I made the node resolve the it's own address from the socket so that the setting would work smoothly on mutlihomed machines (most) without having to work with subnet masks and alike internally. It also resolves it's own address on the incoming connection, and then checks that it is equal to the outgoing, reseting all source pointers to the old network to itself if it isn't. This allows Nodes to work as proxies on ip masq routers and alike (and should even allow (if you write the transport code) for nodes that have one foot on the Internet and one foot on another network effortlessly).
If somebody knows a better way of doing this I am all ears - I could of course allow a simple configuration setting that overrides the resolution if set. How desirable is that? On Wed, Aug 16, 2000 at 10:50:43AM +1200, Stephen Blackheath wrote: > All, > > I'd like to be able to do exactly this too, because my firewall runs on a > shitbox that hasn't enough memory or hard disk to run a Freenet node. I > don't want put my big box directly on the Internet for semi-obvious > reasons that would be too long and boring to explain here. > > I can't see how (if you're using port forwards) the Freenet node can > definitively know what IP address and port it should advertise itself as > unless you tell it. Surely the same problem exists if you've got a > dynamic IP address - you'll want to advertise yourself as a name, not an > IP address. > > ? > > > Steve -- \oskar _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
