[00:04:10] *** bryan986 has joined #bittorrent
[00:13:40] *** The_8472 has quit IRC
[00:15:52] *** kwinz2 has quit IRC
[00:27:27] *** ajaya has quit IRC
[00:37:25] *** kwinz2 has joined #bittorrent
[00:49:51] <klapaucjusz> The_8472: delete sounds like a good idea, but I'm not sure it counts as "minor".
[01:34:33] *** kwinz2 has quit IRC
[01:37:29] *** kwinz2 has joined #bittorrent
[01:44:50] *** andar2 has quit IRC
[02:01:08] *** ukdkbr has quit IRC
[02:01:56] *** waldorf_ has joined #bittorrent
[02:03:09] *** waldorf_ has quit IRC
[02:05:49] *** kwinz2 has quit IRC
[02:19:13] *** kwinz2 has joined #bittorrent
[02:32:18] *** Waldorf has quit IRC
[02:33:25] *** Waldorf has joined #bittorrent
[02:33:39] *** Waldorf has quit IRC
[02:34:53] *** Waldorf has joined #bittorrent
[02:35:40] *** Waldorf has quit IRC
[02:39:22] *** Waldorf has joined #bittorrent
[02:40:11] *** Waldorf has quit IRC
[02:58:03] *** klapaucjusz has quit IRC
[03:03:19] *** kwinz2_ has joined #bittorrent
[03:03:40] *** kwinz2 has quit IRC
[03:36:02] *** echelog has joined #bittorrent
[03:57:41] *** goussx has quit IRC
[04:04:22] *** cyb2063_ has joined #bittorrent
[04:17:08] *** cyb2063 has quit IRC
[04:24:47] *** init0 has joined #bittorrent
[04:37:59] *** kwinz2_ has quit IRC
[04:38:52] *** init0_ has quit IRC
[04:39:09] *** kwinz2_ has joined #bittorrent
[04:51:10] *** ukdkbr has joined #bittorrent
[04:57:26] *** kwinz2_ has quit IRC
[05:08:50] *** ryanprior has joined #bittorrent
[05:14:26] *** kwinz2_ has joined #bittorrent
[05:14:35] *** bryan986 has quit IRC
[05:46:19] *** kwinz2_ has quit IRC
[05:55:15] *** kwinz2_ has joined #bittorrent
[06:13:11] *** ukdkbr has quit IRC
[06:35:58] *** Switeck has quit IRC
[06:40:47] *** KyleK_ has joined #bittorrent
[06:49:19] *** MassaRoddel has quit IRC
[06:55:37] *** KyleK_ has quit IRC
[07:05:59] *** GTHK has quit IRC
[07:18:21] *** MassaRoddel has joined #bittorrent
[08:42:37] *** cyb2063_ is now known as cyb2063
[09:02:01] *** goussx has joined #bittorrent
[09:29:07] *** TheSHAD0W has quit IRC
[09:39:43] *** TheSHAD0W has joined #bittorrent
[10:00:44] *** Nakor has joined #bittorrent
[10:15:17] *** _rafi_ has joined #bittorrent
[10:15:49] *** ryanprior has quit IRC
[10:17:16] <_rafi_> alus: nice test-build, fixing the rate_control :)
[10:48:52] *** void^ has quit IRC
[10:49:36] <_rafi_> ...though.. I'm not sure how to observe if it works at all..
[10:56:47] <alus> well, were you having problems before that disabling tcp_rate_control would fix?
[11:14:30] <_rafi_> yes...
[11:15:47] <alus> well, does the new build work, even with tcp_rate_control enabled?
[11:16:52] <_rafi_> alus: ...yes, and I'm sure you've fixed those :) I was just curious if I should be  observing any change in performence with it being true or false
[11:18:38] <_rafi_> http://forum.utorrent.com/viewtopic.php?pid=444629#p444629
[11:20:47] <alus> in some cases, yes you would notice an improvement. but if you don't notice a problem, then don't worry about it
[11:21:33] <_rafi_> k :) now that it's meets my expectations... ;)
[11:22:50] <_rafi_> strange build # though, alus... (regressed from the one you've sent me)
[11:23:40] <alus> it just depends
[11:38:12] *** Andrius has joined #bittorrent
[11:45:23] *** medecau has joined #bittorrent
[11:51:19] *** void^ has joined #bittorrent
[12:05:23] *** bittwist has joined #BitTorrent
[12:17:38] *** Gottaname2 has joined #bittorrent
[12:19:48] *** Gottaname has quit IRC
[12:23:18] *** Gottaname2 has quit IRC
[12:25:08] *** bt42 has quit IRC
[12:41:06] *** Gottaname2 has joined #bittorrent
[12:54:10] *** _rafi2_ has joined #bittorrent
[12:58:39] *** rrr_ has quit IRC
[13:02:34] *** Gottaname2 has quit IRC
[13:10:17] *** The_8472 has joined #bittorrent
[13:10:44] *** _rafi_ has quit IRC
[14:12:59] *** GTHK has joined #bittorrent
[15:36:44] *** medecau has quit IRC
[15:40:43] *** bt42 has joined #BitTorrent
[15:48:57] *** _rafi2_ is now known as _rafi_
[15:59:33] *** bittwist has quit IRC
[16:12:39] *** Gottaname has joined #bittorrent
[16:51:59] *** bt42 has quit IRC
[16:55:41] *** void^ has quit IRC
[17:08:45] *** medecau has joined #bittorrent
[17:11:14] *** stalled has quit IRC
[17:11:41] *** medecau has left #bittorrent
[17:13:45] *** stalled has joined #bittorrent
[17:28:31] <swolchok> The_8472: I don't get it, what is the point of DHT scrape vs. normal use of DHT?
[17:28:51] <The_8472> normal use doesn't give you seed/peer counts
[17:29:03] <The_8472> you only see a subset
[17:29:46] <The_8472> and the purpose is having scrapes for queued torrents so you can sort them in your seeding queue
[17:31:01] *** medecau has joined #bittorrent
[17:31:48] <swolchok> The_8472: how do you only see a subset? shouldn't everyone be listed?
[17:32:06] *** Waldorf has joined #bittorrent
[17:32:16] <swolchok> I know azdht just stashes everyone under SHA-1(infohash); how is mainlne different?
[17:32:23] <swolchok> is it a replication consistency thing?
[17:34:35] <The_8472> nono, when you perform a GET_PEERS query you only get a subset of whatever the node has stored
[17:34:55] <The_8472> not to mention that noise seems to be damn high, so not everyone is putting values into the same nodes
[17:35:03] <The_8472> they're sortof scattered around the target
[17:48:37] <The_8472> swolchok: http://pastebin.com/m41caf8fc
[17:55:11] <swolchok> that's, uh, bizarre?
[17:55:40] <swolchok> routing inconsistency?
[17:55:44] <swolchok> no replication?
[17:55:51] <swolchok> probably both of the above
[18:02:29] <TheSHAD0W> I'm starting to like Kademlia.
[18:02:40] <TheSHAD0W> If it can still work with so many peers screwing up the implementation...
[18:02:47] <TheSHAD0W> It's a pretty robust design.  :-P
[18:06:58] <The_8472> yes, no replication and crappy routing by peers
[18:13:32] *** void^ has joined #bittorrent
[18:19:43] *** KyleK_ has joined #bittorrent
[18:26:13] *** goussx has quit IRC
[18:33:41] *** Waldorf has quit IRC
[18:36:53] *** bittwist has joined #BitTorrent
[18:46:22] *** Miller` has quit IRC
[18:46:39] *** Miller` has joined #bittorrent
[18:47:26] <burris> mainline dht has replication
[18:48:09] <The_8472> no it doesn't
[18:48:57] <burris> every announce is inserted, what, three times?
[18:49:12] <The_8472> that's not replication.
[18:49:18] <burris> :-P
[18:49:21] <The_8472> and it should be K times, i.e. 8
[18:50:12] <The_8472> replication is the K nodes sharing values among each other, not the inserting nodes putting them redundantly
[18:50:42] <burris> whatever you say, boss
[18:52:20] <The_8472> lol
[18:52:34] <The_8472> i'm just saying how things are defined
[18:59:32] *** goussx has joined #bittorrent
[19:00:51] *** andar2 has joined #bittorrent
[19:05:33] *** rrr has joined #bittorrent
[19:11:45] *** r2wj has joined #bittorrent
[19:36:41] *** bittwist has quit IRC
[19:36:49] *** bittwist has joined #BitTorrent
[19:48:01] *** ProperNoun has joined #bittorrent
[19:51:15] *** alienvenom has quit IRC
[19:51:57] *** alienvenom has joined #bittorrent
[20:41:21] *** bt42 has joined #BitTorrent
[21:01:36] *** bittwist has quit IRC
[21:08:10] *** Switeck has joined #bittorrent
[21:11:54] *** burris has quit IRC
[21:12:18] *** burris has joined #bittorrent
[21:37:17] <The_8472> swolchok, good point though about replication... though the question is what would be easier... bloom filters or replication.
[21:41:26] <swolchok> The_8472: bloom filters? are you referencing a doc somewhere?
[21:41:47] <swolchok> it's not immediately obvious what probabilistic set representations have to do with replication :)
[21:42:44] <swolchok> also, is it known which clients do and do not have good DHT impls (e.g., ones that actually replicate to the neighborhood set)?
[21:43:56] <swolchok> argh. just took another look at BEP 5 and it appears to sacrifice the possibility of replication (presumably because of poisoning concerns)
[21:44:50] <swolchok> I'm baffled as to how mainline manages to work :)
[21:46:26] *** cyb2063 has quit IRC
[21:48:42] <swolchok> I guess peers will find each other through PEX or something
[21:49:41] <Switeck> let's hope so...
[21:49:58] <Switeck> I'm seeing bogus ip addresses from udp trackers
[21:50:08] <Switeck> I've also gotten them via PEX
[21:50:28] *** medecau has left #bittorrent
[21:50:32] <Switeck> which makes no sense to me, since PEX is supposed to only be "tested and good" ips. :(
[21:50:36] <swolchok> oh, and sacrificing replication like that doesn't really prevent poisoning.
[21:51:03] <swolchok> the ISP can do it, and you can also poison by actually starting up some nodes and lying
[21:51:44] <Switeck> there are many buggy clients out there as well that no doubt spew bad ips via DHT, PEX, and Local Peer Discovery.
[21:52:06] <Switeck> (they're not maliciously lying, just stupid)
[21:52:19] <swolchok> right
[21:52:28] <swolchok> I'm just trying to work out why on earth you would build a DHT that can't replicate
[21:53:27] <swolchok> I take it mainline is still good enough? I don't actually use bittorrent very much myself.
[21:54:44] <Switeck> replicate connection routes?
[21:57:19] *** yngress has joined #bittorrent
[21:57:51] <yngress> Does anyone know of a torrent video streaming application available for Linux?
[21:58:25] <Switeck> I hope not :P
[21:58:39] <yngress> Care to explain? ;P
[21:58:51] <Switeck> simple: hostile behavior is unwanted.
[21:59:02] <Switeck> especially automated hostile behavior
[21:59:13] <The_8472> swolchok, for scrapes, not for replication.
[21:59:52] <Switeck> video streaming of torrents implies breaking rarest-piece order that torrent swarms use to stay functional.
[22:00:22] <yngress> I see.
[22:00:59] <Switeck> in the case of low seeds...the few seeds may be "hit" by people downloading the same common first few pieces over and over again, get a high ratio, and leave.
[22:01:31] <Switeck> ...And that was happening before torrent streaming was "cool". >:(
[22:02:02] <yngress> :/
[22:02:29] <Switeck> better to just get faster download speeds instead and sort it out when done.
[22:03:01] <Switeck> or if it's really huge but you're not sure it's what you're looking for, do a spot check using a partial file video viewer program...
[22:05:26] <Switeck> FYI: streaming is slower than regular downloading
[22:06:22] <yngress> Hmm I actually hoped to have distributed video streaming... but thanks for the help :)
[22:11:33] <Switeck> if you're on a 100mbit/sec line, you effectively already do XD
[22:12:00] <Switeck> given enough sources, the whole torrent gets downloaded at speeds the rest of us can only dream about
[22:13:56] <yngress> Hmmmm
[22:14:09] <yngress> :D
[22:15:11] <yngress> Oh, why's the forum down?
[22:15:31] <yngress> btw FluxBB ftw :P
[22:17:12] <_rafi_> Switeck, have you tried the latest test-build ? Seems like almost all the open issues were handled in it ... :)
[22:18:01] <Switeck> I haven't tested the special test builds
[22:18:09] <Switeck> haven't rebooted in a week
[22:18:51] <_rafi_> good for you... :)
[22:19:59] <_rafi_> anyways, Firron posted it in the forum
[22:20:07] <Switeck> I downloaded it XD
[22:20:13] <Switeck> just haven't tested it
[22:42:10] <_rafi_> still, we're trying to get an option to include the overhead inside/under  the speed limit. want to vote for it ?
[22:47:50] <Switeck> I vote the overheads should not be counted against the limiter
[22:49:06] <Switeck> but an option to do so is ok, at least if off by default.
[23:04:34] *** KyleK_ has quit IRC
[23:07:50] <_rafi_> k... so I count 5 votes... :)
[23:08:24] <_rafi_> zzz...
[23:08:30] <_rafi_> http://forum.utorrent.com/viewtopic.php?pid=444719#p444719
[23:08:36] *** _rafi_ has quit IRC
[23:08:37] <Switeck> Were such an option on by default, there would be many uploading nothing even while demand is high.
[23:11:52] *** yngress has quit IRC
[23:14:17] *** medecau has joined #bittorrent
[23:21:43] <swolchok> is there a good argument against dropping token verification in get_peers in the DHT?
[23:58:35] *** RandomTime has joined #bittorrent
[23:59:49] *** RandomTime has left #bittorrent





top