Add SignerERC7913 and verifiers for P256 and RSA

[Amxx]

Includes

• contracts/interfaces/IERC7913.sol: ERC-7913 interface
• contracts/utils/cryptography/ERC7913SignatureVerifierP256.sol a stateless verifier for P256 keys. This is unopinionated and production ready, so I don't think it needs to be a mock
• contracts/utils/cryptography/ERC7913SignatureVerifierRSA.sol a stateless verifier for RSA keys. This is unopinionated and production ready, so I don't think it needs to be a mock
• contracts/utils/cryptography/ERC7913Utils.sol library for checking signer's signatures using ERC-7913 (with ECDSA/ERC-1271 fallback)
• contracts/utils/cryptography/SignerERC7913.sol abstract signer that uses an ERC-7913 signer.

[ernestognw]

I think I would prefer having just one implementation of the isValidSignatureNow function. The rationale is that signature always ends up in memory for any valid verification, and a portion of signer will be copied too. So I doubt there's actual benefits in having two versions. I'll simplify for now following #109

[Amxx]

The memory version uses a slice, that does a memory copy.

Overall, calldata versions are often cheaper because they don't copy to memory until the very last point (when the Abi call is encoded) contrary to the memory version that copy to memory at least once more (in this case 2 times)

Overall not a big deal. Having a single "memory" version is probably good enough. Just explaining the initial idea

[ernestognw]

Yeah I saw the memory version is required, so I'd agree with just one "memory" version for now unless there are significant savings measured.

Thanks for answering though, now please you go 👨‍🍼