23,121
社区成员
发帖
与我相关
我的任务
分享
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
//#ifndef _WIN32
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <netdb.h>
//#endif
char* host = "www.baidu.com"; //指定为抓取www.baidu.com页面
int port = 80;
int main(void)
{
char buffer[4096];
int isock;
struct sockaddr_in pin;
struct hostent * remoteHost;
char message[512];
int done = 0;
int chars = 0;
int l = 0;
if( (remoteHost = gethostbyname(host)) == 0 )
{
printf("Error resolving host\n");
exit(1);
}
bzero(message,sizeof(message));
bzero(&pin,sizeof(pin));
pin.sin_family = AF_INET;
pin.sin_port = htons(port);
pin.sin_addr.s_addr = ( (struct in_addr *)(remoteHost->h_addr) )->s_addr;
if( (isock = socket(AF_INET, SOCK_STREAM, 0)) == -1)
{
printf("Error opening socket!\n");
exit(1);
}
sprintf(message, "GET / HTTP/1.1\r\n");
strcat(message, "Host:www.baidu.com\r\n");
strcat(message, "Accept: */*\r\n");
strcat(message, "User-Agent: Mozilla/4.0(compatible)\r\n");
strcat(message, "connection:Keep-Alive\r\n");
strcat(message, "\r\n\r\n");
printf("%s",message);
if( connect(isock, /*(void *)*/ (const sockaddr*)&pin, sizeof(pin)) == -1 )
{
printf("Error connecting to socket\n");
exit(1);
}
if( send(isock, message, strlen(message), 0) == -1)
{
printf("Error in send\n");
exit(1);
}
while(done == 0)
{
l = recv(isock, buffer, 1, 0);
if( l < 0 )
done = 1;
switch(*buffer)
{
case '\r':
break;
case '\n':
if(chars == 0)
done = 1;
chars = 0;
break;
default:
chars++;
break;
}
printf("%c",*buffer);
}
printf("begin to crawling page\n");
do
{
l = recv(isock, buffer, sizeof(buffer) - 1, 0);
if( l < 0 )
break;
*(buffer + l) = 0;
fputs(buffer, stdout);
}while( l > 0 );
// close(isock);
shutdown(isock, SHUT_RDWR);
return 0;
}
while(done == 0)
{
l = recv(isock, buffer, 1, 0); /* 非阻塞模式,当服务器不发送数据时,就阻塞在这里*/
if( l < 0 )
done = 1;
switch(*buffer)
{
case '\r':
break;
case '\n':
if(chars == 0)
done = 1;
chars = 0;
break;
default:
chars++;
break;
}
printf("%c",*buffer);
}