#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <sys/types.h>
#include <sys/stat.h>
// gcc -mno-cygwin -mwindows -o removetag removetag.c
// gcc -mno-cygwin -o removetag removetag.c

#define TAG_MAX_SIZE  	2048
#define TOKEN_MAX_SIZE	512
#define FAVOR     1
#define NOFAVOR   (!FAVOR)

int     link_print_flag = FAVOR ;

unsigned char *buffer ;
struct stat   fileInfo;

unsigned char 	getToken(unsigned char *buff, unsigned char *token, int n)
{
register int 	i, c ;

	for( i=0, *token='\0'; (c=*(buff+i)) != '\0' && i<=n; i++ ) {
		if( c != ' ' && c != '>'&& c != '\r' && c != '\n' ) {
			*(token+i) = tolower(c) ; *(token+i+1) = '\0' ;
		} else return(c) ;
	}
}

int	seekEndToken(unsigned char *buff, unsigned char *token)
{
unsigned char tkn[TOKEN_MAX_SIZE] ;
register int 	i, c ;

	for( i=1; (c=*(buff+i)) != '\0'; i++ ) {
		if ( c == '<' ) {
			getToken((buff+i), tkn, strlen(token)) ;
			strcat( tkn, ">") ;
			if( strcmp( token, tkn ) == 0 ) {
				return( i + strlen(token) -1 ) ;
			}
		}
	}
	return( i ) ;
}

unsigned char *checkStringsInTag(unsigned char *tag, unsigned char *strings)
{
unsigned char lowercase[TAG_MAX_SIZE] ;
register int 	i, length = (TAG_MAX_SIZE -1) > strlen(tag) ? strlen(tag) : TAG_MAX_SIZE -1 ;

	for( i=0; i<length; i++ ) {
		lowercase[i] = tolower(*(tag +i)) ; lowercase[i+1] = '\0';
	}
	return( strstr(lowercase, strings) ) ;
}

int	linkTagOperation(unsigned char *buff, int size, int flag, FILE *fpo)
{
unsigned char link[TAG_MAX_SIZE] ;
register int 	i, j ;

	link[0] = '\0' ;
	for(i=j=0; *(buff+i) != '>' && *(buff+i) != '\0' && !(j>0 && *(buff+i)=='<'); i++) {
		if( j < TAG_MAX_SIZE -3 ) {
			link[j] = *(buff+i); link[j+1] = '>'; link[j+2] = '\0';
			j++;
		}
	}
	if( *(buff+i) == '<' || *(buff+i) == '\0' ) {
		fprintf( stderr, "Bad file format!\n" );
		if( *(buff+i) == '\0' ) {
			exit(1);
		}
	} else if( checkStringsInTag(link, "http") == NULL && flag == FAVOR) {
		fprintf( fpo, "%s", link ) ;
	}
	return( i ) ;
}

execution(unsigned char *buff, int size, FILE *fpo)
{
unsigned char token[TOKEN_MAX_SIZE], link[TAG_MAX_SIZE] ;
register int 	i, j ;

	for( i=0; i<size; i++ ) {
		if( *(buff+i) == '<' &&  *(buff+i+1) != '/' ) {
			getToken((buff+i), token, sizeof(token)-1) ;
			if( strcmp( token, "<script") == 0 ) {
				i += seekEndToken( (buff+i), "</script>") ;
			} else if( strcmp( token, "<noscript") == 0 ) {
				i += seekEndToken( (buff+i), "</noscript>") ;
			} else if( strcmp( token, "<iframe") == 0 ) {
				i += seekEndToken( (buff+i), "</iframe>") ;
			} else if( strcmp( token, "<link") == 0 ) {
				i += linkTagOperation( buff+i,  size -i+1, link_print_flag, fpo ) ;
			} else 	fputc( *(buff+i), fpo ) ;
		} else	fputc( *(buff+i), fpo ) ;
	}
}

read_data(unsigned char *buff, int size, FILE *fpi)
{
register int 	i, c ;

	for( i=0; i<size ; i++ ) {
		if((c=fgetc(fpi)) == EOF ) {
			fprintf( stderr, "Unexpected EOF was found.\n" );
		} else *(buff+i) = c ;
	}
	*(buff+size) = '\0' ;
}

main( int argc, char *argv[])
{
FILE	*fpi, *fpo ;
register int 	data_size = 0 ;

	if( argc != 2 ) {
		fprintf( stderr, "usage: %s html_file\n", argv[0] );
		exit( 1 ) ;
	} else if( (fpi=fopen(argv[1],"rb")) == NULL ) {
		fprintf( stderr, "File '%s' open error!\n", argv[1] );
		exit( 1 ) ;
	}
	if ( fstat(fileno(fpi), &fileInfo) != 0 ) {
		fprintf( stderr, "fstat failed!\n" );
		exit( 1 ) ;
	} else data_size = fileInfo.st_size ;

	buffer = (unsigned char *)malloc(data_size +1) ;
	if(buffer == NULL) {
		fprintf( stderr, "Could not allocate memory!\n" );
		exit( 1 ) ;
	} else read_data(buffer, data_size, fpi) ;
	fclose(fpi) ;

	if((fpo=fopen(argv[1],"wb")) == NULL ) {
		fprintf( stderr, "File '%s' open error!", argv[1] );
		exit( 1 ) ;
	} else execution(buffer, data_size, fpo);
	fclose(fpo) ;

	free(buffer);
	return 0 ;
}