Where Do Schema Come From?

[Klingenstein, Nate]

In the platonic sense, how does the IdP actually, functionally load schema for e.g. parsing attribute-resolver.xml files?  Does it use the schemaLocation URL defined at all, or is that just for human consumption?

I see "Update schema mapping redirect" all over the place in issues.shibboleth.net and a SchemaBuilder in system/conf/global-system.xml that appears to load schema from the classpath, but I can't find any cohesive description of what's going on.

Thanks,
Nate.