1
/* vi: set sw=4 ts=4: */
3
* Unicode support routines.
5
* Copyright (C) 2009 Denys Vlasenko
7
* Licensed under GPLv2, see file LICENSE in this source tree.
12
/* If it's not #defined as a constant in unicode.h... */
13
#ifndef unicode_status
14
uint8_t unicode_status;
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:
22
#if ENABLE_UNICODE_USING_LOCALE
24
/* Unicode support using libc locale support. */
26
void FAST_FUNC reinit_unicode(const char *LANG)
28
static const char unicode_0x394[] = { 0xce, 0x94, 0 };
31
/* We pass "" instead of "C" because some libc's have
32
* non-ASCII default locale for setlocale("") call
33
* (this allows users of such libc to have Unicoded
34
* system without having to mess with env).
36
* We set LC_CTYPE because (a) we may be called with $LC_CTYPE
37
* value in LANG, not with $LC_ALL, (b) internationalized
38
* LC_NUMERIC and LC_TIME are more PITA than benefit
39
* (for one, some utilities have hard time with comma
40
* used as a fractional separator).
42
//TODO: avoid repeated calls by caching last string?
43
setlocale(LC_CTYPE, LANG ? LANG : "");
45
/* In unicode, this is a one character string */
46
width = unicode_strlen(unicode_0x394);
47
unicode_status = (width == 1 ? UNICODE_ON : UNICODE_OFF);
50
void FAST_FUNC init_unicode(void)
52
/* Some people set only $LC_CTYPE, not $LC_ALL, because they want
53
* only Unicode to be activated on their system, not the whole
54
* shebang of wrong decimal points, strange date formats and so on.
56
if (unicode_status == UNICODE_UNKNOWN) {
57
char *s = getenv("LC_ALL");
58
if (!s) s = getenv("LC_CTYPE");
59
if (!s) s = getenv("LANG");
66
/* Homegrown Unicode support. It knows only C and Unicode locales. */
68
# if ENABLE_FEATURE_CHECK_UNICODE_IN_ENV
69
void FAST_FUNC reinit_unicode(const char *LANG)
71
unicode_status = UNICODE_OFF;
72
if (!LANG || !(strstr(LANG, ".utf") || strstr(LANG, ".UTF")))
74
unicode_status = UNICODE_ON;
77
void FAST_FUNC init_unicode(void)
79
if (unicode_status == UNICODE_UNKNOWN) {
80
char *s = getenv("LC_ALL");
81
if (!s) s = getenv("LC_CTYPE");
82
if (!s) s = getenv("LANG");
88
static size_t wcrtomb_internal(char *s, wchar_t wc)
98
/* RFC 3629 says that Unicode ends at 10FFFF,
99
* but we cover entire 32 bits */
101
/* 4000000-FFFFFFFF -> 111111tt 10tttttt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
102
/* 200000-3FFFFFF -> 111110tt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
103
/* 10000-1FFFFF -> 11110zzz 10zzyyyy 10yyyyxx 10xxxxxx */
104
/* 800-FFFF -> 1110yyyy 10yyyyxx 10xxxxxx */
105
/* 80-7FF -> 110yyyxx 10xxxxxx */
107
/* How many bytes do we need? */
109
/* (0x80000000+ would result in n = 7, limiting n to 6) */
110
while (v >= 0x800 && n < 6) {
114
/* Fill bytes n-1..1 */
117
s[i] = (wc & 0x3f) | 0x80;
121
s[0] = wc | (uint8_t)(0x3f00 >> n);
124
size_t FAST_FUNC wcrtomb(char *s, wchar_t wc, mbstate_t *ps UNUSED_PARAM)
126
if (unicode_status != UNICODE_ON) {
131
return wcrtomb_internal(s, wc);
133
size_t FAST_FUNC wcstombs(char *dest, const wchar_t *src, size_t n)
137
if (unicode_status != UNICODE_ON) {
148
while (n >= MB_CUR_MAX) {
150
size_t len = wcrtomb_internal(dest, wc);
158
char tbuf[MB_CUR_MAX];
160
size_t len = wcrtomb_internal(tbuf, wc);
164
memcpy(dest, tbuf, len);
173
# define ERROR_WCHAR (~(wchar_t)0)
175
static const char *mbstowc_internal(wchar_t *res, const char *src)
178
unsigned c = (unsigned char) *src++;
185
/* 80-7FF -> 110yyyxx 10xxxxxx */
186
/* 800-FFFF -> 1110yyyy 10yyyyxx 10xxxxxx */
187
/* 10000-1FFFFF -> 11110zzz 10zzyyyy 10yyyyxx 10xxxxxx */
188
/* 200000-3FFFFFF -> 111110tt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
189
/* 4000000-FFFFFFFF -> 111111tt 10tttttt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */
194
} while ((c & 0x80) && bytes < 6);
196
/* A bare "continuation" byte. Say, 80 */
200
c = (uint8_t)(c) >> bytes;
203
unsigned ch = (unsigned char) *src;
204
if ((ch & 0xc0) != 0x80) {
205
/* Missing "continuation" byte. Example: e0 80 */
209
c = (c << 6) + (ch & 0x3f);
214
/* Need to check that c isn't produced by overlong encoding */
215
/* Example: 11000000 10000000 converts to NUL */
216
/* 11110000 10000000 10000100 10000000 converts to 0x100 */
217
/* correct encoding: 11000100 10000000 */
218
if (c <= 0x7f) { /* crude check */
226
size_t FAST_FUNC mbstowcs(wchar_t *dest, const char *src, size_t n)
230
if (unicode_status != UNICODE_ON) {
232
unsigned char c = *src++;
245
src = mbstowc_internal(&wc, src);
246
if (wc == ERROR_WCHAR) /* error */
250
if (wc == 0) /* end-of-string */
258
int FAST_FUNC iswspace(wint_t wc)
260
return (unsigned)wc <= 0x7f && isspace(wc);
263
int FAST_FUNC iswalnum(wint_t wc)
265
return (unsigned)wc <= 0x7f && isalnum(wc);
268
int FAST_FUNC iswpunct(wint_t wc)
270
return (unsigned)wc <= 0x7f && ispunct(wc);
274
# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
280
/* auxiliary function for binary search in interval table */
281
static int in_interval_table(unsigned ucs, const struct interval *table, unsigned max)
286
if (ucs < table[0].first || ucs > table[max].last)
291
mid = (min + max) / 2;
292
if (ucs > table[mid].last)
294
else if (ucs < table[mid].first)
302
static int in_uint16_table(unsigned ucs, const uint16_t *table, unsigned max)
306
unsigned first, last;
308
first = table[0] >> 2;
309
last = first + (table[0] & 3);
310
if (ucs < first || ucs > last)
315
mid = (min + max) / 2;
316
first = table[mid] >> 2;
317
last = first + (table[mid] & 3);
320
else if (ucs < first)
331
* This is an implementation of wcwidth() and wcswidth() (defined in
332
* IEEE Std 1002.1-2001) for Unicode.
334
* http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
335
* http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html
337
* In fixed-width output devices, Latin characters all occupy a single
338
* "cell" position of equal width, whereas ideographic CJK characters
339
* occupy two such cells. Interoperability between terminal-line
340
* applications and (teletype-style) character terminals using the
341
* UTF-8 encoding requires agreement on which character should advance
342
* the cursor by how many cell positions. No established formal
343
* standards exist at present on which Unicode character shall occupy
344
* how many cell positions on character terminals. These routines are
345
* a first attempt of defining such behavior based on simple rules
346
* applied to data provided by the Unicode Consortium.
348
* For some graphical characters, the Unicode standard explicitly
349
* defines a character-cell width via the definition of the East Asian
350
* FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
351
* In all these cases, there is no ambiguity about which width a
352
* terminal shall use. For characters in the East Asian Ambiguous (A)
353
* class, the width choice depends purely on a preference of backward
354
* compatibility with either historic CJK or Western practice.
355
* Choosing single-width for these characters is easy to justify as
356
* the appropriate long-term solution, as the CJK practice of
357
* displaying these characters as double-width comes from historic
358
* implementation simplicity (8-bit encoded characters were displayed
359
* single-width and 16-bit ones double-width, even for Greek,
360
* Cyrillic, etc.) and not any typographic considerations.
362
* Much less clear is the choice of width for the Not East Asian
363
* (Neutral) class. Existing practice does not dictate a width for any
364
* of these characters. It would nevertheless make sense
365
* typographically to allocate two character cells to characters such
366
* as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
367
* represented adequately with a single-width glyph. The following
368
* routines at present merely assign a single-cell width to all
369
* neutral characters, in the interest of simplicity. This is not
370
* entirely satisfactory and should be reconsidered before
371
* establishing a formal standard in this area. At the moment, the
372
* decision which Not East Asian (Neutral) characters should be
373
* represented by double-width glyphs cannot yet be answered by
374
* applying a simple rule from the Unicode database content. Setting
375
* up a proper standard for the behavior of UTF-8 character terminals
376
* will require a careful analysis not only of each Unicode character,
377
* but also of each presentation form, something the author of these
378
* routines has avoided to do so far.
380
* http://www.unicode.org/unicode/reports/tr11/
382
* Markus Kuhn -- 2007-05-26 (Unicode 5.0)
384
* Permission to use, copy, modify, and distribute this software
385
* for any purpose and without fee is hereby granted. The author
386
* disclaims all warranties with regard to this software.
388
* Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
391
/* Assigned Unicode character ranges:
393
* 0 0000–FFFF Basic Multilingual Plane
394
* 1 10000–1FFFF Supplementary Multilingual Plane
395
* 2 20000–2FFFF Supplementary Ideographic Plane
396
* 3 30000-3FFFF Tertiary Ideographic Plane (no chars assigned yet)
397
* 4-13 40000–DFFFF currently unassigned
398
* 14 E0000–EFFFF Supplementary Special-purpose Plane
399
* 15 F0000–FFFFF Supplementary Private Use Area-A
400
* 16 100000–10FFFF Supplementary Private Use Area-B
402
* "Supplementary Special-purpose Plane currently contains non-graphical
403
* characters in two blocks of 128 and 240 characters. The first block
404
* is for language tag characters for use when language cannot be indicated
405
* through other protocols (such as the xml:lang attribute in XML).
406
* The other block contains glyph variation selectors to indicate
407
* an alternate glyph for a character that cannot be determined by context."
409
* In simpler terms: it is a tool to fix the "Han unification" mess
410
* created by Unicode committee, to select Chinese/Japanese/Korean/Taiwan
411
* version of a character. (They forgot that the whole purpose of the Unicode
412
* was to be able to write all chars in one charset without such tricks).
413
* Until East Asian users say it is actually necessary to support these
414
* code points in console applications like busybox
415
* (i.e. do these chars ever appear in filenames, hostnames, text files
416
* and such?), we are treating these code points as invalid.
418
* Tertiary Ideographic Plane is also ignored for now,
419
* until Unicode committee assigns something there.
421
/* The following two functions define the column width of an ISO 10646
422
* character as follows:
424
* - The null character (U+0000) has a column width of 0.
426
* - Other C0/C1 control characters and DEL will lead to a return
429
* - Non-spacing and enclosing combining characters (general
430
* category code Mn or Me in the Unicode database) have a
433
* - SOFT HYPHEN (U+00AD) has a column width of 1.
435
* - Other format characters (general category code Cf in the Unicode
436
* database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
438
* - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
439
* have a column width of 0.
441
* - Spacing characters in the East Asian Wide (W) or East Asian
442
* Full-width (F) category as defined in Unicode Technical
443
* Report #11 have a column width of 2.
445
* - All remaining characters (including all printable
446
* ISO 8859-1 and WGL4 characters, Unicode control characters,
447
* etc.) have a column width of 1.
449
* This implementation assumes that wchar_t characters are encoded
452
int FAST_FUNC wcwidth(unsigned ucs)
454
# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
455
/* sorted list of non-overlapping intervals of non-spacing characters */
456
/* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
457
# define BIG_(a,b) { a, b },
459
# define ARRAY /* PAIR if < 0x4000 and no more than 4 chars big */ \
460
BIG_(0x0300, 0x036F) \
461
PAIR(0x0483, 0x0486) \
462
PAIR(0x0488, 0x0489) \
463
BIG_(0x0591, 0x05BD) \
464
PAIR(0x05BF, 0x05BF) \
465
PAIR(0x05C1, 0x05C2) \
466
PAIR(0x05C4, 0x05C5) \
467
PAIR(0x05C7, 0x05C7) \
468
PAIR(0x0600, 0x0603) \
469
BIG_(0x0610, 0x0615) \
470
BIG_(0x064B, 0x065E) \
471
PAIR(0x0670, 0x0670) \
472
BIG_(0x06D6, 0x06E4) \
473
PAIR(0x06E7, 0x06E8) \
474
PAIR(0x06EA, 0x06ED) \
475
PAIR(0x070F, 0x070F) \
476
PAIR(0x0711, 0x0711) \
477
BIG_(0x0730, 0x074A) \
478
BIG_(0x07A6, 0x07B0) \
479
BIG_(0x07EB, 0x07F3) \
480
PAIR(0x0901, 0x0902) \
481
PAIR(0x093C, 0x093C) \
482
BIG_(0x0941, 0x0948) \
483
PAIR(0x094D, 0x094D) \
484
PAIR(0x0951, 0x0954) \
485
PAIR(0x0962, 0x0963) \
486
PAIR(0x0981, 0x0981) \
487
PAIR(0x09BC, 0x09BC) \
488
PAIR(0x09C1, 0x09C4) \
489
PAIR(0x09CD, 0x09CD) \
490
PAIR(0x09E2, 0x09E3) \
491
PAIR(0x0A01, 0x0A02) \
492
PAIR(0x0A3C, 0x0A3C) \
493
PAIR(0x0A41, 0x0A42) \
494
PAIR(0x0A47, 0x0A48) \
495
PAIR(0x0A4B, 0x0A4D) \
496
PAIR(0x0A70, 0x0A71) \
497
PAIR(0x0A81, 0x0A82) \
498
PAIR(0x0ABC, 0x0ABC) \
499
BIG_(0x0AC1, 0x0AC5) \
500
PAIR(0x0AC7, 0x0AC8) \
501
PAIR(0x0ACD, 0x0ACD) \
502
PAIR(0x0AE2, 0x0AE3) \
503
PAIR(0x0B01, 0x0B01) \
504
PAIR(0x0B3C, 0x0B3C) \
505
PAIR(0x0B3F, 0x0B3F) \
506
PAIR(0x0B41, 0x0B43) \
507
PAIR(0x0B4D, 0x0B4D) \
508
PAIR(0x0B56, 0x0B56) \
509
PAIR(0x0B82, 0x0B82) \
510
PAIR(0x0BC0, 0x0BC0) \
511
PAIR(0x0BCD, 0x0BCD) \
512
PAIR(0x0C3E, 0x0C40) \
513
PAIR(0x0C46, 0x0C48) \
514
PAIR(0x0C4A, 0x0C4D) \
515
PAIR(0x0C55, 0x0C56) \
516
PAIR(0x0CBC, 0x0CBC) \
517
PAIR(0x0CBF, 0x0CBF) \
518
PAIR(0x0CC6, 0x0CC6) \
519
PAIR(0x0CCC, 0x0CCD) \
520
PAIR(0x0CE2, 0x0CE3) \
521
PAIR(0x0D41, 0x0D43) \
522
PAIR(0x0D4D, 0x0D4D) \
523
PAIR(0x0DCA, 0x0DCA) \
524
PAIR(0x0DD2, 0x0DD4) \
525
PAIR(0x0DD6, 0x0DD6) \
526
PAIR(0x0E31, 0x0E31) \
527
BIG_(0x0E34, 0x0E3A) \
528
BIG_(0x0E47, 0x0E4E) \
529
PAIR(0x0EB1, 0x0EB1) \
530
BIG_(0x0EB4, 0x0EB9) \
531
PAIR(0x0EBB, 0x0EBC) \
532
BIG_(0x0EC8, 0x0ECD) \
533
PAIR(0x0F18, 0x0F19) \
534
PAIR(0x0F35, 0x0F35) \
535
PAIR(0x0F37, 0x0F37) \
536
PAIR(0x0F39, 0x0F39) \
537
BIG_(0x0F71, 0x0F7E) \
538
BIG_(0x0F80, 0x0F84) \
539
PAIR(0x0F86, 0x0F87) \
540
PAIR(0x0FC6, 0x0FC6) \
541
BIG_(0x0F90, 0x0F97) \
542
BIG_(0x0F99, 0x0FBC) \
543
PAIR(0x102D, 0x1030) \
544
PAIR(0x1032, 0x1032) \
545
PAIR(0x1036, 0x1037) \
546
PAIR(0x1039, 0x1039) \
547
PAIR(0x1058, 0x1059) \
548
BIG_(0x1160, 0x11FF) \
549
PAIR(0x135F, 0x135F) \
550
PAIR(0x1712, 0x1714) \
551
PAIR(0x1732, 0x1734) \
552
PAIR(0x1752, 0x1753) \
553
PAIR(0x1772, 0x1773) \
554
PAIR(0x17B4, 0x17B5) \
555
BIG_(0x17B7, 0x17BD) \
556
PAIR(0x17C6, 0x17C6) \
557
BIG_(0x17C9, 0x17D3) \
558
PAIR(0x17DD, 0x17DD) \
559
PAIR(0x180B, 0x180D) \
560
PAIR(0x18A9, 0x18A9) \
561
PAIR(0x1920, 0x1922) \
562
PAIR(0x1927, 0x1928) \
563
PAIR(0x1932, 0x1932) \
564
PAIR(0x1939, 0x193B) \
565
PAIR(0x1A17, 0x1A18) \
566
PAIR(0x1B00, 0x1B03) \
567
PAIR(0x1B34, 0x1B34) \
568
BIG_(0x1B36, 0x1B3A) \
569
PAIR(0x1B3C, 0x1B3C) \
570
PAIR(0x1B42, 0x1B42) \
571
BIG_(0x1B6B, 0x1B73) \
572
BIG_(0x1DC0, 0x1DCA) \
573
PAIR(0x1DFE, 0x1DFF) \
574
BIG_(0x200B, 0x200F) \
575
BIG_(0x202A, 0x202E) \
576
PAIR(0x2060, 0x2063) \
577
BIG_(0x206A, 0x206F) \
578
BIG_(0x20D0, 0x20EF) \
579
BIG_(0x302A, 0x302F) \
580
PAIR(0x3099, 0x309A) \
581
/* Too big to be packed in PAIRs: */ \
582
BIG_(0xA806, 0xA806) \
583
BIG_(0xA80B, 0xA80B) \
584
BIG_(0xA825, 0xA826) \
585
BIG_(0xFB1E, 0xFB1E) \
586
BIG_(0xFE00, 0xFE0F) \
587
BIG_(0xFE20, 0xFE23) \
588
BIG_(0xFEFF, 0xFEFF) \
590
static const struct interval combining[] = { ARRAY };
594
# define PAIR(a,b) (a << 2) | (b-a),
595
static const uint16_t combining1[] = { ARRAY };
598
# define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
599
# define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
600
struct CHECK { ARRAY };
609
/* Test for 8-bit control characters (00-1f, 80-9f, 7f) */
610
if ((ucs & ~0x80) < 0x20 || ucs == 0x7f)
612
/* Quick abort if it is an obviously invalid char */
613
if (ucs > CONFIG_LAST_SUPPORTED_WCHAR)
616
/* Optimization: no combining chars below 0x300 */
617
if (CONFIG_LAST_SUPPORTED_WCHAR < 0x300 || ucs < 0x300)
620
# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
621
/* Binary search in table of non-spacing characters */
622
if (in_interval_table(ucs, combining, ARRAY_SIZE(combining) - 1))
624
if (in_uint16_table(ucs, combining1, ARRAY_SIZE(combining1) - 1))
627
/* Optimization: all chars below 0x1100 are not double-width */
628
if (CONFIG_LAST_SUPPORTED_WCHAR < 0x1100 || ucs < 0x1100)
631
# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x1100
632
/* Invalid code points: */
633
/* High (d800..dbff) and low (dc00..dfff) surrogates (valid only in UTF16) */
634
/* Private Use Area (e000..f8ff) */
635
/* Noncharacters fdd0..fdef */
636
if ((CONFIG_LAST_SUPPORTED_WCHAR >= 0xd800 && ucs >= 0xd800 && ucs <= 0xf8ff)
637
|| (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfdd0 && ucs >= 0xfdd0 && ucs <= 0xfdef)
641
/* 0xfffe and 0xffff in every plane are invalid */
642
if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfffe && ((ucs & 0xfffe) == 0xfffe)) {
646
# if CONFIG_LAST_SUPPORTED_WCHAR >= 0x10000
647
if (ucs >= 0x10000) {
648
/* Combining chars in Supplementary Multilingual Plane 0x1xxxx */
649
static const struct interval combining0x10000[] = {
650
{ 0x0A01, 0x0A03 }, { 0x0A05, 0x0A06 }, { 0x0A0C, 0x0A0F },
651
{ 0x0A38, 0x0A3A }, { 0x0A3F, 0x0A3F }, { 0xD167, 0xD169 },
652
{ 0xD173, 0xD182 }, { 0xD185, 0xD18B }, { 0xD1AA, 0xD1AD },
655
/* Binary search in table of non-spacing characters in Supplementary Multilingual Plane */
656
if (in_interval_table(ucs ^ 0x10000, combining0x10000, ARRAY_SIZE(combining0x10000) - 1))
658
/* Check a few non-spacing chars in Supplementary Special-purpose Plane 0xExxxx */
659
if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xE0001
661
|| (ucs >= 0xE0020 && ucs <= 0xE007F)
662
|| (ucs >= 0xE0100 && ucs <= 0xE01EF)
670
/* If we arrive here, ucs is not a combining or C0/C1 control character.
671
* Check whether it's 1 char or 2-shar wide.
674
( (/*ucs >= 0x1100 &&*/ ucs <= 0x115f) /* Hangul Jamo init. consonants */
675
|| ucs == 0x2329 /* left-pointing angle bracket; also CJK punct. char */
676
|| ucs == 0x232a /* right-pointing angle bracket; also CJK punct. char */
677
|| (ucs >= 0x2e80 && ucs <= 0xa4cf && ucs != 0x303f) /* CJK ... Yi */
678
# if CONFIG_LAST_SUPPORTED_WCHAR >= 0xac00
679
|| (ucs >= 0xac00 && ucs <= 0xd7a3) /* Hangul Syllables */
680
|| (ucs >= 0xf900 && ucs <= 0xfaff) /* CJK Compatibility Ideographs */
681
|| (ucs >= 0xfe10 && ucs <= 0xfe19) /* Vertical forms */
682
|| (ucs >= 0xfe30 && ucs <= 0xfe6f) /* CJK Compatibility Forms */
683
|| (ucs >= 0xff00 && ucs <= 0xff60) /* Fullwidth Forms */
684
|| (ucs >= 0xffe0 && ucs <= 0xffe6)
685
|| ((ucs >> 17) == (2 >> 1)) /* 20000..3ffff: Supplementary and Tertiary Ideographic Planes */
688
# endif /* >= 0x1100 */
689
# endif /* >= 0x300 */
693
# if ENABLE_UNICODE_BIDI_SUPPORT
694
int FAST_FUNC unicode_bidi_isrtl(wint_t wc)
697
* http://www.unicode.org/Public/5.2.0/ucd/extracted/DerivedBidiClass.txt
698
* Bidi_Class=Left_To_Right | Bidi_Class=Arabic_Letter
700
# define BIG_(a,b) { a, b },
703
PAIR(0x0590, 0x0590) \
704
PAIR(0x05BE, 0x05BE) \
705
PAIR(0x05C0, 0x05C0) \
706
PAIR(0x05C3, 0x05C3) \
707
PAIR(0x05C6, 0x05C6) \
708
BIG_(0x05C8, 0x05FF) \
709
PAIR(0x0604, 0x0605) \
710
PAIR(0x0608, 0x0608) \
711
PAIR(0x060B, 0x060B) \
712
PAIR(0x060D, 0x060D) \
713
BIG_(0x061B, 0x064A) \
714
PAIR(0x065F, 0x065F) \
715
PAIR(0x066D, 0x066F) \
716
BIG_(0x0671, 0x06D5) \
717
PAIR(0x06E5, 0x06E6) \
718
PAIR(0x06EE, 0x06EF) \
719
BIG_(0x06FA, 0x070E) \
720
PAIR(0x0710, 0x0710) \
721
BIG_(0x0712, 0x072F) \
722
BIG_(0x074B, 0x07A5) \
723
BIG_(0x07B1, 0x07EA) \
724
PAIR(0x07F4, 0x07F5) \
725
BIG_(0x07FA, 0x0815) \
726
PAIR(0x081A, 0x081A) \
727
PAIR(0x0824, 0x0824) \
728
PAIR(0x0828, 0x0828) \
729
BIG_(0x082E, 0x08FF) \
730
PAIR(0x200F, 0x200F) \
731
PAIR(0x202B, 0x202B) \
732
PAIR(0x202E, 0x202E) \
733
BIG_(0xFB1D, 0xFB1D) \
734
BIG_(0xFB1F, 0xFB28) \
735
BIG_(0xFB2A, 0xFD3D) \
736
BIG_(0xFD40, 0xFDCF) \
737
BIG_(0xFDC8, 0xFDCF) \
738
BIG_(0xFDF0, 0xFDFC) \
739
BIG_(0xFDFE, 0xFDFF) \
741
/* Probably not necessary
754
static const struct interval rtl_b[] = { ARRAY };
758
# define PAIR(a,b) (a << 2) | (b-a),
759
static const uint16_t rtl_p[] = { ARRAY };
762
# define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
763
# define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
764
struct CHECK { ARRAY };
769
if (in_interval_table(wc, rtl_b, ARRAY_SIZE(rtl_b) - 1))
771
if (in_uint16_table(wc, rtl_p, ARRAY_SIZE(rtl_p) - 1))
776
# if ENABLE_UNICODE_NEUTRAL_TABLE
777
int FAST_FUNC unicode_bidi_is_neutral_wchar(wint_t wc)
780
* http://www.unicode.org/Public/5.2.0/ucd/extracted/DerivedBidiClass.txt
781
* Bidi_Classes: Paragraph_Separator, Segment_Separator,
782
* White_Space, Other_Neutral, European_Number, European_Separator,
783
* European_Terminator, Arabic_Number, Common_Separator
785
# define BIG_(a,b) { a, b },
788
BIG_(0x0009, 0x000D) \
789
BIG_(0x001C, 0x0040) \
790
BIG_(0x005B, 0x0060) \
791
PAIR(0x007B, 0x007E) \
792
PAIR(0x0085, 0x0085) \
793
BIG_(0x00A0, 0x00A9) \
794
PAIR(0x00AB, 0x00AC) \
795
BIG_(0x00AE, 0x00B4) \
796
PAIR(0x00B6, 0x00B9) \
797
BIG_(0x00BB, 0x00BF) \
798
PAIR(0x00D7, 0x00D7) \
799
PAIR(0x00F7, 0x00F7) \
800
PAIR(0x02B9, 0x02BA) \
801
BIG_(0x02C2, 0x02CF) \
802
BIG_(0x02D2, 0x02DF) \
803
BIG_(0x02E5, 0x02FF) \
804
PAIR(0x0374, 0x0375) \
805
PAIR(0x037E, 0x037E) \
806
PAIR(0x0384, 0x0385) \
807
PAIR(0x0387, 0x0387) \
808
PAIR(0x03F6, 0x03F6) \
809
PAIR(0x058A, 0x058A) \
810
PAIR(0x0600, 0x0603) \
811
PAIR(0x0606, 0x0607) \
812
PAIR(0x0609, 0x060A) \
813
PAIR(0x060C, 0x060C) \
814
PAIR(0x060E, 0x060F) \
815
BIG_(0x0660, 0x066C) \
816
PAIR(0x06DD, 0x06DD) \
817
PAIR(0x06E9, 0x06E9) \
818
BIG_(0x06F0, 0x06F9) \
819
PAIR(0x07F6, 0x07F9) \
820
PAIR(0x09F2, 0x09F3) \
821
PAIR(0x09FB, 0x09FB) \
822
PAIR(0x0AF1, 0x0AF1) \
823
BIG_(0x0BF3, 0x0BFA) \
824
BIG_(0x0C78, 0x0C7E) \
825
PAIR(0x0CF1, 0x0CF2) \
826
PAIR(0x0E3F, 0x0E3F) \
827
PAIR(0x0F3A, 0x0F3D) \
828
BIG_(0x1390, 0x1400) \
829
PAIR(0x1680, 0x1680) \
830
PAIR(0x169B, 0x169C) \
831
PAIR(0x17DB, 0x17DB) \
832
BIG_(0x17F0, 0x17F9) \
833
BIG_(0x1800, 0x180A) \
834
PAIR(0x180E, 0x180E) \
835
PAIR(0x1940, 0x1940) \
836
PAIR(0x1944, 0x1945) \
837
BIG_(0x19DE, 0x19FF) \
838
PAIR(0x1FBD, 0x1FBD) \
839
PAIR(0x1FBF, 0x1FC1) \
840
PAIR(0x1FCD, 0x1FCF) \
841
PAIR(0x1FDD, 0x1FDF) \
842
PAIR(0x1FED, 0x1FEF) \
843
PAIR(0x1FFD, 0x1FFE) \
844
BIG_(0x2000, 0x200A) \
845
BIG_(0x2010, 0x2029) \
846
BIG_(0x202F, 0x205F) \
847
PAIR(0x2070, 0x2070) \
848
BIG_(0x2074, 0x207E) \
849
BIG_(0x2080, 0x208E) \
850
BIG_(0x20A0, 0x20B8) \
851
PAIR(0x2100, 0x2101) \
852
PAIR(0x2103, 0x2106) \
853
PAIR(0x2108, 0x2109) \
854
PAIR(0x2114, 0x2114) \
855
PAIR(0x2116, 0x2118) \
856
BIG_(0x211E, 0x2123) \
857
PAIR(0x2125, 0x2125) \
858
PAIR(0x2127, 0x2127) \
859
PAIR(0x2129, 0x2129) \
860
PAIR(0x212E, 0x212E) \
861
PAIR(0x213A, 0x213B) \
862
BIG_(0x2140, 0x2144) \
863
PAIR(0x214A, 0x214D) \
864
BIG_(0x2150, 0x215F) \
865
PAIR(0x2189, 0x2189) \
866
BIG_(0x2190, 0x2335) \
867
BIG_(0x237B, 0x2394) \
868
BIG_(0x2396, 0x23E8) \
869
BIG_(0x2400, 0x2426) \
870
BIG_(0x2440, 0x244A) \
871
BIG_(0x2460, 0x249B) \
872
BIG_(0x24EA, 0x26AB) \
873
BIG_(0x26AD, 0x26CD) \
874
BIG_(0x26CF, 0x26E1) \
875
PAIR(0x26E3, 0x26E3) \
876
BIG_(0x26E8, 0x26FF) \
877
PAIR(0x2701, 0x2704) \
878
PAIR(0x2706, 0x2709) \
879
BIG_(0x270C, 0x2727) \
880
BIG_(0x2729, 0x274B) \
881
PAIR(0x274D, 0x274D) \
882
PAIR(0x274F, 0x2752) \
883
BIG_(0x2756, 0x275E) \
884
BIG_(0x2761, 0x2794) \
885
BIG_(0x2798, 0x27AF) \
886
BIG_(0x27B1, 0x27BE) \
887
BIG_(0x27C0, 0x27CA) \
888
PAIR(0x27CC, 0x27CC) \
889
BIG_(0x27D0, 0x27FF) \
890
BIG_(0x2900, 0x2B4C) \
891
BIG_(0x2B50, 0x2B59) \
892
BIG_(0x2CE5, 0x2CEA) \
893
BIG_(0x2CF9, 0x2CFF) \
894
BIG_(0x2E00, 0x2E99) \
895
BIG_(0x2E9B, 0x2EF3) \
896
BIG_(0x2F00, 0x2FD5) \
897
BIG_(0x2FF0, 0x2FFB) \
898
BIG_(0x3000, 0x3004) \
899
BIG_(0x3008, 0x3020) \
900
PAIR(0x3030, 0x3030) \
901
PAIR(0x3036, 0x3037) \
902
PAIR(0x303D, 0x303D) \
903
PAIR(0x303E, 0x303F) \
904
PAIR(0x309B, 0x309C) \
905
PAIR(0x30A0, 0x30A0) \
906
PAIR(0x30FB, 0x30FB) \
907
BIG_(0x31C0, 0x31E3) \
908
PAIR(0x321D, 0x321E) \
909
BIG_(0x3250, 0x325F) \
910
PAIR(0x327C, 0x327E) \
911
BIG_(0x32B1, 0x32BF) \
912
PAIR(0x32CC, 0x32CF) \
913
PAIR(0x3377, 0x337A) \
914
PAIR(0x33DE, 0x33DF) \
915
PAIR(0x33FF, 0x33FF) \
916
BIG_(0x4DC0, 0x4DFF) \
917
BIG_(0xA490, 0xA4C6) \
918
BIG_(0xA60D, 0xA60F) \
919
BIG_(0xA673, 0xA673) \
920
BIG_(0xA67E, 0xA67F) \
921
BIG_(0xA700, 0xA721) \
922
BIG_(0xA788, 0xA788) \
923
BIG_(0xA828, 0xA82B) \
924
BIG_(0xA838, 0xA839) \
925
BIG_(0xA874, 0xA877) \
926
BIG_(0xFB29, 0xFB29) \
927
BIG_(0xFD3E, 0xFD3F) \
928
BIG_(0xFDFD, 0xFDFD) \
929
BIG_(0xFE10, 0xFE19) \
930
BIG_(0xFE30, 0xFE52) \
931
BIG_(0xFE54, 0xFE66) \
932
BIG_(0xFE68, 0xFE6B) \
933
BIG_(0xFF01, 0xFF20) \
934
BIG_(0xFF3B, 0xFF40) \
935
BIG_(0xFF5B, 0xFF65) \
936
BIG_(0xFFE0, 0xFFE6) \
937
BIG_(0xFFE8, 0xFFEE) \
958
static const struct interval neutral_b[] = { ARRAY };
962
# define PAIR(a,b) (a << 2) | (b-a),
963
static const uint16_t neutral_p[] = { ARRAY };
966
# define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
967
# define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
968
struct CHECK { ARRAY };
973
if (in_interval_table(wc, neutral_b, ARRAY_SIZE(neutral_b) - 1))
975
if (in_uint16_table(wc, neutral_p, ARRAY_SIZE(neutral_p) - 1))
981
# endif /* UNICODE_BIDI_SUPPORT */
983
#endif /* Homegrown Unicode support */
986
/* The rest is mostly same for libc and for "homegrown" support */
988
size_t FAST_FUNC unicode_strlen(const char *string)
990
size_t width = mbstowcs(NULL, string, INT_MAX);
991
if (width == (size_t)-1L)
992
return strlen(string);
996
size_t FAST_FUNC unicode_strwidth(const char *string)
999
printable_string2(&uni_stat, string);
1000
return uni_stat.unicode_width;
1003
static char* FAST_FUNC unicode_conv_to_printable2(uni_stat_t *stats, const char *src, unsigned width, int flags)
1010
if (unicode_status != UNICODE_ON) {
1012
if (flags & UNI_FLAG_PAD) {
1013
d = dst = xmalloc(width + 1);
1014
while ((int)--width >= 0) {
1015
unsigned char c = *src;
1019
while ((int)--width >= 0);
1022
*d++ = (c >= ' ' && c < 0x7f) ? c : '?';
1027
d = dst = xstrndup(src, width);
1029
unsigned char c = *d;
1030
if (c < ' ' || c >= 0x7f)
1036
stats->byte_count = (d - dst);
1037
stats->unicode_count = (d - dst);
1038
stats->unicode_width = (d - dst);
1044
uni_count = uni_width = 0;
1050
#if ENABLE_UNICODE_USING_LOCALE
1052
mbstate_t mbst = { 0 };
1053
ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst);
1054
/* If invalid sequence is seen: -1 is returned,
1055
* src points to the invalid sequence, errno = EILSEQ.
1056
* Else number of wchars (excluding terminating L'\0')
1057
* written to dest is returned.
1058
* If len (here: 1) non-L'\0' wchars stored at dest,
1059
* src points to the next char to be converted.
1060
* If string is completely converted: src = NULL.
1062
if (rc == 0) /* end-of-string */
1064
if (rc < 0) { /* error */
1072
src = mbstowc_internal(&wc, src);
1073
/* src is advanced to next mb char
1074
* wc == ERROR_WCHAR: invalid sequence is seen
1077
if (wc == ERROR_WCHAR) /* error */
1079
if (wc == 0) /* end-of-string */
1082
if (CONFIG_LAST_SUPPORTED_WCHAR && wc > CONFIG_LAST_SUPPORTED_WCHAR)
1085
if ((ENABLE_UNICODE_COMBINING_WCHARS && w < 0) /* non-printable wchar */
1086
|| (!ENABLE_UNICODE_COMBINING_WCHARS && w <= 0)
1087
|| (!ENABLE_UNICODE_WIDE_WCHARS && w > 1)
1090
wc = CONFIG_SUBST_WCHAR;
1094
/* Note: if width == 0, we still may add more chars,
1095
* they may be zero-width or combining ones */
1096
if ((int)width < 0) {
1097
/* can't add this wc, string would become longer than width */
1104
dst = xrealloc(dst, dst_len + MB_CUR_MAX);
1105
#if ENABLE_UNICODE_USING_LOCALE
1107
mbstate_t mbst = { 0 };
1108
dst_len += wcrtomb(&dst[dst_len], wc, &mbst);
1111
dst_len += wcrtomb_internal(&dst[dst_len], wc);
1115
/* Pad to remaining width */
1116
if (flags & UNI_FLAG_PAD) {
1117
dst = xrealloc(dst, dst_len + width + 1);
1120
while ((int)--width >= 0) {
1121
dst[dst_len++] = ' ';
1124
if (!dst) /* for example, if input was "" */
1126
dst[dst_len] = '\0';
1128
stats->byte_count = dst_len;
1129
stats->unicode_count = uni_count;
1130
stats->unicode_width = uni_width;
1135
char* FAST_FUNC unicode_conv_to_printable(uni_stat_t *stats, const char *src)
1137
return unicode_conv_to_printable2(stats, src, INT_MAX, 0);
1139
char* FAST_FUNC unicode_conv_to_printable_fixedwidth(/*uni_stat_t *stats,*/ const char *src, unsigned width)
1141
return unicode_conv_to_printable2(/*stats:*/ NULL, src, width, UNI_FLAG_PAD);
1145
char* FAST_FUNC unicode_conv_to_printable_maxwidth(uni_stat_t *stats, const char *src, unsigned maxwidth)
1147
return unicode_conv_to_printable2(stats, src, maxwidth, 0);
1150
unsigned FAST_FUNC unicode_padding_to_width(unsigned width, const char *src)
1152
if (unicode_status != UNICODE_ON) {
1153
return width - strnlen(src, width);
1160
#if ENABLE_UNICODE_USING_LOCALE
1162
mbstate_t mbst = { 0 };
1163
ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst);
1164
if (rc <= 0) /* error, or end-of-string */
1168
src = mbstowc_internal(&wc, src);
1169
if (wc == ERROR_WCHAR || wc == 0) /* error, or end-of-string */
1173
if (w < 0) /* non-printable wchar */
1176
if ((int)width <= 0) /* string is longer than width */