(how is it possible that my spamfilter eats valid important bugreports, but mails with strange words (aka tags) added to these bugs are considered ham…)
So now that i had a short look at it i think we need to change the lockstep handling introduced to fix debbug 618288 (in APT 0.8.15) a bit further by doing the unpack first for all packages which are (now) M-A:same and already installed and only after that work on the non-installed silbings.
Attached patch should do this from my branch, but i don't have the time to test that properly currently. You might need to backport the other packagemanager.cc fix in my branch, too (i guess lp will establish the link soon). I will try to work that out with Michael then he is back in action (and i am not mostly occupied by exams).
(how is it possible that my spamfilter eats valid important bugreports, but mails with strange words (aka tags) added to these bugs are considered ham…)
So now that i had a short look at it i think we need to change the lockstep handling introduced to fix debbug 618288 (in APT 0.8.15) a bit further by doing the unpack first for all packages which are (now) M-A:same and already installed and only after that work on the non-installed silbings.
Attached patch should do this from my branch, but i don't have the time to test that properly currently. You might need to backport the other packagemanager.cc fix in my branch, too (i guess lp will establish the link soon). I will try to work that out with Michael then he is back in action (and i am not mostly occupied by exams).