Bug 191783 – REGRESSION (r238294): TestWebKitAPI.WKNavigation.ProcessCrashDuringCallback failing on iOS

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Ryan Haddad]

The following API test is failing on iOS after https://trac.webkit.org/changeset/238294:

    TestWebKitAPI.WKNavigation.ProcessCrashDuringCallback
        
        /Volumes/Data/slave/ios-simulator-12-release/build/Tools/TestWebKitAPI/Tests/WebKitCocoa/WebContentProcessDidTerminate.mm:269
        Expected equality of these values:
          WKErrorWebContentProcessTerminated
            Which is: 2
          error.code
            Which is: 3
        
        
        /Volumes/Data/slave/ios-simulator-12-release/build/Tools/TestWebKitAPI/Tests/WebKitCocoa/WebContentProcessDidTerminate.mm:262
        Expected equality of these values:
          WKErrorWebContentProcessTerminated
            Which is: 2
          error.code
            Which is: 3
        
        
        /Volumes/Data/slave/ios-simulator-12-release/build/Tools/TestWebKitAPI/Tests/WebKitCocoa/WebContentProcessDidTerminate.mm:255
        Expected equality of these values:
          WKErrorWebContentProcessTerminated
            Which is: 2
          error.code
            Which is: 3
        
https://build.webkit.org/builders/Apple%20iOS%2012%20Simulator%20Release%20WK2%20%28Tests%29/builds/1015/steps/run-api-tests/logs/stdio

[Ryosuke Niwa]

Looking into this now.

[Ryosuke Niwa]

Hm... I can't reproduce this failure locally.

[Chris Dumez]

(In reply to Ryosuke Niwa from comment #2)
> Hm... I can't reproduce this failure locally.

Probably flaky?

[Ryosuke Niwa]

(In reply to Chris Dumez from comment #3)
> (In reply to Ryosuke Niwa from comment #2)
> > Hm... I can't reproduce this failure locally.
> 
> Probably flaky?

Possibly. I've ran it like 5 times and couldn't repro. I am using debug instead of release builds so I'm gonna try release now. Sigh... another 40min wait...

[Chris Dumez]

(In reply to Ryosuke Niwa from comment #4)
> (In reply to Chris Dumez from comment #3)
> > (In reply to Ryosuke Niwa from comment #2)
> > > Hm... I can't reproduce this failure locally.
> > 
> > Probably flaky?
> 
> Possibly. I've ran it like 5 times and couldn't repro. I am using debug
> instead of release builds so I'm gonna try release now. Sigh... another
> 40min wait...

Honestly, the test is calling [view _close] so getting a WKErrorWebViewInvalidated error instead of WKErrorWebContentProcessTerminated is not technically wrong.

I think we should allow both. It is not the first time this test has started failing with WKErrorWebViewInvalidated errors, see Bug 189148.

[Chris Dumez]

(In reply to Chris Dumez from comment #5)
> (In reply to Ryosuke Niwa from comment #4)
> > (In reply to Chris Dumez from comment #3)
> > > (In reply to Ryosuke Niwa from comment #2)
> > > > Hm... I can't reproduce this failure locally.
> > > 
> > > Probably flaky?
> > 
> > Possibly. I've ran it like 5 times and couldn't repro. I am using debug
> > instead of release builds so I'm gonna try release now. Sigh... another
> > 40min wait...
> 
> Honestly, the test is calling [view _close] so getting a
> WKErrorWebViewInvalidated error instead of
> WKErrorWebContentProcessTerminated is not technically wrong.
> 
> I think we should allow both. It is not the first time this test has started
> failing with WKErrorWebViewInvalidated errors, see Bug 189148.

I wrote the test and what it was really testing was not the error code but the fact that all callbacks get invalidated.

[Chris Dumez]

(In reply to Chris Dumez from comment #6)
> (In reply to Chris Dumez from comment #5)
> > (In reply to Ryosuke Niwa from comment #4)
> > > (In reply to Chris Dumez from comment #3)
> > > > (In reply to Ryosuke Niwa from comment #2)
> > > > > Hm... I can't reproduce this failure locally.
> > > > 
> > > > Probably flaky?
> > > 
> > > Possibly. I've ran it like 5 times and couldn't repro. I am using debug
> > > instead of release builds so I'm gonna try release now. Sigh... another
> > > 40min wait...
> > 
> > Honestly, the test is calling [view _close] so getting a
> > WKErrorWebViewInvalidated error instead of
> > WKErrorWebContentProcessTerminated is not technically wrong.
> > 
> > I think we should allow both. It is not the first time this test has started
> > failing with WKErrorWebViewInvalidated errors, see Bug 189148.
> 
> I wrote the test and what it was really testing was not the error code but
> the fact that all callbacks get invalidated.

Whether we get WKErrorWebContentProcessTerminated or WKErrorWebViewInvalidated error seems to be a matter of timing.

[Ryosuke Niwa]

Created attachment 355176 [details]
Allow both errors

[Chris Dumez]

Comment on attachment 355176 [details]
Allow both errors

r=me

[Ryosuke Niwa]

Comment on attachment 355176 [details]
Allow both errors

Clearing flags on attachment: 355176

Committed r238340: <https://trac.webkit.org/changeset/238340>

[Ryosuke Niwa]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/46146920>