Bug 206614 – REGRESSION: [ Mac wk2 ] compositing/video/video-border-radius-clipping.html is a flaky image failure

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: Compositing

[Truitt Savell]

compositing/video/video-border-radius-clipping.html

Description:
This test is a flaky failure on Mac WK2. It is a known constant failure on Catalina wk1 from webkit.org/b/205226
The diff shows a 1 when it should be a 4 with a lesser blue curve on the left side.
I am able to reproduce this locally by just running this test in iterations. 
It looks like this started around r252246 which is the earliest failure.

History:
https://results.webkit.org/?suite=layout-tests&test=compositing%2Fvideo%2Fvideo-border-radius-clipping.html

Diff:
https://build.webkit.org/results/Apple-Catalina-Release-WK2-Tests/r254939%20(2328)/compositing/video/video-border-radius-clipping-diffs.html

[Radar WebKit Bug Importer]

<rdar://problem/58814182>

[Simon Fraser (smfr)]

We're getting a different video frame.

[Alexey Proskuryakov]

results.webkit.org shows a clear regression in early November, but internal bots have been flakily failing throughout the history.

Why is this test that is about clipping so flaky? Can we just have a video whose frames are all the same perhaps?

[Simon Fraser (smfr)]

It's flakey because it's apparently impossible to have a non-flakey test that actually shows video frames.

[Peng Liu]

*** This bug has been marked as a duplicate of bug 205226 ***