> OK - So ho do I do that?

> When I run squid on its own (on machine 192.168.123.101) and a user on
(say) 
> machine 192.168.123.103 accesses a blocked site, squid pops up a
message 
> which is (physically) held in a file on m/c
192.168/123.101/usr/share/errors/ERR_Somemessage 
> - so squid *can* access the local drive on 192.168.123.101.

> SquidGuard - running on the same machine together with the same
instance of squid - can't 
> access the same file - nor can it run the cgi file in
/var/www/cgi-bin.

> I guess my question boils down to - why can't squidGuard use the same
redirect file as squid?

> I know that my dumb questions must be frustrating for you guys - but
not half as frustrating
> as it is for me being so close to a solution yet so far.

The short answer is squid's redirector interface doesn't work that way.
AFAIK there isn't anyway for a redirector to inform squid to return one
of its error messages.  Perhaps it could be done, but isn't really
neccesary.

Squid/squidguard is not a web server so they definatly wouldn't be able
to access or run a CGI.  Changing them to do that would greatly increase
how complex they would need to be.  With complexness you always get more
bugs and a greature chance for security issues.

I would suggest that you simply run a small web server on your
squid+squidguard host.  I run copy of webfs ( a small http server for
static files only ) on my squid+squidguard host bound to port 9000.
Squidguard is set to redirect to the server on 127.0.0.1:9000.  If you
wanted to run the CGI you would need something a little smarter then
webfs though.


Oh, as some of the previous people mentioned you would definatly want to
setup an acl in squid+squidguard to allow all ruquests to
127.0.0.1:9000, or whatever the URL is for your redirection.


Chris

Reply via email to