I noticed that with the standard vpnc scripts, connecting to a VPN (VPN B) that is only available via another VPN (VPN A) causes a loss of connectivity to VPN B. This is because the route to the IP address for VPN B is hard fixed to the system default gw, which is not (or at least not always) routed through VPN A. A simple fix seems to be to simply look up the route (route get on OSX, ip route get on Linux I think) and add the host to that route instead. Is there a reason we don't do this? I can submit a patch if this sounds sane. Jeremy
