Version 6.1.4.1, tag libreoffice-6.1.4.1
[LibreOffice.git] / i18npool / source / breakiterator / breakiteratorImpl.cxx
blobe923fbd99eaa5ac3e45182b51613b3e8e1f0ba75
1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
2 /*
3 * This file is part of the LibreOffice project.
5 * This Source Code Form is subject to the terms of the Mozilla Public
6 * License, v. 2.0. If a copy of the MPL was not distributed with this
7 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
9 * This file incorporates work covered by the following license notice:
11 * Licensed to the Apache Software Foundation (ASF) under one or more
12 * contributor license agreements. See the NOTICE file distributed
13 * with this work for additional information regarding copyright
14 * ownership. The ASF licenses this file to you under the Apache
15 * License, Version 2.0 (the "License"); you may not use this file
16 * except in compliance with the License. You may obtain a copy of
17 * the License at http://www.apache.org/licenses/LICENSE-2.0 .
19 #include <config_locales.h>
21 #include <breakiteratorImpl.hxx>
22 #include <cppuhelper/supportsservice.hxx>
23 #include <unicode/uchar.h>
24 #include <i18nutil/unicode.hxx>
25 #include <rtl/ustrbuf.hxx>
27 using namespace ::com::sun::star;
28 using namespace ::com::sun::star::uno;
29 using namespace ::com::sun::star::i18n;
30 using namespace ::com::sun::star::lang;
32 namespace i18npool {
34 BreakIteratorImpl::BreakIteratorImpl( const Reference < XComponentContext >& rxContext ) : m_xContext( rxContext )
38 BreakIteratorImpl::BreakIteratorImpl()
42 BreakIteratorImpl::~BreakIteratorImpl()
46 #define LBI getLocaleSpecificBreakIterator(rLocale)
48 sal_Int32 SAL_CALL BreakIteratorImpl::nextCharacters( const OUString& Text, sal_Int32 nStartPos,
49 const Locale &rLocale, sal_Int16 nCharacterIteratorMode, sal_Int32 nCount, sal_Int32& nDone )
51 if (nCount < 0) throw RuntimeException();
53 return LBI->nextCharacters( Text, nStartPos, rLocale, nCharacterIteratorMode, nCount, nDone);
56 sal_Int32 SAL_CALL BreakIteratorImpl::previousCharacters( const OUString& Text, sal_Int32 nStartPos,
57 const Locale& rLocale, sal_Int16 nCharacterIteratorMode, sal_Int32 nCount, sal_Int32& nDone )
59 if (nCount < 0) throw RuntimeException();
61 return LBI->previousCharacters( Text, nStartPos, rLocale, nCharacterIteratorMode, nCount, nDone);
64 #define isZWSP(c) (ch == 0x200B)
66 static sal_Int32 skipSpace(const OUString& Text, sal_Int32 nPos, sal_Int32 len, sal_Int16 rWordType, bool bDirection)
68 sal_uInt32 ch=0;
69 sal_Int32 pos=nPos;
70 switch (rWordType) {
71 case WordType::ANYWORD_IGNOREWHITESPACES:
72 if (bDirection)
73 while (nPos < len && (u_isWhitespace(ch = Text.iterateCodePoints(&pos)) || isZWSP(ch))) nPos=pos;
74 else
75 while (nPos > 0 && (u_isWhitespace(ch = Text.iterateCodePoints(&pos, -1)) || isZWSP(ch))) nPos=pos;
76 break;
77 case WordType::DICTIONARY_WORD:
78 if (bDirection)
79 while (nPos < len && (u_isWhitespace(ch = Text.iterateCodePoints(&pos)) || isZWSP(ch) ||
80 ! (ch == 0x002E || u_isalnum(ch)))) nPos=pos;
81 else
82 while (nPos > 0 && (u_isWhitespace(ch = Text.iterateCodePoints(&pos, -1)) || isZWSP(ch) ||
83 ! (ch == 0x002E || u_isalnum(ch)))) nPos=pos;
84 break;
85 case WordType::WORD_COUNT:
86 if (bDirection)
87 while (nPos < len && (u_isUWhiteSpace(ch = Text.iterateCodePoints(&pos)) || isZWSP(ch))) nPos=pos;
88 else
89 while (nPos > 0 && (u_isUWhiteSpace(ch = Text.iterateCodePoints(&pos, -1)) || isZWSP(ch))) nPos=pos;
90 break;
92 return nPos;
95 Boundary SAL_CALL BreakIteratorImpl::nextWord( const OUString& Text, sal_Int32 nStartPos,
96 const Locale& rLocale, sal_Int16 rWordType )
98 sal_Int32 len = Text.getLength();
99 if( nStartPos < 0 || len == 0 )
100 result.endPos = result.startPos = 0;
101 else if (nStartPos >= len)
102 result.endPos = result.startPos = len;
103 else {
104 result = LBI->nextWord(Text, nStartPos, rLocale, rWordType);
106 nStartPos = skipSpace(Text, result.startPos, len, rWordType, true);
108 if ( nStartPos != result.startPos) {
109 if( nStartPos >= len )
110 result.startPos = result.endPos = len;
111 else {
112 result = LBI->getWordBoundary(Text, nStartPos, rLocale, rWordType, true);
113 // i88041: avoid startPos goes back to nStartPos when switching between Latin and CJK scripts
114 if (result.startPos < nStartPos) result.startPos = nStartPos;
118 return result;
121 static inline bool isCJK( const Locale& rLocale ) {
122 return rLocale.Language == "zh" || rLocale.Language == "ja" || rLocale.Language == "ko";
125 Boundary SAL_CALL BreakIteratorImpl::previousWord( const OUString& Text, sal_Int32 nStartPos,
126 const Locale& rLocale, sal_Int16 rWordType)
128 sal_Int32 len = Text.getLength();
129 if( nStartPos <= 0 || len == 0 ) {
130 result.endPos = result.startPos = 0;
131 return result;
132 } else if (nStartPos > len) {
133 result.endPos = result.startPos = len;
134 return result;
137 sal_Int32 nPos = skipSpace(Text, nStartPos, len, rWordType, false);
139 // if some spaces are skipped, and the script type is Asian with no CJK rLocale, we have to return
140 // (nStartPos, -1) for caller to send correct rLocale for loading correct dictionary.
141 result.startPos = nPos;
142 if (nPos != nStartPos && nPos > 0 && !isCJK(rLocale) && getScriptClass(Text.iterateCodePoints(&nPos, -1)) == ScriptType::ASIAN) {
143 result.endPos = -1;
144 return result;
147 return LBI->previousWord(Text, result.startPos, rLocale, rWordType);
151 Boundary SAL_CALL BreakIteratorImpl::getWordBoundary( const OUString& Text, sal_Int32 nPos, const Locale& rLocale,
152 sal_Int16 rWordType, sal_Bool bDirection )
154 sal_Int32 len = Text.getLength();
155 if( nPos < 0 || len == 0 )
156 result.endPos = result.startPos = 0;
157 else if (nPos > len)
158 result.endPos = result.startPos = len;
159 else {
160 sal_Int32 next, prev;
161 next = skipSpace(Text, nPos, len, rWordType, true);
162 prev = skipSpace(Text, nPos, len, rWordType, false);
163 if (prev == 0 && next == len) {
164 result.endPos = result.startPos = nPos;
165 } else if (prev == 0 && ! bDirection) {
166 result.endPos = result.startPos = 0;
167 } else if (next == len && bDirection) {
168 result.endPos = result.startPos = len;
169 } else {
170 if (next != prev) {
171 if (next == nPos && next != len)
172 bDirection = true;
173 else if (prev == nPos && prev != 0)
174 bDirection = false;
175 else
176 nPos = bDirection ? next : prev;
178 result = LBI->getWordBoundary(Text, nPos, rLocale, rWordType, bDirection);
181 return result;
184 sal_Bool SAL_CALL BreakIteratorImpl::isBeginWord( const OUString& Text, sal_Int32 nPos,
185 const Locale& rLocale, sal_Int16 rWordType )
187 sal_Int32 len = Text.getLength();
189 if (nPos < 0 || nPos >= len) return false;
191 sal_Int32 tmp = skipSpace(Text, nPos, len, rWordType, true);
193 if (tmp != nPos) return false;
195 result = getWordBoundary(Text, nPos, rLocale, rWordType, true);
197 return result.startPos == nPos;
200 sal_Bool SAL_CALL BreakIteratorImpl::isEndWord( const OUString& Text, sal_Int32 nPos,
201 const Locale& rLocale, sal_Int16 rWordType )
203 sal_Int32 len = Text.getLength();
205 if (nPos <= 0 || nPos > len) return false;
207 sal_Int32 tmp = skipSpace(Text, nPos, len, rWordType, false);
209 if (tmp != nPos) return false;
211 result = getWordBoundary(Text, nPos, rLocale, rWordType, false);
213 return result.endPos == nPos;
216 sal_Int32 SAL_CALL BreakIteratorImpl::beginOfSentence( const OUString& Text, sal_Int32 nStartPos,
217 const Locale &rLocale )
219 if (nStartPos < 0 || nStartPos > Text.getLength())
220 return -1;
221 if (Text.isEmpty()) return 0;
222 return LBI->beginOfSentence(Text, nStartPos, rLocale);
225 sal_Int32 SAL_CALL BreakIteratorImpl::endOfSentence( const OUString& Text, sal_Int32 nStartPos,
226 const Locale &rLocale )
228 if (nStartPos < 0 || nStartPos > Text.getLength())
229 return -1;
230 if (Text.isEmpty()) return 0;
231 return LBI->endOfSentence(Text, nStartPos, rLocale);
234 LineBreakResults SAL_CALL BreakIteratorImpl::getLineBreak( const OUString& Text, sal_Int32 nStartPos,
235 const Locale& rLocale, sal_Int32 nMinBreakPos, const LineBreakHyphenationOptions& hOptions,
236 const LineBreakUserOptions& bOptions )
238 return LBI->getLineBreak(Text, nStartPos, rLocale, nMinBreakPos, hOptions, bOptions);
241 sal_Int16 SAL_CALL BreakIteratorImpl::getScriptType( const OUString& Text, sal_Int32 nPos )
243 return (nPos < 0 || nPos >= Text.getLength()) ? ScriptType::WEAK :
244 getScriptClass(Text.iterateCodePoints(&nPos, 0));
248 /** Increments/decrements position first, then obtains character.
249 @return current position, may be -1 or text length if string was consumed.
251 static sal_Int32 iterateCodePoints(const OUString& Text, sal_Int32 &nStartPos, sal_Int32 inc, sal_uInt32& ch) {
252 sal_Int32 nLen = Text.getLength();
253 if (nStartPos + inc < 0 || nStartPos + inc >= nLen) {
254 ch = 0;
255 nStartPos = nStartPos + inc < 0 ? -1 : nLen;
256 } else {
257 ch = Text.iterateCodePoints(&nStartPos, inc);
258 // Fix for #i80436#.
259 // erAck: 2009-06-30T21:52+0200 This logic looks somewhat
260 // suspicious as if it cures a symptom.. anyway, had to add
261 // nStartPos < Text.getLength() to silence the (correct) assertion
262 // in rtl_uString_iterateCodePoints() if Text was one character
263 // (codepoint) only, made up of a surrogate pair.
264 //if (inc > 0 && nStartPos < Text.getLength())
265 // ch = Text.iterateCodePoints(&nStartPos, 0);
266 // With surrogates, nStartPos may actually point behind string
267 // now, even if inc is only +1
268 if (inc > 0)
269 ch = (nStartPos < nLen ? Text.iterateCodePoints(&nStartPos, 0) : 0);
271 return nStartPos;
275 sal_Int32 SAL_CALL BreakIteratorImpl::beginOfScript( const OUString& Text,
276 sal_Int32 nStartPos, sal_Int16 ScriptType )
278 if (nStartPos < 0 || nStartPos >= Text.getLength())
279 return -1;
281 if(ScriptType != getScriptClass(Text.iterateCodePoints(&nStartPos, 0)))
282 return -1;
284 if (nStartPos == 0) return 0;
285 sal_uInt32 ch=0;
286 while (iterateCodePoints(Text, nStartPos, -1, ch) >= 0 && ScriptType == getScriptClass(ch)) {
287 if (nStartPos == 0) return 0;
290 return iterateCodePoints(Text, nStartPos, 1, ch);
293 sal_Int32 SAL_CALL BreakIteratorImpl::endOfScript( const OUString& Text,
294 sal_Int32 nStartPos, sal_Int16 ScriptType )
296 if (nStartPos < 0 || nStartPos >= Text.getLength())
297 return -1;
299 if(ScriptType != getScriptClass(Text.iterateCodePoints(&nStartPos, 0)))
300 return -1;
302 sal_Int32 strLen = Text.getLength();
303 sal_uInt32 ch=0;
304 while(iterateCodePoints(Text, nStartPos, 1, ch) < strLen ) {
305 sal_Int16 currentCharScriptType = getScriptClass(ch);
306 if(ScriptType != currentCharScriptType && currentCharScriptType != ScriptType::WEAK)
307 break;
309 return nStartPos;
312 sal_Int32 SAL_CALL BreakIteratorImpl::previousScript( const OUString& Text,
313 sal_Int32 nStartPos, sal_Int16 ScriptType )
315 if (nStartPos < 0)
316 return -1;
317 if (nStartPos > Text.getLength())
318 nStartPos = Text.getLength();
320 sal_Int16 numberOfChange = (ScriptType == getScriptClass(Text.iterateCodePoints(&nStartPos, 0))) ? 3 : 2;
322 sal_uInt32 ch=0;
323 while (numberOfChange > 0 && iterateCodePoints(Text, nStartPos, -1, ch) >= 0) {
324 if (((numberOfChange % 2) == 0) != (ScriptType != getScriptClass(ch)))
325 numberOfChange--;
326 else if (nStartPos == 0) {
327 return -1;
330 return numberOfChange == 0 ? iterateCodePoints(Text, nStartPos, 1, ch) : -1;
333 sal_Int32 SAL_CALL BreakIteratorImpl::nextScript( const OUString& Text, sal_Int32 nStartPos,
334 sal_Int16 ScriptType )
337 if (nStartPos < 0)
338 nStartPos = 0;
339 sal_Int32 strLen = Text.getLength();
340 if (nStartPos >= strLen)
341 return -1;
343 sal_Int16 numberOfChange = (ScriptType == getScriptClass(Text.iterateCodePoints(&nStartPos, 0))) ? 2 : 1;
345 sal_uInt32 ch=0;
346 while (numberOfChange > 0 && iterateCodePoints(Text, nStartPos, 1, ch) < strLen) {
347 sal_Int16 currentCharScriptType = getScriptClass(ch);
348 if ((numberOfChange == 1) ? (ScriptType == currentCharScriptType) :
349 (ScriptType != currentCharScriptType && currentCharScriptType != ScriptType::WEAK))
350 numberOfChange--;
352 return numberOfChange == 0 ? nStartPos : -1;
355 sal_Int32 SAL_CALL BreakIteratorImpl::beginOfCharBlock( const OUString& Text, sal_Int32 nStartPos,
356 const Locale& /*rLocale*/, sal_Int16 CharType )
358 if (CharType == CharType::ANY_CHAR) return 0;
359 if (nStartPos < 0 || nStartPos >= Text.getLength()) return -1;
360 if (CharType != static_cast<sal_Int16>(u_charType( Text.iterateCodePoints(&nStartPos, 0)))) return -1;
362 sal_Int32 nPos=nStartPos;
363 while(nStartPos > 0 && CharType == static_cast<sal_Int16>(u_charType(Text.iterateCodePoints(&nPos, -1)))) { nStartPos=nPos; }
364 return nStartPos; // begin of char block is inclusive
367 sal_Int32 SAL_CALL BreakIteratorImpl::endOfCharBlock( const OUString& Text, sal_Int32 nStartPos,
368 const Locale& /*rLocale*/, sal_Int16 CharType )
370 sal_Int32 strLen = Text.getLength();
372 if (CharType == CharType::ANY_CHAR) return strLen; // end of char block is exclusive
373 if (nStartPos < 0 || nStartPos >= strLen) return -1;
374 if (CharType != static_cast<sal_Int16>(u_charType(Text.iterateCodePoints(&nStartPos, 0)))) return -1;
376 sal_uInt32 ch=0;
377 while(iterateCodePoints(Text, nStartPos, 1, ch) < strLen && CharType == static_cast<sal_Int16>(u_charType(ch))) {}
378 return nStartPos; // end of char block is exclusive
381 sal_Int32 SAL_CALL BreakIteratorImpl::nextCharBlock( const OUString& Text, sal_Int32 nStartPos,
382 const Locale& /*rLocale*/, sal_Int16 CharType )
384 if (CharType == CharType::ANY_CHAR) return -1;
385 if (nStartPos < 0 || nStartPos >= Text.getLength()) return -1;
387 sal_Int16 numberOfChange = (CharType == static_cast<sal_Int16>(u_charType(Text.iterateCodePoints(&nStartPos, 0)))) ? 2 : 1;
388 sal_Int32 strLen = Text.getLength();
390 sal_uInt32 ch=0;
391 while (numberOfChange > 0 && iterateCodePoints(Text, nStartPos, 1, ch) < strLen) {
392 if ((CharType != static_cast<sal_Int16>(u_charType(ch))) != (numberOfChange == 1))
393 numberOfChange--;
395 return numberOfChange == 0 ? nStartPos : -1;
398 sal_Int32 SAL_CALL BreakIteratorImpl::previousCharBlock( const OUString& Text, sal_Int32 nStartPos,
399 const Locale& /*rLocale*/, sal_Int16 CharType )
401 if(CharType == CharType::ANY_CHAR) return -1;
402 if (nStartPos < 0 || nStartPos >= Text.getLength()) return -1;
404 sal_Int16 numberOfChange = (CharType == static_cast<sal_Int16>(u_charType(Text.iterateCodePoints(&nStartPos, 0)))) ? 3 : 2;
406 sal_uInt32 ch=0;
407 while (numberOfChange > 0 && iterateCodePoints(Text, nStartPos, -1, ch) >= 0) {
408 if (((numberOfChange % 2) == 0) != (CharType != static_cast<sal_Int16>(u_charType(ch))))
409 numberOfChange--;
410 if (nStartPos == 0 && numberOfChange > 0) {
411 numberOfChange--;
412 if (numberOfChange == 0) return nStartPos;
415 return numberOfChange == 0 ? iterateCodePoints(Text, nStartPos, 1, ch) : -1;
419 sal_Int16 SAL_CALL BreakIteratorImpl::getWordType( const OUString& /*Text*/,
420 sal_Int32 /*nPos*/, const Locale& /*rLocale*/ )
422 return 0;
425 namespace
427 sal_Int16 getScriptClassByUAX24Script(sal_uInt32 currentChar)
429 int32_t script = u_getIntPropertyValue(currentChar, UCHAR_SCRIPT);
430 return unicode::getScriptClassFromUScriptCode(static_cast<UScriptCode>(script));
433 struct UBlock2Script
435 UBlockCode from;
436 UBlockCode to;
437 sal_Int16 script;
440 static const UBlock2Script scriptList[] =
442 {UBLOCK_NO_BLOCK, UBLOCK_NO_BLOCK, ScriptType::WEAK},
443 {UBLOCK_BASIC_LATIN, UBLOCK_SPACING_MODIFIER_LETTERS, ScriptType::LATIN},
444 {UBLOCK_GREEK, UBLOCK_ARMENIAN, ScriptType::LATIN},
445 {UBLOCK_HEBREW, UBLOCK_MYANMAR, ScriptType::COMPLEX},
446 {UBLOCK_GEORGIAN, UBLOCK_GEORGIAN, ScriptType::LATIN},
447 {UBLOCK_HANGUL_JAMO, UBLOCK_HANGUL_JAMO, ScriptType::ASIAN},
448 {UBLOCK_ETHIOPIC, UBLOCK_ETHIOPIC, ScriptType::COMPLEX},
449 {UBLOCK_CHEROKEE, UBLOCK_RUNIC, ScriptType::LATIN},
450 {UBLOCK_KHMER, UBLOCK_MONGOLIAN, ScriptType::COMPLEX},
451 {UBLOCK_LATIN_EXTENDED_ADDITIONAL, UBLOCK_GREEK_EXTENDED, ScriptType::LATIN},
452 {UBLOCK_NUMBER_FORMS, UBLOCK_NUMBER_FORMS, ScriptType::WEAK},
453 {UBLOCK_CJK_RADICALS_SUPPLEMENT, UBLOCK_HANGUL_SYLLABLES, ScriptType::ASIAN},
454 {UBLOCK_CJK_COMPATIBILITY_IDEOGRAPHS, UBLOCK_CJK_COMPATIBILITY_IDEOGRAPHS, ScriptType::ASIAN},
455 {UBLOCK_ARABIC_PRESENTATION_FORMS_A, UBLOCK_ARABIC_PRESENTATION_FORMS_A, ScriptType::COMPLEX},
456 {UBLOCK_CJK_COMPATIBILITY_FORMS, UBLOCK_CJK_COMPATIBILITY_FORMS, ScriptType::ASIAN},
457 {UBLOCK_ARABIC_PRESENTATION_FORMS_B, UBLOCK_ARABIC_PRESENTATION_FORMS_B, ScriptType::COMPLEX},
458 {UBLOCK_HALFWIDTH_AND_FULLWIDTH_FORMS, UBLOCK_HALFWIDTH_AND_FULLWIDTH_FORMS, ScriptType::ASIAN},
459 {UBLOCK_CJK_UNIFIED_IDEOGRAPHS_EXTENSION_B, UBLOCK_CJK_COMPATIBILITY_IDEOGRAPHS_SUPPLEMENT, ScriptType::ASIAN},
460 {UBLOCK_CJK_STROKES, UBLOCK_CJK_STROKES, ScriptType::ASIAN},
461 {UBLOCK_LATIN_EXTENDED_C, UBLOCK_LATIN_EXTENDED_D, ScriptType::LATIN}
464 #define scriptListCount SAL_N_ELEMENTS(scriptList)
466 //always sets rScriptType
468 //returns true for characters historically explicitly assigned to
469 //latin/weak/asian
471 //returns false for characters that historically implicitly assigned to
472 //weak as unknown
473 bool getCompatibilityScriptClassByBlock(sal_uInt32 currentChar, sal_Int16 &rScriptType)
475 bool bKnown = true;
476 //handle specific characters always as weak:
477 // 0x01 - this breaks a word
478 // 0x02 - this can be inside a word
479 // 0x20 & 0xA0 - Bug 102975, declare western space and non-break space as WEAK char.
480 if( 0x01 == currentChar || 0x02 == currentChar || 0x20 == currentChar || 0xA0 == currentChar)
481 rScriptType = ScriptType::WEAK;
482 // Few Spacing Modifier Letters that can be Bopomofo tonal marks.
483 else if ( 0x2CA == currentChar || 0x2CB == currentChar || 0x2C7 == currentChar || 0x2D9 == currentChar )
484 rScriptType = ScriptType::WEAK;
485 // workaround for Coptic
486 else if ( 0x2C80 <= currentChar && 0x2CE3 >= currentChar)
487 rScriptType = ScriptType::LATIN;
488 else
490 UBlockCode block=ublock_getCode(currentChar);
491 size_t i = 0;
492 while (i < scriptListCount)
494 if (block <= scriptList[i].to)
495 break;
496 ++i;
498 if (i < scriptListCount && block >= scriptList[i].from)
499 rScriptType = scriptList[i].script;
500 else
502 rScriptType = ScriptType::WEAK;
503 bKnown = false;
506 return bKnown;
510 sal_Int16 BreakIteratorImpl::getScriptClass(sal_uInt32 currentChar)
512 static sal_uInt32 lastChar = 0;
513 static sal_Int16 nRet = ScriptType::WEAK;
515 if (currentChar != lastChar)
517 lastChar = currentChar;
519 if (!getCompatibilityScriptClassByBlock(currentChar, nRet))
520 nRet = getScriptClassByUAX24Script(currentChar);
523 return nRet;
526 bool BreakIteratorImpl::createLocaleSpecificBreakIterator(const OUString& aLocaleName)
528 // to share service between same Language but different Country code, like zh_CN and zh_TW
529 for (lookupTableItem& listItem : lookupTable) {
530 if (aLocaleName == listItem.aLocale.Language) {
531 xBI = listItem.xBI;
532 return true;
536 #if !WITH_LOCALE_ALL && !WITH_LOCALE_ja
537 if (aLocaleName == "ja")
538 return false;
539 #endif
540 #if !WITH_LOCALE_ALL && !WITH_LOCALE_zh
541 if (aLocaleName == "zh" || aLocaleName == "zh_TW")
542 return false;
543 #endif
544 #if !WITH_LOCALE_ALL && !WITH_LOCALE_ko
545 if (aLocaleName == "ko")
546 return false;
547 #endif
548 #if !WITH_LOCALE_ALL && !WITH_LOCALE_th
549 if (aLocaleName == "th")
550 return false;
551 #endif
553 Reference < uno::XInterface > xI = m_xContext->getServiceManager()->createInstanceWithContext(
554 "com.sun.star.i18n.BreakIterator_" + aLocaleName, m_xContext);
556 if ( xI.is() ) {
557 xBI.set(xI, UNO_QUERY);
558 if (xBI.is()) {
559 lookupTable.emplace_back(Locale(aLocaleName, aLocaleName, aLocaleName), xBI);
560 return true;
563 return false;
566 Reference < XBreakIterator >
567 BreakIteratorImpl::getLocaleSpecificBreakIterator(const Locale& rLocale)
569 if (xBI.is() && rLocale == aLocale)
570 return xBI;
571 else if (m_xContext.is()) {
572 aLocale = rLocale;
574 for (lookupTableItem& listItem : lookupTable) {
575 if (rLocale == listItem.aLocale)
576 return xBI = listItem.xBI;
579 sal_Unicode under = '_';
581 sal_Int32 l = rLocale.Language.getLength();
582 sal_Int32 c = rLocale.Country.getLength();
583 sal_Int32 v = rLocale.Variant.getLength();
584 OUStringBuffer aBuf(l+c+v+3);
586 if ((l > 0 && c > 0 && v > 0 &&
587 // load service with name <base>_<lang>_<country>_<variant>
588 createLocaleSpecificBreakIterator(aBuf.append(rLocale.Language).append(under).append(
589 rLocale.Country).append(under).append(rLocale.Variant).makeStringAndClear())) ||
590 (l > 0 && c > 0 &&
591 // load service with name <base>_<lang>_<country>
592 createLocaleSpecificBreakIterator(aBuf.append(rLocale.Language).append(under).append(
593 rLocale.Country).makeStringAndClear())) ||
594 (l > 0 && c > 0 && rLocale.Language == "zh" &&
595 (rLocale.Country == "HK" ||
596 rLocale.Country == "MO" ) &&
597 // if the country code is HK or MO, one more step to try TW.
598 createLocaleSpecificBreakIterator(aBuf.append(rLocale.Language).append(under).append(
599 "TW").makeStringAndClear())) ||
600 (l > 0 &&
601 // load service with name <base>_<lang>
602 createLocaleSpecificBreakIterator(rLocale.Language)) ||
603 // load default service with name <base>_Unicode
604 createLocaleSpecificBreakIterator("Unicode")) {
605 lookupTable.emplace_back( aLocale, xBI );
606 return xBI;
609 throw RuntimeException();
612 OUString SAL_CALL
613 BreakIteratorImpl::getImplementationName()
615 return OUString("com.sun.star.i18n.BreakIterator");
618 sal_Bool SAL_CALL
619 BreakIteratorImpl::supportsService(const OUString& rServiceName)
621 return cppu::supportsService(this, rServiceName);
624 Sequence< OUString > SAL_CALL
625 BreakIteratorImpl::getSupportedServiceNames()
627 Sequence< OUString > aRet { "com.sun.star.i18n.BreakIterator" };
628 return aRet;
633 extern "C" SAL_DLLPUBLIC_EXPORT css::uno::XInterface *
634 com_sun_star_i18n_BreakIterator_get_implementation(
635 css::uno::XComponentContext *context,
636 css::uno::Sequence<css::uno::Any> const &)
638 return cppu::acquire(new i18npool::BreakIteratorImpl(context));
641 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */