transex3/source/txtconv.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: txtconv.cxx,v $
  10  * $Revision: 1.8 $
  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_transex3.hxx"
  33 #include <stdio.h>
  34 #include <tools/fsys.hxx>
  35 #include <tools/stream.hxx>
  36
  37 // local includes
  38 #include "utf8conv.hxx"
  39
  40 extern void ConvertHalfwitdhToFullwidth( String& rString );
  41
  42 /*****************************************************************************/
  43 void Help()
  44 /*****************************************************************************/
  45 {
  46     fprintf( stdout, "\n" );
  47     fprintf( stdout, "txtconv (c)2001 by StarOffice Entwicklungs GmbH\n" );
  48     fprintf( stdout, "===============================================\n" );
  49     fprintf( stdout, "\n" );
  50     fprintf( stdout, "txtconv converts textfiles from or to UTF-8\n" );
  51     fprintf( stdout, "\n" );
  52     fprintf( stdout, "Syntax: txtconv -t|-f charset filename (destinationfile)\n" );
  53     fprintf( stdout, "Switches: -t   => conversion from charset to UTF-8\n" );
  54     fprintf( stdout, "          -f   => conversion from UTF-8 to charset\n" );
  55     fprintf( stdout, "\n" );
  56     fprintf( stdout, "Allowed charsets:\n" );
  57     fprintf( stdout, "          MS_932  => Japanese\n" );
  58     fprintf( stdout, "          MS_936  => Chinese Simplified\n" );
  59     fprintf( stdout, "          MS_949  => Korean\n" );
  60     fprintf( stdout, "          MS_950  => Chinese Traditional\n" );
  61     fprintf( stdout, "          MS_1250 => East Europe\n" );
  62     fprintf( stdout, "          MS_1251 => Cyrillic\n" );
  63     fprintf( stdout, "          MS_1252 => West Europe\n" );
  64     fprintf( stdout, "          MS_1253 => Greek\n" );
  65     fprintf( stdout, "          MS_1254 => Turkish\n" );
  66     fprintf( stdout, "          MS_1255 => Hebrew\n" );
  67     fprintf( stdout, "          MS_1256 => Arabic\n" );
  68     fprintf( stdout, "          HW2FW   => Only with -t, converts half to full width katakana" );
  69     fprintf( stdout, "\n" );
  70 }
  71
  72 /*****************************************************************************/
  73 #if defined(UNX) || defined(OS2)
  74 int main( int argc, char *argv[] )
  75 #else
  76 int _cdecl main( int argc, char *argv[] )
  77 #endif
  78 /*****************************************************************************/
  79 {
  80     if (( argc != 4 ) && ( argc != 5 )) {
  81         Help();
  82         exit ( 0 );
  83     }
  84
  85     if ( ByteString( argv[ 1 ] ) == "-t" || ByteString( argv[ 1 ] ) == "-f" ) {
  86         rtl_TextEncoding nEncoding = RTL_TEXTENCODING_MS_1252;
  87
  88         BOOL bHW2FW = FALSE;
  89
  90         ByteString sCharset( argv[ 2 ] );
  91         sCharset.ToUpperAscii();
  92
  93         if              ( sCharset == "MS_932" )        nEncoding = RTL_TEXTENCODING_MS_932;
  94         else if ( sCharset == "MS_936" )        nEncoding = RTL_TEXTENCODING_MS_936;
  95         else if ( sCharset == "MS_949" )        nEncoding = RTL_TEXTENCODING_MS_949;
  96         else if ( sCharset == "MS_950" )        nEncoding = RTL_TEXTENCODING_MS_950;
  97         else if ( sCharset == "MS_1250" )       nEncoding = RTL_TEXTENCODING_MS_1250;
  98         else if ( sCharset == "MS_1251" )       nEncoding = RTL_TEXTENCODING_MS_1251;
  99         else if ( sCharset == "MS_1252" )       nEncoding = RTL_TEXTENCODING_MS_1252;
 100         else if ( sCharset == "MS_1253" )       nEncoding = RTL_TEXTENCODING_MS_1253;
 101         else if ( sCharset == "MS_1254" )       nEncoding = RTL_TEXTENCODING_MS_1254;
 102         else if ( sCharset == "MS_1255" )       nEncoding = RTL_TEXTENCODING_MS_1255;
 103         else if ( sCharset == "MS_1256" )       nEncoding = RTL_TEXTENCODING_MS_1256;
 104         else if ( sCharset == "MS_1257" )       nEncoding = RTL_TEXTENCODING_MS_1257;
 105         else if (( sCharset == "HW2FW" ) && ( ByteString( argv[ 1 ] ) == "-t" )) bHW2FW = TRUE;
 106
 107         else {
 108             Help();
 109             exit ( 1 );
 110         }
 111
 112         DirEntry aSource = DirEntry( String( argv[ 3 ], RTL_TEXTENCODING_ASCII_US ));
 113         if ( !aSource.Exists()) {
 114             fprintf( stderr, "\nERROR: File %s not found!\n\n", ByteString( argv[ 3 ] ).GetBuffer());
 115             exit ( 2 );
 116         }
 117
 118         String sOutput;
 119         SvFileStream aOutput;
 120         if ( argc == 5 ) {
 121             sOutput= String( argv[ 4 ], RTL_TEXTENCODING_ASCII_US );
 122             aOutput.Open( sOutput, STREAM_STD_WRITE | STREAM_TRUNC );
 123             if ( !aOutput.IsOpen()) {
 124                 fprintf( stderr, "\nERROR: Could not open output file %s!\n\n", argv[ 4 ]);
 125                 exit ( 3 );
 126             }
 127         }
 128
 129         String sGSI( argv[ 3 ], RTL_TEXTENCODING_ASCII_US );
 130         SvFileStream aGSI( sGSI, STREAM_STD_READ );
 131         if ( !aGSI.IsOpen()) {
 132             fprintf( stderr, "\nERROR: Could not open input file %s!\n\n", argv[ 3 ]);
 133             exit ( 3 );
 134         }
 135
 136         ByteString sGSILine;
 137         while ( !aGSI.IsEof()) {
 138
 139             aGSI.ReadLine( sGSILine );
 140             if ( bHW2FW ) {
 141                 String sConverter( sGSILine, RTL_TEXTENCODING_UTF8 );
 142                 ConvertHalfwitdhToFullwidth( sConverter );
 143                 sGSILine = ByteString( sConverter, RTL_TEXTENCODING_UTF8 );
 144             }
 145             else {
 146                 if ( ByteString( argv[ 1 ] ) == "-t" )
 147                     sGSILine = UTF8Converter::ConvertToUTF8( sGSILine, nEncoding );
 148                 else
 149                     sGSILine = UTF8Converter::ConvertFromUTF8( sGSILine, nEncoding );
 150             }
 151
 152             if ( aOutput.IsOpen())
 153                 aOutput.WriteLine( sGSILine );
 154             else
 155                 fprintf( stdout, "%s\n", sGSILine.GetBuffer());
 156         }
 157
 158         aGSI.Close();
 159         if ( aOutput.IsOpen())
 160             aOutput.Close();
 161     }
 162     else {
 163         Help();
 164         exit( 1 );
 165     }
 166
 167     return 0;
 168 }