Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
> Date: Fri, 25 Feb 2000 13:11:33 -0800 > From: Paul Vojta > To: [EMAIL PROTECTED], [EMAIL PROTECTED] > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad code] > Cc: debian-sparc@lists.debian.org, [EMAIL PROTECTED] > > > Date: Fri, 25 Feb 2000 13:58:40 -0500 > > From: Ben Collins <[EMAIL PROTECTED]> > > To: Wichert Akkerman <[EMAIL PROTECTED]> > > Cc: debian-sparc@lists.debian.org, [EMAIL PROTECTED] > > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad > > code] > > > > Can you give us your kernel version and paste some of the error messages > > from the failures you are seeing please? > > Hm... I booted up an old 2.3.1 kernel (grabbed from cvs last May) > and the problem does not occur. I'll try grabbing a current kernel. Well, I grabbed and compiled 2.2.15-pre10 from cvs on openserver, and it does not exhibit the bug either. So it must have been a bad kernel. --Paul Vojta, [EMAIL PROTECTED]
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
> Date: Fri, 25 Feb 2000 16:17:30 -0500 > From: Ben Collins <[EMAIL PROTECTED]> > To: Paul Vojta <[EMAIL PROTECTED]> > Cc: [EMAIL PROTECTED], debian-sparc@lists.debian.org > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad code] > > > _tashkent# cd /var/cache/apt/archives > > _tashkent# dpkg --install findutils_4.1-37_sparc.deb > > (Reading database ... 16703 files and directories currently installed.) > > Preparing to replace findutils 4.1-37 (using findutils_4.1-37_sparc.deb) ... > > dpkg: error processing findutils_4.1-37_sparc.deb (--install): > > subprocess pre-installation script killed by signal (Segmentation fault) > > Errors were encountered while processing: > > findutils_4.1-37_sparc.deb > > This is the subprocess dying (the preinst script), not dpkg. What happens > if the run the preinst script manually? Have you done an strace of the > process aswell? That's what I initially thought, but some of the scripts that were 'dying' were pretty simple. Also, I haven't had the problem with dpkg-1.6.6. So I think the problem is occurring after the fork() but before (or during) execv(). This may all be moot, though, since it doesn't occur on the older kernel that I'm running right now. --Paul Vojta, [EMAIL PROTECTED]
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
On Fri, Feb 25, 2000 at 12:59:31PM -0800, Paul Vojta wrote: > > Date: Fri, 25 Feb 2000 13:58:40 -0500 > > From: Ben Collins <[EMAIL PROTECTED]> > > To: Wichert Akkerman <[EMAIL PROTECTED]> > > Cc: debian-sparc@lists.debian.org, [EMAIL PROTECTED] > > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad > > code] > > > > Can you give us your kernel version and paste some of the error messages > > from the failures you are seeing please? > > - > > _tashkent# cd /var/cache/apt/archives > _tashkent# dpkg --install findutils_4.1-37_sparc.deb > (Reading database ... 16703 files and directories currently installed.) > Preparing to replace findutils 4.1-37 (using findutils_4.1-37_sparc.deb) ... > dpkg: error processing findutils_4.1-37_sparc.deb (--install): > subprocess pre-installation script killed by signal (Segmentation fault) > Errors were encountered while processing: > findutils_4.1-37_sparc.deb This is the subprocess dying (the preinst script), not dpkg. What happens if the run the preinst script manually? Have you done an strace of the process aswell? -- ---===-=-==-=---==-=-- / Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \ ` [EMAIL PROTECTED] -- [EMAIL PROTECTED] -- [EMAIL PROTECTED] ' `---=--===-=-=-=-===-==---=--=---'
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
> Date: Fri, 25 Feb 2000 13:58:40 -0500 > From: Ben Collins <[EMAIL PROTECTED]> > To: Wichert Akkerman <[EMAIL PROTECTED]> > Cc: debian-sparc@lists.debian.org, [EMAIL PROTECTED] > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad code] > > Can you give us your kernel version and paste some of the error messages > from the failures you are seeing please? Hm... I booted up an old 2.3.1 kernel (grabbed from cvs last May) and the problem does not occur. I'll try grabbing a current kernel. --Paul Vojta, [EMAIL PROTECTED]
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
> Date: Fri, 25 Feb 2000 13:58:40 -0500 > From: Ben Collins <[EMAIL PROTECTED]> > To: Wichert Akkerman <[EMAIL PROTECTED]> > Cc: debian-sparc@lists.debian.org, [EMAIL PROTECTED] > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad code] > > Can you give us your kernel version and paste some of the error messages > from the failures you are seeing please? - _tashkent# cd /var/cache/apt/archives _tashkent# dpkg --install findutils_4.1-37_sparc.deb (Reading database ... 16703 files and directories currently installed.) Preparing to replace findutils 4.1-37 (using findutils_4.1-37_sparc.deb) ... dpkg: error processing findutils_4.1-37_sparc.deb (--install): subprocess pre-installation script killed by signal (Segmentation fault) Errors were encountered while processing: findutils_4.1-37_sparc.deb - I am running 2.2.13 from the cvs server, grabbed on 15 Nov. --Paul Vojta, [EMAIL PROTECTED]
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
> Date: Fri, 25 Feb 2000 09:47:50 -0500 > From: Ben Collins <[EMAIL PROTECTED]> > To: Wichert Akkerman <[EMAIL PROTECTED]> > Cc: debian-sparc@lists.debian.org > Subject: Re: [EMAIL PROTECTED]: [sparc32] compilation of dpkg gives bad code] > > On Fri, Feb 25, 2000 at 10:32:45AM +0100, Wichert Akkerman wrote: > > > > Can someone comment on this? I haven't heard of it before.. > > I've compiled 1.6.9 with gcc 2.95.2-4, with no problems. I don't think > there is anything in 2.95.2-6 that would cause this. I'll try a test > compile with 2.95.2-6 though and make sure. The bug also occurs when I compile dpkg help.c with gcc-2.95.2-4. That's the oldest gcc .deb I have still lying around. --Paul Vojta, [EMAIL PROTECTED]
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
On Fri, Feb 25, 2000 at 09:47:50AM -0500, Ben Collins wrote: > On Fri, Feb 25, 2000 at 10:32:45AM +0100, Wichert Akkerman wrote: > > > > Can someone comment on this? I haven't heard of it before.. > > I've compiled 1.6.9 with gcc 2.95.2-4, with no problems. I don't think > there is anything in 2.95.2-6 that would cause this. I'll try a test > compile with 2.95.2-6 though and make sure. Ok, I did a compile of dpkg 1.6.9 using gcc 2.95.2-6. I Then ran heavy upgrades on a sun4u, sun4m and sun4c. None of them exhibited any problems. Can you give us your kernel version and paste some of the error messages from the failures you are seeing please? -- ---===-=-==-=---==-=-- / Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \ ` [EMAIL PROTECTED] -- [EMAIL PROTECTED] -- [EMAIL PROTECTED] ' `---=--===-=-=-=-===-==---=--=---'
Re: [vojta@math.berkeley.edu: [sparc32] compilation of dpkg gives bad code]
On Fri, Feb 25, 2000 at 10:32:45AM +0100, Wichert Akkerman wrote: > > Can someone comment on this? I haven't heard of it before.. I've compiled 1.6.9 with gcc 2.95.2-4, with no problems. I don't think there is anything in 2.95.2-6 that would cause this. I'll try a test compile with 2.95.2-6 though and make sure. -- ---===-=-==-=---==-=-- / Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \ ` [EMAIL PROTECTED] -- [EMAIL PROTECTED] -- [EMAIL PROTECTED] ' `---=--===-=-=-=-===-==---=--=---'
