REGRESSION: animations/fill-mode-forwards.html is a flaky failure

• Component: Animations
• Status: NEW ---

[Eric Hutchison]

animations/fill-mode-forwards.html

is a flaky failure on BigSur wk2 Release and iOS15.

History: https://results.webkit.org/?suite=layout-tests&test=animations%2Ffill-mode-forwards.html

Results: https://build.webkit.org/results/Apple-BigSur-Release-WK2-Tests/r281633%20(4747)/results.html

Diff: 
--- /Volumes/Data/worker/bigsur-release-tests-wk2/build/layout-test-results/animations/fill-mode-forwards-expected.txt
+++ /Volumes/Data/worker/bigsur-release-tests-wk2/build/layout-test-results/animations/fill-mode-forwards-actual.txt
@@ -1 +1 @@
-PASS - Test working
+FAIL - pop animation should not be finished after pulse animation

[Radar WebKit Bug Importer]

<rdar://problem/82404950>

[Wenson Hsieh]

http://trac.webkit.org/changeset/281633/webkit is clearly unrelated.

[Eric Hutchison]

Unable to reproduce locally.

Updated test expectations at https://trac.webkit.org/changeset/281651/webkit.

[Fujii Hironori]

Still failing.
https://results.webkit.org/?suite=layout-tests&test=animations%2Ffill-mode-forwards.html

But, not only flaky text failure, but also flaky timeout.
Not only Release, but also Debug.
Not only wk2, but also wk1.

I think we should move the marker to the top TestExpectations.
> webkit.org/b/229584 animations/fill-mode-forwards.html [ Pass Failure Timeout ]

[Ahmad Saleem]

This test was removed from Blink here - https://codereview.chromium.org/29363005

*****

This test relies on receiving a specific number of animationiteration events.
However, in both the legacy and Web Animations implementations, we do not
guarantee that the correct number of these events is fired. See
crbug.com/275263. We have no immediate plans to fix this and the flakiness
dashboard shows timeouts for this test with both implementations.

The test was added to test a particular bug fix in the legacy animations
implementation (http://trac.webkit.org/changeset/117216) and has little to do
with forwards fill.

Most cases of forwards-fill testing (excluding play-state and events) is
covered by timing-model.html

This test therefore adds little value and should be removed. We can consider
reinstating it if and when the underlying bug is fixed.

****