Welcome to the GitHub repository for our paper titled "On the Coupling between Vulnerabilities and LLM-generated Mutants: A Study on Vul4J dataset" which has been accepted in the 17th IEEE International Conference on Software Testing, Verification and Validation (ICST 2024).
The accepted paper is available here:
The link to the proceedings of ICST 2024 conference, related to our paper is available here:
In our study, we utilized the Vul4J dataset, and accordingly, the directory names adhere to the naming convention as provided in the mapping file.
The dataset includes the following components:
-
Source Code: The source code of all projects, including both vulnerable and fixed code, provided in zipped format. Please note that for larger projects, there may be multiple zip files, each with a maximum size of 100 MB.
-
Patches: Individual classes modified during the fix, available as patches for both vulnerable and fixed versions.
-
Mutated Classes: Individual fixed classes that underwent mutation.
-
Mutants: All the mutants generated for each project using µBert.
-
Vulnerability-Fix Information: Comprehensive information on vulnerabilities and their fixes, including sentence-level changes, commit dates, and more, located in the patches directory.
-
Scripts: A collection of scripts combined into a single Python script.