clang-r353983 emits error when building Skia for Windows. Do not build
it for now until the compiler issue is addressed.
Bug: 126457671
Change-Id: I2942aadb10f5956d72f064870653694a7e1e847c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/201836
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Test: Ran gn_to_bp.py and built Android
Bug: b/123953909
Change-Id: Icfb7f5d6ff55f4837d1d227cb2956530d492c745
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/199920
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Stan Iliev <stani@google.com>
Bug: b/123085931
Bug: b/117921091
Based on ag/6614841
Don't use the default Windows FontMgr/FontHost. Android will
handle fonts.
Remove _HAS_EXCEPTIONS=0, which was added by gn/BUILD.gn
Change-Id: I93a72a550b2dbb9adc498638853f6e95c5d4e94f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/197041
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
Bug: b/117921091
Bug: b/123085931
Test: mmma external/skia
Based on ag/6098473. Generate an SkUserConfig for Windows, and update
Android.bp with Windows in mind. Disable xps on Windows, as the
Android Windows build system does not provide FontSub.h
Change-Id: I755d225004ccbe273099889fea8dda01d23d03fc
Reviewed-on: https://skia-review.googlesource.com/c/195881
Auto-Submit: Leon Scroggins <scroggo@google.com>
Reviewed-by: Jerome Gaillard <jgaillard@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Presumably if we don't need these in Skia,
we don't need them in Android?
Change-Id: Iff18066c1c2a52c999beabbd9239d4bc96d0ba6d
Reviewed-on: https://skia-review.googlesource.com/c/190438
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
libicuuc and libicui18n are moved into APEX, but
they have no stable ABI due to the version suffix.
Use libandroidicu which provides stable symbol.
See http://go/apex-stable-icu4c-interface for the design.
Bug: b/117094880
Change-Id: Id27cf49b9caad84080a3e09e39243573d921b17f
Reviewed-on: https://skia-review.googlesource.com/c/182202
Reviewed-by: Ben Wagner <bungeman@google.com>
Bug: b/118742766
This does not change Android.bp or (any) SkUserConfig.h. But it
consolidates the arguments for the different platforms into one method,
which makes it clear what they share and where they differ. It also no
longer sets skia_enable_tools for the non-Android build, where it
doesn't need to be set to true (this makes no difference in the final
output nor the time GN takes to generate JSON).
Change-Id: I22af345aaf3878227d4f42afc2861e49ec2691ec
Reviewed-on: https://skia-review.googlesource.com/c/173238
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Based on ag/5616352, patch set 5.
Bug: b/118742766
Test: lunch sdk && mmma external/skia on Mac
Similar to the linux build, use a new set of GN arguments to create a
json object to fill in the updated template.
Add the new include/config/mac folder for the newly generated mac-
specific SkUserConfig.h.
Include stdarg.h in SkTraceEventCommon, which is necessary for the mac
build.
Add a new GN arg (skia_use_fonthost_mac) to allow framework build to
manually not build SkFontHost_mac.cpp, and use the same font host
as the other builds.
Change-Id: I654ba496306a3f3591c3937ad5524cd45e49dd65
Reviewed-on: https://skia-review.googlesource.com/c/173183
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Bug: b/118742766
Update gn_to_bp to write an Android.bp file that will build a host-side
Skia library.
Switch some methods from SK_BUILD_FOR_ANDROID to
SK_BUILD_FOR_ANDROID_FRAMEWORK.
Prior reviews were done at ag/5482397.
gn_to_bp.py:
- Run GN twice - once for android and once for linux
- Disable GPU (depends on a to-be-written host side GL target) and HEIF
(which relies on Android hardware) on linux
- TODO: Turn on GPU on linux
- Split sources into everywhere, android-only, and linux-only.
It seems that Android.bp does not allow using the same cpp
file in multiple targets.
- note that we currently *only* divide out the sources. The cflags are
the same (except for a couple manual ones) and include
directories are mostly the same (again, except for manual ones).
Android has a "gpu" include directory, which I don't expect to
make a difference to the linux build, which isn't using GPU (yet).
- Use the same "custom empty" font manager on the host as on Android
- Write separate SkUserConfig files; one for android and one for linux.
This allows libskia to force libraries that use it to use the right
defines by setting export_include_dirs.
- Add extra checks to SkUserConfig.h to ensure we have only the
appropriate SK_BUILD_FOR macro defined
- Add host_supported: true for libskia
gn_to_bp_utils.py:
- Switch SkUserConfig.h from include guards to pragma once so it is
easier to append to the end. This matches how Android generally
includes headers.
BUILD.gn:
- Add skia_use_fixed_gamma_text so host build can use the same SK_GAMMA
defines as the device.
SkPreConfig.h:
- Stop making SK_BUILD_FOR_ANDROID_FRAMEWORK imply SK_BUILD_FOR_ANDROID.
The host build needs the former defined but not the latter.
SkRegion.cpp/.h:
- Make toString() SK_BUILD_FOR_ANDROID_FRAMEWORK so it can be called on
the host.
SkCamera.h/.cpp:
- Switch methods to SK_BUILD_FOR_ANDROID_FRAMEWORK so they can be called
on the host.
- Make getCameraLocation*() const. They are logically const, and this
allows removing a const_cast + TODO in hwui.
Change-Id: I771f825d06380e01c0488fd1c00df1d8a2454dc0
Reviewed-on: https://skia-review.googlesource.com/c/171231
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Bug: b/118742766
The Android Framework does not need SkFontMgr_android*. Switch it to
using SkFontMgr_custom*. This will also be the SkFontMgr to use when
we're building Skia as a host-side library.
BUILD.gn:
- Rename fontmgr_fuchsia to fontmgr_custom_empty so it can be selected
specifically by the Android Framework build. Default condition is
unchanged, so fuchsia will still build it.
- Similarly, create a new argument for fontmgr_android and
fontmgr_custom, so they can be omitted by the Android Framework build.
Again, default condition is unchanged, so no other builds should be
affected.
gn_to_bp.py:
- Set the new arguments to build the files we want.
Change-Id: Ic85db8acd64f625b00fe2c9bc451ae8814d97a59
Reviewed-on: https://skia-review.googlesource.com/c/170725
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Android's Clang now correctly handles the default alignment for
new/malloc. The flag is still set in BUILD.gn since outside of an
Android tree, we don't have control over the Clang used to build Skia.
Bug: skia:
Change-Id: Ifc71c033c18232a3b0d2faed6f1ecb7eee8b364e
Reviewed-on: https://skia-review.googlesource.com/c/163647
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Chrome's analyze step uses GN and doesn't see changes
to these files as mattering to anything.
I've kept the ones particular to old style opts in
each of their particular targets, and the ones that
are included multiple times (including from outside
src/opts) in core. Kind of arbitrary, but it's at
least close to the right slicing.
Change-Id: I2a5aaeca5b6287c13d7365ec1b63158f48e4e84a
Reviewed-on: https://skia-review.googlesource.com/152040
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Core Skia will always include our own SkiaVulkan.h which basically just
includes vulkan_core.h. All platform vulkan specific stuff must be
include by the client in their own vulkan.h file. Our public interface
is set up that we only use vulkan objects that will be present in all
versions of vulkan headers that the client could include.
Bug: skia:
Change-Id: I6673fd91498eabcc923d65d20f2b5e0a89b4ccf6
Reviewed-on: https://skia-review.googlesource.com/142985
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Change-Id: I6b08cd586d313e3bc41c0da90698fc26ae1a8bb8
Reviewed-on: https://skia-review.googlesource.com/130822
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This reverts commit 98bddf930e.
Reason for revert: <INSERT REASONING HERE>
Original change's description:
> Revert "Create API for GrVkMemoryAllocator and impliment use of AMD VulkanMemoryAllocator on this API."
>
> This reverts commit 26c0e4c1f5.
>
> Reason for revert: breaks roll into fuchsia
>
> Original change's description:
> > Create API for GrVkMemoryAllocator and impliment use of AMD VulkanMemoryAllocator on this API.
> >
> > Bug: skia:
> > Change-Id: I1e122e1b11ab308c2f83cb98c36c81511f4507d0
> > Reviewed-on: https://skia-review.googlesource.com/129980
> > Commit-Queue: Greg Daniel <egdaniel@google.com>
> > Reviewed-by: Brian Salomon <bsalomon@google.com>
> > Reviewed-by: Jim Van Verth <jvanverth@google.com>
>
>
>
> Change-Id: I6c74978f778987c422e6162e7dd85ea9c6baa0e4
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: skia:
> Reviewed-on: https://skia-review.googlesource.com/130182
> Reviewed-by: Greg Daniel <egdaniel@google.com>
> Commit-Queue: Greg Daniel <egdaniel@google.com>
Bug: skia:
Change-Id: I422ffb3562da567f2e85c806286ad1a17c3862cd
Reviewed-on: https://skia-review.googlesource.com/130183
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
This reverts commit 26c0e4c1f5.
Reason for revert: breaks roll into fuchsia
Original change's description:
> Create API for GrVkMemoryAllocator and impliment use of AMD VulkanMemoryAllocator on this API.
>
> Bug: skia:
> Change-Id: I1e122e1b11ab308c2f83cb98c36c81511f4507d0
> Reviewed-on: https://skia-review.googlesource.com/129980
> Commit-Queue: Greg Daniel <egdaniel@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Jim Van Verth <jvanverth@google.com>
TBR=egdaniel@google.com,jvanverth@google.com,bsalomon@google.com,bungeman@google.com
Change-Id: I6c74978f778987c422e6162e7dd85ea9c6baa0e4
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:
Reviewed-on: https://skia-review.googlesource.com/130182
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Bug: skia:
Change-Id: I1e122e1b11ab308c2f83cb98c36c81511f4507d0
Reviewed-on: https://skia-review.googlesource.com/129980
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
This is a reland of 05b5e40519
Original change's description:
> Retry enabling skcms on Android
>
> Since we tried this last, DM has a direct skcms dependency
> via tests/ColorSpaceTest.cpp, so I've rearranged gn_to_bp.py slightly.
>
> We need to keep our eyes out for BitmapRGBAF16Test.testGetPixel in CTS.
>
> Change-Id: I3630bbebb4574522c43362c359d1a14da3bdd480
> Reviewed-on: https://skia-review.googlesource.com/124720
> Commit-Queue: Mike Klein <mtklein@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
Change-Id: I67041fb00687f789c80b7ba41acd3bf2cf257cf3
Reviewed-on: https://skia-review.googlesource.com/126825
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Make Skottie truly optional (own flag), and disable in framework builds
(to unblock landing the RapidJson refactoring).
Change-Id: I4611f915e43fe11c1f6754ab4a9f63e45af2f8d3
Reviewed-on: https://skia-review.googlesource.com/125872
Commit-Queue: Florin Malita <fmalita@chromium.org>
Reviewed-by: Mike Klein <mtklein@google.com>
This reverts commit 05b5e40519.
Reason for revert:
BitmapColorSpaceTest.android.graphics.cts.BitmapColorSpaceTest.inColorSpaceP3ToSRGB (android.graphics.cts) 06.949
java.lang.AssertionError: expected:<67043583> but was:<117375231>
at org.junit.Assert.fail(Assert.java:88)
at org.junit.Assert.failNotEquals(Assert.java:834)
at org.junit.Assert.assertEquals(Assert.java:645)
at org.junit.Assert.assertEquals(Assert.java:631)
at android.graphics.cts.BitmapColorSpaceTest.verifyGetPixel(BitmapColorSpaceTest.java:301)
at android.graphics.cts.BitmapColorSpaceTest.inColorSpaceP3ToSRGB(BitmapColorSpaceTest.java:612)
at java.lang.reflect.Method.invoke(Native Method)
at org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:50)
at org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12)
at org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:52)
at org.junit.internal.runners.statements.InvokeMethod.evaluate(InvokeMethod.java:17)
at org.junit.internal.runners.statements.FailOnTimeout$CallableStatement.call(FailOnTimeout.java:148)
at org.junit.internal.runners.statements.FailOnTimeout$CallableStatement.call(FailOnTimeout.java:142)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.lang.Thread.run(Thread.java:764)
>>> hex(117375231)
'0x6ff00ff'
>>> hex(67043583)
'0x3ff00ff'
Original change's description:
> Retry enabling skcms on Android
>
> Since we tried this last, DM has a direct skcms dependency
> via tests/ColorSpaceTest.cpp, so I've rearranged gn_to_bp.py slightly.
>
> We need to keep our eyes out for BitmapRGBAF16Test.testGetPixel in CTS.
>
> Change-Id: I3630bbebb4574522c43362c359d1a14da3bdd480
> Reviewed-on: https://skia-review.googlesource.com/124720
> Commit-Queue: Mike Klein <mtklein@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
TBR=mtklein@chromium.org,mtklein@google.com,brianosman@google.com
Change-Id: Ib11393c427e9c5fea645de6e6e9a3034dcf06d02
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/125041
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Since we tried this last, DM has a direct skcms dependency
via tests/ColorSpaceTest.cpp, so I've rearranged gn_to_bp.py slightly.
We need to keep our eyes out for BitmapRGBAF16Test.testGetPixel in CTS.
Change-Id: I3630bbebb4574522c43362c359d1a14da3bdd480
Reviewed-on: https://skia-review.googlesource.com/124720
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
We need to update BitmapRGBAF16Test.testGetPixel in CTS.
We're getting off-by-a-bit errors when enabling skcms.
I think we allow a per-byte tolerance of 1 in each of red,green,blue?
I'm not convinced the values in the test or the new ones are any more
correct than the others.
Change-Id: I5196f9315aa8187239eabd58e003d4cb6b603e39
Reviewed-on: https://skia-review.googlesource.com/123246
Reviewed-by: Mike Klein <mtklein@google.com>
Change-Id: Icca4144d2093076e91e0a23d4224695c3a1f51b2
Reviewed-on: https://skia-review.googlesource.com/123180
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
This effectively switches Flutter and Android Framework to use skcms for
color space transformation.
Change-Id: I254f6a8614d6976ff13979ae0e745c49c334e5c6
Reviewed-on: https://skia-review.googlesource.com/123080
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Also use the same profile file as Android's hwui (hwui/hwui.profdata).
Since hwui and skia both also use LTO, the profile files need to be the
same to satisfy the Clang LTO plugin.
Test: Verify that the miscompiles are fixed.
Change-Id: I2a5217b64966144319e63b83351dea2388b60250
Reviewed-on: https://skia-review.googlesource.com/112106
Reviewed-by: Zhizhou Yang <zhizhouy@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
This is a reland of 78cb579f33
This time, lowp stages are controlled by !defined(JUMPER_IS_SCALAR), not
by defined(__clang__). The two are usually the same, except when we opt
Clang builds into JUMPER_IS_SCALAR artificially.
Some Google3 builds use compilers old enough that they barf when
compiling our NEON code. It's conceivably also possible to define
JUMPER_IS_SCALAR yourself, but I don't think anyone does that.
Original change's description:
> Reland "make SkJumper stages normal Skia code"
>
> This is a reland of 22e536e3a1
>
> Now with fixed #include paths in SkRasterPipeline_opts.h,
> and -ffp-contract=fast for the :hsw target to minimize
> diffs on non-Windows Clang AVX2/AVX-512 bots.
>
> Original change's description:
> > make SkJumper stages normal Skia code
> >
> > Enough clients are using Clang now that we can say, use Clang to build
> > if you want these software pipeline stages to go fast.
> >
> > This lets us drop the offline build aspect of SkJumper stages, instead
> > building as part of Skia using the SkOpts framework.
> >
> > I think everything should work, except I've (temporarily) removed
> > AVX-512 support. I will put this back in a follow up.
> >
> > I have had to drop Windows down to __vectorcall and our narrower
> > stage calling convention that keeps the d-registers on the stack.
> > I tried forcing sysv_abi, but that crashed Clang. :/
> >
> > Added a TODO to up the same narrower stage calling convention
> > for lowp stages... we just *don't* today, for no good reason.
> >
> > Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> > Reviewed-on: https://skia-review.googlesource.com/110641
> > Commit-Queue: Mike Klein <mtklein@chromium.org>
> > Reviewed-by: Herb Derby <herb@google.com>
> > Reviewed-by: Florin Malita <fmalita@chromium.org>
>
> Change-Id: I44f2c03d33958e3807747e40904b6351957dd448
> Reviewed-on: https://skia-review.googlesource.com/112742
> Reviewed-by: Mike Klein <mtklein@chromium.org>
Change-Id: I3d71197d4bbb19ca4a94961a97fa2e54d5cbfb0d
Reviewed-on: https://skia-review.googlesource.com/112744
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This reverts commit 78cb579f33.
Reason for revert: lowp should be controlled by defined(JUMPER_IS_SCALAR), not defined(__clang__). So close.
Original change's description:
> Reland "make SkJumper stages normal Skia code"
>
> This is a reland of 22e536e3a1
>
> Now with fixed #include paths in SkRasterPipeline_opts.h,
> and -ffp-contract=fast for the :hsw target to minimize
> diffs on non-Windows Clang AVX2/AVX-512 bots.
>
> Original change's description:
> > make SkJumper stages normal Skia code
> >
> > Enough clients are using Clang now that we can say, use Clang to build
> > if you want these software pipeline stages to go fast.
> >
> > This lets us drop the offline build aspect of SkJumper stages, instead
> > building as part of Skia using the SkOpts framework.
> >
> > I think everything should work, except I've (temporarily) removed
> > AVX-512 support. I will put this back in a follow up.
> >
> > I have had to drop Windows down to __vectorcall and our narrower
> > stage calling convention that keeps the d-registers on the stack.
> > I tried forcing sysv_abi, but that crashed Clang. :/
> >
> > Added a TODO to up the same narrower stage calling convention
> > for lowp stages... we just *don't* today, for no good reason.
> >
> > Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> > Reviewed-on: https://skia-review.googlesource.com/110641
> > Commit-Queue: Mike Klein <mtklein@chromium.org>
> > Reviewed-by: Herb Derby <herb@google.com>
> > Reviewed-by: Florin Malita <fmalita@chromium.org>
>
> Change-Id: I44f2c03d33958e3807747e40904b6351957dd448
> Reviewed-on: https://skia-review.googlesource.com/112742
> Reviewed-by: Mike Klein <mtklein@chromium.org>
TBR=mtklein@chromium.org,herb@google.com,fmalita@chromium.org
Change-Id: Ie64da98f5187d44e03c0ce05d7cb189d4a6e6663
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/112743
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This is a reland of 22e536e3a1
Now with fixed #include paths in SkRasterPipeline_opts.h,
and -ffp-contract=fast for the :hsw target to minimize
diffs on non-Windows Clang AVX2/AVX-512 bots.
Original change's description:
> make SkJumper stages normal Skia code
>
> Enough clients are using Clang now that we can say, use Clang to build
> if you want these software pipeline stages to go fast.
>
> This lets us drop the offline build aspect of SkJumper stages, instead
> building as part of Skia using the SkOpts framework.
>
> I think everything should work, except I've (temporarily) removed
> AVX-512 support. I will put this back in a follow up.
>
> I have had to drop Windows down to __vectorcall and our narrower
> stage calling convention that keeps the d-registers on the stack.
> I tried forcing sysv_abi, but that crashed Clang. :/
>
> Added a TODO to up the same narrower stage calling convention
> for lowp stages... we just *don't* today, for no good reason.
>
> Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> Reviewed-on: https://skia-review.googlesource.com/110641
> Commit-Queue: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Herb Derby <herb@google.com>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
Change-Id: I44f2c03d33958e3807747e40904b6351957dd448
Reviewed-on: https://skia-review.googlesource.com/112742
Reviewed-by: Mike Klein <mtklein@chromium.org>
This reverts commit 22e536e3a1.
Reason for revert: wrong include path :/
Original change's description:
> make SkJumper stages normal Skia code
>
> Enough clients are using Clang now that we can say, use Clang to build
> if you want these software pipeline stages to go fast.
>
> This lets us drop the offline build aspect of SkJumper stages, instead
> building as part of Skia using the SkOpts framework.
>
> I think everything should work, except I've (temporarily) removed
> AVX-512 support. I will put this back in a follow up.
>
> I have had to drop Windows down to __vectorcall and our narrower
> stage calling convention that keeps the d-registers on the stack.
> I tried forcing sysv_abi, but that crashed Clang. :/
>
> Added a TODO to up the same narrower stage calling convention
> for lowp stages... we just *don't* today, for no good reason.
>
> Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> Reviewed-on: https://skia-review.googlesource.com/110641
> Commit-Queue: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Herb Derby <herb@google.com>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
TBR=mtklein@chromium.org,herb@google.com,fmalita@chromium.org
Change-Id: I2bdc709c80cdfa6b13ff24e024b3721bef887f46
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/112741
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Enough clients are using Clang now that we can say, use Clang to build
if you want these software pipeline stages to go fast.
This lets us drop the offline build aspect of SkJumper stages, instead
building as part of Skia using the SkOpts framework.
I think everything should work, except I've (temporarily) removed
AVX-512 support. I will put this back in a follow up.
I have had to drop Windows down to __vectorcall and our narrower
stage calling convention that keeps the d-registers on the stack.
I tried forcing sysv_abi, but that crashed Clang. :/
Added a TODO to up the same narrower stage calling convention
for lowp stages... we just *don't* today, for no good reason.
Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
Reviewed-on: https://skia-review.googlesource.com/110641
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Android media framework is expected to be run with the background
thread pool for handling binder RPC calls.
Test: skia_dm does not hang with heif files.
Bug: 65463215 Bug: 72869975
Change-Id: I2333d5a13d6145e9468464fa93807851e3a17bc8
Reviewed-on: https://skia-review.googlesource.com/108141
Reviewed-by: Chong Zhang <chz@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
This refactors the current gn_to_bp script and moves the common
components into a utils class. Both gn_to_bp scripts now also
accept an optional param (--gn <gn_path>) to make it easier to run
within the Android tree where we don't have the DEPS synced.
Change-Id: Idc4de7b3993e63e21a5b7137b1873d82a8e1843e
Reviewed-on: https://skia-review.googlesource.com/102184
Commit-Queue: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Bug: b/63909536
SkAnimatedImage is a simple drawable for animating a GIF. Thread-safety
is left up to the client. At most two bitmaps are stored in the
drawable; one for the current frame and one for a frame that may need to
be restored. The backup frame prevents some cases where we would
otherwise have to re-decode from the beginning of the image.
The API lets the client set the time value, and decodes to match that
time.
TODO:
- Callback for when the animation is complete
- Ability to use SkAndroidCodec
- Modify the loop count (or leave that up to client?)
- Better and/or client-specific caching
Other changes:
- Add a sample which animates a GIF
- Reenable SK_CODEC_PRINTF for debug builds and Android
Change-Id: I945ffbccdb6008f2a05ed4d9b2af869a261fb300
Reviewed-on: https://skia-review.googlesource.com/93420
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Doing so installs packages the resource directory during a 'tests'
build. This lets us run skia_nanobench in the internal test infra,
which doesn't have access to the source tree.
Test: Check that all files in "resources" is packaged during 'make tests
dist' in an Android tree.
Change-Id: Ia547cf09087e5795bdd825d0d48f179086a9ca4a
Reviewed-on: https://skia-review.googlesource.com/93060
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Stephen Hines <srhines@google.com>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
This patch is to turn on PGO for skia on Android, which will provide a
performance improvement of 7.6% for hwui when applied PGO for both.
The patch specified a skia.profdata file, which locates in internal
google_data/pgo_profile directory, to work as the profile to feed PGO.
This profdata can be re-collected with PGO build system support.
PGO can be turned off by setting ANDROID_PGO_NO_PROFILE_USE environment
variable or set variable to false.
Test: Build skia successfully and verified the performance improvement
on device through benchmark.
BUG=skia:
Change-Id: I826f417569b2853630f6d4fcce236b5bc36547fa
Reviewed-on: https://skia-review.googlesource.com/82880
Commit-Queue: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Reviewed-by: John Reck <jreck@google.com>
Follow up to I04a5b013dd0ec8f162046388e89cf08cb9476bab
Change-Id: I5ab77564fe867ad1fef48fb322ff75869c025776
Reviewed-on: https://skia-review.googlesource.com/79040
Commit-Queue: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Speculative fix for Android roll, broken here: ag/3293969.
We used to make an exception for this warning on Skia builds, until I
disabled it in https://skia-review.googlesource.com/c/skia/+/60530,
because it no longer seemed necessary: Turning on most warnings without
this exception in Android framework builds, as I did in
https://skia-review.googlesource.com/55703 did not break Android.
I'm not sure why Skia's waterfall did not show the problem, but this
should fix it on the Android framework side.
TBR=mtklein@chromium.org
Change-Id: I04a5b013dd0ec8f162046388e89cf08cb9476bab
Reviewed-on: https://skia-review.googlesource.com/78580
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Bug: b/66996870
Android wants to build with warnings as errors. Keep the external/skia
warnings in sync with other Skia builds by pulling them from the json
generated by GN.
Fix a couple small errors that show up in the framework build.
Uploaded in response to AOSP's
https://android-review.googlesource.com/#/c/platform/external/skia/+/498211/
Change-Id: I4d791d43a9b00f9d6b79ecf16839716f241cba99
Reviewed-on: https://skia-review.googlesource.com/55703
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
This CL will enable Android to statically link libskia into libhwui
while still exposing the symbols needed by other components of the
system.
Bug: b/31971097
Change-Id: Ib6c87331dbe456e247b46a34c38da4863dfe02f0
Reviewed-on: https://skia-review.googlesource.com/49766
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
This reverts commit c902ff8fc8.
Reason for revert: not gonna do this afterall.
Original change's description:
> turn exceptions on in test tools on Android
>
> This is a follow-up to https://skia-review.googlesource.com/c/34982
> which did the same for locally built test tools.
>
> Change-Id: Id97841a64521fda99cb952a1a751ffc10f636f53
> Reviewed-on: https://skia-review.googlesource.com/35162
> Commit-Queue: Mike Klein <mtklein@google.com>
> Reviewed-by: Leon Scroggins <scroggo@google.com>
TBR=mtklein@chromium.org,mtklein@google.com,herb@google.com,scroggo@google.com
Change-Id: Iae6c817d625989e3a427dbc18328477382d892bd
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/35524
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Since both armv7-a-neon and 32-bit armv8-a have NEON, we can treat them
the same in Android.bp.
Bug: b/62895439
Corresponds to https://android-review.googlesource.com/c/423660/3
This change will generate the change to Android.bp described there.
Change-Id: Icae9b5b79093d6f2886da39771d4fbe901be237a
Reviewed-on: https://skia-review.googlesource.com/33000
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
CQ_INCLUDE_TRYBOTS=skia.primary:Test-Debian9-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD
Bug: skia:
Change-Id: If6f0d0a57463bf99a66d674e65a62ce3931d0116
Reviewed-on: https://skia-review.googlesource.com/24644
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Remove #defines that lived in gn_to_bp.py and
android_framework_defines.gni. These have been moved into a new file
in Android, SkUserConfigManual.h, in https://googleplex-android-review.git.corp.google.com/#/c/2519600/
Update gn_to_bp.py to include SkUserConfigManual.h, so it will still
result in using the same #defines.
Lately, we've found it difficult to guard changes behind a flag. e.g.
a change to drawing causes a CTS failure in Android, so we have to do
the following:
- put the change behind a flag, and add it to gn_to_bp.py or
android_framework_defines.gni
- generate new images on Android (by running CTS with external/skia
modified to not define the flag)
- create a CL in CTS that uses the new images
- land a CL in Skia that stops defining the flag
- when the Skia change lands, wait for the auto-roller to create a CL
that includes the change, stop the auto-roller, add the topic to the CTS
CL so the two can land at the same time
- land both Android changes (with TreeHugger)
- restart the Android auto-roller
With SkUserConfigManual.h (which lives in Android), the process will
be similar to Chromium:
- land a CL in Android's external/skia that defines a flag e.g.
SK_SUPPORT_LEGACY_FEATURE. Land without TreeHugger because it isn't used
in Skia and does not do anything
- land a change in Skia that changes behavior unless
SK_SUPPORT_LEGACY_FEATURE is defined. This will safely go through the
Android roll and not change any behavior for Android
- create two Android CLs - one in CTS to use the new images, and one in
external/skia to delete SK_SUPPORT_LEGACY_FEATURE. Set them to the same
topic and land them with TreeHugger
In the new process, there is no need to mess with the Android roll.
A downside to the new process is that we cannot test the android
framework defines without checking in to Android. But given how much
we've progressed in automating Android testing, this is fine.
Bug: b/63429612
Change-Id: Idfbaef2f4cae641a75fb6e7bf70428733a441336
Reviewed-on: https://skia-review.googlesource.com/22072
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Eric Boren <borenet@google.com>
Create a new SkImage public API to make an image from an Android
hardware buffer. Implementation is using a SkImageGenerator
derived class GrAndroidBufferImageGenerator.
A new EGLImage texture is created, which is then wrapped with
GrTextureProxy.
Bug: skia:
Change-Id: I610a4c5a58198686ce7c03e9a0adad3f9d2342e0
Reviewed-on: https://skia-review.googlesource.com/17789
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Stan Iliev <stani@google.com>
This reverts commit 3a3bc42b7d.
Reason for revert: still breaking android
Original change's description:
> Check-in vulkan.h into third_party and use that instead of local sdk vulkan.h
>
> This change is needed since once we start getting support for varrying of extensions
> and newer version support in general, we need a common vulkan header to compile off of.
> Otherwise we will run into problems if clients have older headers that don't include
> functions/symbols we are trying to use.
>
> Additionally it has the benefit of not needing to add if SK_VULKAN around code in
> include which wants to use vulkan symbols.
>
> This is a reupload of CL: https://skia-review.googlesource.com/13651
>
> Bug: skia:
> Change-Id: I091f526b8c4a61774c34834cd7bfb7e2c822ff5c
> Reviewed-on: https://skia-review.googlesource.com/13804
> Reviewed-by: Derek Sollenberger <djsollen@google.com>
> Commit-Queue: Greg Daniel <egdaniel@google.com>
>
TBR=djsollen@google.com,egdaniel@google.com,mtklein@google.com,jvanverth@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: Ic595e32005761170156499cfb6efc1acfce96001
Reviewed-on: https://skia-review.googlesource.com/13806
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
This change is needed since once we start getting support for varrying of extensions
and newer version support in general, we need a common vulkan header to compile off of.
Otherwise we will run into problems if clients have older headers that don't include
functions/symbols we are trying to use.
Additionally it has the benefit of not needing to add if SK_VULKAN around code in
include which wants to use vulkan symbols.
This is a reupload of CL: https://skia-review.googlesource.com/13651
Bug: skia:
Change-Id: I091f526b8c4a61774c34834cd7bfb7e2c822ff5c
Reviewed-on: https://skia-review.googlesource.com/13804
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Rebased on top of _win.S change, which I hope lands first.
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel,win_chrome_official,win_chromium_dbg_ng,win_chromium_rel_ng,win_chromium_compile_dbg_ng,win_chromium_compile_rel_ng,win_clang;master.tryserver.chromium.android:android_optional_gpu_tests_rel
Change-Id: I30e97dbb55b8a42251b46f5607096ca12bc670b9
Reviewed-on: https://skia-review.googlesource.com/10462
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
This reverts commit 135555101c.
Reason for revert: forgot g3, seems to break Chrome NaCl builders?
https://build.chromium.org/p/chromium.chrome/builders/Google%20Chrome%20Win/builds/15969/steps/compile/logs/stdio
Original change's description:
> Turn on SkJumper all the time.
>
> If the previous CL sticks, it's now easy to turn on SkJumper everywhere:
> I was mostly holding back because of build system complexity.
>
> This has the main effect of turning on SkJumper in Chromium.
> It's already been on on our local test bots and on Android framework.
>
> Change-Id: I7fbfc6aaaa7dace9c3f2cb509583c69b10997dbf
> Reviewed-on: https://skia-review.googlesource.com/9380
> Reviewed-by: Mike Klein <mtklein@chromium.org>
> Commit-Queue: Mike Klein <mtklein@chromium.org>
>
TBR=mtklein@chromium.org,herb@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: I69747e9e75b2a8e3d78f6c150c9b6d6453e00632
Reviewed-on: https://skia-review.googlesource.com/9444
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
If the previous CL sticks, it's now easy to turn on SkJumper everywhere:
I was mostly holding back because of build system complexity.
This has the main effect of turning on SkJumper in Chromium.
It's already been on on our local test bots and on Android framework.
Change-Id: I7fbfc6aaaa7dace9c3f2cb509583c69b10997dbf
Reviewed-on: https://skia-review.googlesource.com/9380
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
This makes is_official_build turn off all development targets and
features in Skia, including building third-party dependencies from
source.
This will intentionally break some external users, who will find
themselves no longer able to find third-party headers or link against
third-party libraries. These users have been building with our testing
third-party dependencies unknowingly. They'll need to either explicitly
turn back on building each dependency from source
(skia_use_system_foo=false) or disable that dependency entirely
(skia_use_foo=false).
is_skia_standalone is now basically !is_official_build, so I've
propagated that through, removing is_skia_standalone. In a few places
we were using it as a stand-in for defined(ndk), so I've just written
defined(ndk) there. Duh.
gn_to_bp:
is_offical_build's new strength also makes gn_to_bp.py simpler to
write. In spirit, Android builds are official Skia builds that also
build DM and nanobench.
It seems that SkJumper (src/jumper/*) is (unintentionally) enabled
on Android. Switching to an is_official_build would have disabled
that. But as that accidental launch seems to have gone fine, I've
kept it explicitly enabled.
In the end, no changes to Android.bp or its SkUserConfig.h.
The -Mini builder no longer needs to explicitly disable tools.
CQ_INCLUDE_TRYBOTS=skia.primary:Build-Ubuntu-Clang-x86_64-Release-Mini
Change-Id: Id06e53268a5caf55c6046ada354a0863c3031c73
Reviewed-on: https://skia-review.googlesource.com/9190
Reviewed-by: Leon Scroggins <scroggo@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Currently gn_to_bp GrabDependentSrcs only grabs the top level deps of
a target. In a proposed change there are more levels of deps to
clarify the real dependencies. To make this work well, a number of
lists are now sets to avoid including the same file multiple times.
Change-Id: I5005718e42762fe411215920e8991ab27138c7eb
Reviewed-on: https://skia-review.googlesource.com/8936
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
This change was made in the client master, but, since the bp file is
autogenerated, it has broken all subsequent rolls.
BUG=skia:
Change-Id: I722c020414f2700c50bf4a1b5936d993e44864a1
Reviewed-on: https://skia-review.googlesource.com/8730
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Matt Sarett <msarett@google.com>
SkJumper's looking promising enough that I want this further out of my face.
Change-Id: I6dbe71aeabe32f7f4258ba157460e6985733b0ce
Reviewed-on: https://skia-review.googlesource.com/8528
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Generate targets for dm and nanobench from ninja and add them to the
generated Android.bp file.
Remove nanobenchAndroid and SkAndroidSDKCanvas. These rely on HWUI
internals and are currently unused.
Update gyp file references to removed files, just in case.
Change-Id: Ic6ae18a70bfd0c33804e7996d077f2081dfdfe07
Reviewed-on: https://skia-review.googlesource.com/7635
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Leon Scroggins <scroggo@google.com>
- only depend on cpu-features when we know we have an NDK
- disable SkSplicer
Change-Id: I89e4cc70d6ddf0ebb7005d1cb453768d946cd205
Reviewed-on: https://skia-review.googlesource.com/7060
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
There are only a couple of these, primarily focused on index8 srcs and 565 dsts. The burden's starting to outweigh the benefit. No one on the team knows MIPS assembly.
If we're going to try this again, I'd rather we try some sort of SkNx / compiler-intrinsic based approach, probably targeting MIPS SIMD (MSA), not this older instruction set.
We already ignore these files for 64-bit MIPS. This just closes the loop on 32-bit MIPS.
CQ_INCLUDE_TRYBOTS=skia.primary:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD,Build-Ubuntu-Clang-mips64el-Debug-Android,Build-Ubuntu-Clang-mips64el-Release-Android,Build-Ubuntu-Clang-mipsel-Debug-Android,Build-Ubuntu-Clang-mipsel-Release-Android
BUG=skia:6065
Change-Id: Iecac15b56f59625b2e743ea36e7791b90bb0b422
Reviewed-on: https://skia-review.googlesource.com/6353
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
This reverts commit eeb7137a0b.
Reason for revert: well, duh, I guess we'd better update the GYP and Google3 builds...
Original change's description:
> Do not build the ktx encoder for android framework
>
> Move SkKTXImageEncoder.cpp into an optional block, and disable that
> block for the android framework. Use a new define to determine whether
> to define the entry point, rather than using
> SK_BUILD_FOR_ANDROID_FRAMEWORK.
>
> Change-Id: I41103459135af744cf5715f27783c63dc37a7ad1
> Reviewed-on: https://skia-review.googlesource.com/5982
> Commit-Queue: Leon Scroggins <scroggo@google.com>
> Commit-Queue: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Mike Klein <mtklein@chromium.org>
>
TBR=mtklein@chromium.org,mtklein@google.com,scroggo@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: I8da75db31884b5148f7f85a6a0c3e6913b71cfa8
Reviewed-on: https://skia-review.googlesource.com/6021
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Move SkKTXImageEncoder.cpp into an optional block, and disable that
block for the android framework. Use a new define to determine whether
to define the entry point, rather than using
SK_BUILD_FOR_ANDROID_FRAMEWORK.
Change-Id: I41103459135af744cf5715f27783c63dc37a7ad1
Reviewed-on: https://skia-review.googlesource.com/5982
Commit-Queue: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Instead of relying on cpu-features.c, just do what it does.
Good reading: http://man7.org/linux/man-pages/man3/getauxval.3.html
While it's nice to use the headers when possible, should either of these headers not be available, we can fall back to doing it all manually:
extern "C" uint32_t getauxval(uint32_t)
static const int AT_HWCAP = 16;
static const int HWCAP_CRC32 = (1<<7);
To keep things simple I've slimmed cpu feature detection down to just the features we actually make use of. This removes all runtime feature detection for ARMv7... we expect NEON to be globally available, and so far we haven't used the other FMA/FP16 bits on ARMv7. ARMv8 feature dection remains the same, CRC32 before, CRC32 after. x86 (cpuid-based detection) and MIPS (nothing) are untouched.
We need to keep //third_party/cpu-features for //third_party/libwebp.
Change-Id: I6c96df9a09ae68c8c0e54c1152aa177ba9bafc83
Reviewed-on: https://skia-review.googlesource.com/5800
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
in response to 5784
Change-Id: I3ad34a30743e7ffbd04767668c288a4f884eb19c
Reviewed-on: https://skia-review.googlesource.com/5732
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
The general idea here is, run GN in --ide=json mode to get most information.
Then, read a couple .gni files to get the rest (platform specific source lists, Android framework defines).
For now, I'm generating Android.bp and SkUserConfig.h. I figure we can do DM and nanobench once these work?
Change-Id: I8e7f60d6572f2d4769760cf872895518a15d841b
Reviewed-on: https://skia-review.googlesource.com/5554
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
