REGRESSION (r237735?): [ MacOS WK2 ] Layout Test http/tests/resourceLoadStatistics/remove-blocking-in-redirect.html is a flaky timeout

[Truitt Savell]

The following layout test is flaky on MacOS WK2

http/tests/resourceLoadStatistics/remove-blocking-in-redirect.html

Probable cause:

This test has been slightly flakey for a while, within the last few revisions it has become significantly flakey. r237735 is around where this increased flakiness began. 

Flakiness Dashboard:

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

[Truitt Savell]

looks like this issue came from https://trac.webkit.org/changeset/237735/webkit

[Chris Dumez]

http/tests/resourceLoadStatistics/remove-blocking-in-redirect.html is indeed impacted by r237735 as the test navigates from origin A to origin B and then back to A. We are now using the same process both both loads of origin A instead of 2 separate processes.

That said, my change only impacted the timing of things by reusing an existing process instead of spinning up a fresh new one. 

The test was already flaky so patches impacting timing can definitely make flaky tests more or less flaky. That said, the issue is not with r237735 but rather with the test being flaky in the first place.

[Radar WebKit Bug Importer]

<rdar://problem/45818606>

[John Wilander]

Created attachment 359094 [details]
Patch

[John Wilander]

Uploaded an attempt to address the flakiness.

[John Wilander]

mac-wk2 failing tests are unrelated.

[John Wilander]

Created attachment 359100 [details]
Patch for landing

[WebKit Commit Bot]

Comment on attachment 359100 [details]
Patch for landing

Clearing flags on attachment: 359100

Committed r239966: <https://trac.webkit.org/changeset/239966>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[John Wilander]

I just had a look at the flakiness dashboard and the test has improved significantly after the change.