W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2008

Re: Deploying new expectation-extensions

From: Jamie Lokier <jamie@shareable.org>
Date: Sun, 6 Apr 2008 15:22:50 +0100
To: Henrik Nordstrom <henrik@henriknordstrom.net>
Cc: HTTP Working Group <ietf-http-wg@w3.org>, google-gears-eng@googlegroups.com
Message-ID: <20080406142250.GA6944@shareable.org>

Henrik Nordstrom wrote:
> fre 2008-04-04 klockan 01:54 +0100 skrev Jamie Lokier:
> > My reading of "Expect: 100-continue" is that the proxies _themselves_
> > can return "100 Continue" responses (as well as eating the one
> > produced by the next hop, if any.
> 
> My reading is that proxies can not inject 100 Continue, only forward
> them or eat them..

If there's one thing becoming clear, it's that different people have
different ideas about the behaviour of Expect and about 100 Continue,
while reading the same spec.

> An 100 Continue is a guarantee that the client can continue sending the
> request entity. 

The client can always send the request entity, because if there's a
HTTP/1.0 proxy in the chain, the client won't receive 100 Continue, so
HTTP/1.1 clients have an unspecified short timeout and send it anyway.

The 100 Continue is a hint to continue immediately, instead of after a
short timeout.  The idea is if the server sends an error response
(quickly), the client can abort the request and close the connection
without transmitting any part of the entity.

At least, that's how _I_ understand the spec!

-- Jamie
Received on Sunday, 6 April 2008 14:23:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:46 GMT