The reason I'm asking this is that I have several cooperating sites that
want to allow each others users to login on their pms.
I was thinking of doing it this way:
1. if the login-id is foo-user, and `foo-user' isn't
a valid user, `foo' is looked up in a table of
string-IP pairs, and a radius server at the IP
corresponding to `foo' is queried for info on
the user `user'.
2. if the server gets an Access-Accept back for `user', it
sends the info back to the original client
I'm assuming that this has been done before, but I just want to make sure
that portmasters work the way that I think they do before I start writing
code.
Are there any pitfalls/restrictions that I should watch out for? For
example, what is the limit on the length of usernames that a pm will
accept?
Also, if someone has a patch for radiusd that does this, I'd be
interested in that also.
Thanks,
Ron
-- Hagbard@LambdaMOO % I mean, it's the 90's. You can't just go Ron@JaysHouseMOO % around trusting every layer you interface Ron@MediaMOO % with. - tommyd@microsoft.com <A HREF="http://www.nmia.com/~tapia/">My home page</A>