On Tuesday 03 April 2001 13:31, you wrote: > > On Tue, Apr 03, 2001 at 06:47:22PM +0200, Stefan Reich wrote: > > >How does you verification work BTW? Do you download the whole key or > > >abort the second something comes through? > > > > Currently I'm doing it the simple & stupid way - URLConnection to fproxy. > > I will change this to use the official library. > > > > fxproy always retrieves the whole file, even though I close the > > connection after 10.000 bytes. Does anyone know if the library behaves > > differently? This would speed up key verification a lot. > > You could with libfreenet Use the freenet_request_buffer function to > request the keys and set the buffer size to 0 bytes. It'll abort with > the FNS_BUFFER_TOO_SMALL error code the second it finds that the size > of the file is too big. > > However last time GJ tried the same thing the routing got screwed up > for a few days every time he verified a big pile of keys... He aborted > early too... Maybe aborting early screws things up?
I think the problem is that the request is only "aborted" from the client's perspective. It still continues on in the upstream nodes. At least that's how I understand it works. I'm sure someone will let me know if I've got this wrong. Routing got screwed up? That was my paranoia but I'm not sure that it ever actually happened. The mechanism that I imaged is that upstream nodes would get so saturated with requests (that the client had already dropped) that they would start failing to serve other legitimate inbound requests and that the nodes making the requests would drop the saturated nodes from their routing as a result. Also remember that this was back when the network was very weak. --gj P.S. I kind of like your version though. "gj, the guy who's client app took down freenet... " ;-) -- Web page inside Freenet: freenet:MSK at SSK@enI8YFo3gj8UVh-Au0HpKMftf6QQAgE/homepage// _______________________________________________ Devl mailing list Devl at freenetproject.org http://lists.freenetproject.org/mailman/listinfo/devl
