Bug 223283 – REGRESSION (r274461): [macOS wk1] fast/repaint/canvas-object-fit.html is flakey text failing

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: New Bugs

[Robert Jenner]

fast/repaint/canvas-object-fit.html

is a flakey text failure in macOS Catalina and BigSur in release and debug on Intel and Apple Silicon only in wk1. 


HISTORY URL:
https://results.webkit.org/?suite=layout-tests&test=fast%2Frepaint%2Fcanvas-object-fit.html&platform=mac

TEXT DIFF:
@@ -10,9 +10,6 @@
           (position 8.00 8.00)
           (bounds 202.00 102.00)
           (drawsContent 1)
-          (repaint rects
-            (rect 70.00 20.00 7.00 7.00)
-          )
         )
       )
     )

It is a flakey text failure, but the failures to appear to happen more often than not.

[Radar WebKit Bug Importer]

<rdar://problem/75495803>

[Robert Jenner]

Was able to reproduce and bisect to a regression point.


I reproduced the failure using the following test:
run-webkit-test fast/repaint/canvas-object-fit.html --iterations 100 -f -1

The test produced the text failure at ToT, and 274461. Failures did not occur at 274460. 

Changes in 274460 do appear to be related to the test that is failing. 
https://trac.webkit.org/changeset/274461/webkit

[Robert Jenner]

Updated test expectations for this test in wk1 to Pass Failure while test is being reviewed: 
https://trac.webkit.org/changeset/274525/webkit

[Simon Fraser (smfr)]

Bad:

16:23:02.570 41683   Page::updateRendering
16:23:02.570 41683   CanvasRenderingContext2DBase::didDraw
16:23:02.570 41683   HTMLCanvasElement::didDraw - 0.0,0.0 200.0x200.0
16:23:02.570 41683   CanvasRenderingContext2DBase::didDraw
16:23:02.570 41683   HTMLCanvasElement::didDraw - null rect
16:23:02.570 41683   Page::updateRendering
16:23:02.570 41683   HTMLCanvasElement 0x160430108 paint
16:23:02.570 41683   Page::updateRendering

Good:

16:23:02.499 41683   CanvasRenderingContext2DBase::didDraw
16:23:02.499 41683   HTMLCanvasElement::didDraw - 0.0,0.0 200.0x200.0
16:23:02.499 41683   HTMLCanvasElement 0x1609f2850 paint
16:23:02.499 41683   CanvasRenderingContext2DBase::didDraw
16:23:02.499 41683   HTMLCanvasElement::didDraw - 40.0,40.0 10.0x10.0
16:23:02.499 41683   Page::updateRendering
16:23:02.499 41683   HTMLCanvasElement 0x1609f2850 paint
16:23:02.499 41683   Page::updateRendering

Note the "HTMLCanvasElement 0x1609f2850 paint" is missing in the bad case.

[Simon Fraser (smfr)]

Bad:
16:36:19.749 46316   WebViewLayerFlushScheduler::schedule()
16:36:19.749 46316   WebViewLayerFlushScheduler::layerFlushCallback()
16:36:19.749 46316   LayerFlushController::flushLayers()
16:36:19.749 46316   WebView 0x7f800b40bec0 _updateRendering
16:36:19.750 46316   Page::updateRendering
16:36:19.750 46316   CanvasRenderingContext2DBase::didDraw
16:36:19.750 46316   HTMLCanvasElement::didDraw - 0.0,0.0 200.0x200.0
16:36:19.750 46316   CanvasRenderingContext2DBase::didDraw
16:36:19.750 46316   HTMLCanvasElement::didDraw - null rect
16:36:19.750 46316   WebViewLayerFlushScheduler::schedule()
16:36:19.750 46316   _forceRepaintForTesting
16:36:19.750 46316   WebView 0x7f800b40bec0 _updateRendering
16:36:19.750 46316   Page::updateRendering
16:36:19.750 46316   _forceRepaintForTesting - flushing
16:36:19.750 46316   WebSimpleLayer display
16:36:19.750 46316   HTMLCanvasElement 0x15df38210 paint
16:36:19.750 46316   WebViewLayerFlushScheduler::layerFlushCallback()
16:36:19.750 46316   LayerFlushController::flushLayers()
16:36:19.750 46316   WebView 0x7f800b40bec0 _updateRendering
16:36:19.750 46316   Page::updateRendering

Good:
16:36:19.675 46316   WebViewLayerFlushScheduler::schedule()
16:36:19.675 46316   CanvasRenderingContext2DBase::didDraw
16:36:19.675 46316   HTMLCanvasElement::didDraw - 0.0,0.0 200.0x200.0
16:36:19.675 46316   WebSimpleLayer display
16:36:19.675 46316   HTMLCanvasElement 0x15df38108 paint
16:36:19.675 46316   WebViewLayerFlushScheduler::schedule()
16:36:19.675 46316   CanvasRenderingContext2DBase::didDraw
16:36:19.675 46316   HTMLCanvasElement::didDraw - 40.0,40.0 10.0x10.0
16:36:19.675 46316   WebViewLayerFlushScheduler::schedule()
16:36:19.675 46316   _forceRepaintForTesting
16:36:19.675 46316   WebView 0x7f800b40bec0 _updateRendering
16:36:19.675 46316   Page::updateRendering
16:36:19.675 46316   _forceRepaintForTesting - flushing
16:36:19.675 46316   WebSimpleLayer display
16:36:19.675 46316   HTMLCanvasElement 0x15df38108 paint
16:36:19.675 46316   WebViewLayerFlushScheduler::layerFlushCallback()
16:36:19.675 46316   LayerFlushController::flushLayers()
16:36:19.675 46316   WebView 0x7f800b40bec0 _updateRendering
16:36:19.675 46316   Page::updateRendering

[Simon Fraser (smfr)]

Created attachment 423413 [details]
Patch

[Simon Fraser (smfr)]

Created attachment 423416 [details]
Patch

[Ryan Haddad]

Is this good to land?

[EWS]

Tools/Scripts/svn-apply failed to apply attachment 423416 [details] to trunk.
Please resolve the conflicts and upload a new patch.

[Ryan Haddad]

Created attachment 427600 [details]
Patch for landing

[EWS]

Committed r276924 (237266@main): <https://commits.webkit.org/237266@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 427600 [details].

[Truitt Savell]

This change https://trac.webkit.org/changeset/276924/webkit

unfortunately did not fix this test. History still shows flaky failures and the same diff.

[Ryan Haddad]

Marked test as flaky again in https://trac.webkit.org/changeset/278938/webkit

[Simon Fraser (smfr)]

Reproduced with run-webkit-tests -1 --iterations=300 --force fast/repaint/canvas-object-fit.html

[Hercules Hjalmarsson]

*** Bug 246653 has been marked as a duplicate of this bug. ***

[Hercules Hjalmarsson]

*** Bug 245902 has been marked as a duplicate of this bug. ***

[Said Abou-Hallawa]

This should be fixed by the fix of bug 254200.

*** This bug has been marked as a duplicate of bug 254200 ***