common/wcwidth.c

   1 /*
   2  * This is an implementation of wcwidth() and wcswidth() as defined in
   3  * "The Single UNIX Specification, Version 2, The Open Group, 1997"
   4  * <http://www.UNIX-systems.org/online.html>
   5  *
   6  * Markus Kuhn -- 2000-02-08 -- public domain
   7  */
   8
   9 #include <wchar.h>
  10
  11 /* These functions define the column width of an ISO 10646 character
  12  * as follows:
  13  *
  14  *    - The null character (U+0000) has a column width of 0.
  15  *
  16  *    - Other C0/C1 control characters and DEL will lead to a return
  17  *      value of -1.
  18  *
  19  *    - Non-spacing and enclosing combining characters (general
  20  *      category code Mn or Me in the Unicode database) have a
  21  *      column width of 0.
  22  *
  23  *    - Spacing characters in the East Asian Wide (W) or East Asian
  24  *      FullWidth (F) category as defined in Unicode Technical
  25  *      Report #11 have a column width of 2.
  26  *
  27  *    - All remaining characters (including all printable
  28  *      ISO 8859-1 and WGL4 characters, Unicode control characters,
  29  *      etc.) have a column width of 1.
  30  *
  31  * This implementation assumes that wchar_t characters are encoded
  32  * in ISO 10646.
  33  */
  34
  35 int ucswidth(wchar_t ucs)
  36 {
  37   /* sorted list of non-overlapping intervals of non-spacing characters */
  38   static const struct interval {
  39     unsigned short first;
  40     unsigned short last;
  41   } combining[] = {
  42     { 0x0300, 0x034E }, { 0x0360, 0x0362 }, { 0x0483, 0x0486 },
  43     { 0x0488, 0x0489 }, { 0x0591, 0x05A1 }, { 0x05A3, 0x05B9 },
  44     { 0x05BB, 0x05BD }, { 0x05BF, 0x05BF }, { 0x05C1, 0x05C2 },
  45     { 0x05C4, 0x05C4 }, { 0x064B, 0x0655 }, { 0x0670, 0x0670 },
  46     { 0x06D6, 0x06E4 }, { 0x06E7, 0x06E8 }, { 0x06EA, 0x06ED },
  47     { 0x0711, 0x0711 }, { 0x0730, 0x074A }, { 0x07A6, 0x07B0 },
  48     { 0x0901, 0x0902 }, { 0x093C, 0x093C }, { 0x0941, 0x0948 },
  49     { 0x094D, 0x094D }, { 0x0951, 0x0954 }, { 0x0962, 0x0963 },
  50     { 0x0981, 0x0981 }, { 0x09BC, 0x09BC }, { 0x09C1, 0x09C4 },
  51     { 0x09CD, 0x09CD }, { 0x09E2, 0x09E3 }, { 0x0A02, 0x0A02 },
  52     { 0x0A3C, 0x0A3C }, { 0x0A41, 0x0A42 }, { 0x0A47, 0x0A48 },
  53     { 0x0A4B, 0x0A4D }, { 0x0A70, 0x0A71 }, { 0x0A81, 0x0A82 },
  54     { 0x0ABC, 0x0ABC }, { 0x0AC1, 0x0AC5 }, { 0x0AC7, 0x0AC8 },
  55     { 0x0ACD, 0x0ACD }, { 0x0B01, 0x0B01 }, { 0x0B3C, 0x0B3C },
  56     { 0x0B3F, 0x0B3F }, { 0x0B41, 0x0B43 }, { 0x0B4D, 0x0B4D },
  57     { 0x0B56, 0x0B56 }, { 0x0B82, 0x0B82 }, { 0x0BC0, 0x0BC0 },
  58     { 0x0BCD, 0x0BCD }, { 0x0C3E, 0x0C40 }, { 0x0C46, 0x0C48 },
  59     { 0x0C4A, 0x0C4D }, { 0x0C55, 0x0C56 }, { 0x0CBF, 0x0CBF },
  60     { 0x0CC6, 0x0CC6 }, { 0x0CCC, 0x0CCD }, { 0x0D41, 0x0D43 },
  61     { 0x0D4D, 0x0D4D }, { 0x0DCA, 0x0DCA }, { 0x0DD2, 0x0DD4 },
  62     { 0x0DD6, 0x0DD6 }, { 0x0E31, 0x0E31 }, { 0x0E34, 0x0E3A },
  63     { 0x0E47, 0x0E4E }, { 0x0EB1, 0x0EB1 }, { 0x0EB4, 0x0EB9 },
  64     { 0x0EBB, 0x0EBC }, { 0x0EC8, 0x0ECD }, { 0x0F18, 0x0F19 },
  65     { 0x0F35, 0x0F35 }, { 0x0F37, 0x0F37 }, { 0x0F39, 0x0F39 },
  66     { 0x0F71, 0x0F7E }, { 0x0F80, 0x0F84 }, { 0x0F86, 0x0F87 },
  67     { 0x0F90, 0x0F97 }, { 0x0F99, 0x0FBC }, { 0x0FC6, 0x0FC6 },
  68     { 0x102D, 0x1030 }, { 0x1032, 0x1032 }, { 0x1036, 0x1037 },
  69     { 0x1039, 0x1039 }, { 0x1058, 0x1059 }, { 0x17B7, 0x17BD },
  70     { 0x17C6, 0x17C6 }, { 0x17C9, 0x17D3 }, { 0x18A9, 0x18A9 },
  71     { 0x20D0, 0x20E3 }, { 0x302A, 0x302F }, { 0x3099, 0x309A },
  72     { 0xFB1E, 0xFB1E }, { 0xFE20, 0xFE23 }
  73   };
  74   int min = 0;
  75   int max = sizeof(combining) / sizeof(struct interval) - 1;
  76   int mid;
  77
  78   /* test for 8-bit control characters */
  79   if (ucs == 0)
  80     return 0;
  81   if (ucs < 32 || (ucs >= 0x7f && ucs < 0xa0))
  82     return -1;
  83
  84   /* first quick check for Latin-1 etc. characters */
  85   if (ucs < combining[0].first)
  86     return 1;
  87
  88   /* binary search in table of non-spacing characters */
  89   while (max >= min) {
  90     mid = (min + max) / 2;
  91     if (combining[mid].last < ucs)
  92       min = mid + 1;
  93     else if (combining[mid].first > ucs)
  94       max = mid - 1;
  95     else if (combining[mid].first <= ucs && combining[mid].last >= ucs)
  96       return 0;
  97   }
  98
  99   /* if we arrive here, ucs is not a combining or C0/C1 control character */
 100
 101   /* fast test for majority of non-wide scripts */
 102   if (ucs < 0x1100)
 103     return 1;
 104
 105   return 1 +
 106     ((ucs >= 0x1100 && ucs <= 0x115f) || /* Hangul Jamo */
 107      (ucs >= 0x2e80 && ucs <= 0xa4cf && (ucs & ~0x0011) != 0x300a &&
 108       ucs != 0x303f) ||                  /* CJK ... Yi */
 109      (ucs >= 0xac00 && ucs <= 0xd7a3) || /* Hangul Syllables */
 110      (ucs >= 0xf900 && ucs <= 0xfaff) || /* CJK Compatibility Ideographs */
 111      (ucs >= 0xfe30 && ucs <= 0xfe6f) || /* CJK Compatibility Forms */
 112      (ucs >= 0xff00 && ucs <= 0xff5f) || /* Fullwidth Forms */
 113      (ucs >= 0xffe0 && ucs <= 0xffe6));
 114 }
 115
 116
 117 #if 0
 118 int wcswidth(const wchar_t *pwcs, size_t n)
 119 {
 120   int w, width = 0;
 121
 122   for (;*pwcs && n-- > 0; pwcs++)
 123     if ((w = wcwidth(*pwcs)) < 0)
 124       return -1;
 125     else
 126       width += w;
 127
 128   return width;
 129 }
 130 #endif