Fwd: draft-ietf-http-v11-spec-rev-04 comments

Some loose ends from previous topics.


> From: "Roy T. Fielding" <fielding@kiwi.ics.uci.edu>
> Date: Wed, 09 Sep 1998 00:41:41 -0700
> To: Jim Gettys <jg@pa.dec.com>
> Subject: Re: Fwd: draft-ietf-http-v11-spec-rev-04 comments
> -----
> It looks like the second-to-last paragraph in section 5.1:
> 
>    In requests that they forward, transparent proxies MUST NOT rewrite the
>    "abs_path" part of a Request-URI in any way except as noted above to
>    replace a null abs_path with "*", no matter what the proxy does in its
>    internal implementation.
> 
> should be replaced by
> 
>    A transparent proxy MUST NOT rewrite the "abs_path" part of the
>    received Request-URI when forwarding it to the next inbound server,
>    except as noted above to replace a null abs_path with "/".
> 

I've changed all references to inbound to upstream and outbound to downstream
in the document.

> I checked and all the other leftovers with the "*" were removed.
> 
> On a slightly unrelated note, I just noticed that the frequently used
> terms inbound, outbound, upstream, and downstream are never defined.
> Can they be assumed, or should they be added to the terminology section?
> 

Gotta stop polishing this apple; I think it is pretty obvious from the
context what upstream and downstream mean.

> Another continuing problem is that the spec only allows the
> syntax used by the CONNECT method by accident (basically, because
> the host:port syntax that CONNECT uses for a Request-URI also happens
> to be a legitimate scheme:scheme-specific absoluteURI).  My suggestion
> of replacing Request-URI with Request-Target was turned down due to
> the editing cost.  We could bypass it by adding
> 
>        Request-URI    = "*" | absoluteURI | abs_path | authority
> 
> with a paragraph saying that the authority form is only used with CONNECT.
> Speaking of which, the following
> 
>    9.9 CONNECT
> 
>    This specification reserves the method name CONNECT for use by SSL
>    tunneling. [44]
> 
> should not mention SSL by name.  It should just be "for use with a
> proxy that can dynamically switch to being a tunnel."  Fortunately,
> I defined what "tunnel" means long before the CONNECT trick came out.
> 

OK. I'll fix.

Thanks for the final set of nits.  I'm now starting what I hope to be
revision 5 production.

				- Jim

Received on Wednesday, 9 September 1998 07:25:58 UTC