current community

your communities

more stack exchange communities

Stack Exchange
tour help
stack overflow careers
Take the 2-minute tour ×
Stack Overflow is a question and answer site for professional and enthusiast programmers. It's 100% free.
up vote 1 down vote favorite

I wish to validate an XML file (which has no namespace, nor schema declaration in the XML file) against an XSD file.

The XML file looks like:

<?xml version="1.0" encoding="UTF-8"?>
<report> ... </report>

And the XSD file, is:

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:element name="report">
    ...
  </xs:element>
</xs:schema>

The code is similar to the following:

SchemaFactory xsdFac = SchemaFactory.newInstance(
   XMLConstants.W3C_XML_SCHEMA_NS_URI);
xsdFac.setErrorHandler( ... );
Schema xsd = xsdFac.newSchema(xsdUrl);  // xsdUrl works
Validator xsdValidator = xsd.newValidator();
xsdValidator.validate(new DOMSource(doc));  // doc is correct

I have spent 3 hours on this, as it works locally on my PC, it used to work on the server (I think), but now it does not work on the server. I tried to identify differences between my PC and the server, but they both use the same JARs, and so on.

Anyway, I have identified the following difference. I was not passing a class name to the SchemaFactory.newInstance method (almost certainly a mistake), when i printed out the class name of the xsdFac, I saw it was different between locally and on the server. I don't think I want to get in to why that was the case (I have no idea), I think it's better to find one which works, and explicitly specify it.

  • On my PC (works) it was com.sun.org.apache.xerces.internal.jaxp.validation.XMLSchemaFactory. If I explicitly specify that on the newInstance call, then it works on the PC and on the server.
  • On the server (didn't work), it was org.apache.xerces.jaxp.validation.XMLSchemaFactory. If I explicitly set that on the newInstance call, then I get the same error on both the PC and the server.

Notice the com.sun at the start of the one that works.

So at least I have a solution, that's good. But, I think I shouldn't be using com.sun classes explicitly in my code?

Other information:

  • The error I get is: org.xml.sax.SAXParseException: cvc-elt.1: Cannot find the declaration of element 'report'
  • This is the same error I get when, on the working version, I change the element name to something else, something that doesn't exist in the schema. So I think it just means, "I understand your schema, and I understand the XML, and you've used an element that isn't declared anywhere in the schema."
  • xercesImpl-2.9.1.jar is in my project (both on the server and on the PC).
  • This JAR contains an XMLSchemaFactory.class under the package specified on the non-working version (i.e. would lead me to believe it's there, findable, and should work, and after all I don't get any exceptions relating to classes not found)
  • The doc object I am parsing is, in both cases, a org.apache.xerces.dom.DeferredDocumentImpl

So, my question is: I would like to use the explicit Xerces implementation of the schema factory (I think), as I am including the JAR and I have a Document object from Xerces and basically I am using the Xerces validator anyway (in the working case of com.sun).

Has anyone has a similar experience?

share|improve this question

1 Answer 1

up vote 1 down vote

I would hazard a guess that you don't, in fact, have Xerces on the classpath in the failing case.

-Djaxp.debug will tell you for sure what implementation is active when you call the JAXP APIS that you are calling.

If this is so, and it isn't just the result of a mistake in your classpath, you can indeed specifically 'new' the Xerces classes instead of calling the JAXP generic factory methods. I've done it.

edit

comments indicate that this is a webapp. The JAXP classes are one of those things which are painful to put in a webapp, given the class loading rules. You should at least try putting it in the 'system' classpath of tomcat and see what happens.

share|improve this answer
    
OK thanks I'll give that a go and get back to you! But the system (in the failure case) did instantiate the org.apache.xerces.jaxp.validation.XMLSchemaFactory (I got that class name by doing a System.out.println(fac.getClass()), so I think it must be on the classpath? –  Adrian Smith Feb 19 '11 at 23:49
    
Yes, that does sound promising. But you might have somehow contrived to get a strange mixture of Xerces and non-Xerces. Wait, I misread. –  bmargulies Feb 19 '11 at 23:50
    
Yes I do feel that a "strange mixture" is exactly what's going on, although I've no way to prove it. I mean saying the doc does not conform to the schema (but no other errors such as class not found), when the XML and the XSD file are available, and another implementation says everything's OK, .... hmm. –  Adrian Smith Feb 19 '11 at 23:56
    
Or could it be a classloading issue? Do I need to "endorse" the Xerces JAR in Tomcat etc., otherwise the class loader associated with SchemaFactory.newInstance (system class) cannot find my Xerces JAR (in my WAR)? –  Adrian Smith Feb 20 '11 at 0:05
    
Oh, a WAR file. There you go. The simplest thing is to put Xerces in the system class loader for tomcat, not in the webapp. –  bmargulies Feb 20 '11 at 0:20

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.