The drive/mounts work added new behavior that ships with thin regression coverage. The pre-merge fixes (#5411/#5412) add tests for the specific bugs they close, but two gaps remain worth pinning.
The goal is to lock the highest-value untested behaviors against regression.
Follow-up from the #5400 backend review (test-scaffolding fixes stacked in #5411/#5412). Cited findings: 2.4 (the new endpoints' behavior was untested and the PR body overstated coverage) and 3.3 (_rollup_recent_entries is clever, silent in edge cases, and had zero tests). The review report is not in the repo, so this summary stands alone.
The drive/mounts work added new behavior that ships with thin regression coverage. The pre-merge fixes (#5411/#5412) add tests for the specific bugs they close, but two gaps remain worth pinning.
The goal is to lock the highest-value untested behaviors against regression.
api/oss/tests/pytest/acceptance/mounts/); today it gained nothing for them._rollup_recent_entriesif the pre-merge tests did not already cover it — a pureList[MountFile] -> List[MountFile]function with subtle boundary semantics.Follow-up from the #5400 backend review (test-scaffolding fixes stacked in #5411/#5412). Cited findings: 2.4 (the new endpoints' behavior was untested and the PR body overstated coverage) and 3.3 (
_rollup_recent_entriesis clever, silent in edge cases, and had zero tests). The review report is not in the repo, so this summary stands alone.