概要: Automated program repair (APR) is a concept of automati- cally fixing bugs in source code to free developers from the burden of debugging. One of the issues facing search-based APR is that repaired code contains wasteful or meaningless statements that do not affect external behavior. This paper proposes a concept named source code tidying that eliminates wasteful statements in source code repaired by search-based APR. Our proposed method applies pre-defined tidying rules to repaired code and evaluates the effect of tidying using source code metrics such as lines of code. By repeating this process based on a genetic algorithm, unnatural and full of wasteful source code is gradually brought close to natural with preserving its behavior. Our method will be involved in a process of APR by improving the readability of repaired code. タグ: PROFES

@inproceedings{TakumiIwase2022,
  author = {Takumi Iwase and Shinsuke Matsumoto and Kusumoto Shinji},
  title = {Tidy Up Your Source Code! Eliminating Wasteful Statements in Automatically Repaired Source Code},
  booktitle = {Proceedings of the 23rd International Conference on Product-Focused Software Process Improvement (PROFES2022)},
  pages = {544--550},
  year = {2022},
  month = {nov}
}