Add support for surrogates when serializing and deserializing for native ports

runtime/vm/unicode.h

Index: runtime/vm/unicode.h
diff --git a/runtime/vm/unicode.h b/runtime/vm/unicode.h
index 03a4b29d879898c2c83f06aac4a175d4729db209..d922ce4d430750dc5aeceb68e84a1597f522af1f 100644
--- a/runtime/vm/unicode.h
+++ b/runtime/vm/unicode.h
@@ -1,4 +1,4 @@
-// Copyright (c) 2011, the Dart project authors.  Please see the AUTHORS file
+// Copyright (c) 2012, 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.
 
@@ -32,6 +32,11 @@ class Utf8 : AllStatic {
   // Returns true if 'utf8_array' is a valid UTF-8 string.
   static bool IsValid(const uint8_t* utf8_array, intptr_t array_len);
 
+  // Returns true if 'utf8_array' is a valid UTF-8 string allowing
+  // UTF-8 encoded UTF-16 surrogate code units.
+  static bool IsValidAllowSurrogates(
+      const uint8_t* utf8_array, intptr_t array_len);
+
   static intptr_t Length(int32_t ch);
   static intptr_t Length(const String& str);
 
@@ -41,6 +46,9 @@ class Utf8 : AllStatic {
   static intptr_t Decode(const uint8_t* utf8_array,
                          intptr_t array_len,
                          int32_t* ch);
+  static intptr_t DecodeAllowSurrogates(const uint8_t* utf8_array,
+                                        intptr_t array_len,
+                                        int32_t* ch);
 
   static bool DecodeToLatin1(const uint8_t* utf8_array,
                              intptr_t array_len,
@@ -50,6 +58,10 @@ class Utf8 : AllStatic {
                             intptr_t array_len,
                             uint16_t* dst,
                             intptr_t len);
+  static bool DecodeToUTF16AllowSurrogates(const uint8_t* utf8_array,
+                                           intptr_t array_len,
+                                           uint16_t* dst,
+                                           intptr_t len);
   static bool DecodeToUTF32(const uint8_t* utf8_array,
                             intptr_t array_len,
                             int32_t* dst,
@@ -73,6 +85,31 @@ class Utf16 : AllStatic {
 
   static const int32_t kSurrogateOffset = (0x10000 - (0xD800 << 10) - 0xDC00);
 
+  class CodePointIterator {
+   public:
+    explicit CodePointIterator(const uint16_t* utf16_array,
+                               intptr_t array_len)

[siva, 2012/11/27 03:00:25]

Has two parameters in the constructor why is explicit needed?

[Søren Gjesse, 2012/11/27 11:35:54]

Not needed - removed.

+        : utf16_array_(utf16_array),
+          array_len_(array_len),
+          index_(-1),
+          ch_(-1) {
+    }
+
+    int32_t Current() {

[siva, 2012/11/27 03:00:25]

int32_t Current() const {

[Søren Gjesse, 2012/11/27 11:35:54]

Done (added for String::CodePointIterator as well).

+      ASSERT(index_ >= 0);
+      ASSERT(index_ < array_len_);
+      return ch_;
+    }
+
+    bool Next();
+
+   private:
+    const uint16_t* utf16_array_;
+    intptr_t array_len_;
+    intptr_t index_;
+    int32_t ch_;

[siva, 2012/11/27 03:00:25]

Missing DISALLOW stuff?

[Søren Gjesse, 2012/11/27 11:35:54]

Done.

+  };
+
   // Returns the length of the code point in UTF-16 code units.
   static intptr_t Length(int32_t ch) {
     return (ch <= kMaxBmpCodepoint) ? 1 : 2;