experimental/xps_to_png: pass in DPI as program argument

NOTRY=true
Change-Id: Ieb4f43c69a4f185d8e7877c9c736bd6a7d7b1eac
Reviewed-on: https://skia-review.googlesource.com/4761
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

Start hooking shaders into SkRasterPipelineBlitter.

Here first just the simplest, constant-color shaders.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4743

Change-Id: I92c6523660e21a1e2aa353524570230282ba5dfe
Reviewed-on: https://skia-review.googlesource.com/4743
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Start plumbing deferral of GPU resources in the image filters

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4688

Change-Id: I904d999ea87a55bee73765aa96a2f8aef7a0d61c
Reviewed-on: https://skia-review.googlesource.com/4688
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

Apple's Clang can't assemble jsimd_arm_...

This should fix Build-Mac-Clang-arm64-Debug-GN_iOS_NoBuildbot.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4755

Change-Id: Ie49ce5642fb2d373102c9309074e13ee3035a569
Reviewed-on: https://skia-review.googlesource.com/4755
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Always build the ANGLE test code. Always build ANGLE on windows and linux."

This reverts commit 238b820369c4b556c2e7c8277855e6950de3d9eb.

Reason for revert: breaks 32-bit bots, iOS bots, Google3 roll, -ASAN bot.

Original change's description:
> Always build the ANGLE test code. Always build ANGLE on windows and linux.
>
> Make ANGLE test code independent of having ANGLE lib. Make ANGLE test code not include EGL headers.
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4040
>
> DOCS_PREVIEW= https://skia.org/?cl=4040
>
> Change-Id: I7b857e9785246743f53fb969647b1162ce7419ab
> Reviewed-on: https://skia-review.googlesource.com/4040
> Commit-Queue: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Mike Klein <mtklein@chromium.org>
>

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

Change-Id: I19bab8c93baebf032f8a4cefbedfe7359317e806
Reviewed-on: https://skia-review.googlesource.com/4758
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

Always build the ANGLE test code. Always build ANGLE on windows and linux.

Make ANGLE test code independent of having ANGLE lib. Make ANGLE test code not include EGL headers.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4040

DOCS_PREVIEW= https://skia.org/?cl=4040

Change-Id: I7b857e9785246743f53fb969647b1162ce7419ab
Reviewed-on: https://skia-review.googlesource.com/4040
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>

Documentation:  move some things around

NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=4750

Change-Id: I0d1bc77a533ede8e0514ff0b680a5aa851051fc6
Reviewed-on: https://skia-review.googlesource.com/4750
Reviewed-by: Heather Miller <hcm@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

Mark inline functions in SkTypes.h as static.

If these are not marked static (or SK_FORCE_INLINE) they become extremely dangerous to use from files built with different optimization flags than the baseline.  The One Definition Rule becomes trivially easy to violate, and the linker can easily pick a CPU-specific version of the function.

TBR=reed@google.com

BUG=chromium:664864

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4752

Change-Id: If3044a32229db1a1e3a8b83aace6047d92b36d05
Reviewed-on: https://skia-review.googlesource.com/4752
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

GN: turn on easy libjpeg-turbo SIMD backends (ARMv7 and ARMv8)

These are so easy we might as well...

I did a quick check of relevant-looking defines:
   - GYP defined WITH_SIMD, but it looks like that's already defined (by jconfig.h?);
   - GYP defined RGBX_FILLER_0XFF, but that affects only x86/x86-64;
   - GYP defined STRICT_MEMORY_ACCESS, which does nothing;
   - GYP defined MOTION_JPEG_SUPPORTED, which does nothing (and we'd probably not care anyway).

BUG=skia:5875

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4745

Change-Id: Ib1f28d354630be472c4d9648d5ade74a452a9e24
Reviewed-on: https://skia-review.googlesource.com/4745
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert[2] "Change SkCanvas to *not* inherit from SkRefCnt"

Changes over original:
- conditionalize ownership in SkPictureRecorder
- conditionalize ownership in SkCanvasStateUtils

This reverts commit b613c266df48cf45296ecc23d1bd7098c84bb7ba.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4742

Change-Id: Ib25514d6f546c69b6650b5c957403b04f7380dc2
Reviewed-on: https://skia-review.googlesource.com/4742
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>

Revert of Make SkSmallAllocator obey the RAII invariants and be expandable (patchset #15 id:280001 of https://codereview.chromium.org/2488523003/ )

Reason for revert:
bots crashing / asserting

Original issue's description:
> Make SkSmallAllocator obey the RAII invariants and move to heap structures when needed.
>
> The biggest change is to the API which allowed code to bypass the
> destruction invariants. This destruction bypass feature was needed in
> only one use, and is totally encapsulated using createWithIniterT.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2488523003
>
> Committed: https://skia.googlesource.com/skia/+/d5dc657b8c3ac916f98005dafdedafe02f023449
> Committed: https://skia.googlesource.com/skia/+/c18b5f8f57a4efc5d5d1e399ed8bd3bd02c592ab

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

Review-Url: https://codereview.chromium.org/2494353002

Make SkSmallAllocator obey the RAII invariants and move to heap structures when needed.

The biggest change is to the API which allowed code to bypass the
destruction invariants. This destruction bypass feature was needed in
only one use, and is totally encapsulated using createWithIniterT.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2488523003

Committed: https://skia.googlesource.com/skia/+/d5dc657b8c3ac916f98005dafdedafe02f023449
Review-Url: https://codereview.chromium.org/2488523003

[gn] Disable use of dng SDK in Fuchsia

These libraries aren't configured in Fuchsia and would need some GN
reworking to enable. We can wire this up in Fuchsia if/when this
capability when needed.

Change-Id: Iec2d339da07f44516592d6114563a8a4fb59d952
Reviewed-on: https://skia-review.googlesource.com/4744
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

SkPath::fIsVolatile does not need to be mutable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4696

Change-Id: I48b438ab1d99033ab2a422fb28e976cdf9188365
Reviewed-on: https://skia-review.googlesource.com/4696
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Fix SkModeColorFilter in 565

It has been incorrectly interpreting its SkColor as sRGB all the time.  Now, we plumb through the destintation color space and some scratch space, letting it decide how to interpret its SkColor later when it knows about the dst color space.  The scratch space is blitter scoped, which lets this be thread safe (this is much like SkShader::Context).

This only corrects the gamma transformation for now.  I've kept my previous TODO about gamut transformation.  Everything assumes sRGB gamut for now.

Shaders will get the same treatement in this pipeline.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4725

Change-Id: I55b0c7d5db9ad8d7dcdd6295c9dac61d10aeaed4
Reviewed-on: https://skia-review.googlesource.com/4725
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Revert "Annotate benign race on SkPath::fConvexity."

This reverts commit c51a3a4e3c81a1d4100d5d29d31bf3a02eda2a7c.

Reason for revert: had to revert this in the copy constructor.  Might as well revert here too for consistency.

Original change's description:
> Annotate benign race on SkPath::fConvexity.
>
> We're tired of this making our TSAN bot flaky, and equally tired of trying to fix it.
>
> BUG=skia:5776
>
> CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-TSAN-Trybot
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4691
>
> Change-Id: Ib6005282c514a6d785fd4fffe5387cbb1caccbe1
> Reviewed-on: https://skia-review.googlesource.com/4691
> Commit-Queue: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Ben Wagner <bungeman@google.com>
>

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

Change-Id: I906b1e017471d9e00fc58402136497459d881f72
Reviewed-on: https://skia-review.googlesource.com/4740
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

Revert "Annotate SkPath::fCovexity benign race in the other constructor too."

This reverts commit f978f12af553f899cfed7078ab63d3a71c6fbcb1.

Reason for revert: slowing the TSAN bot down to the point of timeouts.

Original change's description:
> Annotate SkPath::fCovexity benign race in the other constructor too.
>
> I overlooked the copy constructor last time.
>
> BUG=skia:5776
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4730
>
> Change-Id: I30bc89e1472dd48badf57664cfae8899f44bca9e
> Reviewed-on: https://skia-review.googlesource.com/4730
> Reviewed-by: Mike Klein <mtklein@chromium.org>
> Commit-Queue: Mike Klein <mtklein@chromium.org>
>

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

Change-Id: I6e34b1fc80fd98387d79521ce24f4545b990eb25
Reviewed-on: https://skia-review.googlesource.com/4739
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>

Initial implementation of a SkColorSpace_A2B xform

There is support for all features of SkColorSpace_A2B.

Tests for these functionality were adapted from
the XYZ xform, plus a CLUT-specific test was added.

Shared functions used by both SkColorSpaceXform_XYZ and SkColorSpaceXform_A2B
have been moved into a shared header.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2449243003
CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2449243003

Clean up glyph id handling.

Extract SkPackedID and its strongly typed subclasses SkPackedGlyphID and
SkPackedUnicharID out of SkGlyph. This simplifies the code handling
these types, as well as making it clearer that we wouuld eventually like
to get away from this scheme.

Changes SkScalerContext::getPath to take SkPackedGlyphID.
Changes SkScalerContext::generatePath to take SkGlyphID.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4722

Change-Id: I365c0c618b7ae0d348272155fac7761a69faa920
Reviewed-on: https://skia-review.googlesource.com/4722
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Herb Derby <herb@google.com>

Revert "Revert "Add integer texture support.""

This reverts commit 9c7edb8311409a141b0dc1d5e480c68c629f1997.

Fixes ASAN errors

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4736

Change-Id: I1b1dae754d357b01da7169c8e7c59d7d8d8a10f6
Reviewed-on: https://skia-review.googlesource.com/4736
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>

Revert of Set analytic AA as default. (patchset #8 id:140001 of https://codereview.chromium.org/2471133002/ )

Reason for revert:
SkASSERT triggered in some perf tests. Roll back.

Original issue's description:
> Set analytic AA as default.
>
> We'll use https://codereview.chromium.org/2471123002/ to keep Chrome from using
> analytic AA until we fix all the Chrome tests.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2471133002
>
> Committed: https://skia.googlesource.com/skia/+/dca4f6530013cf43a2557ccb07f5cb4fd916b8e8

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

Review-Url: https://codereview.chromium.org/2499723002

re-re-land of skslc now uses standard Skia caps

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4732

Change-Id: I144110bf66f67a28da7ad333173db43bddf9e8d0
Reviewed-on: https://skia-review.googlesource.com/4732
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>

Set analytic AA as default.

We'll use https://codereview.chromium.org/2471123002/ to keep Chrome from using
analytic AA until we fix all the Chrome tests.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2471133002

Review-Url: https://codereview.chromium.org/2471133002

Annotate SkPath::fCovexity benign race in the other constructor too.

I overlooked the copy constructor last time.

BUG=skia:5776

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4730

Change-Id: I30bc89e1472dd48badf57664cfae8899f44bca9e
Reviewed-on: https://skia-review.googlesource.com/4730
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Clean up validation layers

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4727

Change-Id: I92cef6fd23cd629a7a68e6ecb9e048674556517b
Reviewed-on: https://skia-review.googlesource.com/4727
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>

Snap y when updating cubic

If we don't snap, we could have end up with many tiny line segments
wihtin a single row of pixels. That combined with our partial alpha
rounding code (https://codereview.chromium.org/2483523002/) could
accumlate so much error that violates our alpha <= 256 assert.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4724

Change-Id: I58b02cffc8a4dc9f541e194d61a2f6fef7b85c97
Reviewed-on: https://skia-review.googlesource.com/4724
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Yuqian Li <liyuqian@google.com>

Re-purpose color space tool

(1) Compile tool with GN.
(2) Rename tool to colorspaceinfo.
(3) Support both images and icc profiles as input.
(4) Print out information on transfer fn in addition to gamut.

TODO: Output graphs of the three transfer fn curves.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4721

Change-Id: I2e75139685ea41446d3ae6f9803c8068ea05661a
Reviewed-on: https://skia-review.googlesource.com/4721
Commit-Queue: Matt Sarett <msarett@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Brian Osman <brianosman@google.com>

Revert "re-land of skslc now uses standard Skia caps"

This reverts commit 498d403f7703cb2157bf3c877b84906db5a06cd4.

Reason for revert: Breaking compile bot:
https://uberchromegw.corp.google.com/i/client.skia.compile/builders/Build-Ubuntu-GCC-x86_64-Release-NoGPU/builds/10116/steps/compile_skia%20on%20Ubuntu/logs/stdio

undefined reference to `GrGLSLCaps::GrGLSLCaps(GrContextOptions const&)

Original change's description:
> re-land of skslc now uses standard Skia caps
>
> BUG=skia:
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4726
>
> Change-Id: Ib1ea5d4269396e6c57ee1f0d7209aa9a9f2798e8
> Reviewed-on: https://skia-review.googlesource.com/4726
> Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
>

TBR=bsalomon@google.com,benjaminwagner@google.com,kjlubick@google.com,ethannicholas@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I7ee96d6c658be6b7b79027c147b8950c78c79a70
Reviewed-on: https://skia-review.googlesource.com/4729
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>

Add explicit UniqueID classes for GrGpuResource & GrSurfaceProxy

This sets the stage for using the Proxy's/RenderTargetContext's ID above the flush and the RenderTarget's/GrGpuResource's below the flush.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4650

Change-Id: I9f1e6b00c02a0691d90b58c49e1d8c60684884c1
Reviewed-on: https://skia-review.googlesource.com/4650
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

re-land of skslc now uses standard Skia caps

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4726

Change-Id: Ib1ea5d4269396e6c57ee1f0d7209aa9a9f2798e8
Reviewed-on: https://skia-review.googlesource.com/4726
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>

Tighten SkAAClip Bounds by Path Bounds

This bug chromium:662780 exists after our original fix (https://codereview.chromium.org/2477393002/) because this path (added in unit test) is calling blitAntiRect rather than blitAntiH when the path is drifted across the boundary. (The quadratic edge drifts across the boundary after an update and sets a dX=0 line segment which triggers blitAntiRect.)

Note that I didn't assert for the dLeft = dRite = 0 case because the left/right there won't drift after the SkTMin/SkTMax in line 964/966.

Theoretically we can revert the relaxation in https://codereview.chromium.org/2477393002/ (that's only a relaxation for analytic AA, not supersampled AA). However, consider that the initial landing of analytic AA is so painful, I decide to revert that relaxation only after our successful landing...

BUG=chromium:662780, chromium:662862
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2482193004

Review-Url: https://codereview.chromium.org/2482193004

GN: build Lua tools when skia_use_lua.

When we opt into Lua, this builds SampleLua into SampleApp, and the lua_app and lua_pictures tools.

I've tested this builds with and without skia_use_system_lua on my Mac laptop and Linux desktop.

I've made lua_pictures.cpp's flags static to avoid conflicts with flags in SkCommonFlags.cpp.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4699

DOCS_PREVIEW= https://skia.org/?cl=4699

Change-Id: I8176fd51d8a38746e7d730cfcce66da42b9a015a
Reviewed-on: https://skia-review.googlesource.com/4699
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

Documentation: more gn, less gyp

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4698
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=4698

Change-Id: I03100542752a769060a7f0c9671cc44acbea2e48
Reviewed-on: https://skia-review.googlesource.com/4698
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Hal Canary <halcanary@google.com>

SkFixedAlloc

Looking at SkSmallAlloc hasn't left me terribly impressed.  I think we can replace it with something a lot simpler to work with.

That simpler thing's core would be something like SkFixedAlloc, which allocates objects out of a fixed sized buffer, and cleans them up when done.

If needed, we can wrap that with logic to try to allocate out of an SkFixedAlloc, falling back on mallc() when exhausted.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4658

Change-Id: I8d94156ddf98802e42ec0890cff0f06b21f073b0
Reviewed-on: https://skia-review.googlesource.com/4658
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Reed <reed@google.com>

Skip left/right if alpha = 0 in SkRectClipCheckBlitter::blitAntiRect

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4684

Change-Id: I8c5e5f460dba76f9d74718f7ea104cb1f3cea993
Reviewed-on: https://skia-review.googlesource.com/4684
Commit-Queue: Yuqian Li <liyuqian@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>

Catch width overflow

BUG=chromium:662730

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4628

Change-Id: Iaf3a30d39fda3166a6f8fc62a30580629418dc88
Reviewed-on: https://skia-review.googlesource.com/4628
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Yuqian Li <liyuqian@google.com>

Revert "skslc now uses standard Skia caps"

This reverts commit 8af38a96475ac5ce83b20c16d9cf82bf1006b8c4.

Reason for revert: Breaking compile bots, e.g. from https://uberchromegw.corp.google.com/i/client.skia.compile/builders/Build-Ubuntu-GCC-x86_64-Debug-NoGPU/builds/10029/steps/compile_skia%20on%20Ubuntu/logs/stdio :

In function `sk_make_sp<GrGLSLCaps, GrContextOptions>'
undefined reference to `GrGLSLCaps::GrGLSLCaps(GrContextOptions const&)

In function `_Z10sk_make_spI10GrGLSLCapsI16GrContextOptionsEE5sk_spIT_EDpOT0_':
undefined reference to `GrGLSLCaps::GrGLSLCaps(GrContextOptions const&)

Original change's description:
> skslc now uses standard Skia caps
>
> BUG=skia:
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4660
>
> Change-Id: Idaedae3f81426b97f5052bb872cdf0610e47a84f
> Reviewed-on: https://skia-review.googlesource.com/4660
> Reviewed-by: Ben Wagner <benjaminwagner@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
>

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

Change-Id: Ic7f987f5c050ac2e333f5a0f16c8de85c1047a74
Reviewed-on: https://skia-review.googlesource.com/4697
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>

skslc now uses standard Skia caps

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4660

Change-Id: Idaedae3f81426b97f5052bb872cdf0610e47a84f
Reviewed-on: https://skia-review.googlesource.com/4660
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>

Annotate benign race on SkPath::fConvexity.

We're tired of this making our TSAN bot flaky, and equally tired of trying to fix it.

BUG=skia:5776

CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Release-TSAN-Trybot

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4691

Change-Id: Ib6005282c514a6d785fd4fffe5387cbb1caccbe1
Reviewed-on: https://skia-review.googlesource.com/4691
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Ben Wagner <bungeman@google.com>

Avoid platform-dependent function params in Fuzzer

We use this approach instead of T next() because different compilers
evaluate function parameters in different orders. If fuzz->next()
returned 5 and then 7, foo(fuzz->next(), fuzz->next()) would be
foo(5, 7) when compiled on GCC and foo(7, 5) when compiled on Clang.
By requiring params to be passed in, we avoid the temptation to call
next() in a way that does not consume fuzzed bytes in a single
platform-independent order.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4392

Change-Id: I35de849f82e8be45378f662a48100eb732fa8895
Reviewed-on: https://skia-review.googlesource.com/4392
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Kevin Lubick <kjlubick@google.com>

Check and report FreeType error.

BUG=chromium:648959

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4685

Change-Id: Ibc3cd5ddf37926accf1b7bf0959668510de903a8
Reviewed-on: https://skia-review.googlesource.com/4685
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Ben Wagner <bungeman@google.com>

Revert "Change SkCanvas to *not* inherit from SkRefCnt"

This reverts commit 824075071885b6b741c141cbe2134d8345d34589.

Reason for revert: Breaking WebView (chromium:663959)

Original change's description:
> Change SkCanvas to *not* inherit from SkRefCnt
>
> Definitely tricky for classes like SkNWayCanvas, where the caller (today)
> need not pay attention to ownership of the canvases it gave the NWay
> (after this CL, the caller *must* managed ownership)
>
> BUG=skia:
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4441
>
> DOCS_PREVIEW= https://skia.org/?cl=4441
>
> Change-Id: Ib1ac07a3cdf0686d78e7aaa4735d45cc90bea081
> Reviewed-on: https://skia-review.googlesource.com/4441
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
>

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

Change-Id: I5e3b3e876b7d2c09833cf841801321033b6b968b
Reviewed-on: https://skia-review.googlesource.com/4687
Commit-Queue: Heather Miller <hcm@google.com>
Reviewed-by: Heather Miller <hcm@google.com>

Revert "Fix BUG=chromium:664036"

This reverts commit 805eb6c7779eb855a26d0f1e821d58ce2bf5ffd2.

Reason for revert: Backing up to original change causing issues w/ WebView (chromium:663959)

Original change's description:
> Fix BUG=chromium:664036
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4659
>
> Change-Id: Idac00fef9cac023649a50d9c70d107f988fd2510
> Reviewed-on: https://skia-review.googlesource.com/4659
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Hal Canary <halcanary@google.com>
>

TBR=halcanary@google.com,reed@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Change-Id: I1c3258a3d46a713e9fa60fef829b8b6c6d73b2c4
Reviewed-on: https://skia-review.googlesource.com/4686
Commit-Queue: Heather Miller <hcm@google.com>
Reviewed-by: Heather Miller <hcm@google.com>

skpbench: add utility to format results for Skia Perf

Adds skiaperf.py for formatting skpbench.py outputs for Skia Perf.
Also renames parseskpbench.py to sheet.py.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4657

Change-Id: I758678e1c589b15ec2d07c43e4921663e919b47b
Reviewed-on: https://skia-review.googlesource.com/4657
Commit-Queue: Kevin Lubick <kjlubick@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>

Fix BUG=skia:5662

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4661

Change-Id: I5f5b10387be5d538347a3626c165cc17af154212
Reviewed-on: https://skia-review.googlesource.com/4661
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

Revert "Add integer texture support."

This reverts commit 434c534bd0ec08cb355fecfb6d2110197b523e74.

Reason for revert: Undefined behavior. From

https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-Clang-Golo-GPU-GT610-x86_64-Debug-ASAN/builds/1272/steps/test_skia%20on%20Ubuntu/logs/stdio

../../../tests/IntTextureTest.cpp:51:44: runtime error: left shift of negative value -1
    #0 0x2257480 in test_IntTexture(skiatest::Reporter*, sk_gpu_test::ContextInfo const&) (/b/swarm_slave/w/ir0aO1sa/out/Debug/dm+0x2257480)
    #1 0x1ca1066 in skiatest::RunWithGPUTestContexts(void (*)(skiatest::Reporter*, sk_gpu_test::ContextInfo const&), bool (*)(sk_gpu_test::GrContextFactory::ContextType), skiatest::Reporter*, sk_gpu_test::GrContextFactory*) (/b/swarm_slave/w/ir0aO1sa/out/Debug/dm+0x1ca1066)
    #2 0x1ca080d in run_test(skiatest::Test) (/b/swarm_slave/w/ir0aO1sa/out/Debug/dm+0x1ca080d)
    #3 0x1c9e5e9 in dm_main() (/b/swarm_slave/w/ir0aO1sa/out/Debug/dm+0x1c9e5e9)
    #4 0x7f2d2ba8df44 in __libc_start_main /build/eglibc-oGUzwX/eglibc-2.19/csu/libc-start.c:287
    #5 0x1bb3028 in _start (/b/swarm_slave/w/ir0aO1sa/out/Debug/dm+0x1bb3028)

SUMMARY: AddressSanitizer: undefined-behavior ../../../tests/IntTextureTest.cpp:51:44 in
step returned non-zero exit code: 1

Original change's description:
> Add integer texture support.
>
> This allows us to create integer textures and sample them from a GrProcessor's code.
>
> Filtering is limited to NEAREST.
>
> Adds tests for reading/writing pixels, copying, and drawing. These operations are not allowed to convert to fixed/float configs.
>
> Vulkan support is TBD.
>
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4348
>
> Change-Id: If38d89a03285d4bd98d1f14f9638b0320977e43d
> Reviewed-on: https://skia-review.googlesource.com/4348
> Commit-Queue: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Chris Dalton <csmartdalton@google.com>
>

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

Change-Id: I39f1a0a0dd3e6cde1143c8cc1217d2e3d5977b21
Reviewed-on: https://skia-review.googlesource.com/4663
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>

Fix BUG=chromium:664036

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4659

Change-Id: Idac00fef9cac023649a50d9c70d107f988fd2510
Reviewed-on: https://skia-review.googlesource.com/4659
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

SkPDF: paths as rects when possible

http://crrev.com/2269593003

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4652

Change-Id: I12e238e8f00ac1c54a1b8dcf82a5502240563288
Reviewed-on: https://skia-review.googlesource.com/4652
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>

Add integer texture support.

This allows us to create integer textures and sample them from a GrProcessor's code.

Filtering is limited to NEAREST.

Adds tests for reading/writing pixels, copying, and drawing. These operations are not allowed to convert to fixed/float configs.

Vulkan support is TBD.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4348

Change-Id: If38d89a03285d4bd98d1f14f9638b0320977e43d
Reviewed-on: https://skia-review.googlesource.com/4348
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>

Simplify SkBitmapProvider

We're no longer using SkBitmap sources for SkBitmapProvider, so we can
restrict it to SkImage only.

At this point we could also remove it, but I think it'll come in handy
when we add shader subset support.

BUG=skia:5806
R=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4655

Change-Id: I298da253cc8f7c1205b543521e62060a202a9f78
Reviewed-on: https://skia-review.googlesource.com/4655
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>

Update skia.org docs to reference Gerrit instead of Rietveld

BUG=skia:5958

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4648
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=4648

Change-Id: I6a57b7f9b92f6b769ba2b78f8e0ab9f4ae7bb24e
Reviewed-on: https://skia-review.googlesource.com/4648
Commit-Queue: Ravi Mistry <rmistry@google.com>
Reviewed-by: Heather Miller <hcm@google.com>

skpbench: add "resultsfile" option

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4654

Change-Id: I1a26eddb40de1398cad5348d3fe0ba397a87dbb0
Reviewed-on: https://skia-review.googlesource.com/4654
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>

Snug the clip with path bounds inside SkAAClip

https://codereview.chromium.org/2482193004/ is related with this patch.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4636

Change-Id: Ibd6d64578e08c75a3cfcef45489afc9b59ad40be
Reviewed-on: https://skia-review.googlesource.com/4636
Commit-Queue: Yuqian Li <liyuqian@google.com>
Reviewed-by: Mike Reed <reed@google.com>

Make SSE2/Neon convolution functions not to read extra bytes

This change makes SSE2/Neon horizontal convolution functions do not read
extra pixels past the end of the buffer. So we can remove all the SIMD
specific logic in SkConvolver to deal with last couple of rows and also
avoid applying padding to convolution filters.

Performance impact is small. Nanobench time change:
                              SSE2    NEON
bitmap_scale_filter_64_256     1%     -2%
bitmap_scale_filter_256_64     1%      2%
bitmap_scale_filter_90_10      1%     -1%
bitmap_scale_filter_90_30      1%      0%
bitmap_scale_filter_90_80      1%      0%
bitmap_scale_filter_90_90      1%      1%
bitmap_scale_filter_80_90      0%      0%
bitmap_scale_filter_30_90      3%      6%
bitmap_scale_filter_10_90      0%      2%

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2481733003
CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2481733003

Clamp parametric gamma values to 0-1 range

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4641

Change-Id: I1c6963da956500deea559d5ece31529add89980a
Reviewed-on: https://skia-review.googlesource.com/4641
Reviewed-by: Robert Aftias <raftias@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

skpbench: remove --path flag

Replaces the confusing --path flag with a positional argument that
gives the path to the skpbench binary.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4635

Change-Id: I2b5eeca61ec85e7fe45fd3370625eddf34a2fd0e
Reviewed-on: https://skia-review.googlesource.com/4635
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Kevin Lubick <kjlubick@google.com>

[nobuildbot] Copy timeouts from swarm_trigger

BUG=skia:

TBR=benjaminwagner
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4642

Change-Id: Ia81609af8ae11c0171c99c459e8711fa786c1e20
Reviewed-on: https://skia-review.googlesource.com/4642
Commit-Queue: Eric Boren <borenet@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>

Remove call to getMetrics from GlyphGenerator.

GlyphGenerator::generatePath does not need to call getMetrics before
calling getPath. This is wasted effort, especially since the
SkScalerContext is being used directly instead of going through a glyph
cache (which would require this ordering).

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4620

Change-Id: I140a6715a1d1692b161ee3dae48e5482b6004b63
Reviewed-on: https://skia-review.googlesource.com/4620
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>

Revert of Make SkSmallAllocator obey the RAII invariants and be expandable (patchset #6 id:100001 of https://codereview.chromium.org/2488523003/ )

Reason for revert:
Crashing Mac Perf and Test bots.

This is a flaky but extremely likely crash.  I've only seen one Mac Perf or Test bot that had this patch that didn't crash.

This should be easy to reproduce like this:
  $ gn gen out --args=is_debug=false
  $ ninja -C out dm
  $ out/dm -m xfermodes3 --config gpu

This is crashing every time I run it on my laptop, and never when I revert this CL.

Building in release and running --config gpu probably don't matter.

Original issue's description:
> Make SkSmallAllocator obey the RAII invariants and move to heap structures when needed.
>
> The biggest change is to the API which allowed code to bypass the
> destruction invariants. This destruction bypass feature was needed in
> only one use, and is totally encapsulated using createWithIniterT.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2488523003
>
> Committed: https://skia.googlesource.com/skia/+/d5dc657b8c3ac916f98005dafdedafe02f023449

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

Review-Url: https://codereview.chromium.org/2485853005

Android builders do not need host toolchains.

This trims a few packages.  We do the same in buildbot.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4634

Change-Id: I4c76d744917cb244ca3390d55d9ea602cda2dd58
Reviewed-on: https://skia-review.googlesource.com/4634
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

auto-add Build jobs

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4639

Change-Id: I61acce39abf08c4307e5a7ec990ec8857539f943
Reviewed-on: https://skia-review.googlesource.com/4639
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

skpbench: call adb wait-for-device after root

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4632

Change-Id: I310600e79141d2eafd668c76b97a017b151ac0fd
Reviewed-on: https://skia-review.googlesource.com/4632
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>

added constant folding & branch elimination to skslc

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2489673002

Committed: https://skia.googlesource.com/skia/+/6136310ee8f43247548bcefcaeca6d43023c10aa
Review-Url: https://codereview.chromium.org/2489673002

skpbench: filter out junk spewed by Pixel C driver

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2488043003

Review-Url: https://codereview.chromium.org/2488043003

Fix implicit nullability cast of [NSScreen mainScreen].

Without this patch I get this warning-as-error with XCode 8.1 (8B62):

../src/views/mac/SkOSWindow_Mac.mm:88:43: error: implicit conversion from nullable pointer 'NSScreen * _Nullable' to non-nullable pointer type 'NSScreen * _Nonnull' [-Werror,-Wnullable-to-nonnull-conversion]
    [(SkNSView*)fHWND enterFullScreenMode:[NSScreen mainScreen] withOptions:nil];

Seems like the thing to do is explicitly null check [NSScreen mainScreen] before calling enterFullScreenMode?

(google.com/images?q=i+have+no+idea+what+i+am+doing)

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4631

Change-Id: I20c4a2b559a347a6a128479b81cb515692832e72
Reviewed-on: https://skia-review.googlesource.com/4631
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Jim Van Verth <jvanverth@google.com>

Revert "add alias for xfermode::getproc to allow legacy g client to bulid"

Alias no longer needed

This reverts commit 40c07dc797ada471ba89cb81acc4b7b93f7374d7.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4630

Change-Id: Ia3d13d6b1055ecb4b569c99f6047e4ab3d21f107
Reviewed-on: https://skia-review.googlesource.com/4630
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>

[nobuildbot] AndroidOne bots

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4627

Change-Id: I9479ba49cc75f658b0cf0ac724c6462cb06598ea
Reviewed-on: https://skia-review.googlesource.com/4627
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Commit-Queue: Eric Boren <borenet@google.com>

Make SkSmallAllocator obey the RAII invariants and move to heap structures when needed.

The biggest change is to the API which allowed code to bypass the
destruction invariants. This destruction bypass feature was needed in
only one use, and is totally encapsulated using createWithIniterT.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2488523003

Review-Url: https://codereview.chromium.org/2488523003

add debugging wrapper blitter to ensure unclipped scan-conversion is safe

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4629

Change-Id: Ie6c35b28a6e8e87bf230e50f9940d3924d12969a
Reviewed-on: https://skia-review.googlesource.com/4629
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Yuqian Li <liyuqian@google.com>

Include what you use with signbit.

Include cmath in a few source files which use signbit and a relying on
magic to happen to use it.

Also Fix nuttiness in SampleClip. No need to #define single character
identifiers.

Change-Id: Iae3352d0cab9aaa6c37d6424f064b3d86fa2e011
Reviewed-on: https://skia-review.googlesource.com/4626
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>

skpbench: use 3 cores instead of 4 on Pixel C

Hopefully with less cores running we will produce less heat. 3 cores
should be enough to run Ganesh, the graphics driver, and the OS.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2491633002

Review-Url: https://codereview.chromium.org/2491633002

Revert of added constant folding & branch elimination to skslc (patchset #3 id:40001 of https://codereview.chromium.org/2489673002/ )

Reason for revert:
Breaking the DEPS roll:

../../third_party/skia/src/sksl/SkSLIRGenerator.cpp:257:70: error: chosen constructor is explicit in copy-initialization
            return std::unique_ptr<Statement>(new Block(s.fPosition, { }, fSymbolTable));
                                                                     ^~~
../../build/linux/debian_wheezy_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/4.6/../../../../include/c++/4.6/debug/vector:77:7: note: constructor declared here
      vector(const _Allocator& __a = _Allocator())
      ^
../../third_party/skia/src/sksl/ir/SkSLBlock.h:20:70: note: passing argument to parameter 'statements' here
    Block(Position position, std::vector<std::unique_ptr<Statement>> statements,
                                                                     ^

Original issue's description:
> added constant folding & branch elimination to skslc
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2489673002
>
> Committed: https://skia.googlesource.com/skia/+/6136310ee8f43247548bcefcaeca6d43023c10aa

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

Review-Url: https://codereview.chromium.org/2494523002

Fix fuzzer assert in GradientShaderBase4fContext::TSampler

The arithmetic in tileProc<kMirror> may cause the result to snap to the
open interval value - which violates invariants down the line.

We need to clamp the result to nextafterf(2, 0) to ensure it stays less
than two.

BUG=skia:5913
R=reed@google.com,herb@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2472763002

Review-Url: https://codereview.chromium.org/2472763002

Change iOS bots to new-style, no-buildbot bots

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4582

Change-Id: Ied11c727a491382fce113b6e02831a67d944ca2f
Reviewed-on: https://skia-review.googlesource.com/4582
Commit-Queue: Eric Boren <borenet@google.com>
Reviewed-by: Ben Wagner <benjaminwagner@google.com>
Reviewed-by: Stephan Altmueller <stephana@google.com>

Change SkCanvas to *not* inherit from SkRefCnt

Definitely tricky for classes like SkNWayCanvas, where the caller (today)
need not pay attention to ownership of the canvases it gave the NWay
(after this CL, the caller *must* managed ownership)

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4441

DOCS_PREVIEW= https://skia.org/?cl=4441

Change-Id: Ib1ac07a3cdf0686d78e7aaa4735d45cc90bea081
Reviewed-on: https://skia-review.googlesource.com/4441
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Reviewed-by: Robert Phillips <robertphillips@google.com>

Do not call blitV with 0 height

This is causing SkAAClipBlitter to crash.

BUG=chromium:662952
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2485243002

Review-Url: https://codereview.chromium.org/2485243002

add alias for xfermode::getproc to allow legacy g client to bulid

BUG=skia:
NOTRY=True

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4623

Change-Id: Ib8108c6e2f2ca4592428711c2ae8a0bd1fa7963d
Reviewed-on: https://skia-review.googlesource.com/4623
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Reed <reed@google.com>

Make uploading to Gerrit the default for Skia

BUG=skia:5612

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4621
# Does not get tested with CQ.
NOTRY=true

Change-Id: I8d3fbfc6e557948cff7722f0b42ea09dc9be68e1
Reviewed-on: https://skia-review.googlesource.com/4621
Commit-Queue: Ravi Mistry <rmistry@google.com>
Reviewed-by: Eric Boren <borenet@google.com>

Missing SkOSPath include in lua_pictures.cpp

TBR=bungeman@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2494443002

Review-Url: https://codereview.chromium.org/2494443002

added constant folding & branch elimination to skslc

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2489673002

Review-Url: https://codereview.chromium.org/2489673002

remove use of xfermode* in procs

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4592

Change-Id: I99f35924ff5325dfac527bb573a86d2d0366e0b3
Reviewed-on: https://skia-review.googlesource.com/4592
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Reed <reed@google.com>

Enable SkImage_pinAsTexture to signal if the operation succeeded.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4488

Change-Id: Id667f385b172e4a56faa9324066f57295af30ef3
Reviewed-on: https://skia-review.googlesource.com/4488
Commit-Queue: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Fix the meaning of stop_y

stop_y means that we should stop exactly at stop_y, so the last row should be
[stop_y - 1, stop_y], not [stop_y, stop_y + 1].

Somehow this misunderstanding didn't trigger any issue until Chrome exercises
SkAAClip with some websites (e.g.,
http://www.lemonde.fr/elections-americaines/article/2016/11/07/deux-programmes-deux-visions-de-l-amerique_5026444_829254.html).
When we blitter the extra row [stop_y, stop_y + 1], the SkAAClip will return
nullptr by findRow. Later when that nullptr row is used to findX, the crash
happened.

BUG=chromium:662925, chromium:662776
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2481703004

Review-Url: https://codereview.chromium.org/2481703004

skpbench: simplify adb and reduce number of invocations

This change reduces a bit of startup overhead by condensing more bash
into fewer invocations of 'adb shell'.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2481413003

Review-Url: https://codereview.chromium.org/2481413003

Check negative overflow of quickSkFDot6Div

The following fuzz html reveals the bug in chromium content_shell

===html=begins===
<style>
*{min-width:4%;-webkit-border-radius:+256%;}
.CLASS11{text-decoration:rgba(128%,16129%,1%,0.0000000004656612317904879831274006762300773920593144339363789186) dotted blink;vertical-align:124px;-webkit-column-count:2147483655 !important;</style>
<h1 class="CLASS11 CLASS1">
>
B
<button>
<h4 class="CLASS11 CLASS12">
</h4>
<p>
c C
<table>
<caption class="CLASS11">
>
<ruby class="CLASS11 CLASS12">
</ruby>
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA x
===html===ends===

BUG=chromium:662905
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2482863004

Review-Url: https://codereview.chromium.org/2482863004

Fix assets.py args

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4588

Change-Id: Id2ddbabf8920e59341555f7f4c65a9b1ccabf4d6
Reviewed-on: https://skia-review.googlesource.com/4588
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Eric Boren <borenet@google.com>

Move GrRenderTargetPriv::maxWindowRectangles to GrRenderTargetContextPriv & GrRenderTargetProxy

This removes a reason to call accessRenderTarget on the GrRenderTargetContext

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4583

Change-Id: I6e8a53ffd5c1fea80f542b70e05744e2991f70f8
Reviewed-on: https://skia-review.googlesource.com/4583
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

Fix fuzzer bug

BUG=663687

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4580

Change-Id: I9e914a1134b8049cf62bb36d4a6145eb8487d4e6
Reviewed-on: https://skia-review.googlesource.com/4580
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>

Consolidate GrSurfaceProxy Make methods

I think GrSurfaceDesc is still the most compact way to communicate the deferred GrSurface's settings to the Proxy but this CL, at least, reduces where it is used.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4559

Change-Id: Ica599c28a5aef1ed4094f47a4ac119e2d204d652
Reviewed-on: https://skia-review.googlesource.com/4559
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

Rename SkSourceGammaTreatment to SkDestinationSurfaceColorMode

This is much more explicit about what that type represents (are we in
legacy mode or not), which also makes it suitable for other (upcoming)
usage.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4529

Change-Id: Iacb397c34e7765f1ca86c0195bc622b2be4d9acf
Reviewed-on: https://skia-review.googlesource.com/4529
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>

skpbench: fix typo in 6P bash script

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2486973002

Review-Url: https://codereview.chromium.org/2486973002

Refactor RGBA/BGRA xform logic in SkCodecs

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4554

Change-Id: Ic9105a2806b915fc56b6810a80dd444561d0d959
Reviewed-on: https://skia-review.googlesource.com/4554
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Relax the extra span's alpha

BUG=chrome:662862
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2477393002

Review-Url: https://codereview.chromium.org/2477393002

Do not call blitAntiRect for empty rect

BUG=chromium:662800
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2490673002

Review-Url: https://codereview.chromium.org/2490673002

Fail jpeg decodes on too many progressive scans

BUG:642462

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4560

Change-Id: I22891ce1e0b3a1bedefc34dadd5cf34dfc301b79
Reviewed-on: https://skia-review.googlesource.com/4560
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>

Fix query for GR_GL_IMPLEMENTATION_COLOR_READ_FORMAT for GrPixelConfigs that aren't renderable

We recently separated renderabibilty by Skia from FBO-bindable ( https://skia-review.googlesource.com/4345). Overlooked in that change was the necessity to be able to query GR_GL_IMPLEMENTATION_COLOR_READ_FORMAT and _TYPE for formats that we don't support rendering to Skia but are legal color attachment formats.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2489643002

Review-Url: https://codereview.chromium.org/2489643002

Move SkOSPath out of include/core.

It is moved to src/utils. It is almost a tool, but has two uses in
src/ports.

The existing SkOSFile.cpp is left empty for the time being since it is
mentioned in Chromium's BUILD.gn for Skia.

Change-Id: I3bb7f7c4214359eb6ab906bfe76737d20bf1d6c7
Reviewed-on: https://skia-review.googlesource.com/4536
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Ben Wagner <bungeman@google.com>

Make the infra tests part of the CQ

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4550

Change-Id: I3ed720e5608282cd9fa14365e6f4fa4338c81ef0
Reviewed-on: https://skia-review.googlesource.com/4550
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Eric Boren <borenet@google.com>

Smart pointers for GrBatchAtlas.

While navigating this code the ownership rules were found to not be
entirely clear. This updates GrBatchAtlas:: fPlotArray and fTexture to
be smart pointers and propagates some clarifying changes.

Change-Id: I6ca67247575c2d7c4e7986c10b948201fe0080f1
Reviewed-on: https://skia-review.googlesource.com/4549
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>

Add -j as alias for --threads.

I keep typing this...

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4553

Change-Id: Ib3e387054810cebae468addc001a1e148e5af155
Reviewed-on: https://skia-review.googlesource.com/4553
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>