Re: draft-ietf-httpbis-encryption-encoding-08, "6.1. Normative References"

On 13/03/2017 1:10 p.m., Martin Thomson wrote:
> On 12 March 2017 at 04:42, Julian Reschke wrote:
>> As far as I can tell, the reference RFC7515 can be moved to informative;
>> it's only needed to define base64url for examples.
> 
> Ack, done in the editor's copy.
> 

If that is the only reason, why not use *actual* base64url as defined in
RFC 4648?

The statement "Base64url encoding is defined in Section 2 of [RFC7515]."
is a lie anyway. RFC 7515 simply adds some JSON-centric restrictions to
the RFC 4648 definition. Which do not seem particularly relevant to HTTP
headers.

That would move the reference to RFC 4648 and get one step further away
from JSON abominations. People are free to use the JSON sub-set on what
they send - that sis valid base64url, so long as they accept and discard
the '=' padding from those who dont use JSON encoders.

Amos

Received on Wednesday, 15 March 2017 12:44:10 UTC