dce/rpc "client" api

[Andrew Tridgell]

> yes, of course.  however, it seems that you have both missed the point of
> the "logical extreme" i am describing.
> 
> on the basis the argument that jeremy is applying to justify that
> NET_USER_INFO_3 should be removed, surely the presence of, and the code
> that manages, user_struct should _also_ be removed?

that's an "illogical extreme" :)

the vuid is a token that allows access to a user_struct, much like a
cnum gives access to a connection structure. If you deleted the
user_struct management code then the vuid would be useless.

or maybe logic doesn't work quite the same way in my cubical as yours :)