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

Re: Likely OT: "closing the connection" (was: i28 proposed replacement text)

From: Henrik Nordstrom <henrik@henriknordstrom.net>
Date: Thu, 15 May 2008 21:58:23 +0200
To: Frank Ellermann <hmdmhdfmhdjmzdtjmzdtzktdkztdjz@gmail.com>
Cc: ietf-http-wg@w3.org
Message-Id: <1210881503.17190.10.camel@henriknordstrom.net>

On tor, 2008-05-15 at 21:04 +0200, Frank Ellermann wrote:

> Remotely related, I use a HTTP/1.0 GET script, and this
> script, after sending its request, closed the sending
> direction to indicate that it now only waits for an
> answer.  Recently I stumbled over a HTTP/1.1 server
> sending only what would be the first chunk in HTTP/1.1,
> instead of the complete file.
> 
> Who got it wrong, my script or this server ?  I got the
> complete file when I disabled the code terminating the
> sending direction (from my POV).

Behaviour is somewhat unspecified if the client closes it's transmit
channel while still waiting for a response, so don't do it. Some servers
will give you the whole response, some will assume the client aborted
and abort the transaction.

There is arguments in favor for both server behaviours.

In general clients should only close it's transmit channel in response
to the server closing the connection.

"Connection: close" is used by the client if it want to signal that
there is no further requests to be expected on this connection. (or in
case of HTTP/1.0 no Connection: keep-alive header..)

Regards
Henrik
Received on Thursday, 15 May 2008 19:59:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:47 GMT