i18npool/inc/cclass_unicode.hxx

   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: cclass_unicode.hxx,v $
  10  * $Revision: 1.12 $
  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 #ifndef _I18N_CCLASS_UNICODE_HXX_
  31 #define _I18N_CCLASS_UNICODE_HXX_
  32
  33 #include <com/sun/star/i18n/XNativeNumberSupplier.hpp>
  34 #include <com/sun/star/i18n/XCharacterClassification.hpp>
  35 #include <com/sun/star/i18n/XLocaleData.hpp>
  36 #include <com/sun/star/lang/XMultiServiceFactory.hpp>
  37 #include <cppuhelper/implbase1.hxx> // helper for implementations
  38 #include <com/sun/star/lang/XServiceInfo.hpp>
  39
  40 #define TRANSLITERATION_casemapping
  41 #include <transliteration_body.hxx>
  42
  43 namespace com { namespace sun { namespace star { namespace i18n {
  44
  45 typedef sal_uInt32 UPT_FLAG_TYPE;
  46
  47 class cclass_Unicode : public cppu::WeakImplHelper1 < XCharacterClassification >
  48 {
  49 public:
  50     cclass_Unicode(com::sun::star::uno::Reference < com::sun::star::lang::XMultiServiceFactory > xSMgr );
  51     ~cclass_Unicode();
  52
  53     virtual rtl::OUString SAL_CALL toUpper( const rtl::OUString& Text, sal_Int32 nPos, sal_Int32 nCount,
  54         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  55     virtual rtl::OUString SAL_CALL toLower( const rtl::OUString& Text, sal_Int32 nPos, sal_Int32 nCount,
  56         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  57     virtual rtl::OUString SAL_CALL toTitle( const rtl::OUString& Text, sal_Int32 nPos, sal_Int32 nCount,
  58         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  59     virtual sal_Int16 SAL_CALL getType( const rtl::OUString& Text, sal_Int32 nPos )  throw(com::sun::star::uno::RuntimeException);
  60     virtual sal_Int16 SAL_CALL getCharacterDirection( const rtl::OUString& Text, sal_Int32 nPos )
  61         throw(com::sun::star::uno::RuntimeException);
  62     virtual sal_Int16 SAL_CALL getScript( const rtl::OUString& Text, sal_Int32 nPos ) throw(com::sun::star::uno::RuntimeException);
  63     virtual sal_Int32 SAL_CALL getCharacterType( const rtl::OUString& text, sal_Int32 nPos,
  64         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  65     virtual sal_Int32 SAL_CALL getStringType( const rtl::OUString& text, sal_Int32 nPos, sal_Int32 nCount,
  66         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  67     virtual ParseResult SAL_CALL parseAnyToken( const rtl::OUString& Text, sal_Int32 nPos,
  68         const com::sun::star::lang::Locale& rLocale, sal_Int32 nStartCharFlags, const rtl::OUString& userDefinedCharactersStart,
  69         sal_Int32 nContCharFlags, const rtl::OUString& userDefinedCharactersCont ) throw(com::sun::star::uno::RuntimeException);
  70     virtual ParseResult SAL_CALL parsePredefinedToken( sal_Int32 nTokenType, const rtl::OUString& Text,
  71         sal_Int32 nPos, const com::sun::star::lang::Locale& rLocale, sal_Int32 nStartCharFlags,
  72         const rtl::OUString& userDefinedCharactersStart, sal_Int32 nContCharFlags,
  73         const rtl::OUString& userDefinedCharactersCont ) throw(com::sun::star::uno::RuntimeException);
  74
  75     //XServiceInfo
  76     virtual rtl::OUString SAL_CALL getImplementationName() throw( com::sun::star::uno::RuntimeException );
  77     virtual sal_Bool SAL_CALL supportsService(const rtl::OUString& ServiceName) throw( com::sun::star::uno::RuntimeException );
  78     virtual com::sun::star::uno::Sequence< rtl::OUString > SAL_CALL getSupportedServiceNames() throw( com::sun::star::uno::RuntimeException );
  79
  80 protected:
  81     const sal_Char *cClass;
  82
  83 private:
  84     Transliteration_casemapping *trans;
  85
  86 // --- parser specific (implemented in cclass_unicode_parser.cxx) ---
  87
  88     enum ScanState
  89     {
  90         ssGetChar,
  91         ssGetValue,
  92         ssGetWord,
  93         ssGetWordFirstChar,
  94         ssGetString,
  95         ssGetBool,
  96         ssRewindFromValue,
  97         ssIgnoreLeadingInRewind,
  98         ssStopBack,
  99         ssBounce,
 100         ssStop
 101     };
 102
 103     static const sal_uInt8              nDefCnt;
 104     static const UPT_FLAG_TYPE  pDefaultParserTable[];
 105     static const sal_Int32              pParseTokensType[];
 106
 107     /// Flag values of table.
 108     static const UPT_FLAG_TYPE  TOKEN_ILLEGAL;
 109     static const UPT_FLAG_TYPE  TOKEN_CHAR;
 110     static const UPT_FLAG_TYPE  TOKEN_CHAR_BOOL;
 111     static const UPT_FLAG_TYPE  TOKEN_CHAR_WORD;
 112     static const UPT_FLAG_TYPE  TOKEN_CHAR_VALUE;
 113     static const UPT_FLAG_TYPE  TOKEN_CHAR_STRING;
 114     static const UPT_FLAG_TYPE  TOKEN_CHAR_DONTCARE;
 115     static const UPT_FLAG_TYPE  TOKEN_BOOL;
 116     static const UPT_FLAG_TYPE  TOKEN_WORD;
 117     static const UPT_FLAG_TYPE  TOKEN_WORD_SEP;
 118     static const UPT_FLAG_TYPE  TOKEN_VALUE;
 119     static const UPT_FLAG_TYPE  TOKEN_VALUE_SEP;
 120     static const UPT_FLAG_TYPE  TOKEN_VALUE_EXP;
 121     static const UPT_FLAG_TYPE  TOKEN_VALUE_SIGN;
 122     static const UPT_FLAG_TYPE  TOKEN_VALUE_EXP_VALUE;
 123     static const UPT_FLAG_TYPE  TOKEN_VALUE_DIGIT;
 124     static const UPT_FLAG_TYPE  TOKEN_NAME_SEP;
 125     static const UPT_FLAG_TYPE  TOKEN_STRING_SEP;
 126     static const UPT_FLAG_TYPE  TOKEN_EXCLUDED;
 127
 128     /// If and where c occurs in pStr
 129     static      const sal_Unicode*      StrChr( const sal_Unicode* pStr, sal_Unicode c );
 130
 131
 132     com::sun::star::uno::Reference < com::sun::star::lang::XMultiServiceFactory > xMSF;
 133
 134     /// used for parser only
 135     com::sun::star::lang::Locale        aParserLocale;
 136     com::sun::star::uno::Reference < XLocaleData > xLocaleData;
 137     com::sun::star::uno::Reference < com::sun::star::i18n::XNativeNumberSupplier > xNatNumSup;
 138     rtl::OUString             aStartChars;
 139     rtl::OUString             aContChars;
 140     UPT_FLAG_TYPE*                              pTable;
 141     UPT_FLAG_TYPE*                              pStart;
 142     UPT_FLAG_TYPE*                              pCont;
 143     sal_Int32                                   nStartTypes;
 144     sal_Int32                                   nContTypes;
 145     ScanState                                   eState;
 146     sal_Unicode                                 cGroupSep;
 147     sal_Unicode                                 cDecimalSep;
 148
 149     /// Get corresponding KParseTokens flag for a character
 150     sal_Int32 getParseTokensType( const sal_Unicode* aStr, sal_Int32 nPos );
 151
 152     /// Access parser table flags.
 153     UPT_FLAG_TYPE getFlags( const sal_Unicode* aStr, sal_Int32 nPos );
 154
 155     /// Access parser flags via International and special definitions.
 156     UPT_FLAG_TYPE getFlagsExtended( const sal_Unicode* aStr, sal_Int32 nPos );
 157
 158     /// Access parser table flags for user defined start characters.
 159     UPT_FLAG_TYPE getStartCharsFlags( sal_Unicode c );
 160
 161     /// Access parser table flags for user defined continuation characters.
 162     UPT_FLAG_TYPE getContCharsFlags( sal_Unicode c );
 163
 164     /// Setup parser table. Calls initParserTable() only if needed.
 165     void setupParserTable( const com::sun::star::lang::Locale& rLocale, sal_Int32 startCharTokenType,
 166         const rtl::OUString& userDefinedCharactersStart, sal_Int32 contCharTokenType,
 167         const rtl::OUString& userDefinedCharactersCont );
 168
 169     /// Init parser table.
 170     void initParserTable( const com::sun::star::lang::Locale& rLocale, sal_Int32 startCharTokenType,
 171         const rtl::OUString& userDefinedCharactersStart, sal_Int32 contCharTokenType,
 172         const rtl::OUString& userDefinedCharactersCont );
 173
 174     /// Destroy parser table.
 175     void destroyParserTable();
 176
 177     /// Parse a text.
 178     void parseText( ParseResult& r, const rtl::OUString& rText, sal_Int32 nPos,
 179         sal_Int32 nTokenType = 0xffffffff );
 180
 181     /// Setup International class, new'ed only if different from existing.
 182     sal_Bool setupInternational( const com::sun::star::lang::Locale& rLocale );
 183
 184     /// Implementation of getCharacterType() for one single character
 185     sal_Int32 SAL_CALL getCharType( const rtl::OUString& Text, sal_Int32 *nPos, sal_Int32 increment);
 186
 187 };
 188
 189 } } } }
 190
 191 #endif