It seems to me that the situation where an incoming requests cannot find a single routable node on the first pass should be so rare that it hardly matters.
The change to make maxRoutingSteps count the number of nodes passed rather than keys was definitely right - giving up after just one or two failed connects is not a good idea at any level. I think adding a second "desperate" pass (which is how I understood Matthew's description) is probably an unnecessary complication though. I think you are worrying a little more than necessary though, gj. The situation with the network now if not what it was last spring - nodes are not being "hammered" with connections they cannot accept. Nodes ever being forced to reject connections is more or less a thing of the past from all my observations. On Sun, Oct 27, 2002 at 06:52:42PM -0500, Gianni Johansson wrote: > Mathew: > Did you run this by Oskar? > > Why is your approach better than just increasing maxRoutingSteps? > > If you are concerned about local users getting angry because their requests > RNF, why not just have a separate, higher maxRoutingSteps for local requests? > That way the node could work harder just for the local user without utterly > hammering overloaded noderefs for every incoming FNP request. > > -- gj > > > _______________________________________________ > devl mailing list > devl at freenetproject.org > http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl -- Oskar Sandberg oskar at freenetproject.org _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
