Algorithm Of Multiple Faults Localization In Software Debugging

Testing, debugging and verification of software accounted for the vast majority of the cost of the software development process. This is largely due to the process requires a lot of staff to take part in, especially diagnosed with the faults which lead to the execution of test cases failed. Locating the faults in software has a direct impact on the effectiveness of software and correcting the faults in software has a direct impact on the reliability of the software. Therefore, technology which can automatically diagnose faults is bound to help developers improve the quality of the software, and significant savings in development costs. In recent years, many experts and scholars engaged in research of fault localization technology, and propose many methods which are widely used in the field of software engineering automation. Among them, Spectrum-Based the Fault the Localization is a lightweight, diagnostic technique. By running a large number of test cases, SFL locates false statements by comparing the difference between the spectrum of successful test cases and the spectrum of false test cases. SFL can cope with real programs, and thus shortens the cycle of testing, diagnosing and repairing by reducing the time overhead of the debugging process.This paper summaries the development of SFL technology, proposes an improved multiple faults localization technique that utilizes program’s run-time dependencies like data dependence and control dependence, instead of the statement coverage information. The process goes as follows:we first get the dependence information for a dynamic program by running the test case, and establish a program model based on the dependence information, then calculates the suspicious value of each dependence, finally assign the suspicious value of each statement using the proposed method to help locate the multiple faults in program and to modify the faults.In this thesis, we implement the multiple faults localization algorithm on the basis of WET tools and LOUPE code. Running our algorithm against the Siemens suite, we do comparison with existing fault localization technology to analyze the effect of the proposed method. The results show that the proposed method works well, and has some practical value.