[1921] in athena10
dpkg and conffiles bug
daemon@ATHENA.MIT.EDU (Tim Abbott)
Sat Mar 28 13:30:39 2009
Date: Sat, 28 Mar 2009 13:29:58 -0400 (EDT)
From: Tim Abbott <tabbott@MIT.EDU>
To: debathena@mit.edu
cc: Alex T Prengel <alexp@mit.edu>
Message-ID: <alpine.DEB.1.10.0903261622110.7289@vinegar-pot.mit.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
So, I attempted to reproduce the issue that alexp had on a cluster machine
with the double-diversion of /etc/gdm/gdm.conf-custom under the assumption
that it was the result of the dpkg conffile bug
(http://bugs.debian.org/476899). While I can get the system into an
inappropriate state such as this (aka #476899):
test:~/dpkg-test# ls /etc/inetd.conf* -l
lrwxrwxrwx 1 root root 20 2009-03-28 12:50 /etc/inetd.conf -> inetd.conf.debathena
lrwxrwxrwx 1 root root 30 2009-03-28 12:50 /etc/inetd.conf.debathena -> inetd.conf.debathena.debathena
-rw-r--r-- 1 root root 96 2009-03-28 12:18 /etc/inetd.conf.debathena.debathena-orig
-rw-r--r-- 1 root root 102 2009-03-28 12:51 /etc/inetd.conf.debathena.debathena-orig.dpkg-new
-rw-r--r-- 1 root root 1110 2009-03-28 12:13 /etc/inetd.conf.debathena-orig
However, I've been unable to get dpkg to prompt me about upgrading a
conffile.
Have we seen this issue on any machine other than alexp's? At this point
my best guess is that his issue is actually a result of the
debathena-gdm-config-athena10 -> debathena-gdm-config transition; did
alexp's cluster test machine start out pointed at athena10.mit.edu?
-Tim Abbott
