[surlymochan]

Fixes #3210.

Why (in your own words)

gstack-pregate --only lanes exits PREGATE_NO_LANE when the only touched files are planning or evidence Markdown and the base branch has no .gstack/pregate.json to mark them as an empty dependency. A .work plan and a redproof note are documentation of the change, not test subjects, so a docs-only branch should not fail the lane check. Those paths are now an explicit docs/no-tests empty lane. A source file that still maps to no test keeps failing.

Live evidence

Before, on a fixture whose base has no .gstack/pregate.json, gstack-pregate --base main --allow-repo-commands --only lanes after committing .work/round-3-plan.md, notes/redproof.md, and docs/evidence.md:

tests      lanes     fail    3 touched file(s) with no test lane: .work/round-3-plan.md docs/evidence.md notes/redproof.md (PREGATE_NO_LANE)
    lanes: FAIL lanes — .work/round-3-plan.md has no test lane
    lanes: FAIL lanes — docs/evidence.md has no test lane
    lanes: FAIL lanes — notes/redproof.md has no test lane

After the same command:

tests      lanes     pass    docs/no-tests: 3 documentation file(s); nothing touched selects a test
    lanes: docs/no-tests (empty lane): .work/round-3-plan.md docs/evidence.md notes/redproof.md
PREGATE: pass

An unmapped lib/orphan.js beside that plan still fails, and the plan is not in the failure:

tests      lanes     fail    1 touched file(s) with no test lane: lib/orphan.js (PREGATE_NO_LANE)
    lanes: FAIL lanes — lib/orphan.js has no test lane
    lanes: docs/no-tests (empty lane): .work/round-3-plan.md

A test that names a Markdown file still selects and runs that test. empty_lanes in .gstack/pregate.json adds more documentation globs. It cannot clear a source file or a file under a test directory.

Scope

• Changed: built-in docs/no-tests classification for planning and evidence Markdown (docs trees, .work plans, plan / redproof / evidence names, README and the usual root prose), plus optional empty_lanes globs. lanes prints the class and does not emit PREGATE_NO_LANE for it.
• Verified live by: the fixture commands above; bun test test/pregate-selection.test.ts (23 pass); adjacent pregate, module-size, and troubleshooting-anchor tests (182 pass); bun run typecheck clean.
• Did NOT test: paid evals. bun run test:quick was run here and failed on missing Playwright/Chromium, an unbuilt CSO helper, and git timeouts. None of the failing files are this lane change.

Liveness proof

No live GSTACK PR screenshot from the contributor's machine is attached with this open. Add that if the evidence bar requires it before review.

Checklist

• Liveness screenshot attached (GSTACK PR typed live into a real surface) or PR author is the repository owner
• This is not a generated-file-only diff (I edited the source/template and regenerated)
• No ETHOS.md edits, and no changes to voice / founder perspective / YC references
• New public command / external service / host adapter has an accepted issue linked (or N/A)
• Linked issue or reproduction: pregate lane check rejects documentation-only changes without a lane #3210

  

[trunk-io]

Merging to main in this repository is managed by Trunk.

• To merge this pull request, check the box to the left or comment /trunk merge below.

After your PR is submitted to the merge queue, this comment will be automatically updated with its status. If the PR fails, failure details will also be posted here

[capy-ai]

Triage (fix wave 5, GSTA-37): thanks. Treating docs as an empty lane is the right direction for #3210. It can't land in this wave: 7 of its files (lanes.ts, config.ts, lane-selection.ts, docs/pregate.md, docs/troubleshooting.md, lib/result-codes.ts, the size ratchet) are being changed by open PRs #3190 and #3209. It'll be reviewed against main after those land. Expect a merge from main first, and one design point: an unmarked .md that a test actually reads still needs to select that test.

[surlymochan]

Thanks! Pushed 4f422e5: lane selection checks the import graph first, so a .md that a test references now selects that test. Only Markdown with no test mapping falls into the docs/no-tests lane, and unmapped source still fails with PREGATE_NO_LANE. I'll rebase once #3190 lands.

[capy-ai]

Thanks for this. Both commits (e8793e5 and 7aabbdd) were reviewed and are being carried unchanged, with Cursor Agent kept as author, in the field-list-2 wave B PR, together with an added test for the reviewer edges (test/pregate-docs-lane-review.test.ts). The fix lands through wave B instead of separately; this PR will be closed as superseded, with credit, when B merges. I'll link B here once it's open.

[capy-ai]

Wave B is open as #3235; it carries this PR's two commits unchanged with Cursor Agent as author.

[capy-ai]

Closing as superseded by #3235 (v1.91.116.0). Both commits from this PR (e8793e5 and 7aabbdd, the pregate docs/no-tests empty lane) are carried there unchanged, with Cursor Agent kept as the author, plus a reviewer-edge test. Thank you for the fix. Per GStack Minder, it lands through wave B rather than merging separately.