Revert of Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader (patchset #11 id:200001 of https://codereview.chromium.org/2062703003/ )

Reason for revert:
Causing the Test-Ubuntu-GCC-ShuttleA-GPU-GTX550Ti-x86_64-Release-Valgrind builder to fail.

Started failing at this build:
https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-ShuttleA-GPU-GTX550Ti-x86_64-Release-Valgrind/builds/1214

Snippet from build log:

==3016== Invalid read of size 4
==3016==    at 0xCD1901: GrFragmentProcessor::registerChildProcessor(sk_sp<GrFragmentProcessor>) (SkTArray.h:163)
==3016==    by 0xA58B47: _Z10sk_make_spI11NormalMapFPI5sk_spI19GrFragmentProcessorERK8SkMatrixEES1_IT_EDpOT0_ (SkNormalSource.cpp:84)
==3016==    by 0xA582E8: NormalMapSourceImpl::asFragmentProcessor(GrContext*, SkMatrix const&, SkMatrix const*, SkFilterQuality, SkSourceGammaTreatment) const (SkNormalSource.cpp:187)
==3016==    by 0xA0ADB5: SkLightingShaderImpl::asFragmentProcessor(GrContext*, SkMatrix const&, SkMatrix const*, SkFilterQuality, SkSourceGammaTreatment) const (SkLightingShader.cpp:268)
==3016==    by 0xDA17E6: skpaint_to_grpaint_impl(GrContext*, SkPaint const&, SkMatrix const&, sk_sp<GrFragmentProcessor>*, SkXfermode::Mode*, bool, bool, GrPaint*) (SkGr.cpp:552)
==3016==    by 0xDA4532: SkPaintToGrPaint(GrContext*, SkPaint const&, SkMatrix const&, bool, GrPaint*) (SkGr.cpp:668)
==3016==    by 0xD99B51: SkGpuDevice::drawRect(SkDraw const&, SkRect const&, SkPaint const&) (SkGpuDevice.cpp:534)
==3016==    by 0x9DDB1C: SkCanvas::onDrawRect(SkRect const&, SkPaint const&) (SkCanvas.cpp:2148)
==3016==    by 0x9DBF40: SkCanvas::drawRect(SkRect const&, SkPaint const&) (SkCanvas.cpp:1919)
==3016==    by 0x8D2A55: skiagm::LightingShaderGM::onDraw(SkCanvas*) (lightingshader.cpp:110)
==3016==    by 0x625659: skiagm::GM::drawContent(SkCanvas*) (gm.cpp:32)
==3016==    by 0x6256AD: skiagm::GM::draw(SkCanvas*) (gm.cpp:24)
==3016==    by 0x61B747: DM::GMSrc::draw(SkCanvas*) const (DMSrcSink.cpp:61)
==3016==    by 0x61F37B: DM::GPUSink::draw(DM::Src const&, SkBitmap*, SkWStream*, SkString*) const (DMSrcSink.cpp:1115)
==3016==    by 0x619261: dm_main() (DM.cpp:1016)
==3016==    by 0x619841: main (DM.cpp:1409)
==3016==  Address 0xf4 is not stack'd, malloc'd or (recently) free'd
==3016==

Caught signal 11 [Segmentation fault], was running:
gpu gm  lightingshader

Original issue's description:
> Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader
>
> Will not run until after landing https://codereview.chromium.org/2106893003/
>
> This CL's base is the CL for the addition of NormalSource to the API: https://codereview.chromium.org/2063793002/
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062703003
>
> Committed: https://skia.googlesource.com/skia/+/3f58cd0eb8bf4499c4f179aa6111d7e005809df8

TBR=egdaniel@google.com,robertphillips@google.com,dvonbeck@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:
NOTRY=true

Review-Url: https://codereview.chromium.org/2137513002

Use sRGB pixel config for ARGB masks (color emoji)

Fixes over-bright color emoji and GM:colortype in sRGB mode

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129913004

Committed: https://skia.googlesource.com/skia/+/920af9fac3d9e275a93910d9c626ee7bc23ed17a
Review-Url: https://codereview.chromium.org/2129913004

Reland of Better encapsulate oval/rrect batchs. (patchset #1 id:1 of https://codereview.chromium.org/2121313004/ )

Reason for revert:
Workaround for compiler/stdlib bug landed: https://codereview.chromium.org/2135453002

Original issue's description:
> Revert of Better encapsulate oval/rrect batchs. (patchset #5 id:80001 of https://codereview.chromium.org/2104423003/ )
>
> Reason for revert:
> This CL is suspected to break compilation on arm64.
> https://build.chromium.org/p/chromium.android/builders/Android%20arm64%20Builder%20%28dbg%29/builds/5851/steps/compile/logs/stdio
>
> Original issue's description:
> > Better encapsulate oval/rrect batchs.
> > GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2104423003
> >
> > Committed: https://skia.googlesource.com/skia/+/5fd209e8ee477c703bc5c11b008f247d515fc0fc
> > Committed: https://skia.googlesource.com/skia/+/7f06c6947a3bef78dc57b9252779567c33604c90
> > Committed: https://skia.googlesource.com/skia/+/11bf8b2eae7d1780cb969146422a2ab3b933047a
>
> TBR=robertphillips@google.com,bsalomon@google.com
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://skia.googlesource.com/skia/+/c79fc33d3e83664ffbd3651df3131c3e9e0b635c

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

Review-Url: https://codereview.chromium.org/2134693002

update path->rsxform to correctly interp dy

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2131043002

TBR=

Review-Url: https://codereview.chromium.org/2131043002

Promote the GCC/libc++/<memory> 'typedef float float32_t' workaround to SkTypes.h

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2135453002

Review-Url: https://codereview.chromium.org/2135453002

Make all parameters reqiured to GrReducedClip::ReduceClipStack
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2130903002

Review-Url: https://codereview.chromium.org/2130903002

Revert of Move sRGB <-> linear conversion components to their own files. (patchset #5 id:80001 of https://codereview.chromium.org/2128893002/ )

Reason for revert:
Monotonicity assert is failing on ARM.  (Different rsqrt() and invert() precision?)  Will investigate a bit tomorrow... might reland with the test TODO.

Original issue's description:
> Move sRGB <-> linear conversion components to their own files.
>
> This makes them a little easier to use outside SkColorXform code.
>
> I've added some notes about how best to use them and their eccentricities, and added a test.
>
> Ultimately any software sRGB <-> linear conversion should funnel somehow through here.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128893002
> CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/45e58c8807179638980aae8503573b950b844e4c

TBR=reed@google.com,msarett@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/2131793002

make some canvas utils virtual (for proxy based subclasses)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2130973002

Review-Url: https://codereview.chromium.org/2130973002

fix typos and capitalization on drawTextRSXform CL

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2123923006

Review-Url: https://codereview.chromium.org/2123923006

Update markdown to use absolute image paths, fixing broken images

Markdown can be served with or without a trailing slash but when a
trailing slash is present, relative image links were broken.

The following example works as expected:
url: https://skia.org/dev/design/pdftheory
image url: PdfLogicalDocumentStructure.png
resolves to: https://skia.org/dev/design/PdfLogicalDocumentStructure.png

The following example was broken before this patch:
url: https://skia.org/dev/design/pdftheory/
image url: PdfLogicalDocumentStructure.png
resolves to: https://skia.org/dev/design/pdftheory/PdfLogicalDocumentStructure.png

The PdfLogicalDocumentStructure.png file lives in /dev/design/, not
/dev/design/pdftheory/. This patch updates all images in our markdown
files to be served with the full path so a trailing slash no longer
breaks images.

I hit this bug when promoting the new Skia debugger:
https://groups.google.com/a/chromium.org/d/msg/paint-dev/QJ8xpu8dNyA/F4KX2O4ZBwAJ

BUG=skia:5493
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128013003
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=2128013003

Review-Url: https://codereview.chromium.org/2128013003

Move sRGB <-> linear conversion components to their own files.

This makes them a little easier to use outside SkColorXform code.

I've added some notes about how best to use them and their eccentricities, and added a test.

Ultimately any software sRGB <-> linear conversion should funnel somehow through here.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128893002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review-Url: https://codereview.chromium.org/2128893002

add bounds to textblob

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2124113003

TBR=

Review-Url: https://codereview.chromium.org/2124113003

Revert of Better encapsulate oval/rrect batchs. (patchset #5 id:80001 of https://codereview.chromium.org/2104423003/ )

Reason for revert:
This CL is suspected to break compilation on arm64.
https://build.chromium.org/p/chromium.android/builders/Android%20arm64%20Builder%20%28dbg%29/builds/5851/steps/compile/logs/stdio

Original issue's description:
> Better encapsulate oval/rrect batchs.
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2104423003
>
> Committed: https://skia.googlesource.com/skia/+/5fd209e8ee477c703bc5c11b008f247d515fc0fc
> Committed: https://skia.googlesource.com/skia/+/7f06c6947a3bef78dc57b9252779567c33604c90
> Committed: https://skia.googlesource.com/skia/+/11bf8b2eae7d1780cb969146422a2ab3b933047a

TBR=robertphillips@google.com,bsalomon@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/2121313004

Revert of Use sRGB pixel config for ARGB masks (color emoji) (patchset #1 id:1 of https://codereview.chromium.org/2129913004/ )

Reason for revert:
Color text rendering failing on ANGLE

Original issue's description:
> Use sRGB pixel config for ARGB masks (color emoji)
>
> Fixes over-bright color emoji and GM:colortype in sRGB mode
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129913004
>
> Committed: https://skia.googlesource.com/skia/+/920af9fac3d9e275a93910d9c626ee7bc23ed17a

TBR=bungeman@google.com,jvanverth@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/2128583007

drawTextRSXform

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2130643004

Review-Url: https://codereview.chromium.org/2130643004

SkPDF: Fix encoding of unichr outside of basic plane

In ToUnicode table, write unicode codepoints as one or two UTF16BE
values, rather than a single hex, as the standard requires.

Factor out uint16 -> big-endian hex code.

SkUtils is now a namespace.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2120533002

Review-Url: https://codereview.chromium.org/2120533002

SkPDF: fix PDFA-only metadata error

Review-Url: https://codereview.chromium.org/2119883002

Use sRGB pixel config for ARGB masks (color emoji)

Fixes over-bright color emoji and GM:colortype in sRGB mode

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129913004

Review-Url: https://codereview.chromium.org/2129913004

Check granularity in render pass bounds

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2126303002

Review-Url: https://codereview.chromium.org/2126303002

Add offsets to GrVkBuffer.

BUG=skia:5031
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2127183002

Review-Url: https://codereview.chromium.org/2127183002

Begin instanced rendering for simple shapes

Adds a module that performs instanced rendering and starts using it
for a select subset of draws on Mac GL platforms. The instance
processor can currently handle rects, ovals, round rects, and double
round rects. It can generalize shapes as round rects in order to
improve batching. The instance processor also employs new drawing
algorithms, irrespective of instanced rendering, that improve GPU-side
performance (e.g. sample mask, different triangle layouts, etc.).

This change only scratches the surface of instanced rendering. The
majority of draws still only have one instance. Future work may
include:

 * Passing coord transforms through the texel buffer.
 * Sending FP uniforms through instanced vertex attribs.
 * Using instanced rendering for more draws (stencil writes,
   drawAtlas, etc.).
 * Adding more shapes to the instance processor’s repertoire.
 * Batching draws that have mismatched scissors (analyzing draw
   bounds, inserting clip planes, etc.).
 * Bindless textures.
 * Uber shaders.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2066993003

Committed: https://skia.googlesource.com/skia/+/42eafa4bc00354b132ad114d22ed6b95d8849891
Review-Url: https://codereview.chromium.org/2066993003

Only check resource tracking on program shutdown, not context shutdown

BUG=skia:5496
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128273002

Review-Url: https://codereview.chromium.org/2128273002

Fix vulkan compile

TBR=jvanverth@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2122263003

Review-Url: https://codereview.chromium.org/2122263003

centralize VECTORCALL as SK_VECTORCALL

Gonna start using this, might as well define it once centrally.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2126253002

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

Review-Url: https://codereview.chromium.org/2126253002

Correctly handle mat2x2 in Vulkan

BUG=skia:5497
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129953002

Review-Url: https://codereview.chromium.org/2129953002

Better encapsulate oval/rrect batchs.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2104423003

Committed: https://skia.googlesource.com/skia/+/5fd209e8ee477c703bc5c11b008f247d515fc0fc
Committed: https://skia.googlesource.com/skia/+/7f06c6947a3bef78dc57b9252779567c33604c90
Review-Url: https://codereview.chromium.org/2104423003

Remove fColor from PathRenderer DrawPathArgs struct

This was getting in the way of other refactorings.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2126083002

Review-Url: https://codereview.chromium.org/2126083002

Pull out freelist allocation from GrVkSubHeap

BUG=skia:5031
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2128673002

Review-Url: https://codereview.chromium.org/2128673002

fix for SkSL producing invalid branch after an OpKill
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129913002

Review-Url: https://codereview.chromium.org/2129913002

Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader

Will not run until after landing https://codereview.chromium.org/2106893003/

This CL's base is the CL for the addition of NormalSource to the API: https://codereview.chromium.org/2063793002/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2062703003

Review-Url: https://codereview.chromium.org/2062703003

Hide NonAAStrokeRectBatch Geometry class and cleanup class.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2126893002

Review-Url: https://codereview.chromium.org/2126893002

Fix bounds computation for non-aa hairlines when snapping to pixel centers

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2126913002

Review-Url: https://codereview.chromium.org/2126913002

Deleting void pointers is bad, mmmkay?

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129483003

Review-Url: https://codereview.chromium.org/2129483003

More accurate render time and continuous fresh

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129613002

Review-Url: https://codereview.chromium.org/2129613002

API change to allow for NormalSource selection at the user level.

This CL's base is the CL for CPU handling: https://codereview.chromium.org/2050773002/

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2063793002

Review-Url: https://codereview.chromium.org/2063793002

Hide geometry class in AAStrokeRectBatch and remove unused Append function.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2119813004

Review-Url: https://codereview.chromium.org/2119813004

Vulkan fixes for TesselatingPathRenderer test

- Be sure to release all secondard command buffers on shutdown
- Allow mapping of static buffers
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2123323002

Review-Url: https://codereview.chromium.org/2123323002

check for zero scale in SkDraw::ComputeResScaleForStroking

If the transform used to compute the res scale has both 0 scale and skew,
then it will currently compute a 0 res scale. This causes getFillPath
to trigger an assertion while using SkStroke since that can't handle
a 0 res scale.

Downstream Firefox bug report: https://bugzilla.mozilla.org/show_bug.cgi?id=1284356

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2125933003

Review-Url: https://codereview.chromium.org/2125933003

Grow descriptor pools by 1.5

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2121243002

Review-Url: https://codereview.chromium.org/2121243002

Fixed crash when RunInSeriesFP's color input is nil

SeriesFragmentProcessor would call emitChild with "" input color as opposed to nil input color when ginven a nil input color.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2106893003

Review-Url: https://codereview.chromium.org/2106893003

Add gm for various font rendering combinations.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2080453002

Review-Url: https://codereview.chromium.org/2080453002

Add gm that tests shaded stroked rectangles.

Fix GPU handling of previously untested cases.

Move rect->path fallback from SkGpuDevice to GrDrawContext.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2125663003

Review-Url: https://codereview.chromium.org/2125663003

Update rebaseline instructions.

BUG=skia:2805

NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=2127953002

Review-Url: https://codereview.chromium.org/2127953002

Fix caching of sample locations

The original caching logic for sample locations wishfully assumed that
the GPU would always use the same sample pattern for render targets
that had the same number of samples. It turns out we can't rely on
that. This change improves the caching logic to handle mismatched
simple patterns with the same count, and adds a unit test that
emulates different sample patterns observed on real hardware.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2111423002

Committed: https://skia.googlesource.com/skia/+/09d49a3bfe2d1e652a648ce1ea0962b38d10d166
Review-Url: https://codereview.chromium.org/2111423002

Move GrNonAAFillRectPerspectiveBatch to its own file

Get this out of the way of the upcoming changes (and add a test for it)

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2108403005

Review-Url: https://codereview.chromium.org/2108403005

Check fIgnoresCoverage in GrPipeline::AreEqual

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2120153002

Review-Url: https://codereview.chromium.org/2120153002

Experiment: add flag for finish-recording to return null

BUG=skia:5495
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2106843004

Review-Url: https://codereview.chromium.org/2106843004

Add resource tracking output and command buffer recycling

BUG=skia:5042
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115993002

Review-Url: https://codereview.chromium.org/2115993002

Compile Vulkan when building for the Android framework.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2129523002

Review-Url: https://codereview.chromium.org/2129523002

Fix vulkan image sampling

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2118733005

Review-Url: https://codereview.chromium.org/2118733005

Try fix the mac build bot after removing Android SampleApp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2117793002

Review-Url: https://codereview.chromium.org/2117793002

Increase batching for AA fill rects.

This allows batching of rects provided without a local matrix when local coords are required and when the view matrix changes.

It also allows batching of rects with a local matrix with rects without a local matrix.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116823002

Committed: https://skia.googlesource.com/skia/+/e525ecaf63f225f1da6e9834f7a291c06ad44d23
Review-Url: https://codereview.chromium.org/2116823002

fix for Clang 3.8 warning in SkSL
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2118033002

Review-Url: https://codereview.chromium.org/2118033002

Blacklist fontscalerdistortable on Win8.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2109583003

Review-Url: https://codereview.chromium.org/2109583003

Enlarge strokerect GM size to reveal hidden content
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2126723002

Review-Url: https://codereview.chromium.org/2126723002

Document SkSurface::MakeRaster's memory initialization

So clients don't go clearing w/ SK_ColorTRANSPARENT unnecessarily.

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2066903003

Review-Url: https://codereview.chromium.org/2066903003

Update dev docs to reference the new skia debugger (skiaserve)

This patch adds a new markdown file describing the new debugger along
with instructions on running it locally and a sweet screenshot. The
screenshot has been sized to not be clipped when viewed in production.

The old documentation for the qt-based debugger has been marked as
deprecated and all docs referencing the old debugger have been updated.

BUG=skia:5493
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2121673002
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=2121673002

Review-Url: https://codereview.chromium.org/2121673002

make setScaleTranslate public

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2122093002

TBR=

Review-Url: https://codereview.chromium.org/2122093002

Revert of Begin instanced rendering for simple shapes (patchset #20 id:380001 of https://codereview.chromium.org/2066993003/ )

Reason for revert:
This caused static initializer regressions in Chromium (crbug.com/625728).

Relevant build logs here:
Linux:
https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/21849

Mac:
https://build.chromium.org/p/chromium/builders/Mac/builds/17350

Relevant lines from the error log:
Linux:
# InstanceProcessor.cpp GrUniqueKey::GenerateDomain()
# InstanceProcessor.cpp gr_instanced::kShapeBufferDomain
FAILED linux-release-64/sizes/nacl_helper-si/initializers: actual 8, expected 7, better lower
FAILED linux-release-64/sizes/chrome-si/initializers: actual 8, expected 7, better lower

Mac:
FAILED mac-release/sizes/chrome-si/initializers: actual 2, expected 0, better lower

Original issue's description:
> Begin instanced rendering for simple shapes
>
> Adds a module that performs instanced rendering and starts using it
> for a select subset of draws on Mac GL platforms. The instance
> processor can currently handle rects, ovals, round rects, and double
> round rects. It can generalize shapes as round rects in order to
> improve batching. The instance processor also employs new drawing
> algorithms, irrespective of instanced rendering, that improve GPU-side
> performance (e.g. sample mask, different triangle layouts, etc.).
>
> This change only scratches the surface of instanced rendering. The
> majority of draws still only have one instance. Future work may
> include:
>
>  * Passing coord transforms through the texel buffer.
>  * Sending FP uniforms through instanced vertex attribs.
>  * Using instanced rendering for more draws (stencil writes,
>    drawAtlas, etc.).
>  * Adding more shapes to the instance processor’s repertoire.
>  * Batching draws that have mismatched scissors (analyzing draw
>    bounds, inserting clip planes, etc.).
>  * Bindless textures.
>  * Uber shaders.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2066993003
>
> Committed: https://skia.googlesource.com/skia/+/42eafa4bc00354b132ad114d22ed6b95d8849891

NOTREECHECKS=true
TBR=bsalomon@google.com,egdaniel@google.com,robertphillips@google.com,csmartdalton@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:

Review-Url: https://codereview.chromium.org/2123693002

Revert of Fix caching of sample locations (patchset #3 id:40001 of https://codereview.chromium.org/2111423002/ )

Reason for revert:
Seems to be causing Chromium roll failures:
* https://codereview.chromium.org/2120373003
* https://codereview.chromium.org/2117193002
* https://codereview.chromium.org/2124653002

Original issue's description:
> Fix caching of sample locations
>
> The original caching logic for sample locations wishfully assumed that
> the GPU would always use the same sample pattern for render targets
> that had the same number of samples. It turns out we can't rely on
> that. This change improves the caching logic to handle mismatched
> simple patterns with the same count, and adds a unit test that
> emulates different sample patterns observed on real hardware.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2111423002
>
> Committed: https://skia.googlesource.com/skia/+/09d49a3bfe2d1e652a648ce1ea0962b38d10d166

TBR=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/2120403002

Fix caching of sample locations

The original caching logic for sample locations wishfully assumed that
the GPU would always use the same sample pattern for render targets
that had the same number of samples. It turns out we can't rely on
that. This change improves the caching logic to handle mismatched
simple patterns with the same count, and adds a unit test that
emulates different sample patterns observed on real hardware.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2111423002

Review-Url: https://codereview.chromium.org/2111423002

all dumpcanvas to be built in release

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2123633002

TBR=

Review-Url: https://codereview.chromium.org/2123633002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2119053002

Review-Url: https://codereview.chromium.org/2119053002

Revert of Increase batching for AA fill rects. (patchset #4 id:60001 of https://codereview.chromium.org/2116823002/ )

Reason for revert:
I believe this is causing the Chromium DEPS roll to fail due to linux_blink_rel. Reverting for now to get the roll going again.

Original issue's description:
> Increase batching for AA fill rects.
>
> This allows batching of rects provided without a local matrix when local coords are required and when the view matrix changes.
>
> It also allows batching of rects with a local matrix with rects without a local matrix.
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116823002
>
> Committed: https://skia.googlesource.com/skia/+/e525ecaf63f225f1da6e9834f7a291c06ad44d23

TBR=robertphillips@google.com,bsalomon@google.com
NOTREECHECKS=true
# Not skipping CQ checks because original CL landed more than 1 days ago.

Review-Url: https://codereview.chromium.org/2124603002

Exclude SkSL in public.bzl.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2122463002
TBR=ethannicholas
NOTREECHECKS=true

Review-Url: https://codereview.chromium.org/2122463002

fixed SkSL memory leak
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2119903002
NOTREECHECKS=true

Review-Url: https://codereview.chromium.org/2119903002

fixed non-GPU build
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2120543003
NOTREECHECKS=true

Review-Url: https://codereview.chromium.org/2120543003

Update Win toolchain

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116173002
NOTREECHECKS=true
NOTRY=true
TBR=rmistry@google.com

Review-Url: https://codereview.chromium.org/2116173002

Specify new Windows Toolchain isolated hash.

NOTREECHECKS=true
NOTRY=true
TBR=rmistry@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2121443002

Review-Url: https://codereview.chromium.org/2121443002

Increase batching for AA fill rects.

This allows batching of rects provided without a local matrix when local coords are required and when the view matrix changes.

It also allows batching of rects with a local matrix with rects without a local matrix.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116823002

Review-Url: https://codereview.chromium.org/2116823002

Split apart the GP construction for the NonAAFillRect Batches

This sets up for adding yet another batch type that specializes for rectStaysRect

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2106063005

Review-Url: https://codereview.chromium.org/2106063005

using_skia_and_harfbuzz: use default typeface

Also: add little script to generate utf-8 test text.

BUG=skia:5434
TBR=bungeman@google.com

Review-Url: https://codereview.chromium.org/2118833002

initial checkin of SkSL compiler

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1984363002
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac-Clang-x86_64-Release-CMake-Trybot

Review-Url: https://codereview.chromium.org/1984363002

Use bounds of batches for render pass bounds

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2113303002

Review-Url: https://codereview.chromium.org/2113303002

Fix vulkan ImageView swizzling to match spec

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115903002

Review-Url: https://codereview.chromium.org/2115903002

Cleanup non-AA fill rect batch construction.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2119853002

Review-Url: https://codereview.chromium.org/2119853002

Simplify AA fill rect batch construction and rename Geometry->RectInfo
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116723003

Review-Url: https://codereview.chromium.org/2116723003

Remove unneeded helpers in AA and non-AA rect GrBatch subclasses.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115673002

Review-Url: https://codereview.chromium.org/2115673002

Cleanup PLSPathBatch
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116543002

Review-Url: https://codereview.chromium.org/2116543002

Disable GPU dithering in Android framework builds
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115683002

Review-Url: https://codereview.chromium.org/2115683002

Remove unused batch rect append functions
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2106283003

Review-Url: https://codereview.chromium.org/2106283003

Remove GrTInstanceBatch

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2114663003

Review-Url: https://codereview.chromium.org/2114663003

Fix bug where ovals' AA exceed bounds by .5 pixel

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2106953009

Review-Url: https://codereview.chromium.org/2106953009

Begin instanced rendering for simple shapes

Adds a module that performs instanced rendering and starts using it
for a select subset of draws on Mac GL platforms. The instance
processor can currently handle rects, ovals, round rects, and double
round rects. It can generalize shapes as round rects in order to
improve batching. The instance processor also employs new drawing
algorithms, irrespective of instanced rendering, that improve GPU-side
performance (e.g. sample mask, different triangle layouts, etc.).

This change only scratches the surface of instanced rendering. The
majority of draws still only have one instance. Future work may
include:

 * Passing coord transforms through the texel buffer.
 * Sending FP uniforms through instanced vertex attribs.
 * Using instanced rendering for more draws (stencil writes,
   drawAtlas, etc.).
 * Adding more shapes to the instance processor’s repertoire.
 * Batching draws that have mismatched scissors (analyzing draw
   bounds, inserting clip planes, etc.).
 * Bindless textures.
 * Uber shaders.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2066993003

Review-Url: https://codereview.chromium.org/2066993003

Hide GrNonAANinePatchBatch::Geometry and rename to Patch
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2111943002

Review-Url: https://codereview.chromium.org/2111943002

Hide MSAAPathBatch::Geometry and rename to PathInfo
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116533002

Review-Url: https://codereview.chromium.org/2116533002

SkPDF: Glyph Useage Map improvements

Instead of having a fFontGlyphUsage on each device and one on each
document, just have the one on the document, and never merge.

Make fGlyphUsage accesible on SkPDFDocument.

Remove SkPDFGlyphSetMap::merge, ::reset, and SkPDFGlyphSet::merge.

SkPDFGlyphSetMap has an TArray of SkPDFGlyphSet, not TDArray of
SkPDFGlyphSet pointers.  SkPDFGlyphSet and SkPDFBitset get move
constructors.

All tests produce exactly identical output PDFs.

BUG=skia:5434
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2112943002

Review-Url: https://codereview.chromium.org/2112943002

Remove support for GL discards

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2112693003

Review-Url: https://codereview.chromium.org/2112693003

Fix png encoding in skia debugger

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2119513002

Review-Url: https://codereview.chromium.org/2119513002

Add cubic example and draw original path

- Add cubic example
- Draw original path
- Also added dumpHex to stderr option on 'D'

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2106063004

Review-Url: https://codereview.chromium.org/2106063004

skiaserve: Use snakeCase for JSON identifiers.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2108823006

Review-Url: https://codereview.chromium.org/2108823006

Add static buffer support

Put resource tracking check in the right place

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2112653002

Review-Url: https://codereview.chromium.org/2112653002

Add discard support to Vulkan

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2104353002

Review-Url: https://codereview.chromium.org/2104353002

SkPDF: allow overriding Producer metadata

I recommend not using this functionality.

Also, some documentation.

BUG=skia:5436

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2074583003

Review-Url: https://codereview.chromium.org/2074583003

use mapRectScaleTranslate for common case in clipRect

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2119473002

Review-Url: https://codereview.chromium.org/2119473002

Hide GrDrawVerticesBatch::Geometry and rename to Mesh
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2104383003

Review-Url: https://codereview.chromium.org/2104383003

debugger: Allow for larger images.

Also differentiate between max dimensions and default dimensions.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2113693002

Review-Url: https://codereview.chromium.org/2113693002

inline very common intersect+rect_rgn

Extracted from previous compound CL (with other changes).

This is the very common case for clips, and inlining this eliminates a series of function calls.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2116513002

TBR=mtklein

Review-Url: https://codereview.chromium.org/2116513002

Update skia milestone to next

No API changes, just moving version fwd.

TBR=reed@google.com

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2115573002

Review-Url: https://codereview.chromium.org/2115573002

Even more hiding of Geometry structs in GrBatch subclasses.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2110853004

Review-Url: https://codereview.chromium.org/2110853004