JL4 will webrtc be the transport mechanism for yjs
Only if you install that linked extension: core will continue to do… whatever it’s doing. Really a lot going on there.
Definitely not pushing that linked WebRTC provider as having MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
: it was extracted from jupyterlite as it was a way to leverage free (as in beer) signaling-servers-of-opportunity. To my knowledge, it has only been evaluated for small demos.
At scale, I imagine an yjs-xmpp
-powered provider backed with ejabberd
would probably be the most robust: a server NIC will give out before that piece of code will, and presumably the yjs piece will support it well.
deviate from the intended product line
This little post isn’t likely to give you the detail you want, and I definitely don’t know all the stuff. You’ll want to track:
- open PRs on the main
jupyterlab/jupyterlab
repo- we’re well into the alpha phase, and probably a beta phase will hopefully start soon once some of the big rocks, like rtc, codemirror 6, etc. have settled a little more
- team-compass
- Jupyter Community Calendar