Mining sequences of changed-files from version histories
Mining sequences of changed-files from version histories
复制标题
从版本历史中挖掘更改文件的序列
DOI：
10.1145/1137983.1137996
复制
发表时间：
2006
期刊：
影响因子：
--
通讯作者：
Jonathan I. Maletic
中科院分区：
文献类型：
--
作者：
Huzefa H. Kagdi;Shehnaaz Yusuf;Jonathan I. Maletic
现代的源代码控制系统，如Subversion，将文件的变更集保存为原子提交。但是，在这些源代码存储库中通常找不到更改文件的特定顺序信息。本文提出了一组启发式方法，用于对源代码存储库中的变更集（即日志条目）进行分组。给定这样一组更改集，就会发现经常一起更改的文件序列。这种方法不仅提供（无序的）文件集，而且用（部分时序的）排序信息补充它们。在KDE源代码存储库的一个子集上演示了该技术。结果表明，该方法能够找到被修改文件的序列。
Modern source-control systems, such as Subversion, preserve change-sets of files as atomic commits. However, the specific ordering information in which files were changed is typically not found in these source-code repositories. In this paper, a set of heuristics for grouping change-sets (i.e., log-entries) found in source-code repositories is presented. Given such groups of change-sets, sequences of files that frequently change together are uncovered. This approach not only gives the (unordered) sets of files but supplements them with (partial temporal) ordering information. The technique is demonstrated on a subset of KDE source-code repository. The results show that the approach is able to find sequences of changed-files.