Hello, Let's define this query: authorize_reply_query = "\ SELECT 1, '%{SQL-User-Name}', 'Reply-Message', 'Hey', ':=' \ UNION \ SELECT 2, '%{SQL-User-Name}', 'Reply-Message', 'Hello', ':=' \ ORDER BY 1" In radiusd -X output, on may then read: Sending Access-Accept Id 201 from 127.0.0.1:1812 to 127.0.0.1:62417 Reply-Message = 'Hey' Reply-Message = 'Hello' A Reply-Message may indeed appear multiple times in an Acces-Accept (or -Reject or -Challenge) packet. Now, according to http://wiki.freeradius.org/config/Operators, operator ":=" has the same meaning as a reply item than as a check item. As a check item, it "replaces in the configuration items any attribute of the same name. If no attribute of that name appears in the request, then this attribute is added." How should that idea of replacement of "any attribute of the same name" then be interpreted? TIA, Axel
Axel Luttgens wrote:
authorize_reply_query = "\ SELECT 1, '%{SQL-User-Name}', 'Reply-Message', 'Hey', ':=' \ UNION \ SELECT 2, '%{SQL-User-Name}', 'Reply-Message', 'Hello', ':=' \ ORDER BY 1"
In radiusd -X output, on may then read:
Sending Access-Accept Id 201 from 127.0.0.1:1812 to 127.0.0.1:62417 Reply-Message = 'Hey' Reply-Message = 'Hello'
The issue is that the operators don't look at the new attributes being added. They only apply to the attributes in the original list. Alan DeKok.
Le 21 août 2014 à 13:45, Alan DeKok a écrit :
[...] The issue is that the operators don't look at the new attributes being added. They only apply to the attributes in the original list.
Hello Alan, Many tanks for your reply. Should "issue" be understood as "explanation", so that one is facing an intended behavior? Or does "issue" really mean "issue"? If the latter, I am correct at believing that operator "+=" would be logically better suited for building those duplicate Reply-Messages? Axel
Axel Luttgens wrote:
Le 21 août 2014 à 13:45, Alan DeKok a écrit :
[...] The issue is that the operators don't look at the new attributes being added. They only apply to the attributes in the original list.
Should "issue" be understood as "explanation", so that one is facing an intended behavior?
It's intended behavior.
Or does "issue" really mean "issue"?
It means that there is an *underlying* reason for the behavior. If we were to change the way it worked, we would get *another* kind of odd behavior. There is no "perfect" solution. Alan DeKok.
participants (2)
-
Alan DeKok -
Axel Luttgens