SPARQL Parser handles escaping incorrectly

[joka921]

The SPARQL-parser currently has one big regex for <IRIREF>, prefixed:name, and one for all Literals, including langtags and datatypes.

prefixes are then much later expanded by the ParsedQuery class.

In addition, there is a separate PropertyPathParser, which again splits a complete PropertyPath again into parts.

This should all be done properly by the SPARQL-Parser.