1 #include "wcwidth.h"
2 
3 /*
4  * This is an implementation of wcwidth() and wcswidth() (defined in
5  * IEEE Std 1002.1-2001) for Unicode.
6  *
7  * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
8  * http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html
9  *
10  * In fixed-width output devices, Latin characters all occupy a single
11  * "cell" position of equal width, whereas ideographic CJK characters
12  * occupy two such cells. Interoperability between terminal-line
13  * applications and (teletype-style) character terminals using the
14  * UTF-8 encoding requires agreement on which character should advance
15  * the cursor by how many cell positions. No established formal
16  * standards exist at present on which Unicode character shall occupy
17  * how many cell positions on character terminals. These routines are
18  * a first attempt of defining such behavior based on simple rules
19  * applied to data provided by the Unicode Consortium.
20  *
21  * For some graphical characters, the Unicode standard explicitly
22  * defines a character-cell width via the definition of the East Asian
23  * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
24  * In all these cases, there is no ambiguity about which width a
25  * terminal shall use. For characters in the East Asian Ambiguous (A)
26  * class, the width choice depends purely on a preference of backward
27  * compatibility with either historic CJK or Western practice.
28  * Choosing single-width for these characters is easy to justify as
29  * the appropriate long-term solution, as the CJK practice of
30  * displaying these characters as double-width comes from historic
31  * implementation simplicity (8-bit encoded characters were displayed
32  * single-width and 16-bit ones double-width, even for Greek,
33  * Cyrillic, etc.) and not any typographic considerations.
34  *
35  * Much less clear is the choice of width for the Not East Asian
36  * (Neutral) class. Existing practice does not dictate a width for any
37  * of these characters. It would nevertheless make sense
38  * typographically to allocate two character cells to characters such
39  * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
40  * represented adequately with a single-width glyph. The following
41  * routines at present merely assign a single-cell width to all
42  * neutral characters, in the interest of simplicity. This is not
43  * entirely satisfactory and should be reconsidered before
44  * establishing a formal standard in this area. At the moment, the
45  * decision which Not East Asian (Neutral) characters should be
46  * represented by double-width glyphs cannot yet be answered by
47  * applying a simple rule from the Unicode database content. Setting
48  * up a proper standard for the behavior of UTF-8 character terminals
49  * will require a careful analysis not only of each Unicode character,
50  * but also of each presentation form, something the author of these
51  * routines has avoided to do so far.
52  *
53  * http://www.unicode.org/unicode/reports/tr11/
54  *
55  * Markus Kuhn -- 2007-05-26 (Unicode 5.0)
56  *
57  * Permission to use, copy, modify, and distribute this software
58  * for any purpose and without fee is hereby granted. The author
59  * disclaims all warranties with regard to this software.
60  *
61  * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
62  */
63 
64 #include <cstdint>
65 
66 struct interval {
67     uint32_t first;
68     uint32_t last;
69 };
70 
71 /* auxiliary function for binary search in interval table */
bisearch(uint32_t ucs,const struct interval * table,uint32_t max)72 static uint32_t bisearch( uint32_t ucs, const struct interval *table, uint32_t max )
73 {
74     uint32_t min = 0;
75     uint32_t mid = 0;
76 
77     if( ucs < table[0].first || ucs > table[max].last ) {
78         return 0;
79     }
80     while( max >= min ) {
81         mid = ( min + max ) / 2;
82         if( ucs > table[mid].last ) {
83             min = mid + 1;
84         } else if( ucs < table[mid].first ) {
85             max = mid - 1;
86         } else {
87             return 1;
88         }
89     }
90 
91     return 0;
92 }
93 
mk_wcwidth(uint32_t ucs)94 int mk_wcwidth( uint32_t ucs )
95 {
96     /* sorted list of non-overlapping intervals of non-spacing characters */
97     /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
98     static const struct interval combining[] = {
99         { 0x0300, 0x036F }, { 0x0483, 0x0486 }, { 0x0488, 0x0489 },
100         { 0x0591, 0x05BD }, { 0x05BF, 0x05BF }, { 0x05C1, 0x05C2 },
101         { 0x05C4, 0x05C5 }, { 0x05C7, 0x05C7 }, { 0x0600, 0x0603 },
102         { 0x0610, 0x0615 }, { 0x064B, 0x065E }, { 0x0670, 0x0670 },
103         { 0x06D6, 0x06E4 }, { 0x06E7, 0x06E8 }, { 0x06EA, 0x06ED },
104         { 0x070F, 0x070F }, { 0x0711, 0x0711 }, { 0x0730, 0x074A },
105         { 0x07A6, 0x07B0 }, { 0x07EB, 0x07F3 }, { 0x0901, 0x0902 },
106         { 0x093C, 0x093C }, { 0x0941, 0x0948 }, { 0x094D, 0x094D },
107         { 0x0951, 0x0954 }, { 0x0962, 0x0963 }, { 0x0981, 0x0981 },
108         { 0x09BC, 0x09BC }, { 0x09C1, 0x09C4 }, { 0x09CD, 0x09CD },
109         { 0x09E2, 0x09E3 }, { 0x0A01, 0x0A02 }, { 0x0A3C, 0x0A3C },
110         { 0x0A41, 0x0A42 }, { 0x0A47, 0x0A48 }, { 0x0A4B, 0x0A4D },
111         { 0x0A70, 0x0A71 }, { 0x0A81, 0x0A82 }, { 0x0ABC, 0x0ABC },
112         { 0x0AC1, 0x0AC5 }, { 0x0AC7, 0x0AC8 }, { 0x0ACD, 0x0ACD },
113         { 0x0AE2, 0x0AE3 }, { 0x0B01, 0x0B01 }, { 0x0B3C, 0x0B3C },
114         { 0x0B3F, 0x0B3F }, { 0x0B41, 0x0B43 }, { 0x0B4D, 0x0B4D },
115         { 0x0B56, 0x0B56 }, { 0x0B82, 0x0B82 }, { 0x0BC0, 0x0BC0 },
116         { 0x0BCD, 0x0BCD }, { 0x0C3E, 0x0C40 }, { 0x0C46, 0x0C48 },
117         { 0x0C4A, 0x0C4D }, { 0x0C55, 0x0C56 }, { 0x0CBC, 0x0CBC },
118         { 0x0CBF, 0x0CBF }, { 0x0CC6, 0x0CC6 }, { 0x0CCC, 0x0CCD },
119         { 0x0CE2, 0x0CE3 }, { 0x0D41, 0x0D43 }, { 0x0D4D, 0x0D4D },
120         { 0x0DCA, 0x0DCA }, { 0x0DD2, 0x0DD4 }, { 0x0DD6, 0x0DD6 },
121         { 0x0E31, 0x0E31 }, { 0x0E34, 0x0E3A }, { 0x0E47, 0x0E4E },
122         { 0x0EB1, 0x0EB1 }, { 0x0EB4, 0x0EB9 }, { 0x0EBB, 0x0EBC },
123         { 0x0EC8, 0x0ECD }, { 0x0F18, 0x0F19 }, { 0x0F35, 0x0F35 },
124         { 0x0F37, 0x0F37 }, { 0x0F39, 0x0F39 }, { 0x0F71, 0x0F7E },
125         { 0x0F80, 0x0F84 }, { 0x0F86, 0x0F87 }, { 0x0F90, 0x0F97 },
126         { 0x0F99, 0x0FBC }, { 0x0FC6, 0x0FC6 }, { 0x102D, 0x1030 },
127         { 0x1032, 0x1032 }, { 0x1036, 0x1037 }, { 0x1039, 0x1039 },
128         { 0x1058, 0x1059 }, { 0x1160, 0x11FF }, { 0x135F, 0x135F },
129         { 0x1712, 0x1714 }, { 0x1732, 0x1734 }, { 0x1752, 0x1753 },
130         { 0x1772, 0x1773 }, { 0x17B4, 0x17B5 }, { 0x17B7, 0x17BD },
131         { 0x17C6, 0x17C6 }, { 0x17C9, 0x17D3 }, { 0x17DD, 0x17DD },
132         { 0x180B, 0x180D }, { 0x18A9, 0x18A9 }, { 0x1920, 0x1922 },
133         { 0x1927, 0x1928 }, { 0x1932, 0x1932 }, { 0x1939, 0x193B },
134         { 0x1A17, 0x1A18 }, { 0x1B00, 0x1B03 }, { 0x1B34, 0x1B34 },
135         { 0x1B36, 0x1B3A }, { 0x1B3C, 0x1B3C }, { 0x1B42, 0x1B42 },
136         { 0x1B6B, 0x1B73 }, { 0x1DC0, 0x1DCA }, { 0x1DFE, 0x1DFF },
137         { 0x200B, 0x200F }, { 0x202A, 0x202E }, { 0x2060, 0x2063 },
138         { 0x206A, 0x206F }, { 0x20D0, 0x20EF }, { 0x302A, 0x302F },
139         { 0x3099, 0x309A }, { 0xA806, 0xA806 }, { 0xA80B, 0xA80B },
140         { 0xA825, 0xA826 }, { 0xFB1E, 0xFB1E }, { 0xFE00, 0xFE0F },
141         { 0xFE20, 0xFE23 }, { 0xFEFF, 0xFEFF }, { 0xFFF9, 0xFFFB },
142         { 0x10A01, 0x10A03 }, { 0x10A05, 0x10A06 }, { 0x10A0C, 0x10A0F },
143         { 0x10A38, 0x10A3A }, { 0x10A3F, 0x10A3F }, { 0x1D167, 0x1D169 },
144         { 0x1D173, 0x1D182 }, { 0x1D185, 0x1D18B }, { 0x1D1AA, 0x1D1AD },
145         { 0x1D242, 0x1D244 }, { 0xE0001, 0xE0001 }, { 0xE0020, 0xE007F },
146         { 0xE0100, 0xE01EF }
147     };
148 
149     /* test for 8-bit control characters */
150     if( ucs == 0 ) {
151         return 0;
152     }
153     if( ( ucs < 32 ) || ( ucs >= 0x7f && ucs < 0xa0 ) ) {
154         return -1;
155     }
156 
157     /* binary search in table of non-spacing characters */
158     if( bisearch( ucs, combining,
159                   sizeof( combining ) / sizeof( struct interval ) - 1 ) ) {
160         return 0;
161     }
162 
163     /* if we arrive here, ucs is not a combining or C0/C1 control character */
164 
165     return 1 +
166            ( ucs >= 0x1100 &&
167              ( ucs <= 0x115f ||                   /* Hangul Jamo init. consonants */
168                ucs == 0x2329 || ucs == 0x232a ||
169                ( ucs >= 0x2e80 && ucs <= 0xa4cf &&
170                  ucs != 0x303f ) ||                 /* CJK ... Yi */
171                ( ucs >= 0xac00 && ucs <= 0xd7a3 ) || /* Hangul Syllables */
172                ( ucs >= 0xf900 && ucs <= 0xfaff ) || /* CJK Compatibility Ideographs */
173                ( ucs >= 0xfe10 && ucs <= 0xfe19 ) || /* Vertical forms */
174                ( ucs >= 0xfe30 && ucs <= 0xfe6f ) || /* CJK Compatibility Forms */
175                ( ucs >= 0xff00 && ucs <= 0xff60 ) || /* Fullwidth Forms */
176                ( ucs >= 0xffe0 && ucs <= 0xffe6 ) ||
177                ( ucs >= 0x20000 && ucs <= 0x2fffd ) ||
178                ( ucs >= 0x30000 && ucs <= 0x3fffd ) ) );
179 }
180