Conversation
| by other means such as TLS; the `none` algorithm MUST NOT be used or accepted. | ||
|
|
||
| The Transmitter and Receiver MUST use a mutually supported signing algorithm, | ||
| established by means outside the scope of this profile. Receivers MUST reject |
There was a problem hiding this comment.
Would that wording still allow SSF signature negotiation mechanisms as suggested by Joseph? #355
There was a problem hiding this comment.
Signature negotiation is general aspect of JWKS based validations, imo, we should not take dependency on it and add any specific language around it
| established by means outside the scope of this profile. Receivers MUST reject | ||
| events signed with an algorithm they are not configured to accept. |
There was a problem hiding this comment.
This feel like it should be a part of the base spec, not the interop profile.
There was a problem hiding this comment.
base spec calls SET signatures as optional, we are making it mandatory in interop spec, hence added here
| {{event-signatures}}). This allows deployments to adopt stronger algorithms | ||
| and retire weaker ones. Receivers remain responsible for enforcing their own |
There was a problem hiding this comment.
This allows deployments to adopt stronger algorithms and retire weaker ones.
I think that's an over claim. It doesn't provide any mechanism at all for retiring apart from out-of-band negotiation, which is how you do it without the profile. Nor does it have any mechanism for showing a stronger algorithm is available.
There was a problem hiding this comment.
Are you suggesting dropping this sentennce? I wanted to provide forward compatibility when I added it
This branch has not been deployed
3 participants
Closed #350