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
14 uint8_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 
reinit_unicode(const char * LANG)26 void FAST_FUNC reinit_unicode(const char *LANG)
27 {
28 	static const char unicode_0x394[] ALIGN1 = { 0xce, 0x94, 0 };
29 	size_t width;
30 
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).
35 	 *
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).
41 	 */
42 //TODO: avoid repeated calls by caching last string?
43 	setlocale(LC_CTYPE, LANG ? LANG : "");
44 
45 	/* In unicode, this is a one character string */
46 	width = unicode_strlen(unicode_0x394);
47 	unicode_status = (width == 1 ? UNICODE_ON : UNICODE_OFF);
48 }
49 
init_unicode(void)50 void FAST_FUNC init_unicode(void)
51 {
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.
55 	 */
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");
60 		reinit_unicode(s);
61 	}
62 }
63 
64 #else
65 
66 /* Homegrown Unicode support. It knows only C and Unicode locales. */
67 
68 # if ENABLE_FEATURE_CHECK_UNICODE_IN_ENV
reinit_unicode(const char * LANG)69 void FAST_FUNC reinit_unicode(const char *LANG)
70 {
71 	unicode_status = UNICODE_OFF;
72 	if (!LANG || !(strstr(LANG, ".utf") || strstr(LANG, ".UTF")))
73 		return;
74 	unicode_status = UNICODE_ON;
75 }
76 
init_unicode(void)77 void FAST_FUNC init_unicode(void)
78 {
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");
83 		reinit_unicode(s);
84 	}
85 }
86 # endif
87 
wcrtomb_internal(char * s,wchar_t wc)88 static size_t wcrtomb_internal(char *s, wchar_t wc)
89 {
90 	int n, i;
91 	uint32_t v = wc;
92 
93 	if (v <= 0x7f) {
94 		*s = v;
95 		return 1;
96 	}
97 
98 	/* RFC 3629 says that Unicode ends at 10FFFF,
99 	 * but we cover entire 32 bits */
100 
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 */
106 
107 	/* How many bytes do we need? */
108 	n = 2;
109 	/* (0x80000000+ would result in n = 7, limiting n to 6) */
110 	while (v >= 0x800 && n < 6) {
111 		v >>= 5;
112 		n++;
113 	}
114 	/* Fill bytes n-1..1 */
115 	i = n;
116 	while (--i) {
117 		s[i] = (wc & 0x3f) | 0x80;
118 		wc >>= 6;
119 	}
120 	/* Fill byte 0 */
121 	s[0] = wc | (uint8_t)(0x3f00 >> n);
122 	return n;
123 }
wcrtomb(char * s,wchar_t wc,mbstate_t * ps UNUSED_PARAM)124 size_t FAST_FUNC wcrtomb(char *s, wchar_t wc, mbstate_t *ps UNUSED_PARAM)
125 {
126 	if (unicode_status != UNICODE_ON) {
127 		*s = wc;
128 		return 1;
129 	}
130 
131 	return wcrtomb_internal(s, wc);
132 }
wcstombs(char * dest,const wchar_t * src,size_t n)133 size_t FAST_FUNC wcstombs(char *dest, const wchar_t *src, size_t n)
134 {
135 	size_t org_n = n;
136 
137 	if (unicode_status != UNICODE_ON) {
138 		while (n) {
139 			wchar_t c = *src++;
140 			*dest++ = c;
141 			if (c == 0)
142 				break;
143 			n--;
144 		}
145 		return org_n - n;
146 	}
147 
148 	while (n >= MB_CUR_MAX) {
149 		wchar_t wc = *src++;
150 		size_t len = wcrtomb_internal(dest, wc);
151 
152 		if (wc == L'\0')
153 			return org_n - n;
154 		dest += len;
155 		n -= len;
156 	}
157 	while (n) {
158 		char tbuf[MB_CUR_MAX];
159 		wchar_t wc = *src++;
160 		size_t len = wcrtomb_internal(tbuf, wc);
161 
162 		if (len > n)
163 			break;
164 		memcpy(dest, tbuf, len);
165 		if (wc == L'\0')
166 			return org_n - n;
167 		dest += len;
168 		n -= len;
169 	}
170 	return org_n - n;
171 }
172 
173 # define ERROR_WCHAR (~(wchar_t)0)
174 
mbstowc_internal(wchar_t * res,const char * src)175 static const char *mbstowc_internal(wchar_t *res, const char *src)
176 {
177 	int bytes;
178 	unsigned c = (unsigned char) *src++;
179 
180 	if (c <= 0x7f) {
181 		*res = c;
182 		return src;
183 	}
184 
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 */
190 	bytes = 0;
191 	do {
192 		c <<= 1;
193 		bytes++;
194 	} while ((c & 0x80) && bytes < 6);
195 	if (bytes == 1) {
196 		/* A bare "continuation" byte. Say, 80 */
197 		*res = ERROR_WCHAR;
198 		return src;
199 	}
200 	c = (uint8_t)(c) >> bytes;
201 
202 	while (--bytes) {
203 		unsigned ch = (unsigned char) *src;
204 		if ((ch & 0xc0) != 0x80) {
205 			/* Missing "continuation" byte. Example: e0 80 */
206 			*res = ERROR_WCHAR;
207 			return src;
208 		}
209 		c = (c << 6) + (ch & 0x3f);
210 		src++;
211 	}
212 
213 	/* TODO */
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 */
219 		*res = ERROR_WCHAR;
220 		return src;
221 	}
222 
223 	*res = c;
224 	return src;
225 }
mbstowcs(wchar_t * dest,const char * src,size_t n)226 size_t FAST_FUNC mbstowcs(wchar_t *dest, const char *src, size_t n)
227 {
228 	size_t org_n = n;
229 
230 	if (unicode_status != UNICODE_ON) {
231 		while (n) {
232 			unsigned char c = *src++;
233 
234 			if (dest)
235 				*dest++ = c;
236 			if (c == 0)
237 				break;
238 			n--;
239 		}
240 		return org_n - n;
241 	}
242 
243 	while (n) {
244 		wchar_t wc;
245 		src = mbstowc_internal(&wc, src);
246 		if (wc == ERROR_WCHAR) /* error */
247 			return (size_t) -1L;
248 		if (dest)
249 			*dest++ = wc;
250 		if (wc == 0) /* end-of-string */
251 			break;
252 		n--;
253 	}
254 
255 	return org_n - n;
256 }
257 
iswspace(wint_t wc)258 int FAST_FUNC iswspace(wint_t wc)
259 {
260 	return (unsigned)wc <= 0x7f && isspace(wc);
261 }
262 
iswalnum(wint_t wc)263 int FAST_FUNC iswalnum(wint_t wc)
264 {
265 	return (unsigned)wc <= 0x7f && isalnum(wc);
266 }
267 
iswpunct(wint_t wc)268 int FAST_FUNC iswpunct(wint_t wc)
269 {
270 	return (unsigned)wc <= 0x7f && ispunct(wc);
271 }
272 
273 
274 # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
275 struct interval {
276 	uint16_t first;
277 	uint16_t last;
278 };
279 
280 /* auxiliary function for binary search in interval table */
in_interval_table(unsigned ucs,const struct interval * table,unsigned max)281 static int in_interval_table(unsigned ucs, const struct interval *table, unsigned max)
282 {
283 	unsigned min;
284 	unsigned mid;
285 
286 	if (ucs < table[0].first || ucs > table[max].last)
287 		return 0;
288 
289 	min = 0;
290 	while (max >= min) {
291 		mid = (min + max) / 2;
292 		if (ucs > table[mid].last)
293 			min = mid + 1;
294 		else if (ucs < table[mid].first)
295 			max = mid - 1;
296 		else
297 			return 1;
298 	}
299 	return 0;
300 }
301 
in_uint16_table(unsigned ucs,const uint16_t * table,unsigned max)302 static int in_uint16_table(unsigned ucs, const uint16_t *table, unsigned max)
303 {
304 	unsigned min;
305 	unsigned mid;
306 	unsigned first, last;
307 
308 	first = table[0] >> 2;
309 	if (ucs < first)
310 		return 0;
311 	last = (table[max] >> 2) + (table[max] & 3);
312 	if (ucs > last)
313 		return 0;
314 
315 	min = 0;
316 	while (max >= min) {
317 		mid = (min + max) / 2;
318 		first = table[mid] >> 2;
319 		last = first + (table[mid] & 3);
320 		if (ucs > last)
321 			min = mid + 1;
322 		else if (ucs < first)
323 			max = mid - 1;
324 		else
325 			return 1;
326 	}
327 	return 0;
328 }
329 # endif
330 
331 
332 /*
333  * This is an implementation of wcwidth() and wcswidth() (defined in
334  * IEEE Std 1002.1-2001) for Unicode.
335  *
336  * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html
337  * http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html
338  *
339  * In fixed-width output devices, Latin characters all occupy a single
340  * "cell" position of equal width, whereas ideographic CJK characters
341  * occupy two such cells. Interoperability between terminal-line
342  * applications and (teletype-style) character terminals using the
343  * UTF-8 encoding requires agreement on which character should advance
344  * the cursor by how many cell positions. No established formal
345  * standards exist at present on which Unicode character shall occupy
346  * how many cell positions on character terminals. These routines are
347  * a first attempt of defining such behavior based on simple rules
348  * applied to data provided by the Unicode Consortium.
349  *
350  * For some graphical characters, the Unicode standard explicitly
351  * defines a character-cell width via the definition of the East Asian
352  * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes.
353  * In all these cases, there is no ambiguity about which width a
354  * terminal shall use. For characters in the East Asian Ambiguous (A)
355  * class, the width choice depends purely on a preference of backward
356  * compatibility with either historic CJK or Western practice.
357  * Choosing single-width for these characters is easy to justify as
358  * the appropriate long-term solution, as the CJK practice of
359  * displaying these characters as double-width comes from historic
360  * implementation simplicity (8-bit encoded characters were displayed
361  * single-width and 16-bit ones double-width, even for Greek,
362  * Cyrillic, etc.) and not any typographic considerations.
363  *
364  * Much less clear is the choice of width for the Not East Asian
365  * (Neutral) class. Existing practice does not dictate a width for any
366  * of these characters. It would nevertheless make sense
367  * typographically to allocate two character cells to characters such
368  * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be
369  * represented adequately with a single-width glyph. The following
370  * routines at present merely assign a single-cell width to all
371  * neutral characters, in the interest of simplicity. This is not
372  * entirely satisfactory and should be reconsidered before
373  * establishing a formal standard in this area. At the moment, the
374  * decision which Not East Asian (Neutral) characters should be
375  * represented by double-width glyphs cannot yet be answered by
376  * applying a simple rule from the Unicode database content. Setting
377  * up a proper standard for the behavior of UTF-8 character terminals
378  * will require a careful analysis not only of each Unicode character,
379  * but also of each presentation form, something the author of these
380  * routines has avoided to do so far.
381  *
382  * http://www.unicode.org/unicode/reports/tr11/
383  *
384  * Markus Kuhn -- 2007-05-26 (Unicode 5.0)
385  *
386  * Permission to use, copy, modify, and distribute this software
387  * for any purpose and without fee is hereby granted. The author
388  * disclaims all warranties with regard to this software.
389  *
390  * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
391  */
392 
393 /* Assigned Unicode character ranges:
394  * Plane Range
395  * 0       0000–FFFF   Basic Multilingual Plane
396  * 1      10000–1FFFF  Supplementary Multilingual Plane
397  * 2      20000–2FFFF  Supplementary Ideographic Plane
398  * 3      30000-3FFFF  Tertiary Ideographic Plane (no chars assigned yet)
399  * 4-13   40000–DFFFF  currently unassigned
400  * 14     E0000–EFFFF  Supplementary Special-purpose Plane
401  * 15     F0000–FFFFF  Supplementary Private Use Area-A
402  * 16    100000–10FFFF Supplementary Private Use Area-B
403  *
404  * "Supplementary Special-purpose Plane currently contains non-graphical
405  * characters in two blocks of 128 and 240 characters. The first block
406  * is for language tag characters for use when language cannot be indicated
407  * through other protocols (such as the xml:lang  attribute in XML).
408  * The other block contains glyph variation selectors to indicate
409  * an alternate glyph for a character that cannot be determined by context."
410  *
411  * In simpler terms: it is a tool to fix the "Han unification" mess
412  * created by Unicode committee, to select Chinese/Japanese/Korean/Taiwan
413  * version of a character. (They forgot that the whole purpose of the Unicode
414  * was to be able to write all chars in one charset without such tricks).
415  * Until East Asian users say it is actually necessary to support these
416  * code points in console applications like busybox
417  * (i.e. do these chars ever appear in filenames, hostnames, text files
418  * and such?), we are treating these code points as invalid.
419  *
420  * Tertiary Ideographic Plane is also ignored for now,
421  * until Unicode committee assigns something there.
422  */
423 /* The following two functions define the column width of an ISO 10646
424  * character as follows:
425  *
426  *    - The null character (U+0000) has a column width of 0.
427  *
428  *    - Other C0/C1 control characters and DEL will lead to a return
429  *      value of -1.
430  *
431  *    - Non-spacing and enclosing combining characters (general
432  *      category code Mn or Me in the Unicode database) have a
433  *      column width of 0.
434  *
435  *    - SOFT HYPHEN (U+00AD) has a column width of 1.
436  *
437  *    - Other format characters (general category code Cf in the Unicode
438  *      database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
439  *
440  *    - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
441  *      have a column width of 0.
442  *
443  *    - Spacing characters in the East Asian Wide (W) or East Asian
444  *      Full-width (F) category as defined in Unicode Technical
445  *      Report #11 have a column width of 2.
446  *
447  *    - All remaining characters (including all printable
448  *      ISO 8859-1 and WGL4 characters, Unicode control characters,
449  *      etc.) have a column width of 1.
450  *
451  * This implementation assumes that wchar_t characters are encoded
452  * in ISO 10646.
453  */
wcwidth(unsigned ucs)454 int FAST_FUNC wcwidth(unsigned ucs)
455 {
456 # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
457 	/* sorted list of non-overlapping intervals of non-spacing characters */
458 	/* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */
459 #  define BIG_(a,b) { a, b },
460 #  define PAIR(a,b)
461 #  define ARRAY /* PAIR if < 0x4000 and no more than 4 chars big */ \
462 		BIG_(0x0300, 0x036F) \
463 		PAIR(0x0483, 0x0486) \
464 		PAIR(0x0488, 0x0489) \
465 		BIG_(0x0591, 0x05BD) \
466 		PAIR(0x05BF, 0x05BF) \
467 		PAIR(0x05C1, 0x05C2) \
468 		PAIR(0x05C4, 0x05C5) \
469 		PAIR(0x05C7, 0x05C7) \
470 		PAIR(0x0600, 0x0603) \
471 		BIG_(0x0610, 0x0615) \
472 		BIG_(0x064B, 0x065E) \
473 		PAIR(0x0670, 0x0670) \
474 		BIG_(0x06D6, 0x06E4) \
475 		PAIR(0x06E7, 0x06E8) \
476 		PAIR(0x06EA, 0x06ED) \
477 		PAIR(0x070F, 0x070F) \
478 		PAIR(0x0711, 0x0711) \
479 		BIG_(0x0730, 0x074A) \
480 		BIG_(0x07A6, 0x07B0) \
481 		BIG_(0x07EB, 0x07F3) \
482 		PAIR(0x0901, 0x0902) \
483 		PAIR(0x093C, 0x093C) \
484 		BIG_(0x0941, 0x0948) \
485 		PAIR(0x094D, 0x094D) \
486 		PAIR(0x0951, 0x0954) \
487 		PAIR(0x0962, 0x0963) \
488 		PAIR(0x0981, 0x0981) \
489 		PAIR(0x09BC, 0x09BC) \
490 		PAIR(0x09C1, 0x09C4) \
491 		PAIR(0x09CD, 0x09CD) \
492 		PAIR(0x09E2, 0x09E3) \
493 		PAIR(0x0A01, 0x0A02) \
494 		PAIR(0x0A3C, 0x0A3C) \
495 		PAIR(0x0A41, 0x0A42) \
496 		PAIR(0x0A47, 0x0A48) \
497 		PAIR(0x0A4B, 0x0A4D) \
498 		PAIR(0x0A70, 0x0A71) \
499 		PAIR(0x0A81, 0x0A82) \
500 		PAIR(0x0ABC, 0x0ABC) \
501 		BIG_(0x0AC1, 0x0AC5) \
502 		PAIR(0x0AC7, 0x0AC8) \
503 		PAIR(0x0ACD, 0x0ACD) \
504 		PAIR(0x0AE2, 0x0AE3) \
505 		PAIR(0x0B01, 0x0B01) \
506 		PAIR(0x0B3C, 0x0B3C) \
507 		PAIR(0x0B3F, 0x0B3F) \
508 		PAIR(0x0B41, 0x0B43) \
509 		PAIR(0x0B4D, 0x0B4D) \
510 		PAIR(0x0B56, 0x0B56) \
511 		PAIR(0x0B82, 0x0B82) \
512 		PAIR(0x0BC0, 0x0BC0) \
513 		PAIR(0x0BCD, 0x0BCD) \
514 		PAIR(0x0C3E, 0x0C40) \
515 		PAIR(0x0C46, 0x0C48) \
516 		PAIR(0x0C4A, 0x0C4D) \
517 		PAIR(0x0C55, 0x0C56) \
518 		PAIR(0x0CBC, 0x0CBC) \
519 		PAIR(0x0CBF, 0x0CBF) \
520 		PAIR(0x0CC6, 0x0CC6) \
521 		PAIR(0x0CCC, 0x0CCD) \
522 		PAIR(0x0CE2, 0x0CE3) \
523 		PAIR(0x0D41, 0x0D43) \
524 		PAIR(0x0D4D, 0x0D4D) \
525 		PAIR(0x0DCA, 0x0DCA) \
526 		PAIR(0x0DD2, 0x0DD4) \
527 		PAIR(0x0DD6, 0x0DD6) \
528 		PAIR(0x0E31, 0x0E31) \
529 		BIG_(0x0E34, 0x0E3A) \
530 		BIG_(0x0E47, 0x0E4E) \
531 		PAIR(0x0EB1, 0x0EB1) \
532 		BIG_(0x0EB4, 0x0EB9) \
533 		PAIR(0x0EBB, 0x0EBC) \
534 		BIG_(0x0EC8, 0x0ECD) \
535 		PAIR(0x0F18, 0x0F19) \
536 		PAIR(0x0F35, 0x0F35) \
537 		PAIR(0x0F37, 0x0F37) \
538 		PAIR(0x0F39, 0x0F39) \
539 		BIG_(0x0F71, 0x0F7E) \
540 		BIG_(0x0F80, 0x0F84) \
541 		PAIR(0x0F86, 0x0F87) \
542 		PAIR(0x0FC6, 0x0FC6) \
543 		BIG_(0x0F90, 0x0F97) \
544 		BIG_(0x0F99, 0x0FBC) \
545 		PAIR(0x102D, 0x1030) \
546 		PAIR(0x1032, 0x1032) \
547 		PAIR(0x1036, 0x1037) \
548 		PAIR(0x1039, 0x1039) \
549 		PAIR(0x1058, 0x1059) \
550 		BIG_(0x1160, 0x11FF) \
551 		PAIR(0x135F, 0x135F) \
552 		PAIR(0x1712, 0x1714) \
553 		PAIR(0x1732, 0x1734) \
554 		PAIR(0x1752, 0x1753) \
555 		PAIR(0x1772, 0x1773) \
556 		PAIR(0x17B4, 0x17B5) \
557 		BIG_(0x17B7, 0x17BD) \
558 		PAIR(0x17C6, 0x17C6) \
559 		BIG_(0x17C9, 0x17D3) \
560 		PAIR(0x17DD, 0x17DD) \
561 		PAIR(0x180B, 0x180D) \
562 		PAIR(0x18A9, 0x18A9) \
563 		PAIR(0x1920, 0x1922) \
564 		PAIR(0x1927, 0x1928) \
565 		PAIR(0x1932, 0x1932) \
566 		PAIR(0x1939, 0x193B) \
567 		PAIR(0x1A17, 0x1A18) \
568 		PAIR(0x1B00, 0x1B03) \
569 		PAIR(0x1B34, 0x1B34) \
570 		BIG_(0x1B36, 0x1B3A) \
571 		PAIR(0x1B3C, 0x1B3C) \
572 		PAIR(0x1B42, 0x1B42) \
573 		BIG_(0x1B6B, 0x1B73) \
574 		BIG_(0x1DC0, 0x1DCA) \
575 		PAIR(0x1DFE, 0x1DFF) \
576 		BIG_(0x200B, 0x200F) \
577 		BIG_(0x202A, 0x202E) \
578 		PAIR(0x2060, 0x2063) \
579 		BIG_(0x206A, 0x206F) \
580 		BIG_(0x20D0, 0x20EF) \
581 		BIG_(0x302A, 0x302F) \
582 		PAIR(0x3099, 0x309A) \
583 		/* Too big to be packed in PAIRs: */ \
584 		BIG_(0xA806, 0xA806) \
585 		BIG_(0xA80B, 0xA80B) \
586 		BIG_(0xA825, 0xA826) \
587 		BIG_(0xFB1E, 0xFB1E) \
588 		BIG_(0xFE00, 0xFE0F) \
589 		BIG_(0xFE20, 0xFE23) \
590 		BIG_(0xFEFF, 0xFEFF) \
591 		BIG_(0xFFF9, 0xFFFB)
592 	static const struct interval combining[] ALIGN4 = { ARRAY };
593 #  undef BIG_
594 #  undef PAIR
595 #  define BIG_(a,b)
596 #  define PAIR(a,b) (a << 2) | (b-a),
597 	static const uint16_t combining1[] ALIGN2 = { ARRAY };
598 #  undef BIG_
599 #  undef PAIR
600 #  define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
601 #  define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
602 	struct CHECK { ARRAY };
603 #  undef BIG_
604 #  undef PAIR
605 #  undef ARRAY
606 # endif
607 
608 	if (ucs == 0)
609 		return 0;
610 
611 	/* Test for 8-bit control characters (00-1f, 80-9f, 7f) */
612 	if ((ucs & ~0x80) < 0x20 || ucs == 0x7f)
613 		return -1;
614 	/* Quick abort if it is an obviously invalid char */
615 	if (ucs > CONFIG_LAST_SUPPORTED_WCHAR)
616 		return -1;
617 
618 	/* Optimization: no combining chars below 0x300 */
619 	if (CONFIG_LAST_SUPPORTED_WCHAR < 0x300 || ucs < 0x300)
620 		return 1;
621 
622 # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300
623 	/* Binary search in table of non-spacing characters */
624 	if (in_interval_table(ucs, combining, ARRAY_SIZE(combining) - 1))
625 		return 0;
626 	if (in_uint16_table(ucs, combining1, ARRAY_SIZE(combining1) - 1))
627 		return 0;
628 
629 	/* Optimization: all chars below 0x1100 are not double-width */
630 	if (CONFIG_LAST_SUPPORTED_WCHAR < 0x1100 || ucs < 0x1100)
631 		return 1;
632 
633 #  if CONFIG_LAST_SUPPORTED_WCHAR >= 0x1100
634 	/* Invalid code points: */
635 	/* High (d800..dbff) and low (dc00..dfff) surrogates (valid only in UTF16) */
636 	/* Private Use Area (e000..f8ff) */
637 	/* Noncharacters fdd0..fdef */
638 	if ((CONFIG_LAST_SUPPORTED_WCHAR >= 0xd800 && ucs >= 0xd800 && ucs <= 0xf8ff)
639 	 || (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfdd0 && ucs >= 0xfdd0 && ucs <= 0xfdef)
640 	) {
641 		return -1;
642 	}
643 	/* 0xfffe and 0xffff in every plane are invalid */
644 	if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfffe && ((ucs & 0xfffe) == 0xfffe)) {
645 		return -1;
646 	}
647 
648 #   if CONFIG_LAST_SUPPORTED_WCHAR >= 0x10000
649 	if (ucs >= 0x10000) {
650 		/* Combining chars in Supplementary Multilingual Plane 0x1xxxx */
651 		static const struct interval combining0x10000[] ALIGN4 = {
652 			{ 0x0A01, 0x0A03 }, { 0x0A05, 0x0A06 }, { 0x0A0C, 0x0A0F },
653 			{ 0x0A38, 0x0A3A }, { 0x0A3F, 0x0A3F }, { 0xD167, 0xD169 },
654 			{ 0xD173, 0xD182 }, { 0xD185, 0xD18B }, { 0xD1AA, 0xD1AD },
655 			{ 0xD242, 0xD244 }
656 		};
657 		/* Binary search in table of non-spacing characters in Supplementary Multilingual Plane */
658 		if (in_interval_table(ucs ^ 0x10000, combining0x10000, ARRAY_SIZE(combining0x10000) - 1))
659 			return 0;
660 		/* Check a few non-spacing chars in Supplementary Special-purpose Plane 0xExxxx */
661 		if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xE0001
662 		 && (  ucs == 0xE0001
663 		    || (ucs >= 0xE0020 && ucs <= 0xE007F)
664 		    || (ucs >= 0xE0100 && ucs <= 0xE01EF)
665 		    )
666 		) {
667 			return 0;
668 		}
669 	}
670 #   endif
671 
672 	/* If we arrive here, ucs is not a combining or C0/C1 control character.
673 	 * Check whether it's 1 char or 2-shar wide.
674 	 */
675 	return 1 +
676 		(  (/*ucs >= 0x1100 &&*/ ucs <= 0x115f) /* Hangul Jamo init. consonants */
677 		|| ucs == 0x2329 /* left-pointing angle bracket; also CJK punct. char */
678 		|| ucs == 0x232a /* right-pointing angle bracket; also CJK punct. char */
679 #   if CONFIG_LAST_SUPPORTED_WCHAR >= 0x2e80
680 		|| (ucs >= 0x2e80 && ucs <= 0xa4cf && ucs != 0x303f) /* CJK ... Yi */
681 #   endif
682 #   if CONFIG_LAST_SUPPORTED_WCHAR >= 0xac00
683 		|| (ucs >= 0xac00 && ucs <= 0xd7a3) /* Hangul Syllables */
684 #   endif
685 #   if CONFIG_LAST_SUPPORTED_WCHAR >= 0xf900
686 		|| (ucs >= 0xf900 && ucs <= 0xfaff) /* CJK Compatibility Ideographs */
687 		|| (ucs >= 0xfe10 && ucs <= 0xfe19) /* Vertical forms */
688 		|| (ucs >= 0xfe30 && ucs <= 0xfe6f) /* CJK Compatibility Forms */
689 		|| (ucs >= 0xff00 && ucs <= 0xff60) /* Fullwidth Forms */
690 		|| (ucs >= 0xffe0 && ucs <= 0xffe6)
691 #   endif
692 #   if CONFIG_LAST_SUPPORTED_WCHAR >= 0x20000
693 		|| ((ucs >> 17) == (2 >> 1)) /* 20000..3ffff: Supplementary and Tertiary Ideographic Planes */
694 #   endif
695 		);
696 #  endif /* >= 0x1100 */
697 # endif /* >= 0x300 */
698 }
699 
700 
701 # if ENABLE_UNICODE_BIDI_SUPPORT
unicode_bidi_isrtl(wint_t wc)702 int FAST_FUNC unicode_bidi_isrtl(wint_t wc)
703 {
704 	/* ranges taken from
705 	 * http://www.unicode.org/Public/5.2.0/ucd/extracted/DerivedBidiClass.txt
706 	 * Bidi_Class=Left_To_Right | Bidi_Class=Arabic_Letter
707 	 */
708 #  define BIG_(a,b) { a, b },
709 #  define PAIR(a,b)
710 #  define ARRAY \
711 		PAIR(0x0590, 0x0590) \
712 		PAIR(0x05BE, 0x05BE) \
713 		PAIR(0x05C0, 0x05C0) \
714 		PAIR(0x05C3, 0x05C3) \
715 		PAIR(0x05C6, 0x05C6) \
716 		BIG_(0x05C8, 0x05FF) \
717 		PAIR(0x0604, 0x0605) \
718 		PAIR(0x0608, 0x0608) \
719 		PAIR(0x060B, 0x060B) \
720 		PAIR(0x060D, 0x060D) \
721 		BIG_(0x061B, 0x064A) \
722 		PAIR(0x065F, 0x065F) \
723 		PAIR(0x066D, 0x066F) \
724 		BIG_(0x0671, 0x06D5) \
725 		PAIR(0x06E5, 0x06E6) \
726 		PAIR(0x06EE, 0x06EF) \
727 		BIG_(0x06FA, 0x070E) \
728 		PAIR(0x0710, 0x0710) \
729 		BIG_(0x0712, 0x072F) \
730 		BIG_(0x074B, 0x07A5) \
731 		BIG_(0x07B1, 0x07EA) \
732 		PAIR(0x07F4, 0x07F5) \
733 		BIG_(0x07FA, 0x0815) \
734 		PAIR(0x081A, 0x081A) \
735 		PAIR(0x0824, 0x0824) \
736 		PAIR(0x0828, 0x0828) \
737 		BIG_(0x082E, 0x08FF) \
738 		PAIR(0x200F, 0x200F) \
739 		PAIR(0x202B, 0x202B) \
740 		PAIR(0x202E, 0x202E) \
741 		BIG_(0xFB1D, 0xFB1D) \
742 		BIG_(0xFB1F, 0xFB28) \
743 		BIG_(0xFB2A, 0xFD3D) \
744 		BIG_(0xFD40, 0xFDCF) \
745 		BIG_(0xFDC8, 0xFDCF) \
746 		BIG_(0xFDF0, 0xFDFC) \
747 		BIG_(0xFDFE, 0xFDFF) \
748 		BIG_(0xFE70, 0xFEFE)
749 		/* Probably not necessary
750 		{0x10800, 0x1091E},
751 		{0x10920, 0x10A00},
752 		{0x10A04, 0x10A04},
753 		{0x10A07, 0x10A0B},
754 		{0x10A10, 0x10A37},
755 		{0x10A3B, 0x10A3E},
756 		{0x10A40, 0x10A7F},
757 		{0x10B36, 0x10B38},
758 		{0x10B40, 0x10E5F},
759 		{0x10E7F, 0x10FFF},
760 		{0x1E800, 0x1EFFF}
761 		*/
762 	static const struct interval rtl_b[] ALIGN4 = { ARRAY };
763 #  undef BIG_
764 #  undef PAIR
765 #  define BIG_(a,b)
766 #  define PAIR(a,b) (a << 2) | (b-a),
767 	static const uint16_t rtl_p[] ALIGN2 = { ARRAY };
768 #  undef BIG_
769 #  undef PAIR
770 #  define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
771 #  define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
772 	struct CHECK { ARRAY };
773 #  undef BIG_
774 #  undef PAIR
775 #  undef ARRAY
776 
777 	if (in_interval_table(wc, rtl_b, ARRAY_SIZE(rtl_b) - 1))
778 		return 1;
779 	if (in_uint16_table(wc, rtl_p, ARRAY_SIZE(rtl_p) - 1))
780 		return 1;
781 	return 0;
782 }
783 
784 #  if ENABLE_UNICODE_NEUTRAL_TABLE
unicode_bidi_is_neutral_wchar(wint_t wc)785 int FAST_FUNC unicode_bidi_is_neutral_wchar(wint_t wc)
786 {
787 	/* ranges taken from
788 	 * http://www.unicode.org/Public/5.2.0/ucd/extracted/DerivedBidiClass.txt
789 	 * Bidi_Classes: Paragraph_Separator, Segment_Separator,
790 	 * White_Space, Other_Neutral, European_Number, European_Separator,
791 	 * European_Terminator, Arabic_Number, Common_Separator
792 	 */
793 #  define BIG_(a,b) { a, b },
794 #  define PAIR(a,b)
795 #  define ARRAY \
796 		BIG_(0x0009, 0x000D) \
797 		BIG_(0x001C, 0x0040) \
798 		BIG_(0x005B, 0x0060) \
799 		PAIR(0x007B, 0x007E) \
800 		PAIR(0x0085, 0x0085) \
801 		BIG_(0x00A0, 0x00A9) \
802 		PAIR(0x00AB, 0x00AC) \
803 		BIG_(0x00AE, 0x00B4) \
804 		PAIR(0x00B6, 0x00B9) \
805 		BIG_(0x00BB, 0x00BF) \
806 		PAIR(0x00D7, 0x00D7) \
807 		PAIR(0x00F7, 0x00F7) \
808 		PAIR(0x02B9, 0x02BA) \
809 		BIG_(0x02C2, 0x02CF) \
810 		BIG_(0x02D2, 0x02DF) \
811 		BIG_(0x02E5, 0x02FF) \
812 		PAIR(0x0374, 0x0375) \
813 		PAIR(0x037E, 0x037E) \
814 		PAIR(0x0384, 0x0385) \
815 		PAIR(0x0387, 0x0387) \
816 		PAIR(0x03F6, 0x03F6) \
817 		PAIR(0x058A, 0x058A) \
818 		PAIR(0x0600, 0x0603) \
819 		PAIR(0x0606, 0x0607) \
820 		PAIR(0x0609, 0x060A) \
821 		PAIR(0x060C, 0x060C) \
822 		PAIR(0x060E, 0x060F) \
823 		BIG_(0x0660, 0x066C) \
824 		PAIR(0x06DD, 0x06DD) \
825 		PAIR(0x06E9, 0x06E9) \
826 		BIG_(0x06F0, 0x06F9) \
827 		PAIR(0x07F6, 0x07F9) \
828 		PAIR(0x09F2, 0x09F3) \
829 		PAIR(0x09FB, 0x09FB) \
830 		PAIR(0x0AF1, 0x0AF1) \
831 		BIG_(0x0BF3, 0x0BFA) \
832 		BIG_(0x0C78, 0x0C7E) \
833 		PAIR(0x0CF1, 0x0CF2) \
834 		PAIR(0x0E3F, 0x0E3F) \
835 		PAIR(0x0F3A, 0x0F3D) \
836 		BIG_(0x1390, 0x1400) \
837 		PAIR(0x1680, 0x1680) \
838 		PAIR(0x169B, 0x169C) \
839 		PAIR(0x17DB, 0x17DB) \
840 		BIG_(0x17F0, 0x17F9) \
841 		BIG_(0x1800, 0x180A) \
842 		PAIR(0x180E, 0x180E) \
843 		PAIR(0x1940, 0x1940) \
844 		PAIR(0x1944, 0x1945) \
845 		BIG_(0x19DE, 0x19FF) \
846 		PAIR(0x1FBD, 0x1FBD) \
847 		PAIR(0x1FBF, 0x1FC1) \
848 		PAIR(0x1FCD, 0x1FCF) \
849 		PAIR(0x1FDD, 0x1FDF) \
850 		PAIR(0x1FED, 0x1FEF) \
851 		PAIR(0x1FFD, 0x1FFE) \
852 		BIG_(0x2000, 0x200A) \
853 		BIG_(0x2010, 0x2029) \
854 		BIG_(0x202F, 0x205F) \
855 		PAIR(0x2070, 0x2070) \
856 		BIG_(0x2074, 0x207E) \
857 		BIG_(0x2080, 0x208E) \
858 		BIG_(0x20A0, 0x20B8) \
859 		PAIR(0x2100, 0x2101) \
860 		PAIR(0x2103, 0x2106) \
861 		PAIR(0x2108, 0x2109) \
862 		PAIR(0x2114, 0x2114) \
863 		PAIR(0x2116, 0x2118) \
864 		BIG_(0x211E, 0x2123) \
865 		PAIR(0x2125, 0x2125) \
866 		PAIR(0x2127, 0x2127) \
867 		PAIR(0x2129, 0x2129) \
868 		PAIR(0x212E, 0x212E) \
869 		PAIR(0x213A, 0x213B) \
870 		BIG_(0x2140, 0x2144) \
871 		PAIR(0x214A, 0x214D) \
872 		BIG_(0x2150, 0x215F) \
873 		PAIR(0x2189, 0x2189) \
874 		BIG_(0x2190, 0x2335) \
875 		BIG_(0x237B, 0x2394) \
876 		BIG_(0x2396, 0x23E8) \
877 		BIG_(0x2400, 0x2426) \
878 		BIG_(0x2440, 0x244A) \
879 		BIG_(0x2460, 0x249B) \
880 		BIG_(0x24EA, 0x26AB) \
881 		BIG_(0x26AD, 0x26CD) \
882 		BIG_(0x26CF, 0x26E1) \
883 		PAIR(0x26E3, 0x26E3) \
884 		BIG_(0x26E8, 0x26FF) \
885 		PAIR(0x2701, 0x2704) \
886 		PAIR(0x2706, 0x2709) \
887 		BIG_(0x270C, 0x2727) \
888 		BIG_(0x2729, 0x274B) \
889 		PAIR(0x274D, 0x274D) \
890 		PAIR(0x274F, 0x2752) \
891 		BIG_(0x2756, 0x275E) \
892 		BIG_(0x2761, 0x2794) \
893 		BIG_(0x2798, 0x27AF) \
894 		BIG_(0x27B1, 0x27BE) \
895 		BIG_(0x27C0, 0x27CA) \
896 		PAIR(0x27CC, 0x27CC) \
897 		BIG_(0x27D0, 0x27FF) \
898 		BIG_(0x2900, 0x2B4C) \
899 		BIG_(0x2B50, 0x2B59) \
900 		BIG_(0x2CE5, 0x2CEA) \
901 		BIG_(0x2CF9, 0x2CFF) \
902 		BIG_(0x2E00, 0x2E99) \
903 		BIG_(0x2E9B, 0x2EF3) \
904 		BIG_(0x2F00, 0x2FD5) \
905 		BIG_(0x2FF0, 0x2FFB) \
906 		BIG_(0x3000, 0x3004) \
907 		BIG_(0x3008, 0x3020) \
908 		PAIR(0x3030, 0x3030) \
909 		PAIR(0x3036, 0x3037) \
910 		PAIR(0x303D, 0x303D) \
911 		PAIR(0x303E, 0x303F) \
912 		PAIR(0x309B, 0x309C) \
913 		PAIR(0x30A0, 0x30A0) \
914 		PAIR(0x30FB, 0x30FB) \
915 		BIG_(0x31C0, 0x31E3) \
916 		PAIR(0x321D, 0x321E) \
917 		BIG_(0x3250, 0x325F) \
918 		PAIR(0x327C, 0x327E) \
919 		BIG_(0x32B1, 0x32BF) \
920 		PAIR(0x32CC, 0x32CF) \
921 		PAIR(0x3377, 0x337A) \
922 		PAIR(0x33DE, 0x33DF) \
923 		PAIR(0x33FF, 0x33FF) \
924 		BIG_(0x4DC0, 0x4DFF) \
925 		BIG_(0xA490, 0xA4C6) \
926 		BIG_(0xA60D, 0xA60F) \
927 		BIG_(0xA673, 0xA673) \
928 		BIG_(0xA67E, 0xA67F) \
929 		BIG_(0xA700, 0xA721) \
930 		BIG_(0xA788, 0xA788) \
931 		BIG_(0xA828, 0xA82B) \
932 		BIG_(0xA838, 0xA839) \
933 		BIG_(0xA874, 0xA877) \
934 		BIG_(0xFB29, 0xFB29) \
935 		BIG_(0xFD3E, 0xFD3F) \
936 		BIG_(0xFDFD, 0xFDFD) \
937 		BIG_(0xFE10, 0xFE19) \
938 		BIG_(0xFE30, 0xFE52) \
939 		BIG_(0xFE54, 0xFE66) \
940 		BIG_(0xFE68, 0xFE6B) \
941 		BIG_(0xFF01, 0xFF20) \
942 		BIG_(0xFF3B, 0xFF40) \
943 		BIG_(0xFF5B, 0xFF65) \
944 		BIG_(0xFFE0, 0xFFE6) \
945 		BIG_(0xFFE8, 0xFFEE) \
946 		BIG_(0xFFF9, 0xFFFD)
947 		/*
948 		{0x10101, 0x10101},
949 		{0x10140, 0x1019B},
950 		{0x1091F, 0x1091F},
951 		{0x10B39, 0x10B3F},
952 		{0x10E60, 0x10E7E},
953 		{0x1D200, 0x1D241},
954 		{0x1D245, 0x1D245},
955 		{0x1D300, 0x1D356},
956 		{0x1D6DB, 0x1D6DB},
957 		{0x1D715, 0x1D715},
958 		{0x1D74F, 0x1D74F},
959 		{0x1D789, 0x1D789},
960 		{0x1D7C3, 0x1D7C3},
961 		{0x1D7CE, 0x1D7FF},
962 		{0x1F000, 0x1F02B},
963 		{0x1F030, 0x1F093},
964 		{0x1F100, 0x1F10A}
965 		*/
966 	static const struct interval neutral_b[] ALIGN4 = { ARRAY };
967 #  undef BIG_
968 #  undef PAIR
969 #  define BIG_(a,b)
970 #  define PAIR(a,b) (a << 2) | (b-a),
971 	static const uint16_t neutral_p[] ALIGN2 = { ARRAY };
972 #  undef BIG_
973 #  undef PAIR
974 #  define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1];
975 #  define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1];
976 	struct CHECK { ARRAY };
977 #  undef BIG_
978 #  undef PAIR
979 #  undef ARRAY
980 
981 	if (in_interval_table(wc, neutral_b, ARRAY_SIZE(neutral_b) - 1))
982 		return 1;
983 	if (in_uint16_table(wc, neutral_p, ARRAY_SIZE(neutral_p) - 1))
984 		return 1;
985 	return 0;
986 }
987 #  endif
988 
989 # endif /* UNICODE_BIDI_SUPPORT */
990 
991 #endif /* Homegrown Unicode support */
992 
993 
994 /* The rest is mostly same for libc and for "homegrown" support */
995 
unicode_strlen(const char * string)996 size_t FAST_FUNC unicode_strlen(const char *string)
997 {
998 	size_t width = mbstowcs(NULL, string, INT_MAX);
999 	if (width == (size_t)-1L)
1000 		return strlen(string);
1001 	return width;
1002 }
1003 
unicode_strwidth(const char * string)1004 size_t FAST_FUNC unicode_strwidth(const char *string)
1005 {
1006 	uni_stat_t uni_stat;
1007 	printable_string2(&uni_stat, string);
1008 	return uni_stat.unicode_width;
1009 }
1010 
unicode_conv_to_printable2(uni_stat_t * stats,const char * src,unsigned width,int flags)1011 static char* FAST_FUNC unicode_conv_to_printable2(uni_stat_t *stats, const char *src, unsigned width, int flags)
1012 {
1013 	char *dst;
1014 	unsigned dst_len;
1015 	unsigned uni_count;
1016 	unsigned uni_width;
1017 
1018 	if (unicode_status != UNICODE_ON) {
1019 		char *d;
1020 		if (flags & UNI_FLAG_PAD) {
1021 			d = dst = xmalloc(width + 1);
1022 			while ((int)--width >= 0) {
1023 				unsigned char c = *src;
1024 				if (c == '\0') {
1025 					do
1026 						*d++ = ' ';
1027 					while ((int)--width >= 0);
1028 					break;
1029 				}
1030 				*d++ = (c >= ' ' && c < 0x7f) ? c : '?';
1031 				src++;
1032 			}
1033 			*d = '\0';
1034 		} else {
1035 			d = dst = xstrndup(src, width);
1036 			while (*d) {
1037 				unsigned char c = *d;
1038 				if (c < ' ' || c >= 0x7f)
1039 					*d = '?';
1040 				d++;
1041 			}
1042 		}
1043 		if (stats) {
1044 			stats->byte_count = (d - dst);
1045 			stats->unicode_count = (d - dst);
1046 			stats->unicode_width = (d - dst);
1047 		}
1048 		return dst;
1049 	}
1050 
1051 	dst = NULL;
1052 	uni_count = uni_width = 0;
1053 	dst_len = 0;
1054 	while (1) {
1055 		int w;
1056 		wchar_t wc;
1057 
1058 #if ENABLE_UNICODE_USING_LOCALE
1059 		{
1060 			mbstate_t mbst = { 0 };
1061 			ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst);
1062 			/* If invalid sequence is seen: -1 is returned,
1063 			 * src points to the invalid sequence, errno = EILSEQ.
1064 			 * Else number of wchars (excluding terminating L'\0')
1065 			 * written to dest is returned.
1066 			 * If len (here: 1) non-L'\0' wchars stored at dest,
1067 			 * src points to the next char to be converted.
1068 			 * If string is completely converted: src = NULL.
1069 			 */
1070 			if (rc == 0) /* end-of-string */
1071 				break;
1072 			if (rc < 0) { /* error */
1073 				src++;
1074 				goto subst;
1075 			}
1076 			if (!iswprint(wc))
1077 				goto subst;
1078 		}
1079 #else
1080 		src = mbstowc_internal(&wc, src);
1081 		/* src is advanced to next mb char
1082 		 * wc == ERROR_WCHAR: invalid sequence is seen
1083 		 * else: wc is set
1084 		 */
1085 		if (wc == ERROR_WCHAR) /* error */
1086 			goto subst;
1087 		if (wc == 0) /* end-of-string */
1088 			break;
1089 #endif
1090 		if (CONFIG_LAST_SUPPORTED_WCHAR && wc > CONFIG_LAST_SUPPORTED_WCHAR)
1091 			goto subst;
1092 		w = wcwidth(wc);
1093 		if ((ENABLE_UNICODE_COMBINING_WCHARS && w < 0) /* non-printable wchar */
1094 		 || (!ENABLE_UNICODE_COMBINING_WCHARS && w <= 0)
1095 		 || (!ENABLE_UNICODE_WIDE_WCHARS && w > 1)
1096 		) {
1097  subst:
1098 			wc = CONFIG_SUBST_WCHAR;
1099 			w = 1;
1100 		}
1101 		width -= w;
1102 		/* Note: if width == 0, we still may add more chars,
1103 		 * they may be zero-width or combining ones */
1104 		if ((int)width < 0) {
1105 			/* can't add this wc, string would become longer than width */
1106 			width += w;
1107 			break;
1108 		}
1109 
1110 		uni_count++;
1111 		uni_width += w;
1112 		dst = xrealloc(dst, dst_len + MB_CUR_MAX);
1113 #if ENABLE_UNICODE_USING_LOCALE
1114 		{
1115 			mbstate_t mbst = { 0 };
1116 			dst_len += wcrtomb(&dst[dst_len], wc, &mbst);
1117 		}
1118 #else
1119 		dst_len += wcrtomb_internal(&dst[dst_len], wc);
1120 #endif
1121 	}
1122 
1123 	/* Pad to remaining width */
1124 	if (flags & UNI_FLAG_PAD) {
1125 		dst = xrealloc(dst, dst_len + width + 1);
1126 		uni_count += width;
1127 		uni_width += width;
1128 		while ((int)--width >= 0) {
1129 			dst[dst_len++] = ' ';
1130 		}
1131 	}
1132 	if (!dst) /* for example, if input was "" */
1133 		dst = xzalloc(1);
1134 	dst[dst_len] = '\0';
1135 	if (stats) {
1136 		stats->byte_count = dst_len;
1137 		stats->unicode_count = uni_count;
1138 		stats->unicode_width = uni_width;
1139 	}
1140 
1141 	return dst;
1142 }
unicode_conv_to_printable(uni_stat_t * stats,const char * src)1143 char* FAST_FUNC unicode_conv_to_printable(uni_stat_t *stats, const char *src)
1144 {
1145 	return unicode_conv_to_printable2(stats, src, INT_MAX, 0);
1146 }
unicode_conv_to_printable_fixedwidth(const char * src,unsigned width)1147 char* FAST_FUNC unicode_conv_to_printable_fixedwidth(/*uni_stat_t *stats,*/ const char *src, unsigned width)
1148 {
1149 	return unicode_conv_to_printable2(/*stats:*/ NULL, src, width, UNI_FLAG_PAD);
1150 }
1151 
1152 #ifdef UNUSED
unicode_conv_to_printable_maxwidth(uni_stat_t * stats,const char * src,unsigned maxwidth)1153 char* FAST_FUNC unicode_conv_to_printable_maxwidth(uni_stat_t *stats, const char *src, unsigned maxwidth)
1154 {
1155 	return unicode_conv_to_printable2(stats, src, maxwidth, 0);
1156 }
1157 
unicode_padding_to_width(unsigned width,const char * src)1158 unsigned FAST_FUNC unicode_padding_to_width(unsigned width, const char *src)
1159 {
1160 	if (unicode_status != UNICODE_ON) {
1161 		return width - strnlen(src, width);
1162 	}
1163 
1164 	while (1) {
1165 		int w;
1166 		wchar_t wc;
1167 
1168 #if ENABLE_UNICODE_USING_LOCALE
1169 		{
1170 			mbstate_t mbst = { 0 };
1171 			ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst);
1172 			if (rc <= 0) /* error, or end-of-string */
1173 				return width;
1174 		}
1175 #else
1176 		src = mbstowc_internal(&wc, src);
1177 		if (wc == ERROR_WCHAR || wc == 0) /* error, or end-of-string */
1178 			return width;
1179 #endif
1180 		w = wcwidth(wc);
1181 		if (w < 0) /* non-printable wchar */
1182 			return width;
1183 		width -= w;
1184 		if ((int)width <= 0) /* string is longer than width */
1185 			return 0;
1186 	}
1187 }
1188 #endif
1189