Add support for surrogates when serializing and deserializing for native ports

runtime/vm/dart_api_impl_test.cc

 // 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.
 
 #include "include/dart_api.h"
 #include "platform/assert.h"
 #include "platform/json.h"
 #include "platform/utils.h"
 #include "vm/class_finalizer.h"
 #include "vm/dart_api_impl.h"
@@ skipping 532 matching lines @@
 
   int32_t data32[] = { 'f', 'o', 'u', 'r', 0x10FFFF };
 
   Dart_Handle str32 = Dart_NewStringFromUTF32(data32, ARRAY_SIZE(data32));
   EXPECT_VALID(str32);
   EXPECT(Dart_IsString(str32));
   EXPECT(!Dart_IsExternalString(str32));
 }
 
 
+static void CheckToUTF8(Dart_Handle str,
+                        intptr_t expected_len,
+                        uint8_t* expected_utf8,
+                        intptr_t expected_utf8_len) {
+  intptr_t len;
+  EXPECT_VALID(Dart_StringLength(str, &len));
+  EXPECT_EQ(expected_len, len);
+  uint8_t* utf8_out;
+  EXPECT_VALID(Dart_StringToUTF8(str, &utf8_out, &len));
+  EXPECT_EQ(expected_utf8_len, len);
+  EXPECT(!memcmp(expected_utf8, utf8_out, expected_utf8_len));
+}
+
+
 TEST_CASE(NewString) {
   const char* ascii = "string";
   Dart_Handle ascii_str = NewString(ascii);
   EXPECT_VALID(ascii_str);
   EXPECT(Dart_IsString(ascii_str));
 
   const char* null = NULL;
   Dart_Handle null_str = NewString(null);
   EXPECT(Dart_IsError(null_str));
 
-  uint8_t data[] = { 0xE4, 0xBA, 0x8c };  // U+4E8C.
-  Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
-  EXPECT_VALID(utf8_str);
-  EXPECT(Dart_IsString(utf8_str));
+  {
+    uint8_t data[] = { 0xE4, 0xBA, 0x8c };  // U+4E8C.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    CheckToUTF8(utf8_str, 1, data, sizeof(data));
+  }
 
-  uint8_t invalid[] = { 0xE4, 0xBA };  // underflow.
-  Dart_Handle invalid_str = Dart_NewStringFromUTF8(invalid,
-                                                   ARRAY_SIZE(invalid));
-  EXPECT(Dart_IsError(invalid_str));
+  {
+    uint8_t data[] = { 0xE4, 0xBA, 0x8c };  // U+4E8C.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    CheckToUTF8(utf8_str, 1, data, sizeof(data));
+  }

[siva, 2012/11/28 18:22:46]

This test and the one above seem to test identical values (U+4E8C), why repeat
the test?

[Søren Gjesse, 2012/11/29 09:06:14]

This is all reverted.

+
+  {
+    uint8_t data[] = {0xf0, 0x90, 0x80, 0x80,   // U+10000.
+                      0xf0, 0x9f, 0x98, 0x81,   // U+1F601.
+                      0xf0, 0x9f, 0x98, 0xb7,   // U+1F637.
+                      0xf0, 0xa0, 0x80, 0x80};  // U+20000.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    CheckToUTF8(utf8_str, 8, data, sizeof(data));
+  }
+
+  {
+    uint8_t data[] = {0xed, 0xa0, 0x80};   // U+D800.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    CheckToUTF8(utf8_str, 1, data, sizeof(data));
+  }
+
+  {
+    uint8_t data[] = {0xed, 0xb0, 0x80};   // U+DC00.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    CheckToUTF8(utf8_str, 1, data, sizeof(data));
+  }
+
+  {
+    uint8_t data[] = {0xed, 0xa0, 0x80,   // U+D800.
+                      0xed, 0xb0, 0x80};  // U+DC00.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    uint8_t expected_out[] = {0xf0, 0x90, 0x80, 0x80};  // U+10000.
+    CheckToUTF8(utf8_str, 2, expected_out, sizeof(expected_out));
+  }
+
+  {
+    uint8_t data[] = {0xed, 0xb0, 0x80,   // U+DC00.
+                      0xed, 0xa0, 0x80};  // U+D800.
+    Dart_Handle utf8_str = Dart_NewStringFromUTF8(data, ARRAY_SIZE(data));
+    EXPECT_VALID(utf8_str);
+    EXPECT(Dart_IsString(utf8_str));
+    CheckToUTF8(utf8_str, 2, data, sizeof(data));
+  }
+
+  {
+    uint8_t invalid[] = { 0xE4, 0xBA };  // underflow.
+    Dart_Handle invalid_str = Dart_NewStringFromUTF8(invalid,
+                                                     ARRAY_SIZE(invalid));
+    EXPECT(Dart_IsError(invalid_str));
+  }
 }
 
 
 TEST_CASE(ExternalStringGetPeer) {
   Dart_Handle result;
 
   uint8_t data8[] = { 'o', 'n', 'e', 0x7F };
   int peer_data = 123;
   void* peer = NULL;
 
@@ skipping 6798 matching lines @@
   Dart_LoadSource(TestCase::lib(), url, source);
 
   dart_args[0] = Dart_NewInteger(1);
   result = Dart_Invoke(lib1, NewString("start"), 1, dart_args);
   EXPECT_VALID(result);
 }
 
 #endif  // defined(TARGET_ARCH_IA32) || defined(TARGET_ARCH_X64).
 
 }  // namespace dart