Hi Elena,

--On Montag, 28. Januar 2002 13:05 -0500 Elena Litani <[EMAIL PROTECTED]> 
wrote:

> Christian,
>
> I could reproduce the problem last week, but we can't reproduce it with
> X2 build today.
> Could you try the latest X2 build and see it is works for you?
> http://gump.covalent.net/jars/latest/
>
> Please, let us know, since we are planning to release the first
> production release of X2 ASAP.

Thanks for fixing that. This 'doubled' path from the Entity Resolver is OK 
now.

BUT: ;-)) another problem remains which breaks the xml-security project if 
xerces-2 is used: I parse the following 'file':

// The text will be a space (20) and the CR/LR combi
byte[] d = { (byte) 0x0d, (byte) 0x0a };
String returnStr = new String(d, "UTF-8");
String str = "<Elem xmlns='http://xml.apache.org/#Sample'>"
+ " " + returnStr + "</Elem>";

If you now say:

Text t = (Text) doc.getDocumentElement().getFirstChild();
byte[] data = t.getData().getBytes("UTF-8");
for (int i=0; i<data.length; i++) {
  System.out.println(i + " " + data[i]);
}

the results differ under Xerces-1.4.4 and the Xerces-2 from the GUMP run:

Xerces 1.4.4
0 32
1 10

Xerces 2
0 32
1 13
2 10

This means v1.4.4 omits the 0x0d but v2 doesn't. I attach a JAVA src for 
demonstrating that bug.

Best regards,
Christian

BTW: Sam, if this is fixed, I assume that xml-security's GUMP will unittest 
under Xerces2, too.

Attachment: TestXerces2.java
Description: Binary data

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to