Mike Glover <mpg4 at duluoz.net> wrote: > On Tue, 11 Apr 2000, Lee Daniel Crocker wrote: > > > Lee Daniel Crocker <lee at piclab.com> wrote: > > > > (1) "Depth" is gone now, right? I am assuming that replies just > > > > create their own HopsToLive value so that the replying node can't > > > > be identified from depth information. > > > > > > Oskar pointed out that you still need Depth on the initial messages > > > from clients so that you can set HopsToLive on reply, in case the > > > route back gets messed up for some reason. > > > > I don't see the message where he said that, and I still don't see the > > need. Why can't replies just set their own HTL as long as it is > > sufficient for any reasonable route back? > > Better question: if the route back gets messed up for some reason, how is > a "Depth" header going to help us find our way back? Either we have a > specific route back to the source of the request, or we need to broadcast > our reply to every node to be sure it gets where it belongs.
It's not supposed to help you find your way back, it just kills you off so you don't wander around forever. But yeah, it seems like you could just (arbitrarily) set HTL to e.g. 50 on replies, although that would place a hard limit on HTL for requests. That is, you could never specify -htl > 50, which isn't so hot. theo _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
