Date: Fri, 26 Mar 2004 00:04:05 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 325 ************************************************** Thursday 25 March 2004 Number 325 ************************************************** Subjects for today 1 Re: Building gcc 3.2.2 : John Poltorak 2 Re: Posix/2 headers : John Poltorak 3 FreeBSD : Andrea Venturoli 4 Re: Posix/2 headers : Holger Veit 5 Re: /vfs? : Dave Saville" 6 Re: /vfs? : John Poltorak 7 Berkeley DB 4.1 : John Poltorak 8 Re: Posix/2 headers : John Poltorak 9 Re: *.h,v files : John Poltorak 10 Re: Building gcc 3.2.2 : Steve Wendt" 11 Re: *.h,v files : Dave and Natalie" 12 Re: *.h,v files : John Poltorak 13 Re: *.h,v files : Steven Levine" 14 Re: *.h,v files : John Poltorak 15 Re: Building gcc 3.2.2 : John Poltorak 16 Re: File v4.08 : Steve Wendt 17 libgimp122.lib -> libgimp122.a : Franz Bakan" 18 Re: *.h,v files : Steven Levine" 19 Berkeley DB - Make install : John Poltorak 20 Re: libgimp122.lib -> libgimp122.a : Dave and Natalie" 21 Re: /vfs? : Holger Veit 22 Re: /vfs? : Johannes Fahrenkrug 23 Re: /vfs? : John Poltorak 24 Building libc : John Poltorak 25 Building UNZIP using Posix/2 : John Poltorak 26 Re: Building libc : Knut Stange Osmundsen 27 Re: Building libc : John Poltorak 1 Re: Building gcc 3.2.2 : Knut Stange Osmundsen **= Email 1 ==========================** Date: Wed, 24 Mar 2004 12:55:09 +0000 From: John Poltorak Subject: Re: Building gcc 3.2.2 On Wed, Mar 24, 2004 at 01:23:58PM +0100, Knut Stange Osmundsen wrote: > For LIBC I've always tried to use sources and headers from the latest > 5.x release under the assumption that 5.x is more conforming to late > standards (like C99). I've two ways of getting the stuff, checking out > the release in CVS and zipping up /usr/include on an installed system. I have managed to get hold of a copy now from the ISO image of CD-2 and have been doing some comparisons... There appear to be some descrepancies in /usr/include/objc:- You have:- NXConstS.h objc-lis.h typedstr.h FreeBSD has:- NXConstStr.h objc-list.h typedstream.h I don't know the significance of these differences. > As you might have noticed the headers I've updated or added to LIBC > includes a note about what it was based on and hints about the changes. Wouldn't it be better to keep notes in the revision history, ie under RCS or CVS? Some headers such dir.h are identical apart from inclusion of lines such as:- /** at file * FreeBSD 5.1 */ Personally I would prefer them to be unaltered if they are identical to the original because a simple comparison would suggest they needed to be changed to work correctly on OS/2... > Kind Regards, > knut -- John **= Email 2 ==========================** Date: Wed, 24 Mar 2004 13:09:47 +0000 From: John Poltorak Subject: Re: Posix/2 headers On Wed, Mar 24, 2004 at 01:38:14PM +0100, Knut Stange Osmundsen wrote: > John Poltorak wrote: > > In the absence of any activity on the Posix/2 list, I hope it is OK > > to discuss Posix/2 here, since I hope most people who are still > > interested, will be on this list... > > > > AIUI Posix/2 is based on *BSD, and having just got hold of > > /usr/include from FreeBSD for comparison purposes, I note the absence > > of the directories asm & i386 and wonder where these originated and > > if they are necessary. It looks like a lot of work has gone into > > Posix/2 and it would be nice to see it become the basis of a new libc > > for OS/2. So, if FreeBSD is to be a guide for our libc, is it > > possible to get it completely inline? I notice that all the headers > > under asm are just redirected to i386 and wondered which programs > > actually required asm or i386... > > The directory perhaps corresponding to i386 and asm is called machine on > FreeBSD. Any BSD dependencies on these directories should go away if the > code is reprogrammed to a more current version of BSD (preferably have > 'Free' in front of that 'BSD' IMHO). I think we have agreed that FreeBSD should be the definitive guide for headers. If your libc adheres to that structure it ought to be simpler to get apps to build on OS/2 without too much reworking. What I'd like to know is how do you find which programs require asm and 386 or i386? > The LIBC should be able to shop code from both BSD and LGPL worlds, > perhaps prefering BSD if having multiple options. (The best place of > getting a GNU extension is usually to GLIBC, not *BSD.) > Thus the basic layout of headers should be BSD like, while LGPL stuff > would have to be made fit in somehow. I've heard people suggest that GLIBC is a can of worms and should be avoided at all costs. How true that is I don't know. > Kind Regards, > knut -- John **= Email 3 ==========================** Date: Wed, 24 Mar 2004 15:27:35 EST From: Andrea Venturoli Subject: FreeBSD ** Reply to note from John Poltorak Tue, 23 Mar 2004 20:54:34 +0000 >If anyone on this list is familiar with FreeBSD could you say whereabouts >on the CD you could expect to find the headers which get installed into >/usr/include? I read that you already found them, however... They are in the bin tarball, which is split in several files named bin.aa, bin.ab, bin.ac, ... You'll need to concatenate them and extract as you would do from a normal tarball e.g.: cat bin.*|tar xzf - (provided they are already sorted alfabetically). bye av. **= Email 4 ==========================** Date: Wed, 24 Mar 2004 15:24:26 +0100 From: Holger Veit Subject: Re: Posix/2 headers On Wed, Mar 24, 2004 at 01:09:47PM +0000, John Poltorak wrote: > On Wed, Mar 24, 2004 at 01:38:14PM +0100, Knut Stange Osmundsen wrote: [...] > > The directory perhaps corresponding to i386 and asm is called machine on > > FreeBSD. Any BSD dependencies on these directories should go away if the > > code is reprogrammed to a more current version of BSD (preferably have > > 'Free' in front of that 'BSD' IMHO). > > I think we have agreed that FreeBSD should be the definitive guide for > headers. If your libc adheres to that structure it ought to be simpler to > get apps to build on OS/2 without too much reworking. > > What I'd like to know is how do you find which programs require asm and > 386 or i386? Well-written software shouldn't access 386 or i386 directly, otherwise they are nonportable. BSD uses, as Knut correctly remarked, /usr/include/machine which is a platform-dependent symbolic link to the architecture subdirectory, e.g. to 'alpha' on DEC alphas, or 386 on Intels, or mips on MIPS based machines. As OS/2 is Intel only (forget about the PowerPC), it is consequent, given the BSD paradigma, to have all these i386 machine dependent files under /usr/include/machine. If other BSD header files refer to such machine dependent headers they usually do it by including . Prominent exception maybe: headers that are in <386> itself calling other 386 headers. That has to be corrected then. > > The LIBC should be able to shop code from both BSD and LGPL worlds, > > perhaps prefering BSD if having multiple options. (The best place of > > getting a GNU extension is usually to GLIBC, not *BSD.) > > Thus the basic layout of headers should be BSD like, while LGPL stuff > > would have to be made fit in somehow. > > I've heard people suggest that GLIBC is a can of worms and should be > avoided at all costs. How true that is I don't know. So did I hear. Not only that, it is quite difficult to port it as it has been over-configured in an attempt to adapt to almost all awkward Unix-lookalike dialect under the sun, with the exception of the one that you just want to port to ;-( OTOH, it is maybe the best source for GNU extensions. OTTH, then, if you try to pick up one of these fine extensions, it is like a ball of spaghetti: you might end up pulling the whole stuff into your boat. Holger **= Email 5 ==========================** Date: Wed, 24 Mar 2004 14:28:20 +0000 (GMT) From: "Dave Saville" Subject: Re: /vfs? I like to keep the OS and the rest on different drives. But if this were not an issue you could use LVM to mash all your drives into one huge C: and then you could forget about the drive letter as /something always works assuming one is on the correct drive - and in this case there would be only one. -- Regards Dave Saville **= Email 6 ==========================** Date: Wed, 24 Mar 2004 14:34:58 +0000 From: John Poltorak Subject: Re: /vfs? On Wed, Mar 24, 2004 at 02:28:20PM +0000, Dave Saville wrote: > I like to keep the OS and the rest on different drives. But if this > were not an issue you could use LVM to mash all your drives into one > huge C: and then you could forget about the drive letter as > /something always works assuming one is on the correct drive - and in > this case there would be only one. ....until you NET USE x: \\someotherserver\c$ > -- > Regards > > Dave Saville > -- John **= Email 7 ==========================** Date: Wed, 24 Mar 2004 15:01:25 +0000 From: John Poltorak Subject: Berkeley DB 4.1 Can anyone suggest how I can find out the contents of a Berkeley DB 4.1 installable package? I'd like to create an equivalent one for OS/2. -- John **= Email 8 ==========================** Date: Wed, 24 Mar 2004 15:22:53 +0000 From: John Poltorak Subject: Re: Posix/2 headers On Wed, Mar 24, 2004 at 03:24:26PM +0100, Holger Veit wrote: > On Wed, Mar 24, 2004 at 01:09:47PM +0000, John Poltorak wrote: > > What I'd like to know is how do you find which programs require asm and > > 386 or i386? > > Well-written software shouldn't access 386 or i386 directly, otherwise > they are nonportable. I guess these directories have been created to accommodate non portable software... But do we know which software we are talking about? By continuing to provide these directories we are unlikely to trap the software which needs changing. I suppose a version of Posix/2 without asm and i386 could be used to discover what breaks... But first I need to know how to build Posix/2. Are there any simple build instructions around? > usually do it by including . Prominent exception maybe: > headers that are in <386> itself calling other 386 headers. > That has to be corrected then. Need to find them first... > Holger -- John **= Email 9 ==========================** Date: Wed, 24 Mar 2004 17:04:32 +0000 From: John Poltorak Subject: Re: *.h,v files On Tue, Mar 23, 2004 at 07:19:14PM -0800, Steven Levine wrote: > http://www.freebsd.org/support.html#cvs > > If it were me, I would just use anon cvs to pull the /usr/include part of > the source tree rather than creating a local respository. Instructions > for anon cvs are at: > > http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/anoncvs.html > > and there's a one page getting started guide at: > > http://www.scoug.com/os24u/2003/scoug305.mrkia.html > > Once you get logged in: > > cvs co usr/src/include I'm not getting logged in. I get no response after entering 'anoncvs' as the password. > HTH, > > Steven > > -- > ---------------------------------------------------------------------- > "Steven Levine" MR2/ICE 2.41 #10183 Warp4/FP15/14.093c_W4 > www.scoug.com irc.webbnet.info irc.fyrelizard.org #scoug (Wed 7pm PST) > ---------------------------------------------------------------------- > -- John **= Email 10 ==========================** Date: Wed, 24 Mar 2004 09:36:10 -0800 (PST) From: "Steve Wendt" Subject: Re: Building gcc 3.2.2 On Wed, 24 Mar 2004 12:55:09 +0000, John Poltorak wrote: >You have:- > >NXConstS.h >objc-lis.h >typedstr.h > >FreeBSD has:- > >NXConstStr.h >objc-list.h >typedstream.h > >I don't know the significance of these differences. Looks like 8.3 conformance to me. >Wouldn't it be better to keep notes in the revision history, ie under RCS >or CVS? Having it in the source makes it available to someone who isn't using the source control system. ----------- "Good people do not need laws to tell them to act responsibly, while bad people will find a way around the laws." - Plato (427-347 B.C.) **= Email 11 ==========================** Date: Wed, 24 Mar 2004 09:50:17 -0800 From: "Dave and Natalie" Subject: Re: *.h,v files On Wed, 24 Mar 2004 17:04:32 +0000, John Poltorak wrote: >> Once you get logged in: >> >> cvs co usr/src/include > > >I'm not getting logged in. I get no response after entering 'anoncvs' as >the password. IIRC you don't get any response from a good password. Try entering cvs co usr/src/inlude. Dave **= Email 12 ==========================** Date: Wed, 24 Mar 2004 18:45:58 +0000 From: John Poltorak Subject: Re: *.h,v files On Wed, Mar 24, 2004 at 09:50:17AM -0800, Dave and Natalie wrote: > On Wed, 24 Mar 2004 17:04:32 +0000, John Poltorak wrote: > > >> Once you get logged in: > >> > >> cvs co usr/src/include > > > > > >I'm not getting logged in. I get no response after entering 'anoncvs' as > >the password. > > IIRC you don't get any response from a good password. Try entering cvs co usr/src/inlude. It looks like it must have been a glitch - I've manage to login now. If anyone wants to try you need a variable set to: :pserver:anoncvs at anoncvs.FreeBSD.org:/home/ncvs and this value should also be in your ~/.cvspass file. You can grab the headers using:- cvs co src/include One thing I notice is that it only has the arpa, protocols, rpc and rpcsvc directories... > Dave > -- John **= Email 13 ==========================** Date: Wed, 24 Mar 2004 10:43:17 -0800 From: "Steven Levine" Subject: Re: *.h,v files In <20040324175018.39E48B7982 at joseph.ncoldns.com>, on 03/24/04 at 09:50 AM, "Dave and Natalie" said: >IIRC you don't get any response from a good password. This is true. You will get a warning the first time that cvs creates $HOME\.cvspass. Either John already had a cvs setup or he neglected to mention the message. >Try entering cvs co >usr/src/inlude. Dave I tested this. The FreeBSD documentation led me astray. The checkout command John needs is: cvs co src/include and there's a makefile to install to /usr/include. FWIW, the first thing I usually do when going to a new cvs repository is: cvs co CVSROOT Not every site allows access to this directory, but most do. CVSROOT contains the administrative files. In particular the modules file is handy for understanding the repository layout. HTH, Steven -- ---------------------------------------------------------------------- "Steven Levine" MR2/ICE 2.41 #10183 Warp4/FP15/14.093c_W4 www.scoug.com irc.webbnet.info irc.fyrelizard.org #scoug (Wed 7pm PST) ---------------------------------------------------------------------- **= Email 14 ==========================** Date: Wed, 24 Mar 2004 19:15:08 +0000 From: John Poltorak Subject: Re: *.h,v files On Wed, Mar 24, 2004 at 10:43:17AM -0800, Steven Levine wrote: > In <20040324175018.39E48B7982 at joseph.ncoldns.com>, on 03/24/04 > at 09:50 AM, "Dave and Natalie" said: > > >IIRC you don't get any response from a good password. > > This is true. What I meant was there was no return to the command prompt. It is as if the system was waiting for some response from the remote host which wasn't forthcoming - but this must have been a temporary fault which has now gone. > You will get a warning the first time that cvs creates > $HOME\.cvspass. Either John already had a cvs setup or he neglected to > mention the message. > > >Try entering cvs co > >usr/src/inlude. Dave > > I tested this. The FreeBSD documentation led me astray. The checkout > command John needs is: > > cvs co src/include > > and there's a makefile to install to /usr/include. Is there something different about *BSD Makefiles? I remember looking at one some time back and couldn't work out what it was supposed to be doing at all... > FWIW, the first thing I usually do when going to a new cvs repository is: > > cvs co CVSROOT > > Not every site allows access to this directory, but most do. CVSROOT > contains the administrative files. In particular the modules file is > handy for understanding the repository layout. Thanks for those handy tips. I've never really spent much time with CVS, but I suspet it is something which will be using more and more in the future. > HTH, > > Steven > > -- > ---------------------------------------------------------------------- > "Steven Levine" MR2/ICE 2.41 #10183 Warp4/FP15/14.093c_W4 > www.scoug.com irc.webbnet.info irc.fyrelizard.org #scoug (Wed 7pm PST) > ---------------------------------------------------------------------- > -- John **= Email 15 ==========================** Date: Wed, 24 Mar 2004 20:24:49 +0000 From: John Poltorak Subject: Re: Building gcc 3.2.2 On Wed, Mar 24, 2004 at 09:36:10AM -0800, Steve Wendt wrote: > On Wed, 24 Mar 2004 12:55:09 +0000, John Poltorak wrote: > > >You have:- > > > >NXConstS.h > >objc-lis.h > >typedstr.h > > > >FreeBSD has:- > > > >NXConstStr.h > >objc-list.h > >typedstream.h > > > >I don't know the significance of these differences. > > Looks like 8.3 conformance to me. But is it intentional? -- John **= Email 16 ==========================** Date: Wed, 24 Mar 2004 13:05:42 -0800 (PST) From: Steve Wendt Subject: Re: File v4.08 > /usr/share/misc/magic Yes, that one is a symlink to /usr/share/magic About your db4 question, maybe this will help you: ~>rpm -ql db4 /lib/libdb-4.0.so /usr/lib/libdb_cxx-4.0.so /usr/share/doc/db4-4.0.14 /usr/share/doc/db4-4.0.14/LICENSE /usr/share/doc/db4-4.0.14/README /usr/share/doc/db4-4.0.14/images /usr/share/doc/db4-4.0.14/images/api.gif /usr/share/doc/db4-4.0.14/images/next.gif /usr/share/doc/db4-4.0.14/images/prev.gif /usr/share/doc/db4-4.0.14/images/ps.gif /usr/share/doc/db4-4.0.14/images/ref.gif /usr/share/doc/db4-4.0.14/images/sleepycat.gif ~>rpm -ql db4-utils /usr/bin/berkeley_db_svc /usr/bin/db_archive /usr/bin/db_checkpoint /usr/bin/db_deadlock /usr/bin/db_dump /usr/bin/db_dump185 /usr/bin/db_load /usr/bin/db_printlog /usr/bin/db_recover /usr/bin/db_stat /usr/bin/db_upgrade /usr/bin/db_verify /usr/lib/libdb_tcl-4.0.a /usr/lib/libdb_tcl-4.0.la /usr/lib/libdb_tcl-4.0.so /usr/lib/libdb_tcl.so /usr/share/doc/db4-utils-4.0.14 /usr/share/doc/db4-utils-4.0.14/utility /usr/share/doc/db4-utils-4.0.14/utility/berkeley_db_svc.html /usr/share/doc/db4-utils-4.0.14/utility/db_archive.html /usr/share/doc/db4-utils-4.0.14/utility/db_checkpoint.html /usr/share/doc/db4-utils-4.0.14/utility/db_deadlock.html /usr/share/doc/db4-utils-4.0.14/utility/db_dump.html /usr/share/doc/db4-utils-4.0.14/utility/db_load.html /usr/share/doc/db4-utils-4.0.14/utility/db_printlog.html /usr/share/doc/db4-utils-4.0.14/utility/db_recover.html /usr/share/doc/db4-utils-4.0.14/utility/db_stat.html /usr/share/doc/db4-utils-4.0.14/utility/db_upgrade.html /usr/share/doc/db4-utils-4.0.14/utility/db_verify.html /usr/share/doc/db4-utils-4.0.14/utility/index.html ~>rpm -ql db4-devel|more /lib/libdb.so /usr/include/cxx_common.h /usr/include/cxx_except.h /usr/include/db.h /usr/include/db4 /usr/include/db4/cxx_common.h /usr/include/db4/cxx_except.h /usr/include/db4/db.h /usr/include/db4/db_185.h /usr/include/db4/db_cxx.h /usr/include/db_185.h /usr/include/db_cxx.h /usr/lib/libdb-4.0.a /usr/lib/libdb-4.0.la /usr/lib/libdb-4.0.so /usr/lib/libdb_cxx-4.0.a /usr/lib/libdb_cxx-4.0.la /usr/lib/libdb_cxx.so /usr/share/doc/db4-devel-4.0.14 /usr/share/doc/db4-devel-4.0.14/api_c .... and lots more docs ... **= Email 17 ==========================** Date: Wed, 24 Mar 2004 22:53:15 +0100 (CET) From: "Franz Bakan" Subject: libgimp122.lib -> libgimp122.a Hi, For building xsane as gimp-plugin it would be nice to have libgimp122.a which is not included in the packages available for download. I allready have libgimp122.lib but can not use -Zomf because xsane uses fork(). As a workaround I can link with gimp121.a and it works too, but having a libgimp122.a would be nicer. Any chance to convert either libgimp122.lib or gimp122.dll to libgimp122.a ? Or does anyone having a libgimp122.a laying around ? Franz **= Email 18 ==========================** Date: Wed, 24 Mar 2004 14:35:14 -0800 From: "Steven Levine" Subject: Re: *.h,v files In <20040324191508.D42296 at warpix.org>, on 03/24/04 at 07:15 PM, John Poltorak said: >What I meant was there was no return to the command prompt. It is as if >the system was waiting for some response from the remote host which >wasn't forthcoming - but this must have been a temporary fault which has >now gone. The site is slow. It took about 15 seconds for my password to be processed. It don't know if this is temporary or the server is just busy. cvs puts a pretty good sized load on a server. When we were stress testing the last set of Andrew's cvs patches I was hitting a server with 6-8 concurrent checkouts of a 1.2GB respository. The server was a reasonable quick Intellistation with 512MB of RAM. This was enoungh to max the performance meter. >Is there something different about *BSD Makefiles? I remember looking at >one some time back and couldn't work out what it was supposed to be doing > at all... Since IMO, there's no such thing as a standard makefile, the answer is probably yes. I looked the provided Makefile. If not standard, but then again, it doesn't use any features I haven't seen implemented elsewhere, although not always with the same syntax. >Thanks for those handy tips. I've never really spent much time with CVS, >but I suspet it is something which will be using more and more in the >future. Well, if you work with open source, it's pretty hard to avoid for long. For open source, the current major players are probably CVS, perforce and subversion. Over time, I expect subversion of replace CVS. It will be better, but CVS has a lot of history and a lot of tools. Steven -- ---------------------------------------------------------------------- "Steven Levine" MR2/ICE 2.41 #10183 Warp4/FP15/14.093c_W4 www.scoug.com irc.webbnet.info irc.fyrelizard.org #scoug (Wed 7pm PST) ---------------------------------------------------------------------- **= Email 19 ==========================** Date: Wed, 24 Mar 2004 23:26:52 +0000 From: John Poltorak Subject: Berkeley DB - Make install As part of building Berkeley DB v4.1.25 I run 'make install' and it appears to install properly even though it doesn't get built correctly. However in looking through the Makefile, I don't see an install: target. If there isn't one should I expect make to fail? Maybe there is another Makefile somewhere and I'm looking at the wrong one... -- John **= Email 20 ==========================** Date: Wed, 24 Mar 2004 16:55:48 -0800 From: "Dave and Natalie" Subject: Re: libgimp122.lib -> libgimp122.a On Wed, 24 Mar 2004 22:53:15 +0100 (CET), Franz Bakan wrote: > >Any chance to convert either libgimp122.lib or gimp122.dll to libgimp122.a ? > >Or does anyone having a libgimp122.a laying around ? I've changed *.lib to *.a before. IIRC emximp -o libgimp122.a libgimp122.lib Dave ps if you have a gimp121.dll and/or gimp.dll could you send it to me privately? **= Email 21 ==========================** Date: Thu, 25 Mar 2004 08:57:24 +0100 From: Holger Veit Subject: Re: /vfs? On Wed, Mar 24, 2004 at 02:34:58PM +0000, John Poltorak wrote: > On Wed, Mar 24, 2004 at 02:28:20PM +0000, Dave Saville wrote: > > I like to keep the OS and the rest on different drives. But if this > > were not an issue you could use LVM to mash all your drives into one > > huge C: and then you could forget about the drive letter as > > /something always works assuming one is on the correct drive - and in > > this case there would be only one. > > > ...until you NET USE x: \\someotherserver\c$ Not only then. The real problem with LVM is that you have to plan your system accordingly in advance. Then you can combine multiple disk into one large drive. This will no longer work if you already have several partitioned and formatted disks, perhaps even with different types of filesystems on it. You won't get rid of those drive letters then with LVM. Holger **= Email 22 ==========================** Date: Thu, 25 Mar 2004 09:59:24 GMT From: Johannes Fahrenkrug Subject: Re: /vfs? Hi! I'm new to this list and the first thing I read in the welcome mail was, that this isn't the right mailing list for me :) I have problems compiling subversion under OS/2. I know this is a porting question, so I'd appreciate it if one of you nice contemporaries could point me to the right mailing list. -Johannes. **= Email 23 ==========================** Date: Thu, 25 Mar 2004 09:53:18 +0000 From: John Poltorak Subject: Re: /vfs? Hi Johannes, Welcome to the list. I'm not sure what subversion is but if it is a Unix app which you want to get working on OS/2 then this is definitely the right list. Maybe you could outline the sort of problems you are having... Also, there doesn't seem much point in starting a thread about subversion using a Subject of /vfs? On Thu, Mar 25, 2004 at 09:59:24AM +0000, Johannes Fahrenkrug wrote: > Hi! > > I'm new to this list and the first thing I read in the welcome mail > was, that this isn't the right mailing list for me :) > I have problems compiling subversion under OS/2. I know this is a > porting question, so I'd appreciate it if one of you nice > contemporaries could point me to the right mailing list. > > -Johannes. > > > -- John **= Email 24 ==========================** Date: Thu, 25 Mar 2004 10:26:48 +0000 From: John Poltorak Subject: Building libc Can anyone give me any guidance on building libc? I'd like to try and combine Innotek's libc with Posix/2's libcExt but am not sure how to go about it. -- John **= Email 25 ==========================** Date: Thu, 25 Mar 2004 10:56:50 +0000 From: John Poltorak Subject: Building UNZIP using Posix/2 When attempting to build InfoZip's UNZIP I get this error when using Posix/2:- [C:\ux2bs\workdir\unzip-5.50]make -f os2\makefile.os2 gcc make -f os2/makefile.os2 all \ CC="gcc -Zomf -O -I." \ CFLAGS="-Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK" \ NFLAGS="" \ DLLFLAG="" \ AS="gcc -Zomf" \ ASFLAGS="-Di386" \ LDFLAGS="-o ./" \ LDFLAGS2="-Zsmall-conv -Zstack 320 -Zsys -s" \ OUT="-o" \ OBJ=".obj" \ CRC32="crc_gcc" \ CRC32F="crc_gcc" \ CRC32X="crc_gcc" \ DEF="os2/unzip.def" make[1]: Entering directory `/ux2bs/workdir/unzip-5.50' gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK unzip.c gcc -Zomf -Di386 -x assembler-with-cpp -c -o crc_gcc.obj crc_i386.S gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK crctab.c gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK crypt.c gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK envargs.c gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK explode.c gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK extract.c gcc -Zomf -O -I. -c -Wall -DOS2 -DASM_CRC -DOS2_EAS -DUSE_UNSHRINK fileio.c fileio.c: In function `dos_to_unix_time': fileio.c:1743: structure has no member named `tm_gmtoff' make[1]: *** [fileio.obj] Error 1 make[1]: Leaving directory `/ux2bs/workdir/unzip-5.50' make: *** [gcc] Error 2 Here is the relevant code:- /*--------------------------------------------------------------------------- Adjust for local standard timezone offset. ---------------------------------------------------------------------------*/ #if (!defined(MACOS) && !defined(RISCOS) && !defined(QDOS) && !defined(TANDEM)) #ifdef WIN32 /* account for timezone differences */ res = GetTimeZoneInformation(&tzinfo); if (res != TIME_ZONE_ID_INVALID) { m_time += 60*(tzinfo.Bias); #else /* !WIN32 */ #if (defined(BSD) || defined(MTS) || defined(__GO32__)) #ifdef BSD4_4 if ( (dosdatetime >= DOSTIME_2038_01_18) && (m_time < (time_t)0x70000000L) ) m_time = U_TIME_T_MAX; /* saturate in case of (unsigned) overflow */ if (m_time < (time_t)0L) /* a converted DOS time cannot be negative */ m_time = S_TIME_T_MAX; /* -> saturate at max signed time_t value */ if ((tm = localtime(&m_time)) != (struct tm *)NULL) /* line 1743 ====> */ m_time -= tm->tm_gmtoff; /* sec. EAST of GMT: subtr. */ #else /* !(BSD4_4 */ ftime(&tbp); /* get `timezone' */ m_time += tbp.timezone * 60L; /* seconds WEST of GMT: add */ #endif /* ?(BSD4_4 || __EMX__) */ #else /* !(BSD || MTS || __GO32__) */ /* tzset was already called at start of process_zipfiles() */ /* tzset(); */ /* set `timezone' variable */ #ifndef __BEOS__ /* BeOS DR8 has no timezones... */ m_time += timezone; /* seconds WEST of GMT: add */ #endif #endif /* ?(BSD || MTS || __GO32__) */ #endif /* ?WIN32 */ TTrace((stderr, " m_time after timezone = %lu\n", (ulg)m_time)); This builds fine without Posix/2. Can anyone suggest why it won't build with it? -- John **= Email 26 ==========================** Date: Thu, 25 Mar 2004 12:20:56 +0100 From: Knut Stange Osmundsen Subject: Re: Building libc John Poltorak wrote: > > Can anyone give me any guidance on building libc? Sure, you'll need make 3.8x, GCC 3.2.2, ilink v5, masm v6, and the usual set of unix utils (sed, awk, cat, cp, sh,...). The location of masm v6 can be specified using ASM=d:/where/ever/masm/v6/is/ml.exe either as an environment variable or an argument to make. Then going to src/emx and running make will give you a help about the available targets. "make all" is probably what you want the first time. All generated files will be found in out/dbg/ or out/rel/, depending on the value of MODE which by default is dbg. Make install will default to out/$(MODE)/install/, this can be overrided by specifying INS=myinstaldirectory/ as an argument to make install. There is an description of the MoBuS (build system) internals in build.txt. Kind Regards, knut **= Email 27 ==========================** Date: Thu, 25 Mar 2004 12:11:12 +0000 From: John Poltorak Subject: Re: Building libc On Thu, Mar 25, 2004 at 12:20:56PM +0100, Knut Stange Osmundsen wrote: > John Poltorak wrote: > > > > Can anyone give me any guidance on building libc? > Sure, you'll need make 3.8x, GCC 3.2.2, ilink v5, masm v6, and the usual > set of unix utils (sed, awk, cat, cp, sh,...). The location of masm v6 > can be specified using ASM=d:/where/ever/masm/v6/is/ml.exe either as an > environment variable or an argument to make. MASM v6?!!! Haven't touched that in years - I'll need to dig deep to lay my hands on it again... > Then going to src/emx and running make will give you a help about the > available targets. "make all" is probably what you want the first time. > There is an description of the MoBuS (build system) internals in build.txt. Which package are you talking about? I don't suppose it is this one:- ... http://download.innotek.de/gccos2/3.2.2-Beta4/gcc-3.2.2-beta4-src-emx.zip > Kind Regards, > knut -- John **= Email 1 ==========================** Date: Thu, 25 Mar 2004 11:58:25 +0100 From: Knut Stange Osmundsen Subject: Re: Building gcc 3.2.2 Steve Wendt wrote: > On Wed, 24 Mar 2004 12:55:09 +0000, John Poltorak wrote: > > >> You have:- >> >> NXConstS.h >> objc-lis.h >> typedstr.h >> >> FreeBSD has:- >> >> NXConstStr.h >> objc-list.h >> typedstream.h >> >> I don't know the significance of these differences. Very in siginificant I'd say, especially when considering that I'm not shipping them... >> Wouldn't it be better to keep notes in the revision history, ie >> under RCS or CVS? > > > Having it in the source makes it available to someone who isn't using > the source control system. That's exactly my intention. Also all the changes are marked so anyone can look at it and/or compare it with the original and see what's done. Kind Regards, knut