<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>代码审查 on TheFreeMeal</title><link>https://blog.thefreemeal.com/zh/tags/%E4%BB%A3%E7%A0%81%E5%AE%A1%E6%9F%A5/</link><description>Recent content in 代码审查 on TheFreeMeal</description><generator>Hugo -- 0.146.0</generator><language>zh</language><lastBuildDate>Wed, 29 Jul 2026 17:14:28 +0800</lastBuildDate><atom:link href="https://blog.thefreemeal.com/zh/tags/%E4%BB%A3%E7%A0%81%E5%AE%A1%E6%9F%A5/index.xml" rel="self" type="application/rss+xml"/><item><title>《代码审查的终结：编程 agent 取代人工检查》笔记</title><link>https://blog.thefreemeal.com/zh/posts/2026-07-29-end-of-code-review-paper-notes/</link><pubDate>Wed, 29 Jul 2026 17:14:28 +0800</pubDate><guid>https://blog.thefreemeal.com/zh/posts/2026-07-29-end-of-code-review-paper-notes/</guid><description>一篇 6 月的 arXiv 论文主张取消「每个改动都要人逐行看」这条默认规矩。这是它的论点、论据、自己承认的软肋，以及对照我三轮四个 agent 的评审实践。</description></item><item><title>AI 生成的代码，如何进行评审</title><link>https://blog.thefreemeal.com/zh/posts/2026-07-27-should-humans-review-ai-code/</link><pubDate>Mon, 27 Jul 2026 18:44:32 +0800</pubDate><guid>https://blog.thefreemeal.com/zh/posts/2026-07-27-should-humans-review-ai-code/</guid><description>Uncle Bob 不读 agent 写的代码，Booch 全审，Monperrus 主张让 agent 审 agent。我们在 botmux 上跑了第三条路，用两个 PR 说明它一定程度可行，以及「一定程度」的边界在哪。</description></item></channel></rss>