platform/upstream/libSkiaSharp.git
10 years agoRebaseline strokes_round and strokes3 on SGX devices
bsalomon@google.com [Thu, 24 Oct 2013 20:03:27 +0000 (20:03 +0000)]
Rebaseline strokes_round and strokes3 on SGX devices

BUG=skia:1752

Review URL: https://codereview.chromium.org/41083003

git-svn-id: http://skia.googlecode.com/svn/trunk@11953 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoChanges bench_graph_svg.py test to check_bench_regressions.py test; deletes deprecate...
bensong@google.com [Thu, 24 Oct 2013 18:08:45 +0000 (18:08 +0000)]
Changes bench_graph_svg.py test to check_bench_regressions.py test; deletes deprecated chunks in syn_bucket_subdir.
Review URL: https://codereview.chromium.org/27377002

git-svn-id: http://skia.googlecode.com/svn/trunk@11952 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoTile large bitmaps that are clipped.
bsalomon@google.com [Thu, 24 Oct 2013 17:52:07 +0000 (17:52 +0000)]
Tile large bitmaps that are clipped.

R=robertphillips@google.com

Review URL: https://codereview.chromium.org/31033002

git-svn-id: http://skia.googlecode.com/svn/trunk@11951 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoDM: print cache stats after every GPU GM when GR_CACHE_STATS is set
commit-bot@chromium.org [Thu, 24 Oct 2013 17:44:43 +0000 (17:44 +0000)]
DM: print cache stats after every GPU GM when GR_CACHE_STATS is set

BUG=
R=bsalomon@google.com

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/39743003

git-svn-id: http://skia.googlecode.com/svn/trunk@11950 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoWe want to give SkPixelRef a way to signal over to GrResourceCache that it's become...
commit-bot@chromium.org [Thu, 24 Oct 2013 17:44:27 +0000 (17:44 +0000)]
We want to give SkPixelRef a way to signal over to GrResourceCache that it's become pointless to keep around textures based on that SkPixelRef when its pixels change, so that it can be a good citizen and free those textures.

This adds an invalidation listener mechanism to SkPixelRef to let it send this message while still staying ignorant of who's listening.

These messages are tricky to deliver.  The SkPixelRefs they originates from and the GrResourceCaches they ultimately end up at may be on different threads; neither class is threadsafe; their object lifetimes are totally independent; it's a many-senders-to-many-receivers relation; and neither codebase should really know about the other.

So I've added a per-message-type global message bus to broadcast messages to threadsafe inboxes.  Anyone can post() a message, which will show up in all the inboxes of that type, read whenever the inbox's owner calls poll().  The implementation is _dumb_; it can be improved in several dimensions (inbox size limits, lock-free message delivery) if we find the need.

I took some care to make sure not to send the invalidation message for any SkPixelRef that's sharing a generation ID with another SkPixelRef.

BUG=
R=bsalomon@google.com, scroggo@google.com, reed@google.com

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/26734003

git-svn-id: http://skia.googlecode.com/svn/trunk@11949 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years ago[PDF] Rebaseline mac expectations after fixing image unpremtuliplication in r11945
vandebo@chromium.org [Thu, 24 Oct 2013 16:40:12 +0000 (16:40 +0000)]
[PDF] Rebaseline mac expectations after fixing image unpremtuliplication in r11945

Review URL: https://codereview.chromium.org/40263003

git-svn-id: http://skia.googlecode.com/svn/trunk@11948 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix drawBitmap() of scratch texture.
senorblanco@chromium.org [Thu, 24 Oct 2013 15:59:31 +0000 (15:59 +0000)]
Fix drawBitmap() of scratch texture.

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/40163004

git-svn-id: http://skia.googlecode.com/svn/trunk@11947 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agorebaseline_server: cleanup of HTML/Javascript
epoger@google.com [Thu, 24 Oct 2013 15:38:27 +0000 (15:38 +0000)]
rebaseline_server: cleanup of HTML/Javascript

No behavioral change at all.
(SkipBuildbotRuns)

R=borenet@google.com

Review URL: https://codereview.chromium.org/31583007

git-svn-id: http://skia.googlecode.com/svn/trunk@11946 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years ago[PDF] Add unpremultiply support and a GM (try5)
vandebo@chromium.org [Thu, 24 Oct 2013 15:20:00 +0000 (15:20 +0000)]
[PDF] Add unpremultiply support and a GM (try5)

This is a resubmit of https://codereview.chromium.org/22831039 which was
reverted because it causes issues on Windows.

BUG=chromium:175548
R=edisonn@google.com

Review URL: https://codereview.chromium.org/37873003

git-svn-id: http://skia.googlecode.com/svn/trunk@11945 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoWhitespace change to trigger buildbots
epoger@google.com [Thu, 24 Oct 2013 15:05:13 +0000 (15:05 +0000)]
Whitespace change to trigger buildbots

git-svn-id: http://skia.googlecode.com/svn/trunk@11944 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAdd note to ignored-tests.txt
epoger@google.com [Thu, 24 Oct 2013 14:42:19 +0000 (14:42 +0000)]
Add note to ignored-tests.txt
BUG=skia:1748
(SkipBuildbotRuns)

R=scroggo@google.com

Review URL: https://codereview.chromium.org/40353002

git-svn-id: http://skia.googlecode.com/svn/trunk@11943 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoupdate pdf-mac gms, after adding duport for perspective shaders in pdf
edisonn@google.com [Thu, 24 Oct 2013 14:24:59 +0000 (14:24 +0000)]
update pdf-mac gms, after adding duport for perspective shaders in pdf

Review URL: https://codereview.chromium.org/40313002

git-svn-id: http://skia.googlecode.com/svn/trunk@11942 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoBorrow Android's ARMv6 memory barrier for SkOnce.
commit-bot@chromium.org [Thu, 24 Oct 2013 14:25:22 +0000 (14:25 +0000)]
Borrow Android's ARMv6 memory barrier for SkOnce.

BUG=skia:1743
R=bungeman@google.com

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/34353009

git-svn-id: http://skia.googlecode.com/svn/trunk@11941 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoMinor changes
robertphillips@google.com [Thu, 24 Oct 2013 14:03:01 +0000 (14:03 +0000)]
Minor changes

https://codereview.chromium.org/39973004/

git-svn-id: http://skia.googlecode.com/svn/trunk@11940 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAlways create an instance when testing GrEffects for SkXfermode and SkArithmeticMode
commit-bot@chromium.org [Thu, 24 Oct 2013 13:33:32 +0000 (13:33 +0000)]
Always create an instance when testing GrEffects for SkXfermode and SkArithmeticMode

Previously the TestCreate methods of the effects would create a static
instance and return that for all invocations. This is probably a
copy-paste error going back to initial TestCreate method in r8449. At
that time, the effect was always the same and thus static usage made
sense.

R=bsalomon@google.com, senorblanco@google.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/39393002

git-svn-id: http://skia.googlecode.com/svn/trunk@11939 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix ModeColorFilterEffect assert during GLProgramsTest
commit-bot@chromium.org [Thu, 24 Oct 2013 13:29:38 +0000 (13:29 +0000)]
Fix ModeColorFilterEffect assert during GLProgramsTest

Select random filter mode for the test-case until the mode is not
Dst. SkModeColorFilter with Dst mode will not create a
ModeColorFilterEffect, and thus the case is not implemented.

Also changes it so that each call to ModeColorFilterEffect::TestCreate
creates a new GrEffect. Previously only one static effect was ever
created, due to copy-paste error.

R=bsalomon@google.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/39403002

git-svn-id: http://skia.googlecode.com/svn/trunk@11938 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoPDF: support perspective in simple shaders. (this version does not work well with...
edisonn@google.com [Thu, 24 Oct 2013 13:19:28 +0000 (13:19 +0000)]
PDF: support perspective in simple shaders. (this version does not work well with tilling)

R=vandebo@chromium.org

Review URL: https://codereview.chromium.org/26389006

git-svn-id: http://skia.googlecode.com/svn/trunk@11937 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoReplacing SkTDArray with SkAutoSTArray to reduce memory allocation overhead in SkTile...
commit-bot@chromium.org [Thu, 24 Oct 2013 11:24:38 +0000 (11:24 +0000)]
Replacing SkTDArray with SkAutoSTArray to reduce memory allocation overhead in SkTileGrid::search

BUG=https://code.google.com/p/skia/issues/detail?id=1735
R=tomhudson@chromium.org

Author: junov@chromium.org

Review URL: https://codereview.chromium.org/35633002

git-svn-id: http://skia.googlecode.com/svn/trunk@11936 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoExpose SkPicture::willPlayBackBitmaps()
tomhudson@google.com [Thu, 24 Oct 2013 11:12:47 +0000 (11:12 +0000)]
Expose SkPicture::willPlayBackBitmaps()

This returns true if (1) the picture has finished recording and
(2) this picture or any picture drawn into it refers to any bitmaps.
It allows clients doing complicated manipulations of the picture to
early-out when there are no bitmaps present.

BUG=303281
R=reed@google.com

git-svn-id: http://skia.googlecode.com/svn/trunk@11935 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Thu, 24 Oct 2013 07:01:59 +0000 (07:01 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@11934 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoApply matrix early in draw bitmap
commit-bot@chromium.org [Thu, 24 Oct 2013 01:46:11 +0000 (01:46 +0000)]
Apply matrix early in draw bitmap

Committed: http://code.google.com/p/skia/source/detail?r=11930

Reverted: https://code.google.com/p/skia/source/detail?r=11932

R=robertphillips@google.com, senorblanco@chromium.org

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/30593003

git-svn-id: http://skia.googlecode.com/svn/trunk@11933 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "Apply matrix early in draw bitmap"
scroggo@google.com [Wed, 23 Oct 2013 22:19:06 +0000 (22:19 +0000)]
Revert "Apply matrix early in draw bitmap"

This reverts commit 4c446c8eb50b7a6e6f917fae7b4c4ffe5658a949.

Fix build.

git-svn-id: http://skia.googlecode.com/svn/trunk@11932 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "[PDF] Add unpremultiply support and a GM (try4)"
vandebo@chromium.org [Wed, 23 Oct 2013 21:36:43 +0000 (21:36 +0000)]
Revert "[PDF] Add unpremultiply support and a GM (try4)"

This reverts commit r11928

gm seems to crash on windows.

Review URL: https://codereview.chromium.org/38153003

git-svn-id: http://skia.googlecode.com/svn/trunk@11931 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoApply matrix early in draw bitmap
commit-bot@chromium.org [Wed, 23 Oct 2013 21:06:51 +0000 (21:06 +0000)]
Apply matrix early in draw bitmap

R=robertphillips@google.com, senorblanco@chromium.org

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/30593003

git-svn-id: http://skia.googlecode.com/svn/trunk@11930 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years ago[PDF] Add unpremultiply support and a GM (try4)
vandebo@chromium.org [Wed, 23 Oct 2013 20:17:29 +0000 (20:17 +0000)]
[PDF] Add unpremultiply support and a GM (try4)

This is a resubmit of https://codereview.chromium.org/22831039 which was
reverted because it causes issues on Windows (really, release builds).

BUG=chromium:175548
R=edisonn@google.com

Review URL: https://codereview.chromium.org/33493003

git-svn-id: http://skia.googlecode.com/svn/trunk@11928 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agorebaseline scaled tilemodes gms on the GPU
humper@google.com [Wed, 23 Oct 2013 19:41:08 +0000 (19:41 +0000)]
rebaseline scaled tilemodes gms on the GPU

git-svn-id: http://skia.googlecode.com/svn/trunk@11927 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSplit up SkXfermode::asNewEffectOrCoeff() into asNewEffect(), asCoeff().
senorblanco@chromium.org [Wed, 23 Oct 2013 18:41:36 +0000 (18:41 +0000)]
Split up SkXfermode::asNewEffectOrCoeff() into asNewEffect(), asCoeff().

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/37593002

git-svn-id: http://skia.googlecode.com/svn/trunk@11926 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFollow up to serialization validation code
commit-bot@chromium.org [Wed, 23 Oct 2013 18:33:18 +0000 (18:33 +0000)]
Follow up to serialization validation code

1 ) Added check for bool to make sure is it either 0 or 1 and not garbage

2 ) Added more solid kernel size checks in SkMatrixConvolutionImageFilter

3 ) Make sure array size is validated in SkMergeImageFilter

BUG=
R=reed@google.com, mtklein@google.com, senorblanco@google.com, senorblanco@chromium.org

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/23548034

git-svn-id: http://skia.googlecode.com/svn/trunk@11925 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoInitial error handling code
commit-bot@chromium.org [Wed, 23 Oct 2013 17:06:21 +0000 (17:06 +0000)]
Initial error handling code

I made it as simple as possible. The impact seems minimal and it should do what's necessary to make this code secure.

BUG=

Committed: http://code.google.com/p/skia/source/detail?r=11247

R=reed@google.com, scroggo@google.com, djsollen@google.com, sugoi@google.com, bsalomon@google.com, mtklein@google.com, senorblanco@google.com, senorblanco@chromium.org

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/23021015

git-svn-id: http://skia.googlecode.com/svn/trunk@11922 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agouse the supplied matrix instead of the texture w/h div for the bicubic GPU effect
commit-bot@chromium.org [Wed, 23 Oct 2013 16:18:50 +0000 (16:18 +0000)]
use the supplied matrix instead of the texture w/h div for the bicubic GPU effect

BUG=
R=bsalomon@google.com, caryclark@google.com

Author: humper@google.com

Review URL: https://codereview.chromium.org/37003005

git-svn-id: http://skia.googlecode.com/svn/trunk@11921 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoLighting filter uses input on raster path
commit-bot@chromium.org [Wed, 23 Oct 2013 15:33:44 +0000 (15:33 +0000)]
Lighting filter uses input on raster path

BUG=
R=senorblanco@chromium.org

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/35703002

git-svn-id: http://skia.googlecode.com/svn/trunk@11918 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAdd support for reading non-rendertarget textures.
commit-bot@chromium.org [Wed, 23 Oct 2013 15:32:39 +0000 (15:32 +0000)]
Add support for reading non-rendertarget textures.

Also removes rendertarget flag from GrAtlas texture creation (no
longer needed) and re-enables GrFontCache::dump().

R=robertphillips@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/29263004

git-svn-id: http://skia.googlecode.com/svn/trunk@11917 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoTSAN build fix: forgot to pass the address into ANNOTATE_BENIGN_RACE.
mtklein@google.com [Wed, 23 Oct 2013 15:24:27 +0000 (15:24 +0000)]
TSAN build fix: forgot to pass the address into ANNOTATE_BENIGN_RACE.

BUG=

Review URL: https://codereview.chromium.org/37403002

git-svn-id: http://skia.googlecode.com/svn/trunk@11916 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agorebaseline_server: add tabs, and ability to submit new baselines to the server
epoger@google.com [Wed, 23 Oct 2013 15:07:26 +0000 (15:07 +0000)]
rebaseline_server: add tabs, and ability to submit new baselines to the server

Tabs allow the user to divide the tests into groups:
hide these for now, approve these, etc.

(SkipBuildbotRuns)

R=borenet@google.com

Review URL: https://codereview.chromium.org/28903008

git-svn-id: http://skia.googlecode.com/svn/trunk@11915 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSK_ONCE for SkData and SkPathRef
commit-bot@chromium.org [Wed, 23 Oct 2013 14:44:08 +0000 (14:44 +0000)]
SK_ONCE for SkData and SkPathRef

Adds SK_ONCE_FRIEND, to allow SK_DEF_ONCE code to be friends with a class.  This had to go in include/core to be visible to headers there.

BUG=
R=reed@google.com, bungeman@google.com

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/26491003

git-svn-id: http://skia.googlecode.com/svn/trunk@11914 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRemove GrContext from SkXfermode::[Aa]sNewEffectOrCoeff() and all subclasses,
senorblanco@chromium.org [Wed, 23 Oct 2013 14:03:22 +0000 (14:03 +0000)]
Remove GrContext from SkXfermode::[Aa]sNewEffectOrCoeff() and all subclasses,
since it's unused.

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/31853003

git-svn-id: http://skia.googlecode.com/svn/trunk@11913 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoImplement SkColorFilter as a GrGLEffect
commit-bot@chromium.org [Wed, 23 Oct 2013 05:42:03 +0000 (05:42 +0000)]
Implement SkColorFilter as a GrGLEffect

Adds GrEffect::willUseInputColor() which indicates whether or not the
input color affects the output of the effect. This is needed for
certain Xfermodes, such as kSrc_Mode. For these modes the color filter
will not use the input color.

An effect with GrEffect::willUseInputColor() true will cause all color
or coverage effects before it to be discarded, as their computations
cannot affect the output. In these cases program is marked as having
white input color.

This fixes an assert when Skia is compiled in a mode that prefers
using uniforms instead of attributes for constants. (Flags
GR_GL_USE_NV_PATH_RENDERING or GR_GL_NO_CONSTANT_ATTRIBUTES). Using
attributes hides the problem where the fragment shader does not need
input color for color filters that ignore DST part of the filter. The
assert would be hit when uniform manager tries to bind an uniform which
has been optimized away by the shader compiler.

Adds specific GrGLSLExpr4 and GrGLSLExpr1 classes. This way the GLSL
expressions like "(v - src.a)" can remain somewhat readable in form of
"(v - src.a())". The GrGLSLExpr<typename> template implements the
generic functionality, GrGLSLExprX is the specialization that exposes
the type-safe interface to this functionality.

Also adds operators so that GLSL binary operators of the form
"(float * vecX)" can be expressed in C++. Before only the equivalent
"(vecX * float)" was possible. This reverts the common blending
calculations to more conventional order, such as "(1-a) * c" instead of
"c * (1-a)".

Changes GrGLSLExpr1::OnesStr from 1 to 1.0 in order to preserve the
color filter blending formula string the same (with the exception of
variable name change).

Shaders change in case of input color being needed:
 -   vec4 filteredColor;
 -   filteredColor = (((1.0 - uFilterColor.a) * output_Stage0) + uFilterColor);
 -   fsColorOut = filteredColor;
 +   vec4 output_Stage1;
 +   { // Stage 1: ModeColorFilterEffect
 +   output_Stage1 = (((1.0 - uFilterColor_Stage1.a) * output_Stage0) + uFilterColor_Stage1);
 +   }
 +   fsColorOut = output_Stage1;

Shaders change in case of input color being not needed:
 -uniform vec4 uFilterColor;
 -in vec4 vColor;
 +uniform vec4 uFilterColor_Stage0;
  out vec4 fsColorOut;
  void main() {
 -   vec4 filteredColor;
 -   filteredColor = uFilterColor;
 -   fsColorOut = filteredColor;
 +   vec4 output_Stage0;
 +   { // Stage 0: ModeColorFilterEffect
 +   output_Stage0 = uFilterColor_Stage0;
 +   }
 +   fsColorOut = output_Stage0;
  }

R=bsalomon@google.com, robertphillips@google.com, jvanverth@google.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/25023003

git-svn-id: http://skia.googlecode.com/svn/trunk@11912 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoNew gm expectations for imagemagnifier
sugoi@google.com [Tue, 22 Oct 2013 19:42:10 +0000 (19:42 +0000)]
New gm expectations for imagemagnifier

git-svn-id: http://skia.googlecode.com/svn/trunk@11911 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix a test warning on some linux machines
edisonn@google.com [Tue, 22 Oct 2013 18:33:21 +0000 (18:33 +0000)]
Fix a test warning on some linux machines

R=sugoi@google.com

Review URL: https://codereview.chromium.org/34803004

git-svn-id: http://skia.googlecode.com/svn/trunk@11910 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "If the path is a rect, call drawRect to raster the geometry in SkCanvas:...
bsalomon@google.com [Tue, 22 Oct 2013 16:54:15 +0000 (16:54 +0000)]
Revert "If the path is a rect, call drawRect to raster the geometry in SkCanvas::drawPath to get better performance."

This reverts commit r11904

Review URL: https://codereview.chromium.org/35543002

git-svn-id: http://skia.googlecode.com/svn/trunk@11909 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoif fBitmap is still NULL, return false
reed@google.com [Tue, 22 Oct 2013 16:17:29 +0000 (16:17 +0000)]
if fBitmap is still NULL, return false

git-svn-id: http://skia.googlecode.com/svn/trunk@11908 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoadd tests for SkDocument
commit-bot@chromium.org [Tue, 22 Oct 2013 14:49:03 +0000 (14:49 +0000)]
add tests for SkDocument

R=reed@google.com, vandebo@chromium.org

Author: edisonn@google.com

Review URL: https://codereview.chromium.org/33423002

git-svn-id: http://skia.googlecode.com/svn/trunk@11907 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAdd strokerect to ignored tests list
bsalomon@google.com [Tue, 22 Oct 2013 13:52:47 +0000 (13:52 +0000)]
Add strokerect to ignored tests list

Review URL: https://codereview.chromium.org/35223003

git-svn-id: http://skia.googlecode.com/svn/trunk@11906 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoIf the path is a rect, call drawRect to raster the geometry in SkCanvas::drawPath...
bsalomon@google.com [Tue, 22 Oct 2013 13:19:12 +0000 (13:19 +0000)]
If the path is a rect, call drawRect to raster the geometry in SkCanvas::drawPath to get better performance.

Committed: http://code.google.com/p/skia/source/detail?r=11842

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/23484007

git-svn-id: http://skia.googlecode.com/svn/trunk@11904 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Tue, 22 Oct 2013 07:02:30 +0000 (07:02 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@11903 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoMore work to integrate skimage with rebaseline tools.
scroggo@google.com [Tue, 22 Oct 2013 00:42:46 +0000 (00:42 +0000)]
More work to integrate skimage with rebaseline tools.

tools/skimage_main.cpp:
Add the ability to write the results to checksum based filenames,
much like GM uses. This will allow using the skpdiff server to
rebaseline images.
Write the keys in the JSON file as <original image>_<pref config>.png,
so it matches gm_json.IMAGE_FILENAME_PATTERN. Also replace '_' with
'-' in the original file name, to avoid confusing the pattern matcher.
The '_' to '-' replacement also happens on the output filename.
Read the keys in a similar manner.
In make_outname, no longer remove a suffix. This fixes a bug where
subset decoding writes multiple subsets to the same file.

tools/rebaseline.py:
Since the filenames written to json files now match
gm_json.IMAGE_FILENAME_PATTERN, enable the option to match based
on configs/tests when rebaselining skimage.

test json files:
Update to match the new format of output.

gm/gm_expectations:
Add a constructor that takes a BitmapAndDigest as input.

tools/tests/skimage_self_test.py:
Test that reading the expectations file just created by skimage with
the same file actually compares to the original file (rather than just
succeeding because expectations were missing).

Change the expectations files to match the new format.

Will require a buildbot change to use the new flag: https://codereview.chromium.org/27389002/

BUG=1466
R=epoger@google.com

Review URL: https://codereview.chromium.org/26297004

git-svn-id: http://skia.googlecode.com/svn/trunk@11902 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoXoom rebaselines.
senorblanco@chromium.org [Mon, 21 Oct 2013 22:00:47 +0000 (22:00 +0000)]
Xoom rebaselines.

Note: Some of these had no base versions in GS corresponding to the hashes in the JSON, causing the svndiff script to barf. I hacked it up to let me finish, but it should probably handle the case of a missing hash more gracefully.

TBR=epoger

Review URL: https://codereview.chromium.org/33303007

git-svn-id: http://skia.googlecode.com/svn/trunk@11900 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert r11880 (Apply matrix early in draw bitmap - https://codereview.chromium.org...
robertphillips@google.com [Mon, 21 Oct 2013 21:30:49 +0000 (21:30 +0000)]
Revert r11880 (Apply matrix early in draw bitmap - https://codereview.chromium.org/30593003) due to layout test failure

git-svn-id: http://skia.googlecode.com/svn/trunk@11899 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoGrTHashCache -> GrTHashTable
mtklein@google.com [Mon, 21 Oct 2013 21:04:06 +0000 (21:04 +0000)]
GrTHashCache -> GrTHashTable

The class is Table, but the file's Cache.  That's confusing.

BUG=
R=bsalomon@google.com

Review URL: https://codereview.chromium.org/32523004

git-svn-id: http://skia.googlecode.com/svn/trunk@11898 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoImage decoder fixes (mostly) around A8.
scroggo@google.com [Mon, 21 Oct 2013 20:47:31 +0000 (20:47 +0000)]
Image decoder fixes (mostly) around A8.

Opaquness:
When decoding Gray to A8 in libpng, set reallyHasAlpha to true
and add a comment why we do not check for the real answer.

Add comments in jpeg decoder explaining why A8 is not opaque.

Fix a bug where an A8 subset is considered to be opaque.

Other fixes:
In SkJPEGImageDecoder, only allocate as much memory as
needed for each source row, based on the input config.
Also pull out common code into a static function.

When performing the check for requiring unpremultiplied colors,
allow A8 to succeed, since that setting should have no effect on
A8.

Add the check for requiring unpremultiplied colors to subset
decoding.

Fix a bug where attempting to sample gray to A8 does not sample.

R=reed@google.com

Review URL: https://codereview.chromium.org/26210007

git-svn-id: http://skia.googlecode.com/svn/trunk@11897 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "Revert "cache SkImage::Info calculation in lazypixelref""
reed@google.com [Mon, 21 Oct 2013 20:13:42 +0000 (20:13 +0000)]
Revert "Revert "cache SkImage::Info calculation in lazypixelref""

This reverts commit 81eba32ab10f9210c742938819cf1218be5611c9.

ImageDecoder is changed to allow info to be NULL, since it is an output-only parameter.

R=scroggo@google.com

Review URL: https://codereview.chromium.org/33573002

git-svn-id: http://skia.googlecode.com/svn/trunk@11896 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "cache SkImage::Info calculation in lazypixelref"
reed@google.com [Mon, 21 Oct 2013 19:45:38 +0000 (19:45 +0000)]
Revert "cache SkImage::Info calculation in lazypixelref"

This reverts commit a2537480710aa10b9ab9f8aec98538b79539f47e.

git-svn-id: http://skia.googlecode.com/svn/trunk@11894 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agocache SkImage::Info calculation in lazypixelref
reed@google.com [Mon, 21 Oct 2013 18:58:17 +0000 (18:58 +0000)]
cache SkImage::Info calculation in lazypixelref

BUG=
R=scroggo@google.com

Review URL: https://codereview.chromium.org/28493003

git-svn-id: http://skia.googlecode.com/svn/trunk@11893 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRoll gyp from r1684 to r1765.
bungeman@google.com [Mon, 21 Oct 2013 18:49:43 +0000 (18:49 +0000)]
Roll gyp from r1684 to r1765.

R=epoger@google.com

Review URL: https://codereview.chromium.org/32963002

git-svn-id: http://skia.googlecode.com/svn/trunk@11892 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoChromeOS and IntelRhb rebaselines.
senorblanco@chromium.org [Mon, 21 Oct 2013 18:46:49 +0000 (18:46 +0000)]
ChromeOS and IntelRhb rebaselines.

TBR=robertphillips

Review URL: https://codereview.chromium.org/32533004

git-svn-id: http://skia.googlecode.com/svn/trunk@11891 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoDM: add --serialize
commit-bot@chromium.org [Mon, 21 Oct 2013 18:40:25 +0000 (18:40 +0000)]
DM: add --serialize

Plus:
  - minor ReplayTask refactoring to share code with SerializeTask
  - move --replay to ReplayTask and --serialize to SerializeTask like WriteTask
  - when --writePath is given, write failures for Replay and Serialize tasks
  - function names have fewer blatant Skia style violations

BUG=
R=bsalomon@google.com

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/32613003

git-svn-id: http://skia.googlecode.com/svn/trunk@11890 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoThird round of matrixconvolution rebaselines, and removal from ignored-tests.txt.
senorblanco@chromium.org [Mon, 21 Oct 2013 18:23:30 +0000 (18:23 +0000)]
Third round of matrixconvolution rebaselines, and removal from ignored-tests.txt.

"We have normality. I repeat, we have normality. Anything you still can't cope with is therefore your own problem."

TBR=robertphillips
BUG=

Review URL: https://codereview.chromium.org/33323002

git-svn-id: http://skia.googlecode.com/svn/trunk@11889 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix out of order destruction error
robertphillips@google.com [Mon, 21 Oct 2013 17:37:28 +0000 (17:37 +0000)]
Fix out of order destruction error

https://codereview.chromium.org/33203002/

git-svn-id: http://skia.googlecode.com/svn/trunk@11888 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoGM: remove deprecated --ignoreTests flag
epoger@google.com [Mon, 21 Oct 2013 17:25:26 +0000 (17:25 +0000)]
GM: remove deprecated --ignoreTests flag
BUG=skia:1730
R=borenet@google.com

Review URL: https://codereview.chromium.org/27983004

git-svn-id: http://skia.googlecode.com/svn/trunk@11887 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAdd DPI stettings to SkDocument::CreatePDF(). Tests will be added in a future cl...
commit-bot@chromium.org [Mon, 21 Oct 2013 17:14:37 +0000 (17:14 +0000)]
Add DPI stettings to SkDocument::CreatePDF(). Tests will be added in a future cl, once DPI will be used in SkPDFDevice

R=reed@google.com, bungeman@google.com, vandebo@chromium.org

Author: edisonn@google.com

Review URL: https://codereview.chromium.org/32233003

git-svn-id: http://skia.googlecode.com/svn/trunk@11886 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSecond round of matrixconvolution rebaselines.
senorblanco@chromium.org [Mon, 21 Oct 2013 17:04:11 +0000 (17:04 +0000)]
Second round of matrixconvolution rebaselines.

TBR=robertphillips
BUG=

Review URL: https://codereview.chromium.org/33103002

git-svn-id: http://skia.googlecode.com/svn/trunk@11885 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoMore clang warning fixes.
commit-bot@chromium.org [Mon, 21 Oct 2013 16:41:00 +0000 (16:41 +0000)]
More clang warning fixes.

Mostly unused functions and variables removed.

BUG=None
TEST=ninja -C out/Debug most
     ninja -C out/Release most

R=bsalomon@google.com, caryclark@google.com, robertphillips@google.com

Author: tfarina@chromium.org

Review URL: https://codereview.chromium.org/27933002

git-svn-id: http://skia.googlecode.com/svn/trunk@11884 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoMagnifier gpu path fix
commit-bot@chromium.org [Mon, 21 Oct 2013 15:59:26 +0000 (15:59 +0000)]
Magnifier gpu path fix

The y axis was inverted for the offset, which should have been computed on the other side of the magnifying rectangle.

BUG=
R=bsalomon@google.com, senorblanco@google.com, senorblanco@chromium.org

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/28273005

git-svn-id: http://skia.googlecode.com/svn/trunk@11883 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFirst round of matrixconvolution rebaselines.
senorblanco@chromium.org [Mon, 21 Oct 2013 15:41:55 +0000 (15:41 +0000)]
First round of matrixconvolution rebaselines.

TBR=robertphillips

Review URL: https://codereview.chromium.org/32823002

git-svn-id: http://skia.googlecode.com/svn/trunk@11882 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix GrProgramsTest to not generate tests with too many TexCoord references
commit-bot@chromium.org [Mon, 21 Oct 2013 14:48:23 +0000 (14:48 +0000)]
Fix GrProgramsTest to not generate tests with too many TexCoord references

Make GrProgramsTest check how many texture coordinate sets are
available and select random effects up until the amount runs out.

Otherwise, following effect sequence would fail the shader compilation
when Skia is compiled with nv_path_rendering on (eg. when fixed
function codepath is used):
 * Stage 0: TextureDomain (1 texcoord)
 * Stage 1: Convolution (1 texcoord)
 * Stage 2: Bitmap Alpha Threshold (2 texcoords)
 * Stage 3: DisplacementMap (2 texcoords)
 * Stage 4: Config Conversion (1 texcoords)
 * Stage 5: Two-Point Conical Gradient (2 texcoords)

This would use more texture coordinate sets than 8, which is fairly
common amount currently.

R=bsalomon@google.com, cdalton@nvidia.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/32403002

git-svn-id: http://skia.googlecode.com/svn/trunk@11881 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoApply matrix early in draw bitmap
commit-bot@chromium.org [Mon, 21 Oct 2013 14:48:17 +0000 (14:48 +0000)]
Apply matrix early in draw bitmap

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/30593003

git-svn-id: http://skia.googlecode.com/svn/trunk@11880 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoscope ComputeIsOpaque with SkBitmap
reed@google.com [Mon, 21 Oct 2013 14:31:20 +0000 (14:31 +0000)]
scope ComputeIsOpaque with SkBitmap

BUG=

Review URL: https://codereview.chromium.org/32653002

git-svn-id: http://skia.googlecode.com/svn/trunk@11879 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agochange ERROR label to BAD_CONFIG -- avoid windows complaint
reed@google.com [Mon, 21 Oct 2013 14:12:13 +0000 (14:12 +0000)]
change ERROR label to BAD_CONFIG -- avoid windows complaint

BUG=

Review URL: https://codereview.chromium.org/32623002

git-svn-id: http://skia.googlecode.com/svn/trunk@11878 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agostore SkAlphaType inside SkBitmap, on road to support unpremul
reed@google.com [Mon, 21 Oct 2013 14:00:07 +0000 (14:00 +0000)]
store SkAlphaType inside SkBitmap, on road to support unpremul

BUG=
R=bsalomon@google.com, scroggo@google.com

Review URL: https://codereview.chromium.org/25275004

git-svn-id: http://skia.googlecode.com/svn/trunk@11877 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoReverting revert of r11862 (Constify SkFlattenable::getFactory()) or reverting r11871.
robertphillips@google.com [Mon, 21 Oct 2013 12:26:10 +0000 (12:26 +0000)]
Reverting revert of r11862 (Constify SkFlattenable::getFactory()) or reverting r11871.

git-svn-id: http://skia.googlecode.com/svn/trunk@11876 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoA few stragglers for r1187 (Fix cubic->quad RR corner regression)
robertphillips@google.com [Mon, 21 Oct 2013 00:15:15 +0000 (00:15 +0000)]
A few stragglers for r1187 (Fix cubic->quad RR corner regression)

https://codereview.chromium.org/31333002/

git-svn-id: http://skia.googlecode.com/svn/trunk@11875 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoMore rebaselines for r1187 (Fix cubic->quad RR corner regression)
robertphillips@google.com [Sun, 20 Oct 2013 23:53:11 +0000 (23:53 +0000)]
More rebaselines for r1187 (Fix cubic->quad RR corner regression)

https://codereview.chromium.org/31323002/

git-svn-id: http://skia.googlecode.com/svn/trunk@11874 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agorebaselines after r1187 (Fix cubic->quad RR corner regression)
robertphillips@google.com [Sun, 20 Oct 2013 23:06:40 +0000 (23:06 +0000)]
rebaselines after r1187 (Fix cubic->quad RR corner regression)

https://codereview.chromium.org/31313002/

git-svn-id: http://skia.googlecode.com/svn/trunk@11873 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix for cropped matrix convolution for BottomLeft render targets.
senorblanco@chromium.org [Sun, 20 Oct 2013 20:21:19 +0000 (20:21 +0000)]
Fix for cropped matrix convolution for BottomLeft render targets.

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/31083002

git-svn-id: http://skia.googlecode.com/svn/trunk@11872 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoReverted due to Chromium compilation failures. Please see:
robertphillips@google.com [Sun, 20 Oct 2013 13:34:21 +0000 (13:34 +0000)]
Reverted due to Chromium compilation failures. Please see:

https://codereview.chromium.org/30973002/

git-svn-id: http://skia.googlecode.com/svn/trunk@11871 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoFix cubic->quad RR corner regression
robertphillips@google.com [Sun, 20 Oct 2013 13:11:16 +0000 (13:11 +0000)]
Fix cubic->quad RR corner regression

https://codereview.chromium.org/29673002/

git-svn-id: http://skia.googlecode.com/svn/trunk@11870 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Sat, 19 Oct 2013 07:02:35 +0000 (07:02 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@11869 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "[PDF] Add unpremultiply support and a GM (try3)."
vandebo@chromium.org [Fri, 18 Oct 2013 21:04:51 +0000 (21:04 +0000)]
Revert "[PDF] Add unpremultiply support and a GM (try3)."

This is crashing in some configurations, can now repro.

This reverts commit 12add177d0ab7bbac97f5093c9382924d88bde8c.

Review URL: https://codereview.chromium.org/29903002

git-svn-id: http://skia.googlecode.com/svn/trunk@11868 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoDM: add --writePath
mtklein@google.com [Fri, 18 Oct 2013 20:52:44 +0000 (20:52 +0000)]
DM: add --writePath

For fun, make the output prettier, hiding ", N failures" unless there is one.

BUG=
R=bsalomon@google.com

Review URL: https://codereview.chromium.org/29293003

git-svn-id: http://skia.googlecode.com/svn/trunk@11867 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years ago[PDF] Add unpremultiply support and a GM (try3).
vandebo@chromium.org [Fri, 18 Oct 2013 20:18:29 +0000 (20:18 +0000)]
[PDF] Add unpremultiply support and a GM (try3).

This is a resubmit of https://codereview.chromium.org/22831039 which was
reverted because it causes issues on Windows.

BUG=chromium:175548
R=edisonn@google.com

Review URL: https://codereview.chromium.org/29193003

git-svn-id: http://skia.googlecode.com/svn/trunk@11866 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "PDF: support perspective in simple shaders. (this version does not work well...
vandebo@chromium.org [Fri, 18 Oct 2013 20:16:34 +0000 (20:16 +0000)]
Revert "PDF: support perspective in simple shaders. (this version does not work well with tilling)"

Accidentally committed the wrong git branch.

This reverts commit 76839c65670df8a863041a82a4b61502e78af043.

Review URL: https://codereview.chromium.org/29743002

git-svn-id: http://skia.googlecode.com/svn/trunk@11865 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoPDF: support perspective in simple shaders. (this version does not work well with...
vandebo@chromium.org [Fri, 18 Oct 2013 20:14:08 +0000 (20:14 +0000)]
PDF: support perspective in simple shaders. (this version does not work well with tilling)

Review URL: https://codereview.chromium.org/26389006

git-svn-id: http://skia.googlecode.com/svn/trunk@11864 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoImplement crop rect support for SkMatrixConvolutionImageFilter.
senorblanco@chromium.org [Fri, 18 Oct 2013 20:08:14 +0000 (20:08 +0000)]
Implement crop rect support for SkMatrixConvolutionImageFilter.

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/27471002

git-svn-id: http://skia.googlecode.com/svn/trunk@11863 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoConstify SkFlattenable::getFactory().
commit-bot@chromium.org [Fri, 18 Oct 2013 19:37:55 +0000 (19:37 +0000)]
Constify SkFlattenable::getFactory().

This allows us to also constify the argument of
SkFlattenableWriteBuffer::writeFlattenable() - which I've recently found
myself const-casting for no apparent good reason.

R=bungeman@google.com, mtklein@google.com, reed@google.com

Author: fmalita@chromium.org

Review URL: https://codereview.chromium.org/29143005

git-svn-id: http://skia.googlecode.com/svn/trunk@11862 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert r11860, which was supposed to just be a whitespace change (argh)
epoger@google.com [Fri, 18 Oct 2013 18:36:25 +0000 (18:36 +0000)]
Revert r11860, which was supposed to just be a whitespace change (argh)
(SkipBuildbotRuns)

Review URL: https://codereview.chromium.org/29603002

git-svn-id: http://skia.googlecode.com/svn/trunk@11861 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoWhitespace change to trigger buildbots after https://code.google.com/p/skia/source...
epoger@google.com [Fri, 18 Oct 2013 18:33:03 +0000 (18:33 +0000)]
Whitespace change to trigger buildbots after https://code.google.com/p/skia/source/detail?r=11859

git-svn-id: http://skia.googlecode.com/svn/trunk@11860 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert "[PDF] Add unpremultiply support and a GM."
vandebo@chromium.org [Fri, 18 Oct 2013 16:10:03 +0000 (16:10 +0000)]
Revert "[PDF] Add unpremultiply support and a GM."

This reverts commit f746273b8ba282f548a6eaffb45b11b6f468aa8e.

Review URL: https://codereview.chromium.org/29343002

git-svn-id: http://skia.googlecode.com/svn/trunk@11858 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years ago[PDF] Add unpremultiply support and a GM.
vandebo@chromium.org [Fri, 18 Oct 2013 15:53:03 +0000 (15:53 +0000)]
[PDF] Add unpremultiply support and a GM.

This is a resubmit of https://codereview.chromium.org/22831039 which was
reverted because it causes issues on Windows.

BUG=chromium:175548
R=edisonn@google.com

Review URL: https://codereview.chromium.org/27518003

git-svn-id: http://skia.googlecode.com/svn/trunk@11857 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agorebaseline gpu complex clip tests
bsalomon@google.com [Fri, 18 Oct 2013 15:47:28 +0000 (15:47 +0000)]
rebaseline gpu complex clip tests

Review URL: https://codereview.chromium.org/29303002

git-svn-id: http://skia.googlecode.com/svn/trunk@11856 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoGM: add --ignoreFailuresFile as an alternative to --ignoreTests
epoger@google.com [Fri, 18 Oct 2013 14:55:47 +0000 (14:55 +0000)]
GM: add --ignoreFailuresFile as an alternative to --ignoreTests

(moved from old review at https://codereview.chromium.org/27752002/ )

This will allow the user to specify the tests to ignore failure on as a file,
rather than listing them out in the command line (and thus busting the
line length limit on Android).
BUG=skia:1730
R=scroggo@google.com

Review URL: https://codereview.chromium.org/29203002

git-svn-id: http://skia.googlecode.com/svn/trunk@11855 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSkThreadPool: allow for Runnables that add other Runnables to the pool.
commit-bot@chromium.org [Fri, 18 Oct 2013 14:19:19 +0000 (14:19 +0000)]
SkThreadPool: allow for Runnables that add other Runnables to the pool.

There's a scenario that we're currently not allowing for, but I'd really like to use in DM:

1) client calls add(SomeRunnable*) several times
2) client calls wait()
3) any of the runnables added by the client _themselves_ call add(SomeOtherRunnable*)
4-inf) maybe those SomeOtherRunnables too call add(SomeCrazyThirdRunnable*), etc.

Right now in this scenario we'll assert in debug mode in step 3) when we call
add() and we're waiting to stop, and do strange unspecified things in release
mode.

The old threadpool had basically two states: running, and waiting to stop.  If
a thread saw we were waiting to stop and the queue was empty, that thread shut
down.  This wasn't accounting for any work that other threads might be doing;
potentially they were about to add to the queue.

So now we have three states: running, waiting, and halting.  When the client
calls wait() (or the destructor triggers), we move into waiting.  When a thread
notices we're _really_ done, that is, have an empty queue and there are no
active threads, we move into halting.  The halting state actually triggers the
threads to stop, which wait() is patiently join()ing on.

BUG=
R=bungeman@google.com, bsalomon@google.com

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/26389005

git-svn-id: http://skia.googlecode.com/svn/trunk@11852 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Fri, 18 Oct 2013 07:01:59 +0000 (07:01 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@11850 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoMake SkImageFilter::asNewEffect() (and all derived-class overrides) protected.
senorblanco@chromium.org [Thu, 17 Oct 2013 19:35:09 +0000 (19:35 +0000)]
Make SkImageFilter::asNewEffect() (and all derived-class overrides) protected.

R=bsalomon@google.com, bsalomon

Review URL: https://codereview.chromium.org/26937006

git-svn-id: http://skia.googlecode.com/svn/trunk@11847 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAdd texture filtering override to debugger
robertphillips@google.com [Thu, 17 Oct 2013 17:56:10 +0000 (17:56 +0000)]
Add texture filtering override to debugger

https://codereview.chromium.org/27716003/

git-svn-id: http://skia.googlecode.com/svn/trunk@11846 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert r11842 (call drawRect to try GrAARectRenderer if the path is a rect - https...
robertphillips@google.com [Thu, 17 Oct 2013 17:34:20 +0000 (17:34 +0000)]
Revert r11842 (call drawRect to try GrAARectRenderer if the path is a rect - https://codereview.chromium.org/23484007) due to changes to the following GM images:

inverse_paths
pathopsinverse

git-svn-id: http://skia.googlecode.com/svn/trunk@11845 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoAdd support for GL_CHROMIUM_renderbuffer_format_BGRA8888
commit-bot@chromium.org [Thu, 17 Oct 2013 16:29:41 +0000 (16:29 +0000)]
Add support for GL_CHROMIUM_renderbuffer_format_BGRA8888

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/27525006

git-svn-id: http://skia.googlecode.com/svn/trunk@11844 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoARM Skia NEON patches - 30 - Xfermode: NEON modeprocs
commit-bot@chromium.org [Thu, 17 Oct 2013 16:29:34 +0000 (16:29 +0000)]
ARM Skia NEON patches - 30 - Xfermode: NEON modeprocs

Xfermode: NEON implementation of SIMD procs

This patch contains a NEON implementation for a number of Xfermodes.
It provides a big speedup on Xfermode benchmarks (currently up to 3x
with gcc4.7 but up to 10x when gcc produces optimal code for it).

Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
BUG=

Committed: http://code.google.com/p/skia/source/detail?r=11777

Committed: http://code.google.com/p/skia/source/detail?r=11813

R=djsollen@google.com, mtklein@google.com, reed@google.com, robertphillips@google.com

Author: kevin.petit.arm@gmail.com

Review URL: https://codereview.chromium.org/26627004

git-svn-id: http://skia.googlecode.com/svn/trunk@11843 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoIf the path is a rect, call drawRect to raster the geometry in SkCanvas::drawPath...
commit-bot@chromium.org [Thu, 17 Oct 2013 16:29:21 +0000 (16:29 +0000)]
If the path is a rect, call drawRect to raster the geometry in SkCanvas::drawPath to get better performance.

R=bsalomon@google.com, robertphillips@google.com, reed@google.com

Author: yunchao.he@intel.com

Review URL: https://codereview.chromium.org/23484007

git-svn-id: http://skia.googlecode.com/svn/trunk@11842 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoDisable r11835 (Replace cubic round rect corners with quads) due to performance regre...
robertphillips@google.com [Thu, 17 Oct 2013 16:29:25 +0000 (16:29 +0000)]
Disable r11835 (Replace cubic round rect corners with quads) due to performance regression

git-svn-id: http://skia.googlecode.com/svn/trunk@11841 2bbb7eff-a529-9590-31e7-b0007b416f81

10 years agoRevert 11839 (rebaseline images after r11835) due to performance regression in 11835
robertphillips@google.com [Thu, 17 Oct 2013 16:28:11 +0000 (16:28 +0000)]
Revert 11839 (rebaseline images after r11835) due to performance regression in 11835

git-svn-id: http://skia.googlecode.com/svn/trunk@11840 2bbb7eff-a529-9590-31e7-b0007b416f81