Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

StoneDetector: Conventional and versatile code clone detection for Java

Thomas S. Heinze, André Schäfer, Wolfram Amme

TL;DR

StoneDetector addresses code clone detection in Java by encoding control-flow information through dominator-tree paths and performing approximate textual matching on encoded path descriptions. The approach maintains a conventional, language-agnostic clone-detection core that supports Java source and Bytecode, with configurable string metrics and hashing to balance recall, precision, and scalability. Thorough evaluations against BigCloneBench and additional benchmarks show competitive performance for exact and near-miss clones, and notable strength in harder-to-detect Type 3/4 clones, including large-scale codebases. The tool is publicly available and designed for extensibility, enabling experiments with alternative representations, metrics, and Bytecode-based analysis, with promising directions for integrating Deep Learning and multi-language support in future work.

Abstract

Copy & paste is a widespread practice when developing software and, thus, duplicated and subsequently modified code occurs frequently in software projects. Since such code clones, i.e., identical or similar fragments of code, can bloat software projects and cause issues like bug or vulnerability propagation, their identification is of importance. In this paper, we present StoneDetector and its underlying method for finding code clones in Java source and Bytecode. StoneDetector implements a conventional clone detection approach based upon the textual comparison of paths derived from the code's representation by dominator trees. In this way, the tool does not only find exact and syntactically similar near-miss code clones, but also code clones that are harder to detect due to their larger variety in the syntax. We demonstrate StoneDetector's versatility as a conventional clone detection tool and analyze its various available configuration parameters, including the usage of different string metrics, hashing algorithms, etc. In our exhaustive evaluation with other conventional clone detectors on several state-of-the-art benchmarks, we can show StoneDetector's performance and scalability in finding code clones in both, Java source and Bytecode.

StoneDetector: Conventional and versatile code clone detection for Java

TL;DR

StoneDetector addresses code clone detection in Java by encoding control-flow information through dominator-tree paths and performing approximate textual matching on encoded path descriptions. The approach maintains a conventional, language-agnostic clone-detection core that supports Java source and Bytecode, with configurable string metrics and hashing to balance recall, precision, and scalability. Thorough evaluations against BigCloneBench and additional benchmarks show competitive performance for exact and near-miss clones, and notable strength in harder-to-detect Type 3/4 clones, including large-scale codebases. The tool is publicly available and designed for extensibility, enabling experiments with alternative representations, metrics, and Bytecode-based analysis, with promising directions for integrating Deep Learning and multi-language support in future work.

Abstract

Copy & paste is a widespread practice when developing software and, thus, duplicated and subsequently modified code occurs frequently in software projects. Since such code clones, i.e., identical or similar fragments of code, can bloat software projects and cause issues like bug or vulnerability propagation, their identification is of importance. In this paper, we present StoneDetector and its underlying method for finding code clones in Java source and Bytecode. StoneDetector implements a conventional clone detection approach based upon the textual comparison of paths derived from the code's representation by dominator trees. In this way, the tool does not only find exact and syntactically similar near-miss code clones, but also code clones that are harder to detect due to their larger variety in the syntax. We demonstrate StoneDetector's versatility as a conventional clone detection tool and analyze its various available configuration parameters, including the usage of different string metrics, hashing algorithms, etc. In our exhaustive evaluation with other conventional clone detectors on several state-of-the-art benchmarks, we can show StoneDetector's performance and scalability in finding code clones in both, Java source and Bytecode.

Paper Structure

This paper contains 32 sections, 6 figures, 10 tables.

Table of Contents

  1. Introduction
  2. Overall Approach and Architecture
  3. Extensibility of StoneDetector
  4. Clone Detection using Dominator Trees
  5. Control flow and Dominator Trees
  6. Finding Similar Code Fragments
  7. Metrics for Path Comparison
  8. Fingerprints and Hashing
  9. Comparison of Description Sets
  10. Evaluation
  11. Benchmarks
  12. Research Question 1: Recall and Precision
  13. Qualitative Analysis
  14. StoneDetector vs NiCad
  15. StoneDetector vs CloneWorks
  16. ...and 17 more sections

Figures (6)

  • Figure 1: Variations of code clones: Examples of Type 1, 2, 3, and 4 code clones
  • Figure 2: Example of two code fragments that are more difficult to identify as code clone
  • Figure 3: Overview of the tool StoneDetector
  • Figure 4: Control flow graph, dominator tree, description set, and fingerprints of example
  • Figure 5: Distance metrics for calculating path similarity
  • ...and 1 more figures