Manual Skia DEPS roll 5744bbe -> b328158

Manual Skia DEPS roll 5744bbe -> b328158

control (whitespace CL) is at https://codereview.chromium.org/305183005

BUG=skia:2610
NOTRY=True

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=274114

[epoger, 2014-05-30 22:20:17]

Looks mostly good so far, with a few seemingly new webkit_test failures:

mac_blink_dbg: 
  virtual/softwarecompositing/lots-of-img-layers-with-opacity.html
  virtual/softwarecompositing/lots-of-img-layers.html

mac_blink_rel:
  virtual/softwarecompositing/lots-of-img-layers-with-opacity.html
  virtual/softwarecompositing/lots-of-img-layers.html

win_blink_dbg: [still running]

win_blink_rel: 
  compositing/rtl/rtl-iframe-absolute-overflow-scrolled.html


Do these failures ring a bell for anyone?  (Ideas as to what in the roll would
cause them?)

Should I go ahead and land it, asking a webkit gardener to rebaseline these
results?

[reed1, 2014-05-30 23:01:37]

lots-of-img-layers.html -- seems tiny, and could be rebaselined
lots-of-img-layers-with-opacity.html -- seems huge, something definitely
changed.

I can't see the html from here, so I don't know what is happening in the opacity
one. Funny its only on Mac (true?). 

I'm home, or would try to repro locally on a desktop.

[epoger, 2014-05-31 12:27:07]

Full analysis, now that all the results are in:


These failures are ignorable:

android_dbg_triggered_tests: looks like infra failure, not code problem
linux_chromeos_asan: seems about the same as control CL
linux_layout_rel: seems about the same as control CL
mac: infra failure, not code problem
mac_asan: same as control CL
win: same as control CL


Here are the "real" failures, all in webkit_tests:

mac_blink_dbg: 
  virtual/softwarecompositing/lots-of-img-layers-with-opacity.html
  virtual/softwarecompositing/lots-of-img-layers.html

mac_blink_rel:
  virtual/softwarecompositing/lots-of-img-layers-with-opacity.html
  virtual/softwarecompositing/lots-of-img-layers.html

win_blink_dbg:
  virtual/threaded/animations/end-exclusive.html

win_blink_rel: 
  compositing/rtl/rtl-iframe-absolute-overflow-scrolled.html


Looking at
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#group=%40...

it seems to me that all of these tests EXCEPT for
virtual/threaded/animations/end-exclusive.html have already been failing or
flaky, so we can ignore them.

My inclination is to go ahead and land in spite of the single remaining failure
(virtual/threaded/animations/end-exclusive.html).  Given that it is the only
unexplained failure, I bet it's a flake.


If you agree, gimme an LG... otherwise, let me know what might be a better
course of action!

[epoger2 (DO NOT USE), 2014-06-01 12:15:39]

lgtm

[epoger2 (DO NOT USE), 2014-06-01 12:15:53]

The CQ bit was checked by epoger@chromium.org

[commit-bot: I haz the power, 2014-06-01 12:17:43]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/epoger@google.com/309503006/1

[commit-bot: I haz the power, 2014-06-01 12:26:29]

Change committed as 274114