Fix input and output to handle UTF16 surrogate pairs.

src/debug-agent.cc

 // Copyright 2009 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 354 matching lines @@
 }
 
 
 bool DebuggerAgentUtil::SendMessage(const Socket* conn,
                                     const Vector<uint16_t> message) {
   static const int kBufferSize = 80;
   char buffer[kBufferSize];  // Sending buffer both for header and body.
 
   // Calculate the message size in UTF-8 encoding.
   int utf8_len = 0;
+  int previous = unibrow::Utf8::kNoPreviousCharacter;
   for (int i = 0; i < message.length(); i++) {
-    utf8_len += unibrow::Utf8::Length(message[i]);
+    uint16_t character = message[i];
+    utf8_len += unibrow::Utf8::Length(character, previous);
+    previous = character;
   }
 
   // Send the header.
   int len;
   len = OS::SNPrintF(Vector<char>(buffer, kBufferSize),
                      "%s: %d\r\n", kContentLength, utf8_len);
   conn->Send(buffer, len);
 
   // Terminate header with empty line.
   len = OS::SNPrintF(Vector<char>(buffer, kBufferSize), "\r\n");
   conn->Send(buffer, len);
 
   // Send message body as UTF-8.
   int buffer_position = 0;  // Current buffer position.
+  previous = unibrow::Utf8::kNoPreviousCharacter;
   for (int i = 0; i < message.length(); i++) {
     // Write next UTF-8 encoded character to buffer.
+    uint16_t character = message[i];
     buffer_position +=
-        unibrow::Utf8::Encode(buffer + buffer_position, message[i]);
+        unibrow::Utf8::Encode(buffer + buffer_position, character, previous);
     ASSERT(buffer_position < kBufferSize);
 
     // Send buffer if full or last character is encoded.
     if (kBufferSize - buffer_position < 3 || i == message.length() - 1) {
-      conn->Send(buffer, buffer_position);
-      buffer_position = 0;
+      if (unibrow::Utf16::IsLeadSurrogate(character)) {
+        int kEncodedSurrogateLength = 3;

[rossberg, 2012/03/07 13:32:47]

Maybe move this declaration up and use it instead of literal 3 in the above
condition. Also, make it const for clarity.

[Erik Corry, 2012/03/11 19:29:22]

The 3 above is actually a different 3.  I gave them both names in unicode.h.

+        ASSERT(buffer_position >= kEncodedSurrogateLength);
+        conn->Send(buffer, buffer_position - kEncodedSurrogateLength);
+        for (int i = 0; i < kEncodedSurrogateLength; i++) {
+          buffer[i] = buffer[buffer_position + i];
+        }
+        buffer_position = kEncodedSurrogateLength;
+      } else {
+        conn->Send(buffer, buffer_position);
+        buffer_position = 0;
+      }
     }
+    previous = character;
   }
 
   return true;
 }
 
 
 bool DebuggerAgentUtil::SendMessage(const Socket* conn,
                                     const v8::Handle<v8::String> request) {
   static const int kBufferSize = 80;
   char buffer[kBufferSize];  // Sending buffer both for header and body.
@@ skipping 27 matching lines @@
       return total_received;
     }
     total_received += received;
   }
   return total_received;
 }
 
 } }  // namespace v8::internal
 
 #endif  // ENABLE_DEBUGGER_SUPPORT