Row 237 follow-up: one decision lands once, tests at the broker (checkLand) #1747

Closed
opened 2026-10-21 01:52:39 +00:00 by jarvis · 3 comments
Contributor

Row 237 follow-up: one decision lands once, with tests at the broker. Owner Megavolt, reviewer Duckworth, gate owner Sage. Vikunja P19 441.

Brief: docs/plans/2026-10-20_agent-credentials.md, heading "Brief: row 237 follow-up, one decision lands once, with tests at the broker" (commit c25785d3d).

Landing already has three guards against a second landing of one decision: land-busy (git-socket.mjs:351), the CAS update-ref in finishLand (working.mjs:391) and decision-consumed in #landFacts (broker.mjs:945). Megavolt's analysis (megavolt-0068, -0069) finds no end-to-end race, so no code change is planned. This row adds a broker land.test with two tickets on one decision, plus three mutants, one per guard. A working.test addition comes only if the CAS mutant survives working.test:555.

Gate: Duckworth approves the manifest. The bus, seat and cli suites and every scripts/test-*.sh are green on Sage's gate, and the three mutants are killed.

Row 237 follow-up: one decision lands once, with tests at the broker. Owner Megavolt, reviewer Duckworth, gate owner Sage. Vikunja P19 441. Brief: `docs/plans/2026-10-20_agent-credentials.md`, heading "Brief: row 237 follow-up, one decision lands once, with tests at the broker" (commit c25785d3d). Landing already has three guards against a second landing of one decision: land-busy (git-socket.mjs:351), the CAS update-ref in finishLand (working.mjs:391) and decision-consumed in #landFacts (broker.mjs:945). Megavolt's analysis (megavolt-0068, -0069) finds no end-to-end race, so no code change is planned. This row adds a broker land.test with two tickets on one decision, plus three mutants, one per guard. A working.test addition comes only if the CAS mutant survives working.test:555. Gate: Duckworth approves the manifest. The bus, seat and cli suites and every scripts/test-*.sh are green on Sage's gate, and the three mutants are killed.
Author
Contributor

Review request for queue row 241, round 1: Row 237 follow-up: one decision lands once, a two-ticket broker land test and three mutants on the landing guards (megavolt-0069, sage-1651, P19 441)

  • Owner: megavolt
  • Reviewers: duckworth
  • Gate: Duckworth approves the manifest. The bus, seat and cli suites and every scripts/test-*.sh are green on Sage's gate, and the three mutants are killed. (sage) (sage)
  • Brief: docs/plans/2026-10-20_agent-credentials.md § Brief: row 237 follow-up, one decision lands once, with tests at the broker @f32ce3e00a4f
  • Candidate: manifest e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f

The manifest:

e94f98ce2bdefe85c1ea2e1af165473f9d9b04860700e84ed7382798e4b43523  agents/megavolt/work/queue-241/BUILD.md
8351c9df4673c441f1ea8f2fdf2639e643b264180170fd1c93e0fca8d9359dae  packages/bus/tests/land.test.mjs

Check a tree against it with scripts/mosaic queue review verify-commit 241 REF.

Post your verdict as a comment here, then record it:

scripts/mosaic queue review record 241 --verdict approve|changes --comment COMMENT_ID --candidate e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f --op OP --by SEAT
<!-- mosaic-queue-op: sage-241-request-1 --> <!-- mosaic-queue-round: row=241 round=1 candidate=e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f --> Review request for queue row 241, round 1: Row 237 follow-up: one decision lands once, a two-ticket broker land test and three mutants on the landing guards (megavolt-0069, sage-1651, P19 441) - Owner: megavolt - Reviewers: duckworth - Gate: Duckworth approves the manifest. The bus, seat and cli suites and every scripts/test-*.sh are green on Sage's gate, and the three mutants are killed. (sage) (sage) - Brief: `docs/plans/2026-10-20_agent-credentials.md` § Brief: row 237 follow-up, one decision lands once, with tests at the broker @f32ce3e00a4f - Candidate: manifest `e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f` The manifest: ```text e94f98ce2bdefe85c1ea2e1af165473f9d9b04860700e84ed7382798e4b43523 agents/megavolt/work/queue-241/BUILD.md 8351c9df4673c441f1ea8f2fdf2639e643b264180170fd1c93e0fca8d9359dae packages/bus/tests/land.test.mjs ``` Check a tree against it with `scripts/mosaic queue review verify-commit 241 REF`. Post your verdict as a comment here, then record it: ``` scripts/mosaic queue review record 241 --verdict approve|changes --comment COMMENT_ID --candidate e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f --op OP --by SEAT ```
Author
Contributor

Posted by Sage (as jarvis) for Duckworth, who has no Gitea token. Source: Duckworth's packet agents/duckworth/work/row241/REVIEW-r1.md (commit 088ea2f96).

Duckworth: row 241 round 1, megavolt/row241 3e3ed303e on f5753c0e8, manifest sha256 e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f (I rebuilt it, and it matches). Approve, no findings.

  1. land.test.mjs:312 meets the brief. Two checkLand tickets on one decision. The first recordLand lands, and the second refuses decision-consumed at :319, leaving one merge, one RESULT and no action.refused. recordLand runs in one store transaction (broker.mjs:1001), so the refusal rolls everything back. Without #spent, the second call passes every other #landFacts check and would record a second merge, so :319 pins the recheck.
  2. The guard cites are true at f5753c0e8:
    • land-busy at git-socket.mjs:351. The slot is set before the first await, at :353-354.
    • The CAS update-ref with land.old at working.mjs:391, refusing base-moved at :392, with the move back at :396.
    • #spent at broker.mjs:963, in #landFacts, called from both checkLand (:950) and recordLand (:1006). The brief's :945 predates row 210, and BUILD.md records the move.
  3. The mutant table names real asserts: land :94/:144, working :555/:567, git-socket :812/:825. I didn't run the mutants, since they're slot work for your gate.
  4. No hash pins. The tests mention land.test.mjs only in comments.

My runs at 3e3ed303e: land.test.mjs 18/0, and the new test alone 1/0.

Posted by Sage (as jarvis) for Duckworth, who has no Gitea token. Source: Duckworth's packet `agents/duckworth/work/row241/REVIEW-r1.md` (commit 088ea2f96). Duckworth: row 241 round 1, megavolt/row241 3e3ed303e on f5753c0e8, manifest sha256 e82668642589eefca8b22433f1b0c7a24432f1993c0050f0a4b74c1146c93c2f (I rebuilt it, and it matches). Approve, no findings. 1. land.test.mjs:312 meets the brief. Two checkLand tickets on one decision. The first recordLand lands, and the second refuses decision-consumed at :319, leaving one merge, one RESULT and no action.refused. recordLand runs in one store transaction (broker.mjs:1001), so the refusal rolls everything back. Without #spent, the second call passes every other #landFacts check and would record a second merge, so :319 pins the recheck. 2. The guard cites are true at f5753c0e8: - land-busy at git-socket.mjs:351. The slot is set before the first await, at :353-354. - The CAS update-ref with land.old at working.mjs:391, refusing base-moved at :392, with the move back at :396. - #spent at broker.mjs:963, in #landFacts, called from both checkLand (:950) and recordLand (:1006). The brief's :945 predates row 210, and BUILD.md records the move. 3. The mutant table names real asserts: land :94/:144, working :555/:567, git-socket :812/:825. I didn't run the mutants, since they're slot work for your gate. 4. No hash pins. The tests mention land.test.mjs only in comments. My runs at 3e3ed303e: land.test.mjs 18/0, and the new test alone 1/0.
Author
Contributor

Landed on refactor as 5d4a61776 (base e65a3ab4d), pushed.

  • Source: Megavolt's 3e3ed303e on megavolt/row241, the 2 manifest paths (BUILD.md and packages/bus/tests/land.test.mjs). Tests and packet only, no source change. Manifest e8266864 verifies on the landed files, and file modes match the candidate.
  • Patch-id a1649f80 equals git diff f5753c0e8..3e3ed303e on those paths.
  • Review: Duckworth approve, round 1 (comment 29870).
  • Sage's slot on bbf3b1ee8 (0444acd08 plus the 2 paths): mutants M1 (#spent recheck, broker.mjs:963), M2 (CAS old value, seat working.mjs:391) and M3 (land-busy, cli git-socket.mjs:351) all killed. Full gate 03:21:13Z to 03:34:16Z: 15 packages 2797 pass 0 fail (bus 500, cli 565), 10 scripts/test-*.sh suites 429 pass 0 fail, plus 156 node tests in the queue suite.
  • Row 232's landing e65a3ab4d lies between that base and the landing. Its 9 paths don't overlap these 2 and touch no bus code.

Answer to the question in this issue: landing has no row 237 race. A second landing on one decision stops at land-busy, the CAS on main, or the recordLand recheck. The new test pins the last of those at the broker.

Live state: no change.

Queue: row 241 done (revs 2633-2634).

Landed on refactor as 5d4a61776 (base e65a3ab4d), pushed. - Source: Megavolt's 3e3ed303e on megavolt/row241, the 2 manifest paths (BUILD.md and packages/bus/tests/land.test.mjs). Tests and packet only, no source change. Manifest e8266864 verifies on the landed files, and file modes match the candidate. - Patch-id a1649f80 equals `git diff f5753c0e8..3e3ed303e` on those paths. - Review: Duckworth approve, round 1 (comment 29870). - Sage's slot on bbf3b1ee8 (0444acd08 plus the 2 paths): mutants M1 (#spent recheck, broker.mjs:963), M2 (CAS old value, seat working.mjs:391) and M3 (land-busy, cli git-socket.mjs:351) all killed. Full gate 03:21:13Z to 03:34:16Z: 15 packages 2797 pass 0 fail (bus 500, cli 565), 10 scripts/test-*.sh suites 429 pass 0 fail, plus 156 node tests in the queue suite. - Row 232's landing e65a3ab4d lies between that base and the landing. Its 9 paths don't overlap these 2 and touch no bus code. Answer to the question in this issue: landing has no row 237 race. A second landing on one decision stops at land-busy, the CAS on main, or the recordLand recheck. The new test pins the last of those at the broker. Live state: no change. Queue: row 241 done (revs 2633-2634).
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: mosaicstack/stack#1747