A Novel Fault Localization Method with Fault Propagation Context Analysis

A variety of Graph-based fault localization methods are always applied to abstract the relationships between program entities, and thereby facilitate program debugging and understanding. They assess the fault suspiciousness of individual program entities and rank the statements in descending order according to their suspiciousness scores to help identify faults in programs. However, many of these methods focus on assessing the suspiciousness of individual program entities while ignore the propagation of infected program states among them. They could not locate the true fault. In this paper, we consider the fault in a statement may be propagated to its subsequent statements via its data flow edges. We propose a novel CDN-based fault localization method. It includes two steps. First is fault-related statements localization and second is fault comprehension. It calculates the combined dependence probability of each statement to find the fault-related statements and then analyzes the propagation contexts of the statements to locate the true fault. The experimental results show that our approach is superior to other fault localization methods on both localization effectiveness and stability.