@parsonsmatt
I think implementing timeouts this way is not very good idea. There are many libraries use pool, so there are many cases you create pool used by external library. But a library author don't (and, maybe, shouldn't?) think whether to use with timeout version or basic one. So I think you should be able to inject timeout inside of pool.

Update: Because of reasons I've written above my issue #10 is not solved by this PR (because I don't have access to code which takes resource.) And also your PR can be implemented in beyond of the library. It's just a sugar for existing interface. So in my PR #11 I added new feature allows to inject timeout, which can't be implemented beyond.