transex3/source/hw2fw.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: hw2fw.cxx,v $
  10  * $Revision: 1.4 $
  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
  32 // MARKER(update_precomp.py): autogen include statement, do not remove
  33 #include "precompiled_transex3.hxx"
  34 #include <tools/string.hxx>
  35
  36 struct hw_pair
  37 {
  38     sal_Unicode                 nFrom;
  39     sal_Unicode                 nTo;
  40 };
  41
  42 #define MAKE_PAIR(a,b) { a, b }
  43
  44 static struct hw_pair aHWPairs[] =
  45 {
  46   MAKE_PAIR( 0xFF65, 0x30FB ),  // HALFWIDTH KATAKANA MIDDLE DOT --> KATAKANA MIDDLE DOT
  47   MAKE_PAIR( 0xFF66, 0x30F2 ),  // HALFWIDTH KATAKANA LETTER WO --> KATAKANA LETTER WO
  48   MAKE_PAIR( 0xFF67, 0x30A1 ),  // HALFWIDTH KATAKANA LETTER SMALL A --> KATAKANA LETTER SMALL A
  49   MAKE_PAIR( 0xFF68, 0x30A3 ),  // HALFWIDTH KATAKANA LETTER SMALL I --> KATAKANA LETTER SMALL I
  50   MAKE_PAIR( 0xFF69, 0x30A5 ),  // HALFWIDTH KATAKANA LETTER SMALL U --> KATAKANA LETTER SMALL U
  51   MAKE_PAIR( 0xFF6A, 0x30A7 ),  // HALFWIDTH KATAKANA LETTER SMALL E --> KATAKANA LETTER SMALL E
  52   MAKE_PAIR( 0xFF6B, 0x30A9 ),  // HALFWIDTH KATAKANA LETTER SMALL O --> KATAKANA LETTER SMALL O
  53   MAKE_PAIR( 0xFF6C, 0x30E3 ),  // HALFWIDTH KATAKANA LETTER SMALL YA --> KATAKANA LETTER SMALL YA
  54   MAKE_PAIR( 0xFF6D, 0x30E5 ),  // HALFWIDTH KATAKANA LETTER SMALL YU --> KATAKANA LETTER SMALL YU
  55   MAKE_PAIR( 0xFF6E, 0x30E7 ),  // HALFWIDTH KATAKANA LETTER SMALL YO --> KATAKANA LETTER SMALL YO
  56   MAKE_PAIR( 0xFF6F, 0x30C3 ),  // HALFWIDTH KATAKANA LETTER SMALL TU --> KATAKANA LETTER SMALL TU
  57   MAKE_PAIR( 0xFF70, 0x30FC ),  // HALFWIDTH KATAKANA-HIRAGANA PROLONGED SOUND MARK --> KATAKANA-HIRAGANA PROLONGED SOUND MARK
  58   MAKE_PAIR( 0xFF71, 0x30A2 ),  // HALFWIDTH KATAKANA LETTER A --> KATAKANA LETTER A
  59   MAKE_PAIR( 0xFF72, 0x30A4 ),  // HALFWIDTH KATAKANA LETTER I --> KATAKANA LETTER I
  60   MAKE_PAIR( 0xFF73, 0x30A6 ),  // HALFWIDTH KATAKANA LETTER U --> KATAKANA LETTER U
  61   MAKE_PAIR( 0xFF74, 0x30A8 ),  // HALFWIDTH KATAKANA LETTER E --> KATAKANA LETTER E
  62   MAKE_PAIR( 0xFF75, 0x30AA ),  // HALFWIDTH KATAKANA LETTER O --> KATAKANA LETTER O
  63   MAKE_PAIR( 0xFF76, 0x30AB ),  // HALFWIDTH KATAKANA LETTER KA --> KATAKANA LETTER KA
  64   MAKE_PAIR( 0xFF77, 0x30AD ),  // HALFWIDTH KATAKANA LETTER KI --> KATAKANA LETTER KI
  65   MAKE_PAIR( 0xFF78, 0x30AF ),  // HALFWIDTH KATAKANA LETTER KU --> KATAKANA LETTER KU
  66   MAKE_PAIR( 0xFF79, 0x30B1 ),  // HALFWIDTH KATAKANA LETTER KE --> KATAKANA LETTER KE
  67   MAKE_PAIR( 0xFF7A, 0x30B3 ),  // HALFWIDTH KATAKANA LETTER KO --> KATAKANA LETTER KO
  68   MAKE_PAIR( 0xFF7B, 0x30B5 ),  // HALFWIDTH KATAKANA LETTER SA --> KATAKANA LETTER SA
  69   MAKE_PAIR( 0xFF7C, 0x30B7 ),  // HALFWIDTH KATAKANA LETTER SI --> KATAKANA LETTER SI
  70   MAKE_PAIR( 0xFF7D, 0x30B9 ),  // HALFWIDTH KATAKANA LETTER SU --> KATAKANA LETTER SU
  71   MAKE_PAIR( 0xFF7E, 0x30BB ),  // HALFWIDTH KATAKANA LETTER SE --> KATAKANA LETTER SE
  72   MAKE_PAIR( 0xFF7F, 0x30BD ),  // HALFWIDTH KATAKANA LETTER SO --> KATAKANA LETTER SO
  73   MAKE_PAIR( 0xFF80, 0x30BF ),  // HALFWIDTH KATAKANA LETTER TA --> KATAKANA LETTER TA
  74   MAKE_PAIR( 0xFF81, 0x30C1 ),  // HALFWIDTH KATAKANA LETTER TI --> KATAKANA LETTER TI
  75   MAKE_PAIR( 0xFF82, 0x30C4 ),  // HALFWIDTH KATAKANA LETTER TU --> KATAKANA LETTER TU
  76   MAKE_PAIR( 0xFF83, 0x30C6 ),  // HALFWIDTH KATAKANA LETTER TE --> KATAKANA LETTER TE
  77   MAKE_PAIR( 0xFF84, 0x30C8 ),  // HALFWIDTH KATAKANA LETTER TO --> KATAKANA LETTER TO
  78   MAKE_PAIR( 0xFF85, 0x30CA ),  // HALFWIDTH KATAKANA LETTER NA --> KATAKANA LETTER NA
  79   MAKE_PAIR( 0xFF86, 0x30CB ),  // HALFWIDTH KATAKANA LETTER NI --> KATAKANA LETTER NI
  80   MAKE_PAIR( 0xFF87, 0x30CC ),  // HALFWIDTH KATAKANA LETTER NU --> KATAKANA LETTER NU
  81   MAKE_PAIR( 0xFF88, 0x30CD ),  // HALFWIDTH KATAKANA LETTER NE --> KATAKANA LETTER NE
  82   MAKE_PAIR( 0xFF89, 0x30CE ),  // HALFWIDTH KATAKANA LETTER NO --> KATAKANA LETTER NO
  83   MAKE_PAIR( 0xFF8A, 0x30CF ),  // HALFWIDTH KATAKANA LETTER HA --> KATAKANA LETTER HA
  84   MAKE_PAIR( 0xFF8B, 0x30D2 ),  // HALFWIDTH KATAKANA LETTER HI --> KATAKANA LETTER HI
  85   MAKE_PAIR( 0xFF8C, 0x30D5 ),  // HALFWIDTH KATAKANA LETTER HU --> KATAKANA LETTER HU
  86   MAKE_PAIR( 0xFF8D, 0x30D8 ),  // HALFWIDTH KATAKANA LETTER HE --> KATAKANA LETTER HE
  87   MAKE_PAIR( 0xFF8E, 0x30DB ),  // HALFWIDTH KATAKANA LETTER HO --> KATAKANA LETTER HO
  88   MAKE_PAIR( 0xFF8F, 0x30DE ),  // HALFWIDTH KATAKANA LETTER MA --> KATAKANA LETTER MA
  89   MAKE_PAIR( 0xFF90, 0x30DF ),  // HALFWIDTH KATAKANA LETTER MI --> KATAKANA LETTER MI
  90   MAKE_PAIR( 0xFF91, 0x30E0 ),  // HALFWIDTH KATAKANA LETTER MU --> KATAKANA LETTER MU
  91   MAKE_PAIR( 0xFF92, 0x30E1 ),  // HALFWIDTH KATAKANA LETTER ME --> KATAKANA LETTER ME
  92   MAKE_PAIR( 0xFF93, 0x30E2 ),  // HALFWIDTH KATAKANA LETTER MO --> KATAKANA LETTER MO
  93   MAKE_PAIR( 0xFF94, 0x30E4 ),  // HALFWIDTH KATAKANA LETTER YA --> KATAKANA LETTER YA
  94   MAKE_PAIR( 0xFF95, 0x30E6 ),  // HALFWIDTH KATAKANA LETTER YU --> KATAKANA LETTER YU
  95   MAKE_PAIR( 0xFF96, 0x30E8 ),  // HALFWIDTH KATAKANA LETTER YO --> KATAKANA LETTER YO
  96   MAKE_PAIR( 0xFF97, 0x30E9 ),  // HALFWIDTH KATAKANA LETTER RA --> KATAKANA LETTER RA
  97   MAKE_PAIR( 0xFF98, 0x30EA ),  // HALFWIDTH KATAKANA LETTER RI --> KATAKANA LETTER RI
  98   MAKE_PAIR( 0xFF99, 0x30EB ),  // HALFWIDTH KATAKANA LETTER RU --> KATAKANA LETTER RU
  99   MAKE_PAIR( 0xFF9A, 0x30EC ),  // HALFWIDTH KATAKANA LETTER RE --> KATAKANA LETTER RE
 100   MAKE_PAIR( 0xFF9B, 0x30ED ),  // HALFWIDTH KATAKANA LETTER RO --> KATAKANA LETTER RO
 101   MAKE_PAIR( 0xFF9C, 0x30EF ),  // HALFWIDTH KATAKANA LETTER WA --> KATAKANA LETTER WA
 102   MAKE_PAIR( 0xFF9D, 0x30F3 ),  // HALFWIDTH KATAKANA LETTER N --> KATAKANA LETTER N
 103   MAKE_PAIR( 0xFF9E, 0x3099 ),  // HALFWIDTH KATAKANA VOICED SOUND MARK --> COMBINING KATAKANA-HIRAGANA VOICED SOUND MARK
 104   MAKE_PAIR( 0xFF9F, 0x309A )   // HALFWIDTH KATAKANA SEMI-VOICED SOUND MARK --> COMBINING KATAKANA-
 105 };
 106
 107
 108 static struct hw_pair aCombine3099[] =
 109 {
 110     { 0x30a6, 0x30f4 },
 111     { 0x30ab, 0x30ac },
 112     { 0x30ad, 0x30ae },
 113     { 0x30af, 0x30b0 },
 114     { 0x30b1, 0x30b2 },
 115     { 0x30b3, 0x30b4 },
 116     { 0x30b5, 0x30b6 },
 117     { 0x30b7, 0x30b8 },
 118     { 0x30b9, 0x30ba },
 119     { 0x30bb, 0x30bc },
 120     { 0x30bd, 0x30be },
 121     { 0x30bf, 0x30c0 },
 122     { 0x30c1, 0x30c2 },
 123     { 0x30c4, 0x30c5 },
 124     { 0x30c6, 0x30c7 },
 125     { 0x30c8, 0x30c9 },
 126     { 0x30cf, 0x30d0 },
 127     { 0x30d2, 0x30d3 },
 128     { 0x30d5, 0x30d6 },
 129     { 0x30d8, 0x30d9 },
 130     { 0x30db, 0x30dc },
 131     { 0x30ef, 0x30f7 },
 132     { 0x30f0, 0x30f8 },
 133     { 0x30f1, 0x30f9 },
 134     { 0x30f2, 0x30fa },
 135     { 0x30fd, 0x30fe }
 136 };
 137
 138 static struct hw_pair aCombine309A[] =
 139 {
 140     { 0x30cf, 0x30d1 },
 141     { 0x30d2, 0x30d4 },
 142     { 0x30d5, 0x30d7 },
 143     { 0x30d8, 0x30da },
 144     { 0x30db, 0x30dd }
 145 };
 146
 147 USHORT ImplReplaceFullWidth( sal_Unicode* pString, USHORT nLen )
 148 {
 149     sal_Unicode* pRead  = pString;
 150     sal_Unicode* pWrite = pRead;
 151     USHORT nNewLen = nLen;
 152
 153     while( (pRead - pString) < nLen )
 154     {
 155         if( pWrite != pRead )
 156             *pWrite = *pRead;
 157
 158         if( *pRead >= 0xff65 && *pRead <= 0xff9f )
 159         {
 160             *pWrite = aHWPairs[ *pRead - 0xff65 ].nTo;
 161
 162             struct hw_pair* pTable = NULL;
 163             int nTableEntries = 0;
 164             if( *pWrite == 0x3099 )
 165             {
 166                 // replace 0x3099 combinations
 167                 pTable = aCombine3099;
 168                 nTableEntries = sizeof(aCombine3099)/sizeof(aCombine3099[0]);
 169             }
 170             else if( *pWrite == 0x309a )
 171             {
 172                 // replace 0x309a combinations
 173                 pTable = aCombine309A;
 174                 nTableEntries = sizeof(aCombine309A)/sizeof(aCombine309A[0]);
 175             }
 176             if( pTable )
 177             {
 178                 sal_Unicode c = pWrite[-1];
 179                 for( int i = 0; i < nTableEntries; i++ )
 180                     if( c == pTable[i].nFrom )
 181                     {
 182                         pWrite--;
 183                         *pWrite = pTable[i].nTo;
 184                         nNewLen--;
 185                         break;
 186                     }
 187             }
 188         }
 189         pRead++;
 190         pWrite++;
 191     }
 192     if( pWrite < pRead )
 193         *pWrite = 0;
 194
 195     return nNewLen;
 196 }
 197
 198 void ConvertHalfwitdhToFullwidth( String& rString )
 199 {
 200     USHORT nNewLen = ImplReplaceFullWidth( rString.GetBufferAccess(), rString.Len() );
 201     rString.ReleaseBufferAccess( nNewLen );
 202 }