[scanner] Split SkipWhiteSpace into fast and slow path

This places the hot part of SkipWhiteSpace in the header, allowing it to be
inlined, and leaves a slow path to handle the rest. This improves comment
scanning overall by ~10%.

Bug: v8:7926
Change-Id: I2e2ebbbae0d1af619b161397712fdf667d078884
Reviewed-on: https://chromium-review.googlesource.com/1150230
Commit-Queue: Florian Sattler <sattlerf@google.com>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54744}

src/parsing/scanner-inl.h

+// Copyright 2018 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.
+
+#ifndef V8_PARSING_SCANNER_INL_H_
+#define V8_PARSING_SCANNER_INL_H_
+
+#include "src/parsing/scanner.h"
+#include "src/unicode-cache-inl.h"
+
+namespace v8 {
+namespace internal {
+
+V8_INLINE Token::Value Scanner::SkipWhiteSpace() {
+  int start_position = source_pos();
+
+  SkipWhiteSpaceImpl();
+
+  // If there is an HTML comment end '-->' at the beginning of a
+  // line (with only whitespace in front of it), we treat the rest
+  // of the line as a comment. This is in line with the way
+  // SpiderMonkey handles it.
+  if (c0_ != '-' || !has_line_terminator_before_next_) {
+    // Return whether or not we skipped any characters.
+    if (source_pos() == start_position) {
+      return Token::ILLEGAL;
+    }
+
+    return Token::WHITESPACE;
+  }
+
+  return TryToSkipHTMLCommentAndWhiteSpaces(start_position);
+}
+
+V8_INLINE void Scanner::SkipWhiteSpaceImpl() {
+  while (true) {
+    // We won't skip behind the end of input.
+    DCHECK(!unicode_cache_->IsWhiteSpace(kEndOfInput));
+
+    // Advance as long as character is a WhiteSpace or LineTerminator.
+    // Remember if the latter is the case.
+    if (unibrow::IsLineTerminator(c0_)) {
+      has_line_terminator_before_next_ = true;
+    } else if (!unicode_cache_->IsWhiteSpace(c0_)) {
+      break;
+    }
+    Advance();
+  }
+}
+}  // namespace internal
+}  // namespace v8
+
+#endif  // V8_PARSING_SCANNER_INL_H_

src/parsing/scanner.cc

@@ -15,7 +15,7 @@
 #include "src/conversions-inl.h"
 #include "src/objects/bigint.h"
 #include "src/parsing/duplicate-finder.h"  // For Scanner::FindSymbol
-#include "src/unicode-cache-inl.h"
+#include "src/parsing/scanner-inl.h"
 
 namespace v8 {
 namespace internal {
@@ -420,32 +420,13 @@ Token::Value Scanner::PeekAhead() {
   return ret;
 }
 
-
-Token::Value Scanner::SkipWhiteSpace() {
-  int start_position = source_pos();
-
+Token::Value Scanner::TryToSkipHTMLCommentAndWhiteSpaces(int start_position) {
   while (true) {
-    while (true) {
-      // We won't skip behind the end of input.
-      DCHECK(!unicode_cache_->IsWhiteSpace(kEndOfInput));
-
-      // Advance as long as character is a WhiteSpace or LineTerminator.
-      // Remember if the latter is the case.
-      if (unibrow::IsLineTerminator(c0_)) {
-        has_line_terminator_before_next_ = true;
-      } else if (!unicode_cache_->IsWhiteSpace(c0_)) {
-        break;
-      }
-      Advance();
-    }
+    Advance();
 
     // If there is an HTML comment end '-->' at the beginning of a
-    // line (with only whitespace in front of it), we treat the rest
-    // of the line as a comment. This is in line with the way
-    // SpiderMonkey handles it.
-    if (c0_ != '-' || !has_line_terminator_before_next_) break;
-
-    Advance();
+    // line, we treat the rest of the line as a comment. This is in line with
+    // the way SpiderMonkey handles it.
     if (c0_ != '-') {
       PushBack('-');  // undo Advance()
       break;
@@ -462,6 +443,12 @@ Token::Value Scanner::SkipWhiteSpace() {
     if (token == Token::ILLEGAL) {
       return token;
     }
+
+    // Skip remaining whitespaces after the HTML comment.
+    SkipWhiteSpaceImpl();
+
+    // Only repeat loop if we find another HTML comment
+    if (c0_ != '-' || !has_line_terminator_before_next_) break;
   }
 
   // Return whether or not we skipped any characters.

src/parsing/scanner.h

@@ -728,7 +728,9 @@ class Scanner {
   // Scans a single JavaScript token.
   void Scan();
 
-  Token::Value SkipWhiteSpace();
+  V8_INLINE Token::Value SkipWhiteSpace();
+  V8_INLINE void SkipWhiteSpaceImpl();
+  Token::Value TryToSkipHTMLCommentAndWhiteSpaces(int start_position);
   Token::Value SkipSingleHTMLComment();
   Token::Value SkipSingleLineComment();
   Token::Value SkipSourceURLComment();