Implement convertToBlob() in OffscreenCanvas

Implement convertToBlob() in OffscreenCanvas

This CL implements convertToBlob on OffscreenCanvas for both
worker and main threads.

The OffscreenCanvasRenderingContext2D cases are complete.
For the Webgl rendering context, currently convertToBlob()
can only work for PNG image type.

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel

Committed: https://crrev.com/c9e12fb8f608b3d0e204b167459bf9ea999e068c
Cr-Commit-Position: refs/heads/master@{#427153}

[xlai (Olivia), 2016-10-14 21:40:21]

Description was changed from

==========
[WIP] Implement convertToBlob() in OffscreenCanvas

[not ready for review yet]
COMMIT=false

BUG=563830
==========

to

==========
[WIP] Implement convertToBlob() in OffscreenCanvas

[not ready for review yet]
COMMIT=false

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

[xlai (Olivia), 2016-10-19 15:46:32]

Patchset #3 (id:40001) has been deleted

[xlai (Olivia), 2016-10-19 18:46:31]

Patchset #6 (id:120001) has been deleted

[xlai (Olivia), 2016-10-19 18:50:24]

Description was changed from

==========
[WIP] Implement convertToBlob() in OffscreenCanvas

[not ready for review yet]
COMMIT=false

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

to

==========
Implement convertToBlob() in OffscreenCanvas

This CL implements convertToBlob on OffscreenCanvas for both
worker and main threads.

The OffscreenCanvasRenderingContext2D cases are complete.
For the Webgl rendering context, currently convertToBlob()
can only work for PNG image type.

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

[xlai (Olivia), 2016-10-19 18:51:05]

xlai@chromium.org changed reviewers:
+ junov@chromium.org

[xlai (Olivia), 2016-10-19 18:51:07]

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Lay...
File third_party/WebKit/LayoutTests/VirtualTestSuites (right):

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/VirtualTestSuites:16: "args":
["--enable-threaded-compositing"]
I just realize that I still need this flag. The reason why canvas.toBlob no
longer needs it is because it implements the enforcing mechanism when idle task
is not run.

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Lay...
File
third_party/WebKit/LayoutTests/fast/idleToBlob/OffscreenCanvas-convertToBlob-2d-main-expected.html
(right):

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/fast/idleToBlob/OffscreenCanvas-convertToBlob-2d-main-expected.html:29:
var canvas = document.createElement("canvas");
This file is exactly same as
OffscreenCanvas-convertToBlob-2d-worker-expected.html.

They both use canvas.toBlob as a reference image test.

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/fast/idleToBlob/OffscreenCanvas-convertToBlob-2d-main-expected.html:50:
}, "image/jpeg", 1.0);
As written in my email, the default quality of jpeg/webp in canvas.toBlob is
different from the default quality in OffscreenCanvas.convertToBlob. Whether or
not this should be changed to ensure consistency between canvas.toBlob and
OffscreenCanvas.convertToBlob is still subject to discussion. For now, I'm
putting the value explicitly to ensure tests run correctly.

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp
(left):

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:200:
ASSERT(isMainThread());
All these assertion of isMainThread() need to be removed because the function
can now happen on worker thread.

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Sou...
File
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp
(right):

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:38:
ImageData* OffscreenCanvasModules::toImageData(OffscreenCanvas& offscreenCanvas,
This function requires a higher-level view of OffscreenCanvasRenderingContext2D,
WebglRenderingContextBase and so it is placed in modules/

https://codereview.chromium.org/2420203002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:46:
// images. See crbug.com/657531.
I might be missing something here in converting webgl rendering context to
imagedata, such that only toBlob(png) works and not for jpeg and webp. Will do
it in next patch.

[Justin Novosad, 2016-10-19 19:59:33]

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Lay...
File third_party/WebKit/LayoutTests/virtual/threaded/fast/idleToBlob/README.txt
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/virtual/threaded/fast/idleToBlob/README.txt:1: #
This suite runs the tests in virtual/threaded/fast/canvas with
--enable-threaded-compositing.
I think you mean fast/idleToBlob

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:91: //
TODO: Add histograms for OffscreenCanvas.convertToBlob.
Preferred syntax is // TODO(crbug...):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:115: //
TODO: Add histograms for OffscreenCanvas.convertToBlob.
TODO(crbug...):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:199:
m_parentFrameTaskRunner = ParentFrameTaskRunners::create(document->frame());
DCHECK(m_callback);

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:201: if
(m_scriptPromiseResolver) {
Not necessary.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:215:
m_callback.clear();
How about the promise resolver?

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:26: static
CanvasAsyncBlobCreator* create(
Since the callback and the promise resolver are mutually exclusive, it would be
clearer if we had two separate flavors of create(). The constructor is private
so no need to duplicate it. The main thing is to avoid the use of nullptr
literals at the external call sites, which is bad for readability.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:49:
IdleTaskSwitchedToMainThreadTask,
This name is obsolete (not necessarily main thread). Even before the name
"MainThreadTask" was not the best description. Perhaps we should call this
"IdleTaskSwitchedToImmediateTask"

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:80: virtual
void postDelayedTaskToMainThread(const WebTraceLocation&,
This also needs to be renamed because it is longer necessarily the Main thread.
Perhaps we could call it the ScriptExecutionThread, or maybe you have a better
idea?

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:110:
ToBlobFunctionType m_functionType;
This additional state is not necessary, you could just have this: 
bool usingPromise() {
  DCHECK(!m_callback ^ !m_scriptPromiseResolver);
  DCHECK(!m_document == !m_callback);
  return m_scriptPromiseResolver;
}

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:122:
encodeRowsPngOnMainThread();  // Similar to idleEncodeRowsPng without deadline
Here too, not necessarily MainThread

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:126: void
encodeRowsJpegOnMainThread();  // Similar to idleEncodeRowsJpeg without
and here

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:44:
if (offscreenCanvas.renderingContext()->is3d()) {
This is not a great pattern.  You should make just call
offscreenCanvas.renderingContext()->getImage() to acquire a snapshot.  This way,
this function does not need to know about specific context types and then you
can put it directly in OffscreenCanvas instead of OffscreenCanvasModules.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:86:
ScriptPromise OffscreenCanvasModules::convertToBlob(
Move this method into core as well, once toImageData is in Core.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h:162:
DrawingBuffer* drawingBuffer() const;
This no longer needs to be public once other comments are addressed.

[xlai (Olivia), 2016-10-20 15:54:11]

Patchset #7 (id:160001) has been deleted

[xlai (Olivia), 2016-10-20 15:54:24]

Patchset #7 (id:180001) has been deleted

[xlai (Olivia), 2016-10-20 15:58:54]

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:91: //
TODO: Add histograms for OffscreenCanvas.convertToBlob.
On 2016/10/19 19:59:32, Justin Novosad wrote:
> Preferred syntax is // TODO(crbug...):

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:115: //
TODO: Add histograms for OffscreenCanvas.convertToBlob.
On 2016/10/19 19:59:32, Justin Novosad wrote:
> TODO(crbug...):

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:199:
m_parentFrameTaskRunner = ParentFrameTaskRunners::create(document->frame());
On 2016/10/19 19:59:32, Justin Novosad wrote:
> DCHECK(m_callback);

If I do a DCHECK here, the unit test will fail...unit test assumes no Callback.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.cpp:215:
m_callback.clear();
On 2016/10/19 19:59:32, Justin Novosad wrote:
> How about the promise resolver?

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:26: static
CanvasAsyncBlobCreator* create(
On 2016/10/19 19:59:32, Justin Novosad wrote:
> Since the callback and the promise resolver are mutually exclusive, it would
be
> clearer if we had two separate flavors of create(). The constructor is private
> so no need to duplicate it. The main thing is to avoid the use of nullptr
> literals at the external call sites, which is bad for readability.

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:49:
IdleTaskSwitchedToMainThreadTask,
On 2016/10/19 19:59:32, Justin Novosad wrote:
> This name is obsolete (not necessarily main thread). Even before the name
> "MainThreadTask" was not the best description. Perhaps we should call this
> "IdleTaskSwitchedToImmediateTask"

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:80: virtual
void postDelayedTaskToMainThread(const WebTraceLocation&,
On 2016/10/19 19:59:33, Justin Novosad wrote:
> This also needs to be renamed because it is longer necessarily the Main
thread.
> Perhaps we could call it the ScriptExecutionThread, or maybe you have a better
> idea?

This function is related to the mechanism of "enforcing toBlob to happen if idle
task is waiting for too long" which has not been implemented in this CL yet.
Maybe we
think of it in the next CL?

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:110:
ToBlobFunctionType m_functionType;
On 2016/10/19 19:59:33, Justin Novosad wrote:
> This additional state is not necessary, you could just have this: 
> bool usingPromise() {
>   DCHECK(!m_callback ^ !m_scriptPromiseResolver);
>   DCHECK(!m_document == !m_callback);
>   return m_scriptPromiseResolver;
> }

I think this additional state is necessary for the UMA histogram. It needs
to distinguish between canvas.toBlob and OffscreenCanvas.convertToBlob with
an explicit enum type.

And also for the unit test.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasAsyncBlobCreator.h:122:
encodeRowsPngOnMainThread();  // Similar to idleEncodeRowsPng without deadline
On 2016/10/19 19:59:33, Justin Novosad wrote:
> Here too, not necessarily MainThread

This one is not executed in the OffscreenCanvas case in this CL yet.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:44:
if (offscreenCanvas.renderingContext()->is3d()) {
On 2016/10/19 19:59:33, Justin Novosad wrote:
> This is not a great pattern.  You should make just call
> offscreenCanvas.renderingContext()->getImage() to acquire a snapshot.  This
way,
> this function does not need to know about specific context types and then you
> can put it directly in OffscreenCanvas instead of OffscreenCanvasModules.

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:86:
ScriptPromise OffscreenCanvasModules::convertToBlob(
On 2016/10/19 19:59:33, Justin Novosad wrote:
> Move this method into core as well, once toImageData is in Core.

Done.

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h
(right):

https://codereview.chromium.org/2420203002/diff/140001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h:162:
DrawingBuffer* drawingBuffer() const;
On 2016/10/19 19:59:33, Justin Novosad wrote:
> This no longer needs to be public once other comments are addressed.

Done.

[Justin Novosad, 2016-10-20 16:18:54]

lgtm with follow-up request

https://codereview.chromium.org/2420203002/diff/200001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasRenderingContext.h
(right):

https://codereview.chromium.org/2420203002/diff/200001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasRenderingContext.h:91: virtual
ImageData* toImageData(SnapshotReason reason) const {
Having this is a good idea. Please create a crbug to refactor around this in a
follow-up CL:
1. Make this method pure virtual (and implement in in all context classes)
2. Augment this method so that it can take the same arguments as
HTMLCanvasElement::toImageData
3. Add a convertToBlob entry to the "SnapshotReason" enum, and add it to the
aprropriate UMA histograms
4. Make HTMLCanvasElement call this method instead using its own toImageData
5. Delete HTMLCanvasElement::toImageData from the code

[xlai (Olivia), 2016-10-20 16:26:26]

xlai@chromium.org changed reviewers:
+ haraken@chromium.org, kbr@chromium.org

[xlai (Olivia), 2016-10-20 16:26:27]

Sending to more reviewers.

haraken@: Please review WebKit/Source/bindings/. ImageEncodeOptions is added
based on newly modified spec.

kbr@: Please review WebglRenderingContextBase.

Thanks!

https://codereview.chromium.org/2420203002/diff/200001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/canvas/CanvasRenderingContext.h
(right):

https://codereview.chromium.org/2420203002/diff/200001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/canvas/CanvasRenderingContext.h:91: virtual
ImageData* toImageData(SnapshotReason reason) const {
On 2016/10/20 16:18:53, Justin Novosad wrote:
> Having this is a good idea. Please create a crbug to refactor around this in a
> follow-up CL:
> 1. Make this method pure virtual (and implement in in all context classes)
> 2. Augment this method so that it can take the same arguments as
> HTMLCanvasElement::toImageData
> 3. Add a convertToBlob entry to the "SnapshotReason" enum, and add it to the
> aprropriate UMA histograms
> 4. Make HTMLCanvasElement call this method instead using its own toImageData
> 5. Delete HTMLCanvasElement::toImageData from the code

Thanks! Creating crbug.com/657870 for record.

[haraken, 2016-10-20 16:29:07]

bindings/ LGTM

[Ken Russell (switch to Gerrit), 2016-10-22 00:57:08]

webgl lgtm

[xlai (Olivia), 2016-10-24 15:09:04]

The CQ bit was checked by xlai@chromium.org

[xlai (Olivia), 2016-10-24 15:09:05]

The patchset sent to the CQ was uploaded after l-g-t-m from
haraken@chromium.org, junov@chromium.org
Link to the patchset: https://codereview.chromium.org/2420203002/#ps220001
(title: "rebase and fix global-interface-listing-service-worker-expected.txt")

[commit-bot: I haz the power, 2016-10-24 15:09:33]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-24 16:17:37]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-24 16:17:38]

Try jobs failed on following builders:
  linux_chromium_chromeos_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[xlai (Olivia), 2016-10-24 19:15:19]

The CQ bit was checked by xlai@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-24 19:16:14]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-24 20:53:29]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-24 20:53:30]

Dry run: This issue passed the CQ dry run.

[xlai (Olivia), 2016-10-24 20:58:31]

The CQ bit was checked by xlai@chromium.org

[commit-bot: I haz the power, 2016-10-24 20:59:14]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-24 21:20:21]

Description was changed from

==========
Implement convertToBlob() in OffscreenCanvas

This CL implements convertToBlob on OffscreenCanvas for both
worker and main threads.

The OffscreenCanvasRenderingContext2D cases are complete.
For the Webgl rendering context, currently convertToBlob()
can only work for PNG image type.

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

to

==========
Implement convertToBlob() in OffscreenCanvas

This CL implements convertToBlob on OffscreenCanvas for both
worker and main threads.

The OffscreenCanvasRenderingContext2D cases are complete.
For the Webgl rendering context, currently convertToBlob()
can only work for PNG image type.

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

[commit-bot: I haz the power, 2016-10-24 21:20:23]

Committed patchset #8 (id:220001)

[commit-bot: I haz the power, 2016-10-24 21:26:26]

Description was changed from

==========
Implement convertToBlob() in OffscreenCanvas

This CL implements convertToBlob on OffscreenCanvas for both
worker and main threads.

The OffscreenCanvasRenderingContext2D cases are complete.
For the Webgl rendering context, currently convertToBlob()
can only work for PNG image type.

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

to

==========
Implement convertToBlob() in OffscreenCanvas

This CL implements convertToBlob on OffscreenCanvas for both
worker and main threads.

The OffscreenCanvasRenderingContext2D cases are complete.
For the Webgl rendering context, currently convertToBlob()
can only work for PNG image type.

BUG=563830
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel

Committed: https://crrev.com/c9e12fb8f608b3d0e204b167459bf9ea999e068c
Cr-Commit-Position: refs/heads/master@{#427153}
==========

[commit-bot: I haz the power, 2016-10-24 21:26:27]

Patchset 8 (id:??) landed as
https://crrev.com/c9e12fb8f608b3d0e204b167459bf9ea999e068c
Cr-Commit-Position: refs/heads/master@{#427153}