Merge pull request ruby#214 from yui-knk/extract_process_rhs

Extract `rhs` replacement logic into `#process_rhs`
ydah · Nov 10, 2023 · eddf032 · eddf032
2 parents 04010c2 + 49c43c0
commit eddf032
Show file tree

Hide file tree

Showing 2 changed files with 35 additions and 13 deletions.
diff --git a/lib/lrama/grammar/rule_builder.rb b/lib/lrama/grammar/rule_builder.rb
@@ -14,7 +14,6 @@ def initialize(rule_counter, midrule_action_counter)
         @user_code = nil
         @precedence_sym = nil
         @line = nil
-        @code_to_new_token = {}
       end
 
       def add_rhs(rhs)
@@ -55,20 +54,15 @@ def preprocess_references
       end
 
       def midrule_action_rules
-        @midrule_action_rules ||= rhs.select do |token|
-          token.is_a?(Lrama::Lexer::Token::UserCode)
-        end.each_with_index.map do |code, i|
-          prefix = code.referred ? "@" : "$@"
-          new_token = Lrama::Lexer::Token::Ident.new(s_value: prefix + @midrule_action_counter.increment.to_s)
-          @code_to_new_token[code] = new_token
-          Rule.new(id: @rule_counter.increment, lhs: new_token, rhs: [], token_code: code, lineno: code.line)
-        end
+        process_rhs
+
+        @midrule_action_rules
       end
 
       def rhs_with_new_tokens
-        rhs.map do |token|
-          @code_to_new_token[token] || token
-        end
+        process_rhs
+
+        @replaced_rhs
       end
 
       def build_rules
@@ -84,6 +78,34 @@ def build_rules
 
       private
 
+      # rhs is a mixture of variety type of tokens like `Ident`, `Parameterizing`, `UserCode` and so on.
+      # `#process_rhs` replaces some kind of tokens to `Ident` so that all `@replaced_rhs` are `Ident` or `Char`.
+      def process_rhs
+        return @replaced_rhs if @replaced_rhs
+
+        @replaced_rhs = []
+        @midrule_action_rules = []
+
+        rhs.each_with_index do |token|
+          case token
+          when Lrama::Lexer::Token::Char
+            @replaced_rhs << token
+          when Lrama::Lexer::Token::Ident
+            @replaced_rhs << token
+          when Lrama::Lexer::Token::Parameterizing
+            # TODO: Expand Parameterizing here
+            @replaced_rhs << token
+          when Lrama::Lexer::Token::UserCode
+            prefix = token.referred ? "@" : "$@"
+            new_token = Lrama::Lexer::Token::Ident.new(s_value: prefix + @midrule_action_counter.increment.to_s)
+            @replaced_rhs << new_token
+            @midrule_action_rules << Rule.new(id: @rule_counter.increment, lhs: new_token, rhs: [], token_code: token, lineno: token.line)
+          else
+            raise "Unexpected token. #{token}"
+          end
+        end
+      end
+
       def expand_parameterizing_rules
         rhs = rhs_with_new_tokens
         rules = []

diff --git a/sig/lrama/grammar/rule_builder.rbs b/sig/lrama/grammar/rule_builder.rbs
@@ -9,7 +9,6 @@ module Lrama
       attr_accessor extracted_action_number: Integer
 
       @user_code: Lexer::Token::UserCode?
-      @code_to_new_token: Hash[Lexer::Token, Lexer::Token]
 
       def initialize: (Counter rule_counter, Counter midrule_action_counter) -> void
       def add_rhs: (Lexer::Token rhs) -> void
@@ -22,6 +21,7 @@ module Lrama
 
       private
 
+      def process_rhs: () -> void
       def expand_parameterizing_rules: () -> Array[Rule]
       def numberize_references: () -> void
       def setup_references: () -> void