BannerXE, unscoped entityid

[Cantor, Scott]

On 11/21/16, 7:46 PM, "users on behalf of Jeffrey Crawford" <users-bounces at shibboleth.net on behalf of jeffreyc at ucsc.edu> wrote:

> This would not be able to be in the InCommon metadata obviously, but could we run into trouble otherwise?

Assuming you trust all your other metadata sources not under explicit control not to register such an ID, then you're in control of whether you violate the uniqueness constraint or not.

The deeper concern is that as a SAML implementation, I'm not prepared to promise that Shibboleth will always and forever permit invalid entityIDs. It does today, and it probably always will, but it doesn't have to, and somebody needs to tell these guys to fix their code.

-- Scott