/**
 * Convert a file of UTF-8 to ISO-8859-1.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <unistd.h>

static int tolerance = 100;
static char * defaultValue;

int getUnicodePoint( FILE * infp ) {
    int c;
    int answer;
    int mask;
    int topmask;
    int bytes;

    c = fgetc( infp );
    answer = c;
    if ( EOF != c && 0x80 & c ) {
        bytes = 0;
        topmask = 0xff;
        for ( mask = 0x80; c & mask; mask >>= 1 ) {
            topmask >>= 1;
            bytes++;
        }
        bytes--;
        answer = ( c & topmask ) << ( bytes * 6 );
        while ( 0 < bytes ) {
            c = fgetc( infp );
            bytes--;
            answer += ( c & 0x3f ) << ( bytes * 6 );
        }
        assert( EOF != answer );
    }
    return answer;
}

void replace( int c ) {
    fprintf( stderr,
        "Unicode code point %d has no equivalent in ISO-8859-1.\n", c );
    if ( 0 == tolerance ) {
        fprintf( stderr, "Too many unmappable characters. Quitting.\n" );
        exit( 1 );
    }
    tolerance--;
    printf( defaultValue, c );
}

void putUnicodePoint( int c ) {
    if ( c < 0  || 255 < c ) {
        replace( c );
    }
    else {
        putchar( c );
    }
}

void doFiles( int argc, char **argv ) {

    FILE *infp;
    char *infile;
    int c;

    while ( --argc > 0 ) {
        infile = *++argv;

        if ( 0 == strcmp( infile, "-" ) ) {
            infp = stdin;
        }
        /* Must open input in binary mode so we can catch chars with high bit set. */
        else if ( ( infp = fopen( infile, "rb" ) ) == NULL ) {
            fprintf( stderr, "Unable to open input file \"%s\".\n", infile );
            continue;
        }
        while ( EOF != ( c = getUnicodePoint( infp ) ) ) {
            putUnicodePoint( c );
        }
    }

}

int main( int argc, char **argv ) {

    extern int optind;
    int i, j, c;
    char * oneName[2];

    defaultValue = "?";

    while ( -1 != ( c = getopt( argc, argv, "t:d:" ) ) ) {
        switch (c) {
        case 't':
            fprintf( stderr, "Tolerance given as %s.\n", optarg );
            tolerance = atoi(optarg);
            break;
        case 'd':
            defaultValue = optarg;
            break;
        default:
            fprintf( stderr, "usage: utf2iso [-t n] [-d s] [files]\n" );
            fprintf( stderr, " where n is how many errors to tolerate\n" );
            fprintf( stderr, " and s is a default string" );
            fprintf( stderr, " for unmappable chars.\n" );
            fprintf( stderr, " s can contain something like %%d" );
            fprintf( stderr, " or be empty.\n" );
            exit( 1 );
        }
    }
    i = 1;
    j = optind;
    while ( j < argc ) {
        argv[i++] = argv[j++];
    }
    argc -= ( j - i );

    if ( argc < 2 ) {
        oneName[0] = argv[0];
        oneName[1] = "-";
        doFiles( 2, oneName );
    }
    else {
        doFiles( argc, argv );
    }

    return 0;
}