locale/programs/3level.h

   1 /* Copyright (C) 2000-2013 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3    Contributed by Bruno Haible <haible@clisp.cons.org>, 2000.
   4
   5    This program is free software; you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published
   7    by the Free Software Foundation; version 2 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
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program; if not, see <http://www.gnu.org/licenses/>.  */
  17
  18 /* Construction of sparse 3-level tables.
  19    See wchar-lookup.h or coll-lookup.h for their structure and the
  20    meaning of p and q.
  21
  22    Before including this file, set
  23      TABLE        to the name of the structure to be defined
  24      ELEMENT      to the type of every entry
  25      DEFAULT      to the default value for empty entries
  26      ITERATE      if you want the TABLE_iterate function to be defined
  27      NO_FINALIZE  if you don't want the TABLE_finalize function to be defined
  28
  29    This will define
  30
  31      struct TABLE;
  32      void TABLE_init (struct TABLE *t);
  33      ELEMENT TABLE_get (struct TABLE *t, uint32_t wc);
  34      void TABLE_add (struct TABLE *t, uint32_t wc, ELEMENT value);
  35      void TABLE_iterate (struct TABLE *t,
  36                          void (*fn) (uint32_t wc, ELEMENT value));
  37      void TABLE_finalize (struct TABLE *t);
  38 */
  39
  40 #define CONCAT(a,b) CONCAT1(a,b)
  41 #define CONCAT1(a,b) a##b
  42
  43 struct TABLE
  44 {
  45   /* Parameters.  */
  46   unsigned int p;
  47   unsigned int q;
  48   /* Working representation.  */
  49   size_t level1_alloc;
  50   size_t level1_size;
  51   uint32_t *level1;
  52   size_t level2_alloc;
  53   size_t level2_size;
  54   uint32_t *level2;
  55   size_t level3_alloc;
  56   size_t level3_size;
  57   ELEMENT *level3;
  58   /* Compressed representation.  */
  59   size_t result_size;
  60   char *result;
  61 };
  62
  63 /* Initialize.  Assumes t->p and t->q have already been set.  */
  64 static inline void
  65 CONCAT(TABLE,_init) (struct TABLE *t)
  66 {
  67   t->level1 = NULL;
  68   t->level1_alloc = t->level1_size = 0;
  69   t->level2 = NULL;
  70   t->level2_alloc = t->level2_size = 0;
  71   t->level3 = NULL;
  72   t->level3_alloc = t->level3_size = 0;
  73 }
  74
  75 /* Marker for an empty slot.  This has the value 0xFFFFFFFF, regardless
  76    whether 'int' is 16 bit, 32 bit, or 64 bit.  */
  77 #define EMPTY ((uint32_t) ~0)
  78
  79 /* Retrieve an entry.  */
  80 static inline ELEMENT
  81 __attribute ((always_inline))
  82 CONCAT(TABLE,_get) (struct TABLE *t, uint32_t wc)
  83 {
  84   uint32_t index1 = wc >> (t->q + t->p);
  85   if (index1 < t->level1_size)
  86     {
  87       uint32_t lookup1 = t->level1[index1];
  88       if (lookup1 != EMPTY)
  89         {
  90           uint32_t index2 = ((wc >> t->p) & ((1 << t->q) - 1))
  91                             + (lookup1 << t->q);
  92           uint32_t lookup2 = t->level2[index2];
  93           if (lookup2 != EMPTY)
  94             {
  95               uint32_t index3 = (wc & ((1 << t->p) - 1))
  96                                 + (lookup2 << t->p);
  97               ELEMENT lookup3 = t->level3[index3];
  98
  99               return lookup3;
 100             }
 101         }
 102     }
 103   return DEFAULT;
 104 }
 105
 106 /* Add one entry.  */
 107 static void
 108 CONCAT(TABLE,_add) (struct TABLE *t, uint32_t wc, ELEMENT value)
 109 {
 110   uint32_t index1 = wc >> (t->q + t->p);
 111   uint32_t index2 = (wc >> t->p) & ((1 << t->q) - 1);
 112   uint32_t index3 = wc & ((1 << t->p) - 1);
 113   size_t i, i1, i2;
 114
 115   if (value == CONCAT(TABLE,_get) (t, wc))
 116     return;
 117
 118   if (index1 >= t->level1_size)
 119     {
 120       if (index1 >= t->level1_alloc)
 121         {
 122           size_t alloc = 2 * t->level1_alloc;
 123           if (alloc <= index1)
 124             alloc = index1 + 1;
 125           t->level1 = (uint32_t *) xrealloc ((char *) t->level1,
 126                                              alloc * sizeof (uint32_t));
 127           t->level1_alloc = alloc;
 128         }
 129       while (index1 >= t->level1_size)
 130         t->level1[t->level1_size++] = EMPTY;
 131     }
 132
 133   if (t->level1[index1] == EMPTY)
 134     {
 135       if (t->level2_size == t->level2_alloc)
 136         {
 137           size_t alloc = 2 * t->level2_alloc + 1;
 138           t->level2 = (uint32_t *) xrealloc ((char *) t->level2,
 139                                              (alloc << t->q) * sizeof (uint32_t));
 140           t->level2_alloc = alloc;
 141         }
 142       i1 = t->level2_size << t->q;
 143       i2 = (t->level2_size + 1) << t->q;
 144       for (i = i1; i < i2; i++)
 145         t->level2[i] = EMPTY;
 146       t->level1[index1] = t->level2_size++;
 147     }
 148
 149   index2 += t->level1[index1] << t->q;
 150
 151   if (t->level2[index2] == EMPTY)
 152     {
 153       if (t->level3_size == t->level3_alloc)
 154         {
 155           size_t alloc = 2 * t->level3_alloc + 1;
 156           t->level3 = (ELEMENT *) xrealloc ((char *) t->level3,
 157                                             (alloc << t->p) * sizeof (ELEMENT));
 158           t->level3_alloc = alloc;
 159         }
 160       i1 = t->level3_size << t->p;
 161       i2 = (t->level3_size + 1) << t->p;
 162       for (i = i1; i < i2; i++)
 163         t->level3[i] = DEFAULT;
 164       t->level2[index2] = t->level3_size++;
 165     }
 166
 167   index3 += t->level2[index2] << t->p;
 168
 169   t->level3[index3] = value;
 170 }
 171
 172 #ifdef ITERATE
 173 /* Apply a function to all entries in the table.  */
 174 static void
 175 CONCAT(TABLE,_iterate) (struct TABLE *t,
 176                         void (*fn) (uint32_t wc, ELEMENT value))
 177 {
 178   uint32_t index1;
 179   for (index1 = 0; index1 < t->level1_size; index1++)
 180     {
 181       uint32_t lookup1 = t->level1[index1];
 182       if (lookup1 != EMPTY)
 183         {
 184           uint32_t lookup1_shifted = lookup1 << t->q;
 185           uint32_t index2;
 186           for (index2 = 0; index2 < (1 << t->q); index2++)
 187             {
 188               uint32_t lookup2 = t->level2[index2 + lookup1_shifted];
 189               if (lookup2 != EMPTY)
 190                 {
 191                   uint32_t lookup2_shifted = lookup2 << t->p;
 192                   uint32_t index3;
 193                   for (index3 = 0; index3 < (1 << t->p); index3++)
 194                     {
 195                       ELEMENT lookup3 = t->level3[index3 + lookup2_shifted];
 196                       if (lookup3 != DEFAULT)
 197                         fn ((((index1 << t->q) + index2) << t->p) + index3,
 198                             lookup3);
 199                     }
 200                 }
 201             }
 202         }
 203     }
 204 }
 205 #endif
 206
 207 #ifndef NO_FINALIZE
 208 /* Finalize and shrink.  */
 209 static void
 210 CONCAT(TABLE,_finalize) (struct TABLE *t)
 211 {
 212   size_t i, j, k;
 213   uint32_t reorder3[t->level3_size];
 214   uint32_t reorder2[t->level2_size];
 215   uint32_t level1_offset, level2_offset, level3_offset, last_offset;
 216
 217   /* Uniquify level3 blocks.  */
 218   k = 0;
 219   for (j = 0; j < t->level3_size; j++)
 220     {
 221       for (i = 0; i < k; i++)
 222         if (memcmp (&t->level3[i << t->p], &t->level3[j << t->p],
 223                     (1 << t->p) * sizeof (ELEMENT)) == 0)
 224           break;
 225       /* Relocate block j to block i.  */
 226       reorder3[j] = i;
 227       if (i == k)
 228         {
 229           if (i != j)
 230             memcpy (&t->level3[i << t->p], &t->level3[j << t->p],
 231                     (1 << t->p) * sizeof (ELEMENT));
 232           k++;
 233         }
 234     }
 235   t->level3_size = k;
 236
 237   for (i = 0; i < (t->level2_size << t->q); i++)
 238     if (t->level2[i] != EMPTY)
 239       t->level2[i] = reorder3[t->level2[i]];
 240
 241   /* Uniquify level2 blocks.  */
 242   k = 0;
 243   for (j = 0; j < t->level2_size; j++)
 244     {
 245       for (i = 0; i < k; i++)
 246         if (memcmp (&t->level2[i << t->q], &t->level2[j << t->q],
 247                     (1 << t->q) * sizeof (uint32_t)) == 0)
 248           break;
 249       /* Relocate block j to block i.  */
 250       reorder2[j] = i;
 251       if (i == k)
 252         {
 253           if (i != j)
 254             memcpy (&t->level2[i << t->q], &t->level2[j << t->q],
 255                     (1 << t->q) * sizeof (uint32_t));
 256           k++;
 257         }
 258     }
 259   t->level2_size = k;
 260
 261   for (i = 0; i < t->level1_size; i++)
 262     if (t->level1[i] != EMPTY)
 263       t->level1[i] = reorder2[t->level1[i]];
 264
 265   /* Create and fill the resulting compressed representation.  */
 266   last_offset =
 267     5 * sizeof (uint32_t)
 268     + t->level1_size * sizeof (uint32_t)
 269     + (t->level2_size << t->q) * sizeof (uint32_t)
 270     + (t->level3_size << t->p) * sizeof (ELEMENT);
 271   t->result_size = (last_offset + 3) & ~3ul;
 272   t->result = (char *) xmalloc (t->result_size);
 273
 274   level1_offset =
 275     5 * sizeof (uint32_t);
 276   level2_offset =
 277     5 * sizeof (uint32_t)
 278     + t->level1_size * sizeof (uint32_t);
 279   level3_offset =
 280     5 * sizeof (uint32_t)
 281     + t->level1_size * sizeof (uint32_t)
 282     + (t->level2_size << t->q) * sizeof (uint32_t);
 283
 284   ((uint32_t *) t->result)[0] = t->q + t->p;
 285   ((uint32_t *) t->result)[1] = t->level1_size;
 286   ((uint32_t *) t->result)[2] = t->p;
 287   ((uint32_t *) t->result)[3] = (1 << t->q) - 1;
 288   ((uint32_t *) t->result)[4] = (1 << t->p) - 1;
 289
 290   for (i = 0; i < t->level1_size; i++)
 291     ((uint32_t *) (t->result + level1_offset))[i] =
 292       (t->level1[i] == EMPTY
 293        ? 0
 294        : (t->level1[i] << t->q) * sizeof (uint32_t) + level2_offset);
 295
 296   for (i = 0; i < (t->level2_size << t->q); i++)
 297     ((uint32_t *) (t->result + level2_offset))[i] =
 298       (t->level2[i] == EMPTY
 299        ? 0
 300        : (t->level2[i] << t->p) * sizeof (ELEMENT) + level3_offset);
 301
 302   for (i = 0; i < (t->level3_size << t->p); i++)
 303     ((ELEMENT *) (t->result + level3_offset))[i] = t->level3[i];
 304
 305   if (last_offset < t->result_size)
 306     memset (t->result + last_offset, 0, t->result_size - last_offset);
 307
 308   if (t->level1_alloc > 0)
 309     free (t->level1);
 310   if (t->level2_alloc > 0)
 311     free (t->level2);
 312   if (t->level3_alloc > 0)
 313     free (t->level3);
 314 }
 315 #endif
 316
 317 #undef EMPTY
 318 #undef TABLE
 319 #undef ELEMENT
 320 #undef DEFAULT
 321 #undef ITERATE
 322 #undef NO_FINALIZE