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 // prevent internal compiler error with MSVC6SP3 32 #include <utility> 33 34 #include <i18nutil/oneToOneMapping.hxx> 35 #define TRANSLITERATION_IterationMark_ja_JP 36 #include <transliteration_Ignore.hxx> 37 38 using namespace com::sun::star::uno; 39 using namespace com::sun::star::lang; 40 using namespace rtl; 41 42 namespace com { namespace sun { namespace star { namespace i18n { 43 44 OneToOneMappingTable_t ignoreIterationMark_ja_JP_mappingTable[] = { 45 MAKE_PAIR( 0x3046, 0x3094 ), // HIRAGANA LETTER U --> HIRAGANA LETTER VU 46 MAKE_PAIR( 0x304B, 0x304C ), // HIRAGANA LETTER KA --> HIRAGANA LETTER GA 47 MAKE_PAIR( 0x304D, 0x304E ), // HIRAGANA LETTER KI --> HIRAGANA LETTER GI 48 MAKE_PAIR( 0x304F, 0x3050 ), // HIRAGANA LETTER KU --> HIRAGANA LETTER GU 49 MAKE_PAIR( 0x3051, 0x3052 ), // HIRAGANA LETTER KE --> HIRAGANA LETTER GE 50 MAKE_PAIR( 0x3053, 0x3054 ), // HIRAGANA LETTER KO --> HIRAGANA LETTER GO 51 MAKE_PAIR( 0x3055, 0x3056 ), // HIRAGANA LETTER SA --> HIRAGANA LETTER ZA 52 MAKE_PAIR( 0x3057, 0x3058 ), // HIRAGANA LETTER SI --> HIRAGANA LETTER ZI 53 MAKE_PAIR( 0x3059, 0x305A ), // HIRAGANA LETTER SU --> HIRAGANA LETTER ZU 54 MAKE_PAIR( 0x305B, 0x305C ), // HIRAGANA LETTER SE --> HIRAGANA LETTER ZE 55 MAKE_PAIR( 0x305D, 0x305E ), // HIRAGANA LETTER SO --> HIRAGANA LETTER ZO 56 MAKE_PAIR( 0x305F, 0x3060 ), // HIRAGANA LETTER TA --> HIRAGANA LETTER DA 57 MAKE_PAIR( 0x3061, 0x3062 ), // HIRAGANA LETTER TI --> HIRAGANA LETTER DI 58 MAKE_PAIR( 0x3064, 0x3065 ), // HIRAGANA LETTER TU --> HIRAGANA LETTER DU 59 MAKE_PAIR( 0x3066, 0x3067 ), // HIRAGANA LETTER TE --> HIRAGANA LETTER DE 60 MAKE_PAIR( 0x3068, 0x3069 ), // HIRAGANA LETTER TO --> HIRAGANA LETTER DO 61 MAKE_PAIR( 0x306F, 0x3070 ), // HIRAGANA LETTER HA --> HIRAGANA LETTER BA 62 MAKE_PAIR( 0x3072, 0x3073 ), // HIRAGANA LETTER HI --> HIRAGANA LETTER BI 63 MAKE_PAIR( 0x3075, 0x3076 ), // HIRAGANA LETTER HU --> HIRAGANA LETTER BU 64 MAKE_PAIR( 0x3078, 0x3079 ), // HIRAGANA LETTER HE --> HIRAGANA LETTER BE 65 MAKE_PAIR( 0x307B, 0x307C ), // HIRAGANA LETTER HO --> HIRAGANA LETTER BO 66 MAKE_PAIR( 0x309D, 0x309E ), // HIRAGANA ITERATION MARK --> HIRAGANA VOICED ITERATION MARK 67 MAKE_PAIR( 0x30A6, 0x30F4 ), // KATAKANA LETTER U --> KATAKANA LETTER VU 68 MAKE_PAIR( 0x30AB, 0x30AC ), // KATAKANA LETTER KA --> KATAKANA LETTER GA 69 MAKE_PAIR( 0x30AD, 0x30AE ), // KATAKANA LETTER KI --> KATAKANA LETTER GI 70 MAKE_PAIR( 0x30AF, 0x30B0 ), // KATAKANA LETTER KU --> KATAKANA LETTER GU 71 MAKE_PAIR( 0x30B1, 0x30B2 ), // KATAKANA LETTER KE --> KATAKANA LETTER GE 72 MAKE_PAIR( 0x30B3, 0x30B4 ), // KATAKANA LETTER KO --> KATAKANA LETTER GO 73 MAKE_PAIR( 0x30B5, 0x30B6 ), // KATAKANA LETTER SA --> KATAKANA LETTER ZA 74 MAKE_PAIR( 0x30B7, 0x30B8 ), // KATAKANA LETTER SI --> KATAKANA LETTER ZI 75 MAKE_PAIR( 0x30B9, 0x30BA ), // KATAKANA LETTER SU --> KATAKANA LETTER ZU 76 MAKE_PAIR( 0x30BB, 0x30BC ), // KATAKANA LETTER SE --> KATAKANA LETTER ZE 77 MAKE_PAIR( 0x30BD, 0x30BE ), // KATAKANA LETTER SO --> KATAKANA LETTER ZO 78 MAKE_PAIR( 0x30BF, 0x30C0 ), // KATAKANA LETTER TA --> KATAKANA LETTER DA 79 MAKE_PAIR( 0x30C1, 0x30C2 ), // KATAKANA LETTER TI --> KATAKANA LETTER DI 80 MAKE_PAIR( 0x30C4, 0x30C5 ), // KATAKANA LETTER TU --> KATAKANA LETTER DU 81 MAKE_PAIR( 0x30C6, 0x30C7 ), // KATAKANA LETTER TE --> KATAKANA LETTER DE 82 MAKE_PAIR( 0x30C8, 0x30C9 ), // KATAKANA LETTER TO --> KATAKANA LETTER DO 83 MAKE_PAIR( 0x30CF, 0x30D0 ), // KATAKANA LETTER HA --> KATAKANA LETTER BA 84 MAKE_PAIR( 0x30D2, 0x30D3 ), // KATAKANA LETTER HI --> KATAKANA LETTER BI 85 MAKE_PAIR( 0x30D5, 0x30D6 ), // KATAKANA LETTER HU --> KATAKANA LETTER BU 86 MAKE_PAIR( 0x30D8, 0x30D9 ), // KATAKANA LETTER HE --> KATAKANA LETTER BE 87 MAKE_PAIR( 0x30DB, 0x30DC ), // KATAKANA LETTER HO --> KATAKANA LETTER BO 88 MAKE_PAIR( 0x30EF, 0x30F7 ), // KATAKANA LETTER WA --> KATAKANA LETTER VA 89 MAKE_PAIR( 0x30F0, 0x30F8 ), // KATAKANA LETTER WI --> KATAKANA LETTER VI 90 MAKE_PAIR( 0x30F1, 0x30F9 ), // KATAKANA LETTER WE --> KATAKANA LETTER VE 91 MAKE_PAIR( 0x30F2, 0x30FA ), // KATAKANA LETTER WO --> KATAKANA LETTER VO 92 MAKE_PAIR( 0x30FD, 0x30FE ) // KATAKANA ITERATION MARK --> KATAKANA VOICED ITERATION MARK 93 }; 94 95 96 OUString SAL_CALL 97 ignoreIterationMark_ja_JP::folding( const OUString& inStr, sal_Int32 startPos, sal_Int32 nCount, Sequence< sal_Int32 >& offset ) 98 throw(RuntimeException) 99 { 100 oneToOneMapping aTable(ignoreIterationMark_ja_JP_mappingTable, sizeof(ignoreIterationMark_ja_JP_mappingTable)); 101 102 // Create a string buffer which can hold nCount + 1 characters. 103 // The reference count is 0 now. 104 rtl_uString * newStr = x_rtl_uString_new_WithLength( nCount ); // defined in x_rtl_ustring.h 105 sal_Unicode * dst = newStr->buffer; 106 const sal_Unicode * src = inStr.getStr() + startPos; 107 108 sal_Int32 * p = 0; 109 sal_Int32 position = 0; 110 if (useOffset) { 111 // Allocate nCount length to offset argument. 112 offset.realloc( nCount ); 113 p = offset.getArray(); 114 position = startPos; 115 } 116 117 // 118 sal_Unicode previousChar = *src ++; 119 sal_Unicode currentChar; 120 121 // Conversion 122 while (-- nCount > 0) { 123 currentChar = *src ++; 124 125 switch ( currentChar ) { 126 case 0x30fd: // KATAKANA ITERATION MARK 127 case 0x309d: // HIRAGANA ITERATION MARK 128 case 0x3005: // IDEOGRAPHIC ITERATION MARK 129 currentChar = previousChar; 130 break; 131 case 0x30fe: // KATAKANA VOICED ITERATION MARK 132 case 0x309e: // HIRAGANA VOICED ITERATION MARK 133 currentChar = aTable[ previousChar ]; 134 break; 135 } 136 if (useOffset) 137 *p ++ = position ++; 138 *dst ++ = previousChar; 139 previousChar = currentChar; 140 } 141 142 if (nCount == 0) { 143 if (useOffset) 144 *p = position; 145 *dst ++ = previousChar; 146 } 147 148 *dst = (sal_Unicode) 0; 149 150 newStr->length = sal_Int32(dst - newStr->buffer); 151 if (useOffset) 152 offset.realloc(newStr->length); 153 return OUString( newStr ); // defined in rtl/usrting. The reference count is increased from 0 to 1. 154 155 } 156 157 } } } } 158