Skip to content

v1.92.75.0 fix(ship): the receipt counts every blocker the readers found (GST-951) - #3845

Merged
capy-ai[bot] merged 6 commits into
mainfrom
gst-951-receipt-counts-blockers
Nov 2, 2026
Merged

capy-ai[bot] merged 6 commits into
mainfrom
gst-951-receipt-counts-blockers

Conversation

@garrytan

@garrytan garrytan commented Nov 2, 2026 •

Copy link
Copy Markdown
Owner
{
  "schema_version": 1,
  "pr": 3845,
  "head": "b593334e44099ed5a8e49f746e4472914465616e",
  "version": "1.92.61.0",
  "base": "main",
  "gate": {
    "remote_gate": {
      "pass": 25289,
      "fail": 0
    }
  },
  "session_kind": "unattended",
  "head_tree": "e496909f1f05b05377103804a498e9c3a114a439",
  "product_patch_id": "22dd6a60ca4370e903355bb878ba163589d3a7c6",
  "gstack_version": "1.92.61.0",
  "review_depth": {
    "depth": "full",
    "class": "gate",
    "runtime_lines": 203,
    "rules": "551486253e156572",
    "base": "0c84f4ee14c32734313e68b700417e0748ad1d55",
    "registry": "0c84f4ee14c32734313e68b700417e0748ad1d55",
    "gate_scan": {
      "tagged": 0,
      "head": "b593334e44099ed5a8e49f746e4472914465616e"
    },
    "reasons": [
      "gate-class",
      "risk:money",
      "runtime=203"
    ],
    "risk": [
      "money"
    ]
  },
  "pregate": "requires-remote cleared @e496909f1f05 (1 checks, current for b593334e4409)",
  "requires_remote": [
    "free-tests.yml/cso-macos-launcher: https://github.com/garrytan/gstack/actions/runs/41412339960",
    "free-tests.yml/macos-named-regressions: https://github.com/garrytan/gstack/actions/runs/41412339960",
    "windows-smoke.yml/windows-smoke: https://github.com/garrytan/gstack/actions/runs/41412340028",
    "pregate.json/remote_gate: remote-gate tree=e496909f1f05 tests=25289"
  ],
  "issues": {
    "fixes": [],
    "refs": [],
    "flakes": []
  },
  "docs": {
    "status": "updated",
    "scope": "docs/ and ship/sections/ (GST-951 review-scope ledger docs, git diff origin/main...HEAD)"
  },
  "review_loop": {
    "passes": 2,
    "blocking_found": 12,
    "deferred": 0,
    "noted": 7,
    "fixed": 15,
    "scope": "frozen",
    "fix_checks": 1,
    "rounds": 3,
    "unrecorded": 4,
    "ledgers_read": 4,
    "dispatches": {
      "specialist": 2,
      "red-team": 2,
      "adversarial-native": 2,
      "outside-voice": 2,
      "targeted": 2,
      "fix-check": 1
    }
  },
  "review_loop_exception": {
    "accepted_by": "GStack Minder (Captain, GSTA-36)",
    "date": "2026-11-02",
    "via": "the GSTA-167 lane parent relaying the Captain's rule 23 grant",
    "reason": "one scoped fix-check after round 2 on the five counting gaps; its findings ship as open notes",
    "fix_check": {
      "head": "15823cc46656deb6c2e1b17c355321feffdb801b",
      "findings": 3
    }
  },
  "review_targeted": {
    "classes": [
      "money"
    ],
    "rules": "c82addff6fb60fa0",
    "brief_sha": "7f92a0d0a8319c3c0e076fb3b1a4689c45120b2be2c14dfc7d4cb03aafa9d6a5",
    "product_patch_id": "22dd6a60ca4370e903355bb878ba163589d3a7c6",
    "status": "done",
    "findings": 0,
    "raised_by": "targeted"
  }
}

What changed

The receipt's blocker count now includes what the readers found in the dispatch ledgers: 9 becomes 12, and 4 blockers with no recorded disposition are named

The receipt's blocker count now includes what the readers found in the dispatch ledgers: 9 becomes 12 on this PR's own review, and 4 blockers with no recorded disposition are named.

Text version: on origin/main dd9834d the summary counts nine blockers for this PR's own review. On this branch it counts twelve, names four with no recorded disposition, and says it read four ledgers.

Why

The receipts for #3841, #3838 and #3837 said no blockers were found, even though their readers raised real CRITICALs that got fixed. The summary only counted review-log rows stamped blocking, while /ship's readers report through the dispatch ledgers. Now the receipt counts every blocker any reader found, and it refuses to write a count when it can't read the ledgers it needs.

Blast radius

This touches every gstack-review-scope summary caller: ship/DRIVER.md now passes --run, the Claude Code pr-body section passes --no-ledger, and the four existing test call sites got --no-ledger; a caller that passes none of the three now gets REVIEW_SCOPE_LEDGER instead of a quiet undercount. The remote pregate ran the whole selected free tier (25289 tests, 0 failed on tree e496909f), and the receipt reader accepts receipts without the two new keys.

Approvals

Date Asked Exact words Outcome Covers Still pending
2026-11-02 One scoped fix-check after round 2 on 15823cc (five counting gaps) “the Captain chose A and granted it in writing under rule 23. Run one scoped fix-check on 15823cc ... covering only the five counting gaps ... Any finding after that ships as an open note in the PR body and READY, with no further rounds.” (GStack Minder (via the GSTA-167 lane parent)) approved GST-951 review loop none

Source: gstack-approvals record for gst-951-receipt-counts-blockers (display only; not an approval citation).

Summary

summary also reads every ledger's collected and judged CRITICAL findings (ledgerBlockers in lib/review-ledger.ts). Each one pairs with at most one row finding from the same reviewer at the same path:line, so a blocker held in both places counts once and one a row classified below the floor does not; an unclassified CRITICAL row counts as blocking. It prints unrecorded (blockers no row gives a disposition) and ledgers_read, and the schema, types and parseReviewLoop carry both. It needs --run <dir>, --ledger <f> or --no-ledger; a missing choice, a run dir with no ledger, an unreadable ledger, merged file or judged set, or a judged set changed since judge is the new REVIEW_SCOPE_LEDGER result code with nothing written (troubleshooting and doctor anchors added). DRIVER.md stays at 12262 bytes.

Test Coverage

test/review-scope-ledger-count.test.ts derives every count from ledgers and logs (the PR cases are reconstructions, labeled as such). Regression proof: on origin/main dd9834d it fails 10 of 10; at this head it passes 10 of 10.

Case Found before → after Fixed after Unrecorded after
#3841 as /ship logs it 0 → 1 1 0
#3841 as the lane logged it 0 → 1 0 1 (no log holds the fix)
#3838 0 → 2 4 1
#3837 0 → 7 0 7

Tests ran: remote_gate 25289

Pre-Landing Review

Two review rounds on the full panel (gate-class), then the Captain's granted fix-check on 15823cc; the targeted money reader found nothing on the final tree.

REVIEW_LOOP: passes=2 blocking_found=12 deferred=0 noted=7 fixed=15 scope=frozen unrecorded=4 ledgers_read=4 fix_checks=1 dispatches=11

Review-loop exception: accepted by GStack Minder (Captain, GSTA-36) 2026-11-02 via the GSTA-167 lane parent relaying the Captain's rule 23 grant: one scoped fix-check after round 2 on the five counting gaps; its findings ship as open notes (fix-check at 15823cc, 3 findings)

Review notes

Open notes from the granted fix-check (no further rounds):

  1. lib/review-scope.ts:363: rows raised by core, unknown or no raised_by pair with any reviewer's ledger blocker, so a below-floor core row can still absorb another reviewer's CRITICAL at that line.
  2. lib/review-scope.ts:362: raiserOf skips raisedBy's normalization (lowercasing, the adversarial alias, the specialist field), so a row raised by adversarial and its adversarial-native ledger finding count twice and show one false unrecorded.
  3. lib/review-scope.ts:420: a fix-check CRITICAL at a line an earlier row logged fixed pairs with that stale row, so a failed fix reads as fixed instead of unrecorded.
  4. Known residual: two distinct blockers from one reviewer at the same line with one row pair once.

No network calls, secrets, permissions, workflows, install scripts or dependencies change.

Documentation

Status: updated. Scope: docs/ and ship/sections/ (git diff origin/main...HEAD). docs/review-policy.md's code map lists unrecorded, ledgers_read and REVIEW_SCOPE_LEDGER; docs/ship-policy.md's refusal list links REVIEW_SCOPE_LEDGER.

Test plan

  • Remote pregate (UBI_OWNER=gsta167): 25289 tests, 0 failed on tree e496909f, $0.08
  • bun test test/review-scope-ledger-count.test.ts: all ten green
  • CI at b593334: all checks green

garrytan and others added 5 commits November 2, 2026 07:18
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
…and tests from review pass 1 (GST-951)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
…sification wins; ledgers_read (GST-951 Step 11 findings)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
…less loci; pr-body names no targeted ledger (GST-951 round-2 findings, pending grant)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
@trunk-io

trunk-io Bot commented Nov 2, 2026 •

Copy link
Copy Markdown

😎 This pull request was merged.

@github-actions github-actions Bot changed the title fix(ship): the receipt counts every blocker the readers found (GST-951) v1.92.75.0 fix(ship): the receipt counts every blocker the readers found (GST-951) Nov 2, 2026
@capy-ai
capy-ai Bot merged commit 82e9d7e into main Nov 2, 2026
56 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant