Skia DEPS roll 3100->3124

Skia DEPS roll 3100->3124
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=120068

[epoger, 2012-01-31 20:12:57]

control group is https://chromiumcodereview.appspot.com/9317006

will request review once the trybots complete

[reed1, 2012-01-31 21:09:41]

local run of DRT on windows yields:

compositing/shadows/shadow-drawing.html
fast/box-shadow/box-shadow-clipped-slices.html
svg/css/group-with-shadow.svg

I think these three should be rebaselined, as the differences are in the low bit

[epoger, 2012-01-31 21:39:52]

I have re-launched win_layout_rel and linux trybots because of suspiciously
early errors.

Either later tonight, or first thing tomorrow, I will examine all the trybot
results and add test expectations as Mike has mentioned.

[bsalomon, 2012-01-31 22:04:14]

On 2012/01/31 21:39:52, epoger wrote:
> I have re-launched win_layout_rel and linux trybots because of suspiciously
> early errors.
> 
> Either later tonight, or first thing tomorrow, I will examine all the trybot
> results and add test expectations as Mike has mentioned.

The failure in linux_layout of canvas-modify-emptyPath.html should be fixed by
r3125.

[epoger, 2012-02-01 14:37:19]

Below is the summary of trybot results.  I think the only failures of concern
are:
  fast/box-shadow/box-shadow-clipped-slices.html : add to test_expectations and
rebaseline later
  fast/canvas/canvas-modify-emptyPath.html : roll to 3125 instead of 3124 ??

linux : green
linux_chromeos : 1 NEW UI_TEST FAILURE, but it's probably a flake:
  ChromeMainTest.SecondLaunchFromIncognitoWithNormalUrl
linux_layout : 1 NEW WEBKIT_TEST FAILURE AND 1 NEW WEBKIT_GPU_TEST FAILURE:
  fast/box-shadow/box-shadow-clipped-slices.html = IMAGE
  fast/canvas/canvas-modify-emptyPath.html = CRASH   (webkit_gpu_test)
linux_layout_rel : 1 NEW WEBKIT_TEST FAILURE:
  fast/box-shadow/box-shadow-clipped-slices.html = IMAGE
linux_shared : green

mac : green
mac_layout : 3 NEW WEBKIT_TEST FAILURES:
  compositing/iframes/scrolling-iframe.html = TEXT
  fast/box-shadow/box-shadow-clipped-slices.html = IMAGE
  animations/animation-iteration-event-destroy-renderer.html = TIMEOUT
mac_layout_rel : 1 NEW WEBKIT_TEST FAILURE:
  fast/box-shadow/box-shadow-clipped-slices.html = IMAGE

win : green
win_layout : 1 NEW WEBKIT_UNIT_TEST FAILURE AND 2 NEW WEBKIT_TEST FAILURES:
  CCLayerTreeHostTestAtomicCommit.runMultiThread did not complete
(webkit_unit_test; probably a flake)
  accessibility/aria-checkbox-sends-notification.html = TEXT (probably a flake)
  http/tests/css/link-css-disabled-value-with-slow-loading-sheet.html = TEXT
(probably a flake)
win_layout_rel : compile step timed out (as usual)

[epoger, 2012-02-01 14:51:58]

Patchset 2 rolls Skia to r3125 instead of 3124, which should fix
fast/canvas/canvas-modify-emptyPath.html .

We will not bother suppressing fast/box-shadow/box-shadow-clipped-slices.html ;
instead, I will watch the results in the waterfall and plan to rebaseline the
results for that test later today.

[reed1, 2012-02-01 15:03:31]

lgtm