source: branches/3.2/mindi-busybox/libbb/unicode.c @ 3232

Last change on this file since 3232 was 3232, checked in by bruno, 5 years ago
  • Update mindi-busybox to 1.21.1
  • Property svn:eol-style set to native
File size: 30.8 KB
Line 
1/* vi: set sw=4 ts=4: */
2/*
3 * Unicode support routines.
4 *
5 * Copyright (C) 2009 Denys Vlasenko
6 *
7 * Licensed under GPLv2, see file LICENSE in this source tree.
8 */
9#include "libbb.h"
10#include "unicode.h"
11
12/* If it's not #defined as a constant in unicode.h... */
13#ifndef unicode_status
14uint8_t unicode_status;
15#endif
16
17/* This file is compiled only if UNICODE_SUPPORT is on.
18 * We check other options and decide whether to use libc support
19 * via locale, or use our own logic:
20 */
21
22#if ENABLE_UNICODE_USING_LOCALE
23
24/* Unicode support using libc locale support. */
25
26void FAST_FUNC reinit_unicode(const char *LANG)
27{
28    static const char unicode_0x394[] = { 0xce, 0x94, 0 };
29    size_t width;
30
31//TODO: avoid repeated calls by caching last string?
32    setlocale(LC_ALL, (LANG && LANG[0]) ? LANG : "C");
33
34    /* In unicode, this is a one character string */
35// can use unicode_strlen(string) too, but otherwise unicode_strlen() is unused
36    width = mbstowcs(NULL, unicode_0x394, INT_MAX);
37    unicode_status = (width == 1 ? UNICODE_ON : UNICODE_OFF);
38}
39
40void FAST_FUNC init_unicode(void)
41{
42    if (unicode_status == UNICODE_UNKNOWN)
43        reinit_unicode(getenv("LANG"));
44}
45
46#else
47
48/* Homegrown Unicode support. It knows only C and Unicode locales. */
49
50# if ENABLE_FEATURE_CHECK_UNICODE_IN_ENV
51void FAST_FUNC reinit_unicode(const char *LANG)
52{
53    unicode_status = UNICODE_OFF;
54    if (!LANG || !(strstr(LANG, ".utf") || strstr(LANG, ".UTF")))
55        return;
56    unicode_status = UNICODE_ON;
57}
58
59void FAST_FUNC init_unicode(void)
60{
61    if (unicode_status == UNICODE_UNKNOWN)
62        reinit_unicode(getenv("LANG"));
63}
64# endif
65
66static size_t wcrtomb_internal(char *s, wchar_t wc)
67{
68    int n, i;
69    uint32_t v = wc;
70
71    if (v <= 0x7f) {
72        *s = v;
73        return 1;
74    }
75
76    /* RFC 3629 says that Unicode ends at 10FFFF,
77     * but we cover entire 32 bits */
78
79    /* 4000000-FFFFFFFF -> 111111tt 10tttttt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
80    /* 200000-3FFFFFF -> 111110tt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
81    /* 10000-1FFFFF -> 11110zzz 10zzyyyy 10yyyyxx 10xxxxxx */
82    /* 800-FFFF -> 1110yyyy 10yyyyxx 10xxxxxx */
83    /* 80-7FF -> 110yyyxx 10xxxxxx */
84
85    /* How many bytes do we need? */
86    n = 2;
87    /* (0x80000000+ would result in n = 7, limiting n to 6) */
88    while (v >= 0x800 && n < 6) {
89        v >>= 5;
90        n++;
91    }
92    /* Fill bytes n-1..1 */
93    i = n;
94    while (--i) {
95        s[i] = (wc & 0x3f) | 0x80;
96        wc >>= 6;
97    }
98    /* Fill byte 0 */
99    s[0] = wc | (uint8_t)(0x3f00 >> n);
100    return n;
101}
102size_t FAST_FUNC wcrtomb(char *s, wchar_t wc, mbstate_t *ps UNUSED_PARAM)
103{
104    if (unicode_status != UNICODE_ON) {
105        *s = wc;
106        return 1;
107    }
108
109    return wcrtomb_internal(s, wc);
110}
111size_t FAST_FUNC wcstombs(char *dest, const wchar_t *src, size_t n)
112{
113    size_t org_n = n;
114
115    if (unicode_status != UNICODE_ON) {
116        while (n) {
117            wchar_t c = *src++;
118            *dest++ = c;
119            if (c == 0)
120                break;
121            n--;
122        }
123        return org_n - n;
124    }
125
126    while (n >= MB_CUR_MAX) {
127        wchar_t wc = *src++;
128        size_t len = wcrtomb_internal(dest, wc);
129
130        if (wc == L'\0')
131            return org_n - n;
132        dest += len;
133        n -= len;
134    }
135    while (n) {
136        char tbuf[MB_CUR_MAX];
137        wchar_t wc = *src++;
138        size_t len = wcrtomb_internal(tbuf, wc);
139
140        if (len > n)
141            break;
142        memcpy(dest, tbuf, len);
143        if (wc == L'\0')
144            return org_n - n;
145        dest += len;
146        n -= len;
147    }
148    return org_n - n;
149}
150
151# define ERROR_WCHAR (~(wchar_t)0)
152
153static const char *mbstowc_internal(wchar_t *res, const char *src)
154{
155    int bytes;
156    unsigned c = (unsigned char) *src++;
157
158    if (c <= 0x7f) {
159        *res = c;
160        return src;
161    }
162
163    /* 80-7FF -> 110yyyxx 10xxxxxx */
164    /* 800-FFFF -> 1110yyyy 10yyyyxx 10xxxxxx */
165    /* 10000-1FFFFF -> 11110zzz 10zzyyyy 10yyyyxx 10xxxxxx */
166    /* 200000-3FFFFFF -> 111110tt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
167    /* 4000000-FFFFFFFF -> 111111tt 10tttttt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
168    bytes = 0;
169    do {
170        c <<= 1;
171        bytes++;
172    } while ((c & 0x80) && bytes < 6);
173    if (bytes == 1) {
174        /* A bare "continuation" byte. Say, 80 */
175        *res = ERROR_WCHAR;
176        return src;
177    }
178    c = (uint8_t)(c) >> bytes;
179
180    while (--bytes) {
181        unsigned ch = (unsigned char) *src;
182        if ((ch & 0xc0) != 0x80) {
183            /* Missing "continuation" byte. Example: e0 80 */
184            *res = ERROR_WCHAR;
185            return src;
186        }
187        c = (c << 6) + (ch & 0x3f);
188        src++;
189    }
190
191    /* TODO */
192    /* Need to check that c isn't produced by overlong encoding */
193    /* Example: 11000000 10000000 converts to NUL */
194    /* 11110000 10000000 10000100 10000000 converts to 0x100 */
195    /* correct encoding: 11000100 10000000 */
196    if (c <= 0x7f) { /* crude check */
197        *res = ERROR_WCHAR;
198        return src;
199    }
200
201    *res = c;
202    return src;
203}
204size_t FAST_FUNC mbstowcs(wchar_t *dest, const char *src, size_t n)
205{
206    size_t org_n = n;
207
208    if (unicode_status != UNICODE_ON) {
209        while (n) {
210            unsigned char c = *src++;
211
212            if (dest)
213                *dest++ = c;
214            if (c == 0)
215                break;
216            n--;
217        }
218        return org_n - n;
219    }
220
221    while (n) {
222        wchar_t wc;
223        src = mbstowc_internal(&wc, src);
224        if (wc == ERROR_WCHAR) /* error */
225            return (size_t) -1L;
226        if (dest)
227            *dest++ = wc;
228        if (wc == 0) /* end-of-string */
229            break;
230        n--;
231    }
232
233    return org_n - n;
234}
235
236int FAST_FUNC iswspace(wint_t wc)
237{
238    return (unsigned)wc <= 0x7f && isspace(wc);
239}
240
241int FAST_FUNC iswalnum(wint_t wc)
242{
243    return (unsigned)wc <= 0x7f && isalnum(wc);
244}
245
246int FAST_FUNC iswpunct(wint_t wc)
247{
248    return (unsigned)wc <= 0x7f && ispunct(wc);
249}
250
251
252# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
253struct interval {
254    uint16_t first;
255    uint16_t last;
256};
257
258/* auxiliary function for binary search in interval table */
259static int in_interval_table(unsigned ucs, const struct interval *table, unsigned max)
260{
261    unsigned min;
262    unsigned mid;
263
264    if (ucs < table[0].first || ucs > table[max].last)
265        return 0;
266
267    min = 0;
268    while (max >= min) {
269        mid = (min + max) / 2;
270        if (ucs > table[mid].last)
271            min = mid + 1;
272        else if (ucs < table[mid].first)
273            max = mid - 1;
274        else
275            return 1;
276    }
277    return 0;
278}
279
280static int in_uint16_table(unsigned ucs, const uint16_t *table, unsigned max)
281{
282    unsigned min;
283    unsigned mid;
284    unsigned first, last;
285
286    first = table[0] >> 2;
287    last = first + (table[0] & 3);
288    if (ucs < first || ucs > last)
289        return 0;
290
291    min = 0;
292    while (max >= min) {
293        mid = (min + max) / 2;
294        first = table[mid] >> 2;
295        last = first + (table[mid] & 3);
296        if (ucs > last)
297            min = mid + 1;
298        else if (ucs < first)
299            max = mid - 1;
300        else
301            return 1;
302    }
303    return 0;
304}
305# endif
306
307
308/*
309 * This is an implementation of wcwidth() and wcswidth() (defined in
310 * IEEE Std 1002.1-2001) for Unicode.
311 *
312 * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
313 * http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html
314 *
315 * In fixed-width output devices, Latin characters all occupy a single
316 * "cell" position of equal width, whereas ideographic CJK characters
317 * occupy two such cells. Interoperability between terminal-line
318 * applications and (teletype-style) character terminals using the
319 * UTF-8 encoding requires agreement on which character should advance
320 * the cursor by how many cell positions. No established formal
321 * standards exist at present on which Unicode character shall occupy
322 * how many cell positions on character terminals. These routines are
323 * a first attempt of defining such behavior based on simple rules
324 * applied to data provided by the Unicode Consortium.
325 *
326 * For some graphical characters, the Unicode standard explicitly
327 * defines a character-cell width via the definition of the East Asian
328 * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
329 * In all these cases, there is no ambiguity about which width a
330 * terminal shall use. For characters in the East Asian Ambiguous (A)
331 * class, the width choice depends purely on a preference of backward
332 * compatibility with either historic CJK or Western practice.
333 * Choosing single-width for these characters is easy to justify as
334 * the appropriate long-term solution, as the CJK practice of
335 * displaying these characters as double-width comes from historic
336 * implementation simplicity (8-bit encoded characters were displayed
337 * single-width and 16-bit ones double-width, even for Greek,
338 * Cyrillic, etc.) and not any typographic considerations.
339 *
340 * Much less clear is the choice of width for the Not East Asian
341 * (Neutral) class. Existing practice does not dictate a width for any
342 * of these characters. It would nevertheless make sense
343 * typographically to allocate two character cells to characters such
344 * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
345 * represented adequately with a single-width glyph. The following
346 * routines at present merely assign a single-cell width to all
347 * neutral characters, in the interest of simplicity. This is not
348 * entirely satisfactory and should be reconsidered before
349 * establishing a formal standard in this area. At the moment, the
350 * decision which Not East Asian (Neutral) characters should be
351 * represented by double-width glyphs cannot yet be answered by
352 * applying a simple rule from the Unicode database content. Setting
353 * up a proper standard for the behavior of UTF-8 character terminals
354 * will require a careful analysis not only of each Unicode character,
355 * but also of each presentation form, something the author of these
356 * routines has avoided to do so far.
357 *
358 * http://www.unicode.org/unicode/reports/tr11/
359 *
360 * Markus Kuhn -- 2007-05-26 (Unicode 5.0)
361 *
362 * Permission to use, copy, modify, and distribute this software
363 * for any purpose and without fee is hereby granted. The author
364 * disclaims all warranties with regard to this software.
365 *
366 * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
367 */
368
369/* Assigned Unicode character ranges:
370 * Plane Range
371 * 0       0000–FFFF   Basic Multilingual Plane
372 * 1      10000–1FFFF  Supplementary Multilingual Plane
373 * 2      20000–2FFFF  Supplementary Ideographic Plane
374 * 3      30000-3FFFF  Tertiary Ideographic Plane (no chars assigned yet)
375 * 4-13   40000–DFFFF  currently unassigned
376 * 14     E0000–EFFFF  Supplementary Special-purpose Plane
377 * 15     F0000–FFFFF  Supplementary Private Use Area-A
378 * 16    100000–10FFFF Supplementary Private Use Area-B
379 *
380 * "Supplementary Special-purpose Plane currently contains non-graphical
381 * characters in two blocks of 128 and 240 characters. The first block
382 * is for language tag characters for use when language cannot be indicated
383 * through other protocols (such as the xml:lang  attribute in XML).
384 * The other block contains glyph variation selectors to indicate
385 * an alternate glyph for a character that cannot be determined by context."
386 *
387 * In simpler terms: it is a tool to fix the "Han unification" mess
388 * created by Unicode committee, to select Chinese/Japanese/Korean/Taiwan
389 * version of a character. (They forgot that the whole purpose of the Unicode
390 * was to be able to write all chars in one charset without such tricks).
391 * Until East Asian users say it is actually necessary to support these
392 * code points in console applications like busybox
393 * (i.e. do these chars ever appear in filenames, hostnames, text files
394 * and such?), we are treating these code points as invalid.
395 *
396 * Tertiary Ideographic Plane is also ignored for now,
397 * until Unicode committee assigns something there.
398 */
399/* The following two functions define the column width of an ISO 10646
400 * character as follows:
401 *
402 *    - The null character (U+0000) has a column width of 0.
403 *
404 *    - Other C0/C1 control characters and DEL will lead to a return
405 *      value of -1.
406 *
407 *    - Non-spacing and enclosing combining characters (general
408 *      category code Mn or Me in the Unicode database) have a
409 *      column width of 0.
410 *
411 *    - SOFT HYPHEN (U+00AD) has a column width of 1.
412 *
413 *    - Other format characters (general category code Cf in the Unicode
414 *      database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
415 *
416 *    - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
417 *      have a column width of 0.
418 *
419 *    - Spacing characters in the East Asian Wide (W) or East Asian
420 *      Full-width (F) category as defined in Unicode Technical
421 *      Report #11 have a column width of 2.
422 *
423 *    - All remaining characters (including all printable
424 *      ISO 8859-1 and WGL4 characters, Unicode control characters,
425 *      etc.) have a column width of 1.
426 *
427 * This implementation assumes that wchar_t characters are encoded
428 * in ISO 10646.
429 */
430int FAST_FUNC wcwidth(unsigned ucs)
431{
432# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
433    /* sorted list of non-overlapping intervals of non-spacing characters */
434    /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
435#  define BIG_(a,b) { a, b },
436#  define PAIR(a,b)
437#  define ARRAY /* PAIR if < 0x4000 and no more than 4 chars big */ \
438        BIG_(0x0300, 0x036F) \
439        PAIR(0x0483, 0x0486) \
440        PAIR(0x0488, 0x0489) \
441        BIG_(0x0591, 0x05BD) \
442        PAIR(0x05BF, 0x05BF) \
443        PAIR(0x05C1, 0x05C2) \
444        PAIR(0x05C4, 0x05C5) \
445        PAIR(0x05C7, 0x05C7) \
446        PAIR(0x0600, 0x0603) \
447        BIG_(0x0610, 0x0615) \
448        BIG_(0x064B, 0x065E) \
449        PAIR(0x0670, 0x0670) \
450        BIG_(0x06D6, 0x06E4) \
451        PAIR(0x06E7, 0x06E8) \
452        PAIR(0x06EA, 0x06ED) \
453        PAIR(0x070F, 0x070F) \
454        PAIR(0x0711, 0x0711) \
455        BIG_(0x0730, 0x074A) \
456        BIG_(0x07A6, 0x07B0) \
457        BIG_(0x07EB, 0x07F3) \
458        PAIR(0x0901, 0x0902) \
459        PAIR(0x093C, 0x093C) \
460        BIG_(0x0941, 0x0948) \
461        PAIR(0x094D, 0x094D) \
462        PAIR(0x0951, 0x0954) \
463        PAIR(0x0962, 0x0963) \
464        PAIR(0x0981, 0x0981) \
465        PAIR(0x09BC, 0x09BC) \
466        PAIR(0x09C1, 0x09C4) \
467        PAIR(0x09CD, 0x09CD) \
468        PAIR(0x09E2, 0x09E3) \
469        PAIR(0x0A01, 0x0A02) \
470        PAIR(0x0A3C, 0x0A3C) \
471        PAIR(0x0A41, 0x0A42) \
472        PAIR(0x0A47, 0x0A48) \
473        PAIR(0x0A4B, 0x0A4D) \
474        PAIR(0x0A70, 0x0A71) \
475        PAIR(0x0A81, 0x0A82) \
476        PAIR(0x0ABC, 0x0ABC) \
477        BIG_(0x0AC1, 0x0AC5) \
478        PAIR(0x0AC7, 0x0AC8) \
479        PAIR(0x0ACD, 0x0ACD) \
480        PAIR(0x0AE2, 0x0AE3) \
481        PAIR(0x0B01, 0x0B01) \
482        PAIR(0x0B3C, 0x0B3C) \
483        PAIR(0x0B3F, 0x0B3F) \
484        PAIR(0x0B41, 0x0B43) \
485        PAIR(0x0B4D, 0x0B4D) \
486        PAIR(0x0B56, 0x0B56) \
487        PAIR(0x0B82, 0x0B82) \
488        PAIR(0x0BC0, 0x0BC0) \
489        PAIR(0x0BCD, 0x0BCD) \
490        PAIR(0x0C3E, 0x0C40) \
491        PAIR(0x0C46, 0x0C48) \
492        PAIR(0x0C4A, 0x0C4D) \
493        PAIR(0x0C55, 0x0C56) \
494        PAIR(0x0CBC, 0x0CBC) \
495        PAIR(0x0CBF, 0x0CBF) \
496        PAIR(0x0CC6, 0x0CC6) \
497        PAIR(0x0CCC, 0x0CCD) \
498        PAIR(0x0CE2, 0x0CE3) \
499        PAIR(0x0D41, 0x0D43) \
500        PAIR(0x0D4D, 0x0D4D) \
501        PAIR(0x0DCA, 0x0DCA) \
502        PAIR(0x0DD2, 0x0DD4) \
503        PAIR(0x0DD6, 0x0DD6) \
504        PAIR(0x0E31, 0x0E31) \
505        BIG_(0x0E34, 0x0E3A) \
506        BIG_(0x0E47, 0x0E4E) \
507        PAIR(0x0EB1, 0x0EB1) \
508        BIG_(0x0EB4, 0x0EB9) \
509        PAIR(0x0EBB, 0x0EBC) \
510        BIG_(0x0EC8, 0x0ECD) \
511        PAIR(0x0F18, 0x0F19) \
512        PAIR(0x0F35, 0x0F35) \
513        PAIR(0x0F37, 0x0F37) \
514        PAIR(0x0F39, 0x0F39) \
515        BIG_(0x0F71, 0x0F7E) \
516        BIG_(0x0F80, 0x0F84) \
517        PAIR(0x0F86, 0x0F87) \
518        PAIR(0x0FC6, 0x0FC6) \
519        BIG_(0x0F90, 0x0F97) \
520        BIG_(0x0F99, 0x0FBC) \
521        PAIR(0x102D, 0x1030) \
522        PAIR(0x1032, 0x1032) \
523        PAIR(0x1036, 0x1037) \
524        PAIR(0x1039, 0x1039) \
525        PAIR(0x1058, 0x1059) \
526        BIG_(0x1160, 0x11FF) \
527        PAIR(0x135F, 0x135F) \
528        PAIR(0x1712, 0x1714) \
529        PAIR(0x1732, 0x1734) \
530        PAIR(0x1752, 0x1753) \
531        PAIR(0x1772, 0x1773) \
532        PAIR(0x17B4, 0x17B5) \
533        BIG_(0x17B7, 0x17BD) \
534        PAIR(0x17C6, 0x17C6) \
535        BIG_(0x17C9, 0x17D3) \
536        PAIR(0x17DD, 0x17DD) \
537        PAIR(0x180B, 0x180D) \
538        PAIR(0x18A9, 0x18A9) \
539        PAIR(0x1920, 0x1922) \
540        PAIR(0x1927, 0x1928) \
541        PAIR(0x1932, 0x1932) \
542        PAIR(0x1939, 0x193B) \
543        PAIR(0x1A17, 0x1A18) \
544        PAIR(0x1B00, 0x1B03) \
545        PAIR(0x1B34, 0x1B34) \
546        BIG_(0x1B36, 0x1B3A) \
547        PAIR(0x1B3C, 0x1B3C) \
548        PAIR(0x1B42, 0x1B42) \
549        BIG_(0x1B6B, 0x1B73) \
550        BIG_(0x1DC0, 0x1DCA) \
551        PAIR(0x1DFE, 0x1DFF) \
552        BIG_(0x200B, 0x200F) \
553        BIG_(0x202A, 0x202E) \
554        PAIR(0x2060, 0x2063) \
555        BIG_(0x206A, 0x206F) \
556        BIG_(0x20D0, 0x20EF) \
557        BIG_(0x302A, 0x302F) \
558        PAIR(0x3099, 0x309A) \
559        /* Too big to be packed in PAIRs: */ \
560        BIG_(0xA806, 0xA806) \
561        BIG_(0xA80B, 0xA80B) \
562        BIG_(0xA825, 0xA826) \
563        BIG_(0xFB1E, 0xFB1E) \
564        BIG_(0xFE00, 0xFE0F) \
565        BIG_(0xFE20, 0xFE23) \
566        BIG_(0xFEFF, 0xFEFF) \
567        BIG_(0xFFF9, 0xFFFB)
568    static const struct interval combining[] = { ARRAY };
569#  undef BIG_
570#  undef PAIR
571#  define BIG_(a,b)
572#  define PAIR(a,b) (a << 2) | (b-a),
573    static const uint16_t combining1[] = { ARRAY };
574#  undef BIG_
575#  undef PAIR
576#  define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
577#  define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
578    struct CHECK { ARRAY };
579#  undef BIG_
580#  undef PAIR
581#  undef ARRAY
582# endif
583
584    if (ucs == 0)
585        return 0;
586
587    /* Test for 8-bit control characters (00-1f, 80-9f, 7f) */
588    if ((ucs & ~0x80) < 0x20 || ucs == 0x7f)
589        return -1;
590    /* Quick abort if it is an obviously invalid char */
591    if (ucs > CONFIG_LAST_SUPPORTED_WCHAR)
592        return -1;
593
594    /* Optimization: no combining chars below 0x300 */
595    if (CONFIG_LAST_SUPPORTED_WCHAR < 0x300 || ucs < 0x300)
596        return 1;
597
598# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
599    /* Binary search in table of non-spacing characters */
600    if (in_interval_table(ucs, combining, ARRAY_SIZE(combining) - 1))
601        return 0;
602    if (in_uint16_table(ucs, combining1, ARRAY_SIZE(combining1) - 1))
603        return 0;
604
605    /* Optimization: all chars below 0x1100 are not double-width */
606    if (CONFIG_LAST_SUPPORTED_WCHAR < 0x1100 || ucs < 0x1100)
607        return 1;
608
609#  if CONFIG_LAST_SUPPORTED_WCHAR >= 0x1100
610    /* Invalid code points: */
611    /* High (d800..dbff) and low (dc00..dfff) surrogates (valid only in UTF16) */
612    /* Private Use Area (e000..f8ff) */
613    /* Noncharacters fdd0..fdef */
614    if ((CONFIG_LAST_SUPPORTED_WCHAR >= 0xd800 && ucs >= 0xd800 && ucs <= 0xf8ff)
615     || (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfdd0 && ucs >= 0xfdd0 && ucs <= 0xfdef)
616    ) {
617        return -1;
618    }
619    /* 0xfffe and 0xffff in every plane are invalid */
620    if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfffe && ((ucs & 0xfffe) == 0xfffe)) {
621        return -1;
622    }
623
624#   if CONFIG_LAST_SUPPORTED_WCHAR >= 0x10000
625    if (ucs >= 0x10000) {
626        /* Combining chars in Supplementary Multilingual Plane 0x1xxxx */
627        static const struct interval combining0x10000[] = {
628            { 0x0A01, 0x0A03 }, { 0x0A05, 0x0A06 }, { 0x0A0C, 0x0A0F },
629            { 0x0A38, 0x0A3A }, { 0x0A3F, 0x0A3F }, { 0xD167, 0xD169 },
630            { 0xD173, 0xD182 }, { 0xD185, 0xD18B }, { 0xD1AA, 0xD1AD },
631            { 0xD242, 0xD244 }
632        };
633        /* Binary search in table of non-spacing characters in Supplementary Multilingual Plane */
634        if (in_interval_table(ucs ^ 0x10000, combining0x10000, ARRAY_SIZE(combining0x10000) - 1))
635            return 0;
636        /* Check a few non-spacing chars in Supplementary Special-purpose Plane 0xExxxx */
637        if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xE0001
638         && (  ucs == 0xE0001
639            || (ucs >= 0xE0020 && ucs <= 0xE007F)
640            || (ucs >= 0xE0100 && ucs <= 0xE01EF)
641            )
642        ) {
643            return 0;
644        }
645    }
646#   endif
647
648    /* If we arrive here, ucs is not a combining or C0/C1 control character.
649     * Check whether it's 1 char or 2-shar wide.
650     */
651    return 1 +
652        (  (/*ucs >= 0x1100 &&*/ ucs <= 0x115f) /* Hangul Jamo init. consonants */
653        || ucs == 0x2329 /* left-pointing angle bracket; also CJK punct. char */
654        || ucs == 0x232a /* right-pointing angle bracket; also CJK punct. char */
655        || (ucs >= 0x2e80 && ucs <= 0xa4cf && ucs != 0x303f) /* CJK ... Yi */
656#   if CONFIG_LAST_SUPPORTED_WCHAR >= 0xac00
657        || (ucs >= 0xac00 && ucs <= 0xd7a3) /* Hangul Syllables */
658        || (ucs >= 0xf900 && ucs <= 0xfaff) /* CJK Compatibility Ideographs */
659        || (ucs >= 0xfe10 && ucs <= 0xfe19) /* Vertical forms */
660        || (ucs >= 0xfe30 && ucs <= 0xfe6f) /* CJK Compatibility Forms */
661        || (ucs >= 0xff00 && ucs <= 0xff60) /* Fullwidth Forms */
662        || (ucs >= 0xffe0 && ucs <= 0xffe6)
663        || ((ucs >> 17) == (2 >> 1)) /* 20000..3ffff: Supplementary and Tertiary Ideographic Planes */
664#   endif
665        );
666#  endif /* >= 0x1100 */
667# endif /* >= 0x300 */
668}
669
670
671# if ENABLE_UNICODE_BIDI_SUPPORT
672int FAST_FUNC unicode_bidi_isrtl(wint_t wc)
673{
674    /* ranges taken from
675     * http://www.unicode.org/Public/5.2.0/ucd/extracted/DerivedBidiClass.txt
676     * Bidi_Class=Left_To_Right | Bidi_Class=Arabic_Letter
677     */
678#  define BIG_(a,b) { a, b },
679#  define PAIR(a,b)
680#  define ARRAY \
681        PAIR(0x0590, 0x0590) \
682        PAIR(0x05BE, 0x05BE) \
683        PAIR(0x05C0, 0x05C0) \
684        PAIR(0x05C3, 0x05C3) \
685        PAIR(0x05C6, 0x05C6) \
686        BIG_(0x05C8, 0x05FF) \
687        PAIR(0x0604, 0x0605) \
688        PAIR(0x0608, 0x0608) \
689        PAIR(0x060B, 0x060B) \
690        PAIR(0x060D, 0x060D) \
691        BIG_(0x061B, 0x064A) \
692        PAIR(0x065F, 0x065F) \
693        PAIR(0x066D, 0x066F) \
694        BIG_(0x0671, 0x06D5) \
695        PAIR(0x06E5, 0x06E6) \
696        PAIR(0x06EE, 0x06EF) \
697        BIG_(0x06FA, 0x070E) \
698        PAIR(0x0710, 0x0710) \
699        BIG_(0x0712, 0x072F) \
700        BIG_(0x074B, 0x07A5) \
701        BIG_(0x07B1, 0x07EA) \
702        PAIR(0x07F4, 0x07F5) \
703        BIG_(0x07FA, 0x0815) \
704        PAIR(0x081A, 0x081A) \
705        PAIR(0x0824, 0x0824) \
706        PAIR(0x0828, 0x0828) \
707        BIG_(0x082E, 0x08FF) \
708        PAIR(0x200F, 0x200F) \
709        PAIR(0x202B, 0x202B) \
710        PAIR(0x202E, 0x202E) \
711        BIG_(0xFB1D, 0xFB1D) \
712        BIG_(0xFB1F, 0xFB28) \
713        BIG_(0xFB2A, 0xFD3D) \
714        BIG_(0xFD40, 0xFDCF) \
715        BIG_(0xFDC8, 0xFDCF) \
716        BIG_(0xFDF0, 0xFDFC) \
717        BIG_(0xFDFE, 0xFDFF) \
718        BIG_(0xFE70, 0xFEFE)
719        /* Probably not necessary
720        {0x10800, 0x1091E},
721        {0x10920, 0x10A00},
722        {0x10A04, 0x10A04},
723        {0x10A07, 0x10A0B},
724        {0x10A10, 0x10A37},
725        {0x10A3B, 0x10A3E},
726        {0x10A40, 0x10A7F},
727        {0x10B36, 0x10B38},
728        {0x10B40, 0x10E5F},
729        {0x10E7F, 0x10FFF},
730        {0x1E800, 0x1EFFF}
731        */
732    static const struct interval rtl_b[] = { ARRAY };
733#  undef BIG_
734#  undef PAIR
735#  define BIG_(a,b)
736#  define PAIR(a,b) (a << 2) | (b-a),
737    static const uint16_t rtl_p[] = { ARRAY };
738#  undef BIG_
739#  undef PAIR
740#  define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
741#  define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
742    struct CHECK { ARRAY };
743#  undef BIG_
744#  undef PAIR
745#  undef ARRAY
746
747    if (in_interval_table(wc, rtl_b, ARRAY_SIZE(rtl_b) - 1))
748        return 1;
749    if (in_uint16_table(wc, rtl_p, ARRAY_SIZE(rtl_p) - 1))
750        return 1;
751    return 0;
752}
753
754#  if ENABLE_UNICODE_NEUTRAL_TABLE
755int FAST_FUNC unicode_bidi_is_neutral_wchar(wint_t wc)
756{
757    /* ranges taken from
758     * http://www.unicode.org/Public/5.2.0/ucd/extracted/DerivedBidiClass.txt
759     * Bidi_Classes: Paragraph_Separator, Segment_Separator,
760     * White_Space, Other_Neutral, European_Number, European_Separator,
761     * European_Terminator, Arabic_Number, Common_Separator
762     */
763#  define BIG_(a,b) { a, b },
764#  define PAIR(a,b)
765#  define ARRAY \
766        BIG_(0x0009, 0x000D) \
767        BIG_(0x001C, 0x0040) \
768        BIG_(0x005B, 0x0060) \
769        PAIR(0x007B, 0x007E) \
770        PAIR(0x0085, 0x0085) \
771        BIG_(0x00A0, 0x00A9) \
772        PAIR(0x00AB, 0x00AC) \
773        BIG_(0x00AE, 0x00B4) \
774        PAIR(0x00B6, 0x00B9) \
775        BIG_(0x00BB, 0x00BF) \
776        PAIR(0x00D7, 0x00D7) \
777        PAIR(0x00F7, 0x00F7) \
778        PAIR(0x02B9, 0x02BA) \
779        BIG_(0x02C2, 0x02CF) \
780        BIG_(0x02D2, 0x02DF) \
781        BIG_(0x02E5, 0x02FF) \
782        PAIR(0x0374, 0x0375) \
783        PAIR(0x037E, 0x037E) \
784        PAIR(0x0384, 0x0385) \
785        PAIR(0x0387, 0x0387) \
786        PAIR(0x03F6, 0x03F6) \
787        PAIR(0x058A, 0x058A) \
788        PAIR(0x0600, 0x0603) \
789        PAIR(0x0606, 0x0607) \
790        PAIR(0x0609, 0x060A) \
791        PAIR(0x060C, 0x060C) \
792        PAIR(0x060E, 0x060F) \
793        BIG_(0x0660, 0x066C) \
794        PAIR(0x06DD, 0x06DD) \
795        PAIR(0x06E9, 0x06E9) \
796        BIG_(0x06F0, 0x06F9) \
797        PAIR(0x07F6, 0x07F9) \
798        PAIR(0x09F2, 0x09F3) \
799        PAIR(0x09FB, 0x09FB) \
800        PAIR(0x0AF1, 0x0AF1) \
801        BIG_(0x0BF3, 0x0BFA) \
802        BIG_(0x0C78, 0x0C7E) \
803        PAIR(0x0CF1, 0x0CF2) \
804        PAIR(0x0E3F, 0x0E3F) \
805        PAIR(0x0F3A, 0x0F3D) \
806        BIG_(0x1390, 0x1400) \
807        PAIR(0x1680, 0x1680) \
808        PAIR(0x169B, 0x169C) \
809        PAIR(0x17DB, 0x17DB) \
810        BIG_(0x17F0, 0x17F9) \
811        BIG_(0x1800, 0x180A) \
812        PAIR(0x180E, 0x180E) \
813        PAIR(0x1940, 0x1940) \
814        PAIR(0x1944, 0x1945) \
815        BIG_(0x19DE, 0x19FF) \
816        PAIR(0x1FBD, 0x1FBD) \
817        PAIR(0x1FBF, 0x1FC1) \
818        PAIR(0x1FCD, 0x1FCF) \
819        PAIR(0x1FDD, 0x1FDF) \
820        PAIR(0x1FED, 0x1FEF) \
821        PAIR(0x1FFD, 0x1FFE) \
822        BIG_(0x2000, 0x200A) \
823        BIG_(0x2010, 0x2029) \
824        BIG_(0x202F, 0x205F) \
825        PAIR(0x2070, 0x2070) \
826        BIG_(0x2074, 0x207E) \
827        BIG_(0x2080, 0x208E) \
828        BIG_(0x20A0, 0x20B8) \
829        PAIR(0x2100, 0x2101) \
830        PAIR(0x2103, 0x2106) \
831        PAIR(0x2108, 0x2109) \
832        PAIR(0x2114, 0x2114) \
833        PAIR(0x2116, 0x2118) \
834        BIG_(0x211E, 0x2123) \
835        PAIR(0x2125, 0x2125) \
836        PAIR(0x2127, 0x2127) \
837        PAIR(0x2129, 0x2129) \
838        PAIR(0x212E, 0x212E) \
839        PAIR(0x213A, 0x213B) \
840        BIG_(0x2140, 0x2144) \
841        PAIR(0x214A, 0x214D) \
842        BIG_(0x2150, 0x215F) \
843        PAIR(0x2189, 0x2189) \
844        BIG_(0x2190, 0x2335) \
845        BIG_(0x237B, 0x2394) \
846        BIG_(0x2396, 0x23E8) \
847        BIG_(0x2400, 0x2426) \
848        BIG_(0x2440, 0x244A) \
849        BIG_(0x2460, 0x249B) \
850        BIG_(0x24EA, 0x26AB) \
851        BIG_(0x26AD, 0x26CD) \
852        BIG_(0x26CF, 0x26E1) \
853        PAIR(0x26E3, 0x26E3) \
854        BIG_(0x26E8, 0x26FF) \
855        PAIR(0x2701, 0x2704) \
856        PAIR(0x2706, 0x2709) \
857        BIG_(0x270C, 0x2727) \
858        BIG_(0x2729, 0x274B) \
859        PAIR(0x274D, 0x274D) \
860        PAIR(0x274F, 0x2752) \
861        BIG_(0x2756, 0x275E) \
862        BIG_(0x2761, 0x2794) \
863        BIG_(0x2798, 0x27AF) \
864        BIG_(0x27B1, 0x27BE) \
865        BIG_(0x27C0, 0x27CA) \
866        PAIR(0x27CC, 0x27CC) \
867        BIG_(0x27D0, 0x27FF) \
868        BIG_(0x2900, 0x2B4C) \
869        BIG_(0x2B50, 0x2B59) \
870        BIG_(0x2CE5, 0x2CEA) \
871        BIG_(0x2CF9, 0x2CFF) \
872        BIG_(0x2E00, 0x2E99) \
873        BIG_(0x2E9B, 0x2EF3) \
874        BIG_(0x2F00, 0x2FD5) \
875        BIG_(0x2FF0, 0x2FFB) \
876        BIG_(0x3000, 0x3004) \
877        BIG_(0x3008, 0x3020) \
878        PAIR(0x3030, 0x3030) \
879        PAIR(0x3036, 0x3037) \
880        PAIR(0x303D, 0x303D) \
881        PAIR(0x303E, 0x303F) \
882        PAIR(0x309B, 0x309C) \
883        PAIR(0x30A0, 0x30A0) \
884        PAIR(0x30FB, 0x30FB) \
885        BIG_(0x31C0, 0x31E3) \
886        PAIR(0x321D, 0x321E) \
887        BIG_(0x3250, 0x325F) \
888        PAIR(0x327C, 0x327E) \
889        BIG_(0x32B1, 0x32BF) \
890        PAIR(0x32CC, 0x32CF) \
891        PAIR(0x3377, 0x337A) \
892        PAIR(0x33DE, 0x33DF) \
893        PAIR(0x33FF, 0x33FF) \
894        BIG_(0x4DC0, 0x4DFF) \
895        BIG_(0xA490, 0xA4C6) \
896        BIG_(0xA60D, 0xA60F) \
897        BIG_(0xA673, 0xA673) \
898        BIG_(0xA67E, 0xA67F) \
899        BIG_(0xA700, 0xA721) \
900        BIG_(0xA788, 0xA788) \
901        BIG_(0xA828, 0xA82B) \
902        BIG_(0xA838, 0xA839) \
903        BIG_(0xA874, 0xA877) \
904        BIG_(0xFB29, 0xFB29) \
905        BIG_(0xFD3E, 0xFD3F) \
906        BIG_(0xFDFD, 0xFDFD) \
907        BIG_(0xFE10, 0xFE19) \
908        BIG_(0xFE30, 0xFE52) \
909        BIG_(0xFE54, 0xFE66) \
910        BIG_(0xFE68, 0xFE6B) \
911        BIG_(0xFF01, 0xFF20) \
912        BIG_(0xFF3B, 0xFF40) \
913        BIG_(0xFF5B, 0xFF65) \
914        BIG_(0xFFE0, 0xFFE6) \
915        BIG_(0xFFE8, 0xFFEE) \
916        BIG_(0xFFF9, 0xFFFD)
917        /*
918        {0x10101, 0x10101},
919        {0x10140, 0x1019B},
920        {0x1091F, 0x1091F},
921        {0x10B39, 0x10B3F},
922        {0x10E60, 0x10E7E},
923        {0x1D200, 0x1D241},
924        {0x1D245, 0x1D245},
925        {0x1D300, 0x1D356},
926        {0x1D6DB, 0x1D6DB},
927        {0x1D715, 0x1D715},
928        {0x1D74F, 0x1D74F},
929        {0x1D789, 0x1D789},
930        {0x1D7C3, 0x1D7C3},
931        {0x1D7CE, 0x1D7FF},
932        {0x1F000, 0x1F02B},
933        {0x1F030, 0x1F093},
934        {0x1F100, 0x1F10A}
935        */
936    static const struct interval neutral_b[] = { ARRAY };
937#  undef BIG_
938#  undef PAIR
939#  define BIG_(a,b)
940#  define PAIR(a,b) (a << 2) | (b-a),
941    static const uint16_t neutral_p[] = { ARRAY };
942#  undef BIG_
943#  undef PAIR
944#  define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
945#  define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
946    struct CHECK { ARRAY };
947#  undef BIG_
948#  undef PAIR
949#  undef ARRAY
950
951    if (in_interval_table(wc, neutral_b, ARRAY_SIZE(neutral_b) - 1))
952        return 1;
953    if (in_uint16_table(wc, neutral_p, ARRAY_SIZE(neutral_p) - 1))
954        return 1;
955    return 0;
956}
957#  endif
958
959# endif /* UNICODE_BIDI_SUPPORT */
960
961#endif /* Homegrown Unicode support */
962
963
964/* The rest is mostly same for libc and for "homegrown" support */
965
966#if 0 // UNUSED
967size_t FAST_FUNC unicode_strlen(const char *string)
968{
969    size_t width = mbstowcs(NULL, string, INT_MAX);
970    if (width == (size_t)-1L)
971        return strlen(string);
972    return width;
973}
974#endif
975
976size_t FAST_FUNC unicode_strwidth(const char *string)
977{
978    uni_stat_t uni_stat;
979    printable_string(&uni_stat, string);
980    return uni_stat.unicode_width;
981}
982
983static char* FAST_FUNC unicode_conv_to_printable2(uni_stat_t *stats, const char *src, unsigned width, int flags)
984{
985    char *dst;
986    unsigned dst_len;
987    unsigned uni_count;
988    unsigned uni_width;
989
990    if (unicode_status != UNICODE_ON) {
991        char *d;
992        if (flags & UNI_FLAG_PAD) {
993            d = dst = xmalloc(width + 1);
994            while ((int)--width >= 0) {
995                unsigned char c = *src;
996                if (c == '\0') {
997                    do
998                        *d++ = ' ';
999                    while ((int)--width >= 0);
1000                    break;
1001                }
1002                *d++ = (c >= ' ' && c < 0x7f) ? c : '?';
1003                src++;
1004            }
1005            *d = '\0';
1006        } else {
1007            d = dst = xstrndup(src, width);
1008            while (*d) {
1009                unsigned char c = *d;
1010                if (c < ' ' || c >= 0x7f)
1011                    *d = '?';
1012                d++;
1013            }
1014        }
1015        if (stats) {
1016            stats->byte_count = (d - dst);
1017            stats->unicode_count = (d - dst);
1018            stats->unicode_width = (d - dst);
1019        }
1020        return dst;
1021    }
1022
1023    dst = NULL;
1024    uni_count = uni_width = 0;
1025    dst_len = 0;
1026    while (1) {
1027        int w;
1028        wchar_t wc;
1029
1030#if ENABLE_UNICODE_USING_LOCALE
1031        {
1032            mbstate_t mbst = { 0 };
1033            ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst);
1034            /* If invalid sequence is seen: -1 is returned,
1035             * src points to the invalid sequence, errno = EILSEQ.
1036             * Else number of wchars (excluding terminating L'\0')
1037             * written to dest is returned.
1038             * If len (here: 1) non-L'\0' wchars stored at dest,
1039             * src points to the next char to be converted.
1040             * If string is completely converted: src = NULL.
1041             */
1042            if (rc == 0) /* end-of-string */
1043                break;
1044            if (rc < 0) { /* error */
1045                src++;
1046                goto subst;
1047            }
1048            if (!iswprint(wc))
1049                goto subst;
1050        }
1051#else
1052        src = mbstowc_internal(&wc, src);
1053        /* src is advanced to next mb char
1054         * wc == ERROR_WCHAR: invalid sequence is seen
1055         * else: wc is set
1056         */
1057        if (wc == ERROR_WCHAR) /* error */
1058            goto subst;
1059        if (wc == 0) /* end-of-string */
1060            break;
1061#endif
1062        if (CONFIG_LAST_SUPPORTED_WCHAR && wc > CONFIG_LAST_SUPPORTED_WCHAR)
1063            goto subst;
1064        w = wcwidth(wc);
1065        if ((ENABLE_UNICODE_COMBINING_WCHARS && w < 0) /* non-printable wchar */
1066         || (!ENABLE_UNICODE_COMBINING_WCHARS && w <= 0)
1067         || (!ENABLE_UNICODE_WIDE_WCHARS && w > 1)
1068        ) {
1069 subst:
1070            wc = CONFIG_SUBST_WCHAR;
1071            w = 1;
1072        }
1073        width -= w;
1074        /* Note: if width == 0, we still may add more chars,
1075         * they may be zero-width or combining ones */
1076        if ((int)width < 0) {
1077            /* can't add this wc, string would become longer than width */
1078            width += w;
1079            break;
1080        }
1081
1082        uni_count++;
1083        uni_width += w;
1084        dst = xrealloc(dst, dst_len + MB_CUR_MAX);
1085#if ENABLE_UNICODE_USING_LOCALE
1086        {
1087            mbstate_t mbst = { 0 };
1088            dst_len += wcrtomb(&dst[dst_len], wc, &mbst);
1089        }
1090#else
1091        dst_len += wcrtomb_internal(&dst[dst_len], wc);
1092#endif
1093    }
1094
1095    /* Pad to remaining width */
1096    if (flags & UNI_FLAG_PAD) {
1097        dst = xrealloc(dst, dst_len + width + 1);
1098        uni_count += width;
1099        uni_width += width;
1100        while ((int)--width >= 0) {
1101            dst[dst_len++] = ' ';
1102        }
1103    }
1104    dst[dst_len] = '\0';
1105    if (stats) {
1106        stats->byte_count = dst_len;
1107        stats->unicode_count = uni_count;
1108        stats->unicode_width = uni_width;
1109    }
1110
1111    return dst;
1112}
1113char* FAST_FUNC unicode_conv_to_printable(uni_stat_t *stats, const char *src)
1114{
1115    return unicode_conv_to_printable2(stats, src, INT_MAX, 0);
1116}
1117char* FAST_FUNC unicode_conv_to_printable_fixedwidth(/*uni_stat_t *stats,*/ const char *src, unsigned width)
1118{
1119    return unicode_conv_to_printable2(/*stats:*/ NULL, src, width, UNI_FLAG_PAD);
1120}
1121
1122#ifdef UNUSED
1123char* FAST_FUNC unicode_conv_to_printable_maxwidth(uni_stat_t *stats, const char *src, unsigned maxwidth)
1124{
1125    return unicode_conv_to_printable2(stats, src, maxwidth, 0);
1126}
1127
1128unsigned FAST_FUNC unicode_padding_to_width(unsigned width, const char *src)
1129{
1130    if (unicode_status != UNICODE_ON) {
1131        return width - strnlen(src, width);
1132    }
1133
1134    while (1) {
1135        int w;
1136        wchar_t wc;
1137
1138#if ENABLE_UNICODE_USING_LOCALE
1139        {
1140            mbstate_t mbst = { 0 };
1141            ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst);
1142            if (rc <= 0) /* error, or end-of-string */
1143                return width;
1144        }
1145#else
1146        src = mbstowc_internal(&wc, src);
1147        if (wc == ERROR_WCHAR || wc == 0) /* error, or end-of-string */
1148            return width;
1149#endif
1150        w = wcwidth(wc);
1151        if (w < 0) /* non-printable wchar */
1152            return width;
1153        width -= w;
1154        if ((int)width <= 0) /* string is longer than width */
1155            return 0;
1156    }
1157}
1158#endif
Note: See TracBrowser for help on using the repository browser.