platform/upstream/libSkiaSharp.git
10 years agobegin clean up for scalar is double option
reed [Fri, 31 Oct 2014 13:55:41 +0000 (06:55 -0700)]
begin clean up for scalar is double option

BUG=skia:

Review URL: https://codereview.chromium.org/694693002

10 years agoTurn SkTaskGroups back on.
mtklein [Fri, 31 Oct 2014 13:38:24 +0000 (06:38 -0700)]
Turn SkTaskGroups back on.

Revert "Disable SkTaskGroup in SkMultiPictureDraw temporarily."
Revert "Revert harder, removing SkTaskGroup.cpp from core temporarily."

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/687263007

10 years agoUpgrade GPUOptimize
robertphillips [Thu, 30 Oct 2014 23:45:02 +0000 (16:45 -0700)]
Upgrade GPUOptimize

This will be a bit hairy to review.

The FillBounds and CollectLayers code has diverged significantly resulting in the rendering path seeing different bounds than the hoisting path. This CL merges the FillBounds changes into CollectLayers. A follow on CL will, hopefully, find a way to layer CollectLayers on top of FillBounds.

The only code in CollectLayers that is different from FillBounds is bracketed by "LAYER HOISTING" comments.

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/685263004

10 years agoTrying to be consistent when members are allowed to be NULL
sugoi [Thu, 30 Oct 2014 21:05:14 +0000 (14:05 -0700)]
Trying to be consistent when members are allowed to be NULL

Got a few crashes running the fuzzer locally, all related to handling NULL members/parameters in an inconsistent way.

BUG=skia:

Review URL: https://codereview.chromium.org/675013003

10 years agoUpdate the Android NDK to r10c
djsollen [Thu, 30 Oct 2014 20:27:26 +0000 (13:27 -0700)]
Update the Android NDK to r10c

Review URL: https://codereview.chromium.org/689943002

10 years agoBlind GM rebaseline for Galaxy S4
borenet [Thu, 30 Oct 2014 20:27:20 +0000 (13:27 -0700)]
Blind GM rebaseline for Galaxy S4

BUG=skia:3032

Review URL: https://codereview.chromium.org/694573002

10 years agoCleanup conversion that the new NDK compiler tries to over optimize.
djsollen [Thu, 30 Oct 2014 18:54:43 +0000 (11:54 -0700)]
Cleanup conversion that the new NDK compiler tries to over optimize.

This allows us to upgrade to NDK r10c as the dftext gm no longer fails
when compiled with -O2.

Review URL: https://codereview.chromium.org/690053002

10 years agoAdd return value on GrRecordReplaceDraw
robertphillips [Thu, 30 Oct 2014 18:54:31 +0000 (11:54 -0700)]
Add return value on GrRecordReplaceDraw

This assists debugging layer hoisting errors (e.g., when a layer is hoisted but not actually used in subsequent rendering).

Review URL: https://codereview.chromium.org/694533004

10 years agoFix member variable naming
robertphillips [Thu, 30 Oct 2014 18:49:37 +0000 (11:49 -0700)]
Fix member variable naming

This renaming was missed in (MultiPictureDraw is taskgroup aware - https://codereview.chromium.org/684923002)

TBR=reed@google.com

Review URL: https://codereview.chromium.org/654293005

10 years agoCombine nameFontStyleProc and FindByNameStyle.
bungeman [Thu, 30 Oct 2014 18:49:27 +0000 (11:49 -0700)]
Combine nameFontStyleProc and FindByNameStyle.

These two functions do the same thing, combine them.

Review URL: https://codereview.chromium.org/688133003

10 years agoAdd skimage expectations for GalaxyS4.
scroggo [Thu, 30 Oct 2014 18:44:27 +0000 (11:44 -0700)]
Add skimage expectations for GalaxyS4.

Fixes build.

Review URL: https://codereview.chromium.org/692913003

10 years agoMake RTree handle the case where the playback canvas has empty clip bounds
junov [Thu, 30 Oct 2014 18:44:19 +0000 (11:44 -0700)]
Make RTree handle the case where the playback canvas has empty clip bounds

BUG=skia:3074

Review URL: https://codereview.chromium.org/678303004

10 years agoFix bug in aggressive layer cache purging
robertphillips [Thu, 30 Oct 2014 18:39:19 +0000 (11:39 -0700)]
Fix bug in aggressive layer cache purging

A picture may possess many layers that get placed in one plot of the atlas. In this case we can only remove the plot from the plotUsage tracking structure when all the layers belonging to the picture in that plot have been removed.

Review URL: https://codereview.chromium.org/654463004

10 years agoCleanup android to ensure it can compile with clang
djsollen [Thu, 30 Oct 2014 18:39:13 +0000 (11:39 -0700)]
Cleanup android to ensure it can compile with clang

Review URL: https://codereview.chromium.org/694533002

10 years agoallow SkUserConfig.h to control NOGPU
reed [Thu, 30 Oct 2014 18:34:15 +0000 (11:34 -0700)]
allow SkUserConfig.h to control NOGPU

BUG=skia:

Review URL: https://codereview.chromium.org/685243002

10 years agoUpdate SKP version
skia.buildbots [Thu, 30 Oct 2014 18:34:04 +0000 (11:34 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/688793003

10 years agotell our grcontext to abandon its backend
reed [Thu, 30 Oct 2014 18:29:08 +0000 (11:29 -0700)]
tell our grcontext to abandon its backend

BUG=skia:

Review URL: https://codereview.chromium.org/692803003

10 years agoAdd SkDocument::getStream() method.
halcanary [Thu, 30 Oct 2014 18:29:00 +0000 (11:29 -0700)]
Add SkDocument::getStream() method.

Motivation: SkDocument_PDF can be refactored to stream all
PDF objects as SkCanvas calls are made, to save memory!

BUG=skia:3030
BUG=skia:2683

Review URL: https://codereview.chromium.org/691783002

10 years agoRevert of scale xmin and xmax for FontMetrics (patchset #1 id:1 of https://codereview...
robertphillips [Thu, 30 Oct 2014 16:59:43 +0000 (09:59 -0700)]
Revert of scale xmin and xmax for FontMetrics (patchset #1 id:1 of https://codereview.chromium.org/685153004/)

Reason for revert:
Experimental revert to see if this is blocking the DEPS roll

Original issue's description:
> scale xmin and xmax for FontMetrics
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/79c276e4fd93d9c9fd57a45bc6929c48e4415b2f

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

Review URL: https://codereview.chromium.org/691033002

10 years agoRevert harder, removing SkTaskGroup.cpp from core temporarily.
mtklein [Thu, 30 Oct 2014 15:19:19 +0000 (08:19 -0700)]
Revert harder, removing SkTaskGroup.cpp from core temporarily.

Not all linkers are seeing the SkThread code is not needed.

BUG=skia:
NOTREECHECKS=true

Review URL: https://codereview.chromium.org/673903007

10 years agoDisable SkTaskGroup in SkMultiPictureDraw temporarily.
mtklein [Thu, 30 Oct 2014 14:19:11 +0000 (07:19 -0700)]
Disable SkTaskGroup in SkMultiPictureDraw temporarily.

Think this should be enough to get rolling again.

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/687273006

10 years agoUse the most recent version of ADB from the SDK if necessary.
djsollen [Thu, 30 Oct 2014 12:57:31 +0000 (05:57 -0700)]
Use the most recent version of ADB from the SDK if necessary.

This removes ADB from the Skia checkout and also does some quick
checks to make sure the version of ADB used meets a minimum threshold

Review URL: https://codereview.chromium.org/692513003

10 years agoSupport multiple scales for dfpaths.
jvanverth [Thu, 30 Oct 2014 12:57:21 +0000 (05:57 -0700)]
Support multiple scales for dfpaths.

Adds miplevel as part of dfpath key, and scale factor so we know
how much to adjust to fit desired scale.

BUG=skia:2935

Review URL: https://codereview.chromium.org/687283002

10 years agoIncrease wait time for spreadsheetdashed pageset
rmistry [Thu, 30 Oct 2014 12:52:25 +0000 (05:52 -0700)]
Increase wait time for spreadsheetdashed pageset

BUG=skia:3049
TBR=
NOTRY=true

Review URL: https://codereview.chromium.org/692843002

10 years agoWhitespace change to see how the Android bots behave
borenet [Thu, 30 Oct 2014 12:52:16 +0000 (05:52 -0700)]
Whitespace change to see how the Android bots behave

NOTRY=true
TBR=rmistry
NOTREECHECKS=true
BUG=skia:

Review URL: https://codereview.chromium.org/693483003

10 years agoIncrease timeout of blogger pageset and re-enable weather pageset
Ravi Mistry [Thu, 30 Oct 2014 11:37:07 +0000 (07:37 -0400)]
Increase timeout of blogger pageset and re-enable weather pageset

BUG=skia:3049

10 years agobreak transitions into separate file
reed [Thu, 30 Oct 2014 03:36:05 +0000 (20:36 -0700)]
break transitions into separate file

BUG=skia:
TBR=
NOTRY=True

Review URL: https://codereview.chromium.org/692543004

10 years agocleanup reentrance logic in CMM
joshualitt [Thu, 30 Oct 2014 01:29:27 +0000 (18:29 -0700)]
cleanup reentrance logic in CMM

BUG=skia:

Review URL: https://codereview.chromium.org/684223002

10 years agoUpdate SKP version
skia.buildbots [Wed, 29 Oct 2014 23:51:53 +0000 (16:51 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/665233005

10 years agoRevert of Use SkTypeface::getBounds() in bounding-box calculations. (patchset #5...
mtklein [Wed, 29 Oct 2014 23:07:10 +0000 (16:07 -0700)]
Revert of Use SkTypeface::getBounds() in bounding-box calculations. (patchset #5 id:80001 of https://codereview.chromium.org/680363003/)

Reason for revert:
Mac mini asserting

Original issue's description:
> Use SkTypeface::getBounds() in bounding-box calculations.
>
> This should produce tighter conservative bounding boxes for text than the
> approximation code it replaces.
>
> Recording performance is neutral on my desktop.  Playback performance
> improves by up to 15% on text heavy pages, e.g.
>
>   desk_pokemonwiki.skp_1 3.24ms -> 2.83ms  0.87x
>         desk_baidu.skp_1 1.91ms -> 1.58ms  0.83x
>
> Committed: https://skia.googlesource.com/skia/+/bf8dc343df4fbdcb8af546eb68b640e011a33489
>
> CQ_EXTRA_TRYBOTS=client.skia:Test-Win7-ShuttleA-HD2000-x86-Debug-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/c51add674dfb89b988a7fbc05f41838c203f9dcd

TBR=reed@google.com,mtklein@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/690833002

10 years agouse SkTaskGroup::batch
reed [Wed, 29 Oct 2014 22:44:25 +0000 (15:44 -0700)]
use SkTaskGroup::batch

BUG=skia:

Review URL: https://codereview.chromium.org/688753002

10 years agoUse SkTypeface::getBounds() in bounding-box calculations.
mtklein [Wed, 29 Oct 2014 22:36:38 +0000 (15:36 -0700)]
Use SkTypeface::getBounds() in bounding-box calculations.

This should produce tighter conservative bounding boxes for text than the
approximation code it replaces.

Recording performance is neutral on my desktop.  Playback performance
improves by up to 15% on text heavy pages, e.g.

  desk_pokemonwiki.skp_1 3.24ms -> 2.83ms  0.87x
        desk_baidu.skp_1 1.91ms -> 1.58ms  0.83x

Committed: https://skia.googlesource.com/skia/+/bf8dc343df4fbdcb8af546eb68b640e011a33489

CQ_EXTRA_TRYBOTS=client.skia:Test-Win7-ShuttleA-HD2000-x86-Debug-Trybot

Review URL: https://codereview.chromium.org/680363003

10 years agoIncrease timeouts of wowwiki and youtube
Ravi Mistry [Wed, 29 Oct 2014 21:39:15 +0000 (17:39 -0400)]
Increase timeouts of wowwiki and youtube

10 years agoSkTaskGroup::batch(fn, args, N)
mtklein [Wed, 29 Oct 2014 21:17:13 +0000 (14:17 -0700)]
SkTaskGroup::batch(fn, args, N)

Porting QuiltTask isn't important in itself; this is mostly an API feeler.

BUG=skia:

Review URL: https://codereview.chromium.org/689673003

10 years agoTry out SkTree in nanobench.
mtklein [Wed, 29 Oct 2014 21:15:10 +0000 (14:15 -0700)]
Try out SkTree in nanobench.

Looks like a fairly large recording speed win with no playback cost.

BUG=skia:

Review URL: https://codereview.chromium.org/653023003

10 years agoMoving weather pageset since it is crashing
Ravi Mistry [Wed, 29 Oct 2014 21:03:22 +0000 (17:03 -0400)]
Moving weather pageset since it is crashing

10 years agoDiscard atlas after every MultiPictureDraw::draw
robertphillips [Wed, 29 Oct 2014 20:56:02 +0000 (13:56 -0700)]
Discard atlas after every MultiPictureDraw::draw

This is intended to prevent ghosting on tiled architectures.

This CL also defers creation of the atlas (and its texture) until it is actually needed.

Committed: https://skia.googlesource.com/skia/+/6d5b5455743414ddb11d2b8c1fe9d7959f2b853d

Review URL: https://codereview.chromium.org/678403002

10 years agoscale xmin and xmax for FontMetrics
reed [Wed, 29 Oct 2014 20:43:50 +0000 (13:43 -0700)]
scale xmin and xmax for FontMetrics

BUG=skia:

Review URL: https://codereview.chromium.org/685153004

10 years agoAdd a helper script to expose which version of adb the android scripts use.
djsollen [Wed, 29 Oct 2014 19:59:50 +0000 (12:59 -0700)]
Add a helper script to expose which version of adb the android scripts use.

Review URL: https://codereview.chromium.org/692633002

10 years agoRevert of scale xmin/xmax etc in fontmetrics (patchset #1 id:1 of https://codereview...
bungeman [Wed, 29 Oct 2014 19:46:52 +0000 (12:46 -0700)]
Revert of scale xmin/xmax etc in fontmetrics (patchset #1 id:1 of https://codereview.chromium.org/690533004/)

Reason for revert:
Breaking tests in Chromium, and not in a good way.

Original issue's description:
> scale xmin/xmax etc in fontmetrics
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/f8c2feffd68b3262f3c039ec1bd517c5f6cc9e4c

TBR=reed@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/686263002

10 years agoSeparate out GrSurfaceConfig's fields from other structs used to create GrGL* subclas...
bsalomon [Wed, 29 Oct 2014 19:41:57 +0000 (12:41 -0700)]
Separate out GrSurfaceConfig's fields from other structs used to create GrGL* subclasses of GrSurface.

BUG=skia:2889

Review URL: https://codereview.chromium.org/691523002

10 years agoremove now unused (by Chrome) ifdef
humper [Wed, 29 Oct 2014 19:41:39 +0000 (12:41 -0700)]
remove now unused (by Chrome) ifdef

BUG=skia:

Review URL: https://codereview.chromium.org/688043002

10 years agoMultiPictureDraw is taskgroup aware.
reed [Wed, 29 Oct 2014 19:36:45 +0000 (12:36 -0700)]
MultiPictureDraw is taskgroup aware.
SampleApp is multipicturedraw aware.

BUG=skia:

Review URL: https://codereview.chromium.org/684923002

10 years agoFix bounds computation of all 0-input filters.
senorblanco [Wed, 29 Oct 2014 19:36:32 +0000 (12:36 -0700)]
Fix bounds computation of all 0-input filters.

The SkRectShaderImageFilter had the same bug as previously fixed for
SkBitmapSource and SkPictureImageFilter. Rather than copy-and-paste
the implementation, this change makes all filters with 0 inputs return
their source bounds, instead of returning false.

BUG=427251

Review URL: https://codereview.chromium.org/681643003

10 years agoIssue acquire barrier in SkRefCnt::unique().
bungeman [Wed, 29 Oct 2014 19:31:33 +0000 (12:31 -0700)]
Issue acquire barrier in SkRefCnt::unique().

When unique() returns true, it must also issue an acquire barrier.
Note that this change may adversly impact SkPath performance,
but editing SkPaths is already a performance issue.

BUG=chromium:258499

No API changes.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/687293002

10 years agoBUG=skia:
joshualitt [Wed, 29 Oct 2014 19:31:28 +0000 (12:31 -0700)]
BUG=skia:

Review URL: https://codereview.chromium.org/683933003

10 years agoMoving wikipedia pageset since it is crashing
Ravi Mistry [Wed, 29 Oct 2014 18:36:15 +0000 (14:36 -0400)]
Moving wikipedia pageset since it is crashing

10 years agoChange extension of crashing page set
Ravi Mistry [Wed, 29 Oct 2014 17:55:23 +0000 (13:55 -0400)]
Change extension of crashing page set

10 years agoMoving vnexpress pageset since it is crashing
Ravi Mistry [Wed, 29 Oct 2014 17:45:17 +0000 (13:45 -0400)]
Moving vnexpress pageset since it is crashing

10 years agoWhitespace change to trigger Android builds
borenet [Wed, 29 Oct 2014 17:20:48 +0000 (10:20 -0700)]
Whitespace change to trigger Android builds

BUG=skia:
NOTREECHECKS=true

Review URL: https://codereview.chromium.org/688033002

10 years agodelete webtry from main skia repo; it's been moved to buildbots
Greg Humphreys [Wed, 29 Oct 2014 16:30:16 +0000 (12:30 -0400)]
delete webtry from main skia repo; it's been moved to buildbots

BUG=skia:

Review URL: https://codereview.chromium.org/688713002

10 years agoIncrease wait times for androidpolice and nytimes
Ravi Mistry [Wed, 29 Oct 2014 16:25:34 +0000 (12:25 -0400)]
Increase wait times for androidpolice and nytimes

10 years agoSkTaskGroup::add(fn, arg)
mtklein [Wed, 29 Oct 2014 16:21:47 +0000 (09:21 -0700)]
SkTaskGroup::add(fn, arg)

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/687953003

10 years agoclear stencil clip on draw target
joshualitt [Wed, 29 Oct 2014 15:48:18 +0000 (08:48 -0700)]
clear stencil clip on draw target

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/683673002

10 years agoUpdate adb on linux to 1.0.32
Derek Sollenberger [Wed, 29 Oct 2014 15:41:17 +0000 (11:41 -0400)]
Update adb on linux to 1.0.32

Review URL: https://codereview.chromium.org/684203002

10 years agoRevert of Use SkTypeface::getBounds() in bounding-box calculations. (patchset #3...
mtklein [Wed, 29 Oct 2014 15:33:38 +0000 (08:33 -0700)]
Revert of Use SkTypeface::getBounds() in bounding-box calculations. (patchset #3 id:40001 of https://codereview.chromium.org/680363003/)

Reason for revert:
http://build.chromium.org/p/client.skia/builders/Test-Win7-ShuttleA-HD2000-x86-Debug/builds/97/steps/nanobench/logs/stdio

Original issue's description:
> Use SkTypeface::getBounds() in bounding-box calculations.
>
> This should produce tighter conservative bounding boxes for text than the
> approximation code it replaces.
>
> Recording performance is neutral on my desktop.  Playback performance
> improves by up to 15% on text heavy pages, e.g.
>
>   desk_pokemonwiki.skp_1 3.24ms -> 2.83ms  0.87x
>         desk_baidu.skp_1 1.91ms -> 1.58ms  0.83x
>
> Committed: https://skia.googlesource.com/skia/+/bf8dc343df4fbdcb8af546eb68b640e011a33489

TBR=reed@google.com,mtklein@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/685173002

10 years agopagesets are still crashing in RecreateSKPs bot, try adding waits to most pagesets
rmistry [Wed, 29 Oct 2014 15:19:04 +0000 (08:19 -0700)]
pagesets are still crashing in RecreateSKPs bot, try adding waits to most pagesets

BUG=skia:3049
NOTRY=true
TBR=borenet

Review URL: https://codereview.chromium.org/682573004

10 years agoRevert of Discard atlas after every MultiPictureDraw::draw (patchset #4 id:60001...
robertphillips [Wed, 29 Oct 2014 15:17:35 +0000 (08:17 -0700)]
Revert of Discard atlas after every MultiPictureDraw::draw (patchset #4 id:60001 of https://codereview.chromium.org/678403002/)

Reason for revert:
Breaking bots

Original issue's description:
> Discard atlas after every MultiPictureDraw::draw
>
> This is intended to prevent ghosting on tiled architectures.
>
> This CL also defers creation of the atlas (and its texture) until it is actually needed.
>
> Committed: https://skia.googlesource.com/skia/+/6d5b5455743414ddb11d2b8c1fe9d7959f2b853d

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

Review URL: https://codereview.chromium.org/687233002

10 years agoUse SkTypeface::getBounds() in bounding-box calculations.
mtklein [Wed, 29 Oct 2014 15:12:07 +0000 (08:12 -0700)]
Use SkTypeface::getBounds() in bounding-box calculations.

This should produce tighter conservative bounding boxes for text than the
approximation code it replaces.

Recording performance is neutral on my desktop.  Playback performance
improves by up to 15% on text heavy pages, e.g.

  desk_pokemonwiki.skp_1 3.24ms -> 2.83ms  0.87x
        desk_baidu.skp_1 1.91ms -> 1.58ms  0.83x

Review URL: https://codereview.chromium.org/680363003

10 years agoDiscard atlas after every MultiPictureDraw::draw
robertphillips [Wed, 29 Oct 2014 15:05:21 +0000 (08:05 -0700)]
Discard atlas after every MultiPictureDraw::draw

This is intended to prevent ghosting on tiled architectures.

This CL also defers creation of the atlas (and its texture) until it is actually needed.

Review URL: https://codereview.chromium.org/678403002

10 years agofix issues with simultaneous configs and security jail on skfiddle
Greg Humphreys [Wed, 29 Oct 2014 14:57:10 +0000 (14:57 +0000)]
fix issues with simultaneous configs and security jail on skfiddle

BUG=skia:
TBR=jcgregorio

Review URL: https://codereview.chromium.org/688003002

10 years agoscale xmin/xmax etc in fontmetrics
reed [Wed, 29 Oct 2014 14:53:24 +0000 (07:53 -0700)]
scale xmin/xmax etc in fontmetrics

BUG=skia:

Review URL: https://codereview.chromium.org/690533004

10 years agorework GUI, scripts, and fiddle main for multiple simultaneous configs
Greg Humphreys [Wed, 29 Oct 2014 14:31:33 +0000 (10:31 -0400)]
rework GUI, scripts, and fiddle main for multiple simultaneous configs

You can now select Raster, Ganesh, or both (PDF coming soon), and see
all output simultaneously.

BUG=skia:
R=jcgregorio@google.com
TBR=jcgregorio

Review URL: https://codereview.chromium.org/688433002

10 years agoAdd wait to amazon page set to avoid tab crashes
rmistry [Wed, 29 Oct 2014 13:40:27 +0000 (06:40 -0700)]
Add wait to amazon page set to avoid tab crashes

BUG=skia:3049
TBR=borenet
NOTRY=true

Review URL: https://codereview.chromium.org/686133002

10 years agoRename GalaxyNexus GM and SkImage expectations for GalaxyS4, remove NexusS
borenet [Wed, 29 Oct 2014 13:21:27 +0000 (06:21 -0700)]
Rename GalaxyNexus GM and SkImage expectations for GalaxyS4, remove NexusS

GMs will surely need rebaselining.

BUG=skia:3032

Review URL: https://codereview.chromium.org/683853006

10 years agoSkV8Sample: Now with Path2D and Path2DBuilder.
jcgregorio [Wed, 29 Oct 2014 12:33:27 +0000 (05:33 -0700)]
SkV8Sample: Now with Path2D and Path2DBuilder.

BUG=skia:

Review URL: https://codereview.chromium.org/661033005

10 years agoOptState owns program descriptor
joshualitt [Wed, 29 Oct 2014 00:59:26 +0000 (17:59 -0700)]
OptState owns program descriptor

BUG=skia:

Review URL: https://codereview.chromium.org/674543004

10 years agoone more test updated
joshualitt [Wed, 29 Oct 2014 00:23:35 +0000 (17:23 -0700)]
one more test updated

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/681373003

10 years agorebaselines
joshualitt [Wed, 29 Oct 2014 00:07:44 +0000 (17:07 -0700)]
rebaselines

TBR=
BUG=skia:

Review URL: https://codereview.chromium.org/682123003

10 years agofix typo in legacy flag alias
bsalomon [Tue, 28 Oct 2014 22:17:38 +0000 (15:17 -0700)]
fix typo in legacy flag alias

TBR=joshualitt@google.com
NOTRY=true

Review URL: https://codereview.chromium.org/675173005

10 years agoExtend SkFontMgr_Custom to cover ttc, otf, pfb.
bungeman [Tue, 28 Oct 2014 22:07:23 +0000 (15:07 -0700)]
Extend SkFontMgr_Custom to cover ttc, otf, pfb.

Adds ability to test ScanFont with Type1 (pfb) fonts
and fonts with multiple faces (ttc).

Committed: https://skia.googlesource.com/skia/+/ee0c2e4fd429424beaa35f29e7f656997ba3f115

Review URL: https://codereview.chromium.org/672723002

10 years agorename GrTextureDesc->GrSurfaceDesc, GrTextureFlags->GrSurfaceFlags
bsalomon [Tue, 28 Oct 2014 21:33:06 +0000 (14:33 -0700)]
rename GrTextureDesc->GrSurfaceDesc, GrTextureFlags->GrSurfaceFlags

Review URL: https://codereview.chromium.org/682223002

10 years agoUpdated all page sets to capture new set of webpage archives.
rmistry [Tue, 28 Oct 2014 21:25:17 +0000 (14:25 -0700)]
Updated all page sets to capture new set of webpage archives.

Fixed the record_wpr retry mechanism in webpages_playback.py.

Updated all page sets to remove the no longer needed RunSmoothness and RunNavigateSteps.

Updated the key_silk_cases pagesets to point to a non svn location:
* skia_fontwipe_desktop.py
* skia_silkfinance_desktop.py

Removed mercurynews, record_wpr was timing out because of a popup.

Added back the following pagesets which were previously disabled but are now working again:
* skia_cnn_nexus10.py
* skia_wordpress_desktop.py
* skia_transformice_nexus10.py

BUG=skia:3049

Review URL: https://codereview.chromium.org/684863002

10 years agomore ignores
joshualitt [Tue, 28 Oct 2014 20:22:56 +0000 (13:22 -0700)]
more ignores
TBR=

BUG=skia:

Review URL: https://codereview.chromium.org/688443002

10 years agoRevert of Extend SkFontMgr_Custom to cover ttc, otf, pfb. (patchset #2 id:20001 of...
bungeman [Tue, 28 Oct 2014 19:03:55 +0000 (12:03 -0700)]
Revert of Extend SkFontMgr_Custom to cover ttc, otf, pfb. (patchset #2 id:20001 of https://codereview.chromium.org/672723002/)

Reason for revert:
Reverting as this appears to be causing issues with color fonts on Android.

Original issue's description:
> Extend SkFontMgr_Custom to cover ttc, otf, pfb.
>
> Adds ability to test ScanFont with Type1 (pfb) fonts
> and fonts with multiple faces (ttc).
>
> Committed: https://skia.googlesource.com/skia/+/ee0c2e4fd429424beaa35f29e7f656997ba3f115

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

Review URL: https://codereview.chromium.org/685803003

10 years agoRemoving dead code from CanvasTest
piotaixr [Tue, 28 Oct 2014 18:13:59 +0000 (11:13 -0700)]
Removing dead code from CanvasTest

BUG=skia:3054

Review URL: https://codereview.chromium.org/681083002

10 years agoDistance field path optimizations and clean up.
jvanverth [Tue, 28 Oct 2014 18:10:48 +0000 (11:10 -0700)]
Distance field path optimizations and clean up.

Adds the following:
- Use cached geometry processor rather than recreating all the time.
- Use context's quad index buffer.

Review URL: https://codereview.chromium.org/683923002

10 years agoRevert of harden pathops for pathological test (patchset #19 id:410001 of https:...
hcm [Tue, 28 Oct 2014 17:55:54 +0000 (10:55 -0700)]
Revert of harden pathops for pathological test (patchset #19 id:410001 of https://codereview.chromium.org/633393002/)

Reason for revert:
Compile errors on bots

Original issue's description:
> These tests stress pathops by describing the union of circle-like paths that have tiny line segments embedded and double back to create near-coincident conditions.
>
> The fixes include
> - detect when finding the active top loops between two possible answers
> - preflight chasing winding to ensure answer is consistent
> - binary search more often when quadratic intersection fails
> - add more failure paths when an intersect is missed
>
> While this fixes the chrome bug, reenabling path ops in svg should be deferred until additional fixes are landed.
>
> TBR=
> BUG=421132
>
> Committed: https://skia.googlesource.com/skia/+/6f726addf3178b01949bb389ef83cf14a1d7b6b2

TBR=caryclark@google.com
NOTREECHECKS=true
NOTRY=true
BUG=421132

Review URL: https://codereview.chromium.org/686843002

10 years agoExtend SkFontMgr_Custom to cover ttc, otf, pfb.
bungeman [Tue, 28 Oct 2014 17:46:06 +0000 (10:46 -0700)]
Extend SkFontMgr_Custom to cover ttc, otf, pfb.

Adds ability to test ScanFont with Type1 (pfb) fonts
and fonts with multiple faces (ttc).

Review URL: https://codereview.chromium.org/672723002

10 years agoremove the GPU checkbox from the hash / database
Greg Humphreys [Tue, 28 Oct 2014 17:33:54 +0000 (13:33 -0400)]
remove the GPU checkbox from the hash / database

We decided last week that GPU rendering wouldn't be associated
with the fiddle itself; options in the left bar will be split into
ones that are bound to the fiddle (so appear in the hash and database)
and ones that are run-specific (like the rendering mode).

This will leave us better placed to do things like rendering in multiple
modes at once.

BUG=skia:
R=jcgregorio@google.com

Review URL: https://codereview.chromium.org/685543002

10 years agoThese tests stress pathops by describing the union of circle-like paths that have...
caryclark [Tue, 28 Oct 2014 17:33:09 +0000 (10:33 -0700)]
These tests stress pathops by describing the union of circle-like paths that have tiny line segments embedded and double back to create near-coincident conditions.

The fixes include
- detect when finding the active top loops between two possible answers
- preflight chasing winding to ensure answer is consistent
- binary search more often when quadratic intersection fails
- add more failure paths when an intersect is missed

While this fixes the chrome bug, reenabling path ops in svg should be deferred until additional fixes are landed.

TBR=
BUG=421132

Review URL: https://codereview.chromium.org/633393002

10 years agoRemove unneeded file.
scroggo [Tue, 28 Oct 2014 17:27:50 +0000 (10:27 -0700)]
Remove unneeded file.

public_headers.gypi was never needed by Skia proper, and was almost
always out of date. Its actual user no longer needs it, so stop
updating it with every other change to public headers.

BUG=skia:2350

Review URL: https://codereview.chromium.org/683123003

10 years agoWorking patch to remove constant attributes. This may cause some gm mismatches,...
joshualitt [Tue, 28 Oct 2014 17:27:44 +0000 (10:27 -0700)]
Working patch to remove constant attributes.  This may cause some gm mismatches, I will rebaseline tonight.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/84c94c0dfd1e12e97d8a835882dda575f36e41d2

Committed: https://skia.googlesource.com/skia/+/95f5194abce19e8ed875f3495fd16c79a9b931b4

Review URL: https://codereview.chromium.org/678073005

10 years agoRevert of Patch to remove constant attributes (patchset #8 id:120002 of https://coder...
joshualitt [Tue, 28 Oct 2014 16:08:35 +0000 (09:08 -0700)]
Revert of Patch to remove constant attributes (patchset #8 id:120002 of https://codereview.chromium.org/678073005/)

Reason for revert:
I'll checkin tonight when the tree is quieter

Original issue's description:
> Working patch to remove constant attributes.  This may cause some gm mismatches, I will rebaseline tonight.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/84c94c0dfd1e12e97d8a835882dda575f36e41d2
>
> Committed: https://skia.googlesource.com/skia/+/95f5194abce19e8ed875f3495fd16c79a9b931b4

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

Review URL: https://codereview.chromium.org/683203002

10 years agoRevert of Update the Android NDK to r10c (patchset #1 id:1 of https://codereview...
hcm [Tue, 28 Oct 2014 15:25:31 +0000 (08:25 -0700)]
Revert of Update the Android NDK to r10c (patchset #1 id:1 of https://codereview.chromium.org/687603002/)

Reason for revert:
Spooky behavior on Nexus GPUs

Original issue's description:
> Update the Android NDK to r10c
>
> Committed: https://skia.googlesource.com/skia/+/a01a17b3f2f225d9499627e20f9b42ce351bd0eb

TBR=tomhudson@chromium.org,borenet@google.com,tomhudson@google.com,djsollen@google.com
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/685753003

10 years agoaa rect renderer takes a gpu on construction
joshualitt [Tue, 28 Oct 2014 15:12:18 +0000 (08:12 -0700)]
aa rect renderer takes a gpu on construction

BUG=skia:

Review URL: https://codereview.chromium.org/683483004

10 years agoOverride SkCanvas::drawImage[Rect]() in SkProxyCanvas
piotaixr [Tue, 28 Oct 2014 15:10:42 +0000 (08:10 -0700)]
Override SkCanvas::drawImage[Rect]() in SkProxyCanvas

BUG=skia:2947

Review URL: https://codereview.chromium.org/679293002

10 years agoWorking patch to remove constant attributes. This may cause some gm mismatches,...
joshualitt [Tue, 28 Oct 2014 14:54:54 +0000 (07:54 -0700)]
Working patch to remove constant attributes.  This may cause some gm mismatches, I will rebaseline tonight.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/84c94c0dfd1e12e97d8a835882dda575f36e41d2

Review URL: https://codereview.chromium.org/678073005

10 years agoAlter layer hoisting to only hoist layers for one canvas at a time
robertphillips [Tue, 28 Oct 2014 14:21:44 +0000 (07:21 -0700)]
Alter layer hoisting to only hoist layers for one canvas at a time

This CL alters layer hoisting to defer creation of the free floating layers until they are actually needed (rather than creating _all_ the hoisted layers at the start).

It also fixes a pre vs. post Concat bug with how matrices were being accumulated.

BUG=skia:2315

Review URL: https://codereview.chromium.org/657383004

10 years agoAdd SK_API to global scope operators of SkMatrix
junov [Tue, 28 Oct 2014 14:14:58 +0000 (07:14 -0700)]
Add SK_API to global scope operators of SkMatrix

Review URL: https://codereview.chromium.org/678153002

10 years agonot a failure if we didn't load the test file
reed [Tue, 28 Oct 2014 14:12:37 +0000 (07:12 -0700)]
not a failure if we didn't load the test file

BUG=skia:

Review URL: https://codereview.chromium.org/687453005

10 years agoAdd Geometry processor to srcAlphaWillBeOne check.
egdaniel [Tue, 28 Oct 2014 14:06:36 +0000 (07:06 -0700)]
Add Geometry processor to srcAlphaWillBeOne check.

BUG=skia:

Review URL: https://codereview.chromium.org/681333002

10 years agoUpdate the Android NDK to r10c
djsollen [Tue, 28 Oct 2014 13:53:08 +0000 (06:53 -0700)]
Update the Android NDK to r10c

Review URL: https://codereview.chromium.org/687603002

10 years agoReplace color_wheel.* with beautiful versions I made with Skia.
halcanary [Tue, 28 Oct 2014 13:33:07 +0000 (06:33 -0700)]
Replace color_wheel.* with beautiful versions I made with Skia.

NOTRY=true

Review URL: https://codereview.chromium.org/683113002

10 years agoRevert of Use approximate scratch textures for image filters. (patchset #2 id:20001...
hcm [Tue, 28 Oct 2014 04:43:30 +0000 (21:43 -0700)]
Revert of Use approximate scratch textures for image filters. (patchset #2 id:20001 of https://codereview.chromium.org/679113004/)

Reason for revert:
Continued GM failures

Original issue's description:
> Use approximate scratch textures for image filters.
>
> Committed: https://skia.googlesource.com/skia/+/96c60686eef6fc514feba6136f1b475f71dbf3ec
>
> Committed: https://skia.googlesource.com/skia/+/50a140563c5960acb2479d2be7c437d146963b3f

TBR=senorblanco@google.com,senorblanco@chromium.org,bsalomon@google.com
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/681183002

10 years agoRevert of Fix bounds computation of all 0-input filters. (patchset #2 id:20001 of...
hcm [Tue, 28 Oct 2014 03:55:01 +0000 (20:55 -0700)]
Revert of Fix bounds computation of all 0-input filters. (patchset #2 id:20001 of https://codereview.chromium.org/681643003/)

Reason for revert:
try again

Original issue's description:
> Fix bounds computation of all 0-input filters.
>
> The SkRectShaderImageFilter had the same bug as previously fixed for
> SkBitmapSource and SkPictureImageFilter. Rather than copy-and-paste
> the implementation, this change makes all filters with 0 inputs return
> their source bounds, instead of returning false.
>
> BUG=427251
>
> Committed: https://skia.googlesource.com/skia/+/ba036cc82b5a543a13cafd11a19ba0e3087fca38

TBR=bsalomon@google.com,senorblanco@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=427251

Review URL: https://codereview.chromium.org/678273002

10 years agoRevert of Rebaseline imagemagnifier and dropshadowimagefilter tests. (patchset #3...
hcm [Tue, 28 Oct 2014 03:53:25 +0000 (20:53 -0700)]
Revert of Rebaseline imagemagnifier and dropshadowimagefilter tests. (patchset #3 id:40001 of https://codereview.chromium.org/685523002/)

Reason for revert:
Getting to revert for failures on image filter tests in Skia and Blink

Original issue's description:
> Rebaseline imagemagnifier and dropshadowimagefilter tests.
>
> BUG=skia:
> TBR=hcm@google.com
> NOTRY=true
>
> Committed: https://skia.googlesource.com/skia/+/0fc71ea432d3895ede198a24080f877e607cd264

TBR=senorblanco@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/678043003

10 years agoRevert of Re-add suppressions for magnifier and dropshadow GMs. (patchset #1 id:1...
hcm [Tue, 28 Oct 2014 03:48:57 +0000 (20:48 -0700)]
Revert of Re-add suppressions for magnifier and dropshadow GMs. (patchset #1 id:1 of https://codereview.chromium.org/686573002/)

Reason for revert:
Trying to get revert of patchset potentially causing failures on image filter tests

Original issue's description:
> Re-add suppressions for magnifier and dropshadow GMs.
>
> Brian's change has invalidated the new baselines, so they'll need
> to be rebaselined again once his changed has worked its way through.
>
> TBR=bsalomon
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/fa4689d56384da5bb264a480e636fac126fb93ca

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

Review URL: https://codereview.chromium.org/679413002

10 years agoWhitespace change to trigger builds
Ravi Mistry [Tue, 28 Oct 2014 01:21:16 +0000 (21:21 -0400)]
Whitespace change to trigger builds

10 years agoRevert of Patch to remove constant attributes (patchset #6 id:100001 of https://coder...
joshualitt [Tue, 28 Oct 2014 00:56:18 +0000 (17:56 -0700)]
Revert of Patch to remove constant attributes (patchset #6 id:100001 of https://codereview.chromium.org/678073005/)

Reason for revert:
Breaks the K1

Original issue's description:
> Working patch to remove constant attributes.  This may cause some gm mismatches, I will rebaseline tonight.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/84c94c0dfd1e12e97d8a835882dda575f36e41d2

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

Review URL: https://codereview.chromium.org/679363002