← 研踪 PaperTrailarXiv:2611.13395 · v1
A Taxonomy of Failing Bug Reproduction Tests on SWT-Bench
首次提交 2026-11-10 08:17
版本更新 2026-11-10 08:17
雷达观测 2026-11-11 11:04
摘要
Automatic Program Repair (APR) tools require high-quality test generation models to validate candidate patches. Despite the advancements of automated bug reproduction testing tools, they continue to struggle with issues of higher complexity. These struggles may be caused by a framework's lack of problem-specific context. We address the unclassified nature of these framework failures by providing a taxonomy of bug reproduction tests. This study uses the AssertFlip and OpenHands frameworks, the two highest-scoring open-source tools on SWT-Bench Verified. The significance of model selection on the causes of failure is also explored by utilizing GPT-5-mini and GPT-4o-mini on both tools. The error traces of these frameworks are categorized to determine granular modes of failure. Through this process, mechanical and implementation failures were the most common failure classes. Overall, a tool's generation approach has a significant impact on precision and coverage, while the selection of the model alters the quality and quantity of generated tests. Specifically, more powerful models with higher reasoning capability tended to present fewer failures due to a lack of technical or problem-specific context. Our experiment also shows evidence that our taxonomy can be used as concrete guidance to augment the issue description to improve the generation quality.
作者备注：To be published in CASCON 2026 proceedings
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-10 08:17 · 最新观测版本
A Taxonomy of Failing Bug Reproduction Tests on SWT-Bench
Automatic Program Repair (APR) tools require high-quality test generation models to validate candidate patches. Despite the advancements of automated bug reproduction testing tools, they continue to struggle with issues of higher complexity. These struggles may be caused by a framework's lack of problem-specific context. We address the unclassified nature of these framework failures by providing a taxonomy of bug reproduction tests. This study uses the AssertFlip and OpenHands frameworks, the two highest-scoring open-source tools on SWT-Bench Verified. The significance of model selection on the causes of failure is also explored by utilizing GPT-5-mini and GPT-4o-mini on both tools. The error traces of these frameworks are categorized to determine granular modes of failure. Through this process, mechanical and implementation failures were the most common failure classes. Overall, a tool's generation approach has a significant impact on precision and coverage, while the selection of the model alters the quality and quantity of generated tests. Specifically, more powerful models with higher reasoning capability tended to present fewer failures due to a lack of technical or problem-specific context. Our experiment also shows evidence that our taxonomy can be used as concrete guidance to augment the issue description to improve the generation quality.
查看此版本解读 →查看此版本原文 ↗