Thanks Michael, Adding a default namespace to the input schema did the trick and my code is working now - so, this isn't a bug. Cheers,
David _____ From: Michael Glavassevich [mailto:[EMAIL PROTECTED] Sent: Friday, 14 November 2008 15:47 PM To: j-users@xerces.apache.org Subject: Re: error when constructing XSModelGroupImpl and XSObjectListImpl - invalid particle count Hi David, The schema you've attached is invalid. The value of the "ref" attribute is a QName. Since you have no declaration for the default namespace (i.e. xmlns="http://foo.com/#";) in your schema those QNames have no namespace and thus point to nonexistent declarations which couldn't be referred to even if they existed without a <xs:import>. If you give the XSLoader a broken schema don't be surprised if you get a busted XSModel back. The behaviour of such a model is undefined. Thanks. Michael Glavassevich XML Parser Development IBM Toronto Lab E-mail: [EMAIL PROTECTED] E-mail: [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote on 11/13/2008 11:01:14 PM: > In using Xerces2-J 2.9.0 to parse and query XML Schema components in > memory (ref: http://xerces.apache.org/xerces2-j/faq-xs.html#faq-9), > I have noticed what seems to be a bug when parsing the attached > schema, test.xsl. > > The instance of org.apache.xerces.impl.xs.XSModelGroupImpl created > for the fragment: > <xs:sequence> > <xs:element ref="a"/> > <xs:element ref="b" minOccurs="0" maxOccurs="1"/> > <xs:element ref="c" minOccurs="1" maxOccurs="unbounded"/> > </xs:sequence> > ...of test.xsl contains a ternary array XSParticleDecl[] fParticles > containing instances for each of the three xs:element declarations > in the sequence, however the corresponding fParticleCount attribute > (recording the number of particles in fParticles) seems to have not > been set, defaulting to 0. > > I discovered the problem via a call to getParticles() on org.apache. > xerces.impl.xs.XSModelGroupImpl, which builds a new XSObjectListImpl > via constructor XSObjectListImpl(XSObject[] array, int length), > where array is set to fParticles and length to fParticleCount. Using > fParticles.length for the latter (in this case) solves the problem. > Attached is a test file to reproduce this behaviour. > > If someone can confirm this as a bug, I'll submit it... thanks! > > David
smime.p7s
Description: S/MIME cryptographic signature
