Activity

1. github-actions

   added

   triage-neededNeeds assignment to the proper sub-teamNeeds assignment to the proper sub-team

   on Feb 28, 2026
2. Eleanor Boyd (eleanorjboyd)

   self-assigned this

   on Mar 3, 2026
3. Eleanor Boyd (eleanorjboyd)

   added

   bugIssue identified by VS Code Team member as probable bugIssue identified by VS Code Team member as probable bug

   area-testing

   needs PRReady to be worked onReady to be worked on

   and removed

   triage-neededNeeds assignment to the proper sub-teamNeeds assignment to the proper sub-team

   on Mar 3, 2026
4. jaap3 commented on Mar 5, 2026

   

   Jaap Roes (jaap3)

   on Mar 5, 2026

   More actions

   Seeing the same: crosslinking pytest-dev/pytest-subtests#184 which was originally created by Eleanor Boyd (@eleanorjboyd) in the now archived pytest-subtests repository
5. jaap3 commented on Mar 5, 2026

   

   Jaap Roes (jaap3)

   on Mar 5, 2026

   More actions

   For now it is possible to disable subtests in vscode using the following snippet (inspired by the workaround in the pytest-subtests issue)

   # tests/conftest.py
   
   import importlib.util
   import contextlib
   
   @pytest.fixture
   def subtests(subtests):
       if importlib.util.find_spec('debugpy') is not None:
           # VS Code's test runner doesn't report failed subtests, disable them to avoid confusion.
           # See: https://github.com/microsoft/vscode-python/issues/25824
           subtests.test = contextlib.nullcontext
       yield subtests
6. loic-simon commented on May 15, 2026

   

   Loïc Simon (loic-simon)

   on May 15, 2026

   Last edited by loic-simon

   More actions

   Another workaround, that allows subtests to work and report a failure if any subtest failed:

   # tests/conftest.py
   
   from collections.abc import Generator
   import pytest
   
   _calling = set[str]()
   
   @pytest.hookimpl(wrapper=True)
   def pytest_runtest_call(item: pytest.Item) -> Generator[None]:
       """https://docs.pytest.org/en/stable/reference/reference.html#pytest.hookspec.pytest_runtest_call"""
       _calling.add(item.nodeid)
       yield
       _calling.remove(item.nodeid)
   
   @pytest.hookimpl(wrapper=True)
   def pytest_runtest_makereport(item: pytest.Item, call: pytest.CallInfo) -> Generator[None, pytest.TestReport, pytest.TestReport]:
       """https://docs.pytest.org/en/stable/reference/reference.html#pytest.hookspec.pytest_runtest_makereport"""
       report = yield
       if call.when == "call" and item.nodeid in _calling:
           report.nodeid += "(subtest)"
       return report

   That work because subtests are (AFAIK) the only time where reports are generated while the item is still running, so we detect that and alter the reported nodeid, so VSCode doesn't take it as the item result, but wait for the final result.

   I made this quickly by playing with pytest hooks, I can't guarantee there are no side-effects, but so far it worked great!

   👍React with 👍3Reacted by Jaap Roes, Michael Brundage and Matt Huisman
7. amesser-hilscher commented on Jul 3, 2026

   

   Andreas Messer (amesser-hilscher)

   on Jul 3, 2026

   More actions

   Another workaround, that allows subtests to work and report a failure if any subtest failed:

   This breaks junit output of pytest and probably even more. Tests without subtest now show up twice in junit xml files here.