Merge pull request #367 from msgpack/use-replace-mode-in-packer

Use REPLACE mode when converting java string into UTF8.

Author: xerial

msgpack-core/src/main/java/org/msgpack/core/MessagePacker.java

@@ -27,6 +27,7 @@
 import java.nio.charset.CharacterCodingException;
 import java.nio.charset.CharsetEncoder;
 import java.nio.charset.CoderResult;
+import java.nio.charset.CodingErrorAction;
 
 import static org.msgpack.core.MessagePack.Code.ARRAY16;
 import static org.msgpack.core.MessagePack.Code.ARRAY32;
@@ -447,7 +448,20 @@ private void packStringWithGetBytes(String s)
     private void prepareEncoder()
     {
         if (encoder == null) {
-            this.encoder = MessagePack.UTF8.newEncoder();
+            /**
+             * Even if String object contains invalid UTF-8 characters, we should not throw any exception.
+             *
+             * The following exception has happened before:
+             *
+             * org.msgpack.core.MessageStringCodingException: java.nio.charset.MalformedInputException: Input length = 1
+             *     at org.msgpack.core.MessagePacker.encodeStringToBufferAt(MessagePacker.java:467) ~[msgpack-core-0.8.6.jar:na]
+             *     at org.msgpack.core.MessagePacker.packString(MessagePacker.java:535) ~[msgpack-core-0.8.6.jar:na]
+             *
+             * This happened on JVM 7. But no ideas how to reproduce.
+             */
+            this.encoder = MessagePack.UTF8.newEncoder()
+                    .onMalformedInput(CodingErrorAction.REPLACE)
+                    .onUnmappableCharacter(CodingErrorAction.REPLACE);
         }
         encoder.reset();
     }