From 9896068726bc9d0ab866e545dc9e36433505b48a Mon Sep 17 00:00:00 2001
From: ydah <t.yudai92@gmail.com>
Date: Fri, 4 Sep 2026 21:28:43 +0900
Subject: [PATCH] Ignore references in C literals and comments

---
 lib/lrama/lexer/token/user_code.rb            | 22 ++++++++++++++++++-
 sig/generated/lrama/lexer/token/user_code.rbs |  3 +++
 spec/lrama/lexer/token/user_code_spec.rb      | 15 +++++++++++++
 3 files changed, 39 insertions(+), 1 deletion(-)

diff --git a/lib/lrama/lexer/token/user_code.rb b/lib/lrama/lexer/token/user_code.rb
index 166f04954..ea1a178bc 100644
--- a/lib/lrama/lexer/token/user_code.rb
+++ b/lib/lrama/lexer/token/user_code.rb
@@ -26,7 +26,11 @@ def _references
             when reference = scan_reference(scanner)
               references << reference
             when scanner.scan(/\/\*/)
-              scanner.scan_until(/\*\//)
+              scanner.scan_until(/\*\//) || scanner.terminate
+            when scanner.scan(%r{//})
+              scanner.scan_until(/\n/) || scanner.terminate
+            when quote = scanner.scan(/["']/)
+              skip_quoted_literal(scanner, quote)
             else
               scanner.getch
             end
@@ -35,6 +39,22 @@ def _references
           references
         end
 
+        # @rbs (StringScanner scanner, String quote) -> void
+        def skip_quoted_literal(scanner, quote)
+          escaped = false
+
+          until scanner.eos?
+            char = scanner.getch
+            if escaped
+              escaped = false
+            elsif char == "\\"
+              escaped = true
+            elsif char == quote
+              break
+            end
+          end
+        end
+
         # @rbs (StringScanner scanner) -> Lrama::Grammar::Reference?
         def scan_reference(scanner)
           start = scanner.pos
diff --git a/sig/generated/lrama/lexer/token/user_code.rbs b/sig/generated/lrama/lexer/token/user_code.rbs
index a2ce8af81..b58486001 100644
--- a/sig/generated/lrama/lexer/token/user_code.rbs
+++ b/sig/generated/lrama/lexer/token/user_code.rbs
@@ -14,6 +14,9 @@ module Lrama
         # @rbs () -> Array[Lrama::Grammar::Reference]
         def _references: () -> Array[Lrama::Grammar::Reference]
 
+        # @rbs (StringScanner scanner, String quote) -> void
+        def skip_quoted_literal: (StringScanner scanner, String quote) -> void
+
         # @rbs (StringScanner scanner) -> Lrama::Grammar::Reference?
         def scan_reference: (StringScanner scanner) -> Lrama::Grammar::Reference?
       end
diff --git a/spec/lrama/lexer/token/user_code_spec.rb b/spec/lrama/lexer/token/user_code_spec.rb
index 5607c6813..657905991 100644
--- a/spec/lrama/lexer/token/user_code_spec.rb
+++ b/spec/lrama/lexer/token/user_code_spec.rb
@@ -58,5 +58,20 @@
       expect(references.count).to eq 1
       expect(references[0]).to eq Lrama::Grammar::Reference.new(type: :at, name: "expr.right", ex_tag: nil, first_column: 1, last_column: 14)
     end
+
+    it "ignores references in C literals and comments" do
+      code = <<~'C'
+        $1;
+        printf("total $2 and @2 \"done\"");
+        int character = '$3';
+        // note $3 and @3
+        $4; @4;
+        /* note $5 and @5 */
+      C
+
+      references = Lrama::Lexer::Token::UserCode.new(s_value: code, location: location).references
+
+      expect(references.map { |ref| [ref.type, ref.index] }).to eq([[:dollar, 1], [:dollar, 4], [:at, 4]])
+    end
   end
 end