【24h】

Evaluation of Test Code Quality with Aspect-Oriented Mutations

机译:使用面向方面的变异评估测试代码质量

获取原文
获取原文并翻译 | 示例

摘要

To evaluate this approach, we have built a prototype based on AspectJ compiler to build code and tests and with JUnit as the testing library. Early experiments show that it appears to generate and run the mutants a few orders of magnitude faster that the popular Jester. The savings result mainly from the fact that the tool does not require multiple mutant compilations, reduces the number of equivalent and transparent mutants, and preserves the syntactic correctness of the mutated code. However, it differs from Jester in that it learns the code usage from existing test cases, and then mutates the code. Jester, on the other hand, mutates the code insight into test cases, which allows for assessing the code coverage, but also leads to redundant or transparent mutants. Currently the prototype deals only with primitive Java types and null values for objects. In future, we plan to employ an on-fly object creation with dynamic proxies and implement other mutation operators as well as perform a larger scale evaluation.
机译:为了评估这种方法,我们已经建立了一个基于AspectJ编译器的原型来构建代码和测试,并使用JUnit作为测试库。早期实验表明,它似乎比流行的Jester快几个数量级地产生和运行突变体。节省的主要原因是该工具不需要多次编写突变体,减少了等效和透明突变体的数量,并保留了突变代码的语法正确性。但是,它与Jester的不同之处在于,它从现有测试用例中学习代码用法,然后对代码进行突变。另一方面,Jester将代码洞察力突变为测试用例,从而可以评估代码覆盖率,但也会导致冗余或透明的突变体。当前,原型仅处理原始Java类型和对象的空值。将来,我们计划采用带有动态代理的动态对象创建方法,并实施其他变异算子以及进行更大规模的评估。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号