Not sure why I'm cc'd on this bug. Maybe, you wanted to cc someone else?

Thanks for the report.

> This flakiness is much worse than the baseline flakiness of the non-virtual test.

Can you tell me what you mean by "much worse"? https://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=ontimeout-event-override-after-failure.html shows both are flaky.

I think you are right to question "much worse". Looking at it again, I don't think the data says either way. The recent failures could be a result of the underlying flakiness.

Thank you, then I will add a TestExpectation entry.

If the pass/fail result of the test is not important, can we just remove the test entirely?

Alternatively, the test depends on setTimeout which is known to be a flaky approach to testing. Can we adjust how the test works to reduce flakiness? Maybe we could increase the xhr timeout value of 100?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1a8d96b6418a9dda53ef147a95ef721ad5daaa6e

commit 1a8d96b6418a9dda53ef147a95ef721ad5daaa6e
Author: yhirano <yhirano@chromium.org>
Date: Tue Nov 15 09:13:18 2016

Add expectations on virtual/mojo-loading tests whose base tests have expectations

If a test X has a TestExpecation entry, virtual/mojo-loading/X should have an
entry as well.

BUG= 659917 ,  520194 , 518987, 446385,  665165 

Review-Url: https://codereview.chromium.org/2502913002
Cr-Commit-Position: refs/heads/master@{#432150}

[modify] https://crrev.com/1a8d96b6418a9dda53ef147a95ef721ad5daaa6e/third_party/WebKit/LayoutTests/TestExpectations

#5: Sounds reasonable, but it should be addressed in  issue 464736 .

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/613135a62748f6d2f875f11b9096c368a3873714

commit 613135a62748f6d2f875f11b9096c368a3873714
Author: pdr <pdr@chromium.org>
Date: Wed Nov 16 06:45:55 2016

Deflake ontimeout-event-override-after-failure by waiting longer for http

This test flakes when the http server does not respond quick enough. I
suspect this happens on bots that are busy running other tests in
parallel. I was able to simulate this flakiness by lowering the XHR
timeout to 13ms.

This patch increases the XHR timeout from 100ms to 180ms which should
give more time for the test http server to respond. This has no effect
on overall test time because the pass condition is behind a 200ms timer.

After this patch lands, we need to monitor the following:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fxmlhttprequest%2Fontimeout-event-override-after-failure.html
If the flakiness disappears, the lines in TestExpectations should be removed.

BUG= 464736 , 665165 

Review-Url: https://codereview.chromium.org/2504713003
Cr-Commit-Position: refs/heads/master@{#432409}

[modify] https://crrev.com/613135a62748f6d2f875f11b9096c368a3873714/third_party/WebKit/LayoutTests/http/tests/xmlhttprequest/ontimeout-event-override-after-failure.html

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ff4bdcb669f08b5119a63d8eb60c40bb5c4b89c4

commit ff4bdcb669f08b5119a63d8eb60c40bb5c4b89c4
Author: pdr <pdr@chromium.org>
Date: Thu Nov 24 21:43:44 2016

Rewrite ontimeout-event-override-after-failure for less flakes and better errors

This patch rewrites ontimeout-event-override-after-failure with these changes:
1) The test no longer depends on setTimeout racing the network to pass. Instead
of having an async setTimeout check for success, we rely on async callbacks in
xhr's onerror and ontimeout. This should make the test faster in the common case,
and still pass if the test server is slow.
2) The xhr timeout has been increased to 500ms. If this test is flaky due to the
test server not responding fast enough, this change should deflake it.
3) The two failure modes of this test now print different errors which should
help in debugging this if it is still flaky.
4) The test has been reformated to 80cols.

This test is flaky but only very rarely:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fxmlhttprequest%2Fontimeout-event-override-after-failure.html

A previous attempt at deflaking this test by increasing the timeout failed:
https://crrev.com/613135a62748f6d2f875f11b9096c368a3873714

After this patch lands, we need to monitor the following:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fxmlhttprequest%2Fontimeout-event-override-after-failure.html
If the flakiness disappears, the lines in TestExpectations should be removed.

BUG= 464736 , 665165 

Review-Url: https://codereview.chromium.org/2518383003
Cr-Commit-Position: refs/heads/master@{#434390}

[modify] https://crrev.com/ff4bdcb669f08b5119a63d8eb60c40bb5c4b89c4/third_party/WebKit/LayoutTests/http/tests/xmlhttprequest/ontimeout-event-override-after-failure.html

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f26356128e7e04ff3793f93ffc98a89e7d25b710

commit f26356128e7e04ff3793f93ffc98a89e7d25b710
Author: pdr <pdr@chromium.org>
Date: Wed Dec 07 06:57:08 2016

Mark ontimeout-event-override-after-failure.html as not flaky

This test is no longer flaky:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fxmlhttprequest%2Fontimeout-event-override-after-failure.html

BUG= 464736 , 665165 

Review-Url: https://codereview.chromium.org/2560723002
Cr-Commit-Position: refs/heads/master@{#436881}

[modify] https://crrev.com/f26356128e7e04ff3793f93ffc98a89e7d25b710/third_party/WebKit/LayoutTests/TestExpectations