Research And Application Of Software Fault Localization Technique Based On Optimized Test Suit

Software debugging is an important link in the entire software development process,but usually it takes a lot of effort and time to find the location of the fault.Software fault localization technique enables fast fault localization,saves developer workload,and improves software development efficiency.Spectrum-based fault localization is one of software fault localization techniques.Developers can realize rapid fault localization with providing the fault test suites.But for spectrum-based software fault localization,different quality test suites will bring different impacts.A high-quality test suite not only enables developers to locate the fault location more accurately,but also saves time for locating.This thesis conducts research based on two types of problems that appear in the test suite and the conclusion is verified on the Defects4J dataset.Firstly,for the problem of redundant information in the test suite,this thesis uses the failed execution set on the spectrum information to simplify the redundancy in the spectrum dimension,and then uses the clustering algorithm to remove the redundant information from the successful test cases.Compared with the PTD-TO algorithm and various clustering algorithms,the de-redundancy strategy based on Chameleon algorithm proposed in this thesis not only has higher accuracy,but also maintains better efficiency.Secondly,for the problem of too few failed test cases in the test suite,this thesis proposes an adaptive weighted fault location algorithm based on the problems of the cloning strategy.Through comparative experiments,the algorithm can achieve a more efficient localizing while improving the accuracy than the cloning strategy.Finally,this thesis combines the optimization test suite strategy and fault localization technique to design and implement a fault localization system,and completes a user-friendly interactive interface for fault localization.