get_proxy_address.c

获得ip代理

#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <errno.h>
#include <netdb.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>

#define HTTP_REQ "GET /proxy1.html HTTP/1.1\r\n"\
"Host: www.cnproxy.com:80\r\n"\
"User-Agent: Mozilla/5.0\r\n"\
"Accept: text/xml,text/html;*/*\r\n\r\n"


int main(int argc, char *argv[])
{
	int sockfd;
	struct sockaddr_in servaddr;
	struct hostent *host = NULL;
	char **ptr = NULL, *p = NULL, *q = NULL;
	char buffer[65535];
	//char http_answer[4096];
	char html[65535];
	int recv_len, i, recved_len, table_num = 0;
	int code;
	char ip_port[30];
	unsigned long ip;

	sockfd = socket(AF_INET, SOCK_STREAM, 0);
	if (sockfd < 0) {
		perror("socket");
		return 0;
	}
	host = gethostbyname("www.cnproxy.com");
	if (host == NULL) {
		perror("gethostbyname");
		return 0;
	}
	if (host->h_addrtype != AF_INET) {
		printf("Address type error!\n");
		return 0;
	}
	bzero(&servaddr, sizeof(servaddr));
	for (ptr = host->h_addr_list; *ptr != NULL; ptr++) {
		//printf("address %s\n", inet_ntop(host->h_addrtype, *ptr, buffer, sizeof(buffer)));
		memmove(&ip, *ptr, host->h_length);
		//ip = ntohl(ip);
		memmove(&servaddr.sin_addr, &ip, sizeof(ip));
		break;
	}
	servaddr.sin_family = AF_INET;
	servaddr.sin_port = htons(80);

	if (connect(sockfd, (struct sockaddr *)&servaddr, sizeof(servaddr)) < 0) {
		perror("connect");
		return 0;
	}

	send(sockfd, HTTP_REQ, sizeof(HTTP_REQ), 0);
	recved_len = 0;
	html[0] = '\0';
	code = 0;
	while (1) {
rerecv:
		recv_len = recv(sockfd, buffer, sizeof(buffer), 0);
		//printf("%s\n", buffer);
		p = buffer;
		//strncpy(http_answer, buffer, sizeof(http_answer));
		if (code == 0) {
			p = strchr(buffer, ' ');
			p++;
			sscanf(p, "%d", &code);
			if (code != 200) {
				printf("Be rejected!\n");
				return 0;
			}
			p = strstr(buffer, "Content-Length");
			p += strlen("Content-Length");
			p++;
			sscanf(p, "%d", &recv_len);
			p = strstr(buffer, "\r\n\r\n");
			p += strlen("\r\n\r\n");
		}
		recved_len += strlen(p);
		while (1) {
			if (table_num == 3)
				break;
			p = strstr(p, "<table>");
			if (p == NULL)
				goto rerecv;
			p += strlen("<table>");
			table_num++;
			if (table_num == 3) {
				p += strlen("\r\n");
				p = strstr(p, "\r\n");
				if (p == NULL)
					goto rerecv;
				p += strlen("\r\n");
			}
		}
		q = strstr(p, "</table>");
		if (q == NULL) {
			strcat(html, p);
			goto rerecv;
		}
		else {
			strcat(html, p);
			break;
		}
	}
	//p = strstr(html, "\r\n");
	//p += strlen("\r\n");
	p = html;
	q = strstr(html, "</table>");
	//printf("%s\n*********************\n", p);
	while (p < q) {
		p = strstr(p, "<tr><td>");
		p += strlen("<tr><td>");
		for (i = 0; *p != '&'; p++, i++) {
			if (i >= 30)
				break;
			ip_port[i] = *p;
		}
		if (i >= 30)
			continue;
		ip_port[i] = ':';
		i++;
		p = strchr(p, ':');
		p++;
		for (; *p != '<'; p++, i++) {
			if (i >= 30)
				break;
			ip_port[i] = *p;
		}
		if (i >= 30)
			continue;
		ip_port[i] = '\0';
		if (judge_ip_port(ip_port) == 0)
			printf("%s\n", ip_port);
		p = strstr(p, "\r\n");
		p += strlen("\r\n");
	}
	close(sockfd);
	return 0;
}

int judge_ip_port(const char *ip_port)
{
	int ret = -1;
	const char *p = NULL;
	int i = 0;
	int tmp;
	
	p = ip_port;
	while (1) {
		p = strchr(ip_port, '.');
		if (p != NULL) {
			i++;
			if (sscanf(ip_port, "%d", &tmp) != 1) {
				goto end;
			}
			if (tmp >= 255 || tmp < 0)
				goto end;
			if (i >= 4)
				goto end;
			ip_port = p + 1;
		}
		else {
			break;
		}
	}
	if (i != 3)
		goto end;
	p = strchr(ip_port, ':');
	if (p == NULL)
		goto end;
	if (sscanf(++p, "%d", &tmp) != 1)
		goto end;
	if (tmp < 0 || tmp > 65535)
		goto end;
	ret = 0;
end:
	return ret;
}