Fold alpha to the draw in savelayer-draw-restore patterns with non-opaque draw

Fold alpha of a save layer call to the subsequent draw call paint, even when
the draw call paint color is already non-opaque.

Comparing the difference of the unoptimized and the optimized call pattern
with all (layer alpha, draw alpha) combinations, this produces
off-by-one pixels ~50% of the time.

Reduces layers of current desk_carsvg.skp (recorded with cross-process
picture image filters allowed) from 122 to 115.

BUG=skia:3119

Review URL: https://codereview.chromium.org/840483005

add gm to show miplevels (patchset #2 id:20001 of https://codereview.chromium.org/849173004/)"

fixed memory leak

This reverts commit 04f07fcbfddbe8ca97d29053554a01c66d44ece2.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/854163002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/855003003

Revert of Add temporary blacklist debugging. (patchset #1 id:1 of https://codereview.chromium.org/805643004/)

Reason for revert:
fixed!

Original issue's description:
> Add temporary blacklist debugging.
>
> Not sure why blacklists don't seem to work on bots.  They look right.
>
> NOTREECHECKS=true
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/f77156e1925cd5d5504070fe5533a4c456732c1b

TBR=mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/851903003

Revert of add gm to show miplevels (patchset #2 id:20001 of https://codereview.chromium.org/849173004/)

Reason for revert:
==31530==ERROR: LeakSanitizer: detected memory leaks

Direct leak of 224 byte(s) in 2 object(s) allocated from:
    #0 0x7febb223f65b in operator new(unsigned long) (/home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/dm+0x35865b)
    #1 0x7febb3483c56 in SkMipMap::Build(SkBitmap const&, SkDiscardableMemory* (*)(unsigned long)) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/core/SkMipMap.cpp:181:9
    #2 0x7febb2c37914 in ShowMipLevels::onDraw(SkCanvas*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../gm/downsamplebitmap.cpp:228:24
    #3 0x7febb2295736 in skiagm::GM::drawContent(SkCanvas*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../gm/gm.cpp:32:5
    #4 0x7febb2295019 in skiagm::GM::draw(SkCanvas*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../gm/gm.cpp:24:5
    #5 0x7febb226722f in DM::GMSrc::draw(SkCanvas*) const /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../dm/DMSrcSink.cpp:22:5
    #6 0x7febb226d929 in DM::RasterSink::draw(DM::Src const&, SkBitmap*, SkWStream*) const /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../dm/DMSrcSink.cpp:230:12
    #7 0x7febb224e004 in Task::Run(Task*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../dm/DM.cpp:254:25
    #8 0x7febb3805e4c in (anonymous namespace)::ThreadPool::Loop(void*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/core/SkTaskGroup.cpp:149:13
    #9 0x7febb460dc7c in thread_start(void*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/utils/SkThreadUtils_pthread.cpp:66:9
    #10 0x7febb1251f6d in start_thread /build/buildd/eglibc-2.17/nptl/pthread_create.c:311

Indirect leak of 699528 byte(s) in 2 object(s) allocated from:
    #0 0x7febb22214ab in __interceptor_malloc (/home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/dm+0x33a4ab)
    #1 0x7febb473e3c6 in sk_malloc_flags(unsigned long, unsigned int) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/ports/SkMemory_malloc.cpp:54:15
    #2 0x7febb473e19a in sk_malloc_throw(unsigned long) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/ports/SkMemory_malloc.cpp:40:12
    #3 0x7febb3483c94 in SkMipMap::Build(SkBitmap const&, SkDiscardableMemory* (*)(unsigned long)) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/core/SkMipMap.cpp:181:18
    #4 0x7febb2c37914 in ShowMipLevels::onDraw(SkCanvas*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../gm/downsamplebitmap.cpp:228:24
    #5 0x7febb2295736 in skiagm::GM::drawContent(SkCanvas*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../gm/gm.cpp:32:5
    #6 0x7febb2295019 in skiagm::GM::draw(SkCanvas*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../gm/gm.cpp:24:5
    #7 0x7febb226722f in DM::GMSrc::draw(SkCanvas*) const /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../dm/DMSrcSink.cpp:22:5
    #8 0x7febb226d929 in DM::RasterSink::draw(DM::Src const&, SkBitmap*, SkWStream*) const /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../dm/DMSrcSink.cpp:230:12
    #9 0x7febb224e004 in Task::Run(Task*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../dm/DM.cpp:254:25
    #10 0x7febb3805e4c in (anonymous namespace)::ThreadPool::Loop(void*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/core/SkTaskGroup.cpp:149:13
    #11 0x7febb460dc7c in thread_start(void*) /home/default/storage/skia-repo/buildbot/skiabot-linux-xsan-001/build/slave/Test-Ubuntu13_10-GCE-NoGPU-x86_64-Debug-ASAN/build/skia/out/Debug/../../src/utils/SkThreadUtils_pthread.cpp:66:9
    #12 0x7febb1251f6d in start_thread /build/buildd/eglibc-2.17/nptl/pthread_create.c:311

Original issue's description:
> add gm to show miplevels
>
> BUG=skia:
> TBR=
>
> NOTREECHECKS=true
>
> Committed: https://skia.googlesource.com/skia/+/a598f4b773f066a939e0216a116c179b90550727

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

Review URL: https://codereview.chromium.org/854103004

force a build with fixed --blacklist

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/850413003

add gm to show miplevels

BUG=skia:
TBR=

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/849173004

Add temporary blacklist debugging.

Not sure why blacklists don't seem to work on bots.  They look right.

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/805643004

newlines after failures

NOTREECHECKS=true

BUG=skia:3255

Review URL: https://codereview.chromium.org/853123003

Totally serialize gpu tests on Windows.

NOTREECHECKS=true

BUG=skia:3255

Review URL: https://codereview.chromium.org/845193006

Turn back on image and subset sources now that they're blacklisted on Android.

To land after https://codereview.chromium.org/850403002/

NOTREECHECKS=true

BUG=skia:3255

Review URL: https://codereview.chromium.org/857483003

DM: add basic --blacklist <config> <srcType> <name> functionality.

NOTREECHECKS=true

BUG=skia:3255

Review URL: https://codereview.chromium.org/817573005

remove suppressions (CL was reverted)

BUG=skia:
TBR=
NOTREECHECKS=True
NOTRY=True

Review URL: https://codereview.chromium.org/837633007

turn back on gpu tests

NOTREECHECKS=true

BUG=skia:3255

Review URL: https://codereview.chromium.org/817573004

Revert "choose better mip level"

This reverts commit 09544fcf9611be6efc0c38ab9d9a3f82181341fb.

BUG=skia:
TBR=
NOTREECHECKS=True

Review URL: https://codereview.chromium.org/849173003

Remove obsolete mac_gpu

BUG=chromium:426897
R=kbr@google.com,phajdan@google.com,reed@google.com

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/812963003

Sketch DM refactor.

BUG=skia:3255

I think this supports everything DM used to, but has completely refactored how
it works to fit the design in the bug.

Configs like "tiles-gpu" are automatically wired up.

I wouldn't suggest looking at this as a diff.  There's just a bunch of deleted
files, a few new files, and one new file that shares a name with a deleted file
(DM.cpp).

NOTREECHECKS=true

Committed: https://skia.googlesource.com/skia/+/709d2c3e5062c5b57f91273bfc11a751f5b2bb88

Review URL: https://codereview.chromium.org/788243008

Move XferEffects class to GrCustomXfermode file

BUG=skia:

Review URL: https://codereview.chromium.org/844913003

SkPictureShader should handle negative scaling gracefully.

Let's not choke on inverted tiles.

BUG=chromium:447707
R=robertphillips@google.com,reed@google.com

Review URL: https://codereview.chromium.org/852213002

Revert of Sketch DM refactor. (patchset #45 id:850001 of https://codereview.chromium.org/788243008/)

Reason for revert:
plenty of data

Original issue's description:
> Sketch DM refactor.
>
> BUG=skia:3255
>
>
> I think this supports everything DM used to, but has completely refactored how
> it works to fit the design in the bug.
>
> Configs like "tiles-gpu" are automatically wired up.
>
> I wouldn't suggest looking at this as a diff.  There's just a bunch of deleted
> files, a few new files, and one new file that shares a name with a deleted file
> (DM.cpp).
>
> NOTREECHECKS=true
>
> Committed: https://skia.googlesource.com/skia/+/709d2c3e5062c5b57f91273bfc11a751f5b2bb88

TBR=bsalomon@google.com,mtklein@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:3255

Review URL: https://codereview.chromium.org/853883004

more GMs to rebase?

TBR=reed@google.com

BUG=skia:

Review URL: https://codereview.chromium.org/789933003

Sketch DM refactor.

BUG=skia:3255

I think this supports everything DM used to, but has completely refactored how
it works to fit the design in the bug.

Configs like "tiles-gpu" are automatically wired up.

I wouldn't suggest looking at this as a diff.  There's just a bunch of deleted
files, a few new files, and one new file that shares a name with a deleted file
(DM.cpp).

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/788243008

these GM names are ridiculous

TBR=reed@google.com

BUG=skia:

Review URL: https://codereview.chromium.org/852193003

Why are we afraid of saveLayer() bounds?

They're only hints.  We can ignore them in this optimization.
My only hesitation is that this pattern doesn't seem to happen in our .skps.

BUG=skia:

Review URL: https://codereview.chromium.org/849243002

choose better mip level

BUG=skia:

Review URL: https://codereview.chromium.org/845793005

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/845873005

Add include before SK_SUPPORT_GPU in SkArithmeticMode_gpu.h to fix android build

BUG=skia:

Review URL: https://codereview.chromium.org/852193002

Use device-space stroke width for SkDraw::drawRect() quick-reject

The stroke width needs to be CTM-adjusted when applied to device space
rects.

BUG=skia:3313
R=reed@google.com

Review URL: https://codereview.chromium.org/801353008

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/852163002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/855473003

Rebaseline 6 GMs after crrev.com/841763005

BUG=skia:3299
TBR=djsollen@google.com

Review URL: https://codereview.chromium.org/852053004

a few ignores for nvprmsaa4

TBR=
NOTRY=True
BUG=skia:

Review URL: https://codereview.chromium.org/850143003

update images after texture cache change

Review URL: https://codereview.chromium.org/852913002

fMaxTexCoords is unused in Release builds

BUG=skia:

Review URL: https://codereview.chromium.org/853643005

possible bug fix for nvpr

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/850953003

add sample for filter quality

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/788163007

Add Arithmetic mode xp.

BUG=skia:

Review URL: https://codereview.chromium.org/837633005

separate out the scaline-choosing to simplify chooseProcs

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/796063007

fix link reference in docs

BUG=skia:
NOTRIES=1
TBR=halcanary

Review URL: https://codereview.chromium.org/853703003

Add reference material for the style of markdown we use.

Preview here: http://skiadocs.com:8000/dev/tools/markdown?cl=853493004

Note that the image won't show up, that's a known issue with the way patching in CLs works.

BUG=skia:

Review URL: https://codereview.chromium.org/853493004

Move Gpu ArithmeticMode xfer effect to _gpu files.

BUG=skia:

Review URL: https://codereview.chromium.org/827163004

Make uncached textures uncached from the get go.

This avoids the problem of a newly created uncached texture causing a purge of cached resources.

BUG=chromium:445885

Review URL: https://codereview.chromium.org/846303002

Add repro GM for  GPU clipped-AA vs. non-AA drawRect discrepancy

In the clip stack we were manually rounding out non-AA clip rects but leaving the hardening of non-AA drawRects up to the GPU. In some border cases the GPU can truncate rather than round out resulting in visual discrepancies.

BUG=423834

Committed: https://skia.googlesource.com/skia/+/933a03fecb65c83f81cf65d5cf9870c69aa379ff

Review URL: https://codereview.chromium.org/839883003

DM can see into src/gpu.  No need for relative include.

BUG=skia:

Review URL: https://codereview.chromium.org/853613004

Remove self-assignment in a test.

BUG=skia:

Review URL: https://codereview.chromium.org/853723002

More changes to bring together path / geo procs

BUG=skia:

Review URL: https://codereview.chromium.org/820783005

Switch to a more complete method of filtering hoisted layers

Committed: https://skia.googlesource.com/skia/+/74668a8e818fd891a08c4966cc5973b04c1ad992

Review URL: https://codereview.chromium.org/842323003

Do some minor pre cleanup work before converting all xfermodes to XPs.

BUG=skia:

Review URL: https://codereview.chromium.org/853543003

upstream hooking into google3 crash handler for better stacktraces

BUG=skia:

Review URL: https://codereview.chromium.org/821643006

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/812283007

upstream google3 include paths for jsoncpp

BUG=skia:

No public API changes
TBR=reed@google.com

Review URL: https://codereview.chromium.org/853613002

Remove incremental draw optimization from SkDebugCanvas

The SkDebugCanvas can be (or is currently) being used to draw to multiple
different canvases. If this use-case is intended, then storing draw
-related state in the canvas causes bugs.

Remove incremental draw from SkDebugCanvas. It can only optimize the
case where draw command index is advanced, no other changes are done
and no visualization is used. This case is not that critical, as it
happens only once per new frame at most. It causes bugs, because
one SkDebugCanvas is used to draw to multiple canvases.

Leave the draw to canvas in same state as it was passed in.

This fixes the debugger bugs where:
 * Old, stale image would stay on the raster canvas when a new image is
 loaded. Also happened with resizes.

 * Proper image for overdraw filter would be visible only for the first
   frame

Review URL: https://codereview.chromium.org/844493003

debugger: SW rasterize the picture only when it changes

Rasterize the picture only when it changes, not every time
raster widget paints itself.

Removes include SkForceLinking.h while organizing the SkDebuggerGUI.cpp
includes.

This is part of the work trying to remove bugs that come from
SkDebugCanvas stateful draw. Part of the state comes from some
optimization that this tries to reproduce in a different way.

Review URL: https://codereview.chromium.org/839743003

delete the dreaded TransitionView

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/850843002

rebase + ignores

NOTRY=True
TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/841873007

rebase of expectations

TBR=
NOTRY=True
BUG=skia:

Review URL: https://codereview.chromium.org/847313002

fix for winvdows nvpr

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/804023005

Move most of the transform logic into the primitive processors

BUG=skia:

Review URL: https://codereview.chromium.org/822423004

rename blitrow::proc and add (uncalled) hook for colorproc16

BUG=skia:3302

Review URL: https://codereview.chromium.org/847443003

Allow -DGOOGLE3 to bypass our normal platform intrinsic dispatch.

The current code that pound-includes another define doesn't seem to be working there.

BUG=skia:

No API changes.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/820643005

fix for FB fetch on nexus 10 ES3.0

BUG=skia:3304

Review URL: https://codereview.chromium.org/807143006

Add pageset for ugamsolutions.com

BUG=skia:3306
NOTRY=true

Review URL: https://codereview.chromium.org/851743004

Revert of Switch to a more complete method of filtering hoisted layers (patchset #1 id:1 of https://codereview.chromium.org/842323003/)

Reason for revert:
nanobench woes

Original issue's description:
> Switch to a more complete method of filtering hoisted layers
>
> Committed: https://skia.googlesource.com/skia/+/74668a8e818fd891a08c4966cc5973b04c1ad992

TBR=bsalomon@google.com
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/851773002

allow simple

BUG=skia:

Review URL: https://codereview.chromium.org/824753006

Switch to a more complete method of filtering hoisted layers

Review URL: https://codereview.chromium.org/842323003

namespace {} trick for SK_DECLARE_STATIC_ONCE

Like all our other SK_DECLARE_STATIC_*, it's usually not a thread-safe
thing to put inside a function.  Adding namespace {} prevents that
syntactically.

Needs https://codereview.chromium.org/841263004/ to land first.

BUG=chromium:447890

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

Review URL: https://codereview.chromium.org/806473006

Hide ARM assembly memset symbols.

Prevent these symbols from being exported from any library which
includes skia by marking them .hidden, as they are implementation
details of skia.

BUG=skia:3303

Review URL: https://codereview.chromium.org/831663004

Simplify SkInstCnt

This code requires fewer macros to use it (just one), has less code in macro
definitions, and has simpler synchronization code (just atomic ints, no SkOnce,
no SkMutex, etc.)

A minor downside, we lose indentation and reverse-ordering in the final report:
  Leaked SkRefCntBase: 7
     Leaked SkFontMgr: 1
     Leaked SkWeakRefCnt: 1
         Leaked SkTypeface: 1
     Leaked SkFlattenable: 3
         Leaked SkXfermode: 3
     Leaked SkPathRef: 1
     Leaked SkPixelRef: 1
         Leaked SkMallocPixelRef: 1
becomes
  Leaked SkXfermode: 3
  Leaked SkMallocPixelRef: 1
  Leaked SkPixelRef: 1
  Leaked SkPathRef: 1
  Leaked SkFlattenable: 3
  Leaked SkTypeface: 1
  Leaked SkWeakRefCnt: 1
  Leaked SkFontMgr: 1
  Leaked SkRefCntBase: 7

This is motivated by wanting to land https://codereview.chromium.org/806473006/,
which makes sure all static use of SkOnce are in global scope.  The current
implementation of SkInstCnt uses them in function scope, which isn't safe.
BUG=skia:

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

Review URL: https://codereview.chromium.org/841263004

Generic PDF shader fallback

Instead of ignoring unsupported shaders (and essentially filling with
solid black), convert them to bitmap shaders using on-the-fly
rasterization.

BUG=skia:3299
R=reed@google.com,halcanary@google.com

Review URL: https://codereview.chromium.org/841763005

In SkPDFDocument::emitPDF(), stop pre-calculating file offsets.

* Add Streamer utility class which measures the current
  pdf offset by calling SkWStream::bytesWritten(). Calls
  SkPDFCatalog::setFileOffset() and SkPDFObject::emit() at
  the same time to guarantee that everything works out.

* SkPDFCatalog::setFileOffset() no longer calculates the
  object's size.

* SkPDFCatalog::setSubstituteResourcesOffsets() removed.

* SkPDFCatalog::emitSubstituteResources() removed and
  getSubstituteList() made public in its place.

* Remove SkPDFPage::getPageSize and SkPDFPage::emitPage.
  Replace with SkPDFPage::getContentStream().

* SkPDFObject::getOutputSize no longer virtual, only used in
  unit tests.  All SkPDFObject subclasses getOutputSize()
  overrides removed.

* SkPDFObject::getIndirectOutputSize removed.

* PDFPrimitivesTest updated for new functions.

Review URL: https://codereview.chromium.org/846023003

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/819403005

remove NewTexture?

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/831323005

Revert of Fix GPU clipped-AA vs. non-AA drawRect discrepancy (patchset #2 id:20001 of https://codereview.chromium.org/839883003/)

Reason for revert:
This CL introduces rendering conflicts with hairlines (i.e., the hairlines get overwritten). These conflicts are particularly visible on the following GMs (for the Ubuntu and Android gpu configs):

coloremoji & complexclip2_rrect_bw

Original issue's description:
> Fix GPU clipped-AA vs. non-AA drawRect discrepancy
>
> In the clip stack we were manually rounding out non-AA clip rects but leaving the hardening of non-AA drawRects up to the GPU. In some border cases the GPU can truncate rather than round out resulting in visual discrepancies.
>
> BUG=423834
>
> Committed: https://skia.googlesource.com/skia/+/933a03fecb65c83f81cf65d5cf9870c69aa379ff

TBR=bsalomon@google.com,jvanverth@google.com
NOTREECHECKS=true
NOTRY=true
BUG=423834

Review URL: https://codereview.chromium.org/847033002

Fix GPU clipped-AA vs. non-AA drawRect discrepancy

In the clip stack we were manually rounding out non-AA clip rects but leaving the hardening of non-AA drawRects up to the GPU. In some border cases the GPU can truncate rather than round out resulting in visual discrepancies.

BUG=423834

Review URL: https://codereview.chromium.org/839883003

SIMPLE_GM generates less code

Review URL: https://codereview.chromium.org/811703005

Change function signature of SkPDFObject::emitObject.

Replace virutal SkPDFObject::emitObject(s, c, true) with non-virtual
SkPDFObject::emitIndirectObject(s, c), since none of the subclasses do
(or should do) anything different.

Replace object->emitObject(s, c, false) with object->emitObject(s, c)

This is one step in simplifying the SkPDFObject interface to allow for
the next step in refactoring.

Review URL: https://codereview.chromium.org/827733004

Fix overlap test for 64-bit pointers in SkMatrix

BUG=skia:3297

Review URL: https://codereview.chromium.org/803493008

Remove SK_SUPPORT_LEGACY_GRADIENT_PRECISION

No longer used by Chromium.

BUG=447870
R=reed@google.com

Review URL: https://codereview.chromium.org/824063004

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/846843002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/828243004

add stroke params to c api

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/848553002

add --rasterPDF flag to DM

If no rasterizer is compiled in, this flag does nothing.  Default
value (true) gives the same behavior as before.

Review URL: https://codereview.chromium.org/830333005

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/845093003

remove (dead) SK_SUPPORT_LEGACY_DRAWDATA and SK_SUPPORT_LEGACY_CANVAS_VIRTUAL

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/844053002

Document some skia perf interesting queries.

Preview here: http://skiadocs.com:8000/dev/tools/skiaperf?cl=839303002

BUG=skia:

Review URL: https://codereview.chromium.org/839303002

Fix up all the easy virtual ... SK_OVERRIDE cases.

This fixes every case where virtual and SK_OVERRIDE were on the same line,
which should be the bulk of cases.  We'll have to manually clean up the rest
over time unless I level up in regexes.

for f in (find . -type f); perl -p -i -e 's/virtual (.*)SK_OVERRIDE/\1SK_OVERRIDE/g' $f; end

BUG=skia:

Review URL: https://codereview.chromium.org/806653007

Can't include SkTypes.h from ports/

BUG=skia:

Review URL: https://codereview.chromium.org/847533002

Use error-checking mutexes in Debug builds.

Upside:   our deadlock detection is now thread-safe.
Downside: static mutexes on Macs won't have deadlock detection until we drop 10.6.

BUG=skia:3294

Review URL: https://codereview.chromium.org/816833005

rebaseline after conic change

BUG=skia:
TBR=
NOTRY=True

Review URL: https://codereview.chromium.org/840333002

PictureNestingBench: stay in ints.

BUG=skia:

Review URL: https://codereview.chromium.org/784173004

Upstream custom freetype embolding strength from Android

BUG=skia:2377

Review URL: https://codereview.chromium.org/839243002

Remove SkTileGrid (except for TileGridInfo).

TBR=reed@google.com

BUG=skia:3085

Review URL: https://codereview.chromium.org/845623002

rebaseline NexusPlayer after conic fix

BUG=skia:
TBR=
NOTRY=True

Review URL: https://codereview.chromium.org/843063003

remove unneeded flags

BUG=skia:

Review URL: https://codereview.chromium.org/842203002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/843933002

Add a favicon for docs.

BUG=skia:

Review URL: https://codereview.chromium.org/845593002

rebaseline after crrev.com/840653002

NOTRY=true
TBR=

Review URL: https://codereview.chromium.org/838183003

Modify transparency_check GM to always tile correctly.

Review URL: https://codereview.chromium.org/842013002

Add Linux quick start docs

Preview: http://skiadocs.com:8000/user/quick/linux?cl=846523002

BUG=skia:

Review URL: https://codereview.chromium.org/846523002

rebaseline after change to GM:c_gms

BUG=skia:
TBR=
NOTRY=True

Review URL: https://codereview.chromium.org/843843002