Testing helps assure software quality by executing a program and uncovering bugs. Scientific software developers often find it challenging to carry out systematic and automated testing due to reasons like inherent model uncertainties and complex floating-point computations. Extending the recent work on analyzing the unit tests written by the developers of the Storm Water Management Model (SWMM) [32], we report in this paper the investigation of both unit and regression tests of SWMM. The results show that the 2953 unit tests of SWMM have a 39.7% statement-level code coverage and a 82.4% user manual coverage. Meanwhile, an examination of 58 regression tests of SWMM shows a 44.9% statement-level code coverage and a near 100% user manual coverage. We also observe a “getter-setter-getter” testing pattern from the SWMM unit tests, and suggest a diversified way of executing regression tests.

测试通过执行程序和发现错误来帮助确保软件质量。由于固有的模型不确定性和复杂的浮点计算等原因，科学软件开发人员经常发现进行系统和自动化测试具有挑战性。扩展最近对雨水管理模型（SWMM）开发人员编写的单元测试进行分析的工作[32]，我们在本文中报告了对 SWMM 的单元测试和回归测试的调查。结果显示，SWMM的2953个单元测试的语句级代码覆盖率为39.7%，用户手册覆盖率为82.4%。同时，对SWMM的58个回归测试的检查显示，语句级代码覆盖率为44.9%，用户手册覆盖率接近100%。我们还从 SWMMH 单元测试中观察到“getter-setter-getter”测试模式，并提出了执行回归测试的多样化方法。