A Study To Improve CFL Based On TheDiscovery Of Coincidental Correctness Test Cases

Software fault localization is a significant research topic of software engineering field.Coverage-based fault localization (CFL) is an important kind of approach to locate faults insoftware; some articles proved that CFL is also effective when there are multiple faults inprograms. The Coincidental Correctness Test Cases bring lots of negative influence to CFL;therefore the way to find or prevent these test cases has huge significance to improve theperformance of CFL.This article discovered an approach to find coincidental correctness test cases with nonefalse positive by analyzing the influence brought by coincidental correctness test cases.Moreover, this article studies on the way to modify CFL after the coincidental correctness testcases are founded, then three approaches of CFL based on the discovery of coincidentalcorrectness test cases were proposed. The first one is proposed by finding the short of the existedmethod, based on the value IDLO(V)/WIDLO to improve the CFL method. The second one is proposedby analysis of the correct statement, finding a way based on the value MYCC(s)/TMYCCtolocate the faults. The third one is to use my method combined with another method, which candecrease the correct statement’s suspiciousness, to locate the faults.Through the experiments on Siemens test suits with122fault versions, it shows that allthree approaches can generally improve the existed CFL’s results. In additional, the emphasesand difficulties of building the experimental environment are in this article, they may helpresearchers to dig in fault localization field or modify existed platform.