/*
BIG5	19782
        1st	81      FE				total = 126
        2nd	40      7E		63
                A1      FE		94		total = 157

        USR3    8140    8DFE    2041
        USR2    8E40    A0FE    2983
        SPC1    A140    A3BF     408
        CTRL    A3C0    A3E0      33
        ????    A3E1    A3FE      30
        STD1    A440    C67E    5401
        SPC2    C6A1    C8FE     408
        STD2    C940    F9D5    7652
        ETEN    F9D6    F9FE      41
        USR1    FA40    FEFE     785
*/
#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>

typedef unsigned char	byte;
typedef unsigned short	word;
typedef unsigned long	dword;
typedef union {
	word W;
	struct {
		byte L, H;
	} B;
} WB;

const char* crlf = "\r\n";

word hex_atoi( char *s )
{
	word    rc = 0;
	byte    ch;

	while( *s ) {
		ch = toupper( *s );
		s++;
		if( ( ch >= 'A' ) && ( ch <= 'F' ) )  {
			rc <<= 4;
			rc += ch - 'A' + 0x0A;
		}
		else if( ( ch >= '0' ) && ( ch <= '9' ) ) {
			rc <<= 4;
			rc += ch - '0';
		}
		else if( ch != '.' )
			break;
	}
	return rc;
}

// wkliang: 20100325 - org: B5PG = 0x81
#define B5PG	0xA1

int b5_cod2ord( byte c1, byte c2 )
{
	if( ( c1 < B5PG ) || ( c1 > 0xFE ) )
		return -1;
	c1 = c1 - B5PG;				// 126

	if( ( c2 >= 0x40 ) && ( c2 <= 0x7E ) )
		c2 = c2 - 0x40;			// 63
	else if( ( c2 >= 0xA1 ) && ( c2 <= 0xFE ) )
		c2 = c2 - 0xA1 + 63;	// 94
	else
		return -1;

	return( c1 * 157 + c2 );
}

void b5_ord2cod( int ord, byte *p1, byte *p2 )
{
	*p1 = ord / 157 + B5PG;
	*p2 = ord % 157;
	if( *p2 < 63 )
		*p2 += 0x40;
	else
		*p2 += -63 + 0xA1;
	return ;
}

int b5tbl_out( FILE* fp, int ord, int end, int esc )
{
	byte c1, c2;
	int i;

	b5_ord2cod( ord, &c1, &c2 );
	fprintf( stderr, "b5tbl_out fm %d:%02X%02X - ", ord, c1, c2 );

	b5_ord2cod( end, &c1, &c2 );
	fprintf( stderr, "to %d:%02X%02X - ", end, c1, c2 );

	for( i = 0; ord <= end; i++, ord++ ) {
		if ( esc >= 0 )
			b5_ord2cod( esc, &c1, &c2 );
	
		else
			b5_ord2cod( ord, &c1, &c2 );
		fwrite( &c1, 1, 1, fp );
		fwrite( &c2, 1, 1, fp );
		// printf( "%5d:%02X%02X\r", i, c1, c2 );
		// fflush( stdout );
		// sleep( 1 );
//		if( ( i % 38 ) == 37 )
//			fwrite( crlf, 1, 2, fp );
	}
//	fwrite( crlf, 1, 2, fp );
	fprintf( stderr, "generate: %d.\n", i );
	return i;

}

int b5tbl_gen( int argc, char *argv[] )
{
	FILE    *fp;
	int cnt = 0;

	if( argc < 2 ) {
		fprintf( stderr, "%s outfile\n", argv[0] );
		return EXIT_FAILURE;
	}

	if( ( fp = fopen( argv[1], "wb+" ) ) == NULL ) {
		perror( argv[1] );
		return EXIT_FAILURE;
	}
	// SPC1
	cnt += b5tbl_out( fp, b5_cod2ord(0xA1,0x40), b5_cod2ord(0xA3,0xBF), -1 );

	// CTRL + ???? should be escaped
	cnt += b5tbl_out( fp, b5_cod2ord(0xA3,0xC0), b5_cod2ord(0xA3,0xE0), 0 );
	cnt += b5tbl_out( fp, b5_cod2ord(0xA3,0xE1), b5_cod2ord(0xA3,0xFE), 0 );

	// STD1 + SPC2 + STD2 + ETEN
	cnt += b5tbl_out( fp, b5_cod2ord(0xA4,0x40), b5_cod2ord(0xC6,0x7E), -1 );
	cnt += b5tbl_out( fp, b5_cod2ord(0xC6,0xA1), b5_cod2ord(0xC8,0xFE), -1 );
	cnt += b5tbl_out( fp, b5_cod2ord(0xC9,0x40), b5_cod2ord(0xF9,0xD5), -1 );
	cnt += b5tbl_out( fp, b5_cod2ord(0xF9,0xD6), b5_cod2ord(0xF9,0xFE), -1 );

	fclose( fp );
	fprintf( stderr, "total = %d\n", cnt );
	return EXIT_SUCCESS;
}

/*
 * iconv -f big5 -t unicode infile >outfile
 */

int b52uni_load( char* fname, WB** plst )
{
	FILE    *fp;
	int     i, rc, num;
	WB*	list;

	if( ( fp = fopen( fname, "rb" ) ) == NULL ) {
		perror( fname );
		return -1;
	}
	fseek( fp, 0, SEEK_END );
	num = ftell( fp ) / 2;

	if( (*plst = (WB *)malloc( num * 2 )) == NULL ) {
		return -1;
	}
	fseek( fp, 2, SEEK_SET );	/* skip leading two bytes */
	rc = fread( (byte*)*plst, 2, num, fp );

	fclose( fp );
	return rc;
}

WB*	wlist;
int	wcnt;

int b52uni( char* istr, char* ostr, int ocnt )
{
	int idx;

	while( *istr ) {
		idx = b5_cod2ord( istr[0], istr[1] );
		if( idx < 0 ) {
			if( --ocnt < 0 )
				break;
			*ostr++ = *istr++;
			if( --ocnt < 0 )
				break;
			*ostr++ = 0x00;
		}
		else {
			if( idx > wcnt )
				idx = 0;
			if( --ocnt < 0 )
				break;
			// *ostr++ = ((byte *)&(wlist[idx]))[0];
			*ostr++ = wlist[idx].B.L;
			if( --ocnt < 0 )
				break;
			// *ostr++ = ((byte *)&(wlist[idx]))[1];
			*ostr++ = wlist[idx].B.H;
			istr += 2;

		}
	}
/*
	if( --ocnt < 0 )
		return;
	*ostr++ = 0x00;
	if( --ocnt < 0 )
		return;
	*ostr++ = 0x00;
*/
	return ocnt;
}

int b52uni_test( char* fname )
{
	FILE* fp;
	char ibuf[ BUFSIZ ], obuf[ BUFSIZ * 2];
	int rc;

	if( ( fp = fopen( fname, "r" ) ) == NULL ) {
		perror( fname );
		return __LINE__;
	}

	// unicode file leading code
	obuf[0] = 0xFF;
	obuf[1] = 0xFE;
	fwrite( obuf, 1, 2, stdout );

	while( fgets(ibuf, BUFSIZ, fp) != NULL ) {
		rc = b52uni( ibuf, obuf, sizeof(obuf) );
		fwrite( obuf, 1, sizeof(obuf)-rc, stdout );
	}
	fclose( fp );
	return 0;
}

int main( int argc, char *argv[] )
{
	return b5tbl_gen( argc, argv );

	if( argc < 2 ) {
		fprintf( stderr, "%s textfile\n", argv[0] );
		return EXIT_FAILURE;
	}
	wcnt = b52uni_load( "b52uni.tab", &wlist );
	fprintf( stderr, "b52uni_load %s = %d.\n", "b52uni.tab", wcnt );

	b52uni_test( argv[1] );

	return EXIT_SUCCESS;
}