1 2013-03-13 00:00:04 <sudofox> Hi
 2 2013-03-13 00:00:11 <sudofox> Who wants free testnet BTC
 3 2013-03-13 00:00:19 <sudofox> I've got tons of it
 4 2013-03-13 00:00:39 <buddyrandom> this guy needs some http://tpfaucet.appspot.com/
 5 2013-03-13 00:00:43 <RoboTeddy> tjader: I see, that could make sense, thanks.
 6 2013-03-13 00:00:44 <K1773R> sudofox, donate it to tpfacute pls
 7 2013-03-13 00:01:06 <K1773R> wow thats really dry
 8 2013-03-13 00:01:13 <K1773R> did someone start selling TBTC again? *sigh*
 9 2013-03-13 00:01:35 <sudofox> @buddyrandom I'll send 'em 1000 TBTC
10 2013-03-13 00:01:39 <sudofox> Thanks
11 2013-03-13 00:01:47 <K1773R> thanks to you ;)
12 2013-03-13 00:01:53 <tjader> RoboTeddy: I have no idea if it's true or not :p
13 2013-03-13 00:02:26 <RoboTeddy> tjader: I know ;) and if it is true, it still doesn't necessarily kill the idea -- it would then depend on the length of the list of nodes each client compiled before selecting ones to connect to
14 2013-03-13 00:02:33 <buddyrandom> woot :)
15 2013-03-13 00:02:40 <RoboTeddy> I guess I should go, you know, read the code..
16 2013-03-13 00:03:26 <sudofox> Sent.
17 2013-03-13 00:03:58 <tjader> RoboTeddy: if it works as I said it would kill the idea.
18 2013-03-13 00:05:13 <RoboTeddy> tjader: well, if clients compiled a list of 20 /24s on average, and connected to 10 of them, and one of your nodes is definitely in the list, then you'd be connected to about half the time
19 2013-03-13 00:05:31 <RoboTeddy> (I have no idea what the numbers look like in actuality)
20 2013-03-13 00:05:37 <MKCoin> <K1773R> did someone start selling TBTC again? *sigh*
21 2013-03-13 00:05:49 <MKCoin> sell them for what? I made 50 tbtc earlier /by accident/
22 2013-03-13 00:06:07 <tjader> RoboTeddy: but the amound of nodes inside a single /24 wouldn't matter for the likelyhood of it connecting to the /24
23 2013-03-13 00:06:12 <K1773R> ppl tryd to trade TBTC for BTC. there was even once an exchange for it!
24 2013-03-13 00:06:16 <MKCoin> lawl
25 2013-03-13 00:06:39 <RoboTeddy> tjader: right, but maybe the client is only choosing between 20 /24s, and you're practically guaranteed to always be one of the /24s they're choosing between
26 2013-03-13 00:06:40 <K1773R> maybe some stupid altcoincollector who thingks TBTC is worth something
27 2013-03-13 00:06:53 <K1773R> -g
28 2013-03-13 00:07:08 <tjader> RoboTeddy: you're a few orders of magnitude off.
29 2013-03-13 00:07:37 <tjader> there are 16 million /24s.
30 2013-03-13 00:07:52 <sudofox> @K1773R I google tpfacute, didn't find anything
31 2013-03-13 00:08:08 <K1773R> sudofox, the same as buddyrandom did link u
32 2013-03-13 00:08:23 <RoboTeddy> tjader: yes, but the node might only select from the subset of /24s that have public nodes in them, and the subset of those that the node has happened to actually learn about
33 2013-03-13 00:08:38 <K1773R> there are 2 faucets for TBTC, let me search the second
34 2013-03-13 00:08:42 <phantomcircuit> RoboTeddy, there are ~500k ip:port combinations being passed around the network by addr messages
35 2013-03-13 00:08:52 <phantomcircuit> RoboTeddy, there are ~4k nodes which accept connections
36 2013-03-13 00:08:55 <tjader> I'm not very familiar with the peer discovery protocol, but 20 seems quite low.
37 2013-03-13 00:09:05 <phantomcircuit> you're going to need a lot more than 20 ipv4 addresses
38 2013-03-13 00:09:15 <sudofox> Ohhh tpfaucet makes more sense
39 2013-03-13 00:09:16 <RoboTeddy> phantomcircuit: ok, true story! thanks for the numbers
40 2013-03-13 00:09:29 <MKCoin> http://testnet.mojocoin.com/
41 2013-03-13 00:09:33 <phantomcircuit> and you're probably going to have to actually run bitcoin nodes on the ip addresses for anybody to pass your address around as valid
42 2013-03-13 00:09:34 <phantomcircuit> so
43 2013-03-13 00:10:17 <K1773R> sudofox: here is the second -> http://testnet.mojocoin.com/
44 2013-03-13 00:10:21 <phantomcircuit> you're talking more like 500 ipv4 addresses and actually to get good coverage probably more like 1000
45 2013-03-13 00:10:33 <K1773R> argl MKCoin
46 2013-03-13 00:10:33 <phantomcircuit> and they would all have to be actual nodes that actually returned useful results
47 2013-03-13 00:10:34 <phantomcircuit> so
48 2013-03-13 00:10:37 <MKCoin> X)
49 2013-03-13 00:10:37 <phantomcircuit> good luck with that
50 2013-03-13 00:10:46 <sudofox> They have >3000 TBTC, I think they're fine
51 2013-03-13 00:11:05 <sudofox> tpfaucet has 0.95004663 TestNet Bitcoins
52 2013-03-13 00:11:22 <sudofox> I sent 1000 TBTC
53 2013-03-13 00:12:40 <RoboTeddy> phantomcircuit: you could run one actual node, with 500-1000 "proxy" nodes that each have their own IP. seems doable for about 20 bucks an hour
54 2013-03-13 00:12:50 <phantomcircuit> RoboTeddy, ahaha no you couldn't
55 2013-03-13 00:12:58 <phantomcircuit> your actual node would fucking melt
56 2013-03-13 00:13:13 <RoboTeddy> phantomcircuit: how much does a node actually have to do to have its address be passed around?
57 2013-03-13 00:13:16 <K1773R> RoboTeddy: you cant, most proxy dont allow bind tunnels (expect some few SOCKS5 proxys)
58 2013-03-13 00:13:17 <phantomcircuit> you'd need at least 10 actual nodes
59 2013-03-13 00:13:39 <phantomcircuit> K1773R, shh that's the wrong criticism
60 2013-03-13 00:13:44 <K1773R> LOL
61 2013-03-13 00:13:51 <phantomcircuit> K1773R, haproxy
62 2013-03-13 00:14:32 <phantomcircuit> RoboTeddy, it's not simple but the gist of it is you need a node which will consistently and reliably handle the version/verack over a long period of time
63 2013-03-13 00:14:33 <K1773R> u forgot he dosnt know how the IP nor the TCP stack works, so haproxy is way out of range
64 2013-03-13 00:14:44 <phantomcircuit> and which doesn't trigger and of the DoS conditionals all over the codebase
65 2013-03-13 00:15:06 <phantomcircuit> the main problem is that 1000 ips would cost like
66 2013-03-13 00:15:06 <RoboTeddy> phantomcircuit: version/verack is pretty simple. do you have to actively gossip transactions and block information too?
67 2013-03-13 00:15:09 <phantomcircuit> $2000/month
68 2013-03-13 00:15:17 <phantomcircuit> and the information you get is basically useless
69 2013-03-13 00:15:40 <RoboTeddy> K1773R: errm I use haproxy :)
70 2013-03-13 00:15:45 <phantomcircuit> RoboTeddy, you'd have to check all of the DoS conditionals im pretty sure there are some that basically count against you for not responding to certain things
71 2013-03-13 00:15:56 <RoboTeddy> phantomcircuit: yeah, the IPs do seem like the expensive part
72 2013-03-13 00:16:09 <RoboTeddy> phantomcircuit: thanks for the lesson
73 2013-03-13 00:16:12 <K1773R> if (networkmagic->req() == GETVERSION) { printf "DoS detected, crash imminent"; } ROFL
74 2013-03-13 00:16:26 <phantomcircuit> yeah there's stuff like that all over the place
75 2013-03-13 00:18:20 <K1773R> RoboTeddy, i told you already way ealier its WAY to expensive
76 2013-03-13 00:18:28 <K1773R> why didnt u want to believe me huh?
77 2013-03-13 00:18:36 <phantomcircuit> numbers
78 2013-03-13 00:18:37 <RoboTeddy> cause I like reasons for things :)
79 2013-03-13 00:18:39 <phantomcircuit> you needed numbers
80 2013-03-13 00:19:18 <K1773R> i told you wont have somany IPs nor servers to do it...
81 2013-03-13 00:19:27 <K1773R> well, good to see it sorted out
82 2013-03-13 00:19:57 <RoboTeddy> K1773R: right, but at that point I was under the assumption you could run 60k servers off each port of a single ip
83 2013-03-13 00:20:26 <RoboTeddy> and holding onto TCP connections is super easy these days, and I wasn't aware you might need to actually do work to avoid the DOS protections
84 2013-03-13 00:25:08 <K1773R> of course you CAN
85 2013-03-13 00:25:13 <K1773R> but no nodes wil connect to them :P
86 2013-03-13 00:25:28 <K1773R> anyway, gtg. cya later around
87 2013-03-13 00:44:52 <[Tycho]> Anyone here ?
88 2013-03-13 00:45:06 <deadweasel> yo.
89 2013-03-13 00:45:16 <deadweasel> I only have a question though.
90 2013-03-13 00:46:02 <randy-waterhouse> quiet huh?
91 2013-03-13 00:46:15 <randy-waterhouse> almost too quiet
92 2013-03-13 00:46:35 <randy-waterhouse> hey [Tycho]
93 2013-03-13 00:46:53 <[Tycho]> Yes.
94 2013-03-13 00:48:06 <randy-waterhouse> how's deepbit doing these days?
95 2013-03-13 00:48:41 <[Tycho]> Not very fine because of that big reorg.