From c98912e0aec64f91108b89b917c68cbc761a5dc1 Mon Sep 17 00:00:00 2001 From: fc_botelho Date: Fri, 29 Jul 2005 03:09:31 +0000 Subject: [PATCH] it was fixed more mistakes in BRZ algorithm --- src/brz.c | 47 +++++++++++++++++++++------------------------- src/djb2_hash.c | 7 +++++++ src/djb2_hash.h | 1 + src/fnv_hash.c | 7 +++++++ src/fnv_hash.h | 1 + src/hash.c | 24 +++++++++++++++++++++++ src/hash.h | 1 + src/jenkins_hash.c | 11 +++++++++++ src/jenkins_hash.h | 1 + src/sdbm_hash.c | 7 +++++++ src/sdbm_hash.h | 1 + 11 files changed, 82 insertions(+), 26 deletions(-) diff --git a/src/brz.c b/src/brz.c index 4c0756d..fb27277 100755 --- a/src/brz.c +++ b/src/brz.c @@ -19,7 +19,7 @@ static int brz_before_gen_graphs(cmph_config_t *mph, cmph_uint32 * disksize, cmph_uint32 * diskoffset); static void brz_gen_graphs(cmph_config_t *mph, cmph_uint32 * disksize, cmph_uint32 * diskoffset, FILE * graphs_fd); -static char ** brz_read_keys_vd(FILE * graphs_fd, cmph_uint8 nkeys, cmph_uint32 h3); +static char ** brz_read_keys_vd(FILE * graphs_fd, cmph_uint8 nkeys); static void brz_destroy_keys_vd(char ** keys_vd, cmph_uint8 nkeys); static void brz_copy_partial_mphf(brz_config_data_t *brz, bmz_data_t * bmzf, cmph_uint32 index); @@ -128,7 +128,8 @@ cmph_t *brz_new(cmph_config_t *mph, float c) return NULL; } - graphs_fd = fopen("/colecao/fbotelho/cmph.tmp", "wb"); +// graphs_fd = fopen("/colecao/fbotelho/cmph.tmp", "wb"); + graphs_fd = fopen("cmph.tmp", "wb"); if (graphs_fd == NULL) { free(brz->size); @@ -144,7 +145,8 @@ cmph_t *brz_new(cmph_config_t *mph, float c) free(diskoffset); DEBUGP("Graphs generated\n"); fclose(graphs_fd); - graphs_fd = fopen("/colecao/fbotelho/cmph.tmp", "rb"); +// graphs_fd = fopen("/colecao/fbotelho/cmph.tmp", "rb"); + graphs_fd = fopen("cmph.tmp", "rb"); // codigo do algoritmo... brz->h1 = (hash_state_t **)malloc(sizeof(hash_state_t *)*brz->k); brz->h2 = (hash_state_t **)malloc(sizeof(hash_state_t *)*brz->k); @@ -155,7 +157,7 @@ cmph_t *brz_new(cmph_config_t *mph, float c) cmph_uint32 j; bmz_data_t * bmzf = NULL; if (brz->size[i] == 0) continue; - keys_vd = brz_read_keys_vd(graphs_fd, brz->size[i], i); + keys_vd = brz_read_keys_vd(graphs_fd, brz->size[i]); // Source of keys source = cmph_io_vector_adapter(keys_vd, (cmph_uint32)brz->size[i]); config = cmph_config_new(source); @@ -214,11 +216,11 @@ static int brz_before_gen_graphs(cmph_config_t *mph, cmph_uint32 * disksize, cmp char *key; mph->key_source->read(mph->key_source->data, &key, &keylen); h3 = hash(brz->h3, key, keylen) % brz->k; - if(h3 == 6) - { - DEBUGP("key = %s\n", key); - DEBUGP("keylen = %u\n", keylen + 1); - } +// if(h3 == 6) +// { +// DEBUGP("key = %s\n", key); +// DEBUGP("keylen = %u\n", keylen + 1); +// } mph->key_source->dispose(mph->key_source->data, key, keylen); if (brz->size[h3] == 255) return 0; @@ -230,13 +232,13 @@ static int brz_before_gen_graphs(cmph_config_t *mph, cmph_uint32 * disksize, cmp // } } - fprintf(stderr, "size:%u offset: %u\n", brz->size[0], brz->offset[0]); +// DEBUGP("size:%u offset: %u\n", brz->size[0], brz->offset[0]); for (e = 1; e < brz->k; ++e) { brz->offset[e] = brz->size[e-1] + brz->offset[e-1]; diskoffset[e] = disksize[e-1] + diskoffset[e-1]; - DEBUGP("disksize[%u]=%u diskoffset[%u]: %u\n", e, disksize[e], e, diskoffset[e]); - DEBUGP("size[%u]=%u offset[%u]: %u\n", e, brz->size[e], e, brz->offset[e]); +/* DEBUGP("disksize[%u]=%u diskoffset[%u]: %u\n", e, disksize[e], e, diskoffset[e]); + DEBUGP("size[%u]=%u offset[%u]: %u\n", e, brz->size[e], e, brz->offset[e]);*/ } return 1; } @@ -255,23 +257,23 @@ static void brz_gen_graphs(cmph_config_t *mph, cmph_uint32 * disksize, cmph_uint char *key; mph->key_source->read(mph->key_source->data, &key, &keylen); h3 = hash(brz->h3, key, keylen) % brz->k; - if(h3 == 6) +/* if(h3 == 6) { DEBUGP("key = %s\n", key); DEBUGP("keylen = %u\n", keylen + 1); - } + }*/ fseek(graphs_fd, diskoffset[h3], SEEK_SET); fwrite(key, sizeof(char), keylen + 1, graphs_fd); - if(h3 == 6) +/* if(h3 == 6) { DEBUGP("diskoffset[%u]=%u \n", h3, diskoffset[h3]); - } + }*/ diskoffset[h3] = diskoffset[h3] + keylen + 1; mph->key_source->dispose(mph->key_source->data, key, keylen); } } -static char ** brz_read_keys_vd(FILE * graphs_fd, cmph_uint8 nkeys, cmph_uint32 h3) +static char ** brz_read_keys_vd(FILE * graphs_fd, cmph_uint8 nkeys) { char ** keys_vd = (char **)malloc(sizeof(char *)*nkeys); cmph_uint8 i; @@ -291,7 +293,6 @@ static char ** brz_read_keys_vd(FILE * graphs_fd, cmph_uint8 nkeys, cmph_uint32 } keys_vd[i] = (char *)malloc(strlen(buf) + 1); strcpy(keys_vd[i], buf); - if(h3 == 6) DEBUGP("key = %s\n", keys_vd[i]); free(buf); } return keys_vd; @@ -311,8 +312,8 @@ static void brz_copy_partial_mphf(brz_config_data_t *brz, bmz_data_t * bmzf, cmp { brz->g[index][i] = (cmph_uint8) bmzf->g[i]; } - brz->h1[index] = bmzf->hashes[0]; - brz->h2[index] = bmzf->hashes[1]; + brz->h1[index] = hash_state_copy(bmzf->hashes[0]); + brz->h2[index] = hash_state_copy(bmzf->hashes[1]); } int brz_dump(cmph_t *mphf, FILE *fd) @@ -411,19 +412,13 @@ void brz_destroy(cmph_t *mphf) { cmph_uint32 i; brz_data_t *data = (brz_data_t *)mphf->data; - fprintf(stderr, "MERDAAAAAA %u\n", data->h3->hashfunc); for(i = 0; i < data->k; i++) { free(data->g[i]); - fprintf(stderr, "MERDAAAAAA1 %u\n", data->h3->hashfunc); hash_state_destroy(data->h1[i]); - fprintf(stderr, "MERDAAAAAA2 %u\n", data->h3->hashfunc); hash_state_destroy(data->h2[i]); - fprintf(stderr, "MERDAAAAAA3 %u\n", data->h3->hashfunc); } - fprintf(stderr, "MERDAAAAAA %u\n", data->h3->hashfunc); hash_state_destroy(data->h3); - fprintf(stderr, "MERDAAAAAA1FAB\n"); free(data->g); free(data->h1); free(data->h2); diff --git a/src/djb2_hash.c b/src/djb2_hash.c index 5bbd9bd..923bbd8 100644 --- a/src/djb2_hash.c +++ b/src/djb2_hash.c @@ -34,6 +34,13 @@ void djb2_state_dump(djb2_state_t *state, char **buf, cmph_uint32 *buflen) return; } +djb2_state_t *djb2_state_copy(djb2_state_t *src_state) +{ + djb2_state_t *dest_state = (djb2_state_t *)malloc(sizeof(djb2_state_t)); + dest_state->hashfunc = src_state->hashfunc; + return dest_state; +} + djb2_state_t *djb2_state_load(const char *buf, cmph_uint32 buflen) { djb2_state_t *state = (djb2_state_t *)malloc(sizeof(djb2_state_t)); diff --git a/src/djb2_hash.h b/src/djb2_hash.h index 3fef71d..dda97e3 100644 --- a/src/djb2_hash.h +++ b/src/djb2_hash.h @@ -11,6 +11,7 @@ typedef struct __djb2_state_t djb2_state_t *djb2_state_new(); cmph_uint32 djb2_hash(djb2_state_t *state, const char *k, cmph_uint32 keylen); void djb2_state_dump(djb2_state_t *state, char **buf, cmph_uint32 *buflen); +djb2_state_t *djb2_state_copy(djb2_state_t *src_state); djb2_state_t *djb2_state_load(const char *buf, cmph_uint32 buflen); void djb2_state_destroy(djb2_state_t *state); diff --git a/src/fnv_hash.c b/src/fnv_hash.c index 5bf8db2..aeaca8f 100644 --- a/src/fnv_hash.c +++ b/src/fnv_hash.c @@ -38,6 +38,13 @@ void fnv_state_dump(fnv_state_t *state, char **buf, cmph_uint32 *buflen) return; } +fnv_state_t * fnv_state_copy(fnv_state_t *src_state) +{ + fnv_state_t *dest_state = (fnv_state_t *)malloc(sizeof(fnv_state_t)); + dest_state->hashfunc = src_state->hashfunc; + return dest_state; +} + fnv_state_t *fnv_state_load(const char *buf, cmph_uint32 buflen) { fnv_state_t *state = (fnv_state_t *)malloc(sizeof(fnv_state_t)); diff --git a/src/fnv_hash.h b/src/fnv_hash.h index e73e0f6..7f57946 100644 --- a/src/fnv_hash.h +++ b/src/fnv_hash.h @@ -11,6 +11,7 @@ typedef struct __fnv_state_t fnv_state_t *fnv_state_new(); cmph_uint32 fnv_hash(fnv_state_t *state, const char *k, cmph_uint32 keylen); void fnv_state_dump(fnv_state_t *state, char **buf, cmph_uint32 *buflen); +fnv_state_t *fnv_state_copy(fnv_state_t *src_state); fnv_state_t *fnv_state_load(const char *buf, cmph_uint32 buflen); void fnv_state_destroy(fnv_state_t *state); diff --git a/src/hash.c b/src/hash.c index 868e25c..f09c47b 100644 --- a/src/hash.c +++ b/src/hash.c @@ -86,6 +86,30 @@ void hash_state_dump(hash_state_t *state, char **buf, cmph_uint32 *buflen) return; } +hash_state_t * hash_state_copy(hash_state_t *src_state) +{ + hash_state_t *dest_state = NULL; + switch (src_state->hashfunc) + { + case CMPH_HASH_JENKINS: + dest_state = (hash_state_t *)jenkins_state_copy((jenkins_state_t *)src_state); + break; + case CMPH_HASH_DJB2: + dest_state = (hash_state_t *)djb2_state_copy((djb2_state_t *)src_state); + break; + case CMPH_HASH_SDBM: + dest_state = (hash_state_t *)sdbm_state_copy((sdbm_state_t *)src_state); + break; + case CMPH_HASH_FNV: + dest_state = (hash_state_t *)fnv_state_copy((fnv_state_t *)src_state); + break; + default: + assert(0); + } + dest_state->hashfunc = src_state->hashfunc; + return dest_state; +} + hash_state_t *hash_state_load(const char *buf, cmph_uint32 buflen) { cmph_uint32 i; diff --git a/src/hash.h b/src/hash.h index 2397c33..2735a22 100644 --- a/src/hash.h +++ b/src/hash.h @@ -8,6 +8,7 @@ typedef union __hash_state_t hash_state_t; hash_state_t *hash_state_new(CMPH_HASH, cmph_uint32 hashsize); cmph_uint32 hash(hash_state_t *state, const char *key, cmph_uint32 keylen); void hash_state_dump(hash_state_t *state, char **buf, cmph_uint32 *buflen); +hash_state_t * hash_state_copy(hash_state_t *src_state); hash_state_t *hash_state_load(const char *buf, cmph_uint32 buflen); void hash_state_destroy(hash_state_t *state); diff --git a/src/jenkins_hash.c b/src/jenkins_hash.c index c849ec8..4ed64f5 100644 --- a/src/jenkins_hash.c +++ b/src/jenkins_hash.c @@ -178,6 +178,17 @@ void jenkins_state_dump(jenkins_state_t *state, char **buf, cmph_uint32 *buflen) return; } + +jenkins_state_t *jenkins_state_copy(jenkins_state_t *src_state) +{ + jenkins_state_t *dest_state = (jenkins_state_t *)malloc(sizeof(jenkins_state_t)); + dest_state->hashfunc = src_state->hashfunc; + dest_state->seed = src_state->seed; + dest_state->nbits = src_state->nbits; + dest_state->size = src_state->size; + return dest_state; +} + jenkins_state_t *jenkins_state_load(const char *buf, cmph_uint32 buflen) { jenkins_state_t *state = (jenkins_state_t *)malloc(sizeof(jenkins_state_t)); diff --git a/src/jenkins_hash.h b/src/jenkins_hash.h index baa6209..e259ee2 100644 --- a/src/jenkins_hash.h +++ b/src/jenkins_hash.h @@ -14,6 +14,7 @@ typedef struct __jenkins_state_t jenkins_state_t *jenkins_state_new(cmph_uint32 size); //size of hash table cmph_uint32 jenkins_hash(jenkins_state_t *state, const char *k, cmph_uint32 keylen); void jenkins_state_dump(jenkins_state_t *state, char **buf, cmph_uint32 *buflen); +jenkins_state_t *jenkins_state_copy(jenkins_state_t *src_state); jenkins_state_t *jenkins_state_load(const char *buf, cmph_uint32 buflen); void jenkins_state_destroy(jenkins_state_t *state); diff --git a/src/sdbm_hash.c b/src/sdbm_hash.c index bddcce6..b854478 100644 --- a/src/sdbm_hash.c +++ b/src/sdbm_hash.c @@ -34,6 +34,13 @@ void sdbm_state_dump(sdbm_state_t *state, char **buf, cmph_uint32 *buflen) return; } +sdbm_state_t *sdbm_state_copy(sdbm_state_t *src_state) +{ + sdbm_state_t *dest_state = (sdbm_state_t *)malloc(sizeof(sdbm_state_t)); + dest_state->hashfunc = src_state->hashfunc; + return dest_state; +} + sdbm_state_t *sdbm_state_load(const char *buf, cmph_uint32 buflen) { sdbm_state_t *state = (sdbm_state_t *)malloc(sizeof(sdbm_state_t)); diff --git a/src/sdbm_hash.h b/src/sdbm_hash.h index a05daae..f44b2f1 100644 --- a/src/sdbm_hash.h +++ b/src/sdbm_hash.h @@ -11,6 +11,7 @@ typedef struct __sdbm_state_t sdbm_state_t *sdbm_state_new(); cmph_uint32 sdbm_hash(sdbm_state_t *state, const char *k, cmph_uint32 keylen); void sdbm_state_dump(sdbm_state_t *state, char **buf, cmph_uint32 *buflen); +sdbm_state_t *sdbm_state_copy(sdbm_state_t *src_state); sdbm_state_t *sdbm_state_load(const char *buf, cmph_uint32 buflen); void sdbm_state_destroy(sdbm_state_t *state);