Caching interferes with detecting a new python interpreter

Issue Type: Bug

I’m changing my desired Python interpreter by putting it in $PATH. Because I’ve already used vscode for my workspace, the Python extension appears to be using the cached python interpreter (afaict that’s coming from state.vscdb -> ms-python.python -> PYTHON_ENV_INFO_CACHE). I don’t see any way to disable that caching. Ideally there would be some way to configure the extension to respect $PATH or disable caching.

Diagnostic data

• Python version (& distribution if applicable, e.g. Anaconda): 3.9.12
• Type of virtual environment used (e.g. conda, venv, virtualenv, etc.): Global
• Value of the python.languageServer setting: Default

User Settings

defaultLS: {"defaultLSType":"Jedi"}

downloadLanguageServer: true

envFile: "<placeholder>"

venvPath: "<placeholder>"

venvFolders: "<placeholder>"

condaPath: "<placeholder>"

pipenvPath: "<placeholder>"

poetryPath: "<placeholder>"

languageServer: "Jedi"

linting
• enabled: true
• cwd: "<placeholder>"
• Flake8Args: "<placeholder>"
• flake8Enabled: false
• flake8Path: "<placeholder>"
• lintOnSave: true
• banditArgs: "<placeholder>"
• banditEnabled: false
• banditPath: "<placeholder>"
• mypyArgs: "<placeholder>"
• mypyEnabled: false
• mypyPath: "<placeholder>"
• pycodestyleArgs: "<placeholder>"
• pycodestyleEnabled: false
• pycodestylePath: "<placeholder>"
• prospectorArgs: "<placeholder>"
• prospectorEnabled: false
• prospectorPath: "<placeholder>"
• pydocstyleArgs: "<placeholder>"
• pydocstyleEnabled: false
• pydocstylePath: "<placeholder>"
• pylamaArgs: "<placeholder>"
• pylamaEnabled: false
• pylamaPath: "<placeholder>"
• pylintArgs: "<placeholder>"
• pylintPath: "<placeholder>"

sortImports
• args: "<placeholder>"
• path: "<placeholder>"

formatting
• autopep8Args: "<placeholder>"
• autopep8Path: "<placeholder>"
• provider: "autopep8"
• blackArgs: "<placeholder>"
• blackPath: "<placeholder>"
• yapfArgs: "<placeholder>"
• yapfPath: "<placeholder>"

testing
• cwd: "<placeholder>"
• debugPort: 3000
• nosetestArgs: "<placeholder>"
• nosetestsEnabled: undefined
• nosetestPath: "<placeholder>"
• promptToConfigure: true
• pytestArgs: "<placeholder>"
• pytestEnabled: false
• pytestPath: "<placeholder>"
• unittestArgs: "<placeholder>"
• unittestEnabled: false
• autoTestDiscoverOnSaveEnabled: true

terminal
• activateEnvironment: true
• executeInFileDir: "<placeholder>"
• launchArgs: "<placeholder>"

experiments
• enabled: true
• optInto: []
• optOutFrom: []

insidersChannel: "off"

tensorBoard
• logDirectory: "<placeholder>"

Extension version: 2022.0.1814523869 VS Code version: Code 1.67.1 (da15b6fd3ef856477bf6f4fb29ba1b7af717770d, 2022-05-06T12:37:14.619Z) OS version: Darwin arm64 21.4.0 Restricted Mode: No

System Info

Item	Value
CPUs	Apple M1 Pro (10 x 24)
GPU Status	2d_canvas: enabled canvas_oop_rasterization: disabled_off direct_rendering_display_compositor: disabled_off_ok gpu_compositing: enabled metal: disabled_off multiple_raster_threads: enabled_on oop_rasterization: enabled opengl: enabled_on rasterization: enabled raw_draw: disabled_off_ok skia_renderer: enabled_on video_decode: enabled video_encode: enabled webgl: enabled webgl2: enabled
Load (avg)	3, 3, 2
Memory (System)	16.00GB (0.44GB free)
Process Argv	–extensions-dir /nix/store/53k4y7l6nb68m0j9iqmdzvgg6rxp2jsa-vscode-extensions/share/vscode/extensions . --crash-reporter-id 06f4509c-da16-4f2d-ac50-cffb467932e2
Screen Reader	no
VM	0%

A/B Experiments

vsliv368cf:30146710
vsreu685:30147344
python383:30185418
vspor879:30202332
vspor708:30202333
vspor363:30204092
pythonvspyl392:30443607
pythontb:30283811
pythonptprofiler:30281270
vshan820:30294714
vstes263:30335439
vscorecescf:30445987
pythondataviewer:30285071
vscod805:30301674
binariesv615:30325510
bridge0708:30335490
bridge0723:30353136
vsaa593cf:30376535
vsc1dst:30438360
pythonvs932:30410667
wslgetstarted:30449410
vscscmwlcmt:30465135
cppdebug:30492333
vsclangdf:30486550

[karrtikr]

Sure, I think it’s better if you open it as a separate issue because of the amount of the verbose in this one. So as I understand your feature request would be: Respect PATH while selecting the interpreter if no interpreter has been explicitly specified by the user

You can list alternative options in the issue description, thanks!

[mkenigs]

So is that desirable behavior? Switching interpreters requires manually clearing extension storage and restarting extension host?

We just want a way to specify the interpreter for a single instance of code - is there any way to do that you would recommend or support? Ideally we could set the interpreter through python from PATH, or a command line flag or an env var would also work. A possible workaround would even be respecting python.defaultInterpreterPath in .vscode/settings.json. Right now, even that setting won’t be respected without clearing workspace interpreter settings.