Errors validating redefined attributeGroup
Added by Anonymous almost 13 years ago
Legacy ID: #11004304 Legacy Poster: Emma Burrows (emmaburrows)
Hello. When validating the following (test) schemas with saxon EE 9.3.0.5 in Oxygen 12: [b]testBase.xsd:[/b] <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified"> <xs:simpleType name="yesno-att.class"> <xs:restriction base="xs:string"> <xs:enumeration value="yes"/> <xs:enumeration value="no"/> </xs:restriction> </xs:simpleType> <xs:attributeGroup name="image.attributes"> <xs:attribute name="href" type="xs:string"/> <xs:attribute name="scalefit" type="yesno-att.class"/> </xs:attributeGroup> <xs:element name="test"> xs:complexType <xs:attributeGroup ref="image.attributes"/> </xs:complexType> </xs:element> </xs:schema> [b]test.xsd[/b] <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified"> <xs:redefine schemaLocation="testBase.xsd"> <xs:attributeGroup name="image.attributes"> <xs:attribute name="href" type="xs:string" [u]use="required"[/u]/> <xs:attribute name="scalefit" type="yesno-att.class"/> </xs:attributeGroup> </xs:redefine> </xs:schema> I get the following error message: Invalid restriction: attribute @scalefit is inconsistent with the base type. Type yesno-att.class is not validly derived from type yesno-att.class Xerces reports no error given the same input and the error message appears spurious in any case. Is this a known issue with Saxon 9.3?
Replies (2)
RE: Errors validating redefined attributeGroup - Added by Anonymous almost 13 years ago
Legacy ID: #11004527 Legacy Poster: Michael Kay (mhkay)
Thanks. You are right, this is a bug. The bug had already been fixed in the 9.4 branch, but I will now issue a patch for 9.3. Sorry about the inconvenience. It's one of those bugs where it's quite hard to explain the circumstances in user-oriented terms. During schema processing Saxon creates a temporary SimpleTypeDefinition object for a simple type, and later turns this into an AtomicType, ListType, or UnionType when its variety becomes clear. The problem occurs because sometimes after the conversion, another schema component is left pointing to the old (pre-conversion) SimpleTypeDefinition. It all depends on the order in which components are processed, which is itself rather random, as it involves iterating over a HashMap.
RE: Errors validating redefined attributeGroup - Added by Anonymous almost 13 years ago
Legacy ID: #11004626 Legacy Poster: Emma Burrows (emmaburrows)
Thanks for the swift response and the explanation. George Bina from SyncRO Soft actually spotted that this was a Saxon bug when I reported the behaviour on the DITA list. I suppose it's one of those things that doesn't crop up all that often!
Please register to reply