Random order of test execution #422
On of the principles of xunit test approaches is that tests should be independent and so relying on ordering of tests (outside of before or beforeall annotation) is considered "a bad thing".
Maintaining independent tests means that it is easier to diagnose failed test since the programmer only needs to look at a single test scope to understand everything needed to diagnose.
Enabling this capability feels like a change in direction of ethos of a test suite.
Just my 0.2c
That is why it should be a runtime parameter, not an annotation I think.
I love the idea of random or even simply reverse order of execution in suite.