This limit is arbitrary, but hopefully prevents pathological (or
malicious) SVG content from consuming huge amounts of CPU/GPU time,
without impacting any legitimate uses of feMorphology. (Typical usage
has a much smaller radius).
Bug: chromium:1123035
Change-Id: I4405bc595128e9a6287eb5efa1be14621baa3a00
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315219
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This is a reland of ed63444587
Original change's description:
> Add idea of DataType to SkYUVAPixmapInfo.
>
> DataType describes the data type of YUVA channels
> independent of how they are grouped into planes.
>
> Adds mapping functions between SkColorType/channel count
> and DataType.
>
> SkYUVAPixmapInfo can be constructed from DataType and will
> choose appropriate SkColorTypes for each plane.
>
> Valid SkYUVAPixmapInfos now have the same DataType for each
> plane (could relax this in the future, esp for alpha plane).
>
> SkYUVAPixmapInfo::SupportedDataTypes specifies the supported
> combinations of SkYUVAInfo::PlanarConfig and
> kYUVAPixmapInfo::DataType supported by a GrContext (based on
> supported texture formats).
>
> SkImageGenerator/SkCodec YUVA query API now takes a
> SupportedDataTypes.
>
> Change-Id: I8791234638e6ba3396d1e7960b7bc210edc6dd57
> Bug: skia:10632
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314276
> Reviewed-by: Leon Scroggins <scroggo@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
> Commit-Queue: Brian Salomon <bsalomon@google.com>
Bug: skia:10632
Change-Id: I35b55b7477c11c822fdb3729a9f84acff1eb785d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315284
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Blur and stroke information were kept out of line with the key.
Make them part of the key, and update operator==.
Change-Id: I2e7669aae4e9c3243c078b039ae796a7e371ef8e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315282
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This is a shorthand function for verifying the kind of IRNode you have.
Change-Id: Ifc53100b6d29d9fcd02e90e0658fda971f136210
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315256
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
It has been largely superseded by the SkottieFrames varient.
The GPU version is still important because the Puppeteer
tracing gives us insights we don't yet get with SkottieFrames.
Furthermore, it is timing out for reasons unknown.
Change-Id: Ibdf2b5308e026e93502a597404b3d1214a9e6598
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315276
Reviewed-by: Ravi Mistry <rmistry@google.com>
Commit-Queue: Ravi Mistry <rmistry@google.com>
Commit-Queue: Kevin Lubick <kjlubick@google.com>
Auto-Submit: Kevin Lubick <kjlubick@google.com>
I added 3 tests, one using the drawRect API naively,
one using drawRect with a Malloc'd array, and one using
the drawRect4f.
rough local measurements:
- [baseline with tip of tree code]: 50ms
- naive drawRect: 40ms
- drawRect with malloc: 28ms
- drawRect4f: 27ms
I also tried the benchmarks locally with taking in paint
as a const reference. I did not see any changes, but that
could just be small sample size. I plan to land the code
as is for now, collect a bit of data in Perf and then try
landing the const reference stuff and see if we get
something measurable.
To aid this, I added in a helper list of tests to only run
some benchmarks easily.
Change-Id: I2d8c5296e93f05be45fc12059955fb9d9e339d83
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315143
Reviewed-by: Kevin Lubick <kjlubick@google.com>
This is a reland of bdc214a50e
The CPU version of SkottieWasm is timing out for reasons unknown,
but the GPU version is happy. I think we can get rid of the CPU
version of the job since it has been more or less superseded by
the SkottieFrames one (and the latter is more stable).
Original change's description:
> [canvaskit] Change SkRects to be arrays, not objects.
>
> This changes several APIs, so there are lots of breaking
> notes in the Changelog.
>
> This made the "draw 100 colored regions" benchmark about
> 20% faster (1ms -> .8ms).
>
> In theory, rendering should stay the same.
>
> Change-Id: Ib80b15e2d980ad5d568fff4460d2b529766c1b36
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/312491
> Reviewed-by: Nathaniel Nifong <nifong@google.com>
Change-Id: I674aba85ecfb30b72e94cbaf89b2d97bfae3b7a4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315142
Reviewed-by: Nathaniel Nifong <nifong@google.com>
This reverts commit 370de72763.
Reason for revert: Unrelated to breakage
Original change's description:
> Revert "[canvaskit] Replace RRect objects with TypedArrays."
>
> Revert submission 314622
>
> Reason for revert: breaking wasm bots
> Reverted Changes:
> Ia1ba13814:[canvaskit] Replace RRect objects with TypedArrays...
> Ib80b15e2d:[canvaskit] Change SkRects to be arrays, not objec...
> I790b2d6fc:[canvaskit] Add drawRect4f as example 'fast path' ...
>
> Change-Id: I3912ffeb446b497e4e03a6260e51cae1e664bd83
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314894
> Reviewed-by: Michael Ludwig <michaelludwig@google.com>
TBR=kjlubick@google.com,michaelludwig@google.com,nifong@google.com
# Not skipping CQ checks because this is a reland.
Change-Id: I9b1e38fd36f635146b3b518142bbc4626007700f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315141
Reviewed-by: Kevin Lubick <kjlubick@google.com>
6eeb3d8bf6..0c25fbe3e3
2020-09-03 syoussefi@chromium.org Vulkan: Remove accidental padding in push constants
2020-09-02 syoussefi@chromium.org Vulkan: Allow pbuffer usage if no window support
2020-09-02 geofflang@chromium.org Add a frontend feature to disable anisotropic filtering.
2020-09-02 syoussefi@chromium.org Vulkan: Don't break the RP on masked/scissored clears
2020-09-02 sugoi@google.com Make copyTexImage2D robust when source area is out of bounds
2020-09-02 syoussefi@chromium.org Vulkan: Fix resolve with different formats
2020-09-02 syoussefi@chromium.org Vulkan: Fix resolve-with-subpass with non-zero read attachment
2020-09-02 jonahr@google.com Workaround to disable OES_get_program_binary
2020-09-02 syoussefi@chromium.org Gracefully fail end2end tests if no window support
2020-09-02 sugoi@chromium.org Revert "Make copyTexImage2D robust when source area is out of bounds"
2020-09-02 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Chromium from 863342438ec1 to 09df13c25e1b (356 revisions)
2020-09-02 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Vulkan-ValidationLayers from f1315472247d to b0337a548273 (2 revisions)
2020-09-02 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SPIRV-Tools from 8a0ebd40f86d to c278dada96d9 (5 revisions)
2020-09-02 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from a2e6c1a149f4 to e1051cbaad46 (5 revisions)
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 michaelludwig@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:1034840,chromium:1123524,chromium:1124379
Tbr: michaelludwig@google.com
Change-Id: I06ef5d3cb92e50c66347dccdf0c1b0dd89efe036
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315083
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Mechanically updated via Xcode "Replace Regular Expression":
typedef (.*) INHERITED;
-->
using INHERITED = $1;
The ClangTidy approach generated an even larger CL which would have
required a significant amount of hand-tweaking to be usable.
Change-Id: I671dc9d9efdf6d60151325c8d4d13fad7e10a15b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314999
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
All tests are now GLSL-based instead of .fp based, for ease of
readability. DEF_TESTs which called test() more than once have been
split up into a separate individual DEF_TEST for each test() call, to
simplify fixups going forward.
Change-Id: I9202e95e9cafff3efc0c32aa3f3ca2eea04719fe
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314886
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
This reverts commit ed63444587.
Reason for revert: trying to unblock android roller, which incorrectly thinks this has a merge conflict: https://b.corp.google.com/issues/167576324
Original change's description:
> Add idea of DataType to SkYUVAPixmapInfo.
>
> DataType describes the data type of YUVA channels
> independent of how they are grouped into planes.
>
> Adds mapping functions between SkColorType/channel count
> and DataType.
>
> SkYUVAPixmapInfo can be constructed from DataType and will
> choose appropriate SkColorTypes for each plane.
>
> Valid SkYUVAPixmapInfos now have the same DataType for each
> plane (could relax this in the future, esp for alpha plane).
>
> SkYUVAPixmapInfo::SupportedDataTypes specifies the supported
> combinations of SkYUVAInfo::PlanarConfig and
> kYUVAPixmapInfo::DataType supported by a GrContext (based on
> supported texture formats).
>
> SkImageGenerator/SkCodec YUVA query API now takes a
> SupportedDataTypes.
>
> Change-Id: I8791234638e6ba3396d1e7960b7bc210edc6dd57
> Bug: skia:10632
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314276
> Reviewed-by: Leon Scroggins <scroggo@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
> Commit-Queue: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,robertphillips@google.com,scroggo@google.com
Change-Id: I72c39539a4766f10cac3ca3cdef6c503a8319ff1
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:10632
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314895
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Revert submission 314622
Reason for revert: breaking wasm bots
Reverted Changes:
Ia1ba13814:[canvaskit] Replace RRect objects with TypedArrays...
Ib80b15e2d:[canvaskit] Change SkRects to be arrays, not objec...
I790b2d6fc:[canvaskit] Add drawRect4f as example 'fast path' ...
Change-Id: I3912ffeb446b497e4e03a6260e51cae1e664bd83
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314894
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Revert submission 314622
Reason for revert: breaking wasm bots
Reverted Changes:
Ia1ba13814:[canvaskit] Replace RRect objects with TypedArrays...
Ib80b15e2d:[canvaskit] Change SkRects to be arrays, not objec...
I790b2d6fc:[canvaskit] Add drawRect4f as example 'fast path' ...
Change-Id: Ie6e4c57ba412ca9ff8e4446b06681b51029da2d6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314893
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Revert submission 314622
Reason for revert: breaking wasm bots
Reverted Changes:
Ia1ba13814:[canvaskit] Replace RRect objects with TypedArrays...
Ib80b15e2d:[canvaskit] Change SkRects to be arrays, not objec...
I790b2d6fc:[canvaskit] Add drawRect4f as example 'fast path' ...
Change-Id: I295791532a2d27bff41f706fa8ed37e0499c27d7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314892
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
These checks don't appear to be useful; the methods in question are
overloaded anyway, and don't access the INHERITED:: methods.
Change-Id: I2ae637b7fb7f2281dc00fe148fbbbccb464cded9
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314885
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Seems this was fixed in canvaskit, but overlooked in debugger.
Change-Id: I326b809628a8531a521a6246e65798d37dcce1b7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314457
Commit-Queue: Nathaniel Nifong <nifong@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>
I added 3 tests, one using the drawRect API naively,
one using drawRect with a Malloc'd array, and one using
the drawRect4f.
rough local measurements:
- [baseline with tip of tree code]: 50ms
- naive drawRect: 40ms
- drawRect with malloc: 28ms
- drawRect4f: 27ms
I also tried the benchmarks locally with taking in paint
as a const reference. I did not see any changes, but that
could just be small sample size. I plan to land the code
as is for now, collect a bit of data in Perf and then try
landing the const reference stuff and see if we get
something measurable.
To aid this, I added in a helper list of tests to only run
some benchmarks easily.
Change-Id: I790b2d6fc32c125a4d371768fffb19ab22633d4e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314622
Reviewed-by: Kevin Lubick <kjlubick@google.com>
This changes several APIs, so there are lots of breaking
notes in the Changelog.
This made the "draw 100 colored regions" benchmark about
20% faster (1ms -> .8ms).
In theory, rendering should stay the same.
Change-Id: Ib80b15e2d980ad5d568fff4460d2b529766c1b36
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/312491
Reviewed-by: Nathaniel Nifong <nifong@google.com>
This is about 2.5x faster.
Will try to do the same for Rects in a followup CL.
Change-Id: Ia1ba1381435f4eee2fcf4f0cc2738e3306edd5f9
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/312304
Reviewed-by: Nathaniel Nifong <nifong@google.com>
Make direct2D concrete. Simplify the math.
Change-Id: I4cfa3263cedcd53108e47cfb4f503060ea13a7d8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314458
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
- Section is only allowed in .fp files
- Interface blocks, extensions, and layout qualifiers (not attached
to a declaration) are only allowed in fs/vs/gs.
- Enums are not allowed in runtime effects
Bug: skia:10626
Change-Id: I3de8c6894fa18340f5da942fb7d995e5426ef0d9
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314878
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
The bulk of the diff is just reordering existing code; the logical
changes are very small.
Change-Id: I3b918e64f5229da43d43f0922e8b59a007a6ad3e
Bug: skia:10687
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314882
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
It turns out the parser was accepting empty declarations, turning them
into an empty modifiers declaration. This was causing problems with a
CL that was going to disallow declarations in contexts where they
weren't allowed. Since this was never the intended behavior, we now
disallow it.
Change-Id: Iea56529c76a946e8002ce1e929790aec488fd4f5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314879
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Auto-Submit: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Change-Id: I9d9a18f9d37cac385921fc0bd759407ea2f3db31
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314881
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
The following tests show incorrect results from the inliner:
- SkSLFPInlinedIfBodyMustBeInAScope
- SkSLFPInlinedElseBodyMustBeInAScope
Change-Id: Iafc567f9a97f67f9b734edd348ee25a14939592a
Bug: skia:10687
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314880
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>
Change-Id: I50476e8b8bc9796656cb87cdbd6ec0fe83819f0e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314496
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Auto-Submit: Greg Daniel <egdaniel@google.com>
Change-Id: I43479d8543ea4860be45614a65cf8ad4cec307d8
Bug: skia:10687
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314877
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
I'm not sure why I left this out, or why for so long.
This fixes a few GMs, especially F16 where we're not
otherwise clamped.
Creates a little `Color clamp01(Color)` helper to
dedup this logic a little.
Bug: skia:10689
Change-Id: If879f75b219f910ac3e1491625674bcfc38d36bf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314863
Reviewed-by: Florin Malita <fmalita@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
There's an _mm_movemask_ps() intrinsic that gets at the movmskps
instruction, which grabs the top (sign) bit of each float directly
without needing to reinterpret them as bytes.
I wouldn't really have done this but I think Chrome's clang is
miscompiling the version at head that uses _mm_movemask_epi8(). The
SkNx<2,float> `!(a+b == a*b).anyTrue()` test case fails when I use that
compiler, and spooky things like adding SkDebugf() make it pass again.
Change-Id: Idd0698d46ccfe9a00909faca1c6693a70e91157a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314860
Auto-Submit: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
The plan is to use this table to evaluate the vertical blur
when creating the blurred rrect mask on the cpu.
Bug: 1108408
Change-Id: Ie322b9e2924d7b0f2242b28daebfcf21c8062aa8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314716
Reviewed-by: Adlai Holler <adlai@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This prevents unnecessary work from being done, and also prevents us
from executing side-effects from the wrong side.
Change-Id: I4dbf3974388807f15e9eadb2abf1b1243d047ce2
Bug: skia:10688
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314797
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>
Some of these tests worked perfectly, but others expose latent inlining
bugs. The following tests show incorrect results from the inliner:
- SkSLFPTernaryExpressionsShouldNotInlineResults
- SkSLFPInlinedWhileBodyMustBeInAScope
- SkSLFPInlinedDoWhileBodyMustBeInAScope
- SkSLFPInlinedForBodyMustBeInAScope
Change-Id: I523e2e3272dea01d5c194a478df6f39ecadf2f5c
Bug: skia:10687, skia:10688
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314796
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
- remove some workarounds
- more SI/SIN/SIT/SINT use
- rewrap a lot of code to 100 cols
- etc. misc.
Change-Id: I78b7ff272afcbb8658cf147aad8af85d0e2acf42
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314676
Auto-Submit: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Prepare this method for cpu-side generation of the rrect blur mask in DDL-mode
Bug: 1108408
Change-Id: I902ad507cd53f5f5b1d418a8b52f96dddffd0b18
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314621
Reviewed-by: Adlai Holler <adlai@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
We've been stifling this on Clang builds but not GCC,
for no good reason.
Change-Id: I49d6059dcf7547d25ec83c507c92f115ab199b79
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/314677
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
