Matthew Toseland schrieb: > > On Wed, Jan 29, 2003 at 09:18:50PM +0100, Michael Schierl wrote: > > hi matthew, > > > > Matthew Toseland schrieb: > > > > Another idea. Or switching it on automatically, if you have as many > > > > failures that you cannot complete the file and no successful block. > > > > > > I don't get it. > > > > set paranoid checking to true. try to dl the splitfile. > > > > if it fails because there were too few blocks decodable, and if there > > was no single block decodable, retry with paranoia off. > > With any splitfile inserted with FIW, all the blocks will be regarded as > corrupted.
Yep, right; and the probability that you see that somewhere in the wild with your code is nearly zero. So I thought it's a good check for heuristics - if checking whether a block has metadata is too difficult. > > > Mostly, this would be illegal. > > > > Why? Does GJ live in a country where free speech is not permitted? > Sorry, I misread. Most splitfiles inserted with FIW are likely to be > illegal. they are likely, but are all of them? Of course that depends on the lawyers as well. www.xenu.net is seen as illegal by some of them as well - they don't like one picture, and when it's removed they don't like another one... My point is, transporting files w/ freenet is not illegal, storing files is not either, healing files is not either, so if you set up a script that converts old splitfiles into new ones and you don't look at their contents, where is the difference? > Hey, we almost completely ignore Frost, mostly because they completely > ignore us (thankfully this is changing... there is a beta with FEC > support), despite the fact that Frost comprizes a LOT of freenet > traffic, probably more than the whole freenet web put together (the > spikes at 2k, 16k, 256k in your datastore are Frost, mostly). > Yeah. The incentive is purely that the splitfiles it inserts will work > significantly better through healing. Time will show that. I also think it is a good idea to heal files - but not ignore unhealable files completely. mihi _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
