Re: HTTP/1.1 Contradiction

>19.4.6 Introduction of Transfer-Encoding
>... Proxies/gateways MUST remove any transfer coding prior to forwarding a
>message via a MIME-compliant protocol.
>
>(Sounds good, not only can we remove it [per 4.3], we MUST remove it, but)

IIRC, HTTP is not a MIME compliant protocol.  The spec writers are talking
here about proxies that forward messages to a mail gateway.  (And yeah, a
proxy that has that feature would need to remove Trasnfer-Encoding, since
such a beast will choke any mail software out there.)

>13.5.2 Non-modifiable Headers
>... A cache or non-caching proxy MUST NOT modify or add any of the following
>fields in a response that contains  the no-transform Cache-Control
>directive, or in any request:
>Content-Length
>
>Now if we MUST remove Transfer coding, seems like we MUST add a
>Content-Length header?

Seems to me if you remove the transfer encoding, you would have to add a
Content-Length.  I don't understand why they included that header amongst
those that couldn't be added/modified couldn't be added.  Let's ask them....

Hey HTTP-WG, why can't proxies modify/change Content-Length on no-transform
responses?

-----
Daniel DuBois, Traveling Coderman        www.spyglass.com/~ddubois
   o  The Heroes of Might and Magic II Bible is here!
      http://www.spyglass.com/~ddubois/HOMM2.html