]>
Commit | Line | Data |
---|---|---|
b75a7d8f A |
1 | /* |
2 | ******************************************************************************* | |
3 | * | |
374ca955 | 4 | * Copyright (C) 2001-2004, International Business Machines |
b75a7d8f A |
5 | * Corporation and others. All Rights Reserved. |
6 | * | |
7 | ******************************************************************************* | |
8 | * file name: ustrcase.c | |
9 | * encoding: US-ASCII | |
10 | * tab size: 8 (not used) | |
11 | * indentation:4 | |
12 | * | |
13 | * created on: 2002feb20 | |
14 | * created by: Markus W. Scherer | |
15 | * | |
16 | * Implementation file for string casing C API functions. | |
17 | * Uses functions from uchar.c for basic functionality that requires access | |
18 | * to the Unicode Character Database (uprops.dat). | |
19 | */ | |
20 | ||
21 | #include "unicode/utypes.h" | |
374ca955 | 22 | #include "unicode/uloc.h" |
b75a7d8f A |
23 | #include "unicode/ustring.h" |
24 | #include "unicode/ubrk.h" | |
25 | #include "cmemory.h" | |
374ca955 | 26 | #include "ucase.h" |
b75a7d8f A |
27 | #include "unormimp.h" |
28 | #include "ustr_imp.h" | |
29 | ||
30 | /* string casing ------------------------------------------------------------ */ | |
31 | ||
374ca955 A |
32 | /* append a full case mapping result, see UCASE_MAX_STRING_LENGTH */ |
33 | static U_INLINE int32_t | |
34 | appendResult(UChar *dest, int32_t destIndex, int32_t destCapacity, | |
35 | int32_t result, const UChar *s) { | |
36 | UChar32 c; | |
37 | int32_t length; | |
38 | ||
39 | /* decode the result */ | |
40 | if(result<0) { | |
41 | /* (not) original code point */ | |
42 | c=~result; | |
43 | length=-1; | |
44 | } else if(result<=UCASE_MAX_STRING_LENGTH) { | |
45 | c=U_SENTINEL; | |
46 | length=result; | |
47 | } else { | |
48 | c=result; | |
49 | length=-1; | |
50 | } | |
51 | ||
52 | if(destIndex<destCapacity) { | |
53 | /* append the result */ | |
54 | if(length<0) { | |
55 | /* code point */ | |
56 | UBool isError=FALSE; | |
57 | U16_APPEND(dest, destIndex, destCapacity, c, isError); | |
58 | if(isError) { | |
59 | /* overflow, nothing written */ | |
60 | destIndex+=U16_LENGTH(c); | |
61 | } | |
62 | } else { | |
63 | /* string */ | |
64 | if((destIndex+length)<=destCapacity) { | |
65 | while(length>0) { | |
66 | dest[destIndex++]=*s++; | |
67 | --length; | |
68 | } | |
69 | } else { | |
70 | /* overflow */ | |
71 | destIndex+=length; | |
72 | } | |
73 | } | |
74 | } else { | |
75 | /* preflight */ | |
76 | if(length<0) { | |
77 | destIndex+=U16_LENGTH(c); | |
78 | } else { | |
79 | destIndex+=length; | |
80 | } | |
81 | } | |
82 | return destIndex; | |
83 | } | |
84 | ||
85 | static UChar32 U_CALLCONV | |
86 | utf16_caseContextIterator(void *context, int8_t dir) { | |
87 | UCaseContext *csc=(UCaseContext *)context; | |
88 | UChar32 c; | |
89 | ||
90 | if(dir<0) { | |
91 | /* reset for backward iteration */ | |
92 | csc->index=csc->cpStart; | |
93 | csc->dir=dir; | |
94 | } else if(dir>0) { | |
95 | /* reset for forward iteration */ | |
96 | csc->index=csc->cpLimit; | |
97 | csc->dir=dir; | |
98 | } else { | |
99 | /* continue current iteration direction */ | |
100 | dir=csc->dir; | |
101 | } | |
102 | ||
103 | if(dir<0) { | |
104 | if(csc->start<csc->index) { | |
105 | U16_PREV((const UChar *)csc->p, csc->start, csc->index, c); | |
106 | return c; | |
107 | } | |
108 | } else { | |
109 | if(csc->index<csc->limit) { | |
110 | U16_NEXT((const UChar *)csc->p, csc->index, csc->limit, c); | |
111 | return c; | |
112 | } | |
113 | } | |
114 | return U_SENTINEL; | |
115 | } | |
116 | ||
117 | typedef int32_t U_CALLCONV | |
118 | UCaseMapFull(const UCaseProps *csp, UChar32 c, | |
119 | UCaseContextIterator *iter, void *context, | |
120 | const UChar **pString, | |
121 | const char *locale, int32_t *locCache); | |
122 | ||
123 | /* | |
124 | * Lowercases [srcStart..srcLimit[ but takes | |
125 | * context [0..srcLength[ into account. | |
126 | */ | |
127 | static int32_t | |
128 | _caseMap(UCaseProps *csp, UCaseMapFull *map, | |
129 | UChar *dest, int32_t destCapacity, | |
130 | const UChar *src, UCaseContext *csc, | |
131 | int32_t srcStart, int32_t srcLimit, | |
132 | const char *locale, int32_t *locCache, | |
133 | UErrorCode *pErrorCode) { | |
134 | const UChar *s; | |
135 | UChar32 c; | |
136 | int32_t srcIndex, destIndex; | |
137 | ||
138 | /* case mapping loop */ | |
139 | srcIndex=srcStart; | |
140 | destIndex=0; | |
141 | while(srcIndex<srcLimit) { | |
142 | csc->cpStart=srcIndex; | |
143 | U16_NEXT(src, srcIndex, srcLimit, c); | |
144 | csc->cpLimit=srcIndex; | |
145 | c=map(csp, c, utf16_caseContextIterator, csc, &s, locale, locCache); | |
146 | destIndex=appendResult(dest, destIndex, destCapacity, c, s); | |
147 | } | |
148 | ||
149 | if(destIndex>destCapacity) { | |
150 | *pErrorCode=U_BUFFER_OVERFLOW_ERROR; | |
151 | } | |
152 | return destIndex; | |
153 | } | |
154 | ||
b75a7d8f A |
155 | #if !UCONFIG_NO_BREAK_ITERATION |
156 | ||
157 | /* | |
374ca955 | 158 | * Internal titlecasing function. |
b75a7d8f A |
159 | * |
160 | * Must get titleIter!=NULL. | |
161 | */ | |
374ca955 A |
162 | static int32_t |
163 | _toTitle(UCaseProps *csp, | |
164 | UChar *dest, int32_t destCapacity, | |
165 | const UChar *src, UCaseContext *csc, | |
166 | int32_t srcLength, | |
167 | UBreakIterator *titleIter, | |
168 | const char *locale, int32_t *locCache, | |
169 | UErrorCode *pErrorCode) { | |
170 | const UChar *s; | |
b75a7d8f | 171 | UChar32 c; |
374ca955 | 172 | int32_t prev, index, destIndex; |
b75a7d8f A |
173 | UBool isFirstIndex; |
174 | ||
175 | /* set up local variables */ | |
b75a7d8f A |
176 | destIndex=0; |
177 | prev=0; | |
178 | isFirstIndex=TRUE; | |
179 | ||
180 | /* titlecasing loop */ | |
181 | while(prev<srcLength) { | |
182 | /* find next index where to titlecase */ | |
183 | if(isFirstIndex) { | |
184 | isFirstIndex=FALSE; | |
185 | index=ubrk_first(titleIter); | |
186 | } else { | |
187 | index=ubrk_next(titleIter); | |
188 | } | |
189 | if(index==UBRK_DONE || index>srcLength) { | |
190 | index=srcLength; | |
191 | } | |
192 | ||
193 | /* lowercase [prev..index[ */ | |
194 | if(prev<index) { | |
374ca955 A |
195 | destIndex+= |
196 | _caseMap( | |
197 | csp, ucase_toFullLower, | |
198 | dest+destIndex, destCapacity-destIndex, | |
199 | src, csc, | |
200 | prev, index, | |
201 | locale, locCache, | |
202 | pErrorCode); | |
b75a7d8f A |
203 | } |
204 | ||
205 | if(index>=srcLength) { | |
206 | break; | |
207 | } | |
208 | ||
209 | /* titlecase the character at the found index */ | |
374ca955 A |
210 | csc->cpStart=index; |
211 | U16_NEXT(src, index, srcLength, c); | |
212 | csc->cpLimit=index; | |
213 | c=ucase_toFullTitle(csp, c, utf16_caseContextIterator, csc, &s, locale, locCache); | |
214 | destIndex=appendResult(dest, destIndex, destCapacity, c, s); | |
b75a7d8f A |
215 | |
216 | prev=index; | |
217 | } | |
218 | ||
374ca955 A |
219 | if(destIndex>destCapacity) { |
220 | *pErrorCode=U_BUFFER_OVERFLOW_ERROR; | |
221 | } | |
b75a7d8f A |
222 | return destIndex; |
223 | } | |
224 | ||
374ca955 A |
225 | U_CFUNC int32_t |
226 | ustr_toTitle(UCaseProps *csp, | |
227 | UChar *dest, int32_t destCapacity, | |
228 | const UChar *src, int32_t srcLength, | |
229 | UBreakIterator *titleIter, | |
230 | const char *locale, | |
231 | UErrorCode *pErrorCode) { | |
232 | UCaseContext csc={ NULL }; | |
233 | int32_t locCache; | |
234 | ||
235 | csc.p=(void *)src; | |
236 | csc.limit=srcLength; | |
237 | locCache=0; | |
238 | ||
239 | return _toTitle(csp, | |
240 | dest, destCapacity, | |
241 | src, &csc, srcLength, | |
242 | titleIter, locale, &locCache, pErrorCode); | |
243 | } | |
244 | ||
b75a7d8f A |
245 | #endif |
246 | ||
374ca955 A |
247 | /* functions available in the common library (for unistr_case.cpp) */ |
248 | ||
249 | U_CFUNC int32_t | |
250 | ustr_toLower(UCaseProps *csp, | |
251 | UChar *dest, int32_t destCapacity, | |
252 | const UChar *src, int32_t srcLength, | |
253 | const char *locale, | |
254 | UErrorCode *pErrorCode) { | |
255 | UCaseContext csc={ NULL }; | |
256 | int32_t locCache; | |
257 | ||
258 | csc.p=(void *)src; | |
259 | csc.limit=srcLength; | |
260 | locCache=0; | |
261 | ||
262 | return _caseMap(csp, ucase_toFullLower, | |
263 | dest, destCapacity, | |
264 | src, &csc, 0, srcLength, | |
265 | locale, &locCache, pErrorCode); | |
266 | } | |
267 | ||
268 | U_CFUNC int32_t | |
269 | ustr_toUpper(UCaseProps *csp, | |
270 | UChar *dest, int32_t destCapacity, | |
271 | const UChar *src, int32_t srcLength, | |
272 | const char *locale, | |
273 | UErrorCode *pErrorCode) { | |
274 | UCaseContext csc={ NULL }; | |
275 | int32_t locCache; | |
276 | ||
277 | csc.p=(void *)src; | |
278 | csc.limit=srcLength; | |
279 | locCache=0; | |
280 | ||
281 | return _caseMap(csp, ucase_toFullUpper, | |
282 | dest, destCapacity, | |
283 | src, &csc, 0, srcLength, | |
284 | locale, &locCache, pErrorCode); | |
285 | } | |
286 | ||
287 | U_CFUNC int32_t | |
288 | ustr_foldCase(UCaseProps *csp, | |
289 | UChar *dest, int32_t destCapacity, | |
290 | const UChar *src, int32_t srcLength, | |
291 | uint32_t options, | |
292 | UErrorCode *pErrorCode) { | |
293 | int32_t srcIndex, destIndex; | |
294 | ||
295 | const UChar *s; | |
296 | UChar32 c; | |
297 | ||
298 | /* case mapping loop */ | |
299 | srcIndex=destIndex=0; | |
300 | while(srcIndex<srcLength) { | |
301 | U16_NEXT(src, srcIndex, srcLength, c); | |
302 | c=ucase_toFullFolding(csp, c, &s, options); | |
303 | destIndex=appendResult(dest, destIndex, destCapacity, c, s); | |
304 | } | |
305 | ||
306 | if(destIndex>destCapacity) { | |
307 | *pErrorCode=U_BUFFER_OVERFLOW_ERROR; | |
308 | } | |
309 | return destIndex; | |
310 | } | |
311 | ||
b75a7d8f A |
312 | /* |
313 | * Implement argument checking and buffer handling | |
314 | * for string case mapping as a common function. | |
315 | */ | |
316 | enum { | |
317 | TO_LOWER, | |
318 | TO_UPPER, | |
319 | TO_TITLE, | |
320 | FOLD_CASE | |
321 | }; | |
322 | ||
374ca955 A |
323 | /* common internal function for public API functions */ |
324 | ||
b75a7d8f | 325 | static int32_t |
374ca955 A |
326 | caseMap(UChar *dest, int32_t destCapacity, |
327 | const UChar *src, int32_t srcLength, | |
328 | UBreakIterator *titleIter, | |
329 | const char *locale, | |
330 | uint32_t options, | |
331 | int32_t toWhichCase, | |
332 | UErrorCode *pErrorCode) { | |
b75a7d8f A |
333 | UChar buffer[300]; |
334 | UChar *temp; | |
374ca955 A |
335 | |
336 | UCaseProps *csp; | |
337 | ||
b75a7d8f A |
338 | int32_t destLength; |
339 | UBool ownTitleIter; | |
340 | ||
341 | /* check argument values */ | |
342 | if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { | |
343 | return 0; | |
344 | } | |
345 | if( destCapacity<0 || | |
346 | (dest==NULL && destCapacity>0) || | |
347 | src==NULL || | |
348 | srcLength<-1 | |
349 | ) { | |
350 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; | |
351 | return 0; | |
352 | } | |
353 | ||
374ca955 A |
354 | csp=ucase_getSingleton(pErrorCode); |
355 | if(U_FAILURE(*pErrorCode)) { | |
356 | return 0; | |
357 | } | |
358 | ||
b75a7d8f A |
359 | /* get the string length */ |
360 | if(srcLength==-1) { | |
361 | srcLength=u_strlen(src); | |
362 | } | |
363 | ||
364 | /* check for overlapping source and destination */ | |
365 | if( dest!=NULL && | |
366 | ((src>=dest && src<(dest+destCapacity)) || | |
367 | (dest>=src && dest<(src+srcLength))) | |
368 | ) { | |
369 | /* overlap: provide a temporary destination buffer and later copy the result */ | |
370 | if(destCapacity<=(sizeof(buffer)/U_SIZEOF_UCHAR)) { | |
371 | /* the stack buffer is large enough */ | |
372 | temp=buffer; | |
373 | } else { | |
374 | /* allocate a buffer */ | |
375 | temp=(UChar *)uprv_malloc(destCapacity*U_SIZEOF_UCHAR); | |
376 | if(temp==NULL) { | |
377 | *pErrorCode=U_MEMORY_ALLOCATION_ERROR; | |
378 | return 0; | |
379 | } | |
380 | } | |
381 | } else { | |
382 | temp=dest; | |
383 | } | |
384 | ||
385 | ownTitleIter=FALSE; | |
386 | destLength=0; | |
387 | ||
374ca955 A |
388 | if(toWhichCase==FOLD_CASE) { |
389 | destLength=ustr_foldCase(csp, temp, destCapacity, src, srcLength, | |
390 | options, pErrorCode); | |
391 | } else { | |
392 | UCaseContext csc={ NULL }; | |
393 | int32_t locCache; | |
394 | ||
395 | csc.p=(void *)src; | |
396 | csc.limit=srcLength; | |
397 | locCache=0; | |
398 | ||
399 | /* the internal functions require locale!=NULL */ | |
400 | if(locale==NULL) { | |
401 | locale=uloc_getDefault(); | |
b75a7d8f | 402 | } |
374ca955 A |
403 | |
404 | if(toWhichCase==TO_LOWER) { | |
405 | destLength=_caseMap(csp, ucase_toFullLower, | |
406 | temp, destCapacity, | |
407 | src, &csc, | |
408 | 0, srcLength, | |
409 | locale, &locCache, pErrorCode); | |
410 | } else if(toWhichCase==TO_UPPER) { | |
411 | destLength=_caseMap(csp, ucase_toFullUpper, | |
412 | temp, destCapacity, | |
413 | src, &csc, | |
414 | 0, srcLength, | |
415 | locale, &locCache, pErrorCode); | |
416 | } else /* if(toWhichCase==TO_TITLE) */ { | |
417 | #if UCONFIG_NO_BREAK_ITERATION | |
418 | *pErrorCode=U_UNSUPPORTED_ERROR; | |
419 | #else | |
420 | if(titleIter==NULL) { | |
421 | titleIter=ubrk_open(UBRK_WORD, locale, | |
422 | src, srcLength, | |
423 | pErrorCode); | |
424 | ownTitleIter=(UBool)U_SUCCESS(*pErrorCode); | |
425 | } | |
426 | if(U_SUCCESS(*pErrorCode)) { | |
427 | destLength=_toTitle(csp, temp, destCapacity, | |
428 | src, &csc, srcLength, | |
429 | titleIter, locale, &locCache, pErrorCode); | |
430 | } | |
431 | #endif | |
b75a7d8f | 432 | } |
b75a7d8f A |
433 | } |
434 | if(temp!=dest) { | |
435 | /* copy the result string to the destination buffer */ | |
436 | if(destLength>0) { | |
437 | int32_t copyLength= destLength<=destCapacity ? destLength : destCapacity; | |
438 | if(copyLength>0) { | |
439 | uprv_memmove(dest, temp, copyLength*U_SIZEOF_UCHAR); | |
440 | } | |
441 | } | |
442 | if(temp!=buffer) { | |
443 | uprv_free(temp); | |
444 | } | |
445 | } | |
446 | ||
447 | #if !UCONFIG_NO_BREAK_ITERATION | |
448 | if(ownTitleIter) { | |
449 | ubrk_close(titleIter); | |
450 | } | |
451 | #endif | |
452 | ||
453 | return u_terminateUChars(dest, destCapacity, destLength, pErrorCode); | |
454 | } | |
455 | ||
374ca955 A |
456 | /* public API functions */ |
457 | ||
b75a7d8f A |
458 | U_CAPI int32_t U_EXPORT2 |
459 | u_strToLower(UChar *dest, int32_t destCapacity, | |
460 | const UChar *src, int32_t srcLength, | |
461 | const char *locale, | |
462 | UErrorCode *pErrorCode) { | |
374ca955 A |
463 | return caseMap(dest, destCapacity, |
464 | src, srcLength, | |
465 | NULL, locale, 0, | |
466 | TO_LOWER, pErrorCode); | |
b75a7d8f A |
467 | } |
468 | ||
469 | U_CAPI int32_t U_EXPORT2 | |
470 | u_strToUpper(UChar *dest, int32_t destCapacity, | |
471 | const UChar *src, int32_t srcLength, | |
472 | const char *locale, | |
473 | UErrorCode *pErrorCode) { | |
374ca955 A |
474 | return caseMap(dest, destCapacity, |
475 | src, srcLength, | |
476 | NULL, locale, 0, | |
477 | TO_UPPER, pErrorCode); | |
b75a7d8f A |
478 | } |
479 | ||
374ca955 A |
480 | #if !UCONFIG_NO_BREAK_ITERATION |
481 | ||
b75a7d8f A |
482 | U_CAPI int32_t U_EXPORT2 |
483 | u_strToTitle(UChar *dest, int32_t destCapacity, | |
484 | const UChar *src, int32_t srcLength, | |
485 | UBreakIterator *titleIter, | |
486 | const char *locale, | |
487 | UErrorCode *pErrorCode) { | |
374ca955 A |
488 | return caseMap(dest, destCapacity, |
489 | src, srcLength, | |
490 | titleIter, locale, 0, | |
491 | TO_TITLE, pErrorCode); | |
b75a7d8f A |
492 | } |
493 | ||
374ca955 A |
494 | #endif |
495 | ||
b75a7d8f A |
496 | U_CAPI int32_t U_EXPORT2 |
497 | u_strFoldCase(UChar *dest, int32_t destCapacity, | |
498 | const UChar *src, int32_t srcLength, | |
499 | uint32_t options, | |
500 | UErrorCode *pErrorCode) { | |
374ca955 A |
501 | return caseMap(dest, destCapacity, |
502 | src, srcLength, | |
503 | NULL, NULL, options, | |
504 | FOLD_CASE, pErrorCode); | |
505 | } | |
506 | ||
507 | /* case-insensitive string comparisons -------------------------------------- */ | |
508 | ||
509 | /* | |
510 | * This function is a copy of unorm_cmpEquivFold() minus the parts for | |
511 | * canonical equivalence. | |
512 | * Keep the functions in sync, and see there for how this works. | |
513 | * The duplication is for modularization: | |
514 | * It makes caseless (but not canonical caseless) matches independent of | |
515 | * the normalization code. | |
516 | */ | |
517 | ||
518 | /* stack element for previous-level source/decomposition pointers */ | |
519 | struct CmpEquivLevel { | |
520 | const UChar *start, *s, *limit; | |
521 | }; | |
522 | typedef struct CmpEquivLevel CmpEquivLevel; | |
523 | ||
524 | /* internal function */ | |
525 | U_CFUNC int32_t | |
526 | u_strcmpFold(const UChar *s1, int32_t length1, | |
527 | const UChar *s2, int32_t length2, | |
528 | uint32_t options, | |
529 | UErrorCode *pErrorCode) { | |
530 | UCaseProps *csp; | |
531 | ||
532 | /* current-level start/limit - s1/s2 as current */ | |
533 | const UChar *start1, *start2, *limit1, *limit2; | |
534 | ||
535 | /* case folding variables */ | |
536 | const UChar *p; | |
537 | int32_t length; | |
538 | ||
539 | /* stacks of previous-level start/current/limit */ | |
540 | CmpEquivLevel stack1[2], stack2[2]; | |
541 | ||
542 | /* case folding buffers, only use current-level start/limit */ | |
543 | UChar fold1[UCASE_MAX_STRING_LENGTH+1], fold2[UCASE_MAX_STRING_LENGTH+1]; | |
544 | ||
545 | /* track which is the current level per string */ | |
546 | int32_t level1, level2; | |
547 | ||
548 | /* current code units, and code points for lookups */ | |
549 | UChar32 c1, c2, cp1, cp2; | |
550 | ||
551 | /* no argument error checking because this itself is not an API */ | |
552 | ||
553 | /* | |
554 | * assume that at least the option U_COMPARE_IGNORE_CASE is set | |
555 | * otherwise this function would have to behave exactly as uprv_strCompare() | |
556 | */ | |
557 | csp=ucase_getSingleton(pErrorCode); | |
558 | if(U_FAILURE(*pErrorCode)) { | |
559 | return 0; | |
560 | } | |
561 | ||
562 | /* initialize */ | |
563 | start1=s1; | |
564 | if(length1==-1) { | |
565 | limit1=NULL; | |
566 | } else { | |
567 | limit1=s1+length1; | |
568 | } | |
569 | ||
570 | start2=s2; | |
571 | if(length2==-1) { | |
572 | limit2=NULL; | |
573 | } else { | |
574 | limit2=s2+length2; | |
575 | } | |
576 | ||
577 | level1=level2=0; | |
578 | c1=c2=-1; | |
579 | ||
580 | /* comparison loop */ | |
581 | for(;;) { | |
582 | /* | |
583 | * here a code unit value of -1 means "get another code unit" | |
584 | * below it will mean "this source is finished" | |
585 | */ | |
586 | ||
587 | if(c1<0) { | |
588 | /* get next code unit from string 1, post-increment */ | |
589 | for(;;) { | |
590 | if(s1==limit1 || ((c1=*s1)==0 && (limit1==NULL || (options&_STRNCMP_STYLE)))) { | |
591 | if(level1==0) { | |
592 | c1=-1; | |
593 | break; | |
594 | } | |
595 | } else { | |
596 | ++s1; | |
597 | break; | |
598 | } | |
599 | ||
600 | /* reached end of level buffer, pop one level */ | |
601 | do { | |
602 | --level1; | |
603 | start1=stack1[level1].start; | |
604 | } while(start1==NULL); | |
605 | s1=stack1[level1].s; | |
606 | limit1=stack1[level1].limit; | |
607 | } | |
608 | } | |
609 | ||
610 | if(c2<0) { | |
611 | /* get next code unit from string 2, post-increment */ | |
612 | for(;;) { | |
613 | if(s2==limit2 || ((c2=*s2)==0 && (limit2==NULL || (options&_STRNCMP_STYLE)))) { | |
614 | if(level2==0) { | |
615 | c2=-1; | |
616 | break; | |
617 | } | |
618 | } else { | |
619 | ++s2; | |
620 | break; | |
621 | } | |
622 | ||
623 | /* reached end of level buffer, pop one level */ | |
624 | do { | |
625 | --level2; | |
626 | start2=stack2[level2].start; | |
627 | } while(start2==NULL); | |
628 | s2=stack2[level2].s; | |
629 | limit2=stack2[level2].limit; | |
630 | } | |
631 | } | |
632 | ||
633 | /* | |
634 | * compare c1 and c2 | |
635 | * either variable c1, c2 is -1 only if the corresponding string is finished | |
636 | */ | |
637 | if(c1==c2) { | |
638 | if(c1<0) { | |
639 | return 0; /* c1==c2==-1 indicating end of strings */ | |
640 | } | |
641 | c1=c2=-1; /* make us fetch new code units */ | |
642 | continue; | |
643 | } else if(c1<0) { | |
644 | return -1; /* string 1 ends before string 2 */ | |
645 | } else if(c2<0) { | |
646 | return 1; /* string 2 ends before string 1 */ | |
647 | } | |
648 | /* c1!=c2 && c1>=0 && c2>=0 */ | |
649 | ||
650 | /* get complete code points for c1, c2 for lookups if either is a surrogate */ | |
651 | cp1=c1; | |
652 | if(U_IS_SURROGATE(c1)) { | |
653 | UChar c; | |
654 | ||
655 | if(U_IS_SURROGATE_LEAD(c1)) { | |
656 | if(s1!=limit1 && U16_IS_TRAIL(c=*s1)) { | |
657 | /* advance ++s1; only below if cp1 decomposes/case-folds */ | |
658 | cp1=U16_GET_SUPPLEMENTARY(c1, c); | |
659 | } | |
660 | } else /* isTrail(c1) */ { | |
661 | if(start1<=(s1-2) && U16_IS_LEAD(c=*(s1-2))) { | |
662 | cp1=U16_GET_SUPPLEMENTARY(c, c1); | |
663 | } | |
664 | } | |
665 | } | |
666 | ||
667 | cp2=c2; | |
668 | if(U_IS_SURROGATE(c2)) { | |
669 | UChar c; | |
670 | ||
671 | if(U_IS_SURROGATE_LEAD(c2)) { | |
672 | if(s2!=limit2 && U16_IS_TRAIL(c=*s2)) { | |
673 | /* advance ++s2; only below if cp2 decomposes/case-folds */ | |
674 | cp2=U16_GET_SUPPLEMENTARY(c2, c); | |
675 | } | |
676 | } else /* isTrail(c2) */ { | |
677 | if(start2<=(s2-2) && U16_IS_LEAD(c=*(s2-2))) { | |
678 | cp2=U16_GET_SUPPLEMENTARY(c, c2); | |
679 | } | |
680 | } | |
681 | } | |
682 | ||
683 | /* | |
684 | * go down one level for each string | |
685 | * continue with the main loop as soon as there is a real change | |
686 | */ | |
687 | ||
688 | if( level1==0 && | |
689 | (length=ucase_toFullFolding(csp, (UChar32)cp1, &p, options))>=0 | |
690 | ) { | |
691 | /* cp1 case-folds to the code point "length" or to p[length] */ | |
692 | if(U_IS_SURROGATE(c1)) { | |
693 | if(U_IS_SURROGATE_LEAD(c1)) { | |
694 | /* advance beyond source surrogate pair if it case-folds */ | |
695 | ++s1; | |
696 | } else /* isTrail(c1) */ { | |
697 | /* | |
698 | * we got a supplementary code point when hitting its trail surrogate, | |
699 | * therefore the lead surrogate must have been the same as in the other string; | |
700 | * compare this decomposition with the lead surrogate in the other string | |
701 | * remember that this simulates bulk text replacement: | |
702 | * the decomposition would replace the entire code point | |
703 | */ | |
704 | --s2; | |
705 | c2=*(s2-1); | |
706 | } | |
707 | } | |
708 | ||
709 | /* push current level pointers */ | |
710 | stack1[0].start=start1; | |
711 | stack1[0].s=s1; | |
712 | stack1[0].limit=limit1; | |
713 | ++level1; | |
714 | ||
715 | /* copy the folding result to fold1[] */ | |
716 | if(length<=UCASE_MAX_STRING_LENGTH) { | |
717 | u_memcpy(fold1, p, length); | |
718 | } else { | |
719 | int32_t i=0; | |
720 | U16_APPEND_UNSAFE(fold1, i, length); | |
721 | length=i; | |
722 | } | |
723 | ||
724 | /* set next level pointers to case folding */ | |
725 | start1=s1=fold1; | |
726 | limit1=fold1+length; | |
727 | ||
728 | /* get ready to read from decomposition, continue with loop */ | |
729 | c1=-1; | |
730 | continue; | |
731 | } | |
732 | ||
733 | if( level2==0 && | |
734 | (length=ucase_toFullFolding(csp, (UChar32)cp2, &p, options))>=0 | |
735 | ) { | |
736 | /* cp2 case-folds to the code point "length" or to p[length] */ | |
737 | if(U_IS_SURROGATE(c2)) { | |
738 | if(U_IS_SURROGATE_LEAD(c2)) { | |
739 | /* advance beyond source surrogate pair if it case-folds */ | |
740 | ++s2; | |
741 | } else /* isTrail(c2) */ { | |
742 | /* | |
743 | * we got a supplementary code point when hitting its trail surrogate, | |
744 | * therefore the lead surrogate must have been the same as in the other string; | |
745 | * compare this decomposition with the lead surrogate in the other string | |
746 | * remember that this simulates bulk text replacement: | |
747 | * the decomposition would replace the entire code point | |
748 | */ | |
749 | --s1; | |
750 | c1=*(s1-1); | |
751 | } | |
752 | } | |
753 | ||
754 | /* push current level pointers */ | |
755 | stack2[0].start=start2; | |
756 | stack2[0].s=s2; | |
757 | stack2[0].limit=limit2; | |
758 | ++level2; | |
759 | ||
760 | /* copy the folding result to fold2[] */ | |
761 | if(length<=UCASE_MAX_STRING_LENGTH) { | |
762 | u_memcpy(fold2, p, length); | |
763 | } else { | |
764 | int32_t i=0; | |
765 | U16_APPEND_UNSAFE(fold2, i, length); | |
766 | length=i; | |
767 | } | |
768 | ||
769 | /* set next level pointers to case folding */ | |
770 | start2=s2=fold2; | |
771 | limit2=fold2+length; | |
772 | ||
773 | /* get ready to read from decomposition, continue with loop */ | |
774 | c2=-1; | |
775 | continue; | |
776 | } | |
777 | ||
778 | /* | |
779 | * no decomposition/case folding, max level for both sides: | |
780 | * return difference result | |
781 | * | |
782 | * code point order comparison must not just return cp1-cp2 | |
783 | * because when single surrogates are present then the surrogate pairs | |
784 | * that formed cp1 and cp2 may be from different string indexes | |
785 | * | |
786 | * example: { d800 d800 dc01 } vs. { d800 dc00 }, compare at second code units | |
787 | * c1=d800 cp1=10001 c2=dc00 cp2=10000 | |
788 | * cp1-cp2>0 but c1-c2<0 and in fact in UTF-32 it is { d800 10001 } < { 10000 } | |
789 | * | |
790 | * therefore, use same fix-up as in ustring.c/uprv_strCompare() | |
791 | * except: uprv_strCompare() fetches c=*s while this functions fetches c=*s++ | |
792 | * so we have slightly different pointer/start/limit comparisons here | |
793 | */ | |
794 | ||
795 | if(c1>=0xd800 && c2>=0xd800 && (options&U_COMPARE_CODE_POINT_ORDER)) { | |
796 | /* subtract 0x2800 from BMP code points to make them smaller than supplementary ones */ | |
797 | if( | |
798 | (c1<=0xdbff && s1!=limit1 && U16_IS_TRAIL(*s1)) || | |
799 | (U16_IS_TRAIL(c1) && start1!=(s1-1) && U16_IS_LEAD(*(s1-2))) | |
800 | ) { | |
801 | /* part of a surrogate pair, leave >=d800 */ | |
802 | } else { | |
803 | /* BMP code point - may be surrogate code point - make <d800 */ | |
804 | c1-=0x2800; | |
805 | } | |
806 | ||
807 | if( | |
808 | (c2<=0xdbff && s2!=limit2 && U16_IS_TRAIL(*s2)) || | |
809 | (U16_IS_TRAIL(c2) && start2!=(s2-1) && U16_IS_LEAD(*(s2-2))) | |
810 | ) { | |
811 | /* part of a surrogate pair, leave >=d800 */ | |
812 | } else { | |
813 | /* BMP code point - may be surrogate code point - make <d800 */ | |
814 | c2-=0x2800; | |
815 | } | |
816 | } | |
817 | ||
818 | return c1-c2; | |
819 | } | |
b75a7d8f A |
820 | } |
821 | ||
374ca955 | 822 | /* public API functions */ |
b75a7d8f A |
823 | |
824 | U_CAPI int32_t U_EXPORT2 | |
825 | u_strCaseCompare(const UChar *s1, int32_t length1, | |
826 | const UChar *s2, int32_t length2, | |
827 | uint32_t options, | |
828 | UErrorCode *pErrorCode) { | |
829 | /* argument checking */ | |
830 | if(pErrorCode==0 || U_FAILURE(*pErrorCode)) { | |
831 | return 0; | |
832 | } | |
833 | if(s1==NULL || length1<-1 || s2==NULL || length2<-1) { | |
834 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; | |
835 | return 0; | |
836 | } | |
374ca955 A |
837 | return u_strcmpFold(s1, length1, s2, length2, |
838 | options|U_COMPARE_IGNORE_CASE, | |
839 | pErrorCode); | |
b75a7d8f A |
840 | } |
841 | ||
842 | U_CAPI int32_t U_EXPORT2 | |
843 | u_strcasecmp(const UChar *s1, const UChar *s2, uint32_t options) { | |
844 | UErrorCode errorCode=U_ZERO_ERROR; | |
374ca955 A |
845 | return u_strcmpFold(s1, -1, s2, -1, |
846 | options|U_COMPARE_IGNORE_CASE, | |
847 | &errorCode); | |
b75a7d8f A |
848 | } |
849 | ||
850 | U_CAPI int32_t U_EXPORT2 | |
851 | u_memcasecmp(const UChar *s1, const UChar *s2, int32_t length, uint32_t options) { | |
852 | UErrorCode errorCode=U_ZERO_ERROR; | |
374ca955 A |
853 | return u_strcmpFold(s1, length, s2, length, |
854 | options|U_COMPARE_IGNORE_CASE, | |
855 | &errorCode); | |
b75a7d8f A |
856 | } |
857 | ||
858 | U_CAPI int32_t U_EXPORT2 | |
859 | u_strncasecmp(const UChar *s1, const UChar *s2, int32_t n, uint32_t options) { | |
860 | UErrorCode errorCode=U_ZERO_ERROR; | |
374ca955 A |
861 | return u_strcmpFold(s1, n, s2, n, |
862 | options|(U_COMPARE_IGNORE_CASE|_STRNCMP_STYLE), | |
863 | &errorCode); | |
b75a7d8f | 864 | } |