memory/codepage.c

   1 /*
   2  * Code page functions
   3  *
   4  * Copyright 2000 Alexandre Julliard
   5  *
   6  * This library is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * This library is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with this library; if not, write to the Free Software
  18  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  19  */
  20
  21 #include <assert.h>
  22 #include <stdio.h>
  23 #include <stdlib.h>
  24 #include <string.h>
  25
  26 #include "winbase.h"
  27 #include "winerror.h"
  28 #include "winnls.h"
  29 #include "wine/unicode.h"
  30 #include "thread.h"
  31 #include "wine/debug.h"
  32
  33 WINE_DEFAULT_DEBUG_CHANNEL(string);
  34
  35 /* current code pages */
  36 static const union cptable *ansi_cptable;
  37 static const union cptable *oem_cptable;
  38 static const union cptable *mac_cptable;
  39 static LCID default_lcid = MAKELCID( MAKELANGID(LANG_ENGLISH,SUBLANG_DEFAULT), SORT_DEFAULT );
  40
  41 /* setup default codepage info before we can get at the locale stuff */
  42 static void init_codepages(void)
  43 {
  44     ansi_cptable = cp_get_table( 1252 );
  45     oem_cptable  = cp_get_table( 437 );
  46     mac_cptable  = cp_get_table( 10000 );
  47     assert( ansi_cptable );
  48     assert( oem_cptable );
  49     assert( mac_cptable );
  50 }
  51
  52 /* find the table for a given codepage, handling CP_ACP etc. pseudo-codepages */
  53 static const union cptable *get_codepage_table( unsigned int codepage )
  54 {
  55     const union cptable *ret = NULL;
  56
  57     if (!ansi_cptable) init_codepages();
  58
  59     switch(codepage)
  60     {
  61     case CP_ACP:
  62         return ansi_cptable;
  63     case CP_OEMCP:
  64         return oem_cptable;
  65     case CP_MACCP:
  66         return mac_cptable;
  67     case CP_UTF7:
  68     case CP_UTF8:
  69         break;
  70     case CP_THREAD_ACP:
  71         if (!(codepage = NtCurrentTeb()->code_page)) return ansi_cptable;
  72         /* fall through */
  73     default:
  74         if (codepage == ansi_cptable->info.codepage) return ansi_cptable;
  75         if (codepage == oem_cptable->info.codepage) return oem_cptable;
  76         if (codepage == mac_cptable->info.codepage) return mac_cptable;
  77         ret = cp_get_table( codepage );
  78         break;
  79     }
  80     return ret;
  81 }
  82
  83 /* initialize default code pages from locale info */
  84 /* FIXME: should be done in init_codepages, but it can't right now */
  85 /* since it needs KERNEL32 to be loaded for the locale info. */
  86 void CODEPAGE_Init( UINT ansi, UINT oem, UINT mac, LCID lcid )
  87 {
  88     extern void __wine_init_codepages( const union cptable *ansi, const union cptable *oem );
  89     const union cptable *table;
  90
  91     default_lcid = lcid;
  92     if (!ansi_cptable) init_codepages();  /* just in case */
  93
  94     if ((table = cp_get_table( ansi ))) ansi_cptable = table;
  95     if ((table = cp_get_table( oem ))) oem_cptable = table;
  96     if ((table = cp_get_table( mac ))) mac_cptable = table;
  97     __wine_init_codepages( ansi_cptable, oem_cptable );
  98
  99     TRACE( "ansi=%03d oem=%03d mac=%03d\n", ansi_cptable->info.codepage,
 100            oem_cptable->info.codepage, mac_cptable->info.codepage );
 101 }
 102
 103 /******************************************************************************
 104  *              GetACP   (KERNEL32.@)
 105  *
 106  * RETURNS
 107  *    Current ANSI code-page identifier, default if no current defined
 108  */
 109 UINT WINAPI GetACP(void)
 110 {
 111     if (!ansi_cptable) return 1252;
 112     return ansi_cptable->info.codepage;
 113 }
 114
 115
 116 /***********************************************************************
 117  *              GetOEMCP   (KERNEL32.@)
 118  */
 119 UINT WINAPI GetOEMCP(void)
 120 {
 121     if (!oem_cptable) return 437;
 122     return oem_cptable->info.codepage;
 123 }
 124
 125
 126 /***********************************************************************
 127  *           IsValidCodePage   (KERNEL32.@)
 128  */
 129 BOOL WINAPI IsValidCodePage( UINT codepage )
 130 {
 131     switch(codepage) {
 132     case CP_SYMBOL:
 133         return FALSE;
 134     case CP_UTF7:
 135     case CP_UTF8:
 136         return TRUE;
 137     default:
 138         return cp_get_table( codepage ) != NULL;
 139     }
 140 }
 141
 142
 143 /***********************************************************************
 144  *              GetUserDefaultLangID (KERNEL32.@)
 145  */
 146 LANGID WINAPI GetUserDefaultLangID(void)
 147 {
 148     return LANGIDFROMLCID(default_lcid);
 149 }
 150
 151
 152 /***********************************************************************
 153  *              GetSystemDefaultLangID (KERNEL32.@)
 154  */
 155 LANGID WINAPI GetSystemDefaultLangID(void)
 156 {
 157     return GetUserDefaultLangID();
 158 }
 159
 160
 161 /***********************************************************************
 162  *              GetUserDefaultLCID (KERNEL32.@)
 163  */
 164 LCID WINAPI GetUserDefaultLCID(void)
 165 {
 166     return default_lcid;
 167 }
 168
 169
 170 /***********************************************************************
 171  *              GetSystemDefaultLCID (KERNEL32.@)
 172  */
 173 LCID WINAPI GetSystemDefaultLCID(void)
 174 {
 175     return GetUserDefaultLCID();
 176 }
 177
 178
 179 /***********************************************************************
 180  *           IsDBCSLeadByteEx   (KERNEL32.@)
 181  */
 182 BOOL WINAPI IsDBCSLeadByteEx( UINT codepage, BYTE testchar )
 183 {
 184     const union cptable *table = get_codepage_table( codepage );
 185     return table && is_dbcs_leadbyte( table, testchar );
 186 }
 187
 188
 189 /***********************************************************************
 190  *           IsDBCSLeadByte   (KERNEL32.@)
 191  *           IsDBCSLeadByte   (KERNEL.207)
 192  */
 193 BOOL WINAPI IsDBCSLeadByte( BYTE testchar )
 194 {
 195     if (!ansi_cptable) return FALSE;
 196     return is_dbcs_leadbyte( ansi_cptable, testchar );
 197 }
 198
 199
 200 /***********************************************************************
 201  *           GetCPInfo   (KERNEL32.@)
 202  */
 203 BOOL WINAPI GetCPInfo( UINT codepage, LPCPINFO cpinfo )
 204 {
 205     const union cptable *table = get_codepage_table( codepage );
 206
 207     if (!table)
 208     {
 209         SetLastError( ERROR_INVALID_PARAMETER );
 210         return FALSE;
 211     }
 212     if (table->info.def_char & 0xff00)
 213     {
 214         cpinfo->DefaultChar[0] = table->info.def_char & 0xff00;
 215         cpinfo->DefaultChar[1] = table->info.def_char & 0x00ff;
 216     }
 217     else
 218     {
 219         cpinfo->DefaultChar[0] = table->info.def_char & 0xff;
 220         cpinfo->DefaultChar[1] = 0;
 221     }
 222     if ((cpinfo->MaxCharSize = table->info.char_size) == 2)
 223         memcpy( cpinfo->LeadByte, table->dbcs.lead_bytes, sizeof(cpinfo->LeadByte) );
 224     else
 225         cpinfo->LeadByte[0] = cpinfo->LeadByte[1] = 0;
 226
 227     return TRUE;
 228 }
 229
 230
 231 /***********************************************************************
 232  *              EnumSystemCodePagesA   (KERNEL32.@)
 233  */
 234 BOOL WINAPI EnumSystemCodePagesA( CODEPAGE_ENUMPROCA lpfnCodePageEnum, DWORD flags )
 235 {
 236     const union cptable *table;
 237     char buffer[10];
 238     int index = 0;
 239
 240     for (;;)
 241     {
 242         if (!(table = cp_enum_table( index++ ))) break;
 243         sprintf( buffer, "%d", table->info.codepage );
 244         if (!lpfnCodePageEnum( buffer )) break;
 245     }
 246     return TRUE;
 247 }
 248
 249
 250 /***********************************************************************
 251  *              EnumSystemCodePagesW   (KERNEL32.@)
 252  */
 253 BOOL WINAPI EnumSystemCodePagesW( CODEPAGE_ENUMPROCW lpfnCodePageEnum, DWORD flags )
 254 {
 255     const union cptable *table;
 256     WCHAR buffer[10], *p;
 257     int page, index = 0;
 258
 259     for (;;)
 260     {
 261         if (!(table = cp_enum_table( index++ ))) break;
 262         p = buffer + sizeof(buffer)/sizeof(WCHAR);
 263         *--p = 0;
 264         page = table->info.codepage;
 265         do
 266         {
 267             *--p = '0' + (page % 10);
 268             page /= 10;
 269         } while( page );
 270         if (!lpfnCodePageEnum( p )) break;
 271     }
 272     return TRUE;
 273 }
 274
 275
 276 /***********************************************************************
 277  *              MultiByteToWideChar   (KERNEL32.@)
 278  *
 279  * PARAMS
 280  *   page [in]    Codepage character set to convert from
 281  *   flags [in]   Character mapping flags
 282  *   src [in]     Source string buffer
 283  *   srclen [in]  Length of source string buffer
 284  *   dst [in]     Destination buffer
 285  *   dstlen [in]  Length of destination buffer
 286  *
 287  * NOTES
 288  *   The returned length includes the null terminator character.
 289  *
 290  * RETURNS
 291  *   Success: If dstlen > 0, number of characters written to destination
 292  *            buffer.  If dstlen == 0, number of characters needed to do
 293  *            conversion.
 294  *   Failure: 0. Occurs if not enough space is available.
 295  *
 296  * ERRORS
 297  *   ERROR_INSUFFICIENT_BUFFER
 298  *   ERROR_INVALID_PARAMETER
 299  *   ERROR_NO_UNICODE_TRANSLATION
 300  *
 301  */
 302 INT WINAPI MultiByteToWideChar( UINT page, DWORD flags, LPCSTR src, INT srclen,
 303                                 LPWSTR dst, INT dstlen )
 304 {
 305     const union cptable *table;
 306     int ret;
 307
 308     if (!src || (!dst && dstlen))
 309     {
 310         SetLastError( ERROR_INVALID_PARAMETER );
 311         return 0;
 312     }
 313
 314     if (srclen == -1) srclen = strlen(src) + 1;
 315
 316     if (flags & MB_USEGLYPHCHARS) FIXME("MB_USEGLYPHCHARS not supported\n");
 317
 318     switch(page)
 319     {
 320     case CP_UTF7:
 321         FIXME("UTF not supported\n");
 322         SetLastError( ERROR_CALL_NOT_IMPLEMENTED );
 323         return 0;
 324     case CP_UTF8:
 325         ret = utf8_mbstowcs( flags, src, srclen, dst, dstlen );
 326         break;
 327     default:
 328         if (!(table = get_codepage_table( page )))
 329         {
 330             SetLastError( ERROR_INVALID_PARAMETER );
 331             return 0;
 332         }
 333         ret = cp_mbstowcs( table, flags, src, srclen, dst, dstlen );
 334         break;
 335     }
 336
 337     if (ret < 0)
 338     {
 339         switch(ret)
 340         {
 341         case -1: SetLastError( ERROR_INSUFFICIENT_BUFFER ); break;
 342         case -2: SetLastError( ERROR_NO_UNICODE_TRANSLATION ); break;
 343         }
 344         ret = 0;
 345     }
 346     return ret;
 347 }
 348
 349
 350 /***********************************************************************
 351  *              WideCharToMultiByte   (KERNEL32.@)
 352  *
 353  * PARAMS
 354  *   page [in]    Codepage character set to convert to
 355  *   flags [in]   Character mapping flags
 356  *   src [in]     Source string buffer
 357  *   srclen [in]  Length of source string buffer
 358  *   dst [in]     Destination buffer
 359  *   dstlen [in]  Length of destination buffer
 360  *   defchar [in] Default character to use for conversion if no exact
 361  *                  conversion can be made
 362  *   used [out]   Set if default character was used in the conversion
 363  *
 364  * NOTES
 365  *   The returned length includes the null terminator character.
 366  *
 367  * RETURNS
 368  *   Success: If dstlen > 0, number of characters written to destination
 369  *            buffer.  If dstlen == 0, number of characters needed to do
 370  *            conversion.
 371  *   Failure: 0. Occurs if not enough space is available.
 372  *
 373  * ERRORS
 374  *   ERROR_INSUFFICIENT_BUFFER
 375  *   ERROR_INVALID_PARAMETER
 376  */
 377 INT WINAPI WideCharToMultiByte( UINT page, DWORD flags, LPCWSTR src, INT srclen,
 378                                 LPSTR dst, INT dstlen, LPCSTR defchar, BOOL *used )
 379 {
 380     const union cptable *table;
 381     int ret, used_tmp;
 382
 383     if (!src || (!dst && dstlen))
 384     {
 385         SetLastError( ERROR_INVALID_PARAMETER );
 386         return 0;
 387     }
 388
 389     if (srclen == -1) srclen = strlenW(src) + 1;
 390
 391     switch(page)
 392     {
 393     case CP_UTF7:
 394         FIXME("UTF-7 not supported\n");
 395         SetLastError( ERROR_CALL_NOT_IMPLEMENTED );
 396         return 0;
 397     case CP_UTF8:
 398         ret = utf8_wcstombs( src, srclen, dst, dstlen );
 399         break;
 400     default:
 401         if (!(table = get_codepage_table( page )))
 402         {
 403             SetLastError( ERROR_INVALID_PARAMETER );
 404             return 0;
 405         }
 406         ret = cp_wcstombs( table, flags, src, srclen, dst, dstlen,
 407                            defchar, used ? &used_tmp : NULL );
 408         if (used) *used = used_tmp;
 409         break;
 410     }
 411
 412     if (ret == -1)
 413     {
 414         SetLastError( ERROR_INSUFFICIENT_BUFFER );
 415         ret = 0;
 416     }
 417     return ret;
 418 }