RE: [squid-users] Connection: close

From: Robert Collins <robert.collins@dont-contact.us>
Date: Mon, 20 May 2002 11:07:09 +1000

> -----Original Message-----
> From: Warrick FitzGerald [mailto:wfitzgerald@LiveTechnology.com]
> Sent: Monday, May 20, 2002 11:00 AM
> To: Robert Collins; Squid-Users
> Subject: Re: [squid-users] Connection: close
>
>
> Ahh, ok thanks guys I'm starting to see the light :)
>
> The other situation I am then faced with is that these boxes
> accelerate more than one site. I now realize that I can not
> get the realm as such, but is it possible to retrieve the URL
> that the client is requesting .... just because he has access
> to one of the URL's I am acceleration does not mean I want to
> give him access to all.

I think you are making it hard for yourself. Rather than trying to get
all the different info, consider using squid's very comprehensive ACLs
to do all the tests in squid itself. It can check against the requested
site, the username, the url path and more.
 
> I have no working knowledge of SquidGaurd at all, but on
> their site they speak of a "Squid standard redirector
> interface" ... but it has a broken link that that is not of
> much use. If I can not get the URL the client is requesting
> via the authenticate_program, would I be able to do it this way ?

Yes, a redirector is often used for complex reverse proxy scenarios. See
the list archives for some examples and ideas. Also the rproxy tweaks on
devel.squid-cache.org may be of use - but beware they are still in
'development'.
 
> -- One last and not so great method may then of course be to
> use a normal web server like apache etc. to handle the
> authentication .... make a private request to squid on the
> backend for the cached content and then deliver it to the
> client as if it has generated the content itself.

There's no need to do that. Squid should be able to do it all.
 
Rob
Received on Sun May 19 2002 - 19:07:15 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:08:10 MST