[feat] EdDSA algorithm

[beanow-at-crabnebula]

Has some great properties, like being deterministic, why you might choose it over the ECDSA family.

https://www.rfc-editor.org/rfc/rfc8037

[beanow-at-crabnebula]

I should add, would you accept PRs on this feature? And are there any gotchas in contributing you can think of? :]

[rib]

Thanks. Yeah, I'd be happy to review a PR for adding EdDSA support. I'm not familiar with EdDSA atm but wouldn't expect any major gotchas, assuming it's supported in ring.

[beanow-at-crabnebula]

Ring does support Ed25519 / X25519, but not Ed448 / X448.

• https://docs.rs/ring/0.16.20/ring/signature/index.html#signing-and-verifying-with-ed25519
• https://github.com/briansmith/ring/issues/463
• https://github.com/briansmith/ring/issues/464

The way I'm reading the RFC8037, it doesn't seem required to implement both. It seems to stress interoperability first.

[beanow-at-crabnebula]

Sorry I haven't picked this up. I had conflated this library and https://github.com/Keats/jsonwebtoken Currently evaluating their crate. Since they also use ring, there's support only for Ed25519 but not Ed448.

[rib]

:) no worries.