Make OffscreenCanvas an EventTarget

Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be
upstreamed to khronos group on github later on once the spec for
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759, 630515
CQ_INCLUDE_TRYBOTS=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

Committed: https://crrev.com/8372014fc16a378fc7452164db9b3b89fbd62909
Cr-Commit-Position: refs/heads/master@{#432478}

[xidachen, 2016-11-08 03:04:05]

Description was changed from

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we need the
OffscreenCanvasRenderingContext2d and WebGLRenderingContextBase keep a
Member<ExecutionContext>. We do not keep this Member in the other canvas
rendering contexts because this is very specific to OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. For 2D, the tests will follow later on.
The layout tests here should be upstreamed to khronos group on github
later on once the spec for OffscreenCanvas is landed.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270
==========

to

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we need the
OffscreenCanvasRenderingContext2d and WebGLRenderingContextBase keep a
Member<ExecutionContext>. We do not keep this Member in the other canvas
rendering contexts because this is very specific to OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. For 2D, the tests will follow later on.
The layout tests here should be upstreamed to khronos group on github
later on once the spec for OffscreenCanvas is landed.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270
CQ_INCLUDE_TRYBOTS=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
==========

[xidachen, 2016-11-13 01:45:45]

Description was changed from

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we need the
OffscreenCanvasRenderingContext2d and WebGLRenderingContextBase keep a
Member<ExecutionContext>. We do not keep this Member in the other canvas
rendering contexts because this is very specific to OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. For 2D, the tests will follow later on.
The layout tests here should be upstreamed to khronos group on github
later on once the spec for OffscreenCanvas is landed.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270
CQ_INCLUDE_TRYBOTS=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
==========

to

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we need the
OffscreenCanvasRenderingContext2d and WebGLRenderingContextBase keep a
Member<ExecutionContext>. We do not keep this Member in the other canvas
rendering contexts because this is very specific to OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759
CQ_INCLUDE_TRYBOTS=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
==========

[xidachen, 2016-11-13 01:45:45]

xidachen@chromium.org changed reviewers:
+ junov@chromium.org, kbr@chromium.org, zmo@chromium.org

[xidachen, 2016-11-13 01:53:41]

PTAL

https://codereview.chromium.org/2490443002/diff/20001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/fast/canvas/webgl/offscreenCanvas-context-lost-restored.html
(right):

https://codereview.chromium.org/2490443002/diff/20001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/fast/canvas/webgl/offscreenCanvas-context-lost-restored.html:78:
//    assert_false(gl.isContextLost());
kbr@, zmo@: could you please provide some suggestions? This test fails at this
statement, I am not sure why.

[Justin Novosad, 2016-11-14 16:41:00]

https://codereview.chromium.org/2490443002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h
(right):

https://codereview.chromium.org/2490443002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h:1599:
Member<ExecutionContext> m_executionContext;
Storing this in the rendering context is unnecessary complexity IMHO. You should
just store is in OffscreenCanvas.

[xidachen, 2016-11-14 18:45:42]

Description was changed from

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we need the
OffscreenCanvasRenderingContext2d and WebGLRenderingContextBase keep a
Member<ExecutionContext>. We do not keep this Member in the other canvas
rendering contexts because this is very specific to OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759
CQ_INCLUDE_TRYBOTS=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
==========

to

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759
CQ_INCLUDE_TRYBOTS=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
==========

[xidachen, 2016-11-14 18:48:12]

https://codereview.chromium.org/2490443002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h
(right):

https://codereview.chromium.org/2490443002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.h:1599:
Member<ExecutionContext> m_executionContext;
On 2016/11/14 16:41:00, Justin Novosad wrote:
> Storing this in the rendering context is unnecessary complexity IMHO. You
should
> just store is in OffscreenCanvas.

Done. It does make things much simpler.

[Justin Novosad, 2016-11-14 19:45:08]

lgtm with nits

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/offscreencanvas/OffscreenCanvas.h (right):

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/offscreencanvas/OffscreenCanvas.h:125:
Member<ExecutionContext> m_executionContext;
Nit: This reference is not need for lifetime management purposes, so you could
make this a WeakMember, then you won't have to trace into it.

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Sour...
File
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp
(right):

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/offscreencanvas/OffscreenCanvasModules.cpp:26:
offscreenCanvas.setExecutionContext(scriptState->getExecutionContext());
Add a comment: OffscreenCanvas cannot be transferred after getContext, so this
execution context will always be the right one from here on.

[Ken Russell (switch to Gerrit), 2016-11-14 19:54:16]

WebGL changes lgtm. I defer to junov@'s review of the rest.

[xidachen, 2016-11-14 19:56:38]

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/fast/canvas/webgl/offscreenCanvas-context-lost-restored.html
(right):

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/fast/canvas/webgl/offscreenCanvas-context-lost-restored.html:78:
//    assert_false(gl.isContextLost());
Ken, this test fails here if I un-comment this line. It appears that the context
is still lost after WEBGL_lose_context.restoreContext() has been called. Any
suggestion why this happens?

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/offscreencanvas/OffscreenCanvas.h (right):

https://codereview.chromium.org/2490443002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/offscreencanvas/OffscreenCanvas.h:125:
Member<ExecutionContext> m_executionContext;
On 2016/11/14 19:45:08, Justin Novosad wrote:
> Nit: This reference is not need for lifetime management purposes, so you could
> make this a WeakMember, then you won't have to trace into it.

Changed to WeakMember. But still need to trace it, otherwise compiler complains:
[blink-gc] Class 'OffscreenCanvas' has untraced fields that require tracing.

[xidachen, 2016-11-15 17:02:42]

Description was changed from

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759
CQ_INCLUDE_TRYBOTS=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
==========

to

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759, 630515
CQ_INCLUDE_TRYBOTS=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
==========

[xidachen, 2016-11-15 17:04:11]

kbr@, junov@: could either one of you please take a look at the webgl/ changes?
I made some more changes before the context restoration event was not handled
properly in previous PS. It should be fine now indicated by all the passing
tests.

[xidachen, 2016-11-15 17:12:57]

On 2016/11/15 17:04:11, xidachen wrote:
> kbr@, junov@: could either one of you please take a look at the webgl/
changes?
> I made some more changes before the context restoration event was not handled
> properly in previous PS. It should be fine now indicated by all the passing
> tests.

Awww, I just found my previous statement very confusing. PS#6 should work, I put
some more changes under webgl/ to handle context restoration. These changes
should be reviewed.

[Justin Novosad, 2016-11-15 19:54:33]

https://codereview.chromium.org/2490443002/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp
(right):

https://codereview.chromium.org/2490443002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp:7365:
DrawingBuffer::DisallowChromiumImage);
Why?  Perhaps add a comment.

[xidachen, 2016-11-15 20:14:19]

https://codereview.chromium.org/2490443002/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp
(right):

https://codereview.chromium.org/2490443002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp:7365:
DrawingBuffer::DisallowChromiumImage);
On 2016/11/15 19:54:33, Justin Novosad wrote:
> Why?  Perhaps add a comment.

Comment added in the new patch.

Long story short: at some point the WebGL's transferToImageBitmap gives black
regardless what I have drawn. I freaked out and filed this bug:
crbug.com/649668. People offered great help and we made the change using
DisallowChromiumImage. Detailed discussions can be found in crbug.com/649668 :)

[Justin Novosad, 2016-11-15 20:22:24]

lgtm

[xidachen, 2016-11-15 20:22:36]

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

[commit-bot: I haz the power, 2016-11-15 20:23:10]

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

[xidachen, 2016-11-15 20:29:29]

The CQ bit was unchecked by xidachen@chromium.org

[Ken Russell (switch to Gerrit), 2016-11-15 21:33:45]

Still lgtm

[xidachen, 2016-11-15 23:35:19]

The CQ bit was checked by xidachen@chromium.org

[commit-bot: I haz the power, 2016-11-15 23:36:11]

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

[xidachen, 2016-11-16 00:53:41]

The CQ bit was unchecked by xidachen@chromium.org

[xidachen, 2016-11-16 02:16:35]

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

[commit-bot: I haz the power, 2016-11-16 02:17:02]

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

[xidachen, 2016-11-16 02:23:34]

The CQ bit was unchecked by xidachen@chromium.org

[xidachen, 2016-11-16 02:23:44]

The CQ bit was checked by xidachen@chromium.org

[xidachen, 2016-11-16 02:23:44]

The patchset sent to the CQ was uploaded after l-g-t-m from kbr@chromium.org,
junov@chromium.org
Link to the patchset: https://codereview.chromium.org/2490443002/#ps140001
(title: "pass API layout tests")

[commit-bot: I haz the power, 2016-11-16 02:24:12]

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

[xidachen, 2016-11-16 03:18:08]

The CQ bit was unchecked by xidachen@chromium.org

[xidachen, 2016-11-16 13:00:20]

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

[commit-bot: I haz the power, 2016-11-16 13:00:39]

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

[xidachen, 2016-11-16 14:15:59]

The CQ bit was unchecked by xidachen@chromium.org

[xidachen, 2016-11-16 14:16:07]

The CQ bit was checked by xidachen@chromium.org

[commit-bot: I haz the power, 2016-11-16 14:16:34]

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, 2016-11-16 14:20:42]

Description was changed from

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759, 630515
CQ_INCLUDE_TRYBOTS=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
==========

to

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759, 630515
CQ_INCLUDE_TRYBOTS=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
==========

[commit-bot: I haz the power, 2016-11-16 14:20:44]

Committed patchset #8 (id:140001)

[commit-bot: I haz the power, 2016-11-16 14:23:59]

Description was changed from

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be 
upstreamed to khronos group on github later on once the spec for 
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759, 630515
CQ_INCLUDE_TRYBOTS=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
==========

to

==========
Make OffscreenCanvas an EventTarget

Right now OffscreenCanvas is not an EventTarget, which means that it
cannot listen to any event such as webgl context lost and restoration.

This CL makes OffscreenCanvas an EventTarget. In order to override the
pure virtual function getExecutionContext in EventTarget, we keep a
Member<ExecutionContext> in OffscreenCanvas.

We also added some layout tests here. In this CL, we only tests webgl
context lost and restoration. The layout tests here should be
upstreamed to khronos group on github later on once the spec for
OffscreenCanvas is landed. Also, notice that the context lost test in
worker verifies that all WebGL's API works in a worker.

Note that the tests we have here is pretty much copied from the webgl's
conformance tests.

BUG=655270, 610759, 630515
CQ_INCLUDE_TRYBOTS=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

Committed: https://crrev.com/8372014fc16a378fc7452164db9b3b89fbd62909
Cr-Commit-Position: refs/heads/master@{#432478}
==========

[commit-bot: I haz the power, 2016-11-16 14:24:00]

Patchset 8 (id:??) landed as
https://crrev.com/8372014fc16a378fc7452164db9b3b89fbd62909
Cr-Commit-Position: refs/heads/master@{#432478}