platform/upstream/libSkiaSharp.git
8 years agoSkPDF/Deflate: clean up old SkFlate code
halcanary [Tue, 18 Aug 2015 20:30:25 +0000 (13:30 -0700)]
SkPDF/Deflate: clean up old SkFlate code

Factor out some of https://crrev.com/1227913008

BUG=skia:3030

Review URL: https://codereview.chromium.org/1298243002

8 years agoPath stenciling in GrBatch
bsalomon [Tue, 18 Aug 2015 20:28:19 +0000 (13:28 -0700)]
Path stenciling in GrBatch

Review URL: https://codereview.chromium.org/1293973003

8 years agoTest scaling for small images
msarett [Tue, 18 Aug 2015 20:22:46 +0000 (13:22 -0700)]
Test scaling for small images

We don't want to test small images on Gold because they are
not interested to look at.  Instead, I wrote a unit test to
verify that scaling small images does not cause crashes.

BUG=skia:

Review URL: https://codereview.chromium.org/1287863004

8 years agoTry enabling all nanobenches on Nexus7
egdaniel [Tue, 18 Aug 2015 20:08:36 +0000 (13:08 -0700)]
Try enabling all nanobenches on Nexus7

BUG=skia:2774

Review URL: https://codereview.chromium.org/1290223005

8 years agohandle no shader from image in drawAtlas
reed [Tue, 18 Aug 2015 19:50:24 +0000 (12:50 -0700)]
handle no shader from image in drawAtlas

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1298233003

8 years agoRemove Cmd in GrTargetCommands that is no longer needed with GrCopySurfaceBatch
bsalomon [Tue, 18 Aug 2015 19:30:02 +0000 (12:30 -0700)]
Remove Cmd in GrTargetCommands that is no longer needed with GrCopySurfaceBatch

TBR=joshualitt@google.com

Review URL: https://codereview.chromium.org/1292623004

8 years agoTry again to put SkXfermode_opts in SK_OPTS_NS
mtklein [Tue, 18 Aug 2015 19:23:28 +0000 (12:23 -0700)]
Try again to put SkXfermode_opts in SK_OPTS_NS

Remember failed attempt https://codereview.chromium.org/1286093004/ ?  I think this one is simpler and safer and even technically legal C++.

BUG=skia:4117

Review URL: https://codereview.chromium.org/1296183004

8 years agoAdd skeleton json support to visualbench
joshualitt [Tue, 18 Aug 2015 19:13:34 +0000 (12:13 -0700)]
Add skeleton json support to visualbench

BUG=skia:

Review URL: https://codereview.chromium.org/1298213002

8 years agoGrCopySurfaceBatch
bsalomon [Tue, 18 Aug 2015 19:12:35 +0000 (12:12 -0700)]
GrCopySurfaceBatch

Review URL: https://codereview.chromium.org/1289673004

8 years agoadded emitChild() to GrGLFragmentProcessor; removed AutoFragmentChildProcAdvance...
wangyix [Tue, 18 Aug 2015 19:00:12 +0000 (12:00 -0700)]
added emitChild() to GrGLFragmentProcessor; removed AutoFragmentChildProcAdvance class

BUG=skia:4182

Review URL: https://codereview.chromium.org/1301523003

8 years agoAdd support for non-mappable vert buffers to tessellating path renderer.
senorblanco [Tue, 18 Aug 2015 18:46:28 +0000 (11:46 -0700)]
Add support for non-mappable vert buffers to tessellating path renderer.

Use malloc-ed memory and the updateData() call instead.

BUG=skia:4215

Review URL: https://codereview.chromium.org/1288683004

8 years agoRevert of Update skia to test with libpng to 1.6.10 (patchset #3 id:60001 of https...
msarett [Tue, 18 Aug 2015 18:36:44 +0000 (11:36 -0700)]
Revert of Update skia to test with libpng to 1.6.10 (patchset #3 id:60001 of https://codereview.chromium.org/1301523005/ )

Reason for revert:
DM is failing on gm tests on Android bots.  Cause is not yet clear.

Original issue's description:
> Update skia to test with libpng to 1.6.10
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/f272bb03df9b86e7ea2cf23fb4d5cc56624e0118

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

Review URL: https://codereview.chromium.org/1298223002

8 years agoWhen getGLInstance is called on a frag proc, the resulting GrGLFragmentProcessor...
wangyix [Tue, 18 Aug 2015 18:29:31 +0000 (11:29 -0700)]
When getGLInstance is called on a frag proc, the resulting GrGLFragmentProcessor will be the root of a tree of GrGLFragmentProcessors that mirrors the GrFragmentProcessor's tree. This allows setData() to be called recursively (removing the responsibility from compose shader) and allows gl instances direct access to their children gl instances so they can emit their code.

BUG=skia:4182

Review URL: https://codereview.chromium.org/1287023009

8 years agoAdd subsets to SkImageGenerator and SkImageCacherator
reed [Tue, 18 Aug 2015 18:16:09 +0000 (11:16 -0700)]
Add subsets to SkImageGenerator and SkImageCacherator
... to support subsets in SkImage!

BUG=skia:

Review URL: https://codereview.chromium.org/1301633002

8 years agoUpdate skia to test with libpng to 1.6.10
msarett [Tue, 18 Aug 2015 17:46:01 +0000 (10:46 -0700)]
Update skia to test with libpng to 1.6.10

BUG=skia:

Review URL: https://codereview.chromium.org/1301523005

8 years agoAdd internal links section to skia.org
hcm [Tue, 18 Aug 2015 17:38:00 +0000 (10:38 -0700)]
Add internal links section to skia.org

BUG=skia:4142
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=1284043004

Review URL: https://codereview.chromium.org/1284043004

8 years agoClearStencilClip in GrBatch
bsalomon [Tue, 18 Aug 2015 17:33:30 +0000 (10:33 -0700)]
ClearStencilClip in GrBatch

Review URL: https://codereview.chromium.org/1288963004

8 years agoMove GrTBatchTesselator to its own file
joshualitt [Tue, 18 Aug 2015 17:16:01 +0000 (10:16 -0700)]
Move GrTBatchTesselator to its own file

BUG=skia:

Review URL: https://codereview.chromium.org/1298983002

8 years agoUpdate SkOpts namespaces.
mtklein [Tue, 18 Aug 2015 17:00:29 +0000 (10:00 -0700)]
Update SkOpts namespaces.

portable -> default, and everyone gets an sk_ prefix.

BUG=skia:4117

Review URL: https://codereview.chromium.org/1299013003

8 years agoAdd missing overrides to clear and discard batch dumpInfo()
Brian Salomon [Tue, 18 Aug 2015 16:46:51 +0000 (12:46 -0400)]
Add missing overrides to clear and discard batch dumpInfo()

TBR=joshualitt@google.com

Review URL: https://codereview.chromium.org/1303533003

8 years agoPatches on top of Radu's latest.
mtklein [Tue, 18 Aug 2015 16:43:28 +0000 (09:43 -0700)]
Patches on top of Radu's latest.

    patch from issue 1273033005 at patchset 120001 (http://crrev.com/1273033005#ps120001)

BUG=skia:

Review URL: https://codereview.chromium.org/1288323004

8 years agoUse portable code for family names with DirectWrite.
bungeman [Tue, 18 Aug 2015 16:34:07 +0000 (09:34 -0700)]
Use portable code for family names with DirectWrite.

IDWriteFamily::GetFamilyNames appears to be 'helpful' and removes
parts of family names that look like style names. Since the iterator
is supposed to return the actual names and not just the name the
platform thinks the name is (as getFamilyName does), try returning
the raw names when possible.

BUG=skia:4217

Review URL: https://codereview.chromium.org/1302573002

8 years agoPut clear and discard into GrBatch.
bsalomon [Tue, 18 Aug 2015 16:20:09 +0000 (09:20 -0700)]
Put clear and discard into GrBatch.

Review URL: https://codereview.chromium.org/1293563003

8 years agoUpdate BUILD.public a bit.
mtklein [Tue, 18 Aug 2015 15:51:49 +0000 (08:51 -0700)]
Update BUILD.public a bit.

 - SSE2 files are unfortunately now mixed-case, _SSE2 or _sse2.
 - Adds lists for SSSE3 and SSE4 files.
 - Remove SkDocument_PDF_None.cpp
 - Remove a few more references to animator.
 - Exclude private headers from HDRS.
 - Formatting and notes.

BUG=skia:

Review URL: https://codereview.chromium.org/1298003007

8 years agoClean up BUILD_simulator.py
halcanary [Tue, 18 Aug 2015 15:35:45 +0000 (08:35 -0700)]
Clean up BUILD_simulator.py

Review URL: https://codereview.chromium.org/1298813002

8 years agoDeduplicate typefaces across sub-pictures
mtklein [Tue, 18 Aug 2015 15:29:59 +0000 (08:29 -0700)]
Deduplicate typefaces across sub-pictures

Old flow to serialize a picture:
   1) serialize picture ops
   2) serialize all sub pictures recursively
   3) flatten the rest of this picture into a buffer, deduping flattenable factories and typefaces as we go
   4) serialize the factories and typefaces
   5) serialize the bytes from 3)

This allows the data in step 5) to refer to the deduplicated factories and typefaces from step 4).  But, each sub picture in step 2) is completely siloed, so they can't dedup with the parent picture or each other.

New flow:
   1) serialize picture ops
   2) flatten the rest of this picture into a buffer, deduping flattenable factories and typefaces as we go
   3) dummy-serialize sub pictures into /dev/null, with the effect of adding any new typefaces to our dedup set
   4) serialize the factories and typefaces
   5) serialize the bytes from 2)
   6) serialize all sub pictures recursively, with perfect deduplication because of step 3).

Now all typefaces in the top-level picture and all sub pictures recursively should end up deduplicated in the top-level typeface set.

Decoding changes are similar: we just thread through the top-level typefaces to the sub pictures.  What's convenient / surprising is that this new code correctly reads old pictures if we just have each picture prefer its local typeface set over the top-level one: old pictures always just use their own typefaces, and new pictures always use the top-level ones.

BUG=skia:4092

Review URL: https://codereview.chromium.org/1233953004

8 years agoPrefer native scaling to sampling
msarett [Tue, 18 Aug 2015 15:03:58 +0000 (08:03 -0700)]
Prefer native scaling to sampling

In the cases we have come across so far, native scaling has better performance
and correctness than sampling.  If native scaling is supported we want to use
it.

Jpegs native scale rounds up.
Ex: An 11x11 image with sampleSize=8 scales to 2x2.

SkScaledCodec rounds down.
Ex: An 11x11 image with sampleSize=8 scales to 1x1.

Before the CL, we would choose to use SkScaledCodec because it scales closer to
the "ideal" scale.

I think we want to go with the native option as long as its within 1 of the
ideal value.
BUG=skia:

Review URL: https://codereview.chromium.org/1284243004

8 years agoRegenerate LCD text blobs if GrPaint's color changes, not SkPaint.
jvanverth [Tue, 18 Aug 2015 14:44:22 +0000 (07:44 -0700)]
Regenerate LCD text blobs if GrPaint's color changes, not SkPaint.

The GrPaint's color takes into account shaders and color filters, so is a
more accurate picture of the color state.

BUG=chromium:511787

Review URL: https://codereview.chromium.org/1297053004

8 years agoMade isEqual in GrFragmentProcessor recursive
wangyix [Tue, 18 Aug 2015 14:39:33 +0000 (07:39 -0700)]
Made isEqual in GrFragmentProcessor recursive

Added comment about how computeInvariantOutput() is non-recursive in GrFragmentProcessor

Made isEqual() recursive in GrFragmentProcessor

BUG=skia:4182

Review URL: https://codereview.chromium.org/1287343005

8 years agoGrAAFillRectBatch holds onto GrPipelineOptimizations
joshualitt [Tue, 18 Aug 2015 14:25:38 +0000 (07:25 -0700)]
GrAAFillRectBatch holds onto GrPipelineOptimizations

BUG=skia:

Review URL: https://codereview.chromium.org/1301603004

8 years agoAdded class AutoFragmentChildProcAdvance to be constructed before a child emitCode...
wangyix [Tue, 18 Aug 2015 14:24:29 +0000 (07:24 -0700)]
Added class AutoFragmentChildProcAdvance to be constructed before a child emitCode and destructed after

Fixed wrong indent

Changed auto child advance back to backwards linear search for getting subset of coords and samplers array of a child

Used offset from parent instead of backwards linear search to find a child proc's coords and transforms in Auto...Advance

append mangleString to variable name in nameVariable()

BUILDS! Added AutoFragmentChildProcAdvance class; fixed a few errors from previous commits

BUG=skia:4182

Review URL: https://codereview.chromium.org/1286293002

8 years agofix pathops fuzz failures
caryclark [Tue, 18 Aug 2015 14:12:43 +0000 (07:12 -0700)]
fix pathops fuzz failures

If a curve has the identical start and control points, the
initial or final tangent can't be trivally determined. The
perpendicular to the tangent is used to measure coincidence.

Add logic for cubics, quadratics, and conics, to use the
secondary control points or the end points if the initial
control point alone can't determine the tangent.

Add debugging (currently untriggered by exhaustive testing)
to detect zero-length tangents which are not at the curve
endpoints.

Increase the number of temporary intersecions gathered from
10 to 12 but reduce the max passed in by cubic intersection from
27 to 12. Also, add checks if the max passed exceeds the
storage allocated.

When cleaning up parallel lines, choose the intersection which
is on the end of both segments over the intersection which
is on the end of a single segment.

TBR=reed@google.com
BUG=425140,516266

Review URL: https://codereview.chromium.org/1288863004

8 years agoReland of Fix a few bugs in the google3 sync scripts. (patchset #1 id:1 of https...
benjaminwagner [Tue, 18 Aug 2015 13:25:14 +0000 (06:25 -0700)]
Reland of Fix a few bugs in the google3 sync scripts. (patchset #1 id:1 of https://codereview.chromium.org/1295213002/ )

Reason for revert:
Retry.

Original issue's description:
> Revert of Fix a few bugs in the google3 sync scripts. (patchset #1 id:1 of https://codereview.chromium.org/1291343006/ )
>
> Reason for revert:
> MacMini buildbots were unable to apply the patch.
> fatal: Could not parse object '18af0a0080cc2b2b3464292d35a0886cdade551f'.
> http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release/builds/1580
> http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Debug/builds/1474
> http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Release/builds/1566
> http://build.chromium.org/p/client.skia/builders/Perf-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Release/builds/1480
> http://build.chromium.org/p/client.skia/builders/Perf-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release/builds/1456
> http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Debug/builds/1364
>
>
> Original issue's description:
> > Fix a few bugs in the google3 sync scripts.
> >
> > BUG=skia:
> >
> > Committed: https://skia.googlesource.com/skia/+/18af0a0080cc2b2b3464292d35a0886cdade551f
>
> TBR=mtklein@google.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/e349d6b92574c1b0acdb4ba1ed8ff905ea830131

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

Review URL: https://codereview.chromium.org/1284093003

8 years agoPrivatize GrBatch subclass overrides
bsalomon [Tue, 18 Aug 2015 13:05:14 +0000 (06:05 -0700)]
Privatize GrBatch subclass overrides

Review URL: https://codereview.chromium.org/1301663002

8 years agoRemove a couple zombie pathops files.
mtklein [Tue, 18 Aug 2015 12:51:56 +0000 (05:51 -0700)]
Remove a couple zombie pathops files.

These files don't compile and are not referenced by any GYP file.

BUG=skia:

Review URL: https://codereview.chromium.org/1299773002

8 years agoRemove SkOpts_sse2.cpp.
mtklein [Tue, 18 Aug 2015 12:18:52 +0000 (05:18 -0700)]
Remove SkOpts_sse2.cpp.

It's sort of pointless: all our clients that will have SSE2 at runtime have it
unconditionally at compile time, so the functions in namespace portable will
pick up the SSE2 code.  The procs in SkOpts_sse2.o were just duplicate code.

A couple of the procs we had in _sse2.cpp can benefit slightly when compiled
with SSSE3.  I've moved those to _ssse3.cpp.  This should lead to small speedups
on platforms like Linux and Windows that have a baseline of SSE2.

Similarly, I've removed the call to Init_neon() when NEON is available globally... it's a no-op.

Renaming namespace portable to something clearer is TBD.

BUG=skia:4117

Review URL: https://codereview.chromium.org/1294213002

8 years agoRevert[8] "move some public headers into private"
reed [Tue, 18 Aug 2015 01:29:47 +0000 (18:29 -0700)]
Revert[8] "move some public headers into private"

This reverts commit fb28cd2b13d76c324570341e1155aaadc1b532b6.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1298833002

8 years agoMove normal map creation methods to sk_tools
robertphillips [Mon, 17 Aug 2015 22:04:47 +0000 (15:04 -0700)]
Move normal map creation methods to sk_tools

Split off of https://codereview.chromium.org/1261433009/ (Add SkCanvas::drawLitAtlas call)

Review URL: https://codereview.chromium.org/1298763002

8 years agoAdd a simple script to simulate BUILD file glob() expansion.
mtklein [Mon, 17 Aug 2015 22:02:57 +0000 (15:02 -0700)]
Add a simple script to simulate BUILD file glob() expansion.

This lets us test changes to BUILD.public.

This is not yet automated in any way.  My hope is to trigger it quietly via the presubmit for any CL that adds or removes a file, or changes BUILD.public.

BUG=skia:

Review URL: https://codereview.chromium.org/1290833003

8 years agoRevert of Fix a few bugs in the google3 sync scripts. (patchset #1 id:1 of https...
benjaminwagner [Mon, 17 Aug 2015 21:44:01 +0000 (14:44 -0700)]
Revert of Fix a few bugs in the google3 sync scripts. (patchset #1 id:1 of https://codereview.chromium.org/1291343006/ )

Reason for revert:
MacMini buildbots were unable to apply the patch.
fatal: Could not parse object '18af0a0080cc2b2b3464292d35a0886cdade551f'.
http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release/builds/1580
http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Debug/builds/1474
http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Release/builds/1566
http://build.chromium.org/p/client.skia/builders/Perf-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Release/builds/1480
http://build.chromium.org/p/client.skia/builders/Perf-Mac10.9-Clang-MacMini6.2-CPU-AVX-x86_64-Release/builds/1456
http://build.chromium.org/p/client.skia/builders/Test-Mac10.9-Clang-MacMini6.2-GPU-HD4000-x86_64-Debug/builds/1364

Original issue's description:
> Fix a few bugs in the google3 sync scripts.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/18af0a0080cc2b2b3464292d35a0886cdade551f

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

Review URL: https://codereview.chromium.org/1295213002

8 years agoFix a few bugs in the google3 sync scripts.
benjaminwagner [Mon, 17 Aug 2015 21:23:03 +0000 (14:23 -0700)]
Fix a few bugs in the google3 sync scripts.

BUG=skia:

Review URL: https://codereview.chromium.org/1291343006

8 years agoAdd files added to Skia HEAD in google3 to the Git repo.
benjaminwagner [Mon, 17 Aug 2015 19:58:10 +0000 (12:58 -0700)]
Add files added to Skia HEAD in google3 to the Git repo.

LICENSE is modified in google3 to mention some code copied from other projects.

BUG=skia:

Review URL: https://codereview.chromium.org/1291603003

8 years agoMake GrVertexBatch objects hold their own draws during GrDrawTarget flush
bsalomon [Mon, 17 Aug 2015 19:55:38 +0000 (12:55 -0700)]
Make GrVertexBatch objects hold their own draws during GrDrawTarget flush

NO_MERGE_BUILDS

Review URL: https://codereview.chromium.org/1286043004

8 years agoStop ping ponging AA in SkGpuDevice
joshualitt [Mon, 17 Aug 2015 18:53:44 +0000 (11:53 -0700)]
Stop ping ponging AA in SkGpuDevice

BUG=skia:

Review URL: https://codereview.chromium.org/1299703003

8 years agoRevert of Depend on ETC1 via DEPS instead of a direct third_party checkin. (patchset...
mtklein [Mon, 17 Aug 2015 18:15:00 +0000 (11:15 -0700)]
Revert of Depend on ETC1 via DEPS instead of a direct third_party checkin. (patchset #1 id:1 of https://codereview.chromium.org/1296253003/ )

Reason for revert:
This will make the google3 roll awkward.  Reverting.

Original issue's description:
> Depend on ETC1 via DEPS instead of a direct third_party checkin.
>
> Yes, this does mean we'll check out ~40M of code for 2 files...
>
> I think these trybots are moot... they gclient sync before patching.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/ca01d179347b5e848ded97c8228b79ffc6b5e7bf

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

Review URL: https://codereview.chromium.org/1290913008

8 years agodrawBitmapImage can batch across AA rects
joshualitt [Mon, 17 Aug 2015 17:51:22 +0000 (10:51 -0700)]
drawBitmapImage can batch across AA rects

BUG=464835

Committed: https://skia.googlesource.com/skia/+/0d4bcea0d570041434ac3de2df2bd9063138fdb5

Review URL: https://codereview.chromium.org/1293543002

8 years agofix typo in android_framework_lib.gyp
Derek Sollenberger [Mon, 17 Aug 2015 17:39:03 +0000 (13:39 -0400)]
fix typo in android_framework_lib.gyp

BUG=skia:

Review URL: https://codereview.chromium.org/1293183002 .

8 years agoAdd GPU to the android include path.
djsollen [Mon, 17 Aug 2015 17:20:46 +0000 (10:20 -0700)]
Add GPU to the android include path.

Review URL: https://codereview.chromium.org/1301623002

8 years agoDepend on ETC1 via DEPS instead of a direct third_party checkin.
mtklein [Mon, 17 Aug 2015 17:20:16 +0000 (10:20 -0700)]
Depend on ETC1 via DEPS instead of a direct third_party checkin.

Yes, this does mean we'll check out ~40M of code for 2 files...

I think these trybots are moot... they gclient sync before patching.

BUG=skia:

Review URL: https://codereview.chromium.org/1296253003

8 years agoremove unused code
caryclark [Mon, 17 Aug 2015 17:01:23 +0000 (10:01 -0700)]
remove unused code

R=mtklein@google.com

Review URL: https://codereview.chromium.org/1297013002

8 years agoSkPaintFilterCanvas should inherit the target canvas state
fmalita [Mon, 17 Aug 2015 15:05:13 +0000 (08:05 -0700)]
SkPaintFilterCanvas should inherit the target canvas state

Currently, SkPaintFilterCanvas does not provide any help in cloning
target canvas state.  While that could be handled in subclasses, it is
easy to miss (see linked bug).

This CL adds a new constructor variant which ensures that the initial
matrix and clip bounds are inherited from the target canvas.

BUG=516790
R=reed@google.com,robertphillips@google.com

Review URL: https://codereview.chromium.org/1294013002

8 years agoFix webp compile warnings on windows
msarett [Mon, 17 Aug 2015 14:49:37 +0000 (07:49 -0700)]
Fix webp compile warnings on windows

This fix was landed with:
https://codereview.chromium.org/1280073002/

The above CL was reverted due to an unrelated bug in libwebp.

The above CL contains multiple components, and I think that
reverting this part of the change was unintentional.

BUG=skia:

Review URL: https://codereview.chromium.org/1286903003

8 years agoRevert of drawBitmapImage can batch across AA rects (patchset #5 id:80001 of https...
joshualitt [Mon, 17 Aug 2015 14:20:58 +0000 (07:20 -0700)]
Revert of drawBitmapImage can batch across AA rects (patchset #5 id:80001 of https://codereview.chromium.org/1293543002/ )

Reason for revert:
a follow up patch exposes a bug

Original issue's description:
> drawBitmapImage can batch across AA rects
>
> BUG=464835
>
> Committed: https://skia.googlesource.com/skia/+/0d4bcea0d570041434ac3de2df2bd9063138fdb5

TBR=bsalomon@google.com,robertphillips@google.com,joshualitt@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=464835

Review URL: https://codereview.chromium.org/1291663005

8 years agodrawBitmapImage can batch across AA rects
joshualitt [Mon, 17 Aug 2015 13:59:36 +0000 (06:59 -0700)]
drawBitmapImage can batch across AA rects

BUG=464835

Review URL: https://codereview.chromium.org/1293543002

8 years agoUpdate SKP version
rmistry [Sun, 16 Aug 2015 07:20:54 +0000 (00:20 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS

Review URL: https://codereview.chromium.org/1296513004

8 years agosupport both box and circle lcd
reed [Sat, 15 Aug 2015 21:02:29 +0000 (14:02 -0700)]
support both box and circle lcd

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1292943003

8 years agoshow LCD on max fat-zoom
reed [Sat, 15 Aug 2015 18:57:23 +0000 (11:57 -0700)]
show LCD on max fat-zoom

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1295483003

8 years agoUpdate SKP version
rmistry [Sat, 15 Aug 2015 18:00:57 +0000 (11:00 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=
NO_MERGE_BUILDS

Review URL: https://codereview.chromium.org/1291273003

8 years agoRevert[7] "move some public headers into private"
reed [Sat, 15 Aug 2015 15:46:27 +0000 (08:46 -0700)]
Revert[7] "move some public headers into private"

This reverts commit e02716908fe82c7c4ae9b415793277620a22bcd6.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1300523002

8 years agoincrease zoom in fatbits
reed [Sat, 15 Aug 2015 14:42:31 +0000 (07:42 -0700)]
increase zoom in fatbits

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1300513002

8 years agoRevert[6] "move some public headers into private"
reed [Sat, 15 Aug 2015 14:28:34 +0000 (07:28 -0700)]
Revert[6] "move some public headers into private"

This reverts commit cc1229c9d5e57b45305ba8f99ae33e92db0032bc.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1295003002

8 years agoRevert of stop dropping AA when rect stays rect (patchset #8 id:140001 of https:...
joshualitt [Fri, 14 Aug 2015 18:36:03 +0000 (11:36 -0700)]
Revert of stop dropping AA when rect stays rect (patchset #8 id:140001 of https://codereview.chromium.org/1295523002/ )

Reason for revert:
breaks layout tests

Original issue's description:
> stop dropping AA when rect stays rect
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/1bef9f59c566cc54c2259cc4d0171c115157cd1c

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

Review URL: https://codereview.chromium.org/1295813005

8 years agoSkPDF: Move SkFlate into src/pdf (part 3/3)
halcanary [Fri, 14 Aug 2015 18:26:33 +0000 (11:26 -0700)]
SkPDF: Move SkFlate into src/pdf (part 3/3)

This must land after http://crrev.com/1288053002

Motivation: Skia only uses SkFlate in the SkPDF module.

Review URL: https://codereview.chromium.org/1288063002

8 years agoA bit of misc cleanup.
mtklein [Fri, 14 Aug 2015 17:03:45 +0000 (10:03 -0700)]
A bit of misc cleanup.

 - Update comments about skia_shared_lib
 - Remove dead code (system_preferences)
 - Remove non-building code (techtalk1)
 - Turn gm/annotated_text back on.

BUG=skia:

Review URL: https://codereview.chromium.org/1286133004

8 years agofix valgrind issue
joshualitt [Fri, 14 Aug 2015 16:01:21 +0000 (09:01 -0700)]
fix valgrind issue

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1289203004

8 years agofix viewmatrix mapping in AAFillRect
joshualitt [Fri, 14 Aug 2015 15:45:39 +0000 (08:45 -0700)]
fix viewmatrix mapping in AAFillRect

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1293823003

8 years agoAdd include paths in gputest.gyp for use by additional gpu tests.
egdaniel [Fri, 14 Aug 2015 15:37:50 +0000 (08:37 -0700)]
Add include paths in gputest.gyp for use by additional gpu tests.

BUG=skia:

Review URL: https://codereview.chromium.org/1290553005

8 years agoBuild all GMs on Android.
mtklein [Fri, 14 Aug 2015 15:33:50 +0000 (08:33 -0700)]
Build all GMs on Android.

Many of these were excluded from the build for bots that no longer exist.

Will fix up anything that starts failing using dm_flags.py.

BUG=skia:2326,skia:2313, skia:1978

Review URL: https://codereview.chromium.org/1290943004

8 years agoSupport more swizzles to 565 in SkCodec
scroggo [Fri, 14 Aug 2015 15:32:46 +0000 (08:32 -0700)]
Support more swizzles to 565 in SkCodec

Add more swizzling functions for swizzling to 565. Much of this
code was revived from crrev.com/1055743003 (for BMP). Also added
swizzling functions for WBMP.

Consolidate the static function conversion_possible.

In SkCodec::getPixels, check that the alphatype corresponds to the
colorType. This prevents requesting 565 + non-opaque.

In SkIcoCodec, report that the image is unpremul (instead of
whatever the largest embedded codec thinks), but modify the
requested info to have the alpha type expected/required by the
embedded codec.

Add tests for decoding to 565.

BUG=skia:3257
BUG=skia:3683

Review URL: https://codereview.chromium.org/1277213002

8 years agoSkScaledCodec class
emmaleer [Fri, 14 Aug 2015 14:44:46 +0000 (07:44 -0700)]
SkScaledCodec class

This class does scaling by using a scanlineDecoder.
getScanlines and skipScanlines are used for y sampling,
the swizzler is used for x sampling

this class is currently only working for png and jpeg images
I will update other Codec types to work soon

For SkJpegCodec to implement width wise swizzling it now
uses a swizzler. I ran performance tests on this change.
Here are the performance test results:
https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/0944100ac89f797714eeae0cf2875e2335ff52ee

Committed: https://skia.googlesource.com/skia/+/d518ea7927f9f4e0ed5b4134d1b4f48243855a47

Committed: https://skia.googlesource.com/skia/+/b157917507d4f7d2651f0aeb566d31603cc02240

Review URL: https://codereview.chromium.org/1260673002

8 years agostop dropping AA when rect stays rect
joshualitt [Fri, 14 Aug 2015 14:10:57 +0000 (07:10 -0700)]
stop dropping AA when rect stays rect

BUG=skia:

Review URL: https://codereview.chromium.org/1295523002

8 years agoRevert of SkScaledCodec class (patchset #35 id:680001 of https://codereview.chromium...
egdaniel [Fri, 14 Aug 2015 13:37:37 +0000 (06:37 -0700)]
Revert of SkScaledCodec class (patchset #35 id:680001 of https://codereview.chromium.org/1260673002/ )

Reason for revert:
breaking ubuntu bots

Original issue's description:
> SkScaledCodec class
>
> This class does scaling by using a scanlineDecoder.
> getScanlines and skipScanlines are used for y sampling,
> the swizzler is used for x sampling
>
> this class is currently only working for png and jpeg images
> I will update other Codec types to work soon
>
> For SkJpegCodec to implement width wise swizzling it now
> uses a swizzler. I ran performance tests on this change.
> Here are the performance test results:
> https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/0944100ac89f797714eeae0cf2875e2335ff52ee
>
> Committed: https://skia.googlesource.com/skia/+/d518ea7927f9f4e0ed5b4134d1b4f48243855a47
>
> Committed: https://skia.googlesource.com/skia/+/b157917507d4f7d2651f0aeb566d31603cc02240

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

Review URL: https://codereview.chromium.org/1285173003

8 years agoAdd GM to test drawing atlas case
joshualitt [Fri, 14 Aug 2015 13:30:50 +0000 (06:30 -0700)]
Add GM to test drawing atlas case

BUG=skia:

Review URL: https://codereview.chromium.org/1294513002

8 years agoRevert of Paranoid temporary check that Chrome's building with SSE2 on all Windows...
mtklein [Fri, 14 Aug 2015 12:49:17 +0000 (05:49 -0700)]
Revert of Paranoid temporary check that Chrome's building with SSE2 on all Windows builders. (patchset #1 id:1 of https://codereview.chromium.org/1283183004/ )

Reason for revert:
Nothing broke.  Experiment is over... SSE2 is present.

Original issue's description:
> Paranoid temporary check that Chrome's building with SSE2 on all Windows builders.
>
> BUG=chromium:475596
>
> Committed: https://skia.googlesource.com/skia/+/e8445307ac33eddcf04457ae21040e47dc9c2a97

TBR=egdaniel@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:475596

Review URL: https://codereview.chromium.org/1296593002

8 years agoSkScaledCodec class
emmaleer [Fri, 14 Aug 2015 12:46:08 +0000 (05:46 -0700)]
SkScaledCodec class

This class does scaling by using a scanlineDecoder.
getScanlines and skipScanlines are used for y sampling,
the swizzler is used for x sampling

this class is currently only working for png and jpeg images
I will update other Codec types to work soon

For SkJpegCodec to implement width wise swizzling it now
uses a swizzler. I ran performance tests on this change.
Here are the performance test results:
https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/0944100ac89f797714eeae0cf2875e2335ff52ee

Committed: https://skia.googlesource.com/skia/+/d518ea7927f9f4e0ed5b4134d1b4f48243855a47

Review URL: https://codereview.chromium.org/1260673002

8 years agoParanoid temporary check that Chrome's building with SSE2 on all Windows builders.
mtklein [Fri, 14 Aug 2015 00:53:28 +0000 (17:53 -0700)]
Paranoid temporary check that Chrome's building with SSE2 on all Windows builders.

BUG=chromium:475596

Review URL: https://codereview.chromium.org/1283183004

8 years agoSplit GrDrawBatch and GrVertexBatch into their own files
bsalomon [Thu, 13 Aug 2015 21:55:50 +0000 (14:55 -0700)]
Split GrDrawBatch and GrVertexBatch into their own files

TBR=joshualitt@google.com

Review URL: https://codereview.chromium.org/1296483002

8 years agosome catchup CL revisions
reed [Thu, 13 Aug 2015 21:06:46 +0000 (14:06 -0700)]
some catchup CL revisions

BUG=skia:

Review URL: https://codereview.chromium.org/1295593002

8 years agoEvery pixel ref gets its own mutex.
mtklein [Thu, 13 Aug 2015 21:02:06 +0000 (14:02 -0700)]
Every pixel ref gets its own mutex.

Seems like a memory-saving flourish follow up would be to use SkSpinlock.

BUG=Florin's email.

Review URL: https://codereview.chromium.org/1289623004

8 years agoRevert of SkScaledCodec class (patchset #34 id:660001 of https://codereview.chromium...
emmaleer [Thu, 13 Aug 2015 20:59:21 +0000 (13:59 -0700)]
Revert of SkScaledCodec class (patchset #34 id:660001 of https://codereview.chromium.org/1260673002/ )

Reason for revert:
Seg-faulting:
http://build.chromium.org/p/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86-Release/builds/1608/steps/dm/logs/stdio

Original issue's description:
> SkScaledCodec class
>
> This class does scaling by using a scanlineDecoder.
> getScanlines and skipScanlines are used for y sampling,
> the swizzler is used for x sampling
>
> this class is currently only working for png and jpeg images
> I will update other Codec types to work soon
>
> For SkJpegCodec to implement width wise swizzling it now
> uses a swizzler. I ran performance tests on this change.
> Here are the performance test results:
> https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/0944100ac89f797714eeae0cf2875e2335ff52ee
>
> Committed: https://skia.googlesource.com/skia/+/d518ea7927f9f4e0ed5b4134d1b4f48243855a47

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

Review URL: https://codereview.chromium.org/1294613002

8 years agoBelt and suspenders for colors in DrawAtlasBatch test
jvanverth [Thu, 13 Aug 2015 20:41:03 +0000 (13:41 -0700)]
Belt and suspenders for colors in DrawAtlasBatch test

Review URL: https://codereview.chromium.org/1290253002

8 years agoIntroduce GrBatch subclasses GrDrawBatch and GrVertexBatch to prepare for non-drawing...
bsalomon [Thu, 13 Aug 2015 20:34:48 +0000 (13:34 -0700)]
Introduce GrBatch subclasses GrDrawBatch and GrVertexBatch to prepare for non-drawing batches

Review URL: https://codereview.chromium.org/1293583002

8 years agoExtend SkImageGenerator to support natively generated GrTextures. As part of this...
reed [Thu, 13 Aug 2015 20:32:39 +0000 (13:32 -0700)]
Extend SkImageGenerator to support natively generated GrTextures. As part of this, added uniqueID() to the generator, and made it be in the same namespace is bitmaps, pixelrefs, images.

To do this, create SkImageCacherator, which wraps a generator and provides an
interface to get a cached answer for either the raster or texture output of
the generator.

BUG=skia:

Review URL: https://codereview.chromium.org/1291803002

8 years agoNormalize SkXfermode_opts.h argument order as d,s[,aa].
mtklein [Thu, 13 Aug 2015 20:10:30 +0000 (13:10 -0700)]
Normalize SkXfermode_opts.h argument order as d,s[,aa].

At head they're s,d[,aa] in SkXfermode_opts.h but Sk4px::Map* expect d,s[,aa]
so we ended up having to write weird little lambda shims to match impedance.

There's no reason for these to disagree, and d,s[,aa] is the One True Order
(because no matter what you're doing in graphics, there's always a dst).

Should be no perf or image diff, though I'm suspicious it might help MSVC code generation.

BUG=skia:4117

Committed: https://skia.googlesource.com/skia/+/6028a8476504022fe40b6870b1460b5e4a80969f

CQ_EXTRA_TRYBOTS=client.skia:Test-Win8-MSVC-ShuttleB-CPU-AVX2-x86-Release-Trybot

Review URL: https://codereview.chromium.org/1289903002

8 years agoSkScaledCodec class
emmaleer [Thu, 13 Aug 2015 20:07:03 +0000 (13:07 -0700)]
SkScaledCodec class

This class does scaling by using a scanlineDecoder.
getScanlines and skipScanlines are used for y sampling,
the swizzler is used for x sampling

this class is currently only working for png and jpeg images
I will update other Codec types to work soon

For SkJpegCodec to implement width wise swizzling it now
uses a swizzler. I ran performance tests on this change.
Here are the performance test results:
https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/0944100ac89f797714eeae0cf2875e2335ff52ee

Review URL: https://codereview.chromium.org/1260673002

8 years agoFix for VALGRIND DrawAtlas test issue.
jvanverth [Thu, 13 Aug 2015 18:53:39 +0000 (11:53 -0700)]
Fix for VALGRIND DrawAtlas test issue.

Don't pre-alloc color array.

Review URL: https://codereview.chromium.org/1289003002

8 years agoRevert of SkScaledCodec class (patchset #32 id:620001 of https://codereview.chromium...
emmaleer [Thu, 13 Aug 2015 18:44:05 +0000 (11:44 -0700)]
Revert of SkScaledCodec class (patchset #32 id:620001 of https://codereview.chromium.org/1260673002/ )

Reason for revert:
Segfaulting:
http://build.chromium.org/p/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86-Release/builds/1605/steps/dm/logs/stdio

Original issue's description:
> SkScaledCodec class
>
> This class does scaling by using a scanlineDecoder.
> getScanlines and skipScanlines are used for y sampling,
> the swizzler is used for x sampling
>
> this class is currently only working for png and jpeg images
> I will update other Codec types to work soon
>
> For SkJpegCodec to implement width wise swizzling it now
> uses a swizzler. I ran performance tests on this change.
> Here are the performance test results:
> https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/0944100ac89f797714eeae0cf2875e2335ff52ee

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

Review URL: https://codereview.chromium.org/1294593002

8 years agoSkScaledCodec class
emmaleer [Thu, 13 Aug 2015 18:26:57 +0000 (11:26 -0700)]
SkScaledCodec class

This class does scaling by using a scanlineDecoder.
getScanlines and skipScanlines are used for y sampling,
the swizzler is used for x sampling

this class is currently only working for png and jpeg images
I will update other Codec types to work soon

For SkJpegCodec to implement width wise swizzling it now
uses a swizzler. I ran performance tests on this change.
Here are the performance test results:
https://docs.google.com/a/google.com/spreadsheets/d/1D7-Q_GXD_dI68LZO005NNvb8Wq2Ee0wEBEPG72671yw/edit?usp=sharing

BUG=skia:

Review URL: https://codereview.chromium.org/1260673002

8 years agofix up batch_spew
joshualitt [Thu, 13 Aug 2015 17:19:32 +0000 (10:19 -0700)]
fix up batch_spew

TBR=bsalomon@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/1292113002

8 years agoMove GrRectBatch to GrBWFillRectBatch and make internal
joshualitt [Thu, 13 Aug 2015 17:05:51 +0000 (10:05 -0700)]
Move GrRectBatch to GrBWFillRectBatch and make internal

BUG=skia:

Review URL: https://codereview.chromium.org/1294553003

8 years agoadd SkImage::NewFromPicture and a GM to test it
reed [Thu, 13 Aug 2015 16:37:45 +0000 (09:37 -0700)]
add SkImage::NewFromPicture and a GM to test it

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1288403002

8 years agoSimplify parameters passed to path renderers
robertphillips [Thu, 13 Aug 2015 14:57:10 +0000 (07:57 -0700)]
Simplify parameters passed to path renderers

It appears that CanDrawPathArgs::fTarget was only ever used to access the shaderCaps and no instances of onGetStencilSupport ever used the drawTarget and pipelineBuilder parameters.

Review URL: https://codereview.chromium.org/1291803006

8 years agoadd asserts on Sk4px::Map functions that our arrays are non-null.
mtklein [Thu, 13 Aug 2015 14:51:29 +0000 (07:51 -0700)]
add asserts on Sk4px::Map functions that our arrays are non-null.

There's a crasher (null dereference it looks like) in Clank with an impossible stack, but with blit mask on top (apparently sk_free() calls blit mask...).  Can't hurt to check that all these Sk4px-using methods are iterating over non-null arrays.

(This is that bug I mentioned that got me thinking about blit mask in the first place.)

BUG=520354

Review URL: https://codereview.chromium.org/1295443002

8 years agoMinor fix to not save the modeling space rect
joshualitt [Thu, 13 Aug 2015 14:12:50 +0000 (07:12 -0700)]
Minor fix to not save the modeling space rect

BUG=skia:

Review URL: https://codereview.chromium.org/1292543002

8 years agoMove willColorBlendWithDst from pipeline builder to GrPipelineOptimizations.
bsalomon [Thu, 13 Aug 2015 14:08:31 +0000 (07:08 -0700)]
Move willColorBlendWithDst from pipeline builder to GrPipelineOptimizations.

Review URL: https://codereview.chromium.org/1285193004

8 years agoAdded mangleString member and onBefore*, onAfter* functions to GrGLFragmentShaderBuilder
wangyix [Thu, 13 Aug 2015 13:51:35 +0000 (06:51 -0700)]
Added mangleString member and  onBefore*, onAfter* functions to GrGLFragmentShaderBuilder

BUILDS! Added mangleString, onBefore, and onAfterChildProcEmitCode() to GrGLFragmentShaderBuilder.cpp

BUG=skia:4182

Review URL: https://codereview.chromium.org/1288723002

8 years agoFail early in GrPipeline::CreateAt() if XP creation fails
bsalomon [Thu, 13 Aug 2015 13:44:04 +0000 (06:44 -0700)]
Fail early in GrPipeline::CreateAt() if XP creation fails

R=egdaniel@google.com

Review URL: https://codereview.chromium.org/1286173006

8 years agoDefer flushes if kPreferNoIO is specified
robertphillips [Thu, 13 Aug 2015 12:19:14 +0000 (05:19 -0700)]
Defer flushes if kPreferNoIO is specified

Prior to this patch clients who were solely uploading to textures (e.g., SW Mask Mgr) would cause extra flushes b.c., even though kPreferNoIO was being specified, resources with pending IO would still be returned even though there was plenty of space in the resource cache.

Review URL: https://codereview.chromium.org/1286203002

8 years agoRevert "Revert[4] "move some public headers into private""
reed [Thu, 13 Aug 2015 01:58:52 +0000 (18:58 -0700)]
Revert "Revert[4] "move some public headers into private""

This reverts commit 33e00db73d5f1836481cf4053c64e4ed20537662.

BUG=skia:
TBR=
NOTREECHECKS=True

FAILED: /b/build/goma/gomacc ../../third_party/llvm-build/Release+Asserts/bin/clang++ -MMD -MF obj/components/user_manager/user_manager_test_support.fake_user_manager.o.d -DV8_DEPRECATION_WARNINGS -DCLD_VERSION=2 -D_FILE_OFFSET_BITS=64 -DCHROMIUM_BUILD -DCR_CLANG_REVISION=242792-1 -DCOMPONENT_BUILD -DUI_COMPOSITOR_IMAGE_TRANSPORT -DUSE_AURA=1 -DUSE_ASH=1 -DUSE_PANGO=1 -DUSE_CAIRO=1 -DUSE_DEFAULT_RENDER_THEME=1 -DUSE_LIBJPEG_TURBO=1 -DUSE_X11=1 -DIMAGE_LOADER_EXTENSION=1 -DENABLE_WEBRTC=1 -DENABLE_MEDIA_ROUTER=1 -DUSE_PROPRIETARY_CODECS -DENABLE_PEPPER_CDMS -DENABLE_CONFIGURATION_POLICY -DENABLE_NOTIFICATIONS -DENABLE_HIDPI=1 -DENABLE_TOPCHROME_MD=1 -DUSE_UDEV -DDONT_EMBED_BUILD_METADATA -DDCHECK_ALWAYS_ON=1 -DFIELDTRIAL_TESTING_ENABLED -DENABLE_TASK_MANAGER=1 -DENABLE_EXTENSIONS=1 -DENABLE_PLUGINS=1 -DENABLE_SESSION_SERVICE=1 -DENABLE_THEMES=1 -DENABLE_AUTOFILL_DIALOG=1 -DENABLE_BACKGROUND=1 -DENABLE_GOOGLE_NOW=1 -DENABLE_PRINTING=1 -DENABLE_PRINT_PREVIEW=1 -DENABLE_SPELLCHECK=1 -DENABLE_CAPTIVE_PORTAL_DETECTION=1 -DENABLE_APP_LIST=1 -DENABLE_SUPERVISED_USERS=1 -DENABLE_MDNS=1 -DENABLE_SERVICE_DISCOVERY=1 -DV8_USE_EXTERNAL_STARTUP_DATA -DFULL_SAFE_BROWSING -DSAFE_BROWSING_CSD -DSAFE_BROWSING_DB_LOCAL -DSAFE_BROWSING_SERVICE -DGTEST_HAS_POSIX_RE=0 -DGTEST_LANG_CXX11=0 -DUNIT_TEST -DGTEST_HAS_RTTI=0 -DUSE_LIBPCI=1 -DUSE_OPENSSL=1 -DUSE_GLIB=1 -DUSE_NSS_CERTS=1 -DOS_CHROMEOS=1 -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -DDYNAMIC_ANNOTATIONS_ENABLED=1 -DWTF_USE_DYNAMIC_ANNOTATIONS=1 -D_DEBUG -D_GLIBCXX_DEBUG=1 -Igen -I../.. -I../../skia/config -I../../testing/gmock/include -I../../testing/gtest/include -Werror -pthread -fno-strict-aliasing -Wall -Wno-unused-parameter -Wno-missing-field-initializers -fvisibility=hidden -pipe -fPIC -Wno-reserved-user-defined-literal -Xclang -load -Xclang /b/build/slave/linux_chromeos/build/src/third_party/llvm-build/Release+Asserts/lib/libFindBadConstructs.so -Xclang -add-plugin -Xclang find-bad-constructs -Xclang -plugin-arg-find-bad-constructs -Xclang check-templates -fcolor-diagnostics -B/b/build/slave/linux_chromeos/build/src/third_party/binutils/Linux_x64/Release/bin -Wheader-hygiene -Wno-char-subscripts -Wno-unneeded-internal-declaration -Wno-covered-switch-default -Wstring-conversion -Wno-c++11-narrowing -Wno-deprecated-register -Wno-inconsistent-missing-override -Wno-shift-negative-value -pthread -I/usr/include/glib-2.0 -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -m64 -march=x86-64 -O0 -g -funwind-tables -gsplit-dwarf -Wno-undefined-bool-conversion -Wno-tautological-undefined-compare -g0 -fno-exceptions -fno-rtti -fno-threadsafe-statics -fvisibility-inlines-hidden -Wsign-compare -std=gnu++11  -c ../../components/user_manager/fake_user_manager.cc -o obj/components/user_manager/user_manager_test_support.fake_user_manager.o
In file included from ../../components/user_manager/fake_user_manager.cc:5:
In file included from ../../components/user_manager/fake_user_manager.h:12:
In file included from ../../components/user_manager/user.h:14:
In file included from ../../components/user_manager/user_image/user_image.h:12:
In file included from ../../ui/gfx/image/image_skia.h:15:
In file included from ../../ui/gfx/image/image_skia_rep.h:8:
In file included from ../../third_party/skia/include/core/SkBitmap.h:11:
In file included from ../../third_party/skia/include/core/SkColor.h:11:
In file included from ../../third_party/skia/include/core/SkScalar.h:12:
../../third_party/skia/include/core/../private/SkFloatingPoint.h:13:10: fatal error: 'SkTypes.h' file not found
#include "SkTypes.h"

ninja -w dupbuild=err -C /b/build/slave/linux_chromeos/build/src/out/Debug accessibility_unittests accessibility_unittests_run app_list_unittests app_list_unittests_run app_shell_unittests app_shell_unittests_run ash_unittests ash_unittests_run aura_builder aura_unittests aura_unittests_run base_unittests base_unittests_run browser_tests browser_tests_run cacheinvalidation_unittests cacheinvalidation_unittests_run chromeos_unittests chromeos_unittests_run components_unittests components_unittests_run compositor_unittests compositor_unittests_run content_browsertests content_browsertests_run content_unittests content_unittests_run crypto_unittests crypto_unittests_run dbus_unittests device_unittests device_unittests_run display_unittests display_unittests_run events_unittests events_unittests_run extensions_browsertests extensions_browsertests_run extensions_unittests extensions_unittests_run gcm_unit_tests gcm_unit_tests_run gfx_unittests gfx_unittests_run google_apis_unittests google_apis_unittests_run gpu_unittests gpu_unittests_run interactive_ui_tests interactive_ui_tests_run ipc_tests ipc_tests_run jingle_unittests jingle_unittests_run media_unittests media_unittests_run message_center_unittests message_center_unittests_run midi_unittests midi_unittests_run nacl_helper_nonsfi_unittests nacl_loader_unittests nacl_loader_unittests_run net_unittests net_unittests_run ppapi_unittests printing_unittests printing_unittests_run remoting_unittests remoting_unittests_run sandbox_linux_unittests sandbox_linux_unittests_run skia_unittests skia_unittests_run sql_unittests sql_unittests_run sync_unit_tests sync_unit_tests_run ui_base_unittests ui_base_unittests_run ui_chromeos_unittests ui_touch_selection_unittests ui_touch_selection_unittests_run unit_tests unit_tests_run url_unittests url_unittests_run views_unittests wm_unittests wm_unittests_run -j50
ninja: Entering directory `/b/build/slave/linux_chromeos/build/src/out/Debug'

Review URL: https://codereview.chromium.org/1288863002

8 years agoRevert of Normalize SkXfermode_opts.h argument order as d,s[,aa]. (patchset #1 id...
mtklein [Wed, 12 Aug 2015 20:51:27 +0000 (13:51 -0700)]
Revert of Normalize SkXfermode_opts.h argument order as d,s[,aa]. (patchset #1 id:1 of https://codereview.chromium.org/1289903002/ )

Reason for revert:
?

Original issue's description:
> Normalize SkXfermode_opts.h argument order as d,s[,aa].
>
> At head they're s,d[,aa] in SkXfermode_opts.h but Sk4px::Map* expect d,s[,aa]
> so we ended up having to write weird little lambda shims to match impedance.
>
> There's no reason for these to disagree, and d,s[,aa] is the One True Order
> (because no matter what you're doing in graphics, there's always a dst).
>
> Should be no perf or image diff, though I'm suspicious it might help MSVC code generation.
>
> BUG=skia:4117
>
> Committed: https://skia.googlesource.com/skia/+/6028a8476504022fe40b6870b1460b5e4a80969f

TBR=djsollen@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:4117

Review URL: https://codereview.chromium.org/1284363002