]>
Commit | Line | Data |
---|---|---|
1 | /* | |
2 | ******************************************************************************* | |
3 | * Copyright (C) 2013-2014, International Business Machines | |
4 | * Corporation and others. All Rights Reserved. | |
5 | ******************************************************************************* | |
6 | * collationsets.cpp | |
7 | * | |
8 | * created on: 2013feb09 | |
9 | * created by: Markus W. Scherer | |
10 | */ | |
11 | ||
12 | #include "unicode/utypes.h" | |
13 | ||
14 | #if !UCONFIG_NO_COLLATION | |
15 | ||
16 | #include "unicode/ucharstrie.h" | |
17 | #include "unicode/uniset.h" | |
18 | #include "unicode/unistr.h" | |
19 | #include "unicode/ustringtrie.h" | |
20 | #include "collation.h" | |
21 | #include "collationdata.h" | |
22 | #include "collationsets.h" | |
23 | #include "normalizer2impl.h" | |
24 | #include "uassert.h" | |
25 | #include "utf16collationiterator.h" | |
26 | #include "utrie2.h" | |
27 | ||
28 | U_NAMESPACE_BEGIN | |
29 | ||
30 | U_CDECL_BEGIN | |
31 | ||
32 | static UBool U_CALLCONV | |
33 | enumTailoredRange(const void *context, UChar32 start, UChar32 end, uint32_t ce32) { | |
34 | if(ce32 == Collation::FALLBACK_CE32) { | |
35 | return TRUE; // fallback to base, not tailored | |
36 | } | |
37 | TailoredSet *ts = (TailoredSet *)context; | |
38 | return ts->handleCE32(start, end, ce32); | |
39 | } | |
40 | ||
41 | U_CDECL_END | |
42 | ||
43 | void | |
44 | TailoredSet::forData(const CollationData *d, UErrorCode &ec) { | |
45 | if(U_FAILURE(ec)) { return; } | |
46 | errorCode = ec; // Preserve info & warning codes. | |
47 | data = d; | |
48 | baseData = d->base; | |
49 | U_ASSERT(baseData != NULL); | |
50 | utrie2_enum(data->trie, NULL, enumTailoredRange, this); | |
51 | ec = errorCode; | |
52 | } | |
53 | ||
54 | UBool | |
55 | TailoredSet::handleCE32(UChar32 start, UChar32 end, uint32_t ce32) { | |
56 | U_ASSERT(ce32 != Collation::FALLBACK_CE32); | |
57 | if(Collation::isSpecialCE32(ce32)) { | |
58 | ce32 = data->getIndirectCE32(ce32); | |
59 | if(ce32 == Collation::FALLBACK_CE32) { | |
60 | return U_SUCCESS(errorCode); | |
61 | } | |
62 | } | |
63 | do { | |
64 | uint32_t baseCE32 = baseData->getFinalCE32(baseData->getCE32(start)); | |
65 | // Do not just continue if ce32 == baseCE32 because | |
66 | // contractions and expansions in different data objects | |
67 | // normally differ even if they have the same data offsets. | |
68 | if(Collation::isSelfContainedCE32(ce32) && Collation::isSelfContainedCE32(baseCE32)) { | |
69 | // fastpath | |
70 | if(ce32 != baseCE32) { | |
71 | tailored->add(start); | |
72 | } | |
73 | } else { | |
74 | compare(start, ce32, baseCE32); | |
75 | } | |
76 | } while(++start <= end); | |
77 | return U_SUCCESS(errorCode); | |
78 | } | |
79 | ||
80 | void | |
81 | TailoredSet::compare(UChar32 c, uint32_t ce32, uint32_t baseCE32) { | |
82 | if(Collation::isPrefixCE32(ce32)) { | |
83 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); | |
84 | ce32 = data->getFinalCE32(CollationData::readCE32(p)); | |
85 | if(Collation::isPrefixCE32(baseCE32)) { | |
86 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); | |
87 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); | |
88 | comparePrefixes(c, p + 2, q + 2); | |
89 | } else { | |
90 | addPrefixes(data, c, p + 2); | |
91 | } | |
92 | } else if(Collation::isPrefixCE32(baseCE32)) { | |
93 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); | |
94 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); | |
95 | addPrefixes(baseData, c, q + 2); | |
96 | } | |
97 | ||
98 | if(Collation::isContractionCE32(ce32)) { | |
99 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); | |
100 | if((ce32 & Collation::CONTRACT_SINGLE_CP_NO_MATCH) != 0) { | |
101 | ce32 = Collation::NO_CE32; | |
102 | } else { | |
103 | ce32 = data->getFinalCE32(CollationData::readCE32(p)); | |
104 | } | |
105 | if(Collation::isContractionCE32(baseCE32)) { | |
106 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); | |
107 | if((baseCE32 & Collation::CONTRACT_SINGLE_CP_NO_MATCH) != 0) { | |
108 | baseCE32 = Collation::NO_CE32; | |
109 | } else { | |
110 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); | |
111 | } | |
112 | compareContractions(c, p + 2, q + 2); | |
113 | } else { | |
114 | addContractions(c, p + 2); | |
115 | } | |
116 | } else if(Collation::isContractionCE32(baseCE32)) { | |
117 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); | |
118 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); | |
119 | addContractions(c, q + 2); | |
120 | } | |
121 | ||
122 | int32_t tag; | |
123 | if(Collation::isSpecialCE32(ce32)) { | |
124 | tag = Collation::tagFromCE32(ce32); | |
125 | U_ASSERT(tag != Collation::PREFIX_TAG); | |
126 | U_ASSERT(tag != Collation::CONTRACTION_TAG); | |
127 | // Currently, the tailoring data builder does not write offset tags. | |
128 | // They might be useful for saving space, | |
129 | // but they would complicate the builder, | |
130 | // and in tailorings we assume that performance of tailored characters is more important. | |
131 | U_ASSERT(tag != Collation::OFFSET_TAG); | |
132 | } else { | |
133 | tag = -1; | |
134 | } | |
135 | int32_t baseTag; | |
136 | if(Collation::isSpecialCE32(baseCE32)) { | |
137 | baseTag = Collation::tagFromCE32(baseCE32); | |
138 | U_ASSERT(baseTag != Collation::PREFIX_TAG); | |
139 | U_ASSERT(baseTag != Collation::CONTRACTION_TAG); | |
140 | } else { | |
141 | baseTag = -1; | |
142 | } | |
143 | ||
144 | // Non-contextual mappings, expansions, etc. | |
145 | if(baseTag == Collation::OFFSET_TAG) { | |
146 | // We might be comparing a tailoring CE which is a copy of | |
147 | // a base offset-tag CE, via the [optimize [set]] syntax | |
148 | // or when a single-character mapping was copied for tailored contractions. | |
149 | // Offset tags always result in long-primary CEs, | |
150 | // with common secondary/tertiary weights. | |
151 | if(!Collation::isLongPrimaryCE32(ce32)) { | |
152 | add(c); | |
153 | return; | |
154 | } | |
155 | int64_t dataCE = baseData->ces[Collation::indexFromCE32(baseCE32)]; | |
156 | uint32_t p = Collation::getThreeBytePrimaryForOffsetData(c, dataCE); | |
157 | if(Collation::primaryFromLongPrimaryCE32(ce32) != p) { | |
158 | add(c); | |
159 | return; | |
160 | } | |
161 | } | |
162 | ||
163 | if(tag != baseTag) { | |
164 | add(c); | |
165 | return; | |
166 | } | |
167 | ||
168 | if(tag == Collation::EXPANSION32_TAG) { | |
169 | const uint32_t *ce32s = data->ce32s + Collation::indexFromCE32(ce32); | |
170 | int32_t length = Collation::lengthFromCE32(ce32); | |
171 | ||
172 | const uint32_t *baseCE32s = baseData->ce32s + Collation::indexFromCE32(baseCE32); | |
173 | int32_t baseLength = Collation::lengthFromCE32(baseCE32); | |
174 | ||
175 | if(length != baseLength) { | |
176 | add(c); | |
177 | return; | |
178 | } | |
179 | for(int32_t i = 0; i < length; ++i) { | |
180 | if(ce32s[i] != baseCE32s[i]) { | |
181 | add(c); | |
182 | break; | |
183 | } | |
184 | } | |
185 | } else if(tag == Collation::EXPANSION_TAG) { | |
186 | const int64_t *ces = data->ces + Collation::indexFromCE32(ce32); | |
187 | int32_t length = Collation::lengthFromCE32(ce32); | |
188 | ||
189 | const int64_t *baseCEs = baseData->ces + Collation::indexFromCE32(baseCE32); | |
190 | int32_t baseLength = Collation::lengthFromCE32(baseCE32); | |
191 | ||
192 | if(length != baseLength) { | |
193 | add(c); | |
194 | return; | |
195 | } | |
196 | for(int32_t i = 0; i < length; ++i) { | |
197 | if(ces[i] != baseCEs[i]) { | |
198 | add(c); | |
199 | break; | |
200 | } | |
201 | } | |
202 | } else if(tag == Collation::HANGUL_TAG) { | |
203 | UChar jamos[3]; | |
204 | int32_t length = Hangul::decompose(c, jamos); | |
205 | if(tailored->contains(jamos[0]) || tailored->contains(jamos[1]) || | |
206 | (length == 3 && tailored->contains(jamos[2]))) { | |
207 | add(c); | |
208 | } | |
209 | } else if(ce32 != baseCE32) { | |
210 | add(c); | |
211 | } | |
212 | } | |
213 | ||
214 | void | |
215 | TailoredSet::comparePrefixes(UChar32 c, const UChar *p, const UChar *q) { | |
216 | // Parallel iteration over prefixes of both tables. | |
217 | UCharsTrie::Iterator prefixes(p, 0, errorCode); | |
218 | UCharsTrie::Iterator basePrefixes(q, 0, errorCode); | |
219 | const UnicodeString *tp = NULL; // Tailoring prefix. | |
220 | const UnicodeString *bp = NULL; // Base prefix. | |
221 | // Use a string with a U+FFFF as the limit sentinel. | |
222 | // U+FFFF is untailorable and will not occur in prefixes. | |
223 | UnicodeString none((UChar)0xffff); | |
224 | for(;;) { | |
225 | if(tp == NULL) { | |
226 | if(prefixes.next(errorCode)) { | |
227 | tp = &prefixes.getString(); | |
228 | } else { | |
229 | tp = &none; | |
230 | } | |
231 | } | |
232 | if(bp == NULL) { | |
233 | if(basePrefixes.next(errorCode)) { | |
234 | bp = &basePrefixes.getString(); | |
235 | } else { | |
236 | bp = &none; | |
237 | } | |
238 | } | |
239 | if(tp == &none && bp == &none) { break; } | |
240 | int32_t cmp = tp->compare(*bp); | |
241 | if(cmp < 0) { | |
242 | // tp occurs in the tailoring but not in the base. | |
243 | addPrefix(data, *tp, c, (uint32_t)prefixes.getValue()); | |
244 | tp = NULL; | |
245 | } else if(cmp > 0) { | |
246 | // bp occurs in the base but not in the tailoring. | |
247 | addPrefix(baseData, *bp, c, (uint32_t)basePrefixes.getValue()); | |
248 | bp = NULL; | |
249 | } else { | |
250 | setPrefix(*tp); | |
251 | compare(c, (uint32_t)prefixes.getValue(), (uint32_t)basePrefixes.getValue()); | |
252 | resetPrefix(); | |
253 | tp = NULL; | |
254 | bp = NULL; | |
255 | } | |
256 | } | |
257 | } | |
258 | ||
259 | void | |
260 | TailoredSet::compareContractions(UChar32 c, const UChar *p, const UChar *q) { | |
261 | // Parallel iteration over suffixes of both tables. | |
262 | UCharsTrie::Iterator suffixes(p, 0, errorCode); | |
263 | UCharsTrie::Iterator baseSuffixes(q, 0, errorCode); | |
264 | const UnicodeString *ts = NULL; // Tailoring suffix. | |
265 | const UnicodeString *bs = NULL; // Base suffix. | |
266 | // Use a string with two U+FFFF as the limit sentinel. | |
267 | // U+FFFF is untailorable and will not occur in contractions except maybe | |
268 | // as a single suffix character for a root-collator boundary contraction. | |
269 | UnicodeString none((UChar)0xffff); | |
270 | none.append((UChar)0xffff); | |
271 | for(;;) { | |
272 | if(ts == NULL) { | |
273 | if(suffixes.next(errorCode)) { | |
274 | ts = &suffixes.getString(); | |
275 | } else { | |
276 | ts = &none; | |
277 | } | |
278 | } | |
279 | if(bs == NULL) { | |
280 | if(baseSuffixes.next(errorCode)) { | |
281 | bs = &baseSuffixes.getString(); | |
282 | } else { | |
283 | bs = &none; | |
284 | } | |
285 | } | |
286 | if(ts == &none && bs == &none) { break; } | |
287 | int32_t cmp = ts->compare(*bs); | |
288 | if(cmp < 0) { | |
289 | // ts occurs in the tailoring but not in the base. | |
290 | addSuffix(c, *ts); | |
291 | ts = NULL; | |
292 | } else if(cmp > 0) { | |
293 | // bs occurs in the base but not in the tailoring. | |
294 | addSuffix(c, *bs); | |
295 | bs = NULL; | |
296 | } else { | |
297 | suffix = ts; | |
298 | compare(c, (uint32_t)suffixes.getValue(), (uint32_t)baseSuffixes.getValue()); | |
299 | suffix = NULL; | |
300 | ts = NULL; | |
301 | bs = NULL; | |
302 | } | |
303 | } | |
304 | } | |
305 | ||
306 | void | |
307 | TailoredSet::addPrefixes(const CollationData *d, UChar32 c, const UChar *p) { | |
308 | UCharsTrie::Iterator prefixes(p, 0, errorCode); | |
309 | while(prefixes.next(errorCode)) { | |
310 | addPrefix(d, prefixes.getString(), c, (uint32_t)prefixes.getValue()); | |
311 | } | |
312 | } | |
313 | ||
314 | void | |
315 | TailoredSet::addPrefix(const CollationData *d, const UnicodeString &pfx, UChar32 c, uint32_t ce32) { | |
316 | setPrefix(pfx); | |
317 | ce32 = d->getFinalCE32(ce32); | |
318 | if(Collation::isContractionCE32(ce32)) { | |
319 | const UChar *p = d->contexts + Collation::indexFromCE32(ce32); | |
320 | addContractions(c, p + 2); | |
321 | } | |
322 | tailored->add(UnicodeString(unreversedPrefix).append(c)); | |
323 | resetPrefix(); | |
324 | } | |
325 | ||
326 | void | |
327 | TailoredSet::addContractions(UChar32 c, const UChar *p) { | |
328 | UCharsTrie::Iterator suffixes(p, 0, errorCode); | |
329 | while(suffixes.next(errorCode)) { | |
330 | addSuffix(c, suffixes.getString()); | |
331 | } | |
332 | } | |
333 | ||
334 | void | |
335 | TailoredSet::addSuffix(UChar32 c, const UnicodeString &sfx) { | |
336 | tailored->add(UnicodeString(unreversedPrefix).append(c).append(sfx)); | |
337 | } | |
338 | ||
339 | void | |
340 | TailoredSet::add(UChar32 c) { | |
341 | if(unreversedPrefix.isEmpty() && suffix == NULL) { | |
342 | tailored->add(c); | |
343 | } else { | |
344 | UnicodeString s(unreversedPrefix); | |
345 | s.append(c); | |
346 | if(suffix != NULL) { | |
347 | s.append(*suffix); | |
348 | } | |
349 | tailored->add(s); | |
350 | } | |
351 | } | |
352 | ||
353 | ContractionsAndExpansions::CESink::~CESink() {} | |
354 | ||
355 | U_CDECL_BEGIN | |
356 | ||
357 | static UBool U_CALLCONV | |
358 | enumCnERange(const void *context, UChar32 start, UChar32 end, uint32_t ce32) { | |
359 | ContractionsAndExpansions *cne = (ContractionsAndExpansions *)context; | |
360 | if(cne->checkTailored == 0) { | |
361 | // There is no tailoring. | |
362 | // No need to collect nor check the tailored set. | |
363 | } else if(cne->checkTailored < 0) { | |
364 | // Collect the set of code points with mappings in the tailoring data. | |
365 | if(ce32 == Collation::FALLBACK_CE32) { | |
366 | return TRUE; // fallback to base, not tailored | |
367 | } else { | |
368 | cne->tailored.add(start, end); | |
369 | } | |
370 | // checkTailored > 0: Exclude tailored ranges from the base data enumeration. | |
371 | } else if(start == end) { | |
372 | if(cne->tailored.contains(start)) { | |
373 | return TRUE; | |
374 | } | |
375 | } else if(cne->tailored.containsSome(start, end)) { | |
376 | cne->ranges.set(start, end).removeAll(cne->tailored); | |
377 | int32_t count = cne->ranges.getRangeCount(); | |
378 | for(int32_t i = 0; i < count; ++i) { | |
379 | cne->handleCE32(cne->ranges.getRangeStart(i), cne->ranges.getRangeEnd(i), ce32); | |
380 | } | |
381 | return U_SUCCESS(cne->errorCode); | |
382 | } | |
383 | cne->handleCE32(start, end, ce32); | |
384 | return U_SUCCESS(cne->errorCode); | |
385 | } | |
386 | ||
387 | U_CDECL_END | |
388 | ||
389 | void | |
390 | ContractionsAndExpansions::forData(const CollationData *d, UErrorCode &ec) { | |
391 | if(U_FAILURE(ec)) { return; } | |
392 | errorCode = ec; // Preserve info & warning codes. | |
393 | // Add all from the data, can be tailoring or base. | |
394 | if(d->base != NULL) { | |
395 | checkTailored = -1; | |
396 | } | |
397 | data = d; | |
398 | utrie2_enum(data->trie, NULL, enumCnERange, this); | |
399 | if(d->base == NULL || U_FAILURE(errorCode)) { | |
400 | ec = errorCode; | |
401 | return; | |
402 | } | |
403 | // Add all from the base data but only for un-tailored code points. | |
404 | tailored.freeze(); | |
405 | checkTailored = 1; | |
406 | data = d->base; | |
407 | utrie2_enum(data->trie, NULL, enumCnERange, this); | |
408 | ec = errorCode; | |
409 | } | |
410 | ||
411 | void | |
412 | ContractionsAndExpansions::forCodePoint(const CollationData *d, UChar32 c, UErrorCode &ec) { | |
413 | if(U_FAILURE(ec)) { return; } | |
414 | errorCode = ec; // Preserve info & warning codes. | |
415 | uint32_t ce32 = d->getCE32(c); | |
416 | if(ce32 == Collation::FALLBACK_CE32) { | |
417 | d = d->base; | |
418 | ce32 = d->getCE32(c); | |
419 | } | |
420 | data = d; | |
421 | handleCE32(c, c, ce32); | |
422 | ec = errorCode; | |
423 | } | |
424 | ||
425 | void | |
426 | ContractionsAndExpansions::handleCE32(UChar32 start, UChar32 end, uint32_t ce32) { | |
427 | for(;;) { | |
428 | if((ce32 & 0xff) < Collation::SPECIAL_CE32_LOW_BYTE) { | |
429 | // !isSpecialCE32() | |
430 | if(sink != NULL) { | |
431 | sink->handleCE(Collation::ceFromSimpleCE32(ce32)); | |
432 | } | |
433 | return; | |
434 | } | |
435 | switch(Collation::tagFromCE32(ce32)) { | |
436 | case Collation::FALLBACK_TAG: | |
437 | return; | |
438 | case Collation::RESERVED_TAG_3: | |
439 | case Collation::BUILDER_DATA_TAG: | |
440 | case Collation::LEAD_SURROGATE_TAG: | |
441 | if(U_SUCCESS(errorCode)) { errorCode = U_INTERNAL_PROGRAM_ERROR; } | |
442 | return; | |
443 | case Collation::LONG_PRIMARY_TAG: | |
444 | if(sink != NULL) { | |
445 | sink->handleCE(Collation::ceFromLongPrimaryCE32(ce32)); | |
446 | } | |
447 | return; | |
448 | case Collation::LONG_SECONDARY_TAG: | |
449 | if(sink != NULL) { | |
450 | sink->handleCE(Collation::ceFromLongSecondaryCE32(ce32)); | |
451 | } | |
452 | return; | |
453 | case Collation::LATIN_EXPANSION_TAG: | |
454 | if(sink != NULL) { | |
455 | ces[0] = Collation::latinCE0FromCE32(ce32); | |
456 | ces[1] = Collation::latinCE1FromCE32(ce32); | |
457 | sink->handleExpansion(ces, 2); | |
458 | } | |
459 | // Optimization: If we have a prefix, | |
460 | // then the relevant strings have been added already. | |
461 | if(unreversedPrefix.isEmpty()) { | |
462 | addExpansions(start, end); | |
463 | } | |
464 | return; | |
465 | case Collation::EXPANSION32_TAG: | |
466 | if(sink != NULL) { | |
467 | const uint32_t *ce32s = data->ce32s + Collation::indexFromCE32(ce32); | |
468 | int32_t length = Collation::lengthFromCE32(ce32); | |
469 | for(int32_t i = 0; i < length; ++i) { | |
470 | ces[i] = Collation::ceFromCE32(*ce32s++); | |
471 | } | |
472 | sink->handleExpansion(ces, length); | |
473 | } | |
474 | // Optimization: If we have a prefix, | |
475 | // then the relevant strings have been added already. | |
476 | if(unreversedPrefix.isEmpty()) { | |
477 | addExpansions(start, end); | |
478 | } | |
479 | return; | |
480 | case Collation::EXPANSION_TAG: | |
481 | if(sink != NULL) { | |
482 | int32_t length = Collation::lengthFromCE32(ce32); | |
483 | sink->handleExpansion(data->ces + Collation::indexFromCE32(ce32), length); | |
484 | } | |
485 | // Optimization: If we have a prefix, | |
486 | // then the relevant strings have been added already. | |
487 | if(unreversedPrefix.isEmpty()) { | |
488 | addExpansions(start, end); | |
489 | } | |
490 | return; | |
491 | case Collation::PREFIX_TAG: | |
492 | handlePrefixes(start, end, ce32); | |
493 | return; | |
494 | case Collation::CONTRACTION_TAG: | |
495 | handleContractions(start, end, ce32); | |
496 | return; | |
497 | case Collation::DIGIT_TAG: | |
498 | // Fetch the non-numeric-collation CE32 and continue. | |
499 | ce32 = data->ce32s[Collation::indexFromCE32(ce32)]; | |
500 | break; | |
501 | case Collation::U0000_TAG: | |
502 | U_ASSERT(start == 0 && end == 0); | |
503 | // Fetch the normal ce32 for U+0000 and continue. | |
504 | ce32 = data->ce32s[0]; | |
505 | break; | |
506 | case Collation::HANGUL_TAG: | |
507 | if(sink != NULL) { | |
508 | // TODO: This should be optimized, | |
509 | // especially if [start..end] is the complete Hangul range. (assert that) | |
510 | UTF16CollationIterator iter(data, FALSE, NULL, NULL, NULL); | |
511 | UChar hangul[1] = { 0 }; | |
512 | for(UChar32 c = start; c <= end; ++c) { | |
513 | hangul[0] = (UChar)c; | |
514 | iter.setText(hangul, hangul + 1); | |
515 | int32_t length = iter.fetchCEs(errorCode); | |
516 | if(U_FAILURE(errorCode)) { return; } | |
517 | // Ignore the terminating non-CE. | |
518 | U_ASSERT(length >= 2 && iter.getCE(length - 1) == Collation::NO_CE); | |
519 | sink->handleExpansion(iter.getCEs(), length - 1); | |
520 | } | |
521 | } | |
522 | // Optimization: If we have a prefix, | |
523 | // then the relevant strings have been added already. | |
524 | if(unreversedPrefix.isEmpty()) { | |
525 | addExpansions(start, end); | |
526 | } | |
527 | return; | |
528 | case Collation::OFFSET_TAG: | |
529 | // Currently no need to send offset CEs to the sink. | |
530 | return; | |
531 | case Collation::IMPLICIT_TAG: | |
532 | // Currently no need to send implicit CEs to the sink. | |
533 | return; | |
534 | } | |
535 | } | |
536 | } | |
537 | ||
538 | void | |
539 | ContractionsAndExpansions::handlePrefixes( | |
540 | UChar32 start, UChar32 end, uint32_t ce32) { | |
541 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); | |
542 | ce32 = CollationData::readCE32(p); // Default if no prefix match. | |
543 | handleCE32(start, end, ce32); | |
544 | if(!addPrefixes) { return; } | |
545 | UCharsTrie::Iterator prefixes(p + 2, 0, errorCode); | |
546 | while(prefixes.next(errorCode)) { | |
547 | setPrefix(prefixes.getString()); | |
548 | // Prefix/pre-context mappings are special kinds of contractions | |
549 | // that always yield expansions. | |
550 | addStrings(start, end, contractions); | |
551 | addStrings(start, end, expansions); | |
552 | handleCE32(start, end, (uint32_t)prefixes.getValue()); | |
553 | } | |
554 | resetPrefix(); | |
555 | } | |
556 | ||
557 | void | |
558 | ContractionsAndExpansions::handleContractions( | |
559 | UChar32 start, UChar32 end, uint32_t ce32) { | |
560 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); | |
561 | if((ce32 & Collation::CONTRACT_SINGLE_CP_NO_MATCH) != 0) { | |
562 | // No match on the single code point. | |
563 | // We are underneath a prefix, and the default mapping is just | |
564 | // a fallback to the mappings for a shorter prefix. | |
565 | U_ASSERT(!unreversedPrefix.isEmpty()); | |
566 | } else { | |
567 | ce32 = CollationData::readCE32(p); // Default if no suffix match. | |
568 | U_ASSERT(!Collation::isContractionCE32(ce32)); | |
569 | handleCE32(start, end, ce32); | |
570 | } | |
571 | UCharsTrie::Iterator suffixes(p + 2, 0, errorCode); | |
572 | while(suffixes.next(errorCode)) { | |
573 | suffix = &suffixes.getString(); | |
574 | addStrings(start, end, contractions); | |
575 | if(!unreversedPrefix.isEmpty()) { | |
576 | addStrings(start, end, expansions); | |
577 | } | |
578 | handleCE32(start, end, (uint32_t)suffixes.getValue()); | |
579 | } | |
580 | suffix = NULL; | |
581 | } | |
582 | ||
583 | void | |
584 | ContractionsAndExpansions::addExpansions(UChar32 start, UChar32 end) { | |
585 | if(unreversedPrefix.isEmpty() && suffix == NULL) { | |
586 | if(expansions != NULL) { | |
587 | expansions->add(start, end); | |
588 | } | |
589 | } else { | |
590 | addStrings(start, end, expansions); | |
591 | } | |
592 | } | |
593 | ||
594 | void | |
595 | ContractionsAndExpansions::addStrings(UChar32 start, UChar32 end, UnicodeSet *set) { | |
596 | if(set == NULL) { return; } | |
597 | UnicodeString s(unreversedPrefix); | |
598 | do { | |
599 | s.append(start); | |
600 | if(suffix != NULL) { | |
601 | s.append(*suffix); | |
602 | } | |
603 | set->add(s); | |
604 | s.truncate(unreversedPrefix.length()); | |
605 | } while(++start <= end); | |
606 | } | |
607 | ||
608 | U_NAMESPACE_END | |
609 | ||
610 | #endif // !UCONFIG_NO_COLLATION |