vutf8.c

/* UTF8 related functions.
 *
 * Copyright (C) Navaneeth.K.N
 *
 * This is part of libvarnam. See LICENSE.txt for the license
 */

/* This UTF8 decoder is by JSON.org. I have modified this to add reentrancy support */

#include "vutf8.h"

/*
    Very Strict UTF-8 Decoder

    UTF-8 is a multibyte character encoding of Unicode. A character can be
    represented by 1-4 bytes. The bit pattern of the first byte indicates the
    number of continuation bytes.

    Most UTF-8 decoders tend to be lenient, attempting to recover as much
    information as possible, even from badly encoded input. This UTF-8
    decoder is not lenient. It will reject input which does not include
    proper continuation bytes. It will reject aliases (or suboptimal
    codings). It will reject surrogates. (Surrogate encoding should only be
    used with UTF-16.)

    Code     Contination Minimum Maximum
    0xxxxxxx           0       0     127
    10xxxxxx       error
    110xxxxx           1     128    2047
    1110xxxx           2    2048   65535 excluding 55296 - 57343
    11110xxx           3   65536 1114111
    11111xxx       error
*/


/*
    Get the next byte. It returns UTF8_END if there are no more bytes.
*/
static int
get(utf8_decoder *decoder)
{
    int c;
    if (decoder->the_index >= decoder->the_length) {
        return UTF8_END;
    }
    c = decoder->the_input[decoder->the_index] & 0xFF;
    decoder->the_index += 1;
    return c;
}


/*
    Get the 6-bit payload of the next continuation byte.
    Return UTF8_ERROR if it is not a contination byte.
*/
static int
cont(utf8_decoder *decoder)
{
    int c = get(decoder);
    return ((c & 0xC0) == 0x80) ? (c & 0x3F) : UTF8_ERROR;
}

/*
    Initialize the UTF-8 decoder with the specified values
*/
void
utf8_decode_init(char p[], int length, utf8_decoder *decoder)
{
    decoder->the_index = 0;
    decoder->the_input = p;
    decoder->the_length = length;
    decoder->the_char = 0;
    decoder->the_byte = 0;
}

/*
    Get the current byte offset. This is generally used in error reporting.
*/
int
utf8_decode_at_byte(utf8_decoder *decoder)
{
    return decoder->the_byte;
}


/*
    Get the current character offset. This is generally used in error reporting.
    The character offset matches the byte offset if the text is strictly ASCII.
*/
int
utf8_decode_at_character(utf8_decoder *decoder)
{
    return decoder->the_char > 0 ? decoder->the_char - 1 : 0;
}


/*
    Extract the next character.
    Returns: the character (between 0 and 1114111)
         or  UTF8_END   (the end)
         or  UTF8_ERROR (error)
*/
int
utf8_decode_next(utf8_decoder *decoder)
{
    int c;  /* the first byte of the character */
    int r;  /* the result */

    if (decoder->the_index >= decoder->the_length) {
        return decoder->the_index == decoder->the_length ? UTF8_END : UTF8_ERROR;
    }
    decoder->the_byte = decoder->the_index;
    decoder->the_char += 1;
    c = get(decoder);
/*
    Zero continuation (0 to 127)
*/
    if ((c & 0x80) == 0) {
        return c;
    }
/*
    One contination (128 to 2047)
*/
    if ((c & 0xE0) == 0xC0) {
        int c1 = cont(decoder);
        if (c1 < 0) {
            return UTF8_ERROR;
        }
        r = ((c & 0x1F) << 6) | c1;
        return r >= 128 ? r : UTF8_ERROR;
    }
/*
    Two continuation (2048 to 55295 and 57344 to 65535)
*/
    if ((c & 0xF0) == 0xE0) {
        int c1 = cont(decoder);
        int c2 = cont(decoder);
        if (c1 < 0 || c2 < 0) {
            return UTF8_ERROR;
        }
        r = ((c & 0x0F) << 12) | (c1 << 6) | c2;
        return r >= 2048 && (r < 55296 || r > 57343) ? r : UTF8_ERROR;
    }
/*
    Three continuation (65536 to 1114111)
*/
    if ((c & 0xF8) == 0xF0) {
        int c1 = cont(decoder);
        int c2 = cont(decoder);
        int c3 = cont(decoder);
        if (c1 < 0 || c2 < 0 || c3 < 0) {
            return UTF8_ERROR;
        }
        r = ((c & 0x0F) << 18) | (c1 << 12) | (c2 << 6) | c3;
        return r >= 65536 && r <= 1114111 ? r : UTF8_ERROR;
    }
    return UTF8_ERROR;
}