platform/upstream/libSkiaSharp.git
9 years agoRejigger keep-alive thread so TSAN and LSAN don't complain.
mtklein [Thu, 12 Mar 2015 14:16:56 +0000 (07:16 -0700)]
Rejigger keep-alive thread so TSAN and LSAN don't complain.

BUG=skia:

Review URL: https://codereview.chromium.org/1006443002

9 years agoAdd #include first in Sk4x_* files
rmistry [Thu, 12 Mar 2015 14:09:25 +0000 (07:09 -0700)]
Add #include first in Sk4x_* files

BUG=skia:3362

Review URL: https://codereview.chromium.org/1004693002

9 years agoDM: add keepalive for Valgrind bot.
mtklein [Thu, 12 Mar 2015 13:28:54 +0000 (06:28 -0700)]
DM: add keepalive for Valgrind bot.

BUG=skia:

Review URL: https://codereview.chromium.org/998263003

9 years agoClean up SkDynamicAnnotations.
mtklein [Thu, 12 Mar 2015 12:27:46 +0000 (05:27 -0700)]
Clean up SkDynamicAnnotations.

    Unprotected reads -> relaxed reads.
    Unprotected write -> relaxed write.

The only unprotected write we had was in SkTraceEvent, which it looks like we nabbed from Chrome at some point and changed only to silence TSAN.  Chrome's version uses AtomicWord / NoBarrier_Load / NoBarrier_Store, which boils down to the same as here, intptr_t / relaxed load / relaxed store.

This leaves one place where we're lying a bit to TSAN, in include/core/SkLazyPtr.h where we're doing a data-dependent consume load.  We're telling TSAN it's consume, but telling any other compiler to compile it as relaxed, given how they all upgrade consume to acquire.  This eliminates a barrier for us on ARM.  How do you guys deal with this?  Just use a consume memory order, take the hit, and hope compilers get smarter one day?

BUG=chromium:465721

No public API changes.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/996763002

9 years agoFix HWUI Sink for DM
Derek Sollenberger [Thu, 12 Mar 2015 11:52:22 +0000 (07:52 -0400)]
Fix HWUI Sink for DM

Review URL: https://codereview.chromium.org/996413002

9 years agoAdd terrible workaround to fix stencil clear crash on S3.
bsalomon [Wed, 11 Mar 2015 22:49:54 +0000 (15:49 -0700)]
Add terrible workaround to fix stencil clear crash on S3.

BUG=skia:3525

Review URL: https://codereview.chromium.org/1001693003

9 years agoGlyph positions maintain 32 bit integer part.
bungeman [Wed, 11 Mar 2015 21:05:29 +0000 (14:05 -0700)]
Glyph positions maintain 32 bit integer part.

A glyph position when mapped from canvas space to device space may land
outside the bounds of the current 16 bit integer part of device space.
Device space is already limited to 32 bits for the integer part,
but for a short space in drawText and drawPosText it is currently
limited to 16 bits (SkFixed). Raise this limit by moving to 48.16.
This matches the current similar fix for measureText.

BUG=chromium:375322

Review URL: https://codereview.chromium.org/977623002

9 years agoFix Windows build.
scroggo [Wed, 11 Mar 2015 20:58:17 +0000 (13:58 -0700)]
Fix Windows build.

SK_BUILD_FOR_WIN is no longer a valid way to check for building on
Windows (go figure). Build everywhere.

Remove the REPORTER_ASSERT, which was the failing part. It also isn't
necessary for the test, which is just that we are not leaking an
SkColorTable.

Also fix indentation.

TBR=bungeman@google.com,mtklein@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

BUG=skia:3457

Review URL: https://codereview.chromium.org/1002583002

9 years agoSince SkDrawFilter has serious architectural issues and may be
tomhudson [Wed, 11 Mar 2015 19:21:08 +0000 (12:21 -0700)]
Since SkDrawFilter has serious architectural issues and may be
deprecated, use a proxy SkCanvas for the same end: in every draw
call, inspect the paint, changing it to fit within the capabilities
of the Android Java (HWUI) drawing API.

Verified that this allows us to reenable all our ColorFilter tests.

R=djsollen@google.com

Review URL: https://codereview.chromium.org/997183003

9 years agodelete dead code
mtklein [Wed, 11 Mar 2015 19:17:00 +0000 (12:17 -0700)]
delete dead code

BUG=skia:

Review URL: https://codereview.chromium.org/1003493002

9 years agoFor safety include SkTypes.h in SkFontHost_mac.cpp
mtklein [Wed, 11 Mar 2015 19:01:25 +0000 (12:01 -0700)]
For safety include SkTypes.h in SkFontHost_mac.cpp

BUG=skia:3362

Review URL: https://codereview.chromium.org/996113003

9 years agoClean up poppler code.
mtklein [Wed, 11 Mar 2015 18:55:13 +0000 (11:55 -0700)]
Clean up poppler code.

We're not actually using it.

BUG=skia:3362
DOCS_PREVIEW= https://skia.org/?cl=1002493002

Review URL: https://codereview.chromium.org/1002493002

9 years agoAdjust atlas sizes to fix Mali400 precision issues
jvanverth [Wed, 11 Mar 2015 18:18:11 +0000 (11:18 -0700)]
Adjust atlas sizes to fix Mali400 precision issues

The previous change to atlas sizes introduced an issue where texture
coordinates for glyphs stored on the right side of the atlas were being
computed in an imprecise manner on Mali400 chips.

The only reasonable fix appears to be to use power-of-two texture sizes.
This widens the glyph atlas to the next power of 2 (or 2048) for A8 glyphs
only, and widens each GrPlot by 2x as well. By doing this, we can fit 3-4 large
distance field glyphs into a single GrPlot, which gives us 100-128 total large
glyphs at one time.

The 565 and 8888 glyph atlases are kept in their original size to preserve
space.

BUG=skia:3523

Review URL: https://codereview.chromium.org/994303003

9 years agoFix debugger's overdraw color filter
robertphillips [Wed, 11 Mar 2015 18:18:00 +0000 (11:18 -0700)]
Fix debugger's overdraw color filter

The colors seen in the overdraw color filter are not guaranteed to remain the colors from its own table. Looping through the table to find the old overdraw level is error prone and expensive.

Review URL: https://codereview.chromium.org/997913002

9 years agoUse proper SkFDot6 instead of int
qiankun.miao [Wed, 11 Mar 2015 18:12:59 +0000 (11:12 -0700)]
Use proper SkFDot6 instead of int

BUG=skia:

Review URL: https://codereview.chromium.org/960353005

9 years agoFix a memory leak when decoding corrupted indexed PNGs.
dml [Wed, 11 Mar 2015 18:12:54 +0000 (11:12 -0700)]
Fix a memory leak when decoding corrupted indexed PNGs.
Commit to branch refs/heads/png-leak

BUG=skia:3457

Committed: https://skia.googlesource.com/skia/+/561a1ca9559a1ea7589ab93350124284fcef3315

Review URL: https://codereview.chromium.org/951663002

9 years agoSimplify clip inside a layer -- will be applied in restore
reed [Wed, 11 Mar 2015 17:39:04 +0000 (10:39 -0700)]
Simplify clip inside a layer -- will be applied in restore

This reverts commit 44124650895833378938b4b174302fbf64403fe2.

NOTREECHECKS=True
BUG=skia:

Review URL: https://codereview.chromium.org/996233002

9 years agofix for valgrind preAbandonGpuContext
joshualitt [Wed, 11 Mar 2015 16:11:19 +0000 (09:11 -0700)]
fix for valgrind preAbandonGpuContext

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/997923002

9 years agooptimization/fix: dirty the clip-bounds when we mod the clip in savelayer
reed [Wed, 11 Mar 2015 15:47:12 +0000 (08:47 -0700)]
optimization/fix: dirty the clip-bounds when we mod the clip in savelayer

Before the fix, we could use a stale cache of the clipbounds in quickReject. Often this could return false negatives, meaning we would try to draw more than we should (it would eventually be really clipped). Occasionally this could also report false positives (if the layer were outside of the normal canvas bounds, e.g. a layer with an offset imagefilter).

BUG=skia:
NOTREECHECKS=True

Review URL: https://codereview.chromium.org/983243003

9 years agoSkip WritePixels on GalaxyS3
mtklein [Wed, 11 Mar 2015 15:43:43 +0000 (08:43 -0700)]
Skip WritePixels on GalaxyS3

TBR=bsalomon@google.com

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/994343003

9 years agoDM: show skipped and --blacklist notes only in verbose mode.
mtklein [Wed, 11 Mar 2015 14:19:15 +0000 (07:19 -0700)]
DM: show skipped and --blacklist notes only in verbose mode.

In non-verbose mode, these notes will spin away too fast to read anyway,
unless they're so long they end up leaving junk on the terminal.

NOTREECHECKS=true

BUG=skia:

Review URL: https://codereview.chromium.org/989083002

9 years agoRevert of simplify clip inside a layer -- will be applied in restore (patchset #2...
reed [Wed, 11 Mar 2015 11:02:20 +0000 (04:02 -0700)]
Revert of simplify clip inside a layer -- will be applied in restore (patchset #2 id:20001 of https://codereview.chromium.org/997763002/)

Reason for revert:
slight rebaselines needed in layouttests

http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/51755

Original issue's description:
> simplify clip inside a layer -- will be applied in restore
>
> BUG=skia:
> TBR=
> NOTREECHECKS=True
>
> speculative land to measure performance
>
> Committed: https://skia.googlesource.com/skia/+/c275493d4247a2c5ae845d340c4297258f4684bc

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

Review URL: https://codereview.chromium.org/997103002

9 years agosimplify clip inside a layer -- will be applied in restore
reed [Wed, 11 Mar 2015 02:57:10 +0000 (19:57 -0700)]
simplify clip inside a layer -- will be applied in restore

BUG=skia:
TBR=
NOTREECHECKS=True

speculative land to measure performance

Review URL: https://codereview.chromium.org/997763002

9 years agoremove unused params to DeviceCM constructor
reed [Wed, 11 Mar 2015 00:30:07 +0000 (17:30 -0700)]
remove unused params to DeviceCM constructor

BUG=skia:
TBR=
NOTREECHECKS=True

Review URL: https://codereview.chromium.org/999593002

9 years agoDon't run ResourceCache unit test on NexusPlayer.
mtklein [Tue, 10 Mar 2015 18:55:18 +0000 (11:55 -0700)]
Don't run ResourceCache unit test on NexusPlayer.

TBR=bsalomon@google.com

BUG=skia:

Review URL: https://codereview.chromium.org/992203005

9 years agoAdjust distance field glyph and font atlas sizes to maximize storage.
jvanverth [Tue, 10 Mar 2015 18:53:39 +0000 (11:53 -0700)]
Adjust distance field glyph and font atlas sizes to maximize storage.

Because of high DPI devices, we need more room in the glyph
atlas for the larger glyphs. These settings will allow 4 of
the distance field glyphs to fit in one Plot (increasing
the storage from 32 large glyphs to 128), and still permit
us to handle glyphs up to a 312 point size.

BUG=chromium:458791

Committed: https://skia.googlesource.com/skia/+/d2737ad7dd8f4ea94a74034027014fd3d78923cb

Committed: https://skia.googlesource.com/skia/+/07e245ffe1ca7f781590fcff904c2922c2269781

Review URL: https://codereview.chromium.org/986343002

9 years agoRevert of Adjust distance field glyph and font atlas sizes to maximize storage. ...
jvanverth [Tue, 10 Mar 2015 15:23:07 +0000 (08:23 -0700)]
Revert of Adjust distance field glyph and font atlas sizes to maximize storage. (patchset #8 id:140001 of https://codereview.chromium.org/986343002/)

Reason for revert:
Still seeing bad text rendering in GMs with Nexus 2 and Nexus 3.

Original issue's description:
> Adjust distance field glyph and font atlas sizes to maximize storage.
>
> Because of high DPI devices, we need more room in the glyph
> atlas for the larger glyphs. These settings will allow 4 of
> the distance field glyphs to fit in one Plot (increasing
> the storage from 32 large glyphs to 128), and still permit
> us to handle glyphs up to a 312 point size.
>
> BUG=chromium:458791
>
> Committed: https://skia.googlesource.com/skia/+/d2737ad7dd8f4ea94a74034027014fd3d78923cb
>
> Committed: https://skia.googlesource.com/skia/+/07e245ffe1ca7f781590fcff904c2922c2269781

TBR=joshualitt@google.com,bsalomon@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:458791

Review URL: https://codereview.chromium.org/995863002

9 years agoFix bug in Valgrind flags config.
mtklein [Tue, 10 Mar 2015 15:03:26 +0000 (08:03 -0700)]
Fix bug in Valgrind flags config.

'GPU' is in 'Test-Ubuntu14-GCE-NoGPU-x86_64-Release-Valgrind_CPU' too.

This means we're building it in no-GPU mode, and running it in no-CPU mode.
At least it finishes quite quickly this way (~10 seconds).

BUG=skia:

Review URL: https://codereview.chromium.org/992203004

9 years agoAdjust distance field glyph and font atlas sizes to maximize storage.
jvanverth [Tue, 10 Mar 2015 13:53:28 +0000 (06:53 -0700)]
Adjust distance field glyph and font atlas sizes to maximize storage.

Because of high DPI devices, we need more room in the glyph
atlas for the larger glyphs. These settings will allow 4 of
the distance field glyphs to fit in one Plot (increasing
the storage from 32 large glyphs to 128), and still permit
us to handle glyphs up to a 312 point size.

BUG=chromium:458791

Committed: https://skia.googlesource.com/skia/+/d2737ad7dd8f4ea94a74034027014fd3d78923cb

Review URL: https://codereview.chromium.org/986343002

9 years agoUpdate SKP version
skia.buildbots [Tue, 10 Mar 2015 13:20:37 +0000 (06:20 -0700)]
Update SKP version

Automatic commit by the RecreateSKPs bot.

TBR=

Review URL: https://codereview.chromium.org/991263002

9 years agoAdd pageset for theverge
rmistry [Tue, 10 Mar 2015 12:00:08 +0000 (05:00 -0700)]
Add pageset for theverge

BUG=skia:3234

TBR=robertphillips
NOTRY=true
NOPRESUBMIT=true

Review URL: https://codereview.chromium.org/994923002

9 years agoIncrease maximum gpu resources from 1024 to 4096
kkinnunen [Tue, 10 Mar 2015 05:12:29 +0000 (22:12 -0700)]
Increase maximum gpu resources from 1024 to 4096

Pictures with many drawn paths benefit from the increase of the
count. As an example, chalkboard testcase has ~2500 paths, tiger
has ~300.

Bigger objects, like textures, tend to be limited by the cache
byte limit, not item count.

Review URL: https://codereview.chromium.org/991783003

9 years agoUse SkTMin and SkTMax for clamp/pin.
bungeman [Mon, 9 Mar 2015 20:40:15 +0000 (13:40 -0700)]
Use SkTMin and SkTMax for clamp/pin.

This changes SkScalarClampMax and SkScalarPin to use SkTMin and SkTMax.
This change allows compilers to more easily transform these operations
into fast max/min operarations as opposed to conditional branches.

Review URL: https://codereview.chromium.org/993593002

9 years agoRevert of Adjust distance field glyph and font atlas sizes to maximize storage. ...
jvanverth [Mon, 9 Mar 2015 19:17:07 +0000 (12:17 -0700)]
Revert of Adjust distance field glyph and font atlas sizes to maximize storage. (patchset #5 id:80001 of https://codereview.chromium.org/986343002/)

Reason for revert:
There's some weirdness with Tegra2 and Tegra3 GMs (Xoom and Nexus7) that may be related to this.

Original issue's description:
> Adjust distance field glyph and font atlas sizes to maximize storage.
>
> Because of high DPI devices, we need more room in the glyph
> atlas for the larger glyphs. These settings will allow 4 of
> the distance field glyphs to fit in one Plot (increasing
> the storage from 32 large glyphs to 128), and still permit
> us to handle glyphs up to a 312 point size.
>
> BUG=chromium:458791
>
> Committed: https://skia.googlesource.com/skia/+/d2737ad7dd8f4ea94a74034027014fd3d78923cb

TBR=joshualitt@google.com,bsalomon@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:458791

Review URL: https://codereview.chromium.org/986333005

9 years agoUse global GrMemoryPools protected by mutex for GrProcessor/GrBatch
bsalomon [Mon, 9 Mar 2015 19:15:53 +0000 (12:15 -0700)]
Use global GrMemoryPools protected by mutex for GrProcessor/GrBatch

BUG=chromium:464892

Review URL: https://codereview.chromium.org/991943002

9 years agoAdjust distance field glyph and font atlas sizes to maximize storage.
jvanverth [Mon, 9 Mar 2015 16:45:40 +0000 (09:45 -0700)]
Adjust distance field glyph and font atlas sizes to maximize storage.

Because of high DPI devices, we need more room in the glyph
atlas for the larger glyphs. These settings will allow 4 of
the distance field glyphs to fit in one Plot (increasing
the storage from 32 large glyphs to 128), and still permit
us to handle glyphs up to a 312 point size.

BUG=chromium:458791

Review URL: https://codereview.chromium.org/986343002

9 years agoAllow user to pass in a path for extra dependencies for gpu
egdaniel [Mon, 9 Mar 2015 16:40:42 +0000 (09:40 -0700)]
Allow user to pass in a path for extra dependencies for gpu

BUG=skia:

Review URL: https://codereview.chromium.org/986363002

9 years agoRemove SK_SUPPORT_LEGACY_ADDOVAL
fmalita [Mon, 9 Mar 2015 16:40:37 +0000 (09:40 -0700)]
Remove SK_SUPPORT_LEGACY_ADDOVAL

The flad has been retired from Chromium.

BUG=chromium:458898
R=reed@google.com

Review URL: https://codereview.chromium.org/994533002

9 years agoRun Valgrind bots with --nogpu and --nocpu as appropriate
borenet [Mon, 9 Mar 2015 13:59:16 +0000 (06:59 -0700)]
Run Valgrind bots with --nogpu and --nocpu as appropriate

BUG=skia:3506
NOPRESUBMIT=true

Review URL: https://codereview.chromium.org/986583004

9 years agoRevert of use colorfilter instead of imagefilter when possible (patchset #6 id:100001...
reed [Mon, 9 Mar 2015 01:47:13 +0000 (18:47 -0700)]
Revert of use colorfilter instead of imagefilter when possible (patchset #6 id:100001 of https://codereview.chromium.org/966233002/)

Reason for revert:
Lots of PDF failures -- need to investigate

Original issue's description:
> use colorfilter instead of imagefilter when possible
>
> big speed win w/ nanobench --match colorfilter_
>
> BUG=skia:
> TBR=
>
> Will require rebaselining, so landing temporarily to trigger layouttests.
>
> Committed: https://skia.googlesource.com/skia/+/f641bd076d29666a5d2a10d5384e4924b0a0e524

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

Review URL: https://codereview.chromium.org/991693002

9 years agouse colorfilter instead of imagefilter when possible
reed [Mon, 9 Mar 2015 01:34:37 +0000 (18:34 -0700)]
use colorfilter instead of imagefilter when possible

big speed win w/ nanobench --match colorfilter_

BUG=skia:
TBR=

Will require rebaselining, so landing temporarily to trigger layouttests.

Review URL: https://codereview.chromium.org/966233002

9 years agoUse ComposColorFilter to collaps hierarchy (when possible).
reed [Sun, 8 Mar 2015 11:42:52 +0000 (04:42 -0700)]
Use ComposColorFilter to collaps hierarchy (when possible).

Clarify asColorFilter ...
1. Rename to isColorFilterNode for DAG reduction
2. Add asAColorFilter for removing the imagefilter entirely (future use-case)

Need layouttest rebaseline suppression before this can land in chrome...
https://codereview.chromium.org/984023004/

BUG=skia:

Review URL: https://codereview.chromium.org/982933002

9 years agoRevert of For consistency, use our homebrew zlib everywhere possible. (patchset ...
mtklein [Sat, 7 Mar 2015 15:47:51 +0000 (07:47 -0800)]
Revert of For consistency, use our homebrew zlib everywhere possible. (patchset #6 id:100001 of https://codereview.chromium.org/971673005/)

Reason for revert:
Reverting in case this is the cause of the non-Windows failures.

Original issue's description:
> For consistency, use our homebrew zlib everywhere possible.
>
> This switches when we build our own zlib from "just Windows" to "everyone, but
> not Android framework of course".
>
> I tested this by building DM for my Mac and for an Android bot config.
> It took minor tweaks to the GYP to get ARM builds working.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/5a8f2257b0b0f954fb74f65e7ea3ada772ed9240

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

Review URL: https://codereview.chromium.org/989873002

9 years agoRevert of Fix a memory leak when decoding corrupted indexed PNGs. (patchset #2 id...
reed [Sat, 7 Mar 2015 14:05:51 +0000 (06:05 -0800)]
Revert of Fix a memory leak when decoding corrupted indexed PNGs. (patchset #2 id:20001 of https://codereview.chromium.org/951663002/)

Reason for revert:
speculative revert due to many failures in tree

Original issue's description:
> Fix a memory leak when decoding corrupted indexed PNGs.
> Commit to branch refs/heads/png-leak
>
> BUG=skia:3457
>
> Committed: https://skia.googlesource.com/skia/+/561a1ca9559a1ea7589ab93350124284fcef3315

TBR=scroggo@google.com,dml@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:3457

Review URL: https://codereview.chromium.org/980203005

9 years agoFix a memory leak when decoding corrupted indexed PNGs.
dml [Sat, 7 Mar 2015 00:25:28 +0000 (16:25 -0800)]
Fix a memory leak when decoding corrupted indexed PNGs.
Commit to branch refs/heads/png-leak

BUG=skia:3457

Review URL: https://codereview.chromium.org/951663002

9 years agoUpdate SkPicture cull rects with RTree information
schenney [Sat, 7 Mar 2015 00:20:28 +0000 (16:20 -0800)]
Update SkPicture cull rects with RTree information

When computed, the RTree for an SkPicture will have a root
bounds that reflects the best bounding information available,
rather than the best estimate at the time the picture recorder
is created. Given that creators frequently don't know ahead of
time what will be drawn, the RTree bound is often tighter.

Perf testing on Chrome indicates a small raster performance
advantage. For upcoming painting changes in Chrome the
performance advantage is much larger.

BUG=

Committed: https://skia.googlesource.com/skia/+/2dd3b6647dc726f36fd8774b3d0d2e83b493aeac

Review URL: https://codereview.chromium.org/971803002

9 years agoMake Sk4f(float) constructor explicit.
mtklein [Sat, 7 Mar 2015 00:20:22 +0000 (16:20 -0800)]
Make Sk4f(float) constructor explicit.

BUG=skia:

Review URL: https://codereview.chromium.org/985003003

9 years agoFix bit rot in SkGlyphCache validate code.
herb [Sat, 7 Mar 2015 00:15:23 +0000 (16:15 -0800)]
Fix bit rot in SkGlyphCache validate code.

TESTED
CFLAGS=-DSK_DEBUG_GLYPH_CACHE CPPFLAGS=-DSK_DEBUG_GLYPH_CACHE ./gyp_skia

ninja -C out/Debug dm
out/Debug/dm

Review URL: https://codereview.chromium.org/989683002

9 years agoFor consistency, use our homebrew zlib everywhere possible.
mtklein [Sat, 7 Mar 2015 00:15:16 +0000 (16:15 -0800)]
For consistency, use our homebrew zlib everywhere possible.

This switches when we build our own zlib from "just Windows" to "everyone, but
not Android framework of course".

I tested this by building DM for my Mac and for an Android bot config.
It took minor tweaks to the GYP to get ARM builds working.

BUG=skia:

Review URL: https://codereview.chromium.org/971673005

9 years agoClip scissor to RT bounds
bsalomon [Fri, 6 Mar 2015 16:42:34 +0000 (08:42 -0800)]
Clip scissor to RT bounds

R=joshualitt@google.com

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/985783002

9 years agoRemove rect clip type from grclip.
bsalomon [Fri, 6 Mar 2015 16:21:38 +0000 (08:21 -0800)]
Remove rect clip type from grclip.

R=joshualitt@google.com
BUG=skia:
NOTREECHECKS=true

Review URL: https://codereview.chromium.org/989443003

9 years agoRead pixels in BGRA non-premul mode in few tests
kkinnunen [Fri, 6 Mar 2015 15:16:00 +0000 (07:16 -0800)]
Read pixels in BGRA non-premul mode in few tests

The tests assert on Sk_ColorGREEN, which is in BRGA non-premul. Read the
pixels in same color format.

Currently the tests pass on all platforms because GREEN is fully opaque
and the component stays in the same place both on BGRA and
RGBA. However, hypothetically somebody could copy-paste the assertion
for further tests but use, say, RED. This creates latent error that is
only visible on some platforms like mac.

Review URL: https://codereview.chromium.org/989463002

9 years agoFix rect clips
bsalomon [Fri, 6 Mar 2015 15:13:01 +0000 (07:13 -0800)]
Fix rect clips

R=joshualitt@google.com

Review URL: https://codereview.chromium.org/987653004

9 years agoSKPMFloat: we can beat the naive loops when clamping
mtklein [Fri, 6 Mar 2015 14:15:44 +0000 (06:15 -0800)]
SKPMFloat: we can beat the naive loops when clamping

Clamping 4 at a time is now about 15% faster than 1 at a time with SSSE3.
Clamping 4 at a time is now about 20% faster with SSE2,
and this applies to non-clamping too (we still just clamp there).

In all cases, 4 at a time is never worse than 1 at a time,
and not clamping is never slower than clamping.

Here's all the bench results, with the numbers for portable code as a fun point
of reference:

SSSE3:
maxrss  loops   min median  mean    max stddev  samples     config  bench
  10M   2291    4.66ns  4.66ns  4.66ns  4.68ns  0%  ▆█▁▁▁▇▁▇▁▃  nonrendering    SkPMFloat_get_1x
  10M   2040    5.29ns  5.3ns   5.3ns   5.32ns  0%  ▃▆▃▃▁▁▆▃▃█  nonrendering    SkPMFloat_clamp_1x
  10M   7175    4.62ns  4.62ns  4.62ns  4.63ns  0%  ▁▄▃████▃▄▇  nonrendering    SkPMFloat_get_4x
  10M   5801    4.89ns  4.89ns  4.89ns  4.91ns  0%  █▂▄▃▁▃▄█▁▁  nonrendering    SkPMFloat_clamp_4x

SSE2:
maxrss  loops   min median  mean    max stddev  samples     config  bench
  10M   1601    6.02ns  6.05ns  6.04ns  6.08ns  0%  █▅▄▅▄▂▁▂▂▂  nonrendering    SkPMFloat_get_1x
  10M   2918    6.05ns  6.06ns  6.05ns  6.06ns  0%  ▂▇▁▇▇▁▇█▇▂  nonrendering    SkPMFloat_clamp_1x
  10M   3569    5.43ns  5.45ns  5.44ns  5.45ns  0%  ▄█▂██▇▁▁▇▇  nonrendering    SkPMFloat_get_4x
  10M   4168    5.43ns  5.43ns  5.43ns  5.44ns  0%  █▄▇▁▇▄▁▁▁▁  nonrendering    SkPMFloat_clamp_4x

Portable:
maxrss  loops   min median  mean    max stddev  samples     config  bench
  10M   500     27.8ns  28.1ns  28ns    28.2ns  0%  ▃█▆▃▇▃▆▁▇▂  nonrendering    SkPMFloat_get_1x
  10M   770     40.1ns  40.2ns  40.2ns  40.3ns  0%  ▅▁▃▂▆▄█▂▅▂  nonrendering    SkPMFloat_clamp_1x
  10M   1269    28.4ns  28.8ns  29.1ns  32.7ns  4%  ▂▂▂█▂▁▁▂▁▁  nonrendering    SkPMFloat_get_4x
  10M   1439    40.2ns  40.4ns  40.4ns  40.5ns  0%  ▆▆▆█▁▆▅█▅▆  nonrendering    SkPMFloat_clamp_4x

SkPMFloat_neon.h is still one big TODO as far as 4-at-a-time APIs go.

BUG=skia:

Review URL: https://codereview.chromium.org/982123002

9 years agoWe had zero coverage for SkDrawFilter. This new GM draws two rectangles.
tomhudson [Fri, 6 Mar 2015 14:15:20 +0000 (06:15 -0800)]
We had zero coverage for SkDrawFilter. This new GM draws two rectangles.
If SkDrawFilter is working, they will not match (have different colors
and one will be blurred). If SkDrawFilter is broken, they will match.

R=scroggo@google.com

Review URL: https://codereview.chromium.org/984883003

9 years agoAdding linear interpolation to rgb->yuv conversion
sugoi [Fri, 6 Mar 2015 13:16:52 +0000 (05:16 -0800)]
Adding linear interpolation to rgb->yuv conversion

When the UV planes are smaller than the Y plane, doing the upscaling in nearest mode was creating artefacts, so I changed it to use linear interpolation to fix the issue.

BUG=460380

Committed: https://skia.googlesource.com/skia/+/cd9d42c5167a50f1bf20e969343556d61354171b

Review URL: https://codereview.chromium.org/973563002

9 years agoRevert "Don't require AA in order to use shader-based clips"
bsalomon [Fri, 6 Mar 2015 01:43:09 +0000 (17:43 -0800)]
Revert "Don't require AA in order to use shader-based clips"

This reverts commit 259656779334689ab1624ec4e2e234b35fe4024b.

TBR=robertphillips@google.com

Review URL: https://codereview.chromium.org/981723009

9 years agoRevert of check for inputs before reporting asColorFilter (patchset #7 id:110001...
reed [Thu, 5 Mar 2015 22:47:29 +0000 (14:47 -0800)]
Revert of check for inputs before reporting asColorFilter (patchset #7 id:110001 of https://codereview.chromium.org/967833003/)

Reason for revert:
Need to suppress these for rebaselining, so reverting for now.

Regressions: Unexpected image-only failures (5)
  css3/filters/effect-brightness-clamping-hw.html [ ImageOnlyFailure ]
  css3/filters/effect-combined-hw.html [ ImageOnlyFailure ]
  virtual/slimmingpaint/css3/filters/effect-brightness-clamping-hw.html [ ImageOnlyFailure ]
  virtual/slimmingpaint/css3/filters/effect-combined-hw.html [ ImageOnlyFailure ]

Original issue's description:
> Use ComposeColorFilter in factory to collapse consecutive filters (when possible).
> Change asColorFilter to reflect its reliance on the new factory behavior.
>
> patch from issue 967143002 at patchset 80001 (http://crrev.com/967143002#ps80001)
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/dac843bf046c2cd79fd955cb177aee241d7a4b0c

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

Review URL: https://codereview.chromium.org/978923005

9 years agohandle null vertex or index buffers in batch
joshualitt [Thu, 5 Mar 2015 22:33:41 +0000 (14:33 -0800)]
handle null vertex or index buffers in batch

BUG=skia:

Review URL: https://codereview.chromium.org/979343002

9 years agoIncrease default tile sizes in nanobench
bsalomon [Thu, 5 Mar 2015 21:42:27 +0000 (13:42 -0800)]
Increase default tile sizes in nanobench

R=mtklein@google.com

Review URL: https://codereview.chromium.org/982863003

9 years agoWhen running under Valgrind, don't draw .webps into .pdfs.
mtklein [Thu, 5 Mar 2015 21:38:17 +0000 (13:38 -0800)]
When running under Valgrind, don't draw .webps into .pdfs.

This blacklist entry bans any test with 'pdf' config, any source type, whose
name has '.webp' in it.  In practice, that's 'image' or 'subset' source type
decoding some WEBP file.

BUG=skia:3505

Review URL: https://codereview.chromium.org/982163002

9 years agoDisable --config pdf temporarily on Valgrind bot.
mtklein [Thu, 5 Mar 2015 21:06:30 +0000 (13:06 -0800)]
Disable --config pdf temporarily on Valgrind bot.

BUG=skia:3505

Review URL: https://codereview.chromium.org/987533002

9 years agoRemove SkPorterDuff now that Android has been updated.
djsollen [Thu, 5 Mar 2015 20:42:29 +0000 (12:42 -0800)]
Remove SkPorterDuff now that Android has been updated.

BUG=skia:3178

Review URL: https://codereview.chromium.org/983743003

9 years agoAdd flag to force gpu trace markers to be used and printed out for debugging.
egdaniel [Thu, 5 Mar 2015 20:19:17 +0000 (12:19 -0800)]
Add flag to force gpu trace markers to be used and printed out for debugging.

BUG=skia:

Review URL: https://codereview.chromium.org/976413002

9 years agoRevert of Update SkPicture cull rects with RTree information (patchset #6 id:140001...
joshualitt [Thu, 5 Mar 2015 20:18:38 +0000 (12:18 -0800)]
Revert of Update SkPicture cull rects with RTree information (patchset #6 id:140001 of https://codereview.chromium.org/971803002/)

Reason for revert:
Might be breaking deps roll

Original issue's description:
> Update SkPicture cull rects with RTree information
>
> When computed, the RTree for an SkPicture will have a root
> bounds that reflects the best bounding information available,
> rather than the best estimate at the time the picture recorder
> is created. Given that creators frequently don't know ahead of
> time what will be drawn, the RTree bound is often tighter.
>
> Perf testing on Chrome indicates a small raster performance
> advantage. For upcoming painting changes in Chrome the
> performance advantage is much larger.
>
> BUG=
>
> Committed: https://skia.googlesource.com/skia/+/2dd3b6647dc726f36fd8774b3d0d2e83b493aeac

TBR=mtklein@google.com,schenney@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/977413003

9 years agoFix uninitialized memory bug in the SkGlyphCache.
herb [Thu, 5 Mar 2015 19:51:11 +0000 (11:51 -0800)]
Fix uninitialized memory bug in the SkGlyphCache.

The core of the problem is that the system is asked to lookup the metrics for a character with id == 0. This causes a hit in the fCharToGlyphHash matching the sentinel glyph. This happens because fCharToGlpyhHash is initialized with all zeros, therefore, the fID is zero matching the char with id == 0. The fAdvanceX field of the sentinel glyph is in fact not initialized.

The bigger question is now did a zero character get passed to getUnicharMetrics?

The breaking code is basically as follows:
wchar_t glyph = L'S';
paint.measureText(&glyph, 2);

This get mischaracterized as a utf8 string instead of a utf16(?) string. Because of the little endian ordering, this is the character string 'L' '\0'. Since the size of the original string is two bytes (but a single character) the '\0' is treated as its own character and past to getUnicharMetrics.

TEST:
On windows failed using DrMemory. With this change does not fail.

BUG=463204

Review URL: https://codereview.chromium.org/977063002

9 years agoDM: Use the new non-fatal errors for invalid color conversions.
scroggo [Thu, 5 Mar 2015 19:46:40 +0000 (11:46 -0800)]
DM: Use the new non-fatal errors for invalid color conversions.

Also allow incomplete to be considered successful.

Do not attempt to draw transparent images on 565.

BUG=skia:3475

Review URL: https://codereview.chromium.org/978413002

9 years ago4-at-a-time SkPMColor -> SkPMFloat API.
mtklein [Thu, 5 Mar 2015 19:31:59 +0000 (11:31 -0800)]
4-at-a-time SkPMColor -> SkPMFloat API.

Please see if this looks usable.  It may even give a perf boost if you use it, even without custom implementations for each instruction set.

I've been trying this morning to beat this naive loop implementation, but so far no luck with either _SSE2.h or _SSSE3.h.  It's possible this is an artifact of the microbenchmark, because we're not doing anything between the conversions.  I'd like to see how this fits into real code, what assembly's generated, what the hot spots are, etc.

I've updated the tests to test these new APIs, and splintered off a pair of new benchmarks that use the new APIs.  This required some minor rejiggering in the benches.

BUG=skia:

Review URL: https://codereview.chromium.org/978213003

9 years agoWhitespace change to test new CQ trybot
rmistry [Thu, 5 Mar 2015 18:26:04 +0000 (10:26 -0800)]
Whitespace change to test new CQ trybot

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/987433002

9 years agoUse ComposeColorFilter in factory to collapse consecutive filters (when possible).
reed [Thu, 5 Mar 2015 18:22:20 +0000 (10:22 -0800)]
Use ComposeColorFilter in factory to collapse consecutive filters (when possible).
Change asColorFilter to reflect its reliance on the new factory behavior.

patch from issue 967143002 at patchset 80001 (http://crrev.com/967143002#ps80001)

BUG=skia:

Review URL: https://codereview.chromium.org/967833003

9 years agoUpdate gpu trace marker system for MultiDrawBuffer world
robertphillips [Thu, 5 Mar 2015 17:17:17 +0000 (09:17 -0800)]
Update gpu trace marker system for MultiDrawBuffer world

The main thrust of this CL is to remove the currCmdMarker variable from GrTargetCommands::flush. In a MultiDrawBuffer world the Cmds need not be execute in the order of their issuance.

Review URL: https://codereview.chromium.org/978363002

9 years agoAdd custom dm upload script to be used by the android framework
djsollen [Thu, 5 Mar 2015 17:13:58 +0000 (09:13 -0800)]
Add custom dm upload script to be used by the android framework

Review URL: https://codereview.chromium.org/979153002

9 years agoFix GPU trace marker bug
robertphillips [Thu, 5 Mar 2015 16:55:28 +0000 (08:55 -0800)]
Fix GPU trace marker bug

Without this patch the GPU trace markers can be potentially unbalanced when batching occurs.

Review URL: https://codereview.chromium.org/981973002

9 years agoDM: support non-fatal errors
mtklein [Thu, 5 Mar 2015 16:40:28 +0000 (08:40 -0800)]
DM: support non-fatal errors

Tasks that produce a non-fatal error will bail out before writing their output to
disk and hash to dm.json, but not count as failures.

This also makes true failures bail out before writing their results.  If the DM
program failed, we probably don't want to triage that image result.

We use this new feature first to skip image subset decoding when we detect it's
not supported.  Here's a snippet of an example run, where in this case only
.webp are subset decodable:

...
(  15MB    12) 172µs 8888 subset color_wheel.jpg (skipped: Subset decoding not supported.)
(  15MB    11) 9.05ms 8888 subset randPixels.webp
(  16MB    10) 863µs 8888 subset baby_tux.png (skipped: Subset decoding not supported.)
...

Only outputs corresponding to the .webp show up, both on disk and in the .json.

BUG=skia:

Review URL: https://codereview.chromium.org/980333002

9 years agoViaAndroidSDK for DM
tomhudson [Thu, 5 Mar 2015 16:01:07 +0000 (08:01 -0800)]
ViaAndroidSDK for DM

Make a Via for DM which transforms a set of draws to be more like what
we'd see through the Android Framework's HWUI API. Only built inside
Android's framework because we depend on HWUI classes for half of
those transformations.

Tested with --config androidsdk-8888 and --config androidsdk-hwui.

R=djsollen@google.com,mtklein@google.com,reed@google.com
BUG=skia:

Review URL: https://codereview.chromium.org/974913002

9 years agoUpdate SkPicture cull rects with RTree information
schenney [Thu, 5 Mar 2015 15:43:10 +0000 (07:43 -0800)]
Update SkPicture cull rects with RTree information

When computed, the RTree for an SkPicture will have a root
bounds that reflects the best bounding information available,
rather than the best estimate at the time the picture recorder
is created. Given that creators frequently don't know ahead of
time what will be drawn, the RTree bound is often tighter.

Perf testing on Chrome indicates a small raster performance
advantage. For upcoming painting changes in Chrome the
performance advantage is much larger.

BUG=

Review URL: https://codereview.chromium.org/971803002

9 years agoadd impl limit for number of leaf-nodes in composecolorfilter
reed [Thu, 5 Mar 2015 15:21:02 +0000 (07:21 -0800)]
add impl limit for number of leaf-nodes in composecolorfilter

BUG=skia:

Review URL: https://codereview.chromium.org/972153010

9 years agoBatchify the tessellated path renderer.
senorblanco [Thu, 5 Mar 2015 15:13:42 +0000 (07:13 -0800)]
Batchify the tessellated path renderer.

Review URL: https://codereview.chromium.org/976103002

9 years agoAdd some METADATA files to control docs layout.
jcgregorio [Thu, 5 Mar 2015 15:11:13 +0000 (07:11 -0800)]
Add some METADATA files to control docs layout.

BUG=skia:3484
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=976303002

Review URL: https://codereview.chromium.org/976303002

9 years agoRevert of Add image as a draw type that can be filtered (patchset #4 id:60001 of...
kkinnunen [Thu, 5 Mar 2015 14:33:54 +0000 (06:33 -0800)]
Revert of Add image as a draw type that can be filtered (patchset #4 id:60001 of https://codereview.chromium.org/960783003/)

Reason for revert:
Fails on mac for some reason.
Also is a bit wrong, but this should not be reason for the failure..

Original issue's description:
> Add image as a draw type that can be filtered
>
> Add image as a draw type that can be filtered.
>
> This is needed when SkImage is added as an object to be drawn so that
> the draw is forwarded to SkBaseDevice. This would be used in making
> filters use SkImages.
>
> BUG=skia:3388
>
> Committed: https://skia.googlesource.com/skia/+/fa77eb1e51b9317ff993d1be504ada173b561e5f

TBR=reed@google.com,bsalomon@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:3388

Review URL: https://codereview.chromium.org/980273002

9 years agoAdd image as a draw type that can be filtered
kkinnunen [Thu, 5 Mar 2015 08:39:45 +0000 (00:39 -0800)]
Add image as a draw type that can be filtered

Add image as a draw type that can be filtered.

This is needed when SkImage is added as an object to be drawn so that
the draw is forwarded to SkBaseDevice. This would be used in making
filters use SkImages.

BUG=skia:3388

Review URL: https://codereview.chromium.org/960783003

9 years ago4x library for NEON
msarett [Wed, 4 Mar 2015 23:55:54 +0000 (15:55 -0800)]
4x library for NEON

CQ_EXTRA_TRYBOTS=client.skia.android:Test-Android-Nexus5-Adreno330-Arm7-Debug-Trybot

BUG=skia:

Review URL: https://codereview.chromium.org/975303003

9 years agoDM: no PDF for Daisy either
mtklein [Wed, 4 Mar 2015 21:58:05 +0000 (13:58 -0800)]
DM: no PDF for Daisy either

BUG=skia:

Review URL: https://codereview.chromium.org/974413002

9 years agoudpate dox for colorfilter fragment processor requirements
reed [Wed, 4 Mar 2015 21:30:50 +0000 (13:30 -0800)]
udpate dox for colorfilter fragment processor requirements

BUG=skia:
NOTRY=True

Review URL: https://codereview.chromium.org/961933003

9 years ago(starting with mtklein's 966503002 mod to 963593003)
caryclark [Wed, 4 Mar 2015 20:32:22 +0000 (12:32 -0800)]
(starting with mtklein's 966503002 mod to 963593003)
fix casts

Committed: https://skia.googlesource.com/skia/+/9e73a84deb5e6c201081b1d30175863898543e8d

Review URL: https://codereview.chromium.org/979453004

9 years agoImprove tracking of bound FBOs in GrGLGpu.
bsalomon [Wed, 4 Mar 2015 19:57:37 +0000 (11:57 -0800)]
Improve tracking of bound FBOs in GrGLGpu.

Committed: https://skia.googlesource.com/skia/+/d2ad8eb5801e2c8c0fa544a6a776bb46eedde2a0

Committed: https://skia.googlesource.com/skia/+/b2af2d8b83ca4774c3b3bb1e49bc72605faa9589

Committed: https://skia.googlesource.com/skia/+/0b70b86a7e9fda52ee7ebc1b9897eeaa09b9abef

Review URL: https://codereview.chromium.org/949263002

9 years agoDM: run PDF except on Android bots.
mtklein [Wed, 4 Mar 2015 19:47:11 +0000 (11:47 -0800)]
DM: run PDF except on Android bots.

While we work out RAM issues on Android, run everywhere else.

BUG=skia:

Review URL: https://codereview.chromium.org/971463004

9 years agoRevert of fix casts (patchset #1 id:1 of https://codereview.chromium.org/979453004/)
mtklein [Wed, 4 Mar 2015 19:46:27 +0000 (11:46 -0800)]
Revert of fix casts (patchset #1 id:1 of https://codereview.chromium.org/979453004/)

Reason for revert:
64-bit windows build warnings-as-errors

Original issue's description:
> (starting with mtklein's 966503002 mod to 963593003)
> fix casts
>
> Committed: https://skia.googlesource.com/skia/+/9e73a84deb5e6c201081b1d30175863898543e8d

TBR=caryclark@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/977583007

9 years agoUpdate SkPMFloat API a bit.
mtklein [Wed, 4 Mar 2015 19:25:27 +0000 (11:25 -0800)]
Update SkPMFloat API a bit.

Instead of set(SkPMColor), add a constructor SkPMFloat(SkPMColor).
Replace setA(), setR(), etc. with a 4 float constructor.

And, promise to stick to SkPMColor order.

BUG=skia:

Review URL: https://codereview.chromium.org/977773002

9 years ago(starting with mtklein's 966503002 mod to 963593003)
caryclark [Wed, 4 Mar 2015 19:22:05 +0000 (11:22 -0800)]
(starting with mtklein's 966503002 mod to 963593003)
fix casts

Review URL: https://codereview.chromium.org/979453004

9 years agoPass blob bounds to the looper, if available
fmalita [Wed, 4 Mar 2015 19:20:12 +0000 (11:20 -0800)]
Pass blob bounds to the looper, if available

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

Review URL: https://codereview.chromium.org/979943002

9 years agosimplify logic in SkPerliNoiseShader for GPU when constant color
bsalomon [Wed, 4 Mar 2015 19:03:52 +0000 (11:03 -0800)]
simplify logic in SkPerliNoiseShader for GPU when constant color

BUG=skia:3488

Review URL: https://codereview.chromium.org/974183002

9 years agoskia: Add tracing for skia's budget
hendrikw [Wed, 4 Mar 2015 18:33:49 +0000 (10:33 -0800)]
skia: Add tracing for skia's budget

We've run into several places where GPU rasterization slows down a lot,
and in some cases, it's due to use reaching skia's budget.  This shows a
graph of skia's used and free budgeted memory.

Review URL: https://codereview.chromium.org/977143002

9 years agoAdd SSSE3 implementation for SkPMFloat, with faster get() and set().
mtklein [Wed, 4 Mar 2015 18:18:10 +0000 (10:18 -0800)]
Add SSSE3 implementation for SkPMFloat, with faster get() and set().

With SSSE3, we can use the Swiss Army Knife byte shuffler pshufb,
a.k.a. _mm_shuffle_epi8(), to jump directly between 32 and 128 bits.

In microbench isolation, this looks like an additional 10-15% speedup:

SkPMFloat_get:   2.35ns -> 1.98ns
SkPMFloat_clamp: 2.35ns -> 2.18ns

Before this CL, get() and clamp() were identical code.  The _get benchmark improves because both set() and get() become faster; the _clamp benchmark shows the improvement from set() getting faster with clamp() staying the same.

BUG=skia:

Review URL: https://codereview.chromium.org/976493002

9 years agoRevert of Adding linear interpolation to rgb->yuv conversion (patchset #1 id:1 of...
joshualitt [Wed, 4 Mar 2015 17:56:09 +0000 (09:56 -0800)]
Revert of Adding linear interpolation to rgb->yuv conversion (patchset #1 id:1 of https://codereview.chromium.org/973563002/)

Reason for revert:
Speculative revert to see if this unblocks the deps roll

Original issue's description:
> Adding linear interpolation to rgb->yuv conversion
>
> When the UV planes are smaller than the Y plane, doing the upscaling in nearest mode was creating artefacts, so I changed it to use linear interpolation to fix the issue.
>
> BUG=460380
>
> Committed: https://skia.googlesource.com/skia/+/cd9d42c5167a50f1bf20e969343556d61354171b

TBR=bsalomon@google.com,scroggo@google.com,reed@google.com,sugoi@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=460380

Review URL: https://codereview.chromium.org/977133002

9 years agodisable chatting debugf
reed [Wed, 4 Mar 2015 15:51:37 +0000 (07:51 -0800)]
disable chatting debugf

BUG=skia:
TBR=bsalomon@google.com

Review URL: https://codereview.chromium.org/974883005

9 years agoFix uninitialized variable in GrGpu
hendrikw [Wed, 4 Mar 2015 14:22:18 +0000 (06:22 -0800)]
Fix uninitialized variable in GrGpu

Stumbled on this while running a trace.

Review URL: https://codereview.chromium.org/975993002

9 years agoText blob run paints should be filtered.
fmalita [Wed, 4 Mar 2015 03:08:17 +0000 (19:08 -0800)]
Text blob run paints should be filtered.

We're constructing blob run paints after the canvas draw filter has been
applied.

Instead, we need do defer text blob draw filters until we have access
to the full run paints.

BUG=skia:3494
R=reed@google.com,mtklein@google.com

Review URL: https://codereview.chromium.org/973973003

9 years agoAdding linear interpolation to rgb->yuv conversion
sugoi [Tue, 3 Mar 2015 21:28:30 +0000 (13:28 -0800)]
Adding linear interpolation to rgb->yuv conversion

When the UV planes are smaller than the Y plane, doing the upscaling in nearest mode was creating artefacts, so I changed it to use linear interpolation to fix the issue.

BUG=460380

Review URL: https://codereview.chromium.org/973563002

9 years agoSplit GrTargetCommands into its own files
robertphillips [Tue, 3 Mar 2015 20:40:49 +0000 (12:40 -0800)]
Split GrTargetCommands into its own files

Review URL: https://codereview.chromium.org/979493002