diff options
Diffstat (limited to 'libs/libcurl/src/hash.c')
-rw-r--r-- | libs/libcurl/src/hash.c | 400 |
1 files changed, 400 insertions, 0 deletions
diff --git a/libs/libcurl/src/hash.c b/libs/libcurl/src/hash.c new file mode 100644 index 0000000000..4a12e1a7b9 --- /dev/null +++ b/libs/libcurl/src/hash.c @@ -0,0 +1,400 @@ +/*************************************************************************** + * _ _ ____ _ + * Project ___| | | | _ \| | + * / __| | | | |_) | | + * | (__| |_| | _ <| |___ + * \___|\___/|_| \_\_____| + * + * Copyright (C) 1998 - 2013, Daniel Stenberg, <daniel@haxx.se>, et al. + * + * This software is licensed as described in the file COPYING, which + * you should have received as part of this distribution. The terms + * are also available at http://curl.haxx.se/docs/copyright.html. + * + * You may opt to use, copy, modify, merge, publish, distribute and/or sell + * copies of the Software, and permit persons to whom the Software is + * furnished to do so, under the terms of the COPYING file. + * + * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY + * KIND, either express or implied. + * + ***************************************************************************/ + +#include "curl_setup.h" + +#include "hash.h" +#include "llist.h" + +#define _MPRINTF_REPLACE /* use our functions only */ +#include <curl/mprintf.h> + +#include "curl_memory.h" +/* The last #include file should be: */ +#include "memdebug.h" + +static void +hash_element_dtor(void *user, void *element) +{ + struct curl_hash *h = (struct curl_hash *) user; + struct curl_hash_element *e = (struct curl_hash_element *) element; + + Curl_safefree(e->key); + + if(e->ptr) { + h->dtor(e->ptr); + e->ptr = NULL; + } + + e->key_len = 0; + + free(e); +} + +/* return 1 on error, 0 is fine */ +int +Curl_hash_init(struct curl_hash *h, + int slots, + hash_function hfunc, + comp_function comparator, + curl_hash_dtor dtor) +{ + int i; + + if(!slots || !hfunc || !comparator ||!dtor) { + return 1; /* failure */ + } + + h->hash_func = hfunc; + h->comp_func = comparator; + h->dtor = dtor; + h->size = 0; + h->slots = slots; + + h->table = malloc(slots * sizeof(struct curl_llist *)); + if(h->table) { + for(i = 0; i < slots; ++i) { + h->table[i] = Curl_llist_alloc((curl_llist_dtor) hash_element_dtor); + if(!h->table[i]) { + while(i--) { + Curl_llist_destroy(h->table[i], NULL); + h->table[i] = NULL; + } + free(h->table); + h->table = NULL; + h->slots = 0; + return 1; /* failure */ + } + } + return 0; /* fine */ + } + else { + h->slots = 0; + return 1; /* failure */ + } +} + +struct curl_hash * +Curl_hash_alloc(int slots, + hash_function hfunc, + comp_function comparator, + curl_hash_dtor dtor) +{ + struct curl_hash *h; + + if(!slots || !hfunc || !comparator ||!dtor) { + return NULL; /* failure */ + } + + h = malloc(sizeof(struct curl_hash)); + if(h) { + if(Curl_hash_init(h, slots, hfunc, comparator, dtor)) { + /* failure */ + free(h); + h = NULL; + } + } + + return h; +} + + + +static struct curl_hash_element * +mk_hash_element(const void *key, size_t key_len, const void *p) +{ + struct curl_hash_element *he = malloc(sizeof(struct curl_hash_element)); + + if(he) { + void *dupkey = malloc(key_len); + if(dupkey) { + /* copy the key */ + memcpy(dupkey, key, key_len); + + he->key = dupkey; + he->key_len = key_len; + he->ptr = (void *) p; + } + else { + /* failed to duplicate the key, free memory and fail */ + free(he); + he = NULL; + } + } + return he; +} + +#define FETCH_LIST(x,y,z) x->table[x->hash_func(y, z, x->slots)] + +/* Insert the data in the hash. If there already was a match in the hash, + * that data is replaced. + * + * @unittest: 1305 + */ +void * +Curl_hash_add(struct curl_hash *h, void *key, size_t key_len, void *p) +{ + struct curl_hash_element *he; + struct curl_llist_element *le; + struct curl_llist *l = FETCH_LIST (h, key, key_len); + + for(le = l->head; le; le = le->next) { + he = (struct curl_hash_element *) le->ptr; + if(h->comp_func(he->key, he->key_len, key, key_len)) { + Curl_llist_remove(l, le, (void *)h); + --h->size; + break; + } + } + + he = mk_hash_element(key, key_len, p); + if(he) { + if(Curl_llist_insert_next(l, l->tail, he)) { + ++h->size; + return p; /* return the new entry */ + } + /* + * Couldn't insert it, destroy the 'he' element and the key again. We + * don't call hash_element_dtor() since that would also call the + * "destructor" for the actual data 'p'. When we fail, we shall not touch + * that data. + */ + free(he->key); + free(he); + } + + return NULL; /* failure */ +} + +/* remove the identified hash entry, returns non-zero on failure */ +int Curl_hash_delete(struct curl_hash *h, void *key, size_t key_len) +{ + struct curl_llist_element *le; + struct curl_hash_element *he; + struct curl_llist *l = FETCH_LIST(h, key, key_len); + + for(le = l->head; le; le = le->next) { + he = le->ptr; + if(h->comp_func(he->key, he->key_len, key, key_len)) { + Curl_llist_remove(l, le, (void *) h); + --h->size; + return 0; + } + } + return 1; +} + +void * +Curl_hash_pick(struct curl_hash *h, void *key, size_t key_len) +{ + struct curl_llist_element *le; + struct curl_hash_element *he; + struct curl_llist *l; + + if(h) { + l = FETCH_LIST(h, key, key_len); + for(le = l->head; le; le = le->next) { + he = le->ptr; + if(h->comp_func(he->key, he->key_len, key, key_len)) { + return he->ptr; + } + } + } + + return NULL; +} + +#if defined(DEBUGBUILD) && defined(AGGRESIVE_TEST) +void +Curl_hash_apply(curl_hash *h, void *user, + void (*cb)(void *user, void *ptr)) +{ + struct curl_llist_element *le; + int i; + + for(i = 0; i < h->slots; ++i) { + for(le = (h->table[i])->head; + le; + le = le->next) { + curl_hash_element *el = le->ptr; + cb(user, el->ptr); + } + } +} +#endif + +void +Curl_hash_clean(struct curl_hash *h) +{ + int i; + + for(i = 0; i < h->slots; ++i) { + Curl_llist_destroy(h->table[i], (void *) h); + h->table[i] = NULL; + } + + Curl_safefree(h->table); + h->size = 0; + h->slots = 0; +} + +void +Curl_hash_clean_with_criterium(struct curl_hash *h, void *user, + int (*comp)(void *, void *)) +{ + struct curl_llist_element *le; + struct curl_llist_element *lnext; + struct curl_llist *list; + int i; + + if(!h) + return; + + for(i = 0; i < h->slots; ++i) { + list = h->table[i]; + le = list->head; /* get first list entry */ + while(le) { + struct curl_hash_element *he = le->ptr; + lnext = le->next; + /* ask the callback function if we shall remove this entry or not */ + if(comp(user, he->ptr)) { + Curl_llist_remove(list, le, (void *) h); + --h->size; /* one less entry in the hash now */ + } + le = lnext; + } + } +} + +void +Curl_hash_destroy(struct curl_hash *h) +{ + if(!h) + return; + + Curl_hash_clean(h); + + free(h); +} + +size_t Curl_hash_str(void* key, size_t key_length, size_t slots_num) +{ + const char* key_str = (const char *) key; + const char *end = key_str + key_length; + unsigned long h = 5381; + + while(key_str < end) { + h += h << 5; + h ^= (unsigned long) *key_str++; + } + + return (h % slots_num); +} + +size_t Curl_str_key_compare(void*k1, size_t key1_len, void*k2, size_t key2_len) +{ + char *key1 = (char *)k1; + char *key2 = (char *)k2; + + if(key1_len == key2_len && + *key1 == *key2 && + memcmp(key1, key2, key1_len) == 0) { + return 1; + } + + return 0; +} + +void Curl_hash_start_iterate(struct curl_hash *hash, + struct curl_hash_iterator *iter) +{ + iter->hash = hash; + iter->slot_index = 0; + iter->current_element = NULL; +} + +struct curl_hash_element * +Curl_hash_next_element(struct curl_hash_iterator *iter) +{ + int i; + struct curl_hash *h = iter->hash; + + /* Get the next element in the current list, if any */ + if(iter->current_element) + iter->current_element = iter->current_element->next; + + /* If we have reached the end of the list, find the next one */ + if(!iter->current_element) { + for(i = iter->slot_index;i < h->slots;i++) { + if(h->table[i]->head) { + iter->current_element = h->table[i]->head; + iter->slot_index = i+1; + break; + } + } + } + + if(iter->current_element) { + struct curl_hash_element *he = iter->current_element->ptr; + return he; + } + else { + iter->current_element = NULL; + return NULL; + } +} + +#if 0 /* useful function for debugging hashes and their contents */ +void Curl_hash_print(struct curl_hash *h, + void (*func)(void *)) +{ + struct curl_hash_iterator iter; + struct curl_hash_element *he; + int last_index = -1; + + if(!h) + return; + + fprintf(stderr, "=Hash dump=\n"); + + Curl_hash_start_iterate(h, &iter); + + he = Curl_hash_next_element(&iter); + while(he) { + if(iter.slot_index != last_index) { + fprintf(stderr, "index %d:", iter.slot_index); + if(last_index >= 0) { + fprintf(stderr, "\n"); + } + last_index = iter.slot_index; + } + + if(func) + func(he->ptr); + else + fprintf(stderr, " [%p]", (void *)he->ptr); + + he = Curl_hash_next_element(&iter); + } + fprintf(stderr, "\n"); +} +#endif |