Improved Method to Generate Path-Wise Test Data

Gupta et al., proposed a method, which is referred to as the Iterative Relaxation Method, to generate test data for a given path in a program by linearizing the predicate functions. In this paper, a model language is presented and the properties of static and dynamic data dependencies are investigated. The notions in the Iterative Relaxation Method are defined formally. The predicate slice proposed by Gupta et al. is extended to path-wise static slice. The correctness of the constructional algorithm is proved afterward. The improvement shows that the constructions of predicate slice and input dependency set can be omitted. The equivalence of systems of constraints generated by both methods is proved. The prototype of path-wise test data generator is presented in this paper. The experiments show that our method is practical, and fits the path-wise automatic generation of test data for both white-box testing and black-box testing.