Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8997544322276301152

=== Auto-CCing suspected CL author liyuqian@google.com ===

Hi liyuqian@google.com, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Turn on Analytic AA in Chrome
Author  : Yuqian Li
Commit description:
  
Update cc_unittests png files to reflect our changes.

We also downgraded some ref tests to pixel tests because
there are some (dR <= 1, dG <= 1, dB <= 1) pixel difference on the
circle boundaries. This seems to be reasonable as the test html usually
has more layers than the ref html. In the future, we may set a
configurable threshold for ref tests so we don't have to downgrade ref
tests for this reason.

BUG=

patch from issue 2407393003 at patchset 20001 (http://crrev.com/2407393003#ps20001)
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
R=ajuma@chromium.org, fmalita@chromium.org

Review URL: https://codereview.chromium.org/2458433002 .

Patch from Yuqian Li <liyuqian@google.com>.

Cr-Commit-Position: refs/heads/master@{#427919}
Commit  : a947a329b3030259fbd042458dd6f31b1fe79dcf
Date    : Thu Oct 27 02:24:39 2016


===== TESTED REVISIONS =====
Revision         Exit Code  Std Dev  N  Good?
chromium@427897  0          N/A      2  good
chromium@427910  0          N/A      2  good
chromium@427917  0          N/A      2  good
chromium@427918  0          N/A      2  good
chromium@427919  1          N/A      2  bad    <--
chromium@427920  1          N/A      2  bad
chromium@427923  1          N/A      2  bad

Bisect job ran on: mac_10_10_perf_bisect
Bug ID: 660394

Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests memory.long_running_idle_gmail_tbmv2
Test Metric: v8-gc-incremental-step_avg/v8-gc-incremental-step_avg
Relative Change: Zero to non-zero
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_10_10_perf_bisect/builds/2458
Job details: https://chromeperf.appspot.com/buildbucket_job_status/8997544322276301152


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5877486120861696

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

My CL has already been reverted by Alan (CC'ed). So it shouldn't be the reason of the performance regression?

This wasn't a regression, it was a test failure.

The revert is in the current run, so if it passes I'll mark as fixed.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8997534545311931296

Above bisect is on different test that started to fail at the same time as this did, and on linux started to pass when the suspected CL was reverted. long_running never failed on linux. This is to test if we can use this test to repro locally the failure.

After the revert rolled onto the bot, the test started passing. lowering priority and keeping open to track relanding. If relanding isn't going to happen feel free to close.

Is there a way to run this bot before relanding?  I don't see it in the list of trybots.

===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Turn on Analytic AA in Chrome
Author  : Yuqian Li
Commit description:
  
Update cc_unittests png files to reflect our changes.

We also downgraded some ref tests to pixel tests because
there are some (dR <= 1, dG <= 1, dB <= 1) pixel difference on the
circle boundaries. This seems to be reasonable as the test html usually
has more layers than the ref html. In the future, we may set a
configurable threshold for ref tests so we don't have to downgrade ref
tests for this reason.

BUG=

patch from issue 2407393003 at patchset 20001 (http://crrev.com/2407393003#ps20001)
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
R=ajuma@chromium.org, fmalita@chromium.org

Review URL: https://codereview.chromium.org/2458433002 .

Patch from Yuqian Li <liyuqian@google.com>.

Cr-Commit-Position: refs/heads/master@{#427919}
Commit  : a947a329b3030259fbd042458dd6f31b1fe79dcf
Date    : Thu Oct 27 02:24:39 2016


===== TESTED REVISIONS =====
Revision         Exit Code  Std Dev  N  Good?
chromium@427915  0          N/A      2  good
chromium@427918  0          N/A      2  good
chromium@427919  1          N/A      2  bad    <--
chromium@427920  1          N/A      2  bad
chromium@427925  1          N/A      2  bad

Bisect job ran on: linux_perf_bisect
Bug ID: 660394

Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests page_cycler_v2.typical_25
Test Metric: timeToFirstContentfulPaint_avg/timeToFirstContentfulPaint_avg
Relative Change: Zero to non-zero
Score: 99.9

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/linux_perf_bisect/builds/6806
Job details: https://chromeperf.appspot.com/buildbucket_job_status/8997534545311931296


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5787377706991616

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

The following revision refers to this bug:
  https://skia.googlesource.com/skia.git/+/a3316adf63ca168c9125df965cfa85bdd7b50a1c

commit a3316adf63ca168c9125df965cfa85bdd7b50a1c
Author: liyuqian <liyuqian@google.com>
Date: Sat Oct 29 00:16:53 2016

Prevent overflow by falling back to non-AA

This piece of code is directly copied from our old supersampling AA
code. I didn not preserve this code because it's not triggered by any Skia
tests. However, this will be critical to Chromium's tests as some
websites will generate huge paths!

I'm not so sure whether the long_running_idle_gmail_background_tbmv2 test
failed because of this. But I'm sure that www.nationalgeographic.com from
page_cycler_v2.typical_25 failed because of this.

BUG= chromium:660394 

TBR=reed@google.com,caryclark@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2461133002

Review-Url: https://codereview.chromium.org/2461133002

[modify] https://crrev.com/a3316adf63ca168c9125df965cfa85bdd7b50a1c/src/core/SkScan_AAAPath.cpp

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c80fe5eb22d56e95c196c00a5ad5fdebd52fe214

commit c80fe5eb22d56e95c196c00a5ad5fdebd52fe214
Author: skia-deps-roller <skia-deps-roller@chromium.org>
Date: Sat Oct 29 01:44:20 2016

Roll src/third_party/skia/ 1f05f44e8..a3316adf6 (1 commit).

https://chromium.googlesource.com/skia.git/+log/1f05f44e8fc0..a3316adf63ca

$ git log 1f05f44e8..a3316adf6 --date=short --no-merges --format='%ad %ae %s'
2016-10-28 liyuqian Prevent overflow by falling back to non-AA

BUG= 660394 

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
TBR=reed@google.com

Review-Url: https://codereview.chromium.org/2453683012
Cr-Commit-Position: refs/heads/master@{#428583}

[modify] https://crrev.com/c80fe5eb22d56e95c196c00a5ad5fdebd52fe214/DEPS