Class PathMap


public class PathMap extends DTM.SinglePathVisitor
Encoding of the target language parse automaton as states and relations. Required to work in deterministic and non-deterministic mode for Fragmented Validation of XSLT code. Collects all possible paths through content models for one given dtd. The algorithm is a variant of the tpath algorithm described in the "absynt" paper. Here all content model positions are encoded as integers. We need three relations:
  1. between parsing positions and the element/pcdata/attributes which must be consumed to reach them,
  2. between parsing positions and the content model they belong to (currently diagnosis only!),
  3. among parsing positions, representing legal transitions.
Addtionally we have the set of accepting states and one start position for each content model.
For uniform treatment, the toplevel document and all pc data contents are represented by one pseudo element each.