A review saved with decision Approved leaves 结论时间 (decided_at) empty — the deviation stamps it, the review does not

Filing gate: ① user-visible defect, reproduced in the browser on @objectstack 18.2.0 (hotclm PR #155 branch).

Symptom. Legal counsel creates a review on 审查与偏离 with 结论 = 通过 and saves. The row reads decision: approved, decided_at: null; 结论时间 stays empty on the page.

Where. clm_review.decided_at (src/objects/review.object.ts) has no stamping hook and no description. clm_deviation stamps its own decided_at / decided_by when it is decided (src/objects/contract.hook.ts, the deviation guard); clm_signature stamps completed_at since #145. DESIGN.md §03 lists decided_at on the review with no "manual" note.

Expected. Saving a review whose decision leaves pending stamps decided_at (unless the user typed one), the way the deviation and the signing round do.

Business reason. The review page and any report on review turnaround read when legal decided; with the field empty, a decided review carries no decision time at all and the lawyer has to type it by hand.

Reader. hotclm execution seat (PM seat #36).

Duplicate search. hotclm issues, all states, "decided_at review": 0 hits; #138/#145 (signature completed_at) is the sibling, closed.

Triage — 2026-10-27

Filed from the 18.2.0 browser pass (hotclm PR #155, landed as 21f2319). Defect, user-visible, reproduced: accepted.

Route: app-side, in the deviation and signing-round pattern. Stamp clm_review.decided_at when decision leaves pending, unless the user typed one, and give the field a description. DESIGN.md §03 needs no change.

State: → pm:queue + bug.

Generated by Claude Code

Claim: PM loop round 17
Ruling-ref: none
Session: session_017nzWFRUo2U7evMkPdVniAn
Account: zhuangjianguo (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-159-review-decided-at
Worktree: hotclm-issue-159
Domain: repo:hotclm (single-lane sister repo — no domain:*)
Seat: repo:hotclm#1
File surface: src/objects/contract.hook.ts (review stamping, in the deviation pattern) · src/objects/review.object.ts (description of decided_at) · translations for that description if the i18n gate asks — stop on breach; explain in the report
Container & model: S, mode:subagent, model: opus (TIER_DEFAULT; hand-derived for a sister repo), effort: medium
Clause-②: no
Metadata-protocol: none
Responsibility: this app's review record (when legal decided) | none | every lawyer deciding a review
Thread-read: 6437440077
Premise: re-measured by the dev on origin/main 21f2319 (@objectstack 18.2.0) with README personas first
Serial constraints cleared: Triaged 2026-10-27; round 17 dispatches #134, #127, #64 and #159 together on disjoint files.

Generated by Claude Code

os-dev-report

{
  "issue": 159,
  "status": "done",
  "branch": "claude/issue-159-review-decided-at",
  "pr": "https://github.com/objectstack-ai/hotclm/pull/163",
  "session": "session_017nzWFRUo2U7evMkPdVniAn",
  "premise_still_valid": true,
  "summary": "Premise re-measured on origin/main 21f2319 (18.2.0) as Legal Counsel 1 in zh-CN Chromium: a review created with 结论 = 通过 stored decided_at null, and the 结论时间 column was absent from the contract's review grid. New hook review_decision_stamp in src/objects/contract.hook.ts (clm_review, beforeInsert + beforeUpdate, priority 200), in the deviation pattern: when a write gives the review a decision (approved / changes_requested / rejected) different from its previous one and no decided_at is present in the write or on the row, it stamps now. It never overwrites or clears a value. decided_at got a description in review.object.ts, and the i18n gate asked for its zh-CN help, which was added along with the en line. No decided_by exists on clm_review or in DESIGN.md §03 (the reviewer decides), so none was added. Back to pending: allowed (§03 gives the review decision no transition table) and the stamp is left in place, matching the deviation's fill-only, never-erase stamp. A re-decision after that keeps the first time; see open_questions. One step beyond the card's wording: a decided-to-decided change on a row with an empty decided_at (the rows this defect left behind) is stamped too.",
  "tests": "No unit-test harness in this app; behaviour measured on 18.2.0, pnpm demo:zh --port 3144, fresh DB before (21f2319) and after (4041dd1). Persona: admin create-user Legal Counsel 1 -> 200 (mustChangePassword true), sys_user_position clm_legal_counsel -> 201, persona change-password -> 200, security/explain clm_review create allowed. BROWSER before: DPA-2026-0005 › 审查与偏离 › 新建, 结论=通过, POST clm_review 201, row {decision: approved, decided_at: null}; the record page shows no 结论时间. BROWSER after: same steps, decided_at 2026-10-27T09:04:05.678Z; the record page reads 结论时间 2026/10/27 上午9:04, and the grid shows the column. Typed 2026/10/01 09:30 -> kept as 2026-10-01T09:30:00.000Z. Pending (待定) -> decided_at null. Edit pending -> 通过 -> stamped 09:05:42.961Z, and the page shows it. Edit back to 待定 -> stamp left in place. API before/after as the persona: POST approved null -> stamped; POST approved + typed kept/kept; POST pending and POST with no decision null/null; PATCH pending->changes_requested null -> stamped; PATCH ->pending null -> left in place; PATCH pending->rejected + typed kept/kept; PATCH comments only unchanged/unchanged; after only: re-decide after the back-edge keeps the first time; approved->rejected on a cleared row -> stamped. Seed: 67 reviews (60 seeded + 7 measured); exactly 4 have decided_at on or after 2026-10-27, all from the measurement; 0 decided reviews are empty. The seed summary is identical before and after (inserted 830, errored 404: the README's documented unnamed personas). Browser console: the same 404 and 401 resource lines before and after. No ablation: there is no test file to pin; before/after is the proof.",
  "gates": "Final head 4041dd1, exit codes captured before any pipe: pnpm install EXIT=0; pnpm validate EXIT=0 (Validation passed); pnpm lint EXIT=0 (1 warning(s), 6 suggestion(s): the one warning is main's security-controlled-by-parent-owner-policy-without-opt-in, so no new warning); pnpm typecheck EXIT=0; pnpm lint:i18n-gate EXIT=0 (0 missing keys across 2 locale(s)); CLM_DEMO_SEED=1 pnpm validate EXIT=0. Mid-way, before the translation: lint:i18n-gate EXIT=1 on [zh-CN] objects.clm_review.fields.decided_at.help, and lint showed a second warning for the same key; both cleared. PR check-runs on 4041dd1: Validate in_progress at report time. No dispatch-gates derivation: this is a sister repo with no scripts/pm/dispatch-gates.mjs, and the gate list is AGENTS.md's.",
  "dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
  "mcp_calls": "0",
  "api_writes": "2 — fleet-write relay: pr_create objectstack-ai/hotclm (#163); comment on objectstack-ai/hotclm#159 (this report). git push of the branch is not a REST write.",
  "open_questions": [
    {
      "question": "When a review's decision goes back to pending, should decided_at stay (as shipped), be cleared, or should the back-edge be refused?",
      "options": [
        "A keep the stamp (shipped): matches the deviation's fill-only stamp and never erases a typed time, but a review decided again after pending keeps its first decision time",
        "B clear decided_at on the edge into pending, unless the same write sets one: a later re-decision gets a fresh time, but a typed time is erased and a hook writes null into a user-editable field",
        "C refuse decided -> pending, the deviation's 'decided once' rule: a new transition table on clm_review.decision that DESIGN.md §03 does not give, which is product semantics and a §03 change"
      ],
      "recommendation": "A, because it is the deviation's stamping behaviour, needs no §03 change, and the lawyer can correct the editable field. Revisit with B if review-turnaround reports show re-decided reviews."
    }
  ],
  "out_of_scope_findings": [
    "class: none (renderer observation) · reach: browser, contract › 审查与偏离 review grid: the configured decided_at column is omitted while every row's value is empty (before the fix), and appears once one row has a value · carrier: none — noted in PR Acceptance notes, not filed"
  ]
}

A decided review records when legal decided (#163)

Ruling-ref: none (triaged defect, comment 6437440077)

Landed — PR #163 · 2026-10-27

reading	value
PR #163	`MERGED` (squash on `main`), merged at its ACCEPT head `4041dd1`; `Fixes #159` closes this card
authority	maintainer, project thread, verbatim: 「你可以自己审核的就自己合并」

Release: a review saved with a decision now records 结论时间 automatically; a typed time is kept. State: pm:dispatched removed, assignee cleared.

Generated by Claude Code

Navigation Menu

Repository navigation

A review saved with decision Approved leaves 结论时间 (decided_at) empty — the deviation stamps it, the review does not #159

Triage — 2026-10-27

Landed — PR #163 · 2026-10-27

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

A review saved with decision Approved leaves 结论时间 (decided_at) empty — the deviation stamps it, the review does not #159

Description

Activity

objectstack-fleet commented on Oct 27, 2026

Triage — 2026-10-27

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

Landed — PR #163 · 2026-10-27

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions