Replace redpallas stub in zebra-chain with full implementation

[dconnolly]

Is your feature request related to a problem? Please describe.
To suffice types in the sapling->orchard port in zebra-chain, we created redpallas types but they are functionally incomplete. They need to be functional to verify SpendAuth and Binding signatures for NU5.

Describe the solution you'd like
@str4d has implemented reddsa as a fork of redjubjub to abstract the common types and traits into a dependency for the redpallas types in the orchard crate. The reddsa orchard branch there defines the Orchard basepoints, which is not my favorite.

Should the redjubjub crate turn into something a more shared reddsa, that includes the redjubjub and reddsa? With FROST, it may be annoying to have 3 crates to juggle: reddsa as a dependency, redjubjub, redpallas, with boilerplate across them.

Or should we just duplicate where necessary?

[daira]

RedDSA was designed to operate over a generic represented group. To my mind, it would be very unfortunate if we ended up duplicating effort. A small amount of boilerplate might be ok, but I think it's preferable to merge implementations as far as possible.

[str4d]

The reddsa orchard branch there defines the Orchard basepoints, which is not my favorite.

That was the simplest path forward that made the smallest change to the existing redjubjub code (to make it easy for you to review). I'd be happy with a further refactor that enabled reddsa to be completely generic; I just don't know whether that is compatible with the strict typing that redjubjub has. Given that I had to use sealed traits in redpallas anyway though, I think it's possible to remove them from reddsa, but I won't commit to doing that work unless it will actually be used by Zebra.