X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Ftermlists.c;h=b40e5b026cf413e0edab895be20b731e282bebbb;hb=8351b07cbbd0deab140944ce7ededa3d52ddb41c;hp=3af94df30abe748bdba62bee3726d6d53e1b1dfd;hpb=49088b6680bed3c069be10ff0cb7cc16f3733729;p=pazpar2-moved-to-github.git

diff --git a/src/termlists.c b/src/termlists.c
index 3af94df..b40e5b0 100644
--- a/src/termlists.c
+++ b/src/termlists.c
@@ -1,5 +1,5 @@
 /* This file is part of Pazpar2.
-   Copyright (C) 2006-2008 Index Data
+   Copyright (C) 2006-2010 Index Data
 
 Pazpar2 is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free
@@ -26,6 +26,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/yaz-util.h>
 
 #include "termlists.h"
+#include "jenkins_hash.h"
 
 // Discussion:
 // As terms are found in incoming records, they are added to (or updated in) a
@@ -41,8 +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;
@@ -52,42 +52,14 @@ struct termlist
     NMEM nmem;
 };
 
-
-// Jenkins one-at-a-time hash (from wikipedia)
-static unsigned int hash(const unsigned char *key)
-{
-    unsigned int hash = 0;
-
-    while (*key)
-    {
-        hash += *(key++);
-        hash += (hash << 10);
-        hash ^= (hash >> 6);
-    }
-    hash += (hash << 3);
-    hash ^= (hash >> 11);
-    hash += (hash << 15);
-    return hash;
-}
-
-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;
@@ -148,7 +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 = 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))
@@ -186,7 +158,9 @@ struct termlist_score **termlist_highscore(struct termlist *tl, int *len)
 /*
  * Local variables:
  * c-basic-offset: 4
+ * c-file-style: "Stroustrup"
  * indent-tabs-mode: nil
  * End:
  * vim: shiftwidth=4 tabstop=8 expandtab
  */
+