Repository navigation
VPython hangs in Jupyter notebook inside VS Code #281
Description
Activity
Two-part answer to this issue is underway:
-
The silent hang is fixed in Fail loudly instead of hanging forever when no frontend connects (#281) #292:
import vpythonnow fails after 30s with a clear error naming the environment (VS Code detected via its env vars) and pointing at what works, instead of hanging the kernel forever.VPYTHON_CONNECT_TIMEOUTraises the limit for genuinely slow first-runs. -
Actual VS Code support is being designed: a VS Code extension using the Notebook Renderer API + Jupyter kernel comm messaging (the mechanism ipywidgets uses), rendering through the same host-agnostic
glowcomm_host.jsfrontend recently factored out on thepyodide-packagingbranch (Pyodide/wasm support: pure-Python wheel, worker transport, GlowScript host factory #291). No timeline promised yet, but the hang at least stops being silent now.
🤖 Generated with Claude Code
-
- added a commit that references this issue
on Aug 16, 2026 Update: VPython scenes now render live in VS Code notebooks — working prototype, end to end. Objects,
rate()animation, zoom/orbit,text()fonts, and textures all verified.Two pieces:
- Kernel side: PR VS Code + Colab frontends, with kernel-side scene replay (#281) #293 — a websocket-only frontend that vpython auto-selects under VS Code (the whole wire protocol rides the tornado websocket the kernel already runs; no Comm, no nbextension JS).
- Renderer side: https://github.com/vpython/vscode-vpython — a VS Code notebook renderer extension that loads GlowScript in the output webview and connects to that websocket. Not on the marketplace yet; installable from a packaged vsix.
With the extension absent,
import vpythonfails loudly per #292 instead of hanging. Remote kernels (WSL/SSH/Codespaces) are the main known gap — needs the extension to ask VS Code for port forwarding.- added a commit that references this issue
on Aug 16, 2026
When running VPython inside a Jupyter notebook in VS Code, the code hangs indefinitely.
Minimal working example:
Expected behavior:
A 3D canvas should open and display a sphere.
Actual behavior:
The line with sphere() runs forever without producing output. No error message appears.
Environment:
Additional info: