Hang on server failure

[trsmith2]

I'm seeing a timing dependent hang when the gremlin server goes away for some reason. It looks like the response handler sits on a channel waiting for a response with no timeout. If we get to this blocking call as the server goes down, there is never a response and so this call never unblocks.

https://github.com/qasaur/gremgo/blob/master/response.go#L73

Maybe I missed some setting that would avoid this? Otherwise, my temporary fix is simply to introduce a select/case style timeout, which I've hard coded to a second but I guess could come from some config option.

If you agree I'd be happy to do a PR.