Added support for simulcast and TWCC to Duktape and Lua plugins

[lminiero]

This patch adds support for both simulcast and TWCC (transport-wide CC, i.e., sender side bandwidth estimation) negotiation to the Duktape and Lua plugins. This was made possible thanks to a generous sponsorship by our good friends at Voicenter, so make sure to properly thank them if you needed these features in your scripts! ❤️

In case you want to start using the simulcast features in a Duktape/Lua script, while negotiation and RTP management is automatic there's a couple of things you'll need to do:

1. You can use the setSubstream(id, substream) and setTemporalLayer(id, temporal) methods from your scripts to configure which layer you're interested in: by default both plugins will assume you'll want the highest quality (2, even though the maximum temporal layer is really 1 right now). Notice that it's up to you to ask for a keyframe after a layer change request, which you can do using the already existing sendPli(id) request.
2. You can be notified when a substream/layer actually changes via the substreamChanged(id, substream) and temporalLayerChanged(id, temporal) callbacks: this is especially useful when you want to be aware of when a layer change request was actually done (which may not be right away, e.g., if a keyframe was missing), and/or to notify users about when things change (which is what we do in our demos, for instance).

You can find an example of how those new methods and callbacks are used in Duktape's echotest.js and Lua's echotest.lua scripts, which will work nicely with the existing EchoTest demo if you have it connect to those plugins rather than the C EchoTest one.

For what concerns TWCC, instead, we now automatically negotiate a few RTP extensions in both plugins, namely mid, rid and repaired-rid: the TWCC RTP extension is also negotiated by default, but you can disable it by passing disableTwcc: true when calling generateAnswer(). I'd always suggest negotiating it, as it will make publisher streams more stable, but if for some reason you'd rather disable it, that's the way to do it.

I only tested this with the EchoTest demo, since that's the script available in both plugins, and it seems to be working as expected. Looking forward to your feedback to see if there's anything that needs to be fixed, tweaked or improved: the sooner that happens, the sooner we can merge!

[lminiero]

Merging in master, thanks again to Voicenter for making this possible!