Implement 1:1 Calls

[Airyzz]

No description provided.

[Airyzz]

looks like flutter_webrtc package is causing integration test to fail, even if it isnt being used in the actual app. Going to have to disable integration test for now, and investigate this issue later

[notramo]

@Airyzz, this is really cool, is there any ETA for this?

[Airyzz]

@Airyzz, this is really cool, is there any ETA for this?

Thanks! No ETA at the moment as I need to evaluate what changes I need to make for MatrixRTC support

[theotheroracle]

this is immensely epic

[notramo]

@Airyzz as MatrixRTC is still in the making, I would recommend adding calls using the current spec (probably behind a labs switch, or a notification when first used), for multiple reasons:

• Gain some popularity for this client (calls are one of the most requested features, for most client, directly 2nd place behind threading)
• See possible pitfalls in the implementation (mostly network or event ordering/sending corner cases, that should be considered when designing). The Element and Matrix.org team often puts out barely working demos solely for the purpose of experimenting.
• Some clients will probably lack behind MatrixRTC support and use current VoIP spec for a few months.

(Sorry if my comment is dumb, I haven't yet catched up with the MatrixRTC spec, and don't exactly know what is included in it and what is not, neither the differences from the current VoIP spec.)

[Airyzz]

@notramo I think its a good point - it does still need evaluating though because I believe the matrix SDK has had some changes since I last worked on this

[Kommynct]

I think it is no longer worth working on the old calls system,

https://element.io/blog/we-have-lift-off-element-x-call-and-server-suite-are-ready/

element x is ready, so, there is a perfectly good matrixrtc mobile client, matrix 2.0 is here:

https://matrix.org/blog/2024/10/29/matrix-2.0-is-here/

and I think that would just be a massive duplication of effort.

There are so few clients that actually support calls, the old system probably just isn't worth supporting, especially given the small size of this project. It would be a lot easier to implement a warning that it only works with clients that support matrixrtc.

[theotheroracle]

awe, sad but I understand

[Kommynct]

Well, to be clear I'm a nobody that has helped in no way, I'm just giving my 2c on it

[theotheroracle]

I would personally really enjoy p2p calls, since a few clients support it, nheko, fluffy, element, and much fewer clients support the new stuff

[Kommynct]

Element is already planning on switching, and to be clear the new system is also p2p, if element switches, fluffy and nheko will soon as well, you'd be supporting a deprecated feature for a year at best

[Airyzz]

Realistically, we are probably going to support whichever method of calling is supported by matrix-dart-sdk. as it stands now, i will likely merge the current implementation relatively soon, so we can start testing it in pre-release builds.

I dont think switching from the current implementation to MatrixRTC will be a significant amount of work from client side, as most of the implementation detail is handled by sdk.