Bug 120922
| Summary: | compositing/overlap-blending/children-opacity-huge.html and children-opacity-no-overlap.html are failing | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Ryosuke Niwa <rniwa> |
| Component: | Layout and Rendering | Assignee: | Nobody <webkit-unassigned> |
| Status: | NEW | ||
| Severity: | Normal | CC: | ap, dino, simon.fraser, thorton |
| Priority: | P2 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Ryosuke Niwa
http://build.webkit.org/results/Apple%20MountainLion%20Release%20WK1%20(Tests)/r155235%20(12619)/results.html
http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=compositing%2Foverlap-blending%2Fchildren-opacity-huge.html%2Ccompositing%2Foverlap-blending%2Fchildren-opacity-no-overlap.html
It seems like we've introduced some nondeterministic behavior in WebKit. Random compositing tests are failing on each run. This is bad. It's going to clog EWS and Commit Queue completely.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Ryosuke Niwa
Committed r155237: <http://trac.webkit.org/changeset/155237>
Simon Fraser (smfr)
Did it start with http://trac.webkit.org/changeset/155140 ?
Ryosuke Niwa
(In reply to comment #2)
> Did it start with http://trac.webkit.org/changeset/155140 ?
It's hard to tell. Because there is no single test that has been failing.
http://build.webkit.org/builders/Apple%20MountainLion%20Release%20WK2%20%28Tests%29?numbuilds=100
Ryosuke Niwa
It seems like media/video-object-fit.html started failing after that patch on Lion:
http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=media%2Fvideo-object-fit.html
Ryosuke Niwa
EWS seems to indicate something bad happened about 10 hours ago.
Simon Fraser (smfr)
That would implicate http://trac.webkit.org/changeset/155196
Ryosuke Niwa
(In reply to comment #6)
> That would implicate http://trac.webkit.org/changeset/155196
I've rolled that change out in http://trac.webkit.org/changeset/155235 but that change was in response to EWS getting behind so the underlying issue must be somewhere else.
Alexey Proskuryakov
These two tests are failing for me locally with r155076.