Analysis on audit "3.4 unwrap() risks"

[oxarbitrage]

I was checking all the unwrap()s we have in this codebase and i found:

• some are unavoidable and safe to use like the ones as: https://github.com/ZcashFoundation/redjubjub/blob/main/src/batch.rs#L208-L215. Here the code will unwrap() after checking there is some value. Is the correct way for CtOption, they also have a comment XXX-jubjub: should not use CtOption here however i have no idea what is the status of that.
• some others are safe but they need some comment.
• some can be changed and use the ? operator instead.
• some can be replaced with expect("some more clear message")
• some are in tests so we don't really care.

This PR makes some changes where some improvement can be made.

The PR can have conflicts with #88 so we probably need to rebase if we consider to merge.

[teor2345]

I don't think we should return a Result from the basepoint calculation.

It's ok to panic here, because an invalid basepoint is a serious bug in the program.

[daira]

This PR should be closed; the non-test code that it's changing doesn't exist in this crate any more.

[conradoplg]

Closing since it oudated, I opened ZcashFoundation/reddsa#61 instead