Created attachment 453722 [details]
Patch

Comment on attachment 453722 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=453722&action=review

> Source/WebKit/WebProcess/GPU/webrtc/LibWebRTCCodecs.cpp:525
> +            auto* encoder = m_encoders.get(encoderIdentifier);

If this is guaranteed non-null, then I suggest putting in a reference, not a pointer. Maybe also add an assertion? Not clear to me why this is strongly guaranteed.

(In reply to Darin Adler from comment #2)
> Comment on attachment 453722 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=453722&action=review
> 
> > Source/WebKit/WebProcess/GPU/webrtc/LibWebRTCCodecs.cpp:525
> > +            auto* encoder = m_encoders.get(encoderIdentifier);
> 
> If this is guaranteed non-null, then I suggest putting in a reference, not a
> pointer. Maybe also add an assertion? Not clear to me why this is strongly
> guaranteed.

Will do.
It is guaranteed to be non null as creation/deletion of the pointer is doing the same journey (hop to main-thread then hop to work queue).
I'll update the code to directly use the LibWebRTCCodecs connection instead, which is more efficient anyway.

Created attachment 453818 [details]
Patch for landing

Committed r290829 (248065@main): <https://commits.webkit.org/248065@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 453818 [details].

<rdar://problem/89809391>