HTTP GET 명령을 통해 얻은 html 문서, 중간에 끊김

뭐가 잘못된 건지 도통 모르겠습니다..

무엇보다.. 아래 코드로 해서 돌아갔었거든요.. 해당 웹 사이트의 html 문서를 끝까지 다 읽어 왔었습니다..
그런데.. 되다가 왜 안 되는거죠?
lacovnk 님께서 지적해 주신 문제들 때문이라면, 아예 되지를 않아야 할텐데.. 한 두번 정상적으로 작동하다가 마네요.. 원래 HTTP 의 정상적인 작동 방식인가요?

아직 초보라.. 코드가 좀 엉성하지만.. 참고하시라고 한 번 올려봅니다..
ret_html 이라는 함수를 다른 함수에서 호출해서 사용합니다..

/**************************************************************************************
ret_html.c

DESCRIPTION: retrieves web page from web site which is designated by user in command line
DATE : 2005/2~
**************************************************************************************/

#include "engine_header.h"

//PORT : port number through which we can access target site
#define PORT 80

//MAXBUFSIZE : maximum capacity of buffer for retrieving web page from remote site
#define MAXBUFSIZE 20000

int ret_html(char *target_site,char *msg_to_server,pid_t crawler_pid,int doc_num)
{
int sockfd,numbytes;
char buf[MAXBUFSIZE+1]; //buffer for retrieving web text
struct hostent *he;
struct sockaddr_in their_addr;

char *prefix="txt_from_html"; //prefix to be attached to the output filename
char tmpfile[50]; //Temporary file storing raw html document
char txtfile[50];
//Text file which resulted from temporary html document by eliminating tags

//-------------------------------------------------------------------------------
//Create filename : tmp__.txt for raw html file
// __.txt for tag_eliminated file
//To show which crawler has retrieved that page, attach crawler_pid to filename
//To show document number , attach doc_num to file name
//To show this file is text file, attach expension '.txt' to the file name
//-------------------------------------------------------------------------------

sprintf(tmpfile,"%s%d%s%d%s","html_",(int)crawler_pid,"_",doc_num,".txt");
//is a temporary file name storing raw html document
sprintf(txtfile,"%s%s%d%s%d%s",prefix,"_",(int)crawler_pid,"_",doc_num,".txt");
//is a tag-eliminated text file name

//------------------------------------------------------------------------------
FILE *raw_html_fd; //FILE structure for raw html document
FILE *tag_eliminated_fd; //FILE structure for tag-eliminated html document

if((raw_html_fd=fopen(tmpfile,"w+"))==NULL)
{
fatal("Could not open temporary file for storing raw html document");
}

if((tag_eliminated_fd=fopen(txtfile,"w"))==NULL)
{
fatal("Could not open tag-eliminated file");
}
//------------------------------------------------------------------------------

if((he=gethostbyname(target_site))==NULL) {
herror("gethostbyname in ret_html()");
printf("%d\n",h_errno);
exit(1);
}

if((sockfd=socket(AF_INET,SOCK_STREAM,0))==-1)
{
fatal("socket in ret_html()");
}

their_addr.sin_family=AF_INET;
their_addr.sin_port=htons(PORT);
their_addr.sin_addr=*((struct in_addr *)he->h_addr);
bzero(&(their_addr.sin_zero),8);

if(connect(sockfd,(struct sockaddr *)&their_addr,sizeof(struct sockaddr))==-1)
{
fatal("connect in ret_html()");
}

int bytes_sent;

if((bytes_sent=send(sockfd,msg_to_server,strlen(msg_to_server),0))==-1)
{
fatal("send in ret_html()");
}

//----------------------------------------------------------------
//Following while loop retrieve numbytes bytes from target site
//Each time search engine retrieves web page from target site,
//----------------------------------------------------------------

int i; //for couting the number of zero bytes receptions

int totalbytes=0; //for total bytes read

while((numbytes=recv(sockfd,buf,MAXBUFSIZE,0)) != -1)
{

//Test ----------------------------------------
//printf("numbytes read : %d\n",numbytes);
//Test ----------------------------------------

if(numbytes == 0)
{
i++;

if(i == 5)
{
printf("\nZero bytes received over the predefined limit\n");
break;
}
else
{
// printf("Zero bytes received\n");
sleep(1);
continue;
}
}
else
{

totalbytes += numbytes;

if(fwrite((void *)buf,sizeof(char),numbytes,raw_html_fd) != numbytes)
{
fatal("fputs in ret_html()");
}

//insert '\n' to output file to avoid 'line too long' error
putc('\n',raw_html_fd);

//Wait for receiving new message from remote web server
//(Essential part for retrieving full html document)
sleep(1);

}
}

//Test ------------------------------
printf("totalbytes : %d\n",totalbytes);
printf("numbytes : %d\n",numbytes);
//Test ------------------------------

//----------------------------------------
//Eliminate tags from raw html document
//----------------------------------------
tag_eliminator(raw_html_fd,tag_eliminated_fd);

//Test --------------
printf("raw HTML file : %s\n",tmpfile);
//------------------

//----------------------
//Close file descriptor
//----------------------
fclose(raw_html_fd);
fclose(tag_eliminated_fd);

//---------------------------
//Close socket descriptor
//----------------------------
close(sockfd);

return 0;
}