Spectrum-based Multiple Fault Localization

Fault localization is one of time-consuming and labor-intensive activity in the debugging process. The traditional spectrum-based fault localization technology can be classified into two categories: based on component and based on statement. The former is too coarse to locate the accurate place; while the latter is too fine to decrease the computation complexity. A good software fault localization technology can save funds, save human resources, and accelerate the progress of software development. Therefore, how to solve these problems to improve the positioning has a great significance.In this thesis, we propose the method of fault localization based on Double-Times-Locating strategy(DTL). The main idea of this method consists of two parts: for the first time locating, we use existing tools to generate a program function call graph, build program spectrum from the information of function coverage that can get the relationship with the faults, and then use model-based diagnosis(MBD) to calculate with probability possible functions candidates based on logical reasoning and behavior proposition pattern, and sort of all. For the second time locating, we use the technology of fault localization based on DStar to locate faults in the statements of the candidate sets. We conduct experiments using Siemens suits, and compare the results of traditional methods which are less effective than our method. In addition, this paper also introduces the building experimental platform that can help readers to better understand.The contributions of this paper are listed below:·We propose a new idea which uses a function call graph to locate software faults, because it can reflect the real internal relations of functions;·We propose to use Double-Times-Locating strategy to locate software faults, and we compare it with the traditional spectrum-based fault localization technology, validate its feasibility and effectiveness;·We build experiment platform of Double-Times-Locating strategy for fault localization to facilitate the traditional spectrum-based fault localization technology for comparison and analysis.