karmazilla (Christian Vest Hansen)

put contention handler test cases in their own package

implement getOriginalPoolSize in the JMX interface so the ActiveResizingContentionHandler mentioned in #10 can use it as a minimum pool size mark

exclude the transitive hamcrest-core test dependency because we already have hamcrest-all

Fsm no longer unconditionally marks a Connector as AVAILABLE if it fails to getConnection, but will now respect OUTDATED and SHUTDOWN states and leave them untouched. Closes #13

begun work on the javadoc for the TimingHook, plus a minor refactoring

completed the javadoc for the JulLogHook

complete the javadoc for the Hook interface

refactoring to remove the ConnectionPoolDataSource parameter from the Hook interface and its runHook method - I doubt it will be missed

formatting in the Settings javadoc

bit of javadoc for the Hook interface

add note on a possible data race bug

a bit of javadoc to the hooks.EventType

more spit, wax and polish for the Settings javadocs

javadoc for @return on getters in Settings

javadoc and test for CountUpHook, plus some wax on CountDownHook

added javadoc and tests to the CountDownHook

tiny bit of hooks javadoc

renamed FsmMixin to Fsm

let the Settings class use good old synchronization instead of that hairy lock-free scheme it had going

completed the javadocs for the Settings class

add the beginnings of a python script that will eventually use the PyCha charting library to generate really nice charts from the output of benchmark runs

wax

wax

add a todo on a potential race condition

add a why-note on recent change

organize imports

more wax - no idea if this is faster or not

wax

reduced the number of arguments to Connector.getConnection

now makes sure that pre-connect hooks are only run once. closes #10