RE: FW: Digest mess

I'd also note that this is 1/2 way (or so) to a way to avoid much or all of
the problem of proxies fiddling with the headers -- create the digest over a
cononical form of the values of the header, instead of what's actually sent
or received.

> 1.  Compose the subset of headers to be digested
> 2.  Combine into a single string with CR/LF between headers
3.  Canonicalize (reduce LWS to a single space -- anything else?)
4. Digest it

Paul

> ----------
> From: 	Jeffrey Mogul[SMTP:mogul@pa.dec.com]
> Sent: 	Wednesday, December 31, 1997 3:55 PM
> To: 	'ietf-http-wg@w3.org'; http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
> Subject: 	Re: FW: Digest mess 
> 
> Ben Laurie writes:
> 
>     David W. Morris wrote:
>     > Also, they encoding rule should probably be something like:
>     > 
>     > 1.  Compose the subset of headers to be digested
>     > 2.  Combine into a single string with CR/LF between headers
>     > 3.  Encode the whole string.
>     
>     Isn't that exactly what Jeffrey was suggesting?
> 
> Exactly.  Except that I simply asserted that this was easy to
> specify; David actually wrote the spec :-)
> 
> -Jeff
>

Received on Wednesday, 7 January 1998 22:19:44 UTC