Uh oh!
There was an error while loading. Please reload this page.
feat(spec): icontains joins the view and infix filter vocabularies (#8934) - #9020
Merged
Merged
Conversation
…8934) Maintainer ruling (Option A on #8934): the case-insensitive contains every driver executes as $icontains (#5702/#6520) becomes authorable from all three filter dialects instead of exactly one. - ui/view.zod.ts: VIEW_FILTER_OPERATORS gains 'icontains'; no alias rows — the alias table bridges spellings already in stored metadata, and a new canonical operator has none. - data/filter.zod.ts: AST_OPERATOR_MAP gains 'icontains' -> '$icontains' (extending the #7536 boundary prose: escaped substring, never a raw pattern, never aliased onto ilike/$ilike in either direction); matching CANONICAL_INFIX row so canonicalAstOperator round-trips generically. - filter-view-operator-parity.test.ts: #8934 block — lowering, %-literal comparand, no ilike/contains collapse, no negative form. - filter-text-conformance.ts: one case computed THROUGH the lowering, so a future icontains->$ilike fold goes red on every backend running the table. - No not_icontains: the $ dialect has no $notIcontains; the authoring vocabularies mirror the executed set. - content/docs/references regenerated via check:generated --fix (gen:docs was the one proved-stale artifact; authorable-surface key set unchanged). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01225pUjnCKWqxcc1PeqKFUq
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
Contributor
📓 Docs Drift CheckThis PR changes 1 package(s): 106 hand-written doc(s) reference the affected code and may need an implementation-accuracy re-verification:
⛔ 7 release-owned page(s) also reference the affected code. These are read-only:
|
…s after #8934 The new FILTER_TEXT_CASES entry (infix icontains, %-literal) lands as $icontains post-lowering, so it joins the analytics face's expressible subset automatically — the count pin moves 11 -> 12, exactly the mechanism the selection note promises. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01225pUjnCKWqxcc1PeqKFUq
…ntains-dialect-parity
…8934) os-regen relay: merge of origin/main (6468df1) committed first with main's side of the generated artifacts; gen:docs re-run on the merged tree restores the icontains rows into main's current references. gen:openapi re-run after (the gen:schema-cleanup sister trap). check:generated: all 13 up to date. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01225pUjnCKWqxcc1PeqKFUq
os-zhuang
marked this pull request as ready for review
August 16, 2026 07:31
Uh oh!
There was an error while loading. Please reload this page.
This was referenced Aug 16, 2026
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 freeto 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#8934
Implements the maintainer's Option A ruling (issue comment, 2026-08-16T04:25Z): the case-insensitive contains that every driver and evaluation face already executes as
$icontains(#5702/#6520) becomes authorable from all three filter dialects instead of exactly one.What changed
packages/spec/src/ui/view.zod.ts—VIEW_FILTER_OPERATORSgainsicontains. No alias rows, deliberately: the alias-table convention, measured, is a migration bridge for spellings already living in stored metadata (historical shorthand likeeq, camelCase likenotEquals); a canonical operator born after the table has no legacy spellings to bridge, and single-token canonicals (contains,in,between) carry no folds either. Inventing synonyms would widen the authoring surface, not bridge a legacy one. The curation decision is recorded on the table's docblock.packages/spec/src/data/filter.zod.ts—AST_OPERATOR_MAPgainsicontains→$icontains(soVALID_AST_OPERATORS,isFilterASTandparseFilterASTall pick it up — the set is derived from the map's keys, verified: no second hand-maintained set exists). The new row extends the Filter AST:likeis folded to$containsat the wire — wildcards bind as literals and driver-sql'slike/ilikearm is unreachable #7536 boundary prose in place rather than duplicating it:$icontainsLIKE-escapes the comparand (a caller's%is a LITERAL),$ilikepasses a raw pattern — a semantic boundary, so neither spelling is aliased onto the other in either vocabulary. A matchingCANONICAL_INFIXrow makescanonicalAstOperatorround-trip it through the generic path (same shape as$like/$ilike) instead of the not-a-known-operator fallback.not_icontains— ruled out by name: the$dialect has no$notIcontains, and the authoring vocabularies mirror the executed set rather than widening it. Pinned by test.filter-view-operator-parity.test.ts—$icontainsjoins the real-lowering KNOWN set; new$icontainshas no counterpart inVIEW_FILTER_OPERATORSorVALID_AST_OPERATORS, so it is authorable only in the MongoDB-style dialect #8934 block pins: the lowering (['name','icontains','acme']→{ name: { $icontains: 'acme' } }), case-insensitive spelling fold, the%-literal comparand case the ruling required, no collapse ontoilikeorcontainsin either direction, generic canonicalisation round-trip, and the absence of a negative form.filter-text-conformance.ts— one new case computed THROUGHparseFilterAST(['name','icontains','100%']). Today it is byte-equal at runtime to the existing$icontains%-literal case, on purpose: if the infix spelling is ever folded onto$ilike, the raw-pattern reading of100%also matches the100X matchfixture row and the case goes red on every backend that imports the table (all five drivers plus objectql's having face) — the boundary is enforced where it executes.driver-memory/src/memory-filter-text-conformance.test.ts— the analytics-face case-count pin moves 11 → 12: the new table case is$icontainspost-lowering, so it joins that face's expressible subset automatically (the exact mechanism the test's own selection note promises). Caught by CI on the first push; count and prose updated with a$icontainshas no counterpart inVIEW_FILTER_OPERATORSorVALID_AST_OPERATORS, so it is authorable only in the MongoDB-style dialect #8934 pointer.$icontainshas been aFieldOperatorsSchemakey since spec:$regex按 ADR-0049 退役 +$icontains入算子词表与 FILTER_LOGIC_CASES(#4706 裁决 B 案 · 契约半边,先行) #5701, soSCALAR_COMPARAND_OPERATORSand the reconciliation test infilter-comparand-type.test.tsalready cover it; the comparand-type conformance table judges TYPE policy on positions, not per-operator spellings (per-operator value rules like non-string$icontainsareFILTER_TEXT_CASES' subject and already pinned there).content/docs/references/ui/*.mdxregenerated viacheck:generated --fix(gen:docswas the one proved-stale artifact; the authorable-surface key set is unchanged — the widening is enum values, not keys).@objectstack/specminor (acceptance-set widening; matches the recent filter-vocabulary changeset convention).Per-face impact (the compile/evaluation faces the two vocabularies feed)
metadata-protocol:isFilterAST→parseFilterAST)icontains, lowers to$icontains; error messages listing the vocabulary are rendered fromVALID_AST_OPERATORSdynamicallyobjectql)$icontainsexecution already landed (#6520)ViewFilterRuleSchema)icontainsparses; value-shape refinement treats it as an unshaped scalar operator (correct — its comparand is a scalar);graftNormalizedOperatorsis identity on a canonical spelling$icontainssince #5702/#6520 — no driver change needed;driver-memory's vocabulary-parity suite (it.each(VALID_AST_OPERATORS)) picks the new spelling up automatically and passes through theparseFilterASTroutedriver-memorylegacy comparison-node face (direct-caller shape{ type: 'comparison', operator }, not reachable through any platform door)icontainshits the loudunsupportedFilterErrordefault arm — a loud refusal, not a silent drop; the platform doors lower the spelling before any driver. Left as-is: extending that legacy face is a driver-scope decision outside this card's surfacelintvalidate-preset-comparandscanonicalAstOperator; judges ordering operators only —icontainsis not one; unaffectedVerification — union run at
ab81100(git rev-parse --short HEADfrom the run itself; working tree clean)pnpm --filter @objectstack/spec test— 406 files / 10753 tests passed (full suite; spec content identical to head).@objectstack/spec✓ (test layer undertsconfig.test.json, debt ledger untouched),@objectstack/driver-memory✓.check:generated— all 13 artifacts up to date after--fix.node scripts/pm/dispatch-gates.mjs, re-derived against the actual diff — it added changeset/docs gates beyond the dispatch list):check:changeset-gate-self-tests,check:cross-package-test-inputs(both spellings),check:docs-audit-scope,check:merge-driver,check:objectui-changeset,check:quick-reference-counts,check:role-word,check:spec-parsed-alias,check:test-source-alias,check:type-source-resolution,check:nul-bytes,check-adr-0087-registration,check-changeset-no-major,check-empty-changeset,check-dev-prereqs,check:doc-formula-expressions,check:query-options-erasure,check:engine-double-contract,check:where-matcher,check:type-check-coverage,check:type-check-debt(ratchet re-measure on the built closure: 33 entries, none above recorded) — all green atab81100.'icontains': '$icontains'map row turned 3 assertions red —VALID_AST_OPERATORS covers every canonical view operator,icontains survives isFilterAST as a bare triple, and the$icontainshas no counterpart inVIEW_FILTER_OPERATORSorVALID_AST_OPERATORS, so it is authorable only in the MongoDB-style dialect #8934 block's vocabulary-membership pin — then the row was restored from the commit (tree byte-clean). Honest nuance: the lowering assertions themselves stay green under that ablation becauseconvertComparison's deliberate$${op}fallback coincidentally produces$icontains; the enforcing guard for the row is theisFilterASTvocabulary gate, and the guard against the dangerous direction (mis-aliasing onto$ilike) is the computed conformance case plus the explicit non-aliasing pins.Generated by Claude Code