Bug 298903 – REGRESSION(299383@main): [ iOS Release, macOS wk2 ] webrtc/video-rotation-black.html is a flaky text failure

• Component: WebRTC
• Status: RESOLVED FIXED

[Bryan Medina]

webrtc/video-rotation-black.html 

is a flaky failure on iOS Release and macOS wk2, although it is more flaky on Mac than it is on iOS.

HISTORY:
https://results.webkit.org/?suite=layout-tests&test=webrtc%2Fvideo-rotation-black.html

TEXT DIFF:
--- /Volumes/Data/worker/Apple-iPadOS-18-Simulator-Debug-WK2-Tests/build/layout-test-results/webrtc/video-rotation-black-expected.txt
+++ /Volumes/Data/worker/Apple-iPadOS-18-Simulator-Debug-WK2-Tests/build/layout-test-results/webrtc/video-rotation-black-actual.txt
@@ -1,4 +1,4 @@
 
 
-PASS Rotated disabled track should keep sending the same intrinsic size video if CVO is on
+FAIL Rotated disabled track should keep sending the same intrinsic size video if CVO is on promise_test: Unhandled rejection with value: "checkVideoBlack timed out for 'canvas0', expected true"
 
TEXT DIFF URL:
https://build.webkit.org/results/Apple-iPadOS-18-Simulator-Debug-WK2-Tests/299962@main%20(1738)/webrtc/video-rotation-black-diff.txt

[Radar WebKit Bug Importer]

<rdar://problem/160646554>

[Bryan Medina]

I was able to reproduce the text failure locally on my device running iOS Debug ToT using the following command:

run-webkit-tests webrtc/video-rotation-black.html --iOS-simulator --debug --iterations=5

[Bryan Medina]

It seems that the test started to flakily fail much more after 299383@main. I've found that testing at 299383@main also produces failures, whereas testing at 299382@main does not, leading me to believe that 299383 is the regression point.

[EWS]

Test gardening commit 299988@main (07401dec43b8): <https://commits.webkit.org/299988@main>

Reviewed commits have been landed. Closing PR #50776 and removing active labels.

[youenn fablet]

Pull request: https://github.com/WebKit/WebKit/pull/50815

[EWS]

Committed 300148@main (94ae61ff0973): <https://commits.webkit.org/300148@main>

Reviewed commits have been landed. Closing PR #50815 and removing active labels.