spec(record-surface): a field-heavy create/edit flow on desktop derives a bounded modal, not full-screen; mobile keeps full (objectui#12521 letter B)

Filing gate: the spec half of a maintainer ruling — objectui#12521, batch #352 item 2, letter B (the ruling comment on that card). Filed by the director seat (summon #42, session_01AKVFxqmRLqktJspVXwPE9A). ⛔ Not a claim; the lane and grade are the ruling's; triage adds target: as it sees fit.

What the ruling binds

packages/spec/src/data/record-surface.ts deriveRecordFlowSurface: a task flow (create / edit / child-*) whose base surface is page (field-heavy, RECORD_SURFACE_PAGE_THRESHOLD = 12) derives on desktop a bounded modal, { container: 'overlay', surface: 'modal', size: 'auto' }, instead of size: 'full' (:150 on origin/main 9bb93706fb); a mobile posture keeps size: 'full'. record-surface.test.ts, the docblock and the docs move with it. The objectstack#2604 D1 note the docblock cites gets a dated amendment line naming the ruling, ⛔ not rewritten.

Premise the spec seat measures first

A field-heavy task flow WITH an embedded child or line-items table. If #2604 D1 chose full-screen for that case, the bounded modal applies only to forms without an embedded subtable, and this card's PR states that split in the derivation and its test; the fork is reported on objectui#12521.

Downstream

objectui AppContent.tsx:1200-1206 keys on size === 'full', so the console needs no code: the spec bump carries it. objectui#12521 is pm:blocked on this card; objectui PR #12546 (the bounded default when the host gives no size) lands after the spec is installable at objectui's pin.

Clause-②: no (no key moves; a published function's derived default changes, with a BREAKING line in the changeset for a consumer that relied on size: 'full' for heavy desktop task flows). Contract review at CONTRACT_REVIEW_TIER by face (packages/spec/src/**).

Dedupe: objectstack open titles read for record-surface, deriveRecordFlowSurface, full-screen and modal in this act: none.

First touch: lane and grade confirmed; `target:v18` added · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-15T23:04Z

Read in this act: the body in full, 0 comments, and the ruling on objectstack-ai/objectui#12521.

Lane: domain:spec, confirmed. The card lands in packages/spec/src/data/record-surface.ts with its test and docs.
Grade: p1, confirmed. This is protocol-first: a published spec derivation's default changes, and the changeset carries a BREAKING line.
target:v18 added. The class takes the in-development major, and the body leaves target: to triage.
State: pm:queue stands. console record form: a create/edit dialog with no authored layout opens at 95vw with one column of unequal field widths — a bounded dialog with a two-column field grid and one width per field kind objectui#12521 is pm:blocked on this card, with its body line in place.

⛔ Not a claim.

Claim: PM loop round 2 · 2026-10-15T23:10Z
Session: session_01HjR8RUwTPQ3TgbfKihmxp9 (the dispatching seat; the dev runs as a cloud session the seat post names)
Account: os-elon-musk (the seat's linked user as GET /user answers it; this card's assignee)
Branch: claude/issue-23734-heavy-flow-bounded-modal
Worktree: objectstack-issue-23734 (in the cloud session's container)
Domain: domain:spec
Seat: domain:spec#1 (seat post #6017)
File surface (at origin/main 3b46855fb or later; stop on a breach and explain it in the report):

packages/spec/src/data/record-surface.ts: deriveRecordFlowSurface's page arm (:150) and its docblock, with a dated amendment line on the objectstack#2604 D1 note it cites (⛔ the note is not rewritten).
packages/spec/src/data/record-surface.test.ts, the docs that state the derivation, the regenerated packages/spec artifacts check:generated names, and one .changeset/23734-*.md with a BREAKING line.
⛔ No objectui file: console record form: a create/edit dialog with no authored layout opens at 95vw with one column of unequal field widths — a bounded dialog with a two-column field grid and one width per field kind objectui#12521 and PR fix(plugin-form): a record dialog with no authored layout opens bounded with a two-column grid (objectui#12521) objectui#12546 carry that half.
Container & model: M, mode:cloud (the maintainer's word in this session's chat: 「batch 调到 5, 可以开云卡」), model: default tier (dispatch-gates --tier --repo objectstack-ai/objectstack: no path-derived mandate; the contract review owed before the queue is at CONTRACT_REVIEW_TIER by face, packages/spec/src/**)
Clause-②: no
Responsibility: n/a — not a defect card (a maintainer ruling's spec half)
Thread-read: 6204108463
Ruling-ref: console record form: a create/edit dialog with no authored layout opens at 95vw with one column of unequal field widths — a bounded dialog with a two-column field grid and one width per field kind objectui#12521 ruling 6203721639 (letter B, maintainer 「B (推荐)」).
Serial constraints cleared: the file lists of all 5 open PRs and the claims of all 16 open pm:dispatched cards were read in this act; none names record-surface.ts. Downstream: console record form: a create/edit dialog with no authored layout opens at 95vw with one column of unequal field widths — a bounded dialog with a two-column field grid and one width per field kind objectui#12521 (p1) is pm:blocked on this card.

Release note from the domain:ui seat 1 (session_01JZnppgHRp4sREcouhkQYDB), for the seat that lands this card: the same objectui card (objectui#12521) now also waits on #23740, FormSectionSchema.columns dropping default(1) (a maintainer ruling recorded on objectui#12521). It is a separate file (view.zod.ts) and a separate card, but shipping both in one spec release saves objectui a second bump. If both can land before the next next publish, please say so on #23740. No ask beyond that, and no change to this card's scope.

Generated by Claude Code

os-dev-report

{
  "issue": 23734,
  "status": "done",
  "branch": "claude/issue-23734-heavy-flow-bounded-modal",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/23747",
  "session": "session_01FaR7ExTuo2XxtaZH3xKJ1e",
  "premise_still_valid": true,
  "summary": "deriveRecordFlowSurface (packages/spec/src/data/record-surface.ts) now answers a field-heavy desktop task flow (create/edit/child-create/child-edit) with { container: 'overlay', surface: 'modal', size: 'auto' }; viewport 'mobile' keeps size 'full'. No new parameter, export or type member (Zone 2 hypotheses 1 and 2 held: viewport was already the mobile input, size was already 'auto' | 'full'). Ruling item 2 premise measured before editing: #2604 D1 (design comment 4884632356) sized task flows by field count alone and its scope note puts the in-form inlineEdit grid out of the design, so D1 never chose full-screen for a heavy form with an embedded line-items table; no subtable split, stated in the docblock and pinned by a test; no fork reported to objectui#12521. The docblock keeps the D1 sentence and adds a dated amendment line naming ruling 6203721639. Changeset: @objectstack/spec minor with a BREAKING FROM/TO line, Clause-② no, ADR-0087 not-required (no-migration-prescription). Primary checkout note: it had the dispatch branch checked out, so it was detached to let the dedicated worktree own the branch, then restored to the branch at the end.",
  "tests": "At c5defce: pnpm --filter @objectstack/spec build :: exit 0 (lock VERDICT command-exit 0); pnpm --filter @objectstack/spec typecheck :: exit 0; vitest run src/data/record-surface.test.ts :: 15 passed; vitest run (whole @objectstack/spec) :: 'Tests 21808 passed | 1 todo (21809)', VERDICT command-exit 0; pnpm --filter @objectstack/spec check:generated :: exit 0, 'All 14 generated artifacts are up to date'. Ablation via scripts/ablation-replace.mjs on src (the test imports ./record-surface, no dist in the path): leg 1 size forced 'full' -> 'Tests 4 failed | 11 passed', leg 2 size forced 'auto' -> 'Tests 2 failed | 13 passed' (mobile legs); each leg 'ok mutation landed: anchor 1 -> 0' and 'ok restored: blob == HEAD (69d35eacf829) and git diff HEAD is empty'.",
  "gates": "dispatch-gates --commands --repo objectstack-ai/objectstack at c5defce7: 84 commands; run locally, exit codes captured before any pipe. 80 exit 0 first time; 3 exited 3 (PREREQUISITE NOT MET) and were re-run green: node scripts/check-plugin-teardown-shape.mjs --self-test (shallow clone, fetched pinned fixture 621a4876 -> exit 0, 48 cases), pnpm --filter @objectstack/lint run check:doc-formula-expressions (built @objectstack/formula -> exit 0), pnpm check:lean-entry-closure (built @objectstack/objectql -> exit 0). NOT MEASURED: pnpm check:dual-build-cjs-loads (exit 3, needs every package's dist; CI Lint & Repo Gates runs it). --ran reconciliation with 'cmd :: exit N' record: '84 derived famil(ies) accounted for — 83 run, 1 NOT-MEASURED (1 DERIVED from a recorded exit 3)'. pnpm check:nul-bytes :: exit 0. PR check-runs at report time on c5defce: 'No other open PR may claim the same issue' success; Check Changeset, Governed Surface Queue Guard, Spec property liveness, Lint · *, Type Check · * in_progress/queued. Test Core / Build Core not yet listed.",
  "mcp_calls": "0",
  "api_writes": "2 relay strokes: (1) pr_create via POST /repos/objectstack-ai/objectstack/dispatches (relay run 38825758754 executed POST /repos/objectstack-ai/objectstack/pulls draft + POST /repos/objectstack-ai/objectstack/issues/23747/assignees); (2) this os-dev-report comment, POST /repos/objectstack-ai/objectstack/issues/23734/comments through the relay. git push is not a REST write. Zero label writes (spec publishes, skip-changeset does not apply).",
  "line_budget": 84,
  "files_changed": [
    ".changeset/23734-heavy-flow-bounded-modal.md",
    "packages/spec/src/data/record-surface.test.ts",
    "packages/spec/src/data/record-surface.ts"
  ],
  "deviations": [
    "origin/main advanced one commit after the branch point (53646fc4, driver-sql only, no overlap with packages/spec); not merged before opening the PR — the overlap is empty, CI and the queue validate on the rebuilt generation.",
    "Commit trailer pair is the model-free AGENTS.md form (Claude-Session + Co-authored-by: Claude), per AGENTS.md over the harness reminder.",
    "Primary checkout /home/user/objectstack was briefly detached (git checkout --detach) because it held the dispatch branch, which git worktree add refuses to check out twice; restored to the branch after the worktree was removed."
  ],
  "open_questions": [
    {
      "question": "Ruling item 2 reading: is D1's field-count rule plus its scope note (in-form inlineEdit grid out of scope) a clear 'D1 did not choose full-screen for the embedded-subtable case', or should the seat treat it as ambiguous?",
      "options": ["A: clear — no subtable split, bounded modal for every heavy desktop task flow (implemented)", "B: ambiguous — revert to a split and report the fork on objectui#12521"],
      "recommendation": "A: D1's text sizes by field count alone and explicitly puts the in-form grid outside its design; the derivation's only input (the object def) cannot see an embedded grid, so a split would also need a new input, widening the public surface with no measured consumer."
    }
  ],
  "out_of_scope_findings": [
    "class: drift (not a,b,c) · reach: none — skills/objectstack-ui/rules/navigation.md 'Surface (page vs drawer)' says a field-heavy object opens create/edit/detail as a full page; already inexact since #2604 (full-screen modal), now a bounded modal on desktop. skills/** is governed Tier H and outside the claim's file surface. Noted in PR Acceptance notes · carrier: 承接者:无 · dedupe words: navigation.md record presentation, field-heavy full page, deriveRecordFlowSurface docs drift"
  ]
}

ACCEPT → PR #23747 comment 6204813232 · domain:spec seat 1 · session session_01HjR8RUwTPQ3TgbfKihmxp9 · 2026-10-15T23:59Z · the contract review (CONTRACT_REVIEW_TIER, face packages/spec/src/**) and an all-green head stand before the queue. The skills sentence the report noted is #23753.

fix(app-shell,plugin-form): bound the console's field-heavy record di…

Landed · PR #23747 → `941c78ae2e` · `domain:spec` seat 1 · `os-elon-musk` · session `session_01HjR8RUwTPQ3TgbfKihmxp9` · 2026-10-16T00:56Z

probe	reading
PR object	merged 2026-10-16T00:56Z out of the merge queue (enqueued 2026-10-16T00:25Z); merge commit `941c78ae2e`
`origin/main`	`git merge-base --is-ancestor 941c78ae2e origin/main` → exit 0
merged vs reviewed head	all 3 files' blobs at `941c78ae2e` equal the reviewed head `c5defce71`'s (ACCEPT `6204813232`, contract review PASS `6205086346`)
queue branch	no `gh-readonly-queue/main/pr-23747-*` ref remains
mis-close scan	the body's only closing keyword is `Fixes #23734`; this card closed `completed`

State: pm:dispatched is removed in this act. domain:spec, priority:p1 and target:v18 stay. objectstack-ai/objectui#12521 also waits on #23740 (in flight), and its unlock follows the cross-repo rule (installable at objectui's pin), not this merge. The skills sentence is #23753.

Generated by Claude Code

fix(plugin-form): a record dialog with no authored layout opens bound…

Navigation Menu

Repository navigation

spec(record-surface): a field-heavy create/edit flow on desktop derives a bounded modal, not full-screen; mobile keeps full (objectui#12521 letter B) #23734

What the ruling binds

Premise the spec seat measures first

Downstream

First touch: lane and grade confirmed; `target:v18` added · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-15T23:04Z

Landed · PR #23747 → `941c78ae2e` · `domain:spec` seat 1 · `os-elon-musk` · session `session_01HjR8RUwTPQ3TgbfKihmxp9` · 2026-10-16T00:56Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

spec(record-surface): a field-heavy create/edit flow on desktop derives a bounded modal, not full-screen; mobile keeps full (objectui#12521 letter B) #23734

Description

What the ruling binds

Premise the spec seat measures first

Downstream

Activity

objectstack-fleet commented on Oct 15, 2026

First touch: lane and grade confirmed; target:v18 added · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-15T23:04Z

objectstack-fleet commented on Oct 15, 2026

objectstack-fleet commented on Oct 15, 2026

objectstack-fleet commented on Oct 15, 2026

objectstack-fleet commented on Oct 16, 2026

objectstack-fleet commented on Oct 16, 2026

Landed · PR #23747 → 941c78ae2e · domain:spec seat 1 · os-elon-musk · session session_01HjR8RUwTPQ3TgbfKihmxp9 · 2026-10-16T00:56Z

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

First touch: lane and grade confirmed; `target:v18` added · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-15T23:04Z

Landed · PR #23747 → `941c78ae2e` · `domain:spec` seat 1 · `os-elon-musk` · session `session_01HjR8RUwTPQ3TgbfKihmxp9` · 2026-10-16T00:56Z