ACL based on local port

Here's the situation:

I have a single "terminal server" running portslave, in to which all my
customers dial. Some of my customers should be allowed HTTP access, others
only limited HTTP access. The terminal server is running transproxy,
forcing all customers to use the squid server. This works fine.

The trouble is, the squid server has no way of knowing whether a request
coming from the terminal server is a customer who should be granted full
HTTP access, or limited HTTP access. I thought it might work to have squid
listen on two ports (3128 and 3129) for HTTP requests, then based on which
port it receives the request on, have it perform a different action.
Apparantly there's no option for this in the ACL setup, though.

Are there any other established methods by which to differenciate between
requests coming from the same transparant proxying host?

I suppose if all else fails, I can bind two IP addresses to the terminal
server, and tell one transproxy instance to bind to the second IP address so
squid will know the difference between requests. That seems like more than
should be required, though...

Any thoughts or suggestions are much appreciated.

--
"The wireless music box has no imaginable commercial value. Who would pay
for a message sent to nobody in particular?" -- David Sarnoff's associates
in response to his urgings for investment in the radio in the 1920s.
--
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
  Jonathan Hall  *  jonhall@futureks.net  *  PGP public key available
 Systems Admin, Future Internet Services; Goessel, KS * (316) 367-2487
         http://www.futureks.net  *  PGP Key ID: FE 00 FD 51
                  -=  Running Debian GNU/Linux  =-
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -