During the execution of the Node2XMLStreamReader transform operation,
a code path compiles properties into a list, namely the different
namespaces and their prefixes. The code allows for the special case
for the default namespace (in
org.apache.tuscany.core.databinding.xml.DOMXMLStreamReader: line
1224):
if
(XMLConstants.XMLNS_ATTRIBUTE_NS_URI.equals(attr.getNamespaceURI()))
{
// Skip xmlns:xxx
if (!attr.getName().equals(attr.getLocalName())) {
// Skip xmlns="..."
registerNamespace(attr.getLocalName(), attr.getValue());
} else {
registerNamespace(XMLConstants.DEFAULT_NS_PREFIX,
attr.getValue());
}
continue;
}
In the integration branch
(http://svn.apache.org/repos/asf/incubator/tuscany/branches/sca-java-integration),
however, the tabulation doesn't account for this case:
if (XMLNS_ATTRIBUTE_NS_URI.equals(ns)) {
nsList.add(new QName(attr.getValue(), "", attr.getLocalName()));
}
When this transformation occurs, it tabulates the default namespace as
a prefix of "xmlns" resulting in a SOAP message with following
namespace definitions:
xmlns="..." xmlns:xmlns="..."
Was there a reason for this change in the integration branch or is
this a problem that needs to be fixed?
Kapish Aggarwal
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
