platform/upstream/libSkiaSharp.git
11 years agoAdded support for Chrome's gpu command buffer extension BindUniformLocation.
commit-bot@chromium.org [Thu, 7 Nov 2013 22:06:08 +0000 (22:06 +0000)]
Added support for Chrome's gpu command buffer extension BindUniformLocation.

R=bsalomon@google.com, bsalomon

Author: skaslev@chromium.org

Review URL: https://codereview.chromium.org/62163004

git-svn-id: http://skia.googlecode.com/svn/trunk@12178 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agopdfviewer: fix typos (|| instead of |)
edisonn@google.com [Thu, 7 Nov 2013 20:33:17 +0000 (20:33 +0000)]
pdfviewer: fix typos (|| instead of |)

Review URL: https://codereview.chromium.org/65263002

git-svn-id: http://skia.googlecode.com/svn/trunk@12176 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix broken build on mac due to compiler warning
djsollen@google.com [Thu, 7 Nov 2013 19:40:35 +0000 (19:40 +0000)]
fix broken build on mac due to compiler warning

R=epoger@google.com

Review URL: https://codereview.chromium.org/64423004

git-svn-id: http://skia.googlecode.com/svn/trunk@12175 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoupdate skpdiff visualization (image magnification with alpha mask)
djsollen@google.com [Thu, 7 Nov 2013 19:24:06 +0000 (19:24 +0000)]
update skpdiff visualization (image magnification with alpha mask)

R=epoger@google.com

Review URL: https://codereview.chromium.org/29103005

git-svn-id: http://skia.googlecode.com/svn/trunk@12174 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoImplement the y-transpose optimization in image (RGBA) blurs. This gives ~38% perfor...
senorblanco@chromium.org [Thu, 7 Nov 2013 18:35:12 +0000 (18:35 +0000)]
Implement the y-transpose optimization in image (RGBA) blurs.  This gives ~38% performance improvement on my Linux box. Also unify the box blur functions with a template.

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

Review URL: https://codereview.chromium.org/59133006

git-svn-id: http://skia.googlecode.com/svn/trunk@12171 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix Mac warning in PdfViewer.
scroggo@google.com [Thu, 7 Nov 2013 18:22:18 +0000 (18:22 +0000)]
Fix Mac warning in PdfViewer.

Move a function only used by code guarded by #ifdef PDF_REPORT inside
guard.

R=edisonn@google.com

Review URL: https://codereview.chromium.org/60763003

git-svn-id: http://skia.googlecode.com/svn/trunk@12170 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoMark Xoom scaled_tilemodes_npot GPU images as ignore
bsalomon@google.com [Thu, 7 Nov 2013 16:36:46 +0000 (16:36 +0000)]
Mark Xoom scaled_tilemodes_npot GPU images as ignore

BUG=skia:1814

Review URL: https://codereview.chromium.org/61763011

git-svn-id: http://skia.googlecode.com/svn/trunk@12169 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix for thin filled and hairline roundrects.
commit-bot@chromium.org [Thu, 7 Nov 2013 16:34:53 +0000 (16:34 +0000)]
Fix for thin filled and hairline roundrects.

The roundrect renderer was dropping into the
ConvexPathRenderer when it didn't need to -- if a stroke is
just on the edge of being acceptable it will now use the
roundrect renderer.

BUG=Skia:1802
R=robertphillips@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/59903005

git-svn-id: http://skia.googlecode.com/svn/trunk@12168 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert "Revert "speed up A8 by creating a new entry-point in SkDraw that blits the...
reed@google.com [Thu, 7 Nov 2013 16:06:53 +0000 (16:06 +0000)]
Revert "Revert "speed up A8 by creating a new entry-point in SkDraw that blits the path's coverage directly into an A8 target, regardless of the previous pixel values.""

This reverts commit 3c77887b3eb2d32ab668ab4e5f2f9e79103956e8.

BUG=

Review URL: https://codereview.chromium.org/50673005

git-svn-id: http://skia.googlecode.com/svn/trunk@12167 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoARM Skia NEON patches - 33 - Convolution filter
commit-bot@chromium.org [Thu, 7 Nov 2013 15:57:32 +0000 (15:57 +0000)]
ARM Skia NEON patches - 33 - Convolution filter
NEON version of the convolutionProcs

The bitmap_scale benchmark is now twice as fast on ARM.

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

Committed: http://code.google.com/p/skia/source/detail?r=12154

R=djsollen@google.com, mtklein@google.com, humper@google.com, epoger@google.com

Author: kevin.petit.arm@gmail.com

Review URL: https://codereview.chromium.org/27533004

git-svn-id: http://skia.googlecode.com/svn/trunk@12166 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoWhitespace change to trigger builds after r12163
borenet@google.com [Thu, 7 Nov 2013 13:43:03 +0000 (13:43 +0000)]
Whitespace change to trigger builds after r12163

Review URL: https://codereview.chromium.org/64283003

git-svn-id: http://skia.googlecode.com/svn/trunk@12165 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoWhitespace change to test master responsiveness.
rmistry@google.com [Thu, 7 Nov 2013 12:18:35 +0000 (12:18 +0000)]
Whitespace change to test master responsiveness.

git-svn-id: http://skia.googlecode.com/svn/trunk@12162 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Thu, 7 Nov 2013 07:02:19 +0000 (07:02 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@12161 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agowhitespace change after killing trybots
epoger@google.com [Thu, 7 Nov 2013 04:22:46 +0000 (04:22 +0000)]
whitespace change after killing trybots
BUG=skia:1810

git-svn-id: http://skia.googlecode.com/svn/trunk@12160 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years ago2nd whitespace change after git-repo filtering
epoger@google.com [Wed, 6 Nov 2013 20:58:25 +0000 (20:58 +0000)]
2nd whitespace change after git-repo filtering
BUG=skia:1808

git-svn-id: http://skia.googlecode.com/svn/trunk@12159 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoWhitespace commit after git-repo filtering
epoger@google.com [Wed, 6 Nov 2013 20:44:12 +0000 (20:44 +0000)]
Whitespace commit after git-repo filtering
BUG=skia:1808

git-svn-id: http://skia.googlecode.com/svn/trunk@12158 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoUpdate gm flags so that it can be run without a pdf rasterizer.
vandebo@chromium.org [Wed, 6 Nov 2013 16:32:15 +0000 (16:32 +0000)]
Update gm flags so that it can be run without a pdf rasterizer.

BUG=1806
R=edisonn@google.com

Review URL: https://codereview.chromium.org/61643004

git-svn-id: http://skia.googlecode.com/svn/trunk@12157 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert r12154
epoger@google.com [Wed, 6 Nov 2013 15:49:04 +0000 (15:49 +0000)]
Revert r12154
BUG=skia:1807

git-svn-id: http://skia.googlecode.com/svn/trunk@12156 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoNew SkRTConf macro SK_CONF_TRY_SET: no complaint on missing configuration
halcanary@google.com [Wed, 6 Nov 2013 15:07:44 +0000 (15:07 +0000)]
New SkRTConf macro SK_CONF_TRY_SET: no complaint on missing configuration

SK_CONF_TRY_SET() is like SK_CONF_SET(), but doesn't complain if
confname can't be found.  This is useful if the SK_CONF_DECLARE is
inside a source file whose linkage is dependent on the system.

Internally to the SkRTConf system, SkRTConfRegistry::set() was given
an additional parameter controling wanrings.

A new RuntimeConfig unit test was introduced.  It should run silently.
In the future, it should be expanded to cover all of the SkRTConf
functionality.

(For example, the images.jpeg.suppressDecoderWarnings variable is
defined and used only in SkImageDecoder_libjpeg.cpp, but on MacOS, we
use Core Graphics via SkImageDecoder_CG.cpp - SkImageDecoder_libjpeg
is never linked in.  The same is true of the Windows Imaging Component
on Windows.)

BUG=
R=reed@google.com

Review URL: https://codereview.chromium.org/54503007

git-svn-id: http://skia.googlecode.com/svn/trunk@12155 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoARM Skia NEON patches - 33 - Convolution filter
commit-bot@chromium.org [Wed, 6 Nov 2013 15:02:46 +0000 (15:02 +0000)]
ARM Skia NEON patches - 33 - Convolution filter
NEON version of the convolutionProcs

The bitmap_scale benchmark is now twice as fast on ARM.

Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
BUG=
R=djsollen@google.com, mtklein@google.com, humper@google.com

Author: kevin.petit.arm@gmail.com

Review URL: https://codereview.chromium.org/27533004

git-svn-id: http://skia.googlecode.com/svn/trunk@12154 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix installing the release build sample app apk
commit-bot@chromium.org [Wed, 6 Nov 2013 15:01:37 +0000 (15:01 +0000)]
Fix installing the release build sample app apk

The android_setup.sh parses command line arguments and stores the ones
it does not use to APP_ARGS variable.

R=djsollen@google.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/61553002

git-svn-id: http://skia.googlecode.com/svn/trunk@12153 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix a bug which caused crash in SkRTConfig whith string values in environment variables
halcanary@google.com [Wed, 6 Nov 2013 14:18:46 +0000 (14:18 +0000)]
Fix a bug which caused crash in SkRTConfig whith string values in environment variables

A minor change to SkRTConfRegistry::parse() fixes this

BUG=skia:1717
R=scroggo@google.com

Review URL: https://codereview.chromium.org/26373008

git-svn-id: http://skia.googlecode.com/svn/trunk@12152 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoPerlin issue fixed
commit-bot@chromium.org [Wed, 6 Nov 2013 10:08:38 +0000 (10:08 +0000)]
Perlin issue fixed

BUG=skia:1799
R=sugoi@chromium.org, sugoi@google.com, senorblanco@chromium.org

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/59173005

git-svn-id: http://skia.googlecode.com/svn/trunk@12150 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoBreak up SkLazyPixelRef functionally into class hierarchy.
commit-bot@chromium.org [Wed, 6 Nov 2013 10:08:30 +0000 (10:08 +0000)]
Break up SkLazyPixelRef functionally into class hierarchy.

The reason for this CL is to allow greater decoder flexibility.
Chrome currently uses its own decoding functions.  These allow for
greater flexibility in dealing with images with multiple frames or
partial data.  The DecodeProc function was not flexible enough to
handle these.  Instead of asking the decoder to squeeze everything
into the DecodeProc, we now ask the downstream library to inherit from
SkCachingPixelRef.  If WebKit's LazyDecodingPixelRef is re-tooled to
inherit from SkCachingPixelRef, then it can make use of Skia's caching
ability while still allowing it to deal with multiple frames, scaling,
subsetting, and partial data.

-   The abstract SkCachingPixelRef class handles caching the decoded
    data in a SkScaledImageCache.  This class relies on the virtual
    functions onDecodeInfo() and onDecode() to do the actual decoding
    of data.

-   The SkLazyCachingPixelRef class is derived from SkCachingPixelRef.
    It provides an implementation of onDecodeInfo() and onDecode() in
    terms of calls to a SkBitmapFactory::DecodeProc function.  It also
    provides an Install() static method which installs a new
    SkLazyCachingPixelRef into a SkBitmap.

SkLazyCachingPixelRef exists for two reasons: to test
SkCachingPixelRef within Skia and as an example for downstream
developers to make their own classes that inherit from
SkCachingPixelRef.

-   The CachedDecodingPixelRefTest was updated to test the
    SkLazyCachingPixelRef class and indirectly the SkCachingPixelRef
    class.

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

Author: halcanary@google.com

Review URL: https://codereview.chromium.org/54203006

git-svn-id: http://skia.googlecode.com/svn/trunk@12149 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd bevel-stroke support in GrAARectRenderer
commit-bot@chromium.org [Wed, 6 Nov 2013 10:08:21 +0000 (10:08 +0000)]
Add bevel-stroke support in GrAARectRenderer

Committed: http://code.google.com/p/skia/source/detail?r=12082

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

Author: yunchao.he@intel.com

Review URL: https://codereview.chromium.org/23712005

git-svn-id: http://skia.googlecode.com/svn/trunk@12148 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Wed, 6 Nov 2013 07:02:11 +0000 (07:02 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@12147 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoblurroundrect gm: disable one set of tests in attempt to fix bug
epoger@google.com [Tue, 5 Nov 2013 22:04:53 +0000 (22:04 +0000)]
blurroundrect gm: disable one set of tests in attempt to fix bug
BUG=skia:1801
R=scroggo@google.com

Review URL: https://codereview.chromium.org/61073002

git-svn-id: http://skia.googlecode.com/svn/trunk@12146 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix overzealous IntToScalar.
scroggo@google.com [Tue, 5 Nov 2013 21:22:04 +0000 (21:22 +0000)]
Fix overzealous IntToScalar.

git-svn-id: http://skia.googlecode.com/svn/trunk@12145 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFixes for blurroundrect gm/bench.
scroggo@google.com [Tue, 5 Nov 2013 21:10:58 +0000 (21:10 +0000)]
Fixes for blurroundrect gm/bench.

Simplify naming/various cases.
Have an atlas of gms, rather than several different images.
In the bench, pull non rendering out of the loop.
Use meaningful enums instead of integers where appropriate.
Add comments.

Addresses comments in https://codereview.chromium.org/52793005/

R=robertphillips@google.com

Review URL: https://codereview.chromium.org/59983004

git-svn-id: http://skia.googlecode.com/svn/trunk@12144 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agorebaseline_server was not allowing tests to be marked ignore-failure
epoger@google.com [Tue, 5 Nov 2013 21:00:24 +0000 (21:00 +0000)]
rebaseline_server was not allowing tests to be marked ignore-failure
(SkipBuildbotRuns)
TBR=jcgregorio

Review URL: https://codereview.chromium.org/59063005

git-svn-id: http://skia.googlecode.com/svn/trunk@12143 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoMark blurs-gpu test as known-to-fail on Nexus4
epoger@google.com [Tue, 5 Nov 2013 20:55:32 +0000 (20:55 +0000)]
Mark blurs-gpu test as known-to-fail on Nexus4
BUG=skia:1791
(SkipBuildbotRuns)

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/59943004

git-svn-id: http://skia.googlecode.com/svn/trunk@12142 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoInt to scalar (again)
scroggo@google.com [Tue, 5 Nov 2013 17:29:23 +0000 (17:29 +0000)]
Int to scalar (again)

git-svn-id: http://skia.googlecode.com/svn/trunk@12139 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAnother int to scalar fix.
scroggo@google.com [Tue, 5 Nov 2013 17:27:01 +0000 (17:27 +0000)]
Another int to scalar fix.

git-svn-id: http://skia.googlecode.com/svn/trunk@12138 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoMore IntToScalars
scroggo@google.com [Tue, 5 Nov 2013 17:02:07 +0000 (17:02 +0000)]
More IntToScalars

Fix build.

git-svn-id: http://skia.googlecode.com/svn/trunk@12137 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAnd fix the similar bench.
scroggo@google.com [Tue, 5 Nov 2013 16:40:15 +0000 (16:40 +0000)]
And fix the similar bench.

git-svn-id: http://skia.googlecode.com/svn/trunk@12136 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoBuild fix number 2.
scroggo@google.com [Tue, 5 Nov 2013 16:28:38 +0000 (16:28 +0000)]
Build fix number 2.

Use SkColors instead of numbers.

git-svn-id: http://skia.googlecode.com/svn/trunk@12135 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix build.
scroggo@google.com [Tue, 5 Nov 2013 16:19:32 +0000 (16:19 +0000)]
Fix build.

Fix some warnings.

git-svn-id: http://skia.googlecode.com/svn/trunk@12134 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd gms and benchmarks for drawing blurry round rects.
scroggo@google.com [Tue, 5 Nov 2013 15:57:21 +0000 (15:57 +0000)]
Add gms and benchmarks for drawing blurry round rects.

Further changes (https://codereview.chromium.org/48623006) will change
the speed at which the bench draws and the drawing of the gm (the gm
change is small).

One of these round rects causes slow drawing in a webpage that we have
observed.

BUG=https://b.corp.google.com/issue?id=11174385

Review URL: https://codereview.chromium.org/52793005

git-svn-id: http://skia.googlecode.com/svn/trunk@12133 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd SkRRect::transform.
scroggo@google.com [Tue, 5 Nov 2013 15:54:42 +0000 (15:54 +0000)]
Add SkRRect::transform.

Much like SkPath::transform, it transforms an SkRRect based on an
SkMatrix. Unlike SkPath::transform, it will fail for matrices that
contain perspective or skewing.

Required by a future change (https://codereview.chromium.org/48623006)
to speed up drawing large blurry rounded rectangles by using ninepatches.

TODO: This could easily support 90 degree rotations, if desired.

BUG=https://b.corp.google.com/issue?id=11174385
R=reed@google.com, robertphillips@google.com

Review URL: https://codereview.chromium.org/52703003

git-svn-id: http://skia.googlecode.com/svn/trunk@12132 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoCheck that Mode is valid for ModeColorFilterEffect
commit-bot@chromium.org [Tue, 5 Nov 2013 15:47:48 +0000 (15:47 +0000)]
Check that Mode is valid for ModeColorFilterEffect

R=robertphillips@google.com

Author: bsalomon@google.com

Review URL: https://codereview.chromium.org/53983005

git-svn-id: http://skia.googlecode.com/svn/trunk@12131 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoChecking structure sizes before reading them from memory to avoid overflowing the...
commit-bot@chromium.org [Tue, 5 Nov 2013 15:46:56 +0000 (15:46 +0000)]
Checking structure sizes before reading them from memory to avoid overflowing the buffer's stream.

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

Committed: https://code.google.com/p/skia/source/detail?r=12114

Committed: https://code.google.com/p/skia/source/detail?r=12119

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/41253002

git-svn-id: http://skia.googlecode.com/svn/trunk@12130 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd SK_PREFETCH and use in SkBlurImageFilter.
commit-bot@chromium.org [Tue, 5 Nov 2013 15:03:26 +0000 (15:03 +0000)]
Add SK_PREFETCH and use in SkBlurImageFilter.

Relative speed is 1.2-1.6x on desktop, 1.0-1.2x on Nexus 4.

(Division remains the bottleneck, now more so.)

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

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/57823003

git-svn-id: http://skia.googlecode.com/svn/trunk@12129 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoMake it possible to pass GYP_DEFINES to Android builds
commit-bot@chromium.org [Tue, 5 Nov 2013 15:03:19 +0000 (15:03 +0000)]
Make it possible to pass GYP_DEFINES to Android builds

Pass possibly existing GYP_DEFINES from shell to gyp when building
for Android. Makes it possible to use custom build flags such as
skia_nv_path_rendering=1.

R=djsollen@google.com

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/51003007

git-svn-id: http://skia.googlecode.com/svn/trunk@12128 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAvoid re-rendering stencil clip for every draw with reducable clip stack
commit-bot@chromium.org [Tue, 5 Nov 2013 15:03:08 +0000 (15:03 +0000)]
Avoid re-rendering stencil clip for every draw with reducable clip stack

Fixes the cases where clip stack reduction would cause clip to be
re-rendered to stencil for each draw call. This causes unneeded
slowdown.

Stencil cache would not be used because the clip stack generation id communicated
by the clip stack element list would be invalid. This happended due to

 a) clip stack reduction creating new elements in the element list.

 b) purging logic removing the generation id, but reduction logic
    selecting already purged element, and thus the generation id, as
    the representative state of the clip.

Cases of a) where reduction would flatten the stack to a single new
element were fixed by assigning the generation id of the top-most
element of the clip stack as the generation id of the new
element. This is not strictly minimal, but enables more caching than
using invalid id.

Cases of a) where reduction would substitute a stack element with a
new element the generation id of the substituted element is used.

The b) part was fixed by removing the purging logic. It was not
exactly correct, as the previously purged states were actually
used. The purging was not used for anything.

Changes SkClipStack API to highlight that invalid generation id is
never returned by SkClipStack. Empty stacks are wide open. Changes the
clients to reflect this.

Fixes a crash when not passing anti-alias out parameter to
GrReducedClip::ReduceClipStack. The crash is not exercised in the
current code.

Committed: http://code.google.com/p/skia/source/detail?r=12084

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

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/48593003

git-svn-id: http://skia.googlecode.com/svn/trunk@12127 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert "Checking structure sizes before reading them from memory to avoid overflowing...
reed@google.com [Mon, 4 Nov 2013 21:35:55 +0000 (21:35 +0000)]
Revert "Checking structure sizes before reading them from memory to avoid overflowing the buffer's stream."

This reverts commit 6bc22e8ef1ea70a1b58409aa21254358c50f149a.

git-svn-id: http://skia.googlecode.com/svn/trunk@12124 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoWe always read out of bounds here. Fix that.
mtklein@google.com [Mon, 4 Nov 2013 21:22:45 +0000 (21:22 +0000)]
We always read out of bounds here.  Fix that.

Turns out the comment wasn't lying!

BUG=
R=bungeman@google.com

Review URL: https://codereview.chromium.org/58473002

git-svn-id: http://skia.googlecode.com/svn/trunk@12123 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert "speed up A8 by creating a new entry-point in SkDraw that blits the path's...
reed@google.com [Mon, 4 Nov 2013 20:57:36 +0000 (20:57 +0000)]
Revert "speed up A8 by creating a new entry-point in SkDraw that blits the path's coverage directly into an A8 target, regardless of the previous pixel values."

This reverts commit d615e839b71f75df895de6850b774a9e1c28ad2a.

Revert "must initialize SkDraw.fClip"

This reverts commit 108e46d29b5f57927fc8b8c403bb52019d8cb16d.

BUG=

Review URL: https://codereview.chromium.org/57883006

git-svn-id: http://skia.googlecode.com/svn/trunk@12122 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agorebaseline scaled_tilemode_npot gpu images on RHB and rebaseline SGX tilemode gpu
bsalomon@google.com [Mon, 4 Nov 2013 20:38:18 +0000 (20:38 +0000)]
rebaseline scaled_tilemode_npot gpu images on RHB and rebaseline SGX tilemode gpu

BUG=skia:1789

Review URL: https://codereview.chromium.org/47263003

git-svn-id: http://skia.googlecode.com/svn/trunk@12121 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agomust initialize SkDraw.fClip
reed@google.com [Mon, 4 Nov 2013 20:31:40 +0000 (20:31 +0000)]
must initialize SkDraw.fClip

git-svn-id: http://skia.googlecode.com/svn/trunk@12120 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoChecking structure sizes before reading them from memory to avoid overflowing the...
sugoi@google.com [Mon, 4 Nov 2013 20:28:23 +0000 (20:28 +0000)]
Checking structure sizes before reading them from memory to avoid overflowing the buffer's stream.

BUG=
R=reed@google.com

Committed: https://code.google.com/p/skia/source/detail?r=12114

Review URL: https://codereview.chromium.org/41253002

git-svn-id: http://skia.googlecode.com/svn/trunk@12119 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agospeed up A8 by creating a new entry-point in SkDraw that blits the path's coverage...
reed@google.com [Mon, 4 Nov 2013 20:10:33 +0000 (20:10 +0000)]
speed up A8 by creating a new entry-point in SkDraw that blits the path's coverage directly into an A8 target, regardless of the previous pixel values.

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

Review URL: https://codereview.chromium.org/56453002

git-svn-id: http://skia.googlecode.com/svn/trunk@12118 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agorebaseline blurs_gpu
epoger@google.com [Mon, 4 Nov 2013 18:13:00 +0000 (18:13 +0000)]
rebaseline blurs_gpu
BUG=skia:1791
TBR=robertphillips

Review URL: https://codereview.chromium.org/57433005

git-svn-id: http://skia.googlecode.com/svn/trunk@12116 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert r12114 due to https://code.google.com/p/skia/issues/detail?id=1794 ('Assertion...
epoger@google.com [Mon, 4 Nov 2013 18:06:12 +0000 (18:06 +0000)]
Revert r12114 due to https://code.google.com/p/skia/issues/detail?id=1794 ('Assertion failures on various buildbots as of r12114')

git-svn-id: http://skia.googlecode.com/svn/trunk@12115 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoChecking structure sizes before reading them from memory to avoid overflowing the...
sugoi@google.com [Mon, 4 Nov 2013 16:18:15 +0000 (16:18 +0000)]
Checking structure sizes before reading them from memory to avoid overflowing the buffer's stream.

BUG=
R=reed@google.com

Review URL: https://codereview.chromium.org/41253002

git-svn-id: http://skia.googlecode.com/svn/trunk@12114 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoChange SkBlurImageFilter to use fixed-point division. Yields ~1.8X speedup on
senorblanco@chromium.org [Mon, 4 Nov 2013 16:07:33 +0000 (16:07 +0000)]
Change SkBlurImageFilter to use fixed-point division. Yields ~1.8X speedup on
MacBookPro.

Note: this will require defining SK_DISABLE_BLUR_DIVISION_OPTIMIZATION in
Chrome, a la https://codereview.chromium.org/49973005.

R=mtklein@google.com

Review URL: https://codereview.chromium.org/57513002

git-svn-id: http://skia.googlecode.com/svn/trunk@12113 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix baselines for arcofzorro hairlines shadertext shadertext2
caryclark@google.com [Mon, 4 Nov 2013 12:57:42 +0000 (12:57 +0000)]
fix baselines for arcofzorro hairlines shadertext shadertext2

git-svn-id: http://skia.googlecode.com/svn/trunk@12109 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix for Chromium-side compilation complaint:
robertphillips@google.com [Sun, 3 Nov 2013 13:40:34 +0000 (13:40 +0000)]
Fix for Chromium-side compilation complaint:

src/pathops/SkIntersectionHelper.h:10:13: error: #if with no expression

git-svn-id: http://skia.googlecode.com/svn/trunk@12108 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Sat, 2 Nov 2013 07:02:02 +0000 (07:02 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix more expectations for shadertext shadertext2 hairlines
caryclark@google.com [Fri, 1 Nov 2013 21:15:01 +0000 (21:15 +0000)]
fix more expectations for shadertext shadertext2 hairlines

git-svn-id: http://skia.googlecode.com/svn/trunk@12099 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoUpdate missed expectations for xfermodes
vandebo@chromium.org [Fri, 1 Nov 2013 21:10:17 +0000 (21:10 +0000)]
Update missed expectations for xfermodes

ref: r12034, r12053

Review URL: https://codereview.chromium.org/56473003

git-svn-id: http://skia.googlecode.com/svn/trunk@12098 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoupdate ignored tests with rebaselines
caryclark@google.com [Fri, 1 Nov 2013 20:45:49 +0000 (20:45 +0000)]
update ignored tests with rebaselines

git-svn-id: http://skia.googlecode.com/svn/trunk@12097 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix shadertext2 expectations
caryclark@google.com [Fri, 1 Nov 2013 20:34:24 +0000 (20:34 +0000)]
fix shadertext2 expectations

git-svn-id: http://skia.googlecode.com/svn/trunk@12096 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix shadedtext expectations
caryclark@google.com [Fri, 1 Nov 2013 20:25:47 +0000 (20:25 +0000)]
fix shadedtext expectations

git-svn-id: http://skia.googlecode.com/svn/trunk@12095 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix hairline expectations
caryclark@google.com [Fri, 1 Nov 2013 20:16:50 +0000 (20:16 +0000)]
fix hairline expectations

git-svn-id: http://skia.googlecode.com/svn/trunk@12094 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix expectations for arcofzorro
caryclark@google.com [Fri, 1 Nov 2013 20:02:16 +0000 (20:02 +0000)]
fix expectations for arcofzorro

git-svn-id: http://skia.googlecode.com/svn/trunk@12093 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoignore gm roundrect errors including arcofzorro
caryclark@google.com [Fri, 1 Nov 2013 19:28:33 +0000 (19:28 +0000)]
ignore gm roundrect errors including arcofzorro

git-svn-id: http://skia.googlecode.com/svn/trunk@12092 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoReverting r12082 (Add bevel-stroke support in GrAARectRenderer) due to GM failures
robertphillips@google.com [Fri, 1 Nov 2013 19:22:42 +0000 (19:22 +0000)]
Reverting r12082 (Add bevel-stroke support in GrAARectRenderer) due to GM failures

git-svn-id: http://skia.googlecode.com/svn/trunk@12091 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoignore gm roundrect errors
caryclark@google.com [Fri, 1 Nov 2013 19:18:59 +0000 (19:18 +0000)]
ignore gm roundrect errors

git-svn-id: http://skia.googlecode.com/svn/trunk@12090 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agopathops work in progress
caryclark@google.com [Fri, 1 Nov 2013 17:36:03 +0000 (17:36 +0000)]
pathops work in progress

BUG=

Review URL: https://codereview.chromium.org/52653002

git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agorebaseline_server: make default landing page list supported query URLs
epoger@google.com [Fri, 1 Nov 2013 16:46:41 +0000 (16:46 +0000)]
rebaseline_server: make default landing page list supported query URLs
(SkipBuildbotRuns)

Originally, the default landing page redirected to view.html?resultsToLoad=all
(which displays all results, and can take a while).
Instead, start with a fast-loading page that shows the user all the result
sets that are available.

R=scroggo@google.com

Review URL: https://codereview.chromium.org/46413004

git-svn-id: http://skia.googlecode.com/svn/trunk@12088 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert "Avoid re-rendering stencil clip for every draw with reducable clip stack"
reed@google.com [Fri, 1 Nov 2013 16:25:59 +0000 (16:25 +0000)]
Revert "Avoid re-rendering stencil clip for every draw with reducable clip stack"

This reverts commit 92a7d4bf6a371f1f864154be902e8d86938e560b.

Revert "fix mac 10.6 build"

This reverts commit 114cd1a9f2734aaed6914718814364811b78bd7f.

BUG=

Review URL: https://codereview.chromium.org/54543008

git-svn-id: http://skia.googlecode.com/svn/trunk@12087 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agofix mac 10.6 build
reed@google.com [Fri, 1 Nov 2013 16:15:01 +0000 (16:15 +0000)]
fix mac 10.6 build
- double -> float implicit conversion (for literals like 1.3)
- locally defined (and named) struct

R=scroggo@google.com

Review URL: https://codereview.chromium.org/56103002

git-svn-id: http://skia.googlecode.com/svn/trunk@12086 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoConstruct round rects with perpendicular tangents.
commit-bot@chromium.org [Fri, 1 Nov 2013 15:24:55 +0000 (15:24 +0000)]
Construct round rects with perpendicular tangents.

The round rects are constructed as before out of quadratics,
but without fudging the control points. Instead, the mid on-
curve point is nudged slightly outward to prevent the
convexity test from failing.

The convexity test now includes an error term for sign
inequality after computing the cross product of the control
lines. When the control points are represented as vectors,
the number of bits of precision may be greatly reduced.

Account for this by passing the number of bits available
from the original control point values into the equality
check.

Making round rect construction lines perpendicular improves
the chances of success when path ops encounters clips.

No new tests are needed -- this change is exercised by the
convex Path unit tests and the gm tests arcofzorro and
hairlines.

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

Author: caryclark@google.com

Review URL: https://codereview.chromium.org/48783002

git-svn-id: http://skia.googlecode.com/svn/trunk@12085 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAvoid re-rendering stencil clip for every draw with reducable clip stack
commit-bot@chromium.org [Fri, 1 Nov 2013 15:24:20 +0000 (15:24 +0000)]
Avoid re-rendering stencil clip for every draw with reducable clip stack

Fixes the cases where clip stack reduction would cause clip to be
re-rendered to stencil for each draw call. This causes unneeded
slowdown.

Stencil cache would not be used because the clip stack generation id communicated
by the clip stack element list would be invalid. This happended due to

 a) clip stack reduction creating new elements in the element list.

 b) purging logic removing the generation id, but reduction logic
    selecting already purged element, and thus the generation id, as
    the representative state of the clip.

Cases of a) where reduction would flatten the stack to a single new
element were fixed by assigning the generation id of the top-most
element of the clip stack as the generation id of the new
element. This is not strictly minimal, but enables more caching than
using invalid id.

Cases of a) where reduction would substitute a stack element with a
new element the generation id of the substituted element is used.

The b) part was fixed by removing the purging logic. It was not
exactly correct, as the previously purged states were actually
used. The purging was not used for anything.

Changes SkClipStack API to highlight that invalid generation id is
never returned by SkClipStack. Empty stacks are wide open. Changes the
clients to reflect this.

Fixes a crash when not passing anti-alias out parameter to
GrReducedClip::ReduceClipStack. The crash is not exercised in the
current code.

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

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/48593003

git-svn-id: http://skia.googlecode.com/svn/trunk@12084 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoMake GrContext cache the gpu paths
commit-bot@chromium.org [Fri, 1 Nov 2013 15:23:44 +0000 (15:23 +0000)]
Make GrContext cache the gpu paths

Creating paths for nv_path_rendering is costly. Try to reduce this
cost by caching paths based on the SkPath "hash" (i.e. SkPathRef
generation id) and stroke properties.

Adds the paths to GrContext::fTextureCache instance. Later this should
be renamed and the GrContext API should reflect the nature of the cache
better.

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

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/26557003

git-svn-id: http://skia.googlecode.com/svn/trunk@12083 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd bevel-stroke support in GrAARectRenderer
commit-bot@chromium.org [Fri, 1 Nov 2013 15:23:17 +0000 (15:23 +0000)]
Add bevel-stroke support in GrAARectRenderer

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

Author: yunchao.he@intel.com

Review URL: https://codereview.chromium.org/23712005

git-svn-id: http://skia.googlecode.com/svn/trunk@12082 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRebaseline improved N4 images
robertphillips@google.com [Fri, 1 Nov 2013 15:11:18 +0000 (15:11 +0000)]
Rebaseline improved N4 images

https://codereview.chromium.org/49983005/

git-svn-id: http://skia.googlecode.com/svn/trunk@12081 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoOnly disable scratch texture reuse on the N10 and N4
robertphillips@google.com [Fri, 1 Nov 2013 15:03:34 +0000 (15:03 +0000)]
Only disable scratch texture reuse on the N10 and N4

https://codereview.chromium.org/48903017/

git-svn-id: http://skia.googlecode.com/svn/trunk@12080 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agorebaseline_server: right-align "select/clear/toggle all tests" buttons with results...
epoger@google.com [Fri, 1 Nov 2013 14:29:03 +0000 (14:29 +0000)]
rebaseline_server: right-align "select/clear/toggle all tests" buttons with results table
(SkipBuildbotRuns)

R=edisonn@google.com

Review URL: https://codereview.chromium.org/55073003

git-svn-id: http://skia.googlecode.com/svn/trunk@12079 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agomove SkImage::ColorType into SkColorType
reed@google.com [Fri, 1 Nov 2013 13:46:54 +0000 (13:46 +0000)]
move SkImage::ColorType into SkColorType

objective -- move clients over to SkImage

tasks
- use SkImageInfo instead of SkBitmap::Config
- add support for colortables to SkImage
- add drawImage to SkCanvas
- return SkImage from readPixels

This CL works towards the first task

R=robertphillips@google.com

Review URL: https://codereview.chromium.org/54363008

git-svn-id: http://skia.googlecode.com/svn/trunk@12077 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoUpdate results for a Nexus4 tests that don't render well with perspective
edisonn@google.com [Fri, 1 Nov 2013 12:33:43 +0000 (12:33 +0000)]
Update results for a Nexus4 tests that don't render well with perspective

Review URL: https://codereview.chromium.org/55633003

git-svn-id: http://skia.googlecode.com/svn/trunk@12076 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdding a way to disable SkOffsetImageFilter's optimized path in blink/chrome
commit-bot@chromium.org [Fri, 1 Nov 2013 11:55:06 +0000 (11:55 +0000)]
Adding a way to disable SkOffsetImageFilter's optimized path in blink/chrome

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

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/54543005

git-svn-id: http://skia.googlecode.com/svn/trunk@12075 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix two inconsistencies in handling fixed-function texture coordinate set generation
commit-bot@chromium.org [Fri, 1 Nov 2013 11:54:54 +0000 (11:54 +0000)]
Fix two inconsistencies in handling fixed-function texture coordinate set generation

Prevent off-by-one bug leaving the texture coordinate set 0 always enabled if
it ever was used. This makes the code consistent with its apparent purpose.

When enabling / disabling texture coordinate sets, call glPathTexGen
only if path rendering is applicable. This makes the functions
consistent with GrGpuGL::resetContext.

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

Author: kkinnunen@nvidia.com

Review URL: https://codereview.chromium.org/54403003

git-svn-id: http://skia.googlecode.com/svn/trunk@12074 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoadd some GM tests for rects
commit-bot@chromium.org [Fri, 1 Nov 2013 11:54:29 +0000 (11:54 +0000)]
add some GM tests for rects

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

Author: yunchao.he@intel.com

Review URL: https://codereview.chromium.org/54213002

git-svn-id: http://skia.googlecode.com/svn/trunk@12073 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoSanitizing source files in Housekeeper-Nightly
skia.committer@gmail.com [Fri, 1 Nov 2013 07:02:15 +0000 (07:02 +0000)]
Sanitizing source files in Housekeeper-Nightly

git-svn-id: http://skia.googlecode.com/svn/trunk@12071 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoFix for r12069 (actually turn on don't-reuse-scratch-textures feature)
robertphillips@google.com [Fri, 1 Nov 2013 01:55:40 +0000 (01:55 +0000)]
Fix for r12069 (actually turn on don't-reuse-scratch-textures feature)

git-svn-id: http://skia.googlecode.com/svn/trunk@12070 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoTurn on don't-reuse-scratch-textures for all Android devices
robertphillips@google.com [Fri, 1 Nov 2013 00:40:47 +0000 (00:40 +0000)]
Turn on don't-reuse-scratch-textures for all Android devices

https://codereview.chromium.org/54493005/

git-svn-id: http://skia.googlecode.com/svn/trunk@12069 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAnother rebaseline
robertphillips@google.com [Thu, 31 Oct 2013 23:20:03 +0000 (23:20 +0000)]
Another rebaseline

https://codereview.chromium.org/46043021/

git-svn-id: http://skia.googlecode.com/svn/trunk@12068 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoDisabling SK_ATTR_DEPRECATED the hard way
robertphillips@google.com [Thu, 31 Oct 2013 22:42:33 +0000 (22:42 +0000)]
Disabling SK_ATTR_DEPRECATED the hard way

git-svn-id: http://skia.googlecode.com/svn/trunk@12067 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoMore image rebaselines
robertphillips@google.com [Thu, 31 Oct 2013 22:36:37 +0000 (22:36 +0000)]
More image rebaselines

git-svn-id: http://skia.googlecode.com/svn/trunk@12066 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoYet more rebaselines
robertphillips@google.com [Thu, 31 Oct 2013 22:21:25 +0000 (22:21 +0000)]
Yet more rebaselines

https://codereview.chromium.org/55303004/

git-svn-id: http://skia.googlecode.com/svn/trunk@12065 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd can-ignore-rect hint to clear call
robertphillips@google.com [Thu, 31 Oct 2013 21:44:25 +0000 (21:44 +0000)]
Add can-ignore-rect hint to clear call

https://codereview.chromium.org/53823003/

git-svn-id: http://skia.googlecode.com/svn/trunk@12064 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdd suppressions for Issue 1784
robertphillips@google.com [Thu, 31 Oct 2013 21:27:44 +0000 (21:27 +0000)]
Add suppressions for Issue 1784

https://codereview.chromium.org/49533006/

git-svn-id: http://skia.googlecode.com/svn/trunk@12063 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoPartial fix for Skia issue 1769: GenerateGMs crashing (out of memory)
commit-bot@chromium.org [Thu, 31 Oct 2013 20:22:47 +0000 (20:22 +0000)]
Partial fix for Skia issue 1769: GenerateGMs crashing (out of memory)

GenerateGMs was crashing because we were adding genID listeners for
textures even when creation failed -- in this case the key generated
is not valid, so when it comes time to purge it, it triggers an
assert and then an apparent crash. The solution is to not add
a genID listener in this case.

This doesn't solve the problem of running out of memory, it just prevents it
from causing a crash in Debug builds.

BUG=
R=bsalomon@google.com, robertphillips@google.com, mtklein@google.com

Author: jvanverth@google.com

Review URL: https://codereview.chromium.org/55023005

git-svn-id: http://skia.googlecode.com/svn/trunk@12062 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRebaseline images that improved with 12037
robertphillips@google.com [Thu, 31 Oct 2013 19:43:24 +0000 (19:43 +0000)]
Rebaseline images that improved with 12037

https://codereview.chromium.org/55193002/

git-svn-id: http://skia.googlecode.com/svn/trunk@12061 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRebaseline lingering RoundRect RHB images
robertphillips@google.com [Thu, 31 Oct 2013 19:36:42 +0000 (19:36 +0000)]
Rebaseline lingering RoundRect RHB images

https://codereview.chromium.org/52713011/

git-svn-id: http://skia.googlecode.com/svn/trunk@12060 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoAdding size parameter to read array functions
commit-bot@chromium.org [Thu, 31 Oct 2013 18:37:50 +0000 (18:37 +0000)]
Adding size parameter to read array functions

In some cases, the allocated array into which the data will be read is using getArrayCount() to allocate itself, which should be safe, but some cases use fixed length arrays or compute the array size before reading, which could overflow if the stream is compromised.

To prevent that from happening, I added a check that will verify that the number of bytes to read will not exceed the capacity of the input buffer argument passed to all the read...Array() functions.

I chose to use the byte array for this initial version, so that "size" represents the same value across all read...Array() functions, but I could also use the element count, if it is preferred.

Note : readPointArray and writePointArray are unused, so I could also remove them

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

Author: sugoi@chromium.org

Review URL: https://codereview.chromium.org/37803002

git-svn-id: http://skia.googlecode.com/svn/trunk@12058 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert "Revert "add SK_ATTR_DEPRECATED -- will need to disable for chrome, since...
reed@google.com [Thu, 31 Oct 2013 17:28:30 +0000 (17:28 +0000)]
Revert "Revert "add SK_ATTR_DEPRECATED -- will need to disable for chrome, since it triggers a warning""

This reverts commit 1e787c38fa71f2a21fd728f1b1d620b9b09b0d3d.

BUG=

Review URL: https://codereview.chromium.org/54603004

git-svn-id: http://skia.googlecode.com/svn/trunk@12057 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoRevert "add SK_ATTR_DEPRECATED -- will need to disable for chrome, since it triggers...
reed@google.com [Thu, 31 Oct 2013 16:31:11 +0000 (16:31 +0000)]
Revert "add SK_ATTR_DEPRECATED -- will need to disable for chrome, since it triggers a warning"

This reverts commit 1d22c4aaf9d8f053f25194a1ed74b137bfb19497.

git-svn-id: http://skia.googlecode.com/svn/trunk@12056 2bbb7eff-a529-9590-31e7-b0007b416f81

11 years agoadd SK_ATTR_DEPRECATED -- will need to disable for chrome, since it triggers a warning
reed@google.com [Thu, 31 Oct 2013 16:24:08 +0000 (16:24 +0000)]
add SK_ATTR_DEPRECATED -- will need to disable for chrome, since it triggers a warning

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

Review URL: https://codereview.chromium.org/51033004

git-svn-id: http://skia.googlecode.com/svn/trunk@12055 2bbb7eff-a529-9590-31e7-b0007b416f81