newlib/libc/string/wcwidth.c

   1 /*
   2 FUNCTION
   3         <<wcwidth>>---number of column positions of a wide-character code
   4
   5 INDEX
   6         wcwidth
   7
   8 SYNOPSIS
   9         #include <wchar.h>
  10         int wcwidth(const wint_t <[wc]>);
  11
  12 DESCRIPTION
  13         The <<wcwidth>> function shall determine the number of column
  14         positions required for the wide character <[wc]>. The application
  15         shall ensure that the value of <[wc]> is a character representable
  16         as a wint_t (combining Unicode surrogate pairs into single 21-bit
  17         Unicode code points), and is a wide-character code corresponding to a
  18         valid character in the current locale.
  19
  20 RETURNS
  21         The <<wcwidth>> function shall either return 0 (if <[wc]> is a null
  22         wide-character code), or return the number of column positions to
  23         be occupied by the wide-character code <[wc]>, or return -1 (if <[wc]>
  24         does not correspond to a printable wide-character code).
  25
  26 PORTABILITY
  27 <<wcwidth>> has been introduced in the Single UNIX Specification Volume 2.
  28 <<wcwidth>> has been marked as an extension in the Single UNIX Specification Volume 3.
  29 */
  30
  31 /*
  32  * This is an implementation of wcwidth() (defined in
  33  * IEEE Std 1002.1-2001) for Unicode.
  34  *
  35  * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
  36  *
  37  * In fixed-width output devices, Latin characters all occupy a single
  38  * "cell" position of equal width, whereas ideographic CJK characters
  39  * occupy two such cells. Interoperability between terminal-line
  40  * applications and (teletype-style) character terminals using the
  41  * UTF-8 encoding requires agreement on which character should advance
  42  * the cursor by how many cell positions. No established formal
  43  * standards exist at present on which Unicode character shall occupy
  44  * how many cell positions on character terminals. These routines are
  45  * a first attempt of defining such behavior based on simple rules
  46  * applied to data provided by the Unicode Consortium.
  47  *
  48  * For some graphical characters, the Unicode standard explicitly
  49  * defines a character-cell width via the definition of the East Asian
  50  * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
  51  * In all these cases, there is no ambiguity about which width a
  52  * terminal shall use. For characters in the East Asian Ambiguous (A)
  53  * class, the width choice depends purely on a preference of backward
  54  * compatibility with either historic CJK or Western practice.
  55  * Choosing single-width for these characters is easy to justify as
  56  * the appropriate long-term solution, as the CJK practice of
  57  * displaying these characters as double-width comes from historic
  58  * implementation simplicity (8-bit encoded characters were displayed
  59  * single-width and 16-bit ones double-width, even for Greek,
  60  * Cyrillic, etc.) and not any typographic considerations.
  61  *
  62  * Much less clear is the choice of width for the Not East Asian
  63  * (Neutral) class. Existing practice does not dictate a width for any
  64  * of these characters. It would nevertheless make sense
  65  * typographically to allocate two character cells to characters such
  66  * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
  67  * represented adequately with a single-width glyph. The following
  68  * routines at present merely assign a single-cell width to all
  69  * neutral characters, in the interest of simplicity. This is not
  70  * entirely satisfactory and should be reconsidered before
  71  * establishing a formal standard in this area. At the moment, the
  72  * decision which Not East Asian (Neutral) characters should be
  73  * represented by double-width glyphs cannot yet be answered by
  74  * applying a simple rule from the Unicode database content. Setting
  75  * up a proper standard for the behavior of UTF-8 character terminals
  76  * will require a careful analysis not only of each Unicode character,
  77  * but also of each presentation form, something the author of these
  78  * routines has avoided to do so far.
  79  *
  80  * http://www.unicode.org/unicode/reports/tr11/
  81  *
  82  * Markus Kuhn -- 2007-05-26 (Unicode 5.0)
  83  *
  84  * Permission to use, copy, modify, and distribute this software
  85  * for any purpose and without fee is hereby granted. The author
  86  * disclaims all warranties with regard to this software.
  87  *
  88  * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
  89  */
  90
  91 #include <_ansi.h>
  92 #include <wchar.h>
  93 #ifndef _MB_CAPABLE
  94 #include <wctype.h> /* iswprint, iswcntrl */
  95 #endif
  96 #include "local.h"
  97
  98 #ifdef _MB_CAPABLE
  99 struct interval
 100 {
 101   int first;
 102   int last;
 103 };
 104
 105 /* auxiliary function for binary search in interval table */
 106 static int
 107 bisearch(wint_t ucs, const struct interval *table, int max)
 108 {
 109   int min = 0;
 110   int mid;
 111
 112   if (ucs < table[0].first || ucs > table[max].last)
 113     return 0;
 114   while (max >= min)
 115     {
 116       mid = (min + max) / 2;
 117       if (ucs > table[mid].last)
 118         min = mid + 1;
 119       else if (ucs < table[mid].first)
 120         max = mid - 1;
 121       else
 122         return 1;
 123     }
 124
 125   return 0;
 126 }
 127 #endif /* _MB_CAPABLE */
 128
 129 /* The following function defines the column width of an ISO 10646
 130  * character as follows:
 131  *
 132  *    - The null character (U+0000) has a column width of 0.
 133  *
 134  *    - Other C0/C1 control characters and DEL will lead to a return
 135  *      value of -1.
 136  *
 137  *    - If the current language is recognized as a language usually using
 138  *      CJK fonts, spacing characters in the East Asian Ambiguous (A)
 139  *      category as defined in Unicode Technical Report #11 have a column
 140  *      width of 2.
 141  *
 142  *    - Non-spacing and enclosing combining characters (general
 143  *      category code Mn or Me in the Unicode database) have a
 144  *      column width of 0.
 145  *
 146  *    - SOFT HYPHEN (U+00AD) has a column width of 1.
 147  *
 148  *    - Other format characters (general category code Cf in the Unicode
 149  *      database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
 150  *
 151  *    - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
 152  *      have a column width of 0.
 153  *
 154  *    - Spacing characters in the East Asian Wide (W) or East Asian
 155  *      Full-width (F) category as defined in Unicode Technical
 156  *      Report #11 have a column width of 2.
 157  *
 158  *    - All remaining characters (including all printable
 159  *      ISO 8859-1 and WGL4 characters, Unicode control characters,
 160  *      etc.) have a column width of 1.
 161  *
 162  * This implementation assumes that wint_t characters are encoded
 163  * in ISO 10646.
 164  */
 165
 166 int
 167 __wcwidth (const wint_t ucs)
 168 {
 169 #ifdef _MB_CAPABLE
 170   /* sorted list of non-overlapping intervals of East Asian Ambiguous chars */
 171   static const struct interval ambiguous[] =
 172 #include "ambiguous.t"
 173
 174   /* sorted list of non-overlapping intervals of non-spacing characters */
 175   static const struct interval combining[] =
 176 #include "combining.t"
 177
 178   /* sorted list of non-overlapping intervals of wide characters,
 179      ranges extended to Blocks where possible
 180    */
 181   static const struct interval wide[] =
 182 #include "wide.t"
 183
 184   /* Test for NUL character */
 185   if (ucs == 0)
 186     return 0;
 187
 188   /* Test for printable ASCII characters */
 189   if (ucs >= 0x20 && ucs < 0x7f)
 190     return 1;
 191
 192   /* Test for control characters */
 193   if (ucs < 0xa0)
 194     return -1;
 195
 196   /* Test for surrogate pair values. */
 197   if (ucs >= 0xd800 && ucs <= 0xdfff)
 198     return -1;
 199
 200   /* check CJK width mode (1: ambiguous-wide, 0: normal, -1: disabled) */
 201   int cjk_lang = __locale_cjk_lang ();
 202
 203   /* binary search in table of ambiguous characters */
 204   if (cjk_lang > 0
 205       && bisearch(ucs, ambiguous,
 206                   sizeof(ambiguous) / sizeof(struct interval) - 1))
 207     return 2;
 208
 209   /* binary search in table of non-spacing characters */
 210   if (bisearch(ucs, combining,
 211                sizeof(combining) / sizeof(struct interval) - 1))
 212     return 0;
 213
 214   /* if we arrive here, ucs is not a combining or C0/C1 control character */
 215
 216   /* binary search in table of wide character codes */
 217   if (cjk_lang >= 0
 218       && bisearch(ucs, wide,
 219                sizeof(wide) / sizeof(struct interval) - 1))
 220     return 2;
 221   else
 222     return 1;
 223 #else /* !_MB_CAPABLE */
 224   if (iswprint (ucs))
 225     return 1;
 226   if (iswcntrl (ucs) || ucs == L'\0')
 227     return 0;
 228   return -1;
 229 #endif /* _MB_CAPABLE */
 230 }
 231
 232 int
 233 wcwidth (const wint_t wc)
 234 {
 235   wint_t wi = wc;
 236
 237 #ifdef _MB_CAPABLE
 238   wi = _jp2uc (wi);
 239 #endif /* _MB_CAPABLE */
 240   return __wcwidth (wi);
 241 }