lib/unistr/u8-mbtouc-aux.c

   1 /* Conversion UTF-8 to UCS-4.
   2    Copyright (C) 2001-2002, 2006-2007, 2009-2010 Free Software Foundation, Inc.
   3    Written by Bruno Haible <bruno@clisp.org>, 2001.
   4
   5    This program is free software: you can redistribute it and/or modify it
   6    under the terms of the GNU Lesser General Public License as published
   7    by the Free Software Foundation; either version 3 of the License, or
   8    (at your option) any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public License
  16    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  17
  18 #include <config.h>
  19
  20 /* Specification.  */
  21 #include "unistr.h"
  22
  23 #if defined IN_LIBUNISTRING || HAVE_INLINE
  24
  25 int
  26 u8_mbtouc_aux (ucs4_t *puc, const uint8_t *s, size_t n)
  27 {
  28   uint8_t c = *s;
  29
  30   if (c >= 0xc2)
  31     {
  32       if (c < 0xe0)
  33         {
  34           if (n >= 2)
  35             {
  36               if ((s[1] ^ 0x80) < 0x40)
  37                 {
  38                   *puc = ((unsigned int) (c & 0x1f) << 6)
  39                          | (unsigned int) (s[1] ^ 0x80);
  40                   return 2;
  41                 }
  42               /* invalid multibyte character */
  43             }
  44           else
  45             {
  46               /* incomplete multibyte character */
  47               *puc = 0xfffd;
  48               return 1;
  49             }
  50         }
  51       else if (c < 0xf0)
  52         {
  53           if (n >= 3)
  54             {
  55               if ((s[1] ^ 0x80) < 0x40)
  56                 {
  57                   if ((s[2] ^ 0x80) < 0x40)
  58                     {
  59                       if ((c >= 0xe1 || s[1] >= 0xa0)
  60                           && (c != 0xed || s[1] < 0xa0))
  61                         {
  62                           *puc = ((unsigned int) (c & 0x0f) << 12)
  63                                  | ((unsigned int) (s[1] ^ 0x80) << 6)
  64                                  | (unsigned int) (s[2] ^ 0x80);
  65                           return 3;
  66                         }
  67                       /* invalid multibyte character */
  68                       *puc = 0xfffd;
  69                       return 3;
  70                     }
  71                   /* invalid multibyte character */
  72                   *puc = 0xfffd;
  73                   return 2;
  74                 }
  75               /* invalid multibyte character */
  76             }
  77           else
  78             {
  79               /* incomplete multibyte character */
  80               *puc = 0xfffd;
  81               if (n == 1 || (s[1] ^ 0x80) >= 0x40)
  82                 return 1;
  83               else
  84                 return 2;
  85             }
  86         }
  87       else if (c < 0xf8)
  88         {
  89           if (n >= 4)
  90             {
  91               if ((s[1] ^ 0x80) < 0x40)
  92                 {
  93                   if ((s[2] ^ 0x80) < 0x40)
  94                     {
  95                       if ((s[3] ^ 0x80) < 0x40)
  96                         {
  97                           if ((c >= 0xf1 || s[1] >= 0x90)
  98 #if 1
  99                               && (c < 0xf4 || (c == 0xf4 && s[1] < 0x90))
 100 #endif
 101                              )
 102                             {
 103                               *puc = ((unsigned int) (c & 0x07) << 18)
 104                                      | ((unsigned int) (s[1] ^ 0x80) << 12)
 105                                      | ((unsigned int) (s[2] ^ 0x80) << 6)
 106                                      | (unsigned int) (s[3] ^ 0x80);
 107                               return 4;
 108                             }
 109                           /* invalid multibyte character */
 110                           *puc = 0xfffd;
 111                           return 4;
 112                         }
 113                       /* invalid multibyte character */
 114                       *puc = 0xfffd;
 115                       return 3;
 116                     }
 117                   /* invalid multibyte character */
 118                   *puc = 0xfffd;
 119                   return 2;
 120                 }
 121               /* invalid multibyte character */
 122             }
 123           else
 124             {
 125               /* incomplete multibyte character */
 126               *puc = 0xfffd;
 127               if (n == 1 || (s[1] ^ 0x80) >= 0x40)
 128                 return 1;
 129               else if (n == 2 || (s[2] ^ 0x80) >= 0x40)
 130                 return 2;
 131               else
 132                 return 3;
 133             }
 134         }
 135 #if 0
 136       else if (c < 0xfc)
 137         {
 138           if (n >= 5)
 139             {
 140               if ((s[1] ^ 0x80) < 0x40)
 141                 {
 142                   if ((s[2] ^ 0x80) < 0x40)
 143                     {
 144                       if ((s[3] ^ 0x80) < 0x40)
 145                         {
 146                           if ((s[4] ^ 0x80) < 0x40)
 147                             {
 148                               if (c >= 0xf9 || s[1] >= 0x88)
 149                                 {
 150                                   *puc = ((unsigned int) (c & 0x03) << 24)
 151                                          | ((unsigned int) (s[1] ^ 0x80) << 18)
 152                                          | ((unsigned int) (s[2] ^ 0x80) << 12)
 153                                          | ((unsigned int) (s[3] ^ 0x80) << 6)
 154                                          | (unsigned int) (s[4] ^ 0x80);
 155                                   return 5;
 156                                 }
 157                               /* invalid multibyte character */
 158                               *puc = 0xfffd;
 159                               return 5;
 160                             }
 161                           /* invalid multibyte character */
 162                           *puc = 0xfffd;
 163                           return 4;
 164                         }
 165                       /* invalid multibyte character */
 166                       *puc = 0xfffd;
 167                       return 3;
 168                     }
 169                   /* invalid multibyte character */
 170                   return 2;
 171                 }
 172               /* invalid multibyte character */
 173             }
 174           else
 175             {
 176               /* incomplete multibyte character */
 177               *puc = 0xfffd;
 178               return n;
 179             }
 180         }
 181       else if (c < 0xfe)
 182         {
 183           if (n >= 6)
 184             {
 185               if ((s[1] ^ 0x80) < 0x40)
 186                 {
 187                   if ((s[2] ^ 0x80) < 0x40)
 188                     {
 189                       if ((s[3] ^ 0x80) < 0x40)
 190                         {
 191                           if ((s[4] ^ 0x80) < 0x40)
 192                             {
 193                               if ((s[5] ^ 0x80) < 0x40)
 194                                 {
 195                                   if (c >= 0xfd || s[1] >= 0x84)
 196                                     {
 197                                       *puc = ((unsigned int) (c & 0x01) << 30)
 198                                              | ((unsigned int) (s[1] ^ 0x80) << 24)
 199                                              | ((unsigned int) (s[2] ^ 0x80) << 18)
 200                                              | ((unsigned int) (s[3] ^ 0x80) << 12)
 201                                              | ((unsigned int) (s[4] ^ 0x80) << 6)
 202                                              | (unsigned int) (s[5] ^ 0x80);
 203                                       return 6;
 204                                     }
 205                                   /* invalid multibyte character */
 206                                   *puc = 0xfffd;
 207                                   return 6;
 208                                 }
 209                               /* invalid multibyte character */
 210                               *puc = 0xfffd;
 211                               return 5;
 212                             }
 213                           /* invalid multibyte character */
 214                           *puc = 0xfffd;
 215                           return 4;
 216                         }
 217                       /* invalid multibyte character */
 218                       *puc = 0xfffd;
 219                       return 3;
 220                     }
 221                   /* invalid multibyte character */
 222                   return 2;
 223                 }
 224               /* invalid multibyte character */
 225             }
 226           else
 227             {
 228               /* incomplete multibyte character */
 229               *puc = 0xfffd;
 230               return n;
 231             }
 232         }
 233 #endif
 234     }
 235   /* invalid multibyte character */
 236   *puc = 0xfffd;
 237   return 1;
 238 }
 239
 240 #endif