nameid-format:unspecified for relying party

[Cantor, Scott]

> Am I misunderstanding though that I should be able to specify the
> NameIDFormat via a RelyingParty override rather than specifying it in the SP's
> metadata?

Yes, as long as you're not trying to use the one "format that isn't a format" that people keep trying to use that's referenced in the subject line.

> Per your ArcGIS documentation this works as expected.

Being out of position in the metadata tends not to matter as much to the IdP if it's not schema validating, but I would be surprised that quoting it like that worked. That would be an interesting bug.

> It generates a transient format NameID in the Subject once again. The
> documentation I'm using as references describes the RelyingParty
> nameIDFormatPrecedence attribute as "A space delimited, ordered list of
> name identifier formats" [1] and the logic (I think) to select the
> NameIDFormat [2][3].

They are equivalent and if you get different results, there's some fact not in evidence that means it isn't using the configuration you think it is.

-- Scott