Christian, Yes, the attachment made it to the list though haven't had an opportunity to look at it yet.
Thanks. Michael Glavassevich XML Technologies and WAS Development IBM Toronto Lab E-mail: mrgla...@ca.ibm.com E-mail: mrgla...@apache.org Christian Greene <christian.gre...@oracle.com> wrote on 01/11/2012 03:39:23 PM: > Michael, did you get this previous email or did it bounce because of > the zip attachement? > > Thanks > -Christian > > From: Christian Greene > Sent: Monday, January 09, 2012 3:48 PM > To: mrgla...@ca.ibm.com > Cc: j-users@xerces.apache.org > Subject: RE: Validation of Types has changed from 2.9.1 > > Hi Michael, > > As requested, I have produced a unit test that produces the behavior > (attached zip). Place Java class and XSD into a path like src/test/ > xml/ and run the Java class as a junit test and it will fail like ? > cvc-complex-type.2.4.a: Invalid content was found starting with > element 'b:InnerTypeExtra1'?? > > With 2.11.0 it fails. With 2.9.1, it passes. > > Cheers, > -Christian > > Hi Christian, > > Do you have a complete test case (in addition to the XSDs you provided,= > > Java code + XML document) that you can share? > > Thanks. > > Michael Glavassevich > XML Technologies and WAS Development > IBM Toronto Lab > E-mail: mrgla...@ca.ibm.com > E-mail: mrgla...@apache.org > > > From: Christian Greene > Sent: Wednesday, January 04, 2012 2:36 PM > To: j-users@xerces.apache.org > Subject: Validation of Types has changed from 2.9.1 > > Hi, > > This is not an issue when I use Xerces-J 2.9.1, but because of a > known security issue our group is supposed to be using 2.11.0. When > I execute validation unit tests with the 2.11.0 Xerces library, > tests that passed validation before do not now. In advance, I will > admit the schemas being used are complex, but it was working and I > am wondering if anyone knows what changed and if the behavior is > configurable or pluggable in any way. > > The following snippets layout the issue: > > SchemaA > xmlns:b="http://www.acme.com/schemas"; > <xs:import namespace="http://www.acme.com/schemas"; schemaLocation= > "SchemaB.xsd"/> > > <xs:complexType name="MainType"> > <xs:complexContent> > <xs:extension base="MainType"> > <xs:group ref="b:MainTypeExtraElements"/> > </xs:extension> > </xs:complexContent> > </xs:complexType> > > <xs:complexType name="InnerType"> > <xs:complexContent> > <xs:extension base="InnerType"> > <xs:group ref="b:InnerTypeExtraElements"/> > </xs:extension> > </xs:complexContent> > </xs:complexType> > > SchemaB > <xs:import namespace="http://www.3rdparty.com/namespace"; schemaLocation= > "SchemaC.xsd"/> > <xs:import namespace="http://www.3rdparty.com/namespace"; schemaLocation= > "SchemaD.xsd"/> > > <xs:group name="MainTypeExtraElements"> > <xs:sequence> > <xs:element name="MainTypeExtra1" type="xs:string" minOccurs="0"/> > <xs:element name="MainTypeExtra2" type="xs:string" minOccurs="0"/> > </xs:sequence> > </xs:group> > > <xs:group name="InnerTypeExtraElements"> > <xs:sequence> > <xs:element name="InnerTypeExtra1" type="xs:string" minOccurs="0"/> > </xs:sequence> > </xs:group> > > SchemaC > <xs:complexType name="MainType"> > <xs:sequence> > <xs:element name="MainElement1" type="xs:string"/> > <xs:element name="MainElement2" type="xs:boolean" minOccurs="0"/> > <xs:element name="InnerTypeElement" type="InnerType" minOccurs="0"/> > </xs:sequence> > </xs:complexType> > > SchemaD > <xs:complexType name="InnerType"> > <xs:sequence> > <xs:element name="InnerElement1" type="xs:string"/> > <xs:element name="InnerElement2" type="xs:string"/> > </xs:sequence> > </xs:complexType> > > In 2.9.1, when XML was validated that had extra elements in the > inner type ?InnerType? it would validate fine. With 2.11.0, the > validation fails once InnerTypeExtra1 is traversed complaining that only > InnerElement1 or InnerElement2 is expected. So before when SchemaC > would reference InnerType, the definition from SchemaA was used, but > now it seems that the definition from SchemaD is used. MainTypeExtra1 > would likely fail as well if the validation ever got that far. > > Does that follow? Anyone have a solution? > > Thanks, > -Christian
