Make SpeechRecognitionController per frame instead of per page.

Make SpeechRecognitionController per frame instead of per page.

This is still using the WebViewClient::speechRecognizer() if
WebFrameClient::speechRecognizer() returns null. After part 2
lands, part 3 will remove it.

Part 1: <this>
Part 2: https://codereview.chromium.org/636863003/
Part 3: https://codereview.chromium.org/752303003/

BUG=390749
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=186174

[mlamouri (slow - plz ping), 2014-11-10 20:04:48]

mlamouri@chromium.org changed reviewers:
+ dcheng@chromium.org, mkwst@chromium.org

[mlamouri (slow - plz ping), 2014-11-10 20:04:48]

mlamouri@chromium.org changed required reviewers:
+ mkwst@chromium.org

[Mike West, 2014-11-11 05:50:48]

LGTM.

https://codereview.chromium.org/636863002/diff/20001/Source/web/WebLocalFrame...
File Source/web/WebLocalFrameImpl.cpp (right):

https://codereview.chromium.org/636863002/diff/20001/Source/web/WebLocalFrame...
Source/web/WebLocalFrameImpl.cpp:1597:
SpeechRecognitionController::provideTo(*m_frame,
SpeechRecognitionClientProxy::create(m_client ? m_client->speechRecognizer() :
0));
s/0/nullptr/g

You could fix this whole method, if you like. :)

[dcheng, 2014-11-11 08:09:57]

dcheng@chromium.org changed required reviewers:
- mkwst@chromium.org

[dcheng, 2014-11-11 08:09:58]

https://codereview.chromium.org/636863002/diff/20001/Source/modules/speech/Sp...
File Source/modules/speech/SpeechRecognitionController.h (right):

https://codereview.chromium.org/636863002/diff/20001/Source/modules/speech/Sp...
Source/modules/speech/SpeechRecognitionController.h:39: public
WillBeHeapSupplement<LocalFrame> {
Alignment. I think git cl format mostly handles Blink style so you can just try
using that to fix this.

https://codereview.chromium.org/636863002/diff/20001/public/web/WebViewClient.h
File public/web/WebViewClient.h (left):

https://codereview.chromium.org/636863002/diff/20001/public/web/WebViewClient...
public/web/WebViewClient.h:242: // Speech
--------------------------------------------------------------
These sort of things usually require a 3-way patch. Will doing this all in one
patch break any layout tests?

[mlamouri (slow - plz ping), 2014-11-28 10:43:05]

https://codereview.chromium.org/636863002/diff/20001/Source/modules/speech/Sp...
File Source/modules/speech/SpeechRecognitionController.h (right):

https://codereview.chromium.org/636863002/diff/20001/Source/modules/speech/Sp...
Source/modules/speech/SpeechRecognitionController.h:39: public
WillBeHeapSupplement<LocalFrame> {
On 2014/11/11 at 08:09:58, dcheng wrote:
> Alignment. I think git cl format mostly handles Blink style so you can just
try using that to fix this.

Done.

https://codereview.chromium.org/636863002/diff/20001/Source/web/WebLocalFrame...
File Source/web/WebLocalFrameImpl.cpp (right):

https://codereview.chromium.org/636863002/diff/20001/Source/web/WebLocalFrame...
Source/web/WebLocalFrameImpl.cpp:1597:
SpeechRecognitionController::provideTo(*m_frame,
SpeechRecognitionClientProxy::create(m_client ? m_client->speechRecognizer() :
0));
On 2014/11/11 at 05:50:47, Mike West (OOO) wrote:
> s/0/nullptr/g
> 
> You could fix this whole method, if you like. :)

Done.

https://codereview.chromium.org/636863002/diff/20001/public/web/WebViewClient.h
File public/web/WebViewClient.h (left):

https://codereview.chromium.org/636863002/diff/20001/public/web/WebViewClient...
public/web/WebViewClient.h:242: // Speech
--------------------------------------------------------------
On 2014/11/11 at 08:09:58, dcheng wrote:
> These sort of things usually require a 3-way patch. Will doing this all in one
patch break any layout tests?

That's correct. I use this patch with a chromium patch. I'm trying to figure out
how to handle the three-way patch. Do you have any suggestion?

[mlamouri (slow - plz ping), 2014-11-28 11:07:15]

Mike, could you PTAL. I have changed the CL slightly to make sure it can land.

Basically, I keep the WebViewClient::speechRecognizer() and I use that
SpeechRecognizerClient() if it is available and the one on the WebFrameClient
isn't. Then Chromium can implement WebFrameClient::speechRecognizer() which will
enables the other path. At which point I will be able to entirely remove the
WebViewClient::speechRecognizer() method.

[Mike West, 2014-11-28 11:15:12]

Still LGTM. Sorry I didn't note this in the initial review; I should have been
more careful.

Do you already have the removal CL up? It's nice to be able to list all the CLs
in the description, similar to https://codereview.chromium.org/736233004/.

[mlamouri (slow - plz ping), 2014-11-28 11:23:28]

On 2014/11/28 at 11:15:12, mkwst wrote:
> Still LGTM. Sorry I didn't note this in the initial review; I should have been
more careful.
> 
> Do you already have the removal CL up? It's nice to be able to list all the
CLs in the description, similar to https://codereview.chromium.org/736233004/.

Done. Thanks :)

[mlamouri (slow - plz ping), 2014-11-28 11:23:35]

The CQ bit was checked by mlamouri@chromium.org

[commit-bot: I haz the power, 2014-11-28 11:24:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/636863002/40001

[commit-bot: I haz the power, 2014-11-28 12:10:20]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=186174

[mlamouri (slow - plz ping), 2015-01-22 10:39:17]

A revert of this CL (patchset #3 id:40001) has been created in
https://codereview.chromium.org/863213002/ by mlamouri@chromium.org.

The reason for reverting is: https://crbug.com/449739.