Add support of UTF-8!

message_generation/src/main/java/org/ros/internal/message/field/PrimitiveFieldType.java

@@ -28,6 +28,7 @@
 
 /**
  * @author damonkohler@google.com (Damon Kohler)
+ * @author mick.gaillard@gmail.com (Mickael Gaillard)
  */
 public enum PrimitiveFieldType implements FieldType {
 
@@ -571,7 +572,7 @@ public int getSerializedSize() {
     @Override
     public <T> void serialize(T value, ChannelBuffer buffer) {
       Preconditions.checkArgument(value instanceof String);
-      byte[] bytes = ((String) value).getBytes();
+      byte[] bytes = ((String) value).getBytes(DEFAULT_CHARSET);
       buffer.writeInt(bytes.length);
       buffer.writeBytes(bytes);
     }
@@ -581,7 +582,7 @@ public <T> void serialize(T value, ChannelBuffer buffer) {
     public String deserialize(ChannelBuffer buffer) {
       int length = buffer.readInt();
       ByteBuffer stringBuffer = buffer.readSlice(length).toByteBuffer();
-      return Charset.forName("US-ASCII").decode(stringBuffer).toString();
+      return DEFAULT_CHARSET.decode(stringBuffer).toString();
     }
 
     @SuppressWarnings("unchecked")
@@ -678,6 +679,7 @@ public String getJavaTypeName() {
     }
   };
 
+  private static final Charset DEFAULT_CHARSET = Charset.forName("UTF-8");
   private static final ImmutableSet<String> TYPE_NAMES;
 
   static {

message_generation/src/test/java/org/ros/internal/message/RawMessageSerializationTest.java

@@ -30,6 +30,7 @@
 
 /**
  * @author damonkohler@google.com (Damon Kohler)
+ * @author mick.gaillard@gmail.com (Mickael Gaillard)
  */
 public class RawMessageSerializationTest {
 
@@ -138,6 +139,71 @@ public void testString() {
     rawMessage.setString("data", "Hello, ROS!");
     checkSerializeAndDeserialize(rawMessage);
   }
+
+  @Test
+  public void testStringUTF8() {
+    RawMessage rawMessage = messageFactory.newFromType("std_msgs/String");
+    rawMessage.setString("data", "éêè €àáßëœ 文字化け");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // i18n test case
+    // base on http://www.inter-locale.com/whitepaper/learn/learn-to-test.html
+
+    // Combining Marks and Accents test
+    rawMessage.setString("data", "àéîōũ");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // DOS 860 test
+    rawMessage.setString("data", "você nós mãe avô irmã criança");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Windows-1252 test
+    rawMessage.setString("data", "€ŒœŠš™©‰ƒ");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Turkish test
+    rawMessage.setString("data", "ışık bir İyi Günler");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Dakuten and handakuten marks test
+    rawMessage.setString("data", "がざばだぱか゛さ゛た゛は");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Combining Grapheme Joiner character
+    rawMessage.setString("data", "אִ͏ַ");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Bidi with Latin test
+    rawMessage.setString("data", "abcאבגדabc ");
+    checkSerializeAndDeserialize(rawMessage);
+
+    rawMessage.setString("data", "אבגדabcאבגד");
+    checkSerializeAndDeserialize(rawMessage);
+
+    rawMessage.setString("data", "אבגד012אבגד");
+    checkSerializeAndDeserialize(rawMessage);
+
+    rawMessage.setString("data", "אבגד 012 012");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Complex Scripts test
+    rawMessage.setString("data", "สวัสดี");
+    checkSerializeAndDeserialize(rawMessage);
+
+    rawMessage.setString("data", "டாஹ்கோ");
+    checkSerializeAndDeserialize(rawMessage);
+
+    rawMessage.setString("data", "بِسْمِ اللّهِ الرَّحْمـَنِ الرَّحِيمِ");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Numeric Shaping test
+    rawMessage.setString("data", "عدد مارس ١٩٩٨");
+    checkSerializeAndDeserialize(rawMessage);
+
+    // Common Scripts and Encodings test
+    rawMessage.setString("data", "Слава Жанна Ювеналий Ярополк");
+    checkSerializeAndDeserialize(rawMessage);
+  }
 
   @Test
   public void testTime() {