v1.94.11.0 fix(ship): a granted fix-check's range counts at finish; fix-check rows carry completed/converged (GST-1130) (#3916)

* fix(ship-release): a completed fix-check's range counts as reviewed at finish; fix-check rows carry completed/converged (GST-1130)

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

* fix(ship-release): fix-check range base must hold the reviewed tree; docs keep the recheck (review round 1)

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

* fix(ship-release): count a fix-check row only under a recorded review_loop_grant (GST-1130 granted round)

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

* docs(changelog): GST-1130 fragment wording

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>

CHANGELOG.md

@@ -1,5 +1,17 @@
 # Changelog
 
+## [1.94.11.0] - 2026-11-03
+
+**A granted fix-check's clean read now counts at Step 16.** On gstack-evals#168 the owner granted one scoped fix-check, the reader read `9b0c56c..7ea1907` and found nothing, and `gstack-ship-release finish` still stopped with `STOP: behavior, tests or build inputs changed since review (tests/panel-kit-v3-1.test.ts)`, sending the test file the fix-check had read back through Steps 5 to 11.5. The review-loop exception could not help, because it is the owner accepting unreviewed edits and it never accepts a gate-edited test. The fix-check row also read `UNVERIFIED`, because it was logged without its `completed` and `converged` fields (GST-1130, GEVA-218).
+
+### Fixed
+- **`gstack-ship-release finish`** reads the newest `review-fix-check` row. When the row is a completed clean read on committed content, its `range_base` is a branch commit holding the reviewed tree, and the owner's recorded `review_loop_grant` names a stop head inside that range, every code or test path inside that range that is unchanged since routes as `fix_checked=` on the `ROUTE:` line, with a `FIX_CHECK: reviewed <paths>` line. That includes a gate-edited test. Authored docs still get the plan-item recheck, and a path edited after the fix-check still stops. A row that does not qualify prints `FIX_CHECK: not counted (<reason>)` and changes nothing.
+- **`gstack-review-log`** refuses a `review-fix-check` row that does not carry `completed` and `converged` as true or false, and keeps the start token for the retry. The refusal names the missing fields, so the row is fixed when it is written instead of reading `UNVERIFIED` at Step 11.5.
+
+### Tests
+- `test/ship-rereview-triggers.test.ts`: a fixture shaped like gstack-evals#168 (a round-2 pass, then a fix commit that edits the source and a gate-edited test) stops on main and passes finish once a clean fix-check row with its range base is logged through the real `gstack-review-log`. Negative controls: an edit after the fix-check still stops, and an ungranted row, a grant for another stop head, an incomplete row, a row with no range base, a range base that does not hold the reviewed tree, or a row logged on uncommitted content is not counted.
+- `test/review-log-findings.test.ts`: a `review-fix-check` row missing either field, or with a non-boolean value, is refused and keeps its start token; the complete row binds as verified.
+
 ## [1.94.10.0] - 2026-11-03
 
 **Windows checks now run by path.** Windows Setup E2E, the 10-to-25-minute install test, no longer starts its Windows jobs on a pull request whose only install change is `package.json`'s version line, which the merge-train stamp puts on every pull request before it merges. Of the 61 pull requests in its last 100 runs (as of 2026-11-03), 37 changed no install file besides `package.json`. Its path list also gains `bun.lock` and the `bin/` scripts that the test or `./setup` runs. A new `Windows Scoped Tests` workflow runs the Windows-safe test files a pull request changed, on one Windows job, when the pull request touches path, file-name or process code, such as a hook script or a new `spawnSync` call. On the last 120 commits to main (through 923fc0a) it would have run on 50, with a median of one test file. It catches both Windows breaks the plan names: /freeze's path check (#3758) and a test that started Bun with `--config=/dev/null` (#3829). A red run blocks its pull request like any failed check, so this release also fixes #3829's test, which has failed every nightly Windows run since it landed. `windows-smoke` still runs on every pull request, and the full Windows suite still runs nightly.

VERSION

@@ -1 +1 @@
-1.94.10.0
+1.94.11.0

agents-digest/gstack-AGENTS.md

@@ -1,4 +1,4 @@
-# gstack digest v1.94.10.0 — regenerate/re-copy after upgrading gstack
+# gstack digest v1.94.11.0 — regenerate/re-copy after upgrading gstack
 
 Behavioral rules from gstack (https://github.com/garrytan/gstack), compressed
 for agent hosts without a full skill install. The full skills add workflows,

docs/review-policy.md

@@ -493,6 +493,21 @@ through the docs stage. A path outside that scope and a gate-edited test still
 route as behavior, with `--reviewed-base` (as /ship passes it) so does a file
 too large for the scoped review, and without the record nothing changes.
 
+A completed fix-check is a review, not an owner acceptance (GST-1130). Finish
+reads the newest `review-fix-check` row from `gstack-review-read`. When it is a
+completed clean read on committed content (freshness `CURRENT` against its own
+tree) whose commit is on the branch, whose `range_base` is a branch commit
+holding the reviewed tree, and for which an approved `review_loop_grant` in
+`gstack-approvals` names a stop head inside `range_base..commit`, each code or test path inside `range_base..commit`
+that is unchanged since routes as `fix_checked=` in `ROUTE:` and is listed on a
+`FIX_CHECK: reviewed` line, a gate-edited test included. Authored docs keep the
+stage 2.2 recheck, and a path edited after the fix-check still routes by the
+rules above. Any other row prints
+`FIX_CHECK: not counted (<reason>)` and changes nothing. `gstack-review-log`
+refuses a `review-fix-check` row without boolean `completed` and `converged`,
+before it consumes the start token, because without them the row reads
+`UNVERIFIED` and finish cannot count it.
+
 The receipt's summary line prints `passes=<n> blocking_found=<k> deferred=<m>`,
 which is what a scorecard tracks (target: a median of 1 pass per PR). A
 receipt carried across a base merge or a test-only fix

lib/review-evidence.ts

@@ -272,6 +272,9 @@ export function bindReview(rec: Record<string, any>, token: string, env = proces
   if (env.GSTACK_STAMP_TREE) rec.tree = env.GSTACK_STAMP_TREE;
   if (env.GSTACK_STAMP_DIRTY) rec.dirty = env.GSTACK_STAMP_DIRTY === 'true';
   stampGateFindings(rec.findings, env.GSTACK_STAMP_SESSION_KIND);
+  if (rec.skill === 'review-fix-check' && (typeof rec.completed !== 'boolean' || typeof rec.converged !== 'boolean')) {
+    throw new Error('review-log: a review-fix-check row must carry completed and converged as true or false (review-army item 6 fields; without them gstack-review-read marks the row UNVERIFIED and finish cannot count it)');
+  }
   // A /ship pass whose findings were all fixed, or deferred or noted below the floor, left no open defect (review-army
   // item 6), so a claimed issues_found is recorded clean with zero counts and Step 11.5 can bind it (GST-430). Only the
   // logger-stamped action counts (never a caller's status), and the findings must account for every claimed issue.

lib/ship-release.ts

@@ -15,6 +15,8 @@ import { mergeChangelogs, normalizeChangelog } from './changelog-check';
 import { versionAtRef } from './restamp';
 import { classifyPath } from './review-depth';
 import { comparePath, contextPacket, gateEditedPaths, isTestCode, PACKET_SECTION_CAP, readScopedReview, touchedPaths, type PatchSides } from './ship-patch';
+import { locate, readRecord } from './approval-record';
+import { reviewFreshness } from './review-evidence';
 import { parseReviewLoopException, type ReviewLoopException } from './ship-receipt-depth';
 import { parseReceiptDocs, type ReceiptDocs } from './ship-receipt-docs';
 import { resolveStateRoot } from './state-root';
@@ -183,8 +185,14 @@ function harnessRepin(cwd: string, from: string, to: string, p: string): boolean
     || (e?.since === b.version && typeof e?.nonBehavioral === 'string' && e.nonBehavioral.trim().length >= 10));
 }
 
-/** Step 16 stage 2's route. With --reviewed-base, a path whose branch patch is the reviewed one counts as base (lib/ship-patch.ts). */
-type Routed = Record<'behavior' | 'docs' | 'metadata' | 'tests' | 'base' | 'context' | 'accepted', string[]>;
+/**
+ * Step 16 stage 2's route. With --reviewed-base, a path whose branch patch is the reviewed one counts as base (lib/ship-patch.ts).
+ * A code or test path the newest completed clean review-fix-check row read (inside its range_base..commit range, where
+ * range_base holds the reviewed tree, and unchanged since) counts as fix_checked when an approved review_loop_grant names a stop
+ * head inside range_base..commit: the owner's granted fix-check after Step 11 is a review (GST-1130). Authored docs keep the
+ * stage 2.2 recheck.
+ */
+type Routed = Record<'behavior' | 'docs' | 'metadata' | 'tests' | 'base' | 'context' | 'accepted' | 'fix_checked', string[]>;
 function route(cwd: string, base: string, reviewed: string, reviewedBase: string | undefined, release: string[], opts: { docsRechecked: boolean; contextReviewed?: string; loop?: ReviewLoopException }, r: Report): Routed | null {
   const now = bin(cwd, 'gstack-wtree', [], 60_000).stdout.trim();
   if (!/^[0-9a-f]{40,64}$/.test(now)) { r.stop('gstack-wtree printed no tree; the stage 2 comparison is blocked'); return null; }
@@ -204,7 +212,7 @@ function route(cwd: string, base: string, reviewed: string, reviewedBase: string
     sides = { reviewedBase: rb, reviewed, currentBase: cb, current: now };
     touched = new Set([...a, ...b]);
   }
-  const by: Routed = { behavior: [], docs: [], metadata: [], tests: [], base: [], context: [], accepted: [] };
+  const by: Routed = { behavior: [], docs: [], metadata: [], tests: [], base: [], context: [], accepted: [], fix_checked: [] };
   for (const p of [...changedSet].sort()) {
     const verdict = sides ? (touched.has(p) ? comparePath(cwd, sides, p) : 'same') : 'changed';
     by[verdict === 'same' ? 'base' : release.includes(p) || harnessRepin(cwd, reviewed, now, p) ? 'metadata' : sides && isTestCode(p) ? 'tests' : classifyPath(p, cwd) === 'docs' ? 'docs' : verdict === 'context' ? 'context' : 'behavior'].push(p);
@@ -219,6 +227,30 @@ function route(cwd: string, base: string, reviewed: string, reviewedBase: string
     by.context = by.context.filter(p => !packetBody.over.includes(p));
     by.tests = by.tests.filter(p => !packetBody.over.includes(p));
   }
+  const fixRow = bin(cwd, 'gstack-review-read', [], 60_000).stdout.split('\n---CONFIG---')[0].split('\n')
+    .flatMap(l => { try { const v = JSON.parse(l); return v?.skill === 'review-fix-check' ? [v] : []; } catch { return []; } }).at(-1);
+  const commitOf = (c: unknown) => typeof c === 'string' && /^[0-9a-f]{7,64}$/.test(c) ? git(cwd, ['rev-parse', '--verify', '--quiet', `${c}^{commit}`]).stdout.trim() : '';
+  const onBranch = (c: string) => !!c && git(cwd, ['merge-base', '--is-ancestor', c, 'HEAD']).status === 0;
+  const fixCommit = commitOf(fixRow?.commit_full), fixFrom = commitOf(fixRow?.range_base);
+  const fresh = fixRow && reviewFreshness(fixRow, fixRow.wtree);
+  const ancestor = (a: string, d: string) => git(cwd, ['merge-base', '--is-ancestor', a, d]).status === 0;
+  const grants = fixRow ? (loc => 'file' in loc ? readRecord(loc).rows : [])(locate({ cwd, env: process.env })) : [];
+  const grant = grants.find(g => g.outcome === 'approved' && `${g.ask} ${g.scope}`.includes('review_loop_grant')
+    && (`${g.ask} ${g.scope}`.match(/\b[0-9a-f]{7,40}\b/g) ?? []).some(s => { const c = commitOf(s); return !!c && !!fixFrom && !!fixCommit && ancestor(fixFrom, c) && ancestor(c, fixCommit); }));
+  const notCounted = !fixRow ? '' : fresh?.status !== 'CURRENT' ? fresh?.reason ?? 'not a review row'
+    : fixRow.tree !== fixRow.wtree || !onBranch(fixCommit) ? 'the row was logged on content no branch commit holds'
+    : !onBranch(fixFrom) ? 'the row names no range_base commit on this branch'
+    : git(cwd, ['rev-parse', `${fixFrom}^{tree}`]).stdout.trim() !== reviewed ? `range_base ${fixFrom.slice(0, 12)} does not hold the reviewed tree`
+    : !grant ? 'no approved review_loop_grant names a stop head inside its range_base..commit' : '';
+  if (notCounted) r.add(`FIX_CHECK: not counted (${notCounted})`);
+  const counted = !!fixRow && !notCounted;
+  if (counted) {
+    const range = touchedPaths(cwd, fixFrom, fixCommit), after = touchedPaths(cwd, fixCommit, now);
+    if (!range || !after) { r.stop(`git could not diff the fix-check range ${fixFrom.slice(0, 12)}..${fixCommit.slice(0, 12)} against the reviewed and current trees: stage 2 is blocked`); return null; }
+    by.fix_checked = [...by.behavior, ...by.tests, ...by.context].filter(p => range.has(p) && !after.has(p)).sort();
+    for (const k of ['behavior', 'tests', 'context'] as const) by[k] = by[k].filter(p => !by.fix_checked.includes(p));
+    r.add(`FIX_CHECK: reviewed ${by.fix_checked.length ? short(by.fix_checked) : 'none'} (fix-check at ${fixCommit.slice(0, 12)}, range from ${fixFrom.slice(0, 12)}, grant ${grant!.id})`);
+  }
   const loop = opts.loop;
   if (loop) {
     const head = loop.fix_check.head;
@@ -233,7 +265,7 @@ function route(cwd: string, base: string, reviewed: string, reviewedBase: string
     by.behavior = by.behavior.filter(p => !by.accepted.includes(p));
     by.docs = by.docs.filter(p => !by.accepted.includes(p));
   }
-  r.add(`ROUTE: reviewed=${reviewed.slice(0, 12)} current=${now.slice(0, 12)} behavior=${by.behavior.length} docs=${by.docs.length} metadata=${by.metadata.length} tests=${by.tests.length} base=${sides ? by.base.length : 'unknown'} context=${by.context.length}${loop ? ` accepted=${by.accepted.length}` : ''}`);
+  r.add(`ROUTE: reviewed=${reviewed.slice(0, 12)} current=${now.slice(0, 12)} behavior=${by.behavior.length} docs=${by.docs.length} metadata=${by.metadata.length} tests=${by.tests.length} base=${sides ? by.base.length : 'unknown'} context=${by.context.length}${counted ? ` fix_checked=${by.fix_checked.length}` : ''}${loop ? ` accepted=${by.accepted.length}` : ''}`);
   if (loop) r.add(`REVIEW_LOOP_EXCEPTION: accepted ${by.accepted.length ? short(by.accepted) : 'none'} (changed since fix-check ${loop.fix_check.head.slice(0, 12)}; accepted by ${loop.accepted_by} ${loop.date} via ${loop.via})`);
   if (by.behavior.length) { r.stop(`behavior, tests or build inputs changed since review (${short(by.behavior.sort())}): insert 5–11.5 → 12–14 → 16 before Step 17 (stage 2.1)`); return null; }
   const scoped = [...by.context, ...by.tests].sort();

package.json

@@ -1,6 +1,6 @@
 {
   "name": "gstack",
-  "version": "1.94.10",
+  "version": "1.94.11",
   "description": "Garry's Stack — Claude Code skills + fast headless browser. One repo, one install, entire AI engineering workflow.",
   "license": "MIT",
   "type": "module",

test/review-log-findings.test.ts

@@ -191,6 +191,23 @@ describe('lib: deriveFromFindings, bindReview order and reviewFreshness', () =>
     expect(reviewFreshness(bound, 'd'.repeat(64))?.status).toBe('STALE');
   });
 
+  // Value: protects=a fix-check row is written with the fields review-read verifies (GEVA-218: the row lacked them, so it read
+  //   UNVERIFIED and finish could not count it); fails_when=the logger accepts a review-fix-check row without completed and
+  //   converged, or consumes the start token while refusing it; seam=bindReview
+  test('a review-fix-check row without completed and converged is refused and keeps its start token (GST-1130)', () => {
+    const dir = path.join(home, 'starts', '.review-starts');
+    fs.mkdirSync(dir, { recursive: true });
+    const token = '77777777-2222-3333-4444-555555555555';
+    const w = 'e'.repeat(64);
+    fs.writeFileSync(path.join(dir, `${token}.json`), JSON.stringify({ skill: 'review-fix-check', repo: '/r', branch: 'b', wtree: w, started_at: 't' }));
+    const env = { GSTACK_REVIEW_DIR: path.join(home, 'starts'), GSTACK_REVIEW_REPO: '/r', GSTACK_REVIEW_BRANCH: 'b', GSTACK_STAMP_WTREE: w };
+    for (const row of [{}, { completed: true }, { converged: true }, { completed: 'yes', converged: true }]) {
+      expect(() => bindReview({ skill: 'review-fix-check', status: 'clean', ...row }, token, env)).toThrow('review-log: a review-fix-check row must carry completed and converged as true or false');
+    }
+    expect(fs.existsSync(path.join(dir, `${token}.json`))).toBe(true);
+    expect(bindReview({ skill: 'review-fix-check', status: 'clean', completed: true, converged: true }, token, env).review_binding.state).toBe('verified');
+  });
+
   test('reviewFreshness reads findings_open: resolved criticals are CURRENT, claimed issues_found stays conservative', () => {
     const w = 'b'.repeat(64);
     const base = { skill: 'review', completed: true, converged: true, wtree: w, review_binding: { state: 'verified', start_wtree: w, end_wtree: w } };

test/ship-rereview-triggers.test.ts

@@ -757,3 +757,83 @@ describe('finish with an owner-accepted edit after a fix-check (GST-965, GBRA-27
   });
 });
 
+
+// Value: protects=a granted fix-check's clean read clears the paths it reviewed at Step 16, a gate-edited test included, so
+//   finish does not send them back through Steps 5–11.5 (GEVA-218, gstack-evals#168); fails_when=finish routes a path inside
+//   the fix-check's range as behavior, counts a path edited after the fix-check, or counts a row that is not a completed clean
+//   fix-check with a range base; seam=finish's route reading the review log; no_claim=Step 11.5's binding
+describe('finish after a completed fix-check (GST-1130, GEVA-218)', () => {
+  const TEST_V1 = "import { expect, test } from 'bun:test';\ntest('a', () => { expect(2).toBe(2); });\n";
+  const TEST_V2 = "import { expect, test } from 'bun:test';\ntest('a', () => { expect(2).toBe(2); }, 1_000);\ntest('b finishes fast', () => { expect(1 + 1).toBe(2); }, 1_000);\n";
+  const logFixCheck = (f: ShipFx, row: Record<string, unknown>) => {
+    const start = f.run(path.join(BIN, 'gstack-review-log'), ['--start', 'review-fix-check']);
+    expect(start.status).toBe(0);
+    return f.run(path.join(BIN, 'gstack-review-log'), [JSON.stringify({ skill: 'review-fix-check', timestamp: '2026-11-02T00:00:00Z', via: 'ship', review_scope: 'frozen', status: 'clean', issues_found: 0, critical: 0, ...row }), '--finish', start.stdout.trim()]);
+  };
+  const grant = (f: ShipFx, stopHead: string) => {
+    const a = f.run(path.join(BIN, 'gstack-approvals'), ['add', '--ask', `Owner grant: one scoped fix-check after round 2 (review_loop_grant ${stopHead.slice(0, 12)})`, '--words', 'one scoped fix-check after round 2', '--by', 'GStack Minder', '--scope', 'this /ship invocation', '--json'], { GSTACK_APPROVALS_MIRROR: '0' });
+    expect(a.status).toBe(0);
+    return JSON.parse(a.stdout).added[0] as string;
+  };
+  const afterFixCheck = () => {
+    const f = fx({ 'src/x.ts': 'export const x = 1;\n' }, { 'src/x.ts': 'export const x = 2;\n', 'test/x.test.ts': TEST_V1 });
+    const passHead = f.git('rev-parse', 'HEAD').stdout.trim();
+    const reviewed = wtree(f);
+    f.write({ 'src/x.ts': 'export const x = 3;\n', 'test/x.test.ts': TEST_V2 }); f.commit('fix the round-2 blocker and bound its case');
+    return { f, passHead, fixHead: f.git('rev-parse', 'HEAD').stdout.trim(), reviewed, post: snapshot(f) };
+  };
+
+  test('paths inside a completed clean fix-check range count as reviewed, a gate-edited test included; a later edit still stops', () => {
+    const { f, passHead, fixHead, reviewed, post } = afterFixCheck();
+    const plain = finish(f, reviewed, post);
+    expect([plain.status, plain.stdout.includes('STOP: behavior, tests or build inputs changed since review (src/x.ts,test/x.test.ts)')]).toEqual([3, true]);
+    const id = grant(f, passHead);
+    expect(logFixCheck(f, { completed: true, converged: true, range_base: passHead }).status).toBe(0);
+    const r = finish(f, reviewed, post);
+    expect(lineOf(r.stdout, 'FIX_CHECK')).toBe(`FIX_CHECK: reviewed src/x.ts,test/x.test.ts (fix-check at ${fixHead.slice(0, 12)}, range from ${passHead.slice(0, 12)}, grant ${id})`);
+    expect(lineOf(r.stdout, 'ROUTE')).toMatch(/ behavior=0 docs=0 metadata=0 tests=0 base=unknown context=0 fix_checked=2$/);
+    expect([r.status, lineOf(r.stdout, 'EVIDENCE')]).toEqual([0, expect.stringMatching(/^EVIDENCE: FRESH label=tests /)]);
+    f.write({ 'src/x.ts': 'export const x = 4;\n' }); f.commit('an edit after the fix-check');
+    const late = finish(f, reviewed, post);
+    expect(lineOf(late.stdout, 'FIX_CHECK')).toStartWith('FIX_CHECK: reviewed test/x.test.ts (');
+    expect([late.status, late.stdout.includes('STOP: behavior, tests or build inputs changed since review (src/x.ts)')]).toEqual([3, true]);
+  });
+
+  test('a clean fix-check row with no recorded review_loop_grant for a stop head in its range is not counted', () => {
+    const { f, passHead, reviewed, post } = afterFixCheck();
+    expect(logFixCheck(f, { completed: true, converged: true, range_base: passHead }).status).toBe(0);
+    const ungranted = finish(f, reviewed, post);
+    expect(lineOf(ungranted.stdout, 'FIX_CHECK')).toBe('FIX_CHECK: not counted (no approved review_loop_grant names a stop head inside its range_base..commit)');
+    expect([ungranted.status, ungranted.stdout.includes('STOP: behavior, tests or build inputs changed since review (src/x.ts,test/x.test.ts)')]).toEqual([3, true]);
+    grant(f, mergeBase(f));
+    const elsewhere = finish(f, reviewed, post);
+    expect([elsewhere.status, lineOf(elsewhere.stdout, 'FIX_CHECK')]).toEqual([3, 'FIX_CHECK: not counted (no approved review_loop_grant names a stop head inside its range_base..commit)']);
+  });
+
+  test('a fix-check row that did not finish clean, or names no range base, is not counted', () => {
+    const { f, passHead, reviewed, post } = afterFixCheck();
+    expect(logFixCheck(f, { completed: false, converged: false, range_base: passHead }).status).toBe(0);
+    const incomplete = finish(f, reviewed, post);
+    expect(lineOf(incomplete.stdout, 'FIX_CHECK')).toBe('FIX_CHECK: not counted (missing start capture or incomplete/nonconverged pass)');
+    expect([incomplete.status, incomplete.stdout.includes('STOP: behavior, tests or build inputs changed since review (src/x.ts,test/x.test.ts)')]).toEqual([3, true]);
+    expect(logFixCheck(f, { completed: true, converged: true }).status).toBe(0);
+    const unranged = finish(f, reviewed, post);
+    expect(lineOf(unranged.stdout, 'FIX_CHECK')).toBe('FIX_CHECK: not counted (the row names no range_base commit on this branch)');
+    expect(unranged.status).toBe(3);
+    const fork = mergeBase(f);
+    expect(logFixCheck(f, { completed: true, converged: true, range_base: fork }).status).toBe(0);
+    const early = finish(f, reviewed, post);
+    expect(lineOf(early.stdout, 'FIX_CHECK')).toBe(`FIX_CHECK: not counted (range_base ${fork.slice(0, 12)} does not hold the reviewed tree)`);
+    expect(early.status).toBe(3);
+  });
+
+  test('a fix-check logged on uncommitted content is not counted', () => {
+    const { f, passHead, reviewed, post } = afterFixCheck();
+    f.write({ 'src/x.ts': 'export const x = 9;\n' });
+    expect(logFixCheck(f, { completed: true, converged: true, range_base: passHead }).status).toBe(0);
+    f.git('checkout', '--', 'src/x.ts');
+    const r = finish(f, reviewed, post);
+    expect(lineOf(r.stdout, 'FIX_CHECK')).toBe('FIX_CHECK: not counted (the row was logged on content no branch commit holds)');
+    expect(r.status).toBe(3);
+  });
+});