kafka生产者发送消息的时候避免粘包的代码

通过观察构建ClientRequest的过程可知，其有一个RequestSend，其构造函数如下：

    public RequestSend(String destination, RequestHeader header, Struct body) {
    super(destination, serialize(header, body));
    this.header = header;
    this.body = body;
}

public static ByteBuffer serialize(RequestHeader header, Struct body) {
    ByteBuffer buffer = ByteBuffer.allocate(header.sizeOf() + body.sizeOf());
    header.writeTo(buffer);
    body.writeTo(buffer);
    buffer.rewind();
    return buffer;
}

那么，其extends的super干了啥

public NetworkSend(String destination, ByteBuffer... buffers) {
    super(destination, sizeDelimit(buffers));
}

private static ByteBuffer[] sizeDelimit(ByteBuffer[] buffers) {
    int size = 0;
    for (int i = 0; i < buffers.length; i++)
        size += buffers[i].remaining();
    ByteBuffer[] delimited = new ByteBuffer[buffers.length + 1];
    delimited[0] = ByteBuffer.allocate(4);//4字节的缓冲区
    delimited[0].putInt(size);
    delimited[0].rewind();
    System.arraycopy(buffers, 0, delimited, 1, buffers.length);
    return delimited;
}

继续看NetworkSend执行的super

//这个数组一共是3个元素，第一个是header+body的长度，为int，占四个字节,第二个是header的byteBuffer，第三个是body的ByteBuffer
//这个实现在NetworkSend内,这也是为什么需要实现分散和聚集的接口

protected final ByteBuffer[] buffers;//最终写的是这个ByteBuffer数组


public ByteBufferSend(String destination, ByteBuffer... buffers) {
    super();
    this.destination = destination;
    this.buffers = buffers;
    for (int i = 0; i < buffers.length; i++)
        remaining += buffers[i].remaining();
    this.size = remaining;
}

那nio里怎么写ByteBuffer数组？

使用分散和聚集！即：ScatteringByteChannel and GatheringByteChannel

所以，我们看到KafkaChannel的TransportLayer实现了ScatteringByteChannel and GatheringByteChannel