Experimental parser: add break instruction

tools/lexer_generator/nfa.py

 # Copyright 2013 the V8 project authors. All rights reserved.
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions are
 # met:
 #
 #     * Redistributions of source code must retain the above copyright
 #       notice, this list of conditions and the following disclaimer.
 #     * Redistributions in binary form must reproduce the above
 #       copyright notice, this list of conditions and the following
 #       disclaimer in the documentation and/or other materials provided
@@ skipping 203 matching lines @@
     (start, ends) = self.__process(graph[1])
     state = self.__peek_state()
     if not state['start_node']:
       state['start_node'] = self.__new_state()
     end = self.__new_state()
     self.__patch_ends(ends, end)
     ends = [end]
     end.add_epsilon_transition(state['start_node'])
     return (start, ends)
 
+  def __break(self, graph):
+    (start, ends) = self.__process(graph[1])
+    self.__peek_state()['unpatched_ends'] += ends
+    new_end = self.__new_state()
+    for end in ends:
+      end.add_epsilon_transition(new_end)
+    return (start, [new_end])
+
   def __join(self, graph):
     (graph, name, subgraph, modifier) = graph[1:]
     subgraphs = self.__peek_state()['subgraphs']
     if not name in subgraphs:
       subgraphs[name] = self.__nfa(subgraph)
     (subgraph_start, subgraph_end, nodes_in_subgraph) = subgraphs[name]
     (start, ends) = self.__process(graph)
     if modifier:
       assert modifier == 'ZERO_OR_MORE'
       for end in ends:
@@ skipping 13 matching lines @@
     return self.__operation_map[method](graph)
 
   def __patch_ends(self, ends, new_end):
     for end in ends:
       end.close(new_end)
 
   def __push_state(self):
     self.__states.append({
       'start_node' : None,
       'subgraphs' : {},
+      'unpatched_ends' : [],
     })
 
   def __pop_state(self):
     return self.__states.pop()
 
   def __peek_state(self):
     return self.__states[len(self.__states) - 1]
 
   def __nfa(self, graph):
     start_node_number = self.__node_number
     self.__push_state()
     (start, ends) = self.__process(graph)
     state = self.__pop_state()
     if state['start_node']:
       state['start_node'].close(start)
       start = state['start_node']
     for k, subgraph in state['subgraphs'].items():
       subgraph[1].close(None)
     end =  self.__new_state()
+    if self.__states:
+      self.__peek_state()['unpatched_ends'] += state['unpatched_ends']
+    else:
+      self.__patch_ends(state['unpatched_ends'], end)
     self.__patch_ends(ends, end)
     return (start, end, self.__node_number - start_node_number)
 
   def nfa(self, graph):
     (start, end, nodes_created) = self.__nfa(graph)
     end.close(None)
     return Nfa(start, end, nodes_created)
 
   @staticmethod
   def add_action(graph, action):
     return ('ACTION', graph, action)
 
   @staticmethod
   def add_continue(graph):
     return ('CONTINUE', graph)
 
   @staticmethod
+  def add_break(graph):
+    return ('BREAK', graph)
+
+  @staticmethod
   def join_subgraph(graph, name, subgraph, modifier):
     if modifier:
       modifier = NfaBuilder.__modifer_map[modifier]
     return ('JOIN', graph, name, subgraph, modifier)
 
   @staticmethod
   def or_graphs(graphs):
     return reduce(lambda acc, g: ('OR', acc, g), graphs)
 
   @staticmethod
@@ skipping 94 matching lines @@
   def compute_dfa(self):
     self.__compute_epsilon_closures()
     dfa_nodes = {}
     start_name = self.__to_dfa(set([self.__start]), dfa_nodes, self.__end)
     return (start_name, dfa_nodes)
 
   def to_dot(self):
     iterator = lambda visitor, state: self.__visit_all_edges(visitor, state)
     state_iterator = lambda x : x
     return self.generate_dot(self.__start, set([self.__end]), iterator, state_iterator)