Alternative Builds dataset for Levels of Binary Equivalence paper

The availability of multiple binaries built from the same sources creates new challenges and opportunities, and raises questions such as: “Does build A confirm the integrity of build B?” or “Can build A reveal a compromised build B?”. We introduce levels of equivalence, inspired by clone detection types, and demonstrate the value of these new levels through several experiments. For this purpose, we construct a dataset consisting of Java binaries (jar files) built from the same sources independently by different providers, resulting in 14,156 pairs of binaries in total.多份由同一源代码构建的二进制文件的存在，既带来了全新的挑战与机遇，也引出了诸如「构建产物A能否验证构建产物B的完整性？」或「构建产物A能否揭露已被篡改的构建产物B？」这类问题。借鉴克隆检测的分类思路，我们提出了一系列等价性分级体系，并通过多组实验验证了该新型分级体系的应用价值。为此，我们构建了一套数据集，该数据集收录了由不同构建方基于同一源代码独立编译得到的Java二进制文件（jar文件），最终总计形成14156组二进制文件对。