Handle SUT panic gracefully

[symbiont-daniel-gustafsson]

Currently if SUT panics the executor will also panic, it would be good to handle this a bit more gracefully.

The executor should detect that if SUT panics, if so it should return to scheduler with information to Scheduler that the SUT crashed. We can also create a new event, that is emitted from executor/scheduler when SUT crash.
The executor could try to get the last state of the reactor that crashed and emit that.
The scheduler could emit it current state, and then finish

All checkers should probably check if such an event was emitted, if so fail directly.