Research On Defect Localization Of Program Slicing Based On Spectrum

Software plays a great role in computer systems,and the reliability of software products is very important,so the precise positioning of the hidden defects in the computer software can guarantee the quality of the software.However,nowadays,with the computer system getting more complex and larger,the existing software defect localization technology is facing severe challenges.How to design and improve software defect localization methods effectively is one of the problems to be solved in software testing.In view of the shortcomings of existing methods,this paper proposes a program slicing defect location framework based on spectrum on the basis of previous work,which is used for the research of software defect location.A program slicing defect location method based on spectrum is to locate the defect function that causes the program abnormal by using the dynamic spectrum information obtained by executing the program.As a spectrum based program slicing defect location method needs to cover a large number of paths,a large number of test cases will be generated.In order to increase the efficiency of the defect location,first of all,we cluster the test cases with similar execution failure to get the example set of similar error testing,and then use the program dynamic slicing algorithm to construct the test case.A function crash table is built to slice the granularity of the function block to obtain the defect skeptical function set.Finally,the dubious algorithm is designed to sort the suspicious function within the range of the defect path,and finally the most defective function is located.The program we used was Siemens with defects which is input manually.With the function call relationship of the source program which is automatically generated,we used the cross table method to compare with the algorithm of Tarantula,Ochial and Jaccard.The result of experiment showed that the proposed defect location method is better than the above algorithms.At the same time,this method can reduce the scope of code review effectively and improve the accuracy and efficiency of software defect location.The result also provide a new idea for the implementation of automated software testing defect location.