All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=815146

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=218e9dc7001672ba42e41c165c739815b0eb6fb0f7d4b29079757e15b0955962


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

android-webview-nexus6

📍 Pinpoint job started.
https://chromeperf.appspot.com/job/1796e50f840000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12c98a48440000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/11aa5831440000

Trying another bisect, expanding the range.

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/11aa5831440000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/13d0ee23440000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/13d0ee23440000

[notification_helper.exe] Add installation and uninstallation support by chengx@chromium.org
https://chromium.googlesource.com/chromium/src/+/7fe85d7b9f83a059d57f2b69633812ba27caa54f

Avoid creating AXObject at unsafe times by dmazzoni@chromium.org
https://chromium.googlesource.com/chromium/src/+/eebef6ebc57245087743bcdc5b99dfa106843109

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

My change would only affect the accessibility perf tests, if anything.

If you're seeing a change for any other test it doesn't make sense that it's my change.

@chengx?

Seems that the regression is on android. My changes is Windows desktop only.

I'm not sure I trust pinpoint's results. It looks to me like it was unable to compile/test between 537796 and 537961. The problem could be anywhere in that range.

Simon, is this due to the perf builder renaming breakage? Anything we can do to bisect?

Yes, I believe these were from the builder migration, I'll try re-kicking this since they should build now.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/17beb7fcc40000

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/17beb7fcc40000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1485b47ac40000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/1485b47ac40000

WebRequest+NetworkService: Declarative APIs, response redirects by rockot@chromium.org
https://chromium.googlesource.com/chromium/src/+/0b74695cb7788315744b0ca5a91c88e54a575930

WebView: enable cc::switches::kCheckDamageEarly by jamwalla@chromium.org
https://chromium.googlesource.com/chromium/src/+/05381aa3a9570439a16774f120bc3c0d20c9de81

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Looks like both the CLs in #17 actually did increase memory. Assigning to jamwalla because the regression at "WebView: enable cc::switches::kCheckDamageEarly" is a little larger.

this is the same java heap one caused by begin frame loop that's already been fixed. jamwalla can merge this to the other bug