Study On Defect Localization Based On Spectrum

Software debugging can be divided into four steps:defect reproduction,defect location,defect repair and defect verification.Defect location is the most difficult and time-consuming step.Therefore,an effective software defect localization method is helpful to find potential problems in software and improve the reliability of software.At present,in the research of defect localization based on program spectrum,the effectiveness of existing defect localization methods still has room for improvement.On the other side,the application of reduction strategy to regression testing is less.In order to solve this problem,this paper carries out relevant research from three aspects:defect localization method,use case reduction and the influence of spectrum factor on defect localization.The main work is as follows:1.Spectrum-based defect localization method:At first,we analyze the influence of the number of test cases(spectrum factor T_ep),with the passed execution results and cover statement s,on the defect localization effectiveness.Then,an effective defect location method is proposed,which reduces the influence of excessive T_epp on suspicious values and improves the effectiveness of defect localization effect.2.Study on test cases reduction strategy:In-depth studying of reduction strategy VR,we found that an effective defect localization method is less affected by redundant test cases.The reduced test cases set generated by reduction strategy VR is applied to regression testing.Empirical research shows that the reduction method can greatly reduce the cost of regression testing on the premise of ensuring the effect of defect localization.3.Research on the effectiveness of spectrum factor on defect localization:In order to highlight the importance of test cases(spectrum factor T_nf),with the failed execution results and cover statement s,a new defect localization method is proposed.By analyzing the influence of the number of different spectrum factors on the defect localization,a Kulczynski coefficient is obtained when the defect localization method achieves the best effectiveness.