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 "wine/debug.h"
  31
  32 WINE_DEFAULT_DEBUG_CHANNEL(string);
  33
  34 /* current code pages */
  35 static const union cptable *ansi_cptable;
  36 static const union cptable *oem_cptable;
  37 static const union cptable *mac_cptable;
  38
  39 /* retrieve a code page table from the locale info */
  40 static const union cptable *get_locale_cp( LCID lcid, LCTYPE type )
  41 {
  42     const union cptable *table = NULL;
  43     char buf[32];
  44
  45     if (GetLocaleInfoA( lcid, type, buf, sizeof(buf) )) table = cp_get_table( atoi(buf) );
  46     return table;
  47 }
  48
  49 /* setup default codepage info before we can get at the locale stuff */
  50 static void init_codepages(void)
  51 {
  52     ansi_cptable = cp_get_table( 1252 );
  53     oem_cptable  = cp_get_table( 437 );
  54     mac_cptable  = cp_get_table( 10000 );
  55     assert( ansi_cptable );
  56     assert( oem_cptable );
  57     assert( mac_cptable );
  58 }
  59
  60 /* find the table for a given codepage, handling CP_ACP etc. pseudo-codepages */
  61 static const union cptable *get_codepage_table( unsigned int codepage )
  62 {
  63     const union cptable *ret = NULL;
  64
  65     if (!ansi_cptable) init_codepages();
  66
  67     switch(codepage)
  68     {
  69     case CP_ACP:        return ansi_cptable;
  70     case CP_OEMCP:      return oem_cptable;
  71     case CP_MACCP:      return mac_cptable;
  72     case CP_THREAD_ACP: return get_locale_cp( GetThreadLocale(), LOCALE_IDEFAULTANSICODEPAGE );
  73     case CP_UTF7:
  74     case CP_UTF8:
  75         break;
  76     default:
  77         if (codepage == ansi_cptable->info.codepage) return ansi_cptable;
  78         if (codepage == oem_cptable->info.codepage) return oem_cptable;
  79         if (codepage == mac_cptable->info.codepage) return mac_cptable;
  80         ret = cp_get_table( codepage );
  81         break;
  82     }
  83     return ret;
  84 }
  85
  86 /* initialize default code pages from locale info */
  87 /* FIXME: should be done in init_codepages, but it can't right now */
  88 /* since it needs KERNEL32 to be loaded for the locale info. */
  89 void CODEPAGE_Init(void)
  90 {
  91     extern void __wine_init_codepages( const union cptable *ansi, const union cptable *oem );
  92     const union cptable *table;
  93     LCID lcid = GetUserDefaultLCID();
  94
  95     if (!ansi_cptable) init_codepages();  /* just in case */
  96
  97     if ((table = get_locale_cp( lcid, LOCALE_IDEFAULTANSICODEPAGE ))) ansi_cptable = table;
  98     if ((table = get_locale_cp( lcid, LOCALE_IDEFAULTMACCODEPAGE ))) mac_cptable = table;
  99     if ((table = get_locale_cp( lcid, LOCALE_IDEFAULTCODEPAGE ))) oem_cptable = table;
 100     __wine_init_codepages( ansi_cptable, oem_cptable );
 101
 102     TRACE( "ansi=%03d oem=%03d mac=%03d\n", ansi_cptable->info.codepage,
 103            oem_cptable->info.codepage, mac_cptable->info.codepage );
 104 }
 105
 106 /******************************************************************************
 107  *              GetACP   (KERNEL32.@)
 108  *
 109  * RETURNS
 110  *    Current ANSI code-page identifier, default if no current defined
 111  */
 112 UINT WINAPI GetACP(void)
 113 {
 114     if (!ansi_cptable) init_codepages();
 115     return ansi_cptable->info.codepage;
 116 }
 117
 118
 119 /***********************************************************************
 120  *              GetOEMCP   (KERNEL32.@)
 121  */
 122 UINT WINAPI GetOEMCP(void)
 123 {
 124     if (!oem_cptable) init_codepages();
 125     return oem_cptable->info.codepage;
 126 }
 127
 128
 129 /***********************************************************************
 130  *           IsValidCodePage   (KERNEL32.@)
 131  */
 132 BOOL WINAPI IsValidCodePage( UINT codepage )
 133 {
 134     return cp_get_table( codepage ) != NULL;
 135 }
 136
 137
 138 /***********************************************************************
 139  *           IsDBCSLeadByteEx   (KERNEL32.@)
 140  */
 141 BOOL WINAPI IsDBCSLeadByteEx( UINT codepage, BYTE testchar )
 142 {
 143     const union cptable *table = get_codepage_table( codepage );
 144     return table && is_dbcs_leadbyte( table, testchar );
 145 }
 146
 147
 148 /***********************************************************************
 149  *           IsDBCSLeadByte   (KERNEL32.@)
 150  *           IsDBCSLeadByte   (KERNEL.207)
 151  */
 152 BOOL WINAPI IsDBCSLeadByte( BYTE testchar )
 153 {
 154     if (!ansi_cptable) init_codepages();
 155     return is_dbcs_leadbyte( ansi_cptable, testchar );
 156 }
 157
 158
 159 /***********************************************************************
 160  *           GetCPInfo   (KERNEL32.@)
 161  */
 162 BOOL WINAPI GetCPInfo( UINT codepage, LPCPINFO cpinfo )
 163 {
 164     const union cptable *table = get_codepage_table( codepage );
 165
 166     if (!table)
 167     {
 168         SetLastError( ERROR_INVALID_PARAMETER );
 169         return FALSE;
 170     }
 171     if (table->info.def_char & 0xff00)
 172     {
 173         cpinfo->DefaultChar[0] = table->info.def_char & 0xff00;
 174         cpinfo->DefaultChar[1] = table->info.def_char & 0x00ff;
 175     }
 176     else
 177     {
 178         cpinfo->DefaultChar[0] = table->info.def_char & 0xff;
 179         cpinfo->DefaultChar[1] = 0;
 180     }
 181     if ((cpinfo->MaxCharSize = table->info.char_size) == 2)
 182         memcpy( cpinfo->LeadByte, table->dbcs.lead_bytes, sizeof(cpinfo->LeadByte) );
 183     else
 184         cpinfo->LeadByte[0] = cpinfo->LeadByte[1] = 0;
 185
 186     return TRUE;
 187 }
 188
 189
 190 /***********************************************************************
 191  *              EnumSystemCodePagesA   (KERNEL32.@)
 192  */
 193 BOOL WINAPI EnumSystemCodePagesA( CODEPAGE_ENUMPROCA lpfnCodePageEnum, DWORD flags )
 194 {
 195     const union cptable *table;
 196     char buffer[10];
 197     int index = 0;
 198
 199     for (;;)
 200     {
 201         if (!(table = cp_enum_table( index++ ))) break;
 202         sprintf( buffer, "%d", table->info.codepage );
 203         if (!lpfnCodePageEnum( buffer )) break;
 204     }
 205     return TRUE;
 206 }
 207
 208
 209 /***********************************************************************
 210  *              EnumSystemCodePagesW   (KERNEL32.@)
 211  */
 212 BOOL WINAPI EnumSystemCodePagesW( CODEPAGE_ENUMPROCW lpfnCodePageEnum, DWORD flags )
 213 {
 214     const union cptable *table;
 215     WCHAR buffer[10], *p;
 216     int page, index = 0;
 217
 218     for (;;)
 219     {
 220         if (!(table = cp_enum_table( index++ ))) break;
 221         p = buffer + sizeof(buffer)/sizeof(WCHAR);
 222         *--p = 0;
 223         page = table->info.codepage;
 224         do
 225         {
 226             *--p = '0' + (page % 10);
 227             page /= 10;
 228         } while( page );
 229         if (!lpfnCodePageEnum( p )) break;
 230     }
 231     return TRUE;
 232 }
 233
 234
 235 /***********************************************************************
 236  *              MultiByteToWideChar   (KERNEL32.@)
 237  *
 238  * PARAMS
 239  *   page [in]    Codepage character set to convert from
 240  *   flags [in]   Character mapping flags
 241  *   src [in]     Source string buffer
 242  *   srclen [in]  Length of source string buffer
 243  *   dst [in]     Destination buffer
 244  *   dstlen [in]  Length of destination buffer
 245  *
 246  * NOTES
 247  *   The returned length includes the null terminator character.
 248  *
 249  * RETURNS
 250  *   Success: If dstlen > 0, number of characters written to destination
 251  *            buffer.  If dstlen == 0, number of characters needed to do
 252  *            conversion.
 253  *   Failure: 0. Occurs if not enough space is available.
 254  *
 255  * ERRORS
 256  *   ERROR_INSUFFICIENT_BUFFER
 257  *   ERROR_INVALID_PARAMETER
 258  *   ERROR_NO_UNICODE_TRANSLATION
 259  *
 260  */
 261 INT WINAPI MultiByteToWideChar( UINT page, DWORD flags, LPCSTR src, INT srclen,
 262                                 LPWSTR dst, INT dstlen )
 263 {
 264     const union cptable *table;
 265     int ret;
 266
 267     if (!src || (!dst && dstlen))
 268     {
 269         SetLastError( ERROR_INVALID_PARAMETER );
 270         return 0;
 271     }
 272
 273     if (srclen == -1) srclen = strlen(src) + 1;
 274
 275     if (flags & MB_USEGLYPHCHARS) FIXME("MB_USEGLYPHCHARS not supported\n");
 276
 277     switch(page)
 278     {
 279     case CP_UTF7:
 280         FIXME("UTF not supported\n");
 281         SetLastError( ERROR_CALL_NOT_IMPLEMENTED );
 282         return 0;
 283     case CP_UTF8:
 284         ret = utf8_mbstowcs( flags, src, srclen, dst, dstlen );
 285         break;
 286     default:
 287         if (!(table = get_codepage_table( page )))
 288         {
 289             SetLastError( ERROR_INVALID_PARAMETER );
 290             return 0;
 291         }
 292         ret = cp_mbstowcs( table, flags, src, srclen, dst, dstlen );
 293         break;
 294     }
 295
 296     if (ret < 0)
 297     {
 298         switch(ret)
 299         {
 300         case -1: SetLastError( ERROR_INSUFFICIENT_BUFFER ); break;
 301         case -2: SetLastError( ERROR_NO_UNICODE_TRANSLATION ); break;
 302         }
 303         ret = 0;
 304     }
 305     return ret;
 306 }
 307
 308
 309 /***********************************************************************
 310  *              WideCharToMultiByte   (KERNEL32.@)
 311  *
 312  * PARAMS
 313  *   page [in]    Codepage character set to convert to
 314  *   flags [in]   Character mapping flags
 315  *   src [in]     Source string buffer
 316  *   srclen [in]  Length of source string buffer
 317  *   dst [in]     Destination buffer
 318  *   dstlen [in]  Length of destination buffer
 319  *   defchar [in] Default character to use for conversion if no exact
 320  *                  conversion can be made
 321  *   used [out]   Set if default character was used in the conversion
 322  *
 323  * NOTES
 324  *   The returned length includes the null terminator character.
 325  *
 326  * RETURNS
 327  *   Success: If dstlen > 0, number of characters written to destination
 328  *            buffer.  If dstlen == 0, number of characters needed to do
 329  *            conversion.
 330  *   Failure: 0. Occurs if not enough space is available.
 331  *
 332  * ERRORS
 333  *   ERROR_INSUFFICIENT_BUFFER
 334  *   ERROR_INVALID_PARAMETER
 335  */
 336 INT WINAPI WideCharToMultiByte( UINT page, DWORD flags, LPCWSTR src, INT srclen,
 337                                 LPSTR dst, INT dstlen, LPCSTR defchar, BOOL *used )
 338 {
 339     const union cptable *table;
 340     int ret, used_tmp;
 341
 342     if (!src || (!dst && dstlen))
 343     {
 344         SetLastError( ERROR_INVALID_PARAMETER );
 345         return 0;
 346     }
 347
 348     if (srclen == -1) srclen = strlenW(src) + 1;
 349
 350     switch(page)
 351     {
 352     case CP_UTF7:
 353         FIXME("UTF-7 not supported\n");
 354         SetLastError( ERROR_CALL_NOT_IMPLEMENTED );
 355         return 0;
 356     case CP_UTF8:
 357         ret = utf8_wcstombs( src, srclen, dst, dstlen );
 358         break;
 359     default:
 360         if (!(table = get_codepage_table( page )))
 361         {
 362             SetLastError( ERROR_INVALID_PARAMETER );
 363             return 0;
 364         }
 365         ret = cp_wcstombs( table, flags, src, srclen, dst, dstlen,
 366                            defchar, used ? &used_tmp : NULL );
 367         if (used) *used = used_tmp;
 368         break;
 369     }
 370
 371     if (ret == -1)
 372     {
 373         SetLastError( ERROR_INSUFFICIENT_BUFFER );
 374         ret = 0;
 375     }
 376     return ret;
 377 }
 378
 379
 380 /******************************************************************************
 381  *              GetStringTypeW   (KERNEL32.@)
 382  *
 383  */
 384 BOOL WINAPI GetStringTypeW( DWORD type, LPCWSTR src, INT count, LPWORD chartype )
 385 {
 386     if (count == -1) count = strlenW(src) + 1;
 387     switch(type)
 388     {
 389     case CT_CTYPE1:
 390         while (count--) *chartype++ = get_char_typeW( *src++ ) & 0xfff;
 391         break;
 392     case CT_CTYPE2:
 393         while (count--) *chartype++ = get_char_typeW( *src++ ) >> 12;
 394         break;
 395     case CT_CTYPE3:
 396         FIXME("CT_CTYPE3 not supported.\n");
 397     default:
 398         SetLastError( ERROR_INVALID_PARAMETER );
 399         return FALSE;
 400     }
 401     return TRUE;
 402 }
 403
 404
 405 /******************************************************************************
 406  *              GetStringTypeExW   (KERNEL32.@)
 407  */
 408 BOOL WINAPI GetStringTypeExW( LCID locale, DWORD type, LPCWSTR src, INT count, LPWORD chartype )
 409 {
 410     /* locale is ignored for Unicode */
 411     return GetStringTypeW( type, src, count, chartype );
 412 }
 413
 414 /******************************************************************************
 415  *              GetStringTypeA  [KERNEL32.@]
 416  */
 417 BOOL WINAPI GetStringTypeA(LCID locale, DWORD type, LPCSTR src, INT count, LPWORD chartype)
 418 {
 419     char buf[20];
 420     UINT cp;
 421     INT countW;
 422     LPWSTR srcW;
 423     BOOL ret = FALSE;
 424
 425     if(count == -1) count = strlen(src) + 1;
 426
 427     if(!GetLocaleInfoA(locale, LOCALE_IDEFAULTANSICODEPAGE | LOCALE_NOUSEROVERRIDE,
 428                    buf, sizeof(buf)))
 429     {
 430         FIXME("For locale %04lx using current ANSI code page\n", locale);
 431         cp = GetACP();
 432     }
 433     else
 434         cp = atoi(buf);
 435
 436     countW = MultiByteToWideChar(cp, 0, src, count, NULL, 0);
 437     if((srcW = HeapAlloc(GetProcessHeap(), 0, countW * sizeof(WCHAR))))
 438     {
 439         MultiByteToWideChar(cp, 0, src, count, srcW, countW);
 440         ret = GetStringTypeW(type, srcW, count, chartype);
 441         HeapFree(GetProcessHeap(), 0, srcW);
 442     }
 443     return ret;
 444 }
 445
 446 /******************************************************************************
 447  *              GetStringTypeExA        [KERNEL32.@]
 448  */
 449 BOOL WINAPI GetStringTypeExA(LCID locale, DWORD type, LPCSTR src, INT count, LPWORD chartype)
 450 {
 451     return GetStringTypeA(locale, type, src, count, chartype);
 452 }