File: | out/../deps/icu-small/source/i18n/transreg.cpp |
Warning: | line 360, column 5 Returning null reference |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
1 | // © 2016 and later: Unicode, Inc. and others. | |||
2 | // License & terms of use: http://www.unicode.org/copyright.html | |||
3 | /* | |||
4 | ********************************************************************** | |||
5 | * Copyright (c) 2001-2014, International Business Machines | |||
6 | * Corporation and others. All Rights Reserved. | |||
7 | ********************************************************************** | |||
8 | * Date Name Description | |||
9 | * 08/10/2001 aliu Creation. | |||
10 | ********************************************************************** | |||
11 | */ | |||
12 | ||||
13 | #include "unicode/utypes.h" | |||
14 | ||||
15 | #if !UCONFIG_NO_TRANSLITERATION0 | |||
16 | ||||
17 | #include "unicode/translit.h" | |||
18 | #include "unicode/resbund.h" | |||
19 | #include "unicode/uniset.h" | |||
20 | #include "unicode/uscript.h" | |||
21 | #include "rbt.h" | |||
22 | #include "cpdtrans.h" | |||
23 | #include "nultrans.h" | |||
24 | #include "transreg.h" | |||
25 | #include "rbt_data.h" | |||
26 | #include "rbt_pars.h" | |||
27 | #include "tridpars.h" | |||
28 | #include "charstr.h" | |||
29 | #include "uassert.h" | |||
30 | #include "locutil.h" | |||
31 | ||||
32 | // Enable the following symbol to add debugging code that tracks the | |||
33 | // allocation, deletion, and use of Entry objects. BoundsChecker has | |||
34 | // reported dangling pointer errors with these objects, but I have | |||
35 | // been unable to confirm them. I suspect BoundsChecker is getting | |||
36 | // confused with pointers going into and coming out of a UHashtable, | |||
37 | // despite the hinting code that is designed to help it. | |||
38 | // #define DEBUG_MEM | |||
39 | #ifdef DEBUG_MEM | |||
40 | #include <stdio.h> | |||
41 | #endif | |||
42 | ||||
43 | // UChar constants | |||
44 | static const UChar LOCALE_SEP = 95; // '_' | |||
45 | //static const UChar ID_SEP = 0x002D; /*-*/ | |||
46 | //static const UChar VARIANT_SEP = 0x002F; // '/' | |||
47 | ||||
48 | // String constants | |||
49 | static const UChar ANY[] = { 0x41, 0x6E, 0x79, 0 }; // Any | |||
50 | static const UChar LAT[] = { 0x4C, 0x61, 0x74, 0 }; // Lat | |||
51 | ||||
52 | // empty string | |||
53 | #define NO_VARIANTUnicodeString() UnicodeString() | |||
54 | ||||
55 | // initial estimate for specDAG size | |||
56 | // ICU 60 Transliterator::countAvailableSources() | |||
57 | #define SPECDAG_INIT_SIZE149 149 | |||
58 | ||||
59 | // initial estimate for number of variant names | |||
60 | #define VARIANT_LIST_INIT_SIZE11 11 | |||
61 | #define VARIANT_LIST_MAX_SIZE31 31 | |||
62 | ||||
63 | // initial estimate for availableIDs count (default estimate is 8 => multiple reallocs) | |||
64 | // ICU 60 Transliterator::countAvailableIDs() | |||
65 | #define AVAILABLE_IDS_INIT_SIZE641 641 | |||
66 | ||||
67 | // initial estimate for number of targets for source "Any", "Lat" | |||
68 | // ICU 60 Transliterator::countAvailableTargets("Any")/("Latn") | |||
69 | #define ANY_TARGETS_INIT_SIZE125 125 | |||
70 | #define LAT_TARGETS_INIT_SIZE23 23 | |||
71 | ||||
72 | /** | |||
73 | * Resource bundle key for the RuleBasedTransliterator rule. | |||
74 | */ | |||
75 | //static const char RB_RULE[] = "Rule"; | |||
76 | ||||
77 | U_NAMESPACE_BEGINnamespace icu_71 { | |||
78 | ||||
79 | //------------------------------------------------------------------ | |||
80 | // Alias | |||
81 | //------------------------------------------------------------------ | |||
82 | ||||
83 | TransliteratorAlias::TransliteratorAlias(const UnicodeString& theAliasID, | |||
84 | const UnicodeSet* cpdFilter) : | |||
85 | ID(), | |||
86 | aliasesOrRules(theAliasID), | |||
87 | transes(0), | |||
88 | compoundFilter(cpdFilter), | |||
89 | direction(UTRANS_FORWARD), | |||
90 | type(TransliteratorAlias::SIMPLE) { | |||
91 | } | |||
92 | ||||
93 | TransliteratorAlias::TransliteratorAlias(const UnicodeString& theID, | |||
94 | const UnicodeString& idBlocks, | |||
95 | UVector* adoptedTransliterators, | |||
96 | const UnicodeSet* cpdFilter) : | |||
97 | ID(theID), | |||
98 | aliasesOrRules(idBlocks), | |||
99 | transes(adoptedTransliterators), | |||
100 | compoundFilter(cpdFilter), | |||
101 | direction(UTRANS_FORWARD), | |||
102 | type(TransliteratorAlias::COMPOUND) { | |||
103 | } | |||
104 | ||||
105 | TransliteratorAlias::TransliteratorAlias(const UnicodeString& theID, | |||
106 | const UnicodeString& rules, | |||
107 | UTransDirection dir) : | |||
108 | ID(theID), | |||
109 | aliasesOrRules(rules), | |||
110 | transes(0), | |||
111 | compoundFilter(0), | |||
112 | direction(dir), | |||
113 | type(TransliteratorAlias::RULES) { | |||
114 | } | |||
115 | ||||
116 | TransliteratorAlias::~TransliteratorAlias() { | |||
117 | delete transes; | |||
118 | } | |||
119 | ||||
120 | ||||
121 | Transliterator* TransliteratorAlias::create(UParseError& pe, | |||
122 | UErrorCode& ec) { | |||
123 | if (U_FAILURE(ec)) { | |||
124 | return 0; | |||
125 | } | |||
126 | Transliterator *t = NULL__null; | |||
127 | switch (type) { | |||
128 | case SIMPLE: | |||
129 | t = Transliterator::createInstance(aliasesOrRules, UTRANS_FORWARD, pe, ec); | |||
130 | if(U_FAILURE(ec)){ | |||
131 | return 0; | |||
132 | } | |||
133 | if (compoundFilter != 0) | |||
134 | t->adoptFilter(compoundFilter->clone()); | |||
135 | break; | |||
136 | case COMPOUND: | |||
137 | { | |||
138 | // the total number of transliterators in the compound is the total number of anonymous transliterators | |||
139 | // plus the total number of ID blocks-- we start by assuming the list begins and ends with an ID | |||
140 | // block and that each pair anonymous transliterators has an ID block between them. Then we go back | |||
141 | // to see whether there really are ID blocks at the beginning and end (by looking for U+FFFF, which | |||
142 | // marks the position where an anonymous transliterator goes) and adjust accordingly | |||
143 | int32_t anonymousRBTs = transes->size(); | |||
144 | int32_t transCount = anonymousRBTs * 2 + 1; | |||
145 | if (!aliasesOrRules.isEmpty() && aliasesOrRules[0] == (UChar)(0xffff)) | |||
146 | --transCount; | |||
147 | if (aliasesOrRules.length() >= 2 && aliasesOrRules[aliasesOrRules.length() - 1] == (UChar)(0xffff)) | |||
148 | --transCount; | |||
149 | UnicodeString noIDBlock((UChar)(0xffff)); | |||
150 | noIDBlock += ((UChar)(0xffff)); | |||
151 | int32_t pos = aliasesOrRules.indexOf(noIDBlock); | |||
152 | while (pos >= 0) { | |||
153 | --transCount; | |||
154 | pos = aliasesOrRules.indexOf(noIDBlock, pos + 1); | |||
155 | } | |||
156 | ||||
157 | UVector transliterators(uprv_deleteUObjectuprv_deleteUObject_71, nullptr, ec); | |||
158 | UnicodeString idBlock; | |||
159 | int32_t blockSeparatorPos = aliasesOrRules.indexOf((UChar)(0xffff)); | |||
160 | while (blockSeparatorPos >= 0) { | |||
161 | aliasesOrRules.extract(0, blockSeparatorPos, idBlock); | |||
162 | aliasesOrRules.remove(0, blockSeparatorPos + 1); | |||
163 | if (!idBlock.isEmpty()) | |||
164 | transliterators.adoptElement(Transliterator::createInstance(idBlock, UTRANS_FORWARD, pe, ec), ec); | |||
165 | if (!transes->isEmpty()) | |||
166 | transliterators.adoptElement(transes->orphanElementAt(0), ec); | |||
167 | blockSeparatorPos = aliasesOrRules.indexOf((UChar)(0xffff)); | |||
168 | } | |||
169 | if (!aliasesOrRules.isEmpty()) | |||
170 | transliterators.adoptElement(Transliterator::createInstance(aliasesOrRules, UTRANS_FORWARD, pe, ec), ec); | |||
171 | while (!transes->isEmpty()) | |||
172 | transliterators.adoptElement(transes->orphanElementAt(0), ec); | |||
173 | transliterators.setDeleter(nullptr); | |||
174 | ||||
175 | if (U_SUCCESS(ec)) { | |||
176 | t = new CompoundTransliterator(ID, transliterators, | |||
177 | (compoundFilter ? compoundFilter->clone() : nullptr), | |||
178 | anonymousRBTs, pe, ec); | |||
179 | if (t == 0) { | |||
180 | ec = U_MEMORY_ALLOCATION_ERROR; | |||
181 | return 0; | |||
182 | } | |||
183 | } else { | |||
184 | for (int32_t i = 0; i < transliterators.size(); i++) | |||
185 | delete (Transliterator*)(transliterators.elementAt(i)); | |||
186 | } | |||
187 | } | |||
188 | break; | |||
189 | case RULES: | |||
190 | UPRV_UNREACHABLE_EXITabort(); // don't call create() if isRuleBased() returns TRUE! | |||
191 | } | |||
192 | return t; | |||
193 | } | |||
194 | ||||
195 | UBool TransliteratorAlias::isRuleBased() const { | |||
196 | return type == RULES; | |||
197 | } | |||
198 | ||||
199 | void TransliteratorAlias::parse(TransliteratorParser& parser, | |||
200 | UParseError& pe, UErrorCode& ec) const { | |||
201 | U_ASSERT(type == RULES)(void)0; | |||
202 | if (U_FAILURE(ec)) { | |||
203 | return; | |||
204 | } | |||
205 | ||||
206 | parser.parse(aliasesOrRules, direction, pe, ec); | |||
207 | } | |||
208 | ||||
209 | //---------------------------------------------------------------------- | |||
210 | // class TransliteratorSpec | |||
211 | //---------------------------------------------------------------------- | |||
212 | ||||
213 | /** | |||
214 | * A TransliteratorSpec is a string specifying either a source or a target. In more | |||
215 | * general terms, it may also specify a variant, but we only use the | |||
216 | * Spec class for sources and targets. | |||
217 | * | |||
218 | * A Spec may be a locale or a script. If it is a locale, it has a | |||
219 | * fallback chain that goes xx_YY_ZZZ -> xx_YY -> xx -> ssss, where | |||
220 | * ssss is the script mapping of xx_YY_ZZZ. The Spec API methods | |||
221 | * hasFallback(), next(), and reset() iterate over this fallback | |||
222 | * sequence. | |||
223 | * | |||
224 | * The Spec class canonicalizes itself, so the locale is put into | |||
225 | * canonical form, or the script is transformed from an abbreviation | |||
226 | * to a full name. | |||
227 | */ | |||
228 | class TransliteratorSpec : public UMemory { | |||
229 | public: | |||
230 | TransliteratorSpec(const UnicodeString& spec); | |||
231 | ~TransliteratorSpec(); | |||
232 | ||||
233 | const UnicodeString& get() const; | |||
234 | UBool hasFallback() const; | |||
235 | const UnicodeString& next(); | |||
236 | void reset(); | |||
237 | ||||
238 | UBool isLocale() const; | |||
239 | ResourceBundle& getBundle() const; | |||
240 | ||||
241 | operator const UnicodeString&() const { return get(); } | |||
242 | const UnicodeString& getTop() const { return top; } | |||
243 | ||||
244 | private: | |||
245 | void setupNext(); | |||
246 | ||||
247 | UnicodeString top; | |||
248 | UnicodeString spec; | |||
249 | UnicodeString nextSpec; | |||
250 | UnicodeString scriptName; | |||
251 | UBool isSpecLocale; // TRUE if spec is a locale | |||
252 | UBool isNextLocale; // TRUE if nextSpec is a locale | |||
253 | ResourceBundle* res; | |||
254 | ||||
255 | TransliteratorSpec(const TransliteratorSpec &other); // forbid copying of this class | |||
256 | TransliteratorSpec &operator=(const TransliteratorSpec &other); // forbid copying of this class | |||
257 | }; | |||
258 | ||||
259 | TransliteratorSpec::TransliteratorSpec(const UnicodeString& theSpec) | |||
260 | : top(theSpec), | |||
261 | res(0) | |||
262 | { | |||
263 | UErrorCode status = U_ZERO_ERROR; | |||
264 | Locale topLoc(""); | |||
265 | LocaleUtility::initLocaleFromName(theSpec, topLoc); | |||
266 | if (!topLoc.isBogus()) { | |||
267 | res = new ResourceBundle(U_ICUDATA_TRANSLIT"icudt" "71" "l" "-" "translit", topLoc, status); | |||
268 | /* test for NULL */ | |||
269 | if (res == 0) { | |||
270 | return; | |||
271 | } | |||
272 | if (U_FAILURE(status) || status == U_USING_DEFAULT_WARNING) { | |||
273 | delete res; | |||
274 | res = 0; | |||
275 | } | |||
276 | } | |||
277 | ||||
278 | // Canonicalize script name -or- do locale->script mapping | |||
279 | status = U_ZERO_ERROR; | |||
280 | static const int32_t capacity = 10; | |||
281 | UScriptCode script[capacity]={USCRIPT_INVALID_CODE}; | |||
282 | int32_t num = uscript_getCodeuscript_getCode_71(CharString().appendInvariantChars(theSpec, status).data(), | |||
283 | script, capacity, &status); | |||
284 | if (num > 0 && script[0] != USCRIPT_INVALID_CODE) { | |||
285 | scriptName = UnicodeString(uscript_getNameuscript_getName_71(script[0]), -1, US_INVicu::UnicodeString::kInvariant); | |||
286 | } | |||
287 | ||||
288 | // Canonicalize top | |||
289 | if (res
| |||
290 | // Canonicalize locale name | |||
291 | UnicodeString locStr; | |||
292 | LocaleUtility::initNameFromLocale(topLoc, locStr); | |||
293 | if (!locStr.isBogus()) { | |||
294 | top = locStr; | |||
295 | } | |||
296 | } else if (scriptName.length() != 0) { | |||
297 | // We are a script; use canonical name | |||
298 | top = scriptName; | |||
299 | } | |||
300 | ||||
301 | // assert(spec != top); | |||
302 | reset(); | |||
303 | } | |||
304 | ||||
305 | TransliteratorSpec::~TransliteratorSpec() { | |||
306 | delete res; | |||
307 | } | |||
308 | ||||
309 | UBool TransliteratorSpec::hasFallback() const { | |||
310 | return nextSpec.length() != 0; | |||
311 | } | |||
312 | ||||
313 | void TransliteratorSpec::reset() { | |||
314 | if (spec != top) { | |||
315 | spec = top; | |||
316 | isSpecLocale = (res != 0); | |||
317 | setupNext(); | |||
318 | } | |||
319 | } | |||
320 | ||||
321 | void TransliteratorSpec::setupNext() { | |||
322 | isNextLocale = FALSE0; | |||
323 | if (isSpecLocale) { | |||
324 | nextSpec = spec; | |||
325 | int32_t i = nextSpec.lastIndexOf(LOCALE_SEP); | |||
326 | // If i == 0 then we have _FOO, so we fall through | |||
327 | // to the scriptName. | |||
328 | if (i > 0) { | |||
329 | nextSpec.truncate(i); | |||
330 | isNextLocale = TRUE1; | |||
331 | } else { | |||
332 | nextSpec = scriptName; // scriptName may be empty | |||
333 | } | |||
334 | } else { | |||
335 | // spec is a script, so we are at the end | |||
336 | nextSpec.truncate(0); | |||
337 | } | |||
338 | } | |||
339 | ||||
340 | // Protocol: | |||
341 | // for(const UnicodeString& s(spec.get()); | |||
342 | // spec.hasFallback(); s(spec.next())) { ... | |||
343 | ||||
344 | const UnicodeString& TransliteratorSpec::next() { | |||
345 | spec = nextSpec; | |||
346 | isSpecLocale = isNextLocale; | |||
347 | setupNext(); | |||
348 | return spec; | |||
349 | } | |||
350 | ||||
351 | const UnicodeString& TransliteratorSpec::get() const { | |||
352 | return spec; | |||
353 | } | |||
354 | ||||
355 | UBool TransliteratorSpec::isLocale() const { | |||
356 | return isSpecLocale; | |||
357 | } | |||
358 | ||||
359 | ResourceBundle& TransliteratorSpec::getBundle() const { | |||
360 | return *res; | |||
| ||||
361 | } | |||
362 | ||||
363 | //---------------------------------------------------------------------- | |||
364 | ||||
365 | #ifdef DEBUG_MEM | |||
366 | ||||
367 | // Vector of Entry pointers currently in use | |||
368 | static UVector* DEBUG_entries = NULL__null; | |||
369 | ||||
370 | static void DEBUG_setup() { | |||
371 | if (DEBUG_entries == NULL__null) { | |||
372 | UErrorCode ec = U_ZERO_ERROR; | |||
373 | DEBUG_entries = new UVector(ec); | |||
374 | } | |||
375 | } | |||
376 | ||||
377 | // Caller must call DEBUG_setup first. Return index of given Entry, | |||
378 | // if it is in use (not deleted yet), or -1 if not found. | |||
379 | static int DEBUG_findEntry(TransliteratorEntry* e) { | |||
380 | for (int i=0; i<DEBUG_entries->size(); ++i) { | |||
381 | if (e == (TransliteratorEntry*) DEBUG_entries->elementAt(i)) { | |||
382 | return i; | |||
383 | } | |||
384 | } | |||
385 | return -1; | |||
386 | } | |||
387 | ||||
388 | // Track object creation | |||
389 | static void DEBUG_newEntry(TransliteratorEntry* e) { | |||
390 | DEBUG_setup(); | |||
391 | if (DEBUG_findEntry(e) >= 0) { | |||
392 | // This should really never happen unless the heap is broken | |||
393 | printf("ERROR DEBUG_newEntry duplicate new pointer %08X\n", e); | |||
394 | return; | |||
395 | } | |||
396 | UErrorCode ec = U_ZERO_ERROR; | |||
397 | DEBUG_entries->addElement(e, ec); | |||
398 | } | |||
399 | ||||
400 | // Track object deletion | |||
401 | static void DEBUG_delEntry(TransliteratorEntry* e) { | |||
402 | DEBUG_setup(); | |||
403 | int i = DEBUG_findEntry(e); | |||
404 | if (i < 0) { | |||
405 | printf("ERROR DEBUG_delEntry possible double deletion %08X\n", e); | |||
406 | return; | |||
407 | } | |||
408 | DEBUG_entries->removeElementAt(i); | |||
409 | } | |||
410 | ||||
411 | // Track object usage | |||
412 | static void DEBUG_useEntry(TransliteratorEntry* e) { | |||
413 | if (e == NULL__null) return; | |||
414 | DEBUG_setup(); | |||
415 | int i = DEBUG_findEntry(e); | |||
416 | if (i < 0) { | |||
417 | printf("ERROR DEBUG_useEntry possible dangling pointer %08X\n", e); | |||
418 | } | |||
419 | } | |||
420 | ||||
421 | #else | |||
422 | // If we're not debugging then make these macros into NOPs | |||
423 | #define DEBUG_newEntry(x) | |||
424 | #define DEBUG_delEntry(x) | |||
425 | #define DEBUG_useEntry(x) | |||
426 | #endif | |||
427 | ||||
428 | //---------------------------------------------------------------------- | |||
429 | // class Entry | |||
430 | //---------------------------------------------------------------------- | |||
431 | ||||
432 | /** | |||
433 | * The Entry object stores objects of different types and | |||
434 | * singleton objects as placeholders for rule-based transliterators to | |||
435 | * be built as needed. Instances of this struct can be placeholders, | |||
436 | * can represent prototype transliterators to be cloned, or can | |||
437 | * represent TransliteratorData objects. We don't support storing | |||
438 | * classes in the registry because we don't have the rtti infrastructure | |||
439 | * for it. We could easily add this if there is a need for it in the | |||
440 | * future. | |||
441 | */ | |||
442 | class TransliteratorEntry : public UMemory { | |||
443 | public: | |||
444 | enum Type { | |||
445 | RULES_FORWARD, | |||
446 | RULES_REVERSE, | |||
447 | LOCALE_RULES, | |||
448 | PROTOTYPE, | |||
449 | RBT_DATA, | |||
450 | COMPOUND_RBT, | |||
451 | ALIAS, | |||
452 | FACTORY, | |||
453 | NONE // Only used for uninitialized entries | |||
454 | } entryType; | |||
455 | // NOTE: stringArg cannot go inside the union because | |||
456 | // it has a copy constructor | |||
457 | UnicodeString stringArg; // For RULES_*, ALIAS, COMPOUND_RBT | |||
458 | int32_t intArg; // For COMPOUND_RBT, LOCALE_RULES | |||
459 | UnicodeSet* compoundFilter; // For COMPOUND_RBT | |||
460 | union { | |||
461 | Transliterator* prototype; // For PROTOTYPE | |||
462 | TransliterationRuleData* data; // For RBT_DATA | |||
463 | UVector* dataVector; // For COMPOUND_RBT | |||
464 | struct { | |||
465 | Transliterator::Factory function; | |||
466 | Transliterator::Token context; | |||
467 | } factory; // For FACTORY | |||
468 | } u; | |||
469 | TransliteratorEntry(); | |||
470 | ~TransliteratorEntry(); | |||
471 | void adoptPrototype(Transliterator* adopted); | |||
472 | void setFactory(Transliterator::Factory factory, | |||
473 | Transliterator::Token context); | |||
474 | ||||
475 | private: | |||
476 | ||||
477 | TransliteratorEntry(const TransliteratorEntry &other); // forbid copying of this class | |||
478 | TransliteratorEntry &operator=(const TransliteratorEntry &other); // forbid copying of this class | |||
479 | }; | |||
480 | ||||
481 | TransliteratorEntry::TransliteratorEntry() { | |||
482 | u.prototype = 0; | |||
483 | compoundFilter = NULL__null; | |||
484 | entryType = NONE; | |||
485 | DEBUG_newEntry(this); | |||
486 | } | |||
487 | ||||
488 | TransliteratorEntry::~TransliteratorEntry() { | |||
489 | DEBUG_delEntry(this); | |||
490 | if (entryType == PROTOTYPE) { | |||
491 | delete u.prototype; | |||
492 | } else if (entryType == RBT_DATA) { | |||
493 | // The data object is shared between instances of RBT. The | |||
494 | // entry object owns it. It should only be deleted when the | |||
495 | // transliterator component is being cleaned up. Doing so | |||
496 | // invalidates any RBTs that the user has instantiated. | |||
497 | delete u.data; | |||
498 | } else if (entryType == COMPOUND_RBT) { | |||
499 | while (u.dataVector != NULL__null && !u.dataVector->isEmpty()) | |||
500 | delete (TransliterationRuleData*)u.dataVector->orphanElementAt(0); | |||
501 | delete u.dataVector; | |||
502 | } | |||
503 | delete compoundFilter; | |||
504 | } | |||
505 | ||||
506 | void TransliteratorEntry::adoptPrototype(Transliterator* adopted) { | |||
507 | if (entryType == PROTOTYPE) { | |||
508 | delete u.prototype; | |||
509 | } | |||
510 | entryType = PROTOTYPE; | |||
511 | u.prototype = adopted; | |||
512 | } | |||
513 | ||||
514 | void TransliteratorEntry::setFactory(Transliterator::Factory factory, | |||
515 | Transliterator::Token context) { | |||
516 | if (entryType == PROTOTYPE) { | |||
517 | delete u.prototype; | |||
518 | } | |||
519 | entryType = FACTORY; | |||
520 | u.factory.function = factory; | |||
521 | u.factory.context = context; | |||
522 | } | |||
523 | ||||
524 | // UObjectDeleter for Hashtable::setValueDeleter | |||
525 | U_CDECL_BEGINextern "C" { | |||
526 | static void U_CALLCONV | |||
527 | deleteEntry(void* obj) { | |||
528 | delete (TransliteratorEntry*) obj; | |||
529 | } | |||
530 | U_CDECL_END} | |||
531 | ||||
532 | //---------------------------------------------------------------------- | |||
533 | // class TransliteratorRegistry: Basic public API | |||
534 | //---------------------------------------------------------------------- | |||
535 | ||||
536 | TransliteratorRegistry::TransliteratorRegistry(UErrorCode& status) : | |||
537 | registry(TRUE1, status), | |||
538 | specDAG(TRUE1, SPECDAG_INIT_SIZE149, status), | |||
539 | variantList(VARIANT_LIST_INIT_SIZE11, status), | |||
540 | availableIDs(AVAILABLE_IDS_INIT_SIZE641, status) | |||
541 | { | |||
542 | registry.setValueDeleter(deleteEntry); | |||
543 | variantList.setDeleter(uprv_deleteUObjectuprv_deleteUObject_71); | |||
544 | variantList.setComparer(uhash_compareCaselessUnicodeStringuhash_compareCaselessUnicodeString_71); | |||
545 | UnicodeString *emptyString = new UnicodeString(); | |||
546 | if (emptyString != NULL__null) { | |||
547 | variantList.adoptElement(emptyString, status); | |||
548 | } | |||
549 | availableIDs.setDeleter(uprv_deleteUObjectuprv_deleteUObject_71); | |||
550 | availableIDs.setComparer(uhash_compareCaselessUnicodeStringuhash_compareCaselessUnicodeString_71); | |||
551 | specDAG.setValueDeleter(uhash_deleteHashtableuhash_deleteHashtable_71); | |||
552 | } | |||
553 | ||||
554 | TransliteratorRegistry::~TransliteratorRegistry() { | |||
555 | // Through the magic of C++, everything cleans itself up | |||
556 | } | |||
557 | ||||
558 | Transliterator* TransliteratorRegistry::get(const UnicodeString& ID, | |||
559 | TransliteratorAlias*& aliasReturn, | |||
560 | UErrorCode& status) { | |||
561 | U_ASSERT(aliasReturn == NULL)(void)0; | |||
562 | TransliteratorEntry *entry = find(ID); | |||
563 | return (entry == 0) ? 0 | |||
564 | : instantiateEntry(ID, entry, aliasReturn, status); | |||
565 | } | |||
566 | ||||
567 | Transliterator* TransliteratorRegistry::reget(const UnicodeString& ID, | |||
568 | TransliteratorParser& parser, | |||
569 | TransliteratorAlias*& aliasReturn, | |||
570 | UErrorCode& status) { | |||
571 | U_ASSERT(aliasReturn == NULL)(void)0; | |||
572 | TransliteratorEntry *entry = find(ID); | |||
| ||||
573 | ||||
574 | if (entry == 0) { | |||
575 | // We get to this point if there are two threads, one of which | |||
576 | // is instantiating an ID, and another of which is removing | |||
577 | // the same ID from the registry, and the timing is just right. | |||
578 | return 0; | |||
579 | } | |||
580 | ||||
581 | // The usage model for the caller is that they will first call | |||
582 | // reg->get() inside the mutex, they'll get back an alias, they call | |||
583 | // alias->isRuleBased(), and if they get TRUE, they call alias->parse() | |||
584 | // outside the mutex, then reg->reget() inside the mutex again. A real | |||
585 | // mess, but it gets things working for ICU 3.0. [alan]. | |||
586 | ||||
587 | // Note: It's possible that in between the caller calling | |||
588 | // alias->parse() and reg->reget(), that another thread will have | |||
589 | // called reg->reget(), and the entry will already have been fixed up. | |||
590 | // We have to detect this so we don't stomp over existing entry | |||
591 | // data members and potentially leak memory (u.data and compoundFilter). | |||
592 | ||||
593 | if (entry->entryType == TransliteratorEntry::RULES_FORWARD || | |||
594 | entry->entryType == TransliteratorEntry::RULES_REVERSE || | |||
595 | entry->entryType == TransliteratorEntry::LOCALE_RULES) { | |||
596 | ||||
597 | if (parser.idBlockVector.isEmpty() && parser.dataVector.isEmpty()) { | |||
598 | entry->u.data = 0; | |||
599 | entry->entryType = TransliteratorEntry::ALIAS; | |||
600 | entry->stringArg = UNICODE_STRING_SIMPLE("Any-NULL")icu::UnicodeString(true, u"Any-NULL", -1); | |||
601 | } | |||
602 | else if (parser.idBlockVector.isEmpty() && parser.dataVector.size() == 1) { | |||
603 | entry->u.data = (TransliterationRuleData*)parser.dataVector.orphanElementAt(0); | |||
604 | entry->entryType = TransliteratorEntry::RBT_DATA; | |||
605 | } | |||
606 | else if (parser.idBlockVector.size() == 1 && parser.dataVector.isEmpty()) { | |||
607 | entry->stringArg = *(UnicodeString*)(parser.idBlockVector.elementAt(0)); | |||
608 | entry->compoundFilter = parser.orphanCompoundFilter(); | |||
609 | entry->entryType = TransliteratorEntry::ALIAS; | |||
610 | } | |||
611 | else { | |||
612 | entry->entryType = TransliteratorEntry::COMPOUND_RBT; | |||
613 | entry->compoundFilter = parser.orphanCompoundFilter(); | |||
614 | entry->u.dataVector = new UVector(status); | |||
615 | // TODO ICU-21701: missing check for nullptr and failed status. | |||
616 | // Unclear how best to bail out. | |||
617 | entry->stringArg.remove(); | |||
618 | ||||
619 | int32_t limit = parser.idBlockVector.size(); | |||
620 | if (parser.dataVector.size() > limit) | |||
621 | limit = parser.dataVector.size(); | |||
622 | ||||
623 | for (int32_t i = 0; i < limit; i++) { | |||
624 | if (i < parser.idBlockVector.size()) { | |||
625 | UnicodeString* idBlock = (UnicodeString*)parser.idBlockVector.elementAt(i); | |||
626 | if (!idBlock->isEmpty()) | |||
627 | entry->stringArg += *idBlock; | |||
628 | } | |||
629 | if (!parser.dataVector.isEmpty()) { | |||
630 | TransliterationRuleData* data = (TransliterationRuleData*)parser.dataVector.orphanElementAt(0); | |||
631 | entry->u.dataVector->addElement(data, status); | |||
632 | if (U_FAILURE(status)) { | |||
633 | delete data; | |||
634 | } | |||
635 | entry->stringArg += (UChar)0xffff; // use U+FFFF to mark position of RBTs in ID block | |||
636 | } | |||
637 | } | |||
638 | } | |||
639 | } | |||
640 | ||||
641 | Transliterator *t = | |||
642 | instantiateEntry(ID, entry, aliasReturn, status); | |||
643 | return t; | |||
644 | } | |||
645 | ||||
646 | void TransliteratorRegistry::put(Transliterator* adoptedProto, | |||
647 | UBool visible, | |||
648 | UErrorCode& ec) | |||
649 | { | |||
650 | TransliteratorEntry *entry = new TransliteratorEntry(); | |||
651 | if (entry == NULL__null) { | |||
652 | ec = U_MEMORY_ALLOCATION_ERROR; | |||
653 | return; | |||
654 | } | |||
655 | entry->adoptPrototype(adoptedProto); | |||
656 | registerEntry(adoptedProto->getID(), entry, visible); | |||
657 | } | |||
658 | ||||
659 | void TransliteratorRegistry::put(const UnicodeString& ID, | |||
660 | Transliterator::Factory factory, | |||
661 | Transliterator::Token context, | |||
662 | UBool visible, | |||
663 | UErrorCode& ec) { | |||
664 | TransliteratorEntry *entry = new TransliteratorEntry(); | |||
665 | if (entry == NULL__null) { | |||
666 | ec = U_MEMORY_ALLOCATION_ERROR; | |||
667 | return; | |||
668 | } | |||
669 | entry->setFactory(factory, context); | |||
670 | registerEntry(ID, entry, visible); | |||
671 | } | |||
672 | ||||
673 | void TransliteratorRegistry::put(const UnicodeString& ID, | |||
674 | const UnicodeString& resourceName, | |||
675 | UTransDirection dir, | |||
676 | UBool readonlyResourceAlias, | |||
677 | UBool visible, | |||
678 | UErrorCode& ec) { | |||
679 | TransliteratorEntry *entry = new TransliteratorEntry(); | |||
680 | if (entry == NULL__null) { | |||
681 | ec = U_MEMORY_ALLOCATION_ERROR; | |||
682 | return; | |||
683 | } | |||
684 | entry->entryType = (dir == UTRANS_FORWARD) ? TransliteratorEntry::RULES_FORWARD | |||
685 | : TransliteratorEntry::RULES_REVERSE; | |||
686 | if (readonlyResourceAlias) { | |||
687 | entry->stringArg.setTo(TRUE1, resourceName.getBuffer(), -1); | |||
688 | } | |||
689 | else { | |||
690 | entry->stringArg = resourceName; | |||
691 | } | |||
692 | registerEntry(ID, entry, visible); | |||
693 | } | |||
694 | ||||
695 | void TransliteratorRegistry::put(const UnicodeString& ID, | |||
696 | const UnicodeString& alias, | |||
697 | UBool readonlyAliasAlias, | |||
698 | UBool visible, | |||
699 | UErrorCode& /*ec*/) { | |||
700 | TransliteratorEntry *entry = new TransliteratorEntry(); | |||
701 | // Null pointer check | |||
702 | if (entry != NULL__null) { | |||
703 | entry->entryType = TransliteratorEntry::ALIAS; | |||
704 | if (readonlyAliasAlias) { | |||
705 | entry->stringArg.setTo(TRUE1, alias.getBuffer(), -1); | |||
706 | } | |||
707 | else { | |||
708 | entry->stringArg = alias; | |||
709 | } | |||
710 | registerEntry(ID, entry, visible); | |||
711 | } | |||
712 | } | |||
713 | ||||
714 | void TransliteratorRegistry::remove(const UnicodeString& ID) { | |||
715 | UnicodeString source, target, variant; | |||
716 | UBool sawSource; | |||
717 | TransliteratorIDParser::IDtoSTV(ID, source, target, variant, sawSource); | |||
718 | // Only need to do this if ID.indexOf('-') < 0 | |||
719 | UnicodeString id; | |||
720 | TransliteratorIDParser::STVtoID(source, target, variant, id); | |||
721 | registry.remove(id); | |||
722 | removeSTV(source, target, variant); | |||
723 | availableIDs.removeElement((void*) &id); | |||
724 | } | |||
725 | ||||
726 | //---------------------------------------------------------------------- | |||
727 | // class TransliteratorRegistry: Public ID and spec management | |||
728 | //---------------------------------------------------------------------- | |||
729 | ||||
730 | /** | |||
731 | * == OBSOLETE - remove in ICU 3.4 == | |||
732 | * Return the number of IDs currently registered with the system. | |||
733 | * To retrieve the actual IDs, call getAvailableID(i) with | |||
734 | * i from 0 to countAvailableIDs() - 1. | |||
735 | */ | |||
736 | int32_t TransliteratorRegistry::countAvailableIDs(void) const { | |||
737 | return availableIDs.size(); | |||
738 | } | |||
739 | ||||
740 | /** | |||
741 | * == OBSOLETE - remove in ICU 3.4 == | |||
742 | * Return the index-th available ID. index must be between 0 | |||
743 | * and countAvailableIDs() - 1, inclusive. If index is out of | |||
744 | * range, the result of getAvailableID(0) is returned. | |||
745 | */ | |||
746 | const UnicodeString& TransliteratorRegistry::getAvailableID(int32_t index) const { | |||
747 | if (index < 0 || index >= availableIDs.size()) { | |||
748 | index = 0; | |||
749 | } | |||
750 | return *(const UnicodeString*) availableIDs[index]; | |||
751 | } | |||
752 | ||||
753 | StringEnumeration* TransliteratorRegistry::getAvailableIDs() const { | |||
754 | return new Enumeration(*this); | |||
755 | } | |||
756 | ||||
757 | int32_t TransliteratorRegistry::countAvailableSources(void) const { | |||
758 | return specDAG.count(); | |||
759 | } | |||
760 | ||||
761 | UnicodeString& TransliteratorRegistry::getAvailableSource(int32_t index, | |||
762 | UnicodeString& result) const { | |||
763 | int32_t pos = UHASH_FIRST(-1); | |||
764 | const UHashElement *e = 0; | |||
765 | while (index-- >= 0) { | |||
766 | e = specDAG.nextElement(pos); | |||
767 | if (e == 0) { | |||
768 | break; | |||
769 | } | |||
770 | } | |||
771 | if (e == 0) { | |||
772 | result.truncate(0); | |||
773 | } else { | |||
774 | result = *(UnicodeString*) e->key.pointer; | |||
775 | } | |||
776 | return result; | |||
777 | } | |||
778 | ||||
779 | int32_t TransliteratorRegistry::countAvailableTargets(const UnicodeString& source) const { | |||
780 | Hashtable *targets = (Hashtable*) specDAG.get(source); | |||
781 | return (targets == 0) ? 0 : targets->count(); | |||
782 | } | |||
783 | ||||
784 | UnicodeString& TransliteratorRegistry::getAvailableTarget(int32_t index, | |||
785 | const UnicodeString& source, | |||
786 | UnicodeString& result) const { | |||
787 | Hashtable *targets = (Hashtable*) specDAG.get(source); | |||
788 | if (targets == 0) { | |||
789 | result.truncate(0); // invalid source | |||
790 | return result; | |||
791 | } | |||
792 | int32_t pos = UHASH_FIRST(-1); | |||
793 | const UHashElement *e = 0; | |||
794 | while (index-- >= 0) { | |||
795 | e = targets->nextElement(pos); | |||
796 | if (e == 0) { | |||
797 | break; | |||
798 | } | |||
799 | } | |||
800 | if (e == 0) { | |||
801 | result.truncate(0); // invalid index | |||
802 | } else { | |||
803 | result = *(UnicodeString*) e->key.pointer; | |||
804 | } | |||
805 | return result; | |||
806 | } | |||
807 | ||||
808 | int32_t TransliteratorRegistry::countAvailableVariants(const UnicodeString& source, | |||
809 | const UnicodeString& target) const { | |||
810 | Hashtable *targets = (Hashtable*) specDAG.get(source); | |||
811 | if (targets == 0) { | |||
812 | return 0; | |||
813 | } | |||
814 | uint32_t varMask = targets->geti(target); | |||
815 | int32_t varCount = 0; | |||
816 | while (varMask > 0) { | |||
817 | if (varMask & 1) { | |||
818 | varCount++; | |||
819 | } | |||
820 | varMask >>= 1; | |||
821 | } | |||
822 | return varCount; | |||
823 | } | |||
824 | ||||
825 | UnicodeString& TransliteratorRegistry::getAvailableVariant(int32_t index, | |||
826 | const UnicodeString& source, | |||
827 | const UnicodeString& target, | |||
828 | UnicodeString& result) const { | |||
829 | Hashtable *targets = (Hashtable*) specDAG.get(source); | |||
830 | if (targets == 0) { | |||
831 | result.truncate(0); // invalid source | |||
832 | return result; | |||
833 | } | |||
834 | uint32_t varMask = targets->geti(target); | |||
835 | int32_t varCount = 0; | |||
836 | int32_t varListIndex = 0; | |||
837 | while (varMask > 0) { | |||
838 | if (varMask & 1) { | |||
839 | if (varCount == index) { | |||
840 | UnicodeString *v = (UnicodeString*) variantList.elementAt(varListIndex); | |||
841 | if (v != NULL__null) { | |||
842 | result = *v; | |||
843 | return result; | |||
844 | } | |||
845 | break; | |||
846 | } | |||
847 | varCount++; | |||
848 | } | |||
849 | varMask >>= 1; | |||
850 | varListIndex++; | |||
851 | } | |||
852 | result.truncate(0); // invalid target or index | |||
853 | return result; | |||
854 | } | |||
855 | ||||
856 | //---------------------------------------------------------------------- | |||
857 | // class TransliteratorRegistry::Enumeration | |||
858 | //---------------------------------------------------------------------- | |||
859 | ||||
860 | TransliteratorRegistry::Enumeration::Enumeration(const TransliteratorRegistry& _reg) : | |||
861 | index(0), reg(_reg) { | |||
862 | } | |||
863 | ||||
864 | TransliteratorRegistry::Enumeration::~Enumeration() { | |||
865 | } | |||
866 | ||||
867 | int32_t TransliteratorRegistry::Enumeration::count(UErrorCode& /*status*/) const { | |||
868 | return reg.availableIDs.size(); | |||
869 | } | |||
870 | ||||
871 | const UnicodeString* TransliteratorRegistry::Enumeration::snext(UErrorCode& status) { | |||
872 | // This is sloppy but safe -- if we get out of sync with the underlying | |||
873 | // registry, we will still return legal strings, but they might not | |||
874 | // correspond to the snapshot at construction time. So there could be | |||
875 | // duplicate IDs or omitted IDs if insertions or deletions occur in one | |||
876 | // thread while another is iterating. To be more rigorous, add a timestamp, | |||
877 | // which is incremented with any modification, and validate this iterator | |||
878 | // against the timestamp at construction time. This probably isn't worth | |||
879 | // doing as long as there is some possibility of removing this code in favor | |||
880 | // of some new code based on Doug's service framework. | |||
881 | if (U_FAILURE(status)) { | |||
882 | return NULL__null; | |||
883 | } | |||
884 | int32_t n = reg.availableIDs.size(); | |||
885 | if (index > n) { | |||
886 | status = U_ENUM_OUT_OF_SYNC_ERROR; | |||
887 | } | |||
888 | // index == n is okay -- this means we've reached the end | |||
889 | if (index < n) { | |||
890 | // Copy the string! This avoids lifetime problems. | |||
891 | unistr = *(const UnicodeString*)reg.availableIDs[index++]; | |||
892 | return &unistr; | |||
893 | } else { | |||
894 | return NULL__null; | |||
895 | } | |||
896 | } | |||
897 | ||||
898 | void TransliteratorRegistry::Enumeration::reset(UErrorCode& /*status*/) { | |||
899 | index = 0; | |||
900 | } | |||
901 | ||||
902 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TransliteratorRegistry::Enumeration)UClassID TransliteratorRegistry::Enumeration::getStaticClassID () { static char classID = 0; return (UClassID)&classID; } UClassID TransliteratorRegistry::Enumeration::getDynamicClassID () const { return TransliteratorRegistry::Enumeration::getStaticClassID (); } | |||
903 | ||||
904 | //---------------------------------------------------------------------- | |||
905 | // class TransliteratorRegistry: internal | |||
906 | //---------------------------------------------------------------------- | |||
907 | ||||
908 | /** | |||
909 | * Convenience method. Calls 6-arg registerEntry(). | |||
910 | */ | |||
911 | void TransliteratorRegistry::registerEntry(const UnicodeString& source, | |||
912 | const UnicodeString& target, | |||
913 | const UnicodeString& variant, | |||
914 | TransliteratorEntry* adopted, | |||
915 | UBool visible) { | |||
916 | UnicodeString ID; | |||
917 | UnicodeString s(source); | |||
918 | if (s.length() == 0) { | |||
919 | s.setTo(TRUE1, ANY, 3); | |||
920 | } | |||
921 | TransliteratorIDParser::STVtoID(source, target, variant, ID); | |||
922 | registerEntry(ID, s, target, variant, adopted, visible); | |||
923 | } | |||
924 | ||||
925 | /** | |||
926 | * Convenience method. Calls 6-arg registerEntry(). | |||
927 | */ | |||
928 | void TransliteratorRegistry::registerEntry(const UnicodeString& ID, | |||
929 | TransliteratorEntry* adopted, | |||
930 | UBool visible) { | |||
931 | UnicodeString source, target, variant; | |||
932 | UBool sawSource; | |||
933 | TransliteratorIDParser::IDtoSTV(ID, source, target, variant, sawSource); | |||
934 | // Only need to do this if ID.indexOf('-') < 0 | |||
935 | UnicodeString id; | |||
936 | TransliteratorIDParser::STVtoID(source, target, variant, id); | |||
937 | registerEntry(id, source, target, variant, adopted, visible); | |||
938 | } | |||
939 | ||||
940 | /** | |||
941 | * Register an entry object (adopted) with the given ID, source, | |||
942 | * target, and variant strings. | |||
943 | */ | |||
944 | void TransliteratorRegistry::registerEntry(const UnicodeString& ID, | |||
945 | const UnicodeString& source, | |||
946 | const UnicodeString& target, | |||
947 | const UnicodeString& variant, | |||
948 | TransliteratorEntry* adopted, | |||
949 | UBool visible) { | |||
950 | UErrorCode status = U_ZERO_ERROR; | |||
951 | registry.put(ID, adopted, status); | |||
952 | if (visible) { | |||
953 | registerSTV(source, target, variant); | |||
954 | if (!availableIDs.contains((void*) &ID)) { | |||
955 | UnicodeString *newID = ID.clone(); | |||
956 | // Check to make sure newID was created. | |||
957 | if (newID != NULL__null) { | |||
958 | // NUL-terminate the ID string | |||
959 | newID->getTerminatedBuffer(); | |||
960 | availableIDs.adoptElement(newID, status); | |||
961 | } | |||
962 | } | |||
963 | } else { | |||
964 | removeSTV(source, target, variant); | |||
965 | availableIDs.removeElement((void*) &ID); | |||
966 | } | |||
967 | } | |||
968 | ||||
969 | /** | |||
970 | * Register a source-target/variant in the specDAG. Variant may be | |||
971 | * empty, but source and target must not be. | |||
972 | */ | |||
973 | void TransliteratorRegistry::registerSTV(const UnicodeString& source, | |||
974 | const UnicodeString& target, | |||
975 | const UnicodeString& variant) { | |||
976 | // assert(source.length() > 0); | |||
977 | // assert(target.length() > 0); | |||
978 | UErrorCode status = U_ZERO_ERROR; | |||
979 | Hashtable *targets = (Hashtable*) specDAG.get(source); | |||
980 | if (targets == 0) { | |||
981 | int32_t size = 3; | |||
982 | if (source.compare(ANY,3) == 0) { | |||
983 | size = ANY_TARGETS_INIT_SIZE125; | |||
984 | } else if (source.compare(LAT,3) == 0) { | |||
985 | size = LAT_TARGETS_INIT_SIZE23; | |||
986 | } | |||
987 | targets = new Hashtable(TRUE1, size, status); | |||
988 | if (U_FAILURE(status) || targets == NULL__null) { | |||
989 | return; | |||
990 | } | |||
991 | specDAG.put(source, targets, status); | |||
992 | } | |||
993 | int32_t variantListIndex = variantList.indexOf((void*) &variant, 0); | |||
994 | if (variantListIndex < 0) { | |||
995 | if (variantList.size() >= VARIANT_LIST_MAX_SIZE31) { | |||
996 | // can't handle any more variants | |||
997 | return; | |||
998 | } | |||
999 | UnicodeString *variantEntry = new UnicodeString(variant); | |||
1000 | if (variantEntry != NULL__null) { | |||
1001 | variantList.adoptElement(variantEntry, status); | |||
1002 | if (U_SUCCESS(status)) { | |||
1003 | variantListIndex = variantList.size() - 1; | |||
1004 | } | |||
1005 | } | |||
1006 | if (variantListIndex < 0) { | |||
1007 | return; | |||
1008 | } | |||
1009 | } | |||
1010 | uint32_t addMask = 1 << variantListIndex; | |||
1011 | uint32_t varMask = targets->geti(target); | |||
1012 | targets->puti(target, varMask | addMask, status); | |||
1013 | } | |||
1014 | ||||
1015 | /** | |||
1016 | * Remove a source-target/variant from the specDAG. | |||
1017 | */ | |||
1018 | void TransliteratorRegistry::removeSTV(const UnicodeString& source, | |||
1019 | const UnicodeString& target, | |||
1020 | const UnicodeString& variant) { | |||
1021 | // assert(source.length() > 0); | |||
1022 | // assert(target.length() > 0); | |||
1023 | UErrorCode status = U_ZERO_ERROR; | |||
1024 | Hashtable *targets = (Hashtable*) specDAG.get(source); | |||
1025 | if (targets == NULL__null) { | |||
1026 | return; // should never happen for valid s-t/v | |||
1027 | } | |||
1028 | uint32_t varMask = targets->geti(target); | |||
1029 | if (varMask == 0) { | |||
1030 | return; // should never happen for valid s-t/v | |||
1031 | } | |||
1032 | int32_t variantListIndex = variantList.indexOf((void*) &variant, 0); | |||
1033 | if (variantListIndex < 0) { | |||
1034 | return; // should never happen for valid s-t/v | |||
1035 | } | |||
1036 | int32_t remMask = 1 << variantListIndex; | |||
1037 | varMask &= (~remMask); | |||
1038 | if (varMask != 0) { | |||
1039 | targets->puti(target, varMask, status); | |||
1040 | } else { | |||
1041 | targets->remove(target); // should delete variants | |||
1042 | if (targets->count() == 0) { | |||
1043 | specDAG.remove(source); // should delete targets | |||
1044 | } | |||
1045 | } | |||
1046 | } | |||
1047 | ||||
1048 | /** | |||
1049 | * Attempt to find a source-target/variant in the dynamic registry | |||
1050 | * store. Return 0 on failure. | |||
1051 | * | |||
1052 | * Caller does NOT own returned object. | |||
1053 | */ | |||
1054 | TransliteratorEntry* TransliteratorRegistry::findInDynamicStore(const TransliteratorSpec& src, | |||
1055 | const TransliteratorSpec& trg, | |||
1056 | const UnicodeString& variant) const { | |||
1057 | UnicodeString ID; | |||
1058 | TransliteratorIDParser::STVtoID(src, trg, variant, ID); | |||
1059 | TransliteratorEntry *e = (TransliteratorEntry*) registry.get(ID); | |||
1060 | DEBUG_useEntry(e); | |||
1061 | return e; | |||
1062 | } | |||
1063 | ||||
1064 | /** | |||
1065 | * Attempt to find a source-target/variant in the static locale | |||
1066 | * resource store. Do not perform fallback. Return 0 on failure. | |||
1067 | * | |||
1068 | * On success, create a new entry object, register it in the dynamic | |||
1069 | * store, and return a pointer to it, but do not make it public -- | |||
1070 | * just because someone requested something, we do not expand the | |||
1071 | * available ID list (or spec DAG). | |||
1072 | * | |||
1073 | * Caller does NOT own returned object. | |||
1074 | */ | |||
1075 | TransliteratorEntry* TransliteratorRegistry::findInStaticStore(const TransliteratorSpec& src, | |||
1076 | const TransliteratorSpec& trg, | |||
1077 | const UnicodeString& variant) { | |||
1078 | TransliteratorEntry* entry = 0; | |||
1079 | if (src.isLocale()) { | |||
1080 | entry = findInBundle(src, trg, variant, UTRANS_FORWARD); | |||
1081 | } else if (trg.isLocale()) { | |||
1082 | entry = findInBundle(trg, src, variant, UTRANS_REVERSE); | |||
1083 | } | |||
1084 | ||||
1085 | // If we found an entry, store it in the Hashtable for next | |||
1086 | // time. | |||
1087 | if (entry != 0) { | |||
1088 | registerEntry(src.getTop(), trg.getTop(), variant, entry, FALSE0); | |||
1089 | } | |||
1090 | ||||
1091 | return entry; | |||
1092 | } | |||
1093 | ||||
1094 | // As of 2.0, resource bundle keys cannot contain '_' | |||
1095 | static const UChar TRANSLITERATE_TO[] = {84,114,97,110,115,108,105,116,101,114,97,116,101,84,111,0}; // "TransliterateTo" | |||
1096 | ||||
1097 | static const UChar TRANSLITERATE_FROM[] = {84,114,97,110,115,108,105,116,101,114,97,116,101,70,114,111,109,0}; // "TransliterateFrom" | |||
1098 | ||||
1099 | static const UChar TRANSLITERATE[] = {84,114,97,110,115,108,105,116,101,114,97,116,101,0}; // "Transliterate" | |||
1100 | ||||
1101 | /** | |||
1102 | * Attempt to find an entry in a single resource bundle. This is | |||
1103 | * a one-sided lookup. findInStaticStore() performs up to two such | |||
1104 | * lookups, one for the source, and one for the target. | |||
1105 | * | |||
1106 | * Do not perform fallback. Return 0 on failure. | |||
1107 | * | |||
1108 | * On success, create a new Entry object, populate it, and return it. | |||
1109 | * The caller owns the returned object. | |||
1110 | */ | |||
1111 | TransliteratorEntry* TransliteratorRegistry::findInBundle(const TransliteratorSpec& specToOpen, | |||
1112 | const TransliteratorSpec& specToFind, | |||
1113 | const UnicodeString& variant, | |||
1114 | UTransDirection direction) | |||
1115 | { | |||
1116 | UnicodeString utag; | |||
1117 | UnicodeString resStr; | |||
1118 | int32_t pass; | |||
1119 | ||||
1120 | for (pass=0; pass<2; ++pass) { | |||
1121 | utag.truncate(0); | |||
1122 | // First try either TransliteratorTo_xxx or | |||
1123 | // TransliterateFrom_xxx, then try the bidirectional | |||
1124 | // Transliterate_xxx. This precedence order is arbitrary | |||
1125 | // but must be consistent and documented. | |||
1126 | if (pass
| |||
1127 | utag.append(direction
| |||
1128 | TRANSLITERATE_TO : TRANSLITERATE_FROM, -1); | |||
1129 | } else { | |||
1130 | utag.append(TRANSLITERATE, -1); | |||
1131 | } | |||
1132 | UnicodeString s(specToFind.get()); | |||
1133 | utag.append(s.toUpper("")); | |||
1134 | UErrorCode status = U_ZERO_ERROR; | |||
1135 | ResourceBundle subres(specToOpen.getBundle().get( | |||
1136 | CharString().appendInvariantChars(utag, status).data(), status)); | |||
1137 | if (U_FAILURE(status) || status == U_USING_DEFAULT_WARNING) { | |||
1138 | continue; | |||
1139 | } | |||
1140 | ||||
1141 | s.truncate(0); | |||
1142 | if (specToOpen.get() != LocaleUtility::initNameFromLocale(subres.getLocale(), s)) { | |||
1143 | continue; | |||
1144 | } | |||
1145 | ||||
1146 | if (variant.length() != 0) { | |||
1147 | status = U_ZERO_ERROR; | |||
1148 | resStr = subres.getStringEx( | |||
1149 | CharString().appendInvariantChars(variant, status).data(), status); | |||
1150 | if (U_SUCCESS(status)) { | |||
1151 | // Exit loop successfully | |||
1152 | break; | |||
1153 | } | |||
1154 | } else { | |||
1155 | // Variant is empty, which means match the first variant listed. | |||
1156 | status = U_ZERO_ERROR; | |||
1157 | resStr = subres.getStringEx(1, status); | |||
1158 | if (U_SUCCESS(status)) { | |||
1159 | // Exit loop successfully | |||
1160 | break; | |||
1161 | } | |||
1162 | } | |||
1163 | } | |||
1164 | ||||
1165 | if (pass==2) { | |||
1166 | // Failed | |||
1167 | return NULL__null; | |||
1168 | } | |||
1169 | ||||
1170 | // We have succeeded in loading a string from the locale | |||
1171 | // resources. Create a new registry entry to hold it and return it. | |||
1172 | TransliteratorEntry *entry = new TransliteratorEntry(); | |||
1173 | if (entry != 0) { | |||
1174 | // The direction is always forward for the | |||
1175 | // TransliterateTo_xxx and TransliterateFrom_xxx | |||
1176 | // items; those are unidirectional forward rules. | |||
1177 | // For the bidirectional Transliterate_xxx items, | |||
1178 | // the direction is the value passed in to this | |||
1179 | // function. | |||
1180 | int32_t dir = (pass == 0) ? UTRANS_FORWARD : direction; | |||
1181 | entry->entryType = TransliteratorEntry::LOCALE_RULES; | |||
1182 | entry->stringArg = resStr; | |||
1183 | entry->intArg = dir; | |||
1184 | } | |||
1185 | ||||
1186 | return entry; | |||
1187 | } | |||
1188 | ||||
1189 | /** | |||
1190 | * Convenience method. Calls 3-arg find(). | |||
1191 | */ | |||
1192 | TransliteratorEntry* TransliteratorRegistry::find(const UnicodeString& ID) { | |||
1193 | UnicodeString source, target, variant; | |||
1194 | UBool sawSource; | |||
1195 | TransliteratorIDParser::IDtoSTV(ID, source, target, variant, sawSource); | |||
1196 | return find(source, target, variant); | |||
1197 | } | |||
1198 | ||||
1199 | /** | |||
1200 | * Top-level find method. Attempt to find a source-target/variant in | |||
1201 | * either the dynamic or the static (locale resource) store. Perform | |||
1202 | * fallback. | |||
1203 | * | |||
1204 | * Lookup sequence for ss_SS_SSS-tt_TT_TTT/v: | |||
1205 | * | |||
1206 | * ss_SS_SSS-tt_TT_TTT/v -- in hashtable | |||
1207 | * ss_SS_SSS-tt_TT_TTT/v -- in ss_SS_SSS (no fallback) | |||
1208 | * | |||
1209 | * repeat with t = tt_TT_TTT, tt_TT, tt, and tscript | |||
1210 | * | |||
1211 | * ss_SS_SSS-t/ * | |||
1212 | * ss_SS-t/ * | |||
1213 | * ss-t/ * | |||
1214 | * sscript-t/ * | |||
1215 | * | |||
1216 | * Here * matches the first variant listed. | |||
1217 | * | |||
1218 | * Caller does NOT own returned object. Return 0 on failure. | |||
1219 | */ | |||
1220 | TransliteratorEntry* TransliteratorRegistry::find(UnicodeString& source, | |||
1221 | UnicodeString& target, | |||
1222 | UnicodeString& variant) { | |||
1223 | ||||
1224 | TransliteratorSpec src(source); | |||
1225 | TransliteratorSpec trg(target); | |||
1226 | TransliteratorEntry* entry; | |||
1227 | ||||
1228 | // Seek exact match in hashtable. Temporary fix for ICU 4.6. | |||
1229 | // TODO: The general logic for finding a matching transliterator needs to be reviewed. | |||
1230 | // ICU ticket #8089 | |||
1231 | UnicodeString ID; | |||
1232 | TransliteratorIDParser::STVtoID(source, target, variant, ID); | |||
1233 | entry = (TransliteratorEntry*) registry.get(ID); | |||
1234 | if (entry != 0) { | |||
1235 | // std::string ss; | |||
1236 | // std::cout << ID.toUTF8String(ss) << std::endl; | |||
1237 | return entry; | |||
1238 | } | |||
1239 | ||||
1240 | if (variant.length() != 0) { | |||
1241 | ||||
1242 | // Seek exact match in hashtable | |||
1243 | entry = findInDynamicStore(src, trg, variant); | |||
1244 | if (entry != 0) { | |||
1245 | return entry; | |||
1246 | } | |||
1247 | ||||
1248 | // Seek exact match in locale resources | |||
1249 | entry = findInStaticStore(src, trg, variant); | |||
1250 | if (entry != 0) { | |||
1251 | return entry; | |||
1252 | } | |||
1253 | } | |||
1254 | ||||
1255 | for (;;) { | |||
1256 | src.reset(); | |||
1257 | for (;;) { | |||
1258 | // Seek match in hashtable | |||
1259 | entry = findInDynamicStore(src, trg, NO_VARIANTUnicodeString()); | |||
1260 | if (entry != 0) { | |||
1261 | return entry; | |||
1262 | } | |||
1263 | ||||
1264 | // Seek match in locale resources | |||
1265 | entry = findInStaticStore(src, trg, NO_VARIANTUnicodeString()); | |||
1266 | if (entry != 0) { | |||
1267 | return entry; | |||
1268 | } | |||
1269 | if (!src.hasFallback()) { | |||
1270 | break; | |||
1271 | } | |||
1272 | src.next(); | |||
1273 | } | |||
1274 | if (!trg.hasFallback()) { | |||
1275 | break; | |||
1276 | } | |||
1277 | trg.next(); | |||
1278 | } | |||
1279 | ||||
1280 | return 0; | |||
1281 | } | |||
1282 | ||||
1283 | /** | |||
1284 | * Given an Entry object, instantiate it. Caller owns result. Return | |||
1285 | * 0 on failure. | |||
1286 | * | |||
1287 | * Return a non-empty aliasReturn value if the ID points to an alias. | |||
1288 | * We cannot instantiate it ourselves because the alias may contain | |||
1289 | * filters or compounds, which we do not understand. Caller should | |||
1290 | * make aliasReturn empty before calling. | |||
1291 | * | |||
1292 | * The entry object is assumed to reside in the dynamic store. It may be | |||
1293 | * modified. | |||
1294 | */ | |||
1295 | Transliterator* TransliteratorRegistry::instantiateEntry(const UnicodeString& ID, | |||
1296 | TransliteratorEntry *entry, | |||
1297 | TransliteratorAlias* &aliasReturn, | |||
1298 | UErrorCode& status) { | |||
1299 | Transliterator *t = 0; | |||
1300 | U_ASSERT(aliasReturn == 0)(void)0; | |||
1301 | ||||
1302 | switch (entry->entryType) { | |||
1303 | case TransliteratorEntry::RBT_DATA: | |||
1304 | t = new RuleBasedTransliterator(ID, entry->u.data); | |||
1305 | if (t == 0) { | |||
1306 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1307 | } | |||
1308 | return t; | |||
1309 | case TransliteratorEntry::PROTOTYPE: | |||
1310 | t = entry->u.prototype->clone(); | |||
1311 | if (t == 0) { | |||
1312 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1313 | } | |||
1314 | return t; | |||
1315 | case TransliteratorEntry::ALIAS: | |||
1316 | aliasReturn = new TransliteratorAlias(entry->stringArg, entry->compoundFilter); | |||
1317 | if (aliasReturn == 0) { | |||
1318 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1319 | } | |||
1320 | return 0; | |||
1321 | case TransliteratorEntry::FACTORY: | |||
1322 | t = entry->u.factory.function(ID, entry->u.factory.context); | |||
1323 | if (t == 0) { | |||
1324 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1325 | } | |||
1326 | return t; | |||
1327 | case TransliteratorEntry::COMPOUND_RBT: | |||
1328 | { | |||
1329 | UVector* rbts = new UVector(uprv_deleteUObjectuprv_deleteUObject_71, nullptr, entry->u.dataVector->size(), status); | |||
1330 | // Check for null pointer | |||
1331 | if (rbts == NULL__null) { | |||
1332 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1333 | return NULL__null; | |||
1334 | } | |||
1335 | int32_t passNumber = 1; | |||
1336 | for (int32_t i = 0; U_SUCCESS(status) && i < entry->u.dataVector->size(); i++) { | |||
1337 | // TODO: Should passNumber be turned into a decimal-string representation (1 -> "1")? | |||
1338 | Transliterator* tl = new RuleBasedTransliterator(UnicodeString(CompoundTransliterator::PASS_STRING) + UnicodeString(passNumber++), | |||
1339 | (TransliterationRuleData*)(entry->u.dataVector->elementAt(i)), FALSE0); | |||
1340 | if (tl == 0) | |||
1341 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1342 | else | |||
1343 | rbts->adoptElement(tl, status); | |||
1344 | } | |||
1345 | if (U_FAILURE(status)) { | |||
1346 | delete rbts; | |||
1347 | return 0; | |||
1348 | } | |||
1349 | rbts->setDeleter(nullptr); | |||
1350 | aliasReturn = new TransliteratorAlias(ID, entry->stringArg, rbts, entry->compoundFilter); | |||
1351 | } | |||
1352 | if (aliasReturn == 0) { | |||
1353 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1354 | } | |||
1355 | return 0; | |||
1356 | case TransliteratorEntry::LOCALE_RULES: | |||
1357 | aliasReturn = new TransliteratorAlias(ID, entry->stringArg, | |||
1358 | (UTransDirection) entry->intArg); | |||
1359 | if (aliasReturn == 0) { | |||
1360 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1361 | } | |||
1362 | return 0; | |||
1363 | case TransliteratorEntry::RULES_FORWARD: | |||
1364 | case TransliteratorEntry::RULES_REVERSE: | |||
1365 | // Process the rule data into a TransliteratorRuleData object, | |||
1366 | // and possibly also into an ::id header and/or footer. Then | |||
1367 | // we modify the registry with the parsed data and retry. | |||
1368 | { | |||
1369 | TransliteratorParser parser(status); | |||
1370 | ||||
1371 | // We use the file name, taken from another resource bundle | |||
1372 | // 2-d array at static init time, as a locale language. We're | |||
1373 | // just using the locale mechanism to map through to a file | |||
1374 | // name; this in no way represents an actual locale. | |||
1375 | //CharString ch(entry->stringArg); | |||
1376 | //UResourceBundle *bundle = ures_openDirect(0, ch, &status); | |||
1377 | UnicodeString rules = entry->stringArg; | |||
1378 | //ures_close(bundle); | |||
1379 | ||||
1380 | //if (U_FAILURE(status)) { | |||
1381 | // We have a failure of some kind. Remove the ID from the | |||
1382 | // registry so we don't keep trying. NOTE: This will throw off | |||
1383 | // anyone who is, at the moment, trying to iterate over the | |||
1384 | // available IDs. That's acceptable since we should never | |||
1385 | // really get here except under installation, configuration, | |||
1386 | // or unrecoverable run time memory failures. | |||
1387 | // remove(ID); | |||
1388 | //} else { | |||
1389 | ||||
1390 | // If the status indicates a failure, then we don't have any | |||
1391 | // rules -- there is probably an installation error. The list | |||
1392 | // in the root locale should correspond to all the installed | |||
1393 | // transliterators; if it lists something that's not | |||
1394 | // installed, we'll get an error from ResourceBundle. | |||
1395 | aliasReturn = new TransliteratorAlias(ID, rules, | |||
1396 | ((entry->entryType == TransliteratorEntry::RULES_REVERSE) ? | |||
1397 | UTRANS_REVERSE : UTRANS_FORWARD)); | |||
1398 | if (aliasReturn == 0) { | |||
1399 | status = U_MEMORY_ALLOCATION_ERROR; | |||
1400 | } | |||
1401 | //} | |||
1402 | } | |||
1403 | return 0; | |||
1404 | default: | |||
1405 | UPRV_UNREACHABLE_EXITabort(); // can't get here | |||
1406 | } | |||
1407 | } | |||
1408 | U_NAMESPACE_END} | |||
1409 | ||||
1410 | #endif /* #if !UCONFIG_NO_TRANSLITERATION */ | |||
1411 | ||||
1412 | //eof |