W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 1998

Re: Microsoft proxy

From: Roy T. Fielding <fielding@kiwi.ics.uci.edu>
Date: Thu, 01 Oct 1998 18:07:28 -0700
To: "David W. Morris" <dwm@xpasc.com>
Cc: http working group <http-wg@hplb.hpl.hp.com>
Message-Id: <9810011807.aa11820@paris.ics.uci.edu>
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/87
>On the other hand, the proxy detected an error and attempted to report 
>it.  The question to me would hinge on whether the proxy then did
>a connection reset or normal close. 

The proxy is broken.  The only thing it needs to do to "report"
a dropped connection is to close its own connection with the client.
That is why we have both content-length and the chunked encoding.

>I don't believe that HTTP is well architected for recovering 
>partial transfers or in some cases even detecting incomplete transfers
>except those initiated by the client. Without architected error recovery,
>its all up to the implementation.

It is all up to the implementation in any case.  Partial transfers are
easily and completely recoverable for HTTP/1.1.  The only thing it is
missing is some indication of the reason why the connection was dropped,
but that is hard to do without multiplexing or a separate control channel.
That was wisely left for a major protocol revision.

Received on Thursday, 1 October 1998 18:14:58 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:43:05 UTC