Spectrum-Based Incremental Defect Location Method

Abstract

The invention discloses an incremental defect positioning method based on a frequency spectrum. The method adopts an incremental stepwise iteration method, and finds a program defect statement or predicts a range in which the defect statement possibly exists by employing program information and test information. The method comprises the following steps: firstly, running a test case, collecting coverage information, i.e., frequency spectrum information and operation result information, which is executed by the test case in a program to generate a coverage information table; then, carrying out statistic analysis on the coverage information table, calculating the suspicious degree of the program statement, and sorting the program statement according to the suspicious degree to obtain a defect positioning sequence; and according to the statement arrangement in a positioning sequence, carrying out troubleshooting one by one until the statement which causes program exception is found. According to the method, by adopting an optimizing strategy from multiple angles, the software defect positioning efficiency is improved, the coverage rate of the test case is increased, the cost for frequency spectrum collection is favorably reduced, and the accuracy of a suspicious degree algorithm is improved.

Description

technical field [0001] The invention relates to a spectrum-based incremental defect location method, which belongs to the technical field of software testing. Background technique [0002] Over the years, people have proposed many methods in the study of defect location, mainly using the static information and dynamic information of the program to locate program errors. However, the overhead of obtaining static information is relatively large. For large-scale software, comprehensive static analysis is even unrealistic, while the collection of dynamic information is mainly to run test cases, which will not bring too much overhead to the test. At the same time, since the dynamic information includes the information when the program is running, it can provide more accurate results compared with the method using static information. [0003] Using program spectrum information to locate defects is a more practical and effective software defect location method at present. Program ...

AI Technical Summary

Problems solved by technology

However, the overhead of obtaining static information is relatively high. For large-scale software, comprehensive static analysis is even unrealistic, and the collection of dynamic information is mainly to run test cases, which will not bring too much overhead to the test.

Images