@techspokes/typescript-wsdl-client 1.1.4 → 1.1.5

package/docs/content-model-s06-pw-01.md

@@ -0,0 +1,350 @@
+# PW01: Group-to-Wildcard Restriction
+
+Research contract, competing interpretations and independent evidence for S06-PW-01. See the root [README](../README.md).
+
+## Disposition and decision request
+
+Recovery #270 moves the shared group-to-wildcard control flow into `src/compiler/particleWildcardRelation.ts`.
+The fixed `restrictParticleToWildcard` source entry uses the caller's authenticated context/budget, accepted #178 exact arithmetic and #269's normative wildcard relation; the historical research entry becomes an adapter.
+Its result is a conditional local relation with explicit normalized-view, occurrence, construction and attribution prerequisites, not full particle/schema validity.
+
+The source entry does not infer the actual ur-type wildcard identity from caller-supplied spelling.
+Recovery now implements normalization, full restriction dispatch, raw extension and UPA/EDC in separate shared owners, while aggregate candidate acceptance remains #271.
+The complete universal witness theorem remains #234.
+
+### Recovery particle authorities
+
+`particleSemanticView.ts` distinguishes authenticated declared views from restriction-normalized comparison views.
+It uses the existing #178 occurrence implementation with the caller's shared budget; it does not replace formal ranges with accepted-count intervals.
+`particleRestriction.ts` implements the specified dispatch and ordered/unmatched mappings, using the shared scalar/type and wildcard owners.
+The [accepted implicit-order interpretation](content-model-s06-re-01.md#selected-raw-matching-implicit-order-and-generated-scalar-properties) asks whether some coherent enumeration satisfies the ordered mapping.
+Explicit lists keep their original order, and one implicit head uses one enumeration throughout a root comparison under the shared budget.
+
+`particleExtension.ts` checks the original raw extension relation.
+The same particle identity is a primitive case; a unit sequence with the required recursively equal first member is checked through concrete component properties.
+Annotations and diagnostic storage do not substitute for semantic properties, and normalized language equivalence does not establish raw extension.
+`schemaComponentEquality.ts` owns the recursive property comparison, with scalar and component-validity prerequisites kept explicit.
+
+`particleAttribution.ts` constructs one marked counter automaton from particle-use positions and exact occurrence bounds.
+Finite counts remain `bigint` guards and counter values; numeric bounds do not duplicate event structure.
+Unbounded counters saturate only after their minimum, where the transition guards no longer distinguish larger values.
+The finite alphabet partition distinguishes named elements, remaining names in each mentioned namespace and unmentioned namespaces.
+Effective substitution members and wildcard namespace operations use the same context-qualified owners as restriction.
+Attribution checks overlapping marked transitions at every reachable prefix, including branches whose later required suffix cannot complete.
+The normative requirement to choose without the remainder is corroborated by Appendix H's untrimmed automaton construction; Appendix H is explicitly nonnormative guidance.
+A suffix behind an unreachable required empty choice has no reachable transition and does not create a prefix ambiguity.
+
+Determinization explores exact counter configurations under the shared budget.
+Reviewed acceleration skips only a deterministic counter cycle below its next minimum guard, with the same consuming position, unchanged other counters and no intervening reset or exit.
+It is not an arbitrary cap on repetition counts.
+All finite counters have finite value domains, unbounded counters have the justified saturated domain, and the structural state graph is finite; exploration therefore terminates with sufficient resources.
+Resource interruption yields no UPA verdict or accepted candidate.
+
+`particleDeclarationConsistency.ts` checks EDC across positions, including implicit substitution members and named-type identity.
+`particleLocalAssessment.ts` combines local all/UPA/EDC checks without assuming the containing schema is valid.
+The aggregate owner also checks unused original named groups and compositors; legality cannot depend on whether a complex type happens to reference them.
+
+Independent restriction, extension, EDC and UPA suites use real canonical/resolved inputs.
+The nested-repeat ambiguity `((a?, b, a?)){1,2}` exposed an unsound path heuristic; exact counted attribution now checks that contrast and huge-range boundaries without event duplication.
+Finite language comparisons remain separate bounded review evidence, not proofs of full RE01 coverage.
+
+[Issue #235](https://github.com/TechSpokes/typescript-wsdl-client/issues/235) has a complete research contract and executable evidence.
+The maintainer [selected option B on October 10, 2026](https://github.com/TechSpokes/typescript-wsdl-client/issues/235#issuecomment-6095922074); final independent semantic review and delivery govern local acceptance.
+The published XSD 1.0 text and proposed R-240 correction yield different answers for 13 cases below.
+The literal alternative remains comparison evidence, while B is the selected project interpretation.
+
+The selected project interpretation is option B, the proposed R-240 zero member minimum.
+Its authority is a documented WG proposal, not an adopted XSD 1.0 erratum.
+Its explicit selection does not authorize product guard removal before #179 integration.
+
+| Option | Member comparisons | Whole retained group | Compatibility consequence |
+|---|---|---|---|
+| A: Literal 2004 member bounds | Original wildcard minimum and maximum | Original wildcard range | Rejects optional/unit members even when the total fits |
+| B: Proposed R-240 zero member minimum | Zero minimum; original maximum | Original wildcard range | Admits those contrasts; preserves namespace, maximum and processing checks |
+
+Option B is monotone relative to A on the same normalized operands: lowering only member-view minima cannot turn a passing A comparison into a failure.
+It preserves every result in the already discharged original `0..unbounded` subset and additionally settles finite and positive-minimum cases once selected.
+It also admits the formally dead `dead-sequence` contrast; accepting B does not authorize a payload-language claim.
+
+A term-only interpretation that drops both member bounds differs from the actual R-240 proposal and has no adopted wording here.
+Adopting the XSD 1.1 restriction system changes the approved profile and is outside this decision.
+No narrower permanent profile, persisted format change, public activation or removal of the #179 guard is delivered.
+
+The [joint handoff](content-model-s06-research-handoff.md) owns exact delivery/review revisions and final combined validation.
+Production implementation remains #179; [RE01 #234](https://github.com/TechSpokes/typescript-wsdl-client/issues/234) receives the parameterized predicate below.
+#180/#181 retain payload matching, count gaps and independent model enumeration.
+
+## Source authority and applicable clauses
+
+The governing source is [XSD 1.0 Structures, Recommendation 28 October 2004](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/).
+The complete XML source was read through the [pinned mirror](https://github.com/jacoelho/xsd/blob/142f25ee187e17f041af87b6bfcfc82385254520/docs/spec/xml/structures.xml), SHA-256 `e496af408b14853e6169ac7c1fca09d55a81bd73771758ad6be020960e1317ba`.
+Editorial deletions in that source are excluded from the final clauses.
+
+[NSRecurseCheckCardinality](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#rcase-NSRecurseCheckCardinality) invokes “Every member of the {particles} of the group” and separately the “effective total range of the group”.
+Its first clause calls the member a “valid restriction” of the wildcard through `cos-particle-restrict`, although that relation compares particles.
+This particle/term ambiguity is the specific defect recorded by R-240.
+
+| Clause | Operand and required comparison |
+|---|---|
+| [NSCompat](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#rcase-NSCompat) | Element particle: namespace membership and local occurrence range |
+| [NSSubset](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#rcase-NSSubset) | Wildcard particle: occurrence range, namespace subset and processing strength |
+| [NSRecurseCheckCardinality](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#rcase-NSRecurseCheckCardinality) | Group particle: recursively valid members and effective total range |
+| [Occurrence Range OK](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#range-ok) | Exact lower-bound and upper-bound containment |
+| [Sequence/all range](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#cos-seq-range) | Formal sum followed by the group's local repetition |
+| [Choice range](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#cos-choice-range) | Formal alternative extrema followed by local repetition |
+| [Particle restriction](https://www.w3.org/TR/2004/REC-xmlschema-1-20041028/#cos-particle-restrict) | Component mapping and specified pointless-compositor rules precede dispatch |
+
+The [published second-edition errata](https://www.w3.org/2004/03/xmlschema-errata), including E1-56, do not amend this restriction rule.
+[WG issue 2232](https://www.w3.org/Bugs/Public/show_bug.cgi?id=2232), reported September 14, 2005, remains `ASSIGNED` and `1.0 only` in the archived tracker.
+Its proposed constructed comparison particle has “min occurs: 0” and “max occurs: the same as the max occurs of the wildcard particle”.
+
+The September 23, 2006 WG comment says the constraint was deleted from XSD 1.1.
+That statement records a later-version change; it does not adopt either proposed correction into XSD 1.0.
+Official errata and WG pages were independently retrieved through the web connector on October 10, 2026 after direct W3C shell retrieval returned HTTP 403.
+
+## Pure predicate consumed by RE01 and #179
+
+Let `P` be a restriction-normalized element, wildcard or group particle and `W` the original base wildcard particle.
+`within(R, W)` means exact `R.min >= W.min` and `R.max <= W.max`, with `unbounded` a distinct maximum.
+For a group, `ETR(P)` is the formal effective total range supplied by [#178](content-model-analysis.md), not its accepted-count interval.
+
+```text
+restrictToWildcard(P, W, interpretation):
+  element:
+    within(P.localOccurs, W) AND namespaceAllowed(P.elementNamespace, W.term)
+  wildcard:
+    within(P.localOccurs, W)
+    AND namespaceSubset(P.term, W.term)
+    AND (W is actual ur-type content wildcard OR P.process >= W.process)
+  sequence / choice / legal all:
+    within(ETR(P), W)
+    AND every member C satisfies restrictToWildcard(C, memberView(W), interpretation)
+
+memberView(W):
+  option A: W
+  option B: original term/context/max, minimum replaced by zero
+```
+
+Every nested group receives its current member-view wildcard for its own total comparison.
+Under B, only the outer selected group checks the original positive minimum; nested totals and terminal members compare against `0..original W.max`.
+Declared group/use-site/member bounds never change, and maximum checks are never omitted.
+Under A, every recursive comparison retains the original minimum.
+
+The complete preconditions are independently validated/resolved graphs, legal source-local ranges, UPA/EDC and legal `all` placement, cycle-free containment/group edges, and the owning derivation's other constraints.
+Legal element/type recursion stops at the element terminal; this predicate never descends into its value type.
+`inputNodes` must be the actual immutable accepted graph node count supplied by the input owner, rather than a guessed reachable count.
+
+The caller supplies the restriction-normalized particle view, its #178 formal ranges, original operands/provenance, and exact namespace constraints.
+The base wildcard's identity distinguishes the actual ur-type exception; neither `##any` spelling nor `lax` alone implies that exception.
+A positive relation result certifies this particle relation only, not complete schema legality, runtime support, finite recursive payload existence or language inclusion.
+
+The historical [prototype adapter](../test/research/s06-pw01/predicate.ts) now imports the shared source control flow.
+Its [test adapter](../test/research/s06-pw01/predicate.test.ts) injects accepted main's occurrence arithmetic and namespace relations as the subject under test; all expected outcomes are independently authored in the fixture file.
+Neither file imports the unaccepted #231 assessment code, recalculates effective ranges with a second algorithm, or creates the #181 enumeration harness.
+
+The current [TypeScript reference suite](../test/conformance/reference/s06-pw01-contract.test.ts) also invokes the pure prototype with an independent [fixture syntax adapter](../test/conformance/reference/scoped-pw.ts).
+That adapter computes exact ranges and namespace constraints only for these fixed source fixtures, using original source identities and literal expectations.
+It supplies scoped reference assertions; #179 still consumes the single #178 evaluator for production occurrence analysis.
+
+The prototype returns `answer {validRestriction, steps}` or `failure {diagnostic, steps}`.
+Malformed operand callbacks report `invalid-schema`; exhaustion reports `resource-limit` with original component/source and no relation result.
+A research consumer uses selected option B; production retains its qualification until #179 implements and validates that contract.
+Budget exhaustion must never become an invalidity answer.
+
+## Normalization, emptiness and original operands
+
+XSD component mapping removes zero-maximum particles before the particle relation, while their original syntax remains available for diagnostics and reference/cycle checks.
+`zero-group` maps both contents to empty and is checked by the owning complex-content restriction rule, rather than by pretending an absent wildcard has members.
+The fixed fixture adapter takes that route explicitly.
+
+A required empty choice is not pointless: its formal `ETR=0..0` and `schemaEmptiable=true` coexist with `nullable=false` and `hasRealization=false`.
+Against an original `0..unbounded` wildcard its universal member check is vacuous and its formal total fits, so both candidate relations pass.
+Against a positive-minimum wildcard both fail the original total check.
+
+An epsilon alternative built from a non-disabled optional element remains a real choice alternative.
+`epsilon-choice-failure` therefore retains `ETR=0..1` and fails original minimum 1 under both candidates.
+`nested-epsilon-choice` has outer total `1..2`; its nested `0..1` group fails A and passes B.
+
+`removable-epsilon-control` uses an empty sequence plus a single element alternative.
+The published pointless-compositor rules ignore that empty sequence and singleton choice for restriction, dispatching the surviving element through NSCompat.
+Its original `ETR=0..1` and original epsilon payload alternative remain recorded; both candidate relations pass without inferring payload containment.
+
+#179 must distinguish original grammar/ranges from any normalized comparison view.
+If normalization retains a group but changes its formal range, obtain that view's range from the single #178 evaluator over a non-mutating comparison graph; do not reuse a stale original range or introduce another occurrence algorithm.
+The fixed probe's retained groups have unchanged ranges; singleton dispatch is independently covered by `removable-epsilon-control` and `zero-disabled-choice`.
+
+Named-group definition bounds stay `1..1`; a group reference supplies its own repetition and provenance.
+`shared-group-uses` and `repeated-group-use` cover shared definitions and different use-site multiplicities without expanding finite repetitions.
+Forbidden group cycles remain S05 invalid schemas even beneath zero bounds; element/type recursion stays legal.
+
+## Namespace and processing requirements
+
+The base wildcard is decoded using its original lexical/effective namespace context, including imports and chameleon inclusion.
+`##other` excludes that effective namespace and the absent namespace; it is never rebound to the restricting schema's namespace.
+Imported `##other` therefore accepts `urn:pw01`, excludes `urn:pw01:base` and excludes unqualified elements in the three imported controls.
+
+The chameleon control binds `##other` to its included effective `urn:pw01` namespace and rejects the caller's qualified members.
+Prefixes, declaration identity, source path and the original lexical wildcard are retained beside the decoded namespace constraint.
+No flattened S05 attribute-wildcard processing plan is substituted for these particle operands.
+
+NSCompat imposes namespace and occurrence comparisons on an element particle; it does not compare an element's nonexistent `processContents` property.
+Nested wildcard members use NSSubset, with `strict > lax > skip` and the precise ur-type exception.
+Optional members still require namespace and wildcard processing checks; their zero minimum does not excuse a disallowed namespace or weakened mode.
+
+The existing [S05 imported/processing evidence](content-model-composition.md#independent-disagreements-and-remaining-obligations) remains separate, particularly attribute-group/local wildcard processing qualifications.
+These PW01 particle controls do not settle those other obligations or change their owners.
+
+## Independent fixtures and engine investigation
+
+[The fixed fixture expectations](../test/conformance/fixtures/xsd/s06-pw01/expectations.json) separate `literal2004`, `proposed2232`, `primarySchema`, and historical engine answers.
+`primarySchema=unresolved` means the candidate rules differ; neither engine observation fills that field.
+Payload primary entries describe the original particle language independently of whether a validator accepted the derivation declaration.
+
+The exact [pinned nested fixture](../test/conformance/fixtures/xsd/s06-pw01/pinned-nested-optional.xsd) was copied from #231 revision `460f5b8379c68ffef79917284e445b5ab046429e`, path `test/conformance/fixtures/xsd/assessment/nested-group-wildcard-cardinality.xsd`.
+Its SHA-256 is `1795040f9bdb7b592f389340262167d56927e9ec4a5735b1955366e305ed93ca`.
+All other fixtures in this family are independently authored research contrasts.
+
+Historical XMLSchema 4.2.0 observations accept `direct-optional-member` but reject the otherwise equivalent `nested-optional-member` and the pinned fixture.
+Its [groups.py at v4.2.0](https://github.com/sissaschool/xmlschema/blob/v4.2.0/xmlschema/validators/groups.py) `XsdGroup.is_element_restriction`, lines 685-714, first checks group occurrence compatibility, then rejects a retained `XsdGroup` member at lines 697-698.
+The nested fixture reaches that branch with direct members `[sequence 0..1, element 1..1, element 1..1]`; the rejection is not proof that the published rule requires each leaf minimum 2.
+
+That same method bypasses member restriction checking for a zero-minimum terminal at line 699.
+The optional namespace and wildcard-strength controls expose its resulting acceptance of primary-invalid comparisons.
+Empty-group early success also explains acceptance of `empty-required-positive-min` despite its zero formal total.
+
+Historical libxml2 2.14.6 observations accept the nested fixture and every ordinary negative particle restriction control here.
+Its [versioned xmlschemas.c](https://github.com/GNOME/libxml2/blob/v2.14.6/xmlschemas.c), lines 16521-16528, leaves the owning derivation particle check as an urgent TODO.
+The separate particle restriction code is guarded by `ENABLE_PARTICLE_RESTRICTION`; `xmlSchemaCheckRCaseNSRecurseCheckCardinality` is additionally under `#if 0` at lines 17026-17079.
+
+The fetched libxml2 source SHA-256 is `a74cd24a7ab2835a27abd5a0da7ff9343301aa284d29ef5ecd0b725fd00f433c`.
+The historically inspected XMLSchema `groups.py` SHA-256 is `c49a47da49d3146bbfd692da02e138f084cd50b8d52212e24cd4b63d2de9e063`.
+Source inspection and independently executed negative controls explain the disagreement without promoting either validator to semantic authority.
+
+Historical libxml2 2.14.6 rejected the huge exact fixture during `maxOccurs` parsing; the archived reference record calls this `parse-range-limit` rather than a validity verdict.
+The source integer is legal and exact; the TypeScript predicate and accepted #178 arithmetic decide the proposed comparison without repetition expansion.
+Reference engines' finite representations do not redefine the schema's allowed integer domain.
+
+### Fixed schema truth table
+
+`A/B` are relation outcomes after owning component mapping and restriction normalization.
+`ETR` is the independently specified original formal range; the removable epsilon row deliberately dispatches a surviving element instead.
+Engine columns retain historical schema-load observations, including cases both candidates classify invalid.
+
+| Case | Original ETR | A | B | XMLSchema 4.2.0 | libxml2 2.14.6 |
+|---|---|---|---|---|---|
+| `zero-min-control` | 2..3 | true | true | accepted | accepted |
+| `direct-optional-member` | 2..3 | false | true | accepted | accepted |
+| `nested-optional-member` | 2..3 | false | true | rejected | accepted |
+| `literal-positive-control` | 4..4 | true | true | accepted | accepted |
+| `whole-minimum-failure` | 1..2 | false | false | rejected | accepted |
+| `whole-maximum-failure` | 4..4 | false | false | rejected | accepted |
+| `choice-positive-control` | 2..3 | true | true | accepted | accepted |
+| `repeated-choice` | 2..2 | false | true | accepted | accepted |
+| `epsilon-choice-failure` | 0..1 | false | false | rejected | accepted |
+| `nested-epsilon-choice` | 1..2 | false | true | rejected | accepted |
+| `removable-epsilon-control` | 0..1 | true | true | accepted | accepted |
+| `required-wildcard-strength-failure` | 2..2 | false | false | rejected | accepted |
+| `optional-namespace-failure` | 1..2 | false | false | accepted | accepted |
+| `all-optional-member` | 2..3 | false | true | accepted | accepted |
+| `all-positive-control` | 2..2 | true | true | accepted | accepted |
+| `repeated-sequence` | 4..6 | false | true | rejected | accepted |
+| `shared-group-uses` | 4..4 | false | true | rejected | accepted |
+| `count-gap` | 2..4 | false | true | accepted | accepted |
+| `unbounded-control` | 1..unbounded | true | true | rejected | accepted |
+| `unbounded-member-failure` | 1..unbounded | false | false | rejected | accepted |
+| `empty-required-choice` | 0..0 | true | true | accepted | accepted |
+| `empty-required-positive-min` | 0..0 | false | false | accepted | accepted |
+| `dead-sequence` | 2..2 | false | true | accepted | accepted |
+| `zero-disabled-choice` | 1..1 | true | true | rejected | accepted |
+| `zero-group` | 0..0 | true | true | rejected | accepted |
+| `wildcard-strength-valid` | 1..2 | true | true | accepted | accepted |
+| `wildcard-strength-failure` | 1..2 | false | false | accepted | accepted |
+| `namespace-failure` | 2..2 | false | false | rejected | accepted |
+| `recursive-element-control` | 1..2 | true | true | accepted | accepted |
+| `huge-exact-repetition` | 2H..2H | false | true | rejected | parse-range-limit |
+| `imported-other-valid` | 2..2 | true | true | accepted | accepted |
+| `imported-other-base-failure` | 2..2 | false | false | rejected | accepted |
+| `imported-other-absent-failure` | 2..2 | false | false | rejected | accepted |
+| `chameleon-other-failure` | 2..2 | false | false | rejected | accepted |
+| `pinned-nested-optional` | 2..3 | false | true | rejected | accepted |
+| `all-optional-root-failure` | 0..2 | false | false | rejected | accepted |
+| `positive-unbounded-members` | 2..2 | false | true | accepted | accepted |
+| `repeated-group-use` | 4..6 | false | true | rejected | accepted |
+
+`H=900719925474099312345678901234567890`; `2H=1801439850948198624691357802469135780`.
+The count-gap fixture admits counts 2 and 4 while rejecting count 3 despite its formal `2..4` range.
+Its repeated sequence also retains order; no interval proves admission of `a,a,b,b`.
+
+The eight fixed payload contrasts include the required empty choice, the dead sequence, zero content and recursive element values.
+In the historical observations, both engines reject the isolated required empty choice payload, while XMLSchema accepts the dead sequence payload that the primary language rule and libxml2 reject.
+Historical XMLSchema could not run the two zero-group payload checks because it rejected that schema; those two observations remain explicitly `unavailable-schema`, not skipped passes.
+
+### Current reference evidence
+
+The accepted [NT-CONT-01 evidence policy](reference-validation.md) runs `libxml2-wasm@0.7.2` with libxml2 2.15.1 as the current primary, scoped TypeScript assertions of selected option B, and pinned historical external observations.
+The current reference suite attempts all 38 original schemas and validates all eight original payloads after their schemas construct.
+The primary classifies the original huge occurrence as `unsupported-capability`; the selected exact predicate still evaluates its original integer without expansion.
+
+Current primary results, selected contract answers, historical XMLSchema/libxml2 answers and unqualified capabilities have separate identities in the report.
+The historical source and tables remain hash-verified [non-executable data](../test/conformance/reference/legacy-source-snapshot.json).
+Fresh second-engine full-schema construction, including validity of the original huge schema, remains unqualified under the accepted policy.
+
+## Bounded implementation and completeness
+
+For either selected candidate, an iterative postorder traversal checks every retained member and uses supplied exact formal ranges.
+Memoization keys are particle use identity, original base wildcard identity and member-view minimum; there are at most two minimum views per base pair.
+The prototype fixes one original base per invocation and retains that identity implicitly; #179 must include it explicitly in any cross-call cache.
+
+The prevalidated group/containment DAG is finite and acyclic, so each pair finishes after finitely many outgoing members.
+Shared groups reuse established pair results; element/type recursion ends at a terminal.
+Given sufficient resources this procedure exhaustively decides the stated relation for arbitrary legal sequence/choice/all operands, finite or unbounded counts, empty required groups and original namespace contexts.
+
+Every frame push, memo insertion, member visit and exact arithmetic digit comparison is charged before allocation or operation.
+The injected namespace adapter charges character scans, conservative membership work and shared-helper set copying before executing those operations.
+The prototype consumes supplied views; bounded normalization/indexing/copying remains a separately charged #179 integration obligation, not hidden prototype work.
+
+Inclusive defaults remain independently applied `100,000` input nodes and `1,000,000` work steps.
+Exhaustion produces no relation answer and never proves invalidity, even when one candidate finished earlier.
+S03/S04 loading/catalog limits, conservative cyclic depth, D09 owner boundaries, later matcher/payload/output budgets and #208 production qualification remain unchanged.
+
+| Measurement | Result | Charged work | Observed elapsed |
+|---|---|---|---|
+| Exact huge fixture, 10 nodes / 726 allowed steps | B relation true | 726 | 0.041 ms |
+| Same fixture, 725 allowed steps | resource-limit | 725 | 0.060 ms |
+| Same fixture, 9 allowed nodes | resource-limit before traversal | 0 | 0.131 ms |
+| Synthetic source array, 100,000 nodes | B relation true | 726 | 0.087 ms |
+| Synthetic source array, 100,001 nodes | resource-limit before traversal | 0 | 0.096 ms |
+| Synthetic padded callback at default work limit | B relation true | 1,000,000 | 1.700 ms |
+| Supplied 500,000-digit operand beyond default work | resource-limit | 1,000,000 | 10.487 ms |
+
+These are narrow development measurements, not production capacity claims.
+The synthetic node and padded callback rows test the prototype's supplied-count/work contract; their input construction is outside the measured predicate.
+The exact huge fixture is a useful nontrivial successful decision, unlike a probe that returns resource-limit for every difficult operand.
+
+## Reproduction and handoff
+
+Use Node 24 as the supported floor and Node 26 for current-line qualification from a fresh delivery checkout.
+The current [reference setup and evidence contract](reference-validation.md) uses normal `npm ci`; both reference commands discover the strict TypeScript suite.
+The committed literal fixtures and historical source snapshot establish the fixed results without an ignored log.
+
+```bash
+npm ci
+npm run typecheck:research
+npm run typecheck:reference
+npx vitest run test/conformance/reference/s06-pw01-contract.test.ts test/research/s06-pw01
+npm run test:reference:full
+npm run ci
+npm run test:conformance
+```
+
+At the historical leaf checkpoint, scoped TypeScript passed and focused Vitest passed 40 tests in one file.
+The retired reference suite passed two methods with 76 historical schema-load observations and 16 historical payload observations; two are explicit unavailable-schema results.
+The current TypeScript suite preserves those two method families and every historical answer while adding current primary and independent scoped checks.
+
+The historical leaf environment used Node 24.19.0, npm 11.9.0, TypeScript 6.0.3, Vitest 5.0.3, Python 3.12.14, XMLSchema 4.2.0, lxml 6.1.0 and libxml2 2.14.6.
+The exact final combined commands, versions, result counts and reviewer revision belong to the [joint execution record](content-model-s06-research-handoff.md).
+IDE inspections were unavailable; terminal TypeScript, fixture/reference and repository documentation checks supply local verification.
+
+The delivered research does not claim #179 production guard removal, complete S06 acceptance, #180 matching, #181 enumeration, #184 payload enforcement or downstream activation.
+To resume the PW01 part of #179 after local acceptance, implement selected B in the existing assessment path, preserve original/normalized ranges, and requalify the affected selected-operation plans.
+The final issue handoff records independent semantic review, exact delivered revisions and normal delivery acceptance; sibling or parent closure is not a local prerequisite.