Move SkPixelXorXfermode.h to include/client/android

This intended to make clients aware that the SkPixelXorXfermode is Android-specific

This needs to land after the Chromium CL https://codereview.chromium.org/1647953006/ (Add include/client/android to skia .gn & .gypi files)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1645093002

Committed: https://skia.googlesource.com/skia/+/aa9cb62901075e9d2858353cb041e5e6d4719614

Review URL: https://codereview.chromium.org/1645093002

Fix S32_D32_constX_shaderproc assert

After https://crrev.com/1642273002, fFilterOneY includes a rounding bias
which is not accounted for in the debug-mode S32_D32_constX_shaderproc
check.

Update the S32_D32_constX_shaderproc assert to use the (biased)
SkBitmapProcStateAutoMapper utility.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1655823002
R=reed@google.com
BUG=skia:4874

Review URL: https://codereview.chromium.org/1655823002

batch merge new edges

Find where newly introduced edges go in the edge list once,
then stitch all of them into the edge list.

R=reed@google.com
BUG=573166
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1647223002

Committed: https://skia.googlesource.com/skia/+/ae658e15477df86d1a864feb48d0274af2784f40

Review URL: https://codereview.chromium.org/1647223002

Suppress SkBitmapProcState assert temporarily

TBR=jcgregorio@google.com,reed@google.com
BUG=skia:4874
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1658573002

Review URL: https://codereview.chromium.org/1658573002

Only define NO_POSIX_MEMALIGN for arm

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1647893003

Review URL: https://codereview.chromium.org/1647893003

Remove SkDEPRECATED_InstallDiscardablePixelRef from SkPictureShader

Use an SkImage generator/shader instead.

R=reed@google.com

Committed: https://skia.googlesource.com/skia/+/fa694ad067976a5b713eec0b6f92ba5efeb0aa9d

Review URL: https://codereview.chromium.org/1405103003

SkNx Load/store: take any pointer.

This means we can remove a lot of explicit casts in code that uses SkNx.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1650653002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1650653002

Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1651803002

Review URL: https://codereview.chromium.org/1651803002

Remove SkBitmapProcStateAutoMapper's overflow check

(follow-up to https://codereview.chromium.org/1642273002)

Add an optional SkPoint outparam, and relocate the overflow check to
the only client which needs it.

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1650403002

Review URL: https://codereview.chromium.org/1650403002

float components in xfermodes

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1623483002

TBR=mtklein

Review URL: https://codereview.chromium.org/1634273002

combine vertical overlapping edges

Paths outside clips, and sometimes paths inside clips, devolve
to multiple adjacent or overlapping vertical edges. Combine
these edges when possible to reduce the overall edge count.

R=reed@google.com
BUG=573166
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1654433002

Review URL: https://codereview.chromium.org/1654433002

Revert of batch merge new edges (patchset #1 id:1 of https://codereview.chromium.org/1647223002/ )

Reason for revert:
Breaks the roll.

Original issue's description:
> batch merge new edges
>
> Find where newly introduced edges go in the edge list once,
> then stitch all of them into the edge list.
>
> R=reed@google.com
> BUG=573166
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1647223002
>
> Committed: https://skia.googlesource.com/skia/+/ae658e15477df86d1a864feb48d0274af2784f40

TBR=reed@google.com,caryclark@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=573166

Review URL: https://codereview.chromium.org/1654593002

add new tests

These tests are for upcoming changes to optimize
the path edge list.

TBR=reed@google.com
BUG=573166
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1651573002

Review URL: https://codereview.chromium.org/1651573002

Add sampler bias for the nofilter/translate specializations

Convert SkBitmapProcState::setupForTranslate() to use
SkBitmapProcStateAutoMapper.  This adds bias for the translate procs:
Clamp_S32_D32_nofilter_trans_shaderproc,
Repeat_S32_D32_nofilter_trans_shaderproc

Since the original impl checks for int overflow, extend
SkBitmapProcStateAutoMapper to detect this condition.

BUG=chromium:581870
R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1642273002

Review URL: https://codereview.chromium.org/1642273002

Use proper row step in SkRawCodec
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1643373002

Review URL: https://codereview.chromium.org/1643373002

Add SkImage factory method that forces image to be resolved to a texture.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1631053003

Review URL: https://codereview.chromium.org/1631053003

allow the caller to specified raster-surface rowbytes.
along the way, simplify how we copy the surface's bitmap

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1643873002

Review URL: https://codereview.chromium.org/1643873002

batch merge new edges

Find where newly introduced edges go in the edge list once,
then stitch all of them into the edge list.

R=reed@google.com
BUG=573166
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1647223002

Review URL: https://codereview.chromium.org/1647223002

added support for PLS path rendering

BUG=skia:3555
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1541903002

Committed: https://skia.googlesource.com/skia/+/7df3f5e127f8016d17b637cc48a6a4718f1a6822

Review URL: https://codereview.chromium.org/1541903002

Revert of Treat bad values passed to --images as a fatal error (patchset #17 id:320001 of https://codereview.chromium.org/1611323004/ )

Reason for revert:
Speculative to fix windows bots

Original issue's description:
> Treat bad values passed to --images as a fatal error
>
> If an option is passed to --images that is either a non-existent path or
> a folder with no images matching the supported types, assume this is
> an error and exit, so they can supply a valid path instead.
>
> Share code between DM and nanobench in SkCommonFlags.
>
> nanobench now behaves more like DM - it will check a directory for
> images that match the supported extensions.
>
> Only consider image paths ending in RAW suffixes as images if
> SK_CODE_DECODES_RAW is defined. This prevents us from seeing failure
> to decode errors on platforms that cannot decode it.
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1611323004
>
> Committed: https://skia.googlesource.com/skia/+/7579786f3bd5a8fda84a1abc45b16213c3371f93

TBR=mtklein@google.com,borenet@google.com,msarett@google.com

NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
# Not skipping CQ checks because original CL landed more than 1 days ago.

Review URL: https://codereview.chromium.org/1653543002

Revert of Update SK_IMAGE_VERSION to test RAW (patchset #1 id:1 of https://codereview.chromium.org/1612113002/ )

Reason for revert:
https://skia.googlesource.com/skia/+/7579786f3bd5a8fda84a1abc45b16213c3371f93 seems to have caused problems, but it allows us to test RAW. In order to reland this, we'll need to reland that one, or have a simpler fix to disable RAW on platforms where it's not supported (e.g. ChromeOS)

Original issue's description:
> Update SK_IMAGE_VERSION to test RAW
>
> Once https://codereview.chromium.org/1520403003/ lands, we can start
> testing RAW images. I have already added RAW images to a folder named
> v2/; this updates the version so the script (once written) will check
> the new version.
>
> Depends on skbug.com/4558
>
> BUG=skia:4829
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1612113002
>
> Committed: https://skia.googlesource.com/skia/+/317a45e879889f9ed876549a10bbfd51aab41b45

TBR=borenet@google.com,msarett@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.

# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4829

Review URL: https://codereview.chromium.org/1651603004

fix teeny dashed path bug

If the path dashed is sufficiently small, there may be no
segments generated to dash. Check for an empty segment list.

R=reed@google.com
BUG=skia:4871
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1645613006

Review URL: https://codereview.chromium.org/1645613006

Consolidate SK_CRASH and sk_throw into SK_ABORT

TBR=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1648343003

Review URL: https://codereview.chromium.org/1648343003

Move SkAvoidXfermode over from Android

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1649503002

Review URL: https://codereview.chromium.org/1649503002

fix for skiaserve build error
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1651443002

TBR=jcgregorio@google.com

Review URL: https://codereview.chromium.org/1651443002

Print a message when DM finishes

When running from the command line, it is obvious when DM finishes,
since it brings the prompt back up. But in XCode, it just sits
there, so my glance doesn't let me know that it's finished. Make
it more obvious that DM is done.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1650113002

Review URL: https://codereview.chromium.org/1650113002

added support for more features in JSON (blurs, dashing, different path fill types, etc.)
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1644903003

Review URL: https://codereview.chromium.org/1644903003

Revert of Move SkPixelXorXfermode.h to include/client/android (patchset #5 id:80001 of https://codereview.chromium.org/1645093002/ )

Reason for revert:
GN woes

Original issue's description:
> Move SkPixelXorXfermode.h to include/client/android
>
> This intended to make clients aware that the SkPixelXorXfermode is Android-specific
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1645093002
>
> Committed: https://skia.googlesource.com/skia/+/aa9cb62901075e9d2858353cb041e5e6d4719614

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

Review URL: https://codereview.chromium.org/1644903004

starter procs for blending with pm4f

curr/maxrss loops min median mean max stddev samples    config bench
   8/8   MB 4 87.1µs 91µs 89.8µs 92µs 2% ▇▇▇▇█▇▅▁▁▁ nonrendering xfer4f_srcover_N_opaque_linear
   9/9   MB 2 196µs 196µs 215µs 383µs 27% ▁▁▁▁█▁▁▁▁▁ nonrendering xfer4f_srcover_N_opaque_srgb
   9/9   MB 1 313µs 313µs 313µs 313µs 0% ▁▄▅▅▅▂████ nonrendering xfer4f_srcover_N_alpha_linear
   9/9   MB 1 580µs 580µs 582µs 602µs 1% ▁▁▁▁▁▁▂▁▁█ nonrendering xfer4f_srcover_N_alpha_srgb
   9/9   MB 23 13.1µs 13.1µs 13.1µs 13.1µs 0% ▆▄▄█▂▂▂▁▂▁ nonrendering xfer4f_srcover_1_opaque_linear
   9/9   MB 23 13.2µs 13.2µs 13.2µs 13.2µs 0% █▄▂▁▃▁▂▂▂▂ nonrendering xfer4f_srcover_1_opaque_srgb
   9/9   MB 2 178µs 183µs 183µs 185µs 1% ▇▇▇█▇▇▇▇▇▁ nonrendering xfer4f_srcover_1_alpha_linear
   9/9   MB 1 517µs 517µs 517µs 517µs 0% ▇█▄▃▄▁▂▁▂▄ nonrendering xfer4f_srcover_1_alpha_srgb

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1642703003

TBR=
landing now so these incremental types/functions can be used to collaborate with herb's work. nothing is active at this point

Review URL: https://codereview.chromium.org/1642703003

Move SkPixelXorXfermode.h to include/client/android

This intended to make clients aware that the SkPixelXorXfermode is Android-specific

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1645093002

Review URL: https://codereview.chromium.org/1645093002

Fix broken Android framework builds
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641173003

Review URL: https://codereview.chromium.org/1641173003

SaveFlags begone!

This is now dead code (post https://codereview.chromium.org/1636103002).

R=reed@google.com,robertphillips@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1648773002

Review URL: https://codereview.chromium.org/1648773002

Fix typo on trooper page.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1642183002
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1642183002

Review URL: https://codereview.chromium.org/1642183002

Work around exasperating Google3 Android build error.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1649493005

Review URL: https://codereview.chromium.org/1649493005

Update SK_IMAGE_VERSION to test RAW

Once https://codereview.chromium.org/1520403003/ lands, we can start
testing RAW images. I have already added RAW images to a folder named
v2/; this updates the version so the script (once written) will check
the new version.

Depends on skbug.com/4558

BUG=skia:4829
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1612113002

Review URL: https://codereview.chromium.org/1612113002

SkPDF: roll back 53b1c09 and 05b48e2

See:
https://crrev.com/1552263003
https://crrev.com/1438503002

BUG=skia:257
BUG=skia:734
  BUG=577794

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1645803003

Review URL: https://codereview.chromium.org/1645803003

tools/git-sync-deps: fix error in git syntax for changing origin

Error introduced in https://crrev.com/1475513004

NOTRY=true

Review URL: https://codereview.chromium.org/1646863002

Fix SkPixelXorXfermode serialization bug

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1647963002

Review URL: https://codereview.chromium.org/1647963002

Treat bad values passed to --images as a fatal error

If an option is passed to --images that is either a non-existent path or
a folder with no images matching the supported types, assume this is
an error and exit, so they can supply a valid path instead.

Share code between DM and nanobench in SkCommonFlags.

nanobench now behaves more like DM - it will check a directory for
images that match the supported extensions.

Only consider image paths ending in RAW suffixes as images if
SK_CODE_DECODES_RAW is defined. This prevents us from seeing failure
to decode errors on platforms that cannot decode it.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1611323004

Review URL: https://codereview.chromium.org/1611323004

Image filters: Make a recursive, forward-mapping bounds
traversal which respects the CropRect. This is useful when
you want the device-space bounds of a primitive after
filtering. (This may also eventually subsume
computeFastBounds()).

This CL generalizes filterBounds() and onFilterBounds() to
take a mapping direction. It also makes filterBounds()
responsible for calling onFilterNodeBounds() and applying
the crop rect, simplifying onFilterBounds().

BUG=skia:4627
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1612953004

Review URL: https://codereview.chromium.org/1612953004

Disable verbose mode for DNG SDK on mac/ios.

The issue was, that the qDNGReportErrors was not disabled in xcode_settings.

With that also removed the complexety of the dng_sdk.gyp and unified the flags handling.

BUG=26821211
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641553004

Review URL: https://codereview.chromium.org/1641553004

Revert "Enable GL_CHROMIUM_bind_uniform_location, command buffer is fixed"

This reverts commit 1dc85cda5e990f379b891bb7b0356f796045e7f6.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1645013002

TBR=

Review URL: https://codereview.chromium.org/1645013002

Make SkPixelXorXfermode's opColor be SkPMColor

Xoring an SkColor with 2 SkPMColors creates rendering inconsistencies on Macs and some Android devices in 8888.

AFAICT Android doesn't compensate for this so we may be changing SkPixelXorXfermode's behavior on Android.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1646453003

Review URL: https://codereview.chromium.org/1646453003

Add a background timing thread to kilobench

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1612513002

Review URL: https://codereview.chromium.org/1612513002

skiaserve: Clean up flags.

Remove unused flags.
Add --port and --source.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1643963002

Review URL: https://codereview.chromium.org/1643963002

Adding support for parsing extended gpu config parameters in nanobench

BUG=skia:2992
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1603063005

Review URL: https://codereview.chromium.org/1603063005

skiaserve: Add CORS header, and debugger code is served over HTTPS.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1646683002

Review URL: https://codereview.chromium.org/1646683002

Remove SK_SUPPORT_LEGACY_SAVEFLAGS define on Android

Android is the only platform still using SK_SUPPORT_LEGACY_SAVEFLAGS
(indirectly, due to SK_SUPPORT_LEGACY_CLIPTOLAYERFLAG).

At this point it should be safe to remove the define.

R=djsollen@google.com,scroggo@google.com,reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1636103002

Review URL: https://codereview.chromium.org/1636103002

SkPngFilters: guard with libpng >= 1.5.7

The read_filter array was added in libpng-1.5.7.
No point to these filter methods if we can't plug them into libpng.

TBR=msarett@google.com
NOTRY=true
NOPRESUBMIT=true

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1640933004

Review URL: https://codereview.chromium.org/1640933004

SkPngFilters: use normal pointer types.

TBR=msarett@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641793004

Review URL: https://codereview.chromium.org/1641793004

kill SkValue

This is clearly not what we're going to do.

TBR=reed@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1643753002

Review URL: https://codereview.chromium.org/1643753002

sketch hooking into PNG_FILTER_OPTIMIZATIONS

Local timing says this 4-byte Paeth function takes about 0.3x the time the serial libpng code does, dropping from ~10 cycles per byte to ~2.9.

bpp=4 is mainly an easy demo.  This approach can work for any bpp up to 16, 1 pixel at a time, at roughly the same cost per pixel.  Doing more than 1 pixel at a time is a tricky math problem I have yet to attempt to solve.

Everything here can be trivially downgraded to MMX, supporting bpp up to 8.  It seems to be a little slower (~3.5 cycles per byte), but it would make the code compatible with every x86 that can still power on.

I've tried four approaches:
  - this way;
  - doing things naively in 16-bit;
  - a 16-bit version that requires division by 3 (i.e. mulhi_epu16(..., 0x5580) );
  - a mostly 8-bit version of the same.

They're all fine, but this one is consistently the fastest I've measured.
I'd be happy to settle on the naive 16-bit version too, which would have a very clear implementation that's only minorly slower than this version.  The other two are way more complicated, and would require us to draw some serious ASCII diagrams to explain.

I have learned that the .skp serialization tests (serialize-8888) have a nice side effect of testing the correctness of these filters!

(Since writing the description above, I've bumped things up to {Paeth,Sub,Avg} x { 3 bpp, 4 bpp }.)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1573943002

Review URL: https://codereview.chromium.org/1573943002

skiaserve: Enable /cmd endpoint.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1639403002

Review URL: https://codereview.chromium.org/1639403002

Revert of Consolidate SkStream copying methods (patchset #1 id:1 of https://codereview.chromium.org/1640793002/ )

Reason for revert:
Test to see if this fixes the bots.

Original issue's description:
> Consolidate SkStream copying methods
>
> Make SkCopyStreamToData call SkStreamCopy, removing duplicate code.
>
> The former still has its own method of copying with a length, since
> it saves one copy.
>
> BUG=skia:4788
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1640793002
>
> Committed: https://skia.googlesource.com/skia/+/440c5a98dee428c661b77d149e30c794d264b8cd

TBR=halcanary@google.com,scroggo@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4788

Review URL: https://codereview.chromium.org/1641853002

Suppress jpeg valgrind errors

Do a better job suppressing these errors until I can prioritize fixing
them.

BUG=skia:4550
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641763002

Review URL: https://codereview.chromium.org/1641763002

skiaserve: Set 303 status code on redirect.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641823002

Review URL: https://codereview.chromium.org/1641823002

Make fuzz output the words success for a 'clean exit'

The analysis looks for "Success" to determine if the fuzz should be considered
a clean exit or not.  It classifies clean exists as "grey" fuzzes, i.e. fixed.

TBR=mtklein@google.com
BUG=skia:4438
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1646603002

Review URL: https://codereview.chromium.org/1646603002

Wire up JsonCanvas in skiaserve

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1640333002

Review URL: https://codereview.chromium.org/1640333002

Consolidate SkStream copying methods

Make SkCopyStreamToData call SkStreamCopy, removing duplicate code.

The former still has its own method of copying with a length, since
it saves one copy.

BUG=skia:4788
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1640793002

Review URL: https://codereview.chromium.org/1640793002

switched to SkJSONCPP for JSON output
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1639333002

Review URL: https://codereview.chromium.org/1639333002

Fix Windows build

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1633413003

Review URL: https://codereview.chromium.org/1633413003

Stop testing SkImageDecoder in DM/nanobench

We have already used it for comparison, and are switching forward
to using SkCodec. This also allows us to simplify the code for checking
the extensions we support for images.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641663002

Review URL: https://codereview.chromium.org/1641663002

Enable RAW codec for Windows

* Use new DNG SDK version from ASOP, which fixed some build issues.
* Fix SkRawCodec.
* Fix gyp files.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1641533004

Review URL: https://codereview.chromium.org/1641533004

It derives the dng_host and overrides PerformAreaTask() to split the task into several sub-tasks which get added to SkTaskGroup.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1634763002

Review URL: https://codereview.chromium.org/1634763002

Fix complier issue [-Werror, -Wvla]

BUG=skia:4861
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1638323002

Review URL: https://codereview.chromium.org/1638323002

Add primitive url handling system to skiaserve

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1644633002

Review URL: https://codereview.chromium.org/1644633002

Update DEPS to use new PIEX version from AOSP.

Highlight:
* Better support for reading GPS tags

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1646433003

Review URL: https://codereview.chromium.org/1646433003

Allow clients to upload skps to skiaserve

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1638883002

Review URL: https://codereview.chromium.org/1638883002

Add gpu implementation of OverdrawXfermode
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1607253002

Committed: https://skia.googlesource.com/skia/+/8bc3cf88bbf5e5d5724356f076931bb70a6117ba

Review URL: https://codereview.chromium.org/1607253002

move more geometry to simd

Remove duplicate quad and cubic code around
computing the polynomial coefficients, and
use common SIMD-based code instead.

R=reed@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1633143002

Review URL: https://codereview.chromium.org/1633143002

Add gpu backend for SkPixelXorXfermode
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1631873002

Review URL: https://codereview.chromium.org/1631873002

remove unused sample

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1634683006

TBR=

Review URL: https://codereview.chromium.org/1634683006

Cleaning up trailing spaces and typos.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1631993002

Review URL: https://codereview.chromium.org/1631993002

Image filters: fix srcOffset handling in asFragmentProcessor() path.

Filters such as SkMatrixConvolutionImageFilter which use the
asFragmentProcessor() path were not correctly handling srcOffset.
It is correctly applied to the bounds, but the srcRect and dstRect
were computed from the pre-offset bounds.

The fix is to move them to just above where they're used in drawing.

Note: this change adds a new test case to the imagefiltersgraph GM,
so it will have to be rebaselined post-landing.

BUG=skia:4855
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1637443003

Review URL: https://codereview.chromium.org/1637443003

Initial support for turning Skia draws into a JSON document and vice versa.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1636563002

Committed: https://skia.googlesource.com/skia/+/3cb582f688822461efa5a034e18008bf2f11e4f8

Review URL: https://codereview.chromium.org/1636563002

de-proc sk_float_rsqrt

This is the first of many little baby steps to have us stop runtime-detecting NEON.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1616013003
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Committed: https://skia.googlesource.com/skia/+/efcc125acd2d71eb077caf6db65fdd6b9eb1dc0d

Review URL: https://codereview.chromium.org/1616013003

check for both __ARM_NEON__ and __ARM_NEON

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

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1630903004

Review URL: https://codereview.chromium.org/1630903004

Make A8 readbacks work from non-A8 surfaces

Expand testing of A8 readbacks.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1613903005

Review URL: https://codereview.chromium.org/1613903005

Remove SkLerpXfermode

This relies on the Chromium CL https://codereview.chromium.org/1610573004/ (Replace use of SkLerpXfermode with SkArithmeticMode)

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1611633002

Review URL: https://codereview.chromium.org/1611633002

Revert of Initial support for turning Skia draws into a JSON document and vice versa. (patchset #3 id:80001 of https://codereview.chromium.org/1636563002/ )

Reason for revert:
Breaking the CMake build.

Original issue's description:
> Initial support for turning Skia draws into a JSON document and vice versa.
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1636563002
>
> Committed: https://skia.googlesource.com/skia/+/3cb582f688822461efa5a034e18008bf2f11e4f8

TBR=jcgregorio@google.com,joshualitt@google.com,mtklein@google.com,ethannicholas@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/1637643002

Initial support for turning Skia draws into a JSON document and vice versa.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1636563002

Review URL: https://codereview.chromium.org/1636563002

Disable RAW test when SkRawCodec is not compiled

TBR=scroggo@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1639533002

Review URL: https://codereview.chromium.org/1639533002

Add pre_log option to nanobench; make this option default true in nanobench_flags.py and dm_flags.py.

BUG=skia:4826
BUG=skia:4827
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1612483002

Review URL: https://codereview.chromium.org/1612483002

Work around vs2013sp2-3 bug in skstd::unique_ptr.

When compiling skstd::unique_ptr::compressed_base with the vc++ in
some intermediate service pack versions of Visual Studio 2013 the
compiler will crash with an internal compiler error. In the interest of
reducing headaches, work around this issue in skstd::unique_ptr until
std::unique_ptr can be used.

BUG=skia:4564

Review URL: https://codereview.chromium.org/1636503002

Fix gyp files so that gyp_to_Android.mk succeeds

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1630723003

Review URL: https://codereview.chromium.org/1630723003

Fix memory leak in pixelxorxfermode GM
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1633773002

TBR=jvanverth@google.com

Review URL: https://codereview.chromium.org/1633773002

Changes to public.bzl to support building on iOS with objc_library bazel target.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1639443002

Review URL: https://codereview.chromium.org/1639443002

Disable RAW test on Chrome OS, Windows

TBR=scroggo@google.com

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1633763002

Review URL: https://codereview.chromium.org/1633763002

Fix leak of SkImageGenerator

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1638463002

Review URL: https://codereview.chromium.org/1638463002

Build and link microhttpd from gyp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1628363002

Review URL: https://codereview.chromium.org/1628363002

Remove RAW Codec from Google3 build.

This should fix the build after https://codereview.chromium.org/1520403003 and allow the autoroller to complete. I will follow up to see if it will be simple to re-enable this codec.
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1628393002

Review URL: https://codereview.chromium.org/1628393002

spin off some safe parts from AVX2 CL

(reviewed here https://codereview.chromium.org/1532613002/)

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1628333003
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1628333003

Revert of AVX 2 SrcOver blits: color32, blitmask. (patchset #24 id:450001 of https://codereview.chromium.org/1532613002/ )

Reason for revert:
Bot failures

Original issue's description:
> AVX 2 SrcOver blits: color32, blitmask.
>
> As a follow up to the SSE 4.1 CL, this should look pretty familiar.
>
> I've made some organizational changes around how we load, store, pack, and unpack data that I think makes things clearer and more orthogonal, and it'll make it easier to try out a pmaddubsw lerp.  I have backported these changes to the SSE 4.1 code, and I hope that I can actually get a lot of this code templated for sharing between the two later.
>
> Perf changes (relative to SSE 4.1):
> Xfermode_SrcOver:      1650 -> 1180  (0.71x)  // large opaque blit
> Xfermode_SrcOver_aa:   1794 -> 1653  (0.92x)  // large opaque + small transparent
> text_16_AA_{FF,BK,WT}: 1.72 -> 1.59  (0.92x)  // small opaque blit
> text_16_AA_88:         1.83 -> 1.77  (0.97x)  // small transparent blit
>
> This should be a big throughout win, and a small latency win.
> This should all be pixel-exact to the previous SSE 4.1 code.
>
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1532613002
> CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot;client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/5d2117015eb271e09faf4a7ddd89093c9d618a36

TBR=herb@google.com,mtklein@google.com,mtklein@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/1632713002

Add GM for SkPixelXorXfermode

This is in preparation for adding a gpu implementation

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1601743009

Review URL: https://codereview.chromium.org/1601743009

AVX 2 SrcOver blits: color32, blitmask.

As a follow up to the SSE 4.1 CL, this should look pretty familiar.

I've made some organizational changes around how we load, store, pack, and unpack data that I think makes things clearer and more orthogonal, and it'll make it easier to try out a pmaddubsw lerp.  I have backported these changes to the SSE 4.1 code, and I hope that I can actually get a lot of this code templated for sharing between the two later.

Perf changes (relative to SSE 4.1):
Xfermode_SrcOver:      1650 -> 1180  (0.71x)  // large opaque blit
Xfermode_SrcOver_aa:   1794 -> 1653  (0.92x)  // large opaque + small transparent
text_16_AA_{FF,BK,WT}: 1.72 -> 1.59  (0.92x)  // small opaque blit
text_16_AA_88:         1.83 -> 1.77  (0.97x)  // small transparent blit

This should be a big throughout win, and a small latency win.
This should all be pixel-exact to the previous SSE 4.1 code.

GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1532613002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot;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/1532613002

Add RAW decoding into Skia.

TBR=reed@google.com

BUG=skia:

(Based on the work from ebrauer in https://codereview.chromium.org/1459473007)
(Based on the work from adaubert in https://codereview.chromium.org/1494003003)
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1520403003

Committed: https://skia.googlesource.com/skia/+/6bd8639f8c142eedf543f4e5f3b02d2bf11df308

Review URL: https://codereview.chromium.org/1520403003

Do not test images with alpha to 565

This should handle the untriaged images in Gold.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1635453002

Review URL: https://codereview.chromium.org/1635453002

Fixed compilation on mingw.

Found and tested in mingw Firefox build. There are two problems:

- GCC doesn't support __vectorcall calling convention, so its usage needs to be #ifdefed
- GetProcAddress returns FARPROC type (a function pointer) and GCC requires an explicit cast to void*

Review URL: https://codereview.chromium.org/1589933002

remove legacy defines

The defines

SK_SUPPORT_LEGACY_ARCTO
SK_SUPPORT_LEGACY_CONIC_MEASURE
SK_SUPPORT_LEGACY_DASH_MEASURE
SK_SUPPORT_LEGACY_HAIR_IGNORES_CAPS
SK_SUPPORT_LEGACY_PATH_MEASURE_TVALUE

have been removed from Chrome. This removes the obsolete code
from Skia as well.

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1627703002

Review URL: https://codereview.chromium.org/1627703002

Delete 2 legacy flags

No longer used in Chrome/Android:

  SK_SUPPORT_LEGACY_ANISOTROPIC_MIPMAPS
  SK_SUPPORT_LEGACY_PRESCALE_SEMANTICS

R=caryclark@google.com,reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1632633002

Review URL: https://codereview.chromium.org/1632633002