The Viewer slide list stores the values from repeated calls to
Slide::getName().c_str() in a large list for filtering purposes.
However, the temporary SkString was being destroyed before the list was
complete. This luckily happened to work because SkString's refcounting
behavior kept the internal buffer alive past the lifetime of the
temporary SkString object, but it's not intended to work.
Returning a const-ref to the internal SkString is more efficient, and
gives us long-lived SkStrings to work with.
Change-Id: I958148ba46332f9ec576ac60d0a9b916284c4761
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306256
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Since subsetting may require rasterizing/resolving the generator, we may
also need access to the GrDirectContext. To simplify apis, rely on
makeSubset() for that.
Related: https://skia-review.googlesource.com/c/skia/+/305970
Change-Id: I1980e3c823fb6cf54f197c350942c2f82b03e20f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306136
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This reverts commit 143646297e.
Reason for revert: Have to fix it and reland again
Original change's description:
> Revert "Attach whitespaces to the neighbor unresolved blocks"
>
> This reverts commit 131c5ad6f1.
>
> Reason for revert: Blocking the G3 roll
>
> Original change's description:
> > Attach whitespaces to the neighbor unresolved blocks
> >
> > Fix the situation when an unresolved {arabic} text is broken into
> > many small runs by resolved english spaces.
> >
> > Bug: skia:10487
> > Change-Id: I3a739501c0fb7e0fc845e68392e1d214df9302db
> > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/304000
> > Commit-Queue: Julia Lavrova <jlavrova@google.com>
> > Reviewed-by: Ben Wagner <bungeman@google.com>
>
> TBR=bungeman@google.com,jlavrova@google.com
>
> Change-Id: Iaa338dd5fb5c9962df2ee32bafbc089da0e2b8a1
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: skia:10487
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305797
> Reviewed-by: Robert Phillips <robertphillips@google.com>
> Commit-Queue: Robert Phillips <robertphillips@google.com>
TBR=bungeman@google.com,robertphillips@google.com,jlavrova@google.com
Bug: skia:10487
Change-Id: If8a85254fd536f465d80766de0406053d90c96a3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306062
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Julia Lavrova <jlavrova@google.com>
Commit-Queue: Julia Lavrova <jlavrova@google.com>
No principled reason these values have to be > -SK_ScalarNearlyZero
Speculative fix for:
Bug: chromium:1096001
Change-Id: I53cb7ec2aa46f3d0507e6010130be8925e587238
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305565
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
This fixes a Chrome canary failure in skia CL 305970 where
Chrome has the flag set, and so we can't internally call the
GrDirectContext variant.
This matches the more recent migrations we've done. Hopefully
soon Chrome's GrDirectContext CL will get its final approvals and
all of this can go away.
Bug: skia:104662
Change-Id: I09cdff54e01c22578a970e5967815a2e470752bf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306185
Commit-Queue: Adlai Holler <adlai@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Auto-Submit: Adlai Holler <adlai@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Bug: chromium:1101491
Bug: b/161896447
Found using
git grep -wiEIl \ '(he)|(she)|(his)|(hers)|(him)|(her)|(guy)|(guys)'
Change-Id: I6b91853de067fd4c2e84f7ec70275522ce6c8bfc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306186
Commit-Queue: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Auto-Submit: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
They have been replaced with equivalent constants (and given the
requisite k prefix).
Change-Id: I70907eec234e0861cc97aac1ca03086faca42f9a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306160
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Reed <reed@google.com>
The more we can test these configs the better.
Change-Id: I2d86985770047937d7f31c9659ad6984caafe11d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305569
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Adlai Holler <adlai@google.com>
Bug: skia:10540
Change-Id: I5625b3785ced1951a6a91e62e7aca3279e309989
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306183
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Auto-Submit: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
No longer needed as GrSurface and GrRenderTarget are private.
Change-Id: I2ec653b2d9daa115233bb7eaa1f2b7f880772c0a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305730
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: chromium:1101491
Bug: b/161896447
The motivation here was to remove uses of the word "master". But this
code is in experimental/, and it doesn't look to be used (based on
code searches in Chromium, Android, and Google3). The last meaningful
changes were in 2011; other changes have just been upkeep.
Change-Id: I5993402008877ec8c12c0bca4cf09c02dc6de9ec
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305965
Auto-Submit: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Bug: chromium:1101491
Bug: b/161896447
Switch to more inclusive language, like "main", or remove where
simply unnecessary.
Change-Id: I36ef6ec631eb991f54f42b98887333f07c0984c2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306060
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>
3f851efa2a..de309a4238
2020-07-28 cclao@google.com Vulkan: Make staging buffer per context
2020-07-28 nguyenmh@google.com Serializes GL context states and reformats frame_capture_utils
2020-07-27 jmadill@chromium.org Disable GPU time tracking in trace tests.
2020-07-27 cclao@google.com Vulkan: Add cache for driver uniform descriptor set
2020-07-27 syoussefi@chromium.org Vulkan: Fix render target's tracking of content defined
2020-07-27 m.maiya@samsung.com Vulkan: Enable VK_KHR_get_memory_requirements2 extension
2020-07-27 cclao@google.com Vulkan: Increase default uniform buffer size to 64K
2020-07-27 jmadill@chromium.org Tests: Add getProcAddress method to GLWindowBase.
2020-07-27 tobine@google.com Vulkan:Enable Validation best practices layer
2020-07-27 cclao@google.com Vulkan: Make mDefaultUniformStorage per ContextVk
2020-07-27 flibitijibibo@gmail.com Add GGP to Platform Support table
2020-07-27 msisov@igalia.com X11: cast native display type to XDisplay.
2020-07-27 nguyenmh@google.com Update CaptureAndReplay.md and Python script with links
2020-07-27 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from 6a8a74986c35 to 3b2cd31c7400 (9 revisions)
2020-07-27 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Vulkan-ValidationLayers from adf938680697 to 361fb5311b9b (13 revisions)
2020-07-27 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Vulkan-Headers from 83825d55c7d5 to a0f1f5149d91 (1 revision)
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-skia-autoroll
Please CC borenet@google.com on the revert to ensure that a human
is aware of the problem.
To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md
Cq-Include-Trybots: skia/skia.primary:Build-Debian10-Clang-x86_64-Release-ANGLE;skia/skia.primary:Test-Win10-Clang-AlphaR2-GPU-RadeonR9M470X-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-Golo-GPU-QuadroP400-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC5i7RYH-GPU-IntelIris6100-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC6i5SYK-GPU-IntelIris540-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC8i5BEK-GPU-IntelIris655-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUCD34010WYKH-GPU-IntelHD4400-x86_64-Debug-All-ANGLE
Bug: chromium:1085700
Tbr: borenet@google.com
Change-Id: Idfdb1345be0e0584b7923e2bc2240fb44002d96b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306198
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Previously, assigning a SkBitSet to itself would zero out its size.
Change-Id: I2179e93ebc054626dee8366beaad78aa3d62b8c6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306159
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
It's hard to imagine a scenario where it makes sense to recompile a
shader for every possible point or rect.
Change-Id: Ic4ca9a4826c4ae4b604705549170503c8ec580a6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306075
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Previously, DataEntry supported only move-construction but not move-
assignment.
Change-Id: I1dc1b2bc90a3dfe4cd74d4a6c2986b7d777f84d3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306156
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Previously, floats would be converted to int32 using a C-style cast when
added to a processor key, truncating any fractional component of the
value. This would cause two processors compiled with different floats to
generate the same key.
The float is now sk_bit_cast to a uint32 instead; this will preserve the
uniqueness of all values.
Additionally, the CPP generator will now abort if asked to add an
unsupported type to the processor key. Previously, it would do a C-style
cast to int32 for all unrecognized types and hope for the best.
Change-Id: I270ae8b3036497a9e9a77747255f763d9aad1927
Bug: skia:10486
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305572
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This isn't used, and doesn't seem like we will have a use for it going
forward. The SkString class should be used to allocate strings.
Change-Id: I024d1d7edb5d4946fa9ecfd4e10af84afc8b5df8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306065
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
These missing items were discovered when attempting a refactor.
Change-Id: Ibf905dacba14ebf5d3174f11c557741c7ab7848f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306061
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
While doing a refactor, Clang issued some warnings about having copy-
constructors without assignment operators in some TextStyle classes.
Rewrote the code to minimize custom constructor/copy/assignment methods,
favoring C++11 member initialization syntax instead.
Change-Id: I4f0361ce38d4b0ada346b0ede4fcad4e2b686d67
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305959
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Julia Lavrova <jlavrova@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Just passing the time while waiting on some bots.
Change-Id: I4ca823732bbc2a408f582508fb9d5c69689a029f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306056
Commit-Queue: Adlai Holler <adlai@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Auto-Submit: Adlai Holler <adlai@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Bug: chromium:1101491
Bug: b/161896447
We're moving to using more inclusive language, and thus removing
the word "master". Now these are just "GM" and "DM", which is how
we frequently refer to them anyway.
Change-Id: I844aeb4fc1d97a58a669b4f04821fe06f45d4c73
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306059
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Stumbled into this when I made a function that took a float3x3.
Change-Id: I35a8eef84fefe41da79e1082663937f126899160
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305728
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This should be the last batch of tests. All the remaining uses of
GrContext should be resolved when SkImage no longer requires a context.
Change-Id: I47eeb3b74c28f483c20d9bec4daecbdb6d2cb982
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305541
Reviewed-by: Adlai Holler <adlai@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
OpChains contain an internal List which does bookkeeping/validation when
moved, but our SkTArray was skipping over this validation step because
MEM_MOVE was set to true.
Change-Id: I4a4f1f0dff96076dfa3719467f4393eb7e958c8f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305958
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
This is a reland of 7f5e99465d
Original change's description:
> implement working CorrectnessActivity to generate lottie keyframe PNGs
>
>
> make correctness activity loop through all available lotties
>
>
> implement tests for lottie correctness through Android Activity
>
> Change-Id: I8b5d1dfc0641e3227f23a4d4560faa756b245ede
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305216
> Reviewed-by: Stan Iliev <stani@google.com>
> Commit-Queue: Jorge Betancourt <jmbetancourt@google.com>
Change-Id: I59f4a02acfbddd2c8bbee0adcda1d737c9566c0b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305601
Reviewed-by: Stan Iliev <stani@google.com>
Commit-Queue: Jorge Betancourt <jmbetancourt@google.com>
These missing includes were discovered when attempting a refactor.
Change-Id: I8aad5610de2337de6e55885b5868e606c1514eaf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305997
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Bug: b/161858189
This matches the format of the old traces in HWUI.
Change-Id: Ieec91d4ecea9d2f37d40f4f56a09634905c07b61
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305687
Reviewed-by: Stan Iliev <stani@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
The current build configuration (empty fontmgr) doesn't support loading
custom fonts.
For full/npm builds, use the custom fontmgr.
Change-Id: Ifd623fdbac2d5723a04add6fc455188fb3dc96e1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305816
Commit-Queue: Kevin Lubick <kjlubick@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Auto-Submit: Florin Malita <fmalita@chromium.org>
df5ed61ddb..3f851efa2a
2020-07-27 syoussefi@chromium.org Vulkan: Fix overflow in maxCombinedUniformComponents
2020-07-26 jmadill@chromium.org Capture/Replay: Update headers.
2020-07-26 syoussefi@chromium.org Vulkan: Generate mipmap in compute
2020-07-25 timvp@google.com Enable KHR GLES31 tests blocked on Issue 4146
2020-07-25 timvp@google.com Remove third_party/angle/ from generate_android_bp.py
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-skia-autoroll
Please CC borenet@google.com on the revert to ensure that a human
is aware of the problem.
To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md
Cq-Include-Trybots: skia/skia.primary:Build-Debian10-Clang-x86_64-Release-ANGLE;skia/skia.primary:Test-Win10-Clang-AlphaR2-GPU-RadeonR9M470X-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-Golo-GPU-QuadroP400-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC5i7RYH-GPU-IntelIris6100-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC6i5SYK-GPU-IntelIris540-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC8i5BEK-GPU-IntelIris655-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUCD34010WYKH-GPU-IntelHD4400-x86_64-Debug-All-ANGLE
Bug: None
Tbr: borenet@google.com
Test: Test: KHR-GLES31.core.shader_storage_buffer_object.basic-syntaxSSOTest: Test: KHR-GLES31.core.vertex_attrib_binding.advanced-bindingUpdateTest: Test: KHR-GLES31.core.vertex_attrib_binding.advanced-largeStrideAndOffsetsNewAndLegacyAPITest: Test: Manual script execution and building in AOSP
Change-Id: I0c4411b66584d8401bedf81de89148ab4cf99292
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305928
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
* transformed masks -> GrTransformedMaskSubRun
* GrMaskSubRun is SDFT -> GrSDFTSubRun
There is much duplicate code, which will be cleaned up shortly.
Change-Id: If54e388e89f7db15ddfd4b4354f0a1b4d6f5e36b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305686
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Herb Derby <herb@google.com>
We seed the builder with the imageinfo of an image, but after it is
built we then have to "attach" it to an image (or more than one).
Check that the mip chain is compatible with the parent image we are
attaching to.
Change-Id: Iead6aee54861fdd4910f38f89a0364ed80c341d8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305680
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
BM can now export inline fonts, encoded as data-uris.
Extend our DataURIResourceProviderProxy helper to also intercept and
decode inline typefaces.
Change-Id: Iaf1be9db2fd32383af78bc351b1228fe6b3b64bf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305685
Commit-Queue: Florin Malita <fmalita@chromium.org>
Commit-Queue: Ben Wagner <bungeman@google.com>
Auto-Submit: Florin Malita <fmalita@chromium.org>
Reviewed-by: Ben Wagner <bungeman@google.com>
This is a quality-of-life improvement when debugging processor
unit tests.
Change-Id: I6eba2ea17710e224852e4688c5cc491ed09461ce
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305690
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
See linked bugs, but there were mixed results on whether this was a
performance win. My guess is that for simple, easy-to-analyze clips
that satisfied a round rect clipped to itself it was a win, but when
the clip stack was large or not actually a round rect, it slowed down
the analysis.
This undoes that optimization that was turned on as part of
https://skia.googlesource.com/skia/+show/4e3cab7919258df8d6fd884298e28b20e4288060
to see if that's the cause.
Bug: skia:10462
Bug: chromium:1101513
Change-Id: I11e75a7a643337850766ed51956608b76a4a0501
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305681
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Auto-Submit: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
We will rely on the tracking of GrBuffers on the command list to manage
lifetime of d3d buffers. This also means we won't destroy the underlying
d3d buffer resource until the destructor.
Change-Id: I26084b8fffc65038891a0e48662d6a864c5dd8c1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305719
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
The code was not accounting for the GrGlyphVector data properly. Add
that back in. Make a type AllSubRuns that accounts for the max size
of all the SubRuns.
Change-Id: I9ce702d24bc4f9efbfcdb4a0f14494f53d1b701c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305568
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Herb Derby <herb@google.com>
