Issue #12376 has been merged into this issue.

cc @ricowind.

Did we file a chromium bug for this?
This is happening on windows and mac builders right? These run stable chrome and this seems like very relevant info to pass along? Do we have a reproduction that is easy to verify?

I spent about a half day investigating but could not come up with a good repro. From what I can tell:

• I could only get the issue to repro when running multiple unit tests, I assume there's something with the timing between two tests
• We also got network errors from data- URIs (!!!) (see https://code.google.com/p/dart/issues/detail?id=11836)

I was unable to repro on any setup other than via the Selenium test harness.

Issue #12592 has been merged into this issue.

Issue #11836 has been merged into this issue.

Wondering if Rico or Martin would have time to investigate this further- I think this comes down to Chrome behavior when running multiple tests at the same time, but I'm not sure what the best way is to simplify the environment for a Chrome repro.

Bug 11836 is probably the simplest repro code-wise as it has a data-uri failing with a network error.

Pete: Just for my education and to not spend unnecessary time on this
Did we still never see this when not running with selenium?

cc @mkustermann.

I could never repro outside of selenium, but not sure if it's related to selenium or just that selenium makes it easy to hit the timing issues.

For example, the test files are htmlIndividualConfig tests with multiple groups, so they execute in parallel.
  - If I comment out all of the groups except for the problematic one, then the bug does not repro.
  - If I leave the group with the error, and add one empty group, then the bug does repro.
  - If I add a delay before the test "test('delay', () => Future.delay(new Duration(seconds: 3)));" then it does not repro.
  - Manually running in a browser window never repros

I just tried making the simplified repro's that I was using when investigating earlier and I can no longer repro locally- could be that the issue has already been fixed in Chrome (29.0.1547.57)?

Thx Pete
That may very well be. Alternatively I hope we can retire selenium really soon!

Issue #11998 has been merged into this issue.

Issue #11884 has been merged into this issue.

Removed this from the M6 milestone.
Added this to the M7 milestone.

I saw this on Dartium Mac with canvasrenderingcontext2d_test/drawImage_video_element_dataUrl

Removed this from the M7 milestone.
Added this to the M8 milestone.

This may be related to the dartium import loading issues where there is an ongoing investigation.

I sent Martin a patch with the same workaround that we used for this, and it didn't help for the import loading issue. But I agree, that it sounds similar.

Thanks for reminding me, I was sitting at his computer when we tried it :-). I would not rule out that the workaround is just eliminating this particular issue for other timing related issues on this test, but I can't of course be sure. We did see xhr related flake on dartium lately as well

Moving to later as I don't think we have any concrete changes to the DOM classes for this and don't have bandwidth for investigation in M8.

Removed this from the M8 milestone.
Added this to the Later milestone.