From: Ripduman Sohan <[email protected]>

It's cheaper and simpler to do a comparison based on a match with the
(unique) memory address of the h_next item as opposed to doing a
comparison of the hash value and string comparison of the key

Signed-off-by: Ripduman Sohan <[email protected]>
---
 engines/default_engine/assoc.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/engines/default_engine/assoc.c b/engines/default_engine/assoc.c
index 74ac12d..3d30647 100644
--- a/engines/default_engine/assoc.c
+++ b/engines/default_engine/assoc.c
@@ -117,9 +117,8 @@ static inline hash_item** _hashitem_before(struct 
default_engine *engine,
                                     hash_item *it) {
     hash_item **pos = &bucket->head;
 
-    while (*pos && ((it->hash != (*pos)->hash) || memcmp(item_get_key(it), 
item_get_key(*pos), it->nkey))) {
+    while (*pos && *pos != it)
         pos = &(*pos)->h_next;
-    }
     return pos;
 }
 
-- 
1.7.1

Reply via email to