As discussed at today's team meeting, over the weekend there was an [active discussion](https://discussion.fedoraproject.org/t/is-most-recent-mesa-push-to-stable-mesa-25-2-7-2-fc43-reasonable/173585) about a [mesa update](https://bodhi.fedoraproject.org/updates/FEDORA-2025-82b66363b4) which proved to have some problems.

This is my best understanding of what, broadly speaking, happened:

1. airlied (Dave Airlie, RH employee, mesa maintainer) submitted a mesa update for F43 which included a backport of a PR he had submitted upstream but which had not yet been approved and merged. He didn't tag any specific bug (downstream or upstream) in the update or the upstream PR. He disabled autopush on the update.
2. rpmfusion created a matching updated to their mesa 'freeworld' packages.
3. The Fedora update got quite a lot of negative feedback, mainly from folks reporting some variation on "games don't work in Steam on my NVIDIA card any more".
4. The rpmfusion update went stable.
5. leigh123linux (Leigh Scott, not RH employee, provenpackager but apparently with prior approval to do stuff to mesa related to staying in sync with rpmfusion) pushed the Fedora update stable (with no comment on the negative feedback).
6. The thread got started to complain about the update going stable and breaking stuff.
7. The issue was confirmed to be real, airlied sent out a [new update with the fix](https://bodhi.fedoraproject.org/updates/FEDORA-2025-2f4ba7cd17).

The biggest concerns I see with what happened here are:

1. The fact that there was a backport of an unreviewed and unmerged upstream PR without a clear explanation either upstream or downstream as to exactly what the update was meant to fix. All I've found is "This also helps fix a deadlock seen where a compositor has WAYLAND_DISPLAY set and tries to initialise zink and device select tries to connect to the wayland display" in the upstream PR, but there's no link to an actual issue report so we could get any kind of sense of why this backport was happening at all, and why it would be important enough to justify doing a backport before it was fully reviewed/merged upstream.
2. The fact that the update was pushed stable despite substantial negative feedback and with no explanation to justify that choice.

In the meeting we agreed that an appropriate response would be to contact both involved maintainers and ask if they have any plans to avoid a recurrence. I said I'd draft such a message, so here it is:

===

Subject: Follow-up on the mesa-25.2.7-2.fc43 update that caused some problems
Body:
Hi folks! We (the Fedora QA team) noticed the situation with the mesa-25.2.7-2.fc43 update - https://bodhi.fedoraproject.org/updates/FEDORA-2025-82b66363b4 . Thanks for promptly sending out the follow-up update to fix the problem. We wanted to check in and see if you've got any thoughts or plans for preventing anything like this happening again in future. Thanks!

===

How does that sound? And how should we deliver it? At a quick check it looks like Leigh is pretty active on discourse, Dave has posted one time. They're both active on mailing lists. We could send it to test@ and both maintainers, or we could create a thread on discourse and tag both of them. Or we could do...something else. Thoughts?

Thanks folks!