On Fri, 28 Nov 2003 [EMAIL PROTECTED] wrote:

> I am curious what method(s) you used to analyze this problem and make
> your determination?

I collected both the request send by Mozilla when talking directly to the 
site and the same request when proxied via Squid, I then manually repeated 
variations of the difference between these two until finding the parameter 
which made a significant difference.

The request details was collected with tcpdump or ngrep. The netcat was 
used for manually sending the modified requests to the server, simulating 
normal browser requests but in a fully controlled manner.  In parallell 
tcpdump was used for monitoring the traffic to determine when/how/why 
connections was terminated.

Regards
Henrik

Reply via email to