[00:04:25] *** burris has joined #bittorrent
[00:11:50] *** bbelt16ag has quit IRC
[00:15:10] *** burris has quit IRC
[00:19:37] *** burris has joined #bittorrent
[00:23:46] *** h2ro has joined #bittorrent
[00:52:49] *** init0_ has joined #bittorrent
[00:59:13] <Switeck> HOSTILE ips - BAYTSP:  http://pastebin.com/f54c2cd55
[00:59:27] <Switeck> do note first 2 lines are to block bogons and multicast!
[01:00:13] <h2ro> baytsp?
[01:02:34] <Switeck> google/wiki them :P
[01:05:07] *** init0 has quit IRC
[01:11:15] <h2ro> i'm asking myself: do you guys really use these lists
[01:11:32] <h2ro> or is it only to silence the stupid paranoids?
[01:16:23] <chelz> the second
[01:16:52] <chelz> people develop superstitions around things they don't understand. it's a digital form of a talisman.
[01:19:10] <Switeck> I use that list
[01:19:25] <Switeck> Got a problem with it?
[01:19:34] <h2ro> I guess not much has changed since the dark ages
[01:19:38] <Andrius> burn heretics!
[01:19:53] <Switeck> Do you even know what BAYTSP is?
[01:20:10] <h2ro> I just looked it up as you commanded me to do
[01:20:27] <Switeck> good, then why would you want them connecting to you?
[01:20:41] <h2ro> Huh, I don't mind anyone connecting to me
[01:20:45] <h2ro> as long as they do it digitally
[01:20:49] <h2ro> and don't dos me
[01:20:54] <Switeck> they do DOS attacks
[01:21:03] <Switeck> DDOS to be specific
[01:21:22] <h2ro> well, I would not be happy about that
[01:21:27] <Switeck> I'm not
[01:21:31] <h2ro> but they haven't attacked me at all
[01:21:36] <h2ro> never
[01:21:55] <chelz> i'm pretty sure the point of a DDoS is that it can't be blocked
[01:21:56] <Switeck> but I had to do research to figure out what ranges are really theirs instead of trusting Peer Guardian's extremely huge list
[01:22:08] <Switeck> chelz, depends on the type
[01:23:05] <h2ro> why don't you guys just sue these fuckers
[01:23:48] <h2ro> you have their ip adress, you have the guys responsible for it, what is stopping you?
[01:23:51] <chelz> similar reasons to why really illegal stuff can be hosted offshore i think
[01:24:13] <h2ro> so how can you prove baytsp is behind all this
[01:25:45] <h2ro> you guys are probably just scaremongering, or even worse, perhaps you *are* Baytsp and try to dos us by putting the big seeder's adresses on these lists
[01:27:17] <Switeck> pinging someone isn't illegal
[01:27:19] <Switeck> it's just annoying
[01:27:35] <Switeck> h2ro, do the research...
[01:27:46] <Switeck> I'll give you a link or 2
[01:28:15] <Switeck> http://www.cidr-report.org/cgi-bin/as-report?as=14478&view=2.0&v=4
[01:28:30] <Switeck> http://www.dslreports.com/whois
[01:28:35] <Switeck> http://www.ripe.net/ripencc/pub-services/db/whois/whois.html
[01:28:40] <Switeck> http://www.apnic.net/apnic-bin/whois.pl
[01:28:57] <chelz> call me reckless but i'm not going to ban any IPs until i start getting DDoS'd. even then i'd just call up my ISP to get a new IP.
[01:28:59] <Switeck> If they say someone/something is registered to a certain AS or ip address, it is.
[01:29:35] <Switeck> I'm not saying you have to or even NEED to use that range
[01:29:42] <Switeck> for me, it's just an annoyance blocker
[01:30:05] <Switeck> If they're poisoning the torrents I'm on, I get the torrent a little faster this way
[01:30:17] <Switeck> otherwise, i have to wait till auto-banning occurs.
[01:30:28] <Switeck> (which is forgotten if I restart the program)
[01:31:22] <chelz> yeah that forgetting is unfortunate
[01:31:41] <chelz> would be nice if remembering was added in as an option into clients
[01:32:00] <Switeck> thus my short list
[01:34:28] <Switeck> I was surprised with changing of hands that many of the ips remained owned by the same guys, different name.
[01:34:37] <Switeck> http://torrentfreak.com/meet-dtecnet-riaas-new-anti-piracy-partners-090113/
[01:34:45] <Switeck> http://torrentfreak.com/riaa-anti-piracy-partner-clueless-about-bittorrent-091028/
[01:35:29] *** goussx has joined #bittorrent
[01:44:12] <Switeck> the list is not good for "hiding" from monitoring, only from reducing poisoners a little
[01:58:16] <anakh> baytsp has their own asn ?
[01:58:33] <Switeck> seems so :P
[01:58:50] <Switeck> they're contributing to IPv4 exhaustion.
[01:59:03] <anakh> should setup live monitoring of new prefixes they announce :p
[01:59:05] *** goussx has quit IRC
[01:59:25] <anakh> even worse they seem to be announcing a lot of /24's
[01:59:45] <anakh> which means that their fucking networks unneccessarily take up memory/resources in EVERY CORE ROUTER ON THE INTERNET
[02:00:30] <Switeck> It'd be worthwhile to "untangle" a few large ISPs that have 1000's of tiny ip ranges and give each an equal sized continuous space.
[02:00:38] <Switeck> But changing out all the ips would be immense :(
[02:01:16] <anakh> it'd be nice if everyone just refused to peer with them :PP
[02:01:34] <anakh> so they can sit there and waste forwarding table entries in their OWN routers
[02:01:44] <Switeck> what creeps me out is there's ASNs not associated with ip addresses
[02:01:55] <Switeck> Sony's got a few :P
[02:06:13] *** burris has quit IRC
[02:08:21] *** burris has joined #bittorrent
[02:09:15] <anakh> ..or they could just avoid having their asn in the aspaths :p
[02:11:09] <Switeck> Would those by their very nature be internal networks only?
[02:11:34] <Switeck> or could they still be outward-facing, complete with either their own ips (IPv4 or IPv6)?
[02:13:38] *** goussx has joined #bittorrent
[02:20:41] *** nGTHK has joined #bittorrent
[02:22:23] <The_8472> <h2ro> but they haven't attacked me at all <- i saw their ranges recently in a torrent
[02:22:41] <The_8472> the odd thing was that the torrent was already on its way towards the end of its lifecycle
[02:23:04] <The_8472> but they flooded it with many fake seeds
[02:23:25] <The_8472> when i saw it i banned that range and the torrent started to hum along as usual
[02:41:07] *** GTHK has quit IRC
[03:00:27] <The_8472> anakh, i can maded you forum section, but then i hided it
[03:01:04] <anakh> parse error
[03:01:21] <The_8472> ^^
[03:01:36] <The_8472> http://forum.bittorrent.org/viewtopic.php?id=136
[03:04:31] *** nGTHK is now known as GTHK
[03:06:30] <anakh> i wanna locate the MAFIAAs sabotage clients in DHT and perform certain offensive maneuvers :p
[03:07:56] <anakh> consider it network self defence
[03:08:10] <The_8472> to use one of my favorite ambigous phrases: we cannot sanction such behavior
[03:08:11] <The_8472> ^^
[03:10:48] <anakh> sure, dht is vulnerable, so lets do a good preemptive strike :)
[03:11:03] <Switeck> I don't want to teach them how to do more damage with DHT
[03:11:08] <Switeck> better to give them a silent treatment
[03:11:17] <Switeck> connect, but send nothing
[03:11:29] <Switeck> and toss out everything they send you
[03:11:35] <anakh> i think they would have to reconsider their operations if their clients well.. misbehaved :)
[03:11:57] <Switeck> no they don't
[03:12:12] <Switeck> they simply don't care until FBI comes to pay them an unfriendly visit
[03:12:32] <Switeck> "oh my bad, wasn't aware it was DDOSing that website for the last week")
[03:12:53] <anakh> wouldnt they if their antip2p bots started DoSing .. :)
[03:13:12] <anakh> exactly.. thats one of the more uhm violent strategies
[03:13:18] <Switeck> http://revision3.com/blog/2008/05/29/inside-the-attack-that-crippled-revision3/
[03:13:29] <Switeck> and DONE!
[03:13:42] <anakh> although it'd probably be better for everyone if they just attacked themselves :)
[03:15:09] <anakh> does anyone have any hard data on how my darlings behave on DHT ?
[03:15:21] <anakh> T mentioned incorrect versions, lots of id changes, etc
[03:16:45] *** bbelt16ag has joined #bittorrent
[03:17:28] <anakh> should do a bit of non-invasive monitoring and see what pop ups around torrents they might be after
[03:18:18] <TheSHAD0W> Regarding the Revision3 attack...
[03:18:21] <The_8472> well, i never looked into anti-p2p nodes, just misbebhaving nodes in general
[03:18:31] <TheSHAD0W> I can't see why shutting out torrents would result in a syn attack.
[03:18:57] <TheSHAD0W> Peers running on their tracker would be able to succeed in connecting, then get tossed.
[03:19:11] <Switeck> the SYN attack was used to shut the website down
[03:19:19] <Switeck> because it was hosting questionable torrents
[03:19:52] <Switeck> (both unwittingly and unknowingly!)
[03:20:10] <TheSHAD0W> What I'm saying is that it had to be an attack, rather than a flood of peers trying to get on the tracker.
[03:20:12] <anakh> without knowing the details i dont think that was a purposeful attack actually
[03:20:35] <TheSHAD0W> Even if the peers kept retrying, it would not have been a syn attack.
[03:20:37] <anakh> noone in their right state of mind would send 8000 non-spoofed SYNs/sec as a DoS attack
[03:20:40] <Switeck> anakh, being that the attack vector was not standard BitTorrent client behavior...it was an attack!
[03:20:57] <anakh> more like severely buggy software
[03:21:14] <The_8472> mediadefender wrote their own poisoning software
[03:21:19] <Switeck> can't that still be criminal negligence?
[03:21:21] <The_8472> so it's possible that they just wrote it badly
[03:21:36] <anakh> switeck: yes
[03:21:58] <anakh> i know from tpb that even a lot of normal clients behave like a fucking pain in the ass
[03:21:59] <Switeck> especially when hosted on >1 gbit/sec (combined) bandwidth access
[03:22:10] <Switeck> BitComet and clones are junk
[03:22:14] <TheSHAD0W> I can't see how that sort of behavior would occur assuming they used standard tcp libraries - and I can't see why they would've reimplemented it.
[03:22:17] <Switeck> retry times on ips is nuts
[03:22:23] <anakh> ya just a bit messed up retry/timeout logic and boom
[03:22:40] *** DHE has joined #bittorrent
[03:23:04] <anakh> ive screwed up in ways similar to that (none related to BT though :P)
[03:23:33] <anakh> mostly with udp or other connectionless protos though
[03:24:02] <anakh> but i _can_ see it happening if they have tons of servers doing async io with high fd limits etc
[03:24:57] <anakh> back when i was a ddos kid one of the attacks i implemented was a nonblocking connect() thing ... under win95
[03:25:01] <Switeck> even assuming it's tracker updates/scrapes...that's nuts
[03:25:08] <anakh> and it was very much effective
[03:25:22] <The_8472> btw... if anyone needs to be shoveled into the developers group on the bt.org forums, i can do tha
[03:25:24] <The_8472> t
[03:25:40] <anakh> just lemme get my user activated
[03:25:44] <anakh> username Choy Kloun
[03:25:53] <Switeck> How effective is a DHT based attack against someone also on the swarm that has DHT disabled?
[03:26:09] <Switeck> swarm = same torrent swarm
[03:45:01] <anakh> not directly unless you like cause the clients to crash / DoS / etc ..
[03:47:47] <Switeck> I was thinking more along the lines of giving them junk ips that don't exist or hostile ips
[03:55:46] *** bittwist has joined #bittorrent
[03:59:31] <anakh> aint getting any junk ip addrs via dht with dht disabled :>
[04:00:53] *** The_8472 has quit IRC
[04:01:14] *** wadim has joined #bittorrent
[04:01:16] *** wadim is now known as The_8472
[04:01:43] <The_8472> DHT as it is right now does more good than harm. i don't see why you would want to disable it
[04:02:21] <K`Tetch> because your torrents will leak onto it, and people will track your ass back using it and you can also be classes as a terrorist organisation ZOMG!
[04:02:47] <The_8472> yeah, as if that couldn't happen with regular trackers.
[04:03:25] <The_8472> and i'm not a terrorist, i'm a part of the resistance against the corrupt... brbfbi
[04:07:37] <K`Tetch> yeah, i know, i'm just giving you the usually parroted reasons
[04:08:09] <K`Tetch> by people who have more keys on their keyboard, than braincells in their head
[04:08:14] <Switeck> I can upload the torrent faster with DHT disabled
[04:11:04] <Switeck> reasons I've heard for disabling DHT:  My router doesn't respond well to DHT traffic or my software firewall goes nuts
[04:11:32] *** anakh has quit IRC
[04:12:24] <The_8472> well the router thingy is true if you have a crappy router
[04:12:33] <The_8472> but software firewalls can be easily replaced
[04:12:40] <The_8472> or thrown out of the window(s)
[04:13:06] <DHE> but that subverts the usual reason for even having a router
[04:13:23] <DHE> now, openwrt with careful use of the NOTRACK rule.. hmm...
[04:13:46] <The_8472> ah yes... notrack *sigh*
[04:13:53] <The_8472> i should hammer that into dd-wrt at some point
[04:13:55] <DHE> wait, is that even compatible with port forwarding?
[04:14:05] <The_8472> i hope
[04:14:20] <DHE> can I say "forward this PACKET" (but not this stream, since it won't be locking onto it)
[04:14:25] <DHE> I'll try it when I get home
[04:14:54] <The_8472> i mean... packet based snat+dnat rules mirroring each other should be stateless
[04:15:06] <The_8472> since you forward all traffic to that particular port
[04:15:47] <DHE> my worry is that if the NOTRACK rule is matched, the packet might bypass the `nat` table entirely. a little experimentation can quickly determine if this is the case, but still.
[04:16:01] <The_8472> maybe you could do it with one of the mangle filters instead of the nat table
[04:18:04] <DHE> all the default mangle rules are things like TOS fields and packet marks. no IP rewrites...
[04:19:15] <DHE> and DNAT won't work outside the nat table
[04:19:23] *** bbelt16ag has quit IRC
[04:19:29] <The_8472> apparently you can create stateless nat with the ip route command
[04:20:09] <The_8472> question is if you can do it on a port-by-port basis
[04:20:37] <DHE> not necessarily, but the ip route command can do selective operations by marks, which can be done with iptables
[04:22:00] <DHE> still it looks promising
[04:22:39] <The_8472> now the question is if the gimped kernel on the routers can do that too
[04:22:54] <DHE> if the Advanced Routing feature is enabled, it should be enough.
[04:23:22] <DHE> I think it's needed for any of the advanced ip tool features, like ip rule
[04:23:47] <The_8472> well, then mark + ip route nat might to the trick
[04:24:08] <The_8472> + notrack
[04:24:55] <DHE> let's see if we can't make it work first...
[04:25:45] <The_8472> hrrm... speaking of which... i have to add a notrack rule for 6to4 too
[04:25:58] <The_8472> it fries conntrack, since it doesn't know how to real with it
[04:26:10] <DHE> *Deal ?
[04:26:44] <The_8472> well... 6to4 is a raw ipv6-over-ipv4 encapsulation. protocol 41 or so. so the usual TCP or UDP-specific rules don't apply
[04:26:50] <The_8472> it just tracks it as generic connection
[04:27:55] <DHE> and each packet is a unique connection? or each host?
[04:28:38] <The_8472> no, but ipv6 dht via 6to4 means it'll track the "connections"
[04:28:42] *** ChoyKloun has joined #bittorrent
[04:28:43] <The_8472> which are very shortlived
[04:28:50] <ChoyKloun> hm maybe i should use this nick more :P
[04:28:51] <The_8472> but it doesn't know how shortlived they're supposed to be
[04:29:07] <The_8472> so... conntrack overflow
[04:29:14] <DHE> okay, yeah I can see that wreaking havoc
[04:29:53] <The_8472> just like the normal DHT leads to conntrack overflow if the UDP pseudo-connection lifetimes are too high
[04:30:01] <The_8472> *configured too high
[04:30:30] <DHE> and if it doesn't know, it probably uses a longer lifetime field, like a TCP timeout. Oh joy.
[04:31:03] <The_8472> there is some generic_timeout, but messing with it seems to have other sideeffects
[04:31:29] <DHE> like ICMP?
[04:31:48] <The_8472> not sure, it caused problems at least when i lowered it ^^
[04:35:28] <ChoyKloun> gargh.. i shouldnt be allowed to code sometimes... wellwell, stupid bug, now my dht announce works :p
[04:39:24] <ChoyKloun> something that _really_ annoys me are all dht nodes that cant/wont/etc respond .. someone should add some simple NAT detection so they dont end up in _my_ back yard :)
[04:41:33] <The_8472> normally NATed nodes should fall out of a well-maintained routing table quite fast
[04:41:41] *** K`Tetch has quit IRC
[04:43:01] <ChoyKloun> they still annoy me when bootstrapping :p
[04:48:17] *** K`Tetch has joined #bittorrent
[04:48:28] <ChoyKloun> and they certainly dont fall out of _other_ peoples routing tables as i keep seeing them
[04:49:09] *** KyleK_ has joined #bittorrent
[04:50:27] <Switeck> why would DHT nodes be trying to connect to a peer/seed on anything other than their listening port?
[04:51:04] <Switeck> [16:31]	<spikeysnack>	My (upstream) firewall is rejecting all kinds of DHT udp packets on different ports forwarding only my one BT port which works correctly. DHT works on some torrents fine. Am I missing out or are other clients just spewing out on arbitrary ports?
[04:51:49] <ChoyKloun> are you sure your NAT box is keeping the udp source port propery
[04:51:53] <ChoyKloun> properly*
[04:52:15] <ChoyKloun> and not assigning new ones dynamically for some/all of the packets
[04:54:20] *** K`Tetch has quit IRC
[04:59:29] <Switeck> I don't know what spikeysnack had
[05:00:10] <Switeck> but if this was even an occasional thing, I'd expect BT Devs to have seen it.
[05:00:26] *** K`Tetch has joined #bittorrent
[05:02:05] <ChoyKloun> my bet is on your NAT box reassigning ports dynamically
[05:02:12] <ChoyKloun> might have run out of state or something
[05:02:30] <ChoyKloun> as most nat devices keep the src port when possible
[05:02:31] <Switeck> my NAT box?
[05:02:33] <Switeck> no
[05:02:38] <Switeck> it's not me
[05:03:12] <Switeck> I'm reporting something I heard second-hand
[05:03:13] <ChoyKloun> replace 'your' with 'spikeysnack' then :P
[05:03:34] <Switeck> and I didn't get much of anything for details
[05:03:54] <Switeck> [16:42]	<spikeysnack>	Well my tcp stateful packet setup works fine and all and I just would like to stop the banging against my firewall. What if I redirected the udp to my BT port? would that be bad?
[05:04:15] <Switeck> [16:46]	<spikeysnack>	I can send on any upper port and establish stateful connections on both tcp and udp outgoing. I would refusre to open up all upper udp ports to my windows machine inside.
[05:04:26] <Switeck> [16:47]	<Switeck>	are you tracking incoming ips' ports or the ports they're trying to connect to on your end?
[05:04:34] <Switeck> [16:50]	<spikeysnack>	No I track on outgoing connections and block attempts on weird ports. I accept no traffic from brazil, korea or china. I use ipfilter.dat and peerblock. Firewall is OpenBSD. IP is charter.net cable modem.
[05:04:55] <ChoyKloun> he should check the outgoing traffic from the fw and see if the src port always stays the same
[05:05:16] <Switeck> that was over 4 hours ago, he's long gone :(
[05:05:18] <The_8472> uhm, not if his router is doing the mangling
[05:06:17] <ChoyKloun> setup a box as a bridge inbetween or whatever :p
[05:06:18] <Switeck> I don't know what's going on with him and cannot ask any more questions as the conversation took place over 4 hours ago
[05:07:03] <ChoyKloun> i once had to switch ports on a vpn that ran over udp because i kept getting random game traffic destined for some other person on the same connection :p
[05:15:46] *** h2ro has quit IRC
[05:20:00] <ChoyKloun> hm how does my signature on the forum appear to others ? http://forum.bittorrent.org/profile.php?id=90637
[05:20:08] <ChoyKloun> how many ppl actually have the proper font installed
[05:25:37] *** KyleK_ has quit IRC
[05:25:59] <ChoyKloun> it means appx 'you dont like p2p? go to hell!' but the exact level of rudeness isnt possible to reproduce in english :p
[05:38:33] <The_8472> characters are very small ^^
[05:39:04] <ChoyKloun> but you see the actual characters if you increase font size ?
[05:39:26] <The_8472> and the letters remind me of some very old german printing fonts
[05:39:29] <The_8472> yes
[05:39:54] <The_8472> ????????????????????????????!
[05:39:59] <The_8472> hrrrm
[05:40:17] <The_8472> i don't even know what language that is
[05:40:20] <ChoyKloun> i actually think many recent distros etc do ship with khmer fonts
[05:40:45] <The_8472> i see
[05:41:37] <The_8472> well, windows 7 got pretty decent unicode coverage
[05:42:04] <ChoyKloun> worlds longest alphabet.. kinda fun to type on a keyboard :)
[05:45:42] <ChoyKloun> i kinda like DHT btw
[05:45:46] <ChoyKloun> its refreshingly simple
[05:46:10] <ChoyKloun> compared to the self-organizing systems ive played with / designed in the past
[05:46:13] <The_8472> the protocol is simple, the algorithms how to maintain your routing table properly and do things efficiently are sometimes a bit... subtle
[05:46:35] <ChoyKloun> heh yeah
[05:47:31] <ChoyKloun> i kinda liked my self-organizing routing algorithm/protocol where each node always knew where to forward a message to reach the target node
[05:47:53] <The_8472> http://infinite-source.de/az/whitepapers/kademlia_optimized.pdf <- must-read if you want to implement a proper DHT node
[05:48:05] <ChoyKloun> different system though.. amore of a mesh network than a distributed tracker
[05:48:10] <The_8472> BEP5 is insufficient
[05:49:54] <ChoyKloun> im not a big fan of reading first .. thats the last resort when my code blows up :)
[05:50:47] <The_8472> well, there is a big difference between a working and a proper implementation when it comes to DHT nodes
[05:51:06] <The_8472> it "works" usually means it's inefficient and thus you're burdening others
[05:51:22] <The_8472> in p2p environments you have to think of your impact on the whole system
[05:51:30] <ChoyKloun> hehe yeah
[05:53:40] *** ProperNoun has joined #bittorrent
[06:05:36] <Switeck> in a more bite-sized chunk, just think about how the other guy would see you.
[06:08:57] *** burris has quit IRC
[06:09:00] *** burris has joined #bittorrent
[06:30:57] *** burris has quit IRC
[06:31:06] *** burris has joined #bittorrent
[06:46:14] *** MassaRoddel has quit IRC
[06:46:32] *** chelz has quit IRC
[06:47:16] *** bittwist is now known as bitty
[06:47:26] *** bitty is now known as bittwist
[06:49:23] *** init0 has joined #bittorrent
[07:03:32] *** init0_ has quit IRC
[07:19:13] *** Switeck has quit IRC
[07:44:22] *** bbelt16ag has joined #bittorrent
[08:33:30] *** nGTHK has joined #bittorrent
[08:34:27] *** GTHK has quit IRC
[09:01:58] *** GTHK has joined #bittorrent
[09:07:37] *** goussx has quit IRC
[09:15:00] *** goussx has joined #bittorrent
[09:15:21] *** MassaRoddel has joined #bittorrent
[09:25:11] *** GTHK has quit IRC
[09:33:53] *** nGTHK has quit IRC
[09:34:17] *** goussx has quit IRC
[10:11:50] *** MassaRoddel has quit IRC
[10:12:39] *** PN has joined #bittorrent
[10:15:05] *** MassaRoddel has joined #bittorrent
[10:25:20] *** ProperNoun has quit IRC
[10:38:10] *** rrr_ has quit IRC
[11:22:51] *** Miller` has quit IRC
[12:16:25] *** Miller` has joined #bittorrent
[12:43:32] *** razvand has joined #bittorrent
[13:46:40] *** Miller` has quit IRC
[14:08:04] *** KyleK_ has joined #bittorrent
[14:33:54] <ChoyKloun> argh
[14:33:56] <ChoyKloun> i hate bencode
[14:33:58] <ChoyKloun> hate hate hate
[14:34:12] <ChoyKloun> i should write a proposal of switching bt to esbuf_serialize :p
[14:35:48] <The_8472> a serious case of NIH?
[14:36:52] <ChoyKloun> the knights who say NIH! ?
[14:37:03] <The_8472> not invented here
[14:37:28] <The_8472> http://en.wikipedia.org/wiki/Not_invented_here
[14:37:30] <ChoyKloun> more like IAP
[14:37:36] <ChoyKloun> It's A PITA
[14:37:37] <ChoyKloun> :P
[14:37:46] <DHE> Invent Another Protocol?
[14:38:02] <The_8472> you're the first dev to complain about it
[14:38:05] <ChoyKloun> wasnt that what i just said i felt like doing :PP
[14:38:07] <DHE> and I still say bencode was pretty easy to do
[14:38:47] <DHE> wrote my own in a langauge I was just learning, it worked with minimal kinks, and was faster than what was already available for the same language.
[14:38:49] <DHE> win win for me
[14:38:58] <DHE> surely it can't be that hard
[14:39:08] <ChoyKloun> it's just so much more work to implement in C than other marshalling
[14:39:25] <ChoyKloun> nah im just fixing a detail in my bencode parser and it reminded me of my burning hate for it :p
[14:39:42] <The_8472> well, use a C library with dynamic data structures and you're golden
[14:40:06] <ChoyKloun> i do, i use esbuf
[14:40:15] <The_8472> which i've never heard of
[14:40:23] <ChoyKloun> http://estoykh.com/opensource/ :)
[14:40:59] <The_8472> which i haven't heard of either
[14:41:00] <ChoyKloun> latest ver isnt up yet though
[14:41:27] <ChoyKloun> ive used it in every major piece of code ive written since 2003
[14:41:58] <The_8472> let me guess... written by... you?
[14:42:03] <ChoyKloun> bingo :)
[14:42:18] <ChoyKloun> i guess i do suffer from a bit of NIH as you can see in that directory that i've also written my own XML lib .. :)
[14:42:37] <ChoyKloun> although main reason for that was that i needed it to use esbuf
[14:42:38] <DHE> despite several being available already for years
[14:43:49] <ChoyKloun> well we can all share information with gopher and ftp
[14:43:53] <ChoyKloun> so why work on bittorrent
[14:43:54] <ChoyKloun> :P
[14:44:20] <The_8472> because bittorrent offers features that neither gopher nor ftp do
[14:44:39] <ChoyKloun> and esbuf offers features other buffer/mgmt libs dont
[14:44:39] <ChoyKloun> :p
[14:44:49] <The_8472> for the endusers? i do not think so
[14:45:18] <ChoyKloun> :-)
[14:45:33] <mpl> try to share as much data with as many ppl as efficiently with ftp...
[14:45:42] <ChoyKloun> yeah exactly
[14:45:44] <ChoyKloun> QED :)
[14:45:44] * DHE fires up an FTP server on his DSL line
[14:45:54] * DHE has quit (Ping timeout)
[14:45:58] <The_8472> maybe there is a shortcoming in esbuf if it leads you to complain about implementing bencoding
[14:46:05] <ChoyKloun> well ftp actually works better than bt on my connection
[14:46:07] <The_8472> since other C-client devs havn't complained about it
[14:46:08] <ChoyKloun> but then im on 256/128 heh
[14:46:18] <ChoyKloun> maybe im just better at complaining? :)
[14:46:24] <DHE> thou needst QOS
[14:46:39] <ChoyKloun> ya
[14:46:45] <The_8472> circumstanical evidence points in the NIH direction
[14:46:51] <ChoyKloun> havent had the time to toss out the chinese plastic currently doing the routing
[14:46:59] <ChoyKloun> i mean i have several soekris boxes sitting right here
[14:47:14] <ChoyKloun> i dont even need NAT or any crap like that
[14:47:33] <ChoyKloun> well MSS clamping is nice coz of stupid pppoe
[14:48:05] <DHE> you only have one PC?
[14:48:09] <ChoyKloun> no
[14:48:11] <ChoyKloun> i have a /24 at home
[14:48:16] <DHE> oh...
[14:48:19] <DHE> I only have a /31
[14:48:27] <ChoyKloun> check addr i irc from :p
[14:48:45] <The_8472> i have a /48
[14:48:54] <ChoyKloun> inetnum:        194.71.126.0 - 194.71.126.255
[14:48:54] <ChoyKloun> netname:        ESTOYKH-NET
[14:48:55] <ChoyKloun> descr:          Estoy Ltd. Cambodian network
[14:49:02] <ChoyKloun> i have an accute NAT allergy
[14:49:14] <ChoyKloun> and i know the guys at my isp :p
[14:49:17] <The_8472> ohh, you're talking about ipv4 :P
[14:49:26] <DHE> The_8472: oh, well, so do I in that case... but worldwide accessability is a problem. so I got a /31 to ease the ipv4 suffering.
[14:49:46] <ChoyKloun> oh i read /31 as /30 at first
[14:49:58] <ChoyKloun>  /31 wouldnt be much of a usable ipv4 prefix heh
[14:50:02] <DHE> you run a small datacenter from home or something?
[14:50:08] <ChoyKloun> nah
[14:50:21] <ChoyKloun> power here is too awful for that anyway
[14:50:27] <DHE> /31 works just fine. I don't need 4 addresses. 2 is enough.
[14:50:48] <ChoyKloun> well normally lowest+highest arent usable
[14:50:51] <ChoyKloun> although f.ex. linux can do it
[14:51:25] <DHE> oh, that. no. In that case, a /30 only has 1 usable address (always count the router). I'm treating the /31 as two /32 addresses
[14:51:52] <ChoyKloun> well /30 are normally used for exactly that
[14:51:54] <DHE> my ISP's static space has gotten a little fragmented so I asked for one of the /31 blocks
[14:51:55] <ChoyKloun> point-to-point links
[14:51:55] <ChoyKloun> :p
[14:52:03] <kjetilho> you don't really need a broadcast address
[14:52:19] <ChoyKloun> nah but many ip stacks complain loudly
[14:52:26] <DHE> ifconfig eth0 up 1.2.3.4 netmask 255.255.255.255;  route add -host 192.168.0.1 dev eth0;  route add default gw 192.168.0.1
[14:52:30] <DHE> tada!
[14:52:32] <The_8472> not on PPP links.
[14:52:36] <ChoyKloun> i recally windows until xp or so even refused to talk to ANY .255/.0 addr
[14:53:02] <ChoyKloun> ppp links arent really comparable though
[14:53:13] <ChoyKloun> you dont even need a gateway address
[14:53:41] <The_8472> you only need 1 address, which is the least wasteful way to assign a single address... soo...
[14:54:14] <DHE> yeah ipv4 is still at a premium. and buying IP blocks from the regisrars like ARIN is an administrative pain
[14:54:15] <ChoyKloun> for customers i always use a large subnet + proxyarp
[14:54:15] <The_8472> a /32 basically
[14:54:18] <ChoyKloun> and do /32 routes
[14:55:27] * The_8472 asks RIPE if i can get a /8 for my toaster. and another one for the fridge
[14:55:48] <ChoyKloun> you can buy space on the grey market :p
[14:56:07] <DHE> not with some registrars you can't...
[14:56:26] <ChoyKloun> hey i use RIPE space registered to an african company announced from asia
[14:56:37] <ChoyKloun> so thats more of a finer detail :p
[14:56:40] <ChoyKloun> btw regarding ppp
[14:56:52] <ChoyKloun> set user name="pppin" ipaddr=ippool ipmask=255.255.255.255
[14:56:52] <ChoyKloun> set user name="pppin" localipaddr=10.141.49.1 netrouting=off
[14:56:53] <DHE> but someone has to pay to renew the block...
[14:57:03] <ChoyKloun> #> show ippool
[14:57:03] <ChoyKloun>  ip base address     count
[14:57:03] <ChoyKloun>   10.141.49.32        32
[14:57:06] <ChoyKloun> dhe: no
[14:57:14] <ChoyKloun> end user doesnt pay
[14:57:22] <ChoyKloun> and in RIPE land there are no per-block charges
[14:57:28] <DHE> really...
[14:57:42] <ChoyKloun> however the local internet registry pays a yearly fee thats based on its size including # of bloks
[14:59:40] <ChoyKloun> anyway
[15:00:15] <ChoyKloun> what i SHOULD write up thats actually important for something is a "how to avoid having your DHT implementation being usable in DDoS attacks" summary
[15:01:05] <ChoyKloun> oh and i also propose a small protocol extension
[15:01:43] <The_8472> yes?
[15:01:45] <ChoyKloun> that any udp packet can be responded to with a "shut the fuck up" and then that host will be totally ignored by the client in the future
[15:02:05] <ChoyKloun> so theres an easy way to stop any unwelcome udp traffic from dht clients
[15:02:32] <The_8472> and there i thought we had ICMP for that ^^
[15:02:33] <DHE> well, the UDP layer itself does define one, but typically reserved for OS use
[15:03:27] <ChoyKloun> how many dht clients stop all udp communications because of an icmp response .. :)
[15:03:40] <The_8472> none, probably
[15:03:48] <The_8472> but then again it's also unlikely that they'll keep state for that
[15:03:52] <ChoyKloun> and im talking of adding the ip addr to an actual "ignore" lits
[15:03:53] <ChoyKloun> list
[15:04:32] <ChoyKloun> not only that port, since then there are 65535-1 ports left to get unwanted traffic on ..
[15:04:47] <DHE> and the OS takes care of the rest via said ICMP
[15:04:54] <DHE> so really
[15:05:01] <The_8472> so i could just spoof packets and kill the DHT by sending STFU-packets for all DHT nodes?
[15:05:03] <ChoyKloun> uhm it doesnt work that way in reality
[15:05:16] <ChoyKloun> have you ever been the recipient of a large amplifier udp-based ddos atack
[15:05:22] <ChoyKloun> the_8472: obviously you need some auth
[15:05:35] <ChoyKloun> that was kinda implied
[15:05:39] <The_8472> and suddenly the small extension grows ^^
[15:05:55] <DHE> no, I like the spoofing idea better
[15:05:55] <ChoyKloun> its a shame dht doesnt already have a mandatory cookie in all packets
[15:06:35] <ChoyKloun> however if clients implementing this also started using non-predictable transids that could be used
[15:07:15] <The_8472> considering that only 1-4byte transaction IDs are used...
[15:07:26] <The_8472> by most clients
[15:07:43] <ChoyKloun> easy change to add enough entropy
[15:07:54] <The_8472> true
[15:08:01] <The_8472> but you would have to mandate it for the STFU-packet
[15:08:19] <ChoyKloun> dhe: i can promise you that you can send all the icmp you want, you're still getting 3+Gbps of udp responses :p
[15:08:36] <The_8472> anyway, that DoS amplification attack you're speaking of... the best you can do is increase the packet size. you can't multiply the number of packets
[15:08:54] <The_8472> and since the DHT is lossy and nodes are possibly rate-limited...
[15:09:00] <DHE> if I'm getting 3 Gbps of DHT, I don't think sending stfu will get me anywhere either
[15:09:01] <ChoyKloun> ill see if i cant think of a way to actually get multiplication
[15:09:24] <ChoyKloun> dhe: i'd have loved to have a STFU feature in DNS when we got huge DNS amp attacks :p
[15:10:56] <ChoyKloun> would have involved less network voodoo and less being on the phone with upstreams instructing them on what ACLs to insert in their edge routers :p
[15:11:36] <ChoyKloun> the_8472: DHT also gives an attacker excellent spreading of attack sources
[15:11:53] <The_8472> ISP-side configureable firewalls for endusers would be awesome. this way a DoS couldn't congest the last mile
[15:12:02] <ChoyKloun> like i used to irc from addrs that were only reachable from .SE
[15:12:16] <ChoyKloun> could easily collect enough .se dht nodes to blast me to hell
[15:12:39] <ChoyKloun> the_8472: well if you have good isps you kinda have it, but it involes middle-of-the-night phonecalls :)
[15:13:04] <The_8472> yeah, i meant via some standardized protocol
[15:13:20] <ChoyKloun> well once we had a us backbone provider contacting us and asking whether we wanted that gig of traffic they saw filtered .. :)
[15:13:38] <The_8472> hahaha
[15:13:51] <ChoyKloun> note that we had no relationship with this provider
[15:14:00] <ChoyKloun> just hosting a bunch of ddos magnets :)
[15:20:39] <The_8472> btw, shouldn't routers raise all kinds of red flags if you're sending to millions of IPs with spoofed source addresses?
[15:21:57] <kjetilho> how do you recognise spoofing?
[15:22:21] <The_8472> that the packets from the spoofed address normally wouldn't go along that route?
[15:22:32] <kjetilho> universal egressfiltering would nice, wouldn't it
[15:22:38] <ChoyKloun> the_8472: routers route .... :)
[15:22:49] <ChoyKloun> in the core the routing is all done in hardware anyways
[15:22:50] <kjetilho> core routers have no idea what happens towards the edges
[15:22:58] <ChoyKloun> though you can keep a certain level of monitoring with f.ex. netflow
[15:23:11] <ChoyKloun> and some isps do actually detect weird traffic patterns
[15:23:39] <ChoyKloun> ya filtering spoofed packets is more or less undoable except on the edge, from customers
[15:23:39] <The_8472> i'm just saying, with a faked source + millions of different destination addresses some routes must seem fishy
[15:24:29] <The_8472> as in... "look, you're sending me traffic from prefixes you're not even announcing via BGP"
[15:26:00] <DHE> and cisco has a low-cpu-cost filter available that does pretty much that exact thing. though it's probably best used on edge devices in the wrong direction. which means more people need to use it.
[15:26:07] <ChoyKloun> asymmetric routing is very common...
[15:26:25] <The_8472> well, at the core probably, yes
[15:26:53] <The_8472> but it should raise flags on peering providers and other more leafy ASNs
[15:27:02] <DHE> So turn it on at your distribution routers facing the end devices.
[15:27:19] <ChoyKloun> not only at the core
[15:27:47] <DHE> The_8472: I'm just worried about how it would get applied in BGP scenarios. From what I read in the cisco docs, it's based on the routing table so if a route isn't selected for the main routing table, I'm worried about false positives
[15:27:56] <DHE> *main routing table
[15:28:07] <ChoyKloun> well end story is, its not doable in reall ife
[15:47:20] *** waldorf_ has joined #bittorrent
[15:54:37] <ChoyKloun> global internet routing kinda resembles strings and tincans
[15:54:41] <ChoyKloun> its a wonder it doesnt collapse more often
[15:55:01] <DHE> maybe, but the devices on the other ends tend to be smart enough
[15:55:07] <ChoyKloun> last year a single small isp in bumfucksville caused a global incident which affected like 70% of all traffic
[15:55:25] <ChoyKloun> by announcing an extremely long as-path
[15:55:54] <DHE> oh, that was an OS bug in routers though. crashing I think.
[15:56:02] <ChoyKloun> apparently some older cisco routers (amongst others) drop the bgp session when they see it
[15:56:02] <ChoyKloun> so we had core routers all over the world tearing down their bgp sessions
[15:56:22] <ChoyKloun> well another good example is when some turkish telco blocked youtube internally
[15:56:26] <ChoyKloun> but the prefix leaked
[15:56:39] <ChoyKloun> so a large part of the global users couldnt reach youtube :P
[15:57:05] <DHE> that's a matter of trusting the prefixes offered. IP spoof filtering wouldn't have done you any good there.
[15:57:05] <ChoyKloun> also last year at prq we had a bgp-based DoS attack
[15:57:09] <ChoyKloun> they couldnt bring down some site via other means
[15:57:26] <ChoyKloun> so they used a hacked isp in the us to announce a more specific prefix for the subnet the site was in
[15:57:44] <ChoyKloun> ya but its good examples of the state of global internet routing :P
[15:58:12] <DHE> so maybe we need routers to tie into IANA to validate the routes announced?
[15:58:46] <ChoyKloun> there are tools to do it
[15:58:52] <ChoyKloun> but its simply not feasable in real life
[15:59:09] <The_8472> i think they're planning on securing BGP like they did with DNSSEC
[15:59:22] <ChoyKloun> there are some proposals
[15:59:23] <ChoyKloun> but well
[15:59:30] <ChoyKloun> 'not reasable in real life' :P
[15:59:48] <DHE> dnssec is still coming, but hasn't arrived yet
[16:00:07] <DHE> I mean BIND supports it, but the internet...
[16:00:27] <The_8472> well, some ccNICs already support it, vista/7/linux support it...
[16:00:37] <The_8472> it's a start
[16:01:04] <The_8472> and it went through faster than IPv6 ^^
[16:01:21] <ChoyKloun> also the global bgp routing table is growing like mad
[16:01:29] <DHE> so what's the BGPsec protocol like? a signed statement from ARIN/RIPE/etc saying you're allowed to advertise this block?
[16:01:54] <The_8472> don't ask me, i already forgot the details
[16:02:07] <ChoyKloun> there are a coupleo f different ideas
[16:02:13] <ChoyKloun> none of which would work outside a lab
[16:02:24] <The_8472> but i think the first step was actually authenticating the propagation of routes, not the origin
[16:04:00] <DHE> but the problem with routing seems to be idiot admins making mistakes, not malicious BGP peering applications.
[16:04:25] <ChoyKloun> well i just gave you an example of malicious bgp use :)
[16:04:28] <DHE> I could hack Quagga to do evil things, but I don't think anybody actually does that. the youtube fiasco was more an administrative error
[16:05:53] <The_8472> if anyone would consistently do malicious things via BGP then they'd get de-peered or heavily filtered
[16:06:12] <The_8472> depending on how much weight each side has ^^
[16:06:43] <ChoyKloun> well then they could just sign up with / hack another isp ..
[16:07:46] <DHE> word of idiocy causing internet outages spread pretty fast.
[16:07:57] <ChoyKloun> yet they happen .. :)
[16:08:07] <ChoyKloun> but dont worry, no maliciousness needed
[16:08:14] <ChoyKloun> bgp will implode of its own weight quite soon
[16:08:20] <ChoyKloun> just watch the growth of the global bgp table ..
[16:08:28] <DHE> I am familiar with it
[16:08:34] <The_8472> then again ipv4 will explode quite soon too
[16:08:48] <ChoyKloun> well the bgp is the same in v6 ..
[16:08:49] <The_8472> and ipv6 is supposed to simplify routing...
[16:08:57] <ChoyKloun> actually almost all v6 peerign is done over v4
[16:08:58] <DHE> I'm responsible for monitoring at my ISP. I tend to know how trending is going
[16:09:13] <ChoyKloun> ya v6 will help with the table size
[16:09:30] <ChoyKloun> just hope that the tier1 actors keep tight prefix size filters
[16:09:39] <DHE> does ipv6 allow you to have a subnet size smaller than /64?
[16:09:40] <ChoyKloun> so we dont see morons deaggregating like they do in ipv4 today
[16:11:33] <The_8472> problem with BGP is that it's not a decentralized route finding algorithm... instead if keeps almost absolute knowldge of the entire graph in every vertex
[16:11:58] <ChoyKloun> ya
[16:12:26] <ChoyKloun> though you can setup your bgp in a way that you dont need the full table everywhere
[16:13:34] <The_8472> but as long as even unicast causes that much trouble multicast stays a pipe dream *sigh*
[16:15:51] <The_8472> if deployed internet-wide or at least in large regions we could cut down significantly on overall transit traffic AND on last mile upload...
[16:16:14] <ChoyKloun> all isps should have a multicast domain for p2p :)
[16:16:45] <DHE> multicast download of fedora 13 at 300 kb/sec... I think that could work
[16:17:12] <The_8472> by multicast domain you mean ... ?
[16:17:55] <ChoyKloun> i was just thinking of tracking etc
[16:18:20] <The_8472> tracking is a non-issue. we have DHT for that. i'm talking about the payload
[16:18:52] <ChoyKloun> though i do recall that there used to be some satellite service that gave you lots of warez downloads multicast-style
[16:19:36] <The_8472> that's probably via classic ASM, i'm talking about SSM
[16:23:18] *** rrr_ has joined #bittorrent
[16:25:49] <ChoyKloun> like users could queue stuff for download
[16:26:14] <ChoyKloun> but there was a lot of warez doods using it so you always got the latest stuff automagically
[16:58:08] *** PN is now known as ProperNoun
[17:00:54] *** Elrohir has joined #bittorrent
[17:02:37] *** Miller` has joined #bittorrent
[17:03:42] *** lioux has joined #bittorrent
[17:07:59] *** Elrohir has quit IRC
[17:42:27] <ChoyKloun> proposal: all clients should start including a cookie in the dht udp queries they send
[17:42:35] <ChoyKloun> and echo it in responses if present
[17:43:17] *** goussx has joined #bittorrent
[17:43:20] <The_8472> how would that be different from the transaction id?
[17:43:29] <Andrius> I'm going to eat your cookies
[17:43:35] <Andrius> and dht will crumble
[17:43:49] <ChoyKloun> its not actually
[17:43:59] <ChoyKloun> the above implied strong entropy though
[17:44:29] <The_8472> then just changing the definition of the transaction ID would be better
[17:44:32] <Andrius> It's obvious that I'm not going to eat transaction id, but cookies are fine
[17:45:00] <ChoyKloun> ya
[17:45:03] <ChoyKloun> clearly one way to do it
[17:45:34] <ChoyKloun> hm would there be any point in making an extension for encrypting the traffic
[17:46:41] <The_8472> if ISPs start to intercept and manipulate the traffic, then yes
[17:49:36] <ChoyKloun> my suggestion would be to use curve25519 key exchange
[17:49:40] <ChoyKloun> its extremey fast
[17:50:31] <ChoyKloun> and the data sent by each end is just 32 bytes
[17:55:54] *** KyleK__ has joined #bittorrent
[18:03:55] *** KyleK_ has quit IRC
[18:03:55] *** KyleK__ is now known as KyleK_
[18:16:40] *** goussx has quit IRC
[18:49:39] *** waldorf_ has quit IRC
[18:56:36] *** KyleK_ has quit IRC
[19:13:52] <SundanceKid> hello ppl. I'm looking for a good bdecoder for PHP
[19:14:09] <DHE> is mine any good?
[19:14:15] <SundanceKid> please give me a link if you know...
[19:14:25] <DHE> dehacked.2y.net/BT/  link on the bottom
[19:14:29] <ChoyKloun> oh fuck
[19:14:30] <ChoyKloun> i hate bencode
[19:14:32] <ChoyKloun> and i hate php
[19:14:47] <DWKnight> then go jump off a cliff
[19:14:49] * DHE takes a shot of the Potion of Apathy
[19:15:00] <SundanceKid> :-)
[19:15:07] * DWKnight chugs a bottle
[19:15:08] <SundanceKid> I develop my own tracker
[19:15:13] <DHE> oh, so did I
[19:15:36] <SundanceKid> lol)
[19:15:39] <SundanceKid> torrentparse :))
[19:16:27] <ChoyKloun> you would hate php too if it was your dayjob :P
[19:16:34] <ChoyKloun> well i dont actually write code most of the time
[19:16:38] <SundanceKid> I've seen it already. Bullshit :)
[19:16:39] <ChoyKloun> but im the boss of 5 php developers
[19:16:48] <SundanceKid> ChoyKloun: lol.
[19:16:54] <SundanceKid> http://phpdaemon.googlecode.com
[19:17:02] <mpl> ChoyKloun: if you're the boss, make them switchto something else ;P
[19:17:15] <SundanceKid> php rules
[19:17:17] <ChoyKloun> they only know php
[19:17:29] <ChoyKloun> well php and flash/actionscript and javascript etc
[19:17:32] *** DHE has quit IRC
[19:17:36] <ChoyKloun> but if they started to develop all websites in flash i'd fucking kill them
[19:17:54] <SundanceKid> DHE: thank you. do you know any another?
[19:20:51] *** lioux_ has joined #bittorrent
[19:21:48] <ChoyKloun> the mitigating factor is that its quite good business.. total salaries for 5 univ graduate developers + 2 office staff is less than what ONE developer would get in europe
[19:23:41] <mpl> ChoyKloun: in other words you're running a sweatshop?
[19:23:53] <ChoyKloun> well we do have air condition?
[19:24:00] <ChoyKloun> so no sweat!
[19:24:16] <ChoyKloun> but yeah business model is pretty similar
[19:24:30] <ChoyKloun> except our labour isnt unqualified/manual work
[19:25:17] <ChoyKloun> and we pay far more than what they'd earn at most other jobs
[19:25:31] <ChoyKloun> average income here is $600 ... per year
[19:27:16] <mpl> :(
[19:28:47] <ChoyKloun> you prefer developing countries to remain poor or what ?
[19:29:22] <ChoyKloun> or maybe you want to 'hire' our female staff for the job they'd do if they didnt have this (prostitution)
[19:29:46] <mpl> ChoyKloun: no, I was making the sad face about your last comment: 19:25 < ChoyKloun> average income here is $600 ... per year
[19:30:34] <mpl> ChoyKloun: so, good for you if you think you're offering better working conditions.
[19:30:34] <ChoyKloun> ah k
[19:30:51] <ChoyKloun> this isn't even the third world.. its the 4th :)
[19:30:58] <mpl> where is it?
[19:31:10] <ChoyKloun> preah reach nachakr kampuchea !
[19:31:44] <mpl> cambodia?
[19:31:57] <ChoyKloun> yep
[19:32:01] <mpl> ok
[19:36:02] <ChoyKloun> the nicest and most interesting country in the world <3
[19:36:11] <mpl> ChoyKloun: 19:29 < ChoyKloun> you prefer developing countries to remain poor or what ? <-- I bought fair trade pants and underwear from china and india yesterday ;)
[19:36:20] <ChoyKloun> fair trade is fucking crap
[19:36:48] <mpl> ChoyKloun: easy to say that, but at least it's a step better than the rest imho.
[19:37:22] <ChoyKloun> the workers and countries benefit more WITHOUT "fair trade" ..
[19:37:33] <mpl> hah, riiight.
[19:37:39] <ChoyKloun> just buy products from developing countries... any products
[19:38:44] <ChoyKloun> mpl: read the book In defense of global capitalism by Norberg
[19:38:49] <mpl> right, and the conditions will get better for workers because most of the money is not going into the pocket of bosses?
[19:38:53] <ChoyKloun> it details the problems with 'fair trade' very well
[19:39:37] <mpl> ChoyKloun: I will, thx for the reference.
[19:39:46] <ChoyKloun> you do know that FT charges quite hefty licensing fees from the workers/farmers, right? :>
[19:40:03] <mpl> yes.
[19:40:20] *** lioux has quit IRC
[19:40:56] <ChoyKloun> and uhm if you actually studied the real world you'd see people DREAMING about getting jobs in sweatshops
[19:41:17] <mpl> because they have no other choice.
[19:41:24] <mpl> it's that or die starving.
[19:41:31] <mpl> or prostitution as you said it.
[19:42:18] <ChoyKloun>  http://estoykh.com/hellonearth/dsc00241.jpg you think those kids share the worldview of western liberal do-gooders ?
[19:42:58] <mpl> ChoyKloun: which is exactly what I just said above...
[19:43:51] <ChoyKloun> oh well, read the book :) the issue is a bit too complicated for a short irc chat :p
[19:44:02] <mpl> agreed.
[19:44:35] <ChoyKloun> living in developing countries is a good eye-opener too
[19:44:54] <charles> those are your php hackers?  buy them some shirts
[19:45:01] <ChoyKloun> charles: hahah
[19:45:02] <ChoyKloun> yeah
[19:45:04] <ChoyKloun> thats our office !
[19:45:28] <ChoyKloun> it smells a bit bad and the rats and stray dogs are a bit of a nuisance
[19:45:37] <charles> is that you lying on the ground by the bike?
[19:46:11] <ChoyKloun> lying on the ground by the bike?
[19:46:21] <charles> there are three people in that picture
[19:46:31] <charles> one's resting/asleep/dead on the ground by the bike
[19:46:38] <charles> if they're dead, it's probably not you
[19:46:40] <ChoyKloun> btw... http://estoykh.com/fiber/ there are some signs of civilization here after all :)
[19:46:48] <mpl> yeah he's the boss, he's just sleeping while the other ones are working ;)
[19:46:59] <ChoyKloun> ah yeah i see it now
[19:47:15] <ChoyKloun> if he was dead i think someone would have stolen his bike and/or his clothes :)
[19:48:05] <mpl> ChoyKloun: what's the second pic, the sewage system?
[19:48:12] <TheSHAD0W> Oh god they did it again!
[19:48:14] <TheSHAD0W> http://reason.com/archives/2009/11/27/miniature-monsters-attack-amer
[19:48:18] <ChoyKloun> mpl: check all the images
[19:48:30] <ChoyKloun> the dirty tunnel is where they're installing the cables
[19:48:41] <mpl> ah, fiber installation?
[19:48:44] <ChoyKloun> ya
[19:48:47] <mpl> ok
[19:49:58] <charles> well, I'm convinced.  If there are two kids without shirts somewhere in the world, fair trade must be bullshit cooked up by libr'l do-gooders
[19:50:20] <ChoyKloun> hey someone get me a fire extinguisher quick
[19:50:28] <ChoyKloun> the strawman charles built is on fire !
[19:50:51] * charles throws the burning strawman at ChoyKloun
[19:51:39] * mpl hugs the burning strawman
[19:51:57] * ChoyKloun turns it into a Burning Man festival
[19:52:09] <mpl> ow ow ow!
[19:52:21] * ChoyKloun Front 242 - Don't Crash
[19:54:10] <charles> ah, back when f242 was still good
[20:00:43] <ChoyKloun> they did a good live show in 2007 actually
[20:01:03] <ChoyKloun> nitzer ebb (at the same event) was much better though
[20:01:29] <ChoyKloun> but the fucking power failed during the chorus of join in the chant.. was a "bit" of an anticlimax...
[20:06:41] <charles> they were never the same after Richard 23 left the first time around...
[20:06:54] <charles> Pulse was a big improvement though
[20:07:39] <ChoyKloun> pulse was a bit .. well, boring
[20:07:44] <ChoyKloun> good music but lacks a soul
[20:09:26] <charles> Tyranny was the last really strong one
[20:11:50] <ChoyKloun> weirdest music video ever lol
[20:13:24] <ChoyKloun> no EBM clubs here :( :( :(
[20:14:30] *** waldorf_ has joined #bittorrent
[20:15:00] <ChoyKloun> also my connection is too slow to stream mp3 radios reliably :( :(
[20:15:15] <ChoyKloun> can do it at the office though at off-peak hours
[20:16:51] *** razvand has quit IRC
[20:19:49] *** waldorf_ has quit IRC
[20:51:58] <SundanceKid> :)
[20:52:17] <SundanceKid> folks, please advice tell me what about order of info's items
[20:52:26] <SundanceKid> in info_hash
[20:52:53] <SundanceKid> I guess it may cause many troubles
[20:53:16] <SundanceKid> if the order will be changed
[20:53:24] <The_8472> dictionaries are sorted by binary sorting of the keys
[20:53:38] <SundanceKid> binary sorting?
[20:54:01] <The_8472> well, not unicode-compare of the characters but sorting by the bytes of the keys
[20:54:07] *** razvand has joined #bittorrent
[20:54:11] <The_8472> binary, not strings
[20:54:24] <SundanceKid> thank you very muchg
[20:54:26] <SundanceKid> *much
[20:54:39] <The_8472> the bencoding spec mentions that somewhere btw
[20:55:01] <SundanceKid> theory's wiki doesn't say it...
[20:55:02] <The_8472> oh, and lists have no well-defined order, but if you dump them into numerically indexed arrays they should retain their order
[20:55:04] <SundanceKid> AFAIK
[20:55:10] <The_8472> that should be fixed
[20:55:34] <The_8472> http://bittorrent.org/beps/bep_0003.html <- the official spec does though
[20:56:09] <The_8472> http://wiki.theory.org/BitTorrentSpecification#dictionaries <- it does too
[20:56:18] <The_8472> "Keys must be strings and appear in sorted order (sorted as raw strings, not alphanumerics). The strings should be compared using a binary comparison, not a culture-specific "natural" comparison. "
[21:27:33] *** Switeck has joined #bittorrent
[21:44:41] *** KyleK_ has joined #bittorrent
[21:58:14] *** razvand has quit IRC
[22:03:35] *** GTHK has joined #bittorrent
[23:23:51] *** Andrius has quit IRC
[23:49:22] *** KyleK_ has quit IRC





top