Provide a code point iterator to the String class to simplify iteration.

runtime/vm/unicode.h

 // Copyright (c) 2011, the Dart project authors.  Please see the AUTHORS file
 // for details. All rights reserved. Use of this source code is governed by a
 // BSD-style license that can be found in the LICENSE file.
 
 #ifndef VM_UNICODE_H_
 #define VM_UNICODE_H_
 
 #include "vm/allocation.h"
 #include "vm/globals.h"
 
 namespace dart {
 
 class String;
 
 class Utf8 : AllStatic {
  public:
   enum Type {
     kLatin1 = 0,  // Latin-1 character set.
     kBMP,  // Basic Multilingual Plane.
     kSMP,  // Supplementary Multilingual Plane.
   };
 
   static const intptr_t kMaxOneByteChar   = 0x7F;
   static const intptr_t kMaxTwoByteChar   = 0x7FF;
   static const intptr_t kMaxThreeByteChar = 0xFFFF;
   static const intptr_t kMaxFourByteChar  = 0x10FFFF;
-  static const intptr_t kMaxBmpCodepoint  = 0xffff;
-  static const int32_t kLeadOffset = (0xD800 - (0x10000 >> 10));
-  static const int32_t kSurrogateOffset = (0x10000 - (0xD800 << 10) - 0xDC00);
 
-  static void ConvertUTF32ToUTF16(int32_t codepoint, uint16_t* dst);
   static intptr_t CodePointCount(const uint8_t* utf8_array,
                                  intptr_t array_len,
                                  Type* type);
 
   // Returns true if 'utf8_array' is a valid UTF-8 string.
   static bool IsValid(const uint8_t* utf8_array, intptr_t array_len);
 
   static intptr_t Length(int32_t ch);
   static intptr_t Length(const String& str);
 
@@ skipping 20 matching lines @@
                                    uint32_t* dst,
                                    intptr_t len) {
     ASSERT(str != NULL);
     intptr_t array_len = strlen(str);
     const uint8_t* utf8_array = reinterpret_cast<const uint8_t*>(str);
     return DecodeToUTF32(utf8_array, array_len, dst, len);
   }
 };
 
 
+class Utf16 : AllStatic {
+ public:
+  static const int32_t kMaxBmpCodepoint  = 0xFFFF;
+
+  static const int32_t kLeadSurrogateOffset = (0xD800 - (0x10000 >> 10));
+
+  static const int32_t kSurrogateOffset = (0x10000 - (0xD800 << 10) - 0xDC00);
+
+  // Returns the length of the code point in UTF-16 code units.
+  static intptr_t Length(int32_t ch) {
+    return (ch <= kMaxBmpCodepoint) ? 1 : 2;
+  }
+
+  // Returns true if ch is a lead or trail surrogate.
+  static bool IsSurrogate(int32_t ch) {
+    return (ch & 0xFFFFF800) == 0xD800;
+  }
+
+  // Returns true if ch is a lead surrogate.
+  static bool IsLeadSurrogate(int32_t ch) {
+    return (ch & 0xFFFFFC00) == 0xD800;
+  }
+
+  // Returns true if ch is a low surrogate.
+  static bool IsTrailSurrogate(int32_t ch) {
+    return (ch & 0xFFFFFC00) == 0xDC00;
+  }
+
+  // Decodes a surrogate pair into a supplementary code point.
+  static int32_t Decode(int32_t lead, int32_t trail) {
+    return 0x10000 + ((lead & 0x3FF) << 10) + (trail & 0x3FF);
+  }
+
+  // Encodes a single code point.
+  static void Encode(int32_t codepoint, uint16_t* dst);
+};
+
+
 class CaseMapping : AllStatic {
  public:
   // Maps a code point to uppercase.
   static int32_t ToUpper(int32_t code_point) {
     return Convert(code_point, kUppercase);
   }
 
   // Maps a code point to lowercase.
   static int32_t ToLower(int32_t code_point) {
     return Convert(code_point, kLowercase);
@@ skipping 46 matching lines @@
   // Data for small code points with one mapping
   static const int16_t stage2_[];
 
   // Data for large code points or code points with both mappings.
   static const int32_t stage2_exception_[][2];
 };
 
 }  // namespace dart
 
 #endif  // VM_UNICODE_H_