"The version of Python associated with the selected kernel is no longer supported."

[luukburger]

Applies To

• Notebooks (.ipynb files)
• Interactive Window and/or Cell Scripts (.py files with #%% markers)

What happened?

When I open a Notebook VS Code displays a message in the bottom right corner saying:

"The version of Python associated with the selected kernel is no longer supported. Please consider selecting a different kernel."

I'm using a Miniconda environment running Python 3.12.1.

VS Code Version

Version: 1.86.0 (Universal) Commit: 05047486b6df5eb8d44b2ecd70ea3bdf775fd937 Date: 2024-01-31T10:29:15.765Z Electron: 27.2.3 ElectronBuildId: 26495564 Chromium: 118.0.5993.159 Node.js: 18.17.1 V8: 11.8.172.18-electron.0 OS: Darwin arm64 23.2.0

Jupyter Extension Version

v2024.1.0

Jupyter logs

No response

Coding Language and Runtime Version

Python 3.12.1

Language Extension Version (if applicable)

v2024.0.1

Anaconda Version (if applicable)

Conda 23.11..0 (Miniconda)

Running Jupyter locally or remotely?

Local

[luukburger]

I see this in my logs:

20:45:44.076 [info] Starting Kernel startUsingPythonInterpreter, .jvsc74a57bd06aa9a930b94239702003b1d5483aca273205853dc0a175a13e8fde3b74aafeaf.~/miniconda3/envs/python311/python.~/miniconda3/envs/python311/python.-m#ipykernel_launcher (Python Path: ~/miniconda3/envs/python311/bin/python, Conda, python311, 3.1.0) for '~/Code/crypto-quotes/notebooks/2. Data analysis/Analyse crypto quotes.ipynb' (disableUI=true)

Notice the "3.1.0". This is weird, because my Conda environment runs version 3.12. I also completely reinstalled Miniconda and setup multiple environments using Python 3.10, 3.11 and 3.12 and all give the same error.

[luukburger]

And now I also completely removed and reinstalled VS Code and the issue persists.

This only started occurring suddenly today, so maybe it has to do with an upgrade of the extension or of VS Code...?

[luukburger]

Hmm, you also see the version being interpreted incorrectly in the kernel selector. Notice the 3.1...

[DonJayamanne]

Transferring to Python extension repo as this version information is returned by Python extensxion.
@karrtikr A similar issue was reported by @luabud , looks like the version information returned is incorrect at some stage.

[kjarus1]

Suffering from the same problem. It also appears to result in the kernel crashing after a while with miniconda, whereas with venv environments everything is OK.

[MarkSomogyvari]

I had the same issue. I rolled back to the previous jupyter extension version in VSCode (v2024.1.0), and that solved the problem for me for now.

[karrtikr]

For folks running into the issue, can you please provide the debug logs?

• Run Developer: Set log level command
• Select Trace
• Run Developer: Reload Window command
• Provide the output:

Output for Python in the Output panel (View→Output, change the drop-down the upper-right of the Output panel to Python)

XXX

[OrhanGuen]

I had the same issue and also rolled-back to the previous extension version of jupyter Notebook in Visual Studio Code v2024.1.2024012502 (pre-release) that worked for me too.

[DonJayamanne]

@OrhanGuen @MarkSomogyvari
Does this work when you go back to the latest (stable) version of the Jupyter extension.
I believe it will,
I suspect you are running into a race condition or caching issue, i.e. by the time you revert to an older version, the race condition/caches have been resolved, and you'll get the right version.

Please do share the logs requested here if you run into this again #22877 (comment)

[kksoo1769]

I have same issues too. For python's version issue, I removed the file python31 in my env folder. Now everything is working fine, but the message "Python version associated with the selected kernel is no longer supported" still appears.

[DonJayamanne]

@kksoo1769 please can you share the logs as requested in the previous post

[kksoo1769]

@DonJayamanne Here are the logs you requested.
16:14:20.832 [info] Starting Kernel startUsingPythonInterpreter, .jvsc74a57bd0d34c88351d00784b174a0644aef35c2f82b95779f6031b06671f577ef55ac3d9.~\anaconda3\python.exe.~\anaconda3\python.exe.-m#ipykernel_launcher (Python Path: ~\anaconda3\python.exe, Conda, base, 0.0.0) for '~\Desktop\work\test.ipynb' (disableUI=true) 16:14:22.263 [info] Process Execution: ~\anaconda3\python.exe -c "import ipykernel; print(ipykernel.__version__); print("5dc3a68c-e34e-4080-9c3e-2a532b2ccb4d"); print(ipykernel.__file__)" 16:14:22.270 [info] Process Execution: ~\anaconda3\python.exe -m pip list 16:14:22.278 [info] Process Execution: ~\anaconda3\python.exe -m ipykernel_launcher --f=~\AppData\Roaming\jupyter\runtime\kernel-v2-8108IQ3NM62Xp9Mc.json > cwd: ~\Desktop\work 16:14:22.278 [info] Kernel process 4072. 16:14:23.218 [warn] StdErr from Kernel Process 0.00s - Debugger warning: It seems that frozen modules are being used, which may 0.00s - make the debugger miss breakpoints. Please pass -Xfrozen_modules=off 0.00s - to python to disable frozen modules. 0.00s - Note: Debugging will proceed. Set PYDEVD_DISABLE_FILE_VALIDATION=1 to disable this validation. 16:14:23.612 [info] Registering Kernel Completion Provider from kernel base (Python) for language python

[DonJayamanne]

@kksoo1769 please share the logs as per instructions here
#22877 (comment)

[luukburger]

Sorry, I have been on holidays, so no time to report back. I've just started VS Code again and my plugins got updated to:

• Jupyter: v2024.1.1 (last updated 2024-02-08, 19:47:15)
• Python: v2024.0.1 (last updated 2024-02-07, 21:00:51)

Now the issue doesn't occur for me anymore...

[OrhanGuen]

@OrhanGuen @MarkSomogyvari Does this work when you go back to the latest (stable) version of the Jupyter extension. I believe it will, I suspect you are running into a race condition or caching issue, i.e. by the time you revert to an older version, the race condition/caches have been resolved, and you'll get the right version.

Please do share the logs requested here if you run into this again #22877 (comment)

I changed it back to the current version of Jupyter v2024.1.1. Now there is no error anymore.

[DonJayamanne]

I changed it back to the current version of Jupyter v2024.1.1. Now there is no error anymore.

Yes, I suspect this is an issue with the cache.
I.e. first time there's no cache and the information recieved is incorrect. when you reload VS Code, the cache has been updated with the right values.

@karrtikr is there something else users can to do get better logs and replicate this issue, e.g. clera the cache or re-create a new environment? Please advice, as we have a few uesrs running into this and @luabud too ran into this.

[ProcedureOriented]

I guess it is a bug when comparing python version string in extension v2024.1.1. Or does it just get the first digit after dot as version? It disappeared after rolling back to v2024.1.0.

[king159]

Agree with @ProcedureOriented .This bug vanishes after rolling back to v2024.1.0

[DonJayamanne]

@ProcedureOriented @king159
Does it reappear when you install the latest version again 2024.1.1
Please can you try that, if it does repro then this is definitely a regression
Else it’s likely an issue with caching or the like

once again than you for getting back without the necessary information, really helps get to the bottom of this

[ProcedureOriented]

@DonJayamanne
Hi, thank you for your patience. I think the cache influenced its behavior, so it strangely showed this notification again on v2024.1.0 and even v2023.11.1100101639 when I tried after receiving your reply.

Then I tried to quit VS Code after installing each version of Jupyter extension, now I am sure it is a problem only on v2024.1.1, even I installed the version again.

What's more, I found it has another indicator on this problem, on v2024.1.1, python 3.12 version will not show in the bracket after the environment's name, like this

while the v2024.1.0 shows correctly:

Actually, it seems like an occasional problem. When I stay in v2024.1.1, it shows 2 times in I tried 10 times restarting VS Code... I don't know what makes it perform like this.

The output log after changing kernel from 'base' to 'py312' by clicking is here, FYI.

2024-02-20 15:06:49.965 [debug] Found cached env for /usr/bin/python3
2024-02-20 15:06:49.969 [debug] Found cached env for /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python
2024-02-20 15:06:49.970 [error] Invalid version for /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python: {"name":"py312","location":"/opt/homebrew/Caskroom/miniforge/base/envs/py312","kind":"virt-conda","executable":{"filename":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","sysPrefix":"","ctime":1708016995993,"mtime":1708016995993},"display":"Python 3.1.x ('py312')","version":{"major":3,"minor":1,"micro":-1},"arch":1,"distro":{"org":"Anaconda, Inc."},"source":[],"type":"Conda","id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","detailedDisplayName":"Python 3.1.x ('py312': conda)"}
2024-02-20 15:06:49.977 [debug] Found cached env for /usr/bin/python3
2024-02-20 15:06:49.978 [debug] Found cached env for /usr/bin/python3
2024-02-20 15:06:49.984 [debug] Found cached env for /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python
2024-02-20 15:06:49.984 [debug] Found cached env for /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python
2024-02-20 15:06:49.984 [error] Invalid version for /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python: {"name":"py312","location":"/opt/homebrew/Caskroom/miniforge/base/envs/py312","kind":"virt-conda","executable":{"filename":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","sysPrefix":"","ctime":1708016995993,"mtime":1708016995993},"display":"Python 3.1.x ('py312')","version":{"major":3,"minor":1,"micro":-1},"arch":1,"distro":{"org":"Anaconda, Inc."},"source":[],"type":"Conda","id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","detailedDisplayName":"Python 3.1.x ('py312': conda)"}
2024-02-20 15:06:49.984 [error] Invalid version for /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python: {"name":"py312","location":"/opt/homebrew/Caskroom/miniforge/base/envs/py312","kind":"virt-conda","executable":{"filename":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","sysPrefix":"","ctime":1708016995993,"mtime":1708016995993},"display":"Python 3.1.x ('py312')","version":{"major":3,"minor":1,"micro":-1},"arch":1,"distro":{"org":"Anaconda, Inc."},"source":[],"type":"Conda","id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","detailedDisplayName":"Python 3.1.x ('py312': conda)"}
2024-02-20 15:06:50.013 [debug] Extension ms-python.vscode-pylance accessed getActiveEnvironmentPath with args: undefined
2024-02-20 15:06:50.017 [debug] Extension ms-python.vscode-pylance accessed getEnvironmentVariables with args: undefined
2024-02-20 15:06:50.019 [debug] Extension ms-python.vscode-pylance accessed resolveEnvironment with args: {"id":"/usr/bin/python3","path":"/usr/bin/python3"}
2024-02-20 15:06:50.020 [debug] Extension ms-toolsai.jupyter accessed known with args: undefined
2024-02-20 15:06:50.025 [debug] Extension ms-toolsai.jupyter accessed resolveEnvironment with args: {"internal":{"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","executable":{"uri":{"$mid":1,"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","scheme":"file"},"bitness":"Unknown","sysPrefix":""},"environment":{"type":"Conda","name":"py312","folderUri":{"$mid":1,"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312","scheme":"file"}},"version":{"major":3,"minor":1,"micro":-1},"tools":["Conda"]},"id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python"}
2024-02-20 15:06:50.026 [debug] Extension ms-toolsai.jupyter accessed getEnvironmentVariables with args: undefined
2024-02-20 15:06:50.027 [debug] Extension ms-python.vscode-pylance accessed getActiveEnvironmentPath with args: undefined
2024-02-20 15:06:50.031 [debug] Extension ms-python.vscode-pylance accessed resolveEnvironment with args: {"id":"/usr/bin/python3","path":"/usr/bin/python3"}
2024-02-20 15:06:50.032 [debug] Extension ms-python.vscode-pylance accessed getEnvironmentVariables with args: undefined
2024-02-20 15:06:50.033 [debug] Extension ms-python.vscode-pylance accessed getActiveEnvironmentPath with args: undefined
2024-02-20 15:06:50.035 [debug] Extension ms-python.vscode-pylance accessed resolveEnvironment with args: {"id":"/usr/bin/python3","path":"/usr/bin/python3"}
2024-02-20 15:06:50.037 [debug] Extension ms-toolsai.jupyter accessed known with args: undefined
2024-02-20 15:06:50.040 [debug] Extension ms-toolsai.jupyter accessed known with args: undefined
2024-02-20 15:06:50.040 [debug] Extension ms-toolsai.jupyter accessed getEnvironmentVariables with args: undefined
2024-02-20 15:06:50.041 [debug] Extension ms-toolsai.jupyter accessed resolveEnvironment with args: {"internal":{"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","executable":{"uri":{"$mid":1,"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","scheme":"file"},"bitness":"Unknown","sysPrefix":""},"environment":{"type":"Conda","name":"py312","folderUri":{"$mid":1,"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312","scheme":"file"}},"version":{"major":3,"minor":1,"micro":-1},"tools":["Conda"]},"id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python"}
2024-02-20 15:06:50.042 [debug] Extension ms-toolsai.jupyter accessed resolveEnvironment with args: {"internal":{"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","executable":{"uri":{"$mid":1,"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python","scheme":"file"},"bitness":"Unknown","sysPrefix":""},"environment":{"type":"Conda","name":"py312","folderUri":{"$mid":1,"path":"/opt/homebrew/Caskroom/miniforge/base/envs/py312","scheme":"file"}},"version":{"major":3,"minor":1,"micro":-1},"tools":["Conda"]},"id":"/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python"}
2024-02-20 15:06:51.071 [debug] Clearing context for python dependencies not installed: ms-python.python.Python

[karrtikr]

Hi all, can anyone please provide the complete Python logs as requested here: #22877 (comment).

Current info:

• From the partial log in "The version of Python associated with the selected kernel is no longer supported." #22877 (comment) provided by @ProcedureOriented, it seems to be happening with miniforge which we may not support yet: Add support for micromamba environments #20919
• Problem started happening starting v2024.1.1. This seems strange as the only changes we made were related to activation: https://github.com/microsoft/vscode-python/releases/tag/v2024.0.1.
• I notice the logs contain {"major":3,"minor":1,"micro":-1}, where micro version contains "-1", based on the complete logs I can investigate why that is the case.

[karrtikr]

A potential fix should be out in the pre-release version of the extension, use the following to try it out:

Please report back with "debug" level logs if the issue re-occurs.

[DonJayamanne]

an you provide the trace/debug level logs please?

Sorry, it was a different machine and I was doing some other work when I ran into this.
Once again, didn't have trace logging enabled, sorry.

[ProcedureOriented]

@karrtikr
Hi, here is the log with "[error] Invalid version for ...",
Python.log

And then I installed the latest pre-release Python extension, the problem solved when editing .py files, but the notification still pop up when open jupyter notebook files. However I could not find the same string in the log whatever actions I try.
Python_withPR.log

I noticed the Jupyter log contains

"envs/py312/python.-m#ipykernel_launcher (Python Path: /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python, Conda, py312, 3.1.0) for ..."

I don't know why "3.1.0" appears, so I attach the jupyter log if it helps.
1-Jupyter.log

More, I found the kernel display in notebook seems different from the release version Python before.
All kernels lose version number after running Python: Clear cache and reload window,

but they shows correctly in the list chose by: "Select Another Kernel..." > "Python Environments ...", and if I choose 'py312' again, the kernel shows version number in change list. But the notification still pops up when I switch to it from other kernel.

I also found if I restart VS Code once or twice without running Python: Clear cache and reload window again, the problem won't occur after that, so I think it might not confuse me when working with the pre-release version now. 🤔

[karrtikr]

Thanks for the info @ProcedureOriented. Based on the logs Python extension seems to resolve the version correctly as "3.12.1", when using resolveEnvironment API.

Starting Kernel startUsingPythonInterpreter, ... "envs/py312/python.-m#ipykernel_launcher (Python Path: /opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python, Conda, py312, 3.1.0) for ..."

@DonJayamanne In the logs pointed out containing "3.1.0", could it be the case that Jupyter is not using the results from "resolveEnvironment" API to get this version? As results from API seem to be correct based on the logs.

[karrtikr]

FWIW Python extension might be returning "3.1.0" version via .known API at one point. Before starting a kernel, I think if we make sure to use version returned by "resolveEnvironment" API, it should solve the problem.

[karrtikr]

It also might be the case that Jupyter has its own cache containing "3.1.0" which isn't cleared?

[DonJayamanne]

It also might be the case that Jupyter has its own cache containing "3.1.0" which isn't cleared?

If the details are available in environments.known, we use that, else we get it from resolveEnvironment..
There's no caching in Jupyter.
There is, but unlikely to be the cause, as we've always call resolve & handle change events...

Also, the versions are not cached, version information is always retrieved from API,
we only cache the display name, id, uri, envname, envpath (this cache be removed soon, part of microsoft/vscode-jupyter#14038)

[karrtikr]

If the details are available in environments.known, we use that, else we get it from resolveEnvironment..

environments.known may contain partial details (in this case, incorrect, which seems to be a bug). For the selected kernel please always use resolveEnvironment to get the version, and that should solve the issue.

[DonJayamanne]

environments.known may contain partial details (in this case, incorrect, which seems to be a bug). For the selected kernel please always use resolveEnvironment to get the version, and that should solve the issue.

I do not think I can use resolve Environment easily, that will require a big change.
I'm trying to avoid using resolveEnv if we already have the infonrmation.

Thanks, I'd rather wait for the fix, rather than patching our code as work around (that just adds more debt, as it is, we have a few layers from the old Pthon API types that we're still trying to get rid of)

[karrtikr]

I am not sure of what exactly causes us to return "3.1.0" in the first place, we approximately figure out version based on nearby files, conda-meta etc. Version returned by looking around is not guaranteed to be reliable, by design.

Consumer should be await on resolveEnvironment API for complete information, before making important decisions such as what is the Python version of a kernel before running. I don't consider this being a patch on Jupyter extension's side.

[DonJayamanne]

I am not sure of what exactly causes us to return "3.1.0" in the first place, we approximately figure out version based on nearby files, conda-meta etc. Version returned by looking around is not guaranteed to be reliable, by design.

Isn't this a regression? this used to work?

[karrtikr]

@ProcedureOriented Can you please install the latest pre-release and provide me with the Python debug level logs after running Python: Clear cache command?

[ProcedureOriented]

@karrtikr
Hi, here are Python and Jupyter logs, log level is Trace.
Python: v2024.3.10581547 (pre-release)
Jupyter: v2024.1.2024012502 (pre-release)

Python.log
1-Jupyter.log

[karrtikr]

@ProcedureOriented Great, it seems that this portion is returning the version as "{"major":3,"minor":1,"micro":-1}" when "/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin/python" is passed into it which is the cause.

vscode-python/src/client/pythonEnvironments/common/commonUtils.ts

Lines 218 to 233 in d0a07de

	async function getPythonVersionFromNearByFiles(interpreterPath: string): Promise<PythonVersion> {
	const root = path.dirname(interpreterPath);
	let version = UNKNOWN_PYTHON_VERSION;
	for await (const entry of findInterpretersInDir(root)) {
	const { filename } = entry;
	try {
	const curVersion = parseVersion(path.basename(filename));
	if (comparePythonVersionSpecificity(curVersion, version) > 0) {
	version = curVersion;
	}
	} catch (ex) {
	// Ignore any parse errors
	}
	}
	return version;
	}

Can you provide the contents of "/opt/homebrew/Caskroom/miniforge/base/envs/py312/bin" directory? We're primarily interested in all the python binaries present inside the directory, based on whose filenames, we figure an approximate version.

[ProcedureOriented]

@karrtikr
Here it is. Btw I noticed there are alias named python, python3, and python3.1 in it. I guess whether the alias cause the bug.
ls_la.txt

[karrtikr]

Yep, "python3.1" is an incorrect naming for that alias which is causing the bug. The tool which created the alias likely didn't account for multidigit minor versions, which is leading to the bug. I advise to either change the name or file an issue on miniforge.

This is why version can only be best effort guess and not always reliable cc/ @DonJayamanne

Closing this as cause for the bug lies somewhere else, thanks for all the help @ProcedureOriented .