Speed up identifier, keyword and smi parsing

src/scanner.cc

 // Copyright 2011 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 // Features shared by parsing and pre-parsing scanners.
 
 #include <stdint.h>
 
 #include <cmath>
 
@@ skipping 965 matching lines @@
         }
       }
     }
 
     // Parse decimal digits and allow trailing fractional part.
     if (kind == DECIMAL) {
       if (at_start) {
         int value = 0;
         while (IsDecimalDigit(c0_)) {
           value = 10 * value + (c0_ - '0');
-          AddLiteralCharAdvance();
+
+          uc32 first_char = c0_;
+          Advance<false, false>();
+          AddLiteralChar(first_char);
         }
 
         if (next_.literal_chars->one_byte_literal().length() < 10 &&
             c0_ != '.' && c0_ != 'e' && c0_ != 'E') {
           smi_value_ = value;
           literal.Complete();
+          HandleLeadSurrugate();
+
           return Token::SMI;
         }
+        HandleLeadSurrugate();
       }
 
       ScanDecimalDigits();  // optional
       if (c0_ == '.') {
         AddLiteralCharAdvance();
         ScanDecimalDigits();  // optional
       }
     }
   }
 
@@ skipping 181 matching lines @@
   return Token::FUTURE_STRICT_RESERVED_WORD ==
          KeywordOrIdentifierToken(string->raw_data(), string->length(),
                                   harmony_scoping_, harmony_modules_,
                                   harmony_classes_);
 }
 
 
 Token::Value Scanner::ScanIdentifierOrKeyword() {
   DCHECK(unicode_cache_->IsIdentifierStart(c0_));
   LiteralScope literal(this);
-  // Scan identifier start character.
-  if (c0_ == '\\') {
+  if (IsInRange(c0_, 'a', 'z')) {
+    do {
+      uc32 first_char = c0_;
+      Advance<false, false>();
+      AddLiteralChar(first_char);
+    } while (IsInRange(c0_, 'a', 'z'));
+
+    if (IsDecimalDigit(c0_) || IsInRange(c0_, 'A', 'Z') || c0_ == '_' ||
+        c0_ == '$') {
+      // Identifier starting with lowercase.
+      uc32 first_char = c0_;
+      Advance<false, false>();
+      AddLiteralChar(first_char);
+      while (IsAsciiIdentifier(c0_)) {
+        uc32 first_char = c0_;
+        Advance<false, false>();
+        AddLiteralChar(first_char);
+      }
+      if (c0_ <= 127 && c0_ != '\\') {
+        literal.Complete();
+        return Token::IDENTIFIER;
+      }
+    } else if (c0_ <= 127 && c0_ != '\\') {
+      // Only a-z+: could be a keyword or identifier.
+      literal.Complete();
+      Vector<const uint8_t> chars = next_.literal_chars->one_byte_literal();
+      return KeywordOrIdentifierToken(chars.start(), chars.length(),
+                                      harmony_scoping_, harmony_modules_,
+                                      harmony_classes_);
+    }
+
+    HandleLeadSurrugate();
+  } else if (IsInRange(c0_, 'A', 'Z') || c0_ == '_' || c0_ == '$') {
+    do {
+      uc32 first_char = c0_;
+      Advance<false, false>();
+      AddLiteralChar(first_char);
+    } while (IsAsciiIdentifier(c0_));
+
+    if (c0_ <= 127 && c0_ != '\\') {
+      literal.Complete();
+      return Token::IDENTIFIER;
+    }
+
+    HandleLeadSurrugate();
+  } else if (c0_ == '\\') {
+    // Scan identifier start character.
     uc32 c = ScanIdentifierUnicodeEscape();
     // Only allow legal identifier start characters.
     if (c < 0 ||
         c == '\\' ||  // No recursive escapes.
         !unicode_cache_->IsIdentifierStart(c)) {
       return Token::ILLEGAL;
     }
     AddLiteralChar(c);
     return ScanIdentifierSuffix(&literal);
+  } else {
+    uc32 first_char = c0_;
+    Advance();
+    AddLiteralChar(first_char);
   }
 
-  uc32 first_char = c0_;
-  Advance();
-  AddLiteralChar(first_char);
-
   // Scan the rest of the identifier characters.
   while (c0_ >= 0 && unicode_cache_->IsIdentifierPart(c0_)) {
     if (c0_ != '\\') {
       uc32 next_char = c0_;
       Advance();
       AddLiteralChar(next_char);
       continue;
     }
     // Fallthrough if no longer able to complete keyword.
     return ScanIdentifierSuffix(&literal);
   }
 
   literal.Complete();
 
   if (next_.literal_chars->is_one_byte()) {
     Vector<const uint8_t> chars = next_.literal_chars->one_byte_literal();
     return KeywordOrIdentifierToken(chars.start(),
                                     chars.length(),
                                     harmony_scoping_,
                                     harmony_modules_,
                                     harmony_classes_);
   }
-
   return Token::IDENTIFIER;
 }
 
 
 Token::Value Scanner::ScanIdentifierSuffix(LiteralScope* literal) {
   // Scan the rest of the identifier characters.
   while (c0_ >= 0 && unicode_cache_->IsIdentifierPart(c0_)) {
     if (c0_ == '\\') {
       uc32 c = ScanIdentifierUnicodeEscape();
       // Only allow legal identifier part characters.
@@ skipping 258 matching lines @@
     }
     backing_store_.Add(static_cast<uint8_t>((one_byte_length >> 7) | 0x80u));
   }
   backing_store_.Add(static_cast<uint8_t>(one_byte_length & 0x7f));
 
   backing_store_.AddBlock(bytes);
   return backing_store_.EndSequence().start();
 }
 
 } }  // namespace v8::internal