Research on infeasible branch-based infeasible path in program

Usually, it is a tired time-consuming process to consider path condition produced from an infeasible path. Analyzing these infeasible paths can not only save test resources but also improve test efficiency, at the same time, it helps to provide valid information to optimize program structure. Based on the Dominator Tree and valid information produced during the process of program compiling, this paper presents a new technique can reduce the cost to analyze infeasible paths. To find the infeasible paths produced by data dependency, the technique first refines the Program Control Graph (PCG) and determines the interesting field by immediate dominate node of a control node, then, in this field analyzes infeasible path by the demand-driven solution. And to find the infeasible path produced by predicate subsumption, the technique first constructs the correlative predicate information pool for correlative control nodes, and finds out infeasible branches of paths by these valid information obtained in the data flow dependency and control flow dependency. During the testing, if a testing path including an infeasible branch or more, the testing path is infeasible. The experiment results show that this infeasible branch method can find more infeasible paths in the unit time, and consumes almost the same memory as other methods.