Fixed size hash for some structures.

Fixed size hash lists for termlists, reclists.. Remove global variable
parameters.toget.

diff --git a/src/client.c b/src/client.c
index 95e130a..08cc491 100644 (file)
--- a/src/client.c
+++ b/src/client.c
@@ -520,11 +520,8 @@ void client_start_search(struct client *cl)
     if (*opt_maxrecs)
         ZOOM_connection_option_set(link, "count", opt_maxrecs);
     else
-    {
-        char n[128];
-        sprintf(n, "%d", global_parameters.toget);
-        ZOOM_connection_option_set(link, "count", n);
-    }
+        ZOOM_connection_option_set(link, "count", "100");
+
     if (databaseName)
         ZOOM_connection_option_set(link, "databaseName", databaseName);
 
@@ -721,8 +718,7 @@ int client_parse_query(struct client *cl, const char *query)
         extract_terms(se->nmem, cn, p);
         se->relevance = relevance_create(
             se->service->relevance_pct,
-            se->nmem, (const char **) p,
-            se->expected_maxrecs);
+            se->nmem, (const char **) p);
     }
 
     ccl_rpn_delete(cn);

diff --git a/src/logic.c b/src/logic.c
index 550222e..3ce3091 100644 (file)
--- a/src/logic.c
+++ b/src/logic.c
@@ -77,8 +77,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 struct parameters global_parameters = 
 {
     0,   // dump_records
-    0,   // debug_mode
-    100,
+    0    // debug_mode
 };
 
 static void log_xml_doc(xmlDoc *doc)
@@ -146,8 +145,7 @@ static void add_facet(struct session *s, const char *type, const char *value)
 
         s->termlists[i].name = nmem_strdup(s->nmem, type);
         s->termlists[i].termlist 
-            = termlist_create(s->nmem, s->expected_maxrecs,
-                              TERMLIST_HIGH_SCORE);
+            = termlist_create(s->nmem, TERMLIST_HIGH_SCORE);
         s->num_termlists = i + 1;
     }
     termlist_insert(s->termlists[i].termlist, value);
@@ -493,9 +491,7 @@ enum pazpar2_error_code search(struct session *se,
     live_channels = select_targets(se, criteria);
     if (live_channels)
     {
-        int maxrecs = live_channels * global_parameters.toget; // This is buggy!!!
-        se->reclist = reclist_create(se->nmem, maxrecs);
-        se->expected_maxrecs = maxrecs;
+        se->reclist = reclist_create(se->nmem);
     }
     else
         return PAZPAR2_NO_TARGETS;
@@ -663,7 +659,6 @@ struct session *new_session(NMEM nmem, struct conf_service *service)
     session->num_termlists = 0;
     session->reclist = 0;
     session->clients = 0;
-    session->expected_maxrecs = 0;
     session->session_nmem = nmem;
     session->nmem = nmem_create();
     session->wrbuf = wrbuf_alloc();

diff --git a/src/parameters.h b/src/parameters.h
index 412ed0d..22c65a1 100644 (file)
--- a/src/parameters.h
+++ b/src/parameters.h
@@ -26,7 +26,6 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 struct parameters {
     int dump_records;
     int debug_mode;
-    int toget;
 };
 extern struct parameters global_parameters;
 

diff --git a/src/pazpar2.h b/src/pazpar2.h
index 14815cd..738774e 100644 (file)
--- a/src/pazpar2.h
+++ b/src/pazpar2.h
@@ -125,7 +125,6 @@ struct session {
     struct relevance *relevance;
     struct reclist *reclist;
     struct session_watchentry watchlist[SESSION_WATCH_MAX + 1];
-    int expected_maxrecs;
     int total_hits;
     int total_records;
     int total_merged;

diff --git a/src/reclists.c b/src/reclists.c
index 70b091f..64fafbb 100644 (file)
--- a/src/reclists.c
+++ b/src/reclists.c
@@ -32,8 +32,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 struct reclist
 {
     struct reclist_bucket **hashtable;
-    int hashtable_size;
-    int hashmask;
+    unsigned hash_size;
 
     int num_records;
     struct reclist_bucket *sorted_list;
@@ -217,21 +216,14 @@ void reclist_rewind(struct reclist *l)
         l->sorted_ptr = l->sorted_list;
 }
 
-struct reclist *reclist_create(NMEM nmem, int numrecs)
+struct reclist *reclist_create(NMEM nmem)
 {
-    int hashsize = 1;
-    struct reclist *res;
-
-    assert(numrecs);
-    while (hashsize < numrecs)
-        hashsize <<= 1;
-    res = nmem_malloc(nmem, sizeof(struct reclist));
+    struct reclist *res = nmem_malloc(nmem, sizeof(struct reclist));
+    res->hash_size = 399;
     res->hashtable 
-        = nmem_malloc(nmem, hashsize * sizeof(struct reclist_bucket*));
-    memset(res->hashtable, 0, hashsize * sizeof(struct reclist_bucket*));
-    res->hashtable_size = hashsize;
+        = nmem_malloc(nmem, res->hash_size * sizeof(struct reclist_bucket*));
+    memset(res->hashtable, 0, res->hash_size * sizeof(struct reclist_bucket*));
     res->nmem = nmem;
-    res->hashmask = hashsize - 1; // Creates a bitmask
 
     res->sorted_ptr = 0;
     res->sorted_list = 0;
@@ -264,7 +256,7 @@ struct record_cluster *reclist_insert( struct reclist *l,
     assert(merge_key);
     assert(total);
 
-    bucket = jenkins_hash((unsigned char*) merge_key) & l->hashmask;
+    bucket = jenkins_hash((unsigned char*) merge_key) % l->hash_size;
 
     for (p = &l->hashtable[bucket]; *p; p = &(*p)->hnext)
     {

diff --git a/src/reclists.h b/src/reclists.h
index d49acff..a356431 100644 (file)
--- a/src/reclists.h
+++ b/src/reclists.h
@@ -34,7 +34,7 @@ struct reclist_sortparms
     struct reclist_sortparms *next;
 };
 
-struct reclist *reclist_create(NMEM, int numrecs);
+struct reclist *reclist_create(NMEM);
 struct record_cluster *reclist_insert( struct reclist *tl,
                                        struct conf_service *service,
                                        struct record  *record,

diff --git a/src/relevance.c b/src/relevance.c
index ffcda2c..f2a58e7 100644 (file)
--- a/src/relevance.c
+++ b/src/relevance.c
@@ -241,7 +241,7 @@ void relevance_countwords(struct relevance *r, struct record_cluster *cluster,
 
 
 struct relevance *relevance_create(pp2_charset_t pct,
-                                   NMEM nmem, const char **terms, int numrecs)
+                                   NMEM nmem, const char **terms)
 {
     struct relevance *res = nmem_malloc(nmem, sizeof(struct relevance));
     const char **p;

diff --git a/src/relevance.h b/src/relevance.h
index 4fba31e..e271d8c 100644 (file)
--- a/src/relevance.h
+++ b/src/relevance.h
@@ -28,7 +28,7 @@ struct record_cluster;
 struct reclist;
 
 struct relevance *relevance_create(pp2_charset_t pct,
-                                   NMEM nmem, const char **terms, int numrecs);
+                                   NMEM nmem, const char **terms);
 void relevance_newrec(struct relevance *r, struct record_cluster *cluster);
 void relevance_countwords(struct relevance *r, struct record_cluster *cluster,
         const char *words, int multiplier);

diff --git a/src/termlists.c b/src/termlists.c
index 892e6d9..3646d67 100644 (file)
--- a/src/termlists.c
+++ b/src/termlists.c
@@ -42,8 +42,7 @@ struct termlist_bucket
 struct termlist
 {
     struct termlist_bucket **hashtable;
-    int hashtable_size;
-    int hashmask;
+    unsigned hash_size;
 
     struct termlist_score **highscore;
     int highscore_size;
@@ -53,24 +52,14 @@ struct termlist
     NMEM nmem;
 };
 
-struct termlist *termlist_create(NMEM nmem, int numterms, int highscore_size)
+struct termlist *termlist_create(NMEM nmem, int highscore_size)
 {
-    int hashsize = 1;
-    int halfnumterms;
-    struct termlist *res;
-
-    // Calculate a hash size smallest power of 2 larger than 50% of expected numterms
-    halfnumterms = numterms >> 1;
-    if (halfnumterms < 0)
-        halfnumterms = 1;
-    while (hashsize < halfnumterms)
-        hashsize <<= 1;
-    res = nmem_malloc(nmem, sizeof(struct termlist));
-    res->hashtable = nmem_malloc(nmem, hashsize * sizeof(struct termlist_bucket*));
-    memset(res->hashtable, 0, hashsize * sizeof(struct termlist_bucket*));
-    res->hashtable_size = hashsize;
+    struct termlist *res = nmem_malloc(nmem, sizeof(struct termlist));
+    res->hash_size = 399;
+    res->hashtable =
+        nmem_malloc(nmem, res->hash_size * sizeof(struct termlist_bucket*));
+    memset(res->hashtable, 0, res->hash_size * sizeof(struct termlist_bucket*));
     res->nmem = nmem;
-    res->hashmask = hashsize - 1; // Creates a bitmask
 
     res->highscore = nmem_malloc(nmem, highscore_size * sizeof(struct termlist_score *));
     res->highscore_size = highscore_size;
@@ -131,7 +120,7 @@ void termlist_insert(struct termlist *tl, const char *term)
     for (cp = buf + strlen(buf); cp != buf && strchr(",. -", cp[-1]); cp--)
         cp[-1] = '\0';
     
-    bucket = jenkins_hash((unsigned char *)buf) & tl->hashmask;
+    bucket = jenkins_hash((unsigned char *)buf) % tl->hash_size;
     for (p = &tl->hashtable[bucket]; *p; p = &(*p)->next)
     {
         if (!strcmp(buf, (*p)->term.term))

diff --git a/src/termlists.h b/src/termlists.h
index 1bd27b1..d3a5b1b 100644 (file)
--- a/src/termlists.h
+++ b/src/termlists.h
@@ -30,7 +30,7 @@ struct termlist_score
 
 struct termlist;
 
-struct termlist *termlist_create(NMEM nmem, int numterms, int highscore_size);
+struct termlist *termlist_create(NMEM nmem, int highscore_size);
 void termlist_insert(struct termlist *tl, const char *term);
 struct termlist_score **termlist_highscore(struct termlist *tl, int *len);