Bug 181482 – REGRESSION: Layout Test http/tests/resourceLoadStatistics/grandfathering.html is a flaky timeout

• Component: New Bugs
• Status: RESOLVED FIXED

[Ryan Haddad]

The following layout test is flaky on  macOS WK2

http/tests/resourceLoadStatistics/grandfathering.html

Probable cause:

The first macOS WK2 timeout visible on the dashboard blames this range: https://trac.webkit.org/log/webkit/?action=stop_on_copy&mode=stop_on_copy&rev=226543&stop_rev=226541&limit=100&verbose=on

A few commits before that, I skipped the test listed right above this one in TestExpectations: https://trac.webkit.org/changeset/226538/webkit


Flakiness Dashboard:

https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=http%2Ftests%2FresourceLoadStatistics%2Fgrandfathering.html

[Matt Lewis]

Previously the test expectations were [ Pass Failure Timeout ] and were changed with many others in https://trac.webkit.org/changeset/225640/webkit however according to the dashboard as Ryan said, the test seems to have recently become a flaky timeout again.

[Matt Lewis]

I wonder if another resourceLoadStatistics test was timing out for a similar or an underlying issue in the bug related via see also. 

diff for this test:
--- /Volumes/Data/slave/highsierra-release-tests-wk2/build/layout-test-results/http/tests/resourceLoadStatistics/grandfathering-expected.txt
+++ /Volumes/Data/slave/highsierra-release-tests-wk2/build/layout-test-results/http/tests/resourceLoadStatistics/grandfathering-actual.txt
@@ -1,2 +1,5 @@
-PASS Grandfathered cookie was not purged.
+#PID UNRESPONSIVE - com.apple.WebKit.WebContent.Development (pid 1666)
+FAIL: Timed out waiting for notifyDone to be called
 
+#EOF
+#EOF

[Matt Lewis]

Skipped in https://trac.webkit.org/changeset/226896/webkit

[Ryan Haddad]

(In reply to Matt Lewis from comment #3)
> Skipped in https://trac.webkit.org/changeset/226896/webkit

http/tests/resourceLoadStatistics/non-prevalent-resource-with-user-interaction.html started timing out after this, so something weird is going on:

https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=http%2Ftests%2FresourceLoadStatistics%2Fnon-prevalent-resource-with-user-interaction.html

[Radar WebKit Bug Importer]

<rdar://problem/36549026>

[John Wilander]

Hopefully resolved in https://trac.webkit.org/changeset/227223, https://bugs.webkit.org/show_bug.cgi?id=181822.

[Ryan Haddad]

Still happening.

[Matt Lewis]

Skipped the test again in https://trac.webkit.org/changeset/227368/webkit

Seeing if this is just now only the test or if it is still the same issue as before.

[Ryan Haddad]

(In reply to Matt Lewis from comment #8)
> Skipped the test again in https://trac.webkit.org/changeset/227368/webkit
> 
> Seeing if this is just now only the test or if it is still the same issue as
> before.

http/tests/resourceLoadStatistics/non-prevalent-resource-with-user-interaction.html is timing out frequently, so this requires more investigation.

John, can you take a look?

[John Wilander]

Created attachment 332581 [details]
Patch for landing

[John Wilander]

Here's to hoping. 100 iterations on my laptop without timeout.

[John Wilander]

Committed r227761: <https://trac.webkit.org/changeset/227761>