Repro
```
git checkout main
pytest tests/test_setup_runtime_checks.py::SetupReportTest::test_run_setup_reports_runtime_and_release_notes
```
Fails on origin/main alone.
Symptom
```
AssertionError: 'big release' not found in ()
```
`report.release_notes` is an empty tuple even though the test fixture writes a `CHANGELOG.md` containing `'## 9.9.9\n- big release\n'` and calls `run_setup(... last_seen_version='0.0.1')`.
What I'd guess (haven't dug)
Either `run_setup` doesn't read CHANGELOG.md anymore, or the version comparison `9.9.9 > 0.0.1` is failing the gate that decides what counts as 'new release notes', or the parsing was refactored without updating the test.
The runtime-checks half of the same test (`assertGreaterEqual(len(report.runtime_checks), 3)`) appears to pass — the failure is specifically the release-notes path.
Why I'm filing
Hit this while verifying merge-conflict resolutions on a feature branch (PR #32). Reproduces on pure origin/main with no local changes.
Repro
```
git checkout main
pytest tests/test_setup_runtime_checks.py::SetupReportTest::test_run_setup_reports_runtime_and_release_notes
```
Fails on origin/main alone.
Symptom
```
AssertionError: 'big release' not found in ()
```
`report.release_notes` is an empty tuple even though the test fixture writes a `CHANGELOG.md` containing `'## 9.9.9\n- big release\n'` and calls `run_setup(... last_seen_version='0.0.1')`.
What I'd guess (haven't dug)
Either `run_setup` doesn't read CHANGELOG.md anymore, or the version comparison `9.9.9 > 0.0.1` is failing the gate that decides what counts as 'new release notes', or the parsing was refactored without updating the test.
The runtime-checks half of the same test (`assertGreaterEqual(len(report.runtime_checks), 3)`) appears to pass — the failure is specifically the release-notes path.
Why I'm filing
Hit this while verifying merge-conflict resolutions on a feature branch (PR #32). Reproduces on pure origin/main with no local changes.