From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Thu, 20 Feb 2003 04:53:49 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 50 ************************************************** Wednesday 19 February 2003 Number 50 ************************************************** Subjects for today 1 CVS-20030219 : Andrew Belov" 2 Re: ln : Hannes Hromadka 3 Apache2 mod_ssl ? : Hannes Hromadka 4 New RAMFS : John Poltorak 5 Re: Announcement: bash 2.05a : John Poltorak 6 'tee' and exit code : Maynard" 7 Re: Announcement: bash 2.05a : Jeff Robinson 8 Re: Announcement: bash 2.05a : Stefan Neis 9 Re: Announcement: bash 2.05a : John Poltorak 10 Re: Announcement: bash 2.05a : Jeff Robinson 11 Re: 'tee' and exit code : Henry Sobotka 12 Re: 'tee' and exit code : Maynard" 13 Re: TMPDIR : Andreas Buening 14 Re: Make v3.79.1 bug? : Andreas Buening 15 Zope 2.6.1 The Latest and the Greatest! : Ted Sikora **= Email 1 ==========================** Date: Thu, 20 Feb 2003 02:06:15 +0300 (MSK) From: "Andrew Belov" Subject: CVS-20030219 I'm uploading a refreshed build of CVS v 1.12.0.1 to the same location: ftp://ftp.os2ports.com/incoming/cvs11202.zip (or is there a way to overwrite/move files?) This one should have several RSH problems resolved, thanks to Christoph Schulte Moenting for reporting them. Packaging has also been rectified. The infamous "SSH -b" issue historically present in the EMX-new branch has now the following resolution: we use "-b" only if no CVS_RSH was supplied. Then CVS runs "rsh -b", and the IBM's RSH.EXE is meant to be located through PATH. Otherwise we just invoke %CVS_RSH% without any intervening "-b" parameters. **= Email 2 ==========================** Date: Thu, 20 Feb 2003 08:21:13 +0100 From: Hannes Hromadka Subject: Re: ln On Wed, Feb 19, 2003 at 02:23:59PM +0000, John Poltorak wrote: > > The problem here is, that I never succeeded to build slang by > > configure > > make > > I have managed to build a libslang.a by this method, but it wasn't usable > for creating any SLANG based apps. > That's what I had too. (as I mentioned in a mail before in this thread). It should read like this. "I never succeeded to build a _working_ slang by configure make > This is only partially correct. The Makefile is not only used for building > apps, it is often used for installing them. At least those created through > the configure & make process. Yes, but what benefit do I have of installing a broken lib ? > I don't know if it is possible to incorporate the PC specific parts of the > SLANG build into configure. Possibly not, but I do think it would be > useful to be able to run 'Make install' and get everything installed > automatically. IMHO it should be possible, but I used up allready all my time I dedicated to this task. (Should be read: I was not able to do so, and I don't have enough time to dig into the deepness of the slang source) > I'd also like to see a family mode version of mkmake included so that I > wasn't reliant on DOS software... I did manage to build an OS/2 version > once but it was almost ten times as big as the DOS version. Ciao, Hannes -- Johannes Hromadka | Email Office: HromadkaJ at gmx.at | Home : Johannes.Hromadka at gmx.net Vienna/Austria/Europe | OECC: http://www.oecc.org/ >>> Rust never sleeps (borrowed from Neil YOUNG) <<< **= Email 3 ==========================** Date: Thu, 20 Feb 2003 08:27:59 +0100 From: Hannes Hromadka Subject: Apache2 mod_ssl ? Hello: Has anybody tried to build Apache2 with SSL support yet ? I tried and failed. I need the mod_ssl and mod_rexx modules for Apache2. I anyone needs mod_rexx for Apache 1.3.27 SSL, I have one now. (easy to compile) Ciao, Hannes -- Johannes Hromadka | Email Office: HromadkaJ at gmx.at | Home : Johannes.Hromadka at gmx.net Vienna/Austria/Europe | OECC: http://www.oecc.org/ >>> Rust never sleeps (borrowed from Neil YOUNG) <<< **= Email 4 ==========================** Date: Thu, 20 Feb 2003 11:29:48 +0000 From: John Poltorak Subject: New RAMFS There's a new RAMFS (v1.06) at Hobbes. I've found RAMFS indispensible for building apps - it keeps CPU pegged at almost 100% for compiles because there is no time lost for I/0. I just wonder how much it would speed up builds on a system with a fast CPU... -- John **= Email 5 ==========================** Date: Thu, 20 Feb 2003 14:51:51 +0000 From: John Poltorak Subject: Re: Announcement: bash 2.05a On Wed, Feb 19, 2003 at 11:05:51PM -0600, Jeff Robinson wrote: > Hi folks, > > Just leaving a notice that I've uploaded bash-2.05a to UnixOS2.org (I > don't know if it has shown up in /incoming yet or not). It's there now, and I'm having a quick look to see if I can integrate it into my build framework. The first thing I notice is a *15,000* line patch for configure, and I'm wondering if there isn't a better way to build it... With that enhancements made to the GNU Auto tools over the last year we ought to be able to recreate an appropriate configure script by running Autoconf rather than patching it. I'll try building it by running autoconf, configure & make after applying your patches and see how far I get... > Jeff > -- > ---------------- > Whatza JamochaMUD? > http://jamochamud.anecho.mb.ca > > Or other stuff: http://www.anecho.mb.ca/~jeffnik > ----------------------------------------------------------- -- John **= Email 6 ==========================** Date: Thu, 20 Feb 2003 16:32:29 -0600 (CST) From: "Maynard" Subject: 'tee' and exit code Hi, Use of 'tee' forces exit code 0. Witness: # garbage; echo exit code $? sh: garbage: not found exit code 127 # garbage |tee log; echo exit code $? sh: garbage: not found exit code 0 Is there a way to retain the exit code of 'garbage' while 'tee'ing its output? Thanks, -- Maynard **= Email 7 ==========================** Date: Thu, 20 Feb 2003 16:46:32 -0600 From: Jeff Robinson Subject: Re: Announcement: bash 2.05a John Poltorak wrote: > On Wed, Feb 19, 2003 at 11:05:51PM -0600, Jeff Robinson wrote: > >>Hi folks, >> >>Just leaving a notice that I've uploaded bash-2.05a to UnixOS2.org (I >>don't know if it has shown up in /incoming yet or not). > > > It's there now, and I'm having a quick look to see if I can integrate it > into my build framework. > > The first thing I notice is a *15,000* line patch for configure, and I'm > wondering if there isn't a better way to build it... > > With that enhancements made to the GNU Auto tools over the last year we > ought to be able to recreate an appropriate configure script by running > Autoconf rather than patching it. > > I'll try building it by running autoconf, configure & make after applying > your patches and see how far I get... > Yeah, that is a pretty mighty patch, now that I look at it. I think what I should've done (in retrospect) is removed the generated configures from the new source before creating the patch... considering that you should run autoconf after patching it anyways. I wanted to have this done in a much cleaner way, but I also wanted to get something done relatively quickly so that Javier could test building Mozilla with this version of bash. I haven't heard back from him lately so either it works alright or bash set his workstation on fire. What I would really like to do (like you suggested) is have most things automatically defined by running autoconf... but I couldn't figure out a clean way off adding additional flags... if I do it in the config.os2 (config_site) file then it over-rides *ALL* the autoconf'd flags. Is there a way to just append flags to what autoconf comes up with? Jeff -- ---------------- Whatza JamochaMUD? http://jamochamud.anecho.mb.ca Or other stuff: http://www.anecho.mb.ca/~jeffnik ----------------------------------------------------------- **= Email 8 ==========================** Date: Thu, 20 Feb 2003 17:00:12 +0100 (CET) From: Stefan Neis Subject: Re: Announcement: bash 2.05a On Thu, 20 Feb 2003, John Poltorak wrote: > The first thing I notice is a *15,000* line patch for configure, and I'm > wondering if there isn't a better way to build it... With that size, it has got to have been regenerated by autoconf. Still, from the user perspective, I prefer applying a huge patch to the existing configure sciprt over installing and running autoconf to get a working configure... ;-) And diff, of course, won't care how the file was generated, it will just tell you the differences ... Regards, Stefan **= Email 9 ==========================** Date: Thu, 20 Feb 2003 17:16:10 +0000 From: John Poltorak Subject: Re: Announcement: bash 2.05a On Thu, Feb 20, 2003 at 05:00:12PM +0100, Stefan Neis wrote: > On Thu, 20 Feb 2003, John Poltorak wrote: > > > The first thing I notice is a *15,000* line patch for configure, and I'm > > wondering if there isn't a better way to build it... > > With that size, it has got to have been regenerated by autoconf. I have tried, and I don't think it is too far from working. In fact I did manage to create four binaries, including something called bashversion.exe which was 60kB! Why you need a 60kB program to simply tell you which version it is, I'll never know.. One of the warnings, which occurs about 70 times is like this:- gcc -c -DHAVE_CONFIG_H -DSHELL -I. -I.. -I.. -I../include -I../lib -I. -g -O2 trap.c || ( rm -f trap.c ; exit 1 ) In file included from ../shell.h:39, from ./trap.def:56: ../sig.h:33: warning: `RETSIGTYPE' redefined ..\config.h:419: warning: this is the location of the previous definition Any idea what is causing that? There is also this error:- In file included from ./ulimit.def:76: pipesize.h:9: parse error before `/' make[1]: *** [ulimit.o] Error 1 make[1]: Leaving directory `U:/unixos2/workdir/bash-2.05a/builtins' make: *** [builtins/libbuiltins.a] Error 1 Here it looks as though '\b' has been interpreted as some sort of escape sequence. Instead of 'workdir\bash' the file has 'workdir*ash' where * is a backspace char. I don't know how to oversome errors like that. > Still, from the user perspective, I prefer applying a huge patch to the > existing configure sciprt over installing and running autoconf to get a > working configure... ;-) Sure, if it works, it works. But in the longer term it would be nice to get OS/2 patches rolled into the main version, so that we don't have to port the next release, and the one after that, and it just isn't going to happen if 15,000 line patches for a configure program are going to be submitted to GNU for approval. > Regards, > Stefan -- John **= Email 10 ==========================** Date: Thu, 20 Feb 2003 17:18:38 -0600 From: Jeff Robinson Subject: Re: Announcement: bash 2.05a John Poltorak wrote: > > > I have tried, and I don't think it is too far from working. In fact I did > manage to create four binaries, including something called bashversion.exe > which was 60kB! Why you need a 60kB program to simply tell you which > version it is, I'll never know.. > It must be full of bash-magic! Okay, honestly I don't know what else it does. > > One of the warnings, which occurs about 70 times is like this:- > > gcc -c -DHAVE_CONFIG_H -DSHELL -I. -I.. -I.. -I../include -I../lib -I. > -g -O2 trap.c || ( rm -f trap.c ; exit 1 ) > In file included from ../shell.h:39, > from ./trap.def:56: > ../sig.h:33: warning: `RETSIGTYPE' redefined > ..\config.h:419: warning: this is the location of the previous definition > At this point I was just letting sleeping warnings lie, though I'll take a closer look at this when I go after bash-2.05b. Admittedly, most of these patches were taken straight from Jun's work so I'm not 100% familiar with some of the choices he made. > > There is also this error:- > > In file included from ./ulimit.def:76: > pipesize.h:9: parse error before `/' > make[1]: *** [ulimit.o] Error 1 > make[1]: Leaving directory `U:/unixos2/workdir/bash-2.05a/builtins' > make: *** [builtins/libbuiltins.a] Error 1 > > > Here it looks as though '\b' has been interpreted as some sort of escape > sequence. Instead of 'workdir\bash' the file has 'workdir*ash' where * is > a backspace char. I don't know how to oversome errors like that. > My pipesize.h looks like this: /* * pipesize.h * * This file is automatically generated by psize.sh * Do not edit! */ #define PIPESIZE ./psize.sh: ./psize.aux: cannot execute binary file Heh... obviously something went wrong on my machine that worked... but went right on your machine and failed. Or something like that. Huh. Jeff -- ---------------- Whatza JamochaMUD? http://jamochamud.anecho.mb.ca Or other stuff: http://www.anecho.mb.ca/~jeffnik ----------------------------------------------------------- **= Email 11 ==========================** Date: Thu, 20 Feb 2003 18:01:31 -0500 From: Henry Sobotka Subject: Re: 'tee' and exit code Maynard wrote: > > Is there a way to retain the exit code of 'garbage' while 'tee'ing its > output? Just move it up before the pipe: [borges|c:/]garbage; echo exit code $? | tee log sh: garbage: command not found exit code 127 [borges|c:/]cat log exit code 127 h~ **= Email 12 ==========================** Date: Thu, 20 Feb 2003 18:06:07 -0600 (CST) From: "Maynard" Subject: Re: 'tee' and exit code Henry, >Just move it up before the pipe: Thanks; that helped a lot in redefining the problem for me. I've more pondering to do ;-} -- Maynard **= Email 13 ==========================** Date: Thu, 20 Feb 2003 21:35:43 +0100 From: Andreas Buening Subject: Re: TMPDIR Holger Veit wrote: > > On Wed, Feb 19, 2003 at 11:09:57AM +0000, John Poltorak wrote: > > How common is it to have $TMPDIR defined on Unix system? > > > > Should we have it on UnixOS/2? [history of temp files and anachronisms of certain Unixes] > Actually, the presence > of any of these environment variables is an indication > that the software relying on it, is broken, poor designed > or otherwise lousy. > > So, to summarize, whatever other Unixes use - lousy software > is a fact, so be prepared that you might need any of these > variables, as well as any of the api functions (whether or not > they are POSIX). I don't know what other Unixes use but I know that a few GNU tools use $TMPDIR especially for DOS-like systems. So I'd say the answer to John's question is "yes". Nevertheless, it's not an anachronism that you can define a tmp dir by an env. var., IMHO. /tmp is writable to all users, i.e. everybody can pollute this dir. This might be no problem on a single user linux out of the box installation on a new nnn GB harddisk with a n GB /tmp partition where the only user has also root access if he needs it. On a real multiuser system the admins tend to go into weekend just 5 minutes before somebody decides to screw up /tmp. ;-) Apart from that there are several OSes that might have no /tmp. And I can't see any disadvantage if I'm able to specify the temp dir manually. 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 Mordor where the Shadows lie. **= Email 14 ==========================** Date: Thu, 20 Feb 2003 21:35:58 +0100 From: Andreas Buening Subject: Re: Make v3.79.1 bug? John Poltorak wrote: > > Can anyone with Make v3.79.1 try building InfoZip's UNZIP? > > The source is available here:- > > ftp://ftp.info-zip.org/pub/infozip/src/unzip550.tar.gz > > The build fails when using using this verion of Make, but works with > earlier versions. I've just uploaded a patched version: sources: http://unix.os2site.com/sw/pub/source/make/make-3_79_2a1-r2.zip binaries: http://unix.os2site.com/sw/pub/binary/make/make-3_79_2a1-r2-bin.zip 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 Mordor where the Shadows lie. **= Email 15 ==========================** Date: Thu, 20 Feb 2003 22:39:53 -0500 From: Ted Sikora Subject: Zope 2.6.1 The Latest and the Greatest! New Zope 2.6.1 bin complete with Python 2.2.2 with many refinements in both Python and Zope. Includes LocalFS and Jeff Robinson's LFSBrowser. Mailman ready ...too! Python now includes all relevant src files for any/all Python development work. WarpIn package will follow tomorrow. ftp://os2ports.com/pub/os2/unix/internet/Zope/Zope261-os2emx-030220.zip Viva Zopista! -- Ted Sikora tsikora at ntplx.net