Perplexing Squid Problem

[davet246]

I'm running a Squid proxy out to a VPN with it's own proxy to the internet. All is fine on my server box. From my clients however...

When I enter a web address from a client browser (Konqueror - I'm using KDE), it makes contact with the VPN proxy server, gets the username and password (First contact only. After that it remembers me.) and loads the web page. So far, so good. Click on any link made up of variables (like a google search or amazon links), however, and it times out and give a "connection failed" error. If I reconfigure Konqueror to bypass Squid and go directly to the VPN proxy, everything is fine, so the problem seems to be in my Squid configuration.

The big question is: Why does it work for direct access to web pages, but not for addresses created 'on the fly' within web pages?

Any ideas appreciated.

[davet246]

In squid.conf, the default value for the hierarchy_stoplist tag was set to "cgi -bin ?". Once I removed the "?", everything works great.

Thanks to snyone who thought about the problem.