LAUSR

With the development of information technology such as cloud computing, IoT, etc, software becomes the infrastructure. On the one hand, it is critical to ensure the reliability of software, on the other, sample code can be mined from open source software to provide reference for automatic debugging. Most of existing automated debugging researches are based on the assumption that defect programs are nearly correct, therefore they can successfully pass some test cases and fail to execute others. However, this assumption sometimes does not hold. In view of the fact that a programs may fail for all the given test cases in the test suite, but there are a large number of examples available for reference, a sample based fault localization method is studied. A fault localization method by analyzing the context of failure propagation is proposed, which locates suspicious statements by identifying the execution status and structural semantics differences between the defective program and sample program. Through the interactive analysis of value sequences and structure semantics, the influence of code variations and failure propagation is reduced. The experimental results have shown that the method can effectively locate suspicious statements and provide assistance for defect repair when there are enough sample programs.