Allow escape sequences in Constructor/Prototype tokens in PreParserTraits::GetSymbol()

src/scanner.h

Index: src/scanner.h
diff --git a/src/scanner.h b/src/scanner.h
index 356c8e4a549b0e41b753ab95db4891f164e0bc3d..e1e8f0f560037f67a2b5ace3b157aa14e0928ff5 100644
--- a/src/scanner.h
+++ b/src/scanner.h
@@ -394,6 +394,18 @@ class Scanner {
     }
     return false;
   }
+
+  bool LiteralMatches(const char* data, int length) {
+    if (current_token() == Token::STRING &&

[caitp (gmail), 2014/10/01 20:52:18]

The check for Token::STRING is simply because I don't fully understand the
scanner yet, it was a really quick look to see how this could work. It may not
be necessary.

+        is_literal_one_byte() &&
+        literal_length() == length) {
+      const char* token =
+          reinterpret_cast<const char*>(literal_one_byte_string().start());
+      return !strncmp(token, data, length);
+    }
+    return UnescapedLiteralMatches(data, length);

[caitp (gmail), 2014/10/01 20:52:18]

But if it does turn out to be necessary, I want to use the
unescapedliteralmatches to verify non-string-literals.

[arv (Not doing code reviews), 2014/10/01 21:00:35]

Is the only difference here is the literal_contains_escapes? If so, maybe
refactor to share the code?

+  }
+
   void IsGetOrSet(bool* is_get, bool* is_set) {
     if (is_literal_one_byte() &&
         literal_length() == 3 &&