fix(pm): dispatch-gates' reconciliation line enumerates every OUTSIDE block, not three of five

[hotlong]

Fixes #16398

The reconciliation line closes with the sentence that tells a harvester what its
total does not cover. It named three blocks while the same run printed five.

What changed

familyReconciliationLines now builds that enumeration once, as
outsideBlockNames, from two block counts carried on recon — the lengths of
the very arrays artifactRosterLines and widePopulationLines are rendered
from, so the list cannot name a set the output does not contain. A count of 0
drops the name, because at zero rows those two render nothing and pointing a
reader "below" at an absent heading is the same defect facing the other way.

Three consequences, all of them the point:

• Both branches of the rendering use that one expression. The zero-total
  branch used to spell the claim separately and named a single block; it now
  renders the same list.
• The docblock no longer restates the list. It named the same three blocks
  in prose — one claim written out twice, and both copies were wrong. It now
  names outsideBlockNames as the only place the list exists.
• The names are in print order, so a reader walking down the output meets
  the blocks in the order this line promised them.

⛔ The five blocks themselves, the conservation line, every verdict and every
exit code are untouched. familyReconciliation gains two counts that are not
terms of the total and never enter its closure assertion — they sit beside
ciOnly and notRunnable, which are outside-the-total counts already.

Readings — node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack scripts/pm/dispatch-gates.mjs

BEFORE (on origin/main 941232040):

  ⛔ 30 is what THIS CARD owes by path and kind — NOT a complete account of what CI runs on the PR.
  The pending-changeset families, the unreachable listing and the always-runs tail below are each
  OUTSIDE it, each with its own count.

AFTER (this branch):

  ⛔ 30 is what THIS CARD owes by path and kind — NOT a complete account of what CI runs on the PR.
  The 44 artifact-roster famil(ies), the 10 declared WIDE-population famil(ies), the
  pending-changeset families, the unreachable listing and the always-runs tail below are each
  OUTSIDE it, each with its own count.

The two new counts are the two headings the same run prints below that line:

Artifact rosters — 44 famil(ies) whose `silent` verdict is a fact about a LIST, not about your paths:
Declared WIDE population — 10 famil(ies) DECLARE that CI runs them over a population too wide to place, …

Tests

The self-test case that pinned this sentence asked only for the substring
always-runs tail, so it stayed green through the entire defect. It now asserts
each name separately, pins the whole phrase in print order, asserts that neither
block is named on a run that printed neither, and asserts the zero-total branch
renders the same list.

Ablation — the mutation is the card's own defect: delete the artifact-roster
entry from outsideBlockNames, so the enumeration again omits a block the same
run printed.

leg	blob of scripts/pm/dispatch-gates.mjs	pnpm check:pm-dispatch-gates
mutated (artifact-roster entry deleted)	e933f57c47ffb2d18d42575230856b627795e9c4	exit 1 — ✗ dispatch-gates self-test: 3 of 1561 case(s) failed.
restored (HEAD)	b2232d75d7cace982fedffd1686677a64b196cb3	exit 0 — ✓ dispatch-gates self-test: 1561 cases pass.

Restore was git checkout HEAD -- scripts/pm/dispatch-gates.mjs, proven by the
blob hash matching HEAD's and by an empty git status --porcelain, never by an
exit code.

Gates: the 30 families node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derives for this diff were run, all exit 0 — the
derived list is byte-identical to the one the dispatching seat derived, so no
family was added by the re-derivation. pnpm exec eslint scripts/pm/dispatch-gates.mjs exits 0 over 1 file, 0 errors, 0 warnings;
type-aware linting is not enabled in eslint.config.mjs (no
parserOptions.project), so this diff cannot move the verdict on a file it does
not touch. The repo-wide pnpm lint is CI's run, not claimed here.

node scripts/pm/check-governed-merges.mjs --test scripts/pm/dispatch-gates.mjs
→ ✅ NOT governed — ordinary queue landing applies to a PR with exactly this file list.

验收备注

• The same claim is spelled twice more, on the other two output lanes, and both
  are still short. --ran (runReconciliationLines) names three of the five
  blocks; --commands / --json (machineReadableOutput) never discloses the
  declared WIDE population at all — not as a block, not as a count, not in its
  closing disclaimer, which names one block only. Filed as [finding] dispatch-gates' other two output lanes disclaim an incomplete OUTSIDE list too — --ran names three of five blocks, and --commands never discloses the WIDE population at all #16795 with a repro
  rather than folded in: each lane needs the counts threaded through a different
  structure, which is a different verification surface from the one this card's
  claim declared. Both can read outsideBlockNames once this lands.
• The triage ruling said the sentence was written twice in this file. It is
  written four times in total; the two loci the ruling named are the ones this
  PR owns, and the other two are [finding] dispatch-gates' other two output lanes disclaim an incomplete OUTSIDE list too — --ran names three of five blocks, and --commands never discloses the WIDE population at all #16795.
• Declared rider reading, measurement only, no implementation: dispatch-gates: STAGE-THEN-RUN reaches a program by an edge neither follow traverses — check:objectui-changeset inherits nothing from scripts/bump-objectui.sh #14290's re-check
  on this branch, node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --residue scripts/objectui-changeset-digest.mjs,
  exits 0 and still names its family —
  pnpm check:objectui-changeset [lint.yml] matched via scripts/objectui-changeset-digest.mjs ⇢ gate script scripts/objectui-changeset-digest.mjs. [finding] dispatch-gates residue: after #12514 the extensionless-target reason is only reachable for a BARE hint, where it names the wrong cause — latent wrong-reason row, zero instances today #12797 and [finding] check-cross-package-test-inputs' docblock states the single-segment hint refusal without its one exception #12808 are held riders on
  this surface and are not folded: a different defect shape, each another
  lane's card needing its own verification.
• skip-changeset: nothing published moves. scripts/pm/** is on the
  fast-track list — it is not in any package's files[] and ships to no
  consumer.
• Merged cleanly against origin/main 6d4d5d3dd, probed from a throwaway bare
  clone sharing the object store with no merge driver registered, so the reading
  is GitHub's question and not the local os-regen driver's.

---

Generated by Claude Code