You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Debugging leaves terminal in conda's Base environment instead of selected (interpreter's) conda environment #896
I am debugging my (pytest) test using Testing > select test > Debug Test GUI.
Test is executing
After test finishes it leaves terminal in Base conda environment instead of selected (interpreter's) environment.
As I end up in base environment, sometimes I execute poetry install which installs current project's dependencies in base environment. Sometimes I run pytest in base environment but it is not installed there and so on.
After interpreter is selected, whatever we do, we should do it inside that selected environment.
Any terminal windows that are opening during any process should use selected environment.
We should never end up in terminal in Base conda's environment.
Currently, I can't afford using experimental extension at work.
Until this problem is solved I have to execute exit in terminal whenever I want to do something in project's environment.
Really annoying.
Thanks again for reporting this issue. Based on what you described, this appears to touch the same parts of our environment/activation logic that we’ve recently replaced in our new Python Environments extension. A lot of the behavior in this area now follows a completely different activation and environment-resolution path, so this may already be fixed.
We’d really appreciate it if you could try the new extension and let us know—this helps us confirm that the new implementation addresses older issues like yours.
👉 Enable the new extension with "python.useEnvironmentsExtension": true (then you should see a Python icon in the sidebar.
📫 More details on enablement here
📝 Useful background documentation on activation here
If you do run into issues, please include as much info and reproduction steps or the "Report issue..." command inside VS Code!
Thanks again for reporting this originally, and for helping us validate the new environment direction for Python in VS Code 🙌
Activity
karthiknadig commentedon Apr 30, 2025
zljubisic Can you try with the https://marketplace.visualstudio.com/items?itemName=ms-python.vscode-python-envs extension? You may need to set
python.condaPathtocondawith it. That is temporary.That is a experimental extension that will eventually replace environment management and handling, it should do a bit better than this extension.
zljubisic commentedon Apr 30, 2025
Currently, I can't afford using experimental extension at work.
Until this problem is solved I have to execute exit in terminal whenever I want to do something in project's environment.
Really annoying.
zljubisic commentedon May 6, 2025
I believe that we can conclude here that whatever we do within our project should be done in selected environment.
mumiao2000 commentedon May 11, 2025
No help with experimental or historical extension. It seems due to VSCode 1.100.0 version.
Everything goes well except the terminal.
zljubisic commentedon Oct 8, 2025
Karthik Nadig (@karthiknadig) Any progress with this issue?
eleanorjboyd commentedon Dec 12, 2025
Hi there! 👋
Thanks again for reporting this issue. Based on what you described, this appears to touch the same parts of our environment/activation logic that we’ve recently replaced in our new Python Environments extension. A lot of the behavior in this area now follows a completely different activation and environment-resolution path, so this may already be fixed.
We’d really appreciate it if you could try the new extension and let us know—this helps us confirm that the new implementation addresses older issues like yours.
👉 Enable the new extension with
"python.useEnvironmentsExtension": true(then you should see aPythonicon in the sidebar.📫 More details on enablement here
📝 Useful background documentation on activation here
If you do run into issues, please include as much info and reproduction steps or the "Report issue..." command inside VS Code!
Thanks again for reporting this originally, and for helping us validate the new environment direction for Python in VS Code 🙌