I'm trying to simplify the workflow for using C++ in Jupyter Notebooks through VS Code on Windows 11. Here's my current setup:
Environment:
- Windows 11 Pro Workstation (Version 24H2, Build 26100.2894)
- WSL 2 (Ubuntu 22.04.5 LTS)
- VS Code 1.97.0 with Jupyter/Jupyter Notebook & C/C++ extensions
- Conda 25.1.1 with xeus-cling 0.15.0 installed in virtual environment
Current Workflow:
- Start WSL instance
- Launch Jupyter server via jupyter notebook --no-browser in WSL conda env
- Manually copy the WSL-generated URL (e.g., http://localhost:8888/?token=...)
- Paste URL into VS Code's "Enter the URL of the running Jupyter server" prompt
This works but becomes tedious compared to native Windows kernels (Python/R).
What I've Tried:
- WSL-based Solutions:
- VS Code's "Remote - WSL" extension (kernel connection instability)
Question:
Is there a way to achieve either:
- A persistent/automated connection to the WSL-based Jupyter server, or
- natively run the xeus-cling kernel on Windows?
I am specifically looking for a way to:
- Auto-start WSL Jupyter service on VS Code launch
- Automatically handle URL/token authentication
- Maintain kernel stability (current manual setup occasionally requires port changes)