Fix #2459

lib/index.js

@@ -72,12 +72,54 @@ exports.encodePacket = function (packet, supportsBinary, utf8encode, callback) {
 
   // data fragment is optional
   if (undefined !== packet.data) {
-    encoded += utf8encode ? utf8.encode(String(packet.data)) : String(packet.data);
+    if (utf8encode) {
+      encoded += utf8.encode(sanitizeLoneSurrogates(String(packet.data)));
+    } else {
+      encoded += String(packet.data);
+    }
   }
 
   return callback('' + encoded);
 };
 
+/**
+ * Sanitize a WTF-8 string, replacing lone surrogates with
+ * U+FFFD 'REPLACEMENT CHARACTER'
+ */
+function sanitizeLoneSurrogates(str) {
+  var i = 0;
+  var output = '';
+  while (i < str.length) {
+    var code = str.charCodeAt(i);
+    if (code >= 0xd800 && code <= 0xdbff) {
+      // High surrogate -- expect next char to be a low surrogate
+      if (i + 1 < str.length) {
+        var nextCode = str.charCodeAt(i + 1);
+        if (nextCode >= 0xdc00 && nextCode <= 0xdfff) {
+          // Valid surrogate pair
+          output += str[i];
+          output += str[i + 1];
+          i++;
+        } else {
+          // High surrogate was not followed by a low surrogate
+          output += '\ufffd';
+        }
+      } else {
+        // High surrogate was at the end of a string (and therefore
+        // not followed by a low surrogate)
+        output += '\ufffd';
+      }
+    } else if (code >= 0xdc00 && code <= 0xdfff) {
+      // Low surrogate by itself -- invalid
+      output += '\ufffd';
+    } else {
+      output += str[i];
+    }
+    i++;
+  }
+  return output;
+}
+
 /**
  * Encode Buffer data
  */

test/node/index.js

@@ -25,7 +25,7 @@ describe('parser', function() {
     var data = new Buffer(5);
     for (var i = 0; i < data.length; i++) { data[i] = i; }
     encode({ type: 'message', data: data }, function(encoded) {
-      expect(decode(encoded)).to.eql({ type: 'message', data: data }); 
+      expect(decode(encoded)).to.eql({ type: 'message', data: data });
       done();
     });
   });
@@ -130,6 +130,22 @@ describe('parser', function() {
     });
   });
 
+  it('should encode a string message with lone surrogates replaced', function(done) {
+    var data = String.fromCharCode(0xd800);
+    encode({ type: 'message', data: data }, null, true, function(encoded) {
+      expect(decode(encoded, null, true)).to.eql({ type: 'message', data: '\ufffd' });
+      done();
+    });
+  });
+
+  it('should encode a string message with valid surrogate pairs', function(done) {
+    var data = String.fromCharCode(0xd800) + String.fromCharCode(0xdc00);
+    encode({ type: 'message', data: data }, null, true, function(encoded) {
+      expect(decode(encoded, null, true)).to.eql({ type: 'message', data: data });
+      done();
+    });
+  });
+
 });
 
 function areArraysEqual(x, y) {