[bmcnaboe]

Regression tests for the hook augment payload-cwd -> project resolution across a linked git worktree, on a real worktree fixture:

• payload cwd inside an indexed linked worktree (and a subdir) resolves to that worktree own project, never a sibling checkout or nothing;
• an unindexed worktree gets a deliberate "no indexed graph project matched" notice, never a silent 0 bytes;
• main-checkout resolution is unchanged.

Current main already resolves worktrees correctly; this adds the coverage that was missing so it cannot regress. Opened as a draft.

Fixes #1752

[github-actions]

Thanks for opening this — it has been seen, and it is queued.

This note is automated, but it is not a brush-off: it exists so you know where your PR stands instead of having to guess from silence.

Current review status: working through a backlog. 0.9.1-rc.1 is out, so the release freeze that held reviews is over — but it left a large queue of open pull requests behind it, and we are reading through them oldest-first. The background is in discussion #1144.

What that means for this PR, concretely:

• It will not be closed for inactivity. No stale bot touches pull requests here.
• It may still sit a while before a human reads it. That is on us, not on you.
• Older PRs are read first, so a recent one is not being skipped — it is behind a queue.

Things that will genuinely speed it up whenever review does happen:

• Keep it rebased on main — the tree is moving quickly right now, and a conflicting branch cannot be reviewed as the diff you intended.
• Get CI green, or say which failures you believe are pre-existing.
• Keep the change to one claim. Bundled features and refactors get split before they get merged, which costs you a round trip.
• Every commit needs a sign-off (git commit -s) — CI enforces DCO.

If this fixes a bug, a reproduction we can run is worth more than a description of the symptom.

Thanks for contributing, and sorry in advance for the wait.

[DeusData]

Thank you for adding deterministic linked-worktree coverage while leaving current production behavior unchanged. The positive, unindexed, and main-checkout controls give this a clear maintenance boundary. The contribution queue is quite full, so review may take a little time, but we will come back with feedback as soon as possible.

[DeusData]

Picking this back up, with thanks for the patience: the branch now conflicts with main (test_cli.c moved a lot underneath it in the last two weeks), so it cannot be evaluated as-is. Could you rebase onto current main and keep the sign-off? Once it applies cleanly I will take the host-side clearance for the four system() sites in the test (they are the same shape as the fourteen existing ones in the suite) and give it a proper review straight away.

[bmcnaboe]

Rebased onto current main (b5185d1), sign-off kept. Same two tests, one fewer hunk: the forward declaration of cbm_project_name_from_path is gone since pipeline.h already provides it. cli suite passes locally (311). The earlier test-windows-guards failure was the guard's own fixture ("ASCII baseline did not index"), not these tests, which are compiled out on Windows.

[DeusData]

Merged — thank you, and sorry for the long wait after you had the last word.

Verified on the merge result with the current main: the cli suite passes, including both of your new tests, and main has not touched tests/test_cli.c in a way that overlaps them. The shape is right for what they pin: each fixture works in its own mkdtemp directory, cleans up before its assertions, and carries a control on a plain main checkout, so a failure points at the behaviour rather than at the fixture.

One small follow-up worth knowing, not a blocker: the fixtures' git commit inherits the user's global git configuration, so on a machine with commit.gpgsign=true they would fail rather than skip. Passing -c commit.gpgsign=false (or setting it in the fixture repository's local config) would make them independent of whoever runs them. Happy to take that in a tiny follow-up if you prefer not to.