Shibboleth Developers

[Tom Scavo <>]

On Mon, 1 Nov 2004 16:49:35 -0500, Scott Cantor 
<>
 wrote:
> > In the SAML 1.1 spec, the target parameter is alternately referred to
> > as "TARGET" and "Target".  In Shibboleth, the parameter name is
> > "target".  Is this inconsistency cause for concern or do
> > implementations routinely work around case sensitivity of parameters?
> 
> There is no parameter in SAML that corresponds to the one in Shibboleth
> called "target" because that's a request parameter and there is no request
> in SAML. The one that is defined in SAML is called "TARGET" and Shibboleth
> matches this.

I must be missing something because SAML 1.1 *does* define
"TARGET/Target", which *does* appear to coincide with Shib's use of
"target" (in the new architectural doc).  The latter may originate at
the SP, but it eventually finds its way to the IdP, at which point the
name clashes with SAML 1.1.

> There's a non-normative bit in the artifact profile that mentions something
> called "TARGET" and "Target" in the same step, but it isn't a required part
> of the spec and is internal to the IdP anyway, since it doesn't carry enough
> information to drive a SAML response.

No, the "TARGET" parameter is *required* in steps 2 and 3 of the
browser/artifact profile of SAML 1.1.

> I would not personally write any code that tries to accommodate case in this
> area.

That's my point.  Shouldn't Shib adhere to "TARGET" (which is what the
SAML spec is trying to say, I think).  A small point, admittedly, but
it affects interoperability (if that's an issue).

> -- Scott

Tom