libs/unicode/mbtowc.c

   1 /*
   2  * MultiByteToWideChar implementation
   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 <string.h>
  22
  23 #include "wine/unicode.h"
  24
  25 /* get the decomposition of a Unicode char */
  26 int get_decomposition( WCHAR src, WCHAR *dst, unsigned int dstlen )
  27 {
  28     extern const WCHAR unicode_decompose_table[];
  29     const WCHAR *ptr = unicode_decompose_table;
  30     int res;
  31
  32     *dst = src;
  33     ptr = unicode_decompose_table + ptr[src >> 8];
  34     ptr = unicode_decompose_table + ptr[(src >> 4) & 0x0f] + 2 * (src & 0x0f);
  35     if (!*ptr) return 1;
  36     if (dstlen <= 1) return 0;
  37     /* apply the decomposition recursively to the first char */
  38     if ((res = get_decomposition( *ptr, dst, dstlen-1 ))) dst[res++] = ptr[1];
  39     return res;
  40 }
  41
  42 /* check src string for invalid chars; return non-zero if invalid char found */
  43 static inline int check_invalid_chars_sbcs( const struct sbcs_table *table,
  44                                             const unsigned char *src, unsigned int srclen )
  45 {
  46     const WCHAR * const cp2uni = table->cp2uni;
  47     while (srclen)
  48     {
  49         if (cp2uni[*src] == table->info.def_unicode_char && *src != table->info.def_char)
  50             break;
  51         src++;
  52         srclen--;
  53     }
  54     return srclen;
  55 }
  56
  57 /* mbstowcs for single-byte code page */
  58 /* all lengths are in characters, not bytes */
  59 static inline int mbstowcs_sbcs( const struct sbcs_table *table,
  60                                  const unsigned char *src, unsigned int srclen,
  61                                  WCHAR *dst, unsigned int dstlen )
  62 {
  63     const WCHAR * const cp2uni = table->cp2uni;
  64     int ret = srclen;
  65
  66     if (dstlen < srclen)
  67     {
  68         /* buffer too small: fill it up to dstlen and return error */
  69         srclen = dstlen;
  70         ret = -1;
  71     }
  72
  73     for (;;)
  74     {
  75         switch(srclen)
  76         {
  77         default:
  78         case 16: dst[15] = cp2uni[src[15]];
  79         case 15: dst[14] = cp2uni[src[14]];
  80         case 14: dst[13] = cp2uni[src[13]];
  81         case 13: dst[12] = cp2uni[src[12]];
  82         case 12: dst[11] = cp2uni[src[11]];
  83         case 11: dst[10] = cp2uni[src[10]];
  84         case 10: dst[9]  = cp2uni[src[9]];
  85         case 9:  dst[8]  = cp2uni[src[8]];
  86         case 8:  dst[7]  = cp2uni[src[7]];
  87         case 7:  dst[6]  = cp2uni[src[6]];
  88         case 6:  dst[5]  = cp2uni[src[5]];
  89         case 5:  dst[4]  = cp2uni[src[4]];
  90         case 4:  dst[3]  = cp2uni[src[3]];
  91         case 3:  dst[2]  = cp2uni[src[2]];
  92         case 2:  dst[1]  = cp2uni[src[1]];
  93         case 1:  dst[0]  = cp2uni[src[0]];
  94         case 0: break;
  95         }
  96         if (srclen < 16) return ret;
  97         dst += 16;
  98         src += 16;
  99         srclen -= 16;
 100     }
 101 }
 102
 103 /* mbstowcs for single-byte code page with char decomposition */
 104 static int mbstowcs_sbcs_decompose( const struct sbcs_table *table,
 105                                     const unsigned char *src, unsigned int srclen,
 106                                     WCHAR *dst, unsigned int dstlen )
 107 {
 108     const WCHAR * const cp2uni = table->cp2uni;
 109     unsigned int len;
 110
 111     if (!dstlen)  /* compute length */
 112     {
 113         WCHAR dummy[4]; /* no decomposition is larger than 4 chars */
 114         for (len = 0; srclen; srclen--, src++)
 115             len += get_decomposition( cp2uni[*src], dummy, 4 );
 116         return len;
 117     }
 118
 119     for (len = dstlen; srclen && len; srclen--, src++)
 120     {
 121         int res = get_decomposition( cp2uni[*src], dst, len );
 122         if (!res) break;
 123         len -= res;
 124         dst += res;
 125     }
 126     if (srclen) return -1;  /* overflow */
 127     return dstlen - len;
 128 }
 129
 130 /* query necessary dst length for src string */
 131 static inline int get_length_dbcs( const struct dbcs_table *table,
 132                                    const unsigned char *src, unsigned int srclen )
 133 {
 134     const unsigned char * const cp2uni_lb = table->cp2uni_leadbytes;
 135     int len;
 136
 137     for (len = 0; srclen; srclen--, src++, len++)
 138     {
 139         if (cp2uni_lb[*src])
 140         {
 141             if (!--srclen) break;  /* partial char, ignore it */
 142             src++;
 143         }
 144     }
 145     return len;
 146 }
 147
 148 /* check src string for invalid chars; return non-zero if invalid char found */
 149 static inline int check_invalid_chars_dbcs( const struct dbcs_table *table,
 150                                             const unsigned char *src, unsigned int srclen )
 151 {
 152     const WCHAR * const cp2uni = table->cp2uni;
 153     const unsigned char * const cp2uni_lb = table->cp2uni_leadbytes;
 154
 155     while (srclen)
 156     {
 157         unsigned char off = cp2uni_lb[*src];
 158         if (off)  /* multi-byte char */
 159         {
 160             if (srclen == 1) break;  /* partial char, error */
 161             if (cp2uni[(off << 8) + src[1]] == table->info.def_unicode_char &&
 162                 ((src[0] << 8) | src[1]) != table->info.def_char) break;
 163             src++;
 164             srclen--;
 165         }
 166         else if (cp2uni[*src] == table->info.def_unicode_char &&
 167                  *src != table->info.def_char) break;
 168         src++;
 169         srclen--;
 170     }
 171     return srclen;
 172 }
 173
 174 /* mbstowcs for double-byte code page */
 175 /* all lengths are in characters, not bytes */
 176 static inline int mbstowcs_dbcs( const struct dbcs_table *table,
 177                                  const unsigned char *src, unsigned int srclen,
 178                                  WCHAR *dst, unsigned int dstlen )
 179 {
 180     const WCHAR * const cp2uni = table->cp2uni;
 181     const unsigned char * const cp2uni_lb = table->cp2uni_leadbytes;
 182     unsigned int len;
 183
 184     if (!dstlen) return get_length_dbcs( table, src, srclen );
 185
 186     for (len = dstlen; srclen && len; len--, srclen--, src++, dst++)
 187     {
 188         unsigned char off = cp2uni_lb[*src];
 189         if (off)
 190         {
 191             if (!--srclen) break;  /* partial char, ignore it */
 192             src++;
 193             *dst = cp2uni[(off << 8) + *src];
 194         }
 195         else *dst = cp2uni[*src];
 196     }
 197     if (srclen) return -1;  /* overflow */
 198     return dstlen - len;
 199 }
 200
 201
 202 /* mbstowcs for double-byte code page with character decomposition */
 203 static int mbstowcs_dbcs_decompose( const struct dbcs_table *table,
 204                                     const unsigned char *src, unsigned int srclen,
 205                                     WCHAR *dst, unsigned int dstlen )
 206 {
 207     const WCHAR * const cp2uni = table->cp2uni;
 208     const unsigned char * const cp2uni_lb = table->cp2uni_leadbytes;
 209     unsigned int len;
 210     WCHAR ch;
 211     int res;
 212
 213     if (!dstlen)  /* compute length */
 214     {
 215         WCHAR dummy[4]; /* no decomposition is larger than 4 chars */
 216         for (len = 0; srclen; srclen--, src++)
 217         {
 218             unsigned char off = cp2uni_lb[*src];
 219             if (off)
 220             {
 221                 if (!--srclen) break;  /* partial char, ignore it */
 222                 src++;
 223                 ch = cp2uni[(off << 8) + *src];
 224             }
 225             else ch = cp2uni[*src];
 226             len += get_decomposition( ch, dummy, 4 );
 227         }
 228         return len;
 229     }
 230
 231     for (len = dstlen; srclen && len; srclen--, src++)
 232     {
 233         unsigned char off = cp2uni_lb[*src];
 234         if (off)
 235         {
 236             if (!--srclen) break;  /* partial char, ignore it */
 237             src++;
 238             ch = cp2uni[(off << 8) + *src];
 239         }
 240         else ch = cp2uni[*src];
 241         if (!(res = get_decomposition( ch, dst, len ))) break;
 242         dst += res;
 243         len -= res;
 244     }
 245     if (srclen) return -1;  /* overflow */
 246     return dstlen - len;
 247 }
 248
 249
 250 /* return -1 on dst buffer overflow, -2 on invalid input char */
 251 int wine_cp_mbstowcs( const union cptable *table, int flags,
 252                       const char *src, int srclen,
 253                       WCHAR *dst, int dstlen )
 254 {
 255     if (table->info.char_size == 1)
 256     {
 257         if (flags & MB_ERR_INVALID_CHARS)
 258         {
 259             if (check_invalid_chars_sbcs( &table->sbcs, src, srclen )) return -2;
 260         }
 261         if (!(flags & MB_COMPOSITE))
 262         {
 263             if (!dstlen) return srclen;
 264             return mbstowcs_sbcs( &table->sbcs, src, srclen, dst, dstlen );
 265         }
 266         return mbstowcs_sbcs_decompose( &table->sbcs, src, srclen, dst, dstlen );
 267     }
 268     else /* mbcs */
 269     {
 270         if (flags & MB_ERR_INVALID_CHARS)
 271         {
 272             if (check_invalid_chars_dbcs( &table->dbcs, src, srclen )) return -2;
 273         }
 274         if (!(flags & MB_COMPOSITE))
 275             return mbstowcs_dbcs( &table->dbcs, src, srclen, dst, dstlen );
 276         else
 277             return mbstowcs_dbcs_decompose( &table->dbcs, src, srclen, dst, dstlen );
 278     }
 279 }
 280
 281 /* CP_SYMBOL implementation */
 282 /* return -1 on dst buffer overflow */
 283 int wine_cpsymbol_mbstowcs( const char *src, int srclen, WCHAR *dst, int dstlen)
 284 {
 285     int len, i;
 286     if( dstlen == 0) return srclen;
 287     len = dstlen > srclen ? srclen : dstlen;
 288     for( i = 0; i < len; i++)
 289     {
 290         unsigned char c = src [ i ];
 291         if( c < 0x20 )
 292             dst[i] = c;
 293         else
 294             dst[i] = c + 0xf000;
 295     }
 296     if( srclen > len) return -1;
 297     return len;
 298 }