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) 2001-2014, International Business Machines
7* Corporation and others. All Rights Reserved.
8*
9*******************************************************************************
10* file name: unormcmp.cpp
11* encoding: UTF-8
12* tab size: 8 (not used)
13* indentation:4
14*
15* created on: 2004sep13
16* created by: Markus W. Scherer
17*
18* unorm_compare() function moved here from unorm.cpp for better modularization.
19* Depends on both normalization and case folding.
20* Allows unorm.cpp to not depend on any character properties code.
21*/
22
23#include "unicode/utypes.h"
24
25#if !UCONFIG_NO_NORMALIZATION
26
27#include "unicode/unorm.h"
28#include "unicode/ustring.h"
29#include "cmemory.h"
30#include "normalizer2impl.h"
31#include "ucase.h"
32#include "uprops.h"
33#include "ustr_imp.h"
34
35U_NAMESPACE_USE
36
37/* compare canonically equivalent ------------------------------------------- */
38
39/*
40 * Compare two strings for canonical equivalence.
41 * Further options include case-insensitive comparison and
42 * code point order (as opposed to code unit order).
43 *
44 * In this function, canonical equivalence is optional as well.
45 * If canonical equivalence is tested, then both strings must fulfill
46 * the FCD check.
47 *
48 * Semantically, this is equivalent to
49 * strcmp[CodePointOrder](NFD(foldCase(s1)), NFD(foldCase(s2)))
50 * where code point order, NFD and foldCase are all optional.
51 *
52 * String comparisons almost always yield results before processing both strings
53 * completely.
54 * They are generally more efficient working incrementally instead of
55 * performing the sub-processing (strlen, normalization, case-folding)
56 * on the entire strings first.
57 *
58 * It is also unnecessary to not normalize identical characters.
59 *
60 * This function works in principle as follows:
61 *
62 * loop {
63 * get one code unit c1 from s1 (-1 if end of source)
64 * get one code unit c2 from s2 (-1 if end of source)
65 *
66 * if(either string finished) {
67 * return result;
68 * }
69 * if(c1==c2) {
70 * continue;
71 * }
72 *
73 * // c1!=c2
74 * try to decompose/case-fold c1/c2, and continue if one does;
75 *
76 * // still c1!=c2 and neither decomposes/case-folds, return result
77 * return c1-c2;
78 * }
79 *
80 * When a character decomposes, then the pointer for that source changes to
81 * the decomposition, pushing the previous pointer onto a stack.
82 * When the end of the decomposition is reached, then the code unit reader
83 * pops the previous source from the stack.
84 * (Same for case-folding.)
85 *
86 * This is complicated further by operating on variable-width UTF-16.
87 * The top part of the loop works on code units, while lookups for decomposition
88 * and case-folding need code points.
89 * Code points are assembled after the equality/end-of-source part.
90 * The source pointer is only advanced beyond all code units when the code point
91 * actually decomposes/case-folds.
92 *
93 * If we were on a trail surrogate unit when assembling a code point,
94 * and the code point decomposes/case-folds, then the decomposition/folding
95 * result must be compared with the part of the other string that corresponds to
96 * this string's lead surrogate.
97 * Since we only assemble a code point when hitting a trail unit when the
98 * preceding lead units were identical, we back up the other string by one unit
99 * in such a case.
100 *
101 * The optional code point order comparison at the end works with
102 * the same fix-up as the other code point order comparison functions.
103 * See ustring.c and the comment near the end of this function.
104 *
105 * Assumption: A decomposition or case-folding result string never contains
106 * a single surrogate. This is a safe assumption in the Unicode Standard.
107 * Therefore, we do not need to check for surrogate pairs across
108 * decomposition/case-folding boundaries.
109 *
110 * Further assumptions (see verifications tstnorm.cpp):
111 * The API function checks for FCD first, while the core function
112 * first case-folds and then decomposes. This requires that case-folding does not
113 * un-FCD any strings.
114 *
115 * The API function may also NFD the input and turn off decomposition.
116 * This requires that case-folding does not un-NFD strings either.
117 *
118 * TODO If any of the above two assumptions is violated,
119 * then this entire code must be re-thought.
120 * If this happens, then a simple solution is to case-fold both strings up front
121 * and to turn off UNORM_INPUT_IS_FCD.
122 * We already do this when not both strings are in FCD because makeFCD
123 * would be a partial NFD before the case folding, which does not work.
124 * Note that all of this is only a problem when case-folding _and_
125 * canonical equivalence come together.
126 * (Comments in unorm_compare() are more up to date than this TODO.)
127 */
128
129/* stack element for previous-level source/decomposition pointers */
130struct CmpEquivLevel {
131 const char16_t *start, *s, *limit;
132};
133typedef struct CmpEquivLevel CmpEquivLevel;
134
135/**
136 * Internal option for unorm_cmpEquivFold() for decomposing.
137 * If not set, just do strcasecmp().
138 */
139#define _COMPARE_EQUIV 0x80000
140
141/* internal function */
142static int32_t
143unorm_cmpEquivFold(const char16_t *s1, int32_t length1,
144 const char16_t *s2, int32_t length2,
145 uint32_t options,
146 UErrorCode *pErrorCode) {
147 const Normalizer2Impl *nfcImpl;
148
149 /* current-level start/limit - s1/s2 as current */
150 const char16_t *start1, *start2, *limit1, *limit2;
151
152 /* decomposition and case folding variables */
153 const char16_t *p;
154 int32_t length;
155
156 /* stacks of previous-level start/current/limit */
157 CmpEquivLevel stack1[2], stack2[2];
158
159 /* buffers for algorithmic decompositions */
160 char16_t decomp1[4], decomp2[4];
161
162 /* case folding buffers, only use current-level start/limit */
163 char16_t fold1[UCASE_MAX_STRING_LENGTH+1], fold2[UCASE_MAX_STRING_LENGTH+1];
164
165 /* track which is the current level per string */
166 int32_t level1, level2;
167
168 /* current code units, and code points for lookups */
169 UChar32 c1, c2, cp1, cp2;
170
171 /* no argument error checking because this itself is not an API */
172
173 /*
174 * assume that at least one of the options _COMPARE_EQUIV and U_COMPARE_IGNORE_CASE is set
175 * otherwise this function must behave exactly as uprv_strCompare()
176 * not checking for that here makes testing this function easier
177 */
178
179 /* normalization/properties data loaded? */
180 if((options&_COMPARE_EQUIV)!=0) {
181 nfcImpl=Normalizer2Factory::getNFCImpl(*pErrorCode);
182 } else {
183 nfcImpl=nullptr;
184 }
185 if(U_FAILURE(*pErrorCode)) {
186 return 0;
187 }
188
189 /* initialize */
190 start1=s1;
191 if(length1==-1) {
192 limit1=nullptr;
193 } else {
194 limit1=s1+length1;
195 }
196
197 start2=s2;
198 if(length2==-1) {
199 limit2=nullptr;
200 } else {
201 limit2=s2+length2;
202 }
203
204 level1=level2=0;
205 c1=c2=-1;
206
207 /* comparison loop */
208 for(;;) {
209 /*
210 * here a code unit value of -1 means "get another code unit"
211 * below it will mean "this source is finished"
212 */
213
214 if(c1<0) {
215 /* get next code unit from string 1, post-increment */
216 for(;;) {
217 if(s1==limit1 || ((c1=*s1)==0 && (limit1==nullptr || (options&_STRNCMP_STYLE)))) {
218 if(level1==0) {
219 c1=-1;
220 break;
221 }
222 } else {
223 ++s1;
224 break;
225 }
226
227 /* reached end of level buffer, pop one level */
228 do {
229 --level1;
230 start1=stack1[level1].start; /*Not uninitialized*/
231 } while(start1==nullptr);
232 s1=stack1[level1].s; /*Not uninitialized*/
233 limit1=stack1[level1].limit; /*Not uninitialized*/
234 }
235 }
236
237 if(c2<0) {
238 /* get next code unit from string 2, post-increment */
239 for(;;) {
240 if(s2==limit2 || ((c2=*s2)==0 && (limit2==nullptr || (options&_STRNCMP_STYLE)))) {
241 if(level2==0) {
242 c2=-1;
243 break;
244 }
245 } else {
246 ++s2;
247 break;
248 }
249
250 /* reached end of level buffer, pop one level */
251 do {
252 --level2;
253 start2=stack2[level2].start; /*Not uninitialized*/
254 } while(start2==nullptr);
255 s2=stack2[level2].s; /*Not uninitialized*/
256 limit2=stack2[level2].limit; /*Not uninitialized*/
257 }
258 }
259
260 /*
261 * compare c1 and c2
262 * either variable c1, c2 is -1 only if the corresponding string is finished
263 */
264 if(c1==c2) {
265 if(c1<0) {
266 return 0; /* c1==c2==-1 indicating end of strings */
267 }
268 c1=c2=-1; /* make us fetch new code units */
269 continue;
270 } else if(c1<0) {
271 return -1; /* string 1 ends before string 2 */
272 } else if(c2<0) {
273 return 1; /* string 2 ends before string 1 */
274 }
275 /* c1!=c2 && c1>=0 && c2>=0 */
276
277 /* get complete code points for c1, c2 for lookups if either is a surrogate */
278 cp1=c1;
279 if(U_IS_SURROGATE(c1)) {
280 char16_t c;
281
282 if(U_IS_SURROGATE_LEAD(c1)) {
283 if(s1!=limit1 && U16_IS_TRAIL(c=*s1)) {
284 /* advance ++s1; only below if cp1 decomposes/case-folds */
285 cp1=U16_GET_SUPPLEMENTARY(c1, c);
286 }
287 } else /* isTrail(c1) */ {
288 if(start1<=(s1-2) && U16_IS_LEAD(c=*(s1-2))) {
289 cp1=U16_GET_SUPPLEMENTARY(c, c1);
290 }
291 }
292 }
293
294 cp2=c2;
295 if(U_IS_SURROGATE(c2)) {
296 char16_t c;
297
298 if(U_IS_SURROGATE_LEAD(c2)) {
299 if(s2!=limit2 && U16_IS_TRAIL(c=*s2)) {
300 /* advance ++s2; only below if cp2 decomposes/case-folds */
301 cp2=U16_GET_SUPPLEMENTARY(c2, c);
302 }
303 } else /* isTrail(c2) */ {
304 if(start2<=(s2-2) && U16_IS_LEAD(c=*(s2-2))) {
305 cp2=U16_GET_SUPPLEMENTARY(c, c2);
306 }
307 }
308 }
309
310 /*
311 * go down one level for each string
312 * continue with the main loop as soon as there is a real change
313 */
314
315 if( level1==0 && (options&U_COMPARE_IGNORE_CASE) &&
316 (length=ucase_toFullFolding((UChar32)cp1, &p, options))>=0
317 ) {
318 /* cp1 case-folds to the code point "length" or to p[length] */
319 if(U_IS_SURROGATE(c1)) {
320 if(U_IS_SURROGATE_LEAD(c1)) {
321 /* advance beyond source surrogate pair if it case-folds */
322 ++s1;
323 } else /* isTrail(c1) */ {
324 /*
325 * we got a supplementary code point when hitting its trail surrogate,
326 * therefore the lead surrogate must have been the same as in the other string;
327 * compare this decomposition with the lead surrogate in the other string
328 * remember that this simulates bulk text replacement:
329 * the decomposition would replace the entire code point
330 */
331 --s2;
332 c2=*(s2-1);
333 }
334 }
335
336 /* push current level pointers */
337 stack1[0].start=start1;
338 stack1[0].s=s1;
339 stack1[0].limit=limit1;
340 ++level1;
341
342 /* copy the folding result to fold1[] */
343 if(length<=UCASE_MAX_STRING_LENGTH) {
344 u_memcpy(fold1, p, length);
345 } else {
346 int32_t i=0;
347 U16_APPEND_UNSAFE(fold1, i, length);
348 length=i;
349 }
350
351 /* set next level pointers to case folding */
352 start1=s1=fold1;
353 limit1=fold1+length;
354
355 /* get ready to read from decomposition, continue with loop */
356 c1=-1;
357 continue;
358 }
359
360 if( level2==0 && (options&U_COMPARE_IGNORE_CASE) &&
361 (length=ucase_toFullFolding((UChar32)cp2, &p, options))>=0
362 ) {
363 /* cp2 case-folds to the code point "length" or to p[length] */
364 if(U_IS_SURROGATE(c2)) {
365 if(U_IS_SURROGATE_LEAD(c2)) {
366 /* advance beyond source surrogate pair if it case-folds */
367 ++s2;
368 } else /* isTrail(c2) */ {
369 /*
370 * we got a supplementary code point when hitting its trail surrogate,
371 * therefore the lead surrogate must have been the same as in the other string;
372 * compare this decomposition with the lead surrogate in the other string
373 * remember that this simulates bulk text replacement:
374 * the decomposition would replace the entire code point
375 */
376 --s1;
377 c1=*(s1-1);
378 }
379 }
380
381 /* push current level pointers */
382 stack2[0].start=start2;
383 stack2[0].s=s2;
384 stack2[0].limit=limit2;
385 ++level2;
386
387 /* copy the folding result to fold2[] */
388 if(length<=UCASE_MAX_STRING_LENGTH) {
389 u_memcpy(fold2, p, length);
390 } else {
391 int32_t i=0;
392 U16_APPEND_UNSAFE(fold2, i, length);
393 length=i;
394 }
395
396 /* set next level pointers to case folding */
397 start2=s2=fold2;
398 limit2=fold2+length;
399
400 /* get ready to read from decomposition, continue with loop */
401 c2=-1;
402 continue;
403 }
404
405 if( level1<2 && (options&_COMPARE_EQUIV) &&
406 0!=(p=nfcImpl->getDecomposition((UChar32)cp1, decomp1, length))
407 ) {
408 /* cp1 decomposes into p[length] */
409 if(U_IS_SURROGATE(c1)) {
410 if(U_IS_SURROGATE_LEAD(c1)) {
411 /* advance beyond source surrogate pair if it decomposes */
412 ++s1;
413 } else /* isTrail(c1) */ {
414 /*
415 * we got a supplementary code point when hitting its trail surrogate,
416 * therefore the lead surrogate must have been the same as in the other string;
417 * compare this decomposition with the lead surrogate in the other string
418 * remember that this simulates bulk text replacement:
419 * the decomposition would replace the entire code point
420 */
421 --s2;
422 c2=*(s2-1);
423 }
424 }
425
426 /* push current level pointers */
427 stack1[level1].start=start1;
428 stack1[level1].s=s1;
429 stack1[level1].limit=limit1;
430 ++level1;
431
432 /* set empty intermediate level if skipped */
433 if(level1<2) {
434 stack1[level1++].start=nullptr;
435 }
436
437 /* set next level pointers to decomposition */
438 start1=s1=p;
439 limit1=p+length;
440
441 /* get ready to read from decomposition, continue with loop */
442 c1=-1;
443 continue;
444 }
445
446 if( level2<2 && (options&_COMPARE_EQUIV) &&
447 0!=(p=nfcImpl->getDecomposition((UChar32)cp2, decomp2, length))
448 ) {
449 /* cp2 decomposes into p[length] */
450 if(U_IS_SURROGATE(c2)) {
451 if(U_IS_SURROGATE_LEAD(c2)) {
452 /* advance beyond source surrogate pair if it decomposes */
453 ++s2;
454 } else /* isTrail(c2) */ {
455 /*
456 * we got a supplementary code point when hitting its trail surrogate,
457 * therefore the lead surrogate must have been the same as in the other string;
458 * compare this decomposition with the lead surrogate in the other string
459 * remember that this simulates bulk text replacement:
460 * the decomposition would replace the entire code point
461 */
462 --s1;
463 c1=*(s1-1);
464 }
465 }
466
467 /* push current level pointers */
468 stack2[level2].start=start2;
469 stack2[level2].s=s2;
470 stack2[level2].limit=limit2;
471 ++level2;
472
473 /* set empty intermediate level if skipped */
474 if(level2<2) {
475 stack2[level2++].start=nullptr;
476 }
477
478 /* set next level pointers to decomposition */
479 start2=s2=p;
480 limit2=p+length;
481
482 /* get ready to read from decomposition, continue with loop */
483 c2=-1;
484 continue;
485 }
486
487 /*
488 * no decomposition/case folding, max level for both sides:
489 * return difference result
490 *
491 * code point order comparison must not just return cp1-cp2
492 * because when single surrogates are present then the surrogate pairs
493 * that formed cp1 and cp2 may be from different string indexes
494 *
495 * example: { d800 d800 dc01 } vs. { d800 dc00 }, compare at second code units
496 * c1=d800 cp1=10001 c2=dc00 cp2=10000
497 * cp1-cp2>0 but c1-c2<0 and in fact in UTF-32 it is { d800 10001 } < { 10000 }
498 *
499 * therefore, use same fix-up as in ustring.c/uprv_strCompare()
500 * except: uprv_strCompare() fetches c=*s while this functions fetches c=*s++
501 * so we have slightly different pointer/start/limit comparisons here
502 */
503
504 if(c1>=0xd800 && c2>=0xd800 && (options&U_COMPARE_CODE_POINT_ORDER)) {
505 /* subtract 0x2800 from BMP code points to make them smaller than supplementary ones */
506 if(
507 (c1<=0xdbff && s1!=limit1 && U16_IS_TRAIL(*s1)) ||
508 (U16_IS_TRAIL(c1) && start1!=(s1-1) && U16_IS_LEAD(*(s1-2)))
509 ) {
510 /* part of a surrogate pair, leave >=d800 */
511 } else {
512 /* BMP code point - may be surrogate code point - make <d800 */
513 c1-=0x2800;
514 }
515
516 if(
517 (c2<=0xdbff && s2!=limit2 && U16_IS_TRAIL(*s2)) ||
518 (U16_IS_TRAIL(c2) && start2!=(s2-1) && U16_IS_LEAD(*(s2-2)))
519 ) {
520 /* part of a surrogate pair, leave >=d800 */
521 } else {
522 /* BMP code point - may be surrogate code point - make <d800 */
523 c2-=0x2800;
524 }
525 }
526
527 return c1-c2;
528 }
529}
530
531static
532UBool _normalize(const Normalizer2 *n2, const char16_t *s, int32_t length,
533 UnicodeString &normalized, UErrorCode *pErrorCode) {
534 UnicodeString str(length<0, s, length);
535
536 // check if s fulfill the conditions
537 int32_t spanQCYes=n2->spanQuickCheckYes(str, *pErrorCode);
538 if (U_FAILURE(*pErrorCode)) {
539 return false;
540 }
541 /*
542 * ICU 2.4 had a further optimization:
543 * If both strings were not in FCD, then they were both NFD'ed,
544 * and the _COMPARE_EQUIV option was turned off.
545 * It is not entirely clear that this is valid with the current
546 * definition of the canonical caseless match.
547 * Therefore, ICU 2.6 removes that optimization.
548 */
549 if(spanQCYes<str.length()) {
550 UnicodeString unnormalized=str.tempSubString(spanQCYes);
551 normalized.setTo(false, str.getBuffer(), spanQCYes);
552 n2->normalizeSecondAndAppend(normalized, unnormalized, *pErrorCode);
553 if (U_SUCCESS(*pErrorCode)) {
554 return true;
555 }
556 }
557 return false;
558}
559
560U_CAPI int32_t U_EXPORT2
561unorm_compare(const char16_t *s1, int32_t length1,
562 const char16_t *s2, int32_t length2,
563 uint32_t options,
564 UErrorCode *pErrorCode) {
565 /* argument checking */
566 if(U_FAILURE(*pErrorCode)) {
567 return 0;
568 }
569 if(s1==0 || length1<-1 || s2==0 || length2<-1) {
570 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
571 return 0;
572 }
573
574 UnicodeString fcd1, fcd2;
575 int32_t normOptions=(int32_t)(options>>UNORM_COMPARE_NORM_OPTIONS_SHIFT);
576 options|=_COMPARE_EQUIV;
577
578 /*
579 * UAX #21 Case Mappings, as fixed for Unicode version 4
580 * (see Jitterbug 2021), defines a canonical caseless match as
581 *
582 * A string X is a canonical caseless match
583 * for a string Y if and only if
584 * NFD(toCasefold(NFD(X))) = NFD(toCasefold(NFD(Y)))
585 *
586 * For better performance, we check for FCD (or let the caller tell us that
587 * both strings are in FCD) for the inner normalization.
588 * BasicNormalizerTest::FindFoldFCDExceptions() makes sure that
589 * case-folding preserves the FCD-ness of a string.
590 * The outer normalization is then only performed by unorm_cmpEquivFold()
591 * when there is a difference.
592 *
593 * Exception: When using the Turkic case-folding option, we do perform
594 * full NFD first. This is because in the Turkic case precomposed characters
595 * with 0049 capital I or 0069 small i fold differently whether they
596 * are first decomposed or not, so an FCD check - a check only for
597 * canonical order - is not sufficient.
598 */
599 if(!(options&UNORM_INPUT_IS_FCD) || (options&U_FOLD_CASE_EXCLUDE_SPECIAL_I)) {
600 const Normalizer2 *n2;
601 if(options&U_FOLD_CASE_EXCLUDE_SPECIAL_I) {
602 n2=Normalizer2::getNFDInstance(*pErrorCode);
603 } else {
604 n2=Normalizer2Factory::getFCDInstance(*pErrorCode);
605 }
606 if (U_FAILURE(*pErrorCode)) {
607 return 0;
608 }
609
610 if(normOptions&UNORM_UNICODE_3_2) {
611 const UnicodeSet *uni32=uniset_getUnicode32Instance(*pErrorCode);
612 FilteredNormalizer2 fn2(*n2, *uni32);
613 if(_normalize(&fn2, s1, length1, fcd1, pErrorCode)) {
614 s1=fcd1.getBuffer();
615 length1=fcd1.length();
616 }
617 if(_normalize(&fn2, s2, length2, fcd2, pErrorCode)) {
618 s2=fcd2.getBuffer();
619 length2=fcd2.length();
620 }
621 } else {
622 if(_normalize(n2, s1, length1, fcd1, pErrorCode)) {
623 s1=fcd1.getBuffer();
624 length1=fcd1.length();
625 }
626 if(_normalize(n2, s2, length2, fcd2, pErrorCode)) {
627 s2=fcd2.getBuffer();
628 length2=fcd2.length();
629 }
630 }
631 }
632
633 if(U_SUCCESS(*pErrorCode)) {
634 return unorm_cmpEquivFold(s1, length1, s2, length2, options, pErrorCode);
635 } else {
636 return 0;
637 }
638}
639
640#endif /* #if !UCONFIG_NO_NORMALIZATION */
641