From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Thu, 21 Feb 2002 04:16:12 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 142 ************************************************** Wednesday 20 February 2002 Number 142 ************************************************** Subjects for today 1 Re: EMACS build error : Andreas Buening 2 Re: mod_ssl : Stepan Kazakov 3 Re: mod_ssl : Brian Havard" 4 Re: EMACS build error : Masaru Nomiya 5 Re: make 3.79.1 : John Poltorak 6 Re: OSFree... : Jack Troughton 7 Re: terminfo database : James Cannon 8 Re: URL for CVS tarball : James Cannon 9 Re: EMACS build error : Holger Veit 10 Re: Re: EMACS build error : John Poltorak 11 Re: mod_ssl : John Poltorak 12 Re: make 3.79.1 : John Poltorak 13 Re: terminfo database : email at eracc.hypermart.net 14 Re: Has anyone ported nmap? : email at eracc.hypermart.net 15 c++filt : John Poltorak 16 Re: terminfo database : Thomas E. Dickey" 17 Re: OSFree... : Mentore Siesto 18 Re: emx\lib\*.obj ? : John Poltorak 19 Re: mod_ssl : Jack Troughton 20 Re: OSFree... : Jack Troughton 21 emacs v21.1 : John Poltorak 22 Re: Re: EMACS build error : John Poltorak 23 Courier IMAP : John Poltorak 24 Re: mod_ssl : Jack Troughton 25 Re: OSFree... : Sebastian Wittmeier (ShadoW)" 26 Re: Re: EMACS build error : Holger Veit 27 Re: OSFree... : Jack Troughton 28 Re: emx\lib\*.obj ? : Holger Veit 29 Re: OSFree... : Kris Steenhaut 30 URL for CVS tarball : John Poltorak 31 terminfo database : John Poltorak 32 bfd.dll : John Poltorak 33 UnixOS/2 distro : John Poltorak 34 Re: URL for CVS tarball : Andrew Belov" 35 Re: UnixOS/2 distro : Thomas Dickey 36 Re: EMACS build error : Masaru Nomiya 37 Re: Building Perl with gcc v3.0.3 : Henry Sobotka 38 Another Patch (WAs : Re: EMACS build error) : Masaru Nomiya 39 Re: EMACS build error : Andrew MacIntyre 40 (un)protoize : John Poltorak 41 Re: EMACS build error : Masaru Nomiya 42 Re: make 3.79.1 : Andreas Buening 43 Re: EMACS build error : Andreas Buening 44 Building Perl with gcc v3.0.3 : John Poltorak 45 Re: (un)protoize : Sebastian Wittmeier (ShadoW)" **= Email 1 ==========================** Date: Thu, 21 Feb 2002 00:48:09 +0100 From: Andreas Buening Subject: Re: EMACS build error Holger Veit wrote: > > On Tue, Feb 19, 2002 at 11:44:12PM +0100, Andreas Buening wrote: [fchown] > > Only an idea: Would it be possible to define these function > > (fchown(), link(), ) somewhere > > in the EMX headers? I'm thinking about something like > > > > #ifdef __DUMMY_FUNCTIONS__ > > int fchown(...) { errno = WHATEVER; return -1; } > > int link(...) { errno = WHATEVER; return -1; } > > ... > > #endif > > > > > > This could help in porting applications that use those functions > > but don't really depend on them. (at least every Unix program > > _should_ be able to handle a failure of those functions > > because the user really might have no permission to execute them). > > Using -D__DUMMY_FUNCTIONS__ as compiler flag would be much easier > > than patching every source file. And, the emx* libraries wouldn't > > need a recompilation. It's just an idea. > > Yes, and no. Partly this was attempted with POSIX/2, but the problem is > that it will work for some functions only: you can fake fchown() by > returning a success code - no security here, so accept whatever owner > the file should adopt. Yes, and therefore this shouldn't become the default. But if the programmer requests it explicitely by using -DDUMMY_FUNCTIONS then it's another story. He should know what he is doing. > OTOH, link() or chroot() are functions which > do essential things, and which must succeed, otherwise the surrounding > code might make the whole program fail. The functions are in the original > program for a reason. If a program requires that link() must succeed then it's broken. In general the only file system that supports hard links is /tmp on Unix systems (to my knowledge). symlink() might be another story. chown(), lchown(), fchown() are currently quite senseless, lstat() would be identical to stat(), mkfifo() and mknod() are currently not supported, and so on, and last but not least there are the all times favorites strcasecmp() and strncasecmp(). There could be two different implementations: one that lets these functions succeed and one that lets them fail. To make clearer what I'm thinking about: #ifdef EMULATE_POSIX_SUCCESS /* let them succeed and hope that it is not tested by the code */ # ifndef chown /* for the case config.h already handles chown() */ # define chown(x,y,z) (0) # endif # ifndef symlink /* for the case config.h already handles chown() */ inline int symlink(x,y,z) { /* copy the file? */ } # endif ... and so on #elif defined(EMULATE_POSIX_FAILURE) /* !EMULATE_POSIX_SUCCESS */ /* let them fail and hope that the code can handle this */ # ifndef chown inline int chown(x,y,z) { errno = EPERN; return -1; } # endif # ifndef symlink inline int symlink(x,y,z) { errno = EPERM; return -1; } # endif ... and so on #endif I know that this cannot be a general solution but in some cases it might help. Even if it wouldn't help in the most cases then it wouldn't harm because you would have to request this feature explicitely by -DEMULATE_POSIX_SUCCESS or -DEMULATE_POSIX_FAILURE. bye, Andreas -- One OS to rule them all, One OS to find them, One OS to bring them all and in the darkness bind them In the Land of Redmond where the Shadows lie. **= Email 2 ==========================** Date: Thu, 21 Feb 2002 01:04:13 -0500 From: Stepan Kazakov Subject: Re: mod_ssl to John Poltorak: look also at http://avn.nikoil.ru/ for openssl, apache-ssl & other stuff.. afaik there are multithreading problems in current openssl builds for os/2, so stunnel/2 does not working in threaded mode.. -- madded. [Red Hot Chili Hackers] **= Email 3 ==========================** Date: Thu, 21 Feb 2002 01:24:57 +1000 (EST) From: "Brian Havard" Subject: Re: mod_ssl On Wed, 20 Feb 2002 12:50:45 +0000, John Poltorak wrote: >Does anyone know of an OS/2 port of mod_ssl? http://silk.apana.org.au/apache/apache-ssl.html -- ______________________________________________________________________________ | Brian Havard | "He is not the messiah! | | brianh at kheldar.apana.org.au | He's a very naughty boy!" - Life of Brian | ------------------------------------------------------------------------------ **= Email 4 ==========================** Date: Thu, 21 Feb 2002 01:50:52 +0900 From: Masaru Nomiya Subject: Re: EMACS build error Hello, In the Message; Subject : Re: EMACS build error Message-ID : <20020220161238.J23629 at eyup.org> Date & Time: Wed, 20 Feb 2002 16:12:38 +0000 [John] == John Poltorak has written: Me> I have never used PGCC, so I have no idea. John> I've just tried another LD.EXE from here:- John> ftp://ftp.netlabs.org/pub/gcc/gcc-os2-3.0.3-beta-gcc.zip John> and this one works OK. Nice! Me> There doesn't exist "text.ptr", and it doen't make any .ptr file. Me> As you might know, make operation produce the file named emacs.res. Me> I suppose this error depends on your emx+gcc environment. John> I don't see the files emacs.res or os2/test*.ptr. But there are John> cross* and scroll* ptr files. Is it? There must exist text1.ptr and text2.ptr before make. The Patch-KIT make them as follows; 95-05-10 23:55 874 0 text1.ptr 95-05-10 23:56 874 0 text2.ptr Me> I have just made Emacs 20.7 to check your trouble. Me> But, I haven't got any error. John> If you run:- John> make 2>&1 | tee make.log John> could you leave make.log somewhere? It may give me a clue about what is John> missing. OK! I will send this to you. # Now, I can read MS Office and PDF files on the Emacs 20.7. # How nice! Don't you think so? --- Masaru Nomiya mail-to: nomiya at ttmy.ne.jp "No WIndows, no gains!" ..... "Why, I am wrong?" -- Bill -- **= Email 5 ==========================** Date: Thu, 21 Feb 2002 08:20:43 +0000 From: John Poltorak Subject: Re: make 3.79.1 On Wed, Feb 20, 2002 at 07:35:21PM +0100, Andreas Buening wrote: > John Poltorak wrote: > > > > make[2]: $SHELL changed (was `$', now `sh') > > make[2]: $SHELL changed (was `echo 2.52f|sed 's/\./\\./g'', now `sh') > > make[2]: Nothing to be done for `install-exec-am'. > > > > ???... > > Should be fixed now: > ftp://ftp.unixos2.com/incoming/make-3_79_1-bin-beta2.zip I still get similar errors. Also, I can't use this make for building emacs 20.7 (see ongoing thread about emacs), whereas 3.76.1 works OK. > bye, > Andreas > > -- > One OS to rule them all, One OS to find them, > One OS to bring them all and in the darkness bind them > In the Land of Redmond where the Shadows lie. -- John **= Email 6 ==========================** Date: Thu, 21 Feb 2002 08:42:28 -0500 From: Jack Troughton Subject: Re: OSFree... "Sebastian Wittmeier (ShadoW)" wrote: > > On Thu, 21 Feb 2002 13:25:36 -0500, Jack Troughton wrote: > > >in any way whatsoever. Do you remember the hullabaloo over the Warpmedia > >player some months ago? > [...] > >That they > >weren't is of no consequence; they apologised, made the source > >available, and everybody involved dropped the issue as not worth > >pursuing any further. > > That means, Warpmedia source is released? > Several people asked for it, but the developers never wanted. I dunno... might have been the other player (Warpflash? Warp something...). However, the one in question did release the source. They had to. The original authors were asking them for it, and the license clearly states the terms, so... Regards, Jack **= Email 7 ==========================** Date: Thu, 21 Feb 2002 09:04:28 -0800 (PST) From: James Cannon Subject: Re: terminfo database > +> What is the "correct" location for a terminfo > database? > +> > +> /usr/share/terminfo ?... > > +looks ok to me > > On UnixWare 7+ it is '/usr/share/lib/terminfo', > OpenServer 5+ uses > '/usr/lib/terminfo/' and Linux uses > '/usr/share/terminfo/' from what > I can determine. I don't have access to Solaris but > I'd be willing to > bet SUN has their own pet location for terminfo. So > the location is > not fixed across platforms and there is no real > "correct" location. > Pick one, use it and stick to it for unixos2. On SunOS 5.6 (SPARC station 20) it is located in /usr/share/lib/terminfo/*/ > Gene > -- > +=========================-=>Unix & > OS/2<=-=========================+ > # Owner and C.E.O. - ERA Computer Consulting - > Jackson, TN USA # > # OS/2, UnixWare, OpenServer & Linux Business > Computing Solutions # > # Please visit our www pages at > http://eracc.hypermart.net/ # > +===================================================================+ > We run IBM OS/2 v.4.00, Revision > 9.036 > Sysinfo: 40 Processes, 157 Threads, uptime is 9d > 1h 6m 26s 307ms > > ===== Sincerely, James Cannon Using OS/2 Warp in the beautiful Wine Country of Northen California! __________________________________________________ Do You Yahoo!? Yahoo! Sports - Coverage of the 2002 Olympic Games http://sports.yahoo.com **= Email 8 ==========================** Date: Thu, 21 Feb 2002 09:06:13 -0800 (PST) From: James Cannon Subject: Re: URL for CVS tarball --- John Poltorak wrote: > > Can someone provide me with a URL for the latest CVS > tarball? > > > I have managed to retrieve it using:- > > http://ccvs.cvshome.org/servlets/ProjectDownloadList?action=download&dlID=41 > > > but it get saved as:- > > cvs-1.11.1p1.tar.gz%3FJServSessionIdservlets=a2ondtxti1 Can you not just rename it to: cvs-1.11.1p1.tar.gz ? > > > I'd prefer an FTP reference, but can't find one. > > > -- > John > > ===== Sincerely, James Cannon Using OS/2 Warp in the beautiful Wine Country of Northen California! __________________________________________________ Do You Yahoo!? Yahoo! Sports - Coverage of the 2002 Olympic Games http://sports.yahoo.com **= Email 9 ==========================** Date: Thu, 21 Feb 2002 09:31:18 +0100 From: Holger Veit Subject: Re: EMACS build error On Thu, Feb 21, 2002 at 12:48:09AM +0100, Andreas Buening wrote: > > Yes, and no. Partly this was attempted with POSIX/2, but the problem is > > that it will work for some functions only: you can fake fchown() by > > returning a success code - no security here, so accept whatever owner > > the file should adopt. > > Yes, and therefore this shouldn't become the default. But if > the programmer requests it explicitely by using -DDUMMY_FUNCTIONS > then it's another story. He should know what he is doing. The problem is, you need to recompile the code when you want to migrate to real functions. There is still stoneold code out there with severe bugs due to such quick hacks, which we won't get rid of (think of the persistency of broken makes and shells). > > OTOH, link() or chroot() are functions which > > do essential things, and which must succeed, otherwise the surrounding > > code might make the whole program fail. The functions are in the original > > program for a reason. > > If a program requires that link() must succeed then it's broken. > In general the only file system that supports hard links is /tmp > on Unix systems (to my knowledge). symlink() might be another story. You are wrong. Any Unix file system supports hard links, and has done so for decades. The thing that makes link() unattractive compared to symlink(), though, is that symlink can span over several mounted file systems, whereas a link is (due to its mechanism - is is basically an alias to an inode) restricted to the filesystem where the original file is located. The latter is not much a problem, because nowadays we have large disks, so splitting into /var, /usr, /opt, /home filesystems is often no longer done (except on larger servers for good reasons). > chown(), lchown(), fchown() are currently quite senseless, > lstat() would be identical to stat(), mkfifo() and mknod() > are currently not supported, and so on, and last but not least > there are the all times favorites strcasecmp() and strncasecmp(). > There could be two different implementations: one that lets these > functions succeed and one that lets them fail. To make clearer > what I'm thinking about: > > #ifdef EMULATE_POSIX_SUCCESS > /* let them succeed and hope that it is not tested by the code */ > # ifndef chown /* for the case config.h already handles chown() */ > # define chown(x,y,z) (0) > # endif > > # ifndef symlink /* for the case config.h already handles chown() */ > inline int symlink(x,y,z) { /* copy the file? */ } > # endif Separation between *SUCCESS and *FAILURE is too complicated, IMHO, because you would sometimes need both: let fchown succeed and symlink fail, for instance. You'll encounter a clash if both #ifdefs check for #ifndef fchown. You also need some configuration file that will disable the replacement code by #define fchown if you have a real version. *Sometimes* (namely in the autoconf case), you have a config.h file which usually declares #define HAS_FCHOWN or similar; so it is a better idea to use these as switcher symbols. Holger -- Please update your tables to my new e-mail address: holger.veit$ais.fhg.de (replace the '$' with ' at ' -- spam-protection) **= Email 10 ==========================** Date: Thu, 21 Feb 2002 09:34:11 +0000 From: John Poltorak Subject: Re: Re: EMACS build error On Thu, Feb 21, 2002 at 01:50:52AM +0900, Masaru Nomiya wrote: > Is it? > > There must exist text1.ptr and text2.ptr before make. > The Patch-KIT make them as follows; > > 95-05-10 23:55 874 0 text1.ptr > 95-05-10 23:56 874 0 text2.ptr > > Me> I have just made Emacs 20.7 to check your trouble. > Me> But, I haven't got any error. > John> If you run:- > John> make 2>&1 | tee make.log > John> could you leave make.log somewhere? It may give me a clue about what is > John> missing. > > OK! > I will send this to you. Thanks. I think I'm nearly done now, but I get a core dump before make has finished. Any ideas on this one:- Loading vc-hooks... Loading ediff-hook... ((30564 . 1688) (6955 . 0) (453 . 41) 58035 49793 (7 . 5) (0 . 0)) Finding pointers to doc strings... Finding pointers to doc strings...done Wrote c:/eval/emacs/emacs-20.7/lib-src/fns-20.7.1.el Dumping under names emacs and emacs-20.7.1 emxbind 0.9d -- Copyright (c) 1991-1997 by Eberhard Mattes 7 resources in 2 objects 506864 pure bytes used make6[1]: Leaving directory `/eval/emacs/emacs-20.7/src' cd leim && make6 make6[1]: Entering directory `/eval/emacs/emacs-20.7/leim' ../src/emacs.exe -batch --no-init-file --no-site-file -l ../lisp/international/titdic-cnv \ --eval '(batch-titdic-convert t)' -dir quail ./CXTERM-DIC Process terminated by SIGSEGV core dumped make6[1]: *** [quail/CTLau.el] Error 3 make6[1]: Leaving directory `/eval/emacs/emacs-20.7/leim' make6: *** [leim] Error 2 By this point a number of exe's including emacs, temacs, emacsclient and emacsserver have already been built, so should I have something usable now? > --- > Masaru Nomiya mail-to: nomiya at ttmy.ne.jp > > "No WIndows, no gains!" ..... "Why, I am wrong?" > > -- Bill -- -- John **= Email 11 ==========================** Date: Thu, 21 Feb 2002 09:44:59 +0000 From: John Poltorak Subject: Re: mod_ssl On Thu, Feb 21, 2002 at 01:04:13AM -0500, Stepan Kazakov wrote: > to John Poltorak: > look also at http://avn.nikoil.ru/ for openssl, apache-ssl & other stuff.. including a nice picture of a pussy cat :-)... I also noticed a build of SOCKS5. I've never managed to get that working... Do you have any notes for setting it up? > -- > madded. [Red Hot Chili Hackers] -- John **= Email 12 ==========================** Date: Thu, 21 Feb 2002 10:03:09 +0000 From: John Poltorak Subject: Re: make 3.79.1 On Wed, Feb 20, 2002 at 07:35:21PM +0100, Andreas Buening wrote: > Should be fixed now: > ftp://ftp.unixos2.com/incoming/make-3_79_1-bin-beta2.zip Can anyone get this to work with the Makefile in? :- emx\bsd\libc > bye, > Andreas > > -- > One OS to rule them all, One OS to find them, > One OS to bring them all and in the darkness bind them > In the Land of Redmond where the Shadows lie. -- John **= Email 13 ==========================** Date: Thu, 21 Feb 2002 10:13:06 -0600 From: email at eracc.hypermart.net Subject: Re: terminfo database In: On: Thu, 21 Feb 2002 10:48:43 -0500 (EST) Screaming: Re: terminfo database "Thomas E. Dickey" did rant: +On Thu, 21 Feb 2002, John Poltorak wrote: +> What is the "correct" location for a terminfo database? +> +> /usr/share/terminfo ?... +looks ok to me On UnixWare 7+ it is '/usr/share/lib/terminfo', OpenServer 5+ uses '/usr/lib/terminfo/' and Linux uses '/usr/share/terminfo/' from what I can determine. I don't have access to Solaris but I'd be willing to bet SUN has their own pet location for terminfo. So the location is not fixed across platforms and there is no real "correct" location. Pick one, use it and stick to it for unixos2. Gene -- +=========================-=>Unix & OS/2<=-=========================+ # Owner and C.E.O. - ERA Computer Consulting - Jackson, TN USA # # OS/2, UnixWare, OpenServer & Linux Business Computing Solutions # # Please visit our www pages at http://eracc.hypermart.net/ # +===================================================================+ We run IBM OS/2 v.4.00, Revision 9.036 Sysinfo: 40 Processes, 157 Threads, uptime is 9d 1h 6m 26s 307ms **= Email 14 ==========================** Date: Thu, 21 Feb 2002 10:29:54 -0600 From: email at eracc.hypermart.net Subject: Re: Has anyone ported nmap? In: <200201310942.0000000L at warped.netlabs.org> On: Thu, 31 Jan 2002 09:46:06 +0100 (CET) Screaming: Re: Has anyone ported nmap? "Adrian Gschwend" did rant: +On Wed, 30 Jan 2002 22:53:11 +0000, John Poltorak wrote: +>nmap has not been ported AFAICT, although a number of people would like to +>see an OS/2 version, so please go ahead and feel free to ask for any +>assistance. +I just talked to Brian Smith (aka nuke), he ported it but does not want +to maintain it. However, he will send me all the source and I will upload +that to netlabs later. Will be a good base I think. +I will let you know as soon as I got it +cu +Adrian Hi Adrian, I *have* to ask ... have you got it yet? I've been holding on to your message to remember to check for 'nmap' and, like a kid anticipating Christmas, I am "dying" for it to get here. :-) Gene -- +=========================-=>Unix & OS/2<=-=========================+ # Owner and C.E.O. - ERA Computer Consulting - Jackson, TN USA # # OS/2, UnixWare, OpenServer & Linux Business Computing Solutions # # Please visit our www pages at http://eracc.hypermart.net/ # +===================================================================+ We run IBM OS/2 v.4.00, Revision 9.036 Sysinfo: 41 Processes, 166 Threads, uptime is 9d 1h 23m 16s 244ms **= Email 15 ==========================** Date: Thu, 21 Feb 2002 10:41:14 +0000 From: John Poltorak Subject: c++filt Which package should c++filt appear in? It's part of GNU gcc 3.0.3 but not included in the OS/2 port, however it is included in the OS/2 port of binutils, but not the original. I'm tempted to move it from BINUTILS to GCC as far as UnixOS/2 packages are concerned. Is this going to cause any problems? -- John **= Email 16 ==========================** Date: Thu, 21 Feb 2002 10:48:43 -0500 (EST) From: "Thomas E. Dickey" Subject: Re: terminfo database On Thu, 21 Feb 2002, John Poltorak wrote: > What is the "correct" location for a terminfo database? > > /usr/share/terminfo ?... looks ok to me > I assume that the definitive terminfo database is the one included in > ncurses... yes. > Does it contain anything related to OS/2? I've just notice entries for > 'emx-base' and 'os2' and was never aware of them previously. In terminfo.src there are some relatively-obsolete entries for OS/2; emx.src has the updated ones. > What are they for? some entries are used as building blocks (not intended to be used as a complete entry). -- T.E.Dickey http://invisible-island.net ftp://invisible-island.net **= Email 17 ==========================** Date: Thu, 21 Feb 2002 10:49:35 +0100 (CET) From: Mentore Siesto Subject: Re: OSFree... On Wed, 20 Feb 2002, Kris Steenhaut wrote: KS >Mentore Siesto schreef: KS > KS >> KS > KS >> KS >Could only be done in court. KS >> KS >> But - I'm afraid - it CAN definitely be done. KS >You damn well know you can't go to court in Europe about patented KS >software, as software can't be patented here. bsa seems to be working KS >on it, but that's quite another matter. Didn't know about situation in Europe. But Hobbes is in New Mexico... And OSFree is a Russian project. If IBM can sue someone, it will do this I'm afraid. -- Mentore Siesto Team OS/2 Italia (http://www.teamos2.it) Home page: http://www.geocities.com/mentoruccio/ **= Email 18 ==========================** Date: Thu, 21 Feb 2002 10:56:01 +0000 From: John Poltorak Subject: Re: emx\lib\*.obj ? On Mon, Feb 18, 2002 at 01:30:35PM +0100, Holger Veit wrote: > On Mon, Feb 18, 2002 at 12:09:54PM +0000, John Poltorak wrote: > > > > Why do we have .o and .obj files such as crt0.obj in emx\lib ? > > We need them. crt0.* is the startup code for a program which does > some initialization, and then calls main(). This code must be outside > a library, so that it gets linked into the executable in the first place. Is there a Makefile for them anywhere? I'd like to try building them with gcc 3.0.3. > Holger > > -- > Please update your tables to my new e-mail address: > holger.veit$ais.fhg.de (replace the '$' with ' at ' -- spam-protection) > -- John **= Email 19 ==========================** Date: Thu, 21 Feb 2002 10:59:14 -0500 From: Jack Troughton Subject: Re: mod_ssl Lyn St George wrote: > >Everything apart from usr/libexec/libssl.so looks like text, scripts or > >data files. This seems to be the only binary, so I guess I'm only looking > >for a libssl.a.... > > I'm not sure that I understand the point of a binary distribution like this, and > if it can be done on OS/2 then it would need to be a .dll. Actually, you know what would be really excellent? If this ssl could be compiled into a rexx callable dll, that could be used with rxsock to create secure connections to other systems. I don't have the "skillz" to do this, but sure wish I did:) I know this doesn't have any direct bearing on unix-os2, but I just thought I'd pass that one in there... Regards, Jack **= Email 20 ==========================** Date: Thu, 21 Feb 2002 11:01:00 -0500 From: Jack Troughton Subject: Re: OSFree... Kris Steenhaut wrote: > > Mentore Siesto schreef: > > > KS > > > KS >Could only be done in court. > > > > But - I'm afraid - it CAN definitely be done. > > > > You damn well know you can't go to court in Europe about patented software, as > software can't be patented here. bsa seems to be working on it, but that's quite > another matter. It wouldn't be a patent issue; it would be a copyright issue. AFAIK, the common market is still signatory to the Berne Convention. Regards, Jack **= Email 21 ==========================** Date: Thu, 21 Feb 2002 11:07:02 +0000 From: John Poltorak Subject: emacs v21.1 I see that emacs 21.1 has been available since October whereas 20.7 has been out since June 2000. Has anyone tried applying the OS/2 patches to the most recent version? -- John **= Email 22 ==========================** Date: Thu, 21 Feb 2002 12:15:18 +0000 From: John Poltorak Subject: Re: Re: EMACS build error On Thu, Feb 21, 2002 at 07:07:43PM +0900, Masaru Nomiya wrote: > > John> Process terminated by SIGSEGV > John> core dumped > John> make6[1]: *** [quail/CTLau.el] Error 3 > John> make6[1]: Leaving directory `/eval/emacs/emacs-20.7/leim' > John> make6: *** [leim] Error 2 > > Haven't you forgotten to input > > SET NAME=John Poltorak > > before starting make? > You must write this in your CONFIG.SYS, unless you can't use Emacs. I didn't think that would affect the build... Does it mean my name gets compiled in or something? I have now got it built, with much help from you. Many thanks! BTW can I run it in text mode? > Regards, > > --- > Masaru Nomiya mail-to: nomiya at ttmy.ne.jp > > "No WIndows, no gains!" ..... "Why, I am wrong?" > > -- Bill -- -- John **= Email 23 ==========================** Date: Thu, 21 Feb 2002 12:57:34 +0000 From: John Poltorak Subject: Courier IMAP Has anyone ever attempted to port the Courier IMAP server? :- http://www.inter7.com/courierimap/ -- John **= Email 24 ==========================** Date: Thu, 21 Feb 2002 13:12:09 -0500 From: Jack Troughton Subject: Re: mod_ssl Lyn St George wrote: > > On Thu, 21 Feb 2002 10:59:14 -0500, Jack Troughton wrote: > > >Lyn St George wrote: > > > > > > > >> >Everything apart from usr/libexec/libssl.so looks like text, scripts or > >> >data files. This seems to be the only binary, so I guess I'm only looking > >> >for a libssl.a.... > >> > >> I'm not sure that I understand the point of a binary distribution like this, and > >> if it can be done on OS/2 then it would need to be a .dll. > > > >Actually, you know what would be really excellent? > > > >If this ssl could be compiled into a rexx callable dll, that could be > >used with rxsock to create secure connections to other systems. > > Wouldn't this be openssl, rather than mod_ssl (which is only for > Apache)? When you build openssl, you get both openssl.dll and > sslcrypt.dll (plus various .a's and .lib's and the .exe). Can't these > be called by Rexx? (I know nothing about Rexx, and often wish > I did) You're right... it would be openssl. No, you can't call any dll from rexx; it has to be built specifically for it. There are some APIs available for doing this; you can find them in the OS/2 programming reference docs. I imagine that I _could_ do it, but the amount of stuff I'd have to learn to do it would be pretty extreme... and I don't have the time to do that learning right now. Maybe someday... Regards, Jack **= Email 25 ==========================** Date: Thu, 21 Feb 2002 13:21:17 +0100 (CET) From: "Sebastian Wittmeier (ShadoW)" Subject: Re: OSFree... On Thu, 21 Feb 2002 13:25:36 -0500, Jack Troughton wrote: >in any way whatsoever. Do you remember the hullabaloo over the Warpmedia >player some months ago? [...] >That they >weren't is of no consequence; they apologised, made the source >available, and everybody involved dropped the issue as not worth >pursuing any further. That means, Warpmedia source is released? Several people asked for it, but the developers never wanted. Sebastian **= Email 26 ==========================** Date: Thu, 21 Feb 2002 13:23:17 +0100 From: Holger Veit Subject: Re: Re: EMACS build error On Thu, Feb 21, 2002 at 07:07:43PM +0900, Masaru Nomiya wrote: [...] > Haven't you forgotten to input > > SET NAME=John Poltorak > > before starting make? > You must write this in your CONFIG.SYS, unless you can't use Emacs. Who invented this? There are already several defined environment variables for such a purpose: LOGNAME, USER, so why a new one? And why a fullname? To make GNUS or movemail happy? Holger -- Please update your tables to my new e-mail address: holger.veit$ais.fhg.de (replace the '$' with ' at ' -- spam-protection) **= Email 27 ==========================** Date: Thu, 21 Feb 2002 13:25:36 -0500 From: Jack Troughton Subject: Re: OSFree... Kris Steenhaut wrote: > > Jack Troughton schreef: > > > Kris Steenhaut wrote: > > > > > > Mentore Siesto schreef: > > > > > > > KS > > > > > KS >Could only be done in court. > > > > > > > > But - I'm afraid - it CAN definitely be done. > > > > > > > > > > You damn well know you can't go to court in Europe about patented software, as > > > software can't be patented here. bsa seems to be working on it, but that's quite > > > another matter. > > > > It wouldn't be a patent issue; it would be a copyright issue. AFAIK, the > > common market is still signatory to the Berne Convention. > > > > In a nutshell: programs can be copyrighted, source codes can't (The usa and Japan > being the exception). No, that's not true. Source code is automatically copyrighted to the author as soon as they create it. This is definitely the case in Canada, and is the case in every country that is signatory to the Berne Convention on Intellectual Property Rights. In fact, if there's a conceptual problem, it's on the protection offered to binaries, not to source code; however, common practice means that this isn't a problem (and in every country that uses the common law approach to civil legal matters, common practice is very important... and common law is certainly used in more countries than just the US and Japan). If there's no copyright on source code, how is the GPL enforceable? It is, after all, a set of rules about what can and cannot be done with source code and the binaries that are derived from that source code. If there was no copyright on source code, the GPL wouldn't be enforceable in any way whatsoever. Do you remember the hullabaloo over the Warpmedia player some months ago? The authors used some GPL source in it and didn't make their source available at the same time that they made the binaries available. They got drawn over the coals for that one, and could have been sued by the original authors of the code they used for violating the license agreement accompanying the source code. That they weren't is of no consequence; they apologised, made the source available, and everybody involved dropped the issue as not worth pursuing any further. If there was no copyright protection offered on source code, they could have thumbed their noses at the original authors without fear of legal punishment. In short, the distinction you draw between "programs" and "source codes" is an artificial one at best... even in the legal sense, and has no bearing whatsoever on the question at hand regarding leaked IBM source code for components of OS/2. Regards, Jack **= Email 28 ==========================** Date: Thu, 21 Feb 2002 13:26:39 +0100 From: Holger Veit Subject: Re: emx\lib\*.obj ? On Thu, Feb 21, 2002 at 10:56:01AM +0000, John Poltorak wrote: > On Mon, Feb 18, 2002 at 01:30:35PM +0100, Holger Veit wrote: > > On Mon, Feb 18, 2002 at 12:09:54PM +0000, John Poltorak wrote: > > > > > > Why do we have .o and .obj files such as crt0.obj in emx\lib ? > > > > We need them. crt0.* is the startup code for a program which does > > some initialization, and then calls main(). This code must be outside > > a library, so that it gets linked into the executable in the first place. > > Is there a Makefile for them anywhere? > > I'd like to try building them with gcc 3.0.3. They get built during the build of the EMX libraries, so the code is in the EMX source. But this source is a major mess which requires dmake.exe. I have isolated them for libemu; have to look up the makefile at home later. But you don't need to verify them with gcc-3.0.3 - they are assembler code. Holger -- Please update your tables to my new e-mail address: holger.veit$ais.fhg.de (replace the '$' with ' at ' -- spam-protection) **= Email 29 ==========================** Date: Thu, 21 Feb 2002 13:42:33 +0100 From: Kris Steenhaut Subject: Re: OSFree... Mentore Siesto schreef: > > Didn't know about situation in Europe. But Hobbes is in New Mexico... And > OSFree is a Russian project. If IBM can sue someone, it will do this I'm > afraid. > Quite right. They could sue Hobbes maybe, they can't sue in Europe (and Russia) about patented software. It's as simple as that. Meanwhile the bsa is working on it. But they won't have it all their way. No way! http://www.freepatents.org/ http://swpat.ffii.org/vreji/papri/eubsa-swpat0202/ -- Groeten uit Gent, Kris **= Email 30 ==========================** Date: Thu, 21 Feb 2002 13:48:34 +0000 From: John Poltorak Subject: URL for CVS tarball Can someone provide me with a URL for the latest CVS tarball? I have managed to retrieve it using:- http://ccvs.cvshome.org/servlets/ProjectDownloadList?action=download&dlID=41 but it get saved as:- cvs-1.11.1p1.tar.gz%3FJServSessionIdservlets=a2ondtxti1 I'd prefer an FTP reference, but can't find one. -- John **= Email 31 ==========================** Date: Thu, 21 Feb 2002 15:17:47 +0000 From: John Poltorak Subject: terminfo database What is the "correct" location for a terminfo database? /usr/share/terminfo ?... I assume that the definitive terminfo database is the one included in ncurses... Does it contain anything related to OS/2? I've just notice entries for 'emx-base' and 'os2' and was never aware of them previously. What are they for? -- John **= Email 32 ==========================** Date: Thu, 21 Feb 2002 16:44:06 +0000 From: John Poltorak Subject: bfd.dll Would it be possible to use the same BFD.DLL in GCC 3.0.3 as in the BINUTILS? Is there a makefile available, so that I can try rebuilding it? -- John **= Email 33 ==========================** Date: Thu, 21 Feb 2002 17:12:05 +0000 From: John Poltorak Subject: UnixOS/2 distro Here is a list of the most recent versions of many of the important apps in Slackware. Some of these are not in the latest Slackware distro yet. It would be nice to get as many of these into UnixOS/2 as possible. In fact, most of these apps are already available on OS/2, although a few are just slightly back levelled, so we are not far off having a credible distro... gawk: ftp://ftp.gnu.org/pub/gnu/gawk/gawk-3.1.0.tar.gz grep: ftp://ftp.gnu.org/pub/gnu/grep/grep-2.4.2.tar.gz sed: ftp://ftp.gnu.org/pub/gnu/sed/sed-3.02.tar.gz patch: ftp://ftp.gnu.org/pub/gnu/patch/patch-2.5.4.tar.gz m4: ftp://ftp.gnu.org/pub/gnu/m4/m4-1.4.tar.gz gmake: ftp://ftp.gnu.org/pub/gnu/make/make-3.79.1.tar.gz gettext: ftp://ftp.gnu.org/pub/gnu/gettext/gettext-0.11.tar.gz fileutls: ftp://ftp.gnu.org/pub/gnu/fileutils/fileutils-4.1.tar.gz find: ftp://ftp.gnu.org/pub/gnu/findutils/findutils-4.1.tar.gz sh_utils: ftp://ftp.gnu.org/pub/gnu/sh-utils/sh-utils-2.0.tar.gz txtutils: ftp://ftp.gnu.org/pub/gnu/textutils/textutils-2.0.tar.gz autoconf: ftp://alpha.gnu.org/pub/gnu/autoconf/autoconf-2.52h.tar.gz automake: ftp://ftp.gnu.org/pub/gnu/automake/automake-1.5.tar.gz cpio: ftp://ftp.gnu.org/pub/gnu/cpio/cpio-2.4.2.tar.gz diff: ftp://ftp.gnu.org/pub/gnu/diffutils/diffutils-2.7.tar.gz bison: ftp://ftp.gnu.org/pub/gnu/bison/bison-1.33.tar.gz bash: ftp://ftp.gnu.org/pub/gnu/bash/bash-2.05a.tar.gz bash1: ftp://ftp.gnu.org/pub/gnu/bash/bash-1.14.7.tar.gz tcsh: ftp://ftp.astron.com/pub/tcsh/tcsh-6.11.tar.gz gzip: ftp://gatekeeper.dec.com/pub/GNU/gzip/gzip-1.2.4.tar.gz bzip2: ftp://sources.redhat.com/pub/bzip2/v102/bzip2-1.0.2.tar.gz less: http://www.greenwoodsoftware.com/less/less-374.tar.gz binutils: ftp://ftp.gnu.org/pub/gnu/binutils/binutils-2.11.2.tar.gz gcc: ftp://ftp.gnu.org/pub/gnu/gcc/gcc-3.0.3/gcc-3.0.3.tar.gz gcc_g77: ftp://ftp.gnu.org/pub/gnu/gcc/gcc-3.0.3/gcc-g77-3.0.3.tar.gz gcc_objc: ftp://ftp.gnu.org/pub/gnu/gcc/gcc-3.0.3/gcc-objc-3.0.3.tar.gz perl: http://www.cpan.org/src/stable.tar.gz python: ftp://ftp.python.org/pub/python/2.2/Python-2.2.tgz flex: ftp://ftp.gnu.org/pub/gnu/non-gnu/flex/flex-2.5.4a.tar.gz ncurses: ftp://invisible-island.net/ncurses/ncurses-5.2.tar.gz slang: ftp://space.mit.edu/pub/davis/slang/v1.4/slang-1.4.5.tar.gz apache: http://www.apache.org/dist/httpd/apache_1.3.22.tar.gz lynx: ftp://lynx.isc.org/lynx/lynx2.8.4/lynx2-8-4.tar.gz openssl: http://www.openssl.org/source/openssl-engine-0.9.6b.tar.gz tin: ftp://ftp.tin.org/pub/news/clients/tin/unstable/tin-1.5.11.tar.gz bind: ftp://ftp.isc.org/isc/bind/src/8.3.1/bind-src.tar.gz ftchmail: http://www.tuxedo.org/~esr/fetchmail/fetchmail-5.9.8.tar.gz wget: ftp://ftp.gnu.org/pub/gnu/wget/wget-1.8.1.tar.gz man: ftp://ftp.win.tue.nl/pub/linux-local/utils/man/man-1.5j.tar.gz groff: ftp://ftp.ffii.org/pub/groff/groff-1.17.2.tar.gz texinfo: ftp://ftp.gnu.org/pub/gnu/texinfo/texinfo-4.0.tar.gz manpages: ftp://ftp.win.tue.nl/pub/linux-local/manpages/man-pages-1.47.tar.gz rcs: http://www.cs.purdue.edu/homes/trinkle/RCS/rcs-5.7.tar.Z sox: http://prdownloads.sourceforge.net/sox/sox-12.17.3.tar.gz a2ps: ftp://ftp.enst.fr/pub/unix/a2ps/a2ps-4.13b.tar.gz zlib: ftp://ftp.info-zip.org/pub/infozip/zlib/zlib-1.1.3.tar.gz zip: ftp://ftp.info-zip.org/pub/infozip/src/zip23.tar.gz unzip: ftp://ftp.info-zip.org/pub/infozip/src/unzip542.tar.gz jed: ftp://space.mit.edu/pub/davis/jed/v0.99/jed-B0.99-15.tar.gz oggutils: http://www.vorbis.com/files/rc3/unix/vorbis-tools-1.0rc3.tar.gz bitchx: ftp://ftp.bitchx.org/pub/BitchX/source/ircii-pana-1.0c18.tar.gz vim: ftp://ftp.vim.org/pub/vim/unix/vim-6.0.tar.bz2 elvis: ftp://ftp.fh-wedel.de/pub/fh-wedel/staff/herbert/elvis/elvis-2.2g-os2.tar.gz emacleim: ftp://ftp.gnu.org/pub/gnu/emacs/leim-21.1.tar.gz emacsbin: ftp://ftp.gnu.org/pub/gnu/emacs/emacs-21.1.tar.gz ghostscr: http://prdownloads.sourceforge.net/ghostscript/ghostscript-7.04.tar.gz mysql: http://www.mysql.com/Downloads/MySQL-3.23/mysql-3.23.49.tar.gz readline: ftp://ftp.gnu.org/pub/gnu/readline/readline-4.2a.tar.gz pine: ftp://ftp.cac.washington.edu/pine/pine.tar.Z cvs: ftp://ftp.cvshome.org/pub/cvs-1.11.1/cvs-1.11.1p1.tar.gz What have I missed (apart from XFree86 and X apps) ? -- John **= Email 34 ==========================** Date: Thu, 21 Feb 2002 17:14:59 +0300 (MSK) From: "Andrew Belov" Subject: Re: URL for CVS tarball On Thu, 21 Feb 2002 13:48:34 +0000, John Poltorak wrote: >Can someone provide me with a URL for the latest CVS tarball? IMHO, there were none beyond the 1.11 "patch" that shortly followed the 1.11 release. I had to checkout the "current" tree from :pserver:anonymous at cvs.cvshome.org:/home2/cvsroot. It seemed to be under heavy reconstruction for over a year now, at least it now requires automake. **= Email 35 ==========================** Date: Thu, 21 Feb 2002 18:59:10 -0500 From: Thomas Dickey Subject: Re: UnixOS/2 distro On Thu, Feb 21, 2002 at 05:12:05PM +0000, John Poltorak wrote: > > Here is a list of the most recent versions of many of the important apps > in Slackware. Some of these are not in the latest Slackware distro yet. It > would be nice to get as many of these into UnixOS/2 as possible. In fact, > most of these apps are already available on OS/2, although a few are just > slightly back levelled, so we are not far off having a credible distro... > > > > ncurses: ftp://invisible-island.net/ncurses/ncurses-5.2.tar.gz the official page for ncurses is on ftp.gnu.org (though I encourage people to go to my ftp area for patches, of course). -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 36 ==========================** Date: Thu, 21 Feb 2002 19:07:43 +0900 From: Masaru Nomiya Subject: Re: EMACS build error Hello, In the Message; Subject : Re: Re: EMACS build error Message-ID : <20020221093411.P23629 at eyup.org> Date & Time: Thu, 21 Feb 2002 09:34:11 +0000 [John] == John Poltorak has written: John> I think I'm nearly done now, but I get a core dump before make has John> finished. Any ideas on this one:- [...] John> make6[1]: Leaving directory `/eval/emacs/emacs-20.7/src' John> cd leim && make6 John> make6[1]: Entering directory `/eval/emacs/emacs-20.7/leim' John> ../src/emacs.exe -batch --no-init-file --no-site-file -l ../lisp/international/titdic-cnv \ John> --eval '(batch-titdic-convert t)' -dir quail ./CXTERM-DIC John> Process terminated by SIGSEGV John> core dumped John> make6[1]: *** [quail/CTLau.el] Error 3 John> make6[1]: Leaving directory `/eval/emacs/emacs-20.7/leim' John> make6: *** [leim] Error 2 Haven't you forgotten to input SET NAME=John Poltorak before starting make? You must write this in your CONFIG.SYS, unless you can't use Emacs. Regards, --- Masaru Nomiya mail-to: nomiya at ttmy.ne.jp "No WIndows, no gains!" ..... "Why, I am wrong?" -- Bill -- **= Email 37 ==========================** Date: Thu, 21 Feb 2002 19:26:27 -0500 From: Henry Sobotka Subject: Re: Building Perl with gcc v3.0.3 John Poltorak wrote: > > I tried building Perl using gcc 3.0.3 but it fails with repeated errors > Does this indicate a problem with gcc 3.0.3? No such problem here, just the *.o vs. *.obj one I mentioned in a previous post. h~ **= Email 38 ==========================** Date: Thu, 21 Feb 2002 19:54:06 +0900 From: Masaru Nomiya Subject: Another Patch (WAs : Re: EMACS build error) Hello, I'm so soory, I have fogotten to inform another patch. Mr. Nakagawa made this patch. By applying this, you can use ange-ftp like this. C-x C-f /anonymous at unixos2.com:/incoming Soon, file list will be shown. Then select any file (ex. osf_en.zip) and push Enter key. You can see the contents of os2_en.zip. But you can't download by this. Regards, -- Masary Nomiya mail-to: nomiya at ttmy.ne.jp Emacs broden your world! ^L diff -ru emacs-20.7-tmp/src/os2.c emacs-20.7-tmp2/src/os2.c --- emacs-20.7-tmp/src/os2.c Tue Jun 5 11:11:42 2001 +++ emacs-20.7-tmp2/src/os2.c Mon Jun 25 14:03:28 2001 at at -585,6 +585,15 at at return NULL; } +char * +getwd (pathname) + char *pathname; +{ + _getcwd2(pathname, MAXPATHLEN); + + return pathname; +} + DEFUN ("os2-current-directory", Fos2_current_directory, Sos2_current_directory, 0, 1, 0, "Return the current directory path, as a string.\n\ diff -ru emacs-20.7-tmp/src/s/os2.h emacs-20.7-tmp2/src/s/os2.h --- emacs-20.7-tmp/src/s/os2.h Fri Jun 15 14:34:30 2001 +++ emacs-20.7-tmp2/src/s/os2.h Mon Jun 25 14:07:32 2001 at at -180,7 +180,7 at at #if 0 #define HAVE_DYNAMIC_LOADER #define HAVE_DLD -#endif +#endif /* 0 */ /* beginning of code from dgux4.h */ #ifndef NOT_C_CODE at at -236,10 +236,8 at at extern void DebugSnap( ) ; -#if 0 /* Tell emacs.c not to define abort. */ #define NO_ABORT -#endif /* 0 */ #define HAVE_WAIT_HEADER #define WAITTYPE int at at -252,4 +250,6 at at #define WCOREDUMP(w) ((w&0200) != 0) #define POSIX_SIGNALS + +#define HAVE_GETWD diff -ru emacs-20.7-tmp/src/sysdep.c emacs-20.7-tmp2/src/sysdep.c --- emacs-20.7-tmp/src/sysdep.c Fri Jun 15 14:14:22 2001 +++ emacs-20.7-tmp2/src/sysdep.c Mon Jun 25 14:23:30 2001 at at -593,7 +593,7 at at #endif #ifdef OS2 - s.main.c_lflag &= ~IDEFAULT; /* Enable termio */ + s.main.c_lflag |= IDEFAULT; /* Disable termio */ #endif /* OS2 */ s.main.c_lflag |= ICANON; /* Enable erase/kill and eof processing */ s.main.c_cc[VEOF] = 04; /* insure that EOF is Control-D */ **= Email 39 ==========================** Date: Thu, 21 Feb 2002 19:55:58 +1100 (EDT) From: Andrew MacIntyre Subject: Re: EMACS build error On Thu, 21 Feb 2002, Andreas Buening wrote: > In general the only file system that supports hard links is /tmp > on Unix systems (to my knowledge). Completely false. -- Andrew I MacIntyre "These thoughts are mine alone..." E-mail: andymac at bullseye.apana.org.au | Snail: PO Box 370 andymac at pcug.org.au | Belconnen ACT 2616 Web: http://www.andymac.org/ | Australia **= Email 40 ==========================** Date: Thu, 21 Feb 2002 20:41:52 +0000 From: John Poltorak Subject: (un)protoize gcc v3.0.3 has a couple of programs, protoize and unprotoize, which are missing from the OS/2 port. I have no idea what they are but would prefer to include them in a GCC package if possible. Does anyone know where I coud get them? -- John **= Email 41 ==========================** Date: Thu, 21 Feb 2002 22:22:24 +0900 From: Masaru Nomiya Subject: Re: EMACS build error Hello, In the Message; Subject : Re: Re: EMACS build error Message-ID : <20020221121518.V23629 at eyup.org> Date & Time: Thu, 21 Feb 2002 12:15:18 +0000 [John] == John Poltorak has written: John> I didn't think that would affect the build... Does it mean my name gets John> compiled in or something? In the Message; Subject : Re: Re: EMACS build error Message-ID : <20020221132317.A1378 at simi.gmd.de> Date & Time: Thu, 21 Feb 2002 13:23:17 +0100 [Holger] == Holger Veit has written: Me> SET NAME=John Poltorak [...] Holger> Who invented this? There are already several defined Holger> environment variables for such a purpose: LOGNAME, USER, so why a Holger> new one? And why a fullname? To make GNUS or movemail happy? This is caused by Mr. Nakgawa's treatment of getpwnam() in os2.c. He said he would fix this..... John> I have now got it built, with much help from you. Great! John> Many thanks! Not at all. John> BTW can I run it in text mode? What do you mean "in text mode"? ..... I made emacs.cmd like this; g:\usr\local\emacs-20.7\src\emacs.exe -bg Black -fg Green2 -cr Blue then make new program icon for emacs.cmd. And set minimization at start on property, relate to nice icon in the emacs-20.7/os2 directory. Another way, at ../emacs-20.7/src make emacspm.exe produce PM version (emacspm.exe) from emacs,exe. Make a program icon for emacspm.exe same as emacs.cmd. Regards, g:\usr\local\emacs-20.7\src\emacs.exe -bg Black -fg Green2 -cr Blue --- Masaru Nomiya mail-to: nomiya at ttmy.ne.jp "No WIndows, no gains!" ..... "Why, I am wrong?" -- Bill -- **= Email 42 ==========================** Date: Thu, 21 Feb 2002 23:12:59 +0100 From: Andreas Buening Subject: Re: make 3.79.1 Mikus Grinbergs wrote: > > On Sat, 16 Feb 2002 21:00:31 +0100 Andreas Buening wrote: > > > > I've just uploaded a new make 3.79.1 release. I treat this as > > beta, so don't expect to much. But I would appreciate if > > some people try whether it works for them. You can get it from > > ftp://ftp.unixos2.com/incoming/make-3_79_1-bin-beta1.zip > > Tried make 3.79.1 with my application. It failed to find > _nl_msg_cat_cntr in my runtime intl.dll (the most recent > versions of intl.dll no longer have that entry point). Even the most recent gettext 0.11 uses that variable. > The documentation says that I should be using gettext 0.10.39 > -- the 0.10.39 package which I downloaded happened __not__ to > include an intl.dll (so presumably the 09 OCT 2001 version of > intl.dll is what is supposed to be used with the new make). This refers to gettext-0_10_39-r2-bin.zip on hobbes. That's the latest version I've uploaded and I still use but in principle you can use any newer (and most likely also older) version if you want. > My point is that there have been TWO distributions since > gettext 0.10.39, both of which provide more recent intl.dll > files. I'm surprised that this new make version depends > upon a back-level gettext package. No, it doesn't. Where exactly did you get your intl.dll? Which version is it? And what error message do you get? > Went back to using the previous make -- it does not care > whether or not I am using a newer intl.dll in my system. bye, Andreas -- One OS to rule them all, One OS to find them, One OS to bring them all and in the darkness bind them In the Land of Redmond where the Shadows lie. **= Email 43 ==========================** Date: Thu, 21 Feb 2002 23:14:01 +0100 From: Andreas Buening Subject: Re: EMACS build error Holger Veit wrote: > > On Thu, Feb 21, 2002 at 12:48:09AM +0100, Andreas Buening wrote: > > > Yes, and no. Partly this was attempted with POSIX/2, but the problem is > > > that it will work for some functions only: you can fake fchown() by > > > returning a success code - no security here, so accept whatever owner > > > the file should adopt. > > > > Yes, and therefore this shouldn't become the default. But if > > the programmer requests it explicitely by using -DDUMMY_FUNCTIONS > > then it's another story. He should know what he is doing. > > The problem is, you need to recompile the code when you want to > migrate to real functions. That's for sure because you would have to use a new or at least an updated libc. > There is still stoneold code out there > with severe bugs due to such quick hacks, which we won't get rid of > (think of the persistency of broken makes and shells). Okay, but this would be no hack (no source code change) because it would be simply a more or less hidden feature of the emx header that can be requested by defining a macro. > > > OTOH, link() or chroot() are functions which > > > do essential things, and which must succeed, otherwise the surrounding > > > code might make the whole program fail. The functions are in the original > > > program for a reason. > > > > If a program requires that link() must succeed then it's broken. > > In general the only file system that supports hard links is /tmp > > on Unix systems (to my knowledge). symlink() might be another story. > > You are wrong. Any Unix file system supports hard links, and has done > so for decades. The thing that makes link() unattractive compared to > symlink(), though, is that symlink can span over several mounted file > systems, whereas a link is (due to its mechanism - is is basically > an alias to an inode) restricted to the filesystem where the original > file is located. The latter is not much a problem, because nowadays > we have large disks, so splitting into /var, /usr, /opt, /home > filesystems is often no longer done (except on larger servers for > good reasons). Okay, to be more precise: At work I cannot do the following in my home directory (it works only in /tmp): touch file1 md subdir ln file1 subdir/file2 Hard links are only possible _within_ the same directory and therefore quite useless, I've been told that this is an AFS feature, so better every serious Unix software should be able to handle a failing link() without a core dump. > > chown(), lchown(), fchown() are currently quite senseless, > > lstat() would be identical to stat(), mkfifo() and mknod() > > are currently not supported, and so on, and last but not least > > there are the all times favorites strcasecmp() and strncasecmp(). > > There could be two different implementations: one that lets these > > functions succeed and one that lets them fail. To make clearer > > what I'm thinking about: > > > > #ifdef EMULATE_POSIX_SUCCESS > > /* let them succeed and hope that it is not tested by the code */ > > # ifndef chown /* for the case config.h already handles chown() */ > > # define chown(x,y,z) (0) > > # endif > > > > # ifndef symlink /* for the case config.h already handles chown() */ > > inline int symlink(x,y,z) { /* copy the file? */ } > > # endif > > Separation between *SUCCESS and *FAILURE is too complicated, IMHO, because > you would sometimes need both: let fchown succeed and symlink fail, for > instance. You'll encounter a clash if both #ifdefs check for > #ifndef fchown. You also need some configuration file > that will disable the replacement code by #define fchown if you have > a real version. *Sometimes* (namely in the autoconf case), you have a > config.h file which usually declares #define HAS_FCHOWN or similar; > so it is a better idea to use these as switcher symbols. If config.h handles a missing lchown() it often uses stuff like #define lchown repl_lchown which would be handled by "#ifndef lchown" in the code above, but in general you're right. I do not say that there could be a all in one solution of this problem. I only say that there are packages that could benefit from such a treatment (maybe less than 50%, I don't know). bye, Andreas -- One OS to rule them all, One OS to find them, One OS to bring them all and in the darkness bind them In the Land of Redmond where the Shadows lie. **= Email 44 ==========================** Date: Thu, 21 Feb 2002 23:14:22 +0000 From: John Poltorak Subject: Building Perl with gcc v3.0.3 I tried building Perl using gcc 3.0.3 but it fails with repeated errors similar to:- make[1]: Entering directory `/tmp/perl1/perl-5.6.1' sh writemain lib/auto/DynaLoader/DynaLoader.lib > writemain.tmp sh mv-if-diff writemain.tmp perlmain.c echo malloc.c av.c scope.c op.c doop.c doio.c dump.c hv.c mg.c perl.c perly.c pp.c pp_hot.c pp_ctl.c pp_sys.c regcomp.c regexec.c utf8.c gv.c sv.c taint.c toke.c util.c deb.c run.c universal.c xsutils.c globals.c perlio.c perlapi.c miniperlmain.c perlmain.c | tr ' ' '\n\r' >.clist make[1]: Leaving directory `/tmp/perl1/perl-5.6.1' Finding dependencies for malloc.obj. cpp0.exe: warning: is shorter than expected malloc.c:505:5: #endif without #if malloc.c:524:2: #endif without #if Finding dependencies for av.obj. cpp0.exe: warning: is shorter than expected In file included from perl.h:1639, from av.c:3: os2ish.h:412:19: operator 'EOL' has no left operand av.c:17:2: #endif without #if Finding dependencies for scope.obj. cpp0.exe: warning: is shorter than expected In file included from perl.h:1639, from scope.c:3: Does this indicate a problem with gcc 3.0.3? What is this 'shorter than expected' referring to? -- John **= Email 45 ==========================** Date: Thu, 21 Feb 2002 23:53:23 +0100 (CET) From: "Sebastian Wittmeier (ShadoW)" Subject: Re: (un)protoize On Thu, 21 Feb 2002 20:41:52 +0000, John Poltorak wrote: "The program protoize is an optional part of GCC. You can use it to add prototypes to a program, thus converting the program to ISO C in one respect. The companion program unprotoize does the reverse: it removes argument types from any prototypes that are found." >gcc v3.0.3 has a couple of programs, protoize and unprotoize, which are >missing from the OS/2 port. I have no idea what they are but would prefer >to include them in a GCC package if possible. Does anyone know where I >coud get them? In the original gnu gcc.tar.gz archive? Sebastian