Clarify key selection for client_metadata jwks

[deshmukhrajvardhan]

PR Checkpoints

• official html rendering, which is generated using mmark - Download it from the GitHub action tab, https://github.com/openid/OpenID4VP/pull//checks - by clicking on 'artifacts' and then 'output'
• >=4 approvals OR >=2 if editorial PR
• >=1 week since PR's final state
• Only Chairs can merge the PR

Summary

Clarify key selection for client_metadata jwks like described in FAPI: https://openid.net/specs/openid-financial-api-part-2-1_0.html#duplicate-key-identifiers

Related Issue(s)

resolves #438

Special notes for your reviewer

[bc-pi]

Honestly not sure this works well with anticipated changes from #477

[Sakurann]

Suggested change

	### Key Selection for `jwks` Client Metadata Parameter {#client_metadata_key_selection}
	JWK sets SHOULD NOT contain multiple keys with the same `kid`. However, when there are multiple keys with the same `kid`, the verifier shall consider other JWK attributes, such as `kty`, `use`, `alg`, etc., when selecting the verification key for the particular JWS message.
	### Key Selection for `jwks` Client Metadata Parameter {#client_metadata_key_selection}
	JWK sets SHOULD NOT contain multiple keys with the same `kid`. However, when there are multiple keys with the same `kid`, the verifier shall consider other JWK attributes, such as `kty`, `use`, `alg`, etc., when selecting the verification key for the particular JWS message.

[Sakurann]

@bc-pi umm this is based on the feedback from the implementers who have implemented only response encryption (no JARM kind of signing at all) so i still think this PR is valuable regardless of #477

[bc-pi]

@bc-pi umm this is based on the feedback from the implementers who have implemented only response encryption (no JARM kind of signing at all) so i still think this PR is valuable regardless of #477

I think I was largly reacting to the text (now gone) that was borrowed from FAPI. But the "selecting the verification key for the particular JWS message" in this PR is in a new little section that is refrenced only from the context of the client_metadata/jwks param where it says "Public keys included in this parameter MUST NOT be used to verify the signature of signed Authorization Requests".

Admittedly I don't know specifically what that feedback from the implementers was but the text introduced here is incorrect so probably doesn't meaningfully address it.

[bc-pi]

see prior comments #483 (comment) & #483 (comment)

[Sakurann]

superseded by #477