Bug 206958 – [ macOS wk2 ] http/tests/media/media-stream/get-display-media-prompt.html is flaky failure

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Jacob Uphoff]

http/tests/media/media-stream/get-display-media-prompt.html 

this test is a flaky failure on MacOS 

test is also flaky timing out 

I was able to reproduce the failure by running the test in iterations

History:
https://results.webkit.org/?suite=layout-tests&test=http%2Ftests%2Fmedia%2Fmedia-stream%2Fget-display-media-prompt.html
Diff:
--- /Volumes/Data/slave/catalina-release-tests-wk2/build/layout-test-results/http/tests/media/media-stream/get-display-media-prompt-expected.txt
+++ /Volumes/Data/slave/catalina-release-tests-wk2/build/layout-test-results/http/tests/media/media-stream/get-display-media-prompt-actual.txt
@@ -40,12 +40,12 @@
 PASS err.name is "TypeError"
 
 ** Request a stream with valid constraints, the user should be prompted **
-PASS numberOfTimesGetUserMediaPromptHasBeenCalled() is 3
-PASS stream.getAudioTracks().length is 0
-PASS stream.getVideoTracks().length is 1
+FAIL numberOfTimesGetUserMediaPromptHasBeenCalled() should be 3. Was 2.
+FAIL stream.getAudioTracks().length should be 0. Threw exception TypeError: undefined is not an object (evaluating 'stream.getAudioTracks')
+FAIL stream.getVideoTracks().length should be 1. Threw exception TypeError: undefined is not an object (evaluating 'stream.getVideoTracks')
 
 ** Request a stream with an exact audio constraint, it should be ignored **
-PASS numberOfTimesGetUserMediaPromptHasBeenCalled() is 4
+FAIL numberOfTimesGetUserMediaPromptHasBeenCalled() should be 4. Was 3.
 PASS stream.getAudioTracks().length is 0
 PASS stream.getVideoTracks().length is 1

[Jacob Uphoff]

here is the timeout bug https://bugs.webkit.org/show_bug.cgi?id=206959

[Radar WebKit Bug Importer]

<rdar://problem/59003765>

[Jacob Uphoff]

Created attachment 389177 [details]
Update Test Expectations

[Truitt Savell]

Comment on attachment 389177 [details]
Update Test Expectations

Clearing flags on attachment: 389177

Committed r255379: <https://trac.webkit.org/changeset/255379>

[Alexey Proskuryakov]

*** Bug 206959 has been marked as a duplicate of this bug. ***

[Alexey Proskuryakov]

No failures before 2019-11-05. First failure was with r252047, which was a media related change, but failures were not frequent enough to positively blame.

[youenn fablet]

Created attachment 389504 [details]
Patch

[Eric Carlson]

Comment on attachment 389504 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=389504&action=review

> Source/WebCore/ChangeLog:11
> +        For that reason, we were storing a pointer to the gesture event and resetting in case the current gesture event was equal to the stored pointer.
> +        Instead, store a WeakPtr which ensures that if the previous event is destroyed, the weak pointer will be null and so will not have the same value as the new gesture event.

Nit: these lines are really long and should be wrapped.

[youenn fablet]

Created attachment 389512 [details]
Patch for landing

[WebKit Commit Bot]

Comment on attachment 389512 [details]
Patch for landing

Clearing flags on attachment: 389512

Committed r255582: <https://trac.webkit.org/changeset/255582>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[youenn fablet]

Removing flaky test expectation in https://bugs.webkit.org/show_bug.cgi?id=207893.