Shibboleth Developers

["Howard Gilbert" <>]

> 
> What we need to test:
>   - Individual classes, or small groups of highly inter-related classes
>   - Subsystems (e.g. the AA, AAP, Trust engine)
>   - Round trip functional test (e.g. making an attribute query and
> getting a valid response)
> 
> Does this accurately cover what needs to be tested?

In some cases the round trip test is functional. I would describe this as a
test that a particular attribute that you feed in at the beginning goes
through ARP, SAML, AAP, and the Filter and gets mapped into the dummy
"shib-foo" HTTP header as seen by the Resource at the end of all that
processing.
 
However although it would still be called a functional test, I would like to
place some emphasis on the end-to-end test as also testing Protocol
sequences. My first effort was to create examples of all the major sequences
in the Protocol spec, and it may be possible to develop tests driven by
specification elements. We can even test edge conditions in the spec (say
where the Attributes were pushed but for some reason the SP decides to do an
AA query anyway.