Towards Discovery of Differences of Source Code Through Semantic Difference Checking

Abstract

Source code difference analysis plays an important role in software development, maintenance and regression testing. Many existing research methods are text-based, however, they are subject to code movement and alignment limitations, resulting in imprecise difference analysis. In addition, some methods are lexical and syntax-based, which ignore the key semantic information of the code and make the analysis results imprecise. In this paper, we propose a source code difference analysis technique based on semantic difference checking to improve the problem of inaccurate difference analysis. We implement the prototype, and exploit static analysis and textual difference detection tools to pro-process the programs, then perform semantic difference checking on the code with textual differences, and define the semantic differences. We conduct experiments on two C programs with differences and the results show that our approach can effectively and accurately detect code differences.