Automated test data generation and stubbing method for C/C++ embedded projects

Abstract

Automated test data generation for unit testing C/C++ functions using concolic testing has been known for improving software quality while reducing human testing effort. However, concolic testing could face challenging problems when tackling complex practical projects. This paper proposes a concolic-based method named Automated Unit Testing and Stubbing (AUTS) for automated test data and stub generation. The key idea of the proposed method is to apply the concolic testing approach with three major improvements. Firstly, the test data generation, which includes two path search strategies, not only is able to avoid infeasible paths but also achieves higher code coverage. Secondly, AUTS generates appropriate values for specialized data types to cover more test scenarios. Finally, the proposed method integrates automatic stub preparation and generation to reduce the costs of human effort. The method even works on incomplete source code or missing libraries. AUTS is implemented in a tool to test various C/C++ industrial and open-source projects. The experimental results show that the proposed method significantly improves the coverage of the generated test data in comparison with other existing methods.