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