jeffrey-p-mcateer / a-pod Goto Github PK

See https://developer.mozilla.org/en-US/docs/Web/API/Screen_Capture_API/Using_Screen_Capture

We should have separate buttons for all our capture possibilities; one on the follower UI (for permissions + sending to leader) and one on the leader UI (for recording).

This will be kind of a large change; at the moment our p2p connection is very simple and we need to extend it to handle forwarding connection details to all connected parties without duplicating media channels.

Plan:

• extend webserver::GlobalData to include client metadata
  • name
  • list of media being saved: ["webcam", "screen"] for now (possibly model an audio-only one someday, but for now do both)
• Extend client JS code to automatically broadcast webcam + screen to leader and re-connect if connection is lost

Relevant docs:
https://webrtc.github.io/samples/src/content/peerconnection/multiple/
https://stackoverflow.com/questions/40065536/webrtc-multi-peer-connection

a-pod needs remote machines to be able to talk to it, and some OSes (primarily macOS and Ubuntu, also Fedora) ship with firewalls that disallow all traffic by default. Writing logic to interact with all the possible OS firewalls is time-consuming and unportable, so the best strategy is simply detect blocked ports and report that to the user. Then it's the operator's problem.

Process:

• detect own IP address
• bind to 0.0.0.0:PORT
• fire off HTTPS request to IP:PORT/conn-test and expect the result to be ok
• If no response/timeout (we can use 5 seconds for the timeout) tell the leader nobody will be able to access the follower UI.