i18npool/source/breakiterator/breakiterator_cjk.cxx

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2008 by Sun Microsystems, Inc.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * $RCSfile: breakiterator_cjk.cxx,v $
  10  * $Revision: 1.17.16.1 $
  11  *
  12  * This file is part of OpenOffice.org.
  13  *
  14  * OpenOffice.org is free software: you can redistribute it and/or modify
  15  * it under the terms of the GNU Lesser General Public License version 3
  16  * only, as published by the Free Software Foundation.
  17  *
  18  * OpenOffice.org is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21  * GNU Lesser General Public License version 3 for more details
  22  * (a copy is included in the LICENSE file that accompanied this code).
  23  *
  24  * You should have received a copy of the GNU Lesser General Public License
  25  * version 3 along with OpenOffice.org.  If not, see
  26  * <http://www.openoffice.org/license.html>
  27  * for a copy of the LGPLv3 License.
  28  *
  29  ************************************************************************/
  30
  31 // MARKER(update_precomp.py): autogen include statement, do not remove
  32 #include "precompiled_i18npool.hxx"
  33
  34 #define BREAKITERATOR_ALL
  35 #include <breakiterator_cjk.hxx>
  36 #include <localedata.hxx>
  37 #include <i18nutil/unicode.hxx>
  38
  39 using namespace ::com::sun::star::uno;
  40 using namespace ::com::sun::star::lang;
  41 using namespace ::rtl;
  42
  43 namespace com { namespace sun { namespace star { namespace i18n {
  44
  45 //      ----------------------------------------------------
  46 //      class BreakIterator_CJK
  47 //      ----------------------------------------------------;
  48
  49 BreakIterator_CJK::BreakIterator_CJK() :
  50     dict( NULL ),
  51     hangingCharacters()
  52 {
  53         cBreakIterator = "com.sun.star.i18n.BreakIterator_CJK";
  54 }
  55
  56 Boundary SAL_CALL
  57 BreakIterator_CJK::previousWord(const OUString& text, sal_Int32 anyPos,
  58         const lang::Locale& nLocale, sal_Int16 wordType) throw(RuntimeException)
  59 {
  60         if (dict) {
  61             result = dict->previousWord(text, anyPos, wordType);
  62             // #109813# for non-CJK, single character word, fallback to ICU breakiterator.
  63             if (result.endPos - result.startPos != 1 ||
  64                     getScriptType(text, result.startPos) == ScriptType::ASIAN)
  65                 return result;
  66             result = BreakIterator_Unicode::getWordBoundary(text, result.startPos, nLocale, wordType, true);
  67             if (result.endPos < anyPos)
  68                 return result;
  69         }
  70         return BreakIterator_Unicode::previousWord(text, anyPos, nLocale, wordType);
  71 }
  72
  73 Boundary SAL_CALL
  74 BreakIterator_CJK::nextWord(const OUString& text, sal_Int32 anyPos,
  75         const lang::Locale& nLocale, sal_Int16 wordType) throw(RuntimeException)
  76 {
  77         if (dict) {
  78             result = dict->nextWord(text, anyPos, wordType);
  79             // #109813# for non-CJK, single character word, fallback to ICU breakiterator.
  80             if (result.endPos - result.startPos != 1 ||
  81                     getScriptType(text, result.startPos) == ScriptType::ASIAN)
  82                 return result;
  83             result = BreakIterator_Unicode::getWordBoundary(text, result.startPos, nLocale, wordType, true);
  84             if (result.startPos > anyPos)
  85                 return result;
  86         }
  87         return BreakIterator_Unicode::nextWord(text, anyPos, nLocale, wordType);
  88 }
  89
  90 Boundary SAL_CALL
  91 BreakIterator_CJK::getWordBoundary( const OUString& text, sal_Int32 anyPos,
  92         const lang::Locale& nLocale, sal_Int16 wordType, sal_Bool bDirection )
  93         throw(RuntimeException)
  94 {
  95         if (dict) {
  96             result = dict->getWordBoundary(text, anyPos, wordType, bDirection);
  97             // #109813# for non-CJK, single character word, fallback to ICU breakiterator.
  98             if (result.endPos - result.startPos != 1 ||
  99                     getScriptType(text, result.startPos) == ScriptType::ASIAN)
 100                 return result;
 101         }
 102         return BreakIterator_Unicode::getWordBoundary(text, anyPos, nLocale, wordType, bDirection);
 103 }
 104
 105 LineBreakResults SAL_CALL BreakIterator_CJK::getLineBreak(
 106         const OUString& Text, sal_Int32 nStartPos,
 107         const lang::Locale& /*rLocale*/, sal_Int32 /*nMinBreakPos*/,
 108         const LineBreakHyphenationOptions& /*hOptions*/,
 109         const LineBreakUserOptions& bOptions ) throw(RuntimeException)
 110 {
 111         LineBreakResults lbr;
 112
 113         if (bOptions.allowPunctuationOutsideMargin &&
 114                 hangingCharacters.indexOf(Text[nStartPos]) != -1 &&
 115                 (Text.iterateCodePoints( &nStartPos, 1), nStartPos == Text.getLength())) {
 116             ; // do nothing
 117         } else if (bOptions.applyForbiddenRules && 0 < nStartPos && nStartPos < Text.getLength()) {
 118             while (nStartPos > 0 &&
 119                     (bOptions.forbiddenBeginCharacters.indexOf(Text[nStartPos]) != -1 ||
 120                     bOptions.forbiddenEndCharacters.indexOf(Text[nStartPos-1]) != -1))
 121                 Text.iterateCodePoints( &nStartPos, -1);
 122         }
 123
 124         lbr.breakIndex = nStartPos;
 125         lbr.breakType = BreakType::WORDBOUNDARY;
 126         return lbr;
 127 }
 128
 129 #define LOCALE(language, country) lang::Locale(OUString::createFromAscii(language), OUString::createFromAscii(country), OUString())
 130 //      ----------------------------------------------------
 131 //      class BreakIterator_zh
 132 //      ----------------------------------------------------;
 133 BreakIterator_zh::BreakIterator_zh()
 134 {
 135         dict = new xdictionary("zh");
 136         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("zh", "CN"));
 137         cBreakIterator = "com.sun.star.i18n.BreakIterator_zh";
 138 }
 139
 140 BreakIterator_zh::~BreakIterator_zh()
 141 {
 142         delete dict;
 143 }
 144
 145 //      ----------------------------------------------------
 146 //      class BreakIterator_zh_TW
 147 //      ----------------------------------------------------;
 148 BreakIterator_zh_TW::BreakIterator_zh_TW()
 149 {
 150         dict = new xdictionary("zh");
 151         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("zh", "TW"));
 152         cBreakIterator = "com.sun.star.i18n.BreakIterator_zh_TW";
 153 }
 154
 155 BreakIterator_zh_TW::~BreakIterator_zh_TW()
 156 {
 157         delete dict;
 158 }
 159
 160 //      ----------------------------------------------------
 161 //      class BreakIterator_ja
 162 //      ----------------------------------------------------;
 163 BreakIterator_ja::BreakIterator_ja()
 164 {
 165         dict = new xdictionary("ja");
 166         dict->setJapaneseWordBreak();
 167         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("ja", "JP"));
 168         cBreakIterator = "com.sun.star.i18n.BreakIterator_ja";
 169 }
 170
 171 BreakIterator_ja::~BreakIterator_ja()
 172 {
 173         delete dict;
 174 }
 175
 176 //      ----------------------------------------------------
 177 //      class BreakIterator_ko
 178 //      ----------------------------------------------------;
 179 BreakIterator_ko::BreakIterator_ko()
 180 {
 181         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("ko", "KR"));
 182         cBreakIterator = "com.sun.star.i18n.BreakIterator_ko";
 183 }
 184
 185 BreakIterator_ko::~BreakIterator_ko()
 186 {
 187 }
 188
 189 } } } }