Hi, Thanks for following up.
David Baron wrote: > [Subject: Re: Bug#629983: [libc6] I also had this on possible upgrade to > 2.13-6] Please keep in mind that these show up as emails, so the subject line is a good place to put a message in context. > Problem is that, whatever be the offending file, the only alternatives are to > go ahead anyway or abort ALL upgrades. This bug report is about a situation in which libc6.preinst cancels the upgrade because there are files around that would interfere with it. (Some of the possible situations in which it can trigger have been tracked down and will be corrected automatically in the next upload.) In particular, it was about the lack of clarity in the error message, such that it seemed to imply one fix, renaming to add a .old suffix, would work when it didn't. As far as I can tell, you are talking about something else. The usual dpkg behavior (as described in the Debian policy manual) after "preinst upgrade" fails is for the old "postinst abort-upgrade" will be called and the upgrade of this package to be cancelled. Then the higher-level package manager takes over and has a chance to finish unwinding the upgrade and put things in a sane state. Do I understand correctly that you are saying that last step didn't proceed to your satisfaction? (For example, maybe it downgraded too many packages, or maybe it didn't downgrade enough.) > On my system, six packages are pulled > with lib6 upgrades. The mix of new and old will crash out X, for instance. I've never heard of this behavior. Please file a new bug. > A version of this bug was listed as grave and then withdrawn. Probably you are referring to Bug#629534 of 2.13-5, which was fixed (not withdrawn :)) in 2.13-6 by adding the preinst check. Hope that helps, Jonathan -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/20110612203156.GA21473@elie
