sax/inc/xml2utf.hxx

   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: xml2utf.hxx,v $
  10  * $Revision: 1.3.10.1 $
  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 // TODO: Woher?
  32 #define Max( a, b )             (((a)>(b)) ? (a) : (b) )
  33 #define Min( a, b )             (((a)<(b)) ? (a) : (b) )
  34
  35 /*
  36 *
  37 * Text2UnicodeConverter
  38 *
  39 **/
  40 namespace sax_expatwrap {
  41
  42 class Text2UnicodeConverter
  43 {
  44
  45 public:
  46     Text2UnicodeConverter( const ::rtl::OString & sEncoding );
  47     ~Text2UnicodeConverter();
  48
  49     ::com::sun::star::uno::Sequence < sal_Unicode > convert( const ::com::sun::star::uno::Sequence<sal_Int8> & );
  50     sal_Bool canContinue() {  return m_bCanContinue; }
  51
  52 private:
  53     void init( rtl_TextEncoding encoding );
  54
  55     rtl_TextToUnicodeConverter  m_convText2Unicode;
  56     rtl_TextToUnicodeContext    m_contextText2Unicode;
  57     sal_Bool                                    m_bCanContinue;
  58     sal_Bool                                    m_bInitialized;
  59     rtl_TextEncoding                    m_rtlEncoding;
  60     ::com::sun::star::uno::Sequence<sal_Int8> m_seqSource;
  61 };
  62
  63 /*----------------------------------------
  64 *
  65 * Unicode2TextConverter
  66 *
  67 **-----------------------------------------*/
  68 class Unicode2TextConverter
  69 {
  70 public:
  71     Unicode2TextConverter( rtl_TextEncoding encoding );
  72     ~Unicode2TextConverter();
  73
  74     inline ::com::sun::star::uno::Sequence<sal_Int8> convert( const ::rtl::OUString &s )
  75         {
  76             return convert( s.getStr() , s.getLength() );
  77         }
  78     ::com::sun::star::uno::Sequence<sal_Int8> convert( const sal_Unicode * , sal_Int32 nLength );
  79     sal_Bool canContinue() {  return m_bCanContinue; }
  80
  81 private:
  82     void init( rtl_TextEncoding encoding );
  83
  84     rtl_UnicodeToTextConverter  m_convUnicode2Text;
  85     rtl_UnicodeToTextContext    m_contextUnicode2Text;
  86     sal_Bool                                    m_bCanContinue;
  87     sal_Bool                                    m_bInitialized;
  88     rtl_TextEncoding                    m_rtlEncoding;
  89     ::com::sun::star::uno::Sequence<sal_Unicode>                m_seqSource;
  90 };
  91
  92
  93
  94 /*----------------------------------------
  95 *
  96 * XMLFile2UTFConverter
  97 *
  98 **-----------------------------------------*/
  99 class XMLFile2UTFConverter
 100 {
 101 public:
 102     XMLFile2UTFConverter( ):
 103         m_bStarted( sal_False ),
 104         m_pText2Unicode( 0 ),
 105         m_pUnicode2Text( 0 )
 106         {}
 107
 108     ~XMLFile2UTFConverter();
 109
 110     void setInputStream( ::com::sun::star::uno::Reference< ::com::sun::star::io::XInputStream > &r ) { m_in = r; }
 111     void setEncoding( const ::rtl::OString &s ) { m_sEncoding = s; }
 112
 113
 114
 115     // @param nMaxToRead The number of chars, that should be read. Note that this is no exact number. There
 116     //                   may be returned less or more bytes than ordered.
 117     sal_Int32 readAndConvert( ::com::sun::star::uno::Sequence<sal_Int8> &seq , sal_Int32 nMaxToRead )
 118         throw ( ::com::sun::star::io::IOException,
 119                 ::com::sun::star::io::NotConnectedException ,
 120                 ::com::sun::star::io::BufferSizeExceededException ,
 121                 ::com::sun::star::uno::RuntimeException );
 122
 123 private:
 124
 125     // Called only on first Sequence of bytes. Tries to figure out file format and encoding information.
 126     // @return TRUE, when encoding information could be retrieved
 127     // @return FALSE, when no encoding information was found in file
 128     sal_Bool scanForEncoding( ::com::sun::star::uno::Sequence<sal_Int8> &seq );
 129
 130     // Called only on first Sequence of bytes. Tries to figure out
 131     // if enough data is available to scan encoding
 132     // @return TRUE, when encoding is retrievable
 133     // @return FALSE, when more data is needed
 134     sal_Bool isEncodingRecognizable( const ::com::sun::star::uno::Sequence< sal_Int8 > & seq );
 135
 136     // When encoding attribute is within the text (in the first line), it is removed.
 137     void removeEncoding( ::com::sun::star::uno::Sequence<sal_Int8> &seq );
 138
 139     // Initializes decoding depending on m_sEncoding setting
 140     void initializeDecoding();
 141 private:
 142     ::com::sun::star::uno::Reference< ::com::sun::star::io::XInputStream >  m_in;
 143
 144     sal_Bool m_bStarted;
 145     ::rtl::OString m_sEncoding;
 146
 147     Text2UnicodeConverter *m_pText2Unicode;
 148     Unicode2TextConverter *m_pUnicode2Text;
 149 };
 150 }