All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=667890

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg68XwqQoM


Bot(s) for this bug's original alert(s):

android-one

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8995259250928540704

=== Auto-CCing suspected CL author kochi@chromium.org ===

Hi kochi@chromium.org, 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 : Parser support for >>> (shadow-piercing descendant) combinator.
Author  : kochi
Commit description:
  
'>>>' (shadow-piercing descendant) can only be parsed in
static profile. The parsing is guarded by a runtime enabled flag.

Matching part is done in the next CL:
https://codereview.chromium.org/2496123002/

BUG= 633007 

Review-Url: https://codereview.chromium.org/2500813003
Cr-Commit-Position: refs/heads/master@{#433133}
Commit  : 69635dc029d7ce77e7db12669a994295d5f93952
Date    : Fri Nov 18 06:40:50 2016


===== TESTED REVISIONS =====
Revision         Mean     Std Dev  N   Good?
chromium@433111  415.151  30.7911  25  good
chromium@433127  402.596  39.2616  25  good
chromium@433131  403.049  29.2489  25  good
chromium@433132  403.199  28.856   25  good
chromium@433133  364.828  28.4302  25  bad    <--
chromium@433135  371.461  28.9763  25  bad
chromium@433143  373.645  22.1005  25  bad

Bisect job ran on: android_one_perf_bisect
Bug ID: 667890

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.css
Test Metric: PseudoClassSelectors/PseudoClassSelectors
Relative Change: 10.00%

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_one_perf_bisect/builds/1822
Job details: https://chromeperf.appspot.com/buildbucket_job_status/8995259250928540704


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5305510880870400

| 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!

Taking a look.  Assigning Blink>DOM, as it affects querySelector
(might be CSS, but for now).

Analysis:

The affected benchmark is
third_party/WebKit/PerformanceTests/CSS/PseudoClassSelectors.html
https://cs.chromium.org/chromium/src/third_party/WebKit/PerformanceTests/CSS/PseudoClassSelectors.html?q=PseudoClassSelectors.html&sq=package:chromium&l=1

The test is basically a microbenchmark for running querySelector
for selectors containing pseudo classes (:first-child et al.).

My addition in the CL adds one new *combinator*, so it shouldn't affect
directly to the performance of matching *pseudo classes*.

I don't have reproduction locally on Linux (it needs Android one?).

It could have happened due to some code organization change and
cache hit/miss changes.

I'm not sure this happens on any other platforms, and if Android one
is the only platform that regressed for this micro benchmark,
I'd like to close this as WONTFIX.

Oops, forgot to close this.

cc-ing rune, blink_perf.css owner, as FYI