Lets tests with 'testability=0' return 'cantTell' or notApplicale'

[hannolans]

We have tests that can't tell if something is implemented right or wrong. At the moment QUAIL returns a testability=0.
But in that case it's still interesting that there is a certain situation. Therefore the test can return one of these two outcomes:

• cantTell, if the situation occurs, but we don't know for sure if it passes or fails
• notApplicable, if the situation doesn't occur, so no additional manual testing is needed

[jessebeach]

This will need to be implemented on a test-by-test basis. I'll need to first compile a list of tests that have a testability of zero, then of those, determine which ones can be marked as cantTell and which as notApplicable -- one, the other or either-or depending on the test.