Bug 172650 – Layout test workers/wasm-hashset.html is flaky

• Status: RESOLVED
• Resolution: CONFIGURATION CHANGED
• Component: New Bugs

[Matt Lewis]

workers/wasm-hashset.html is a flaky timeout. 
On the flakiness dashboard the test is timing out on both Mac wk1 debugs and the Sierra WK2 Release.
https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=workers%2Fwasm-hashset.html

currently the test is skipped on iOS and windows.

The test usually runs quickly with a 1-5 second run time but each time the test times out, it runs for 155 seconds.
The test was added with https://trac.webkit.org/changeset/215353/webkit
It was marked as slow in the same revision r215353.

https://build.webkit.org/results/Apple%20El%20Capitan%20Debug%20WK1%20(Tests)/r217490%20(1478)/results.html
https://build.webkit.org/builders/Apple%20El%20Capitan%20Debug%20WK1%20(Tests)/builds/1478

[Saam Barati]

(In reply to Matt Lewis from comment #0)
> workers/wasm-hashset.html is a flaky timeout. 
> On the flakiness dashboard the test is timing out on both Mac wk1 debugs and
> the Sierra WK2 Release.
> https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.
> html#showAllRuns=true&tests=workers%2Fwasm-hashset.html
> 
> currently the test is skipped on iOS and windows.
> 
> The test usually runs quickly with a 1-5 second run time but each time the
> test times out, it runs for 155 seconds.
> The test was added with https://trac.webkit.org/changeset/215353/webkit
> It was marked as slow in the same revision r215353.
> 
> https://build.webkit.org/results/
> Apple%20El%20Capitan%20Debug%20WK1%20(Tests)/r217490%20(1478)/results.html
> https://build.webkit.org/builders/
> Apple%20El%20Capitan%20Debug%20WK1%20(Tests)/builds/1478

Hmm, perhaps it's worth looking into this since the jump from 5 seconds to 155 seconds could implicate a deadlock somewhere. I'll look into it.

[Matt Lewis]

This test now intermittently crashes on High Sierra and Sierra release testers:

https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=workers%2Fwasm-hashset.html

https://build.webkit.org/builders/Apple%20High%20Sierra%20Release%20WK2%20(Tests)/builds/669
https://build.webkit.org/results/Apple%20High%20Sierra%20Release%20WK2%20(Tests)/r223914%20(669)/results.html

[JF Bastien]

Might be unrelated, but my latest checkout has a 100% reproducible debug assert about a refcount somewhere in initNormalWorld's callees. I'll try to bisect to see where that started, it might lead to a release crash in the future.

[JF Bastien]

(In reply to JF Bastien from comment #3)
> Might be unrelated, but my latest checkout has a 100% reproducible debug
> assert about a refcount somewhere in initNormalWorld's callees. I'll try to
> bisect to see where that started, it might lead to a release crash in the
> future.

I think my local build was messed up? Ignore me for now, will report back later...

[JF Bastien]

Did this stop flaking after Saam reverted the bmalloc patch? I don't seem to have access to https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=workers%2Fwasm-hashset.html

[Matt Lewis]

This has since stoped flaking according to the flakiness dashboard.

[Radar WebKit Bug Importer]

<rdar://problem/35568980>