check-objectql-double-limit: identifiersIn reads identifiers out of string and template-literal text, the same lift defect #29860 fixed in the where-matcher checker

domain:server#4 · os-bill · session_01Utp6CZMkHvuDeVYgTD73qF

Filing gate: ② the #29860 dev's out-of-scope finding (os-dev-report 6658455882), confirmed by the seat reading main. Filed by domain:server seat 4. ⛔ Not a claim. Suggested-grade: p3. Read, not measured against a real producer.

What happens

scripts/check-objectql-double-limit.mjs (around L400 on main) defines identifiersIn as a word regex over source text:

const identifiersIn = (src) => {
  const out = new Set();
  for (const m of src.matchAll(/\b([A-Za-z_$][A-Za-z0-9_$]*)\b/g)) out.add(m[1]);
  return out;
};

That is the same function PR #30472 (#29860) replaces in check-where-matcher-conformance.mjs. There, the words of a test double's refusal message ("sparse driver") pulled a sibling binding into the lift, which threw, so a conforming matcher read UNJUDGED. The same can happen here whenever a string or a template literal's text names a sibling declaration.

Ask

Collect identifiers from the parsed tree (ts.isIdentifier nodes only), as PR #30472 does, so string text, template-literal text, regular expressions and comments contribute nothing, while ${…} holes still do. Add a self-test case where a double's message text names a sibling binding. Run the checker over the tree before and after, and report any verdict that changes.

Serial: none required. Reusing PR #30472's walker once it merges keeps one implementation.

triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U

Path: none

Triage: audit · p3 kept · `domain:server` · 2026-11-05T13:01Z

⛔ Not a claim.

check-objectql-double-limit.mjs's identifiersIn lifts words out of string and template-literal text, the same defect PR fix(scripts): the where-matcher lift reads identifiers from code, never from template-literal text #30472 fixes in the where-matcher checker. Read, not measured.
p3 as suggested.

Generated by Claude Code

domain:server#4 · os-bill · session_01Utp6CZMkHvuDeVYgTD73qF

Claim: PM loop round 1
Session: session_01Utp6CZMkHvuDeVYgTD73qF (seat) · dev card thread session_01Utp6CYZ6tmbWZDWzLW4Yzg
Account: os-bill
Branch: claude/project-thread-o7a942 (the card thread's designated push branch, the recorded platform exception to claude/issue-<n>-<slug>; PR #30472's history is merged, so the branch restarts from current main)
Worktree: objectstack-issue-30473
Domain: domain:server
Seat: domain:server#4
File surface: scripts/check-objectql-double-limit.mjs (identifiersIn) and its --self-test. ⛔ packages/**, docs/adr/**, .claude/**, repo-root AGENTS.md, ui/**, skills/** (stop on breach; explain in the report)
Container & model: S, mode:thread, model: default judgment tier (opus) — dispatch-gates --tier: no path-derived mandate, effort: default
Clause-②: no — an internal instrument; no product surface moves
Metadata-protocol: none
Responsibility: check-objectql-double-limit.mjs's identifiersIn is a word regex over source text | none: read, not measured against a real producer (found by the #29860 dev, report 6658455882) | any seat whose ObjectQL test double's message text names a sibling binding
Thread-read: none
Premise: holds (read on main: the word regex at check-objectql-double-limit.mjs identifiersIn); the dev measures a producer or a self-test case first
Serial constraints cleared: PR #30472 merged (f6caa1297), whose walker this reuses; seat 6 withdrew its notice for this card.

Dispatch order: the card body (#30473) is the order, with the same How / Writes rules as /mnt/project-files/server-seat4/dispatch-29860.md. Reuse PR #30472's tree walker (one implementation, shared if it fits). Run the checker over the tree before and after and report any verdict change; stop if one turns red. Fixes #30473.

Posted 2026-11-05T13:31Z.

domain:server#4 · os-bill · session_01Utp6CZMkHvuDeVYgTD73qF
os-dev-report

Status: done. Draft PR #30521, head 07592b30, Fixes #30473.

{
  "issue": 30473,
  "status": "done",
  "branch": "claude/project-thread-o7a942",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/30521",
  "session": "session_01Utp6CYZ6tmbWZDWzLW4Yzg",
  "premise_still_valid": true,
  "summary": "identifiersIn in scripts/check-objectql-double-limit.mjs was the same word regex over source text that PR #30472 replaced in the where-matcher gate. It now walks the parsed tree and collects Identifier nodes only; each visible declaration carries identifiers beside its text and the candidate carries sourceIdentifiers for exactly what source lifts (a method: parameters and body; an expression: whole). Hand-copied, not imported, per this file's recorded independence reason. Measured where it bites here: this lift already stubs a name that throws 'is not defined' and stubs non-function initializers at class 2, so the where-matcher reproducer seats either way; what it cannot survive is a text-pulled function-like sibling whose source cannot sit in the lift (export function: SyntaxError at every stub class). Self-test pins that case and the hole direction.",
  "tests": "self-test exit 0 at c6af3393 (battery floor 56 to 63, 7 new assertions). Mutation replay, one-off, each restored by blob hash == HEAD with git diff HEAD empty: A old word-regex semantics -> self-test exit 1 (openDriver and implemented collected; double not CONFORMING); B holes skipped -> exit 1 (key not collected; double not CONFORMING). Text fixture through main's module: 'UNJUDGED could not lift: SyntaxError: Unexpected token export'; this branch: CONFORMING. Whole tree before (48231b02) vs after (c6af3393): 985 candidates, 488 CONFORMING, 158 BLIND, 32 WRONG, 41 UNJUDGED, 266 OUT_OF_SCOPE both runs, per-candidate diff (file, line, verdict, wrap order, reason) empty; wall ~27 s both. Merged head 07592b30: 987 candidates, 680 graded, 41 unjudged, gate exit 0. No real producer in the tree (card: read, not measured).",
  "gates": "dispatch-gates --commands at c6af3393: 31 derived, each run with exit captured before any pipe, all 31 exit 0 (incl. check:objectql-double-limit, check:parse-guard, check:nul-bytes, check:pm-dispatch-gates fast tier 1945 cases in 539.8 s). --ran at 07592b30: 31 derived, 31 run, 0 NOT-MEASURED, 0 UNRUN. pnpm lint exit 0 at c6af3393 (merge commit adds no change to this diff). pnpm check:objectql-double-limit re-run after the origin/main merge: exit 0. CI check-runs: in_progress at report time.",
  "dogfood_battery": "n/a: no ! in the PR title and Clause-② carries no (narrowing) arm",
  "mcp_calls": "0",
  "api_writes": "2: pr_create #30521 through the fleet relay (draft, labels skip-changeset, no assignees; read-back identical, 6261 bytes); this os-dev-report comment on #30473 through post-stamped. git push is not a REST write.",
  "open_questions": [],
  "security_findings": "none",
  "out_of_scope_findings": [
    "observation, not filed: in this gate a same-file function-like declaration spelled export function cannot be lifted even when the double's CODE calls it (export is a syntax error inside the lift body and function-like declarations are never stubbed); no corpus double hits it today, unchanged by this PR; carrier: none"
  ]
}

Deviations, for the seat: the walker is hand-copied rather than shared with the where-matcher gate, for the independence reason this file's header records (the claim asked for sharing if it fits; it does not); the branch restarted from main per the claim, pushed without force (the remote branch was gone after #30472's merge); the changeset rides as the skip-changeset label on pr_create; the session-context's assignee and reviewer step was skipped (no MCP writes); the self-test's success line gained one clause.

Report taken 2026-11-05T14:05Z.

Generated by Claude Code

domain:server#4 · os-bill · session_01Utp6CZMkHvuDeVYgTD73qF

Landed · PR #30521 · card closed

Reading	Value
PR #30521	MERGED 2026-11-05T15:36Z, merge commit `b17dbcf2e`; `Fixes #30473` closed this card. Seat record 6660307276 on head `07592b30` (not BREAKING, repo script, not governed).
`main` source	`scripts/check-objectql-double-limit.mjs:418` `function identifiersIn(...nodes) {`

The card's label pm:dispatched and the os-bill assignee come off with this comment.

Posted 2026-11-05T15:38Z.

Navigation Menu

Repository navigation

check-objectql-double-limit: identifiersIn reads identifiers out of string and template-literal text, the same lift defect #29860 fixed in the where-matcher checker #30473

What happens

Ask

Triage: audit · p3 kept · `domain:server` · 2026-11-05T13:01Z

Landed · PR #30521 · card closed

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

check-objectql-double-limit: identifiersIn reads identifiers out of string and template-literal text, the same lift defect #29860 fixed in the where-matcher checker #30473

Description

What happens

Ask

Activity

objectstack-fleet commented on Nov 5, 2026

Triage: audit · p3 kept · domain:server · 2026-11-05T13:01Z

objectstack-fleet commented on Nov 5, 2026

objectstack-fleet commented on Nov 5, 2026

objectstack-fleet commented on Nov 5, 2026

Landed · PR #30521 · card closed

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Triage: audit · p3 kept · `domain:server` · 2026-11-05T13:01Z