Re: [squid-users] 301 Moved Permanently: bug?

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Tue, 16 Jul 2002 00:26:05 +0200

On Monday 15 July 2002 23.27, Florin Andrei wrote:

> Ah, yes, i see what you mean.
> Nope, i don't see any actual "HTTP data" (that's in
> Ethereal-speak). It's only the directive and the headers.

The very important thing to look for at the packet level is if the
server sent a FIN indicating the end of the reply or not..

> But _theoretically_ the browser should follow the headers even when
> the entity is missing, right?

That is up to the browser. The HTTP message as such does not end until
the end of the reply entity which is signalled by the server sending
a FIN (HTTP/1.0, no Content-Length header).

> At least, IE does that when i browse
> without a proxy. Only when browsing through the proxy it gets
> stuck. I cannot test it with other browsers because it's one of
> those "IE only" web applications. :-(

One thing you can verify to see if there is a Squid problem is if the
required blank line after the headers is present in the Squid reply..
if it is not there then there is a Squid problem.

If the blank line is there then I would say you have either (probably
both) of

a) A broken browser who for some reason hangs on the Squid reply
headers. Some simple experiments using netcat to duplicate the reply
by Squid should tell fairly quickly by varying the response slightly
to see if the browser behaves differently..

b) A broken server being confused by Squid to act differently than
when your browser is talking to it direcly. Compare traces..

Regards
Henrik
Received on Tue Jul 16 2002 - 00:50:31 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:09:15 MST