On 2006-10-02 18:22 +0200, Frank Küster wrote: > Hi, look at this. Maybe it's just cosmetics, but it might also point to > an underlying important problem: > > Preparing to replace tex-common 0.29~bpo.1 (using > .../tex-common_0.30~bpo.1_all.deb) ... > Unpacking replacement tex-common ... > Preparing to replace tetex-base 3.0-21~bpo.1 (using > .../tetex-base_3.0-23~bpo.1_all.deb) ... > Unpacking replacement tetex-base ... > Preparing to replace tetex-doc 3.0-21~bpo.1 (using > .../tetex-doc_3.0-23~bpo.1_all.deb) ... > Unpacking replacement tetex-doc ... > Preparing to replace tetex-extra 3.0-21~bpo.1 (using > .../tetex-extra_3.0-23~bpo.1_all.deb) ... > Unpacking replacement tetex-extra ... > Setting up libpoppler-glib-dev (0.4.5-3bpo1) ... > Setting up tex-common (0.30~bpo.2) ... > > Setting up tetex-base (3.0-23~bpo.1) ... > > Note that in the first line, it says it is using version 0.30~bpo.1, > whereas when in the configure phase, it says 0.30~bpo.2.
One explanation is that the "Preparing to replace..." message displays (part of) a file name in the brackets whereas the "Setting up..." message displays a package version. The file name does not need to match the package version. > And the second one must have been correct, because bpo.1 had an error in > its postinst, which would not run. > > I uploaded both 30~bpo.* versions today, and it might have to do with > apt-proxy running between me and the actual mirrors. But still I think > this should never happen. Hm, is it possible that you mixed up the file names when copying the packages and that your ~bpo.2 package inadvertently became a ~bpo.1_all.deb? That would be an explanation, and really the only one I can think of. Can these packages be found anywhere today? In the unlikely case that you have the aptitude and/or dpkg logs still available, they could shed more light on the issue, but otherwise it might be best to just close the bug, since there's probably nothing that can be done about it (apart from improving dpkg's messages, which has already been suggested -- see #71106, for instance). Kind regards, Sven -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
