Thank you, I tried to make the code as readable as I could. And yes, since the *nix model for (sub)processes is cleaner than Cygwin-on-Windows it's kind of surprising that that gap is what it is.