Fault Localization Using a Failed Execution Slice

Spectrum-based fault localization (SFL) is amongst the most promising methods. To further improve the effectiveness of traditional SFL methods, a novel approach is proposed in this paper for software fault localization by incorporating a failed execution slice (FES). The proposed method first calculates the suspiciousness scores of statements in the program under test using existing risk evaluation formulas. Since a selective failed execution slice can contribute to reduce the code detection scope in fault localization, a utility evaluation function f is then defined to choose a key FES from all failed execution slices. Finally, all statements contained in the generated key FES are sorted by their suspiciousness scores in descending order and then examined to locate the bug. Experiments on 11 well-known benchmark programs and 12 classic risk evaluation formulas are reported to demonstrate the effectiveness of our proposed method.