1 2013-03-13 00:00:30 <sipa> lianj: and an infinite improbability drive
2 2013-03-13 00:00:44 Boydy has joined
3 2013-03-13 00:00:46 <K1773R> gavinandresen, http://unicornify.appspot.com/whats-this
4 2013-03-13 00:01:01 <K1773R> saw that on SE meta
5 2013-03-13 00:01:02 joe_k1 has joined
6 2013-03-13 00:01:19 Tron- has joined
7 2013-03-13 00:01:20 bangarang has quit (Ping timeout: 260 seconds)
8 2013-03-13 00:01:59 keystroke has joined
9 2013-03-13 00:01:59 keystroke has quit (Changing host)
10 2013-03-13 00:01:59 keystroke has joined
11 2013-03-13 00:02:13 <ProfMac> is there a #bitcoin-<whatever> for currency trading software development
12 2013-03-13 00:02:32 <ProfMac> or how do I find all the #bitcoin- related discussions?
13 2013-03-13 00:02:32 drizzt_ has quit (Remote host closed the connection)
14 2013-03-13 00:03:11 <K1773R> ProfMac: /list
15 2013-03-13 00:03:25 <RoboTeddy> K1773R: that is an interesting idea. if we'd have bitcoinds connected to every node in the network, could we have estimated the hash power associated with each branch of the fork more effectively than just waiting to see block frequency?
16 2013-03-13 00:03:26 <ProfMac> Thats it! The whole explaination!! Yesterday was Douglas Adams' birthday, and the event was improbable !
17 2013-03-13 00:03:33 i2pRelay has quit (Remote host closed the connection)
18 2013-03-13 00:03:43 <ProfMac> (it was also 42 years since my firstborn was born)
19 2013-03-13 00:04:02 joe_k has quit (Ping timeout: 264 seconds)
20 2013-03-13 00:04:02 i2pRelay has joined
21 2013-03-13 00:04:10 joe_k1 is now known as joe_k
22 2013-03-13 00:04:35 <tjader> ProfMac: "/list -re bitcoin"
23 2013-03-13 00:04:40 <K1773R> RoboTeddy, no need to be connected to all nodes for that (which wont happen anyway since not everyone node has an reachable port)
24 2013-03-13 00:04:41 <Boydy> ProfMac: congrats
25 2013-03-13 00:05:19 showard has joined
26 2013-03-13 00:05:49 rincewind has joined
27 2013-03-13 00:06:29 gotnate has joined
28 2013-03-13 00:07:47 lantastic has quit (Read error: Connection reset by peer)
29 2013-03-13 00:07:52 rincewind has left ()
30 2013-03-13 00:08:48 Tritonio has left ()
31 2013-03-13 00:09:11 ben-abuya has left ()
32 2013-03-13 00:09:12 Cryptoman has joined
33 2013-03-13 00:09:12 <RoboTeddy> K1773R: is there some way to induce nodes without open ports to connect to you? or make it likely to happen? (e.g. by running a ton of nodes yourself)
34 2013-03-13 00:09:33 Shrugged has quit (Quit: Shrugged)
35 2013-03-13 00:10:37 rbecker is now known as RBecker
36 2013-03-13 00:10:53 <K1773R> no, if there is no port to listen on ("no open port") then they cant connect to you
37 2013-03-13 00:10:59 TD has quit (Quit: TD)
38 2013-03-13 00:11:30 FredEE has joined
39 2013-03-13 00:11:34 i2pRelay has quit (Remote host closed the connection)
40 2013-03-13 00:11:38 <K1773R> in case ur interested about these network things, read a documentation about TCP and NAT
41 2013-03-13 00:11:41 diatonic has left ()
42 2013-03-13 00:12:06 i2pRelay has joined
43 2013-03-13 00:12:07 <RoboTeddy> K1773R: I mean, if you ran a ton of bitcoin nodes that /were/ listening, would bitcoin nodes that /aren't/ listening be statistically likely to connect to you?
44 2013-03-13 00:12:44 hsmiths7 has joined
45 2013-03-13 00:12:52 <K1773R> u dont have to run alot of nodes, 1 nodes with an open port is enough
46 2013-03-13 00:13:06 <K1773R> wait some time and the connection count increases ;)
47 2013-03-13 00:13:14 bristol has left ()
48 2013-03-13 00:13:17 CaptainBlaze has quit (Quit: CaptainBlaze)
49 2013-03-13 00:13:41 <RoboTeddy> K1773R: that's enough for normal usage, but I'm wondering about mapping all bitcoin nodes, including ones that aren't listening publicly. if you ran 200k nodes that accepted incoming connections, chances are that nearly all nodes would end up connected to at least one of your nodes
50 2013-03-13 00:14:18 <K1773R> you can only map all nodes if your the only guy who has a reachable port
51 2013-03-13 00:14:34 <K1773R> but this will never happen, that would mean bitcoin is dead
52 2013-03-13 00:15:24 <RoboTeddy> K1773R: what if 50% of nodes with reachable ports are yours? that would mean that even if nodes that aren't listening publicly connected to only 1 public peer on average, you'd capture half of them
53 2013-03-13 00:15:27 <K1773R> if you want to try such things use the altcoins CLC/GRC (i run a public node), there are only max 7 cons for me, if you dont have an open port sometimes (if my node is down) u wont get any cons at all
54 2013-03-13 00:15:30 showard has quit (Quit: Ex-Chat)
55 2013-03-13 00:15:43 <K1773R> test with these altcoins to see how its behaving ;)
56 2013-03-13 00:16:10 <RoboTeddy> would this be an acceptable use of testnet? or nah
57 2013-03-13 00:16:14 <K1773R> so, and what u wanna do then? whats the goal of it?
58 2013-03-13 00:16:51 <K1773R> testnet has enough public nodes, these altcoins dont (sometimes)
59 2013-03-13 00:16:53 <RoboTeddy> K1773R: well, you could map all nodes! it'd be cool! and you could perhaps learn how many nodes are following a given fork, to get a better idea of hash power associated with a given fork
60 2013-03-13 00:17:37 <K1773R> you would have to publish somany servers with bitcoind running that you cant afford to do it
61 2013-03-13 00:18:25 <TheSeven> hm, why does blockchain.info's mining pool distribution page show two "unknown" portions?
62 2013-03-13 00:18:34 paraipan has joined
63 2013-03-13 00:18:36 <RoboTeddy> K1773R: hmm, is the gossip protocol that expensive? you wouldn't have to mine or anything, or necessarily even verify
64 2013-03-13 00:18:40 <ProfMac> can't you get a lot of that from grep "version message" < debug.log
65 2013-03-13 00:18:45 <K1773R> blockchains graphs are broken...
66 2013-03-13 00:19:05 rafsoaken has quit (Quit: rafsoaken)
67 2013-03-13 00:19:07 <RoboTeddy> K1773R: also, are these kinds of questions acceptable for #bitcoin-dev, or should I stfu?
68 2013-03-13 00:19:09 <K1773R> ProfMac: yes, on connect they tell u the version
69 2013-03-13 00:19:30 <ProfMac> and debug.log tells all the IPs you have connected to...
70 2013-03-13 00:19:36 i2pRelay has quit (Remote host closed the connection)
71 2013-03-13 00:19:48 eennaam has left ()
72 2013-03-13 00:20:07 i2pRelay has joined
73 2013-03-13 00:20:13 <Vinnie_win> Is it true that Bitcoin will never have a fixed specification and the only client will consist of the C++ hairball?
74 2013-03-13 00:20:21 <K1773R> RoboTeddy: because you would have to host atleast somany nodes that are currently publicly run (ie reachable port). if you CAN afford somany servers then u could invest ur money better or ask the big pools/projects to gather the data u need for you
75 2013-03-13 00:21:07 <K1773R> Vinnie_win, there is a Java and C daemon/SDK already (maybe even more)
76 2013-03-13 00:21:30 shlm has quit (Ping timeout: 252 seconds)
77 2013-03-13 00:21:32 <Vinnie_win> K1773R: No, the guy who started the C implementation gave up and the Java implementation is not a full node
78 2013-03-13 00:21:42 <Luke-Jr> there's Haskell :P
79 2013-03-13 00:21:47 jine has quit (Ping timeout: 245 seconds)
80 2013-03-13 00:21:56 jine has joined
81 2013-03-13 00:21:56 <RoboTeddy> K1773R: holding onto a tcp connection is cheap these days; one system can hold like 200k connections if you tune it a bit. any idea how much work you'd have to do per-connection? (e.g. packets per second, or computation requirements) -- remember that these could be "fake" nodes just for recon, that don't actually do much processing
82 2013-03-13 00:22:06 rich__ has quit (Ping timeout: 272 seconds)
83 2013-03-13 00:22:11 shlm has joined
84 2013-03-13 00:22:24 <K1773R> Luke-Jr: really? nice!
85 2013-03-13 00:22:35 <Luke-Jr> and pynode
86 2013-03-13 00:22:44 <Luke-Jr> and bitsofproof is a full Java impl I think
87 2013-03-13 00:22:54 <K1773R> Vinnie_win, https://bitcointalk.org/index.php?topic=122013.0 <-- Java + JNI, https://bitcointalk.org/index.php?topic=123488.0 <-- C
88 2013-03-13 00:23:01 Shrugged has joined
89 2013-03-13 00:24:00 <sipa> K1773R: purecoin, by roconnor
90 2013-03-13 00:24:01 <ProfMac> "/list" and "/list -ie bitcoin" give me a list of plugins. I don't grok how this is a list of other discussion threads.
91 2013-03-13 00:24:07 <Vinnie_win> K1773R: Thanks, I will read
92 2013-03-13 00:24:19 <K1773R> RoboTeddy, i dont know how many nodes are around but lets say its 1 million nodes, then you would have to have 1 million nodes (preferably everyone with a own IP) and since this wont run on 1 server its, you would have to buy thousands of servers
93 2013-03-13 00:24:25 rich__ has joined
94 2013-03-13 00:24:40 <K1773R> ProfMac, /QUOTE LIST
95 2013-03-13 00:25:09 pacpac has joined
96 2013-03-13 00:25:11 <K1773R> in your case your IRC client dosnt send the LIST command to the IRC server, it process it internally. with QOUTE you can force to send it to the IRC server
97 2013-03-13 00:25:22 <RoboTeddy> K1773R: well, there's EC2, and you might not have to leave them up for that long ;) -- I don't think there are that many public nodes, though -- https://en.bitcoin.it/wiki/Bitcoin_Map -- only 60k unique ips in last 3 hrs
98 2013-03-13 00:25:26 <lianj> K1773R: more like 5 servers
99 2013-03-13 00:25:46 <tjader> ProfMac: try /quote /list -re bitcoin
100 2013-03-13 00:25:49 theymos has quit (Remote host closed the connection)
101 2013-03-13 00:26:04 <K1773R> lianj, can i buy a server from you where u can run 1 million full bitcoin nodes?
102 2013-03-13 00:26:10 Shrugged has quit (Client Quit)
103 2013-03-13 00:26:17 <tjader> ProfMac: actually, /quote list -re bitcoin
104 2013-03-13 00:26:20 <K1773R> tjader, not with the /list
105 2013-03-13 00:26:23 <K1773R> err...
106 2013-03-13 00:26:32 <Luke-Jr> /msg alis list bitcoin*
107 2013-03-13 00:26:39 redeper has quit (Read error: Operation timed out)
108 2013-03-13 00:26:39 <Luke-Jr> err
109 2013-03-13 00:26:43 <Luke-Jr> /msg alis list #bitcoin*
110 2013-03-13 00:26:54 <K1773R> LOL somany typos today
111 2013-03-13 00:27:11 <RoboTeddy> it would cost 1.2k to put 60k micro instances up on EC2 for one hour. might be enough to map the whole network
112 2013-03-13 00:27:37 i2pRelay has quit (Remote host closed the connection)
113 2013-03-13 00:27:40 <RoboTeddy> (since you'd be 1 in every 2 public nodes)
114 2013-03-13 00:28:09 i2pRelay has joined
115 2013-03-13 00:28:40 <K1773R> one hour? you need more time for a sync and if you want to gather data you have to do this over a longer period than an hour...
116 2013-03-13 00:28:44 tyn has joined
117 2013-03-13 00:28:57 <lianj> K1773R: no, but why you need full nodes? or even nodes. 1m tcp connections feeding into a loadbalanced handful of full nodes should be enough
118 2013-03-13 00:29:28 hsmiths7 has quit (Remote host closed the connection)
119 2013-03-13 00:29:33 <RoboTeddy> the expensive part might be renting all the public ip addresses
120 2013-03-13 00:29:48 <RoboTeddy> and right: no need to sync, is there?
121 2013-03-13 00:30:08 <Luke-Jr> last night in a nutshell: http://i.imgur.com/NncJ5fI.jpg
122 2013-03-13 00:30:14 int03h has joined
123 2013-03-13 00:30:18 <ProfMac> yeah!! [19:05] Luke-Jr. That worked.
124 2013-03-13 00:30:18 <lianj> Luke-Jr: :)
125 2013-03-13 00:30:21 <K1773R> lianj, well you could a own daemon which only gets the version string, still 1M different IPs (or lets say 60k daemons per IP, the rest of the ports being reserved) its still a huge project
126 2013-03-13 00:30:31 Casimir1904 has joined
127 2013-03-13 00:30:41 <RoboTeddy> K1773R: ohh there's not a required port? that makes it way easier.
128 2013-03-13 00:30:52 forever-d has quit (Read error: Connection reset by peer)
129 2013-03-13 00:30:58 <RoboTeddy> there only seem to be 60k public nodes at a time -- not 1m. so you could do it off like a few machines with a few ips.
130 2013-03-13 00:31:13 <K1773R> the 1m is just an example
131 2013-03-13 00:31:20 <RoboTeddy> right, but in practice, it might be cheap
132 2013-03-13 00:31:34 <K1773R> in this case, do it?
133 2013-03-13 00:31:50 <RoboTeddy> let's hope I don't get too bored next weekend :D
134 2013-03-13 00:31:55 <K1773R> somuch effort for a fork detection (which was the goal at the start), thats expensive!
135 2013-03-13 00:31:56 t7 has quit (Quit: Konversation terminated!)
136 2013-03-13 00:32:01 <RoboTeddy> I also don't want to mess anything up. I would map testnet first
137 2013-03-13 00:32:22 <RoboTeddy> K1773R: well, are there any other things that you could figure out from knowing the state of every node?
138 2013-03-13 00:33:01 <K1773R> create sweet graphs, exploit bugs on all the vulnerable nodes, et
139 2013-03-13 00:33:03 <K1773R> *etc
140 2013-03-13 00:33:24 <K1773R> data itself dosnt do anything, its just information. what you do with it is part of ur imagination
141 2013-03-13 00:33:28 <lianj> doesn't blockchain.info already try and often does a good job as mapping the nodes?
142 2013-03-13 00:33:48 <K1773R> depends what you understand with "mapping"
143 2013-03-13 00:34:43 <RoboTeddy> is there somewhere I can look to see exactly what nodes are willing to tell each other?
144 2013-03-13 00:34:54 <RoboTeddy> or must I go read codes :(
145 2013-03-13 00:35:10 stretchwarren has quit (Ping timeout: 255 seconds)
146 2013-03-13 00:35:14 Ukto has joined
147 2013-03-13 00:35:15 <K1773R> tell each other what?
148 2013-03-13 00:35:40 i2pRelay has quit (Remote host closed the connection)
149 2013-03-13 00:35:48 <RoboTeddy> well, we're talking about getting a connection to every node. that lets us count nodes, but what information do nodes pass over connections? the more they tell us, the more cool things we can maybe figure out
150 2013-03-13 00:35:58 manet_ has joined
151 2013-03-13 00:36:08 <K1773R> start with this: $ grep version ~/.bitcoin/debug.log
152 2013-03-13 00:36:08 <lianj> RoboTeddy: version pkt tells stuff, and addr pkt tells neighbors to connect to next
153 2013-03-13 00:36:12 i2pRelay has joined
154 2013-03-13 00:36:26 <lianj> RoboTeddy: nothing very exciting
155 2013-03-13 00:36:30 rdponticelli has joined
156 2013-03-13 00:36:35 easye has quit (Read error: Operation timed out)
157 2013-03-13 00:36:40 prolxray has joined
158 2013-03-13 00:36:51 <RoboTeddy> lianj: but don't nodes gossip to pass around blocks and things?
159 2013-03-13 00:37:22 <RoboTeddy> (could you tell which nodes have seen a given block, and which ones haven't?)
160 2013-03-13 00:37:42 <lianj> gossip? they annonce what unconfirmed txs they have and what blocks they have, if the other node likes, he asks for them to be send over then
161 2013-03-13 00:37:46 meefozio has joined
162 2013-03-13 00:37:54 Conflict has quit (Remote host closed the connection)
163 2013-03-13 00:38:00 <phantomcircuit> lianj, they only do that when they receive the tx though
164 2013-03-13 00:38:01 manet has quit (Ping timeout: 252 seconds)
165 2013-03-13 00:38:05 stretchwarren has joined
166 2013-03-13 00:38:25 <phantomcircuit> if a node receives a tx and puts it into it's mempool, then another node connects, it doesn't tell the connected node about the tx's in it's mempool
167 2013-03-13 00:38:28 <lianj> right, you would have to gather the data over time
168 2013-03-13 00:38:29 B0g4r7 has joined
169 2013-03-13 00:38:33 <K1773R> RoboTeddy: yes, just ask to send them the block you want to check for
170 2013-03-13 00:38:37 marvborg has joined
171 2013-03-13 00:38:37 ThomasV has quit (Quit: Quitte)
172 2013-03-13 00:38:45 <sipa> phantomcircuit: you can ask them, via the mempool P2P command
173 2013-03-13 00:38:54 Boydy has quit (Remote host closed the connection)
174 2013-03-13 00:39:11 Conflict has joined
175 2013-03-13 00:39:16 <phantomcircuit> sipa, wouldn't it be so much easier if nodes just sent an inv of all their mempool tx's whenever a peer sent 'verack'
176 2013-03-13 00:39:17 <ProfMac> RoboTeddy, I think that peers.dat has a lot of information about which nodes a certain copy of bitcoind knows about. If you hae a program to parse that, you have a start on what you want. Then you simply connect to each, and download all the nodes that they know about...
177 2013-03-13 00:39:26 <K1773R> RoboTeddy: ie. -> hello RoboTeddy, gimme the block with the followhing hash: <somebtcblockhash>, if you say you dont have it i know you dont have it, so simple
178 2013-03-13 00:39:30 <phantomcircuit> that should be fairly simple to code and couldn't be expensive for either node
179 2013-03-13 00:39:44 <sipa> phantomcircuit: perhaps, but today that would be very dangerous i think
180 2013-03-13 00:39:44 meefozio has left ()
181 2013-03-13 00:39:52 <phantomcircuit> sipa, really? why
182 2013-03-13 00:40:00 <sipa> we sort of rely on mempools not being persistent to clear out conflicts
183 2013-03-13 00:40:03 <phantomcircuit> (genuine curiosity :P)
184 2013-03-13 00:40:07 <phantomcircuit> oooh
185 2013-03-13 00:40:26 <RoboTeddy> these recon nodes could avoid telling other nodes what tx it has learned about
186 2013-03-13 00:40:30 <phantomcircuit> so if you generated conflicting transactions you'd be in trouble if neither of them ever confirmed?
187 2013-03-13 00:40:34 da2ce7 has joined
188 2013-03-13 00:40:38 <sipa> sometimes, you really want to do a new version of a transactions, and thanks to mempools dying, that is actually possible after some time
189 2013-03-13 00:40:40 <phantomcircuit> so what you're saying is i shouldn't write a node that does that
190 2013-03-13 00:40:45 * phantomcircuit whistles
191 2013-03-13 00:40:47 <RoboTeddy> so it wouldn't increase persistence of mempools
192 2013-03-13 00:40:50 <sipa> it's bad that we rely on that
193 2013-03-13 00:40:54 <sipa> so we need a solution
194 2013-03-13 00:41:09 <sipa> like allowing replacement, and killing off too old mempool entries
195 2013-03-13 00:41:09 dpyle_ has quit (Quit: dpyle_)
196 2013-03-13 00:41:13 <phantomcircuit> sipa, when a transaction is confirmed in a block it clears conflicting transactions from the mempool right?
197 2013-03-13 00:41:18 <sipa> yes
198 2013-03-13 00:41:25 <sipa> since 0.8 :p
199 2013-03-13 00:41:33 <phantomcircuit> that doesn't happen in 0.7?
200 2013-03-13 00:41:36 <RoboTeddy> sipa: would it be OK to launch something like this on testnet? I don't want to screw anything up
201 2013-03-13 00:41:43 <sipa> phantomcircuit: no
202 2013-03-13 00:41:54 <phantomcircuit> sipa, with that behaviour it would actually be better that it's hard to replace transactions
203 2013-03-13 00:42:05 <phantomcircuit> it makes it much more difficult to actually execute a double spend
204 2013-03-13 00:42:12 <phantomcircuit> even if there was another fork
205 2013-03-13 00:42:19 drapetomano has joined
206 2013-03-13 00:42:22 <sipa> replacement should only be possible if it's not a conflict
207 2013-03-13 00:42:30 <sipa> like pay out to the same outputs at least
208 2013-03-13 00:42:37 <phantomcircuit> right but because this bug caused everybody to restart
209 2013-03-13 00:42:44 <phantomcircuit> the collective mempool was erased
210 2013-03-13 00:42:46 <da2ce7> having a 'revoke transaction' message it too hard? or open to attack?
211 2013-03-13 00:42:54 <phantomcircuit> which is probably the only reason that 211 btc double spend was possible
212 2013-03-13 00:43:01 <sipa> da2ce7: impossible to enforce
213 2013-03-13 00:43:17 <phantomcircuit> sipa, oh you mean if the result of the transaction was largely the same
214 2013-03-13 00:43:23 <sipa> phantomcircuit: yes
215 2013-03-13 00:43:27 one_zero has quit ()
216 2013-03-13 00:43:27 thepok has quit (Ping timeout: 264 seconds)
217 2013-03-13 00:43:33 <phantomcircuit> that would need to be implemented very carefully
218 2013-03-13 00:43:36 <lianj> phantomcircuit: who suffered from the doublespend?
219 2013-03-13 00:43:43 i2pRelay has quit (Remote host closed the connection)
220 2013-03-13 00:43:43 <phantomcircuit> lianj, lol btc-e
221 2013-03-13 00:43:53 <phantomcircuit> couldn't have happened to someone more deserving
222 2013-03-13 00:43:54 <RoboTeddy> sipa: sorry to bug, but is it OK to launch a ton of recon nodes on testnet and see if I can map it?
223 2013-03-13 00:44:06 <lianj> phantomcircuit: ha, yikes
224 2013-03-13 00:44:08 <RoboTeddy> I find bitcoin fascinating, but don't want to cause any problems
225 2013-03-13 00:44:13 i2pRelay has joined
226 2013-03-13 00:45:08 Jazb has quit (Ping timeout: 245 seconds)
227 2013-03-13 00:45:08 <da2ce7> well you have transactions where people can keep on adding inpute (muti sig, or reciver pays fees). Then sometimes there is a need to for a completly new tx.
228 2013-03-13 00:45:14 <sipa> RoboTeddy: that's what testnet is for, imho
229 2013-03-13 00:45:29 * RoboTeddy grins
230 2013-03-13 00:45:47 piuk has joined
231 2013-03-13 00:45:50 <da2ce7> maybe a 'I do not want this tansaction to be included' message, could be usefull in the second case.
232 2013-03-13 00:46:13 <da2ce7> even if it wasn't possible to enforce.
233 2013-03-13 00:46:40 <lianj> da2ce7: add more fees
234 2013-03-13 00:46:58 <lianj> (and hope)
235 2013-03-13 00:47:14 one_zero has joined
236 2013-03-13 00:47:24 <da2ce7> lianj: yes, implicitly that will doubble spend your old tx. However we do not do any checking if the outputs are the same or not.
237 2013-03-13 00:47:48 <sipa> ehm, you can't replace transactions right now, period
238 2013-03-13 00:47:57 <sipa> the code has been disabled since forever
239 2013-03-13 00:48:15 <sipa> but we can re-enable it, and with semantics we choose, basically
240 2013-03-13 00:48:27 <lianj> but isnt doublespending with a higher fee a bit like replacing it?
241 2013-03-13 00:48:42 <sipa> yes, but the higher fee is irrelevant then
242 2013-03-13 00:48:54 <sipa> it's just as easy to do that with an entirely different transaction
243 2013-03-13 00:48:57 <da2ce7> lianj: well you can have 'part-made' transactions, where you add more inputs later. afaik.
244 2013-03-13 00:49:03 <da2ce7> they are still the same tx.
245 2013-03-13 00:49:11 myrond_ has left ()
246 2013-03-13 00:50:05 Cryptoman has quit ()
247 2013-03-13 00:50:51 <lianj> ok well replacing tx may be a bad term then, but you can try to spent those txs inputs in another txs you create first, that includes higher fees then
248 2013-03-13 00:51:08 one_zero has quit (Client Quit)
249 2013-03-13 00:51:11 <lianj> s/txs/tx/
250 2013-03-13 00:51:22 <andrew12> this was probably already posted here: http://i.imgur.com/NncJ5fI.jpg
251 2013-03-13 00:51:36 cap2002 has quit ()
252 2013-03-13 00:51:38 <lianj> andrew12: yes, but still nice
253 2013-03-13 00:51:45 i2pRelay has quit (Remote host closed the connection)
254 2013-03-13 00:52:08 B0g4r7 has quit (Ping timeout: 276 seconds)
255 2013-03-13 00:52:15 i2pRelay has joined
256 2013-03-13 00:52:38 <da2ce7> need to go soon: But there is two different things: adding a new input, or changing a output. We should treet these cases very diffeently.
257 2013-03-13 00:52:41 tvisdog has joined
258 2013-03-13 00:54:18 <RoboTeddy> is there a way to ask a node which block is at the head of its longest chain?
259 2013-03-13 00:54:38 <lianj> da2ce7: in any case, its not really the same tx than anymore because it changes the hash
260 2013-03-13 00:54:40 <RoboTeddy> looking at https://en.bitcoin.it/wiki/Protocol_specification and don't see an obvious way
261 2013-03-13 00:54:52 <lianj> da2ce7: just a new tx that uses the same inputs
262 2013-03-13 00:55:06 <lianj> RoboTeddy: version pkt
263 2013-03-13 00:55:08 int03h has quit (Ping timeout: 258 seconds)
264 2013-03-13 00:55:31 canoon has quit (Ping timeout: 258 seconds)
265 2013-03-13 00:56:14 twmz has joined
266 2013-03-13 00:56:28 <lianj> RoboTeddy: tells you the height, not the hash though
267 2013-03-13 00:56:31 <RoboTeddy> version pkt looks like it contains start_height, int32_t, "The last block received by the emitting node" -- yeah, just the height
268 2013-03-13 00:57:04 <RoboTeddy> if you knew about the heights of the various forks, you could tell which fork that node is on in cases other than ties
269 2013-03-13 00:57:06 <lianj> combined with getheaderes and a state on your side you can get the hash
270 2013-03-13 00:57:14 da2ce7 has quit (Ping timeout: 250 seconds)
271 2013-03-13 00:57:55 <etotheipi_> is it normal for 0.8 bitcoind to generate a .bitcoin directory that Bitcoin-Qt rejects? It's telling me to backup that directory and delete everything except wallet.dat
272 2013-03-13 00:58:12 <RoboTeddy> lianj: sorry I'm a total newb at the protocol; how would that work?
273 2013-03-13 00:58:27 yellowhat has joined
274 2013-03-13 00:58:37 yellowhat has quit (Changing host)
275 2013-03-13 00:58:37 yellowhat has joined
276 2013-03-13 00:58:37 yellowhat has quit (Changing host)
277 2013-03-13 00:58:37 yellowhat has joined
278 2013-03-13 00:58:38 rdponticelli has quit (Ping timeout: 276 seconds)
279 2013-03-13 00:58:47 twmz has quit (Client Quit)
280 2013-03-13 00:59:18 twmz has joined
281 2013-03-13 00:59:18 <RoboTeddy> the satoshi client apparently happily reports back headers of blocks it considers invalid; does that make getheaders less useful for figuring out which chain a node believes is valid?
282 2013-03-13 00:59:38 one_zero has joined
283 2013-03-13 00:59:49 i2pRelay has quit (Remote host closed the connection)
284 2013-03-13 01:00:21 i2pRelay has joined
285 2013-03-13 01:00:29 <jgarzik> RoboTeddy: details?
286 2013-03-13 01:00:35 <lianj> why would getheaders report back headers it think are not on the main chain?
287 2013-03-13 01:00:42 NestedUniverse has quit (Quit: Page closed)
288 2013-03-13 01:00:51 <RoboTeddy> just reading what is claimed here: https://en.bitcoin.it/wiki/Protocol_specification#getheaders
289 2013-03-13 01:01:06 sgornick has joined
290 2013-03-13 01:01:32 <lianj> "if the block locator object contains a hash on the invalid branch."
291 2013-03-13 01:01:36 <lianj> read to the end
292 2013-03-13 01:01:59 B0g4r7 has joined
293 2013-03-13 01:02:11 <yellowhat> anyone around with bitcoinJ skills? why does this code fail:
294 2013-03-13 01:02:12 <yellowhat> http://pastebin.com/iX11KhT2
295 2013-03-13 01:02:12 hsmiths has quit (Quit: Nettalk6 - www.ntalk.de)
296 2013-03-13 01:02:19 <RoboTeddy> lianj: ah, so one could just specify a locator hash that's sufficiently old as to not be involved in a fork?
297 2013-03-13 01:02:37 <yellowhat> java.io.IOException: Failed to rename
298 2013-03-13 01:03:21 <lianj> RoboTeddy: yes
299 2013-03-13 01:04:02 RazielZ has quit (Ping timeout: 264 seconds)
300 2013-03-13 01:04:05 <RoboTeddy> jgarzik: thinking about running enough public servers (60k ish?) s.t. every node in the bitcoin network ends up connected to at least one, and using this to determine both the number of nodes in the whole network, as well as the count of nodes on each fork (in the event of a fork)
301 2013-03-13 01:04:12 <TheSeven> which line does it happen at?
302 2013-03-13 01:04:42 <jgarzik> RoboTeddy: That's why the code specifies a lot of hashes, when building a 'getheaders' packet
303 2013-03-13 01:05:09 <TheSeven> RoboTeddy: you will never find all nodes
304 2013-03-13 01:05:14 Ukto has left ()
305 2013-03-13 01:05:33 <jgarzik> RoboTeddy: you have access to 60k Class C networks?
306 2013-03-13 01:05:49 <TheSeven> some of them are firewalled in ways so that you just can't reach them no matter what you try
307 2013-03-13 01:06:00 <RoboTeddy> jgarzik: I heard you can run bitcoind on whatever port you want, so one server could get you 60k instances or so
308 2013-03-13 01:06:01 mortikia has quit (Disconnected by services)
309 2013-03-13 01:06:01 hsmiths has joined
310 2013-03-13 01:06:03 * jgarzik wonders about the total number of Autonomous Systems on the internet
311 2013-03-13 01:06:06 <TheSeven> jgarzik: ipv6? :P
312 2013-03-13 01:06:10 Boydy has joined
313 2013-03-13 01:06:13 <jgarzik> RoboTeddy: oh, don't bother
314 2013-03-13 01:06:23 <phantomcircuit> RoboTeddy, the client wont connect to the same /24 more than once
315 2013-03-13 01:06:31 <lianj> RoboTeddy: ^
316 2013-03-13 01:06:36 <jgarzik> phantomcircuit: shhhh. better to let him waste effort ;p
317 2013-03-13 01:06:38 <phantomcircuit> so you would need 60k ipv4 addresses in different /24s
318 2013-03-13 01:06:54 <RoboTeddy> phantomcircuit: that's OK, you would only need each client to connect to one of your recon nodes, no?
319 2013-03-13 01:06:54 <K1773R> or ipv6's
320 2013-03-13 01:06:58 zooko_afk has joined
321 2013-03-13 01:07:15 <RoboTeddy> jgarzik: hey I at least have good intentions here, would've been useful to know how many nodes were on each fork yesterday :)
322 2013-03-13 01:07:15 <K1773R> but i guess the ipv6 net is little
323 2013-03-13 01:07:18 klmist has quit (Ping timeout: 256 seconds)
324 2013-03-13 01:07:29 <K1773R> we did know?
325 2013-03-13 01:07:43 <K1773R> 0.8 bitcoind was obviously on the 0.8 chain and all others on the 0.7
326 2013-03-13 01:07:46 mortikia has joined
327 2013-03-13 01:07:49 i2pRelay has quit (Remote host closed the connection)
328 2013-03-13 01:08:03 <Ssateneth> i havent touched my bitcoin client and its 0.8
329 2013-03-13 01:08:18 <RoboTeddy> but we didn't have a count of 0.8 bitcoind and 0.7 bitcoind. we kind of knew since we sort've knew where each pool was, and that accounted for a lot
330 2013-03-13 01:08:19 <K1773R> Ssateneth, its over
331 2013-03-13 01:08:21 i2pRelay has joined
332 2013-03-13 01:08:22 <Ssateneth> the larger chain (that appears valid to the client) will win, and 0.7 is larger
333 2013-03-13 01:08:26 Conflict has quit (Remote host closed the connection)
334 2013-03-13 01:08:41 <Ssateneth> K1773R no shit :p
335 2013-03-13 01:08:45 <Ssateneth> oh, im in the wrong channel
336 2013-03-13 01:08:49 <Ssateneth> i thought i was replying to synops
337 2013-03-13 01:08:52 <K1773R> no, 0.8 was larger due to not enough locks in BDB being able to import the block
338 2013-03-13 01:09:12 <K1773R> since 0.8 uses leveldb it was no problem for it, so it was 1 block ahead
339 2013-03-13 01:09:18 <Ssateneth> 0.7 overtook 0.8 something like 3 am though
340 2013-03-13 01:09:25 Guest51255 has quit (Read error: Operation timed out)
341 2013-03-13 01:09:30 <K1773R> 0.7 and below didnt agree on these blocks and tada, "CHAIN FORK"
342 2013-03-13 01:09:34 <Ssateneth> yes i know
343 2013-03-13 01:09:38 <Ssateneth> you dont need to explain to me
344 2013-03-13 01:09:40 <K1773R> ok, just in case :P
345 2013-03-13 01:09:48 * warren wonders how many times this will be explained.
346 2013-03-13 01:09:49 eric___ has quit (Ping timeout: 255 seconds)
347 2013-03-13 01:10:01 hsmiths has quit (Read error: Connection reset by peer)
348 2013-03-13 01:10:02 <TheSeven> after pushing like 15TH/s over the 0.7 chain won
349 2013-03-13 01:10:03 <K1773R> warren, i hoped i dont have to do it again, pissed me off yesterday
350 2013-03-13 01:10:03 <Ssateneth> luke-jr already gave me the rundown, and I also read the various pages explaining what happened too
351 2013-03-13 01:10:07 Zaba_ has joined
352 2013-03-13 01:10:16 <warren> K1773R: isn't there a URL we can just paste every time?
353 2013-03-13 01:10:25 <K1773R> /dev/null
354 2013-03-13 01:10:29 <warren> K1773R: unfortunately, the media has been getting it wrong in their reporting =(
355 2013-03-13 01:10:38 <RoboTeddy> jgarzik: if you ran 60k public nodes from one IP, you could easily end up with at least 1 incoming connection per node in the network, right? (there are only ~60k existing nodes that accept incoming connections). or am I misunderstanding something?
356 2013-03-13 01:10:41 <K1773R> thats why its called "media"
357 2013-03-13 01:11:21 <tjader> warren: http://www.reddit.com/r/Bitcoin/comments/1a51xx/now_that_its_over_the_blockchain_fork_explained/
358 2013-03-13 01:11:39 <jgarzik> RoboTeddy: no, not at all
359 2013-03-13 01:12:06 <RoboTeddy> jgarzik: sorry, is that "no, that's not possible" or "no, you don't misunderstand" ?
360 2013-03-13 01:12:11 <K1773R> Regular users are not affected. Their transactions are included in both ledgers and don't need to change any programs. <-- wrong, a ugly double spend did happen
361 2013-03-13 01:12:12 MashRinx has quit (Quit: Page closed)
362 2013-03-13 01:12:16 <jgarzik> RoboTeddy: bitcoin tries to spread connections across as many _networks_ as possible. It doesn't matter if you have 1,000,000 nodes on that IP.
363 2013-03-13 01:12:22 <K1773R> slight chance... ouch
364 2013-03-13 01:12:25 <warren> tjader: is the author of that here?
365 2013-03-13 01:12:37 <tjader> Dunno.
366 2013-03-13 01:12:37 hsmiths has joined
367 2013-03-13 01:12:56 <K1773R> a slight chance for double spends (25 block difference), thats a fking huge chance...
368 2013-03-13 01:13:09 <jgarzik> K1773R: link to malicious double-spend?
369 2013-03-13 01:13:12 <RoboTeddy> jgarzik: right, but if each bitcoin client has, say... 7 out going connections, and you're running 1 in 2 publicly accessible nodes, wouldn't most clients end up hitting you once? (never twice, since they spread connections across networks)
370 2013-03-13 01:13:29 <K1773R> let me search
371 2013-03-13 01:13:52 <warren> K1773R: a double spend happened because the merchant was accepting the fork that lost?
372 2013-03-13 01:13:52 <jgarzik> RoboTeddy: no
373 2013-03-13 01:13:53 whiterabbit has joined
374 2013-03-13 01:14:12 <K1773R> jgarzik: https://bitcointalk.org/index.php?topic=152348.0
375 2013-03-13 01:14:51 ic3 has quit (Quit: Lost terminal)
376 2013-03-13 01:14:53 <RoboTeddy> why not? do nodes tell each other who they're connected to, in an effort for the network as a whole to avoid connecting to one /24 too much?
377 2013-03-13 01:15:11 <K1773R> jgarzik / warren: http://blockchain.info/tx/12814b8ad57ce5654ba69eb26a52ddae1bff42093ca20cef3ad96fe7fd85d195
378 2013-03-13 01:15:12 HiWEB has joined
379 2013-03-13 01:15:26 <RoboTeddy> (if A is connected to B, and A connects to me, is B less likely to also connect to me?)
380 2013-03-13 01:15:31 <owowo> ^^that's why I hate mac
381 2013-03-13 01:15:32 wrabbit has quit (Ping timeout: 276 seconds)
382 2013-03-13 01:15:34 whiterabbit is now known as wrabbit
383 2013-03-13 01:15:50 i2pRelay has quit (Remote host closed the connection)
384 2013-03-13 01:16:20 i2pRelay has joined
385 2013-03-13 01:17:28 <Luke-Jr> lol over
386 2013-03-13 01:17:28 rdponticelli has joined
387 2013-03-13 01:18:28 Zaba_ has quit (Ping timeout: 245 seconds)
388 2013-03-13 01:18:55 <RoboTeddy> it'd be cool if someone explained to me why I'm being stupid before I go waste a bunch of time ;)
389 2013-03-13 01:19:07 <tjader> RoboTeddy:
390 2013-03-13 01:19:35 <tjader> It could be that the client first sorts all nodes into /24s, picks a few of those and then picks a random node from each /24
391 2013-03-13 01:19:56 <tjader> Instead of picking a random node from the whole pool then blacklisting the /24 and repeating
392 2013-03-13 01:20:52 sudofox has joined
393 2013-03-13 01:20:56 <sudofox> Hi
394 2013-03-13 01:21:03 <sudofox> Who wants free testnet BTC
395 2013-03-13 01:21:11 <sudofox> I've got tons of it
396 2013-03-13 01:21:31 <buddyrandom> this guy needs some http://tpfaucet.appspot.com/
397 2013-03-13 01:21:35 <RoboTeddy> tjader: I see, that could make sense, thanks.
398 2013-03-13 01:21:37 <K1773R> sudofox, donate it to tpfacute pls
399 2013-03-13 01:21:38 elkingrey has joined
400 2013-03-13 01:21:40 int03h has joined
401 2013-03-13 01:21:58 <K1773R> wow thats really dry
402 2013-03-13 01:22:05 <K1773R> did someone start selling TBTC again? *sigh*
403 2013-03-13 01:22:27 <sudofox> @buddyrandom I'll send 'em 1000 TBTC
404 2013-03-13 01:22:31 <sudofox> Thanks
405 2013-03-13 01:22:39 <K1773R> thanks to you ;)
406 2013-03-13 01:22:46 <tjader> RoboTeddy: I have no idea if it's true or not :p
407 2013-03-13 01:23:18 <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
408 2013-03-13 01:23:26 <buddyrandom> woot :)
409 2013-03-13 01:23:31 <RoboTeddy> I guess I should go, you know, read the code..
410 2013-03-13 01:23:51 i2pRelay has quit (Remote host closed the connection)
411 2013-03-13 01:24:02 JDuke128 has quit (Quit: [BB])
412 2013-03-13 01:24:18 <sudofox> Sent.
413 2013-03-13 01:24:20 i2pRelay has joined
414 2013-03-13 01:24:47 canoon has joined
415 2013-03-13 01:24:51 <tjader> RoboTeddy: if it works as I said it would kill the idea.
416 2013-03-13 01:25:23 pacpac has quit (Quit: Leaving)
417 2013-03-13 01:26:05 <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
418 2013-03-13 01:26:23 <RoboTeddy> (I have no idea what the numbers look like in actuality)
419 2013-03-13 01:26:30 <MKCoin> <K1773R> did someone start selling TBTC again? *sigh*
420 2013-03-13 01:26:30 jtimon has quit (Ping timeout: 260 seconds)
421 2013-03-13 01:26:41 <MKCoin> sell them for what? I made 50 tbtc earlier /by accident/
422 2013-03-13 01:26:59 <tjader> RoboTeddy: but the amound of nodes inside a single /24 wouldn't matter for the likelyhood of it connecting to the /24
423 2013-03-13 01:27:04 <K1773R> ppl tryd to trade TBTC for BTC. there was even once an exchange for it!
424 2013-03-13 01:27:09 <MKCoin> lawl
425 2013-03-13 01:27:31 <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
426 2013-03-13 01:27:32 <K1773R> maybe some stupid altcoincollector who thingks TBTC is worth something
427 2013-03-13 01:27:46 <K1773R> -g
428 2013-03-13 01:27:53 Conflict has joined
429 2013-03-13 01:28:00 <tjader> RoboTeddy: you're a few orders of magnitude off.
430 2013-03-13 01:28:29 <tjader> there are 16 million /24s.
431 2013-03-13 01:28:35 elkingrey has quit (Quit: Leaving)
432 2013-03-13 01:28:44 <sudofox> @K1773R I google tpfacute, didn't find anything
433 2013-03-13 01:29:00 <K1773R> sudofox, the same as buddyrandom did link u
434 2013-03-13 01:29:15 <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
435 2013-03-13 01:29:30 <K1773R> there are 2 faucets for TBTC, let me search the second
436 2013-03-13 01:29:34 <phantomcircuit> RoboTeddy, there are ~500k ip:port combinations being passed around the network by addr messages
437 2013-03-13 01:29:44 <phantomcircuit> RoboTeddy, there are ~4k nodes which accept connections
438 2013-03-13 01:29:48 <tjader> I'm not very familiar with the peer discovery protocol, but 20 seems quite low.
439 2013-03-13 01:29:57 <phantomcircuit> you're going to need a lot more than 20 ipv4 addresses
440 2013-03-13 01:30:07 <sudofox> Ohhh tpfaucet makes more sense
441 2013-03-13 01:30:08 <RoboTeddy> phantomcircuit: ok, true story! thanks for the numbers
442 2013-03-13 01:30:14 piuk has quit (Quit: Page closed)
443 2013-03-13 01:30:21 <MKCoin> http://testnet.mojocoin.com/
444 2013-03-13 01:30:25 <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
445 2013-03-13 01:30:26 <phantomcircuit> so
446 2013-03-13 01:31:09 <K1773R> sudofox: here is the second -> http://testnet.mojocoin.com/
447 2013-03-13 01:31:11 johnsoft has quit (Ping timeout: 252 seconds)
448 2013-03-13 01:31:13 <phantomcircuit> you're talking more like 500 ipv4 addresses and actually to get good coverage probably more like 1000
449 2013-03-13 01:31:25 <phantomcircuit> and they would all have to be actual nodes that actually returned useful results
450 2013-03-13 01:31:25 <K1773R> argl MKCoin
451 2013-03-13 01:31:26 <phantomcircuit> so
452 2013-03-13 01:31:29 <MKCoin> X)
453 2013-03-13 01:31:29 <phantomcircuit> good luck with that
454 2013-03-13 01:31:39 <sudofox> They have >3000 TBTC, I think they're fine
455 2013-03-13 01:31:53 i2pRelay has quit (Remote host closed the connection)
456 2013-03-13 01:31:57 <sudofox> tpfaucet has 0.95004663 TestNet Bitcoins
457 2013-03-13 01:32:14 <sudofox> I sent 1000 TBTC
458 2013-03-13 01:32:16 BTC_Bear is now known as hbrntng!~BTC_Bear@unaffiliated/btc-bear/x-5233302|BTC_Bear
459 2013-03-13 01:32:23 i2pRelay has joined
460 2013-03-13 01:33:01 johnsoft_ has joined
461 2013-03-13 01:33:16 edward_h has joined
462 2013-03-13 01:33:32 <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
463 2013-03-13 01:33:42 <phantomcircuit> RoboTeddy, ahaha no you couldn't
464 2013-03-13 01:33:50 <phantomcircuit> your actual node would fucking melt
465 2013-03-13 01:34:05 <RoboTeddy> phantomcircuit: how much does a node actually have to do to have its address be passed around?
466 2013-03-13 01:34:09 <K1773R> RoboTeddy: you cant, most proxy dont allow bind tunnels (expect some few SOCKS5 proxys)
467 2013-03-13 01:34:09 <phantomcircuit> you'd need at least 10 actual nodes
468 2013-03-13 01:34:17 coolsa has quit (Ping timeout: 245 seconds)
469 2013-03-13 01:34:30 <phantomcircuit> K1773R, shh that's the wrong criticism
470 2013-03-13 01:34:36 <K1773R> LOL
471 2013-03-13 01:34:43 <phantomcircuit> K1773R, haproxy
472 2013-03-13 01:35:24 <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
473 2013-03-13 01:35:25 <K1773R> u forgot he dosnt know how the IP nor the TCP stack works, so haproxy is way out of range
474 2013-03-13 01:35:36 <phantomcircuit> and which doesn't trigger and of the DoS conditionals all over the codebase
475 2013-03-13 01:35:58 <phantomcircuit> the main problem is that 1000 ips would cost like
476 2013-03-13 01:35:58 <RoboTeddy> phantomcircuit: version/verack is pretty simple. do you have to actively gossip transactions and block information too?
477 2013-03-13 01:36:01 <phantomcircuit> $2000/month
478 2013-03-13 01:36:09 <phantomcircuit> and the information you get is basically useless
479 2013-03-13 01:36:32 <RoboTeddy> K1773R: errm I use haproxy :)
480 2013-03-13 01:36:37 <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
481 2013-03-13 01:36:48 <RoboTeddy> phantomcircuit: yeah, the IPs do seem like the expensive part
482 2013-03-13 01:37:01 <RoboTeddy> phantomcircuit: thanks for the lesson
483 2013-03-13 01:37:05 <K1773R> if (networkmagic->req() == GETVERSION) { printf "DoS detected, crash imminent"; } ROFL
484 2013-03-13 01:37:17 Stilger has left ("Leaving")
485 2013-03-13 01:37:19 <phantomcircuit> yeah there's stuff like that all over the place
486 2013-03-13 01:38:30 reardencode has left ()
487 2013-03-13 01:39:12 <K1773R> RoboTeddy, i told you already way ealier its WAY to expensive
488 2013-03-13 01:39:20 <K1773R> why didnt u want to believe me huh?
489 2013-03-13 01:39:28 <phantomcircuit> numbers
490 2013-03-13 01:39:29 <RoboTeddy> cause I like reasons for things :)
491 2013-03-13 01:39:31 <phantomcircuit> you needed numbers
492 2013-03-13 01:39:57 i2pRelay has quit (Remote host closed the connection)
493 2013-03-13 01:39:59 Roi has joined
494 2013-03-13 01:40:08 sudofox has quit (Ping timeout: 250 seconds)
495 2013-03-13 01:40:08 Boydy has quit (Remote host closed the connection)
496 2013-03-13 01:40:10 <K1773R> i told you wont have somany IPs nor servers to do it...
497 2013-03-13 01:40:19 <K1773R> well, good to see it sorted out
498 2013-03-13 01:40:27 i2pRelay has joined
499 2013-03-13 01:40:49 <RoboTeddy> K1773R: right, but at that point I was under the assumption you could run 60k servers off each port of a single ip
500 2013-03-13 01:41:01 leviathanbaphz has joined
501 2013-03-13 01:41:15 int03h2 has joined
502 2013-03-13 01:41:18 <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
503 2013-03-13 01:41:55 Leviathanzz has quit (Ping timeout: 252 seconds)
504 2013-03-13 01:42:00 Paul_Hartman has joined
505 2013-03-13 01:43:17 Shrugged has joined
506 2013-03-13 01:43:40 Conflict has quit (Remote host closed the connection)
507 2013-03-13 01:44:05 Test157382 has joined
508 2013-03-13 01:44:16 Xeno-Genesis has quit (Remote host closed the connection)
509 2013-03-13 01:44:43 tockitj has quit (Read error: Operation timed out)
510 2013-03-13 01:44:47 johnsoft_1 has joined
511 2013-03-13 01:45:14 whiterabbit has joined
512 2013-03-13 01:45:29 wrabbit has quit (Remote host closed the connection)
513 2013-03-13 01:45:32 whiterabbit is now known as wrabbit
514 2013-03-13 01:46:00 <K1773R> of course you CAN
515 2013-03-13 01:46:06 <K1773R> but no nodes wil connect to them :P
516 2013-03-13 01:46:08 Test157382 has quit (Remote host closed the connection)
517 2013-03-13 01:46:21 <K1773R> anyway, gtg. cya later around
518 2013-03-13 01:46:53 johnsoft_ has quit (Ping timeout: 258 seconds)
519 2013-03-13 01:47:00 JDuke128 has joined
520 2013-03-13 01:47:25 robwerks has joined
521 2013-03-13 01:47:38 grubles has quit (Quit: leaving)
522 2013-03-13 01:47:58 i2pRelay has quit (Remote host closed the connection)
523 2013-03-13 01:48:31 i2pRelay has joined
524 2013-03-13 01:48:55 Test157382 has joined
525 2013-03-13 01:48:59 Boydy has joined
526 2013-03-13 01:48:59 Boydy has quit (Changing host)
527 2013-03-13 01:48:59 Boydy has joined
528 2013-03-13 01:49:17 saivann has quit ()
529 2013-03-13 01:50:33 Roi has quit (Ping timeout: 245 seconds)
530 2013-03-13 01:50:48 Test157382 has quit (Remote host closed the connection)
531 2013-03-13 01:51:19 nsillik_ has joined
532 2013-03-13 01:51:57 kadoban has quit (Ping timeout: 252 seconds)
533 2013-03-13 01:55:53 robwerks has left ("Leaving")
534 2013-03-13 01:56:03 i2pRelay has quit (Remote host closed the connection)
535 2013-03-13 01:56:25 randy-waterhouse has joined
536 2013-03-13 01:56:33 i2pRelay has joined
537 2013-03-13 01:58:34 marvborg has quit (Quit: Page closed)
538 2013-03-13 01:59:27 dvide has quit ()
539 2013-03-13 01:59:55 MobiusL has quit (Quit: Ex-Chat)
540 2013-03-13 01:59:57 pobri19` is now known as pobri19
541 2013-03-13 02:01:04 HiWEB has quit (Ping timeout: 258 seconds)
542 2013-03-13 02:01:07 Shaded has quit (Read error: Connection reset by peer)
543 2013-03-13 02:01:45 nimdAHK has joined
544 2013-03-13 02:01:48 epylar_ has quit (Ping timeout: 245 seconds)
545 2013-03-13 02:01:50 int03h2 has quit (Ping timeout: 258 seconds)
546 2013-03-13 02:01:52 jaake has left ()
547 2013-03-13 02:03:29 marvinborg has joined
548 2013-03-13 02:03:44 COGSMITH has quit (Read error: Connection reset by peer)
549 2013-03-13 02:04:07 i2pRelay has quit (Remote host closed the connection)
550 2013-03-13 02:04:23 ToryJujube has joined
551 2013-03-13 02:04:41 i2pRelay has joined
552 2013-03-13 02:05:21 nicocesar has joined
553 2013-03-13 02:05:25 dudr has joined
554 2013-03-13 02:05:43 <[Tycho]> Anyone here ?
555 2013-03-13 02:05:44 dudr has quit (Client Quit)
556 2013-03-13 02:05:47 COGSMITH has joined
557 2013-03-13 02:05:58 <deadweasel> yo.
558 2013-03-13 02:06:08 <deadweasel> I only have a question though.
559 2013-03-13 02:06:54 <randy-waterhouse> quiet huh?
560 2013-03-13 02:07:07 <randy-waterhouse> almost too quiet
561 2013-03-13 02:07:27 <randy-waterhouse> hey [Tycho]
562 2013-03-13 02:07:45 <[Tycho]> Yes.
563 2013-03-13 02:07:55 undecim has quit (Ping timeout: 260 seconds)
564 2013-03-13 02:08:58 <randy-waterhouse> how's deepbit doing these days?
565 2013-03-13 02:09:33 <[Tycho]> Not very fine because of that big reorg.
566 2013-03-13 02:09:48 <randy-waterhouse> oh ... running 0.8 was it?
567 2013-03-13 02:09:57 Mandrius has quit (Ping timeout: 248 seconds)
568 2013-03-13 02:10:04 <[Tycho]> Looks like I need to write my own DB layer instead of that buggy BDB
569 2013-03-13 02:10:14 <randy-waterhouse> BDB is not buggy
570 2013-03-13 02:10:23 <[Tycho]> No, my code is based on 0.3.4.
571 2013-03-13 02:10:23 <randy-waterhouse> bitcoiners were not using it right ...
572 2013-03-13 02:10:34 <[Tycho]> Maybe not buggy, but causes problems.
573 2013-03-13 02:10:43 <[Tycho]> Lots of.
574 2013-03-13 02:10:45 <randy-waterhouse> well if you don't know what youo are doing maybe
575 2013-03-13 02:10:59 <doublec> even if you'd written your own you'd have got stuck unless it reproduced the bdb issue the older bitcoin clients had
576 2013-03-13 02:11:00 <[Tycho]> That wasn't me who did it.
577 2013-03-13 02:11:06 <randy-waterhouse> pointing fingers at this point is not going to solve anything anyway
578 2013-03-13 02:11:11 marvinborg has quit (Remote host closed the connection)
579 2013-03-13 02:11:22 <doublec> well, not stuck, you'd have been on the wrong side of the fork
580 2013-03-13 02:11:41 <randy-waterhouse> [Tycho]: did you lose money?
581 2013-03-13 02:11:50 <randy-waterhouse> ah, bitcoins rather
582 2013-03-13 02:11:53 <[Tycho]> doublec: my current problem was caused by the reorg.
583 2013-03-13 02:12:03 i2pRelay has quit (Remote host closed the connection)
584 2013-03-13 02:12:16 <[Tycho]> Yes, I think I will because I'm planning to pay my users anyway.
585 2013-03-13 02:12:33 i2pRelay has joined
586 2013-03-13 02:12:43 <deadweasel> it's smarter not to upgrade first in 90% of mission critical applications.
587 2013-03-13 02:12:57 <randy-waterhouse> pool operation just got a whole lot more risky that's for sure
588 2013-03-13 02:12:58 <deadweasel> every idiot is techinically "testing in production"
589 2013-03-13 02:13:19 <[Tycho]> deadweasel: I didn't. See abobe. I'm running heavily customized 0.3.4
590 2013-03-13 02:13:35 <[Tycho]> *above
591 2013-03-13 02:13:37 undecim has joined
592 2013-03-13 02:13:37 <deadweasel> I know, i was just seeing the damned sense in that.
593 2013-03-13 02:14:04 <randy-waterhouse> [Tycho]: what are you block size limit settings?
594 2013-03-13 02:14:05 <[Tycho]> deadweasel: I just got hit by ANOTHER side of that problem - vulnerability to big reorgs.
595 2013-03-13 02:14:22 <deadweasel> every time i encouter mission critical software it's old as fuck because they don't dare touch it. because it ain't broke yet.
596 2013-03-13 02:14:36 Ssateneth has left ()
597 2013-03-13 02:14:40 <doublec> [Tycho]: how did the reorg cause problems?
598 2013-03-13 02:14:57 protus has joined
599 2013-03-13 02:15:14 <warren> [Tycho]: are you a pool op?
600 2013-03-13 02:15:15 tyn has quit (Ping timeout: 264 seconds)
601 2013-03-13 02:15:17 <[Tycho]> randy-waterhouse: AFAIR I'm taking about 50 Kb of free TXes. TXes are considered free when paying less than 0.01 per Kb.
602 2013-03-13 02:15:21 <[Tycho]> warren: yes.
603 2013-03-13 02:15:38 <[Tycho]> doublec: BDB can't allocate memory.
604 2013-03-13 02:15:46 pobri19` has joined
605 2013-03-13 02:16:18 <doublec> [Tycho]: sounds familiar. There were big reorgs on an alt chain once that caused similar issues.
606 2013-03-13 02:16:20 <randy-waterhouse> [Tycho]: you might want to read up on BDB limits ... goes for all pool operators using pre-0.8 (or derivative) code
607 2013-03-13 02:16:38 <randy-waterhouse> the devs cannot be expected to carry the water for everybody
608 2013-03-13 02:16:50 <warren> [Tycho]: you may be interested in Luke-Jr's safety measure. Operate multiple versions of bitcoind and query them to check for inconsistencies, and auto-stop if there is a problem.
609 2013-03-13 02:17:01 <[Tycho]> Good idea.
610 2013-03-13 02:17:19 <doublec> [Tycho]: can you run an unmodified bitcoin that can handle the reorg to get past it, then continue with your modified one?
611 2013-03-13 02:17:24 holorga_ has joined
612 2013-03-13 02:17:37 <doublec> or just download the blockchain again
613 2013-03-13 02:17:47 <[Tycho]> I'm redownloading.
614 2013-03-13 02:18:10 <doublec> I guess that knocks your pool offline for 24 hours or so
615 2013-03-13 02:18:14 HackFisher has joined
616 2013-03-13 02:18:14 <[Tycho]> I wonder where is that bug...
617 2013-03-13 02:18:39 pobri19 has quit (Ping timeout: 272 seconds)
618 2013-03-13 02:19:00 HackFisher has left ()
619 2013-03-13 02:19:35 <doublec> [Tycho]: there was a commit with this message "Prevent stuck block download in large reorganisations"
620 2013-03-13 02:19:39 <randy-waterhouse> [Tycho]: https://bitcointalk.org/index.php?topic=152208.0 <---- did you see this thread?
621 2013-03-13 02:19:46 <doublec> [Tycho]: b2fe3a5ca690299a55f52beee3932f0126b79363
622 2013-03-13 02:20:05 i2pRelay has quit (Remote host closed the connection)
623 2013-03-13 02:20:07 <randy-waterhouse> set lock limit greater than 10,000
624 2013-03-13 02:20:29 <[Tycho]> I think that my bitcoind doesn't works with these settings.
625 2013-03-13 02:20:32 holorga has quit (Ping timeout: 276 seconds)
626 2013-03-13 02:20:37 i2pRelay has joined
627 2013-03-13 02:20:37 <doublec> [Tycho]: might not be your issue though https://github.com/bitcoin/bitcoin/pull/948
628 2013-03-13 02:20:39 <randy-waterhouse> like 40,000 ... the offensive block triggered 5,000 lock limit (*2n)
629 2013-03-13 02:21:36 <randy-waterhouse> but you know your own code better than anybody .. so might be unhelpful
630 2013-03-13 02:22:01 <[Tycho]> doublec: I think it's another issue.
631 2013-03-13 02:22:23 <[Tycho]> ************************
632 2013-03-13 02:22:23 <[Tycho]> EXCEPTION: 11DbException
633 2013-03-13 02:22:23 <[Tycho]> Db::del: Cannot allocate memory
634 2013-03-13 02:22:25 <[Tycho]> bitcoin in ProcessMessage()
635 2013-03-13 02:22:38 <[Tycho]> That's how it failed on reorg.
636 2013-03-13 02:23:26 Anto_- has quit (Read error: Connection reset by peer)
637 2013-03-13 02:23:42 Ant0 has joined
638 2013-03-13 02:24:32 BTCTrader2 has quit (Quit: BTCTrader2)
639 2013-03-13 02:24:50 Casimir1904 has quit (Ping timeout: 258 seconds)
640 2013-03-13 02:25:46 Lexx_ has joined
641 2013-03-13 02:28:09 i2pRelay has quit (Remote host closed the connection)
642 2013-03-13 02:28:22 debiantoruser has joined
643 2013-03-13 02:28:41 i2pRelay has joined
644 2013-03-13 02:28:47 debiantoruser is now known as Guest15459
645 2013-03-13 02:28:58 Lexx has quit (Ping timeout: 245 seconds)
646 2013-03-13 02:29:48 josephin has quit (Ping timeout: 245 seconds)
647 2013-03-13 02:29:50 sudofox has joined
648 2013-03-13 02:30:13 edward_h has quit ()
649 2013-03-13 02:31:01 josephin has joined
650 2013-03-13 02:31:35 <sudofox> Hi guys!
651 2013-03-13 02:31:59 <sudofox> Check out TP testnet faucet now!
652 2013-03-13 02:33:19 Linkandzelda_ has joined
653 2013-03-13 02:33:52 <buddyrandom> sweet
654 2013-03-13 02:34:00 hkfhskjd has left ("Leaving")
655 2013-03-13 02:34:14 tjb0607 has left ("Segmentation fault")
656 2013-03-13 02:34:27 <randy-waterhouse> any new work on appropriate fee levels?
657 2013-03-13 02:35:08 Guest16726 has quit (Quit: Saliendo)
658 2013-03-13 02:35:25 amantonop has joined
659 2013-03-13 02:35:45 erska has quit (Read error: Operation timed out)
660 2013-03-13 02:36:12 i2pRelay has quit (Remote host closed the connection)
661 2013-03-13 02:36:19 erska has joined
662 2013-03-13 02:36:24 <sudofox> Anyone want/need some testnet BTC? I'm giving it out free
663 2013-03-13 02:36:43 i2pRelay has joined
664 2013-03-13 02:37:06 <warren> I'll get some eventually at 9KH/sec. I hope.
665 2013-03-13 02:37:20 Linkandzelda has quit (Ping timeout: 250 seconds)
666 2013-03-13 02:37:42 <sudofox> @warren I have 3864.38399995 BTC
667 2013-03-13 02:38:09 <HM> is the current best strategy to run 0.7.2?
668 2013-03-13 02:38:42 <HM> or has githead been fixed
669 2013-03-13 02:38:45 Uthark has quit (Quit: Leaving)
670 2013-03-13 02:38:55 <MKCoin> 'We are giving away 50 BTC per request' I guess the amount given is dynamic?
671 2013-03-13 02:39:02 <sudofox> Yes
672 2013-03-13 02:39:04 <gmaxwell> HM: it's fine to run 0.8 unless you're a major pool.
673 2013-03-13 02:39:10 <randy-waterhouse> HM: neither is "wrong" per se .. they just diagree
674 2013-03-13 02:39:19 <sudofox> gotta go
675 2013-03-13 02:39:20 <randy-waterhouse> disagree
676 2013-03-13 02:39:33 <gmaxwell> (and really even the major pools can use 0.8 now, but a bit better if they stay on 0.7.2 for the moment)
677 2013-03-13 02:40:03 <randy-waterhouse> gmaxwell: how do you mean "a bit better" ?
678 2013-03-13 02:40:47 <warren> sudofox: mobrqr83QWM9iaavMsrVHZ1Mi3WH334sv9
679 2013-03-13 02:42:17 <tre-spective> 182m1p8VoNMDJaF9EfPGu15uL4EiB92gmw
680 2013-03-13 02:42:28 <gmaxwell> randy-waterhouse: less risk of failure. There is still some risk from people mining with non-default settings on 0.8 until we get updates tested and out, that risk is removed by having a supermajority of hashpower stay on 0.7.2
681 2013-03-13 02:43:04 <drago777> what is testnet guys
682 2013-03-13 02:43:22 <randy-waterhouse> gmaxwell: thanks ... any ideas on what will the updates do specifically?
683 2013-03-13 02:43:34 <buddyrandom> https://en.bitcoin.it/wiki/Testnet
684 2013-03-13 02:43:47 Lannoc has quit ()
685 2013-03-13 02:43:50 <drago777> thanks buddy
686 2013-03-13 02:44:13 i2pRelay has quit (Remote host closed the connection)
687 2013-03-13 02:44:14 sudofox has quit (Ping timeout: 264 seconds)
688 2013-03-13 02:44:44 i2pRelay has joined
689 2013-03-13 02:45:30 <gmaxwell> randy-waterhouse: make 0.8 reject all the blocks 0.7.2 nodes reject potentially until some future date after which everyone will need to be updated.
690 2013-03-13 02:46:35 <randy-waterhouse> gmaxwell: so 0.8 (levelDB) will 'emulate' BDB lock limit behaviour ... boradly speaking?
691 2013-03-13 02:46:42 <randy-waterhouse> s/broadly
692 2013-03-13 02:47:43 <joe_k> well ive started reading a protocol document and writing an implementation in F#
693 2013-03-13 02:47:48 <gmaxwell> randy-waterhouse: Right. In order to make the rule simpler we'll likely have it more agressive and reject some things 0.7.2 would not, and also issue an 0.7.3 which imposes the same limit.
694 2013-03-13 02:48:23 <randy-waterhouse> ok, 0.7.3 that imposes the limits explicitly rather than by defualt ... ok, I'll be watching .. sounds like aplan
695 2013-03-13 02:48:55 johnsoft_ has joined
696 2013-03-13 02:48:58 <gmaxwell> (rejecting more is safeâ we call that a soft fork, and it's an okay thing to have so long as a supermajority of hashpower imposes it)
697 2013-03-13 02:48:59 <randy-waterhouse> can test also
698 2013-03-13 02:49:00 JWU42 has joined
699 2013-03-13 02:49:22 johnsoft_1 has quit (Ping timeout: 258 seconds)
700 2013-03-13 02:49:43 drapetomano has quit (Ping timeout: 245 seconds)
701 2013-03-13 02:50:03 <randy-waterhouse> now somme thought on fee structure is needed quickly imo so that these new limits can be enforced correctly without losing legitimate transactions to spam
702 2013-03-13 02:51:25 jankins has joined
703 2013-03-13 02:52:15 i2pRelay has quit (Remote host closed the connection)
704 2013-03-13 02:52:46 i2pRelay has joined
705 2013-03-13 02:55:07 undecim has quit (Ping timeout: 258 seconds)
706 2013-03-13 02:55:12 brwyatt is now known as Away!~brwyatt@brwyatt.net|brwyatt
707 2013-03-13 02:55:22 jankins has quit (Client Quit)
708 2013-03-13 02:57:18 Boydy has quit (Read error: Connection reset by peer)
709 2013-03-13 02:57:27 RBecker is now known as rbecker
710 2013-03-13 02:59:55 undecim has joined
711 2013-03-13 03:00:16 i2pRelay has quit (Remote host closed the connection)
712 2013-03-13 03:00:16 FredEE has quit (Quit: FredEE)
713 2013-03-13 03:00:49 i2pRelay has joined
714 2013-03-13 03:01:09 jMyles has quit (Ping timeout: 264 seconds)
715 2013-03-13 03:01:27 paybitcoin1 has quit (Read error: Connection reset by peer)
716 2013-03-13 03:01:39 jaeyjay has joined
717 2013-03-13 03:01:56 jaeyjay has quit (Client Quit)
718 2013-03-13 03:03:51 moarrr has quit ()
719 2013-03-13 03:05:47 klmist has joined
720 2013-03-13 03:06:07 tyn has joined
721 2013-03-13 03:07:01 ikeewee has quit (Quit: Page closed)
722 2013-03-13 03:07:13 Shrugged has quit (Quit: Shrugged)
723 2013-03-13 03:08:16 i2pRelay has quit (Remote host closed the connection)
724 2013-03-13 03:08:46 i2pRelay has joined
725 2013-03-13 03:10:34 tbcoin has quit (Remote host closed the connection)
726 2013-03-13 03:14:36 tyn has quit (Ping timeout: 250 seconds)
727 2013-03-13 03:16:16 nodroids has joined
728 2013-03-13 03:16:17 i2pRelay has quit (Remote host closed the connection)
729 2013-03-13 03:16:47 i2pRelay has joined
730 2013-03-13 03:17:22 canoon has quit (Ping timeout: 252 seconds)
731 2013-03-13 03:18:01 AMDchild__ has joined
732 2013-03-13 03:24:18 i2pRelay has quit (Remote host closed the connection)
733 2013-03-13 03:24:18 Derringer has quit (Read error: Connection reset by peer)
734 2013-03-13 03:24:49 i2pRelay has joined
735 2013-03-13 03:25:28 brson has quit (Quit: leaving)
736 2013-03-13 03:27:18 sudofox has joined
737 2013-03-13 03:27:44 <sudofox> Hi again
738 2013-03-13 03:27:50 Boydy_ has joined
739 2013-03-13 03:28:06 <sudofox> Still trying to give out TBTC lol
740 2013-03-13 03:28:18 <sudofox> Any other channels beside this and #bitcoin I can try?
741 2013-03-13 03:28:28 drivelights has left ("Leaving")
742 2013-03-13 03:30:27 gntbynynybt has joined
743 2013-03-13 03:30:48 <sudofox> Anyone want testnet btc for free? I have 3864.38399995 BTC.
744 2013-03-13 03:32:03 sudofox has quit (Client Quit)
745 2013-03-13 03:32:21 i2pRelay has quit (Remote host closed the connection)
746 2013-03-13 03:32:50 i2pRelay has joined
747 2013-03-13 03:33:27 <scintill> sudofox: I'll take some
748 2013-03-13 03:35:04 komododragon_ has joined
749 2013-03-13 03:35:12 tyn has joined
750 2013-03-13 03:35:51 Boydy_ has quit (Ping timeout: 240 seconds)
751 2013-03-13 03:36:37 <scintill> mvs8tdDDFBMoxMLstxLVUkDHMo3Sek8MNF - never used testnet before, but having some coins could be handy
752 2013-03-13 03:36:47 has quit (Clown|!Clown@static-87-79-93-140.netcologne.de|Remote host closed the connection)
753 2013-03-13 03:36:58 h4ckm3 has joined
754 2013-03-13 03:37:06 has joined
755 2013-03-13 03:38:27 MobGod has joined
756 2013-03-13 03:39:27 Ashaman has joined
757 2013-03-13 03:40:23 i2pRelay has quit (Remote host closed the connection)
758 2013-03-13 03:40:55 i2pRelay has joined
759 2013-03-13 03:42:31 donut123456789 has quit (Quit: Page closed)
760 2013-03-13 03:43:27 jine has quit (Ping timeout: 245 seconds)
761 2013-03-13 03:43:34 jine has joined
762 2013-03-13 03:44:16 <kuzetsa> scintill: well, there's #bitcoin-otc, channels for some bitcoin related projects (at least one exchange, and at least 1-2 pools) at least one or two linux distros have a channel on freenode... you didn't specify a whole lot of filters when you said "other channels beside this"
763 2013-03-13 03:44:25 <kuzetsa> ... freenode kinda has hundreds (err... thousands) of channels
764 2013-03-13 03:44:38 gasteve has joined
765 2013-03-13 03:44:42 zooko_afk has quit (Remote host closed the connection)
766 2013-03-13 03:44:59 darwin_ has quit (Remote host closed the connection)
767 2013-03-13 03:45:31 <scintill> kuzetsa: I didn't ask about the channels, sudofox did
768 2013-03-13 03:45:39 nicocesar has quit (Quit: Konversation terminated!)
769 2013-03-13 03:46:17 <kuzetsa> scintill: sorry, I just woke up and highlighted the wrong person... wand sudofox isn't in here now, huh.
770 2013-03-13 03:46:38 <kuzetsa> (Local NY time) [23:10:43] * sudofox (~sudofox@207.179.89.27) Quit (Client Quit) [23:12:07] <scintill> sudofox: I'll take some
771 2013-03-13 03:46:45 JDuke128 has quit (Quit: [BB])
772 2013-03-13 03:47:05 <kuzetsa> ad a couple minutes after that you posted a testnet address to send to :)
773 2013-03-13 03:47:09 <kuzetsa> *and
774 2013-03-13 03:47:25 DarkGhost is now known as DarkGhost`
775 2013-03-13 03:47:27 <amantonop> sudofox keeps coming and going
776 2013-03-13 03:47:43 Goonie has quit (Ping timeout: 245 seconds)
777 2013-03-13 03:47:44 <scintill> indeed. I didn't notice, I don't use irc much
778 2013-03-13 03:47:52 <amantonop> he may be back again
779 2013-03-13 03:48:17 <scintill> cool, thanks
780 2013-03-13 03:48:25 i2pRelay has quit (Remote host closed the connection)
781 2013-03-13 03:48:55 i2pRelay has joined
782 2013-03-13 03:49:27 <kuzetsa> gmaxwell ... what? there was talk about 0.7.3 while I napped?
783 2013-03-13 03:49:47 zooko has joined
784 2013-03-13 03:51:14 <scintill> it was about adding an explicit limit to a new 0.8.0 release, and putting the same one in a 0.7 release to make sure there aren't hidden bdb/leveldb discrepancies
785 2013-03-13 03:53:17 <kuzetsa> uhg :(
786 2013-03-13 03:54:10 jabawok has left ("Leaving")
787 2013-03-13 03:54:25 dawei101 has joined
788 2013-03-13 03:54:26 <amantonop> that was an interesting discussion
789 2013-03-13 03:54:55 fiesh has quit (Ping timeout: 272 seconds)
790 2013-03-13 03:54:59 <amantonop> essentially a filter that works in a predictable way in both versions
791 2013-03-13 03:55:46 <amantonop> it might be a good idea to have validation rules that are explicit, rather than probing the edge of the parameter space by trial and error
792 2013-03-13 03:55:55 AndChat has joined
793 2013-03-13 03:56:28 i2pRelay has quit (Remote host closed the connection)
794 2013-03-13 03:56:49 <amantonop> does anyone here know anything about history of alternate clients? Is there a python client that could be developed to be compatible with 0.7-0.8?
795 2013-03-13 03:56:58 i2pRelay has joined
796 2013-03-13 03:57:22 <amantonop> By client, I mean non-mining node.
797 2013-03-13 03:58:47 fiesh has joined
798 2013-03-13 03:59:11 <weex> i don't know of a full python client
799 2013-03-13 03:59:21 <weex> is the language the key thing?
800 2013-03-13 03:59:46 <amantonop> do you know of a anything-other-than-full python client?
801 2013-03-13 03:59:51 <weex> pynode is pretty good though for more network, db- level stuff
802 2013-03-13 04:00:09 <amantonop> I don't mean GUI, sorry... I mean bitcoind
803 2013-03-13 04:00:12 <amantonop> a node
804 2013-03-13 04:00:25 <amantonop> ah pynode
805 2013-03-13 04:00:37 <amantonop> I thought that was 2 years old. Have you any experience with pynode?
806 2013-03-13 04:01:11 beshini has quit (Quit: Page closed)
807 2013-03-13 04:01:23 <weex> i've run it, it's not 2 years old...hang on
808 2013-03-13 04:01:31 <amantonop> never mind, I see!
809 2013-03-13 04:01:37 <etotheipi_> amantonop: I don't know what you're looking for exactly
810 2013-03-13 04:01:44 <amantonop> I must've mixed them up. jeff is working on it I see....
811 2013-03-13 04:01:45 <etotheipi_> but Armory is kind of python
812 2013-03-13 04:01:59 <etotheipi_> I mean, it's fully python outer-layer, C++ under the hood for all the blockchain utilities
813 2013-03-13 04:02:04 komododragon_ has left ()
814 2013-03-13 04:02:06 <amantonop> I'm looking for infrastructure-class node, to build backend
815 2013-03-13 04:02:22 <amantonop> I need the opposite or just an inner core of Python
816 2013-03-13 04:02:39 <amantonop> I think pynode and python-bitcoinlib are what I'm looking for
817 2013-03-13 04:02:40 gasteve has quit (Quit: leaving)
818 2013-03-13 04:02:43 <amantonop> thank you weex!
819 2013-03-13 04:02:48 <weex> yw
820 2013-03-13 04:02:58 <amantonop> I thought those had been abandoned
821 2013-03-13 04:03:24 <amantonop> now I see it is garzik, so I guess the exact opposite of abandoned...
822 2013-03-13 04:04:09 OldEnK has joined
823 2013-03-13 04:04:31 i2pRelay has quit (Remote host closed the connection)
824 2013-03-13 04:05:02 i2pRelay has joined
825 2013-03-13 04:06:07 <amantonop> thanks etotheipi_
826 2013-03-13 04:06:29 Darkneo has joined
827 2013-03-13 04:06:32 <etotheipi_> amantonop: if you look at it, be aware that it does no validation at all
828 2013-03-13 04:06:41 <etotheipi_> it's intended to piggyback on bitcoind/bitcoin-qt
829 2013-03-13 04:06:53 Detritus has quit (Ping timeout: 245 seconds)
830 2013-03-13 04:07:03 <amantonop> etotheipi_ which are you referring to?
831 2013-03-13 04:07:22 <etotheipi_> Armory
832 2013-03-13 04:07:25 <etotheipi_> and armoryengine.py
833 2013-03-13 04:08:02 <amantonop> etotheipi_ it is precisely that issue I am looking to address. I want an alternative core protocol engine
834 2013-03-13 04:08:41 <kuzetsa> oh
835 2013-03-13 04:09:17 <kuzetsa> I don't know of any python-based bitcoin stuff which DOESN'T require the RPC api stuffs from the bitcoind / bitcoin-qt to provide the inner workings
836 2013-03-13 04:10:09 <amantonop> so this: https://github.com/jgarzik/python-bitcoinlib is RPC based?
837 2013-03-13 04:11:05 noosemoose has joined
838 2013-03-13 04:11:07 fooooooo has joined
839 2013-03-13 04:11:10 alna has joined
840 2013-03-13 04:11:32 <amantonop> ok got it, it is RPC based I see
841 2013-03-13 04:11:38 <amantonop> so that won't work either...
842 2013-03-13 04:11:59 zoinky has quit (Quit: Leaving.)
843 2013-03-13 04:12:15 <etotheipi_> amantonop: I guess I still don't understand.... you want a python library that does no validation, and not accessible via RPC?
844 2013-03-13 04:12:21 <amantonop> Am I to understand correctly that the only nodes running on the bitcoin network are based on bitcoind
845 2013-03-13 04:12:32 i2pRelay has quit (Remote host closed the connection)
846 2013-03-13 04:12:33 Detritus has joined
847 2013-03-13 04:12:41 <amantonop> I want a core protocol node that is *not* bitcoind
848 2013-03-13 04:13:04 i2pRelay has joined
849 2013-03-13 04:13:05 <amantonop> I want to talk bitcoin protocol with code that is not bitcoind code
850 2013-03-13 04:13:22 <kuzetsa> so you said language doesn't matter?
851 2013-03-13 04:13:27 <amantonop> I want a python re-write of bitcoind, which may also have an RPC interface or not
852 2013-03-13 04:13:48 <amantonop> depends on the language ;-)
853 2013-03-13 04:14:11 <etotheipi_> I mean, Armory speaks the language... it just has zero protections in place to protect itself from a malicious node
854 2013-03-13 04:14:23 <amantonop> sorry, I don't mean to sound demanding with all the "I want". I mean to say "I would like to know if such a thing exists before I go start writing it myself"
855 2013-03-13 04:14:44 <etotheipi_> it's not intended to be robust... but it does evaluate most scripts, it does very light network communication
856 2013-03-13 04:14:55 <kuzetsa> ... not bitcoind / bitcoin-qt under the hood or as a dependancy... huh, I'm really not sure that exists :|
857 2013-03-13 04:15:24 <etotheipi_> and it's very actively developed :)
858 2013-03-13 04:15:27 <amantonop> I see. I'm beginning to understand the whole argument in the forums
859 2013-03-13 04:16:23 <kuzetsa> etotheipi_: I thought armory required a node to talk to via RPC and couldn't natively use the bitcoin p2p to other nodes?
860 2013-03-13 04:16:26 <amantonop> etotheipi_ but it communicates via RPC with bitcoind underneath, correct?
861 2013-03-13 04:16:32 <etotheipi_> amantonop: no
862 2013-03-13 04:16:36 <etotheipi_> it connects as a P2P peer
863 2013-03-13 04:16:37 <kuzetsa> what? since when?
864 2013-03-13 04:16:40 <amantonop> or does it natively speak bitcoin protocol
865 2013-03-13 04:16:42 <etotheipi_> since always
866 2013-03-13 04:16:45 <etotheipi_> it's never used RPC
867 2013-03-13 04:17:00 gotnate has left ()
868 2013-03-13 04:17:02 <kuzetsa> back in june 2012 I was under the mistaken impression that armory required bitcoin-qt / bitcoind to be installed in order to use it?
869 2013-03-13 04:17:06 <amantonop> etotheipi_: ok thanks
870 2013-03-13 04:17:09 <etotheipi_> kuzetsa:
871 2013-03-13 04:17:14 <etotheipi_> it DOES use bitcoin-qt
872 2013-03-13 04:17:18 <etotheipi_> (or bitcoind)
873 2013-03-13 04:17:22 <amantonop> etotheipi_: I will look again, the design was pretty secure, I liked the principles
874 2013-03-13 04:17:27 <etotheipi_> but it doesn't communicate with it through RPC
875 2013-03-13 04:17:33 * kuzetsa facepalms
876 2013-03-13 04:17:40 <etotheipi_> it connects as a regular peer
877 2013-03-13 04:18:04 <kuzetsa> "it connects as a regular peer" --> so why doesn't it just bootstrap to the network like a regular
878 2013-03-13 04:18:09 John____ has joined
879 2013-03-13 04:18:13 <etotheipi_> kuzetsa: because it would die
880 2013-03-13 04:18:17 <kuzetsa> huh?
881 2013-03-13 04:18:23 <etotheipi_> it has no validation
882 2013-03-13 04:18:23 <kuzetsa> what's that supposed to mean?
883 2013-03-13 04:18:34 epylar has joined
884 2013-03-13 04:18:40 <etotheipi_> the code is written assuming that the single connected node is honest
885 2013-03-13 04:18:47 <etotheipi_> and filtering garbage for me
886 2013-03-13 04:19:05 <kuzetsa> ooooh, so it's not actually required, just "recommended because there aren't any sanity checks"
887 2013-03-13 04:19:12 <etotheipi_> basically
888 2013-03-13 04:19:14 <kuzetsa> the documentation for armory never made that clear
889 2013-03-13 04:19:34 <kuzetsa> ... kinda implied that using p2p network directly (bootstraping or otherwise) wasn't possible on a protocol level
890 2013-03-13 04:20:02 <etotheipi_> kuzetsa: you should *never* use Armory with any non-trusted peers
891 2013-03-13 04:20:04 <amantonop> etotheipi_: are you the author?
892 2013-03-13 04:20:10 <etotheipi_> amantonop: yes
893 2013-03-13 04:20:25 <etotheipi_> kuzetsa: I didn't even want to imply that was possible/feasible
894 2013-03-13 04:20:26 <amantonop> excellent
895 2013-03-13 04:20:30 <etotheipi_> because it's a recipe for disaster
896 2013-03-13 04:20:34 i2pRelay has quit (Remote host closed the connection)
897 2013-03-13 04:20:38 <gwillen> etotheipi_: tjat
898 2013-03-13 04:20:46 <amantonop> let me go look at the code
899 2013-03-13 04:20:47 Anduck has joined
900 2013-03-13 04:20:49 <kuzetsa> amantonop: yeah, etotheipi_ has a habit of centralized control over ideas / decisions to a greater extent than the bitcoind / bitcoin-qt team though :(
901 2013-03-13 04:20:49 <gwillen> etotheipi_: that's interesting -- I had also assumed it used RPC.
902 2013-03-13 04:20:54 <gwillen> you have hidden your secret well. ;-)
903 2013-03-13 04:20:55 <kuzetsa> *cough* wallet imports
904 2013-03-13 04:21:06 i2pRelay has joined
905 2013-03-13 04:21:12 Ashaman has quit (Read error: Connection reset by peer)
906 2013-03-13 04:21:27 <amantonop> we're only reliant on specific implementations because of a lack of robust protocol definition and portable protocol engine
907 2013-03-13 04:21:27 fooooooo has quit (Quit: Page closed)
908 2013-03-13 04:21:29 <etotheipi_> haha... well part of it was that I didn't want to mess with setting up the user's RPC interface
909 2013-03-13 04:21:30 molec has quit (Ping timeout: 260 seconds)
910 2013-03-13 04:21:50 <amantonop> so, if etotheipi_ is developing a *core engine*, he/she is a hero
911 2013-03-13 04:22:05 <etotheipi_> amantonop: I don't know what "core engine" means
912 2013-03-13 04:22:07 <amantonop> and opens the door for many implementations that are independent
913 2013-03-13 04:22:19 <etotheipi_> I think of it as an extension of Bitcoin_Qt/bitcoind
914 2013-03-13 04:22:45 <etotheipi_> (well, it doesn't feel like a simple extension, but I don't really do much duplication of work)
915 2013-03-13 04:22:59 <amantonop> Well... let me see the code... pardon me for a few min. Code speaks louder than IRC?
916 2013-03-13 04:23:10 <etotheipi_> hah, let me know if you have any questions
917 2013-03-13 04:23:56 <amantonop> you've all been superbly helpful, thank you! I will return with more questions!
918 2013-03-13 04:23:57 <etotheipi_> everything you're looking for is probably in armoryengine.py
919 2013-03-13 04:24:08 <amantonop> https://github.com/etotheipi/BitcoinArmory/blob/master/armoryengine.py
920 2013-03-13 04:24:15 <amantonop> I ended up there in three clicks ;-)
921 2013-03-13 04:24:24 <etotheipi_> just be aware there's 10k lines of C++ in the background
922 2013-03-13 04:24:32 gonzoflip has quit (Quit: Leaving.)
923 2013-03-13 04:24:37 <kuzetsa> etotheipi_: what license is armory anyway?
924 2013-03-13 04:24:42 <etotheipi_> AGPLv3
925 2013-03-13 04:24:47 TBZ1-SanDiego has joined
926 2013-03-13 04:24:53 <kuzetsa> like... bitcoind / bitcoin-qt is MIT so... hmm
927 2013-03-13 04:24:55 <kuzetsa> ok well huh
928 2013-03-13 04:25:40 <kuzetsa> I was entertaining the idea of ripping some of the working "engine" (what amantonop calls core) from bitcoin-qt / bitcoin-d and sticking it in armory to handle the "validation" which it lacks
929 2013-03-13 04:25:57 <etotheipi_> kuzetsa: I have considered that in the past
930 2013-03-13 04:26:13 <etotheipi_> kuzetsa: but the problem is that I don't want to be responsible for reintegrating everything every new release
931 2013-03-13 04:26:16 <kuzetsa> etotheipi_: I can't stand python though :(
932 2013-03-13 04:26:17 <etotheipi_> it creates a ton of work for me
933 2013-03-13 04:26:29 <etotheipi_> kuzetsa: that's a shame... it's the most delightful language ever created!
934 2013-03-13 04:26:31 <kuzetsa> yeah, good poiint
935 2013-03-13 04:26:36 <kuzetsa> what? no it's not
936 2013-03-13 04:26:39 * kuzetsa hisses
937 2013-03-13 04:26:44 <etotheipi_> haha
938 2013-03-13 04:27:35 <etotheipi_> I do credit its "success" on python
939 2013-03-13 04:27:48 <etotheipi_> getting features implemented is 4x more efficient than C++
940 2013-03-13 04:27:49 lb4956 has quit (Ping timeout: 255 seconds)
941 2013-03-13 04:28:37 i2pRelay has quit (Remote host closed the connection)
942 2013-03-13 04:28:56 <amantonop> etotheipi_: yup, python rocks
943 2013-03-13 04:29:07 i2pRelay has joined
944 2013-03-13 04:29:11 <amantonop> etotheipi_: you've done a very nice job with all the support functions in python too
945 2013-03-13 04:29:42 <etotheipi_> amantonop: except for one thing that still annoys me... endianness
946 2013-03-13 04:29:47 <etotheipi_> but thank you :)
947 2013-03-13 04:29:48 smelul8r_ has quit (Ping timeout: 245 seconds)
948 2013-03-13 04:30:22 <etotheipi_> very early on I realized that all serialization and everything seemed to be little-endian.... and then kind of based everything to be LE by default
949 2013-03-13 04:30:30 <etotheipi_> but then later found out everyone likes using BE
950 2013-03-13 04:30:41 <epylar> can the core not be written in C/C++ and armory call it via a python interface?
951 2013-03-13 04:31:02 <epylar> and then bitcoind/bitcoin-qt share this core code?
952 2013-03-13 04:31:06 <etotheipi_> all the blockchain scanning stuff is in C++, turned into a python-accessible module via swig
953 2013-03-13 04:31:08 <epylar> update in only one place
954 2013-03-13 04:31:11 <amantonop> etotheipi_: are you parsing the block chain index files that bitcoind creates?
955 2013-03-13 04:31:17 <etotheipi_> amantonop: yes
956 2013-03-13 04:31:25 <amantonop> Nice.
957 2013-03-13 04:31:28 <amantonop> So you're 50% there
958 2013-03-13 04:31:35 <randy-waterhouse> amantonop: oes anyone here know anything about history of alternate clients? Is there a python client that could be developed to be compatible with 0.7-0.8?"
959 2013-03-13 04:31:42 <randy-waterhouse> electrum server
960 2013-03-13 04:31:47 <randy-waterhouse> might be what you are looking for
961 2013-03-13 04:31:49 <etotheipi_> in fact, I do it every time... which was okay back in the day before multi-gigabyte blockfiles
962 2013-03-13 04:31:57 <amantonop> randy-waterhouse: Yes, I'm looking at Stratum too
963 2013-03-13 04:31:57 <kuzetsa> I'm a proponent of the "structured BASIC" family of languages <--- "the most delightful language(s) ever created!" (to gently modify your quote with with a plural "s" thingy)
964 2013-03-13 04:31:57 <randy-waterhouse> it uses ABE db I think
965 2013-03-13 04:32:10 <kuzetsa> etotheipi_ :)
966 2013-03-13 04:32:16 <randy-waterhouse> k, just so you know
967 2013-03-13 04:32:19 TBZ1-SanDiego has left ()
968 2013-03-13 04:32:32 <amantonop> randy-waterhouse: Have you used Stratum?
969 2013-03-13 04:32:40 bitcoin-dev623 has quit (Quit: Page closed)
970 2013-03-13 04:32:47 <etotheipi_> right now, I'm getting ready to start having Armory handle its own blockchain files
971 2013-03-13 04:32:47 <amantonop> randy-waterhouse: I am very impressed with the concept, have not looked at the server yet.
972 2013-03-13 04:32:59 <amantonop> etotheipi_: what data store do you plan to use?
973 2013-03-13 04:33:31 <etotheipi_> so far, leveldb is winning the simplicity/speed/bundle-ability/license race
974 2013-03-13 04:34:05 <amantonop> and if you ever need to import/export from >0.8 that's not a bad benefit either
975 2013-03-13 04:34:12 <epylar> as bitcoin grows, would it be reasonable to assume the block chain size will grow exponentially over time?
976 2013-03-13 04:34:23 smelul8r has joined
977 2013-03-13 04:34:39 <amantonop> epylar: only to an extent
978 2013-03-13 04:34:39 <etotheipi_> but I'm still open to suggestions... I just haven't found another embeddable DB engine that has all the right properties
979 2013-03-13 04:34:43 John____ has quit (Ping timeout: 245 seconds)
980 2013-03-13 04:35:00 <amantonop> epylar: it grows more than linear but not exponential to the number of transactions
981 2013-03-13 04:35:19 <amantonop> epylar: right now the # of transactions is grown expo, so... you can't really see the diff
982 2013-03-13 04:35:44 <amantonop> epylar: storage and bandwidth costs are dropping faster. Not a problem
983 2013-03-13 04:35:44 <epylar> so maybe n^(something) where something > 1
984 2013-03-13 04:35:58 <epylar> storage is growing ridiculously fast
985 2013-03-13 04:36:01 <epylar> i can't believe we have 4 TB disks
986 2013-03-13 04:36:19 <amantonop> I've been around since the 1KB days and punch card...
987 2013-03-13 04:36:25 <amantonop> so... yeah
988 2013-03-13 04:36:31 lb4956 has joined
989 2013-03-13 04:36:38 <epylar> my smallest disk was a 20 MB disk
990 2013-03-13 04:36:38 i2pRelay has quit (Remote host closed the connection)
991 2013-03-13 04:36:40 <amantonop> the one thing you can't apply Moore's law too is people
992 2013-03-13 04:37:03 <epylar> compute power still seems to be doubling on a reliable rate, for the time being, but only if you can run your algorithms across many cores
993 2013-03-13 04:37:06 i2pRelay has joined
994 2013-03-13 04:37:07 <etotheipi_> I'm pretty sure, in partso f the world, you can
995 2013-03-13 04:37:08 <epylar> or many machines
996 2013-03-13 04:37:41 <epylar> although in the past decade we've done a good job increasing instructions per clock using more transistors
997 2013-03-13 04:37:50 <epylar> but the future is really in large numbers of not very fast cores
998 2013-03-13 04:38:02 perezd has quit (Quit: perezd)
999 2013-03-13 04:38:11 watchcomp has quit (Read error: Connection reset by peer)
1000 2013-03-13 04:38:41 <HM> great, where can I invest?
1001 2013-03-13 04:38:46 <HM> we can get rich together
1002 2013-03-13 04:39:20 perezd has joined
1003 2013-03-13 04:39:27 <epylar> if there was anything to invest in, it would be in programming techniques/tools that let us use more cores
1004 2013-03-13 04:39:39 Tykling has quit (Ping timeout: 252 seconds)
1005 2013-03-13 04:40:00 <epylar> although i suppose handling interconnect is still an open problem
1006 2013-03-13 04:40:01 loltu has quit (Ping timeout: 252 seconds)
1007 2013-03-13 04:40:46 Darkneo1 has joined
1008 2013-03-13 04:41:26 Darkneo has quit (Ping timeout: 276 seconds)
1009 2013-03-13 04:41:33 Darkneo1 is now known as Darkneo
1010 2013-03-13 04:41:45 <HM> people were predicting single core performance ceilings when the pentium 4s hit 4 Ghz
1011 2013-03-13 04:41:47 <amantonop> had to step away, sorry
1012 2013-03-13 04:41:56 <HM> or whatever insane rates we were clocking in at
1013 2013-03-13 04:41:57 <epylar> HM: we're still stuck around 3-4 ghz
1014 2013-03-13 04:42:04 <epylar> we've increased cores and instructions per clock
1015 2013-03-13 04:42:07 <amantonop> people have been predicting performance ceiling every year since forever
1016 2013-03-13 04:42:10 <epylar> also added more cache and interconnects
1017 2013-03-13 04:42:13 <HM> I'm pretty sure someone once overclocked an Intel CPU to 5 Ghz
1018 2013-03-13 04:42:19 <epylar> order of magnitude
1019 2013-03-13 04:42:26 <etotheipi_> HM: they got an AMD chip up to 8GHz
1020 2013-03-13 04:42:38 <HM> even so, cycles per instruction have dropped, the world of single core performance didn't end
1021 2013-03-13 04:42:39 <epylar> most of the stuff bitcoin does that's expected to grow can pretty trivially be run in parallel.
1022 2013-03-13 04:42:45 <epylar> HM: true
1023 2013-03-13 04:42:58 <amantonop> etotheipi_: what are the limitations of Affero?
1024 2013-03-13 04:43:13 <amantonop> etotheipi_: i'm not familiar, is there a reason behind it?
1025 2013-03-13 04:43:21 <etotheipi_> amantonop: it's GPL with an extra clause for network-connected software
1026 2013-03-13 04:43:38 <etotheipi_> it's pretty restrictive ... if that's what you're looking for
1027 2013-03-13 04:43:48 <etotheipi_> (looking for me to tell you)
1028 2013-03-13 04:43:48 <HM> basically if you run a server that faces the public you must make it open source
1029 2013-03-13 04:44:08 <amantonop> "it" being what part of it?
1030 2013-03-13 04:44:23 <epylar> i would assume any part required to offer the same service including whatever changes you've made
1031 2013-03-13 04:44:40 i2pRelay has quit (Remote host closed the connection)
1032 2013-03-13 04:44:40 <HM> probably
1033 2013-03-13 04:44:54 <HM> licenses are a soulsuck
1034 2013-03-13 04:45:04 <amantonop> hypothetical: server based on a core engine on the network (open source part), as a library to a service (non open source), is that cool?
1035 2013-03-13 04:45:09 i2pRelay has joined
1036 2013-03-13 04:45:33 <randy-waterhouse> amantonop: spunds cool
1037 2013-03-13 04:45:36 <randy-waterhouse> sounds
1038 2013-03-13 04:45:37 <amantonop> like, you have Armoury GUI on top. If I built a different GUI or a web service or a shop or a re-mixer, using your network node
1039 2013-03-13 04:45:38 <epylar> it talks about Corresponding Source..
1040 2013-03-13 04:45:49 <amantonop> would I need to open source the stuff on top, or only the network node?
1041 2013-03-13 04:45:57 <epylar> this is the key part:
1042 2013-03-13 04:45:59 <epylar> The "Corresponding Source" for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities. However, it does not include the work's System Libraries, or general-purpose tools or generally available free programs which are used unmodified in performing those activities but which are not part
1043 2013-03-13 04:46:01 <randy-waterhouse> depends what it does most likely
1044 2013-03-13 04:46:06 <epylar> For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work.
1045 2013-03-13 04:46:39 <amantonop> ok, that sounds fair enugh
1046 2013-03-13 04:46:51 <amantonop> it's downstream code, not upstream code
1047 2013-03-13 04:47:11 <amantonop> If I build a poker site on top of it, you won't ask me to open source my poker engine, more or less.
1048 2013-03-13 04:47:13 <etotheipi_> amantonop: I'm pretty sure that as long as you open-source any changes you make to armoryengine, you can use it
1049 2013-03-13 04:47:25 <etotheipi_> but if you build your poker site into armoryengine... then yeah
1050 2013-03-13 04:47:29 <amantonop> hehe
1051 2013-03-13 04:47:38 <randy-waterhouse> and shared libraries
1052 2013-03-13 04:47:56 <HM> read the FSF FAQ
1053 2013-03-13 04:48:05 <HM> their out of license explanation is usually pretty good
1054 2013-03-13 04:48:05 <etotheipi_> if you are going to do any major modifications to it that you don't want to open-source, then you should come talk to me about dual-licensing :)
1055 2013-03-13 04:48:28 <amantonop> HM: This was the quick intro. My lawyer will be looking at licenses, of course
1056 2013-03-13 04:48:38 zoinky has joined
1057 2013-03-13 04:48:39 <amantonop> etotheipi_: cool, thanks
1058 2013-03-13 04:49:04 tyn has quit (Ping timeout: 250 seconds)
1059 2013-03-13 04:49:20 sawtooth has joined
1060 2013-03-13 04:49:25 <amantonop> does anyone know if ThomasV comes here?
1061 2013-03-13 04:49:42 hmmmstrange has joined
1062 2013-03-13 04:49:47 pacpac has joined
1063 2013-03-13 04:50:10 <amantonop> The electrum/Stratum approach is fascinating
1064 2013-03-13 04:50:12 <epylar> i wonder if tux gets nervous handling 77 million dollars worth of bitcoin transactions in a month
1065 2013-03-13 04:50:15 stretchwarren has quit (Ping timeout: 264 seconds)
1066 2013-03-13 04:51:52 <keystroke> topic needs updating
1067 2013-03-13 04:51:53 rcorreia has quit (Read error: Connection reset by peer)
1068 2013-03-13 04:51:57 loltu has joined
1069 2013-03-13 04:52:01 <amantonop> +1 on that
1070 2013-03-13 04:52:08 <amantonop> update the topic!!!!
1071 2013-03-13 04:52:17 <keystroke> we don't want people to panic :D
1072 2013-03-13 04:52:30 rcorreia has joined
1073 2013-03-13 04:52:42 i2pRelay has quit (Remote host closed the connection)
1074 2013-03-13 04:52:44 <amantonop> especially since by now there's nothing to panic about. At this point, if you didn't notice, you don't need to
1075 2013-03-13 04:53:13 i2pRelay has joined
1076 2013-03-13 04:53:28 Tykling has joined
1077 2013-03-13 04:53:37 stretchwarren has joined
1078 2013-03-13 04:53:59 OldEnK has quit (Quit: Leaving)
1079 2013-03-13 04:54:45 buddyrandom has quit (Ping timeout: 256 seconds)
1080 2013-03-13 04:55:30 protojay has joined
1081 2013-03-13 04:55:31 <kuzetsa> hmmmm
1082 2013-03-13 04:55:55 <hmmmstrange> ?
1083 2013-03-13 04:56:01 <kuzetsa> oh lol
1084 2013-03-13 04:56:11 <kuzetsa> I didn't mean to ... hi hmmmstrange
1085 2013-03-13 04:56:28 <kuzetsa> I'm just making "thinking" err... "sounds"
1086 2013-03-13 04:56:30 <hmmmstrange> hi
1087 2013-03-13 04:56:33 <randy-waterhouse> amantonop: join #electrum and/or electrum topic on bitcointalk.org to find ThomasV ....
1088 2013-03-13 04:56:49 <amantonop> etotheipi_: thank you again for your help and guidance. I'm still early-stage but I'm getting to know the code and community.
1089 2013-03-13 04:56:57 <amantonop> randy-waterhouse: thanks!
1090 2013-03-13 04:57:22 <kuzetsa> etotheipi_ there's no armory channel hmm
1091 2013-03-13 04:57:45 <etotheipi_> there is
1092 2013-03-13 04:57:48 <etotheipi_> #bitcoin-armory
1093 2013-03-13 04:57:51 <kuzetsa> oooh
1094 2013-03-13 04:57:52 <etotheipi_> there's not much activity on it
1095 2013-03-13 04:58:03 <etotheipi_> but I try to keep it open in case people wander over there
1096 2013-03-13 04:58:29 <etotheipi_> amantonop: best of luck to you... let me know if there's anything I can do to help you understand/use the code
1097 2013-03-13 04:58:33 sawtooth has quit (Ping timeout: 258 seconds)
1098 2013-03-13 04:59:22 TheSeven has quit (Disconnected by services)
1099 2013-03-13 04:59:32 [7] has joined
1100 2013-03-13 04:59:38 owowo has quit (Quit: sayonara)
1101 2013-03-13 05:00:29 grau has joined
1102 2013-03-13 05:00:43 i2pRelay has quit (Remote host closed the connection)
1103 2013-03-13 05:01:10 i2pRelay has joined
1104 2013-03-13 05:01:12 owowo has joined
1105 2013-03-13 05:01:45 Anduck has quit (Ping timeout: 245 seconds)
1106 2013-03-13 05:02:17 puhc_ has quit (Read error: Connection reset by peer)
1107 2013-03-13 05:02:21 TradeFortress has joined
1108 2013-03-13 05:02:27 hsmiths has quit (Read error: Connection reset by peer)
1109 2013-03-13 05:03:13 puhc_ has joined
1110 2013-03-13 05:03:23 gntbynynybt has quit (Read error: Connection reset by peer)
1111 2013-03-13 05:03:31 cads has joined
1112 2013-03-13 05:03:34 gntbynynybt has joined
1113 2013-03-13 05:04:56 dust-otc has joined
1114 2013-03-13 05:05:01 hsmiths has joined
1115 2013-03-13 05:06:34 kadoban has joined
1116 2013-03-13 05:06:57 noosemoose has quit (Remote host closed the connection)
1117 2013-03-13 05:08:45 i2pRelay has quit (Remote host closed the connection)
1118 2013-03-13 05:09:15 i2pRelay has joined
1119 2013-03-13 05:10:11 lb4956 has quit (Ping timeout: 276 seconds)
1120 2013-03-13 05:11:43 derfy_ is now known as derfy
1121 2013-03-13 05:12:46 <neofutur> amantonop: ehy not opensource you rpoker engine : https://github.com/hippich/Bitcoin-Poker-Room
1122 2013-03-13 05:12:55 <JyZyXEL> is there a limit protocol wise how many transactions a block can contain?
1123 2013-03-13 05:13:03 <neofutur> _why_ not
1124 2013-03-13 05:13:37 xempew has quit (Quit: xempew)
1125 2013-03-13 05:15:06 i2pRelay has quit (Remote host closed the connection)
1126 2013-03-13 05:16:53 amiller has quit (Changing host)
1127 2013-03-13 05:16:54 amiller has joined
1128 2013-03-13 05:17:17 i2pRelay has joined
1129 2013-03-13 05:18:12 stretchwarren has quit (Quit: Textual IRC Client: www.textualapp.com)
1130 2013-03-13 05:18:42 int03h has quit (Quit: ChatZilla 0.9.90 [Firefox 22.0a1/20130311030946])
1131 2013-03-13 05:19:14 lb4956 has joined
1132 2013-03-13 05:22:03 Namworld has quit ()
1133 2013-03-13 05:23:03 <amantonop> neofutur: I'm not building a poker engine, that was a hypothetical
1134 2013-03-13 05:23:26 <amantonop> neofutur: and I may open source everything (if I build a company), but I just wanted to understand the demarcation line.
1135 2013-03-13 05:24:18 epylar has quit (Ping timeout: 245 seconds)
1136 2013-03-13 05:24:49 i2pRelay has quit (Remote host closed the connection)
1137 2013-03-13 05:25:14 <grau> neofutur: you are building the poker engine for fun or is it a business?
1138 2013-03-13 05:25:23 zooko has quit (Ping timeout: 258 seconds)
1139 2013-03-13 05:25:24 i2pRelay has joined
1140 2013-03-13 05:25:51 copumpkin has quit (Ping timeout: 240 seconds)
1141 2013-03-13 05:25:59 tyn has joined
1142 2013-03-13 05:26:00 TheXev has joined
1143 2013-03-13 05:26:11 <neofutur> grau: I just tried it havent coded that, my github is https://github.com/neofutur , not https://github.com/hippich
1144 2013-03-13 05:26:23 <neofutur> http://betco.in/
1145 2013-03-13 05:26:35 copumpkin has joined
1146 2013-03-13 05:26:37 <neofutur> ( the poker site have been hacked )
1147 2013-03-13 05:27:17 TheXev has left ()
1148 2013-03-13 05:27:37 Lepton has quit (Ping timeout: 245 seconds)
1149 2013-03-13 05:27:40 <neofutur> I just play a little bit of poker sometimes on http://gw.gd/bc_poker
1150 2013-03-13 05:27:43 TBZ1-SanDiego has joined
1151 2013-03-13 05:27:49 TBZ1-SanDiego is now known as TheButterZone
1152 2013-03-13 05:28:32 <TheButterZone> it appears blockchain.info isnt allowing external pushtx, from the brainwallet.org github source, specifically
1153 2013-03-13 05:28:46 Cache_Money has joined
1154 2013-03-13 05:29:22 <grau> nefutur: yeah, also my passion
1155 2013-03-13 05:30:00 <TheButterZone> nevermind, after 4 or so attempts, it finally popped up. but no beep
1156 2013-03-13 05:30:16 <chrisb> some guy had an opensource poker engine in erlang
1157 2013-03-13 05:30:17 <neofutur> not a passion here, but an interesting game, not just randomness and gamble imo
1158 2013-03-13 05:30:48 jonwaller has joined
1159 2013-03-13 05:31:03 Burninate has left ("Leaving")
1160 2013-03-13 05:31:16 <doublec> chrisb: that was https://github.com/wagerlabs/openpoker
1161 2013-03-13 05:32:17 <grau> neufuture: yes, it is not a game of chance. I also do not play online, just face to face.
1162 2013-03-13 05:32:50 i2pRelay has quit (Remote host closed the connection)
1163 2013-03-13 05:32:51 <grau> neufutur: since I rather want to learn reading people than computing.
1164 2013-03-13 05:33:23 i2pRelay has joined
1165 2013-03-13 05:33:38 <phantomcircuit> anybody in here running testnet with 0.7.2?
1166 2013-03-13 05:34:14 <grau> I am however interested to explore AI questions it raises.
1167 2013-03-13 05:34:57 MasterDebater has joined
1168 2013-03-13 05:36:52 <MasterDebater> Did the event/fork really end?
1169 2013-03-13 05:37:32 <jonwaller> We should update the title.
1170 2013-03-13 05:38:18 <MasterDebater> So, is it safe to upgrade to 0.8 now?
1171 2013-03-13 05:38:39 <SomeoneWeird> not entirely sure, check the url in topic
1172 2013-03-13 05:38:41 <SomeoneWeird> might have more info
1173 2013-03-13 05:38:47 B0g4r7 has quit (Ping timeout: 276 seconds)
1174 2013-03-13 05:39:32 <MasterDebater> I did
1175 2013-03-13 05:39:55 <doublec> if you're not a miner you're fine to upgrade iirc
1176 2013-03-13 05:40:52 i2pRelay has quit (Remote host closed the connection)
1177 2013-03-13 05:41:22 i2pRelay has joined
1178 2013-03-13 05:42:09 [Author] has joined
1179 2013-03-13 05:43:16 amantonop has quit (Quit: amantonop)
1180 2013-03-13 05:44:07 TheButterZone has left ()
1181 2013-03-13 05:44:34 <neofutur> MasterDebater: http://www.reddit.com/r/Bitcoin/comments/1a51xx/now_that_its_over_the_blockchain_fork_explained/
1182 2013-03-13 05:46:19 <MasterDebater> So, it's a "yes".
1183 2013-03-13 05:48:05 B0g4r7 has joined
1184 2013-03-13 05:48:54 i2pRelay has quit (Remote host closed the connection)
1185 2013-03-13 05:49:11 brwyatt is now known as brwyatt|Away
1186 2013-03-13 05:49:28 i2pRelay has joined
1187 2013-03-13 05:50:04 pacpac has quit (Ping timeout: 264 seconds)
1188 2013-03-13 05:52:52 pacpac has joined
1189 2013-03-13 05:56:55 i2pRelay has quit (Remote host closed the connection)
1190 2013-03-13 05:57:00 <nanotube> MasterDebater: yes
1191 2013-03-13 05:57:28 i2pRelay has joined
1192 2013-03-13 05:58:34 <randy-waterhouse> MasterDebater: it will still work but if another fork happens you will be on a different branch to what the majority of mining nodes is using
1193 2013-03-13 05:58:42 Diablo-D3 has quit (Quit: This computer has gone to sleep)
1194 2013-03-13 05:59:12 <randy-waterhouse> just so you know
1195 2013-03-13 05:59:17 <kuzetsa> haha I'm just remembering someone in my guild in everquest (like 10 years ago when it was still a common mmorpg) someone in our guild had the custom title "masterbaiter" ... like... because they had maxed out fishing
1196 2013-03-13 05:59:34 <kuzetsa> :) hi MasterDebater
1197 2013-03-13 05:59:57 dgriffi has quit (Ping timeout: 252 seconds)
1198 2013-03-13 06:00:58 <MasterDebater> I must debate!
1199 2013-03-13 06:01:43 dgriffi has joined
1200 2013-03-13 06:04:33 BTCTrader2 has joined
1201 2013-03-13 06:04:57 i2pRelay has quit (Remote host closed the connection)
1202 2013-03-13 06:05:27 i2pRelay has joined
1203 2013-03-13 06:08:52 CodeShark has quit (Remote host closed the connection)
1204 2013-03-13 06:12:14 <nanotube> randy-waterhouse: as long as mining power is on 0.7, it'll only be a 1-deep fork, so irrelevant.
1205 2013-03-13 06:12:27 CodeShark has joined
1206 2013-03-13 06:13:08 i2pRelay has quit (Remote host closed the connection)
1207 2013-03-13 06:13:41 Cache_Money has quit (Ping timeout: 258 seconds)
1208 2013-03-13 06:13:52 i2pRelay has joined
1209 2013-03-13 06:15:34 <randy-waterhouse> nanotube: I would put 0.8 upgrades on hold ... in case their occur more unknown unknowns
1210 2013-03-13 06:15:43 <randy-waterhouse> seems prudent
1211 2013-03-13 06:16:09 <nanotube> yea i guess that can't hurt. but "as far as we know" it should be fine :)
1212 2013-03-13 06:16:33 <randy-waterhouse> as far we knew it was fine until block 245430 yesterday ....
1213 2013-03-13 06:16:47 <randy-waterhouse> 225430 rather
1214 2013-03-13 06:17:07 <randy-waterhouse> but 0.7 is killing my hard drives ...
1215 2013-03-13 06:17:31 moleccc has joined
1216 2013-03-13 06:18:13 <nanotube> hm, i had no prob with 0.7.2. only upgraded to see "what all the hype was about" :)
1217 2013-03-13 06:18:40 pacpac has quit (Ping timeout: 258 seconds)
1218 2013-03-13 06:19:10 amiller has quit (Quit: ZNC - http://znc.sourceforge.net)
1219 2013-03-13 06:20:41 kalleboo is now known as kalleboo|away
1220 2013-03-13 06:20:48 owowo has quit (Quit: sayonara)
1221 2013-03-13 06:21:03 i2pRelay has quit (Remote host closed the connection)
1222 2013-03-13 06:21:12 amiller has joined
1223 2013-03-13 06:21:45 i2pRelay has joined
1224 2013-03-13 06:25:32 Nicksasa has joined
1225 2013-03-13 06:25:44 MWNinja has joined
1226 2013-03-13 06:26:37 LacLeman has quit (Quit: Computer has gone to sleep.)
1227 2013-03-13 06:26:45 thefinn93 has left ("WeeChat 0.4.0")
1228 2013-03-13 06:26:49 BTCTrader2 has quit (Quit: BTCTrader2)
1229 2013-03-13 06:27:19 yeahrock has joined
1230 2013-03-13 06:27:25 panzer has quit (Ping timeout: 256 seconds)
1231 2013-03-13 06:28:35 panzer has joined
1232 2013-03-13 06:29:14 i2pRelay has quit (Remote host closed the connection)
1233 2013-03-13 06:29:17 gwinter has joined
1234 2013-03-13 06:29:51 i2pRelay has joined
1235 2013-03-13 06:30:02 tyn has quit (Ping timeout: 250 seconds)
1236 2013-03-13 06:32:23 easye has joined
1237 2013-03-13 06:32:52 panzer has quit (Ping timeout: 252 seconds)
1238 2013-03-13 06:33:45 panzer has joined
1239 2013-03-13 06:34:00 gwinter has left ()
1240 2013-03-13 06:36:15 lb4956 has quit (Ping timeout: 252 seconds)
1241 2013-03-13 06:37:11 i2pRelay has quit (Remote host closed the connection)
1242 2013-03-13 06:37:40 i2pRelay has joined
1243 2013-03-13 06:39:53 ThomasV has joined
1244 2013-03-13 06:40:33 nodroids has quit (Ping timeout: 245 seconds)
1245 2013-03-13 06:41:01 Conflict has joined
1246 2013-03-13 06:41:03 MasterDebater has quit (Quit: Leaving)
1247 2013-03-13 06:41:41 Ant0 has quit (Ping timeout: 252 seconds)
1248 2013-03-13 06:43:40 Ashaman has joined
1249 2013-03-13 06:45:06 i2pRelay has quit (Remote host closed the connection)
1250 2013-03-13 06:45:35 i2pRelay has joined
1251 2013-03-13 06:47:47 snakie has quit (Ping timeout: 255 seconds)
1252 2013-03-13 06:48:43 shlm has quit (Ping timeout: 252 seconds)
1253 2013-03-13 06:49:11 ThomasV has quit (Quit: Quitte)
1254 2013-03-13 06:52:20 RazielZ has joined
1255 2013-03-13 06:53:09 i2pRelay has quit (Remote host closed the connection)
1256 2013-03-13 06:53:39 i2pRelay has joined
1257 2013-03-13 06:54:12 shlm has joined
1258 2013-03-13 06:56:06 BTCTrader2 has joined
1259 2013-03-13 06:56:54 FredEE has joined
1260 2013-03-13 06:57:32 johnsoft_ is now known as johnsoft
1261 2013-03-13 06:58:51 perezd has quit (Quit: perezd)
1262 2013-03-13 06:59:09 yeahrock_ has joined
1263 2013-03-13 07:00:27 yeahrock has quit (Ping timeout: 264 seconds)
1264 2013-03-13 07:00:27 yeahrock_ is now known as yeahrock
1265 2013-03-13 07:01:11 i2pRelay has quit (Remote host closed the connection)
1266 2013-03-13 07:01:51 i2pRelay has joined
1267 2013-03-13 07:09:13 i2pRelay has quit (Remote host closed the connection)
1268 2013-03-13 07:09:46 i2pRelay has joined
1269 2013-03-13 07:14:11 defunctzombie is now known as defunctzombie_zz
1270 2013-03-13 07:14:40 o8gf5 has quit (Ping timeout: 250 seconds)
1271 2013-03-13 07:17:16 i2pRelay has quit (Remote host closed the connection)
1272 2013-03-13 07:17:46 i2pRelay has joined
1273 2013-03-13 07:24:19 zappsio has joined
1274 2013-03-13 07:25:17 i2pRelay has quit (Remote host closed the connection)
1275 2013-03-13 07:25:36 whiterabbit has joined
1276 2013-03-13 07:25:46 keystroke has quit ()
1277 2013-03-13 07:25:48 i2pRelay has joined
1278 2013-03-13 07:26:37 saulimus has joined
1279 2013-03-13 07:27:03 <zappsio> 0.8 is crashing on Mac when reindexing blocks. Too many open files. Looking at it in Activity Monitor, it doesn't appear to be closing any block or chain state .sst files.
1280 2013-03-13 07:27:21 wrabbit has quit (Ping timeout: 276 seconds)
1281 2013-03-13 07:27:23 whiterabbit is now known as wrabbit
1282 2013-03-13 07:28:05 cads has left ("Leaving")
1283 2013-03-13 07:29:01 t4nk084 has joined
1284 2013-03-13 07:29:02 t4nk084 has quit (Client Quit)
1285 2013-03-13 07:29:29 karmic has quit ()
1286 2013-03-13 07:29:51 dimethylZ has quit (Remote host closed the connection)
1287 2013-03-13 07:29:53 t4nk444 has joined
1288 2013-03-13 07:30:29 tockitj has joined
1289 2013-03-13 07:31:04 derfy has quit (Quit: ZNC - http://znc.in)
1290 2013-03-13 07:31:34 derfy has joined
1291 2013-03-13 07:32:07 nethershaw has quit (Read error: Connection reset by peer)
1292 2013-03-13 07:32:23 nethershaw has joined
1293 2013-03-13 07:33:18 i2pRelay has quit (Remote host closed the connection)
1294 2013-03-13 07:33:49 grau has quit (Remote host closed the connection)
1295 2013-03-13 07:33:51 i2pRelay has joined
1296 2013-03-13 07:37:30 Squidicuz has quit (Quit: Leaving)
1297 2013-03-13 07:37:36 <t4nk444> Good job, all you!
1298 2013-03-13 07:38:36 t4nk444 has quit (Quit: Page closed)
1299 2013-03-13 07:39:04 <zappsio> any devs about? the Mac client is crashing
1300 2013-03-13 07:39:22 amantonop has joined
1301 2013-03-13 07:41:19 i2pRelay has quit (Remote host closed the connection)
1302 2013-03-13 07:41:32 Ashaman has quit (Read error: Connection reset by peer)
1303 2013-03-13 07:41:39 yeahrock has quit (Ping timeout: 276 seconds)
1304 2013-03-13 07:41:48 i2pRelay has joined
1305 2013-03-13 07:42:13 witwit has quit (Ping timeout: 245 seconds)
1306 2013-03-13 07:43:01 davos has quit ()
1307 2013-03-13 07:44:33 <BTC_Bear> zappsio: did you get EXC_BAD_ACCESS (SIGSEGV)
1308 2013-03-13 07:45:16 <zappsio> BTC_Bear: Too many open files
1309 2013-03-13 07:45:30 <zappsio> BTC_Bear: 0.8 is crashing on Mac when reindexing blocks. Too many open files. Looking at it in Activity Monitor, it doesn't appear to be closing any block or chain state .sst files.
1310 2013-03-13 07:46:14 whiterabbit has joined
1311 2013-03-13 07:46:17 wubino has quit (Quit: Leaving)
1312 2013-03-13 07:46:18 <zappsio> BTC_Bear: I've upped kern.maxfiles and ulimit file limits so I can get through a reindex.
1313 2013-03-13 07:46:53 <BTC_Bear> I had a KERN_INVALID_ADDRESS at 0x00000000a02c7ac5 which closed the -qt client
1314 2013-03-13 07:47:30 wrabbit has quit (Ping timeout: 276 seconds)
1315 2013-03-13 07:47:32 whiterabbit is now known as wrabbit
1316 2013-03-13 07:47:47 Ant0 has joined
1317 2013-03-13 07:49:08 ciphermonk has joined
1318 2013-03-13 07:49:09 johnsoft has quit (Read error: Connection reset by peer)
1319 2013-03-13 07:49:11 kalleboo is now known as away!~kalleboo@i125-204-93-9.s11.a046.ap.plala.or.jp|kalleboo
1320 2013-03-13 07:49:11 kalleboo has quit (Excess Flood)
1321 2013-03-13 07:49:20 i2pRelay has quit (Remote host closed the connection)
1322 2013-03-13 07:49:22 FredEE has quit (Quit: FredEE)
1323 2013-03-13 07:49:32 <BTC_Bear> zappsio: what version OSX?
1324 2013-03-13 07:49:59 i2pRelay has joined
1325 2013-03-13 07:49:59 <zappsio> Latest. 10.8.2
1326 2013-03-13 07:50:10 <BTC_Bear> Yep... ok thnx
1327 2013-03-13 07:50:34 n1c has joined
1328 2013-03-13 07:53:01 Hashdog has joined
1329 2013-03-13 07:53:01 safra has joined
1330 2013-03-13 07:53:44 <zappsio> BTC_Bear: I had a bunch of other stuff running the first time it happened, so I rebooted and tried again, running only the client. Happened a second time. Looked at log, saw "too many open files". Now trying for a third time with ridiculously high limits.
1331 2013-03-13 07:55:35 andrew12 has quit (Ping timeout: 272 seconds)
1332 2013-03-13 07:55:35 bitcoin-dev000 has joined
1333 2013-03-13 07:56:02 Duly has quit (Read error: Connection reset by peer)
1334 2013-03-13 07:57:05 mogri has quit (Ping timeout: 255 seconds)
1335 2013-03-13 07:57:08 <zappsio> BTC_Bear: I should maybe mention that previously I was running version 0.6.3. Hadn't used it for a month or so. So this is an upgrade from 0.6.3 to 0.8.
1336 2013-03-13 07:57:22 BTC_Bear has quit (Ping timeout: 248 seconds)
1337 2013-03-13 07:57:23 i2pRelay has quit (Remote host closed the connection)
1338 2013-03-13 07:57:55 i2pRelay has joined
1339 2013-03-13 08:00:16 buddyrandom has joined
1340 2013-03-13 08:02:47 Darkneo1 has joined
1341 2013-03-13 08:03:09 Goonie has joined
1342 2013-03-13 08:04:23 Darkneo has quit (Ping timeout: 256 seconds)
1343 2013-03-13 08:05:13 Ant0 has quit (Ping timeout: 258 seconds)
1344 2013-03-13 08:05:27 i2pRelay has quit (Remote host closed the connection)
1345 2013-03-13 08:05:35 bitafterbit has quit (Read error: Connection reset by peer)
1346 2013-03-13 08:05:59 i2pRelay has joined
1347 2013-03-13 08:06:06 Mandrius has joined
1348 2013-03-13 08:06:30 Mandrius is now known as Guest38056
1349 2013-03-13 08:06:46 zappsio has left ()
1350 2013-03-13 08:13:14 JDuke128 has joined
1351 2013-03-13 08:13:29 i2pRelay has quit (Remote host closed the connection)
1352 2013-03-13 08:13:44 Boydy_ has joined
1353 2013-03-13 08:14:01 i2pRelay has joined
1354 2013-03-13 08:14:30 <muhoo> has mining gone slightly insane on testnet3? seems like blocks are being mined every 30 seconds
1355 2013-03-13 08:14:33 oleganza has joined
1356 2013-03-13 08:14:53 mappum has quit (Ping timeout: 252 seconds)
1357 2013-03-13 08:14:54 <muhoo> https://www.refheap.com/paste/12483
1358 2013-03-13 08:15:11 lantastic has joined
1359 2013-03-13 08:16:31 oleganza has quit (Client Quit)
1360 2013-03-13 08:17:33 Ant0 has joined
1361 2013-03-13 08:18:15 zoinky has quit (Quit: Leaving.)
1362 2013-03-13 08:20:33 Grishnakh_ has quit (Read error: Connection reset by peer)
1363 2013-03-13 08:21:20 * muhoo backscrolls
1364 2013-03-13 08:21:32 Guest38056 is now known as Mandrius
1365 2013-03-13 08:21:32 i2pRelay has quit (Remote host closed the connection)
1366 2013-03-13 08:22:04 i2pRelay has joined
1367 2013-03-13 08:22:28 Hashdog has quit (Remote host closed the connection)
1368 2013-03-13 08:23:01 scintill has quit (Ping timeout: 260 seconds)
1369 2013-03-13 08:25:04 BTCTrader2 has quit (Quit: BTCTrader2)
1370 2013-03-13 08:25:33 BTCTrader2 has joined
1371 2013-03-13 08:27:56 thepok has joined
1372 2013-03-13 08:28:29 protojay has quit (Read error: Connection reset by peer)
1373 2013-03-13 08:28:54 protojay has joined
1374 2013-03-13 08:29:04 protojay is now known as moarrr
1375 2013-03-13 08:29:06 toffoo has quit ()
1376 2013-03-13 08:29:29 necrobunny has joined
1377 2013-03-13 08:29:34 i2pRelay has quit (Remote host closed the connection)
1378 2013-03-13 08:29:42 <necrobunny> what happened to bitcoins
1379 2013-03-13 08:30:05 i2pRelay has joined
1380 2013-03-13 08:30:47 <jgarzik> muhoo: sigh. might have to reset, if some miner insists upon spamming.
1381 2013-03-13 08:33:50 OPrime has joined
1382 2013-03-13 08:34:06 pete2 has quit (Ping timeout: 260 seconds)
1383 2013-03-13 08:34:59 amantonop has quit (Quit: amantonop)
1384 2013-03-13 08:35:07 testnode9 has quit (Ping timeout: 245 seconds)
1385 2013-03-13 08:35:34 testnode9 has joined
1386 2013-03-13 08:36:16 ovidiusoft has joined
1387 2013-03-13 08:36:41 tvisdog has quit (Ping timeout: 255 seconds)
1388 2013-03-13 08:36:42 pete2 has joined
1389 2013-03-13 08:37:08 <petertodd> Odd, just lots of 0.001 value txouts from what I see, but not enough to call it a proper attack test or anything.
1390 2013-03-13 08:37:21 sgornick has quit (Ping timeout: 240 seconds)
1391 2013-03-13 08:37:25 necrobunny has left ()
1392 2013-03-13 08:37:35 i2pRelay has quit (Remote host closed the connection)
1393 2013-03-13 08:38:09 i2pRelay has joined
1394 2013-03-13 08:39:17 buddyrandom has quit (Quit: Leaving)
1395 2013-03-13 08:40:16 sgornick has joined
1396 2013-03-13 08:43:24 sacredchao has quit (Ping timeout: 276 seconds)
1397 2013-03-13 08:43:35 LightRider has quit ()
1398 2013-03-13 08:44:08 sacredchao has joined
1399 2013-03-13 08:44:12 prolxray has quit ()
1400 2013-03-13 08:44:19 Ant0 has quit (Ping timeout: 258 seconds)
1401 2013-03-13 08:45:38 i2pRelay has quit (Remote host closed the connection)
1402 2013-03-13 08:46:07 i2pRelay has joined
1403 2013-03-13 08:46:30 vigilyn3 has joined
1404 2013-03-13 08:46:56 vigilyn2 has quit (Read error: Connection reset by peer)
1405 2013-03-13 08:47:00 ^V has quit ()
1406 2013-03-13 08:47:43 nsillik_ has quit (Ping timeout: 252 seconds)
1407 2013-03-13 08:48:17 <Mandrius> h
1408 2013-03-13 08:49:16 da2ce7 has joined
1409 2013-03-13 08:50:50 undecim has quit (Ping timeout: 258 seconds)
1410 2013-03-13 08:51:48 nsillik_ has joined
1411 2013-03-13 08:53:39 i2pRelay has quit (Remote host closed the connection)
1412 2013-03-13 08:53:55 midnightmagic has quit (Read error: Operation timed out)
1413 2013-03-13 08:54:06 i2pRelay has joined
1414 2013-03-13 08:54:36 Ant0 has joined
1415 2013-03-13 08:55:36 TD_ has joined
1416 2013-03-13 08:55:44 midnightmagic has joined
1417 2013-03-13 08:55:53 ocminer has quit (Read error: Operation timed out)
1418 2013-03-13 08:56:13 ocminer has joined
1419 2013-03-13 08:57:10 fishfish has joined
1420 2013-03-13 08:57:32 kazimir42 has joined
1421 2013-03-13 08:58:27 testnode9 has quit (Ping timeout: 245 seconds)
1422 2013-03-13 09:01:00 testnode9 has joined
1423 2013-03-13 09:01:40 i2pRelay has quit (Remote host closed the connection)
1424 2013-03-13 09:02:13 i2pRelay has joined
1425 2013-03-13 09:02:35 ocminer has quit (Ping timeout: 250 seconds)
1426 2013-03-13 09:04:12 da2ce7 has quit (Ping timeout: 276 seconds)
1427 2013-03-13 09:04:14 <EvilPete> Hmm. Deepbit still seems to be produing blocks on the orphaned 0.8 chain, is this known?
1428 2013-03-13 09:04:29 <[Tycho]> I don't think so.
1429 2013-03-13 09:04:29 BlackPrapor has joined
1430 2013-03-13 09:04:44 midnightmagic has quit (Ping timeout: 250 seconds)
1431 2013-03-13 09:04:54 grau has joined
1432 2013-03-13 09:05:07 <EvilPete> gah, misread UTC timestamp. never mind me.
1433 2013-03-13 09:05:22 <[Tycho]> Also it wasn't 0.8 chain, it was our "own" branch :)
1434 2013-03-13 09:05:37 <randy-waterhouse> heh
1435 2013-03-13 09:06:31 oracleCoin has joined
1436 2013-03-13 09:06:39 <EvilPete> I bet that caused a few wtf's..
1437 2013-03-13 09:06:48 snakie has joined
1438 2013-03-13 09:09:47 i2pRelay has quit (Remote host closed the connection)
1439 2013-03-13 09:10:37 i2pRelay has joined
1440 2013-03-13 09:10:46 midnightmagic has joined
1441 2013-03-13 09:12:14 stalled_ has quit (Ping timeout: 255 seconds)
1442 2013-03-13 09:14:39 manet_ has quit (Ping timeout: 252 seconds)
1443 2013-03-13 09:15:19 fishfish is now known as fishfish|AFK
1444 2013-03-13 09:16:14 <UukGoblin> hm, block download stopped at 225459 on 0.8.0 (I haven't run it for a ~week)
1445 2013-03-13 09:16:16 JZavala has joined
1446 2013-03-13 09:16:48 <jgarzik> UukGoblin: stopped, for how long? sometimes it will get stuck, and then un-stuck
1447 2013-03-13 09:16:56 banghouse has quit (Remote host closed the connection)
1448 2013-03-13 09:16:57 <UukGoblin> jgarzik, few minutes at most
1449 2013-03-13 09:17:28 TD_ has quit (Quit: TD_)
1450 2013-03-13 09:17:42 <jgarzik> UukGoblin: typically you would want to wait 1-2 network blocks before calling it stuck. I presume you have plenty of peers connected, not just 1 or 2?
1451 2013-03-13 09:17:46 i2pRelay has quit (Remote host closed the connection)
1452 2013-03-13 09:18:18 i2pRelay has joined
1453 2013-03-13 09:18:57 dlunch has quit (Remote host closed the connection)
1454 2013-03-13 09:19:01 fishfish is now known as AFK!~fishfish2@109.238.67.130|fishfish
1455 2013-03-13 09:19:22 dlunch has joined
1456 2013-03-13 09:20:52 <UukGoblin> got 8, yeah
1457 2013-03-13 09:21:04 <UukGoblin> k, restarted it and it seems un-stuck now
1458 2013-03-13 09:21:37 yesminister has joined
1459 2013-03-13 09:24:15 manet has joined
1460 2013-03-13 09:25:45 ovidiusoft has quit (Ping timeout: 264 seconds)
1461 2013-03-13 09:25:46 Darkneo1 is now known as Darkneo
1462 2013-03-13 09:25:48 i2pRelay has quit (Remote host closed the connection)
1463 2013-03-13 09:26:21 i2pRelay has joined
1464 2013-03-13 09:27:07 ovidiusoft has joined
1465 2013-03-13 09:29:33 andzib has joined
1466 2013-03-13 09:30:53 nelisky has joined
1467 2013-03-13 09:31:09 t7 has joined
1468 2013-03-13 09:33:50 i2pRelay has quit (Remote host closed the connection)
1469 2013-03-13 09:34:14 kwatz has joined
1470 2013-03-13 09:34:22 i2pRelay has joined
1471 2013-03-13 09:34:57 pete3 has joined
1472 2013-03-13 09:36:27 stalled has joined
1473 2013-03-13 09:38:18 kwatz is now known as destructure
1474 2013-03-13 09:38:25 destructure has quit (Changing host)
1475 2013-03-13 09:38:25 destructure has joined
1476 2013-03-13 09:38:29 pete2 has quit (Ping timeout: 248 seconds)
1477 2013-03-13 09:38:40 tonikt has joined
1478 2013-03-13 09:41:53 i2pRelay has quit (Remote host closed the connection)
1479 2013-03-13 09:42:25 i2pRelay has joined
1480 2013-03-13 09:43:49 btcur has joined
1481 2013-03-13 09:44:06 pete78 has joined
1482 2013-03-13 09:47:39 blkashdla has quit ()
1483 2013-03-13 09:48:12 newbie has joined
1484 2013-03-13 09:48:19 Linkandzelda has joined
1485 2013-03-13 09:48:21 Neskia has joined
1486 2013-03-13 09:49:14 pierre`_ has joined
1487 2013-03-13 09:49:31 o2_ has joined
1488 2013-03-13 09:49:50 n0g4n0o has joined
1489 2013-03-13 09:49:55 i2pRelay has quit (Remote host closed the connection)
1490 2013-03-13 09:49:57 Leviathanzz has joined
1491 2013-03-13 09:49:59 sl1982 has joined
1492 2013-03-13 09:50:02 thepok2222 has joined
1493 2013-03-13 09:50:03 dawei101 has quit (Read error: Connection reset by peer)
1494 2013-03-13 09:50:12 dawei101 has joined
1495 2013-03-13 09:50:19 fposdkgs has joined
1496 2013-03-13 09:50:25 i2pRelay has joined
1497 2013-03-13 09:50:29 OPrime_ has joined
1498 2013-03-13 09:50:36 sdfsdf has joined
1499 2013-03-13 09:50:38 jayne_ has joined
1500 2013-03-13 09:51:18 a5m0_ has joined
1501 2013-03-13 09:51:24 Spooktooth has joined
1502 2013-03-13 09:51:49 enquirer4 has joined
1503 2013-03-13 09:51:53 dawei101 has quit (Read error: Connection reset by peer)
1504 2013-03-13 09:51:59 dawei101 has joined
1505 2013-03-13 09:52:17 intx_ has joined
1506 2013-03-13 09:52:29 dawei101 has quit (Client Quit)
1507 2013-03-13 09:52:46 jaromil has joined
1508 2013-03-13 09:52:52 doublec_ has joined
1509 2013-03-13 09:52:58 gonffen_ has joined
1510 2013-03-13 09:53:11 gwollon has joined
1511 2013-03-13 09:53:12 gyver has joined
1512 2013-03-13 09:53:13 gwollon has quit (Changing host)
1513 2013-03-13 09:53:13 gwollon has joined
1514 2013-03-13 09:53:26 pizzaman2337 has joined
1515 2013-03-13 09:53:29 nowan_ has joined
1516 2013-03-13 09:53:41 nexes_ has joined
1517 2013-03-13 09:54:05 Conflict has quit (Ping timeout: 258 seconds)
1518 2013-03-13 09:55:39 Boydy_ has quit (Remote host closed the connection)
1519 2013-03-13 09:56:46 cyphase has quit (Read error: Operation timed out)
1520 2013-03-13 09:57:22 swhitt_ has joined
1521 2013-03-13 09:57:45 swhitt_ is now known as swhitt
1522 2013-03-13 09:57:45 OPrime_ is now known as OPrime
1523 2013-03-13 09:57:45 intx_ is now known as intx
1524 2013-03-13 09:57:45 intx has quit (Changing host)
1525 2013-03-13 09:57:56 i2pRelay has quit (Remote host closed the connection)
1526 2013-03-13 09:58:03 moarrr has quit (Read error: Connection reset by peer)
1527 2013-03-13 09:58:20 moarrr has joined
1528 2013-03-13 09:58:25 thepok2222 has quit (Ping timeout: 245 seconds)
1529 2013-03-13 09:58:26 i2pRelay has joined
1530 2013-03-13 09:58:31 intx has joined
1531 2013-03-13 10:00:16 wyrag has joined
1532 2013-03-13 10:00:22 unknown45682 has quit (Ping timeout: 246 seconds)
1533 2013-03-13 10:00:43 occulta has joined
1534 2013-03-13 10:01:08 pierre`_ is now known as pierre`
1535 2013-03-13 10:01:46 techlife has quit (Ping timeout: 252 seconds)
1536 2013-03-13 10:02:40 valparaiso_ has joined
1537 2013-03-13 10:03:57 valparaiso has quit (Ping timeout: 256 seconds)
1538 2013-03-13 10:03:58 valparaiso_ is now known as valparaiso
1539 2013-03-13 10:04:02 lady_awk has joined
1540 2013-03-13 10:06:00 i2pRelay has quit (Remote host closed the connection)
1541 2013-03-13 10:06:25 epilido has joined
1542 2013-03-13 10:06:25 Aaron_TangCryp has joined
1543 2013-03-13 10:06:25 evan_ has joined
1544 2013-03-13 10:06:32 techlife has joined
1545 2013-03-13 10:06:32 i2pRelay has joined
1546 2013-03-13 10:07:00 nexes has joined
1547 2013-03-13 10:08:56 PlantMan has joined
1548 2013-03-13 10:09:02 ovidiusoft has quit (Ping timeout: 252 seconds)
1549 2013-03-13 10:09:13 cyphase has joined
1550 2013-03-13 10:09:21 pete78 has quit (Quit: Page closed)
1551 2013-03-13 10:09:39 pete76 has joined
1552 2013-03-13 10:10:14 valparaiso has quit ()
1553 2013-03-13 10:10:54 ovidiusoft has joined
1554 2013-03-13 10:10:58 valparaiso has joined
1555 2013-03-13 10:10:58 valparaiso has quit (Changing host)
1556 2013-03-13 10:10:58 valparaiso has joined
1557 2013-03-13 10:11:48 dawei101 has joined
1558 2013-03-13 10:13:00 JDuke128 has quit (Quit: [BB])
1559 2013-03-13 10:13:08 wizkid057 has quit (Ping timeout: 252 seconds)
1560 2013-03-13 10:14:02 i2pRelay has quit (Remote host closed the connection)
1561 2013-03-13 10:14:32 i2pRelay has joined
1562 2013-03-13 10:15:20 Boydy has joined
1563 2013-03-13 10:15:20 Boydy has quit (Changing host)
1564 2013-03-13 10:15:20 Boydy has joined
1565 2013-03-13 10:18:28 chmod755 has joined
1566 2013-03-13 10:18:44 wizkid057 has joined
1567 2013-03-13 10:19:46 Nicksasa has quit (Ping timeout: 258 seconds)
1568 2013-03-13 10:22:04 i2pRelay has quit (Remote host closed the connection)
1569 2013-03-13 10:22:38 i2pRelay has joined
1570 2013-03-13 10:23:12 andrew12 has joined
1571 2013-03-13 10:24:48 CaptainBlaze has joined
1572 2013-03-13 10:25:59 PlantMan has quit (Read error: Connection reset by peer)
1573 2013-03-13 10:26:06 B0g4r7 has quit (Ping timeout: 276 seconds)
1574 2013-03-13 10:26:26 PlantMan has joined
1575 2013-03-13 10:26:31 fishfish is now known as fishfish|AFK
1576 2013-03-13 10:26:32 randy-waterhouse has quit (Quit: Out.)
1577 2013-03-13 10:27:32 gjs278 has quit (Quit: Konversation terminated!)
1578 2013-03-13 10:27:49 gjs278 has joined
1579 2013-03-13 10:28:15 nelisky has quit (Quit: nelisky)
1580 2013-03-13 10:29:59 one_zero has quit ()
1581 2013-03-13 10:30:06 i2pRelay has quit (Remote host closed the connection)
1582 2013-03-13 10:30:39 i2pRelay has joined
1583 2013-03-13 10:31:25 doublec_ is now known as doublec
1584 2013-03-13 10:31:28 eric____ has joined
1585 2013-03-13 10:31:30 doublec has quit (Changing host)
1586 2013-03-13 10:31:31 ville-_ has joined
1587 2013-03-13 10:32:01 doublec has joined
1588 2013-03-13 10:34:23 LainZ has quit (Ping timeout: 276 seconds)
1589 2013-03-13 10:34:37 JDuke128 has joined
1590 2013-03-13 10:37:10 Boydy has quit (Remote host closed the connection)
1591 2013-03-13 10:38:09 i2pRelay has quit (Remote host closed the connection)
1592 2013-03-13 10:38:39 i2pRelay has joined
1593 2013-03-13 10:39:27 fishfish is now known as AFK!~fishfish2@109.238.67.130|fishfish
1594 2013-03-13 10:40:42 ciphermonk has quit (Remote host closed the connection)
1595 2013-03-13 10:41:22 ciphermonk has joined
1596 2013-03-13 10:42:33 sacredchao has quit (Disconnected by services)
1597 2013-03-13 10:44:02 Guest79369 has quit (Quit: quit)
1598 2013-03-13 10:44:04 darxun has joined
1599 2013-03-13 10:45:38 KornKage2 has joined
1600 2013-03-13 10:45:42 KornKage has quit (Ping timeout: 248 seconds)
1601 2013-03-13 10:46:05 darxun has left ()
1602 2013-03-13 10:46:10 i2pRelay has quit (Remote host closed the connection)
1603 2013-03-13 10:46:41 i2pRelay has joined
1604 2013-03-13 10:47:15 pete78 has joined
1605 2013-03-13 10:47:24 pete76 has quit (Quit: Page closed)
1606 2013-03-13 10:47:38 fiesh has quit (Ping timeout: 255 seconds)
1607 2013-03-13 10:49:38 jonwaller has quit (Remote host closed the connection)
1608 2013-03-13 10:50:42 paraipan has quit (Remote host closed the connection)
1609 2013-03-13 10:50:42 rdponticelli has quit (Remote host closed the connection)
1610 2013-03-13 10:51:28 paraipan has joined
1611 2013-03-13 10:51:56 pete3 has quit (Ping timeout: 252 seconds)
1612 2013-03-13 10:51:59 Mandrius has quit (Quit: KVIrc 4.0.4 Insomnia http://www.kvirc.net/)
1613 2013-03-13 10:52:10 rdponticelli has joined
1614 2013-03-13 10:53:22 kytv has quit (Remote host closed the connection)
1615 2013-03-13 10:53:45 kytv has joined
1616 2013-03-13 10:54:02 a5m0_ is now known as a5m0
1617 2013-03-13 10:54:10 fiesh has joined
1618 2013-03-13 10:54:11 i2pRelay has quit (Remote host closed the connection)
1619 2013-03-13 10:54:21 a5m0 has quit (Changing host)
1620 2013-03-13 10:54:22 pete2 has joined
1621 2013-03-13 10:54:51 a5m0 has joined
1622 2013-03-13 10:54:57 fishfish is now known as fishfish|AFK
1623 2013-03-13 10:55:08 i2pRelay has joined
1624 2013-03-13 10:56:20 manet has quit (Ping timeout: 252 seconds)
1625 2013-03-13 10:56:20 pete2 has quit (Read error: Connection reset by peer)
1626 2013-03-13 10:56:26 pete3 has joined
1627 2013-03-13 10:57:26 datagutt has joined
1628 2013-03-13 11:00:38 Grouver has joined
1629 2013-03-13 11:01:38 thepok has joined
1630 2013-03-13 11:02:39 i2pRelay has quit (Remote host closed the connection)
1631 2013-03-13 11:03:13 i2pRelay has joined
1632 2013-03-13 11:03:22 fishfish is now known as AFK!~fishfish2@109.238.67.130|fishfish
1633 2013-03-13 11:03:43 holorga_ has quit (Remote host closed the connection)
1634 2013-03-13 11:03:43 i2pRelay has quit (Remote host closed the connection)
1635 2013-03-13 11:04:39 i2pRelay has joined
1636 2013-03-13 11:04:45 holorga_ has joined
1637 2013-03-13 11:04:59 enquirer4 has quit (Quit: Nettalk6 - www.ntalk.de)
1638 2013-03-13 11:05:20 rbecker is now known as RBecker
1639 2013-03-13 11:05:36 VoteGoat_ has joined
1640 2013-03-13 11:05:36 pete3 has quit (Ping timeout: 260 seconds)
1641 2013-03-13 11:06:09 <VoteGoat_> Hi, I have a noob question. I am doing research on how to start a bitcoin business and was wondering, how would one manage hundreds of thousands of bitcoin addresses, is that done in one wallet? Does anyone know how the current players are doing it?
1642 2013-03-13 11:06:46 <ligar> script it :P
1643 2013-03-13 11:07:20 <VoteGoat_> I mean are there any issues with a wallet that size?
1644 2013-03-13 11:07:26 <ligar> you can make calls to bitcoind and it does the work
1645 2013-03-13 11:07:28 <ligar> no
1646 2013-03-13 11:07:42 <ligar> not really
1647 2013-03-13 11:07:54 Ashaman has joined
1648 2013-03-13 11:08:10 Ashaman has quit (Remote host closed the connection)
1649 2013-03-13 11:08:27 pete2 has joined
1650 2013-03-13 11:09:28 <VoteGoat_> looking into that now, thanks for the responce, again sorry for the noob question
1651 2013-03-13 11:10:03 ThomasV has joined
1652 2013-03-13 11:11:07 Ashaman has joined
1653 2013-03-13 11:12:10 i2pRelay has quit (Remote host closed the connection)
1654 2013-03-13 11:12:40 i2pRelay has joined
1655 2013-03-13 11:16:29 JZavala has quit (Ping timeout: 256 seconds)
1656 2013-03-13 11:17:52 pete2 has quit (Read error: Connection reset by peer)
1657 2013-03-13 11:18:52 fishfish is now known as fishfish|AFK
1658 2013-03-13 11:19:47 joehoyle has joined
1659 2013-03-13 11:20:13 i2pRelay has quit (Remote host closed the connection)
1660 2013-03-13 11:20:45 i2pRelay has joined
1661 2013-03-13 11:22:36 pete2 has joined
1662 2013-03-13 11:23:57 rdponticelli has quit (Ping timeout: 276 seconds)
1663 2013-03-13 11:24:26 RBecker is now known as rbecker
1664 2013-03-13 11:24:36 ciphermonk has quit (Ping timeout: 276 seconds)
1665 2013-03-13 11:25:32 MobiusL has joined
1666 2013-03-13 11:26:43 nelisky has joined
1667 2013-03-13 11:26:57 fishfish is now known as AFK!~fishfish2@109.238.67.130|fishfish
1668 2013-03-13 11:28:15 i2pRelay has quit (Remote host closed the connection)
1669 2013-03-13 11:28:30 <Ant0> what happened finally with the fork bug? fixed in 0.8 client?
1670 2013-03-13 11:28:39 techlife has quit (Ping timeout: 264 seconds)
1671 2013-03-13 11:28:46 i2pRelay has joined
1672 2013-03-13 11:28:57 <ligar> Ant0: ran the 7.0 chain forcing reconverge
1673 2013-03-13 11:29:30 <ligar> up in the air as far as i know as the perm fix.
1674 2013-03-13 11:29:58 <ligar> 7.0 =0.7
1675 2013-03-13 11:30:27 chenshaoju has joined
1676 2013-03-13 11:30:59 brendyn has quit (Ping timeout: 252 seconds)
1677 2013-03-13 11:31:03 nowan_ has quit (Ping timeout: 252 seconds)
1678 2013-03-13 11:31:47 pete2 has quit (Ping timeout: 252 seconds)
1679 2013-03-13 11:32:05 brendyn has joined
1680 2013-03-13 11:32:50 nowan has joined
1681 2013-03-13 11:32:52 VoteGoat_ has quit (Quit: Page closed)
1682 2013-03-13 11:35:07 pete2 has joined
1683 2013-03-13 11:36:16 i2pRelay has quit (Remote host closed the connection)
1684 2013-03-13 11:36:48 i2pRelay has joined
1685 2013-03-13 11:37:39 brendyn has quit (Ping timeout: 252 seconds)
1686 2013-03-13 11:38:37 techlife has joined
1687 2013-03-13 11:39:21 pete2 has quit (Ping timeout: 240 seconds)
1688 2013-03-13 11:40:00 <Goonie> does anyone know about the current state of testnet3? it appears to be very forked as well since a few days.
1689 2013-03-13 11:40:39 <gmaxwell> Goonie: why do you say that?
1690 2013-03-13 11:40:59 brendyn has joined
1691 2013-03-13 11:40:59 pgp has quit ()
1692 2013-03-13 11:41:10 <Goonie> because I get different results on blockexplorer and in Bitcoin Wallet
1693 2013-03-13 11:41:14 pgp has joined
1694 2013-03-13 11:41:20 rdponticelli has joined
1695 2013-03-13 11:41:29 <gmaxwell> Goonie: blockexplorer gets stuck whenever there is a reorg of more than a couple blocks
1696 2013-03-13 11:41:29 <Goonie> and also I get user reports of tx missing between Bitcoin Wallet and bitcoind
1697 2013-03-13 11:42:22 jdnavarro has joined
1698 2013-03-13 11:42:34 <Goonie> ok indeed blockexplorer is stuck. I'll email theymous
1699 2013-03-13 11:42:54 pete2 has joined
1700 2013-03-13 11:43:49 CaptainBlaze has quit (Quit: CaptainBlaze)
1701 2013-03-13 11:44:19 i2pRelay has quit (Remote host closed the connection)
1702 2013-03-13 11:44:33 <Goonie> guess I'll add a simple block explorer to Bitcoin Wallet
1703 2013-03-13 11:44:48 i2pRelay has joined
1704 2013-03-13 11:45:22 wrabbit has left ()
1705 2013-03-13 11:46:21 pete3 has joined
1706 2013-03-13 11:47:56 chenshaoju has left ("Once you know what it is you want to be true, instinct is a very useful device for enabling you to know that it is")
1707 2013-03-13 11:49:35 pete2 has quit (Ping timeout: 246 seconds)
1708 2013-03-13 11:51:52 pete2 has joined
1709 2013-03-13 11:52:21 i2pRelay has quit (Remote host closed the connection)
1710 2013-03-13 11:52:51 daybyter has joined
1711 2013-03-13 11:52:53 i2pRelay has joined
1712 2013-03-13 11:53:01 tre-spective has quit (Ping timeout: 260 seconds)
1713 2013-03-13 11:53:50 viperhr has quit (Ping timeout: 245 seconds)
1714 2013-03-13 11:53:58 fishfish is now known as fishfish|AFK
1715 2013-03-13 11:55:00 OPrime has joined
1716 2013-03-13 11:55:39 pete3 has quit (Ping timeout: 264 seconds)
1717 2013-03-13 11:59:08 terry has joined
1718 2013-03-13 11:59:09 nick has joined
1719 2013-03-13 11:59:12 Casimir1904 has joined
1720 2013-03-13 11:59:33 nick is now known as Guest92523
1721 2013-03-13 12:00:23 i2pRelay has quit (Remote host closed the connection)
1722 2013-03-13 12:00:56 i2pRelay has joined
1723 2013-03-13 12:01:15 da2ce7_d is now known as da2ce7
1724 2013-03-13 12:01:52 drizztbsd has joined
1725 2013-03-13 12:01:52 drizztbsd has quit (Changing host)
1726 2013-03-13 12:01:52 drizztbsd has joined
1727 2013-03-13 12:03:22 nivoc has joined
1728 2013-03-13 12:03:28 pete2 has quit (Ping timeout: 260 seconds)
1729 2013-03-13 12:04:33 dawei101 has quit (Quit: Leaving.)
1730 2013-03-13 12:05:39 pete2 has joined
1731 2013-03-13 12:05:48 tre-spective has joined
1732 2013-03-13 12:06:16 rkosten has joined
1733 2013-03-13 12:07:17 RoboTeddy has quit (Remote host closed the connection)
1734 2013-03-13 12:08:25 i2pRelay has quit (Remote host closed the connection)
1735 2013-03-13 12:08:57 i2pRelay has joined
1736 2013-03-13 12:08:57 fishfish is now known as AFK!~fishfish2@109.238.67.130|fishfish
1737 2013-03-13 12:09:32 PhantomSpark has quit (Ping timeout: 246 seconds)
1738 2013-03-13 12:10:45 pete2 has quit (Ping timeout: 276 seconds)
1739 2013-03-13 12:11:43 da2ce7 has quit (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
1740 2013-03-13 12:11:45 Guest92523 is now known as Nicksasa
1741 2013-03-13 12:11:46 stochasm has quit (Ping timeout: 248 seconds)
1742 2013-03-13 12:11:47 Nicksasa has quit (Changing host)
1743 2013-03-13 12:11:47 Nicksasa has joined
1744 2013-03-13 12:14:02 da2ce7 has joined
1745 2013-03-13 12:15:13 MobiusL has quit (Quit: Ex-Chat)
1746 2013-03-13 12:16:26 i2pRelay has quit (Remote host closed the connection)
1747 2013-03-13 12:16:58 i2pRelay has joined
1748 2013-03-13 12:16:59 da2ce7 has quit (Read error: Connection reset by peer)
1749 2013-03-13 12:17:08 MobiusL has joined
1750 2013-03-13 12:17:21 chmod755 has quit (Quit: chmod755)
1751 2013-03-13 12:18:30 da2ce7 has joined
1752 2013-03-13 12:19:23 TradeFortress has quit (Ping timeout: 252 seconds)
1753 2013-03-13 12:20:01 pete2 has joined
1754 2013-03-13 12:21:42 manet has joined
1755 2013-03-13 12:21:44 hydrogenesis has joined
1756 2013-03-13 12:23:11 Grouver has quit (Quit: HydraIRC -> http://www.hydrairc.com <- Chicks dig it)
1757 2013-03-13 12:23:29 Paul_Hartman has quit (Ping timeout: 252 seconds)
1758 2013-03-13 12:23:43 agricocb has quit (Quit: Leaving.)
1759 2013-03-13 12:24:17 gimlet has joined
1760 2013-03-13 12:24:27 i2pRelay has quit (Remote host closed the connection)
1761 2013-03-13 12:24:38 pizzaman2337 has quit (Read error: Operation timed out)
1762 2013-03-13 12:25:00 i2pRelay has joined
1763 2013-03-13 12:25:41 mogri has joined
1764 2013-03-13 12:26:39 smelul8r has quit (Ping timeout: 258 seconds)
1765 2013-03-13 12:27:07 smelul8r has joined
1766 2013-03-13 12:28:06 gimlet90210 has quit (Ping timeout: 256 seconds)
1767 2013-03-13 12:28:36 alezakos has joined
1768 2013-03-13 12:28:55 alezakos has left ("Leaving")
1769 2013-03-13 12:30:13 pizzaman1337 has joined
1770 2013-03-13 12:31:55 fishfish has quit (Quit: Bye!)
1771 2013-03-13 12:32:34 i2pRelay has quit (Remote host closed the connection)
1772 2013-03-13 12:33:11 i2pRelay has joined
1773 2013-03-13 12:33:12 bukaj has joined
1774 2013-03-13 12:34:15 TD_ has joined
1775 2013-03-13 12:34:23 JDuke128 has quit (Quit: [BB])
1776 2013-03-13 12:34:24 TD_ is now known as TD
1777 2013-03-13 12:36:15 unknown45682 has joined
1778 2013-03-13 12:38:16 MC-Droid has quit (Read error: Connection reset by peer)
1779 2013-03-13 12:38:31 MC-Droid has joined
1780 2013-03-13 12:38:42 stochasm has joined
1781 2013-03-13 12:40:25 sanford has joined
1782 2013-03-13 12:40:39 i2pRelay has quit (Remote host closed the connection)
1783 2013-03-13 12:41:09 i2pRelay has joined
1784 2013-03-13 12:41:15 sanford has quit (Quit: Page closed)
1785 2013-03-13 12:42:25 Perlboy has joined
1786 2013-03-13 12:42:28 SchmalzTech has joined
1787 2013-03-13 12:43:48 hydrogenesis has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
1788 2013-03-13 12:44:00 m00p has joined
1789 2013-03-13 12:44:26 polpaul has quit (Quit: Page closed)
1790 2013-03-13 12:44:29 lightlord has quit (Quit: Page closed)
1791 2013-03-13 12:45:42 OPrime has quit (Ping timeout: 248 seconds)
1792 2013-03-13 12:46:33 albert0 has joined
1793 2013-03-13 12:48:28 Forte has left ("WeeChat 0.4.0")
1794 2013-03-13 12:48:39 i2pRelay has quit (Remote host closed the connection)
1795 2013-03-13 12:49:13 i2pRelay has joined
1796 2013-03-13 12:49:39 gimlet is now known as gimlet90210
1797 2013-03-13 12:52:51 gollonline has joined
1798 2013-03-13 12:53:36 gollonline has quit (Client Quit)
1799 2013-03-13 12:53:51 Lepton has joined
1800 2013-03-13 12:56:38 i2pRelay has quit (Remote host closed the connection)
1801 2013-03-13 12:57:08 i2pRelay has joined
1802 2013-03-13 12:58:26 tsche has quit (Ping timeout: 245 seconds)
1803 2013-03-13 12:58:56 cyphase has quit (Read error: Connection reset by peer)
1804 2013-03-13 12:59:04 tsche has joined
1805 2013-03-13 12:59:37 tonikt has quit (Ping timeout: 258 seconds)
1806 2013-03-13 13:01:38 viperhr has joined
1807 2013-03-13 13:02:23 <topi`_> so, the dust has settled now, is anyone missing some btc because of orphaned forks?
1808 2013-03-13 13:03:13 <Luke-Jr> just OKPAY
1809 2013-03-13 13:03:42 agricocb has joined
1810 2013-03-13 13:04:04 <jouke> and that has been resolved as well
1811 2013-03-13 13:04:41 i2pRelay has quit (Remote host closed the connection)
1812 2013-03-13 13:05:12 i2pRelay has joined
1813 2013-03-13 13:05:30 <epscy> so the issue was caused by a BDB querying limitation when validating blocks?
1814 2013-03-13 13:05:53 <jouke> setting
1815 2013-03-13 13:05:56 <jouke> afaik
1816 2013-03-13 13:06:23 <epscy> i'm just wondering if bitcoind couldn't benefit from some abstraction
1817 2013-03-13 13:06:36 nexes has quit (Remote host closed the connection)
1818 2013-03-13 13:06:40 <epscy> though as always it would probably come at the cost of performance
1819 2013-03-13 13:06:51 Spooktooth has quit (Remote host closed the connection)
1820 2013-03-13 13:06:57 <pete78> <+sipa> there are up to 2n locks necessary for n updates to a btree, and our limit was 10000 locks
1821 2013-03-13 13:07:03 <epscy> and wouldn't solve the problem that there are tons of BDB clients still out in the wild
1822 2013-03-13 13:07:18 <gmaxwell> epscy: it's a limitation on the size of an atomic update to bdb.
1823 2013-03-13 13:07:27 <gmaxwell> (given our setting)
1824 2013-03-13 13:07:32 <pete78> <+sipa> and the offending block modified slightly over 5000 transaction index entries
1825 2013-03-13 13:08:17 <gmaxwell> epscy: thats not something that can be easily abstracted, as the change must be atomic.
1826 2013-03-13 13:08:55 Ashaman has quit (Remote host closed the connection)
1827 2013-03-13 13:09:25 <Luke-Jr> ;;bc,blocks
1828 2013-03-13 13:09:25 <gribble> 225643
1829 2013-03-13 13:10:36 pete2 has quit (Read error: Connection reset by peer)
1830 2013-03-13 13:12:27 <epscy> gmaxwell: i see
1831 2013-03-13 13:12:41 i2pRelay has quit (Remote host closed the connection)
1832 2013-03-13 13:13:12 i2pRelay has joined
1833 2013-03-13 13:13:52 <sipa> with large -dbcache, 0.8.0 can do up to millions of modified transactions in one atomic change
1834 2013-03-13 13:14:06 kaptah has joined
1835 2013-03-13 13:14:16 <sipa> so it seems no such limitation exits in leveldb
1836 2013-03-13 13:14:28 <sipa> (which is expected, based on how it works)
1837 2013-03-13 13:15:56 pete2 has joined
1838 2013-03-13 13:16:04 <sipa> block 191652 modified 3813 transactions
1839 2013-03-13 13:16:56 polpaul has joined
1840 2013-03-13 13:17:44 <thermoman> sipa: will upgrading from 0.7 to 0.8 besindes losing bdb for blockchain improve overall performance?
1841 2013-03-13 13:17:51 Guest94161 is now known as MC1984
1842 2013-03-13 13:18:00 MC1984 has quit (Quit: Leaving)
1843 2013-03-13 13:18:13 * Luke-Jr sends 0.8.1 proposal to ML
1844 2013-03-13 13:18:14 MC1984 has joined
1845 2013-03-13 13:18:27 <sipa> thermoman: depending on the speed of I/O and your CPU, it will improve performance somewhere between moderately and ridiculously
1846 2013-03-13 13:18:47 <thermoman> sipa: i guess 0.8 is multithreaded, right?
1847 2013-03-13 13:18:49 <sipa> yes
1848 2013-03-13 13:18:55 <sipa> for signature validation
1849 2013-03-13 13:19:03 <sipa> so only after the last checkpoint (=216116)
1850 2013-03-13 13:19:09 OPrime has joined
1851 2013-03-13 13:19:23 <thermoman> we have performance issues for rpc calls querying wallet stuff (balance for labels, transactions, ...) will this improve?
1852 2013-03-13 13:19:31 <sipa> no
1853 2013-03-13 13:19:34 <thermoman> :(
1854 2013-03-13 13:19:51 <sipa> the wallet has some serious performance issues, but those weren't improved in 0.8
1855 2013-03-13 13:19:58 JDuke128 has joined
1856 2013-03-13 13:20:01 <Luke-Jr> I think he meant future versions :P
1857 2013-03-13 13:20:12 <Luke-Jr> thermoman: probably not at a level suitable to big merchants, for a while
1858 2013-03-13 13:20:13 <thermoman> is it planned to move away from bdb completely?
1859 2013-03-13 13:20:15 thezerg_ has quit (Remote host closed the connection)
1860 2013-03-13 13:20:17 <sipa> thermoman: YES
1861 2013-03-13 13:20:42 i2pRelay has quit (Remote host closed the connection)
1862 2013-03-13 13:20:43 <thermoman> :)
1863 2013-03-13 13:21:14 i2pRelay has joined
1864 2013-03-13 13:22:35 kanoi_ has joined
1865 2013-03-13 13:25:27 <Scrat> sipa: but queries are sometimes slow because of the inefficiencies of internal structures and/or locks, since everything is in ram, not BDB. right?
1866 2013-03-13 13:25:45 <sipa> Scrat: the slowless of the wallet has nothing to do with BDB
1867 2013-03-13 13:25:50 <sipa> just inefficient implementation
1868 2013-03-13 13:25:53 <Scrat> yep
1869 2013-03-13 13:26:04 yesminister has quit (Ping timeout: 258 seconds)
1870 2013-03-13 13:26:11 kanoi has quit (Ping timeout: 272 seconds)
1871 2013-03-13 13:26:32 <grau> You might want to experiment with bitsofproof if you need quick query of tx by address and similar.
1872 2013-03-13 13:26:55 Boydy has joined
1873 2013-03-13 13:27:05 Boydy has quit (Remote host closed the connection)
1874 2013-03-13 13:27:38 kadoban has quit (Ping timeout: 245 seconds)
1875 2013-03-13 13:28:46 i2pRelay has quit (Remote host closed the connection)
1876 2013-03-13 13:29:15 i2pRelay has joined
1877 2013-03-13 13:30:56 <topi`_> what's the main thing that slows down the wallet? is there something of complexity O^2 or even O^3 ?
1878 2013-03-13 13:31:20 <Luke-Jr> topi`_: IIRC it scans every transaction to calculate the output in question
1879 2013-03-13 13:31:22 <sipa> no, O(n), where it should be O(1) or O(log n) for many things
1880 2013-03-13 13:32:04 <Luke-Jr> users in #bitcoin are complaining that we should have a remote trigger for safemode again :P
1881 2013-03-13 13:32:56 <Luke-Jr> maybe we should have a -alertsafemode option (default disabled) to restore such behaviour?
1882 2013-03-13 13:33:04 <gmaxwell> sipa: there is some bad complexity stuff in the wallet.
1883 2013-03-13 13:33:19 <sipa> ok, coin selection is bad
1884 2013-03-13 13:33:24 <gmaxwell> IsMine/IsConfirmed on unconfirmed change is ugly.
1885 2013-03-13 13:33:54 <TD> Luke-Jr: it'd be nice if the alert system had a bit more structure
1886 2013-03-13 13:33:54 <gmaxwell> ;;bc,blocks
1887 2013-03-13 13:33:55 <gribble> 225647
1888 2013-03-13 13:34:04 <TD> Luke-Jr: so alerts could trigger different behavior.
1889 2013-03-13 13:34:12 <Luke-Jr> TD: didn't it?
1890 2013-03-13 13:34:20 <Luke-Jr> I thought there was a flag for safemode
1891 2013-03-13 13:34:38 <TD> hmm
1892 2013-03-13 13:34:41 <TD> don't think so
1893 2013-03-13 13:34:42 <TD> https://github.com/bitcoin/bitcoin/blob/master/src/alert.h
1894 2013-03-13 13:34:46 <TD> there's a lot of stuff in there
1895 2013-03-13 13:34:56 <TD> but ultimately it's just a few strings
1896 2013-03-13 13:35:06 <TD> perhaps strReserved could hold another data structure that contains extra commands
1897 2013-03-13 13:35:46 <TD> i mean there can be non-urgent alerts to advise users of upgrades, urgent "upgrade soon or be hard forked off" type alerts, "emergency panic stop trading" type alerts ...
1898 2013-03-13 13:36:00 saulimus has quit (Quit: saulimus)
1899 2013-03-13 13:36:00 <TD> and that's before we even get into, wouldn't it be nice if other types of client could have their own alerts too
1900 2013-03-13 13:36:23 AndChat has joined
1901 2013-03-13 13:36:37 <Luke-Jr> TD: yes - they can, but getting them relayed is troublesome
1902 2013-03-13 13:36:48 i2pRelay has quit (Remote host closed the connection)
1903 2013-03-13 13:36:48 HANTI is now known as hanti
1904 2013-03-13 13:36:50 <Luke-Jr> TD: relaying regardless of key creates spam opportunities
1905 2013-03-13 13:36:52 <TD> well, right.
1906 2013-03-13 13:36:54 <TD> it's a hard problem
1907 2013-03-13 13:37:05 <Luke-Jr> but really, no reason to use p2p for alerts?
1908 2013-03-13 13:37:09 <gmaxwell> obviously you just need to use fidelity bonded alerts. Alerts only get relayed if their timestamp is accurate to at least 12 hours... and you lose your alert bond if you issue more than 1 per day. :P
1909 2013-03-13 13:37:18 i2pRelay has joined
1910 2013-03-13 13:37:33 <sipa> ;;bc,blocks
1911 2013-03-13 13:37:34 <gribble> 225648
1912 2013-03-13 13:37:53 <sipa> op to block 221000, no block modified over 3813 transactions
1913 2013-03-13 13:38:04 <Scrat> I'm starting to think fidelity bonds are like DHT (solves everything)
1914 2013-03-13 13:38:04 <Luke-Jr> but I mean, as long as it's a centralizing thing, why not hit the project website?
1915 2013-03-13 13:38:25 <gmaxwell> Scrat: I make that kind of suggestion as half a joke along those lines.
1916 2013-03-13 13:38:30 pete2 has quit (Ping timeout: 248 seconds)
1917 2013-03-13 13:38:40 <TD> haha
1918 2013-03-13 13:38:51 <Luke-Jr> bitcoind could be polling https://github.com/bitcoin/bitcoin/blob/alerts/alerts.dat
1919 2013-03-13 13:38:56 <Luke-Jr> but then I guess polling is bad
1920 2013-03-13 13:38:56 <TD> Luke-Jr: yeah, sure, i mean it's a longer term thing. i didn't mean to imply we should try to solve that right now
1921 2013-03-13 13:39:01 <gmaxwell> Scrat: Although ignoring the @#$#@mess of engineering complexity, it would actually achieve the desired properties. Everyone can make alerts, if they want.. but there is no risk of floods.
1922 2013-03-13 13:39:01 banghouse has joined
1923 2013-03-13 13:39:01 banghouse is now known as Guest83324
1924 2013-03-13 13:39:09 <TD> satoshi was very very keen on having no SPOFs at all, even for somewhat trivial things like the alerts
1925 2013-03-13 13:39:18 <TD> i suspect most clients won't be that hard core
1926 2013-03-13 13:39:18 <sipa> SPOF?
1927 2013-03-13 13:39:21 <TD> single point of failure
1928 2013-03-13 13:39:23 yesminister has joined
1929 2013-03-13 13:39:39 <sipa> i believe he was even reluctant about something like alerts
1930 2013-03-13 13:39:42 <gmaxwell> Scrat: it would allow things like exchanges to send daily quote messages.. but still be confident that no one would flood.
1931 2013-03-13 13:39:51 <TD> yeah
1932 2013-03-13 13:40:23 <gmaxwell> yea, the alerts are ... ugh. And we see that they often get invoked when people want to argue that bitcoin is really centeralized afterall.
1933 2013-03-13 13:40:44 <gmaxwell> but too useful to toss at the same time, quite obviously.
1934 2013-03-13 13:41:15 daybyter has quit (Quit: Konversation terminated!)
1935 2013-03-13 13:41:27 <gmaxwell> If the internet had a attack resistant decenteralized broadcast medium we would just use that and call it done.
1936 2013-03-13 13:41:33 <jouke> well, about the alert, my .7 client showed the warning and not the alert, I am not sure I agree with that prioritization.
1937 2013-03-13 13:41:55 ciphermonk has joined
1938 2013-03-13 13:42:29 <SomeoneWeird> gmaxwell, but it doesn't
1939 2013-03-13 13:42:29 <SomeoneWeird> lol
1940 2013-03-13 13:42:50 <TD> the "centralization" of alerts comes from the small number of people with the key
1941 2013-03-13 13:42:57 joe_k has quit (Quit: Leaving.)
1942 2013-03-13 13:43:16 <TD> i think a lot of the complaints would go away if doing an alert required a 2-of-10 threshold or something
1943 2013-03-13 13:44:30 <Luke-Jr> TD: probably, but how can we change/upgrade the 2-of-10 on need? :P
1944 2013-03-13 13:44:33 <sipa> the reorganization 25-block deep from the 0.8/0.7 fork results in a database transaction affecting 26522 transactions
1945 2013-03-13 13:44:37 <gmaxwell> TD: 2 of 10 is still centeralized it's just distributedâ an improvement, but you still have a small clique with special authority. E.g. FooClient is disenfranchized by not being able to participate in this game until its quite popular.
1946 2013-03-13 13:44:42 <TD> an alert can contain a command to change the accepted key set
1947 2013-03-13 13:44:48 AndChat has quit (417489!~AndChat41@184-197-106-63.pools.spcsdns.net|Quit: Bye)
1948 2013-03-13 13:44:54 i2pRelay has quit (Remote host closed the connection)
1949 2013-03-13 13:45:05 * TD shrugs
1950 2013-03-13 13:45:10 <Luke-Jr> how about a 10-of-10 master key, that signs a script hash + expiration timestamp to verify the real sigs? :P
1951 2013-03-13 13:45:15 <TD> 99% of users will be on apps developed by a small clique of people anyway
1952 2013-03-13 13:45:28 i2pRelay has joined
1953 2013-03-13 13:45:46 thoughtcourier has joined
1954 2013-03-13 13:45:50 <TD> sipa: so the temp rule has to be something like "don't create a block with more than 4000 inputs in it"?
1955 2013-03-13 13:45:53 <Luke-Jr> hmm, seems at least one user would want the signatures to be necessarily multiple jurisdictions
1956 2013-03-13 13:46:01 <gmaxwell> TD: I don't think it's much of a pratical issue, but philosophically its not good.
1957 2013-03-13 13:46:11 <Luke-Jr> TD: see ML post
1958 2013-03-13 13:46:15 <TD> yeah, i think jurisdictional diversity is a good thing, if/when we split the android app RSA key i'll be looking for that.
1959 2013-03-13 13:46:20 <TD> but it's hard
1960 2013-03-13 13:46:51 <sipa> ok, just reindexed the entire main chain with -nocheckpoints and my secp256k1: 53 minute wall clock time; 3.5h CPU time; 13 minutes wall clock after the last checkpoint
1961 2013-03-13 13:46:55 COGSMITH has quit (Read error: Connection reset by peer)
1962 2013-03-13 13:46:56 pete2 has joined
1963 2013-03-13 13:47:24 <sipa> TD: the temp rule can count the number of modified transactions per block exactly
1964 2013-03-13 13:47:59 <TD> sipa: wow, nice.
1965 2013-03-13 13:47:59 <gmaxwell> sipa: very cool.
1966 2013-03-13 13:48:03 COGSMITH has joined
1967 2013-03-13 13:48:04 <TD> sipa: what tx/sec rate is that?
1968 2013-03-13 13:48:30 <sipa> we're at 14.3M transactions total
1969 2013-03-13 13:48:37 joehoyle has quit (Quit: Leaving...)
1970 2013-03-13 13:48:50 defekt has quit (Ping timeout: 248 seconds)
1971 2013-03-13 13:48:54 alna has quit (Ping timeout: 245 seconds)
1972 2013-03-13 13:49:17 <MC1984> impressiv
1973 2013-03-13 13:49:19 <sipa> so 4500 tx/s
1974 2013-03-13 13:49:35 defekt has joined
1975 2013-03-13 13:49:41 <TD> not bad
1976 2013-03-13 13:49:56 kanoi_ is now known as kanoi
1977 2013-03-13 13:50:04 <TD> sipa: i'm curious if ecdsa is really the main bottleneck at this point or whether it shifted back to disk / "misc" cpu
1978 2013-03-13 13:50:17 <TD> i was thinking this morning that at some point we'll discover it spends most of its time in malloc
1979 2013-03-13 13:50:20 <sipa> it is the bottleneck still
1980 2013-03-13 13:50:22 <TD> ok
1981 2013-03-13 13:50:32 <sipa> at least here
1982 2013-03-13 13:50:35 <TD> that's with 4 cores?
1983 2013-03-13 13:51:00 <sipa> 4 hyperthreading cores, so 8 threads
1984 2013-03-13 13:51:09 <gmaxwell> Scaling to more cores is still somewhat limited.
1985 2013-03-13 13:51:11 <sipa> but i never get above 400-450% CPU usage
1986 2013-03-13 13:51:19 <TD> we should grab a 32-core machine and see when it falls over and dies.
1987 2013-03-13 13:51:26 <gmaxwell> TD: I did.
1988 2013-03-13 13:51:29 <TD> well, i suspect hyperthreading doesn't help with ecdsa much
1989 2013-03-13 13:51:45 <sipa> TD: it does, actually, at least on OpenSSL
1990 2013-03-13 13:51:51 <gmaxwell> I did a bunch of testing and benchmarks on a 32 core system for the original par patches.
1991 2013-03-13 13:52:00 <sipa> far from a factor 2, but at least a non-trivial improvement
1992 2013-03-13 13:52:00 <TD> gmaxwell: how did it do?
1993 2013-03-13 13:52:42 <gmaxwell> IIRC past 8 cores or so it stopped getting faster. Though that was older code, sipa has debottlenecked a little since then.
1994 2013-03-13 13:52:52 i2pRelay has quit (Remote host closed the connection)
1995 2013-03-13 13:53:03 <sipa> well, currently it seems even 4 cores is the limit
1996 2013-03-13 13:53:06 <_dr> be sure to pin the threads if you use more than 4
1997 2013-03-13 13:53:17 <_dr> otherwise cache misses will limit your scalability
1998 2013-03-13 13:53:20 <TD> sipa: you're not expecting to see 800% cpu usage on a 4 core machine even with hyperthreading, right?
1999 2013-03-13 13:53:22 i2pRelay has joined
2000 2013-03-13 13:53:25 <sipa> TD: you do
2001 2013-03-13 13:53:51 ovidiusoft has quit (Ping timeout: 264 seconds)
2002 2013-03-13 13:53:58 <gmaxwell> _dr: sure, I made sure there was no thread bouncing killing it.
2003 2013-03-13 13:53:59 <sipa> TD: the wall clock time per serial execution will go up, but it will be measured as loading all cores 100%
2004 2013-03-13 13:54:09 <TD> ok
2005 2013-03-13 13:54:12 <sipa> in case of perfect scheduling
2006 2013-03-13 13:54:27 <gmaxwell> sipa: thoughts on using your blacklist reorg code to make all this validation async?
2007 2013-03-13 13:54:33 <_dr> if you need machines to run some scaling test let me know
2008 2013-03-13 13:55:12 <sipa> gmaxwell: the blacklist reorg code was easy, because ultraprune at one point had built-in asynchronous parallel verification (which was buggy and slow)
2009 2013-03-13 13:55:18 joehoyle has joined
2010 2013-03-13 13:55:36 ovidiusoft has joined
2011 2013-03-13 13:56:15 <_dr> did you have a chance to look at intel's optimization yet? or are they already in the latest openssl?
2012 2013-03-13 13:56:21 <gmaxwell> my 32 way hosts are offline (due to my previously mentioned move) so I can't do more tests nowâ and probably not for some weeks.
2013 2013-03-13 13:56:23 <_dr> it should give another major boost
2014 2013-03-13 13:56:33 b00tkitz has joined
2015 2013-03-13 13:56:58 <gmaxwell> _dr: the intel stuff I've seen has been for binary curves, not prime-k ones.
2016 2013-03-13 13:56:59 <sipa> but before going into full async validation, i think we first need a block cache manager (so the blocks being worked on are refcount-cached as long as not all their txn are validated), which would also reduce the time spent in single-threaded code (as writing/syncing the database is single-threaded, while it could be asynchronous in parallel with validation of the next block)
2017 2013-03-13 13:57:21 <_dr> gmaxwell: the paper i showed to sipa was independent of the curve
2018 2013-03-13 13:57:23 <sipa> or alternatively, use the undo data to retroactively validate blocks, even after all other processing is done
2019 2013-03-13 13:57:44 <_dr> gmaxwell: it was about a very efficient implementation of point doubling/adding
2020 2013-03-13 13:57:59 <_dr> simply put: very fast 256bit mult
2021 2013-03-13 13:58:10 <sipa> _dr: i know way too little about assembly code to grok that
2022 2013-03-13 13:58:13 <TD> sipa: hmm, which part is single threaded?
2023 2013-03-13 13:58:24 <TD> sipa: bear in mind that leveldb more or less immediately offloads the heavy lifting to other threads.
2024 2013-03-13 13:58:35 <TD> sipa: you mean building the db batch?
2025 2013-03-13 13:58:49 <sipa> TD: yes, but that is only done every N blocks anyway
2026 2013-03-13 13:58:52 <_dr> sipa: i already asked, but would you guys be interested in a gsoc student doing that?
2027 2013-03-13 13:59:05 <sipa> TD: when the utxo cache overflows
2028 2013-03-13 13:59:42 <TD> sipa: although it's neat to optimize initial sync time, i think longer term it makes sense to try and optimize steady state performance. at which point blocks won't do a whole lot except hit the sig cache and write through to the db
2029 2013-03-13 13:59:44 <sipa> TD: the part that is single-threaded per block is writing the block data to disk and updating the txindex (which is a few bytes of data only)
2030 2013-03-13 13:59:55 <sipa> TD: agree
2031 2013-03-13 14:00:17 <TD> sipa: we're using non-sync writes, i thought? so the write() to the log file just pushes data into the kernel cache and then returns. or should do unless the kernel ran out of space.
2032 2013-03-13 14:00:22 <TD> _dr: it sounds interesting!
2033 2013-03-13 14:00:31 <gmaxwell> _dr: I'm still thinking that the curve specific optimization we have (which converts things into 64,64->128 multiplies) is likely to be faster than whatever generic thing intel is doing.
2034 2013-03-13 14:00:38 <sipa> TD: where possible
2035 2013-03-13 14:00:53 i2pRelay has quit (Remote host closed the connection)
2036 2013-03-13 14:01:11 <_dr> gmaxwell: it's not even EC specific
2037 2013-03-13 14:01:21 i2pRelay has joined
2038 2013-03-13 14:01:28 <TD> _dr: is it using quite recent opcodes?
2039 2013-03-13 14:01:34 TonyBit has joined
2040 2013-03-13 14:01:35 <_dr> gmaxwell: it's about efficiently implementing the 256bit integer mult. you still have that in your code, right?
2041 2013-03-13 14:01:40 <_dr> TD: aes-ni
2042 2013-03-13 14:01:47 <_dr> pclmuldq or something like that
2043 2013-03-13 14:01:52 <TD> right, pcm.... right
2044 2013-03-13 14:02:02 <sipa> TD: a block cache manager could optimize that a lot, by doing all writes in a separate thread, taken the dependencies between the data into account (so no changing a txindex before the data it refers to is synced to disk, for example)
2045 2013-03-13 14:02:04 <TD> i think most modern cores support aes-ni
2046 2013-03-13 14:02:06 <_dr> pcl, CarryLess
2047 2013-03-13 14:02:17 <_dr> TD: yeah, since the 'core' trademark
2048 2013-03-13 14:02:26 pete3 has joined
2049 2013-03-13 14:03:15 putridp has joined
2050 2013-03-13 14:03:25 <TD> sipa: well, i'm leery of complicating the block handling logic even further for that case, approaches like the sig cache or sig checking in a thread pool make more sense to me as it's easier to see they're correct.
2051 2013-03-13 14:03:26 <MC1984> theres nothing fundamental limiting scaling of verification with core count is there?
2052 2013-03-13 14:03:47 <TD> sipa: i'd imagine that the bottleneck with larger blocks would become looking up all the connected outputs and that could benefit from being done in parallel
2053 2013-03-13 14:03:52 <_dr> MC1984: well, there's only so many transactions in a block to verify :)
2054 2013-03-13 14:03:57 <TD> sipa: as then the kernel can overlap/re-order reads
2055 2013-03-13 14:04:01 _934TSX has joined
2056 2013-03-13 14:04:18 <TD> MC1984: depends what you mean by "fundamental". is the software we use today fundamental? there is always a scaling limit somewhere in any system.
2057 2013-03-13 14:04:30 <TD> MC1984: yes in theory it scales linearly with cores. in practice it doesn't.
2058 2013-03-13 14:04:31 <gmaxwell> MC1984: depends on what you mean by verification. Some things like making sure transactions in a block don't spend the same inputs is fundimentally somewhat serial, but also trivial.
2059 2013-03-13 14:04:36 <sipa> TD: the current parallel code starts doing verifications as soon as inputs are known
2060 2013-03-13 14:04:45 _934TSX has quit (Client Quit)
2061 2013-03-13 14:04:45 <sipa> TD: but fetching the inputs is single-threaded by itself
2062 2013-03-13 14:04:51 pete2 has quit (Ping timeout: 252 seconds)
2063 2013-03-13 14:04:55 <TD> you mean fetching the outputs?
2064 2013-03-13 14:05:08 <sipa> fetching the prevouts referred to by the inputs :)
2065 2013-03-13 14:05:11 <TD> right :)
2066 2013-03-13 14:05:27 <TD> leveldb is thread safe. so it might be as easy as moving the code that reads the prevouts into the thread pool too
2067 2013-03-13 14:05:36 <Luke-Jr> sigh, ASICMiner spamming dust again
2068 2013-03-13 14:05:37 <MC1984> so making a decent 1024 core verifier in a few years is possible but hard
2069 2013-03-13 14:05:47 <TD> i'm not sure how much benefit can really be obtained today because the whole utxo set fits in ram anyway
2070 2013-03-13 14:05:58 <TD> so lookups should all be "free"? but in future it might help
2071 2013-03-13 14:06:03 <gmaxwell> MC1984: 'decent' â you'd waste all your time on syncs.
2072 2013-03-13 14:06:09 <sipa> TD: you can't really, unless with complex code that checks for dependencies between the transactions
2073 2013-03-13 14:06:14 <TD> MC1984: you'd need to have the storage be held in fast RAM too
2074 2013-03-13 14:06:25 <sipa> TD: as inputs can refer to outputs within the same block
2075 2013-03-13 14:06:40 <TD> sipa: just check a hash_map first before hitting leveldb?
2076 2013-03-13 14:06:47 tre-spective has quit (Ping timeout: 246 seconds)
2077 2013-03-13 14:06:47 <gmaxwell> If only someone in this discussion mentioned that above. :P
2078 2013-03-13 14:07:01 <sipa> TD: that always happens
2079 2013-03-13 14:07:02 <TD> heh
2080 2013-03-13 14:07:41 <gmaxwell> Luke-Jr: work on messaging transactions?
2081 2013-03-13 14:07:41 <sipa> TD: the UTXO set is cached unpacked in RAM, backed by LevelDB
2082 2013-03-13 14:07:44 <TD> well, i suppose the write batch can be built and examined serially but it's all in ram anyway
2083 2013-03-13 14:08:01 <TD> so you can just use the write batch as the hashmap. if it needs to be abandoned for some reason, no problem
2084 2013-03-13 14:08:10 <TD> sipa: indeed
2085 2013-03-13 14:08:41 thepok has quit (Ping timeout: 264 seconds)
2086 2013-03-13 14:08:50 <MC1984> hmm well clock speed is pretty much stuck where it is, and the chain only ever gets bigger. I suppose there is implications there for the great blocksize debate too
2087 2013-03-13 14:08:50 tonikt has joined
2088 2013-03-13 14:08:54 i2pRelay has quit (Remote host closed the connection)
2089 2013-03-13 14:09:00 pete3 has quit (Ping timeout: 258 seconds)
2090 2013-03-13 14:09:04 <sipa> TD: constructing the write batch means serializing the data, and as transactions are often modified by several blocks, that's something you better postpone
2091 2013-03-13 14:09:04 <TD> clock speed doesn't go up much but performance has done, for a long time still
2092 2013-03-13 14:09:22 <gmaxwell> MC1984: clockspeed has been increasing, but yea as td says...
2093 2013-03-13 14:09:24 i2pRelay has joined
2094 2013-03-13 14:09:27 <TD> intel and amd have a ton of clever engineers who keep finding ways to optimize things. eg, the aes-ni extensions _dr referred to
2095 2013-03-13 14:09:28 <MC1984> yeah but even that will slow down i think
2096 2013-03-13 14:09:36 <gmaxwell> MC1984: did you see me mocking you have HD sizes in the backscroll?
2097 2013-03-13 14:09:36 <TD> could be. but then, cores are getting cheaper too
2098 2013-03-13 14:09:51 <TD> gmaxwell has 32 core machines at home, it seems. ok, that's gmaxwell :)
2099 2013-03-13 14:09:52 <gmaxwell> s/have/about/
2100 2013-03-13 14:09:52 <MC1984> EG i dont think intels next cpu is going to be +20% better this time, like that last few generations
2101 2013-03-13 14:10:01 <TD> but, you know, it's not like he's a walking fab
2102 2013-03-13 14:10:07 <gmaxwell> TD: yea, 5 of them. :P
2103 2013-03-13 14:10:10 <TD> haha
2104 2013-03-13 14:10:13 <Scrat> gigahertz not increasing, IPC is
2105 2013-03-13 14:10:16 <_dr> cores won't scale in the future, that i can tell you :)
2106 2013-03-13 14:10:25 <_dr> just like ghz didn't in the past
2107 2013-03-13 14:10:28 <sipa> _dr: famous last words?
2108 2013-03-13 14:10:41 <TD> somehow google managed to scale "number of cores" :)
2109 2013-03-13 14:10:41 pete2 has joined
2110 2013-03-13 14:10:44 <_dr> sipa: i bet gmaxwell's 32 core machine is a n-socket node
2111 2013-03-13 14:10:47 <MC1984> yeahh i sawy the 4tb disk
2112 2013-03-13 14:10:53 <_dr> TD: no i meant per core
2113 2013-03-13 14:10:55 <TD> anyway, you can certainly build a sharded node
2114 2013-03-13 14:11:01 <TD> right
2115 2013-03-13 14:11:01 <_dr> because people keep asking for their 1000core cpus
2116 2013-03-13 14:11:04 <gmaxwell> _dr: sure, 8. But fortunately our problems are embarassingly parallel, at least the hard parts.
2117 2013-03-13 14:11:06 <TD> yes, per core performance will eventually hit a wall too
2118 2013-03-13 14:11:07 <_dr> cache coherence will kill you
2119 2013-03-13 14:11:11 <Scrat> avx2 has 256 integer ops :)
2120 2013-03-13 14:11:22 <sipa> _dr: the number of cores per core won't go up; i'm sure the number 1 won't change indeed :p
2121 2013-03-13 14:11:28 <MC1984> what the hell chip has 32 cores alrady?
2122 2013-03-13 14:11:38 <TD> 32 core machines aren't rare at all MC1984
2123 2013-03-13 14:11:42 <TD> they're not all on one chip
2124 2013-03-13 14:11:44 <gmaxwell> Scrat: 256 bit wide vector operations. Not the same thing as 256 bit multiplies. :P
2125 2013-03-13 14:11:49 zooko has joined
2126 2013-03-13 14:11:51 <flyingkiwiguy> can buy one from Dell for $50K
2127 2013-03-13 14:12:04 <flyingkiwiguy> quad sock, 12 core
2128 2013-03-13 14:12:09 <_dr> sipa: cores per cpu of course :)
2129 2013-03-13 14:12:12 <flyingkiwiguy> 256GB memory
2130 2013-03-13 14:12:16 <MC1984> oh i thought they stopped making multi socket servers
2131 2013-03-13 14:12:16 ovidiusoft has quit (Ping timeout: 264 seconds)
2132 2013-03-13 14:12:20 <gmaxwell> flyingkiwiguy: thats overpriced to an extreme. :P
2133 2013-03-13 14:12:27 <flyingkiwiguy> yup
2134 2013-03-13 14:12:31 <gmaxwell> MC1984: no way, though it seems that 8 socket boards are seasonal. :P
2135 2013-03-13 14:12:39 <Scrat> gmaxwell: that's AVX1
2136 2013-03-13 14:12:49 <_dr> the current trend for speed improvement is focussing on vectorization
2137 2013-03-13 14:12:53 <gmaxwell> Scrat: no, avx1 has no integer operations at all.
2138 2013-03-13 14:12:53 <flyingkiwiguy> Intel MIC is 32 core @ 1.2GHz per socket
2139 2013-03-13 14:13:02 <_dr> e.g. intels push for 256bits or 512bits on the xeon phi
2140 2013-03-13 14:13:10 sjn8rzsw7 has quit (Read error: Connection reset by peer)
2141 2013-03-13 14:13:11 <_dr> they'll go the way of the gpu, which they so despise
2142 2013-03-13 14:13:13 <flyingkiwiguy> ditto
2143 2013-03-13 14:13:15 <MC1984> oh shit yeah that thing
2144 2013-03-13 14:13:21 BlackPrapor has quit (Ping timeout: 240 seconds)
2145 2013-03-13 14:13:26 <MC1984> hows it run on a xeon phi
2146 2013-03-13 14:13:36 <gmaxwell> MC1984: send me one and I'll tell you.
2147 2013-03-13 14:13:38 <MC1984> i doubt that anyone has one to try
2148 2013-03-13 14:13:45 <_dr> yes i do
2149 2013-03-13 14:13:47 <TD> sipa: so does pclmulqdq look useful?
2150 2013-03-13 14:13:53 * flyingkiwiguy has access to a 32 core quad socket w/256GB RAM if some test code needs to be profiled
2151 2013-03-13 14:13:56 ovidiusoft has joined
2152 2013-03-13 14:14:02 <_dr> we even have a 'sold' phi system now apart from the early-access one
2153 2013-03-13 14:14:23 <MC1984> _dr can you run a chain sync on it? mite b cool
2154 2013-03-13 14:14:35 <_dr> MC1984: probably you wouldn't want that ;)
2155 2013-03-13 14:14:51 <_dr> it's very slow with vanialla x86 instructions
2156 2013-03-13 14:15:07 <_dr> it only get's fast using the VPU (as i said, 512bits)
2157 2013-03-13 14:15:15 <MC1984> i thought it was just 100 penitum cores or whatever
2158 2013-03-13 14:15:18 Guest83324 has quit (Remote host closed the connection)
2159 2013-03-13 14:15:27 <_dr> 61 pentium cores, 4way smt, and the vpu
2160 2013-03-13 14:15:27 pete2 has quit (Ping timeout: 264 seconds)
2161 2013-03-13 14:15:31 <gmaxwell> MC1984: the cores are slow, but they have fast vector units.
2162 2013-03-13 14:15:47 <gmaxwell> but that means you need ecdsa code that uses the vector units.
2163 2013-03-13 14:16:15 <_dr> gmaxwell: and imci doesn't have support for 64bit ints :*
2164 2013-03-13 14:16:21 <TD> sipa: paper is here: http://www.intel.com/content/www/us/en/intelligent-systems/wireless-infrastructure/polynomial-multiplication-instructions-paper.html
2165 2013-03-13 14:16:51 <flyingkiwiguy> SMT likely won't give you much, unless you are 100% memory bound
2166 2013-03-13 14:16:57 i2pRelay has quit (Remote host closed the connection)
2167 2013-03-13 14:17:14 <_dr> TD: in a nutshell they use multiple karatsubas to do the 256mult with 64 bit mults. the clever thing is they use the large vector register (256 avx) to hold the intermediate results
2168 2013-03-13 14:17:26 <sipa> TD: i don't think that's useful; we need multiplication with carry
2169 2013-03-13 14:17:27 i2pRelay has joined
2170 2013-03-13 14:17:41 <sipa> secp256k1 is not a GF(2^n) based field
2171 2013-03-13 14:17:43 <gmaxwell> As I said above, it's useful for the binary fields.
2172 2013-03-13 14:17:49 <gmaxwell> Which we do not use.
2173 2013-03-13 14:17:51 Cache_Money has joined
2174 2013-03-13 14:17:55 <TD> sipa: the one they experimented with was in GFp
2175 2013-03-13 14:17:56 <flyingkiwiguy> is it parrellisable?
2176 2013-03-13 14:18:02 <TD> sipa: but the version of openssl they used was apparently not using jacobian coords
2177 2013-03-13 14:18:17 pete2 has joined
2178 2013-03-13 14:19:11 senseless has quit (Read error: Connection reset by peer)
2179 2013-03-13 14:19:16 <gmaxwell> _dr: ah, does avx1 have useful mask and permutes? (I haven't looked since floating point SIMD is only of mild interest to me)
2180 2013-03-13 14:19:28 <sipa> TD: ok so they speed up modular inversion using it for GFp curves
2181 2013-03-13 14:19:45 <sipa> TD: that's cheating, because with jacobian coordinates you only need 1 modular inversion per verify operation anyway
2182 2013-03-13 14:19:51 <jaakkos> are some of the sec curves GF2?
2183 2013-03-13 14:19:55 <sipa> yes
2184 2013-03-13 14:20:05 <Scrat> gmaxwell: you made me read the pdf again, dammit. you're right. 256 bit wide 32/64 bit elements
2185 2013-03-13 14:20:06 <_dr> gmaxwell: not really. they do have some mask instructions but shuffle is really bad. essentially it's two sse units glued together, which doesn't allow random cross-lane shuffling
2186 2013-03-13 14:20:09 <TD> sipa: oh right, yes, i misread one of the sentences at the top
2187 2013-03-13 14:20:09 <jaakkos> might be easier to speed up those...
2188 2013-03-13 14:20:18 defekt has quit (Ping timeout: 248 seconds)
2189 2013-03-13 14:20:27 <jaakkos> but hard fork ;)
2190 2013-03-13 14:20:29 <sipa> TD: which is still dog slow in OpenSSL; my current naive implementation using a precomputed ladder is even faster
2191 2013-03-13 14:20:39 <gmaxwell> jaakkos: lots of the curves are gf2 ... as are almost all the patents on ECC. :P
2192 2013-03-13 14:20:42 <sipa> TD: but GMP is even faster at doing that
2193 2013-03-13 14:20:56 sanchaz has quit (Ping timeout: 255 seconds)
2194 2013-03-13 14:21:03 defekt has joined
2195 2013-03-13 14:21:04 zveda has joined
2196 2013-03-13 14:21:06 <jaakkos> gmaxwell: was the choice made to avoid patent issues?
2197 2013-03-13 14:21:08 tre-spective has joined
2198 2013-03-13 14:21:08 tre-spective has quit (Changing host)
2199 2013-03-13 14:21:08 tre-spective has joined
2200 2013-03-13 14:21:10 Hashdog has joined
2201 2013-03-13 14:21:20 <sipa> TD: 110us for the add/multiply of EC points, and then 26us in OpenSSL for the field inversion, 12us in my own code, or 3us in GMP
2202 2013-03-13 14:21:24 vbuterin has joined
2203 2013-03-13 14:21:31 <flyingkiwiguy> is there some test code I can d/l and experiment with?
2204 2013-03-13 14:21:32 <sipa> i haven't investigated why GMP is so fast
2205 2013-03-13 14:21:52 senseless has joined
2206 2013-03-13 14:22:21 <_dr> so the stuff from their paper is totally useless for our curve? don't you have to perform 256bit mults in your code?
2207 2013-03-13 14:22:33 sanchaz has joined
2208 2013-03-13 14:22:34 <sipa> you certainly do
2209 2013-03-13 14:23:12 <CodeShark> are we comparing GF(p) to GF(2^n) fields?
2210 2013-03-13 14:23:46 <sipa> _dr: i have just looked at the page in the paper referred to by 256x256->512 multiplication, and all i see is something about speeding up modular inverses
2211 2013-03-13 14:24:39 m00p has quit (Ping timeout: 276 seconds)
2212 2013-03-13 14:24:39 rdponticelli has quit (Ping timeout: 276 seconds)
2213 2013-03-13 14:24:52 <TD> i wonder if FMA4 might help
2214 2013-03-13 14:24:53 <TD> http://en.wikipedia.org/wiki/FMA4_instruction_set
2215 2013-03-13 14:24:58 i2pRelay has quit (Remote host closed the connection)
2216 2013-03-13 14:25:11 senseless1 has joined
2217 2013-03-13 14:25:30 i2pRelay has joined
2218 2013-03-13 14:25:39 Phraust has quit (Ping timeout: 264 seconds)
2219 2013-03-13 14:25:48 <_dr> FMA is usually for floats
2220 2013-03-13 14:25:55 Phraust has joined
2221 2013-03-13 14:25:55 Phraust has quit (Changing host)
2222 2013-03-13 14:25:55 Phraust has joined
2223 2013-03-13 14:25:57 fredsted has quit (Ping timeout: 245 seconds)
2224 2013-03-13 14:26:32 senseless has quit (Ping timeout: 256 seconds)
2225 2013-03-13 14:26:38 MrSnoopy has joined
2226 2013-03-13 14:27:15 <MrSnoopy> hi
2227 2013-03-13 14:27:38 <MrSnoopy> i cannot build bitcoind. i cloned the git, tried to make and ran into an error
2228 2013-03-13 14:27:39 <MrSnoopy> alert.cpp:5:29: fatal error: boost/foreach.hpp: No such file or directory
2229 2013-03-13 14:28:00 <CodeShark> have you installed boost?
2230 2013-03-13 14:28:09 banghouse has joined
2231 2013-03-13 14:28:21 <Scrat> libboost-all-dev
2232 2013-03-13 14:28:22 <_dr> http://www.intel.de/content/dam/www/public/us/en/documents/white-papers/carry-less-multiplication-instruction-in-gcm-mode-paper.pdf p12 might be interesting
2233 2013-03-13 14:28:33 banghouse is now known as Guest29939
2234 2013-03-13 14:28:37 fredsted has joined
2235 2013-03-13 14:28:38 fredsted has quit (Excess Flood)
2236 2013-03-13 14:28:49 bernard75 has joined
2237 2013-03-13 14:28:49 bernard75 has quit (Changing host)
2238 2013-03-13 14:28:49 bernard75 has joined
2239 2013-03-13 14:29:04 <sipa> TD: anyway, performance numbers for a typical single block (in IBD), with my secp256k1 code: 11ms for fetching all inputs (mostly from CCoinsCache, not from LevelDB, i think), 46 for verifying inputs (multithreaded, and simultaneously with the fetching), 20ms for writing block data/index to disk/leveldb (sequentially, after validation completes), 3ms for pushing the modified utxo entries to the ccoinscache
2240 2013-03-13 14:29:17 <CodeShark> make sure it's in your include path, MrSnoopy :p
2241 2013-03-13 14:29:32 gimme_bottles has joined
2242 2013-03-13 14:29:33 <MrSnoopy> iwhat that boost thing ?
2243 2013-03-13 14:29:36 <CodeShark> yes
2244 2013-03-13 14:29:48 <sipa> TD: that 20ms should and could be done in parallel with the rest of the next block
2245 2013-03-13 14:30:11 fredsted has joined
2246 2013-03-13 14:30:11 fredsted has quit (Excess Flood)
2247 2013-03-13 14:30:18 <TD> yes but in the steady state writing block data will be very rare
2248 2013-03-13 14:30:26 <sipa> huh?
2249 2013-03-13 14:30:33 putridp has quit (Quit: leaving)
2250 2013-03-13 14:30:36 <sipa> oh you mean mempool tx processing?
2251 2013-03-13 14:30:38 <TD> i mean, i'm thinking to the future
2252 2013-03-13 14:30:38 <TD> yeah
2253 2013-03-13 14:30:52 Guest29939 is now known as banghouse
2254 2013-03-13 14:30:53 <sipa> yeah, that's something else; i'm still only focussing on block processing
2255 2013-03-13 14:31:10 <sipa> there is parallellism possible too, of course for mempool transactions
2256 2013-03-13 14:31:25 <TD> naturally.
2257 2013-03-13 14:31:38 fredsted has joined
2258 2013-03-13 14:31:38 fredsted has quit (Excess Flood)
2259 2013-03-13 14:31:57 <sipa> when verification is completely asynchronous, it's much more easy to make mempool processing use it too
2260 2013-03-13 14:32:12 <sipa> as now it is very tied to the block processing stages
2261 2013-03-13 14:32:36 <TD> anyway
2262 2013-03-13 14:32:37 fredsted has joined
2263 2013-03-13 14:32:38 fredsted has quit (Excess Flood)
2264 2013-03-13 14:32:44 <TD> current speeds are more than good enough for a long, long time
2265 2013-03-13 14:32:48 <sipa> yeah
2266 2013-03-13 14:32:59 i2pRelay has quit (Remote host closed the connection)
2267 2013-03-13 14:33:30 i2pRelay has joined
2268 2013-03-13 14:33:37 fredsted has joined
2269 2013-03-13 14:33:38 fredsted has quit (Excess Flood)
2270 2013-03-13 14:33:45 <sipa> _dr: as i haven't answered before: if someone wants to work on (and maintain) specific assembly-optimized operations in 'my' secp256k1 library, i'm perfectly fine with that... but it's something i'd rather not spend time on myself now
2271 2013-03-13 14:34:11 <jotik> nice: http://nakedsecurity.sophos.com/2013/03/13/anatomy-of-a-problem-bitcoin-loses-25-percent-in-value/
2272 2013-03-13 14:34:35 <sipa> first thing to do is add a shitload of unit tests, for every operation at the field/group/scalar level... so swapping on part with anything becomes relatively easy and trustable
2273 2013-03-13 14:34:37 fredsted has joined
2274 2013-03-13 14:34:38 fredsted has quit (Excess Flood)
2275 2013-03-13 14:35:35 Duly842 has joined
2276 2013-03-13 14:35:37 fredsted has joined
2277 2013-03-13 14:35:38 fredsted has quit (Excess Flood)
2278 2013-03-13 14:36:06 KornKage2 has quit (Ping timeout: 248 seconds)
2279 2013-03-13 14:36:12 <CodeShark> that should even be done in C++, sipa :p
2280 2013-03-13 14:36:19 <sipa> CodeShark: sure
2281 2013-03-13 14:36:28 <sipa> i wasn't suggesting otherwise
2282 2013-03-13 14:36:29 bernard75 has quit ()
2283 2013-03-13 14:36:36 bernard75 has joined
2284 2013-03-13 14:36:36 bernard75 has quit (Changing host)
2285 2013-03-13 14:36:36 bernard75 has joined
2286 2013-03-13 14:36:41 fredsted has joined
2287 2013-03-13 14:36:41 fredsted has quit (Excess Flood)
2288 2013-03-13 14:36:44 <CodeShark> but yeah, good luck debugging a large asm routine without isolating it :p
2289 2013-03-13 14:37:10 <sipa> i mean: we need unit tests for every operation before even considering changing the implementation of one operation by something in assembly
2290 2013-03-13 14:37:27 <CodeShark> absolutely
2291 2013-03-13 14:37:36 <MrSnoopy> db.h:14:20: fatal error: db_cxx.h: No such file or directory
2292 2013-03-13 14:37:39 <MrSnoopy> what this again ?
2293 2013-03-13 14:37:45 <CodeShark> you're missing bdb
2294 2013-03-13 14:37:48 andytoshi has quit (Quit: WeeChat 0.4.0)
2295 2013-03-13 14:37:52 m00p has joined
2296 2013-03-13 14:38:08 fredsted has joined
2297 2013-03-13 14:38:08 fredsted has quit (Excess Flood)
2298 2013-03-13 14:38:38 joe_k has joined
2299 2013-03-13 14:39:07 fredsted has joined
2300 2013-03-13 14:39:08 fredsted has quit (Excess Flood)
2301 2013-03-13 14:39:35 <CodeShark> sipa: the testing should include both deliberately constructed pathological cases that test all the branch possibilities (or as many as possible) as well as randomized blackbox testing
2302 2013-03-13 14:39:38 <MrSnoopy> why is there not a requirement of what you need to build bitcoin
2303 2013-03-13 14:40:07 fredsted has joined
2304 2013-03-13 14:40:08 fredsted has quit (Excess Flood)
2305 2013-03-13 14:40:20 <CodeShark> MrSnoopy, have you tried the install nodes?
2306 2013-03-13 14:40:27 <CodeShark> look in the document directory :p
2307 2013-03-13 14:40:39 <CodeShark> it's all there
2308 2013-03-13 14:40:56 <CodeShark> unless you're porting it to some exotic system
2309 2013-03-13 14:41:03 i2pRelay has quit (Remote host closed the connection)
2310 2013-03-13 14:41:09 fredsted has joined
2311 2013-03-13 14:41:10 fredsted has quit (Excess Flood)
2312 2013-03-13 14:41:20 <MrSnoopy> ok right
2313 2013-03-13 14:41:22 <MrSnoopy> found it tkx
2314 2013-03-13 14:41:32 i2pRelay has joined
2315 2013-03-13 14:41:46 KornKage has joined
2316 2013-03-13 14:42:14 vbuterin has quit (Ping timeout: 245 seconds)
2317 2013-03-13 14:42:15 <joe_k> err
2318 2013-03-13 14:42:37 fredsted has joined
2319 2013-03-13 14:42:38 fredsted has quit (Excess Flood)
2320 2013-03-13 14:43:21 <sipa> CodeShark: whitebox tests for the very-low level operations (e.g. specific ones that cause edge-cases in the 5x52 field representation, for example), and blackbox end-to-end tests with random data (e.g. ECDSA validation with random correct pubkey/sig/message)
2321 2013-03-13 14:43:38 fredsted has joined
2322 2013-03-13 14:43:39 fredsted has quit (Excess Flood)
2323 2013-03-13 14:44:36 blaap has joined
2324 2013-03-13 14:44:38 fredsted has joined
2325 2013-03-13 14:44:39 fredsted has quit (Excess Flood)
2326 2013-03-13 14:44:50 senseless1 has quit (Read error: Connection reset by peer)
2327 2013-03-13 14:45:15 occulta has quit (Quit: .)
2328 2013-03-13 14:45:38 fredsted has joined
2329 2013-03-13 14:45:38 <CodeShark> yes, sipa - and preferably compared against at least two reference implementations
2330 2013-03-13 14:45:38 fredsted has quit (Excess Flood)
2331 2013-03-13 14:46:40 fredsted has joined
2332 2013-03-13 14:46:40 destructure has quit (Quit: Lost terminal)
2333 2013-03-13 14:46:41 fredsted has quit (Excess Flood)
2334 2013-03-13 14:46:42 lantastic has quit (Quit: ãó-ãà -ãà )
2335 2013-03-13 14:47:06 <CodeShark> for correctness, not performance
2336 2013-03-13 14:47:34 senseless has joined
2337 2013-03-13 14:47:40 glitch003_ has quit (Quit: glitch003_)
2338 2013-03-13 14:47:42 senseless has quit (Read error: Connection reset by peer)
2339 2013-03-13 14:48:07 fredsted has joined
2340 2013-03-13 14:48:08 fredsted has quit (Excess Flood)
2341 2013-03-13 14:48:12 <sipa> CodeShark: the unit tests i have were generated by a python implementation :)
2342 2013-03-13 14:49:01 i2pRelay has quit (Remote host closed the connection)
2343 2013-03-13 14:49:08 fredsted has joined
2344 2013-03-13 14:49:09 fredsted has quit (Excess Flood)
2345 2013-03-13 14:49:33 i2pRelay has joined
2346 2013-03-13 14:50:08 fredsted has joined
2347 2013-03-13 14:50:08 fredsted has quit (Excess Flood)
2348 2013-03-13 14:50:23 <CodeShark> I was even using someone's online EC curve calculator at one point as a reference
2349 2013-03-13 14:50:32 <CodeShark> web-based
2350 2013-03-13 14:50:35 <CodeShark> lol
2351 2013-03-13 14:50:55 <_dr> a cybercalculator :)
2352 2013-03-13 14:51:08 fredsted has joined
2353 2013-03-13 14:51:09 fredsted has quit (Excess Flood)
2354 2013-03-13 14:51:14 <gmaxwell> CodeShark: what I'd do is something like build cases by hand to hit each of the carries, max and min values, etc. then use those as the starting point for both randomized tests.
2355 2013-03-13 14:51:25 senseless has joined
2356 2013-03-13 14:51:30 <sipa> gmaxwell: pull requests welcome!
2357 2013-03-13 14:51:40 <_dr> nothing in openssl?
2358 2013-03-13 14:51:48 senseless is now known as Guest2667
2359 2013-03-13 14:52:09 fredsted has joined
2360 2013-03-13 14:52:10 fredsted has quit (Excess Flood)
2361 2013-03-13 14:52:14 <sipa> i'm sure openssl has unit tests too, but none that will hit our edge cases
2362 2013-03-13 14:52:21 <gmaxwell> sipa: waiting on you to finish your implementation!
2363 2013-03-13 14:52:29 Guest2667 has left ()
2364 2013-03-13 14:52:32 <sipa> gmaxwell: well, it "works" :p
2365 2013-03-13 14:52:48 <CodeShark> it would be nice to get rid of the GMP dependencies, though
2366 2013-03-13 14:52:53 <sipa> yeah
2367 2013-03-13 14:52:55 <CodeShark> but GMP is just so damn fast
2368 2013-03-13 14:52:59 <CodeShark> it's hard to beat it
2369 2013-03-13 14:53:09 <Luke-Jr> GMP is part of any Linux system anyway
2370 2013-03-13 14:53:38 fredsted has joined
2371 2013-03-13 14:53:38 fredsted has quit (Excess Flood)
2372 2013-03-13 14:54:23 glitch004 has joined
2373 2013-03-13 14:54:37 fredsted has joined
2374 2013-03-13 14:54:38 fredsted has quit (Excess Flood)
2375 2013-03-13 14:55:54 flyingkiwiguy has quit (Quit: leaving)
2376 2013-03-13 14:56:09 drizzt_ has joined
2377 2013-03-13 14:56:12 <MrSnoopy> what about the blockchain fork ? is it over?
2378 2013-03-13 14:56:15 joehoyle has quit (Quit: Leaving...)
2379 2013-03-13 14:56:21 flyingkiwiguy has joined
2380 2013-03-13 14:56:21 drizztbsd has quit (Ping timeout: 264 seconds)
2381 2013-03-13 14:56:25 <CodeShark> for now :)
2382 2013-03-13 14:56:28 <gmaxwell> MrSnoopy: been over for two days.
2383 2013-03-13 14:56:34 <MrSnoopy> good
2384 2013-03-13 14:56:36 <gmaxwell> (almost)
2385 2013-03-13 14:56:48 <gmaxwell> MrSnoopy: it was just a short term thing, went on for a few hours and then it was over.
2386 2013-03-13 14:56:58 <MrSnoopy> 6 confirms is enough then ?
2387 2013-03-13 14:57:04 i2pRelay has quit (Remote host closed the connection)
2388 2013-03-13 14:57:07 rdponticelli has joined
2389 2013-03-13 14:57:11 <CodeShark> no, this fork was like 25 blocks
2390 2013-03-13 14:57:30 <CodeShark> oh, you mean right now?
2391 2013-03-13 14:57:32 <CodeShark> or in general?
2392 2013-03-13 14:57:35 <gmaxwell> "6 confirms is enough" is just a rule of thumb in any case, it depends on what you're doing.
2393 2013-03-13 14:57:37 <MrSnoopy> yeah now
2394 2013-03-13 14:57:37 i2pRelay has joined
2395 2013-03-13 14:57:46 gimme_bottles_ has joined
2396 2013-03-13 14:57:50 somequestions has joined
2397 2013-03-13 14:58:02 <MrSnoopy> say a transfer from a to b
2398 2013-03-13 14:58:02 <CodeShark> depends on how big the amounts are how much risk you're comfortable with
2399 2013-03-13 14:58:03 dust-otc has quit (Remote host closed the connection)
2400 2013-03-13 14:58:13 gimme_bottles_ has quit (Client Quit)
2401 2013-03-13 14:58:15 <MrSnoopy> ok
2402 2013-03-13 14:58:29 <MrSnoopy> did the miner that inserted the large block lost the mined btc btw ?
2403 2013-03-13 14:58:39 <somequestions> Hi, I'm writing a story for endthelie.com on the last few days in Bitcoin and wanted to get some comments on the OKPay doublespend - Is there someone in particular I should talk to?
2404 2013-03-13 14:59:03 <gmaxwell> somequestions: the person who did it, and OKPAY.
2405 2013-03-13 14:59:06 <CodeShark> as well as all who mined on the abandoned branch, snoopy
2406 2013-03-13 14:59:16 <MrSnoopy> how much did they lose ?
2407 2013-03-13 14:59:17 skeledrew has quit (Ping timeout: 256 seconds)
2408 2013-03-13 14:59:22 <somequestions> I already have their side of the story, I'm interested in issues like "can this happen again"
2409 2013-03-13 14:59:24 <tockitj> somequestions, is it going to be favorable for bitcoin ?
2410 2013-03-13 14:59:24 <CodeShark> 25 btc per block
2411 2013-03-13 14:59:28 Swappermall has joined
2412 2013-03-13 14:59:32 <somequestions> it was a $10,000 transaction roughly
2413 2013-03-13 14:59:36 <somequestions> it's been refunded
2414 2013-03-13 14:59:50 <MrSnoopy> refunded ?
2415 2013-03-13 14:59:50 <tockitj> somequestions, talk to the devs
2416 2013-03-13 14:59:55 <helo> ...
2417 2013-03-13 14:59:57 <somequestions> That's why i'm in bitcoindev
2418 2013-03-13 15:00:07 <somequestions> It's going to be real about bitcoin, personally I am a proponant
2419 2013-03-13 15:00:23 <MrSnoopy> who paid for the refund?
2420 2013-03-13 15:00:34 <somequestions> the recipient
2421 2013-03-13 15:00:36 <grau> somequestions: Things like this happen here every 4 years :)
2422 2013-03-13 15:00:50 <somequestions> Is that true given how fast things are scaling?
2423 2013-03-13 15:00:52 <gmaxwell> somequestions: You should ask questions. People will answer them.
2424 2013-03-13 15:00:58 <somequestions> Ok
2425 2013-03-13 15:01:02 <MrSnoopy> refund in btc only depend on the good will of the recipient with btc right ?
2426 2013-03-13 15:01:04 <kinlo> somequestions: altough you can never ensure no bugs will re-appear, this particular incident will not happen again, as the big mining pools make sure that it will not happen.
2427 2013-03-13 15:01:17 skeledrew has joined
2428 2013-03-13 15:01:19 <CodeShark> it is impossible to completely prevent forks
2429 2013-03-13 15:01:33 <somequestions> You can read the article in progress here, any corrections to the explanation are appreciated as it gets a bit into the mud - This is still a first draft so I haven't edited yet https://docs.google.com/document/d/1H2eKJIfkbhRqp936zDXW3i9kW43Re7cS3sNMDDTo3oA/edit?usp=sharing
2430 2013-03-13 15:01:33 <CodeShark> this is one of many possible issues that could cause a fork
2431 2013-03-13 15:01:34 <shlm> gotta say i think the devs handled it good
2432 2013-03-13 15:01:39 gimme_bottles has quit (Ping timeout: 264 seconds)
2433 2013-03-13 15:02:11 <CodeShark> forks are inevitable - the trick is trying to get them to be as little disruptive as possible
2434 2013-03-13 15:02:12 <gmaxwell> somequestions: "the 6 required confirmations" is kinda odd. Required by what?
2435 2013-03-13 15:02:16 <CodeShark> preferably engineered to be such
2436 2013-03-13 15:02:24 <somequestions> that's the standard,isn't it?
2437 2013-03-13 15:02:33 <helo> the double spend was made possible by a very lucky coincidence of events that probably can't be triggered at will by anyone
2438 2013-03-13 15:02:35 <somequestions> it's not *required* by the protocol, but by most merchants
2439 2013-03-13 15:02:41 PiZZaMaN2K has joined
2440 2013-03-13 15:02:42 <kinlo> somequestions: the 6 required confirmations is not something set in stone, it's just something some people live by
2441 2013-03-13 15:02:48 PiZZaMaN2K has quit (Changing host)
2442 2013-03-13 15:02:48 PiZZaMaN2K has joined
2443 2013-03-13 15:02:50 <somequestions> I'm trying to keep it simple
2444 2013-03-13 15:03:06 <kinlo> somequestions: also, the limit of 512kb as bloksize is plain wrong. if you want to keep it simple, just say "too large"
2445 2013-03-13 15:03:12 <gmaxwell> "a confirmation is a return message" ugh, this isn't technically true and the difference is very relevant in understanding cases like this.
2446 2013-03-13 15:03:36 <somequestions> Is that accurate? I thought you couldn't have atomic changes in BDB over 512k without manually changing the client?
2447 2013-03-13 15:03:44 <gmaxwell> as kinlo says too. E.g. load testnet, you'll see there are 1MB blocks that 0.7.2 is perfectly happy with.
2448 2013-03-13 15:03:58 <somequestions> so why did it break this time?
2449 2013-03-13 15:04:17 <imsaguy> did those testnet 1MB blocks have the same number of txs as prodnet's failure?
2450 2013-03-13 15:04:19 <helo> a technical problem occurred that required many nodes to be reset, so they lost their knowledge of unconfirmed transactions. this allowed the double spend to propagate without being dropped, as it would normally be.
2451 2013-03-13 15:04:25 glitch004 has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
2452 2013-03-13 15:04:27 <kinlo> somequestions: as I already replied recently on #bitcoin, because there weren't enough locks in the code foreseen to process the block
2453 2013-03-13 15:04:27 <gmaxwell> "rejected by miners running on 0.7", this is true, but it's not the most critical element. It was rejected by ALL nodes, not running 0.8.
2454 2013-03-13 15:04:32 <gmaxwell> Not just rejected by miners.
2455 2013-03-13 15:04:58 <gmaxwell> imsaguy: They didn't hit as many locks, which is the relevant thing.
2456 2013-03-13 15:04:59 <somequestions> I'm using the term "participants" interchangably with nodes for simplicity
2457 2013-03-13 15:05:06 i2pRelay has quit (Remote host closed the connection)
2458 2013-03-13 15:05:08 <somequestions> I don't understand the "locks" concept
2459 2013-03-13 15:05:15 <kinlo> somequestions: but that's probably too complex for your article, so either be vague, or be too technical for your audience
2460 2013-03-13 15:05:29 <somequestions> OK, i'll simplify that part
2461 2013-03-13 15:05:36 i2pRelay has joined
2462 2013-03-13 15:05:46 <gmaxwell> somequestions: the development isn't "all-volunteer", Gavin is paid by the bitcoin foundation, which is a professional org funded by large bitcoin businesses.
2463 2013-03-13 15:06:01 <somequestions> Ah, I did not realize he was being compensated these days
2464 2013-03-13 15:06:08 <sipa> gmaxwell, Luke-Jr: block 225203 modified 4154 transactions
2465 2013-03-13 15:06:09 <somequestions> That wasn't the case 18 months ago, right?
2466 2013-03-13 15:06:10 defekt has quit (Ping timeout: 248 seconds)
2467 2013-03-13 15:06:15 <CodeShark> is he the only dev paid, gmaxwell?
2468 2013-03-13 15:06:17 <sipa> gmaxwell, Luke-Jr: interesting test case, i guess
2469 2013-03-13 15:06:32 <gmaxwell> CodeShark: AFAIK.
2470 2013-03-13 15:06:43 <MrSnoopy> compilation worked. thanks for your help
2471 2013-03-13 15:07:00 <MrSnoopy> now i'm 0.8 è
2472 2013-03-13 15:07:05 defekt has joined
2473 2013-03-13 15:07:08 <gmaxwell> I don't know how important the detail is, just pedantry.
2474 2013-03-13 15:07:08 MKCoin has quit (Ping timeout: 252 seconds)
2475 2013-03-13 15:07:18 <somequestions> thats fine, I want to be correct
2476 2013-03-13 15:07:36 HiWEB has joined
2477 2013-03-13 15:07:39 <somequestions> I appreciate the editing ;) I've been into bitcoin for 2 years but some of the nitty gritty escapes me still
2478 2013-03-13 15:08:00 <tockitj> i wish developers could review all technical articles like this
2479 2013-03-13 15:08:08 <gmaxwell> somequestions: In any case... Bitcoin is a consensus systems, confirmations are a rough measure of the amount of computing power which is supporting a particular position. (and thus the amount of computing power an attacker would need to undo a position and replace it)
2480 2013-03-13 15:08:08 wesgarrison has left ()
2481 2013-03-13 15:08:46 <Luke-Jr> tockitj: I try to, for Bitcoin Magazine
2482 2013-03-13 15:09:01 bernard75 has left ()
2483 2013-03-13 15:09:15 <MrSnoopy> can i replace the bitcoind i have from the new one ? no need to redownload the blockchain or evthg ?
2484 2013-03-13 15:09:24 <MrSnoopy> wallet version ?
2485 2013-03-13 15:09:36 <kinlo> just upgrade your client, it will work
2486 2013-03-13 15:09:41 <Luke-Jr> MrSnoopy: the 0.8 upgrade will reindex
2487 2013-03-13 15:09:42 <MrSnoopy> ok cool
2488 2013-03-13 15:09:44 <Luke-Jr> takes a few hours
2489 2013-03-13 15:09:49 <CodeShark> MrSnoopy: you will need to sync again...but it will be faster than 0.7
2490 2013-03-13 15:09:50 <MrSnoopy> :/
2491 2013-03-13 15:09:57 <Luke-Jr> CodeShark: not sync, just inded
2492 2013-03-13 15:09:59 <Luke-Jr> index
2493 2013-03-13 15:10:01 <MrSnoopy> i hope so. it took me ages to download the blockchain
2494 2013-03-13 15:10:02 joehoyle has joined
2495 2013-03-13 15:10:03 <imsaguy> gmaxwell: has a new test been added (or existing test updated) to test for locking in the future? I know bdb has been replaced in .8 but does it make sense to be testing for this in the future?
2496 2013-03-13 15:10:12 <tockitj> Luke-Jr, Bitcoin Mag is nice
2497 2013-03-13 15:10:21 <kinlo> MrSnoopy: 0.8 is a LOT faster in downloading the blockchain
2498 2013-03-13 15:10:21 <tockitj> (just checking)
2499 2013-03-13 15:10:24 <Luke-Jr> tockitj: Bitcoin is bigger than the blockchain
2500 2013-03-13 15:10:27 <somequestions> So how about this for the explanation, omitting the 512k number ."..not everyone upgraded their software. Berkeley Database (BDB) was used for versions of the bitcoin software .7 and below. Some time ago it was discovered that BDB will accept a maximum entry size per block"
2501 2013-03-13 15:10:31 <MrSnoopy> good news
2502 2013-03-13 15:10:39 <gmaxwell> imsaguy: We will add tests for blocks changing lots of transactions of course.
2503 2013-03-13 15:10:49 <MrSnoopy> and about the wallet ? no need to upgrade or anything?
2504 2013-03-13 15:10:56 <kinlo> somequestions: what do you mean by "some time ago it was discovered" ?
2505 2013-03-13 15:10:56 <CodeShark> somequestions: that's not strictly true - it was actually a particular edge case
2506 2013-03-13 15:11:07 <Scrat> somequestions: <+sipa> there are up to 2n locks necessary for n updates to a btree, and our limit was 10000 locks
2507 2013-03-13 15:11:11 <Luke-Jr> tockitj: 6 confirms takes on average 1 hour
2508 2013-03-13 15:11:12 <somequestions> Well it was fixed in .8 so obviously it was realized it might be an issue
2509 2013-03-13 15:11:12 <gmaxwell> somequestions: 'Some time ago' noâ we didn't know about this particular limit, or 0.8 would have obeyed it.
2510 2013-03-13 15:11:24 <kinlo> somequestions: AFAIK nobody knew this was going to happen
2511 2013-03-13 15:11:26 <Luke-Jr> err
2512 2013-03-13 15:11:28 <Scrat> that's the limit, not block size
2513 2013-03-13 15:11:33 <Luke-Jr> somequestions: last 2 comments for you
2514 2013-03-13 15:11:40 <somequestions> why was it changed from .7 to .8 if nobody thought it was a problem?
2515 2013-03-13 15:11:42 <imsaguy> gmaxwell: you say 'of course' but if it was so common sense, then why wasn't it done?
2516 2013-03-13 15:11:44 <gmaxwell> We had previously thought that the locks used in prior versions were equal to 'infinite'.
2517 2013-03-13 15:11:56 <Luke-Jr> somequestions: no, the fix was accidental
2518 2013-03-13 15:12:01 <gmaxwell> imsaguy: the of course is _now_. We did not know there was a relevant limit there.
2519 2013-03-13 15:12:04 <kinlo> somequestions: it was a limitation in BDB, and 0.8 doesn't use BDB for that code
2520 2013-03-13 15:12:11 <Luke-Jr> somequestions: also note that MOST nodes had not upgraded..
2521 2013-03-13 15:12:15 <Luke-Jr> only a minority were using 0.8
2522 2013-03-13 15:12:19 <gmaxwell> somequestions: 0.8 is basically a rewrite of the storage and validation engine.
2523 2013-03-13 15:12:20 <kinlo> somequestions: as luke says, the fix for that limitation was purely accidental
2524 2013-03-13 15:12:32 <somequestions> OK
2525 2013-03-13 15:12:36 <imsaguy> Luke-Jr: that isn't a good thing. :-/
2526 2013-03-13 15:12:51 <gmaxwell> somequestions: part of the problem here is that most nodes are <0.8 but most _mining_ was on 0.8 because mining is not as well distributed as the whole system.
2527 2013-03-13 15:13:06 i2pRelay has quit (Remote host closed the connection)
2528 2013-03-13 15:13:16 <gmaxwell> (as right now nearly half the hashpower is held by a single party. :( )
2529 2013-03-13 15:13:37 i2pRelay has joined
2530 2013-03-13 15:13:48 <somequestions> well damn
2531 2013-03-13 15:13:54 <somequestions> that doesn't sound very good to be honest
2532 2013-03-13 15:14:02 <Luke-Jr> somequestions: "Why'd it happen? Because not everyone upgraded their software" is wrong; it's because an unknown protocol rule was changed by the development team accidentally
2533 2013-03-13 15:14:20 MKCoin has joined
2534 2013-03-13 15:14:25 <somequestions> So I know its unforeseeable, but this doesn't say good things about the future as bitcoin scales
2535 2013-03-13 15:14:27 <gmaxwell> (and mining was fast to upgrade because 0.8 is such an enormous improvement to block processing speed, so upgrading to 0.8 was a competative advantage for miners)
2536 2013-03-13 15:14:27 moartr4dez is now known as IRSAgent
2537 2013-03-13 15:14:32 <somequestions> It was handled well but its a small community
2538 2013-03-13 15:14:43 <gmaxwell> somequestions: I'm not sure I follow your thinking there.
2539 2013-03-13 15:15:03 <gmaxwell> I think the concerns here are mostly scale invariant.
2540 2013-03-13 15:15:06 <somequestions> Well the doublespend would have been a big deal if people hadn't been involved with the community
2541 2013-03-13 15:15:11 t7 has quit (Quit: Leaving)
2542 2013-03-13 15:15:27 <kinlo> gmaxwell: half the hashpower?
2543 2013-03-13 15:15:30 <somequestions> so if scaling causes these problems, and they're not foreseen beforehand, it just seems like lacking a more comprehensive alert system it could lead to real doublespends
2544 2013-03-13 15:15:34 n1c has quit (Quit: peace out)
2545 2013-03-13 15:15:46 <somequestions> this is just conversational, not going in the article
2546 2013-03-13 15:15:59 <epscy> "off the record"
2547 2013-03-13 15:16:00 <gmaxwell> somequestions: I don't think that you can say scaling causes these problem.
2548 2013-03-13 15:16:01 <somequestions> i'm not taking any direct quotes form this fyi
2549 2013-03-13 15:16:04 flyingkiwiguy has quit (Quit: leaving)
2550 2013-03-13 15:16:12 <somequestions> gmaxwell what caused this problem?
2551 2013-03-13 15:16:13 sl1982 has quit (Read error: Connection reset by peer)
2552 2013-03-13 15:16:21 <Luke-Jr> somequestions: unknown protocol rules
2553 2013-03-13 15:16:28 IRSAgent is now known as moartr4dez
2554 2013-03-13 15:16:28 <somequestions> so that seems like a problem
2555 2013-03-13 15:16:37 <epscy> there known unknowns, and unknown unknowns
2556 2013-03-13 15:16:40 <somequestions> why were there unknown protocol rules in an open source protocol?
2557 2013-03-13 15:16:41 <gmaxwell> The problems were caused by inadequate testing of 0.8 and prior versions which failed to uncover unintended behavior in the software.
2558 2013-03-13 15:16:43 <Luke-Jr> yes, that's why we've been working on getting unit tests and documenting the protocol
2559 2013-03-13 15:16:54 flyingkiwiguy has joined
2560 2013-03-13 15:16:58 <Scrat> that double spend had something to do with miners shutting down 0.7 nodes, not the chain fork. someone correct me if I'm wrong
2561 2013-03-13 15:17:11 <gmaxwell> somequestions: because they arose implicitly out of complicated interactions in the software, part of it not written as a part of the bitcoin project.
2562 2013-03-13 15:17:14 <Luke-Jr> Scrat: you're wrong
2563 2013-03-13 15:17:19 <somequestions> ah
2564 2013-03-13 15:17:29 flyingkiwiguy has quit (Client Quit)
2565 2013-03-13 15:17:36 <somequestions> So for example, BDB wasn't written for the bitcoin protocol
2566 2013-03-13 15:17:38 gntbynynybt has quit (Ping timeout: 245 seconds)
2567 2013-03-13 15:17:42 <somequestions> it's just an element used by it
2568 2013-03-13 15:17:53 <somequestions> the unknown was in the element, not the bitcoin protocol itself
2569 2013-03-13 15:17:55 <Luke-Jr> correct
2570 2013-03-13 15:17:58 <gmaxwell> Scrat: responsible vs responsible, the fact of the fork alone made it possible to get conflicting spends with >=6 confirms.
2571 2013-03-13 15:18:03 flyingkiwiguy has joined
2572 2013-03-13 15:18:09 <Luke-Jr> somequestions: well, it's part of the protocol
2573 2013-03-13 15:18:17 <somequestions> but not a proprietary part
2574 2013-03-13 15:18:24 <gmaxwell> somequestions: right. It's part of the protocol to the extend that it does things we don't expect it to do.
2575 2013-03-13 15:18:26 <somequestions> which is what the bitcoin dev focuses on, right?
2576 2013-03-13 15:18:28 <somequestions> right
2577 2013-03-13 15:18:33 <gmaxwell> somequestions: well no, 'a proprietary part'
2578 2013-03-13 15:18:53 awsas has joined
2579 2013-03-13 15:18:57 sl1982 has joined
2580 2013-03-13 15:19:07 <gmaxwell> somequestions: bitcoin is a consensus system, basically all these hundred thousand validating nodes act like _one computer_ exactly executing this great big algorithim and all must come to _exactly_ the same decision.
2581 2013-03-13 15:19:08 <somequestions> when I say proprietary, I mean someone coded it specifically for bitcoin vs. it being included because it already existed and performed desired functions
2582 2013-03-13 15:19:44 JDuke128 has quit (Quit: [BB])
2583 2013-03-13 15:19:45 <somequestions> thats an interesting explanation
2584 2013-03-13 15:19:52 <gmaxwell> somequestions: so BDB's "locking" behavior was able to influence this decisionâ it shouldn't have been, but BDB behaved in ways that were not forseen and so it did.
2585 2013-03-13 15:20:07 <somequestions> ok
2586 2013-03-13 15:20:09 t7 has joined
2587 2013-03-13 15:20:42 <gmaxwell> The only way to avoid that is to be smart and careful and do a lot of testingâ of course we _do_ do a lot of testing, but one problems with testing is that when you're testing for things you don't know you don't know you sometimes miss things.
2588 2013-03-13 15:20:55 <somequestions> So getting back to my earlier questions - Do you think we'll see more doublespends going forward, or are they literally once-in-a-billion confluences of circumstance
2589 2013-03-13 15:20:57 <gmaxwell> We tested very large blocks. But not blocks that were large in the right way to trigger the misbehavior.
2590 2013-03-13 15:21:09 i2pRelay has quit (Remote host closed the connection)
2591 2013-03-13 15:21:11 sl1982 has quit (Read error: Connection reset by peer)
2592 2013-03-13 15:21:32 <awsas> helllo
2593 2013-03-13 15:21:41 i2pRelay has joined
2594 2013-03-13 15:21:50 <gmaxwell> somequestions: there are "doublespends" every day. You probably mean to ask if doublespends with >5 confirms? I would be surprised if we never had any >5 block reorgs in the future.
2595 2013-03-13 15:21:51 manet has quit (Ping timeout: 276 seconds)
2596 2013-03-13 15:22:04 <somequestions> I mean doublespends that work
2597 2013-03-13 15:22:08 <Luke-Jr> somequestions: if someone had known about this divergence beforehand, they could have orchestrated an elaborate double-spend to rip everyone off - but once fixed, that risk is gone
2598 2013-03-13 15:22:25 <gmaxwell> somequestions: When people accept 0 confirmed transactions doublespends that work happen easily.
2599 2013-03-13 15:22:39 <somequestions> do people do that?
2600 2013-03-13 15:22:39 sl1982 has joined
2601 2013-03-13 15:22:50 <gmaxwell> somequestions: If OKPAY had required 20 confirms on a $10,000 deposit there would have been no doublespend here (for example)
2602 2013-03-13 15:23:12 <gmaxwell> somequestions: some people do, yes... because they've decided the risk is worth it. Or because they don't understand the risk.
2603 2013-03-13 15:23:14 <somequestions> I believe that transaction got 30+ confirmations
2604 2013-03-13 15:23:36 <gmaxwell> somequestions: Did it? I thought it had 12.
2605 2013-03-13 15:23:43 <Luke-Jr> gmaxwell: probably Deepbit
2606 2013-03-13 15:23:47 m00p has quit (Ping timeout: 246 seconds)
2607 2013-03-13 15:23:49 <somequestions> I could be wrong, read that thread last night
2608 2013-03-13 15:24:01 <somequestions> either way, it seemed like more than most businesses I work with require
2609 2013-03-13 15:24:11 <gmaxwell> The question is how many it had before it was reorged out, not how many that chain had eventually.
2610 2013-03-13 15:24:13 <imsaguy> All of those 0 conf mtgox deposit services could have been screwed royally
2611 2013-03-13 15:24:44 <gmaxwell> imsaguy: they can already be screwed royally. We get 1 block reorgs every day.
2612 2013-03-13 15:24:45 <somequestions> So given what happened here, do you think it's safe to operate a 0 confirmation service in the bitcoin world?
2613 2013-03-13 15:24:46 <Luke-Jr> imsaguy: no
2614 2013-03-13 15:24:55 <somequestions> Seems like we have to learn through trials by fire given all the thefts
2615 2013-03-13 15:25:04 <Cusipzzz> somequestions: it was never safe, even before this incident
2616 2013-03-13 15:25:08 <Luke-Jr> imsaguy: wait, what do you mean? people who accept 0-conf from mtgox would be fine
2617 2013-03-13 15:25:12 <imsaguy> no
2618 2013-03-13 15:25:17 <gmaxwell> somequestions: I don't think there is anything special learned from this wrt number of confirms.
2619 2013-03-13 15:25:20 <imsaguy> Luke-Jr, you're not getting it
2620 2013-03-13 15:25:33 <imsaguy> Person A sends 10 coins to 0confbusiness
2621 2013-03-13 15:25:35 <gmaxwell> Luke-Jr: there are services where you give them 0 conf bitcoin and they give you mtgox btc.
2622 2013-03-13 15:25:41 <somequestions> OK
2623 2013-03-13 15:25:46 <Luke-Jr> gmaxwell: wtf
2624 2013-03-13 15:25:46 awsas has quit (Quit: Page closed)
2625 2013-03-13 15:25:47 <gmaxwell> Luke-Jr: these services are moronic and are probably scame.
2626 2013-03-13 15:25:49 <gmaxwell> er scams.
2627 2013-03-13 15:25:52 MrSnoopy has quit (Quit: Page closed)
2628 2013-03-13 15:25:54 <imsaguy> 0confbusiness transfers 10 coins to Person A's mtgox
2629 2013-03-13 15:26:00 <imsaguy> gmaxwell: coingenuity runs one
2630 2013-03-13 15:26:07 <gmaxwell> (scames as in long cons that will eventually doublespend themselves to make off with inflight money)
2631 2013-03-13 15:26:21 <gmaxwell> imsaguy: I will yell at him them.
2632 2013-03-13 15:26:28 <k9quaint> I am mildly surprised at the general lack of media coverage of the rollback
2633 2013-03-13 15:26:38 <gmaxwell> k9quaint: too freeking complicated.
2634 2013-03-13 15:26:48 <Scrat> gmaxwell: this is his http://fasterco.in/
2635 2013-03-13 15:26:48 <gmaxwell> We can hardly get people who come in here to understand it. :P
2636 2013-03-13 15:26:52 <somequestions> Ok.
2637 2013-03-13 15:27:08 <somequestions> It is very complicated
2638 2013-03-13 15:27:10 <helo> good luck distilling this down in a useful way :)
2639 2013-03-13 15:27:14 <somequestions> But there is interest out there
2640 2013-03-13 15:27:30 <gmaxwell> somequestions: in any case, certantly there are things that can be improved in how merchants act. They should freeze withdraws when something weird starts happening. As MTGOX did, for example.
2641 2013-03-13 15:27:55 <somequestions> It's all good. I'm also starting thedailybitcoin.com, bitcoin is obviously the future :)
2642 2013-03-13 15:28:05 <gmaxwell> (er I meant hold deposits not freeze withdraws)
2643 2013-03-13 15:28:17 dust-otc has joined
2644 2013-03-13 15:28:18 <somequestions> Is there any intent or talk about formalizing a notification system when things like this happen?
2645 2013-03-13 15:28:22 <imsaguy> gmaxwell: well realistically, both are probably prudent.
2646 2013-03-13 15:28:27 <Ant0> gmaxwell the reason of the fork is know now? I see the 0.8 message has been removed from bitcoin.org
2647 2013-03-13 15:28:33 cheako has joined
2648 2013-03-13 15:28:42 <imsaguy> somequestions: there already is a notification system built in
2649 2013-03-13 15:28:48 <somequestions> Did it work?
2650 2013-03-13 15:28:49 <imsaguy> not all clients expose it.
2651 2013-03-13 15:28:51 JDuke128 has joined
2652 2013-03-13 15:28:52 <gmaxwell> somequestions: it's like writing articles that explain activities in parlementary government to the public. The system has arcane rules that no one cares about except the people deeply involved, because the rules seldom matter except when they do.
2653 2013-03-13 15:28:55 <somequestions> clearly OKpay didnt get the memo
2654 2013-03-13 15:29:12 i2pRelay has quit (Remote host closed the connection)
2655 2013-03-13 15:29:48 i2pRelay has joined
2656 2013-03-13 15:29:52 <imsaguy> somequestions, there will always be those instances where people didn't get the memo.
2657 2013-03-13 15:30:00 <sipa> Ant0: there's a very likely theory
2658 2013-03-13 15:30:03 <gmaxwell> Ant0: It's all fine now, the network reorged onto the chain acceptable to all nodes. We'll be issuing some updates to prevent this problem from happening again. (right now it can't happen again because enough mining is on pre-0.8 again)
2659 2013-03-13 15:30:11 <somequestions> So no plans or talk of changing the way word gets out?
2660 2013-03-13 15:30:28 <Ant0> ah great
2661 2013-03-13 15:30:41 <gmaxwell> somequestions: gavin has been working on a postmortem document who's first point is going to be able improving some of that.
2662 2013-03-13 15:30:52 <Ant0> and is finally anything going to be done with SatoshiDice or just leave it as is?
2663 2013-03-13 15:30:54 <helo> i suspect there will be a "defense in depth" reponse to all of the lessons learned
2664 2013-03-13 15:30:54 <Luke-Jr> somequestions: it used to be, that Gavin could disable clients remotely - that might get reconsidered as an optional feature, maybe
2665 2013-03-13 15:30:56 <somequestions> Great
2666 2013-03-13 15:31:02 <imsaguy> Ant0: leave as is.
2667 2013-03-13 15:31:03 <gmaxwell> helo: indeed.
2668 2013-03-13 15:31:08 <Ant0> ah
2669 2013-03-13 15:31:14 HiWEB has quit (Read error: Connection reset by peer)
2670 2013-03-13 15:31:31 <imsaguy> satoshidice is a good test of things.
2671 2013-03-13 15:31:46 <Luke-Jr> imsaguy: not really
2672 2013-03-13 15:31:50 <gmaxwell> Ant0: thats a question for other channels. Nothing is even suggested by anyone credible as something to do in the software.
2673 2013-03-13 15:31:53 <helo> kind of like a home intruder is a good test of things
2674 2013-03-13 15:31:59 <Luke-Jr> ^
2675 2013-03-13 15:32:01 agath_pd has joined
2676 2013-03-13 15:32:03 agath has quit (Ping timeout: 256 seconds)
2677 2013-03-13 15:32:19 <gmaxwell> imsaguy: a good test would be a transaction flood on tesetnet. Blowing up the production network is not a test. :P (though SD isn't the proximal cause here)
2678 2013-03-13 15:32:27 <imsaguy> helo: to date, they aren't breaking any rules
2679 2013-03-13 15:32:31 <Luke-Jr> imsaguy: they are so
2680 2013-03-13 15:32:32 hanti is now known as HANTI
2681 2013-03-13 15:32:33 <imsaguy> so its more like a whitehat hacker doing pentesting
2682 2013-03-13 15:32:34 <gmaxwell> stop
2683 2013-03-13 15:32:43 <gmaxwell> I'm going to start kicking, take that crap to #bitcoin
2684 2013-03-13 15:32:45 glitch004 has joined
2685 2013-03-13 15:33:05 <imsaguy> my bad
2686 2013-03-13 15:33:26 <Ant0> ok gmaxwell sorry
2687 2013-03-13 15:34:04 <Scrat> Luke-Jr: you can easily accept 0conf if you're not selling anything physical. If you detect a double spend just close their account
2688 2013-03-13 15:34:10 <gmaxwell> It's okay. It's an intersting subject, I'm opinionated too.. but as a matter of transparency we do really need to keep this channel more focused on technology and the technical maintance of the bitcoin infrastructure.
2689 2013-03-13 15:34:27 <Luke-Jr> Scrat: on the contrary, 0conf is more acceptable for physical products
2690 2013-03-13 15:34:28 <gmaxwell> Scrat: better to generalize to "not taking an irreversable action"
2691 2013-03-13 15:34:35 <Luke-Jr> Scrat: since you have time before shipping
2692 2013-03-13 15:35:16 <Dyaheon> disabling clients remotely? couldn't the clients themselves detect forks and stop to wait for user input if there was a say 5 blocks fork?
2693 2013-03-13 15:35:30 <k9quaint> Luke-Jr: but if you have time before shipping, why would you "accept" 0 conf?
2694 2013-03-13 15:35:30 <gmaxwell> Scrat: you should have an adeqaute amount of confirmation before you take an irreversable action. Adequate depends on your risks (value of action, ability to punish cheating, state of the network).
2695 2013-03-13 15:36:04 <imsaguy> It all boils down to acceptable risk.
2696 2013-03-13 15:36:05 <Luke-Jr> k9quaint: well, point is you act immediately on 0conf and tell the user "ok", then do a second check as you're about to ship
2697 2013-03-13 15:36:16 <gmaxwell> Dyaheon: no, that doesn't really helpâ you can't reliably detect. And then it turns non-malicious things into bad dos attacks. Somewhat paradoxically when there is consensus ambiguity the important thing to do is resolve it as fast as possible.
2698 2013-03-13 15:36:38 Diablo-D3 has joined
2699 2013-03-13 15:36:43 <imsaguy> uh oh
2700 2013-03-13 15:37:02 <Dyaheon> but you need a lot of power to create a 5 block fork, right? those shouldn't happen very often
2701 2013-03-13 15:37:14 i2pRelay has quit (Remote host closed the connection)
2702 2013-03-13 15:37:38 <gmaxwell> Dyaheon: by that same argument you don't need to guard against it. :P having the whole network shut down randomly because of something rare doesn't actually help! :P
2703 2013-03-13 15:37:48 i2pRelay has joined
2704 2013-03-13 15:38:04 <gmaxwell> Dyaheon: now, if instead you propose being more agressive about discovering long forks and alerting users and applications that something weird is going onâ thats interesting and would be useful.
2705 2013-03-13 15:38:34 <k9quaint> gmaxwell: make fork detection like the terror alert, its orange in the blockchain and nobody knows what that means ;)
2706 2013-03-13 15:38:37 <Luke-Jr> gmaxwell: to be fair, I think it would be wise to enable safe mode if a 5-deep fork with reasonable difficulty is detected
2707 2013-03-13 15:38:54 <Luke-Jr> gmaxwell: say, if it began in the last 100 blocks
2708 2013-03-13 15:39:20 protus has quit (Remote host closed the connection)
2709 2013-03-13 15:39:43 <imsaguy> How would you go about detecting whether one of your peers is lying to you about seeing a different fork? Consensus among your local peers?
2710 2013-03-13 15:40:09 wyrag has quit (Ping timeout: 245 seconds)
2711 2013-03-13 15:40:18 <gmaxwell> Luke-Jr: forks cannot be reliably detected right now.
2712 2013-03-13 15:40:30 <gmaxwell> Luke-Jr: if all your peers were 0.7 or all were 0.8 you wouldn't have seen it.
2713 2013-03-13 15:40:40 <k9quaint> gmaxwell: you can detect if people think there is a fork, but not from inside the blockchain
2714 2013-03-13 15:40:42 <Luke-Jr> gmaxwell: does that make a difference?
2715 2013-03-13 15:40:42 pegu has quit (Read error: Connection reset by peer)
2716 2013-03-13 15:40:58 <Scrat> split brain scenario requires an arbiter :p
2717 2013-03-13 15:41:07 <Luke-Jr> gmaxwell: imperfect safeguards must die because they're not 100% effective? :p
2718 2013-03-13 15:41:25 <somequestions> OK, so here's the rewrite on the problem without the 512k number - Does this jive with reality? Call it growing pains - Berkeley Database (BDB) was used for versions of the bitcoin software .7 and below and as luck would have it, will accept a maximum size per block. The size of a block is determined in large part by how many transactions it includes, so as transaction volume has gone up so has the potential block size.
2719 2013-03-13 15:41:29 <gmaxwell> Luke-Jr: Sometimes! they give false security, so that must be weighed.
2720 2013-03-13 15:41:36 ZedsterX has quit (Read error: Connection reset by peer)
2721 2013-03-13 15:41:46 <gmaxwell> somequestions: it's not a maximum size. :(
2722 2013-03-13 15:41:53 <somequestions> What is it then?
2723 2013-03-13 15:41:58 ZedsterX has joined
2724 2013-03-13 15:42:15 <gmaxwell> somequestions: it's an internal limit to the number of changed transactions. E.g. the number of old txn being consumed and new txn being created.
2725 2013-03-13 15:42:20 SchmalzTech has quit ()
2726 2013-03-13 15:42:30 <somequestions> ok, thats actually simpler
2727 2013-03-13 15:43:13 <gmaxwell> Luke-Jr: also, why safe mode? I don't want _my_ nodes going into safe mode. Why not just alert the caller and freeze confirmation counts displayed in the wallet?
2728 2013-03-13 15:43:43 <gmaxwell> and let people behave intelligently. Safe mode doesn't prevent you from shipping a good in any case. :P
2729 2013-03-13 15:44:02 <Luke-Jr> gmaxwell: -disablesafemode then
2730 2013-03-13 15:44:15 <Luke-Jr> gmaxwell: I don't have context for your request on -dev :/
2731 2013-03-13 15:44:17 <Luke-Jr> ML
2732 2013-03-13 15:44:25 GMP has quit (Remote host closed the connection)
2733 2013-03-13 15:44:26 <Luke-Jr> Peter Todd's email did not make it to me
2734 2013-03-13 15:44:27 <somequestions> gmaxwell: what's the desired time between issuance of blocks?
2735 2013-03-13 15:44:37 <gmaxwell> Luke-Jr: he's debating the blocksize increase.
2736 2013-03-13 15:45:06 <Jezzz> it would be trivial to query a few nodes of different versions for forks
2737 2013-03-13 15:45:06 <gmaxwell> somequestions: the system is designed to achieve a mean time of 10 minutes between blocks. (but thats a mean, it can be substantially faster or slower)
2738 2013-03-13 15:45:13 <somequestions> thats fine
2739 2013-03-13 15:45:16 i2pRelay has quit (Remote host closed the connection)
2740 2013-03-13 15:45:18 Descry has joined
2741 2013-03-13 15:45:32 rdymac has joined
2742 2013-03-13 15:45:47 i2pRelay has joined
2743 2013-03-13 15:46:25 <gmaxwell> Jezzz: If you are connected to people with a fork you'll know about it. But you have to be connected. If the fork is only accepted by a minority of nodes there is a good chance you won't know.
2744 2013-03-13 15:46:32 protus has joined
2745 2013-03-13 15:46:40 <egecko> you could if the blocks had some form of data saying what chain a miner was mining when the block was made
2746 2013-03-13 15:46:41 egecko has quit (Read error: Connection reset by peer)
2747 2013-03-13 15:46:45 <gmaxwell> Jezzz: we could forward notices of such things, but thought would need to be given to prevent dos attack.
2748 2013-03-13 15:46:47 <Jezzz> gmaxwell: good point
2749 2013-03-13 15:47:06 <gmaxwell> egecko: uhhh. they _do_. 0_o.
2750 2013-03-13 15:47:27 <Jezzz> gmaxwell: but if it's only accepted by a minority of nodes, then it's not really a worry :P
2751 2013-03-13 15:47:33 <helo> bitcoin-alert@lists.sourceforge.net
2752 2013-03-13 15:47:42 <gmaxwell> the bigger problem is that because such events are by definition rare its the case that all these sites will handle them poorly if at all.
2753 2013-03-13 15:47:42 <somequestions> Call it growing pains - Berkeley Database (BDB) was used for versions of the bitcoin software .7 and below and as luck would have it, will accept a maximum number of transactions per block. The protocolâs self adjusting difficulty seeks to issue a block every 10 minutes, so as bitcoin usage has increased, so has the number of transactions contained in the average block.
2754 2013-03-13 15:48:02 <gmaxwell> Jezzz: yes, it is. This fork was accepted by a minority of _nodes_ but it was an issue because it was accepted by a majority of miners.
2755 2013-03-13 15:48:21 <Jezzz> gmaxwell: right. my mistake
2756 2013-03-13 15:48:27 manet has joined
2757 2013-03-13 15:48:34 npouillard has quit (Remote host closed the connection)
2758 2013-03-13 15:48:47 pete78 is now known as pjr
2759 2013-03-13 15:48:56 <gmaxwell> Jezzz: amiller had proposed some ideas for self adjusting block time which would incentivize miners to include notices of other forks in their blocks.
2760 2013-03-13 15:49:07 npouillard has joined
2761 2013-03-13 15:49:18 <gmaxwell> but thats rocket science right now, not something related to the pratical bitcoin system, at least today.
2762 2013-03-13 15:49:31 pjr is now known as pjr1972
2763 2013-03-13 15:49:43 <lianj> gmaxwell: is eventually a hard fork planned in the future?
2764 2013-03-13 15:49:47 <helo> this is assuming the next problem will be almost exactly like the last problem... "TSA body scanners"
2765 2013-03-13 15:49:48 rdponticelli has quit (Ping timeout: 276 seconds)
2766 2013-03-13 15:50:01 glitch004 has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
2767 2013-03-13 15:50:13 pjr1972 is now known as pjr
2768 2013-03-13 15:50:14 egecko has joined
2769 2013-03-13 15:50:32 <gmaxwell> lianj: Nothing is planned right now along those lines, of course ones will eventually be neededâ for crypto upgrades even if nothing else. I assume any that are successful will be broadly uncontroversial and acceptable to virtually everyone.
2770 2013-03-13 15:50:55 <gmaxwell> helo: needing to know about forks is a _general_ concern. Not just a one time issue.
2771 2013-03-13 15:51:13 bitcoolman2 has joined
2772 2013-03-13 15:51:35 gntbynynybt has joined
2773 2013-03-13 15:51:39 pjr is now known as _pjr
2774 2013-03-13 15:51:44 <helo> nods
2775 2013-03-13 15:51:55 <gmaxwell> helo: most byzantine faults result in forks that you could learn about and change your behavior... though some cases of attacks result in forks that can't be detected until they're too late.
2776 2013-03-13 15:51:55 <egecko> add some kind of source blockchain identifier to the block and make fork detection easier and managable
2777 2013-03-13 15:52:08 <gmaxwell> egecko: 08:25 <+gmaxwell> egecko: uhhh. they _do_. 0_o.
2778 2013-03-13 15:52:19 <gmaxwell> egecko: every block uniquely identifies the chain that its a part of.
2779 2013-03-13 15:52:45 <gmaxwell> thats so fundimental to bitcoin ... uh. Blonde day? :P
2780 2013-03-13 15:53:18 i2pRelay has quit (Remote host closed the connection)
2781 2013-03-13 15:53:31 <egecko> then wtf is the problem detecting a fork?
2782 2013-03-13 15:53:43 <somequestions> gmaxwell: is this accurate regarding doublespends
2783 2013-03-13 15:53:50 i2pRelay has joined
2784 2013-03-13 15:53:52 <somequestions> oh snap, can't paste that much
2785 2013-03-13 15:54:10 <somequestions> please take a look at https://docs.google.com/document/d/1H2eKJIfkbhRqp936zDXW3i9kW43Re7cS3sNMDDTo3oA/edit?usp=sharing the section entitled "Schrodingers Bitcoin"
2786 2013-03-13 15:54:27 _pjr is now known as pjr
2787 2013-03-13 15:54:30 <egecko> if everything to determine block X came from chain A and block X2 came form B then why cant we get some fork detection put in and then it really wouldnt matter if a hardfork happened inadvertantly or as a matter of choice.
2788 2013-03-13 15:54:51 pjr is now known as _pjr
2789 2013-03-13 15:55:08 HANTI is now known as hanti
2790 2013-03-13 15:55:21 defunctzombie_zz is now known as defunctzombie
2791 2013-03-13 15:55:25 MrSnoopy has joined
2792 2013-03-13 15:55:37 <skinnkavaj> gmaxwell: i like competition its for us customers. i just wanna ask about the other coins out there. do they have any functions that cannot be implemented in bitcoin?
2793 2013-03-13 15:55:42 Namworld has joined
2794 2013-03-13 15:55:46 <MrSnoopy> help ! i have upgrade to bitcoind 0.8 and now my balance is now 0.0 !!
2795 2013-03-13 15:56:08 <sipa> MrSnoopy: is it reindexing?
2796 2013-03-13 15:56:10 <somequestions> mrsnoopy: are you sync'd?
2797 2013-03-13 15:56:10 <helo> egecko: i suspect there is a reasonable chance that a particular node won't see both chains
2798 2013-03-13 15:56:37 <MrSnoopy> how can i know it's reindexing ?
2799 2013-03-13 15:56:47 <MrSnoopy> its running in process list
2800 2013-03-13 15:57:01 <MrSnoopy> and my accounts are visible so wallet should be ok
2801 2013-03-13 15:57:14 <CodeShark> check the debug.log file
2802 2013-03-13 15:57:16 voodoo has quit (Read error: Connection reset by peer)
2803 2013-03-13 15:57:24 <MrSnoopy> "version" : 80000, "protocolversion" : 70001, "walletversion" : 60000, "balance" : 0.00000000,
2804 2013-03-13 15:57:27 <MrSnoopy> is this ok ?
2805 2013-03-13 15:57:28 <Dyaheon> so 0.7 basically ignored the 0.8 fork block completely? what if it stored & quarantined it (or rather it's height at least) provided the difficulty is right so it couldn't be spam?
2806 2013-03-13 15:57:54 <gmaxwell> skinnkavaj: Kinda OT in here, other coins could have interesting functionality we can't offer in Bitcoin, but I don't believe that any do today. Most are very minor modifications and are otherwise the same as bitcoin.
2807 2013-03-13 15:58:00 <CodeShark> Dyaheon - it specifically rejected a block which means the height was meaningless to 0.7
2808 2013-03-13 15:58:01 <helo> MrSnoopy: you are safe, just be patient. you did back your wallet up first though, right?
2809 2013-03-13 15:58:06 <egecko> helo: sure, but in that case it has to assume that whatever chains it does know about are the only ones that exist, it isn't until it detects a new chain that is a fork that it needs to possibly take any action
2810 2013-03-13 15:58:11 <MrSnoopy> nope
2811 2013-03-13 15:58:20 <helo> MrSnoopy: you only had one job! :P
2812 2013-03-13 15:58:26 <Dyaheon> CodeShark: yeah, but I meant for fork detecting purposes
2813 2013-03-13 15:58:42 <Dyaheon> could be useful to know that such a fork exists
2814 2013-03-13 15:58:43 <gmaxwell> egecko: nodes only forward blocks on their best chain, if not for that (or some other protection) it would be easy to DOS attack.
2815 2013-03-13 15:58:48 <helo> MrSnoopy: you're still almost 100% surely safe... but back up regularly, and particularly before changing anything!
2816 2013-03-13 15:58:50 <MrSnoopy> i just changed bitcoind, not the wallet
2817 2013-03-13 15:59:01 <MrSnoopy> ok
2818 2013-03-13 15:59:08 <MrSnoopy> so i need to wait then ?
2819 2013-03-13 15:59:18 _pjr is now known as _pjr_
2820 2013-03-13 15:59:24 tonikt has quit (Ping timeout: 258 seconds)
2821 2013-03-13 15:59:31 <gmaxwell> Dyaheon: pre-0.8 nodes would detect the fork, but only if they had some 0.8 peers that had it.
2822 2013-03-13 15:59:41 <gmaxwell> Dyaheon: ones that did displayed a warning.
2823 2013-03-13 15:59:45 <lianj> gmaxwell: ok that makes sense. explains why .8 nodes only saw all new .7 blocks once it got longer
2824 2013-03-13 15:59:50 <Dyaheon> alright
2825 2013-03-13 15:59:52 <MrSnoopy> debug.log show some accepted transactions from june 2011
2826 2013-03-13 16:00:05 <MrSnoopy> is it reindexing ?
2827 2013-03-13 16:00:07 _pjr_ is now known as _pjr
2828 2013-03-13 16:00:10 <helo> yeah.
2829 2013-03-13 16:00:12 <egecko> gmaxwell: but when a new peer is detected, it could very easily attempt to figure out if it needs to do any action - not that it _does_ this right now, but it should is all my position is
2830 2013-03-13 16:00:23 <MrSnoopy> so when my transactions will be processed my balance should change then ?
2831 2013-03-13 16:00:32 <lianj> MrSnoopy: yes
2832 2013-03-13 16:00:37 <MrSnoopy> ok i understand
2833 2013-03-13 16:01:02 _pjr is now known as _pr
2834 2013-03-13 16:01:14 <egecko> anyway, code calls
2835 2013-03-13 16:01:19 <gmaxwell> egecko: it doesâ as mentioned pre-0.8 nodes displayed a warning while the 0.8 fork they were rejecting was longer. But only if they were connected to a 0.8 node offering the fork.
2836 2013-03-13 16:01:20 <MrSnoopy> will my new blk000 files now be 128MB instead of 2GB ?
2837 2013-03-13 16:01:20 i2pRelay has quit (Remote host closed the connection)
2838 2013-03-13 16:01:27 <MrSnoopy> and the old will remaing 2GB ?
2839 2013-03-13 16:01:37 <helo> MrSnoopy: you should be able to still see all of the transactions in your wallet with "listtransactions"
2840 2013-03-13 16:01:43 <helo> and they'll all have 0 confirms
2841 2013-03-13 16:01:51 i2pRelay has joined
2842 2013-03-13 16:02:11 <MrSnoopy> helo: correct
2843 2013-03-13 16:02:19 <MrSnoopy> transactions are logged in the wallet then
2844 2013-03-13 16:02:20 <MrSnoopy> ?
2845 2013-03-13 16:02:29 MiningBuddy has joined
2846 2013-03-13 16:02:42 autismnode has joined
2847 2013-03-13 16:02:43 <helo> yeah. transactions relevant in determining your balance are stored there permanently.
2848 2013-03-13 16:03:00 <MrSnoopy> cool
2849 2013-03-13 16:03:04 <CodeShark> even transactions no longer relevant to that are stored there permanently :p
2850 2013-03-13 16:03:15 <helo> yes heh
2851 2013-03-13 16:04:03 <MrSnoopy> is it just a log
2852 2013-03-13 16:04:10 <MrSnoopy> or sthg important in the wallet
2853 2013-03-13 16:04:32 saulimus has joined
2854 2013-03-13 16:05:20 <MrSnoopy> from what i understand only the privkeys are needed to get back your btc
2855 2013-03-13 16:05:22 <Dyaheon> maybe nodes could relay some fork heights then, providing difficulty was right of course? so you wouldn't need a direct connection
2856 2013-03-13 16:05:26 <MrSnoopy> so this should not be really needed
2857 2013-03-13 16:06:01 rdponticelli has joined
2858 2013-03-13 16:06:58 glitch004 has joined
2859 2013-03-13 16:07:09 <CodeShark> MrSnoopy: you can recover all your bitcoins from the blockchain just with the privkeys - but you don't want to have to scan the whole block chain each time you want to spend
2860 2013-03-13 16:07:20 <CodeShark> the wallet also allows you to quickly query history
2861 2013-03-13 16:07:30 <MrSnoopy> ok so its a log
2862 2013-03-13 16:07:35 thepok has joined
2863 2013-03-13 16:07:57 <MrSnoopy> can you fool the system by modifying this wallet to increase your btc and try to spend them ?
2864 2013-03-13 16:08:16 <CodeShark> you'd only fool your own node
2865 2013-03-13 16:08:23 <CodeShark> the rest of the network would not be fooled
2866 2013-03-13 16:08:28 <MrSnoopy> ok the others would not confirm
2867 2013-03-13 16:08:28 jarpiain has quit (Ping timeout: 245 seconds)
2868 2013-03-13 16:08:36 <deadweasel> if you play wiht your node too much, you'll grow hair on the palms of your hands
2869 2013-03-13 16:08:40 <CodeShark> others wouldn't even relay
2870 2013-03-13 16:08:46 <glitch004> Yeah you would need 51% to confirm it
2871 2013-03-13 16:09:15 jarpiain has joined
2872 2013-03-13 16:09:22 i2pRelay has quit (Remote host closed the connection)
2873 2013-03-13 16:09:31 <gmaxwell> glitch004: no, thats not how bitcoin works. :(
2874 2013-03-13 16:09:32 <MrSnoopy> fine
2875 2013-03-13 16:09:43 <MrSnoopy> hm ?
2876 2013-03-13 16:09:53 i2pRelay has joined
2877 2013-03-13 16:10:17 <gmaxwell> glitch004: Bitcoin is primarily a zero trust systemâ every full node checks everything for itselfâ except ordering, which uses a computational vote because you can't autonomously validate that.
2878 2013-03-13 16:10:41 <glitch004> Ahh I see.
2879 2013-03-13 16:10:56 <gmaxwell> glitch004: so if a miner produces a block that makes up fake bitcoin or steals from someoneâ everyone will just ignore it, it's just like that block doesn't exist.. even if "51%"â or 80% or whatever of miners are agreeing.
2880 2013-03-13 16:10:56 <sipa> "Bitcoin is consensus, not democracy."
2881 2013-03-13 16:11:18 <somequestions> consensus reality
2882 2013-03-13 16:11:31 <gmaxwell> If we could achieve autonomous validation of ordering we'd do that too. sadly physics doesn't seem to permit it. :(
2883 2013-03-13 16:11:54 <sipa> gmaxwell: can't we get congress to change that stupid physics law?
2884 2013-03-13 16:12:03 <CodeShark> damn relativity
2885 2013-03-13 16:12:09 <sipa> it'd make things so much easier
2886 2013-03-13 16:12:21 <imsaguy> can we get congress to pass any laws right now?
2887 2013-03-13 16:12:22 <imsaguy> :-x
2888 2013-03-13 16:12:27 <imsaguy> (wrong channel)
2889 2013-03-13 16:14:06 <MrSnoopy> everybody from usa it seems here
2890 2013-03-13 16:14:20 <glitch004> So I guess my question is how do the other nodes know its a fake block? Hashes are wrong somehow?
2891 2013-03-13 16:14:25 FredEE has joined
2892 2013-03-13 16:14:34 <sipa> MrSnoopy: not at all, it's just easier to talk from a USA-centered perspective sometimes
2893 2013-03-13 16:15:37 <gmaxwell> glitch004: they apply the rulesâ adding up the transactions, checking that each spends only what it's permitted to, with the permissions being controlled by digital signatures.
2894 2013-03-13 16:15:52 rkosten has quit (Quit: ChatZilla 0.9.90 [Firefox 19.0.2/20130307122903])
2895 2013-03-13 16:16:02 <glitch004> Ahh ok that makes perfect sense
2896 2013-03-13 16:16:05 <CodeShark> and they also check proof-of-work
2897 2013-03-13 16:16:47 <CodeShark> which is perhaps what glitch004 was alluding too
2898 2013-03-13 16:16:49 <CodeShark> *to
2899 2013-03-13 16:16:51 bukaj has quit (Quit: Page closed)
2900 2013-03-13 16:17:03 tonikt has joined
2901 2013-03-13 16:17:14 aa has joined
2902 2013-03-13 16:17:23 i2pRelay has quit (Remote host closed the connection)
2903 2013-03-13 16:17:45 tyn has joined
2904 2013-03-13 16:17:53 <gmaxwell> CodeShark: yea, though POW is not needed for the autonomous validation. His 'fake block' was one letting you "fool the system by modifying this wallet to increase your btc and try to spend them"
2905 2013-03-13 16:17:54 i2pRelay has joined
2906 2013-03-13 16:18:07 rdymac has quit (Quit: This computer has gone to sleep)
2907 2013-03-13 16:18:19 <MrSnoopy> can i remove blk00*.dat now that they are located in blocks/blk0*.dat ?
2908 2013-03-13 16:18:52 <gmaxwell> MrSnoopy: yes, you can remove the top level ones if you won't downgrade to 0.7 again. Though they don't really take up any spaceâ they're hardlinks.
2909 2013-03-13 16:19:01 <MrSnoopy> ah ok
2910 2013-03-13 16:19:16 flyingkiwiguy has quit (Ping timeout: 252 seconds)
2911 2013-03-13 16:19:32 PiZZaMaN2K is now known as PiZZaMaN2K|away
2912 2013-03-13 16:19:44 bitcoin-dev000 has quit (Ping timeout: 245 seconds)
2913 2013-03-13 16:20:09 <MrSnoopy> can i remove blkindex.dat too ?
2914 2013-03-13 16:20:10 hydrogenesis has joined
2915 2013-03-13 16:20:28 <gmaxwell> MrSnoopy: yes, and that indeed will save you space.
2916 2013-03-13 16:20:34 <MrSnoopy> great
2917 2013-03-13 16:20:42 <MrSnoopy> coz im short of space
2918 2013-03-13 16:20:56 rdymac has joined
2919 2013-03-13 16:21:06 flyingkiwiguy has joined
2920 2013-03-13 16:21:21 <MrSnoopy> and addr.dat should be removed too
2921 2013-03-13 16:21:41 FredEE has quit (Quit: FredEE)
2922 2013-03-13 16:21:42 <MrSnoopy> but not peers.dat
2923 2013-03-13 16:21:57 <gmaxwell> MrSnoopy: yea, we have used addr.dat for a long time.
2924 2013-03-13 16:22:17 <MrSnoopy> looks clean now
2925 2013-03-13 16:22:31 <gavinandresen> contrib/tidy_datadir.sh will tidy things up foryou
2926 2013-03-13 16:24:15 ciphermonk has quit (Ping timeout: 276 seconds)
2927 2013-03-13 16:24:33 kadoban has joined
2928 2013-03-13 16:24:53 bitcoolman2 has quit (Quit: Page closed)
2929 2013-03-13 16:24:53 <MrSnoopy> thanks but too late i did it by hand
2930 2013-03-13 16:25:14 chrisb has quit (Ping timeout: 252 seconds)
2931 2013-03-13 16:25:24 i2pRelay has quit (Remote host closed the connection)
2932 2013-03-13 16:25:37 andytoshi has joined
2933 2013-03-13 16:25:51 parus has quit (Read error: Operation timed out)
2934 2013-03-13 16:25:52 i2pRelay has joined
2935 2013-03-13 16:25:56 aa has quit (Quit: Page closed)
2936 2013-03-13 16:26:51 rdponticelli has quit (Ping timeout: 276 seconds)
2937 2013-03-13 16:26:53 Cache_Money has quit (Ping timeout: 252 seconds)
2938 2013-03-13 16:27:04 gwollon is now known as gwillen
2939 2013-03-13 16:27:12 parus has joined
2940 2013-03-13 16:27:57 zoinky has joined
2941 2013-03-13 16:28:35 <somequestions> gmaxwell: Can I paraphrase our earlier conversation, or would you like to put together a more concise quote addressing the question "Will there be a next time, and is this situation repeatable"? I'd like to quote you.
2942 2013-03-13 16:29:10 EasyAt has joined
2943 2013-03-13 16:29:27 cap2002 has joined
2944 2013-03-13 16:29:48 <gmaxwell> somequestions: You can do whatever you like not attributing to me, if you'd like to attribute to me I'd appreciate it if I can see the paraphrase first. :)
2945 2013-03-13 16:30:03 <somequestions> I'd like to attribute it to you
2946 2013-03-13 16:30:28 glitch004 has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
2947 2013-03-13 16:31:46 <gmaxwell> (Having been misquoted by reporters before, I know to at least be warry of 'quotes' :P but yea, I'm perfectly fine with that.)
2948 2013-03-13 16:32:06 blaap has left ()
2949 2013-03-13 16:32:23 zoinky has quit (Ping timeout: 252 seconds)
2950 2013-03-13 16:32:24 <somequestions> gmaxwell I'd be happy to give you edit rights and let you insert the quote yourself
2951 2013-03-13 16:32:45 <somequestions> It does no good to be wrong
2952 2013-03-13 16:33:27 i2pRelay has quit (Remote host closed the connection)
2953 2013-03-13 16:33:35 <gmaxwell> I'm always happy to have someone help me communicate better. Can I pick this up again after lunch (in an hour or two?)
2954 2013-03-13 16:33:36 joehoyle- has joined
2955 2013-03-13 16:33:39 <gavinandresen> I strongly disagree, being wrong frequently is good practice.
2956 2013-03-13 16:33:49 <MrSnoopy> among my peers some are running Satoshi:0.6.3
2957 2013-03-13 16:33:59 i2pRelay has joined
2958 2013-03-13 16:34:23 <somequestions> being wrong in an experimental sense is fine, i'm talking about a journalistic article talking to people not involved with bitcoin
2959 2013-03-13 16:34:39 forever-d has joined
2960 2013-03-13 16:34:42 <somequestions> If you'd like to take a look at the article in progress it's here https://docs.google.com/document/d/1H2eKJIfkbhRqp936zDXW3i9kW43Re7cS3sNMDDTo3oA/edit?usp=sharing
2961 2013-03-13 16:34:42 <gavinandresen> Oh, yeah, journalists should be perfect....
2962 2013-03-13 16:34:43 <MrSnoopy> will they eventually run into problems ?
2963 2013-03-13 16:34:53 perezd has joined
2964 2013-03-13 16:35:01 <MrSnoopy> when most will run 0.8
2965 2013-03-13 16:35:03 forever-d is now known as Guest16536
2966 2013-03-13 16:35:07 <gmaxwell> If a journalist is wrong its best if they are grossly wrong rather than subtly. :P
2967 2013-03-13 16:35:40 <gavinandresen> Actually, my beef is with hit-and-run journalists who won't correct mistakes if they're pointed out
2968 2013-03-13 16:35:42 TonyBit has quit ()
2969 2013-03-13 16:35:44 zoinky has joined
2970 2013-03-13 16:36:00 <gavinandresen> Being wrong is OK, if you admit when you were wrong and correct your mistake
2971 2013-03-13 16:36:05 _pr has quit (Quit: Page closed)
2972 2013-03-13 16:36:09 <somequestions> I'm looking for an attributable quote that the double spending question "Will there be a next time, is this situation repeatable?"
2973 2013-03-13 16:36:14 <gavinandresen> Software development is no different....
2974 2013-03-13 16:36:23 tyn has quit (Ping timeout: 245 seconds)
2975 2013-03-13 16:36:27 <somequestions> that *answers
2976 2013-03-13 16:36:52 Swappermall has quit (Remote host closed the connection)
2977 2013-03-13 16:36:58 joehoyle has quit (Ping timeout: 258 seconds)
2978 2013-03-13 16:37:02 <_dr> ``probably'' --murphy
2979 2013-03-13 16:37:02 <Vinnie_win> somequestions: If you need some headlines -> https://bitcointalk.org/index.php?topic=152040.0
2980 2013-03-13 16:37:03 <MrSnoopy> are you in trouble if you still run btc v 0.6 ?
2981 2013-03-13 16:37:11 <MC1984> is a fork repeatable?
2982 2013-03-13 16:37:17 <MC1984> its not the first time its happened
2983 2013-03-13 16:37:45 <CodeShark> and it won't be the last
2984 2013-03-13 16:37:57 <somequestions> I sold the article on the premise "how does crisis look in decentralized bitcoin vs. central bank operated currencies"
2985 2013-03-13 16:38:07 <kinlo> MC1984: this fork will not happen again
2986 2013-03-13 16:38:07 <CodeShark> hopefully each time we can learn to deal with it better so that it never becomes too disruptive
2987 2013-03-13 16:38:07 <somequestions> haven't even gotten to that part yet, not sure if I will in this article
2988 2013-03-13 16:38:11 thepok has quit (Ping timeout: 255 seconds)
2989 2013-03-13 16:38:13 <CodeShark> (or we learn to better preempt it)
2990 2013-03-13 16:38:13 <helo> somequestions: i think consensus is still that bitcoin is still an experiment, and has the potential to fail or succeed spectacularly in a number of ways
2991 2013-03-13 16:38:14 <kinlo> MrSnoopy: no you are not in trouble
2992 2013-03-13 16:38:30 <MC1984> this particular fork maybe not
2993 2013-03-13 16:38:36 gribble has quit (Remote host closed the connection)
2994 2013-03-13 16:38:41 CaptainBlaze has joined
2995 2013-03-13 16:38:42 <HM> lol experiment or not, it's a piece of software
2996 2013-03-13 16:38:44 <MC1984> but they happen all the time, and infrequently need manual fiing
2997 2013-03-13 16:38:49 <HM> software always fails spectacularly
2998 2013-03-13 16:39:11 <kinlo> MC1984: it's how bitcoin works...
2999 2013-03-13 16:39:15 <MrSnoopy> if you are a merchant and run 0.6 you might not be able to validate some transactions
3000 2013-03-13 16:39:15 <MrSnoopy> ?
3001 2013-03-13 16:39:18 Belkaar has quit (Ping timeout: 248 seconds)
3002 2013-03-13 16:39:20 <MC1984> somequestions you ould reference the RBS outage in the UK last year
3003 2013-03-13 16:39:25 <somequestions> it's a piece of software with a decent amount of money moving through it minute to minute, so while bugs are fine I think some reassurance would be appropriate that this is a very very very rare and quickly resolved issue should it occur again
3004 2013-03-13 16:39:27 <MC1984> and the natwest one here sometime last week
3005 2013-03-13 16:39:31 <kinlo> MrSnoopy: no, it's completly safe now
3006 2013-03-13 16:39:31 <somequestions> but i'm writing the article, not a member of the dev team
3007 2013-03-13 16:39:46 <somequestions> so I need someone else to tell me that's right
3008 2013-03-13 16:39:47 <MrSnoopy> but what when new large blocks will be generated ?
3009 2013-03-13 16:39:51 <MrSnoopy> like 2 d ago
3010 2013-03-13 16:40:01 <somequestions> MC1984 that's my intention
3011 2013-03-13 16:40:10 <kinlo> MrSnoopy: you should try to upgrade more often, but except during the window of the fork, which was more then 24h ago, there is no problem
3012 2013-03-13 16:40:17 <somequestions> the details of the outage just turned out to be more complicated and detailed than I understood
3013 2013-03-13 16:40:22 <MrSnoopy> ok
3014 2013-03-13 16:40:29 MWNinja has quit (Remote host closed the connection)
3015 2013-03-13 16:40:29 <kinlo> MrSnoopy: they won't be. They are not accepted anymore
3016 2013-03-13 16:40:31 <HM> there's only bitcoins moving through bitcoin. the market sets the value. if it's not assured it can devalue bitcoin
3017 2013-03-13 16:40:55 Belkaar has joined
3018 2013-03-13 16:40:58 <Luke-Jr> somequestions: are you going to post an updated version when you've made revisions based on the input here?
3019 2013-03-13 16:41:21 <somequestions> I've been making revisions all along, that document is live
3020 2013-03-13 16:41:28 <MrSnoopy> i really hope btc rise to 10.000 $ / btc
3021 2013-03-13 16:41:30 i2pRelay has quit (Remote host closed the connection)
3022 2013-03-13 16:41:31 <Luke-Jr> somequestions: I don't see most things fixed..
3023 2013-03-13 16:41:33 <kinlo> MrSnoopy: in order to exploit the current situation you need a lot of hashpower, and the top 5 pools are all aware of the issue and are using their hashpower to prevent any new forks, so it's safe
3024 2013-03-13 16:41:45 ZedsterX has quit (Read error: Connection reset by peer)
3025 2013-03-13 16:41:46 <somequestions> if there are corrections that need to be made please comment on the document or message me here
3026 2013-03-13 16:41:54 * jgarzik wakes up, and looks
3027 2013-03-13 16:41:57 ZedsterX has joined
3028 2013-03-13 16:42:01 i2pRelay has joined
3029 2013-03-13 16:42:09 <MrSnoopy> or maybe 100000$/btc
3030 2013-03-13 16:42:10 <sipa> it has awoken!
3031 2013-03-13 16:42:14 fposdkgs is now known as Transisto
3032 2013-03-13 16:42:26 <kinlo> oh noes, it's awake! :)
3033 2013-03-13 16:42:28 <jgarzik> somequestions: *chuckle* it was hardly a calamity
3034 2013-03-13 16:42:29 <MC1984> everyone act normal
3035 2013-03-13 16:42:43 <MrSnoopy> i want to become rich :)
3036 2013-03-13 16:42:48 <jgarzik> somequestions: sometimes the chain forks
3037 2013-03-13 16:43:05 rdponticelli has joined
3038 2013-03-13 16:43:06 <sipa> well, it *was* the largest fork in two years
3039 2013-03-13 16:43:18 <kinlo> but still only limited :)
3040 2013-03-13 16:43:36 <MC1984> is it important to point out that its not like bitcoin would have destroyed itself without the manual intervention
3041 2013-03-13 16:43:54 <MC1984> it would have got on fine actually, the decision was taken to avoid fucking over merchants and such
3042 2013-03-13 16:43:57 <jgarzik> *nod* But "calamity"?
3043 2013-03-13 16:43:58 bitcoinmike___ has joined
3044 2013-03-13 16:43:58 <jgarzik> hardly.
3045 2013-03-13 16:44:07 HM has quit (Read error: Connection reset by peer)
3046 2013-03-13 16:44:07 <somequestions> jgarzik - someone doublespent and was paid for $10,000 worth of bitcoins with 12+ confirmations on the transaction - That seems like a big deal to me if the whole point of the currency is to eliminate doublespends
3047 2013-03-13 16:44:19 <sipa> MC1984: if it had taken a day, i think the consequences would have been worse
3048 2013-03-13 16:44:49 <MC1984> it would have knocked bitcoin for sure
3049 2013-03-13 16:45:07 <MC1984> but most hashpower was on 0.8 already right
3050 2013-03-13 16:45:15 <MC1984> good excuse to upgrade
3051 2013-03-13 16:45:41 <jgarzik> The whole point of the currency is not to eliminate double-spends. People try to double spend all the time. c.f. SatoshiDICE
3052 2013-03-13 16:45:46 <donpdonp> it would have set bitcoin back a couple years, imho. its always been promoted as 'wait for X confirmations and you're safe', and then we'd have to say oh you're on the 0.7 chain, thats not safe
3053 2013-03-13 16:45:57 <jgarzik> Confirmations make double-spends increasingly more unlikely, statistically
3054 2013-03-13 16:46:19 Guest15459 is now known as debiantoruser
3055 2013-03-13 16:46:22 <jgarzik> If you are dumb enough to accept zero-conf transactions, double spends are something you must defend against
3056 2013-03-13 16:46:24 <sipa> jgarzik: i disagree; preventing double-spends (however not perfectly, but exponentially harder given confirmations at least) is the point of bitcoin
3057 2013-03-13 16:46:24 rdymac has quit (Quit: This computer has gone to sleep)
3058 2013-03-13 16:46:27 gribble has joined
3059 2013-03-13 16:46:29 rich__ has quit (Ping timeout: 264 seconds)
3060 2013-03-13 16:46:34 <donpdonp> though i agree people would have upgraded to 0.8 and the 0.8 chain would have continued normally
3061 2013-03-13 16:46:35 <sipa> jgarzik: without that requirement, we wouldn't need a p2p network at all
3062 2013-03-13 16:46:37 <somequestions> so doublespends are theoretically possible even in a non-forked situation?
3063 2013-03-13 16:46:40 <MC1984> im assuming the 0.7 chain would have become a pariah pretty quickly out of mutual rational self interest of everyone in the system
3064 2013-03-13 16:46:46 <MC1984> maybe thats wrong
3065 2013-03-13 16:47:05 <Luke-Jr> somequestions: not with 12 confirms
3066 2013-03-13 16:47:08 <somequestions> You said people try to do it all the time, do they ever succeed?
3067 2013-03-13 16:47:08 <MC1984> but i did see a lot of mutual rational self interest in the move back to 0.7
3068 2013-03-13 16:47:14 <jgarzik> somequestions: one single block chain on one computer will not contain a double spend. It is getting consensus for that snapshot
3069 2013-03-13 16:47:19 <Luke-Jr> somequestions: it's easy to double-spend with 0 confirms
3070 2013-03-13 16:47:24 <jgarzik> (or any other)
3071 2013-03-13 16:47:25 <MC1984> even from SD which offered to suspend itself to help out (!)
3072 2013-03-13 16:47:50 <jgarzik> somequestions: people have successfully double-spent against zero-confirmation businesses, yes
3073 2013-03-13 16:48:06 <gavinandresen> somequestions: double-spend risk is not a "yes or no" -- like most risks, it is "more risky" or "less risky", down to "So little risk I won't worry about it"
3074 2013-03-13 16:48:13 <somequestions> Right, but everyone agrees only stupid people accept with 0 confirms
3075 2013-03-13 16:48:18 <MrSnoopy> suppose i sell one day stuffs for btc. how many confs should i need ?
3076 2013-03-13 16:48:18 rdponticelli has quit (Ping timeout: 276 seconds)
3077 2013-03-13 16:48:25 <somequestions> what about the OKpay doublespend that had 12 confirms, is that possible outside a fork?
3078 2013-03-13 16:48:36 zer0def has quit (Ping timeout: 252 seconds)
3079 2013-03-13 16:48:36 <gwillen> somequestions: there are limited occasions where you can reasonably accept zero confirms.
3080 2013-03-13 16:48:37 <gavinandresen> somequestions: One of the reasons I still tell people "bitcoin is an experiment" is because everybody is still figuring out where the edge-case risks are.
3081 2013-03-13 16:48:37 <MC1984> not really
3082 2013-03-13 16:48:42 <MrSnoopy> looks like 12 is not enough ?
3083 2013-03-13 16:48:51 <jgarzik> yes, OKPAY was extraordinary
3084 2013-03-13 16:48:54 <gwillen> somequestions: suppose you're a movie theater, so whoever pays you is going to be within arresting distance for the next 2 hours? :-)
3085 2013-03-13 16:48:58 <jgarzik> thankfully, it was paid back
3086 2013-03-13 16:49:02 <jgarzik> not really malicious double-spend
3087 2013-03-13 16:49:13 <somequestions> gavin that's a little bit of a cop-out, you've got to admit
3088 2013-03-13 16:49:22 <gavinandresen> somequestions: it's the truth
3089 2013-03-13 16:49:29 <somequestions> it may be an experiment but alot of people are using it on the claim I've heard many times that doublespends are impossible in real life
3090 2013-03-13 16:49:31 i2pRelay has quit (Remote host closed the connection)
3091 2013-03-13 16:49:33 bahkls has joined
3092 2013-03-13 16:49:43 <MC1984> its not a cop out!
3093 2013-03-13 16:49:53 <jgarzik> double-spends are not impossible, just successively harder with each confirmation.
3094 2013-03-13 16:50:00 i2pRelay has joined
3095 2013-03-13 16:50:02 <gavinandresen> somequestions: "okey dokey" -- I can't control what people claim about bitcoin
3096 2013-03-13 16:50:16 <MC1984> its always been described as an experiment, others have chosen to found startups around it
3097 2013-03-13 16:50:32 <somequestions> *skeptical face* i've been paying attention for two years and while you devs may say its statistically highly impropable but not impossible, I think the common perception out there is this just does not ever happen
3098 2013-03-13 16:50:34 kazimir42 has quit (Remote host closed the connection)
3099 2013-03-13 16:50:35 <somequestions> and that's clearly not the case
3100 2013-03-13 16:50:41 <gavinandresen> The Internet was an experiment when it started, too⦠some would say it still is.
3101 2013-03-13 16:50:44 <CodeShark> there's always risk associated with new technologies
3102 2013-03-13 16:50:47 <somequestions> anyhow
3103 2013-03-13 16:50:48 <somequestions> moving on
3104 2013-03-13 16:50:54 tyn has joined
3105 2013-03-13 16:50:56 discrete has joined
3106 2013-03-13 16:50:58 <CodeShark> bitcoin's still in the wild west
3107 2013-03-13 16:51:05 <cap2002> Is v1.0 then no longer an experiment?
3108 2013-03-13 16:51:10 <MrSnoopy> :)
3109 2013-03-13 16:51:24 <gavinandresen> v1.0 will be "safe enough for my grandma to use"
3110 2013-03-13 16:51:55 <gavinandresen> (well, that's my criteria⦠I might be out-voted on what we call 1.0)
3111 2013-03-13 16:52:03 <CodeShark> grandma's more likely to get hurt falling down the stairs than using bitcoin
3112 2013-03-13 16:52:07 <cap2002> great - any idea on a release date, ca.?
3113 2013-03-13 16:52:22 <sipa> cap2002: somewhere between 2014 and 4012
3114 2013-03-13 16:52:29 <cap2002> haha
3115 2013-03-13 16:52:35 <donpdonp> it showed me the 51% attack is more achievable that i thought, as the fix was a "0.7 attack" on the 0.8 chain.
3116 2013-03-13 16:52:37 <MC1984> somequestions it doesnt happen when the system is working as intended, which it does the great majority of the time
3117 2013-03-13 16:52:40 <jgarzik> somequestions: OKPAY is _not_ out $10,000
3118 2013-03-13 16:52:45 <jgarzik> somequestions: factually untrue
3119 2013-03-13 16:52:45 <donpdonp> (though 51% is probably too low)
3120 2013-03-13 16:53:14 <jgarzik> somequestions: OKPAY situation resolved itself, everybody got paid back. Read that thread started by "macbook-air" all the way to the end.
3121 2013-03-13 16:53:25 <donpdonp> jgarzik: thats good to know.
3122 2013-03-13 16:53:28 <moartr4dez> donpdonp: it was't as much an attack as it was the 0.8 folks switching sides and joining hashing power with folks on the 0.7 side
3123 2013-03-13 16:53:42 <gavinandresen> donpdonp: ⦠but that was a "tail risk" event, because we just happened to be in a situation where the network was split already at close to 50/50 old/new.
3124 2013-03-13 16:53:47 <donpdonp> moartr4dez: potatoe, po-tah-toe :)
3125 2013-03-13 16:54:05 <jgarzik> somequestions: and this is definitely not the "first successful double-spend" That happened long ago.
3126 2013-03-13 16:54:15 <MrSnoopy> when people will run more btc nodes, will it make this fork issue less likely ?
3127 2013-03-13 16:54:18 <gavinandresen> Ideally in the future we'll have more diversity, so a bug in any one implementation affects a solid minority of merchants/miners/etc
3128 2013-03-13 16:54:21 <moartr4dez> but it does reveal an interesting case when we go through version changes/upgrades - there is much greater potential during the transition for the network to be split in half
3129 2013-03-13 16:54:38 <jgarzik> gavinandresen: you mean "does not impact"?
3130 2013-03-13 16:54:39 <Luke-Jr> donpdonp: there was no "0.7 attack"
3131 2013-03-13 16:55:22 <MC1984> would you say some of the 0.8 miners who mined blocks on the sidechain and chose not to switch back felt attacked?
3132 2013-03-13 16:55:27 <donpdonp> gavinandresen: so alternate implementations should be encouraged?
3133 2013-03-13 16:55:51 <gavinandresen> donpdonp: yes, I think alternate implementations should be encouraged. I think diversity is good for the network.
3134 2013-03-13 16:56:02 <jgarzik> donpdonp: gavinandresen has always been encouraging of alternate implementations
3135 2013-03-13 16:56:06 Nick___ has joined
3136 2013-03-13 16:56:06 <jgarzik> and kudos for him
3137 2013-03-13 16:56:06 <gavinandresen> .. diversity will mean more, smaller failures rather than fewer, larger failures
3138 2013-03-13 16:56:10 <jgarzik> +1
3139 2013-03-13 16:56:16 <donpdonp> great!
3140 2013-03-13 16:56:17 <Nick___> #join #bitcoin-bots
3141 2013-03-13 16:56:17 <moartr4dez> indeed
3142 2013-03-13 16:56:23 <CodeShark> it would be a good idea for miners to be encouraged to cross-validate blocks with other implementations prior to mining, though
3143 2013-03-13 16:56:27 <helo> as long by diversity you mean "is exactly the same" ;)
3144 2013-03-13 16:56:36 <jgarzik> CodeShark: ...and within ourselves
3145 2013-03-13 16:56:38 <moartr4dez> there are chain forks quite routinely - they just don't normally last that long
3146 2013-03-13 16:56:47 <jgarzik> CodeShark: i.e. run multiple versions of our own software
3147 2013-03-13 16:57:03 <moartr4dez> the more diverse the network the less likely a major split. It makes sense
3148 2013-03-13 16:57:16 <jgarzik> heterogenous > homogenous
3149 2013-03-13 16:57:23 firstworld has joined
3150 2013-03-13 16:57:25 <helo> and the more impossible any future changes will be to enact
3151 2013-03-13 16:57:33 i2pRelay has quit (Remote host closed the connection)
3152 2013-03-13 16:57:51 <moartr4dez> if the majority of hashing power accepts the change, what's the problem?
3153 2013-03-13 16:58:04 i2pRelay has joined
3154 2013-03-13 16:58:13 <moartr4dez> tyranny of the majority... but oh well!
3155 2013-03-13 16:58:15 <donpdonp> the client is so difficult to get right, and with the 'standard' client being freely available, there isnt much encouragement for alternate implementations.
3156 2013-03-13 16:58:15 <helo> a bunch of vendors that get doublespent
3157 2013-03-13 16:58:39 <sipa> moartr4dez: a majority of hash power cannot arbitrarily change any rule
3158 2013-03-13 16:58:55 <CodeShark> jgarzik: agreed.
3159 2013-03-13 16:59:04 HM has joined
3160 2013-03-13 16:59:28 <jgarzik> moarrr: a majority of hash power can accept/deny sets of transactions, but cannot change the rules
3161 2013-03-13 16:59:31 <moartr4dez> sipa: why not? If they all agree on the change - won't they be able to build the block chain faster than others?
3162 2013-03-13 16:59:41 <jgarzik> moarrr: all clients validate their work
3163 2013-03-13 16:59:42 <jgarzik> er
3164 2013-03-13 16:59:48 <sipa> moartr4dez: yes, but others will simply ignore their blocks
3165 2013-03-13 16:59:49 <jgarzik> moartr4dez: all clients validate their work
3166 2013-03-13 16:59:57 <moartr4dez> right, which creates a fork
3167 2013-03-13 17:00:04 <jgarzik> moartr4dez: a miner-only fork
3168 2013-03-13 17:00:11 <jgarzik> moartr4dez: which clients then ignore
3169 2013-03-13 17:00:13 <sipa> moartr4dez: so they'll start building their own fork, effectively removing their haspower from the view of everyone else
3170 2013-03-13 17:00:13 cap2002 has quit (Quit: cap2002)
3171 2013-03-13 17:00:17 <gavinandresen> somequestions: â¦. so, RE: can "it" ever happen again: We'll work hard to try to make sure it doesn't. I believe that as bitcoin grows and gets more diverse this type of chain-split problem will be much less likely, but I've been around long enough to never say "never"
3172 2013-03-13 17:00:36 <jgarzik> moartr4dez: Any such miner simply knocks themselves off the network that way
3173 2013-03-13 17:00:45 <CodeShark> a supermajority of miners can arbitrarily choose not to accept blocks but they cannot choose to create blocks that regular nodes won't accept
3174 2013-03-13 17:00:52 <sipa> moartr4dez: or again: Bitcoin is consensus, not democracy.
3175 2013-03-13 17:00:52 <moartr4dez> sipa: well true - but then what if the non-miners also went along with the new rules?
3176 2013-03-13 17:00:54 <jgarzik> correct
3177 2013-03-13 17:01:06 <jgarzik> moartr4dez: then that's a hard fork
3178 2013-03-13 17:01:10 <moartr4dez> then the new-rule chain would become the standard
3179 2013-03-13 17:01:18 <sipa> moartr4dez: if _everyone_ agrees on a change, then they just start using software that follows those rules, duh
3180 2013-03-13 17:01:33 <sipa> moartr4dez: but any change that is not accepted by someone else, means just creating a fork
3181 2013-03-13 17:01:35 <donpdonp> another interesting effect that was critical to the recovery (from what ive heard people say) is when the miners noticed the switch in chain forks, they resubmitted the orphaned transactions from the old chain - brilliant!
3182 2013-03-13 17:01:46 ThomasV_ has joined
3183 2013-03-13 17:02:00 <moartr4dez> sipa: well yes, I agree - there can always be stragglers going off on their own side of the fork
3184 2013-03-13 17:02:08 <MC1984> that was automatic somehow right?
3185 2013-03-13 17:02:13 ThomasV has quit (Remote host closed the connection)
3186 2013-03-13 17:02:21 ThomasV_ has left ()
3187 2013-03-13 17:02:25 rich__ has joined
3188 2013-03-13 17:02:33 prolxray has joined
3189 2013-03-13 17:02:41 ThomasV has joined
3190 2013-03-13 17:02:54 <somequestions> jgarzik: I'm not claiming the money was kept, only that a doublespend successfully happened. I did read that thread all the way through and it was resolved because both participants were known community members who did the right thing, which seems silly to count on in future instances.
3191 2013-03-13 17:03:03 <CodeShark> donpdonp: most of the transactions on the network had been confirmed in both chains already
3192 2013-03-13 17:03:29 <donpdonp> CodeShark: oh i see. the distribution of unconfirmed transactions is something i have yet to look into in detail.
3193 2013-03-13 17:03:43 <helo> you can count on some people thinking about how to maliciously double spend if this happens again
3194 2013-03-13 17:03:53 tyn has quit (Ping timeout: 264 seconds)
3195 2013-03-13 17:04:26 <moartr4dez> I guess it's more a matter of how resilient does the code need to be to take into account potentially "long-lived" forks.
3196 2013-03-13 17:04:50 <moartr4dez> right now it seems not so resilient, since there was a panic/rush to re-merge and orphan one of the sides of the fork.
3197 2013-03-13 17:05:07 <MC1984> i suprised only that one guy did a double spend tbh
3198 2013-03-13 17:05:07 <CodeShark> if you have a bunch of confirmations for a transaction in two chains, either way the network goes it will confirm...but until then you won't really be able to spend it
3199 2013-03-13 17:05:15 <MC1984> and that it seemed to be opportunistic
3200 2013-03-13 17:05:18 itarkus has joined
3201 2013-03-13 17:05:18 itarkus has quit (Remote host closed the connection)
3202 2013-03-13 17:05:36 i2pRelay has quit (Remote host closed the connection)
3203 2013-03-13 17:05:47 hydrogenesis has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
3204 2013-03-13 17:05:48 <MC1984> saying that, engineering a long chain fork to do a double spend seems outside the realm of the doable
3205 2013-03-13 17:06:09 i2pRelay has joined
3206 2013-03-13 17:06:58 hydrogenesis has joined
3207 2013-03-13 17:07:03 <donpdonp> can the average bitcoin client detect the presence of multiple chains? might be a trigger for extra caution
3208 2013-03-13 17:07:15 <donpdonp> s/chains/forks/
3209 2013-03-13 17:07:16 <MrSnoopy> right
3210 2013-03-13 17:07:22 i2pRelay has quit (Remote host closed the connection)
3211 2013-03-13 17:07:37 <moartr4dez> donpdonp: I think the client just holds both sides of the fork - until the fork is resolved/merged.
3212 2013-03-13 17:07:45 <MrSnoopy> i guess it depends on your peers
3213 2013-03-13 17:07:46 <sipa> donpdonp: it detects when an invalid chain is longer
3214 2013-03-13 17:08:17 <jgarzik> somequestions: No argument, it was a successful double-spend (though not the first)
3215 2013-03-13 17:08:22 <moartr4dez> I shouldn't say "merged" I guess... you're right; one side of the fork is just orphaned (the "invalid" side)
3216 2013-03-13 17:08:30 FredEE has joined
3217 2013-03-13 17:08:31 <moartr4dez> invalid = shorter
3218 2013-03-13 17:08:47 Ant0 has quit (Ping timeout: 258 seconds)
3219 2013-03-13 17:08:58 <MrSnoopy> honestly i m surprised the btc xchange rate recovered that fast
3220 2013-03-13 17:09:13 <MrSnoopy> i thought it would have crashed flat
3221 2013-03-13 17:09:19 <moartr4dez> MrSnoopy - I'm not... talking to a couple bitcoin user's the next day - they didn't even notice it had happened.
3222 2013-03-13 17:09:29 <jgarzik> indeed
3223 2013-03-13 17:09:32 <EasyAt> When a fork finally outgrows the other, the TX's in the orphan that are not in the longest chain are automatically queued to the front, right?
3224 2013-03-13 17:09:38 <moartr4dez> and really - that price swing was more confined to Mt. Gox which froze withdrawals for a while I think
3225 2013-03-13 17:09:42 <MrSnoopy> yeah but now its in on news sites so ppl know it
3226 2013-03-13 17:09:48 <moartr4dez> I didn't see nearly as big a swing on other exchanges
3227 2013-03-13 17:10:03 <donpdonp> EasyAt: i dont think there is any queueing to the 'front' as they're distributed in a shotgun style, but yeah they get put back in the pool
3228 2013-03-13 17:10:04 HM has quit (Read error: Connection reset by peer)
3229 2013-03-13 17:10:16 <MrSnoopy> i would have guessed ppl would have panicked and sold
3230 2013-03-13 17:10:22 ToryJujube_ has joined
3231 2013-03-13 17:10:35 <moartr4dez> just goes to show how "strong" the demand is.
3232 2013-03-13 17:10:47 <Luke-Jr> meh, I don't regret selling anyhow
3233 2013-03-13 17:10:57 <Luke-Jr> I was holding too much bitcoins
3234 2013-03-13 17:11:05 <MrSnoopy> er how much ? :)
3235 2013-03-13 17:11:07 <moartr4dez> strong hands can ride out volatility... folks that borrowed on credit card to buy BC on the other hand = weak.
3236 2013-03-13 17:11:09 <CodeShark> I regret having been busy with something else at the moment when it dipped below 40 :p
3237 2013-03-13 17:11:10 ToryJujube_ has quit (Remote host closed the connection)
3238 2013-03-13 17:11:20 <CodeShark> I'd been looking for an opportunity to buy back
3239 2013-03-13 17:11:21 <somequestions> jgarzik - any idea when the last one was? Also, are you guys OK with me talking about the fact that doublespends are going to happen? That seems like a potential public relations problem
3240 2013-03-13 17:11:38 draradech has joined
3241 2013-03-13 17:11:51 <sipa> somequestions: within one bitcoin blockchain there are no double spends possible, at all
3242 2013-03-13 17:12:07 ToryJujube has quit (Remote host closed the connection)
3243 2013-03-13 17:12:22 ovidiusoft has quit (Ping timeout: 248 seconds)
3244 2013-03-13 17:12:28 <donpdonp> sipa: that's a technicallity thats hard to explain to a general audience thats aware of the okpay event. :)
3245 2013-03-13 17:12:30 ToryJujube has joined
3246 2013-03-13 17:12:35 <Luke-Jr> sipa: 0-confirm double spends are
3247 2013-03-13 17:12:38 <moartr4dez> sipa: well sure they are possible - but not if you are doing due diligence (waiting for sufficient # of confirmations)
3248 2013-03-13 17:12:46 <jgarzik> somequestions: What sipa said. Double spends have occurred at zero confirmations, a couple at 1-2 confirmations, and way back in history, when the chain was much more weak (less hash power), some people played with double-spends with more confirmations.
3249 2013-03-13 17:12:48 <sipa> somequestions: the question is 1) whether the chain can get reverted (which is what happened here, and is exceedingly rare in normal circumstance for more than 1-2 blocks) and 2) whether transactions in the memory pool can get reverted
3250 2013-03-13 17:13:06 <moartr4dez> 1-confirmation double spends can only happen if a block gets orphaned I think, no?
3251 2013-03-13 17:13:10 <jgarzik> double-spends today beyond 1-2 blocks are incredibly rare
3252 2013-03-13 17:13:16 <jgarzik> and probably only happen during forks
3253 2013-03-13 17:13:18 <jgarzik> like this event
3254 2013-03-13 17:13:21 <CodeShark> *argh* at the bastardization of the term "orphan" :p
3255 2013-03-13 17:13:22 <sipa> if you accept transactions in the memory pool as valid (0 confirmations), then you are taking the risk of it getting reverted, yes
3256 2013-03-13 17:13:23 HM has joined
3257 2013-03-13 17:13:26 <moartr4dez> during a "routine" fork (two blocks being discovered almost simultaneously)
3258 2013-03-13 17:13:37 <donpdonp> moartr4dez: that too is the great double-cross of the event, confirmation count was no guarantee of validity
3259 2013-03-13 17:13:38 HM has quit (Read error: Connection reset by peer)
3260 2013-03-13 17:13:41 <MrSnoopy> i guess one day some insurance companies will offer product to cover for the risk of double spending. kind of insurance like any risk can be covered
3261 2013-03-13 17:13:49 Descry has quit (Ping timeout: 260 seconds)
3262 2013-03-13 17:13:51 <sipa> moartr4dez: "within one chain" implies i'm not talking about the fact where we switch from one chain to another
3263 2013-03-13 17:13:59 i2pRelay has joined
3264 2013-03-13 17:14:04 <CodeShark> I suggest the use of the term "abandoned"
3265 2013-03-13 17:14:10 saulimus has quit (Quit: saulimus)
3266 2013-03-13 17:14:19 <sipa> CodeShark: abandoned/stale/extinct/... all better terms than orphan
3267 2013-03-13 17:14:24 <MrSnoopy> if you cannot reduce a risk to 0 and are not willing to accept it then transfer it through insurance
3268 2013-03-13 17:14:27 <CodeShark> absolutely, sipa :)
3269 2013-03-13 17:14:28 <moartr4dez> CodeShark: I like "abandoned" - I was using the term "orphan" for blocks because that's what blockchain.info calls them
3270 2013-03-13 17:14:31 <moartr4dez> :-)
3271 2013-03-13 17:14:34 <somequestions> So besides the doublespend, were *any* payments on either blockchain not processed properly, or reverted? I'm talking with Luke about this and he's saying the only transaction that was reverted is the doublespend on one fork?
3272 2013-03-13 17:15:04 <moartr4dez> somebody should tell blockchain.info that ^
3273 2013-03-13 17:15:12 <CodeShark> I've told him but he never changed it
3274 2013-03-13 17:15:18 <moartr4dez> l4me
3275 2013-03-13 17:15:44 <moartr4dez> https://blockchain.info/block-height/225637
3276 2013-03-13 17:15:58 <moartr4dez> 225637 (Orphaned)
3277 2013-03-13 17:16:02 <Luke-Jr> sipa: Eligius stats are using "stale" now :P
3278 2013-03-13 17:16:15 hydrogenesis has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
3279 2013-03-13 17:16:15 <sipa> "orphan" has become common terminology, because abandoned chains result in coinbase transactions that the reference client called "orphan"
3280 2013-03-13 17:16:17 <donpdonp> somequestions: "reverted" seems like the wrong term.
3281 2013-03-13 17:16:17 <Luke-Jr> moartr4dez: blockchain.info is full of misleading "info"
3282 2013-03-13 17:16:18 <moartr4dez> "Invalid Fork"
3283 2013-03-13 17:16:34 <moartr4dez> Luke-Jr: Agreed :-)
3284 2013-03-13 17:16:42 <sipa> but orphan mean "without parent", and internally in the software, there is something else called orphan blocks, namely blocks whose parent isn't known
3285 2013-03-13 17:17:04 <donpdonp> the abandoned fork is 'without children'
3286 2013-03-13 17:17:50 HM2 has joined
3287 2013-03-13 17:17:54 t7 has quit (Quit: Leaving)
3288 2013-03-13 17:17:59 da2ce7_d has joined
3289 2013-03-13 17:18:02 <sipa> donpdonp: not necessarily
3290 2013-03-13 17:18:24 zoinky has quit (Read error: Connection reset by peer)
3291 2013-03-13 17:18:34 <somequestions> sipa used the word "reverted" in the same context I intend it, is that not correct?
3292 2013-03-13 17:18:52 <donpdonp> sipa: i see. in either case 'abandoned' > 'orphaned'
3293 2013-03-13 17:19:23 MrSnoopy has quit (Quit: Page closed)
3294 2013-03-13 17:19:45 <sipa> somequestions: imho your question makes sense, but in discussions about the meaning of words, you'll likely get disagreement in any case
3295 2013-03-13 17:19:50 <CodeShark> sipa: the term "orphan" is also used within the satoshi source and log file to describe transactions that don't yet connect
3296 2013-03-13 17:19:54 <donpdonp> somequestions: i suppose its alright. to me reverted means a new transaction that counteracts the previous one, which is not the case.
3297 2013-03-13 17:20:18 da2ce7 has quit (Ping timeout: 256 seconds)
3298 2013-03-13 17:20:20 <Luke-Jr> donpdonp: a new block DOES counteract this one in question
3299 2013-03-13 17:20:24 <CodeShark> typically, the "orphan" status changes to "accepted" status once it does connect
3300 2013-03-13 17:20:29 <sipa> somequestions: there was a mail on the mailinglist about that, with statistics about reverted transactions
3301 2013-03-13 17:20:42 autismnode has quit (Excess Flood)
3302 2013-03-13 17:20:51 tyn has joined
3303 2013-03-13 17:21:00 <CodeShark> "orphan" just means "we're going to hold this thing in memory for a bit to see if we can place it somewhere..."
3304 2013-03-13 17:21:01 autismnode has joined
3305 2013-03-13 17:21:36 <CodeShark> "...once we learn of its dependencies."
3306 2013-03-13 17:21:37 HM2 has quit (Read error: Connection reset by peer)
3307 2013-03-13 17:21:57 <donpdonp> Luke-Jr: assuming a balance existed pre-fork, and a payment was made to okpay in 0.8chain, then the 0.7fork takes over and has its own payment to okpay. thats a different spend but not an un-doing of the 0.8chain transaction. for example the 0.7chain transaction could have been for a different amount.
3308 2013-03-13 17:22:20 <donpdonp> s/un-doing/negation/
3309 2013-03-13 17:22:43 <somequestions> sipa: Can you link me? I'm not on the mailing list
3310 2013-03-13 17:22:46 <Luke-Jr> â¦
3311 2013-03-13 17:23:54 HM2 has joined
3312 2013-03-13 17:26:46 SchmalzTech has joined
3313 2013-03-13 17:28:28 tyn has quit (Ping timeout: 245 seconds)
3314 2013-03-13 17:28:29 ic3 has joined
3315 2013-03-13 17:28:36 JDuke128 has quit (Quit: [BB])
3316 2013-03-13 17:29:06 jdnavarro has quit (Ping timeout: 258 seconds)
3317 2013-03-13 17:30:07 thoughtcourier has quit (Remote host closed the connection)
3318 2013-03-13 17:31:00 CodeShark has quit (Remote host closed the connection)
3319 2013-03-13 17:31:18 zoinky has joined
3320 2013-03-13 17:31:30 Ant0 has joined
3321 2013-03-13 17:32:32 <gavinandresen> somequestions: http://sourceforge.net/mailarchive/message.php?msg_id=30591726
3322 2013-03-13 17:32:59 pete2 has quit (Quit: bye!)
3323 2013-03-13 17:33:15 rdymac has joined
3324 2013-03-13 17:35:04 Transisto has quit (Read error: Connection reset by peer)
3325 2013-03-13 17:37:32 Transisto has joined
3326 2013-03-13 17:40:35 gntbynynybt has quit (Ping timeout: 258 seconds)
3327 2013-03-13 17:40:40 _pr has joined
3328 2013-03-13 17:41:04 josephin has quit (Quit: leaving)
3329 2013-03-13 17:41:04 <gavinandresen> BlueMatt: ping
3330 2013-03-13 17:41:20 _pr has quit (Client Quit)
3331 2013-03-13 17:41:24 <MC-Droid> 2 confirm double spend
3332 2013-03-13 17:41:26 Joric has joined
3333 2013-03-13 17:41:33 <MC-Droid> thats the finney attack right
3334 2013-03-13 17:41:42 <TD> no
3335 2013-03-13 17:41:58 <TD> finney attack means "mine a block on the best chain, broadcast a double spend, once the merchant gives you what you want, broadcast your new best block"
3336 2013-03-13 17:42:06 CaptainBlaze has quit (Quit: CaptainBlaze)
3337 2013-03-13 17:42:11 <MC-Droid> i know
3338 2013-03-13 17:42:12 <TD> if you mine a side-chain and broadcast it causing a re-org, that's just a Plain Old Double Spend
3339 2013-03-13 17:42:13 cheako has quit (Quit: Client exiting)
3340 2013-03-13 17:42:23 <TD> so you can't finney attack an already confirmed transaction
3341 2013-03-13 17:42:35 HM2 has quit (Read error: Connection reset by peer)
3342 2013-03-13 17:43:09 Transisto has quit (Read error: Connection reset by peer)
3343 2013-03-13 17:43:20 HM2 has joined
3344 2013-03-13 17:43:29 gntbynynybt has joined
3345 2013-03-13 17:43:31 <MC-Droid> how do you reverse 2 confirms then
3346 2013-03-13 17:43:36 <MC-Droid> brute force?
3347 2013-03-13 17:43:40 moleccc has quit (Ping timeout: 258 seconds)
3348 2013-03-13 17:43:41 _pr has joined
3349 2013-03-13 17:43:55 pobri19` has quit (Quit: Leaving)
3350 2013-03-13 17:44:30 _pr has quit (Client Quit)
3351 2013-03-13 17:45:37 Transisto has joined
3352 2013-03-13 17:45:59 somequestions has quit (Ping timeout: 245 seconds)
3353 2013-03-13 17:46:02 h4ckm3 has quit (Quit: Leaving)
3354 2013-03-13 17:46:11 <jgarzik> MC-Droid: with sufficient hashing power
3355 2013-03-13 17:46:22 _pr has joined
3356 2013-03-13 17:46:31 _pr has quit (Client Quit)
3357 2013-03-13 17:47:12 <HM2> the double spend that happened during the fork
3358 2013-03-13 17:47:37 user has joined
3359 2013-03-13 17:47:38 <HM2> i maintain that nodes need to shut themselves down if forks get too long
3360 2013-03-13 17:48:21 int03h_ has joined
3361 2013-03-13 17:48:45 <jgarzik> HM2: yes
3362 2013-03-13 17:48:51 <jgarzik> HM2: that is why alerts went out immediately
3363 2013-03-13 17:49:16 <jgarzik> HM2: but ideally, merchants would have sufficient infrastructure to notice chain forks, in a sufficiently mature payment ecosystem
3364 2013-03-13 17:49:28 <jgarzik> so alerts would be redundant in that case
3365 2013-03-13 17:49:37 <MC-Droid> is there any reliable way for a node to independently detect a chain fork
3366 2013-03-13 17:50:01 <MC-Droid> not a network fork obv
3367 2013-03-13 17:50:04 <jgarzik> MC-Droid: run multiple versions at multiple network sites
3368 2013-03-13 17:50:09 chicyanww has joined
3369 2013-03-13 17:50:28 <jgarzik> MC-Droid: parallel chains are by definition impossible _on a single node_
3370 2013-03-13 17:50:43 <HM2> surely any node that detects 2 'valid' and growing long chains should assume a hard fork and shut down transaction processing automatically
3371 2013-03-13 17:50:47 <jgarzik> Though single nodes do notice and collect orphan blocks.
3372 2013-03-13 17:51:04 protus has quit (Quit: protus)
3373 2013-03-13 17:51:07 <jgarzik> HM2: any one node only has one notion of "valid"
3374 2013-03-13 17:51:13 <MC-Droid> but the .7 nodes were still receiving and rejecting a string of blocks right
3375 2013-03-13 17:51:18 nivoc has quit (Quit: nivoc)
3376 2013-03-13 17:51:23 Lepton has quit (Ping timeout: 245 seconds)
3377 2013-03-13 17:51:24 ThiagoCMC has quit (Quit: Leaving)
3378 2013-03-13 17:51:40 o8gf5 has joined
3379 2013-03-13 17:51:44 <jgarzik> .7 nodes were storing as orphans all-but-the-bad-block in the 0.8 fork
3380 2013-03-13 17:51:49 <comboy> jgarzik: all orphan blocks are collected by every node?
3381 2013-03-13 17:51:58 <jgarzik> comboy: up to a limit yes
3382 2013-03-13 17:52:08 _pr has joined
3383 2013-03-13 17:52:31 <comboy> what is the limit? I mean if I would start breadcasting side chain starting at some very low difficulty, these blocks wouldn't be collected, right?
3384 2013-03-13 17:52:32 <jgarzik> orphan blocks, blocks on weak side chains, and blocks on the main chain (trunk) are all received
3385 2013-03-13 17:52:38 <MC-Droid> yes so when lots of consecutive orphans come in cant a node be all like, wait a minute
3386 2013-03-13 17:52:40 <jgarzik> side chain block != orphan block
3387 2013-03-13 17:53:07 <jgarzik> MC-Droid: who is to say that's not simply an Attack Of The Orphans ;p
3388 2013-03-13 17:53:10 Transisto has quit (Read error: Connection reset by peer)
3389 2013-03-13 17:53:35 <MC-Droid> it could be, but after 3 or 4 it becomes unlikey right
3390 2013-03-13 17:53:43 zoinky has quit (Quit: Leaving.)
3391 2013-03-13 17:53:54 <MC-Droid> ask okpay, better safe than sorry?
3392 2013-03-13 17:54:20 user has quit (Quit: bye!)
3393 2013-03-13 17:54:36 <HM2> How did a fault in the database even create a fork
3394 2013-03-13 17:54:52 <HM2> it'd have been better if it just segfaulted and died
3395 2013-03-13 17:54:53 zooko has quit (Ping timeout: 264 seconds)
3396 2013-03-13 17:55:19 <MC-Droid> bdb didnt even have the good grace to do that
3397 2013-03-13 17:55:20 <HM2> or thrown an exception and died
3398 2013-03-13 17:55:25 <jgarzik> MC-Droid: no, an orphan chain could be anything. until connected, that's garbage data.
3399 2013-03-13 17:55:40 Transisto has joined
3400 2013-03-13 17:55:51 user has joined
3401 2013-03-13 17:55:51 <comboy> I mean if they are stored, isn't it possible to SPAM network with big orphan blocks at very low difficulty?
3402 2013-03-13 17:56:11 <MC-Droid> oh rifght so its a dos risk
3403 2013-03-13 17:56:23 user has quit (Client Quit)
3404 2013-03-13 17:56:45 <comboy> s/SPAM/flood I guess, but mostly I mean wasted storage on nodes
3405 2013-03-13 17:56:54 <HM2> there needs to be code to take action automatically if there's another hard fork, even if it means the entire network grinds to a halt
3406 2013-03-13 17:57:21 OPrime has quit (Quit: OPrime)
3407 2013-03-13 17:57:51 <comboy> HM2: but wouldn't that mean that % much smaller than 50 of total network hashing power would be enough to effectively ddos bitcoin?
3408 2013-03-13 17:58:04 <HM2> i don't know
3409 2013-03-13 17:58:07 chicyanww has quit (Quit: Page closed)
3410 2013-03-13 17:58:10 int03h_ has left ()
3411 2013-03-13 17:58:29 <HM2> i'm just saying there must be a heuristic that can determine if a fork is beyond the realm of realistic possibility and at least throw a big fat warning
3412 2013-03-13 17:58:57 <HM2> you know the difficulty of each block
3413 2013-03-13 17:59:21 user has joined
3414 2013-03-13 17:59:38 user has quit (Client Quit)
3415 2013-03-13 18:00:36 pete78 has joined
3416 2013-03-13 18:00:36 <MC-Droid> yeah as long as nodes are receiving something unusual for an extended time there must be a way to tell
3417 2013-03-13 18:00:36 whooper has quit (Ping timeout: 245 seconds)
3418 2013-03-13 18:01:02 pete78 has quit (Client Quit)
3419 2013-03-13 18:01:42 _pr2 has joined
3420 2013-03-13 18:01:58 _pr2 has quit (Client Quit)
3421 2013-03-13 18:02:02 <MC-Droid> blocks with a non consecutive height?
3422 2013-03-13 18:02:48 <HM2> another idea is to have a watchdog process that can watch several nodes at once
3423 2013-03-13 18:02:54 <HM2> possibly running different versions
3424 2013-03-13 18:03:02 <MC-Droid> sigh im sure this is babbys first node security from me
3425 2013-03-13 18:03:11 Transisto has quit (Read error: Connection reset by peer)
3426 2013-03-13 18:03:48 pete78 has joined
3427 2013-03-13 18:04:10 joehoyle has joined
3428 2013-03-13 18:04:34 <HM2> meh
3429 2013-03-13 18:04:45 toffoo has joined
3430 2013-03-13 18:05:04 <HM2> I suppose the typical economic response is for payment processors to offer gaurantees and insurances for merchants
3431 2013-03-13 18:05:27 <HM2> just like VISA you get chargeback cover, with bitcoin you can get double spend cover
3432 2013-03-13 18:05:34 dust-otc has quit (Remote host closed the connection)
3433 2013-03-13 18:05:40 Transisto has joined
3434 2013-03-13 18:06:09 <HM2> simpler and cheaper than writing robust software
3435 2013-03-13 18:06:32 <MC-Droid> and then visa and friends undercut thr bitcoin merchant flow and its over
3436 2013-03-13 18:06:33 <HM2> (not that i'm saying bitcoin isn't particularly robust)
3437 2013-03-13 18:06:47 zooko has joined
3438 2013-03-13 18:06:56 pete78 has quit (Client Quit)
3439 2013-03-13 18:07:02 joehoyle- has quit (Ping timeout: 258 seconds)
3440 2013-03-13 18:07:03 _pr is now known as pete78
3441 2013-03-13 18:07:10 o8gf5 has quit (Ping timeout: 264 seconds)
3442 2013-03-13 18:07:28 <MC-Droid> they could probably drop thier fees under 1pc they just have no reason to
3443 2013-03-13 18:08:00 _pr has joined
3444 2013-03-13 18:08:40 _pr has quit (Client Quit)
3445 2013-03-13 18:09:20 _pr has joined
3446 2013-03-13 18:09:38 <prolxray> Sounds much better to have the network come to a grinding halt, than to have double spends going on, if that's the choice given.
3447 2013-03-13 18:09:44 KornKage has quit (Ping timeout: 252 seconds)
3448 2013-03-13 18:09:48 juanperez has joined
3449 2013-03-13 18:09:51 pete78 has quit (Quit: Page closed)
3450 2013-03-13 18:10:44 <MC-Droid> of course it is
3451 2013-03-13 18:10:58 b00tkitz has quit (Remote host closed the connection)
3452 2013-03-13 18:12:14 bitafterbit has joined
3453 2013-03-13 18:12:15 b00tkitz has joined
3454 2013-03-13 18:12:23 joehoyle has quit (Quit: Linkinus - http://linkinus.com)
3455 2013-03-13 18:12:41 cande has joined
3456 2013-03-13 18:12:42 glitch003_ has joined
3457 2013-03-13 18:13:02 <cande> hi guys
3458 2013-03-13 18:13:10 nealmcb has quit (Ping timeout: 248 seconds)
3459 2013-03-13 18:13:11 Transisto has quit (Read error: Connection reset by peer)
3460 2013-03-13 18:13:13 pete78 has joined
3461 2013-03-13 18:13:15 <cande> well done with fixing the fork
3462 2013-03-13 18:13:52 <cande> i have a qeustion regarding receiving several transactions to the same bitcoin adress
3463 2013-03-13 18:14:13 nealmcb has joined
3464 2013-03-13 18:14:36 <cande> how do i easy retreive all tx to a specific bitcoin address?
3465 2013-03-13 18:15:09 <MC-Droid> look it up on blockchain.info?
3466 2013-03-13 18:15:42 Transisto has joined
3467 2013-03-13 18:16:15 amantonop has joined
3468 2013-03-13 18:16:15 _pr has quit (Ping timeout: 258 seconds)
3469 2013-03-13 18:16:23 eculver has joined
3470 2013-03-13 18:16:35 <sipa> cande: getreceivedbyaddress RPC?
3471 2013-03-13 18:17:02 Joric has quit ()
3472 2013-03-13 18:18:37 zer0def has joined
3473 2013-03-13 18:19:08 cyphase has joined
3474 2013-03-13 18:21:09 blaap has joined
3475 2013-03-13 18:22:39 TD has quit (Quit: Leaving)
3476 2013-03-13 18:23:15 Transisto has quit (Read error: Connection reset by peer)
3477 2013-03-13 18:24:02 Anduck has joined
3478 2013-03-13 18:24:03 Anduck has quit (Changing host)
3479 2013-03-13 18:24:03 Anduck has joined
3480 2013-03-13 18:24:30 b00tkitz has quit (Ping timeout: 276 seconds)
3481 2013-03-13 18:25:36 Transisto has joined
3482 2013-03-13 18:26:14 b00tkitz has joined
3483 2013-03-13 18:26:40 KornKage has joined
3484 2013-03-13 18:26:52 brson has joined
3485 2013-03-13 18:27:56 adamblevine has joined
3486 2013-03-13 18:28:18 <adamblevine> <----somequestions from earlier, my IRC crashed while writing
3487 2013-03-13 18:28:22 <pete78> how does the blockchain checkpoint feature work?
3488 2013-03-13 18:28:28 Lepton has joined
3489 2013-03-13 18:28:29 Lepton has quit (Changing host)
3490 2013-03-13 18:28:29 Lepton has joined
3491 2013-03-13 18:28:59 <prolxray> pete78, you can learn lots by searching at http://bitcoin.stackexchange.com.
3492 2013-03-13 18:29:07 perezd has quit (Quit: perezd)
3493 2013-03-13 18:29:18 prolxray has left ()
3494 2013-03-13 18:29:20 <cande> sipa: it only shows the amount of bitcoin received, i want some juicy tx info
3495 2013-03-13 18:29:25 prolxray has joined
3496 2013-03-13 18:29:33 perezd has joined
3497 2013-03-13 18:29:55 <Ant0> umm can I delete everything from the bitcoin directory leaving only blocks and wallet.dat?
3498 2013-03-13 18:29:56 <cande> MC-Droid: i'm using bitcoind
3499 2013-03-13 18:30:21 <cande> Ant0: i would first copy wallet.dat, and then delete :-)
3500 2013-03-13 18:30:32 <Ant0> yeah I have several backups of the wallet
3501 2013-03-13 18:30:32 <adamblevine> jgarzik: you still around? have a few quotes I need to finish
3502 2013-03-13 18:31:01 <Ant0> but i wont be able to download everything again as im using a 3g connection
3503 2013-03-13 18:31:01 <adamblevine> I'm using Gavin's quote from earlier "double-spend risk is not a "yes or no" -- like most risks, it is "more risky" or "less risky", down to "So little risk I won't worry about it. One of the reasons I still tell people "bitcoin is an experiment" is because everybody is still figuring out where the edge-case risks are."
3504 2013-03-13 18:31:15 PlantMan has quit (Read error: Connection reset by peer)
3505 2013-03-13 18:31:34 <moartr4dez> anybody familiar with the work being done at John Hopkins?
3506 2013-03-13 18:31:39 <moartr4dez> http://www.newscientist.com/blogs/onepercent/2013/03/bitcoin-zerocoin.html
3507 2013-03-13 18:31:40 PlantMan has joined
3508 2013-03-13 18:31:41 jtimon has joined
3509 2013-03-13 18:31:47 <adamblevine> and have two followup questions "For the average user or merchant, how much risk is there in every-day use?" and then "Are there any steps, or particular thresholds of confirmations at which a user can be assured they're safe under all the but most extreme of circumstances?"
3510 2013-03-13 18:31:53 <moartr4dez> might be an interesting feature to incorporate in bitcoin/main if it turns out to be true/useful
3511 2013-03-13 18:32:00 <MC-Droid> i dont think bitcoin .8
3512 2013-03-13 18:32:00 <MC-Droid> onwards maintains a full txn index anymore?
3513 2013-03-13 18:32:14 bitcoinmike___ has quit (Ping timeout: 245 seconds)
3514 2013-03-13 18:32:29 <MC-Droid> or full spent output index
3515 2013-03-13 18:32:36 <adamblevine> gavinandresen: I'd take a response from you too
3516 2013-03-13 18:32:48 <MC-Droid> i think theres a way to enable it
3517 2013-03-13 18:32:53 <MC-Droid> is that relevant?
3518 2013-03-13 18:32:53 <adamblevine> gavinandresen: "For the average user or merchant, how much risk is there in every-day use?" and then "Are there any steps, or particular thresholds of confirmations at which a user can be assured they're safe under all the but most extreme of circumstances?"
3519 2013-03-13 18:33:10 Transisto has quit (Read error: Connection reset by peer)
3520 2013-03-13 18:33:32 lantastic has joined
3521 2013-03-13 18:33:40 <Luke-Jr> adamblevine: ask your question or get ignored
3522 2013-03-13 18:33:47 <adamblevine> I did, twice
3523 2013-03-13 18:33:57 <Luke-Jr> 2 quotes != question :P
3524 2013-03-13 18:34:03 <gavinandresen> adamblevine: average user or merchant delivering a physical product to customers: I'd say very close to zero risk after 3 confirmations.
3525 2013-03-13 18:34:03 <adamblevine> those are my quotes
3526 2013-03-13 18:34:18 <moartr4dez> so know one has heard of the supposed innovation from John Hopkins adding better privacy to Bitcoin?
3527 2013-03-13 18:34:19 <Luke-Jr> oh, I get it
3528 2013-03-13 18:34:21 <moartr4dez> no one
3529 2013-03-13 18:34:50 <Luke-Jr> thought you were quoting from the hardfork FAQ :P
3530 2013-03-13 18:34:54 <gavinandresen> adamblevine: exchanges or merchants that deliver "cash-like" high-value products or services have their own particular set of risks, and they need to be more careful.
3531 2013-03-13 18:35:31 molec has joined
3532 2013-03-13 18:35:45 Transisto has joined
3533 2013-03-13 18:36:27 <gavinandresen> adamblevine: merchants that deliver services that have a zero marginal cost to them (e.g. "Subscribe to my e-magazine") have even less risk than users or merchants delivering physical products.
3534 2013-03-13 18:37:01 pacpac has joined
3535 2013-03-13 18:37:28 <adamblevine> So is it safe to say the only risky situations are ones where tangible, irreversible value is sent before the transaction has been confirmed by say, 6 nodes? Is there a magic number of confirmations that makes you personally feel safe for any kind of transaction?
3536 2013-03-13 18:37:46 <gavinandresen> adamblevine: And all of that is modified if you have some trust in the person you're transacting with. I recently sold a couple thousand dollars worth of bitcoin to a friend who promised to send me a check in the mail, and I was happy to do that.
3537 2013-03-13 18:38:08 Maged has quit (Quit: ChatZilla 0.9.90 [Firefox 18.0.2/20130201065344])
3538 2013-03-13 18:38:59 <gavinandresen> adamblevine: it is not the number of nodes confirming, it is how long ago the transaction was included in the block chain. Longer == more trust that the transaction will not get reversed.
3539 2013-03-13 18:39:21 <adamblevine> So waiting two hours is better than waiting for 12 confirmations on an absolute level?
3540 2013-03-13 18:39:30 <adamblevine> (each confirmation assumed at 10m)
3541 2013-03-13 18:39:41 <gavinandresen> adamblevine: for ultimate trust transacting a huge amount of value to somebody who I think will try to rip me off-- 24 hours would be the "100% safe" time.
3542 2013-03-13 18:40:25 <gavinandresen> adamblevine: it really is confirmations and not time, but unless something wacky is happening those two are equal after 6 confirmations or so
3543 2013-03-13 18:41:21 <adamblevine> ok
3544 2013-03-13 18:41:32 <gavinandresen> adamblevine: â¦. and I pick 24 hours because that would be 120+ confirmations
3545 2013-03-13 18:41:33 licnep has quit (Ping timeout: 260 seconds)
3546 2013-03-13 18:42:28 <gavinandresen> I can't conceive of a situation where "we" (the shared consensus of the bitcoin network) would allow a 120-block fork to happen that double-spent.
3547 2013-03-13 18:43:08 <hmmmstrange> where do we get the wacky status from?
3548 2013-03-13 18:43:13 Transisto has quit (Read error: Connection reset by peer)
3549 2013-03-13 18:44:02 <adamblevine> How do you think the "we" has handled this crisis? Any comments on that?
3550 2013-03-13 18:44:04 <gavinandresen> hmmmstrange: if you don't get 100+ confirmations in the last 24 hours then something wacky is happening to you-- you're almost certainly disconnected from the real bitcoin network
3551 2013-03-13 18:44:38 protus has joined
3552 2013-03-13 18:44:47 gasteve has joined
3553 2013-03-13 18:45:18 <gavinandresen> adamblevine: I think "we" handled the crisis pretty well, but there are things that we could do better and will do better in the future.
3554 2013-03-13 18:45:43 Transisto has joined
3555 2013-03-13 18:46:14 <adamblevine> anything specific you want to share?
3556 2013-03-13 18:46:22 <adamblevine> things you'll do differently next time?
3557 2013-03-13 18:46:31 skeledrew has quit (Ping timeout: 258 seconds)
3558 2013-03-13 18:46:35 drizzt_ has quit (Quit: Konversation terminated!)
3559 2013-03-13 18:46:37 <adamblevine> any lessons learned/
3560 2013-03-13 18:47:40 <adamblevine> I know you're doing a postmortem, just hoping to get something positive to tie it all up with
3561 2013-03-13 18:48:15 copumpkin has quit (Ping timeout: 252 seconds)
3562 2013-03-13 18:48:20 <zveda> bitcoin 0.7 has a bug !
3563 2013-03-13 18:48:24 perezd has quit (Quit: perezd)
3564 2013-03-13 18:48:26 <gavinandresen> adamblevine: none I'm ready to share just yet⦠short term focus is on what to do today/tomorrow, I'd like to make the post-mortem public later
3565 2013-03-13 18:48:56 copumpkin has joined
3566 2013-03-13 18:49:00 <adamblevine> fair enough, I appreciate your time in responding
3567 2013-03-13 18:49:17 <gavinandresen> adamblevine: actually one really stupid/simple thing 'we' should have done long ago: make it easy for bitcoin services to get alerts sent to their email/SMS/etc
3568 2013-03-13 18:49:55 <gasteve> i'll wire mine up to a cowbell next to my bed
3569 2013-03-13 18:50:00 <gavinandresen> adamblevine: ⦠I'm kicking myself for not doing that a couple months ago when it occurred to me, but hi-ho
3570 2013-03-13 18:50:07 <zveda> who was the least cooperative mining pool ? :D
3571 2013-03-13 18:50:52 asa1024 has left ()
3572 2013-03-13 18:51:35 drizztbsd has joined
3573 2013-03-13 18:51:36 drizztbsd has quit (Changing host)
3574 2013-03-13 18:51:36 drizztbsd has joined
3575 2013-03-13 18:51:47 <gavinandresen> adamblevine: the post-mortem TODOs will concentrate on things like better communication, because no two crises are exactly alike and shutting the barn door after the pigs fly and all that....
3576 2013-03-13 18:52:14 BTCTrader2 has quit (Quit: BTCTrader2)
3577 2013-03-13 18:52:24 <zveda> is there any way to get in touch with botnet operator miners to get them to switch ?
3578 2013-03-13 18:52:34 ovidiusoft has joined
3579 2013-03-13 18:52:34 adamblevineagain has joined
3580 2013-03-13 18:52:42 <zveda> even tho they are evil they are still part of our community :D
3581 2013-03-13 18:52:43 <adamblevineagain> irc dropped me again, did the link to the working document go through?
3582 2013-03-13 18:52:47 alexkravets has joined
3583 2013-03-13 18:52:54 <gavinandresen> zveda: botnet miners will be going away with ASICs, but they should get alerts like everybody else....
3584 2013-03-13 18:53:03 <adamblevineagain> gavinandresen and sorry but if you responded to my followup question I did not see it
3585 2013-03-13 18:53:27 Transisto has quit (Read error: Connection reset by peer)
3586 2013-03-13 18:53:29 adamblevine has quit (Ping timeout: 245 seconds)
3587 2013-03-13 18:53:41 <Vinnie_win> Did I just hear that SMS/email/text message allerts are high priority?
3588 2013-03-13 18:53:54 <adamblevineagain> you said :actually one really stupid/simple thing 'we' should have done long ago: make it easy for bitcoin services to get alerts sent to their email/SMS/etc" and I asked "any specific plans to do something about it soon?" or something to that effect
3589 2013-03-13 18:53:57 <gavinandresen> adamblevineagain: the post-mortem TODOs will concentrate on things like better communication, because no two crises are exactly alike and shutting the barn door after the pigs fly and all that....
3590 2013-03-13 18:53:58 <gavinandresen> [2:30pm]
3591 2013-03-13 18:54:12 skeledrew has joined
3592 2013-03-13 18:54:35 Hashdog has quit (Ping timeout: 256 seconds)
3593 2013-03-13 18:54:35 <petertodd> re: alerts, it occured to me that regular transactions can be used for them. It'd be good for client-defined alerts; say Armory has a dangerous bug, but only armory. Works fine with the bloom filters too, just add an entry for your particular applications alert scriptPubKey form.
3594 2013-03-13 18:54:50 <adamblevineagain> Can you say that in a way that's more of a contained sentence? I can paraphrase but thats a little rough.
3595 2013-03-13 18:55:01 <adamblevineagain> happy to paraphrase, just don't want to misquote
3596 2013-03-13 18:55:32 <Vinnie_win> Am I hearing this right, using transactions in the block chain to send alerts and messages which are vendor-specific?
3597 2013-03-13 18:55:56 <petertodd> Vinnie_win: Exactly.
3598 2013-03-13 18:56:06 CaptainBlaze has joined
3599 2013-03-13 18:56:16 Transisto has joined
3600 2013-03-13 18:56:19 <Vinnie_win> Nice. Maybe we can better support developer salaries with some clickable banner ads in the block chain too?
3601 2013-03-13 18:56:21 <gavinandresen> adamblevineagain: mmmm⦠"We'll be concentrating on things like improving communication during crise s, because no two crises are exactly alike."
3602 2013-03-13 18:56:30 <adamblevineagain> Thank you
3603 2013-03-13 18:56:38 diatonic has joined
3604 2013-03-13 18:57:13 <petertodd> Vinnie_win: Lol, we have those, they're called donation addresses...
3605 2013-03-13 18:57:15 <gavinandresen> adamblevineagain: re: specific tasks: implementing a -alertnotify=<run some command> is very high on my TODO list
3606 2013-03-13 18:57:16 Duly842 has quit (Ping timeout: 258 seconds)
3607 2013-03-13 18:57:37 lodse has joined
3608 2013-03-13 18:57:58 <gavinandresen> adamblevineagain: (that's easy to do, and is the thing I'm kicking myself for not doing sooner)
3609 2013-03-13 18:58:19 <petertodd> gavinandresen: Also, -testalertkey=somekey and similar obvious testing stuff.
3610 2013-03-13 18:58:24 <gavinandresen> petertodd: RE: using transactions: good idea
3611 2013-03-13 18:58:32 <Vinnie_win> Instead of polluting the One True Client with extraneous features, can the Bitcoin Foundation contract with a third party company to provide alert subscription services for mining pool operators and users?
3612 2013-03-13 18:59:00 <prolxray> I don't see how that sort of functionality is "extraneous" at this point.
3613 2013-03-13 18:59:02 <adamblevineagain> So that means the ability to send commands remotely over the network?
3614 2013-03-13 18:59:22 <petertodd> gavinandresen: It's a legit OP_DROP use-case.
3615 2013-03-13 18:59:31 <adamblevineagain> or is it just communication in that a message will pop up on the end-users side?
3616 2013-03-13 18:59:33 <gavinandresen> adamblevineagain: no, that means "if you get an alert, run this command to tell me about it"
3617 2013-03-13 18:59:42 <BlueMatt> gavinandresen: pong
3618 2013-03-13 18:59:49 <gavinandresen> ⦠where "this command" is specified by the service operator -- send email, send sms, whatever
3619 2013-03-13 19:00:03 <adamblevineagain> Thanks for the clarification
3620 2013-03-13 19:00:09 <BlueMatt> gavinandresen: re: kill, the script is correct, in bash if command; will execute when command exits with 0
3621 2013-03-13 19:00:21 viperhr1 has joined
3622 2013-03-13 19:00:31 <gavinandresen> BlueMatt: right, but kill exits with 0 if it manages to kill something
3623 2013-03-13 19:00:50 <gavinandresen> BlueMatt: waitâ¦.
3624 2013-03-13 19:00:56 <gavinandresen> BlueMatt: too many negatives....
3625 2013-03-13 19:01:02 <BlueMatt> yes...
3626 2013-03-13 19:01:20 <gavinandresen> BlueMatt: ok, I see. bash if is backwards from every other if in the world....
3627 2013-03-13 19:01:44 viperhr has quit (Ping timeout: 255 seconds)
3628 2013-03-13 19:02:12 <adamblevineagain> gavinandresen When you say service operator, do you mean online wallets, normal wallets, or individual users, or miners? I'm not clear where this is set up in the infrastructure
3629 2013-03-13 19:02:16 <BlueMatt> gavinandresen: yep
3630 2013-03-13 19:02:29 <BlueMatt> gavinandresen: (the assumption is exit code 0 is always success, non-zero is failure)
3631 2013-03-13 19:02:46 Hashdog has joined
3632 2013-03-13 19:02:59 <Vinnie_win> Wouldn't it be better to use existing technologies for an out-of-band notification system instead of grafting it on top of a cryptocurrency ledger?
3633 2013-03-13 19:03:49 <Vinnie_win> What if someone isn't running a full node and just using a brain wallet, how will they get notifications?
3634 2013-03-13 19:03:58 benjam55 has joined
3635 2013-03-13 19:04:11 <prolxray> That's a good point.
3636 2013-03-13 19:04:23 <gavinandresen> BlueMatt: sigh, ok⦠I'll have to fire up a Linux Vm and figure out what's different (works ok on my Mac)
3637 2013-03-13 19:04:25 <benjam55> Hello:)
3638 2013-03-13 19:05:02 <gavinandresen> adamblevineagain: service operators means anybody running the bitcoind daemon-- so mining pools, individual miners, online wallets, merchants, exchangesâ¦.
3639 2013-03-13 19:05:14 <gavinandresen> adamblevineagain: ⦠pretty much everybody except individual users
3640 2013-03-13 19:05:34 fposdkgs has joined
3641 2013-03-13 19:05:48 <gavinandresen> adamblevineagain: individual users see the alerts on their screen, so they're already covered
3642 2013-03-13 19:06:24 adamblevineagain has quit (Ping timeout: 245 seconds)
3643 2013-03-13 19:06:30 <zveda> can the ability to send notifications ever fall into the wrong hands ?
3644 2013-03-13 19:06:44 <gavinandresen> Vinnie_win: go for it, you should start a service that runs bitcoind -alertnotify=⦠and routes the messages to a list of people who subscribe
3645 2013-03-13 19:06:46 Transisto has quit (Ping timeout: 245 seconds)
3646 2013-03-13 19:06:48 <Vinnie_win> So use the client to send critical notifications about bugs in the client. Wouldn't it be better to isolate the two systems?
3647 2013-03-13 19:06:50 bernard75 has joined
3648 2013-03-13 19:07:07 <BlueMatt> gavinandresen: heh, have fun
3649 2013-03-13 19:07:09 <Vinnie_win> zveda: The presumption is that the alerts are protected by some hard coded public keys I guess
3650 2013-03-13 19:07:35 <petertodd> Vinnie_win: Realisticly users aren't going to run a totally separate "alert client", although they could.
3651 2013-03-13 19:07:54 <gavinandresen> zveda: if the alert key is compromised then I send a special alert that says "the alert key is compromised" and we release new software with a new alert key.
3652 2013-03-13 19:08:12 <zveda> icic
3653 2013-03-13 19:08:26 D34TH has joined
3654 2013-03-13 19:08:27 D34TH has quit (Changing host)
3655 2013-03-13 19:08:27 D34TH has joined
3656 2013-03-13 19:08:31 <Vinnie_win> petertodd: There are a zillion companies that offer low cost "corporate sms/email" services to get alert messages out to hundreds, thousands, or tens of thousands of people when necessary. There's no need to reinvent the wheel or add to the hairball.
3657 2013-03-13 19:09:01 <Vinnie_win> petertodd: The advantage of outsourcing it is that it would be a separate, isolated system that consumes no developer resources
3658 2013-03-13 19:09:08 kadoban has quit (Quit: bye)
3659 2013-03-13 19:09:22 <Vinnie_win> petertodd: Plus, anyone could subscribe. They dont even need to have a wallet or run the client. The press, for example.
3660 2013-03-13 19:09:31 kadoban has joined
3661 2013-03-13 19:09:38 <gavinandresen> Vinnie_win: ⦠the disadvantage is paranoid people who don't want to be on a centralized list of "people who care about Bitcoin" won't use it
3662 2013-03-13 19:10:05 <petertodd> Vinnie_win: The point is, Bitcoin itself doesn't need any extra machinery to implement this. All I'll do is write up a BIP to make it a standard, make sure the idea is thought out carefully, and let client developers do what they may. Also, what gavin said. Additionally, for some clients, their Bitcoin connection might be the only data they have access to at all.
3663 2013-03-13 19:10:08 skinnkavaj has quit (Quit: - nbs-irc 2.39 - www.nbs-irc.net -)
3664 2013-03-13 19:11:18 <petertodd> Vinnie_win: The only ugly bit is it has to be done such that the scriptPubKey is fixed, which means basically you want to be able to get the full transaction for an unspent txout; not all future UTXO pruning systems will allow that.
3665 2013-03-13 19:11:53 joshsm has quit (Ping timeout: 260 seconds)
3666 2013-03-13 19:12:20 joshsm has joined
3667 2013-03-13 19:12:46 <Vinnie_win> petertodd: couldn't this be done using relay messages only
3668 2013-03-13 19:13:15 rbecker is now known as RBecker
3669 2013-03-13 19:13:46 <Vinnie_win> If using the block chain to notify gamblers of their losses using single-satoshi outputs is bad then isn't using the block chain to let every vendor send their pet notifications ("new version, less spyware, download now!") also bad?
3670 2013-03-13 19:14:05 blkashdla has joined
3671 2013-03-13 19:14:22 <petertodd> Vinnie_win: Anti-DoS is a hard problem, and Bitcoin nodes only relay messages with signatures from the given set of pubkeys.
3672 2013-03-13 19:15:07 <petertodd> Vinnie_win: Yeah, it's a trade-off, although the message would go in the prunable scriptSig, so it's only an ugly trade-off.
3673 2013-03-13 19:15:33 lodse has left ()
3674 2013-03-13 19:15:42 Transisto has joined
3675 2013-03-13 19:15:51 pete78 has quit (Ping timeout: 240 seconds)
3676 2013-03-13 19:15:51 fposdkgs has quit (Ping timeout: 276 seconds)
3677 2013-03-13 19:16:34 BTCTrader2 has joined
3678 2013-03-13 19:17:00 <Vinnie_win> If I were a government I would round all you guys up and lock you in a room, take over the location where the binary download is stored and put my own version there, and then use the messaging system to tell the Bitcoin world to upgrade
3679 2013-03-13 19:17:21 ville-_ has quit (Read error: Operation timed out)
3680 2013-03-13 19:17:25 _pr has joined
3681 2013-03-13 19:17:35 <petertodd> Vinnie_win: They don't need the messaging system to pull that stunt off.
3682 2013-03-13 19:17:55 fishfish has joined
3683 2013-03-13 19:18:02 <prolxray> Split the public keys across multiple trusted developers in different countries
3684 2013-03-13 19:18:16 <prolxray> To reduce 'government risk'
3685 2013-03-13 19:18:29 <petertodd> anyway, I'm off
3686 2013-03-13 19:18:34 eennaam has joined
3687 2013-03-13 19:18:49 mappum has joined
3688 2013-03-13 19:19:10 TD has joined
3689 2013-03-13 19:19:18 perezd has joined
3690 2013-03-13 19:19:43 joshsm has quit (Ping timeout: 245 seconds)
3691 2013-03-13 19:20:08 <oracleCoin> Hi
3692 2013-03-13 19:20:15 <oracleCoin> I am trying to compile bitcoin windows
3693 2013-03-13 19:20:20 <oracleCoin> on mingw client
3694 2013-03-13 19:20:22 <oracleCoin> i get this
3695 2013-03-13 19:20:24 <oracleCoin> bogj@CHZRHFPWKS500 ~/Downloads/db-4.8.30/build_unix $ make ./libtool --mode=compile cc -c -I. -I../dist/.. -O3 ../dist/../mutex/mut_win32.c libtool: compile: cc -c -I. -I../dist/.. -O3 ../dist/../mutex/mut_win32.c -DDLL_EXPORT -DPIC -o .libs/mut_win32.o In file included from ./db_int.h:886:0, from ../dist/../mutex/mut_win32.c:12: ../dist/../dbinc/repmgr.h:502:13: error: two or more data types in declara
3696 2013-03-13 19:20:32 <oracleCoin> wtf?
3697 2013-03-13 19:20:41 <Vinnie_win> Can't you just open the Visual Studio project and press "Build" ?
3698 2013-03-13 19:21:27 <oracleCoin> configure works , following instructions https://github.com/bitcoin/bitcoin/blob/master/doc/build-msw.txt
3699 2013-03-13 19:21:37 <oracleCoin> can someone give some pointers?
3700 2013-03-13 19:22:02 <oracleCoin> berkley db build is failing
3701 2013-03-13 19:22:33 <oracleCoin> ../dist/../dbinc/repmgr.h:502:13: error: two or more data types in declaration specifiers ../dist/../dbinc/repmgr.h:502:1: warning: useless type name in empty declaration [enabled by default] make: *** [mut_win32.lo] Error 1
3702 2013-03-13 19:22:47 <oracleCoin> wot?^^^^
3703 2013-03-13 19:22:57 <Vinnie_win> ugh...bitcoin uses that "configure" shit?
3704 2013-03-13 19:23:09 <sipa> no, he isn't building bitcoin
3705 2013-03-13 19:23:12 <sipa> he's building BDB
3706 2013-03-13 19:23:14 MC1984 has quit (Read error: Connection reset by peer)
3707 2013-03-13 19:23:23 MC1984 has joined
3708 2013-03-13 19:23:28 <Vinnie_win> he said he's trying to compile bitcoin windows
3709 2013-03-13 19:23:32 <oracleCoin> yeah but i need it right sipa?
3710 2013-03-13 19:23:38 <sipa> you do, yes
3711 2013-03-13 19:23:42 adamblevine has joined
3712 2013-03-13 19:23:44 <oracleCoin> just following the build instructions
3713 2013-03-13 19:24:17 <oracleCoin> sipa, did you build bdb successfull on windows?
3714 2013-03-13 19:24:27 <oracleCoin> using mingw?
3715 2013-03-13 19:24:28 <adamblevine> first draft of "What Just Happened: Financial Crisis in the World of Bitcoin" is ready and if anyone wants to do a final fact check before I take off from IRC, this is your last chance https://docs.google.com/document/d/1H2eKJIfkbhRqp936zDXW3i9kW43Re7cS3sNMDDTo3oA/edit?usp=sharing
3716 2013-03-13 19:24:32 <sipa> the last windows i used was somewhere last millenium
3717 2013-03-13 19:24:44 bahkls has quit (Ping timeout: 245 seconds)
3718 2013-03-13 19:24:56 Duly678 has joined
3719 2013-03-13 19:25:08 <sipa> adamblevine: i'll have a look
3720 2013-03-13 19:25:15 <adamblevine> if you find something that needs correcting, you can highlight and right click on the passage to leave a comment
3721 2013-03-13 19:25:49 fposdkgs has joined
3722 2013-03-13 19:25:50 joshsm has joined
3723 2013-03-13 19:26:20 Transisto has quit (Ping timeout: 246 seconds)
3724 2013-03-13 19:27:45 <oracleCoin> guys this is bitcoin-dev, take this to bitcoin [normal]
3725 2013-03-13 19:27:56 rdymac has quit (Quit: This computer has gone to sleep)
3726 2013-03-13 19:28:06 <oracleCoin> I have some coding issues here goddamned
3727 2013-03-13 19:28:12 <oracleCoin> ../dist/../dbinc/repmgr.h:502:13: error: two or more data types in declaration specifiers ../dist/../dbinc/repmgr.h:502:1: warning: useless type name in empty declaration [enabled by default] make: *** [mut_win32.lo] Error 1
3728 2013-03-13 19:28:33 <oracleCoin> ^^ cant compile the DBD part with minGw as described in the GIT readme
3729 2013-03-13 19:28:45 <oracleCoin> wasup with that?
3730 2013-03-13 19:29:26 PhantomSpark has joined
3731 2013-03-13 19:29:38 PhantomSpark has quit (Remote host closed the connection)
3732 2013-03-13 19:30:11 <Vinnie_win> " the last windows i used was somewhere last millenium" - Not surprising the Windows build is painful and prone to error.
3733 2013-03-13 19:30:14 <sipa> oracleCoin: building on windows has always been a paid, sorry for that
3734 2013-03-13 19:30:27 <sipa> as none of the core developers are into windows development, really
3735 2013-03-13 19:30:35 <sipa> *pain
3736 2013-03-13 19:30:50 <sipa> feel free to send a correction if you figure things out
3737 2013-03-13 19:30:56 <moartr4dez> I'm surprised that anything useful gets done on Windows - the developments tools from Microsoft are particularly awful
3738 2013-03-13 19:31:01 <Vinnie_win> I'm not into Mac or Linux development but my projects build flawlessly and I haven't even tested them but once or twice. You don't have to be "into windows development" you just need to make a well groomed project a priority.
3739 2013-03-13 19:31:49 adamblevine has quit (Ping timeout: 245 seconds)
3740 2013-03-13 19:32:06 ville-_ has joined
3741 2013-03-13 19:32:07 d4de has quit (Ping timeout: 260 seconds)
3742 2013-03-13 19:32:25 <alexkravets> +sipe: what you just said is the same as saying: "Supporting Chinese local has always been a pain in the ass, Chinese people should feel free to go off and do their own bitcoin" Do you not see how such attitude is counterproductive ?
3743 2013-03-13 19:33:43 <sipa> alexkravets: as i said, i'm sorry about it (btw, i have already spent hours fixing windows-only issues), but i'm a volunteer and so are most other developers
3744 2013-03-13 19:34:02 <K1773R> alexkravets: cross compiling works better and is more easy than (not even to talk about the gitian-builder) so why should building with "messy/deprecated" tools?
3745 2013-03-13 19:34:11 saulimus has joined
3746 2013-03-13 19:34:13 <sipa> yes, it would be better if building was easier, but it isn't now
3747 2013-03-13 19:34:33 <K1773R> argl, did edit my sentences wrong, there goes the grammar...
3748 2013-03-13 19:34:36 <gavinandresen> for the record, I would love to see somebody step up and maintain a Visual Studio project for Bitcoin-Qt/bitcoind
3749 2013-03-13 19:34:49 <gavinandresen> it just isn't high enough on my TODO list to do it myself
3750 2013-03-13 19:34:55 <Vinnie_win> eliminate all external deps
3751 2013-03-13 19:35:00 <gavinandresen> (and I only run windows in virtual machines)
3752 2013-03-13 19:35:30 <Vinnie_win> How was LevelDB integrated?
3753 2013-03-13 19:35:34 <gavinandresen> Vinnie_win: eliminating all external dependencies ain't gonna happen. You might want to help jgarzik with his 'picocoin' project if you're looking for minimal dependencies
3754 2013-03-13 19:35:42 Transisto has joined
3755 2013-03-13 19:35:59 cande has quit (Ping timeout: 245 seconds)
3756 2013-03-13 19:36:14 <Vinnie_win> Was LevelDB brought into the repo using git-subtree and added to the makefile for the executable or would someone have to separately "install" the LevelDB into some global shared folder and build it separately
3757 2013-03-13 19:36:24 nimdAHK has quit (Ping timeout: 245 seconds)
3758 2013-03-13 19:36:27 Ashaman has joined
3759 2013-03-13 19:36:36 <Vinnie_win> gavinandresen: Oh! Cool...looking at picocoin now
3760 2013-03-13 19:36:42 <gavinandresen> Vinnie_win: you'll make us grumpy if you keep asking questions that are easy to figure out with a little work
3761 2013-03-13 19:36:49 alexkravets has quit (Ping timeout: 245 seconds)
3762 2013-03-13 19:36:49 AMDchild__ has quit (Ping timeout: 245 seconds)
3763 2013-03-13 19:37:09 <Vinnie_win> wasn't expecting an answer but I will look now.
3764 2013-03-13 19:37:14 juanperez has quit (Ping timeout: 245 seconds)
3765 2013-03-13 19:37:14 Nick___ has quit (Ping timeout: 245 seconds)
3766 2013-03-13 19:37:14 AMDchild has quit (Ping timeout: 245 seconds)
3767 2013-03-13 19:37:14 coinUsr has quit (Ping timeout: 245 seconds)
3768 2013-03-13 19:37:21 <fiesh> ... then why did you ask?!
3769 2013-03-13 19:37:28 oleganza has joined
3770 2013-03-13 19:37:30 <gavinandresen> â¦. you'll make us even grumpier asking rhetorical questions....
3771 2013-03-13 19:37:39 yellowhat has quit (Ping timeout: 245 seconds)
3772 2013-03-13 19:37:39 fposdkgs has quit (Ping timeout: 264 seconds)
3773 2013-03-13 19:37:58 <gavinandresen> ⦠and we're probably already overly grumpy because of lack of sleep from the chainfork crisis....
3774 2013-03-13 19:38:08 alexkravets has joined
3775 2013-03-13 19:38:11 <Vinnie_win> hey look leveldb was added to the repo!
3776 2013-03-13 19:38:20 <Vinnie_win> my apologies
3777 2013-03-13 19:39:00 <sipa> gavinandresen: i'd _love_ to drop OpenSSL though :)
3778 2013-03-13 19:39:16 ville-_ has quit (Ping timeout: 245 seconds)
3779 2013-03-13 19:39:17 joshsm has quit (Ping timeout: 246 seconds)
3780 2013-03-13 19:39:21 prolxray has left ()
3781 2013-03-13 19:39:30 <Vinnie_win> I brought this up a year ago...doesn't cryptopp do most everything that is needed?
3782 2013-03-13 19:39:57 <Vinnie_win> If this thing wasn't such a pain in the ass to build I would run a node with a debugger attached 24/7
3783 2013-03-13 19:40:08 <sipa> cryptopp is too slow
3784 2013-03-13 19:40:16 Ashaman_ has joined
3785 2013-03-13 19:40:20 <sipa> but yes
3786 2013-03-13 19:40:22 <gavinandresen> sipa: dropping openssl from bitcoind is probably possible, but Qt likes OpenSSL
3787 2013-03-13 19:40:22 <Vinnie_win> wha....?
3788 2013-03-13 19:40:28 <Vinnie_win> too slow?
3789 2013-03-13 19:40:41 <sipa> yes, it's far slower than OpenSSL for ECDSA verifications
3790 2013-03-13 19:40:47 <gmaxwell> gavinandresen: I don't think dependencies for the GUI are the same as dependencies for chain validation.
3791 2013-03-13 19:40:52 ville-_ has joined
3792 2013-03-13 19:40:57 PlantMan has quit (Ping timeout: 258 seconds)
3793 2013-03-13 19:41:08 <gavinandresen> gmaxwell: sure sureâ¦
3794 2013-03-13 19:41:13 <gmaxwell> (though for compiling is compiling, but I don't care about that.)
3795 2013-03-13 19:41:25 <Vinnie_win> sipa: Did you tell the author to speed it up?
3796 2013-03-13 19:41:41 <sipa> Vinnie_win: no, i implemented it from scratch myself :p
3797 2013-03-13 19:41:51 Ashaman has quit (Ping timeout: 264 seconds)
3798 2013-03-13 19:42:01 <Vinnie_win> sipa: Ermm....what do you need openssl for then
3799 2013-03-13 19:42:12 <Vinnie_win> sipa: Or are you joking about re-implementing ECDSA verification
3800 2013-03-13 19:42:34 <sipa> i'm not joking, but i'm also not as stupid to think that i can do that in a bug-free way
3801 2013-03-13 19:43:36 <Vinnie_win> sipa: Okay I fired off an email to Wei Dai
3802 2013-03-13 19:43:39 <gavinandresen> bitcoind uses openssl to get cryptographic random numbers and to implement RPC-over-SSL and probably a few other things I'm forgetting
3803 2013-03-13 19:43:49 <sipa> gavinandresen: also, OpenSSL as a dependency is easier than ECDSA in OpenSSL, which several distributions remove
3804 2013-03-13 19:44:17 <Vinnie_win> Cryptographic random numbers is easy, you can just crib the yarrow sources from schneier
3805 2013-03-13 19:45:16 manacit is now known as t0astiest
3806 2013-03-13 19:45:25 joshsm has joined
3807 2013-03-13 19:45:26 t0astiest is now known as manacit
3808 2013-03-13 19:45:38 Habbie has quit (Ping timeout: 248 seconds)
3809 2013-03-13 19:45:51 fposdkgs has joined
3810 2013-03-13 19:45:55 maaku has joined
3811 2013-03-13 19:46:05 <Vinnie_win> RPC over SSL is not so straightforward
3812 2013-03-13 19:46:10 <maaku> is there a BIP for document timestamp outputs?
3813 2013-03-13 19:46:44 stalled has quit (Ping timeout: 255 seconds)
3814 2013-03-13 19:47:21 Transisto has quit (Ping timeout: 260 seconds)
3815 2013-03-13 19:48:18 <sipa> maaku: ?
3816 2013-03-13 19:48:24 Habbie has joined
3817 2013-03-13 19:48:40 thepok has joined
3818 2013-03-13 19:48:57 o8gf5 has joined
3819 2013-03-13 19:49:06 <maaku> i know people have discussed using the bitcoin block chain as a document timestamping service - has anyone formalized a common mechanism for doing so?
3820 2013-03-13 19:49:36 <sipa> iirc there are even two implementations
3821 2013-03-13 19:49:37 <maaku> e.g, OP_PUSH hash160(document) OP_DROP with nValue=0, etc.
3822 2013-03-13 19:50:15 <grau> maaku: P2SH format will be accepted by the network. However you burn the coins by that.
3823 2013-03-13 19:50:27 <sipa> please don't do that
3824 2013-03-13 19:50:43 <sipa> the best solution is using the coinbase to commit to data, like committing to merge-mined chains
3825 2013-03-13 19:50:51 <sipa> chronobit uses that, via p2pool, afaik
3826 2013-03-13 19:51:15 <maaku> this is an alt-chain application
3827 2013-03-13 19:51:15 <gavinandresen> ⦠but there appears to be approximately zero demand for a decentralized timestamping service right now.
3828 2013-03-13 19:51:42 <petertodd> maaku: try https://github.com/opentimestamps/opentimestamps-client
3829 2013-03-13 19:51:53 <maaku> we're looking at timestamping ripple commits as part of a colored coin proposal for freicoin
3830 2013-03-13 19:52:10 <maaku> petertodd: thank you
3831 2013-03-13 19:52:23 <petertodd> gavinandresen: someone gave me $150 to turn my timestamper back on, and they appear to be doing a timestamp every hour or so
3832 2013-03-13 19:52:49 amantonop has quit (Quit: amantonop)
3833 2013-03-13 19:52:55 <gavinandresen> petertodd: $150 is approximately zero when compared to developer's salary....
3834 2013-03-13 19:53:13 <grau> sipa: just curious. What is the reason for being picky about what transactions mean?
3835 2013-03-13 19:53:18 <gavinandresen> ⦠err, I mean, "Cool! good for you!"
3836 2013-03-13 19:53:44 zoinky has joined
3837 2013-03-13 19:53:55 <petertodd> gavinandresen: part of why I wrote opentimestamps was because I knew people would use the ugly way; that person said they were doing unspendable txouts
3838 2013-03-13 19:54:24 <sipa> grau: hmm, am i?
3839 2013-03-13 19:54:39 <grau> gavinandersen: If I my pay would ave been set in BTC month ago I would also ignorant of small sums :)
3840 2013-03-13 19:54:42 fishfish is now known as fishfish|AFK
3841 2013-03-13 19:55:21 fposdkgs has quit (Ping timeout: 240 seconds)
3842 2013-03-13 19:55:23 <sipa> grau: (sorry, i'm just not sure what you're referring to)
3843 2013-03-13 19:55:27 tyn has joined
3844 2013-03-13 19:55:43 <grau> sipa: you said do not do that for using P2SH to store hash
3845 2013-03-13 19:55:48 Transisto has joined
3846 2013-03-13 19:56:00 <sipa> grau: i'm absolutely against abusing the blockchain as a communication channel
3847 2013-03-13 19:56:14 <gmaxwell> grau: because that has poor scalablity for something that can be done with infinite scalablity.
3848 2013-03-13 19:56:23 sudofox has joined
3849 2013-03-13 19:56:27 <sipa> well, yes, it's always about pro/cons
3850 2013-03-13 19:56:33 <grau> sipa: I am against abusing too, but it is a communication channel by its purpose
3851 2013-03-13 19:56:44 <sipa> grau: it is not; it is a consensus mechanism
3852 2013-03-13 19:56:51 <sipa> that requires some minimal communication
3853 2013-03-13 19:56:55 <gmaxwell> Timestamping via coinbase merging adds no data per timestamped element.
3854 2013-03-13 19:56:56 <sudofox> Hello all
3855 2013-03-13 19:57:05 <sipa> but purely as a communication medium, it absolutely sucks: it's slow, expensive and unreliable
3856 2013-03-13 19:57:12 <grau> sipa: you also communicate what you consent on.
3857 2013-03-13 19:57:29 sgornick has quit (Ping timeout: 256 seconds)
3858 2013-03-13 19:57:30 <sipa> yes, not disagreeing with the fact that there is communication involved (duh!)
3859 2013-03-13 19:57:32 <grau> sipa: storing a hash is a rather concise store of a document
3860 2013-03-13 19:57:33 <gmaxwell> grau: thats why it can communicate at all. if we could remove the communication from it we would. :P
3861 2013-03-13 19:57:44 <sipa> grau: and storing none is even shorter :)
3862 2013-03-13 19:57:48 <gmaxwell> grau: it is infinitely worse than a possible alternative.
3863 2013-03-13 19:58:05 <gmaxwell> grau: if you only want a timestamp you can do that with _zero_ marginal bytes added.
3864 2013-03-13 19:58:09 <maaku> petertodd: how/where does opentimestamps interface with the blockchain?
3865 2013-03-13 19:58:14 <MC-Droid> someone tried that once i remember
3866 2013-03-13 19:58:14 <grau> ok, I do not prepare to do that just do not see it as a sin
3867 2013-03-13 19:58:45 <oracleCoin> sipa: thanks
3868 2013-03-13 19:58:51 <grau> comared to SD using a tx to tell you lost thats nothing
3869 2013-03-13 19:58:51 drizztbsd has quit (Remote host closed the connection)
3870 2013-03-13 19:58:57 <oracleCoin> I will look into a nativeVisual Studio Build
3871 2013-03-13 19:59:00 <oracleCoin> for the client
3872 2013-03-13 19:59:10 <petertodd> maaku: It relys on a (currently) central server that aggregates multiple time stamps, builds a merkle tree, and then timestamps just that digest in a spendable txout. (which gets spent)
3873 2013-03-13 19:59:12 <gmaxwell> grau: there must be some motivation to encourage people to not externalize costs which they could make zero. We can't tecnically prevent some things, but we can certantly socially discourage them.
3874 2013-03-13 19:59:41 <petertodd> maaku: It's not perfect, but for the client doing it the right way in coinbases is a transparent upgrade.
3875 2013-03-13 19:59:45 <grau> gmaxwell: you did that with me now. But it does not scale.
3876 2013-03-13 20:00:00 <maaku> i see; that's a different application, but some parts might be applicable
3877 2013-03-13 20:00:14 <gmaxwell> grau: If we do a good enough job at it then you will also carry on the discouragement, and so that scales. :)
3878 2013-03-13 20:00:18 <petertodd> maaku: What are you trying to do exactly?
3879 2013-03-13 20:00:26 <sipa> see it this way: anything you do with the block chain forces costs on all validation nodes (bandwidth, storage), which are not payed for that; sure, they usually run their client voluntarily, but they do so because they value the idea of decentralized monetary transfer that bitcoin provides; using it for another purpose may push them away
3880 2013-03-13 20:00:45 <Vinnie_win> Is it correct to say that although a transaction fee is paid exactly once, the total carrying cost of that transaction steadily increases with both the number of full nodes and the amount of time that the outputs remain unspent?
3881 2013-03-13 20:01:10 <sipa> grau: how do you mean, doesn't scale? you can commit to an arbitrary number of arbitrarily-sized pieces of data in O(1) cost to the chain
3882 2013-03-13 20:01:11 fishfish is now known as AFK!~fishfish2@cpc10-mort5-2-0-cust225.19-2.cable.virginmedia.com|fishfish
3883 2013-03-13 20:01:14 <gmaxwell> ^ which is also a reason that bitcoin users should build things like scalable timestamping, even though they may not want it themselves, just to prevent the less efficient usage.
3884 2013-03-13 20:01:24 fishfish has quit (Quit: Bye!)
3885 2013-03-13 20:01:27 <gmaxwell> sipa: he means that pressuring people to not build bad systems does not scale.
3886 2013-03-13 20:01:41 <sipa> well, maybe it doesn't, but then bitcoin is doomed :)
3887 2013-03-13 20:01:44 <grau> sipa: I meant it does not scale to socially discuorage misuse
3888 2013-03-13 20:01:54 <sipa> maybe not
3889 2013-03-13 20:02:35 <maaku> petertodd: composable transactions (SINGLE|ANYONECANPAY), that are combined together into an atomic transaction, allowing some outputs to be records of off-chain transactions
3890 2013-03-13 20:03:07 mow has quit (Remote host closed the connection)
3891 2013-03-13 20:03:45 <maaku> in other words, the SIGHASH_SINGLE matched with a document timestamp becomes "I send my coins so long as you include this document timestamp in the chain" - the document being some off-chain contract
3892 2013-03-13 20:04:03 <maaku> like a private, 2-phase ripple transaction
3893 2013-03-13 20:04:10 <grau> Talking about true costs of a transaction. Did someone estimate it? eg. Inflation + routing costs + storage until virtually eternity?
3894 2013-03-13 20:04:29 <sudofox> Hi, does anyone want or need Testnet BTC? I'm giving it out for free just for the feels.
3895 2013-03-13 20:04:40 <petertodd> maaku: opentimestamps will work for what you are doing. It's basically a generic way of describing hash proofs that happen to lead up to Bitcoin block headers, as well as a way of getting those proofs made.
3896 2013-03-13 20:04:42 <sudofox> Of course, it's not worth anything, so..
3897 2013-03-13 20:04:50 <Vinnie_win> grau: Is that in response to my question?
3898 2013-03-13 20:05:21 randy-waterhouse has joined
3899 2013-03-13 20:05:40 Luke-Jr has quit (Excess Flood)
3900 2013-03-13 20:05:42 <grau> Vinnie_win: no, I have not scrolled back... just doing
3901 2013-03-13 20:05:48 fposdkgs has joined
3902 2013-03-13 20:06:00 stalled has joined
3903 2013-03-13 20:06:04 Luke-Jr has joined
3904 2013-03-13 20:06:16 <helo> mailing list... can't... keep... up.
3905 2013-03-13 20:06:46 Transisto has quit (Ping timeout: 248 seconds)
3906 2013-03-13 20:06:53 zer0def has quit (Ping timeout: 252 seconds)
3907 2013-03-13 20:07:11 <grau> Vinnie_win: sorry cant locate the context.
3908 2013-03-13 20:07:17 <gmaxwell> grau: I suspect that limit is not very useful... :( the storage cost is infinite unless you assuming exponentially decreasing storage cost, which is questionable since physics seems to forbid storage from becoming cheaper at some point :P.
3909 2013-03-13 20:07:20 <Vinnie_win> grau: "Is it correct to say that although a transaction fee is paid exactly once, the total carrying cost of that transaction steadily increases with both the number of full nodes and the amount of time that the outputs remain unspent? "
3910 2013-03-13 20:07:31 <Vinnie_win> grau: It seems that eventually, fees will never cover the true cost of transactions
3911 2013-03-13 20:07:54 <gmaxwell> Vinnie_win: marginal storage costs appear to be going down over time.
3912 2013-03-13 20:08:38 <grau> Vinnie_win: I guess without a mechanism of pruning true costs are not converging.
3913 2013-03-13 20:08:41 <Vinnie_win> gmaxwell: But you agree that the carrying cost of a transaction is steadily increasing with the # of nodes and the time unspent right?
3914 2013-03-13 20:09:22 <gmaxwell> grau: even pruning doesn't fix that, since some outputs are unspendable... and the full history can't be completely forgotten.. though its a big improvement.
3915 2013-03-13 20:10:00 Diablo-D3 has quit (Quit: This computer has gone to sleep)
3916 2013-03-13 20:10:06 fposdkgs has quit (Ping timeout: 245 seconds)
3917 2013-03-13 20:10:15 <gmaxwell> Vinnie_win: well it's increasing while cost per byte decreases. If the first is faster than the other in the long run then bitcoin must fail (or become something else).
3918 2013-03-13 20:10:37 <Vinnie_win> gmaxwell: But also the number of nodes. If the number of nodes doubles, then the cost for carrying that unspent tx doubles
3919 2013-03-13 20:10:41 <grau> gmaxwell: theoretical limits by side, I was curious how big costs currently are.
3920 2013-03-13 20:10:50 i2pRelay has quit (Remote host closed the connection)
3921 2013-03-13 20:11:15 Transisto has joined
3922 2013-03-13 20:11:20 i2pRelay has joined
3923 2013-03-13 20:11:25 <gmaxwell> PRab: have an update of your txn processing cost spreadsheet?
3924 2013-03-13 20:11:32 tonikt has quit (Quit: Leaving)
3925 2013-03-13 20:12:03 ZedsterX has quit (Read error: Connection reset by peer)
3926 2013-03-13 20:12:21 ZedsterX has joined
3927 2013-03-13 20:12:57 Morblias has joined
3928 2013-03-13 20:14:36 <gmaxwell> Vinnie_win: I think the cost per person using the system is perhaps more interesting than the total cost, however... I mean $1 per person per year is a _really_ cheap price for a good distributed currencyâ way less than anyone in the first world loses to inflation (so long as your geometric mean cash net worth is >$35)â but it's actually a lot of money when you multiply by 7 billion people.
3929 2013-03-13 20:15:08 LainZ has joined
3930 2013-03-13 20:15:25 <gmaxwell> the freerider problem is a sad thing, however. :(
3931 2013-03-13 20:16:28 zoinky has quit (Quit: Leaving.)
3932 2013-03-13 20:16:53 <Vinnie_win> gmaxwell: It seems that if the total amount of fees+subsidy collected over time ever becomes less than the total carrying cost of all unspent transactions, then the system would collapse
3933 2013-03-13 20:17:17 <gmaxwell> Vinnie_win: can we move to #bitcoin?
3934 2013-03-13 20:17:21 <Vinnie_win> sure
3935 2013-03-13 20:17:24 <Vinnie_win> sorry about that
3936 2013-03-13 20:17:31 pacpac has quit (Read error: Connection reset by peer)
3937 2013-03-13 20:17:34 gasteve has quit (Quit: leaving)
3938 2013-03-13 20:18:00 pacpac has joined
3939 2013-03-13 20:18:22 CaptainBlaze has quit (Quit: CaptainBlaze)
3940 2013-03-13 20:18:50 <grau> gmaxwell: since SD is strongly profitable and need the blockchain to work they might price out everybody else soon. We all might find ourselves working for SD as volunteers.
3941 2013-03-13 20:19:01 tyn has quit (Read error: Connection reset by peer)
3942 2013-03-13 20:19:17 _pr has quit (Ping timeout: 258 seconds)
3943 2013-03-13 20:19:24 zoinky has joined
3944 2013-03-13 20:19:36 BTCTrader2 has quit (Quit: BTCTrader2)
3945 2013-03-13 20:19:52 alexkravets has quit (Quit: Page closed)
3946 2013-03-13 20:20:09 <sudofox> Who else wants Testnet BTC? I have a bunch, giving it out for free just for the feels ^-^
3947 2013-03-13 20:20:13 _pr has joined
3948 2013-03-13 20:21:30 fposdkgs has joined
3949 2013-03-13 20:22:24 PlantMan has joined
3950 2013-03-13 20:22:30 nivoc has joined
3951 2013-03-13 20:22:39 zoinky has quit (Remote host closed the connection)
3952 2013-03-13 20:22:40 zooko has quit (Ping timeout: 264 seconds)
3953 2013-03-13 20:22:44 Transisto has quit (Ping timeout: 252 seconds)
3954 2013-03-13 20:27:08 * jgarzik scrolls back
3955 2013-03-13 20:27:13 <jgarzik> cryptopp does ECDSA?
3956 2013-03-13 20:27:17 <jgarzik> learn something new every day.
3957 2013-03-13 20:27:20 _pr has quit (Read error: Connection reset by peer)
3958 2013-03-13 20:30:32 zooko has joined
3959 2013-03-13 20:30:53 _pr has joined
3960 2013-03-13 20:31:32 Transisto has joined
3961 2013-03-13 20:32:02 hsmithsN7 has joined
3962 2013-03-13 20:32:19 gntbynynybt has quit (Ping timeout: 258 seconds)
3963 2013-03-13 20:32:38 fposdkgs has quit (Ping timeout: 255 seconds)
3964 2013-03-13 20:32:52 bitcoin-dev537 has joined
3965 2013-03-13 20:35:13 bitcoin-dev537 has quit (Client Quit)
3966 2013-03-13 20:36:10 lantastic has quit (Read error: Connection reset by peer)
3967 2013-03-13 20:36:35 pacpac has quit (Ping timeout: 256 seconds)
3968 2013-03-13 20:38:52 licnep has joined
3969 2013-03-13 20:40:32 <sudofox> Still giving out Testnet BTC for free here
3970 2013-03-13 20:40:51 alexwaters has joined
3971 2013-03-13 20:40:57 <sudofox> Am I spamming by saying that? People in #bitcoin don't seem to mind too much.
3972 2013-03-13 20:40:59 BTC_Bear has joined
3973 2013-03-13 20:41:33 fposdkgs has joined
3974 2013-03-13 20:42:36 D34TH has quit (Quit: Leaving)
3975 2013-03-13 20:43:21 Transisto has quit (Ping timeout: 240 seconds)
3976 2013-03-13 20:43:55 datagutt has quit (Quit: kthxbai)
3977 2013-03-13 20:44:18 eckey has joined
3978 2013-03-13 20:44:21 hsmithsN7 has quit (Ping timeout: 255 seconds)
3979 2013-03-13 20:45:00 owowo has joined
3980 2013-03-13 20:46:18 hsmithsN7 has joined
3981 2013-03-13 20:46:29 <eckey> is testNet3 running? i'm getting "connection refused" from PeerHandler. (this is the first time i've tried this in two weeks)
3982 2013-03-13 20:46:34 space_cadet has joined
3983 2013-03-13 20:46:34 space_cadet has quit (Changing host)
3984 2013-03-13 20:46:34 space_cadet has joined
3985 2013-03-13 20:46:37 <eckey> bitcoinj
3986 2013-03-13 20:47:16 vigilyn3 has quit (Read error: Connection reset by peer)
3987 2013-03-13 20:47:26 vigilyn3 has joined
3988 2013-03-13 20:48:07 amiller has quit (Excess Flood)
3989 2013-03-13 20:48:22 smakatak has quit (Ping timeout: 240 seconds)
3990 2013-03-13 20:48:34 norill has joined
3991 2013-03-13 20:48:45 Guest49455 has joined
3992 2013-03-13 20:48:54 BTCTrader2 has joined
3993 2013-03-13 20:49:41 <norill> is it possible to speed up initial blockchain synchronization?
3994 2013-03-13 20:49:55 Guest49455 is now known as amliler
3995 2013-03-13 20:49:56 <norill> i have a slow HDD
3996 2013-03-13 20:50:02 amliler is now known as amiller_
3997 2013-03-13 20:50:10 <sipa> norill: which version?
3998 2013-03-13 20:50:14 <sipa> of the client
3999 2013-03-13 20:50:24 <norill> v0.8.0-beta
4000 2013-03-13 20:50:26 Diablo-D3 has joined
4001 2013-03-13 20:50:43 hsmithsN7 has quit (Ping timeout: 258 seconds)
4002 2013-03-13 20:50:55 <sipa> one thing you can do is increase the database cache (-dbcache option)
4003 2013-03-13 20:51:09 <sipa> that reduces disk writes duromg ibd
4004 2013-03-13 20:51:17 <norill> ok, thank you
4005 2013-03-13 20:51:20 <jgarzik> Probably won't help, but,
4006 2013-03-13 20:51:26 <jgarzik> Blockchain torrent: https://bitcointalk.org/index.php?topic=145386.0
4007 2013-03-13 20:51:35 <sipa> it dependa where the problem lies
4008 2013-03-13 20:51:44 <jgarzik> (that's more for people having download/network problems, than slow drive)
4009 2013-03-13 20:52:08 <eckey> does anyone know the status of testNet3? thx
4010 2013-03-13 20:52:15 <TD> eckey: it should be working.
4011 2013-03-13 20:52:18 <sipa> and if the bottleneck is CPU (likely in the last 10000 blocks), there isn't much that will help
4012 2013-03-13 20:52:27 <TD> norill: you could also try multibit or a lighter client
4013 2013-03-13 20:53:35 <MC-Droid> thermoman i'm getting invalid chain and "no available lock entries" in db.log
4014 2013-03-13 20:53:46 <MC-Droid> thats how it started last night lol
4015 2013-03-13 20:53:52 <sipa> yes
4016 2013-03-13 20:53:55 <grau> sipa: Could checkpoints help us organizing protocol cleansing forks? I mean scripts are not evaluated before them right ?
4017 2013-03-13 20:54:14 <MC-Droid> then growing horror as we realise it wasnt just that one guy lol
4018 2013-03-13 20:54:22 <MC-Droid> logs are fun
4019 2013-03-13 20:54:23 kadoban has quit (Read error: Operation timed out)
4020 2013-03-13 20:54:26 <EvilPete> sipa: the bottleneck for OSX client is fsync() on hfs.. apple hfs is really bad with fsync() on a real HDD
4021 2013-03-13 20:55:29 zooko has quit (Ping timeout: 252 seconds)
4022 2013-03-13 20:55:33 Luke-Jr has quit (Excess Flood)
4023 2013-03-13 20:55:54 Luke-Jr has joined
4024 2013-03-13 20:56:24 Detritus has quit (Remote host closed the connection)
4025 2013-03-13 20:57:12 Xeno-Genesis has joined
4026 2013-03-13 20:57:36 Detritus has joined
4027 2013-03-13 20:57:57 <sipa> grau: such a rule can simplify code, as you don't need to support all earlier variations of validity rules
4028 2013-03-13 20:58:16 <sipa> grau: though i personally like the idea of being able to run in no-checkpoints mode
4029 2013-03-13 20:58:56 <gmaxwell> 13:30 < eckey> does anyone know the status of testNet3? thx
4030 2013-03-13 20:59:03 <gmaxwell> I think you're asking because blockexplorer is broken.
4031 2013-03-13 20:59:09 <gmaxwell> Thats just blockexplorer. The network is fine.
4032 2013-03-13 20:59:37 <grau> sipa: I also like the idea of simplifying the protocol for the sake of reliability
4033 2013-03-13 20:59:58 <sipa> grau: agree, but that is a hard thing to do
4034 2013-03-13 21:00:15 discrete has quit ()
4035 2013-03-13 21:00:37 <sipa> i'd like to push the enforcing of canonical pubkeys/signatures for example
4036 2013-03-13 21:00:51 <grau> sipa: The checkpoints act like genesis hash -lite. Scripts that are not checked do not really matter if they valid
4037 2013-03-13 21:01:04 <gwillen> sipa: canonical?
4038 2013-03-13 21:01:12 zooko has joined
4039 2013-03-13 21:01:32 <sipa> gwillen: no non-standard variations allowed, even tbough oopenssl eaccepts them
4040 2013-03-13 21:01:41 <gwillen> ahh, *nods*
4041 2013-03-13 21:01:45 <sipa> sorry for my typish, i'm on a phone
4042 2013-03-13 21:01:50 <gwillen> np
4043 2013-03-13 21:01:54 <sipa> lol, typiNG
4044 2013-03-13 21:02:04 <eckey> gmaxwell: i modified a running app from prodNet() to testNet3() using bitcoinj-0.6.1. it isn't able to connect to nodes because of refusal, timeouts, and no route.
4045 2013-03-13 21:02:08 <gwillen> is it the case that these are accepted now; and does anybody ever create them?
4046 2013-03-13 21:02:19 <TD> eckey: you probably forgot to change the peer discovery
4047 2013-03-13 21:02:19 <sipa> they are constantly created
4048 2013-03-13 21:02:30 <sipa> and accepted
4049 2013-03-13 21:02:39 <TD> eckey: you need to use IrcDiscovery("#bitcoinTEST3") instead of DnsDiscovery()
4050 2013-03-13 21:02:59 <randy-waterhouse> sipa: maybe bitcoin will eventually encorporate all the bits of the dependency packages it needs natively and leave out all those nasty one it doesn't want when it grows up ;)
4051 2013-03-13 21:03:06 <eckey> yikes!
4052 2013-03-13 21:03:08 <petertodd> eckey: For a quick test you can also just directly try connecting to the DNS address testnet-seed.bitcoin.petertodd.org which is the testnet dns seed
4053 2013-03-13 21:03:16 <eckey> ok, sorry
4054 2013-03-13 21:03:29 <phantomcircuit> randy-waterhouse, good lucky importing the openssl ASN.1 logic
4055 2013-03-13 21:03:30 <phantomcircuit> lol
4056 2013-03-13 21:03:33 Xeno-Genesis has quit (Remote host closed the connection)
4057 2013-03-13 21:03:57 ShaTwo has joined
4058 2013-03-13 21:04:29 thoughtcourier has joined
4059 2013-03-13 21:04:44 whooper has joined
4060 2013-03-13 21:04:57 <grau> sipa: we might tighten the rules if we forgo the option of full revalidating before the checkpoint, right? This would simplify code to the extent it is better captured, less vulnerable.
4061 2013-03-13 21:04:59 discrete has joined
4062 2013-03-13 21:05:01 <gmaxwell> phantomcircuit: see sipa's canonicazation stuff.
4063 2013-03-13 21:05:47 <eckey> thanks guys... my bad
4064 2013-03-13 21:06:23 moopie has joined
4065 2013-03-13 21:06:25 juanperez has joined
4066 2013-03-13 21:06:28 o2_ is now known as Blitzboom
4067 2013-03-13 21:06:35 Blitzboom has quit (Changing host)
4068 2013-03-13 21:06:35 Blitzboom has joined
4069 2013-03-13 21:08:08 Xeno-Genesis has joined
4070 2013-03-13 21:08:22 gmatteson_ has quit (Ping timeout: 260 seconds)
4071 2013-03-13 21:08:40 <grau> was this already discussed, or is blasphemy ?
4072 2013-03-13 21:09:37 saulimus has quit (Quit: saulimus)
4073 2013-03-13 21:09:48 <gmaxwell> grau: I don't believe we could simplify anything that way right now.
4074 2013-03-13 21:10:45 <Vinnie_win> sipa: Wei Dei got back to me he wants an idea of the performance difference between OpenSSL and CryptoPP's ECDSA verifications do you have any data?
4075 2013-03-13 21:10:49 <gmaxwell> It's really ungood if people cannot validate for themselves with transparent source code ... (though the fact that they could fetch an older version that does is probably good enough in principle)
4076 2013-03-13 21:10:50 zer0def has joined
4077 2013-03-13 21:10:57 <midnightmagic> +1 wei dai
4078 2013-03-13 21:11:53 <grau> gmaxwell: The POW gives enough confidence on the validity of the past if it is deep enough behind.
4079 2013-03-13 21:11:59 <gmaxwell> Vinnie_win: Sipa's implementation is now something like 4.5x faster than openssl. So I'm not sure what comparing CryptoPP's and openssl's is worth?
4080 2013-03-13 21:12:08 <gmaxwell> grau: it doesn't. :(
4081 2013-03-13 21:12:09 ZedsterX has quit (Read error: Connection reset by peer)
4082 2013-03-13 21:12:11 <gmaxwell> thats SPV security.
4083 2013-03-13 21:12:12 <midnightmagic> he fixed an assembly bug in his sha routines I helped discover in less than.. perhaps 2 hours.
4084 2013-03-13 21:12:24 ZedsterX has joined
4085 2013-03-13 21:12:30 <Vinnie_win> gmaxwell: As he said, he doesn't have a lot of confidence that it is bug-free. But regardless, Wei is offering to improve the performance of cryptopp ecdsa routines in a way that costs us nothing
4086 2013-03-13 21:12:33 <gmaxwell> grau: part of the reason that you can trust the pow at all is that rule validation bounds the incentive to cheat it.
4087 2013-03-13 21:13:25 maaku has left ()
4088 2013-03-13 21:13:27 <gmaxwell> grau: e.g. that fact that I can't spend a bunch on POW and as a result gain accounts with a trillion additional bitcoin is part of the reason it's reasonable for SPV clients to accept the POW alone.
4089 2013-03-13 21:13:55 <grau> gmaxwell: even if you have full blocks ?
4090 2013-03-13 21:14:02 <gmaxwell> Vinnie_win: well perhaps what we should do is direct wei to sipa's code and see if he can match that. :P
4091 2013-03-13 21:14:22 <Vinnie_win> gmaxwell: hah. is the bitcoin client using the sipa code currently?
4092 2013-03-13 21:14:25 <gmaxwell> grau: having them doesn't help if you're not validating anything before the checkpoint, which was what I thought you were suggesting.
4093 2013-03-13 21:15:04 <grau> I suggest the same validation as is done now before checkpoint, that IMHO does not include scripts
4094 2013-03-13 21:15:18 <gmaxwell> Vinnie_win: No, this is new work. Having wei dai work on itâ or even merge some implementation of these techniques would be interesting!
4095 2013-03-13 21:15:53 <grau> That means script oddities could be cleaned up if we forgo the option to full validate from genesis
4096 2013-03-13 21:16:14 <gmaxwell> grau: yes indeed. Which is enough to prevent inflation. It's a little unfortunate to not check at all, what I think we'd prefer is randomized checking but we don't have infrastructure to make use of randomized checking yet.
4097 2013-03-13 21:16:28 <gmaxwell> grau: I don't know what oddity you think could be cleaned up?
4098 2013-03-13 21:16:29 Lowflow has joined
4099 2013-03-13 21:16:40 <gmaxwell> Just the p2sh branch, but its a single test.
4100 2013-03-13 21:16:42 <sudofox> /clear
4101 2013-03-13 21:16:44 <sudofox> derp
4102 2013-03-13 21:17:14 <grau> gmaxwell: like extra signature formats, additional pops
4103 2013-03-13 21:17:56 <sudofox> There we go. | Does anyone want testnet btc? Giving it out still
4104 2013-03-13 21:18:19 <grau> one could simplify code
4105 2013-03-13 21:18:46 <petertodd> sudofox: Go away.
4106 2013-03-13 21:19:06 <sudofox> Okay
4107 2013-03-13 21:19:13 sudofox has quit (Quit: Have a nice day!)
4108 2013-03-13 21:20:08 <Magma_> I have a question about the Bitcoin protocol. It's possible to track the start of a transaction by looking at the network packets in Wireshark?
4109 2013-03-13 21:20:47 RBecker is now known as rbecker
4110 2013-03-13 21:21:30 <Xeno-Genesis> Magma_, why not?
4111 2013-03-13 21:22:13 <gmaxwell> grau: I'm still confused. Right now there is almost no simplification possible-- as all the stuff actually in the chain is valid under the currently effective rules. Maybe in the future that will be the case, and perhaps it would be worthwhile.
4112 2013-03-13 21:22:19 <Xeno-Genesis> Magma_, the Bitcoin software uses the Bitcoin protocol, and you can intercept it
4113 2013-03-13 21:22:30 <MC-Droid> bitcoin chatter is not obfuscated at all
4114 2013-03-13 21:22:39 <Magma_> Xeno-Genesis: Any good place for me to look?
4115 2013-03-13 21:22:44 bitafterbit has quit (Remote host closed the connection)
4116 2013-03-13 21:22:55 <Xeno-Genesis> I think there is a Wireshark plugin already AFAIK
4117 2013-03-13 21:23:29 <Magma_> In this specific case, I want to write a Snort module to track initialized Bitcoin transactions
4118 2013-03-13 21:24:59 Guest16536 has quit (Ping timeout: 246 seconds)
4119 2013-03-13 21:26:00 <BTC_Bear> Magma_: What's the Use Case?
4120 2013-03-13 21:26:03 SchmalzTech has quit ()
4121 2013-03-13 21:26:13 scintill has joined
4122 2013-03-13 21:27:14 <Magma_> BTC_Bear: Trading tracking in a local network
4123 2013-03-13 21:27:23 <gmaxwell> Magma_: you can't distinguish "initialized" bitcoin transactions from ones relayed. Even if you look at all traffic you might miss ones that are coming in via ways you miss, like tor.
4124 2013-03-13 21:28:34 <Magma_> gmaxwell: But if I know the address of trader A and trader B?
4125 2013-03-13 21:29:23 <gmaxwell> Magma_: oh in that case it usually just easier to modifiy bitcoin to log the transactions. or even make them appear in the wallet.
4126 2013-03-13 21:29:57 <BTC_Bear> Magma_: if the local network is exposed, trader A could trade with trader B without himself initializing a transaction from a local client.
4127 2013-03-13 21:30:01 nivoc has quit (Quit: nivoc)
4128 2013-03-13 21:30:23 PlantMan has quit (Remote host closed the connection)
4129 2013-03-13 21:30:37 PlantMan has joined
4130 2013-03-13 21:30:58 _pr has quit (Ping timeout: 258 seconds)
4131 2013-03-13 21:31:13 <TD> my bitcoind has reached 2G virt/785mb resident and it seems to be going up rather fast
4132 2013-03-13 21:31:35 <thermoman> MC-Droid: hey my heartbeat just jumped up when you hilighted me ... stop that!
4133 2013-03-13 21:31:39 <thermoman> :)
4134 2013-03-13 21:33:30 <helo> www.newscientist.com/blogs/onepercent/2013/03/bitcoin-zerocoin.html
4135 2013-03-13 21:33:46 <helo> oh, MC-Droid's in here
4136 2013-03-13 21:33:47 <diatonic> ;;getrating defaced
4137 2013-03-13 21:33:47 <gribble> Currently authenticated from hostmask defaced!4ac4307e@gateway/web/freenode/ip.74.196.48.126 . User defaced, rated since Tue Sep 20 15:24:20 2011. Cumulative rating 14, from 10 total ratings. Received ratings: 10 positive, 0 negative. Sent ratings: 11 positive, 0 negative. Details: http://bitcoin-otc.com/viewratingdetail.php?nick=defaced
4138 2013-03-13 21:35:14 <MC-Droid> thermoman: sorry it was in the log lol
4139 2013-03-13 21:35:28 <MC-Droid> and sorry again
4140 2013-03-13 21:35:40 nivoc has joined
4141 2013-03-13 21:37:00 scatha has joined
4142 2013-03-13 21:37:17 Hashdog has quit (Remote host closed the connection)
4143 2013-03-13 21:37:46 Lowflow has quit (Remote host closed the connection)
4144 2013-03-13 21:39:48 nivoc has quit (Client Quit)
4145 2013-03-13 21:41:08 grau has quit (Remote host closed the connection)
4146 2013-03-13 21:42:14 <topi`_> I upgraded from 0.6.2 to 0.8.0 and now it is "reindexing blocks on disk", whatever that means. probably re-doing the blkindex files?
4147 2013-03-13 21:42:19 Darin has joined
4148 2013-03-13 21:42:29 <topi`_> so it has to scan all that raw block data from blk0001.dat
4149 2013-03-13 21:43:03 <sipa> topi`_: correct
4150 2013-03-13 21:43:04 Darin has quit (Client Quit)
4151 2013-03-13 21:43:11 PatrikR has quit (Ping timeout: 246 seconds)
4152 2013-03-13 21:43:24 Darin has joined
4153 2013-03-13 21:43:29 Darin is now known as dman
4154 2013-03-13 21:43:37 dman is now known as thedman
4155 2013-03-13 21:44:09 <topi`_> sipa: has the indexing changed or why?
4156 2013-03-13 21:44:28 <randy-waterhouse> topi`_: it's a whole new database
4157 2013-03-13 21:44:29 <sipa> topi`_: the database layout and engine changed very substantially
4158 2013-03-13 21:44:31 <topi`_> ah, this must be the new leveldb
4159 2013-03-13 21:44:35 <sipa> indeed
4160 2013-03-13 21:44:46 <sipa> but also what and how the data is stored inside the database has changed
4161 2013-03-13 21:44:46 <topi`_> so you rebuild everything, but using the old raw block data
4162 2013-03-13 21:44:50 <sipa> indeed
4163 2013-03-13 21:44:53 JDuke128 has joined
4164 2013-03-13 21:45:24 <topi`_> are there checks that the rebuilding process works always correctly?
4165 2013-03-13 21:45:47 <topi`_> if it hasn't been tested, one could always revert to empty blk000x.dat
4166 2013-03-13 21:46:31 <sipa> it does the exact same thing as if those blocks were received from network
4167 2013-03-13 21:46:56 <topi`_> it seems to be receiving new blocks while this db building is happening. does it just receive blocks and append them to blk000x.dat in a separate thread?
4168 2013-03-13 21:47:45 <topi`_> for me, thread-programming is a bit no-no. it just gives you subtle bugs that are impossible to debug
4169 2013-03-13 21:48:53 scatha has quit (Quit: Konversation terminated!)
4170 2013-03-13 21:49:05 <Vinnie_win> sipa: Wei asked "why not use Pieter's code" (lol)
4171 2013-03-13 21:50:22 <sipa> topi`_: it shouldn't, but it's possible that it receives some things as they are received from the network
4172 2013-03-13 21:50:31 <sipa> topi`_: but they are just held in ram or dropped in that case
4173 2013-03-13 21:50:39 <sipa> Vinnie_win: wow, you actually mailed wei dai?
4174 2013-03-13 21:51:02 <Vinnie_win> sipa: Yeah, why not
4175 2013-03-13 21:51:14 m00p has joined
4176 2013-03-13 21:51:21 <Vinnie_win> sipa: He said that CrytoPP did not optimize specifically for secp256k1
4177 2013-03-13 21:51:42 <sipa> it's an obscure curve, that's only known because bitcoin uses it :)
4178 2013-03-13 21:52:00 <Vinnie_win> sipa: is the new hand rolled version optimized for that curve?
4179 2013-03-13 21:52:06 <sipa> yes
4180 2013-03-13 21:52:07 mow has joined
4181 2013-03-13 21:52:53 <sipa> https://github.com/sipa/secp256k1
4182 2013-03-13 21:53:23 <Vinnie_win> includes <openssl/...> ...
4183 2013-03-13 21:53:36 <sipa> for scalar operations, for now
4184 2013-03-13 21:53:44 <sipa> it can also use gmp
4185 2013-03-13 21:53:50 <Vinnie_win> so it's using openssl's "bignums"
4186 2013-03-13 21:53:53 rdymac has joined
4187 2013-03-13 21:54:05 <thedman> any bitcoin developers in the seattle area that understand the protocol? I need some help with a project, looking for a partner
4188 2013-03-13 21:54:16 <sipa> Vinnie_win: yes, but the bulk of the work is field/group operations, which are entirely hand-written
4189 2013-03-13 21:54:35 <sipa> Vinnie_win: and i hope to remove the dependency on a bignum library for the scalar operations too
4190 2013-03-13 21:54:41 <Vinnie_win> sipa: nice work
4191 2013-03-13 21:55:02 <sipa> thx :)
4192 2013-03-13 21:55:41 <gmaxwell> sipa: perhaps we should get wei to review BIP32. :P
4193 2013-03-13 21:55:48 <sipa> idea!
4194 2013-03-13 21:57:04 <sipa> gmaxwell: regarding BIP32, a derived key would be invalid if I_L is 0
4195 2013-03-13 21:57:17 <gmaxwell> sipa: yep.
4196 2013-03-13 21:57:40 o8gf5 has quit (Quit: Konversation terminated!)
4197 2013-03-13 21:57:47 discrete has quit ()
4198 2013-03-13 21:58:06 <sipa> gmaxwell: the chance for a 256-bit part of the output of SHA512 being zero is obviously ridiculously rare, so i wonder about either writing "the derived key is invalid if I_L is zero; we rely on SHA512 for this not to happen", or using something like 1+(...) mod (n-1)
4199 2013-03-13 21:58:54 <sipa> seems like a needless technical complication for something that's even more rare than the security level we aim for
4200 2013-03-13 21:59:00 <gmaxwell> Does ed25519 have this problem?
4201 2013-03-13 21:59:31 <gmaxwell> (as it makes a point from a hash, no?)
4202 2013-03-13 22:00:10 <gmaxwell> The bummer is that it breaks the whole chain, not just a single key. I'd feel less bad about 'we rely on SHA512 for this not to happen' if it was just a broken key...
4203 2013-03-13 22:01:35 <sipa> ed25519 uses G*(2^constant + secretkey), so no
4204 2013-03-13 22:01:57 scatha has joined
4205 2013-03-13 22:03:22 <weex> sorry to interrupt but i have been told that litecoin's implementation of scrypt makes the network vulnerable any single party with the first asic. gmaxwell and sipa do you have any opinion on this topic? trying to do an honest evaluation but i don't understand asic limitations well enough
4206 2013-03-13 22:03:36 <weex> to any single party*
4207 2013-03-13 22:03:50 zahhaatorute has joined
4208 2013-03-13 22:05:11 <Vinnie_win> someone said Litecoin
4209 2013-03-13 22:05:13 cap2002 has joined
4210 2013-03-13 22:05:34 <weex> yeah pretty much
4211 2013-03-13 22:05:36 <Vinnie_win> weex: I can answer you in #bitcoin
4212 2013-03-13 22:06:42 rdymac has quit (Ping timeout: 255 seconds)
4213 2013-03-13 22:07:41 asdbasjkd has joined
4214 2013-03-13 22:09:22 <asdbasjkd> hi
4215 2013-03-13 22:09:24 OneFixt has quit (Read error: Connection reset by peer)
4216 2013-03-13 22:09:27 Monarch__ has joined
4217 2013-03-13 22:09:43 gritcoin has joined
4218 2013-03-13 22:12:02 <BTC_Bear> Someone track this down: 1DfhsLMsiiTZueRu3TR4aL3JabsVYpvXpz was for 0.1035 during the fork from a -qt client.
4219 2013-03-13 22:12:25 joe_k has quit (Quit: Leaving.)
4220 2013-03-13 22:12:50 <asdbasjkd> good or bad time to buy bitcoins?
4221 2013-03-13 22:13:12 <BTC_Bear> nvm, just got 1 of 6... it worked.
4222 2013-03-13 22:13:49 cap2002 has quit (Quit: cap2002)
4223 2013-03-13 22:14:06 rdymac has joined
4224 2013-03-13 22:14:12 asdbasjkd has quit (Quit: Page closed)
4225 2013-03-13 22:15:28 Monarch__ has quit (Quit: Page closed)
4226 2013-03-13 22:15:29 btcur has quit (Quit: Leaving)
4227 2013-03-13 22:15:36 denisx has joined
4228 2013-03-13 22:16:02 ovidiusoft has quit (Quit: leaving)
4229 2013-03-13 22:16:17 rdymac has quit (Read error: Connection reset by peer)
4230 2013-03-13 22:17:07 scatha has quit (Ping timeout: 248 seconds)
4231 2013-03-13 22:18:51 scatha has joined
4232 2013-03-13 22:19:29 one_zero has joined
4233 2013-03-13 22:20:44 perezd_ has joined
4234 2013-03-13 22:20:45 <PRab> gmaxwell: not really. Was there something specific you were looking for?
4235 2013-03-13 22:20:46 gritcoin has quit (Quit: gritcoin)
4236 2013-03-13 22:20:58 safra has quit (Ping timeout: 245 seconds)
4237 2013-03-13 22:21:34 perezd has quit (Ping timeout: 258 seconds)
4238 2013-03-13 22:21:36 perezd_ is now known as perezd
4239 2013-03-13 22:22:27 dgriffi has quit (Ping timeout: 255 seconds)
4240 2013-03-13 22:24:08 dgriffi has joined
4241 2013-03-13 22:25:49 gritcoin has joined
4242 2013-03-13 22:28:30 <sipa> gmaxwell: still, i'd rather not change BIP32 at this point, certainly for something that's supposed to be a astronomical coincidence
4243 2013-03-13 22:29:04 amiller_ is now known as amiller
4244 2013-03-13 22:29:26 WebDev has joined
4245 2013-03-13 22:29:34 amiller is now known as Guest94877
4246 2013-03-13 22:29:36 <WebDev> Anyone interested in www.btccoin.com ???
4247 2013-03-13 22:29:49 WebDev is now known as Guest82088
4248 2013-03-13 22:29:58 <scatha> WebDev, sell at bitmit.net
4249 2013-03-13 22:29:59 space_cadet has quit (Remote host closed the connection)
4250 2013-03-13 22:30:12 <Guest82088> scatha how much u think i will get for it?
4251 2013-03-13 22:30:20 <scatha> no idea....
4252 2013-03-13 22:30:31 <Guest82088> is it an OG name though? :P
4253 2013-03-13 22:30:49 <jrmithdobbs> gmaxwell: aye, ed25519 doesn't have that problem
4254 2013-03-13 22:31:23 freewil has quit (Read error: Connection reset by peer)
4255 2013-03-13 22:32:52 <TD> heap profiling makes things really slow
4256 2013-03-13 22:33:49 <Guest82088> http://cryptofunds.com/
4257 2013-03-13 22:34:02 <Guest82088> Check my site out, Developers are needed :)
4258 2013-03-13 22:34:12 <jrmithdobbs> gmaxwell: well, at least it doesn't as used with the curved25519 signing scheme, pretty sure it doesn't in general though
4259 2013-03-13 22:34:27 <jrmithdobbs> curve not curved*
4260 2013-03-13 22:35:28 agricocb has quit (Quit: Leaving.)
4261 2013-03-13 22:35:40 unclemantis has joined
4262 2013-03-13 22:36:20 <unclemantis> how does the satoshidice pay to the same address that it received payment from? I want to set up a similar return payment service.
4263 2013-03-13 22:36:31 <jgarzik> sigh
4264 2013-03-13 22:36:45 <Scrat> :D
4265 2013-03-13 22:36:48 <jrmithdobbs> no you don't
4266 2013-03-13 22:37:00 <unclemantis> -_O
4267 2013-03-13 22:37:26 <TD> unclemantis: it just assumes the first input is owned by the sender
4268 2013-03-13 22:37:37 zoinky has joined
4269 2013-03-13 22:37:40 <TD> (hence the warning that users have to test it with a small amount of money first)
4270 2013-03-13 22:37:41 <doublec> Someone needs to write a "I want to return payments to the sender" faq that people can just point to
4271 2013-03-13 22:38:14 <unclemantis> TD, can you show me an example of which address it would send to based on where the bitcoin came from the client?
4272 2013-03-13 22:38:32 <unclemantis> doublec, I agree. Would be nice
4273 2013-03-13 22:38:48 <jouke> there are allready
4274 2013-03-13 22:38:48 <sipa> unclemantis: i figure doublec means the advise should be "Don't do it"
4275 2013-03-13 22:38:56 <doublec> unclemantis: say I send coins from a webwallet to your service. If you return them then I wouldn't get them.
4276 2013-03-13 22:39:07 <doublec> unclemantis: since the "sender" address is owned by the wallet
4277 2013-03-13 22:39:15 <unclemantis> Correct
4278 2013-03-13 22:39:15 <TD> unclemantis: what is your use case? giving people refunds?
4279 2013-03-13 22:39:21 <sipa> if you need to do refunds to a client, ask for a refund address
4280 2013-03-13 22:39:39 <sipa> (or soon, hopefully just use the payment protocol which should handle that for you)
4281 2013-03-13 22:39:53 <unclemantis> If I am setting up a pay per play game service. I suppose sending back to the same address would limit my clientel base
4282 2013-03-13 22:40:29 <unclemantis> so I guess I can say return address is OPTIONAL but not require or just make it required?
4283 2013-03-13 22:41:27 <Scrat> unclemantis: that's frowned upon (and sometimes wrong). just ask for an address when they click on withdraw
4284 2013-03-13 22:41:49 <TD> yeah if somebody asks for a refund for some reason just ask where you should send it
4285 2013-03-13 22:41:58 <TD> the payment protocol is supposed to optimize this once it's done i think
4286 2013-03-13 22:42:10 lb4956 has joined
4287 2013-03-13 22:42:11 <TD> as users wallets can upload a refund address automatically. but it's some ways out
4288 2013-03-13 22:42:17 <unclemantis> Scrat, so Sataoshi Dice is the EXCEPTION not the rule
4289 2013-03-13 22:42:27 <doublec> unclemantis: read this https://bitcointalk.org/index.php?topic=138752.0
4290 2013-03-13 22:42:46 <doublec> unclemantis: read all the comments, not just the first few that tell you how to do it. The followups explain why not to do it.
4291 2013-03-13 22:43:01 <unclemantis> I will just NOT DO IT
4292 2013-03-13 22:43:04 PlantMan has quit (Remote host closed the connection)
4293 2013-03-13 22:43:07 <unclemantis> Thanks for the insight folks
4294 2013-03-13 22:44:02 unclemantis has quit (Quit: Leaving)
4295 2013-03-13 22:45:25 _pr has joined
4296 2013-03-13 22:45:59 brwyatt is now known as Away!~brwyatt@brwyatt.net|brwyatt
4297 2013-03-13 22:46:23 tg has quit (Ping timeout: 245 seconds)
4298 2013-03-13 22:48:03 Anduck has quit (Ping timeout: 256 seconds)
4299 2013-03-13 22:48:09 tg has joined
4300 2013-03-13 22:49:27 Potato5491 has joined
4301 2013-03-13 22:50:59 PlantMan has joined
4302 2013-03-13 22:51:12 PlantMan has quit (Remote host closed the connection)
4303 2013-03-13 22:51:46 PlantMan has joined
4304 2013-03-13 22:51:49 PlantMan has quit (Remote host closed the connection)
4305 2013-03-13 22:52:10 PlantMan has joined
4306 2013-03-13 22:52:15 <Prattler> quick question... why does bitcoin uses/stores outputs and not total address balances? What does this achieve that balances would not?
4307 2013-03-13 22:52:25 pete78 has joined
4308 2013-03-13 22:53:29 Potato5491 has quit (Ping timeout: 245 seconds)
4309 2013-03-13 22:53:42 <Prattler> A has 1.5 BTC. Send all A coins to B 1.0 and A 0.5. Why are outputs used instead?
4310 2013-03-13 22:54:20 gritcoin has quit (Quit: gritcoin)
4311 2013-03-13 22:54:44 <gavinandresen> because outputs are more flexible than addresses.
4312 2013-03-13 22:54:56 <gavinandresen> outputs allow concepts like "either A or B can spend this"
4313 2013-03-13 22:55:04 _pr has quit (Ping timeout: 264 seconds)
4314 2013-03-13 22:55:30 CodeShark has joined
4315 2013-03-13 22:55:40 <sipa> not entirely true; you could do the same with addresses
4316 2013-03-13 22:55:40 <Prattler> bingo! thanks
4317 2013-03-13 22:55:44 <Prattler> aarg
4318 2013-03-13 22:55:54 <Prattler> hm..
4319 2013-03-13 22:56:11 <gavinandresen> sipa: yeah, I realized I was wrong as soon as I hit the second return
4320 2013-03-13 22:56:12 <Prattler> I understand they should be more flexible, but couldn't think of an example on my own
4321 2013-03-13 22:56:31 <sipa> but coins are more low-level: they exist or they don't exist; you need to store state for each
4322 2013-03-13 22:56:32 <Prattler> they - outputs
4323 2013-03-13 22:56:50 <sipa> plus the fact that the system is designed to use a new address for every transaction anyway
4324 2013-03-13 22:57:13 <sipa> *no need to store state
4325 2013-03-13 22:57:14 CaptainBlaze has joined
4326 2013-03-13 22:57:28 bolapara has quit (Ping timeout: 264 seconds)
4327 2013-03-13 22:57:35 agricocb has joined
4328 2013-03-13 22:59:09 <Prattler> well total current account balance data would be equivalent to UTXO data, but there would be no bloat from 0.00000001. Sounds simple... Outputs seem like they should be better, but I still don't see why
4329 2013-03-13 22:59:21 CaptainBlaze has quit (Client Quit)
4330 2013-03-13 22:59:48 <sipa> why wouldn't there be bloat from 0.00000001 ?
4331 2013-03-13 23:00:03 rbecker is now known as RBecker
4332 2013-03-13 23:00:49 <gavinandresen> Prattler: the real answer is "I don't want to think about it that deeply right now", but seems to me if you referred to spends by address then you'd have trouble matching up a sequence of fund/spend/fund/spent transactions
4333 2013-03-13 23:01:11 <gmaxwell> Prattler: if ~every interaction isn't to a seperate address the system loses any semblance of privacy. A lack of transactions forming a chain would also make it impossible to sort out which transactions survive a double spend reliably.
4334 2013-03-13 23:01:15 <Prattler> same block transactions would probably be impossible
4335 2013-03-13 23:02:15 <sipa> well, first of all, it would be to scripts, not to addresses
4336 2013-03-13 23:02:22 <TD> if you don't have a list of transactions nodes that go offline cannot catch up
4337 2013-03-13 23:02:24 <sipa> addresses simply don't exist at that level
4338 2013-03-13 23:02:58 <sipa> but there would also not be any need for matching up output transactions with inputs
4339 2013-03-13 23:03:04 <gmaxwell> Prattler: for example, say you have 6 btc and you send txn for 3 (A) and 3 (B) .... but then 3 (A)'s reciever is mad that the confirmation is taking a while so you double spend 3 (A) with fees as 3(A'). but then the miner manages to confirm 3(A) and 3(A') and 3(B) gets rejected.
4340 2013-03-13 23:03:24 <gmaxwell> so you've failed to pay someone you thought you paid, and double paid A.
4341 2013-03-13 23:03:27 <sipa> i figure the real reason is what gmaxwell says: you're not supposed to reuse addresses anyway, so there is no reason for the system to optize that behaviour
4342 2013-03-13 23:03:51 <gavinandresen> nah, B would just wait until there were more coins there and then retransmit 3(B) to get themâ¦
4343 2013-03-13 23:03:52 p8m has quit (Read error: No route to host)
4344 2013-03-13 23:03:56 <sipa> hmm, tracking confirmations is indeed another reason!
4345 2013-03-13 23:03:57 <gavinandresen> :)
4346 2013-03-13 23:04:24 Guest94877 is now known as amiller_
4347 2013-03-13 23:04:26 <gmaxwell> gavinandresen: yea, of course before that happened you paid (B) from another account ...
4348 2013-03-13 23:04:59 juanperez has quit (Quit: Page closed)
4349 2013-03-13 23:05:47 <Prattler> gmaxwell/gavinandresen, you're right. Balances would not allow chaining unconfirmed transactions, as transaction order is important. That doesn't sound like a huge limitation though.
4350 2013-03-13 23:06:15 <CodeShark> it's a significant limitation for some applications I've worked on :)
4351 2013-03-13 23:06:21 <sipa> it would be somewhat different system, i'm sure it would have some advantages, but certainly its own challenges too
4352 2013-03-13 23:06:44 <gmaxwell> Prattler: I think it is. right now we can tell people 'bitcoin txn are instant, they just take time to become irreversable' but that wouldn't be true if you had to hold transactions until you got your last one mined!
4353 2013-03-13 23:06:51 <CodeShark> for instance, say you want to distribute received funds to exchanges ASAP for liquidity purposes
4354 2013-03-13 23:07:04 <randy-waterhouse> seems like a moot discussion?
4355 2013-03-13 23:07:05 <CodeShark> not being able to spend unconfirmed coins is a big problem
4356 2013-03-13 23:07:08 <gmaxwell> To overcome that limitation you'd need to use different accounts for everything and .. you'd be basically back to the current model! :P
4357 2013-03-13 23:07:13 aethero has joined
4358 2013-03-13 23:07:19 <aethero> http://bitcoin.stackexchange.com/questions/8384/custom-filter-for-mtgox-depth-api-request <-- anyone know what programming language this is?
4359 2013-03-13 23:07:23 <Prattler> gmaxwell, ok.. 1 account wouldn't be able to do more than 1 transaction per block.. That is limiting..
4360 2013-03-13 23:07:32 pete79 has joined
4361 2013-03-13 23:07:34 <sipa> Prattler: i don't see why?
4362 2013-03-13 23:07:39 <Scrat> aethero: its json. not a language
4363 2013-03-13 23:07:43 <aethero> Gracias
4364 2013-03-13 23:07:47 aethero has left ("Once you know what it is you want to be true, instinct is a very useful device for enabling you to know that it is")
4365 2013-03-13 23:07:47 <sipa> 'balance' would be a function of depth
4366 2013-03-13 23:07:56 <gmaxwell> Prattler: well, and you've ignored the whole privacy point I made. Bitcoin's privacy is quite fragle. In some ways its the weakest aspect of the system.
4367 2013-03-13 23:08:38 <Prattler> I also agree that if addresses are never reused, this system is equivalent to outputs.
4368 2013-03-13 23:09:04 <Prattler> balances are equivalent to outputs, because there would never be more than 1 output with the same address.
4369 2013-03-13 23:09:23 <sipa> randy-waterhouse: very moot, but perhaps enlightening :_
4370 2013-03-13 23:09:27 <gmaxwell> BIP32 will close off a bunch of reasons people reuse addresses. Hopefully theymos will even get BIP32 support for forum signatures. :P
4371 2013-03-13 23:09:55 <sipa> gmaxwell: you're ok with simply saying "a key is invalid if I_L is 0?"
4372 2013-03-13 23:10:13 <Luke-Jr> gmaxwell: thoughts on a new address form that finds the next unused "child" key and sends to it? :P
4373 2013-03-13 23:10:19 <sipa> s/\?"/"\?/
4374 2013-03-13 23:10:57 <sipa> Luke-Jr: publically revealing parent extended keys can be dangerous
4375 2013-03-13 23:10:59 <randy-waterhouse> any future plans to put p2p mixing into the protocol?
4376 2013-03-13 23:10:59 <CodeShark> the substitution rule is more confusing than the original typo
4377 2013-03-13 23:11:15 CaptainBlaze has joined
4378 2013-03-13 23:11:17 <sipa> CodeShark: not entirely unintentionally, i might add :p
4379 2013-03-13 23:11:20 <randy-waterhouse> considering privacy is "fragile"
4380 2013-03-13 23:11:25 <Luke-Jr> sipa: how dangerous?
4381 2013-03-13 23:11:39 <Luke-Jr> randy-waterhouse: it's been discussed
4382 2013-03-13 23:11:53 <Luke-Jr> randy-waterhouse: but not mixing like laundry services
4383 2013-03-13 23:11:55 <sipa> Luke-Jr: knowing a parent extended pubkey + knowing any child private key means being able to compute the parent private key
4384 2013-03-13 23:12:11 <sipa> Luke-Jr: it's the one thing that makes me unhappy about BIP32
4385 2013-03-13 23:12:24 <Luke-Jr> sipa: why would someone know the child privkey?
4386 2013-03-13 23:12:33 <gmaxwell> sipa: but how do we keep that from killing the chain?
4387 2013-03-13 23:12:33 eric____ has quit (Quit: leaving)
4388 2013-03-13 23:12:37 <MC-Droid> some prof from baltimore university thinks he has a way
4389 2013-03-13 23:12:47 <CodeShark> what would be a use case where you want someone to know the child privkey but not the parent privkey?
4390 2013-03-13 23:12:56 <MC-Droid> calling it zerocoin, but details are nil
4391 2013-03-13 23:12:57 <sipa> gmaxwell: why should it - this is something that's deemed impossible
4392 2013-03-13 23:13:39 <MC-Droid> as an "add on" to bitcoin(?)
4393 2013-03-13 23:13:51 <sipa> CodeShark: accounts, and give someone authority to spend from one account but not all
4394 2013-03-13 23:14:00 <randy-waterhouse> Luke-Jr: but no concrete plans ... i keep seeing snippets here and there and wondered if there was proposal or spec. somwhere i was missing?
4395 2013-03-13 23:14:31 <TD> no plans yet
4396 2013-03-13 23:14:46 <TD> there have always been higher priorities
4397 2013-03-13 23:14:51 <gmaxwell> sipa: meh. I'd feel happier if you could just skip that key, but you can't. I suppose your stuck wallet will be a small price to pay for the great fame you can get by knowing the SHA512 preimage for all zeros.
4398 2013-03-13 23:14:56 <TD> if you want to prototype it go ahead, i doubt anyone will be competing with you on that
4399 2013-03-13 23:15:13 <gmaxwell> (well I suppose 256 bits of zeros :P )
4400 2013-03-13 23:15:38 enikanorov_ has quit (Remote host closed the connection)
4401 2013-03-13 23:16:00 enikanorov has joined
4402 2013-03-13 23:16:00 <gmaxwell> Luke-Jr: say you decide you want your webserver to be able to sign for some addresses in a common wallet, so you load it with a child chain, and a full extended private key for its root.
4403 2013-03-13 23:16:23 <randy-waterhouse> TD: noone fro google anyway
4404 2013-03-13 23:16:28 <randy-waterhouse> s/fro/from
4405 2013-03-13 23:16:53 <CodeShark> keeping bitcoin privkeys on webservers is probably not such a good idea to begin with :p
4406 2013-03-13 23:16:58 <gmaxwell> randy-waterhouse: it doesn't need to be put into the protocol.
4407 2013-03-13 23:17:06 <gmaxwell> CodeShark: right sure, but it could be worse than the user expects.
4408 2013-03-13 23:17:18 <TD> randy-waterhouse: what's google got to do with anything?
4409 2013-03-13 23:17:32 <randy-waterhouse> they have a less than stellar record on privacy
4410 2013-03-13 23:17:35 <TD> i think having it as a feature of the p2p protocol is a good idea
4411 2013-03-13 23:17:38 <randy-waterhouse> is all I was saying
4412 2013-03-13 23:17:39 <gmaxwell> randy-waterhouse: p2p mixing should really be external to the bitcoin protocol for a number of reasonsâ one being that it should probably be conducted over tor.
4413 2013-03-13 23:17:53 <TD> as you'd need a p2p protocol anywy
4414 2013-03-13 23:18:02 moopie has quit (Quit: Page closed)
4415 2013-03-13 23:18:10 <TD> (so people who want to swap coins can find each other)
4416 2013-03-13 23:18:25 <gonffen_> randy-waterhouse: nice nick ;)
4417 2013-03-13 23:19:27 <gmaxwell> TD: I don't agree. The probablity of any particular bitcoin peer you'd connect to being interested in this stuff is ~0. And bitcoin's protocol has poor privacy properties.
4418 2013-03-13 23:20:03 <gmaxwell> Stuffing everything in one protocol isn't a good idea when the participants don't want everything.
4419 2013-03-13 23:20:25 <TD> i was thinking that if you have some coins and want to make a combined large transaction with other people to obfuscate which inputs map to which outputs, you need a way to find other people who are online at the time in order to make that tx
4420 2013-03-13 23:20:36 <TD> which means some kind of "i am here and interested, is anyone out there?" type mechanism
4421 2013-03-13 23:20:49 <gmaxwell> Sure. You need a group broadcast medium.
4422 2013-03-13 23:21:02 <CodeShark> like IRC :)
4423 2013-03-13 23:21:12 <TD> once nodes that want to mix rendezvous they can either just use direct IP or Tor indeed
4424 2013-03-13 23:21:13 <sipa> haha
4425 2013-03-13 23:21:17 <randy-waterhouse> https://bitcointalk.org/index.php?topic=138752.msg1479445#msg1479445 ... gavin mentions p2p in here so was thinking it might be in long-term scheme ... since fungibility and privacy become inseparable in digital world why not make it part of the protocol?
4426 2013-03-13 23:21:30 <jrmithdobbs> TD: that sounds like inviting all sorts of abuse
4427 2013-03-13 23:21:32 <gmaxwell> And then you need a secure mixing protocol (which are all high bandwidth), or you need your broadcast medium to be strongly anonymous.
4428 2013-03-13 23:22:29 <TD> i suppose it depends who implements it and how much effort they make.
4429 2013-03-13 23:22:33 Spunkie_ is now known as Spunkie
4430 2013-03-13 23:22:39 Spunkie has quit (Changing host)
4431 2013-03-13 23:22:40 Spunkie has joined
4432 2013-03-13 23:23:00 brson_ has joined
4433 2013-03-13 23:23:17 <gmaxwell> (e.g. multiparty reencryption mixes are O(peers^2) bandwidth, if you don't have a mixer to hide particpants from each other there is an incentive to participate to just monitor people :P )
4434 2013-03-13 23:23:52 brson has quit (Ping timeout: 252 seconds)
4435 2013-03-13 23:24:28 COGSMITH has quit (Quit: .)
4436 2013-03-13 23:24:38 <TD> something like tor is a double edged sword. with direct IP connections, if someone tries to monitor every mix, or a lot of them, people can notice the same IP cropping up over and over again
4437 2013-03-13 23:24:38 Xeno-Genesis has quit (Ping timeout: 248 seconds)
4438 2013-03-13 23:24:47 <TD> with tor, they can take part in every mix and nobody would ever know
4439 2013-03-13 23:25:30 <MC-Droid> i thought gavin didnt care for any mewsures to make bitcoin more anon
4440 2013-03-13 23:25:31 gritcoin has joined
4441 2013-03-13 23:25:44 <gavinandresen> what made you think that?
4442 2013-03-13 23:25:46 <TD> i think it makes sense for tor to be available. mandating it might be a bit much. a lot of people have tor-like privacy already without even realizing due to the prevalence of carrier nat
4443 2013-03-13 23:26:02 <gavinandresen> I just dont' want half-measures that give users a false sense of anonymity
4444 2013-03-13 23:26:08 OneMiner is now known as richmanmoneybags
4445 2013-03-13 23:26:36 CaptainBlaze_ has joined
4446 2013-03-13 23:26:41 richmanmoneybags is now known as OneMiner
4447 2013-03-13 23:27:10 pete79 has quit (Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org)
4448 2013-03-13 23:27:24 <MC-Droid> i....cant remember
4449 2013-03-13 23:27:47 CaptainBlaze has quit (Ping timeout: 260 seconds)
4450 2013-03-13 23:27:47 CaptainBlaze_ is now known as CaptainBlaze
4451 2013-03-13 23:27:51 <MC-Droid> maybe something you wrote that i read once
4452 2013-03-13 23:28:20 <gmaxwell> TD: you can connect to tor from one app without running tor for anything else. Effectively using it just like a special network protocol. This is what torchat does, for example.
4453 2013-03-13 23:28:57 pete79 has joined
4454 2013-03-13 23:29:01 <TD> perhaps that time he wrote, "i dislike privacy, it bothers me that terrorists and money launders who use fake names like satoshi could use bitcoin. in future we should add an 'uploadpassport' command to the protocol"
4455 2013-03-13 23:29:29 <jouke> I saw that one.
4456 2013-03-13 23:29:33 <gmaxwell> TD: it is utterly trivial to get a ton of non-tor IPs... e.g. using amazon, or if you're willing to spend more than a tiny amount spread out across many networks. There are commercial p2p poisoning services based on this.
4457 2013-03-13 23:29:40 <TD> gmaxwell: yeah. i think tor will become a lot more common in future. but i think the p2p protocol could still use a rendezvous mechanism in it. even if all it does it let some people agree on a hidden service to use to run the mix
4458 2013-03-13 23:29:46 <MC-Droid> that sounds pretty sarcastic
4459 2013-03-13 23:29:52 <randy-waterhouse> i liked the hookers 'n blow one better
4460 2013-03-13 23:29:53 <TD> MC-Droid: heh
4461 2013-03-13 23:30:21 <TD> gmaxwell: in particular i think web sites that take deposits should run their core logic+hot wallet behind a hidden service so it's harder to do VPS infrastructure attacks
4462 2013-03-13 23:30:31 <TD> but this is all quite some way out yet
4463 2013-03-13 23:30:38 <gmaxwell> Well I don't care about anonymity I care about privacy. Anonymity is really really really hard. Privacy is a basic feature which any ecommerce system should have. It just so happens we use anonymity to get privacy.
4464 2013-03-13 23:30:59 <MC-Droid> whats the difference
4465 2013-03-13 23:31:08 <TD> ah. good question. this should be in the faq
4466 2013-03-13 23:31:17 <TD> imagine satoshi cashes out tomorrow
4467 2013-03-13 23:31:24 <TD> sells all his bitcoins for dollars
4468 2013-03-13 23:31:28 <CodeShark> strictly speaking, we use pseudonymity - not anonymity :)
4469 2013-03-13 23:31:36 <TD> we can see that in the block chain. he does not have financial privacy
4470 2013-03-13 23:31:45 <gmaxwell> CodeShark: I am pedantic about that 99/100 times ... give me a break. :P
4471 2013-03-13 23:31:48 <TD> however we don't know who he really is. so he is anonymous (somewhat)
4472 2013-03-13 23:32:16 <TD> clearly it's possible to suffer privacy violations, despite being anonymous (unless you have no linkable actions or identity at all)
4473 2013-03-13 23:32:35 <TD> most people are (a) afraid of anonymity and (b) very keen on privacy
4474 2013-03-13 23:32:51 <TD> so bitcoin is sort of backwards, today. it's anonymous by default and it's remarkably easy to leak sensitive financial information
4475 2013-03-13 23:32:57 <MC-Droid> the way the internet is going anonymity is privacy
4476 2013-03-13 23:33:02 <TD> in future i think it'll get better at offering tradeoffs
4477 2013-03-13 23:33:09 <gmaxwell> MC-Droid: when I trade with you, you know who I am. (or at least nothing about bitcoin prevents you from knowing who I am) But I _do not_ want you to know how wealthy I am, or what other transactions I've been making, or if I like bondage porn.
4478 2013-03-13 23:33:30 <TD> MC-Droid: hardly! the big trend of the last 10 years is the rise of Facebook in which people use their real names and have a billion different privacy controls
4479 2013-03-13 23:33:35 <MC-Droid> im not judging
4480 2013-03-13 23:34:10 <TD> it turned out that for most people, anonymity (being identified as PinkPoney84) was kind of useless/annoying, but privacy (only my friends can see my photos) was a killer feature
4481 2013-03-13 23:34:12 <gmaxwell> In some ways things like facebook exploit privacy controls to actually make people feel more private (because they have lots of knobs) while making them less private. People at least like the appearance of privacy.
4482 2013-03-13 23:34:14 <jrmithdobbs> gmaxwell: and so long as you don't re-use addresses (and aren't spending coinbases) it all works out pretty fine as is
4483 2013-03-13 23:34:37 <TD> facebook definitely gave people more privacy from each other than what came before. it made people less private to facebook itself.
4484 2013-03-13 23:34:39 <gmaxwell> And most facebook persononas are carefully construted lies in so many ways.
4485 2013-03-13 23:34:56 <MC-Droid> tru dat
4486 2013-03-13 23:35:15 <TD> but overall, a well known corporation with a brand to protect is usually a bit more trustable than $JOE_PERVERT that someone met at a bar last night. so it's a good tradeoff.
4487 2013-03-13 23:35:18 <MC-Droid> td fb is a bad example for this
4488 2013-03-13 23:35:35 <MC-Droid> you have privacy fromnyour friends list but facrbook sees all
4489 2013-03-13 23:35:38 <jrmithdobbs> gmaxwell: what specifically is at issue with bip32?
4490 2013-03-13 23:35:43 <gmaxwell> jrmithdobbs: yes, used well bitcoin can achieve privacy. But sometimes it makes sense to use it poorly.. e.g. a public donation address is useful if you have a one way communication channel. But then you need to disassociate that to avoid further privacy problems.
4491 2013-03-13 23:35:45 bernard75 has quit ()
4492 2013-03-13 23:35:45 <TD> yes, i just said that :)
4493 2013-03-13 23:35:45 <jrmithdobbs> that is causing this conversation
4494 2013-03-13 23:36:03 <gmaxwell> MC-Droid: I think thats a perfect example actually.
4495 2013-03-13 23:36:28 <MC-Droid> fuck i cant type fast enough on this thing
4496 2013-03-13 23:36:29 <jrmithdobbs> gmaxwell: do we really need a solution to this or does someone need to create a better donation service?
4497 2013-03-13 23:36:46 <TD> facebook can be seen as a "social data bank". it knows everything and we trust it to compartmentalize the data from each other
4498 2013-03-13 23:36:50 <gmaxwell> MC-Droid: I feel private with my bank accountâ even though my bank sees all. I don't care about my bank seeing all, and if they tell anyone without a court order they've violated their agreement with me so I trust that to be enough for my needs of not letting you know about my fettish for ponys.
4499 2013-03-13 23:36:54 <jrmithdobbs> gmaxwell: eg, i bet joepie would be interested in helping people do it "right" for donations with his redonate project
4500 2013-03-13 23:37:01 <TD> right
4501 2013-03-13 23:37:20 <MC-Droid> funny you should say that gmaxwell
4502 2013-03-13 23:37:22 <gmaxwell> jrmithdobbs: well, I specified "one way channel" for a reason. :P I don't think you can do better than static addresses for one way broadcast.
4503 2013-03-13 23:37:30 <MC-Droid> wheres that story....
4504 2013-03-13 23:37:39 <jrmithdobbs> gmaxwell: well, i'm coming at the problem from a different angle
4505 2013-03-13 23:37:45 <randy-waterhouse> gmaxwell: many countries have changed laws to allow police to view bank accounts without any court orders or notifying the account holders
4506 2013-03-13 23:37:49 <jrmithdobbs> gmaxwell: eg, right now say you want to donate to the FreeBSD foundation using usd
4507 2013-03-13 23:38:12 <jrmithdobbs> gmaxwell: do you just shoot a check to FreeBSD Foundation, Some street, Some city usa ? no
4508 2013-03-13 23:38:17 <MC-Droid> story about fincen opening up to every agency by default now
4509 2013-03-13 23:38:33 <jrmithdobbs> gmaxwell: you contact the organization (in this case through a form on their website) that gives you donation instructions
4510 2013-03-13 23:38:37 <gmaxwell> randy-waterhouse: I only would mind that to the extent that idiotic misunderstandings and stupid machine identification could cause that access to turn into a investigation that wastes my time.
4511 2013-03-13 23:38:46 <jrmithdobbs> gmaxwell: the same could be done generating random donate addresses
4512 2013-03-13 23:38:57 <MC-Droid> also thanks usa for taking my SWIFT data sincerely yurop
4513 2013-03-13 23:39:37 <gmaxwell> jrmithdobbs: yes that could, but people do donate out of effective broadcast mediums. Doubly so because indiviguals don't tend to have stable and secure websites (except hosted ones).
4514 2013-03-13 23:39:38 <jrmithdobbs> what I'm asking is, should we work around people's misuse of the infrastructure, and even encourage it by standardizing their one-offs, or should we find ways to help people use it as originally designed?
4515 2013-03-13 23:39:52 <randy-waterhouse> gmaxwell: i know what you mean ... it is going to become a giant trawl fest now they have opened up the door to such activity ... you can only hope you don't get enmeshed in some bureacratic mess up
4516 2013-03-13 23:40:31 slavik0329 has joined
4517 2013-03-13 23:40:38 <gmaxwell> jrmithdobbs: but sure, we should have a good bip32 address producer cgi. And the bitcointalk forum should even support oneâ since I don't mind if theymos can enumerate addresses that I recieved forum donations at.
4518 2013-03-13 23:41:10 ThomasV has quit (Ping timeout: 248 seconds)
4519 2013-03-13 23:41:14 <sipa> gmaxwell: any other suggestions regarding the I_L==0 case?
4520 2013-03-13 23:41:18 <slavik0329> Anyone know how I can get the input addresses from a transaction using RCP?
4521 2013-03-13 23:41:27 <slavik0329> RPC*
4522 2013-03-13 23:41:29 <Scrat> how about a centralized mixing service operated by the Foundation. with a small randomized fee that will act as a donation
4523 2013-03-13 23:41:40 <jrmithdobbs> no
4524 2013-03-13 23:41:46 <jrmithdobbs> just no
4525 2013-03-13 23:41:47 <gmaxwell> Scrat: egads.
4526 2013-03-13 23:42:07 <gmaxwell> First principle of not making a freeking mess: do not conflate things that don't need to be conflated.
4527 2013-03-13 23:42:12 <jrmithdobbs> their was enough vitroil targeted at people in here when they pulled off the amazing the other night
4528 2013-03-13 23:42:22 <jrmithdobbs> can you imagine the backlash when something goes wrong with the foundation mixer? ha
4529 2013-03-13 23:42:38 <gmaxwell> Scrat: besides, centralized mixing is idiotic.
4530 2013-03-13 23:42:44 <jrmithdobbs> vitriol*
4531 2013-03-13 23:42:47 <gmaxwell> Zero trust mixing works fine.
4532 2013-03-13 23:43:05 <MC-Droid> dammit i really cant type fast enough on this
4533 2013-03-13 23:43:08 <gmaxwell> sipa: no, it's just invalid, skip that one.
4534 2013-03-13 23:43:11 <gavinandresen> foundation mixer won't happen with the current Board, too risk-averse
4535 2013-03-13 23:43:24 <gavinandresen> ⦠and bad PR potential, too
4536 2013-03-13 23:43:49 <gmaxwell> gavinandresen: It would be a really bad idea. I consider it a good sign that you think they're risk adverse.
4537 2013-03-13 23:44:12 <gmaxwell> There is no need to have something like that run by a trusted entity. ... and no reason you'd want the foundation brought into the line of fire if someone decided the mixer was unlawful.
4538 2013-03-13 23:44:13 Pucilowski has quit (Read error: Operation timed out)
4539 2013-03-13 23:44:38 JDuke128 has quit (Quit: [BB])
4540 2013-03-13 23:45:04 <jrmithdobbs> i can see the headlines now .... "Bitcoin Foundation caught laundering money for child pornographers"
4541 2013-03-13 23:45:10 <jrmithdobbs> ...
4542 2013-03-13 23:45:19 <MC-Droid> e foundation will probably be mainly to deflect accusations of pedoterrorcoin, not bolster them
4543 2013-03-13 23:45:20 Pucilowski has joined
4544 2013-03-13 23:45:27 <MC-Droid> bad idea
4545 2013-03-13 23:45:40 <CodeShark> let's move on - not worth discussing this idea any further :p
4546 2013-03-13 23:46:21 <MC-Droid> jrmithdobbs: hsbc got off scott, why not us :p
4547 2013-03-13 23:46:58 oleganza has quit (Quit: oleganza)
4548 2013-03-13 23:47:16 <randy-waterhouse> has anyone here studied Watson Ladd's OP SIG_FUNGIBLE proposal? http://wbl.github.com/bitcoinanon.pdf
4549 2013-03-13 23:47:58 <eckey> aethero: looks like a JSON object
4550 2013-03-13 23:48:02 <randy-waterhouse> seems like it raises more questions than answers
4551 2013-03-13 23:48:31 benjam55 has quit (Quit: Leaving)
4552 2013-03-13 23:48:34 <randy-waterhouse> something for the alt coins
4553 2013-03-13 23:48:58 <gavinandresen> yeah, why haven't any of the alt coins started doing stuff like that?
4554 2013-03-13 23:49:16 <randy-waterhouse> gavinandresen: because all the money is in bitcoin?
4555 2013-03-13 23:49:42 <gavinandresen> mmm. And I suppose you could just as easily if (fTestnet) â¦. the code and use testnet like an altcoin
4556 2013-03-13 23:50:41 <gmaxwell> randy-waterhouse: search the development mailing list.
4557 2013-03-13 23:50:43 <TD> hmm
4558 2013-03-13 23:50:51 <TD> gavinandresen: have you ever looked at a heap profile of bitcoind?
4559 2013-03-13 23:51:03 <randy-waterhouse> gmaxwell: link t mailing list?
4560 2013-03-13 23:51:11 <gavinandresen> TD: nope
4561 2013-03-13 23:51:29 <sipa> TD: i tried using valgrind's heap profiler once, and found the result very uninteresting
4562 2013-03-13 23:51:35 <TD> gavinandresen: i see nearly 62% of memory allocated to what appears to be CNode::vSend
4563 2013-03-13 23:51:52 <sipa> TD: yeah, network buffers; expected
4564 2013-03-13 23:52:08 <sipa> jeff's patches fix that in a very nice way, but alas, they segfault
4565 2013-03-13 23:52:23 <randy-waterhouse> gmaxwell: this one? http://sourceforge.net/mailarchive/forum.php?forum_name=bitcoin-development
4566 2013-03-13 23:52:24 <TD> for sending data? why? messages are supposed to be serialized, written to the socket and then freed, i'd have thought.
4567 2013-03-13 23:52:42 <TD> satoshis code is doing something odd with CDataStream where begin() doesn't actually return the start of the buffer
4568 2013-03-13 23:52:44 kadoban has joined
4569 2013-03-13 23:52:44 <sipa> TD: the network sender thread is separate from the message processing thread
4570 2013-03-13 23:52:46 rdymac has joined
4571 2013-03-13 23:52:53 <gmaxwell> randy-waterhouse: yes.
4572 2013-03-13 23:52:56 <randy-waterhouse> ta
4573 2013-03-13 23:53:25 <sipa> TD: jeff's code changes that to first check whether the buffer is empty (which it will be in many many cases) and just send immediately in that case
4574 2013-03-13 23:53:48 <Scrat> thanks for killing my dream guys :p
4575 2013-03-13 23:54:03 <TD> yes, but i don't think that can explain 62% of all memory being used by this. memory should only be sitting in the buffer temporarily. the cpu was not heavily loaded
4576 2013-03-13 23:54:18 <Scrat> gmaxwell: decentralized mixing does work great but it would be too cumbersome even with a dedicated application
4577 2013-03-13 23:54:21 <TD> sipa: which buffer?
4578 2013-03-13 23:54:24 <TD> vSend?
4579 2013-03-13 23:54:29 <sipa> actually
4580 2013-03-13 23:54:45 <MC-Droid> it works?
4581 2013-03-13 23:54:49 <sipa> just doing a vSend.swap(std::vector()) when it is empty would help a lot
4582 2013-03-13 23:54:54 <gmaxwell> Scrat: I can assert random things too! Your mother is an icosahedron!
4583 2013-03-13 23:55:10 <sipa> as std::vector::clear doesn't actually free memory
4584 2013-03-13 23:55:32 <Scrat> gmaxwell: she's a mathematician, how'd you guess
4585 2013-03-13 23:55:37 <Scrat> gmaxwell confirmed for wizard
4586 2013-03-13 23:55:52 <HM2> C++ introduced strink_to_fit()
4587 2013-03-13 23:55:56 <HM2> C++11*
4588 2013-03-13 23:56:36 <sipa> i know
4589 2013-03-13 23:56:47 <sipa> but we're not on C++11 (yet) :(
4590 2013-03-13 23:56:49 <TD> sipa: yeah
4591 2013-03-13 23:56:53 <Scrat> gmaxwell: im a UX design guy. distill it down for me
4592 2013-03-13 23:56:59 <gmaxwell> Scrat: With an application it would be indistiguishable from using a serviceâ better, the app could check every time you make a transaction and mix opportunistically when there are available partners for the right involved values.
4593 2013-03-13 23:57:15 <TD> sipa: actually, CDataStream would need to have its erase(begin(),end()) calls replaced with that
4594 2013-03-13 23:57:29 <gmaxwell> Scrat: UX design? pftt. Has your mother disowned you? :P
4595 2013-03-13 23:57:33 <sipa> TD: vSend should be a linked list of CNetMessage's
4596 2013-03-13 23:58:27 <HM2> vectors are usually fine
4597 2013-03-13 23:58:38 bolapara has joined
4598 2013-03-13 23:58:44 <MC-Droid> she just considers him imaginary. shes a mathermatician.
4599 2013-03-13 23:58:44 rcorreia has quit (Read error: Connection reset by peer)
4600 2013-03-13 23:58:58 <TD> sipa: yeah. it's a weird design.
4601 2013-03-13 23:59:12 <TD> sipa: i suppose it works that way because send() can return any arbitrary number of bytes
4602 2013-03-13 23:59:37 <sipa> TD: if you feel like looking at Jeff's patches... i've already spent hours on trying to find the segfault it causes
4603 2013-03-13 23:59:44 <TD> Scrat: i think wallets will do that kind of thing automatically in future
4604 2013-03-13 23:59:47 <TD> sipa: where are they?