Use of classpath*: / Windows

[Tom Zeller]

> I know that at one point I located the bit of code that was breaking when I reverted this to use classpath: in Main.java and undid the Spring context class change you made. And I know it had something to do with loading additional web flows in a test somewhere. But at the moment I can't find it again.
>
> The idp-conf tests all run without classpath*. So this must be an integration test thing?

Yes.

To test the local storage read and write subflows, I "needed" test
flows which call those subflows. With idp.home = classpath*:, those
test flows can reside in idp-conf:/src/test/resources/flows. Which I
think would help anyone writing a custom flow.

I put quotes around 'needed' because, of course, in the idp-conf tests
we create a flow repository/registry (I forget which) specific for
each test. However, for local storage, I want to test using a browser,
which we don't do in the idp-conf tests. I guess we could, but so much
would get mocked up I don't know if it would be worth the effort. Or,
we could remove the impediment to running subflows directly.

There are other ways to do what I did, which essentially is just
loading a flow into the IdP. But it seemed like supporting classpath*:
was a good-idea, especially if we want to move system flows into a
JAR.

But, the base-path in the SWF repository/registry is weird, those
locations aren't resources. I'd have to work through the details again
to say more.