1/*
2 * Copyright (c) 2009 Public Software Group e. V., Berlin, Germany
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
--- 24 unchanged lines hidden (view full) ---
34/*
35 * File name: utf8proc.c
36 *
37 * Description:
38 * Implementation of libutf8proc.
39 */
40
41
42#include "utf8proc.h"
43#include "utf8proc_data.c"
44
45
46UTF8PROC_DATA
47const int8_t utf8proc_utf8class[256] = {
48 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
49 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
50 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
51 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
52 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
53 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
54 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
55 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
--- 21 unchanged lines hidden (view full) ---
77#define UTF8PROC_HANGUL_L_FILLER 0x115F
78#define UTF8PROC_HANGUL_V_START 0x1160
79#define UTF8PROC_HANGUL_V_END 0x11A3
80#define UTF8PROC_HANGUL_T_START 0x11A8
81#define UTF8PROC_HANGUL_T_END 0x11FA
82#define UTF8PROC_HANGUL_S_START 0xAC00
83#define UTF8PROC_HANGUL_S_END 0xD7A4
84
85
86#define UTF8PROC_BOUNDCLASS_START 0
87#define UTF8PROC_BOUNDCLASS_OTHER 1
88#define UTF8PROC_BOUNDCLASS_CR 2
89#define UTF8PROC_BOUNDCLASS_LF 3
90#define UTF8PROC_BOUNDCLASS_CONTROL 4
91#define UTF8PROC_BOUNDCLASS_EXTEND 5
92#define UTF8PROC_BOUNDCLASS_L 6
93#define UTF8PROC_BOUNDCLASS_V 7
94#define UTF8PROC_BOUNDCLASS_T 8
95#define UTF8PROC_BOUNDCLASS_LV 9
96#define UTF8PROC_BOUNDCLASS_LVT 10
97
98
99UTF8PROC_API
100const char *utf8proc_version(void) {
101 return "1.1.5";
102}
103
104/*
105 * This macro tells translators that string X should be translated,
106 * but does not look up the translation at run time. This is standard
107 * GNU gettext notation for annotating compile-time constant strings.
108 */
109#ifndef N_
110#define N_(x) x
111#endif
112
113UTF8PROC_API
114const char *utf8proc_errmsg(ssize_t errcode) {
115 switch (errcode) {
116 case UTF8PROC_ERROR_NOMEM:
117 return N_("Memory for processing UTF-8 data could not be allocated.");
118 case UTF8PROC_ERROR_OVERFLOW:
119 return N_("UTF-8 string is too long to be processed.");
120 case UTF8PROC_ERROR_INVALIDUTF8:
121 return N_("Invalid UTF-8 string");
122 case UTF8PROC_ERROR_NOTASSIGNED:
123 return N_("Unassigned Unicode code point found in UTF-8 string.");
124 case UTF8PROC_ERROR_INVALIDOPTS:
125 return N_("Invalid options for UTF-8 processing chosen.");
126 default:
127 return N_("An unknown error occured while processing UTF-8 data.");
128 }
129}
130
131UTF8PROC_API
132ssize_t utf8proc_iterate(
133 const uint8_t *str, ssize_t strlen, int32_t *dst
134) {
135 int length;
136 int i;
137 int32_t uc = -1;
138 *dst = -1;
139 if (!strlen) return 0;
140 length = utf8proc_utf8class[str[0]];
141 if (!length) return UTF8PROC_ERROR_INVALIDUTF8;
142 if (strlen >= 0 && length > strlen) return UTF8PROC_ERROR_INVALIDUTF8;
143 for (i=1; i<length; i++) {
144 if ((str[i] & 0xC0) != 0x80) return UTF8PROC_ERROR_INVALIDUTF8;
145 }
146 switch (length) {
147 case 1:
148 uc = str[0];
149 break;
150 case 2:
151 uc = ((str[0] & 0x1F) << 6) + (str[1] & 0x3F);
152 if (uc < 0x80) uc = -1;
153 break;
154 case 3:
155 uc = ((str[0] & 0x0F) << 12) + ((str[1] & 0x3F) << 6)
156 + (str[2] & 0x3F);
157 if (uc < 0x800 || (uc >= 0xD800 && uc < 0xE000) ||
158 (uc >= 0xFDD0 && uc < 0xFDF0)) uc = -1;
159 break;
160 case 4:
161 uc = ((str[0] & 0x07) << 18) + ((str[1] & 0x3F) << 12)
162 + ((str[2] & 0x3F) << 6) + (str[3] & 0x3F);
163 if (uc < 0x10000 || uc >= 0x110000) uc = -1;
164 break;
165 }
166 if (uc < 0 || ((uc & 0xFFFF) >= 0xFFFE))
167 return UTF8PROC_ERROR_INVALIDUTF8;
168 *dst = uc;
169 return length;
170}
171
172UTF8PROC_API
173bool utf8proc_codepoint_valid(int32_t uc) {
174 if (uc < 0 || uc >= 0x110000 ||
175 ((uc & 0xFFFF) >= 0xFFFE) || (uc >= 0xD800 && uc < 0xE000) ||
176 (uc >= 0xFDD0 && uc < 0xFDF0)) return false;
177 else return true;
178}
179
180UTF8PROC_API
181ssize_t utf8proc_encode_char(int32_t uc, uint8_t *dst) {
182 if (uc < 0x00) {
183 return 0;
184 } else if (uc < 0x80) {
185 dst[0] = (uint8_t)uc;
186 return 1;
187 } else if (uc < 0x800) {
188 dst[0] = 0xC0 + (uint8_t)(uc >> 6);
189 dst[1] = 0x80 + (uc & 0x3F);
190 return 2;
191 } else if (uc == 0xFFFF) {
192 dst[0] = 0xFF;
193 return 1;
194 } else if (uc == 0xFFFE) {
195 dst[0] = 0xFE;
196 return 1;
197 } else if (uc < 0x10000) {
198 dst[0] = 0xE0 + (uint8_t)(uc >> 12);
199 dst[1] = 0x80 + ((uc >> 6) & 0x3F);
200 dst[2] = 0x80 + (uc & 0x3F);
201 return 3;
202 } else if (uc < 0x110000) {
203 dst[0] = 0xF0 + (uint8_t)(uc >> 18);
204 dst[1] = 0x80 + ((uc >> 12) & 0x3F);
205 dst[2] = 0x80 + ((uc >> 6) & 0x3F);
206 dst[3] = 0x80 + (uc & 0x3F);
207 return 4;
208 } else return 0;
209}
210
211UTF8PROC_API
212const utf8proc_property_t *utf8proc_get_property(int32_t uc) {
213 /* ASSERT: uc >= 0 && uc < 0x110000 */
214 return utf8proc_properties + (
215 utf8proc_stage2table[
216 utf8proc_stage1table[uc >> 8] + (uc & 0xFF)
217 ]
218 );
219}
220
221#define utf8proc_decompose_lump(replacement_uc) \
222 return utf8proc_decompose_char((replacement_uc), dst, bufsize, \
223 options & ~UTF8PROC_LUMP, last_boundclass)
224
225UTF8PROC_API
226ssize_t utf8proc_decompose_char(int32_t uc, int32_t *dst, ssize_t bufsize,
227 int options, int *last_boundclass) {
228 /* ASSERT: uc >= 0 && uc < 0x110000 */
229 const utf8proc_property_t *property;
230 utf8proc_propval_t category;
231 int32_t hangul_sindex;
232 property = utf8proc_get_property(uc);
233 category = property->category;
234 hangul_sindex = uc - UTF8PROC_HANGUL_SBASE;
235 if (options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) {
236 if (hangul_sindex >= 0 && hangul_sindex < UTF8PROC_HANGUL_SCOUNT) {
237 int32_t hangul_tindex;
238 if (bufsize >= 1) {
239 dst[0] = UTF8PROC_HANGUL_LBASE +
240 hangul_sindex / UTF8PROC_HANGUL_NCOUNT;
241 if (bufsize >= 2) dst[1] = UTF8PROC_HANGUL_VBASE +
242 (hangul_sindex % UTF8PROC_HANGUL_NCOUNT) / UTF8PROC_HANGUL_TCOUNT;
243 }
244 hangul_tindex = hangul_sindex % UTF8PROC_HANGUL_TCOUNT;
245 if (!hangul_tindex) return 2;
--- 34 unchanged lines hidden (view full) ---
280 }
281 }
282 if (options & UTF8PROC_STRIPMARK) {
283 if (category == UTF8PROC_CATEGORY_MN ||
284 category == UTF8PROC_CATEGORY_MC ||
285 category == UTF8PROC_CATEGORY_ME) return 0;
286 }
287 if (options & UTF8PROC_CASEFOLD) {
288 if (property->casefold_mapping) {
289 const int32_t *casefold_entry;
290 ssize_t written = 0;
291 for (casefold_entry = property->casefold_mapping;
292 *casefold_entry >= 0; casefold_entry++) {
293 written += utf8proc_decompose_char(*casefold_entry, dst+written,
294 (bufsize > written) ? (bufsize - written) : 0, options,
295 last_boundclass);
296 if (written < 0) return UTF8PROC_ERROR_OVERFLOW;
297 }
298 return written;
299 }
300 }
301 if (options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) {
302 if (property->decomp_mapping &&
303 (!property->decomp_type || (options & UTF8PROC_COMPAT))) {
304 const int32_t *decomp_entry;
305 ssize_t written = 0;
306 for (decomp_entry = property->decomp_mapping;
307 *decomp_entry >= 0; decomp_entry++) {
308 written += utf8proc_decompose_char(*decomp_entry, dst+written,
309 (bufsize > written) ? (bufsize - written) : 0, options,
310 last_boundclass);
311 if (written < 0) return UTF8PROC_ERROR_OVERFLOW;
312 }
313 return written;
314 }
315 }
316 if (options & UTF8PROC_CHARBOUND) {
317 bool boundary;
318 int tbc, lbc;
319 tbc =
320 (uc == 0x000D) ? UTF8PROC_BOUNDCLASS_CR :
321 (uc == 0x000A) ? UTF8PROC_BOUNDCLASS_LF :
322 ((category == UTF8PROC_CATEGORY_ZL ||
323 category == UTF8PROC_CATEGORY_ZP ||
324 category == UTF8PROC_CATEGORY_CC ||
325 category == UTF8PROC_CATEGORY_CF) &&
326 !(uc == 0x200C || uc == 0x200D)) ? UTF8PROC_BOUNDCLASS_CONTROL :
327 property->extend ? UTF8PROC_BOUNDCLASS_EXTEND :
328 ((uc >= UTF8PROC_HANGUL_L_START && uc < UTF8PROC_HANGUL_L_END) ||
329 uc == UTF8PROC_HANGUL_L_FILLER) ? UTF8PROC_BOUNDCLASS_L :
330 (uc >= UTF8PROC_HANGUL_V_START && uc < UTF8PROC_HANGUL_V_END) ?
331 UTF8PROC_BOUNDCLASS_V :
332 (uc >= UTF8PROC_HANGUL_T_START && uc < UTF8PROC_HANGUL_T_END) ?
333 UTF8PROC_BOUNDCLASS_T :
334 (uc >= UTF8PROC_HANGUL_S_START && uc < UTF8PROC_HANGUL_S_END) ? (
335 ((uc-UTF8PROC_HANGUL_SBASE) % UTF8PROC_HANGUL_TCOUNT == 0) ?
336 UTF8PROC_BOUNDCLASS_LV : UTF8PROC_BOUNDCLASS_LVT
337 ) :
338 UTF8PROC_BOUNDCLASS_OTHER;
339 lbc = *last_boundclass;
340 boundary =
341 (tbc == UTF8PROC_BOUNDCLASS_EXTEND) ? false :
342 (lbc == UTF8PROC_BOUNDCLASS_START) ? true :
343 (lbc == UTF8PROC_BOUNDCLASS_CR &&
344 tbc == UTF8PROC_BOUNDCLASS_LF) ? false :
345 (lbc == UTF8PROC_BOUNDCLASS_CONTROL) ? true :
346 (tbc == UTF8PROC_BOUNDCLASS_CONTROL) ? true :
347 (lbc == UTF8PROC_BOUNDCLASS_L &&
348 (tbc == UTF8PROC_BOUNDCLASS_L ||
349 tbc == UTF8PROC_BOUNDCLASS_V ||
350 tbc == UTF8PROC_BOUNDCLASS_LV ||
351 tbc == UTF8PROC_BOUNDCLASS_LVT)) ? false :
352 ((lbc == UTF8PROC_BOUNDCLASS_LV ||
353 lbc == UTF8PROC_BOUNDCLASS_V) &&
354 (tbc == UTF8PROC_BOUNDCLASS_V ||
355 tbc == UTF8PROC_BOUNDCLASS_T)) ? false :
356 ((lbc == UTF8PROC_BOUNDCLASS_LVT ||
357 lbc == UTF8PROC_BOUNDCLASS_T) &&
358 tbc == UTF8PROC_BOUNDCLASS_T) ? false :
359 true;
360 *last_boundclass = tbc;
361 if (boundary) {
362 if (bufsize >= 1) dst[0] = 0xFFFF;
363 if (bufsize >= 2) dst[1] = uc;
364 return 2;
365 }
366 }
367 if (bufsize >= 1) *dst = uc;
368 return 1;
369}
370
371UTF8PROC_API
372ssize_t utf8proc_decompose(
373 const uint8_t *str, ssize_t strlen,
374 int32_t *buffer, ssize_t bufsize, int options
375) {
376 /* strlen will be ignored, if UTF8PROC_NULLTERM is set in options */
377 ssize_t wpos = 0;
378 if ((options & UTF8PROC_COMPOSE) && (options & UTF8PROC_DECOMPOSE))
379 return UTF8PROC_ERROR_INVALIDOPTS;
380 if ((options & UTF8PROC_STRIPMARK) &&
381 !(options & UTF8PROC_COMPOSE) && !(options & UTF8PROC_DECOMPOSE))
382 return UTF8PROC_ERROR_INVALIDOPTS;
383 {
384 int32_t uc;
385 ssize_t rpos = 0;
386 ssize_t decomp_result;
387 int boundclass = UTF8PROC_BOUNDCLASS_START;
388 while (1) {
389 if (options & UTF8PROC_NULLTERM) {
390 rpos += utf8proc_iterate(str + rpos, -1, &uc);
391 /* checking of return value is not neccessary,
392 as 'uc' is < 0 in case of error */
393 if (uc < 0) return UTF8PROC_ERROR_INVALIDUTF8;
394 if (rpos < 0) return UTF8PROC_ERROR_OVERFLOW;
395 if (uc == 0) break;
396 } else {
397 if (rpos >= strlen) break;
398 rpos += utf8proc_iterate(str + rpos, strlen - rpos, &uc);
399 if (uc < 0) return UTF8PROC_ERROR_INVALIDUTF8;
400 }
401 decomp_result = utf8proc_decompose_char(
402 uc, buffer + wpos, (bufsize > wpos) ? (bufsize - wpos) : 0, options,
403 &boundclass
404 );
405 if (decomp_result < 0) return decomp_result;
406 wpos += decomp_result;
407 /* prohibiting integer overflows due to too long strings: */
408 if (wpos < 0 || wpos > SSIZE_MAX/sizeof(int32_t)/2)
409 return UTF8PROC_ERROR_OVERFLOW;
410 }
411 }
412 if ((options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) && bufsize >= wpos) {
413 ssize_t pos = 0;
414 while (pos < wpos-1) {
415 int32_t uc1, uc2;
416 const utf8proc_property_t *property1, *property2;
417 uc1 = buffer[pos];
418 uc2 = buffer[pos+1];
419 property1 = utf8proc_get_property(uc1);
420 property2 = utf8proc_get_property(uc2);
421 if (property1->combining_class > property2->combining_class &&
422 property2->combining_class > 0) {
423 buffer[pos] = uc2;
424 buffer[pos+1] = uc1;
425 if (pos > 0) pos--; else pos++;
426 } else {
427 pos++;
428 }
429 }
430 }
431 return wpos;
432}
433
434UTF8PROC_API
435ssize_t utf8proc_reencode(int32_t *buffer, ssize_t length, int options) {
436 /* UTF8PROC_NULLTERM option will be ignored, 'length' is never ignored
437 ASSERT: 'buffer' has one spare byte of free space at the end! */
438 if (options & (UTF8PROC_NLF2LS | UTF8PROC_NLF2PS | UTF8PROC_STRIPCC)) {
439 ssize_t rpos;
440 ssize_t wpos = 0;
441 int32_t uc;
442 for (rpos = 0; rpos < length; rpos++) {
443 uc = buffer[rpos];
444 if (uc == 0x000D && rpos < length-1 && buffer[rpos+1] == 0x000A) rpos++;
445 if (uc == 0x000A || uc == 0x000D || uc == 0x0085 ||
446 ((options & UTF8PROC_STRIPCC) && (uc == 0x000B || uc == 0x000C))) {
447 if (options & UTF8PROC_NLF2LS) {
448 if (options & UTF8PROC_NLF2PS) {
449 buffer[wpos++] = 0x000A;
--- 12 unchanged lines hidden (view full) ---
462 if (uc == 0x0009) buffer[wpos++] = 0x0020;
463 } else {
464 buffer[wpos++] = uc;
465 }
466 }
467 length = wpos;
468 }
469 if (options & UTF8PROC_COMPOSE) {
470 int32_t *starter = NULL;
471 int32_t current_char;
472 const utf8proc_property_t *starter_property = NULL, *current_property;
473 utf8proc_propval_t max_combining_class = -1;
474 ssize_t rpos;
475 ssize_t wpos = 0;
476 int32_t composition;
477 for (rpos = 0; rpos < length; rpos++) {
478 current_char = buffer[rpos];
479 current_property = utf8proc_get_property(current_char);
480 if (starter && current_property->combining_class > max_combining_class) {
481 /* combination perhaps possible */
482 int32_t hangul_lindex;
483 int32_t hangul_sindex;
484 hangul_lindex = *starter - UTF8PROC_HANGUL_LBASE;
485 if (hangul_lindex >= 0 && hangul_lindex < UTF8PROC_HANGUL_LCOUNT) {
486 int32_t hangul_vindex;
487 hangul_vindex = current_char - UTF8PROC_HANGUL_VBASE;
488 if (hangul_vindex >= 0 && hangul_vindex < UTF8PROC_HANGUL_VCOUNT) {
489 *starter = UTF8PROC_HANGUL_SBASE +
490 (hangul_lindex * UTF8PROC_HANGUL_VCOUNT + hangul_vindex) *
491 UTF8PROC_HANGUL_TCOUNT;
492 starter_property = NULL;
493 continue;
494 }
495 }
496 hangul_sindex = *starter - UTF8PROC_HANGUL_SBASE;
497 if (hangul_sindex >= 0 && hangul_sindex < UTF8PROC_HANGUL_SCOUNT &&
498 (hangul_sindex % UTF8PROC_HANGUL_TCOUNT) == 0) {
499 int32_t hangul_tindex;
500 hangul_tindex = current_char - UTF8PROC_HANGUL_TBASE;
501 if (hangul_tindex >= 0 && hangul_tindex < UTF8PROC_HANGUL_TCOUNT) {
502 *starter += hangul_tindex;
503 starter_property = NULL;
504 continue;
505 }
506 }
507 if (!starter_property) {
508 starter_property = utf8proc_get_property(*starter);
509 }
510 if (starter_property->comb1st_index >= 0 &&
511 current_property->comb2nd_index >= 0) {
512 composition = utf8proc_combinations[
513 starter_property->comb1st_index +
514 current_property->comb2nd_index
515 ];
516 if (composition >= 0 && (!(options & UTF8PROC_STABLE) ||
517 !(utf8proc_get_property(composition)->comp_exclusion))) {
518 *starter = composition;
519 starter_property = NULL;
520 continue;
521 }
522 }
523 }
524 buffer[wpos] = current_char;
525 if (current_property->combining_class) {
526 if (current_property->combining_class > max_combining_class) {
527 max_combining_class = current_property->combining_class;
528 }
529 } else {
530 starter = buffer + wpos;
531 starter_property = NULL;
532 max_combining_class = -1;
533 }
534 wpos++;
535 }
536 length = wpos;
537 }
538 {
539 ssize_t rpos, wpos = 0;
540 int32_t uc;
541 for (rpos = 0; rpos < length; rpos++) {
542 uc = buffer[rpos];
543 wpos += utf8proc_encode_char(uc, ((uint8_t *)buffer) + wpos);
544 }
545 ((uint8_t *)buffer)[wpos] = 0;
546 return wpos;
547 }
548}
549
550UTF8PROC_API
551ssize_t utf8proc_map(
552 const uint8_t *str, ssize_t strlen, uint8_t **dstptr, int options
553) {
554 int32_t *buffer;
555 ssize_t result;
556 *dstptr = NULL;
557 result = utf8proc_decompose(str, strlen, NULL, 0, options);
558 if (result < 0) return result;
559 buffer = malloc(result * sizeof(int32_t) + 1);
560 if (!buffer) return UTF8PROC_ERROR_NOMEM;
561 result = utf8proc_decompose(str, strlen, buffer, result, options);
562 if (result < 0) {
563 free(buffer);
564 return result;
565 }
566 result = utf8proc_reencode(buffer, result, options);
567 if (result < 0) {
568 free(buffer);
569 return result;
570 }
571 {
572 int32_t *newptr;
573 newptr = realloc(buffer, (size_t)result+1);
574 if (newptr) buffer = newptr;
575 }
576 *dstptr = (uint8_t *)buffer;
577 return result;
578}
579
580UTF8PROC_API
581uint8_t *utf8proc_NFD(const uint8_t *str) {
582 uint8_t *retval;
583 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
584 UTF8PROC_DECOMPOSE);
585 return retval;
586}
587
588UTF8PROC_API
589uint8_t *utf8proc_NFC(const uint8_t *str) {
590 uint8_t *retval;
591 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
592 UTF8PROC_COMPOSE);
593 return retval;
594}
595
596UTF8PROC_API
597uint8_t *utf8proc_NFKD(const uint8_t *str) {
598 uint8_t *retval;
599 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
600 UTF8PROC_DECOMPOSE | UTF8PROC_COMPAT);
601 return retval;
602}
603
604UTF8PROC_API
605uint8_t *utf8proc_NFKC(const uint8_t *str) {
606 uint8_t *retval;
607 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
608 UTF8PROC_COMPOSE | UTF8PROC_COMPAT);
609 return retval;
610}
611
2 * Copyright (c) 2009 Public Software Group e. V., Berlin, Germany
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
--- 24 unchanged lines hidden (view full) ---
34/*
35 * File name: utf8proc.c
36 *
37 * Description:
38 * Implementation of libutf8proc.
39 */
40
41
42#include "utf8proc.h"
43#include "utf8proc_data.c"
44
45
46UTF8PROC_DATA
47const int8_t utf8proc_utf8class[256] = {
48 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
49 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
50 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
51 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
52 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
53 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
54 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
55 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
--- 21 unchanged lines hidden (view full) ---
77#define UTF8PROC_HANGUL_L_FILLER 0x115F
78#define UTF8PROC_HANGUL_V_START 0x1160
79#define UTF8PROC_HANGUL_V_END 0x11A3
80#define UTF8PROC_HANGUL_T_START 0x11A8
81#define UTF8PROC_HANGUL_T_END 0x11FA
82#define UTF8PROC_HANGUL_S_START 0xAC00
83#define UTF8PROC_HANGUL_S_END 0xD7A4
84
85
86#define UTF8PROC_BOUNDCLASS_START 0
87#define UTF8PROC_BOUNDCLASS_OTHER 1
88#define UTF8PROC_BOUNDCLASS_CR 2
89#define UTF8PROC_BOUNDCLASS_LF 3
90#define UTF8PROC_BOUNDCLASS_CONTROL 4
91#define UTF8PROC_BOUNDCLASS_EXTEND 5
92#define UTF8PROC_BOUNDCLASS_L 6
93#define UTF8PROC_BOUNDCLASS_V 7
94#define UTF8PROC_BOUNDCLASS_T 8
95#define UTF8PROC_BOUNDCLASS_LV 9
96#define UTF8PROC_BOUNDCLASS_LVT 10
97
98
99UTF8PROC_API
100const char *utf8proc_version(void) {
101 return "1.1.5";
102}
103
104/*
105 * This macro tells translators that string X should be translated,
106 * but does not look up the translation at run time. This is standard
107 * GNU gettext notation for annotating compile-time constant strings.
108 */
109#ifndef N_
110#define N_(x) x
111#endif
112
113UTF8PROC_API
114const char *utf8proc_errmsg(ssize_t errcode) {
115 switch (errcode) {
116 case UTF8PROC_ERROR_NOMEM:
117 return N_("Memory for processing UTF-8 data could not be allocated.");
118 case UTF8PROC_ERROR_OVERFLOW:
119 return N_("UTF-8 string is too long to be processed.");
120 case UTF8PROC_ERROR_INVALIDUTF8:
121 return N_("Invalid UTF-8 string");
122 case UTF8PROC_ERROR_NOTASSIGNED:
123 return N_("Unassigned Unicode code point found in UTF-8 string.");
124 case UTF8PROC_ERROR_INVALIDOPTS:
125 return N_("Invalid options for UTF-8 processing chosen.");
126 default:
127 return N_("An unknown error occured while processing UTF-8 data.");
128 }
129}
130
131UTF8PROC_API
132ssize_t utf8proc_iterate(
133 const uint8_t *str, ssize_t strlen, int32_t *dst
134) {
135 int length;
136 int i;
137 int32_t uc = -1;
138 *dst = -1;
139 if (!strlen) return 0;
140 length = utf8proc_utf8class[str[0]];
141 if (!length) return UTF8PROC_ERROR_INVALIDUTF8;
142 if (strlen >= 0 && length > strlen) return UTF8PROC_ERROR_INVALIDUTF8;
143 for (i=1; i<length; i++) {
144 if ((str[i] & 0xC0) != 0x80) return UTF8PROC_ERROR_INVALIDUTF8;
145 }
146 switch (length) {
147 case 1:
148 uc = str[0];
149 break;
150 case 2:
151 uc = ((str[0] & 0x1F) << 6) + (str[1] & 0x3F);
152 if (uc < 0x80) uc = -1;
153 break;
154 case 3:
155 uc = ((str[0] & 0x0F) << 12) + ((str[1] & 0x3F) << 6)
156 + (str[2] & 0x3F);
157 if (uc < 0x800 || (uc >= 0xD800 && uc < 0xE000) ||
158 (uc >= 0xFDD0 && uc < 0xFDF0)) uc = -1;
159 break;
160 case 4:
161 uc = ((str[0] & 0x07) << 18) + ((str[1] & 0x3F) << 12)
162 + ((str[2] & 0x3F) << 6) + (str[3] & 0x3F);
163 if (uc < 0x10000 || uc >= 0x110000) uc = -1;
164 break;
165 }
166 if (uc < 0 || ((uc & 0xFFFF) >= 0xFFFE))
167 return UTF8PROC_ERROR_INVALIDUTF8;
168 *dst = uc;
169 return length;
170}
171
172UTF8PROC_API
173bool utf8proc_codepoint_valid(int32_t uc) {
174 if (uc < 0 || uc >= 0x110000 ||
175 ((uc & 0xFFFF) >= 0xFFFE) || (uc >= 0xD800 && uc < 0xE000) ||
176 (uc >= 0xFDD0 && uc < 0xFDF0)) return false;
177 else return true;
178}
179
180UTF8PROC_API
181ssize_t utf8proc_encode_char(int32_t uc, uint8_t *dst) {
182 if (uc < 0x00) {
183 return 0;
184 } else if (uc < 0x80) {
185 dst[0] = (uint8_t)uc;
186 return 1;
187 } else if (uc < 0x800) {
188 dst[0] = 0xC0 + (uint8_t)(uc >> 6);
189 dst[1] = 0x80 + (uc & 0x3F);
190 return 2;
191 } else if (uc == 0xFFFF) {
192 dst[0] = 0xFF;
193 return 1;
194 } else if (uc == 0xFFFE) {
195 dst[0] = 0xFE;
196 return 1;
197 } else if (uc < 0x10000) {
198 dst[0] = 0xE0 + (uint8_t)(uc >> 12);
199 dst[1] = 0x80 + ((uc >> 6) & 0x3F);
200 dst[2] = 0x80 + (uc & 0x3F);
201 return 3;
202 } else if (uc < 0x110000) {
203 dst[0] = 0xF0 + (uint8_t)(uc >> 18);
204 dst[1] = 0x80 + ((uc >> 12) & 0x3F);
205 dst[2] = 0x80 + ((uc >> 6) & 0x3F);
206 dst[3] = 0x80 + (uc & 0x3F);
207 return 4;
208 } else return 0;
209}
210
211UTF8PROC_API
212const utf8proc_property_t *utf8proc_get_property(int32_t uc) {
213 /* ASSERT: uc >= 0 && uc < 0x110000 */
214 return utf8proc_properties + (
215 utf8proc_stage2table[
216 utf8proc_stage1table[uc >> 8] + (uc & 0xFF)
217 ]
218 );
219}
220
221#define utf8proc_decompose_lump(replacement_uc) \
222 return utf8proc_decompose_char((replacement_uc), dst, bufsize, \
223 options & ~UTF8PROC_LUMP, last_boundclass)
224
225UTF8PROC_API
226ssize_t utf8proc_decompose_char(int32_t uc, int32_t *dst, ssize_t bufsize,
227 int options, int *last_boundclass) {
228 /* ASSERT: uc >= 0 && uc < 0x110000 */
229 const utf8proc_property_t *property;
230 utf8proc_propval_t category;
231 int32_t hangul_sindex;
232 property = utf8proc_get_property(uc);
233 category = property->category;
234 hangul_sindex = uc - UTF8PROC_HANGUL_SBASE;
235 if (options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) {
236 if (hangul_sindex >= 0 && hangul_sindex < UTF8PROC_HANGUL_SCOUNT) {
237 int32_t hangul_tindex;
238 if (bufsize >= 1) {
239 dst[0] = UTF8PROC_HANGUL_LBASE +
240 hangul_sindex / UTF8PROC_HANGUL_NCOUNT;
241 if (bufsize >= 2) dst[1] = UTF8PROC_HANGUL_VBASE +
242 (hangul_sindex % UTF8PROC_HANGUL_NCOUNT) / UTF8PROC_HANGUL_TCOUNT;
243 }
244 hangul_tindex = hangul_sindex % UTF8PROC_HANGUL_TCOUNT;
245 if (!hangul_tindex) return 2;
--- 34 unchanged lines hidden (view full) ---
280 }
281 }
282 if (options & UTF8PROC_STRIPMARK) {
283 if (category == UTF8PROC_CATEGORY_MN ||
284 category == UTF8PROC_CATEGORY_MC ||
285 category == UTF8PROC_CATEGORY_ME) return 0;
286 }
287 if (options & UTF8PROC_CASEFOLD) {
288 if (property->casefold_mapping) {
289 const int32_t *casefold_entry;
290 ssize_t written = 0;
291 for (casefold_entry = property->casefold_mapping;
292 *casefold_entry >= 0; casefold_entry++) {
293 written += utf8proc_decompose_char(*casefold_entry, dst+written,
294 (bufsize > written) ? (bufsize - written) : 0, options,
295 last_boundclass);
296 if (written < 0) return UTF8PROC_ERROR_OVERFLOW;
297 }
298 return written;
299 }
300 }
301 if (options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) {
302 if (property->decomp_mapping &&
303 (!property->decomp_type || (options & UTF8PROC_COMPAT))) {
304 const int32_t *decomp_entry;
305 ssize_t written = 0;
306 for (decomp_entry = property->decomp_mapping;
307 *decomp_entry >= 0; decomp_entry++) {
308 written += utf8proc_decompose_char(*decomp_entry, dst+written,
309 (bufsize > written) ? (bufsize - written) : 0, options,
310 last_boundclass);
311 if (written < 0) return UTF8PROC_ERROR_OVERFLOW;
312 }
313 return written;
314 }
315 }
316 if (options & UTF8PROC_CHARBOUND) {
317 bool boundary;
318 int tbc, lbc;
319 tbc =
320 (uc == 0x000D) ? UTF8PROC_BOUNDCLASS_CR :
321 (uc == 0x000A) ? UTF8PROC_BOUNDCLASS_LF :
322 ((category == UTF8PROC_CATEGORY_ZL ||
323 category == UTF8PROC_CATEGORY_ZP ||
324 category == UTF8PROC_CATEGORY_CC ||
325 category == UTF8PROC_CATEGORY_CF) &&
326 !(uc == 0x200C || uc == 0x200D)) ? UTF8PROC_BOUNDCLASS_CONTROL :
327 property->extend ? UTF8PROC_BOUNDCLASS_EXTEND :
328 ((uc >= UTF8PROC_HANGUL_L_START && uc < UTF8PROC_HANGUL_L_END) ||
329 uc == UTF8PROC_HANGUL_L_FILLER) ? UTF8PROC_BOUNDCLASS_L :
330 (uc >= UTF8PROC_HANGUL_V_START && uc < UTF8PROC_HANGUL_V_END) ?
331 UTF8PROC_BOUNDCLASS_V :
332 (uc >= UTF8PROC_HANGUL_T_START && uc < UTF8PROC_HANGUL_T_END) ?
333 UTF8PROC_BOUNDCLASS_T :
334 (uc >= UTF8PROC_HANGUL_S_START && uc < UTF8PROC_HANGUL_S_END) ? (
335 ((uc-UTF8PROC_HANGUL_SBASE) % UTF8PROC_HANGUL_TCOUNT == 0) ?
336 UTF8PROC_BOUNDCLASS_LV : UTF8PROC_BOUNDCLASS_LVT
337 ) :
338 UTF8PROC_BOUNDCLASS_OTHER;
339 lbc = *last_boundclass;
340 boundary =
341 (tbc == UTF8PROC_BOUNDCLASS_EXTEND) ? false :
342 (lbc == UTF8PROC_BOUNDCLASS_START) ? true :
343 (lbc == UTF8PROC_BOUNDCLASS_CR &&
344 tbc == UTF8PROC_BOUNDCLASS_LF) ? false :
345 (lbc == UTF8PROC_BOUNDCLASS_CONTROL) ? true :
346 (tbc == UTF8PROC_BOUNDCLASS_CONTROL) ? true :
347 (lbc == UTF8PROC_BOUNDCLASS_L &&
348 (tbc == UTF8PROC_BOUNDCLASS_L ||
349 tbc == UTF8PROC_BOUNDCLASS_V ||
350 tbc == UTF8PROC_BOUNDCLASS_LV ||
351 tbc == UTF8PROC_BOUNDCLASS_LVT)) ? false :
352 ((lbc == UTF8PROC_BOUNDCLASS_LV ||
353 lbc == UTF8PROC_BOUNDCLASS_V) &&
354 (tbc == UTF8PROC_BOUNDCLASS_V ||
355 tbc == UTF8PROC_BOUNDCLASS_T)) ? false :
356 ((lbc == UTF8PROC_BOUNDCLASS_LVT ||
357 lbc == UTF8PROC_BOUNDCLASS_T) &&
358 tbc == UTF8PROC_BOUNDCLASS_T) ? false :
359 true;
360 *last_boundclass = tbc;
361 if (boundary) {
362 if (bufsize >= 1) dst[0] = 0xFFFF;
363 if (bufsize >= 2) dst[1] = uc;
364 return 2;
365 }
366 }
367 if (bufsize >= 1) *dst = uc;
368 return 1;
369}
370
371UTF8PROC_API
372ssize_t utf8proc_decompose(
373 const uint8_t *str, ssize_t strlen,
374 int32_t *buffer, ssize_t bufsize, int options
375) {
376 /* strlen will be ignored, if UTF8PROC_NULLTERM is set in options */
377 ssize_t wpos = 0;
378 if ((options & UTF8PROC_COMPOSE) && (options & UTF8PROC_DECOMPOSE))
379 return UTF8PROC_ERROR_INVALIDOPTS;
380 if ((options & UTF8PROC_STRIPMARK) &&
381 !(options & UTF8PROC_COMPOSE) && !(options & UTF8PROC_DECOMPOSE))
382 return UTF8PROC_ERROR_INVALIDOPTS;
383 {
384 int32_t uc;
385 ssize_t rpos = 0;
386 ssize_t decomp_result;
387 int boundclass = UTF8PROC_BOUNDCLASS_START;
388 while (1) {
389 if (options & UTF8PROC_NULLTERM) {
390 rpos += utf8proc_iterate(str + rpos, -1, &uc);
391 /* checking of return value is not neccessary,
392 as 'uc' is < 0 in case of error */
393 if (uc < 0) return UTF8PROC_ERROR_INVALIDUTF8;
394 if (rpos < 0) return UTF8PROC_ERROR_OVERFLOW;
395 if (uc == 0) break;
396 } else {
397 if (rpos >= strlen) break;
398 rpos += utf8proc_iterate(str + rpos, strlen - rpos, &uc);
399 if (uc < 0) return UTF8PROC_ERROR_INVALIDUTF8;
400 }
401 decomp_result = utf8proc_decompose_char(
402 uc, buffer + wpos, (bufsize > wpos) ? (bufsize - wpos) : 0, options,
403 &boundclass
404 );
405 if (decomp_result < 0) return decomp_result;
406 wpos += decomp_result;
407 /* prohibiting integer overflows due to too long strings: */
408 if (wpos < 0 || wpos > SSIZE_MAX/sizeof(int32_t)/2)
409 return UTF8PROC_ERROR_OVERFLOW;
410 }
411 }
412 if ((options & (UTF8PROC_COMPOSE|UTF8PROC_DECOMPOSE)) && bufsize >= wpos) {
413 ssize_t pos = 0;
414 while (pos < wpos-1) {
415 int32_t uc1, uc2;
416 const utf8proc_property_t *property1, *property2;
417 uc1 = buffer[pos];
418 uc2 = buffer[pos+1];
419 property1 = utf8proc_get_property(uc1);
420 property2 = utf8proc_get_property(uc2);
421 if (property1->combining_class > property2->combining_class &&
422 property2->combining_class > 0) {
423 buffer[pos] = uc2;
424 buffer[pos+1] = uc1;
425 if (pos > 0) pos--; else pos++;
426 } else {
427 pos++;
428 }
429 }
430 }
431 return wpos;
432}
433
434UTF8PROC_API
435ssize_t utf8proc_reencode(int32_t *buffer, ssize_t length, int options) {
436 /* UTF8PROC_NULLTERM option will be ignored, 'length' is never ignored
437 ASSERT: 'buffer' has one spare byte of free space at the end! */
438 if (options & (UTF8PROC_NLF2LS | UTF8PROC_NLF2PS | UTF8PROC_STRIPCC)) {
439 ssize_t rpos;
440 ssize_t wpos = 0;
441 int32_t uc;
442 for (rpos = 0; rpos < length; rpos++) {
443 uc = buffer[rpos];
444 if (uc == 0x000D && rpos < length-1 && buffer[rpos+1] == 0x000A) rpos++;
445 if (uc == 0x000A || uc == 0x000D || uc == 0x0085 ||
446 ((options & UTF8PROC_STRIPCC) && (uc == 0x000B || uc == 0x000C))) {
447 if (options & UTF8PROC_NLF2LS) {
448 if (options & UTF8PROC_NLF2PS) {
449 buffer[wpos++] = 0x000A;
--- 12 unchanged lines hidden (view full) ---
462 if (uc == 0x0009) buffer[wpos++] = 0x0020;
463 } else {
464 buffer[wpos++] = uc;
465 }
466 }
467 length = wpos;
468 }
469 if (options & UTF8PROC_COMPOSE) {
470 int32_t *starter = NULL;
471 int32_t current_char;
472 const utf8proc_property_t *starter_property = NULL, *current_property;
473 utf8proc_propval_t max_combining_class = -1;
474 ssize_t rpos;
475 ssize_t wpos = 0;
476 int32_t composition;
477 for (rpos = 0; rpos < length; rpos++) {
478 current_char = buffer[rpos];
479 current_property = utf8proc_get_property(current_char);
480 if (starter && current_property->combining_class > max_combining_class) {
481 /* combination perhaps possible */
482 int32_t hangul_lindex;
483 int32_t hangul_sindex;
484 hangul_lindex = *starter - UTF8PROC_HANGUL_LBASE;
485 if (hangul_lindex >= 0 && hangul_lindex < UTF8PROC_HANGUL_LCOUNT) {
486 int32_t hangul_vindex;
487 hangul_vindex = current_char - UTF8PROC_HANGUL_VBASE;
488 if (hangul_vindex >= 0 && hangul_vindex < UTF8PROC_HANGUL_VCOUNT) {
489 *starter = UTF8PROC_HANGUL_SBASE +
490 (hangul_lindex * UTF8PROC_HANGUL_VCOUNT + hangul_vindex) *
491 UTF8PROC_HANGUL_TCOUNT;
492 starter_property = NULL;
493 continue;
494 }
495 }
496 hangul_sindex = *starter - UTF8PROC_HANGUL_SBASE;
497 if (hangul_sindex >= 0 && hangul_sindex < UTF8PROC_HANGUL_SCOUNT &&
498 (hangul_sindex % UTF8PROC_HANGUL_TCOUNT) == 0) {
499 int32_t hangul_tindex;
500 hangul_tindex = current_char - UTF8PROC_HANGUL_TBASE;
501 if (hangul_tindex >= 0 && hangul_tindex < UTF8PROC_HANGUL_TCOUNT) {
502 *starter += hangul_tindex;
503 starter_property = NULL;
504 continue;
505 }
506 }
507 if (!starter_property) {
508 starter_property = utf8proc_get_property(*starter);
509 }
510 if (starter_property->comb1st_index >= 0 &&
511 current_property->comb2nd_index >= 0) {
512 composition = utf8proc_combinations[
513 starter_property->comb1st_index +
514 current_property->comb2nd_index
515 ];
516 if (composition >= 0 && (!(options & UTF8PROC_STABLE) ||
517 !(utf8proc_get_property(composition)->comp_exclusion))) {
518 *starter = composition;
519 starter_property = NULL;
520 continue;
521 }
522 }
523 }
524 buffer[wpos] = current_char;
525 if (current_property->combining_class) {
526 if (current_property->combining_class > max_combining_class) {
527 max_combining_class = current_property->combining_class;
528 }
529 } else {
530 starter = buffer + wpos;
531 starter_property = NULL;
532 max_combining_class = -1;
533 }
534 wpos++;
535 }
536 length = wpos;
537 }
538 {
539 ssize_t rpos, wpos = 0;
540 int32_t uc;
541 for (rpos = 0; rpos < length; rpos++) {
542 uc = buffer[rpos];
543 wpos += utf8proc_encode_char(uc, ((uint8_t *)buffer) + wpos);
544 }
545 ((uint8_t *)buffer)[wpos] = 0;
546 return wpos;
547 }
548}
549
550UTF8PROC_API
551ssize_t utf8proc_map(
552 const uint8_t *str, ssize_t strlen, uint8_t **dstptr, int options
553) {
554 int32_t *buffer;
555 ssize_t result;
556 *dstptr = NULL;
557 result = utf8proc_decompose(str, strlen, NULL, 0, options);
558 if (result < 0) return result;
559 buffer = malloc(result * sizeof(int32_t) + 1);
560 if (!buffer) return UTF8PROC_ERROR_NOMEM;
561 result = utf8proc_decompose(str, strlen, buffer, result, options);
562 if (result < 0) {
563 free(buffer);
564 return result;
565 }
566 result = utf8proc_reencode(buffer, result, options);
567 if (result < 0) {
568 free(buffer);
569 return result;
570 }
571 {
572 int32_t *newptr;
573 newptr = realloc(buffer, (size_t)result+1);
574 if (newptr) buffer = newptr;
575 }
576 *dstptr = (uint8_t *)buffer;
577 return result;
578}
579
580UTF8PROC_API
581uint8_t *utf8proc_NFD(const uint8_t *str) {
582 uint8_t *retval;
583 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
584 UTF8PROC_DECOMPOSE);
585 return retval;
586}
587
588UTF8PROC_API
589uint8_t *utf8proc_NFC(const uint8_t *str) {
590 uint8_t *retval;
591 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
592 UTF8PROC_COMPOSE);
593 return retval;
594}
595
596UTF8PROC_API
597uint8_t *utf8proc_NFKD(const uint8_t *str) {
598 uint8_t *retval;
599 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
600 UTF8PROC_DECOMPOSE | UTF8PROC_COMPAT);
601 return retval;
602}
603
604UTF8PROC_API
605uint8_t *utf8proc_NFKC(const uint8_t *str) {
606 uint8_t *retval;
607 utf8proc_map(str, 0, &retval, UTF8PROC_NULLTERM | UTF8PROC_STABLE |
608 UTF8PROC_COMPOSE | UTF8PROC_COMPAT);
609 return retval;
610}
611