Signature malleability

[mrwillis]

I believe that as Edge is currently written, it allows for signature malleability in its verification logic, in the sense that both versions of the proposer and sealer signatures will be valid. It might not cause bugs, if there is no consequence to having a signature be verified twice, but it might not be intended.

go-ethereum rejects s values in the upper range

https://github.com/ethereum/go-ethereum/blob/56dec25ae26bf749b93c3ea69538fabea60c5768/crypto/crypto.go#L263

Discussion in tendermint:

tendermint/tendermint#1958

Open-zeppelin rejects one of the transformations

https://github.com/OpenZeppelin/openzeppelin-contracts/blob/master/contracts/utils/cryptography/ECDSA.sol#L132. Related issue that spurred this: OpenZeppelin/openzeppelin-contracts#1622

We should consider adding this functionality in a breaking change (it would have to be a fork I believe) for completeness