1 | // © 2016 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html |
3 | /* |
4 | ******************************************************************************* |
5 | * |
6 | * Copyright (C) 2005-2016, International Business Machines |
7 | * Corporation and others. All Rights Reserved. |
8 | * |
9 | ******************************************************************************* |
10 | * file name: ucasemap.cpp |
11 | * encoding: UTF-8 |
12 | * tab size: 8 (not used) |
13 | * indentation:4 |
14 | * |
15 | * created on: 2005may06 |
16 | * created by: Markus W. Scherer |
17 | * |
18 | * Case mapping service object and functions using it. |
19 | */ |
20 | |
21 | #include "unicode/utypes.h" |
22 | #include "unicode/brkiter.h" |
23 | #include "unicode/bytestream.h" |
24 | #include "unicode/casemap.h" |
25 | #include "unicode/edits.h" |
26 | #include "unicode/stringoptions.h" |
27 | #include "unicode/stringpiece.h" |
28 | #include "unicode/ubrk.h" |
29 | #include "unicode/uloc.h" |
30 | #include "unicode/ustring.h" |
31 | #include "unicode/ucasemap.h" |
32 | #if !UCONFIG_NO_BREAK_ITERATION |
33 | #include "unicode/utext.h" |
34 | #endif |
35 | #include "unicode/utf.h" |
36 | #include "unicode/utf8.h" |
37 | #include "unicode/utf16.h" |
38 | #include "bytesinkutil.h" |
39 | #include "cmemory.h" |
40 | #include "cstring.h" |
41 | #include "uassert.h" |
42 | #include "ucase.h" |
43 | #include "ucasemap_imp.h" |
44 | #include "ustr_imp.h" |
45 | |
46 | U_NAMESPACE_USE |
47 | |
48 | /* UCaseMap service object -------------------------------------------------- */ |
49 | |
50 | UCaseMap::UCaseMap(const char *localeID, uint32_t opts, UErrorCode *pErrorCode) : |
51 | #if !UCONFIG_NO_BREAK_ITERATION |
52 | iter(NULL), |
53 | #endif |
54 | caseLocale(UCASE_LOC_UNKNOWN), options(opts) { |
55 | ucasemap_setLocale(this, localeID, pErrorCode); |
56 | } |
57 | |
58 | UCaseMap::~UCaseMap() { |
59 | #if !UCONFIG_NO_BREAK_ITERATION |
60 | delete iter; |
61 | #endif |
62 | } |
63 | |
64 | U_CAPI UCaseMap * U_EXPORT2 |
65 | ucasemap_open(const char *locale, uint32_t options, UErrorCode *pErrorCode) { |
66 | if(U_FAILURE(*pErrorCode)) { |
67 | return NULL; |
68 | } |
69 | UCaseMap *csm = new UCaseMap(locale, options, pErrorCode); |
70 | if(csm==NULL) { |
71 | *pErrorCode = U_MEMORY_ALLOCATION_ERROR; |
72 | return NULL; |
73 | } else if (U_FAILURE(*pErrorCode)) { |
74 | delete csm; |
75 | return NULL; |
76 | } |
77 | return csm; |
78 | } |
79 | |
80 | U_CAPI void U_EXPORT2 |
81 | ucasemap_close(UCaseMap *csm) { |
82 | delete csm; |
83 | } |
84 | |
85 | U_CAPI const char * U_EXPORT2 |
86 | ucasemap_getLocale(const UCaseMap *csm) { |
87 | return csm->locale; |
88 | } |
89 | |
90 | U_CAPI uint32_t U_EXPORT2 |
91 | ucasemap_getOptions(const UCaseMap *csm) { |
92 | return csm->options; |
93 | } |
94 | |
95 | U_CAPI void U_EXPORT2 |
96 | ucasemap_setLocale(UCaseMap *csm, const char *locale, UErrorCode *pErrorCode) { |
97 | if(U_FAILURE(*pErrorCode)) { |
98 | return; |
99 | } |
100 | if (locale != NULL && *locale == 0) { |
101 | csm->locale[0] = 0; |
102 | csm->caseLocale = UCASE_LOC_ROOT; |
103 | return; |
104 | } |
105 | |
106 | int32_t length=uloc_getName(locale, csm->locale, (int32_t)sizeof(csm->locale), pErrorCode); |
107 | if(*pErrorCode==U_BUFFER_OVERFLOW_ERROR || length==sizeof(csm->locale)) { |
108 | *pErrorCode=U_ZERO_ERROR; |
109 | /* we only really need the language code for case mappings */ |
110 | length=uloc_getLanguage(locale, csm->locale, (int32_t)sizeof(csm->locale), pErrorCode); |
111 | } |
112 | if(length==sizeof(csm->locale)) { |
113 | *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
114 | } |
115 | if(U_SUCCESS(*pErrorCode)) { |
116 | csm->caseLocale=UCASE_LOC_UNKNOWN; |
117 | csm->caseLocale = ucase_getCaseLocale(csm->locale); |
118 | } else { |
119 | csm->locale[0]=0; |
120 | csm->caseLocale = UCASE_LOC_ROOT; |
121 | } |
122 | } |
123 | |
124 | U_CAPI void U_EXPORT2 |
125 | ucasemap_setOptions(UCaseMap *csm, uint32_t options, UErrorCode *pErrorCode) { |
126 | if(U_FAILURE(*pErrorCode)) { |
127 | return; |
128 | } |
129 | csm->options=options; |
130 | } |
131 | |
132 | /* UTF-8 string case mappings ----------------------------------------------- */ |
133 | |
134 | /* TODO(markus): Move to a new, separate utf8case.cpp file. */ |
135 | |
136 | namespace { |
137 | |
138 | /* append a full case mapping result, see UCASE_MAX_STRING_LENGTH */ |
139 | inline UBool |
140 | appendResult(int32_t cpLength, int32_t result, const UChar *s, |
141 | ByteSink &sink, uint32_t options, icu::Edits *edits, UErrorCode &errorCode) { |
142 | U_ASSERT(U_SUCCESS(errorCode)); |
143 | |
144 | /* decode the result */ |
145 | if(result<0) { |
146 | /* (not) original code point */ |
147 | if(edits!=NULL) { |
148 | edits->addUnchanged(cpLength); |
149 | } |
150 | if((options & U_OMIT_UNCHANGED_TEXT) == 0) { |
151 | ByteSinkUtil::appendCodePoint(cpLength, ~result, sink); |
152 | } |
153 | } else { |
154 | if(result<=UCASE_MAX_STRING_LENGTH) { |
155 | // string: "result" is the UTF-16 length |
156 | return ByteSinkUtil::appendChange(cpLength, s, result, sink, edits, errorCode); |
157 | } else { |
158 | ByteSinkUtil::appendCodePoint(cpLength, result, sink, edits); |
159 | } |
160 | } |
161 | return TRUE; |
162 | } |
163 | |
164 | // See unicode/utf8.h U8_APPEND_UNSAFE(). |
165 | inline uint8_t getTwoByteLead(UChar32 c) { return (uint8_t)((c >> 6) | 0xc0); } |
166 | inline uint8_t getTwoByteTrail(UChar32 c) { return (uint8_t)((c & 0x3f) | 0x80); } |
167 | |
168 | UChar32 U_CALLCONV |
169 | utf8_caseContextIterator(void *context, int8_t dir) { |
170 | UCaseContext *csc=(UCaseContext *)context; |
171 | UChar32 c; |
172 | |
173 | if(dir<0) { |
174 | /* reset for backward iteration */ |
175 | csc->index=csc->cpStart; |
176 | csc->dir=dir; |
177 | } else if(dir>0) { |
178 | /* reset for forward iteration */ |
179 | csc->index=csc->cpLimit; |
180 | csc->dir=dir; |
181 | } else { |
182 | /* continue current iteration direction */ |
183 | dir=csc->dir; |
184 | } |
185 | |
186 | if(dir<0) { |
187 | if(csc->start<csc->index) { |
188 | U8_PREV((const uint8_t *)csc->p, csc->start, csc->index, c); |
189 | return c; |
190 | } |
191 | } else { |
192 | if(csc->index<csc->limit) { |
193 | U8_NEXT((const uint8_t *)csc->p, csc->index, csc->limit, c); |
194 | return c; |
195 | } |
196 | } |
197 | return U_SENTINEL; |
198 | } |
199 | |
200 | /** |
201 | * caseLocale >= 0: Lowercases [srcStart..srcLimit[ but takes context [0..srcLength[ into account. |
202 | * caseLocale < 0: Case-folds [srcStart..srcLimit[. |
203 | */ |
204 | void toLower(int32_t caseLocale, uint32_t options, |
205 | const uint8_t *src, UCaseContext *csc, int32_t srcStart, int32_t srcLimit, |
206 | icu::ByteSink &sink, icu::Edits *edits, UErrorCode &errorCode) { |
207 | const int8_t *latinToLower; |
208 | if (caseLocale == UCASE_LOC_ROOT || |
209 | (caseLocale >= 0 ? |
210 | !(caseLocale == UCASE_LOC_TURKISH || caseLocale == UCASE_LOC_LITHUANIAN) : |
211 | (options & _FOLD_CASE_OPTIONS_MASK) == U_FOLD_CASE_DEFAULT)) { |
212 | latinToLower = LatinCase::TO_LOWER_NORMAL; |
213 | } else { |
214 | latinToLower = LatinCase::TO_LOWER_TR_LT; |
215 | } |
216 | const UTrie2 *trie = ucase_getTrie(); |
217 | int32_t prev = srcStart; |
218 | int32_t srcIndex = srcStart; |
219 | for (;;) { |
220 | // fast path for simple cases |
221 | int32_t cpStart; |
222 | UChar32 c; |
223 | for (;;) { |
224 | if (U_FAILURE(errorCode) || srcIndex >= srcLimit) { |
225 | c = U_SENTINEL; |
226 | break; |
227 | } |
228 | uint8_t lead = src[srcIndex++]; |
229 | if (lead <= 0x7f) { |
230 | int8_t d = latinToLower[lead]; |
231 | if (d == LatinCase::EXC) { |
232 | cpStart = srcIndex - 1; |
233 | c = lead; |
234 | break; |
235 | } |
236 | if (d == 0) { continue; } |
237 | ByteSinkUtil::appendUnchanged(src + prev, srcIndex - 1 - prev, |
238 | sink, options, edits, errorCode); |
239 | char ascii = (char)(lead + d); |
240 | sink.Append(&ascii, 1); |
241 | if (edits != nullptr) { |
242 | edits->addReplace(1, 1); |
243 | } |
244 | prev = srcIndex; |
245 | continue; |
246 | } else if (lead < 0xe3) { |
247 | uint8_t t; |
248 | if (0xc2 <= lead && lead <= 0xc5 && srcIndex < srcLimit && |
249 | (t = src[srcIndex] - 0x80) <= 0x3f) { |
250 | // U+0080..U+017F |
251 | ++srcIndex; |
252 | c = ((lead - 0xc0) << 6) | t; |
253 | int8_t d = latinToLower[c]; |
254 | if (d == LatinCase::EXC) { |
255 | cpStart = srcIndex - 2; |
256 | break; |
257 | } |
258 | if (d == 0) { continue; } |
259 | ByteSinkUtil::appendUnchanged(src + prev, srcIndex - 2 - prev, |
260 | sink, options, edits, errorCode); |
261 | ByteSinkUtil::appendTwoBytes(c + d, sink); |
262 | if (edits != nullptr) { |
263 | edits->addReplace(2, 2); |
264 | } |
265 | prev = srcIndex; |
266 | continue; |
267 | } |
268 | } else if ((lead <= 0xe9 || lead == 0xeb || lead == 0xec) && |
269 | (srcIndex + 2) <= srcLimit && |
270 | U8_IS_TRAIL(src[srcIndex]) && U8_IS_TRAIL(src[srcIndex + 1])) { |
271 | // most of CJK: no case mappings |
272 | srcIndex += 2; |
273 | continue; |
274 | } |
275 | cpStart = --srcIndex; |
276 | U8_NEXT(src, srcIndex, srcLimit, c); |
277 | if (c < 0) { |
278 | // ill-formed UTF-8 |
279 | continue; |
280 | } |
281 | uint16_t props = UTRIE2_GET16(trie, c); |
282 | if (UCASE_HAS_EXCEPTION(props)) { break; } |
283 | int32_t delta; |
284 | if (!UCASE_IS_UPPER_OR_TITLE(props) || (delta = UCASE_GET_DELTA(props)) == 0) { |
285 | continue; |
286 | } |
287 | ByteSinkUtil::appendUnchanged(src + prev, cpStart - prev, |
288 | sink, options, edits, errorCode); |
289 | ByteSinkUtil::appendCodePoint(srcIndex - cpStart, c + delta, sink, edits); |
290 | prev = srcIndex; |
291 | } |
292 | if (c < 0) { |
293 | break; |
294 | } |
295 | // slow path |
296 | const UChar *s; |
297 | if (caseLocale >= 0) { |
298 | csc->cpStart = cpStart; |
299 | csc->cpLimit = srcIndex; |
300 | c = ucase_toFullLower(c, utf8_caseContextIterator, csc, &s, caseLocale); |
301 | } else { |
302 | c = ucase_toFullFolding(c, &s, options); |
303 | } |
304 | if (c >= 0) { |
305 | ByteSinkUtil::appendUnchanged(src + prev, cpStart - prev, |
306 | sink, options, edits, errorCode); |
307 | appendResult(srcIndex - cpStart, c, s, sink, options, edits, errorCode); |
308 | prev = srcIndex; |
309 | } |
310 | } |
311 | ByteSinkUtil::appendUnchanged(src + prev, srcIndex - prev, |
312 | sink, options, edits, errorCode); |
313 | } |
314 | |
315 | void toUpper(int32_t caseLocale, uint32_t options, |
316 | const uint8_t *src, UCaseContext *csc, int32_t srcLength, |
317 | icu::ByteSink &sink, icu::Edits *edits, UErrorCode &errorCode) { |
318 | const int8_t *latinToUpper; |
319 | if (caseLocale == UCASE_LOC_TURKISH) { |
320 | latinToUpper = LatinCase::TO_UPPER_TR; |
321 | } else { |
322 | latinToUpper = LatinCase::TO_UPPER_NORMAL; |
323 | } |
324 | const UTrie2 *trie = ucase_getTrie(); |
325 | int32_t prev = 0; |
326 | int32_t srcIndex = 0; |
327 | for (;;) { |
328 | // fast path for simple cases |
329 | int32_t cpStart; |
330 | UChar32 c; |
331 | for (;;) { |
332 | if (U_FAILURE(errorCode) || srcIndex >= srcLength) { |
333 | c = U_SENTINEL; |
334 | break; |
335 | } |
336 | uint8_t lead = src[srcIndex++]; |
337 | if (lead <= 0x7f) { |
338 | int8_t d = latinToUpper[lead]; |
339 | if (d == LatinCase::EXC) { |
340 | cpStart = srcIndex - 1; |
341 | c = lead; |
342 | break; |
343 | } |
344 | if (d == 0) { continue; } |
345 | ByteSinkUtil::appendUnchanged(src + prev, srcIndex - 1 - prev, |
346 | sink, options, edits, errorCode); |
347 | char ascii = (char)(lead + d); |
348 | sink.Append(&ascii, 1); |
349 | if (edits != nullptr) { |
350 | edits->addReplace(1, 1); |
351 | } |
352 | prev = srcIndex; |
353 | continue; |
354 | } else if (lead < 0xe3) { |
355 | uint8_t t; |
356 | if (0xc2 <= lead && lead <= 0xc5 && srcIndex < srcLength && |
357 | (t = src[srcIndex] - 0x80) <= 0x3f) { |
358 | // U+0080..U+017F |
359 | ++srcIndex; |
360 | c = ((lead - 0xc0) << 6) | t; |
361 | int8_t d = latinToUpper[c]; |
362 | if (d == LatinCase::EXC) { |
363 | cpStart = srcIndex - 2; |
364 | break; |
365 | } |
366 | if (d == 0) { continue; } |
367 | ByteSinkUtil::appendUnchanged(src + prev, srcIndex - 2 - prev, |
368 | sink, options, edits, errorCode); |
369 | ByteSinkUtil::appendTwoBytes(c + d, sink); |
370 | if (edits != nullptr) { |
371 | edits->addReplace(2, 2); |
372 | } |
373 | prev = srcIndex; |
374 | continue; |
375 | } |
376 | } else if ((lead <= 0xe9 || lead == 0xeb || lead == 0xec) && |
377 | (srcIndex + 2) <= srcLength && |
378 | U8_IS_TRAIL(src[srcIndex]) && U8_IS_TRAIL(src[srcIndex + 1])) { |
379 | // most of CJK: no case mappings |
380 | srcIndex += 2; |
381 | continue; |
382 | } |
383 | cpStart = --srcIndex; |
384 | U8_NEXT(src, srcIndex, srcLength, c); |
385 | if (c < 0) { |
386 | // ill-formed UTF-8 |
387 | continue; |
388 | } |
389 | uint16_t props = UTRIE2_GET16(trie, c); |
390 | if (UCASE_HAS_EXCEPTION(props)) { break; } |
391 | int32_t delta; |
392 | if (UCASE_GET_TYPE(props) != UCASE_LOWER || (delta = UCASE_GET_DELTA(props)) == 0) { |
393 | continue; |
394 | } |
395 | ByteSinkUtil::appendUnchanged(src + prev, cpStart - prev, |
396 | sink, options, edits, errorCode); |
397 | ByteSinkUtil::appendCodePoint(srcIndex - cpStart, c + delta, sink, edits); |
398 | prev = srcIndex; |
399 | } |
400 | if (c < 0) { |
401 | break; |
402 | } |
403 | // slow path |
404 | csc->cpStart = cpStart; |
405 | csc->cpLimit = srcIndex; |
406 | const UChar *s; |
407 | c = ucase_toFullUpper(c, utf8_caseContextIterator, csc, &s, caseLocale); |
408 | if (c >= 0) { |
409 | ByteSinkUtil::appendUnchanged(src + prev, cpStart - prev, |
410 | sink, options, edits, errorCode); |
411 | appendResult(srcIndex - cpStart, c, s, sink, options, edits, errorCode); |
412 | prev = srcIndex; |
413 | } |
414 | } |
415 | ByteSinkUtil::appendUnchanged(src + prev, srcIndex - prev, |
416 | sink, options, edits, errorCode); |
417 | } |
418 | |
419 | } // namespace |
420 | |
421 | #if !UCONFIG_NO_BREAK_ITERATION |
422 | |
423 | U_CFUNC void U_CALLCONV |
424 | ucasemap_internalUTF8ToTitle( |
425 | int32_t caseLocale, uint32_t options, BreakIterator *iter, |
426 | const uint8_t *src, int32_t srcLength, |
427 | ByteSink &sink, icu::Edits *edits, |
428 | UErrorCode &errorCode) { |
429 | if (!ustrcase_checkTitleAdjustmentOptions(options, errorCode)) { |
430 | return; |
431 | } |
432 | |
433 | /* set up local variables */ |
434 | UCaseContext csc=UCASECONTEXT_INITIALIZER; |
435 | csc.p=(void *)src; |
436 | csc.limit=srcLength; |
437 | int32_t prev=0; |
438 | UBool isFirstIndex=TRUE; |
439 | |
440 | /* titlecasing loop */ |
441 | while(prev<srcLength) { |
442 | /* find next index where to titlecase */ |
443 | int32_t index; |
444 | if(isFirstIndex) { |
445 | isFirstIndex=FALSE; |
446 | index=iter->first(); |
447 | } else { |
448 | index=iter->next(); |
449 | } |
450 | if(index==UBRK_DONE || index>srcLength) { |
451 | index=srcLength; |
452 | } |
453 | |
454 | /* |
455 | * Segment [prev..index[ into 3 parts: |
456 | * a) skipped characters (copy as-is) [prev..titleStart[ |
457 | * b) first letter (titlecase) [titleStart..titleLimit[ |
458 | * c) subsequent characters (lowercase) [titleLimit..index[ |
459 | */ |
460 | if(prev<index) { |
461 | /* find and copy skipped characters [prev..titleStart[ */ |
462 | int32_t titleStart=prev; |
463 | int32_t titleLimit=prev; |
464 | UChar32 c; |
465 | U8_NEXT(src, titleLimit, index, c); |
466 | if ((options&U_TITLECASE_NO_BREAK_ADJUSTMENT)==0) { |
467 | // Adjust the titlecasing index to the next cased character, |
468 | // or to the next letter/number/symbol/private use. |
469 | // Stop with titleStart<titleLimit<=index |
470 | // if there is a character to be titlecased, |
471 | // or else stop with titleStart==titleLimit==index. |
472 | UBool toCased = (options&U_TITLECASE_ADJUST_TO_CASED) != 0; |
473 | while (toCased ? UCASE_NONE==ucase_getType(c) : !ustrcase_isLNS(c)) { |
474 | titleStart=titleLimit; |
475 | if(titleLimit==index) { |
476 | break; |
477 | } |
478 | U8_NEXT(src, titleLimit, index, c); |
479 | } |
480 | if (prev < titleStart) { |
481 | if (!ByteSinkUtil::appendUnchanged(src+prev, titleStart-prev, |
482 | sink, options, edits, errorCode)) { |
483 | return; |
484 | } |
485 | } |
486 | } |
487 | |
488 | if(titleStart<titleLimit) { |
489 | /* titlecase c which is from [titleStart..titleLimit[ */ |
490 | if(c>=0) { |
491 | csc.cpStart=titleStart; |
492 | csc.cpLimit=titleLimit; |
493 | const UChar *s; |
494 | c=ucase_toFullTitle(c, utf8_caseContextIterator, &csc, &s, caseLocale); |
495 | if (!appendResult(titleLimit-titleStart, c, s, sink, options, edits, errorCode)) { |
496 | return; |
497 | } |
498 | } else { |
499 | // Malformed UTF-8. |
500 | if (!ByteSinkUtil::appendUnchanged(src+titleStart, titleLimit-titleStart, |
501 | sink, options, edits, errorCode)) { |
502 | return; |
503 | } |
504 | } |
505 | |
506 | /* Special case Dutch IJ titlecasing */ |
507 | if (titleStart+1 < index && |
508 | caseLocale == UCASE_LOC_DUTCH && |
509 | (src[titleStart] == 0x0049 || src[titleStart] == 0x0069)) { |
510 | if (src[titleStart+1] == 0x006A) { |
511 | ByteSinkUtil::appendCodePoint(1, 0x004A, sink, edits); |
512 | titleLimit++; |
513 | } else if (src[titleStart+1] == 0x004A) { |
514 | // Keep the capital J from getting lowercased. |
515 | if (!ByteSinkUtil::appendUnchanged(src+titleStart+1, 1, |
516 | sink, options, edits, errorCode)) { |
517 | return; |
518 | } |
519 | titleLimit++; |
520 | } |
521 | } |
522 | |
523 | /* lowercase [titleLimit..index[ */ |
524 | if(titleLimit<index) { |
525 | if((options&U_TITLECASE_NO_LOWERCASE)==0) { |
526 | /* Normal operation: Lowercase the rest of the word. */ |
527 | toLower(caseLocale, options, |
528 | src, &csc, titleLimit, index, |
529 | sink, edits, errorCode); |
530 | if(U_FAILURE(errorCode)) { |
531 | return; |
532 | } |
533 | } else { |
534 | /* Optionally just copy the rest of the word unchanged. */ |
535 | if (!ByteSinkUtil::appendUnchanged(src+titleLimit, index-titleLimit, |
536 | sink, options, edits, errorCode)) { |
537 | return; |
538 | } |
539 | } |
540 | } |
541 | } |
542 | } |
543 | |
544 | prev=index; |
545 | } |
546 | } |
547 | |
548 | #endif |
549 | |
550 | U_NAMESPACE_BEGIN |
551 | namespace GreekUpper { |
552 | |
553 | UBool isFollowedByCasedLetter(const uint8_t *s, int32_t i, int32_t length) { |
554 | while (i < length) { |
555 | UChar32 c; |
556 | U8_NEXT(s, i, length, c); |
557 | int32_t type = ucase_getTypeOrIgnorable(c); |
558 | if ((type & UCASE_IGNORABLE) != 0) { |
559 | // Case-ignorable, continue with the loop. |
560 | } else if (type != UCASE_NONE) { |
561 | return TRUE; // Followed by cased letter. |
562 | } else { |
563 | return FALSE; // Uncased and not case-ignorable. |
564 | } |
565 | } |
566 | return FALSE; // Not followed by cased letter. |
567 | } |
568 | |
569 | // Keep this consistent with the UTF-16 version in ustrcase.cpp and the Java version in CaseMap.java. |
570 | void toUpper(uint32_t options, |
571 | const uint8_t *src, int32_t srcLength, |
572 | ByteSink &sink, Edits *edits, |
573 | UErrorCode &errorCode) { |
574 | uint32_t state = 0; |
575 | for (int32_t i = 0; i < srcLength;) { |
576 | int32_t nextIndex = i; |
577 | UChar32 c; |
578 | U8_NEXT(src, nextIndex, srcLength, c); |
579 | uint32_t nextState = 0; |
580 | int32_t type = ucase_getTypeOrIgnorable(c); |
581 | if ((type & UCASE_IGNORABLE) != 0) { |
582 | // c is case-ignorable |
583 | nextState |= (state & AFTER_CASED); |
584 | } else if (type != UCASE_NONE) { |
585 | // c is cased |
586 | nextState |= AFTER_CASED; |
587 | } |
588 | uint32_t data = getLetterData(c); |
589 | if (data > 0) { |
590 | uint32_t upper = data & UPPER_MASK; |
591 | // Add a dialytika to this iota or ypsilon vowel |
592 | // if we removed a tonos from the previous vowel, |
593 | // and that previous vowel did not also have (or gain) a dialytika. |
594 | // Adding one only to the final vowel in a longer sequence |
595 | // (which does not occur in normal writing) would require lookahead. |
596 | // Set the same flag as for preserving an existing dialytika. |
597 | if ((data & HAS_VOWEL) != 0 && (state & AFTER_VOWEL_WITH_ACCENT) != 0 && |
598 | (upper == 0x399 || upper == 0x3A5)) { |
599 | data |= HAS_DIALYTIKA; |
600 | } |
601 | int32_t numYpogegrammeni = 0; // Map each one to a trailing, spacing, capital iota. |
602 | if ((data & HAS_YPOGEGRAMMENI) != 0) { |
603 | numYpogegrammeni = 1; |
604 | } |
605 | // Skip combining diacritics after this Greek letter. |
606 | int32_t nextNextIndex = nextIndex; |
607 | while (nextIndex < srcLength) { |
608 | UChar32 c2; |
609 | U8_NEXT(src, nextNextIndex, srcLength, c2); |
610 | uint32_t diacriticData = getDiacriticData(c2); |
611 | if (diacriticData != 0) { |
612 | data |= diacriticData; |
613 | if ((diacriticData & HAS_YPOGEGRAMMENI) != 0) { |
614 | ++numYpogegrammeni; |
615 | } |
616 | nextIndex = nextNextIndex; |
617 | } else { |
618 | break; // not a Greek diacritic |
619 | } |
620 | } |
621 | if ((data & HAS_VOWEL_AND_ACCENT_AND_DIALYTIKA) == HAS_VOWEL_AND_ACCENT) { |
622 | nextState |= AFTER_VOWEL_WITH_ACCENT; |
623 | } |
624 | // Map according to Greek rules. |
625 | UBool addTonos = FALSE; |
626 | if (upper == 0x397 && |
627 | (data & HAS_ACCENT) != 0 && |
628 | numYpogegrammeni == 0 && |
629 | (state & AFTER_CASED) == 0 && |
630 | !isFollowedByCasedLetter(src, nextIndex, srcLength)) { |
631 | // Keep disjunctive "or" with (only) a tonos. |
632 | // We use the same "word boundary" conditions as for the Final_Sigma test. |
633 | if (i == nextIndex) { |
634 | upper = 0x389; // Preserve the precomposed form. |
635 | } else { |
636 | addTonos = TRUE; |
637 | } |
638 | } else if ((data & HAS_DIALYTIKA) != 0) { |
639 | // Preserve a vowel with dialytika in precomposed form if it exists. |
640 | if (upper == 0x399) { |
641 | upper = 0x3AA; |
642 | data &= ~HAS_EITHER_DIALYTIKA; |
643 | } else if (upper == 0x3A5) { |
644 | upper = 0x3AB; |
645 | data &= ~HAS_EITHER_DIALYTIKA; |
646 | } |
647 | } |
648 | |
649 | UBool change; |
650 | if (edits == nullptr && (options & U_OMIT_UNCHANGED_TEXT) == 0) { |
651 | change = TRUE; // common, simple usage |
652 | } else { |
653 | // Find out first whether we are changing the text. |
654 | U_ASSERT(0x370 <= upper && upper <= 0x3ff); // 2-byte UTF-8, main Greek block |
655 | change = (i + 2) > nextIndex || |
656 | src[i] != getTwoByteLead(upper) || src[i + 1] != getTwoByteTrail(upper) || |
657 | numYpogegrammeni > 0; |
658 | int32_t i2 = i + 2; |
659 | if ((data & HAS_EITHER_DIALYTIKA) != 0) { |
660 | change |= (i2 + 2) > nextIndex || |
661 | src[i2] != (uint8_t)u8"\u0308" [0] || |
662 | src[i2 + 1] != (uint8_t)u8"\u0308" [1]; |
663 | i2 += 2; |
664 | } |
665 | if (addTonos) { |
666 | change |= (i2 + 2) > nextIndex || |
667 | src[i2] != (uint8_t)u8"\u0301" [0] || |
668 | src[i2 + 1] != (uint8_t)u8"\u0301" [1]; |
669 | i2 += 2; |
670 | } |
671 | int32_t oldLength = nextIndex - i; |
672 | int32_t newLength = (i2 - i) + numYpogegrammeni * 2; // 2 bytes per U+0399 |
673 | change |= oldLength != newLength; |
674 | if (change) { |
675 | if (edits != NULL) { |
676 | edits->addReplace(oldLength, newLength); |
677 | } |
678 | } else { |
679 | if (edits != NULL) { |
680 | edits->addUnchanged(oldLength); |
681 | } |
682 | // Write unchanged text? |
683 | change = (options & U_OMIT_UNCHANGED_TEXT) == 0; |
684 | } |
685 | } |
686 | |
687 | if (change) { |
688 | ByteSinkUtil::appendTwoBytes(upper, sink); |
689 | if ((data & HAS_EITHER_DIALYTIKA) != 0) { |
690 | sink.Append(u8"\u0308" , 2); // restore or add a dialytika |
691 | } |
692 | if (addTonos) { |
693 | sink.Append(u8"\u0301" , 2); |
694 | } |
695 | while (numYpogegrammeni > 0) { |
696 | sink.Append(u8"\u0399" , 2); |
697 | --numYpogegrammeni; |
698 | } |
699 | } |
700 | } else if(c>=0) { |
701 | const UChar *s; |
702 | c=ucase_toFullUpper(c, NULL, NULL, &s, UCASE_LOC_GREEK); |
703 | if (!appendResult(nextIndex - i, c, s, sink, options, edits, errorCode)) { |
704 | return; |
705 | } |
706 | } else { |
707 | // Malformed UTF-8. |
708 | if (!ByteSinkUtil::appendUnchanged(src+i, nextIndex-i, |
709 | sink, options, edits, errorCode)) { |
710 | return; |
711 | } |
712 | } |
713 | i = nextIndex; |
714 | state = nextState; |
715 | } |
716 | } |
717 | |
718 | } // namespace GreekUpper |
719 | U_NAMESPACE_END |
720 | |
721 | static void U_CALLCONV |
722 | ucasemap_internalUTF8ToLower(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_UNUSED |
723 | const uint8_t *src, int32_t srcLength, |
724 | icu::ByteSink &sink, icu::Edits *edits, |
725 | UErrorCode &errorCode) { |
726 | UCaseContext csc=UCASECONTEXT_INITIALIZER; |
727 | csc.p=(void *)src; |
728 | csc.limit=srcLength; |
729 | toLower( |
730 | caseLocale, options, |
731 | src, &csc, 0, srcLength, |
732 | sink, edits, errorCode); |
733 | } |
734 | |
735 | static void U_CALLCONV |
736 | ucasemap_internalUTF8ToUpper(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_UNUSED |
737 | const uint8_t *src, int32_t srcLength, |
738 | icu::ByteSink &sink, icu::Edits *edits, |
739 | UErrorCode &errorCode) { |
740 | if (caseLocale == UCASE_LOC_GREEK) { |
741 | GreekUpper::toUpper(options, src, srcLength, sink, edits, errorCode); |
742 | } else { |
743 | UCaseContext csc=UCASECONTEXT_INITIALIZER; |
744 | csc.p=(void *)src; |
745 | csc.limit=srcLength; |
746 | toUpper( |
747 | caseLocale, options, |
748 | src, &csc, srcLength, |
749 | sink, edits, errorCode); |
750 | } |
751 | } |
752 | |
753 | static void U_CALLCONV |
754 | ucasemap_internalUTF8Fold(int32_t /* caseLocale */, uint32_t options, UCASEMAP_BREAK_ITERATOR_UNUSED |
755 | const uint8_t *src, int32_t srcLength, |
756 | icu::ByteSink &sink, icu::Edits *edits, |
757 | UErrorCode &errorCode) { |
758 | toLower( |
759 | -1, options, |
760 | src, nullptr, 0, srcLength, |
761 | sink, edits, errorCode); |
762 | } |
763 | |
764 | void |
765 | ucasemap_mapUTF8(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_PARAM |
766 | const char *src, int32_t srcLength, |
767 | UTF8CaseMapper *stringCaseMapper, |
768 | icu::ByteSink &sink, icu::Edits *edits, |
769 | UErrorCode &errorCode) { |
770 | /* check argument values */ |
771 | if (U_FAILURE(errorCode)) { |
772 | return; |
773 | } |
774 | if ((src == nullptr && srcLength != 0) || srcLength < -1) { |
775 | errorCode = U_ILLEGAL_ARGUMENT_ERROR; |
776 | return; |
777 | } |
778 | |
779 | // Get the string length. |
780 | if (srcLength == -1) { |
781 | srcLength = (int32_t)uprv_strlen((const char *)src); |
782 | } |
783 | |
784 | if (edits != nullptr && (options & U_EDITS_NO_RESET) == 0) { |
785 | edits->reset(); |
786 | } |
787 | stringCaseMapper(caseLocale, options, UCASEMAP_BREAK_ITERATOR |
788 | (const uint8_t *)src, srcLength, sink, edits, errorCode); |
789 | sink.Flush(); |
790 | if (U_SUCCESS(errorCode)) { |
791 | if (edits != nullptr) { |
792 | edits->copyErrorTo(errorCode); |
793 | } |
794 | } |
795 | } |
796 | |
797 | int32_t |
798 | ucasemap_mapUTF8(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_PARAM |
799 | char *dest, int32_t destCapacity, |
800 | const char *src, int32_t srcLength, |
801 | UTF8CaseMapper *stringCaseMapper, |
802 | icu::Edits *edits, |
803 | UErrorCode &errorCode) { |
804 | /* check argument values */ |
805 | if(U_FAILURE(errorCode)) { |
806 | return 0; |
807 | } |
808 | if( destCapacity<0 || |
809 | (dest==NULL && destCapacity>0) || |
810 | (src==NULL && srcLength!=0) || srcLength<-1 |
811 | ) { |
812 | errorCode=U_ILLEGAL_ARGUMENT_ERROR; |
813 | return 0; |
814 | } |
815 | |
816 | /* get the string length */ |
817 | if(srcLength==-1) { |
818 | srcLength=(int32_t)uprv_strlen((const char *)src); |
819 | } |
820 | |
821 | /* check for overlapping source and destination */ |
822 | if( dest!=NULL && |
823 | ((src>=dest && src<(dest+destCapacity)) || |
824 | (dest>=src && dest<(src+srcLength))) |
825 | ) { |
826 | errorCode=U_ILLEGAL_ARGUMENT_ERROR; |
827 | return 0; |
828 | } |
829 | |
830 | CheckedArrayByteSink sink(dest, destCapacity); |
831 | if (edits != nullptr && (options & U_EDITS_NO_RESET) == 0) { |
832 | edits->reset(); |
833 | } |
834 | stringCaseMapper(caseLocale, options, UCASEMAP_BREAK_ITERATOR |
835 | (const uint8_t *)src, srcLength, sink, edits, errorCode); |
836 | sink.Flush(); |
837 | if (U_SUCCESS(errorCode)) { |
838 | if (sink.Overflowed()) { |
839 | errorCode = U_BUFFER_OVERFLOW_ERROR; |
840 | } else if (edits != nullptr) { |
841 | edits->copyErrorTo(errorCode); |
842 | } |
843 | } |
844 | return u_terminateChars(dest, destCapacity, sink.NumberOfBytesAppended(), &errorCode); |
845 | } |
846 | |
847 | /* public API functions */ |
848 | |
849 | U_CAPI int32_t U_EXPORT2 |
850 | ucasemap_utf8ToLower(const UCaseMap *csm, |
851 | char *dest, int32_t destCapacity, |
852 | const char *src, int32_t srcLength, |
853 | UErrorCode *pErrorCode) { |
854 | return ucasemap_mapUTF8( |
855 | csm->caseLocale, csm->options, UCASEMAP_BREAK_ITERATOR_NULL |
856 | dest, destCapacity, |
857 | src, srcLength, |
858 | ucasemap_internalUTF8ToLower, NULL, *pErrorCode); |
859 | } |
860 | |
861 | U_CAPI int32_t U_EXPORT2 |
862 | ucasemap_utf8ToUpper(const UCaseMap *csm, |
863 | char *dest, int32_t destCapacity, |
864 | const char *src, int32_t srcLength, |
865 | UErrorCode *pErrorCode) { |
866 | return ucasemap_mapUTF8( |
867 | csm->caseLocale, csm->options, UCASEMAP_BREAK_ITERATOR_NULL |
868 | dest, destCapacity, |
869 | src, srcLength, |
870 | ucasemap_internalUTF8ToUpper, NULL, *pErrorCode); |
871 | } |
872 | |
873 | U_CAPI int32_t U_EXPORT2 |
874 | ucasemap_utf8FoldCase(const UCaseMap *csm, |
875 | char *dest, int32_t destCapacity, |
876 | const char *src, int32_t srcLength, |
877 | UErrorCode *pErrorCode) { |
878 | return ucasemap_mapUTF8( |
879 | UCASE_LOC_ROOT, csm->options, UCASEMAP_BREAK_ITERATOR_NULL |
880 | dest, destCapacity, |
881 | src, srcLength, |
882 | ucasemap_internalUTF8Fold, NULL, *pErrorCode); |
883 | } |
884 | |
885 | U_NAMESPACE_BEGIN |
886 | |
887 | void CaseMap::utf8ToLower( |
888 | const char *locale, uint32_t options, |
889 | StringPiece src, ByteSink &sink, Edits *edits, |
890 | UErrorCode &errorCode) { |
891 | ucasemap_mapUTF8( |
892 | ustrcase_getCaseLocale(locale), options, UCASEMAP_BREAK_ITERATOR_NULL |
893 | src.data(), src.length(), |
894 | ucasemap_internalUTF8ToLower, sink, edits, errorCode); |
895 | } |
896 | |
897 | void CaseMap::utf8ToUpper( |
898 | const char *locale, uint32_t options, |
899 | StringPiece src, ByteSink &sink, Edits *edits, |
900 | UErrorCode &errorCode) { |
901 | ucasemap_mapUTF8( |
902 | ustrcase_getCaseLocale(locale), options, UCASEMAP_BREAK_ITERATOR_NULL |
903 | src.data(), src.length(), |
904 | ucasemap_internalUTF8ToUpper, sink, edits, errorCode); |
905 | } |
906 | |
907 | void CaseMap::utf8Fold( |
908 | uint32_t options, |
909 | StringPiece src, ByteSink &sink, Edits *edits, |
910 | UErrorCode &errorCode) { |
911 | ucasemap_mapUTF8( |
912 | UCASE_LOC_ROOT, options, UCASEMAP_BREAK_ITERATOR_NULL |
913 | src.data(), src.length(), |
914 | ucasemap_internalUTF8Fold, sink, edits, errorCode); |
915 | } |
916 | |
917 | int32_t CaseMap::utf8ToLower( |
918 | const char *locale, uint32_t options, |
919 | const char *src, int32_t srcLength, |
920 | char *dest, int32_t destCapacity, Edits *edits, |
921 | UErrorCode &errorCode) { |
922 | return ucasemap_mapUTF8( |
923 | ustrcase_getCaseLocale(locale), options, UCASEMAP_BREAK_ITERATOR_NULL |
924 | dest, destCapacity, |
925 | src, srcLength, |
926 | ucasemap_internalUTF8ToLower, edits, errorCode); |
927 | } |
928 | |
929 | int32_t CaseMap::utf8ToUpper( |
930 | const char *locale, uint32_t options, |
931 | const char *src, int32_t srcLength, |
932 | char *dest, int32_t destCapacity, Edits *edits, |
933 | UErrorCode &errorCode) { |
934 | return ucasemap_mapUTF8( |
935 | ustrcase_getCaseLocale(locale), options, UCASEMAP_BREAK_ITERATOR_NULL |
936 | dest, destCapacity, |
937 | src, srcLength, |
938 | ucasemap_internalUTF8ToUpper, edits, errorCode); |
939 | } |
940 | |
941 | int32_t CaseMap::utf8Fold( |
942 | uint32_t options, |
943 | const char *src, int32_t srcLength, |
944 | char *dest, int32_t destCapacity, Edits *edits, |
945 | UErrorCode &errorCode) { |
946 | return ucasemap_mapUTF8( |
947 | UCASE_LOC_ROOT, options, UCASEMAP_BREAK_ITERATOR_NULL |
948 | dest, destCapacity, |
949 | src, srcLength, |
950 | ucasemap_internalUTF8Fold, edits, errorCode); |
951 | } |
952 | |
953 | U_NAMESPACE_END |
954 | |