>>>>> "Paul" == Paul Gevers <elb...@debian.org> writes:
Paul> Just to elaborate, that set is frozen because bits that Paul> influence a lot of builds can become difficult to manage if a Paul> bug sneaks in. So it's good to discuss here how much pam can Paul> influence the built artifacts. Hence, it's not so much we Paul> want to prevent it in bullseye, but it's potential impact for Paul> the release process is already in unstable. Nod. Having broken pam for everyone once, I'm very aware of that kind of issue. I certainly wouldn't change pam for non-important issues at this point. My concern would be that we might break things and impact development or buildds *or* that we might break upgrades. (Doc changes are of course fine--the standard freeze policy stuff). At one level, pam probably doesn't matter that much for buildds because most things don't involve authentication. And yet enough links against libpam that it can be a big deal if it fails. Also, if it fails to upgrade, it could create a mess that way. --Sam