Closed #1346 as completed via #3004.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/1346#event-12352564431
You are receiving this because you are subscribed to this thread.
Message ID:
___
Thinking some more on this, I believe the right thing here is treating regular
(as in not having (post/pre) type qualifiers) weak dependencies as implicit
"meta" dependencies, which is exactly how they are *normally* used (eg that
fwupd case).
For the rare case where packagers do want them to
Yet another practical example from Bugzilla:
https://bugzilla.redhat.com/show_bug.cgi?id=2252661
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/1346#issuecomment-1978531887
You are receiving this because you are subscribed to this
This seems like a rather important issue to fix... it could help eliminate lots
of otherwise-unnecessary `%posttrans` and `%pretrans` scripts that exist only
due to dependency-order unpredictability.
--
You are receiving this because you are subscribed to this thread.
Reply to this email
Yup, weak dependencies can and do create massive loops that do not exist
without them as weak dependencies are more liberally sprinkled around. They
should be taken into account in ordering but not equal to hard dependencies
(including pre- and other variants and their side-effects)
--
You
RPM now considers weak dependencies during the transaction ordering. However
weak dependencies were sometimes also used as a way to actually break the
dependency loops between the packages.
For that reason the transaction ordering algorithm should break dependency
loops at weak dependencies if