Automated Debugging System: Implementing Program Spectrum Analysis and Information Retrieval on Fault Localization
Yin-Siang Hao · International journal of high school research · 2023
Debugging is often the most time-consuming phase during program development, lengthening the development time and lowering efficiency.Eventhough there are currently existing tools that help raise debugging efficiency, their functions are largely limited, as they only present a more comprehensive analysis behind the compiling and running processes to save complicated steps in debugging; however, developers are still required to test and verify the reasonability of every line of code.Unlike the current manual debuggers, this research aims to build a system that automatically detects bugsa in the programs through program spectrum analysis and information retrieval.In the section of program spectrum analysis, the system will statistically analyze every code block's probability of containing bugs in the input source code file according to the provided test cases, later forming an initial suspiciousness ranking based on previous calculations.Afterward, the system retrieves historical files that resemble the current source code and uses their suspiciousness rankings to modify the initial suspiciousness ranking, generating the final suspiciousness ranking as the system's output.This research integrates the two techniques and optimizes the formula of forming the suspiciousness values in program spectrum analysis and the comparison mechanisms in information retrieval, reaching performances of higher comprehensiveness and precision.a.The word "bugs" here refers to logic errors that cause wrong answers or runtime errors, not syntax errors that lead to compile errors.