CoreIO : XML export can be broken if the document fields contains prohibited characters

CoreIO XML format directly dumps the String encoding of the Document properties in XML Text node.
But these text node are not wrapped in CDATA, as a result, in some cases (for example if the fields contains HTML or XML), you can end up with a document that is not correctly imported, because the actual field content is parsed at the same time as the envelope.

=> need to use CDATA when creating the export

Since the resuling text node will be read using the same API, there should not be any compatibility issue.