BUMP

BUMP是一个包含571个可复现的破坏性依赖更新的基准数据集，由瑞典皇家理工学院创建。该数据集收集自153个使用Maven构建系统的Java项目，确保了依赖更新的长期可复现性，并提供了稳定的科学工件。BUMP通过Docker镜像存储所有破坏性更新，支持离线运行，适用于研究依赖工程、破坏性依赖更新及可复现性等问题。数据集的应用领域主要集中在软件依赖工程研究，旨在解决依赖更新中出现的构建失败问题。

BUMP is a benchmark dataset consisting of 571 reproducible dependency-breaking updates, created by the Royal Institute of Technology (KTH) in Sweden. This dataset is collected from 153 Java projects that use the Maven build system, ensuring long-term reproducibility of dependency updates and providing stable scientific artifacts. BUMP stores all dependency-breaking updates via Docker images, supports offline execution, and is applicable to research on topics including dependency engineering, dependency-breaking updates and reproducibility. The dataset is mainly focused on software dependency engineering research, aiming to address build failure issues caused by dependency updates.