[Bug]: One-click install of DeepSeek Harness (dsh) in Docker fails with npm ETARGET

Aperta Adatta ai principianti
  • Qualcosa non va?
#475 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
2/5
Tempo stimato
1-3 ore
Idoneità per principianti
75/100
Tipo di issue
Bug
Chiarezza
Specificata chiaramente
Stato di attività
Attiva
Stack tecnologico
docker, go, nodejs
Ambito
cli, devops

Direzione di ricerca

Inizia da internal/runtime/cli_ondemand.go: la voce dsh alla riga 43 imposta Package su @deepseek-ai/dsh, e le righe 147-153 costruiscono il comando npm install senza dist-tag. Decidi se la specifica debba ricevere @latest in quel punto, oppure se il Dockerfile debba preinstallare il pacchetto come avviene per Codex e CodeBuddy. È completato quando l'installazione di dsh riesce all'interno del container loopforge-platform con npm 10.9.9 e non compare alcun errore ETARGET.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

Summary

Clicking one-click install for DeepSeek Harness (dsh) in the Docker deployment (loopforge-platform:local built from Dockerfile based on node:22-bookworm) fails with HTTP 502 and npm error ETARGET: No matching version found for @deepseek-ai/dsh@*..

Note for maintainers: When evaluating whether to fix this issue, please re-test and reproduce this bug in a Docker container environment first to confirm the behavior on your side.

Environment
  • Commit: b751e09 (v0.3.6-29-gb751e09)
  • Deployment: Docker Compose (./start.sh dev, base image node:22-bookworm)
  • Container npm version: 10.9.9 (contrast with macOS host npm 11.13.0, where installing @deepseek-ai/dsh without @latest succeeds)
  • Registry dist-tags for @deepseek-ai/dsh: 
    { alpha: '0.2.2-alpha.3', latest: '0.2.2-rc.1', next: '0.2.2-rc.3' }
Reproduction Steps
  1. Start LoopForge via Docker (./start.sh dev).
  2. Open the Studio Assistant UI, select DeepSeek Harness (dsh), and trigger one-click install (POST /api/assistant/cli-install).
  3. Observe the UI error and check docker logs loopforge-platform-1.
Actual Behavior

POST /api/assistant/cli-install returns 502:

2026/11/13 08:37:18 assistant CLI install dsh: 安装 DeepSeek Harness 失败：npm error code ETARGET npm error notarget No matching version found for @deepseek-ai/dsh@*.
2026/11/13 08:37:18 [HTTP] POST /api/assistant/cli-install -> 502 (204 bytes, 1.435s)
Root Cause Analysis
  1. Dockerfile preinstalls @openai/codex and @tencent-ai/codebuddy-code, but does not preinstall @deepseek-ai/dsh, so dsh relies on runtime on-demand installation via internal/runtime/cli_ondemand.go.
  2. In internal/runtime/cli_ondemand.go:
    • Line 43 defines "dsh": {Engine: "dsh", Binary: "dsh", Package: "@deepseek-ai/dsh", RelaxEngines: true}.
    • Lines 147–153 run npm install ... @deepseek-ai/dsh without an explicit @latest dist-tag.
  3. Because @deepseek-ai/dsh currently only has prerelease versions (its latest tag points to 0.2.2-rc.1), npm 10.9.9 inside node:22-bookworm resolves an untagged package spec as @deepseek-ai/dsh@* and rejects prerelease versions with ETARGET.
  4. Verified inside the same loopforge-platform-1 container that appending @latest (@deepseek-ai/dsh@latest) with --dry-run resolves add @deepseek-ai/dsh 0.2.2-rc.1 (added 724 packages, exit code 0).
Possible Fix Directions
  • Append @latest to Package for "dsh" (or when no explicit version/tag is specified in installNPMCLI) in internal/runtime/cli_ondemand.go, or
  • Preinstall @deepseek-ai/dsh@latest in Dockerfile (similar to Codex / CodeBuddy) or upgrade npm in the base image.
中文说明
问题概述

在 Docker 部署环境（基于 node:22-bookworm 构建的 loopforge-platform:local）中，在界面点击一键安装 DeepSeek Harness (dsh) 会直接失败，返回 HTTP 502，底层 npm 报错 ETARGET: No matching version found for @deepseek-ai/dsh@*.。

评估说明：在评估是否需要修复该问题时，希望能够重新在 Docker 环境中复测并复现一次这个 Bug，以确认实际表现。

复现环境
  • 代码版本：b751e09（v0.3.6-29-gb751e09）
  • 部署方式：Docker Compose（./start.sh dev，基础镜像 node:22-bookworm）
  • 容器内 npm 版本：10.9.9（对比 macOS 宿主机 npm 11.13.0 不带 @latest 可正常安装）
  • @deepseek-ai/dsh 的 dist-tags： 
    { alpha: '0.2.2-alpha.3', latest: '0.2.2-rc.1', next: '0.2.2-rc.3' }
复现步骤
  1. 通过 Docker 启动 LoopForge（./start.sh dev）。
  2. 打开 Studio 助手界面，选择 DeepSeek Harness (dsh) 并触发一键安装（POST /api/assistant/cli-install）。
  3. 观察界面报错并查看 docker logs loopforge-platform-1。
实际表现

POST /api/assistant/cli-install 返回 502：

2026/11/13 08:37:18 assistant CLI install dsh: 安装 DeepSeek Harness 失败：npm error code ETARGET npm error notarget No matching version found for @deepseek-ai/dsh@*.
2026/11/13 08:37:18 [HTTP] POST /api/assistant/cli-install -> 502 (204 bytes, 1.435s)
根因分析
  1. Dockerfile 仅预装了 @openai/codex 和 @tencent-ai/codebuddy-code，未预装 @deepseek-ai/dsh，因此 dsh 依赖 internal/runtime/cli_ondemand.go 在运行时按需安装。
  2. internal/runtime/cli_ondemand.go 中：
    • 第 43 行定义 "dsh": {Engine: "dsh", Binary: "dsh", Package: "@deepseek-ai/dsh", RelaxEngines: true}。
    • 第 147–153 行直接执行 npm install ... @deepseek-ai/dsh，未显式带 @latest 标签。
  3. 由于 @deepseek-ai/dsh 当前只发布了预发布版本（latest 指向 0.2.2-rc.1），node:22-bookworm 自带的 npm 10.9.9 会把未指定标签的包名按 @deepseek-ai/dsh@* 解析，从而拒绝预发布版本并报 ETARGET。
  4. 已在同一个 loopforge-platform-1 容器内使用 --dry-run 实测验证：改传 @deepseek-ai/dsh@latest 后可正常解析出 add @deepseek-ai/dsh 0.2.2-rc.1（added 724 packages，退出码 0）。
修复方向参考
  • 在 internal/runtime/cli_ondemand.go 中为 dsh（或未显式指定版本的包）补全 @latest；或者
  • 在 Dockerfile 中预装 @deepseek-ai/dsh@latest（与 Codex / CodeBuddy 一致），或升级基础镜像中的 npm 版本。
Lingua principale
Go
Stelle
141
Fork
28
Merge medio
1g 15h
PR unite (30g)
113

Preparare l'ambiente

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di Tencent/LoopForge

Tutte le issue di Tencent/LoopForge

Issue simili

Altre issue su Go

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.