Performing high efficiency source code static analysis with intelligent extensions

Abstract

This paper presents an industry practice for highly efficient source code analysis to promote software quality. As a continuous work of previously reported source code analysis system, we researched and developed a few engineering-oriented intelligent extensions to implement more cost-effective extended code static analysis and engineering processes. These include an integrated empirical scan and filtering tool for highly accurate noise reduction, and a new code checking test tool to detect function call mismatch problems, which may lead to many severe software defects. We also extended the system with an automated defect filing and verification procedure. The results show that, for a huge code base of millions of lines, our intelligent extensions not only contribute to the completeness and effectiveness of static analysis, but also establish significant engineering productivity.