fix(plugin-kanban): the card title-dedupe skip set reads the shared name-field resolver, not a key nothing emits - #8433
Merged
Conversation
…ame-field resolver, not a key nothing emits A kanban card printed its record title twice — once as the card heading, once as the first row of the card body. `ObjectKanban` resolves each heading through ADR-0079's `getRecordDisplayName`, then builds `titleFieldsToSkip` so the title field's raw value is not rendered again as a card field. That skip set read `objectDef.NAME_FIELD_KEY`, a key NOTHING produces: `@objectstack/spec@17`'s object schema declares `nameField` (canonical) and `displayNameField` (its deprecated alias), and `NAME_FIELD_KEY` occurs 0 times in the framework tree and 0 times in the published package — this repo reads it only as the last rung of `declaredNameField` in `record-title.ts`, and never emits it. So the read was always `undefined` and the skip set collapsed to its five hard-coded literals (`name` / `full_name` / `title` / `subject` / `display_name`). Any object whose name field is spelled otherwise duplicated its title, which made the defect universal on AI-built apps — whose objects name fields `visit_title`, `owner_name`, `<entity>_name` — and invisible on hand-built objects whose name field is literally `name`. The skip set now reads `@object-ui/core`'s `resolveNameField`, the name-space twin of the resolver that produced the heading, so the two agree on WHICH field titles the object. Deliberately ONE rung, unlike the same dedupe in `record-details.tsx` (objectui#8175), which also carries `deriveTitleField`: that ladder filters a SYNTHESIZED field list, whereas this one filters an AUTHOR-DECLARED `cardFields`, where dropping a field the author asked for is a worse failure than a repeated title. The fourth test is the guard — an object whose declared (`code`) and derived (`owner_name`) pointers disagree — and it goes RED if the second rung is ever added. Closes #8400 Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Contributor
✅ Console Performance Budget
The eager closure is every chunk the entry reaches through static imports — what the browser fetches and parses before the app renders. The entry chunk on its own is a small fraction of it. 📦 Bundle Size Report
Size Limits
|
Contributor
Author
补上你声明缺的那一块：活栈界面证据你在汇报里如实写了「测试是 jsdom 里真实渲染看板卡片 DOM，不是活栈截图……如果你要界面证据，需要在 rig 上跑一遍」。我跑了，修复在真实栈上生效。 装置：把本分支（ DOM 计数（不是看截图数）： // 每个标题字符串在叶子节点里出现的次数
{ "小虎犬牙洁治评估": 1 }, // 修复前 2
{ "豆包皮肤瘙痒复诊": 1 } // 修复前 2卡片正文完整性（证明没有过滤过头）： 标题一次、其余四个字段一个不少。对照 issue 正文里修复前那张卡（标题连着印两遍、然后才是「可乐」），这是同一张卡的前后对照。 你那个「只用一级梯子」的判断，我同意
两处你纠正我的地方，我确认收到
CI 还在跑（Lint / Type Check / 4 个 test shard）。绿了我入队。 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Kanban cards printed the record title twice — once as the card heading, once as the first row of the card body.
Root cause
ObjectKanbanresolves each heading through ADR-0079'sgetRecordDisplayName, then buildstitleFieldsToSkipso the title field's raw value is not rendered again as a card field.packages/plugin-kanban/src/ObjectKanban.tsx:418fed that set fromobjectDef?.NAME_FIELD_KEY— a key nothing produces, so the set collapsed to its five hard-coded literals and the title field flowed straight into the body.The three premises, re-verified independently (not taken from the issue)
NAME_FIELD_KEYgit grep -i NAME_FIELD_KEYinobjectstackat the cloud pinafbf2711and at frameworkmain; positive controlnameFieldmatches in bothgrep -raovernode_modules/@objectstack/spec@17.0.0(dist + json-schema + api-surface)nameFieldanddisplayNameFieldObjectKanban.tsx:418 → :542-549 → :568;titleFieldsToSkipis consulted only in the explicit-cardFieldsbranch, and its other members areexplicitTitleField(unset unless the view declarescardTitle/titleField) plus 5 literalsresolveNameField()(packages/core/src/utils/record-title.ts:346) is this repo's shared name-space resolver, and it is correct: it delegates the declared ladder todeclaredNameField()(:320,nameField ?? displayNameField ?? NAME_FIELD_KEY— the identical??chaingetRecordDisplayNamereads at:551) and then falls through toderiveTitleField().Blast radius: the repo-wide grep
NAME_FIELD_KEYhas exactly two read sites in source:packages/core/src/utils/record-title.ts:321— the correct back-compat fallback. Left alone.packages/plugin-kanban/src/ObjectKanban.tsx:418— the defect. Fixed.Every other occurrence is prose (docblocks, changesets) or the
record-titlealias test. The only sibling of this dedupe,record-details.tsx, was already moved onto the shared resolver by objectui#8175, so no third site needs the same change.Why one rung, not two
objectui#8175 unrolls the same ladder into two candidates —
resolveNameField()andderiveTitleField()— so a declared-but-blank pointer still dedupes against the derivation its value-keyed header fell through to. That second rung is deliberately not copied here:record-detailsfilters a synthesized field list, whereas this set filters an author-declaredcardFields. Carrying the derivation alongside a declared pointer would drop a field the author explicitly asked for whenever the two disagree (nameField: 'code'titles the card while the derivation answersowner_name) — on a four-field card that is a worse failure than a repeated title. The fourth test pins that boundary and goes RED if the second rung is ever added.Ablation (measured, both directions)
New file:
packages/plugin-kanban/src/__tests__/ObjectKanban.nameFieldSkipSet-8400.test.tsx, written to the AI-built shape from the report — an object withnameField: 'visit_title'and a board listingvisit_titleincardFields. Assertions count occurrences rather than asserting presence, because the defect is a duplicate and an absence assertion would bless the wrong fix.Fix reverted (one token:
resolveNameField(objectDef)→objectDef?.NAME_FIELD_KEY):Fix restored:
The one test green in both runs is the positive control (
pet_name/visit_atstill render) — without it every count assertion would be satisfied by a board that dropped the body entirely.Wider verification
vitest run packages/plugin-kanban/— 31 files / 208 tests, all passvitest run packages/core/src/utils packages/plugin-view packages/plugin-detail— 224/225 files pass. The one failure iscore/src/utils/__tests__/date-display.optionsStyle-7745.test.ts(2 assertions), pre-existing and timezone-dependent: it is green underTZ=UTC(15/15) and this diff does not touch that file.turbo run build --filter=@object-ui/plugin-kanban...— 13/13 packages build (this is the real typecheck path; baretsc -pin the package cannot resolve workspace@object-ui/*types without the build).eslinton both changed files — 0 errors (only the file's pre-existingno-explicit-anywarnings).Note on the upstream attribution
cloud#2073 attributed this to
nameFieldnot being persisted intosys_metadata. That attribution is wrong and withdrawn —objectDefcarries the correct answer at this call site; this one line simply did not read it.Closes #8400
🤖 Generated with Claude Code