2004-05-07 Atsushi Enomoto <atsushi@ximian.com>
* XmlDataInferenceReader.cs : Added. It does InferXmlSchema() (it is
also expected to do ReadXml(), but I will remove that feature and
support column reordering).
* XmlSchemaDataImporter.cs : Always initialize dataset name.
Always set Nested and primary key for parent table.
* XmlDataReader.cs : Now it is used.
Top level element might not match to dataset name nor any table
name, but still possible to read content tables.
Handle empty element correctly in some places.
Handle Fragment mode (read up XmlReader to the end).
* DataSet.cs : Modified InferXmlSchema() and ReadXml() to use
XmlDataInferenceLoader and XmlDataReader.
svn path=/trunk/mcs/; revision=26889