A New Clone Group Mapping Algorithm for Extracting Clone Genealogy on Multi-version Software

Research on code clone evolution is very hot, and it can contribute to the investigation of the characteristics, the maintenance, the refactoring, and the harmfulness evaluation of code clones. In clone evolution research, extracting clone genealogies is the key technique, and mapping clone groups between neighboring software versions is a thorny issue for extracting clone genealogies. The efficiency and robustness of the current method are not satisfactory. To address this problem, this paper presents a new clone group mapping algorithm in order to extract the clone genealogy. The algorithm takes NICAD´s results as input, and describes the information of code clones with CRD, and maps the clone groups as well as the clone fragments according to the CRD matching level, the location overlapping rate and if necessary the text similarity rate. We evaluate our method on three software systems of different size and written in different languages. The results show that our method works well in clone group mapping under various circumstances, and it also can as a reliable foundation for constructing clone genealogies.