blob: 6bdb666c9cc98dae24479fb03673c1baa5282843
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 | #ifdef __BIONIC__ |
18 | # define VOID |
19 | #else |
20 | # define VOID void |
21 | #endif |
22 | |
23 | /* This file is compiled only if UNICODE_SUPPORT is on. |
24 | * We check other options and decide whether to use libc support |
25 | * via locale, or use our own logic: |
26 | */ |
27 | |
28 | #if ENABLE_UNICODE_USING_LOCALE |
29 | |
30 | /* Unicode support using libc locale support. */ |
31 | |
32 | void FAST_FUNC reinit_unicode(const char *LANG) |
33 | { |
34 | static const char unicode_0x394[] = { 0xce, 0x94, 0 }; |
35 | size_t width; |
36 | |
37 | /* We pass "" instead of "C" because some libc's have |
38 | * non-ASCII default locale for setlocale("") call |
39 | * (this allows users of such libc to have Unicoded |
40 | * system without having to mess with env). |
41 | * |
42 | * We set LC_CTYPE because (a) we may be called with $LC_CTYPE |
43 | * value in LANG, not with $LC_ALL, (b) internationalized |
44 | * LC_NUMERIC and LC_TIME are more PITA than benefit |
45 | * (for one, some utilities have hard time with comma |
46 | * used as a fractional separator). |
47 | */ |
48 | //TODO: avoid repeated calls by caching last string? |
49 | setlocale(LC_CTYPE, LANG ? LANG : ""); |
50 | |
51 | /* In unicode, this is a one character string */ |
52 | width = unicode_strlen(unicode_0x394); |
53 | unicode_status = (width == 1 ? UNICODE_ON : UNICODE_OFF); |
54 | } |
55 | |
56 | void FAST_FUNC init_unicode(VOID) |
57 | { |
58 | /* Some people set only $LC_CTYPE, not $LC_ALL, because they want |
59 | * only Unicode to be activated on their system, not the whole |
60 | * shebang of wrong decimal points, strange date formats and so on. |
61 | */ |
62 | if (unicode_status == UNICODE_UNKNOWN) { |
63 | char *s = getenv("LC_ALL"); |
64 | if (!s) s = getenv("LC_CTYPE"); |
65 | if (!s) s = getenv("LANG"); |
66 | reinit_unicode(s); |
67 | } |
68 | } |
69 | |
70 | #else |
71 | |
72 | /* Homegrown Unicode support. It knows only C and Unicode locales. */ |
73 | |
74 | # if ENABLE_FEATURE_CHECK_UNICODE_IN_ENV |
75 | void FAST_FUNC reinit_unicode(const char *LANG) |
76 | { |
77 | unicode_status = UNICODE_OFF; |
78 | if (!LANG || !(strstr(LANG, ".utf") || strstr(LANG, ".UTF"))) |
79 | return; |
80 | unicode_status = UNICODE_ON; |
81 | } |
82 | |
83 | void FAST_FUNC init_unicode(VOID) |
84 | { |
85 | if (unicode_status == UNICODE_UNKNOWN) { |
86 | char *s = getenv("LC_ALL"); |
87 | if (!s) s = getenv("LC_CTYPE"); |
88 | if (!s) s = getenv("LANG"); |
89 | reinit_unicode(s); |
90 | } |
91 | } |
92 | # endif |
93 | |
94 | static size_t wcrtomb_internal(char *s, wchar_t wc) |
95 | { |
96 | int n, i; |
97 | uint32_t v = wc; |
98 | |
99 | if (v <= 0x7f) { |
100 | *s = v; |
101 | return 1; |
102 | } |
103 | |
104 | /* RFC 3629 says that Unicode ends at 10FFFF, |
105 | * but we cover entire 32 bits */ |
106 | |
107 | /* 4000000-FFFFFFFF -> 111111tt 10tttttt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */ |
108 | /* 200000-3FFFFFF -> 111110tt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */ |
109 | /* 10000-1FFFFF -> 11110zzz 10zzyyyy 10yyyyxx 10xxxxxx */ |
110 | /* 800-FFFF -> 1110yyyy 10yyyyxx 10xxxxxx */ |
111 | /* 80-7FF -> 110yyyxx 10xxxxxx */ |
112 | |
113 | /* How many bytes do we need? */ |
114 | n = 2; |
115 | /* (0x80000000+ would result in n = 7, limiting n to 6) */ |
116 | while (v >= 0x800 && n < 6) { |
117 | v >>= 5; |
118 | n++; |
119 | } |
120 | /* Fill bytes n-1..1 */ |
121 | i = n; |
122 | while (--i) { |
123 | s[i] = (wc & 0x3f) | 0x80; |
124 | wc >>= 6; |
125 | } |
126 | /* Fill byte 0 */ |
127 | s[0] = wc | (uint8_t)(0x3f00 >> n); |
128 | return n; |
129 | } |
130 | size_t FAST_FUNC wcrtomb(char *s, wchar_t wc, mbstate_t *ps UNUSED_PARAM) |
131 | { |
132 | if (unicode_status != UNICODE_ON) { |
133 | *s = wc; |
134 | return 1; |
135 | } |
136 | |
137 | return wcrtomb_internal(s, wc); |
138 | } |
139 | size_t FAST_FUNC wcstombs(char *dest, const wchar_t *src, size_t n) |
140 | { |
141 | size_t org_n = n; |
142 | |
143 | if (unicode_status != UNICODE_ON) { |
144 | while (n) { |
145 | wchar_t c = *src++; |
146 | *dest++ = c; |
147 | if (c == 0) |
148 | break; |
149 | n--; |
150 | } |
151 | return org_n - n; |
152 | } |
153 | |
154 | while (n >= MB_CUR_MAX) { |
155 | wchar_t wc = *src++; |
156 | size_t len = wcrtomb_internal(dest, wc); |
157 | |
158 | if (wc == L'\0') |
159 | return org_n - n; |
160 | dest += len; |
161 | n -= len; |
162 | } |
163 | while (n) { |
164 | char tbuf[MB_CUR_MAX]; |
165 | wchar_t wc = *src++; |
166 | size_t len = wcrtomb_internal(tbuf, wc); |
167 | |
168 | if (len > n) |
169 | break; |
170 | memcpy(dest, tbuf, len); |
171 | if (wc == L'\0') |
172 | return org_n - n; |
173 | dest += len; |
174 | n -= len; |
175 | } |
176 | return org_n - n; |
177 | } |
178 | |
179 | # define ERROR_WCHAR (~(wchar_t)0) |
180 | |
181 | static const char *mbstowc_internal(wchar_t *res, const char *src) |
182 | { |
183 | int bytes; |
184 | unsigned c = (unsigned char) *src++; |
185 | |
186 | if (c <= 0x7f) { |
187 | *res = c; |
188 | return src; |
189 | } |
190 | |
191 | /* 80-7FF -> 110yyyxx 10xxxxxx */ |
192 | /* 800-FFFF -> 1110yyyy 10yyyyxx 10xxxxxx */ |
193 | /* 10000-1FFFFF -> 11110zzz 10zzyyyy 10yyyyxx 10xxxxxx */ |
194 | /* 200000-3FFFFFF -> 111110tt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */ |
195 | /* 4000000-FFFFFFFF -> 111111tt 10tttttt 10zzzzzz 10zzyyyy 10yyyyxx 10xxxxxx */ |
196 | bytes = 0; |
197 | do { |
198 | c <<= 1; |
199 | bytes++; |
200 | } while ((c & 0x80) && bytes < 6); |
201 | if (bytes == 1) { |
202 | /* A bare "continuation" byte. Say, 80 */ |
203 | *res = ERROR_WCHAR; |
204 | return src; |
205 | } |
206 | c = (uint8_t)(c) >> bytes; |
207 | |
208 | while (--bytes) { |
209 | unsigned ch = (unsigned char) *src; |
210 | if ((ch & 0xc0) != 0x80) { |
211 | /* Missing "continuation" byte. Example: e0 80 */ |
212 | *res = ERROR_WCHAR; |
213 | return src; |
214 | } |
215 | c = (c << 6) + (ch & 0x3f); |
216 | src++; |
217 | } |
218 | |
219 | /* TODO */ |
220 | /* Need to check that c isn't produced by overlong encoding */ |
221 | /* Example: 11000000 10000000 converts to NUL */ |
222 | /* 11110000 10000000 10000100 10000000 converts to 0x100 */ |
223 | /* correct encoding: 11000100 10000000 */ |
224 | if (c <= 0x7f) { /* crude check */ |
225 | *res = ERROR_WCHAR; |
226 | return src; |
227 | } |
228 | |
229 | *res = c; |
230 | return src; |
231 | } |
232 | size_t FAST_FUNC mbstowcs(wchar_t *dest, const char *src, size_t n) |
233 | { |
234 | size_t org_n = n; |
235 | |
236 | if (!src) return 0; |
237 | |
238 | if (unicode_status != UNICODE_ON) { |
239 | while (n) { |
240 | unsigned char c = *src++; |
241 | |
242 | if (dest) |
243 | *dest++ = c; |
244 | if (c == 0) |
245 | break; |
246 | n--; |
247 | } |
248 | return org_n - n; |
249 | } |
250 | |
251 | while (n) { |
252 | wchar_t wc; |
253 | src = mbstowc_internal(&wc, src); |
254 | if (wc == ERROR_WCHAR) /* error */ |
255 | return (size_t) -1L; |
256 | if (dest) |
257 | *dest++ = wc; |
258 | if (wc == 0) /* end-of-string */ |
259 | break; |
260 | n--; |
261 | } |
262 | |
263 | return org_n - n; |
264 | } |
265 | |
266 | int FAST_FUNC iswspace(wint_t wc) |
267 | { |
268 | return (unsigned)wc <= 0x7f && isspace(wc); |
269 | } |
270 | |
271 | int FAST_FUNC iswalnum(wint_t wc) |
272 | { |
273 | return (unsigned)wc <= 0x7f && isalnum(wc); |
274 | } |
275 | |
276 | int FAST_FUNC iswpunct(wint_t wc) |
277 | { |
278 | return (unsigned)wc <= 0x7f && ispunct(wc); |
279 | } |
280 | |
281 | |
282 | # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300 |
283 | struct interval { |
284 | uint16_t first; |
285 | uint16_t last; |
286 | }; |
287 | |
288 | /* auxiliary function for binary search in interval table */ |
289 | static int in_interval_table(unsigned ucs, const struct interval *table, unsigned max) |
290 | { |
291 | unsigned min; |
292 | unsigned mid; |
293 | |
294 | if (ucs < table[0].first || ucs > table[max].last) |
295 | return 0; |
296 | |
297 | min = 0; |
298 | while (max >= min) { |
299 | mid = (min + max) / 2; |
300 | if (ucs > table[mid].last) |
301 | min = mid + 1; |
302 | else if (ucs < table[mid].first) |
303 | max = mid - 1; |
304 | else |
305 | return 1; |
306 | } |
307 | return 0; |
308 | } |
309 | |
310 | static int in_uint16_table(unsigned ucs, const uint16_t *table, unsigned max) |
311 | { |
312 | unsigned min; |
313 | unsigned mid; |
314 | unsigned first, last; |
315 | |
316 | first = table[0] >> 2; |
317 | last = first + (table[0] & 3); |
318 | if (ucs < first || ucs > last) |
319 | return 0; |
320 | |
321 | min = 0; |
322 | while (max >= min) { |
323 | mid = (min + max) / 2; |
324 | first = table[mid] >> 2; |
325 | last = first + (table[mid] & 3); |
326 | if (ucs > last) |
327 | min = mid + 1; |
328 | else if (ucs < first) |
329 | max = mid - 1; |
330 | else |
331 | return 1; |
332 | } |
333 | return 0; |
334 | } |
335 | # endif |
336 | |
337 | |
338 | /* |
339 | * This is an implementation of wcwidth() and wcswidth() (defined in |
340 | * IEEE Std 1002.1-2001) for Unicode. |
341 | * |
342 | * http://www.opengroup.org/onlinepubs/007904975/functions/wcwidth.html |
343 | * http://www.opengroup.org/onlinepubs/007904975/functions/wcswidth.html |
344 | * |
345 | * In fixed-width output devices, Latin characters all occupy a single |
346 | * "cell" position of equal width, whereas ideographic CJK characters |
347 | * occupy two such cells. Interoperability between terminal-line |
348 | * applications and (teletype-style) character terminals using the |
349 | * UTF-8 encoding requires agreement on which character should advance |
350 | * the cursor by how many cell positions. No established formal |
351 | * standards exist at present on which Unicode character shall occupy |
352 | * how many cell positions on character terminals. These routines are |
353 | * a first attempt of defining such behavior based on simple rules |
354 | * applied to data provided by the Unicode Consortium. |
355 | * |
356 | * For some graphical characters, the Unicode standard explicitly |
357 | * defines a character-cell width via the definition of the East Asian |
358 | * FullWidth (F), Wide (W), Half-width (H), and Narrow (Na) classes. |
359 | * In all these cases, there is no ambiguity about which width a |
360 | * terminal shall use. For characters in the East Asian Ambiguous (A) |
361 | * class, the width choice depends purely on a preference of backward |
362 | * compatibility with either historic CJK or Western practice. |
363 | * Choosing single-width for these characters is easy to justify as |
364 | * the appropriate long-term solution, as the CJK practice of |
365 | * displaying these characters as double-width comes from historic |
366 | * implementation simplicity (8-bit encoded characters were displayed |
367 | * single-width and 16-bit ones double-width, even for Greek, |
368 | * Cyrillic, etc.) and not any typographic considerations. |
369 | * |
370 | * Much less clear is the choice of width for the Not East Asian |
371 | * (Neutral) class. Existing practice does not dictate a width for any |
372 | * of these characters. It would nevertheless make sense |
373 | * typographically to allocate two character cells to characters such |
374 | * as for instance EM SPACE or VOLUME INTEGRAL, which cannot be |
375 | * represented adequately with a single-width glyph. The following |
376 | * routines at present merely assign a single-cell width to all |
377 | * neutral characters, in the interest of simplicity. This is not |
378 | * entirely satisfactory and should be reconsidered before |
379 | * establishing a formal standard in this area. At the moment, the |
380 | * decision which Not East Asian (Neutral) characters should be |
381 | * represented by double-width glyphs cannot yet be answered by |
382 | * applying a simple rule from the Unicode database content. Setting |
383 | * up a proper standard for the behavior of UTF-8 character terminals |
384 | * will require a careful analysis not only of each Unicode character, |
385 | * but also of each presentation form, something the author of these |
386 | * routines has avoided to do so far. |
387 | * |
388 | * http://www.unicode.org/unicode/reports/tr11/ |
389 | * |
390 | * Markus Kuhn -- 2007-05-26 (Unicode 5.0) |
391 | * |
392 | * Permission to use, copy, modify, and distribute this software |
393 | * for any purpose and without fee is hereby granted. The author |
394 | * disclaims all warranties with regard to this software. |
395 | * |
396 | * Latest version: http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c |
397 | */ |
398 | |
399 | /* Assigned Unicode character ranges: |
400 | * Plane Range |
401 | * 0 0000–FFFF Basic Multilingual Plane |
402 | * 1 10000–1FFFF Supplementary Multilingual Plane |
403 | * 2 20000–2FFFF Supplementary Ideographic Plane |
404 | * 3 30000-3FFFF Tertiary Ideographic Plane (no chars assigned yet) |
405 | * 4-13 40000–DFFFF currently unassigned |
406 | * 14 E0000–EFFFF Supplementary Special-purpose Plane |
407 | * 15 F0000–FFFFF Supplementary Private Use Area-A |
408 | * 16 100000–10FFFF Supplementary Private Use Area-B |
409 | * |
410 | * "Supplementary Special-purpose Plane currently contains non-graphical |
411 | * characters in two blocks of 128 and 240 characters. The first block |
412 | * is for language tag characters for use when language cannot be indicated |
413 | * through other protocols (such as the xml:lang attribute in XML). |
414 | * The other block contains glyph variation selectors to indicate |
415 | * an alternate glyph for a character that cannot be determined by context." |
416 | * |
417 | * In simpler terms: it is a tool to fix the "Han unification" mess |
418 | * created by Unicode committee, to select Chinese/Japanese/Korean/Taiwan |
419 | * version of a character. (They forgot that the whole purpose of the Unicode |
420 | * was to be able to write all chars in one charset without such tricks). |
421 | * Until East Asian users say it is actually necessary to support these |
422 | * code points in console applications like busybox |
423 | * (i.e. do these chars ever appear in filenames, hostnames, text files |
424 | * and such?), we are treating these code points as invalid. |
425 | * |
426 | * Tertiary Ideographic Plane is also ignored for now, |
427 | * until Unicode committee assigns something there. |
428 | */ |
429 | /* The following two functions define the column width of an ISO 10646 |
430 | * character as follows: |
431 | * |
432 | * - The null character (U+0000) has a column width of 0. |
433 | * |
434 | * - Other C0/C1 control characters and DEL will lead to a return |
435 | * value of -1. |
436 | * |
437 | * - Non-spacing and enclosing combining characters (general |
438 | * category code Mn or Me in the Unicode database) have a |
439 | * column width of 0. |
440 | * |
441 | * - SOFT HYPHEN (U+00AD) has a column width of 1. |
442 | * |
443 | * - Other format characters (general category code Cf in the Unicode |
444 | * database) and ZERO WIDTH SPACE (U+200B) have a column width of 0. |
445 | * |
446 | * - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF) |
447 | * have a column width of 0. |
448 | * |
449 | * - Spacing characters in the East Asian Wide (W) or East Asian |
450 | * Full-width (F) category as defined in Unicode Technical |
451 | * Report #11 have a column width of 2. |
452 | * |
453 | * - All remaining characters (including all printable |
454 | * ISO 8859-1 and WGL4 characters, Unicode control characters, |
455 | * etc.) have a column width of 1. |
456 | * |
457 | * This implementation assumes that wchar_t characters are encoded |
458 | * in ISO 10646. |
459 | */ |
460 | int FAST_FUNC wcwidth(unsigned ucs) |
461 | { |
462 | # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300 |
463 | /* sorted list of non-overlapping intervals of non-spacing characters */ |
464 | /* generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c" */ |
465 | # define BIG_(a,b) { a, b }, |
466 | # define PAIR(a,b) |
467 | # define ARRAY /* PAIR if < 0x4000 and no more than 4 chars big */ \ |
468 | BIG_(0x0300, 0x036F) \ |
469 | PAIR(0x0483, 0x0486) \ |
470 | PAIR(0x0488, 0x0489) \ |
471 | BIG_(0x0591, 0x05BD) \ |
472 | PAIR(0x05BF, 0x05BF) \ |
473 | PAIR(0x05C1, 0x05C2) \ |
474 | PAIR(0x05C4, 0x05C5) \ |
475 | PAIR(0x05C7, 0x05C7) \ |
476 | PAIR(0x0600, 0x0603) \ |
477 | BIG_(0x0610, 0x0615) \ |
478 | BIG_(0x064B, 0x065E) \ |
479 | PAIR(0x0670, 0x0670) \ |
480 | BIG_(0x06D6, 0x06E4) \ |
481 | PAIR(0x06E7, 0x06E8) \ |
482 | PAIR(0x06EA, 0x06ED) \ |
483 | PAIR(0x070F, 0x070F) \ |
484 | PAIR(0x0711, 0x0711) \ |
485 | BIG_(0x0730, 0x074A) \ |
486 | BIG_(0x07A6, 0x07B0) \ |
487 | BIG_(0x07EB, 0x07F3) \ |
488 | PAIR(0x0901, 0x0902) \ |
489 | PAIR(0x093C, 0x093C) \ |
490 | BIG_(0x0941, 0x0948) \ |
491 | PAIR(0x094D, 0x094D) \ |
492 | PAIR(0x0951, 0x0954) \ |
493 | PAIR(0x0962, 0x0963) \ |
494 | PAIR(0x0981, 0x0981) \ |
495 | PAIR(0x09BC, 0x09BC) \ |
496 | PAIR(0x09C1, 0x09C4) \ |
497 | PAIR(0x09CD, 0x09CD) \ |
498 | PAIR(0x09E2, 0x09E3) \ |
499 | PAIR(0x0A01, 0x0A02) \ |
500 | PAIR(0x0A3C, 0x0A3C) \ |
501 | PAIR(0x0A41, 0x0A42) \ |
502 | PAIR(0x0A47, 0x0A48) \ |
503 | PAIR(0x0A4B, 0x0A4D) \ |
504 | PAIR(0x0A70, 0x0A71) \ |
505 | PAIR(0x0A81, 0x0A82) \ |
506 | PAIR(0x0ABC, 0x0ABC) \ |
507 | BIG_(0x0AC1, 0x0AC5) \ |
508 | PAIR(0x0AC7, 0x0AC8) \ |
509 | PAIR(0x0ACD, 0x0ACD) \ |
510 | PAIR(0x0AE2, 0x0AE3) \ |
511 | PAIR(0x0B01, 0x0B01) \ |
512 | PAIR(0x0B3C, 0x0B3C) \ |
513 | PAIR(0x0B3F, 0x0B3F) \ |
514 | PAIR(0x0B41, 0x0B43) \ |
515 | PAIR(0x0B4D, 0x0B4D) \ |
516 | PAIR(0x0B56, 0x0B56) \ |
517 | PAIR(0x0B82, 0x0B82) \ |
518 | PAIR(0x0BC0, 0x0BC0) \ |
519 | PAIR(0x0BCD, 0x0BCD) \ |
520 | PAIR(0x0C3E, 0x0C40) \ |
521 | PAIR(0x0C46, 0x0C48) \ |
522 | PAIR(0x0C4A, 0x0C4D) \ |
523 | PAIR(0x0C55, 0x0C56) \ |
524 | PAIR(0x0CBC, 0x0CBC) \ |
525 | PAIR(0x0CBF, 0x0CBF) \ |
526 | PAIR(0x0CC6, 0x0CC6) \ |
527 | PAIR(0x0CCC, 0x0CCD) \ |
528 | PAIR(0x0CE2, 0x0CE3) \ |
529 | PAIR(0x0D41, 0x0D43) \ |
530 | PAIR(0x0D4D, 0x0D4D) \ |
531 | PAIR(0x0DCA, 0x0DCA) \ |
532 | PAIR(0x0DD2, 0x0DD4) \ |
533 | PAIR(0x0DD6, 0x0DD6) \ |
534 | PAIR(0x0E31, 0x0E31) \ |
535 | BIG_(0x0E34, 0x0E3A) \ |
536 | BIG_(0x0E47, 0x0E4E) \ |
537 | PAIR(0x0EB1, 0x0EB1) \ |
538 | BIG_(0x0EB4, 0x0EB9) \ |
539 | PAIR(0x0EBB, 0x0EBC) \ |
540 | BIG_(0x0EC8, 0x0ECD) \ |
541 | PAIR(0x0F18, 0x0F19) \ |
542 | PAIR(0x0F35, 0x0F35) \ |
543 | PAIR(0x0F37, 0x0F37) \ |
544 | PAIR(0x0F39, 0x0F39) \ |
545 | BIG_(0x0F71, 0x0F7E) \ |
546 | BIG_(0x0F80, 0x0F84) \ |
547 | PAIR(0x0F86, 0x0F87) \ |
548 | PAIR(0x0FC6, 0x0FC6) \ |
549 | BIG_(0x0F90, 0x0F97) \ |
550 | BIG_(0x0F99, 0x0FBC) \ |
551 | PAIR(0x102D, 0x1030) \ |
552 | PAIR(0x1032, 0x1032) \ |
553 | PAIR(0x1036, 0x1037) \ |
554 | PAIR(0x1039, 0x1039) \ |
555 | PAIR(0x1058, 0x1059) \ |
556 | BIG_(0x1160, 0x11FF) \ |
557 | PAIR(0x135F, 0x135F) \ |
558 | PAIR(0x1712, 0x1714) \ |
559 | PAIR(0x1732, 0x1734) \ |
560 | PAIR(0x1752, 0x1753) \ |
561 | PAIR(0x1772, 0x1773) \ |
562 | PAIR(0x17B4, 0x17B5) \ |
563 | BIG_(0x17B7, 0x17BD) \ |
564 | PAIR(0x17C6, 0x17C6) \ |
565 | BIG_(0x17C9, 0x17D3) \ |
566 | PAIR(0x17DD, 0x17DD) \ |
567 | PAIR(0x180B, 0x180D) \ |
568 | PAIR(0x18A9, 0x18A9) \ |
569 | PAIR(0x1920, 0x1922) \ |
570 | PAIR(0x1927, 0x1928) \ |
571 | PAIR(0x1932, 0x1932) \ |
572 | PAIR(0x1939, 0x193B) \ |
573 | PAIR(0x1A17, 0x1A18) \ |
574 | PAIR(0x1B00, 0x1B03) \ |
575 | PAIR(0x1B34, 0x1B34) \ |
576 | BIG_(0x1B36, 0x1B3A) \ |
577 | PAIR(0x1B3C, 0x1B3C) \ |
578 | PAIR(0x1B42, 0x1B42) \ |
579 | BIG_(0x1B6B, 0x1B73) \ |
580 | BIG_(0x1DC0, 0x1DCA) \ |
581 | PAIR(0x1DFE, 0x1DFF) \ |
582 | BIG_(0x200B, 0x200F) \ |
583 | BIG_(0x202A, 0x202E) \ |
584 | PAIR(0x2060, 0x2063) \ |
585 | BIG_(0x206A, 0x206F) \ |
586 | BIG_(0x20D0, 0x20EF) \ |
587 | BIG_(0x302A, 0x302F) \ |
588 | PAIR(0x3099, 0x309A) \ |
589 | /* Too big to be packed in PAIRs: */ \ |
590 | BIG_(0xA806, 0xA806) \ |
591 | BIG_(0xA80B, 0xA80B) \ |
592 | BIG_(0xA825, 0xA826) \ |
593 | BIG_(0xFB1E, 0xFB1E) \ |
594 | BIG_(0xFE00, 0xFE0F) \ |
595 | BIG_(0xFE20, 0xFE23) \ |
596 | BIG_(0xFEFF, 0xFEFF) \ |
597 | BIG_(0xFFF9, 0xFFFB) |
598 | static const struct interval combining[] = { ARRAY }; |
599 | # undef BIG_ |
600 | # undef PAIR |
601 | # define BIG_(a,b) |
602 | # define PAIR(a,b) (a << 2) | (b-a), |
603 | static const uint16_t combining1[] = { ARRAY }; |
604 | # undef BIG_ |
605 | # undef PAIR |
606 | # define BIG_(a,b) char big_##a[b < 0x4000 && b-a <= 3 ? -1 : 1]; |
607 | # define PAIR(a,b) char pair##a[b >= 0x4000 || b-a > 3 ? -1 : 1]; |
608 | struct CHECK { ARRAY }; |
609 | # undef BIG_ |
610 | # undef PAIR |
611 | # undef ARRAY |
612 | # endif |
613 | |
614 | if (ucs == 0) |
615 | return 0; |
616 | |
617 | /* Test for 8-bit control characters (00-1f, 80-9f, 7f) */ |
618 | if ((ucs & ~0x80) < 0x20 || ucs == 0x7f) |
619 | return -1; |
620 | /* Quick abort if it is an obviously invalid char */ |
621 | if (ucs > CONFIG_LAST_SUPPORTED_WCHAR) |
622 | return -1; |
623 | |
624 | /* Optimization: no combining chars below 0x300 */ |
625 | if (CONFIG_LAST_SUPPORTED_WCHAR < 0x300 || ucs < 0x300) |
626 | return 1; |
627 | |
628 | # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x300 |
629 | /* Binary search in table of non-spacing characters */ |
630 | if (in_interval_table(ucs, combining, ARRAY_SIZE(combining) - 1)) |
631 | return 0; |
632 | if (in_uint16_table(ucs, combining1, ARRAY_SIZE(combining1) - 1)) |
633 | return 0; |
634 | |
635 | /* Optimization: all chars below 0x1100 are not double-width */ |
636 | if (CONFIG_LAST_SUPPORTED_WCHAR < 0x1100 || ucs < 0x1100) |
637 | return 1; |
638 | |
639 | # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x1100 |
640 | /* Invalid code points: */ |
641 | /* High (d800..dbff) and low (dc00..dfff) surrogates (valid only in UTF16) */ |
642 | /* Private Use Area (e000..f8ff) */ |
643 | /* Noncharacters fdd0..fdef */ |
644 | if ((CONFIG_LAST_SUPPORTED_WCHAR >= 0xd800 && ucs >= 0xd800 && ucs <= 0xf8ff) |
645 | || (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfdd0 && ucs >= 0xfdd0 && ucs <= 0xfdef) |
646 | ) { |
647 | return -1; |
648 | } |
649 | /* 0xfffe and 0xffff in every plane are invalid */ |
650 | if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xfffe && ((ucs & 0xfffe) == 0xfffe)) { |
651 | return -1; |
652 | } |
653 | |
654 | # if CONFIG_LAST_SUPPORTED_WCHAR >= 0x10000 |
655 | if (ucs >= 0x10000) { |
656 | /* Combining chars in Supplementary Multilingual Plane 0x1xxxx */ |
657 | static const struct interval combining0x10000[] = { |
658 | { 0x0A01, 0x0A03 }, { 0x0A05, 0x0A06 }, { 0x0A0C, 0x0A0F }, |
659 | { 0x0A38, 0x0A3A }, { 0x0A3F, 0x0A3F }, { 0xD167, 0xD169 }, |
660 | { 0xD173, 0xD182 }, { 0xD185, 0xD18B }, { 0xD1AA, 0xD1AD }, |
661 | { 0xD242, 0xD244 } |
662 | }; |
663 | /* Binary search in table of non-spacing characters in Supplementary Multilingual Plane */ |
664 | if (in_interval_table(ucs ^ 0x10000, combining0x10000, ARRAY_SIZE(combining0x10000) - 1)) |
665 | return 0; |
666 | /* Check a few non-spacing chars in Supplementary Special-purpose Plane 0xExxxx */ |
667 | if (CONFIG_LAST_SUPPORTED_WCHAR >= 0xE0001 |
668 | && ( ucs == 0xE0001 |
669 | || (ucs >= 0xE0020 && ucs <= 0xE007F) |
670 | || (ucs >= 0xE0100 && ucs <= 0xE01EF) |
671 | ) |
672 | ) { |
673 | return 0; |
674 | } |
675 | } |
676 | # endif |
677 | |
678 | /* If we arrive here, ucs is not a combining or C0/C1 control character. |
679 | * Check whether it's 1 char or 2-shar wide. |
680 | */ |
681 | return 1 + |
682 | ( (/*ucs >= 0x1100 &&*/ ucs <= 0x115f) /* Hangul Jamo init. consonants */ |
683 | || ucs == 0x2329 /* left-pointing angle bracket; also CJK punct. char */ |
684 | || ucs == 0x232a /* right-pointing angle bracket; also CJK punct. char */ |
685 | || (ucs >= 0x2e80 && ucs <= 0xa4cf && ucs != 0x303f) /* CJK ... Yi */ |
686 | # if CONFIG_LAST_SUPPORTED_WCHAR >= 0xac00 |
687 | || (ucs >= 0xac00 && ucs <= 0xd7a3) /* Hangul Syllables */ |
688 | || (ucs >= 0xf900 && ucs <= 0xfaff) /* CJK Compatibility Ideographs */ |
689 | || (ucs >= 0xfe10 && ucs <= 0xfe19) /* Vertical forms */ |
690 | || (ucs >= 0xfe30 && ucs <= 0xfe6f) /* CJK Compatibility Forms */ |
691 | || (ucs >= 0xff00 && ucs <= 0xff60) /* Fullwidth Forms */ |
692 | || (ucs >= 0xffe0 && ucs <= 0xffe6) |
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 |
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[] = { 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[] = { 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 |
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[] = { 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[] = { 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 | |
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 | |
1004 | size_t FAST_FUNC unicode_strwidth(const char *string) |
1005 | { |
1006 | uni_stat_t uni_stat; |
1007 | printable_string(&uni_stat, string); |
1008 | return uni_stat.unicode_width; |
1009 | } |
1010 | |
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 | #if !ENABLE_UNICODE_PRESERVE_BROKEN /* Unicode checks are not working in 1.19.0 but can be displayed if not filtered */ |
1038 | unsigned char c = *d; |
1039 | if (c < ' ' || c >= 0x7f) |
1040 | *d = '?'; |
1041 | #endif |
1042 | d++; |
1043 | } |
1044 | } |
1045 | if (stats) { |
1046 | stats->byte_count = (d - dst); |
1047 | stats->unicode_count = (d - dst); |
1048 | stats->unicode_width = (d - dst); |
1049 | } |
1050 | return dst; |
1051 | } |
1052 | |
1053 | dst = NULL; |
1054 | uni_count = uni_width = 0; |
1055 | dst_len = 0; |
1056 | while (1) { |
1057 | int w; |
1058 | wchar_t wc; |
1059 | |
1060 | #if ENABLE_UNICODE_USING_LOCALE |
1061 | { |
1062 | mbstate_t mbst = { 0 }; |
1063 | ssize_t rc = mbsrtowcs(&wc, &src, 1, &mbst); |
1064 | /* If invalid sequence is seen: -1 is returned, |
1065 | * src points to the invalid sequence, errno = EILSEQ. |
1066 | * Else number of wchars (excluding terminating L'\0') |
1067 | * written to dest is returned. |
1068 | * If len (here: 1) non-L'\0' wchars stored at dest, |
1069 | * src points to the next char to be converted. |
1070 | * If string is completely converted: src = NULL. |
1071 | */ |
1072 | if (rc == 0) /* end-of-string */ |
1073 | break; |
1074 | if (rc < 0) { /* error */ |
1075 | src++; |
1076 | goto subst; |
1077 | } |
1078 | if (!iswprint(wc)) |
1079 | goto subst; |
1080 | } |
1081 | #else |
1082 | src = mbstowc_internal(&wc, src); |
1083 | /* src is advanced to next mb char |
1084 | * wc == ERROR_WCHAR: invalid sequence is seen |
1085 | * else: wc is set |
1086 | */ |
1087 | if (wc == ERROR_WCHAR) /* error */ |
1088 | goto subst; |
1089 | if (wc == 0) /* end-of-string */ |
1090 | break; |
1091 | #endif |
1092 | if (CONFIG_LAST_SUPPORTED_WCHAR && wc > CONFIG_LAST_SUPPORTED_WCHAR) |
1093 | goto subst; |
1094 | w = wcwidth(wc); |
1095 | if ((ENABLE_UNICODE_COMBINING_WCHARS && w < 0) /* non-printable wchar */ |
1096 | || (!ENABLE_UNICODE_COMBINING_WCHARS && w <= 0) |
1097 | || (!ENABLE_UNICODE_WIDE_WCHARS && w > 1) |
1098 | ) { |
1099 | subst: |
1100 | wc = CONFIG_SUBST_WCHAR; |
1101 | w = 1; |
1102 | } |
1103 | width -= w; |
1104 | /* Note: if width == 0, we still may add more chars, |
1105 | * they may be zero-width or combining ones */ |
1106 | if ((int)width < 0) { |
1107 | /* can't add this wc, string would become longer than width */ |
1108 | width += w; |
1109 | break; |
1110 | } |
1111 | |
1112 | uni_count++; |
1113 | uni_width += w; |
1114 | dst = xrealloc(dst, dst_len + MB_CUR_MAX); |
1115 | #if ENABLE_UNICODE_USING_LOCALE |
1116 | { |
1117 | mbstate_t mbst = { 0 }; |
1118 | dst_len += wcrtomb(&dst[dst_len], wc, &mbst); |
1119 | } |
1120 | #else |
1121 | dst_len += wcrtomb_internal(&dst[dst_len], wc); |
1122 | #endif |
1123 | } |
1124 | |
1125 | /* Pad to remaining width */ |
1126 | if (flags & UNI_FLAG_PAD) { |
1127 | dst = xrealloc(dst, dst_len + width + 1); |
1128 | uni_count += width; |
1129 | uni_width += width; |
1130 | while ((int)--width >= 0) { |
1131 | dst[dst_len++] = ' '; |
1132 | } |
1133 | } |
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 | } |
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 | } |
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 |
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 | |
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 |