Can't access a live workspace for Julia when using integrated notebooks in VS Code

Tooling VS Code
#1

Well, the title kind of says it. According to the docs for Vs Code Julia support it should be possible to see the Julia workspace while using Notebooks.

I can’t get this to work and I see no way to do so. I am using the integrated VS Code support for notebooks, not Python Jupyter. I am using an IJulia kernel and not the kernel provided by the language server.

The Julia language server’s jupyter kernel constantly hung or crashed, so I stopped using it.

I changed my config for the internal VS Code jupyter notebook, so I will try to LS’s kernel again because I believe that is the only kernel that provides workspace support.

“Wish me luck…!”

1 Like
#2

Great News!

  1. Obviously, one must use the language server kernel for a live workspace on the notebook.
  2. It works.
  3. The settings changes I made in VS Code end the crashes and timeouts for the notebook in VS Code when using either external kernels (IJulia) or the language server kernel.

I am not sure which of these settings made things work, but one or more of them did it:

"jupyter.experiments.enabled": false,
"jupyter.jupyterInterruptTimeout": 50000,
"jupyter.jupyterLaunchTimeout": 20000,
"jupyter.jupyterServerType": "local",
"jupyter.useDefaultConfigForJupyter": false,
#4

Hi lewis,

Even after I applied the settings you posted, I am still getting the following crashes. Could you think of other solutions?

info 21:39:48.592: Cancel pending cells
info 21:39:48.592: Interrupt kernel execution
info 21:39:48.592: Interrupting kernel: julia-1.7
info 21:39:48.592: Interrupting kernel with a shell message
warn 21:40:38.596: Error during interrupt: Error: Kernel interrupt failed. Jupyter server is hung. Please reload VS Code.
#5

Are you on Julia for Mac M1/M2? At the end of the day, I had to go back to intel MacOS with Julia 1.7 and then the crashes and problems stopped.

Julia 1.8 just shipped with improved support for Apple Silicon. I am giving it a try today.

#6

I am using Julia for Windows with an Intel CPU. I also raised this issue on Github for VS Code, and it looks like this problem is coming from the Julia kernel itself:

https://github.com/microsoft/vscode-jupyter/issues/10328

#7

Yes, I’ve seen that issue and it’s not what I mentioned. You should report at the GitHub repository for the Julia VS Code extension. Good luck.

  • Lewis
1 Like