Function use in the IdP

[Brent Putman]

On 11/17/14 12:30 PM, Cantor, Scott wrote:
> On 11/17/14, 5:28 PM, "Brent Putman" <putmanb at georgetown.edu> wrote:
>> Yes, of course.  So we're talking about a scripted way to change the
>> actual configuration, not scripted AttributeDefintion, etc?   If so,
>> then it makes sense now.
> It's a matter of perspective, I guess. Most of the functions run at 
> runtime, but they're injected at configure time. That's not really any 
> different from the AttributeDefinition case.



But that's my point though (and confusion).  By the time something like
a scripted AttributeDefinition gets hold of say a MetadataResolver, it's
already initialized.  So it can't change it.   I suppose it could in
theory build a whole new instance of a MetadataResolver - but I'm not
clear 1) why on earth anyone would want to do that - that's what
configuration is for; and 2) where/how you could use it and store -
don't think it could be persistent (other than in the PRC somewhere), so
you'd be building a new one on every request.

So I guess I'm still not clear what we're talking about here....