All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=879058

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=c91e5755df9156edc95966f7af827da66b988a8e5f84c6e6879e79ac6642a4b1


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

android-nexus5x-perf

v8.browsing_mobile - Benchmark documentation link:
  None

system_health.memory_mobile - Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/11d7402e640000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/11d7402e640000

[Android][Performance] Prevent accessory from over-openening keyboard by fhorschig@chromium.org
https://chromium.googlesource.com/chromium/src/+/760cc2256b4d27f9a01c6102a3392f7840df7a4f
1.573e+07 → 1.706e+07 (+1.326e+06)

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

I'll have a look.

I cannot see an easy fix for this issue. This change is definitely needed:
 - it prevents that other performance tests fail (youtube pinch)
 - it ensures UI correctness in that it prevents the keyboard from popping up without good reason

Actually, it seems quite weird that a single condition seems to be responsible for this regression - the check for |accessory.isShown| is the only added call here. This is a read-only access to an attribute. The condition body is still exactly the same method.

ulan@, am I missing something?

 Issue 880907  has been merged into this issue.