Bug 230847 – REGRESSION (r283102): [ MacOS Debug ] TestWebKitAPI.WebKit2.CaptureIndicatorDelay is failing

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

TestWebKitAPI.WebKit2.CaptureIndicatorDelay 

Is failing on macOS Debug.

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

TestWebKitAPI.WebKit2.CaptureIndicatorDelay
        2021-09-27 08:06:21.207 TestWebKitAPI[35291:3521683] Encountered error: Error Domain=WKErrorDomain Code=4 "A JavaScript exception occurred" UserInfo={WKJavaScriptExceptionLineNumber=22, WKJavaScriptExceptionMessage=TypeError: null is not an object (evaluating 'stream.getTracks'), WKJavaScriptExceptionColumnNumber=23, WKJavaScriptExceptionSourceURL=file:///Volumes/Data/worker/bigsur-debug-tests-wk2/build/WebKitBuild/Debug/TestWebKitAPI.resources/getUserMedia.html, NSLocalizedDescription=A JavaScript exception occurred} while evaluating script: stop()
        2021-09-27 08:06:32.214 TestWebKitAPI[35291:3521683] Expected state _WKMediaCaptureStateDeprecatedNone, but after 10 seconds state is _WKMediaCaptureStateDeprecatedActiveCamera 
        LEAK: 1 WebProcessPool
        LEAK: 1 WebPageProxy
        
        /Volumes/Data/worker/bigsur-debug/build/Tools/TestWebKitAPI/cocoa/TestWKWebView.mm:211
        Value of: !error
          Actual: false
        Expected: true
        
        
        /Volumes/Data/worker/bigsur-debug/build/Tools/TestWebKitAPI/Tests/WebKit/GetUserMedia.mm:405
        Value of: waitUntilCaptureState(webView.get(), _WKMediaCaptureStateDeprecatedNone)
          Actual: false
        Expected: true

[Radar WebKit Bug Importer]

<rdar://problem/83577251>

According to the history, the test started failing at https://trac.webkit.org/changeset/283102/webkit

[Alexey Proskuryakov]

This change was reverted again.

*** This bug has been marked as a duplicate of bug 230854 ***

[Saam Barati]

Relanding unlinked baseline. Need to figure this out

[Saam Barati]

I think my patch just perturbed a pre-existing race.

[Saam Barati]

Created attachment 439421 [details]
patch

[youenn fablet]

Comment on attachment 439421 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=439421&action=review

> Tools/TestWebKitAPI/Tests/WebKit/GetUserMedia.mm:398
> +    [webView stringByEvaluatingJavaScript:@"setTimeout(stop, 0)"];

I think this might still be racy, at least in theory.
Let's make it bullet proof by waiting for getUserMedia promise resolution to actually call stop.

[youenn fablet]

Created attachment 439439 [details]
Patch

[youenn fablet]

Created attachment 439463 [details]
Patch

[youenn fablet]

Created attachment 439470 [details]
Patch

[EWS]

Committed r283172 (242220@main): <https://commits.webkit.org/242220@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 439470 [details].