docs(core): re-state the dataSource.filter note for the two populations the convergence split (objectui#8945) - #11286
Merged
objectstack-fleet[bot] merged 3 commits intoOct 1, 2026
Conversation
…two populations The note said three filter shapes "legitimately reach a renderer here" and typed filter as unknown "rather than the spec's FilterCondition". Since the spec's filter doors converged on the ViewFilterRule array (objectui#6206), that mixes what an author may write (the rule array alone; the record form and AST tuple arrays are refused at ElementDataSourceSchema) with what a renderer may still receive (all three, because metadata at rest is not rewritten). The note now states both, names the spec type as ViewFilterRule[], and keeps the unknown type and its reason. The module example binding writes a rule array instead of the record form. Comments and a patch changeset only. Claude-Session: https://claude.ai/code/session_01VhxTqosz7wn54ahqyxgERT Co-authored-by: Claude <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
|
…used on 17.5.0
The pending objectui#7297 changeset teaches { "assignee": "{record_id}" }
as an element:number filter. @objectstack/spec 17.5.0 refuses that record
form at the filter key (invalid_type, expected array) and accepts the
ViewFilterRule array spelling. Append a dated correction giving the
accepted spelling; frontmatter and existing sentences are unchanged.
Claude-Session: https://claude.ai/code/session_01VhxTqosz7wn54ahqyxgERT
Co-authored-by: Claude <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
Contract reviewServed-tier: Inputs, and nothing else: card #8945 (body and all 11 comments, through the seat's ACCEPT ① Derived judgmentsThe diff publishes no type, export, runtime or accept-set change:
⇒ ①-8 fails the record. What flips it, inside the claim's surface and comments-only: re-word the reason in the ② Semver level
③ Boundary flags
Implemented-by: VERDICT: FAIL Rendered 2026-09-30T23:33Z at CONTRACT_REVIEW_TIER by the isolated reviewer inside |
…rers still see all three filter shapes The re-stated ElementDataSourceConfig note and its changeset said the convergence does not rewrite metadata at rest. The migration entry they cite says the opposite: the D2 conversion page-component-filter-record-to-rule-array rewrites the losslessly mappable record forms and single-level AST tuple arrays on every ObjectStack stored-row read and under os migrate meta --stored, and leaves combinator records, parts with no lossless rule spelling and inline-row components' filters as stored. The reason now says that, plus that this backend-agnostic renderer replays no conversion itself. The conclusion, the unknown type and its cast argument are unchanged. Claude-Session: https://claude.ai/code/session_01VhxTqosz7wn54ahqyxgERT Co-authored-by: Claude <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
Contract reviewServed-tier: Inputs, and nothing else: card #8945 (body and all 13 comments, through the seat's round-3 acceptance ① Derived judgmentsThe diff publishes no type, export, runtime or accept-set change:
② Semver level
③ Boundary flags
Implemented-by: VERDICT: PASS Rendered 2026-09-30T23:49Z at CONTRACT_REVIEW_TIER by the isolated reviewer inside |
objectstack-fleet
Bot
deleted the
claude/issue-8945-element-datasource-filter-note
branch
October 1, 2026 00:08
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
2 participants
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.
Fixes #8945
Clause-②: no
This PR re-states the
ElementDataSourceConfigdocblock in@object-ui/corefor the two populations that the upstreamfilterconvergence split apart. It changes comments and adds a patch changeset. No type, export, runtime behaviour or accepted set moves. Dispatched by thedomain:spec @ objectuiseat, sessionhttps://claude.ai/code/session_01VhxTqosz7wn54ahqyxgERT.What changed
packages/core/src/data-scope/element-data-source.ts, theElementDataSourceConfigdocblock. It said three shapes "legitimately reach a renderer here". It also typedfilterasunknown"rather than the spec'sFilterCondition". The docblock now separates two populations:ViewFilterRulearray alone.ElementDataSourceSchemarefuses the record form by kind. It refuses an AST tuple array at its first member, because each member must be a rule object.page-component-filter-record-to-rule-arrayrewrites a storedfilteronly where the rule array spells it losslessly (a flat record, an operator object whose operators the rule vocabulary spells, several such keys, or a single-level AST tuple array), on every ObjectStack stored-row read and underos migrate meta --stored. It leaves exactly as stored a filter carrying$and/$or/$not, any filter with a part that has no lossless rule spelling, and every filter of a component whose rows are inline. This renderer is backend-agnostic and replays no conversion itself, so all three shapes may still arrive, andmergeFilterNodesstill lowers each shape. (Corrected in round 3 ata9dd51ca0, answering contract review5921608196①-8.)unknowntype stays, and so does its reason: narrowing the type would only move the cast. The spec's type is now namedViewFilterRule[], notFilterCondition."filter": { … }, which is the record form. It now writes"filter": [ … ]. This is a bounded in-place fix outside the docblock the claim named. See "File surface" below..changeset/8945-element-datasource-filter-note.mdadds a patch changeset for@object-ui/core.Premises, measured on this worktree's install
The install is
@objectstack/spec17.5.0, and the branch is based one420df310.Restart-when:probe, run as written from the worktree root, exits 0. Its controls are inside the predicate, so a dark instrument would have exited 9. A widerElementDataSourceSchema.safeParsereading:{object}is ACCEPTED. This is a control.unrecognized_keys. This is a control.{ status: 'open' }is REFUSED withinvalid_typeatfilter, expected an array.{ amount: { $gt: 5 } }is REFUSED the same way.[{ field, operator, value }]is ACCEPTED.[['status','=','open']]is REFUSED withinvalid_typeatfilter.0, expected an object.e420df310. The installed.d.tsdeclaresfilteras an optional array ofViewFilterRuleobjects, and the spec's owndescribe()text for the key names the rule array and says the record form is refused.record-picker.tsxis untouched. I read the three sites the card named, at the card's pin and onmain:filterinput is DECLARED carries no shape claim.filterinput was the site that describedFilterConditionat the pin. It now declarestype: 'array'and a rule-array description.ElementRecordPickerPropsSchemaon 17.5.0. The rule array is ACCEPTED. The record form and$andrecord are REFUSED atfilter, expected an array. A tuple member is REFUSED atfilter.0. The controls are{object}, ACCEPTED, and an unknown key, REFUSED. No sentence there misstates the door.File surface
The claim's surface is the
ElementDataSourceConfigdocblock, arecord-picker.tsxsentence only if one still misstates the door, and one patch changeset. The diff also touches one token in the same file's module docblock: the example binding's"filter": { … }becomes[ … ]. All four bounded-fix conditions hold:Leaving the example as it was would have made the file contradict its own re-stated note. The seat owes a same-round addition of this site to the claim's file surface.
Gates, all run at HEAD
4837c74e6pnpm --filter '@object-ui/core^...' build, under the verify lock: VERDICT command-exit 0. The scope was 2 of 47 projects.pnpm --filter @object-ui/core type-check, under the lock: exit 0. The script name is echoed and bothtsclegs ran, with 0 TS errors.pnpm exec vitest run --maxWorkers=2 packages/core/, under the lock: exit 0.Test Files 193 passed (193),Tests 3797 passed | 27 skipped (3824).pnpm --filter @object-ui/core lint: exit 0. It reported 0 errors and 568 pre-existing warnings, none of them in the touched file.check:*gates all exited 0:check:control-bytescheck:new-line-citations, which printed0 new citation(s), compared against merge basee420df310over 2 files.check:element-data-source-declarationcheck:spec-symbolscheck:installed-pin-claimscheck:changeset-claims, which printedNo pending changeset names a file this change touches.check:pending-changeset-literalscheck:comment-mask-corpus. Its one disagreement is inapps/console/src/pages/DocsIndex.tsx, not in this diff.node scripts/check-changeset-presence.mjs: exit 0. It found 1 released source file and 1 changeset added.node scripts/check-changeset-no-major.mjs: exit 0.node scripts/check-governed-queue-guard.mjs --teston both paths printedNOT GOVERNED.Author population: the card's open measurement
This is a historical reading at HEAD
4837c74e6on spec 17.5.0. Nothing re-derives it.Method. It parses candidates, and no regex is involved.
git ls-tree -r HEADwith a clean-tree assertion. That gives 6742.ts/.tsx/.js/.jsx/.mjs/.cjs/.mts/.cts/.jsonfiles; all 6742 parsed and 0 failed.dataSourceproperty, adataSource={{…}}JSX attribute, or anX.dataSource =assignment, and that carries afilterkey.ElementDataSourceSchema.safeParsethree times: on the value withfilterremoved, which decides whether it is a binding at all; on the full value; and onfilteralone.The instrument controls, rule array ACCEPTED and record REFUSED, are lit. A positive control is also lit: the eight fixtures listed on objectui#11117 were all found, 8 of 8.
Reading. There are 36 candidates, all in test sources and none in non-test source.
equalswith an array comparand. It is a deliberate objectui#8557 fixture, not an orthography author.filterand were not measured.invalid_typeatobject, deliberately..mdand.mdxfences are outside this instrument; they belong to objectui#11117.The upstream entry's "seventeen (fifteen tuple arrays, two records)" was taken at a different pin. It is not comparable, and this PR does not assert a rival reading of it.
These are the 18 files and how many retired-orthography fixtures each holds:
packages/components/src/__tests__/record-picker-element-data-source.test.tsx: 2 tuplepackages/components/src/__tests__/record-picker-malformed-filter-10789.test.tsx: 2 recordpackages/plugin-charts/src/ObjectChart.elementDataSource.test.tsx: 1 tuplepackages/plugin-dashboard/src/ObjectPivot.elementDataSource.test.tsx: 1 tuplepackages/plugin-detail/src/__tests__/RecordRelatedListRenderer.elementDataSource.test.tsx: 1 tuplepackages/plugin-detail/src/renderers/__tests__/record-related-list.propsRefusal-9963.test.tsx: 1 tuplepackages/plugin-form/src/EmbeddableForm.elementDataSource.test.tsx: 1 tuplepackages/plugin-form/src/LineItemsPanel.elementDataSource.test.tsx: 1 tuplepackages/plugin-form/src/MasterDetailForm.elementDataSource.test.tsx: 1 tuplepackages/plugin-gantt/src/ObjectGantt.elementDataSource.test.tsx: 1 tuplepackages/plugin-grid/src/__tests__/ObjectGrid.elementDataSource.test.tsx: 2 tuplepackages/plugin-kanban/src/ObjectKanban.elementDataSource.test.tsx: 2 tuplepackages/plugin-list/src/__tests__/ListView.elementDataSource.test.tsx: 1 recordpackages/plugin-list/src/__tests__/ListView.sharedGate.test.tsx: 1 recordpackages/plugin-map/src/ObjectMap.elementDataSource.test.tsx: 1 tuplepackages/plugin-timeline/src/ObjectTimeline.elementDataSource.test.tsx: 2 tuplepackages/react/src/element-data-source/__tests__/ElementDataSourceGate.malformedFilter-10789.test.tsx: 1 recordpackages/react/src/element-data-source/__tests__/ElementDataSourceGate.test.tsx: 2 tuplePatch round 2, at HEAD
0b4445d5a(written by the seat).changeset/7297-record-id-filter-token.mdgains a dated correction, append-only (+2/−0), with frontmatter md52935e1e656cbeb29b6093519d07dde85before and after.element:numberfilter example{ "assignee": "{record_id}" }is refused on@objectstack/spec17.5.0 (invalid_typeatfilter, expected array).ViewFilterRulearray,[{ "field": "assignee", "operator": "equals", "value": "{record_id}" }]. It was measured withElementNumberPropsSchema.safeParseandElementDataSourceSchema.safeParsebefore writing, with both controls lit.5921413381.0b4445d5a:check:changeset-claims,check:pending-changeset-literals,check-changeset-presence,check-changeset-no-major,check:control-bytesandcheck:new-line-citationseach exit 0. No package source changed since4837c74e6, so round 1's@object-ui/corebuild, type-check, test and lint still stand.Acceptance notes
toFilterNodedoc inpackages/core/src/utils/filter-converter.tssays "three shapes in circulation, all legitimate" about what a view hands a renderer. That statement is about the renderer population and stays accurate, so it is not edited.record-picker.tsxis unchanged because H3 held.Generated by Claude Code