Can I force a parser to use fixed DTD for validation?

Stephen,

I saw your other posting, and here's some code that at least works. It shows you the referenced external entities. I ran this with Xerces 1.4 versions, 2.0.0_alpha has problems with it.

Here's the test program:

import javax.xml.parsers.DocumentBuilder;

import javax.xml.parsers.DocumentBuilderFactory;

import org.w3c.dom.Document;

import org.xml.sax.EntityResolver;

import org.xml.sax.InputSource;

import org.xml.sax.SAXException;

public class JAXPDOMTest1 implements org.xml.sax.EntityResolver{

public void domParse(String url)

{

DocumentBuilder parser;

DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();

factory.setValidating(true);

try {

parser = factory.newDocumentBuilder();

parser.setEntityResolver(this);

Document doc = parser.parse(url);

} catch (Exception e) {

e.printStackTrace();

}

}

public InputSource resolveEntity(java.lang.String publicId,java.lang.String systemId) throws SAXException,java.io.IOException

{

InputSource is=null;

System.out.println("--");

System.out.println("publicId="+publicId);

System.out.println("systemId="+systemId);

System.out.println("--");

// comment the next 2 lines, if you want to see the DTD specified in the XML file.

if(systemId != null && systemId.endsWithIgnoreCase("dtd"))

is = new InputSource("replaced.dtd");

return is;

}

static public void main(String[] args)

{

JAXPDOMTest1 x=new JAXPDOMTest1();

x.domParse("test.xml");

}

}

here's test.xml:

<?xml version="1.0"?>

<!DOCTYPE GREETING SYSTEM "test.dtd">

<GREETING>

&junk_txt;

</GREETING>

and here's test.dtd (you need a junk.txt file):

<!ELEMENT GREETING (#PCDATA)>

<!ENTITY junk_txt SYSTEM "junk.txt">

and here's replace.dtd (you need a xyz.txt file):

<!ELEMENT GREETING (#PCDATA)>

<!ENTITY junk_txt SYSTEM "xyz.txt">

However, the problem is that in resovleEntity() you only get the systemID or publicID, respectively. That's why I included an external entity in the DTD examples. So, you simply cannot assume that the first external entity is the DTD (suppose you receive a file which contains an internal DTD and has an external reference (like junk_txt above). You will replace that entity instead of the DTD). The example above relies on the fact that files end with "dtd", that might not be the case either. And the DTD can be part of the file you receive, which makes the resolveEntity() solution useless.

Protecting against "malicious" DTDs is tough.

So, you can go with the "hack" solution, go with the EntityResolver solution (if it's good enough for you), or just do it right and move over to XML Schema, which solves this - there you can override the Schema location specified in the XML file by your own.

For further comments you can take a look at the XML Bible - look for "Validating the document against the schema" some 200 lines down on this page:

http://www.ibiblio.org/xml/books/bible2/chapters/ch24.html

Hope that helps a little,

Good luck.

lk555 at 2007-6-29 9:14:29 >

HEEEEEEEEEEELP !

I still have the same problem

(-concerning the first post on this topic-).

I still don't find a way to do this...

This is a very simple problematic though !

I am sure this IS possible indeed...

Any Help would be extremely appreciated....

javalova at 2007-6-29 9:14:29 >

All right... so what are these 5 lines ? :oP

I followed your advice and written a correct schema file (.xsd)

describing the syntax of my xml file.

is there some simple sample code for validating the xml file

with my xsd in java? [i'm still roaming the internet looking for docs...]

javalova at 2007-6-29 9:14:29 >

> parser.setProperty(JAXP_SCHEMA_SOURCE, schema);

So countings not my thing.

I the above line, schema is a File object. I haven't tried many other types of Objects but I know InputStreams don't work. I couldn't find any documentation on what types this property can accept.

Good luck.

dubwai at 2007-6-29 9:14:29 >

I believe what you want is an EntityResolver. It is called with the public and or private IDs of an entity being referenced. It returns an InputSource that will be used by the XMLReader to actually read what it wants. If you pass back an InputSource that is created from a StringReader with a single blank character, you can ignore the DTD all together.

If you know where your copy of "the right" DTD is, you can return an InputSource that referes to it and it should work as you wish.

Dave Patterson

d.patterson at 2007-6-29 9:14:29 >

I am having exactly the same problem and would appreciate any help. To recap here is what I want to do:

I want to validate XML files against a DTD I have created. The XML file will not have any reference to the DTD, or I want to ignore any DTD reference on the XML file

I found a way to do this with XML Schema, but I want to use DTD to do validation. If someone has found solution, can you please post a code snippet - I would prefer to use SAX, but DOM might be ok too

codecraker at 2007-7-1 1:44:19 >