Hey, 50/50 chance I can't make it to the meeting in the morning. This hasn't been updated in ~10 months. https://fedoraproject.org/wiki/Changes/WaylandByDefault Tracker bug has a lot of bugs still open (60-70) https://bugzilla.redhat.com/show_bug.cgi?id=1277927 So that's a bit daunting. And then I started to think whether tests need complete duplication to make sure they work on both X and Wayland? And what would those be? And how many test cases need to be modified, or rewritten? Then came a dozen more questions. I gave up on organizing those, as to me it really comes down to just two (big) things: Does this Fedora Final release criterion apply to both Xorg and Wayland? All applications that can be launched using the standard graphical mechanism of a release-blocking desktop after a default installation of that desktop must start successfully and withstand a basic functionality test. i.e. do we do the testing implied by that criterion, twice, once for Xorg and once for Wayland? My answer is yes, otherwise we don't really know where the bodies are buried in Wayland, and any missing testing might expose lower standards for Xorg which needs to be a reliable fallback position. Next is it a blocker? Or is it a "reverter" i.e. a reason why Wayland by default should not proceed for Fedora 25, trigger the contingency and use Xorg by default instead? That's where this gets tricky because by necessity the standard must be lowered: there are definitely going to be bugs in this area, and we know in advance Wayland doesn't have completely parity in all ways to Xorg. So some of this process has to be subjectively determined rather than objective. And subjective != arbitrary, so I make a special point here to not overly resist subjectivity. There's another reason in six months, it's not like Wayland standards will always be as subjective as they are this release. Keypoint? For the purpose of testing and bug reporting, yes I think we probably need to do the "basic functionality test" for all applications included on the live media, with both Xorg and Wayland. Such bugs on Xorg are probably blocking as they have been in the past; and bugs on Wayland will need some subjectivity to determine if they are "revert" worthy or if it's acceptable to say users can't just use Xorg to work around it? That's all I have for now. -- Chris Murphy -- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://lists.fedoraproject.org/admin/lists/test@xxxxxxxxxxxxxxxxxxxxxxx
