Bug 213063 – [Cocoa] Build callOnMainThread on WTF::RunLoop rather than on a timer

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Geoffrey Garen]

[Cocoa] Build callOnMainThread on WTF::RunLoop rather than on a timer

[Geoffrey Garen]

Created attachment 401617 [details]
Patch

[Geoffrey Garen]

Comment on attachment 401617 [details]
Patch

cq+

[EWS]

Committed r262904: <https://trac.webkit.org/changeset/262904>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 401617 [details].

[Radar WebKit Bug Importer]

<rdar://problem/64257294>

[Jason Lawrence]

Reverted r262904 for reason:

This commit broke a test on Mac wk1 Debug.

Committed r262977: <https://trac.webkit.org/changeset/262977>

[Darin Adler]

I wonder which test it broke. I’m sure that information is somewhere an expert can immediately find it, but I don’t know where to look.

[Simon Fraser (smfr)]

It looks like it made media/remoteplayback-target-availability.html flakey?

[Geoffrey Garen]

inspector/canvas/create-context-webgpu.html => fail
media/remoteplayback-target-availability.html => flakey fail if run after media/remoteplayback-prompt.html

[Geoffrey Garen]

Created attachment 402228 [details]
Patch for landing

[EWS]

Committed r263223: <https://trac.webkit.org/changeset/263223>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 402228 [details].

[Truitt Savell]

Reverted r263223 for reason:

Broke compositing/video/video-border-radius-clipping.html on Mac release wk1

Committed r263299: <https://trac.webkit.org/changeset/263299>

[Alexey Proskuryakov]

Regressed tests were rewritten to sidestep problems, but regressions on those tests are still not explained. I think that we need to understand why the changes are beneficial (or at least safe) before re-landing.

[Geoffrey Garen]

(In reply to Alexey Proskuryakov from comment #12)
> Regressed tests were rewritten to sidestep problems, but regressions on
> those tests are still not explained. I think that we need to understand why
> the changes are beneficial (or at least safe) before re-landing.

It's unfortunate that you started this discussion in two different places. I responded to your comment in https://bugs.webkit.org/show_bug.cgi?id=213424, and I'm relanding the patch now, since it has been reviewed and it passes all tests.

[Geoffrey Garen]

Created attachment 402485 [details]
Patch for landing

[EWS]

Committed r263357: <https://trac.webkit.org/changeset/263357>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 402485 [details].

[Geoffrey Garen]

> It's unfortunate that you started this discussion in two different places. I
> responded to your comment in https://bugs.webkit.org/show_bug.cgi?id=213424,

I think I should have said "It's fortunate that you called attention to this topic in both places; now let's put the discussion in https://bugs.webkit.org/show_bug.cgi?id=213424".