i18npool/source/breakiterator/breakiterator_cjk.cxx

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2000, 2010 Oracle and/or its affiliates.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * This file is part of OpenOffice.org.
  10  *
  11  * OpenOffice.org is free software: you can redistribute it and/or modify
  12  * it under the terms of the GNU Lesser General Public License version 3
  13  * only, as published by the Free Software Foundation.
  14  *
  15  * OpenOffice.org is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU Lesser General Public License version 3 for more details
  19  * (a copy is included in the LICENSE file that accompanied this code).
  20  *
  21  * You should have received a copy of the GNU Lesser General Public License
  22  * version 3 along with OpenOffice.org.  If not, see
  23  * <http://www.openoffice.org/license.html>
  24  * for a copy of the LGPLv3 License.
  25  *
  26  ************************************************************************/
  27
  28 // MARKER(update_precomp.py): autogen include statement, do not remove
  29 #include "precompiled_i18npool.hxx"
  30
  31 #define BREAKITERATOR_ALL
  32 #include <breakiterator_cjk.hxx>
  33 #include <localedata.hxx>
  34 #include <i18nutil/unicode.hxx>
  35
  36 using namespace ::com::sun::star::uno;
  37 using namespace ::com::sun::star::lang;
  38 using namespace ::rtl;
  39
  40 namespace com { namespace sun { namespace star { namespace i18n {
  41
  42 //      ----------------------------------------------------
  43 //      class BreakIterator_CJK
  44 //      ----------------------------------------------------;
  45
  46 BreakIterator_CJK::BreakIterator_CJK() :
  47     dict( NULL ),
  48     hangingCharacters()
  49 {
  50         cBreakIterator = "com.sun.star.i18n.BreakIterator_CJK";
  51 }
  52
  53 Boundary SAL_CALL
  54 BreakIterator_CJK::previousWord(const OUString& text, sal_Int32 anyPos,
  55         const lang::Locale& nLocale, sal_Int16 wordType) throw(RuntimeException)
  56 {
  57         if (dict) {
  58             result = dict->previousWord(text, anyPos, wordType);
  59             // #109813# for non-CJK, single character word, fallback to ICU breakiterator.
  60             if (result.endPos - result.startPos != 1 ||
  61                     getScriptType(text, result.startPos) == ScriptType::ASIAN)
  62                 return result;
  63             result = BreakIterator_Unicode::getWordBoundary(text, result.startPos, nLocale, wordType, true);
  64             if (result.endPos < anyPos)
  65                 return result;
  66         }
  67         return BreakIterator_Unicode::previousWord(text, anyPos, nLocale, wordType);
  68 }
  69
  70 Boundary SAL_CALL
  71 BreakIterator_CJK::nextWord(const OUString& text, sal_Int32 anyPos,
  72         const lang::Locale& nLocale, sal_Int16 wordType) throw(RuntimeException)
  73 {
  74         if (dict) {
  75             result = dict->nextWord(text, anyPos, wordType);
  76             // #109813# for non-CJK, single character word, fallback to ICU breakiterator.
  77             if (result.endPos - result.startPos != 1 ||
  78                     getScriptType(text, result.startPos) == ScriptType::ASIAN)
  79                 return result;
  80             result = BreakIterator_Unicode::getWordBoundary(text, result.startPos, nLocale, wordType, true);
  81             if (result.startPos > anyPos)
  82                 return result;
  83         }
  84         return BreakIterator_Unicode::nextWord(text, anyPos, nLocale, wordType);
  85 }
  86
  87 Boundary SAL_CALL
  88 BreakIterator_CJK::getWordBoundary( const OUString& text, sal_Int32 anyPos,
  89         const lang::Locale& nLocale, sal_Int16 wordType, sal_Bool bDirection )
  90         throw(RuntimeException)
  91 {
  92         if (dict) {
  93             result = dict->getWordBoundary(text, anyPos, wordType, bDirection);
  94             // #109813# for non-CJK, single character word, fallback to ICU breakiterator.
  95             if (result.endPos - result.startPos != 1 ||
  96                     getScriptType(text, result.startPos) == ScriptType::ASIAN)
  97                 return result;
  98         }
  99         return BreakIterator_Unicode::getWordBoundary(text, anyPos, nLocale, wordType, bDirection);
 100 }
 101
 102 LineBreakResults SAL_CALL BreakIterator_CJK::getLineBreak(
 103         const OUString& Text, sal_Int32 nStartPos,
 104         const lang::Locale& /*rLocale*/, sal_Int32 /*nMinBreakPos*/,
 105         const LineBreakHyphenationOptions& /*hOptions*/,
 106         const LineBreakUserOptions& bOptions ) throw(RuntimeException)
 107 {
 108         LineBreakResults lbr;
 109
 110         if (bOptions.allowPunctuationOutsideMargin &&
 111                 hangingCharacters.indexOf(Text[nStartPos]) != -1 &&
 112                 (Text.iterateCodePoints( &nStartPos, 1), nStartPos == Text.getLength())) {
 113             ; // do nothing
 114         } else if (bOptions.applyForbiddenRules && 0 < nStartPos && nStartPos < Text.getLength()) {
 115             while (nStartPos > 0 &&
 116                     (bOptions.forbiddenBeginCharacters.indexOf(Text[nStartPos]) != -1 ||
 117                     bOptions.forbiddenEndCharacters.indexOf(Text[nStartPos-1]) != -1))
 118                 Text.iterateCodePoints( &nStartPos, -1);
 119         }
 120
 121         lbr.breakIndex = nStartPos;
 122         lbr.breakType = BreakType::WORDBOUNDARY;
 123         return lbr;
 124 }
 125
 126 #define LOCALE(language, country) lang::Locale(OUString::createFromAscii(language), OUString::createFromAscii(country), OUString())
 127 //      ----------------------------------------------------
 128 //      class BreakIterator_zh
 129 //      ----------------------------------------------------;
 130 BreakIterator_zh::BreakIterator_zh()
 131 {
 132         dict = new xdictionary("zh");
 133         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("zh", "CN"));
 134         cBreakIterator = "com.sun.star.i18n.BreakIterator_zh";
 135 }
 136
 137 BreakIterator_zh::~BreakIterator_zh()
 138 {
 139         delete dict;
 140 }
 141
 142 //      ----------------------------------------------------
 143 //      class BreakIterator_zh_TW
 144 //      ----------------------------------------------------;
 145 BreakIterator_zh_TW::BreakIterator_zh_TW()
 146 {
 147         dict = new xdictionary("zh");
 148         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("zh", "TW"));
 149         cBreakIterator = "com.sun.star.i18n.BreakIterator_zh_TW";
 150 }
 151
 152 BreakIterator_zh_TW::~BreakIterator_zh_TW()
 153 {
 154         delete dict;
 155 }
 156
 157 //      ----------------------------------------------------
 158 //      class BreakIterator_ja
 159 //      ----------------------------------------------------;
 160 BreakIterator_ja::BreakIterator_ja()
 161 {
 162         dict = new xdictionary("ja");
 163         dict->setJapaneseWordBreak();
 164         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("ja", "JP"));
 165         cBreakIterator = "com.sun.star.i18n.BreakIterator_ja";
 166 }
 167
 168 BreakIterator_ja::~BreakIterator_ja()
 169 {
 170         delete dict;
 171 }
 172
 173 //      ----------------------------------------------------
 174 //      class BreakIterator_ko
 175 //      ----------------------------------------------------;
 176 BreakIterator_ko::BreakIterator_ko()
 177 {
 178         hangingCharacters = LocaleData().getHangingCharacters(LOCALE("ko", "KR"));
 179         cBreakIterator = "com.sun.star.i18n.BreakIterator_ko";
 180 }
 181
 182 BreakIterator_ko::~BreakIterator_ko()
 183 {
 184 }
 185
 186 } } } }