Remove SK_IGNORE_GL_TEXTURE_TARGET from skia_for_chromium_defines.gypi

This has been added to Chrome's SkUserConfig.h

Review URL: https://codereview.chromium.org/1503173003

fix funky formatting in SkNVRefCnt

BUG=skia:
TBR=reed@google.com
No API change.

Review URL: https://codereview.chromium.org/1505023002

When was SkPDiff last used?

BUG=skia:1451,skia:1463,skia:1798,skia:1859,skia:2710,skia:2711,skia:2712,skia:2713

Review URL: https://codereview.chromium.org/1502173003

Make GrAtlasTextBlob non-virtual

BUG=skia:

Review URL: https://codereview.chromium.org/1503213003

Minor code cleanups in SkCanvas.

Since the SK_SAVE_LAYER_BOUNDS_ARE_FILTERED path is long gone from
SkCanvas, remove or localize some temporary variables.

Cleanup only; no user-visible changes.

BUG=skia:

Review URL: https://codereview.chromium.org/1508823002

SkPixelSerializer: support indexed pixels

By taking a SkPixmap, SkPixelSerializer::encode() can now handle colortables.

Review URL: https://codereview.chromium.org/1501303002

add signed-integer-overflow to yasm exceptions

BUG=skia:4635

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

Review URL: https://codereview.chromium.org/1509613002

A small cleanup of GrAtlasTextContext

BUG=skia:

Review URL: https://codereview.chromium.org/1502323002

fix coincident fuzzer

This fuzzer has very large Y values that cause the
points to sort incorrectly by t. Exit out as soon
as this is detected.

TBR=reed@google.com
BUG=561121

Review URL: https://codereview.chromium.org/1507803002

Simplify D1G so that it can inline DrawOneGlyph, and fix a bug in codegen
that only happens on ARM64 using GCC 4.9.

Review URL: https://codereview.chromium.org/1507633004

SkImageShaderFactoryToName SkAlphaThresholdFilterFactoryToName

https://gold.skia.org/diff?test=image-shader&left=8807a80c69a5d565821432fe6a7b74ec&top=80222191bf0768b0fc62c8e05b58fb5f

https://gold.skia.org/diff?test=imagealphathreshold&left=fc3fbbfbd1b1e7ec1c33c00c6c22b9a8&top=493096aac6f44b91cd6522c6049d5a56

BUG=skia:4613

Review URL: https://codereview.chromium.org/1499443002

Allow SkStream::peek() to partially succeed

If the stream can peek less than requested, peek that amount. Return
the number of bytes peeked.

This simplifies crrev.com/1472123002. For a stream that is smaller than
14 bytes, it can successfully peek, meaning the client will not need to
fall back to read() + rewind(), which may fail if the stream can peek
but not rewind.

This CL revives code from patch set 3 of crrev.com/1044953002, where I
initially introduced peek() (including tests).

Add a test for SkFrontBufferedStream that verifies that peeking does
not make rewind() fail (i.e. by reading past the internal buffer).

BUG=skia:3257

Review URL: https://codereview.chromium.org/1490923005

Start objectifying GrAtlasTextBlob

BUG=skia:

Review URL: https://codereview.chromium.org/1503193002

clang/win: Let SK_TRACEHR not produce -Wunused-value warnings in release builds.

Fixes warnings like:
..\..\third_party\skia\include\utils\win\SkHRESULT.h(51,23) :  note: expanded from macro 'HRNM'
                      ^~~~~~~~~~~~~~~~~~~~~~~~~
..\..\third_party\skia\include\utils\win\SkHRESULT.h(26,20) :  note: expanded from macro 'HR_GENERAL'
        SK_TRACEHR(_hr, _msg);\
                   ^~~
..\..\third_party\skia\include\utils\win\SkHRESULT.h(20,31) :  note: expanded from macro 'SK_TRACEHR'
                              ^~~

BUG=chromium:505318
TBR=reed
This is a trivial implementation change.

Review URL: https://codereview.chromium.org/1503463004

Fix up signed-integer-overflow warnings

When checking whether a matrix was a pure scale, we subtracted
0x3f800000 from the diagonals; if the diagonal value was already
very negative, we'd underflow. Replace subtraction with XOR.

When dealing with repeating tiled bitmaps, when the bitmap was
very large, we'd multiply an offset by 65535, possibly causing
underflow. Throw in a cast to long (casting to unsigned also
silences the warning and wouldn't involve extension, but I can't
convince myself that it's correct).

BUG=skia:4635
R=mtklein@google.com

Review URL: https://codereview.chromium.org/1504933002

Add SkTileImageFilter sample to filterfastbounds GM.

NOTE: will affect pixel results for filterfastbounds GM.

BUG=skia:3194

Review URL: https://codereview.chromium.org/1500373004

Always use high precision on NDS transform

To fix the Chrome fillRect issue on Galaxy S6.
We should use high precision for position related calculation.

BUG=chromium:552999

Review URL: https://codereview.chromium.org/1500393002

SkAlphaThresholdFilter.h allow flattening

Motivation:  allows this:

    #include "SkAlphaThresholdFilter.h"
    void init() {
      SkAlphaThresholdFilter::InitializeFlattenables();
    }

BUG=skia:4613

Review URL: https://codereview.chromium.org/1500373003

Pin result in SkATan2_255

BUG=555544

Review URL: https://codereview.chromium.org/1506913002

add gm to exercise large sigmas

BUG=skia:4437
TBR=

Review URL: https://codereview.chromium.org/1503143002

better NEON div255

We were doing (x+127)/255 = ((x+128) + (x+128)>>8)>>8 in three instructions:
    1) x += 128
    2) shift x right 8 bits
    3) add x and x>>8 together, then shift right more 8 bits

Now do it as two instructions:
    1) shift (x+128) right 8 bits
    2) add x and (x+128)>>8 and 128 all together, then shift right 8 more bits

On ARM this will be a 5-10% speedup for SrcATop, DstATop, Xor, Multiply, Difference, HardLight, Darken, and Lighten xfermodes.  When we have a mask (e.g. text), *all* xfermodes except Plus will get a similar boost.

This should mean now that (a*b).div255() is the same speed as a.approxMulDiv255(b) on both x86 and ARM, and of course it's perfect instead of approximate.  So we should eliminate approxMulDiv255(), but I'll leave it to another CL, as it'll need Blink rebaselines.

This CL should not change GMs or Blink.
https://gold.skia.org/search2?issue=1502843002&unt=true&query=source_type%3Dgm&master=false

BUG=skia:
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot;client.skia.android:Test-Android-GCC-Nexus9-CPU-Denver-Arm64-Debug-Trybot,Test-Android-GCC-Nexus5-CPU-NEON-Arm7-Release-Trybot

Review URL: https://codereview.chromium.org/1502843002

Matrix convolution bounds fix; affectsTransparentBlack fixes.

Because the convolution kernel is (currently) applied in device space,
there's no way to know which object-space pixels will be touched. So
return false from canComputeFastBounds().

The results from the matrixconvolution GM were actually wrong, since
they were showing edge differences on the clip boundaries, where they
should really only show on crop boundaries. I added a crop to the GM
to keep the results the same (which are useful to test the different
convolution tile modes).

While I was at it, SkImageFilter::affectsTransparentBlack() was
inapplicable on most things except color filters, and its use on
leaf nodes was confusing. So I removed it, and made
SkImageFilter::canComputeFastBounds() virtual instead.

BUG=skia:4630

Review URL: https://codereview.chromium.org/1500923004

Add transfer buffer to GLCaps

Adds a check for PBO/transfer buffer support to GrGLCaps,
and uses that to pick the correct buffer type.

BUG=skia:4604

Review URL: https://codereview.chromium.org/1503593002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS

Review URL: https://codereview.chromium.org/1502893002

detect when we can filter bitmaps/images directly, w/o a tmp layer

visual bench run on Mac Pro

curr/maxrss loops min median mean max stddev samples    config bench
 100/100 MB 16 412µs 413µs 413µs 414µs 0% ▄▁▇▄▄▄▄█▄▃▅ gpu warmupbench
 101/102 MB 32 547µs 548µs 611µs 1.24ms 34% █▁▁▁▁▁▁▁▁▁▁ gpu image-filter-sprite-draw-image
 102/103 MB 32 547µs 548µs 721µs 1.23ms 41% █▁▇▁▁█▁▁▁▁▁ gpu image-filter-sprite-draw-bitmap
 103/103 MB 64 546µs 546µs 546µs 547µs 0% ▆▄▂▁▇█▅▇▅▇▃ gpu image-filter-sprite-draw-sprite

Should have no effect on Chrome while SK_SUPPORT_LEGACY_LAYER_BITMAP_IMAGEFILTERS is defined (which it is in chrome)

BUG=skia:1073

Review URL: https://codereview.chromium.org/1491293002

Revert of Matrix convolution bounds fix; affectsTransparentBlack fixes. (patchset #4 id:60001 of https://codereview.chromium.org/1500923004/ )

Reason for revert:
Introduced memory leak; pixel changes in Chrome.

Original issue's description:
> Matrix convolution bounds fix; affectsTransparentBlack fixes.
>
> Because the convolution kernel is (currently) applied in device space,
> there's no way to know which object-space pixels will be touched. So
> return false from canComputeFastBounds().
>
> The results from the matrixconvolution GM were actually wrong, since
> they were showing edge differences on the clip boundaries, where they
> should really only show on crop boundaries. I added a crop to the GM
> to keep the results the same (which are useful to test the different
> convolution tile modes).
>
> While I was at it, SkImageFilter::affectsTransparentBlack() was
> inapplicable on most things except color filters, and its use on
> leaf nodes was confusing. So I removed it, and made
> SkImageFilter::canComputeFastBounds() virtual instead.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/8705ec80518ef551994b82ca5ccaeb0241d6adec

TBR=reed@google.com,reed@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1497083005

report back colorfilters in lua

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1498293002

Matrix convolution bounds fix; affectsTransparentBlack fixes.

Because the convolution kernel is (currently) applied in device space,
there's no way to know which object-space pixels will be touched. So
return false from canComputeFastBounds().

The results from the matrixconvolution GM were actually wrong, since
they were showing edge differences on the clip boundaries, where they
should really only show on crop boundaries. I added a crop to the GM
to keep the results the same (which are useful to test the different
convolution tile modes).

While I was at it, SkImageFilter::affectsTransparentBlack() was
inapplicable on most things except color filters, and its use on
leaf nodes was confusing. So I removed it, and made
SkImageFilter::canComputeFastBounds() virtual instead.

BUG=skia:

Review URL: https://codereview.chromium.org/1500923004

clarify diff manually, showing black where they differ

BUG=skia:
TBR=robertphilips@google.com

Review URL: https://codereview.chromium.org/1498243002

use RawIter in hairpath (simplifies)

no impact on nanobench

BUG=skia:

Review URL: https://codereview.chromium.org/1497283002

add hairline with caps gm

TBR=reed@google.com

Review URL: https://codereview.chromium.org/1502613002

Add iOS compile bot to the CQ

BUG=skia:4252

Review URL: https://codereview.chromium.org/1501453005

Create a skeleton VisualDebugModule

BUG=skia:

Review URL: https://codereview.chromium.org/1498043002

fix mozilla bug

TBR=bsalomon@google.com
BUG=skia:4621

Review URL: https://codereview.chromium.org/1500023002

Fix overflow caught by ASAN.

BUG=skia:

Review URL: https://codereview.chromium.org/1498923002

Make SkAndroidCodec support ico

BUG=skia:

Review URL: https://codereview.chromium.org/1472933002

Speculative fix for win8 bot crash in ~GrAutoLocaleSetter

Review URL: https://codereview.chromium.org/1493913008

Revert of Make SkAndroidCodec support ico (patchset #7 id:130002 of https://codereview.chromium.org/1472933002/ )

Reason for revert:
Crashing: https://uberchromegw.corp.google.com/i/client.skia.android/builders/Test-Android-GCC-NexusPlayer-CPU-SSE4-x86-Release/builds/1499/steps/dm/logs/stdio

Also, related ASAN failures:
https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-ASAN/builds/3676/steps/dm/logs/stdio

Original issue's description:
> Make SkAndroidCodec support ico
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/1603e9310f62cf0dd543cdb09dea06aa78373f13

TBR=djsollen@google.com,msarett@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1498903004

Add android_skp_capture.py, JSON file for geniewidget

The script runs through the directory of JSON files, capturing SKPs from
the apps specified by the JSON files.

BUG=skia:

Review URL: https://codereview.chromium.org/1483063006

Reduce calls sites where both the GrDrawContext & GrRenderTarget are both passed

TBR=bsalomon@google.com

Review URL: https://codereview.chromium.org/1494083003

simplify nanobench --quiet / -q output

This tries to cut things down to the very minimum you might want to know:

$ out/Release/nanobench --match nytimes --config 8888 gpu -q
Timer overhead: 29.6ns
! -> high variance, ? -> moderate variance
    micros       bench
   2479.05 !    desk_nytimes.skp_1_mpd    8888
   1313.92      desk_nytimes.skp_1_mpd    gpu
   3617.65      desk_nytimes.skp_1    8888
   1158.34      desk_nytimes.skp_1    gpu
   1368.99 !    keymobi_nytimes_com_.skp_1_mpd    8888
    393.40      keymobi_nytimes_com_.skp_1_mpd    gpu
   1179.68 !    keymobi_nytimes_com_.skp_1    8888
    342.74      keymobi_nytimes_com_.skp_1    gpu

All times are printed in microseconds, and high variance runs are marked.

BUG=skia:

Review URL: https://codereview.chromium.org/1493313003

Make SkAndroidCodec support ico

BUG=skia:

Review URL: https://codereview.chromium.org/1472933002

Load ANGLE EGL library correctly on Linux

Assign dynamically loaded ANGLE EGL library to correct variable.

BUG=skia:

Review URL: https://codereview.chromium.org/1494963002

Skip dm GPU configs when context creation fails

Skip dm GPU configs when context creation fails instead of stopping
the whole dm run.

Review URL: https://codereview.chromium.org/1497713002

Add option to draw wireframe batch bounds

Committed: https://skia.googlesource.com/skia/+/26489ef21ff5df33b8cb5943fddfd4604e203960

Review URL: https://codereview.chromium.org/1494473005

add bench for sprite-like drawing w/ imagefilters

visualbench run on Mac Pro

curr/maxrss loops min median mean max stddev samples    config bench
  96/96  MB 16 412µs 413µs 413µs 414µs 0% ▄▁█▅▅▅▅▆▅▃▅ gpu warmupbench
  98/98  MB 4 1.86ms 1.86ms 1.86ms 1.86ms 0% ▁▂▇█▆▅▇▄▃▃▃ gpu image-filter-sprite-draw-image
 100/100 MB 4 1.86ms 1.86ms 1.86ms 1.86ms 0% ▆▇▅▄▁▆▅█▅▄▅ gpu image-filter-sprite-draw-bitmap
 100/100 MB 32 547µs 548µs 590µs 1.01ms 24% █▁▁▁▁▁▁▁▁▁▁ gpu image-filter-sprite-draw-sprite

BUG=skia:

Review URL: https://codereview.chromium.org/1497843003

Add currScanline() getter to SkCodec API

This is more correct than using nextScanline() for the
SkGifCodec scanline decoder (since we will get a strange
result in the interlaced case) and is necessary if we want
to add scanline decoding to SkIcoCodec.

This does not actually fix bugs or change behavior.

BUG=skia:

Review URL: https://codereview.chromium.org/1489163002

Update documentation for SkFontMgr methods.

The appropriate way to get the default system font is to pass |nullptr| as the
familyName. This was not clear from the interface, which may be why ui/gfx/ has
the wrong behavior.

BUG=chromium:564266

Review URL: https://codereview.chromium.org/1494253002

Revert of Add option to draw wireframe batch bounds (patchset #6 id:100001 of https://codereview.chromium.org/1494473005/ )

Reason for revert:
Looks to be responsible for breaking builds e.g. https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Debug/builds/3030/steps/dm/logs/stdio

Original issue's description:
> Add option to draw wireframe batch bounds
>
> Committed: https://skia.googlesource.com/skia/+/26489ef21ff5df33b8cb5943fddfd4604e203960

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

Review URL: https://codereview.chromium.org/1492633007

Stop disabling fixed test

skbug.com/1282 has been fixed, so there is no need to skip these
images

BUG=skia:1282

Review URL: https://codereview.chromium.org/1391143007

Match libjpeg-turbo name change in Android

BUG=skia:

Review URL: https://codereview.chromium.org/1464423003

Add option to draw wireframe batch bounds

Review URL: https://codereview.chromium.org/1494473005

Make building with clang documentation a bit easier to find

BUG=skia:
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1495913002

Review URL: https://codereview.chromium.org/1495913002

Create GLSLUniformHandler class for gpu backend

BUG=skia:

Review URL: https://codereview.chromium.org/1490283004

Add transfer buffer support.

BUG=skia:4604

Committed: https://skia.googlesource.com/skia/+/fa498fe12239988578465d0dee69f0d5645bb361

Review URL: https://codereview.chromium.org/1490473003

Don't use the Sk4f gradient impl without SIMD

Also remove the SK_SUPPORT_LEGACY_LINEAR_GRADIENT_TABLE guard since it is no
longer used in Chromium.

BUG=chromium:563492
R=reed@google.com,mtklein@google.com
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1489233005

Revert of Add transfer buffer support. (patchset #9 id:160001 of https://codereview.chromium.org/1490473003/ )

Reason for revert:
speculative revert for deps roll failures

Original issue's description:
> Add transfer buffer support.
>
> BUG=skia:4604
>
> Committed: https://skia.googlesource.com/skia/+/fa498fe12239988578465d0dee69f0d5645bb361

TBR=bsalomon@google.com,jvanverth@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4604

Review URL: https://codereview.chromium.org/1496843003

Fix ODR violation.

src/ports/SkFontMgr_fontconfig.cpp and src/fonts/SkFontMgr_fontconfig.cpp both define SkFontMgr_fontconfig.

BUG=skia:
TBR=bungeman

Review URL: https://codereview.chromium.org/1491943004

API to support native scaling by image-generator

BUG=skia:

Review URL: https://codereview.chromium.org/1396323007

Wire up resource cache stats dumping in nanobench

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1495543003

Wire up new define in buildbot_spec

BUG=skia:

Review URL: https://codereview.chromium.org/1492133002

reset shader compilations for stats

BUG=skia:

Review URL: https://codereview.chromium.org/1491823003

Add transfer buffer support.

BUG=skia:4604

Review URL: https://codereview.chromium.org/1490473003

Create a define for nanobench builds to dump stats

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1499453002

drawSprite and drawImage+clip should draw the same w/ an imagefilter applied. This gm exercises that invariant.

BUG=skia:

Review URL: https://codereview.chromium.org/1498523002

Revert of skstd -> std for unique_ptr (patchset #17 id:320001 of https://codereview.chromium.org/1436033003/ )

Reason for revert:
Chromium bots can't handle forward declared types, e.g.

struct Foo {
   int x,y;
   std::unique_ptr<Foo> next;
};

https://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_asan_rel_ng/builds/86497

Original issue's description:
> skstd -> std for unique_ptr
>
> TBR=reed@google.com
> No public API changes.
>
> BUG=skia:4564
>
> Committed: https://skia.googlesource.com/skia/+/755c553c17b82bb5de3d9cc8d3b2a866ff9e9e50
>
> CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot,Build-Ubuntu-GCC-x86_64-Debug-CrOS_Link-Trybot;client.skia:Perf-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release-Trybot,Test-iOS-Clang-iPad4-GPU-SGX554-Arm7-Release-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/06189155d987db5c7e69015f6ea87c2168d6a065
>
> Committed: https://skia.googlesource.com/skia/+/70e8dfca4a7f5bce97b8021a6e378c4828b09c8c
>
> Committed: https://skia.googlesource.com/skia/+/dadfc245cc9a0279ff7b73da3344f2ca5d139907

TBR=bungeman@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4564

Review URL: https://codereview.chromium.org/1498583002

Add gpuStatsDump to nanobench GPU bots

BUG=skia:

Review URL: https://codereview.chromium.org/1497533002

skstd -> std for unique_ptr

TBR=reed@google.com
No public API changes.

BUG=skia:4564

Committed: https://skia.googlesource.com/skia/+/755c553c17b82bb5de3d9cc8d3b2a866ff9e9e50

CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot,Build-Ubuntu-GCC-x86_64-Debug-CrOS_Link-Trybot;client.skia:Perf-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release-Trybot,Test-iOS-Clang-iPad4-GPU-SGX554-Arm7-Release-Trybot

Committed: https://skia.googlesource.com/skia/+/06189155d987db5c7e69015f6ea87c2168d6a065

Committed: https://skia.googlesource.com/skia/+/70e8dfca4a7f5bce97b8021a6e378c4828b09c8c

Review URL: https://codereview.chromium.org/1436033003

Reduce use of GrRenderTarget in GrTextContexts

Cleanup prior to switching over to GrRenderTargetProxy.

Review URL: https://codereview.chromium.org/1485973004

Modify iOSShell-gyp to generate unique bundle identifier

BUG=skia:

Review URL: https://codereview.chromium.org/1483143002

Began logging more gpu stats from nanobench

BUG=skia:

Review URL: https://codereview.chromium.org/1489033004

Remove plumbing of clipRect thoughout the TextContexts

The RenderTarget was only being passed around to be able to recreate the clipRect.

Review URL: https://codereview.chromium.org/1486323002

Allow SkAndroidCodec to use SkPngChunkReader

Also update documentation of SkPngChunkReader
in SkCodec.

BUG=skia:

Review URL: https://codereview.chromium.org/1487583003

Revert of Simplify draw one glyph (patchset #3 id:40001 of https://codereview.chromium.org/1403573002/ )

Reason for revert:
This seems to break nexus 9 release.

Original issue's description:
> Replace D1G with a simpler implementation.
>
> Committed: https://skia.googlesource.com/skia/+/001e74426672e00f3f2783ccf728031662d4a358

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

Review URL: https://codereview.chromium.org/1486723004

Add in shader blending for CoverageSetOp XP

BUG=skia:

Review URL: https://codereview.chromium.org/1488213002

Retract GrRenderTarget a bit

These sites don't necessarily need the full power of a GrRenderTarget object. This is a clean up for switching over to GrRenderTargetProxys.

Review URL: https://codereview.chromium.org/1486923004

Update CMakeLists.txt to exclude SkFontConfigInterface_direct_*.

These files are not expected to be used by common users.

Fix Google3 fonts. Use fontconfig rather than custom_directory_factory. Add Google3 font caching.

Some future dependents require these changes.

Depends on internal cl/108287941.

Review URL: https://codereview.chromium.org/1471033002

Replace D1G with a simpler implementation.

Review URL: https://codereview.chromium.org/1403573002

Add an SkPath conic conversion utility.

Expose SkConic::chopIntoQuadsPOW2() as SkPath::ConvertConicToQuads().

BUG=chromium:315277
R=reed@google.com

Review URL: https://codereview.chromium.org/1484373002

CMake, include/:  changes to make fiddle 2.0 better

cmake: remove unused directories from skia.h / ${public_includes}

SkPreConfig.h:  work around buggy `#ifdef linux`

include:  guard many platform-specific headers
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot

Review URL: https://codereview.chromium.org/1488813002

Revert of Fix Android framework build (patchset #1 id:1 of https://codereview.chromium.org/1491563003/ )

Reason for revert:
eeebdb538d476c1bfc8b63a946094ca1b505ecd1 was reverted, so this one needs to be reverted, too

Original issue's description:
> Fix Android framework build
>
> eeebdb538d476c1bfc8b63a946094ca1b505ecd1 did not update a Config that
> only builds if SK_BUILD_FOR_ANDROID_FRAMEWORK. This appears to be the
> necessary parameter that is missing.
>
> BUG=skia:2992
>
> Committed: https://skia.googlesource.com/skia/+/cb6cb21cd9d27054810d2c80ef534dcddd615d4b

TBR=bsalomon@google.com,tomhudson@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:2992

Review URL: https://codereview.chromium.org/1487033002

Revert of Make NVPR a GL context option instead of a GL context (patchset #2 id:20001 of https://codereview.chromium.org/1448883002/ )

Reason for revert:
BUG=skia:4609

skbug.com/4609

Seems like GrContextFactory needs to fail when the NVPR option is requested but the driver version isn't sufficiently high.

Original issue's description:
> Make NVPR a GL context option instead of a GL context
>
> Make NVPR a GL context option instead of a GL context.
> This may enable NVPR to be run with command buffer
> interface.
>
> No functionality change in DM or nanobench. NVPR can
> only be run with normal GL APIs.
>
> BUG=skia:2992
>
> Committed: https://skia.googlesource.com/skia/+/eeebdb538d476c1bfc8b63a946094ca1b505ecd1

TBR=mtklein@google.com,jvanverth@google.com,kkinnunen@nvidia.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:2992

Review URL: https://codereview.chromium.org/1486153002

Fix Android framework build

eeebdb538d476c1bfc8b63a946094ca1b505ecd1 did not update a Config that
only builds if SK_BUILD_FOR_ANDROID_FRAMEWORK. This appears to be the
necessary parameter that is missing.

BUG=skia:2992

Review URL: https://codereview.chromium.org/1491563003

check that SSE2 is available before using __vectorcall

BUG=skia:4563

Review URL: https://codereview.chromium.org/1489873002

Add Sk4f::ToBytes(uint8_t[16], Sk4f, Sk4f, Sk4f, Sk4f)

This is a big speedup for float -> byte.  E.g. gradient_linear_clamp_3color:
 x86-64 147µs -> 103µs  (Broadwell MBP)
 arm64 2.03ms -> 648µs  (Galaxy S6)
 armv7 1.12ms -> 489µs  (Galaxy S6, same device!)

BUG=skia:
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot;client.skia.android:Test-Android-GCC-Nexus9-CPU-Denver-Arm64-Debug-Trybot

Review URL: https://codereview.chromium.org/1483953002

Make NVPR a GL context option instead of a GL context

Make NVPR a GL context option instead of a GL context.
This may enable NVPR to be run with command buffer
interface.

No functionality change in DM or nanobench. NVPR can
only be run with normal GL APIs.

BUG=skia:2992

Review URL: https://codereview.chromium.org/1448883002

Fix stroking of zero length paths with end caps on NVPR

Fix stroking of zero length paths with end caps on NVPR.
In case of such paths, stroke them using Skia and just
fill the path with NVPR.

BUG=skia:4427

Review URL: https://codereview.chromium.org/1471763002

Generate list of GPU contexts outside tests

Use DEF_GPUTEST_FOR_*_CONTEXT macros to obtain the
test GPU context.

Makes changing the context -related classes easier,
since not all tests need to be changed.

BUG=skia:2992

Review URL: https://codereview.chromium.org/1448873002

Disable using dev bounds in Gr clip reduction code.

Enabling this broke GPU raster in Chrome. Disabling until bug diagnosed/fixed.

TBR=joshualitt@googole.com

Review URL: https://codereview.chromium.org/1484973002 .

add more conservative check for wayward divide

When the parallel stroke to the curve can't be computed from
the intersection of the tangent lines, as straight line connects
the two points instead. Allow the intersection to succeed unless
the ratio isn't finite or the contribution of (1 - ratio) isn't
significant.

R=reed@google.com,fmalita@chromium.org
BUG=skia:4603

Review URL: https://codereview.chromium.org/1484873003

Add debug option to clip each GrBatch to its device bounds

Review URL: https://codereview.chromium.org/1471083002

CMake: generate skia.h and skia_{compile|link}_arguments.txt

Motivation: for use by fiddle.

Also, add new files to .gitignore
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot

Review URL: https://codereview.chromium.org/1484053002

Make onPrepareDraws const

BUG=skia:

Review URL: https://codereview.chromium.org/1483103003

CMake: generate SkUserConfig.hh

CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot

Review URL: https://codereview.chromium.org/1483953004

Create a static instances of SrcOver XferProcessor

BUG=skia:

Review URL: https://codereview.chromium.org/1471293003

Add documentation for Status bot filters

BUG=skia:4602
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1489583002

Review URL: https://codereview.chromium.org/1489583002

Fix bug in reads fragment position on gpu.

BUG=skia:

Review URL: https://codereview.chromium.org/1482223004

APIs which took colorPOI / coveragePOI pairs updated to take a GrPipelineOptimizations struct

TBR=bsalomon@google.com

Review URL: https://codereview.chromium.org/1480353002

Revert of skstd -> std for unique_ptr (patchset #16 id:300001 of https://codereview.chromium.org/1436033003/ )

Reason for revert:
Test-iOS-Clang-iPad4-GPU-SGX554-Arm7-Release

Original issue's description:
> skstd -> std for unique_ptr
>
> TBR=reed@google.com
> No public API changes.
>
> BUG=skia:4564
>
> Committed: https://skia.googlesource.com/skia/+/755c553c17b82bb5de3d9cc8d3b2a866ff9e9e50
>
> CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot,Build-Ubuntu-GCC-x86_64-Debug-CrOS_Link-Trybot;client.skia:Perf-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/06189155d987db5c7e69015f6ea87c2168d6a065
>
> Committed: https://skia.googlesource.com/skia/+/70e8dfca4a7f5bce97b8021a6e378c4828b09c8c

TBR=bungeman@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4564

Review URL: https://codereview.chromium.org/1482343002