Since this bug is over a year old and I have since upgraded my version of aptitude to 0.6.3-3.2+squeeze1 and the sources of the packages on my system are a lot less mixed now, I am not a hundred percent certain, but I think the problem was my misunderstanding of the aptitude documentation.
In my naïveté I think I assumed then that "aptitude search '?narrow(?installed,?archive(unstable))'" would show packages that are installed from the unstable repositories. If I understand the documentation correctly now though (and feel free to correct me if I don't), aptitude does not really track from which repositories a given package comes from, it only tracks versions. That is installing version x.y.z of a given package from the testing repositories yields the same state, as far as aptitude is concerned, as installing version x.y.z of the same package from the unstable repositories.(assuming here that the version of our package is the same in testing and unstable) That the other variants, i.e."aptitude search '?and(?installed,?archive(testing))'" etc., did what I expected them to, was probably just coincidence due to the particular combination of installed packages and version matches or mismatches between the different repositories. So I'd say close the bug (it was in all likelihood a 'bug in the user not in the program' ;-) ) and maybe update the aptitude documentation to make it abundantly clear that the source repository of an installed package is not tracked, i.e. "'?narrow(?installed,?archive(unstable))'" matches a package with a version that is both in the unstable repositories as well as installed, but that is not necessarily installed from the unstable repositories. Thank you for your time. Regards _______________________________________________ Aptitude-devel mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/aptitude-devel
