Hi, On Sun, Jul 10, 2011 at 04:31:01PM -0500, Jonathan Nieder wrote: > Rene Engelhard wrote: > > On Sat, Jul 09, 2011 at 03:27:09PM -0500, Jonathan Nieder wrote: > > >> Typical experimental amd64 system. From today's upgrade: > > > > How did you do the upgrade? > > > > And what is a "typical experimental amd64 system"? > > I used cupt;
Ah. Something extravagant I do not and never will care about. > you can reproduce it by using dpkg directly, too. True, but using dpkg is not always senseful. I.e. in lenny->squeeze upgrade where upgrade order by apt etc. might be important. And/or in complex dependencies. > Besides, after your explanation it's obvious in the code where the > problem is. True > And the policy is very clear; "apt happens to choose an > upgrade path that doesn't trigger it" does not make a bug go away. Honestly, I disagree. If dpkg -i *.deb breaks because you install stuff which is Break:'ing or Conflict:'ing each other and that fails it's not a bug in the package. It's dpkg not figuring it out, which is OK. it's apt(itutude)s job. > Anyway, I'm happy to take care of writing a fix and checking that it > fixes the problem, so I am not so inclined to attach dpkg logs etc. I > should be able to send a fix soon. I'll be glad to hear feedback on > it then. I have the check already there, just that the problem is that when what is done the trigger never might run when installing -evolution or -binfilter - effectively breaking them because LibO then doesn't know about them in any way. That's why I asked you how you upgraded. Grüße/Regards, René -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
