Mark,

ยง5.1.1 (where the ABNF for credentials is defined) will move out of the OAuth 
core spec in the doc split.
I would like to see the bearer spec use a non-OAuth scheme name when it defines 
its Authorization header.

I strongly favour restricting tokens to a much smaller alphabet, such as the 66 
unreserved chars (A-Z a-z 0-9 - _ . ~). Anything more means we need escaping 
mechanisms in all the places a token can go. Anything more still doesn't 
support arbitrary binary data or arbitrary Unicode text so you still need an 
encoding such as base64url.

Supporting 'quoted-string' would not be strictly necessary for an access-token 
parameter if values were restricted to those 66 chars. However, I would be 
happy enough to require parsers to support (token | quoted-string).

I don't think we need 'realm' as an auth-key. It is defined for 
WWW-Authenticate response headers, but isn't necessary for Authorization 
request headers.

I'm in favour of only using name=value pairs after the scheme name, even if 
there is only 1 parameter (the bearer token). It enables parameters to be added 
in the future.

My variant of your proposal:

    credentials    = "BEARER" RWS #auth-param
    auth-param     = auth-key "=" ( token | quoted-string )
    auth-key       = "a" | future-key
    future-key     = token

The "a" parameter holds an access token that MUST only consist of chars from 
the set of 66 unreserved chars [should probably express that in the ABNF]. 
Examples:
  Authorization: BEARER a=vF9dft4qmT
  Authorization: BEARER a="vF9dft4qmT"

--
James Manger


-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Mark 
Lentczner
Sent: Thursday, 28 October 2010 10:44 AM
To: [email protected]
Subject: [OAUTH-WG] OAuth v2 Authorization Header Credentials

In my work implementing OAuth v2 to draft 10, I have had to review the
definition of credentials included in an Authorization header. I
believe this current definition has some significant issues for
parsers. However, with some small clean up of the definition, proposed
below, I believe that the credentials can be made robust and more
compatible for existing and future implementations.
...
_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to