i18npool/inc/cclass_unicode.hxx

   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 #ifndef _I18N_CCLASS_UNICODE_HXX_
  28 #define _I18N_CCLASS_UNICODE_HXX_
  29
  30 #include <com/sun/star/i18n/XNativeNumberSupplier.hpp>
  31 #include <com/sun/star/i18n/XCharacterClassification.hpp>
  32 #include <com/sun/star/i18n/XLocaleData.hpp>
  33 #include <com/sun/star/lang/XMultiServiceFactory.hpp>
  34 #include <cppuhelper/implbase1.hxx> // helper for implementations
  35 #include <com/sun/star/lang/XServiceInfo.hpp>
  36
  37 #define TRANSLITERATION_casemapping
  38 #include <transliteration_body.hxx>
  39
  40 namespace com { namespace sun { namespace star { namespace i18n {
  41
  42 typedef sal_uInt32 UPT_FLAG_TYPE;
  43
  44 class cclass_Unicode : public cppu::WeakImplHelper1 < XCharacterClassification >
  45 {
  46 public:
  47     cclass_Unicode(com::sun::star::uno::Reference < com::sun::star::lang::XMultiServiceFactory > xSMgr );
  48     ~cclass_Unicode();
  49
  50     virtual rtl::OUString SAL_CALL toUpper( const rtl::OUString& Text, sal_Int32 nPos, sal_Int32 nCount,
  51         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  52     virtual rtl::OUString SAL_CALL toLower( const rtl::OUString& Text, sal_Int32 nPos, sal_Int32 nCount,
  53         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  54     virtual rtl::OUString SAL_CALL toTitle( const rtl::OUString& Text, sal_Int32 nPos, sal_Int32 nCount,
  55         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  56     virtual sal_Int16 SAL_CALL getType( const rtl::OUString& Text, sal_Int32 nPos )  throw(com::sun::star::uno::RuntimeException);
  57     virtual sal_Int16 SAL_CALL getCharacterDirection( const rtl::OUString& Text, sal_Int32 nPos )
  58         throw(com::sun::star::uno::RuntimeException);
  59     virtual sal_Int16 SAL_CALL getScript( const rtl::OUString& Text, sal_Int32 nPos ) throw(com::sun::star::uno::RuntimeException);
  60     virtual sal_Int32 SAL_CALL getCharacterType( const rtl::OUString& text, sal_Int32 nPos,
  61         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  62     virtual sal_Int32 SAL_CALL getStringType( const rtl::OUString& text, sal_Int32 nPos, sal_Int32 nCount,
  63         const com::sun::star::lang::Locale& rLocale ) throw(com::sun::star::uno::RuntimeException);
  64     virtual ParseResult SAL_CALL parseAnyToken( const rtl::OUString& Text, sal_Int32 nPos,
  65         const com::sun::star::lang::Locale& rLocale, sal_Int32 nStartCharFlags, const rtl::OUString& userDefinedCharactersStart,
  66         sal_Int32 nContCharFlags, const rtl::OUString& userDefinedCharactersCont ) throw(com::sun::star::uno::RuntimeException);
  67     virtual ParseResult SAL_CALL parsePredefinedToken( sal_Int32 nTokenType, const rtl::OUString& Text,
  68         sal_Int32 nPos, const com::sun::star::lang::Locale& rLocale, sal_Int32 nStartCharFlags,
  69         const rtl::OUString& userDefinedCharactersStart, sal_Int32 nContCharFlags,
  70         const rtl::OUString& userDefinedCharactersCont ) throw(com::sun::star::uno::RuntimeException);
  71
  72     //XServiceInfo
  73     virtual rtl::OUString SAL_CALL getImplementationName() throw( com::sun::star::uno::RuntimeException );
  74     virtual sal_Bool SAL_CALL supportsService(const rtl::OUString& ServiceName) throw( com::sun::star::uno::RuntimeException );
  75     virtual com::sun::star::uno::Sequence< rtl::OUString > SAL_CALL getSupportedServiceNames() throw( com::sun::star::uno::RuntimeException );
  76
  77 protected:
  78     const sal_Char *cClass;
  79
  80 private:
  81     Transliteration_casemapping *trans;
  82
  83 // --- parser specific (implemented in cclass_unicode_parser.cxx) ---
  84
  85     enum ScanState
  86     {
  87         ssGetChar,
  88         ssGetValue,
  89         ssGetWord,
  90         ssGetWordFirstChar,
  91         ssGetString,
  92         ssGetBool,
  93         ssRewindFromValue,
  94         ssIgnoreLeadingInRewind,
  95         ssStopBack,
  96         ssBounce,
  97         ssStop
  98     };
  99
 100     static const sal_uInt8      nDefCnt;
 101     static const UPT_FLAG_TYPE  pDefaultParserTable[];
 102     static const sal_Int32      pParseTokensType[];
 103
 104     /// Flag values of table.
 105     static const UPT_FLAG_TYPE  TOKEN_ILLEGAL;
 106     static const UPT_FLAG_TYPE  TOKEN_CHAR;
 107     static const UPT_FLAG_TYPE  TOKEN_CHAR_BOOL;
 108     static const UPT_FLAG_TYPE  TOKEN_CHAR_WORD;
 109     static const UPT_FLAG_TYPE  TOKEN_CHAR_VALUE;
 110     static const UPT_FLAG_TYPE  TOKEN_CHAR_STRING;
 111     static const UPT_FLAG_TYPE  TOKEN_CHAR_DONTCARE;
 112     static const UPT_FLAG_TYPE  TOKEN_BOOL;
 113     static const UPT_FLAG_TYPE  TOKEN_WORD;
 114     static const UPT_FLAG_TYPE  TOKEN_WORD_SEP;
 115     static const UPT_FLAG_TYPE  TOKEN_VALUE;
 116     static const UPT_FLAG_TYPE  TOKEN_VALUE_SEP;
 117     static const UPT_FLAG_TYPE  TOKEN_VALUE_EXP;
 118     static const UPT_FLAG_TYPE  TOKEN_VALUE_SIGN;
 119     static const UPT_FLAG_TYPE  TOKEN_VALUE_EXP_VALUE;
 120     static const UPT_FLAG_TYPE  TOKEN_VALUE_DIGIT;
 121     static const UPT_FLAG_TYPE  TOKEN_NAME_SEP;
 122     static const UPT_FLAG_TYPE  TOKEN_STRING_SEP;
 123     static const UPT_FLAG_TYPE  TOKEN_EXCLUDED;
 124
 125     /// If and where c occurs in pStr
 126     static  const sal_Unicode*  StrChr( const sal_Unicode* pStr, sal_Unicode c );
 127
 128
 129     com::sun::star::uno::Reference < com::sun::star::lang::XMultiServiceFactory > xMSF;
 130
 131     /// used for parser only
 132     com::sun::star::lang::Locale    aParserLocale;
 133     com::sun::star::uno::Reference < XLocaleData > xLocaleData;
 134     com::sun::star::uno::Reference < com::sun::star::i18n::XNativeNumberSupplier > xNatNumSup;
 135     rtl::OUString             aStartChars;
 136     rtl::OUString             aContChars;
 137     UPT_FLAG_TYPE*              pTable;
 138     UPT_FLAG_TYPE*              pStart;
 139     UPT_FLAG_TYPE*              pCont;
 140     sal_Int32                   nStartTypes;
 141     sal_Int32                   nContTypes;
 142     ScanState                   eState;
 143     sal_Unicode                 cGroupSep;
 144     sal_Unicode                 cDecimalSep;
 145
 146     /// Get corresponding KParseTokens flag for a character
 147     sal_Int32 getParseTokensType( const sal_Unicode* aStr, sal_Int32 nPos );
 148
 149     /// Access parser table flags.
 150     UPT_FLAG_TYPE getFlags( const sal_Unicode* aStr, sal_Int32 nPos );
 151
 152     /// Access parser flags via International and special definitions.
 153     UPT_FLAG_TYPE getFlagsExtended( const sal_Unicode* aStr, sal_Int32 nPos );
 154
 155     /// Access parser table flags for user defined start characters.
 156     UPT_FLAG_TYPE getStartCharsFlags( sal_Unicode c );
 157
 158     /// Access parser table flags for user defined continuation characters.
 159     UPT_FLAG_TYPE getContCharsFlags( sal_Unicode c );
 160
 161     /// Setup parser table. Calls initParserTable() only if needed.
 162     void setupParserTable( const com::sun::star::lang::Locale& rLocale, sal_Int32 startCharTokenType,
 163         const rtl::OUString& userDefinedCharactersStart, sal_Int32 contCharTokenType,
 164         const rtl::OUString& userDefinedCharactersCont );
 165
 166     /// Init parser table.
 167     void initParserTable( const com::sun::star::lang::Locale& rLocale, sal_Int32 startCharTokenType,
 168         const rtl::OUString& userDefinedCharactersStart, sal_Int32 contCharTokenType,
 169         const rtl::OUString& userDefinedCharactersCont );
 170
 171     /// Destroy parser table.
 172     void destroyParserTable();
 173
 174     /// Parse a text.
 175     void parseText( ParseResult& r, const rtl::OUString& rText, sal_Int32 nPos,
 176         sal_Int32 nTokenType = 0xffffffff );
 177
 178     /// Setup International class, new'ed only if different from existing.
 179     sal_Bool setupInternational( const com::sun::star::lang::Locale& rLocale );
 180
 181     /// Implementation of getCharacterType() for one single character
 182     sal_Int32 SAL_CALL getCharType( const rtl::OUString& Text, sal_Int32 *nPos, sal_Int32 increment);
 183
 184 };
 185
 186 } } } }
 187
 188 #endif