AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
23,189 Commits 1 Branch 0 Tags 544 MiB
1ab9e59da0
Commit Graph

139 Commits

Author SHA1 Message Date
bsalomon
2259c5f650 Disable CopySurface test on Nexus 10 for flakiness 
BUG=skia:5509
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2169073003

Review-Url: https://codereview.chromium.org/2169073003
 2016-07-22 04:09:47 -07:00
msarett
f461e7c227 Try blacklisting RAW images on Nexus 9 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2141923006

Review-Url: https://codereview.chromium.org/2141923006
 2016-07-13 08:03:35 -07:00
benjaminwagner
c3ea34b1ae Blacklist fontscalerdistortable on Win8. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2109583003

Review-Url: https://codereview.chromium.org/2109583003
 2016-07-06 06:55:04 -07:00
bsalomon
ddbc6e5d27 Exempt GrShape test from running on Android 
BUG=skia:5438
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2074403003

Review-Url: https://codereview.chromium.org/2074403003
 2016-06-20 12:48:33 -07:00
bsalomon
ea406dde3c Add blurcircles to blacklist for macmini4.1 
BUG=skia:5414
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2053373002

Review-Url: https://codereview.chromium.org/2053373002
 2016-06-10 06:58:16 -07:00
mtklein
c9b4ccd771 Skip verylargebitmap serialization on Android too 
I'm trying to figure out why we've been skipping serialize-8888 for these.
They don't draw wrong.  Seems like they just eat up all memory while serializing,
maybe only in Debug mode.

This should fix the red Nexus9 bot.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2050423003

Review-Url: https://codereview.chromium.org/2050423003
 2016-06-09 18:57:21 -07:00
mtklein
96e4e1ca5a Add raw pixel serialization fallback for SkImages that cannot be encoded. 
This fixes serialize-8888 for 2 GMs on Mac that I'm now unblacklisting.
I think another was already fixed, and two more were Windows-only.

Seems safe to use encoded_size=1 as another sentinel here (like we already
use =0); I can't imagine any encoded image format that can encode an image
in a single byte.

I suspect this is the root of the referenced bug too,
but this is a good idea even if not.

BUG=chromium:601851
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2039813007

CQ_EXTRA_TRYBOTS=client.skia:Test-Win-MSVC-GCE-CPU-AVX2-x86_64-Release-Trybot,Test-Mac-Clang-MacMini6.2-CPU-AVX-x86_64-Release-Trybot;client.skia.android:Test-Android-GCC-Nexus5-CPU-NEON-Arm7-Release-Trybot,Test-Android-GCC-Nexus9-CPU-Denver-Arm64-Release-Trybot

Review-Url: https://codereview.chromium.org/2039813007
 2016-06-09 10:38:08 -07:00
scroggo
23b80c402a Revert of Blacklist HTC.dng on 32-bit Win CPU bots (patchset #1 id:1 of https://codereview.chromium.org/1995703002/ ) 
Reason for revert:
We think we have independently fixed this problem by using a smaller image for HTC.dng

Original issue's description:
> Blacklist HTC.dng on 32-bit Win CPU bots
>
> BUG=skia:5318
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1995703002
>
> Committed: https://skia.googlesource.com/skia/+/62efdbde9987df3422f677a1a7c8f355791c1ae4

TBR=msarett@google.com,borenet@google.com,borenet@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:5318

Review-Url: https://codereview.chromium.org/2004203002
 2016-05-23 14:35:52 -07:00
borenet
62efdbde99 Blacklist HTC.dng on 32-bit Win CPU bots 
BUG=skia:5318
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1995703002

Review-Url: https://codereview.chromium.org/1995703002
 2016-05-18 12:23:58 -07:00
borenet
ac26aa630f Skip composeshader_bitmap GM on 64-bit Win CPU Test bots 
Temporary measure to get the bots green while the real issue is sorted out.

BUG=skia:5315
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1989323002

Review-Url: https://codereview.chromium.org/1989323002
 2016-05-18 08:12:51 -07:00
scroggo
00e9452791 Run RAW images serially on Win8-MSVC-ShuttleB 
BUG=skia:5283
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1960153005

Review-Url: https://codereview.chromium.org/1960153005
 2016-05-10 06:11:20 -07:00
stephana
338eaea602 Disable ReadWriteAlpha on TSAN 
BUG=skia:
R=borenet
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1927573002

Review URL: https://codereview.chromium.org/1927573002
 2016-04-27 06:50:03 -07:00
borenet
903ec3ce4d Fix CommandBuffer bot's DM config 
BUG=skia:4957
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1880403002

Review URL: https://codereview.chromium.org/1880403002
 2016-04-13 06:00:42 -07:00
borenet
71e185c060 GYP_DEFINES and DM flags for Vulkan bot 
BUG=skia:5186
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1885563002

Review URL: https://codereview.chromium.org/1885563002
 2016-04-12 08:13:57 -07:00
mtklein
ad6a1901d1 nullptr -> "sans-serif" fixes bold in gm/colorwheel 
This was an odd case where 8888 drew wrong and serialize-8888 was correct.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1880653002

Review URL: https://codereview.chromium.org/1880653002
 2016-04-11 13:20:11 -07:00
borenet
48b88cc1f6 Fixes for CommandBuffer bot 
BUG=skia:4957
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1877883002

Review URL: https://codereview.chromium.org/1877883002
 2016-04-11 10:16:01 -07:00
mtklein
0576aa9c07 Serialize invNormRotation in SkLightingShader. 
This fixes a serialize-8888 / 8888 GM mismatch.

This requires a picture bump so we know not to bother trying to read the
invNormRotation vector from older .SKPs.   Any chance no one's ever made
one?

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1870443002

Just bumping the picture version.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/1870443002
 2016-04-08 12:19:50 -07:00
mtklein
11064dfeb4 Fix lazy coding in ViaSingletonPictures. 
I was using SkRect::MakeLargest() as bounds, which is sort of nutso,
as that clearly is way out of bounds for how big a picture can feasibly
be, i.e. something closer to SkIRect::MakeLargest().

This was causing spurious quick rejects in drawPatch().  I didn't really
look much deeper to figure out why.  It's easy enough to just feed it the
proper bounds of the entire content.

This means patch_primitive draws correctly in sp-8888 mode.
I also noticed the GM was too small... it clipped off most of its content.
So I've made it larger.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1865143002

Review URL: https://codereview.chromium.org/1865143002
 2016-04-06 15:01:57 -07:00
mtklein
48156ed412 path_stroke_with_zero_length: turn off drawFat() by default. 
The replay modes cannot readback, and so cannot draw the fat parts, which
means we've had to exempt them from checking that they draw identically.

I don't mind if we leave it alone, but I thought I'd check to see your opinion.

BUG=skia:4095
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1862113004

Review URL: https://codereview.chromium.org/1862113004
 2016-04-06 14:25:35 -07:00
mtklein
9e65f9399e alternate textblobgeometrychange fix... just make our own surface if needed 
This should decopple fixing the replay modes from revising the API.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1868553005

Review URL: https://codereview.chromium.org/1868553005
 2016-04-06 13:57:38 -07:00
mtklein
3d72c099c7 Remove conditional flavor text on blend GM. 
This was breaking replay modes for no good reason.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1864253002

Review URL: https://codereview.chromium.org/1864253002
 2016-04-06 12:42:49 -07:00
mtklein
083b869f50 Register SkLocalMatrixImageFilter for serialization. 
This fixes one of the zillions of serialize-8888 backlisted tests.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1868453003

Review URL: https://codereview.chromium.org/1868453003
 2016-04-06 12:27:07 -07:00
bsalomon
3ddf967ffe Alter dm_flags.py logic flow for certain configs to be subtractive 
Replace msaa4 on Shield Tablet with glmsaa4.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1868473003

Review URL: https://codereview.chromium.org/1868473003
 2016-04-06 09:22:36 -07:00
bsalomon
b8797bb7b7 Add predfined configs for msaa4, nvpr4, nvpr4dit, and srgb that explictly use OpenGL rather than GLES and use them on the X1 bot 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1853103003

Committed: https://skia.googlesource.com/skia/+/74cb44dee50b5db94c1374cbcf0b0a1babd09d89

Review URL: https://codereview.chromium.org/1853103003
 2016-04-05 08:49:38 -07:00
bsalomon
3b597388b3 Revert of Add predfined configs for msaa4, nvpr4, nvpr4dit, and srgb that explictly use OpenGL rather than GL… (patchset #3 id:40001 of https://codereview.chromium.org/1853103003/ ) 
Reason for revert:
causing crashes on win8 bot?

Original issue's description:
> Add predfined configs for msaa4, nvpr4, nvpr4dit, and srgb that explictly use OpenGL rather than GLES and use them on the X1 bot
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1853103003
>
> Committed: https://skia.googlesource.com/skia/+/74cb44dee50b5db94c1374cbcf0b0a1babd09d89

TBR=egdaniel@google.com,jvanverth@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1857233002
 2016-04-05 07:26:29 -07:00
bsalomon
74cb44dee5 Add predfined configs for msaa4, nvpr4, nvpr4dit, and srgb that explictly use OpenGL rather than GLES and use them on the X1 bot 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1853103003

Review URL: https://codereview.chromium.org/1853103003
 2016-04-05 07:03:42 -07:00
halcanary
94b87bd404 DM/PDF: stop adding unused rasterizer keywords 
We can always add it back if we need it in the future.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1852713002

Review URL: https://codereview.chromium.org/1852713002
 2016-04-04 06:12:15 -07:00
brianosman
61d3b08138 Enable gpusrgb config on bots. 
Don't run the config if we can't get a context with srgb support.
Includes a unit test for that logic.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1846603002

Review URL: https://codereview.chromium.org/1846603002
 2016-03-30 11:19:36 -07:00
cdalton
c28afdb4ef Add test configs for nvpr with atlas text 
Adds nvpr4 and nvpr16, which use nvpr for paths and atlas for text.

Renames nvprmsaa4 and nvprmsaa16 to nvprdit4, nvprdit16. These use nvpr
for both paths and text.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1834693002

Review URL: https://codereview.chromium.org/1834693002
 2016-03-29 20:05:08 -07:00
brianosman
8c7d7ad59e Blacklist giant (4k) wbmp from GPU bots. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1827063004

Review URL: https://codereview.chromium.org/1827063004
 2016-03-24 10:33:42 -07:00
mtklein
bee39c5fbe fix --threads 4, blacklist interlaced1,2,3 on GPU 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1827543002

Review URL: https://codereview.chromium.org/1827543002
 2016-03-22 12:36:01 -07:00
mtklein
d33fe1f1f9 Try throttling x86 desktop bots down with fewer threads. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1825163002

Review URL: https://codereview.chromium.org/1825163002
 2016-03-22 11:08:24 -07:00
msarett
6c962f550f Blacklist RAW images on x86 Ubuntu 
BUG=skia:5115
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1822033002

Review URL: https://codereview.chromium.org/1822033002
 2016-03-22 09:01:20 -07:00
msarett
fc0b6d1053 Add SkImageGeneratorWIC 
This will be a replacement for SkImageDecoder_WIC.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1785613010
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac-Clang-x86_64-Release-CMake-Trybot

Review URL: https://codereview.chromium.org/1785613010
 2016-03-17 13:50:18 -07:00
msarett
5a6098217c Disable bad gif decode on CG 
This will save me the trouble of triaging this bad image over and
over again :).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1811943002

Review URL: https://codereview.chromium.org/1811943002
 2016-03-17 10:45:41 -07:00
msarett
1897631ebe Add an SkImageGeneratorCG 
This will serve as a replacement for SkImageDecoder_CG.

BUG=skia:4914
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718273004

Review URL: https://codereview.chromium.org/1718273004
 2016-03-09 14:20:58 -08:00
mtklein
4675220000 Archive SkRemote-related code. I want to start fresh. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1774013002

Review URL: https://codereview.chromium.org/1774013002
 2016-03-07 17:20:26 -08:00
mtklein
ed854fb897 Disable ignored / moot configs in DM. 
- upright-matrix-* are ignored in Gold.  Let's stop wasting time.

  - twice-8888 is silly.  ViaTwice was added to help locally test caching / batching
    in the GPU backend.  Running it into a software backend is not testing anything
    interesting, and doing that on the bots is doubly uninteresting.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1756333002

Review URL: https://codereview.chromium.org/1756333002
 2016-03-02 17:15:15 -08:00
mtklein
7fe18a8211 turn on f16 and sRGB 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1748113002

Review URL: https://codereview.chromium.org/1748113002
 2016-03-01 12:17:33 -08:00
mtklein
dd54e5d557 Revert of DM: disable ignored and very niche configs on bots. (patchset #1 id:1 of https://codereview.chromium.org/1745073002/ ) 
Reason for revert:
remote-8888 just caught a bug.  I am disproved.

Original issue's description:
> DM: disable ignored and very niche configs on bots.
>
> These are all probably just wasting time:
>   - upright-matrix-{8888,gpu} and remote[_cache] are ignored in Gold
>   - sp and 2ndpic test very niche areas of SkPicture
>     that are not under active change
>   - twice-8888 was never useful... the twice- via was meant for local
>     testing of GPU batching / caching.  Drawing into 8888 was moot.
>
> These configs are all useful to keep around for local testing, but probably are
> not worth the extra runtime or especially burden to have to think about them.
>
> If we should happen to stumble into bugs these configs would have caught,
> let's add them back with a note about those bugs to justify their need to run.
> I'm not convinced any exist today.
>
> This cuts ~30s off the CQ.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1745073002
>
> Committed: https://skia.googlesource.com/skia/+/7b6d3a30cf18377935c800603fc00fe0966d15de

TBR=halcanary@google.com,mtklein@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1749583003
 2016-02-29 09:51:09 -08:00
mtklein
7b6d3a30cf DM: disable ignored and very niche configs on bots. 
These are all probably just wasting time:
  - upright-matrix-{8888,gpu} and remote[_cache] are ignored in Gold
  - sp and 2ndpic test very niche areas of SkPicture
    that are not under active change
  - twice-8888 was never useful... the twice- via was meant for local
    testing of GPU batching / caching.  Drawing into 8888 was moot.

These configs are all useful to keep around for local testing, but probably are
not worth the extra runtime or especially burden to have to think about them.

If we should happen to stumble into bugs these configs would have caught,
let's add them back with a note about those bugs to justify their need to run.
I'm not convinced any exist today.

This cuts ~30s off the CQ.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1745073002

Review URL: https://codereview.chromium.org/1745073002
 2016-02-29 08:52:27 -08:00
yujieqin
f236ee4e44 Enable RAW codec for Windows 
* Fix the exception catching
* Set preprocessor differently for MSVC

BUG=skia:4889(b/26958348)
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1738913002

Committed: https://skia.googlesource.com/skia/+/474e4c3dd28b67f590851321f15d9983ef7fd031

Review URL: https://codereview.chromium.org/1738913002
 2016-02-29 07:14:42 -08:00
mtklein
d9381acb0e Turn off DM --pre_log on the bots. Too spammy. 
BUG=skia:4978
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1734703002

Review URL: https://codereview.chromium.org/1734703002
 2016-02-24 08:58:27 -08:00
mtklein
cbf897802b DM: remove unnecessary use of std::function 
This draw_to_canvas() function doesn't need the power of std::function.

This skips a copy or two, which is nice, and seems to clear
up most of my MSAN problems.  Why?  I do not know.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718633002

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-MSAN-Trybot

Review URL: https://codereview.chromium.org/1718633002
 2016-02-19 14:27:14 -08:00
mtklein
09d06850fa add back some basic configs to MSAN bot 
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-MSAN-Trybot

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1706373004

Review URL: https://codereview.chromium.org/1706373004
 2016-02-19 09:08:50 -08:00
mtklein
517558ad19 MSAN: cut down to green 
I have no idea what's going on with the current bot failures.
This disables lots of things to get it green.
Probably more than necessary.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1707873002

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-MSAN-Trybot

Review URL: https://codereview.chromium.org/1707873002
 2016-02-18 11:42:00 -08:00
mtklein
0da8537555 MSAN: build libc++ with MSAN, link to it. 
The errors that we see should now change to the unexplainable,
rather than the explainable errors at head.

We suspect that the original reason to blacklist was this,
not building libc++ with MSAN.  So unblacklist everything to start fresh.

BUG=skia:4903,skia:4900

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1694353004

Review URL: https://codereview.chromium.org/1694353004
 2016-02-16 14:03:15 -08:00
scroggo
3ac66e98f8 Optionally run RAW images serially 
RAW images use a lot of memory. Add a new FLAG to run one at a time so
we have less risk of running out of memory.

Isolate RAW images to their own thread on particular devices where our
images cause OOM errors.

Locally, this drops the max memory use from 3945 MB to 1664 MB (running only --image --images <RAW images we test>)

BUG=skia:4912
BUG=skia:4878
BUG=b/27035849
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1681553003

Review URL: https://codereview.chromium.org/1681553003
 2016-02-08 15:09:48 -08:00
msarett
772cc404e6 Blacklist RAW images on GPU tests 
BUG=skia:4888
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1677133002

Review URL: https://codereview.chromium.org/1677133002
 2016-02-08 11:43:50 -08:00
msarett
13a036b7eb Test CodecImageGenerator on GPU bots 
In crrev.com/1549473003, CodecImageGenerator implemented getYUV8Planes,
so that we can test on a GPU bot. Update the arguments to DM so that
we run CodecImageGenerator on GPU bots.

This is adapted from:
https://codereview.chromium.org/1663453002/

This will cause many untriaged images on Gold.
The GPU converts YUV->RGBA differently than we
do in software.

BUG=skia:4888
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1676663002

Review URL: https://codereview.chromium.org/1676663002
 2016-02-08 09:10:47 -08:00