gvdb/gvdb-builder.c

   1 /*
   2  * Copyright © 2010 Codethink Limited
   3  *
   4  * This library is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * This library is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with this library; if not, see <http://www.gnu.org/licenses/>.
  16  *
  17  * Author: Ryan Lortie <desrt@desrt.ca>
  18  */
  19
  20 #include "gvdb-builder.h"
  21 #include "gvdb-format.h"
  22
  23 #include <glib.h>
  24 #include <fcntl.h>
  25 #if !defined(G_OS_WIN32) || !defined(_MSC_VER)
  26 #include <unistd.h>
  27 #endif
  28 #include <string.h>
  29
  30
  31 struct _GvdbItem
  32 {
  33   gchar *key;
  34   guint32 hash_value;
  35   guint32_le assigned_index;
  36   GvdbItem *parent;
  37   GvdbItem *sibling;
  38   GvdbItem *next;
  39
  40   /* one of:
  41    * this:
  42    */
  43   GVariant *value;
  44
  45   /* this: */
  46   GHashTable *table;
  47
  48   /* or this: */
  49   GvdbItem *child;
  50 };
  51
  52 static void
  53 gvdb_item_free (gpointer data)
  54 {
  55   GvdbItem *item = data;
  56
  57   g_free (item->key);
  58
  59   if (item->value)
  60     g_variant_unref (item->value);
  61
  62   if (item->table)
  63     g_hash_table_unref (item->table);
  64
  65   g_slice_free (GvdbItem, item);
  66 }
  67
  68 GHashTable *
  69 gvdb_hash_table_new (GHashTable  *parent,
  70                      const gchar *name_in_parent)
  71 {
  72   GHashTable *table;
  73
  74   table = g_hash_table_new_full (g_str_hash, g_str_equal,
  75                                  g_free, gvdb_item_free);
  76
  77   if (parent)
  78     {
  79       GvdbItem *item;
  80
  81       item = gvdb_hash_table_insert (parent, name_in_parent);
  82       gvdb_item_set_hash_table (item, table);
  83     }
  84
  85   return table;
  86 }
  87
  88 static guint32
  89 djb_hash (const gchar *key)
  90 {
  91   guint32 hash_value = 5381;
  92
  93   while (*key)
  94     hash_value = hash_value * 33 + *(signed char *)key++;
  95
  96   return hash_value;
  97 }
  98
  99 GvdbItem *
 100 gvdb_hash_table_insert (GHashTable  *table,
 101                         const gchar *key)
 102 {
 103   GvdbItem *item;
 104
 105   item = g_slice_new0 (GvdbItem);
 106   item->key = g_strdup (key);
 107   item->hash_value = djb_hash (key);
 108
 109   g_hash_table_insert (table, g_strdup (key), item);
 110
 111   return item;
 112 }
 113
 114 void
 115 gvdb_hash_table_insert_string (GHashTable  *table,
 116                                const gchar *key,
 117                                const gchar *value)
 118 {
 119   GvdbItem *item;
 120
 121   item = gvdb_hash_table_insert (table, key);
 122   gvdb_item_set_value (item, g_variant_new_string (value));
 123 }
 124
 125 void
 126 gvdb_item_set_value (GvdbItem *item,
 127                      GVariant *value)
 128 {
 129   g_return_if_fail (!item->value && !item->table && !item->child);
 130
 131   item->value = g_variant_ref_sink (value);
 132 }
 133
 134 void
 135 gvdb_item_set_hash_table (GvdbItem   *item,
 136                           GHashTable *table)
 137 {
 138   g_return_if_fail (!item->value && !item->table && !item->child);
 139
 140   item->table = g_hash_table_ref (table);
 141 }
 142
 143 void
 144 gvdb_item_set_parent (GvdbItem *item,
 145                       GvdbItem *parent)
 146 {
 147   GvdbItem **node;
 148
 149   g_return_if_fail (g_str_has_prefix (item->key, parent->key));
 150   g_return_if_fail (!parent->value && !parent->table);
 151   g_return_if_fail (!item->parent && !item->sibling);
 152
 153   for (node = &parent->child; *node; node = &(*node)->sibling)
 154     if (strcmp ((*node)->key, item->key) > 0)
 155       break;
 156
 157   item->parent = parent;
 158   item->sibling = *node;
 159   *node = item;
 160 }
 161
 162 typedef struct
 163 {
 164   GvdbItem **buckets;
 165   gint n_buckets;
 166 } HashTable;
 167
 168 static HashTable *
 169 hash_table_new (gint n_buckets)
 170 {
 171   HashTable *table;
 172
 173   table = g_slice_new (HashTable);
 174   table->buckets = g_new0 (GvdbItem *, n_buckets);
 175   table->n_buckets = n_buckets;
 176
 177   return table;
 178 }
 179
 180 static void
 181 hash_table_free (HashTable *table)
 182 {
 183   g_free (table->buckets);
 184
 185   g_slice_free (HashTable, table);
 186 }
 187
 188 static void
 189 hash_table_insert (gpointer key,
 190                    gpointer value,
 191                    gpointer data)
 192 {
 193   guint32 hash_value, bucket;
 194   HashTable *table = data;
 195   GvdbItem *item = value;
 196
 197   hash_value = djb_hash (key);
 198   bucket = hash_value % table->n_buckets;
 199   item->next = table->buckets[bucket];
 200   table->buckets[bucket] = item;
 201 }
 202
 203 static guint32_le
 204 item_to_index (GvdbItem *item)
 205 {
 206   if (item != NULL)
 207     return item->assigned_index;
 208
 209   return guint32_to_le (-1u);
 210 }
 211
 212 typedef struct
 213 {
 214   GQueue *chunks;
 215   guint64 offset;
 216   gboolean byteswap;
 217 } FileBuilder;
 218
 219 typedef struct
 220 {
 221   gsize offset;
 222   gsize size;
 223   gpointer data;
 224 } FileChunk;
 225
 226 static gpointer
 227 file_builder_allocate (FileBuilder         *fb,
 228                        guint                alignment,
 229                        gsize                size,
 230                        struct gvdb_pointer *pointer)
 231 {
 232   FileChunk *chunk;
 233
 234   if (size == 0)
 235     return NULL;
 236
 237   fb->offset += (-fb->offset) & (alignment - 1);
 238   chunk = g_slice_new (FileChunk);
 239   chunk->offset = fb->offset;
 240   chunk->size = size;
 241   chunk->data = g_malloc (size);
 242
 243   pointer->start = guint32_to_le (fb->offset);
 244   fb->offset += size;
 245   pointer->end = guint32_to_le (fb->offset);
 246
 247   g_queue_push_tail (fb->chunks, chunk);
 248
 249   return chunk->data;
 250 }
 251
 252 static void
 253 file_builder_add_value (FileBuilder         *fb,
 254                         GVariant            *value,
 255                         struct gvdb_pointer *pointer)
 256 {
 257   GVariant *variant, *normal;
 258   gpointer data;
 259   gsize size;
 260
 261   if (fb->byteswap)
 262     {
 263       value = g_variant_byteswap (value);
 264       variant = g_variant_new_variant (value);
 265       g_variant_unref (value);
 266     }
 267   else
 268     variant = g_variant_new_variant (value);
 269
 270   normal = g_variant_get_normal_form (variant);
 271   g_variant_unref (variant);
 272
 273   size = g_variant_get_size (normal);
 274   data = file_builder_allocate (fb, 8, size, pointer);
 275   g_variant_store (normal, data);
 276   g_variant_unref (normal);
 277 }
 278
 279 static void
 280 file_builder_add_string (FileBuilder *fb,
 281                          const gchar *string,
 282                          guint32_le  *start,
 283                          guint16_le  *size)
 284 {
 285   FileChunk *chunk;
 286   gsize length;
 287
 288   length = strlen (string);
 289
 290   chunk = g_slice_new (FileChunk);
 291   chunk->offset = fb->offset;
 292   chunk->size = length;
 293   chunk->data = g_malloc (length);
 294   if (length != 0)
 295     memcpy (chunk->data, string, length);
 296
 297   *start = guint32_to_le (fb->offset);
 298   *size = guint16_to_le (length);
 299   fb->offset += length;
 300
 301   g_queue_push_tail (fb->chunks, chunk);
 302 }
 303
 304 static void
 305 file_builder_allocate_for_hash (FileBuilder            *fb,
 306                                 gsize                   n_buckets,
 307                                 gsize                   n_items,
 308                                 guint                   bloom_shift,
 309                                 gsize                   n_bloom_words,
 310                                 guint32_le            **bloom_filter,
 311                                 guint32_le            **hash_buckets,
 312                                 struct gvdb_hash_item **hash_items,
 313                                 struct gvdb_pointer    *pointer)
 314 {
 315   guint32_le bloom_hdr, table_hdr;
 316   guchar *data;
 317   gsize size;
 318
 319   g_assert (n_bloom_words < (1u << 27));
 320
 321   bloom_hdr = guint32_to_le (bloom_shift << 27 | n_bloom_words);
 322   table_hdr = guint32_to_le (n_buckets);
 323
 324   size = sizeof bloom_hdr + sizeof table_hdr +
 325          n_bloom_words * sizeof (guint32_le) +
 326          n_buckets     * sizeof (guint32_le) +
 327          n_items       * sizeof (struct gvdb_hash_item);
 328
 329   data = file_builder_allocate (fb, 4, size, pointer);
 330
 331 #define chunk(s) (size -= (s), data += (s), data - (s))
 332   memcpy (chunk (sizeof bloom_hdr), &bloom_hdr, sizeof bloom_hdr);
 333   memcpy (chunk (sizeof table_hdr), &table_hdr, sizeof table_hdr);
 334   *bloom_filter = (guint32_le *) chunk (n_bloom_words * sizeof (guint32_le));
 335   *hash_buckets = (guint32_le *) chunk (n_buckets * sizeof (guint32_le));
 336   *hash_items = (struct gvdb_hash_item *) chunk (n_items *
 337                   sizeof (struct gvdb_hash_item));
 338   g_assert (size == 0);
 339 #undef chunk
 340
 341   memset (*bloom_filter, 0, n_bloom_words * sizeof (guint32_le));
 342
 343   /* NOTE - the code to actually fill in the bloom filter here is missing.
 344    * Patches welcome!
 345    *
 346    * http://en.wikipedia.org/wiki/Bloom_filter
 347    * http://0pointer.de/blog/projects/bloom.html
 348    */
 349 }
 350
 351 static void
 352 file_builder_add_hash (FileBuilder         *fb,
 353                        GHashTable          *table,
 354                        struct gvdb_pointer *pointer)
 355 {
 356   guint32_le *buckets, *bloom_filter;
 357   struct gvdb_hash_item *items;
 358   HashTable *mytable;
 359   GvdbItem *item;
 360   guint32 index;
 361   gint bucket;
 362
 363   mytable = hash_table_new (g_hash_table_size (table));
 364   g_hash_table_foreach (table, hash_table_insert, mytable);
 365   index = 0;
 366
 367   for (bucket = 0; bucket < mytable->n_buckets; bucket++)
 368     for (item = mytable->buckets[bucket]; item; item = item->next)
 369       item->assigned_index = guint32_to_le (index++);
 370
 371   file_builder_allocate_for_hash (fb, mytable->n_buckets, index, 5, 0,
 372                                   &bloom_filter, &buckets, &items, pointer);
 373
 374   index = 0;
 375   for (bucket = 0; bucket < mytable->n_buckets; bucket++)
 376     {
 377       buckets[bucket] = guint32_to_le (index);
 378
 379       for (item = mytable->buckets[bucket]; item; item = item->next)
 380         {
 381           struct gvdb_hash_item *entry = items++;
 382           const gchar *basename;
 383
 384           g_assert (index == guint32_from_le (item->assigned_index));
 385           entry->hash_value = guint32_to_le (item->hash_value);
 386           entry->parent = item_to_index (item->parent);
 387           entry->unused = 0;
 388
 389           if (item->parent != NULL)
 390             basename = item->key + strlen (item->parent->key);
 391           else
 392             basename = item->key;
 393
 394           file_builder_add_string (fb, basename,
 395                                    &entry->key_start,
 396                                    &entry->key_size);
 397
 398           if (item->value != NULL)
 399             {
 400               g_assert (item->child == NULL && item->table == NULL);
 401
 402               file_builder_add_value (fb, item->value, &entry->value.pointer);
 403               entry->type = 'v';
 404             }
 405
 406           if (item->child != NULL)
 407             {
 408               guint32 children = 0, i = 0;
 409               guint32_le *offsets;
 410               GvdbItem *child;
 411
 412               g_assert (item->table == NULL);
 413
 414               for (child = item->child; child; child = child->sibling)
 415                 children++;
 416
 417               offsets = file_builder_allocate (fb, 4, 4 * children,
 418                                                &entry->value.pointer);
 419               entry->type = 'L';
 420
 421               for (child = item->child; child; child = child->sibling)
 422                 offsets[i++] = child->assigned_index;
 423
 424               g_assert (children == i);
 425             }
 426
 427           if (item->table != NULL)
 428             {
 429               entry->type = 'H';
 430               file_builder_add_hash (fb, item->table, &entry->value.pointer);
 431             }
 432
 433           index++;
 434         }
 435     }
 436
 437   hash_table_free (mytable);
 438 }
 439
 440 static FileBuilder *
 441 file_builder_new (gboolean byteswap)
 442 {
 443   FileBuilder *builder;
 444
 445   builder = g_slice_new (FileBuilder);
 446   builder->chunks = g_queue_new ();
 447   builder->offset = sizeof (struct gvdb_header);
 448   builder->byteswap = byteswap;
 449
 450   return builder;
 451 }
 452
 453 static GString *
 454 file_builder_serialise (FileBuilder          *fb,
 455                         struct gvdb_pointer   root)
 456 {
 457   struct gvdb_header header = { { 0, }, };
 458   GString *result;
 459
 460   if (fb->byteswap)
 461     {
 462       header.signature[0] = GVDB_SWAPPED_SIGNATURE0;
 463       header.signature[1] = GVDB_SWAPPED_SIGNATURE1;
 464     }
 465   else
 466     {
 467       header.signature[0] = GVDB_SIGNATURE0;
 468       header.signature[1] = GVDB_SIGNATURE1;
 469     }
 470
 471   result = g_string_new (NULL);
 472
 473   header.root = root;
 474   g_string_append_len (result, (gpointer) &header, sizeof header);
 475
 476   while (!g_queue_is_empty (fb->chunks))
 477     {
 478       FileChunk *chunk = g_queue_pop_head (fb->chunks);
 479
 480       if (result->len != chunk->offset)
 481         {
 482           gchar zero[8] = { 0, };
 483
 484           g_assert (chunk->offset > result->len);
 485           g_assert (chunk->offset - result->len < 8);
 486
 487           g_string_append_len (result, zero, chunk->offset - result->len);
 488           g_assert (result->len == chunk->offset);
 489         }
 490
 491       g_string_append_len (result, chunk->data, chunk->size);
 492       g_free (chunk->data);
 493
 494       g_slice_free (FileChunk, chunk);
 495     }
 496
 497   g_queue_free (fb->chunks);
 498   g_slice_free (FileBuilder, fb);
 499
 500   return result;
 501 }
 502
 503 gboolean
 504 gvdb_table_write_contents (GHashTable   *table,
 505                            const gchar  *filename,
 506                            gboolean      byteswap,
 507                            GError      **error)
 508 {
 509   struct gvdb_pointer root;
 510   gboolean status;
 511   FileBuilder *fb;
 512   GString *str;
 513
 514   fb = file_builder_new (byteswap);
 515   file_builder_add_hash (fb, table, &root);
 516   str = file_builder_serialise (fb, root);
 517
 518   status = g_file_set_contents (filename, str->str, str->len, error);
 519   g_string_free (str, TRUE);
 520
 521   return status;
 522 }