platform/upstream/libSkiaSharp.git
9 years agoRemove a pointless use of SkWeakRefCnt.
mtklein [Fri, 24 Oct 2014 17:43:15 +0000 (10:43 -0700)]
Remove a pointless use of SkWeakRefCnt.

Can't quite get rid of SkWeakRefCnt yet... SkFontMgr_indirect uses it to cache
SkTypefaces, and I don't quite understand it enough yet to cut out the weak refs.

BUG=skia:3065

Review URL: https://codereview.chromium.org/664173003

9 years agoPrint GPU cache stats in nanobench/dm with veryVerbose
bsalomon [Fri, 24 Oct 2014 17:40:50 +0000 (10:40 -0700)]
Print GPU cache stats in nanobench/dm with veryVerbose

Review URL: https://codereview.chromium.org/680553002

9 years agoAdd harfbuzz to DEPS
mtklein [Fri, 24 Oct 2014 17:20:22 +0000 (10:20 -0700)]
Add harfbuzz to DEPS

https://skia.googlesource.com/third_party/harfbuzz/

BUG=skia:3067

Review URL: https://codereview.chromium.org/679733002

9 years agoIgnore simpleaaclip_path on Win7-gpu for now.
jvanverth [Fri, 24 Oct 2014 16:40:43 +0000 (09:40 -0700)]
Ignore simpleaaclip_path on Win7-gpu for now.

BUG=skia:3066
TBR=mtklein@google.com

Review URL: https://codereview.chromium.org/646523005

9 years agoBuild gpu stats tracking in dev builds.
bsalomon [Fri, 24 Oct 2014 16:34:41 +0000 (09:34 -0700)]
Build gpu stats tracking in dev builds.

Fix cache stats, add more info.

BUG=skia:2889

Review URL: https://codereview.chromium.org/655263005

9 years agoadd color wheel to validate decoding colors
Hal Canary [Fri, 24 Oct 2014 16:18:46 +0000 (12:18 -0400)]
add color wheel to validate decoding colors

BUG=skia:3061
R=mtklein@google.com

Review URL: https://codereview.chromium.org/681433002

9 years agoRevert of Added varying struct (patchset #9 id:160001 of https://codereview.chromium...
joshualitt [Fri, 24 Oct 2014 15:24:08 +0000 (08:24 -0700)]
Revert of Added varying struct (patchset #9 id:160001 of https://codereview.chromium.org/671023002/)

Reason for revert:
may have caused gm change on arm

Original issue's description:
> Added varying struct
>
> TBR=
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/852ae80b9c3c6fd53f993ac35133d80863993cbe

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

Review URL: https://codereview.chromium.org/675193002

9 years agoAdd isScale() helper function to SkMatrix44.
tfarina [Fri, 24 Oct 2014 15:07:05 +0000 (08:07 -0700)]
Add isScale() helper function to SkMatrix44.

This will be used later in Chromium to cleanup gfx::Transform::IsScale2d().

BUG=408710,skia:997
TEST=None
R=bsalomon@google.com,danakj@chromium.org

Review URL: https://codereview.chromium.org/676583002

9 years agorebaseline for new shaderproc
mtklein [Fri, 24 Oct 2014 15:05:58 +0000 (08:05 -0700)]
rebaseline for new shaderproc

BUG=skia:

Review URL: https://codereview.chromium.org/674123003

9 years agoAdded varying struct
joshualitt [Fri, 24 Oct 2014 14:56:04 +0000 (07:56 -0700)]
Added varying struct

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/671023002

9 years agoRebaseline lots of images, associate some with bug 3063.
bsalomon [Fri, 24 Oct 2014 14:26:00 +0000 (07:26 -0700)]
Rebaseline lots of images, associate some with bug 3063.

TBR=
NOTRY=true

Review URL: https://codereview.chromium.org/661323003

9 years agoSet temporary paths volatile so we don't cache them.
jvanverth [Fri, 24 Oct 2014 14:12:51 +0000 (07:12 -0700)]
Set temporary paths volatile so we don't cache them.

Any path that is generated frame-to-frame should not be rendered by using the
DistanceFieldPathRenderer, because generating the initial distance field,
uploading it and rendering it takes longer than the SoftwarePathRenderer.

BUG=skia:2935

Review URL: https://codereview.chromium.org/677463002

9 years agoAdd test for new FrontBufferedStream behavior.
scroggo [Fri, 24 Oct 2014 13:55:07 +0000 (06:55 -0700)]
Add test for new FrontBufferedStream behavior.

Test for https://skia.googlesource.com/skia/+/dd5a1e094c19fa10202c37c50a1f799e5af5dac0

Verify that FrontBufferedStream does not attempt to read beyond the
end of its underlying stream.

Make SkStreamToCGImageSource handle an empty stream better.

Review URL: https://codereview.chromium.org/641813009

9 years agoDon't read random data in ICO check.
scroggo [Fri, 24 Oct 2014 13:49:57 +0000 (06:49 -0700)]
Don't read random data in ICO check.

Check to ensure SkStream::read succeeds before checking the result.

Review URL: https://codereview.chromium.org/656673005

9 years agoRe-add in v8 to DEPS.
jcgregorio [Fri, 24 Oct 2014 12:57:37 +0000 (05:57 -0700)]
Re-add in v8 to DEPS.

BUG=skia:

Review URL: https://codereview.chromium.org/673713004

9 years agoUpdate SKP version
skia-deps-roller [Fri, 24 Oct 2014 12:47:24 +0000 (05:47 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/679693002

9 years agoPossible optimization to SkString
joshualitt [Fri, 24 Oct 2014 01:23:32 +0000 (18:23 -0700)]
Possible optimization to SkString

BUG=skia:

Review URL: https://codereview.chromium.org/679433003

9 years agoCleanup: Turn GrReducedClip into a class with a static function.
tfarina [Fri, 24 Oct 2014 00:47:18 +0000 (17:47 -0700)]
Cleanup: Turn GrReducedClip into a class with a static function.

Clean up namespace usage.

Similar to what was done in
https://skia.googlesource.com/skia/+/a5414c4a8efc3119ee20fcee96c0bf68a04909c7

BUG=None
TEST=None
R=bsalomon@google.com

Review URL: https://codereview.chromium.org/653393003

9 years agoAdd flag to skimage for decoding into unpremul.
scroggo [Thu, 23 Oct 2014 22:42:11 +0000 (15:42 -0700)]
Add flag to skimage for decoding into unpremul.

Review URL: https://codereview.chromium.org/666073006

9 years agosquelch filterindiabox too
mtklein [Thu, 23 Oct 2014 22:24:11 +0000 (15:24 -0700)]
squelch filterindiabox too

TBR=bsalomon@google.com

BUG=skia:

Review URL: https://codereview.chromium.org/677463003

9 years agoFix imagemagnifier GM quality and speed.
senorblanco [Thu, 23 Oct 2014 22:00:10 +0000 (15:00 -0700)]
Fix imagemagnifier GM quality and speed.

The imagemagnifier GM was applying the filter both to a saveLayer()
restore() around the whole scene, as well as to each individual
text draw.

Applying the filter only in the saveLayer() makes it run ~20x faster
on my Linux box, and the quality is improved as well (since the
primitives are not double-filtered).

BUG=skia:

Review URL: https://codereview.chromium.org/637283009

9 years agoRevert "Revert of create shaderproc for nofilter-opaque-dx (patchset #7 id:120001...
mtklein [Thu, 23 Oct 2014 21:35:01 +0000 (14:35 -0700)]
Revert "Revert of create shaderproc for nofilter-opaque-dx (patchset #7 id:120001 of https://codereview.chromium.org/664783004/)"

This reverts commit 430b795cc8a1cdbddd8fdc5511a3a523348937f7 and adds suppressions.

BUG=skia:

Review URL: https://codereview.chromium.org/673023002

9 years agoDon't use MIP filter mode on compressed textures
bsalomon [Thu, 23 Oct 2014 21:17:46 +0000 (14:17 -0700)]
Don't use MIP filter mode on compressed textures

Add compressed texture image downsample GMs

BUG=chromium:426331

Review URL: https://codereview.chromium.org/672863003

9 years agoRemove test expectations from ignored-tests.txt.
senorblanco [Thu, 23 Oct 2014 21:01:38 +0000 (14:01 -0700)]
Remove test expectations from ignored-tests.txt.

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

Review URL: https://codereview.chromium.org/659393006

9 years agoAdding an option to render only the shadow in SkDropShadowImageFilter
sugoi [Thu, 23 Oct 2014 20:59:52 +0000 (13:59 -0700)]
Adding an option to render only the shadow in SkDropShadowImageFilter

This is basically how blink uses the filter. Currently, I can't use it for "ShadowOnly" mode with the filter at all, but instead of copying the code and risking to have the codepaths diverge, I'm simply going to add the option here.

BUG=skia:

Review URL: https://codereview.chromium.org/646213004

9 years agoRebaselines for tests affected by image filter PDF change.
senorblanco [Thu, 23 Oct 2014 20:50:54 +0000 (13:50 -0700)]
Rebaselines for tests affected by image filter PDF change.

(aka https://codereview.chromium.org/644323006/)

TBR=mtklein@google.com
NOTRY=true

BUG=skia:

Review URL: https://codereview.chromium.org/671303003

9 years agoAdd utils to better quantize grayscale values to three bit indices while
pavel [Thu, 23 Oct 2014 20:18:50 +0000 (13:18 -0700)]
Add utils to better quantize grayscale values to three bit indices while
compressing coverage masks.

Signed-off-by: Pavel Krajcevski <pavel@cs.unc.edu>
BUG=skia:

Review URL: https://codereview.chromium.org/669243003

9 years agoadd float equivalent in dumpHex output as comment
caryclark [Thu, 23 Oct 2014 20:08:56 +0000 (13:08 -0700)]
add float equivalent in dumpHex output as comment

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

Review URL: https://codereview.chromium.org/676803002

9 years agoRevert of create shaderproc for nofilter-opaque-dx (patchset #7 id:120001 of https...
mtklein [Thu, 23 Oct 2014 19:44:39 +0000 (12:44 -0700)]
Revert of create shaderproc for nofilter-opaque-dx (patchset #7 id:120001 of https://codereview.chromium.org/664783004/)

Reason for revert:
Many GMs fixed.  Needs rebaseline, perhaps layout test rebaselines.

Original issue's description:
> create shaderproc for nofilter-opaque-dx
>
>
> speedup nofilter
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/a40a276bcee2246439dcf816273c1307f5c3c69f

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

Review URL: https://codereview.chromium.org/656913005

9 years agouse SkMaskCache inside SkBlurMaskFilter
reed [Thu, 23 Oct 2014 19:42:46 +0000 (12:42 -0700)]
use SkMaskCache inside SkBlurMaskFilter

depends on https://codereview.chromium.org/670063004/

BUG=skia:

Review URL: https://codereview.chromium.org/669993003

9 years agoCorrect typo in SkRemotableFontMgr_win_dw.cpp.
bungeman [Thu, 23 Oct 2014 19:25:29 +0000 (12:25 -0700)]
Correct typo in SkRemotableFontMgr_win_dw.cpp.

New code was introduced behind a flag, but in that code there is a comma
before a parameter name instead of after.

Review URL: https://codereview.chromium.org/646323003

9 years agocreate shaderproc for nofilter-opaque-dx
reed [Thu, 23 Oct 2014 19:22:40 +0000 (12:22 -0700)]
create shaderproc for nofilter-opaque-dx

speedup nofilter

BUG=skia:

Review URL: https://codereview.chromium.org/664783004

9 years agoRevert of Add test for new FrontBufferedStream behavior. (patchset #1 id:1 of https...
mtklein [Thu, 23 Oct 2014 19:08:03 +0000 (12:08 -0700)]
Revert of Add test for new FrontBufferedStream behavior. (patchset #1 id:1 of https://codereview.chromium.org/641813009/)

Reason for revert:
Test is SkASSERTing on Macs, e.g:

http://chromegw.corp.google.com/i/client.skia/builders/Test-Mac10.6-MacMini4.1-GeForce320M-x86_64-Debug/builds/257/steps/dm/logs/stdio

210 tasks left 1737M peak     1ms test FrontBufferedStream../../src/ports/SkImageDecoder_CG.cpp:43: failed assertion "data"

Signal 11:
_sigtramp (+0x1a)
SkStreamToCGImageSource(SkStream*) (+0x62)
SkImageDecoder_CG::onDecode(SkStream*, SkBitmap*, SkImageDecoder::Mode) (+0x2e)
SkImageDecoder::decode(SkStream*, SkBitmap*, SkColorType, SkImageDecoder::Mode) (+0x81)
SkImageDecoder::DecodeStream(SkStreamRewindable*, SkBitmap*, SkColorType, SkImageDecoder::Mode, SkImageDecoder::Format*) (+0xff)
SkImageDecoder::DecodeStream(SkStreamRewindable*, SkBitmap*) (+0x31)
test_ShortFrontBufferedStream(skiatest::Reporter*) (+0x97)
skiatest::ShortFrontBufferedStreamClass::onRun(skiatest::Reporter*) (+0x19)
skiatest::Test::run() (+0x7c)
DM::CpuTestTask::draw() (+0x5a)
DM::CpuTask::run() (+0x9e)
non-virtual thunk to DM::CpuTask::run() (+0x1c)
(anonymous namespace)::ThreadPool::Loop(void*) (+0xf4)
thread_start(void*) (+0x54)
_pthread_start (+0x14b)

Original issue's description:
> Add test for new FrontBufferedStream behavior.
>
> Test for https://skia.googlesource.com/skia/+/dd5a1e094c19fa10202c37c50a1f799e5af5dac0
>
> Verify that FrontBufferedStream does not attempt to read beyond the
> end of its underlying stream.
>
> Committed: https://skia.googlesource.com/skia/+/da59f05c6738dbb9a92cad21c608cdfae53a76b2

TBR=reed@google.com,scroggo@google.com
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/649553003

9 years agoAllocate only the vertices we need for text.
jvanverth [Thu, 23 Oct 2014 18:57:12 +0000 (11:57 -0700)]
Allocate only the vertices we need for text.

This restructures the vertex allocation for text rendering to compute
the max number of vertices we would need for a line of text, and then
only allocate that much. If this number exceeds the quad index limit,
then it will allocate for the max number of quads, and reallocate for
the rest later.

Review URL: https://codereview.chromium.org/663423003

9 years agoupdating k1 expectations
joshualitt [Thu, 23 Oct 2014 18:34:07 +0000 (11:34 -0700)]
updating k1 expectations

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/642543003

9 years agoRemoved simple tests steps from CanvasTest.
piotaixr [Thu, 23 Oct 2014 17:59:07 +0000 (10:59 -0700)]
Removed simple tests steps from CanvasTest.

CanvasTest only checks if the state of the different canvas
implementations are consistent with the one of the default
implementation.

Because most of the draw calls are stateless and already tested in GMs
the test steps that exercices them are not useful.

BUG=skia:3054

Review URL: https://codereview.chromium.org/672273002

9 years agoRemove android specific genID from SkPaint.
djsollen [Thu, 23 Oct 2014 17:23:07 +0000 (10:23 -0700)]
Remove android specific genID from SkPaint.

Review URL: https://codereview.chromium.org/677453002

9 years agoAdd test for new FrontBufferedStream behavior.
scroggo [Thu, 23 Oct 2014 17:21:28 +0000 (10:21 -0700)]
Add test for new FrontBufferedStream behavior.

Test for https://skia.googlesource.com/skia/+/dd5a1e094c19fa10202c37c50a1f799e5af5dac0

Verify that FrontBufferedStream does not attempt to read beyond the
end of its underlying stream.

Review URL: https://codereview.chromium.org/641813009

9 years agoThis return_false is now only used on Android. Squelch a warning.
mtklein [Thu, 23 Oct 2014 17:13:21 +0000 (10:13 -0700)]
This return_false is now only used on Android.  Squelch a warning.

BUG=skia:

Review URL: https://codereview.chromium.org/659393004

9 years agoRename Nexus9 expectations
borenet [Thu, 23 Oct 2014 15:08:23 +0000 (08:08 -0700)]
Rename Nexus9 expectations

BUG=skia:3047
TBR=djsollen

Review URL: https://codereview.chromium.org/652003003

9 years agoAdd SkMaskCache
qiankun.miao [Thu, 23 Oct 2014 14:58:17 +0000 (07:58 -0700)]
Add SkMaskCache

BUG=skia:

Review URL: https://codereview.chromium.org/670063004

9 years agoRevert of Delete unused method aliases and texture flag. (patchset #1 id:1 of https...
bsalomon [Thu, 23 Oct 2014 14:41:29 +0000 (07:41 -0700)]
Revert of Delete unused method aliases and texture flag. (patchset #1 id:1 of https://codereview.chromium.org/672753002/)

Reason for revert:
Need to remove one more reference to this in blink before relanding.

Original issue's description:
> Delete unused method aliases and texture flag.
>
> Committed: https://skia.googlesource.com/skia/+/7fdffe41494cbd78a8ee2a6ef0509e46ec77039f

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

Review URL: https://codereview.chromium.org/671203002

9 years agoUpdate fontMgr to take list of bcp47 language tags.
bungeman [Thu, 23 Oct 2014 14:08:05 +0000 (07:08 -0700)]
Update fontMgr to take list of bcp47 language tags.

This will enable clients to pass more than one bcp47 tag to ensure
that the most appropriate font is selected.

BUG=chromium:422180

Review URL: https://codereview.chromium.org/670243002

9 years agoDelete unused method aliases and texture flag.
bsalomon [Thu, 23 Oct 2014 13:56:55 +0000 (06:56 -0700)]
Delete unused method aliases and texture flag.

Review URL: https://codereview.chromium.org/672753002

9 years agofix memory leak
joshualitt [Thu, 23 Oct 2014 03:53:24 +0000 (20:53 -0700)]
fix memory leak

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/668333005

9 years agosome more expectations, update ignores
joshualitt [Thu, 23 Oct 2014 03:23:07 +0000 (20:23 -0700)]
some more expectations, update ignores

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/668393002

9 years agoNew baselines
joshualitt [Thu, 23 Oct 2014 02:47:36 +0000 (19:47 -0700)]
New baselines
TBR=

BUG=skia:

Review URL: https://codereview.chromium.org/668823006

9 years agoone more test to ignore
joshualitt [Thu, 23 Oct 2014 00:52:31 +0000 (17:52 -0700)]
one more test to ignore

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/674713002

9 years agoOval and stroke AA rect now batch
joshualitt [Wed, 22 Oct 2014 23:00:29 +0000 (16:00 -0700)]
Oval and stroke AA rect now batch

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/bc54fab1a4e5b51628a8c1557c62bc74e078870d

Committed: https://skia.googlesource.com/skia/+/18055afb838a278b5a8436cd51dbfbb688e1e0a0

Review URL: https://codereview.chromium.org/664193002

9 years agoRevert of Oval and stroke AA rect now batch (patchset #7 id:110001 of https://coderev...
joshualitt [Wed, 22 Oct 2014 22:51:52 +0000 (15:51 -0700)]
Revert of Oval and stroke AA rect now batch (patchset #7 id:110001 of https://codereview.chromium.org/664193002/)

Reason for revert:
whoops, accidental commit

Original issue's description:
> Oval and stroke AA rect now batch
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/bc54fab1a4e5b51628a8c1557c62bc74e078870d
>
> Committed: https://skia.googlesource.com/skia/+/18055afb838a278b5a8436cd51dbfbb688e1e0a0

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

Review URL: https://codereview.chromium.org/675623002

9 years agoOval and stroke AA rect now batch
joshualitt [Wed, 22 Oct 2014 22:48:48 +0000 (15:48 -0700)]
Oval and stroke AA rect now batch

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/bc54fab1a4e5b51628a8c1557c62bc74e078870d

Review URL: https://codereview.chromium.org/664193002

9 years agoRevert of Oval and stroke AA rect now batch (patchset #7 id:110001 of https://coderev...
mtklein [Wed, 22 Oct 2014 22:23:29 +0000 (15:23 -0700)]
Revert of Oval and stroke AA rect now batch (patchset #7 id:110001 of https://codereview.chromium.org/664193002/)

Reason for revert:
Many GMs failing, e.g. http://build.chromium.org/p/client.skia.android/builders/Test-Android-Nexus5-Adreno330-Arm7-Release/builds/95/steps/Compare%20GMs/logs/stdio

Original issue's description:
> Oval and stroke AA rect now batch
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/bc54fab1a4e5b51628a8c1557c62bc74e078870d

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

Review URL: https://codereview.chromium.org/671993003

9 years agoOval and stroke AA rect now batch
joshualitt [Wed, 22 Oct 2014 21:41:44 +0000 (14:41 -0700)]
Oval and stroke AA rect now batch

BUG=skia:

Review URL: https://codereview.chromium.org/664193002

9 years agoAdd SkTypeface::getBounds()
reed [Wed, 22 Oct 2014 20:20:58 +0000 (13:20 -0700)]
Add SkTypeface::getBounds()

mirrored from https://codereview.chromium.org/666303002/

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/676523002

9 years agofix some bit-rot in the ios port of sampleapp
reed [Wed, 22 Oct 2014 20:06:00 +0000 (13:06 -0700)]
fix some bit-rot in the ios port of sampleapp

BUG=skia:

Review URL: https://codereview.chromium.org/637263004

9 years agoQualify the return value of SkImageDecoder::decode
scroggo [Wed, 22 Oct 2014 19:07:00 +0000 (12:07 -0700)]
Qualify the return value of SkImageDecoder::decode

Add a new enum to differentiate between a complete decode and a
partial decode (with the third value being failure). Return this
value from SkImageDecoder::onDecode (in all subclasses, plus
SkImageDecoder_empty) and ::decode.

For convenience, if the enum is treated as a boolean, success and
partial success are both considered true.

Note that the static helper functions (DecodeFile etc) still return
true and false (for one thing, this allows us to continue to use
SkImageDecoder::DecodeMemory as an SkPicture::InstallPixelRefProc in
SkPicture::CreateFromStream).

Also correctly report failure in SkASTCImageDecoder::onDecode when
SkTextureCompressor::DecompressBufferFromFormat fails.

BUG=skia:3037
BUG:b/17419670

Review URL: https://codereview.chromium.org/647023006

9 years agowebtry: Improve error handling/message when creating sqlite tables.
tfarina [Wed, 22 Oct 2014 18:57:36 +0000 (11:57 -0700)]
webtry: Improve error handling/message when creating sqlite tables.

In case they fail, lets fail with an ERROR message rather than an INFO,
and lets make the error message a little bit better.

BUG=None
TEST=None
R=jcgregorio@google.com

Review URL: https://codereview.chromium.org/669643004

9 years agoexplicitly size fType in SkRRect, to ensure the are no uninitialized bytes
reed [Wed, 22 Oct 2014 18:23:56 +0000 (11:23 -0700)]
explicitly size fType in SkRRect, to ensure the are no uninitialized bytes
(so it can be used as a key in a cache)

BUG=skia:

Review URL: https://codereview.chromium.org/670183002

9 years agoSkResourceCache::Key namespace support.
fmalita [Wed, 22 Oct 2014 18:20:40 +0000 (11:20 -0700)]
SkResourceCache::Key namespace support.

Add a unique-per-subclass namespace tag to make Keys from different
domains comparable.

Also drop the SkPictureShader cache and convert to using the global
resource cache instead.

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

Review URL: https://codereview.chromium.org/668223002

9 years agoOverride SkCanvas::drawImage() in SkDeferredCanvas and SkGPipe
piotaixr [Wed, 22 Oct 2014 18:02:32 +0000 (11:02 -0700)]
Override SkCanvas::drawImage() in SkDeferredCanvas and SkGPipe

Depend on https://codereview.chromium.org/663233002

BUG=skia:2947

Committed: https://skia.googlesource.com/skia/+/687732fe046c9cfec940f2d7f661cd97651d10a8

Review URL: https://codereview.chromium.org/613673005

9 years agoGP Emit Code takes a struct
joshualitt [Wed, 22 Oct 2014 17:56:26 +0000 (10:56 -0700)]
GP Emit Code takes a struct

BUG=skia:

Review URL: https://codereview.chromium.org/666693012

9 years agoMore genericity: overload isPaintOpaque(SkPaint, SkBitmap)
piotaixr [Wed, 22 Oct 2014 15:35:44 +0000 (08:35 -0700)]
More genericity: overload isPaintOpaque(SkPaint, SkBitmap)

Instead of taking a Bitmap as an argument this version
takes only the type of content (None/Opaque/Transparent).
This will be used to check the opaqueness of a SkPaint
that draws a SkImage.

BUG=skia:3042

Review URL: https://codereview.chromium.org/663233002

9 years agoUse correct size for font style name search.
bungeman [Wed, 22 Oct 2014 15:25:44 +0000 (08:25 -0700)]
Use correct size for font style name search.

This fixes an out of bounds read introduced with
a4c4a2d8cd65abb1e5ac20813831cdb9ace6c7ee
"Replace SkTypeface::Style with SkFontStyle."

BUG=chromium:425856

Review URL: https://codereview.chromium.org/645223006

9 years agoPicture shader resource caching.
fmalita [Wed, 22 Oct 2014 14:39:08 +0000 (07:39 -0700)]
Picture shader resource caching.

Replace the current/naive shader caching mechanism with a more general
implementation based on SkResourceCache.

Caching the bitmap shader itself (as opposed to just the tile bitmap)
makes for a chunkier key, but OTOH avoids allocating new shaders on
cache hit.

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

Review URL: https://codereview.chromium.org/671683004

9 years agoIgnore failure on internal bot
joshualitt [Wed, 22 Oct 2014 14:29:21 +0000 (07:29 -0700)]
Ignore failure on internal bot

BUG=skia:3023
TBR=egdaniel@google.com

Review URL: https://codereview.chromium.org/640063004

9 years agoAdd back page_sets/skia_googlespreadsheet_desktop
rmistry [Wed, 22 Oct 2014 13:19:51 +0000 (06:19 -0700)]
Add back page_sets/skia_googlespreadsheet_desktop

Absolutely could not get the old page to work so I just created a new page.

NOTRY=true
BUG=skia:3008

Review URL: https://codereview.chromium.org/671883002

9 years agoTrack nested picture xform state for layer hoisting
robertphillips [Wed, 22 Oct 2014 12:33:52 +0000 (05:33 -0700)]
Track nested picture xform state for layer hoisting

The accumulated matrix state of any enclosing SkPictures must be stored separate from the picture-local CTM. Any setMatrix calls inside a layer need to replace the picture-local CTM but concatenate with the enclosing SkPicture transform state (and the transform state needed to translate the layer to the correct location in the cached GrTexture).

Review URL: https://codereview.chromium.org/639863005

9 years agoUpdate webpages_playback script and all page_sets to use new skpicture_printer benchmark
rmistry [Wed, 22 Oct 2014 12:04:43 +0000 (05:04 -0700)]
Update webpages_playback script and all page_sets to use new skpicture_printer benchmark

The purpose of this CL is to turn the RecreateSKPs bot green after
https://codereview.chromium.org/665193002/ lands.

Description of changes in this CL:
* After https://codereview.chromium.org/665193002/ lands we will no longer need to copy the archives and pagesets to telemetry's data directory, I removed the code that did this in https://codereview.chromium.org/655313003/
* page_sets now need unique class names. Updated all page_sets to have unique class names.
* Using the new --page_set_name and --page_set_base_dir and --also-run-disabled-tests parameters

BUG=skia:3046

Review URL: https://codereview.chromium.org/671513009

9 years agoBug fix for es 3.00 fb fetch
joshualitt [Tue, 21 Oct 2014 19:53:15 +0000 (12:53 -0700)]
Bug fix for es 3.00 fb fetch

patch for es 300 shader fb fetch

BUG=skia:

Review URL: https://codereview.chromium.org/665893008

9 years agoadd round/ceil/etc. for SkMScalar
reed [Tue, 21 Oct 2014 19:33:21 +0000 (12:33 -0700)]
add round/ceil/etc. for SkMScalar

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/645793006

9 years agoDraw SKPs in 256x256 tiles in nanobench.
mtklein [Tue, 21 Oct 2014 19:29:25 +0000 (12:29 -0700)]
Draw SKPs in 256x256 tiles in nanobench.

(This CL will certainly trigger performance regression alerts.  Tiled drawing is slower than non-tiled drawing.)

BUG=skia:

Review URL: https://codereview.chromium.org/669983002

9 years agoImplement maxrss for Windows
mtklein [Tue, 21 Oct 2014 19:23:12 +0000 (12:23 -0700)]
Implement maxrss for Windows

No more -1 in the first column on Windows!

E.g. http://build.chromium.org/p/client.skia/builders/Perf-Win8-ShuttleA-HD7770-x86_64-Release-Trybot/builds/0/steps/nanobench/logs/stdio

BUG=skia:

Review URL: https://codereview.chromium.org/661413004

9 years agoSkLazyPtr follow ups
mtklein [Tue, 21 Oct 2014 19:20:04 +0000 (12:20 -0700)]
SkLazyPtr follow ups

 - moves test to LazyPtrTest.cpp
 - adds the ability to use a Create() method with an argument

BUG=skia:

Review URL: https://codereview.chromium.org/669783002

9 years agoWiden the set of nested rects that can be drawn natively on the GPU
robertphillips [Tue, 21 Oct 2014 18:25:37 +0000 (11:25 -0700)]
Widen the set of nested rects that can be drawn natively on the GPU

GrAARectRenderer::fillAANestedRects only handles thin stroked rects correctly if the margins are all the same. It is also correct if all the margins are >= 1.0f. This CL allows such cases to use the fillAANestedRects fast path.

This seems to halve the gpu rendering time of the picture in bug crbug.com/425427.

Review URL: https://codereview.chromium.org/672473002

9 years agosmall tweaks to SkResourceCache::Key
mtklein [Tue, 21 Oct 2014 17:55:22 +0000 (10:55 -0700)]
small tweaks to SkResourceCache::Key
  - remove unused method and obsolete note
  - add a reminder that yes, we are checking counts are equal

BUG=skia:

Review URL: https://codereview.chromium.org/667993002

9 years agouse - and = to zoom text in slides
reed [Tue, 21 Oct 2014 17:46:01 +0000 (10:46 -0700)]
use - and = to zoom text in slides

BUG=skia:
TBR=
NOTRY=True

Review URL: https://codereview.chromium.org/668023002

9 years agoFix debugger's SaveAs functionality
robertphillips [Tue, 21 Oct 2014 17:31:38 +0000 (10:31 -0700)]
Fix debugger's SaveAs functionality

Without this patch the SaveAs functionality only copies the portion of the picture from the current command in the GUI.

Review URL: https://codereview.chromium.org/672453002

9 years agoCheck SkStream::isAtEnd before attempting to read.
scroggo [Tue, 21 Oct 2014 15:06:05 +0000 (08:06 -0700)]
Check SkStream::isAtEnd before attempting to read.

In FrontBufferedStream, when attempting to read and buffer more
data, do not try to read if the underlying stream is at the end.

Prevents a bug where we continue trying to read a stream that
has nothing more to read.

Review URL: https://codereview.chromium.org/665303002

9 years agoAdd check to see if we need to remove fixed funciton VA in OptDrawState
egdaniel [Tue, 21 Oct 2014 14:47:10 +0000 (07:47 -0700)]
Add check to see if we need to remove fixed funciton VA in OptDrawState

BUG=skia:

Review URL: https://codereview.chromium.org/667033003

9 years agoRebaseline dftext GM for new color emoji test.
jvanverth [Tue, 21 Oct 2014 14:39:51 +0000 (07:39 -0700)]
Rebaseline dftext GM for new color emoji test.

BUG=skia:3033
TBR=mtklein@google.com

Review URL: https://codereview.chromium.org/671513007

9 years agoClean up some warnings about unused code.
mtklein [Tue, 21 Oct 2014 14:21:36 +0000 (07:21 -0700)]
Clean up some warnings about unused code.

BUG=skia:

Review URL: https://codereview.chromium.org/663093004

9 years agoadd gradients and drawPaint to lua
reed [Tue, 21 Oct 2014 14:12:58 +0000 (07:12 -0700)]
add gradients and drawPaint to lua

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/667803004

9 years agoAdd an assert that we still have a buffer to readFromBuffer().
mtklein [Tue, 21 Oct 2014 14:12:52 +0000 (07:12 -0700)]
Add an assert that we still have a buffer to readFromBuffer().

sk_free(fBuffer.detach()) means fBuffer will be left as NULL, not a zero-length malloc.

BUG=skia:

Review URL: https://codereview.chromium.org/669813002

9 years agoFix image filters for PDF backend.
senorblanco [Mon, 20 Oct 2014 21:03:12 +0000 (14:03 -0700)]
Fix image filters for PDF backend.

Currently, the PDF backend does not support image filters (since PDF
does not have that functionality), so it simply removes them. This is
causing Chrome print preview to render incorrectly (see bug). The fix
here is to fall back to a raster device for image filters, as we used
to do in Blink. The resulting bitmap will be drawn to the destination
device as a normal main-memory-backed bitmap.

Note: this change invalidates the PDF results of all GMs containing
image filters (since they'll actually be rendered).

BUG=422144

Review URL: https://codereview.chromium.org/644323006

9 years agoSkLazyPtr suitable as a local or class member.
mtklein [Mon, 20 Oct 2014 20:46:11 +0000 (13:46 -0700)]
SkLazyPtr suitable as a local or class member.

BUG=skia:

Review URL: https://codereview.chromium.org/653183006

9 years agoDilate approximated text bounds to squelch recent assertion failure.
mtklein [Mon, 20 Oct 2014 20:44:24 +0000 (13:44 -0700)]
Dilate approximated text bounds to squelch recent assertion failure.

This is once again an issue related to logo fonts, so I don't
see any easy way to add a regression test for this.

BUG=424824

Review URL: https://codereview.chromium.org/665103002

9 years agoReplace SkTypeface::Style with SkFontStyle.
bungeman [Mon, 20 Oct 2014 20:33:19 +0000 (13:33 -0700)]
Replace SkTypeface::Style with SkFontStyle.

Committed: https://skia.googlesource.com/skia/+/43b8b36b20ae00e2d78421c4cda1f3f922983a20

Review URL: https://codereview.chromium.org/488143002

9 years agoAdd color emoji fallback for distance field text.
jvanverth [Mon, 20 Oct 2014 20:12:54 +0000 (13:12 -0700)]
Add color emoji fallback for distance field text.

BUG=skia:3033

Review URL: https://codereview.chromium.org/670533002

9 years agoStop doing at-exit cleanup of lazy pointers.
mtklein [Mon, 20 Oct 2014 20:07:03 +0000 (13:07 -0700)]
Stop doing at-exit cleanup of lazy pointers.

There are only a fixed number of them, and we're already intentionally leaking in Release.

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu13.10-GCE-NoGPU-x86_64-Debug-ASAN-Trybot

BUG=skia:2940

Review URL: https://codereview.chromium.org/606013004

9 years agoRevert of Replace SkTypeface::Style with SkFontStyle. (patchset #9 id:160001 of https...
mtklein [Mon, 20 Oct 2014 19:54:31 +0000 (12:54 -0700)]
Revert of Replace SkTypeface::Style with SkFontStyle. (patchset #9 id:160001 of https://codereview.chromium.org/488143002/)

Reason for revert:
CrOS GM failures:

[*] 2 ExpectationsMismatch: fontmgr_iter_565.png fontmgr_iter_8888.png

Original issue's description:
> Replace SkTypeface::Style with SkFontStyle.
>
> Committed: https://skia.googlesource.com/skia/+/43b8b36b20ae00e2d78421c4cda1f3f922983a20

TBR=reed@google.com,bungeman@google.com
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/667023002

9 years agoReplace SkTypeface::Style with SkFontStyle.
bungeman [Mon, 20 Oct 2014 19:33:59 +0000 (12:33 -0700)]
Replace SkTypeface::Style with SkFontStyle.

Review URL: https://codereview.chromium.org/488143002

9 years agoMove init() to Init() so that flag.Parse() can be called first.
jcgregorio [Mon, 20 Oct 2014 17:44:35 +0000 (10:44 -0700)]
Move init() to Init() so that flag.Parse() can be called first.

Otherwise the --log_dir flag doesn't affect where glog output goes to.

TBR=humper
BUG=skia:

Review URL: https://codereview.chromium.org/667493002

9 years agoRevert of Start to vectorize SkTileGrid. (patchset #48 id:1670001 of https://coderevi...
mtklein [Mon, 20 Oct 2014 17:43:55 +0000 (10:43 -0700)]
Revert of Start to vectorize SkTileGrid. (patchset #48 id:1670001 of https://codereview.chromium.org/634543004/)

Reason for revert:
breaks chrome GPU debug bots

Original issue's description:
> Start to vectorize SkTileGrid.
>
> This adds Sk4x.h to help.
>
> BUG=skia:3041
>
> Committed: https://skia.googlesource.com/skia/+/90c7992bfc6330f070f7704d63372a0ec8410170
>
> CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu12-ShuttleA-GTX660-x86-Debug-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/958e9628d5f9a81aeafa78572cb4afc4b19a455a

TBR=reed@google.com,mtklein@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:3041

Review URL: https://codereview.chromium.org/637863005

9 years agoRemove content from public_headers.
scroggo [Mon, 20 Oct 2014 16:55:28 +0000 (09:55 -0700)]
Remove content from public_headers.

The file is unused, but we cannot safely delete it until we remove
the reference in a chromium gyp file.

BUG=skia:2350

Review URL: https://codereview.chromium.org/663913003

9 years agoAdd clip to layer cache
robertphillips [Mon, 20 Oct 2014 15:52:40 +0000 (08:52 -0700)]
Add clip to layer cache

This CL adds the clip region to the GPU layer hoisting image cache. It also switches back to the old caching behavior of using the entire CTM in the cache key rather then just the upper 2x2. This latter change is to focus more on hoisting rather then caching.

It also includes 2 smaller fixes:
a) layer's that have an image filter are no longer atlased (b.c. doing so complicates applying the image filter)

b) the result of clipping the layer's bounds to the current clip is used as the hoisted layer's size. This reduces the amount of pixels drawn to match a normal (non-hoisted) draw pass.

Review URL: https://codereview.chromium.org/640773004

9 years agoFix clearing issue in render_pictures
robertphillips [Mon, 20 Oct 2014 15:45:57 +0000 (08:45 -0700)]
Fix clearing issue in render_pictures

This bug only crops up when using the "--writeWholeImage" option in tile mode

Review URL: https://codereview.chromium.org/664103003

9 years agoRemove unused parameters from filter_texture
robertphillips [Mon, 20 Oct 2014 15:12:04 +0000 (08:12 -0700)]
Remove unused parameters from filter_texture

Trivial cleanup.

Review URL: https://codereview.chromium.org/665863003

9 years agoStart to vectorize SkTileGrid.
mtklein [Mon, 20 Oct 2014 14:12:34 +0000 (07:12 -0700)]
Start to vectorize SkTileGrid.

This adds Sk4x.h to help.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/90c7992bfc6330f070f7704d63372a0ec8410170

CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu12-ShuttleA-GTX660-x86-Debug-Trybot

Review URL: https://codereview.chromium.org/634543004

9 years agoWhen rendering df fonts, pass drawText() down to drawPosText().
jvanverth [Mon, 20 Oct 2014 13:48:59 +0000 (06:48 -0700)]
When rendering df fonts, pass drawText() down to drawPosText().

First pass at getting color emoji working for distance fields.

BUG=skia:3033

Review URL: https://codereview.chromium.org/660853003

9 years agoRemove obsolete SkRecording.
mtklein [Mon, 20 Oct 2014 13:44:58 +0000 (06:44 -0700)]
Remove obsolete SkRecording.

Can land after https://codereview.chromium.org/664583002/ does.

BUG=skia:

Only API removed.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/660903002