Design And Implementation Of A Testing Based Program Fault Localization System

Debugging is one of the most expensive and time-consuming processes for software developers, especially locating faults in programs usually costs most of time in the debugging process. Clearly, techniques that can reduce the time required to locate faults can have a significant impact on the cost and quality of software development and maintenance. In recent years, there have been a body of work on computer-aided software debugging both in the field of software engineering and artificial intelligence, and a variety of fault localization techniques have been proposed. The testing based fault localization (TBFL) techniques successfully locate faulty statements by comparing the statistical information of program elements in a number of program runs. However, there are still examples such that available approaches yield unsatisfactory results. To be an effective intelligent statistical method, TBFL has broader space for development.This paper surveys the state-art-of TBFL techniques, and presents an improved TBFL technique, which focuses on analyzing control flows and data flows instead of statement coverage. Specifically, we run the test suite gets the spectra of dynamic control and data dependence, build appropriate models with the two kind dependence program spectra to evaluate suspected statements, measured by suspicious scores of their relevant control and data dependences, and developers are then assisted with the ranking of suspicious statements to correct program errors.We degsin and implement a fault localization tool that use the runtime dependence information in program runs to locate the bugs. To check the effectiveness of our 'technique, Siemens benchmark has been performed and some others TBFL techniques have also been implemented for the purpose of comparison.. The empirical results show that our technique is more effective than some others, and the system is proved to have some practical value.