RefDetect: A Multi-Language Refactoring Detection Tool based on String Alignment
This page is provided as a supplementary material for the article: RefDetect: A Multi-Language Refactoring Detection Approach based on String Alignment. More specifically, we provide the results as a downloadable JSON file and also provide Java codes for all examples used in the paper. A link to all repositories used to evaluate the proposed approach is also provided.
In this project, we propose a novel, language-independent technique to identify refactorings in commit histories. The proposed approach is based on an alignment algorithm to detect refactorings applied between two program versions. The proposed approach is independent from any programming language, although Java and C++ languages are used to evaluate the approach. The proposed approach has been implemented in a tool called RefDetect, capable of detecting 27 refactoring types (including composite ones). RefDetect is evaluated, and compared with the current state-of-the-art refactoring detection tool: RMiner [1]. In experiments we applied RefDetect to 514 commits of 185 Java applications containing 5,058 true refactoring instances. To demonstrate the language-independence of RefDetect, we conducted a further study with four C++ applications.
We compare our tool with RMiner based on dataset provided by Dr. Nikolaos Tsantalis and his colleagues. RefDetect is not currently publicly available. However, the latest version of RMiner can be downloaded in this link.
We analyze the detected refactoring as follows: If a refactoring is detected by both tools and is reported as true or false positive instance in the original dataset, we accept that as it is determined in the dataset. However, if a refactoring is detected by one tool, but not by the other, we manually analyzed the refactoring to know it is a true or false positive instance. It is important to note that as we are not the developers of Java applications used in the experiments, so it is possible that we made a mistake in categorizing the detected refactorings as false or true positive instances. However, in situations that RefDetect or RMiner cannot detect a correct refactoring or detect a refactoring incorrectly, where possible we simulated the situation using a simple test case downloadable as zip file in this link.
The results as JSON file can be downloaded in this link (Note that this file is based on RefDetect 1.0 and RMiner 2.0.1).
A list of refactorings detected by RefDiff [2, 3], but not detected by either RefDetect or RMiner can be downloaded in this link.
[1] N. Tsantalis, A. Ketkar, and D. Dig, “RefactoringMiner 2.0,” IEEE Transactions on Software Engineering, 2020.
[2] D. Silva and M. T. Valente, “RefDiff: detecting refactorings in version histories,” in Proceedings of the 14th International Conference on Mining Software Repositories, ser. MSR ’17. Piscataway, NJ, USA: IEEE, 2017, pp. 269–279.
[3] D. Silva, J. Silva, G. Santos, R. Terra, and M. T. Valente, “RefDiff 2.0: A multi-language refactoring detection tool,” IEEE Transactions on Software Engineering, 2020.
