reassign 425799 lib32ncurses5 close 425799 5.6-3 thanks On Thu, May 24, 2007 at 10:23:15AM -0700, Joshua Kwan wrote: > reopen 425799 > reassign 425799 libc6-i386 > thanks homeboy > On Thu, May 24, 2007 at 06:51:10AM +0000, Daniel Baumann wrote: > > This was fixed as of ncurses 5.6-3. > > I understand, but this happened in an upgrade situation wherein: > > libc6-i386 was scheduled to be upgraded to 2.5-9 > lib32ncurses5 was scheduled to be upgraded to 5.6-3 > > Because of how apt makes decisions I assume it chose to upgrade libc6-i386 > first, which is perfectly reasonable. We can't assume people will always > get lib32ncurses5 in first. > > There are two ways to fix this. > 1) Make libc6-i386 depend on lib32ncurses5. This is stupid. > 2) Make libc6-i386 Replaces: lib32ncurses5. This should allow for smooth > upgrades. > > This is to say that the fix can't be accomplished on the ncurses end. > The damage has already been done from 5.6-1 and libc6-i386 must clean > up. > > I understand that this won't be an issue in the coming months since > people will never see these two releases, but it's all about doing > things right. > > Thanks and please don't close the bug again without further discussion. WTF ?! This package (lib32ncurses5) will be updated in unstable and 5.6-3 won't exist anymore, we don't care about it beeing smoothly upgraded or not. unstable is unstable, suffer it, or use stable. No we won't add random replaces for packages that never made it to testing, else we would have all Debian in Replaces, which is a very bad thing (remember, if we put a replace and that we *actually* overwrite a file that is not ours, it's silentely ignored, very bad). So I agree with daniel here, and it's not a glibc bug *AT ALL*. kthxbye -- ·O· Pierre Habouzit ··O madcoder@debian.org OOO http://www.madism.org
Attachment:
pgpITGVWdFB5d.pgp
Description: PGP signature