In context if you read what I said I didn't imply Metux was the best out there; I was highlighting how absolutely terrible Wayland is and how I will take anything at this point.
So, here are a few direct links to the comments/issues/MRs (in chronological order), some very minor stuff, some bigger. I also didn't cherrypick examples where the maintainers acted perfectly, but IMO metux is far more difficult as you can see. Note I didn't search very long, so keep looking if you want to see more before you're sure:
- Small procedural issues: https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests...
- metux being difficult regarding work with nvidia: https://gitlab.freedesktop.org/xorg/xserver/-/commit/095e8ee...
- metux causes regression & it stays unaddressed for weeks, even though he was pushing on other branches: https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests...
- metux is marking threads as resolved when comments are unaddressed: https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests...
- metux causing unnecessary API breakage in commits marked trivial: https://gitlab.freedesktop.org/xorg/xserver/-/work_items/176...
- metux isn't properly testing his PRs: https://gitlab.freedesktop.org/xorg/xserver/-/work_items/179...
- metux breaking downstream again: https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests...
None of these are terrible on their own, but IMO taken together I can't fault the maintainers at all.
Not could break, did break. Multiple times. When a contributor repeatedly submits broken and untested PRs, the correct solution isn't to let him keep trying as often as he wants.