... and would eliminate > the need to subclass if that is the only change you need to make ... This way each subclass has a consistent behaviour regarding activation. BTW, you know that NSView doesn't even have a delegate? regards, erik -- Erik Doernenburg, erik@object-factory.com (office), erik@free.de (private) From mmarkowitz at ceiss.org Fri Sep 24 16:14:07 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:19 2005 Subject: RTF line endings? Message-ID: I've looked in MSDN and some other sources but I can't find a clear answer on this. Does anyone know if the RTF spec defines a line ending? All their examples have breaks, but what you use for the line breaks isn't defined. Is it dependant on the platform the RTF was created on? Maury From phr at projectcenter.ch Fri Sep 24 17:15:56 1999 From: phr at projectcenter.ch (Philippe C.D. Robert) Date: Thu Nov 3 14:42:20 2005 Subject: Nesting views in Interface Builder In-Reply-To: <001d01bf0651$3d3a1ec0$4b2910d0@default> References: <001d01bf0651$3d3a1ec0$4b2910d0@default> Message-ID: <9909250015.AA00315@projectcenter.ch> You wrote: > Of course, I have to ask why you spend time on YellowBox given Apples > history and propensity for breaking promises. What makes you think you will > ever be able to use YB widely other than Apple's promise ? How much faith > are you putting in that promise ? How much time and money are you wagering > ? Sorry, but I don't think this list is the right place for such questions... sweet dreams, Phil --- Philippe C.D. Robert http://www.nice.ch/~phip From sanguish at digifix.com Fri Sep 24 20:30:45 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:20 2005 Subject: makeKey - mouseDown event question In-Reply-To: References: Message-ID: <199909250330.XAA05500@digifix.com> Erik Doernenburg wrote: > > Can anyone give a good explanation of why this isn't a > delegate method? > > Consistency? Users get used to what happens when the click on > various user interface elements. It would certainly be > surprising/annoying if the same user interface elements has different > behaviour in different applications. (We all love EOModeler for > accepting first mouses on the attribute icons...) > There are cases where you may want a view to accept first the mouse click without having the window first brought to the front.. A dock application is an excellent example. > > > It seems like a perfect application for it... and would > eliminate > > the need to subclass if that is the only change you need to make ... > > This way each subclass has a consistent behaviour regarding > activation. BTW, you know that NSView doesn't even have a delegate? Yes.. it was more of a design question... From mirko.viviani at rccr.cremona.it Sat Sep 25 12:32:04 1999 From: mirko.viviani at rccr.cremona.it (Mirko Viviani) Date: Thu Nov 3 14:42:20 2005 Subject: G200 driver help... Message-ID: <199909251932.VAA04145@next.rccr.cremona.it> Ciao... I'm working on a G200 driver for NS3.3 and DR2/intel based on the X11 sources... I already have a compiled version but I have a problem with math calculation. An line like this freeze the kernel: curmode.Clock = (int)(xmode->Clock * 1000.0 + 0.5); Why ? xmode is a valid pointer statically defined in the src and curmode is declared inside the method. Is there a way to use double computation inside the kernel ? xmode is defined as follow: struct XMode { unsigned long memSize; /* The memory necessary for this mode. */ float Clock; /* pixel clock */ int HDisplay; /* horizontal timing */ int HSyncStart; int HSyncEnd; int HTotal; int VDisplay; /* vertical timing */ int VSyncStart; int VSyncEnd; int VTotal; int Flags; }; I have another function, MGACalcClock() which use double, that inside the kernel freeze it and in user mode works without problems. I'm really sure that is this line the problem because I put an IOLog() after every line... Thank you in advance. --- Bye, Mirko (NeXTmail, MIME) From mirko.viviani at rccr.cremona.it Sat Sep 25 14:48:24 1999 From: mirko.viviani at rccr.cremona.it (mirko.viviani@rccr.cremona.it) Date: Thu Nov 3 14:42:20 2005 Subject: G200 driver help... Message-ID: <199909251948.VAA09791@rccr1.rccr.cremona.it> Ciao... I'm working on a G200 driver for NS3.3 and DR2/intel based on the X11 sources... I already have a compiled version but I have a problem with math calculation. An line like this freeze the kernel: curmode.Clock = (int)(xmode->Clock * 1000.0 + 0.5); Why ? xmode is a valid pointer statically defined in the src and curmode is declared inside the method. Is there a way to use double computation inside the kernel ? xmode is defined as follow: struct XMode { unsigned long memSize; /* The memory necessary for this mode. */ float Clock; /* pixel clock */ int HDisplay; /* horizontal timing */ int HSyncStart; int HSyncEnd; int HTotal; int VDisplay; /* vertical timing */ int VSyncStart; int VSyncEnd; int VTotal; int Flags; }; I have another function, MGACalcClock() which use double, that inside the kernel freeze it and in user mode works without problems. I'm really sure that is this line the problem because I put an IOLog() after every line... Thank you in advance. --- Bye, Mirko (NeXTmail, MIME) From bungi at omnigroup.com Sat Sep 25 13:10:34 1999 From: bungi at omnigroup.com (Timothy J. Wood) Date: Thu Nov 3 14:42:20 2005 Subject: G200 driver help... Message-ID: <199909252010.NAA13216@electabuzz.omnigroup.com> A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 1947 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990925/ce875425/attachment.bin From jcr at idiom.com Sun Sep 26 03:18:07 1999 From: jcr at idiom.com (John C. Randolph) Date: Thu Nov 3 14:42:20 2005 Subject: Need to compare audio data. Message-ID: <37EDF2D4.3879FD82@idiom.com> Guys, I'm working with Mac OS X server and/or OpenStep 4.2, and what I need to do is: Take a song from a CD, and the same track that's been recorded on a tape deck, played back and digitized, and compare the two signals in order to quantify the deviation between the two. (The idea is to get hard numbers for the quality of everything in this loop: The tape deck, the A/D and D/A converters, etc.) So, what I want to do is get my hands on the raw samples, so that I can do a brute-force phase alignment/correlation/signal inversion and subtraction. Can anyone suggest any apps that already do this, or apps with source that might give me a good jumping-off point? advTHANKSance, -jcr From leigh at tomandandy.com Sun Sep 26 15:42:47 1999 From: leigh at tomandandy.com (Leigh Smith) Date: Thu Nov 3 14:42:20 2005 Subject: Need to compare audio data Message-ID: <199909262242.SAA09744@ernst.tomandandy.com> Hi John! > So, what I want to do is get my hands on the raw samples, so that I can > do a brute-force phase alignment/correlation/signal inversion and subtraction. > Can anyone suggest any apps that already do this, or apps with source > that might give me a good jumping-off point? Resound.app should do the trick. MOXS/OS4.2, source available, bundles/plugins to add in any custom processing. You will be beholden to the "quality" of Mac A/D hardware as there is no support for Digital audio I/O (S/PDIF or AES/EBU). Tomas Hurka had a version of his AdB card driver running on RDR2/Intel until its demise (thanks Apple). No one has written a PCI audio digital I/O driver (SB Live or Event Gina cards etc) as there is no clear direction from Apple what it is doing with audio. My current solution is to use a Win98 PC to do the audio capture and then transfer all the files to the Mac. It is highly likely Apple will expect developers to resort to using Carbon to achieve remotely pro-audio operations, if the dropping of the SoundKit for NSSound is any indication. I really hope I'm wrong in this regard... -- Leigh Smith leigh@tomandandy.com (MIME) tomandandy +1-212-334-0421 (W) +1-212-334-0422 (F) 89 Greene St. New York, NY 10012, USA http://www.cs.uwa.edu.au/~leigh Microsoft - What do you want to re-install today? From rtthomas at mediaone.net Mon Sep 27 12:55:21 1999 From: rtthomas at mediaone.net (Todd Thomas) Date: Thu Nov 3 14:42:20 2005 Subject: NSDocument hierarchy on OPENSTEP? Message-ID: <9909271455.AA217612@macosx> Hi All, I"m just wondering if anyone has tackled writing NSDocument and it's corresponding classes on OPENSTEP so that apps using the document architecture can be compiled on both OPENSTEP and MacOSX Server.. If not, perhaps I'll give it a shot, though I'll probably only end up implementing the 75% or so of the API I actually need. thanks, todd --- Todd Thomas rtthomas@mediaone.net From thomas.roehl at cai.com Mon Sep 27 13:15:02 1999 From: thomas.roehl at cai.com (Roehl, Thomas) Date: Thu Nov 3 14:42:20 2005 Subject: porting and dlopen/dlclose/etc.... Message-ID: <4C094419B0E6D2119FF100805F85CABE0145289C@usilmse1.cai.com> Does anyone have any experience porting code which uses the dynamic link interface calls? I.e dlopen(), dlcose(), etc. I'm trying to build some code on the MacOS X Client preview and I'm having lots of problems. First off, there is a man page for dlfcn which talks about them but I can't find the appropriate include files. Is this something that is on Server but not in the Client version yet? Or is it also missing from Server as well. Any thoughts/experiences would be greatly appreciated. Tom roeth01@cai.com From pcoskren at bbn.com Mon Sep 27 13:29:53 1999 From: pcoskren at bbn.com (Patrick Coskren) Date: Thu Nov 3 14:42:20 2005 Subject: porting and dlopen/dlclose/etc.... Message-ID: <199909272029.QAA20017@po2.bbn.com> > Does anyone have any experience porting code which uses the dynamic link > interface calls? I.e dlopen(), dlcose(), etc. I'm trying to build some code > on the > MacOS X Client preview and I'm having lots of problems. First off, there is > a man > page for dlfcn which talks about them but I can't find the appropriate > include > files. Is this something that is on Server but not in the Client version > yet? Or > is it also missing from Server as well. Any thoughts/experiences would be > greatly appreciated. Me, too! :-) Please be sure to post replies to the list, as I just ran into this problem trying to build Window Maker. To make this post slightly useful, I can say that the same situation applies on Mac OS X Server. -Patrick From pierce at twinforces.com Mon Sep 27 16:18:18 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:20 2005 Subject: rebuilding Precompiled headers In-Reply-To: <199909272029.QAA20017@po2.bbn.com> References: <199909272029.QAA20017@po2.bbn.com> Message-ID: Since I was crazy enough to install and build egcs, my builds now all complain about old precompiled header versions. Anyone remember the magic incantation to rebuild all this? BTW, egcs binaries are at: ftp://ftp.darwinlinux.com/pub/egcs.tgz Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From wsanchez at apple.com Mon Sep 27 16:28:18 1999 From: wsanchez at apple.com (Wilfredo Sanchez) Date: Thu Nov 3 14:42:20 2005 Subject: porting and dlopen/dlclose/etc.... Message-ID: <199909272328.QAA23038@scv1.apple.com> Look at the source code to apache. I believe the file is src/os/unix/os.c. I added dyld loading support into Apache there. Download it either from Darwin or apache.org. You don't need the error handlers (wrapped in DYLD_CANT_UNLOAD conditionals) unless you are using Mac OS X Server 1.0 without the current patches. (Current versions of dyld support unloading.) Also, there is dyld code in Perl, if you need a second example. -Fred -- Wilfredo Sanchez, wsanchez@apple.com Apple Computer, Inc., Core Operating Systems / BSD Technical Lead, Darwin Project 1 Infinite Loop, 302-4K, Cupertino, CA 95014 From fabien-roy at mediaring.com Mon Sep 27 16:54:52 1999 From: fabien-roy at mediaring.com (Fabien Roy) Date: Thu Nov 3 14:42:20 2005 Subject: porting and dlopen/dlclose/etc... Message-ID: <199909272354.QAA01792@scyther.omnigroup.com> I found the followings: [fabien:~] root# find /System -name dlfcn.h -print /System/Developer/Source/Commands/tcl/tcl/compat/dlfcn.h /System/Developer/Source/Commands/gdb/gdb/tcl/compat/dlfcn.h /System/Developer/Source/Commands/perl/perl/os2/dlfcn.h It seems that there are no references to dlopen in the libraries or in the frameworks. I think that one had to emulate the dlopen etc using the Mach dyld interface or better (from the Darwin project perl): /* dl.c Fabien Roy: Extracted from dl_rhapsody.xs in the Perl distribution. Compiled ok on MacOSXServer.Not tested. Use it at your own risk Needed Framework: System */ /* dl_rhapsody.xs * * Platform: Apple Rhapsody 5.0 * Based on: dl_next.xs by Paul Marquess * Based on: dl_dlopen.xs by Anno Siegel * Created: Aug 15th, 1994 * */ #import #import #import #import typedef void * handle; static char * dl_last_error = (char *) 0; static char *dlerror() { return dl_last_error; } int dlclose(handle theHandle) /* stub only */ { return 0; } enum dyldErrorSource { OFImage, }; static void TranslateError (const char *path, enum dyldErrorSource type, int number) { char buf[8192]; unsigned int index; static char *OFIErrorStrings[] = { "%s(%d): Object Image Load Failure\n", "%s(%d): Object Image Load Success\n", "%s(%d): Not an recognisable object file\n", "%s(%d): No valid architecture\n", "%s(%d): Object image has an invalid format\n", "%s(%d): Invalid access (permissions?)\n", "%s(%d): Unknown error code from NSCreateObjectFileImageFromFile\n", }; #define NUM_OFI_ERRORS (sizeof(OFIErrorStrings) / sizeof(OFIErrorStrings[0])) switch (type) { case OFImage: index = number; if (index > NUM_OFI_ERRORS - 1) index = NUM_OFI_ERRORS - 1; sprintf(buf,OFIErrorStrings[index], path, number); break; default: sprintf(buf,"%s(%d): Totally unknown error type %d\n", path, number, type); break; } if (dl_last_error) free(dl_last_error); dl_last_error = strdup(buf); } static char *dlopen(char *path, int mode /* mode is ignored */) { int dyld_result; NSObjectFileImage ofile; NSModule handle = NULL; dyld_result = NSCreateObjectFileImageFromFile(path, &ofile); if (dyld_result != NSObjectFileImageSuccess) TranslateError(path, OFImage, dyld_result); else { // NSLinkModule will cause the run to abort on any link error's // not very friendly but the error recovery functionality is limited. handle = NSLinkModule(ofile, path, TRUE); } return handle; } void * dlsym(handle theHandle,char * symbol) { void *addr; if (NSIsSymbolNameDefined(symbol)) addr = NSAddressOfSymbol(NSLookupAndBindSymbol(symbol)); else addr = NULL; return addr; } /* Hope that helps Fabien */ -- // Fabien Roy // Senior Manager, MIS // MediaRing Inc. // 99 West Tasman Drive // Suite 280 // San Jose, CA 95134 // fabien-roy@mediaring.com // Tel: (408) 383-9222 // Fax: (408) 383-9223 // http://www.mediaring.com From mtrent at msn.fullfeed.com Mon Sep 27 19:23:01 1999 From: mtrent at msn.fullfeed.com (Mike Trent) Date: Thu Nov 3 14:42:20 2005 Subject: rebuilding Precompiled headers Message-ID: <199909280223.VAA25454@fullfeed.msn.fullfeed.com> Ah, how can I forget? As root, try: fixPrecomps (aka /usr/bin/fixPrecomps). You may also be interested in reading through the post install script for the DeveloperLibs.pkg, a copy of which is nestled in your DeveloperLibs.pkg package receipt. Check /Local/Library/Receipts. Mike Trent ---------- >From: "Pierce T. Wetter III" >To: Multiple recipients of list >Subject: rebuilding Precompiled headers >Date: Mon, Sep 27, 1999, 4:28 PM > > > Since I was crazy enough to install and build egcs, my builds now > all complain about old precompiled header versions. > > Anyone remember the magic incantation to rebuild all this? > > BTW, egcs binaries are at: ftp://ftp.darwinlinux.com/pub/egcs.tgz > > Pierce > ---------------------------------------------------------------- > Pierce T. Wetter III, Director, Twin Forces, Inc. > e-mail: pierce@twinforces.com Phone:520-779-4227 > > U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 > Comment: The only thing worse than X-Windows is Windows 95/98/NT > > From buck.erik at mcleod.net Mon Sep 27 23:47:47 1999 From: buck.erik at mcleod.net (Michelle L. Buck) Date: Thu Nov 3 14:42:20 2005 Subject: Patterns in Display Postscript Message-ID: <001801bf097d$60d07760$022910d0@default> In the past I have created fill patterns in Display Postscript by using a clipping path and a little PS routine to draw the pattern. However, I have a request for a feature that I am having trouble implementing. I am trying to emulate a piece of graphics hardware that can "spatially modulate" fills. Basically area fills can be made against a checker board mask. Every other pixel is colored. I want the Display Postscript preview to look similar to the eventual output on the emulated device. I also want to be able to print the PS preview. Do I have to composite a checker board image on top of a solid color ? Any suggestions ? Hatch patterns etc. are not a problem. It is this silly checker pattern that I am finding difficult. From peter at oops.se Tue Sep 28 01:48:45 1999 From: peter at oops.se (Peter Lindberg) Date: Thu Nov 3 14:42:20 2005 Subject: java.lang.OutOfMemory Message-ID: <007501bf098e$4fc352b0$774c07c3@highart> Hello all. I keep getting java.lang.OutOfMemory exceptions. I am trying to pass a 3MB NSString across the bridge. To pass arguments to the VM I have created a NSJavaVirtualMachine subclass that poses and overrides -initWithClassPath:arguments: to add the following argument: [args setObject:@"10485680" forKey:@"NSJavaMaxHeapSize"] The method is called, alright. In my Java method I print Runtime.getRuntime().totalMemory(), which always is "1048568" no matter what number I set as the maximum heap size. I have also tried to set the minimum heap size. Can I add something to CustomInfo.plist instead? Where do I find info on this matter for WebObjects 4? The issue seems to be addressed in a patch for WO 3.5.1, but that call is deprecated. I am really in need of help here. Peter From peter at oops.se Tue Sep 28 03:25:32 1999 From: peter at oops.se (Peter Lindberg) Date: Thu Nov 3 14:42:20 2005 Subject: SOLVED: java.lang.OutOfMemory Message-ID: <008101bf099b$cd0ac840$774c07c3@highart> I found the info I needed on http://til.info.apple.com/techinfo.nsf/artnum/n70050, but I am curious why my method doesn't work. / Peter. I wrote: > Hello all. > > I keep getting java.lang.OutOfMemory exceptions. I am trying to pass > a 3MB NSString across the bridge. > > To pass arguments to the VM I have created a NSJavaVirtualMachine > subclass that poses and overrides -initWithClassPath:arguments: to > add the following argument: > > [args setObject:@"10485680" forKey:@"NSJavaMaxHeapSize"] > > The method is called, alright. > > In my Java method I print Runtime.getRuntime().totalMemory(), which > always is "1048568" no matter what number I set as the maximum heap > size. I have also tried to set the minimum heap size. > > Can I add something to CustomInfo.plist instead? Where do I find info > on this matter for WebObjects 4? The issue seems to be addressed in > a patch for WO 3.5.1, but that call is deprecated. > > I am really in need of help here. > > Peter From marcel at metaobject.com Tue Sep 28 04:11:33 1999 From: marcel at metaobject.com (Marcel Weiher) Date: Thu Nov 3 14:42:20 2005 Subject: Patterns in Display Postscript Message-ID: > From: "Michelle L. Buck" > > In the past I have created fill patterns in Display Postscript by using a > clipping path and a little PS routine to draw the pattern. However, I have > a request for a feature that I am having trouble implementing. > > I am trying to emulate a piece of graphics hardware that can "spatially > modulate" fills. Basically area fills can be made against a checker board > mask. Every other pixel is colored. I want the Display Postscript preview > to look similar to the eventual output on the emulated device. I also want > to be able to print the PS preview. Do I have to composite a checker board > image on top of a solid color ? Any suggestions ? Hatch patterns etc. are > not a problem. It is this silly checker pattern that I am finding > difficult. Something like this? % define an uncolored pattern /Checker << /PatternType 1 /PaintType 2 /TilingType 1 /BBox [ 0 0 20 20 ] /XStep 20 /YStep 20 /PaintProc { pop 0 0 10 10 rectfill 10 10 20 20 rectfill } bind >> matrix makepattern def % set a color space: pattern + base color space [ /Pattern /DeviceGray ] setcolorspace % set color: this pattern + the color black 0 Checker setcolor % now fill to your heart's content 0 0 200 100 rectfill % RGB colrospace / color [ /Pattern /DeviceRGB ] setcolorspace % set color: this pattern + red 1 0 0 Checker setcolor 0 100 200 100 rectfill Marcel From rolf at mail.radiowan.net.au Tue Sep 28 05:34:07 1999 From: rolf at mail.radiowan.net.au (Rolf Schatzmann) Date: Thu Nov 3 14:42:20 2005 Subject: OSX Routing & Ip Accounting Software? Message-ID: <990928202811.162298c.cb5b433c.ASIP6.2.6114@mail.radiowan.net.au> Does anyone out there know if it is possible (or how hard it would be) to make a B&W running OSX Server a gateway for routing and also an IP accounting box? We are a wireless ISP and desperately need a solution for accounting IP traffic over the network, at the moment we are running a Linux box but it is causing a lot of trouble. Is there a GUI based way of setting up routes in OS X? (Something like IPnetrouter for opentransport) If not are there any developers out there working on something? Is there any IP accounting built-in to the MacOSX Kernel or could it be added or are there any developers out there working on a solution? We would be prepared to pay to have something custom written (or help fund it) if the costs were not too prohibitive. On a side note I am pretty sure there is no opentransport based IP accounting package out there but if anyone does know of a solution please let me know. Any advice or help greatly appreciated. Rolf Schatzmann Radiowan Australia From paulm at omnigroup.com Tue Sep 28 13:35:22 1999 From: paulm at omnigroup.com (Paul Mylchreest) Date: Thu Nov 3 14:42:20 2005 Subject: CThread cleanup problem Message-ID: <199909282035.NAA06213@scyther.omnigroup.com> I have run into a serious problem when trying to stop cthreads. Here is the situation briefly. Our server starts two threads (read/write) per-connection. Eventually, when the connection is dropped, I need to cleanup the two threads with something like this: void tmStopThread(void) { kern_return_t error; struct thread_basic_info info; unsigned int info_count = THREAD_BASIC_INFO_COUNT; error = thread_info( thread_self(), THREAD_BASIC_INFO, (thread_info_t)&info, &info_count); // Remove it from my internal table. RemoveThread(cthread_self()); if (error == KERN_SUCCESS) { thread_suspend(thread_self()); cthread_abort(cthread_self()); thread_resume(thread_self()); cthread_exit(0); } } Now, when the server is running (using a script to simulate connections), the value returned by cthread_count() increases to the limit (200). My internal table count, on the other hand, stays within expected values (I need the internal table to stop the threads when the server stops). Is the above code the correct way of stopping cthreads? Is there any more information besides Mach.pdf on using these constructs? Any help would be appreciated, Thanks, -- Paul Mylchreest, R & D Teamsoft Inc. 50 Queen, bureau #304 Montréal,(Qc), Canada, H3C 2N5 tel.: (514) 875-2231 (ext. 210), fax.: (514) 875-2401 Mylchreest@teamsoft.com http://www.teamsoft.com From troy.stephens at page44.com Tue Sep 28 16:07:44 1999 From: troy.stephens at page44.com (=?utf-8?B?VHJveSBTdGVwaGVucw==?=) Date: Thu Nov 3 14:42:20 2005 Subject: =?utf-8?B?UkU6IFtPU0UgNC4yXSBOU1Njcm9sbFZpZXcgYW5kIGNlbnRlcmlu?= =?utf-8?B?ZyBhZnRlciB6b29tPw==?= Message-ID: Thanks again for the tip! -scrollRectToVisible: did the trick. I was able to simplify things using NSInsetRect(), and ended up with the following in my controller implementation: - (void) zoomIn:sender { unsigned oldMagnification; NSRect oldVisibleRect; NSRect newVisibleRect; // Get visible rectangle before the zoom-in. oldVisibleRect = [scrollView documentVisibleRect]; // Increase magnification by a factor of 2. oldMagnification = [customView magnification]; [customView setMagnification:(oldMagnification * 2)]; // Calculate desired new visible rectangle. newVisibleRect = NSInsetRect( oldVisibleRect, oldVisibleRect.size.width / 4.0f, oldVisibleRect.size.height / 4.0f ); // Scroll as needed to make the rectangle visible. [customView scrollRectToVisible:newVisibleRect]; } - (void) zoomOut:sender { unsigned oldMagnification; NSRect oldVisibleRect; NSRect newVisibleRect; // Get visible rectangle before the zoom-out. oldVisibleRect = [scrollView documentVisibleRect]; // Increase magnification by a factor of 2. oldMagnification = [customView magnification]; if( oldMagnification > 1 ) { [customView setMagnification:(oldMagnification / 2)]; // Calculate desired new visible rectangle. newVisibleRect = NSInsetRect( oldVisibleRect, oldVisibleRect.size.width * -0.5f, oldVisibleRect.size.height * -0.5f ); // Scroll as needed to make the rectangle visible. [customView scrollRectToVisible:newVisibleRect]; } } Troy Stephens Page 44 Studios troy.stephens@page44.com ---------------------------------------------------------------------- -----Original Message----- From: Andrew Stone [mailto:andrew@stone.com] Sent: Friday, September 24, 1999 1:13 PM To: troy.stephens@page44.com Cc: Multiple recipients of list Subject: Re: [OSE 4.2] NSScrollView and centering after zoom? > From: Troy Stephens > [[scrollView contentView] scrollToPoint:newScrollToPoint]; > [scrollView reflectScrolledClipView:[scrollView contentView]]; > } > I usually do this simple trick: [[scrollView contentView] scrollRectToVisible:RectAroundPoint(point,4.0)]; with this handy function which you could make inline if you wanted: NSRect RectAroundPoint(NSPoint pt, float amount) { return NSMakeRect(pt.x - amount,pt.y - amount, 2 * amount, 2 * amount); } > Am I getting warm? > > Thanks! > Troy > > > Troy Stephens > Page 44 Studios > troy.stephens@page44.com > > ---------------------------------------------------------------------- > > === Andrew Stone, QuixMaster Stone Design Corp mailto:andrew@stone.com (505) 345-4800 http://www.stone.com - Create(TM) From bungi at omnigroup.com Tue Sep 28 17:58:30 1999 From: bungi at omnigroup.com (Timothy J. Wood) Date: Thu Nov 3 14:42:20 2005 Subject: porting and dlopen/dlclose/etc... Message-ID: <199909290058.RAA04967@electabuzz.omnigroup.com> A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 4466 bytes Desc: not available Url : /mailman/archive/macosx-dev/attachments/19990928/c8406b01/attachment.bin From rolf at mail.radiowan.net.au Wed Sep 29 01:29:29 1999 From: rolf at mail.radiowan.net.au (Rolf Schatzmann) Date: Thu Nov 3 14:42:20 2005 Subject: OSX Server on G3 upgraded machines? Message-ID: <990929162820.1a443fb.cb5b433c.ASIP6.2.7190@mail.radiowan.net.au> I tried to install OSX server on my 8500 with a 233Mhz Sonnet G3 Upgrade card, of course you need to use the "not recomended install" option. It installed fine but when the machine boots it gives a kernal panic and dies. With a 604e installed the machine boots fine (although slowly!) and I have installed Mac OX X Client on the 8500 with G3 card and that works fine too.... I read somehwere that the problem is that MacOSX server tries to set the backside cache to a ratio too high for some g3 upgrade cards and this is what causes it to die. Does anyone know if this is true and whether there is any way of stopping this from happening? Thanxs as always Rolf From LRenn at etci.com Wed Sep 29 10:06:13 1999 From: LRenn at etci.com (Renn, Luke) Date: Thu Nov 3 14:42:20 2005 Subject: OSX Server on G3 upgraded machines? Message-ID: You have to lower the cache ratio to 5:1. I can only my 300 at 220 because of this. Luke -----Original Message----- From: Rolf Schatzmann [mailto:rolf@mail.radiowan.net.au] Sent: Wednesday, September 29, 1999 1:04 PM To: Multiple recipients of list Subject: OSX Server on G3 upgraded machines? I tried to install OSX server on my 8500 with a 233Mhz Sonnet G3 Upgrade card, of course you need to use the "not recomended install" option. It installed fine but when the machine boots it gives a kernal panic and dies. With a 604e installed the machine boots fine (although slowly!) and I have installed Mac OX X Client on the 8500 with G3 card and that works fine too.... I read somehwere that the problem is that MacOSX server tries to set the backside cache to a ratio too high for some g3 upgrade cards and this is what causes it to die. Does anyone know if this is true and whether there is any way of stopping this from happening? Thanxs as always Rolf -------------- next part -------------- An HTML attachment was scrubbed... URL: /mailman/archive/macosx-dev/attachments/19990929/c0d6bd00/attachment.html From mmarkowitz at ceiss.org Wed Sep 29 10:24:38 1999 From: mmarkowitz at ceiss.org (Maury Markowitz) Date: Thu Nov 3 14:42:20 2005 Subject: Using NSDocument's "data" methods on wrappers? Message-ID: I know that NSDocument providers cover methods for when you might be using wrappers, but it seems to me that with all of the wrappered types out there that they are of limited use (ie, the chance that you'll be able to use them in any real app is low). So I'm playing around and trying to figure out ways around the problem. For instance, when I'm in loadDataRepresentation, I'm handed a type and the data. If I get a known wrapper type, say an rtfd file, what exactly is the data? Is it a NSFileWrapper? Or will this simply not work - and if so, what exactly is the failure? Maury From gstasko at nashville.net Wed Sep 29 10:28:54 1999 From: gstasko at nashville.net (Greg Stasko) Date: Thu Nov 3 14:42:20 2005 Subject: OSX Server on G3 upgraded machines? Message-ID: <199909291729.MAA29573@war.telalink.net> Check with the card vendor. I've had different comments from different manufacturers as to whether or not they've been successful with their cards working with OSX Server. I have an 8500, too, and I experienced similar results with a MacTell card. Newer says that their card will work. I'm awaiting same. Greg ---------- >From: "Rolf Schatzmann" >To: Multiple recipients of list >Subject: OSX Server on G3 upgraded machines? >Date: Wed, Sep 29, 1999, 12:08 PM > > I tried to install OSX server on my 8500 with a 233Mhz Sonnet G3 Upgrade > card, of course you need to use the "not recomended install" option. It > installed fine but when the machine boots it gives a kernal panic and dies. > With a 604e installed the machine boots fine (although slowly!) and I have > installed Mac OX X Client on the 8500 with G3 card and that works fine > too.... > > I read somehwere that the problem is that MacOSX server tries to set the > backside cache to a ratio too high for some g3 upgrade cards and this is > what causes it to die. Does anyone know if this is true and whether there is > any way of stopping this from happening? > > Thanxs as always > > Rolf > From mylchreest at teamsoft.com Wed Sep 29 12:10:08 1999 From: mylchreest at teamsoft.com (Paul Mylchreest) Date: Thu Nov 3 14:42:20 2005 Subject: CThread cleanup problem (repost) Message-ID: <199909291910.MAA28504@scyther.omnigroup.com> This is a repost. Something strange happened with the last one. ---------------------------- I have run into a serious problem when trying to stop cthreads. Here is the situation briefly. Our server starts two threads (read/write) per-connection. Eventually, when the connection is dropped, I need to cleanup the two threads with something like this: void tmStopThread(void) { kern_return_t error; struct thread_basic_info info; unsigned int info_count = THREAD_BASIC_INFO_COUNT; error = thread_info( thread_self(), THREAD_BASIC_INFO, (thread_info_t)&info, &info_count); // Remove it from my internal table. RemoveThread(cthread_self()); if (error == KERN_SUCCESS) { thread_suspend(thread_self()); cthread_abort(cthread_self()); thread_resume(thread_self()); cthread_exit(0); } } Now, when the server is running (using a script to simulate connections), the value returned by cthread_count() increases to the limit (200). My internal table count, on the other hand, stays within expected values (I need the internal table to stop the threads when the server stops). Is the above code the correct way of stopping cthreads? Is there any more information besides Mach.pdf on using these constructs? Any help would be appreciated, Thanks, From julien.rault at ppc.fr Wed Sep 29 18:11:24 1999 From: julien.rault at ppc.fr (Julien RAULT) Date: Thu Nov 3 14:42:20 2005 Subject: HTTP Logs analyse Message-ID: <199909300112.DAA29609@wanadoo.fr> I use a G3 server with Mac OS X server as a web server. I have some problems to produce good access stats: First I use the rotatelogs with parameters 86400 so I have a log file per day but they disappeared. I think it's the daily program that remove them. What I have to do to don't remove this file ?. Second I want to use a program that analyze the logs files and produce html reports page. To do that I first search for a program that works on Mac OS X server. I find nothings. You know a such program? I remembered analog, a program that works on UNIX. So I go to the page http://www.statslab.cam.ac.uk/~sret1/analog/ and I download the sources for UNIX on my develop computer. I had to modify the Make file to use c compiler instead of c ++ and I had to create a UNIX architecture files system to use it /usr/local/etc/httpd/analog. OK it's work! But I can't put the compiled program on my server. I think that there are much file installed than the files in the folder /usr/local/etc/httpd/analog How can I make and transfer compiled version? There is a project to make this program easy to install on MacOS X server (with his file architecture, and a .pkg ...) ? I think that it can be a good tools for Tenon iTools. Cordialement, JR a french ADC student who have a poor english. ;-) From chad at pengar.com Thu Sep 30 05:44:16 1999 From: chad at pengar.com (Chad Leigh, Pengar Enterprises, Inc. and Shire.Net LLC) Date: Thu Nov 3 14:42:20 2005 Subject: OSX Server on G3 upgraded machines? In-Reply-To: <990929162820.1a443fb.cb5b433c.ASIP6.2.7190@mail.radiowan.net.au> Message-ID: <16549000.3147669856@imac.beebe.mk> My 9600 with 266 Railgun G3 running at 260/295 (I remember setting the thing to 260 but OSXS claims 295) runs fine and has never had a problem. Dirk Theisen bought a different Railgun model and had problems but got them resolved. Two more data points Chad --On Wed, Sep 29, 1999 10.04 -0700 Rolf Schatzmann wrote: > I tried to install OSX server on my 8500 with a 233Mhz Sonnet G3 Upgrade > card, of course you need to use the "not recomended install" option. It > installed fine but when the machine boots it gives a kernal panic and > dies. With a 604e installed the machine boots fine (although slowly!) > and I have installed Mac OX X Client on the 8500 with G3 card and that > works fine too.... > > I read somehwere that the problem is that MacOSX server tries to set the > backside cache to a ratio too high for some g3 upgrade cards and this is > what causes it to die. Does anyone know if this is true and whether there > is any way of stopping this from happening? > > Thanxs as always > > Rolf > Pengar Enterprises, Inc. and Shire.Net LLC Web and Macintosh Consulting -- full service web hosting Chad Leigh chad@pengar.com chad@shire.net From lrenn at etci.com Thu Sep 30 09:57:49 1999 From: lrenn at etci.com (Luke Renn) Date: Thu Nov 3 14:42:20 2005 Subject: OSX Server on G3 upgraded machines? (Sorry about the html) References: Message-ID: <00a101bf0b64$edadf8e0$8d7ba8c0@localnet> Just wanted to apologize for the html. my company just switched to exchange server (worst mistake ever) and they had it set up to automatically add an html encoding. yuck. ----- Original Message ----- From: Renn, Luke To: Multiple recipients of list Sent: Wednesday, September 29, 1999 8:18 PM Subject: RE: OSX Server on G3 upgraded machines? > This message is in MIME format. Since your mail reader does not > understand > this format, some or all of this message may not be legible. > > ------_=_NextPart_001_01BF0A9C.EF2E7290 > Content-Type: text/plain; > charset="iso-8859-1" > > You have to lower the cache ratio to 5:1. I can only my 300 at 220 > because > of this. > > Luke > > > -----Original Message----- > From: Rolf Schatzmann [mailto:rolf@mail.radiowan.net.au] > Sent: Wednesday, September 29, 1999 1:04 PM > To: Multiple recipients of list > Subject: OSX Server on G3 upgraded machines? > > > I tried to install OSX server on my 8500 with a 233Mhz Sonnet G3 Upgrade > > card, of course you need to use the "not recomended install" option. It > installed fine but when the machine boots it gives a kernal panic and > dies. > With a 604e installed the machine boots fine (although slowly!) and I > have > installed Mac OX X Client on the 8500 with G3 card and that works fine > too.... > > I read somehwere that the problem is that MacOSX server tries to set the > backside cache to a ratio too high for some g3 upgrade cards and this is > what causes it to die. Does anyone know if this is true and whether > there is > any way of stopping this from happening? > > Thanxs as always > > Rolf > > ------_=_NextPart_001_01BF0A9C.EF2E7290 > Content-Type: text/html; > charset="iso-8859-1" > > > > > > > RE: OSX Server on G3 upgraded machines? > > > >

You have to lower the cache ratio to 5:1.  I can > only my 300 at 220 because of this. >

> >

Luke >

>
> >

-----Original Message----- >
From: Rolf Schatzmann [ HREF="mailto:rolf@mail.radiowan.net.au">mailto:rolf@mail.radiowan.net.au > ] >
Sent: Wednesday, September 29, 1999 1:04 PM >
To: Multiple recipients of list >
Subject: OSX Server on G3 upgraded machines? >

>
> >

I tried to install OSX server on my 8500 with a 233Mhz > Sonnet G3 Upgrade >
card, of course you need to use the "not > recomended install" option. It >
installed fine but when the machine boots it gives a > kernal panic and dies. >
With a 604e installed the machine boots fine (although > slowly!)  and I have >
installed Mac OX X Client on the 8500 with G3 card and > that works fine >
too.... >

> >

I read somehwere that the problem is that MacOSX server > tries to set the >
backside cache to a ratio too high for some g3 upgrade > cards and this is >
what causes it to die. Does anyone know if this is true > and whether there is >
any way of stopping this from happening? >

> >

Thanxs as always >

> >

Rolf >

> > > > ------_=_NextPart_001_01BF0A9C.EF2E7290-- > From bbum at codefab.com Thu Sep 30 16:18:47 1999 From: bbum at codefab.com (Bill Bumgarner) Date: Thu Nov 3 14:42:20 2005 Subject: Stopping a run loop (and reading all data from a filehandle in a tool) Message-ID: <199909302318.TAA10531@bjork.codefab.com> [this is from a fix to mssh -- i fixed the command line module such that it will now correctly work when grabbing the string from a task when the task spews more than a few K of output. Since mssh runs as a tool, there is no run loop. The only way to read data from a file handle in a non-blocking manner (without writing nasty platform dependent file descriptor code) is use the read in background API.] I find it odd that a runloop doesn't have a -terminate or -stop method. I ended up having to do something like (this is copy/pasted from WebScript, btw): ... [runLoop runUntilDate: [NSDate distantPast]]; [defaultCenter removeObserver: self]; NSArray *allModes = [runLoop allModes]; NSEnumerator *modeEnumerator = [allModes objectEnumerator]; NSString *nextMode; while ( (nextMode = [modeEnumerator nextObject]) != nil) { NSArray *ports = [runLoop portsForMode: nextMode]; NSEnumerator *portEnumerator = [ports objectEnumerator]; NSPort *nextPort; while( (nextPort = [portEnumerator nextObject]) != nil) [runLoop removePort: nextPort forMode: nextMode]; } ... I.e. tell the runloop to run until a date long ago, then manually remove all ports that it is actively listening for. I'm not even sure what the (1) port is that was still registered-- maybe from the NSTask object with the subshelled process???? But why would that still be registered *after* the task has terminated (the above is the implementation of the termination notification from the task)? Or could it be from the FileHandle? >From the FileHandle: #0 0x4251c074 in -[NSRunLoop addPort:forMode:] () #1 0x4258dad4 in -[NSConcreteFileHandle setModes:] () #2 0x4258d540 in -[NSConcreteFileHandle performFunction:modes:activity:] () #3 0x4258d8b8 in -[NSConcreteFileHandle readToEndOfFileInBackgroundAndNotifyForModes:] () #4 0x4258d8f0 in -[NSConcreteFileHandle readToEndOfFileInBackgroundAndNotify] () Sending -closeFile to the file handle doesn't actually remove the port from the NSRunLoop (or so it seems). Hence, it appears to be necessary to do the above [rather bogus] step remove all of the ports from the run loop and actually allow the damn thing to stop running. Is this really the case? I'm honestly hoping I missed something completely. --- Anyway: If you want to read *all* of the data from a particular file handle from a Tool, the following works. The first method assumes that you want to read everything from the standardoutput of a task... the assumption is that stderr should continue to be visible to the user. This will *only* work for tasks that actually terminate-- if you are subshelling out to, say, perl to do some text filtering, you will need to write the content to a pipe on Standard Input and then *close* the input filehandle. This will cause the task to [eventually, hopefully] end. This is WebScript under Mssh; hence the scripted perform proxy and the ability to use notification names directly (and not @"..."'d constants). .... ivars .... NSRunLoop *runLoop; NSMutableData *accumulatedData; NSNotificationCenter *defaultCenter; ScriptedPerformProxy *selfProxy; NSFileHandle *standardOutput; ............... .... instance methods ..... - (NSData *) dataFromTask: (NSTask *) aTask; { runLoop = [NSRunLoop currentRunLoop]; selfProxy = [ScriptedPerformProxy scriptedPerformProxyWithObject: self]; accumulatedData = [NSMutableData data]; defaultCenter = [NSNotificationCenter defaultCenter]; standardOutput = [[aTask standardOutput] fileHandleForReading]; [defaultCenter addObserver: selfProxy selector: @"taskDidTerminate:" name: NSTaskDidTerminateNotification object: aTask]; [defaultCenter addObserver: selfProxy selector: @"dataRead:" name: NSFileHandleReadToEndOfFileCompletionNotification object: standardOutput]; [standardOutput readToEndOfFileInBackgroundAndNotify]; [aTask launch]; [runLoop run]; return accumulatedData; } - (void) taskDidTerminate: (NSNotification *) aNotif; { [runLoop runUntilDate: [NSDate distantPast]]; [defaultCenter removeObserver: selfProxy]; NSArray *allModes = [runLoop allModes]; NSEnumerator *modeEnumerator = [allModes objectEnumerator]; NSString *nextMode; while ( (nextMode = [modeEnumerator nextObject]) != nil) { NSArray *ports = [runLoop portsForMode: nextMode]; NSEnumerator *portEnumerator = [ports objectEnumerator]; NSPort *nextPort; while( (nextPort = [portEnumerator nextObject]) != nil) [runLoop removePort: nextPort forMode: nextMode]; } } - (void) dataRead: (NSNotification *) aNotif; { [accumulatedData appendData: [[aNotif userInfo] objectForKey: NSFileHandleNotificationDataItem]]; } ...................... From pierce at twinforces.com Thu Sep 30 15:59:56 1999 From: pierce at twinforces.com (Pierce T. Wetter III) Date: Thu Nov 3 14:42:20 2005 Subject: HTTP Logs analyse In-Reply-To: <199909300112.DAA29609@wanadoo.fr> References: <199909300112.DAA29609@wanadoo.fr> Message-ID: At 3:51 PM -0700 9/30/99, Julien RAULT wrote: >I use a G3 server with Mac OS X server as a web server. > >I have some problems to produce good access stats: > >First I use the rotatelogs with parameters 86400 so I have a log file per >day but they disappeared. I think it's the daily program that remove them. > >What I have to do to don't remove this file ?. > >Second I want to use a program that analyze the logs files and produce html >reports page. To do that I first search for a program that works on Mac OS X >server. I find nothings. Check out urchin. You can get it at www.urchin.com Pierce ---------------------------------------------------------------- Pierce T. Wetter III, Director, Twin Forces, Inc. e-mail: pierce@twinforces.com Phone:520-779-4227 U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001 Comment: The only thing worse than X-Windows is Windows 95/98/NT From sanguish at digifix.com Thu Sep 30 23:47:56 1999 From: sanguish at digifix.com (Scott Anguish) Date: Thu Nov 3 14:42:20 2005 Subject: Building Packages and Installer problems Message-ID: <199910010647.CAA29604@digifix.com> There are two new in-depth articles about Installer.app and packages on Stepwise. The first describes the gotchas and issues that everyone should be aware of. The second gives detailed instructions on building packages, and the precautions to take. both are on http://www.stepwise.com/ Thanks Scott