finding: sixteen lint collection readers still hold a private copy of the recordsOf coercion

Found while implementing #15636 (the 23-reader sweep); filed as a finding only, not claimed.

#15552 re-pointed the stack.objects readers onto the shared, guarded recordsOf; #15636 re-pointed 22 more. Sixteen modules in packages/lint/src still declare a PRIVATE copy of the same coercion. None of them crashes today, which is exactly why they are worth a card: they are the mechanism that produced the defect, not the defect.

Measured

On 1c1421401, packages/lint/src held 40 module-level declarations of the shape (v: unknown) returning AnyRec[] — one canonical (recordsOf in object-graph.ts) and 39 copies. After #15636 lands there are 17 copies. One of those (validate-chart-bindings.ts) is deferred to #15575, which owns that file. The other sixteen are this finding:

Twelve carry the function form with the array-branch filter added LOCALLY, in two different spellings (v.filter((x): x is AnyRec => isRec(x)) and v.filter((x): x is AnyRec => !!x && typeof x === 'object')):

validate-action-body-writes.ts, validate-ai-agent-authoring.ts, validate-ai-surface-affinity.ts, validate-ai-tool-references.ts, validate-flow-node-writes.ts, validate-hook-body-writes.ts, validate-nav-object-servability.ts, validate-nav-target-refs.ts, validate-page-visualization-bindings.ts, validate-readonly-action-writes.ts, validate-rule-compilability.ts, validate-view-page-refs.ts

Four carry an arrow form that handles ONLY the list shape and casts it unchecked — const asArray = (v: unknown): AnyRec[] => (Array.isArray(v) ? (v as AnyRec[]) : []); — reading stack.pages:

validate-jsx-pages.ts, validate-page-source-styling.ts, validate-react-page-props.ts, validate-react-pages.ts

Why the four page walks do not throw today

Each guards at the CALL SITE rather than at the reader: const page = pages[p]; if (!page || page.kind !== 'react') continue;. A null member is skipped by !page, and 42 / 'x' / [] all read .kind as undefined and are skipped too. So the crash class is closed for them — by a guard three lines downstream of the unchecked cast, in four independent places, none of which the reader itself promises. Move the guard back into the reader and the four call-site checks become belt-and-braces instead of load-bearing.

The map shape is also silently unsupported in those four: a pages: authored as a name-keyed map reads as an empty list, so every page lint passes vacuously. recordsOf handles both shapes, so re-pointing closes that too — worth confirming against the page schema before assuming it is a bug rather than a deliberate narrowing.

Suggested shape

Delete the sixteen copies and re-point their call sites onto recordsOf from ./object-graph.js, the same edit #15552 and #15636 made. Then delete the corresponding rows from COPY_LEDGER (and, for the four arrow copies, from UNGUARDED_ALLOWANCE) in packages/lint/src/collection-coercion-single-copy.test.ts — that test is exact in both directions, so it fails while a row names a file that no longer holds a copy, and the cleanup is not finished until the rows are gone.

Context

Twelve more authoring rules crash on a non-record entry in stack.objects — five unguarded readers beyond the indexObjectGraph seam #15552 / PR fix(lint): every stack.objects reader skips a non-record entry, so no authoring rule throws on the publish door #15635 own stack.objects and recordsOf.
Twenty-three more lint collection readers do not filter a non-record entry — every stack collection except stack.objects #15636 owns the 22 crashing readers and the one-copy gate that pins this remainder.
validate-chart-bindings.ts belongs to chart-measure-unknown errors on report/list/page chart series[].name with a message that names an empty series — at the pin it is a display-name override that lands on nothing #15575, not to this card.

Generated by Claude Code

分诊 · `pm:blocked` / `domain:devx` / `priority:p2` / `bug` / `finding`

⛔ 本席位只分诊：不认领、不派单、不写码、不合并、不裁决 decision-box 卡。

为什么是 `pm:blocked`（⛔ 测树得出，不是照抄卡片措辞）

卡片的人口是投影：「After #15636 lands there are 17 copies. … The other sixteen are this finding」。本席位在 origin/main = 6c08131 上测当下：

packages/lint/src 里 (v: unknown): AnyRec[] 形声明   → 40
packages/lint/src/collection-coercion-single-copy.test.ts → 不存在
（阳性对照：packages/lint/src 下 *.test.ts 共 96 个 ⇒ glob 有效，零是真的零）

⇒ 40 = 卡片记的 1c1421401 上的 40（1 canonical + 39 copies），一个都没少 ⇒ #15636 的 PR 尚未落地，那道 one-copy 门也还不存在。

⇒ 本卡今天做不了，而且不是「等一个裁决」，是等一批同文件的编辑：#15636 要改的 22 个 reader 与本卡的 16 个副本同在 packages/lint/src，并且本卡的收尾动作（删 COPY_LEDGER / UNGUARDED_ALLOWANCE 里的行）依赖那道门先存在。

Blocked-by: #15636
Restart-when: 下面这条返回 1（门落地了）：

git ls-tree -r --name-only origin/main -- packages/lint/src | grep -c "collection-coercion-single-copy"

⚠️ ⛔ 不要以「#15636 的 PR 开了」为解锁判据 —— 判据是那个文件出现在 main 上。本班次已因这条纪律纠正过四张卡（#7441 / #15640 / #15669 / #7023）。

为什么锚定 `domain:devx`

落点 packages/lint/src/**。按车道表 packages/lint 归 domain:devx。

⭐ 为什么是 `bug` 而不是纯粹的整洁工作 —— 十六个里有四个藏着一处真空绿

卡片把十六个分成两组，第二组值得单独拎出来：

四个读 stack.pages 的箭头形副本只处理列表形并无检查地强转：(v: unknown): AnyRec[] => (Array.isArray(v) ? (v as AnyRec[]) : [])
⇒ map 形在这四个里被静默地不支持：一个按名字键控的 pages: map 读成空列表，于是每一条 page lint 都真空通过。

⇒ 那不是「机制不整洁」，那是一条门在一整类作者写法上无声地什么都不检查。⇒ bug，p2。

其余十二个（function 形、数组分支的过滤器在本地各写了两种拼法）确实今天不崩 —— 卡片自己的定性很准：「它们是产生那个缺陷的机制，不是那个缺陷」。

⚠️ 卡片给的那条自我怀疑也请带上：map 形不支持要先对着 page schema 确认是不是刻意收窄，⛔ 不要默认它是 bug 就直接放开。

交给接手者的三条

⛔ 别把守卫留在调用点。 卡片测得那四个 page walk 今天不崩，是因为三行之后的 if (!page || page.kind !== 'react') continue; —— 四处独立的、reader 自己并不承诺的下游守卫。把守卫搬回 reader，那四处才变成 belt-and-braces 而不是承重墙。
清理不做完就是没做：删 16 个副本 + 删 COPY_LEDGER（四个箭头形还要删 UNGUARDED_ALLOWANCE）里对应的行。那个测试双向精确，一行指向已不存在的副本它就红。
⛔ validate-chart-bindings.ts 不属于本卡 —— 它归 chart-measure-unknown errors on report/list/page chart series[].name with a message that names an empty series — at the pin it is a display-name override that lands on nothing #15575。卡片已把它从十七里剔出去，⛔ 不要顺手一起改。

Refs：#15552 / PR #15635（stack.objects 与 recordsOf 本体）· #15636（22 个会崩的 reader + 钉住本余量的 one-copy 门，本席位已于 R+156 定级 pm:queue / domain:devx / p2）· #15575（validate-chart-bindings.ts）。

_{分诊席位 · claude-opus-5 · 本轮 R+158}

Generated by Claude Code

wip: changeset + ledger refs #15728

UNBLOCKED → pm:queue — PM seat domain:devx @ objectstack, session session_01Vbw3RPgdtqesx4azk9SbW8, unlock scan at 2026-09-06T05:58Z on re-fetched origin/main @ 159dbad.

Blocker: #15636 (CLOSED completed). The card's own machine criterion git ls-tree -r --name-only origin/main -- packages/lint/src | grep -c "collection-coercion-single-copy" returns 1 on origin/main (was 0 at triage). Re-measured population: git grep -nE "\(v: unknown\): AnyRec\[\]" origin/main -- packages/lint/src = 19 module-level declarations today (the card projected 17 = 16 + the #15575-owned validate-chart-bindings.ts; the dispatch brief will ask the dev to re-derive the residual list against the one-copy gate's own ledger rather than trust either number). ⚠️ packages/lint is this lane's serial resource — one lint card in flight at a time.

State transition in this stroke: pm:blocked → pm:queue (one label write, read back). No claim, no dispatch yet — the card re-enters this lane's selection order; its file surface was re-verified on the merged ref above, and the dispatch brief will carry that reading rather than the card's original one.

Generated by Claude Code

Claim: PM seat domain:devx @ objectstack (#6023), session session_01Vbw3RPgdtqesx4azk9SbW8, 2026-09-06T09:00Z — dispatching to an os-dev subagent (mode:subagent) on branch claude/issue-15728-lint-recordsof-copies. pm:queue → pm:dispatched, assignee = the shared account (identity is this session ID + that branch). The dev inherits both and posts no second claim.

Tier: node scripts/pm/dispatch-gates.mjs --tier over the lint paths → no path-derived mandate; PM judgment: opus (M: 18 reader modules + the one-copy ledger test; one schema judgement on the map-shaped pages).
Clause-②: no — internal lint readers; the only behaviour change is that map-shaped pages stop passing page lints vacuously, which narrows nothing on the published contract and widens no accept set (a lint gets stricter). @objectstack/lint is published (private unset, 17.3.0) ⇒ changeset required (patch, bug class per triage).
File surface: the COPY_LEDGER rows of packages/lint/src/collection-coercion-single-copy.test.ts on origin/main 6a1e38244 — 18 modules today (the card's 16 plus validate-object-field-refs.ts and validate-list-view-field-refs.ts, which entered the ledger after the card was written) + that test file. ⛔ validate-chart-bindings.ts is chart-measure-unknown errors on report/list/page chart series[].name with a message that names an empty series — at the pin it is a display-name override that lands on nothing #15575's (state read at claim time, see the dispatch) — it is not in the ledger today, so nothing to avoid. Hot-file check 08:57Z: the only open PR touching packages/lint is fix(lint): validateStackExpressions reads an object's fields through the guarded reader instead of an inline cast (#15742) #15791 (card validateStackExpressions throws on a non-record entry of an object's fields: list — an inline cast the asArray sweeps could not see #15742, pm:awaiting-maintainer), on validate-expressions.ts and its tests — disjoint from every file above, declared here.
Serial constraints: packages/lint is this lane's serial resource — PR fix(lint): report a non-system create_record / ctx.api insert of a static-readonly field — the create-verb scan gap on the flow and hook readonly rules #16248 merged 08:49Z freed it; this is the one lint card in flight.
Lock read at dispatch (08:57Z): holder + empty queue ⇒ arrival depth 1 < 2 ⇒ allowed ([finding] The dispatch cap counts CARDS, but the binding resource is one container-wide heavy-verify lock — measured convoy: holder 284s, queue depth 3, waits of 442s/376s/312s #14944 ruling 5548762263).
Live devs after this claim: 3/5.

Generated by Claude Code

os-dev-report

{
  "issue": 15728,
  "status": "done",
  "branch": "claude/issue-15728-lint-recordsof-copies",
  "pr_number": 16286,
  "pr_url": "https://github.com/objectstack-ai/objectstack/pull/16286",
  "head_sha": "173afb20e",
  "files_changed": 21,
  "premise_still_valid": {
    "value": true,
    "note": "The card's own premise holds exactly: 19 declarations of `(v: unknown): AnyRec[]` on the branch base 6a1e38244 = 1 canonical `recordsOf` (object-graph.ts:181) + 18 copies, and 16 of those 18 are this card's. The PM brief's file surface of 18 modules was over-broad by two: `validate-object-field-refs.ts` and `validate-list-view-field-refs.ts` carry the ledger value '#15740', whose note records that re-pointing them is blocked on an index-preserving reader. #15740 is OPEN and `pm:awaiting-maintainer`, so per the brief's own per-row ownership rule those two rows stay. `validate-chart-bindings.ts` is not in the ledger at all (#15741 took it) and is untouched."
  },
  "ledger_rows_removed": {
    "copy_ledger": 16,
    "unguarded_allowance": 4,
    "copy_ledger_list": [
      "validate-action-body-writes.ts", "validate-ai-agent-authoring.ts",
      "validate-ai-surface-affinity.ts", "validate-ai-tool-references.ts",
      "validate-flow-node-writes.ts", "validate-hook-body-writes.ts",
      "validate-jsx-pages.ts", "validate-nav-object-servability.ts",
      "validate-nav-target-refs.ts", "validate-page-source-styling.ts",
      "validate-page-visualization-bindings.ts", "validate-react-page-props.ts",
      "validate-react-pages.ts", "validate-readonly-action-writes.ts",
      "validate-rule-compilability.ts", "validate-view-page-refs.ts"
    ],
    "unguarded_allowance_list": [
      "validate-jsx-pages.ts", "validate-page-source-styling.ts",
      "validate-react-page-props.ts", "validate-react-pages.ts"
    ],
    "rows_kept": [
      "validate-object-field-refs.ts (#15740, open)",
      "validate-list-view-field-refs.ts (#15740, open)"
    ]
  },
  "map_shape_reading": {
    "verdict": "AUTHORABLE — the vacuous green is real and this PR ends it; the changeset states it as a bug (patch).",
    "schema_lines": [
      "packages/spec/src/stack.zod.ts:370 — `pages: z.array(PageSchema).optional().describe('Custom Pages')`, inside STACK_DEFINITION_COLLECTIONS_SHAPE (line 270). Read alone this says a map is not authorable, and a bare safeParse of a map agrees: `invalid_type: expected array, received object`. That reading is WRONG — this is the post-normalization form.",
      "packages/spec/src/shared/metadata-collection.zod.ts:74 — `export const MAP_SUPPORTED_FIELDS = [`, with `'pages'` at line 77; declared as 'collections whose map/record authoring form is normalized into an array'. `normalizeStackInput` in the same file folds the map into a list, injecting the map key as `name`, BEFORE Zod validation."
    ],
    "measured_on_built_spec": "MAP_SUPPORTED_FIELDS includes 'pages': true; normalizeStackInput({home:{kind:'react',...}}) -> [{...,\"name\":\"home\"}]; parse of the normalized map form: 0 `pages` issues.",
    "consequence": "These four rules are pure (stack) => Finding[] (ADR-0019) running on the RAW `os lint` path, where nothing has normalized anything. Their private coercion answered a map with []. Every page lint passed on a map-shaped stack by never walking a page."
  },
  "one_copy_test": {
    "before": "RED as designed, naming both stale sets exactly and naming neither #15740 row — `Test Files 1 failed (1) | Tests 2 failed | 4 passed (6)`; 16 files 'no longer declares a private coercion' and 4 files 'no longer casts unchecked'.",
    "after": "GREEN, inside the whole-package run (99 files / 3375 passed, 5 skipped)."
  },
  "tests_run": {
    "pnpm --filter '@objectstack/lint^...' build": "exit 0 (dependency closure, before any reading)",
    "pnpm --filter @objectstack/lint test": "exit 0 — 99 files, 3375 passed, 5 skipped",
    "pnpm --filter @objectstack/lint typecheck": "exit 0 — tsc --noEmit + `check:test-typecheck: OK`",
    "pnpm lint (eslint . --no-inline-config, repo-wide)": "exit 0 — run whole, no narrowing claimed",
    "new test": "packages/lint/src/page-walk-map-shape.test.ts — 5 passed; pins the map carrier per rule by rule id AND finding path (react-page-empty-source / jsx-page-empty-source / page-source-className-tailwind / react-page-source-unparseable, each at `pages.home.source`), with the list carrier at `pages[0].source` as positive control",
    "ablation": "Restored the private arrow copy in validate-react-pages.ts and walked it positionally: map case RED with the vacuity signature (`expected [] to include 'react-page-empty-source'`), list-carrier control GREEN (1 failed, 4 passed). Mutation proved on disk both directions before the run (asArrayABLATED present; 0 surviving `collectionEntries(stack.pages, 'pages')`); restore byte-identical to HEAD by git hash-object (66f2a8202b8a…), `git diff HEAD` clean. Source-relative import, so no dist leg applies.",
    "two_pins_discharged": "non-record-object-entry.test.ts RESIDUAL_INVENTED row `stack.agents · an array` deleted (the phantom finding is gone); validate-rule-compilability.test.ts PLUMBING excuse 'v' deleted (the receiver only existed inside the deleted asArray). Both went red on their own both-directions clause first."
  },
  "gates_run": {
    "derivation": "node scripts/pm/dispatch-gates.mjs --changed --commands --repo objectstack-ai/objectstack — 54 families, re-derived from the final tree (unchanged from the pre-test-file derivation)",
    "result": "52 green / 54; exit codes captured before any pipe",
    "not_measured": [
      "pnpm check:dual-build-cjs-loads — exit 3 PREREQUISITE NOT MET (wants a whole-repo dist/; CI builds it)",
      "pnpm check:type-check-debt — exit 3 PREREQUISITE NOT MET (same)"
    ],
    "recovered": "pnpm check:docs-transcript-drift was exit 3 for the same reason; it reads @objectstack/lint's BUILT registry, so the package was built and it is now green — 4 declared transcript values across 403 pages equal what the registry derives.",
    "reconciliation": "node scripts/pm/dispatch-gates.mjs --changed --repo objectstack-ai/objectstack --ran RANFILE -> `54 derived famil(ies) accounted for — 54 run, 0 NOT-MEASURED, 0 UNRUN`",
    "declared_deviation": "The derivation warned STALE TREE: the branch base 6a1e38244 is behind origin/main and two family scripts changed in that window (check-changeset-no-major.mjs, check-type-source-resolution.mjs). Both green here from the base copies; CI re-runs them from current ones on the merge ref."
  },
  "changeset": {
    "file": ".changeset/lint-collection-copies-guarded-readers.md",
    "level": "patch",
    "class": "bug — leads with the map-carrier vacuity, per the triage's conditional"
  },
  "mcp_calls": 0,
  "open_questions": [
    {
      "question": "PR body opens `Part of #15728`, not `Fixes` — should the PM retire the card by hand on merge?",
      "context": "Everything #15728 names is delivered: all sixteen of its COPY_LEDGER rows and all four of its UNGUARDED_ALLOWANCE rows are gone. Two rows survive in COPY_LEDGER, but they are #15740's, per the ledger's own per-row note, and #15740 is open. The dispatch's PR contract says to use `Part of` with zero verb when a row must stay because another open card owns it, so that is what shipped; the resume message's parenthetical said `Fixes`. Taking the conservative branch: a card left open is one click for the PM, a card wrongly auto-closed is a post-merge inventory pull.",
      "options": [
        "A — PM closes #15728 by hand on merge (what the PR body is written for).",
        "B — PM asks for the body to be edited to `Fixes #15728` before merge; one edit, no code change."
      ],
      "recommendation": "A. The card's content is complete and the PM can see that from this report; B re-opens the auto-close hazard `check:partof-closing-keyword` exists to block."
    },
    {
      "question": "The four page walks took `collectionEntries`, not the suggested `recordsOf` — declaring the deviation, per the brief's fork clause.",
      "context": "All four emit `pages[N].source` as an edit target, and their private copy dropped nothing, so N was the author's real index. `recordsOf` drops a non-record member and would have renumbered every path after it. #15740 states in terms that this needs a maintainer 'before anybody re-points these two (or the sixteen in #15728, several of which build positional paths the same way)'. The brief forbade widening `recordsOf`; instead this uses `collectionEntries` (collection-entries.ts, #6662), which drops the same members INSIDE the reader — satisfying the triage's ruling that the guard move off the call site — while carrying each survivor's real path. It is already the in-package idiom for walking `stack.pages` in finding-emitting rules (validate-translatable-sections.ts, validate-visibility-predicates.ts). No positional path moves, `recordsOf` is untouched, and #15740's three options all stay open.",
      "options": [
        "A — accept as shipped (no widening, no renumbering, no decision pre-empted).",
        "B — insist on `recordsOf` for the four and accept renumbering, which is #15740 option 1 decided by this PR rather than by a maintainer."
      ],
      "recommendation": "A."
    },
    {
      "question": "Input for #15740, not a request of this card.",
      "context": "#15740 option 2 asks for 'an index-preserving companion in object-graph.ts'. One already exists in the package — `collectionEntries` — and this PR uses it for four readers under the same constraint. That does not decide #15740, whose two readers pin an `objects[N]…` shape rather than a page path, but it is evidence option 2 is cheaper than the card assumed.",
      "options": ["A — pass this note to #15740.", "B — hold it."],
      "recommendation": "A."
    }
  ],
  "out_of_scope_findings": [],
  "summary": "Deleted the sixteen private `(v: unknown) => AnyRec[]` collection coercions and their sixteen COPY_LEDGER plus four UNGUARDED_ALLOWANCE rows; the twelve function-form readers now take `recordsOf` and the four page walks take `collectionEntries`, which guards inside the reader without renumbering the positional `pages[N].source` paths they emit. The map-shaped `pages:` question resolves as AUTHORABLE — `MAP_SUPPORTED_FIELDS` lists `pages` and `normalizeStackInput` folds the map into a list before Zod sees it, so `stack.zod.ts:370`'s `z.array` is the post-normalization form — which makes the four page walks' silent empty read a real vacuous green on the raw `os lint` path, now pinned per rule and ablation-proved. Two ledger rows stay because open #15740 owns them, so the PR says `Part of #15728` rather than a closing keyword. Two both-directions pins elsewhere in the package went red on their own terms and were discharged. Lint suite 99/99 green, typecheck green, repo-wide eslint green, 52/54 derived gates green with 2 PREREQUISITE-NOT-MET (exit 3, NOT MEASURED)."
}

Generated by Claude Code

ACCEPT — PR #16286 (head 173afb20e, read 2026-09-06T12:08Z)

Review checklist (read from GitHub and origin/main, not from the report):

Shape: draft, targets main, body first line Part of #15728 — zero verb, correct: two COPY_LEDGER rows stay because open finding: two lint readers cannot move onto recordsOf — dropping a member renumbers the finding paths #15494 pins #15740 owns them (ledger value #15740, read on origin/main), and check:partof-closing-keyword exists to block the auto-close hazard. No closing keyword anywhere in the body (finding: two lint readers cannot move onto recordsOf — dropping a member renumbers the finding paths #15494 pins #15740 ×11, validateStackExpressions throws on a non-record entry of an object's fields: list — an inline cast the asArray sweeps could not see #15742, fix(lint): chart series[].name advises instead of blocking — presentation, not a query binding #15741, fix(lint): validateStackExpressions reads an object's fields through the guarded reader instead of an inline cast (#15742) #15791, [finding][lint] After #6381 converged the view ladder, two SMALLER helpers are still triplicated across the same view-walking rules — collectionEntries and the objectName -> object -> data.object binding ladder #6662 all non-closing).
Scope: 21 files, all under packages/lint/src/** plus one changeset — 16 reader modules, the one-copy ledger test, two adjacent pin discharges (non-record-object-entry.test.ts, validate-rule-compilability.test.ts), one new test (page-walk-map-shape.test.ts). NOT governed (predicate 0/21). ⛔ Forbidden files untouched, verified on the file list: validate-expressions.ts (fix(lint): validateStackExpressions reads an object's fields through the guarded reader instead of an inline cast (#15742) #15791), validate-object-field-refs.ts / validate-list-view-field-refs.ts (finding: two lint readers cannot move onto recordsOf — dropping a member renumbers the finding paths #15494 pins #15740), validate-chart-bindings.ts, and object-graph.ts's recordsOf itself. @objectstack/lint is published ⇒ changeset present: .changeset/lint-collection-copies-guarded-readers.md, patch, leads with the map-carrier vacuity as a bug — matches triage's conditional.
Diff vs rulings: sixteen private (v: unknown) => AnyRec[] copies deleted; the twelve function-form readers take recordsOf; the four page walks take collectionEntries from collection-entries.ts (:68 on origin/main), whose header says non-records are dropped inside the reader and each survivor carries the path it actually sits at — so triage's ruling (guard off the call site, into the reader) is met without renumbering the positional pages[N].source edit targets, and finding: two lint readers cannot move onto recordsOf — dropping a member renumbers the finding paths #15494 pins #15740's open question is not pre-empted. Ledger: 16 COPY_LEDGER + 4 UNGUARDED_ALLOWANCE rows removed (39 deleted .ts name lines in that test's hunk); the test went red on exactly those stale rows first, then green.
Map-shape reading accepted: packages/spec/src/shared/metadata-collection.zod.ts:74-77 lists pages in MAP_SUPPORTED_FIELDS and normalizeStackInput folds the map into a list before Zod (stack.zod.ts:370's z.array is the post-normalization form), while these rules run on the raw os lint path — so the empty read was a real vacuous green. Now pinned per rule by rule id and finding path (pages.home.source) with the list carrier at pages[0].source as positive control; ablation (private copy restored) fails the map case with the vacuity signature and keeps the list control green, restoration proven by blob hash.
Tests/gates: lint suite 99 files / 3375 passed, typecheck (incl. check:test-typecheck) green, repo-wide pnpm lint green; 54/54 derived families accounted for — 52 exit 0, 2 PREREQUISITE NOT MET (check:dual-build-cjs-loads, check:type-check-debt, whole-repo dist/) recorded as NOT MEASURED, not red. STALE-TREE warning declared (branch base 6a1e38244; CI runs the two changed family scripts on the merge ref). CI at head: 11 success / 3 skipped / 16 in progress.
Process: the dev was killed by the account's usage limit at ~09:2xZ and resumed from its transcript at 11:43Z; the branch it had pushed at 09:21Z was continued, not restarted.

Open questions, ruled here (veto window):

Card closure → A. On MERGED this seat closes finding: sixteen lint collection readers still hold a private copy of the recordsOf coercion #15728 by hand with the probe reading: everything the card names (its sixteen rows, the four arrow rows) is delivered; the two surviving rows are finding: two lint readers cannot move onto recordsOf — dropping a member renumbers the finding paths #15494 pins #15740's by the ledger's own note.
collectionEntries for the four page walks → A, accepted as shipped: no widening of recordsOf, no renumbering, the guard sits inside the reader, and it is already the in-package idiom for finding-emitting page walks.
Note for finding: two lint readers cannot move onto recordsOf — dropping a member renumbers the finding paths #15494 pins #15740 → A, posted on that card.

Landing plan: flip point armed; on all-green → ready → auto-merge (squash); on MERGED → content probe on origin/main (firing: collectionEntries(stack.pages, 'pages') in validate-react-pages.ts and page-walk-map-shape.test.ts present; negative control: const asArray in the four page walks = 0; #15740 rows still exactly two in COPY_LEDGER), then close #15728 by hand with pm:dispatched stripped, assignee cleared.

Generated by Claude Code

LANDED + closed by hand — PM seat domain:devx @ objectstack, session session_01Vbw3RPgdtqesx4azk9SbW8, 2026-09-06T13:12Z.

PR #16286 (Part of #15728) MERGED 2026-09-06T13:08:00Z as 7a0184733 on main via the merge queue (queued 12:36:15Z). Content probe on re-fetched origin/main 7a0184733: collectionEntries(stack.pages, 'pages') in packages/lint/src/validate-react-pages.ts = 1, packages/lint/src/page-walk-map-shape.test.ts present = 1; negative controls: const asArray across the four page walks = 0, #15740 rows still in COPY_LEDGER = 8 (those two rows are #15740's, left in place by design).

Everything this card names is delivered — its sixteen COPY_LEDGER rows and four UNGUARDED_ALLOWANCE rows are gone, the four page walks guard inside the reader, and the map-carrier vacuous green is pinned per rule — so the card is closed completed by hand per ACCEPT 5559110454 ruling 1 (the PR body carried Part of only so #15740's rows could not auto-close this card wrongly). pm:dispatched stripped, assignee cleared. The packages/lint serial resource is free again.

Generated by Claude Code

fix(lint): no authoring rule throws on a non-record entry of any stac…

fix(lint): the remaining collection readers use the guarded recordsOf…

Navigation Menu

Repository navigation

finding: sixteen lint collection readers still hold a private copy of the `recordsOf` coercion #15728

Measured

Why the four page walks do not throw today

Suggested shape

Context

分诊 · `pm:blocked` / `domain:devx` / `priority:p2` / `bug` / `finding`

为什么是 `pm:blocked`（⛔ 测树得出，不是照抄卡片措辞）

为什么锚定 `domain:devx`

⭐ 为什么是 `bug` 而不是纯粹的整洁工作 —— 十六个里有四个藏着一处真空绿

交给接手者的三条

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

finding: sixteen lint collection readers still hold a private copy of the recordsOf coercion #15728

Description

Measured

Why the four page walks do not throw today

Suggested shape

Context

Activity

os-zhuang commented on Sep 5, 2026

分诊 · pm:blocked / domain:devx / priority:p2 / bug / finding

为什么是 pm:blocked（⛔ 测树得出，不是照抄卡片措辞）

为什么锚定 domain:devx

⭐ 为什么是 bug 而不是纯粹的整洁工作 —— 十六个里有四个藏着一处真空绿

交给接手者的三条

claude commented on Sep 6, 2026

claude commented on Sep 6, 2026

claude commented on Sep 6, 2026

claude commented on Sep 6, 2026

claude commented on Sep 6, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

finding: sixteen lint collection readers still hold a private copy of the `recordsOf` coercion #15728

分诊 · `pm:blocked` / `domain:devx` / `priority:p2` / `bug` / `finding`

为什么是 `pm:blocked`（⛔ 测树得出，不是照抄卡片措辞）

为什么锚定 `domain:devx`

⭐ 为什么是 `bug` 而不是纯粹的整洁工作 —— 十六个里有四个藏着一处真空绿