Automated test case selection.

Abstract

In this thesis, we present an automatic test case selection system, which aims at selecting test cases for both control and data flow oriented testing. In our system, path expressions are employed, path expression generation and complete path selection algorithms are used to achieve the automation of path selection, and problems with loop iterations are addressed. Partial symbolic execution is applied to reduce the effort wasted in analyzing infeasible paths. Linear programing, path predicate simplification, and path predicate evaluation techniques are employed to solve the inequalities resulted from partial symbolic execution as well as symbolic execution. Our system is also capable of selecting different sets of test cases by applying different test selection criteria to achieve different test coverage which serve different test purposes. In our system, first a flowgraph modeling both control and data flow information contained in the program under test is constructed. Second, by applying a control or data flow oriented test selection criterion to this flowgraph, a set of test units is built. Finally, through path expression generation, partial symbolic execution, symbolic execution. predicate simplification/evaluation, and linear programming, a set of test cases is selected examples produced by ETSG are provided.