Change WebAuthn Signer Structure

[nlordell]

Currently, we have a call structure where WebAuthnSigner calls WebAuthnVerifier which calls a P256Verifier implementation. Overall, the call structure is nice because:

• It avoid deploying large amounts of code in WebAuthnSigner related to computing the WebAuthn signing message at the cost of an extra call. In theory, it would take many signatures before the additional signing gas overhead outweighs the deployment costs (TBD)
• It allows multiple P256Verifier implementations to be deployed and used.

However, there are some drawbacks to exactly how it is implemented. Namely, we require a single WebAuthnVerifier deployment for each P256Verifier that is deployed, but without much benefit (namely, we do an external call anyway - so this can be parameterized in the signature verification and does not need to be an immutable). Additionally, this does not need to be a separate contract to the WebAuthnFactory; in fact the WebAuthnSigners can call the WebAuthnFactory contract in order to perform the signature verification, so all the WebAuthn logic lives in one place.

This issue proposes:

1. Changing WebAuthnVerifier to receive the P256 verifier to use instead of using an immutable
2. Combining the WebAuthnVerifier and WebAuthnSignerFactory into WebAuthn contract for general WebAuthn signature verification using the P-256 curve (names TBD, as we would also want to support EdDSA for example).

[nlordell]

Implemented in #320

[rmeissner]

Can this be closed?