AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
20,130 Commits 1 Branch 0 Tags 544 MiB
bfd5f171e6
Commit Graph

20130 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
reed
bfd5f171e6 Revert[2] of "add backdrop option to SaveLayerRec" 
Reverted because of picture/serialization failure

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1567063002

TBR=mtklein

Review URL: https://codereview.chromium.org/1567063002
 2016-01-07 11:28:08 -08:00
msarett
133eaaacdd Store ninepatch chunks in the png header in CodexTest 
We are making a change in Android to store ninepatch chunks in the
png header.  Thus, our ninepatch test should match the Android
framework's use.
https://googleplex-android-review.git.corp.google.com/#/c/832067/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1553303003

Review URL: https://codereview.chromium.org/1553303003
 2016-01-07 11:03:25 -08:00
fmalita
c7e211acd0 SkTreatAsSprite should take AA into account 
Currently we always call SkTreatAsSprite with 0 subpixel bits, which means
subpixel translations are ignored.  This is incorrect for the anti-aliased
case (drawSprite always pixel-snaps, so we lose edge AA).

The CL updates SkTreatAsSprite to take an SkPaint argument and use 8 subpixel
bits when AA is requested.

Also remove unused SkTreatAsSpriteFilter.

BUG=skia:4761
R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1566943002

Committed: https://skia.googlesource.com/skia/+/983dc2541a729609037a05eba731b3eb9788c517

Review URL: https://codereview.chromium.org/1566943002
 2016-01-07 10:34:46 -08:00
robertphillips
b7f4b8e94e Add sentinel to GrMemoryPool's block header 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1562813002

Review URL: https://codereview.chromium.org/1562813002
 2016-01-07 10:12:17 -08:00
reed
5423ee17ed Revert of SkTreatAsSprite should take AA into account (patchset #5 id:80001 of https://codereview.chromium.org/1566943002/ ) 
Reason for revert:
Need to use SkLeftShift since the arg could be negative

Original issue's description:
> SkTreatAsSprite should take AA into account
>
> Currently we always call SkTreatAsSprite with 0 subpixel bits, which means
> subpixel translations are ignored.  This is incorrect for the anti-aliased
> case (drawSprite always pixel-snaps, so we lose edge AA).
>
> The CL updates SkTreatAsSprite to take an SkPaint argument and use 8 subpixel
> bits when AA is requested.
>
> Also remove unused SkTreatAsSpriteFilter.
>
> BUG=skia:4761
> R=reed@google.com
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1566943002
>
> Committed: https://skia.googlesource.com/skia/+/983dc2541a729609037a05eba731b3eb9788c517

TBR=fmalita@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4761

Review URL: https://codereview.chromium.org/1569873003
 2016-01-07 09:15:20 -08:00
reed
0575cb2def Revert of add backdrop option to SaveLayerRec (patchset #14 id:260001 of https://codereview.chromium.org/1523053003/ ) 
Reason for revert:
serialized != direct, will investigate

Original issue's description:
> add backdrop option to SaveLayerRec
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1523053003
>
> Committed: https://skia.googlesource.com/skia/+/247415969a9a5ed6c83cc09395472416c4b7de7f

TBR=robertphillips@google.com,bsalomon@google.com,mtklein@google.com
# 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/1565203002
 2016-01-07 08:15:10 -08:00
fmalita
983dc2541a SkTreatAsSprite should take AA into account 
Currently we always call SkTreatAsSprite with 0 subpixel bits, which means
subpixel translations are ignored.  This is incorrect for the anti-aliased
case (drawSprite always pixel-snaps, so we lose edge AA).

The CL updates SkTreatAsSprite to take an SkPaint argument and use 8 subpixel
bits when AA is requested.

Also remove unused SkTreatAsSpriteFilter.

BUG=skia:4761
R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1566943002

Review URL: https://codereview.chromium.org/1566943002
 2016-01-07 08:02:26 -08:00
fmalita
c070d1514a Remove SK_SUPPORT_LEGACY_BITMAP_SAMPLER_BIAS 
No longer defined in Chromium.

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1568893002

Review URL: https://codereview.chromium.org/1568893002
 2016-01-07 07:58:10 -08:00
reed
247415969a add backdrop option to SaveLayerRec 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1523053003

Review URL: https://codereview.chromium.org/1523053003
 2016-01-07 07:44:35 -08:00
mtklein
5820fe846f DM: add a pixel check to serialize 
If we're running through serialize and then drawing into some raster target (e.g. serialize-8888 like we run on the bots), make sure the serialized version is identical to what we'd get if we just drew into the rest of the pipeline (i.e. 8888).

Start by blacklisting all current failures.  This at least prevents further regression.

BUG=skia:4095
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1561993003

Committed: https://skia.googlesource.com/skia/+/6eb4e36a93ea695e7adb771ea9ac3326680a8e98

CQ_EXTRA_TRYBOTS=client.skia:Test-Win8-MSVC-ShuttleB-CPU-AVX2-x86_64-Release-Trybot

Review URL: https://codereview.chromium.org/1561993003
 2016-01-07 07:38:29 -08:00
scroggo
38ce0a7fe9 Add missing include for sleep 
This fixes the AOSP mips build. Mirrors the same change needed in DM.cpp at crrev.com/1425843003

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1568883002

Review URL: https://codereview.chromium.org/1568883002
 2016-01-07 07:28:47 -08:00
djsollen
19aa2b0867 Fix crash when ETC1 codec is not supported. 
BUG=skia:4216
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1567893002

Review URL: https://codereview.chromium.org/1567893002
 2016-01-07 07:22:47 -08:00
iroth
76a1225431 Remove codecs and images from iOS Bazel build. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1561163004

Review URL: https://codereview.chromium.org/1561163004
 2016-01-07 07:11:39 -08:00
mtklein
95bc09b666 cmake: set SK_RELEASE to 0 if not a Release build 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1564853002
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot

Review URL: https://codereview.chromium.org/1564853002
 2016-01-07 06:56:12 -08:00
reed
ce2dfae8c5 remove dead code from SkColorMatrixFilter 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1568783002

Review URL: https://codereview.chromium.org/1568783002
 2016-01-07 06:23:48 -08:00
caryclark
60c0a76738 Revert of DM: add a pixel check to serialize (patchset #4 id:60001 of https://codereview.chromium.org/1561993003/ ) 
Reason for revert:
breaks windows bots; additional windows specific failures need to be blacklisted

Original issue's description:
> DM: add a pixel check to serialize
>
> If we're running through serialize and then drawing into some raster target (e.g. serialize-8888 like we run on the bots), make sure the serialized version is identical to what we'd get if we just drew into the rest of the pipeline (i.e. 8888).
>
> Start by blacklisting all current failures.  This at least prevents further regression.
>
> BUG=skia:4095
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1561993003
>
> Committed: https://skia.googlesource.com/skia/+/6eb4e36a93ea695e7adb771ea9ac3326680a8e98

TBR=reed@google.com,mtklein@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4095

Review URL: https://codereview.chromium.org/1563923003
 2016-01-07 05:12:08 -08:00
caryclark
fb82437173 Revert of Broke GrTessellatingPathRenderer's tessellator out into a separate file. (patchset #8 id:240001 of https://codereview.chromium.org/1557083002/ ) 
Reason for revert:
broke valgrind bot

Original issue's description:
> Broke GrTessellatingPathRenderer's tessellator out into a separate file.
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1557083002
>
> Committed: https://skia.googlesource.com/skia/+/8b05cb8a00bdb82e100f1ba74bf4de4a504cceea

TBR=bsalomon@google.com,joshualitt@chromium.org,senorblanco@chromium.org,ethannicholas@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1570503002
 2016-01-07 05:09:48 -08:00
kkinnunen
3405800d7a Make SkGLContext lifetime more well-defined 
Remove refcounting from SkGLContext.

SkGLContext is expected to behave like GrContextFactory would own
it, as implied by the GrContextFactory function.

If it is refcounted, this does not hold.

Also other use sites, such as in SkOSWindow_win (command buffer gl
object), confirm the behavior. The object is explicitly owned and
destroyed, not shared.

Also fixes potential crashes from using GL context of an abandoned
context.

Also fixes potential crashes in DM/nanobench, if the GrContext lives
longer than GLContext through internal refing of GrContext.

Moves the non-trivial implementations from GrContextFactory.h to
.cpp, just for consistency sake.

Changes pathops_unittest.gyp. The pathops_unittest uses
GrContextFactory, but did not link to its implementation. The reason
they worked was that the implementation used (constructors, destructors)
happened to be in the .h file.

This works towards being able to use command buffer and NVPR from
the SampleApp.

BUG=skia:2992
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1511773005

Committed: https://skia.googlesource.com/skia/+/830e012187f951d49d7e46e196ac8d1e653a25da

Review URL: https://codereview.chromium.org/1511773005
 2016-01-06 23:49:31 -08:00
mtklein
6eb4e36a93 DM: add a pixel check to serialize 
If we're running through serialize and then drawing into some raster target (e.g. serialize-8888 like we run on the bots), make sure the serialized version is identical to what we'd get if we just drew into the rest of the pipeline (i.e. 8888).

Start by blacklisting all current failures.  This at least prevents further regression.

BUG=skia:4095
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1561993003

Review URL: https://codereview.chromium.org/1561993003
 2016-01-06 14:20:30 -08:00
benjaminwagner
85c8d06540 For Google3, enable SK_SUPPORT_LEGACY_SAVEFLAGS; required by legacy webkit_headless. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1568633002

Review URL: https://codereview.chromium.org/1568633002
 2016-01-06 12:51:58 -08:00
ethannicholas
8b05cb8a00 Broke GrTessellatingPathRenderer's tessellator out into a separate file. 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1557083002

Review URL: https://codereview.chromium.org/1557083002
 2016-01-06 11:44:34 -08:00
djsollen
4ecb1ad0c1 Fix DM tests in the Android Framework 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1567693002

Review URL: https://codereview.chromium.org/1567693002
 2016-01-06 11:38:14 -08:00
scroggo
3e07d94c95 Reenable testing webp to PDF on Valgrind 
The errors were probably due to skbug.com/4038, which has been fixed.
Reenable the test.

BUG=skia:3505
BUG=skia:4038
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1535133002

Review URL: https://codereview.chromium.org/1535133002
 2016-01-06 11:28:26 -08:00
borenet
4894565528 Add Test-Ubuntu-GCC-ShuttleA-GPU-GTX660-x86_64-Release to the CQ 
BUG=skia:4758
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1564743003

Review URL: https://codereview.chromium.org/1564743003
 2016-01-06 10:54:50 -08:00
scroggo
3ab9f2ede4 Fix some SkCodecPrintfs 
Silence the warning about peeking. Until we implement peeking for the
stream often used, it will always print, cluttering the logs.

Add some more information to the warning about failing to rewind.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1564843002

Review URL: https://codereview.chromium.org/1564843002
 2016-01-06 09:53:34 -08:00
senorblanco
b50b97d70a Optimize SkTileImageFilter destination bitmap size. 
The destination bitmap size was not being clipped by the clip bounds,
so tiled rendering (ie., clipping to a small region and rendering
a SkTileImageFilter with a large dstRect) was much slower than
non-tiled rendering.

Correctness is covered by unit test ImageFilterDrawTiled, and
performance by TileImageFilterBench.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1563873002

Review URL: https://codereview.chromium.org/1563873002
 2016-01-06 09:46:24 -08:00
joshualitt
5b1dec7cde Add additional assert to GrSingleOwner 
TBR=bsalomon@google.com
TBR=mtklein@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1566703003

Review URL: https://codereview.chromium.org/1566703003
 2016-01-06 09:25:14 -08:00
halcanary
53b1c09c13 SkPDF: Change handling of large-numbers in matricies. 
re-do http://crrev.com/1438503002 right.

Also, add a unit test that asserts without this change.

BUG=skia:734
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1552263003

Review URL: https://codereview.chromium.org/1552263003
 2016-01-06 09:02:26 -08:00
bungeman
54e6308219 Add skstd::remove_pointer_t and use it. 
This replaces an existing incomplete implementation. This also makes
it easier to update the current use in the future when switching to
<type_traits>.

Review URL: https://codereview.chromium.org/1555153002
 2016-01-06 08:30:59 -08:00
caryclark
1e1e5094e0 add comment 
TBR=reed@google.com
NOTRY=true
BUG=skia:4757
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1566733002

Review URL: https://codereview.chromium.org/1566733002
 2016-01-06 08:27:44 -08:00
joshualitt
1de610a528 Create debug only SkSingleOwner 
This is so Gpu code can guard against improper multithreaded usage in
debug builds

TBR=bsalomon@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1555953004

Review URL: https://codereview.chromium.org/1555953004
 2016-01-06 08:26:09 -08:00
caryclark
e114dd6316 fix valgrind 
(A recently added fuzzer test broke the valgrind bot, intentionally,
to isolate an uninitialized memory bug.)

If the conic has ordered points, but cannot be broken into parts,
treat it as if it were monotonic.

R=reed@google.com
BUG=skia:4757
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1568513002

Review URL: https://codereview.chromium.org/1568513002
 2016-01-06 08:15:44 -08:00
robertphillips
19c6250cad Add debug sentinel to GrMemoryPool to check for memory stomping 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1553233006

Review URL: https://codereview.chromium.org/1553233006
 2016-01-06 07:04:46 -08:00
benjaminwagner
bc348f41a7 In Google3 build, use SK_SFNTLY_SUBSETTER define to enable/disable sfntly. 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1561753002

Review URL: https://codereview.chromium.org/1561753002
 2016-01-05 17:00:04 -08:00
bungeman
221524de3b Start using <type_traits> and <utility> (C++11). 
SkUtility.h and SkTLogic.h implement a number of type traits now
available through <type_traits> and <utility>. This removes SkUtility.h,
replacing it with <utility>, and moves a number of traits in
SkTLogic.h to use the std:: equivelents. This change only uses C++11
parts of the standard library; SkTLogic.h will continue to provide
C++14 and beyond for now in the skstd namespace.

The changes to SkTLogic.h are being done gradually so that safe changes
may be landed confidently, with more risky changes in the future.

Review URL: https://codereview.chromium.org/1561683002
 2016-01-05 14:59:40 -08:00
senorblanco
3c14681889 Add a bench to test SkTileImageFilter. 
Tests in three modes: no tiling, 32x32 tiled rendering, and 64x64 tiled
rendering.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1559403003

Review URL: https://codereview.chromium.org/1559403003
 2016-01-05 13:27:43 -08:00
caryclark
ba25ddb7f2 fix nexus player 
The path contains test has such large numbers that it is indeterminate
whether it will return true or false. Allow either result, ensuring
that the test itself doesn't crash.

TBR=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1565463002

Review URL: https://codereview.chromium.org/1565463002
 2016-01-05 12:36:09 -08:00
mtklein
226aee25f4 Also ignore sfntly for Google3/iOS. 
It's just complicating their build process for now.

Will follow up by removing sfntly dependency after this rolls.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1559393002

Review URL: https://codereview.chromium.org/1559393002
 2016-01-05 10:55:32 -08:00
jvanverth
17aa047066 More framework support for TransferBuffers 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534123003

Review URL: https://codereview.chromium.org/1534123003
 2016-01-05 10:41:27 -08:00
bsalomon
7928ef613f Move read pixels format info into ConifgInfo in GrGLCaps 
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1542813004

Review URL: https://codereview.chromium.org/1542813004
 2016-01-05 10:26:39 -08:00
reed
aeab8ea4d7 fix linear gradient assert, by explicitly clamping 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1556993003

Review URL: https://codereview.chromium.org/1556993003
 2016-01-05 10:01:38 -08:00
reed
05dd251e5e take gr-context parameter to refEncoded, indicating a desire for only gpu-specific formats 
Prime motivator:
- we always call refEncoded on the generator when trying to upload
   - we call it *before* we ask for raster or YUV
- for blink, this call can be very slow, as they have to cons-up their SkData the first time (and grab a mutex to do it)
- this parameter will indicate to them that we're only interested in gpu formats, which they will know if they have.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1556333004

Review URL: https://codereview.chromium.org/1556333004
 2016-01-05 09:16:19 -08:00
joshualitt
20ccd40de9 Add assert that memory pool is empty when GrTextBlobCache frees 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1555993005

Review URL: https://codereview.chromium.org/1555993005
 2016-01-05 08:56:56 -08:00
bsalomon
45eefcfec2 Add null check for bmp->tex result in SkImageFilter::filterInputGpu 
BUG=chromium:570479
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1555333002

Review URL: https://codereview.chromium.org/1555333002
 2016-01-05 08:39:28 -08:00
mtklein
014f06b726 CMake: generate SkUserConfig.h with configure_file 
1)  It's simpler.
  2)  It should cache correctly (and not trigger total rebuilds everytime we re-run CMake.)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1544493002
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot

Review URL: https://codereview.chromium.org/1544493002
 2016-01-05 08:32:32 -08:00
caryclark
3323b05be5 Add comment for using infinite check in quad root 
TBR=reed@google.com
NOTRY=true
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1556073004

Review URL: https://codereview.chromium.org/1556073004
 2016-01-05 08:20:48 -08:00
joshualitt
22c6f5c731 Return 'correct' size from GrMemoryPool 
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1552283002

Review URL: https://codereview.chromium.org/1552283002
 2016-01-05 08:07:18 -08:00
mtklein
e56c626a44 clean up dead x86 filter opts code 
This is dead after removing shadeSpan16().

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1553233004
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1553233004
 2016-01-05 07:12:50 -08:00
mtklein
cd50bcae10 Spin off more thread-safe work in DM. 
Even in GPU configs, encoding a .png and writing to disk is thread safe.
This CL moves that work to an SkTaskGroup, offloading it from the GPU
thread to a threadpool.

In my brief local testing, this makes DM run in about half the time
for GPU-bound work, e.g.

    $ out/Release/dm --src gm --config gpudft -w output
    Before: 43.79 real        37.01 user         4.28 sys
    After : 23.64 real        39.31 user         4.46 sys

This won't affect the bots much, as they already skip .png generation
when gold.skia.org already has a .png of an image with a given hash.
(Most of the time saved here is under WriteToDisk() after the
gUninterestingHashes check.)

CQ_EXTRA_TRYBOTS=client.skia:Test-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Release-Trybot

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1561553002

Review URL: https://codereview.chromium.org/1561553002
 2016-01-05 06:20:20 -08:00
caryclark
dbaec7323f give up on quad root if infinite 
SkPath::contains() uses SkFindUnitQuadRoots to see
which side of a curve a point is on. If the
intermediate exceeds a float, return no found root.

R=reed@google.com
BUG=572406
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1556353002

Review URL: https://codereview.chromium.org/1556353002
 2016-01-05 06:20:09 -08:00