Mark temp paths in SkBaseDevice::drawArc and ::drawDRRect as volatile
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261123004

Review-Url: https://codereview.chromium.org/2261123004

Fix touch event parameter order

The issue is so obvious... However, all newer Android devices (e.g.,
Nexus 6P, 5X, 9) are immune to this bug. So I didn't notice this until I
tested Viewer app on Nexus 7 yesterday...

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254223002

Review-Url: https://codereview.chromium.org/2254223002

Revert of Moving SkBlurImageFilter into core (patchset #8 id:140001 of https://codereview.chromium.org/2255803003/ )

Reason for revert:
Fixed it wrong, needs to revert to re-discuss and re-land.

Original issue's description:
> Moving SkBlurImageFilter into core
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255803003
>
> Committed: https://skia.googlesource.com/skia/+/dd3259eb95c3b47e11eefa3b176365a112a32b48
> Committed: https://skia.googlesource.com/skia/+/e426babe7552b1cb4e27cdf4e90826feabb9e3b0

TBR=reed@google.com,fmalita@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/2266063002

Moving SkBlurImageFilter into core

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255803003

Committed: https://skia.googlesource.com/skia/+/dd3259eb95c3b47e11eefa3b176365a112a32b48
Review-Url: https://codereview.chromium.org/2255803003

Add stroke-and-fill circular arc test, test near 180deg arc in all circular_arc_tests

BUG=skia:5227
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261233002

Review-Url: https://codereview.chromium.org/2261233002

Detect all named gammas

Our DstColorSpace UMA is showing some named gammas that are
not appropriately detected and placed in named categories.
https://uma.googleplex.com/p/chrome/histograms?endDate=latest&dayCount=1&histograms=Blink.ColorSpace.Destination&fixupData=true&showMax=true&filters=isofficial%2Ceq%2CTrue&implicitFilters=isofficial

This CL should fix that.

I'm not sure (yet) how I feel about this landing permanently.
Seems a little messy.

But it will be interesting to see how this affects the UMA.
My best guess is that we are hitting this case when all
three gammas are "invalid" in different ways.  I'm expecting
to see some profiles end up in the "invalid" category now.

It's also possible that we'll see these cases being absorbed
into sRGB or somewhere else.

BUG=skia:5656
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261213002

Review-Url: https://codereview.chromium.org/2261213002

compress_r11eac_blocks() required more alignment than dst has.

This shouldn't change any behavior except that the stores to dst
will no longer require 8-byte alignment.

Empirically it seems like we can use 4-byte alignment here,
but u8 (i.e. 1-byte alignment) is always safe.

BUG=skia:5637
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2264103002
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/2264103002

Revert of Moving SkBlurImageFilter into core (patchset #6 id:100001 of https://codereview.chromium.org/2255803003/ )

Reason for revert:
It looks like this breaks our roll into Google3:
https://test.corp.google.com/ui#id=OCL:130943857:BASE:130944046:1471881622765:61dbdd3

Original issue's description:
> Moving SkBlurImageFilter into core
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255803003
>
> Committed: https://skia.googlesource.com/skia/+/dd3259eb95c3b47e11eefa3b176365a112a32b48

TBR=reed@google.com,fmalita@chromium.org,vjiaoblack@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/2263283002

Use ARMv8 CRC32 instructions for SkOpts::hash().

For large inputs, this runs ~11x faster than Murmur3.
My bench drops from 1µs to 88ns.

Like x86-64, this runs fastest if we work in 24 byte chunks.  16 byte chunks
run at about 0.75x this speed, 8 byte chunks at about 0.4x (which would still
be about 5x faster than Murmur3).

This'll require plumbing support for opts_crc32 into Chrome first before it can roll.

perf.skia.org charts we want to watch: https://perf.skia.org/#5490

Seach for compute_hash in these logs to see the difference:
baseline: https://luci-milo.appspot.com/swarming/task/30ba22f3dfe30e10/steps/nanobench/0/stdout
trybot: https://luci-milo.appspot.com/swarming/task/30bbc406cbf62d10/steps/nanobench/0/stdout

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2260823002
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/2260823002

Moving SkBlurImageFilter into core

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255803003

Review-Url: https://codereview.chromium.org/2255803003

Support color xforms for kIndex8 pngs

This change started as: "Always use color xforms to
premultiply".  We need to be in a linear space to
premultiply correctly.

It became clear that we also need to support kIndex8
color xforms in order to make this change.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2246143002

Review-Url: https://codereview.chromium.org/2246143002

'g' key toggles showgrid for PerlinPatch

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2269673002

TBR=

Review-Url: https://codereview.chromium.org/2269673002

Supply a color space in sRGB/F16 mode of SampleApp

Ganesh now uses this to decide if it should perform gamma-correct
rendering, so the color space is necessary to get correct results in
S32/F16.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2266763004

Review-Url: https://codereview.chromium.org/2266763004

Roll recipe DEPS

BUG=skia:5627
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2266933002

Review-Url: https://codereview.chromium.org/2266933002

'F' will toggle filtering for the --picture sample

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2263203002

TBR=

Review-Url: https://codereview.chromium.org/2263203002

Fix color xform width bug when scaling/subsetting

This was not caught by the bots because we don't test
color correct modes with our many image decoding tests
(takes too long).

Adding a unit test.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2247743002

Review-Url: https://codereview.chromium.org/2247743002

Build command buffer with is_component_build=false

BUG=skia:5630
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2265913002

Review-Url: https://codereview.chromium.org/2265913002

Adjust angles in circular arcs GMs to test smaller arcs

BUG=skia:5227
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2267713002

Review-Url: https://codereview.chromium.org/2267713002

add SkMatrixPriv for specialized helpers

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2268443002

Review-Url: https://codereview.chromium.org/2268443002

GN: _turbo -> -turbo to match Fuchsia

We might as well match the folks who are using our GN files now.

We've got plenty of strategies in our pocket for when we try to move Chrome
onto our GN files (and who knows, there may be even a new better way by then):
  * Same sort of rename in Chrome's third_party
  * Aliased targets via //build/secondary in Chrome.
  * Indirection via build_overrides

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2265503002

Review-Url: https://codereview.chromium.org/2265503002

Simplify the gradient color stop optimizer

We're just discarding the leftmost/rightmost stops -- this can be
achieved by simply adjusting the start offset and count.

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2265803002

Review-Url: https://codereview.chromium.org/2265803002

Add a gradient edge optimization test

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2267593002

Review-Url: https://codereview.chromium.org/2267593002

make zoomer text clearer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2267623002

TBR=

Review-Url: https://codereview.chromium.org/2267623002

add translate to exercise non-identity ctm performance

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261063002

TBR=

Review-Url: https://codereview.chromium.org/2261063002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261983002

Review-Url: https://codereview.chromium.org/2261983002

remove unused field from SkDrawIter

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2267553002

TBR=

Review-Url: https://codereview.chromium.org/2267553002

store info in basedevice, change getter to non-virtual const&

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261003003

Review-Url: https://codereview.chromium.org/2261003003

src/utils/SkBitSet: simplify

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2265623002

Review-Url: https://codereview.chromium.org/2265623002

Rename circular_arcs_circular_arcs_fill to circular_arcs_fill.

Also fix issue with second paint's alpha getting overwritten.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2265643002

Review-Url: https://codereview.chromium.org/2265643002

Move GrFixedClip into src directory

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254343003

Review-Url: https://codereview.chromium.org/2254343003

Added a small widget framework to the interactive bevel SampleApp

I made a small framework to add slider and radial controls more easily.
The Sample now has controls for light direction and color.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259183003

Review-Url: https://codereview.chromium.org/2259183003

Plumb drawArc to SkDevice.

Plumbs the drawArc canvas method down to SkDevice without converting to a path. Plumbs through the various recording canvas classes.

BUG=skia:5227
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257023003

Review-Url: https://codereview.chromium.org/2257023003

Added SkColor4f::FromColor3f(SkColor3f, float a)

Planning to use this for the ColorDisplay widget in https://codereview.chromium.org/2259183003/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2264643002

Review-Url: https://codereview.chromium.org/2264643002

Add GMs for testing drawArc.

This is in preparation for virtualizing this on SkCanvas/SkDevice.

Tests the following variations:

aa and non-aa
stroke/hairline/fill
cap types
using center point and not

BUG=skia:5227
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2264593002

Review-Url: https://codereview.chromium.org/2264593002

Fast path translate() in SkCanvas and SkLiteDL.

This adds didTranslate() so that SkLiteDL (and other canvas recorders)
can record the translate rather than the full concat.

It also adds a case to SkMatrix::preTranslate() to fast path
translate x translate -> translate (i.e. +=).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255283002

Committed: https://skia.googlesource.com/skia/+/5fa47f4fd13b3158de4599414c86d17649c2dd1c

Misc bots failing in pictureimagefilter replay modes.
https://luci-milo.appspot.com/swarming/task/30b8e53f3a1f4f10/steps/dm/0/stdout

Problem is FMA vs. not.

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

Review-Url: https://codereview.chromium.org/2255283002

Implement gradient simplification for 0,0,1 and 0,1,1 gradients

Depends on https://codereview.chromium.org/2259823005/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2256843004

Review-Url: https://codereview.chromium.org/2256843004

Simplify embeddability test.

There is no reason to read the whole OS/2 table, and no need to spell
out the full names of all the types over and over.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2256783002

Review-Url: https://codereview.chromium.org/2256783002

Fix bug, always keep fIsScaleTranslate in correct state

BUG:639179
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2263513003

Review-Url: https://codereview.chromium.org/2263513003

Step 1 of opts_crc32 rollout.

  - This adds an crc32_sources variable to opts.gypi and hooks it into
    our GYP build with a dummy source file.

Step 2 will hook crc32_sources into Chrome's GN build.
Step 3 will actually land some code using CRC32 instructions.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259233002

Review-Url: https://codereview.chromium.org/2259233002

Some assert fixes for running the Fuzzer sample in GPU mode.

quadraticPointCount, cubicPointCount: passing NaN to GrNextPow2 causes
it to assert, so detect that and return the max points per curve
instead.

Passing a zero scale to GrStyle::applyToPath() makes it assert, so
draw nothing (likely what drawing with a zero scale is going to do
anyway).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257423002

Review-Url: https://codereview.chromium.org/2257423002

Return color shader instead of 2-color gradient when color count is 1

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259823005

Committed: https://skia.googlesource.com/skia/+/da082a5767d7edfd3abe74fc683392422565a606
Review-Url: https://codereview.chromium.org/2259823005

fix fuzzes

TBR=reed@google.com
BUG=639157, 638783
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255243003

Review-Url: https://codereview.chromium.org/2255243003

SkFontMgr_android clean up.

This removes the never used defaultTypeface, uses 'StyleSet' in field
names consistently, fixes potential leaks of names, and moves fields to
sk_sp.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2256253003

Review-Url: https://codereview.chromium.org/2256253003

include extra_cflags* in the step description

This keeps them visible on the bots.
E.g. https://luci-milo.appspot.com/swarming/task/30b9147a09f2ac10/steps/ninja/0/stdout

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254733008

Review-Url: https://codereview.chromium.org/2254733008

must call rand in reliable order for bots

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254403002

TBR=

Review-Url: https://codereview.chromium.org/2254403002

Expand Issue Tracking docs to include triage, Chrome instructions

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255133003
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=2255133003

Review-Url: https://codereview.chromium.org/2255133003

Revert of Fast path translate() in SkCanvas and SkLiteDL. (patchset #5 id:80001 of https://codereview.chromium.org/2255283002/ )

Reason for revert:
speculative

Original issue's description:
> Fast path translate() in SkCanvas and SkLiteDL.
>
> This adds didTranslate() so that SkLiteDL (and other canvas recorders)
> can record the translate rather than the full concat.
>
> It also adds a case to SkMatrix::preTranslate() to fast path
> translate x translate -> translate (i.e. +=).
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255283002
>
> Committed: https://skia.googlesource.com/skia/+/5fa47f4fd13b3158de4599414c86d17649c2dd1c

TBR=herb@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/2264433002

quick check to not use AutoDrawLooper

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2248373004

Review-Url: https://codereview.chromium.org/2248373004

Revert of Return color shader instead of 2-color gradient when color count is 1 (patchset #3 id:40001 of https://codereview.chromium.org/2259823005/ )

Reason for revert:
breaks 2-pt-conical, which can sometimes not draw anything

Original issue's description:
> Return color shader instead of 2-color gradient when color count is 1
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259823005
>
> Committed: https://skia.googlesource.com/skia/+/da082a5767d7edfd3abe74fc683392422565a606

TBR=fmenozzi@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/2257993002

Detect CRC32 instructions on ARMv8.

I have successfully detected CRC32 instruction support on my Nexus 5x.

Use of these instructions to follow... I am not yet sure which compilers if any will give me instrinsics or let me write them in asm.

defined(__ARM_FEATURE_CRC32) should cover users like Android Framework who build with the best settings possible.  cpu-features.h covers use cases like Clank and our bots.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259133002

Review-Url: https://codereview.chromium.org/2259133002

Batched implementation of drawLattice() for GPU

Bechmarks (Nexus 6P):

Src=100x100, Dst=250x250, NumRects=9
Android              77.7us
Skia (without patch) 57.2us
Skia (with patch)    30.9us

Src=100x100, Dst=500x500, NumRects=9
Android              77.0us
Skia (without patch) 56.9us
Skia (with patch)    31.8us

Src=100x100, Dst=1000x1000, NumRects=9
Android              180us
Skia (without patch) 96.8us
Skia (with patch)    70.5us

Src=100x100, Dst=250x250, NumRects=15
Android              208us
Skia (without patch) 155us
Skia (with patch)    38.2us

Src=100x100, Dst=500x500, NumRects=15
Android              207us
Skia (without patch) 152us
Skia (with patch)    38.4us

Src=100x100, Dst=1000x1000, NumRects=15
Android              233us
Skia (without patch) 156us
Skia (with patch)    99.9us

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255963002

Committed: https://skia.googlesource.com/skia/+/93242c4ae50dfcc0d922cdb3ba80bbc7b4bbe93d
Review-Url: https://codereview.chromium.org/2255963002

Fast path translate() in SkCanvas and SkLiteDL.

This adds didTranslate() so that SkLiteDL (and other canvas recorders)
can record the translate rather than the full concat.

It also adds a case to SkMatrix::preTranslate() to fast path
translate x translate -> translate (i.e. +=).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255283002

Review-Url: https://codereview.chromium.org/2255283002

Relax check for gpu use of static src-over XP

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259113002

Review-Url: https://codereview.chromium.org/2259113002

add simplerect gm

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257263002

TBR=bsalomon
NOTRY=True

Review-Url: https://codereview.chromium.org/2257263002

Return color shader instead of 2-color gradient when color count is 1

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259823005

Review-Url: https://codereview.chromium.org/2259823005

Use CSS3 style matching on Android.

Android framework doesn't really use this, the largest user is Chromium.
At the moment this doesn't resolve oblique requests in a nice way, but
the existing code is somewhat close to CSS3 rules already. Instead of
adding oblique handling manually, just use the existing CSS3 style
matching.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2256843003

Review-Url: https://codereview.chromium.org/2256843003

Revert of Batched implementation of drawLattice() for GPU (patchset #7 id:180001 of https://codereview.chromium.org/2255963002/ )

Reason for revert:
Things drawing weird.

Original issue's description:
> Batched implementation of drawLattice() for GPU
>
> Bechmarks (Nexus 6P):
>
> Src=100x100, Dst=250x250, NumRects=9
> Android              77.7us
> Skia (without patch) 57.2us
> Skia (with patch)    34.7us
>
> Src=100x100, Dst=500x500, NumRects=9
> Android              77.0us
> Skia (without patch) 56.9us
> Skia (with patch)    44.5us
>
> Src=100x100, Dst=1000x1000, NumRects=9
> Android              180us
> Skia (without patch) 96.8us
> Skia (with patch)    70.5us
>
> Src=100x100, Dst=250x250, NumRects=15
> Android              208us
> Skia (without patch) 155us
> Skia (with patch)    55.9us
>
> Src=100x100, Dst=500x500, NumRects=15
> Android              207us
> Skia (without patch) 152us
> Skia (with patch)    63.0us
>
> Src=100x100, Dst=1000x1000, NumRects=15
> Android              233us
> Skia (without patch) 156us
> Skia (with patch)    99.9us
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255963002
>
> Committed: https://skia.googlesource.com/skia/+/93242c4ae50dfcc0d922cdb3ba80bbc7b4bbe93d

TBR=bsalomon@google.com,reed@google.com,djsollen@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/2255683004

SkPDF: in-place font subsetting

Motivation: gross code simplification, also no bitset lookups at draw time.

SkPDFFont owns its glyph useage bitset.

SkPDFSubstituteMap goes away.

SkPDFObject interface is simplified.

SkPDFDocument tracks font usage (as hash set), not glyph usage.

SkPDFFont gets a simpler constructor.

SkPDFFont has first and last glyph set in constructor, not adjusted later.

SkPDFFont implementations are simplified.

SkPDFGlyphSet is replaced with simple SkBitSet.

SkPDFFont sizes its SkBitSets based on glyph count.

SkPDFGlyphSetMap goes away.

SkBitSet is now non-copyable.

SkBitSet now how utility methods to match old SkPDFGlyphSet.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253283004

CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Win-MSVC-GCE-CPU-AVX2-x86_64-Release-GDI-Trybot,Test-Win-MSVC-GCE-CPU-AVX2-x86_64-Debug-GDI-Trybot

Review-Url: https://codereview.chromium.org/2253283004

Fix initialization bug for fConservativeIsScaleTranslate

Also don't mark it as conservative, we expect it to be correct all
the time.

TBR=reed@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2261573002

Review-Url: https://codereview.chromium.org/2261573002

Delay creation of cpu-side buffer memory until actually needed

IIUC what is going on, this won't really do anything bad but will defer allocation of the cpu-side buffer until it is actually needed.

BUG=635015
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2248283007

Review-Url: https://codereview.chromium.org/2248283007

Batched implementation of drawLattice() for GPU

Bechmarks (Nexus 6P):

Src=100x100, Dst=250x250, NumRects=9
Android              77.7us
Skia (without patch) 57.2us
Skia (with patch)    34.7us

Src=100x100, Dst=500x500, NumRects=9
Android              77.0us
Skia (without patch) 56.9us
Skia (with patch)    44.5us

Src=100x100, Dst=1000x1000, NumRects=9
Android              180us
Skia (without patch) 96.8us
Skia (with patch)    70.5us

Src=100x100, Dst=250x250, NumRects=15
Android              208us
Skia (without patch) 155us
Skia (with patch)    55.9us

Src=100x100, Dst=500x500, NumRects=15
Android              207us
Skia (without patch) 152us
Skia (with patch)    63.0us

Src=100x100, Dst=1000x1000, NumRects=15
Android              233us
Skia (without patch) 156us
Skia (with patch)    99.9us

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255963002

Review-Url: https://codereview.chromium.org/2255963002

make LayerIter private, and remove skipClip option

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257203002

Review-Url: https://codereview.chromium.org/2257203002

GN: make libjpeg_turbo target Chrome-compatible

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257903002

Review-Url: https://codereview.chromium.org/2257903002

Tessellator: better fix for reused-edges issue.

The GrTessellator fix for doubly-added edges in
https://codereview.chromium.org/2259493002/ could leave
a MonotonePoly with zero edges. This is a problem for
Poly::addEdge(), which assumes that MonotonePolys always have
at least one edge. The fix is to move the check and early-out up to
Poly::addEdge(). This should also tighten up the vertex count.

(Unfortunately, the only repro I have for this issue is very
convoluted, and requires non-landed code.)

BUG=skia:5636
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2251643008

Review-Url: https://codereview.chromium.org/2251643008

SkPDF:  cache metrics once.

Motivation: drawText can look up unicode mapping at draw time to see
if ActualText should be used after crrev.com/2084533004 lands.

For each SkTypeface, only call getAdvancedTypefaceMetrics() once per
document.  Cache the result in the SkPDFCanon, indexed by SkFontID.
Also cache PDF FontDescriptors in the canon, also indexed by SkFontID
(Type1 fonts only).

Simplify PDF font lookup, map SkFontID+SkGlyphID into a uint64_t.  Map
that uint64_t to SkPDFFonts.  Remove SkPDFCanon::findFont(),
SkPDFCanon::addFont(), SkPDFFont::IsMatch(), and enum SkPDFFont::Match.

SkPDFFont no longer holds on to ref of SkAdvancedTypefaceMetrics.
Instead, SkPDFFont::GetFontResource() and SkPDFFont::getFontSubset()
get metrics from canon.

SkPDFFont multybite bool is now a function of type.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253993002

CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Win-MSVC-GCE-CPU-AVX2-x86_64-Release-GDI-Trybot,Test-Win-MSVC-GCE-CPU-AVX2-x86_64-Debug-GDI-Trybot

Review-Url: https://codereview.chromium.org/2253993002

remove disable aa/dither flags from Props -- never used

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257643003

TBR=bsalomon

Review-Url: https://codereview.chromium.org/2257643003

Interactive Bevel Sample App

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2246243002

Review-Url: https://codereview.chromium.org/2246243002

Don't emit dead code in circle shader when not using distance vectors.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254103004

Review-Url: https://codereview.chromium.org/2254103004

Rename WorseCaseSize to WorstCasSize
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259863002

Review-Url: https://codereview.chromium.org/2259863002

Add alpha type to texture producer

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2250403003

Review-Url: https://codereview.chromium.org/2250403003

Add executable bit and shebang to python files

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253373002

Review-Url: https://codereview.chromium.org/2253373002

Revert of Reduce window rectangles cap to 8 (patchset #1 id:1 of https://codereview.chromium.org/2254013002/ )

Reason for revert:
Breaking Ubuntu

Original issue's description:
> Reduce window rectangles cap to 8
>
> Lowers the cap to 8 and adds a warning message if this value is ever
> exceeded. The largest (only) implementation currently supports 8, so
> there isn't yet reason to go higher.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254013002
>
> Committed: https://skia.googlesource.com/skia/+/52d721580ee22525c285e2d13cf3975a7a1b2843

TBR=mjk@nvidia.com,bsalomon@google.com,csmartdalton@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/2253293003

Revert of SkPDF:  cache metrics once. (patchset #4 id:60001 of https://codereview.chromium.org/2253993002/ )

Reason for revert:
because

Original issue's description:
> SkPDF:  cache metrics once.
>
> Motivation: drawText can look up unicode mapping at draw time to see
> if ActualText should be used after crrev.com/2084533004 lands.
>
> For each SkTypeface, only call getAdvancedTypefaceMetrics() once per
> document.  Cache the result in the SkPDFCanon, indexed by SkFontID.
> Also cache PDF FontDescriptors in the canon, also indexed by SkFontID
> (Type1 fonts only).
>
> Simplify PDF font lookup, map SkFontID+SkGlyphID into a uint64_t.  Map
> that uint64_t to SkPDFFonts.  Remove SkPDFCanon::findFont(),
> SkPDFCanon::addFont(), SkPDFFont::IsMatch(), and enum SkPDFFont::Match.
>
> SkPDFFont no longer holds on to ref of SkAdvancedTypefaceMetrics.
> Instead, SkPDFFont::GetFontResource() and SkPDFFont::getFontSubset()
> get metrics from canon.
>
> SkPDFFont multybite bool is now a function of type.
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253993002
>
> Committed: https://skia.googlesource.com/skia/+/0a61270f4ba85d10659fb63a86817b435ec04c94

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

Review-Url: https://codereview.chromium.org/2251813006

[SVGDom] Fix <ellipse> positioning

R=robertphillips@google.com,stephana@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253283002

Review-Url: https://codereview.chromium.org/2253283002

Fix assert caused by floating point error in tessellating path renderer.

In rare cases, floating point error causes the tesselator to add the
same Vertex to more than one Poly on the same side. This was not a big
problem when we were allocating new vertices when constructing Polys,
but after https://codereview.chromium.org/2029243002 it causes more serious
issues, since each Edge can only belong to two Polys, and violating this
condition messes up the linked list of Edges used for left & right Polys
and the associated estimated vertex count.

The fix is to simply let the first Poly win, and skip that vertex for
subsequent Polys. Since this only occurs in cases where vertices are very
close to each other, it should have little visual effect.

This is also exercised by Nebraska-StateSeal.svg.

BUG=skia:5636
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259493002

Review-Url: https://codereview.chromium.org/2259493002

[SVGDom] Add support for assorted absolute units

R=robertphillips@google.com,stephana@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259473002

Review-Url: https://codereview.chromium.org/2259473002

SkPDF:  cache metrics once.

Motivation: drawText can look up unicode mapping at draw time to see
if ActualText should be used after crrev.com/2084533004 lands.

For each SkTypeface, only call getAdvancedTypefaceMetrics() once per
document.  Cache the result in the SkPDFCanon, indexed by SkFontID.
Also cache PDF FontDescriptors in the canon, also indexed by SkFontID
(Type1 fonts only).

Simplify PDF font lookup, map SkFontID+SkGlyphID into a uint64_t.  Map
that uint64_t to SkPDFFonts.  Remove SkPDFCanon::findFont(),
SkPDFCanon::addFont(), SkPDFFont::IsMatch(), and enum SkPDFFont::Match.

SkPDFFont no longer holds on to ref of SkAdvancedTypefaceMetrics.
Instead, SkPDFFont::GetFontResource() and SkPDFFont::getFontSubset()
get metrics from canon.

SkPDFFont multybite bool is now a function of type.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253993002

Review-Url: https://codereview.chromium.org/2253993002

Add alphaType() to SkImage

Keep isOpaque as a convenience method -- many places really only need to
know that for optimization purposes (SrcOver -> Src, etc...).

In all the places where we pull data back out or convert to another
object and need to supply an SkImageInfo, we can avoid losing information
about premulness.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2250663002

Review-Url: https://codereview.chromium.org/2250663002

Fix BUILD.gn for arch other then x86 and x64

opts_gypi should only be defined for x86 and x64 else we get unused
variable error

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2248323002

Review-Url: https://codereview.chromium.org/2248323002

Fix computation of clipped src rect when tiling bitmaps in SkGpuDevice
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2252913004

Review-Url: https://codereview.chromium.org/2252913004

Reduce window rectangles cap to 8

Lowers the cap to 8 and adds a warning message if this value is ever
exceeded. The largest (only) implementation currently supports 8, so
there isn't yet reason to go higher.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254013002

Review-Url: https://codereview.chromium.org/2254013002

Eagerly update the unichar to glyphID mapping.

In addition, some small cleanups.

BUG=chromium:635005
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2249063005

Review-Url: https://codereview.chromium.org/2249063005

fix no-gpu build

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253073003

TBR=robertphilips
NOTREECHECKS=true

Review-Url: https://codereview.chromium.org/2253073003

Fix fb variable name in GLSL

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2252143002

NOTREECHECKS=True

Review-Url: https://codereview.chromium.org/2252143002

SkLiteDL: a little spring cleaning

  - kill off optimizeFor() now that we know it's broken and have a better way;
  - simplify makeThreadsafe() a bit: we're not calling it anyway.

This sort of earns me back some crazy code currency,
which I am going to try to spend making reset() faster.

If anything, this appears to make makeThreadsafe() _faster_.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254913002

Review-Url: https://codereview.chromium.org/2254913002

Simplify adding attributes to GrGeometryProcessor
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2248323003

Review-Url: https://codereview.chromium.org/2248323003

Fix SampleAndroidShadows warnings

R=jvanverth@google.com,robertphillips@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255863002
NOTRY=true

Review-Url: https://codereview.chromium.org/2255863002

pin as texture api

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2241353002

Review-Url: https://codereview.chromium.org/2241353002

Move FamilyClass and Panose enums to enum class.

The IBMFamilyClass and Panose structures are obvious candidates
for simplification now that it is no longer necessary to nest
enums inside structs to avoid name clashes.

Review-Url: https://codereview.chromium.org/2253543004

Fix various issues with framebuffer fetch

This CL does two things. First it fixes a bug of ours where when using
a custom fbFetch variable (es 3.0 and higher), we sometimes overwrite
the out color and then attempt to use it as the dst color later. This is
fixed here by using an intermediate variable.

Secondly I've added a workaround to an andreno fbFetch where reading from
the outColor always returns the original dstColor even if we've overwritten
the value.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2248403003

Review-Url: https://codereview.chromium.org/2248403003

Detemplatize GrGLSLGeometryProcessor::setTransformDataHelper()

Instead take a local matrix parameter.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2250563004

Review-Url: https://codereview.chromium.org/2250563004

SkLiteDL: inline empty()

1.85%     1.85%  RenderThread     6458  6475  /system/lib64/libskia.so  SkLiteDL::empty() const

:/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2254823002

Review-Url: https://codereview.chromium.org/2254823002

Add alternative spot shadow to Android shadow sample

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2246273003

Review-Url: https://codereview.chromium.org/2256713002

Implement difference clip rects with window rectangles

Plumbs the pipeline for window rectangles and uses them for a very
basic implementation of difference clip rects. This puts a common
Blink pattern on fast path, but we will still eventually need to make
more comprehensive use of window rectangles during clipping.

GTX 960 perf result:
                              gpu                     glinst4                 glinst16
desk_jsfiddlebigcar.skp       0.254 -> 0.177 [70%]    0.279 -> 0.197 [71%]    0.577 -> 0.196 [34%]
keymobi_sfgate_com_.skp       0.697 -> 0.513 [74%]    0.766 -> 0.451 [59%]    0.769 -> 0.597 [78%]
keymobi_blogger.skp           0.406 -> 0.314 [77%]    0.436 -> 0.292 [67%]    0.696 -> 0.319 [46%]
desk_pokemonwiki.skp          0.121 -> 0.098 [81%]     0.13 -> 0.105 [81%]    0.216 -> 0.097 [45%]
desk_wikipedia.skp            0.121 -> 0.098 [81%]     0.13 -> 0.104 [80%]    0.199 -> 0.104 [52%]
keymobi_androidpolice_co...   0.443 -> 0.382 [86%]    0.447 -> 0.398 [89%]    0.444 -> 0.396 [89%]
keymobi_booking_com_sear...   1 .15 ->  1.03 [90%]     1.17 ->  1.06 [91%]     1.17 ->  1.05 [90%]
keymobi_theverge_com.skp      0.417 -> 0.396 [95%]    0.426 -> 0.405 [95%]    0.429 ->   0.4 [93%]

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2251573002

Review-Url: https://codereview.chromium.org/2251573002

Make GrReducedClip's gen ID only apply to the element list

Renames fGenID to fElementsGenID and designates this value as
undefined when when the element list is empty.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2244223004

Review-Url: https://codereview.chromium.org/2244223004

Kick the can down the road a bit w.r.t. fuzzer complaint

BUG=635787
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2247033006

Review-Url: https://codereview.chromium.org/2247033006

Fix tile bitmap code in SkGpuDevice to compute correct local coords.

This fixes a bug when using an alpha bitmap/image with a SkShader where the SkShader's GrFragmentProcessor would receive incorrect coordinates.

BUG=skia:5643

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2249163004

Review-Url: https://codereview.chromium.org/2249163004

Minor cleanup of GP classes in GrOvalRenderer

Move GLSL nested classes to private and remove unnecessary public getters.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2253903002

Review-Url: https://codereview.chromium.org/2253903002

Modify SkPngCodec to recognize 565 images from the sBIT chunk

Conveniently, SkPngImageEncoder already writes the sBIT chunk
appropriately.

BUG=skia:5616
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2212563003

Review-Url: https://codereview.chromium.org/2212563003

[SVGDom] Add <line> support

R=robertphillips@google.com,stephana@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2244223005

Review-Url: https://codereview.chromium.org/2244223005