使用HTTP协下载文件

通过发送HTTP请求，下载文件

头文件：

#ifndef __HTTP__
#define __HTTP__

#include <stdio.h>
#include <stdlib.h>
#include <string>
#include <winsock.h>

using namespace std;

/**
 * URL相关数据
 */
typedef struct _URLData {
    string url;     // 完整的url地址
    string path;    // 文件路径
    string host;    // 主机名
    string ip;      // 点分隔的ip
    int    port;    // 端口
    string resName; // 资源名字
} URLData;

/**
 * 解析url
 */
void parseURL(string url, URLData* data);

/**
 * 链接http服务器
 */
bool connectHttpServer(int socketId, string ip, int port);

/**
 * 生成请求文本
 */
void generateRequestTxt(URLData* data, string* requestTxt);

/**
 * 发送http请求
 */
bool sendHttpRequest(int socketId, string* requestTxt);

/**
 * 接受响应数据
 */
void recvHttpResponse(int socketId, string* responseTxt);

/**
 * 保存http数据
 */
bool saveHttpData(string* responseTxt, string filePath = "", string fileName = "");

/**
 * 获取http资源
 */
bool getHttpResource(string url, string filePath = "", string fileName = "");

#endif

代码文件：

#include "http.h"
#include "DNS/DNS.h"
#include "time.h"

/**
 * 解析url
 */
void parseURL(string url, URLData* data)
{
    int startPos = 0;
    string begin = url.substr(0, 4);
    if (begin == "http") {
        startPos = 7;
    }

    int i = startPos;
    int lastSplitPos = 0;
    for (int len = url.length(); i<len; i++) {
        unsigned char ch = url.at(i);

        if (ch == 47) {
            if (data->host.length() == 0) {
                data->host = url.substr(startPos, i - startPos);
                data->path = url.substr(i, url.length() - i);

                break;
            }

            lastSplitPos = i;
        }
    }
    // data->path = url.substr(iPath, url.length() - iPath);
    data->url = url;
    data->port = 80;

    int resNameLastPos = url.length()-1;
    int resNameFirstPos = 0;
    for (int i = url.length()-1; i>=0; i--) {
        unsigned char ch = url.at(i);
        
        if (ch == '.')
            resNameLastPos = i;
        else if (ch == '/') {
            resNameFirstPos = i + 1;
            break;
        }
    }
    data->resName = url.substr(resNameFirstPos, resNameLastPos - resNameFirstPos);
}

/**
 * 链接HTTP服务器
 */
bool connectHttpServer(int socketId, string ip, int port)
{
    // 初始套接字化地址
    struct sockaddr_in serverAddr;
    memset(&serverAddr, 0, sizeof(serverAddr));
    serverAddr.sin_family = AF_INET;
    serverAddr.sin_addr.S_un.S_addr = inet_addr(ip.c_str());
    serverAddr.sin_port = htons(port);

    if (connect(socketId, (struct sockaddr*)&serverAddr, sizeof(serverAddr)) == SOCKET_ERROR) {
        return false;
    }
    return true;
}

/**
 * 生成请求文本
 */
void generateRequestTxt(URLData* data, string* requestTxt)
{
    requestTxt->append("GET "); requestTxt->append(data->path); requestTxt->append(" HTTP/1.1
");
    requestTxt->append("Host: "); requestTxt->append(data->host); requestTxt->append("
");
    requestTxt->append("Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
"); 
    requestTxt->append("Accept-Language: zh-CN,zh;q=0.8
");
    requestTxt->append("User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.72 Safari/537.36
");
    requestTxt->append("Accept-Encoding: deflate,sdch
");
    requestTxt->append("Connection: close

");
}

/**
 * 发送http请求
 */
bool sendHttpRequest(int socketId, string* requestTxt)
{
    int sendBytes = send(socketId, requestTxt->c_str(), requestTxt->length(), 0);
    return sendBytes > 0;
}

/**
 * 接受响应数据
 */
void recvHttpResponse(int socketId, string* responseTxt)
{
    char buff[1024];

    int recvBytes = 0;
    while ((recvBytes = recv(socketId, buff, 1024, 0)) > 0) {
        responseTxt->append(buff, recvBytes);
    }
}

/**
 * 保存http数据
 */
bool saveHttpData(string* responseTxt, string filePath, string fileName)
{
    if (fileName == "") {
        char _fileName[100];
        sprintf(_fileName, "%lld", time(0));
        fileName = _fileName;
    }

    string fileExt = ".";

    int i = 0, len = 0;
    for (i = 0, len = responseTxt->length(); i<len; i++) {
        unsigned char ch = responseTxt->at(i);

        if (ch == 'C') {
            unsigned char metaKey[13], typeMetaKey[13] = "Content-Type";
            memset(metaKey, 0, 13);
            memcpy(metaKey, responseTxt->c_str() + i, 12);

            string metaValue;
            if ( strcmp((char*)metaKey, (char*)typeMetaKey) == 0 ) {
                for (int j = i+13;;j++) {
                    unsigned char ch = responseTxt->at(j);

                    metaValue.push_back(ch);
                    if (ch == '
') {
                        for (int k = 0, len = metaValue.length(); k<len; k++) {
                            if (metaValue.at(k) == '/') {
                                fileExt.append( metaValue.substr(k+1, metaValue.length() - k - 2) );
                                fileName.append(fileExt);

                                break;
                            }
                        }

                        break;
                    }
                }
            }
        }

        if (ch == '
' && 
            responseTxt->at(i+1) == '
' &&
            responseTxt->at(i+2) == '
' &&
            responseTxt->at(i+3) == '
') {
            i += 4;
            break;
        }
    }

    FILE* fp = fopen((filePath + fileName).c_str(), "wb+");
    if (fp) {
        fwrite(responseTxt->c_str() + i, 1, responseTxt->length() - i, fp);
        fclose(fp);

        return true;
    }
    return false;
}

/**
 * 获取http资源
 */
bool getHttpResource(string url, string filePath, string fileName)
{
    printf("-------------------------------------
");

    int socketId = socket(AF_INET, SOCK_STREAM, 0);
    if (socketId == INVALID_SOCKET) {
        printf("Create sokcet failed
");

        return false;
    }

    // 解析url数据
    URLData data;
    parseURL(url, &data);

    // 获取DNS信息
    if (!getIPAddrByDNS(data.host, &data.ip)) {
        printf("Get ip(%s) from DNS server failed
", data.host.c_str());

        return false;
    }

    // 生成请求文本
    string requestTxt;
    generateRequestTxt(&data, &requestTxt);

    // 链接服务器
    if (!connectHttpServer(socketId, data.ip, data.port)) {
        printf("Connect http server failed, ip:%s
", data.ip.c_str());

        return false;
    }

    printf("Send http request:
%s
", requestTxt.c_str());

    // 发送请求
    if (!sendHttpRequest(socketId, &requestTxt)) {
        printf("Send http request failed
");

        return false;
    }

    // 接受响应
    string responseTxt;
    recvHttpResponse(socketId, &responseTxt);

    printf("Recive http response:
%s
", responseTxt.c_str());

    // 保存数据
    if (!saveHttpData(&responseTxt, filePath, (fileName.length() == 0) ? data.resName : fileName)) {
        printf("Save response data failed
");
        return false;
    }

    closesocket(socketId);

    printf("-------------------------------------
");

    return true;
}

使用方法：

string url = "http://static.youku.com/index/img/master.png";
getHttpResource(url);