Rodolfo Ochoa created XERCESC-2269:
--------------------------------------
Summary: Schema full checking rejects restriction of a union-typed
element to a restriction of that union ("does not derive from type of
corresponding element in the base")
Key: XERCESC-2269
URL: https://issues.apache.org/jira/browse/XERCESC-2269
Project: Xerces-C++
Issue Type: Bug
Components: Validating Parser (XML Schema)
Affects Versions: 3.3.0
Reporter: Rodolfo Ochoa
With schema full checking enabled ({{-f}} in the samples,
{{XMLUni::fgXercesSchemaFullChecking}}), Xerces-C rejects a valid XML Schema
1.0 document in which a complex type derived by restriction narrows an element
declared with a union simple type to a simple type that is itself a restriction
of that union:
{noformat}
Message: element 'e' has type that does not derive from type of corresponding
element in the base
{noformat}
The schema is valid per XML Schema 1.0 Part 1, 3.14.6 "Type Derivation OK
(Simple)", clause 2.1 (the derived type's base type definition is the union),
which is what Particle Valid (Restriction) / NameAndTypeOK clause 7 requires.
The W3C XQuery/XPath test suite (QT3,
{{prod/SchemaImport/unionListDefined.xsd}}) uses this pattern; XQuery
processors built on Xerces-C (e.g. Zorba) fail 161 QT3 test-cases because of it.
Cause: {{UnionDatatypeValidator::isSubstitutableBy(toCheck)}}
({{src/xercesc/validators/datatype/UnionDatatypeValidator.hpp}}) returns true
only if {{toCheck == this}} or {{toCheck}} is substitutable for one of the
member types. A restriction of a union is a {{UnionDatatypeValidator}} whose
base validator is the union, and its member types are the union's member types,
so neither test matches: unlike {{DatatypeValidator::isSubstitutableBy}}, the
union override never walks {{toCheck}}'s base chain. The same function is used
for {{xsi:type}} and substitution-group checks, so those paths are affected too.
*Minimal reproduction* ({{union-restriction.xsd}}):
{code:xml}
<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema";>
<xs:simpleType name="approximateDate">
<xs:union memberTypes="xs:date xs:dateTime xs:gYear xs:gYearMonth"/>
</xs:simpleType>
<xs:simpleType name="restrictedUnion">
<xs:restriction base="approximateDate">
<xs:pattern value="20.*"/>
</xs:restriction>
</xs:simpleType>
<xs:complexType name="approximateDateType">
<xs:sequence>
<xs:element name="e" type="approximateDate"/>
</xs:sequence>
</xs:complexType>
<xs:complexType name="restrictedDateType">
<xs:complexContent>
<xs:restriction base="approximateDateType">
<xs:sequence>
<xs:element name="e" type="restrictedUnion"/>
</xs:sequence>
</xs:restriction>
</xs:complexContent>
</xs:complexType>
<xs:element name="restricted" type="restrictedDateType"/>
</xs:schema>
{code}
Instance ({{union-restriction.xml}}):
{code:xml}
<?xml version="1.0" encoding="UTF-8"?>
<restricted xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";
xsi:noNamespaceSchemaLocation="union-restriction.xsd">
<e>2024-02-29</e>
</restricted>
{code}
Steps:
{noformat}
$ DOMCount -n -s -f union-restriction.xml
Error at file .../union-restriction.xsd, line <L>, char <C> (the restricting
complexType)
Message: element 'e' has type that does not derive from type of corresponding
element in the base
{noformat}
Expected: {{union-restriction.xml: N ms (2 elems).}} Without {{-f}} the
document parses, since the check is part of schema full checking.
A fix with a regression test is ready (PR to follow).
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
