platform/upstream/libSkiaSharp.git
8 years agoUse BUILD_TYPE in cmake_build.
rmistry [Mon, 25 Apr 2016 13:29:21 +0000 (06:29 -0700)]
Use BUILD_TYPE in cmake_build.

Build files were directly written to the out directory. This change makes them go to the expected out/{Release, Debug} directory.

BUG=skia:5160

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1916773002
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/1916773002

8 years ago"Using Skia" tutorial: googlecode -> googlesource
borenet [Mon, 25 Apr 2016 13:01:51 +0000 (06:01 -0700)]
"Using Skia" tutorial: googlecode -> googlesource

BUG=skia:5212
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1915013002
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1915013002

Review URL: https://codereview.chromium.org/1915013002

8 years agoBatch multiple single NVPR draw paths to instanced draws
kkinnunen [Mon, 25 Apr 2016 09:16:09 +0000 (02:16 -0700)]
Batch multiple single NVPR draw paths to instanced draws

Batch multiple single NVPR draw paths to instanced draws.
A draw path batch can be combined to other batch if the
batches do not overlap and have same draw characteristics.

Join the batches in linked list and flatten the list to a
path list during draw time.

Replace GrPathRendering::drawPath with GrPathRendering::drawPaths.

Perf changes ARM, Shield TV device, Tegra X1 GPU:
                desk_chalkboard.skp_1 21.5ms -> 17.8ms 0.83x
                    desk_mapsvg.skp_1 7.49ms -> 6.18ms 0.82x
Others results are more volatile.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1908433002

Review URL: https://codereview.chromium.org/1908433002

8 years agoCompute bounds union while combining path range draw path batches
kkinnunen [Mon, 25 Apr 2016 06:35:35 +0000 (23:35 -0700)]
Compute bounds union while combining path range draw path batches

Compute bounds union while combining batches generated by NVPR device
independent text code-path. Otherwise the overlap test of the
GrDrawTarget::recordBatch violate painter's algorithm order.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1894973008

Review URL: https://codereview.chromium.org/1894973008

8 years agoUpdate DNG SDK version
yujieqin [Sun, 24 Apr 2016 15:14:57 +0000 (08:14 -0700)]
Update DNG SDK version

* Add PATENTS file

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897773003

Review URL: https://codereview.chromium.org/1897773003

8 years agoUpdate SKP version
update-skps [Sun, 24 Apr 2016 07:32:19 +0000 (00:32 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1918653002

Review URL: https://codereview.chromium.org/1918653002

8 years agoallow imagefilter to manage CTM decomposition
reed [Sat, 23 Apr 2016 22:19:07 +0000 (15:19 -0700)]
allow imagefilter to manage CTM decomposition

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1908173006

Review URL: https://codereview.chromium.org/1908173006

8 years agoRemove SkCanvas::LayerIter guard
robertphillips [Sat, 23 Apr 2016 15:26:43 +0000 (08:26 -0700)]
Remove SkCanvas::LayerIter guard

The LayerIter is no longer required by the Bazel build.

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1917693002

Review URL: https://codereview.chromium.org/1917693002

8 years agoSupport the non-native (RGBA/BGRA) swizzle
msarett [Fri, 22 Apr 2016 23:27:24 +0000 (16:27 -0700)]
Support the non-native (RGBA/BGRA) swizzle

BUG=skia:4456
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1907593004

Review URL: https://codereview.chromium.org/1907593004

8 years agoUse SkEncodedInfo in place of SkSwizzler::SrcConfig
msarett [Fri, 22 Apr 2016 20:18:37 +0000 (13:18 -0700)]
Use SkEncodedInfo in place of SkSwizzler::SrcConfig

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1911613002

Review URL: https://codereview.chromium.org/1911613002

8 years agoEnable flattening/unflattening with custom unflatten procs
msarett [Fri, 22 Apr 2016 19:43:07 +0000 (12:43 -0700)]
Enable flattening/unflattening with custom unflatten procs

Now flattenables are serialized using a string name, so that
flattenables do not necessarily need to be registered before
serialization.  They just need to override getTypeName().

Allows custom unflatten procs to be set on the SkReadBuffer.
This is optional if the flattenable is registered, but otherwise
must be called.

This was split off from:
https://codereview.chromium.org/1837913003/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1858323002

Review URL: https://codereview.chromium.org/1858323002

8 years agodocs: fix fiddle image links.
jcgregorio [Fri, 22 Apr 2016 18:43:42 +0000 (11:43 -0700)]
docs: fix fiddle image links.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1915493004
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1915493004
TBR=hal

Review URL: https://codereview.chromium.org/1915493004

8 years agoOptionally enable SkValidatingReadBuffer in SkPictureImageFilter
robertphillips [Fri, 22 Apr 2016 18:40:42 +0000 (11:40 -0700)]
Optionally enable SkValidatingReadBuffer in SkPictureImageFilter

The task here is to get the type of ReadBuffer being used in SkPicture::MakeFromBuffer mirrored in the down-stack SkPicturePlayback::draw call.

BUG=skia:5208

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1918433002

Review URL: https://codereview.chromium.org/1918433002

8 years agoskia.org: Update all fiddles for fiddle 2.0.
jcgregorio [Fri, 22 Apr 2016 18:25:43 +0000 (11:25 -0700)]
skia.org: Update all fiddles for fiddle 2.0.

The images won't show up in the preview, which won't happen until I switch over to fiddle 2.0 in production later today.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1910593005
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1910593005

Review URL: https://codereview.chromium.org/1910593005

8 years agoRetract SkDevice a bit more
robertphillips [Fri, 22 Apr 2016 17:57:16 +0000 (10:57 -0700)]
Retract SkDevice a bit more

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1905383002

Review URL: https://codereview.chromium.org/1905383002

8 years agoSkMD5: cleanup header and minor refactor
halcanary [Fri, 22 Apr 2016 17:40:49 +0000 (10:40 -0700)]
SkMD5: cleanup header and minor refactor

Also: I now define a non-virtual function in terms of a final virtual
function.  This reduces the number of actual functions while adding
no overhead.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1911363002

Review URL: https://codereview.chromium.org/1911363002

8 years agoRevert of SkStringPrintf and SkString::printf now are no longer limted by a static...
halcanary [Fri, 22 Apr 2016 15:19:04 +0000 (08:19 -0700)]
Revert of SkStringPrintf and SkString::printf now are no longer limted by a static buffer (patchset #5 id:80001 of https://codereview.chromium.org/1403803002/ )

Reason for revert:
breaking something

Original issue's description:
> SkStringPrintf and SkString::printf now are no longer limted by a static buffer
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1403803002
>
> Committed: https://skia.googlesource.com/skia/+/606cadd5aac62299ef2e277709b3684cae2bf96c

TBR=tomhudson@google.com,reed@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/1908423002

8 years agoSkStringPrintf and SkString::printf now are no longer limted by a static buffer
halcanary [Fri, 22 Apr 2016 15:03:03 +0000 (08:03 -0700)]
SkStringPrintf and SkString::printf now are no longer limted by a static buffer

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1403803002

Review URL: https://codereview.chromium.org/1403803002

8 years agoutils.gypi: remove empty file SkMD5.cpp
halcanary [Fri, 22 Apr 2016 14:45:38 +0000 (07:45 -0700)]
utils.gypi: remove empty file SkMD5.cpp

Review URL: https://codereview.chromium.org/1905363002

8 years agoRemove some unnecessary #includes of SkImageFilter.h
senorblanco [Fri, 22 Apr 2016 14:36:12 +0000 (07:36 -0700)]
Remove some unnecessary #includes of SkImageFilter.h

R=robertphillips@google.com
TBR=reed@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1907283002

Review URL: https://codereview.chromium.org/1907283002

8 years agoFixes for SkRWBuffer
scroggo [Fri, 22 Apr 2016 13:59:01 +0000 (06:59 -0700)]
Fixes for SkRWBuffer

Do not call SkBufferHead::validate in SkROBuffer's destructor, which
may be called in a separate thread from SkRWBuffer::append. validate()
reads SkBufferBlock::fUsed, and append() writes to it, resulting in
a data race.

Update some comments to be more clear about how it is safe to use
these classes across threads.

Test the readers in separate threads.

In addition, make sure it is safe to create a reader even when no
data has been appended. Add tests for this case.

Mark a parameter to SkBufferHead::validate() as const, reflecting
its use.

BUG=chromium:601578
BUG=chromium:605479

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1871953002

Review URL: https://codereview.chromium.org/1871953002

8 years agoSkPDF: PDF/A runtime switch
halcanary [Fri, 22 Apr 2016 13:10:21 +0000 (06:10 -0700)]
SkPDF: PDF/A runtime switch

TODO: remove gyp variable and modify API  in SkDocument.h

SkMD5 now moved into core as pdf depends on it now.

BUG=skia:3110
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1891873002

Committed: https://skia.googlesource.com/skia/+/570f18b43417d73c3fbd113cc0b4258e08b14c82

Review URL: https://codereview.chromium.org/1891873002

8 years agoRefactor to separate backend object lifecycle and GpuResource budget decision
kkinnunen [Fri, 22 Apr 2016 08:48:29 +0000 (01:48 -0700)]
Refactor to separate backend object lifecycle and GpuResource budget decision

Refactor GrGpuResource to contain two different pieces of state:
 a) instance is budgeted or not budgeted
 b) instance references wrapped backend objects or not

The "object lifecycle" was also attached to backend object
handles (ids), which made the code a bit unclear. Backend objects
would be associated with GrGpuResource::LifeCycle, even though
GrGpuResource::LifeCycle refers to the GpuResource, and individual
backend objects in one GpuResource might be governed with different
"lifecycle".

Mark the budgeted/not budgeted with SkBudgeted::kYes, SkBudgeted::kNo.
This was previously GrGpuResource::kCached_LifeCycle,
GrGpuResource::kUncached_LifeCycle.

Mark the "references wrapped object" with boolean. This was previously
GrGpuResource::kBorrowed_LifeCycle,
GrGpuResource::kAdopted_LifeCycle for GrGpuResource.

Associate the backend object ownership status with
GrBackendObjectOwnership for the backend object handles.

The resource type leaf constuctors, such has GrGLTexture or
GrGLTextureRenderTarget take "budgeted" parameter. This parameter
is passed to GrGpuResource::registerWithCache().

The resource type intermediary constructors, such as GrGLTexture
constructors for class GrGLTextureRenderTarget do not take "budgeted"
parameters, intermediary construtors do not call registerWithCache.

Removes the need for tagging GrGpuResource -derived subclass
constructors with "Derived" parameter.

Makes instances that wrap backend objects be registered with
a new function GrGpuResource::registerWithCacheWrapped().

Removes "budgeted" parameter from classes such as StencilAttahment, as
they are always cached and never wrap any external backend objects.

Removes the use of concept "external" from the member function names.
The API refers to the objects as "wrapped", so make all related
functions use the term consistently.

No change in functionality. Resources referencing wrapped objects are
always inserted to the cache with budget decision kNo.

BUG=594928
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1862043002

Review URL: https://codereview.chromium.org/1862043002

8 years agoAdd compile guard for SkCanvas::LayerIter
robertphillips [Thu, 21 Apr 2016 22:22:14 +0000 (15:22 -0700)]
Add compile guard for SkCanvas::LayerIter

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1911893002

Review URL: https://codereview.chromium.org/1911893002

8 years agoRevise WSI setup.
jvanverth [Thu, 21 Apr 2016 18:46:23 +0000 (11:46 -0700)]
Revise WSI setup.

- Set up present queue creation correctly.
- Move WSI functions to VulkanTestContext.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1899213002

Review URL: https://codereview.chromium.org/1899213002

8 years agoGuard against FramebufferStatus checks in GrGLRenderTarget
robertphillips [Thu, 21 Apr 2016 18:27:43 +0000 (11:27 -0700)]
Guard against FramebufferStatus checks in GrGLRenderTarget

BUG=skia:5200

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1905063002

Review URL: https://codereview.chromium.org/1905063002

8 years agoHide SkCanvas::LayerIter
robertphillips [Thu, 21 Apr 2016 18:05:32 +0000 (11:05 -0700)]
Hide SkCanvas::LayerIter

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897863008

Committed: https://skia.googlesource.com/skia/+/a814000cbb0e90624a410b4359fff7f3ec66f0e0

Review URL: https://codereview.chromium.org/1897863008

8 years agoSkPathEffect.h: add SK_API where needed
halcanary [Thu, 21 Apr 2016 17:54:11 +0000 (10:54 -0700)]
SkPathEffect.h: add SK_API where needed

motivation: .so fiddle

TBR=reed@google.com
no api chages, only visibility

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1908893004

Review URL: https://codereview.chromium.org/1908893004

8 years agoFix indentation and casts in SkTArray.
bungeman [Thu, 21 Apr 2016 17:52:03 +0000 (10:52 -0700)]
Fix indentation and casts in SkTArray.

Review URL: https://codereview.chromium.org/1902423007

8 years agoSkCpu w/o static initializer
mtklein [Thu, 21 Apr 2016 17:34:41 +0000 (10:34 -0700)]
SkCpu w/o static initializer

I think I cracked it.

Though, this may not technically be legal C++...
I've only got one definition of SkCpu::gCachedFeatures,
but two different declarations: non-const in SkCpu.cpp, const elsewhere.

Is this...
   - legal C++?
   - not C++ but probably works as I think?
   - not C++ and will probably blow up?
   - who knows, let's see?

I have tested that the features are cached properly, read properly, and that the generated code treats SkCpu::gCachedFeatures as a global constant outside SkCpu.cpp.  So it all observably works optimally.

Expanding testing to more bots.
TBR=reed@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1905683003

Review URL: https://codereview.chromium.org/1905683003

8 years agogDefaultProfileIsSRGB symbol must exist in .so
halcanary [Thu, 21 Apr 2016 16:56:43 +0000 (09:56 -0700)]
gDefaultProfileIsSRGB symbol must exist in .so

fix fiddle
TBR=

Review URL: https://codereview.chromium.org/1912593003

8 years agoRemove obsolete image codec colorspace hacks.
brianosman [Thu, 21 Apr 2016 15:48:18 +0000 (08:48 -0700)]
Remove obsolete image codec colorspace hacks.

With Herb's latest change, the codecs themselves are emitting images
tagged based on the new (separate) global flag. All of the changes we
had here to fixup color profile type are no longer needed (and just
create more confusion). This change effectively reverts the file to
where it was two commits ago.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1904943003

Review URL: https://codereview.chromium.org/1904943003

8 years agoFix code regression to more precise call.
herb [Thu, 21 Apr 2016 15:45:39 +0000 (08:45 -0700)]
Fix code regression to more precise call.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1905993002

Review URL: https://codereview.chromium.org/1905993002

8 years agoSkTArray: fix invalid reinterpret_casts over non-initialized memory.
krasin [Thu, 21 Apr 2016 15:34:00 +0000 (08:34 -0700)]
SkTArray: fix invalid reinterpret_casts over non-initialized memory.

This should fix 'CFI Linux' buildbot, which is currently horribly broken:
https://build.chromium.org/p/chromium.fyi/builders/CFI%20Linux/builds/5115

BUG=605337

Review URL: https://codereview.chromium.org/1908763002

8 years agoInclude scaler context flags (gamma and contrast boost) in the text blob cache key.
brianosman [Thu, 21 Apr 2016 15:29:06 +0000 (08:29 -0700)]
Include scaler context flags (gamma and contrast boost) in the text blob cache key.

Fixes a bug in skiaserve when switching between L32 and S32 modes - we were
reusing blobs created with the other (wrong) masks.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1905683004

Review URL: https://codereview.chromium.org/1905683004

8 years agoUpdate min Vulkan version to 1.0.8.0, and fix various bugs
egdaniel [Thu, 21 Apr 2016 15:03:10 +0000 (08:03 -0700)]
Update min Vulkan version to 1.0.8.0, and fix various bugs

With updating the SDK, the debug layers also showed multiple bugs.
I have fixed those as well in this CL. These include:

1. Incorrectly tracking the allocated descriptor sets from the descriptor pools

2. Using MemoryBarriers inside render passes.

3. Correctly setting the Stencil Image layout anytime we are using a render pass with a stencil attachment

4. Setting the correct aspect mask for Depth/Stencil in a barrier.

TBR=bsalomon@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1906623002

Review URL: https://codereview.chromium.org/1906623002

8 years agoVulkanViewer on Android
djsollen [Thu, 21 Apr 2016 14:59:44 +0000 (07:59 -0700)]
VulkanViewer on Android
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903253003

Review URL: https://codereview.chromium.org/1903253003

8 years agoFix data parameter for createBuffer
cdalton [Thu, 21 Apr 2016 14:41:50 +0000 (07:41 -0700)]
Fix data parameter for createBuffer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1911703002

Review URL: https://codereview.chromium.org/1911703002

8 years agoCall SkGraphics::Init() automatically for Google3 processes.
mtklein [Thu, 21 Apr 2016 14:01:19 +0000 (07:01 -0700)]
Call SkGraphics::Init() automatically for Google3 processes.

This is logically a partial revert of https://codereview.chromium.org/1903503002.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1906813002

Review URL: https://codereview.chromium.org/1906813002

8 years agoImage filters: remove SkBitmap-based cache.
senorblanco [Thu, 21 Apr 2016 13:49:15 +0000 (06:49 -0700)]
Image filters: remove SkBitmap-based cache.

Also some #include cleanup.

TBR=reed@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1905753002

Review URL: https://codereview.chromium.org/1905753002

8 years agoRemove stray files
robertphillips [Thu, 21 Apr 2016 12:17:28 +0000 (05:17 -0700)]
Remove stray files

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1912583002

Review URL: https://codereview.chromium.org/1912583002

8 years agoRevert of Hide SkCanvas::LayerIter (patchset #2 id:20001 of https://codereview.chromi...
fmalita [Thu, 21 Apr 2016 00:03:38 +0000 (17:03 -0700)]
Revert of Hide SkCanvas::LayerIter (patchset #2 id:20001 of https://codereview.chromium.org/1897863008/ )

Reason for revert:
Das roll ist gebusted: https://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_ng/builds/214775/steps/compile%20%28with%20patch%29/logs/stdio

FAILED: /b/build/slave/cache/cipd/goma/gomacc ../../third_party/llvm-build/Release+Asserts/bin/clang++ -MMD -MF obj/third_party/WebKit/Source/platform/graphics/paint/blink_platform_unittests.PaintArtifactToSkCanvasTest.o.d -DV8_DEPRECATION_WARNINGS -DCLD_VERSION=2 -D__ASSERT_MACROS_DEFINE_VERSIONS_WITHOUT_UNDERSCORE=0 -DCHROMIUM_BUILD -DCR_CLANG_REVISION=266460-1 -DUSE_LIBJPEG_TURBO=1 -DENABLE_WEBRTC=1 -DENABLE_MEDIA_ROUTER=1 -DUSE_PROPRIETARY_CODECS -DENABLE_PEPPER_CDMS -DENABLE_NOTIFICATIONS -DENABLE_TOPCHROME_MD=1 -DDCHECK_ALWAYS_ON=1 -DFIELDTRIAL_TESTING_ENABLED -DENABLE_TASK_MANAGER=1 -DENABLE_EXTENSIONS=1 -DENABLE_PDF=1 -DENABLE_PLUGIN_INSTALLATION=1 -DENABLE_PLUGINS=1 -DENABLE_SESSION_SERVICE=1 -DENABLE_THEMES=1 -DENABLE_AUTOFILL_DIALOG=1 -DENABLE_PRINTING=1 -DENABLE_BASIC_PRINTING=1 -DENABLE_PRINT_PREVIEW=1 -DENABLE_SPELLCHECK=1 -DUSE_BROWSER_SPELLCHECKER=1 -DENABLE_CAPTIVE_PORTAL_DETECTION=1 -DENABLE_APP_LIST=1 -DENABLE_SETTINGS_APP=1 -DENABLE_SUPERVISED_USERS=1 -DENABLE_SERVICE_DISCOVERY=1 -DV8_USE_EXTERNAL_STARTUP_DATA -DFULL_SAFE_BROWSING -DSAFE_BROWSING_CSD -DSAFE_BROWSING_DB_LOCAL -DINSIDE_BLINK -DMOJO_USE_SYSTEM_IMPL -DGTEST_HAS_POSIX_RE=0 -DGTEST_LANG_CXX11=0 -DUNIT_TEST -DGTEST_HAS_RTTI=0 -DENABLE_LAYOUT_UNIT_IN_INLINE_BOXES=0 -DENABLE_OILPAN=1 -DWTF_USE_CONCATENATED_IMPULSE_RESPONSES=1 -DENABLE_INPUT_MULTIPLE_FIELDS_UI=1 -DWTF_USE_ICCJPEG=1 -DWTF_USE_QCMSLIB=1 -DU_USING_ICU_NAMESPACE=0 -DU_ENABLE_DYLOAD=0 -DU_NOEXCEPT= -DU_STATIC_IMPLEMENTATION -DSK_SUPPORT_GPU=1 -DSK_IGNORE_LINEONLY_AA_CONVEX_PATH_OPTS -DCHROME_PNG_WRITE_SUPPORT -DPNG_USER_CONFIG -DUSE_LIBPCI=1 -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -DNDEBUG -DNVALGRIND -DDYNAMIC_ANNOTATIONS_ENABLED=0 -D_FORTIFY_SOURCE=2 -Igen -Igen/blink -I../.. -I../../skia/config -I../../third_party/WebKit/Source -I../../third_party/khronos -I../../gpu -Igen/angle -I../../third_party/WebKit -Igen/third_party/WebKit -I../../testing/gmock/include -I../../testing/gtest/include -I../../third_party/icu/source/i18n -I../../third_party/icu/source/common -I../../third_party/skia/include/core -I../../third_party/skia/include/effects -I../../third_party/skia/include/pdf -I../../third_party/skia/include/gpu -I../../third_party/skia/include/lazy -I../../third_party/skia/include/pathops -I../../third_party/skia/include/pipe -I../../third_party/skia/include/ports -I../../third_party/skia/include/utils -I../../third_party/skia/include/utils/mac -I../../skia/ext -I../../third_party/harfbuzz-ng/src -I../../third_party/qcms/src -I../../third_party/libpng -I../../third_party/zlib -I../../third_party/libwebp -I../../third_party/ots/include -I../../v8/include -I../../third_party/iccjpeg -I../../third_party/libjpeg_turbo -isysroot /Applications/Xcode5.1.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.10.sdk -O2 -fvisibility=hidden -Werror -mmacosx-version-min=10.7 -arch x86_64 -Wall -Wextra -Wno-unused-parameter -Wno-missing-field-initializers -Wno-selector-type-mismatch -Wpartial-availability -Wheader-hygiene -Wno-char-subscripts -Wno-unneeded-internal-declaration -Wno-covered-switch-default -Wstring-conversion -Wno-c++11-narrowing -Wno-deprecated-register -Wno-inconsistent-missing-override -Wno-shift-negative-value -std=c++11 -stdlib=libc++ -fno-rtti -fno-exceptions -fvisibility-inlines-hidden -fno-threadsafe-statics -Xclang -load -Xclang /b/build/slave/mac/build/src/third_party/llvm-build/Release+Asserts/lib/libFindBadConstructs.dylib -Xclang -add-plugin -Xclang find-bad-constructs -Xclang -plugin-arg-find-bad-constructs -Xclang check-templates -Xclang -plugin-arg-find-bad-constructs -Xclang follow-macro-expansion -Xclang -plugin-arg-find-bad-constructs -Xclang check-implicit-copy-ctors -fcolor-diagnostics -fno-strict-aliasing -Xclang -load -Xclang /b/build/slave/mac/build/src/third_party/llvm-build/Release+Asserts/lib/libBlinkGCPlugin.dylib -Xclang -add-plugin -Xclang blink-gc-plugin -Xclang -plugin-arg-blink-gc-plugin -Xclang enable-oilpan -Xclang -plugin-arg-blink-gc-plugin -Xclang warn-raw-ptr  -c ../../third_party/WebKit/Source/platform/graphics/paint/PaintArtifactToSkCanvasTest.cpp -o obj/third_party/WebKit/Source/platform/graphics/paint/blink_platform_unittests.PaintArtifactToSkCanvasTest.o
../../third_party/WebKit/Source/platform/graphics/paint/PaintArtifactToSkCanvasTest.cpp:201:22: error: 'LayerIter' is a protected member of 'SkCanvas'
    return SkCanvas::LayerIter(canvas, false).clip();
                     ^
../../third_party/skia/include/core/SkCanvas.h:1230:18: note: declared protected here
    class SK_API LayerIter /*: SkNoncopyable*/ {
                 ^
1 error generated.
ninja: build stopped: subcommand failed.

Original issue's description:
> Hide SkCanvas::LayerIter
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897863008
>
> Committed: https://skia.googlesource.com/skia/+/a814000cbb0e90624a410b4359fff7f3ec66f0e0

TBR=reed@google.com,robertphillips@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/1906573003

8 years agopost apply non-scale transforms after imagefilters have run
reed [Wed, 20 Apr 2016 23:36:51 +0000 (16:36 -0700)]
post apply non-scale transforms after imagefilters have run

may choose to eliminate the final matrix-filter buffer before the sprite blit, but at the moment want to defer that change to a 2nd CL.

heavily inspired by https://codereview.chromium.org/1140943004

Need these CLs to land first:
https://codereview.chromium.org/1898193005/#
https://codereview.chromium.org/1902253003/

BUG=skia:3288
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1899263002

Review URL: https://codereview.chromium.org/1899263002

8 years agoHide SkCanvas::LayerIter
robertphillips [Wed, 20 Apr 2016 22:57:27 +0000 (15:57 -0700)]
Hide SkCanvas::LayerIter

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897863008

Review URL: https://codereview.chromium.org/1897863008

8 years agoMake all the codecs default profiles based on gTreatSkColorAsSRGB.
herb [Wed, 20 Apr 2016 22:07:32 +0000 (15:07 -0700)]
Make all the codecs default profiles based on gTreatSkColorAsSRGB.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1893203006

Review URL: https://codereview.chromium.org/1893203006

8 years agoSkOnce: 2 bytes -> 1 byte
mtklein [Wed, 20 Apr 2016 20:49:15 +0000 (13:49 -0700)]
SkOnce: 2 bytes -> 1 byte

This uses the same logic we worked out for SkOncePtr to reduce
the memory footprint of SkOnce from a done byte and lock byte
to a single 3-state byte:

  - NotStarted: no thread has tried to run fn() yet
  - Active:     a thread is running fn()
  - Done:       fn() is complete

Threads which see Done return immediately.
Threads which see NotStarted try to move to Active, run fn(), then move to Done.
Threads which see Active spin until the active thread moves to Done.

This additionally fixes a too-weak memory order bug in SkOncePtr,
and adds a big note to explain.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1904483003

Committed: https://skia.googlesource.com/skia/+/df02d338be8e3c1c50b48a3a9faa582703a39c07

Review URL: https://codereview.chromium.org/1904483003

8 years agoRevert of SkOnce: 2 bytes -> 1 byte (patchset #4 id:60001 of https://codereview.chrom...
mtklein [Wed, 20 Apr 2016 20:02:08 +0000 (13:02 -0700)]
Revert of SkOnce: 2 bytes -> 1 byte (patchset #4 id:60001 of https://codereview.chromium.org/1904483003/ )

Reason for revert:
bust the roll

Original issue's description:
> SkOnce: 2 bytes -> 1 byte
>
> This uses the same logic we worked out for SkOncePtr to reduce
> the memory footprint of SkOnce from a done byte and lock byte
> to a single 3-state byte:
>
>   - NotStarted: no thread has tried to run fn() yet
>   - Active:     a thread is running fn()
>   - Done:       fn() is complete
>
> Threads which see Done return immediately.
> Threads which see NotStarted try to move to Active, run fn(), then move to Done.
> Threads which see Active spin until the active thread moves to Done.
>
> This additionally fixes a too-weak memory order bug in SkOncePtr,
> and adds a big note to explain.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1904483003
>
> Committed: https://skia.googlesource.com/skia/+/df02d338be8e3c1c50b48a3a9faa582703a39c07

TBR=herb@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/1898413004

8 years agoAdd SkEncodedInfo to report properties of encoded image data
msarett [Wed, 20 Apr 2016 18:53:35 +0000 (11:53 -0700)]
Add SkEncodedInfo to report properties of encoded image data

All this does is build an SkEncodedInfo for each codec, and
then convert it to an SkImageInfo.

In future steps I intend to:
(1) Use SkEncodedInfo in place of SrcConfig in SkSwizzler.
(2) Support more conversions in SkSwizzler (non-native
BGRA/RGBA, 16-bit components, float, fixed point)
(3) Investigate optimizing conversions from encoded data
to linear color spaces.

BUG=skia:4133
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1820073002

Committed: https://skia.googlesource.com/skia/+/f682d9ad70d690a343bc15e26ef321d86770be41

Review URL: https://codereview.chromium.org/1820073002

8 years agoBegin removing deprecated (and now, unused) ImageFilter code paths
robertphillips [Wed, 20 Apr 2016 18:48:36 +0000 (11:48 -0700)]
Begin removing deprecated (and now, unused) ImageFilter code paths

This CL focuses on removing the Proxy objects but takes some other stuff with it.

BUG=skia:4965
TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1896383003

Review URL: https://codereview.chromium.org/1896383003

8 years agoFix ImageFilter fuzzer issue
robertphillips [Wed, 20 Apr 2016 18:43:33 +0000 (11:43 -0700)]
Fix ImageFilter fuzzer issue

What appears to be happening in this fuzz is that a paint index inside the picture of an SkPictureImageFilter is getting changed to be out of range.

BUG=skia:5192
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1893423002

Review URL: https://codereview.chromium.org/1893423002

8 years agoSkOnce: 2 bytes -> 1 byte
mtklein [Wed, 20 Apr 2016 17:54:54 +0000 (10:54 -0700)]
SkOnce: 2 bytes -> 1 byte

This uses the same logic we worked out for SkOncePtr to reduce
the memory footprint of SkOnce from a done byte and lock byte
to a single 3-state byte:

  - NotStarted: no thread has tried to run fn() yet
  - Active:     a thread is running fn()
  - Done:       fn() is complete

Threads which see Done return immediately.
Threads which see NotStarted try to move to Active, run fn(), then move to Done.
Threads which see Active spin until the active thread moves to Done.

This additionally fixes a too-weak memory order bug in SkOncePtr,
and adds a big note to explain.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1904483003

Review URL: https://codereview.chromium.org/1904483003

8 years agoGet skiaserve working on Windows.
brianosman [Wed, 20 Apr 2016 17:52:54 +0000 (10:52 -0700)]
Get skiaserve working on Windows.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903203003

Review URL: https://codereview.chromium.org/1903203003

8 years agoRevert of Use transfer buffer for BatchAtlas texture copies (patchset #9 id:160001...
jvanverth [Wed, 20 Apr 2016 17:33:27 +0000 (10:33 -0700)]
Revert of Use transfer buffer for BatchAtlas texture copies (patchset #9 id:160001 of https://codereview.chromium.org/1888473002/ )

Reason for revert:
Lots of sync issues with paths and text.

Original issue's description:
> Use transfer buffer for BatchAtlas texture copies.
>
> Sets up use of transfer buffer (if available) to do one-copy transfers.
> Get transfer buffers working properly in GL.
> Implement GrVkGpu::onTransferPixels.
> Check caps to ensure we can create a transfer buffer.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1888473002
>
> Committed: https://skia.googlesource.com/skia/+/b0ec9836dbf7f2304a3a29289b818719ca0a39bd

TBR=robertphillips@google.com,egdaniel@google.com,bsalomon@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/1904723003

8 years agoSkTArray movable and swap for move only elements.
bungeman [Wed, 20 Apr 2016 17:22:20 +0000 (10:22 -0700)]
SkTArray movable and swap for move only elements.

SkTArray cannot currently contain move only elements because its swap
currently requires the SkTArray to be copyable. This makes SkTArray
movable and makes its swap move instead of copy.

Review URL: https://codereview.chromium.org/1904663004

8 years agoAdd control of the sRGB global hack flag in SampleApp
brianosman [Wed, 20 Apr 2016 16:20:46 +0000 (09:20 -0700)]
Add control of the sRGB global hack flag in SampleApp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1893313006

Review URL: https://codereview.chromium.org/1893313006

8 years agogyp: add osmesa target to factor out common code
halcanary [Wed, 20 Apr 2016 15:49:53 +0000 (08:49 -0700)]
gyp: add osmesa target to factor out common code

Test this out with `GYP_DEFINES=skia_mesa=1`.

Also uses the `library_dirs` keyword, which is relatively new.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903233002

Review URL: https://codereview.chromium.org/1903233002

8 years agoSkFontHost/FreeType/ToUnicode: add an assert
halcanary [Wed, 20 Apr 2016 15:37:43 +0000 (08:37 -0700)]
SkFontHost/FreeType/ToUnicode: add an assert

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1900193002

Review URL: https://codereview.chromium.org/1900193002

8 years agoAdd instructions for building with MSAN
scroggo [Wed, 20 Apr 2016 15:27:18 +0000 (08:27 -0700)]
Add instructions for building with MSAN

Move static functions only used by GPU into if SK_SUPPORT_GPU guard.
This enables building MSAN, which seems to default to not supporting
GPU.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1908493002
DOCS_PREVIEW= https://skia.org/?cl=1908493002

Review URL: https://codereview.chromium.org/1908493002

8 years agoRefactor how we store and use samplers in Ganesh
egdaniel [Wed, 20 Apr 2016 14:09:46 +0000 (07:09 -0700)]
Refactor how we store and use samplers in Ganesh

The main goal of this refactorization is to allow Vulkan to use separate
sampler and texture objects in the shader and descriptor sets and combine
them into a sampler2d in the shader where needed.

A large part of this is separating how we store samplers and uniforms in the
UniformHandler. We no longer need to store handles to samplers besides when
we are initially emitting code. After we emit code all we ever do is loop over
all samplers and do some processor independent work on them, so we have no need
for direct access to individual samplers.

In the GLProgram all we ever do is set the sampler uniforms in the ctor and never
touch them again, so no need to save sampler info there. The texture access on
program reuse just assume that they come in the same order as we set the texture
units for the samplers

For Vulkan, it is a similar story. We create the descriptor set layouts with the samplers,
then when we get new textures, we just assume they come in in the same order as we
set the samplers on the descriptor sets. Thus no need to save direct vulkan info.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1885863004

Committed: https://skia.googlesource.com/skia/+/45b61a1c4c0be896e7b12fd1405abfece799114f

Review URL: https://codereview.chromium.org/1885863004

8 years agoUse transfer buffer for BatchAtlas texture copies.
jvanverth [Wed, 20 Apr 2016 12:54:01 +0000 (05:54 -0700)]
Use transfer buffer for BatchAtlas texture copies.

Sets up use of transfer buffer (if available) to do one-copy transfers.
Get transfer buffers working properly in GL.
Implement GrVkGpu::onTransferPixels.
Check caps to ensure we can create a transfer buffer.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1888473002

Review URL: https://codereview.chromium.org/1888473002

8 years agoRein in the use of SkDevice.h
robertphillips [Wed, 20 Apr 2016 12:29:51 +0000 (05:29 -0700)]
Rein in the use of SkDevice.h

Nothing interesting.

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1904673003

Review URL: https://codereview.chromium.org/1904673003

8 years agoEnable NV_path_rendering in the Null GL.
svaisanen [Wed, 20 Apr 2016 07:36:53 +0000 (00:36 -0700)]
Enable NV_path_rendering in the Null GL.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1882563002

Review URL: https://codereview.chromium.org/1882563002

8 years agoFix DM blaze build (PNG output is now dependent on picture_utils)
brianosman [Wed, 20 Apr 2016 02:41:54 +0000 (19:41 -0700)]
Fix DM blaze build (PNG output is now dependent on picture_utils)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1902333002

Review URL: https://codereview.chromium.org/1902333002

8 years agoRevert of Refactor how we store and use samplers in Ganesh (patchset #7 id:120001...
egdaniel [Tue, 19 Apr 2016 22:24:29 +0000 (15:24 -0700)]
Revert of Refactor how we store and use samplers in Ganesh (patchset #7 id:120001 of https://codereview.chromium.org/1885863004/ )

Reason for revert:
breaking bots

Original issue's description:
> Refactor how we store and use samplers in Ganesh
>
> The main goal of this refactorization is to allow Vulkan to use separate
> sampler and texture objects in the shader and descriptor sets and combine
> them into a sampler2d in the shader where needed.
>
> A large part of this is separating how we store samplers and uniforms in the
> UniformHandler. We no longer need to store handles to samplers besides when
> we are initially emitting code. After we emit code all we ever do is loop over
> all samplers and do some processor independent work on them, so we have no need
> for direct access to individual samplers.
>
> In the GLProgram all we ever do is set the sampler uniforms in the ctor and never
> touch them again, so no need to save sampler info there. The texture access on
> program reuse just assume that they come in the same order as we set the texture
> units for the samplers
>
> For Vulkan, it is a similar story. We create the descriptor set layouts with the samplers,
> then when we get new textures, we just assume they come in in the same order as we
> set the samplers on the descriptor sets. Thus no need to save direct vulkan info.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1885863004
>
> Committed: https://skia.googlesource.com/skia/+/45b61a1c4c0be896e7b12fd1405abfece799114f

TBR=bsalomon@google.com,jvanverth@google.com,cdalton@nvidia.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/1896013003

8 years agoRefactor how we store and use samplers in Ganesh
egdaniel [Tue, 19 Apr 2016 21:45:57 +0000 (14:45 -0700)]
Refactor how we store and use samplers in Ganesh

The main goal of this refactorization is to allow Vulkan to use separate
sampler and texture objects in the shader and descriptor sets and combine
them into a sampler2d in the shader where needed.

A large part of this is separating how we store samplers and uniforms in the
UniformHandler. We no longer need to store handles to samplers besides when
we are initially emitting code. After we emit code all we ever do is loop over
all samplers and do some processor independent work on them, so we have no need
for direct access to individual samplers.

In the GLProgram all we ever do is set the sampler uniforms in the ctor and never
touch them again, so no need to save sampler info there. The texture access on
program reuse just assume that they come in the same order as we set the texture
units for the samplers

For Vulkan, it is a similar story. We create the descriptor set layouts with the samplers,
then when we get new textures, we just assume they come in in the same order as we
set the samplers on the descriptor sets. Thus no need to save direct vulkan info.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1885863004

Review URL: https://codereview.chromium.org/1885863004

8 years agoAdd explicit clip to SkSpecialSurface's SkCanvas
robertphillips [Tue, 19 Apr 2016 21:23:42 +0000 (14:23 -0700)]
Add explicit clip to SkSpecialSurface's SkCanvas

It seems safer (and is probably more performant) to enforce this internal to SkSpecialSurface.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903553002

Review URL: https://codereview.chromium.org/1903553002

8 years agoskcpu: sse4.1 floor, f16c f16<->f32
mtklein [Tue, 19 Apr 2016 21:21:30 +0000 (14:21 -0700)]
skcpu: sse4.1 floor, f16c f16<->f32

  -  floor with roundps is about 4.5x faster when available
  -  f16 srcover_n is similar to but a little faster than the version in https://codereview.chromium.org/1884683002.  This new one fuses the dst load/stores into the f16<->f32 conversions:

+0x180     movups              (%r15), %xmm1
+0x184     vcvtph2ps           (%rbx), %xmm2
+0x189     movaps              %xmm1, %xmm3
+0x18c     shufps              $255, %xmm3, %xmm3
+0x190     movaps              %xmm0, %xmm4
+0x193     subps               %xmm3, %xmm4
+0x196     mulps               %xmm2, %xmm4
+0x199     addps               %xmm1, %xmm4
+0x19c     vcvtps2ph           $0, %xmm4, (%rbx)
+0x1a2     addq                $16, %r15
+0x1a6     addq                $8, %rbx
+0x1aa     decl                %r14d
+0x1ad     jne                 +0x180

If we decide to land this it'd be a good idea to convert most or all users of SkFloatToHalf_01 and SkHalfToFloat_01 over to the pointer-based versions.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1891513002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Committed: https://skia.googlesource.com/skia/+/cbe3c1af987d622ea67ef560d855b41bb14a0ce9

Committed: https://skia.googlesource.com/skia/+/3faf74b8364491ca806f523fbb1d8a97be592663

Review URL: https://codereview.chromium.org/1891513002

8 years agoadd gm to exercise rotated imagefiltesr w/ and w/o antialiasing
reed [Tue, 19 Apr 2016 21:02:52 +0000 (14:02 -0700)]
add gm to exercise rotated imagefiltesr w/ and w/o antialiasing

BUG=skia:3288
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1902253003

Review URL: https://codereview.chromium.org/1902253003

8 years agoMove CPU feature detection to its own file.
mtklein [Tue, 19 Apr 2016 21:00:13 +0000 (14:00 -0700)]
Move CPU feature detection to its own file.

   - Moves CPU feature detection to its own file.
   - Cleans up some redundant feature detection scattered around core/ and opts/.
   - Can now detect a few new CPU features:
       * F16C     -> Intel f16<->f32 instructions, added between AVX and AVX2
       * FMA      -> Intel FMA instructions, added at the same time as AVX2
       * VFP_FP16 -> ARM f16<->f32 instructions, quite common
       * NEON_FMA -> ARM FMA instructions, also quite common
       * SSE and SSE3... why not?

This new internal API makes it very cheap to do fine-grained runtime CPU
feature detection.  Redundant calls to SkCpu::Supports() should be eliminated
and it's hoistable out of loops.  It compiles away entirely when we have the
appropriate instructions available at compile time.

This means we can call it to guard even a little snippet of 1 or 2 instructions
right where needed and let inlining hoist the check (if any at all) up to
somewhere that doesn't hurt performance.  I've explained how I made this work
in the private section of the new header.

Once this lands and bakes a bit, I'll start following up with CLs to use it more
and to add a bunch of those little 1-2 instruction snippets we've been wanting,
e.g. cvtps2ph, cvtph2ps, ptest, pmulld, pmovzxbd, blendvps, pshufb, roundps
(for floor) on x86, and vcvt.f32.f16, vcvt.f16.f32 on ARM.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1890483002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Committed: https://skia.googlesource.com/skia/+/872ea29357439f05b1f6995dd300fc054733e607

Review URL: https://codereview.chromium.org/1890483002

8 years agoAdding support for controlling the global sRGB SkColor switch.
brianosman [Tue, 19 Apr 2016 19:47:54 +0000 (12:47 -0700)]
Adding support for controlling the global sRGB SkColor switch.

Frontend change: https://codereview.chromium.org/1900233002/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1902143002

Review URL: https://codereview.chromium.org/1902143002

8 years agoRemove static initializer for SkOpts::Init()
mtklein [Tue, 19 Apr 2016 19:42:24 +0000 (12:42 -0700)]
Remove static initializer for SkOpts::Init()

Static initializers run in a confusing unspecified order,
so it's best to have as few of them as possible.

Most tools and clients I can find already call SkGraphics::Init(),
(or equivalently create an SkAutoGraphics) which calls SkOpts::Init():
   - Chrome
   - Chrome renderer
   - Android
   - DM
   - nanobench
   - SampleApp
   - VisualBench
   - the old debugger

Seems like the only thing relying on this static initializer today is
the new debugger, fixed here.

TBR=reed@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903503002

Review URL: https://codereview.chromium.org/1903503002

8 years agoAdd onImageInfo call to SkImage_Base.
herb [Tue, 19 Apr 2016 19:30:22 +0000 (12:30 -0700)]
Add onImageInfo call to SkImage_Base.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903483003

Review URL: https://codereview.chromium.org/1903483003

8 years agoFix createBuffer in Vulkan
egdaniel [Tue, 19 Apr 2016 19:13:41 +0000 (12:13 -0700)]
Fix createBuffer in Vulkan

TBR=bsalomon@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897373002

Review URL: https://codereview.chromium.org/1897373002

8 years agoSwitch SkColorFilterImageFilter over to new onFilterImage interface (again)
robertphillips [Tue, 19 Apr 2016 17:21:02 +0000 (10:21 -0700)]
Switch SkColorFilterImageFilter over to new onFilterImage interface (again)

Back when this was originally reverted I was able to easily repro the perf regression locally. At ToT Skia/Chrome I can no longer repro the perf regression with this CL (in fact there is a modest perf improvement).

I propose landing this and then watching the Chromium perf bots.

BUG=602300,598028
TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1901513002

Review URL: https://codereview.chromium.org/1901513002

8 years agoAdding support for playback to L32/S32/F16 canvas.
brianosman [Tue, 19 Apr 2016 17:16:53 +0000 (10:16 -0700)]
Adding support for playback to L32/S32/F16 canvas.

Playback of my test GM works correctly on both raster and GPU, in all three modes.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1893393002

Review URL: https://codereview.chromium.org/1893393002

8 years agoAdd optional data parameter to createBuffer
cdalton [Tue, 19 Apr 2016 17:00:02 +0000 (10:00 -0700)]
Add optional data parameter to createBuffer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1877073002

Review URL: https://codereview.chromium.org/1877073002

8 years agoMake skpinfo more human friendly
robertphillips [Tue, 19 Apr 2016 16:29:01 +0000 (09:29 -0700)]
Make skpinfo more human friendly
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1901713004

Review URL: https://codereview.chromium.org/1901713004

8 years agoMove DM png code to picture_utils, for use by other tools.
brianosman [Tue, 19 Apr 2016 16:18:11 +0000 (09:18 -0700)]
Move DM png code to picture_utils, for use by other tools.

Planning to re-use this code in skiaserve in an upcoming change.

TBR=mtklein@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1901113002

Review URL: https://codereview.chromium.org/1901113002

8 years agofiddle_build_test/gyp: fix mac mesa build (2)
halcanary [Tue, 19 Apr 2016 15:46:50 +0000 (08:46 -0700)]
fiddle_build_test/gyp: fix mac mesa build (2)

TBR=
NOTRY=true

Review URL: https://codereview.chromium.org/1900173002

8 years agoAdd more diagnostic messages to ReadWriteAlpha test
robertphillips [Tue, 19 Apr 2016 15:32:40 +0000 (08:32 -0700)]
Add more diagnostic messages to ReadWriteAlpha test

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1902063002

Review URL: https://codereview.chromium.org/1902063002

8 years agoRename CodexTest.cpp to CodecTest.cpp
scroggo [Tue, 19 Apr 2016 15:27:39 +0000 (08:27 -0700)]
Rename CodexTest.cpp to CodecTest.cpp

CodexTest confuses me almost every time I try to open it. Renaming it
to what my brain thinks it should be called.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897973004

Review URL: https://codereview.chromium.org/1897973004

8 years agofiddle_build_test/gyp: fix mac mesa build
halcanary [Tue, 19 Apr 2016 15:00:01 +0000 (08:00 -0700)]
fiddle_build_test/gyp: fix mac mesa build

TBR=
NOTRY=true

Review URL: https://codereview.chromium.org/1900143002

8 years agoDelete unused fNamedFactorySet from SkWriteBuffer
msarett [Tue, 19 Apr 2016 14:27:58 +0000 (07:27 -0700)]
Delete unused fNamedFactorySet from SkWriteBuffer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1903483002

Review URL: https://codereview.chromium.org/1903483002

8 years agoRevert of Add SkEncodedInfo to report properties of encoded image data (patchset...
robertphillips [Tue, 19 Apr 2016 11:41:12 +0000 (04:41 -0700)]
Revert of Add SkEncodedInfo to report properties of encoded image data (patchset #6 id:200001 of https://codereview.chromium.org/1820073002/ )

Reason for revert:
Lots of bots failing

Original issue's description:
> Add SkEncodedInfo to report properties of encoded image data
>
> All this does is build an SkEncodedInfo for each codec, and
> then convert it to an SkImageInfo.
>
> In future steps I intend to:
> (1) Use SkEncodedInfo in place of SrcConfig in SkSwizzler.
> (2) Support more conversions in SkSwizzler (non-native
> BGRA/RGBA, 16-bit components, float, fixed point)
> (3) Investigate optimizing conversions from encoded data
> to linear color spaces.
>
> BUG=skia:4133
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1820073002
>
> Committed: https://skia.googlesource.com/skia/+/f682d9ad70d690a343bc15e26ef321d86770be41

TBR=scroggo@google.com,reed@google.com,msarett@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4133

Review URL: https://codereview.chromium.org/1895383002

8 years agoAdd SkEncodedInfo to report properties of encoded image data
msarett [Mon, 18 Apr 2016 23:20:00 +0000 (16:20 -0700)]
Add SkEncodedInfo to report properties of encoded image data

All this does is build an SkEncodedInfo for each codec, and
then convert it to an SkImageInfo.

In future steps I intend to:
(1) Use SkEncodedInfo in place of SrcConfig in SkSwizzler.
(2) Support more conversions in SkSwizzler (non-native
BGRA/RGBA, 16-bit components, float, fixed point)
(3) Investigate optimizing conversions from encoded data
to linear color spaces.

BUG=skia:4133
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1820073002

Review URL: https://codereview.chromium.org/1820073002

8 years agoMake SkImageFilter::Cache purging methods pure virtual
robertphillips [Mon, 18 Apr 2016 21:49:57 +0000 (14:49 -0700)]
Make SkImageFilter::Cache purging methods pure virtual

This is intended to see if the Windows 10 bots ignoring of the purge methods is a compiler bug.

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1901883002

Review URL: https://codereview.chromium.org/1901883002

8 years agoUse libjpeg-turbo for YUV->RGB conversion in jpeg encoder
msarett [Mon, 18 Apr 2016 21:21:55 +0000 (14:21 -0700)]
Use libjpeg-turbo for YUV->RGB conversion in jpeg encoder

This remove the previous codepath, which performed a less
accurate conversion to YUV.

It actually faster to use the more accurate conversion, now
that we have updated to libjpeg-turbo.

z620 Performance Speed-Up:
mandrill.jpg (512x512): 1.13x
color_wheel.jpg (128x128): 1.12x

This should fix:
https://buganizer.corp.google.com/u/0/issues/28161384
https://buganizer.corp.google.com/u/0/issues/21891795

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1886183002

Review URL: https://codereview.chromium.org/1886183002

8 years agoPass through the profile info when creating a high-quality scaled offscreen.
herb [Mon, 18 Apr 2016 21:06:21 +0000 (14:06 -0700)]
Pass through the profile info when creating a high-quality scaled offscreen.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1897043002

Review URL: https://codereview.chromium.org/1897043002

8 years agoUpdate to ToT shaderc
egdaniel [Mon, 18 Apr 2016 20:42:02 +0000 (13:42 -0700)]
Update to ToT shaderc

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1898863002

Review URL: https://codereview.chromium.org/1898863002

8 years agoCleanup android build files and tools
djsollen [Mon, 18 Apr 2016 19:57:35 +0000 (12:57 -0700)]
Cleanup android build files and tools
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1898793004

Review URL: https://codereview.chromium.org/1898793004

8 years agofiddle_build_test depends on skia_mesa
halcanary [Mon, 18 Apr 2016 19:42:05 +0000 (12:42 -0700)]
fiddle_build_test depends on skia_mesa

Builds on Build-Ubuntu-GCC-x86_64-Release-Mesa-Trybot

Shouldn't break anything else.

Review URL: https://codereview.chromium.org/1901823002

8 years agoAdd bench for image encodes
msarett [Mon, 18 Apr 2016 19:37:46 +0000 (12:37 -0700)]
Add bench for image encodes

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1890643003

Review URL: https://codereview.chromium.org/1890643003

8 years agofiddle_build_test: remove from most to fix build
halcanary [Mon, 18 Apr 2016 19:18:30 +0000 (12:18 -0700)]
fiddle_build_test: remove from most to fix build

TBR=jcgregorio@google.com
NOTRY=true

Review URL: https://codereview.chromium.org/1897943002

8 years agoexperimental/fiddle -> tools/fiddle
halcanary [Mon, 18 Apr 2016 18:50:07 +0000 (11:50 -0700)]
experimental/fiddle -> tools/fiddle

also, test building fiddle_main.cpp & draw.cpp
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1895143002

Review URL: https://codereview.chromium.org/1895143002

8 years agoUpdate android toolchain variable to point to root dir and not the bin
djsollen [Mon, 18 Apr 2016 18:01:02 +0000 (11:01 -0700)]
Update android toolchain variable to point to root dir and not the bin
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1900773002

Review URL: https://codereview.chromium.org/1900773002

8 years agoRemove deprecated paths from image filter infrastructure.
senorblanco [Mon, 18 Apr 2016 17:32:02 +0000 (10:32 -0700)]
Remove deprecated paths from image filter infrastructure.

Now that there are no filterImageGPUDeprecated() implementations,
we can being to rip out the deprecated infrastructure.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1888243003

Committed: https://skia.googlesource.com/skia/+/6fb3cd7209849e665635ac17ef4eef4ad63e7f61

Review URL: https://codereview.chromium.org/1888243003

8 years agoSkia Filter Quality and Scaling Metrics
ericrk [Mon, 18 Apr 2016 16:14:00 +0000 (09:14 -0700)]
Skia Filter Quality and Scaling Metrics

Adds histogram metrics to log the filter quality and scale factor of
each image draw. To make the data easier to consume, this is broken
down into a number of individual histograms:
- Filter quality across all draw calls
- Scale amount across all draw calls
- Scale amount per filter quality (4 histograms total)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1686203002

Committed: https://skia.googlesource.com/skia/+/f57b3a6e4a002caf01378832cbd756c6c163a783

Review URL: https://codereview.chromium.org/1686203002

8 years agoexperimental/fiddle: update to new SkImage API
halcanary [Mon, 18 Apr 2016 15:17:56 +0000 (08:17 -0700)]
experimental/fiddle: update to new SkImage API

required by https://crrev.com/1897453002

NOTRY=true
TBR=jcgregorio@google.com

Review URL: https://codereview.chromium.org/1895103002

8 years agoModernize and trim down SkOnce.
mtklein [Mon, 18 Apr 2016 15:09:11 +0000 (08:09 -0700)]
Modernize and trim down SkOnce.

The API and implementation are very much simplified.
You may not want to bother reading the diff.

As is our trend, SkOnce now uses <atomic> directly.

Member initialization means we don't need SK_DECLARE_STATIC_ONCE.
SkSpinlock already works this same way.

All uses of the old API taking an external bool* and Lock* were pessimal,
so I have not carried this sort of API forward.  It's simpler, faster,
and more space-efficient to always use this single SkOnce class interface.

SkOnce weighs 2 bytes: a done bool and an SkSpinlock, also a bool internally.
This API refactoring opens up the opportunity to fuse those into a single
three-state byte if we'd like.

No public API changes.
TBR=reed@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1894893002

Review URL: https://codereview.chromium.org/1894893002

8 years agoSplit AlphaThresholdEffect out into its own file
robertphillips [Mon, 18 Apr 2016 14:20:05 +0000 (07:20 -0700)]
Split AlphaThresholdEffect out into its own file

Mainly mechanical although I did:
   convert Create method to sk_sp Make version
   rm extra '\t's and '\n's in emitCode

TBR=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1893193002

Review URL: https://codereview.chromium.org/1893193002