Elm-phoenix-socket question

I’m looking in to Elm again as I am really not liking javascript, but the last time I touched Elm it was with elm-phoenix-socket and it did not have the capability to re-use the existing javascript socket while also lacking heartbeat. It seems to have heartbeat support now, but still lacking the presence functions that are in the phoenix library (though may be PR’d in soon? Maybe? They seem to be thinking no for some reason…).

Right now the phoenix app I have is built using react + phoenix + bootstrap (bootstrap is required by other business stuff), little else. I would love to write it in a better language, however it seems elm-phoenix-socket is still not capable of re-using the javascript phoenix socket to make new channels. Does anyone know of a work-around?

At the moment I have a single socket that is passed to multiple React components for the few dynamic components on the webpages that I have, need to keep it to one websocket connection as well. So I see two ways to go with this now:

1. Try to finagle Elm into using the javascript web socket via injecting the events and such via its ports interop, seems like it would work, also sounds like a lot of work as some things like phoenix presence might even be more easily rewritten inside Elm instead of linking javascript as Elm seems to hate mutable things like the presence part of the library does to mutate an object as events come in.

2. Write it all in Elm, this involves re-writing a not-insubstantial amount of code, but not a large amount either, however I cannot seem to figure out (still new to Elm and elm-phoenix-socket) if elm-phoenix-socket can spool up a single websocket and share it out to multiple embedded components so they can make their own rooms. As far as I can see each embedded component/application will spool up its own websocket, which is a big no-no for what I need to do.

Anyone have ideas?