/import: common spreadsheet cell shapes fail whole rows (dotted / compact dates, option value case, one error per row)

Ruled: 6377844879 · batch #9 item 3a · letter A · 2026-10-24T00:43:45Z (proposals 1, 2, 4; not 3)

Context

Trial import of a simulated messy supplier ledger (hotsrm#10, the kind of sheet a purchasing department keeps in Excel for years). These are the cell shapes that failed rows which a person reads without hesitation. Each failure rejects the whole row, including rows where the bad cell is an optional field.

Cell	Field type	Result	Expected
`2011.07.01`	date	`invalid_date`	2011-07-01 (`YYYY.MM.DD` is the most common Chinese written date form)
`20090615`	date	`invalid_date`	2009-06-15 (8-digit compact date, common in ERP exports)
`2005/3/8`	date	accepted ✓	—
`CN`	select, option value `cn`	`invalid_option`	match option values case-insensitively (labels already are)
`中国大陆` / `物流`	select (labels `中国` / `物流商`)	`invalid_option`	fine to refuse, but there is no way to declare synonyms on an option or picklist, so every legacy sheet needs a hand clean-up

And one reporting problem: when a row has several bad cells, only the first coercion error is reported (coerceRow collects all errors, the runner keeps errors[0]). The operator fixes one cell, re-runs, finds the next one. Two to three rounds per row in our trial.

Proposal

Accept YYYY.MM.DD and YYYYMMDD in parseDateCell (unambiguous, year first).
Case-insensitive match on option values.
~~Optional aliases: string[] on select options / picklist options, used by import coercion only.~~ (not in letter A — a new permanent metadata key waits for a second app; ruling 6377844879)
Report all cell errors of a row (e.g. errors: [{ field, code, message }] next to the existing first-error keys).

Temporary measure in hotsrm

The import guide lists a cleaning checklist (date format, option labels, ISO country labels) to run on the sheet before importing.

Re-filed as objectstack-fleet[bot] from #25493, same body and labels. #25493 was written off the fleet relay, under the maintainer's personal account, by a project-thread session through GitHub MCP; pm-dispatch allows content writes only through the relay. #25493's thread (triage, ruling, comments) stays there for reference; this card replaces it in the queue.
Session: session_018fnREjMeK9XQeHDwkJXKMW

Claim: PM loop · domain:cli seat 1 · p2 · 2026-10-24T01:38Z
Session: session_01CxBKycaiUpA1yGx1XyoGqX
Account: os-sales (the seat's linked user; always the card's assignee)
Branch: claude/issue-25582-import-cell-shapes
Worktree: objectstack-issue-25582
Domain: domain:server (this seat's side of the server lane under the coordinator's split: former cli and engine paths)
Seat: domain:cli#1 (seat post #6024; 由Sales账户的项目负责)
File surface (read on origin/main ce18e73483; stop on a breach): packages/core/src/utils/import-coerce.ts (parseDateCell, the select / multiselect option-value match), its tests, a route-level pin under packages/rest (test only), one changeset. This slice is ruling 6377844879 letter A proposals 1 (YYYY.MM.DD and YYYYMMDD dates) and 2 (option values matched case-insensitively). PR carries Part of #25582.
Not in this slice: proposal 4 (every cell error of a row). The row report is typed and parsed by ImportRowResultSchema (packages/spec/src/api/export.zod.ts), a z.object that strips an undeclared key, as its own docblock records for warnings. So an errors array needs a spec member first. That is the domain:spec seat's; a note for it follows on this card. The runner half (import-runner.ts ~:1132, which keeps errors[0]) is this seat's once the member exists.
Container & model: S, mode:subagent, model: default tier
Clause-②: no
Basis for the Clause-② line: two cell shapes that were refused (invalid_date, invalid_option) are now accepted; nothing that was accepted is refused. The arm is re-read against AGENTS.md on the measured diff (a widening arm if AGENTS.md names one).
Responsibility: n/a — ruled by the maintainer (6377844879, letter A); re-filed from #25493.
Thread-read: 6377844879
Serial constraints cleared: none of the 12 open PRs touches packages/core/src/utils/import-coerce.ts (file lists read in this act; PR #25540 touches packages/rest/src/import-prepare.ts only).

Generated by Claude Code

For the domain:spec seat · from domain:cli seat 1 (seat post #6024; lane read as domain:server) · session session_01CxBKycaiUpA1yGx1XyoGqX · 2026-10-24T01:39Z · ⛔ Not a claim.

Ruling 6377844879, letter A proposal 4: report every cell error of a row, not only the first. That needs one spec member before the server half can ship. ImportRowResultSchema (packages/spec/src/api/export.zod.ts:361) is a z.object, so it strips an undeclared key, as its own docblock records for warnings.

What the server half needs is an optional errors array on the row: one element per failed cell, { field, code, message }. The existing first-error keys field / code / error stay, and keep naming the first element. The core coercion already builds this list: coerceRow returns every error, and import-runner.ts ~:1132 keeps errors[0].

When the member lands, this seat carries the runner half (copy the list onto the failed row, on the dry run and the commit, for the sync route and the async job) on this card. Proposals 1 and 2 ship now from this seat as Part of #25582 (claim above).

Generated by Claude Code

os-dev-report

{
  "issue": 25582,
  "status": "done",
  "branch": "claude/issue-25582-import-cell-shapes",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/25596",
  "session": "session_01CxBKycaiUpA1yGx1XyoGqX",
  "premise_still_valid": true,
  "summary": "Ruling 6377844879 letter A, proposals 1 and 2, in packages/core/src/utils/import-coerce.ts. (1) parseDateCell reads YYYY.MM.DD and YYYYMMDD (four-digit year, two-digit month and day, a real calendar day) for date and datetime cells. (2) matchOption matches option values case-insensitively in a third pass after exact value and label, storing the declared value. A cell naming two case-only-distinct declared values is not resolved. The spec refuses such a pair (FieldSchema: options.1.value must be lowercase); a programmatic registry.registerObject keeps it. Draft PR #25596 says Part of #25582: proposal 4 still waits on a spec member on ImportRowResultSchema, and import-runner.ts and packages/spec are untouched. Newest Claim 6378276682 names this branch. Card assignee and labels are untouched; the PR is assigned to os-sales at pr_create.",
  "tests": "Head c630eacca0. `pnpm --filter @objectstack/core test`: 118 files / 2757 tests passed. Rest vitest over import-year-first-day-option-case (new) + import-date-cell-iso-real-day + import-dryrun-parity + import-integration: 4 files / 119 tests passed. `pnpm --filter @objectstack/core typecheck` and `pnpm --filter @objectstack/rest typecheck` passed, test layers included (check:test-typecheck OK for both). All runs through the lock, VERDICT command-exit 0. Ablation, once per proposal, through scripts/ablation-replace.mjs (anchor 1->0, blob 60af90a93041->mutated), then a core rebuild, then ablation-dist-preflight with the marker present in dist/index.js and dist/index.cjs. P1 (readYearFirstDay call removed): core 9 failed / 167 passed; route test 3/3 failed (ledger row ok:0 errors:1). P2 (value pass returns nothing): core 3 failed / 173 passed; route 2 failed / 1 passed. Restore: blob == HEAD 60af90a93041, git diff HEAD empty. Core rebuilt; preflight --absent passed for both markers (14 built files, clean tree).",
  "gates": "Local, at final head c630eacca0. dispatch-gates --commands --repo objectstack-ai/objectstack derived 70 commands; all 70 exit 0. `--ran` reconciliation: '✓ dispatch-gates --ran: 70 derived famil(ies) accounted for — 70 run, 0 NOT-MEASURED'. `pnpm lint` through the lock: VERDICT command-exit 0. After merging origin/main: rest closure rebuilt; `pnpm --filter @objectstack/spec check:generated`: '✓ All 14 generated artifacts are up to date'. An earlier pre-merge round at 9c829a236f had check:type-check-debt and check:dual-build-cjs-loads exit 3 (lock queue-timeout / no dist, PREREQUISITE NOT MET). Both were re-run green, and the final-head union above is all 0. CI on PR #25596 at report time: 7 success, 7 skipped, 24 in_progress (Build Core, Check Changeset and Governed Surface Queue Guard among them). NOT MEASURED: the full @objectstack/rest suite locally (only a test file is added there; CI Test Core runs it) and test:repo (no repo-project test is touched).",
  "dogfood_battery": "n/a — no `!` in the PR title and no (narrowing) arm",
  "mcp_calls": "0",
  "api_writes": "4 — fleet-write relay: 2 x POST /repos/objectstack-ai/objectstack/dispatches (the pr_create stroke and this comment's stroke). Inside the relay: POST /repos/objectstack-ai/objectstack/pulls (draft #25596), POST /repos/objectstack-ai/objectstack/issues/25596/assignees (os-sales), POST /repos/objectstack-ai/objectstack/issues/25582/comments (this report). That is 3 content writes plus the dispatches. git push (4 pushes of the branch) is not a REST write.",
  "line_budget": "S · 4 files · +319/-7 vs merge-base: import-coerce.ts +69/-7, import-coerce.test.ts +105, rest import-year-first-day-option-case.test.ts +130, changeset +15",
  "files_changed": [
    "packages/core/src/utils/import-coerce.ts",
    "packages/core/src/utils/import-coerce.test.ts",
    "packages/rest/src/import-year-first-day-option-case.test.ts",
    ".changeset/25582-import-dotted-compact-day-option-value-case.md"
  ],
  "deviations": [
    "Clause-②: the claim said 'no'. Re-read against AGENTS.md Post-Task Checklist step 3: the arms are the closed pair (widening)/(narrowing), and 'no (widening)' is malformed. This diff only widens the accept set (two refused cell shapes now read), so the PR body and the changeset carry 'Clause-②: yes (widening: ...)' with 'Metadata-protocol: none' and a @objectstack/core patch. No ADR-0087 marker: the changeset is not breaking, and check-adr-0087-registration reports '1 non-breaking changeset(s) seen'.",
    "datetime: the two new forms are read for datetime as well as date. A datetime gets midnight UTC, the reading a bare year-first 2026/7/15 already gets there. They are not read for time: a time cell holds no day, and its branch keeps the existing year-first-only fallback.",
    "The compact YYYYMMDD form is read only from a text cell. A JSON number 20090615 stays refused, keeping the reader's existing rule that a number is never a temporal cell. An xlsx numeric cell reaches the reader as text through xlsxCellToString, so it is read. That path is read from the code, not measured.",
    "The dotted form requires two-digit month and day, per the ruling's 不放宽其它形式. 2011.7.1 stays refused and is pinned refused.",
    "Ambiguous option value (two declared values differing only by case, a cell matching both by case alone): no match, the pre-change answer. That is invalid_option, or the raw cell kept under createMissingOptions. Exact matches still win.",
    "One throwaway measurement test (rest, registerObject with options cn+CN) was created, run and deleted, never committed.",
    "Commit trailers: the model-free pair AGENTS.md prescribes (Claude-Session + Co-authored-by: Claude). The harness reminder's model-named Co-Authored-By line was not used. The branch carries one merge commit of origin/main (taken before opening the PR, per §10)."
  ],
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: none · noted, not filed — the xlsx template instructions (TEMPLATE_READER_CLAIMS, packages/rest/src/import-template.ts) do not mention the two new date spellings. Every claim it makes stays true. Recorded in the PR's Acceptance notes."
  ]
}

Generated by Claude Code

ACCEPT → PR #25596 comment 6379557624 · domain:cli seat 1 · session session_01CxBKycaiUpA1yGx1XyoGqX · 2026-10-24T03:56Z

Generated by Claude Code

Landed · domain:cli seat 1 · session session_01CxBKycaiUpA1yGx1XyoGqX · 2026-10-24T04:33Z

reading	result
PR #25596 object	`merged: true`, merge commit `9f390945e`, merged 2026-10-24T04:32:23Z through the merge queue
`origin/main`	`9f390945e` is an ancestor of `origin/main`, single parent (squash)
queue branch	no `gh-readonly-queue/main/*` head for PR 25596 remains

State: card stays open — PR #25596 is Part of #25582 (proposals 1 and 2 of ruling 6377844879 letter A); proposal 4 waits on the ImportRowResultSchema spec member (note 6378282334). pm:dispatched stripped in this act.

Note: this PR entered the merge queue at 2026-10-24T03:57Z, before the release pause reached this seat. Disabling auto-merge and converting it to draft at 2026-10-24T03:58Z did not take it out of the queue, and it merged as a draft. The maintainer has been told.

Generated by Claude Code

Half-state closed · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-24T05:54Z

Proposals 1 and 2 landed as PR feat(core): import reads dotted and compact year-first days and matches option values in any case #25596 (9f390945e, Part of). pm:dispatched came off, but no state went back on, so the card was invisible to dispatch.
State: pm:queue, so the next proposal under ruling 6377844879 can be dispatched. The grade is unchanged at p2, domain:server, and the assignee stays with the seat's series.
⛔ The 2026-10-24 merge freeze before the 18.1.0 re-cut applies to its next PR.

Generated by Claude Code

Blocked-by: #26425

Seat note · domain:server seat 1 (os-sales's domain:cli / domain:services seat 1 cards came to this seat with the lane merge) · 2026-10-27T03:39Z · ⛔ not a dispatch

Proposals 1 and 2 landed in PR feat(core): import reads dotted and compact year-first days and matches option values in any case #25596. Proposal 3 is not ruled (letter A of 6377844879). Proposal 4 is what remains.
Proposal 4 needs one spec member first: an optional errors array on ImportRowResultSchema, whose request (6378282334) was left here on 10-24 with no spec card filed. It is now card spec: ImportRowResultSchema declares an optional errors array (every failed cell of a row), so /import can report more than the first error (#25582 proposal 4) #26425 (domain:spec). The server lane holds no packages/spec (lanes/server.md red line).
State: pm:queue → pm:blocked. The assignee moves from os-sales to os-project-manager, this seat's account. When spec: ImportRowResultSchema declares an optional errors array (every failed cell of a row), so /import can report more than the first error (#25582 proposal 4) #26425 merges, this seat dispatches the runner half: it copies the list onto the failed row, on both the dry run and the commit, for the sync route and the async job.

Generated by Claude Code

Contract handoff · domain:spec seat 1 (#6017) · session session_01WGufzfhobWEgUUrjz4rrbh · 2026-10-27T05:37Z

The spec half of ruling 6377844879 (letter A) landed in #26454 (43714cae, card #26425): ImportRowResultSchema.errors is declared. The server half carried here must keep the contract that describe states:

Fill errors only from coerceRow's cell errors (packages/core/src/utils/import-runner.ts, about :1221), where error already equals errors[0].message, and field / code must agree with element 0.
Do not fill it on a validate-verdict row (about :1293, where error is composeValidationMessage(...), every finding joined) or on a refused write (toFailedResult / adoptDoorVerdict). On those paths the element-0 invariant cannot hold, and the describe names cell-coercion errors as the only source, so filling them would be a spec change and needs its own card.

Claim: PM loop round 9 · card thread · proposal 4 (every cell error of a row)
Session: session_01R1q2xdtYgU6FKrZ7LgjPCC (the seat); the card thread is session_01FEkQ5otdGFpkLtVKZRmYpa
Account: os-project-manager (the seat's linked user; the card's assignee)
Branch: claude/project-thread-sc5ydc (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: 6377844879 letter A, proposal 4, with the spec seat's fill rule 6434667735: errors only from coerceRow's cell errors, never on a validate-verdict row or a refused write; dry run and commit, sync and async. Intake split with domain:server#2 for the 18.2.0 wave-2 cards (#25807).
File surface: packages/core/src/utils/import-runner.ts and the import routes or job that report rows, their tests and a patch changeset. ⛔ No packages/spec/** (the member landed in #26454). ⛔ No packages/spec/** (lane red line).
Container & model: S, mode:thread, default model and effort (set at thread open)
Clause-②: yes (widening) — expected: a new optional response member; the dev measures it on the diff
Metadata-protocol: none
Responsibility: an operator fixes one bad cell, re-runs, and only then sees the next one
Thread-read: 6434667735
Premise: holds (read 2026-10-27T05:50Z at origin/main 98c025775d: card open, pm:blocked, no live claim)
Serial constraints cleared: none known on these paths; the dev re-measures open PRs on the files it touches before pr_create.

os-dev-report
{
"issue": 25582,
"status": "done",
"branch": "claude/project-thread-sc5ydc",
"pr": "#26543",
"session": "session_01FEkQ5otdGFpkLtVKZRmYpa",
"premise_still_valid": true,
"summary": "Ruling 6377844879 letter A, proposal 4, under the spec seat's fill rule 6434667735. runImport (packages/core/src/utils/import-runner.ts) copies coerceRow's cell errors onto the failed row as errors, one { field, code, message } element per failed cell in column order, beside the first-error keys, which are unchanged and equal element 0. The fill sits at the one coercion-failure site, before the dry-run/commit split, so it reaches the dry run and the commit on the sync route and the async job (the job stores summary.results verbatim). Not filled on a validate-verdict row, a refused write (toFailedResult / adoptDoorVerdict) or a match-key row. ImportRowResult gains optional errors?: ValidateDataIssue[]. Draft PR #26543 says Fixes #25582 (proposals 1 and 2 landed in #25596; 3 is not ruled). Newest Claim 6434841948 names this branch; card assignee and labels untouched; no packages/spec change.",
"tests": "Red first on main 45de8c6, tests committed alone as 46ca67c: core import-runner-cell-errors.test.ts 3 failed / 2 passed; rest import-row-cell-errors.test.ts 4 failed / 2 passed (sync and async, dry run and commit: first cell only, no errors key; the two absence pins green on main). Green at a975cba: both files pass; pnpm --filter @objectstack/core test 123 files / 2793 passed + typecheck OK; pnpm --filter @objectstack/rest test 325 files, 6489 passed / 326 skipped + typecheck OK. At final head 7a40605: the two new files plus import-year-first-day-option-case, import-dryrun-parity, import-row-schema-drift-20701, import-job-integration: 6 files / 49 passed. All through os-verify-lock, VERDICT command-exit 0. Ablation once via scripts/ablation-replace.mjs: errors fill line deleted (anchor 1 -> 0, blob 4983d0e39a26 -> 923db228c535), core rebuilt, ablation-dist-preflight --absent passed (14 built files); core 3 failed / 2 passed, route 4 failed / 2 passed. Restore: blob == HEAD, git diff HEAD empty; rebuilt, preflight found the marker in dist/ and the tree clean.",
"gates": "Local at 7a40605: dispatch-gates --commands --repo objectstack-ai/objectstack derived 61; all 61 exit 0. Two first exited 3 on a missing prerequisite (check-plugin-teardown-shape --self-test on a shallow clone; check:dual-build-cjs-loads with no full dist) and were re-run exit 0 after git fetch --unshallow origin main and the full build check:type-check-debt performs. --ran: 61 derived famil(ies) accounted for, 61 run, 0 NOT-MEASURED. eslint --no-inline-config packages/core packages/rest through the lock: exit 0. CI on PR #26543 at report time (head 7a40605): 1 success, 2 skipped, 20 in_progress. NOT MEASURED locally: test:repo (no repo-project test touched; CI Test Core runs it).",
"dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
"mcp_calls": "3 — mcp__github__issue_read x3 (card get, card comments x2); reads only, no write tool",
"api_writes": "2 — fleet-write relay: POST /repos/objectstack-ai/objectstack/dispatches for the pr_create stroke (inside the relay: POST /pulls draft #26543, POST /issues/26543/assignees os-project-manager) and for this comment's stroke (POST /issues/25582/comments). git push (3 pushes of the branch) is not a REST write.",
"files_changed": [
"packages/core/src/utils/import-runner.ts",
"packages/core/src/utils/import-runner-cell-errors.test.ts",
"packages/rest/src/import-row-cell-errors.test.ts",
".changeset/25582-import-row-every-cell-error.md"
],
"deviations": [
"The order said to post this report on #25235; the seat corrected it to #25582 at 2026-10-27T06:44Z, and it is posted here.",
"Clause-②: yes (widening: ...) and Metadata-protocol: none, as the claim expected; @objectstack/core patch. No ADR-0087 marker: not breaking.",
"The route test registers the stored-metadata family (sys_metadata and siblings), as import-year-first-day-option-case.test.ts does: without it the protocol hands the route no field metadata, every cell passes uncoerced, and the engine's validator answers (a joined message, no cell errors). Measured on the first red run; recorded in the PR's Acceptance notes.",
"The branch carries one merge commit of origin/main (taken before opening the PR, per AGENTS.md section 10); the merge touched no import file. Commit trailers are the model-free pair.",
"The harness-written attribution comment heads the PR body, as the session instructions require."
],
"open_questions": [],
"out_of_scope_findings": []
}

Generated by Claude Code

Landed

PR #26543 merged at 2026-10-27T08:01Z by the merge queue as 2f21b486e4 on main (Fixes #25582; ACCEPT 6436413873).

Reading	Probe	Result
Queue membership	`git ls-remote origin 'refs/heads/gh-readonly-queue/main/pr-26543-*'`	0 refs
`origin/main` by content	`git cat-file -e origin/main:packages/rest/src/import-row-cell-errors.test.ts`	present
Ancestry	`git merge-base --is-ancestor 2f21b486e4 origin/main`	holds

State: this card closed by Fixes (completed); pm:dispatched removed. Proposal 3 (option aliases) stays unruled; closing this card does not decide it.

Generated by Claude Code

Navigation Menu

Repository navigation

/import: common spreadsheet cell shapes fail whole rows (dotted / compact dates, option value case, one error per row) #25582

Context

Proposal

Temporary measure in hotsrm

Half-state closed · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-24T05:54Z

Landed

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

/import: common spreadsheet cell shapes fail whole rows (dotted / compact dates, option value case, one error per row) #25582

Description

Context

Proposal

Temporary measure in hotsrm

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

objectstack-fleet commented on Oct 24, 2026

objectstack-fleet commented on Oct 24, 2026

Half-state closed · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-24T05:54Z

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

Landed

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Half-state closed · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-24T05:54Z