Can't see or select Jupyter kernel's from Kernel drop-down

[bnapora]

Issue Type: Bug

Can no longer select jupyter kernel in virtual environments created outside of /usr directory. In previous version other jupyter kernels would appear in list after restarting VSCode. (Seems to have stopped working with version 1.6, but not certain.) Now the only way to get Notebook to work with selected kernel is to first open notebook in Jupyter Labs (or Notebook), then select the appropriate kernel, save notebook, and then open notebook in VSCode.

Steps:
1.) Create 2 or more virtual environments using "venv"
2.) Install ipykernel in each environment
3.) Configure ipykernel environment name
4.) Open Jupyter notebook and attempt to select or locate no environment
5.) New environment is not displayed and no method to locate or add it.

Suggestion:
Configure VSCode to use same kernelspec as Jupyter, for selecting kernels.

Extension version: 2021.8.2041215044
VS Code version: Code 1.60.2 (7f6ab5485bbc008386c4386d08766667e155244e, 2021-09-22T12:00:31.514Z)
OS version: Windows_NT x64 10.0.22458
Restricted Mode: No
Remote OS version: Linux x64 5.10.43.3-microsoft-standard-WSL2

System Info

Item	Value
CPUs	Intel(R) Xeon(R) W-10855M CPU @ 2.80GHz (12 x 2808)
GPU Status	2d_canvas: enabled gpu_compositing: enabled multiple_raster_threads: enabled_on oop_rasterization: enabled opengl: enabled_on rasterization: enabled skia_renderer: enabled_on video_decode: enabled vulkan: disabled_off webgl: enabled webgl2: enabled
Load (avg)	undefined
Memory (System)	63.73GB (44.40GB free)
Process Argv	--crash-reporter-id 751cf961-6f23-4f32-a920-1a73dd86b219
Screen Reader	no
VM	18%

Item	Value
Remote	SSH: mltooling-localhost
OS	Linux x64 5.10.43.3-microsoft-standard-WSL2
CPUs	Intel(R) Xeon(R) W-10855M CPU @ 2.80GHz (12 x 2808)
Memory (System)	31.22GB (24.93GB free)
VM	0%

A/B Experiments

vsliv368:30146709
vsreu685:30147344
python383:30185418
pythonvspyt602:30300191
vspor879:30202332
vspor708:30202333
vspor363:30204092
vswsl492:30256859
vstes627:30244334
pythonvspyt639:30300192
pythontb:30283811
pythonptprofiler:30281270
vshan820:30294714
vstes263:30335439
pythondataviewer:30285071
pythonvsuse255:30340121
vscod805cf:30301675
pythonvspyt200:30340761
binariesv615:30325510
vsccppwtct:30364498
pythonvssor306:30344512
bridge0708:30335490
pygetstartedt2:30371810
dockerwalkthru:30370836
bridge0723:30353136
pythonrunftest32:30373476
pythonf5test824:30373475
javagetstartedt:30364666
pythonvspyt187:30373474
pydsgsc2:30361791

[IanMatthewHuff]

@bnapora Could you please try downloadin VS Code - Insiders (version number should be 1.61.0) and checking this scenario quick? That should get you the newest version of our bits, and I believe this is a bug that has been addressed already. It's a safe SxS install with VS Code - Stable, so it won't mess anything up with your current install.

[bnapora]

@IanMatthewHuff I will have to fix my vscode insiders install before I can test this is issue. All of my python envs are in SSH Targets. Currently can't open any of my SSH Targets in Insiders build, because I get an "EISDIR: illegal operation on a directory, read" when it tries to open the SSH config file.

Thanks for the quick response and for the suggestion

[DonJayamanne]

5.) New environment is not displayed and no method to locate or add it.

As a temporary work around, you can:

• Find where python is installed
• Select the command Python: Select Interpreter
  • Next select Enter interpreter path
  • Next enter the path into the input box and hit Enter

This interpretr will now work in VS Code and you might need to re-load VS Code just once more to get it workign in jupyter extension.
We're currently working on ensuring users don't need to re-load vscode to detect the environments in jupyter #5319

[rchiodo]

This should be fixed now.

[roblourens]

What are the steps to verify what you fixed? The steps in the OP seem pretty generic

[cpoptic]

This issue continues to persist in VS Code Jupyter notebooks, as of June 2022. Please re-open this issue until is is solved in a reliable way. This issue has been present in VS Code Jupyter extension for several years now, as evidenced by a litany of GitHub issues mentioning it.

[rchiodo]

This issue continues to persist in VS Code Jupyter notebooks, as of June 2022. Please re-open this issue until is is solved in a reliable way. This issue has been present in VS Code Jupyter extension for several years now, as evidenced by a litany of GitHub issues mentioning it.

@cpoptic Can you be more specific? Or perhaps open a new issue. What exactly doesn't work?

Creating kernels using this will show them in the kernel dropdown:

python -m ipykernel install --name "custom-kernel-name" --user

Granted it takes a reload but it should find it eventually.

[danielaiyuan]

@rchiodo When I modify the virtual environments through Anaconda, removing one virtual environment and creating one new, the new virtual environments cannot provide a stable jupyter kernel for editing. The new interpreter disappears from the jupyter kernel options every time after creating a new .ipynb file in vs code, restarting vs code or re-activating the virtual environments do not help. Occasionally, everything works normal, for instance, after I run a python file with the new python interpreter in vs code, but this method is not reproducible.

[DonJayamanne]

@danielaiyuan
Please could you try using the latest pre-release version of Jupyter extension and installing the latest insider versino of VS Code.
If that still doesn't work, please file an issue and we'd be happy to look into that.