[WIP] Roadmap about nim-libp2p

[arnetheduck]

Related eth2 ticket: ethresearch/p2p#4

[sinkingsugar]

About Curve25519 if @cheatfate and @dryajov agree I can refactor the current ed25519 code which actually includes (not sure if entirely) Curve25519, although very scalar code.
We also need X25519 for noise btw, again depends on Curve25519.
Let me know what you guys think please.
Or actually including a more optimized C library (since C is nim's assembly) could be another option.

[cheatfate]

@sinkingsugar we do not need optimized C library, because we already has one - nim-bearssl.

Also i'm not sure what you mean under word refactor, because ed25519 is constant-time implementation and so i'm not sure it needs any refactor process. For some reason ed25519 is not part of TLS standard, while Curve25519 is, so i'm not sure is ed25519 implementation is enough to build Curve25519.

[sinkingsugar]

@cheatfate I used bearssl fully anyway so don't worry about it!

[Menduist]

Closing this since almost everything on this list is done now (except exotic multiplexers)