Bug#604128: libreoffice: FTBFS (s390): dpkg-shlibdeps: error: objdump died from signal 6

[Niko Tyni]

On Fri, Dec 31, 2010 at 10:13:34PM +0100, Rene Engelhard wrote:
> On Mon, Nov 22, 2010 at 12:26:47PM +0200, Niko Tyni wrote:
> > On Sat, Nov 20, 2010 at 06:12:42PM +0100, Rene Engelhard wrote:
> > 
> > > On Sat, Nov 20, 2010 at 03:27:44PM +0100, Philipp Kern wrote:
> > > > > dpkg-shlibdeps: warning: Can't extract name and version from library name `libvclplug_genl3.so'
> > > > > dpkg-shlibdeps: warning: Can't extract name and version from library name `libvcll3.so'
> > > > > *** glibc detected *** /usr/bin/perl: corrupted double-linked list: 0x01247c90 ***
> > 
> > > > > dpkg-shlibdeps: error: objdump died from signal 6
> > 
> > > So dpkg-shlibdeps fails (well, as far as I read it perl). So why
> > > is this filed against libreoffice? :)
> > 
> > I've requested the build dependencies of libreoffice to be installed on
> 
> Which probably didn't happen, I got loads of missing ones when I tried
> a few days ago.

Yes, many of them were only present in experimental so they couldn't be
installed in the sid chroot. The DSA folks never responded to my followup
about a dedicated chroot for this.

I then went for the emulator route and hit #605759, so I only recently
got an (unfortunately slow) s390 system up and running.

> > BTW, just use perl at packages.debian.org next time. pkg-perl-maintainers
> > isn't the maintainer for perl itself.
> 
> OK. Shouldn't it be?

The pkg-perl project was never about maintaining the Perl interpreter.
I do welcome maintenance help of course, and there's a separate Alioth
project and mailing list set up for that.

On Fri, Dec 31, 2010 at 11:11:54PM +0100, Rene Engelhard wrote:
> On Fri, Dec 31, 2010 at 10:13:34PM +0100, Rene Engelhard wrote:

> > I have a build running there right now, let's see whether that still fails.
> 
> That build survived dpkg-shlibdeps and the build went correctly to the end
> 
> > (and the s390 buildd also is building it right now.)
> 
> also successful.

While that's generally good news, I suppose it means the perl bug is
unreproducible. I don't see much to work on until it manifests again.
-- 
Niko Tyni   ntyni at debian.org