The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/cbf4515e37ae65e2b73088888e826311a4d20787

commit cbf4515e37ae65e2b73088888e826311a4d20787
Author: engedy <engedy@chromium.org>
Date: Fri Jun 02 20:57:14 2017

Mark certain layout tests slow on MSAN bots.

external/wpt/encoding/textdecoder-fatal-single-byte.html
fast/css/css-selector-deeply-nested.html
http/tests/inspector/forced-layout-in-microtask.html
http/tests/inspector/tracing/timeline-xhr-response-type-blob-event.html
inspector/components/file-path-scoring.html
inspector/elements/styles-4/styles-should-not-force-sync-style-recalc.html
webaudio/mixing.html

BUG=729136
TBR=yoichio@chromium.org

Review-Url: https://codereview.chromium.org/2920703005
Cr-Commit-Position: refs/heads/master@{#476782}

[modify] https://crrev.com/cbf4515e37ae65e2b73088888e826311a4d20787/third_party/WebKit/LayoutTests/MSANExpectations

These tests are still timing out (e.g., https://luci-logdog.appspot.com/v/?s=chromium%2Fbb%2Fchromium.webkit%2FWebKit_Linux_Trusty_MSAN%2F1265%2F%2B%2Frecipes%2Fsteps%2Fwebkit_tests%2F0%2Fstdout). I'm not familiar with layout tests infra, but I see on that log the warning:

"SLOW tests should only be added to SlowTests and not to TestExpectations. fast/css/css-selector-deeply-nested.html"

It's not clear to me whether the CL in c#1 actually had the intended effect or not (I looked through the test setup code a bit start from the warning above, but couldn't ferret out what was actually done with the "Slow" annotation).

Strange, I'm not sure why the change didn't work.

qyearsley, can you look at this?

As far as I understand:

 - The expected behavior is that [ Slow ] expectations in MSANExpectations or ASANExpectations etc. should apply on the relevant bots
 - The actual behavior appears to be that a warning is printed and the slow expectations appear not to apply. For example, inspector/layers/layer-canvas-log.html is supposed be marked Slow in ASANExpectations but https://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=webkit_tests&tests=inspector/layers/layer-canvas-log.html seems to indicate that it (maybe) doesn't actually apply?

The possible workaround, of course, is to put all slow expectations in SlowTests, even if the test only times out on some bots. Although maybe it's useful to be able to say that a certain test is "slow" on MSAN builds but is OK on other builds.

Fixing this will probably involving changing this warning:
https://cs.chromium.org/chromium/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/models/test_expectations.py?l=426

But first it would be helpful to verify whether slow expectations apply or not in MSANExpectations etc.

Removing Blink>CSS component as this seems to be an infra issue.

@qyearsley - can you please disable the tests "properly" in the meantime, if the fix isn't obvious? We shouldn't leave the build failing indefinitely.

Marking them as Timeout on MSAN:

https://chromium-review.googlesource.com/c/525433/

Hopefully this will fix the failing bot while the real fix is worked out.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f82f591dfdcf0cc9a69314ee7702dca8998576fa

commit f82f591dfdcf0cc9a69314ee7702dca8998576fa
Author: Colin Blundell <blundell@chromium.org>
Date: Tue Jun 06 09:46:47 2017

Mark several layout tests as timing out on MSAN

These tests seem to just be now slow enough on MSAN to be "slow tests"
(i.e., they don't actually hang), but it is not clear how to mark them
as slow to the testing infrastructure (see discussion on bug).

Marking them as "Timeout" while the bug is resolved.

TBR=qyearsley

Bug: 729136
Change-Id: I991645790d98318d24f7bd65681e244266eae373
Reviewed-on: https://chromium-review.googlesource.com/525433
Reviewed-by: Colin Blundell <blundell@chromium.org>
Commit-Queue: Colin Blundell <blundell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#477246}
[modify] https://crrev.com/f82f591dfdcf0cc9a69314ee7702dca8998576fa/third_party/WebKit/LayoutTests/MSANExpectations

Thanks Colin for making the above CL -- when I was first looking at the bug I missed the fact that the builders were still failing -- partly because results for MSAN aren't showing up on the flakiness dashboard. (Filed bug 730048 for this.)

inspector/sources/debugger-pause/pause-on-elements-panel.html failed in https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20MSAN/builds/1271

Should we add it to the list of MSANExpectations or is there a broader issue to solve?

I think the failure in #c12 is a flake (it doesn't look like it failed in the prior run, but it did fail once on 5/26).

More thoughts on this:

I think that the current behavior makes sense -- all slow expectations can go in SlowTests; if a test is slow in general it can be marked as slow. If it's too slow on the MSAN bots it could be marked as [ Skip ] or [ Timeout Pass ] in MSANExpectations with a note that it's a slow test.

If this sounds OK, then I think the next thing to do here is just to update the comments in MSANExpectations.

If the test is normally fast but particularly slow under MSAN (i.e., much slower than the normal amount of slowness), then I agree that have a slow entry *particularly* in the MSAN file makes sense.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b2d611a67d4c63138f25197d2d9eed777c8f59a8

commit b2d611a67d4c63138f25197d2d9eed777c8f59a8
Author: Robert Ma <robertma@chromium.org>
Date: Wed Dec 13 13:37:05 2017

Mark parsing.any.html as TIMEOUT only on MSAN bots

For some reason, the two tests are at least a magnitude slower on MSAN
bots than other bots (60+ secs vs. ~10 secs). Narrow down the
expectations to reflect so.

Bug:  793792 , 729136
Change-Id: Ie46aed898996fe07cc7621336ffc5fc699c024ba
Reviewed-on: https://chromium-review.googlesource.com/819596
Commit-Queue: Philip Jägenstedt <foolip@chromium.org>
Reviewed-by: Quinten Yearsley <qyearsley@chromium.org>
Reviewed-by: Philip Jägenstedt <foolip@chromium.org>
Cr-Commit-Position: refs/heads/master@{#523762}
[modify] https://crrev.com/b2d611a67d4c63138f25197d2d9eed777c8f59a8/third_party/WebKit/LayoutTests/MSANExpectations
[modify] https://crrev.com/b2d611a67d4c63138f25197d2d9eed777c8f59a8/third_party/WebKit/LayoutTests/TestExpectations

 Issue 793792  has been merged into this issue.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot