Description

We will publish all our implementations and experimental results at Zenodo (https://zenodo.org) once the paper get accepted for faciliating future replication and comparson.

Redundancy-based automated program repair (APR), which generates patches by referencing existing source code, has gained much attention since they are effective in repairing real-world bugs and have good interpretability. However, since existing approaches either demand the existence of multi-line similar code or generate patches by randomly referencing existing code, they still can repair a small number of bugs with many incorrect patches, hindering their wide application in practice.

In this work, we aim to improve the effectiveness of redundancy-based APR by exploring more effective source code reuse methods for improving the number of correct patches and reducing incorrect patches. Specifically, we have proposed a new repair technique named Repatt, which incorporates a two-level pattern mining process for guiding effective patch generation (i.e., token and expression levels).

Additionally, we have conducted an extensive experiment on the widely-used Defects4J benchmark by comparing Repatt with eight state-of-the-art APR approaches. The results show that our approach complements existing approaches by repairing 15 unique bugs compared with the latest deep learning-based methods and 19 unique bugs compared with traditional repair methods when providing the perfect fault localization. In addition, when the perfect fault localization is unknown in real practice, Repatt significantly outperforms the baseline approaches by achieving much higher patch precision, i.e., 83.8%. Moreover, we further proposed an effective patch ranking strategy for combining the strength of Repatt and the baseline methods. The result shows that it repairs 124 bugs when only considering the Top-1 patches and improves the best repair method by repairing 39 more bugs. The results demonstrate the effectiveness of our approach for practical use.

Project Structure

.
├── APRTool                     # Java Code of PatchRanker(For combination)
	# And for priliminary study
├── RankResultPerfectFL.txt     # Rank Result on Perfect FL
├── SBFLRankResult.txt          # Rank Result on SBFL
├── SBFLPatches                 # Patches generated on SBFL 
├── Patches                     # Patches generated on Perfect FL
├── d4j-info                    # Information about D4j, such as failing tests 
├── expriment_without_skip.zip  # Experiment logs of RQ3
├── lib                         # Essential java libaraies 
├── res                         # Log4j settings
├── repatt.jar                  # Executable Jar File of RePatt
├── sbfl_result.zip             # Experiment of our tool in SBFL scenario
├── src                         # Source code, based on SimFix
├── standard_experiment_log.zip # Experiment logs of RQ1
├── test                        # Unit tests
├── token_freq_2_log.zip        # Experiment logs of RQ4
└── token_freq_5_log.zip        # Experiment logs of RQ4

Reproducibility

The most Easy way to run the artefacts is using our docker images

docker pull veweew/repatt_artefacts

After successfully launching the docker container, the path to the executable artifacts is /artefacts.

To check out the buggy projects, you must place them in the d4j folder. The path should be formatted as /d4j/project/project_id_buggy, with all directory names in lowercase. Three sample projects, Codec 3, JacksonDatbind 96, and CLi 5, have already been checked out in the docker image.

To execute the program, simply run ./test.sh, and the fix process will begin.

The command to run Repatt is as follows:

java -Xmx10240m -jar repatt.jar --proj_home=/artefacts --proj_name=jacksondatabind --bug_id=96

---

If you wish to configure the environment by yourself, you should follow these steps:

1. Install JDK 8.
2. Install the Defects4j Framework.
3. Configure your environment variables by adding JAVA_TOOL_OPTIONS=-Dfile.encoding=UTF8 and DEFECTS4J_HOME=/path_to_d4j.
4. Run the program with arguments, as mentioned above.