]> git.saurik.com Git - apple/icu.git/blame - icuSources/common/brkiter.cpp
ICU-8.11.1.tar.gz
[apple/icu.git] / icuSources / common / brkiter.cpp
CommitLineData
b75a7d8f
A
1/*
2*******************************************************************************
73c04bcf 3* Copyright (C) 1997-2006, International Business Machines Corporation and *
b75a7d8f
A
4* others. All Rights Reserved. *
5*******************************************************************************
6*
7* File TXTBDRY.CPP
8*
9* Modification History:
10*
11* Date Name Description
12* 02/18/97 aliu Converted from OpenClass. Added DONE.
13* 01/13/2000 helena Added UErrorCode parameter to createXXXInstance methods.
14*****************************************************************************************
15*/
16
17// *****************************************************************************
18// This file was generated from the java source file BreakIterator.java
19// *****************************************************************************
20
21#include "unicode/utypes.h"
22
23#if !UCONFIG_NO_BREAK_ITERATION
24
73c04bcf 25#include "unicode/rbbi.h"
b75a7d8f
A
26#include "unicode/brkiter.h"
27#include "unicode/udata.h"
374ca955 28#include "unicode/ures.h"
73c04bcf 29#include "unicode/ustring.h"
374ca955 30#include "ucln_cmn.h"
b75a7d8f
A
31#include "cstring.h"
32#include "mutex.h"
73c04bcf 33#include "servloc.h"
374ca955
A
34#include "locbased.h"
35#include "uresimp.h"
73c04bcf
A
36#include "uassert.h"
37#include "ubrkimpl.h"
b75a7d8f
A
38
39// *****************************************************************************
40// class BreakIterator
41// This class implements methods for finding the location of boundaries in text.
42// Instances of BreakIterator maintain a current position and scan over text
43// returning the index of characters where boundaries occur.
44// *****************************************************************************
45
46U_NAMESPACE_BEGIN
47
b75a7d8f
A
48// -------------------------------------
49
b75a7d8f 50BreakIterator*
73c04bcf 51BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind, UErrorCode &status)
b75a7d8f 52{
374ca955 53 char fnbuff[256];
73c04bcf 54 char ext[4]={'\0'};
374ca955
A
55 char actualLocale[ULOC_FULLNAME_CAPACITY];
56 int32_t size;
57 const UChar* brkfname = NULL;
73c04bcf
A
58 UResourceBundle brkRulesStack;
59 UResourceBundle brkNameStack;
60 UResourceBundle *brkRules = &brkRulesStack;
61 UResourceBundle *brkName = &brkNameStack;
62 RuleBasedBreakIterator *result = NULL;
374ca955 63
b75a7d8f
A
64 if (U_FAILURE(status))
65 return NULL;
66
374ca955
A
67 ures_initStackObject(brkRules);
68 ures_initStackObject(brkName);
69
70 // Get the locale
73c04bcf
A
71 UResourceBundle *b = ures_open(U_ICUDATA_BRKITR, loc.getName(), &status);
72 /* this is a hack for now. Should be fixed when the data is fetched from
73 brk_index.txt */
74 if(status==U_USING_DEFAULT_WARNING){
75 status=U_ZERO_ERROR;
76 ures_openFillIn(b, U_ICUDATA_BRKITR, "", &status);
77 }
374ca955
A
78
79 // Get the "boundaries" array.
80 if (U_SUCCESS(status)) {
81 brkRules = ures_getByKeyWithFallback(b, "boundaries", brkRules, &status);
82 // Get the string object naming the rules file
83 brkName = ures_getByKeyWithFallback(brkRules, type, brkName, &status);
84 // Get the actual string
85 brkfname = ures_getString(brkName, &size, &status);
73c04bcf
A
86 U_ASSERT((size_t)size<sizeof(fnbuff));
87 if ((size_t)size>=sizeof(fnbuff)) {
88 size=0;
89 if (U_SUCCESS(status)) {
90 status = U_BUFFER_OVERFLOW_ERROR;
91 }
92 }
374ca955
A
93
94 // Use the string if we found it
95 if (U_SUCCESS(status) && brkfname) {
96 uprv_strncpy(actualLocale,
97 ures_getLocale(brkName, &status),
98 sizeof(actualLocale)/sizeof(actualLocale[0]));
73c04bcf
A
99
100 UChar* extStart=u_strchr(brkfname, 0x002e);
101 int len = 0;
102 if(extStart!=NULL){
103 len = extStart-brkfname;
104 u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
105 u_UCharsToChars(brkfname, fnbuff, len);
106 }
107 fnbuff[len]=0; // nul terminate
374ca955 108 }
b75a7d8f
A
109 }
110
374ca955
A
111 ures_close(brkRules);
112 ures_close(brkName);
113
73c04bcf 114 UDataMemory* file = udata_open(U_ICUDATA_BRKITR, ext, fnbuff, &status);
b75a7d8f 115 if (U_FAILURE(status)) {
374ca955 116 ures_close(b);
b75a7d8f
A
117 return NULL;
118 }
b75a7d8f 119
73c04bcf
A
120 // Create a RuleBasedBreakIterator
121 result = new RuleBasedBreakIterator(file, status);
374ca955 122
73c04bcf 123 // If there is a result, set the valid locale and actual locale, and the kind
374ca955
A
124 if (U_SUCCESS(status) && result != NULL) {
125 U_LOCALE_BASED(locBased, *result);
126 locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), actualLocale);
73c04bcf 127 result->setBreakType(kind);
b75a7d8f 128 }
374ca955
A
129
130 ures_close(b);
131
132 if (U_FAILURE(status) && result != NULL) { // Sometimes redundant check, but simple
b75a7d8f 133 delete result;
374ca955
A
134 return NULL;
135 }
136
137 if (result == NULL) {
138 udata_close(file);
139 if (U_SUCCESS(status)) {
140 status = U_MEMORY_ALLOCATION_ERROR;
141 }
b75a7d8f
A
142 }
143
144 return result;
145}
146
374ca955
A
147// Creates a break iterator for word breaks.
148BreakIterator* U_EXPORT2
149BreakIterator::createWordInstance(const Locale& key, UErrorCode& status)
150{
151 return createInstance(key, UBRK_WORD, status);
152}
153
b75a7d8f
A
154// -------------------------------------
155
156// Creates a break iterator for line breaks.
374ca955 157BreakIterator* U_EXPORT2
b75a7d8f
A
158BreakIterator::createLineInstance(const Locale& key, UErrorCode& status)
159{
374ca955 160 return createInstance(key, UBRK_LINE, status);
b75a7d8f
A
161}
162
163// -------------------------------------
164
165// Creates a break iterator for character breaks.
374ca955 166BreakIterator* U_EXPORT2
b75a7d8f
A
167BreakIterator::createCharacterInstance(const Locale& key, UErrorCode& status)
168{
374ca955 169 return createInstance(key, UBRK_CHARACTER, status);
b75a7d8f
A
170}
171
172// -------------------------------------
173
174// Creates a break iterator for sentence breaks.
374ca955 175BreakIterator* U_EXPORT2
b75a7d8f
A
176BreakIterator::createSentenceInstance(const Locale& key, UErrorCode& status)
177{
374ca955 178 return createInstance(key, UBRK_SENTENCE, status);
b75a7d8f
A
179}
180
181// -------------------------------------
182
183// Creates a break iterator for title casing breaks.
374ca955 184BreakIterator* U_EXPORT2
b75a7d8f
A
185BreakIterator::createTitleInstance(const Locale& key, UErrorCode& status)
186{
374ca955 187 return createInstance(key, UBRK_TITLE, status);
b75a7d8f
A
188}
189
190// -------------------------------------
191
192// Gets all the available locales that has localized text boundary data.
374ca955 193const Locale* U_EXPORT2
b75a7d8f
A
194BreakIterator::getAvailableLocales(int32_t& count)
195{
196 return Locale::getAvailableLocales(count);
197}
198
199// -------------------------------------
200// Gets the objectLocale display name in the default locale language.
374ca955 201UnicodeString& U_EXPORT2
b75a7d8f
A
202BreakIterator::getDisplayName(const Locale& objectLocale,
203 UnicodeString& name)
204{
205 return objectLocale.getDisplayName(name);
206}
207
208// -------------------------------------
209// Gets the objectLocale display name in the displayLocale language.
374ca955 210UnicodeString& U_EXPORT2
b75a7d8f
A
211BreakIterator::getDisplayName(const Locale& objectLocale,
212 const Locale& displayLocale,
213 UnicodeString& name)
214{
215 return objectLocale.getDisplayName(displayLocale, name);
216}
217
218// ------------------------------------------
219//
220// Default constructor and destructor
221//
222//-------------------------------------------
223
224BreakIterator::BreakIterator()
225{
226 fBufferClone = FALSE;
374ca955 227 *validLocale = *actualLocale = 0;
b75a7d8f
A
228}
229
230BreakIterator::~BreakIterator()
231{
232}
233
234// ------------------------------------------
235//
236// Registration
237//
238//-------------------------------------------
374ca955 239#if !UCONFIG_NO_SERVICE
b75a7d8f
A
240
241static ICULocaleService* gService = NULL;
242
243// -------------------------------------
244
245class ICUBreakIteratorFactory : public ICUResourceBundleFactory {
246protected:
374ca955
A
247 virtual UObject* handleCreate(const Locale& loc, int32_t kind, const ICUService* /*service*/, UErrorCode& status) const {
248 return BreakIterator::makeInstance(loc, kind, status);
249 }
b75a7d8f
A
250};
251
252// -------------------------------------
253
254class ICUBreakIteratorService : public ICULocaleService {
255public:
374ca955
A
256 ICUBreakIteratorService()
257 : ICULocaleService(UNICODE_STRING("Break Iterator", 14))
258 {
259 UErrorCode status = U_ZERO_ERROR;
260 registerFactory(new ICUBreakIteratorFactory(), status);
261 }
262
263 virtual UObject* cloneInstance(UObject* instance) const {
264 return ((BreakIterator*)instance)->clone();
265 }
266
267 virtual UObject* handleDefault(const ICUServiceKey& key, UnicodeString* /*actualID*/, UErrorCode& status) const {
268 LocaleKey& lkey = (LocaleKey&)key;
269 int32_t kind = lkey.kind();
270 Locale loc;
271 lkey.currentLocale(loc);
272 return BreakIterator::makeInstance(loc, kind, status);
273 }
274
275 virtual UBool isDefault() const {
276 return countFactories() == 1;
277 }
b75a7d8f
A
278};
279
280// -------------------------------------
281
374ca955
A
282U_NAMESPACE_END
283
284// defined in ucln_cmn.h
285
286/**
287 * Release all static memory held by breakiterator.
288 */
289U_CDECL_BEGIN
290static UBool U_CALLCONV breakiterator_cleanup(void) {
291#if !UCONFIG_NO_SERVICE
292 if (gService) {
293 delete gService;
294 gService = NULL;
295 }
296#endif
297 return TRUE;
298}
299U_CDECL_END
300U_NAMESPACE_BEGIN
301
b75a7d8f
A
302static ICULocaleService*
303getService(void)
304{
305 UBool needsInit;
306 umtx_lock(NULL);
307 needsInit = (UBool)(gService == NULL);
308 umtx_unlock(NULL);
309
310 if (needsInit) {
311 ICULocaleService *tService = new ICUBreakIteratorService();
312 umtx_lock(NULL);
313 if (gService == NULL) {
314 gService = tService;
315 tService = NULL;
374ca955 316 ucln_common_registerCleanup(UCLN_COMMON_BREAKITERATOR, breakiterator_cleanup);
b75a7d8f
A
317 }
318 umtx_unlock(NULL);
319 delete tService;
320 }
321 return gService;
322}
323
324// -------------------------------------
325
326static UBool
327hasService(void)
328{
374ca955
A
329 Mutex mutex;
330 return gService != NULL;
b75a7d8f
A
331}
332
333// -------------------------------------
334
374ca955 335URegistryKey U_EXPORT2
b75a7d8f
A
336BreakIterator::registerInstance(BreakIterator* toAdopt, const Locale& locale, UBreakIteratorType kind, UErrorCode& status)
337{
374ca955 338 return getService()->registerInstance(toAdopt, locale, kind, status);
b75a7d8f
A
339}
340
341// -------------------------------------
342
374ca955 343UBool U_EXPORT2
b75a7d8f
A
344BreakIterator::unregister(URegistryKey key, UErrorCode& status)
345{
346 if (U_SUCCESS(status)) {
347 if (hasService()) {
348 return gService->unregister(key, status);
349 }
350 status = U_ILLEGAL_ARGUMENT_ERROR;
351 }
352 return FALSE;
353}
354
355// -------------------------------------
356
374ca955 357StringEnumeration* U_EXPORT2
b75a7d8f
A
358BreakIterator::getAvailableLocales(void)
359{
360 return getService()->getAvailableLocales();
361}
374ca955
A
362#endif /* UCONFIG_NO_SERVICE */
363
364// -------------------------------------
365
366BreakIterator*
73c04bcf 367BreakIterator::createInstance(const Locale& loc, int32_t kind, UErrorCode& status)
374ca955
A
368{
369 if (U_FAILURE(status)) {
370 return NULL;
371 }
372
373 u_init(&status);
374#if !UCONFIG_NO_SERVICE
375 if (hasService()) {
73c04bcf 376 Locale actualLoc("");
374ca955
A
377 BreakIterator *result = (BreakIterator*)gService->get(loc, kind, &actualLoc, status);
378 // TODO: The way the service code works in ICU 2.8 is that if
379 // there is a real registered break iterator, the actualLoc
380 // will be populated, but if the handleDefault path is taken
381 // (because nothing is registered that can handle the
382 // requested locale) then the actualLoc comes back empty. In
383 // that case, the returned object already has its actual/valid
384 // locale data populated (by makeInstance, which is what
385 // handleDefault calls), so we don't touch it. YES, A COMMENT
386 // THIS LONG is a sign of bad code -- so the action item is to
387 // revisit this in ICU 3.0 and clean it up/fix it/remove it.
388 if (U_SUCCESS(status) && (result != NULL) && *actualLoc.getName() != 0) {
389 U_LOCALE_BASED(locBased, *result);
390 locBased.setLocaleIDs(actualLoc.getName(), actualLoc.getName());
391 }
392 return result;
393 }
394 else
395#endif
396 {
397 return makeInstance(loc, kind, status);
398 }
399}
b75a7d8f
A
400
401// -------------------------------------
402
403BreakIterator*
404BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
405{
374ca955
A
406
407 if (U_FAILURE(status)) {
408 return NULL;
409 }
410
411 BreakIterator *result = NULL;
b75a7d8f 412 switch (kind) {
374ca955 413 case UBRK_CHARACTER:
73c04bcf 414 result = BreakIterator::buildInstance(loc, "grapheme", kind, status);
374ca955
A
415 break;
416 case UBRK_WORD:
73c04bcf 417 result = BreakIterator::buildInstance(loc, "word", kind, status);
374ca955
A
418 break;
419 case UBRK_LINE:
73c04bcf 420 result = BreakIterator::buildInstance(loc, "line", kind, status);
374ca955
A
421 break;
422 case UBRK_SENTENCE:
73c04bcf 423 result = BreakIterator::buildInstance(loc, "sentence", kind, status);
374ca955
A
424 break;
425 case UBRK_TITLE:
73c04bcf 426 result = BreakIterator::buildInstance(loc, "title", kind, status);
374ca955 427 break;
b75a7d8f 428 default:
374ca955 429 status = U_ILLEGAL_ARGUMENT_ERROR;
b75a7d8f 430 }
b75a7d8f 431
374ca955
A
432 if (U_FAILURE(status)) {
433 return NULL;
434 }
b75a7d8f 435
374ca955
A
436 return result;
437}
b75a7d8f 438
374ca955
A
439Locale
440BreakIterator::getLocale(ULocDataLocaleType type, UErrorCode& status) const {
441 U_LOCALE_BASED(locBased, *this);
442 return locBased.getLocale(type, status);
443}
444
445const char *
446BreakIterator::getLocaleID(ULocDataLocaleType type, UErrorCode& status) const {
447 U_LOCALE_BASED(locBased, *this);
448 return locBased.getLocaleID(type, status);
b75a7d8f
A
449}
450
374ca955
A
451U_NAMESPACE_END
452
b75a7d8f
A
453#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
454
455//eof