string/strcoll.c

   1 /* Copyright (C) 1995, 1996, 1997 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3    Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Library General Public License as
   7    published by the Free Software Foundation; either version 2 of the
   8    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    Library General Public License for more details.
  14
  15    You should have received a copy of the GNU Library General Public
  16    License along with the GNU C Library; see the file COPYING.LIB.  If not,
  17    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  18    Boston, MA 02111-1307, USA.  */
  19
  20 #include <endian.h>
  21 #include <stddef.h>
  22 #include <stdlib.h>
  23 #include <string.h>
  24
  25 #ifndef STRING_TYPE
  26 # define STRING_TYPE char
  27 # define USTRING_TYPE unsigned char
  28 # ifdef USE_IN_EXTENDED_LOCALE_MODEL
  29 #  define STRCOLL __strcoll_l
  30 # else
  31 #  define STRCOLL strcoll
  32 # endif
  33 # define STRCMP strcmp
  34 #endif
  35
  36 /* Include the shared helper functions.  `strxfrm'/`wcsxfrm' also use
  37    these functions.  */
  38 #include "../locale/weight.h"
  39
  40
  41 /* Compare S1 and S2, returning less than, equal to or
  42    greater than zero if the collated form of S1 is lexicographically
  43    less than, equal to or greater than the collated form of S2.  */
  44 #ifndef USE_IN_EXTENDED_LOCALE_MODEL
  45 int
  46 STRCOLL (s1, s2)
  47      const STRING_TYPE *s1;
  48      const STRING_TYPE *s2;
  49 #else
  50 int
  51 STRCOLL (s1, s2, l)
  52      const STRING_TYPE *s1;
  53      const STRING_TYPE *s2;
  54      __locale_t l;
  55 #endif
  56 {
  57 #ifdef USE_IN_EXTENDED_LOCALE_MODEL
  58   struct locale_data *current = l->__locales[LC_COLLATE];
  59 # if BYTE_ORDER == BIG_ENDIAN
  60   const u_int32_t *collate_table = (const u_int32_t *)
  61     current->values[_NL_ITEM_INDEX (_NL_COLLATE_TABLE_EB)].string;
  62   const u_int32_t *collate_extra = (const u_int32_t *)
  63     current->values[_NL_ITEM_INDEX (_NL_COLLATE_EXTRA_EB)].string;
  64 # elif BYTE_ORDER == LITTLE_ENDIAN
  65   const u_int32_t *collate_table = (const u_int32_t *)
  66     current->values[_NL_ITEM_INDEX (_NL_COLLATE_TABLE_EL)].string;
  67   const u_int32_t *collate_extra = (const u_int32_t *)
  68     current->values[_NL_ITEM_INDEX (_NL_COLLATE_EXTRA_EL)].string;
  69 # else
  70 #  error bizarre byte order
  71 # endif
  72 #endif
  73   weight_t *s1forw = NULL;
  74   weight_t *s1backw = NULL;
  75   weight_t *s2forw = NULL;
  76   weight_t *s2backw = NULL;
  77   size_t pass;
  78
  79   /* If the current locale does not specify locale data we use normal
  80      8-bit string comparison.  */
  81   if (collate_nrules == 0)
  82     return STRCMP (s1, s2);
  83
  84   /* Get full information about the strings.  This means we get
  85      information for all passes in a special data structure.  */
  86   get_string (s1, s1forw, s1backw);
  87   get_string (s2, s2forw, s2backw);
  88
  89   /* Now we have all the information.  In at most the given number of
  90      passes we can finally decide about the order.  */
  91   for (pass = 0; pass < collate_nrules; ++pass)
  92     {
  93       int forward = (collate_rules[pass] & sort_forward) != 0;
  94       const weight_t *s1run = forward ? s1forw : s1backw;
  95       const weight_t *s2run = forward ? s2forw : s2backw;
  96       int s1idx = forward ? 0 : s1run->data[pass].number - 1;
  97       int s2idx = forward ? 0 : s2run->data[pass].number - 1;
  98
  99       do
 100         {
 101           int s1ignore = 0;
 102           int s2ignore = 0;
 103           u_int32_t w1, w2;
 104
 105           /* Here we have to check for IGNORE entries.  If these are
 106              found we count them and go on with the next value.  */
 107           while ((w1 = s1run->data[pass].value[s1idx])
 108                  == (u_int32_t) IGNORE_CHAR)
 109             {
 110               ++s1ignore;
 111               if ((forward && ++s1idx >= s1run->data[pass].number)
 112                   || (!forward && --s1idx < 0))
 113                 {
 114                   weight_t *nextp = forward ? s1run->next : s1run->prev;
 115                   if (nextp == NULL)
 116                     {
 117                       w1 = 0;
 118                       break;
 119                     }
 120                   s1run = nextp;
 121                   s1idx = forward ? 0 : s1run->data[pass].number - 1;
 122                 }
 123             }
 124
 125           while ((w2 = s2run->data[pass].value[s2idx])
 126                  == (u_int32_t) IGNORE_CHAR)
 127             {
 128               ++s2ignore;
 129               if ((forward && ++s2idx >= s2run->data[pass].number)
 130                   || (!forward && --s2idx < 0))
 131                 {
 132                   weight_t *nextp = forward ? s2run->next : s2run->prev;
 133                   if (nextp == NULL)
 134                     {
 135                       w2 = 0;
 136                       break;
 137                     }
 138                   s2run = nextp;
 139                   s2idx = forward ? 0 : s2run->data[pass].number - 1;
 140                 }
 141             }
 142
 143           /* Now we have information of the number of ignored
 144              weights and the value of the next weight.  */
 145           if ((collate_rules[pass] & sort_position) != 0
 146               && s1ignore != s2ignore && (w1 != 0 || w2 != 0))
 147             return s1ignore < s2ignore ? -1 : 1;
 148
 149           if (w1 != w2)
 150             return w1 < w2 ? -1 : 1;
 151
 152           /* We have to increment the index counters.  */
 153           if ((forward && ++s1idx >= s1run->data[pass].number)
 154               || (!forward && --s1idx < 0))
 155             if (forward)
 156               {
 157                 s1run = s1run->next;
 158                 s1idx = 0;
 159               }
 160             else
 161               {
 162                 s1run = s1run->prev;
 163                 if (s1run != NULL)
 164                   s1idx = s1run->data[pass].number - 1;
 165               }
 166
 167           if ((forward && ++s2idx >= s2run->data[pass].number)
 168               || (!forward && --s2idx < 0))
 169             if (forward)
 170               {
 171                 s2run = s2run->next;
 172                 s2idx = 0;
 173               }
 174             else
 175               {
 176                 s2run = s2run->prev;
 177                 if (s2run != NULL)
 178                   s2idx = s2run->data[pass].number - 1;
 179               }
 180
 181         }
 182       while (s1run != NULL && s2run != NULL);
 183
 184       if (s1run != s2run)
 185         return s1run != NULL ? 1 : -1;
 186     }
 187
 188   return 0;
 189 }