rebaseline GM results on Android-IntelRhb and Valgrind bots which were not running for a while

rebaseline GM results on Android-IntelRhb and Valgrind bots which were not running for a while

Mostly tiny diffs due to roundrect implementation changes.  Also a couple of
GM tests that are known to have changed.

Visually reviewed by bsalomon and epoger

BUG=skia:1771
TBR=bsalomon

Committed: https://code.google.com/p/skia/source/detail?r=12023

[epoger, 2013-10-30 17:56:20]

Committed patchset #1 manually as r12023 (presubmit successful).

[epoger, 2013-10-30 17:56:44]

committed TBR

[fmalita_google_do_not_use, 2013-10-30 18:02:23]

On 2013/10/30 17:56:44, epoger wrote:
> committed TBR

Awesome, thanks for doing this! I was about to shoot an IntelRhb GM code yellow
email for these failures - they sneaked in while the bots were down for more
than a week, so the blame-list is useless :)

[epoger, 2013-10-30 18:05:20]

On 2013/10/30 18:02:23, fmalita wrote:
> On 2013/10/30 17:56:44, epoger wrote:
> > committed TBR
> 
> Awesome, thanks for doing this! I was about to shoot an IntelRhb GM code
yellow
> email for these failures - they sneaked in while the bots were down for more
> than a week, so the blame-list is useless :)

You bet.  I filed a bunch of BreakingTheBuildbot issues for some of these tests
earlier today (https://code.google.com/p/skia/issues/detail?id=1774 , 1775,
1776, ...) and then I thought "man, there's too many, I'd better just carve out
a bunch at once".

[bsalomon, 2013-10-31 15:39:14]

On 2013/10/30 18:05:20, epoger wrote:
> On 2013/10/30 18:02:23, fmalita wrote:
> > On 2013/10/30 17:56:44, epoger wrote:
> > > committed TBR
> > 
> > Awesome, thanks for doing this! I was about to shoot an IntelRhb GM code
> yellow
> > email for these failures - they sneaked in while the bots were down for more
> > than a week, so the blame-list is useless :)
> 
> You bet.  I filed a bunch of BreakingTheBuildbot issues for some of these
tests
> earlier today (https://code.google.com/p/skia/issues/detail?id=1774 , 1775,
> 1776, ...) and then I thought "man, there's too many, I'd better just carve
out
> a bunch at once".

lgtm