F18 Criterions/Testcases interconnection

[Josef Skladanka]

> Well I see your point, but it kind of cuts both ways - Josef clearly
> got
> confused by a few cases where we overload test cases to test several
> different things, so you can argue that it's actually more
> 'accessible'
> when we try to stick to 'one test case tests one thing'. But your
> approach would achieve what we need to achieve indeed.

/me feels a bit ashamed :D

I'm all for 'one test case tests one thing', even though I understand, that having several 'similar' testcases which differ in just few words is nonsense (as I got equally confused with the _almost the same_ criterions in alpha/beta).

J.