--- Begin Message ---
- To: 360852-done@bugs.debian.org
- Subject: Re: Bug#360852: Unpurgeable: problem with postrm script
- From: David Nusinow <dnusinow@speakeasy.net>
- Date: Thu, 5 Oct 2006 22:15:49 -0400
- Message-id: <20061006021549.GG4766@verizon.net>
- In-reply-to: <20060404234112.6049.88866.reportbug@localhost>
- References: <20060404234112.6049.88866.reportbug@localhost>
On Tue, Apr 04, 2006 at 06:41:12PM -0500, Sukant Hajra wrote:
> Package: xserver-xorg
> Version: 6.9.0.dfsg.1-4
> Severity: normal
>
>
>
> Hi,
>
> I was suprised that this package wasn't purging it's configuration and
> was instead exiting prematurely with a -128 code. I took some time to
> tinker around with a local copy of the postrm from /var/lib/dpkg/info,
> and was able to get through the problem. Before going any further,
> here's the one change that got me through the purge (very small).
>
> $ diff /var/lib/dpkg/info/xserver-xorg.postrm xserver-xorg.postrm.copy
> 969c969
> < update-rc.d xserver-xorg remove
> ---
> > update-rc.d xserver-xorg remove > /dev/null
>
> I wouldn't be surprised if this is not a "fix", but really just a hack
> to get through a problem that deserves more attention. I came upon this
> solution by tracing through /usr/share/debconf/confmodule. The error
> occurs inside of _db_cmd() when it's called by db_purge(). I installed
> an "echo" to see the contents of $_db_internal_line in confmodule, and
> noticed that the stdout of the update_rc.d call (patched above) was
> somehow being interpreted as a command; there was an error that said
> something like, "removing: command not found," which was a byproduct of
> the following line from /usr/sbin/update-rc.d.
>
> print " Removing any system startup links for $initd/$bn ...\n"
>
> I don't know internal debian package scripts that well at all, so again,
> I'm not at all certain that this is the proper solution. However, I'm
> sure that you guys will be able to fix the postrm script in short order.
>
> Feel free to E-mail me if you would like more information. Having fixed
> the problem, I'm not exactly in the mood to reproduce it, so I'm banking
> on the fact that this problem is reproduceable. In that case, I'm
> surprised it's hasn't be reported sooner, because this problem doesn't
> seem like it would vary across installations. Oh well, maybe I just was
> the first to catch it.
This was fixed in the current version of the package in unstable. Closing.
Thanks for your report.
- David Nusinow
--- End Message ---