Start changing WebDOMMessageEvent for using target execution context.

Start changing WebDOMMessageEvent for using target execution context.

Message ports were using the source frame's execution context
rather than the targets when ports were transferred from a frame
to another. This seems to be leaky and also creates difficulties
when implementing message port support for android webview
and OOPIF. Use the target instead.

This is patch 1 of 3. The next CL will move chromium code to
use newly introduced method and the third CL will delete
the old methods.

BUG=473258
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=200735

[sgurun-gerrit only, 2015-08-14 23:26:42]

sgurun@chromium.org changed reviewers:
+ adamk@chromium.org

[sgurun-gerrit only, 2015-08-14 23:28:43]

On 2015/08/14 23:26:42, sgurun wrote:
> mailto:sgurun@chromium.org changed reviewers:
> + mailto:adamk@chromium.org

ptal, thanks! I simply duplicated the method at this time, since I would remove
the old one in a subsequent CL.

[adamk, 2015-08-15 00:01:15]

adamk@chromium.org changed reviewers:
+ alexmos@chromium.org, dcheng@chromium.org

[adamk, 2015-08-15 00:01:16]

+alexmos for the TODO, dcheng for public/OWNERS (and probably a better reviewer
for this change overall)

https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
File Source/web/WebDOMMessageEvent.cpp (right):

https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
Source/web/WebDOMMessageEvent.cpp:52: // TODO(alexmos): Figure out if this is
the right thing to do.
Would hate to leave this TODO in the new version. Perhaps we can get it fixed
right now?

https://codereview.chromium.org/1289033003/diff/1/public/web/WebDOMMessageEve...
File public/web/WebDOMMessageEvent.h (right):

https://codereview.chromium.org/1289033003/diff/1/public/web/WebDOMMessageEve...
public/web/WebDOMMessageEvent.h:50: BLINK_EXPORT void initMessageEvent(const
WebString& type, bool canBubble, bool cancelable, const
WebSerializedScriptValue& messageData, const WebString& origin, const WebFrame*
sourceFrame, const WebDocument& targetDocument, const WebString& lastEventId,
const WebMessagePortChannelArray& channels = WebMessagePortChannelArray());
You'll need someone in public/OWNERS to review this API change.

[alexmos, 2015-08-15 01:07:02]

https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
File Source/web/WebDOMMessageEvent.cpp (right):

https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
Source/web/WebDOMMessageEvent.cpp:52: // TODO(alexmos): Figure out if this is
the right thing to do.
On 2015/08/15 00:01:15, adamk wrote:
> Would hate to leave this TODO in the new version. Perhaps we can get it fixed
> right now?

I think we can just remove this.  dcheng@ originally added it in
https://codereview.chromium.org/683013002.  Unless Daniel meant something else,
I don't see a reason for keeping it, as this already works correctly for both
local and remote cases.

[sgurun-gerrit only, 2015-08-15 01:10:13]

On 2015/08/15 01:07:02, alexmos wrote:
>
https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
> File Source/web/WebDOMMessageEvent.cpp (right):
> 
>
https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
> Source/web/WebDOMMessageEvent.cpp:52: // TODO(alexmos): Figure out if this is
> the right thing to do.
> On 2015/08/15 00:01:15, adamk wrote:
> > Would hate to leave this TODO in the new version. Perhaps we can get it
fixed
> > right now?
> 
> I think we can just remove this.  dcheng@ originally added it in
> https://codereview.chromium.org/683013002.  Unless Daniel meant something
else,
> I don't see a reason for keeping it, as this already works correctly for both
> local and remote cases.

agreed, I don't see anything to fix here anymore. it seems to be ok.

[sgurun-gerrit only, 2015-08-15 01:12:11]

https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
File Source/web/WebDOMMessageEvent.cpp (right):

https://codereview.chromium.org/1289033003/diff/1/Source/web/WebDOMMessageEve...
Source/web/WebDOMMessageEvent.cpp:52: // TODO(alexmos): Figure out if this is
the right thing to do.
On 2015/08/15 01:07:02, alexmos wrote:
> On 2015/08/15 00:01:15, adamk wrote:
> > Would hate to leave this TODO in the new version. Perhaps we can get it
fixed
> > right now?
> 
> I think we can just remove this.  dcheng@ originally added it in
> https://codereview.chromium.org/683013002.  Unless Daniel meant something
else,
> I don't see a reason for keeping it, as this already works correctly for both
> local and remote cases.

Done.

[dcheng, 2015-08-15 05:12:30]

https://codereview.chromium.org/1289033003/diff/20001/Source/web/WebDOMMessag...
File Source/web/WebDOMMessageEvent.cpp (right):

https://codereview.chromium.org/1289033003/diff/20001/Source/web/WebDOMMessag...
Source/web/WebDOMMessageEvent.cpp:65: ASSERT(m_private.get());
Can we re-express this in terms of the new interface? Then it'll be clear
there's no difference between the old/new overloads.

https://codereview.chromium.org/1289033003/diff/20001/public/web/WebDOMMessag...
File public/web/WebDOMMessageEvent.h (right):

https://codereview.chromium.org/1289033003/diff/20001/public/web/WebDOMMessag...
public/web/WebDOMMessageEvent.h:1: /*
What's with the extra space?

https://codereview.chromium.org/1289033003/diff/20001/public/web/WebDOMMessag...
public/web/WebDOMMessageEvent.h:51: BLINK_EXPORT void initMessageEvent(const
WebString& type, bool canBubble, bool cancelable, const
WebSerializedScriptValue& messageData, const WebString& origin, const WebFrame*
sourceFrame, const WebString& lastEventId, const WebMessagePortChannelArray&
channels = WebMessagePortChannelArray());
Mark this as deprecated? Presumably, you're going to remove this overload.

[sgurun-gerrit only, 2015-08-18 15:17:31]

https://codereview.chromium.org/1289033003/diff/20001/Source/web/WebDOMMessag...
File Source/web/WebDOMMessageEvent.cpp (right):

https://codereview.chromium.org/1289033003/diff/20001/Source/web/WebDOMMessag...
Source/web/WebDOMMessageEvent.cpp:65: ASSERT(m_private.get());
On 2015/08/15 05:12:30, dcheng wrote:
> Can we re-express this in terms of the new interface? Then it'll be clear
> there's no difference between the old/new overloads.

Done.

https://codereview.chromium.org/1289033003/diff/20001/public/web/WebDOMMessag...
File public/web/WebDOMMessageEvent.h (right):

https://codereview.chromium.org/1289033003/diff/20001/public/web/WebDOMMessag...
public/web/WebDOMMessageEvent.h:1: /*
On 2015/08/15 05:12:30, dcheng wrote:
> What's with the extra space?

unintended, fixed.

https://codereview.chromium.org/1289033003/diff/20001/public/web/WebDOMMessag...
public/web/WebDOMMessageEvent.h:51: BLINK_EXPORT void initMessageEvent(const
WebString& type, bool canBubble, bool cancelable, const
WebSerializedScriptValue& messageData, const WebString& origin, const WebFrame*
sourceFrame, const WebString& lastEventId, const WebMessagePortChannelArray&
channels = WebMessagePortChannelArray());
On 2015/08/15 05:12:30, dcheng wrote:
> Mark this as deprecated? Presumably, you're going to remove this overload.

Done.

[dcheng, 2015-08-18 16:46:54]

lgtm

https://codereview.chromium.org/1289033003/diff/40001/public/web/WebDOMMessag...
File public/web/WebDOMMessageEvent.h (right):

https://codereview.chromium.org/1289033003/diff/40001/public/web/WebDOMMessag...
public/web/WebDOMMessageEvent.h:50: BLINK_EXPORT void initMessageEvent(const
WebString& type, bool canBubble, bool cancelable, const
WebSerializedScriptValue& messageData, const WebString& origin, const WebFrame*
sourceFrame, const WebDocument& targetDocument, const WebString& lastEventId,
const WebMessagePortChannelArray& channels = WebMessagePortChannelArray());
Optional for this patch, but it'd be nice to introduce a class comment on how to
use this class or what it's for? We can stick this into the cleanup patch to
unblock the Chrome-side work.

[sgurun-gerrit only, 2015-08-18 16:49:07]

https://codereview.chromium.org/1289033003/diff/40001/public/web/WebDOMMessag...
File public/web/WebDOMMessageEvent.h (right):

https://codereview.chromium.org/1289033003/diff/40001/public/web/WebDOMMessag...
public/web/WebDOMMessageEvent.h:50: BLINK_EXPORT void initMessageEvent(const
WebString& type, bool canBubble, bool cancelable, const
WebSerializedScriptValue& messageData, const WebString& origin, const WebFrame*
sourceFrame, const WebDocument& targetDocument, const WebString& lastEventId,
const WebMessagePortChannelArray& channels = WebMessagePortChannelArray());
On 2015/08/18 16:46:54, dcheng wrote:
> Optional for this patch, but it'd be nice to introduce a class comment on how
to
> use this class or what it's for? We can stick this into the cleanup patch to
> unblock the Chrome-side work.

will add it as part of cleanup.

[sgurun-gerrit only, 2015-08-18 16:49:12]

The CQ bit was checked by sgurun@chromium.org

[commit-bot: I haz the power, 2015-08-18 16:49:20]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1289033003/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1289033003/40001

[commit-bot: I haz the power, 2015-08-18 17:57:34]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=200735