On Wed, May 28, 2003 at 11:33:14PM +0100, Dave Hooper wrote: > This was discussed before, you can check the devl archives. The > conclusion then seemed to be "it isn't worth the hassle" which I thought > was strange-- the original complaint was by someone (I forget who) whose > external IP changes daily, if not more often. > I tried to point out that it's more hassle for the end user to update his > config every time his IP address changes, especially when he may not even > know precisely when it has changed (other than by noticing that his > freenet traffic has suddenly dropped to zero). > I don't believe any subsequent conclusion was reached. > > You're right, the current situation is a bit mad. I'd like to see replies > from the other nodes saying "here's the data I'm sending you or a > connection response or whatever but hey, this is what your IP address > looks like to me". Wouldn't that be just super? It means external IP > address detection would even work when you're behind a NAT (though of > course you'd still need to config your NAT to route the port, but that's a > one-off change, and finding out your external IP address for some brands > of 'plug-and-play' DSL/cable routers is way more difficult than asking > said router to forward a single TCPIP port) > > Could we please open this up to discussion again? Why can't the nodes > reflect the 'observed' return IP address in at least some of their > responses? (Or, of course, do they already? Or supposed to?)
Firstly, this is probably a good idea, but implementing it cleanly would require it to be implemented at the session layer, meaning implementing sessionv2, which has many other things we need to deal with. It will be gotten around to eventually. Secondly it is not clear that the numbers of nodes involved are sufficient to justify the effort. Sessionv2 will be implemented eventually for performance reasons and we will include the infrastructure necessary to support remote IP detection. Remote IP detection itself will be implemented at some point after that, if it seems to be justified. > > d -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ GPG key lost in last few weeks, new key on keyservers ICTHUS - Nothing is impossible in him whom we trust
pgp00000.pgp
Description: PGP signature
