Recovering and Tracing Links between Software Codes and Test Codes of the Open Source Projects

One of the most important controversial issues in the design and implementation of software is the functionality of the designed system. With impressive efforts of different software teams in the field of the system, the primary concern of the developers is its proper and error free functioning of the whole system. Therefore, various tests are defined and designed to help software teams to produce error free software or software with minimum error rate. It is difficult but important to find a proper link between written test class and the class under the test. Discovering these links is useful for programmers to perform the Regression Test more efficiently. In this paper, we are trying to propose a model for the recovery of traceable links between test classes and the classes under the test. The presented model comprises four sections. Firstly, we retrieve the name of similar classes between the test class and source class. Afterward, we extract the complexity, Cyclomatic and design metrics from the source codes and the test classes. Finally, after creating a train set, we implement the data mining algorithms to find the potential relationship between unit tests and the classes under the test. One of the advantages of this method is its language independence; furthermore, the preliminary results show that the proposed method has a good performance.

@article{Rasekh2016,
  title     = {Recovering and Tracing Links between Software Codes and Test Codes of the Open Source Projects},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {7},
  number    = {10},
  year      = {2016},
  publisher = {The Science and Information Organization},
  author    = {Amir Hossein Rasekh and Amir Hossein Arshia and Seyed Mostafa Fakhrahmad and Mohammad Hadi Sadreddini},
  doi       = {10.14569/IJACSA.2016.071043},
  url       = {https://doi.org/10.14569/IJACSA.2016.071043}
}