Why does loadXML over-write attributes in data DOM even if 'ignore root element' argument is true?

Question

I'm running Acrobat 9.4 if that makes any difference.  This question was spurred by another discussion on inserting an item into a list.  Basically loadXML is erasing the attributes on the reference node.  For example, if domNode is a node in the data DOM and newXML is a replica of domNode's XML with some child elements changed and newXML_str is the string representation of newXML, then the statement domNode.loadXML(newXML_str,true,true) eliminates the attributes in domNode.  A commenter in the above link states that it happens because the attribute is not stored in the data DOM as an attribute but rather as a child node with prefix '@' but if that were the case it would seem that saveXML should show it as a child element and not as an attribute.  However, saveXML (as shown below) correctly shows the attribute as an attribute.
What I would like to happen is to have the child attribute node(s) preserved and the child element nodes replaced.  It seems to me that if the second argument is true (ignore root element in xmlArg) then the root element in the data DOM (including attributes) should be unchanged but that is not what happens.  
Here is some sample XML:
<docroot>
    <contact uid="29033737">
        <firstName>George</firstName>
        <lastName>Sinkenschneider</lastName>
    </contact>
</docroot>
Here's some JavaScript:
var domNode = xfa.resolveNode("xfa.data.docroot.contact.(uid.value=='29033737')");
var domNodeXML_str = domNode.saveXML();
console.println("  contact data node saveXML (domNode) = ".concat(domNodeXML_str));
console.println("  domNode.uid.value=".concat(domNode.uid.value));
var newXML_str = domNodeXML_str.replace("George", "Johnny");
console.println("  newXML_str = ".concat(newXML_str));
domNode.loadXML(newXML_str, true, true);
console.println("  DOM node after executing loadXML(newXML_str, true, true) = ".concat(domNode.saveXML()));
xfa.form.remerge();
Here's the console output from the script (including all the ugly line-feeds):
  contact data node saveXML (domNode) = <?xml version="1.0" encoding="UTF-8"?>
<contact uid="29033737"
><firstName
>George</firstName
><lastName
>Sinkenschneider</lastName
></contact
>
  domNode.uid.value=29033737
  newXML_str = <?xml version="1.0" encoding="UTF-8"?>
<contact uid="29033737"
><firstName
>Johnny</firstName
><lastName
>Sinkenschneider</lastName
></contact
>
  DOM node after executing loadXML(newXML_str, true, true) = <?xml version="1.0" encoding="UTF-8"?>
<contact
><firstName
>Johnny</firstName
><lastName
>Sinkenschneider</lastName
></contact
>
Note that the uid attribute is missing from the final DOM node.

Answer 1

Looks like you may have found a bug .....If I append the new xml I see that the uid attribute is maintained (oit is not touching the root attribute). If I replace it then it seems that the attribute is ignored when the node is re-written. We will need to open a bug with support ...do you want to do that?
Paul