Updated TODO.
[mpdm.git] / wcwidth.c
blobf23d7ac9c0d73c7b795b81d362e96625a8c9784c
1 /*
2 * This is an implementation of wcwidth() and wcswidth() (defined in
3 * IEEE Std 1002.1-2001) for Unicode.
5 * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
6 * http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html
8 * In fixed-width output devices, Latin characters all occupy a single
9 * "cell" position of equal width, whereas ideographic CJK characters
10 * occupy two such cells. Interoperability between terminal-line
11 * applications and (teletype-style) character terminals using the
12 * UTF-8 encoding requires agreement on which character should advance
13 * the cursor by how many cell positions. No established formal
14 * standards exist at present on which Unicode character shall occupy
15 * how many cell positions on character terminals. These routines are
16 * a first attempt of defining such behavior based on simple rules
17 * applied to data provided by the Unicode Consortium.
19 * For some graphical characters, the Unicode standard explicitly
20 * defines a character-cell width via the definition of the East Asian
21 * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
22 * In all these cases, there is no ambiguity about which width a
23 * terminal shall use. For characters in the East Asian Ambiguous (A)
24 * class, the width choice depends purely on a preference of backward
25 * compatibility with either historic CJK or Western practice.
26 * Choosing single-width for these characters is easy to justify as
27 * the appropriate long-term solution, as the CJK practice of
28 * displaying these characters as double-width comes from historic
29 * implementation simplicity (8-bit encoded characters were displayed
30 * single-width and 16-bit ones double-width, even for Greek,
31 * Cyrillic, etc.) and not any typographic considerations.
33 * Much less clear is the choice of width for the Not East Asian
34 * (Neutral) class. Existing practice does not dictate a width for any
35 * of these characters. It would nevertheless make sense
36 * typographically to allocate two character cells to characters such
37 * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
38 * represented adequately with a single-width glyph. The following
39 * routines at present merely assign a single-cell width to all
40 * neutral characters, in the interest of simplicity. This is not
41 * entirely satisfactory and should be reconsidered before
42 * establishing a formal standard in this area. At the moment, the
43 * decision which Not East Asian (Neutral) characters should be
44 * represented by double-width glyphs cannot yet be answered by
45 * applying a simple rule from the Unicode database content. Setting
46 * up a proper standard for the behavior of UTF-8 character terminals
47 * will require a careful analysis not only of each Unicode character,
48 * but also of each presentation form, something the author of these
49 * routines has avoided to do so far.
51 * http://www.unicode.org/unicode/reports/tr11/
53 * Markus Kuhn -- 2003-05-20 (Unicode 4.0)
55 * Permission to use, copy, modify, and distribute this software
56 * for any purpose and without fee is hereby granted. The author
57 * disclaims all warranties with regard to this software.
59 * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
62 #include <wchar.h>
64 struct interval {
65 int first;
66 int last;
69 /* auxiliary function for binary search in interval table */
70 static int bisearch(wchar_t ucs, const struct interval *table, int max)
72 int min = 0;
73 int mid;
75 if (ucs < table[0].first || ucs > table[max].last)
76 return 0;
77 while (max >= min) {
78 mid = (min + max) / 2;
79 if (ucs > table[mid].last)
80 min = mid + 1;
81 else if (ucs < table[mid].first)
82 max = mid - 1;
83 else
84 return 1;
87 return 0;
91 /* The following two functions define the column width of an ISO 10646
92 * character as follows:
94 * - The null character (U+0000) has a column width of 0.
96 * - Other C0/C1 control characters and DEL will lead to a return
97 * value of -1.
99 * - Non-spacing and enclosing combining characters (general
100 * category code Mn or Me in the Unicode database) have a
101 * column width of 0.
103 * - SOFT HYPHEN (U+00AD) has a column width of 1.
105 * - Other format characters (general category code Cf in the Unicode
106 * database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
108 * - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
109 * have a column width of 0.
111 * - Spacing characters in the East Asian Wide (W) or East Asian
112 * Full-width (F) category as defined in Unicode Technical
113 * Report #11 have a column width of 2.
115 * - All remaining characters (including all printable
116 * ISO 8859-1 and WGL4 characters, Unicode control characters,
117 * etc.) have a column width of 1.
119 * This implementation assumes that wchar_t characters are encoded
120 * in ISO 10646.
123 int mk_wcwidth(wchar_t ucs)
125 /* sorted list of non-overlapping intervals of non-spacing characters */
126 /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
127 static const struct interval combining[] = {
128 {0x0300, 0x0357}, {0x035D, 0x036F}, {0x0483, 0x0486},
129 {0x0488, 0x0489}, {0x0591, 0x05A1}, {0x05A3, 0x05B9},
130 {0x05BB, 0x05BD}, {0x05BF, 0x05BF}, {0x05C1, 0x05C2},
131 {0x05C4, 0x05C4}, {0x0600, 0x0603}, {0x0610, 0x0615},
132 {0x064B, 0x0658}, {0x0670, 0x0670}, {0x06D6, 0x06E4},
133 {0x06E7, 0x06E8}, {0x06EA, 0x06ED}, {0x070F, 0x070F},
134 {0x0711, 0x0711}, {0x0730, 0x074A}, {0x07A6, 0x07B0},
135 {0x0901, 0x0902}, {0x093C, 0x093C}, {0x0941, 0x0948},
136 {0x094D, 0x094D}, {0x0951, 0x0954}, {0x0962, 0x0963},
137 {0x0981, 0x0981}, {0x09BC, 0x09BC}, {0x09C1, 0x09C4},
138 {0x09CD, 0x09CD}, {0x09E2, 0x09E3}, {0x0A01, 0x0A02},
139 {0x0A3C, 0x0A3C}, {0x0A41, 0x0A42}, {0x0A47, 0x0A48},
140 {0x0A4B, 0x0A4D}, {0x0A70, 0x0A71}, {0x0A81, 0x0A82},
141 {0x0ABC, 0x0ABC}, {0x0AC1, 0x0AC5}, {0x0AC7, 0x0AC8},
142 {0x0ACD, 0x0ACD}, {0x0AE2, 0x0AE3}, {0x0B01, 0x0B01},
143 {0x0B3C, 0x0B3C}, {0x0B3F, 0x0B3F}, {0x0B41, 0x0B43},
144 {0x0B4D, 0x0B4D}, {0x0B56, 0x0B56}, {0x0B82, 0x0B82},
145 {0x0BC0, 0x0BC0}, {0x0BCD, 0x0BCD}, {0x0C3E, 0x0C40},
146 {0x0C46, 0x0C48}, {0x0C4A, 0x0C4D}, {0x0C55, 0x0C56},
147 {0x0CBC, 0x0CBC}, {0x0CBF, 0x0CBF}, {0x0CC6, 0x0CC6},
148 {0x0CCC, 0x0CCD}, {0x0D41, 0x0D43}, {0x0D4D, 0x0D4D},
149 {0x0DCA, 0x0DCA}, {0x0DD2, 0x0DD4}, {0x0DD6, 0x0DD6},
150 {0x0E31, 0x0E31}, {0x0E34, 0x0E3A}, {0x0E47, 0x0E4E},
151 {0x0EB1, 0x0EB1}, {0x0EB4, 0x0EB9}, {0x0EBB, 0x0EBC},
152 {0x0EC8, 0x0ECD}, {0x0F18, 0x0F19}, {0x0F35, 0x0F35},
153 {0x0F37, 0x0F37}, {0x0F39, 0x0F39}, {0x0F71, 0x0F7E},
154 {0x0F80, 0x0F84}, {0x0F86, 0x0F87}, {0x0F90, 0x0F97},
155 {0x0F99, 0x0FBC}, {0x0FC6, 0x0FC6}, {0x102D, 0x1030},
156 {0x1032, 0x1032}, {0x1036, 0x1037}, {0x1039, 0x1039},
157 {0x1058, 0x1059}, {0x1160, 0x11FF}, {0x1712, 0x1714},
158 {0x1732, 0x1734}, {0x1752, 0x1753}, {0x1772, 0x1773},
159 {0x17B4, 0x17B5}, {0x17B7, 0x17BD}, {0x17C6, 0x17C6},
160 {0x17C9, 0x17D3}, {0x17DD, 0x17DD}, {0x180B, 0x180D},
161 {0x18A9, 0x18A9}, {0x1920, 0x1922}, {0x1927, 0x1928},
162 {0x1932, 0x1932}, {0x1939, 0x193B}, {0x200B, 0x200F},
163 {0x202A, 0x202E}, {0x2060, 0x2063}, {0x206A, 0x206F},
164 {0x20D0, 0x20EA}, {0x302A, 0x302F}, {0x3099, 0x309A},
165 {0xFB1E, 0xFB1E}, {0xFE00, 0xFE0F}, {0xFE20, 0xFE23},
166 {0xFEFF, 0xFEFF}, {0xFFF9, 0xFFFB}, {0x1D167, 0x1D169},
167 {0x1D173, 0x1D182}, {0x1D185, 0x1D18B}, {0x1D1AA, 0x1D1AD},
168 {0xE0001, 0xE0001}, {0xE0020, 0xE007F}, {0xE0100, 0xE01EF}
171 /* test for 8-bit control characters */
172 if (ucs == 0)
173 return 0;
174 if (ucs < 32 || (ucs >= 0x7f && ucs < 0xa0))
175 return -1;
177 /* binary search in table of non-spacing characters */
178 if (bisearch(ucs, combining,
179 sizeof(combining) / sizeof(struct interval) - 1))
180 return 0;
182 /* if we arrive here, ucs is not a combining or C0/C1 control character */
184 return 1 + (ucs >= 0x1100 && (ucs <= 0x115f || /* Hangul Jamo init. consonants */
185 ucs == 0x2329 || ucs == 0x232a || (ucs >= 0x2e80 && ucs <= 0xa4cf && ucs != 0x303f) || /* CJK ... Yi */
186 (ucs >= 0xac00 && ucs <= 0xd7a3) || /* Hangul Syllables */
187 (ucs >= 0xf900 && ucs <= 0xfaff) || /* CJK Compatibility Ideographs */
188 (ucs >= 0xfe30 && ucs <= 0xfe6f) || /* CJK Compatibility Forms */
189 (ucs >= 0xff00 && ucs <= 0xff60) || /* Fullwidth Forms */
190 (ucs >= 0xffe0 && ucs <= 0xffe6) ||
191 (ucs >= 0x20000 && ucs <= 0x2fffd) ||
192 (ucs >= 0x30000 && ucs <= 0x3fffd)));
196 int mk_wcswidth(const wchar_t * pwcs, size_t n)
198 int w, width = 0;
200 for (; *pwcs && n-- > 0; pwcs++)
201 if ((w = mk_wcwidth(*pwcs)) < 0)
202 return -1;
203 else
204 width += w;
206 return width;
211 * The following functions are the same as mk_wcwidth() and
212 * mk_wcwidth_cjk(), except that spacing characters in the East Asian
213 * Ambiguous (A) category as defined in Unicode Technical Report #11
214 * have a column width of 2. This variant might be useful for users of
215 * CJK legacy encodings who want to migrate to UCS without changing
216 * the traditional terminal character-width behaviour. It is not
217 * otherwise recommended for general use.
219 static int mk_wcwidth_cjk(wchar_t ucs)
221 /* sorted list of non-overlapping intervals of East Asian Ambiguous
222 * characters, generated by "uniset +WIDTH-A -cat=Me -cat=Mn -cat=Cf c" */
223 static const struct interval ambiguous[] = {
224 {0x00A1, 0x00A1}, {0x00A4, 0x00A4}, {0x00A7, 0x00A8},
225 {0x00AA, 0x00AA}, {0x00AE, 0x00AE}, {0x00B0, 0x00B4},
226 {0x00B6, 0x00BA}, {0x00BC, 0x00BF}, {0x00C6, 0x00C6},
227 {0x00D0, 0x00D0}, {0x00D7, 0x00D8}, {0x00DE, 0x00E1},
228 {0x00E6, 0x00E6}, {0x00E8, 0x00EA}, {0x00EC, 0x00ED},
229 {0x00F0, 0x00F0}, {0x00F2, 0x00F3}, {0x00F7, 0x00FA},
230 {0x00FC, 0x00FC}, {0x00FE, 0x00FE}, {0x0101, 0x0101},
231 {0x0111, 0x0111}, {0x0113, 0x0113}, {0x011B, 0x011B},
232 {0x0126, 0x0127}, {0x012B, 0x012B}, {0x0131, 0x0133},
233 {0x0138, 0x0138}, {0x013F, 0x0142}, {0x0144, 0x0144},
234 {0x0148, 0x014B}, {0x014D, 0x014D}, {0x0152, 0x0153},
235 {0x0166, 0x0167}, {0x016B, 0x016B}, {0x01CE, 0x01CE},
236 {0x01D0, 0x01D0}, {0x01D2, 0x01D2}, {0x01D4, 0x01D4},
237 {0x01D6, 0x01D6}, {0x01D8, 0x01D8}, {0x01DA, 0x01DA},
238 {0x01DC, 0x01DC}, {0x0251, 0x0251}, {0x0261, 0x0261},
239 {0x02C4, 0x02C4}, {0x02C7, 0x02C7}, {0x02C9, 0x02CB},
240 {0x02CD, 0x02CD}, {0x02D0, 0x02D0}, {0x02D8, 0x02DB},
241 {0x02DD, 0x02DD}, {0x02DF, 0x02DF}, {0x0391, 0x03A1},
242 {0x03A3, 0x03A9}, {0x03B1, 0x03C1}, {0x03C3, 0x03C9},
243 {0x0401, 0x0401}, {0x0410, 0x044F}, {0x0451, 0x0451},
244 {0x2010, 0x2010}, {0x2013, 0x2016}, {0x2018, 0x2019},
245 {0x201C, 0x201D}, {0x2020, 0x2022}, {0x2024, 0x2027},
246 {0x2030, 0x2030}, {0x2032, 0x2033}, {0x2035, 0x2035},
247 {0x203B, 0x203B}, {0x203E, 0x203E}, {0x2074, 0x2074},
248 {0x207F, 0x207F}, {0x2081, 0x2084}, {0x20AC, 0x20AC},
249 {0x2103, 0x2103}, {0x2105, 0x2105}, {0x2109, 0x2109},
250 {0x2113, 0x2113}, {0x2116, 0x2116}, {0x2121, 0x2122},
251 {0x2126, 0x2126}, {0x212B, 0x212B}, {0x2153, 0x2154},
252 {0x215B, 0x215E}, {0x2160, 0x216B}, {0x2170, 0x2179},
253 {0x2190, 0x2199}, {0x21B8, 0x21B9}, {0x21D2, 0x21D2},
254 {0x21D4, 0x21D4}, {0x21E7, 0x21E7}, {0x2200, 0x2200},
255 {0x2202, 0x2203}, {0x2207, 0x2208}, {0x220B, 0x220B},
256 {0x220F, 0x220F}, {0x2211, 0x2211}, {0x2215, 0x2215},
257 {0x221A, 0x221A}, {0x221D, 0x2220}, {0x2223, 0x2223},
258 {0x2225, 0x2225}, {0x2227, 0x222C}, {0x222E, 0x222E},
259 {0x2234, 0x2237}, {0x223C, 0x223D}, {0x2248, 0x2248},
260 {0x224C, 0x224C}, {0x2252, 0x2252}, {0x2260, 0x2261},
261 {0x2264, 0x2267}, {0x226A, 0x226B}, {0x226E, 0x226F},
262 {0x2282, 0x2283}, {0x2286, 0x2287}, {0x2295, 0x2295},
263 {0x2299, 0x2299}, {0x22A5, 0x22A5}, {0x22BF, 0x22BF},
264 {0x2312, 0x2312}, {0x2460, 0x24E9}, {0x24EB, 0x254B},
265 {0x2550, 0x2573}, {0x2580, 0x258F}, {0x2592, 0x2595},
266 {0x25A0, 0x25A1}, {0x25A3, 0x25A9}, {0x25B2, 0x25B3},
267 {0x25B6, 0x25B7}, {0x25BC, 0x25BD}, {0x25C0, 0x25C1},
268 {0x25C6, 0x25C8}, {0x25CB, 0x25CB}, {0x25CE, 0x25D1},
269 {0x25E2, 0x25E5}, {0x25EF, 0x25EF}, {0x2605, 0x2606},
270 {0x2609, 0x2609}, {0x260E, 0x260F}, {0x2614, 0x2615},
271 {0x261C, 0x261C}, {0x261E, 0x261E}, {0x2640, 0x2640},
272 {0x2642, 0x2642}, {0x2660, 0x2661}, {0x2663, 0x2665},
273 {0x2667, 0x266A}, {0x266C, 0x266D}, {0x266F, 0x266F},
274 {0x273D, 0x273D}, {0x2776, 0x277F}, {0xE000, 0xF8FF},
275 {0xFFFD, 0xFFFD}, {0xF0000, 0xFFFFD}, {0x100000, 0x10FFFD}
278 /* binary search in table of non-spacing characters */
279 if (bisearch(ucs, ambiguous,
280 sizeof(ambiguous) / sizeof(struct interval) - 1))
281 return 2;
283 return mk_wcwidth(ucs);
287 int mk_wcswidth_cjk(const wchar_t * pwcs, size_t n)
289 int w, width = 0;
291 for (; *pwcs && n-- > 0; pwcs++)
292 if ((w = mk_wcwidth_cjk(*pwcs)) < 0)
293 return -1;
294 else
295 width += w;
297 return width;