As someone using this api pointed out: > Whoa, shouldn't we require in dlog_eq that generator is NonZero<Point<E>>? Otherwise the proof is shit...