On adaptive sampling-based testing for software reliability assessment