FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian Development

 
 
LinkBack Thread Tools
 
Old 08-22-2011, 10:02 AM
Cyril Brulebois
 
Default Buildd vs. Cowbuilder: broken symlink in procps dev package on one architecture

Michael Prokop <mika@debian.org> (22/08/2011):
> 1) What's the proper way to address this issue in squeeze?
> a) Build a i386 package with broken symlink?
> b) Build whatever-arch package with working symlink?

Stop using basename on what's in /lib, and do that on what's under
debian/tmp.

> 2) How can we make sure such a bug doesn't happen again
> (besides working towards source-only uploads )?
> Bugreport against buildd.debian.org?

Same answer as for 1).

Mraw,
KiBi.
 
Old 08-22-2011, 10:17 AM
Cyril Brulebois
 
Default Buildd vs. Cowbuilder: broken symlink in procps dev package on one architecture

Cyril Brulebois <kibi@debian.org> (22/08/2011):
> Michael Prokop <mika@debian.org> (22/08/2011):
> > 1) What's the proper way to address this issue in squeeze?
> > a) Build a i386 package with broken symlink?
> > b) Build whatever-arch package with working symlink?
>
> Stop using basename on what's in /lib, and do that on what's under
> debian/tmp.

More to the point now that I played a bit with it:

a) One could try moving PROCLIB's definition to the
override_dh_auto_build target, but then one can see there's no
libproc-*.so at this point. (This could be a variable expansion
issue, after all…)

b) One could try replacing it with a `basename proc/libproc-*.so`
in the (cd … && ln …) call but there's still no libproc-*.so
at this point.

c) One could check it appears when override_dh_auto_build is called,
and add the ln -s call after the .so is available under proc/, and
profit. That should fix the issue.

d) To confirm the difference between procps installed and missing,
add an “exit 42” at the end of the build target, and notice the
presence of the libproc-….so file when procps is installed.
Fixing the difference in behaviour here might be a more elegant
solution, but I'll leave that to the maintainer to decide.

> > 2) How can we make sure such a bug doesn't happen again
> > (besides working towards source-only uploads )?
> > Bugreport against buildd.debian.org?
>
> Same answer as for 1).

That part still holds.

Mraw,
KiBi.
 
Old 08-22-2011, 05:14 PM
Julien Cristau
 
Default Buildd vs. Cowbuilder: broken symlink in procps dev package on one architecture

On Mon, Aug 22, 2011 at 12:19:13 +0200, Michael Prokop wrote:

> But that the procps package is non-essential but seems to be installed
> on (some?) buildds is irrelevant?
>
Yes.

Cheers,
Julien


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110822171443.GL2820@radis.liafa.jussieu.fr">http ://lists.debian.org/20110822171443.GL2820@radis.liafa.jussieu.fr
 
Old 08-22-2011, 06:11 PM
"Adam D. Barratt"
 
Default Buildd vs. Cowbuilder: broken symlink in procps dev package on one architecture

On Mon, 2011-08-22 at 12:19 +0200, Michael Prokop wrote:
> * Cyril Brulebois [Mon Aug 22, 2011 at 12:02:17PM +0200]:
> > Michael Prokop <mika@debian.org> (22/08/2011):
>
> > > 1) What's the proper way to address this issue in squeeze?
> > > a) Build a i386 package with broken symlink?
> > > b) Build whatever-arch package with working symlink?
>
> > Stop using basename on what's in /lib, and do that on what's under
> > debian/tmp.
>
> So 1b, thanks.

Looking at the procps source package in unstable, it appears to have the
same issue. If that's the case, then I'm afraid your question should
really have been "what's the proper way to address this issue in
unstable and subsequently in squeeze".

If I've missed a reason why unstable isn't affected, then great.

Regards,

Adam


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 1314036669.25185.5.camel@hathi.jungle.funky-badger.org">http://lists.debian.org/1314036669.25185.5.camel@hathi.jungle.funky-badger.org
 
Old 08-22-2011, 09:51 PM
Cyril Brulebois
 
Default Buildd vs. Cowbuilder: broken symlink in procps dev package on one architecture

Michael Prokop <mika@debian.org> (22/08/2011):
> * Adam D. Barratt [Mon Aug 22, 2011 at 07:11:08PM +0100]:
> > Looking at the procps source package in unstable, it appears to have the
> > same issue. If that's the case, then I'm afraid your question should
> > really have been "what's the proper way to address this issue in
> > unstable and subsequently in squeeze".
>
> Sure, but fixing the package in unstable doesn't modify the
> behaviour within the stable release, which was the rationale behind
> my question towards the issue in squeeze.

The way to fix packages in stable is to fix them in unstable, wait for
the fix to be tested well enough (reaching testing is usually a good
indicator), polished if needed; then backported to stable. Hence Adam's
reply.

Mraw,
KiBi.
 
Old 08-22-2011, 09:56 PM
Christian Hofstaedtler
 
Default Buildd vs. Cowbuilder: broken symlink in procps dev package on one architecture

* Cyril Brulebois <kibi@debian.org> [110822 23:51]:
> Michael Prokop <mika@debian.org> (22/08/2011):
> > * Adam D. Barratt [Mon Aug 22, 2011 at 07:11:08PM +0100]:
> > > Looking at the procps source package in unstable, it appears to have the
> > > same issue. If that's the case, then I'm afraid your question should
> > > really have been "what's the proper way to address this issue in
> > > unstable and subsequently in squeeze".
> >
> > Sure, but fixing the package in unstable doesn't modify the
> > behaviour within the stable release, which was the rationale behind
> > my question towards the issue in squeeze.
>
> The way to fix packages in stable is to fix them in unstable, wait for
> the fix to be tested well enough (reaching testing is usually a good
> indicator), polished if needed; then backported to stable. Hence Adam's
> reply.

I don't see us going to fix this in stable; I'll see if I can come
up with a reliable patch for unstable, but that's it.

--
christian hofstaedtler
 

Thread Tools




All times are GMT. The time now is 07:39 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org