On Thu, 13 Jun 2013 11:52:52 -0700 Adam Williamson <awilliam@xxxxxxxxxx> wrote: > Note that the situation with sssd is a bit complex. The broken update > was submitted for updates-testing on 06-12 at 12:01, and pushed to > updates-testing on 06-12 at 12:41. It was then marked to be > 'unpushed' - i.e. taken off the mirrors - on 06-13 at 01:28. A fixed > sssd package was then added to the update and submitted for > updates-testing on 06-13 at 10:56, and pushed to updates-testing on > 06-13 at 15:53. > > I'm not sure whether unpushes require any manual action or if they > happen automatically, but if the unpush actually happened, then there > was a time when the 'most current' state of the mirrors would have an > *older* sssd package than a 'less current' state of the mirrors - a > 'less current' mirror would still have the broken update, but a 'more > current' mirror would have had it removed. It's not _always_ the case > that the mirror with a higher-versioned package is the more up to > date. 99% of the time, but not always. Yeah, true. When a package that has been 'pushed' to updates-testing is unpushed, it simply means the next compose that package will be removed. It doesn't do it right when the 'unpush' event happens, if it did it would require a full recompose/push of the updates-testing repo, which isn't practical. So yeah, correct with updates-testing. kevin
Attachment:
signature.asc
Description: PGP signature
-- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
