Add custom attributes to assertion

[Cantor, Scott]

On 10/11/17, 4:30 AM, "dev on behalf of David Huebner" <dev-bounces at shibboleth.net on behalf of david.huebner at daasi.de> wrote:

> I understand "arbitrary" as in there are no hard restrictions on what I 
> can use these attributes for.

There was no intention that those attributes be used to represent *values* of the Attribute. They're for meta-information about the Attribute itself. That's how they're used (and nothing actually uses that extension point, and nothing but Shibboleth even supports it notionally).

> So it's zero or more.

I misremembered that, it must have been fixed in 2. I know SAML 1 required them.

> Which does say nothing about the opposite (i.e. an attribute has values) case.

That was the implication. No AttributeValue means no value.

-- Scott