Called when a new document is being parsed and checks
if the document has already been parsed before.
Used to avoid recursive inclusion. Note that the same
document will be parsed multiple times if they are for different
target namespaces.
Document Graph Model
The challenge we are facing here is that you have a graph of
documents that reference each other. Each document has an unique
URI to identify themselves, and references are done by using those.
The graph may contain cycles.
Our goal here is to parse all the documents in the graph, without
parsing the same document twice. This method implements this check.
One complication is the chameleon schema; a document can be parsed
multiple times if they are under different target namespaces.
Also, note that when you resolve relative URIs in the @schemaLocation,
their base URI is *NOT* the URI of the document.