I agree here, the class comparison is not nice :).

[DurieuxPol]

          I agree here, the class comparison is not nice :).

If tomorrow we subclass it or add a class with similar behavior, this code will change many many times.
Better to delegate to the filter as @PalumboN says.

About the two options, doing it inside filterTests: or validateFailures:, I'm ok with any.

Now looking at the code I'm wondering, does it do what it should? :D
The thing is:

• if I have 10 tests out of which 1 is red
• with the red test filter we should have 9 tests for the analysis, and 1 test ignored

Which means that:

1. the filter cannot be just on the "test references" but on the test + result => I see that the tests have a lastResult, so this is ok!
2. If, after running the filters, (and whatever the filter we have) there were broken tests, we should throw the exception => This means we can get the code of the error outside the filter, right?

Originally posted by @guillep in #96 (comment)

[DurieuxPol]

Move the red test validation in analysis, because if there are red tests they will be filtered anyway