C++模拟Http/Https访问web站点

一、概述

1．Http与Https的区别与联系

在OSI参考模型中Http与Https均属于应用层协议。Http即Hypertext Transfer Protocol，超文本传输协议；而Https为Secure Hypertext Transfer Protocol安全超文本传输协议，它是一个安全通信通道，基于HTTP开发，用于在客户端与服务器之间交换信息，它使用安全套接字层SSL进行信息交换，简单来说它就是HTTP的安全版。

Http默认使用80端口，Https使用443端口。

Http的数据在网络上是明文传输，而Https则是通过加密后的传输，因此它相比http会更加安全，但是由于需要额外加解密操作，因为Https的效率没有那么高。在登录Https站点和Http站点时，可以明显感觉到性能差异。

2．关于web访问的安全性

当前许多的web站点登录时都是采用普通的http进行传输，这种方式有着极大的安全隐患。当前web开发登录系统常用的有以下四种方式：

1) 账号和密码完全没有加密，明文传送。这种方式的安全级别是最低的，它无疑是将自己的账号和密码直接暴露给别人，通过抓包工具（例：WireShark）可以很容易的截获到账号和密码。

2) 密码采用MD5或其它加密方式进行加密，声称不可破解。其实，完全没有必要破解，只要截获加密后的密码串，就可以以你的身份访问服务器，这样也是可以通过认证授权的。这种方式在加密程度上有了一定程度的提高，但仍是不安全的。

3) 客户端在登录前去服务端拿一次密钥，通过该密钥进行加密，而服务器端的密钥是随机生成的，每次访问均会用不同的密钥。这种方式的安全性比较高。

4) 采用“安全性最高”的HTTPS方式传输，客户端与服务端会经过认证，且中间的传输数据全部进行加密。之所以在安全性最高上加引号，是因为它也不是绝对安全的，比如前段时间Openssl曝出安全漏洞，大名鼎鼎的“心脏出血”，黑客利用它的一个memcpy的bug，可以从溢出的内存中拿到64K的用户数据，导致用户信息泄露。但是这个安全性级别相对前面三个是最高的，当前服务端的证书一年收费大约3-5千，用这点钱换来相对安全，是很划算的事情了。

二、SOCKET发送HTTP请求

1．基本流程

无论是Http还是Https都是基于TCP进行传输的，因此使用SOCKET模拟HTTP访问web站点的方式，很简单，就是将头部数据拼接成数据包，发送给服务端，然后接收返回再解析就可以了。

其基本流程和编写普通SOCKET通信是一样的。Windows下的流程为：

a. WSAStartup对Winsock服务进行初始化

b. 建立socket套接字

c. connect连接服务端

d. send发送数据

e. recv接收数据

下面，以某站点的登录为例，利用Fiddler抓到的POST的头部信息如下：

                                             

这样，我们就可以构建这样的数据包发送出去，然后接收响应了，C++实现核心代码请见下部分。

2．核心代码

BOOL SocketClient::ConnectToServer(const CString strServerUrl, const int nPort)
{
    cstrServerUrl = strServerUrl;
    nServerPort = nPort;
    BOOL bRet = FALSE;

    do
    {
        if (!InitializeContext())
        {
            break;
        }

        if(!Connect())
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);
    return bRet;
}

BOOL SocketClient::LoginToServer(const CString strUsername, const CString strPasswd)
{
    cstrUserName = strUsername;
    cstrPassWord = strPasswd;
    BOOL bRet = FALSE;

    do
    {
        if (!SendPostData())
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL SocketClient::LogoutOfServer()
{
    return FALSE;
}

BOOL SocketClient::InitializeContext()
{
    BOOL bRet = FALSE;
    wsaData = new WSADATA;
    WORD wVersion = MAKEWORD(2, 2);

    do
    {
        if(0 != WSAStartup(wVersion, wsaData))
        {
            break;
        }

        if(LOBYTE( wsaData->wVersion ) != 2 || HIBYTE( wsaData->wVersion ) != 2 )
        {
            WSACleanup();
            break;
        }

        LPHOSTENT lpHostTent;
        lpHostTent = gethostbyname(cstrServerUrl);
        if (NULL == lpHostTent)
        {
            break;
        }

        socketClient = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
        if (socketClient == INVALID_SOCKET)
        {
            WSACleanup();
            break;
        }

        socketAddrClient = new SOCKADDR_IN;
        socketAddrClient->sin_family = AF_INET;
        socketAddrClient->sin_port = htons(nServerPort);
        socketAddrClient->sin_addr = *((LPIN_ADDR)*lpHostTent->h_addr_list);
        memset(socketAddrClient->sin_zero, 0, sizeof(socketAddrClient->sin_zero));

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL SocketClient::Connect()
{
    BOOL bRet = FALSE;

    do
    {
        if (SOCKET_ERROR == connect(socketClient, (LPSOCKADDR)socketAddrClient, sizeof(SOCKADDR_IN)))
        {
             int nErrorCode = WSAGetLastError();
            closesocket(socketClient);
            break;
        }

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL SocketClient::SendPostData()
{
    CString cstrSendData;
    CString cstrSendParam = "redirect=&username="+cstrUserName+"&password="+cstrPassWord+"&auto_login=checked&submit=%E7%99%BB%E5%BD%95";
    BOOL bRet = FALSE;

    CString cstrSendParamLen;
    cstrSendParamLen.Format("%d", cstrSendParam.GetLength());

    cstrSendData = "POST http://account.vsochina.com/user/login HTTP/1.1
";
    cstrSendData += "Host: account.vsochina.com
";
    cstrSendData += "User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:35.0) Gecko/20100101 Firefox/35.0
";
    cstrSendData += "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
";
    cstrSendData += "Accept-Language: zh-cn,zh;q=0.8,en-us;q=0.5,en;q=0.3
";
    cstrSendData += "Accept-Encoding: gzip, deflate
";
    cstrSendData += "DNT: 1
";
    cstrSendData += "Referer: http://account.vsochina.com/user/login
";
    cstrSendData += "Connection: keep-alive
";
    cstrSendData += "Content-Type: application/x-www-form-urlencoded
";
    cstrSendData += "Content-Length: " + cstrSendParamLen +"
";
    cstrSendData += "
";
    cstrSendData += cstrSendParam;

    CString cstrRecvData;
    do
    {
        if (-1 == send(socketClient, cstrSendData.GetBuffer(), cstrSendData.GetLength(), 0))
        {
            break;
        }

        char recvData[1000] = {0};
        int nRecvLen;

        while((nRecvLen = recv(socketClient, recvData, sizeof(recvData), 0)) > 0)
        {
            cstrRecvData += recvData;
        }

        if (cstrRecvData.GetLength() == 0)
        {
            break;
        }

        ParseCookieFromRecvData(cstrRecvData);

        //!判断返回的COOKIE信息中，UID是否存在
        if (cstrCookieUid.IsEmpty())
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);


    return bRet;
}

void SocketClient::ParseCookieFromRecvData(const CString cstrRecvData)
{
    list<CString> lstCookiesLine;        //!存放Set-Cookie的一行，例：Set-Cookie: vso_uname=houqd_1111;
    CString cstrFind = "Set-Cookie:";    //!查找标记
    CString cstrSeperator = "
";      //!以"
"分割号来分割字符串

    int nPos = 0;
    int nStart = cstrRecvData.Find(cstrSeperator);

    while(nStart != -1)
    {
        CString cstrSessionLine = cstrRecvData.Mid(nPos, nStart - nPos + 1);

        if (cstrSessionLine.Find(cstrFind) != -1)
        {
            CString cstrRealRecord = cstrSessionLine.Right(cstrSessionLine.GetLength() - cstrFind.GetLength() - 3);
            list<CString>::iterator it = find(lstCookiesLine.begin(), lstCookiesLine.end(), cstrRealRecord);
            if (it == lstCookiesLine.end())
            {
                lstCookiesLine.push_back(cstrRealRecord);
            }
        }

        nPos = nStart;
        nStart = cstrRecvData.Find(cstrSeperator, nPos + 2);
    }

    //!根据每行获取的cookie值，解析为key-value的形式
    vector<CString> vecCookieSet;
    for (list<CString>::iterator it = lstCookiesLine.begin(); it != lstCookiesLine.end(); it++)
    {
        CString cstrCookies = *it;
        CString cstrSeperator = ";";
        StaticUtility::StringSplit(cstrCookies, cstrSeperator, vecCookieSet);
    }

    vector<CString> vecTemp;
    for (vector<CString>::iterator it = vecCookieSet.begin(); it != vecCookieSet.end(); it++)
    {
        vecTemp.clear();
        CString cstrOneCookies = *it;
        CString cstrSeperator = "=";

        StaticUtility::StringSplit(cstrOneCookies, cstrSeperator, vecTemp);
        CString cstrKey = vecTemp[0];
        CString cstrVal = vecTemp[1];

        if(cstrKey.Compare("vso_uid") == 0)
        {
            cstrCookieUid = cstrVal;
            break;
        }
    }
}

通过接收来的头部信息中，将cookie信息解析出来，就可以判断是否登录成功了。然后，如果有或许的操作，在请求中挂上这些cookie信息，就可以获取想要的数据，完成想要的操作了。

三、OpenSSL发送HTTPS请求

1．基本流程

HTTPS=HTTP + SSL，因此利用OpenSSL发送请求给HTTPS站点和第二章的SOCKET发送HTTP是非常相似的，只不过要在原生的套接字上套上SSL层，基本流程如下：

a. WSAStartup对Winsock服务进行初始化

b. 建立socket套接字

c. connect连接服务端

d. 建立SSL上下文

e. 建立SSL

f. 将SSL与前面建立的socket套接字绑定

g. SSL_write()发送数据

h. SSL_read()接收数据

下面以小米官网站点的登录为例，来展示利用OpenSSL如何访问HTTPS站点，模拟登陆，核心代码，见下一章节。

2．核心代码

#pragma comment( lib, "libeay32.lib" )
#pragma comment( lib, "ssleay32.lib" )
HttpsClient::HttpsClient(void):
             wsaData(NULL),
             socketAddrClient(NULL),
             ssl(NULL),
             sslCtx(NULL),
             sslMethod(NULL),
             serverCertification(NULL)
{
    SSL_load_error_strings();
    SSLeay_add_ssl_algorithms();
}


HttpsClient::~HttpsClient(void)
{
    //!清理打开的句柄
    if (NULL != ssl)
    {
        SSL_shutdown(ssl);
        closesocket(socketClient);
        SSL_free(ssl);
        ssl = NULL;
    }

    if (NULL != sslCtx)
    {
        SSL_CTX_free(sslCtx);
    }

    WSACleanup();
}

BOOL HttpsClient::ConnectToServer(const CString strServerUrl, const int nPort)
{
    cstrServerUrl = strServerUrl;
    nServerPort = nPort;
    BOOL bRet = FALSE;

    do
    {
        if (!InitializeSocketContext())
        {
            break;
        }

        if (!SocketConnect())
        {
            break;
        }

        if (!InitializeSslContext())
        {
            break;
        }

        if (!SslConnect())
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);
    return bRet;
}

BOOL HttpsClient::LoginToServer(const CString strUsername, const CString strPasswd)
{
    cstrUserName = strUsername;
    cstrPassWord = strPasswd;
    BOOL bRet = FALSE;

    do
    {
        if (!SendLoginPostData())
        {
            break;
        }

        CString cstrRecvData;
        RecvLoginPostData(cstrRecvData);
        if (cstrRecvData.GetLength() == 0)
        {
            break;
        }

        ParseCookieFromRecvData(cstrRecvData);

        if (cstrCookieUid.IsEmpty() || cstrCookieUid.Compare("EXPIRED") == 0)
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);
    return bRet;
}

BOOL HttpsClient::LogoutOfServer()
{
    return FALSE;
}

BOOL HttpsClient::InitializeSocketContext()
{
    //!初始化winSocket环境
    BOOL bRet = FALSE;
    wsaData = new WSADATA;
    WORD wVersion = MAKEWORD(2, 2);

    do
    {
        if(0 != WSAStartup(wVersion, wsaData))
        {
            break;
        }

        if(LOBYTE( wsaData->wVersion ) != 2 || HIBYTE( wsaData->wVersion ) != 2 )
        {
            WSACleanup();
            break;
        }

        LPHOSTENT lpHostTent;
        lpHostTent = gethostbyname(cstrServerUrl);
        if (NULL == lpHostTent)
        {
            break;
        }

        socketClient = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
        if (socketClient == INVALID_SOCKET)
        {
            WSACleanup();
            break;
        }

        socketAddrClient = new SOCKADDR_IN;
        socketAddrClient->sin_family = AF_INET;
        socketAddrClient->sin_port = htons(nServerPort);
        socketAddrClient->sin_addr = *((LPIN_ADDR)*lpHostTent->h_addr_list);
        memset(socketAddrClient->sin_zero, 0, sizeof(socketAddrClient->sin_zero));

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL HttpsClient::SocketConnect()
{
    //!原生socket连接
    BOOL bRet = FALSE;

    do
    {
        if (SOCKET_ERROR == connect(socketClient, (LPSOCKADDR)socketAddrClient, sizeof(SOCKADDR_IN)))
        {
            int nErrorCode = WSAGetLastError();
            closesocket(socketClient);
            break;
        }

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL HttpsClient::InitializeSslContext()
{
    //!SSL通信初始化
    BOOL bRet = FALSE;

    do
    {
        sslMethod = SSLv23_client_method();
        if(NULL == sslMethod)
        {
            break;
        }

        sslCtx = SSL_CTX_new(sslMethod);
        if (NULL == sslCtx)
        {
            break;
        }

        ssl = SSL_new(sslCtx);
        if (NULL == ssl)
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL HttpsClient::SslConnect()
{
    //!SSL绑定原生socket,并连接服务器
    BOOL bRet = FALSE;

    do
    {
        SSL_set_fd(ssl, socketClient);

        int nRet = SSL_connect(ssl);
        if (-1 == nRet)
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL HttpsClient::SslGetCipherAndCertification()
{
    BOOL bRet = FALSE;

    do
    {
        cstrSslCipher = SSL_get_cipher(ssl);
        serverCertification = SSL_get_certificate(ssl);

        if (NULL == serverCertification)
        {
            break;
        }

        cstrSslSubject = X509_NAME_oneline(X509_get_subject_name(serverCertification), 0, 0);
        cstrSslIssuer = X509_NAME_oneline(X509_get_issuer_name(serverCertification), 0, 0);

        X509_free(serverCertification);

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

BOOL HttpsClient::SendLoginPostData()
{
    CString cstrSendData;
    //CString cstrSendParam = "redirect=&username="+cstrUserName+"&password="+cstrPassWord+"&auto_login=checked&submit=%E7%99%BB%E5%BD%95";
    CString cstrSendParam = "user="+cstrUserName+"&_json=true&pwd="+cstrPassWord+"&callback=http%3A%2F%2Forder.mi.com%2Flogin%2Fcallback%3Ffollowup%3Dhttp%253A%252F%252Fwww.mi.com%252F%26sign%3DNWU4MzRmNjBhZmU4MDRmNmZkYzVjMTZhMGVlMGFmMTllMGY0ZTNhZQ%2C%2C&sid=mi_eshop&qs=%253Fcallback%253Dhttp%25253A%25252F%25252Forder.mi.com%25252Flogin%25252Fcallback%25253Ffollowup%25253Dhttp%2525253A%2525252F%2525252Fwww.mi.com%2525252F%252526sign%25253DNWU4MzRmNjBhZmU4MDRmNmZkYzVjMTZhMGVlMGFmMTllMGY0ZTNhZQ%25252C%25252C%2526sid%253Dmi_eshop&hidden=&_sign=%2Bw73Dr7cAfRlMfOR6fW%2BF0QG4jE%3D&serviceParam=%7B%22checkSafePhone%22%3Afalse%7D&captCode=";
    BOOL bRet = FALSE;

    CString cstrSendParamLen;
    cstrSendParamLen.Format("%d", cstrSendParam.GetLength());

    cstrSendData = "POST https://account.xiaomi.com/pass/serviceLoginAuth2 HTTP/1.1
";
    cstrSendData += "Host: account.xiaomi.com
";
    cstrSendData += "User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:35.0) Gecko/20100101 Firefox/35.0
";
    cstrSendData += "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
";
    cstrSendData += "Accept-Language: zh-cn,zh;q=0.8,en-us;q=0.5,en;q=0.3
";
    cstrSendData += "Accept-Encoding: gzip, deflate
";
    cstrSendData += "DNT: 1
";
    cstrSendData += "Content-Type: application/x-www-form-urlencoded; charset=UTF-8
";
    cstrSendData += "Referer: https://account.xiaomi.com/pass/serviceLogin?callback=http%3A%2F%2Forder.mi.com%2Flogin%2Fcallback%3Ffollowup%3Dhttp%253A%252F%252Fwww.mi.com%252Findex.html%26sign%3DNDRhYjQwYmNlZTg2ZGJhZjI0MTJjY2ZiMTNiZWExODMwYjkwNzg2ZQ%2C%2C&sid=mi_eshop
";
    cstrSendData += "Content-Length: " + cstrSendParamLen +"
";
    cstrSendData += "Connection: keep-alive
";
    cstrSendData += "
";
    cstrSendData += cstrSendParam;

    CString cstrRecvData;
    do
    {
        int nRet = SSL_write(ssl, cstrSendData, cstrSendData.GetLength());

        if(-1 == nRet)
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);

    return bRet;
}

void HttpsClient::RecvLoginPostData(CString &cstrRecvData)
{
    BOOL bRet = FALSE;

    do
    {
        TIMEVAL tval;
        tval.tv_sec = 20;
        tval.tv_usec = 0;

        while(TRUE)
        {
            FD_SET fds;
            FD_ZERO(&fds);
            FD_SET(socketClient, &fds);

            char recvData[1000] = {0};
            int nRecvLen;

            //int nSelect = select(FD_SETSIZE, &fds, NULL, NULL, &tval);
            //if (1 != nSelect)
            //{
            //  break;
            //}

            int nErr = SSL_read(ssl, recvData, sizeof(recvData));
            if (nErr <= 0)
            {
                break;
            }

            cstrRecvData += recvData;
        }

        if (cstrRecvData.GetLength() == 0)
        {
            break;
        }

        bRet = TRUE;
    } while (FALSE);
}

void HttpsClient::ParseCookieFromRecvData(const CString cstrRecvData)
{
    list<CString> lstCookiesLine;        //!存放Set-Cookie的一行，例：Set-Cookie: vso_uname=houqd_1111;
    CString cstrFind = "Set-Cookie:";    //!查找标记
    CString cstrSeperator = "
";      //!以"
"分割号来分割字符串

    int nPos = 0;
    int nStart = cstrRecvData.Find(cstrSeperator);

    while(nStart != -1)
    {
        CString cstrSessionLine = cstrRecvData.Mid(nPos, nStart - nPos + 1);

        if (cstrSessionLine.Find(cstrFind) != -1)
        {
            CString cstrRealRecord = cstrSessionLine.Right(cstrSessionLine.GetLength() - cstrFind.GetLength() - 3);
            list<CString>::iterator it = find(lstCookiesLine.begin(), lstCookiesLine.end(), cstrRealRecord);
            if (it == lstCookiesLine.end())
            {
                lstCookiesLine.push_back(cstrRealRecord);
            }
        }

        nPos = nStart;
        nStart = cstrRecvData.Find(cstrSeperator, nPos + 2);
    }

    //!根据每行获取的cookie值，解析为key-value的形式
    vector<CString> vecCookieSet;
    for (list<CString>::iterator it = lstCookiesLine.begin(); it != lstCookiesLine.end(); it++)
    {
        CString cstrCookies = *it;
        CString cstrSeperator = ";";
        StaticUtility::StringSplit(cstrCookies, cstrSeperator, vecCookieSet);
    }

    vector<CString> vecTemp;
    for (vector<CString>::iterator it = vecCookieSet.begin(); it != vecCookieSet.end(); it++)
    {
        vecTemp.clear();
        CString cstrOneCookies = *it;
        CString cstrSeperator = "=";

        StaticUtility::StringSplit(cstrOneCookies, cstrSeperator, vecTemp);
        CString cstrKey;
        CString cstrVal;

        if (vecTemp.size() == 2)
        {
            cstrKey = vecTemp[0];
            cstrVal = vecTemp[1];
        }

        if(cstrKey.Compare("userId") == 0)
        {
            cstrCookieUid = cstrVal;
            break;
        }
    }
}

同理，判断登录也是在返回的信息中拿cookie信息，再进行下一步操作。

四、参考资料及代码下载链接

1) 使用OpenSSL API进行安全编程：http://www.ibm.com/developerworks/cn/linux/l-openssl.html

2) SSL建立过程分析：http://blog.chinaunix.net/uid-127037-id-2919489.html

3) Openssl使用：http://www.cppblog.com/woomsg/archive/2008/11/03/64508.html

4) 知乎OpenSSL：http://www.zhihu.com/topic/19584279

5) Retrieving a file via HTTP：http://www.codeproject.com/Articles/1876/Retrieving-a-file-via-HTTP

6) Step in a Typical HTTP Client Application：https://msdn.microsoft.com/en-us/library/8yh4zs9e%28v=VS.80%29.aspx

7) Make a POST HTTP request over a socket：http://paul.grozav.info/2013/05/16/c-make-a-post-http-request-over-a-socket/

8) C++ STL中哈希表hash_map介绍：blog.csdn.NET/ddkxddkx/article/details/6555754

9) 我的代码实现的下载链接为：http://download.csdn.net/detail/houqingdong2012/8540317