From b6e3a5c15179260b0c2eae301dc005d6196f22ea Mon Sep 17 00:00:00 2001
From: Josh Holtrop <jholtrop@gmail.com>
Date: Tue, 25 May 2021 16:00:25 -0400
Subject: [PATCH] Record accepting token in DFA state

---
 lib/imbecile/grammar.rb   |  4 +++-
 lib/imbecile/regex/dfa.rb | 17 +++++++++++++++--
 2 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/lib/imbecile/grammar.rb b/lib/imbecile/grammar.rb
index 0b92174..7c7c9cc 100644
--- a/lib/imbecile/grammar.rb
+++ b/lib/imbecile/grammar.rb
@@ -39,9 +39,11 @@ module Imbecile
       end
 
       # Build NFA from each token expression.
+      i = 0
       @tokens.each do |token_name, token_def|
         token_def[:regex] = Regex.new(token_def[:pattern])
-        token_def[:regex].nfa.end_state.accepts = token_name
+        token_def[:regex].nfa.end_state.accepts = "#{i}:#{token_name}"
+        i += 1
       end
       dfa = Regex::DFA.new(@tokens.map {|token_name, token_def| token_def[:regex].nfa})
       puts dfa
diff --git a/lib/imbecile/regex/dfa.rb b/lib/imbecile/regex/dfa.rb
index 2dc006d..65a93cf 100644
--- a/lib/imbecile/regex/dfa.rb
+++ b/lib/imbecile/regex/dfa.rb
@@ -60,13 +60,15 @@ module Imbecile
         end
         rv = ""
         @states.each_with_index do |state, state_id|
-          rv += "#{state_id}:\n"
+          accepts_s = state.accepts ? " (#{state.accepts})" : ""
+          rv += "#{state_id}#{accepts_s}:\n"
           state.transitions.each do |transition|
             range_s = chr[transition.code_point_range.first]
             if transition.code_point_range.size > 1
               range_s += "-" + chr[transition.code_point_range.last]
             end
-            rv += "  #{range_s} => #{transition.destination.id}\n"
+            accepts_s = transition.destination.accepts ? " (#{transition.destination.accepts})" : ""
+            rv += "  #{range_s} => #{transition.destination.id}#{accepts_s}\n"
           end
         end
         rv
@@ -85,6 +87,17 @@ module Imbecile
 
       def process_nfa_state_set(nfa_state_set)
         state = @states[@nfa_state_sets[nfa_state_set]]
+        nfa_state_set.each do |nfa_state|
+          if nfa_state.accepts
+            if state.accepts
+              if nfa_state.accepts.to_i < state.accepts.to_i
+                state.accepts = nfa_state.accepts
+              end
+            else
+              state.accepts = nfa_state.accepts
+            end
+          end
+        end
         transitions = transitions_for(nfa_state_set)
         while transitions.size > 0
           subrange = CodePointRange.first_subrange(transitions.map(&:code_point_range))