mtklein [Mon, 11 Jul 2016 16:59:21 +0000 (09:59 -0700)]
Clean up hyper-local SkCpu feature test experiment.
This removes the code paths where we make SkCpu::Supports() calls
from within a tight loop. It keeps code paths using SkCpu::Supports()
to choose entire routines from src/opts/.
We can't rely on these hyper-local checks to be hoisted up reliably enough.
It worked pretty well with the first couple platforms we tried (e.g. Clang
on Linux/Mac) but we can't gaurantee it works everywhere.
Further, I'm not able to actually do anything fancy with those tests
outside of x86... I've not found a way to get, say, NEON+F16 conversion
code embedded into ordinary NEON code outside writing then entire function
in external assembly.
This whole idea becomes less important now that we've got a way to chain
separate function calls together efficiently. We can now, e.g., use an
AVX+F16C method to load some pixels, then chain that into an ordinary AVX
method to color filter them.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2138073002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Review-Url: https://codereview.chromium.org/
2138073002
fmenozzi [Mon, 11 Jul 2016 16:56:31 +0000 (09:56 -0700)]
Add HardStopGradientBench.cpp
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2131643002
Committed: https://skia.googlesource.com/skia/+/
d0d652e0e0cb4bcecb7ccac9f4f4831fcaccdc45
Review-Url: https://codereview.chromium.org/
2131643002
fmenozzi [Mon, 11 Jul 2016 16:30:25 +0000 (09:30 -0700)]
Revert of Add HardStopGradientBench.cpp (patchset #3 id:40001 of https://codereview.chromium.org/
2131643002/ )
Reason for revert:
Segfaulting on HardStopGradientBench; didn't allocate enough stack memory for color/position arrays.
Original issue's description:
> Add HardStopGradientBench.cpp
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2131643002
>
> Committed: https://skia.googlesource.com/skia/+/
d0d652e0e0cb4bcecb7ccac9f4f4831fcaccdc45
TBR=bsalomon@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review-Url: https://codereview.chromium.org/
2140603003
bsalomon [Mon, 11 Jul 2016 15:31:27 +0000 (08:31 -0700)]
Delay initialization of free list in SkTLList until first use of list
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2138043002
Review-Url: https://codereview.chromium.org/
2138043002
fmenozzi [Mon, 11 Jul 2016 15:31:23 +0000 (08:31 -0700)]
Add HardStopGradientBench.cpp
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2131643002
Review-Url: https://codereview.chromium.org/
2131643002
rmistry [Mon, 11 Jul 2016 15:27:49 +0000 (08:27 -0700)]
Do not use the deprecated CQ_EXTRA_TRYBOTS
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2135063003
Review-Url: https://codereview.chromium.org/
2135063003
mtklein [Mon, 11 Jul 2016 15:27:43 +0000 (08:27 -0700)]
ndk: r11c -> r12b
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2140603002
Review-Url: https://codereview.chromium.org/
2140603002
msarett [Mon, 11 Jul 2016 12:49:17 +0000 (05:49 -0700)]
Make all color xforms 'fast' (step 1)
This refactors opt code to handle arbitrary src and dst
gammas that are specified by tables.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2130013002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Review-Url: https://codereview.chromium.org/
2130013002
dvonbeck [Mon, 11 Jul 2016 01:38:33 +0000 (18:38 -0700)]
Changed SkLightingShader API to take in a shader as color source
This CL's base is the CL for the same modification, but in the implementation side: https://codereview.chromium.org/
2062703003/
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2064153002
Review-Url: https://codereview.chromium.org/
2064153002
robertphillips [Sun, 10 Jul 2016 18:49:39 +0000 (11:49 -0700)]
Fix misplaced guard
This fixes an error in https://codereview.chromium.org/
2125333002 (Add choke point for modifying non-AA rect draws (e.g., applying clipping)).
In GrDrawContext::drawRect, when we need to draw a filled rect, we always want to call drawFilledRect when useDrawInsteadOfClear is false. With the buggy code I believe we were falling back to drawing a path.
TBR=bsalomon@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2135883002
Review-Url: https://codereview.chromium.org/
2135883002
reed [Sun, 10 Jul 2016 18:45:34 +0000 (11:45 -0700)]
Revert of change mapRectScaleTranslate to pass args/ret by value (patchset #1 id:1 of https://codereview.chromium.org/
2137853002/ )
Reason for revert:
Triggered a compiler bug on Android?
FAILED: /usr/bin/ccache /b/work/skia/platform_tools/android/bin/../toolchains/arm-r11c-14/bin/arm-linux-androideabi-g++ -MMD -MF obj/src/core/core.SkMatrix.o.d -DSK_INTERNAL -DSK_GAMMA_APPLY_TO_A8 -DQT_NO_KEYWORDS -DSK_ALLOW_STATIC_GLOBAL_INITIALIZERS=0 -DSK_SUPPORT_GPU=1 -DSK_FORCE_DISTANCE_FIELD_TEXT=0 -DSK_HAS_GIF_LIBRARY -DSK_HAS_JPEG_LIBRARY -DSK_HAS_PNG_LIBRARY -DSK_HAS_WEBP_LIBRARY -DSK_TEST_QCMS -DSK_IS_BOT -DSK_CODEC_DECODES_RAW -DSK_ARM_HAS_NEON -DSK_BUILD_FOR_ANDROID -DSK_GAMMA_EXPONENT=1.4 -DSK_GAMMA_CONTRAST=0.0 -DSKIA_DLL -DSKIA_IMPLEMENTATION=1 -DSK_SUPPORT_LEGACY_CLIPTOLAYERFLAG -DNDEBUG -I../../../include/c -I../../../include/config -I../../../include/core -I../../../include/pathops -I../../../include/ports -I../../../include/private -I../../../include/utils -I../../../include/images -I../../../src/core -I../../../src/sfnt -I../../../src/image -I../../../src/opts -I../../../src/utils -I../../../include/gpu -I../../../src/gpu -I../../../platform_tools/android/third_party/cpufeatures -fPIC -g -fno-exceptions -fstrict-aliasing -Wall -Wextra -Winit-self -Wpointer-arith -Wsign-compare -Wvla -Wno-unused-parameter -Werror -march=armv7-a -mthumb -mfpu=neon -mfloat-abi=softfp -fuse-ld=gold -O2 -std=c++11 -fno-rtti -fno-threadsafe-statics -Wnon-virtual-dtor -c ../../../src/core/SkMatrix.cpp -o obj/src/core/core.SkMatrix.o
../../../src/core/SkMatrix.cpp: In member function 'SkRect SkMatrix::mapRectScaleTranslate(SkRect) const':
../../../src/core/SkMatrix.cpp:1120:1: error: unrecognizable insn:
}
^
(insn 32 31 33 2 (set (reg:V4SF 115 [ D.83077 ])
(unspec:V4SF [
(mem/c:V4SF (reg/f:SI 104 virtual-incoming-args) [0 MEM[(const __builtin_neon_sf[4] *)&src]+0 S16 A32])
] UNSPEC_VLD1)) /b/work/skia/platform_tools/android/toolchains/arm-r11c-14/lib/gcc/arm-linux-androideabi/4.9/include/arm_neon.h:8693 -1
(nil))
../../../src/core/SkMatrix.cpp:1120:1: internal compiler error: in extract_insn, at recog.c:2202
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://source.android.com/source/report-bugs.html> for instructions.
Original issue's description:
> change mapRectScaleTranslate to pass args/ret by value
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2137853002
>
> TBR=mtklein
>
> Committed: https://skia.googlesource.com/skia/+/
14dce6ed5934d7a6e1fac79f8e76e12f5670aae2
TBR=msarett@google.com,mtklein@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review-Url: https://codereview.chromium.org/
2139523002
reed [Sun, 10 Jul 2016 18:35:03 +0000 (11:35 -0700)]
change mapRectScaleTranslate to pass args/ret by value
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2137853002
TBR=mtklein
Review-Url: https://codereview.chromium.org/
2137853002
update-skps [Sun, 10 Jul 2016 07:41:28 +0000 (00:41 -0700)]
Update SKP version
Automatic commit by the RecreateSKPs bot.
TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2137823002
Review-Url: https://codereview.chromium.org/
2137823002
pdr [Sun, 10 Jul 2016 00:08:24 +0000 (17:08 -0700)]
Update markdown docs to specify --local for the local devserver
If the devserver is run locally following the instructions on
https://skia.org/dev/tools/markdown, you'll get an error about influxdb.
It looks like an influxdb username/password is required but using
--local works around this issue.
-------------------8<-------------------
$ ./bin/docserver --preview
E0703 19:33:31.510379 30448 common.go:37] Initializing logserver for log level ERROR.
F0703 19:33:31.516750 30448 common.go:63] metadata.Get() failed to make HTTP request for influxdb_name: Get http://metadata/computeMetadata/v1/project/attributes/influxdb_name: dial tcp: lookup metadata: no such host
goroutine 1 [running]:
github.com/skia-dev/glog.stacks(0x772300, 0x0, 0x0, 0x0)
/src/github.com/skia-dev/glog/glog.go:769 +0xb8
github.com/skia-dev/glog.(*loggingT).output(0x756000, 0xc800000003, 0xc820136000, 0x73613e, 0x9, 0x3f, 0x0)
/src/github.com/skia-dev/glog/glog.go:720 +0x259
github.com/skia-dev/glog.(*loggingT).printDepth(0x756000, 0x3, 0x1, 0xc8200dde28, 0x1, 0x1)
/src/github.com/skia-dev/glog/glog.go:649 +0x1b8
github.com/skia-dev/glog.(*loggingT).print(0x756000, 0x3, 0xc8200dde28, 0x1, 0x1)
/src/github.com/skia-dev/glog/glog.go:640 +0x50
github.com/skia-dev/glog.Fatal(0xc8200dde28, 0x1, 0x1)
/src/github.com/skia-dev/glog/glog.go:1128 +0x49
go.skia.org/infra/go/common.StartMetrics2(0x5103e0, 0x9, 0xc820104400, 0xc820104430, 0xc820104450, 0xc820104470, 0xc8200dde00)
/src/go.skia.org/infra/go/common/common.go:63 +0x192
go.skia.org/infra/go/common.InitWithMetrics2(0x5103e0, 0x9, 0xc820104400, 0xc820104430, 0xc820104450, 0xc820104470, 0xc820104498)
/src/go.skia.org/infra/go/common/common.go:48 +0x65
main.main()
/src/skia.googlesource.com/buildbot.git/doc/go/docserver/main.go:183 +0x95
-------------------8<-------------------
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2118113002
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=
2118113002
Review-Url: https://codereview.chromium.org/
2118113002
dvonbeck [Sun, 10 Jul 2016 00:08:15 +0000 (17:08 -0700)]
Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader
Will not run until after landing https://codereview.chromium.org/
2106893003/
This CL's base is the CL for the addition of NormalSource to the API: https://codereview.chromium.org/
2063793002/
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2062703003
Committed: https://skia.googlesource.com/skia/+/
3f58cd0eb8bf4499c4f179aa6111d7e005809df8
Review-Url: https://codereview.chromium.org/
2062703003
robertphillips [Fri, 8 Jul 2016 21:43:03 +0000 (14:43 -0700)]
Add choke point for modifying non-AA rect draws (e.g., applying clipping)
NOTREECHECKS=true
NOPRESUBMIT=true
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2125333002
Committed: https://skia.googlesource.com/skia/+/
e5e3a7a5203f256083d015c923cc9c662877aba9
Review-Url: https://codereview.chromium.org/
2125333002
robertphillips [Fri, 8 Jul 2016 20:30:04 +0000 (13:30 -0700)]
Revert of Add choke point for modifying non-AA rect draws (e.g., applying clipping) (patchset #6 id:100001 of https://codereview.chromium.org/
2125333002/ )
Reason for revert:
problems
Original issue's description:
> Add choke point for modifying non-AA rect draws (e.g., applying clipping)
>
> NOTREECHECKS=true
> NOPRESUBMIT=true
>
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2125333002
>
> Committed: https://skia.googlesource.com/skia/+/
e5e3a7a5203f256083d015c923cc9c662877aba9
TBR=bsalomon@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review-Url: https://codereview.chromium.org/
2136643002
bungeman [Fri, 8 Jul 2016 19:25:37 +0000 (12:25 -0700)]
Corrected test for gcc/libc++ workaround.
NOTREECHECKS=true
NOTRY=true
The presubmit bot is all hung up.
Review-Url: https://codereview.chromium.org/
2136613002
robertphillips [Fri, 8 Jul 2016 19:00:07 +0000 (12:00 -0700)]
Add choke point for modifying non-AA rect draws (e.g., applying clipping)
NOTREECHECKS=true
NOPRESUBMIT=true
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2125333002
Review-Url: https://codereview.chromium.org/
2125333002
bsalomon [Fri, 8 Jul 2016 18:31:22 +0000 (11:31 -0700)]
Use clipped bounds for reordering decisions
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2137543002
NOTREECHECKS=true
NOPRESUBMIT=true
Review-Url: https://codereview.chromium.org/
2137543002
bsalomon [Fri, 8 Jul 2016 16:14:07 +0000 (09:14 -0700)]
Revert change of bounds computation in MSAA path renderer.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2133773002
Review-Url: https://codereview.chromium.org/
2133773002
dvonbeck [Fri, 8 Jul 2016 15:59:47 +0000 (08:59 -0700)]
Fixed error when normal.z = 1 in GPUs with binary16 floats
Check failed in GPUs with low-precision floats since 0.9999 was too small a difference from 1 to be detected by a 16-bit float comparison. Changed it to 0.999 which fixed the issue and results in indistinguishable correct behavior.
BUG=skia:5499
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2124183004
Review-Url: https://codereview.chromium.org/
2124183004
reed [Fri, 8 Jul 2016 15:43:27 +0000 (08:43 -0700)]
deferred canvas
tries to eliminate unneeded clipRects and concats (and their associated save/restores).
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2120333002
Review-Url: https://codereview.chromium.org/
2120333002
bsalomon [Fri, 8 Jul 2016 14:36:42 +0000 (07:36 -0700)]
Don't scissor draws overlapping the render target's edges
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2135673002
Review-Url: https://codereview.chromium.org/
2135673002
bsalomon [Fri, 8 Jul 2016 13:40:56 +0000 (06:40 -0700)]
Consolidate handling of infinitely thin primitives and aa bloat handing WRT batch bounds.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2127673002
Review-Url: https://codereview.chromium.org/
2127673002
mtklein [Fri, 8 Jul 2016 13:33:16 +0000 (06:33 -0700)]
Move sRGB <-> linear conversion components to their own files.
This makes them a little easier to use outside SkColorXform code.
I've added some notes about how best to use them and their eccentricities, and added a test.
Ultimately any software sRGB <-> linear conversion should funnel somehow through here.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2128893002
CQ_EXTRA_TRYBOTS=client.skia.android:Test-Android-GCC-Nexus5-CPU-NEON-Arm7-Release-Trybot;client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Committed: https://skia.googlesource.com/skia/+/
45e58c8807179638980aae8503573b950b844e4c
Review-Url: https://codereview.chromium.org/
2128893002
rmistry [Fri, 8 Jul 2016 13:23:35 +0000 (06:23 -0700)]
Revert of Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader (patchset #11 id:200001 of https://codereview.chromium.org/
2062703003/ )
Reason for revert:
Causing the Test-Ubuntu-GCC-ShuttleA-GPU-GTX550Ti-x86_64-Release-Valgrind builder to fail.
Started failing at this build:
https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-ShuttleA-GPU-GTX550Ti-x86_64-Release-Valgrind/builds/1214
Snippet from build log:
==3016== Invalid read of size 4
==3016== at 0xCD1901: GrFragmentProcessor::registerChildProcessor(sk_sp<GrFragmentProcessor>) (SkTArray.h:163)
==3016== by 0xA58B47: _Z10sk_make_spI11NormalMapFPI5sk_spI19GrFragmentProcessorERK8SkMatrixEES1_IT_EDpOT0_ (SkNormalSource.cpp:84)
==3016== by 0xA582E8: NormalMapSourceImpl::asFragmentProcessor(GrContext*, SkMatrix const&, SkMatrix const*, SkFilterQuality, SkSourceGammaTreatment) const (SkNormalSource.cpp:187)
==3016== by 0xA0ADB5: SkLightingShaderImpl::asFragmentProcessor(GrContext*, SkMatrix const&, SkMatrix const*, SkFilterQuality, SkSourceGammaTreatment) const (SkLightingShader.cpp:268)
==3016== by 0xDA17E6: skpaint_to_grpaint_impl(GrContext*, SkPaint const&, SkMatrix const&, sk_sp<GrFragmentProcessor>*, SkXfermode::Mode*, bool, bool, GrPaint*) (SkGr.cpp:552)
==3016== by 0xDA4532: SkPaintToGrPaint(GrContext*, SkPaint const&, SkMatrix const&, bool, GrPaint*) (SkGr.cpp:668)
==3016== by 0xD99B51: SkGpuDevice::drawRect(SkDraw const&, SkRect const&, SkPaint const&) (SkGpuDevice.cpp:534)
==3016== by 0x9DDB1C: SkCanvas::onDrawRect(SkRect const&, SkPaint const&) (SkCanvas.cpp:2148)
==3016== by 0x9DBF40: SkCanvas::drawRect(SkRect const&, SkPaint const&) (SkCanvas.cpp:1919)
==3016== by 0x8D2A55: skiagm::LightingShaderGM::onDraw(SkCanvas*) (lightingshader.cpp:110)
==3016== by 0x625659: skiagm::GM::drawContent(SkCanvas*) (gm.cpp:32)
==3016== by 0x6256AD: skiagm::GM::draw(SkCanvas*) (gm.cpp:24)
==3016== by 0x61B747: DM::GMSrc::draw(SkCanvas*) const (DMSrcSink.cpp:61)
==3016== by 0x61F37B: DM::GPUSink::draw(DM::Src const&, SkBitmap*, SkWStream*, SkString*) const (DMSrcSink.cpp:1115)
==3016== by 0x619261: dm_main() (DM.cpp:1016)
==3016== by 0x619841: main (DM.cpp:1409)
==3016== Address 0xf4 is not stack'd, malloc'd or (recently) free'd
==3016==
Caught signal 11 [Segmentation fault], was running:
gpu gm lightingshader
Original issue's description:
> Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader
>
> Will not run until after landing https://codereview.chromium.org/
2106893003/
>
> This CL's base is the CL for the addition of NormalSource to the API: https://codereview.chromium.org/
2063793002/
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2062703003
>
> Committed: https://skia.googlesource.com/skia/+/
3f58cd0eb8bf4499c4f179aa6111d7e005809df8
TBR=egdaniel@google.com,robertphillips@google.com,dvonbeck@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:
NOTRY=true
Review-Url: https://codereview.chromium.org/
2137513002
brianosman [Fri, 8 Jul 2016 13:15:45 +0000 (06:15 -0700)]
Use sRGB pixel config for ARGB masks (color emoji)
Fixes over-bright color emoji and GM:colortype in sRGB mode
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129913004
Committed: https://skia.googlesource.com/skia/+/
920af9fac3d9e275a93910d9c626ee7bc23ed17a
Review-Url: https://codereview.chromium.org/
2129913004
bsalomon [Fri, 8 Jul 2016 11:42:54 +0000 (04:42 -0700)]
Reland of Better encapsulate oval/rrect batchs. (patchset #1 id:1 of https://codereview.chromium.org/
2121313004/ )
Reason for revert:
Workaround for compiler/stdlib bug landed: https://codereview.chromium.org/
2135453002
Original issue's description:
> Revert of Better encapsulate oval/rrect batchs. (patchset #5 id:80001 of https://codereview.chromium.org/
2104423003/ )
>
> Reason for revert:
> This CL is suspected to break compilation on arm64.
> https://build.chromium.org/p/chromium.android/builders/Android%20arm64%20Builder%20%28dbg%29/builds/5851/steps/compile/logs/stdio
>
> Original issue's description:
> > Better encapsulate oval/rrect batchs.
> > GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2104423003
> >
> > Committed: https://skia.googlesource.com/skia/+/
5fd209e8ee477c703bc5c11b008f247d515fc0fc
> > Committed: https://skia.googlesource.com/skia/+/
7f06c6947a3bef78dc57b9252779567c33604c90
> > Committed: https://skia.googlesource.com/skia/+/
11bf8b2eae7d1780cb969146422a2ab3b933047a
>
> TBR=robertphillips@google.com,bsalomon@google.com
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://skia.googlesource.com/skia/+/
c79fc33d3e83664ffbd3651df3131c3e9e0b635c
TBR=robertphillips@google.com,hush@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review-Url: https://codereview.chromium.org/
2134693002
reed [Fri, 8 Jul 2016 10:31:09 +0000 (03:31 -0700)]
update path->rsxform to correctly interp dy
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2131043002
TBR=
Review-Url: https://codereview.chromium.org/
2131043002
bsalomon [Fri, 8 Jul 2016 10:28:42 +0000 (03:28 -0700)]
Promote the GCC/libc++/<memory> 'typedef float float32_t' workaround to SkTypes.h
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2135453002
Review-Url: https://codereview.chromium.org/
2135453002
bsalomon [Fri, 8 Jul 2016 10:28:34 +0000 (03:28 -0700)]
Make all parameters reqiured to GrReducedClip::ReduceClipStack
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2130903002
Review-Url: https://codereview.chromium.org/
2130903002
mtklein [Fri, 8 Jul 2016 01:34:06 +0000 (18:34 -0700)]
Revert of Move sRGB <-> linear conversion components to their own files. (patchset #5 id:80001 of https://codereview.chromium.org/
2128893002/ )
Reason for revert:
Monotonicity assert is failing on ARM. (Different rsqrt() and invert() precision?) Will investigate a bit tomorrow... might reland with the test TODO.
Original issue's description:
> Move sRGB <-> linear conversion components to their own files.
>
> This makes them a little easier to use outside SkColorXform code.
>
> I've added some notes about how best to use them and their eccentricities, and added a test.
>
> Ultimately any software sRGB <-> linear conversion should funnel somehow through here.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2128893002
> CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
>
> Committed: https://skia.googlesource.com/skia/+/
45e58c8807179638980aae8503573b950b844e4c
TBR=reed@google.com,msarett@google.com,mtklein@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review-Url: https://codereview.chromium.org/
2131793002
reed [Thu, 7 Jul 2016 23:44:27 +0000 (16:44 -0700)]
make some canvas utils virtual (for proxy based subclasses)
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2130973002
Review-Url: https://codereview.chromium.org/
2130973002
reed [Thu, 7 Jul 2016 23:32:51 +0000 (16:32 -0700)]
fix typos and capitalization on drawTextRSXform CL
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2123923006
Review-Url: https://codereview.chromium.org/
2123923006
pdr [Thu, 7 Jul 2016 21:48:52 +0000 (14:48 -0700)]
Update markdown to use absolute image paths, fixing broken images
Markdown can be served with or without a trailing slash but when a
trailing slash is present, relative image links were broken.
The following example works as expected:
url: https://skia.org/dev/design/pdftheory
image url: PdfLogicalDocumentStructure.png
resolves to: https://skia.org/dev/design/PdfLogicalDocumentStructure.png
The following example was broken before this patch:
url: https://skia.org/dev/design/pdftheory/
image url: PdfLogicalDocumentStructure.png
resolves to: https://skia.org/dev/design/pdftheory/PdfLogicalDocumentStructure.png
The PdfLogicalDocumentStructure.png file lives in /dev/design/, not
/dev/design/pdftheory/. This patch updates all images in our markdown
files to be served with the full path so a trailing slash no longer
breaks images.
I hit this bug when promoting the new Skia debugger:
https://groups.google.com/a/chromium.org/d/msg/paint-dev/QJ8xpu8dNyA/F4KX2O4ZBwAJ
BUG=skia:5493
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2128013003
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=
2128013003
Review-Url: https://codereview.chromium.org/
2128013003
mtklein [Thu, 7 Jul 2016 21:47:36 +0000 (14:47 -0700)]
Move sRGB <-> linear conversion components to their own files.
This makes them a little easier to use outside SkColorXform code.
I've added some notes about how best to use them and their eccentricities, and added a test.
Ultimately any software sRGB <-> linear conversion should funnel somehow through here.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2128893002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Review-Url: https://codereview.chromium.org/
2128893002
reed [Thu, 7 Jul 2016 21:10:14 +0000 (14:10 -0700)]
add bounds to textblob
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2124113003
TBR=
Review-Url: https://codereview.chromium.org/
2124113003
hush [Thu, 7 Jul 2016 20:58:38 +0000 (13:58 -0700)]
Revert of Better encapsulate oval/rrect batchs. (patchset #5 id:80001 of https://codereview.chromium.org/
2104423003/ )
Reason for revert:
This CL is suspected to break compilation on arm64.
https://build.chromium.org/p/chromium.android/builders/Android%20arm64%20Builder%20%28dbg%29/builds/5851/steps/compile/logs/stdio
Original issue's description:
> Better encapsulate oval/rrect batchs.
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2104423003
>
> Committed: https://skia.googlesource.com/skia/+/
5fd209e8ee477c703bc5c11b008f247d515fc0fc
> Committed: https://skia.googlesource.com/skia/+/
7f06c6947a3bef78dc57b9252779567c33604c90
> Committed: https://skia.googlesource.com/skia/+/
11bf8b2eae7d1780cb969146422a2ab3b933047a
TBR=robertphillips@google.com,bsalomon@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review-Url: https://codereview.chromium.org/
2121313004
brianosman [Thu, 7 Jul 2016 20:18:51 +0000 (13:18 -0700)]
Revert of Use sRGB pixel config for ARGB masks (color emoji) (patchset #1 id:1 of https://codereview.chromium.org/
2129913004/ )
Reason for revert:
Color text rendering failing on ANGLE
Original issue's description:
> Use sRGB pixel config for ARGB masks (color emoji)
>
> Fixes over-bright color emoji and GM:colortype in sRGB mode
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129913004
>
> Committed: https://skia.googlesource.com/skia/+/
920af9fac3d9e275a93910d9c626ee7bc23ed17a
TBR=bungeman@google.com,jvanverth@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review-Url: https://codereview.chromium.org/
2128583007
reed [Thu, 7 Jul 2016 19:47:17 +0000 (12:47 -0700)]
drawTextRSXform
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2130643004
Review-Url: https://codereview.chromium.org/
2130643004
halcanary [Thu, 7 Jul 2016 19:31:55 +0000 (12:31 -0700)]
SkPDF: Fix encoding of unichr outside of basic plane
In ToUnicode table, write unicode codepoints as one or two UTF16BE
values, rather than a single hex, as the standard requires.
Factor out uint16 -> big-endian hex code.
SkUtils is now a namespace.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2120533002
Review-Url: https://codereview.chromium.org/
2120533002
halcanary [Thu, 7 Jul 2016 19:28:30 +0000 (12:28 -0700)]
SkPDF: fix PDFA-only metadata error
Review-Url: https://codereview.chromium.org/
2119883002
brianosman [Thu, 7 Jul 2016 19:22:39 +0000 (12:22 -0700)]
Use sRGB pixel config for ARGB masks (color emoji)
Fixes over-bright color emoji and GM:colortype in sRGB mode
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129913004
Review-Url: https://codereview.chromium.org/
2129913004
egdaniel [Thu, 7 Jul 2016 18:58:35 +0000 (11:58 -0700)]
Check granularity in render pass bounds
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2126303002
Review-Url: https://codereview.chromium.org/
2126303002
jvanverth [Thu, 7 Jul 2016 18:18:46 +0000 (11:18 -0700)]
Add offsets to GrVkBuffer.
BUG=skia:5031
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2127183002
Review-Url: https://codereview.chromium.org/
2127183002
csmartdalton [Thu, 7 Jul 2016 15:49:11 +0000 (08:49 -0700)]
Begin instanced rendering for simple shapes
Adds a module that performs instanced rendering and starts using it
for a select subset of draws on Mac GL platforms. The instance
processor can currently handle rects, ovals, round rects, and double
round rects. It can generalize shapes as round rects in order to
improve batching. The instance processor also employs new drawing
algorithms, irrespective of instanced rendering, that improve GPU-side
performance (e.g. sample mask, different triangle layouts, etc.).
This change only scratches the surface of instanced rendering. The
majority of draws still only have one instance. Future work may
include:
* Passing coord transforms through the texel buffer.
* Sending FP uniforms through instanced vertex attribs.
* Using instanced rendering for more draws (stencil writes,
drawAtlas, etc.).
* Adding more shapes to the instance processor’s repertoire.
* Batching draws that have mismatched scissors (analyzing draw
bounds, inserting clip planes, etc.).
* Bindless textures.
* Uber shaders.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2066993003
Committed: https://skia.googlesource.com/skia/+/
42eafa4bc00354b132ad114d22ed6b95d8849891
Review-Url: https://codereview.chromium.org/
2066993003
jvanverth [Thu, 7 Jul 2016 15:21:48 +0000 (08:21 -0700)]
Only check resource tracking on program shutdown, not context shutdown
BUG=skia:5496
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2128273002
Review-Url: https://codereview.chromium.org/
2128273002
egdaniel [Thu, 7 Jul 2016 15:12:33 +0000 (08:12 -0700)]
Fix vulkan compile
TBR=jvanverth@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2122263003
Review-Url: https://codereview.chromium.org/
2122263003
mtklein [Thu, 7 Jul 2016 15:12:09 +0000 (08:12 -0700)]
centralize VECTORCALL as SK_VECTORCALL
Gonna start using this, might as well define it once centrally.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2126253002
No public API changes.
TBR=reed@google.com
Review-Url: https://codereview.chromium.org/
2126253002
egdaniel [Thu, 7 Jul 2016 15:04:08 +0000 (08:04 -0700)]
Correctly handle mat2x2 in Vulkan
BUG=skia:5497
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129953002
Review-Url: https://codereview.chromium.org/
2129953002
bsalomon [Thu, 7 Jul 2016 14:55:24 +0000 (07:55 -0700)]
Better encapsulate oval/rrect batchs.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2104423003
Committed: https://skia.googlesource.com/skia/+/
5fd209e8ee477c703bc5c11b008f247d515fc0fc
Committed: https://skia.googlesource.com/skia/+/
7f06c6947a3bef78dc57b9252779567c33604c90
Review-Url: https://codereview.chromium.org/
2104423003
robertphillips [Thu, 7 Jul 2016 14:33:13 +0000 (07:33 -0700)]
Remove fColor from PathRenderer DrawPathArgs struct
This was getting in the way of other refactorings.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2126083002
Review-Url: https://codereview.chromium.org/
2126083002
jvanverth [Thu, 7 Jul 2016 14:16:42 +0000 (07:16 -0700)]
Pull out freelist allocation from GrVkSubHeap
BUG=skia:5031
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2128673002
Review-Url: https://codereview.chromium.org/
2128673002
ethannicholas [Thu, 7 Jul 2016 13:30:48 +0000 (06:30 -0700)]
fix for SkSL producing invalid branch after an OpKill
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129913002
Review-Url: https://codereview.chromium.org/
2129913002
dvonbeck [Thu, 7 Jul 2016 03:03:46 +0000 (20:03 -0700)]
Abstracted diffuse color map out of SkLightingShaderImpl into a generic SkShader
Will not run until after landing https://codereview.chromium.org/
2106893003/
This CL's base is the CL for the addition of NormalSource to the API: https://codereview.chromium.org/
2063793002/
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2062703003
Review-Url: https://codereview.chromium.org/
2062703003
bsalomon [Thu, 7 Jul 2016 01:56:04 +0000 (18:56 -0700)]
Hide NonAAStrokeRectBatch Geometry class and cleanup class.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2126893002
Review-Url: https://codereview.chromium.org/
2126893002
egdaniel [Wed, 6 Jul 2016 21:38:34 +0000 (14:38 -0700)]
Fix bounds computation for non-aa hairlines when snapping to pixel centers
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2126913002
Review-Url: https://codereview.chromium.org/
2126913002
jvanverth [Wed, 6 Jul 2016 21:26:53 +0000 (14:26 -0700)]
Deleting void pointers is bad, mmmkay?
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129483003
Review-Url: https://codereview.chromium.org/
2129483003
liyuqian [Wed, 6 Jul 2016 21:11:32 +0000 (14:11 -0700)]
More accurate render time and continuous fresh
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129613002
Review-Url: https://codereview.chromium.org/
2129613002
dvonbeck [Wed, 6 Jul 2016 20:58:36 +0000 (13:58 -0700)]
API change to allow for NormalSource selection at the user level.
This CL's base is the CL for CPU handling: https://codereview.chromium.org/
2050773002/
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2063793002
Review-Url: https://codereview.chromium.org/
2063793002
bsalomon [Wed, 6 Jul 2016 20:06:22 +0000 (13:06 -0700)]
Hide geometry class in AAStrokeRectBatch and remove unused Append function.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2119813004
Review-Url: https://codereview.chromium.org/
2119813004
jvanverth [Wed, 6 Jul 2016 19:56:11 +0000 (12:56 -0700)]
Vulkan fixes for TesselatingPathRenderer test
- Be sure to release all secondard command buffers on shutdown
- Allow mapping of static buffers
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2123323002
Review-Url: https://codereview.chromium.org/
2123323002
lsalzman [Wed, 6 Jul 2016 19:19:56 +0000 (12:19 -0700)]
check for zero scale in SkDraw::ComputeResScaleForStroking
If the transform used to compute the res scale has both 0 scale and skew,
then it will currently compute a 0 res scale. This causes getFillPath
to trigger an assertion while using SkStroke since that can't handle
a 0 res scale.
Downstream Firefox bug report: https://bugzilla.mozilla.org/show_bug.cgi?id=1284356
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2125933003
Review-Url: https://codereview.chromium.org/
2125933003
egdaniel [Wed, 6 Jul 2016 19:00:12 +0000 (12:00 -0700)]
Grow descriptor pools by 1.5
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2121243002
Review-Url: https://codereview.chromium.org/
2121243002
dvonbeck [Wed, 6 Jul 2016 19:00:06 +0000 (12:00 -0700)]
Fixed crash when RunInSeriesFP's color input is nil
SeriesFragmentProcessor would call emitChild with "" input color as opposed to nil input color when ginven a nil input color.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2106893003
Review-Url: https://codereview.chromium.org/
2106893003
bungeman [Wed, 6 Jul 2016 18:55:05 +0000 (11:55 -0700)]
Add gm for various font rendering combinations.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2080453002
Review-Url: https://codereview.chromium.org/
2080453002
bsalomon [Wed, 6 Jul 2016 18:54:59 +0000 (11:54 -0700)]
Add gm that tests shaded stroked rectangles.
Fix GPU handling of previously untested cases.
Move rect->path fallback from SkGpuDevice to GrDrawContext.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2125663003
Review-Url: https://codereview.chromium.org/
2125663003
bungeman [Wed, 6 Jul 2016 17:19:54 +0000 (10:19 -0700)]
Update rebaseline instructions.
BUG=skia:2805
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=
2127953002
Review-Url: https://codereview.chromium.org/
2127953002
csmartdalton [Wed, 6 Jul 2016 16:59:43 +0000 (09:59 -0700)]
Fix caching of sample locations
The original caching logic for sample locations wishfully assumed that
the GPU would always use the same sample pattern for render targets
that had the same number of samples. It turns out we can't rely on
that. This change improves the caching logic to handle mismatched
simple patterns with the same count, and adds a unit test that
emulates different sample patterns observed on real hardware.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2111423002
Committed: https://skia.googlesource.com/skia/+/
09d49a3bfe2d1e652a648ce1ea0962b38d10d166
Review-Url: https://codereview.chromium.org/
2111423002
robertphillips [Wed, 6 Jul 2016 16:58:57 +0000 (09:58 -0700)]
Move GrNonAAFillRectPerspectiveBatch to its own file
Get this out of the way of the upcoming changes (and add a test for it)
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2108403005
Review-Url: https://codereview.chromium.org/
2108403005
csmartdalton [Wed, 6 Jul 2016 16:47:38 +0000 (09:47 -0700)]
Check fIgnoresCoverage in GrPipeline::AreEqual
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2120153002
Review-Url: https://codereview.chromium.org/
2120153002
reed [Wed, 6 Jul 2016 16:29:16 +0000 (09:29 -0700)]
Experiment: add flag for finish-recording to return null
BUG=skia:5495
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2106843004
Review-Url: https://codereview.chromium.org/
2106843004
jvanverth [Wed, 6 Jul 2016 16:24:57 +0000 (09:24 -0700)]
Add resource tracking output and command buffer recycling
BUG=skia:5042
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2115993002
Review-Url: https://codereview.chromium.org/
2115993002
djsollen [Wed, 6 Jul 2016 16:17:35 +0000 (09:17 -0700)]
Compile Vulkan when building for the Android framework.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2129523002
Review-Url: https://codereview.chromium.org/
2129523002
egdaniel [Wed, 6 Jul 2016 15:51:23 +0000 (08:51 -0700)]
Fix vulkan image sampling
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2118733005
Review-Url: https://codereview.chromium.org/
2118733005
liyuqian [Wed, 6 Jul 2016 14:52:08 +0000 (07:52 -0700)]
Try fix the mac build bot after removing Android SampleApp
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2117793002
Review-Url: https://codereview.chromium.org/
2117793002
bsalomon [Wed, 6 Jul 2016 14:03:26 +0000 (07:03 -0700)]
Increase batching for AA fill rects.
This allows batching of rects provided without a local matrix when local coords are required and when the view matrix changes.
It also allows batching of rects with a local matrix with rects without a local matrix.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2116823002
Committed: https://skia.googlesource.com/skia/+/
e525ecaf63f225f1da6e9834f7a291c06ad44d23
Review-Url: https://codereview.chromium.org/
2116823002
ethannicholas [Wed, 6 Jul 2016 13:55:44 +0000 (06:55 -0700)]
fix for Clang 3.8 warning in SkSL
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2118033002
Review-Url: https://codereview.chromium.org/
2118033002
benjaminwagner [Wed, 6 Jul 2016 13:55:04 +0000 (06:55 -0700)]
Blacklist fontscalerdistortable on Win8.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=
2109583003
Review-Url: https://codereview.chromium.org/
2109583003
bsalomon [Wed, 6 Jul 2016 13:30:36 +0000 (06:30 -0700)]
Enlarge strokerect GM size to reveal hidden content
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2126723002
Review-Url: https://codereview.chromium.org/
2126723002
fmalita [Wed, 6 Jul 2016 13:22:06 +0000 (06:22 -0700)]
Document SkSurface::MakeRaster's memory initialization
So clients don't go clearing w/ SK_ColorTRANSPARENT unnecessarily.
R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2066903003
Review-Url: https://codereview.chromium.org/
2066903003
pdr [Wed, 6 Jul 2016 13:10:25 +0000 (06:10 -0700)]
Update dev docs to reference the new skia debugger (skiaserve)
This patch adds a new markdown file describing the new debugger along
with instructions on running it locally and a sweet screenshot. The
screenshot has been sized to not be clipped when viewed in production.
The old documentation for the qt-based debugger has been marked as
deprecated and all docs referencing the old debugger have been updated.
BUG=skia:5493
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2121673002
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=
2121673002
Review-Url: https://codereview.chromium.org/
2121673002
reed [Wed, 6 Jul 2016 03:10:42 +0000 (20:10 -0700)]
make setScaleTranslate public
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2122093002
TBR=
Review-Url: https://codereview.chromium.org/
2122093002
ksakamoto [Tue, 5 Jul 2016 10:54:53 +0000 (03:54 -0700)]
Revert of Begin instanced rendering for simple shapes (patchset #20 id:380001 of https://codereview.chromium.org/
2066993003/ )
Reason for revert:
This caused static initializer regressions in Chromium (crbug.com/625728).
Relevant build logs here:
Linux:
https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/21849
Mac:
https://build.chromium.org/p/chromium/builders/Mac/builds/17350
Relevant lines from the error log:
Linux:
# InstanceProcessor.cpp GrUniqueKey::GenerateDomain()
# InstanceProcessor.cpp gr_instanced::kShapeBufferDomain
FAILED linux-release-64/sizes/nacl_helper-si/initializers: actual 8, expected 7, better lower
FAILED linux-release-64/sizes/chrome-si/initializers: actual 8, expected 7, better lower
Mac:
FAILED mac-release/sizes/chrome-si/initializers: actual 2, expected 0, better lower
Original issue's description:
> Begin instanced rendering for simple shapes
>
> Adds a module that performs instanced rendering and starts using it
> for a select subset of draws on Mac GL platforms. The instance
> processor can currently handle rects, ovals, round rects, and double
> round rects. It can generalize shapes as round rects in order to
> improve batching. The instance processor also employs new drawing
> algorithms, irrespective of instanced rendering, that improve GPU-side
> performance (e.g. sample mask, different triangle layouts, etc.).
>
> This change only scratches the surface of instanced rendering. The
> majority of draws still only have one instance. Future work may
> include:
>
> * Passing coord transforms through the texel buffer.
> * Sending FP uniforms through instanced vertex attribs.
> * Using instanced rendering for more draws (stencil writes,
> drawAtlas, etc.).
> * Adding more shapes to the instance processor’s repertoire.
> * Batching draws that have mismatched scissors (analyzing draw
> bounds, inserting clip planes, etc.).
> * Bindless textures.
> * Uber shaders.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2066993003
>
> Committed: https://skia.googlesource.com/skia/+/
42eafa4bc00354b132ad114d22ed6b95d8849891
NOTREECHECKS=true
TBR=bsalomon@google.com,egdaniel@google.com,robertphillips@google.com,csmartdalton@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:
Review-Url: https://codereview.chromium.org/
2123693002
rmistry [Tue, 5 Jul 2016 00:45:06 +0000 (17:45 -0700)]
Revert of Fix caching of sample locations (patchset #3 id:40001 of https://codereview.chromium.org/
2111423002/ )
Reason for revert:
Seems to be causing Chromium roll failures:
* https://codereview.chromium.org/
2120373003
* https://codereview.chromium.org/
2117193002
* https://codereview.chromium.org/
2124653002
Original issue's description:
> Fix caching of sample locations
>
> The original caching logic for sample locations wishfully assumed that
> the GPU would always use the same sample pattern for render targets
> that had the same number of samples. It turns out we can't rely on
> that. This change improves the caching logic to handle mismatched
> simple patterns with the same count, and adds a unit test that
> emulates different sample patterns observed on real hardware.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2111423002
>
> Committed: https://skia.googlesource.com/skia/+/
09d49a3bfe2d1e652a648ce1ea0962b38d10d166
TBR=bsalomon@google.com,csmartdalton@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review-Url: https://codereview.chromium.org/
2120403002
csmartdalton [Mon, 4 Jul 2016 22:55:17 +0000 (15:55 -0700)]
Fix caching of sample locations
The original caching logic for sample locations wishfully assumed that
the GPU would always use the same sample pattern for render targets
that had the same number of samples. It turns out we can't rely on
that. This change improves the caching logic to handle mismatched
simple patterns with the same count, and adds a unit test that
emulates different sample patterns observed on real hardware.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2111423002
Review-Url: https://codereview.chromium.org/
2111423002
reed [Mon, 4 Jul 2016 21:28:31 +0000 (14:28 -0700)]
all dumpcanvas to be built in release
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2123633002
TBR=
Review-Url: https://codereview.chromium.org/
2123633002
update-skps [Mon, 4 Jul 2016 20:38:56 +0000 (13:38 -0700)]
Update SKP version
Automatic commit by the RecreateSKPs bot.
TBR=
NO_MERGE_BUILDS
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2119053002
Review-Url: https://codereview.chromium.org/
2119053002
benjaminwagner [Mon, 4 Jul 2016 18:30:01 +0000 (11:30 -0700)]
Revert of Increase batching for AA fill rects. (patchset #4 id:60001 of https://codereview.chromium.org/
2116823002/ )
Reason for revert:
I believe this is causing the Chromium DEPS roll to fail due to linux_blink_rel. Reverting for now to get the roll going again.
Original issue's description:
> Increase batching for AA fill rects.
>
> This allows batching of rects provided without a local matrix when local coords are required and when the view matrix changes.
>
> It also allows batching of rects with a local matrix with rects without a local matrix.
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2116823002
>
> Committed: https://skia.googlesource.com/skia/+/
e525ecaf63f225f1da6e9834f7a291c06ad44d23
TBR=robertphillips@google.com,bsalomon@google.com
NOTREECHECKS=true
# Not skipping CQ checks because original CL landed more than 1 days ago.
Review-Url: https://codereview.chromium.org/
2124603002
benjaminwagner [Sat, 2 Jul 2016 12:12:46 +0000 (05:12 -0700)]
Exclude SkSL in public.bzl.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2122463002
TBR=ethannicholas
NOTREECHECKS=true
Review-Url: https://codereview.chromium.org/
2122463002
ethannicholas [Sat, 2 Jul 2016 02:09:27 +0000 (19:09 -0700)]
fixed SkSL memory leak
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2119903002
NOTREECHECKS=true
Review-Url: https://codereview.chromium.org/
2119903002
ethannicholas [Sat, 2 Jul 2016 02:09:15 +0000 (19:09 -0700)]
fixed non-GPU build
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2120543003
NOTREECHECKS=true
Review-Url: https://codereview.chromium.org/
2120543003
skia.buildbots [Sat, 2 Jul 2016 00:29:18 +0000 (17:29 -0700)]
Update Win toolchain
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2116173002
NOTREECHECKS=true
NOTRY=true
TBR=rmistry@google.com
Review-Url: https://codereview.chromium.org/
2116173002
benjaminwagner [Fri, 1 Jul 2016 22:51:41 +0000 (15:51 -0700)]
Specify new Windows Toolchain isolated hash.
NOTREECHECKS=true
NOTRY=true
TBR=rmistry@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2121443002
Review-Url: https://codereview.chromium.org/
2121443002
bsalomon [Fri, 1 Jul 2016 18:02:39 +0000 (11:02 -0700)]
Increase batching for AA fill rects.
This allows batching of rects provided without a local matrix when local coords are required and when the view matrix changes.
It also allows batching of rects with a local matrix with rects without a local matrix.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2116823002
Review-Url: https://codereview.chromium.org/
2116823002
robertphillips [Fri, 1 Jul 2016 16:06:56 +0000 (09:06 -0700)]
Split apart the GP construction for the NonAAFillRect Batches
This sets up for adding yet another batch type that specializes for rectStaysRect
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2106063005
Review-Url: https://codereview.chromium.org/
2106063005
halcanary [Fri, 1 Jul 2016 15:48:12 +0000 (08:48 -0700)]
using_skia_and_harfbuzz: use default typeface
Also: add little script to generate utf-8 test text.
BUG=skia:5434
TBR=bungeman@google.com
Review-Url: https://codereview.chromium.org/
2118833002
ethannicholas [Fri, 1 Jul 2016 15:22:01 +0000 (08:22 -0700)]
initial checkin of SkSL compiler
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=
1984363002
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac-Clang-x86_64-Release-CMake-Trybot
Review-Url: https://codereview.chromium.org/
1984363002
egdaniel [Fri, 1 Jul 2016 15:06:45 +0000 (08:06 -0700)]
Use bounds of batches for render pass bounds
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=
2113303002
Review-Url: https://codereview.chromium.org/
2113303002