source: mainline/common/str.c@ ae787807

/*
 * Copyright (c) 2001-2004 Jakub Jermar
 * Copyright (c) 2005 Martin Decky
 * Copyright (c) 2008 Jiri Svoboda
 * Copyright (c) 2011 Martin Sucha
 * Copyright (c) 2011 Oleg Romanenko
 * Copyright (c) 2025 Jiří Zárevúcky
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * - Redistributions of source code must retain the above copyright
 *   notice, this list of conditions and the following disclaimer.
 * - Redistributions in binary form must reproduce the above copyright
 *   notice, this list of conditions and the following disclaimer in the
 *   documentation and/or other materials provided with the distribution.
 * - The name of the author may not be used to endorse or promote products
 *   derived from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/** @addtogroup libc
 * @{
 */

/**
 * @file
 * @brief String functions.
 *
 * Strings and characters use the Universal Character Set (UCS). The standard
 * strings, called just strings are encoded in UTF-8. Wide strings (encoded
 * in UTF-32) are supported to a limited degree. A single character is
 * represented as char32_t.@n
 *
 * Overview of the terminology:@n
 *
 *  Term                  Meaning
 *  --------------------  ----------------------------------------------------
 *  byte                  8 bits stored in uint8_t (unsigned 8 bit integer)
 *
 *  character             UTF-32 encoded Unicode character, stored in char32_t
 *                        (unsigned 32 bit integer), code points 0 .. 1114111
 *                        are valid
 *
 *                        Note that Unicode characters do not match
 *                        one-to-one with displayed characters or glyphs on
 *                        screen. For that level of precision, look up
 *                        Grapheme Clusters.
 *
 *  ASCII character       7 bit encoded ASCII character, stored in char
 *                        (usually signed 8 bit integer), code points 0 .. 127
 *                        are valid
 *
 *  string                UTF-8 encoded NULL-terminated Unicode string, char *
 *
 *  wide string           UTF-32 encoded NULL-terminated Unicode string,
 *                        char32_t *
 *
 *  [wide] string size    number of BYTES in a [wide] string (excluding
 *                        the NULL-terminator), size_t
 *
 *  [wide] string length  number of CHARACTERS in a [wide] string (excluding
 *                        the NULL-terminator), size_t
 *
 *  [wide] string width   number of display cells on a monospace display taken
 *                        by a [wide] string, size_t
 *
 *                        This is virtually impossible to determine exactly for
 *                        all strings without knowing specifics of the display
 *                        device, due to various factors affecting text output.
 *                        If you have the option to query the terminal for
 *                        position change caused by outputting the string,
 *                        it is preferrable to determine width that way.
 *
 *
 * Overview of string metrics:@n
 *
 *  Metric  Abbrev.  Type     Meaning
 *  ------  ------   ------   -------------------------------------------------
 *  size    n        size_t   number of BYTES in a string (excluding the
 *                            NULL-terminator)
 *
 *  length  l        size_t   number of CHARACTERS in a string (excluding the
 *                            null terminator)
 *
 *  width  w         size_t   number of display cells on a monospace display
 *                            taken by a string
 *
 *
 * Function naming prefixes:@n
 *
 *  chr_    operate on characters
 *  ascii_  operate on ASCII characters
 *  str_    operate on strings
 *  wstr_   operate on wide strings
 *
 *  [w]str_[n|l|w]  operate on a prefix limited by size, length
 *                  or width
 *
 *
 * A specific character inside a [wide] string can be referred to by:@n
 *
 *  pointer (char *, char32_t *)
 *  byte offset (size_t)
 *  character index (size_t)
 *
 */

#include <str.h>

#include <align.h>
#include <assert.h>
#include <ctype.h>
#include <errno.h>
#include <limits.h>
#include <macros.h>
#include <mem.h>
#include <stdbool.h>
#include <stddef.h>
#include <stdint.h>
#include <stdlib.h>
#include <uchar.h>

#if __STDC_HOSTED__
#include <fibril.h>
#endif

static void _set_ilseq()
{
#ifdef errno
        errno = EILSEQ;
#endif
}

/** Byte mask consisting of lowest @n bits (out of 8) */
#define LO_MASK_8(n)  ((uint8_t) ((1 << (n)) - 1))

/** Byte mask consisting of lowest @n bits (out of 32) */
#define LO_MASK_32(n)  ((uint32_t) ((1 << (n)) - 1))

/** Byte mask consisting of highest @n bits (out of 8) */
#define HI_MASK_8(n)  (~LO_MASK_8(8 - (n)))

/** Number of data bits in a UTF-8 continuation byte */
#define CONT_BITS  6

#define UTF8_MASK_INITIAL2  0b00011111
#define UTF8_MASK_INITIAL3  0b00001111
#define UTF8_MASK_INITIAL4  0b00000111
#define UTF8_MASK_CONT      0b00111111

#define CHAR_INVALID ((char32_t) UINT_MAX)

static inline bool _is_ascii(uint8_t b)
{
        return b < 0x80;
}

static inline bool _is_continuation(uint8_t b)
{
        return (b & 0xC0) == 0x80;
}

static inline bool _is_2_byte(uint8_t c)
{
        return (c & 0xE0) == 0xC0;
}

static inline bool _is_3_byte(uint8_t c)
{
        return (c & 0xF0) == 0xE0;
}

static inline bool _is_4_byte(uint8_t c)
{
        return (c & 0xF8) == 0xF0;
}

static inline int _char_continuation_bytes(char32_t c)
{
        if ((c & ~LO_MASK_32(7)) == 0)
                return 0;

        if ((c & ~LO_MASK_32(11)) == 0)
                return 1;

        if ((c & ~LO_MASK_32(16)) == 0)
                return 2;

        if ((c & ~LO_MASK_32(21)) == 0)
                return 3;

        /* Codes longer than 21 bits are not supported */
        return -1;
}

static inline int _continuation_bytes(uint8_t b)
{
        /* 0xxxxxxx */
        if (_is_ascii(b))
                return 0;

        /* 110xxxxx 10xxxxxx */
        if (_is_2_byte(b))
                return 1;

        /* 1110xxxx 10xxxxxx 10xxxxxx */
        if (_is_3_byte(b))
                return 2;

        /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
        if (_is_4_byte(b))
                return 3;

        return -1;
}

static bool _is_non_shortest(const mbstate_t *mb, uint8_t b)
{
        return (mb->state == 0b1111110000000000 && !(b & 0b00100000)) ||
            (mb->state == 0b1111111111110000 && !(b & 0b00110000));
}

static bool _is_surrogate(const mbstate_t *mb, uint8_t b)
{
        return (mb->state == 0b1111110000001101 && b >= 0xa0);
}

#define _likely(expr) __builtin_expect((expr), true)
#define _unlikely(expr) __builtin_expect((expr), false)

#define FAST_PATHS 1

static char32_t _str_decode(const char *s, size_t *offset, size_t size, mbstate_t *mb)
{
        assert(s);
        assert(offset);
        assert(*offset <= size);
        assert(size == STR_NO_LIMIT || s + size >= s);
        assert(mb);

        if (*offset == size)
                return 0;

        if (_likely(!mb->state)) {
                /* Clean slate, read initial byte. */
                uint8_t b = s[(*offset)++];

                /* Fast exit for the most common case. */
                if (_likely(_is_ascii(b)))
                        return b;

                /* unexpected continuation byte */
                if (_unlikely(_is_continuation(b)))
                        return CHAR_INVALID;

                /*
                 * The value stored into `continuation` is designed to have
                 * just enough leading ones that after shifting in one less than
                 * the expected number of continuation bytes, the most significant
                 * bit becomes zero. (The field is 16b wide.)
                 */

                if (_is_2_byte(b)) {
                        /* Reject non-shortest form. */
                        if (_unlikely(!(b & 0b00011110)))
                                return CHAR_INVALID;

#if FAST_PATHS
                        /* We can usually take this exit. */
                        if (_likely(*offset < size && _is_continuation(s[*offset])))
                                return (b & UTF8_MASK_INITIAL2) << 6 |
                                    (s[(*offset)++] & UTF8_MASK_CONT);
#endif

                        /* 2 byte continuation    110xxxxx */
                        mb->state = b ^ 0b0000000011000000;

                } else if (_is_3_byte(b)) {
#if FAST_PATHS
                        /* We can usually take this exit. */
                        if (_likely(*offset + 1 < size && _is_continuation(s[*offset]) && _is_continuation(s[*offset + 1]))) {

                                char32_t ch = (b & UTF8_MASK_INITIAL3) << 12 |
                                    (s[(*offset)] & UTF8_MASK_CONT) << 6 |
                                    (s[(*offset) + 1] & UTF8_MASK_CONT);

                                *offset += 2;

                                /* Reject non-shortest form. */
                                if (_unlikely(!(ch & 0xFFFFF800)))
                                        return CHAR_INVALID;

                                /* Reject surrogates */
                                if (_unlikely(ch >= 0xD800 && ch < 0xE000))
                                        return CHAR_INVALID;

                                return ch;
                        }
#endif

                        /* 3 byte continuation    1110xxxx */
                        mb->state = b ^ 0b1111110011100000;

                } else if (_is_4_byte(b)) {
#if FAST_PATHS
                        /* We can usually take this exit. */
                        if (_likely(*offset + 2 < size && _is_continuation(s[*offset]) &&
                            _is_continuation(s[*offset + 1]) && _is_continuation(s[*offset + 2]))) {

                                char32_t ch = (b & UTF8_MASK_INITIAL4) << 18 |
                                    (s[(*offset)] & UTF8_MASK_CONT) << 12 |
                                    (s[(*offset) + 1] & UTF8_MASK_CONT) << 6 |
                                    (s[(*offset) + 2] & UTF8_MASK_CONT);

                                *offset += 3;

                                /* Reject non-shortest form. */
                                if (_unlikely(!(ch & 0xFFFF0000)))
                                        return CHAR_INVALID;

                                /* Reject out-of-range characters. */
                                if (_unlikely(ch >= 0x110000))
                                        return CHAR_INVALID;

                                return ch;
                        }
#endif

                        /* 4 byte continuation    11110xxx */
                        mb->state = b ^ 0b1111111100000000;
                } else {
                        return CHAR_INVALID;
                }
        }

        /* Deal with the remaining edge and invalid cases. */
        for (; *offset < size; (*offset)++) {
                /* Read continuation bytes. */
                uint8_t b = s[*offset];

                if (!_is_continuation(b) || _is_non_shortest(mb, b) || _is_surrogate(mb, b)) {
                        mb->state = 0;
                        return CHAR_INVALID;
                }

                /* Top bit becomes zero when shifting in the second to last byte. */
                if (!(mb->state & 0x8000)) {
                        char32_t c = ((char32_t) mb->state) << 6 | (b & UTF8_MASK_CONT);
                        mb->state = 0;
                        (*offset)++;
                        return c;
                }

                mb->state = mb->state << 6 | (b & UTF8_MASK_CONT);
        }

        /* Incomplete character. */
        assert(mb->state);
        return 0;
}

/** Standard <uchar.h> function since C11. */
size_t mbrtoc32(char32_t *c, const char *s, size_t n, mbstate_t *mb)
{
#if __STDC_HOSTED__
        static fibril_local mbstate_t global_state = { };

        if (!mb)
                mb = &global_state;
#endif

        if (!s) {
                /* Equivalent to mbrtoc32(NULL, "", 1, mb); */
                c = NULL;
                s = "";
                n = 1;
        }

        size_t offset = 0;
        char32_t ret = _str_decode(s, &offset, n, mb);
        if (ret == CHAR_INVALID) {
                assert(!mb->state);
                _set_ilseq();
                return UCHAR_ILSEQ;
        }
        if (mb->state) {
                assert(ret == 0);
                return UCHAR_INCOMPLETE;
        }

        if (c)
                *c = ret;
        return ret ? offset : 0;
}

/** Decode a single character from a string.
 *
 * Decode a single character from a string of size @a size. Decoding starts
 * at @a offset and this offset is moved to the beginning of the next
 * character. In case of decoding error, offset generally advances at least
 * by one. However, offset is never moved beyond size.
 *
 * @param str    String (not necessarily NULL-terminated).
 * @param offset Byte offset in string where to start decoding.
 * @param size   Size of the string (in bytes).
 *
 * @return Value of decoded character, U_SPECIAL on decoding error or
 *         NULL if attempt to decode beyond @a size.
 *
 */
char32_t str_decode(const char *str, size_t *offset, size_t size)
{
        mbstate_t mb = { };
        char32_t ch = _str_decode(str, offset, size, &mb);

        if (ch == CHAR_INVALID)
                return U_SPECIAL;

        if (mb.state)
                return U_SPECIAL;

        return ch;
}

/** Decode a single character from a string to the left.
 *
 * Decode a single character from a string of size @a size. Decoding starts
 * at @a offset and this offset is moved to the beginning of the previous
 * character. In case of decoding error, offset generally decreases at least
 * by one. However, offset is never moved before 0.
 *
 * @param str    String (not necessarily NULL-terminated).
 * @param offset Byte offset in string where to start decoding.
 * @param size   Size of the string (in bytes).
 *
 * @return Value of decoded character, U_SPECIAL on decoding error or
 *         NULL if attempt to decode beyond @a start of str.
 *
 */
char32_t str_decode_reverse(const char *str, size_t *offset, size_t size)
{
        if (*offset == 0)
                return 0;

        int cbytes = 0;
        /* Continue while continuation bytes found */
        while (*offset > 0 && cbytes < 4) {
                uint8_t b = (uint8_t) str[--(*offset)];

                if (_is_continuation(b)) {
                        cbytes++;
                        continue;
                }

                /* Reject non-shortest form encoding. */
                if (cbytes != _continuation_bytes(b))
                        return U_SPECIAL;

                /* Start byte */
                size_t start_offset = *offset;
                return str_decode(str, &start_offset, size);
        }

        /* Too many continuation bytes */
        return U_SPECIAL;
}

/** Encode a single character to string representation.
 *
 * Encode a single character to string representation (i.e. UTF-8) and store
 * it into a buffer at @a offset. Encoding starts at @a offset and this offset
 * is moved to the position where the next character can be written to.
 *
 * @param ch     Input character.
 * @param str    Output buffer.
 * @param offset Byte offset where to start writing.
 * @param size   Size of the output buffer (in bytes).
 *
 * @return EOK if the character was encoded successfully, EOVERFLOW if there
 *         was not enough space in the output buffer or EINVAL if the character
 *         code was invalid.
 */
errno_t chr_encode(char32_t ch, char *str, size_t *offset, size_t size)
{
        // TODO: merge with c32rtomb()

        if (*offset >= size)
                return EOVERFLOW;

        /* Fast exit for the most common case. */
        if (ch < 0x80) {
                str[(*offset)++] = (char) ch;
                return EOK;
        }

        /* Codes longer than 21 bits are not supported */
        if (!chr_check(ch))
                return EINVAL;

        /* Determine how many continuation bytes are needed */

        unsigned int cbytes = _char_continuation_bytes(ch);
        unsigned int b0_bits = 6 - cbytes;  /* Data bits in first byte */

        /* Check for available space in buffer */
        if (*offset + cbytes >= size)
                return EOVERFLOW;

        /* Encode continuation bytes */
        unsigned int i;
        for (i = cbytes; i > 0; i--) {
                str[*offset + i] = 0x80 | (ch & LO_MASK_32(CONT_BITS));
                ch >>= CONT_BITS;
        }

        /* Encode first byte */
        str[*offset] = (ch & LO_MASK_32(b0_bits)) | HI_MASK_8(8 - b0_bits - 1);

        /* Advance offset */
        *offset += cbytes + 1;

        return EOK;
}

/* Convert in place any bytes that don't form a valid character into replacement. */
static size_t _str_sanitize(char *str, size_t n, uint8_t replacement)
{
        uint8_t *b = (uint8_t *) str;
        size_t count = 0;

        for (; n > 0 && b[0]; b++, n--) {
                int cont = _continuation_bytes(b[0]);
                if (__builtin_expect(cont, 0) == 0)
                        continue;

                if (cont < 0 || n <= (size_t) cont) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                /* Check continuation bytes. */
                bool valid = true;
                for (int i = 1; i <= cont; i++) {
                        if (!_is_continuation(b[i])) {
                                valid = false;
                                break;
                        }
                }

                if (!valid) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                /*
                 * Check for non-shortest form encoding.
                 * See https://www.unicode.org/versions/corrigendum1.html
                 */

                /* 0b110!!!!x 0b10xxxxxx */
                if (cont == 1 && !(b[0] & 0b00011110)) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                /* 0b1110!!!! 0b10!xxxxx 0b10xxxxxx */
                if (cont == 2 && !(b[0] & 0b00001111) && !(b[1] & 0b00100000)) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                /* 0b11110!!! 0b10!!xxxx 0b10xxxxxx 0b10xxxxxx */
                if (cont == 3 && !(b[0] & 0b00000111) && !(b[1] & 0b00110000)) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                /* Check for surrogate character encoding. */
                if (cont == 2 && b[0] == 0xED && b[1] >= 0xA0) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                /* Check for out-of-range code points. */
                if (cont == 3 && (b[0] > 0xF4 || (b[0] == 0xF4 && b[1] >= 0x90))) {
                        b[0] = replacement;
                        count++;
                        continue;
                }

                b += cont;
                n -= cont;
        }

        return count;
}

size_t str_sanitize(char *str, size_t n, uint8_t replacement)
{
        return _str_sanitize(str, n, replacement);
}

static size_t _str_size(const char *str)
{
        size_t size = 0;

        while (*str++ != 0)
                size++;

        return size;
}

/** Get size of string.
 *
 * Get the number of bytes which are used by the string @a str (excluding the
 * NULL-terminator).
 *
 * @param str String to consider.
 *
 * @return Number of bytes used by the string
 *
 */
size_t str_size(const char *str)
{
        return _str_size(str);
}

/** Get size of wide string.
 *
 * Get the number of bytes which are used by the wide string @a str (excluding the
 * NULL-terminator).
 *
 * @param str Wide string to consider.
 *
 * @return Number of bytes used by the wide string
 *
 */
size_t wstr_size(const char32_t *str)
{
        return (wstr_length(str) * sizeof(char32_t));
}

/** Get size of string with length limit.
 *
 * Get the number of bytes which are used by up to @a max_len first
 * characters in the string @a str. If @a max_len is greater than
 * the length of @a str, the entire string is measured (excluding the
 * NULL-terminator).
 *
 * @param str     String to consider.
 * @param max_len Maximum number of characters to measure.
 *
 * @return Number of bytes used by the characters.
 *
 */
size_t str_lsize(const char *str, size_t max_len)
{
        size_t len = 0;
        size_t offset = 0;

        while (len < max_len) {
                if (str_decode(str, &offset, STR_NO_LIMIT) == 0)
                        break;

                len++;
        }

        return offset;
}

static size_t _str_nsize(const char *str, size_t max_size)
{
        size_t size = 0;

        while ((*str++ != 0) && (size < max_size))
                size++;

        return size;
}

/** Get size of string with size limit.
 *
 * Get the number of bytes which are used by the string @a str
 * (excluding the NULL-terminator), but no more than @max_size bytes.
 *
 * @param str      String to consider.
 * @param max_size Maximum number of bytes to measure.
 *
 * @return Number of bytes used by the string
 *
 */
size_t str_nsize(const char *str, size_t max_size)
{
        return _str_nsize(str, max_size);
}

/** Get size of wide string with size limit.
 *
 * Get the number of bytes which are used by the wide string @a str
 * (excluding the NULL-terminator), but no more than @max_size bytes.
 *
 * @param str      Wide string to consider.
 * @param max_size Maximum number of bytes to measure.
 *
 * @return Number of bytes used by the wide string
 *
 */
size_t wstr_nsize(const char32_t *str, size_t max_size)
{
        return (wstr_nlength(str, max_size) * sizeof(char32_t));
}

/** Get size of wide string with length limit.
 *
 * Get the number of bytes which are used by up to @a max_len first
 * wide characters in the wide string @a str. If @a max_len is greater than
 * the length of @a str, the entire wide string is measured (excluding the
 * NULL-terminator).
 *
 * @param str     Wide string to consider.
 * @param max_len Maximum number of wide characters to measure.
 *
 * @return Number of bytes used by the wide characters.
 *
 */
size_t wstr_lsize(const char32_t *str, size_t max_len)
{
        return (wstr_nlength(str, max_len * sizeof(char32_t)) * sizeof(char32_t));
}

/** Get number of characters in a string.
 *
 * @param str NULL-terminated string.
 *
 * @return Number of characters in string.
 *
 */
size_t str_length(const char *str)
{
        size_t len = 0;
        size_t offset = 0;

        while (str_decode(str, &offset, STR_NO_LIMIT) != 0)
                len++;

        return len;
}

/** Get number of characters in a wide string.
 *
 * @param str NULL-terminated wide string.
 *
 * @return Number of characters in @a str.
 *
 */
size_t wstr_length(const char32_t *wstr)
{
        size_t len = 0;

        while (*wstr++ != 0)
                len++;

        return len;
}

/** Get number of characters in a string with size limit.
 *
 * @param str  NULL-terminated string.
 * @param size Maximum number of bytes to consider.
 *
 * @return Number of characters in string.
 *
 */
size_t str_nlength(const char *str, size_t size)
{
        size_t len = 0;
        size_t offset = 0;

        while (str_decode(str, &offset, size) != 0)
                len++;

        return len;
}

/** Get number of characters in a string with size limit.
 *
 * @param str  NULL-terminated string.
 * @param size Maximum number of bytes to consider.
 *
 * @return Number of characters in string.
 *
 */
size_t wstr_nlength(const char32_t *str, size_t size)
{
        size_t len = 0;
        size_t limit = ALIGN_DOWN(size, sizeof(char32_t));
        size_t offset = 0;

        while ((offset < limit) && (*str++ != 0)) {
                len++;
                offset += sizeof(char32_t);
        }

        return len;
}

/** Get character display width on a character cell display.
 *
 * @param ch    Character
 * @return      Width of character in cells.
 */
size_t chr_width(char32_t ch)
{
        return 1;
}

/** Get string display width on a character cell display.
 *
 * @param str   String
 * @return      Width of string in cells.
 */
size_t str_width(const char *str)
{
        size_t width = 0;
        size_t offset = 0;
        char32_t ch;

        while ((ch = str_decode(str, &offset, STR_NO_LIMIT)) != 0)
                width += chr_width(ch);

        return width;
}

/** Check whether character is plain ASCII.
 *
 * @return True if character is plain ASCII.
 *
 */
bool ascii_check(char32_t ch)
{
        if (ch <= 127)
                return true;

        return false;
}

/** Check whether character is valid
 *
 * @return True if character is a valid Unicode code point.
 *
 */
bool chr_check(char32_t ch)
{
        if (ch <= 1114111)
                return true;

        return false;
}

/** Compare two NULL terminated strings.
 *
 * Do a char-by-char comparison of two NULL-terminated strings.
 * The strings are considered equal iff their length is equal
 * and both strings consist of the same sequence of characters.
 *
 * A string S1 is less than another string S2 if it has a character with
 * lower value at the first character position where the strings differ.
 * If the strings differ in length, the shorter one is treated as if
 * padded by characters with a value of zero.
 *
 * @param s1 First string to compare.
 * @param s2 Second string to compare.
 *
 * @return 0 if the strings are equal, -1 if the first is less than the second,
 *         1 if the second is less than the first.
 *
 */
int str_cmp(const char *s1, const char *s2)
{
        /*
         * UTF-8 has the nice property that lexicographic ordering on bytes is
         * the same as the lexicographic ordering of the character sequences.
         */
        while (*s1 == *s2 && *s1 != 0) {
                s1++;
                s2++;
        }

        if (*s1 == *s2)
                return 0;

        return (*s1 < *s2) ? -1 : 1;
}

/** Compare two NULL terminated strings with length limit.
 *
 * Do a char-by-char comparison of two NULL-terminated strings.
 * The strings are considered equal iff
 * min(str_length(s1), max_len) == min(str_length(s2), max_len)
 * and both strings consist of the same sequence of characters,
 * up to max_len characters.
 *
 * A string S1 is less than another string S2 if it has a character with
 * lower value at the first character position where the strings differ.
 * If the strings differ in length, the shorter one is treated as if
 * padded by characters with a value of zero. Only the first max_len
 * characters are considered.
 *
 * @param s1      First string to compare.
 * @param s2      Second string to compare.
 * @param max_len Maximum number of characters to consider.
 *
 * @return 0 if the strings are equal, -1 if the first is less than the second,
 *         1 if the second is less than the first.
 *
 */
int str_lcmp(const char *s1, const char *s2, size_t max_len)
{
        char32_t c1 = 0;
        char32_t c2 = 0;

        size_t off1 = 0;
        size_t off2 = 0;

        size_t len = 0;

        while (true) {
                if (len >= max_len)
                        break;

                c1 = str_decode(s1, &off1, STR_NO_LIMIT);
                c2 = str_decode(s2, &off2, STR_NO_LIMIT);

                if (c1 < c2)
                        return -1;

                if (c1 > c2)
                        return 1;

                if (c1 == 0 || c2 == 0)
                        break;

                ++len;
        }

        return 0;

}

/** Compare two NULL terminated strings in case-insensitive manner.
 *
 * Do a char-by-char comparison of two NULL-terminated strings.
 * The strings are considered equal iff their length is equal
 * and both strings consist of the same sequence of characters
 * when converted to lower case.
 *
 * A string S1 is less than another string S2 if it has a character with
 * lower value at the first character position where the strings differ.
 * If the strings differ in length, the shorter one is treated as if
 * padded by characters with a value of zero.
 *
 * @param s1 First string to compare.
 * @param s2 Second string to compare.
 *
 * @return 0 if the strings are equal, -1 if the first is less than the second,
 *         1 if the second is less than the first.
 *
 */
int str_casecmp(const char *s1, const char *s2)
{
        // FIXME: doesn't work for non-ASCII caseful characters

        char32_t c1 = 0;
        char32_t c2 = 0;

        size_t off1 = 0;
        size_t off2 = 0;

        while (true) {
                c1 = tolower(str_decode(s1, &off1, STR_NO_LIMIT));
                c2 = tolower(str_decode(s2, &off2, STR_NO_LIMIT));

                if (c1 < c2)
                        return -1;

                if (c1 > c2)
                        return 1;

                if (c1 == 0 || c2 == 0)
                        break;
        }

        return 0;
}

/** Compare two NULL terminated strings with length limit in case-insensitive
 * manner.
 *
 * Do a char-by-char comparison of two NULL-terminated strings.
 * The strings are considered equal iff
 * min(str_length(s1), max_len) == min(str_length(s2), max_len)
 * and both strings consist of the same sequence of characters,
 * up to max_len characters.
 *
 * A string S1 is less than another string S2 if it has a character with
 * lower value at the first character position where the strings differ.
 * If the strings differ in length, the shorter one is treated as if
 * padded by characters with a value of zero. Only the first max_len
 * characters are considered.
 *
 * @param s1      First string to compare.
 * @param s2      Second string to compare.
 * @param max_len Maximum number of characters to consider.
 *
 * @return 0 if the strings are equal, -1 if the first is less than the second,
 *         1 if the second is less than the first.
 *
 */
int str_lcasecmp(const char *s1, const char *s2, size_t max_len)
{
        // FIXME: doesn't work for non-ASCII caseful characters

        char32_t c1 = 0;
        char32_t c2 = 0;

        size_t off1 = 0;
        size_t off2 = 0;

        size_t len = 0;

        while (true) {
                if (len >= max_len)
                        break;

                c1 = tolower(str_decode(s1, &off1, STR_NO_LIMIT));
                c2 = tolower(str_decode(s2, &off2, STR_NO_LIMIT));

                if (c1 < c2)
                        return -1;

                if (c1 > c2)
                        return 1;

                if (c1 == 0 || c2 == 0)
                        break;

                ++len;
        }

        return 0;

}

static bool _test_prefix(const char *s, const char *p)
{
        while (*s == *p && *s != 0) {
                s++;
                p++;
        }

        return *p == 0;
}

/** Test whether p is a prefix of s.
 *
 * Do a char-by-char comparison of two NULL-terminated strings
 * and determine if p is a prefix of s.
 *
 * @param s The string in which to look
 * @param p The string to check if it is a prefix of s
 *
 * @return true iff p is prefix of s else false
 *
 */
bool str_test_prefix(const char *s, const char *p)
{
        return _test_prefix(s, p);
}

/** Get a string suffix.
 *
 * Return a string suffix defined by the prefix length.
 *
 * @param s             The string to get the suffix from.
 * @param prefix_length Number of prefix characters to ignore.
 *
 * @return String suffix.
 *
 */
const char *str_suffix(const char *s, size_t prefix_length)
{
        size_t off = 0;
        size_t i = 0;

        while (true) {
                str_decode(s, &off, STR_NO_LIMIT);
                i++;

                if (i >= prefix_length)
                        break;
        }

        return s + off;
}

/** Copy string as a sequence of bytes. */
static void _str_cpy(char *dest, const char *src)
{
        while (*src)
                *(dest++) = *(src++);

        *dest = 0;
}

/** Copy string as a sequence of bytes. */
static void _str_cpyn(char *dest, size_t size, const char *src)
{
        assert(dest && src && size);

        if (!dest || !src || !size)
                return;

        if (size == STR_NO_LIMIT)
                return _str_cpy(dest, src);

        char *dest_top = dest + size - 1;
        assert(size == 1 || dest < dest_top);

        while (*src && dest < dest_top)
                *(dest++) = *(src++);

        *dest = 0;
}

/** Copy string.
 *
 * Copy source string @a src to destination buffer @a dest.
 * No more than @a size bytes are written. If the size of the output buffer
 * is at least one byte, the output string will always be well-formed, i.e.
 * null-terminated and containing only complete characters.
 *
 * @param dest  Destination buffer.
 * @param count Size of the destination buffer (must be > 0).
 * @param src   Source string.
 *
 */
void str_cpy(char *dest, size_t size, const char *src)
{
        /* There must be space for a null terminator in the buffer. */
        assert(size > 0);
        assert(src != NULL);
        assert(dest != NULL);
        assert(size == STR_NO_LIMIT || dest + size > dest);

        /* Copy data. */
        _str_cpyn(dest, size, src);

        /* In-place translate invalid bytes to U_SPECIAL. */
        _str_sanitize(dest, size, U_SPECIAL);
}

/** Copy size-limited substring.
 *
 * Copy prefix of string @a src of max. size @a size to destination buffer
 * @a dest. No more than @a size bytes are written. The output string will
 * always be well-formed, i.e. null-terminated and containing only complete
 * characters.
 *
 * No more than @a n bytes are read from the input string, so it does not
 * have to be null-terminated.
 *
 * @param dest  Destination buffer.
 * @param count Size of the destination buffer (must be > 0).
 * @param src   Source string.
 * @param n     Maximum number of bytes to read from @a src.
 *
 */
void str_ncpy(char *dest, size_t size, const char *src, size_t n)
{
        /* There must be space for a null terminator in the buffer. */
        assert(size > 0);
        assert(src != NULL);

        /* Copy data. */
        _str_cpyn(dest, min(size, n + 1), src);

        /* In-place translate invalid bytes to U_SPECIAL. */
        _str_sanitize(dest, size, U_SPECIAL);
}

/** Append one string to another.
 *
 * Append source string @a src to string in destination buffer @a dest.
 * Size of the destination buffer is @a dest. If the size of the output buffer
 * is at least one byte, the output string will always be well-formed, i.e.
 * null-terminated and containing only complete characters.
 *
 * @param dest   Destination buffer.
 * @param count Size of the destination buffer.
 * @param src   Source string.
 */
void str_append(char *dest, size_t size, const char *src)
{
        assert(src != NULL);
        assert(dest != NULL);
        assert(size > 0);
        assert(size == STR_NO_LIMIT || dest + size > dest);

        size_t dstr_size = _str_nsize(dest, size);
        if (dstr_size < size) {
                _str_cpyn(dest + dstr_size, size - dstr_size, src);
                _str_sanitize(dest + dstr_size, size - dstr_size, U_SPECIAL);
        }
}

/** Convert space-padded ASCII to string.
 *
 * Common legacy text encoding in hardware is 7-bit ASCII fitted into
 * a fixed-width byte buffer (bit 7 always zero), right-padded with spaces
 * (ASCII 0x20). Convert space-padded ascii to string representation.
 *
 * If the text does not fit into the destination buffer, the function converts
 * as many characters as possible and returns EOVERFLOW.
 *
 * If the text contains non-ASCII bytes (with bit 7 set), the whole string is
 * converted anyway and invalid characters are replaced with question marks
 * (U_SPECIAL) and the function returns EIO.
 *
 * Regardless of return value upon return @a dest will always be well-formed.
 *
 * @param dest          Destination buffer
 * @param size          Size of destination buffer
 * @param src           Space-padded ASCII.
 * @param n             Size of the source buffer in bytes.
 *
 * @return              EOK on success, EOVERFLOW if the text does not fit
 *                      destination buffer, EIO if the text contains
 *                      non-ASCII bytes.
 */
errno_t spascii_to_str(char *dest, size_t size, const uint8_t *src, size_t n)
{
        size_t len = 0;

        /* Determine the length of the source string. */
        for (size_t i = 0; i < n; i++) {
                if (src[i] == 0)
                        break;

                if (src[i] != ' ')
                        len = i + 1;
        }

        errno_t result = EOK;
        size_t out_len = min(len, size - 1);

        /* Copy characters */
        for (size_t i = 0; i < out_len; i++) {
                dest[i] = src[i];

                if (dest[i] < 0) {
                        dest[i] = U_SPECIAL;
                        result = EIO;
                }
        }

        dest[out_len] = 0;

        if (out_len < len)
                return EOVERFLOW;

        return result;
}

/** Convert wide string to string.
 *
 * Convert wide string @a src to string. The output is written to the buffer
 * specified by @a dest and @a size. @a size must be non-zero and the string
 * written will always be well-formed.
 *
 * @param dest  Destination buffer.
 * @param size  Size of the destination buffer.
 * @param src   Source wide string.
 */
void wstr_to_str(char *dest, size_t size, const char32_t *src)
{
        char32_t ch;
        size_t src_idx;
        size_t dest_off;

        /* There must be space for a null terminator in the buffer. */
        assert(size > 0);

        src_idx = 0;
        dest_off = 0;

        while ((ch = src[src_idx++]) != 0) {
                if (chr_encode(ch, dest, &dest_off, size - 1) != EOK)
                        break;
        }

        dest[dest_off] = '\0';
}

/** Convert UTF16 string to string.
 *
 * Convert utf16 string @a src to string. The output is written to the buffer
 * specified by @a dest and @a size. @a size must be non-zero and the string
 * written will always be well-formed. Surrogate pairs also supported.
 *
 * @param dest  Destination buffer.
 * @param size  Size of the destination buffer.
 * @param src   Source utf16 string.
 *
 * @return EOK, if success, an error code otherwise.
 */
errno_t utf16_to_str(char *dest, size_t size, const uint16_t *src)
{
        size_t idx = 0, dest_off = 0;
        char32_t ch;
        errno_t rc = EOK;

        /* There must be space for a null terminator in the buffer. */
        assert(size > 0);

        while (src[idx]) {
                if ((src[idx] & 0xfc00) == 0xd800) {
                        if (src[idx + 1] && (src[idx + 1] & 0xfc00) == 0xdc00) {
                                ch = 0x10000;
                                ch += (src[idx] & 0x03FF) << 10;
                                ch += (src[idx + 1] & 0x03FF);
                                idx += 2;
                        } else
                                break;
                } else {
                        ch = src[idx];
                        idx++;
                }
                rc = chr_encode(ch, dest, &dest_off, size - 1);
                if (rc != EOK)
                        break;
        }
        dest[dest_off] = '\0';
        return rc;
}

/** Convert string to UTF16 string.
 *
 * Convert string @a src to utf16 string. The output is written to the buffer
 * specified by @a dest and @a dlen. @a dlen must be non-zero and the string
 * written will always be well-formed. Surrogate pairs also supported.
 *
 * @param dest  Destination buffer.
 * @param dlen  Number of utf16 characters that fit in the destination buffer.
 * @param src   Source string.
 *
 * @return EOK, if success, an error code otherwise.
 */
errno_t str_to_utf16(uint16_t *dest, size_t dlen, const char *src)
{
        errno_t rc = EOK;
        size_t offset = 0;
        size_t idx = 0;
        char32_t c;

        assert(dlen > 0);

        while ((c = str_decode(src, &offset, STR_NO_LIMIT)) != 0) {
                if (c > 0x10000) {
                        if (idx + 2 >= dlen - 1) {
                                rc = EOVERFLOW;
                                break;
                        }
                        c = (c - 0x10000);
                        dest[idx] = 0xD800 | (c >> 10);
                        dest[idx + 1] = 0xDC00 | (c & 0x3FF);
                        idx++;
                } else {
                        dest[idx] = c;
                }

                idx++;
                if (idx >= dlen - 1) {
                        rc = EOVERFLOW;
                        break;
                }
        }

        dest[idx] = '\0';
        return rc;
}

/** Get size of UTF-16 string.
 *
 * Get the number of words which are used by the UTF-16 string @a ustr
 * (excluding the NULL-terminator).
 *
 * @param ustr UTF-16 string to consider.
 *
 * @return Number of words used by the UTF-16 string
 *
 */
size_t utf16_wsize(const uint16_t *ustr)
{
        size_t wsize = 0;

        while (*ustr++ != 0)
                wsize++;

        return wsize;
}

/** Convert wide string to new string.
 *
 * Convert wide string @a src to string. Space for the new string is allocated
 * on the heap.
 *
 * @param src   Source wide string.
 * @return      New string.
 */
char *wstr_to_astr(const char32_t *src)
{
        char dbuf[STR_BOUNDS(1)];
        char *str;
        char32_t ch;

        size_t src_idx;
        size_t dest_off;
        size_t dest_size;

        /* Compute size of encoded string. */

        src_idx = 0;
        dest_size = 0;

        while ((ch = src[src_idx++]) != 0) {
                dest_off = 0;
                if (chr_encode(ch, dbuf, &dest_off, STR_BOUNDS(1)) != EOK)
                        break;
                dest_size += dest_off;
        }

        str = malloc(dest_size + 1);
        if (str == NULL)
                return NULL;

        /* Encode string. */

        src_idx = 0;
        dest_off = 0;

        while ((ch = src[src_idx++]) != 0) {
                if (chr_encode(ch, str, &dest_off, dest_size) != EOK)
                        break;
        }

        str[dest_size] = '\0';
        return str;
}

/** Convert string to wide string.
 *
 * Convert string @a src to wide string. The output is written to the
 * buffer specified by @a dest and @a dlen. @a dlen must be non-zero
 * and the wide string written will always be null-terminated.
 *
 * @param dest  Destination buffer.
 * @param dlen  Length of destination buffer (number of wchars).
 * @param src   Source string.
 */
void str_to_wstr(char32_t *dest, size_t dlen, const char *src)
{
        size_t offset;
        size_t di;
        char32_t c;

        assert(dlen > 0);

        offset = 0;
        di = 0;

        do {
                if (di >= dlen - 1)
                        break;

                c = str_decode(src, &offset, STR_NO_LIMIT);
                dest[di++] = c;
        } while (c != '\0');

        dest[dlen - 1] = '\0';
}

/** Convert string to wide string.
 *
 * Convert string @a src to wide string. A new wide NULL-terminated
 * string will be allocated on the heap.
 *
 * @param src   Source string.
 */
char32_t *str_to_awstr(const char *str)
{
        size_t len = str_length(str);

        char32_t *wstr = calloc(len + 1, sizeof(char32_t));
        if (wstr == NULL)
                return NULL;

        str_to_wstr(wstr, len + 1, str);
        return wstr;
}

static char *_strchr(const char *str, char c)
{
        while (*str != 0 && *str != c)
                str++;

        return (*str == c) ? (char *) str : NULL;
}

/** Find first occurence of character in string.
 *
 * @param str String to search.
 * @param ch  Character to look for.
 *
 * @return Pointer to character in @a str or NULL if not found.
 */
char *str_chr(const char *str, char32_t ch)
{
        /* Fast path for an ASCII character. */
        if (ascii_check(ch))
                return _strchr(str, ch);

        /* Convert character to UTF-8. */
        char utf8[STR_BOUNDS(1) + 1];
        size_t offset = 0;

        if (chr_encode(ch, utf8, &offset, sizeof(utf8)) != EOK || offset == 0)
                return NULL;

        utf8[offset] = '\0';

        /* Find the first byte, then check if all of them are correct. */
        while (*str != 0) {
                str = _strchr(str, utf8[0]);
                if (!str)
                        return NULL;

                if (_test_prefix(str, utf8))
                        return (char *) str;

                str++;
        }

        return NULL;
}

/** Find first occurence of substring in string.
 *
 * @param hs  Haystack (string)
 * @param n   Needle (substring to look for)
 *
 * @return Pointer to character in @a hs or @c NULL if not found.
 */
char *str_str(const char *hs, const char *n)
{
        size_t hsize = _str_size(hs);
        size_t nsize = _str_size(n);

        while (hsize >= nsize) {
                if (_test_prefix(hs, n))
                        return (char *) hs;

                hs++;
                hsize--;
        }

        return NULL;
}

static void _str_rtrim(char *str, char c)
{
        char *last = str;

        while (*str) {
                if (*str != c)
                        last = str;

                str++;
        }

        /* Truncate string. */
        last[1] = 0;
}

/** Removes specified trailing characters from a string.
 *
 * @param str String to remove from.
 * @param ch  Character to remove.
 */
void str_rtrim(char *str, char32_t ch)
{
        /* Fast path for the ASCII case. */
        if (ascii_check(ch)) {
                _str_rtrim(str, ch);
                return;
        }

        size_t off = 0;
        size_t pos = 0;
        char32_t c;
        bool update_last_chunk = true;
        char *last_chunk = NULL;

        while ((c = str_decode(str, &off, STR_NO_LIMIT))) {
                if (c != ch) {
                        update_last_chunk = true;
                        last_chunk = NULL;
                } else if (update_last_chunk) {
                        update_last_chunk = false;
                        last_chunk = (str + pos);
                }
                pos = off;
        }

        if (last_chunk)
                *last_chunk = '\0';
}

static void _str_ltrim(char *str, char c)
{
        char *p = str;

        while (*p == c)
                p++;

        if (str != p)
                _str_cpy(str, p);
}

/** Removes specified leading characters from a string.
 *
 * @param str String to remove from.
 * @param ch  Character to remove.
 */
void str_ltrim(char *str, char32_t ch)
{
        /* Fast path for the ASCII case. */
        if (ascii_check(ch)) {
                _str_ltrim(str, ch);
                return;
        }

        char32_t acc;
        size_t off = 0;
        size_t pos = 0;
        size_t str_sz = str_size(str);

        while ((acc = str_decode(str, &off, STR_NO_LIMIT)) != 0) {
                if (acc != ch)
                        break;
                else
                        pos = off;
        }

        if (pos > 0) {
                memmove(str, &str[pos], str_sz - pos);
                pos = str_sz - pos;
                str[pos] = '\0';
        }
}

static char *_str_rchr(const char *str, char c)
{
        const char *last = NULL;

        while (*str) {
                if (*str == c)
                        last = str;

                str++;
        }

        return (char *) last;
}

/** Find last occurence of character in string.
 *
 * @param str String to search.
 * @param ch  Character to look for.
 *
 * @return Pointer to character in @a str or NULL if not found.
 */
char *str_rchr(const char *str, char32_t ch)
{
        if (ascii_check(ch))
                return _str_rchr(str, ch);

        char32_t acc;
        size_t off = 0;
        size_t last = 0;
        const char *res = NULL;

        while ((acc = str_decode(str, &off, STR_NO_LIMIT)) != 0) {
                if (acc == ch)
                        res = (str + last);
                last = off;
        }

        return (char *) res;
}

/** Insert a wide character into a wide string.
 *
 * Insert a wide character into a wide string at position
 * @a pos. The characters after the position are shifted.
 *
 * @param str     String to insert to.
 * @param ch      Character to insert to.
 * @param pos     Character index where to insert.
 * @param max_pos Characters in the buffer.
 *
 * @return True if the insertion was sucessful, false if the position
 *         is out of bounds.
 *
 */
bool wstr_linsert(char32_t *str, char32_t ch, size_t pos, size_t max_pos)
{
        size_t len = wstr_length(str);

        if ((pos > len) || (pos + 1 > max_pos))
                return false;

        size_t i;
        for (i = len; i + 1 > pos; i--)
                str[i + 1] = str[i];

        str[pos] = ch;

        return true;
}

/** Remove a wide character from a wide string.
 *
 * Remove a wide character from a wide string at position
 * @a pos. The characters after the position are shifted.
 *
 * @param str String to remove from.
 * @param pos Character index to remove.
 *
 * @return True if the removal was sucessful, false if the position
 *         is out of bounds.
 *
 */
bool wstr_remove(char32_t *str, size_t pos)
{
        size_t len = wstr_length(str);

        if (pos >= len)
                return false;

        size_t i;
        for (i = pos + 1; i <= len; i++)
                str[i - 1] = str[i];

        return true;
}

/** Duplicate string.
 *
 * Allocate a new string and copy characters from the source
 * string into it. The duplicate string is allocated via sleeping
 * malloc(), thus this function can sleep in no memory conditions.
 *
 * The allocation cannot fail and the return value is always
 * a valid pointer. The duplicate string is always a well-formed
 * null-terminated UTF-8 string, but it can differ from the source
 * string on the byte level.
 *
 * @param src Source string.
 *
 * @return Duplicate string.
 *
 */
char *str_dup(const char *src)
{
        size_t size = _str_size(src) + 1;
        char *dest = malloc(size);
        if (!dest)
                return NULL;

        memcpy(dest, src, size);
        _str_sanitize(dest, size, U_SPECIAL);
        return dest;
}

/** Duplicate string with size limit.
 *
 * Allocate a new string and copy up to @max_size bytes from the source
 * string into it. The duplicate string is allocated via sleeping
 * malloc(), thus this function can sleep in no memory conditions.
 * No more than @max_size + 1 bytes is allocated, but if the size
 * occupied by the source string is smaller than @max_size + 1,
 * less is allocated.
 *
 * The allocation cannot fail and the return value is always
 * a valid pointer. The duplicate string is always a well-formed
 * null-terminated UTF-8 string, but it can differ from the source
 * string on the byte level.
 *
 * @param src Source string.
 * @param n   Maximum number of bytes to duplicate.
 *
 * @return Duplicate string.
 *
 */
char *str_ndup(const char *src, size_t n)
{
        size_t size = _str_nsize(src, n);

        char *dest = malloc(size + 1);
        if (!dest)
                return NULL;

        memcpy(dest, src, size);
        _str_sanitize(dest, size, U_SPECIAL);
        dest[size] = 0;
        return dest;
}

/** Split string by delimiters.
 *
 * @param s             String to be tokenized. May not be NULL.
 * @param delim         String with the delimiters.
 * @param next          Variable which will receive the pointer to the
 *                      continuation of the string following the first
 *                      occurrence of any of the delimiter characters.
 *                      May be NULL.
 * @return              Pointer to the prefix of @a s before the first
 *                      delimiter character. NULL if no such prefix
 *                      exists.
 */
char *str_tok(char *s, const char *delim, char **next)
{
        char *start, *end;

        if (!s)
                return NULL;

        size_t len = str_size(s);
        size_t cur;
        size_t tmp;
        char32_t ch;

        /* Skip over leading delimiters. */
        tmp = 0;
        cur = 0;
        while ((ch = str_decode(s, &tmp, len)) && str_chr(delim, ch))
                cur = tmp;
        start = &s[cur];

        /* Skip over token characters. */
        tmp = cur;
        while ((ch = str_decode(s, &tmp, len)) && !str_chr(delim, ch))
                cur = tmp;
        end = &s[cur];
        if (next)
                *next = (ch ? &s[tmp] : &s[cur]);

        if (start == end)
                return NULL;    /* No more tokens. */

        /* Overwrite delimiter with NULL terminator. */
        *end = '\0';
        return start;
}

void order_suffix(const uint64_t val, uint64_t *rv, char *suffix)
{
        if (val > UINT64_C(10000000000000000000)) {
                *rv = val / UINT64_C(1000000000000000000);
                *suffix = 'Z';
        } else if (val > UINT64_C(1000000000000000000)) {
                *rv = val / UINT64_C(1000000000000000);
                *suffix = 'E';
        } else if (val > UINT64_C(1000000000000000)) {
                *rv = val / UINT64_C(1000000000000);
                *suffix = 'T';
        } else if (val > UINT64_C(1000000000000)) {
                *rv = val / UINT64_C(1000000000);
                *suffix = 'G';
        } else if (val > UINT64_C(1000000000)) {
                *rv = val / UINT64_C(1000000);
                *suffix = 'M';
        } else if (val > UINT64_C(1000000)) {
                *rv = val / UINT64_C(1000);
                *suffix = 'k';
        } else {
                *rv = val;
                *suffix = ' ';
        }
}

void bin_order_suffix(const uint64_t val, uint64_t *rv, const char **suffix,
    bool fixed)
{
        if (val > UINT64_C(1152921504606846976)) {
                *rv = val / UINT64_C(1125899906842624);
                *suffix = "EiB";
        } else if (val > UINT64_C(1125899906842624)) {
                *rv = val / UINT64_C(1099511627776);
                *suffix = "TiB";
        } else if (val > UINT64_C(1099511627776)) {
                *rv = val / UINT64_C(1073741824);
                *suffix = "GiB";
        } else if (val > UINT64_C(1073741824)) {
                *rv = val / UINT64_C(1048576);
                *suffix = "MiB";
        } else if (val > UINT64_C(1048576)) {
                *rv = val / UINT64_C(1024);
                *suffix = "KiB";
        } else {
                *rv = val;
                if (fixed)
                        *suffix = "B  ";
                else
                        *suffix = "B";
        }
}

/** @}
 */