delegation in IDP3

[Brent Putman]

On 6/9/16 3:56 PM, Cantor, Scott wrote:
>> Brent's message from a while back that I cited earlier explicitly
>> mentioned looking at the schema (in the absence of documentation)
>> to figure out how to do this in a legacy file, so I inferred that
>> the intent, at least, was to support this in a legacy configuration.
> I'm pretty much lost so am going to shut up and basically say I have no idea. I filed a bug and if it's not a bug, they'll close it. The code doesn't look like I think it would have to, but there are comments in it that make me think I no longer understand it like I thought I did, so I'm not going to try until Rod or Brent have a chance to weigh in on it.

It's possible that the legacy parser stuff is broken in some way for the
delegation related things.  However, there are unit tests for all of
that, and they pass.  I recall running them when I made the minor schema
changes to add some stuff.  It's possible the unit tests are somehow
broken too, in a consistent way, but I'm not seeing it.

Fwiw, this is what the passing test has:

<rp:ProfileConfiguration xmlns:rp="urn:mace:shibboleth:2.0:relying-party"
    xmlns:saml="urn:mace:shibboleth:2.0:relying-party:saml"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns:spring="http://www.springframework.org/schema/beans"
    xsi:schemaLocation="urn:mace:shibboleth:2.0:relying-party
http://shibboleth.net/schema/idp/shibboleth-relying-party.xsd
                        urn:mace:shibboleth:2.0:relying-party:saml
http://shibboleth.net/schema/idp/shibboleth-relying-party-saml.xsd
                        http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans.xsd"

    xsi:type="saml:SAML2SSOSProfile"
    delegationPredicateRef="jimmy">
</rp:ProfileConfiguration>


> The bottom line is that it doesn't make any sense to me to support a legacy syntax that isn't actually the legacy syntax you would already be using. And it doesn't support that one, for sure, so whatever you have to change it to, no idea...

Yeah, I agree it's not really "legacy".  But we shipped the schema for
3.0 with this stuff in there, so that probably changes things.