string/strcoll_l.c

   1 /* Copyright (C) 1995-2014 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3    Written by Ulrich Drepper <drepper@gnu.org>, 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 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    <http://www.gnu.org/licenses/>.  */
  18
  19
  20 #include <assert.h>
  21 #include <langinfo.h>
  22 #include <locale.h>
  23 #include <stddef.h>
  24 #include <stdint.h>
  25 #include <string.h>
  26 #include <sys/param.h>
  27
  28 #ifndef STRING_TYPE
  29 # define STRING_TYPE char
  30 # define USTRING_TYPE unsigned char
  31 # define STRCOLL __strcoll_l
  32 # define STRCMP strcmp
  33 # define STRLEN strlen
  34 # define WEIGHT_H "../locale/weight.h"
  35 # define SUFFIX MB
  36 # define L(arg) arg
  37 #endif
  38
  39 #define CONCAT(a,b) CONCAT1(a,b)
  40 #define CONCAT1(a,b) a##b
  41
  42 #include "../locale/localeinfo.h"
  43 #include WEIGHT_H
  44
  45 /* Track status while looking for sequences in a string.  */
  46 typedef struct
  47 {
  48   int len;                      /* Length of the current sequence.  */
  49   size_t val;                   /* Position of the sequence relative to the
  50                                    previous non-ignored sequence.  */
  51   size_t idxnow;                /* Current index in sequences.  */
  52   size_t idxmax;                /* Maximum index in sequences.  */
  53   size_t idxcnt;                /* Current count of indices.  */
  54   size_t backw;                 /* Current Backward sequence index.  */
  55   size_t backw_stop;            /* Index where the backward sequences stop.  */
  56   const USTRING_TYPE *us;       /* The string.  */
  57   unsigned char rule;           /* Saved rule for the first sequence.  */
  58   int32_t idx;                  /* Index to weight of the current sequence.  */
  59   int32_t save_idx;             /* Save looked up index of a forward
  60                                    sequence after the last backward
  61                                    sequence.  */
  62   const USTRING_TYPE *back_us;  /* Beginning of the backward sequence.  */
  63 } coll_seq;
  64
  65 /* Get next sequence.  Traverse the string as required.  */
  66 static __always_inline void
  67 get_next_seq (coll_seq *seq, int nrules, const unsigned char *rulesets,
  68               const USTRING_TYPE *weights, const int32_t *table,
  69               const USTRING_TYPE *extra, const int32_t *indirect,
  70               int pass)
  71 {
  72   size_t val = seq->val = 0;
  73   int len = seq->len;
  74   size_t backw_stop = seq->backw_stop;
  75   size_t backw = seq->backw;
  76   size_t idxcnt = seq->idxcnt;
  77   size_t idxmax = seq->idxmax;
  78   int32_t idx = seq->idx;
  79   const USTRING_TYPE *us = seq->us;
  80
  81   while (len == 0)
  82     {
  83       ++val;
  84       if (backw_stop != ~0ul)
  85         {
  86           /* There is something pushed.  */
  87           if (backw == backw_stop)
  88             {
  89               /* The last pushed character was handled.  Continue
  90                  with forward characters.  */
  91               if (idxcnt < idxmax)
  92                 {
  93                   idx = seq->save_idx;
  94                   backw_stop = ~0ul;
  95                 }
  96               else
  97                 {
  98                   /* Nothing anymore.  The backward sequence ended with
  99                      the last sequence in the string.  Note that len is
 100                      still zero.  */
 101                   idx = 0;
 102                   break;
 103                 }
 104             }
 105           else
 106             {
 107               /* XXX Traverse BACKW sequences from the beginning of
 108                  BACKW_STOP to get the next sequence.  Is ther a quicker way
 109                  to do this?  */
 110               size_t i = backw_stop;
 111               us = seq->back_us;
 112               while (i < backw)
 113                 {
 114                   int32_t tmp = findidx (table, indirect, extra, &us, -1);
 115                   idx = tmp & 0xffffff;
 116                   i++;
 117                 }
 118               --backw;
 119               us = seq->us;
 120             }
 121         }
 122       else
 123         {
 124           backw_stop = idxmax;
 125           int32_t prev_idx = idx;
 126
 127           while (*us != L('\0'))
 128             {
 129               int32_t tmp = findidx (table, indirect, extra, &us, -1);
 130               unsigned char rule = tmp >> 24;
 131               prev_idx = idx;
 132               idx = tmp & 0xffffff;
 133               idxcnt = idxmax++;
 134
 135               /* Save the rule for the first sequence.  */
 136               if (__glibc_unlikely (idxcnt == 0))
 137                 seq->rule = rule;
 138
 139               if ((rulesets[rule * nrules + pass]
 140                    & sort_backward) == 0)
 141                 /* No more backward characters to push.  */
 142                 break;
 143               ++idxcnt;
 144             }
 145
 146           if (backw_stop >= idxcnt)
 147             {
 148               /* No sequence at all or just one.  */
 149               if (idxcnt == idxmax || backw_stop > idxcnt)
 150                 /* Note that len is still zero.  */
 151                 break;
 152
 153               backw_stop = ~0ul;
 154             }
 155           else
 156             {
 157               /* We pushed backward sequences.  If the stream ended with the
 158                  backward sequence, then we process the last sequence we
 159                  found.  Otherwise we process the sequence before the last
 160                  one since the last one was a forward sequence.  */
 161               seq->back_us = seq->us;
 162               seq->us = us;
 163               backw = idxcnt;
 164               if (idxmax > idxcnt)
 165                 {
 166                   backw--;
 167                   seq->save_idx = idx;
 168                   idx = prev_idx;
 169                 }
 170               if (backw > backw_stop)
 171                 backw--;
 172             }
 173         }
 174
 175       len = weights[idx++];
 176       /* Skip over indices of previous levels.  */
 177       for (int i = 0; i < pass; i++)
 178         {
 179           idx += len;
 180           len = weights[idx];
 181           idx++;
 182         }
 183     }
 184
 185   /* Update the structure.  */
 186   seq->val = val;
 187   seq->len = len;
 188   seq->backw_stop = backw_stop;
 189   seq->backw = backw;
 190   seq->idxcnt = idxcnt;
 191   seq->idxmax = idxmax;
 192   seq->us = us;
 193   seq->idx = idx;
 194 }
 195
 196 /* Compare two sequences.  */
 197 static __always_inline int
 198 do_compare (coll_seq *seq1, coll_seq *seq2, int position,
 199             const USTRING_TYPE *weights)
 200 {
 201   int seq1len = seq1->len;
 202   int seq2len = seq2->len;
 203   size_t val1 = seq1->val;
 204   size_t val2 = seq2->val;
 205   int idx1 = seq1->idx;
 206   int idx2 = seq2->idx;
 207   int result = 0;
 208
 209   /* Test for position if necessary.  */
 210   if (position && val1 != val2)
 211     {
 212       result = val1 > val2 ? 1 : -1;
 213       goto out;
 214     }
 215
 216   /* Compare the two sequences.  */
 217   do
 218     {
 219       if (weights[idx1] != weights[idx2])
 220         {
 221           /* The sequences differ.  */
 222           result = weights[idx1] - weights[idx2];
 223           goto out;
 224         }
 225
 226       /* Increment the offsets.  */
 227       ++idx1;
 228       ++idx2;
 229
 230       --seq1len;
 231       --seq2len;
 232     }
 233   while (seq1len > 0 && seq2len > 0);
 234
 235   if (position && seq1len != seq2len)
 236     result = seq1len - seq2len;
 237
 238 out:
 239   seq1->len = seq1len;
 240   seq2->len = seq2len;
 241   seq1->idx = idx1;
 242   seq2->idx = idx2;
 243   return result;
 244 }
 245
 246 int
 247 STRCOLL (const STRING_TYPE *s1, const STRING_TYPE *s2, __locale_t l)
 248 {
 249   struct __locale_data *current = l->__locales[LC_COLLATE];
 250   uint_fast32_t nrules = current->values[_NL_ITEM_INDEX (_NL_COLLATE_NRULES)].word;
 251   /* We don't assign the following values right away since it might be
 252      unnecessary in case there are no rules.  */
 253   const unsigned char *rulesets;
 254   const int32_t *table;
 255   const USTRING_TYPE *weights;
 256   const USTRING_TYPE *extra;
 257   const int32_t *indirect;
 258
 259   if (nrules == 0)
 260     return STRCMP (s1, s2);
 261
 262   /* Catch empty strings.  */
 263   if (__glibc_unlikely (*s1 == '\0') || __glibc_unlikely (*s2 == '\0'))
 264     return (*s1 != '\0') - (*s2 != '\0');
 265
 266   rulesets = (const unsigned char *)
 267     current->values[_NL_ITEM_INDEX (_NL_COLLATE_RULESETS)].string;
 268   table = (const int32_t *)
 269     current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_TABLE,SUFFIX))].string;
 270   weights = (const USTRING_TYPE *)
 271     current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_WEIGHT,SUFFIX))].string;
 272   extra = (const USTRING_TYPE *)
 273     current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_EXTRA,SUFFIX))].string;
 274   indirect = (const int32_t *)
 275     current->values[_NL_ITEM_INDEX (CONCAT(_NL_COLLATE_INDIRECT,SUFFIX))].string;
 276
 277   assert (((uintptr_t) table) % __alignof__ (table[0]) == 0);
 278   assert (((uintptr_t) weights) % __alignof__ (weights[0]) == 0);
 279   assert (((uintptr_t) extra) % __alignof__ (extra[0]) == 0);
 280   assert (((uintptr_t) indirect) % __alignof__ (indirect[0]) == 0);
 281
 282   int result = 0, rule = 0;
 283
 284   coll_seq seq1, seq2;
 285   memset (&seq1, 0, sizeof (seq1));
 286   seq2 = seq1;
 287
 288   for (int pass = 0; pass < nrules; ++pass)
 289     {
 290       seq1.idxcnt = 0;
 291       seq1.idx = 0;
 292       seq2.idx = 0;
 293       seq1.backw_stop = ~0ul;
 294       seq1.backw = ~0ul;
 295       seq2.idxcnt = 0;
 296       seq2.backw_stop = ~0ul;
 297       seq2.backw = ~0ul;
 298
 299       /* We need the elements of the strings as unsigned values since they
 300          are used as indices.  */
 301       seq1.us = (const USTRING_TYPE *) s1;
 302       seq2.us = (const USTRING_TYPE *) s2;
 303
 304       /* We assume that if a rule has defined `position' in one section
 305          this is true for all of them.  Please note that the localedef programs
 306          makes sure that `position' is not used at the first level.  */
 307
 308       int position = rulesets[rule * nrules + pass] & sort_position;
 309
 310       while (1)
 311         {
 312           get_next_seq (&seq1, nrules, rulesets, weights, table,
 313                                     extra, indirect, pass);
 314           get_next_seq (&seq2, nrules, rulesets, weights, table,
 315                                     extra, indirect, pass);
 316           /* See whether any or both strings are empty.  */
 317           if (seq1.len == 0 || seq2.len == 0)
 318             {
 319               if (seq1.len == seq2.len)
 320                 {
 321                   /* Both strings ended and are equal at this level.  Do a
 322                      byte-level comparison to ensure that we don't waste time
 323                      going through multiple passes for totally equal strings
 324                      before proceeding to subsequent passes.  */
 325                   if (pass == 0 && STRCMP (s1, s2) == 0)
 326                     return result;
 327                   else
 328                     break;
 329                 }
 330
 331               /* This means one string is shorter than the other.  Find out
 332                  which one and return an appropriate value.  */
 333               return seq1.len == 0 ? -1 : 1;
 334             }
 335
 336           result = do_compare (&seq1, &seq2, position, weights);
 337           if (result != 0)
 338             return result;
 339         }
 340
 341       rule = seq1.rule;
 342     }
 343
 344   return result;
 345 }
 346 libc_hidden_def (STRCOLL)
 347
 348 #ifndef WIDE_CHAR_VERSION
 349 weak_alias (__strcoll_l, strcoll_l)
 350 #endif