diff --git a/src/messageId.cpp b/src/messageId.cpp
index db59ab6a..72cda2a0 100644
--- a/src/messageId.cpp
+++ b/src/messageId.cpp
@@ -118,7 +118,7 @@ void messageId::parse(const string& buffer, const string::size_type position,
 		// Extract left part
 		const string::size_type leftStart = position + (p - pstart);
 
-		while (p < pend && *p != '@') ++p;
+		while (p < pend && *p != '@' && *p != '>') ++p;
 
 		m_left = string(buffer.begin() + leftStart,
 		                buffer.begin() + position + (p - pstart));
@@ -178,6 +178,9 @@ ref <messageId> messageId::parseNext(const string& buffer, const string::size_ty
 
 const string messageId::getId() const
 {
+	if (m_right.empty())
+		return m_left;
+
 	return (m_left + '@' + m_right);
 }
 
@@ -193,7 +196,9 @@ void messageId::generate(utility::outputStream& os, const string::size_type maxL
 		pos = NEW_LINE_SEQUENCE_LENGTH;
 	}
 
-	os << '<' << m_left << '@' << m_right << '>';
+	os << '<' << m_left;
+	if (m_right != "") os << '@' << m_right;
+	os << '>';
 
 	if (newLinePos)
 		*newLinePos = pos + m_left.length() + m_right.length() + 3;
diff --git a/tests/parser/messageIdTest.cpp b/tests/parser/messageIdTest.cpp
index e70065a2..40e38ebf 100644
--- a/tests/parser/messageIdTest.cpp
+++ b/tests/parser/messageIdTest.cpp
@@ -59,12 +59,12 @@ VMIME_TEST_SUITE_BEGIN
 	{
 		vmime::messageId m1;
 
-		VASSERT_EQ("1", "<@>", m1.generate());
+		VASSERT_EQ("1", "<>", m1.generate());
 
 		vmime::messageId m2;
 		m2.setLeft("a");
 
-		VASSERT_EQ("2", "<a@>", m2.generate());
+		VASSERT_EQ("2", "<a>", m2.generate());
 
 		vmime::messageId m3;
 		m3.setRight("b");