Skip to content
View strange2000's full-sized avatar
  • Hefei School of Economics -Anhui Agricultural University
  • Hefei

Highlights

  • Pro

Block or report strange2000

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
strange2000/README.md

陌生壬

合肥。日常在 Windows 上用 DSH(DeepSeek Harness)搭工作流,也折腾 macOS(Intel)上的一些国产客户端。

习惯:撞到 bug 就查到底 —— 从日志、数据库、二进制里取证,把「能跑的证据」写清楚,再报给上游。

复现材料

报 issue 时,我尽量不只描述现象,而是给出可重跑的取证与对照实验:

📁 strange2000/dsh-repro

编号 上游 issue 一句话
0001 slow-stack/mneme #493 蒸馏解析失败后确定性无限重试同一窗口;三臂对照实验定位到「转录裸放 user」被模型当成对话续写
0002 Ebony-Vinyl/dsh-our-free-model #427 应用内升级总是回滚:热重载窗口内无法确认新版本(激活版本不匹配)
0003 c1overzz/dsh-bili-scroll #47 桌面版装完即不可用的三处 bug:pythonPath 回落 / 设置互冲 / headless 缺凭证

每份材料包含:只读取证脚本(真实数据库 / 日志)+ 可重跑的对照实验 + 最小修法建议。

我报 issue 的三条纪律

  1. 只读取证:先证明事实,再猜原因。审计表、游标、日志逐条对齐,不做「大概是」的推断。
  2. 对照实验:把「环境差异」和「代码缺陷」分开。改一个变量跑一次,用结果说话。
  3. 给最小修法:指出代码位置与判据,而不是把问题原样丢回给维护者。

环境

  • Windows 11 + DSH Desktop(主)
  • macOS(Intel)+ 若干客户端实测

Pinned Loading

  1. dsh-repro dsh-repro Public

    DSH 生态 issue 的可跑复现材料:取证脚本 + 对照实验 + 报告

    JavaScript