[WIP] Implementation of .renegotiate and .negotiate API

[cusspvz]

• Add .renegotiate
• Add .negotiate
• Add negotiate event
• Add .addStream
• Add .removeStream
• Add .addDataStream - have a question if we should allow multiple datastreams
• Add .removeDataStream
• Add tests where a fake MediaStream instance is added into peers post their first established connection
  - [x] Added .editorconfig
  - [x] Added node_modules/ to .gitignore and .npmignore

Current stage:

• firefox is giving an error because of differences on sdp both m lines, unfortunately I don't have the time to finish this implementation [EDIT]

Related issues:

• Create renegotiation method #68
• expose addStream on prototype #50

[cusspvz]

NOTE: I've seen that there are linting issues on the CD, I will fix them just before landing the PR.

[feross]

Please don't commit this file

[feross]

I think it's a mistake to stop using the onnegotiationneeded event. That was required for things to work for me before -- has that changed?

[cusspvz]

I have to finish this up, the only issue I had was with the Firefox implementation, tests were flawless on chrome.
Regarding your question related with onnegotiationneeded, it is called whenever a MediaStream is added, perhaps you had async issues and you were calling offer before having the stream?

Seen that you've closed the PR, are you still interested in having renegotiation here?

[feross]

Sorry for closing the PR -- that was an accident. Before I can merge this, you need to fix the merge conflicts, fix the code style, and fix it on Firefox.

Looks like you also might need to add:

if (renegotiating) return

to the top of the _onIceCandidate function for Chrome.

See: https://www.webrtc-experiment.com/docs/how-to-switch-streams.html

[feross]

Also, the test style is kind of weird. Can you clean it up a bit?

[feross]

Just use if statements. That would be much simpler.

[MatthieuLemoine]

@cusspvz Are you still working on it ?

[princemaple]

Hi @cusspvz , are you not interested in this anymore?
(Oops, didn't see the previous comment from one day ago, sorry about the double ping)

[cusspvz]

@MatthieuLemoine @princemaple I am but not for now, but can arrange some time to inject some changes on the PR.
As far as I can remember, there was an issue on Firefox upon a renegotiation.

Also, seems firefox only accepts one data stream per RTConnection.

@feross if I implement addDataStream should it error whenever it isn't possible?

[jeregrine]

addStream is going to be deprecated in future versions. Firefox supports addTrack. They have a migration plan with chrome back-support here https://developer.mozilla.org/en-US/docs/Web/API/RTCPeerConnection/addStream

[t-mullen]

I think we should leave addStream in the simple-peer API. It's much simpler than adding individual tracks (although a secondary track API couldn't hurt). We already use addTrack internally.

[jeregrine]

addStream is being deprecated. If simple-peer wants to support an addStream that simply calls addTracks on each track then that would be fine. But long term the API is going away.

[cusspvz]

Closing due to #250
Thanks for putting your efforts on it @rationalcoding.