Re: User-Name containing an empty string.

Georges Arnould <[email protected]> wrote:

>     Once my patch finished, I wondered if there was another and smartier way
> to force radiusd to send a Access-Accept answer for null User-Name.
> 
>     Any idea for me ?

Well, the best way would be to apply the following patch to
radiusd/radius.c:

Index: radiusd/radius.c
===================================================================
RCS file: /cvsroot/radius/radius/radiusd/radius.c,v
retrieving revision 1.77
diff -p -u -r1.77 radius.c
--- radiusd/radius.c	2 Nov 2004 08:55:39 -0000	1.77
+++ radiusd/radius.c	26 Nov 2004 08:48:12 -0000
@@ -473,13 +473,13 @@ radius_respond(REQUEST *req)
 
 	radiusd_sql_clear_cache();
 
+        /* Add any specific attributes for this username. */
+        hints_setup(radreq);
+
 	if (radreq->request->code == RT_ACCESS_REQUEST
 	    && rad_auth_check_username(radreq, req->fd))
 	    return 1;
 	
-        /* Add any specific attributes for this username. */
-        hints_setup(radreq);
-
         /* Check if we support this request */
         switch (radreq->request->code) {
         case RT_ACCESS_REQUEST:


and then to use the following rewrite function:

integer
fake_name()
{
        if (!*%[User-Name] || User-Name = "")
           %[User-Name] = "fake-user-name";
        return 0;
}

Regards,
Sergey