Proofs aren't zero-knowledge

[ebfull]

I didn't want to bring in the rand crate and deal with that mess, so the proofs right now are not zero-knowledge as mentioned in this comment: https://github.com/ebfull/sonic/blob/master/src/synthesis.rs#L666

[ebfull]

Actually, to be specific the mess is that there are two different versions of rand being brought in and I didn't want to reconcile them at the time.

[burdges]

It's possible paritytech/schnorrkel@8150ef6#diff-b4aea3e418ccdb71239b96952d9cddb6R260-R280 might help if you care in the short term, but waiting for dependencies to converge sounds better.