Fix Python 3.11 test failures and mask-free CI failure reporting

[Copilot]

Two bugs caused Python 3.11 CI runs to silently report 158 test failures as success.

super() in list comprehension — Python 3.11 incompatibility

In Python ≤ 3.11, list comprehensions run in a nested function scope. Bare super() relies on a __class__ cell that isn't propagated into that scope, raising TypeError: super(type, obj): obj must be an instance or subtype of type. PEP 709 (Python 3.12) inlined comprehensions, which is why this only breaks on 3.11.

Stoner/core/base.py — TypeHintedDict.__getitem__:

# Before — fails on Python 3.11
value = [super().__getitem__(nm) for nm in name]

# After — capture super() in enclosing scope first
_super = super()
value = [_super.__getitem__(nm) for nm in name]

CI exit-code masking

The "Test with xvfb" step runs under bash -l {0} (no -e). When pytest exits non-zero, coverage xml still runs and returns 0, so the step — and the whole job — reports success despite test failures.

.github/workflows/run-tests-action.yaml:

# Before
xvfb-run --auto-servernum pytest -n 2 --cov-report= --cov=Stoner --junitxml pytest.xml
coverage xml

# After — propagate pytest's exit code while still generating coverage
xvfb-run --auto-servernum pytest -n 2 --cov-report= --cov=Stoner --junitxml pytest.xml || PYTEST_EXIT=$?
coverage xml
exit ${PYTEST_EXIT:-0}

---

⌨️ Start Copilot coding agent tasks without leaving your editor — available in VS Code, Visual Studio, JetBrains IDEs and Eclipse.

[github-actions]

Test Results

    4 files  +  1      4 suites  +1   22m 55s ⏱️ + 5m 3s
  337 tests + 46    161 ✅ + 46  0 💤 ±0  158 ❌ ±0  18 🔥 ±0 
1 284 runs  +469  1 103 ✅ +469  0 💤 ±0  158 ❌ ±0  23 🔥 ±0 

For more details on these failures and errors, see this check.

Results for commit 114151a. ± Comparison against base commit 6cb282d.