socket_html.c

利用SOCKET编程做的一个获取html源码的程序

/***************************************************
  Copyright(C), 2008 , JUST
  File name: socket_html.c
  Author: Zhang Haidi   Version: 0.8     Date: 08/09/10
  Description: CSocket.c的头文件
	     完成网页的下载
***************************************************/
#include "./inc/socket_html.h" 
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdarg.h>
#include <string.h>
#include <errno.h>
#include <netdb.h>
#include <fcntl.h>
#include <sys/time.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include "./inc/t2s.h"

/*网页原文件保存名*/
#define DOWNLOAD_FILE_NAME "./wdt/getHtml.htm"
extern char MP3_NAME[100];
int is_mp3(char *filename);
int getMp3FileSon(char *hostip,char *file);
int getHtmlFileSon(char *hostip,char *file);

int getHtmlFile(char *hostip)
{
	char myhost[512];
	char filename[512];
	int i,j; 
	printf("hostip=\n\n%s\n\n\n",hostip);
	//sleep(1);
	if(strstr(hostip,"http")!=NULL)
	{
		//printf("1\n");
		//sleep(1);
		i=0;
		j=0;
		strcpy(myhost,"\0");
		strcpy(filename,"\0");
		while(hostip[i+7]!='/'&&hostip[i+7]!='\0')
		{
			myhost[i]=hostip[i+7];
			i++;
		}
		myhost[i]='\0';
		if(hostip[i+7]=='\0')
			strcpy(filename,"index.html");
		else if(hostip[i+8]=='\0')
			strcpy(filename,"index.html");
		else
		{
			while(hostip[i+7]!='\0')
			{
				filename[j++]=hostip[i+7];
				i++;
			}
			filename[j]='\0';	
		}
		printf("Host:<%s>\nFileName:<%s>\n",myhost,filename);
		//sleep(1);
	}	
	else{
		i=0;
		j=0;
		strcpy(myhost,"\0");
		strcpy(filename,"\0");
		while(hostip[i]!='/')
		{
			myhost[i]=hostip[i];
			i++;
		}
		myhost[i]='\0';
		if(hostip[i+7]=='\0')
			strcpy(filename,"index.html");
		else if(hostip[i+8]=='\0')
			strcpy(filename,"index.html");
		else
		{
			while(hostip[i+7]!='\0')
			{
				filename[j++]=hostip[i+7];
				i++;
			}
			filename[j]='\0';	
		}	
		printf("Host:%s\nFileName:%s\n",myhost,filename);	
	}
	
	printf("myhost=%s\nfilename=%s\n",myhost,filename);
	//sleep(1);
	if(is_mp3(filename))
	{
		strcpy(MP3_NAME,".");
		strcat(MP3_NAME,filename);
		return(getMp3FileSon(myhost,filename));
	}
	else  return(getHtmlFileSon(myhost,filename));
}


int is_mp3(char *filename)
{
	char *ptr;
	ptr=&(filename[strlen(filename)-3]);
	printf("ptr=%s\n",ptr);
	//sleep(1);
	if(strcmp(ptr,"mp3")==0||strcmp(ptr,"MP3")==0)
	{
		printf("is_mp3 return 1\n");
		//sleep(1);
		return 1;
	}
	else 
	{
		printf("is_mp3 return 0\n");
		//sleep(1);
		return 0;
	}
}

int getMp3FileSon(char *hostip,char *file)
{
	int fd;
	int flag=0,debug=0;
	FILE *fp;
	char buf[1024];
	int dnum;
	char quest[1024];
	char filename[60];
	int rc;
	struct sockaddr_in peer;
	char hosturl[40];
	struct hostent *hp; //= gethostbyname(argv[1]);

	printf("start...\n");
	strcpy(hosturl,hostip);
	hp=gethostbyname(hosturl);
	bzero( &peer, sizeof(peer) );
	peer.sin_family = AF_INET;
	printf("why? %s\n",hosturl);
	if ( hp == NULL )
	{
		printf("unknow host: %s\n",hosturl);
		return -1;
	}
	peer.sin_addr = *( ( struct in_addr * )hp->h_addr );
	peer.sin_port = htons(80);
	printf("test...\n");
	printf("conncet to %s :%d \n",inet_ntoa(peer.sin_addr),ntohs( peer.sin_port ));
	
	fd = socket( AF_INET, SOCK_STREAM, 0 );
	if ( fd < 0 )
	{
		printf("socket call failed\n");
		return -1;
	}

	if( connect( fd, (struct sockaddr *)&peer, sizeof( peer ) ) )
	{
		printf("%d connect failed \n",errno);
		close( fd );
		return -1;
	}
	sprintf(filename,file);
	sprintf(quest,"GET /%s HTTP/1.0\r\nUser-agent:Mozilla/4.0\r\nAccept-language:zh-cn\r\n\r\n",file);
	printf("I will read for you! %s \n------------------------\n%s\n",quest,filename);
	if((rc = send(fd, quest, sizeof(quest), 0)) == -1)
	{
		if((errno != EWOULDBLOCK) && (errno != EAGAIN))
		{
			printf("send conn error!\n");
			return -1;
		}
	}
	fp = fopen( MP3_NAME, "w+" );
	if(fp==NULL)
	{
		return -1;
	}
	flag=0;
	while((rc = recv( fd, buf, sizeof(buf) - 1, 0 )) > 0)
	{
		fwrite(buf, sizeof(char), rc, fp);
		printf("%d\x0d",debug++);
		//printf("%s",buf);
		if(flag%1000==0)
			T2S_t2s("下载中");
		flag+=1;
	}
	close( fd );
	fclose( fp );
	if(flag==0)
	{
		strcpy(buf,"rm -rf ");
		strcat(buf,MP3_NAME);
		system(buf);
		return -1;
	}
	else 
	{
		return 2;
	}
	//return flag-1;
}
int getHtmlFileSon(char *hostip,char *file)
{
	int fd;
	int flag=0;
	FILE *fp;
	char buf[1024];
	int dnum;
	char quest[1024];
	char filename[60];
	int rc;
	struct sockaddr_in peer;
	char hosturl[40];
	struct hostent *hp; //= gethostbyname(argv[1]);

	printf("start...\n");
	strcpy(hosturl,hostip);
	hp=gethostbyname(hosturl);
	bzero( &peer, sizeof(peer) );
	peer.sin_family = AF_INET;
	printf("why? %s\n",hosturl);
	if ( hp == NULL )
	{
		printf("unknow host: %s\n",hosturl);
		return 0;
	}
	peer.sin_addr = *( ( struct in_addr * )hp->h_addr );
	peer.sin_port = htons(80);
	printf("test...\n");
	printf("conncet to %s :%d \n",inet_ntoa(peer.sin_addr),ntohs( peer.sin_port ));
	
	fd = socket( AF_INET, SOCK_STREAM, 0 );
	if ( fd < 0 )
	{
		printf("socket call failed\n");
		return 0;
	}

	if( connect( fd, (struct sockaddr *)&peer, sizeof( peer ) ) )
	{
		printf("%d connect failed \n",errno);
		close( fd );
		return 0;
	}
	sprintf(filename,file);
	sprintf(quest,"GET /%s HTTP/1.0\r\nUser-agent:Mozilla/4.0\r\nAccept-language:zh-cn\r\n\r\n",file);
	printf("I will read for you! %s \n------------------------\n%s\n",quest,filename);
	if((rc = send(fd, quest, sizeof(quest), 0)) == -1)
	{
		if((errno != EWOULDBLOCK) && (errno != EAGAIN))
		{
			printf("send conn error!\n");
			return 0;
		}
	}
	fp = fopen( DOWNLOAD_FILE_NAME, "w+" );
	if(fp==NULL)
		return 0;
	flag=0;
	while((rc = recv( fd, buf, sizeof(buf) - 1, 0 )) > 0)
	{
		fwrite(buf, sizeof(char), rc, fp);
		//printf("%s",buf);
		flag+=1;
	}
	close( fd );
	fclose( fp );
	if(flag==0)
		return 0;
	else return 1;
	//return flag-1;
}