FAPI2 JWS alg choices are much wider than FAPI1

[Joseph Heenan]

FAPI1Adv allowed only ‘PS256’ and ‘ES256’.

FAPI2Baseline says:

This seems to allow a much wider choice of algorithms, including possibly RS256 that was explicitly (and somewhat painfully for some implementers/deployments) dropped from FAPI1 due to concerns over RSASSA-PKCS1-v1_5 like those expressed in https://www.rfc-editor.org/rfc/rfc8017#section-8 (I’m not sure if that section is considered mandatory to comply with in FAPI2Baseline, if it is this could be a case where it’d be easier for everyone if a clear requirement similar to that in FAPI1 was used instead).

[Nat Sakimura]

On the March 2 call, it was agreed that this widening was not intentional.

[Dave Tonge]

https://bitbucket.org/openid/fapi/pull-requests/313/restrict-to-ps256-and-es256

[Dave Tonge]

PR merged