Subversion Repositories Kolibri OS

Rev

Blame | Last modification | View Log | RSS feed

  1. /*
  2. FUNCTION
  3.         <<wcwidth>>---number of column positions of a wide-character code
  4.        
  5. INDEX
  6.         wcwidth
  7.  
  8. ANSI_SYNOPSIS
  9.         #include <wchar.h>
  10.         int wcwidth(const wchar_t <[wc]>);
  11.  
  12. TRAD_SYNOPSIS
  13.         #include <wchar.h>
  14.         int wcwidth(<[wc]>)
  15.         wchar_t *<[wc]>;
  16.  
  17. DESCRIPTION
  18.         The <<wcwidth>> function shall determine the number of column
  19.         positions required for the wide character <[wc]>. The application
  20.         shall ensure that the value of <[wc]> is a character representable
  21.         as a wchar_t, and is a wide-character code corresponding to a
  22.         valid character in the current locale.
  23.  
  24. RETURNS
  25.         The <<wcwidth>> function shall either return 0 (if <[wc]> is a null
  26.         wide-character code), or return the number of column positions to
  27.         be occupied by the wide-character code <[wc]>, or return -1 (if <[wc]>
  28.         does not correspond to a printable wide-character code).
  29.  
  30. PORTABILITY
  31. <<wcwidth>> has been introduced in the Single UNIX Specification Volume 2.
  32. <<wcwidth>> has been marked as an extension in the Single UNIX Specification Volume 3.
  33. */
  34.  
  35. /*
  36.  * This is an implementation of wcwidth() (defined in
  37.  * IEEE Std 1002.1-2001) for Unicode.
  38.  *
  39.  * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
  40.  *
  41.  * In fixed-width output devices, Latin characters all occupy a single
  42.  * "cell" position of equal width, whereas ideographic CJK characters
  43.  * occupy two such cells. Interoperability between terminal-line
  44.  * applications and (teletype-style) character terminals using the
  45.  * UTF-8 encoding requires agreement on which character should advance
  46.  * the cursor by how many cell positions. No established formal
  47.  * standards exist at present on which Unicode character shall occupy
  48.  * how many cell positions on character terminals. These routines are
  49.  * a first attempt of defining such behavior based on simple rules
  50.  * applied to data provided by the Unicode Consortium.
  51.  *
  52.  * For some graphical characters, the Unicode standard explicitly
  53.  * defines a character-cell width via the definition of the East Asian
  54.  * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
  55.  * In all these cases, there is no ambiguity about which width a
  56.  * terminal shall use. For characters in the East Asian Ambiguous (A)
  57.  * class, the width choice depends purely on a preference of backward
  58.  * compatibility with either historic CJK or Western practice.
  59.  * Choosing single-width for these characters is easy to justify as
  60.  * the appropriate long-term solution, as the CJK practice of
  61.  * displaying these characters as double-width comes from historic
  62.  * implementation simplicity (8-bit encoded characters were displayed
  63.  * single-width and 16-bit ones double-width, even for Greek,
  64.  * Cyrillic, etc.) and not any typographic considerations.
  65.  *
  66.  * Much less clear is the choice of width for the Not East Asian
  67.  * (Neutral) class. Existing practice does not dictate a width for any
  68.  * of these characters. It would nevertheless make sense
  69.  * typographically to allocate two character cells to characters such
  70.  * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
  71.  * represented adequately with a single-width glyph. The following
  72.  * routines at present merely assign a single-cell width to all
  73.  * neutral characters, in the interest of simplicity. This is not
  74.  * entirely satisfactory and should be reconsidered before
  75.  * establishing a formal standard in this area. At the moment, the
  76.  * decision which Not East Asian (Neutral) characters should be
  77.  * represented by double-width glyphs cannot yet be answered by
  78.  * applying a simple rule from the Unicode database content. Setting
  79.  * up a proper standard for the behavior of UTF-8 character terminals
  80.  * will require a careful analysis not only of each Unicode character,
  81.  * but also of each presentation form, something the author of these
  82.  * routines has avoided to do so far.
  83.  *
  84.  * http://www.unicode.org/unicode/reports/tr11/
  85.  *
  86.  * Markus Kuhn -- 2007-05-26 (Unicode 5.0)
  87.  *
  88.  * Permission to use, copy, modify, and distribute this software
  89.  * for any purpose and without fee is hereby granted. The author
  90.  * disclaims all warranties with regard to this software.
  91.  *
  92.  * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
  93.  */
  94.  
  95. #include <_ansi.h>
  96. #include <wchar.h>
  97. #ifndef _MB_CAPABLE
  98. #include <wctype.h> /* iswprint, iswcntrl */
  99. #endif
  100. #include "local.h"
  101.  
  102. #ifdef _MB_CAPABLE
  103. struct interval
  104. {
  105.   int first;
  106.   int last;
  107. };
  108.  
  109. /* auxiliary function for binary search in interval table */
  110. static int
  111. bisearch(wint_t ucs, const struct interval *table, int max)
  112. {
  113.   int min = 0;
  114.   int mid;
  115.  
  116.   if (ucs < table[0].first || ucs > table[max].last)
  117.     return 0;
  118.   while (max >= min)
  119.     {
  120.       mid = (min + max) / 2;
  121.       if (ucs > table[mid].last)
  122.         min = mid + 1;
  123.       else if (ucs < table[mid].first)
  124.         max = mid - 1;
  125.       else
  126.         return 1;
  127.     }
  128.  
  129.   return 0;
  130. }
  131. #endif /* _MB_CAPABLE */
  132.  
  133. /* The following function defines the column width of an ISO 10646
  134.  * character as follows:
  135.  *
  136.  *    - The null character (U+0000) has a column width of 0.
  137.  *
  138.  *    - Other C0/C1 control characters and DEL will lead to a return
  139.  *      value of -1.
  140.  *
  141.  *    - If the current language is recognized as a language usually using
  142.  *      CJK fonts, spacing characters in the East Asian Ambiguous (A)
  143.  *      category as defined in Unicode Technical Report #11 have a column
  144.  *      width of 2.
  145.  *
  146.  *    - Non-spacing and enclosing combining characters (general
  147.  *      category code Mn or Me in the Unicode database) have a
  148.  *      column width of 0.
  149.  *
  150.  *    - SOFT HYPHEN (U+00AD) has a column width of 1.
  151.  *
  152.  *    - Other format characters (general category code Cf in the Unicode
  153.  *      database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
  154.  *
  155.  *    - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
  156.  *      have a column width of 0.
  157.  *
  158.  *    - Spacing characters in the East Asian Wide (W) or East Asian
  159.  *      Full-width (F) category as defined in Unicode Technical
  160.  *      Report #11 have a column width of 2.
  161.  *
  162.  *    - All remaining characters (including all printable
  163.  *      ISO 8859-1 and WGL4 characters, Unicode control characters,
  164.  *      etc.) have a column width of 1.
  165.  *
  166.  * This implementation assumes that wint_t characters are encoded
  167.  * in ISO 10646.
  168.  */
  169.  
  170. int
  171. _DEFUN (__wcwidth, (ucs),
  172.         _CONST wint_t ucs)
  173. {
  174. #ifdef _MB_CAPABLE
  175.   /* sorted list of non-overlapping intervals of East Asian Ambiguous
  176.    * characters, generated by "uniset +WIDTH-A -cat=Me -cat=Mn -cat=Cf c" */
  177.   static const struct interval ambiguous[] = {
  178.     { 0x00A1, 0x00A1 }, { 0x00A4, 0x00A4 }, { 0x00A7, 0x00A8 },
  179.     { 0x00AA, 0x00AA }, { 0x00AE, 0x00AE }, { 0x00B0, 0x00B4 },
  180.     { 0x00B6, 0x00BA }, { 0x00BC, 0x00BF }, { 0x00C6, 0x00C6 },
  181.     { 0x00D0, 0x00D0 }, { 0x00D7, 0x00D8 }, { 0x00DE, 0x00E1 },
  182.     { 0x00E6, 0x00E6 }, { 0x00E8, 0x00EA }, { 0x00EC, 0x00ED },
  183.     { 0x00F0, 0x00F0 }, { 0x00F2, 0x00F3 }, { 0x00F7, 0x00FA },
  184.     { 0x00FC, 0x00FC }, { 0x00FE, 0x00FE }, { 0x0101, 0x0101 },
  185.     { 0x0111, 0x0111 }, { 0x0113, 0x0113 }, { 0x011B, 0x011B },
  186.     { 0x0126, 0x0127 }, { 0x012B, 0x012B }, { 0x0131, 0x0133 },
  187.     { 0x0138, 0x0138 }, { 0x013F, 0x0142 }, { 0x0144, 0x0144 },
  188.     { 0x0148, 0x014B }, { 0x014D, 0x014D }, { 0x0152, 0x0153 },
  189.     { 0x0166, 0x0167 }, { 0x016B, 0x016B }, { 0x01CE, 0x01CE },
  190.     { 0x01D0, 0x01D0 }, { 0x01D2, 0x01D2 }, { 0x01D4, 0x01D4 },
  191.     { 0x01D6, 0x01D6 }, { 0x01D8, 0x01D8 }, { 0x01DA, 0x01DA },
  192.     { 0x01DC, 0x01DC }, { 0x0251, 0x0251 }, { 0x0261, 0x0261 },
  193.     { 0x02C4, 0x02C4 }, { 0x02C7, 0x02C7 }, { 0x02C9, 0x02CB },
  194.     { 0x02CD, 0x02CD }, { 0x02D0, 0x02D0 }, { 0x02D8, 0x02DB },
  195.     { 0x02DD, 0x02DD }, { 0x02DF, 0x02DF }, { 0x0391, 0x03A1 },
  196.     { 0x03A3, 0x03A9 }, { 0x03B1, 0x03C1 }, { 0x03C3, 0x03C9 },
  197.     { 0x0401, 0x0401 }, { 0x0410, 0x044F }, { 0x0451, 0x0451 },
  198.     { 0x2010, 0x2010 }, { 0x2013, 0x2016 }, { 0x2018, 0x2019 },
  199.     { 0x201C, 0x201D }, { 0x2020, 0x2022 }, { 0x2024, 0x2027 },
  200.     { 0x2030, 0x2030 }, { 0x2032, 0x2033 }, { 0x2035, 0x2035 },
  201.     { 0x203B, 0x203B }, { 0x203E, 0x203E }, { 0x2074, 0x2074 },
  202.     { 0x207F, 0x207F }, { 0x2081, 0x2084 }, { 0x20AC, 0x20AC },
  203.     { 0x2103, 0x2103 }, { 0x2105, 0x2105 }, { 0x2109, 0x2109 },
  204.     { 0x2113, 0x2113 }, { 0x2116, 0x2116 }, { 0x2121, 0x2122 },
  205.     { 0x2126, 0x2126 }, { 0x212B, 0x212B }, { 0x2153, 0x2154 },
  206.     { 0x215B, 0x215E }, { 0x2160, 0x216B }, { 0x2170, 0x2179 },
  207.     { 0x2190, 0x2199 }, { 0x21B8, 0x21B9 }, { 0x21D2, 0x21D2 },
  208.     { 0x21D4, 0x21D4 }, { 0x21E7, 0x21E7 }, { 0x2200, 0x2200 },
  209.     { 0x2202, 0x2203 }, { 0x2207, 0x2208 }, { 0x220B, 0x220B },
  210.     { 0x220F, 0x220F }, { 0x2211, 0x2211 }, { 0x2215, 0x2215 },
  211.     { 0x221A, 0x221A }, { 0x221D, 0x2220 }, { 0x2223, 0x2223 },
  212.     { 0x2225, 0x2225 }, { 0x2227, 0x222C }, { 0x222E, 0x222E },
  213.     { 0x2234, 0x2237 }, { 0x223C, 0x223D }, { 0x2248, 0x2248 },
  214.     { 0x224C, 0x224C }, { 0x2252, 0x2252 }, { 0x2260, 0x2261 },
  215.     { 0x2264, 0x2267 }, { 0x226A, 0x226B }, { 0x226E, 0x226F },
  216.     { 0x2282, 0x2283 }, { 0x2286, 0x2287 }, { 0x2295, 0x2295 },
  217.     { 0x2299, 0x2299 }, { 0x22A5, 0x22A5 }, { 0x22BF, 0x22BF },
  218.     { 0x2312, 0x2312 }, { 0x2460, 0x24E9 }, { 0x24EB, 0x254B },
  219.     { 0x2550, 0x2573 }, { 0x2580, 0x258F }, { 0x2592, 0x2595 },
  220.     { 0x25A0, 0x25A1 }, { 0x25A3, 0x25A9 }, { 0x25B2, 0x25B3 },
  221.     { 0x25B6, 0x25B7 }, { 0x25BC, 0x25BD }, { 0x25C0, 0x25C1 },
  222.     { 0x25C6, 0x25C8 }, { 0x25CB, 0x25CB }, { 0x25CE, 0x25D1 },
  223.     { 0x25E2, 0x25E5 }, { 0x25EF, 0x25EF }, { 0x2605, 0x2606 },
  224.     { 0x2609, 0x2609 }, { 0x260E, 0x260F }, { 0x2614, 0x2615 },
  225.     { 0x261C, 0x261C }, { 0x261E, 0x261E }, { 0x2640, 0x2640 },
  226.     { 0x2642, 0x2642 }, { 0x2660, 0x2661 }, { 0x2663, 0x2665 },
  227.     { 0x2667, 0x266A }, { 0x266C, 0x266D }, { 0x266F, 0x266F },
  228.     { 0x273D, 0x273D }, { 0x2776, 0x277F }, { 0xE000, 0xF8FF },
  229.     { 0xFFFD, 0xFFFD }, { 0xF0000, 0xFFFFD }, { 0x100000, 0x10FFFD }
  230.   };
  231.   /* sorted list of non-overlapping intervals of non-spacing characters */
  232.   /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
  233.   static const struct interval combining[] = {
  234.     { 0x0300, 0x036F }, { 0x0483, 0x0486 }, { 0x0488, 0x0489 },
  235.     { 0x0591, 0x05BD }, { 0x05BF, 0x05BF }, { 0x05C1, 0x05C2 },
  236.     { 0x05C4, 0x05C5 }, { 0x05C7, 0x05C7 }, { 0x0600, 0x0603 },
  237.     { 0x0610, 0x0615 }, { 0x064B, 0x065E }, { 0x0670, 0x0670 },
  238.     { 0x06D6, 0x06E4 }, { 0x06E7, 0x06E8 }, { 0x06EA, 0x06ED },
  239.     { 0x070F, 0x070F }, { 0x0711, 0x0711 }, { 0x0730, 0x074A },
  240.     { 0x07A6, 0x07B0 }, { 0x07EB, 0x07F3 }, { 0x0901, 0x0902 },
  241.     { 0x093C, 0x093C }, { 0x0941, 0x0948 }, { 0x094D, 0x094D },
  242.     { 0x0951, 0x0954 }, { 0x0962, 0x0963 }, { 0x0981, 0x0981 },
  243.     { 0x09BC, 0x09BC }, { 0x09C1, 0x09C4 }, { 0x09CD, 0x09CD },
  244.     { 0x09E2, 0x09E3 }, { 0x0A01, 0x0A02 }, { 0x0A3C, 0x0A3C },
  245.     { 0x0A41, 0x0A42 }, { 0x0A47, 0x0A48 }, { 0x0A4B, 0x0A4D },
  246.     { 0x0A70, 0x0A71 }, { 0x0A81, 0x0A82 }, { 0x0ABC, 0x0ABC },
  247.     { 0x0AC1, 0x0AC5 }, { 0x0AC7, 0x0AC8 }, { 0x0ACD, 0x0ACD },
  248.     { 0x0AE2, 0x0AE3 }, { 0x0B01, 0x0B01 }, { 0x0B3C, 0x0B3C },
  249.     { 0x0B3F, 0x0B3F }, { 0x0B41, 0x0B43 }, { 0x0B4D, 0x0B4D },
  250.     { 0x0B56, 0x0B56 }, { 0x0B82, 0x0B82 }, { 0x0BC0, 0x0BC0 },
  251.     { 0x0BCD, 0x0BCD }, { 0x0C3E, 0x0C40 }, { 0x0C46, 0x0C48 },
  252.     { 0x0C4A, 0x0C4D }, { 0x0C55, 0x0C56 }, { 0x0CBC, 0x0CBC },
  253.     { 0x0CBF, 0x0CBF }, { 0x0CC6, 0x0CC6 }, { 0x0CCC, 0x0CCD },
  254.     { 0x0CE2, 0x0CE3 }, { 0x0D41, 0x0D43 }, { 0x0D4D, 0x0D4D },
  255.     { 0x0DCA, 0x0DCA }, { 0x0DD2, 0x0DD4 }, { 0x0DD6, 0x0DD6 },
  256.     { 0x0E31, 0x0E31 }, { 0x0E34, 0x0E3A }, { 0x0E47, 0x0E4E },
  257.     { 0x0EB1, 0x0EB1 }, { 0x0EB4, 0x0EB9 }, { 0x0EBB, 0x0EBC },
  258.     { 0x0EC8, 0x0ECD }, { 0x0F18, 0x0F19 }, { 0x0F35, 0x0F35 },
  259.     { 0x0F37, 0x0F37 }, { 0x0F39, 0x0F39 }, { 0x0F71, 0x0F7E },
  260.     { 0x0F80, 0x0F84 }, { 0x0F86, 0x0F87 }, { 0x0F90, 0x0F97 },
  261.     { 0x0F99, 0x0FBC }, { 0x0FC6, 0x0FC6 }, { 0x102D, 0x1030 },
  262.     { 0x1032, 0x1032 }, { 0x1036, 0x1037 }, { 0x1039, 0x1039 },
  263.     { 0x1058, 0x1059 }, { 0x1160, 0x11FF }, { 0x135F, 0x135F },
  264.     { 0x1712, 0x1714 }, { 0x1732, 0x1734 }, { 0x1752, 0x1753 },
  265.     { 0x1772, 0x1773 }, { 0x17B4, 0x17B5 }, { 0x17B7, 0x17BD },
  266.     { 0x17C6, 0x17C6 }, { 0x17C9, 0x17D3 }, { 0x17DD, 0x17DD },
  267.     { 0x180B, 0x180D }, { 0x18A9, 0x18A9 }, { 0x1920, 0x1922 },
  268.     { 0x1927, 0x1928 }, { 0x1932, 0x1932 }, { 0x1939, 0x193B },
  269.     { 0x1A17, 0x1A18 }, { 0x1B00, 0x1B03 }, { 0x1B34, 0x1B34 },
  270.     { 0x1B36, 0x1B3A }, { 0x1B3C, 0x1B3C }, { 0x1B42, 0x1B42 },
  271.     { 0x1B6B, 0x1B73 }, { 0x1DC0, 0x1DCA }, { 0x1DFE, 0x1DFF },
  272.     { 0x200B, 0x200F }, { 0x202A, 0x202E }, { 0x2060, 0x2063 },
  273.     { 0x206A, 0x206F }, { 0x20D0, 0x20EF }, { 0x302A, 0x302F },
  274.     { 0x3099, 0x309A }, { 0xA806, 0xA806 }, { 0xA80B, 0xA80B },
  275.     { 0xA825, 0xA826 }, { 0xFB1E, 0xFB1E }, { 0xFE00, 0xFE0F },
  276.     { 0xFE20, 0xFE23 }, { 0xFEFF, 0xFEFF }, { 0xFFF9, 0xFFFB },
  277.     { 0x10A01, 0x10A03 }, { 0x10A05, 0x10A06 }, { 0x10A0C, 0x10A0F },
  278.     { 0x10A38, 0x10A3A }, { 0x10A3F, 0x10A3F }, { 0x1D167, 0x1D169 },
  279.     { 0x1D173, 0x1D182 }, { 0x1D185, 0x1D18B }, { 0x1D1AA, 0x1D1AD },
  280.     { 0x1D242, 0x1D244 }, { 0xE0001, 0xE0001 }, { 0xE0020, 0xE007F },
  281.     { 0xE0100, 0xE01EF }
  282.   };
  283.  
  284.   /* Test for NUL character */
  285.   if (ucs == 0)
  286.     return 0;
  287.  
  288.   /* Test for printable ASCII characters */
  289.   if (ucs >= 0x20 && ucs < 0x7f)
  290.     return 1;
  291.  
  292.   /* Test for control characters */
  293.   if (ucs < 0xa0)
  294.     return -1;
  295.  
  296.   /* Test for surrogate pair values. */
  297.   if (ucs >= 0xd800 && ucs <= 0xdfff)
  298.     return -1;
  299.  
  300.   /* binary search in table of ambiguous characters */
  301.   if (__locale_cjk_lang ()
  302.       && bisearch(ucs, ambiguous,
  303.                   sizeof(ambiguous) / sizeof(struct interval) - 1))
  304.     return 2;
  305.  
  306.   /* binary search in table of non-spacing characters */
  307.   if (bisearch(ucs, combining,
  308.                sizeof(combining) / sizeof(struct interval) - 1))
  309.     return 0;
  310.  
  311.   /* if we arrive here, ucs is not a combining or C0/C1 control character */
  312.  
  313.   return 1 +
  314.     (ucs >= 0x1100 &&
  315.      (ucs <= 0x115f ||                    /* Hangul Jamo init. consonants */
  316.       ucs == 0x2329 || ucs == 0x232a ||
  317.       (ucs >= 0x2e80 && ucs <= 0xa4cf &&
  318.        ucs != 0x303f) ||                  /* CJK ... Yi */
  319.       (ucs >= 0xac00 && ucs <= 0xd7a3) || /* Hangul Syllables */
  320.       (ucs >= 0xf900 && ucs <= 0xfaff) || /* CJK Compatibility Ideographs */
  321.       (ucs >= 0xfe10 && ucs <= 0xfe19) || /* Vertical forms */
  322.       (ucs >= 0xfe30 && ucs <= 0xfe6f) || /* CJK Compatibility Forms */
  323.       (ucs >= 0xff00 && ucs <= 0xff60) || /* Fullwidth Forms */
  324.       (ucs >= 0xffe0 && ucs <= 0xffe6) ||
  325.       (ucs >= 0x20000 && ucs <= 0x2fffd) ||
  326.       (ucs >= 0x30000 && ucs <= 0x3fffd)));
  327. #else /* !_MB_CAPABLE */
  328.   if (iswprint (ucs))
  329.     return 1;
  330.   if (iswcntrl (ucs) || ucs == L'\0')
  331.     return 0;
  332.   return -1;
  333. #endif /* _MB_CAPABLE */
  334. }
  335.  
  336. int    
  337. _DEFUN (wcwidth, (wc),
  338.         _CONST wchar_t wc)
  339. {
  340.   wint_t wi = wc;
  341.  
  342. #ifdef _MB_CAPABLE
  343.   wi = _jp2uc (wi);
  344. #endif /* _MB_CAPABLE */
  345.   return __wcwidth (wi);
  346. }
  347.