All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=747925

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=3f54c0029f653e056676d0084f209181095f36bc10208896516d94c4f4cad686


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

android-webview-nexus6

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8973189717859221040

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

Hi yosin@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : Yoshifumi Inoue
  Commit : a20419dc273b83cd6a3b304363d5b088abf87fa2
  Date   : Wed Jul 19 09:50:48 2017
  Subject: Get rid of redundant update of extent_ in VisibleSelection::AdjustSelectionToAvoidCrossingEditingBoundaries()

Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : memory.top_10_mobile
  Metric       : memory:webview:all_processes:reported_by_os:system_memory:private_dirty_size_avg/foreground/http_www_amazon_com_gp_aw_s_k_nexus
  Change       : 2.35% | 65454506.6667 -> 63915776.0

Revision             Result                  N
chromium@487676      65454507 +- 448111      6      good
chromium@487744      65205333 +- 555629      6      good
chromium@487778      65408768 +- 913656      6      good
chromium@487795      65227861 +- 689697      6      good
chromium@487803      64762965 +- 602061      6      good
chromium@487805      65061291 +- 683564      6      good
chromium@487806      64840789 +- 799269      6      good
chromium@487807      63675477 +- 778723      6      bad       <--
chromium@487811      63915776 +- 933722      6      bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests memory.top_10_mobile

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8973189717859221040


For feedback, file a bug with component Speed>Bisection

Since My patch just removes if-statement and assignment as below, I don't think
my patch increase memory usage.

  // Correct the extent if necessary.
  if (base_editable_ancestor !=
      LowestEditableAncestor(extent_.ComputeContainerNode()))
    extent_ = base_is_first_ ? end_ : start_;


[1] http://crrev.com/c/577340: Get rid of redundant update of extent_ in VisibleSelection::AdjustSelectionToAvoidCrossingEditingBoundaries()

In future if the bisect ends up pointing to the wrong CL you can assign back to the perf sheriff (the other person cc'ed on the bug) and put the status to Untriaged and we can investigate further :)

In this case it looks like a revert since the graphs recovered and the bisect must have just been noise so no further action here.