refactor(ext/crypto): clean up rust side importKey

[lucacasonato]

This commit cleans up the Rust side of import_key by using a bunch of
enums instead of structs with "type" and "data" fields.

This commit does add some duplicated code for the time being, because
a lot of the other ops still need to get the same cleanup treatment.

This is an alternative to the Rust side changes in #13027.

[lucacasonato]

@seanwykes Does the pattern introduced here work for the JWK implementation?

[cryptographix]

@seanwykes Does the pattern introduced here work for the JWK implementation?

Yes, should do, we'll just add two more members to KeyData enum. :)

[lucacasonato]

Yes, should do, we'll just add two more members to KeyData enum. :)

As in two fields inside KeyData::Jwk, or two new enum members? If you mean the latter, what members would those be?

[cryptographix]

Yes, should do, we'll just add two more members to KeyData enum. :)

As in two fields inside KeyData::Jwk, or two new enum members? If you mean the latter, what members would those be?

I personally find new enum members to be cleaner than having a single big ::JwK structure.

In #12022, I use different structs and enum members for Symmetric, RSA and EC JWKs:
JwkSecretKey(SecretKeyComponentB64), JwkRsaKey(RSAKeyComponentsB64) and JwkEcKey(ECKeyComponentsB64).

[lucacasonato]

Makes sense. Let's cross that bridge when we get to it :-)

[littledivy]

neat

[bartlomieju]

LGTM