Enhancing Software Robustness by Detecting and Removing Exception Handling Smells: An Empirical Study

Abstract

We propose a systematic way to uncover and fix bugs through detecting smells associated with exception handling. First, code of software under improvement is scanned for exception handling smells by a static analysis tool. The smells are reviewed for confirming if they are bugs by writing failing tests. Finally, code that contains the smells is refactored until the failing test passes and the smells are removed. We have also conducted an empirical study to demonstrate the efficacy of the proposed approach. In the empirical study, an open source static analysis tool is applied to detect exception handling smells in an open source web application. The result shows that out of the 357 smells reported by the tool, 124 are confirmed to be bugs that could affect the robustness of the web application.