Changes
=======
0.00012 - May 28 2015
- Fixed an integer underflow by completely redoing the binary
search -- this code was stolen from whatever ketama code
I stole the algorithm from, and so it used to exist for the
longest time... and so, in order to not break already hashed
data that probably exists in the wild, the old binary search
is left in tact.
You can change the search mechanism by passing "use_hashfunc"
parameter to the constructor
0.00011 - Feb 18 2015
- The library now properly handles hashing binary data. Previously
it would expect NULL-terminated strings.
- Fixed a bug with the code segfaulting on some keys, e.g. the
pack "H*" versions of 161c6d14dae73a874ac0aa0017fb8340 and
37292b669dd8f7c952cf79ca0dc6c5d7
This was discovered "in the wild" in a system that deals with
billions of keys, htese keys just so happened to trigger a
segfault in the library due to how the binary search works.
0.00010 - Mar 1 2013
- No code change. Just a re-release of 0.00009
0.00009 - Oct 20 2012
- No code change
- Upgrade to Module::Install 1.06
(see http://weblog.bulknews.net/post/33907905561/do-not-ship-modules-with-module-install-1-04)
0.00008 - May 15 2012
* Add hash_with_hashnum(), label_from_hashnum(). This allows clients to
cache the hashed values -- thus avoiding the need to recompute the
MD5 hash. However, you should make sure you know what you're doing
before using these (Steffen Mueller)
(https://github.com/lestrrat/Algorithm-ConsistentHash-Ketama/pull/1)
0.00007 - Jun 3 2010
* How about actually using ppport.h?
0.00006 - May 18 2010
* Fix cloning problems under threads
0.00005 - May 09 2010
* Remove superfulous code from constructor
0.00004 - May 09 2010
* Fix memory leak
0.00003 - May 08 2010
* Add bunch of thread safety changes
0.00002 - May 03 2010
* Upload previous changes as 0.00002
* Add author tests
* Hide some XS methods
* Add more for Bucket.pm
0.00001_03 - May 03 2010
* Allocate temporary variable continuum on heap instead of on stack
* Attempt to include math.h explicitly
0.00001_02 - May 03 2010
* In search of a mysterious failure from cpantesters
* Tweak tests
* Updated Test::More dep
0.00001_01 - May 02 2010
* Fix doc typos
* Apply some XS tweaks
* Add test_requires clause
* Add missing Changes file
0.00001 - Apr 28 2010
* Initial release