On Sat, Feb 18, 2006 at 11:29:41AM -0800, Paul Traina wrote: > Justin Pryzby wrote: > >Do you know if there is something wrong with the control file? What > >it is even? :) > > > >Did the .debs just get reuploaded, without rebuilding?! > > > >Justin > > It's bizzare, the control file uses the same version macro everywhere, > but for the gtk reference, it used the old version. Perhaps there was a > problem in the build depends and I don't understand how the expansion of > $Source_Version is supposed to work. The only thing besides ${Source-Version} in the source control file is:
Package: python-gtk2-dev Replaces: ... python2.3-gtk2 (<< 1.99.17-4) ... Conflicts: ... python2.3-gtk2 (<< 1.99.17-4) ... ${Source-Version} is substituted by debhelper with the most recent version number from the changelog (the upstream portion of which should also be in the directory name). and Conflicts+Replaces means to remove sufficiently old versions of python2.3-gtk in favor of python-gtk2-dev. > In any case, this is a MAJOR problem, as it causes a cascade of failures > leading to the gnome metapackages getting uninstalled with an additional > cascade of uninstalls if that happens. I guess I don't understand this .. I only ran into this problem by chance. If it is a serious problem, then you might copy to: Debian GNOME Maintainers <[EMAIL PROTECTED]> > I know folks don't want to upgrade the severity of the problem, but I > hope to hell that the gnome team is working on fixing this IMMEDIATELY. I only failed to upgrade severity further because RC bugs *prevent* testing migration, which is the whole problem, I guess. pqdo has: * [2006-02-04] pygtk 2.6.3-2 MIGRATED to testing (Britney) * [2006-02-02] pygtk 2.8.2-3 MIGRATED to testing (Britney) which makes no sense to me at all; Makholm's scripts write those messages, not Britney itself, so it could be wrong. But now I've gone and completely confused myself and I'm not even sure if there's a bug, besides that testing is somewhat broken, which isn't a real bug anyway. In fact I just checked and it seems this package *is* installable in unstable, so this may be nothing but a transient testing bug; doh. Do you have reason to think otherwise? Justin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]