Remove check for cmake in android make script.
Update docs explaining how to setup Android SDK.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2084763003
DOCS_PREVIEW= https://skia.org/?cl=2084763003

Review-Url: https://codereview.chromium.org/2084763003

Never use CPU-built mips for Gray_8.

Fixes a bug where Gray_8 textures would still render as Alpha_8 (if the
first draw was detected as needing mips).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2088883002

Review-Url: https://codereview.chromium.org/2088883002

update callers to not use SkColorProfileType

Requires https://codereview.chromium.org/2087833002/ to land first.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2086583002

Review-Url: https://codereview.chromium.org/2086583002

Fix Mac clang linker

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2077353006

Review-Url: https://codereview.chromium.org/2077353006

Integers can now be passed as uniforms; needed for passing color count to fragment shader

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2076143003

Committed: https://skia.googlesource.com/skia/+/cc3a22b369e1a60fa2acf2987f2934baf7c4b198
Review-Url: https://codereview.chromium.org/2076143003

Add --simpleCodec to nanobench flags to run a smaller set of benchmarks

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2084683002

Review-Url: https://codereview.chromium.org/2084683002

Apply canonicalizations to path GrShapes.

The following canonicalizations of path-backed GrShapes are added:

*convex shapes are stored with even/odd (or inv even/odd) fill.

*filled paths are closed.

*dashed paths ignore inverseness of the fill

This will improve the results of queries about the geometry that will be added in a future change.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2067283003

Review-Url: https://codereview.chromium.org/2067283003

Support texel and draw-indirect buffers in null GL contexts

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2081183002

Review-Url: https://codereview.chromium.org/2081183002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2086843002

Review-Url: https://codereview.chromium.org/2086843002

Reorder colorspaces in GM to match previous behavior

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2088493002

Review-Url: https://codereview.chromium.org/2088493002

Benchmark rotated rect with AA/noAA

Using this benchmark, we verify that AA is about 4x slower than noAA in
path_fill_big_rotated_rect. This is what I aim to improve in the next
CL.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2087453003

Review-Url: https://codereview.chromium.org/2087453003

Buildbot spec: Add device_type for Android One

BUG=skia:5185
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2085713002

Review-Url: https://codereview.chromium.org/2085713002

Exempt GrShape test from running on Android

BUG=skia:5438
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2074403003

Review-Url: https://codereview.chromium.org/2074403003

Some simplifications of GrShape reductions/canonicalizations
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2064113004

Review-Url: https://codereview.chromium.org/2064113004

assert that non-empty clips have non-empty bounds

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2084603002

TBR=

Review-Url: https://codereview.chromium.org/2084603002

Respect color type and gamma settings in nanobench GPU configs

We were ignoring the requested properties in things like gpuf16 and
gpusrgb.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2075133003

Review-Url: https://codereview.chromium.org/2075133003

More removal of SkColorProfileType...

Scrub GMs. Remove the gDefaultProfile thing (it's unused now), along with
the command line flag that was setting it in DM and nanobench.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2071393002

Committed: https://skia.googlesource.com/skia/+/944876f2745a62a839e49275daf93a0329372e67
Review-Url: https://codereview.chromium.org/2071393002

Sanitize parsing of color look-up table

BUG=skia:5443
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2079243003

Review-Url: https://codereview.chromium.org/2079243003

Support sRGB dsts in opt code

201295.jpg on HP z620 (300x280)

QCMS Xform                    0.418 ms
Skia NEW Xform                0.378 ms

Vs QCMS                       1.11x

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2078623002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2078623002

Revert of Integers can now be passed as uniforms (patchset #1 id:1 of https://codereview.chromium.org/2076143003/ )

Reason for revert:
Compile failures on Vulkan platforms (eg https://build.chromium.org/p/client.skia.compile/builders/Build-Ubuntu-GCC-Arm7-Debug-Android_Vulkan)

Original issue's description:
> Integers can now be passed as uniforms; needed for passing color count to fragment shader
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2076143003
>
> Committed: https://skia.googlesource.com/skia/+/cc3a22b369e1a60fa2acf2987f2934baf7c4b198

TBR=tomhudson@google.com,bsalomon@google.com,fmenozzi@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:

Review-Url: https://codereview.chromium.org/2080133002

Simplify mask/clip intersection, making sure to explicitly check for an empty mask.

Previously we were only asserting the mask wasn't empty, which isn't necessarily true when we're given pathological float coordinates like +Inf or NaN.

A local run of nanobench --match text_ was not able to show this is faster or slower.

This patch fixed this first Chrome bug on my desktop, and the second is probably a dupe.
BUG=chromium:619378,chromium:613912

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2073873002

Review-Url: https://codereview.chromium.org/2073873002

Integers can now be passed as uniforms; needed for passing color count to fragment shader

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2076143003

Review-Url: https://codereview.chromium.org/2076143003

Revert of More removal of SkColorProfileType... (patchset #2 id:20001 of https://codereview.chromium.org/2071393002/ )

Reason for revert:
This was still used in chrome

Original issue's description:
> More removal of SkColorProfileType...
>
> Scrub GMs. Remove the gDefaultProfile thing (it's unused now), along with
> the command line flag that was setting it in DM and nanobench.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2071393002
>
> Committed: https://skia.googlesource.com/skia/+/944876f2745a62a839e49275daf93a0329372e67

TBR=msarett@google.com,reed@google.com,bsalomon@google.com,brianosman@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/2074103004

Fix mixup where SamplePathOverstroke wasn't added

- My sampleapp for the stroking bugs was inflight or something when
this change got made, so it didn't get added to this file.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2076283002

Review-Url: https://codereview.chromium.org/2076283002

Quick bandaid for chromium:611002.

We're somehow receiving non-premultiplied src inputs like 0x00ffffff to this
SrcOver blend.  That's a bug I intend to follow up on.  But for a quick
compatibility fix, go back to treating values like 0x00ffffff as transparent,
like we used to before crrev.com/1820313002.

This will not affect the correctness of code paths using properly premultiplied colors.

This should not change performance in any meaningful way.

The SIMD code paths (handling strides of 16 pixels at a time) happen to treat
invalid colors like 0x00fffff as transparent already.

BUG=chromium:611002
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2075173002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2075173002

More removal of SkColorProfileType...

Scrub GMs. Remove the gDefaultProfile thing (it's unused now), along with
the command line flag that was setting it in DM and nanobench.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2071393002

Review-Url: https://codereview.chromium.org/2071393002

SkPDF: Use type 2/3 shading for gradient shaders

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1925233003

Review-Url: https://codereview.chromium.org/1925233003

port SkColorXform_opts to Sk4f

I have tested that this compiles.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2078913003
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2078913003

Update GYP for linux Vulkan to support SDK 1.0.13

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2077043002

Review-Url: https://codereview.chromium.org/2077043002

Keep drawer open in xlarge-land and add float fps

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2071333002

Review-Url: https://codereview.chromium.org/2071333002

Enable viewer in non-Vulkan builds

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2067023002

Committed: https://skia.googlesource.com/skia/+/746b792a9aad1230ab69158a51c49b2e5437d95b
Review-Url: https://codereview.chromium.org/2067023002

remove colorprofiletype from imageinfo

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2075853002

Review-Url: https://codereview.chromium.org/2075853002

Don't try to set gamma treatment if texture creation fails.

Guessing this will fix the valgrind failure.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2075133002

Review-Url: https://codereview.chromium.org/2075133002

Quickly fix the Compile Problem of Unused kkImageColorXformMetaData
TBR=msarett@google.com

I don't know why the compiler is complaining that since ImageSlide.cpp is still
using it.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2075873003

Review-Url: https://codereview.chromium.org/2075873003

Move headers in include/utils/win to src/utils/win.

These headers are not intended for external use and are only used in src.

Review-Url: https://codereview.chromium.org/2070983002

Revert of Support pixel antialising in DirectWrite. (patchset #4 id:60001 of https://codereview.chromium.org/2065833002/ )

Reason for revert:
Moving to the new glyph run analysis changes things anyway, and the Chromium Win 10 bots are having issues with this. Revert until I have time to make the suppression wider.

Original issue's description:
> Support pixel antialising in DirectWrite.
>
> DirectWrite2 supports pixel antialiasing and rendering without hinting.
>
> BUG=skia:5416
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2065833002
>
> TBR=reed
> Will move SkTScopedComPtr into src.
>
> Committed: https://skia.googlesource.com/skia/+/bd770d619553a88eeaa64ff29082f62db5c9b4d2

TBR=reed@google.com,mtklein@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:5416

Review-Url: https://codereview.chromium.org/2075913002

Use Offscreen Surface for Split Screen

A dashed line is also added to split the screen. Currently, drawing
images with sRGB seems to still have bugs.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069653002

Review-Url: https://codereview.chromium.org/2069653002

Check for some potential subheap allocation failures.

Need to make sure we can allocate areas larger than our specified
subheap sizes, and deal with potential fragmentation in the main
system heap.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2072763002

Review-Url: https://codereview.chromium.org/2072763002

Add missing include

TBR=egdaniel@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2070953002

Review-Url: https://codereview.chromium.org/2070953002

Lots of progress on switching to SkColorSpace rather than SkColorProfileType

Fixed a bunch of code in Ganesh, as well as usage of SkColorProfileType in most of our tools (DM, SampleApp, Viewer, nanobench, skiaserve, HelloWorld).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069173002

Committed: https://skia.googlesource.com/skia/+/6a61a875467646f8dbc37cfecf49e12d1f475170
Review-Url: https://codereview.chromium.org/2069173002

Refactoring of GPU NormalMap handling out into its own class.

The purpose of this change is to refactor the handling of normal maps out of SkLightingShader, laying the groundwork to eventually allow for multiple normal sources.

What this CL includes:

- Created a new 'NormalMapFP', out of the existing normal map reading behavior in LightingFP.

- Encapsulates this new fragment processor on a new class NormalMapSource.

- Created a NormalSource abstraction that will interface with SkLightingShader.

- Adapted SkLightingShader to use the normals from its NormalSource field ON THE GPU SIDE. No changes done to the CPU side yet.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2043393002

Committed: https://skia.googlesource.com/skia/+/87b0dd00cf9409c5fc990f5d0bb7c0df837f08da
Committed: https://skia.googlesource.com/skia/+/a7d1e2a57aef2aa4913d4380646d60bbab761318
Review-Url: https://codereview.chromium.org/2043393002

Clean up two unlaunched SSE 4.1 8888 blits.

This code was running on our bots but never in Chrome.
That's a bad state to be in.

My plan here use to be to redesign how our 8888 blits worked in SSE 4.1, mainly
for perfect correctness but also for speed, then to spread what I learned there
to SSE2, AVX+, and NEON.

I have since lost interest in changing any aspect of how our legacy 8888 blits
work.  There's not much point in making them a bit or two more correct when the
math is fundamentally wrong.

This will cause many diffs in Gold, none perceptible.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062853002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Committed: https://skia.googlesource.com/skia/+/6e472093009bf2fc4a8e53010b51040efcb71213
Review-Url: https://codereview.chromium.org/2062853002

Fix Vulkan build after API change

TBR=brianosman@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2075713002

Review-Url: https://codereview.chromium.org/2075713002

Revert of Lots of progress switching to SkColorSpace rather than SkColorProfileType (patchset #10 id:180001 of https://codereview.chromium.org/2069173002/ )

Reason for revert:
Mac crashes in GrUploadPixmapToTexture

Original issue's description:
> Lots of progress on switching to SkColorSpace rather than SkColorProfileType
>
> Fixed a bunch of code in Ganesh, as well as usage of SkColorProfileType in most of our tools (DM, SampleApp, Viewer, nanobench, skiaserve, HelloWorld).
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069173002
>
> Committed: https://skia.googlesource.com/skia/+/6a61a875467646f8dbc37cfecf49e12d1f475170

TBR=reed@google.com,herb@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:

Review-Url: https://codereview.chromium.org/2072813002

Lots of progress on switching to SkColorSpace rather than SkColorProfileType

Fixed a bunch of code in Ganesh, as well as usage of SkColorProfileType in most of our tools (DM, SampleApp, Viewer, nanobench, skiaserve, HelloWorld).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069173002

Review-Url: https://codereview.chromium.org/2069173002

Implement fast, correct gamma conversion for color xforms

201295.jpg on HP z620
(300x280, most common form of sRGB profile)

QCMS Xform                    0.495 ms
Skia Old Xform                0.235 ms
Skia NEW Xform                0.423 ms

Vs Old Code                   0.56x
Vs QCMS                       1.17x

So to summarize, we are now much slower than before,
but still a bit faster than QCMS.  And now we are also
far more accurate than QCMS :).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2060823003
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2060823003

Support pixel antialising in DirectWrite.

DirectWrite2 supports pixel antialiasing and rendering without hinting.

BUG=skia:5416
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2065833002

TBR=reed
Will move SkTScopedComPtr into src.

Review-Url: https://codereview.chromium.org/2065833002

Add samples to Viewer.

This adds support with animation, assuming the sample has
implemented onAnimate. Event handling has not been
implemented.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2056343004

Committed: https://skia.googlesource.com/skia/+/76963e73704a42a18c29d6fbdcccb566e5c67658
Review-Url: https://codereview.chromium.org/2056343004

Differentiate between sRGBGamma and 2Dot2Gamma

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2067833003

Review-Url: https://codereview.chromium.org/2067833003

Keep SkColorSpace and SkColorProfileType in sync

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2068743003

Committed: https://skia.googlesource.com/skia/+/04d35bd80dde7cace866781037cbbdcab14ab683
Review-Url: https://codereview.chromium.org/2068743003

Add heap tests for Vulkan.

Also fixes some bugs that were found.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2072453002

Review-Url: https://codereview.chromium.org/2072453002

Revert of Keep SkColorSpace and SkColorProfileType in sync (patchset #2 id:20001 of https://codereview.chromium.org/2068743003/ )

Reason for revert:
breaking roll.

../../third_party/skia/include/core/SkImageInfo.h:202: error: undefined reference to 'SkColorSpace::NewNamed(SkColorSpace::Named)'

Original issue's description:
> Keep SkColorSpace and SkColorProfileType in sync
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2068743003
>
> Committed: https://skia.googlesource.com/skia/+/04d35bd80dde7cace866781037cbbdcab14ab683

TBR=reed@google.com,brianosman@google.com,herb@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:

Review-Url: https://codereview.chromium.org/2073593002

Revert of Refactoring of GPU NormalMap handling out into its own class (patchset #9 id:160001 of https://codereview.chromium.org/2043393002/ )

Reason for revert:
break deps roll

Original issue's description:
> Refactoring of GPU NormalMap handling out into its own class.
>
> The purpose of this change is to refactor the handling of normal maps out of SkLightingShader, laying the groundwork to eventually allow for multiple normal sources.
>
> What this CL includes:
>
> - Created a new 'NormalMapFP', out of the existing normal map reading behavior in LightingFP.
>
> - Encapsulates this new fragment processor on a new class NormalMapSource.
>
> - Created a NormalSource abstraction that will interface with SkLightingShader.
>
> - Adapted SkLightingShader to use the normals from its NormalSource field ON THE GPU SIDE. No changes done to the CPU side yet.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2043393002
>
> Committed: https://skia.googlesource.com/skia/+/87b0dd00cf9409c5fc990f5d0bb7c0df837f08da
> Committed: https://skia.googlesource.com/skia/+/a7d1e2a57aef2aa4913d4380646d60bbab761318

TBR=reed@google.com,dvonbeck@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/2068983005

Revert of Enable viewer in non-Vulkan builds (patchset #2 id:20001 of https://codereview.chromium.org/2067023002/ )

Reason for revert:
breaking compile bots

Original issue's description:
> Enable viewer in non-Vulkan builds
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2067023002
>
> Committed: https://skia.googlesource.com/skia/+/746b792a9aad1230ab69158a51c49b2e5437d95b

TBR=djsollen@google.com,jvanverth@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/2072563002

Disable SkImage_drawAbandonedGpuImage on Vulkan.

It crashes issuing a clear batch.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2060263005

TBR=brianosman@google.com

Review-Url: https://codereview.chromium.org/2060263005

Enable viewer in non-Vulkan builds

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2067023002

Review-Url: https://codereview.chromium.org/2067023002

Workaround NVIDIA driver bug with NV_sample_mask_override_coverage

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2070133002

Review-Url: https://codereview.chromium.org/2070133002

Keep SkColorSpace and SkColorProfileType in sync

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2068743003

Review-Url: https://codereview.chromium.org/2068743003

Refactoring of GPU NormalMap handling out into its own class.

The purpose of this change is to refactor the handling of normal maps out of SkLightingShader, laying the groundwork to eventually allow for multiple normal sources.

What this CL includes:

- Created a new 'NormalMapFP', out of the existing normal map reading behavior in LightingFP.

- Encapsulates this new fragment processor on a new class NormalMapSource.

- Created a NormalSource abstraction that will interface with SkLightingShader.

- Adapted SkLightingShader to use the normals from its NormalSource field ON THE GPU SIDE. No changes done to the CPU side yet.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2043393002

Committed: https://skia.googlesource.com/skia/+/87b0dd00cf9409c5fc990f5d0bb7c0df837f08da
Review-Url: https://codereview.chromium.org/2043393002

Add asset management scripts

These provide an easy way to create assets to be used by bots,
eg. Android SDK.

To create an asset:
$ infra/bots/assets/assets.py add android_sdk
(adds scripts in infra/bots/assets/android_sdk)

To upload a new version of an asset:
$ infra/bots/assets/android_sdk/upload.py -t $ANDROID_SDK_ROOT
(uploads Android SDK to GS, writes a version file)
$ git commit
$ git cl upload

To download the current version of the asset:
$ infra/bots/assets/android_sdk/download.py -t ../tmp

BUG=skia:5427
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069543002

Review-Url: https://codereview.chromium.org/2069543002

Added warning for variable-length arrays to GYP, since they are unsupprted under Win targets

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2068863002

Review-Url: https://codereview.chromium.org/2068863002

Add SampleApp that shows problems in stroking

- When the stroke width gets too big the raster algorithm produces
  incorrect output
- This also causes the GPU path to produce incorrect output if the width
  is larger than an arbitrary limit in GrAALinearizingConvexPathRenderer
  (which is the renderer used for this specific geometry)

- Other non-convex geometries probably also produce strange results

BUG=skia:5405,skia:5406
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2060903002

Review-Url: https://codereview.chromium.org/2060903002

Actually fix leak in deferred_texture_image_data

TBR=egdaniel@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2068863003

Review-Url: https://codereview.chromium.org/2068863003

Add test for rendering an already abandoned texture-backed SkImage.

BUG=chromium:618858
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2067133002

Review-Url: https://codereview.chromium.org/2067133002

Check for abandoned texture in GrTextureAdjuster::refTextureSafeForParams

BUG=chromium:618858
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2063313002

TBR=egdaniel@google.com

Review-Url: https://codereview.chromium.org/2063313002

Revert of Refactoring of GPU NormalMap handling out into its own class (patchset #7 id:120001 of https://codereview.chromium.org/2043393002/ )

Reason for revert:
Breaking build and deps roll. Need to move include of SkBitmapProcShader in SkLightingShader.cpp from gpu include list to general list.

Original issue's description:
> Refactoring of GPU NormalMap handling out into its own class.
>
> The purpose of this change is to refactor the handling of normal maps out of SkLightingShader, laying the groundwork to eventually allow for multiple normal sources.
>
> What this CL includes:
>
> - Created a new 'NormalMapFP', out of the existing normal map reading behavior in LightingFP.
>
> - Encapsulates this new fragment processor on a new class NormalMapSource.
>
> - Created a NormalSource abstraction that will interface with SkLightingShader.
>
> - Adapted SkLightingShader to use the normals from its NormalSource field ON THE GPU SIDE. No changes done to the CPU side yet.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2043393002
>
> Committed: https://skia.googlesource.com/skia/+/87b0dd00cf9409c5fc990f5d0bb7c0df837f08da

TBR=reed@google.com,dvonbeck@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/2062133004

Fix leak of GrContextThreadSafeProxy in deferred_texture_image_data
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069443004

Review-Url: https://codereview.chromium.org/2069443004

Refactoring of GPU NormalMap handling out into its own class.

The purpose of this change is to refactor the handling of normal maps out of SkLightingShader, laying the groundwork to eventually allow for multiple normal sources.

What this CL includes:

- Created a new 'NormalMapFP', out of the existing normal map reading behavior in LightingFP.

- Encapsulates this new fragment processor on a new class NormalMapSource.

- Created a NormalSource abstraction that will interface with SkLightingShader.

- Adapted SkLightingShader to use the normals from its NormalSource field ON THE GPU SIDE. No changes done to the CPU side yet.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2043393002

Review-Url: https://codereview.chromium.org/2043393002

Fix GrShape to preserve inverseness of rrects for strokes but not dashes.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2051263003

Review-Url: https://codereview.chromium.org/2051263003

Bin circular blur profile textures by scale and blur to radius ratio.

This reduces the number of profile textures that will be generated throughout an animation and also caps the texture size. It could probably be better tuned.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062743003

Review-Url: https://codereview.chromium.org/2062743003

SkMultiPictureDocument: don't rely on SkPicture::cullRect

SkPicture::cullRect() is not guaranteed to be the same as the
bounds passed to SkPictureRecorder::beginRecording().

Review-Url: https://codereview.chromium.org/2067473003

Remove command buffer sample app support.

This hasn't compiled for months. We are deemphasizing SampleApp in favor of viewer. Windowed command buffer is not currently something we currently plan to maintain.

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2061233002

Review-Url: https://codereview.chromium.org/2061233002

Store GLSL max samplers using uint32_t rather than uint8_t

TBR=egdaniel@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2061123003

Review-Url: https://codereview.chromium.org/2061123003

Another linux Vulkan fix

BUG=skia:
TBR=egdaniel@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069443003

Review-Url: https://codereview.chromium.org/2069443003

Fix linux compile error

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2069633002
TBR=egdaniel@google.com

Review-Url: https://codereview.chromium.org/2069633002

If we fail to create a window surface with sRGB, try again without.

Fixes crash when switching to sRGB mode on Nexus 9, among others.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2066743003

Review-Url: https://codereview.chromium.org/2066743003

Fix compile error (uninitialized variable use)

switch is already handling all cases, but some compilers don't care.
This suppresses errors about potentially using vkLoadOp and vkStoreOp
uninitialized.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2068523002

Review-Url: https://codereview.chromium.org/2068523002

Quickly fix 0f to 0.0f
TBR=egdaniel@google.com

So Visual Studio won't complain.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2064863002

Review-Url: https://codereview.chromium.org/2064863002

Create free list heap for suballocation

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2029763002

Review-Url: https://codereview.chromium.org/2029763002

Quickly fix the double to SkScalar issue

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2060823004

Review-Url: https://codereview.chromium.org/2060823004

Display JPGs in Viewer with Split Screen and ColorSpaceXform

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2058753002

Review-Url: https://codereview.chromium.org/2058753002

Move create_image_large in ImageTest.cpp inside #if SK_SUPPORT_GPU

TBR=egdaniel@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062943002

Review-Url: https://codereview.chromium.org/2062943002

Revert of Clean up two unlaunched SSE 4.1 8888 blits. (patchset #1 id:1 of https://codereview.chromium.org/2062853002/ )

Reason for revert:
Breaks a couple Google3 goldens.  I need to rebaseline google3 with -DSK_SUPPORT_LEGACY_X86_BLITS first, then reland this.

Original issue's description:
> Clean up two unlaunched SSE 4.1 8888 blits.
>
> This code was running on our bots but never in Chrome.
> That's a bad state to be in.
>
> My plan here use to be to redesign how our 8888 blits worked in SSE 4.1, mainly
> for perfect correctness but also for speed, then to spread what I learned there
> to SSE2, AVX+, and NEON.
>
> I have since lost interest in changing any aspect of how our legacy 8888 blits
> work.  There's not much point in making them a bit or two more correct when the
> math is fundamentally wrong.
>
> This will cause many diffs in Gold, none perceptible.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062853002
> CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/6e472093009bf2fc4a8e53010b51040efcb71213

TBR=reed@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/2066453003

Revert of Add samples to Viewer. (patchset #4 id:60001 of https://codereview.chromium.org/2056343004/ )

Reason for revert:
breaking on android compile

Original issue's description:
> Add samples to Viewer.
>
> This adds support with animation, assuming the sample has
> implemented onAnimate. Event handling has not been
> implemented.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2056343004
>
> Committed: https://skia.googlesource.com/skia/+/76963e73704a42a18c29d6fbdcccb566e5c67658

TBR=bsalomon@google.com,jvanverth@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/2063883002

Revert of Update typeface serialization style. (patchset #3 id:40001 of https://codereview.chromium.org/2030683002/ )

Reason for revert:
serialize-8888 broken for typefacestyles on Mac.  (Bold weighting goes missing.)

Original issue's description:
> Update typeface serialization style.
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2030683002
>
> Committed: https://skia.googlesource.com/skia/+/7ae1c72e593823e6e300e4199558555765bcec17

TBR=reed@google.com,bungeman@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/2067583002

Fix SkImage::asLegacyBitmap() to respect color space info

BUG=skia:5428
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2061783002

Review-Url: https://codereview.chromium.org/2061783002

Add prescale option to deferred params

Currently, Skia always uploads GPU textures at full resolution. This
change allows us to pass a pre-scale mip level to the
deferred texture image logic, which causes us to pre-scale the image
to the given mip level, and upload that mip level instead of the full
image.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2007113008

Review-Url: https://codereview.chromium.org/2007113008

Clean up two unlaunched SSE 4.1 8888 blits.

This code was running on our bots but never in Chrome.
That's a bad state to be in.

My plan here use to be to redesign how our 8888 blits worked in SSE 4.1, mainly
for perfect correctness but also for speed, then to spread what I learned there
to SSE2, AVX+, and NEON.

I have since lost interest in changing any aspect of how our legacy 8888 blits
work.  There's not much point in making them a bit or two more correct when the
math is fundamentally wrong.

This will cause many diffs in Gold, none perceptible.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062853002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2062853002

Update typeface serialization style.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2030683002

Review-Url: https://codereview.chromium.org/2030683002

Add control over whether lines are special cased in SkDashPath. Disable when called from GrShape.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2055253002

Review-Url: https://codereview.chromium.org/2055253002

skiaserve no longer crashes when no X server is present
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2059263002

Review-Url: https://codereview.chromium.org/2059263002

Add samples to Viewer.

This adds support with animation, assuming the sample has
implemented onAnimate. Event handling has not been
implemented.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2056343004

Review-Url: https://codereview.chromium.org/2056343004

Better (?) interface for controlling sRGB-ness of mipmaps on GrTexture

Switch to new SkSourceGammaTreatment type. Move to a separate setter,
rather than an optional argument to dirty. Set the flag correctly after
uploading CPU-built mips.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2058143002

Review-Url: https://codereview.chromium.org/2058143002

Request an sRGB FBO0 on Android. Fixes sRGB GL mode.

Tried to comment my logic, but it's totally possible that we run on some
device where this doesn't work. From a strict reading of the relevant
specs, what I'm doing shouldn't work on the Nexus 5X, but it's necessary
to get correct results.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2054223002

Review-Url: https://codereview.chromium.org/2054223002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2065433003

Review-Url: https://codereview.chromium.org/2065433003

tweak mipmap_srgb gm to use integer coordinates

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2061553002

TBR=

Review-Url: https://codereview.chromium.org/2061553002

Make large array of GrShapes in test code heap allocated

TBR=mtklein@chormium.org
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2058343002

Review-Url: https://codereview.chromium.org/2058343002

Move multisampleDisableSupport into GrCaps

Non-platform code needs this information in order to make rendering
decisions.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062443002

Review-Url: https://codereview.chromium.org/2062443002

Type-erase SkAutoMutexAcquire and SkAutoExclusive.

This is purely for convenience, to not need to write the lock type
in the guard anymore.  This should all inline away.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2055023003

Review-Url: https://codereview.chromium.org/2055023003

fix non-square stroke rects (non-aa)

fiddle showing bug: https://fiddle.skia.org/c/a21ecf5e33da84f52eba26e8c2f3d280

BUG=skia:5408
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2060543002

Review-Url: https://codereview.chromium.org/2060543002