MQTT PUBLISH packet encoder splits the topic and payload into different frames

The [Mqtt encoder](https://github.com/Azure/DotNetty/blob/dev/src/DotNetty.Codecs.Mqtt/MqttEncoder.cs#L234-L273) splits the PUBLISH packet into different frames, one containing the topic and QoS information, the other containing an optional payload data.

This causes the downstream socket implementation to send them as different different packets.
Since the [netty implementation](https://github.com/netty/netty/blob/4.1/codec-mqtt/src/main/java/io/netty/handler/codec/mqtt/MqttEncoder.java#L420-L456) encodes the entire PUBLISH packet into the same byte buffer, is there a particular reason dotnetty chose to implement it this way?

Related issue: https://github.com/Azure/azure-iot-sdk-csharp/issues/1814