|
1 /* |
|
2 ******************************************************************************* |
|
3 * |
|
4 * Copyright (C) 2005-2011, International Business Machines |
|
5 * Corporation and others. All Rights Reserved. |
|
6 * |
|
7 ******************************************************************************* |
|
8 * file name: ucasemap.cpp |
|
9 * encoding: US-ASCII |
|
10 * tab size: 8 (not used) |
|
11 * indentation:4 |
|
12 * |
|
13 * created on: 2005may06 |
|
14 * created by: Markus W. Scherer |
|
15 * |
|
16 * Case mapping service object and functions using it. |
|
17 */ |
|
18 |
|
19 #include "unicode/utypes.h" |
|
20 #include "unicode/brkiter.h" |
|
21 #include "unicode/ubrk.h" |
|
22 #include "unicode/uloc.h" |
|
23 #include "unicode/ustring.h" |
|
24 #include "unicode/ucasemap.h" |
|
25 #if !UCONFIG_NO_BREAK_ITERATION |
|
26 #include "unicode/utext.h" |
|
27 #endif |
|
28 #include "unicode/utf.h" |
|
29 #include "unicode/utf8.h" |
|
30 #include "unicode/utf16.h" |
|
31 #include "cmemory.h" |
|
32 #include "cstring.h" |
|
33 #include "ucase.h" |
|
34 #include "ustr_imp.h" |
|
35 |
|
36 U_NAMESPACE_USE |
|
37 |
|
38 /* UCaseMap service object -------------------------------------------------- */ |
|
39 |
|
40 U_CAPI UCaseMap * U_EXPORT2 |
|
41 ucasemap_open(const char *locale, uint32_t options, UErrorCode *pErrorCode) { |
|
42 UCaseMap *csm; |
|
43 |
|
44 if(U_FAILURE(*pErrorCode)) { |
|
45 return NULL; |
|
46 } |
|
47 |
|
48 csm=(UCaseMap *)uprv_malloc(sizeof(UCaseMap)); |
|
49 if(csm==NULL) { |
|
50 return NULL; |
|
51 } |
|
52 uprv_memset(csm, 0, sizeof(UCaseMap)); |
|
53 |
|
54 csm->csp=ucase_getSingleton(); |
|
55 ucasemap_setLocale(csm, locale, pErrorCode); |
|
56 if(U_FAILURE(*pErrorCode)) { |
|
57 uprv_free(csm); |
|
58 return NULL; |
|
59 } |
|
60 |
|
61 csm->options=options; |
|
62 return csm; |
|
63 } |
|
64 |
|
65 U_CAPI void U_EXPORT2 |
|
66 ucasemap_close(UCaseMap *csm) { |
|
67 if(csm!=NULL) { |
|
68 #if !UCONFIG_NO_BREAK_ITERATION |
|
69 // Do not call ubrk_close() so that we do not depend on all of the BreakIterator code. |
|
70 delete reinterpret_cast<BreakIterator *>(csm->iter); |
|
71 #endif |
|
72 uprv_free(csm); |
|
73 } |
|
74 } |
|
75 |
|
76 U_CAPI const char * U_EXPORT2 |
|
77 ucasemap_getLocale(const UCaseMap *csm) { |
|
78 return csm->locale; |
|
79 } |
|
80 |
|
81 U_CAPI uint32_t U_EXPORT2 |
|
82 ucasemap_getOptions(const UCaseMap *csm) { |
|
83 return csm->options; |
|
84 } |
|
85 |
|
86 U_CAPI void U_EXPORT2 |
|
87 ucasemap_setLocale(UCaseMap *csm, const char *locale, UErrorCode *pErrorCode) { |
|
88 int32_t length; |
|
89 |
|
90 if(U_FAILURE(*pErrorCode)) { |
|
91 return; |
|
92 } |
|
93 |
|
94 length=uloc_getName(locale, csm->locale, (int32_t)sizeof(csm->locale), pErrorCode); |
|
95 if(*pErrorCode==U_BUFFER_OVERFLOW_ERROR || length==sizeof(csm->locale)) { |
|
96 *pErrorCode=U_ZERO_ERROR; |
|
97 /* we only really need the language code for case mappings */ |
|
98 length=uloc_getLanguage(locale, csm->locale, (int32_t)sizeof(csm->locale), pErrorCode); |
|
99 } |
|
100 if(length==sizeof(csm->locale)) { |
|
101 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
|
102 } |
|
103 csm->locCache=0; |
|
104 if(U_SUCCESS(*pErrorCode)) { |
|
105 ucase_getCaseLocale(csm->locale, &csm->locCache); |
|
106 } else { |
|
107 csm->locale[0]=0; |
|
108 } |
|
109 } |
|
110 |
|
111 U_CAPI void U_EXPORT2 |
|
112 ucasemap_setOptions(UCaseMap *csm, uint32_t options, UErrorCode * /*pErrorCode*/) { |
|
113 csm->options=options; |
|
114 } |
|
115 |
|
116 /* UTF-8 string case mappings ----------------------------------------------- */ |
|
117 |
|
118 /* TODO(markus): Move to a new, separate utf8case.c file. */ |
|
119 |
|
120 /* append a full case mapping result, see UCASE_MAX_STRING_LENGTH */ |
|
121 static inline int32_t |
|
122 appendResult(uint8_t *dest, int32_t destIndex, int32_t destCapacity, |
|
123 int32_t result, const UChar *s) { |
|
124 UChar32 c; |
|
125 int32_t length, destLength; |
|
126 UErrorCode errorCode; |
|
127 |
|
128 /* decode the result */ |
|
129 if(result<0) { |
|
130 /* (not) original code point */ |
|
131 c=~result; |
|
132 length=-1; |
|
133 } else if(result<=UCASE_MAX_STRING_LENGTH) { |
|
134 c=U_SENTINEL; |
|
135 length=result; |
|
136 } else { |
|
137 c=result; |
|
138 length=-1; |
|
139 } |
|
140 |
|
141 if(destIndex<destCapacity) { |
|
142 /* append the result */ |
|
143 if(length<0) { |
|
144 /* code point */ |
|
145 UBool isError=FALSE; |
|
146 U8_APPEND(dest, destIndex, destCapacity, c, isError); |
|
147 if(isError) { |
|
148 /* overflow, nothing written */ |
|
149 destIndex+=U8_LENGTH(c); |
|
150 } |
|
151 } else { |
|
152 /* string */ |
|
153 errorCode=U_ZERO_ERROR; |
|
154 u_strToUTF8( |
|
155 (char *)(dest+destIndex), destCapacity-destIndex, &destLength, |
|
156 s, length, |
|
157 &errorCode); |
|
158 destIndex+=destLength; |
|
159 /* we might have an overflow, but we know the actual length */ |
|
160 } |
|
161 } else { |
|
162 /* preflight */ |
|
163 if(length<0) { |
|
164 destIndex+=U8_LENGTH(c); |
|
165 } else { |
|
166 errorCode=U_ZERO_ERROR; |
|
167 u_strToUTF8( |
|
168 NULL, 0, &destLength, |
|
169 s, length, |
|
170 &errorCode); |
|
171 destIndex+=destLength; |
|
172 } |
|
173 } |
|
174 return destIndex; |
|
175 } |
|
176 |
|
177 static UChar32 U_CALLCONV |
|
178 utf8_caseContextIterator(void *context, int8_t dir) { |
|
179 UCaseContext *csc=(UCaseContext *)context; |
|
180 UChar32 c; |
|
181 |
|
182 if(dir<0) { |
|
183 /* reset for backward iteration */ |
|
184 csc->index=csc->cpStart; |
|
185 csc->dir=dir; |
|
186 } else if(dir>0) { |
|
187 /* reset for forward iteration */ |
|
188 csc->index=csc->cpLimit; |
|
189 csc->dir=dir; |
|
190 } else { |
|
191 /* continue current iteration direction */ |
|
192 dir=csc->dir; |
|
193 } |
|
194 |
|
195 if(dir<0) { |
|
196 if(csc->start<csc->index) { |
|
197 U8_PREV((const uint8_t *)csc->p, csc->start, csc->index, c); |
|
198 return c; |
|
199 } |
|
200 } else { |
|
201 if(csc->index<csc->limit) { |
|
202 U8_NEXT((const uint8_t *)csc->p, csc->index, csc->limit, c); |
|
203 return c; |
|
204 } |
|
205 } |
|
206 return U_SENTINEL; |
|
207 } |
|
208 |
|
209 /* |
|
210 * Case-maps [srcStart..srcLimit[ but takes |
|
211 * context [0..srcLength[ into account. |
|
212 */ |
|
213 static int32_t |
|
214 _caseMap(const UCaseMap *csm, UCaseMapFull *map, |
|
215 uint8_t *dest, int32_t destCapacity, |
|
216 const uint8_t *src, UCaseContext *csc, |
|
217 int32_t srcStart, int32_t srcLimit, |
|
218 UErrorCode *pErrorCode) { |
|
219 const UChar *s; |
|
220 UChar32 c, c2 = 0; |
|
221 int32_t srcIndex, destIndex; |
|
222 int32_t locCache; |
|
223 |
|
224 locCache=csm->locCache; |
|
225 |
|
226 /* case mapping loop */ |
|
227 srcIndex=srcStart; |
|
228 destIndex=0; |
|
229 while(srcIndex<srcLimit) { |
|
230 csc->cpStart=srcIndex; |
|
231 U8_NEXT(src, srcIndex, srcLimit, c); |
|
232 csc->cpLimit=srcIndex; |
|
233 if(c<0) { |
|
234 int32_t i=csc->cpStart; |
|
235 while(destIndex<destCapacity && i<srcIndex) { |
|
236 dest[destIndex++]=src[i++]; |
|
237 } |
|
238 continue; |
|
239 } |
|
240 c=map(csm->csp, c, utf8_caseContextIterator, csc, &s, csm->locale, &locCache); |
|
241 if((destIndex<destCapacity) && (c<0 ? (c2=~c)<=0x7f : UCASE_MAX_STRING_LENGTH<c && (c2=c)<=0x7f)) { |
|
242 /* fast path version of appendResult() for ASCII results */ |
|
243 dest[destIndex++]=(uint8_t)c2; |
|
244 } else { |
|
245 destIndex=appendResult(dest, destIndex, destCapacity, c, s); |
|
246 } |
|
247 } |
|
248 |
|
249 if(destIndex>destCapacity) { |
|
250 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
|
251 } |
|
252 return destIndex; |
|
253 } |
|
254 |
|
255 #if !UCONFIG_NO_BREAK_ITERATION |
|
256 |
|
257 U_CFUNC int32_t U_CALLCONV |
|
258 ucasemap_internalUTF8ToTitle(const UCaseMap *csm, |
|
259 uint8_t *dest, int32_t destCapacity, |
|
260 const uint8_t *src, int32_t srcLength, |
|
261 UErrorCode *pErrorCode) { |
|
262 const UChar *s; |
|
263 UChar32 c; |
|
264 int32_t prev, titleStart, titleLimit, idx, destIndex, length; |
|
265 UBool isFirstIndex; |
|
266 |
|
267 if(U_FAILURE(*pErrorCode)) { |
|
268 return 0; |
|
269 } |
|
270 |
|
271 // Use the C++ abstract base class to minimize dependencies. |
|
272 // TODO: Change UCaseMap.iter to store a BreakIterator directly. |
|
273 BreakIterator *bi=reinterpret_cast<BreakIterator *>(csm->iter); |
|
274 |
|
275 /* set up local variables */ |
|
276 int32_t locCache=csm->locCache; |
|
277 UCaseContext csc=UCASECONTEXT_INITIALIZER; |
|
278 csc.p=(void *)src; |
|
279 csc.limit=srcLength; |
|
280 destIndex=0; |
|
281 prev=0; |
|
282 isFirstIndex=TRUE; |
|
283 |
|
284 /* titlecasing loop */ |
|
285 while(prev<srcLength) { |
|
286 /* find next index where to titlecase */ |
|
287 if(isFirstIndex) { |
|
288 isFirstIndex=FALSE; |
|
289 idx=bi->first(); |
|
290 } else { |
|
291 idx=bi->next(); |
|
292 } |
|
293 if(idx==UBRK_DONE || idx>srcLength) { |
|
294 idx=srcLength; |
|
295 } |
|
296 |
|
297 /* |
|
298 * Unicode 4 & 5 section 3.13 Default Case Operations: |
|
299 * |
|
300 * R3 toTitlecase(X): Find the word boundaries based on Unicode Standard Annex |
|
301 * #29, "Text Boundaries." Between each pair of word boundaries, find the first |
|
302 * cased character F. If F exists, map F to default_title(F); then map each |
|
303 * subsequent character C to default_lower(C). |
|
304 * |
|
305 * In this implementation, segment [prev..index[ into 3 parts: |
|
306 * a) uncased characters (copy as-is) [prev..titleStart[ |
|
307 * b) first case letter (titlecase) [titleStart..titleLimit[ |
|
308 * c) subsequent characters (lowercase) [titleLimit..index[ |
|
309 */ |
|
310 if(prev<idx) { |
|
311 /* find and copy uncased characters [prev..titleStart[ */ |
|
312 titleStart=titleLimit=prev; |
|
313 U8_NEXT(src, titleLimit, idx, c); |
|
314 if((csm->options&U_TITLECASE_NO_BREAK_ADJUSTMENT)==0 && UCASE_NONE==ucase_getType(csm->csp, c)) { |
|
315 /* Adjust the titlecasing index (titleStart) to the next cased character. */ |
|
316 for(;;) { |
|
317 titleStart=titleLimit; |
|
318 if(titleLimit==idx) { |
|
319 /* |
|
320 * only uncased characters in [prev..index[ |
|
321 * stop with titleStart==titleLimit==index |
|
322 */ |
|
323 break; |
|
324 } |
|
325 U8_NEXT(src, titleLimit, idx, c); |
|
326 if(UCASE_NONE!=ucase_getType(csm->csp, c)) { |
|
327 break; /* cased letter at [titleStart..titleLimit[ */ |
|
328 } |
|
329 } |
|
330 length=titleStart-prev; |
|
331 if(length>0) { |
|
332 if((destIndex+length)<=destCapacity) { |
|
333 uprv_memcpy(dest+destIndex, src+prev, length); |
|
334 } |
|
335 destIndex+=length; |
|
336 } |
|
337 } |
|
338 |
|
339 if(titleStart<titleLimit) { |
|
340 /* titlecase c which is from [titleStart..titleLimit[ */ |
|
341 csc.cpStart=titleStart; |
|
342 csc.cpLimit=titleLimit; |
|
343 c=ucase_toFullTitle(csm->csp, c, utf8_caseContextIterator, &csc, &s, csm->locale, &locCache); |
|
344 destIndex=appendResult(dest, destIndex, destCapacity, c, s); |
|
345 |
|
346 /* Special case Dutch IJ titlecasing */ |
|
347 if ( titleStart+1 < idx && |
|
348 ucase_getCaseLocale(csm->locale, &locCache) == UCASE_LOC_DUTCH && |
|
349 ( src[titleStart] == 0x0049 || src[titleStart] == 0x0069 ) && |
|
350 ( src[titleStart+1] == 0x004A || src[titleStart+1] == 0x006A )) { |
|
351 c=0x004A; |
|
352 destIndex=appendResult(dest, destIndex, destCapacity, c, s); |
|
353 titleLimit++; |
|
354 } |
|
355 /* lowercase [titleLimit..index[ */ |
|
356 if(titleLimit<idx) { |
|
357 if((csm->options&U_TITLECASE_NO_LOWERCASE)==0) { |
|
358 /* Normal operation: Lowercase the rest of the word. */ |
|
359 destIndex+= |
|
360 _caseMap( |
|
361 csm, ucase_toFullLower, |
|
362 dest+destIndex, destCapacity-destIndex, |
|
363 src, &csc, |
|
364 titleLimit, idx, |
|
365 pErrorCode); |
|
366 } else { |
|
367 /* Optionally just copy the rest of the word unchanged. */ |
|
368 length=idx-titleLimit; |
|
369 if((destIndex+length)<=destCapacity) { |
|
370 uprv_memcpy(dest+destIndex, src+titleLimit, length); |
|
371 } |
|
372 destIndex+=length; |
|
373 } |
|
374 } |
|
375 } |
|
376 } |
|
377 |
|
378 prev=idx; |
|
379 } |
|
380 |
|
381 if(destIndex>destCapacity) { |
|
382 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
|
383 } |
|
384 return destIndex; |
|
385 } |
|
386 |
|
387 #endif |
|
388 |
|
389 static int32_t U_CALLCONV |
|
390 ucasemap_internalUTF8ToLower(const UCaseMap *csm, |
|
391 uint8_t *dest, int32_t destCapacity, |
|
392 const uint8_t *src, int32_t srcLength, |
|
393 UErrorCode *pErrorCode) { |
|
394 UCaseContext csc=UCASECONTEXT_INITIALIZER; |
|
395 csc.p=(void *)src; |
|
396 csc.limit=srcLength; |
|
397 return _caseMap( |
|
398 csm, ucase_toFullLower, |
|
399 dest, destCapacity, |
|
400 src, &csc, 0, srcLength, |
|
401 pErrorCode); |
|
402 } |
|
403 |
|
404 static int32_t U_CALLCONV |
|
405 ucasemap_internalUTF8ToUpper(const UCaseMap *csm, |
|
406 uint8_t *dest, int32_t destCapacity, |
|
407 const uint8_t *src, int32_t srcLength, |
|
408 UErrorCode *pErrorCode) { |
|
409 UCaseContext csc=UCASECONTEXT_INITIALIZER; |
|
410 csc.p=(void *)src; |
|
411 csc.limit=srcLength; |
|
412 return _caseMap( |
|
413 csm, ucase_toFullUpper, |
|
414 dest, destCapacity, |
|
415 src, &csc, 0, srcLength, |
|
416 pErrorCode); |
|
417 } |
|
418 |
|
419 static int32_t |
|
420 utf8_foldCase(const UCaseProps *csp, |
|
421 uint8_t *dest, int32_t destCapacity, |
|
422 const uint8_t *src, int32_t srcLength, |
|
423 uint32_t options, |
|
424 UErrorCode *pErrorCode) { |
|
425 int32_t srcIndex, destIndex; |
|
426 |
|
427 const UChar *s; |
|
428 UChar32 c, c2; |
|
429 int32_t start; |
|
430 |
|
431 /* case mapping loop */ |
|
432 srcIndex=destIndex=0; |
|
433 while(srcIndex<srcLength) { |
|
434 start=srcIndex; |
|
435 U8_NEXT(src, srcIndex, srcLength, c); |
|
436 if(c<0) { |
|
437 while(destIndex<destCapacity && start<srcIndex) { |
|
438 dest[destIndex++]=src[start++]; |
|
439 } |
|
440 continue; |
|
441 } |
|
442 c=ucase_toFullFolding(csp, c, &s, options); |
|
443 if((destIndex<destCapacity) && (c<0 ? (c2=~c)<=0x7f : UCASE_MAX_STRING_LENGTH<c && (c2=c)<=0x7f)) { |
|
444 /* fast path version of appendResult() for ASCII results */ |
|
445 dest[destIndex++]=(uint8_t)c2; |
|
446 } else { |
|
447 destIndex=appendResult(dest, destIndex, destCapacity, c, s); |
|
448 } |
|
449 } |
|
450 |
|
451 if(destIndex>destCapacity) { |
|
452 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
|
453 } |
|
454 return destIndex; |
|
455 } |
|
456 |
|
457 static int32_t U_CALLCONV |
|
458 ucasemap_internalUTF8Fold(const UCaseMap *csm, |
|
459 uint8_t *dest, int32_t destCapacity, |
|
460 const uint8_t *src, int32_t srcLength, |
|
461 UErrorCode *pErrorCode) { |
|
462 return utf8_foldCase(csm->csp, dest, destCapacity, src, srcLength, csm->options, pErrorCode); |
|
463 } |
|
464 |
|
465 U_CFUNC int32_t |
|
466 ucasemap_mapUTF8(const UCaseMap *csm, |
|
467 uint8_t *dest, int32_t destCapacity, |
|
468 const uint8_t *src, int32_t srcLength, |
|
469 UTF8CaseMapper *stringCaseMapper, |
|
470 UErrorCode *pErrorCode) { |
|
471 int32_t destLength; |
|
472 |
|
473 /* check argument values */ |
|
474 if(U_FAILURE(*pErrorCode)) { |
|
475 return 0; |
|
476 } |
|
477 if( destCapacity<0 || |
|
478 (dest==NULL && destCapacity>0) || |
|
479 src==NULL || |
|
480 srcLength<-1 |
|
481 ) { |
|
482 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
|
483 return 0; |
|
484 } |
|
485 |
|
486 /* get the string length */ |
|
487 if(srcLength==-1) { |
|
488 srcLength=(int32_t)uprv_strlen((const char *)src); |
|
489 } |
|
490 |
|
491 /* check for overlapping source and destination */ |
|
492 if( dest!=NULL && |
|
493 ((src>=dest && src<(dest+destCapacity)) || |
|
494 (dest>=src && dest<(src+srcLength))) |
|
495 ) { |
|
496 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
|
497 return 0; |
|
498 } |
|
499 |
|
500 destLength=stringCaseMapper(csm, dest, destCapacity, src, srcLength, pErrorCode); |
|
501 return u_terminateChars((char *)dest, destCapacity, destLength, pErrorCode); |
|
502 } |
|
503 |
|
504 /* public API functions */ |
|
505 |
|
506 U_CAPI int32_t U_EXPORT2 |
|
507 ucasemap_utf8ToLower(const UCaseMap *csm, |
|
508 char *dest, int32_t destCapacity, |
|
509 const char *src, int32_t srcLength, |
|
510 UErrorCode *pErrorCode) { |
|
511 return ucasemap_mapUTF8(csm, |
|
512 (uint8_t *)dest, destCapacity, |
|
513 (const uint8_t *)src, srcLength, |
|
514 ucasemap_internalUTF8ToLower, pErrorCode); |
|
515 } |
|
516 |
|
517 U_CAPI int32_t U_EXPORT2 |
|
518 ucasemap_utf8ToUpper(const UCaseMap *csm, |
|
519 char *dest, int32_t destCapacity, |
|
520 const char *src, int32_t srcLength, |
|
521 UErrorCode *pErrorCode) { |
|
522 return ucasemap_mapUTF8(csm, |
|
523 (uint8_t *)dest, destCapacity, |
|
524 (const uint8_t *)src, srcLength, |
|
525 ucasemap_internalUTF8ToUpper, pErrorCode); |
|
526 } |
|
527 |
|
528 U_CAPI int32_t U_EXPORT2 |
|
529 ucasemap_utf8FoldCase(const UCaseMap *csm, |
|
530 char *dest, int32_t destCapacity, |
|
531 const char *src, int32_t srcLength, |
|
532 UErrorCode *pErrorCode) { |
|
533 return ucasemap_mapUTF8(csm, |
|
534 (uint8_t *)dest, destCapacity, |
|
535 (const uint8_t *)src, srcLength, |
|
536 ucasemap_internalUTF8Fold, pErrorCode); |
|
537 } |