This does not seem strong enough - in case of https://pagure.io/releng/issue/10661/ (which is caused by one package overlapping) I noticed the package name mentioned but didn't have time to consider the impact. In case the updates overlap in more packages it would have been hard to even see the packages impacted.
In case the update containing the newer package gets pushed first (e.g. karma threshold reached first), the update with the older package will then end up getting pushed to the compose later, and because we only keep one updated package, that wins.
We can probably consider a few safety measures. From the easiest to to the hardest: - when composing, filter out updates with lower NEVRA (so in this case, rust-below-0.4.1-4 will not get pushed since 0.5.0-1 is already pushed) - when creating an update, make the warning about overlapping packages more prominent and sticky - also disable auto-push and add a sticky notice / edit the comment to indicate the other update set needs to land first - if an update is a superset of another and has all-newer packages, it should obsolete the other update
When do you need this? (YYYY/MM/DD) N/A - RFE
When is this no longer needed or useful? (YYYY/MM/DD) N/A
If we cannot complete your request, what is the impact? Race conditions between updates, potentially causing older updates to override newer updates in the future
https://github.com/fedora-infra/bodhi/issues/1266 - there is a long standing issue that was filed with bodhi long ago.
Closing the ticket, please follow the upstream ticket.
Metadata Update from @mohanboddu: - Issue close_status updated to: upstream - Issue status updated to: Closed (was: Open)