This is happening mostly (maybe exclusively) on OSX 10.10. The rate is low enough that I'm letting it slide for now in the interests of getting this enabled in production, but we'll need to keep an eye on it and potentially may end up skipping the test if the rate is too high. https://treeherder.mozilla.org/logviewer.html#?job_id=17575331&repo=try 21:07:53 INFO - REFTEST TEST-START | file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html 21:07:53 INFO - REFTEST TEST-LOAD | file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html | 13243 / 13292 (99%) 21:07:53 INFO - REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html | failed reftest-no-paint 21:07:53 INFO - REFTEST TEST-END | file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html 21:07:53 INFO - REFTEST INFO | Saved log: START file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] OnDocumentLoad triggering WaitForTestEnd 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] WaitForTestEnd: Adding listeners 21:07:53 INFO - REFTEST INFO | Saved log: Initializing canvas snapshot 21:07:53 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FIRE_INVALIDATE_EVENT 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: waiting for MozAfterPaint 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] AfterPaintListener in file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] SendUpdateCanvasForEvent with 1 rects 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] Rect: 0 0 800 1000 21:07:53 INFO - REFTEST INFO | Saved log: Updating canvas for invalidation 21:07:53 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FIRE_INVALIDATE_EVENT 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: dispatching MozReftestInvalidate 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] AttrModifiedListener fired 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_REFTEST_WAIT_REMOVAL 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_SPELL_CHECKS 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_APZ_FLUSH 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: done requesting APZ flush 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] AfterPaintListener in file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] SendUpdateCanvasForEvent with 0 rects 21:07:53 INFO - REFTEST INFO | Saved log: Updating canvas for invalidation 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_APZ_FLUSH 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_FOR_APZ_FLUSH 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: apz-repaints-flushed fired 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FINISH 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: waiting for MozAfterPaint 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] AfterPaintListener in file:///builds/slave/test/build/tests/reftest/tests/layout/reftests/invalidation/scroll-inactive-layers.html 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] SendUpdateCanvasForEvent with 2 rects 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] Rect: 0 -100 800 1536 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] Rect: 0 1536 785 2230 21:07:53 INFO - REFTEST INFO | Saved log: Updating canvas for invalidation 21:07:53 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000 21:07:53 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,1000,785,0 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: STATE_WAITING_TO_FINISH 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] MakeProgress: Completed 21:07:53 INFO - REFTEST INFO | Saved log: [CONTENT] RecordResult fired 21:07:53 INFO - REFTEST INFO | Saved log: RecordResult fired 21:07:53 INFO - REFTEST INFO | Saved log: RecordResult fired

this seems to be somehow now very frequent up to perma failure now. Markus, kats could you take a look ?

See https://bugzilla.mozilla.org/show_bug.cgi?id=1192910#c17 I'll try to repro it locally, but in the meantime if you want to back out bug 1192910 feel free.

I can reproduce this locally with the second patch from bug 1192910, I'll investigate.

This is basically the same problem that we fixed in bug https://bugzilla.mozilla.org/show_bug.cgi?id=1209852 for a different test. Patch coming up

The displayport was expanding due to the APZ repaint and triggering an invalidation. I think that's orthogonal to the intent of the test.

Comment on attachment 8727874 [details] [diff] [review] Patch Nice. I wonder if this is necessary for more of the scroll invalidation tests.