markbenvenuto and keithbostic can you review this change? Mark, please compile to make sure I didn't mess anything up. It seems that sleep with a 0 value has special meaning that does not seem to be what we intend with a small timeout value.
https://msdn.microsoft.com/en-us/library/windows/desktop/ms686298%28v=vs.85%29.aspx

Or if I'm off-base we can just toss this branch.