To test stability of Jemmy code it would be really useful to be able to run tests multiple times. It should be done so that results from multiple runs are stored separately.

I would be good if that is implemented with no dependency on external ant libraries or such dependency would be optional and only required if repeated run is performed.