[Bash-completion-devel] Bug#486702: bash-completion: Fails to upgrade

Niko Tyni ntyni at debian.org
Fri Jun 20 12:10:25 UTC 2008

On Fri, Jun 20, 2008 at 12:00:22AM +0200, David Paleino wrote:
> On Thu, 19 Jun 2008 22:21:35 +0200, Helge Kreutzmann wrote:

> I just used "diff" instead of "sdiff" and it worked (as it did
> previously now in the x86 chroot). 

> So, we now have two options:
> 1) it's an ucf bug
> 2) it's a sdiff bug.

> > For completeness, the following simple steps reproduce it:
> > Setup up a etch chroot, e.g. by
> > cdebootstrap --arch=i386 etch /var/chroot/sid http://...
> > 
> > Copy the necessary files inside (fstab, passwd, resolve.conf, ...)
> > Mount the necessary directories (using --bind), typically proc
> > Edit sources.list to point to sid
> > chroot in your new chroot
> > apt-get update
> > apt-get dist-upgrade
> > apt-get dist-upgrade
> > (yes, twice). In the second run, bash_completion will be
> > installed/upgraded, and the dialogue appears. 

> I'm still of the idea that this is not our fault though, thus I'm reassigning
> to ucf. Manoj, please feel free to reassign it elsewhere, if you feel it's not
> your fault either (diff?)

FWIW, I think this is a deadlock in debconf with two processes both
blocking to read from the other. It's already reported as #486094 and
#460916 against debconf (and #478659 against bash-completion). There's
some analysis in #460916 from March, and quoting myself there:

> I assume the deadlock occurs because the pipe buffers fill up. Suggest
> cloning to debconf and cutting overly long diff output in ucf. It's not
> like anyone is going to make an informed decision by reading a 100k diff
> presented by debconf...
Niko Tyni   ntyni at debian.org

More information about the Bash-completion-devel mailing list