[scanner] Push surrogate pair handling down into identifier scanning

Most scanner logic doesn't need to care about surrogate pairs, so we can just
push it down to identifier scanning.

This CL additionally drops some explicit kEndOfInput checks that are subsumed
by predicates seemlessly returning false for kEndOfInput (-1).

Change-Id: If031a9355ab5fbca0c3b647045e3034f42923979
Reviewed-on: https://chromium-review.googlesource.com/1124447Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54172}

src/parsing/scanner.h

@@ -608,26 +608,26 @@ class Scanner {
   }
 
   // Low-level scanning support.
-  template <bool capture_raw = false, bool check_surrogate = true>
+  template <bool capture_raw = false>
   void Advance() {
     if (capture_raw) {
       AddRawLiteralChar(c0_);
     }
     c0_ = source_->Advance();
-    if (check_surrogate) HandleLeadSurrogate();
   }
 
-  void HandleLeadSurrogate() {
+  bool CombineSurrogatePair() {
     DCHECK(!unibrow::Utf16::IsLeadSurrogate(kEndOfInput));
     if (unibrow::Utf16::IsLeadSurrogate(c0_)) {
       uc32 c1 = source_->Advance();
       DCHECK(!unibrow::Utf16::IsTrailSurrogate(kEndOfInput));
-      if (!unibrow::Utf16::IsTrailSurrogate(c1)) {
-        source_->Back();
-      } else {
+      if (unibrow::Utf16::IsTrailSurrogate(c1)) {
         c0_ = unibrow::Utf16::CombineSurrogatePair(c0_, c1);
+        return true;
       }
+      source_->Back();
     }
+    return false;
   }
 
   void PushBack(uc32 ch) {
@@ -752,7 +752,6 @@ class Scanner {
   Token::Value ScanNumber(bool seen_period);
   Token::Value ScanIdentifierOrKeyword();
   Token::Value ScanIdentifierOrKeywordInner(LiteralScope* literal);
-  Token::Value ScanIdentifierSuffix(LiteralScope* literal, bool escaped);
 
   Token::Value ScanString();
   Token::Value ScanPrivateName();

src/unicode.h

@@ -39,8 +39,12 @@ class Predicate {
     inline CacheEntry()
         : bit_field_(CodePointField::encode(0) | ValueField::encode(0)) {}
     inline CacheEntry(uchar code_point, bool value)
-        : bit_field_(CodePointField::encode(code_point) |
-                     ValueField::encode(value)) {}
+        : bit_field_(
+              CodePointField::encode(CodePointField::kMask & code_point) |
+              ValueField::encode(value)) {
+      DCHECK_IMPLIES((CodePointField::kMask & code_point) != code_point,
+                     code_point == static_cast<uchar>(-1));
+    }
 
     uchar code_point() const { return CodePointField::decode(bit_field_); }
     bool value() const { return ValueField::decode(bit_field_); }