Software Bug Localization Based on Key Range Invariants

Abstract

Bug localization is expensive and time-consuming during software debugging. The traditional software bug localization based on range invariants need to monitor all variables in the system, requires a large of runtime overhead. However, this overhead is not necessary. Because only a set of key variables can really affect the results of the system. Therefore, this paper proposes a software bug localization method based on key range invariants. First, add the key variables screening phase in the original method. By combining the dynamic filtering mechanism with the static reduction mechanism, the key variables set of the program are screened. Then, the values of the key variables in all successful test cases are counted to obtain the key range invariants. Finally, bug localization is performed by monitoring the values of the key variables in failure test cases. When we need to minimize the overhead of monitoring variables, we can use this method to ignore variables that are considered unimportant. The experimental results show that, the method can still maintain a good bug localization effect only monitoring the key variable set, which verifies the effectiveness of the method.