如何在读取包头并找出包长度后正确遍历包？

Question

鉴于此标题格式：

我正在尝试遍历并打印 Length 和 Num Market Updates：

std::ifstream fileOpener(m_rawInputFileName, std::ios::binary);
if (!fileOpener.is_open())
{
    std::cout << "Error: could not open file" << std::endl;
    fileOpener.close();
    return;
}

short int packetLength;
short int numOfMarketUpdates;
while (fileOpener.read((char *)&packetLength, 2) &&
        fileOpener.read((char *)&numOfMarketUpdates, 2))
{
    cout << packetLength << ":" << numOfMarketUpdates << endl;
    fileOpener.seekg(-4, std::ios::cur);
    fileOpener.seekg(packetLength, std::ios::cur);
}

我做错了什么，还是数据不准确？

Answer 1

回答：我漏掉了一个重要的细节：“所有数据都是按网络字节顺序排列的。” 因此，正确的代码是：

#include <arpa/inet.h>

std::ifstream fileOpener(m_rawInputFileName, std::ios::binary);
if (!fileOpener.is_open())
{
    std::cout << "Error: could not open file" << std::endl;
    return;
}

short int packetLength;
short int numOfMarketUpdates;
while (fileOpener.read((char *)&packetLength, 2) &&
        fileOpener.read((char *)&numOfMarketUpdates, 2))
{
    packetLength = nthos(packetLength);
    numOfMarketUpdates = nthos(numOfMarketUpdates);
    cout << packetLength << ":" << numOfMarketUpdates << endl;
    fileOpener.seekg(-4, std::ios::cur);
    fileOpener.seekg(packetLength, std::ios::cur);
}

nthos 将网络字节顺序转换为主机字节顺序。网络字节顺序配置为 big-endian，这可能与您的主机字节顺序不同，因此除非您将其转换为主机字节顺序，否则它没有意义。

感谢 user253751 指出这一点。