W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1997


From: Roy T. Fielding <fielding@kiwi.ics.uci.edu>
Date: Sun, 10 Aug 1997 22:40:47 -0700
To: John Franks <john@math.nwu.edu>
Cc: HTTP Working Group <http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com>
Message-Id: <9708102300.aa03527@paris.ics.uci.edu>
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/4157
>A 1.1 proxy cannot request an unchunked
>response from a 1.1 server without violating RFC 2145.  So far I
>have seen two responses to this, 1) It's a feature not a bug, and
>2) Become a 1.0 proxy and send a 1.0 request.

That is true, and it's also on purpose.  In order to reliably transfer
data, it is necessary to know the actual length of the data being
transferred (data error detection is handled at the transport layer).
Failure to supply a length may result in prematurely truncated responses
being misinterpreted as full responses, which was (and still is) a
significant problem with HTTP/1.0.  The minimum change necessary to
enable sending dynamically-generated data in efficient, length-delimited
chunks was to require that all HTTP/1.1 applications be prepared to
receive chunked messages; a server may respond with a Length Required
error (to prevent yet another form of denial of service attack).

In other words, it is not an optional feature of the protocol.  Either
the proxy is willing to accept chunked or it is not an HTTP/1.1 proxy.
I have yet to hear an actual proxy implementer claim that it is an
unreasonable burden.

Received on Sunday, 10 August 1997 23:07:43 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:40:21 UTC