Reland of Disable layer hoisting for non-8888 canvases (patchset #1 id:1 of https://codereview.chromium.org/1961483002/ )

Reason for revert:
This CL was not the culprit in the perf regression.

Original issue's description:
> Revert of Disable layer hoisting for non-8888 canvases (patchset #2 id:20001 of https://codereview.chromium.org/1957433002/ )
>
> Reason for revert:
> Experimental revert to investigate perf regression
>
> Original issue's description:
> > Disable layer hoisting for non-8888 canvases
> >
> > This just stops the bleeding. A real fix would propagate the canvas' backing type down to the layer hoister.
> >
> > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1957433002
> >
> > Committed: https://skia.googlesource.com/skia/+/4e30f27164179d344f0c8efa9a691d1bc9a53e3f
>
> TBR=bsalomon@google.com
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://skia.googlesource.com/skia/+/6b53b9daeb2137a91d8ddb8e890d66c702bfbaa3

TBR=bsalomon@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.

Review-Url: https://codereview.chromium.org/1962603002

SkAdvancedTypefaceMetrics: abstract out linked list

+ use SkSinglyLinkedList<T>
+ move SkSinglyLinkedList.h to core
+ remove SkHackyAutoTDelete
+ getAdvanceData() -> setGlyphWidths()
+ finishRange no longer templated
+ remove unused templated functions

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953153004

Review-Url: https://codereview.chromium.org/1953153004

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=1955283002

Review-Url: https://codereview.chromium.org/1955283002

documentation: SkMilestone

https://skia.org/schedule?cl=1956473006
NOTRY=true
TBR=

Review-Url: https://codereview.chromium.org/1956473006

Add sRGB mode toggle to Viewer.

Unlike SampleApp, this just switches out the format of the window
surface (and then adjusts the gamma-correct flag on the SkSurfaces).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1950983007

Review-Url: https://codereview.chromium.org/1950983007

Revert of Simplify SkGpuBlurUtils::GaussianBlur method (patchset #2 id:20001 of https://codereview.chromium.org/1958603002/ )

Reason for revert:
Looks like it's causing some issues with the bleed_image GM.

Original issue's description:
> Simplify SkGpuBlurUtils::GaussianBlur method
>
> No one was using the canClobberSrc capability and moving the direct filtering case forward makes the rest of the logic simpler.
>
> Split out of: https://codereview.chromium.org/1959493002/ (Retract GrRenderTarget from SkGpuBlurUtils)
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1958603002
>
> Committed: https://skia.googlesource.com/skia/+/56a85e69a8d034e0fdee00e8207cda0a9da06fee

TBR=bsalomon@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/1956023002

Clean up GrVkImage::setLayout()

Fixes issues with mipmap generation, and calculates src information
inside the routine rather than having the client do it.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1950713006

Review-Url: https://codereview.chromium.org/1950713006

Remove Sk(Float|Double)PinToFixed functions, only used in tests.

These were triggering compile-time floating point overflow for one user,
and we don't use them.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1952303005

Review-Url: https://codereview.chromium.org/1952303005

Fix issue where GrStyle::applyToPath exited before applying stroke
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1954123002

Review-Url: https://codereview.chromium.org/1954123002

Fix for Sk4f Ubuntu compile warning

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1958633003
TBR=herbderby@google.com

Review-Url: https://codereview.chromium.org/1958633003

Remove hasMixedSamples() from GrPipelineBuilder

This info is unknown until after any clip has been applied to the
stencil settings, so it's misleading to include in the builder.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1952323002

Review-Url: https://codereview.chromium.org/1952323002

Fix for Ubuntu compile warning

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1955913002

Review-Url: https://codereview.chromium.org/1955913002

Disable skia_warnings_as_errors for Fast bot.

BUG=skia:5257
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1955923002

Review-Url: https://codereview.chromium.org/1955923002

Detect an invalid intervalCount in SkRegion during deserialiation.

BUG=609260
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1961463003

Review-Url: https://codereview.chromium.org/1961463003

SkAdvancedTypefaceMetrics: fail cleanly.

BUG=567031
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1956753002

Review-Url: https://codereview.chromium.org/1956753002

https://groups.google.com/forum/#!topic/skia-discuss/2F2she2nQMg

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

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1957503002

Review-Url: https://codereview.chromium.org/1957503002

Revert of Retract GrRenderTarget a bit within SkGpuDevice (patchset #2 id:20001 of https://codereview.chromium.org/1956473002/ )

Reason for revert:
Experimental revert to track down perf regression

Original issue's description:
> Retract GrRenderTarget a bit within SkGpuDevice
>
> Split off of: https://codereview.chromium.org/1930013002/ ((Mostly) Retract GrRenderTarget from SkGpuDevice)
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1956473002
>
> Committed: https://skia.googlesource.com/skia/+/d215a95882485026d176a81bf58871276b21c89c

TBR=bsalomon@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/1954103002

SkAdvancedTypefaceMetrics: improve robustness

Manually unwind linked list.

BUG=567031
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1955033002

Review-Url: https://codereview.chromium.org/1955033002

Take SkStrokeRec::InitStyle rather than SkPaint::Style in mask filter and DrawMask
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1955633002

Review-Url: https://codereview.chromium.org/1955633002

Compile SkForceLinking on CMake

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1956573002
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/1956573002

Simplify SkGpuBlurUtils::GaussianBlur method

No one was using the canClobberSrc capability and moving the direct filtering case forward makes the rest of the logic simpler.

Split out of: https://codereview.chromium.org/1959493002/ (Retract GrRenderTarget from SkGpuBlurUtils)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1958603002

Review-Url: https://codereview.chromium.org/1958603002

Revert of Disable layer hoisting for non-8888 canvases (patchset #2 id:20001 of https://codereview.chromium.org/1957433002/ )

Reason for revert:
Experimental revert to investigate perf regression

Original issue's description:
> Disable layer hoisting for non-8888 canvases
>
> This just stops the bleeding. A real fix would propagate the canvas' backing type down to the layer hoister.
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1957433002
>
> Committed: https://skia.googlesource.com/skia/+/4e30f27164179d344f0c8efa9a691d1bc9a53e3f

TBR=bsalomon@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/1961483002

Whitespace change to test bots

BUG=skia:4763
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1956713002
NOTREECHECKS=true

Review-Url: https://codereview.chromium.org/1956713002

remove alias for SkMatrix44.h -- chrome callers have been updated

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949113006

TBR=

Review-Url: https://codereview.chromium.org/1949113006

SkOncePtr -> SkOnce

It's always nice to kill off a synchronization primitive.
And while less terse, I think this new code reads more clearly.

... and, SkOncePtr's tests were the only thing now using sk_num_cores()
outside of SkTaskGroup, so I've hidden it as static inside SkTaskGroup.cpp.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953533002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Committed: https://skia.googlesource.com/skia/+/9bd3fc7fa9bb7cad050bd619aa93d4c48ebb5c02

Review-Url: https://codereview.chromium.org/1953533002

Fix unit test crash caused by OSMesa not supporting rendering to GL_ALPHA8
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947373003

Review-Url: https://codereview.chromium.org/1947373003

Revert of Modify LineBench for drawing straight line (patchset #3 id:40001 of https://codereview.chromium.org/1936153002/ )

Reason for revert:
         for (int i = 0; i < loops; i++) {
 54             canvas->drawPoints(SkCanvas::kLines_PointMode, PTS, fPts, paint);  68             canvas->drawLine(fStartPts[i].x(), fStartPts[i].y(), fEndPts[i].x(), fEndPts[i].y(), pai
    nt);

This change means we index arbitrarily far into fStartPts (if loops gets big enough).

Original issue's description:
> Modify LineBench for drawing
>
> Currently we only have benchmark for lines that contains randomly generated
> points. This CL modifies the benchmark which adds cases of drawing straight
> lines. Also, this CL changes the call from drawPoints() to drawLines()
> which measures the performance of line drawing.
>
> BUG=skia:5243
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1936153002
>
> Committed: https://skia.googlesource.com/skia/+/6b27a5e7292d9a18e376f0c229ec62f7b801305a

TBR=bsalomon@chromium.org,robertphillips@chromium.org,robertphillips@google.com,xidachen@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:5243

Review-Url: https://codereview.chromium.org/1952063005

Modify LineBench for drawing

Currently we only have benchmark for lines that contains randomly generated
points. This CL modifies the benchmark which adds cases of drawing straight
lines. Also, this CL changes the call from drawPoints() to drawLines()
which measures the performance of line drawing.

BUG=skia:5243
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1936153002

Review-Url: https://codereview.chromium.org/1936153002

Disable layer hoisting for non-8888 canvases

This just stops the bleeding. A real fix would propagate the canvas' backing type down to the layer hoister.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1957433002

Review-Url: https://codereview.chromium.org/1957433002

More refactoring for Viewer

* Move support files into sk_app and main files up to top directory
* Rename VulkanTestContext and create WindowContext parent class
* Place VulkanWindowContext et al. in sk_app namespace.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1944413005

Review-Url: https://codereview.chromium.org/1944413005

Added --deepColor option to SampleApp, triggers creation of a ten-bit/channel
buffer. (Only on Windows at the moment). Uses new effect to do the final
gamma adjustment

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1919993002

Review-Url: https://codereview.chromium.org/1919993002

Retract GrRenderTarget a bit within SkGpuDevice

Split off of: https://codereview.chromium.org/1930013002/ ((Mostly) Retract GrRenderTarget from SkGpuDevice)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1956473002

Review-Url: https://codereview.chromium.org/1956473002

Bring back sRGB-write-control as a caps bit.

We're not going to support this on Vulkan (yet), but want to still
have it as an option for GL. (The fallback alternative for our YUV use
case involves injecting expensive fragment shader code, which we'd like
to avoid).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1955603003

Review-Url: https://codereview.chromium.org/1955603003

experimental/tools/coreGraphicsPdf2png: remove skia dependency

// experimental-only
TBR=
NOTRY=true

Review-Url: https://codereview.chromium.org/1950353002

Update SK_IMAGE_VERSION to 5

This is should allow us to begin seeing colorspace tests on Gold.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1951393002

Review-Url: https://codereview.chromium.org/1951393002

Remove NEON runtime detection support.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1952953004
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/1952953004

Modernize SkMutex and SkSemaphore.

   - use <atomic>
   - fuse SkMutex and SkBaseMutex
   - fuse SkSemaphore and SkBaseSemaphore

Still TODO:
   - replace SK_DECLARE_STATIC_MUTEX(name) with static SkMutex name

I just didn't want to bother fixing all that up until I know this CL sticks.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947153002

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

Committed: https://skia.googlesource.com/skia/+/427c2819d9237d7d7729c59238036cfc73c072ea

Review-Url: https://codereview.chromium.org/1947153002

Make GrPathRenderer only support fills for path stenciling

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1950893002

Review-Url: https://codereview.chromium.org/1950893002

Stop banning stroke-and-fill for GrStyle and add tests

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949943003

Review-Url: https://codereview.chromium.org/1949943003

Revert of SkOncePtr -> SkOnce (patchset #1 id:1 of https://codereview.chromium.org/1949313003/ )

Reason for revert:
ok, I have no idea what's going on

Original issue's description:
> Reland of SkOncePtr -> SkOnce (patchset #1 id:1 of https://codereview.chromium.org/1945293004/ )
>
> Reason for revert:
> previous revert unneeded (I think)
>
> Original issue's description:
> > Revert of SkOncePtr -> SkOnce (patchset #4 id:60001 of https://codereview.chromium.org/1953533002/ )
> >
> > Reason for revert:
> > broken the Mac and Linux builders, e.g.:
> >
> > https://build.chromium.org/p/chromium/builders/Mac/builds/15151
> > https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052
> >
> > Original issue's description:
> > > SkOncePtr -> SkOnce
> > >
> > > It's always nice to kill off a synchronization primitive.
> > > And while less terse, I think this new code reads more clearly.
> > >
> > > ... and, SkOncePtr's tests were the only thing now using sk_num_cores()
> > > outside of SkTaskGroup, so I've hidden it as static inside SkTaskGroup.cpp.
> > >
> > > BUG=skia:
> > > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953533002
> > > CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
> > >
> > > Committed: https://skia.googlesource.com/skia/+/9bd3fc7fa9bb7cad050bd619aa93d4c48ebb5c02
> >
> > TBR=herb@google.com,mtklein@chromium.org
> > # Skipping CQ checks because original CL landed less than 1 days ago.
> > NOPRESUBMIT=true
> > NOTREECHECKS=true
> > NOTRY=true
> > BUG=skia:
> >
> > Committed: https://skia.googlesource.com/skia/+/7eb33da7eede34c050b865dbb1b60c3dcea7191b
>
> TBR=herb@google.com,mtklein@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/82da9a8aa0ce648f36882830765b42e0ada6c0fa

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

Review-Url: https://codereview.chromium.org/1948313002

Reland of SkOncePtr -> SkOnce (patchset #1 id:1 of https://codereview.chromium.org/1945293004/ )

Reason for revert:
previous revert unneeded (I think)

Original issue's description:
> Revert of SkOncePtr -> SkOnce (patchset #4 id:60001 of https://codereview.chromium.org/1953533002/ )
>
> Reason for revert:
> broken the Mac and Linux builders, e.g.:
>
> https://build.chromium.org/p/chromium/builders/Mac/builds/15151
> https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052
>
> Original issue's description:
> > SkOncePtr -> SkOnce
> >
> > It's always nice to kill off a synchronization primitive.
> > And while less terse, I think this new code reads more clearly.
> >
> > ... and, SkOncePtr's tests were the only thing now using sk_num_cores()
> > outside of SkTaskGroup, so I've hidden it as static inside SkTaskGroup.cpp.
> >
> > BUG=skia:
> > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953533002
> > CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
> >
> > Committed: https://skia.googlesource.com/skia/+/9bd3fc7fa9bb7cad050bd619aa93d4c48ebb5c02
>
> TBR=herb@google.com,mtklein@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/7eb33da7eede34c050b865dbb1b60c3dcea7191b

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

Review-Url: https://codereview.chromium.org/1949313003

Revert of Modernize SkMutex and SkSemaphore. (patchset #2 id:20001 of https://codereview.chromium.org/1947153002/ )

Reason for revert:
This CL seems to have broken the Linux x64 and Mac
bots, e.g.

https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052
https://build.chromium.org/p/chromium/builders/Mac/builds/15151

The error appears to have something to do with new static initializers being
added.

Original issue's description:
> Modernize SkMutex and SkSemaphore.
>
>    - use <atomic>
>    - fuse SkMutex and SkBaseMutex
>    - fuse SkSemaphore and SkBaseSemaphore
>
> Still TODO:
>    - replace SK_DECLARE_STATIC_MUTEX(name) with static SkMutex name
>
> I just didn't want to bother fixing all that up until I know this CL sticks.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947153002
>
> No public API changes.
> TBR=reed@google.com
>
> Committed: https://skia.googlesource.com/skia/+/427c2819d9237d7d7729c59238036cfc73c072ea

TBR=herb@google.com,mtklein@chromium.org,reed@google.com,bsalomon@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=609340

Review-Url: https://codereview.chromium.org/1945343003

Revert of SkOncePtr -> SkOnce (patchset #4 id:60001 of https://codereview.chromium.org/1953533002/ )

Reason for revert:
broken the Mac and Linux builders, e.g.:

https://build.chromium.org/p/chromium/builders/Mac/builds/15151
https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052

Original issue's description:
> SkOncePtr -> SkOnce
>
> It's always nice to kill off a synchronization primitive.
> And while less terse, I think this new code reads more clearly.
>
> ... and, SkOncePtr's tests were the only thing now using sk_num_cores()
> outside of SkTaskGroup, so I've hidden it as static inside SkTaskGroup.cpp.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953533002
> CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/9bd3fc7fa9bb7cad050bd619aa93d4c48ebb5c02

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

Review-Url: https://codereview.chromium.org/1945293004

Revert of SK_DECLARE_STATIC_MUTEX -> static SkMutex (patchset #1 id:1 of https://codereview.chromium.org/1948193002/ )

Reason for revert:
broken the Mac and Linux builders, e.g.:

https://build.chromium.org/p/chromium/builders/Mac/builds/15151
https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052

Original issue's description:
> SK_DECLARE_STATIC_MUTEX -> static SkMutex
>
> There's no need to use a macro to declare static SkMutexes any more
> (and there's likewise no need to restrict them to global scope).
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1948193002
>
> Committed: https://skia.googlesource.com/skia/+/5e56cfd3fa1041dbb83899844fb92fa9a2ef1009

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

Review-Url: https://codereview.chromium.org/1945353003

SK_DECLARE_STATIC_MUTEX -> static SkMutex

There's no need to use a macro to declare static SkMutexes any more
(and there's likewise no need to restrict them to global scope).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1948193002

Review-Url: https://codereview.chromium.org/1948193002

SkOncePtr -> SkOnce

It's always nice to kill off a synchronization primitive.
And while less terse, I think this new code reads more clearly.

... and, SkOncePtr's tests were the only thing now using sk_num_cores()
outside of SkTaskGroup, so I've hidden it as static inside SkTaskGroup.cpp.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953533002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/1953533002

Now we're on MSVC 2015, try using std::chrono for timing again.

2015's std::chrono::high_resolution_clock should have a precision equal to our old custom code.

Tested this locally on Windows 10 and Mac (the two affected platforms).

Long term SkTime::GetNSecs() can probably be phased out for direct <chrono> use.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1950173002

Review-Url: https://codereview.chromium.org/1950173002

Simplify implementation of SkOnce to not need so many comments.

I think this version reads more clearly, and the key invariants are
expressed in code rather than comments:
   - race losers always go through an acquire
   - we never exit the function unless fState is Done

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1951013004

Review-Url: https://codereview.chromium.org/1951013004

Make cap only affect the keys of GrShapes that are possibly-open

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949193002

Review-Url: https://codereview.chromium.org/1949193002

Rename VulkanViewer to Viewer, take 2.

Also puts Window and Application in sk_app namespace.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1945103003

Review-Url: https://codereview.chromium.org/1945103003

Remove executable bit from source code files.

TBR=reed
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1952773002

Review-Url: https://codereview.chromium.org/1952773002

move colorshader into its own .cpp, add color4f variant

Implemented as a different subclass for SkColorShader (which is also private) partly to make the CL clearer/simpler, and partly for flatten/unflatten compatibility. Later I'm sure we could combine these if that proves useful.

Lots of TODOs at the moment, but still valuable to get reviewed.

Note: this ignores the question (for the moment) about how to interpret SkColor in the larger world. That needs to happen, but this CL is more focused on what *else* to do besides handle the old-style input (and old-style pipeline).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1934313002

Review-Url: https://codereview.chromium.org/1934313002

Fix ICO bug exposed by the fuzzer

BUG=skia:5248
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949773002

Review-Url: https://codereview.chromium.org/1949773002

Add Gr*Proxy classes

This isn't wired in anywhere yet.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1937553002

Committed: https://skia.googlesource.com/skia/+/de5bf0cfeca908b81a28cc50065f7bc2da3d2fd1

Committed: https://skia.googlesource.com/skia/+/92605b35efa0155c44d24bd8415b4cc1db8831db

Review-Url: https://codereview.chromium.org/1937553002

Allow stencils to be attached to render targets created via SkSurface::MakeFromBackendTextureAsRenderTarget

This is a regression from "Refactor to separate backend object lifecycle
and GpuResource budget decision".

GrGLRenderer::canAttemptStencilAttachment was incorrectly returning false
for all wrapped render targets. This function should return false only if
the FBO is wrapped (unowned). If the FBO is owned by Skia, we can attach
stencils.

BUG=608238

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1924183003

Review-Url: https://codereview.chromium.org/1941353003

Modernize SkMutex and SkSemaphore.

   - use <atomic>
   - fuse SkMutex and SkBaseMutex
   - fuse SkSemaphore and SkBaseSemaphore

Still TODO:
   - replace SK_DECLARE_STATIC_MUTEX(name) with static SkMutex name

I just didn't want to bother fixing all that up until I know this CL sticks.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947153002

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

Review-Url: https://codereview.chromium.org/1947153002

Disable Ganesh layer hoister in a more revertable form

I have a worry that disabling the layer hoister might cause a perf regression on tiled architectures. This CL will be easier to revert than https://codereview.chromium.org/1950523002/ (Remove GrLayerHoister)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1945413003

Review-Url: https://codereview.chromium.org/1945413003

Prototype code that turns any/every flattenable into JSON

This makes inspecting things in SkDebugger far more useful - any filter
or other complex object on the paint is ultimately visible. You still
have to do some guess work to figure out what the fields actually mean,
but you can at least cross-reference with the code in flatten().

Screenshots:
Before: https://screenshot.googleplex.com/a6JM5HBBe6G.png
After : https://screenshot.googleplex.com/XQfr4YJ6mnH.png

Changes to public API are just removals and changes to make
some functions virtual.

TBR=reed@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1920423002

Review-Url: https://codereview.chromium.org/1920423002

Fix ASAN errors.

These errors occur when building for GPU but running dm with --nogpu. The current ASAN bots do not catch these because 1) the CPU bot compiles with skia_gpu=0 and 2) the GPU bot runs with --nocpu; whereas the new ASAN bots use the same compile for both the CPU and GPU bots.

BUG=skia:5157
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1953473002

Review-Url: https://codereview.chromium.org/1953473002

Revert of Allow stencils to be attached to render targets created via SkSurface::MakeFromBackendTextureAsRend… (patchset #1 id:60001 of https://codereview.chromium.org/1941353003/ )

Reason for revert:
Breaking the ANGLE bots. Message is:

Caught exception 3221225477 EXCEPTION_ACCESS_VIOLATION

8.59m elapsed, 1 active, 82 queued, 315MB RAM, 1442MB peak
unit test  SurfaceAttachStencil_Gpu
step returned non-zero exit code: -1073741819

Original issue's description:
> Allow stencils to be attached to render targets created via SkSurface::MakeFromBackendTextureAsRenderTarget
>
> This is a regression from "Refactor to separate backend object lifecycle
> and GpuResource budget decision".
>
> GrGLRenderer::canAttemptStencilAttachment was incorrectly returning false
> for all wrapped render targets. This function should return false only if
> the FBO is wrapped (unowned). If the FBO is owned by Skia, we can attach
> stencils.
>
> BUG=608238
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1924183003
>
> Committed: https://skia.googlesource.com/skia/+/0736f3386820f19c0fe90b5dda2094e253780071

TBR=bsalomon@google.com,kkinnunen@nvidia.com,ericrk@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=608238

Review-Url: https://codereview.chromium.org/1947143002

tools/using_skia_and_harfbuzz: minor cleanup

TBR=aam@

Review-Url: https://codereview.chromium.org/1947903002

Add some more testing around hairline GrShapes
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949903002

Review-Url: https://codereview.chromium.org/1949903002

experimental/tools/coreGraphicsPdf2png: fix

NOTRY=true
TBR=

Review-Url: https://codereview.chromium.org/1944403002

Expand GrStyle's interface.

Adds some basic getters on GrStyle as well as static const instances for fill and hairline.

Adds the ability to apply a GrStyle to a SkPath to produce an output SkPath.

Moves style key functions from GrShape to GrStyle.

Also fixes some issues with SkPath creation when applying style to GrShapes.

Adds tests that GrShape produces the correct SkPath when its GrStyle is applied.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1951613002

Review-Url: https://codereview.chromium.org/1951613002

Fix type conversion compile errors

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949553003

Review-Url: https://codereview.chromium.org/1949553003

Add XSAN suppression files to infrabots.isolate. Only compile 'dm' and 'nanobench' for XSAN flavors.

Needed for https://codereview.chromium.org/1920283002

BUG=skia:5157
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1859413003

Review-Url: https://codereview.chromium.org/1859413003

Allow stencils to be attached to render targets created via SkSurface::MakeFromBackendTextureAsRenderTarget

This is a regression from "Refactor to separate backend object lifecycle
and GpuResource budget decision".

GrGLRenderer::canAttemptStencilAttachment was incorrectly returning false
for all wrapped render targets. This function should return false only if
the FBO is wrapped (unowned). If the FBO is owned by Skia, we can attach
stencils.

BUG=608238

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1924183003

Review-Url: https://codereview.chromium.org/1941353003

Start building HarfBuzz+Skia example

(prevent bitrot in example code)

tools/using_skia_and_harfbuzz.cpp was written by aam@:
  https://github.com/aam/skiaex

update HarfBuzz to latest release: 1.2.7

BUG=skia:4742
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1940393002

Review-Url: https://codereview.chromium.org/1940393002

Create a single, unique pointer to sRGB color space

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1945693002

Review-Url: https://codereview.chromium.org/1945693002

Revert of Add Gr*Proxy classes (patchset #10 id:220001 of https://codereview.chromium.org/1937553002/ )

Reason for revert:
ASAN

Original issue's description:
> Add Gr*Proxy classes
>
> This isn't wired in anywhere yet.
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1937553002
>
> Committed: https://skia.googlesource.com/skia/+/de5bf0cfeca908b81a28cc50065f7bc2da3d2fd1
>
> Committed: https://skia.googlesource.com/skia/+/92605b35efa0155c44d24bd8415b4cc1db8831db

TBR=bsalomon@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/1944953002

return 4x4 matrix from SkColorSpace
move SkMatrix44 into core (with alias in utils to transition chrome)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1943833002

Review-Url: https://codereview.chromium.org/1943833002

documentation: SkDocument::MakePDF()

also, update some tools.

DOCS_PREVIEW= https://skia.org/user/sample/pdf?cl=1936283003
DOCS_PREVIEW= https://skia.org/user/api/canvas?cl=1936283003

Review-Url: https://codereview.chromium.org/1936283003

Add SK_API to SkFontLCDConfig.

This will allow transition from double deprecated SkFontHost.

TBR=reed
This is obviously needed to remove other code.

Review-Url: https://codereview.chromium.org/1951463002

Set barriers and image layout changes between mipmap blits.

Also fix some issues with mipLayer counts.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1945823002

Review-Url: https://codereview.chromium.org/1945823002

Enable two-sided stencil in Vulkan.

Two-sided stencil is always available in Vulkan.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1949573002

Review-Url: https://codereview.chromium.org/1949573002

Apply setImageLayout() to all of a VkImage's subresources.

This also adds fLevelCount to GrVkImage::Resource, which allows
clients to wrap mipmapped textures.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1943933002
TBR=bsalomon@google.com

Review-Url: https://codereview.chromium.org/1943933002

Documentation: Skia+Harfbuzz example

BUG=skia:4742
NOTRY=true
DOCS_PREVIEW= https://skia.org/user/tips?cl=1939123002#kerning

Review-Url: https://codereview.chromium.org/1939123002

Revert of Make GrStyle more sk_sp savy (patchset #1 id:1 of https://codereview.chromium.org/1947543002/ )

Reason for revert:
Crashing Ubuntu/GCC bot (but not clang)

Original issue's description:
> Make GrStyle more sk_sp savy
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947543002
>
> Committed: https://skia.googlesource.com/skia/+/46db22f1cfc65a7a4f1de13d774718347362f60f

TBR=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/1943973002

Minor refactor of GrLayerHoister

The main point of this CL is to make sure using the rect bounds (rather than the underlying texture's bounds) is acceptable in GrLayerHoister::FilterLayer.

This is split out of https://codereview.chromium.org/1943913002/ (Retract GrRenderTarget from GrLayerHoister)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947593002

Review-Url: https://codereview.chromium.org/1947593002

Make GrStyle more sk_sp savy
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947543002

Review-Url: https://codereview.chromium.org/1947543002

Add Gr*Proxy classes

This isn't wired in anywhere yet.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1937553002

Committed: https://skia.googlesource.com/skia/+/de5bf0cfeca908b81a28cc50065f7bc2da3d2fd1

Review-Url: https://codereview.chromium.org/1937553002

Add 8888 fast SrcOver mode.

The SkLinearBitmapPipeline_tile change is an edge case fix.

This will not function until gDefaultProfileSRGB in SkImageInfo is set to true.

The srcover implementation is not the fastest, and I'm working on a better version.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1886233006

Review-Url: https://codereview.chromium.org/1886233006

Add sync in vulkan when writing to linear texture

Since mapping memory of linear textures happens in real time and not via
a command buffer, we need to flush all work that access the texture
before uploading new data to it. Even if the texture is already in the
correct layout, General, we still need to make sure all work has been
finished.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1943333002

Review-Url: https://codereview.chromium.org/1943333002

Revert of Add Gr*Proxy classes (patchset #9 id:160001 of https://codereview.chromium.org/1937553002/ )

Reason for revert:
Seems to be breaking N7 & ANGLE bots

Original issue's description:
> Add Gr*Proxy classes
>
> This isn't wired in anywhere yet.
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1937553002
>
> Committed: https://skia.googlesource.com/skia/+/de5bf0cfeca908b81a28cc50065f7bc2da3d2fd1

TBR=bsalomon@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/1949463002

Remove SkBitmapHeap and SkBitmapHeapReader. They're unused.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1938033002

Review-Url: https://codereview.chromium.org/1938033002

Revert of Make stencils be attachable to render targets created via SkSurface::MakeFromBackendTextureAsRender… (patchset #9 id:150001 of https://codereview.chromium.org/1924183003/ )

Reason for revert:
Appears to be breaking the roll.

Original issue's description:
> Make stencils be attachable to render targets created via SkSurface::MakeFromBackendTextureAsRenderTarget
>
> This is a regression from "Refactor to separate backend object lifecycle
> and GpuResource budget decision".
>
> GrGLRenderTarget::CreateWrapped creates only render targets that wrap
> the FBO.
> GrGLRenderTargetTexture::CreateWrapped creates render targets that
> wrap the texture.
>
> Use the latter as the implementation for
> SkSurface::MakeFromBackendTextureAsRenderTarget.
>
>
> BUG=608238
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1924183003
>
> Committed: https://skia.googlesource.com/skia/+/fe635fd76bbc375d527d1884af23cc617ca364ee
>
> Committed: https://skia.googlesource.com/skia/+/1cb00e42bff455e23f6883bf09680438fa631224

TBR=bsalomon@google.com,kkinnunen@nvidia.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=608238

Review-Url: https://codereview.chromium.org/1943843002

Add Gr*Proxy classes

This isn't wired in anywhere yet.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1937553002

Review-Url: https://codereview.chromium.org/1937553002

Make stencils be attachable to render targets created via SkSurface::MakeFromBackendTextureAsRenderTarget

This is a regression from "Refactor to separate backend object lifecycle
and GpuResource budget decision".

GrGLRenderTarget::CreateWrapped creates only render targets that wrap
the FBO.
GrGLRenderTargetTexture::CreateWrapped creates render targets that
wrap the texture.

Use the latter as the implementation for
SkSurface::MakeFromBackendTextureAsRenderTarget.

BUG=608238

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1924183003

Committed: https://skia.googlesource.com/skia/+/fe635fd76bbc375d527d1884af23cc617ca364ee

Review-Url: https://codereview.chromium.org/1924183003

Make SkTime SK_API so that SkTime::GetDateTime() is visible to the clients.

# begin mtklein edits...
No public API changes.
TBR=reed@google.com
# end mtklein edits

BUG=skia:5230
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1930863003

Review-Url: https://codereview.chromium.org/1930863003

Make sure that Vulkan debug callback ptr is non-null.

BUG=skia:5244
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1935163002

Review-Url: https://codereview.chromium.org/1935163002

When filling nested rect path check for empty inner and empty outer rects

BUG=skia:5221
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1936073003

Review-Url: https://codereview.chromium.org/1936073003

added GrMSAAPathRenderer
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1834133003

Committed: https://skia.googlesource.com/skia/+/47a2dc8e229e93e1bcf7405747320920da1ab742

Committed: https://skia.googlesource.com/skia/+/b0fabd4ed55abcb6f5ed2f331c96c783f42bd37c

Review-Url: https://codereview.chromium.org/1834133003

Clean up SkFontConfigInterface implementation.

Renames some classes to avoid confusion with FontConfig.

Removed direct calls to FontConfig instead of calling FCI.

Moves the globals and factory to one (optional) file.

Moves font management code from typeface to font manager.

Adds index to fonts created from streams.

Associates FCI typefaces with the FCI instance which provides its identity.

Simplifies the singleton initialization.

Review-Url: https://codereview.chromium.org/1936213002

Add a hook for CPU-optimized sRGB-sRGB srcover.

Herb's really starting to get serious about tweaking this, which becomes
a lot easier when you've got SkOpts' runtime CPU detection. We should be
able to optimize this usefully for SSSE3, SSE4.1, AVX, AVX2, or NEON.
(We can of course implement a subset.)

This function takes two counts to give us flexibility to write src patterns:
   nsrc >= ndst -> the usual srcover function
   nsrc <  ndst -> repeat src until it fills dst
   nsrc << ndst -> possibly preprocess src into registers
   nsrc == 1    -> equivalent of blitrow_color32, srcover_1, etc.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1939783003

Review-Url: https://codereview.chromium.org/1939783003

Introduce SkGammas type to represent ICC gamma curves

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1928123002

Committed: https://skia.googlesource.com/skia/+/7b2c6dd8c918209cb92e1338905d511c68da3eb2

Review-Url: https://codereview.chromium.org/1928123002

Fix VK WritePixels with offset rect

VkUploadPixelsTest was failing because it was reallocating
the image when it didn't need to, because it was not
resetting the rowBytes when it did need to, and because the
origin was incorrect when using bottom-up images.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1940963002

Review-Url: https://codereview.chromium.org/1940963002

Update libjpeg-turbo to 1.4.90

Google3, Chrome, and Android are considering updates.  Let's start
testing Skia on the newest version.
https://critique.corp.google.com/#review/116667446
https://codereview.chromium.org/1934113002
https://buganizer.corp.google.com/u/0/issues/28268702

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1940903002

Committed: https://skia.googlesource.com/skia/+/3994c161ec5a2b83e8f6bab4d9c469eec2668d92

Review-Url: https://codereview.chromium.org/1940903002

Simplify ColorFilter in the add-free case.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1938003002

Review-Url: https://codereview.chromium.org/1938003002

Fix VulkanViewer package name
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1941813002

Review-Url: https://codereview.chromium.org/1941813002

Update gdb_app script to work with NDK r11c
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1936163002

Review-Url: https://codereview.chromium.org/1936163002