2.5.0.1

• Bumps time upper bound to 1.14
• Adds roundtrip test for parsing 0-padded years w/ PostgreSQL per #61

2.5.0.0

Forces ForeignPtr finalization at the time disconnect is called, as opposed to relying on when the Haskell GC frees the ForeignPtr. Usually this isn't a problem, but in the case of long lived connection, it can lead to situations where the ForeignPtr is promoted and major GCs never reach it in scenarios with large working sets. This can lead to scenarios when resource-pool has removed a connection, but the GC hasn't run the finalizer.