EncryptedKey placement configuration

[Brent Putman]

On 5/13/14 9:18 PM, Cantor, Scott wrote:
> Yeah, I thought maybe if the SAML encryption resolver subclass could
> create a subtype of EncryptionParams that might work, but it's not
> structured to do that now.

It's not right now.  But maybe, possibly it could work.  I'll think on
it some.   It doesn't really scale though.  Say we needed in the future
some SOAP-specific options.  So we make a SOAP-specific config, but then
madness ensues with what kind(s) of resolvers can deal with that, and
what to do about the params, etc.

I guess the other easier option would be to just expose either a
Map<String,Object> or a ClassIndexedSet or something on the configs, and
just have the resolvers mirror them straight over to the params -
logically merging the values from all the passed configs.  Or something
like that.  Then the thing using the params would have to know what to
look for.  Don't know what the Spring wiring for the config instances
would look like, though, esp for a ClassIndexedSet.  That doesn't sit
100% well with me though at the moment, although it feels more like the
loosely coupled direction we've been going in v3.


> I was hoping not, but I suspect if it's ever needed, it would be for an
> interop purpose, so that's where it would have to be.


Yes, agree, it would be nice to have it be relying-party-specific.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://shibboleth.net/pipermail/dev/attachments/20140513/e1df43b0/attachment.html