Relying party parsing. Metadata providers

[Cantor, Scott]

On 3/20/14, 2:25 PM, "Rod Widdowson" <rdw at steadingsoftware.com> wrote:
>
>I should add that Ian has just reminded me of a further discomfort - the
>precise package included by idp-profile-spring is opensaml-saml-impl.  So
>this is doubly wrong.  There is an explicit derogation that idp-*-spring
>modules may depend on idp-*-impl modules, but it feels slightly icky to
>depend on an opensaml-*-impl module.

Yeah, but they have to, don't they? I don't think we should wrap the
classes there just to avoid that dependency.

>I'll look as see whether the RelyingParty and ProfileConfiguration need
>idp-saml-impl, and if not I'll suggest growing an idp-saml-spring class.
>Of
>course that then asks the question of whether other SAML specific
>configuration should go there (such as the encoders, the NameID attribute
>resolvers and so on).

I thought that was limited by the fact that we couldn't split the parsing
across two modules for the same namespace.

-- Scott