Question

我正在尝试编写TcpClient实现。我想在处理数据时开始下一次读取，但我不确定何时传入BeginRead的字节数组被修改。这是我目前的实施

private void ProcessData()
{
    byte[] buffer = new byte[_tcpClient.ReceiveBufferSize];
    var networkStream = _tcpClient.GetStream();

    IAsyncResult result = networkStream.BeginRead(buffer, 0, buffer.Length, null, null);

    int read;
    while ((read = networkStream.EndRead(result)) != 0)
    {
        result = networkStream.BeginRead(buffer, 0, buffer.Length, null, null);

        ProcessChunk(buffer, read);
    }  
}

我在长时间运行的函数之前启动了下一个BeginRead因为我希望后台IO在我处理时开始获取下一个数据块。我是否可以只使用一个byte[]缓冲区执行此操作，还是需要每隔BeginRead次呼叫进行备用？

private void ProcessData()
{
    bool bufferChoice = false;
    byte[] bufferA = new byte[_tcpClient.ReceiveBufferSize];
    byte[] bufferB = new byte[_tcpClient.ReceiveBufferSize];
    var networkStream = _tcpClient.GetStream();

    IAsyncResult result = networkStream.BeginRead(bufferA, 0, bufferA.Length, null, null);

    int read;
    while ((read = networkStream.EndRead(result)) != 0)
    {
        if (bufferChoice)
        {
            result = networkStream.BeginRead(bufferA, 0, bufferA.Length, null, null);
            ProcessChunk(bufferB, read); //Call function with B's buffer
        }
        else
        {
            result = networkStream.BeginRead(bufferB, 0, bufferB.Length, null, null);
            ProcessChunk(bufferA, read); //Call function with A's buffer
        }

        bufferChoice = !bufferChoice;
    }
}

Answer 1

通常不保证何时使用传递给BeginRead的缓冲区。从理论上讲，它甚至可以在BeginRead调用期间同步填充。

旁注：您应该编写始终正常运行的代码，因为它比推测特定实现的特定行为更容易推理。在这种情况下，只需复制数据或使用多个缓冲区，这样就不会有出色的IO和处理共享相同的缓冲区。

何时填充BeginRead的缓冲区？

1 个答案: