HTMLMediaElement::virtualHasPendingActivity may keep objects alive unnecessarily
<rdar://78726921>
https://jsfiddle.net/cpx73bry/ is an example where we are keeping the object alive for no good reasons.
Created attachment 430336 [details] Patch
Created attachment 430340 [details] Patch
Comment on attachment 430340 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=430340&action=review > LayoutTests/fast/mediastream/mediaElement-gc.html:41 > +promise_test(async (test) => { New test is flaky on EWS: - PASS GC a video element once its srcObject gets ended + FAIL GC a video element once its srcObject gets ended assert_less_than: expected a number less than 100 but got 100
(In reply to Chris Dumez from comment #5) > Comment on attachment 430340 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=430340&action=review > > > LayoutTests/fast/mediastream/mediaElement-gc.html:41 > > +promise_test(async (test) => { > > New test is flaky on EWS: > - PASS GC a video element once its srcObject gets ended > + FAIL GC a video element once its srcObject gets ended assert_less_than: > expected a number less than 100 but got 100 Hum, not sure why it fails, let's start with the main bug. The ended MediaStream case can be fixed as a follow-up.
Created attachment 430359 [details] Patch
Committed r278359 (238391@main): <https://commits.webkit.org/238391@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 430359 [details].
It looks like the new test fast/mediastream/mediaElement-gc.html added in https://trac.webkit.org/changeset/278359/webkit has been a flaky failure on Mac sense introduction: https://results.webkit.org/?suite=layout-tests&test=fast%2Fmediastream%2FmediaElement-gc.html tracking in: https://bugs.webkit.org/show_bug.cgi?id=226991