5 tests in virtual/gpu/fast/canvas are flaky when run in random order |
|||||||||
Issue descriptionThe tests: virtual/gpu/fast/canvas/canvas-createImageBitmap-webgl.html virtual/gpu/fast/canvas/canvas-drawImage-live-video.html virtual/gpu/fast/canvas/canvas-lose-restore-googol-size.html virtual/gpu/fast/canvas/canvas-lost-gpu-context.html virtual/gpu/fast/canvas/OffscreenCanvas-2d-drawImage.html appear to be flaky when run in random order. Steps to reproduce: This should pass consistently: Tools/Scripts/run-webkit-tests virtual/gpu/fast/canvas This should fail intermittently: Tools/Scripts/run-webkit-tests virtual/gpu/fast/canvas --order=random We're working on removing order dependency from Blink layout tests. More context here: https://docs.google.com/document/d/11-628Ny6hY__kfIv__zVD0m3rA0OB9tQLLC_gJI_PNg/edit
,
Nov 14 2016
,
Jan 17 2017
Feel free to re-assign
,
Jan 19 2017
,
Jan 26 2017
Could we just disable some of the tests above? V8's blink trybot (v8_linux_blink_rel) fails regularly on the canvas-lost-gpu-context.html test causing additional work to check each time what's going on: see https://build.chromium.org/p/tryserver.v8/builders/v8_linux_blink_rel/builds/14754
,
Jan 26 2017
Hm, virtual/gpu/fast/canvas/canvas-lost-gpu-context.html is currently marked as [ Pass Failure ], but in that example it crashes. We could mark it as [ Pass Failure Crash ], but then it's not helping provide any coverage. Maybe that's the only one of these tests that crashes, and maybe that's a separate problem? junov@, what do you think about filing a separate bug and marking that as [ Pass Failure Crash ]? Do you have any ideas about what may cause the crash (example crash https://storage.googleapis.com/chromium-layout-test-archives/v8_linux_blink_rel/14754/layout-test-results/results.html)?
,
Feb 27 2017
Flakiness board is showing all of these tests passing all runs. There are no expectations at this time. I'm closing it.
,
Mar 13 2017
These five tests are still listed as [ Pass Failure ] in TestExpectations, and still refer to this bug. Two of them are definitely still flaky on the bots, presumably even when run in non-random order (canvas-createImage-webgl and OffscreenCanvas-2d-drawImage.html. The other three are green, and could be removed from TestExpectations.
,
Mar 21 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/f5c3d3532d4634d5be040bae58664f68d9470b64 commit f5c3d3532d4634d5be040bae58664f68d9470b64 Author: schenney <schenney@chromium.org> Date: Tue Mar 21 14:39:00 2017 Remove non-flaky tests. Three of the tests under this bug are not flaky, while two are very flaky. Marking it such. TBR=junov@chromium.org BUG=664852 Review-Url: https://codereview.chromium.org/2765463003 Cr-Commit-Position: refs/heads/master@{#458404} [modify] https://crrev.com/f5c3d3532d4634d5be040bae58664f68d9470b64/third_party/WebKit/LayoutTests/TestExpectations
,
Feb 27 2018
Looks like fast/canvas/canvas-drawImage-live-video.html is still flaky: https://ci.chromium.org/buildbot/tryserver.chromium.mac/mac_chromium_rel_ng/659500 Swarming: https://chromium-swarm.appspot.com/task?id=3bed80cf2df2a510&refresh=10&show_raw=1 Swarming logs have multiple lines like this: 16:52:14.851 14579 [7042/10232] fast/canvas/canvas-drawImage-live-video.html failed unexpectedly (asserts failed) This indicates a crash right? Maybe we should mark this test as possibly crashing in TestExpectations?
,
May 8 2018
,
Jul 25
,
Jul 25
|
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by xidac...@chromium.org
, Nov 14 2016Status: Available (was: Untriaged)