On Thu, Jun 3, 2010 at 8:36 PM, Emmanuel Lecharny <elecharny@gmail.com> wrote:
On 6/3/10 7:08 PM, Felix Knecht wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

The constructor accepts a cachesize which is never used. Apart of this I
really can't see, where the caching happens.

Any ideas?
 

Ok, in real world, this cache is just useless. I mean, it's a good idea to think about having a cache, but it has to be implemented correctly, which is not the case here.

The discussions we had a while back about the caches (and there is a JIRA about it) demonstrated that we need a global cache, based on some global system like JCS (apache commons-jcs). It should be linked to the DirectryService, and mus not be a singleton.

Right now, the cache we have are associated with instances of the interceptors, so they are not shared. They are just valid during an user session, AFAICT. Not efficient... (I would ask you to take this with cautious, I may be wrong).



It really would be nice if we can get this cache mess sorted out and have a solid entry cache in place before releasing 2.0. It's a critical part of the server and should not be difficult to do with ehcache or something.  If we can get this JDBM serialization hook working to check the cache first that's another plus.

Thoughts? 

--
Alex Karasulu
My Blog :: http://www.jroller.com/akarasulu/
Apache Directory Server :: http://directory.apache.org
Apache MINA :: http://mina.apache.org
To set up a meeting with me: http://tungle.me/AlexKarasulu