Revert "eliminated GrGLSLExpr"

This reverts commit 93f20f5629e52eed732d2b9d6dbbb351cc30b2cd.

Reason for revert: Mismerge readded deleted files.

Original change's description:
> eliminated GrGLSLExpr
>
> Now that skslc performs all of the optimizations (and then some) that
> GrGLSLExpr is responsible for, it's just extra work for no benefit.
>
> Bug: skia:
> Change-Id: I40b0629e00a33873ed9fc6c0a9f41d8350221f9a
> Reviewed-on: https://skia-review.googlesource.com/14560
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>

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

Change-Id: Ia8b723594527afe34489fc78a4b49039081b6390
Reviewed-on: https://skia-review.googlesource.com/15154
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>

Support numerical transfer functions in readPixels()

Let's do this because:
(1) We can.
(2) Android and Chrome have asked for it.
(3) It will simplify the implementation of SkImage::makeColorSpace().

Bug: skia:
Change-Id: Ia3c322b8a58c79ad67cdebe744e0623bd59dcffd
Reviewed-on: https://skia-review.googlesource.com/15148
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Matt Sarett <msarett@google.com>

Add support for row-by-row jpeg encoding

Reland of:
https://skia-review.googlesource.com/c/14641/

Bug: 713862
Change-Id: I9dca5ede4ebf569c5f80edcfb23a506b6cfa935e
Reviewed-on: https://skia-review.googlesource.com/15144
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Revert "Revert "Remove copyTo() legacy API flag from bzl""

This reverts commit d271ba268876d3feca13e02d0a92ecaf79b6ddfc.

Reason for revert: Client failures should be fixed.  Will keep an eye on the roller.

Original change's description:
> Revert "Remove copyTo() legacy API flag from bzl"
>
> This reverts commit 65083e444c03cedecf45bc38bb8fea27e91e0a9e.
>
> Reason for revert: Causes build failures.
>
> Original change's description:
> > Remove copyTo() legacy API flag from bzl
> >
> > NOTRY=true
> >
> > Bug: skia:6464
> > Change-Id: Ia3935998960054df9e54cfb2aead66549c629c04
> > Reviewed-on: https://skia-review.googlesource.com/15001
> > Reviewed-by: Matt Sarett <msarett@google.com>
> > Commit-Queue: Matt Sarett <msarett@google.com>
> >
>
> TBR=msarett@google.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Change-Id: I9d82c7b699e106238981e32553e975c91e42cfb8
> Reviewed-on: https://skia-review.googlesource.com/15021
> Reviewed-by: Ben Wagner <benjaminwagner@google.com>
> Commit-Queue: Ben Wagner <benjaminwagner@google.com>
>

TBR=benjaminwagner@google.com,msarett@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: Ie9c9c1f8d06c8b422b8902e39645cd0d887d6b49
Reviewed-on: https://skia-review.googlesource.com/15150
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Add wireframe mode to Viewer.

Change-Id: I0ff11088465a4702acf9841a791d76f286ddbaf1
Reviewed-on: https://skia-review.googlesource.com/15147
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>

eliminated GrGLSLExpr

Now that skslc performs all of the optimizations (and then some) that
GrGLSLExpr is responsible for, it's just extra work for no benefit.

Bug: skia:
Change-Id: I40b0629e00a33873ed9fc6c0a9f41d8350221f9a
Reviewed-on: https://skia-review.googlesource.com/14560
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

speedup exp() in gaussian colorfilter

~20% faster
differs in low 1 from prev impl

Bug: skia:
Change-Id: If7aa7f5523f15a5424ebf5a09313c5e1d43e1a1f
Reviewed-on: https://skia-review.googlesource.com/15105
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>

Manually roll skia recipe deps.

This includes:
 recipes-py:  https://chromium.googlesource.com/external/github.com/luci/recipes-py/+log/f5e47cfc28b5964701572db12f96c7a5a3f1821a..554e8878ac6395262dc3c90bbb8a100ee844f766
 depot_tools: https://chromium.googlesource.com/chromium/tools/depot_tools/+log/109907e6a95ae0301b1564c76c4a3e9d15e19366..1e71b8bdf0913b306b28141192272350cf68cdc2
 tools/build: https://chromium.googlesource.com/chromium/tools/build/+log/f7b41106d64b6517c67343f2cfb03f4a9d62afc2..dcca7cbe7ecda2c4d511f38c6f3eab29cad08cf7

The ranges include my generic postprocess/merge change in tools/build
that had been failing to roll over in https://skia-review.googlesource.com/c/13920/
along with a few corresponding fixes.

Change-Id: I7e328bdb3acfc9b4c219e408c226f181f3b05f3d
Reviewed-on: https://skia-review.googlesource.com/14678
Commit-Queue: John Budorick <jbudorick@chromium.org>
Commit-Queue: Eric Boren <borenet@google.com>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Eric Boren <borenet@google.com>

Setup support for UNIFORM_TEXEL_BUFFER descriptor sets in Vulkan

This is the first CL to get support for using texel buffers in vulkan.

Bug: skia:
Change-Id: Iaac5ba4a356b487bc2b63111cca34ed968881f6b
Reviewed-on: https://skia-review.googlesource.com/15100
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

sksl support for buffer blocks

Bug: skia:
Change-Id: Ic2cabaf2c7fb23cec7863f2b6152bbed133e0886
Reviewed-on: https://skia-review.googlesource.com/14947
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>

Remove EXPERIMENTAL_SHADOWING, Part 1

Bug: skia:6557
Change-Id: I6482d74be7b360c93141a73dd80c67854530c7a1
Reviewed-on: https://skia-review.googlesource.com/15101
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Add support for row-by-row jpeg encoding"

This reverts commit 9b848d5749c5e34b56f927a3a3374c8ebafbd9db.

Reason for revert: ASAN reports leaked memory [1]. Google3 reports a "delete size mismatch" [2], which I suspect is the same issue.

[1] https://chromium-swarm.appspot.com/task?id=35e2c9fa9eac6310&refresh=10&show_raw=1
[2] https://test.corp.google.com/ui#cl=154838904&flags=CAMQBQ==&id=OCL:154838904:BASE:154839043:1493741642370:9c96115f&t=//chrome/skia/dm_wrapper:dm_wrapper

Original change's description:
> Add support for row-by-row jpeg encoding
>
> Bug: 713862
> Change-Id: I787b7c49662a00b89ae0ef35845dfbd6be3e6fb1
> Reviewed-on: https://skia-review.googlesource.com/14641
> Commit-Queue: Matt Sarett <msarett@google.com>
> Reviewed-by: Leon Scroggins <scroggo@google.com>
>

TBR=msarett@google.com,scroggo@google.com,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: Ic5a8d67e0d4a7733662586055ceff086a2ab335d
Reviewed-on: https://skia-review.googlesource.com/15140
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>

Add support for row-by-row jpeg encoding

Bug: 713862
Change-Id: I787b7c49662a00b89ae0ef35845dfbd6be3e6fb1
Reviewed-on: https://skia-review.googlesource.com/14641
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>

Fix VkWrapTest

Fix current breakage on bots.

Bug: skia:
Change-Id: I4b60ef85d941a179d76cccc1a468aaa6437caddc
Reviewed-on: https://skia-review.googlesource.com/15104
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Blacklist memory intensive images on a bot

Test-Win8-MSVC-Golo-CPU-AVX-x86-Debug is running out of memory running
interlaced image decoding tests. Stop running them on this particular
bot (others are still running them). For simplicity, also blacklist
the RAW images that have had similar problems on other devices.

Change-Id: I1071ca2f59ef26b9a9798924f2eefa0483017439
Reviewed-on: https://skia-review.googlesource.com/15074
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>

Revert "Remove SkLights include from SkCanvas.h"

This reverts commit 9ff301bf918cd85d9f67047f10631b7a74345a90.

Reason for revert: need to update G3, Flutter.

Original change's description:
> Remove SkLights include from SkCanvas.h
>
> SkLights.h pulls in a bunch of other headers and is not needed (fwdecl
> works fine).
>
> Change-Id: Id2d7176eb3bf4609f72f46d513eebf59318f542f
> Reviewed-on: https://skia-review.googlesource.com/14904
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Florin Malita <fmalita@chromium.org>
>

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

Change-Id: I4799ad5b31aaeaf529c8b912bbe09aa8869a5e6c
Reviewed-on: https://skia-review.googlesource.com/15107
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Reland "Add a clip mask bench"

Change-Id: Iefbfd1ecec38d3db5bac4370c75e6701aa03076f
TBR=reed@google.com
Reviewed-on: https://skia-review.googlesource.com/15102
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Remove SkLights include from SkCanvas.h

SkLights.h pulls in a bunch of other headers and is not needed (fwdecl
works fine).

Change-Id: Id2d7176eb3bf4609f72f46d513eebf59318f542f
Reviewed-on: https://skia-review.googlesource.com/14904
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Fix up shadows in raster.

* Re-enable shadow blurs for raster circles and rrects
* Fix up the tessellation as much as possible to remove skinny triangles

Bug: skia:6425
Change-Id: I6548055084bc8596a052bcd3cec852766e084ba2
Reviewed-on: https://skia-review.googlesource.com/14943
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>

Make GrBackendTexture take Gr*Info refs in ctor, and copy them.

Bug: skia:
Change-Id: Ic05d3384fa07560fc18c52bb8ae03541a72515f7
Reviewed-on: https://skia-review.googlesource.com/14374
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Remove declaration of SkTextureImageSetTexture, which does not exist

Bug: skia:
Change-Id: Ifb7edcae491ef48fbc69324d9dfb97a298104db9
Reviewed-on: https://skia-review.googlesource.com/15073
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

Expose resetForNextPicture in SkNoDrawCanvas

SkRecorder has some private Skia behavior with the way it handles an
SkNoDrawCanvas that is not exposed.  In particular, it's able to create
an arbitrary SkIRect device clip at construction time.

This is a performance optimization when reusing SkNoDrawCanvas across
multiple recordings instead of reconstructing a new canvas every time.

https://codereview.chromium.org/2768143002 is trying to implement
another version of SkPictureBuilder/SkRecorder and so wants to be able
to use similar SkNoDrawCanvas behavior.

Change-Id: I0ac5f0eef8f0548cefd7354062311667cf448b07
Reviewed-on: https://skia-review.googlesource.com/11320
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

colorspaceinfo tool: Add flag to indicate that the input is an icc profile

Bug: skia:
Change-Id: Ieb5cff92bae1aa64df5bc7f3a398514fe5b20f77
Reviewed-on: https://skia-review.googlesource.com/14956
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Remove legacy pixelref flag from bzl

NOTRY=true

Bug: skia:6535
Change-Id: I29d786cef240aaeaca829b4a9a25ff1a8f078588
Reviewed-on: https://skia-review.googlesource.com/15000
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Revert "Remove copyTo() legacy API flag from bzl"

This reverts commit 65083e444c03cedecf45bc38bb8fea27e91e0a9e.

Reason for revert: Causes build failures.

Original change's description:
> Remove copyTo() legacy API flag from bzl
>
> NOTRY=true
>
> Bug: skia:6464
> Change-Id: Ia3935998960054df9e54cfb2aead66549c629c04
> Reviewed-on: https://skia-review.googlesource.com/15001
> Reviewed-by: Matt Sarett <msarett@google.com>
> Commit-Queue: Matt Sarett <msarett@google.com>
>

TBR=msarett@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I9d82c7b699e106238981e32553e975c91e42cfb8
Reviewed-on: https://skia-review.googlesource.com/15021
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Added A2B support to colorspaceinfo tool

Tool now parses A2B images/profiles.
Tool can now display a cross-section based visualization for the color
look-up tables in A2B profiles.

BUG=skia:

Change-Id: I68abb3e947b080c533e283783d7859feea8d35d6
Reviewed-on: https://skia-review.googlesource.com/6119
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>

Remove copyTo() legacy API flag from bzl

NOTRY=true

Bug: skia:6464
Change-Id: Ia3935998960054df9e54cfb2aead66549c629c04
Reviewed-on: https://skia-review.googlesource.com/15001
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Revert "Revert "SaveLayerRec::fClipMask -> raw pointer""

This reverts commit 3354969a4a4adcea159f3a0b01ad0a7c04fc9115.

Reason for revert: not the problem?

Original change's description:
> Revert "SaveLayerRec::fClipMask -> raw pointer"
>
> This reverts commit a6b72cb5729306fdd676d739c1e2c53afa0786a7.
>
> Reason for revert: red bots
>
> Original change's description:
> > SaveLayerRec::fClipMask -> raw pointer
> >
> > Use raw pointers for optional clip mask plumbing, to match the backdrop
> > API.
> >
> > Change-Id: I7eb0ee5896faf34cc05789ba0703f35a4ab6a4f2
> > Reviewed-on: https://skia-review.googlesource.com/14901
> > Reviewed-by: Mike Reed <reed@google.com>
> > Commit-Queue: Florin Malita <fmalita@chromium.org>
> >
>
> TBR=fmalita@chromium.org,reed@google.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Change-Id: I4387620cdc3410018af9cef221e5cf8d09015380
> Reviewed-on: https://skia-review.googlesource.com/14955
> Reviewed-by: Mike Klein <mtklein@chromium.org>
> Commit-Queue: Mike Klein <mtklein@chromium.org>
>

TBR=mtklein@chromium.org,reviews@skia.org,fmalita@chromium.org,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I6dde9e0441928f7a7423139a5c39f520f892f4b5
Reviewed-on: https://skia-review.googlesource.com/14958
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Add a clip mask bench"

This reverts commit c59a38d12dce287427f3d3fe1d4b3ad8052cda35.

Reason for revert: whoops, sorry, this one must have been the problem.

Original change's description:
> Add a clip mask bench
>
> Change-Id: I230729492fc23e290136f7d62610abe5ca51c067
> Reviewed-on: https://skia-review.googlesource.com/14941
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Florin Malita <fmalita@chromium.org>
>

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

Change-Id: Ifbd74a8985ebaa95e54032aaaa2891b0f1b67940
Reviewed-on: https://skia-review.googlesource.com/14957
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "SaveLayerRec::fClipMask -> raw pointer"

This reverts commit a6b72cb5729306fdd676d739c1e2c53afa0786a7.

Reason for revert: red bots

Original change's description:
> SaveLayerRec::fClipMask -> raw pointer
>
> Use raw pointers for optional clip mask plumbing, to match the backdrop
> API.
>
> Change-Id: I7eb0ee5896faf34cc05789ba0703f35a4ab6a4f2
> Reviewed-on: https://skia-review.googlesource.com/14901
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Florin Malita <fmalita@chromium.org>
>

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

Change-Id: I4387620cdc3410018af9cef221e5cf8d09015380
Reviewed-on: https://skia-review.googlesource.com/14955
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Add a clip mask bench

Change-Id: I230729492fc23e290136f7d62610abe5ca51c067
Reviewed-on: https://skia-review.googlesource.com/14941
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

SaveLayerRec::fClipMask -> raw pointer

Use raw pointers for optional clip mask plumbing, to match the backdrop
API.

Change-Id: I7eb0ee5896faf34cc05789ba0703f35a4ab6a4f2
Reviewed-on: https://skia-review.googlesource.com/14901
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

finish up constants

For whatever reason, if I swap the condition in the if_then_else tests
from < to >= and swap the then/else values, I can use constants in
hsl_to_rgb.  Still don't understand why, but I'll take it.  I suspect it
has something to do with SSE, IEEE, and NaN, but I don't care enough to
speculate any more concretely.

This does that, removes C() and _f, updates some comments, and adds a
guard in build_stages.py to yell if it sees trouble like LCPI40_4...

This reminds me to try -ffast-math soon.  I think that was mostly held
back by constants.

Change-Id: I3f8a37a4d4642f77422ce3261b750061e9e604a3
Reviewed-on: https://skia-review.googlesource.com/14942
Reviewed-by: Herb Derby <herb@google.com>

refactor hsl_to_rgb a touch

This rewrites the existing logic to expose more of the symmetries,
and especially to make them clearly identical subexpressions.

I think it's clear that the intent in hue_to_rgb is to wrap the t value
back into 0-1... that's t = fract(t).

No GM diffs.

Change-Id: I9d62d8f80bcb45711ee334f953d3f6410e068ce4
Reviewed-on: https://skia-review.googlesource.com/14940
Reviewed-by: Herb Derby <herb@google.com>

Update Nexus6p to Android O developer release.

No-Try: true
Change-Id: I1b39d3548d0979385b24558170fdcd6d766a63bd
Reviewed-on: https://skia-review.googlesource.com/14900
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

fix t / t2 confusion in hsl_to_rgb

I think the original[1] SkRasterPipeline_opts.h version had a typo that
I faithfully copied over to hsl_to_rgb.  In Hue2RGB[2], the scalar
equivalent of hue_to_rgb, we mutate t to keep it in 0-1 range[3].  In
the SkRasterPipeline_opts.h code we introduced a new value t2 instead,
and then used it everywhere, but accidentally typed 't' in the t < 1/6 case.

The expression "p + (q - p)*6.0f*t" should have been "p + (q - p)*6.0f*t2".

This fixes things by changing t in place, much like Hue2RGB does.

The GM doesn't change anywhere, which is troubling.

[1] https://skia-review.googlesource.com/c/7460/21/src/opts/SkRasterPipeline_opts.h#808
[2] https://skia-review.googlesource.com/c/7460/21/src/effects/SkHighContrastFilter.cpp#26
[3] I think this whole clamp should probably become "t = fract(t)".
    Will follow up.

Change-Id: I3dcc1a79ffae46830178d931844ee3113f8bdfd1
Reviewed-on: https://skia-review.googlesource.com/14910
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

getting close on float constants

I think I'm now down to just the constants used in comparisons in
hsl_to_rgb... so weird.  I don't get what's special about this code.

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Win7-MSVC-Golo-CPU-AVX-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE41,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE2

Change-Id: I398139f00371bdbc45281a49bca56d02ba59cba9
Reviewed-on: https://skia-review.googlesource.com/14909
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

SkFontMgr::matchFamily should not crash on nullptr.

While all systems can resolve a font from just a style request
(without a name) almost no systems specify a default font family.

BUG=skia:6574

Change-Id: If7c81808b62cd5d8212bce2eb4d9c476c45af80a
Reviewed-on: https://skia-review.googlesource.com/14902
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Mike Reed <reed@google.com>

some float constants

Trying to go slowly to find where problems arise.
Weirdly, I think I got everything except hsl_to_rgb.

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Win7-MSVC-Golo-CPU-AVX-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE41,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE2

Change-Id: I4d85a4c1f40bd87e7cb18fc9b5ce020812dc31db
Reviewed-on: https://skia-review.googlesource.com/14905
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Blacklist SkImage_makeTextureImage on NVIDIA Win10 Vulkan

Bug: skia:6554
Change-Id: I7c1086ce6afff84c1b1463ec58c2c6a9949aaa2e
Reviewed-on: https://skia-review.googlesource.com/14907
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>

Fix Creation of GrBackendRenderTarget from GrBackendRenderTargetDesc in GL

Bug: skia:
Change-Id: I970e7bc36ae799996f9b28e70f97cb7ce4382eb2
Reviewed-on: https://skia-review.googlesource.com/14906
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Only store width and height on SkPixelRef (last part)

Relanding https://skia-review.googlesource.com/c/14105/
in pieces to try to diagnose problems with the Chrome
roll.

Bug: skia:6535
Change-Id: Iefe4825b9ce9be999baeec8ab7ae6651f1caf451
Reviewed-on: https://skia-review.googlesource.com/14860
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Blacklist testimgari.jpg on CG

Bug: skia:
Change-Id: I2c471d18f29bb6d0e8ac4a4d0afa1815788bb121
Reviewed-on: https://skia-review.googlesource.com/14825
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Restore legacy SaveLayerRec ctor

Let's not break existing clients.

Change-Id: I056fefefb49d46fb284597c5bcf0319470537524
Reviewed-on: https://skia-review.googlesource.com/14829
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Fix merge conflict.

I accidentally caused a merged conflict between
https://skia-review.googlesource.com/14656 and
https://skia-review.googlesource.com/14821

No-Try: true
Change-Id: Ib00b293019fddc3951a5d413744b9d186dd103d8
Reviewed-on: https://skia-review.googlesource.com/14840
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Remove unnecessary SkFILEStream::getMemoryBase

It overrides the SkStream:: version, but does the same thing.

Change-Id: I857f2119a85bb3f942707ab3cef58b45a2cf4b7a
Reviewed-on: https://skia-review.googlesource.com/14828
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>

Only store width and height on SkPixelRef (part 2)

Relanding https://skia-review.googlesource.com/c/14105/
in pieces to try to diagnose problems with the Chrome
roll.

Bug: skia:6535
Change-Id: Iada034fc41ef315f7f00984d8de9d9cc2f361ad2
Reviewed-on: https://skia-review.googlesource.com/14657
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Reed <reed@google.com>

Omit shapes_mixed_10000_32x33 on GTX960 ANGLE Release

Bug: skia:6534
Change-Id: I5634b35c3a231da8b5714209889921933ca38962
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/14656
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Omit HalfFloatAlphaTextureTest on CommandBuffer bot.

Bug: chromium:697030
No-Try: true
Change-Id: Ib56c88c384e6b363aefe91c7027fa587d0b000b5
Reviewed-on: https://skia-review.googlesource.com/14648
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Add Swarming bot id and task id to recipe expectations.

No-Try: true
Change-Id: Idd748bf61ab3d6611b85472de5d3f3a25d4ae9ee
Reviewed-on: https://skia-review.googlesource.com/14821
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Revert "Revert "Plumb the use of GrBackendRenderTarget throughout Skia""

This reverts commit e3bd422fafc74dd3410c3de24a576635be92c3b4.

Reason for revert: Pre-req changes have all landed in other projects at this point.

Original change's description:
> Revert "Plumb the use of GrBackendRenderTarget throughout Skia"
>
> This reverts commit fdd77daedbba3b7c53be74a82fb9fae891b51696.
>
> Reason for revert: Apparently I have a few more build files to update before this can land.
>
> Original change's description:
> > Plumb the use of GrBackendRenderTarget throughout Skia
> >
> > Bug: skia:
> > Change-Id: Ib99a58d9552f5c7b8d77c09dcc72fa88326c26aa
> > Reviewed-on: https://skia-review.googlesource.com/14148
> > Reviewed-by: Brian Salomon <bsalomon@google.com>
> > Reviewed-by: Robert Phillips <robertphillips@google.com>
> > Commit-Queue: Greg Daniel <egdaniel@google.com>
> >
>
> TBR=egdaniel@google.com,bsalomon@google.com,robertphillips@google.com,reviews@skia.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Change-Id: I984e1909870182474c4c3cce257f01b6a9d8581f
> Reviewed-on: https://skia-review.googlesource.com/14531
> Reviewed-by: Greg Daniel <egdaniel@google.com>
> Commit-Queue: Greg Daniel <egdaniel@google.com>
>

TBR=egdaniel@google.com,bsalomon@google.com,robertphillips@google.com,reviews@skia.org
# Not skipping CQ checks because original CL landed > 1 day ago.

Change-Id: Ib7ab94aada8a7cb80fe38f24daf32f9208c5b169
Reviewed-on: https://skia-review.googlesource.com/14826
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Initial clip-mask-layer support

SkBitmapDevice-only implementation.

Will add A8 fast path specializations in a follow-up.

Change-Id: I2ccb1ffba3689e92ac90a23e94737471dfb121a1
BUG=skia:6005

Change-Id: I2ccb1ffba3689e92ac90a23e94737471dfb121a1
Reviewed-on: https://skia-review.googlesource.com/14183
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Optimizations for analytic shadows.

Lots of changes here:
* Batch circle shadows with rrect shadows
* Avoid checking matrix and path conditions twice
* Remove lots of checks for 1/2 pixel radii
  (needed before to force the rrect through the
   regular GPU path)
* Fix scaling effect on ambient blur width
* Remove unused flags

Bug: skia:6119
Change-Id: If0eb78ec4d19d9f978b19bdbc3a7e558a4db2ed9
Reviewed-on: https://skia-review.googlesource.com/14654
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=update-skps@skia.org
NO_MERGE_BUILDS

Change-Id: Ic3ced4cb5f518abc5d1d0611f68fce4b976f5082
Reviewed-on: https://skia-review.googlesource.com/14679
Reviewed-by: update-skps <update-skps@skia.org>
Commit-Queue: update-skps <update-skps@skia.org>

Remove temporary_internal_describeTopLayer

Change-Id: I035b7db84b524be13fb675b75bc402bc9b733cb5
Reviewed-on: https://skia-review.googlesource.com/13803
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

add drawString helper to canvas

Many tests and examples use drawText with
a guess of how long the text is in bytes,
or a call to strlen(). Add a helper to
SkCanvas to simplify these examples.

Add another helper for SkString.

R=reed@google.com

Change-Id: I0204a31e938f065606f08ee7cd9a6b36db791ee2
Reviewed-on: https://skia-review.googlesource.com/13642
Commit-Queue: Cary Clark <caryclark@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Cary Clark <caryclark@skia.org>

Fix bitmap copy gms

Mark pixels as opaque so copying to 565 continues to work.

Bug: skia:
Change-Id: Id2b0a8c462fe35f7d1cc2ccc32ca10eda18981c8
Reviewed-on: https://skia-review.googlesource.com/14653
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Only store width and height on SkPixelRef (part 1)

Relanding https://skia-review.googlesource.com/c/14105/
in pieces to try to diagnose problems with the Chrome
roll.

Bug: skia:6535
Change-Id: Ic321c437ecd3cb7940a48fd73fc22b192804c67a
Reviewed-on: https://skia-review.googlesource.com/14650
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Reed <reed@google.com>

Combine GPU flush with semaphore API

Bug: skia:
Change-Id: I5f388760d43b19755b2767d95d32065f6d0f926e
Reviewed-on: https://skia-review.googlesource.com/14646
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

remove unused yuv methods

Bug: skia:
Change-Id: Ie1e6c905d8989b49c2cf5c9e5fd1f383268f6480
Reviewed-on: https://skia-review.googlesource.com/14645
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>

change drawAtlas to behave like drawVertices

Bug: skia:6571
Change-Id: If60c30755ec0548df0e9bdeca2f3af0bc87764ca
Reviewed-on: https://skia-review.googlesource.com/14642
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>

Revert "Only store width and height on SkPixelRef"

This reverts commit 2cbb6662e329981840f90ef4edd62f70f69e6030.

Reason for revert: Likely cause of Chromium DEPS roll failure; speculative revert.

Original change's description:
> Only store width and height on SkPixelRef
>
> Bug: skia:6535
> Change-Id: Id91e8d1e82f593be7d4b23ca5abde752f2666a77
> Reviewed-on: https://skia-review.googlesource.com/14105
> Commit-Queue: Matt Sarett <msarett@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
>

TBR=djsollen@google.com,msarett@google.com,reed@google.com,stani@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I12a024a71833f33432d5ea8cffdfc642b8b4240a
Bug: skia:
Reviewed-on: https://skia-review.googlesource.com/14644
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Revert "Revert "Delete copyTo(Allocator), hide copyTo() behind flag""

This reverts commit 0122af08f6af0dee490e1a4f35b552377d0d4753.

Reason for revert: Fixed Android callsite

Original change's description:
> Revert "Delete copyTo(Allocator), hide copyTo() behind flag"
>
> This reverts commit d4a338f4d0a0cdc08d7d3668931c60997f0fa971.
>
> Reason for revert: Looks like I missed something I was supposed to delete in Android.
>
> Original change's description:
> > Delete copyTo(Allocator), hide copyTo() behind flag
> >
> > Replace uses of copyTo() in Skia.
> >
> > Bug: skia:6464
> > Change-Id: I921dc53a1c29a5176d18f05741f7c0b5a008e548
> > Reviewed-on: https://skia-review.googlesource.com/14502
> > Commit-Queue: Matt Sarett <msarett@google.com>
> > Reviewed-by: Mike Reed <reed@google.com>
> >
>
> TBR=msarett@google.com,reed@google.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Change-Id: I4d252940cc6a2462b030007055ea6c229471fc6e
> Reviewed-on: https://skia-review.googlesource.com/14602
> Reviewed-by: Matt Sarett <msarett@google.com>
> Commit-Queue: Matt Sarett <msarett@google.com>
>

TBR=msarett@google.com,reviews@skia.org,reed@google.com

Change-Id: I81659a820f79f1958fda23cb62513065b57db99d
Reviewed-on: https://skia-review.googlesource.com/14640
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>

add helper makeShader for the common Clamp case

Bug: skia:
Change-Id: I22cac56f35fed16daede0cf1b3e34dfa6f5ae3ea
Reviewed-on: https://skia-review.googlesource.com/14625
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>

Fix fast blur for shadow rrects.

* Handles case where blur is greater than
  corner radius.
* Speeds up shader by avoiding a divide.
* Tweaks shadow sample to be more MD-ish.

Bug: skia:
Change-Id: I2990357b39784c22f02a8e51872391fd68e8226a
Reviewed-on: https://skia-review.googlesource.com/14141
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Scrub DeviceCM

* remove unused fMatrix, fMatrixStorage
* fPaint  -> std::unique_ptr<>
* fDevice -> sk_sp<>

Change-Id: I790b54bfc4d8154216e620a8c1642f3955fe7be0
Reviewed-on: https://skia-review.googlesource.com/14624
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Testing permissions

Bug: skia:
Change-Id: I204a0956228de5afde9c5df02dc5afa12ab21bfd
Reviewed-on: https://skia-review.googlesource.com/14609
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Ravi Mistry <rmistry@google.com>

Add non-Valgrind PreAbandonGpuContext bot.

This may help debug timeouts for
Test-Ubuntu-GCC-ShuttleA-GPU-GTX550Ti-x86_64-Release-Valgrind_PreAbandonGpuContext

Also stop uploading *AbandonGpuContext results to Gold.

No-Try: true
Change-Id: Ifa10d47cf27f86424d3936f451d190ee34f5637d
Reviewed-on: https://skia-review.googlesource.com/14525
Commit-Queue: Ben Wagner <benjaminwagner@google.com>
Reviewed-by: Eric Boren <borenet@google.com>

Remove srgb flags from GrProcessorSet

Bug: skia:
Change-Id: Icbb1b2f39cac70c9d74603514786d76b46d0afd9
Reviewed-on: https://skia-review.googlesource.com/14603
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>

additional skslc vector optimizations

Bug: skia:
Change-Id: I845d0952c281835a630882ae4026277c93ccf542
Reviewed-on: https://skia-review.googlesource.com/14406
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>

add GM to compare atlas and vertices

Bug: skia:6571
Change-Id: Ic97101c4623c7557eb33f69796232ed51a89af7e
Reviewed-on: https://skia-review.googlesource.com/14621
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>

handle GrVkTransferBuffer::Create failure

Change-Id: I713dd7b678bfec6808bd6f57bfdf5748258e8b24
Reviewed-on: https://skia-review.googlesource.com/14420
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

[infra/bots] Treat extra_config as a list.

Dependency of https://skia-review.googlesource.com/c/14525/

No-Try: true
Change-Id: I7686576aa48865116fe8a593f08d395f901a1d04
Reviewed-on: https://skia-review.googlesource.com/14524
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Ben Wagner <benjaminwagner@google.com>

Disable deferred proxies in Chrome

We will re-enable once the proxy instantiation is moved past the TextureSamplers

Bug: 715488
Change-Id: I4f0dee18fc191d7fffb6a2f4fedd825729ebb057
Reviewed-on: https://skia-review.googlesource.com/14520
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

Only store width and height on SkPixelRef

Bug: skia:6535
Change-Id: Id91e8d1e82f593be7d4b23ca5abde752f2666a77
Reviewed-on: https://skia-review.googlesource.com/14105
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Reed <reed@google.com>

Revert "Delete copyTo(Allocator), hide copyTo() behind flag"

This reverts commit d4a338f4d0a0cdc08d7d3668931c60997f0fa971.

Reason for revert: Looks like I missed something I was supposed to delete in Android.

Original change's description:
> Delete copyTo(Allocator), hide copyTo() behind flag
>
> Replace uses of copyTo() in Skia.
>
> Bug: skia:6464
> Change-Id: I921dc53a1c29a5176d18f05741f7c0b5a008e548
> Reviewed-on: https://skia-review.googlesource.com/14502
> Commit-Queue: Matt Sarett <msarett@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
>

TBR=msarett@google.com,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I4d252940cc6a2462b030007055ea6c229471fc6e
Reviewed-on: https://skia-review.googlesource.com/14602
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Add instrumentation into SkArenaAlloc.

Add a parameter to the constructor that will have the dtor print out
stats for setting the initial parameters.

Clean up: Move some function so they are in the same order as .h

Change-Id: I19d87dcc9c3b8dcc3e1d4f2ff078b78bbc490d92
Reviewed-on: https://skia-review.googlesource.com/14600
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Herb Derby <herb@google.com>

jumper, remove C(int)

This finishes off integer constants... they should all be normal now.

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Win7-MSVC-Golo-CPU-AVX-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE41,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE2

Change-Id: I66ecc6533807fc59bb5ac9d3c5f7ab9e6e1f0d7f
Reviewed-on: https://skia-review.googlesource.com/14528
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Plumb the use of GrBackendRenderTarget throughout Skia"

This reverts commit fdd77daedbba3b7c53be74a82fb9fae891b51696.

Reason for revert: Apparently I have a few more build files to update before this can land.

Original change's description:
> Plumb the use of GrBackendRenderTarget throughout Skia
>
> Bug: skia:
> Change-Id: Ib99a58d9552f5c7b8d77c09dcc72fa88326c26aa
> Reviewed-on: https://skia-review.googlesource.com/14148
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
> Commit-Queue: Greg Daniel <egdaniel@google.com>
>

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

Change-Id: I984e1909870182474c4c3cce257f01b6a9d8581f
Reviewed-on: https://skia-review.googlesource.com/14531
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Fold SkImageCacherator into SkImage_Lazy

SkImageCacherator still exists, but only as an interface implemented
(solely) by SkImage_Lazy. The only external clients are
GrImageTextureMaker and SkImage_Gpu::getDeferredTextureImageData.

This is probably an improvement, but doesn't go as far as I'd hoped.

Bug: skia:
Change-Id: I6812badfabb6924b025621b21af00cbde9c16cac
Reviewed-on: https://skia-review.googlesource.com/14371
Reviewed-by: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

jumper, replace _i with normal constants

So far I only seem to be encountering constant pools with float
constants, so integer constants should be easy to make normal.

This just removes _i.  There might be a couple integer constants
generated with C() too... they'll be the next CL.

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Win7-MSVC-Golo-CPU-AVX-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE41,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE2

Change-Id: Icc82cbc660d1e33bcdb5282072fb86cb5190d901
Reviewed-on: https://skia-review.googlesource.com/14527
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Plumb the use of GrBackendRenderTarget throughout Skia

Bug: skia:
Change-Id: Ib99a58d9552f5c7b8d77c09dcc72fa88326c26aa
Reviewed-on: https://skia-review.googlesource.com/14148
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Add additional checks when creating vulkan context

Makes sure we get the phsyical devices and queues we expect

Bug: skia:
Change-Id: I13031caa8aa85d680a8e921bee3237d6b53cebf0
Reviewed-on: https://skia-review.googlesource.com/14523
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Don't reboot Pixel C or Nexus Player before task

We reboot this after in swarming with real adb, so this
is unnecessary and could, in fact, be harmful.

Bug: skia:6559
NOTRY=true
Change-Id: Iaf58416994637111c02013fe3c83f4ac92de2a66
Reviewed-on: https://skia-review.googlesource.com/14526
Reviewed-by: Eric Boren <borenet@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Kevin Lubick <kjlubick@google.com>

Delete copyTo(Allocator), hide copyTo() behind flag

Replace uses of copyTo() in Skia.

Bug: skia:6464
Change-Id: I921dc53a1c29a5176d18f05741f7c0b5a008e548
Reviewed-on: https://skia-review.googlesource.com/14502
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Reed <reed@google.com>

In ok thread mode, use wait_util instead of wait_for.

wait_for(delta) is wait_until(steady_clock::now() + delta) under the
hood, so using wait_for() like this implies an extra call to now() that
we can avoid by using wait_until().

We can hoist that call out and just do it once...  the past stays the past.

This is not super important.  Just noticed while profiling.  It's nice
to keep the overhead of the ok tool down so the real work can show.  :)

Change-Id: I89d25a800b63ebcfc229b5b3aa3f2dd621f4e7b4
Reviewed-on: https://skia-review.googlesource.com/14480
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>

clear out C(), _i, and _f constants from SkJumper_vectors.h

I think this is just gonna be a bunch of baby steps for now.
This gets everything in SkJumper_vectors.h.

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Win7-MSVC-Golo-CPU-AVX-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE41,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE2

Change-Id: Ic87faa9bf6380a4fc9a577936dad8c3a9c48472e
Reviewed-on: https://skia-review.googlesource.com/14441
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>

remove legacy code

Bug: skia:
Change-Id: I719f68489e9295ec652ccb04de3b10632df534fa
Reviewed-on: https://skia-review.googlesource.com/14226
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>

Change vulkan image memory allocation back to using sub allocation

Bug: skia:
Change-Id: I8e1352c8d615b5a54ceba42500f0781767a1bd2d
Reviewed-on: https://skia-review.googlesource.com/14401
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>

Finish removing GrContext param from onRefEncodedData

TBR=reed@roogle.com

Bug: skia:5485
Change-Id: Ia4ed45ffc39f2ba9a80d4a1001208079142ae985
Reviewed-on: https://skia-review.googlesource.com/14323
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

Fix color space handling in SkImage_Gpu::getROPixels

The dstColorSpace is a badly named parameter. It's a hint about where/how
the returned pixels are going to be used. Raster and GPU are meant to
ignore that information, codecs use it to drive our decoding heuristic.

This is a re-land of https://skia-review.googlesource.com/c/10109/

Bug: skia:
Change-Id: Iee006a8e014e028b4f0f2471d7152b6bccd72cb2
Reviewed-on: https://skia-review.googlesource.com/14404
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>

add #defines to limit SkCpu

I just noticed we don't really have any CPU test bots that have less
than AVX anymore.  I'd like to make sure we're testing each mode of
SkJumper at least, so I've added flags to let us limit to SSE2 or
SSE4.1, the modes currently missing coverage.

Add the bots to test these modes too.

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE2,Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-SK_CPU_LIMIT_SSE41

Change-Id: I7c2b061332e7f037538488260583076c34ae7b1e
Reviewed-on: https://skia-review.googlesource.com/14405
Reviewed-by: Eric Boren <borenet@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Fix new IT blocks ARMv8

ARMv8 specifies that an IT block should be followed by only one 16-bit instruction.
* SkFloatToFix is back to a C implementation that mirrors the assembly code.

* S32A_D565_Opaque_neon switched the usage of the temporary 'ip' register to let
the compiler choose what is best in the context of the IT block. And replaced
'keep_dst' by 'ip' where low register or high register does not matter.

BUG=skia:

CQ_INCLUDE_TRYBOTS=skia.primary:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD

Change-Id: If587110a0c74b637ae99460419d46cf969c694fc
Reviewed-on: https://skia-review.googlesource.com/9346
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Fix uncached getROPixels from GPU images

We were allocating a local bitmap, but then attempting to read into an
uninitialized local pixmap. The only public API that funnels the caching
hint to this function is scalePixels, so I added a test (which previously
failed).

Bug: skia:
Change-Id: Ib4370350be664935b4c85e34c70b675e6d82ba64
Reviewed-on: https://skia-review.googlesource.com/14402
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>

Remove pre-allocated clip space from GrRenderTargetOpList

It seems like this should be safe (perf-regression-wise) since the arena allocated clips are used infrequently.

In the following, each GM run has 621 srcs and a total of 81087 ops.
Each skp run has 87 srcs and a total of 14709 ops.
The following table shows how many times the arena allocator is called for each config/src.
Overall, I think is safe to only allocate the arena-allocated space on an as needed basis.

gl-gm:       0
gl-skp:      0

gl-msaa4:    0
gl-msaa4:    0

glinst-gm:   1225
glinst-skp:  579

glnvpr4-gm:  623
glnvpr4-skp: 77

Change-Id: I78e04b060b6ad3b78697519249bb4b9c3193fc85
Reviewed-on: https://skia-review.googlesource.com/14392
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

prep for more constants

 - Add -z to print zero bytes instead of ...
 - avx+hsw will create 32-byte constants in .const,
   so we should disassemble those too, and align to 32 bytes.
 - The default _text section on Windows is 16-byte aligned,
   so we make a new one that's 32-byte aligned.

Change-Id: Icb2a962baa4c3735e98a992f2285eaf5cb1680fd
Reviewed-on: https://skia-review.googlesource.com/14364
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>

add SkLightingColorFilter for makeColorSpace()

This implementation should be identical to the previous one,
except we remember that the inputs to our matrix were colors
and transform them if so requested.

This fixes the GM lightingcolorfilter in gbr-8888 config.

Change-Id: Iea0d1079e1c0f5515019b3a4786fc5387dac92ff
Reviewed-on: https://skia-review.googlesource.com/14378
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Matt Sarett <msarett@google.com>

Add releaseAndAbandonGpuContext bot

Bug: skia:5142
Change-Id: I35bd42d1a9da2293092a2f09aaaf523a0f36537a
Reviewed-on: https://skia-review.googlesource.com/14376
Commit-Queue: Eric Boren <borenet@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

SkColorFilter::makeColorSpace()

Change-Id: Idea4afac67ef348c2f4e64f4fe6a03704a86652f
Reviewed-on: https://skia-review.googlesource.com/14375
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

SkPngCodec: Do not return kInvalidConversion on corrupt png

In this case, the fuzzer thinks there is a bug because we are
returning kInvalidConversion for a corrupt png file.

Bug: skia:6550
Change-Id: I33f588442f5eaa8a4d642e9328750779f9a9ef5d
Reviewed-on: https://skia-review.googlesource.com/14324
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>