Bug 98149 – MediaStream API: RTCPeerConnection should send down its handler via the FrameLoaderClient directly after creation.

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[Tommy Widenflycht]

The chromium implementation needs to know which Frame created a PeerConnection so that the right housekeeping can take place correctly.

[Tommy Widenflycht]

Created attachment 166673 [details]
Patch

[Tommy Widenflycht]

Trying this way instead.

[WebKit Review Bot]

Please wait for approval from [e-mail], [e-mail], [e-mail], [e-mail] or [e-mail] before submitting, as this patch contains changes to the Chromium public API. See also https://trac.webkit.org/wiki/ChromiumWebKitAPI.

[Adam Barth]

Comment on attachment 166673 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=166673&action=review

> Source/WebCore/Modules/mediastream/RTCPeerConnection.cpp:139
> +    if (!document || !(document->frame())) {

document will always be non-0 here.

> Source/WebCore/platform/mediastream/chromium/RTCPeerConnectionHandlerChromium.cpp:61
> +WebKit::WebRTCPeerConnectionHandler* RTCPeerConnectionHandlerChromium::toWebRTCPeerConnectionHandler(RTCPeerConnectionHandler* handler)
> +{
> +    return static_cast<RTCPeerConnectionHandlerChromium*>(handler)->m_webHandler.get();
> +}

I probably would have just added a public accessor for m_webHandler and had FrameLoaderClientImpl do the static_cast, but this is ok too.

[Tommy Widenflycht]

Comment on attachment 166673 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=166673&action=review

>> Source/WebCore/Modules/mediastream/RTCPeerConnection.cpp:139
>> +    if (!document || !(document->frame())) {
> 
> document will always be non-0 here.

Removed check for empty document.

>> Source/WebCore/platform/mediastream/chromium/RTCPeerConnectionHandlerChromium.cpp:61
>> +}
> 
> I probably would have just added a public accessor for m_webHandler and had FrameLoaderClientImpl do the static_cast, but this is ok too.

RTCPeerConnectionHandler doesn't know anything about any private implementations, neither should it I think, so I am keeping it like this.

[Tommy Widenflycht]

Created attachment 166830 [details]
Patch for landing

[Adam Barth]

Comment on attachment 166830 [details]
Patch for landing

ok

[WebKit Review Bot]

Comment on attachment 166830 [details]
Patch for landing

Clearing flags on attachment: 166830

Committed r130270: <http://trac.webkit.org/changeset/130270>