[maintainer] refactor(objectql): split src/engine.ts (23,903 lines, +7,501 in 30 days) along its ObjectQL seams, one move-only slice per PR

Filing gate ③: a task the maintainer dispatched, quoted below. Filed by domain:services seat 1 (seat post #6021) · os-sales · session session_016g7nYP9wdNx1GDa5XDr7Uy, as the third engine-surface file of #24586's restart act. #25467 (protocol.ts) went first as the ruling orders, and its cut 1 landed in PR #25494. ⛔ Not a claim.

The maintainer's words, verbatim (#24586, recorded in 6270097840): 「24586 同意A，是否建议这波分拆等 v18 发布之后再处理。」 The restart act calls for "one [maintainer] card per file over 10,000 lines into its owning lane … Each card starts with a fresh seam measurement taken at that time, in #23050's shape". Triage read the restart condition as true in 6291735609.

Reader who acts: the domain:server lane, engine surface. One dispatch per slice. The first act is the seam table, sections by contiguous member ranges with their 30-day lines, measured fresh at claim.

Measured (on `origin/main` `842a0aef9d`, 2026-10-23)

Size and churn:
- 23,903 lines, up from 16,402 thirty days ago (dabf8d795e), a gain of +7,501 (+45.7%).
- 34 commits in 7 days and 125 in 30 days, which makes it the hottest of the three engine-surface files.
- git blame --since=2026-09-23T15:00Z attributes 8,337 lines (34.9%) to the last 30 days.
Structure:
- One ObjectQL class of 18,214 lines with 254 methods: 147 private or protected, 107 public.
- 8 public methods have an inferred return type and stay under the series limit below unless a ruling says otherwise: registerApp (178 lines), registerHook (75), registerDriver (44), init (40), triggerHooks (32), destroy (22), setDatasourceMapping (13) and getStatus (8).
- The rest is the module preamble and the helpers.
What reads it by path:
- the ADR anchor file scripts/adr-anchors/packages__objectql__src__engine.ts.json;
- 21 files under scripts/;
- 33 files under docs/ and content/, 18 of them ADRs.
In flight: none of the 13 open PRs touch the file today. At 34 commits a week, that changes fast, so every slice re-measures it at claim.

The ask (from #24586's ruling and #23050's shape, as on #25467)

Split along the existing seams, one area per PR. Each slice is move-only: no logic changes.
Shared state stays explicit. Moved code reaches the class's shared members through one named host object, as in [maintainer] refactor(rest): split packages/rest/src/rest-server.ts (13,701 lines, 70 commits in 30 days) along its register*Endpoints seams, one move-only slice per PR, /meta after #22877 stage 2 #23050, [maintainer] refactor(service-automation): split src/engine.ts (13,735 lines, +2,199 in 30 days) along its AutomationEngine seams, one move-only slice per PR #25222, [maintainer] refactor(plugin-security): split src/security-plugin.ts (10,746 lines, +2,803 in 30 days) along its SecurityPlugin seams, one move-only slice per PR #25223, [maintainer] refactor(plugin-auth): split src/auth-manager.ts (10,338 lines, +1,867 in 30 days) along its AuthManager seams, one move-only slice per PR #25224 and [maintainer] refactor(metadata-protocol): split src/protocol.ts (33,065 lines, +10,169 in 30 days) along its ObjectStackProtocolImplementation seams, one move-only slice per PR #25467.
- ⛔ Not a second copy of a helper.
- ⛔ Not a re-derivation.
Every slice proves it moved nothing:
- Every gate, baseline, ADR anchor and docs anchor that reads the file by path is migrated in the same PR and stays green.
- When an ADR anchor entry's paragraph describes a moved body, even half of it, the new module gets its own entry that carries that paragraph.
- The package's public .d.ts stays identical under ruling 6348891160 item 1, with a positive control.
- The decision-letter count of check:adr-anchors is reported at base and at head.
- ⛔ A gate whose target moved must not go silently empty.
Serial with the file's in-flight PRs. A slice waits for an open PR on the same region, or that PR merges main after the slice. Every slice measures the file's open PRs again at claim.
⛔ No new gate.

Series limits already measured (carry them forward)

Inferred return types. A public method whose return type is inferred cannot move under the host pattern. The .d.ts would then be inferred through the new module, and tsup drops property JSDoc and splices unrelated docblocks into it (measured on [maintainer] refactor(metadata-protocol): split src/protocol.ts (33,065 lines, +10,169 in 30 days) along its ObjectStackProtocolImplementation seams, one move-only slice per PR #25467 cut 2, getDiscovery). Such a member stays. Annotating it is a signature edit, which needs a ruling.
The tenant-audit census. UNTYPED_RECEIVERS is shrink-only. A body with a this.<field>.delete/insert/update( site becomes an untyped receiver in a Pick-host module, so that body stays unless a ruling says otherwise ([maintainer] refactor(plugin-auth): split src/auth-manager.ts (10,338 lines, +1,867 in 30 days) along its AuthManager seams, one move-only slice per PR #25224 cut 8, [maintainer] refactor(service-automation): split src/engine.ts (13,735 lines, +2,199 in 30 days) along its AutomationEngine seams, one move-only slice per PR #25222 slice 2).
Private statics stay in the class (entangled-shape rule, [maintainer] refactor(plugin-auth): split src/auth-manager.ts (10,338 lines, +1,867 in 30 days) along its AuthManager seams, one move-only slice per PR #25224).

Order and pace

Because of the churn, each slice should take the coldest section the seam table finds, as [maintainer] refactor(metadata-protocol): split src/protocol.ts (33,065 lines, +10,169 in 30 days) along its ObjectStackProtocolImplementation seams, one move-only slice per PR #25467 did with batch and *Many.
A seat may hold this card behind sql-driver.ts's first slices when the file's in-flight PRs make a move-only slice collide.

Not in this card

Any behaviour change, and any design change to the shared members beyond giving them an explicit host.
driver-sql/src/sql-driver.ts, filed beside this one as its own card.

Dedupe

MCP search_issues on this repository, closed issues included: "split objectql engine.ts or sql-driver.ts move-only cut ruling 24586". There were 113 hits. None is a split card for objectql/src/engine.ts. The related ones are #24586 (the ruling), #25467 (protocol.ts), and #25222, #25223 and #25224 (the services files). New.

Generated by Claude Code

Claim: PM loop round 1
Session: session_01JVTd3PprStGa26aez5omLB
Account: sunhaolin (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-25522-objectql-engine-cut-1
Worktree: objectstack-issue-25522
Domain: domain:server
Seat: domain:server#2
File surface: packages/objectql/src/** (engine.ts and the new seam module of this slice), scripts/adr-anchors/** entries for moved bodies, the path readers of engine.ts under scripts/ and content/ that the slice must migrate, one changeset. ⛔ No docs/adr/** (governed) (stop on breach; explain in the report)
Container & model: M (one move-only slice plus the seam table), mode:subagent, model: TIER_DEFAULT (dispatch-gates --tier at 8801f2bb3: "no path-derived mandate … the tier stays the PM's per-card judgment call"), effort: high
Clause-②: no
Metadata-protocol: none
Responsibility: n/a — not a defect card
Thread-read: none
Serial constraints cleared: none. No open PR touches packages/objectql/src/engine.ts (13 open PRs read this fire); last touch on main is 77a77e4 (#25396).
Premise: re-read on origin/main 8801f2bb3: engine.ts is 23,903 lines, as the card measured; no split has started.

Generated by Claude Code

refactor(objectql): move the validation-message i18n helpers out of e…

os-dev-report
{
"issue": 25522,
"status": "done",
"branch": "claude/issue-25522-objectql-engine-cut-1",
"pr": "#25615",
"session": "session_01JVTd3PprStGa26aez5omLB — the dispatching seat's session (mode:subagent), on the Claude-Session: trailer of commit 328b5b1",
"premise_still_valid": true,
"summary": "Seam table measured fresh at origin/main e649c2d (engine.ts 23,903 lines; 8,253 from the last 30 days; 124 commits in 30 days, 33 in 7; class ObjectQL 317 members, 255 methods, 147 private / 108 public; the same 8 inferred-return public methods the card lists). It is in the PR body as 41 contiguous sections, each with its 30-day lines. Cut 1 is the coldest area that meets every series limit and has no path reader outside objectql: the validation-message i18n block (122 lines, 0 from the last 30 days). The private bodies of ObjectQL#validationMessageContext, #negotiatedMessageLocale and #objectDisplayLabel moved unedited to packages/objectql/src/engine-validation-messages.ts through ValidationMessageHost: the i18nService getter plus a late-bound negotiatedMessageLocale closure; objectDisplayLabel reads no member and is called directly. setI18nService stays: it is a two-line field setter. The new module imports ObjectQL as a type only, so there is no runtime cycle. Two spec/system imports with no other reader in engine.ts (objectLabelKey, resolveBundleLocale) left engine.ts. All six published declaration files are byte-identical, the export lists are identical, and the draft PR is Part of #25522 with skip-changeset. Roll-up summaries (0 lines in 30 days) was set aside for cut 1: buildSummaryIndex is the liveness ledger's evidence by path (packages/spec/liveness/field.json, summaryOperations), and recomputeSummaries has a this.update( write and passes this to aggregateSummaryValue. Proposed cut 2: Datasource routing (isSystemLedgerObject, getDriver, resolveDatasourceBinding, resolveEffectiveDatasource, resolveMappedDatasource, resolveDatasourceFromMapping, matchPattern; 343 lines, 1 from the last 30 days; setDatasourceMapping and the two public statics stay; the new module gets its own anchor entry for the ADR-0057 routing sentence). Measure getDriver's spies and doubles before choosing it. Draft only, per the release terms: no ready flip, no auto-merge, no queue, no merge.",
"tests": "At head 328b5b1. (1) pnpm --filter @objectstack/objectql exec vitest run --project local --maxWorkers=2 under os-verify-lock: VERDICT command-exit 0; JSON report: 466 files, 1,899 suites, 8,700 tests, 8,700 passed, 0 failed, 0 skipped. That includes engine-validation-locale.test.ts, engine-validation-locale-negotiation.test.ts and engine-delete-restricted-locale.test.ts, which pin these helpers. (2) pnpm --filter @objectstack/objectql typecheck (tsc --noEmit plus check:test-typecheck) under the lock: VERDICT command-exit 0, 'check:test-typecheck: OK — ... 39 file(s) / 233 error(s) / 64 pinned signature(s) held in test-typecheck-debt.json'. (3) Move proof by AST token (recursive getChildren leaves): validationMessageContext 104, negotiatedMessageLocale 134, objectDisplayLabel 133 tokens, all SAME, every body line the base line less 2 leading spaces. Negative control: one changed token in negotiatedMessageLocale reports DIFFER. Class side: 317 members at base and head, the 314 that did not move byte-identical, the 3 stubs keep docblock and signature byte-identical. Top level: 232 of 233 statements byte-identical (the other is the spec/system import less two names) plus one new import. (4) .d.ts bar, ruling 6348891160: index.d.ts 9ca22071717e3f761198115fa54a143df4151a52950d826a10677420dcad16ae, index.d.mts 69dd1cfe3875024f434b2d7defce771ebc8c7e48a05be7c946fa6ecdbd912b40, core.d.ts d27b72bde2eec9a47c7dd3c104976321533cf397452ca9f84491a455ec28e73b, core.d.mts 5a0cca505296b678b31c742c285e028fe434f5467bb8d27e052e470ce0e09652, util-ByiLm3B6.d.ts = .d.mts 1742a9cb029654d1ca4e8e392e98001eec85963987f516b1b76f672ecd08e5a9. Each is raw = normalised and base = head, raw diff 0 lines, using the ruling's normaliser verbatim. Note: this package prints single-quoted literals (23 single-quoted literal unions, 0 double-quoted), so that normaliser is a no-op here. Positive control: line 7261 private objectDisplayLabel; made public gives raw = normalised 270f913f734273e380af6ab964e8b15206aa7c511892dea58cc8f6fd9adf5ef7, not equal to 1742a9cb. Export lists from the built bundles: '.' 182 ESM = 182 CJS, './core' 54 = 54, identical at base and head (sorted-list sha256 881616cc... and 393e8069...). (5) One-off mutation, not a kept test (move-only, so no ablation is owed): scripts/ablation-replace.mjs replaced the moved 'return resolveBundleLocale(offered, requested) ?? requested;' with 'return requested;' (anchor 1 -> 0, blob 5d9194fa75bb -> 88406d6408d5). engine-validation-locale-negotiation.test.ts went to 3 failed / 3 passed. Restore: blob == HEAD 5d9194fa75bb and git diff HEAD empty. vitest resolves the relative import from src/, so there is no dist/ leg. test:repo was not run locally (CI's Test Core runs it).",
"gates": "Local, on the real diff (2 files, +135/-39): node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derived 56 families at 328b5b1 (stderr first line names the repo and the commit). All 56 were run at head, each exit code captured before any pipe. 53 exit 0. Three are NOT MEASURED: node scripts/check-engine-split-ratio.mjs --days 90 exit 2 (CANNOT COMPUTE: the shallow clone's history starts 2026-09-15, inside the 90-day window); node scripts/check-plugin-teardown-shape.mjs --self-test exit 3 (cannot read pinned positive-control commit 621a487 in this shallow clone); pnpm check:dual-build-cjs-loads exit 3 (PREREQUISITE NOT MET: no dist/ for 70 packages; the export lists in tests (4) answer its question for objectql). Reconciliation: dispatch-gates --ran with ':: exit N' lines prints '56 derived famil(ies) accounted for — 54 run, 2 NOT-MEASURED (2 DERIVED from a recorded exit 3)'. It counts the exit-2 refusal as run; by that script's own header it is CANNOT COMPUTE, so this report counts it NOT MEASURED. Outside the derivation, all exit 0: eslint --no-inline-config packages/objectql; eslint on both files; node scripts/check-adr-anchors.mjs (base and head: 84 anchored files, 1,883 decision-letter citations, 52,465 citations across 6,583 files, the same at both ends); node scripts/check-adr-symbol-anchors.mjs (2,384 anchors resolve); node scripts/check-nul-bytes.mjs; node scripts/check-platform-checklist.mjs --changed-from origin/main ('1 anchored file(s) in the change set, 14 anchor(s) judged, 0 lost'). CI on PR #25615 head 328b5b1 at report time: Check Changeset success; No other open PR may claim the same issue success; Type Check (workspace, debt ledger, source gates, consumer gates), the Lint lanes (PM dispatch-gates, CI release & engine gates, ESLint, repo & tooling gates, docs & code guards, the scoped lanes), Governed Surface Queue Guard, Part-of PR must not also close its card, Check PR Size and the rest: in_progress. Opt-in lanes skipped. Not waited on, per the contract.",
"dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm (Clause-②: no)",
"mcp_calls": "0 — no GitHub MCP tool called",
"api_writes": "2 relay strokes, both through scripts/pm/fleet-write/dispatch.mjs as POST /repos/objectstack-ai/objectstack/dispatches, executed by the relay as objectstack-fleet[bot]. (1) pr_create: POST /repos/objectstack-ai/objectstack/pulls (draft) + POST /issues/25615/assignees [sunhaolin] + POST /issues/25615/labels [skip-changeset]; run 40082586078 success; read-back 18,999 bytes sent = stored, identical; assignee and label listed. (2) This report: POST /repos//issues/25522/comments via scripts/pm/post-stamped.mjs. Not REST: git push to the branch, twice (the empty-branch probe, then commit 328b5b1).",
"open_questions": [
{
"question": "Later cuts of this file: @objectstack/objectql's .d.ts prints string literals in SINGLE quotes, so ruling 6348891160's verbatim normaliser (double quotes only) cannot absorb a union-member-order swap here. This cut is raw byte-identical, so nothing depended on it. A later cut that moves a body the DTS program checks before the rest of engine.ts could reorder a union. That shows as a raw diff the normaliser leaves in place, which under the ruling's item 1 is a stop. Does the ruling seat want to decide this before cut 2 is dispatched?",
"options": [
"A — The ruling seat states that item 1(b)'s normalisation covers single-quoted literal unions too: the same single class (member order within one literal union), spelled with the quote this printer uses, nothing else normalised. Real business need: the maintainer's split order for a file with 124 commits in 30 days. Long-term: the bar stays on the contract (declarations, visibility, exports), not on printer quote style. AI-error prevention: unchanged, because 1(a), 1(d) and the positive control still catch every change of kind. No sprawl: a ruling sentence, no script, no gate. Effort: one ruling comment.",
"B — Keep the normaliser verbatim. Any later cut whose raw diff carries a union-order swap stops and returns needs_decision, and the ruling seat decides then. Real business need: the same. Long-term: an equivalent outcome, one round later. AI-error prevention: the same. No sprawl: none added. Cost: one stopped cut and a ruling lap mid-series, at a cut whose timing nobody chooses."
],
"recommendation": "A, decided by the ruling seat, not by a dev: it is the class ruling A already judged meaningless to every consumer (the quote character is a printer setting), and deciding it before cut 2 costs one sentence instead of a stopped cut. ⛔ Not acted on here: this dev did not widen the normaliser and adds no script (ruling item 3, card ask 5)."
}
],
"out_of_scope_findings": [
"carrier: the ruling seat, through open_questions[0] above · the single-quote blind spot of the ruling's normaliser on @objectstack/objectql · noted, not filed (not a defect, contract violation or authoring trap)",
"carrier: none (承接者:无) · cut-2 pricing note: Roll-up summaries is cold (0 lines in 30 days), but buildSummaryIndex is the liveness evidence pointer for field.summaryOperations (packages/spec/liveness/field.json: evidence packages/objectql/src/engine.ts, key-mention), so that cut owes a packages/spec ledger edit, and recomputeSummaries stays (a this.update( write, and this passed to aggregateSummaryValue) · noted in the PR body, not filed"
],
"line_budget": "174 changed lines (+135 / -39) over 2 files, under the 3,000 human-merge line; engine.ts 23,903 -> 23,881; engine-validation-messages.ts 118",
"files_changed": [
"packages/objectql/src/engine-validation-messages.ts (added)",
"packages/objectql/src/engine.ts (modified)"
],
"deviations": [
"Commit trailers: the model-free pair (Claude-Session + Co-authored-by: Claude) per AGENTS.md and os-dev.md, in place of the harness reminder's model-named Co-Authored-By line (the pre-push hook refuses a model identifier). PR body footer: AGENTS.md's session-URL form, in place of the harness's line.",
"A2: no scripts/adr-anchors entry added for the new module. The engine.ts entry's one paragraph is the transaction seam and describes no moved body, and the moved bodies cite no ADR (the ADR-0053 Phase 2 cite is in a docblock that stays on the stub). Card ask 3 owes an entry only when a paragraph describes a moved body.",
"A3, narrowed and declared: no direct dependant was type-checked locally. Every moved member is private (reachable from no dependant), and all six published declaration files are byte-identical, so the dependants' input bytes are unchanged. CI's TypeScript Type Check lanes cover them.",
"No baseline objectql test run (head only). The base-vs-head per-test comparison PR #25494 carried is not repeated; the head suite is 8,700/8,700.",
"History: a sibling agent's git fetch --shallow-since=2026-09-15 held .git/shallow.lock, so this run waited on that PID (tail --pid) and read 30-day blame from the deepened history. It did not remove the lock or deepen further, so the 90-day split-ratio gate stays NOT MEASURED.",
"Created a fetch ref refs/issue-25522/main in the shared ref namespace to re-read main without moving origin/main; deleted at cleanup."
]
}

Generated by Claude Code

ACCEPT → PR #25615 comment 6382672482 (cut 1, head 328b5b194e; lands once CI is green on that head) · domain:server seat 2 · session_01JVTd3PprStGa26aez5omLB

Generated by Claude Code

Release: domain:server seat 2 · sunhaolin · session session_01JVTd3PprStGa26aez5omLB · 2026-10-24T10:42Z · cause: cut 1 landed (Part of #25522) · destination: pm:queue, unassigned, for cut 2

Landed: PR #25615 → `481ac6211` (#25522 cut 1)

reading	result
PR #25615 object	`merged: true`, merge commit `481ac6211641ccb7292508ed423f02e432322e4a`, merged 2026-10-24T10:42:01Z through the merge queue
`origin/main`	`481ac6211` is an ancestor of `origin/main` (`481ac62116`), single parent (squash)
queue branch	no `gh-readonly-queue/main/*` head for PR 25615 remains

Landed: the validation-message i18n block (validationMessageContext, negotiatedMessageLocale, objectDisplayLabel) lives in packages/objectql/src/engine-validation-messages.ts. engine.ts is 23,881 lines.
Remains: every other area of the 41-section seam table in PR refactor(objectql): move the validation-message i18n helpers out of engine.ts into engine-validation-messages.ts (#25522 cut 1) #25615's body. The dev proposes Datasource routing for cut 2 (measure getDriver's spies and doubles first; the new module owes its own anchor entry for the ADR-0057 routing sentence). Roll-up summaries waits: buildSummaryIndex is the liveness evidence path in packages/spec/liveness/field.json, so that cut is spec-lane-coupled.
State: pm:dispatched → pm:queue, assignee cleared, in this act. The series card stays open. This seat re-claims cut 2 when a dispatch slot frees.

Generated by Claude Code

os-dev-report
{
"issue": 25522,
"status": "done",
"branch": "claude/project-thread-axdp1m",
"pr": "#26932",
"session": "session_01FEkQ5n1LHPdPmHoXe1BGDa",
"premise_still_valid": true,
"summary": "Cut 18 of the engine.ts split: moved count and rejectCredentialAggregation (the count/aggregate area, priced -71 on 1708115, the highest-paying section above -30 once delete-cascade -248 is excluded by the order) into packages/objectql/src/engine-count-aggregate.ts under a CountAggregateHost, move-only with host stubs left in the class. aggregate itself stays in engine.ts because packages/spec/liveness/query.json names engine.ts#aggregate. The 7 helpers shared with staying bodies (ENGINE_WHERE_SLOTS, ENGINE_COUNT_OPTION_KEYS, rejectUnknownEngineOptions, foldEngineOptionAliases, lowerWhereFilterArray, declaredDatetimeLowering, mergeReadContext) are exported in place and value-imported back, no second copy. engine.ts 21,835 -> 21,762 lines (net -73, +33/-106). Runner-up: Query lowering -40.",
"tests": "objectql local suite: 475 files / 8,849 passed. Repo projects: core 6 files/83 tests, objectql 1/5, rest 280 passed + 18 skipped, runtime 948, spec 80 files/1,295, types 1/11, all green. objectql + core typecheck exit 0; eslint packages/objectql exit 0; full build 116/116. Move proof: AST compare count SAME (296 tokens/58 lines), rejectCredentialAggregation SAME (296/38); class 326/326 members, 324 identical + 2 stubs; top level 227/236 identical. Late-binding probe (zz-probe-cut18.mjs): 35 result lines identical base vs head. .d.ts: all 6 files raw byte-identical (chunk util-BaJGwxSm); positive control (drop private on a stub) moves chunk to util-CKT4sa10, normalised 32d9098f != c43cb8b1, surface 4c6ca612 != b264bade; restore leg blob c16b33857e02 rebuilt equal. Each in-place export alone and all 7 together leave .d.ts byte-identical. Runtime export lists equal (182/54); bundle regions +1 only (module import placed after ./secret-fields.js so region order is unchanged; placement at the old StandardErrorCode line reordered secret-fields.ts and was rejected); externals equal (113/96 imports, 148/130 requires). No tested behaviour changed, so no ablation. Evidence in pm-orders/server/evidence/25522/cut18-*.",
"gates": "dispatch-gates.mjs --commands: 48/48 exit 0 under os-verify-lock (slot c18-axdp1m); --ran reconciliation 0 NOT-MEASURED (cut18-ran-final.txt). Hand gates exit 0: check-tenant-audit, check:adr-symbol-anchors, platform-checklist, check-commit-card-trailers, check:adr-anchors, system-context-census. check:adr-anchors decision letters: base 1,897, head 1,896 ([ADR-0139 D2] moved with count into the unanchored module; disclosed in the PR body, same as cut 15). Durability --list base vs head identical with line numbers masked. CI check-runs on #26932 are for the seat to read; not re-run here.",
"dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
"mcp_calls": "10 — reads only: issue_read x5, pull_request_read x3, list_pull_requests x2; no write tool",
"api_writes": "3 — git push x1 (claude/project-thread-axdp1m); relay pr_create x1 via scripts/pm/fleet-write/dispatch.mjs (run 40590873567, #26932, body read back identical); this report comment x1 via scripts/pm/post-stamped.mjs",
"open_questions": [],
"out_of_scope_findings": [
"carrier: this PR body · noted, not filed — ADR decision-letter count drops by 1 (1,897 -> 1,896) because [ADR-0139 D2] now sits in engine-count-aggregate.ts, which check:adr-anchors does not anchor",
"carrier: next cut · noted, not filed — delete-cascade (-248) is the top remaining section once draft #26930 lands; runner-up for this cut was Query lowering (-40)"
]
}

Generated by Claude Code

Landed (cut 18)

2026-10-27T22:42Z · domain:server seat 2 · os-project-manager · session session_0186vfjHHMsRXC9EUT9hTFMo

PR #26932 merged at 2026-10-27T22:41:50Z. The merge queue landed it as beb09fb2f6 on main (Part of #25522; ACCEPT 6455030043; Clause-②: no, so no contract review was owed).

Reading	Probe	Result
Queue membership	`git ls-remote origin 'refs/heads/gh-readonly-queue/main/pr-26932-*'`	0 refs (left the queue)
Ancestry	`git merge-base --is-ancestor beb09fb2f6 origin/main`	holds
`origin/main` by content	`packages/objectql/src/engine-count-aggregate.ts`	carries `CountAggregateHost`
`engine.ts` size	`git show origin/main:packages/objectql/src/engine.ts \| wc -l`	21762 (net −73)

Landed: count and its credential guard rejectCredentialAggregation moved out of engine.ts into engine-count-aggregate.ts behind CountAggregateHost. Seven helpers are exported in place, not copied. The published .d.ts files are byte-identical.
Stays in engine.ts: aggregate, because packages/spec/liveness/query.json names engine.ts#aggregate as a producer.
Next: cut 19 is priced off a table re-measured on this main, excluding the delete section while fix(objectql)!: a predicate delete honours deleteBehavior for every matched row: restrict refuses, master_detail cascades, set_null clears (#26871) #26930 (objectql: a predicate delete (multi: true with a filter) skips the referential walk, so a flow's bulk delete_record bypasses restrict, master_detail cascade and set_null #26871) is open there. It will be dispatched when the lane, now capped at 4, has a free slot.
State: stays pm:dispatched, assigned to os-project-manager.

Generated by Claude Code

Release: series claim (cut 18 claim 6453372157, the last one) · domain:server#2 · os-project-manager · session_0186vfjHHMsRXC9EUT9hTFMo · 2026-10-28T11:52Z

The maintainer, verbatim, 2026-10-28T11:52Z: 「已认领、还没开工的应该放回队列」.

Cut 18 landed (PR refactor(objectql): move count and rejectCredentialAggregation out of engine.ts into engine-count-aggregate.ts (#25522 cut 18) #26932, landing 6455397586). Since then no cut has been claimed or started, and no open PR names this card. So the series card sat at pm:dispatched with nothing in development.
Back to the queue: pm:dispatched → pm:queue, and the assignee is cleared. The next cut gets a fresh claim and order from whichever seat picks the series up. The cut notes are in the landing records on this card.

Generated by Claude Code

domain:server#1 · os-sales · session_01L6bhgXcN9HaZYSEUdJWEjw

Claim: PM loop round 1 · card thread · cut 19 (delete cascade and reference walking) · 2026-10-29T04:10Z
Session: session_01L6bhgXcN9HaZYSEUdJWEjw (the seat); the card thread is session_01L6bhgXPvYEKGGY6sms31q2
Account: os-sales (the seat's linked user; the card's assignee)
Branch: claude/project-thread-f9p63p (the card thread's harness-pinned branch; it names no issue, so this claim is the identity record)
Worktree: the card thread's own container checkout
Domain: domain:server
Seat: domain:server#1
Ruling-ref: #24586 ruling A (6270097840); card body (move-only, one area per PR, one named host); .d.ts bar 6348891160 + 6441032442 letter B; host setter 6430872749; shared helper 6432376812
File surface: packages/objectql/src/engine.ts (the delete-cascade section, planCascadeAtomicity … organizationOwnedRowCounts, minus its stays), one new packages/objectql/src/engine-*.ts, its path readers and tests, scripts/adr-anchors/** entries for moved ADR paragraphs. ⛔ No packages/spec/**, ⛔ no docs/adr/** (stop on breach; explain in the report)
Container & model: M, mode:thread, model: default tier (dispatch-gates.mjs --tier: "no path-derived mandate"), effort: high
Clause-②: no
Metadata-protocol: none
Responsibility: n/a — not a defect card
Thread-read: 6468285951
Premise: corrected (read-only premise reading at origin/main c4c515ad3: landing 6455397586 excluded the delete section while #26930 was open; #26930 and #27073 have merged, so the section is free; its −248 price is stale and is re-priced at claim; engine.ts is 22,110 lines)
Serial constraints cleared: draft #27339 (os-bill, #27186) touches engine.ts imports, insert, update and aggregate, not this section; #27277 (marchtian, pm:epic) declares packages/objectql/src/** with no PR yet, so the dev reads branch claude/project-thread-mlrxgx at start; #27165 (queued) edits find/findOne, disjoint. Lane card threads at claim: 2 of the maintainer's 5.

Generated by Claude Code

domain:server#1 · os-sales · session_01L6bhgXPvYEKGGY6sms31q2

os-dev-report

{
  "issue": 25522,
  "status": "done",
  "branch": "claude/project-thread-f9p63p",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/27391",
  "session": "session_01L6bhgXPvYEKGGY6sms31q2",
  "premise_still_valid": true,
  "summary": "Cut 19 of the engine.ts split: five bodies of the Delete cascade and reference walking section (collectCascadeDeleteSet, refuseRestrictsInsideOwnCascade, deleteRestrictedError, referenceProbeFilter, recordReferenceCheckElevation) move unedited into packages/objectql/src/engine-delete-cascade.ts behind DeleteCascadeHost, stubs left in the class; cascadeReachOf (read only by refuseRestrictsInsideOwnCascade) moves with it; cascadeMemberKey, cascadeSetAdd, cascadeSetHas, declaredMultiValued and CascadeDeleteSet are exported in place and imported back, no copy. engine.ts 22,110 -> 21,874 (net -236, +42/-278), module 429 lines, PR 753 changed lines over 4 files. The order hypothesis (~18 bodies, ~1,069 lines) is falsified by the series limit on private statics: 10 methods of the section read a private static (planCascadeAtomicity, cascadeDeleteRelations, dependentCountIsDisclosable, probeReferencingRows, probeReferencingRowsOfSet, walkReferencingRelations 318 lines, countReferencingRowsOutside, refuseOrganizationDeleteWhileOwned, organizationOwnedRowCounts) and the 7 statics stay; cascadeRelationBehavior stays because federated-injected-column-readers.test.ts checks the organizationOwnedRowCounts via row by reading the helper in the SAME file; 6 small bodies pay 3 lines or less and stay. Readers re-keyed: system-context.mdx register row 29 (recordReferenceCheckElevation) and one platform-checklist citation (records-forms.json engine.ts#required). One PR, no split needed.",
  "tests": "Under os-verify-lock: full turbo build 115/115 exit 0; objectql typecheck (tsc + check:test-typecheck) and core typecheck exit 0; objectql vitest --project local Test Files 479 passed (479), Tests 8942 passed (8942); runtime real-driver cascade integration tests 3 files 15 passed 1 skipped, plugin-security delete-reference-cleanup-system-identity 18 passed; eslint --no-inline-config packages/objectql exit 0. test:repo not run locally (surface byte-identical), CI Test Core. Move proof: AST leaf-token compare 5/5 SAME (97/107/582/810/323 tokens), negative control DIFF; class 330/330 members, 325 byte-identical + 5 stubs; top level 229/238 identical, rest = 2 import-list edits, 5 in-place exports, cascadeReachOf moved byte-identical, class. .d.ts: all 6 files raw byte-identical base c4c515ad3c vs head (util-6U6fCMd3 both); positive control via ablation-replace (drop private on the deleteRestrictedError stub) renames chunk to util-Dd_AzMXV and changes every hash, restore blob == HEAD, rebuild equal on 6 d.ts + 4 bundles. dtsnorm.mjs was not available in this container (other project files); with 0 raw-diff lines nothing needed normalising. Export lists 182/54 ESM=CJS, name+fn.name hash equal base/head. Bundle regions +1 only (engine-delete-cascade.ts), external first-load order equal; +1 import statement (module own renderOperationMessage from spec/system). Late-binding probe zz-probe-cut19.mjs: 16 lines identical base/head (5 timestamped construction lines set aside), stderr empty. No tested behaviour changed, so no ablation.",
  "gates": "dispatch-gates --commands on head b3a0abd0f6 (merge base c4c515ad3): 80 families, each run with exit captured before any pipe, 80/80 exit 0; --ran with exit annotations: 80 derived, 80 run, 0 NOT-MEASURED, 0 UNRUN (a DERIVED zero). Prereqs: full build, git fetch --shallow-since=2026-07-22, fetch 621a4876. Hand gates exit 0: check-adr-anchors (decision letters base 1,900 / head 1,900; citations 60,466 across 8,527 / 8,528 files), check-system-context-census (154 sites), check-tenant-audit-census (245), check-adr-symbol-anchors (2,798), check-platform-checklist --changed-from origin/main and full, check-commit-card-trailers, check-nul-bytes, durability --list base==head masked (165 lines). main 7e200b4a68 merged before pr_create (ui/Studio only, none of this diff paths; change set unchanged). CI check-runs on #27391: in_progress at report time, for the seat to read.",
  "dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
  "mcp_calls": "8 — reads only: issue_read x4 (get x1, get_comments x3), pull_request_read x2 (get #26932, get #27391 read-back), list_pull_requests x2; no write tool",
  "api_writes": "6 — git push x4 (branch create, cut commit, checklist commit, main merge; not REST writes); relay pr_create x1 via scripts/pm/fleet-write/dispatch.mjs (run 40820376505: POST pulls draft + assignees os-sales + labels skip-changeset; body 23133 bytes read back identical); relay pr_request_reviewers x1 (run 40820426781, reviewers os-sales); this os-dev-report via scripts/pm/post-stamped.mjs (POST issues/25522/comments)",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: seat 1 · noted, not filed — PR body inconsistency to fix by relay if wanted: the 维护者速读 line says 区内 24 个成员 while the English table counts 27 (26 methods and REFERENCE_PROBE_BATCH); 27 is right; the dev does not PATCH bodies",
    "carrier: next cut · noted, not filed — the rest of this section (walkReferencingRelations 318, probeReferencingRows 100, planCascadeAtomicity 81, dependentCountIsDisclosable 49, cascadeDeleteRelations 45 ...) is locked by the private-static series limit; only a ruling on passing statics through the host would free it. cascadeRelationBehavior (60) is locked by the reader test same-file via rule",
    "carrier: seat 1 · noted, not filed — the order named branch claude/project-thread-mlrxgx (#27277) to read; it does not exist on the remote (git ls-remote 0 refs)"
  ]
}

domain:server#1 · os-sales · session_01L6bhgXcN9HaZYSEUdJWEjw

Release: cut 19 delivered · back to `pm:queue` · `domain:server#1` · 2026-10-29T05:43Z

⛔ Not a claim. This card is Part of a series; the seat's claim on cut 19 ends with its merge.

Delivered: PR refactor(objectql): move five delete-cascade bodies out of engine.ts into engine-delete-cascade.ts (#25522 cut 19) #27391, merged 2026-10-29T05:42Z (5c0e4fd0c6). Five bodies of the "Delete cascade and reference walking" section moved into packages/objectql/src/engine-delete-cascade.ts: collectCascadeDeleteSet, refuseRestrictsInsideOwnCascade, deleteRestrictedError, referenceProbeFilter and recordReferenceCheckElevation. cascadeReachOf moved with them. engine.ts went from 22,110 to 21,874 lines. The seat record is the ACCEPT 6488660939.
Remaining in this section: 17 members are kept by the private-static series limit, the largest being walkReferencingRelations (318 body lines). cascadeRelationBehavior is kept by the reader test's same-file via rule. None of them can move under the current series limits, so the section is spent until a limit changes.
Next claimant: re-price the next section on fresh origin/main before quoting a size. No branch carries unmerged work for this card (Resume-from: none).

Generated by Claude Code

Navigation Menu

Repository navigation

[maintainer] refactor(objectql): split `src/engine.ts` (23,903 lines, +7,501 in 30 days) along its `ObjectQL` seams, one move-only slice per PR #25522

Measured (on `origin/main` `842a0aef9d`, 2026-10-23)

The ask (from #24586's ruling and #23050's shape, as on #25467)

Series limits already measured (carry them forward)

Order and pace

Not in this card

Dedupe

Landed: PR #25615 → `481ac6211` (#25522 cut 1)

110 remaining items

Landed (cut 18)

Release: cut 19 delivered · back to `pm:queue` · `domain:server#1` · 2026-10-29T05:43Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

[maintainer] refactor(objectql): split src/engine.ts (23,903 lines, +7,501 in 30 days) along its ObjectQL seams, one move-only slice per PR #25522

Description

Measured (on origin/main 842a0aef9d, 2026-10-23)

The ask (from #24586's ruling and #23050's shape, as on #25467)

Series limits already measured (carry them forward)

Order and pace

Not in this card

Dedupe

Activity

objectstack-fleet commented on Oct 24, 2026

objectstack-fleet commented on Oct 24, 2026

objectstack-fleet commented on Oct 24, 2026

objectstack-fleet commented on Oct 24, 2026

Landed: PR #25615 → 481ac6211 (#25522 cut 1)

110 remaining items

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

Landed (cut 18)

objectstack-fleet commented on Oct 28, 2026

objectstack-fleet commented on Oct 29, 2026

objectstack-fleet commented on Oct 29, 2026

objectstack-fleet commented on Oct 29, 2026

Release: cut 19 delivered · back to pm:queue · domain:server#1 · 2026-10-29T05:43Z

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

[maintainer] refactor(objectql): split `src/engine.ts` (23,903 lines, +7,501 in 30 days) along its `ObjectQL` seams, one move-only slice per PR #25522

Measured (on `origin/main` `842a0aef9d`, 2026-10-23)

Landed: PR #25615 → `481ac6211` (#25522 cut 1)

Release: cut 19 delivered · back to `pm:queue` · `domain:server#1` · 2026-10-29T05:43Z