path: root/common/hash.c

/*
 * Originally from apache 2.0
 * Modifications for general use by <nielsen@memberwebs.com>
 */

/* Copyright 2000-2004 The Apache Software Foundation
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#include <sys/types.h>
#include <stdlib.h>
#include "hash.h"

#ifdef HASH_TIMESTAMP
#include <time.h>
#endif

#ifdef HASH_COPYKEYS
  #define KEY_DATA(he) (void*)(((unsigned char*)(he)) + sizeof(*(he)))
#else
  #define KEY_DATA(he)  ((he)->key)
#endif

/*
 * The internal form of a hash table.
 *
 * The table is an array indexed by the hash of the key; collisions
 * are resolved by hanging a linked list of hash entries off each
 * element of the array. Although this is a really simple design it
 * isn't too bad given that pools have a low allocation overhead.
 */

typedef struct hash_entry_t hash_entry_t;

struct hash_entry_t
{
  hash_entry_t* next;
  unsigned int hash;
#ifndef HASH_COPYKEYS
  const void* key;
  size_t klen;
#endif
  const void* val;
#ifdef HASH_TIMESTAMP
  time_t stamp;
#endif
};

/*
 * Data structure for iterating through a hash table.
 *
 * We keep a pointer to the next hash entry here to allow the current
 * hash entry to be freed or otherwise mangled between calls to
 * hash_next().
 */
struct hash_index_t
{
  hash_t* ht;
  hash_entry_t* ths;
  hash_entry_t* next;
  unsigned int index;
};

/*
 * The size of the array is always a power of two. We use the maximum
 * index rather than the size so that we can use bitwise-AND for
 * modular arithmetic.
 * The count of hash entries may be greater depending on the chosen
 * collision rate.
 */
struct hash_t
{
  hash_entry_t** array;
  hash_index_t iterator;  /* For hash_first(...) */
  unsigned int count;
  unsigned int max;
#ifdef HASH_COPYKEYS
  unsigned int klen;
#endif
};

#define INITIAL_MAX 15 /* tunable == 2^n - 1 */


/*
 * Hash creation functions.
 */

static hash_entry_t** alloc_array(hash_t* ht, unsigned int max)
{
  return malloc(sizeof(*(ht->array)) * (max + 1));
}

#ifdef HASH_COPYKEYS
hash_t* hash_create(size_t klen)
#else
hash_t* hash_create()
#endif
{
  hash_t* ht = malloc(sizeof(hash_t));
  if(ht)
  {
    ht->count = 0;
    ht->max = INITIAL_MAX;
    ht->array = alloc_array(ht, ht->max);
#ifdef HASH_COPYKEYS
    ht->klen = klen;
#endif

    if(!ht->array)
    {
      free(ht);
      return NULL;
    }
  }
  return ht;
}

void hash_free(hash_t* ht)
{
  hash_index_t* hi;

  for(hi = hash_first(ht); hi; hi = hash_next(hi))
    free(hi->ths);

  if(ht->array)
    free(ht->array);

  free(ht);
}

/*
 * Hash iteration functions.
 */

hash_index_t* hash_next(hash_index_t* hi)
{
  hi->ths = hi->next;
  while(!hi->ths)
  {
    if(hi->index > hi->ht->max)
      return NULL;

    hi->ths = hi->ht->array[hi->index++];
  }
  hi->next = hi->ths->next;
  return hi;
}

hash_index_t* hash_first(hash_t* ht)
{
  hash_index_t* hi = &ht->iterator;

  hi->ht = ht;
  hi->index = 0;
  hi->ths = NULL;
  hi->next = NULL;
  return hash_next(hi);
}

#ifdef HASH_COPYKEYS
void* hash_this(hash_index_t* hi, const void** key)
#else
void* hash_this(hash_index_t* hi, const void** key, size_t* klen)
#endif
{
  if(key)
    *key = KEY_DATA(hi->ths);

#ifndef HASH_COPYKEYS
  if(klen)
    *klen = hi->ths->klen;
#endif

  return (void*)hi->ths->val;
}


/*
 * Expanding a hash table
 */

static int expand_array(hash_t* ht)
{
  hash_index_t* hi;
  hash_entry_t** new_array;
  unsigned int new_max;

  new_max = ht->max * 2 + 1;
  new_array = alloc_array(ht, new_max);

  if(!new_array)
    return 0;

  for(hi = hash_first(ht); hi; hi = hash_next(hi))
  {
    unsigned int i = hi->ths->hash & new_max;
    hi->ths->next = new_array[i];
    new_array[i] = hi->ths;
  }

  if(ht->array)
    free(ht->array);

  ht->array = new_array;
  ht->max = new_max;
  return 1;
}

/*
 * This is where we keep the details of the hash function and control
 * the maximum collision rate.
 *
 * If val is non-NULL it creates and initializes a new hash entry if
 * there isn't already one there; it returns an updatable pointer so
 * that hash entries can be removed.
 */

#ifdef HASH_COPYKEYS
static hash_entry_t** find_entry(hash_t* ht, const void* key, const void* val)
#else
static hash_entry_t** find_entry(hash_t* ht, const void* key, size_t klen, const void* val)
#endif
{
  hash_entry_t** hep;
  hash_entry_t* he;
  const unsigned char* p;
  unsigned int hash;
  size_t i;

#ifdef HASH_COPYKEYS
  size_t klen = ht->klen;
#endif

  /*
   * This is the popular `times 33' hash algorithm which is used by
   * perl and also appears in Berkeley DB. This is one of the best
   * known hash functions for strings because it is both computed
   * very fast and distributes very well.
   *
   * The originator may be Dan Bernstein but the code in Berkeley DB
   * cites Chris Torek as the source. The best citation I have found
   * is "Chris Torek, Hash function for text in C, Usenet message
   * <27038@mimsy.umd.edu> in comp.lang.c , October, 1990." in Rich
   * Salz's USENIX 1992 paper about INN which can be found at
   * <http://citeseer.nj.nec.com/salz92internetnews.html>.
   *
   * The magic of number 33, i.e. why it works better than many other
   * constants, prime or not, has never been adequately explained by
   * anyone. So I try an explanation: if one experimentally tests all
   * multipliers between 1 and 256 (as I did while writing a low-level
   * data structure library some time ago) one detects that even
   * numbers are not useable at all. The remaining 128 odd numbers
   * (except for the number 1) work more or less all equally well.
   * They all distribute in an acceptable way and this way fill a hash
   * table with an average percent of approx. 86%.
   *
   * If one compares the chi^2 values of the variants (see
   * Bob Jenkins ``Hashing Frequently Asked Questions'' at
   * http://burtleburtle.net/bob/hash/hashfaq.html for a description
   * of chi^2), the number 33 not even has the best value. But the
   * number 33 and a few other equally good numbers like 17, 31, 63,
   * 127 and 129 have nevertheless a great advantage to the remaining
   * numbers in the large set of possible multipliers: their multiply
   * operation can be replaced by a faster operation based on just one
   * shift plus either a single addition or subtraction operation. And
   * because a hash function has to both distribute good _and_ has to
   * be very fast to compute, those few numbers should be preferred.
   *
   *                  -- Ralf S. Engelschall <rse@engelschall.com>
   */
  hash = 0;

#ifndef HASH_COPYKEYS
  if(klen == HASH_KEY_STRING)
  {
    for(p = key; *p; p++)
      hash = hash * 33 + *p;

    klen = p - (const unsigned char *)key;
  }
  else
#endif
  {
    for(p = key, i = klen; i; i--, p++)
      hash = hash * 33 + *p;
  }

  /* scan linked list */
  for(hep = &ht->array[hash & ht->max], he = *hep;
      he; hep = &he->next, he = *hep)
  {
   if(he->hash == hash &&
#ifndef HASH_COPYKEYS
      he->klen == klen &&
#endif
      memcmp(KEY_DATA(he), key, klen) == 0)
     break;
  }

  if(he || !val)
    return hep;

  /* add a new entry for non-NULL val */
#ifdef HASH_COPYKEYS
  he = malloc(sizeof(*he) + klen);
#else
  he = malloc(sizeof(*he));
#endif

  if(he)
  {
#ifdef HASH_COPYKEYS
    /* Key data points past end of entry */
    memcpy(KEY_DATA(he), key, klen);
#else
    /* Key points to external data */
    he->key = key;
    he->klen = klen;
#endif

    he->next = NULL;
    he->hash = hash;
    he->val  = val;

#ifdef HASH_TIMESTAMP
    he->stamp = 0;
#endif

    *hep = he;
    ht->count++;
  }

  return hep;
}

#ifdef HASH_COPYKEYS
void* hash_get(hash_t* ht, const void *key)
{
    hash_entry_t** he = find_entry(ht, key, NULL);
#else
void* hash_get(hash_t* ht, const void *key, size_t klen)
{
    hash_entry_t** he = find_entry(ht, key, klen, NULL);
#endif

    if(he && *he)
      return (void*)((*he)->val);
    else
      return NULL;
}

#ifdef HASH_COPYKEYS
int hash_set(hash_t* ht, const void* key, const void* val)
{
  hash_entry_t** hep = find_entry(ht, key, val);
#else
int hash_set(hash_t* ht, const void* key, size_t klen, const void* val)
{
  hash_entry_t** hep = find_entry(ht, key, klen, val);
#endif

  if(hep && *hep)
  {
    if(val)
    {
      /* replace entry */
      (*hep)->val = val;

#ifdef HASH_TIMESTAMP
      /* Update or set the timestamp */
      (*hep)->stamp = time(NULL);
#endif

      /* check that the collision rate isn't too high */
      if(ht->count > ht->max)
      {
        if(!expand_array(ht))
          return 0;
      }

      return 1;
    }
  }

  return 0;
}

#ifdef HASH_COPYKEYS
void hash_rem(hash_t* ht, const void* key)
{
  hash_entry_t** hep = find_entry(ht, key, NULL);
#else
void hash_rem(hash_t* ht, const void* key, size_t klen)
{
  hash_entry_t** hep = find_entry(ht, key, klen, NULL);
#endif

  if(hep && *hep)
  {
    hash_entry_t* old = *hep;
    *hep = (*hep)->next;
    --ht->count;
    free(old);
  }
}

unsigned int hash_count(hash_t* ht)
{
  return ht->count;
}

#ifdef HASH_TIMESTAMP
int hash_purge(hash_t* ht, time_t stamp)
{
  hash_index_t* hi;
  int r = 0;

  for(hi = hash_first(ht); hi; hi = hash_next(hi))
  {
    if(hi->ths->stamp < stamp)
    {
      /* No need to check for errors as we're deleting */
#ifdef HASH_COPYKEYS
      hash_rem(ht, KEY_DATA(hi->ths));
#else
      hash_rem(ht, hi->ths->key, hi->ths->klen);
#endif

      r++;
    }
  }

  return r;
}

#ifdef HASH_COPYKEYS
void* hash_touch(hash_index_t* hi, const void** key);
{
  hash_entry_t** hep = find_entry(ht, key, NULL);
#else
void* hash_touch(hash_index_t* hi, const void** key, size_t* klen);
{
  hash_entry_t** hep = find_entry(ht, key, klen, NULL);
#endif

  if(he && *he)
    ((*he)->stamp) = time(NULL);
}

#endif /* HASH_TIMESTAMP */