Extending the SP Dynamic Metadata Provider

[Cantor, Scott]

On 10/27/17, 10:57 AM, "dev on behalf of Tom Scavo" <dev-bounces at shibboleth.net on behalf of trscavo at gmail.com> wrote:

> This would allow us to deprecate (but not disallow) file:// URLs.
>
> This feature is intriguing but I still don't understand the use case.

Testshib was always a big one, but another is for curating local/manual sets of metadata from vendors. I have a Partner file containing dozens of manually created metadata instances from vendors, ADFS, etc., and it's a pain in the ass to have to combine it into one file by hand. If I could drop them in separately, I'd have less screwing around with XML to do. That's the IdP use case of course, but one could imagine a use case for the SP dropping one-off IdPs in. Mostly it's just for consistency.

> How do files end up in this directory and what is the file naming convention?

Ideally there isn't one and it simply monitors the directory and loads what's there, but that's work.

> What file naming convention will be used in the cache dir?

Unspecified, I would hope.

> What about discovery? What's your discovery strategy in the face of per-entity metadata? Seems like that should come first.

Not in our power unless we're going to build a new solution and we're not doing that with the current make up of this project. Many SPs don't need discovery so it's not a precondition to what we're doing or what I already did.

-- Scott