Path sensitive analysis based on dataflow information and program slicing

摘要

In the process of program analysis, the error detection usually gathers the information of the target position from the path sensitive analysis, which would include some useless dataflow information, thus leading to low analysis efficiency. In this process, the states of the target position usually are combined to solve the path explosion problem, but this would result to high time consumption and fault positive and fault negative. Aiming to this problem, a method based on dataflow information and program slicing is proposed to deal with the path sensitive analysis. This method creates the slicing criteria via path condition, and obtains the sliced program which is equal to the source program, which can improve the analysis efficiency. We have implemented the analysis demon and the experimental results demonstrate that this method has high detection efficiency and can reduce the fault positive.