No support for ECDSA P-256

[SegalMan]

I'd like to open a PR supporting ECDSA P-256 at least for verification as that is my current need (already using the library with an ed25519 and now having to add a key using this algorithm)

The only main decisions I see from my perspective in that regard:

1. Which create should be used (leading candidates are p256 and ecdsa)?
2. The FailedToVerify error currently having an associated ed25519_dalek::SignatureError, meaning I'd need to either introduce a separate error for verification failure for ECDSA or introduce breaking changes to that error

I didn't want to overstep and directly open a PR, so let's discuss :)

[cameronfyfe]

I need this as well before I can switch from my RFC 9421 implementation to using this crate. Any plans for this, or openness to PRs?

[armfazh]

As this crate follows the development of the specification here
https://github.com/thibmeu/http-message-signatures-directory/issues

it make sense to propose the inclusion of P256 in the spec.

[thibmeu]

if you're able to share, what's your specific use case to use ecdsa-p256-sha256 ?

@armfazh in term of specification, architecture would allow for ecdsa-p256-sha256 to be used as defined in Section 3.3.4 of RFC9421. What the current specification does not provide is test vector for this algorithm, which could be added depending on the use cases.