stdlib/grouping.c

   1 /* Internal header for proving correct grouping in strings of numbers.
   2    Copyright (C) 1995-2023 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library 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
  16    License along with the GNU C Library; if not, see
  17    <https://www.gnu.org/licenses/>.  */
  18
  19 #include <limits.h>
  20 #include <stddef.h>
  21 #include <string.h>
  22
  23 #ifndef MAX
  24 #define MAX(a,b)        ({ typeof(a) _a = (a); typeof(b) _b = (b); \
  25                            _a > _b ? _a : _b; })
  26 #endif
  27
  28 #ifdef USE_WIDE_CHAR
  29 # include <wctype.h>
  30 # define L_(Ch) L##Ch
  31 # define UCHAR_TYPE wint_t
  32 # define STRING_TYPE wchar_t
  33 #else
  34 # define L_(Ch) Ch
  35 # define UCHAR_TYPE unsigned char
  36 # define STRING_TYPE char
  37 #endif
  38
  39 #include "grouping.h"
  40
  41 /* Find the maximum prefix of the string between BEGIN and END which
  42    satisfies the grouping rules.  It is assumed that at least one digit
  43    follows BEGIN directly.  */
  44
  45 const STRING_TYPE *
  46 #ifdef USE_WIDE_CHAR
  47 __correctly_grouped_prefixwc (const STRING_TYPE *begin, const STRING_TYPE *end,
  48                               wchar_t thousands,
  49 #else
  50 __correctly_grouped_prefixmb (const STRING_TYPE *begin, const STRING_TYPE *end,
  51                               const char *thousands,
  52 #endif
  53                               const char *grouping)
  54 {
  55   if (grouping == NULL)
  56     return end;
  57
  58 #ifdef USE_WIDE_CHAR
  59   size_t thousands_len = 1;
  60 #else
  61   size_t thousands_len = strlen (thousands);
  62   int cnt;
  63 #endif
  64
  65   while (end - begin >= thousands_len)
  66     {
  67       const STRING_TYPE *cp = end - thousands_len;
  68       const char *gp = grouping;
  69
  70       /* Check first group.  */
  71       while (cp >= begin)
  72         {
  73 #ifdef USE_WIDE_CHAR
  74           if (*cp == thousands)
  75             break;
  76 #else
  77           if (cp[thousands_len - 1] == *thousands)
  78             {
  79               for (cnt = 1; thousands[cnt] != '\0'; ++cnt)
  80                 if (thousands[cnt] != cp[thousands_len - 1 - cnt])
  81                   break;
  82               if (thousands[cnt] == '\0')
  83                 break;
  84             }
  85 #endif
  86           --cp;
  87         }
  88
  89       /* We allow the representation to contain no grouping at all even if
  90          the locale specifies we can have grouping.  */
  91       if (cp < begin)
  92         return end;
  93
  94       if (end - cp == (int) *gp + 1)
  95         {
  96           /* This group matches the specification.  */
  97
  98           const STRING_TYPE *new_end;
  99
 100           if (cp < begin)
 101             /* There is just one complete group.  We are done.  */
 102             return end;
 103
 104           /* CP points to a thousands separator character.  The preceding
 105              remainder of the string from BEGIN to NEW_END is the part we
 106              will consider if there is a grouping error in this trailing
 107              portion from CP to END.  */
 108           new_end = cp - 1;
 109
 110           /* Loop while the grouping is correct.  */
 111           while (1)
 112             {
 113               /* Get the next grouping rule.  */
 114               ++gp;
 115               if (*gp == 0)
 116                 /* If end is reached use last rule.  */
 117                 --gp;
 118
 119               /* Skip the thousands separator.  */
 120               --cp;
 121
 122               if (*gp == CHAR_MAX
 123 #if CHAR_MIN < 0
 124                   || *gp < 0
 125 #endif
 126                   )
 127                 {
 128                   /* No more thousands separators are allowed to follow.  */
 129                   while (cp >= begin)
 130                     {
 131 #ifdef USE_WIDE_CHAR
 132                       if (*cp == thousands)
 133                         break;
 134 #else
 135                       for (cnt = 0; thousands[cnt] != '\0'; ++cnt)
 136                         if (thousands[cnt] != cp[thousands_len - cnt - 1])
 137                           break;
 138                       if (thousands[cnt] == '\0')
 139                         break;
 140 #endif
 141                       --cp;
 142                     }
 143
 144                   if (cp < begin)
 145                     /* OK, only digits followed.  */
 146                     return end;
 147                 }
 148               else
 149                 {
 150                   /* Check the next group.  */
 151                   const STRING_TYPE *group_end = cp;
 152
 153                   while (cp >= begin)
 154                     {
 155 #ifdef USE_WIDE_CHAR
 156                       if (*cp == thousands)
 157                         break;
 158 #else
 159                       for (cnt = 0; thousands[cnt] != '\0'; ++cnt)
 160                         if (thousands[cnt] != cp[thousands_len - cnt - 1])
 161                           break;
 162                       if (thousands[cnt] == '\0')
 163                         break;
 164 #endif
 165                       --cp;
 166                     }
 167
 168                   if (cp < begin && group_end - cp <= (int) *gp)
 169                     /* Final group is correct.  */
 170                     return end;
 171
 172                   if (cp < begin || group_end - cp != (int) *gp)
 173                     /* Incorrect group.  Punt.  */
 174                     break;
 175                 }
 176             }
 177
 178           /* The trailing portion of the string starting at NEW_END
 179              contains a grouping error.  So we will look for a correctly
 180              grouped number in the preceding portion instead.  */
 181           end = new_end;
 182         }
 183       else
 184         {
 185           /* Even the first group was wrong; determine maximum shift.  */
 186           if (end - cp > (int) *gp + 1)
 187             end = cp + (int) *gp + 1;
 188           else if (cp < begin)
 189             /* This number does not fill the first group, but is correct.  */
 190             return end;
 191           else
 192             /* CP points to a thousands separator character.  */
 193             end = cp;
 194         }
 195     }
 196
 197   return MAX (begin, end);
 198 }