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 <stdarg.h>
  23 #include <stdio.h>
  24 #include <stdlib.h>
  25 #include <string.h>
  26
  27 #include "windef.h"
  28 #include "winbase.h"
  29 #include "winerror.h"
  30 #include "winnls.h"
  31 #include "wine/unicode.h"
  32 #include "thread.h"
  33 #include "wine/debug.h"
  34
  35 WINE_DEFAULT_DEBUG_CHANNEL(string);
  36
  37 /* current code pages */
  38 static const union cptable *ansi_cptable;
  39 static const union cptable *oem_cptable;
  40 static const union cptable *mac_cptable;
  41 static const union cptable *unix_cptable;  /* NULL if UTF8 */
  42 static LCID default_lcid = MAKELCID( MAKELANGID(LANG_ENGLISH,SUBLANG_DEFAULT), SORT_DEFAULT );
  43
  44 /* setup default codepage info before we can get at the locale stuff */
  45 static void init_codepages(void)
  46 {
  47     ansi_cptable = wine_cp_get_table( 1252 );
  48     oem_cptable  = wine_cp_get_table( 437 );
  49     mac_cptable  = wine_cp_get_table( 10000 );
  50     unix_cptable  = wine_cp_get_table( 28591 );
  51     assert( ansi_cptable );
  52     assert( oem_cptable );
  53     assert( mac_cptable );
  54     assert( unix_cptable );
  55 }
  56
  57 /* find the table for a given codepage, handling CP_ACP etc. pseudo-codepages */
  58 static const union cptable *get_codepage_table( unsigned int codepage )
  59 {
  60     const union cptable *ret = NULL;
  61
  62     if (!ansi_cptable) init_codepages();
  63
  64     switch(codepage)
  65     {
  66     case CP_ACP:
  67         return ansi_cptable;
  68     case CP_OEMCP:
  69         return oem_cptable;
  70     case CP_MACCP:
  71         return mac_cptable;
  72     case CP_UTF7:
  73     case CP_UTF8:
  74         break;
  75     case CP_THREAD_ACP:
  76         if (!(codepage = NtCurrentTeb()->code_page)) return ansi_cptable;
  77         /* fall through */
  78     default:
  79         if (codepage == ansi_cptable->info.codepage) return ansi_cptable;
  80         if (codepage == oem_cptable->info.codepage) return oem_cptable;
  81         if (codepage == mac_cptable->info.codepage) return mac_cptable;
  82         ret = wine_cp_get_table( codepage );
  83         break;
  84     }
  85     return ret;
  86 }
  87
  88 /* initialize default code pages from locale info */
  89 /* FIXME: should be done in init_codepages, but it can't right now */
  90 /* since it needs KERNEL32 to be loaded for the locale info. */
  91 void CODEPAGE_Init( UINT ansi_cp, UINT oem_cp, UINT mac_cp, UINT unix_cp, LCID lcid )
  92 {
  93     extern void __wine_init_codepages( const union cptable *ansi_cp, const union cptable *oem_cp );
  94     const union cptable *table;
  95
  96     default_lcid = lcid;
  97     if (!ansi_cptable) init_codepages();  /* just in case */
  98
  99     if ((table = wine_cp_get_table( ansi_cp ))) ansi_cptable = table;
 100     if ((table = wine_cp_get_table( oem_cp ))) oem_cptable = table;
 101     if ((table = wine_cp_get_table( mac_cp ))) mac_cptable = table;
 102     if (unix_cp == CP_UTF8)
 103         unix_cptable = NULL;
 104     else if ((table = wine_cp_get_table( unix_cp )))
 105         unix_cptable = table;
 106
 107     __wine_init_codepages( ansi_cptable, oem_cptable );
 108
 109     TRACE( "ansi=%03d oem=%03d mac=%03d unix=%03d\n",
 110            ansi_cptable->info.codepage, oem_cptable->info.codepage,
 111            mac_cptable->info.codepage, unix_cp );
 112 }
 113
 114 /******************************************************************************
 115  *              GetACP   (KERNEL32.@)
 116  *
 117  * RETURNS
 118  *    Current ANSI code-page identifier, default if no current defined
 119  */
 120 UINT WINAPI GetACP(void)
 121 {
 122     if (!ansi_cptable) return 1252;
 123     return ansi_cptable->info.codepage;
 124 }
 125
 126
 127 /***********************************************************************
 128  *              GetOEMCP   (KERNEL32.@)
 129  */
 130 UINT WINAPI GetOEMCP(void)
 131 {
 132     if (!oem_cptable) return 437;
 133     return oem_cptable->info.codepage;
 134 }
 135
 136
 137 /***********************************************************************
 138  *           IsValidCodePage   (KERNEL32.@)
 139  */
 140 BOOL WINAPI IsValidCodePage( UINT codepage )
 141 {
 142     switch(codepage) {
 143     case CP_SYMBOL:
 144         return FALSE;
 145     case CP_UTF7:
 146     case CP_UTF8:
 147         return TRUE;
 148     default:
 149         return wine_cp_get_table( codepage ) != NULL;
 150     }
 151 }
 152
 153
 154 /***********************************************************************
 155  *              GetUserDefaultLangID (KERNEL32.@)
 156  */
 157 LANGID WINAPI GetUserDefaultLangID(void)
 158 {
 159     return LANGIDFROMLCID(default_lcid);
 160 }
 161
 162
 163 /***********************************************************************
 164  *              GetSystemDefaultLangID (KERNEL32.@)
 165  */
 166 LANGID WINAPI GetSystemDefaultLangID(void)
 167 {
 168     return GetUserDefaultLangID();
 169 }
 170
 171
 172 /***********************************************************************
 173  *              GetUserDefaultLCID (KERNEL32.@)
 174  */
 175 LCID WINAPI GetUserDefaultLCID(void)
 176 {
 177     return default_lcid;
 178 }
 179
 180
 181 /***********************************************************************
 182  *              GetSystemDefaultLCID (KERNEL32.@)
 183  */
 184 LCID WINAPI GetSystemDefaultLCID(void)
 185 {
 186     return GetUserDefaultLCID();
 187 }
 188
 189
 190 /***********************************************************************
 191  *              GetUserDefaultUILanguage (KERNEL32.@)
 192  */
 193 LANGID WINAPI GetUserDefaultUILanguage(void)
 194 {
 195     return GetUserDefaultLangID();
 196 }
 197
 198
 199 /***********************************************************************
 200  *              GetSystemDefaultUILanguage (KERNEL32.@)
 201  */
 202 LANGID WINAPI GetSystemDefaultUILanguage(void)
 203 {
 204     return GetSystemDefaultLangID();
 205 }
 206
 207
 208 /***********************************************************************
 209  *           IsDBCSLeadByteEx   (KERNEL32.@)
 210  *
 211  * Determine if a character is a lead byte in a given code page.
 212  *
 213  * PARAMS
 214  *  codepage [I] Code page for the test.
 215  *  testchar [I] Character to test
 216  *
 217  * RETURNS
 218  *  TRUE, if testchar is a lead byte in codepage,
 219  *  FALSE otherwise.
 220  */
 221 BOOL WINAPI IsDBCSLeadByteEx( UINT codepage, BYTE testchar )
 222 {
 223     const union cptable *table = get_codepage_table( codepage );
 224     return table && is_dbcs_leadbyte( table, testchar );
 225 }
 226
 227
 228 /***********************************************************************
 229  *           IsDBCSLeadByte   (KERNEL32.@)
 230  *           IsDBCSLeadByte   (KERNEL.207)
 231  *
 232  * Determine if a character is a lead byte.
 233  *
 234  * PARAMS
 235  *  testchar [I] Character to test
 236  *
 237  * RETURNS
 238  *  TRUE, if testchar is a lead byte in the Ansii code page,
 239  *  FALSE otherwise.
 240  */
 241 BOOL WINAPI IsDBCSLeadByte( BYTE testchar )
 242 {
 243     if (!ansi_cptable) return FALSE;
 244     return is_dbcs_leadbyte( ansi_cptable, testchar );
 245 }
 246
 247
 248 /***********************************************************************
 249  *           GetCPInfo   (KERNEL32.@)
 250  *
 251  * Get information about a code page.
 252  *
 253  * PARAMS
 254  *  codepage [I] Code page number
 255  *  cpinfo   [O] Destination for code page information
 256  *
 257  * RETURNS
 258  *  Success: TRUE. cpinfo is updated with the information about codepage.
 259  *  Failure: FALSE, if codepage is invalid or cpinfo is NULL.
 260  */
 261 BOOL WINAPI GetCPInfo( UINT codepage, LPCPINFO cpinfo )
 262 {
 263     const union cptable *table = get_codepage_table( codepage );
 264
 265     if (!table)
 266     {
 267         SetLastError( ERROR_INVALID_PARAMETER );
 268         return FALSE;
 269     }
 270     if (table->info.def_char & 0xff00)
 271     {
 272         cpinfo->DefaultChar[0] = table->info.def_char & 0xff00;
 273         cpinfo->DefaultChar[1] = table->info.def_char & 0x00ff;
 274     }
 275     else
 276     {
 277         cpinfo->DefaultChar[0] = table->info.def_char & 0xff;
 278         cpinfo->DefaultChar[1] = 0;
 279     }
 280     if ((cpinfo->MaxCharSize = table->info.char_size) == 2)
 281         memcpy( cpinfo->LeadByte, table->dbcs.lead_bytes, sizeof(cpinfo->LeadByte) );
 282     else
 283         cpinfo->LeadByte[0] = cpinfo->LeadByte[1] = 0;
 284
 285     return TRUE;
 286 }
 287
 288 /***********************************************************************
 289  *           GetCPInfoExA   (KERNEL32.@)
 290  *
 291  * Get extended information about a code page.
 292  *
 293  * PARAMS
 294  *  codepage [I] Code page number
 295  *  dwFlags  [I] Reserved, must to 0.
 296  *  cpinfo   [O] Destination for code page information
 297  *
 298  * RETURNS
 299  *  Success: TRUE. cpinfo is updated with the information about codepage.
 300  *  Failure: FALSE, if codepage is invalid or cpinfo is NULL.
 301  */
 302 BOOL WINAPI GetCPInfoExA( UINT codepage, DWORD dwFlags, LPCPINFOEXA cpinfo )
 303 {
 304     const union cptable *table = get_codepage_table( codepage );
 305
 306     if (!GetCPInfo( codepage, (LPCPINFO)cpinfo ))
 307       return FALSE;
 308
 309     cpinfo->CodePage = codepage;
 310     cpinfo->UnicodeDefaultChar = table->info.def_unicode_char;
 311     strcpy(cpinfo->CodePageName, table->info.name);
 312     return TRUE;
 313 }
 314
 315 /***********************************************************************
 316  *           GetCPInfoExW   (KERNEL32.@)
 317  *
 318  * Unicode version of GetCPInfoExA.
 319  */
 320 BOOL WINAPI GetCPInfoExW( UINT codepage, DWORD dwFlags, LPCPINFOEXW cpinfo )
 321 {
 322     const union cptable *table = get_codepage_table( codepage );
 323
 324     if (!GetCPInfo( codepage, (LPCPINFO)cpinfo ))
 325       return FALSE;
 326
 327     cpinfo->CodePage = codepage;
 328     cpinfo->UnicodeDefaultChar = table->info.def_unicode_char;
 329     MultiByteToWideChar( CP_ACP, 0, table->info.name, -1, cpinfo->CodePageName,
 330                          sizeof(cpinfo->CodePageName)/sizeof(WCHAR));
 331     return TRUE;
 332 }
 333
 334 /***********************************************************************
 335  *              EnumSystemCodePagesA   (KERNEL32.@)
 336  */
 337 BOOL WINAPI EnumSystemCodePagesA( CODEPAGE_ENUMPROCA lpfnCodePageEnum, DWORD flags )
 338 {
 339     const union cptable *table;
 340     char buffer[10];
 341     int index = 0;
 342
 343     for (;;)
 344     {
 345         if (!(table = wine_cp_enum_table( index++ ))) break;
 346         sprintf( buffer, "%d", table->info.codepage );
 347         if (!lpfnCodePageEnum( buffer )) break;
 348     }
 349     return TRUE;
 350 }
 351
 352
 353 /***********************************************************************
 354  *              EnumSystemCodePagesW   (KERNEL32.@)
 355  */
 356 BOOL WINAPI EnumSystemCodePagesW( CODEPAGE_ENUMPROCW lpfnCodePageEnum, DWORD flags )
 357 {
 358     const union cptable *table;
 359     WCHAR buffer[10], *p;
 360     int page, index = 0;
 361
 362     for (;;)
 363     {
 364         if (!(table = wine_cp_enum_table( index++ ))) break;
 365         p = buffer + sizeof(buffer)/sizeof(WCHAR);
 366         *--p = 0;
 367         page = table->info.codepage;
 368         do
 369         {
 370             *--p = '0' + (page % 10);
 371             page /= 10;
 372         } while( page );
 373         if (!lpfnCodePageEnum( p )) break;
 374     }
 375     return TRUE;
 376 }
 377
 378
 379 /***********************************************************************
 380  *              MultiByteToWideChar   (KERNEL32.@)
 381  *
 382  * Convert a multibyte character string into a Unicode string.
 383  *
 384  * PARAMS
 385  *   page   [I] Codepage character set to convert from
 386  *   flags  [I] Character mapping flags
 387  *   src    [I] Source string buffer
 388  *   srclen [I] Length of src, or -1 if src is NUL terminated
 389  *   dst    [O] Destination buffer
 390  *   dstlen [I] Length of dst, or 0 to compute the required length
 391  *
 392  * RETURNS
 393  *   Success: If dstlen > 0, the number of characters written to dst.
 394  *            If dstlen == 0, the number of characters needed to perform the
 395  *            conversion. In both cases the count includes the terminating NUL.
 396  *   Failure: 0. Use GetLastError() to determine the cause. Possible errors are
 397  *            ERROR_INSUFFICIENT_BUFFER, if not enough space is available in dst
 398  *            and dstlen != 0; ERROR_INVALID_PARAMETER,  if an invalid parameter
 399  *            is passed, and ERROR_NO_UNICODE_TRANSLATION if no translation is
 400  *            possible for src.
 401  */
 402 INT WINAPI MultiByteToWideChar( UINT page, DWORD flags, LPCSTR src, INT srclen,
 403                                 LPWSTR dst, INT dstlen )
 404 {
 405     const union cptable *table;
 406     int ret;
 407
 408     if (!src || (!dst && dstlen))
 409     {
 410         SetLastError( ERROR_INVALID_PARAMETER );
 411         return 0;
 412     }
 413
 414     if (srclen < 0) srclen = strlen(src) + 1;
 415
 416     if (flags & MB_USEGLYPHCHARS) FIXME("MB_USEGLYPHCHARS not supported\n");
 417
 418     switch(page)
 419     {
 420     case CP_UTF7:
 421         FIXME("UTF-7 not supported\n");
 422         SetLastError( ERROR_CALL_NOT_IMPLEMENTED );
 423         return 0;
 424     case CP_UNIXCP:
 425         if (unix_cptable)
 426         {
 427             ret = wine_cp_mbstowcs( unix_cptable, flags, src, srclen, dst, dstlen );
 428             break;
 429         }
 430         /* fall through */
 431     case CP_UTF8:
 432         ret = wine_utf8_mbstowcs( flags, src, srclen, dst, dstlen );
 433         break;
 434     default:
 435         if (!(table = get_codepage_table( page )))
 436         {
 437             SetLastError( ERROR_INVALID_PARAMETER );
 438             return 0;
 439         }
 440         ret = wine_cp_mbstowcs( table, flags, src, srclen, dst, dstlen );
 441         break;
 442     }
 443
 444     if (ret < 0)
 445     {
 446         switch(ret)
 447         {
 448         case -1: SetLastError( ERROR_INSUFFICIENT_BUFFER ); break;
 449         case -2: SetLastError( ERROR_NO_UNICODE_TRANSLATION ); break;
 450         }
 451         ret = 0;
 452     }
 453     return ret;
 454 }
 455
 456
 457 /***********************************************************************
 458  *              WideCharToMultiByte   (KERNEL32.@)
 459  *
 460  * Convert a Unicode character string into a multibyte string.
 461  *
 462  * PARAMS
 463  *   page    [I] Code page character set to convert to
 464  *   flags   [I] Mapping Flags (MB_ constants from "winnls.h").
 465  *   src     [I] Source string buffer
 466  *   srclen  [I] Length of src, or -1 if src is NUL terminated
 467  *   dst     [O] Destination buffer
 468  *   dstlen  [I] Length of dst, or 0 to compute the required length
 469  *   defchar [I] Default character to use for conversion if no exact
 470  *                  conversion can be made
 471  *   used    [O] Set if default character was used in the conversion
 472  *
 473  * RETURNS
 474  *   Success: If dstlen > 0, the number of characters written to dst.
 475  *            If dstlen == 0, number of characters needed to perform the
 476  *            conversion. In both cases the count includes the terminating NUL.
 477  *   Failure: 0. Use GetLastError() to determine the cause. Possible errors are
 478  *            ERROR_INSUFFICIENT_BUFFER, if not enough space is available in dst
 479  *            and dstlen != 0, and ERROR_INVALID_PARAMETER, if an invalid
 480  *            parameter was given.
 481  */
 482 INT WINAPI WideCharToMultiByte( UINT page, DWORD flags, LPCWSTR src, INT srclen,
 483                                 LPSTR dst, INT dstlen, LPCSTR defchar, BOOL *used )
 484 {
 485     const union cptable *table;
 486     int ret, used_tmp;
 487
 488     if (!src || (!dst && dstlen))
 489     {
 490         SetLastError( ERROR_INVALID_PARAMETER );
 491         return 0;
 492     }
 493
 494     if (srclen < 0) srclen = strlenW(src) + 1;
 495
 496     switch(page)
 497     {
 498     case CP_UTF7:
 499         FIXME("UTF-7 not supported\n");
 500         SetLastError( ERROR_CALL_NOT_IMPLEMENTED );
 501         return 0;
 502     case CP_UNIXCP:
 503         if (unix_cptable)
 504         {
 505             ret = wine_cp_wcstombs( unix_cptable, flags, src, srclen, dst, dstlen,
 506                                     defchar, used ? &used_tmp : NULL );
 507             break;
 508         }
 509         /* fall through */
 510     case CP_UTF8:
 511         ret = wine_utf8_wcstombs( src, srclen, dst, dstlen );
 512         break;
 513     default:
 514         if (!(table = get_codepage_table( page )))
 515         {
 516             SetLastError( ERROR_INVALID_PARAMETER );
 517             return 0;
 518         }
 519         ret = wine_cp_wcstombs( table, flags, src, srclen, dst, dstlen,
 520                                 defchar, used ? &used_tmp : NULL );
 521         if (used) *used = used_tmp;
 522         break;
 523     }
 524
 525     if (ret == -1)
 526     {
 527         SetLastError( ERROR_INSUFFICIENT_BUFFER );
 528         ret = 0;
 529     }
 530     return ret;
 531 }