The following layout test is flaky on all (chromium?) platforms fast/events/mouse-cursor-image-set.html http://trac.webkit.org/browser/trunk/LayoutTests/fast/events/mouse-cursor-image-set.html http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&tests=fast%2Fevents%2Fmouse-cursor-image-set.html This test has only been around for 7 days.
I'm not sure if it's an image loading issue: TEST CASE: Implicit hot-spot at (5,4) physical for 1x and (28,3) physical for 2x -Cursor Info: type=Custom hotSpot=5,4 image=25x25 +Cursor Info: type=Custom hotSpot=0,0 image=0x0 it always seems to fail the same subtest.
I think it's this typo here: { url: 'resources/greenbox-hotspot35-4.cur' }, the '3'. The loading system seems fragile. Maybe we could read those images from the CSS values?
I'll look into this. Yes if the test fails to reliably preload one of the images then it can get output like this. Ideally I'd be able to assert that an image was already preloaded or something. And yes I thought about trying to combine the lists to avoid duplication and risk of inconsistency but I didn't come up with anything that was clean and simple. I'll re-evaluate. Let me know if you have any suggestions.
Dammit, and yes this is just a typo or copy-paste error with 'hotspot35-4' that should be 'hotspot5-4', dammit - sorry about that. I'll submit a trivial fix now, and file a separate bug to improve the infrastructure to make such mismatches impossible (which I probably won't get to until next week since I'm travelling this week).
Underlying issue tracked in bug 104952
Created attachment 179363 [details] Patch
Comment on attachment 179363 [details] Patch Thank you Rick and Beth!
(In reply to comment #7) > (From update of attachment 179363 [details]) > Thank you Rick and Beth! Thank you for noticing and ccing me! I wish we had a way for test owners to be auto-notified of any failures.
Comment on attachment 179363 [details] Patch Clearing flags on attachment: 179363 Committed r137684: <http://trac.webkit.org/changeset/137684>
All reviewed patches have been landed. Closing bug.