>> You cannot define the field as holding: a JOSE message (compact
>> serialization), or raw string content. Because the raw content might
>> look like a JOSE message!

> If you base64 the content, it will never look like JOSE.  It won't look
> like a compact JOSE object, since it's a base64 string with the wrong
> number of components (1 instead of 3/5).  It won't look like a JSON
> JOSE object, because it won't be JSON.
>
> So, I actually proposed some syntax on this on the OAuth list.  The
> proposal allows a header, but the concept is the same.
> <http://www.ietf.org/mail-archive/web/oauth/current/msg11857.html>

I like that proposal. It should be in a JOSE doc, not JWT.

Either "alg":"none" or "alg" being absent could be used to distinguish the 
unprotect message from JWEs and JWSs. "alg":"none" is more consistent with:
  "alg":"RS256"   -> asym-JWS,
  "alg":"HS256"   -> sym-JWS,
  "alg":"dir"     -> sym-JWE,
  "alg":"ECDH-ES" -> asym-key-agree-JWE,
  "alg":"A128KW"  -> sym-key-wrap-JWE

"T":"none" would be better.

> The important thing here is that however the unprotected stuff is
> represented, it must have the property that if I feed it into a JOSE
> library, the library will say "This is not a JWS or JWE", not "This is
> a valid JWS".

Agree.
Unprotected stuff should have 2 segments (header.payload), which encourages 
implementers not to treat it as a special case of JWS.

--
James Manger
_______________________________________________
jose mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/jose

Reply via email to