Function use in the IdP

[Rod Widdowson]

> Or duh - just fix set*MessageContext on InOutOperationContext to call
setParent(this). 

I had thought about that, but for some reason I was convinced that this
tried to insert the child into the class hash of the parent, but I see that
it doesn't.

Doing this removes a level of reflexive-ness (is that a word?)between
child.getParent(), and parent.getSubContext(), but that is a philosophical
issue.  However, I wouldn't be against this, since it makes life easier for
people with weird conifgurations, but I would hope that it would remain an
edge case.

But.... OOTOH it appears from what you are saying that there are times when
an InboundMessageContext will not have a PRC.  In that case, if we have a
need to call a scripted Function or Predicate before the parent PRC is set,
then we are SOL.  I don't know how realistic that case is, but all in all
it's probably easiest to just add a new scripted Function<MessageContext,
Object> and be done with it.

I've added IDP-512 to capture all the work items being engendered.  I hope
to have details for injectable Functions in OpenSAML and Predicates in both
by Monday at the latest.

Rod