Research On Defect Location Method Based On FCP Path Slice

Software has become the main enabled component in more and more system, the reliability of software products related to people’s livelihood, ensure software quality must be timely and accurate positioning the hidden faults. With the rapid development of the scale and complexity of software systems, the existing software faults localization technology is facing severe challenges. How to effectively improve the automatic faults location method to quickly and accurately locate the faults in the software is a problem to be solved in the field of software testing.The paper based on the technology of Function Calling Path(FCP) to research the automatic faults location method. Static analysis and dynamic analysis of source program to extract the FCP. Based on the failure principle and the differernce of executed paths to design path slicing algorithm and obtain the faults suspected sequence of the system. By calculating the suspect rate of the faults function, analyzing the correlation between the faults, constructing the faults correlation topology to show the faults location results.The execution track of software system can be abstracted as a FCP. A single execution can trigger multiple faults, a faults can also be performed by different executions. The successful executions of the potential of the faults, while the failure executions must contain faults. Based on the failure principle and the difference of the executed paths to design path slicing algorithm. The algorithm gradually reduced the faults location space from the FCP to the path slices then to function nodes, layers of progressive and finally obtained the steady faults suspected sequence.The "ripple effect" of the faults determines that they are not independent, but related to each other.This paper considered the path matching and statistical analysis probability to design the faults suspicion rate formula to calculate the fault suspicion rate, and introduced the association rules of frequent pattern tree(FP- tree) from Data Mine to analyze the correlate degree between the faults. Based on the suspicion degree and correlation degree of faults to build the faults correlation topological graph for performing the positioning effec of associated faults.Finally,this paper designed function-based ranking evaluation method and setted the single fault location contrast experiment and the multi-faults related verification experiment to verify the feasibility and effectiveness of the locating method proposed by this paper. Experimental results show that the FCP path slicing method proposed by this paper can effectively reduce the code review scope when fixed the faults, improve the accuracy and efficiency of the software faults location, and the research results also provides a new thinking of the automated software testing.