A Call Graph Mining and Matching Based Defect Localization Technique

Locating defects in the source code of a software system is one of the most challenging tasks in software debugging. Defect localization tools aim to assist developers in finding the location of defects. Both static and dynamic analysis approaches are used. In the case of dynamic approaches, two different scenarios apply. The first is one in which we have multiple (different) executions that exhibit the faulty behavior. The second is one in which we have just a single faulty execution. This is the focus of this paper. In this paper, we present a novel technique for localization of structure-affecting defects (i.e., defects that make an execution diverge from the expected path, thus creating unexpected dynamic call graphs), using tree mining and tree matching techniques. Given a target system and a failing test case, the proposed technique finds methods in the source code of the system which are likely to have caused the failure (i.e., defective) or lie on the same path in the dynamic call tree representation of the failing execution as the defective methods, and thus can be used as a starting point to find the defective method(s). The proposed defect localization technique is implemented as a prototype and evaluated using four subject programs of various sizes, developed in Java or C. Our experiments show comparable results to similar defect localization tools, but unlike most of its counterparts, our technique does not require the availability of multiple failing executions to localize the defects. We believe that this is a major advantage, since it is often the case that we have only a single failing execution to work with.