platform/upstream/libSkiaSharp.git
10 years agoMIPS: added optimizations for functions from SkBitmapProcState
djordje.pesut [Tue, 8 Jul 2014 09:24:16 +0000 (02:24 -0700)]
MIPS: added optimizations for functions from SkBitmapProcState

gain is ~30%

following functions are optimized:
  SI8_D16_nofilter_DX
  SI8_opaque_D32_nofilter_DX

R=djsollen@google.com, teodora.petrovic@gmail.com

Author: djordje.pesut@imgtec.com

Review URL: https://codereview.chromium.org/336533003

10 years agoUpdate SKP version to 45
borenet [Tue, 8 Jul 2014 06:36:25 +0000 (23:36 -0700)]
Update SKP version to 45

Automatic commit by the RecreateSKPs bot.

TBR=

Author: borenet@google.com

Review URL: https://codereview.chromium.org/379523002

10 years agoFlesh out SkMatrix typemask suppressions.
mtklein [Tue, 8 Jul 2014 00:53:17 +0000 (17:53 -0700)]
Flesh out SkMatrix typemask suppressions.

BUG=skia:
R=mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/374763002

10 years agoAnother suppression, this in SkMatrix caching the typemask.
mtklein [Tue, 8 Jul 2014 00:18:50 +0000 (17:18 -0700)]
Another suppression, this in SkMatrix caching the typemask.

BUG=skia:
R=mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/372103005

10 years agoSuppress race to cache SkPathRef bounds.
mtklein [Mon, 7 Jul 2014 23:28:46 +0000 (16:28 -0700)]
Suppress race to cache SkPathRef bounds.

This should be fixed naturally soon, and it's an innocuous race (some paths
may have their bounds calculated more than once), so we can just quiet it for now.

BUG=skia:
R=mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/376733003

10 years agoTurn on threaded DMQuiltTask drawing for old-format SkPictures too.
mtklein [Mon, 7 Jul 2014 23:17:36 +0000 (16:17 -0700)]
Turn on threaded DMQuiltTask drawing for old-format SkPictures too.

This required the same SkBitmap copying that we added today to SkRecord.

BUG=skia:
R=reed@android.com, robertphillips@google.com, mtklein@google.com, reed@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/375703002

10 years agoremove unneeded flag for android
reed [Mon, 7 Jul 2014 22:54:20 +0000 (15:54 -0700)]
remove unneeded flag for android

BUG=skia:
R=scroggo@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/372133003

10 years agoHandle close button in SkOSWindow_Unix.
bungeman [Mon, 7 Jul 2014 22:46:00 +0000 (15:46 -0700)]
Handle close button in SkOSWindow_Unix.

Currently X with most Window managers calls 'exit' when the close button
is used. This can cause issues as the Window is not properly destroyed.
With this change we can handle this window message and properly exit.

R=reed@google.com

Author: bungeman@google.com

Review URL: https://codereview.chromium.org/377733002

10 years agoSuppress getRefCnt() calls from SkImageFilter.
mtklein [Mon, 7 Jul 2014 22:44:00 +0000 (15:44 -0700)]
Suppress getRefCnt() calls from SkImageFilter.

BUG=skia:2726
R=jcgregorio@google.com, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/373863002

10 years agoRevert of Make GrDrawState and GrPaint take GrEffect* instead of GrEffectRef*. (https...
reed [Mon, 7 Jul 2014 22:17:49 +0000 (15:17 -0700)]
Revert of Make GrDrawState and GrPaint take GrEffect* instead of GrEffectRef*. (https://codereview.chromium.org/377503004/)

Reason for revert:
broke linux builders

Original issue's description:
> Make GrDrawState and GrPaint take GrEffect* instead of GrEffectRef*.
>
> Make Sk-effect virtuals produce GrEffect* rather than GrEffectRef*
>
> Make GrEffectRef a typedef for GrEffect.
>
> Committed: https://skia.googlesource.com/skia/+/2011fe9cdfa63b83489a146cea6a724cede352c8

R=robertphillips@google.com, bsalomon@google.com
TBR=bsalomon@google.com, robertphillips@google.com
NOTREECHECKS=true
NOTRY=true

Author: reed@google.com

Review URL: https://codereview.chromium.org/372053003

10 years agoMake GrDrawState and GrPaint take GrEffect* instead of GrEffectRef*.
bsalomon [Mon, 7 Jul 2014 22:11:31 +0000 (15:11 -0700)]
Make GrDrawState and GrPaint take GrEffect* instead of GrEffectRef*.

Make Sk-effect virtuals produce GrEffect* rather than GrEffectRef*

Make GrEffectRef a typedef for GrEffect.

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/377503004

10 years agoYUV to RGB converter
sugoi [Mon, 7 Jul 2014 22:09:48 +0000 (15:09 -0700)]
YUV to RGB converter

This is a first piece of the GPU YUV decoder, which is the actual effect that performs the conversion. For now, it simply applies the conversion matrix, since it is all I need. I may add modes if different matrices need to be applied or if I add color profile support here.

I'll try to keep these cls short and easy to review, but there should be a few of them coming once this one is in.

BUG=skia:
R=senorblanco@chromium.org, senorblanco@google.com, reed@google.com, bsalomon@google.com

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/378503006

10 years agoguard registering mallocpixelref for chrome
reed [Mon, 7 Jul 2014 21:59:31 +0000 (14:59 -0700)]
guard registering mallocpixelref for chrome

companion to https://codereview.chromium.org/320873003

NOTRY=True
TBR=

Author: reed@google.com

Review URL: https://codereview.chromium.org/373903002

10 years agoremove unneeded flags (have been purged from chrome and android)
reed [Mon, 7 Jul 2014 21:45:11 +0000 (14:45 -0700)]
remove unneeded flags (have been purged from chrome and android)

TBR=

Author: reed@google.com

Review URL: https://codereview.chromium.org/372793003

10 years agochange pixelref to not inherit from SkFlattenable
reed [Mon, 7 Jul 2014 21:32:06 +0000 (14:32 -0700)]
change pixelref to not inherit from SkFlattenable

If I can "inline" MallocPixelRef unflatten, then I think we can delete this code. The only caller today should be unflattening in the legacy path for bitmaps.

R=robertphillips@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/320873003

10 years agoremove SK_SUPPORT_LEGACY_DEVICE_VIRTUAL_ISOPAQUE
reed [Mon, 7 Jul 2014 20:51:48 +0000 (13:51 -0700)]
remove SK_SUPPORT_LEGACY_DEVICE_VIRTUAL_ISOPAQUE

TBR=

Author: reed@google.com

Review URL: https://codereview.chromium.org/373853002

10 years agoremove unneeded flags for android
reed [Mon, 7 Jul 2014 20:48:59 +0000 (13:48 -0700)]
remove unneeded flags for android

R=scroggo@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/369683004

10 years agoSplit SkPicturePlayback out of SkPictureData
robertphillips [Mon, 7 Jul 2014 20:46:35 +0000 (13:46 -0700)]
Split SkPicturePlayback out of SkPictureData

This splits the playback functionality out of SkPictureData. The old SkPictureData::draw method is pulled out along
with its supporting functions as verbatim as possible. Some follow on CLs will be required to:

   re-enable profiling in the debugger (and remove the vestiges of SkTimedPicture)
   re-enable display of command offsets in the picture (this should probably wait until we've switched to SkRecord though)
   Clean up CachedOperationList (maybe fuse with SkPicture::OperationList)
   Split SkPicturePlayback into a base class and two derived classes
   Implement parallel version of GatherGPUInfo for SkRecord

Landing this is blocked on removing Android's use of the abortPlayback entry point.

R=mtklein@google.com, reed@google.com

Author: robertphillips@google.com

Review URL: https://codereview.chromium.org/377623002

10 years agoSuppress new TSAN warnings from DMQuiltTask.
mtklein [Mon, 7 Jul 2014 20:11:29 +0000 (13:11 -0700)]
Suppress new TSAN warnings from DMQuiltTask.

We're racing to invalidate the genID of our pixel ref when multiple
threads write into it, and also to call its genID-changed listeners.

We install no listeners on this particular pixel ref, nor do we ever
care about its genID at all.  So these are benign races, races on
data we never make a decision from.

BUG=skia:2725
R=jcgregorio@google.com, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/370353004

10 years agoRemove SK_DISABLE_PIXELREF_LOCKCOUNT_BALANCE_CHECK.
mtklein [Mon, 7 Jul 2014 19:33:27 +0000 (12:33 -0700)]
Remove SK_DISABLE_PIXELREF_LOCKCOUNT_BALANCE_CHECK.

Skia no longer looks at this define.  It's as if it's always set now.

BUG=skia:
R=scroggo@google.com, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/368953005

10 years agodon't look at SK_SCALAR_IS_FLOAT -- it is true, but no longer defined
reed [Mon, 7 Jul 2014 19:30:28 +0000 (12:30 -0700)]
don't look at SK_SCALAR_IS_FLOAT -- it is true, but no longer defined

TBR=

Author: reed@google.com

Review URL: https://codereview.chromium.org/371033004

10 years agoRemove use of GrEffectRef from draw state and below.
bsalomon [Mon, 7 Jul 2014 18:54:23 +0000 (11:54 -0700)]
Remove use of GrEffectRef from draw state and below.

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/372773002

10 years agoAdd return to SkBoxBlurGetPlatformProcs_SSE4.
scroggo [Mon, 7 Jul 2014 18:41:39 +0000 (11:41 -0700)]
Add return to SkBoxBlurGetPlatformProcs_SSE4.

This fixes Android build.

R=reed@google.com, mtklein@google.com

Author: scroggo@google.com

Review URL: https://codereview.chromium.org/378613002

10 years agogpu and cpu flags for gm and bench.
bsalomon [Mon, 7 Jul 2014 17:46:58 +0000 (10:46 -0700)]
gpu and cpu flags for gm and bench.

BUG=skia:2074
R=borenet@google.com, mtklein@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/376643002

10 years agoRemove GrEffectStage::DeferredStage
bsalomon [Mon, 7 Jul 2014 17:46:32 +0000 (10:46 -0700)]
Remove GrEffectStage::DeferredStage

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/373603005

10 years agoAdd always-threaded SkRecord quilt tests.
mtklein [Mon, 7 Jul 2014 17:41:04 +0000 (10:41 -0700)]
Add always-threaded SkRecord quilt tests.

Now that we're drawing tiles threaded like implside painting, remove the checks
that those lock counts are balanced.  They're just not right for anyone anymore.

SkBitmaps themselves are not threadsafe (even const ones), so shallow copy them
on playback of an SkRecord.  (The underlying SkPixelRefs are threadsafe.)

Simplify quilt drawing by using SkBitmap::extractSubset.  No need for locking.

Bump up to 256x256 tiles.  16x16 tiles just murders performance (way too much
contention).  This has the nice side effect of letting us enable a bunch more
GMs for quilt mode; they drew wrong with small tiles but exactly right with large.

BUG=171776
R=reed@google.com, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/371023005

10 years agoEnable 32/64-bit gyp switch on Android
qiankun.miao [Mon, 7 Jul 2014 16:47:24 +0000 (09:47 -0700)]
Enable 32/64-bit gyp switch on Android

Select 32/64-bit Skia build according skia_arch_width on Android.

BUG=skia:
R=djsollen@google.com, mtklein@google.com

Author: qiankun.miao@intel.com

Review URL: https://codereview.chromium.org/369813003

10 years agoCleanup: Remove TestXCode project.
tfarina [Mon, 7 Jul 2014 16:31:44 +0000 (09:31 -0700)]
Cleanup: Remove TestXCode project.

GYP generates the project files for us now.

$ GYP_GENERATORS=xcode ./gyp_skia

BUG=None
TEST=None
R=mtklein@google.com

Author: tfarina@chromium.org

Review URL: https://codereview.chromium.org/365263006

10 years agorebaseline_server: handle column filtering more generically
epoger [Mon, 7 Jul 2014 15:49:14 +0000 (08:49 -0700)]
rebaseline_server: handle column filtering more generically

This is a step towards making rebaseline_server more flexible, and useful for displaying output of the cluster telemetry server as well as actuals-on-configA-vs-actuals-on-configB.

BUG=skia:2230
R=rmistry@google.com

Author: epoger@google.com

Review URL: https://codereview.chromium.org/369133004

10 years agoUpdate remaining expectations for https://codereview.chromium.org/337603008.
jvanverth [Mon, 7 Jul 2014 15:44:45 +0000 (08:44 -0700)]
Update remaining expectations for https://codereview.chromium.org/337603008.

TBR=bungeman@chromium.org

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/371003004

10 years agoRemove GrDrawState::CommonState
bsalomon [Mon, 7 Jul 2014 15:44:05 +0000 (08:44 -0700)]
Remove GrDrawState::CommonState

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/367163002

10 years agoremove unneeded flags for android
reed [Mon, 7 Jul 2014 15:38:13 +0000 (08:38 -0700)]
remove unneeded flags for android

NOTRY=True
R=scroggo@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/377443008

10 years agoRemove vertical/horizontal metrics selection.
bungeman [Mon, 7 Jul 2014 15:16:37 +0000 (08:16 -0700)]
Remove vertical/horizontal metrics selection.

R=reed@google.com

Author: bungeman@google.com

Review URL: https://codereview.chromium.org/370463002

10 years agoRemove deferred version of GrDrawState.
bsalomon [Mon, 7 Jul 2014 15:15:07 +0000 (08:15 -0700)]
Remove deferred version of GrDrawState.

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/368913003

10 years agoAdd SSE4 version of BlurImage optimizations.
henrik.smiding [Mon, 7 Jul 2014 15:05:40 +0000 (08:05 -0700)]
Add SSE4 version of BlurImage optimizations.

Adds an SSE4.1 version of the existing BlurImage optimizations.
Performance of blur_image_filter_* benchmarks show a 10-50%
improvement on Linux/Ubuntu Core i7.

Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
Committed: https://skia.googlesource.com/skia/+/2830632ce93c97ed7647b13348365ea92e4ea665

R=mtklein@google.com, reed@chromium.org

Author: henrik.smiding@intel.com

Review URL: https://codereview.chromium.org/366593004

10 years agowith no save flag options, we can directly reference matrix and clip in MCRec
reed [Mon, 7 Jul 2014 14:49:34 +0000 (07:49 -0700)]
with no save flag options, we can directly reference matrix and clip in MCRec

patch from issue 364973003

R=fmalita@chromium.org
TBR=fmalita@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/364193005

10 years agoRebaseline Angle gms for recent dashing4 rebaseline
egdaniel [Mon, 7 Jul 2014 14:27:37 +0000 (07:27 -0700)]
Rebaseline Angle gms for recent dashing4 rebaseline
NOTREECHECKS=True
NOTRY=True
R=bsalomon@google.com
TBR=bsalomon@google.com

BUG=skia:

Author: egdaniel@google.com

Review URL: https://codereview.chromium.org/378513004

10 years agoOnly you can prevent memory leaks.
mtklein [Mon, 7 Jul 2014 14:00:14 +0000 (07:00 -0700)]
Only you can prevent memory leaks.

SkPaint::setFoo() refs the foo, leaving the foo with a bonus ref unless you
call ->unref().

Valgrind caught this:
http://108.170.220.120:10117/builders/Test-Ubuntu12-ShuttleA-GTX550Ti-x86_64-Release-Valgrind/builds/317/steps/RunTests/logs/stdio

BUG=skia:
R=tomhudson@chromium.org, mtklein@google.com, tomhudson@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/377523002

10 years agoFix history handling for workspaces, issue with M36
jcgregorio [Mon, 7 Jul 2014 13:56:44 +0000 (06:56 -0700)]
Fix history handling for workspaces, issue with M36

BUG=skia:
R=mtklein@google.com

Author: jcgregorio@google.com

Review URL: https://codereview.chromium.org/363003004

10 years agoRebaseline for gpu dashing4 now doing circle dashes
egdaniel [Mon, 7 Jul 2014 13:28:43 +0000 (06:28 -0700)]
Rebaseline for gpu dashing4 now doing circle dashes
NOTREECHECKS=True
NOTRY=True
R=bsalomon@google.com
TBR=bsalomon@google.com

BUG=skia:

Author: egdaniel@google.com

Review URL: https://codereview.chromium.org/374623003

10 years agoAdd Instant Trace Event for shader compilation in gpu which outputs shader text as...
egdaniel [Mon, 7 Jul 2014 13:17:13 +0000 (06:17 -0700)]
Add Instant Trace Event for shader compilation in gpu which outputs shader text as attribute

BUG=skia:
R=bsalomon@google.com

Author: egdaniel@google.com

Review URL: https://codereview.chromium.org/368943002

10 years agoRevert of Add SSE4 version of BlurImage optimizations. (https://codereview.chromium...
reed [Mon, 7 Jul 2014 01:51:29 +0000 (18:51 -0700)]
Revert of Add SSE4 version of BlurImage optimizations. (https://codereview.chromium.org/366593004/)

Reason for revert:
breaks linker on chrome

[04:36:09.966000] [503/5965] LIB obj\chrome\installer_util.lib
[04:36:10.466000] FAILED: C:\Users\chrome-bot\buildbot\third_party\depot_tools\python276_bin\python.exe gyp-win-tool link-with-manifests environment.x86 True skia.dll "C:\Users\chrome-bot\buildbot\third_party\depot_tools\python276_bin\python.exe gyp-win-tool link-wrapper environment.x86 False link.exe /nologo /IMPLIB:skia.dll.lib /DLL /OUT:skia.dll @skia.dll.rsp" 2 mt.exe rc.exe "obj\skia\skia.skia.dll.intermediate.manifest" obj\skia\skia.skia.dll.generated.manifest
[04:36:10.466000] skia.opts_check_x86.obj : error LNK2019: unresolved external symbol "bool __cdecl SkBoxBlurGetPlatformProcs_SSE4(void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int),void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int),void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int),void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int))" (?SkBoxBlurGetPlatformProcs_SSE4@@YA_NPAP6AXPBIHPAIHHHHH@Z222@Z) referenced in function "bool __cdecl SkBoxBlurGetPlatformProcs(void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int),void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int),void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int),void (__cdecl**)(unsigned int const *,int,unsigned int *,int,int,int,int,int))" (?SkBoxBlurGetPlatformProcs@@YA_NPAP6AXPBIHPAIHHHHH@Z222@Z)
[04:36:10.466000]
[04:36:10.466000] skia.dll : fatal error LNK1120: 1 unresolved externals

Original issue's description:
> Add SSE4 version of BlurImage optimizations.
>
> Adds an SSE4.1 version of the existing BlurImage optimizations.
> Performance of blur_image_filter_* benchmarks show a 10-50%
> improvement on Linux/Ubuntu Core i7.
>
> Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
>
> Committed: https://skia.googlesource.com/skia/+/2830632ce93c97ed7647b13348365ea92e4ea665

R=mtklein@google.com, henrik.smiding@intel.com
TBR=henrik.smiding@intel.com, mtklein@google.com
NOTREECHECKS=true
NOTRY=true

Author: reed@chromium.org

Review URL: https://codereview.chromium.org/375503003

10 years agoLinear-time implementation of willPlaybackBitmaps(), computed & cached
tomhudson [Sat, 5 Jul 2014 20:37:53 +0000 (13:37 -0700)]
Linear-time implementation of willPlaybackBitmaps(), computed & cached
on construction in SkPicture. Unit test.

Template trickery thanks to mtklein@.

BUG=skia:2702
R=mtklein@google.com, reed@android.com, reed@google.com, tomhudson@google.com, mtklein, reed

Author: tomhudson@chromium.org

Review URL: https://codereview.chromium.org/366443002

10 years agoAdd SSE4 version of BlurImage optimizations.
henrik.smiding [Fri, 4 Jul 2014 11:23:17 +0000 (04:23 -0700)]
Add SSE4 version of BlurImage optimizations.

Adds an SSE4.1 version of the existing BlurImage optimizations.
Performance of blur_image_filter_* benchmarks show a 10-50%
improvement on Linux/Ubuntu Core i7.

Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
R=mtklein@google.com

Author: henrik.smiding@intel.com

Review URL: https://codereview.chromium.org/366593004

10 years agoFix SkColorFilterImageFilter matrix optimization.
senorblanco [Thu, 3 Jul 2014 18:13:09 +0000 (11:13 -0700)]
Fix SkColorFilterImageFilter matrix optimization.

The order of matrices passed to multiplication was wrong (apparently,
this optimization was only being tested with matrices which commute).

See Chrome bug http://crbug.com/378362

R=sugoi@chromium.org

Author: senorblanco@chromium.org

Review URL: https://codereview.chromium.org/371523002

10 years agorebaseline_server: if urlSchemaVersion is not specified, assume current
epoger [Thu, 3 Jul 2014 17:50:30 +0000 (10:50 -0700)]
rebaseline_server: if urlSchemaVersion is not specified, assume current

Modification to https://codereview.chromium.org/368933002/ ('rebaseline_server: add urlSchemaVersion parameter')

Means we don't need https://codereview.chromium.org/367173002/ ('add urlSchemaVersion to rebaseline_server link on buildbot console page')

NOTREECHECKS=True
NOTRY=True
R=borenet@google.com
TBR=borenet

Author: epoger@google.com

Review URL: https://codereview.chromium.org/364253003

10 years agoAdding 64 bit checks
sugoi [Thu, 3 Jul 2014 17:44:26 +0000 (10:44 -0700)]
Adding 64 bit checks

Added a few more checks to avoid overflowing 32 bit sizes while computing convolutions.

I also changed a dangerously misleading INHERITED typedef.

BUG=389570
R=senorblanco@google.com, senorblanco@chromium.org

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/361403006

10 years agoRevert of Caching the result of readPixelsSupported (https://codereview.chromium...
reed [Thu, 3 Jul 2014 14:54:08 +0000 (07:54 -0700)]
Revert of Caching the result of readPixelsSupported (https://codereview.chromium.org/364193004/)

Reason for revert:
appears to crash GM on Ubuntu and Win8

http://108.170.220.120:10117/builders/Test-Ubuntu12-ShuttleA-GTX660-x86-Release/builds/1237/steps/GenerateGMs/logs/stdio

Original issue's description:
> Caching the result of readPixelsSupported
>
> The call was calling GR_GL_GetIntegerv 2 times for each readPixels
> and thus was causing a loss of performance
>
> (resubmit of issue 344793008)
>
> Benchmark url: http://packages.gkny.fr/tst/index.html
>
> BUG=skia:2681
>
> Committed: https://skia.googlesource.com/skia/+/753a2964afe5661ce9b2a8ca77ca9d0aabd3173c

R=junov@chromium.org, piotaixr@chromium.org
TBR=junov@chromium.org, piotaixr@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:2681

Author: reed@chromium.org

Review URL: https://codereview.chromium.org/367323003

10 years agomove SkChecksum and SkEmptyShader to private
reed [Thu, 3 Jul 2014 14:26:01 +0000 (07:26 -0700)]
move SkChecksum and SkEmptyShader to private

TBR=reed@google.com

Author: reed@chromium.org

Review URL: https://codereview.chromium.org/366133002

10 years agoCaching the result of readPixelsSupported
piotaixr [Thu, 3 Jul 2014 13:50:50 +0000 (06:50 -0700)]
Caching the result of readPixelsSupported

The call was calling GR_GL_GetIntegerv 2 times for each readPixels
and thus was causing a loss of performance

(resubmit of issue 344793008)

Benchmark url: http://packages.gkny.fr/tst/index.html

BUG=skia:2681
R=junov@chromium.org

Author: piotaixr@chromium.org

Review URL: https://codereview.chromium.org/364193004

10 years agoAdd functions to support NV_path_rendering in OpenGL ES
kkinnunen [Thu, 3 Jul 2014 05:56:35 +0000 (22:56 -0700)]
Add functions to support NV_path_rendering in OpenGL ES

Add OpenGL ES extension functions needed to support NV_path_rendering in OpenGL
ES.

The added glProgramPathFragmentInputGenNV call is defined in NV_path_rendering
revision 30, similar to following:

    Append to the end of the "Shader Inputs" subsection of Section 3.12.2
    "Shader Execution":

    The command

        void ProgramPathFragmentInputGenNV(uint program,
                                           int location,
                                           enum genMode,
                                           int components,
                                           const float *coeffs);

    controls how a user-defined (non-built-in) fragment input of a
    GLSL program object is computed for fragment shading operations that
    occur as a result of CoverFillPathNV or CoverStrokePathNV.

    /program/ names a GLSL program object.  If /program/ has not been
    successfully linked, the error INVALID_OPERATION is generated.

    The given fragment input generation state is loaded into the fragment
    input variable location identified by /location/.  This location
    is a value returned either by GetProgramResourceLocation with a
    /programInterface/ of FRAGMENT_INPUT_NV and a given fragment shader
    input variable name or by GetProgramResourceiv with FRAGMENT_INPUT_NV
    for the /programInterface/ and LOCATION for the property for a given
    fragment input resource index.

    ....

glProgramPathFragmentInputGenNV will be used instead of glPathTexGen,
because the latter depends on fixed function pipeline that is not
exposed in ES.

Also add glGetProgramResourceLocation from OpenGL 4.3 or
ARB_program_interface_query.

Also add FRAGMENT_INPUT define to be used with glGetProgramResourceLocation.

The added functions are not used yet, but they're needed when implementing
NV_path_rendering support for OpenGL ES. They can also be used on OpenGL.

Remove uncalled NV_path_rendering functions, so they do not cause confusion
or take space in the interface definition. The ones that are later used
can be re-added when needed.

Remove definitions NV_path_rendering that are for NV_path_rendering function
parameters that are not used. The ones that are later used
can be re-added when needed.

Committed: https://skia.googlesource.com/skia/+/4a995dfff2ecf91e8bf999d77e3218cec596232c

R=bsalomon@google.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/345723002

10 years agoIgnore some GM images for https://codereview.chromium.org/337603008.
jvanverth [Wed, 2 Jul 2014 22:54:24 +0000 (15:54 -0700)]
Ignore some GM images for https://codereview.chromium.org/337603008.

R=bungeman@google.com
TBR=bungeman@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/361333002

10 years agoCheck pixelRef before attempting to ktx encode.
scroggo [Wed, 2 Jul 2014 22:00:07 +0000 (15:00 -0700)]
Check pixelRef before attempting to ktx encode.

If there is no pixelRef, do not attempt to dereference it.

This was caught running on Android, where we don't have the resources
folder (tracked in b/14406768). We fail to decode anything, and then
attempt to encode an SkBitmap without a pixelRef.

R=krajcevski@google.com

Author: scroggo@google.com

Review URL: https://codereview.chromium.org/362413005

10 years agoGate CrashHandler with SK_CRASH_HANDLER.
mtklein [Wed, 2 Jul 2014 21:30:47 +0000 (14:30 -0700)]
Gate CrashHandler with SK_CRASH_HANDLER.

BUG=skia:
R=bsalomon@google.com, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/364033002

10 years agoMore expectation updates for https://codereview.chromium.org/337603008.
jvanverth [Wed, 2 Jul 2014 21:15:37 +0000 (14:15 -0700)]
More expectation updates for https://codereview.chromium.org/337603008.

R=bungeman@google.com
TBR=bungeman@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/368073002

10 years agorebaseline_server: add urlSchemaVersion parameter
epoger [Wed, 2 Jul 2014 20:51:47 +0000 (13:51 -0700)]
rebaseline_server: add urlSchemaVersion parameter

In https://codereview.chromium.org/369473005 ('rebaseline_server: handle category filtering more generically'), I am changing the meaning of some URL parameters.  If a user pastes an old URL into an updated rebaseline_server instance, they may get unexpected results.  Adding a urlSchemaVersion parameter allows us to automatically convert the URL, or at least warn the user.

BUG=skia:2230
NOTRY=True
R=jcgregorio@google.com

Author: epoger@google.com

Review URL: https://codereview.chromium.org/368933002

10 years agoMake GrAllocator Iter return non-const T
bsalomon [Wed, 2 Jul 2014 20:44:57 +0000 (13:44 -0700)]
Make GrAllocator Iter return non-const T

R=egdaniel@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/362333004

10 years agoRemove SK_IGNORE_FASTER_TEXT_FIX.
Ben Wagner [Wed, 2 Jul 2014 20:09:00 +0000 (16:09 -0400)]
Remove SK_IGNORE_FASTER_TEXT_FIX.

This flag is no longer defined, so remove its use and guarded code.

10 years agoRemove SK_IGNORE_DWRITE_RENDERING_FIX.
Ben Wagner [Wed, 2 Jul 2014 20:04:09 +0000 (16:04 -0400)]
Remove SK_IGNORE_DWRITE_RENDERING_FIX.

This flag is no longer defined anywhere, so remove it's use.

10 years agoRevert of Remove the AA requirement for selecting GrEffect-based clipping. (https...
mtklein [Wed, 2 Jul 2014 19:55:21 +0000 (12:55 -0700)]
Revert of Remove the AA requirement for selecting GrEffect-based clipping. (https://codereview.chromium.org/367013003/)

Reason for revert:
We think this is breaking the roll.

Original issue's description:
> Remove the AA requirement for selecting GrEffect-based clipping.
>
> Also, optionally use the scissor for a bw clip rect element instead of an effect.
>
> Committed: https://skia.googlesource.com/skia/+/a73218bbbdcbe458651d10815e8d3b73d71b8e11
>
> Committed: https://skia.googlesource.com/skia/+/e9a729cb4d3f05b9c750dc1f63a9cc65b5659f04

R=robertphillips@google.com, bsalomon@google.com
TBR=bsalomon@google.com, robertphillips@google.com
NOTREECHECKS=true
NOTRY=true

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/361313002

10 years agogen_bench_expectations: use subprocess instead of shell_utils due to buildbot-side...
borenet [Wed, 2 Jul 2014 19:52:34 +0000 (12:52 -0700)]
gen_bench_expectations: use subprocess instead of shell_utils due to buildbot-side name collision

BUG=skia:2714
R=rmistry@google.com
TBR=rmistry
NOTRY=true

Author: borenet@google.com

Review URL: https://codereview.chromium.org/368043002

10 years agoExclude Clang on Windows too. Comment this up a bit.
mtklein [Wed, 2 Jul 2014 19:38:38 +0000 (12:38 -0700)]
Exclude Clang on Windows too.  Comment this up a bit.

BUG=391016
R=tomhudson@chromium.org, mtklein@google.com, rnk@chromium.org, thakis@chromium.org

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/363983004

10 years agoCombine clip stack and clip origin into one struct in GrIODB.
bsalomon [Wed, 2 Jul 2014 19:11:24 +0000 (12:11 -0700)]
Combine clip stack and clip origin into one struct in GrIODB.

R=egdaniel@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/364823004

10 years agoRemove the AA requirement for selecting GrEffect-based clipping.
bsalomon [Wed, 2 Jul 2014 18:42:29 +0000 (11:42 -0700)]
Remove the AA requirement for selecting GrEffect-based clipping.

Also, optionally use the scissor for a bw clip rect element instead of an effect.

Committed: https://skia.googlesource.com/skia/+/a73218bbbdcbe458651d10815e8d3b73d71b8e11

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/367013003

10 years agoDisable assembly code in MemorySanitizer builds.
Mike Klein [Wed, 2 Jul 2014 18:37:29 +0000 (14:37 -0400)]
Disable assembly code in MemorySanitizer builds.

MemorySanitizer is an unitialized memory use detector which is used in
Chromium, and does not presently support assembly code.

BUG=chromium:344505, chromium:373739
R=mtklein@google.com

Review URL: https://codereview.chromium.org/367973005

10 years agoUpdate some expectations for https://codereview.chromium.org/337603008.
jvanverth [Wed, 2 Jul 2014 18:33:06 +0000 (11:33 -0700)]
Update some expectations for https://codereview.chromium.org/337603008.

R=bungeman@google.com
TBR=bungeman@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/362373002

10 years agoRevert of Remove the AA requirement for selecting GrEffect-based clipping. (https...
bsalomon [Wed, 2 Jul 2014 18:11:30 +0000 (11:11 -0700)]
Revert of Remove the AA requirement for selecting GrEffect-based clipping. (https://codereview.chromium.org/367013003/)

Reason for revert:
Assertion failing.

Original issue's description:
> Remove the AA requirement for selecting GrEffect-based clipping.
>
> Also, optionally use the scissor for a bw clip rect element instead of an effect.
>
> Committed: https://skia.googlesource.com/skia/+/a73218bbbdcbe458651d10815e8d3b73d71b8e11

R=robertphillips@google.com
TBR=robertphillips@google.com
NOTREECHECKS=true
NOTRY=true

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/369613002

10 years agoFix no previous prototype for function error
halcanary [Wed, 2 Jul 2014 17:43:50 +0000 (10:43 -0700)]
Fix no previous prototype for function error

BUG=skia:2720
R=scroggo@google.com, caryclark@google.com

Author: halcanary@google.com

Review URL: https://codereview.chromium.org/363983002

10 years agoRemove the AA requirement for selecting GrEffect-based clipping.
bsalomon [Wed, 2 Jul 2014 17:40:52 +0000 (10:40 -0700)]
Remove the AA requirement for selecting GrEffect-based clipping.

Also, optionally use the scissor for a bw clip rect element instead of an effect.

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/367013003

10 years agoFix RecreateSkps
borenet [Wed, 2 Jul 2014 16:37:03 +0000 (09:37 -0700)]
Fix RecreateSkps

- Use shell_utils to auto-print all subprocesses.
- Only download bench_pictures data.
- Skip or error out on bots which failed BenchPictures, PostBench, or UploadBenchResults.

BUG=skia:2714
R=rmistry@google.com
TBR=rmistry
NOTRY=true

Author: borenet@google.com

Review URL: https://codereview.chromium.org/363833004

10 years agoCTFontCopy*Name may return NULL.
bungeman [Wed, 2 Jul 2014 14:57:59 +0000 (07:57 -0700)]
CTFontCopy*Name may return NULL.

BUG=skia:2708
R=caryclark@google.com, reed@google.com

Author: bungeman@google.com

Review URL: https://codereview.chromium.org/365683004

10 years agoReduce integer divides in GrAllocator
bsalomon [Wed, 2 Jul 2014 14:54:42 +0000 (07:54 -0700)]
Reduce integer divides in GrAllocator

R=egdaniel@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/362023002

10 years agorebaseline_server: use just skpdiff, not Python Image Library
epoger [Wed, 2 Jul 2014 14:43:04 +0000 (07:43 -0700)]
rebaseline_server: use just skpdiff, not Python Image Library

BUG=skia:2414
R=djsollen@google.com, borenet@google.com

Author: epoger@google.com

Review URL: https://codereview.chromium.org/325413003

10 years agoFix int to float bug in shader on gpu Dotted Line cl
egdaniel [Wed, 2 Jul 2014 13:56:30 +0000 (06:56 -0700)]
Fix int to float bug in shader on gpu Dotted Line cl
NOTRY=True
NOTREECHECKS=True
R=bsalomon@google.com
TBR=bsalomon@google.com
BUG=skia:

Author: egdaniel@google.com

Review URL: https://codereview.chromium.org/361253002

10 years agoRevert of Update SKP version to 39 (https://codereview.chromium.org/367863003/)
borenet [Wed, 2 Jul 2014 13:54:31 +0000 (06:54 -0700)]
Revert of Update SKP version to 39 (https://codereview.chromium.org/367863003/)

Reason for revert:
Deleted expectations again. https://code.google.com/p/skia/issues/detail?id=2714

Original issue's description:
> Update SKP version to 39
>
> Automatic commit by the RecreateSKPs bot.
>
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/50b3dab100a6537ace504357d38e49ba0a14de43

TBR=
NOTREECHECKS=true
NOTRY=true

Author: borenet@google.com

Review URL: https://codereview.chromium.org/367983002

10 years agoAdd the ability for gpu to render dotted lines (dashed line, 0 on interval, round...
egdaniel [Wed, 2 Jul 2014 13:21:32 +0000 (06:21 -0700)]
Add the ability for gpu to render dotted lines (dashed line, 0 on interval, round caps)

BUG=skia:
R=bsalomon@google.com

Author: egdaniel@google.com

Review URL: https://codereview.chromium.org/345113003

10 years agoreturn early instead of crashing if fileData fails
reed [Wed, 2 Jul 2014 12:57:50 +0000 (05:57 -0700)]
return early instead of crashing if fileData fails

BUG=skia:
R=krajcevski@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/362833004

10 years agoUpdate SKP version to 39
borenet [Wed, 2 Jul 2014 06:46:11 +0000 (23:46 -0700)]
Update SKP version to 39

Automatic commit by the RecreateSKPs bot.

TBR=

Author: borenet@google.com

Review URL: https://codereview.chromium.org/367863003

10 years agoUses optional mutex to guard construction of the singleton, which
tomhudson [Wed, 2 Jul 2014 01:54:41 +0000 (18:54 -0700)]
Uses optional mutex to guard construction of the singleton, which
initializes the non-threadsafe libfontconfig. Without this change,
Skia's parallel path ops test runner crashes 6/10 and hangs 2/10 on
startup; with this change, 0/10 problems.

BUG=skia:2693
R=mtklein@google.com, bungeman@google.com, reed@google.com

Committed: https://skia.googlesource.com/skia/+/df022f5972ae6a2a1d96d15c50eca52cade3abd8

Committed: https://skia.googlesource.com/skia/+/60b08a0adfe73f593af62c8d3f55958438360e1b

Author: tomhudson@chromium.org

Review URL: https://codereview.chromium.org/355573006

10 years agoRemove A8_LCD luminance hack.
jvanverth [Tue, 1 Jul 2014 20:42:59 +0000 (13:42 -0700)]
Remove A8_LCD luminance hack.

It no longer appears to be necessary, and the powf function is slowing down
Android text rendering.

R=bungeman@google.com, bsalomon@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/337603008

10 years agoRevert of Fix race condition in parallel font initialization. (https://codereview...
mtklein [Tue, 1 Jul 2014 20:33:32 +0000 (13:33 -0700)]
Revert of Fix race condition in parallel font initialization. (https://codereview.chromium.org/355573006/)

Reason for revert:
canaries still slightly broken

Original issue's description:
> Fix race condition in parallel font initialization.
>
> Uses a mutex to guard construction of the singleton, which initialies
> the non-threadsafe libfontconfig.  Without this change, the parallel
> path ops test runner crashes 6/10 and hangs 2/10 on startup; with this
> change, 0/10 problems.
>
> BUG=skia:2693
> R=mtklein@google.com,bungeman@google.com
>
> Committed: https://skia.googlesource.com/skia/+/df022f5972ae6a2a1d96d15c50eca52cade3abd8
>
> Committed: https://skia.googlesource.com/skia/+/60b08a0adfe73f593af62c8d3f55958438360e1b

R=bungeman@google.com, reed@google.com, tomhudson@google.com, tomhudson@chromium.org
TBR=bungeman@google.com, reed@google.com, tomhudson@chromium.org, tomhudson@google.com
NOTREECHECKS=true
NOTRY=true
BUG=skia:2693

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/354133004

10 years agocorrectly plumb through explicit rowbytes for allocPixels
reed [Tue, 1 Jul 2014 19:48:11 +0000 (12:48 -0700)]
correctly plumb through explicit rowbytes for allocPixels

BUG=skia:
R=caryclark@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/357073003

10 years agoCleanup usage of GetResourcePath() after commit bcbc1788b478b1e54079318ad073e8490aa66fae.
tfarina [Tue, 1 Jul 2014 19:35:49 +0000 (12:35 -0700)]
Cleanup usage of GetResourcePath() after commit bcbc1788b478b1e54079318ad073e8490aa66fae.

There was a clean up opportunity left over after
https://skia.googlesource.com/skia/+/bcbc1788b478b1e54079318ad073e8490aa66fae, that could make use of the default parameter of GetResourcePath() function to make some call sites cleaner.

We decided to make it in a separate CL to make reviewer's and author's life easier, so we could catch errors and/or mistakes easily.

BUG=None
TEST=make all && out/Debug/dm && out/Debug/SampleApp
R=mtklein@google.com

Author: tfarina@chromium.org

Review URL: https://codereview.chromium.org/351133003

10 years agoTemporarily resurface matrix/clip save flags enum values.
fmalita [Tue, 1 Jul 2014 18:55:41 +0000 (11:55 -0700)]
Temporarily resurface matrix/clip save flags enum values.

Pending Android build fixes.

R=reed@google.com, scroggo@google.com

Author: fmalita@chromium.org

Review URL: https://codereview.chromium.org/365463005

10 years agoHey, maybe we should _call_ glFinish.
mtklein [Tue, 1 Jul 2014 17:02:42 +0000 (10:02 -0700)]
Hey, maybe we should _call_ glFinish.

BUG=skia:

CQ_EXTRA_TRYBOTS=tryserver.skia:Build-Mac10.7-Clang-x86-Debug-Trybot
R=mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/345063004

10 years agoRename SkPicturePlayback to SkPictureData
robertphillips [Tue, 1 Jul 2014 15:47:04 +0000 (08:47 -0700)]
Rename SkPicturePlayback to SkPictureData

This is in preparation for splitting the playback portion of the new SkPictureData class into a new SkPicturePlayback class.

R=reed@google.com, mtklein@google.com

Author: robertphillips@google.com

Review URL: https://codereview.chromium.org/362773002

10 years agodraft gpu support in nanobench
mtklein [Tue, 1 Jul 2014 15:43:42 +0000 (08:43 -0700)]
draft gpu support in nanobench

BUG=skia:
R=bsalomon@google.com, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/359473004

10 years agoFix race condition in parallel font initialization.
tomhudson [Tue, 1 Jul 2014 15:06:14 +0000 (08:06 -0700)]
Fix race condition in parallel font initialization.

Uses a mutex to guard construction of the singleton, which initialies
the non-threadsafe libfontconfig.  Without this change, the parallel
path ops test runner crashes 6/10 and hangs 2/10 on startup; with this
change, 0/10 problems.

BUG=skia:2693
R=mtklein@google.com, bungeman@google.com, reed@google.com, tomhudson@google.com

Committed: https://skia.googlesource.com/skia/+/df022f5972ae6a2a1d96d15c50eca52cade3abd8

Author: tomhudson@chromium.org

Review URL: https://codereview.chromium.org/355573006

10 years agoSpecify the name of tool for creating a makefile.
scroggo [Tue, 1 Jul 2014 15:02:30 +0000 (08:02 -0700)]
Specify the name of tool for creating a makefile.

When generating tool Android.mks from gyp files, specify the name
of the target needed.

This fixes a bug where nanobench was being added to the same
makefile as bench, which would not compile.

TODO: Make gyp_to_android more general, so that it handles this
case better. See http://skbug.com/2709

R=mtklein@google.com

Author: scroggo@google.com

Review URL: https://codereview.chromium.org/366653002

10 years agoAdd multithreaded mode to quilt mode in DM.
mtklein [Tue, 1 Jul 2014 14:46:50 +0000 (07:46 -0700)]
Add multithreaded mode to quilt mode in DM.

Default off for now.  Something to work toward.

BUG=skia:
R=robertphillips@google.com, tomhudson@chromium.org, mtklein@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/360793002

10 years agoWhen performing offscreen rendering on windows, attempt to use a pbuffer context.
bsalomon [Tue, 1 Jul 2014 14:20:11 +0000 (07:20 -0700)]
When performing offscreen rendering on windows, attempt to use a pbuffer context.

A pbuffer context is less likely to have a blocking SwapBuffers (due to vsync).

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/336863009

10 years agoAdjust the alpha type for pixelRefs.
scroggo [Tue, 1 Jul 2014 14:08:19 +0000 (07:08 -0700)]
Adjust the alpha type for pixelRefs.

Move SkBitmap's validate_alphaType to SkImageInfo, with the new
name SkColorTypeValidateAlphaType. Use it in SkPixelRef's constructors,
as well as in SkDecodingImageGenerator. This fixes a bug where an
SkPixelRef's SkAlphaType could get out of sync with its SkBitmap,
when both were assigned the same SkAlphaType.

R=reed@google.com, halcanary@google.com

Author: scroggo@google.com

Review URL: https://codereview.chromium.org/346593003

10 years agoRevert of Update SKP version to 34 (https://codereview.chromium.org/347303003/)
Eric Boren [Tue, 1 Jul 2014 13:39:00 +0000 (09:39 -0400)]
Revert of Update SKP version to 34 (https://codereview.chromium.org/347303003/)

Reason for revert:
This removed all bench expectations!

Original issue's description:
> Update SKP version to 34
>
> Automatic commit by the RecreateSKPs bot.
>
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/0808c82e83749c24183af92a43c2e0e69b35d745

TBR=
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/362753002

10 years agoGrant independence to SkBaseMutex on Windows.
tomhudson [Tue, 1 Jul 2014 13:26:48 +0000 (06:26 -0700)]
Grant independence to SkBaseMutex on Windows.

Under POSIX, class SkMutex inherits from struct SkBaseMutex.
Since we can't have a POD mutex on Windows, we used to just
typedef SkMutex SkBaseMutex there. However, that makes it
impossible to forward-declare SkBaseMutex consistently across
platforms. With this CL we declare an empty struct SkBaseMutex
on Windows, which should have 0 cost but make the compiler happy.

R=bungeman@google.com, mtklein@google.com, tomhudson@google.com, bungeman, mtklein
BUG=skia:

Author: tomhudson@chromium.org

Review URL: https://codereview.chromium.org/364473002

10 years agoHide symbols in S32A_Opaque_BlitRow32_SSE4
henrik.smiding [Tue, 1 Jul 2014 13:12:46 +0000 (06:12 -0700)]
Hide symbols in S32A_Opaque_BlitRow32_SSE4

Marks the symbols in the S32A_Opaque_BlitRow32_SSE4 files as hidden,
so Chromium can build.
Also enables the optimizations.

Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
R=mtklein@google.com, joakim.landberg@intel.com

Author: henrik.smiding@intel.com

Review URL: https://codereview.chromium.org/368573002

10 years agoremove flags that are now in chrome's SkUserConfig.h
reed [Tue, 1 Jul 2014 09:51:26 +0000 (02:51 -0700)]
remove flags that are now in chrome's SkUserConfig.h

R=reed@google.com
NOTRY=True

Author: reed@chromium.org

Review URL: https://codereview.chromium.org/367483003

10 years agoCleanup: Delete SkJSON header.
tfarina [Tue, 1 Jul 2014 00:11:33 +0000 (17:11 -0700)]
Cleanup: Delete SkJSON header.

It seems it does not have any implementation so it is unlikely it will
have any user, either internal or external.

BUG=None
TEST=make all
R=bsalomon@google.com, mtklein@google.com

Author: tfarina@chromium.org

Review URL: https://codereview.chromium.org/353183006

10 years agoRevert of Fix race condition in parallel font initialization. (https://codereview...
reed [Mon, 30 Jun 2014 23:48:23 +0000 (16:48 -0700)]
Revert of Fix race condition in parallel font initialization. (https://codereview.chromium.org/355573006/)

Reason for revert:
breaks chrome builds

Original issue's description:
> Fix race condition in parallel font initialization.
>
> Uses a mutex to guard construction of the singleton, which initialies
> the non-threadsafe libfontconfig.  Without this change, the parallel
> path ops test runner crashes 6/10 and hangs 2/10 on startup; with this
> change, 0/10 problems.
>
> BUG=skia:2693
> R=mtklein@google.com,bungeman@google.com
>
> Committed: https://skia.googlesource.com/skia/+/df022f5972ae6a2a1d96d15c50eca52cade3abd8

R=mtklein@google.com, bungeman@google.com, reed@google.com, tomhudson@google.com, tomhudson@chromium.org
TBR=bungeman@google.com, mtklein@google.com, reed@google.com, tomhudson@chromium.org, tomhudson@google.com
NOTREECHECKS=true
NOTRY=true
BUG=skia:2693

Author: reed@chromium.org

Review URL: https://codereview.chromium.org/365503003

10 years agoFix race condition in parallel font initialization.
tomhudson [Mon, 30 Jun 2014 21:14:01 +0000 (14:14 -0700)]
Fix race condition in parallel font initialization.

Uses a mutex to guard construction of the singleton, which initialies
the non-threadsafe libfontconfig.  Without this change, the parallel
path ops test runner crashes 6/10 and hangs 2/10 on startup; with this
change, 0/10 problems.

BUG=skia:2693
R=mtklein@google.com, bungeman@google.com, reed@google.com, tomhudson@google.com

Author: tomhudson@chromium.org

Review URL: https://codereview.chromium.org/355573006

10 years agoforce opaque if kCGImageAlpha returns None
reed [Mon, 30 Jun 2014 21:06:00 +0000 (14:06 -0700)]
force opaque if kCGImageAlpha returns None

BUG=skia:2424
R=caryclark@google.com

Author: reed@google.com

Review URL: https://codereview.chromium.org/355403003