RealtimeOutgoingVideoSourceMac should pass a ObjCFrameBuffer buffer

[youenn fablet]

RealtimeOutgoingVideoSourceMac should pass a ObjCFrameBuffer buffer

[youenn fablet]

<rdar://problem/39153262>

[youenn fablet]

Created attachment 337124 [details]
Patch

[youenn fablet]

Created attachment 337129 [details]
Patch

[youenn fablet]

A follow-up should make getUserMedia mock produce kCVPixelFormatType_420YpCbCr8Planar
That way, we would catch this issue in bots.
And we can continue covering ARGB cases with captureCanvas.

[WebKit Commit Bot]

Comment on attachment 337129 [details]
Patch

Clearing flags on attachment: 337129

Committed r230227: <https://trac.webkit.org/changeset/230227>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[youenn fablet]

(In reply to youenn fablet from comment #4)
> A follow-up should make getUserMedia mock produce
> kCVPixelFormatType_420YpCbCr8Planar
> That way, we would catch this issue in bots.
> And we can continue covering ARGB cases with captureCanvas.

Filed bug 184304 as a follow-up

[Jeremy Jones]

Comment on attachment 337129 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=337129&action=review

> Source/ThirdParty/libwebrtc/Source/webrtc/sdk/WebKit/WebKitUtilities.mm:123
> +    return new rtc::RefCountedObject<webrtc::ObjCFrameBuffer>(frameBuffer);

As-is, ObjCFrameBuffer() has semantics like adoptNS, but is not annotated as such. This will cause problems if arc is enabled.
The parameter to ObjCFrameBuffer() should be annotated with NS_RELEASES_ARGUMENT.

The other option is to release or autorelease frameBuffer here, and retain it in ObjCFrameBuffer().

> Source/ThirdParty/libwebrtc/Source/webrtc/sdk/objc/Framework/Classes/Video/objc_frame_buffer.mm:51
>      : frame_buffer_(frame_buffer), width_(frame_buffer.width), height_(frame_buffer.height) {}

This should either add the annotation NS_RELEASES_ARGUMENT to frame_buffer parameter, or [frame_buffer retain] in the body.