[RSDK-8489] Implement local webrtc signaling #347
Conversation
| Err(e) => { | ||
| log::error!("error sending signaling message: {:?}", e); | ||
| Err(WebRtcError::SignalingDisconnected()) | ||
| match self.signaling.as_mut() { |
There was a problem hiding this comment.
could some of this be simplified by a function like fn send_answer_stage(&mut self, uuid: String, stage: AnswerStage) -> Result<(), WebRtcError>?
There was a problem hiding this comment.
@gvaradarajan - I'm not quite sure I see what you are saying. Could you clarify? the Answer stuff doesn't come into play at all for local signaling.
- Don't interpret PEER_CLOSE_NOTIFY as a hard error - Use an unbounded channel with blocking send to forward along call updates - Don't fail CallUpdate RPCs that race with a completed Call RPC.
There was a problem hiding this comment.
LGTM mainly some nits. We should merge and start testing it.
I think further work should focus on removing code duplication in webrtc/api.rs and simplify both the rpc stream and the local signaling machinery so the flow is easier to understand.
I would also like to have local signaling be a configurable option in the builder.
|
Per discussion with @npmenard, I'm merging this in its current state, and I'll address comments and add ticketed TODOs in a near-term follow-up PR. |
| Err(e) => { | ||
| log::error!("error sending signaling message: {:?}", e); | ||
| Err(WebRtcError::SignalingDisconnected()) | ||
| match self.signaling.as_mut() { |
There was a problem hiding this comment.
@gvaradarajan - I'm not quite sure I see what you are saying. Could you clarify? the Answer stuff doesn't come into play at all for local signaling.
|
Changes in response to comments are in #349 |
I will add inline annotations to provide context on the current state of things, and what needs to happen next.