The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/bdadaba043088c87c8f6691c3fdd55dbf671777a

commit bdadaba043088c87c8f6691c3fdd55dbf671777a
Author: bsep <bsep@chromium.org>
Date: Thu Feb 09 20:43:00 2017

Disable layout test text-underline-position-cjk.html.

TBR=kojii@chromium.org
BUG= 690486 

Review-Url: https://codereview.chromium.org/2685783007
Cr-Commit-Position: refs/heads/master@{#449398}

[modify] https://crrev.com/bdadaba043088c87c8f6691c3fdd55dbf671777a/third_party/WebKit/LayoutTests/TestExpectations

Here's a couple example failure, for reference:
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.9/builds/42311
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Win7%20%28dbg%29/builds/8957

Removing from sheriff queue.

qyearsley@, could you advise?

I added this test in http://crrev.com/2647923002 , and ran rebaseline-cl to PS11. This caused Mac and Win7 failures, so I added [ NeedsRebaseline ] an drott@ recommended ask you what's happening.

The landed tests caused this failure on Mac10.9/Win7, so the problem happens not only with rebaseline-cl but also with rebaseline bots too.

The test relies on fonts on the system, so I think this is either some bots having different fonts and the test is flaky, or rebaseline-cl/bot failed to pick different result for Win7/Mac10.9.

I haven't yet looked into this, but I plan to this week; my initial suspicion is that it's flakiness due to some bots with the same OS having different fonts.

Quick question, what is PS11?

Probably PS11 = patch set 11.

Quick follow-up: Flakiness dashboard shows that this test is failing consistently on https://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=webkit_tests&tests=fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk.html

I haven't checked in depth, but since it's failing consistently, maybe the fonts on these systems were different for some time, but now they're consistent, and rebaselining would make the tests pass consistently again?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/18ffa64cff913d0b56d09276fd72257b74278f61

commit 18ffa64cff913d0b56d09276fd72257b74278f61
Author: qyearsley <qyearsley@chromium.org>
Date: Sat Mar 11 22:48:24 2017

Rebaseline layout test text-underline-position-cjk.html

BUG= 690486 

Review-Url: https://codereview.chromium.org/2712443003
Cr-Commit-Position: refs/heads/master@{#456289}

[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/TestExpectations
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk.html
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/linux/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.png
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/linux/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.txt
[add] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/mac-mac10.9/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.png
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/mac/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.png
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/mac/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.txt
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/win/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.png
[modify] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/win/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.txt
[add] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/win7/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.png
[add] https://crrev.com/18ffa64cff913d0b56d09276fd72257b74278f61/third_party/WebKit/LayoutTests/platform/win7/fast/css3-text/css3-text-decoration/text-underline-position/text-underline-position-cjk-expected.txt

What we ended up doing here was to change the test itself to avoid using the font family (Yu Gothic) which had inconsistent results on Windows sometimes.