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 -- 2007-05-26 (Unicode 5.0)
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 /* Assigned Unicode character ranges:
64 * 0 0000–FFFF Basic Multilingual Plane
65 * 1 10000–1FFFF Supplementary Multilingual Plane
66 * 2 20000–2FFFF Supplementary Ideographic Plane
67 * 3 30000-3FFFF Tertiary Ideographic Plane (no chars assigned yet)
68 * 4-13 40000–DFFFF currently unassigned
69 * 14 E0000–EFFFF Supplementary Special-purpose Plane
70 * 15 F0000–FFFFF Supplementary Private Use Area-A
71 * 16 100000–10FFFF Supplementary Private Use Area-B
73 * "Supplementary Special-purpose Plane currently contains non-graphical
74 * characters in two blocks of 128 and 240 characters. The first block
75 * is for language tag characters for use when language cannot be indicated
76 * through other protocols (such as the xml:lang attribute in XML).
77 * The other block contains glyph variation selectors to indicate
78 * an alternate glyph for a character that cannot be determined by context."
80 * In simpler terms: it is a tool to fix the "Han unification" mess
81 * created by Unicode committee, to select Chinese/Japanese/Korean/Taiwan
82 * version of a character. (They forgot that the whole purpose of the Unicode
83 * was to be able to write all chars in one charset without such tricks).
84 * Until East Asian users say it is actually necessary to support these
85 * code points in console applications like busybox
86 * (i.e. do these chars ever appear in filenames, hostnames, text files
87 * and such?), we are treating these code points as invalid.
89 * Tertiary Ideographic Plane is also ignored for now,
90 * until Unicode committee assigns something there.
93 #if LAST_SUPPORTED_WCHAR >= 0x300
99 /* auxiliary function for binary search in interval table */
100 static int in_interval_table(unsigned ucs, const struct interval *table, unsigned max)
105 if (ucs < table[0].first || ucs > table[max].last)
110 mid = (min + max) / 2;
111 if (ucs > table[mid].last)
113 else if (ucs < table[mid].first)
121 static int in_uint16_table(unsigned ucs, const uint16_t *table, unsigned max)
125 unsigned first, last;
127 first = table[0] >> 2;
128 last = first + (table[0] & 3);
129 if (ucs < first || ucs > last)
134 mid = (min + max) / 2;
135 first = table[mid] >> 2;
136 last = first + (table[mid] & 3);
139 else if (ucs < first)
149 /* The following two functions define the column width of an ISO 10646
150 * character as follows:
152 * - The null character (U+0000) has a column width of 0.
154 * - Other C0/C1 control characters and DEL will lead to a return
157 * - Non-spacing and enclosing combining characters (general
158 * category code Mn or Me in the Unicode database) have a
161 * - SOFT HYPHEN (U+00AD) has a column width of 1.
163 * - Other format characters (general category code Cf in the Unicode
164 * database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
166 * - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
167 * have a column width of 0.
169 * - Spacing characters in the East Asian Wide (W) or East Asian
170 * Full-width (F) category as defined in Unicode Technical
171 * Report #11 have a column width of 2.
173 * - All remaining characters (including all printable
174 * ISO 8859-1 and WGL4 characters, Unicode control characters,
175 * etc.) have a column width of 1.
177 * This implementation assumes that wchar_t characters are encoded
180 static int wcwidth(unsigned ucs)
182 #if LAST_SUPPORTED_WCHAR >= 0x300
183 /* sorted list of non-overlapping intervals of non-spacing characters */
184 /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
185 static const struct interval combining[] = {
186 #define BIG_(a,b) { a, b },
188 /* PAIR if < 0x4000 and no more than 4 chars big */
310 /* Too big to be packed in PAIRs: */
322 static const uint16_t combining1[] = {
324 #define PAIR(a,b) (a << 2) | (b-a),
325 /* Exact copy-n-paste of the above: */
451 #define BIG_(a,b) char big##a[b-a <= 3 ? -1 : 1];
452 #define PAIR(a,b) char pair##a[b-a > 3 ? -1 : 1];
453 /* Copy-n-paste it here again to verify correctness */
462 /* Test for 8-bit control characters (00-1f, 80-9f, 7f) */
463 if ((ucs & ~0x80) < 0x20 || ucs == 0x7f)
465 /* Quick abort if it is an obviously invalid char */
466 if (ucs > LAST_SUPPORTED_WCHAR)
469 /* Optimization: no combining chars below 0x300 */
470 if (LAST_SUPPORTED_WCHAR < 0x300 || ucs < 0x300)
473 #if LAST_SUPPORTED_WCHAR >= 0x300
474 /* Binary search in table of non-spacing characters */
475 if (in_interval_table(ucs, combining, ARRAY_SIZE(combining) - 1))
477 if (in_uint16_table(ucs, combining1, ARRAY_SIZE(combining1) - 1))
480 /* Optimization: all chars below 0x1100 are not double-width */
481 if (LAST_SUPPORTED_WCHAR < 0x1100 || ucs < 0x1100)
484 # if LAST_SUPPORTED_WCHAR >= 0x1100
485 /* Invalid code points: */
486 /* High (d800..dbff) and low (dc00..dfff) surrogates (valid only in UTF16) */
487 /* Private Use Area (e000..f8ff) */
488 /* Noncharacters fdd0..fdef */
489 if ((LAST_SUPPORTED_WCHAR >= 0xd800 && ucs >= 0xd800 && ucs <= 0xf8ff)
490 || (LAST_SUPPORTED_WCHAR >= 0xfdd0 && ucs >= 0xfdd0 && ucs <= 0xfdef)
494 /* 0xfffe and 0xffff in every plane are invalid */
495 if (LAST_SUPPORTED_WCHAR >= 0xfffe && ((ucs & 0xfffe) == 0xfffe)) {
499 # if LAST_SUPPORTED_WCHAR >= 0x10000
500 if (ucs >= 0x10000) {
501 /* Combining chars in Supplementary Multilingual Plane 0x1xxxx */
502 static const struct interval combining0x10000[] = {
503 { 0x0A01, 0x0A03 }, { 0x0A05, 0x0A06 }, { 0x0A0C, 0x0A0F },
504 { 0x0A38, 0x0A3A }, { 0x0A3F, 0x0A3F }, { 0xD167, 0xD169 },
505 { 0xD173, 0xD182 }, { 0xD185, 0xD18B }, { 0xD1AA, 0xD1AD },
508 /* Binary search in table of non-spacing characters in Supplementary Multilingual Plane */
509 if (in_interval_table(ucs ^ 0x10000, combining0x10000, ARRAY_SIZE(combining0x10000) - 1))
511 /* Check a few non-spacing chars in Supplementary Special-purpose Plane 0xExxxx */
512 if (LAST_SUPPORTED_WCHAR >= 0xE0001
514 || (ucs >= 0xE0020 && ucs <= 0xE007F)
515 || (ucs >= 0xE0100 && ucs <= 0xE01EF)
523 /* If we arrive here, ucs is not a combining or C0/C1 control character.
524 * Check whether it's 1 char or 2-shar wide.
527 ( (/*ucs >= 0x1100 &&*/ ucs <= 0x115f) /* Hangul Jamo init. consonants */
528 || ucs == 0x2329 /* left-pointing angle bracket; also CJK punct. char */
529 || ucs == 0x232a /* right-pointing angle bracket; also CJK punct. char */
530 || (ucs >= 0x2e80 && ucs <= 0xa4cf && ucs != 0x303f) /* CJK ... Yi */
531 # if LAST_SUPPORTED_WCHAR >= 0xac00
532 || (ucs >= 0xac00 && ucs <= 0xd7a3) /* Hangul Syllables */
533 || (ucs >= 0xf900 && ucs <= 0xfaff) /* CJK Compatibility Ideographs */
534 || (ucs >= 0xfe10 && ucs <= 0xfe19) /* Vertical forms */
535 || (ucs >= 0xfe30 && ucs <= 0xfe6f) /* CJK Compatibility Forms */
536 || (ucs >= 0xff00 && ucs <= 0xff60) /* Fullwidth Forms */
537 || (ucs >= 0xffe0 && ucs <= 0xffe6)
538 || ((ucs >> 17) == (2 >> 1)) /* 20000..3ffff: Supplementary and Tertiary Ideographic Planes */
541 # endif /* >= 0x1100 */
542 #endif /* >= 0x300 */