C++抓网页/获取网页内容
#include<windows.h>
#include<Wininet.h>
#include
#include
#include
#pragma comment(lib,"WinInet.lib")
using namespace std;
int main()
{
HINTERNET hINet, hHttpFile;
char szSizeBuffer[32];
DWORD dwLengthSizeBuffer = sizeof(szSizeBuffer);
hINet等于InternetOpen("IE6.0", INTERNET_OPEN_TYPE_PRECONFIG, NULL, NULL, 0 )；并且对InternetOpen进行WININET.DLL初始化。
string url = "http://www.sina.com"; //抓新浪网
if ( !hINet )
{
cout << "InternetOpen fail" << endl;
return 1;
}
hHttpFile = InternetOpenUrl(hINet, url.c_str(), NULL, 0, 0, 0); //该函数通过网络连接并从服务器接收数据
if(!hHttpFile)
{
cout << "error open url" << endl;
return 1;
}
BOOL bQuery = HttpQueryInfo(hHttpFile,
HTTP_QUERY_CONTENT_LENGTH,
szSizeBuffer,
&dwLengthSizeBuffer, NULL); //得到关于文件的信息
if(bQuery ==false)
{
InternetCloseHandle(hINet);
cout << "error query info" << endl;
return 3;
}
int FileSize=atol(szSizeBuffer); //atol函数把字符串转换成长整型数
string revData;
revData.resize(FileSize);
DWORD dwBytesRead;
BOOL bRead = InternetReadFile(hHttpFile, &revData[0], FileSize, &dwBytesRead); //该函数将循环读取数据块。
if(!bRead)
{
cout << "error to read file" << endl;
return 4;
}
ofstream out_file("duhui.txt");
out_file << revData; //输出到文件
InternetCloseHandle(hHttpFile); //关闭句柄
InternetCloseHandle(hINet);
cout << "抓取成功!/n" << endl;
system("pause");
return 0;
}