Upgrade Wayland in JHBuild moduleset. This is needed for EGLGetDisplay() to work on my computer.
If I'm not wrong wayland is not needed at all to run the tests, becase we use the headless backend, so wayland is just a dep of dyz, and we don't need to build it ourselves, so I would just remove it from the jhbuild moduleset. Unless I'm missing something. . .
Yes,since Wayland is not used for running tests, and since it does not seem to depend on anything in the JHBuild moduleset, I think we can remove it safely without having to worry that the system version of Wayland might use libraries in the jhbuild environment (which would be unsafe).
It's also a dep of wpe-mesa, but that does not matter. It would only be a problem if the dependency was reversed.
Created attachment 327457 [details] Patch
Zan thinks Wayland can affect OpenGL tests, even using headless mode, so I'll upgrade it instead.
Created attachment 327468 [details] Patch
Comment on attachment 327468 [details] Patch Unsetting r? because Carlos Lopez thinks this is a bad idea.
Comment on attachment 327457 [details] Patch Setting r? on this one, since I suspect this is the one we'll have to go with... even if it means we don't get consistent results for WebGL tests. (But I'm still not clear as to why Wayland would affect headless mode results.)
(In reply to Michael Catanzaro from comment #5) > Zan thinks Wayland can affect OpenGL tests, even using headless mode, so > I'll upgrade it instead. mmmm... how this can happen?
I don't know.
(In reply to Carlos Alberto Lopez Perez from comment #9) > (In reply to Michael Catanzaro from comment #5) > > Zan thinks Wayland can affect OpenGL tests, even using headless mode, so > > I'll upgrade it instead. > > mmmm... how this can happen? Zan, any comment on this? If Wayland version can still somehow affect layout tests even though we now have headless mode, then I suggest we build Weston inside the JHBuild environment, and expect developers using the JHBuild environment to use that version of Weston. Otherwise, I think we should commit my patch to drop Wayland from the JHBuild environment.
OK, I was supposed to ensure this didn't get stalled... r?
(In reply to Michael Catanzaro from comment #11) > (In reply to Carlos Alberto Lopez Perez from comment #9) > > (In reply to Michael Catanzaro from comment #5) > > > Zan thinks Wayland can affect OpenGL tests, even using headless mode, so > > > I'll upgrade it instead. > > > > mmmm... how this can happen? > > Zan, any comment on this? > It apparently never happened. False memories on my side.
Comment on attachment 327457 [details] Patch Clearing flags on attachment: 327457 Committed r225327: <https://trac.webkit.org/changeset/225327>
All reviewed patches have been landed. Closing bug.