Assertion is not always printed for non-deterministic tests

[furuholm]

For tests that are not deterministic the cause of the failure will not always be printed. My guess is that this is caused by the fact that after shrinking, RapidCheck reruns the smallest failing test case and uses this for printing. If the failure is not reproduced during this run, the assertion will not be printed. This also happens when noshrink is enabled.

[furuholm]

Maybe this could be solved by implementing #99?

[emil-e]

I think you could solve this by saving the last seen actual failure result and simply use that. However, the problem is that shrinking will be slower since you will have to calculate the counterexample even if there might be an even better counterexample.

In this case, I'd lean more towards actually running the final test a bunch of times until it fails again :)