Wrapper around C++ BLS

[wavepacket0]

Hi guys!

first of all, really good idea to create a Go port of the BLS Chia lib, this could open further possibilities Go side.

Why did you decide to do a complete port (I understand a native Go implementation was missing) instead of using CGO/SWIG and create an interface to the offcial C++ BLS lib?
This could avoid the need for code updates to be on par with the official lib and also use all the speedups due to GMP.

I would also suggest to keep the same naming convention as the official lib, for example using G1Element/G2Element

[chuwt]

Hi friend @wavepacket0.

Thank you for reaching out to me.

And I didn't think about it too much about this repo. I just want to write a pure a lib for go developers to use (including me). I'm glad if this repo helps others. I'll be happy too if others can give more advice and more PRs.

[wavepacket0]

thank you for the answer @chuwt

I would like to see a somewhat complete chia-bls implementation in go, so we have a common goal!
My primary language is not Go, nor im experienced in the field of cryptography, but let's try ;)

As of now Im comparing different bls12-381 Go implementations: kilic's one (the one you're using, CIRCL (Cloudflare's crypo Go toolkit, more or less like Relic Toolkit) and go-ethereum implementation (updated, security audited version forked from kilic repo).

What I see is that no one uses minimum public key sizes (MPL) as Chia does for G1/G2 elements. Given the fact kilic's original repo seems without recent activity and the updated version of it is inside go-ethereum, I think im gonna use that as a base, updating G1/G2 to use MPL as Chia uses and then base your lib on top of that. What do you think?

[chuwt]

@wavepacket0 Sounds great. Also here is the implementation written by Coinbase. Hope it will help.

[wavepacket0]

Thank you for the Coibase impl, missed that. I think i will stick with go-ethereum or Coinbase one as both have passed audit phase @chuwt