All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=872599

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=2e92d072c88d55ded871e3b8a29169a999111d1f770e3da12411857e1c2df607


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

Android Nexus6 WebView Perf

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1196d402640000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/1196d402640000

Retire PaymentDetailsModifierData flag by ericwilligers@chromium.org
https://chromium.googlesource.com/chromium/src/+/af1cfb65aa121b157dfffd1575a36b7a1076e7c5
0.6151 → 0.6218 (+0.006763)

Remove lifecycle updates for preferred content size changes by pdr@chromium.org
https://chromium.googlesource.com/chromium/src/+/601fd07dc26269d097d6ef6bd90f1203d361a72c
0.62 → 0.5206 (-0.09939)

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

This is a real regression. Previously, we'd recompute preferred widths once-per-frame, but we now calculate it synchronously after every layout which can occur many times per frame. I thought clean layout implied clean preferred widths but this is not true because preferred widths are computed lazily and are typically not needed.

 Issue 873149  has been merged into this issue.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/168a9482640000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14df9088640000

📍 Job complete. See results below.
https://pinpoint-dot-chromeperf.appspot.com/job/168a9482640000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12d4786c640000

📍 Job complete. See results below.
https://pinpoint-dot-chromeperf.appspot.com/job/12d4786c640000

📍 Job complete. See results below.
https://pinpoint-dot-chromeperf.appspot.com/job/14df9088640000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/15d4786c640000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14895764640000

📍 Job complete. See results below.
https://pinpoint-dot-chromeperf.appspot.com/job/14895764640000

📍 Job complete. See results below.
https://pinpoint-dot-chromeperf.appspot.com/job/15d4786c640000

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/33cd2bccd7dbd33dfddbe3fb4f4c8a6b1a7e4617

commit 33cd2bccd7dbd33dfddbe3fb4f4c8a6b1a7e4617
Author: Philip Rogers <pdr@chromium.org>
Date: Mon Aug 13 21:29:29 2018

Defer preferred size updates to after commit

https://crrev.com/581383 removed a preferred size timer and started
synchronously updating preferred sizes in DidUpdateMainFrameLayout.
This caused a performance regression because, even though layout is
up-to-date, the preferred contents size may need to be re-computed.
Script can cause many layouts per frame which was a performance
regression because many multiple preferred contents size updates are now
occurring per frame.

This patch adds a bool for if a layout occurred which may have changed
the preferred content size, and defers the update until after the
commit.

Bug:  872599 
Change-Id: Idfe1ba523b1ecfdf8f959bafc2e20ef6a52487bb
Reviewed-on: https://chromium-review.googlesource.com/1172059
Commit-Queue: Philip Rogers <pdr@chromium.org>
Reviewed-by: vmpstr <vmpstr@chromium.org>
Reviewed-by: Bo <boliu@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#582708}
[modify] https://crrev.com/33cd2bccd7dbd33dfddbe3fb4f4c8a6b1a7e4617/android_webview/renderer/aw_render_view_ext.cc
[modify] https://crrev.com/33cd2bccd7dbd33dfddbe3fb4f4c8a6b1a7e4617/android_webview/renderer/aw_render_view_ext.h
[modify] https://crrev.com/33cd2bccd7dbd33dfddbe3fb4f4c8a6b1a7e4617/content/renderer/render_view_impl.cc
[modify] https://crrev.com/33cd2bccd7dbd33dfddbe3fb4f4c8a6b1a7e4617/content/renderer/render_view_impl.h