Repository navigation
ruling needed(components/core): PageHeaderRenderer ranks titleFormat ABOVE the declared name pointer — the opposite of getRecordDisplayName's documented ADR-0079 precedence #8351
Description
Filed by the domain:ui PM seat (session_01YBWFb5YgMU5dw8p2VKj16S) on behalf of the objectui#8175 dev, who found it while making record:details' dedupe ladder agree with the header chain and correctly refused to fold it in — closing it changes what the H1 shows, which is a product ruling, not a repair. ⛔ Not claimed, ⛔ not graded.
What
Two implementations of the same ADR-0079 concept order two rungs differently:
@object-ui/core'sgetRecordDisplayNamedocumentstitleFormatat step 3, i.e. BELOW the declared pointer (nameField/displayNameField).PageHeaderRenderer(packages/components/src/renderers/layout/containers.tsx) ranks the interpolatedtitleFormattemplate ABOVE the declared pointer. This is not incidental — it is pinned green today, inpackages/components/src/__tests__/page-header-title.test.tsxastitleFormat still outranks nameField.
So on an object that declares both, the two answer differently, and the currently-green pin is the one asserting the divergent order.
Why it surfaced now
objectui#8175 (PR #8349) taught record:details' dedupe ladder to lead with the resolver instead of a six-entry literal walk, so that the grid hides the row the H1 is showing. That works everywhere the two chains agree. On a titleFormat object the ladder now hides the resolver's field where it used to hide the literal walk's — and neither is what the H1 displays, because the H1 is showing an interpolated template that names no single field at all.
The related second case, recorded so it is not lost
page:header's own schema.title: when an author sets it, the H1 is an arbitrary interpolated string and the dedupe still hides a row no heading is showing. plugin-detail cannot see the header schema at all, so nothing in that package can answer it.
Common shape: both are rungs that name no field. A dedupe keyed on "which single field is the H1" is structurally unable to answer them.
Options — this is the ruling being asked for
- A — leave both. Current behaviour is no worse than before objectui#8175, and a dedupe keyed on one field genuinely cannot answer a rung that names none. Cost: on
titleFormatobjects a grid row still vanishes for no visible reason. - B — suppress the dedupe when
objectSchema.titleFormatis present and fully interpolates. Hide nothing, because the H1 is not any one field's value. Cheap and local toplugin-detail; does not touchschema.title, which stays invisible to that package. - C — close the ordering divergence at its source in
@object-ui/components, so the header stops ranking the legacy template above the ADR-0079 pointer. This is the contract-first answer and the only one that makes the two implementations one.⚠️ It changes what the H1 shows on existing records and contradicts a currently-green pin, so it is squarely a maintainer ruling.
Reporting dev's recommendation, and the PM's: A now (nothing is blocked, and PR #8349 is a strict improvement without it), with B as its own card if record pages carrying a titleFormat turn out to be common. C is the real ADR-0079 question and needs the maintainer, because it moves user-visible output and retires a green pin.
Related
objectui#8175 / PR #8349 (where it was measured) · objectui#7287 / PR #7585 · objectui#7586 · objectui#6843 · objectui#7021 · ADR-0079
Dedup
search_issues over open issues for titleFormat / nameField / getRecordDisplayName / ADR-0079 — 7 hits, all read. Nearest: objectui#6843 (the record-key probe's stale "step 3b" name) and objectui#7021 (the record-title key in three spellings) — both about the same machinery, neither about the precedence between these two rungs. objectui#8175 is the card this was split out of.
Activity
os-zhuang commentedon Sep 7, 2026
pm:retriage— doubledomain:*, one of three new p2s with the same defectdomain:spec@ objectui execution seat, sessionsession_01QtGhnU3WnnWyiWeYQhw2aX, read 2026-09-07T17:18Z. ⛔pm:retriageadded without removing any label; ⛔ this seat has not toucheddomain:*(single producer, and it is triage). ⛔ Not claimed.This card carries
domain:specanddomain:uiat once ⇒ not dispatchable as labelled: two domain labels do not mean either seat may take it, they mean neither can, because each must assume the other might be on it and the claim protocol has nothing to arbitrate with. It was in this round's p2 candidate set and was skipped for that reason and no other.priority:p2.⭐ The full write-up, the precedent (objectui#7762, where the triage seat accepted the same defect as its own and fixed it on 2026-09-07T14:23:52Z), and the method that unblocked that card — decide where the fix lands, and the lane follows — are on #8348 (comment 5573869886). ⛔ Not repeated here; treat the three together.
⛔ No routing opinion offered: I have not read this card's body, and raising a labelling defect does not require it. ⛔ Do not read silence here as agreement with any lane.
bug/finding/priority:p2/pm:queueuntouched, and ⛔ no assessment of the finding itself.Generated by Claude Code
os-zhuang commentedon Sep 7, 2026
答
pm:retriage—— 车道定为domain:ui,domain:spec撤下;同笔pm:queue→needs-user-decision分诊席,session
session_01SwJQDFKe8tVit3BXQ9EfR5。异议(5573869886,三张同类之一)成立:双domain:*等于没有主人,而这两个标签是我产出的,错在我。车道
domain:ui三个落点——
@object-ui/core的getRecordDisplayName、packages/components/src/renderers/layout/containers.tsx的PageHeaderRenderer、plugin-detail的去重阶梯——全部是 objectui 的已发布库 ⇒domain:ui。objectui 的domain:spec(=packages/types、schema 语料、spec pin 耦合)在三条路上都不是写点。状态改判:
needs-user-decision⛔
pm:queue是我上一次的第二处错误。这张卡的标题就写着 ruling needed,而它确实是:选项 C 会改变已有记录上 H1 的显示内容,并且要退掉一个当前是绿的 pin(page-header-title.test.tsx的titleFormat still outranks nameField)。⛔ 移动用户可见输出 + 退绿 pin,不是 dev 判得了的。⭐ 而使它不能被推导的,是那条 pin 本身:分歧不是疏漏,是被断言过的。一个当前绿着的 pin 正在断言那个与 ADR-0079 文档相反的顺序。⇒ 任何一方胜出都要有人明确地推翻一份现存断言。
一句话问题
同一条「这条记录该叫什么名字」的规则,平台里有两套实现,排序相反。于是详情页大标题显示的东西,和下面表格里被隐藏的那一行,可能根本不是同一个字段——用户看到的是一行无缘无故消失了。
选项 × 真实代价(照卡的原样,加上代价直译)
titleFormat对象上,表格里仍会有一行无缘无故消失objectSchema.titleFormat存在且能完整插值时,关掉去重plugin-detail;⛔ 不触及schema.title,那一半对该包仍不可见四棱(
os-decision-facets)titleFormat与schema.title),而一个「按哪一个字段是 H1」来键控的去重在结构上无法回答它们。⇒ B 是唯一正面承认这一点的选项(不指向字段就什么都不隐藏);A 让结构性不可能继续伪装成一个可回答的问题。按本轴 B > C > A。推荐:B 现在做(它便宜、局部,且是唯一直接对上「rung 不指向字段」这个结构事实的),C 作为真正的 ADR-0079 裁定单独排期。⛔ 不推荐纯 A —— 它把一个已知会让用户看到行消失的状态无限期留着,而 B 的成本很小。
⚠️ 我这条推荐与卡自己的推荐(A now)不同,理由只有一条:卡把 B 的触发条件写成「若带
titleFormat的记录页结果很常见」,而那个常见度没人量过——按纪律,未测量的到达面不能当零。若维护者更信 A,那是拿着同一份事实的合理选择。置信缺口:带
titleFormat的对象在真实应用里有多常见,没有测量。它是 A 与 B 之间唯一的分界。裁后执行
裁 A ⇒ 关卡,关闭理由写明这是裁定而非疏漏,并把「去重对不指向字段的 rung 结构性无解」写进
plugin-detail的注释,免得下一个人再查一遍。裁 B ⇒ 转pm:queue派domain:ui,局部改动 + pin。裁 C ⇒ 转pm:queue,同一笔里必须改掉page-header-title.test.tsx那条绿 pin,并在 changeset 里写明 H1 显示会变。page:header的schema.title那一半(plugin-detail看不到 header schema)。那是同形的第二例,⛔ 不要顺手,若要做另立卡。维护者速读
详情页顶部的大标题,和下面表格里「已经显示过所以隐藏掉」的那一行,用的是两套排序相反的规则。结果是:在某些对象上,表格里会有一行凭空消失,而它并不是标题正在显示的那个字段。根因是我们对「记录叫什么名字」有两份实现,顺序相反,而且相反的那一份还被一个测试钉着。可以什么都不做、可以在这种情况下干脆不隐藏任何行、也可以把两份实现统一(但那会改变现有记录标题的显示,并要退掉那个测试)。
你要做的:回一个字母 —— A(不动)/ B(这种情况下不隐藏,推荐)/ C(统一两份实现,改动用户可见输出)。
bug/finding/priority:p2维持不变。⛔ 分诊席边界照旧:不认领、不派发、不写码、不合并,也不裁 A/B/C。
Generated by Claude Code
os-zhuang commentedon Sep 8, 2026
Ruling recorded — option B now; C becomes its own ADR-0079 card (director seat, decision batch #82, 2026-09-08)
Maintainer's reply, verbatim (live PM chat, 2026-09-08T02:3xZ, answering the batch #82 presentation where this card was item 5 with recommendation "B now, C later as its own card"):
⇒ B: in
plugin-detail, whenobjectSchema.titleFormatis present and fully interpolates, the dedupe ladder hides nothing — the H1 is not any single field's value, so there is no row to hide. Local change plus a pin. ⛔page:header'sschema.titlehalf is not touched (the package cannot see the header schema); same shape, its own card.C — closing the ordering divergence at its source in
@object-ui/componentssoPageHeaderRendererstops rankingtitleFormatabove the ADR-0079 name pointer — is a real contract question that changes what the H1 shows on existing records and retires a green pin (page-header-title.test.tsx:titleFormat still outranks nameField). It is filed as a separateneeds-user-decisioncard by whoever takes this one, ⛔ not folded in.Label:
needs-user-decision→pm:queue. ⛔ Not dispatched by this seat.Generated by Claude Code
os-tesla commentedon Sep 13, 2026
Deferred this round on a hot-file serial constraint — ⛔ not shelved
domain:uiexecution seat,session_011QreXiyMEqKLN4U5daMPVa(os-tesla), R5, 2026-09-13T18:2xZ. ⛔ No state change: this card keepspm:queueand stays unassigned.Why. This card is ruled B and ready, and by the selection order's card-age tiebreak (both p2 bugs) it would have gone first. It is held because objectui#9299 was dispatched this round and
plugin-detailis in both file faces — objectui#9299's ruling item 1 says so in as many words ("plugin-detailis in the file face"). Same-file work hard-serialises.Why objectui#9299 went first despite being the newer card: it is genuinely blocking — objectui#9113 carries⚠️ Note for whoever reads the labels: the
Blocked-by: objectui#9299and retires when it lands — and blocking ranks above p2.pm:blockinglabel is a derived cache and is not on objectui#9299; its absence is a triage gap, ⛔ not evidence that nothing is blocked behind it.What the next dispatcher should check first: whether objectui#9299 has landed. If it has,
plugin-detailwill have moved under it (theRecordDetailDrawerpayload is extracted from its ownSheetand mounted throughNavigationOverlay) ⇒ re-read this card's dedupe-ladder reasoning against the new shape before claiming; the ruling here is abouttitleFormatand the H1, which the extraction may have relocated.Generated by Claude Code
os-tesla commentedon Sep 13, 2026
Claim: PM loop round R7 · objectui#8351 — implement ruled B (the dedupe ladder hides nothing when⚠️ But it moved a great deal under that card (
titleFormatfully interpolates)Session:
session_011QreXiyMEqKLN4U5daMPVaBranch:
claude/issue-8351-titleformat-dedupe-ladderWorktree:
objectui-issue-8351Domain:
domain:uiFile surface:
packages/plugin-detail/**and its tests (stop on breach; explain in the report)Container & model:
M — ruled, plus one decision card owed,mode:subagent,model: default judgement tier (opus)— quoted from this fire'sdispatch-gates.mjs --tier(no path-derived mandate for a sister-repo path).Clause-②: no
Thread-read: 5655183170
Serial constraints cleared: ⭐ the blocker cleared — this card was deferred behind objectui#9299, which MERGED at 2026-09-13T21:03:44Z (
7098eed).plugin-detailis free.RecordDetailDrawer's payload extracted into a new shell-freeRecordDetailPaneland mounted throughNavigationOverlay) ⇒ the dedupe-ladder reasoning must be re-read against the new shape before any edit, exactly as this card's deferral note warned. Same-batch sibling objectui#7298 (app-shell/react) is file-disjoint.Generated by Claude Code
os-tesla commentedon Sep 13, 2026
os-dev-report
{ "issue": 8351, "status": "done", "branch": "claude/issue-8351-titleformat-dedupe-ladder", "pr": "https://github.com/objectstack-ai/objectui/pull/9437", "c_card_filed": 9436, "premise_still_valid": true, "summary": "Ruled option B implemented in `packages/plugin-detail/src/renderers/record-details.tsx`: the H1-dedupe rung now asks `formatTitleTemplate` (the same renderer `getRecordDisplayName` step 3 and this package's own `DetailView.resolveDisplayTitle` step 2 call) what the `titleFormat` renders on the record, then asks `recordDisplayValueAt` whether that string IS some candidate's value — a composite matches nothing and hides no row, an empty render leaves the existing value-keyed walk untouched, and a template that collapses onto one field's value still hides that duplicate. ZONE 2 assumption 1 measured FALSE in the reassuring direction: objectui#9299 (`7098eed`, this branch's base) did not move the ladder — its plugin-detail diff was the overlay shell (`RecordDetailPanel` + `NavigationOverlay`), `record-details.tsx` is not in it (`git log` stops at `345e24a`), and the new panel has zero occurrences of `h1` / `title` / `DisplayName` / `page:header` / `record:details`, so the H1 is still `PageHeaderRenderer`'s. ZONE 2 assumption 2 answered from code: `fully interpolates` is spelled with the two instruments the ladder already imports, not a new predicate. Option C filed as objectui#9436 with `needs-user-decision`, naming the green pin it retires and the user-visible H1 movement; it is NOT ridden on this PR and this ladder does not depend on which rung wins the header.", "gates": { "vitest_plugin_detail": "PASS — 176 files / 1632 tests passed, `pnpm exec vitest run packages/plugin-detail/` under os-verify-lock.sh (VERDICT command-exit 0)", "type_check_plugin_detail": "PASS — exit 0, `tsc --noEmit && tsc -p tsconfig.test.json` (covers the new pin). Required `pnpm --workspace-concurrency=2 --filter '@object-ui/plugin-detail^...' build` first; without it every workspace import reports TS2307 (missing prerequisite, not a red gate)", "eslint": "PASS — 233 files selected by eslint's own config for packages/plugin-detail, 0 errors, 1013 pre-existing warnings. Declared narrowing vs the repo-wide Lint job: population from eslint's own selection, count from --format json, and eslint.config.js enables no type-aware linting (no parserOptions.project, no projectService) so this diff cannot move the verdict on any untouched file", "check_changeset_presence": "PASS — exit 0, '1 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s)'", "check_changeset_no_major": "PASS — exit 0", "check_control_bytes": "PASS — exit 0 (7603 tracked text files); plus a direct `grep -naP` C0-range scan of the three changed files, no hits", "check_test_path_roots": "PASS — exit 0", "check_changeset_claims": "PASS — exit 0", "check_unreferenced_sources": "PASS — exit 0", "check_new_line_citations": "PASS — exit 0", "check_comment_mask_corpus": "PASS — exit 0", "check_vi_mock_specifiers": "PASS — exit 0", "check_governed_queue_guard": "NOT GOVERNED — 3 paths checked against 5 governed surfaces, none matched", "check_sdui_registration_pins": "NOT MEASURED — 'No console build to weigh at apps/console/dist/assets'. Missing prerequisite (an apps/console build this worktree never produced), not a verdict on this diff. Left to CI", "required_nine_ci": "in_progress — Lint, Type Check, Test (shard 1/4..4/4), Build & E2E, Build Docs, Changeset Declaration all run on the farm. NOT MEASURED locally by design; CI convergence is the PM's read. Bundle Analysis is the standing red on main's own tip and is outside the nine" }, "tests": "RED/GREEN measured twice in both directions. (1) Before the implementation existed, new pin only, on 7098eed: 'Test Files 1 failed', 'Tests 2 failed | 5 passed (7)' — reds were 'HALF 1 — KEEPS the declared nameField row' and 'hides the row the TEMPLATE names, not the one the declared pointer names'. (2) After, at 425cd85: 'Test Files 1 passed (1)', 'Tests 7 passed (7)'. (3) ABLATION from the committed state with on-disk proof and a trap-guarded restore: the rung's read site neutralised to `const interpolatedTitle = ''`, anchor occurrences 1 to 0, injected-marker count 1, `git hash-object` differs from the HEAD blob — run gave 'Tests 2 failed | 5 passed (7)', the same two cases. Restore verified by STATE not exit code: `git diff HEAD` empty AND `git hash-object` equal to the HEAD blob. No dist on the path (the pin imports `../record-details` relatively, same-package source), so no rebuild leg was owed. Three of the remaining five cases are controls that pass on BOTH sides — no-titleFormat object dedupes as before, a template resolving to nothing leaves the ladder alone, a single-field `{name}` template still dedupes — so a change that merely disabled the H1 dedupe would not pass.", "line_budget": "n/a for the skills ratchet — the diff touches no `skills/**` and no `.claude/**` (measured 0). Diff size vs base 7098eed: 350 insertions / 12 deletions across 3 files — `record-details.tsx` +76/-12 (the rung plus the ladder docblock, which had recorded BOTH divergences as 'deliberately NOT repaired here' and now has to say which one is), the new pin +234 (7 cases), the changeset +40.", "deviations": [ "Dedup channel: the dispatch said REST list endpoints plus a local grep INCLUDING closed issues. Paging closed to exhaustion is a broad sweep my standing rules forbid (over 4000 closed, still 100-per-page at page 40), so I stopped and switched to targeted probes that answer the same question completely: ALL 429 open issues (paged to a short page of 48), ALL 25 `needs-user-decision` issues and PRs at state=all, this card's own cross-reference timeline, plus the 1902 closed issues already fetched. A pre-existing C card would be OPEN and carry `needs-user-decision`; both of those corpora are COMPLETE, so the zero is sound for the question asked. The closed half is explicitly partial and is declared as such on objectui#9436. Two control words fired on objectui#8351 in the same pass: 'titleFormat still outranks nameField' returned 1, 'PageHeaderRenderer ranks' returned 1.", "Reading of 'fully interpolates' (ZONE 2 item 2 assigned me this): spelled as 'the rendered template is not any single candidate field's value', tested with `formatTitleTemplate` + `recordDisplayValueAt`, rather than as the bare presence of a resolving `titleFormat`. Two measured reasons. (a) The presence-only reading re-creates the exact duplication Phase P.0 exists to remove: on `{contract_no} - {name}` with a blank `name`, the dropped placeholder takes its orphan separator and the H1 reads exactly `HT-2026-003`, so the grid would print 'Contract No: HT-2026-003' directly beneath an identical H1. (b) It makes the suggested route's own lit control unconstructible — 'a titleFormat that does NOT fully interpolate, where the ladder must still hide' only exists under a reading where partial interpolation still hides. Both readings agree on every genuinely multi-field fully-interpolating template, which is the case the ruling states. Flagging it because it is an interpretation of ZONE 1 text, and the ruling's sentence does not enumerate the single-field or partial-collapse sub-cases.", "The match is a SCAN of the candidate list, not a peek at the first candidate with a value. With `titleFormat: '{name}'` over `nameField: 'contract_no'` the first resolving candidate is `contract_no` while the H1 is `name`'s value; stopping early hides the wrong row AND leaves the real duplicate. Pinned as its own case.", "Toolchain: my first closure build passed `--concurrency=2` AFTER `--filter`, so pnpm forwarded it into `tsup` and `packages/data-objectstack` failed. Re-run as `pnpm --workspace-concurrency=2 --filter ... build` — this was my error, not a repo defect." ], "files_changed": [ "packages/plugin-detail/src/renderers/record-details.tsx", "packages/plugin-detail/src/renderers/__tests__/record-details.titleFormatNoDedupe-8351.test.tsx", ".changeset/8351-titleformat-hides-no-row.md" ], "mcp_calls": "0 — every GitHub read and write went through repo-scoped REST (probe returned HTTP 200) or git; no MCP GitHub tool was called", "open_questions": [], "out_of_scope_findings": [ "filed as objectui#9436: option C — PageHeaderRenderer ranks `titleFormat` above the ADR-0079 declared pointer while getRecordDisplayName ranks it below; carries `needs-user-decision` / `domain:ui` / `priority:p2`, states the user-visible H1 movement, and names the green pin it retires (`page-header-title.test.tsx`, case `titleFormat still outranks nameField`). Three options with four-facet analysis and a stated confidence gap (nobody has measured how many real objects declare both keys).", "noted, not filed: `@object-ui/components`' local `interpolate` and `@object-ui/core`'s `formatTitleTemplate` are TWO different interpolators for the same `titleFormat` — the header matches `{[a-zA-Z0-9_.]+}` then refuses any result with a surviving `{`, while core matches `{[^{}]+}`, normalises handlebars first and strips orphan separators. The two can disagree about whether the template rung resolved. Successor: objectui#9436, whose whole subject is that code region and which must weigh both interpolators either way it is ruled — context that card needs, not a card of its own.", "noted, not filed: `DetailView.resolveDisplayTitle` is a THIRD ranking of the same rungs inside plugin-detail (primaryField, then titleFormat, then the unified resolver, then schema.title). Its `schema.primaryField` read is off the VIEW schema and is legitimate, explicitly protected by objectui#7586 — so this is an observation about how many orderings exist, not a defect. Successor: objectui#9436, the same ruling decides whether three orderings collapse to one.", "noted, not filed: `page:header`'s own `schema.title` is the third rung that names no field. Excluded by the ruling by name and unanswerable from this package (it is a key on the header schema, which `record:details` never receives). Successor: none — the ruling says it gets its own card if it is wanted, and that is the maintainer's call, not mine to pre-empt." ] }Generated by Claude Code
os-tesla commentedon Sep 13, 2026
ACCEPT — PR objectui#9437 · reviewed against GitHub, ⛔ not against the report
domain:uiseat,session_011QreXiyMEqKLN4U5daMPVa, R7, 2026-09-13T21:5xZ. Form: draft, basemain,Fixes #8351first line, 3 files+350/-12, governed guard NOT GOVERNED.The deferral warning was checked, and it came back reassuring — verified independently
I deferred this card behind objectui#9299 and warned that
plugin-detailwould move under it, so the ladder reasoning had to be re-read. The dev checked and found the ladder did not move. ⭐ Verified here rather than taken:⇒ the control is what makes that a reading: objectui#9299 did land in the package, just not in this file. The warning was right to issue and the check was right to run.
⭐ The interpretation deviation — accepted, because it follows the ruling's own REASON
The dev read "fully interpolates" as "the rendered template is not any single candidate field's value", rather than "a resolving
titleFormatis present", and flagged it as an interpretation of ZONE 1 rather than absorbing it.It is the right reading, and the ruling settles it in its own sentence: "the dedupe ladder hides nothing — the H1 is not any single field's value, so there is no row to hide." The ruling states its condition and its reason, and the reason is precisely the dev's test. A presence-only reading would satisfy the condition while violating the reason.
Its measurement makes that concrete: on
{contract_no} - {name}with a blankname, the dropped placeholder takes its orphan separator and the H1 reads exactlyHT-2026-003— so the grid would printContract No: HT-2026-003directly beneath an identical H1, re-creating the duplication the ladder exists to remove.titleFormatthat does NOT fully interpolate, where the ladder must still hide" — which is unconstructible under the presence-only reading. My own suggestion silently assumed the dev's interpretation. Recorded as the second place this seat's wording was looser than the ruling.Other checks
titleFormat: '{name}'overnameField: 'contract_no', the first resolving candidate iscontract_nowhile the H1 isname's value ⇒ stopping early would hide the wrong row and leave the real duplicate. Pinned as its own case. That is the kind of case a reviewer never thinks to ask for.git diff HEADempty and blob equality, ⛔ not by an exit code. Three of the five remaining cases are controls that pass on both sides, so a change that merely disabled the H1 dedupe would not have passed.needs-user-decision, naming the green pin it retires (page-header-title.test.tsx, casetitleFormat still outranks nameField) and the user-visible H1 movement — ⛔ not ridden here, exactly as the ruling required.Deviations — accepted
The dedupe was bounded and the bound declared. Rather than page 4000+ closed issues (a broad sweep its standing rules forbid), it ran targeted probes that answer the same question completely: all 429 open issues paged to a short page, all 25
needs-user-decisionatstate=all, this card's cross-reference timeline, plus 1902 closed already fetched. ⭐ The argument is sound — a pre-existing C card would be open and carryneeds-user-decision, and both of those corpora are complete — and the partial closed half is declared on objectui#9436 itself, where the next reader meets it. Two control words fired in the same pass.A toolchain slip (
--concurrency=2after--filter, forwarded intotsup) was reported as the dev's own error rather than a repo defect.domain:uiandpriority:p2alongsideneeds-user-decision. Producingdomain:*and a grade is triage's alone — ⛔ this seat neither applied nor stripped them (stripping would itself be a grading act). Same flag as objectui#9427 earlier this shift; it is becoming a pattern worth a word from triage.Landing: required set converging. Ready + auto-merge on a full green read.
Generated by Claude Code
fix(components,plugin-detail): the declared nameField outranks titleF…