On 3/15/07, Emmanuel Lecharny <[EMAIL PROTECTED]> wrote:
On 3/15/07, Enrique Rodriguez <[EMAIL PROTECTED]> wrote:
> ... the issue is that the base DN for Kerberos searches
> has to exist before the Kerberos PP starts.
Oops, I was not aware of that... So the next sentence is quite right :)
> Of course, this is
> probably bad design and has to be fixed for multi-realm capability.
+1. In 1.5.1 or 1.5.2 maybe ?
OK, put me down for fixing that in 1.5.1. It is a quick lazy-load of
the initial context.
> Today the problem is that the Kerberos PP starts on the first call to
> ServerContextFactory. As Alex noted, this leaves LDIF is the only way
> to load prior to protocol startup. So, without an LDIF those ITest
> examples don't help because I'm doing somethings with Kerberos.
Ok, get it. I don't really like having KDIF files all over, so what I do is
generally to create them on the fly in the setup, using a tmp file. I find
it easier to manager than to have a separate file (you have all the test in
front of you instead of having to look in a resource sub directory, not to
mention the intricacy of class loading issues ...)
Good idea, for both best-practice reasons: encapsulation w.r.t. the
test class and class-loading issues. That makes me feel better about
using LDIF in a test.
Enrique
