W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2012

Re: Content-Length and 304

From: Willy Tarreau <w@1wt.eu>
Date: Thu, 20 Sep 2012 08:23:32 +0200
To: Zhong Yu <zhong.j.yu@gmail.com>
Cc: HTTP Working Group <ietf-http-wg@w3.org>
Message-ID: <20120920062332.GI413@1wt.eu>
On Thu, Sep 20, 2012 at 12:43:58AM -0500, Zhong Yu wrote:
> How about HTTP/1.0 clients, as far as you know - if the connection is
> keep-alive, can they handle 204/304 framing correctly if
> Content-Length is missing, or set to a non-zero value. What's the best
> effort to keep a 1.0 connection alive when responding 204/304? Or is
> it safer to just close the connection after responding.

I don't know but it would be a waste of resource to close on 304, since
it's designed to be fast. However, you should note that keep-alive was
not specified in 1.0 and was just implemented on top of the spec. So it
is very likely that some implementations will differ.

FWIW, on the implementation where I observed the truncated objects
behaviour with "Content-Length: 0", the connection was made in HTTP/1.0
and was not using keep-alive. I *think* it was Apache 1.3.12 with
mod_proxy.

Regards,
Willy
Received on Thursday, 20 September 2012 06:23:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 20 September 2012 06:24:00 GMT