sw/source/core/bastyp/breakit.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: breakit.cxx,v $
  10  * $Revision: 1.16.112.1 $
  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_sw.hxx"
  33
  34 #include "breakit.hxx"
  35 #include <unicode/uchar.h>
  36 #include <com/sun/star/lang/XMultiServiceFactory.hpp>
  37 #ifndef _COM_SUN_STAR_I18N_SCRIPTTYPE_HDL_
  38 #include <com/sun/star/i18n/ScriptType.hdl>
  39 #endif
  40 #include <unotools/localedatawrapper.hxx>
  41
  42 #ifndef _SVX_LINGU_HXX
  43 #include <svx/unolingu.hxx>
  44 #endif
  45 #include <svx/scripttypeitem.hxx>
  46 #include "swtypes.hxx"
  47
  48 using namespace com::sun::star;
  49
  50 SwBreakIt * pBreakIt = 0;
  51
  52 void SwBreakIt::_Create(
  53     const uno::Reference< lang::XMultiServiceFactory > & rxMSF)
  54 {
  55     delete pBreakIt, pBreakIt = new SwBreakIt( rxMSF );
  56 }
  57
  58 void SwBreakIt::_Delete()
  59 {
  60     delete pBreakIt, pBreakIt = 0;
  61 }
  62
  63 SwBreakIt * SwBreakIt::Get()
  64 {
  65     return pBreakIt;
  66 }
  67
  68 SwBreakIt::SwBreakIt(
  69     const uno::Reference< lang::XMultiServiceFactory > & rxMSF)
  70     : m_xMSF( rxMSF ),
  71       m_pLocale( NULL ),
  72       m_pForbidden( NULL ),
  73       aLast( LANGUAGE_DONTKNOW ),
  74       aForbiddenLang( LANGUAGE_DONTKNOW)
  75 {
  76     DBG_ASSERT( m_xMSF.is(), "SwBreakIt: no MultiServiceFactory" );
  77     //if ( m_xMSF.is() )
  78     //{
  79  //       xBreak = uno::Reference< i18n::XBreakIterator >(
  80     //          m_xMSF->createInstance(
  81     //                  rtl::OUString::createFromAscii( "com.sun.star.i18n.BreakIterator" ) ),
  82  //           uno::UNO_QUERY);
  83
  84  //       xCTLDetect = uno::Reference< i18n::XScriptTypeDetector >(
  85  //           m_xMSF->createInstance(
  86  //                rtl::OUString::createFromAscii( "com.sun.star.i18n.ScriptTypeDetector" ) ),
  87  //           uno::UNO_QUERY);
  88  //   }
  89 }
  90
  91 SwBreakIt::~SwBreakIt()
  92 {
  93     delete m_pLocale;
  94     delete m_pForbidden;
  95 }
  96 void SwBreakIt::createBreakIterator() const
  97 {
  98     if ( m_xMSF.is() && !xBreak.is() )
  99         xBreak.set(m_xMSF->createInstance(::rtl::OUString(RTL_CONSTASCII_USTRINGPARAM("com.sun.star.i18n.BreakIterator"))),uno::UNO_QUERY);
 100 }
 101 void SwBreakIt::createScriptTypeDetector()
 102 {
 103     if ( m_xMSF.is() && !xCTLDetect.is() )
 104         xCTLDetect.set(m_xMSF->createInstance(rtl::OUString(RTL_CONSTASCII_USTRINGPARAM( "com.sun.star.i18n.ScriptTypeDetector" ))),uno::UNO_QUERY);
 105 }
 106 void SwBreakIt::_GetLocale( const LanguageType aLang )
 107 {
 108     aLast = aLang;
 109     delete m_pLocale;
 110     m_pLocale = new lang::Locale( SvxCreateLocale( aLast ) );
 111 }
 112
 113 void SwBreakIt::_GetForbidden( const LanguageType aLang )
 114 {
 115     LocaleDataWrapper aWrap( m_xMSF, GetLocale( aLang ) );
 116
 117     aForbiddenLang = aLang;
 118     delete m_pForbidden;
 119     m_pForbidden = new i18n::ForbiddenCharacters( aWrap.getForbiddenCharacters() );
 120 }
 121
 122 USHORT SwBreakIt::GetRealScriptOfText( const String& rTxt,
 123                                         xub_StrLen nPos ) const
 124 {
 125     createBreakIterator();
 126     USHORT nScript = i18n::ScriptType::WEAK;
 127     if( xBreak.is() && rTxt.Len() )
 128     {
 129         if( nPos && nPos == rTxt.Len() )
 130             --nPos;
 131         nScript = xBreak->getScriptType( rTxt, nPos );
 132         sal_Int32 nChgPos = 0;
 133         if ( i18n::ScriptType::WEAK == nScript && nPos + 1 < rTxt.Len() )
 134         {
 135             // A weak character followed by a mark may be meant to combine with
 136             // the mark, so prefer the following character's script
 137             switch ( u_charType(rTxt.GetChar(nPos + 1) ) ) {
 138             case U_NON_SPACING_MARK:
 139             case U_ENCLOSING_MARK:
 140             case U_COMBINING_SPACING_MARK:
 141                 nScript = xBreak->getScriptType( rTxt, nPos+1 );
 142                 break;
 143             }
 144         }
 145         if( i18n::ScriptType::WEAK == nScript && nPos &&
 146             0 < (nChgPos = xBreak->beginOfScript( rTxt, nPos, nScript )) )
 147             nScript = xBreak->getScriptType( rTxt, nChgPos-1 );
 148
 149         if( i18n::ScriptType::WEAK == nScript && rTxt.Len() >
 150             ( nChgPos = xBreak->endOfScript( rTxt, nPos, nScript ) ) &&
 151             0 <= nChgPos )
 152             nScript = xBreak->getScriptType( rTxt, nChgPos );
 153     }
 154     if( i18n::ScriptType::WEAK == nScript )
 155         nScript = GetI18NScriptTypeOfLanguage( (USHORT)GetAppLanguage() );
 156     return nScript;
 157 }
 158
 159 USHORT SwBreakIt::GetAllScriptsOfText( const String& rTxt ) const
 160 {
 161     const USHORT coAllScripts = ( SCRIPTTYPE_LATIN |
 162                                   SCRIPTTYPE_ASIAN |
 163                                   SCRIPTTYPE_COMPLEX );
 164     createBreakIterator();
 165     USHORT nRet = 0, nScript;
 166     if( !xBreak.is() )
 167         nRet = coAllScripts;
 168     else if( rTxt.Len() )
 169     {
 170         for( xub_StrLen n = 0, nEnd = rTxt.Len(); n < nEnd;
 171                 n = static_cast<xub_StrLen>(xBreak->endOfScript( rTxt, n, nScript )) )
 172         {
 173             switch( nScript = xBreak->getScriptType( rTxt, n ) )
 174             {
 175             case i18n::ScriptType::LATIN:       nRet |= SCRIPTTYPE_LATIN;   break;
 176             case i18n::ScriptType::ASIAN:       nRet |= SCRIPTTYPE_ASIAN;   break;
 177             case i18n::ScriptType::COMPLEX: nRet |= SCRIPTTYPE_COMPLEX; break;
 178             case i18n::ScriptType::WEAK:
 179                     if( !nRet )
 180                         nRet |= coAllScripts;
 181                     break;
 182             }
 183             if( coAllScripts == nRet )
 184                 break;
 185         }
 186     }
 187     return nRet;
 188 }
 189