| Commit message (Collapse) | Author | Age |
... | |
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1341043002
|
|
|
|
|
|
|
|
| |
Very right, it's not prepared to handle return-NULL mallocs at all.
BUG=530759
Review URL: https://codereview.chromium.org/1339093002
|
|
|
|
|
|
|
|
| |
Add an assert to make sure we're not calling this unimplemented method.
BUG=skia:3978
Review URL: https://codereview.chromium.org/1341753003
|
|
|
|
|
|
| |
BUG=skia:4334
Review URL: https://codereview.chromium.org/1338373002
|
|
|
|
|
|
| |
BUG=skia:4332
Review URL: https://codereview.chromium.org/1345523002
|
|
|
|
|
|
| |
BUG=skia:4333
Review URL: https://codereview.chromium.org/1339103002
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1340563003
|
|
|
|
|
|
|
|
|
| |
just move block of code to expose it
BUG=skia:4328
TBR=
Review URL: https://codereview.chromium.org/1334033004
|
|
|
|
|
|
|
| |
This will allow us to test this without hacking it in, might be useful
for others too.
Review URL: https://codereview.chromium.org/1338003002
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1333423003
|
|
|
|
|
|
|
| |
BUG=skia:
TBR=scroggo
Review URL: https://codereview.chromium.org/1339753002
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1336763003
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1333273003
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1332923003
|
|
|
|
|
|
|
| |
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/1332393002
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
https://codereview.chromium.org/1264103003/ )
Also reverts https://codereview.chromium.org/1333003002/ which was layered on top.
Reason for revert:
Appears to leak GDI handles: http://build.chromium.org/p/chromium.memory.fyi/builders/Windows%20Unit%20%28DrMemory%20full%29%20%282%29/builds/8247
~~Dr.M~~ Error #1: HANDLE LEAK: GDI handle 0x03050a84 and 3 similar handle(s) were opened but not closed:
~~Dr.M~~ # 0 system call NtGdiCreateDIBSection
~~Dr.M~~ # 1 GDI32.dll!CreateDIBSection +0xdc (0x768ead23 <GDI32.dll+0x1ad23>)
~~Dr.M~~ # 2 skia.dll!HDCOffscreen::draw [third_party\skia\src\ports\skfonthost_win.cpp:499]
~~Dr.M~~ # 3 skia.dll!SkScalerContext_GDI::generateImage [third_party\skia\src\ports\skfonthost_win.cpp:1233]
~~Dr.M~~ # 4 skia.dll!SkScalerContext::getImage [third_party\skia\src\core\skscalercontext.cpp:530]
~~Dr.M~~ # 5 skia.dll!SkGlyphCache::OnceFillInImage [third_party\skia\src\core\skglyphcache.cpp:252]
~~Dr.M~~ # 6 skia.dll!sk_once_slow<> [third_party\skia\include\core\skonce.h:76]
~~Dr.M~~ # 7 skia.dll!SkGlyphCache::findImage [third_party\skia\src\core\skglyphcache.cpp:260]
~~Dr.M~~ # 8 skia.dll!D1G_RectClip [third_party\skia\src\core\skdraw.cpp:1479]
~~Dr.M~~ # 9 skia.dll!SkDraw::drawPosText [third_party\skia\src\core\skdraw.cpp:1838]
~~Dr.M~~ #10 skia.dll!SkBitmapDevice::drawPosText [third_party\skia\src\core\skbitmapdevice.cpp:348]
~~Dr.M~~ #11 skia.dll!SkCanvas::onDrawPosText [third_party\skia\src\core\skcanvas.cpp:2433]
~~Dr.M~~ #12 skia.dll!SkCanvas::drawPosText [third_party\skia\src\core\skcanvas.cpp:2507]
~~Dr.M~~ #13 skia.dll!SkRecords::Draw::draw<> [third_party\skia\src\core\skrecorddraw.cpp:109]
~~Dr.M~~ #14 skia.dll!SkRecord::Record::visit<> [third_party\skia\src\core\skrecord.h:170]
~~Dr.M~~ #15 skia.dll!SkRecordDraw [third_party\skia\src\core\skrecorddraw.cpp:55]
~~Dr.M~~ #16 skia.dll!SkBigPicture::playback [third_party\skia\src\core\skbigpicture.cpp:43]
~~Dr.M~~ #17 skia.dll!SkCanvas::onDrawPicture [third_party\skia\src\core\skcanvas.cpp:2800]
~~Dr.M~~ #18 skia.dll!SkCanvas::drawPicture [third_party\skia\src\core\skcanvas.cpp:2770]
~~Dr.M~~ #19 cc.dll!cc::DrawingDisplayItem::Raster [cc\playback\drawing_display_item.cc:51]
~~Dr.M~~ #20 cc.dll!cc::DisplayItemList::Raster [cc\playback\display_item_list.cc:107]
~~Dr.M~~ #21 cc.dll!cc::DisplayListRasterSource::RasterCommon [cc\playback\display_list_raster_source.cc:122]
~~Dr.M~~ #22 cc.dll!cc::DisplayListRasterSource::PlaybackToCanvas [cc\playback\display_list_raster_source.cc:100]
~~Dr.M~~ #23 cc.dll!cc::TileTaskWorkerPool::PlaybackToMemory [cc\raster\tile_task_worker_pool.cc:208]
~~Dr.M~~ #24 cc.dll!cc::OneCopyTileTaskWorkerPool::PlaybackAndCopyOnWorkerThread [cc\raster\one_copy_tile_task_worker_pool.cc:413]
~~Dr.M~~ #25 cc.dll!cc::`anonymous namespace'::RasterBufferImpl::Playback [cc\raster\one_copy_tile_task_worker_pool.cc:53]
~~Dr.M~~ #26 cc.dll!cc::`anonymous namespace'::RasterTaskImpl::Raster [cc\tiles\tile_manager.cc:131]
~~Dr.M~~ #27 cc.dll!cc::`anonymous namespace'::RasterTaskImpl::RunOnWorkerThread [cc\tiles\tile_manager.cc:90]
~~Dr.M~~ #28 cc.dll!cc::TaskGraphRunner::RunTaskWithLockAcquired [cc\raster\task_graph_runner.cc:418]
~~Dr.M~~ #29 cc.dll!cc::TaskGraphRunner::Run [cc\raster\task_graph_runner.cc:361]
~~Dr.M~~ #30 base.dll!base::SimpleThread::ThreadMain [base\threading\simple_thread.cc:66]
~~Dr.M~~ #31 base.dll!base::`anonymous namespace'::ThreadFunc [base\threading\platform_thread_win.cc:82]
~~Dr.M~~ #32 KERNEL32.dll!BaseThreadInitThunk +0x11 (0x7570337a <KERNEL32.dll+0x1337a>)
~~Dr.M~~ Note: @0:15:51.087 in thread 196
~~Dr.M~~ Note: handles created with the same callstack are closed here:
~~Dr.M~~ Note: # 0 system call NtGdiDeleteObjectApp
~~Dr.M~~ Note: # 1 GDI32.dll!DeleteObject +0x149 (0x768e57d3 <GDI32.dll+0x157d3>)
~~Dr.M~~ Note: # 2 skia.dll!HDCOffscreen::draw [third_party\skia\src\ports\skfonthost_win.cpp:471]
~~Dr.M~~ Note: # 3 skia.dll!SkScalerContext_GDI::generateImage [third_party\skia\src\ports\skfonthost_win.cpp:1233]
~~Dr.M~~ Note: # 4 skia.dll!SkScalerContext::getImage [third_party\skia\src\core\skscalercontext.cpp:530]
~~Dr.M~~ Note: # 5 skia.dll!SkGlyphCache::OnceFillInImage [third_party\skia\src\core\skglyphcache.cpp:252]
~~Dr.M~~ Note: # 6 skia.dll!sk_once_slow<> [third_party\skia\include\core\skonce.h:76]
~~Dr.M~~ Note: # 7 skia.dll!SkGlyphCache::findImage [third_party\skia\src\core\skglyphcache.cpp:260]
~~Dr.M~~ Note: # 8 skia.dll!D1G_RectClip [third_party\skia\src\core\skdraw.cpp:1479]
~~Dr.M~~ Note: # 9 skia.dll!SkDraw::drawPosText [third_party\skia\src\core\skdraw.cpp:1838]
~~Dr.M~~ Note: #10 skia.dll!SkBitmapDevice::drawPosText [third_party\skia\src\core\skbitmapdevice.cpp:348]
~~Dr.M~~ Note: #11 skia.dll!SkCanvas::onDrawPosText [third_party\skia\src\core\skcanvas.cpp:2433]
~~Dr.M~~ Note: #12 skia.dll!SkCanvas::drawPosText [third_party\skia\src\core\skcanvas.cpp:2507]
~~Dr.M~~ Note: #13 skia.dll!SkRecords::Draw::draw<> [third_party\skia\src\core\skrecorddraw.cpp:109]
~~Dr.M~~ Note: #14 skia.dll!SkRecord::Record::visit<> [third_party\skia\src\core\skrecord.h:170]
~~Dr.M~~ Note: #15 skia.dll!SkRecordDraw [third_party\skia\src\core\skrecorddraw.cpp:55]
~~Dr.M~~ Note: #16 skia.dll!SkBigPicture::playback [third_party\skia\src\core\skbigpicture.cpp:43]
~~Dr.M~~ Note: #17 skia.dll!SkCanvas::onDrawPicture [third_party\skia\src\core\skcanvas.cpp:2800]
~~Dr.M~~ Note: #18 skia.dll!SkCanvas::drawPicture [third_party\skia\src\core\skcanvas.cpp:2770]
~~Dr.M~~ Note: #19 cc.dll!cc::DrawingDisplayItem::Raster [cc\playback\drawing_display_item.cc:51]
~~Dr.M~~ Note: #20 cc.dll!cc::DisplayItemList::Raster [cc\playback\display_item_list.cc:107]
~~Dr.M~~ Note: #21 cc.dll!cc::DisplayListRasterSource::RasterCommon [cc\playback\display_list_raster_source.cc:122]
~~Dr.M~~ Note: #22 cc.dll!cc::DisplayListRasterSource::PlaybackToCanvas [cc\playback\display_list_raster_source.cc:100]
~~Dr.M~~ Note: #23 cc.dll!cc::TileTaskWorkerPool::PlaybackToMemory [cc\raster\tile_task_worker_pool.cc:208]
~~Dr.M~~ Note: #24 cc.dll!cc::OneCopyTileTaskWorkerPool::PlaybackAndCopyOnWorkerThread [cc\raster\one_copy_tile_task_worker_pool.cc:413]
~~Dr.M~~ Note: #25 cc.dll!cc::`anonymous namespace'::RasterBufferImpl::Playback [cc\raster\one_copy_tile_task_worker_pool.cc:53]
~~Dr.M~~ Note: #26 cc.dll!cc::`anonymous namespace'::RasterTaskImpl::Raster [cc\tiles\tile_manager.cc:131]
~~Dr.M~~ Note: #27 cc.dll!cc::`anonymous namespace'::RasterTaskImpl::RunOnWorkerThread [cc\tiles\tile_manager.cc:90]
~~Dr.M~~ Note: #28 cc.dll!cc::TaskGraphRunner::RunTaskWithLockAcquired [cc\raster\task_graph_runner.cc:418]
~~Dr.M~~ Note: #29 cc.dll!cc::TaskGraphRunner::Run [cc\raster\task_graph_runner.cc:361]
~~Dr.M~~ Note: #30 base.dll!base::SimpleThread::ThreadMain [base\threading\simple_thread.cc:66]
~~Dr.M~~ Note: #31 base.dll!base::`anonymous namespace'::ThreadFunc [base\threading\platform_thread_win.cc:82]
~~Dr.M~~ Note: #32 KERNEL32.dll!BaseThreadInitThunk +0x11 (0x7570337a <KERNEL32.dll+0x1337a>)
Original issue's description:
> Parallel cache.
>
> TBR=reed@google.com
>
> BUG=skia:1330
>
> Committed: https://skia.googlesource.com/skia/+/6f2a486040cb25465990196c229feb47e668e87f
>
> Committed: https://skia.googlesource.com/skia/+/bf2988833e5a36c6b430da6fdd2cfebd0015adec
TBR=reed@google.com,mtklein@google.com,mtklein@chromium.org,herb@google.com
BUG=skia:1330
[mtklein mucking around]
NOTREECHECKS=true
Review URL: https://codereview.chromium.org/1339493002
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
id:40001 of https://codereview.chromium.org/1333983002/ )
Reason for revert:
Unexpected perf impact, and a whole bunch of new images in gold (mostly invisibly different).
Original issue's description:
> use new shuffle to speed up affine matrix mappts
>
> sse: 25 -> 18
> neon: 95 -> 86
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/e70afc9f48d00828ee6b707899a8ff542b0e8b98
TBR=reed@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/1335003002
|
|
|
|
|
|
|
|
|
|
|
|
| |
Possible follow-up changes to consider
1. Roll SkImage_Raster and _Gpu into _Generator, where the generator (or cacherator) is backed by a pre-existing texture or raster.
2. Evolve SkImageUsageType into a verb requiring stretching, and have the caller (common code) digest the caps() and usage, so that subclasses are just told what to do (stretch or not)
3. Common code/utility to convert an unstretched texture into a stretch one (and cache it) if the generator can only make an unstretched one.
BUG=skia:
Review URL: https://codereview.chromium.org/1282363002
|
|
|
|
|
|
|
|
|
| |
sse: 25 -> 18
neon: 95 -> 86
BUG=skia:
Review URL: https://codereview.chromium.org/1333983002
|
|
|
|
|
|
|
|
| |
This shows up in TSAN runs as a use-after-free warning.
BUG=skia:
Review URL: https://codereview.chromium.org/1333003002
|
|
|
|
|
|
|
|
|
|
|
| |
This allows us to express shuffles more directly in code while also giving us a
convenient point to platform-specify particular shuffles for particular types.
No specializations yet. Everyone just uses the (pretty good) default option.
BUG=skia:
Review URL: https://codereview.chromium.org/1301413006
|
|
|
|
| |
Review URL: https://codereview.chromium.org/1330213004 .
|
|
|
|
|
|
|
|
|
| |
bitmap
BUG= 529995
NOTREECHECKS=True
Review URL: https://codereview.chromium.org/1327313002
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1332853003
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
No changes to the code, just moved around.
This will have the effect of enabling vectorized code on ARMv7.
Should be no effect on ARMv8 or x86, which would have been vectorized already.
nanobench --match mappoints changes on Nexus 5 (ARMv7):
_affine: 132 -> 95
_scale: 118 -> 47
_trans: 60 -> 37
A teaser:
We should next look at the ABCD->BADC shuffle we've noted that we need in _affine. A quick hack showed doing that optimally is another ~35% speedup on x86. Got to figure out how to do it best on ARM though: that same quick hack was a 2x slowdown there. Good reason to resurrect that SkNx_shuffle() CL!
(I believe the answers are vrev64q_f32(v) and _mm_shuffle_ps(v,v, _MM_SHUFFLE(2,3,0,1), but we should probably find out in another CL.)
BUG=skia:4117
Review URL: https://codereview.chromium.org/1320673014
|
|
|
|
| |
Review URL: https://codereview.chromium.org/1330353006
|
|
|
|
|
|
| |
R=joshualitt@google.com
Review URL: https://codereview.chromium.org/1333943002
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1337513002
|
|
|
|
| |
Review URL: https://codereview.chromium.org/1315513008
|
|
|
|
|
|
|
|
|
|
| |
This was a pre-SkOpts attempt that we can bring under its wing now.
This should be a perf no-op, deo volente.
BUG=skia:4117
Review URL: https://codereview.chromium.org/1314863006
|
|
|
|
|
|
| |
R=egdaniel@google.com
Review URL: https://codereview.chromium.org/1334753002
|
|
|
|
|
|
|
|
|
|
| |
We use a temp FB and stencil buffer to test different stencil formats with
a given pixel config. We then keep a map from pixel config to desired stencil
format.
BUG=skia:
Review URL: https://codereview.chromium.org/1317443004
|
|
|
|
|
|
|
|
|
|
|
|
| |
raster threads accessing the same GrContext to read back GPU input data
BUG=chromium:524717
TBR=reed@google.com
Committed: https://skia.googlesource.com/skia/+/eb662bc407cec0585a821946fef123102cae64db
Review URL: https://codereview.chromium.org/1329313002
|
|
|
|
|
|
|
|
|
|
| |
created (old behavior), or a pipeline with a single proc tree (added behavior).
Had to move GrComposeEffect class definition from SkComposeShader.cpp to SkComposeShader.h so that GLProgramsTest can call GrComposeEffect::Create()
BUG=skia:4182
Review URL: https://codereview.chromium.org/1314923002
|
|
|
|
|
|
|
|
| |
The check at line 1259 was failing in Valgrind, and in
the default constructor subrun.fMaskFormat is never inited.
Also set some other variables to avoid future problems.
Review URL: https://codereview.chromium.org/1315773005
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
multiple CPU raster threads accessin… (patchset #1 id:1 of https://codereview.chromium.org/1329313002/ )
Reason for revert:
breaking the bots
Original issue's description:
> Add a mutex to GrContext::readSurfacePixels to protect against multiple CPU raster threads accessing the same GrContext to read back GPU input data
>
> BUG=chromium:524717
>
> TBR=reed@google.com
>
> Committed: https://skia.googlesource.com/skia/+/eb662bc407cec0585a821946fef123102cae64db
TBR=reed@google.com,reed@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:524717
Review URL: https://codereview.chromium.org/1334603002
|
|
|
|
|
|
|
|
|
|
| |
raster threads accessing the same GrContext to read back GPU input data
BUG=chromium:524717
TBR=reed@google.com
Review URL: https://codereview.chromium.org/1329313002
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1327193003
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1324503005
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
of https://codereview.chromium.org/1334543002/ )
Reason for revert:
DEPS should be fixed now
Original issue's description:
> Revert of more zero-length changes for svg compatibility (patchset #6 id:100001 of https://codereview.chromium.org/1330623003/ )
>
> Reason for revert:
> breaks DEPS
>
> Original issue's description:
> > more zero-length changes for svg compatibility
> >
> > If a path contains a moveTo followed by a line or curve,
> > even if the line or curve has no length, SVG expects
> > the end caps to draw if the cap style is round or square.
> >
> > Fredrik Söderquist attached a patch to the chrome bug
> > (slightly modified here) that fixes layout test failures
> > resulting from deleting special-case code in SVG
> > dealing with zero-length path segments.
> >
> > R=reed@google.com,fs@opera.com
> > BUG=22974
> >
> > Committed: https://skia.googlesource.com/skia/+/62fb1ba1786863e545c89839b5706ad5151cec15
>
> TBR=fs@opera.com,reed@google.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=22974
>
> Committed: https://skia.googlesource.com/skia/+/5ca4fa3846067a47e88d35ace895df3ebe3ec2a5
TBR=fs@opera.com,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=22974
Review URL: https://codereview.chromium.org/1314833004
|
|
|
|
|
|
|
|
| |
processors that combine 2 things with a xfermode. This way, we can change the implementation of these factories in the future to not use GrComposeEffect if desired.
BUG=skia:4182
Review URL: https://codereview.chromium.org/1306163005
|
|
|
|
|
|
|
|
|
|
| |
TBR=reed@google.com
BUG=skia:1330
Committed: https://skia.googlesource.com/skia/+/6f2a486040cb25465990196c229feb47e668e87f
Review URL: https://codereview.chromium.org/1264103003
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
SkOncePtr<T[]> is identical to SkOncePtr<T> except we'll default to delete[]
for cleanup.
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-ASAN-Trybot
BUG=skia:
No public API changes.
TBR=reed@google.com
Review URL: https://codereview.chromium.org/1311893010
|
|
|
|
|
|
|
| |
TBR=bsalomon@google.com
BUG=skia:
Review URL: https://codereview.chromium.org/1323963003
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1315563003
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
id:100001 of https://codereview.chromium.org/1330623003/ )
Reason for revert:
breaks DEPS
Original issue's description:
> more zero-length changes for svg compatibility
>
> If a path contains a moveTo followed by a line or curve,
> even if the line or curve has no length, SVG expects
> the end caps to draw if the cap style is round or square.
>
> Fredrik Söderquist attached a patch to the chrome bug
> (slightly modified here) that fixes layout test failures
> resulting from deleting special-case code in SVG
> dealing with zero-length path segments.
>
> R=reed@google.com,fs@opera.com
> BUG=22974
>
> Committed: https://skia.googlesource.com/skia/+/62fb1ba1786863e545c89839b5706ad5151cec15
TBR=fs@opera.com,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=22974
Review URL: https://codereview.chromium.org/1334543002
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This gives SkOncePtr a non-trivial destructor that uses std::default_delete
by default. This is overrideable, as seen in SkColorTable.
SK_DECLARE_STATIC_ONCE_PTR still just leaves its pointers hanging at EOP.
BUG=skia:
No public API changes.
TBR=reed@google.com
Committed: https://skia.googlesource.com/skia/+/a1254acdb344174e761f5061c820559dab64a74c
Review URL: https://codereview.chromium.org/1322933005
|
|
|
|
|
|
| |
BUG=skia:
Review URL: https://codereview.chromium.org/1333623003
|
|
|
|
|
|
|
| |
This removes nesting of cache accesses, which avoids a possible
deadlock condition when the caches are the same.
Review URL: https://codereview.chromium.org/1310513008
|
|
|
|
|
|
|
|
|
|
| |
I'd remove it entirely but Android is calling it explicitly.
BUG=skia:4259
Committed: https://skia.googlesource.com/skia/+/925979f733fe8e70d84627147dee04d030423349
Review URL: https://codereview.chromium.org/1329853005
|