[bugfix] fn:xml-to-json: enforce F&O 3.1 §17.4.2 structural validation (+10 XQTS HEAD)

[joewiz]

Summary

Adds per-element-type structural validation to fn:xml-to-json as required by F&O 3.1 §17.4.2 (XML Representation of JSON) and §17.5.4. Before this change, develop silently accepted inputs that violate the spec's structural rules and produced a JSON result rather than raising FOJS0006.

This is a follow-up to #6342 (which closed the walk-from-doc-root sub-cluster of fn-xml-to-json HEAD failures). It closes the over-permissive-validation sub-cluster identified in the 2026-05-10 triage of the FOJS0006 cluster.

What changed

exist-core/src/main/java/org/exist/xquery/functions/fn/FunXmlToJson.java:

• Reject non-whitespace text node children of <map> and <array> elements (whitespace, comments, and processing instructions are still ignored per spec).
• Reject element children of leaf-type elements (<string>, <number>, <boolean>, <null>).
• Reject no-namespace attributes other than key, escaped-key, and escaped; reject any attribute in the http://www.w3.org/2005/xpath-functions namespace (the schema's anyAttribute namespace=\"##other\").
• Require escaped and escaped-key attribute values to be lexically valid xs:boolean (true/false/1/0).
• Reject element names outside the six allowed local names (map, array, string, number, boolean, null) at start-tag rather than only at end-tag.

Per W3C bug 29917 / qt3tests xml-to-json-065, the escaped attribute is tolerated on non-string elements (treated as a no-op); only the lexical value is enforced.

Foreign-namespace attributes remain ignored, matching the schema's anyAttribute namespace=\"##other\" rule on every element type.

exist-core/src/test/xquery/xquery3/xml-to-json.xql: adds 14 XQSuite regression tests covering each new validation path plus the whitespace-allowed and foreign-namespace-attribute-ignored cases.

Spec references

F&O 3.1 §17.4.2:

The XDM representation of a JSON value … must have the type annotations obtained by validating the untyped representation against the schema given in C.2 Schema for the result of fn:json-to-xml. If it is untyped, then it must be an XDM instance such that validation against this schema would succeed; with the proviso that all attributes other than those in no namespace or in namespace http://www.w3.org/2005/xpath-functions are ignored.

F&O 3.1 §17.5.4 Error Conditions:

A dynamic error is raised [err:FOJS0006] if the value of $input is not a document or element node or is not valid according to the schema for the XML representation of JSON, or if a map element has two children whose normalized key values are the same.

F&O 3.1 Appendix C.2 Schema — stringType declares the escaped attribute; nullType, booleanType, numberType, arrayType, mapType each declare only <xs:anyAttribute processContents=\"skip\" namespace=\"##other\"/>.

XQTS HEAD delta

fn-xml-to-json test set:

	Newly passing	Newly failing
Tests	xml-to-json-{033, 040, 042, 043, 044, 062, 063, 069, 081, 082} (10)	— (0)

Violation kind	Tests
Text child of <map> / <array>	081, 082
Disallowed no-namespace attribute (yek)	033
Attribute in xpath-functions namespace	069
Invalid xs:boolean value for escaped-key	062
Invalid xs:boolean value for escaped	063
Element child of <string>	040
Element child of <boolean>	042
Element child of <null>	043, 044

Baseline: develop @ a3865db (2026-05-11 XQ 3.1 HEAD canonical baseline).

Test plan

• mvn test -pl exist-core -Dtest=xquery.xquery3.XQuery3Tests — 1025 pass (includes 14 new regression cases)
• XQTS HEAD fn-xml-to-json re-run shows +10 newly passing, 0 regressions vs develop baseline
• PMD/Codacy: NPathComplexity stays under the 200 threshold after extracting validateStartElement and validateTextInContext helpers; no new findings (only pre-existing UnusedLocalVariable at line 73 and SimplifyBooleanExpressions at line 207).

🤖 Generated with Claude Code

[line-o]

Double encoded ampersand

[line-o]

s.a.

[line-o]

s.a.

[joewiz]

[This response was co-authored with Claude Code. -Joe]

Fixed in 3576ffc874 — F&O → F&O at lines 267, 288, 329. Matches the bare-ampersand Javadoc convention used elsewhere in org.exist.xquery.functions.fn.* (e.g., FunSum, FunMin, FunParseIetfDate).

[joewiz]

[This response was co-authored with Claude Code. -Joe]

Done in 0f2b5b5751 — converted the START_ELEMENT switch at line 156 to arrow syntax. Kept the END_ELEMENT switch at line 179 unchanged for now (pre-existing, not touched by this PR's diff) — happy to convert it too if you'd like a consistency follow-up.