i18npool/source/transliteration/ignoreIterationMark_ja_JP.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: ignoreIterationMark_ja_JP.cxx,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
  31 // MARKER(update_precomp.py): autogen include statement, do not remove
  32 #include "precompiled_i18npool.hxx"
  33
  34 // prevent internal compiler error with MSVC6SP3
  35 #include <utility>
  36
  37 #include <i18nutil/oneToOneMapping.hxx>
  38 #define TRANSLITERATION_IterationMark_ja_JP
  39 #include <transliteration_Ignore.hxx>
  40
  41 using namespace com::sun::star::uno;
  42 using namespace com::sun::star::lang;
  43 using namespace rtl;
  44
  45 namespace com { namespace sun { namespace star { namespace i18n {
  46
  47 OneToOneMappingTable_t ignoreIterationMark_ja_JP_mappingTable[] = {
  48     MAKE_PAIR( 0x3046, 0x3094 ),  // HIRAGANA LETTER U --> HIRAGANA LETTER VU
  49     MAKE_PAIR( 0x304B, 0x304C ),  // HIRAGANA LETTER KA --> HIRAGANA LETTER GA
  50     MAKE_PAIR( 0x304D, 0x304E ),  // HIRAGANA LETTER KI --> HIRAGANA LETTER GI
  51     MAKE_PAIR( 0x304F, 0x3050 ),  // HIRAGANA LETTER KU --> HIRAGANA LETTER GU
  52     MAKE_PAIR( 0x3051, 0x3052 ),  // HIRAGANA LETTER KE --> HIRAGANA LETTER GE
  53     MAKE_PAIR( 0x3053, 0x3054 ),  // HIRAGANA LETTER KO --> HIRAGANA LETTER GO
  54     MAKE_PAIR( 0x3055, 0x3056 ),  // HIRAGANA LETTER SA --> HIRAGANA LETTER ZA
  55     MAKE_PAIR( 0x3057, 0x3058 ),  // HIRAGANA LETTER SI --> HIRAGANA LETTER ZI
  56     MAKE_PAIR( 0x3059, 0x305A ),  // HIRAGANA LETTER SU --> HIRAGANA LETTER ZU
  57     MAKE_PAIR( 0x305B, 0x305C ),  // HIRAGANA LETTER SE --> HIRAGANA LETTER ZE
  58     MAKE_PAIR( 0x305D, 0x305E ),  // HIRAGANA LETTER SO --> HIRAGANA LETTER ZO
  59     MAKE_PAIR( 0x305F, 0x3060 ),  // HIRAGANA LETTER TA --> HIRAGANA LETTER DA
  60     MAKE_PAIR( 0x3061, 0x3062 ),  // HIRAGANA LETTER TI --> HIRAGANA LETTER DI
  61     MAKE_PAIR( 0x3064, 0x3065 ),  // HIRAGANA LETTER TU --> HIRAGANA LETTER DU
  62     MAKE_PAIR( 0x3066, 0x3067 ),  // HIRAGANA LETTER TE --> HIRAGANA LETTER DE
  63     MAKE_PAIR( 0x3068, 0x3069 ),  // HIRAGANA LETTER TO --> HIRAGANA LETTER DO
  64     MAKE_PAIR( 0x306F, 0x3070 ),  // HIRAGANA LETTER HA --> HIRAGANA LETTER BA
  65     MAKE_PAIR( 0x3072, 0x3073 ),  // HIRAGANA LETTER HI --> HIRAGANA LETTER BI
  66     MAKE_PAIR( 0x3075, 0x3076 ),  // HIRAGANA LETTER HU --> HIRAGANA LETTER BU
  67     MAKE_PAIR( 0x3078, 0x3079 ),  // HIRAGANA LETTER HE --> HIRAGANA LETTER BE
  68     MAKE_PAIR( 0x307B, 0x307C ),  // HIRAGANA LETTER HO --> HIRAGANA LETTER BO
  69     MAKE_PAIR( 0x309D, 0x309E ),  // HIRAGANA ITERATION MARK --> HIRAGANA VOICED ITERATION MARK
  70     MAKE_PAIR( 0x30A6, 0x30F4 ),  // KATAKANA LETTER U --> KATAKANA LETTER VU
  71     MAKE_PAIR( 0x30AB, 0x30AC ),  // KATAKANA LETTER KA --> KATAKANA LETTER GA
  72     MAKE_PAIR( 0x30AD, 0x30AE ),  // KATAKANA LETTER KI --> KATAKANA LETTER GI
  73     MAKE_PAIR( 0x30AF, 0x30B0 ),  // KATAKANA LETTER KU --> KATAKANA LETTER GU
  74     MAKE_PAIR( 0x30B1, 0x30B2 ),  // KATAKANA LETTER KE --> KATAKANA LETTER GE
  75     MAKE_PAIR( 0x30B3, 0x30B4 ),  // KATAKANA LETTER KO --> KATAKANA LETTER GO
  76     MAKE_PAIR( 0x30B5, 0x30B6 ),  // KATAKANA LETTER SA --> KATAKANA LETTER ZA
  77     MAKE_PAIR( 0x30B7, 0x30B8 ),  // KATAKANA LETTER SI --> KATAKANA LETTER ZI
  78     MAKE_PAIR( 0x30B9, 0x30BA ),  // KATAKANA LETTER SU --> KATAKANA LETTER ZU
  79     MAKE_PAIR( 0x30BB, 0x30BC ),  // KATAKANA LETTER SE --> KATAKANA LETTER ZE
  80     MAKE_PAIR( 0x30BD, 0x30BE ),  // KATAKANA LETTER SO --> KATAKANA LETTER ZO
  81     MAKE_PAIR( 0x30BF, 0x30C0 ),  // KATAKANA LETTER TA --> KATAKANA LETTER DA
  82     MAKE_PAIR( 0x30C1, 0x30C2 ),  // KATAKANA LETTER TI --> KATAKANA LETTER DI
  83     MAKE_PAIR( 0x30C4, 0x30C5 ),  // KATAKANA LETTER TU --> KATAKANA LETTER DU
  84     MAKE_PAIR( 0x30C6, 0x30C7 ),  // KATAKANA LETTER TE --> KATAKANA LETTER DE
  85     MAKE_PAIR( 0x30C8, 0x30C9 ),  // KATAKANA LETTER TO --> KATAKANA LETTER DO
  86     MAKE_PAIR( 0x30CF, 0x30D0 ),  // KATAKANA LETTER HA --> KATAKANA LETTER BA
  87     MAKE_PAIR( 0x30D2, 0x30D3 ),  // KATAKANA LETTER HI --> KATAKANA LETTER BI
  88     MAKE_PAIR( 0x30D5, 0x30D6 ),  // KATAKANA LETTER HU --> KATAKANA LETTER BU
  89     MAKE_PAIR( 0x30D8, 0x30D9 ),  // KATAKANA LETTER HE --> KATAKANA LETTER BE
  90     MAKE_PAIR( 0x30DB, 0x30DC ),  // KATAKANA LETTER HO --> KATAKANA LETTER BO
  91     MAKE_PAIR( 0x30EF, 0x30F7 ),  // KATAKANA LETTER WA --> KATAKANA LETTER VA
  92     MAKE_PAIR( 0x30F0, 0x30F8 ),  // KATAKANA LETTER WI --> KATAKANA LETTER VI
  93     MAKE_PAIR( 0x30F1, 0x30F9 ),  // KATAKANA LETTER WE --> KATAKANA LETTER VE
  94     MAKE_PAIR( 0x30F2, 0x30FA ),  // KATAKANA LETTER WO --> KATAKANA LETTER VO
  95     MAKE_PAIR( 0x30FD, 0x30FE )   // KATAKANA ITERATION MARK --> KATAKANA VOICED ITERATION MARK
  96 };
  97
  98
  99 OUString SAL_CALL
 100 ignoreIterationMark_ja_JP::folding( const OUString& inStr, sal_Int32 startPos, sal_Int32 nCount, Sequence< sal_Int32 >& offset )
 101   throw(RuntimeException)
 102 {
 103     oneToOneMapping aTable(ignoreIterationMark_ja_JP_mappingTable, sizeof(ignoreIterationMark_ja_JP_mappingTable));
 104
 105     // Create a string buffer which can hold nCount + 1 characters.
 106     // The reference count is 0 now.
 107     rtl_uString * newStr = x_rtl_uString_new_WithLength( nCount ); // defined in x_rtl_ustring.h
 108     sal_Unicode * dst = newStr->buffer;
 109     const sal_Unicode * src = inStr.getStr() + startPos;
 110
 111     sal_Int32 * p = 0;
 112     sal_Int32 position = 0;
 113     if (useOffset) {
 114         // Allocate nCount length to offset argument.
 115         offset.realloc( nCount );
 116         p = offset.getArray();
 117         position = startPos;
 118     }
 119
 120     //
 121     sal_Unicode previousChar = *src ++;
 122     sal_Unicode currentChar;
 123
 124     // Conversion
 125     while (-- nCount > 0) {
 126         currentChar = *src ++;
 127
 128         switch ( currentChar ) {
 129             case 0x30fd: // KATAKANA ITERATION MARK
 130             case 0x309d: // HIRAGANA ITERATION MARK
 131             case 0x3005: // IDEOGRAPHIC ITERATION MARK
 132                 currentChar = previousChar;
 133                 break;
 134             case 0x30fe: // KATAKANA VOICED ITERATION MARK
 135             case 0x309e: // HIRAGANA VOICED ITERATION MARK
 136                 currentChar = aTable[ previousChar ];
 137                 break;
 138         }
 139         if (useOffset)
 140             *p ++ = position ++;
 141         *dst ++ = previousChar;
 142         previousChar = currentChar;
 143     }
 144
 145     if (nCount == 0) {
 146         if (useOffset)
 147             *p = position;
 148         *dst ++ = previousChar;
 149     }
 150
 151     *dst = (sal_Unicode) 0;
 152
 153     newStr->length = sal_Int32(dst - newStr->buffer);
 154     if (useOffset)
 155         offset.realloc(newStr->length);
 156     return OUString( newStr ); // defined in rtl/usrting. The reference count is increased from 0 to 1.
 157
 158 }
 159
 160 } } } }