Bug 221331 – Test landed flaky: [iOS] TestWebKitAPI.WebKit2.CrashGPUProcessWhileCapturing

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Aakash Jain]

TestWebKitAPI.WebKit2.CrashGPUProcessWhileCapturing seems flaky on iOS. In https://ews-build.webkit.org/#/builders/9/builds/39710, the test Failed in run-api-tests step. However, in the immediately next retry step (re-run-api-tests), it passed. Same thing in https://ews-build.webkit.org/#/builders/9/builds/39698 and https://ews-build.webkit.org/#/builders/9/builds/39699

History: https://results.webkit.org/?suite=api-tests&test=TestWebKitAPI.WebKit2.CrashGPUProcessWhileCapturing

[Aakash Jain]

Seems like it has been flaky since it was added 5 days back in https://trac.webkit.org/r272050

[Aakash Jain]

From https://ews-build.webkit.org/#/builders/9/builds/39710/steps/10/logs/stdio:

    TestWebKitAPI.WebKit2.CrashGPUProcessWhileCapturing
        (webrtc_video_engine.cc:3316): Absent receive stream; ignoring clearing encoded frame sink for ssrc 0
        
        /Volumes/Data/worker/iOS-14-Simulator-Build-EWS/build/Tools/TestWebKitAPI/Tests/WebKit/GetUserMedia.mm:74
        Expected equality of these values:
          @"PASS"
            Which is: "PASS"
          [message body]
            Which is: "FAIL checkVideoStatus"

[Radar WebKit Bug Importer]

<rdar://problem/73935129>

[youenn fablet]

Created attachment 419264 [details]
Patch

[EWS]

Committed r272381: <https://trac.webkit.org/changeset/272381>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 419264 [details].

[youenn fablet]

Reopening since flakiness continues in the bots.

[Ryan Haddad]

Do we think this is a test only issue, or should we revert the original change?

[youenn fablet]

(In reply to Ryan Haddad from comment #7)
> Do we think this is a test only issue, or should we revert the original
> change?

This test identifies an issue in the way we handle GPU process capture on iOS.
I am working on fixing this. Once done, the test should no longer be flaky.
I hope to fix this by end of this week, early next week.
I am fine skipping this test on iOS for now.

The original change is fine.

[youenn fablet]

(In reply to youenn fablet from comment #8)
> (In reply to Ryan Haddad from comment #7)
> > Do we think this is a test only issue, or should we revert the original
> > change?
> 
> This test identifies an issue in the way we handle GPU process capture on
> iOS.
> I am working on fixing this. Once done, the test should no longer be flaky.

Fix in https://bugs.webkit.org/show_bug.cgi?id=221747

[youenn fablet]

Created attachment 420105 [details]
Patch

[EWS]

Committed r272802: <https://commits.webkit.org/r272802>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 420105 [details].