| Tweet | |
| M. Doi, Y. Higo, and S. Kusumoto, "A Code Clone Curation -Towards Scalable and Incremental Clone Detection-," In 7th International Workshop on Quantitative Approaches to Software Quality (QuASoQ2019), 7, pp. 17-24, December 2019. | |
| ID | 618 |
| 分類 | 国際会議 |
| タグ | -towards clone code curation detection- incremental scalable |
| 表題 (title) |
A Code Clone Curation -Towards Scalable and Incremental Clone Detection- |
| 表題 (英文) |
|
| 著者名 (author) |
Masayuki Doi,Yoshiki Higo,Shinji Kusumoto |
| 英文著者名 (author) |
Masayuki Doi,Yoshiki Higo,Shinji Kusumoto |
| 編者名 (editor) |
|
| 編者名 (英文) |
|
| キー (key) |
Masayuki Doi,Yoshiki Higo,Shinji Kusumoto |
| 書籍・会議録表題 (booktitle) |
7th International Workshop on Quantitative Approaches to Software Quality (QuASoQ2019) |
| 書籍・会議録表題(英文) |
|
| 巻数 (volume) |
7 |
| 号数 (number) |
|
| ページ範囲 (pages) |
17-24 |
| 組織名 (organization) |
|
| 出版元 (publisher) |
|
| 出版元 (英文) |
|
| 出版社住所 (address) |
|
| 刊行月 (month) |
12 |
| 出版年 (year) |
2019 |
| 採択率 (acceptance) |
66% |
| URL |
|
| 付加情報 (note) |
|
| 注釈 (annote) |
|
| 内容梗概 (abstract) |
Code clones have a negative impact on software maintenance. Code clone detection on large-scale source code takes a long time, and even worse, such detections occasionally aborted due to too much target size. Herein, we consider that we detect cross-project code clones from a set of many projects. In such a detection situation, even if a project is updated, we need to detect cross-project code clones again from the whole of the projects if we simply use a clone detection tool. Therefore we need a new clone detection scheme that rapidly detects code clones from a set of many projects with incremental detection functionality. In this paper, we propose an approach that rapidly obtains code clones from many projects. Our approach includes a strategy of multi-stage code clone detection unlike other clone detection techniques: in the first-stage detection, code clones are detected from each of the projects, respectively; then in the second-stage detection, the code clones detected in the first-stage are unified by using our clone curation technique. This multistage detection strategy has a capability of incremental clone detection: if the source code of a project is updated, the first-stage detection is applied only to the project and then the second-stage curation is performed. We constructed a software system based on the proposed approach. At present, the system utilizes an existing detection tool, CCFinder. We have applied the system to the large-scale source code (12 million LoC) of 128 projects. We also detected clones from the target source code by simply using CCFinder and compared the detection time. As a result, the clone detection with our system was 17 times shorter than CCFinder’s detection. We also compared detection time under an assumption that each project of the targets is updated once. The experimental results showed that our incremental clone detection shortened the detection time by 91% compared to non-incremental clone detection. |
| 論文電子ファイル | m-doi_QuASoQ2019_camera-ready.pdf (application/pdf) [一般閲覧可] |
| BiBTeXエントリ |
@inproceedings{id618,
title = {A Code Clone Curation -Towards Scalable and Incremental Clone Detection-},
author = {Masayuki Doi and Yoshiki Higo and Shinji Kusumoto},
booktitle = {7th International Workshop on Quantitative Approaches to Software Quality (QuASoQ2019)},
volume = {7},
pages = {17-24},
month = {12},
year = {2019},
acceptance = {66\%},
}
|