From 38ae5ac7a1591747b81d9a133dc44a5396249c30 Mon Sep 17 00:00:00 2001
From: Josh Holtrop <jholtrop@gmail.com>
Date: Thu, 15 Sep 2022 22:46:44 -0400
Subject: [PATCH] Split Token class into Token/Pattern

---
 lib/propane.rb                 |  1 +
 lib/propane/generator.rb       |  2 +-
 lib/propane/grammar.rb         | 11 ++++---
 lib/propane/lexer.rb           |  6 ++--
 lib/propane/lexer/dfa.rb       |  8 ++---
 lib/propane/pattern.rb         | 53 ++++++++++++++++++++++++++++++++++
 lib/propane/token.rb           | 24 ---------------
 spec/propane/lexer/dfa_spec.rb |  5 ++--
 8 files changed, 72 insertions(+), 38 deletions(-)
 create mode 100644 lib/propane/pattern.rb

diff --git a/lib/propane.rb b/lib/propane.rb
index 37236d3..2de0f54 100644
--- a/lib/propane.rb
+++ b/lib/propane.rb
@@ -13,6 +13,7 @@ require_relative "propane/lexer/dfa"
 require_relative "propane/parser"
 require_relative "propane/parser/item"
 require_relative "propane/parser/item_set"
+require_relative "propane/pattern"
 require_relative "propane/regex"
 require_relative "propane/regex/nfa"
 require_relative "propane/regex/unit"
diff --git a/lib/propane/generator.rb b/lib/propane/generator.rb
index 0b3456b..c086cf5 100644
--- a/lib/propane/generator.rb
+++ b/lib/propane/generator.rb
@@ -61,7 +61,7 @@ class Propane
         end
       end
       determine_possibly_empty_rulesets!(rule_sets)
-      @lexer = Lexer.new(@grammar.tokens, @grammar.drop_tokens)
+      @lexer = Lexer.new(@grammar.patterns)
       @parser = Parser.new(@grammar, rule_sets, rule_sets["Start"], @log)
     end
 
diff --git a/lib/propane/grammar.rb b/lib/propane/grammar.rb
index 9015cc6..60a5bd3 100644
--- a/lib/propane/grammar.rb
+++ b/lib/propane/grammar.rb
@@ -3,14 +3,14 @@ class Propane
   class Grammar
 
     attr_reader :classname
-    attr_reader :drop_tokens
     attr_reader :modulename
+    attr_reader :patterns
     attr_reader :rules
     attr_reader :tokens
 
     def initialize(input)
+      @patterns = []
       @tokens = []
-      @drop_tokens = []
       @rules = []
       input = input.gsub("\r\n", "\n")
       parse_grammar(input)
@@ -37,10 +37,13 @@ class Propane
           unless name =~ /^[a-zA-Z_][a-zA-Z_0-9]*$/
             raise Error.new("Invalid token name #{name.inspect}")
           end
-          @tokens << Token.new(name: name, pattern: pattern, id: @tokens.size, line_number: line_number)
+          token = Token.new(name: name, id: @tokens.size, line_number: line_number)
+          @tokens << token
+          pattern = Pattern.new(pattern: pattern, token: token, line_number: line_number)
+          @patterns << pattern
         elsif sliced = input.slice!(/\Adrop\s+(\S+)\s*;/)
           pattern = $1
-          @drop_tokens << Token.new(pattern: pattern, line_number: line_number)
+          @patterns << Pattern.new(pattern: pattern, line_number: line_number, drop: true)
         elsif sliced = input.slice!(/\A(\S+)\s*->\s*(.*?)(?:;|<<\n(.*?)^>>\n)/m)
           rule_name, components, code = $1, $2, $3
           unless rule_name =~ /^[a-zA-Z_][a-zA-Z_0-9]*$/
diff --git a/lib/propane/lexer.rb b/lib/propane/lexer.rb
index 7dc5921..0968308 100644
--- a/lib/propane/lexer.rb
+++ b/lib/propane/lexer.rb
@@ -5,8 +5,8 @@ class Propane
     #   Lexer DFA.
     attr_accessor :dfa
 
-    def initialize(tokens, drop_tokens)
-      @dfa = DFA.new(tokens, drop_tokens)
+    def initialize(patterns)
+      @dfa = DFA.new(patterns)
     end
 
     def build_tables
@@ -20,7 +20,7 @@ class Propane
           elsif state.accepts.drop?
             TOKEN_DROP
           else
-            state.accepts.id
+            state.accepts.token.id
           end
         state_table << {
           transition_table_index: transition_table.size,
diff --git a/lib/propane/lexer/dfa.rb b/lib/propane/lexer/dfa.rb
index b10dee7..68a8bc2 100644
--- a/lib/propane/lexer/dfa.rb
+++ b/lib/propane/lexer/dfa.rb
@@ -3,11 +3,11 @@ class Propane
 
     class DFA < FA
 
-      def initialize(tokens, drop_tokens)
+      def initialize(patterns)
         super()
         start_nfa = Regex::NFA.new
-        (tokens + drop_tokens).each do |token|
-          start_nfa.start_state.add_transition(nil, token.nfa.start_state)
+        patterns.each do |pattern|
+          start_nfa.start_state.add_transition(nil, pattern.nfa.start_state)
         end
         @nfa_state_sets = {}
         @states = []
@@ -40,7 +40,7 @@ class Propane
           nfa_state_set.each do |nfa_state|
             if nfa_state.accepts
               if state.accepts
-                if nfa_state.accepts.id < state.accepts.id
+                if nfa_state.accepts.line_number < state.accepts.line_number
                   state.accepts = nfa_state.accepts
                 end
               else
diff --git a/lib/propane/pattern.rb b/lib/propane/pattern.rb
new file mode 100644
index 0000000..2e9ebae
--- /dev/null
+++ b/lib/propane/pattern.rb
@@ -0,0 +1,53 @@
+class Propane
+
+  class Pattern
+
+    # @return [String, nil]
+    #   Pattern.
+    attr_reader :pattern
+
+    # @return [Token, nil]
+    #   Token to be returned by this pattern.
+    attr_reader :token
+
+    # @return [Integer, nil]
+    #   Line number where the pattern was defined in the input grammar.
+    attr_reader :line_number
+
+    # @return [Regex::NFA, nil]
+    #   Regex NFA for matching the pattern.
+    attr_reader :nfa
+
+    # Construct a Pattern.
+    #
+    # @param options [Hash]
+    #   Optional parameters.
+    # @option options [Boolean] :drop
+    #   Whether this is a drop pattern.
+    # @option options [String, nil] :pattern
+    #   Pattern.
+    # @option options [Token, nil] :token
+    #   Token to be returned by this pattern.
+    # @option options [Integer, nil] :line_number
+    #   Line number where the token was defined in the input grammar.
+    def initialize(options)
+      @drop = options[:drop]
+      @pattern = options[:pattern]
+      @token = options[:token]
+      @line_number = options[:line_number]
+      regex = Regex.new(@pattern)
+      regex.nfa.end_state.accepts = self
+      @nfa = regex.nfa
+    end
+
+    # Whether the pattern is a drop pattern.
+    #
+    # @return [Boolean]
+    #   Whether the pattern is a drop pattern.
+    def drop?
+      @drop
+    end
+
+  end
+
+end
diff --git a/lib/propane/token.rb b/lib/propane/token.rb
index b40af45..52e6774 100644
--- a/lib/propane/token.rb
+++ b/lib/propane/token.rb
@@ -6,10 +6,6 @@ class Propane
     #   Token name.
     attr_reader :name
 
-    # @return [String, nil]
-    #   Token pattern.
-    attr_reader :pattern
-
     # @return [Integer, nil]
     #   Token ID.
     attr_reader :id
@@ -18,46 +14,26 @@ class Propane
     #   Line number where the token was defined in the input grammar.
     attr_reader :line_number
 
-    # @return [Regex::NFA, nil]
-    #   Regex NFA for matching the token.
-    attr_reader :nfa
-
     # Construct a Token.
     #
     # @param options [Hash]
     #   Optional parameters.
     # @option options [String, nil] :name
     #   Token name.
-    # @option options [String, nil] :pattern
-    #   Token pattern.
     # @option options [Integer, nil] :id
     #   Token ID.
     # @option options [Integer, nil] :line_number
     #   Line number where the token was defined in the input grammar.
     def initialize(options)
       @name = options[:name]
-      @pattern = options[:pattern]
       @id = options[:id]
       @line_number = options[:line_number]
-      unless @pattern.nil?
-        regex = Regex.new(@pattern)
-        regex.nfa.end_state.accepts = self
-        @nfa = regex.nfa
-      end
     end
 
     def c_name
       @name.upcase
     end
 
-    # Whether the token is a drop token.
-    #
-    # @return [Boolean]
-    #   Whether the token is a drop token.
-    def drop?
-      @name.nil?
-    end
-
     def to_s
       @name
     end
diff --git a/spec/propane/lexer/dfa_spec.rb b/spec/propane/lexer/dfa_spec.rb
index bb8b177..5fab585 100644
--- a/spec/propane/lexer/dfa_spec.rb
+++ b/spec/propane/lexer/dfa_spec.rb
@@ -35,7 +35,8 @@ class TestLexer
       end
     end
     if last_accepts
-      [last_accepts.name, last_s]
+      name = last_accepts.token ? last_accepts.token.name : nil
+      [name, last_s]
     end
   end
 
@@ -51,7 +52,7 @@ end
 
 def run(grammar, input)
   grammar = Propane::Grammar.new(grammar)
-  token_dfa = Propane::Lexer::DFA.new(grammar.tokens, grammar.drop_tokens)
+  token_dfa = Propane::Lexer::DFA.new(grammar.patterns)
   test_lexer = TestLexer.new(token_dfa)
   test_lexer.lex(input)
 end