libs/convert.h

   1 /*
   2    Copyright (C) 2001-2006, William Joseph.
   3    All Rights Reserved.
   4
   5    This file is part of GtkRadiant.
   6
   7    GtkRadiant is free software; you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation; either version 2 of the License, or
  10    (at your option) any later version.
  11
  12    GtkRadiant is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with GtkRadiant; if not, write to the Free Software
  19    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  20  */
  21
  22 #if !defined( INCLUDED_CONVERT_H )
  23 #define INCLUDED_CONVERT_H
  24
  25 /// \file
  26 /// \brief Character encoding conversion.
  27
  28 #include "debugging/debugging.h"
  29 #include <algorithm>
  30 #include <glib.h>
  31
  32 #include "character.h"
  33
  34 /// \brief Returns the number of bytes required to represent \p character in UTF-8 encoding.
  35 inline std::size_t utf8_character_length( const char* character ){
  36         if ( ( *character & 0xE0 ) == 0xC0 ) { // 110xxxxx
  37                 return 2;
  38         }
  39         else if ( ( *character & 0xF0 ) == 0xE0 ) { // 1110xxxx
  40                 return 3;
  41         }
  42         else if ( ( *character & 0xF8 ) == 0xF0 ) { // 11110xxx
  43                 return 4;
  44         }
  45         else if ( ( *character & 0xFC ) == 0xF8 ) { // 111110xx
  46                 return 5;
  47         }
  48         else if ( ( *character & 0xFE ) == 0xFC ) { // 1111110x
  49                 return 6;
  50         }
  51         ERROR_MESSAGE( "" );
  52         return 0;
  53 }
  54
  55 struct UTF8Character
  56 {
  57         const char* buffer;
  58         std::size_t length;
  59         UTF8Character() : buffer( 0 ), length( 0 ){
  60         }
  61         UTF8Character( const char* bytes ) : buffer( bytes ), length( utf8_character_length( bytes ) ){
  62         }
  63 };
  64
  65 inline bool operator<( const UTF8Character& self, const UTF8Character& other ){
  66         return std::lexicographical_compare( self.buffer, self.buffer + self.length, other.buffer, other.buffer + other.length );
  67 }
  68
  69 /// \brief Writes \p c to \p ostream in Hex form. Useful for debugging.
  70 template<typename TextOutputStreamType>
  71 inline TextOutputStreamType& ostream_write( TextOutputStreamType& ostream, const UTF8Character& c ){
  72         for ( const char* p = c.buffer; p != c.buffer + c.length; ++p )
  73         {
  74                 ostream << HexChar( *p );
  75         }
  76         return ostream;
  77 }
  78
  79
  80
  81 /// \brief The character-set encoding for the current C locale.
  82 ///
  83 /// Obtain the global instance with globalCharacterSet().
  84 class CharacterSet
  85 {
  86 const char* m_charSet;
  87 public:
  88 CharacterSet(){
  89         if ( g_get_charset( &m_charSet ) != FALSE ) {
  90                 m_charSet = 0;
  91         }
  92 }
  93 bool isUTF8() const {
  94         return m_charSet == 0;
  95 }
  96 const char* get() const {
  97         return m_charSet;
  98 }
  99 };
 100
 101 typedef LazyStatic<CharacterSet> GlobalCharacterSet;
 102
 103 /// \brief Returns the global instance of CharacterSet.
 104 inline CharacterSet& globalCharacterSet(){
 105         return GlobalCharacterSet::instance();
 106 }
 107
 108
 109 class UTF8CharacterToExtendedASCII
 110 {
 111 public:
 112 UTF8Character m_utf8;
 113 char m_c;
 114 UTF8CharacterToExtendedASCII() : m_c( '\0' ){
 115 }
 116 UTF8CharacterToExtendedASCII( const UTF8Character& utf8, char c ) : m_utf8( utf8 ), m_c( c ){
 117 }
 118 };
 119
 120 inline bool operator<( const UTF8CharacterToExtendedASCII& self, const UTF8CharacterToExtendedASCII& other ){
 121         return self.m_utf8 < other.m_utf8;
 122 }
 123
 124 inline std::size_t extended_ascii_to_index( char c ){
 125         return static_cast<std::size_t>( c & 0x7F );
 126 }
 127
 128 inline char extended_ascii_for_index( std::size_t i ){
 129         return static_cast<char>( i | 0x80 );
 130 }
 131
 132 /// \brief The active extended-ascii character set encoding.
 133 /// Performs UTF-8 encoding and decoding of extended-ascii characters.
 134 ///
 135 /// Obtain the global instance with globalExtendedASCIICharacterSet().
 136 class ExtendedASCIICharacterSet
 137 {
 138 typedef char UTF8CharBuffer[6];
 139 UTF8CharBuffer m_converted[128];
 140 UTF8Character m_decodeMap[128];
 141 UTF8CharacterToExtendedASCII m_encodeMap[128];
 142 public:
 143 ExtendedASCIICharacterSet(){
 144         if ( !globalCharacterSet().isUTF8() ) {
 145                 GIConv descriptor = g_iconv_open( "UTF-8", globalCharacterSet().get() );
 146                 for ( std::size_t i = 1; i < 128; ++i )
 147                 {
 148                         char c = extended_ascii_for_index( i );
 149                         char* inbuf = &c;
 150                         std::size_t inbytesleft = 1;
 151                         char* outbuf = m_converted[i];
 152                         std::size_t outbytesleft = 6;
 153                         if ( g_iconv( descriptor, &inbuf, &inbytesleft, &outbuf, &outbytesleft ) != (size_t)( -1 ) ) {
 154                                 UTF8Character utf8( m_converted[i] );
 155                                 m_decodeMap[i] = utf8;
 156                                 m_encodeMap[i] = UTF8CharacterToExtendedASCII( utf8, c );
 157                         }
 158                 }
 159                 g_iconv_close( descriptor );
 160                 std::sort( m_encodeMap, m_encodeMap + 128 );
 161         }
 162 }
 163 /// \brief Prints the (up to) 128 characters in the current extended-ascii character set.
 164 /// Useful for debugging.
 165 void print() const {
 166         globalOutputStream() << "UTF-8 conversion required from charset: " << globalCharacterSet().get() << "\n";
 167         for ( std::size_t i = 1; i < 128; ++i )
 168         {
 169                 if ( m_decodeMap[i].buffer != 0 ) {
 170                         globalOutputStream() << extended_ascii_for_index( i ) << " = " << m_decodeMap[i] << "\n";
 171                 }
 172         }
 173 }
 174 /// \brief Returns \p c decoded from extended-ascii to UTF-8.
 175 /// \p c must be an extended-ascii character.
 176 const UTF8Character& decode( char c ) const {
 177         ASSERT_MESSAGE( !globalCharacterSet().isUTF8(), "locale is utf8, no conversion required" );
 178         ASSERT_MESSAGE( !char_is_ascii( c ), "decode: ascii character" );
 179         ASSERT_MESSAGE( m_decodeMap[extended_ascii_to_index( c )].buffer != 0, "decode: invalid character: " << HexChar( c ) );
 180         return m_decodeMap[extended_ascii_to_index( c )];
 181 }
 182 /// \brief Returns \p c encoded to extended-ascii from UTF-8.
 183 /// \p c must map to an extended-ascii character.
 184 char encode( const UTF8Character& c ) const {
 185         ASSERT_MESSAGE( !globalCharacterSet().isUTF8(), "locale is utf8, no conversion required" );
 186         ASSERT_MESSAGE( !char_is_ascii( *c.buffer ), "encode: ascii character" );
 187         std::pair<const UTF8CharacterToExtendedASCII*, const UTF8CharacterToExtendedASCII*> range
 188                 = std::equal_range( m_encodeMap, m_encodeMap + 128, UTF8CharacterToExtendedASCII( c, 0 ) );
 189         ASSERT_MESSAGE( range.first != range.second, "encode: invalid character: " << c );
 190         return ( *range.first ).m_c;
 191 }
 192 };
 193
 194 typedef LazyStatic<ExtendedASCIICharacterSet> GlobalExtendedASCIICharacterSet;
 195
 196 /// \brief Returns the global instance of ExtendedASCIICharacterSet.
 197 inline ExtendedASCIICharacterSet& globalExtendedASCIICharacterSet(){
 198         return GlobalExtendedASCIICharacterSet::instance();
 199 }
 200
 201 class ConvertUTF8ToLocale
 202 {
 203 public:
 204 StringRange m_range;
 205 ConvertUTF8ToLocale( const char* string ) : m_range( StringRange( string, string + strlen( string ) ) ){
 206 }
 207 ConvertUTF8ToLocale( const StringRange& range ) : m_range( range ){
 208 }
 209 };
 210
 211 /// \brief Writes \p convert to \p ostream after encoding each character to extended-ascii from UTF-8.
 212 template<typename TextOutputStreamType>
 213 inline TextOutputStreamType& ostream_write( TextOutputStreamType& ostream, const ConvertUTF8ToLocale& convert ){
 214         if ( globalCharacterSet().isUTF8() ) {
 215                 return ostream << convert.m_range;
 216         }
 217
 218         for ( const char* p = convert.m_range.first; p != convert.m_range.last; )
 219         {
 220                 if ( !char_is_ascii( *p ) ) {
 221                         UTF8Character c( p );
 222                         ostream << globalExtendedASCIICharacterSet().encode( c );
 223                         p += c.length;
 224                 }
 225                 else
 226                 {
 227                         ostream << *p++;
 228                 }
 229         }
 230         return ostream;
 231 }
 232
 233
 234 class ConvertLocaleToUTF8
 235 {
 236 public:
 237 StringRange m_range;
 238 ConvertLocaleToUTF8( const char* string ) : m_range( StringRange( string, string + strlen( string ) ) ){
 239 }
 240 ConvertLocaleToUTF8( const StringRange& range ) : m_range( range ){
 241 }
 242 };
 243
 244 /// \brief Writes \p convert to \p ostream after decoding each character from extended-ascii to UTF-8.
 245 template<typename TextOutputStreamType>
 246 inline TextOutputStreamType& ostream_write( TextOutputStreamType& ostream, const ConvertLocaleToUTF8& convert ){
 247         if ( globalCharacterSet().isUTF8() ) {
 248                 return ostream << convert.m_range;
 249         }
 250
 251         for ( const char* p = convert.m_range.first; p != convert.m_range.last; ++p )
 252         {
 253                 if ( !char_is_ascii( *p ) ) {
 254                         UTF8Character c( globalExtendedASCIICharacterSet().decode( *p ) );
 255                         ostream.write( c.buffer, c.length );
 256                 }
 257                 else
 258                 {
 259                         ostream << *p;
 260                 }
 261         }
 262         return ostream;
 263 }
 264
 265
 266 #endif