[00:05:36] *** rrr_ has quit IRC
[00:12:14] *** stiv2k has joined #bittorrent
[00:12:15] <stiv2k> hi
[00:12:20] <stiv2k> i have a question
[00:12:30] <stiv2k> say that you have 99.8% of a torrent downloaded
[00:12:37] <stiv2k> and there are no seeds
[00:12:54] <stiv2k> and only one other peer that has just as much downloaded as you do
[00:13:08] <stiv2k> is it possible to "brute force" the rest of the bits ?
[00:13:08] <TheSHAD0W> You're stuck.
[00:13:16] <TheSHAD0W> Nope.
[00:13:26] <DWKnight> not practically
[00:13:27] <stiv2k> why? i mean you could verify it with the hash check
[00:13:31] <stiv2k> it might take forever i assume
[00:13:33] <kjetilho> haha
[00:13:35] <TheSHAD0W> You can maybe find a missing file elsewhere.
[00:13:40] <stiv2k> well
[00:13:54] <mpl> then if it's 99.8% of a pr0n movie you're probably not missing much anyway...
[00:13:57] <stiv2k> in my case its every file
[00:13:59] <kjetilho> stiv2k: yes, there are more possible permutations for a single block than atoms in the universe
[00:14:01] <stiv2k> nah its an album =)
[00:14:25] <stiv2k> every file in the torrent i have is <100% done
[00:14:41] <mpl> ouch.
[00:14:42] <kjetilho> sounds like RIAA
[00:14:50] <DWKnight> you only have one practical solution
[00:14:52] <DWKnight> prayer
[00:15:03] <stiv2k> im not going to get into what kind of album or what license its under
[00:15:12] <Switeck> some programs can "repair" damaged audio or video files
[00:15:19] <Switeck> but only by removing the damaged parts
[00:15:40] <stiv2k> damn
[00:15:48] <stiv2k> well that sounds like i wasted ~75 MB
[00:15:54] <TheSHAD0W> A piece varies in size but is usually 2^21 bits.  SHA160 is < 2^11 bits.  That means there are ~ 2^10 possible solutions for a piece that fits the hash.
[00:16:07] <TheSHAD0W> *Finding* such a solution is a bit difficult.
[00:16:17] <mpl> use the force.
[00:16:18] <stiv2k> yes it sounds exhaustive
[00:16:24] <TheSHAD0W> But just any solution won't solve your problem.
[00:16:36] <stiv2k> it was just a theoretical question
[00:16:42] <stiv2k> im not actually going to attempt it
[00:16:53] <stiv2k> that would be rather silly
[00:17:29] <stiv2k> i suppose if you lowered the size of each piece to something more manageable it would be easier then
[00:17:32] <stiv2k> right?
[00:17:35] <TheSHAD0W> Wait.
[00:17:38] <TheSHAD0W> SHA160 is < 2^8 bits.
[00:17:41] <stiv2k> you would just have a *lot* more pieces
[00:17:48] <TheSHAD0W> So there's 2^13 possible solutions.
[00:18:01] <DWKnight> you as a downloader can't modify the piece size of a torrent
[00:18:09] <stiv2k> oh thats right
[00:18:10] <stiv2k> ok
[00:18:33] <kjetilho> TheSHAD0W: so how many different blocks would you expect to try before you find a match?
[00:18:45] <TheSHAD0W> Um.
[00:18:46] <mpl> the current usual size has been to that for efficiency reasons, hasn't it?
[00:18:56] <mpl> *has been set
[00:19:26] <stiv2k> kjetilho: what is the estimate for number of atoms in the universe, for comparison?
[00:19:41] <Switeck> actually, that's probably lower
[00:19:46] <K`Tetch> well, avagadros constant is 6.0223x10^23
[00:20:10] <TheSHAD0W> You'd have to try 2^160 combinations for a 50% probability of finding a solution, going completely brute-force.
[00:20:13] <The_8472> there are about 10^88 atoms in the universe i think.
[00:20:19] <stiv2k> gosh
[00:20:35] <kjetilho> TheSHAD0W: you actually have to find a number which has 2^21 *digits*
[00:20:36] <stiv2k> almost a googol ?
[00:20:38] <The_8472> which is funny, because there are also about 10^88 plack volumes in a cm^3 iirc
[00:20:39] <K`Tetch> avagadros constant is the number of atons in a mole
[00:20:59] <kjetilho> TheSHAD0W: ie. every octet in that block is 8 bits of one long number
[00:21:07] <TheSHAD0W> Right.
[00:21:15] <K`Tetch> (and the muon1 lattiecs have 10^900 permutations)
[00:21:34] <kjetilho> you can divide that by 2^13 for the possible solutions :)
[00:21:37] <TheSHAD0W> I was saying that there were 2^13 possible solutions.
[00:21:52] <kjetilho> yes, I don't think your reasoning was wrong
[00:21:55] *** btipling has quit IRC
[00:21:59] <Switeck> so even a "solution" isn't the correct one
[00:22:05] <TheSHAD0W> Correct.
[00:22:08] <K`Tetch> (and muon1 will crunch through a 10^900 lattice in about 6-9 months)
[00:22:13] <TheSHAD0W> It's ajust a checksum.
[00:22:17] <stiv2k> what is muon1 ?
[00:22:57] <mpl> a muon is an elementary particle, involved in the weak interaction iirc
[00:22:59] <The_8472> something about distributed computing
[00:23:15] <stiv2k> :S
[00:23:26] <The_8472> that's the particle, muon1 is a distributed computing project related to particle physics
[00:23:41] <mpl> oh, ok.
[00:23:49] <stiv2k> cool
[00:23:55] <mpl> lead by CERN?
[00:24:05] <The_8472> idk
[00:24:10] <K`Tetch> no
[00:24:16] <K`Tetch> Rutherford applications Lab
[00:24:20] <K`Tetch> in Oxford
[00:24:25] <mpl> ok
[00:24:27] <K`Tetch> part of the Neutreno Factory
[00:24:43] * The_8472 drops a hydrogen bomb
[00:24:50] <Switeck> too weak
[00:24:53] <The_8472> need more neutrinos?
[00:24:59] <K`Tetch> designing a segment of the accelerator, using 2nd gen distributed computing (evolutionary adapteation)
[00:25:08] <The_8472> fine *smashes 2 neutron stars into each other*
[00:25:31] <Switeck> that'd make a black hole...very little energy might escape
[00:25:34] * TheSHAD0W hands The_8472 a bottle of spin-polarized tetraneutrons
[00:25:36] <K`Tetch> the eventual aim of the neutreno factory is to fire neutrenos from Oxford, to a collector in Japan
[00:25:42] <K`Tetch> and try and understand about them
[00:25:59] <The_8472> Switeck, lots of energy escapes during such events
[00:26:00] <mpl> to kamiokande?
[00:26:18] <mpl> or what was that neutrino detector name...
[00:26:19] <The_8472> they're already working on something bigger. hyperkamiokande
[00:26:26] <mpl> cool.
[00:26:36] <K`Tetch> its not going to be a standard bugbble-chamber or similar
[00:26:37] <Switeck> the gamma ray events prove you right...but they do redshift the energy massively
[00:26:38] <The_8472> and superkamiokande is the current one
[00:26:41] <The_8472> very creative names ~~
[00:26:52] <mpl> hehe
[00:27:17] <The_8472> Switeck, neutron stars already do that by themselves
[00:27:18] <mpl> not sure it beats "godzilla" or "power ranger" ;P
[00:27:27] <K`Tetch> http://stephenbrooks.org/muon1/info/
[00:27:42] <K`Tetch> ^^ official site
[00:27:55] <The_8472> looks like web 1.0
[00:28:04] <mpl> good times.
[00:28:15] <K`Tetch> he spends his time working as a nuclear physist, and coding the client
[00:28:20] <K`Tetch> the websites just got to work
[00:28:47] <uau> the computations above look confused
[00:28:48] <The_8472> that's why you let media-tech students make your website
[00:29:28] <K`Tetch> then it won't work, and won't be usable, The_8472
[00:29:40] <mpl> but it'll be shiny!
[00:29:43] <The_8472> i didn't say designer
[00:29:50] <The_8472> i said media-tech
[00:29:53] <mpl> shiny beats working anytime.
[00:30:17] <uau> i think kjetilho had more of a correct idea, but i don't know where the 'divide by 2^13' came from
[00:30:22] <The_8472> only in science fiction and on the management level
[00:31:34] <kjetilho> uau: since there are 2^13 possible solutions, you should on average be able to find one possible solution after searching 1/2^13th of the trial space
[00:31:49] <TheSHAD0W> Oh.
[00:31:51] <TheSHAD0W> My bad.
[00:31:59] <TheSHAD0W> There are 2^(2^13) possible solutions.
[00:34:09] <Andrius> <mpl> shiny beats working anytime. < that's only in Web 2.0
[00:34:17] <Andrius> afaik 1.9 is still ok
[00:34:25] <kjetilho> TheSHAD0W: yeah.  something like that.  I can't really wrap my head around it.  way too many to try it, anyway :)
[00:34:48] <uau> 2^13? if the piece size is 2^21 bits (256 KiB) and sha1 is less than 2^8 bits then there are more than 2^2^21 / 2^8 possible solutions
[00:34:55] *** klapaucjusz has joined #bittorrent
[00:35:15] <uau> 2^2^21 / 2^2^8 i meant
[00:35:30] <uau> or 2^(2^21 - 2^8)
[00:36:26] <TheSHAD0W> Not 2^(2^21 / 2^8)?
[00:36:45] <TheSHAD0W> 2^(2^(21-8)) ?
[00:36:49] *** bittwist has joined #bittorrent
[00:36:52] <klapaucjusz> What are you counting?
[00:37:09] <TheSHAD0W> Hm.
[00:37:19] <DeHackEd> collisions
[00:37:38] <klapaucjusz> Betwen what?
[00:38:53] <uau> TheSHAD0W: think of it this way: you can't compress the 256 KiB piece much if you only know a 20-byte hash
[00:39:06] <uau> no matter what the computational cost of the "compression" scheme
[00:39:39] <TheSHAD0W> Yeah, it's ~256K-20B possible permutations.  You're right.
[00:39:49] <TheSHAD0W> 262124 possible permutations.
[00:40:00] <uau> unless the data happens to be compressible, in general after you know 20 bytes you still have to get 256 KiB - 20 B of data to know which case matching those 20 bytes it was
[00:40:01] <TheSHAD0W> Er, sorry.
[00:40:01] *** Andrius has quit IRC
[00:40:16] * TheSHAD0W is fuzzy today
[00:40:26] <uau> 8^262124 possiblel cases
[00:40:29] <TheSHAD0W> Yup.
[00:47:10] *** BentMyWookie has quit IRC
[00:49:34] *** BentMyWookie has joined #bittorrent
[00:55:05] *** bt42 has quit IRC
[01:14:32] *** klapaucjusz has quit IRC
[01:22:35] *** Miller` has quit IRC
[01:34:19] *** klapaucjusz has joined #bittorrent
[01:35:20] *** klapaucjusz has left #bittorrent
[01:44:30] *** init0 has joined #bittorrent
[01:47:54] *** init0_ has quit IRC
[01:49:28] <TheSHAD0W> http://it.slashdot.org/article.pl?sid=09/11/16/2327230
[01:52:46] *** ajaya has quit IRC
[01:55:15] *** klapaucjusz has joined #bittorrent
[02:04:49] <TheSHAD0W> There seem to be a lot of router geeks here...  Should I use the HTB or HFSC scheduler?
[02:05:32] <klapaucjusz> Don't bother with either.
[02:05:46] <klapaucjusz> Use plain token bucket with SFB.
[02:06:08] <TheSHAD0W> Well, that's all I have with dd-wrt, and I'm not gonna switch.
[02:06:13] <klapaucjusz> Look ma, no tuning.
[02:06:29] <klapaucjusz> Ah.
[02:06:38] * TheSHAD0W sticks with HTB
[02:07:16] <klapaucjusz> If you want to stay with what's stock...
[02:07:37] <klapaucjusz> I'd say plain HTB with a bunch of RED queues.
[02:07:40] <GTHK> DD-WRT fix that issue where a UPnP portforward of a manually mapped port closes the port but reports to the forwarding application that it was successful? I have an older version I believe.
[02:07:47] <klapaucjusz> Or if you have a fast line,
[02:07:54] <klapaucjusz> simply SFQ over RED.
[02:08:14] <klapaucjusz> Trouble is, RED is a bitch to tune.
[02:08:38] <TheSHAD0W> Well...
[02:09:13] <klapaucjusz> (and don't forget to put ECN on your RED command line)
[02:09:14] <TheSHAD0W> The main thing I use it for is to turn down the bandwidth to my computers, and reserve some space for my VOIP.
[02:09:35] <TheSHAD0W> Meh.
[02:09:45] <TheSHAD0W> I'll just set it for HTB and not screw with RED.
[02:09:56] <klapaucjusz> Oh, don't bother with HTB, then.
[02:10:08] <klapaucjusz> You just want the VoIP to get priority over all the rest?
[02:10:08] <TheSHAD0W> I have exactly two choices.
[02:10:12] <TheSHAD0W> HTB and HFSC.
[02:10:20] <klapaucjusz> And you trust your VoIP users not to starve out the rest?
[02:10:28] <TheSHAD0W> There's only one VOIP user.
[02:10:29] <TheSHAD0W> Me.
[02:10:29] <klapaucjusz> Then just use PRIO.  Simple, comprehensible.
[02:10:39] <klapaucjusz> Do you trust yourself?
[02:10:51] <TheSHAD0W> I have the VOIP set to "premium" priority, which is probably PRIO.
[02:11:06] <klapaucjusz> tc -s qdisc show
[02:11:07] <TheSHAD0W> Or should it be "exempt"?
[02:11:21] <klapaucjusz> No idea.  tc -s qdisc show, paste the result somewhere.
[02:11:40] <TheSHAD0W> Let's see...
[02:12:36] <TheSHAD0W> http://pastebin.ca/1674249
[02:12:46] *** GTHK has quit IRC
[02:14:04] <klapaucjusz> Overkill, but looks okay.
[02:16:46] <klapaucjusz> d1:eli202e12:Server Errore1:t4:fn..1:y1:ee
[02:16:55] <klapaucjusz> That's what I call a heplful message.
[02:18:48] *** Miller` has joined #bittorrent
[02:27:33] <The_8472> no humanreadable string that actually says WHAT the error is?
[02:28:20] <klapaucjusz> it's complete above.
[02:28:30] *** GTHK has joined #bittorrent
[02:28:59] <klapaucjusz> The only information is that t4:fn.. identifies the message in error: it's a find_nodes.  Somebody doesn't like want, I suspect.
[02:30:21] <The_8472> this start to get interesting when someone responds with errors to errors
[02:30:44] <The_8472> bonus points if it sends two error messages per error. to the wrong addresses
[02:31:27] <klapaucjusz> Are you employed by a firewall company?
[02:32:07] <The_8472> no, i just like things going up in flames
[02:35:12] <klapaucjusz> Whoever designed this protocol?
[02:35:17] <klapaucjusz> No id in error messages?
[02:35:30] <TheSHAD0W> LOL
[02:39:51] <The_8472> <klapaucjusz> Whoever designed this protocol? <- people who thought it would be reasonable to leave no room for generic stores and hardcode the source address into PUT operations
[02:43:17] <klapaucjusz> How do you prevent gdb from breaking on USR1?
[02:43:44] *** TheSHAD0W` has joined #bittorrent
[02:50:30] *** TheSHAD0W has quit IRC
[02:50:38] <DeHackEd> handle SIGUSR1 nostop pass
[02:53:25] <klapaucjusz> Thanks
[03:04:19] *** bt42 has joined #bittorrent
[03:09:33] *** klapaucjusz has quit IRC
[03:24:17] *** bittwist has quit IRC
[03:26:30] *** The_8472 has quit IRC
[03:44:30] *** The_8472 has joined #bittorrent
[04:01:45] *** Nolar has quit IRC
[04:01:54] *** Nolar has joined #bittorrent
[04:08:08] *** The_8472 has quit IRC
[04:08:28] *** wadim has joined #bittorrent
[04:08:30] *** wadim is now known as The_8472
[04:33:12] *** Miller` has quit IRC
[04:58:55] *** goussx has quit IRC
[05:09:45] <burris> it was a good idea not to make the mainline dht a generic store, otherwise it would be a lot more prone to abuse
[05:27:50] *** goussx has joined #bittorrent
[06:04:22] <chelz> one thing that's interesting is the lack of time any one seeder would remain on something like a music torrent so compared to other forms of p2p such as limewire, kazaa, etc a person generally would be seeding for less time, so somewhat safer from riaa lawsuits
[06:28:55] *** chelz has quit IRC
[06:30:17] *** Miller` has joined #bittorrent
[07:01:17] *** MassaRoddel has quit IRC
[07:18:18] *** MassaRoddel has joined #bittorrent
[07:27:58] *** Switeck has quit IRC
[07:28:06] *** _rafi2_ has joined #bittorrent
[07:30:45] *** _rafi1_ has joined #bittorrent
[07:49:04] *** _rafi2_ has quit IRC
[08:06:49] *** _rafi_ has joined #bittorrent
[08:24:24] *** _rafi1_ has quit IRC
[08:27:49] *** Andrius has joined #bittorrent
[08:43:53] *** cyb2063 has joined #bittorrent
[09:00:45] *** pajlada has quit IRC
[09:00:45] *** Mazon has quit IRC
[09:02:26] *** Mazon has joined #BitTorrent
[09:06:50] *** Mazon has quit IRC
[09:06:59] *** Mazon has joined #bittorrent
[09:06:59] *** pajlada has joined #bittorrent
[09:07:04] *** pajlada has quit IRC
[09:07:12] *** _Mazon has joined #BitTorrent
[09:08:00] *** Mazon has quit IRC
[09:11:22] *** pajlada has joined #bittorrent
[09:34:51] *** _Mazon is now known as Mazon
[09:43:19] *** razvand has joined #bittorrent
[09:46:43] *** medecau has joined #bittorrent
[10:33:07] *** mxs has quit IRC
[10:33:09] *** Miller` has quit IRC
[10:33:10] *** goussx has quit IRC
[10:33:16] *** TheSHAD0W` has quit IRC
[10:33:16] *** BentMyWookie has quit IRC
[10:33:16] *** waldorf_ has quit IRC
[10:33:17] *** ivan` has quit IRC
[10:33:17] *** fireba11 has quit IRC
[10:33:17] *** SuN has quit IRC
[10:33:18] *** maswan has quit IRC
[10:33:18] *** HandheldPenguin` has quit IRC
[10:33:18] *** uriel has quit IRC
[10:34:48] *** HandheldPenguin` has joined #bittorrent
[10:36:27] *** uriel has joined #bittorrent
[10:37:46] *** mxs has joined #bittorrent
[10:38:59] *** TheSHAD0W has joined #bittorrent
[10:39:29] *** waldorf_ has joined #bittorrent
[10:40:12] *** ivan` has joined #bittorrent
[10:40:14] *** BentMyWookie has joined #bittorrent
[10:40:26] *** goussx has joined #bittorrent
[10:41:25] *** SuN has joined #bittorrent
[10:42:04] *** Miller` has joined #bittorrent
[10:44:17] *** fireba11 has joined #bittorrent
[10:53:44] *** maswan has joined #bittorrent
[10:54:03] *** goussx_ has joined #bittorrent
[11:10:15] *** goussx has quit IRC
[11:10:16] *** goussx_ is now known as goussx
[11:12:52] *** _rafi2_ has joined #bittorrent
[11:15:32] *** GTHK has quit IRC
[11:18:32] *** _rafi_ has quit IRC
[12:33:38] <The_8472> burris, i don't think so. most devs wouldn't even have the motivation to think up a scheme to make some use of a generic store
[12:34:28] <The_8472> and i only said "designing in a forwards compatible way". i didn't say that the spec should have explicit support for it
[13:32:58] *** TheSHAD0W` has joined #bittorrent
[13:32:58] *** TheSHAD0W has quit IRC
[13:41:04] *** HandheldPenguin` is now known as HandheldPenguin
[14:08:46] *** pajlada has quit IRC
[14:10:32] *** pajlada has joined #bittorrent
[14:26:23] *** waldorf_ has quit IRC
[14:35:14] *** pajlada has quit IRC
[14:35:16] *** pajlada_ has joined #bittorrent
[15:14:55] *** Andrius has quit IRC
[15:22:41] *** Andrius has joined #bittorrent
[16:01:52] <cyb2063> oh boy, TPB advertises the use of DHT and magnet links with HEX hashes. some clients cannot handle it yet. I can already see people complaing.
[16:18:50] <The_8472> well, i did say that it's stupid to write the magnet spec with hex hashes when we originally specified them with base32
[16:18:59] 
[16:29:49] <kjetilho> base32?  wtf
[16:30:52] *** TheSHAD0W` has quit IRC
[16:31:21] <The_8472> http://www.azureuswiki.com/index.php/Magnet_link <- we specified it like that, ages ago
[16:31:27] <The_8472> but people can't seem to read
[16:32:23] <kjetilho> base32 is just being different for the sake of being different
[16:32:31] <The_8472> it's shorter
[16:32:34] <kjetilho> why not base17 while you where at it?
[16:32:44] <kjetilho> base64 is even shorter
[16:32:53] <kjetilho> and it's all over the plase
[16:32:54] <The_8472> base64 uses capitalization
[16:33:20] <kjetilho> yes, and "/"
[16:33:30] <The_8472> yup, not good for URLs
[16:37:06] *** Switeck has joined #bittorrent
[16:41:35] *** TheSHAD0W has joined #bittorrent
[16:44:43] *** Elrohir has joined #bittorrent
[17:04:14] *** Elrohir has quit IRC
[17:32:53] *** Switeck has quit IRC
[17:38:52] <DWKnight> The_8472: didn't we have it at base32 for like a year before anyone else supported magnets?
[17:39:43] *** rrr_ has joined #bittorrent
[17:42:00] *** pevangelista has joined #bittorrent
[17:42:08] <DWKnight> I know we had a discussion on using btih instead of sha1 because of conflicts
[17:42:31] <The_8472> well, not sure. but what i can say is that base32 was speced at least in 2005, if not earlier
[17:42:41] <The_8472> the hex spec dates from 2008
[17:49:47] *** KyleK_ has joined #bittorrent
[17:53:37] <pevangelista> Hello, all!
[17:54:39] <pevangelista> When a client is in seed state, does it still requests a peer list to the tracker peridiocally?
[17:55:47] <DWKnight> yes
[17:55:57] <pevangelista> And, in seed state, how does the client determines who should be unchoked if he does not have interest in anyone anymore?
[17:56:25] <DWKnight> have a rotation for unchokes when seeding
[17:57:25] <pevangelista> Like a random selection? It must have some kind of fairness mechanism, righ?
[18:00:23] <mpl> I guess rotation is more like a fifo?
[18:02:24] <pevangelista> I see. Is that so peers cannot be randomly unchoked right after being choked.
[18:02:45] <DWKnight> because of how complex swarm communication can be, most fairness mechanisms that work when downloading don't work when seeding
[18:04:36] <mpl> pevangelista: I like your questions. it's always the ones I wonder about but don't take the time to ask. so, thanks. :)
[18:05:53] <burris> if you assume that tit-for-tat will keep incompletes uploading, then the best strategy for seeders is to upload to whichever incompletes can suck it down the fastestest, with random optimistic unchoking to find fast incompletes
[18:05:54] <pevangelista> Well, thank you! <mpl> Quick question: how do I highlight a message, like you just did?
[18:08:10] <mpl> pevangelista: you just enter the "nick:" before typing in your text, and you can complete with tab on any unix client.
[18:08:43] <mpl> see if I don't type your nick in before it's not hilighted.
[18:09:19] <pevangelista> burris: if the seeder behaves this way, peers with larger bandwidth will download faster, but what about peers with smaller bandwidths? They will hardly be chosen for unchoking, only optimistically.
[18:10:00] <pevangelista> mpl: cool!
[18:10:09] <mpl> pevangelista: but then you get more seeders with fat links in the swarm faster.
[18:10:32] <mpl> so I guess his point is it's better for the swarm overall.
[18:10:38] <burris> yeah, you don't want slow incompletes to dominate limited upload slots of the seeders, slower incompletes can get pieces from peers that are getting them from the seed
[18:11:03] <DWKnight> try to encourage peers to download rarer pieces when you can
[18:11:24] <burris> also, incompletes that download slow also upload slow, so giving out pieces to these peers isn't helping them get to other peers
[18:12:41] <mpl> the longer I hang out here and learn stuff, the neater I find this protocol.
[18:13:02] *** waldorf_ has joined #bittorrent
[18:13:14] <burris> bram is very, very smart
[18:13:28] <burris> and very, very stupid at the same time...
[18:13:41] <mpl> I suppose so. what's he doing now, working on mutorrent?
[18:13:43] <burris> funny how that shit works
[18:13:45] <pevangelista> I think that will work if the seeders keep seeding. In an ideal swarm, the number of seeders tend to get larger, right?
[18:14:45] <DWKnight> it actually has a really weird map of piece distribution that can end up with one of those "uncanny valley" points after the first time getting to self-sustaining
[18:16:32] <The_8472> i have built a piece availability distribution view into vuze, if you gaze at it long enough you can see the evolution of a swarm
[18:16:38] <pevangelista> DWKnight: what do you mean with "the first time getting to self-sustaining"?
[18:16:48] <The_8472> <pevangelista> I think that will work if the seeders keep seeding. In an ideal swarm, the number of seeders tend to get larger, right? <- you're lacking global perspective
[18:16:52] *** bt42 has quit IRC
[18:17:05] <The_8472> the ideal swarm does not exist because more than 1 swarm exists and the number of available uploaders is finite
[18:17:19] <DWKnight> once piece distribution among downloaders (excluding the initial seed) exceeds 1 full copy
[18:17:28] <The_8472> thus they have to quit seeding at some point to tend to newer torrents
[18:17:43] *** bittwist has joined #bittorrent
[18:18:09] <The_8472> torrents undergo a lifecycle with peers dominating at one point, seeds at another one. and both of them tailing of towards zero over time (though that time span can be years for larger torrents)
[18:18:25] <K`Tetch> you know what we need, we need a channel 'site'
[18:18:45] * The_8472 blinks
[18:18:50] <K`Tetch> a more modern version of 'btfaq'
[18:19:22] <The_8472> ah
[18:19:42] <uriel> K`Tetch: what is/was 'btfaq'?
[18:19:47] <K`Tetch> we save answering the same questions over and over, and it's not spread between half a dozen sites
[18:19:48] <pevangelista> like a 'not irc' version?
[18:19:57] <mpl> K`Tetch: that'd be really welcome. Any more docs about BT would be.
[18:20:00] <K`Tetch> btfaq.com uriel
[18:20:11] <uriel> K`Tetch: ah, thanks
[18:20:27] <K`Tetch> and your piece has been pushed back a little, thanks to fucking TPB, the intro has to be re-written
[18:21:10] <K`Tetch> was in the auto-publisher for around now
[18:21:14] <pevangelista> mpl: I agree with that! I had a lot of trouble finding good documentation on BitTorrent.
[18:21:35] <K`Tetch> yeah, it's hard to find, and a lot of the time not written in easily understandble language
[18:22:46] <K`Tetch> I/me has gotten over 40 questions about DHt in the last few hours, for isntance (wonder why...)
[18:22:54] <uriel> K`Tetch: hehe, now you have to say it is a good place for all the tracker-orphans from TPB to get adopted ;P
[18:23:10] <K`Tetch> so I'll write something up to deal with it, but it'll get lost 6 months down the line
[18:23:45] <The_8472> the best knowledge about bittorrent can be found in the heads of the people present in this channel ^^
[18:23:52] * uriel still thinks that moving to a pure tracker-less world is a bad move, dht is great, but has its shortcomings, and there are still many clients that don't support it
[18:24:07] <uriel> (of course, mostly small obscure ones, but diversity is good)
[18:24:09] <The_8472> my only issue with DHT is the lack of scrapes
[18:24:18] <K`Tetch> exactly, the, so lets try and get it written down in a nice easily searchable single location
[18:24:46] <uriel> I might work on something for the FAQ
[18:24:54] <uriel> I needed something like that for trackon.org anyway...
[18:24:57] <The_8472> K`Tetch, answering question is one thing (as annoying as it is), sitting down and writing whole articles is another thing
[18:25:02] <The_8472> it's... like writing documentation
[18:25:18] <uriel> The_8472: worse, it *is* writting documentation :P
[18:25:19] <The_8472> and nobody ever does that unless really pressed to
[18:28:26] <mpl> K`Tetch: just dumping on a wiki Q&A that occur here might be a start.
[18:28:31] <cyb2063> on the other hand people usually do _not_ read documentation.
[18:28:40] <pevangelista> Haha, so true! documentation is a pain to write. You need to be sure that other people can understand what you wrote
[18:29:08] <mpl> cyb2063: I don't see how one can get write any BT related software without reading any doc about it.
[18:29:25] <mpl> s/get//
[18:29:47] <mpl> bah, that's not even specific to BT.
[18:30:00] <pevangelista> mpl: I agree, BT mechanisms are really complex
[18:30:16] <cyb2063> mpl: that's why 'usually'
[18:30:29] <pevangelista> It's not possible to understand it without thinking a lot
[18:30:39] <mpl> pevangelista: I don't think they are complex. but the details are hidden, left to the coder's appreciation.
[18:30:43] <K`Tetch> articles/guides are not like documentation
[18:31:03] <K`Tetch> but I hate writing documentation, i don't mind doing articles
[18:31:08] <K`Tetch> its just the focus of the language
[18:31:09] *** HandheldPenguin is now known as HandheldPenguin`
[18:31:42] <K`Tetch> also, I've been asked for info on magnet link generators either hex or b32 - anyone know of any?
[18:32:10] *** goussx has quit IRC
[18:32:18] <DWKnight> I could probably cobble one together in php in about 10-15 minutes
[18:34:17] <The_8472> <mpl> cyb2063: I don't see how one can get write any BT related software without reading any doc about it. <- people come here and ask absolutely basic questions that should be evident from studying specs and having at least a cursory look at existing clients
[18:35:11] <The_8472> K`Tetch, most clients support magnet links can also generate them for loaded torrents
[18:36:53] <K`Tetch> I think it's more site-based
[18:36:59] <uriel> magnet links are really horrible
[18:37:01] <mpl> The_8472: what you call evident might not be so evident for ppl less knowledgable/smart than you. like me.
[18:37:08] <uriel> a much simpler and saner alternative to magnet links is needed
[18:37:20] <K`Tetch> I don't know the circumstances, I've just been asked
[18:37:23] <The_8472> azureus supports pasting of raw hex infohashes
[18:37:39] <uriel> you want to keep the barrier for people to implement new bt clients as low as possible, and magnetlinks are just a hideous mess
[18:37:41] *** klapaucjusz has joined #bittorrent
[18:37:52] <uriel> The_8472: that is a good improvement
[18:38:20] <klapaucjusz> Folks, http://forum.bittorrent.org/viewtopic.php?id=134
[18:38:22] <mpl> The_8472: I've read the spec on wiki theory countless times, and yet there are a lot of things I still don't understand.
[18:38:25] <uriel> perhaps a really simple uri scheme based on that would be good, perhaps base64-encoded or such
[18:38:28] <klapaucjusz> We really need more nodes.
[18:39:40] <mpl> The_8472: so yeah, I could also read some code. but I would not have to if there was a really complete documentation.
[18:40:29] <MassaRoddel> working code is the best documentation you can get
[18:40:43] <klapaucjusz> MassaRoddel: je couinche.
[18:40:45] <mpl> MassaRoddel: working _and well written code_
[18:41:00] 
[18:41:10] <The_8472> mpl, the thing is that things are not specified because many optimizations emerge from tinkering by client developers... things people would never play around with if things were specified to the last inner loop
[18:41:14] <klapaucjusz> There's no such thing as good human-readable documentation with rationales.
[18:41:21] <klapaucjusz> ...what I've tried to achive with BEP-32.
[18:42:06] <klapaucjusz> The_8472: that's why the spec should be revised after it has been implemented.  You put the optimisations in as implementation notes.
[18:42:13] <mpl> The_8472: I got that. there could still be some documentation about some of those implementations. following them would not be mandatory but it would help a lot the understanding in some cases.
[18:42:25] * klapaucjusz mutters PEX
[18:42:31] <The_8472> klapaucjusz, thing is that different devs do it differently, in mostly compatble ways
[18:42:39] <klapaucjusz> Pex, PEX, pex.
[18:42:39] <The_8472> yeah... pex should be speced
[18:43:00] <The_8472> PEX is not a transparent algorithm that can build on protocol primitives
[18:43:27] <The_8472> piece picking, (un)choking, bad peer detection, reconnect behavior... those things are transparent
[18:43:28] <klapaucjusz> pex egg pex pex bacon and pex
[18:44:03] <klapaucjusz> ..or Lobster Thermidor a Crevette with a mornay sauce served in a Provencale manner with shallots and aubergines garnished with truffle pate, brandy and with a fried egg on top and pex
[18:44:06] <mpl> The_8472: I wouldn't care if that's the specific way azureus or transmission did it. as long as it's a working example it helps a lot.
[18:45:11] <klapaucjusz> I fully agree with mpl.  The implementation notes need not be authoritative, they should just reflect what one half-decent implementation does.
[18:45:33] <klapaucjusz> Romans, countrymen, lend me your ears.
[18:45:52] <klapaucjusz> We need more IPv6 nodes in the DHT.
[18:45:53] * The_8472 cuts off everyone's ears and hands them to klapaucjusz
[18:46:08] <klapaucjusz> We're being spammed by buggy libtorrent (rb) nodes.
[18:46:15] <The_8472> working on it :P
[18:46:21] <klapaucjusz> We need more BEP-32 nodes to put them in a minority.
[18:46:26] <The_8472> it's still only deployed to CVS users
[18:46:31] <klapaucjusz> Spam: http://forum.bittorrent.org/viewtopic.php?id=134
[18:46:46] * klapaucjusz gives back ears
[18:47:22] <mpl> *sigh* I'd love to catch up with btfs and add dht, but I'm not even half done with the basic implementation.
[18:47:47] <klapaucjusz> mpl: in the meantime, just use my implementation to give us a hand.
[18:47:56] <mpl> hah
[18:48:10] <klapaucjusz> mpl: you have looked at my DHT library?  It's easily portable to various implementations of BitTorrent.
[18:48:25] <klapaucjusz> I've included it in both Transmission and Hekate, although they're vastly different designs.
[18:48:39] <The_8472> if peope write in language X they usually don't want to use a library in language Y if they can avoid it
[18:49:09] <mpl> klapaucjusz: I would gladly use it. but as I said, at the moment my client can barely leech and I'm having trouble getting started on the seeding part.
[18:49:27] <klapaucjusz> The_8472: btfs is written in C, no?
[18:49:33] <mpl> yes.
[18:49:34] <The_8472> idk
[18:49:50] <klapaucjusz> mpl: yeah, I understand.
[18:50:38] <The_8472> seeding: N upload slots. 1 (at least) of them optimistic, choke the slowest uploading peer if you change optimistics.
[18:50:50] 
[18:50:54] <The_8472> oh, and disconnect seeds
[18:50:58] <The_8472> beyond that it's like downloading
[18:51:23] <klapaucjusz> What I'm merely saying is that if you need a DHT implementation quick, get my code.
[18:51:45] <klapaucjusz> (Rather than doing the weird multiple-wrapper stuff that LT(RB) does.
[18:51:48] <klapaucjusz> )
[18:51:54] <The_8472> yeah, no objects to that. yours is probably better than many others out there *cough*bitvomit*cough*
[18:52:25] * klapaucjusz doesn't understand the hostility against BitComet.
[18:52:31] <mpl> The_8472: it's very specific details I block on. For example my handshake was working ok against mainline and I could request pieces and all. but I just added the piece of code where I send my bitfield to the other peer, right after the handshake, and it seems to break everything.
[18:52:46] <klapaucjusz> (At least, it's not linked to a media distribution platform)
[18:53:10] <klapaucjusz> mpl: you won't get anywhere unless you show us the code
[18:53:18] <The_8472> hey, i'm not all too happy about that part in vuze either, but it can be turned off at least
[18:53:20] <MassaRoddel> klapaucjusz: does your dht have the "v" field?
[18:53:31] <The_8472> it does
[18:53:34] <klapaucjusz> MassaRoddel: yep.  It's optional.
[18:53:59] <mpl> klapaucjusz: right, I should upload the new version somewhere. lemme see to that.
[18:54:18] <klapaucjusz> mpl: just paste the offending fragment somewhere so we can review it.
[18:54:39] <klapaucjusz> (Not willing to browse through thousands of lines of code, even if it doesn't use delegates.)
[18:55:00] <mpl> klapaucjusz: I can do that, yes. hang on.
[18:55:05] <The_8472> a good repository browser has html anchors for every single line ^^
[18:55:43] <klapaucjusz> MassaRoddel: have a look at the interface on http://www.pps.jussieu.fr/~jch/software/repos/dht/dht.h
[18:56:04] <klapaucjusz> Sample client on http://www.pps.jussieu.fr/~jch/software/repos/dht/dht-example.c
[18:56:19] <mpl> klapaucjusz: please, just don't die from a heart attack at the sight of the code
[18:56:28] <mpl> klapaucjusz: heh, you're in jussieu?
[18:56:34] *** ernest0 has joined #bittorrent
[18:56:43] <klapaucjusz> mpl: I've been correcting first year student projects.  You cannot possibly impress me.
[18:56:55] <klapaucjusz> No, in PRG (13th).
[18:57:04] <mpl> which lab?
[18:57:09] <klapaucjusz> PPS
[18:57:34] <mpl> klapaucjusz: I was working at ipgp until last september.
[18:58:06] <klapaucjusz> Heh.
[18:58:48] <klapaucjusz> If you're in Paris tomorrow, I'll be at http://www.univ-paris-diderot.fr/fscience/MenuSiteFS.php?Site=Paris%20Rive%20Gauche#288
[18:59:03] <mpl> nope, working at iram in grenoble now.
[18:59:49] <The_8472> now... here comes the interesting test. i halted the receieve thread for my DHT
[18:59:58] <The_8472> let's see what happens to the routing table
[19:00:17] <mpl> klapaucjusz: http://pastebin.ca/1675123
[19:00:53] <mpl> klapaucjusz: if I just remove the 4 lines where I send the bitfield all goes fine.
[19:01:40] <mpl> if I don't something goes wrong when I try to send my first request. haven't yet analysed the traffic to check what's going on exactly.
[19:01:45] *** pevangelista has quit IRC
[19:02:11] *** goussx has joined #bittorrent
[19:05:57] *** klapaucjusz has quit IRC
[19:06:20] *** klapaucjusz has joined #bittorrent
[19:06:27] <klapaucjusz> Sorry for that.
[19:06:32] <klapaucjusz> Missed anything?
[19:06:42] <mpl> klapaucjusz: you got my paste?
[19:07:38] <klapaucjusz> Yep.
[19:07:54] <klapaucjusz> It's pretty clean code, but you've got multiple problems in there.
[19:08:03] <klapaucjusz> First of all, you're not dealing with partial writes.
[19:08:17] <klapaucjusz> write/send/sendmsg/sendto can return less than the length.
[19:08:29] <klapaucjusz> Anything between 1 and the length, you need to deal with that.
[19:08:44] <mpl> wait, I thought I had some readn or something
[19:09:23] <mpl> but yeah I'll surely double check all that. however it's not likely to be the cause of my problem here
[19:09:30] <mpl> since it's systematic.
[19:09:33] <klapaucjusz> Indeed.
[19:09:48] <klapaucjusz> How much is BTbitfield?
[19:09:53] <klapaucjusz> Do you set the extra bits to 0?
[19:09:56] <mpl> also I know it's totally wrong that it's all sequential
[19:10:04] <mpl> I should not "expect things"
[19:11:03] <mpl> BTbitfield is in an enum, I used the code in a switch to deal with the cases
[19:11:12] <mpl> probably 5 or 6, I could check.
[19:11:17] <klapaucjusz> Yes, it would be simpler in an event-driven style.
[19:11:18] <klapaucjusz> Ah.
[19:11:27] <mpl> yes, the bits are set to 0
[19:11:44] <mpl> they're all set to zero when I create the bitfield array.
[19:11:55] <The_8472> yus, the routingtable survives!
[19:12:05] <klapaucjusz> i don't get it.
[19:12:16] <klapaucjusz> What's the type of tor->bitfield?
[19:12:23] <mpl> char *
[19:12:39] <klapaucjusz> how does writemsg find out its length?
[19:13:13] *** bittwist has quit IRC
[19:13:16] <klapaucjusz> (Hint to The_8472: C's strings and arrays are a major case of braindamage.  Feel free to flame)
[19:13:24] <mpl> it's figured out when the bitfield is first created, and the initial files scanned to see what parts we already have.
[19:13:35] *** bittwist has joined #bittorrent
[19:13:41] <klapaucjusz> writemsg knows this information?
[19:13:51] <The_8472> klapaucjusz, they're not 0-initialized, i know. and not length-prefixed
[19:13:54] <The_8472> silly design
[19:14:11] <The_8472> though, isn't there some alloc that zeros everything?
[19:14:11] <klapaucjusz> Not zeroed is reasonable.  Not length-prefixed is brain-damaged.
[19:14:56] <mpl> klapaucjusz: here's the corresponding case in writemsg: http://pastebin.ca/1675133
[19:15:16] <The_8472> zeroing things makes things more deterministic, and pointers explode in your face instead of pointing... somewhere... somewhere that might be set to a previously valid value
[19:15:49] <The_8472> especially with that pesky stack allocation stuff
[19:15:51] <klapaucjusz> Heh.  That's nothing compared to the pain of dealing with the stupid \0 termination.
[19:16:04] <The_8472> ah true
[19:16:13] <mpl> The_8472: yeah, there's allocz() afair
[19:16:30] <The_8472> alloc should be deprecated ^^
[19:16:51] <mpl> hmm mallocz() actually.
[19:17:43] <klapaucjusz> The_8472: calloc
[19:19:25] <The_8472> the problem with years of OOP programming is that i always think "can't i put this into a struct. and store the length. and put that into another struct... and..."
[19:19:29] <klapaucjusz> mpl: how do you compute bitfieldsize?
[19:20:17] <MassaRoddel> is dht v=TR Transmission?
[19:20:22] <The_8472> yes
[19:20:38] <MassaRoddel> thx
[19:20:48] <mpl> 	if(tor->piecesnb % 8 != 0)
[19:20:48] <mpl> 		bitfieldsize = tor->piecesnb / 8 + 1;
[19:20:48] <mpl> 	else
[19:20:49] <mpl> 		bitfieldsize = tor->piecesnb / 8;
[19:21:53] <mpl> klapaucjusz: I figure the bitfield is ok, because the other peer doesn't drop the connection right away
[19:22:11] <mpl> klapaucjusz: since the choke and interested messages just afterwards seem to go through
[19:22:30] <mpl> the problem occur a bit later, just after the end of hello()
[19:22:31] <klapaucjusz> So what's the problem, then
[19:23:03] <The_8472> "seem to go through"? you know, it could just be attached to the stream and the other client just hasn't decoded your handshake yet
[19:23:07] <mpl> klapaucjusz: the problem occurs just when I try to send the request for my first piece.
[19:23:15] <mpl> The_8472: indeed.
[19:23:40] <The_8472> just because you can write it to the socket doesn't mean all previous messages were valid
[19:23:56] <mpl> yes
[19:24:15] <klapaucjusz> Line 91: missing brackets?
[19:24:53] <mpl> oh
[19:24:57] <mpl> probably yes.
[19:25:04] <mpl> thx
[19:25:32] <mpl> The_8472: thx, I always get biaised thinking like that, in terms of delays...
[19:25:56] <klapaucjusz> Who's MassaRoddel?
[19:26:18] * klapaucjusz is probably a MAFIAA agent trying to penetrate the BitTorrent community
[19:26:41] <The_8472> doubtful, they have displayed lack of insight in the past
[19:27:18] <klapaucjusz> Who's MassaRoddel, then?
[19:27:54] <The_8472> idk, someone from germany it seems
[19:28:41] <mpl> klapaucjusz: assuming the bitfield I created is the problem, what could be wrong with it if not its length?
[19:29:16] * Andrius Welcome to #bittorrent; topic of the day: "Who's MassaRoddel"
[19:29:26] <mpl> brb
[19:30:45] <The_8472> mpl, have you tried sending an all-0 bitfield and checked if it makes any difference?
[19:30:49] <MassaRoddel> that would be me
[19:31:39] * klapaucjusz shakes hands
[19:32:50] <mpl> The_8472: hmm, I haven't. gonna try right away, thx.
[19:38:22] <mpl> The_8472: hmm, no luck with that.
[19:46:45] *** Miller` has quit IRC
[20:10:27] *** razvand has quit IRC
[20:39:29] *** bittwist has quit IRC
[20:42:07] *** ajaya has joined #bittorrent
[21:13:40] *** razvand has joined #bittorrent
[21:14:05] *** klapaucjusz has quit IRC
[21:26:13] *** ajaya has quit IRC
[21:32:30] *** waldorf_ has quit IRC
[21:41:19] *** Switeck has joined #bittorrent
[21:44:20] *** pvvni has joined #bittorrent
[21:49:41] *** chelz has joined #bittorrent
[21:50:24] <pvvni> Anyone got invites to cinematik?
[21:50:43] <pvvni> I'll trade a google wave invite for one :D (or a waffles/what, but I know that's not gonna happen)
[21:51:25] <Switeck> pvvni, you are in the *WRONG* forum for that
[21:51:44] <K`Tetch> damn, the invites thing has gone from the topic
[21:51:48] <chelz> s/forum//channel
[21:51:55] <pvvni> Switeck: what K`Tetch
[21:52:00] <pvvni> gone from topic -> legit.
[21:52:04] <pvvni> :D
[21:52:08] <chelz> lol
[21:52:16] <chelz> yeaah no
[21:52:28] <Switeck> "Do NOT ask about files you can download, do NOT ask where to find files"
[21:52:35] <Switeck> invites definitely falls under that
[21:52:35] <pvvni> I didn't ask about either.
[21:52:42] <pvvni> I know where to find them
[21:52:47] <pvvni> I just cant get to said places... yet
[21:53:06] <pvvni> I'm like napolean trying to conquer russia
[21:53:09] <pvvni> I know it's up there
[21:53:13] <pvvni> but that damn russian winter.
[21:53:29] <Switeck> seek elsewhere then
[21:53:38] <pvvni> righto, in that case
[21:53:40] <Switeck> both for how to get where you want...and for files you want yesterday
[21:53:48] <pvvni> Anyone recommend good cli clients?
[21:54:03] <pvvni> Used rtorrent, but meh
[21:54:07] <pvvni> not impressed.
[21:54:22] <pvvni> A webui would be cool, but by no means necessary.
[21:54:44] <hlindhe> rtorrent can use webuis
[21:55:41] <pvvni> I just don't like the way it manages torrents. Doesn't seem effective for managing many at once
[21:56:08] <pvvni> ideal would be vi-based bindings for selection/navigation
[22:03:38] *** bittwist has joined #bittorrent
[22:14:02] *** medecau has left #bittorrent
[22:17:14] *** thermal has joined #bittorrent
[22:35:36] <DWKnight> getting required troubleshooting information from users is like pulling teeth sometimes
[22:37:19] *** stalled has quit IRC
[22:38:01] *** stalled_ has joined #bittorrent
[22:39:00] <Switeck> It's a terrorist/hostage negotiation crisis. :P
[22:39:41] *** GTHK has joined #bittorrent
[22:41:47] *** rrr_ has quit IRC
[22:43:06] *** _rafi_ has joined #bittorrent
[22:43:11] <Andrius> it's annoying that users generally have no common sense. Well, like any average internet user...
[22:46:59] <Switeck> XD  http://forum.utorrent.com/viewtopic.php?pid=436766#p436766
[22:51:10] <Andrius> ....
[22:51:26] <Andrius> maybe he followed some speed guide on youtube?
[22:51:33] <Switeck> it's very popular settings...straight from YouTube!
[22:52:51] * The_8472 facepalms
[22:53:20] <Switeck> the most popular video for it only has over a million views
[22:53:29] <Switeck> so it's not like a lot of people use it.
[22:55:16] <pvvni> Mmm
[22:55:20] <pvvni> one size fits all approach
[22:55:22] <pvvni> I like it.
[22:55:32] <pvvni> It's gotta be legit, it has over a million views.
[22:56:39] <Andrius> someone should sue youtube, unless they give bt client authors the right to remove these setup guides
[22:59:34] <Switeck> The settings represent a DDoS attack on the whole internet :P
[22:59:53] <Switeck> they recommend raising half open connection attempts at once to 500.
[23:00:07] <Switeck> that's how many outgoing peer/seed connections to try at once.
[23:00:19] *** _rafi2_ has quit IRC
[23:00:59] *** stalled_ has quit IRC
[23:01:05] <Andrius> "halp me, internet doesn't work when I download torrents!!!"
[23:01:55] <Switeck> In that sense, bad routers and software firewalls "protect" the rest of the internet from some of that by crashing first. :P
[23:02:41] <Andrius> so that's why microsoft limits half open connections...
[23:03:10] <Switeck> that figured at least in small part to why M$ limits half open
[23:04:32] <cyb2063> yet people 'patch' that limit out of windows and mess things up.
[23:05:10] <Andrius> cyb2063, I don't consider patching it a bad idea. If client uses sane settings it won't ddos other people
[23:05:40] <Switeck> Chances are, if you "need" the limit patched...you're not using sane settings.
[23:05:43] *** stalled has joined #bittorrent
[23:05:51] <cyb2063> Switeck: yep
[23:06:14] *** waldorf_ has joined #bittorrent
[23:06:18] <Switeck> A possible exception would be ramp-up for a 100 mbit/sec or 1 gbit/sec seedbox with 100+ active torrents
[23:06:31] <Switeck> but even it doesn't need 10x Microsoft's limit.
[23:07:05] <chelz> uriel: ping
[23:07:21] <uriel> chelz: yes?
[23:08:27] <chelz> uriel: have you thought of allowing adding trackers to Trackon through that submit on the frontpage in the form of many delimited by say something like semicolons?
[23:08:37] <chelz> uriel: if not, that might be cool
[23:09:07] <uriel> chelz: actually it is possible but undocumented, (or it was at some point, not sure if I removed the feature)
[23:09:15] <uriel> chelz: do you really have so many trackers as to need that? :)
[23:09:51] <chelz> uriel: ah no, but it was been pretty much the only limiting factor for me submitting torrent sites
[23:10:07] <chelz> just makes it easier, without having to write some kind of script to submit them all one by one
[23:10:17] <uriel> chelz: how many have you got? :)
[23:10:48] <uriel> ok, just checked, the code still allows you to enter multiple trackers separated by spaces...
[23:11:07] <uriel> but the errors might be a bit off, as it stops if there is a problem with one of them
[23:11:24] <chelz> uriel: not sure, just for fun i occasionally put all the trackers in a few popular torrents i'm running in a textfile. before submitting i'd probably sort them and remove duplicates.
[23:11:33] <chelz> problem? as in not responding?
[23:11:54] <uriel> no, as in malformed url
[23:11:57] <chelz> i just remember semicolons because it's how the Az webgui allows you to add many URLs to add torrents. spaces is a good idea too though. any non-uri/url character
[23:11:59] <chelz> ah
[23:12:20] <uriel> (or perhaps if it is already being tracked, which might be more of an issue in your case)
[23:12:40] <chelz> yeah
[23:13:13] <chelz> there wouldn't happen to be a nicely formatted .txt of all the torrents in trackon already, would there?
[23:13:44] <chelz> also well before documenting it the stopping on a malformed url probably should be remedied, but good to hear it already has support :)
[23:13:58] <chelz> btw congrats on the torrentfreak article
[23:14:07] <uriel> well, it was mostly a hack to bootstrap things with a list of trackers I had around :)
[23:14:10] <uriel> thanks
[23:14:29] <uriel> I'm thinking of perhaps prividing a REST api to submit new trackers, would that work for you?
[23:14:53] <chelz> that's bit over my head, but i'm sure others could find that useful
[23:15:08] <uriel> heh
[23:15:26] <uriel> well, I think I will eventually move the tracker submission to a different page that would allow a more rich UI
[23:15:57] *** _rafi_ has quit IRC
[23:16:00] <uriel> still, part of my idea was to allow entering of extra metadata (like homepage, name, etc) which would not work for submitting a long list of trackers in batch
[23:17:12] <K`Tetch> and adding them singly also means the update is staggered
[23:18:31] <uriel> well, update is staggered already because the actual checking is done asynchronously, in part so even if for whatever reason the tracker happens to be down, we try again in a few minutes
[23:18:44] <The_8472> bulk add screams for abuse ^^
[23:18:54] <K`Tetch> yeah
[23:19:08] <K`Tetch> you just can't be adding THAT many that it's an issue
[23:19:08] <uriel> The_8472: yea, that is also part of the reason...
[23:19:21] <K`Tetch> (that singly is so slow)
[23:19:33] <uriel> well, you could flood the input queue and other trackers would take a long time to get in, but still I think it should be mostly ok
[23:19:39] <chelz> hardly any more abuse than what a person could do already with a script submitting one by one
[23:33:05] *** JudgeSHAD0W has joined #bittorrent
[23:37:29] *** JudgeSHAD0W has quit IRC
[23:58:33] *** razvand has quit IRC





top