14 commits

Author	SHA1	Message	Date
Damian Janowski	5053b327f1	Checking in pushes objects to the head of the queue. ... This means that, when not saturated, objects in the pool will be reused more often. This behavior allows you to create a pool of objects which in turn have some kind of time outs themselves. For instance, objects which keep an underlying TCP connection open until they're idle for some time. An example of this is `Net::HTTP::Persistent`. See also issue #20.	2012-12-18 14:05:13 -03:00
Damian Janowski	f0cb4ea11f	Restore Wrapper inheriting from BasicObject. Closes #16.	2012-12-18 13:50:12 -03:00
Damian Janowski	29cd6d7b56	More robust Wrapper#respond_to?. ... See #25.	2012-12-18 13:41:59 -03:00
Petteri Räty	387013797b	Fix nested checkouts ... If you did nested checkouts the inner use would check the connection back to the pool and then the outer checkout could be using a connection at the same time with something else. See issue #14.	2012-06-18 16:15:50 +03:00
Mike Perham	ad08c55f8b	Wrapper should provide 'with' just like a ConnectionPool instance, minor cleanup.	2012-03-14 08:36:40 -07:00
Damian Janowski	8ec4b4e7c5	Do pass-through via a specialized class.	2012-03-14 10:07:43 -03:00
Pete Higgins	c176da4cc7	Smaller timeout.	2012-02-08 09:45:26 -08:00
Pete Higgins	5977adf0f5	Fix method_missing with block.	2012-02-08 09:44:18 -08:00
Ryan LeCompte	05e6b035fb	fix intermittent failure for issue #2	2012-02-05 19:54:48 -08:00
Mike Perham	b36d8cc056	Suggestions from @brixen	2011-09-25 19:23:35 -07:00
Mike Perham	6f39ecd68e	More correct usage of condition variables, release 0.0.3	2011-09-12 21:01:38 -07:00
Mike Perham	e91483ac6f	Exception upon checkout now causes 100% less corruption!	2011-09-09 14:12:23 -07:00
Mike Perham	5ddef034cf	More testing, threading fixes	2011-05-14 19:42:07 -07:00
Mike Perham	27362471c6	Add project info, tests	2011-05-14 15:36:17 -07:00