include/vlc_charset.h

   1 /*****************************************************************************
   2  * vlc_charset.h: Unicode UTF-8 wrappers function
   3  *****************************************************************************
   4  * Copyright (C) 2003-2005 VLC authors and VideoLAN
   5  * Copyright © 2005-2010 Rémi Denis-Courmont
   6  * $Id$
   7  *
   8  * Author: Rémi Denis-Courmont <rem # videolan,org>
   9  *
  10  * This program is free software; you can redistribute it and/or modify it
  11  * under the terms of the GNU Lesser General Public License as published by
  12  * the Free Software Foundation; either version 2.1 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program 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 for more details.
  19  *
  20  * You should have received a copy of the GNU Lesser General Public License
  21  * along with this program; if not, write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
  23  *****************************************************************************/
  24
  25 #ifndef VLC_CHARSET_H
  26 #define VLC_CHARSET_H 1
  27
  28 /**
  29  * \file
  30  * This files handles locale conversions in vlc
  31  */
  32
  33 /* iconv wrappers (defined in src/extras/libc.c) */
  34 typedef void *vlc_iconv_t;
  35 VLC_API vlc_iconv_t vlc_iconv_open( const char *, const char * ) VLC_USED;
  36 VLC_API size_t vlc_iconv( vlc_iconv_t, const char **, size_t *, char **, size_t * ) VLC_USED;
  37 VLC_API int vlc_iconv_close( vlc_iconv_t );
  38
  39 #include <stdarg.h>
  40
  41 VLC_API int utf8_vfprintf( FILE *stream, const char *fmt, va_list ap );
  42 VLC_API int utf8_fprintf( FILE *, const char *, ... ) VLC_FORMAT( 2, 3 );
  43 VLC_API char * vlc_strcasestr(const char *, const char *) VLC_USED;
  44
  45 VLC_API char * EnsureUTF8( char * );
  46 VLC_API const char * IsUTF8( const char * ) VLC_USED;
  47
  48 VLC_API char * FromCharset( const char *charset, const void *data, size_t data_size ) VLC_USED;
  49 VLC_API void * ToCharset( const char *charset, const char *in, size_t *outsize ) VLC_USED;
  50
  51 #ifdef _WIN32
  52 VLC_USED
  53 static inline char *FromWide (const wchar_t *wide)
  54 {
  55     size_t len = WideCharToMultiByte (CP_UTF8, 0, wide, -1, NULL, 0, NULL, NULL);
  56     if (len == 0)
  57         return NULL;
  58
  59     char *out = (char *)malloc (len);
  60
  61     if (likely(out))
  62         WideCharToMultiByte (CP_UTF8, 0, wide, -1, out, len, NULL, NULL);
  63     return out;
  64 }
  65
  66 VLC_USED
  67 static inline wchar_t *ToWide (const char *utf8)
  68 {
  69     int len = MultiByteToWideChar (CP_UTF8, 0, utf8, -1, NULL, 0);
  70     if (len == 0)
  71         return NULL;
  72
  73     wchar_t *out = (wchar_t *)malloc (len * sizeof (wchar_t));
  74
  75     if (likely(out))
  76         MultiByteToWideChar (CP_UTF8, 0, utf8, -1, out, len);
  77     return out;
  78 }
  79
  80 VLC_USED VLC_MALLOC
  81 static inline char *ToCodePage (unsigned cp, const char *utf8)
  82 {
  83     wchar_t *wide = ToWide (utf8);
  84     if (wide == NULL)
  85         return NULL;
  86
  87     size_t len = WideCharToMultiByte (cp, 0, wide, -1, NULL, 0, NULL, NULL);
  88     if (len == 0)
  89         return NULL;
  90
  91     char *out = (char *)malloc (len);
  92     if (likely(out != NULL))
  93         WideCharToMultiByte (cp, 0, wide, -1, out, len, NULL, NULL);
  94     free (wide);
  95     return out;
  96 }
  97
  98 VLC_USED VLC_MALLOC
  99 static inline char *FromCodePage (unsigned cp, const char *mb)
 100 {
 101     int len = MultiByteToWideChar (cp, 0, mb, -1, NULL, 0);
 102     if (len == 0)
 103         return NULL;
 104
 105     wchar_t *wide = (wchar_t *)malloc (len * sizeof (wchar_t));
 106     if (unlikely(wide == NULL))
 107         return NULL;
 108     MultiByteToWideChar (cp, 0, mb, -1, wide, len);
 109
 110     char *utf8 = FromWide (wide);
 111     free (wide);
 112     return utf8;
 113 }
 114
 115 VLC_USED VLC_MALLOC
 116 static inline char *FromANSI (const char *ansi)
 117 {
 118     return FromCodePage (GetACP (), ansi);
 119 }
 120
 121 VLC_USED VLC_MALLOC
 122 static inline char *ToANSI (const char *utf8)
 123 {
 124     return ToCodePage (GetACP (), utf8);
 125 }
 126
 127 # ifdef UNICODE
 128 #  define FromT FromWide
 129 #  define ToT   ToWide
 130 # else
 131 #  define FromT FromANSI
 132 #  define ToT   ToANSI
 133 # endif
 134 # define FromLocale    FromANSI
 135 # define ToLocale      ToANSI
 136 # define LocaleFree(s) free((char *)(s))
 137 # define FromLocaleDup FromANSI
 138 # define ToLocaleDup   ToANSI
 139
 140 #elif defined(__OS2__)
 141
 142 VLC_USED static inline char *FromLocale (const char *locale)
 143 {
 144     return locale ? FromCharset ((char *)"", locale, strlen(locale)) : NULL;
 145 }
 146
 147 VLC_USED static inline char *ToLocale (const char *utf8)
 148 {
 149     size_t outsize;
 150     return utf8 ? (char *)ToCharset ("", utf8, &outsize) : NULL;
 151 }
 152
 153 VLC_USED static inline void LocaleFree (const char *str)
 154 {
 155     free ((char *)str);
 156 }
 157
 158 VLC_USED static inline char *FromLocaleDup (const char *locale)
 159 {
 160     return FromCharset ("", locale, strlen(locale));
 161 }
 162
 163 VLC_USED static inline char *ToLocaleDup (const char *utf8)
 164 {
 165     size_t outsize;
 166     return (char *)ToCharset ("", utf8, &outsize);
 167 }
 168
 169 #else
 170
 171 # define FromLocale(l) (l)
 172 # define ToLocale(u)   (u)
 173 # define LocaleFree(s) ((void)(s))
 174 # define FromLocaleDup strdup
 175 # define ToLocaleDup   strdup
 176 #endif
 177
 178 /**
 179  * Converts a nul-terminated string from ISO-8859-1 to UTF-8.
 180  */
 181 static inline char *FromLatin1 (const char *latin)
 182 {
 183     char *str = (char *)malloc (2 * strlen (latin) + 1), *utf8 = str;
 184     unsigned char c;
 185
 186     if (str == NULL)
 187         return NULL;
 188
 189     while ((c = *(latin++)) != '\0')
 190     {
 191          if (c >= 0x80)
 192          {
 193              *(utf8++) = 0xC0 | (c >> 6);
 194              *(utf8++) = 0x80 | (c & 0x3F);
 195          }
 196          else
 197              *(utf8++) = c;
 198     }
 199     *(utf8++) = '\0';
 200
 201     utf8 = (char *)realloc (str, utf8 - str);
 202     return utf8 ? utf8 : str;
 203 }
 204
 205 VLC_API double us_strtod( const char *, char ** ) VLC_USED;
 206 VLC_API float us_strtof( const char *, char ** ) VLC_USED;
 207 VLC_API double us_atof( const char * ) VLC_USED;
 208 VLC_API int us_vasprintf( char **, const char *, va_list );
 209 VLC_API int us_asprintf( char **, const char *, ... ) VLC_USED;
 210
 211 #endif