标题： AssertFlip：通过反转LLM生成的通过测试来再现错误 显示英文标题

标题： AssertFlip: Reproducing Bugs via Inversion of LLM-Generated Passing Tests

摘要： 错误重现是软件调试和修复过程中的关键环节，然而在开源和工业环境中，大多数错误在报告时缺乏可执行的测试用例来重现它们，这使得诊断和解决变得更加困难和耗时。 为了解决这一挑战，我们引入了AssertFlip，这是一种使用大型语言模型（LLMs）自动生成可重现错误测试（BRTs）的新技术。 与现有方法试图直接生成失败测试不同，AssertFlip首先在有错误的行为上生成通过的测试，然后将这些测试反转，以便在存在错误时失败。 我们假设LLMs在编写通过的测试方面比那些故意崩溃或失败的测试更好。 我们的结果表明，AssertFlip在SWT-Bench基准测试中优于所有已知技术，该基准测试专门用于BRTs。 具体而言，AssertFlip在SWT-Bench-Verified子集上实现了43.6%的失败到通过的成功率。 显示英文摘要

摘要： Bug reproduction is critical in the software debugging and repair process, yet the majority of bugs in open-source and industrial settings lack executable tests to reproduce them at the time they are reported, making diagnosis and resolution more difficult and time-consuming. To address this challenge, we introduce AssertFlip, a novel technique for automatically generating Bug Reproducible Tests (BRTs) using large language models (LLMs). Unlike existing methods that attempt direct generation of failing tests, AssertFlip first generates passing tests on the buggy behaviour and then inverts these tests to fail when the bug is present. We hypothesize that LLMs are better at writing passing tests than ones that crash or fail on purpose. Our results show that AssertFlip outperforms all known techniques in the leaderboard of SWT-Bench, a benchmark curated for BRTs. Specifically, AssertFlip achieves a fail-to-pass success rate of 43.6% on the SWT-Bench-Verified subset.