Create geometry shader modules in vulkan when using geom shaders

Looks like this was left off the original vulkan Geo shader change.
Bug: skia:
Change-Id: Iec1c09d6d7b347147e4237bad3e6b3eb4684154e
Reviewed-on: https://skia-review.googlesource.com/15813
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>

Fixed distance field interpolation for gamma correct rendering

Bug: skia:
Change-Id: I9037decfeb9b2ab673434bb3c4b7f55aca186d68
Reviewed-on: https://skia-review.googlesource.com/13872
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

Added SkImage::MakeCrossContextFromEncoded

Designed for Flutter's threading architecture, with
an eye to being useful to other clients. Under the
hood, uses a new image generator class to lazily wrap
a texture for multiple GrContexts.

Re-land of https://skia-review.googlesource.com/c/14180/

Bug: skia:
Change-Id: I3dd382640629b79b3058f18fee68d043566e43e5
Reviewed-on: https://skia-review.googlesource.com/15895
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

Manually roll recipes.

Include new --package option for recipes.py.

build:
  https://crrev.com/380216596a663dc45d6562b29bce3719e2620ac0 Bump bot environment "vpython". (dnj@chromium.org)
  https://crrev.com/226d6eadceee58fe6832679e39a040a46d5df989 Bump Kitchen canary, clear PYTHONPATH. (dnj@chromium.org)
  https://crrev.com/45f9beb12883c6b509c699c1dd1694be218c4740 Write uploaded log url to file in goma module (tikuta@google.com)
  https://crrev.com/5c6b5807ee3de9277b8c6e3182f621e4216e2eea WebRTC: Rename tools-webrtc -> tools_webrtc (kjellander@chromium.org)
  https://crrev.com/3090e9f2c42f7254fc9b42b21ed792e19d77e667 Manually roll depot_tools 69a239e:e2f9fee (phajdan.jr@chromium.org)
  https://crrev.com/e845e79d0cd39af876b4fff648c68bd023633c8a chromium_checkout: add non-fatal gclient validate step (phajdan.jr@chromium.org)
  https://crrev.com/3bed2fc03bc435464880411740b20031f80ed41c Remove old GYP compatibility targets from chromium_tests. (RELAND) (jbudorick@chromium.org)
  https://crrev.com/9b44962a1bfb8134ffac7f5606a11cb8ff8489e8 Flutter: Rev Android build tools to 23.0.3 (mit@google.com)
  https://crrev.com/6fa4d503d7ef77a951b446ca01b5b18a3c9fb79c remote_run: promote Kitchen canary to stable. (dnj@chromium.org)

depot_tools:
  https://crrev.com/4ebaaf9d9c1e24548f3c9f9f748e8631d091ecd0 Bump "vpython" version. (dnj@chromium.org)
  https://crrev.com/0bbd1c28d5827379f363f2a9170eeda27fd7b758 git-cl: upload to merge-base with master, not tip of origin/master (agable@chromium.org)
  https://crrev.com/e2f9feecaf4326cef699624d5f52b9f31ecbc104 Add validate command to gclient (phajdan.jr@chromium.org)

recipe_engine:
  https://crrev.com/66338449f4d727da096422c1df1b6654224d6512 [bootstrap] don't change directories when bootstrapping. (iannucci@chromium.org)
  https://crrev.com/487e5e371bf8140b0467c6366f94042581f5c63c [recipes.py] refactor PackageContext to remove unused --deps-path arg. (iannucci@chromium.org)
  https://crrev.com/92c5b73ff81fb30c80a33bab91be51551d5e67b7 Bump "vpython" bootstrap version. (dnj@chromium.org)
  https://crrev.com/d6c4a597be2da8be7971b4d5f66e3d95e8fc325c [recipes.py] refactor PackageContext to remove --no-fetch logic. (iannucci@chromium.org)
  https://crrev.com/cfeeb53051e02856a35395a69b09f822ad35e2e5 [doc/recipes.py] Improve copypasta bootstrap script. (iannucci@chromium.org)
  https://crrev.com/e90c11af854574dd68edd0e7c7a5e04ceb49ffd0 [recipes.cfg] remove support for api_version 1. (iannucci@chromium.org)
  https://crrev.com/497f5b7f92cea8e5273d56cc3cba876fb18b42d9 [doc/recipes.py] fix issue when running recipes.py from . (iannucci@chromium.org)
  https://crrev.com/e7bbbf6cfac944157d80cc92fd78837165e69d3f Fix a race crash when recipe expectations directory exists (phajdan.jr@chromium.org)
  https://crrev.com/bf2ff828987549b946135919bd1bffd6e42fd8fd [doc/recipes.py] Add option to recipes.py to allow it to work without .git (iannucci@chromium.org)

R=borenet@google.com

Bug: skia:
Change-Id: Ic7c20a7beb6165ea79194053016759722fea6a51
Reviewed-on: https://skia-review.googlesource.com/16201
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Eric Boren <borenet@google.com>

Add downsampling option to SkJpegEncoder

Bug: 713862
Change-Id: Ibad6ecf836ccfd355499e1bace7bcd4ba772a97a
Reviewed-on: https://skia-review.googlesource.com/15891
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>

update dox for dithering

Bug: skia:6604
Change-Id: I21b505c54c1d43abf62b7c577e7dd88e53ed30fd
Reviewed-on: https://skia-review.googlesource.com/16156
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Mike Reed <reed@google.com>

move set-text-matrix into private

Had no external callers, and was poorly named.

Bug: skia:
Change-Id: I5bed6853e9aa3b31fa106a436a73297bf55f0503
Reviewed-on: https://skia-review.googlesource.com/16155
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Mike Reed <reed@google.com>

Add jpeg encoder alpha handling option

This instructs us on how to encode jpegs when the src
image has alpha.  The original behavior is to ignore
the alpha channel.  This CL adds the option to blend
the pixels onto opaque black.

Note that kBlendOnBlack and kIgnore are identical
unless the input alpha type is kUnpremul.

Bug: 713862
Bug: skia:1501
Change-Id: I4891c70bb0ccd83f7974c359bd40a2143b5c49ac
Reviewed-on: https://skia-review.googlesource.com/15817
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

make append_gamut_tranform() never fail

The only way it could ostensibly fail is if we get non-XYZ color
spaces, which should just not happen.  Assert that doesn't happen
and safely do nothing instead of failing.

This is one of the leaf nodes to getting SkCreateRasterPipelineBlitter
to never fail.  Next come SkColorFilter:: and SkShader::appendStages().

Change-Id: I5c7a8c63d0a9837e2e55208e1674796d86f45307
Reviewed-on: https://skia-review.googlesource.com/16002
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Matt Sarett <msarett@google.com>

Add support for using texel buffers in Vulkan backend.

Bug: skia:
Change-Id: I5cc7b18263dbe28d5d8d89301111ef240109704a
Reviewed-on: https://skia-review.googlesource.com/15770
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Tune down libpng in DM.

Timing `out/dm --src gm --config srgb -w foo` on my desktop,
   - wall time drops from 3.72 -> 2.17 seconds; and
   - CPU time drops from 64.9 -> 27.9 seconds.

So again, close to 2x speedup.
(In exchange, the .pngs of our GMs grow from 57M to 73M.)

Change-Id: I6e2e8c8bda6371e857c8db41ca3119402a9056ae
Reviewed-on: https://skia-review.googlesource.com/16000
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Move SkPngEncoder into public API

Bug: 713862
Change-Id: I45068ed39affe41ffe0f29bf42c5ea1d9b0247ba
Reviewed-on: https://skia-review.googlesource.com/15897
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Make SkColorFilter::appendStages() not fail.

This makes SkColorFilter::appendStages() first try onAppendStages(),
and if it's unimplemented or fails, fall back to filterSpan4f().

This also makes onAppendStages() private to try to ensure that
appendStages() is now its only caller, ensuring everyone goes
through this fallback path.

The fallback uses the color filter transformed into the dst colorspace
using our new SkColorSpaceXformer... that seem ok Matt?

Change-Id: I4751a6859596fa4f7e844e69ef0d986f005b52c7
Reviewed-on: https://skia-review.googlesource.com/16031
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Avoid assertion in MSAA rendering buffer about indexed draw with zero index count

Also add some GrMesh::validate() calls earlier to make future debugging easier. Previously this only asserted when the draws executed, now it asserts as soon as the draw is prepared in the op subclass.

Bug: skia:
Change-Id: Ibdd4488b9a9263f0ed337a2c8c6066bc36297001
Reviewed-on: https://skia-review.googlesource.com/16029
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>

Add flag to avoid stencil buffers in Skia

Certain systems experience a leak in the GL driver associated with
stencil buffers. Attempts to avoid the leak (while still using stencil
buffers) dind't succeed. This patch adds a GrContextOption
fAvoidStencilBuffers. This disables certain path rendering modes, as
well as stencil based masking/clipping.

Bug: 713854
Change-Id: Ifa6c0f2bd5ee395547bda9165d6c79d197ae8b8b
Reviewed-on: https://skia-review.googlesource.com/15253
Commit-Queue: Eric Karl <ericrk@chromium.org>
Reviewed-by: Eric Karl <ericrk@chromium.org>
Reviewed-by: Brian Salomon <bsalomon@google.com>

SkPerlinNoiseShader: handle pathological inputs

BUG=skia:6429
Change-Id: Ifc20c1f480c1396e40ab7fc3a3789aa7c2ae5d85
Reviewed-on: https://skia-review.googlesource.com/15896
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>

Convert color to dst (once) in color shaders

BUG=skia:

Change-Id: Ideba16972b82fc26766349c808ae406e5ea23163
Reviewed-on: https://skia-review.googlesource.com/9418
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Matt Sarett <msarett@google.com>

Add debug-only helper methods to get #ops & #clips per opList

These are being/will be used to determine the correct amout of memory to preallocate for ops & clips in the opLists.

Change-Id: I98ebaec8a6e72a43d97101aca5fbc58264964ebd
Reviewed-on: https://skia-review.googlesource.com/15882
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

[recipes] Remove inputs to BundleRecipes task

This should enable deduplication.

Bug: skia:
Change-Id: I2be98f648acd666c420885f08377c256f9102a8f
Reviewed-on: https://skia-review.googlesource.com/14651
Commit-Queue: Eric Boren <borenet@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>

Remove 'fDrawFace' from GrPipeline

This will, hopefully, unblock the roll.

Change-Id: I2e33ccca7161334cdecd881e2699ae0c61ba2a31
Reviewed-on: https://skia-review.googlesource.com/16101
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>

Forward getGrContext in color space xform canvas

Bug: chromium:719664
Change-Id: I7477c1eb0479d5305233dc6a643280d88029bd17
Reviewed-on: https://skia-review.googlesource.com/15888
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

remove (possibly slow) call to refEncoded in getDeferredTextureImageData

- explicitly reject already-texture-backed and picture-backed

Needed to add a virtual to image_base and generator to distinguish
generators that can (or cannot) natively "generate" on the gpu (e.g. pictures)
Bug: 646089
Change-Id: I3aea22f89b31009ecbb3bd50d88512e6532f0a0f

Change-Id: I3aea22f89b31009ecbb3bd50d88512e6532f0a0f
Reviewed-on: https://skia-review.googlesource.com/15765
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Add radial gradient stage.

Change-Id: Ie1f9640f5149f21bd8b3b864ff8b176232e1b0a9
Reviewed-on: https://skia-review.googlesource.com/15461
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Herb Derby <herb@google.com>

Roll recipe dependencies (trivial).

This is an automated CL created by the recipe roller. This CL rolls recipe
changes from upstream projects (e.g. depot_tools) into downstream projects
(e.g. tools/build).

More info is at https://goo.gl/zkKdpD. Use https://goo.gl/noib3a to file a bug.
build:
  https://crrev.com/59b248f278e4828bbb0a8d8685975856a7f60509 Stop manually adding webkit_layout_tests if it's already configured src-side. (mcgreevy@chromium.org)
  https://crrev.com/ff38188006831a2f281b68c091a0c02f3fd51ec1 Migrate remaining chromium configs from depot_tools (machenbach@chromium.org)
  https://crrev.com/20950290930a4e7e2fb51f627b0066e883eefb3c Refactoring: rename obsolete v8 config (machenbach@chromium.org)
  https://crrev.com/4ac4debb7b303e5ce213acee4d0d6777299121be remote_run: Bump Kitchen, add LogDog tags. (dnj@chromium.org)
  https://crrev.com/f65b85259d62281f29e37ae28b002f6a98a486d8 Revert Kitchen bump. (dnj@chromium.org)
  https://crrev.com/577ed8c48dcad41653d0559a119abbb84f461afe remote_run: Decouple canary from Kitchen. (dnj@chromium.org)
  https://crrev.com/fa6566763ff505e21cb7a012ae31b363dc08aad6 Promote Win32 ASan to the main waterfall and remove Win SyzyASan. (ehmaldonado@chromium.org)
  https://crrev.com/671dc2edd993a46cb4d34621ed5e4c01b269d3a8 Enforce strict coverage for syzygy recipe module (phajdan.jr@chromium.org)
depot_tools:
  https://crrev.com/e39a9a46be39aac5fad2d66cf11ccdd6f8b9f081 Migrate remaining chromium configs to build repo (machenbach@chromium.org)
  https://crrev.com/b4a79690367881c427c9c5adf614823a01fc9990 Add back helper method still used in build repo (machenbach@chromium.org)

TBR=borenet@google.com

Recipe-Tryjob-Bypass-Reason: Autoroller
Bugdroid-Send-Email: False
Change-Id: I0e5b52a9dac9408276d4adb021a20b1a6ec4c524
Reviewed-on: https://skia-review.googlesource.com/15883
Reviewed-by: Recipe Roller <recipe-roller@chromium.org>
Commit-Queue: Recipe Roller <recipe-roller@chromium.org>

Implement disconnect function for GrVkGpu

Apparently this was not added to Vk when it was added to GL. It may be
the cause of a flutter error they are seeing.
Bug: skia:
Change-Id: I4d4e9d1817b308b6c4f01e83f712eb3437e2ad71
Reviewed-on: https://skia-review.googlesource.com/15887
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>

Minor sksl parser cleanup

This adds a new method "checkNext" which is essentially a conditional
nextToken(). This allows for some simplification of the parser code,
but has no other effect.

Bug: skia:
Change-Id: I4e08fd5f2c69cbcb4c31fd936b3a65dc50347191
Reviewed-on: https://skia-review.googlesource.com/15881
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>

add a guide to contributing to SkJumper

Change-Id: Icd40cf5eff3d2156a3ca00d7950059d5b77f48bf
Reviewed-on: https://skia-review.googlesource.com/15890
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Added SkImage::MakeCrossContextFromEncoded"

This reverts commit 83b1b3db36e8622f4bbc8c391d5c714e258d1e03.

Reason for revert: unit test failing and asan leaks

Original change's description:
> Added SkImage::MakeCrossContextFromEncoded
>
> Designed for Flutter's threading architecture, with
> an eye to being useful to other clients. Under the
> hood, uses a new image generator class to lazily wrap
> a texture for multiple GrContexts.
>
> Bug: skia:
> Change-Id: I6c37b12c8ab5bce94b91190e5f0beb91d31ae81b
> Reviewed-on: https://skia-review.googlesource.com/14180
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>

TBR=egdaniel@google.com,bsalomon@google.com,robertphillips@google.com,brianosman@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I1f0bf580aa0ea2d132e18b64ff610ddac9d073a7
Reviewed-on: https://skia-review.googlesource.com/15892
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>

fixed InstanceProcessor SkSL code

Bug: skia:5935
Change-Id: I9ca0e8a7bd30886d8240e338eefbfab368935ac5
Reviewed-on: https://skia-review.googlesource.com/15627
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>

"can we?" -> "do we want to?" for SkRasterPipelineBlitter

There has been a supported() function in SkRasterPipelineBlitter.cpp for
a long time that's becoming increasingly misnamed.  That blitter ought
to be able to handle all destination formats.

This CL moves that logic outside to the creator of the blitter, changing
it from "can we handle this format?" to "do we want to use this blitter
for this format?".

In other CLs I'm working to make creating a pipeline blitter never fail.

Change-Id: Ie59fb8ec6e63d215d1baef439e464e8f0ab3ae4d
Reviewed-on: https://skia-review.googlesource.com/15842
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Added SkImage::MakeCrossContextFromEncoded

Designed for Flutter's threading architecture, with
an eye to being useful to other clients. Under the
hood, uses a new image generator class to lazily wrap
a texture for multiple GrContexts.

Bug: skia:
Change-Id: I6c37b12c8ab5bce94b91190e5f0beb91d31ae81b
Reviewed-on: https://skia-review.googlesource.com/14180
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=update-skps@skia.org
NO_MERGE_BUILDS

Change-Id: I8919465322d3621689bcab4bb577b42b9fd736d7
Reviewed-on: https://skia-review.googlesource.com/15920
Reviewed-by: update-skps <update-skps@skia.org>
Commit-Queue: update-skps <update-skps@skia.org>

SkAdvancedTypefaceMetrics: pack fields better

    64-bit pointers: 88 -> 80 bytes
    32-bit pointers: 68 -> 64 bytes

Change-Id: I2d6e186d15ad84a3b23bf8f6c816eaf482c3bdd5
Reviewed-on: https://skia-review.googlesource.com/15878
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

fix commented-out gaussian

The original was

  static inline float eval_gaussian(float x) {
      float factor = 1 - x;
      return sk_float_exp(-factor * factor * 4) - 0.018f;
  }

Change-Id: I0590f9b4b41870a0145a4fb7aff5ce3fa4d4e412
Reviewed-on: https://skia-review.googlesource.com/15781
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Remove height functor for shadows and replace with plane equation params

Change-Id: I948eceb2c58dc50468993dba54c209f18e440e48
Reviewed-on: https://skia-review.googlesource.com/15873
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>

vk: don't re-check the primitive type for each batch of a mesh

Bug: skia:
Change-Id: I00561d0a2d6296924ef0d8ee25fc7215832338d8
Reviewed-on: https://skia-review.googlesource.com/15832
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>

Revert "Revert "Re-enable deferred proxies in Chrome""

This reverts commit 05814de6ba5087ad71f189d6413246ef1d518e4b.

Reason for revert: Chrome layout test suppressions have landed

Original change's description:
> Revert "Re-enable deferred proxies in Chrome"
>
> This reverts commit 555c49c3479edf1692579c06a5c13c39d3741692.
>
> Reason for revert: layout tests
>
> Original change's description:
> > Re-enable deferred proxies in Chrome
> >
> > Now that https://skia-review.googlesource.com/c/14605/ (Allow TextureSamplers to have null GrTexture pointer) has landed we should be able to re-enable deferred proxies in Chrome.
> >
> > Bug: 715488
> > Change-Id: Ibd38e5ddf0558b94f09e29f17837e6abfd22fa34
> > Reviewed-on: https://skia-review.googlesource.com/15252
> > Reviewed-by: Brian Salomon <bsalomon@google.com>
> > Commit-Queue: Robert Phillips <robertphillips@google.com>
> >
>
> TBR=bsalomon@google.com,robertphillips@google.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Change-Id: Ie90e64a2e4d5efc4b6909a04f8a862dd3a708d97
> Reviewed-on: https://skia-review.googlesource.com/15809
> Reviewed-by: Robert Phillips <robertphillips@google.com>
> Commit-Queue: Robert Phillips <robertphillips@google.com>
>

TBR=bsalomon@google.com,robertphillips@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I957e8e59af2b2f441c75bae2eda4dcd5ed82b34d
Reviewed-on: https://skia-review.googlesource.com/15876
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

Expand GrTextureProxy to handle highestFilterMode

Once TextureProxies aren't instantiated in the TextureSamplers, the they will need to be able to supply this information.

split out of: https://skia-review.googlesource.com/c/10484/ (Omnibus: Push instantiation of GrTextures later (post TextureSampler))
Change-Id: I66555c0746131f565862f7a30d54ff1d458d2062
Reviewed-on: https://skia-review.googlesource.com/15819
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Minimize computeTotalInverse()'s inputs

The helper doesn't need a full context rec - it only looks at the CTM
and external local matrix.  Pass these explicitly instead.
Change-Id: I6a5884f65cd383c2df0e8d83c9086789bd96f345
Reviewed-on: https://skia-review.googlesource.com/15870
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Florin Malita <fmalita@chromium.org>

GL: track enabled vertex arrays as a count rather than a mask

Bug: skia:
Change-Id: I63e70ab844a7e04df20165aba025b963efcafa9e
Reviewed-on: https://skia-review.googlesource.com/15630
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>

port ok to SkPngEncoder

I can't tell you how excited I am to turn down the libpng settings...
anything that saves me a few seconds is a nice quality of life win.

This CL makes ok run in about half the time when producing .pngs.

Profile running `ok gm srgb png` before:
    10.59   s   16.5%    10.59   s          longest_match
    8.98    s   14.0%    8.98    s          png_setup_paeth_row
    8.93    s   13.9%    8.93    s          skia_png_write_find_filter
    7.75    s   12.1%    7.75    s          deflate_slow
    4.63    s    7.2%    4.63    s          std::wait_until(...) const
    959.00 ms    1.5%    959.00 ms          SkPathRef::validate() const
    935.00 ms    1.4%    935.00 ms          sk_to_srgb_hsw
    ...

After:
    2.35    s    8.7%    2.35    s          std::wait_until(...) const
    1.70    s    6.2%    1.70    s          longest_match
    1.19    s    4.4%    1.19    s          deflate_fast
    931.00 ms    3.4%    931.00 ms          SkPathRef::validate() const
    898.00 ms    3.3%    898.00 ms          sk_to_srgb_hsw
    ...

Change-Id: I425c30b2ecd97a0e4a4392779de6301db473ee47
Reviewed-on: https://skia-review.googlesource.com/15547
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Matt Sarett <msarett@google.com>

Separate lines added via post upload from original description

BUG=skia:6598
NOTRY=true

Change-Id: I23fa415788e91da01d33ab3ab912aa420636574c
Reviewed-on: https://skia-review.googlesource.com/15869
Commit-Queue: Ravi Mistry <rmistry@google.com>
Reviewed-by: Eric Boren <borenet@google.com>

Update minimum non-ES OpenGL spec to 2.0.

As a consequence of removing support for GLs lacking separate stencil settings our effective minimum GL version is 2.0. This formalizes that by explicitly failing if the version in the version string is lower.

As a consequence we will always have wrapping stencil operations available.
Bug: skia:
Change-Id: Ibe302f72a063dbadc773bda811cc0d0f6c08c61a
Reviewed-on: https://skia-review.googlesource.com/15610
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>

Remove support for GLs without separate stencil.

As a consequence we no longer need GrDrawFace.

This effectively raises the minimum bar for non-ES OpenGL to 2.0 as there is no extension that adds the GL 2.0 separate stencil functionality. GL_ATI_separate_stencil is close but it does not have glStencilMaskSeparate.
Bug: skia:
Change-Id: I36d17a69400c8beeacb6dab8d8c5c3317814cfe4
Reviewed-on: https://skia-review.googlesource.com/15603
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>

Add filtering and zlib-level options to SkPngEncoder

Bug: skia:6409
Bug: 713862
Change-Id: If287e2bcad5af990fac11e9091305f45ec903dbf
Reviewed-on: https://skia-review.googlesource.com/15647
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>

Revert "Presubmit: ensure empty line between CL title and description"

This reverts commit 9518830650a45446756d7b621fc87e7a8ff52476.

Reason for revert: <INSERT REASONING HERE>

Original change's description:
> Presubmit: ensure empty line between CL title and description
>
> Bug: skia:
> Change-Id: If1b2d655e9be4b960f76d2d17ce1cf1126f7b2d4
> Reviewed-on: https://skia-review.googlesource.com/15624
> Reviewed-by: Eric Boren <borenet@google.com>
> Commit-Queue: Eric Boren <borenet@google.com>
>

TBR=borenet@google.com,rmistry@google.com
NOTRY=true
# Not skipping CQ checks because original CL landed > 1 day ago.

Change-Id: I311ee790f152dbb8325214650838839f6821c995
Reviewed-on: https://skia-review.googlesource.com/15862
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Ravi Mistry <rmistry@google.com>

Re-land of "eliminated GrGLSLExpr".

This change is exactly the same as the last time it was landed; I believe the
underlying optimizer bug that was causing this to cause problems has been
fixed by a prior CL.
Bug: skia:
Change-Id: I5436422f094ea758caa3cd69e9338db31b1f93fa
Reviewed-on: https://skia-review.googlesource.com/15768
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>

Allow wrapped resources to have unique keys

Previously, wrapped resources were never budgeted. Now we explicitly allow
wrapped, unbudgeted resources with unique keys. This allows code that
wraps (and re-wraps) external resources with a deterministic key to find
the same wrapped resource - saving time and ensuring a single wrapped
copy, to preserve state on the resource (like texture sampler state).

Bug: skia:
Change-Id: I1dd7642f1ed8bb6c620029d46203cf5cb6b3c160
Reviewed-on: https://skia-review.googlesource.com/15241
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

[Bazel] Omit bigrect GM for ASAN.

Added in https://skia-review.googlesource.com/c/14751/; known to fail
float-cast-overflow check but forgot to omit in that CL.

Bug: skia:4632
No-Try: true
Change-Id: Ic62e0cbc00229ab3bdde60b3e7300fc66ad7d8e1
Reviewed-on: https://skia-review.googlesource.com/15815
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

GrTessellator AA: fix 1px wide paths.

Move 0 count check below the call to count outer polygons,
so we don't abort on a path with outer but no inner
geometry.

This fixes the "thin_right_angle" and "thin_rect_and_triangle"
samples when running GM:thinconcavepaths --pr tess.

Change-Id: I9a75711787f729ce3a1d5313daa491098e397184
Reviewed-on: https://skia-review.googlesource.com/15661
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>

Remove Sk4fGradientInterval::fZeroRamp

It's cheap enough to compute when needed, and we don't need it in hot
loops anyway.
Change-Id: Ib0242f98f8bee31bff939cfdc7356d51525092e4
Reviewed-on: https://skia-review.googlesource.com/15764
Commit-Queue: Florin Malita <fmalita@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>

SkImage::isValid

Lets clients know if an image is drawable to a particular GrContext
(or to CPU). Checks for abandoned GrContexts beneath GPU backed
images, as well as context mis-match.

Bug: skia:
Change-Id: Ibe88c7ce8091f965c14f6023a3597be4b70c3f99
Reviewed-on: https://skia-review.googlesource.com/15801
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

jumper, finish blend modes

I've decided to ignore our existing CPU implementations and start from
scratch, mostly referencing the GL ES 3.2 spec and w3 spec.

This implementation ought to look a lot like the reference
implementation I've written in gm/hsl.cpp, with the addition of
handling alpha: unpremul, blend, re-premul with a simple SrcOver alpha.

Change-Id: I38cf6be2dc66a6f46d7b18b91847f6933d2fab62
Reviewed-on: https://skia-review.googlesource.com/15316
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Re-enable deferred proxies in Chrome"

This reverts commit 555c49c3479edf1692579c06a5c13c39d3741692.

Reason for revert: layout tests

Original change's description:
> Re-enable deferred proxies in Chrome
>
> Now that https://skia-review.googlesource.com/c/14605/ (Allow TextureSamplers to have null GrTexture pointer) has landed we should be able to re-enable deferred proxies in Chrome.
>
> Bug: 715488
> Change-Id: Ibd38e5ddf0558b94f09e29f17837e6abfd22fa34
> Reviewed-on: https://skia-review.googlesource.com/15252
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Robert Phillips <robertphillips@google.com>
>

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

Change-Id: Ie90e64a2e4d5efc4b6909a04f8a862dd3a708d97
Reviewed-on: https://skia-review.googlesource.com/15809
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

Tighten up gradient clamp logic.

This is just a little refactor.  Shouldn't change any logic or pixels.

Change-Id: I782df78f2d693dc7e35c286b049730e763aa3dc8
Reviewed-on: https://skia-review.googlesource.com/15653
Reviewed-by: Florin Malita <fmalita@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Add GrVkTexelBuffer class

Bug: skia:
Change-Id: I6111902f58df3af92702d974a52d9c7f267634d4
Reviewed-on: https://skia-review.googlesource.com/15628
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Manually roll recipes.

build:
  crrev.com/50e26da30b6cde158a39b6bdc99ddb2ba1a47614 Enforce strict coverage for isolate recipe module
  crrev.com/9f1038e0bc39f6348ce21f4af7cd9ad07184068d chromium_tests: Fix gtest validation when no valid json is present.
  crrev.com/d262a9497b79aa5352ea85c68f4b132872127e37 Buildbot pubsub: Remove LRU for buildsets
  crrev.com/95f0d5777e7c3db976c941758b2caa447f3c9872 remote_run: add support for new kitchen result format
  crrev.com/a081bfaddc1f94073a0ab73f31e7870dea764e27 V8: Bump shards on slowest bot
  crrev.com/0cd7b48611fc2feadb6348a2e16fcb509d45f44f WebRTC: Add logdog cipd package install for webrtc_perf_tests
  crrev.com/bb9f32300c4ca57516ee1b7c35579fe75553b6c6 Switch Android nexus 5x bot over to swarming
  crrev.com/db4dfab8c7f51f5527ecb0ec8d71bb38a45540b1 master.chromium.infra.cron; fix typo in slave name
  crrev.com/178e486407840b54b93429d92d2629c1ef731c1e CrOS: Update Chromite pin.
  crrev.com/7022f73461b3fa296e0acede710d80b2f5436b14 Add new machine to master.chromiumos
  crrev.com/e26fdac7c1d6928a70bdfee80ae267f506ccc35b Enforce strict coverage for chromium_tests recipe module
  crrev.com/2c630ad197926ac96d6600472d8082a2439322b6 Merge build requests on testers on the chromium.chromiumos waterfall.
  crrev.com/3fbbeef094717d4e371d78ec0fe2ca71bcaac765 Reland #2: Run "supersize archive" on Android perf builders
  crrev.com/2c5dd84cc74d0a650aa4cdc040daa8103124a7c4 Revert "Reland: Run "supersize archive" on Android perf builders"
  crrev.com/08b150c6d6e6f3208803bda4ea6af95012008596 Fetch Telemetry binary dependencies on main Linux and Mac builders.
  crrev.com/56598af93d1cf5bd853039c2de8b14ad358bb996 Bisect - Perf-Try - Don't call bot_update on the patched run.
  crrev.com/cccc78db18bb9c80d1689afe883b8131ec878a1f Reland: Run "supersize archive" on Android perf builders
  crrev.com/1130031d6393e79d1954c3554ce994babc5f3ea5 Revert "Run "supersize archive" on Android perf builders"
  crrev.com/37b16786dee83e4e72e8be310c68d46d3da7b19c Revert "Disable kitchen on Site Isolation Linux for now."
  crrev.com/9d59ac6854b57e8d12ce5fcc3266fd413521cc49 Run "supersize archive" on Android perf builders
  crrev.com/f25c842180da0df876d765503b0edddc0f95e048 Don't pin the SyzyAsan bots to VS2015.
  crrev.com/30d8d501de8b35a7e97060ed6b03f737dd424652 Rename android_webview_test_apk to webview_instrumentation_test_apk
  crrev.com/a53da2efbfa651d6437cbbedbe6222d512d3ef65 disconnect 3 masters from buildbucket
  crrev.com/7a08020351a41f514581c7ca75ac38beaef02975 Dart: allocate Windows slaves to integration builders for Dartium roll
  crrev.com/ce235777011d4277ccd791b8e53ac55e28954f2a WebRTC: Remove --timeout=900 from the recipe.
  crrev.com/b810b4628c808d13c4b6481db5693a1e6e0bd0d2 Add Chromium's gclient configs
  crrev.com/ef5479c25e00f2fe22d7f7011236fb313cb56b16 V8: Add gcc debug bots
  crrev.com/f55ee770c2582d2f27778ea78a7736aa8ef82a45 Add speed-profile in the buildbot steps.
  crrev.com/039b01833d51bfdf37a918a1162ac7b2019fa81d Roll recipes.
  crrev.com/0118b6411bc3ea9406b3e360715577b1f7eaeb4b Set fetch_telemetry_dependencies=1 on GPU FYI Mac/Linux builders.
  crrev.com/09b8f2f7f2de95dc3515487a32fa330667265e20 Simplify code which controls adding swarmed or local layout tests.
  crrev.com/da1809cb625558cf7c29df694587997dde022d3f Revert "Roll recipes."

depot_tools:
  crrev.com/a469aeec53a5983550e4b6a61e569971cb583b36 Remove Chromium's gclient configs from depot_tools
  crrev.com/362bc168c49d9e7b1223d126f5e0a2436eaeb7f3 Revert "[git_bootstrap.py] Remove unused options, use cipd in %PATH%."
  crrev.com/0f2e6c53db1b464e3dc7637c00c6251b864ae539 Add vpython.bat.

recipe_engine:
  crrev.com/5cdf9803d55420f3ae8d2e0dd524bc2de9c7284b Minor loader improvements.
  crrev.com/2dc658ad29ce38a0364540d85d05dc3a2ab87f16 [recipe_engine/step] Add ability for recipe engine invocation to inject paths.
  crrev.com/2dba35d569811216433ebd96760f3e9bde78b7e9 [PROPERTIES] Allow declaration of $package_name/module properties.

R=borenet@google.com

Bug: skia:
Change-Id: Ieea18b43b328ce4bccf63ac97c70c5b25f0111d8
Reviewed-on: https://skia-review.googlesource.com/15656
Commit-Queue: Eric Boren <borenet@google.com>
Reviewed-by: Eric Boren <borenet@google.com>
Reviewed-by: Robbie Iannucci <iannucci@google.com>

Re-enable deferred proxies in Chrome

Now that https://skia-review.googlesource.com/c/14605/ (Allow TextureSamplers to have null GrTexture pointer) has landed we should be able to re-enable deferred proxies in Chrome.

Bug: 715488
Change-Id: Ibd38e5ddf0558b94f09e29f17837e6abfd22fa34
Reviewed-on: https://skia-review.googlesource.com/15252
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

polish up gm/hsl.cpp

   - remove broken clip_color_KHR().
   - rearrange a little to match spec closer
   - remove some TODOs

Change-Id: I2de6aa3138455d5970e2cda74f5da6ffadc3db56
Reviewed-on: https://skia-review.googlesource.com/15681
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

bug fix in hsl GM reference impls

As written the "constants" in clip() and channel() change as we update
each channel... duh.  Thankfully this ought to make most GMs look good.

Change-Id: I904e20e8e5114c827233dd1a93c0c59f7e7790fa
Reviewed-on: https://skia-review.googlesource.com/15680
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

hsl GM: more realistic references for non-legacy backends

Change-Id: I94663a2a3b9080ac318b2805ee8baa09d2985c9a
Reviewed-on: https://skia-review.googlesource.com/15545
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

update hsl GM

  - Compare against reference versions of the blend modes.
  - Point out differences between KHR and web versions.

It looks like the KHR ClipColor() is indeed wrong.

Change-Id: I650821b2d4815c493dddeb836c69875650926d7b
Reviewed-on: https://skia-review.googlesource.com/15654
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

remove old 565 destination opts

This is not an important format, and the code is dead or close to it.
The code is an occasional maintenance burden so I'd like it gone.

Change-Id: I4ad921533abf3211e6a81e6e475b848795eea060
Reviewed-on: https://skia-review.googlesource.com/15600
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

impl 4f version for tricolor shader

Bug: skia:
Change-Id: I43f456aa115ce71a5d1acd9bc94984248a88319a
Reviewed-on: https://skia-review.googlesource.com/15540
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>

Fix color blending for analytic shadows

Change-Id: I74e8c385a87a3f397b2df53679fcfd9120d82954
Reviewed-on: https://skia-review.googlesource.com/15645
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>

Allow custom GrAtlasGlyphCache texture sizes

A single glyph cache size doesn't make sense across the hardware Skia
runs on. This change allows a custom size to be specified (via a byte
limit), allowing cache size to be customized at context creation time.

Bug: 717178
Change-Id: I4f7baddd1897b2eac4f6d6e4fff1f805e1cdd250
Reviewed-on: https://skia-review.googlesource.com/15135
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>

More analytic shadow refactoring.

This is mostly to simplify and reduce redundant calculations.

Change-Id: I895bd43fbc9defd812969870c2145f9f19d69a6b
Reviewed-on: https://skia-review.googlesource.com/15462
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Stub encoding impls for clients without libraries

Bug: skia:
Change-Id: I447b071e50182a569af828dc7f62bf78f47f834d
Reviewed-on: https://skia-review.googlesource.com/15644
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Fix implicit ptr to bool

Bug: skia:
Change-Id: I11ce7858306dfe05162aa84a5f862b1b01c2cee9
Reviewed-on: https://skia-review.googlesource.com/15637
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

composecolorfilter can now append stages

Bug: skia:
Change-Id: I46b54e5fd09de16b467142a5501b226273182d52
Reviewed-on: https://skia-review.googlesource.com/15615
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>

Add bigrect GM.

Previous version in https://codereview.chromium.org/1758113005
was reverted.

Bug: skia:4632
Change-Id: I9a85f11ca3a6fcd5e67016edf67f0538f00b6f01
Reviewed-on: https://skia-review.googlesource.com/14751
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

add a new hsl GM

Change-Id: I2c8b9df3e6bf40de4b38966fa8ea671fc68aa45f
Reviewed-on: https://skia-review.googlesource.com/15635
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

SkTypeface: remove old virtual onGetAdvancedMetrics, since https://crrev.com/2860293003

Change-Id: Iad30b03fc8b34bdea2fe902f83a0dd979c16f1a7
Reviewed-on: https://skia-review.googlesource.com/15633
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

tweak the gaussian quartic

Bug: skia:
Change-Id: I0e50568ddea344443ae80c24aad3eb43736f77bc
Reviewed-on: https://skia-review.googlesource.com/15636
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>

SkEncoder: Rename files, change webp API, for consistency

Bug: skia:
Change-Id: I3dd6feb3d5661dcad3d2388b4d01fa9d3bbb15bb
Reviewed-on: https://skia-review.googlesource.com/15631
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Fix clang compiler error in SkSLFileOutputStream

Bug: skia:
Change-Id: Ic3b62ea82c12c9cc5ee8b56b05573e1d06342c0b
Reviewed-on: https://skia-review.googlesource.com/15626
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Presubmit: ensure empty line between CL title and description

Bug: skia:
Change-Id: If1b2d655e9be4b960f76d2d17ce1cf1126f7b2d4
Reviewed-on: https://skia-review.googlesource.com/15624
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Eric Boren <borenet@google.com>

Revert "GL: track enabled vertex arrays as a count rather than a mask"

This reverts commit 288d041c64322fafc77cfaf23907180ebad933a1.

Reason for revert: GL errors.

Original change's description:
> GL: track enabled vertex arrays as a count rather than a mask
>
> Bug: skia:
> Change-Id: I46ba29cb32960a415ee1993a7b957ec49c0c56d3
> Reviewed-on: https://skia-review.googlesource.com/15520
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>

TBR=egdaniel@google.com,bsalomon@google.com,csmartdalton@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I63b0fd3c6f36326a13bf853d13ce8d28869a32cf
Reviewed-on: https://skia-review.googlesource.com/15625
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>

speed-up 4f version of modecolorfilter and modulate

~20% faster

1. Don't call color4f::premul on every invocation
2. Inline modulate as its own special case (used by shadows)

Bug: skia:
Change-Id: I49ca565f589b9a7caac88e95468da7f8c395804a
Reviewed-on: https://skia-review.googlesource.com/15613
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>

Add support for row-by-row png encodes

Also adds a SkEncoder base class.

Bug: 713862
Change-Id: Ia3f009cd9f376514f6c19396245fab3a43ae6536
Reviewed-on: https://skia-review.googlesource.com/15152
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Remove unnecessary non-UTF8 character

Google3 roll is complaining about this character, so remove it. (It
does not look to be intentional.) The roll is affected because this
file is changed (though this line is not).

TBR=bsalomon@google.com, benjaminwagner@google.com

Change-Id: I639a04c424ea11e1a5e7df10c9d6c7233031bf2b
Reviewed-on: https://skia-review.googlesource.com/15612
Reviewed-by: Leon Scroggins <scroggo@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>

Add postFlush call to GrOnFlushCallbackObject

Adds a new postFlush method and renames the class to
GrOnFlushCallbackObject. Also removes the ref counting in favor of
making the callback object a purely virtual interface. ref/unref on the
callback interface would conflict with existing ref/unref methods on the
subclass. It is now the caller’s responsibility to ensure the lifetime
of the callback is tied to that of the context.

Bug: skia:
Change-Id: I2fc1f98c700032e296a36f3a9a09c0753ab47aea
Reviewed-on: https://skia-review.googlesource.com/15463
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Avoid interpolating color lut with less than 2 points

Bug: 717935
Change-Id: Ibf15b815891eef5a0239bc408bcbfe7c8b1507c5
Reviewed-on: https://skia-review.googlesource.com/15301
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>

SkDiscardableMemoryPool: always have a mutex for testing

Change-Id: I898058f1125f62832f2b2b661c661efb0f84da6d
Reviewed-on: https://skia-review.googlesource.com/14274
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Herb Derby <herb@google.com>

Split tracking of TexelBuffers from normal samplers

This is precursor CL to add support for texel buffers in Vulkan.

This change as includes fixes to the ordering of assigning
locations and texture units so that they match in
GrGLProgramDataManager and GrGLProgram.

Bug: skia:
Change-Id: I30c9578fb7dcb187256f744e07651e8564f93a6b
Reviewed-on: https://skia-review.googlesource.com/15225
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

GL: track enabled vertex arrays as a count rather than a mask

Bug: skia:
Change-Id: I46ba29cb32960a415ee1993a7b957ec49c0c56d3
Reviewed-on: https://skia-review.googlesource.com/15520
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

sksl optimizer fixes

The main issue was a use-after-free due to removing (and thus destroying)
the binary expression prior to re-adding part of it. Also cleaned up the
way dead assignments are handled and added the test that originally
identified these problems.

Bug: skia:
Change-Id: Icda93d69a66c4e57850ecdc88fc4a4f634e1aac2
Reviewed-on: https://skia-review.googlesource.com/15383
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>

optimize gaussian with herbs quartic

Bug: skia:
Change-Id: Ia06ad1023129dfcade3f648087b2070eab29b4b2
Reviewed-on: https://skia-review.googlesource.com/15523
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Reed <reed@google.com>

header cleanup

Change-Id: I9d93add4e7998d8a1aa37cba9e7829fcbda867d5
Reviewed-on: https://skia-review.googlesource.com/15318
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

Revert "Revert "SkTypeface::getAdvancedMetrics(): cleanup""

This reverts commit 59ad782b2b05b07aa6eb961aa4d62e934093cbd1.

   - SkAdvancedTypefaceMetrics is a struct not a class
   - SkTypeface::PerGlyphInfo is gone
   - s/getAdvancedTypefaceMetrics/getAdvancedMetrics/g
   - s/onGetAdvancedTypefaceMetrics/onGetAdvancedMetrics/g
   - [on]getAdvancedMetrics now return unique_ptr rather than bare ptr.
   - [on]getAdvancedMetrics no longer has parameters. (Only caller always
     used same arguments.)
   - SkAdvancedTypefaceMetrics uses C++11 in-class member initializers.
   - SkAdvancedTypefaceMetrics no longer inherits from SkRefCnt

Change-Id: I91b56e60f7d9de7d46c426c6bd34ce124e0cf00e
Reviewed-on: https://skia-review.googlesource.com/15360
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

add -Wno-error=attributes for G3 Android builds

This warning is triggering spuriously on 32-bit Android opt builds.

This CL turns it off for {32,64}x{fastbuild,opt}, but I don't mind.
We can rely on our local builds' much more up to date toolchains.

Change-Id: I5210e87f540a1a425fc0abd07bb3c3986f0f78b0
Reviewed-on: https://skia-review.googlesource.com/15541
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Local matrix support for RP gradients

Change-Id: Id2cef26e0f91f14a0f1370fabe5ca15af9a4c07c
Reviewed-on: https://skia-review.googlesource.com/15317
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Retry "Factor out common code from sweep and linear."

We weren't handing failure correctly, sending radials and two point
conics and any other unsupported cases back too far out, instead of
the default SkShader::onAppendStages().

This reverts commit 02ab8cc2909d99377e4200cd426a02c0b482a040.

Change-Id: I80cb028c37071e0146ee4b48b210fc8bc87b290f
Reviewed-on: https://skia-review.googlesource.com/15481
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Factor out common code from sweep and linear."

This reverts commit b7a330ff0048875e178b2d9a68cd5605d83baf8d.

Reason for revert: Oof, I hate to do this, but we didn't expect this to change any drawing did we?  This draws a bunch of GMs differently in both 565 and sRGB mode, some very noticeably.

Original change's description:
> Factor out common code from sweep and linear.
>
> And, some file cleanup.
>
> Change-Id: I804db924bce3b5834f6cda481315dd2da38df5ca
> Reviewed-on: https://skia-review.googlesource.com/15226
> Commit-Queue: Herb Derby <herb@google.com>
> Reviewed-by: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
>

TBR=mtklein@chromium.org,mtklein@google.com,herb@google.com,fmalita@chromium.org,fmalita@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I75bb6c3e4045a0f70c7e9552b84b22f0fed73b80
Reviewed-on: https://skia-review.googlesource.com/15464
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

CRC32 no longer restricted to ARM64

On a simple benchmark the CRC32 version is about 3x faster on
ARM Cortex A57 (Aarch32) than the Murmur3 scalar version.

BUG=skia:

Change-Id: I71515e8463a33924998b837ff9f32202690dd2fe
Reviewed-on: https://skia-review.googlesource.com/15480
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Fix comma in vulkan

Fixes bug where comma was suppose to be a semi colon.

Bug: skia:
Change-Id: I4bbd890fdef46775a1dd687cca06293114ad8a1c
Reviewed-on: https://skia-review.googlesource.com/15460
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Add support for Geom shaders in Vulkan backend.

Still requires SkSL support before it will work.

The main changes here involve support for uniforms in the geometry shader.
We use the same buffer for vertex and geometry shader stages. These
uniforms are not expected to be updated as often as frag data so we keep them
separate to avoid larger buffer uploads to the gpu.

BUG=skia:

Change-Id: I10b631c24071b6ffa258907a02a009ec6c8accd0
Reviewed-on: https://skia-review.googlesource.com/8413
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>

Factor out common code from sweep and linear.

And, some file cleanup.

Change-Id: I804db924bce3b5834f6cda481315dd2da38df5ca
Reviewed-on: https://skia-review.googlesource.com/15226
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Florin Malita <fmalita@chromium.org>

Add ShadowUtils sample.

Also enables mouse support in Viewer.

Change-Id: Iaed08d42a64f591f0cd9b24684b3aee43404ed94
Reviewed-on: https://skia-review.googlesource.com/15313
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Marker for Windows NVIDIA driver update.

No-Try: true
Change-Id: Ib275307c858e0a1c9a37881700aeeb2a9e8b83ac
Reviewed-on: https://skia-review.googlesource.com/15381
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>