Update SK_IMAGE_VERSION to 3

This includes two new test wbmps (that came from clients of BitmapFactory).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1722263002

Review URL: https://codereview.chromium.org/1722263002

Move SkUtils.h to src/core.

Review URL: https://codereview.chromium.org/1722703003

added 'visible' attribute to skiaserve JSON
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1726663002

Review URL: https://codereview.chromium.org/1726663002

Use new API instead of the deprecated ones.

This change will allow us to remove the old API in follow up clean up.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1728553002

Review URL: https://codereview.chromium.org/1728553002

Update PIEX: Prepares support for uncompressed thumbnails

The new version of PIEX added new API for adding support to thumbnails. It also marks some old API as deprecate.

BUG=b/26894354
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1723123002

Review URL: https://codereview.chromium.org/1723123002

Turn on TURBO_HAS_CROP for Android framework

https://buganizer.corp.google.com/u/0/issues/27290496
BUG=skia:4256
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1721953002

Review URL: https://codereview.chromium.org/1721953002

Make DM failures fatal

We should now only be pushing Srcs for tests that we expect
to succeed.  So any failures should be treated as errors.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719293002

Review URL: https://codereview.chromium.org/1719293002

Add dest type hint to SkShader::ContextRec

Let SkBlitter decide which dst type is optimal (PMColor vs PM4f), and
pass that info to shaders.

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1724503002

Review URL: https://codereview.chromium.org/1724503002

remove unnecessary copies

these used to matter with pass-by-ref

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1721153002

Review URL: https://codereview.chromium.org/1721153002

Run no CPU benches in Debug trybots.

They seem to just be wasting time on the CQ.

Hacking this into nanobench_flags.py makes it easy to back out.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720223002

Review URL: https://codereview.chromium.org/1720223002

Better encapsulation and vector calling convention.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1723683002

Review URL: https://codereview.chromium.org/1723683002

Fix DM_INCLUDES after https://codereview.chromium.org/1688543002

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719253002

Review URL: https://codereview.chromium.org/1719253002

Only use fake gamma with linear devices.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711223003

Review URL: https://codereview.chromium.org/1711223003

Add spans for matrix ops.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1722703002

Review URL: https://codereview.chromium.org/1722703002

fix unsigned/signed warning on linux for GrVkDescriptorPool::numPoolSizes()

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1724493002

Review URL: https://codereview.chromium.org/1724493002

Fix unsigned short UVs for Vulkan.

Modifies the attribute type to match GL.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1723633003

Review URL: https://codereview.chromium.org/1723633003

Fix gyp-and-sync for windows
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715423003

Review URL: https://codereview.chromium.org/1715423003

fix misc asserts and checks found by fuzzer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719993002

Review URL: https://codereview.chromium.org/1719993002

allow move/zero-line/close to draw caps

R=fmalita@chromium.org
BUG=skia:4784
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720953002

Review URL: https://codereview.chromium.org/1720953002

kill gm now that we can test srgb on all gms

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719963004

TBR=

Review URL: https://codereview.chromium.org/1719963004

Use new jpeg_crop_scanlines() API to optimize jpeg subset decodes

This was adapted from:
https://codereview.chromium.org/1530933003

Subset Decode Runtime (Original / Optimized) on Nexus 6P
TopLeft     0.51x
TopRight    0.56x
Middle      0.71x
BottomLeft  0.79x
BottomRight 0.79x

BUG=skia:4256
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719073002

Review URL: https://codereview.chromium.org/1719073002

Incorporate Vulkan files into main gpu.gyp/gpu.gypi

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718273002

Review URL: https://codereview.chromium.org/1718273002

Use unorm shorts for texture coordinates when rendering text.

There are a couple of reasons for this:
- Vulkan does not guarantee conversions from integral vertex attributes
  to floating point shader variables
- This may be faster and more precise on some platforms, as it avoids
  the aforementioned conversion and changes a multiply by a very small
  value to a multiply by a medium-sized value.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713693002
TBR=bsalomon@google.com

Committed: https://skia.googlesource.com/skia/+/e507ff0460f4f878214b9454fb5b9ab8d64d8063

Review URL: https://codereview.chromium.org/1713693002

Simplify calculations in AADistanceFieldPathBatch::writePathVertices.

No API changes.

BUG=skia:4632
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715063002
TBR=reed

Review URL: https://codereview.chromium.org/1715063002

Allow client to force an SkImage snapshot to be unique (and uniquely own its backing store).
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1686163002

Review URL: https://codereview.chromium.org/1686163002

Add ContextRec param to SkShader::contextSize()

To facilitate upcoming context selection changes.

R=reed@google.com,mtklein@google.com,herb@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720933002

Review URL: https://codereview.chromium.org/1720933002

remove dead unpremul checks

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715333004

Review URL: https://codereview.chromium.org/1715333004

Fix to bazel build file to exclude vulkan files

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1723593002

Review URL: https://codereview.chromium.org/1723593002

skia.org internal links: Add link to Google3 Autoroller doc.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720943003
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1720943003

Review URL: https://codereview.chromium.org/1720943003

Initial linear gradient 4f impl

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1688543002

Review URL: https://codereview.chromium.org/1688543002

added mode=cpu|gpu to skiaserve JSON
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719973002

Review URL: https://codereview.chromium.org/1719973002

Fix nanobench SKP flushes between loops

Updates SKPBench to actually flush between draw loops, as this was the
original intent of the flush logic.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715873002

Review URL: https://codereview.chromium.org/1715873002

Add gl_SampleMask functionality to fragment builders

Adds methods for overriding and masking a fragment's sample mask.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1690963003

Committed: https://skia.googlesource.com/skia/+/533cefe5b9c7cec2592fc7ca00ee4cf69a26c094

Review URL: https://codereview.chromium.org/1690963003

Revert of fix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1719913002/ )

Reason for revert:
Looks to be causing failures in LayerTreeHostFilters* tests (https://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_ng/builds/177297/steps/cc_unittests%20%28with%20patch%29/logs/stdio).

Original issue's description:
> fix misc asserts and checks found by fuzzer
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713383002
>
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/653db51b440491b0fb1908bf5a43dcc89c90044d

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/1721743002

Remove vulkan files from cmake build

TBR=bsalomon@google.com
NOTRY=True

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1717313002

Review URL: https://codereview.chromium.org/1717313002

clean-up dead code

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1719773002

Review URL: https://codereview.chromium.org/1719773002

Add vulkan files into skia repo. This is an incomplete backend with only partial functionality at this time.

R=robertphillips@google.com
TBR=bsalomon@google.com

BUG=skia:4955
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718693002

Committed: https://skia.googlesource.com/skia/+/48cf268defad66f58f1aa03b4835e5583be96b2f

Review URL: https://codereview.chromium.org/1718693002

lots of sRGB and F16 blits

- generalize F16 xfermode procs
- spriteblits for F16 and sRGB
- saveLayer now respects colortype and profiletype

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1686013002

Review URL: https://codereview.chromium.org/1685203002

fix misc asserts and checks found by fuzzer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713383002

TBR=

Review URL: https://codereview.chromium.org/1719913002

add SkGraphics::PurageAllCaches

BUG=skia:4755
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1722603002

TBR=

Review URL: https://codereview.chromium.org/1722603002

Revert of Add vulkan files into skia repo. (patchset #2 id:20001 of https://codereview.chromium.org/1718693002/ )

Reason for revert:
breaking builds

Original issue's description:
> Add vulkan files into skia repo. This is an incomplete backend with only partial functionality at this time.
>
> R=robertphillips@google.com
> TBR=bsalomon@google.com
>
> BUG=skia:4955
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718693002
>
> Committed: https://skia.googlesource.com/skia/+/48cf268defad66f58f1aa03b4835e5583be96b2f

TBR=robertphillips@google.com,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:4955

Review URL: https://codereview.chromium.org/1723503002

Add vulkan files into skia repo. This is an incomplete backend with only partial functionality at this time.

R=robertphillips@google.com
TBR=bsalomon@google.com

BUG=skia:4955
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718693002

Review URL: https://codereview.chromium.org/1718693002

Revert of ix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1723473002/ )

Reason for revert:
need to not land the canvas change yet

Original issue's description:
> Reland of ix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1713413002/ )
>
> Reason for revert:
> deps instability seemingly fixed. trying again.
>
> Original issue's description:
> > Revert of fix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1713383002/ )
> >
> > Reason for revert:
> > speculative deps-roll fix
> >
> > Original issue's description:
> > > fix misc asserts and checks found by fuzzer
> > >
> > > BUG=skia:
> > > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713383002
> > >
> > > TBR=robertphilips
> > >
> > > Committed: https://skia.googlesource.com/skia/+/00bea4ad310c4ec4dd95809b47ce3fbfa8fd0e1e
> >
> > TBR=robertphillips@google.com
> > # 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/+/d98ef6377269e3596423225ab922301ed40529a0
>
> TBR=robertphillips@google.com
> # 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/+/75a73297cb44c65a9efcf5f212854cb457140e34

TBR=robertphillips@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/1723483002

Reland of ix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1713413002/ )

Reason for revert:
deps instability seemingly fixed. trying again.

Original issue's description:
> Revert of fix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1713383002/ )
>
> Reason for revert:
> speculative deps-roll fix
>
> Original issue's description:
> > fix misc asserts and checks found by fuzzer
> >
> > BUG=skia:
> > GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713383002
> >
> > TBR=robertphilips
> >
> > Committed: https://skia.googlesource.com/skia/+/00bea4ad310c4ec4dd95809b47ce3fbfa8fd0e1e
>
> TBR=robertphillips@google.com
> # 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/+/d98ef6377269e3596423225ab922301ed40529a0

TBR=robertphillips@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/1723473002

fix race: this kCacheSize is probably not meant to be static

TSAN sees us racing to intialize the static const kCacheSize.

This static const value depends on the runtime value image->getSize(), so there
is a race to set it.  I think this is unintentionally copy-paste from the other
tests that use a constant kCacheSize.  It's weird to intialize a constant based
on the first call to test_internal_purge(), though luckily we're always calling
it with same sized images today.

See these TSAN failures:
    https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-TSAN/builds/4937
    https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-TSAN/builds/4940/steps/dm/logs/stdio

TBR=robertphillips@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720743002

Review URL: https://codereview.chromium.org/1720743002

SkNx: kth<...>() -> [...]

Just some syntax cleanup.  No real change: kth<...>() was calling [...] already.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1714363002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1714363002

Revert of fix misc asserts and checks found by fuzzer (patchset #1 id:1 of https://codereview.chromium.org/1713383002/ )

Reason for revert:
speculative deps-roll fix

Original issue's description:
> fix misc asserts and checks found by fuzzer
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713383002
>
> TBR=robertphilips
>
> Committed: https://skia.googlesource.com/skia/+/00bea4ad310c4ec4dd95809b47ce3fbfa8fd0e1e

TBR=robertphillips@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/1713413002

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=1715213002

Review URL: https://codereview.chromium.org/1715213002

revert fuzzer fixes for canvas, need to investigate more

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1717943002

TBR=

Review URL: https://codereview.chromium.org/1717943002

fix misc asserts and checks found by fuzzer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713383002

TBR=robertphilips

Review URL: https://codereview.chromium.org/1713383002

Fix asserts found in SampleApp and update the tests with additional interesting numbers.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1717743004

Review URL: https://codereview.chromium.org/1717743004

Change type of SkScalerContext_FreeType::fScaleX/Y. All uses want FT_F26Dot6.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711053004

Review URL: https://codereview.chromium.org/1711053004

Add point spans, but fall back for all cases.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711963003

curr/maxrss     loops   min     median  mean    max     stddev  samples                 config  bench
  10/10  MB     1       481µs   504µs   512µs   580µs   6%      ▁▁▄▆▂▃▁▁▁▁▁▁▃▇█▅▃█▃▅    nonrendering    SkBitmapFPAffineXClampYClampFilterOrig
  10/10  MB     1       2.21ms  2.31ms  2.31ms  2.53ms  4%      ▁▁▆▃▃▃▄▅▂▁▄▁█▅▁▁▄▂▆▁    nonrendering    SkBitmapFPAffineXClampYClampFilterLinr
  10/10  MB     1       2.49ms  2.63ms  2.61ms  2.78ms  4%      ▄▃▇▄▅▇▆▁▆▁▁▁▅▂▁▄▃█▆▃    nonrendering    SkBitmapFPAffineXClampYClampFiltersRGB
  10/10  MB     2       150µs   151µs   153µs   162µs   2%      ▅▇▂▂▁▁▂▂▁▁▂▂▁▁▁▁▇█▂▁    nonrendering    SkBitmapFPAffineXClampYClampNearestOrig
  10/10  MB     1       552µs   553µs   562µs   658µs   4%      ▁▁▁▁▁▁▁▁▁▄▃▁▁▁▁▁▁▁▁█    nonrendering    SkBitmapFPAffineXClampYClampNearestLinr
  10/10  MB     1       718µs   742µs   748µs   808µs   4%      ▃▃▁▁▅▁▁▁▁▃▃▃▃▇▇▇▁▁▇█    nonrendering    SkBitmapFPAffineXClampYClampNearestsRGB
  10/10  MB     1       277µs   286µs   292µs   341µs   7%      █▁▁▁▁▁▁▁▁▁▃▂▃▃▆█▄▂▂▂    nonrendering    SkBitmapFPScaleXClampYClampFilterOrig
  10/10  MB     1       2.21ms  2.41ms  2.39ms  2.55ms  4%      ▄▆▃▅▇▂▁▁▇▅▆▅▅▆▄▅█▅▆▅    nonrendering    SkBitmapFPScaleXClampYClampFilterLinr
  10/10  MB     1       2.55ms  2.68ms  2.69ms  2.94ms  4%      ▂▃▁▁▄▄▅▅▄▃█▄▃▇▂▁▃▃▂▂    nonrendering    SkBitmapFPScaleXClampYClampFiltersRGB
  10/10  MB     4       59.9µs  60.6µs  61.6µs  71.1µs  5%      ▂▃▄▁▁▁▁█▇▁▂▁▁▁▁▁▁▁▁▁    nonrendering    SkBitmapFPScaleXClampYClampNearestOrig
  10/10  MB     1       534µs   560µs   559µs   613µs   5%      ▇▂▁▅▅▅█▅▅▃▁▁▁▁▂▄▅▁▁▁    nonrendering    SkBitmapFPScaleXClampYClampNearestLinr
  10/10  MB     1       639µs   658µs   666µs   731µs   5%      ▅█▃▂▃▇▆▁▁▁▁▁▂▁▁▁▁▄█▄    nonrendering    SkBitmapFPScaleXClampYClampNearestsRGB
  10/10  MB     4       61.6µs  65.7µs  65.5µs  70.4µs  5%      ▃▃▃█▆▆▂▂▂▅▄▄▇▄▁▂▁▄██    nonrendering    SkBitmapFPIdentityXClampYClampFilterOrig
  10/10  MB     1       2.22ms  2.34ms  2.35ms  2.57ms  3%      ▂█▅▄▃▃▃▄▂▄▄▄▂▄▁▆▃▁▄▃    nonrendering    SkBitmapFPIdentityXClampYClampFilterLinr
  10/10  MB     1       2.42ms  2.52ms  2.52ms  2.76ms  4%      ▁▁▄▂▄▄▃▅█▄▅▁▃▁▁▃▁▄▃▃    nonrendering    SkBitmapFPIdentityXClampYClampFiltersRGB
  10/10  MB     4       59.7µs  64.7µs  64.8µs  73.3µs  7%      ▂▂▄▁▂▁▁▂▁▁▆▄▆▆▇▄▆▇▂█    nonrendering    SkBitmapFPIdentityXClampYClampNearestOrig
  10/10  MB     1       376µs   388µs   387µs   429µs   4%      ▃▃▆▁▄▄▁▃█▁▁▁▁▁▃▃▄▁▁▁    nonrendering    SkBitmapFPIdentityXClampYClampNearestLinr
  10/10  MB     1       496µs   511µs   518µs   568µs   5%      ▂▆▆█▆▃▂▂█▇▃▁▁▁▁▁▁▁▁▁    nonrendering    SkBitmapFPIdentityXClampYClampNearestsRGB

Review URL: https://codereview.chromium.org/1711963003

DM: remove unnecessary use of std::function

This draw_to_canvas() function doesn't need the power of std::function.

This skips a copy or two, which is nice, and seems to clear
up most of my MSAN problems.  Why?  I do not know.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1718633002

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-MSAN-Trybot

Review URL: https://codereview.chromium.org/1718633002

Changes to merge vulkan code

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711793004

Review URL: https://codereview.chromium.org/1711793004

Calculate translations to apply to vertices in batch

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1716443005

Review URL: https://codereview.chromium.org/1716443005

check pm swizzle when extracting lcd coverage

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1706283005

Review URL: https://codereview.chromium.org/1706283005

remove supports4f flag for colorfilters

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720443002

Review URL: https://codereview.chromium.org/1720443002

Insert an empty image when we cannot decode

In SKP deserialization, if there is a 0 byte encoded image, we insert
an empty image into the output SkPicture so that we can continue
deserializing. Do the same when our decoder cannot decode it. This may
be the result of a bug in our decoder, or in the stream serialization.
But it should not mean the entire stream is invalid.

BUG=skia:4691
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715853002

Review URL: https://codereview.chromium.org/1715853002

gyp_skia: use path python, be quiet

/usr/bin/python might not be python 2.x

I'm getting annoyed by scripts printing messages when there's no problem.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1716493003

Review URL: https://codereview.chromium.org/1716493003

Update libpng to 1.6.22beta

Intel SSE filter optimizations have been upstreamed.  Let's test on
the upstream version (which we can use in Android).

BUG=skia:4573
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1710083007

Review URL: https://codereview.chromium.org/1710083007

Add GPU-backed SkImage case to ImageFilterCache unit test

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713073002

Review URL: https://codereview.chromium.org/1713073002

NEON f32 <-> f16 and f32 <-> u16

Adds f32 <-> f16 ARMv7 and ARMv8 NEON code.
Also adds NEON f32 <-> u16 code to make the comparison fair.

The NDK GCC does not support the ARMv8 NEON intrinsics needed to go fastest, so we use a tiny amount of inline assembly.

The ARMv7 half -> float is different enough from the SSE version that it does not make sense to use SkNx.

Still TODO:
ARMv7 float -> half.  Naively translating the SSE version results in 0x0000 where we'd expect a denormal output.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1700473003
CQ_EXTRA_TRYBOTS=client.skia.android:Test-Android-GCC-Nexus5-CPU-NEON-Arm7-Release-Trybot,Test-Android-GCC-Nexus9-CPU-Denver-Arm64-Release-Trybot;client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1700473003

add back some basic configs to MSAN bot

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-MSAN-Trybot

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1706373004

Review URL: https://codereview.chromium.org/1706373004

Revert of Use unorm shorts for texture coordinates when rendering text. (patchset #3 id:40001 of https://codereview.chromium.org/1713693002/ )

Reason for revert:
Causing issues with text on Mali 400s. Examples:  largeglyphblur, imageblurtiled. It appears that there are precision problems.

Original issue's description:
> Use unorm shorts for texture coordinates when rendering text.
>
> There are a couple of reasons for this:
> - Vulkan does not guarantee conversions from integral vertex attributes
>   to floating point shader variables
> - This may be faster and more precise on some platforms, as it avoids
>   the aforementioned conversion and changes a multiply by a very small
>   value to a multiply by a medium-sized value.
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713693002
> TBR=bsalomon@google.com
>
> Committed: https://skia.googlesource.com/skia/+/e507ff0460f4f878214b9454fb5b9ab8d64d8063

TBR=joshualitt@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/1709133003

added /enableGPU command to skiaserve
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715503004

Review URL: https://codereview.chromium.org/1715503004

Update SkImageFilter's cache to handle SkSpecialImages and add unit test

This is calved off of: https://codereview.chromium.org/1695823002/ (Get OffsetImageFilter really working with SkSpecialImages)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1709263002
TBR=bsalomon@google.com

Review URL: https://codereview.chromium.org/1709263002

fix iter to document close correctly

R=reed@google.com
BUG=skia:4950
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1717623002

Review URL: https://codereview.chromium.org/1717623002

Use unorm shorts for texture coordinates when rendering text.

There are a couple of reasons for this:
- Vulkan does not guarantee conversions from integral vertex attributes
  to floating point shader variables
- This may be faster and more precise on some platforms, as it avoids
  the aforementioned conversion and changes a multiply by a very small
  value to a multiply by a medium-sized value.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713693002
TBR=bsalomon@google.com

Review URL: https://codereview.chromium.org/1713693002

Update the version of libpng used by Skia tests

The version of libpng we were using has some security vulnerabilities.
Update to a newer version - 1.6.20 (http://sourceforge.net/p/libpng/code/ci/047737496a77eeb97f9991919dac08ca3c149711/).

Update pnglibconf.h. The new version is a copy of the newer version of
the file, with our /* custom settings */ tacked onto the end. (The
custom settings are unchanged from the last version.)

Update the LICENSE. The new version is a copy of the LICENSE that
libpng now includes as a separate file.

Update the README file to reflect the change in source of the LICENSE
file.

BUG=skia:4573
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1453373003

Review URL: https://codereview.chromium.org/1453373003

Test build scheduler.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711253003

NOTRY=true
TBR=

Review URL: https://codereview.chromium.org/1711253003

remove xferu64 gm -- not needed now that the blitters are hooked up

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1710283002

TBR=

Review URL: https://codereview.chromium.org/1710283002

fix red/blue order in 64bit blits

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715513004

Review URL: https://codereview.chromium.org/1715513004

Switch to pixmap in API.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1712653002

Review URL: https://codereview.chromium.org/1712653002

GrAuditTrail can now be enabled/disabled at runtime
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1712753002

TBR=bsalomon@google.com
Doesn't actually change the public API

Review URL: https://codereview.chromium.org/1712753002

make SkPM4f private

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1713653002

Review URL: https://codereview.chromium.org/1713653002

Add more specialized fragment builders

Adds specialized fragment builders for primitive and fragment
processors. Removes fragment-specific functionality from the Xfer
fragment builder.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1709153002

Review URL: https://codereview.chromium.org/1709153002

Move SkTInternalLList.h to src/core.

TBR=reed
Not intended for external use.

Review URL: https://codereview.chromium.org/1712563004

MSAN: cut down to green

I have no idea what's going on with the current bot failures.
This disables lots of things to get it green.
Probably more than necessary.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1707873002

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-MSAN-Trybot

Review URL: https://codereview.chromium.org/1707873002

Move SkTDStack.h to src/animator.

This file is only used by animator and may or may not be sane. Move
the file from include/core to src/animator since no one else is using
it.

TBR=reed
Make it go away!

Review URL: https://codereview.chromium.org/1710083002

fixed SkDrawCommand JSON generation
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1715443002

Review URL: https://codereview.chromium.org/1715443002

added server-side breakpoint support to skiaserve
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1707863003

Review URL: https://codereview.chromium.org/1707863003

Add luci-go isolate.sha1 files

NOTRY=true
BUG=skia:3775, skia:4763
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711013002

Review URL: https://codereview.chromium.org/1711013002

Make build_command_buffer.py work on Windows

BUG=skia:4957
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1710983002

Review URL: https://codereview.chromium.org/1710983002

check for zero conic weights in interp

this will fix the current (and uncover the next)
ASAN failure

TBR=tomhudson@google.com
BUG=skia:4549
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711813002

Review URL: https://codereview.chromium.org/1711813002

Move SkTDArray to private.

TBR=reed
Moving to private is good.

Review URL: https://codereview.chromium.org/1707213002

Update SkSpecialImage API

This adds some temporary compatibility features:
   internal_getBM & internal_fromBM - to convert back & forth from Bitmaps in the ImageFilters
   internal_getProxy - to allow the special images to interface with the current ImagerFilter system

It also adds a unique ID, opaque flag and size method.
   The unique ID and size method are needed for the image filter's caching system.
   The opaque flag is need for swapping back & forth from Bitmaps

These are all calved off of https://codereview.chromium.org/1695823002/ (Get OffsetImageFilter really working with SkSpecialImages) which actually makes use of them.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1705113002

Review URL: https://codereview.chromium.org/1705113002

lcd blits for sRGB

Unimplemented for F16 for the moment, but not needed yet.
Surprisingly not much slower than current impl (not srgb correct)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1707883002

Review URL: https://codereview.chromium.org/1707883002

Port Skia recipe to normal Python scripts, move to Skia repo

BUG=skia:4763
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1703663002

Review URL: https://codereview.chromium.org/1703663002

Remove FuzzPaeth now that we have some real Fuzzes.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711753002

Review URL: https://codereview.chromium.org/1711753002

Make build_command_buffer.py work on Mac and Windows.

Also specify GYP_GENERATORS and add flags to gclient sync to make it more likely to succeed.

BUG=skia:4957
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1704193002

Review URL: https://codereview.chromium.org/1704193002

Create ParsePath API fuzz

This is based on https://codereview.chromium.org/1675053002

BUG=skia:4438
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1702383003

Review URL: https://codereview.chromium.org/1702383003

remove legacy flags

R=fmalita@chromium.org
BUG=573166
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1707383002

Review URL: https://codereview.chromium.org/1707383002

fix asan error deref-ing null ptr

TBR=tomhudson@google.com
BUG=skia:4549
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1709023002

Review URL: https://codereview.chromium.org/1709023002

Remove position from FrontBufferedStream

When FrontBufferedStream was written, it implemented getPosition()
and returned true for hasPosition(). The CL that introduced it
(crrev.com/23717055) does not have any indication why, but I'm guessing
it was because it was easy to implement. None of our decoders rely on
this (only some tests do).

Now that we have a decoder (SkRawCodec) that expects to be able to seek
a stream if it has a position (which makes sense - SkStream.h
associates that with SkStreamSeekable, and there is no other way to
check to see if a stream is seekable), it is failing because
FrontBufferedStream reports it has a position and the decoder tries to
seek.

Remove FrontBufferedStream::hasPosition() (reverting to the default,
false) and ::getPosition() (so it will return 0).

Fix tests - do not call FrontBufferedStream::getPosition()

Update CodexTest to test using an FrontBufferedStream, like Android
does.

BUG=b/27218441
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1703293002

Review URL: https://codereview.chromium.org/1703293002

remove legacy bitmap filter

R=fmalita@chromium.org
BUG=skia:2261
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1711683002

Review URL: https://codereview.chromium.org/1711683002

fix comma in GrAuditTrail json

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1704133002

Review URL: https://codereview.chromium.org/1704133002

add interp path

Add path methods to determine if a pair of paths
can be interpolated, and to interpolate them.

R=reed@google.com, robertphillips@google.com
BUG=skia:4549
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1703943003

Review URL: https://codereview.chromium.org/1703943003

Add bilerp filtering.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1705203002

Review URL: https://codereview.chromium.org/1705203002

some modes need clamping due to numerical imprecision

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1710453003

TBR=

Review URL: https://codereview.chromium.org/1710453003