All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=875326

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=879f2e3df14fb8b4993d3aa1cdbe4a604a8f1ba0aa414bd595eb083dc0e05b06


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

Android Nexus5X WebView Perf

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/118ecdec640000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/118ecdec640000

[PE] Stop dirtying compositing inputs on any layout for the root PaintLayer. by chrishtr@chromium.org
https://chromium.googlesource.com/chromium/src/+/46520ae13478333fe882e79041b667e5a3112796
0.04595 → 0.1258 (+0.07989)

Revert "[ChromeDriver] Fix path for old Chromium snapshots" by martiniss@chromium.org
https://chromium.googlesource.com/chromium/src/+/f5e8bf6941f0fb31103ad926e86729d45b276e01
0.1258 → 0.1322 (+0.006324)

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

I'm assigning this to the owner of the CL I reverted. I don't think my revert should have made things slower, but apparently it did?? John, any ideas about what's happening here?

I think it was actually my patch that caused this regression.