Test Automation Meets Static Analysis

In this article we advocate an integrated approach for the automation of module or software integration testing and static analysis. It is illustrated how fundamental methods of static analysis, in particular abstract interpretation by interval analysis, contribute to the solution of problems typically encountered in the field of automated test case/test data generation. Conversely, test data generation algorithms are useful to improve results obtained in static analyses: Potential errors identified in the unit under test (UUT) during an analysis can be confirmed by constructing concrete test data leading to the erroneous UUT state. False alarms resulting from over-approximating abstractions applied during the analyses can be uncovered using test automation algorithms disproving the reachability of associated code portions and program states.