[00:12:53] *** Firon has joined #bittorrent
[00:19:39] *** Andrius has quit IRC
[00:29:48] *** Firon has quit IRC
[00:30:56] *** Firon has joined #bittorrent
[00:32:17] *** KyleK_ has quit IRC
[00:40:23] *** Miller` has quit IRC
[00:49:52] *** Miller` has joined #bittorrent
[00:52:29] *** Miller`` has joined #bittorrent
[00:53:43] *** Miller` has quit IRC
[00:53:47] *** Miller`` is now known as Miller`
[00:54:06] *** Miller`` has joined #bittorrent
[00:57:28] *** Miller`` has quit IRC
[01:06:02] <burris> good to know that if your infringement trial doesn't go as planned and you end up in gitmo you can order micky-d's and have it delivered to your cell
[01:08:05] *** The_8472 has quit IRC
[01:11:27] *** Miller` has quit IRC
[01:14:11] *** The_8472 has joined #bittorrent
[01:35:19] *** jimi_ has joined #bittorrent
[01:35:35] <jimi_> is there a settings calculator?
[01:36:35] <DeHackEd> calculator?
[01:36:39] <DeHackEd> what needs calculating?
[01:36:55] <jimi_> settings, what values i should have, i am maxxing out my upload
[01:37:11] <DeHackEd> whatever the maximum you have is, multiply by 0.85
[01:38:03] <jimi_> well that means my upload should be 34kbps, and i have it set to 15 and my internet crawls
[01:38:16] <The_8472> http://infinite-source.de/az/az-calc.html
[01:38:28] * The_8472 hasn't updated that in ages
[01:41:24] <kjetilho> jimi_: you're on a modem?
[01:41:32] <jimi_> kjetilho, CABLE
[01:41:43] <kjetilho> so you meant 34 KiB/s, then
[01:41:50] <kjetilho> not 34000 bits/second
[01:41:50] <jimi_> :D
[01:42:39] <burris> if you're using uT then you should probably just use automatic since I believe that stuff is now regulated by latency measured from uTP peers
[01:42:43] <jimi_> my dl is 9504Kb and upload is 350Kb
[01:43:03] <jimi_> I am using transmission
[01:43:50] <kjetilho> The_8472: you should add a field for download speed in your calculator
[01:43:51] <burris> transmission used to make my machine lock up so I ditched it
[01:44:32] <kjetilho> jimi_: for every 1500 bytes downloaded, you need a 40 byte ACK packet.  this needs to be subtracted from your upload speed capacity
[01:44:37] <burris> transmission also has crap estimated time to completion
[01:44:47] <jimi_> oh
[01:44:56] <kjetilho> jimi_: 40/1500*9504 = 253
[01:45:28] <kjetilho> so *if* you're downloading at full speed, you can only upload at 350-253 = 100 kbps
[01:46:08] <kjetilho> (obviously this calculation is simplified, in some cases the ACK can piggyback on other traffic so no extra packet is needed.)
[01:46:23] <burris> also mtu might not be 1500 bytes, right?
[01:46:38] <klapaucjusz> kjetilho: negligible in the case of BitTorrent.
[01:46:42] <DeHackEd> 1492 is a pretty minute difference
[01:46:51] <kjetilho> with PPPoE it's 1492, but it doesn't make much diff
[01:47:50] <jimi_> maybe i can surf while torrenting now :D
[01:48:17] <burris> The Bandwidth Utilization Tool From Hell
[01:48:47] <burris> probaly smart to take that out of the readme
[01:48:52] *** Switeck has joined #bittorrent
[01:54:59] <burris> damn I think all the BT v1 code is lost, SF cvs starts with v2
[01:56:26] <The_8472> hehehe, we still have azureus1 code in our repos. someone refurbished it last year to get it working on a basic level again (needed compact announces)
[01:56:34] <The_8472> a single-torrent client!
[01:57:51] <burris> BT v1 has no resemblence at all to what is now known as BT (that starts with v2) but I'd like to have the readme with the tool from hell quote
[02:01:44] <The_8472> well, there is a really simple solution to that bandwidth problem. except for one tiny downside.
[02:02:50] <klapaucjusz> ?
[02:04:53] <The_8472> oh, nothing big. just getting all home routers to support igmpv3 and isps to support large-scale multicast
[02:05:04] <DeHackEd> ... 3com being purchased?
[02:05:30] <burris> which will happen first, multicast or USA switching to metric system?  or ipv6?
[02:05:38] <The_8472> ipv6
[02:05:48] <Switeck> How should a BitTorrent client handle the transition from peer to seed? Should it drop peer connections? Quit tit-for-tat with peers that were previously uploading to it?
[02:05:55] <kjetilho> IPv6 is progressing nicely
[02:06:16] <Switeck> IPv6 is getting big boosts through Teredo and 6to4
[02:06:19] <DeHackEd> Switeck: close connections to other seeds, begin executing its seed behaviour
[02:06:34] * DeHackEd is on ipv6. nice to be able to access laptop behind nat using 6to4 from the office.
[02:06:47] <The_8472> Switeck, you can keep all peer connections. just drop the seed ones. and change to a different unchoking algorithm
[02:07:14] <Switeck> Shouldn't some effort still be made to continue uploading in preference to big contributors?
[02:07:17] <The_8472> i have teredo and 6to4. but it doesn't do all that much good beyond testing bittorrent via v6. because i have a dynamic v4
[02:07:32] <The_8472> Switeck, that's not possible.
[02:07:34] <DeHackEd> Switeck: in theory, but that's awfully hard to measure in practice.
[02:07:47] <DeHackEd> superseeding tries to do that, but that's a special case you don't have the luxery for.
[02:08:05] <Switeck> The data's already collected
[02:08:29] <The_8472> basically there are all kinds of game-theoretic attacks on a seed that would prevent you from building a trust metric that could be used to tell who is a good uploader
[02:08:44] <Switeck> name one
[02:08:52] <The_8472> collusion attacks
[02:09:12] <The_8472> Switeck, only for a tiny subset of all peers. there's churn and conditions change, i.e. you cannot renew your estimates.
[02:09:26] <Switeck> if a seed downloaded 100 MB from a peer, until the peer gets 100 MB back from that seed...I don't see how it's "gamed"
[02:09:34] <The_8472> and only if you do the downloading-seeding transitions. not when you just seed. or restart your client. or peers change their IPs... etc.
[02:09:58] <The_8472> Switeck, you're talking about one insignificant case here
[02:10:07] <The_8472> i'm talking the grand scheme
[02:10:18] <The_8472> you're a seed and there are thousands of peers in the swarm
[02:10:23] <Switeck> Many people stop their torrents within an hour or 2 of finishing downloading.
[02:10:43] <Switeck> in the grand scheme, a 1000-peer swarm is the exception not the rule
[02:10:52] <jimi_> Oh shit, shake that ass ma, move it like a gypsy
[02:10:53] <jimi_> Stop, woah, back it up, now let me see your hips SWING
[02:11:08] * The_8472 glares at jimi_
[02:11:15] <jimi_> :(
[02:11:29] * DeHackEd equips Spaceballs the Flame Thrower   (The kids love it)
[02:11:46] <kjetilho> do you think he will listen to Reason?
[02:11:50] <burris> the very first version of BT v2 used ratios as a seeding strategy... i.e. if you downloaded too much from a seed it would stop uploading to you and start uploading to other peers... it was immediately obvious that was a horrible strategy and bram went from "bringing ratios to the ftp world" to "ratios suck!!"
[02:11:53] <Switeck> right now, I'm seeing seeds get more heavily "gamed" by client bugs :P
[02:12:14] <Switeck> good point, burris
[02:12:36] <burris> not even the first version of BT v2, ratios lasted for a single torrent then they were purged forever
[02:12:39] <The_8472> yeah, it's rate-based tit-for-tat, not volume-based
[02:13:18] <Nolar> seeding mode in az is fair
[02:13:20] <Switeck> I would hope most clients have quit giving favoritism to chronic disconnect/reconnect peers.
[02:13:28] <Nolar> everybody gets an equal chance at my bytes
[02:13:48] <Nolar> just because you have me a lot when i was a peer, doesn't mean it's best for the swarm for me to send everthing back to you
[02:14:12] <burris> what do you mean by "fair" or "equal chance" ??
[02:14:36] <Switeck> I guess that would create too much "trading" and that would result in stratification -- lots of old peers/seeds ignoring new ones.
[02:14:37] <Nolar> you get 30s of my upload and then i give someone else their 30s
[02:14:56] <Switeck> so only optimistic unchoke slot?
[02:15:11] <Nolar> pretty much
[02:15:17] <Switeck> if that's the case, it's *HOPELESS* on big torrents with 4 MB pieces
[02:15:18] <The_8472> Nolar, those are the optimistics. the rest is based on current upload speed. which should essentially be the same for all peers on average. so optimistic unchokes will slowly perturb which peers we're uploading to
[02:15:31] <Nolar> when seeding, you really have no idea who is good or bad
[02:15:43] <Nolar> whereas when downloading, you can just based on tit4tat
[02:15:44] <Switeck> round-robin optimistic unchoke causes really rotten swarm results
[02:15:49] <Nolar> just = judge
[02:15:58] <Switeck> if that's the ONLY upload slot behavior present
[02:16:00] <The_8472> Switeck, not round robin. random.
[02:16:14] <Nolar> why do you say that?
[02:16:34] <Switeck> a 4 MB piece will likely take many optimistic unchoke attempts
[02:16:35] <Nolar> you have a bad downloader impl if you rely on completing pieces from a single peer
[02:16:49] <burris> so if there are 9 dialups and one on a fast connection, the fast downloader gets screwed...  it's fair but not necessarily optimal
[02:16:56] <The_8472> Switeck, multi peer downloading for each piece
[02:16:57] <Switeck> considering each unchoke only lasts 30 seconds and may be separated by many minutes
[02:17:03] <Switeck> a piece can take hours to finish
[02:17:13] <The_8472> you have much to learn young padawan
[02:17:16] <klapaucjusz> Folks, how do you get rid of PEX addresses?
[02:17:17] <Switeck> 1st seed conditions still make it suck ass
[02:17:34] <Nolar> klapaucjusz ?
[02:17:34] <klapaucjusz> There's no hop count in PEX messages, how do you make sure that an old PEX message doesn't get forwarded indefinitely?
[02:17:41] <burris> "rm pex_addresses"
[02:17:44] <Nolar> lol
[02:17:51] <The_8472> pex is not forwarded
[02:17:52] <Switeck> just because it "works" in giant swarms, doesn't mean it's practical or sensible in small swrms.
[02:17:55] <The_8472> it's 0-hop
[02:18:02] <klapaucjusz> Hmm...
[02:18:07] <klapaucjusz> Then Transmission is buggy.
[02:18:13] <Switeck> it also confounds determining who gives you (the most) bad hashes
[02:18:16] <burris> klapaucjusz++
[02:18:18] <The_8472> ... wwaaait
[02:18:22] <The_8472> you're forwarding pex?
[02:18:31] <Switeck> and means a peer/seed handing out bad pieces can corrupt many pieces on many peers
[02:18:57] <The_8472> PEX is meant to allow your peer to get an accurate picture of which peers you're connected to. that's why dropped messages are important too
[02:19:20] <The_8472> Switeck, we have various algorithms against that.
[02:19:40] <Switeck> yes, uTorrent does too
[02:19:48] <The_8472> admittedly there are some pathologic scenarios with 4MB pieces were things can break down, but in the real world it works on almost all torrents
[02:19:57] 
[02:20:05] <Switeck> I am aware of that XD
[02:20:37] <Switeck> the old 5 hash fails drop rule uTorrent used was terrible, easily gamed.
[02:20:43] <The_8472> anyway, regaring uploading unchoke slots:
[02:20:46] <The_8472> <The_8472> Nolar, those are the optimistics. the rest is based on current upload speed. which should essentially be the same for all peers on average. so optimistic unchokes will slowly perturb which peers we're uploading to
[02:21:17] <TheSHAD0W> BitTornado's upload algorithm as a seed is very simple...
[02:21:28] <Switeck> simple may work well enough :)
[02:21:46] <Switeck> Hi, TheShad0w...it's been awhile.
[02:21:53] <The_8472> N optimistics out of M slots. M-N slots are based on current upload speed and thus stick longer with 1 peer than 30seconds. but since they're almost-identical due to TCP distributing the upload fairly the optimistics basically cause churn among the remaining slots
[02:21:54] <TheSHAD0W> It will indeed reward people who've uploaded at the start, but as those stats fade away prior upload ceases to count.
[02:22:17] *** BentMyWookie has quit IRC
[02:22:44] <klapaucjusz> The_8472: what do you mean?
[02:22:47] <TheSHAD0W> After that point, given N slots, it uploads to the N-1 peers who have been able to receive the fastest upload, with one optimistic unchoke.
[02:22:53] <Switeck> I have determined that while seeding, extremely few peer connections are needed. Indeed it doesn't take very many before peers start disconnecting due to inactivity.
[02:23:08] <klapaucjusz> PEX is able to give you extra peers when the tracker/DHT gives you a random sample of the swarm.
[02:23:22] *** stalled has joined #bittorrent
[02:23:25] <klapaucjusz> Forwarding PEX over PEX is useful, but only if you've recently confirmed that the peer is still alive.
[02:23:31] <klapaucjusz> http://trac.transmissionbt.com/attachment/ticket/2574/0001-Don-t-send-PEX-over-PEX-without-reachability-confirm.patch
[02:23:42] <klapaucjusz> Or am I speaking nonsense?
[02:23:45] <Switeck> TheShad0w, do you think there might be conditions where that causes new(er) peers to "starve" for pieces to share?
[02:23:57] <The_8472> klapaucjusz, that violates the assumptions we make in maintaining the pex state
[02:24:10] <The_8472> we're trying to have an accurate picture of each peer's connectedness
[02:24:21] <TheSHAD0W> Switeck: The biggest problem is large piece sizes.  Aside from that, I'd have to say no.
[02:24:30] <Switeck> by connectedness, do you mean if it's firewalled or if it currently exists?
[02:24:30] <The_8472> not of their peers' peers
[02:24:38] <The_8472> neither nor
[02:24:42] <The_8472> of all their connections
[02:24:47] <Switeck> heh
[02:24:47] <klapaucjusz> Hmm...
[02:24:57] <klapaucjusz> Ok, there's a problem here.
[02:25:03] <klapaucjusz> We're interpreting PEX differently.
[02:25:05] <Switeck> I heard that PEX also passes if the peers require encryption to connect.
[02:25:20] <klapaucjusz> Transmission: "hey, try this guy"
[02:25:31] <klapaucjusz> Az: "hey, I'm connected to this guy"
[02:25:40] <The_8472> there are flags in pex to indicate encrpytion, yes
[02:25:48] <klapaucjusz> Since PEX is undocumented, there's no way wto decide who's right.
[02:26:05] <The_8472> klapaucjusz, think a about the purpose of "dropped". it wouldn't make sense if you didn't maintain state
[02:26:18] <klapaucjusz> The_8472: dropped doesn't make sense.
[02:26:31] <The_8472> it does make sense if you preserve an accurate picture
[02:26:39] <klapaucjusz> (Hold on...)
[02:26:42] <The_8472> because you can't just add peers your peer is connected to, you also drop them
[02:27:06] <The_8472> i know what you're going to ask
[02:27:25] <The_8472> and the answer is yes, we keep a list of which peers each peer we're connected to is connected to.
[02:27:29] <Switeck> so the PEX list maintains a rolling subset of the latest peers?
[02:27:36] <The_8472> no
[02:27:40] <klapaucjusz> http://trac.transmissionbt.com/ticket/2430#comment:11
[02:29:49] <The_8472> yes, i remember that conversation
[02:31:29] <The_8472> and you still did the forwarding thing, despite knowing that? *scratching head*
[02:32:13] <klapaucjusz> Hey, that's before I came!
[02:32:37] *** BentMyWookie has joined #bittorrent
[02:33:00] <The_8472> anyway, just because you don't make use of the dropped list doesn't mean others don't either
[02:33:17] <The_8472> thinking beyond one's own... err... how does the idiom go in english?
[02:34:24] <The_8472> we say "to look beyond one's edge of the plate" in german
[02:34:38] <klapaucjusz> The_8472: I agree that Transmission's approach to PEX is broken.
[02:34:46] <The_8472> \o/
[02:34:46] <burris> "use your big head instead of your little one"
[02:34:59] <burris> oh wait, maybe not this time
[02:35:02] <Nolar> the peer-added/peer-dropped approach was taken because traditional gossip protocols tend to suck
[02:35:05] <klapaucjusz> But I'm not going to fix it for 1.80.
[02:35:28] <klapaucjusz> Nolar: ?
[02:35:34] <Switeck> The_8472, something you might want to mention in http://www.azureuswiki.com/index.php/User:The8472/Private_trackers  ...is that firewalled peers/seeds are at a huge disadvantage due to the lack of PEX/DHT.
[02:35:50] <The_8472> klapaucjusz, that might be yet another explanation for it's aggressive connection behavior... i imagine waves of  pex messages rolling through the transmission subgraph
[02:36:21] <The_8472> Switeck, i sortof indirectly mentioned it somewhere at the bottom i think
[02:36:30] *** stalled_ has quit IRC
[02:36:56] <Switeck> being that the firewalled seeds can upload only to peers that have previously contacted the tracker...this means by the time the seeds connect to those peers...those peers may be seeds already also.
[02:37:02] <The_8472> klapaucjusz, wait. i'm not sure... was it you whom i talked with about the connection flood or was it charles?
[02:37:11] <klapaucjusz> both
[02:37:20] <Nolar> klapaucjusz keeping track of currently connected allows us to better choose who to next connect to, i.e. the least connected peer
[02:37:30] <klapaucjusz> Ah.
[02:37:49] <The_8472> and theoretically you could also perform some anti-clustering heuristics...
[02:38:02] <The_8472> well, connecting to the least connected peer is one
[02:38:04] <klapaucjusz> Nolar, The_8472: I'd be really grateful if one of you could explain all of that on trac.transmissionbt.com
[02:38:07] <Nolar> that's one problem with traditional gossip; uneven connections
[02:38:49] <Nolar> it also avoid spamming others with dead peers
[02:39:05] <Nolar> ...and mpaa attacks ;)
[02:39:12] <Switeck> PEX flood attacks?
[02:39:15] <Nolar> yup
[02:39:18] <Nolar> funny story
[02:39:19] <Switeck> "fun"
[02:39:23] <The_8472> mediasentry did that, a lot
[02:39:29] <Nolar> they clearly didnt understand how pex worked
[02:39:31] <The_8472> and they shoot themselves in the foot with it
[02:39:31] <klapaucjusz> Strategy?
[02:39:41] <klapaucjusz> Tell me more, tell me more...
[02:39:41] <Nolar> since they spammed everyone via pex of their fake peers
[02:39:49] <Switeck> I guess a huge array of 100+ poisoners in a narrow netblock would easily fill PEX either on purpose or accident.
[02:39:55] <Nolar> except, our logic connects to the *least* connected peers
[02:40:15] <Nolar> so their spamming in the end just made us less likely to ever connect to their fake peers
[02:40:33] <The_8472> not to mention that they ran into some message-flooding-limites and disconnected themselves
[02:40:42] <Nolar> ya
[02:40:48] <Nolar> they did a terrible job of attacking :)
[02:40:53] <Nolar> <Switeck> I guess a huge array of 100+ poisoners in a narrow netblock would easily fill PEX either on purpose or accident. <<< they did that too
[02:41:00] <Nolar> we do netblock banning ;)
[02:41:26] <Nolar> during their peak, i rarely saw more than 5-10mb of bad pieces
[02:41:29] <Switeck> Nolar, I just meant their sheer "size" as far as number of connected peers relative to total peers in the swarm would cause PEX spamming of each other's ip addresses if nothing else.
[02:41:42] <Nolar> yup
[02:42:17] <Nolar> i havent seen those kind of protocol attacks in years
[02:42:24] <Nolar> now they just sue you
[02:42:36] <Nolar> since they failed miserably at beating the technology itself
[02:43:25] <Switeck> Gnutella was full of similar strategies
[02:43:38] <Switeck> more by spammers than MPAA/RIAA minions
[02:44:13] <Switeck> Alt-locs in Gnutella also had Neg-Alt-Locs passed around too for now-dead file sources.
[02:44:54] 
[02:45:50] <Switeck> used car salesmen use similar strategies
[02:45:59] <Switeck> play up the best points :P
[02:47:28] <burris> it's not just used car salesmen....  surgeons too
[02:48:26] <Switeck> does PEX also pass along if a peer/seed is at (or near) max connections allowed?
[02:48:34] <Nolar> Switeck here's the unchoker code http://azureus.cvs.sourceforge.net/viewvc/azureus/azureus2/com/aelitis/azureus/core/peermanager/unchoker/
[02:48:46] <The_8472> Switeck, no
[02:49:39] <The_8472> Nolar, without IDE support that code is not readable ^^
[02:50:07] <Nolar> without years of java training that code is not readable
[02:50:09] <The_8472> open call hierarchy and open declaration ftw ^^
[02:50:20] <Nolar> although, it's far more commented than most ;)
[02:51:10] <The_8472> and parg's blank lines of doom :/
[02:52:14] <Nolar> i'd point folks to the pex code, but it's rather obfuscated
[02:52:22] <Nolar> much worse than unchoker
[02:53:05] <The_8472> btw, are their brains GPLed now?
[02:54:37] <Nolar> who's?
[02:55:14] <TheSHAD0W> I'd GPL my brain, but I lost the source code.
[02:55:14] <The_8472> whoever read the code. reading and understanding it is creating derivative work!
[02:55:22] <The_8472> ahh, a pity
[02:55:29] <Switeck> <- can't read java :(
[02:55:44] <burris> sorry, reading and understanding doesn't fix the work in a tangible medium of expression
[02:55:47] <The_8472> TheSHAD0W, maybe we should ask sylar if he can understand it?
[02:55:54] * TheSHAD0W is running off the blob right now...
[02:55:59] <Nolar> lol
[02:56:12] <The_8472> burris, neuron interconnections!
[02:56:24] <The_8472> not that different from pits in a pressed CD
[02:57:34] <The_8472> klapaucjusz, you wanted a comment on trac? does it have anonymous login?
[02:58:16] <klapaucjusz> No.
[02:58:24] * The_8472 rages
[02:58:28] <klapaucjusz> You need to register, but you can register anonymously.
[02:58:57] <The_8472> you know how many forums, bugtrackers etc. i always have to register on, mostly for single or once-a-year stuff?
[02:59:07] <The_8472> *singular
[02:59:25] <The_8472> combine that with proper password hygene...
[02:59:40] <klapaucjusz> The canonical password is "password"
[02:59:50] <klapaucjusz> The canonical user name is "user"
[03:00:01] * The_8472 tries that
[03:00:10] <klapaucjusz> The canonical e-mail is obtained from www.jetable.org
[03:00:30] <The_8472> Error
[03:00:31] <The_8472> Invalid username or password
[03:00:32] <The_8472> :(
[03:00:36] <klapaucjusz> Yep.
[03:00:59] <klapaucjusz> just create an acount for "user", or whatever.
[03:01:06] <klapaucjusz> Then forget about it.
[03:01:29] <klapaucjusz> I usually add some large random number at the end of the login, so that I don't use up nice login names.
[03:02:05] <The_8472> that's why i like any system that allows me to post, give a name without registering
[03:02:25] <The_8472> captcha against spambots is fine too
[03:02:36] <klapaucjusz> The_8472: I fully agree, but that's not what we came to speak about.
[03:02:38] * The_8472 is stateless
[03:02:52] <klapaucjusz> We came to speak about how Subversion sucks, and Emacs rules.
[03:03:12] <The_8472> i'm not advocating particular implementations, i'm advocating principles ;)
[03:03:33] <klapaucjusz> Giving distributed access to your source code is a principle.
[03:03:43] * klapaucjusz is in a flamewar mood
[03:03:48] <The_8472> and the circle closes!
[03:04:14] * The_8472 doesn't have anything against distributed version control systems btw. i like them.
[03:04:43] *** GTHK has joined #bittorrent
[03:04:54] <The_8472> btw, user/password works now ^^
[03:05:06] <klapaucjusz> Cool.
[03:05:29] <The_8472> is there a particular entry i could comment on?
[03:06:29] <klapaucjusz> I suggest you create a new entry, and mention that it is related to #2574.
[03:06:38] <klapaucjusz> By creating two entries, Charles can close them independently.
[03:06:46] <The_8472> m'kay
[03:19:00] <The_8472> http://trac.transmissionbt.com/ticket/2575
[03:19:14] <klapaucjusz> Thanks 10^6.
[03:19:50] * The_8472 now rages against the lack of an edit button
[03:20:18] <klapaucjusz> More discussion is needed.
[03:21:18] <klapaucjusz> Consider L (leecher) connected to seed S1.
[03:21:30] <klapaucjusz> S1 in the past connected to S2, but then disconnected.
[03:21:59] <klapaucjusz> If you allow PEXing disconnected but known-good peers, then S1 could point L at S2.
[03:22:18] <The_8472> yes, it could. but it's not necessary
[03:22:36] <klapaucjusz> ?
[03:23:34] *** bittwist has quit IRC
[03:23:54] <The_8472> with a non-clustered graph of peer connections your neighbors are connected to even more neighbors. so your neighbors will pex you a lot more addresses than you actually need
[03:24:15] <The_8472> some additional, possibly outdated information is not going to make it any better
[03:24:16] <klapaucjusz> If swarm is large.
[03:24:26] <The_8472> if it's small you already have a fully connected graph
[03:24:39] <The_8472> since all existing peers will fit into everyone's max connections
[03:24:59] <klapaucjusz> L and S2 are firewalled, only accessible over IPv6.  Tracker and DHT don't support IPv6.
[03:25:14] <The_8472> now you're making up fringe scenarios
[03:25:57] <klapaucjusz> http://trac.transmissionbt.com/ticket/2502
[03:26:24] <The_8472> because, to make it relevant, you'd have to add "S2 is the only seed in the swarm" or something like that
[03:26:32] <The_8472> otherwise L could just connect somewhere else
[03:26:47] <The_8472> and learn of S2 from some other peer, who is directly connected to S2
[03:28:54] <The_8472> not to mention that your NAT scenario would require you to pex IPv6 addresses despite being connected via the nated IPv4 connection and such things
[03:29:02] <The_8472> really
[03:29:37] <klapaucjusz> http://trac.transmissionbt.com/ticket/2575#comment:1
[03:29:41] <The_8472> i haven't seen any disjoint swarms unless i was only one out of... < 5 peers (regardless of seed counts)
[03:29:53] *** init0_ has joined #bittorrent
[03:30:19] <The_8472>  <klapaucjusz> http://trac.transmissionbt.com/ticket/2502 <- doesn't mean you should pex it, since you have not verified yet if that ipv6 address is actually correct/reachable
[03:30:27] <The_8472> otherwise you'd be pexing unverified information
[03:30:41] <The_8472> which... sounds like an attack vector
[03:31:46] <klapaucjusz> No worse than the DHT.
[03:32:13] <The_8472> DHT requires more work because you have to insert yourself around the target keyspace of the lookup
[03:32:29] <The_8472> but yes, fine. it's not relevant to the discussion, you're right
[03:32:45] <klapaucjusz> Nope, you just insert yourself close to some popular torrent.
[03:41:05] *** init0 has quit IRC
[03:42:45] <klapaucjusz> Any native english speakers here?
[03:42:56] <klapaucjusz> A good word that means to remove garbage.
[03:42:59] <klapaucjusz> Cleanse?
[03:43:05] <The_8472> dispose
[03:43:08] <klapaucjusz> (From a data structure, I'm trying to name a function.)
[03:43:19] <The_8472> collectGarbage()
[03:43:20] <The_8472> ^^
[03:43:21] <klapaucjusz> Nah, I'm walking an array and removing garbage from it.
[03:43:33] <klapaucjusz> How poetic.
[03:43:43] <The_8472> those thingies are named garbage collectors for a reason
[03:44:12] <klapaucjusz>  Cull?
[03:45:15] <The_8472> unusual but understandable
[03:45:31] <The_8472> cleanup() would work too
[03:45:48] <The_8472> or purge is popular too
[03:48:05] <Switeck> trash collector?
[03:50:27] * klapaucjusz has just declared a function taking function pointers.  In C.  Without an intermediary typedef.  And got it right.
[03:50:38] <DeHackEd> good
[03:52:11] *** The_8472 has quit IRC
[03:52:29] *** Miller` has joined #bittorrent
[03:52:32] *** wadim has joined #bittorrent
[03:52:34] *** wadim is now known as The_8472
[03:55:15] *** The_8472 has quit IRC
[03:55:36] *** wadim has joined #bittorrent
[03:55:38] *** wadim is now known as The_8472
[03:56:11] <The_8472> rahh...
[03:56:13] * The_8472 stabs his ISP
[03:56:18] <The_8472> |03:59:02| <DeHackEd> good
[03:56:18] <The_8472> |04:00:06| <The_8472> klapaucjusz, you should seek medical attention
[03:56:19] <The_8472> |04:00:11| <The_8472> also http://trac.transmissionbt.com/ticket/2575#comment:2
[03:56:36] <DeHackEd> yeah we missed all that
[03:56:44] <DeHackEd> but you didn't miss anything so...
[04:00:49] <klapaucjusz> Segmentation fault      (core dumped) /usr/local/src/bittorrent/transmission/gtk/transmission
[04:01:15] <The_8472> in java we just use callbacks for that :)
[04:01:27] <DeHackEd> computer over?
[04:01:54] <The_8472> i mean... no juggling with function pointers
[04:02:06] <klapaucjusz> In Haskell we just use a few monadic combinators and we're done.
[04:02:07] <DeHackEd> public class Classname implements Runnable
[04:02:08] <klapaucjusz> Your point being?
[04:03:02] <The_8472> mere dynamic method invocation shouldn't be so fault-prone that it causes runtime errors
[04:03:31] <klapaucjusz> mode flame on
[04:03:50] <klapaucjusz> How long did it take Java to have type-safe polymorphism?
[04:04:12] <klapaucjusz> I was writing type-safe higher-order code when Java was not yet a dream in Gosling's head
[04:04:28] <The_8472> counter: do you still use a handaxe to cut your bread today?
[04:04:38] <klapaucjusz> If you want to use a real language, use Haskell or Caml.
[04:04:45] <klapaucjusz> If you want to write low-level code, use C.
[04:04:53] <klapaucjusz> But what's the point of doing Java?
[04:05:09] <DeHackEd> extreme cross-platform support
[04:05:10] * Nolar is a Scala groupie now
[04:05:25] <DeHackEd> personally that's about the only advantage, but I'm not going to outright say No to java across the board.
[04:05:45] <The_8472> writing low-level code should be optional, not ever-present throughout the program. See what D can do. it has garbage collection, dynamically sized arrays, zeroing of variables. if you want to do lowlevel stuff you can turn that off for specific code sections
[04:05:57] <The_8472> or even embed asm
[04:06:06] <DeHackEd> I'm gonna go over here now.
[04:06:09] <Nolar>  lol
[04:06:11] <The_8472> you can write high level abstraction code on one end, low level on the other
[04:06:47] <DeHackEd> maybe I'll learn to play Nethack...
[04:07:10] <klapaucjusz> flame mode off
[04:07:14] <The_8472> i mean even some embedded systems use C++ these days and only use asm for the drivers
[04:07:35] * klapaucjusz teaches Java, and enjoys the experience quite a lot
[04:07:48] *** bbelt16ag has quit IRC
[04:07:48] *** bittwist has joined #bittorrent
[04:09:08] <The_8472> btw, i do know that java has shortcomings. but i consider C as an anachronism... beautified assembler
[04:09:51] <The_8472> writing applications or even whole kernels in it means writing tons of boilerplate code
[04:10:00] <The_8472> that's exactly the same reason why i don't like J2EE
[04:10:09] <The_8472> lots of high level instead of low level boilerplate code
[04:10:31] <Nolar> you'd like scala
[04:10:44] <The_8472> i like groovy, haven't tried scala
[04:10:45] <Nolar> a lot of the java cruft disappears
[04:10:51] <Nolar> groovy is quite awesome
[04:10:59] <The_8472> it is, indeed
[04:11:10] <Nolar> but i find static typing a godsend as the scale of the project increases
[04:11:29] <klapaucjusz> Funny... I like Java a lot, it improves quite a bit on FORTRAN, what I dislike is the amount of Boilerplate.
[04:11:48] <burris> I hate java, but I love scala which lets me use all the java software
[04:11:53] <klapaucjusz> WRITE(6,1)
[04:11:58] <The_8472> well, if closures would have return types too you could use groovy in a completely static-typed way
[04:12:01] <klapaucjusz> 1 FORMAT H12Hello, world
[04:12:07] <klapaucjusz> STOP
[04:12:08] <klapaucjusz> END
[04:12:15] <klapaucjusz> The equivalent in Java:
[04:12:22] <klapaucjusz> public class Hello {
[04:12:31] <klapaucjusz> public static void main(String[] args)
[04:12:32] <klapaucjusz> {
[04:12:33] <klapaucjusz> ...
[04:12:35] <klapaucjusz> more boilerplate
[04:12:49] <The_8472> well, that gets autogenerated by the IDE
[04:12:53] <The_8472> so i don't count it
[04:13:09] <DeHackEd> Container things = new Hashtable();  // coding done on your behalf, and it works pretty well
[04:13:21] <The_8472> what i mean by boilerplate code is stuff like initiliazing a naming context to get a respository instance, then get a session and then create a query object....
[04:13:37] <burris> groovy is slow, you might as well use jython or jruby
[04:13:45] <klapaucjusz> Can your IDE count the number of characters in a FORMAT statement?
[04:14:06] <The_8472> idk, i don't even know what it does ^^
[04:14:52] <The_8472> verbose syntax != boilerplate code
[04:15:44] <burris> java blows, once you get used to type safe object pattern matching and decomposition you can't go back
[04:16:02] <The_8472> <burris> groovy is slow, you might as well use jython or jruby <- yes. and when it comes to speed they might benefit from the java7 vm. they'll add some dynamic language features to the bytecode language
[04:16:27] <Nolar> if speed matters
[04:16:46] <The_8472> like invokedynamic, although that's already in java6 as experimental feature i think
[04:16:53] <burris> yeah dynamic method invocation will be nice, then I'll take a look at clojure
[04:52:18] *** The_8472 has quit IRC
[05:24:38] *** nGTHK has joined #bittorrent
[05:41:15] *** GTHK has quit IRC
[05:41:44] *** nGTHK is now known as GTHK
[06:00:59] *** jimi_ has quit IRC
[06:15:25] *** klapaucjusz has quit IRC
[06:16:54] *** goussx has quit IRC
[06:33:17] *** MassaRoddel has quit IRC
[06:45:36] *** goussx has joined #bittorrent
[07:08:28] *** Miller` has quit IRC
[07:08:54] *** MassaRoddel has joined #bittorrent
[07:42:38] *** Andrius has joined #bittorrent
[07:53:47] *** bt42 has joined #bittorrent
[08:02:41] *** Miller` has joined #bittorrent
[08:14:57] *** bittwist has quit IRC
[08:40:39] *** cyb2063 has joined #bittorrent
[09:08:51] *** Firon has quit IRC
[09:11:32] *** spoop has quit IRC
[09:22:37] *** Firon has joined #bittorrent
[10:37:00] *** sips1|Work has joined #bittorrent
[10:37:07] <sips1|Work> hi anyone there ?
[10:37:22] <Switeck> yes
[10:38:34] 
[10:38:59] <Switeck> yes
[10:39:04] <Switeck> ban all those ips
[10:39:45] <sips1|Work> that should be done automatically
[10:40:01] <Switeck> that should *NEVER* be done automatically!
[10:40:49] <Switeck> banning an entire country is incredibly contrary to the very concept and purpose of the internet itself
[10:41:02] <Switeck> to do it manually for extremely special-case reasons is one thing
[10:41:33] <Switeck> automatically...is unacceptable.
[10:41:42] 
[10:42:28] <Switeck> shall I ban Canada because many/most of the seeds/peers from there are bad to me?
[10:42:28] <alus> uh, the internet doesn't have a purpose. use it however you like
[10:42:42] 
[10:43:03] <Switeck> if you make it automatic, it sounds like you are no longer the one making that decision.
[10:43:32] <alus> I think he meant automatically finding all the IPs in a given country, and banning them all
[10:43:41] <sips1|Work> right
[10:43:41] <alus> the manual part is selecting which country to ban
[10:43:52] <GTHK> XD
[10:44:52] <Switeck> I was thinking he was meaning to incorporate automatic banning of a country into a (popular?) bittorrent client, thus my objection out of principle.
[10:45:41] <alus> one might ban the country of Hollywood
[10:48:19] <Switeck> you mean like this?:  http://en.wikipedia.org/wiki/Reagan's_%22We_begin_bombing_in_five_minutes%22_joke
[10:49:55] <sips1|Work> imagine that you are in germany and want to download the book "Mein Kampf from Adolf Hitler" which is forbidden in germany... so u dont want to be conneced to german computers.
[10:50:03] <sips1|Work> thats not in my mind... just a example
[10:50:34] <GTHK> Interesting example XD
[10:50:36] <Switeck> German bans that book? censorious bastards want to burn history if they could!
[10:51:27] <Switeck> well, if you mean ban any ip that appears to come from a certain country, it is remotely possible to do but rather difficult
[10:52:14] <Switeck> If you mean actually prevent ALL ips from a certain country connecting to you...or you connecting to them....it is *IMPOSSIBLE*.
[10:52:37] 
[10:52:51] <Switeck> oh?
[10:53:04] <sips1|Work> oh shit i have a meeting now
[10:53:04] <Switeck> so if they're using a proxy or VPN, you know what country they are in?
[10:53:14] <sips1|Work> u can block proxys as well
[10:53:54] <Switeck> but not a decent VPN
[10:54:32] <alus> you can block anyone outside of the latency radius needed to be closer to you than germany :D
[10:54:42] <alus> proxies can't fake the speed of light
[10:55:02] <Switeck> which would probably only work if you were in Germany
[10:55:12] <alus> huh?
[10:55:33] <alus> I could very easily filter Germany. it's very far away from me in time
[10:55:50] <Switeck> Just Germany?
[10:56:01] <kjetilho> but you would filter Poland, too
[10:56:25] <cyb2063> and france and denmark and austria and ...
[10:56:33] <cyb2063> probably whole europe
[10:56:36] <alus> sure.
[10:57:04] <kjetilho> Poland and Denmark were Nazis, too, so I guess it's OK
[10:57:11] <Andrius> or you guys could just unplug network cable or whatever you use there
[10:58:03] <alus> oh come now. much > 50% of my torrent traffic comes from inside the US anyway
[10:58:22] *** goussx_ has joined #bittorrent
[10:58:45] <Switeck> yes, but the question on hand was blocking ONE country...without collateral damage.
[10:59:36] <alus> I don't think that was specified. it was -entirely- blocking one country
[10:59:45] <alus> even given proxies
[11:00:01] <kjetilho> alus: an Austrian could only talk to peers in his own city
[11:00:14] <alus> I can't understand them anyway!
[11:00:17] * alus zing
[11:00:51] <Andrius> Switeck, then you'll have to unplug all network cables in that country
[11:01:13] <Switeck> I believe North Korea is working on that...
[11:01:14] *** goussx has quit IRC
[11:01:14] *** goussx_ is now known as goussx
[11:01:34] <Andrius> lol
[12:07:21] <TheSHAD0W> Uggh...
[12:07:33] <TheSHAD0W> It's too damn earlyyyy...
[12:09:20] <sips1|Work> guys... i found a site wich contains list of nearly all ip ranges coming from one country
[12:09:49] <sips1|Work> its easy to add them on the block list
[12:09:50] <sips1|Work> http://iblocklist.com/lists.php
[12:10:09] <sips1|Work> all connections to germany are banned now :-D
[12:11:06] <DWKnight> probably not
[12:11:13] <DWKnight> most maybe
[12:11:18] <DWKnight> but probably not all
[12:11:19] <sips1|Work> yeah i know
[12:11:29] <sips1|Work> proxys not
[12:13:01] *** jimi_ has joined #bittorrent
[12:13:10] <Switeck> You do realize that even if you've blocked german ips, they can still get your (reported) ip from the tracker
[12:14:26] <sips1|Work> huh?
[12:15:09] <Switeck> If you're connecting to a torrent swarm, others in the swarm can and usually will discover your ip
[12:15:16] <Switeck> even if they cannot connect to you directly
[12:15:38] <sips1|Work> thats ok
[12:15:56] <Switeck> http://dmca.cs.washington.edu/
[12:25:46] <TheSHAD0W> http://newteevee.com/2009/11/11/p2p-villain-or-vilified-bram-cohens-take/
[12:27:29] *** GTHK has quit IRC
[12:35:39] <DWKnight> funi has a clue
[12:35:59] <DWKnight> warner sort of does as well
[12:36:01] <DWKnight> fox doesn't
[12:59:13] *** Andrius[] has joined #bittorrent
[12:59:26] *** Andrius has quit IRC
[13:08:20] *** Andrius[] is now known as Andrius
[13:21:05] *** jimi_ has quit IRC
[14:38:13] *** L337hium has joined #bittorrent
[14:42:18] *** eskil has joined #bittorrent
[14:43:14] <eskil> hey guys. is this a good place to ask questions if i'm attempting to implement my own bt client?
[14:50:53] <Switeck> sure
[14:53:09] <eskil> well. i'm really at the start of this and i'm trying to figure out how the protocol, requests, responses and so on works. i've managed to send a request and got a response. the response however, confuses me. I don't get the part "peers".. I've been reading all I can find in the spec at wiki.theory.org, but it doesn't make any sense considering the response i'm getting..
[14:53:28] <eskil> anyone with nice soures of documentation regarding the subject_
[14:53:29] <eskil> ?
[14:53:48] <Switeck> my exact knowledge of the protocol is weak
[14:54:15] <Switeck> but are you getting a response from a tracker or a peer?
[14:54:21] <eskil> from a tracker
[14:54:24] <Switeck> a peer might reply with PEX info
[14:54:37] <Switeck> 'standard' public tracker?
[14:54:40] <eskil> I'm trying to get a list of IP adresses of the peers seeding a torrent
[14:55:05] <eskil> yep, opentracker
[14:55:38] <Switeck> have you used packet monitoring software?
[14:55:51] <eskil> hm, nope, i havn't
[14:56:05] <Switeck> see how another BT client responds using that
[14:56:31] <eskil> ok, thanks, i'll give that a go :)
[14:56:33] <Switeck> it'll save you a ton of mistakes later
[14:56:55] <Switeck> might even want to use a couple BT clients for that XD
[14:57:05] <Switeck> just to rule out a bad lemon reply.
[14:57:41] <eskil> ah, thx. any suggestions of monitoring softwares btw?
[14:58:51] <Switeck> wireshark is what I've used, though frankly it's pretty terrible.
[14:59:19] <Switeck> very slow
[14:59:41] <eskil> okok, it'l have to do :)
[14:59:42] <Switeck> a problem you may have is incoming connections from peers/seeds once you announce to the tracker
[15:00:01] <Switeck> it's very hard to keep wireshark logs under 10 MB
[15:00:14] <Switeck> because it monitors all network traffic by default
[15:00:30] <Switeck> using your router's firewall to block most everything might help
[15:00:33] <eskil> but i should be able to limit it to a specific port i guess?
[15:00:39] <Switeck> yes
[15:00:49] <eskil> then i should be fine :)
[15:00:51] <Switeck> but tracker updates...are on a random outgoing port
[15:01:38] <Switeck> changing listening port often and firewalling that port using your router may help
[15:02:39] <eskil> thx alot
[15:03:29] <Switeck> good luck
[15:54:49] *** SuN has quit IRC
[16:12:16] *** qe2eqe has quit IRC
[16:41:16] *** eskil has quit IRC
[17:20:27] *** KyleK_ has joined #bittorrent
[17:25:07] <burris> do any clients use endgame mode for completing high priority files?
[17:29:47] *** pevangelista has joined #bittorrent
[17:32:05] *** SuN has joined #bittorrent
[17:32:40] <alus> burris: uT will use endgame if the rest of the files are skipped
[17:35:58] <burris> well of course if it's the very last piece to be downloaded, I think they all will do endgame in that case
[17:37:01] <DWKnight> outside of that, I don't think so
[17:37:28] <DWKnight> bitcomet maybe
[17:37:37] <DWKnight> but bitcomet isn't one to model after
[17:37:50] <burris> bitvomit?
[17:38:58] <burris> does bitcomet let you spoof client id?
[17:50:06] <DWKnight> not sure if it does or not
[17:50:21] *** cyb2063 has quit IRC
[17:55:46] <alus> burris: uT didn't do that for the longest time - it would only go in to endgame if it was the very last piece of the torrent, not the last piece which was to be downloaded
[17:56:10] <alus> so it wouldn't surprise me if other clients had the same problem
[18:19:30] *** MassaRoddel has quit IRC
[18:21:52] *** PN has joined #bittorrent
[18:22:31] *** ProperNoun has quit IRC
[18:24:14] <Andrius> can a single bencoded list contain elements with different types? Like l3:fooi1ee
[18:29:11] <DWKnight> yes
[18:29:22] <Andrius> thanks
[18:33:02] *** MassaRoddel has joined #bittorrent
[18:40:23] *** goussx has quit IRC
[18:45:24] *** K`Tetch has joined #bittorrent
[19:06:17] *** ajaya has joined #bittorrent
[19:11:04] *** goussx has joined #bittorrent
[19:32:33] *** Switeck_ has joined #bittorrent
[19:34:47] *** The_8472 has joined #bittorrent
[19:39:50] *** Switeck has quit IRC
[19:41:49] *** L337hium has quit IRC
[19:52:59] <pevangelista> hello to all
[19:53:15] <pevangelista> Another question today, but about message queueing
[19:53:57] <pevangelista> I read somewere that queueing messages was crucial to network performance, but I didn't quite understand it.
[19:55:24] <pevangelista> for instance, my client just started its operation, and it has no pieces whatsoever. Then it is unchoked by a peer. Should the client queue several piece messages at a time?
[19:55:53] <The_8472> you pipeline requests and thus the other side can queue up piece messages to fill the buffers and send without a break
[19:56:07] <pevangelista> rephrasing, when I am to determine which pieces to download, should I choose several pieces at a time or just one?
[19:56:48] <The_8472> you pick one piece, then pipeline several requests. number of oustanding requests per peer should depend on the current download speed form that peer
[19:58:06] *** jnpplf_ is now known as jnpplf
[19:59:37] <pevangelista> Lets say my queue is five messages long. I would make five piece requests and then send them all in one batch message?
[19:59:48] <pevangelista> Supposing my queue was empty
[20:01:40] <The_8472> not necessarily
[20:01:55] <The_8472> you aren't supposed to queue ALL messages, some should be sent immediately
[20:03:29] <pevangelista> I understand that it is suggested that you queue some messages. This five messages long I used as an example is in this way.
[20:04:39] <pevangelista> What messages should be immediately send? I can think of the following: (un)choke, cancel, have...
[20:04:53] <pevangelista> Or better, queueing works better for piece messages?
[20:05:35] <The_8472> haves can be queued except when the peer is not interested in you because you haven't sent any haves yet that might be interesting for him
[20:07:14] * The_8472 suggests to pevangelista to try to understand the purpose of each part in the protocol before trying to implement it blindly
[20:09:04] <pevangelista> I see, so that he can become aware of the client's pieces more rapidly, correct?
[20:09:44] <The_8472> yes....
[20:10:30] <pevangelista> Actually I haven't started implementing anything yet =)
[20:10:58] <pevangelista> I'm am exactly in the process of understanding the purpose of each part in the protocol
[20:11:40] <DWKnight> there is definitely a priority system for which messages to get out first
[20:13:08] <The_8472> though you should be careful with reordering in the queue. like... if you put a choke in the front you have to clean out all pieces
[20:14:17] <pevangelista> To give context to my questions, I am trying to create a simulation of the BitTorrent protocol.
[20:15:35] <The_8472> most important thing: let's say about 80% of all connections are asymmetric and i guess 40% of them are firewalled (just guesstimates)
[20:16:02] <The_8472> simulating bittorrent with symmetric connections is... well, the perfect working environment for bittorrent, but not realistic
[20:16:18] <The_8472> oh, why do you query me?
[20:17:38] <The_8472> stop PMing me...
[20:18:30] <pevangelista> sure
[20:19:06] <pevangelista> I'm not used to IRC. It's too strange for me, LoL! I'm from the messenger era
[20:20:27] <The_8472> well, most things are discussed publicly on irc. priv messages only make sense if you have lengthy 1 on 1 talks or discuss something that others shouldn't know (but isn't so secret that you should use other protocols instead)
[20:20:55] *** K`Tetch has quit IRC
[20:21:14] <The_8472> to answer your question i suggest you look at my hostname
[20:22:09] *** K`Tetch has joined #bittorrent
[20:22:47] <pevangelista> Good to know. Since we're talking about IRC, how do you sent that suggestion to me? I can only send normal messages
[20:24:09] <The_8472> that wasn't sent to you, it was in this channel too
[20:24:20] <The_8472> i think you call them action
[20:24:39] * The_8472 <what you're doing here>
[20:24:41] <The_8472> /me <what you're doing here>
[20:25:34] <pevangelista> So, you are an azureus developer and also an alien species from the star trek universe (http://en.wikipedia.org/wiki/Species_8472), LOL
[20:25:49] <K`Tetch> hija
[20:25:49] <The_8472> yes yes.
[20:25:59] * pevangelista is learning about IRC!
[20:26:33] <pevangelista> well, nice to meet you =)
[20:36:19] <pevangelista> Another question, The_8472. What time is it there where you are?
[20:36:50] <The_8472> 20:45
[20:37:56] <pevangelista> Here is 17:45. Just to know when it is too late to talk
[20:38:11] <The_8472> :roll:
[20:38:35] <The_8472> you just ask questions any time in the channel. and then someone will answer. maybe in a moment, maybe hours later
[20:39:11] <pevangelista> Ah, so I should leave it open most of the time!
[20:42:17] <DWKnight> I only close for system reboots/power outages/lan parties
[20:43:00] <K`Tetch> I usually have a dedicated system for IRc
[20:43:10] <K`Tetch> I've been on IRC almost continuously for 13 years
[20:45:03] <pevangelista> Wow K`Tech! That's unimaginable to me!!
[20:46:42] <The_8472> about 7 years for me, with a break of a few months
[20:58:16] <pevangelista> cool!
[20:58:32] <pevangelista> Well, see you all tomorrow, probably!
[20:58:42] <pevangelista> Thanks for the help, 8472
[20:59:22] <K`Tetch> i've had a break evry now and then, when i've moved house (or country)
[21:01:04] *** pevangelista has left #bittorrent
[21:49:28] *** ruben23 has joined #bittorrent
[21:49:37] <ruben23> hi
[21:49:48] <ruben23> how do i create a torrent file
[21:50:06] <ruben23> and make it working--> what valid tracker should i used..?
[21:50:26] <K`Tetch> http://www.trackon.org/ <-- trackers
[21:55:20] <ruben23> K`Tetch: is this an working online tracker..?
[21:55:43] <K`Tetch> go to the site
[22:02:16] <ruben23> K`Tetch:  i try to add an excellent tracker on the site but when i test the torrent file its not downloading, what could be the problem..?
[22:02:32] <K`Tetch> since its your torrent, you need to be uploading
[22:02:49] <K`Tetch> and for that you need someone else to upload to
[22:03:35] <ruben23>  K`Tetch: how do i upload it..?
[22:03:45] <K`Tetch> oh boy
[22:03:47] <ruben23> actually ill be sending it to someone
[22:04:04] <ruben23>  K`Tetch:sorry for this but pls help..
[22:04:12] <ruben23> im new to this
[22:04:24] <ruben23> where do i upload..?
[22:04:25] <K`Tetch> http://btfaq.com/serve/cache/15.html
[22:04:39] <K`Tetch> what client are you using?
[22:05:14] <ruben23>  K`Tetch: bittorrent
[22:05:50] <K`Tetch> http://www.utorrent.com/documentation/beginners-guide might be better then
[22:08:00] <ruben23> K`Tetch:..? when i run torretn file it will download directly how to make it upload
[22:08:11] <ruben23> i check the link can see the answer
[22:08:27] <K`Tetch> when you download, it also uploads
[22:09:07] <K`Tetch> here's how to make a torrent, to upload initially - http://btfaq.com/serve/cache/56.html
[22:29:21] <ruben23>  K`Tetch: ill just leave my torrent file on seeding right..?
[22:33:26] <Miller`> ...
[22:34:25] <Miller`> ruben23: bittorrenting isn't like a normal download you'd do from your web browser or something
[22:36:16] <ruben23> Miller`: i just made a torrent file then want to share it to my buddys..
[22:36:41] <Miller`> leave your bittorrent client 'downloading' the file
[22:36:55] <Miller`> and then send (via email?) the .torrent file that you made to your friends
[22:37:37] <ruben23> Miller`:its seeding now my torretn file and its 100% and im not seeing nay upload rate or activity
[22:37:54] <K`Tetch> you won't, unless someone else is also running the torrent file
[22:37:57] <Miller`> Then your buddies aren't running bittorrent also
[22:38:01] <K`Tetch> who are you uploading to otherwise
[22:38:15] <K`Tetch> you're not just uploading into a void
[22:38:56] <Miller`> ( isn't bitcomet the bittorrent void? *cough cough* >.> )
[22:39:11] <ruben23>  K`Tetch:ok ill see it
[22:39:28] <ruben23>  K`Tetch:thanks so much for your assistant..
[22:45:29] *** ruben23 has left #bittorrent
[23:20:58] *** Andrius has quit IRC
[23:39:23] *** Switeck_ has quit IRC





top