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: utrie2.cpp |
11 | * encoding: UTF-8 |
12 | * tab size: 8 (not used) |
13 | * indentation:4 |
14 | * |
15 | * created on: 2008aug16 (starting from a copy of utrie.c) |
16 | * created by: Markus W. Scherer |
17 | * |
18 | * This is a common implementation of a Unicode trie. |
19 | * It is a kind of compressed, serializable table of 16- or 32-bit values associated with |
20 | * Unicode code points (0..0x10ffff). |
21 | * This is the second common version of a Unicode trie (hence the name UTrie2). |
22 | * See utrie2.h for a comparison. |
23 | * |
24 | * This file contains only the runtime and enumeration code, for read-only access. |
25 | * See utrie2_builder.c for the builder code. |
26 | */ |
27 | #include "unicode/utypes.h" |
28 | #ifdef UCPTRIE_DEBUG |
29 | #include "unicode/umutablecptrie.h" |
30 | #endif |
31 | #include "unicode/utf.h" |
32 | #include "unicode/utf8.h" |
33 | #include "unicode/utf16.h" |
34 | #include "cmemory.h" |
35 | #include "utrie2.h" |
36 | #include "utrie2_impl.h" |
37 | #include "uassert.h" |
38 | |
39 | /* Public UTrie2 API implementation ----------------------------------------- */ |
40 | |
41 | static uint32_t |
42 | get32(const UNewTrie2 *trie, UChar32 c, UBool fromLSCP) { |
43 | int32_t i2, block; |
44 | |
45 | if(c>=trie->highStart && (!U_IS_LEAD(c) || fromLSCP)) { |
46 | return trie->data[trie->dataLength-UTRIE2_DATA_GRANULARITY]; |
47 | } |
48 | |
49 | if(U_IS_LEAD(c) && fromLSCP) { |
50 | i2=(UTRIE2_LSCP_INDEX_2_OFFSET-(0xd800>>UTRIE2_SHIFT_2))+ |
51 | (c>>UTRIE2_SHIFT_2); |
52 | } else { |
53 | i2=trie->index1[c>>UTRIE2_SHIFT_1]+ |
54 | ((c>>UTRIE2_SHIFT_2)&UTRIE2_INDEX_2_MASK); |
55 | } |
56 | block=trie->index2[i2]; |
57 | return trie->data[block+(c&UTRIE2_DATA_MASK)]; |
58 | } |
59 | |
60 | U_CAPI uint32_t U_EXPORT2 |
61 | utrie2_get32(const UTrie2 *trie, UChar32 c) { |
62 | if(trie->data16!=nullptr) { |
63 | return UTRIE2_GET16(trie, c); |
64 | } else if(trie->data32!=nullptr) { |
65 | return UTRIE2_GET32(trie, c); |
66 | } else if((uint32_t)c>0x10ffff) { |
67 | return trie->errorValue; |
68 | } else { |
69 | return get32(trie->newTrie, c, true); |
70 | } |
71 | } |
72 | |
73 | U_CAPI uint32_t U_EXPORT2 |
74 | utrie2_get32FromLeadSurrogateCodeUnit(const UTrie2 *trie, UChar32 c) { |
75 | if(!U_IS_LEAD(c)) { |
76 | return trie->errorValue; |
77 | } |
78 | if(trie->data16!=nullptr) { |
79 | return UTRIE2_GET16_FROM_U16_SINGLE_LEAD(trie, c); |
80 | } else if(trie->data32!=nullptr) { |
81 | return UTRIE2_GET32_FROM_U16_SINGLE_LEAD(trie, c); |
82 | } else { |
83 | return get32(trie->newTrie, c, false); |
84 | } |
85 | } |
86 | |
87 | static inline int32_t |
88 | u8Index(const UTrie2 *trie, UChar32 c, int32_t i) { |
89 | int32_t idx= |
90 | _UTRIE2_INDEX_FROM_CP( |
91 | trie, |
92 | trie->data32==nullptr ? trie->indexLength : 0, |
93 | c); |
94 | return (idx<<3)|i; |
95 | } |
96 | |
97 | U_CAPI int32_t U_EXPORT2 |
98 | utrie2_internalU8NextIndex(const UTrie2 *trie, UChar32 c, |
99 | const uint8_t *src, const uint8_t *limit) { |
100 | int32_t i, length; |
101 | i=0; |
102 | /* support 64-bit pointers by avoiding cast of arbitrary difference */ |
103 | if((limit-src)<=7) { |
104 | length=(int32_t)(limit-src); |
105 | } else { |
106 | length=7; |
107 | } |
108 | c=utf8_nextCharSafeBody(src, &i, length, c, -1); |
109 | return u8Index(trie, c, i); |
110 | } |
111 | |
112 | U_CAPI int32_t U_EXPORT2 |
113 | utrie2_internalU8PrevIndex(const UTrie2 *trie, UChar32 c, |
114 | const uint8_t *start, const uint8_t *src) { |
115 | int32_t i, length; |
116 | /* support 64-bit pointers by avoiding cast of arbitrary difference */ |
117 | if((src-start)<=7) { |
118 | i=length=(int32_t)(src-start); |
119 | } else { |
120 | i=length=7; |
121 | start=src-7; |
122 | } |
123 | c=utf8_prevCharSafeBody(start, 0, &i, c, -1); |
124 | i=length-i; /* number of bytes read backward from src */ |
125 | return u8Index(trie, c, i); |
126 | } |
127 | |
128 | U_CAPI UTrie2 * U_EXPORT2 |
129 | utrie2_openFromSerialized(UTrie2ValueBits valueBits, |
130 | const void *data, int32_t length, int32_t *pActualLength, |
131 | UErrorCode *pErrorCode) { |
132 | const UTrie2Header *; |
133 | const uint16_t *p16; |
134 | int32_t actualLength; |
135 | |
136 | UTrie2 tempTrie; |
137 | UTrie2 *trie; |
138 | |
139 | if(U_FAILURE(*pErrorCode)) { |
140 | return 0; |
141 | } |
142 | |
143 | if( length<=0 || (U_POINTER_MASK_LSB(data, 3)!=0) || |
144 | valueBits<0 || UTRIE2_COUNT_VALUE_BITS<=valueBits |
145 | ) { |
146 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
147 | return 0; |
148 | } |
149 | |
150 | /* enough data for a trie header? */ |
151 | if(length<(int32_t)sizeof(UTrie2Header)) { |
152 | *pErrorCode=U_INVALID_FORMAT_ERROR; |
153 | return 0; |
154 | } |
155 | |
156 | /* check the signature */ |
157 | header=(const UTrie2Header *)data; |
158 | if(header->signature!=UTRIE2_SIG) { |
159 | *pErrorCode=U_INVALID_FORMAT_ERROR; |
160 | return 0; |
161 | } |
162 | |
163 | /* get the options */ |
164 | if(valueBits!=(UTrie2ValueBits)(header->options&UTRIE2_OPTIONS_VALUE_BITS_MASK)) { |
165 | *pErrorCode=U_INVALID_FORMAT_ERROR; |
166 | return 0; |
167 | } |
168 | |
169 | /* get the length values and offsets */ |
170 | uprv_memset(&tempTrie, 0, sizeof(tempTrie)); |
171 | tempTrie.indexLength=header->indexLength; |
172 | tempTrie.dataLength=header->shiftedDataLength<<UTRIE2_INDEX_SHIFT; |
173 | tempTrie.index2NullOffset=header->index2NullOffset; |
174 | tempTrie.dataNullOffset=header->dataNullOffset; |
175 | |
176 | tempTrie.highStart=header->shiftedHighStart<<UTRIE2_SHIFT_1; |
177 | tempTrie.highValueIndex=tempTrie.dataLength-UTRIE2_DATA_GRANULARITY; |
178 | if(valueBits==UTRIE2_16_VALUE_BITS) { |
179 | tempTrie.highValueIndex+=tempTrie.indexLength; |
180 | } |
181 | |
182 | /* calculate the actual length */ |
183 | actualLength=(int32_t)sizeof(UTrie2Header)+tempTrie.indexLength*2; |
184 | if(valueBits==UTRIE2_16_VALUE_BITS) { |
185 | actualLength+=tempTrie.dataLength*2; |
186 | } else { |
187 | actualLength+=tempTrie.dataLength*4; |
188 | } |
189 | if(length<actualLength) { |
190 | *pErrorCode=U_INVALID_FORMAT_ERROR; /* not enough bytes */ |
191 | return 0; |
192 | } |
193 | |
194 | /* allocate the trie */ |
195 | trie=(UTrie2 *)uprv_malloc(sizeof(UTrie2)); |
196 | if(trie==nullptr) { |
197 | *pErrorCode=U_MEMORY_ALLOCATION_ERROR; |
198 | return 0; |
199 | } |
200 | uprv_memcpy(trie, &tempTrie, sizeof(tempTrie)); |
201 | trie->memory=(uint32_t *)data; |
202 | trie->length=actualLength; |
203 | trie->isMemoryOwned=false; |
204 | #ifdef UTRIE2_DEBUG |
205 | trie->name="fromSerialized" ; |
206 | #endif |
207 | |
208 | /* set the pointers to its index and data arrays */ |
209 | p16=(const uint16_t *)(header+1); |
210 | trie->index=p16; |
211 | p16+=trie->indexLength; |
212 | |
213 | /* get the data */ |
214 | switch(valueBits) { |
215 | case UTRIE2_16_VALUE_BITS: |
216 | trie->data16=p16; |
217 | trie->data32=nullptr; |
218 | trie->initialValue=trie->index[trie->dataNullOffset]; |
219 | trie->errorValue=trie->data16[UTRIE2_BAD_UTF8_DATA_OFFSET]; |
220 | break; |
221 | case UTRIE2_32_VALUE_BITS: |
222 | trie->data16=nullptr; |
223 | trie->data32=(const uint32_t *)p16; |
224 | trie->initialValue=trie->data32[trie->dataNullOffset]; |
225 | trie->errorValue=trie->data32[UTRIE2_BAD_UTF8_DATA_OFFSET]; |
226 | break; |
227 | default: |
228 | *pErrorCode=U_INVALID_FORMAT_ERROR; |
229 | return 0; |
230 | } |
231 | |
232 | if(pActualLength!=nullptr) { |
233 | *pActualLength=actualLength; |
234 | } |
235 | return trie; |
236 | } |
237 | |
238 | U_CAPI UTrie2 * U_EXPORT2 |
239 | utrie2_openDummy(UTrie2ValueBits valueBits, |
240 | uint32_t initialValue, uint32_t errorValue, |
241 | UErrorCode *pErrorCode) { |
242 | UTrie2 *trie; |
243 | UTrie2Header *; |
244 | uint32_t *p; |
245 | uint16_t *dest16; |
246 | int32_t indexLength, dataLength, length, i; |
247 | int32_t dataMove; /* >0 if the data is moved to the end of the index array */ |
248 | |
249 | if(U_FAILURE(*pErrorCode)) { |
250 | return 0; |
251 | } |
252 | |
253 | if(valueBits<0 || UTRIE2_COUNT_VALUE_BITS<=valueBits) { |
254 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
255 | return 0; |
256 | } |
257 | |
258 | /* calculate the total length of the dummy trie data */ |
259 | indexLength=UTRIE2_INDEX_1_OFFSET; |
260 | dataLength=UTRIE2_DATA_START_OFFSET+UTRIE2_DATA_GRANULARITY; |
261 | length=(int32_t)sizeof(UTrie2Header)+indexLength*2; |
262 | if(valueBits==UTRIE2_16_VALUE_BITS) { |
263 | length+=dataLength*2; |
264 | } else { |
265 | length+=dataLength*4; |
266 | } |
267 | |
268 | /* allocate the trie */ |
269 | trie=(UTrie2 *)uprv_malloc(sizeof(UTrie2)); |
270 | if(trie==nullptr) { |
271 | *pErrorCode=U_MEMORY_ALLOCATION_ERROR; |
272 | return 0; |
273 | } |
274 | uprv_memset(trie, 0, sizeof(UTrie2)); |
275 | trie->memory=uprv_malloc(length); |
276 | if(trie->memory==nullptr) { |
277 | uprv_free(trie); |
278 | *pErrorCode=U_MEMORY_ALLOCATION_ERROR; |
279 | return 0; |
280 | } |
281 | trie->length=length; |
282 | trie->isMemoryOwned=true; |
283 | |
284 | /* set the UTrie2 fields */ |
285 | if(valueBits==UTRIE2_16_VALUE_BITS) { |
286 | dataMove=indexLength; |
287 | } else { |
288 | dataMove=0; |
289 | } |
290 | |
291 | trie->indexLength=indexLength; |
292 | trie->dataLength=dataLength; |
293 | trie->index2NullOffset=UTRIE2_INDEX_2_OFFSET; |
294 | trie->dataNullOffset=(uint16_t)dataMove; |
295 | trie->initialValue=initialValue; |
296 | trie->errorValue=errorValue; |
297 | trie->highStart=0; |
298 | trie->highValueIndex=dataMove+UTRIE2_DATA_START_OFFSET; |
299 | #ifdef UTRIE2_DEBUG |
300 | trie->name="dummy" ; |
301 | #endif |
302 | |
303 | /* set the header fields */ |
304 | header=(UTrie2Header *)trie->memory; |
305 | |
306 | header->signature=UTRIE2_SIG; /* "Tri2" */ |
307 | header->options=(uint16_t)valueBits; |
308 | |
309 | header->indexLength=(uint16_t)indexLength; |
310 | header->shiftedDataLength=(uint16_t)(dataLength>>UTRIE2_INDEX_SHIFT); |
311 | header->index2NullOffset=(uint16_t)UTRIE2_INDEX_2_OFFSET; |
312 | header->dataNullOffset=(uint16_t)dataMove; |
313 | header->shiftedHighStart=0; |
314 | |
315 | /* fill the index and data arrays */ |
316 | dest16=(uint16_t *)(header+1); |
317 | trie->index=dest16; |
318 | |
319 | /* write the index-2 array values shifted right by UTRIE2_INDEX_SHIFT */ |
320 | for(i=0; i<UTRIE2_INDEX_2_BMP_LENGTH; ++i) { |
321 | *dest16++=(uint16_t)(dataMove>>UTRIE2_INDEX_SHIFT); /* null data block */ |
322 | } |
323 | |
324 | /* write UTF-8 2-byte index-2 values, not right-shifted */ |
325 | for(i=0; i<(0xc2-0xc0); ++i) { /* C0..C1 */ |
326 | *dest16++=(uint16_t)(dataMove+UTRIE2_BAD_UTF8_DATA_OFFSET); |
327 | } |
328 | for(; i<(0xe0-0xc0); ++i) { /* C2..DF */ |
329 | *dest16++=(uint16_t)dataMove; |
330 | } |
331 | |
332 | /* write the 16/32-bit data array */ |
333 | switch(valueBits) { |
334 | case UTRIE2_16_VALUE_BITS: |
335 | /* write 16-bit data values */ |
336 | trie->data16=dest16; |
337 | trie->data32=nullptr; |
338 | for(i=0; i<0x80; ++i) { |
339 | *dest16++=(uint16_t)initialValue; |
340 | } |
341 | for(; i<0xc0; ++i) { |
342 | *dest16++=(uint16_t)errorValue; |
343 | } |
344 | /* highValue and reserved values */ |
345 | for(i=0; i<UTRIE2_DATA_GRANULARITY; ++i) { |
346 | *dest16++=(uint16_t)initialValue; |
347 | } |
348 | break; |
349 | case UTRIE2_32_VALUE_BITS: |
350 | /* write 32-bit data values */ |
351 | p=(uint32_t *)dest16; |
352 | trie->data16=nullptr; |
353 | trie->data32=p; |
354 | for(i=0; i<0x80; ++i) { |
355 | *p++=initialValue; |
356 | } |
357 | for(; i<0xc0; ++i) { |
358 | *p++=errorValue; |
359 | } |
360 | /* highValue and reserved values */ |
361 | for(i=0; i<UTRIE2_DATA_GRANULARITY; ++i) { |
362 | *p++=initialValue; |
363 | } |
364 | break; |
365 | default: |
366 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
367 | return 0; |
368 | } |
369 | |
370 | return trie; |
371 | } |
372 | |
373 | U_CAPI void U_EXPORT2 |
374 | utrie2_close(UTrie2 *trie) { |
375 | if(trie!=nullptr) { |
376 | if(trie->isMemoryOwned) { |
377 | uprv_free(trie->memory); |
378 | } |
379 | if(trie->newTrie!=nullptr) { |
380 | uprv_free(trie->newTrie->data); |
381 | #ifdef UCPTRIE_DEBUG |
382 | umutablecptrie_close(trie->newTrie->t3); |
383 | #endif |
384 | uprv_free(trie->newTrie); |
385 | } |
386 | uprv_free(trie); |
387 | } |
388 | } |
389 | |
390 | U_CAPI UBool U_EXPORT2 |
391 | utrie2_isFrozen(const UTrie2 *trie) { |
392 | return (UBool)(trie->newTrie==nullptr); |
393 | } |
394 | |
395 | U_CAPI int32_t U_EXPORT2 |
396 | utrie2_serialize(const UTrie2 *trie, |
397 | void *data, int32_t capacity, |
398 | UErrorCode *pErrorCode) { |
399 | /* argument check */ |
400 | if(U_FAILURE(*pErrorCode)) { |
401 | return 0; |
402 | } |
403 | |
404 | if( trie==nullptr || trie->memory==nullptr || trie->newTrie!=nullptr || |
405 | capacity<0 || (capacity>0 && (data==nullptr || (U_POINTER_MASK_LSB(data, 3)!=0))) |
406 | ) { |
407 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
408 | return 0; |
409 | } |
410 | |
411 | if(capacity>=trie->length) { |
412 | uprv_memcpy(data, trie->memory, trie->length); |
413 | } else { |
414 | *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
415 | } |
416 | return trie->length; |
417 | } |
418 | |
419 | /* enumeration -------------------------------------------------------------- */ |
420 | |
421 | #define MIN_VALUE(a, b) ((a)<(b) ? (a) : (b)) |
422 | |
423 | /* default UTrie2EnumValue() returns the input value itself */ |
424 | static uint32_t U_CALLCONV |
425 | enumSameValue(const void * /*context*/, uint32_t value) { |
426 | return value; |
427 | } |
428 | |
429 | /** |
430 | * Enumerate all ranges of code points with the same relevant values. |
431 | * The values are transformed from the raw trie entries by the enumValue function. |
432 | * |
433 | * Currently requires start<limit and both start and limit must be multiples |
434 | * of UTRIE2_DATA_BLOCK_LENGTH. |
435 | * |
436 | * Optimizations: |
437 | * - Skip a whole block if we know that it is filled with a single value, |
438 | * and it is the same as we visited just before. |
439 | * - Handle the null block specially because we know a priori that it is filled |
440 | * with a single value. |
441 | */ |
442 | static void |
443 | enumEitherTrie(const UTrie2 *trie, |
444 | UChar32 start, UChar32 limit, |
445 | UTrie2EnumValue *enumValue, UTrie2EnumRange *enumRange, const void *context) { |
446 | const uint32_t *data32; |
447 | const uint16_t *idx; |
448 | |
449 | uint32_t value, prevValue, initialValue; |
450 | UChar32 c, prev, highStart; |
451 | int32_t j, i2Block, prevI2Block, index2NullOffset, block, prevBlock, nullBlock; |
452 | |
453 | if(enumRange==nullptr) { |
454 | return; |
455 | } |
456 | if(enumValue==nullptr) { |
457 | enumValue=enumSameValue; |
458 | } |
459 | |
460 | if(trie->newTrie==nullptr) { |
461 | /* frozen trie */ |
462 | idx=trie->index; |
463 | U_ASSERT(idx!=nullptr); /* the following code assumes trie->newTrie is not nullptr when idx is nullptr */ |
464 | data32=trie->data32; |
465 | |
466 | index2NullOffset=trie->index2NullOffset; |
467 | nullBlock=trie->dataNullOffset; |
468 | } else { |
469 | /* unfrozen, mutable trie */ |
470 | idx=nullptr; |
471 | data32=trie->newTrie->data; |
472 | U_ASSERT(data32!=nullptr); /* the following code assumes idx is not nullptr when data32 is nullptr */ |
473 | |
474 | index2NullOffset=trie->newTrie->index2NullOffset; |
475 | nullBlock=trie->newTrie->dataNullOffset; |
476 | } |
477 | |
478 | highStart=trie->highStart; |
479 | |
480 | /* get the enumeration value that corresponds to an initial-value trie data entry */ |
481 | initialValue=enumValue(context, trie->initialValue); |
482 | |
483 | /* set variables for previous range */ |
484 | prevI2Block=-1; |
485 | prevBlock=-1; |
486 | prev=start; |
487 | prevValue=0; |
488 | |
489 | /* enumerate index-2 blocks */ |
490 | for(c=start; c<limit && c<highStart;) { |
491 | /* Code point limit for iterating inside this i2Block. */ |
492 | UChar32 tempLimit=c+UTRIE2_CP_PER_INDEX_1_ENTRY; |
493 | if(limit<tempLimit) { |
494 | tempLimit=limit; |
495 | } |
496 | if(c<=0xffff) { |
497 | if(!U_IS_SURROGATE(c)) { |
498 | i2Block=c>>UTRIE2_SHIFT_2; |
499 | } else if(U_IS_SURROGATE_LEAD(c)) { |
500 | /* |
501 | * Enumerate values for lead surrogate code points, not code units: |
502 | * This special block has half the normal length. |
503 | */ |
504 | i2Block=UTRIE2_LSCP_INDEX_2_OFFSET; |
505 | tempLimit=MIN_VALUE(0xdc00, limit); |
506 | } else { |
507 | /* |
508 | * Switch back to the normal part of the index-2 table. |
509 | * Enumerate the second half of the surrogates block. |
510 | */ |
511 | i2Block=0xd800>>UTRIE2_SHIFT_2; |
512 | tempLimit=MIN_VALUE(0xe000, limit); |
513 | } |
514 | } else { |
515 | /* supplementary code points */ |
516 | if(idx!=nullptr) { |
517 | i2Block=idx[(UTRIE2_INDEX_1_OFFSET-UTRIE2_OMITTED_BMP_INDEX_1_LENGTH)+ |
518 | (c>>UTRIE2_SHIFT_1)]; |
519 | } else { |
520 | i2Block=trie->newTrie->index1[c>>UTRIE2_SHIFT_1]; |
521 | } |
522 | if(i2Block==prevI2Block && (c-prev)>=UTRIE2_CP_PER_INDEX_1_ENTRY) { |
523 | /* |
524 | * The index-2 block is the same as the previous one, and filled with prevValue. |
525 | * Only possible for supplementary code points because the linear-BMP index-2 |
526 | * table creates unique i2Block values. |
527 | */ |
528 | c+=UTRIE2_CP_PER_INDEX_1_ENTRY; |
529 | continue; |
530 | } |
531 | } |
532 | prevI2Block=i2Block; |
533 | if(i2Block==index2NullOffset) { |
534 | /* this is the null index-2 block */ |
535 | if(prevValue!=initialValue) { |
536 | if(prev<c && !enumRange(context, prev, c-1, prevValue)) { |
537 | return; |
538 | } |
539 | prevBlock=nullBlock; |
540 | prev=c; |
541 | prevValue=initialValue; |
542 | } |
543 | c+=UTRIE2_CP_PER_INDEX_1_ENTRY; |
544 | } else { |
545 | /* enumerate data blocks for one index-2 block */ |
546 | int32_t i2, i2Limit; |
547 | i2=(c>>UTRIE2_SHIFT_2)&UTRIE2_INDEX_2_MASK; |
548 | if((c>>UTRIE2_SHIFT_1)==(tempLimit>>UTRIE2_SHIFT_1)) { |
549 | i2Limit=(tempLimit>>UTRIE2_SHIFT_2)&UTRIE2_INDEX_2_MASK; |
550 | } else { |
551 | i2Limit=UTRIE2_INDEX_2_BLOCK_LENGTH; |
552 | } |
553 | for(; i2<i2Limit; ++i2) { |
554 | if(idx!=nullptr) { |
555 | block=(int32_t)idx[i2Block+i2]<<UTRIE2_INDEX_SHIFT; |
556 | } else { |
557 | block=trie->newTrie->index2[i2Block+i2]; |
558 | } |
559 | if(block==prevBlock && (c-prev)>=UTRIE2_DATA_BLOCK_LENGTH) { |
560 | /* the block is the same as the previous one, and filled with prevValue */ |
561 | c+=UTRIE2_DATA_BLOCK_LENGTH; |
562 | continue; |
563 | } |
564 | prevBlock=block; |
565 | if(block==nullBlock) { |
566 | /* this is the null data block */ |
567 | if(prevValue!=initialValue) { |
568 | if(prev<c && !enumRange(context, prev, c-1, prevValue)) { |
569 | return; |
570 | } |
571 | prev=c; |
572 | prevValue=initialValue; |
573 | } |
574 | c+=UTRIE2_DATA_BLOCK_LENGTH; |
575 | } else { |
576 | for(j=0; j<UTRIE2_DATA_BLOCK_LENGTH; ++j) { |
577 | value=enumValue(context, data32!=nullptr ? data32[block+j] : idx[block+j]); |
578 | if(value!=prevValue) { |
579 | if(prev<c && !enumRange(context, prev, c-1, prevValue)) { |
580 | return; |
581 | } |
582 | prev=c; |
583 | prevValue=value; |
584 | } |
585 | ++c; |
586 | } |
587 | } |
588 | } |
589 | } |
590 | } |
591 | |
592 | if(c>limit) { |
593 | c=limit; /* could be higher if in the index2NullOffset */ |
594 | } else if(c<limit) { |
595 | /* c==highStart<limit */ |
596 | uint32_t highValue; |
597 | if(idx!=nullptr) { |
598 | highValue= |
599 | data32!=nullptr ? |
600 | data32[trie->highValueIndex] : |
601 | idx[trie->highValueIndex]; |
602 | } else { |
603 | highValue=trie->newTrie->data[trie->newTrie->dataLength-UTRIE2_DATA_GRANULARITY]; |
604 | } |
605 | value=enumValue(context, highValue); |
606 | if(value!=prevValue) { |
607 | if(prev<c && !enumRange(context, prev, c-1, prevValue)) { |
608 | return; |
609 | } |
610 | prev=c; |
611 | prevValue=value; |
612 | } |
613 | c=limit; |
614 | } |
615 | |
616 | /* deliver last range */ |
617 | enumRange(context, prev, c-1, prevValue); |
618 | } |
619 | |
620 | U_CAPI void U_EXPORT2 |
621 | utrie2_enum(const UTrie2 *trie, |
622 | UTrie2EnumValue *enumValue, UTrie2EnumRange *enumRange, const void *context) { |
623 | enumEitherTrie(trie, 0, 0x110000, enumValue, enumRange, context); |
624 | } |
625 | |
626 | U_CAPI void U_EXPORT2 |
627 | utrie2_enumForLeadSurrogate(const UTrie2 *trie, UChar32 lead, |
628 | UTrie2EnumValue *enumValue, UTrie2EnumRange *enumRange, |
629 | const void *context) { |
630 | if(!U16_IS_LEAD(lead)) { |
631 | return; |
632 | } |
633 | lead=(lead-0xd7c0)<<10; /* start code point */ |
634 | enumEitherTrie(trie, lead, lead+0x400, enumValue, enumRange, context); |
635 | } |
636 | |
637 | /* C++ convenience wrappers ------------------------------------------------- */ |
638 | |
639 | U_NAMESPACE_BEGIN |
640 | |
641 | uint16_t BackwardUTrie2StringIterator::previous16() { |
642 | codePointLimit=codePointStart; |
643 | if(start>=codePointStart) { |
644 | codePoint=U_SENTINEL; |
645 | return static_cast<uint16_t>(trie->errorValue); |
646 | } |
647 | uint16_t result; |
648 | UTRIE2_U16_PREV16(trie, start, codePointStart, codePoint, result); |
649 | return result; |
650 | } |
651 | |
652 | uint16_t ForwardUTrie2StringIterator::next16() { |
653 | codePointStart=codePointLimit; |
654 | if(codePointLimit==limit) { |
655 | codePoint=U_SENTINEL; |
656 | return static_cast<uint16_t>(trie->errorValue); |
657 | } |
658 | uint16_t result; |
659 | UTRIE2_U16_NEXT16(trie, codePointLimit, limit, codePoint, result); |
660 | return result; |
661 | } |
662 | |
663 | U_NAMESPACE_END |
664 | |