1 2011-06-03 00:04:46 <midnightmagic> ;;bc,mtgox
  2 2011-06-03 00:04:47 <gribble> {"ticker":{"high":10.999,"low":9.5,"vol":41952,"buy":10.74,"sell":10.87,"last":10.87}}
  3 2011-06-03 00:13:27 <jrmithdobbs> phantomcircuit: bah, i see what you were saying about miniupnpc ... a whole damned header file is gone
  4 2011-06-03 00:14:46 <phantomcircuit> lol
  5 2011-06-03 00:14:53 <jrmithdobbs> several of them
  6 2011-06-03 00:14:55 <jrmithdobbs> actually
  7 2011-06-03 00:14:57 <jrmithdobbs> wtf
  8 2011-06-03 00:14:58 <phantomcircuit> and their version identifier is fucking 1.5
  9 2011-06-03 00:15:02 <phantomcircuit> "1.5"
 10 2011-06-03 00:15:13 <jgarzik> phantomcircuit: that non-blocking bit flip work on Windows also?  I haven't bothered to check.
 11 2011-06-03 00:15:30 <phantomcircuit> jgarzik, windows uses ioctlsocket
 12 2011-06-03 00:15:32 <jrmithdobbs> ya and macports doesn't have a damned package for 1.5 proper just 1.5.20110527
 13 2011-06-03 00:15:39 <phantomcircuit> also the non blocking connect works perfectly on windows
 14 2011-06-03 00:15:51 <phantomcircuit> i built win32 binaries and have tested them
 15 2011-06-03 00:15:59 <phantomcircuit> you get 8 connections within seconds
 16 2011-06-03 00:16:17 <phantomcircuit> also enabled dnsseed by default since the irc server was randomly dropping connections
 17 2011-06-03 00:16:54 <jrmithdobbs> i think miniupnc's project goal is to make building bitcoin annoying
 18 2011-06-03 00:17:10 <jrmithdobbs> miniupnpc's
 19 2011-06-03 00:17:11 <phantomcircuit> lol
 20 2011-06-03 00:17:24 <phantomcircuit> jgarzik, http://bitcoinconsultancy.com/bitcoin-windows.zip
 21 2011-06-03 00:17:29 <jrmithdobbs> not only are includes gone their *own includes* don't look for each other in the right damned place
 22 2011-06-03 00:17:36 <jrmithdobbs> wtf is this crap
 23 2011-06-03 00:17:45 <phantomcircuit> jrmithdobbs, lol
 24 2011-06-03 00:18:15 <jrmithdobbs> /opt/local/include/miniupnpc/upnpcommands.h:11:30: error: portlistingparse.h: No such file or directory
 25 2011-06-03 00:18:18 <jrmithdobbs> /opt/local/include/miniupnpc/upnpcommands.h:13:28: error: miniupnpctypes.h: No such file or directory
 26 2011-06-03 00:18:19 <phantomcircuit> lol
 27 2011-06-03 00:18:21 <jrmithdobbs> LOL
 28 2011-06-03 00:18:27 <jrmithdobbs> (caps intentional)
 29 2011-06-03 00:18:29 <phantomcircuit> jgarzik, also built with upnp support
 30 2011-06-03 00:19:55 <jrmithdobbs> vi /opt/local/include/miniupnpc/upnpcommands.h
 31 2011-06-03 00:19:58 <jrmithdobbs> err
 32 2011-06-03 00:22:41 <jrmithdobbs> oh nm, this is a fuckup on the port maintainers part
 33 2011-06-03 00:22:59 <jrmithdobbs> he didn't add a bunch of split out includes they added
 34 2011-06-03 00:30:33 <jrmithdobbs> nope, it's the install target in the Makefile that's fucked, blargh
 35 2011-06-03 00:50:41 <cacheson> is there anyway to put a cap on how many connections the bitcoin client can make?
 36 2011-06-03 00:51:46 <jrmithdobbs> cacheson: yes, build limit of 8 outbound -maxconnections for max number of inbound (-8 so it needs to be at least 8)
 37 2011-06-03 00:53:31 <cacheson> I keep ending up with 300-400 connections... I think it's saturating my upstream bandwidth a bit
 38 2011-06-03 00:53:37 <cacheson> I'm still using 0.3.19 since it doesn't enforce transaction fees
 39 2011-06-03 00:54:35 <jrmithdobbs> cacheson: you might as well upgrade since most miners have upgraded at this point and you'll soon stop being able to send those free txns anyways
 40 2011-06-03 00:55:05 <cacheson> they all seem to get through pretty quick, but alright
 41 2011-06-03 00:59:32 <phantomcircuit> jrmithdobbs, lol im consistently getting 8 connections within 30 seconds on linux
 42 2011-06-03 01:00:28 <theymos> I am trying to fix something on the forum. Please click this link and tell me if "unknown" is displayed instead of your IP address: https://forum.bitcoin.org/test.php (also try the non-HTTPS version http://forum.bitcoin.org/test.php ).
 43 2011-06-03 01:01:01 <tazjin> I get unknown on the non-https one
 44 2011-06-03 01:01:20 <theymos> Interesting. Do you get "unknown" twice?
 45 2011-06-03 01:01:55 <YwY> the same
 46 2011-06-03 01:01:56 <tcatm> https: IP twice, http: 1x unknown
 47 2011-06-03 01:02:10 <YwY> one unknown on the non-https page
 48 2011-06-03 01:02:14 <Graet> <tcatm> https: IP twice, http: 1x unknown  << same
 49 2011-06-03 01:02:17 <tazjin> Always get unknown on the non https one
 50 2011-06-03 01:02:30 <tazjin> and always my IP on the https one
 51 2011-06-03 01:03:14 <phantomcircuit> theymos, you run the forums?
 52 2011-06-03 01:03:26 <phantomcircuit> ill file that info away somewhere
 53 2011-06-03 01:04:00 <theymos> Sirius hosts the server. He gave me ssh access because he didn't have time to fix these things.
 54 2011-06-03 01:04:05 <tazjin> same thing using a VPN in a different country btw
 55 2011-06-03 01:05:19 <phantomcircuit> theymos, apache?
 56 2011-06-03 01:05:40 <theymos> No, it started when the forum was moved from Apache to lighttpd.
 57 2011-06-03 01:06:56 <theymos> Ah, I figured out how to reproduce it. Accessing the non-https version through a web proxy does it.
 58 2011-06-03 01:07:35 <phantomcircuit> theymos, lighttpd doesn't handle ssl all that well, usually it's used as a front end proxy for http hosts so it tends to drop stuff
 59 2011-06-03 01:08:51 <jarly> hi
 60 2011-06-03 01:08:59 <jarly> have any of you heard of oclhashcat?
 61 2011-06-03 01:09:26 <jarly> it looks like it would be a ridiculously fast sha256 cracker
 62 2011-06-03 01:09:40 <_Netsni3> http://www.youtube.com/watch?v=yYTqvYqXRbY <-- this is awesome
 63 2011-06-03 01:09:45 <jarly> even though we're not looking at cracking, it seems more efficient than current GPU miners
 64 2011-06-03 01:11:21 <phantomcircuit> jarly, looks like it's basically a hash cracker
 65 2011-06-03 01:11:30 <jarly> yeah
 66 2011-06-03 01:11:42 <phantomcircuit> so it's unliekly to be faster than the specialized bitcoin miners
 67 2011-06-03 01:12:29 <jarly> isn't generating a hash similar to hash cracking except you just want one that has a couple bits = 0?
 68 2011-06-03 01:12:37 <jarly> *the first couple
 69 2011-06-03 01:13:18 <phantomcircuit> yes it's identical
 70 2011-06-03 01:13:32 <phantomcircuit> except that's a generic hash cracker supporting multiple algorithms
 71 2011-06-03 01:13:38 <phantomcircuit> bitcoin miners are specialized
 72 2011-06-03 01:13:44 <phantomcircuit> which do you think is faster? :P
 73 2011-06-03 01:14:06 <jarly> well, how are they specialized? they're just sha-256 hashes with additional bytes attached, right?
 74 2011-06-03 01:14:16 <jarly> the block headers, that is
 75 2011-06-03 01:14:41 <mrb_> actually a good hash cracker should have a sha256 implementation just as optimized as a bitcoin miner
 76 2011-06-03 01:14:45 <phantomcircuit> jarly, the opencl code is specially tuned specifically for running sha256(sha256()) on partially computed sha256 rounds
 77 2011-06-03 01:14:51 <mrb_> I know because I have written both kinds of software
 78 2011-06-03 01:15:08 <mrb_> in fact gat3way has adapted his hash cracker into a bitcoin miner
 79 2011-06-03 01:15:41 <mrb_> another example of someone having written both kinds of sw
 80 2011-06-03 01:15:48 <phantomcircuit> sure
 81 2011-06-03 01:15:52 <phantomcircuit> but as you just said
 82 2011-06-03 01:15:55 <phantomcircuit> he had to adapt it
 83 2011-06-03 01:16:05 <mrb_> of course
 84 2011-06-03 01:16:29 <theymos> Alright, I fixed it! lighttpd was returning an IPv6-encoded IP address for non-https requests for some reason. Thanks for your help.
 85 2011-06-03 01:19:48 <jarly> mrb_: that's linux only, though, isn't it?
 86 2011-06-03 01:21:22 <zooko> Hey folks, I looked at http://maps.google.com/maps?q=https://smsz.net/btcStats/bitcoin.kml and it appeared there were a mere 1277 nodes connected at the time I looked.
 87 2011-06-03 01:21:27 <zooko> Is that the right order of magnitude?//
 88 2011-06-03 01:22:24 <jarly> damn. apparently the oclhashcat dev tweaked it to use BFI_INT and he got 1.27 gh/s with a single 5970
 89 2011-06-03 01:22:39 <jarly> if we could only channel his program for bitcoin mining..
 90 2011-06-03 01:22:49 <tcatm> jarly: BFI_INT is possible with recent miners
 91 2011-06-03 01:23:13 <jarly> yeah, and it does make a difference
 92 2011-06-03 01:23:34 <jarly> but it looks like that guy's hasher is running much, much faster than what bitcoin miners are using
 93 2011-06-03 01:24:37 <tcatm> jarly: my 5970 do 1248 Mh/s and I don't overclock much.
 94 2011-06-03 01:24:57 <jarly> with a bitcoin miner?
 95 2011-06-03 01:25:43 <jarly> ...what are your clocks?
 96 2011-06-03 01:26:21 <tcatm> jarly: 736 Mhz on that card
 97 2011-06-03 01:27:07 <jarly> wow, that's fast
 98 2011-06-03 01:27:25 <jarly> the 5970s on the mining hardware comparison seem to only hit 800
 99 2011-06-03 01:28:53 <tcatm> jarly: also remember, that bitcoin hashes are 2x sha256
100 2011-06-03 01:29:01 <jarly> ah, true
101 2011-06-03 01:29:11 <jarly> ...so they're actually about equal then
102 2011-06-03 01:29:39 <tcatm> 800*2 = 1.6gh/s
103 2011-06-03 01:30:25 <jarly> that card had a clock speed of 850, though, so it's not a fair comparison
104 2011-06-03 01:30:53 <tcatm> use (Mh/s)/MHz to compare them
105 2011-06-03 01:31:01 <jarly> the others get around 600, so it looks like they're equal
106 2011-06-03 01:31:47 <zooko> So, how many nodes do you think are currently replicating the bitcoin block chain?
107 2011-06-03 01:32:25 <tcatm> zooko: my "supernode" is connected to about 800 nodes
108 2011-06-03 01:33:09 <tcatm> i.e. it tries to connect to as many nodes as possible, asking for new addresses each time it connects to a new node
109 2011-06-03 01:34:58 <zooko> tcatm: ah, good data point, thanks.
110 2011-06-03 01:35:27 <tcatm> btw, this is the result of parsing data from all these nodes: http://eu1.bitcoincharts.com/map/
111 2011-06-03 01:35:35 <phantomcircuit> tcatm, modified bitcoind or custom?
112 2011-06-03 01:35:48 <tcatm> phantomcircuit: my python client on asyncore
113 2011-06-03 01:35:53 <phantomcircuit> ah
114 2011-06-03 01:36:12 <jrmithdobbs> phantomcircuit: i had bitcoind modified for a while that'd do that too
115 2011-06-03 01:36:29 <jrmithdobbs> phantomcircuit: just bump outbound connections to ~512 and maxconnections to 1024
116 2011-06-03 01:36:33 <jrmithdobbs> by default
117 2011-06-03 01:36:49 <jrmithdobbs> phantomcircuit: that'll still follow the 1 per class b rule but you'll end up with ~300-400 connections or so
118 2011-06-03 01:37:08 <zooko> tcatm: cool!
119 2011-06-03 01:37:17 <zooko> tcatm: I'm a fan of python and twisted.
120 2011-06-03 01:37:21 <phantomcircuit> well im cool with 8 since the router here keeps crashing under the load of 500KB/s over wifi
121 2011-06-03 01:37:22 <phantomcircuit> sigh
122 2011-06-03 01:37:35 <jrmithdobbs> haha
123 2011-06-03 01:37:56 <jrmithdobbs> ya that's on a box on a 100mbit (burstable, and actually really is) pipe
124 2011-06-03 01:38:05 <jrmithdobbs> i don't have to worry about that
125 2011-06-03 01:38:14 <jarly> tcatm: where should I start if I want to learn more about GPGPU?
126 2011-06-03 01:38:18 <scott`> ;;bc,stats
127 2011-06-03 01:38:20 <gribble> Current Blocks: 128273 | Current Difficulty: 434882.7217497 | Next Difficulty At Block: 129023 | Next Difficulty In: 750 blocks | Next Difficulty In About: 3 days, 15 hours, 55 minutes, and 0 seconds | Next Difficulty Estimate: 519144.38304394
128 2011-06-03 01:38:40 <jarly> i mean, the very basics. assembly?
129 2011-06-03 01:39:10 <Clarence> bitcoin-bears please come out of closet... is there any systemic risk inherent in holding bitcoins, or is it as persistent as, say gold is?
130 2011-06-03 01:39:14 <phantomcircuit> jrmithdobbs, lol yeah i have access to a gbps box but it's got cpanel and shit isntalled on it so i just dont bother
131 2011-06-03 01:39:20 <tcatm> jarly: read all PDFs ATI wrote about OpenCL and Stream :)
132 2011-06-03 01:39:29 <jarly> hnnnnnnnnnnnnnnnng
133 2011-06-03 01:39:31 <jarly> alrighty
134 2011-06-03 01:40:05 <jarly> i'm a pythonista, so this is gonna be torture...
135 2011-06-03 01:40:13 <jarly> but i guess it's worth it, lol
136 2011-06-03 01:40:23 <tcatm> I've had to read them multiple times to write my first miner. GPGPU architectures can be quite confusing if you have no experience with parallel algorithms
137 2011-06-03 01:41:16 <jarly> is there some kind of "Hello World" program you'd suggest?
138 2011-06-03 01:41:28 <tcatm> the ATI SDK has lots of examples
139 2011-06-03 01:56:58 <cornfeed> is there a way to map the nodes connected to the network
140 2011-06-03 01:57:00 <cornfeed> ?
141 2011-06-03 01:59:01 <jrmithdobbs> cool think i've got a working macports build + instructions + makefile + options to toggle i386+x86_64/i386only/x86_64only (needed that since my macports only build x86_64 ;P)
142 2011-06-03 02:04:56 <jrmithdobbs> bleh, mostly working, stupid wx-config in the 2.9 broken and spits out build paths instead of installed paths for the libraries, lol
143 2011-06-03 02:12:58 <Dark_Apostrophe> Hello, is it possible to see the bandwidth usage by the Bitcoin client within it? Upload and download bandwidth. Possibly also latency to the fastest peer... If not, could I request it as a UI feature? It can be configurable, so users wouldn't have to see it all the time.
144 2011-06-03 02:14:51 <lfm> Dark_Apostrophe: anything just about is possible, I dont know of any developers who would be interested in doing that for ya tho. maybe if there was a bounty
145 2011-06-03 02:15:10 <Dark_Apostrophe> I see
146 2011-06-03 02:15:45 <Dark_Apostrophe> Btw, someone was working on a Qt UI, right?
147 2011-06-03 02:16:01 <lfm> your OS may have some tools to see network traffic
148 2011-06-03 02:16:09 <Dark_Apostrophe> How's that progressing, and is it planned for inclusion in the main client at some point?
149 2011-06-03 02:16:43 <lfm> I havnt heard about that qt thing so ...?
150 2011-06-03 02:17:36 <Dark_Apostrophe> saw forum mention of it at some point.. hold on, I'll look around
151 2011-06-03 02:18:10 <lfm> there is that stand alone gui thingy someone made, I havnt looked a it myself tho
152 2011-06-03 02:18:53 <Dark_Apostrophe> http://forum.bitcoin.org/index.php?topic=2022.20
153 2011-06-03 02:22:40 <lfm> Dark_Apostrophe: I dont think it will be combined into the main client, it will stay separate
154 2011-06-03 02:22:56 <Dark_Apostrophe> Damn, okay
155 2011-06-03 02:23:17 <Dark_Apostrophe> The UI just looked so good - Good layout + Qt is generally good looking
156 2011-06-03 02:23:39 <zooko> Clarence: are you kidding? There is definitely systemic risk.
157 2011-06-03 02:23:41 <lfm> well use it then
158 2011-06-03 02:23:59 <Dark_Apostrophe> will see how once it gets more complete
159 2011-06-03 02:24:15 <Clarence> zooko, could you explain?
160 2011-06-03 02:25:10 <Clarence> or give me a link
161 2011-06-03 02:25:42 <lfm> Clarence: bitcoin prices are quite voletile
162 2011-06-03 02:25:45 <lfm> mtgox.com
163 2011-06-03 02:25:49 <Clarence> ya, but that's not the risk I meant
164 2011-06-03 02:26:38 <lfm> what do you mean by systemic then?
165 2011-06-03 02:26:38 <zooko> Clarence: it is a new and untested idea.
166 2011-06-03 02:27:17 <lfm> and bitcoin seems to be following a classic bubble path
167 2011-06-03 02:27:33 <zooko> Clarence: there are very few (I think) people who both understand how it works and know how to break such things.
168 2011-06-03 02:28:03 <lfm> zooko: more every day I think
169 2011-06-03 02:28:30 <zooko> Yes, but more every day isn't sufficient to give me confidence in little systemic risk.
170 2011-06-03 02:29:02 <lfm> yes more that understand btc and more that have ideas to break it
171 2011-06-03 02:29:44 <zooko> I guess it is a matter of how much you are willing to gamble and what your tolerance for risk is.
172 2011-06-03 02:30:27 <dirtyfilthy> the greatest risk imho is mtgox gets shutdown
173 2011-06-03 02:30:56 <lfm> dirtyfilthy: I suspect that would only be a small bump in the road
174 2011-06-03 02:31:21 <dirtyfilthy> it depends, that's a lot of money/bitcoin down the drain
175 2011-06-03 02:31:38 <zooko> It MtGox holding a lot of USD and/or BTC on behalf of its customers?
176 2011-06-03 02:31:41 <zooko> s/It/Is/
177 2011-06-03 02:31:49 <dirtyfilthy> a huge amount
178 2011-06-03 02:31:52 <lfm> but I think other traders would pop up pretty quick to fill in the void
179 2011-06-03 02:32:53 <lfm> zooko: he has none of mine atm so I am philosphical about it
180 2011-06-03 02:32:54 <zooko> Do we know how much?
181 2011-06-03 02:34:45 <lfm> zooko: you can see a sort of minimum amount in the "depth of market" graph. it coulbe a lot more than that tho
182 2011-06-03 02:35:06 <dirtyfilthy> volume in the 24 hours was 41607 * say ten usd a bitcoin (avg price) = 416,070 USD went through the market today
183 2011-06-03 02:35:27 <dirtyfilthy> the amount actually held must be many times that
184 2011-06-03 02:35:53 <lfm> dirtyfilthy: ya but that could be very transient money. in and out very quickly
185 2011-06-03 02:37:24 <zooko> Hm.
186 2011-06-03 02:37:35 <zooko> But still, he could probably make off with something on that order.
187 2011-06-03 02:40:47 <lfm> see all the "bids" on the "depth of market page? multiply the prize by thet amount and add them up, that is dollars he is holding. the for the asks do the same, that is btc he is holding
188 2011-06-03 02:41:08 <lfm> prize -> price
189 2011-06-03 02:41:44 <Clarence> ya, and there's probably 5 times that amount not in orders
190 2011-06-03 02:42:09 <lfm> and then there is the "dark orders"
191 2011-06-03 02:42:53 <Xenland> Backburn? you there mate?
192 2011-06-03 02:47:57 <zooko> Wow. What temptation!
193 2011-06-03 02:48:19 <zooko> But, I actually am thinking about deeper, technical systemic risks.
194 2011-06-03 02:48:55 <zooko> My current thesis is that no matter what spectacular disasters or market crashes were to hit BitCoin, it might (*might*) continue to be useful and used as a currency.
195 2011-06-03 02:49:49 <zooko> So, for example, mtgox absconding with everyone's money, triggering a return to prices of $0.001 == ??1.00. :-)
196 2011-06-03 02:50:17 <zooko> Maybe something like that could happen, and a critical mass of people would continue to rely on BitCoin to buy their coffee every week.
197 2011-06-03 02:50:41 <lfm> zooko: well I spoze there is a tiny chance of a cryptographic breakthru that would make it very easy to make hashes of whatever pattern you want. then someone could start mining and collect huge payoffs (in btc) from mining which would prolly make the value collapse
198 2011-06-03 02:50:46 <zooko> Assuming they had, um, about ??10,000 to spend on coffee.
199 2011-06-03 02:51:14 <Clarence> what about persistent DOS attacks?
200 2011-06-03 02:51:14 <zooko> lfm: yeah, see that's the kind of "deeper" systemic risk I'm talking about. Something that makes it so that Alice and Bob cannot reliably exchange BitCoin payments at all, even if they both want to.
201 2011-06-03 02:51:35 <zooko> Clarence: interesting! I haven't thought about it too much.
202 2011-06-03 02:51:37 <zooko> Could be a problem.
203 2011-06-03 02:52:13 <lfm> well ddos attacks against a distributed p2p net are kinda hard since youd need to ddos thousands of ports
204 2011-06-03 02:52:14 <Clarence> i've just seen it mentioned
205 2011-06-03 02:53:08 <lfm> the pools do make a kinda centrallized target for ddos attacks I spoze but thats not really any permanent damage to bitcoin net itself really
206 2011-06-03 02:58:03 <lfm> if you have intimate knowledge of the connections that nodes have you can isolate and manipulate nodes or small groups of nodes so they are fooled into beleiving things which are not true and get overturned when the net rejoins
207 2011-06-03 02:59:18 <lfm> like if your isp was a bitcoin expert he could really screw around with you
208 2011-06-03 03:06:27 <Clarence> I think a major threat would could be competing digital commodities
209 2011-06-03 03:06:47 <cacheson> lfm: what could they actually do, assuming you're not running a "lite" client?
210 2011-06-03 03:06:56 <lfm> Clarence: there are several competing digital commodities already
211 2011-06-03 03:07:01 <cacheson> if they try to feed you fake transaction data, it'll fail verification
212 2011-06-03 03:07:23 <Clarence> lfm, names please :)
213 2011-06-03 03:07:37 <lfm> cacheson: if they isolate you from the main net I think they can double spend
214 2011-06-03 03:08:08 <lfm> Liberty Reserve, pecunix and others
215 2011-06-03 03:08:18 <cacheson> lfm: they'd have to have a significant mining operation to back it up, wouldn't they?  if you're isolated, there's no miners
216 2011-06-03 03:08:36 <cacheson> and without miners no transactions get processed
217 2011-06-03 03:08:38 <lfm> cacheson: ya prolly, i guess so
218 2011-06-03 03:09:37 <cacheson> maybe there are some attacks they could pull off, but nothing really seems obvious to me
219 2011-06-03 03:10:20 <lfm> cacheson: and if you are vigilant you will notice that most of the world has dissapeared
220 2011-06-03 03:10:38 <cacheson> I could see them taking advantage of isolated miners
221 2011-06-03 03:10:42 <cacheson> since those aren't full clients
222 2011-06-03 03:11:16 <lfm> cacheson: ya, there are I think many ways that pool managers can take eadvantage of miners
223 2011-06-03 03:11:36 <cacheson> an ISP could do it too, unless the pool was using SSL
224 2011-06-03 03:12:43 <lfm> maybe ya, depends how the pool is set up too
225 2011-06-03 03:14:57 <Clarence> could it be said bitcoin has first mover advantage among systems that don't have to 'follow the rules'?
226 2011-06-03 03:15:25 <lfm> not sure what that means
227 2011-06-03 03:16:23 <Clarence> no irs reporting, psudo-anonimity ect
228 2011-06-03 03:16:24 <zooko> Someone could definitely say that if they wanted to.
229 2011-06-03 03:16:45 <cacheson> it is the first decentralized electronic currency that's secure against double-spending
230 2011-06-03 03:16:46 <jgarzik> Clarence: bitcoins do not alleviate your IRS burdens
231 2011-06-03 03:16:52 <lfm> and on irc they prolly will say it sooner or later
232 2011-06-03 03:17:04 <cacheson> other e-currencies are centralized, so you have to trust the issuer
233 2011-06-03 03:17:04 <jgarzik> Clarence: I take a -huge- IRS hit due to bitcoins this year
234 2011-06-03 03:18:03 <kreal-> Can anyone recomment a cheap sms provider?
235 2011-06-03 03:18:44 <lfm> sms just isnt cheap as I understand it
236 2011-06-03 03:19:27 <kreal-> yup
237 2011-06-03 03:19:42 <kreal-> dont want sms notifications to ruin my pool.
238 2011-06-03 03:19:59 <jgarzik> https://smsz.net/ - 0.5 BTC per SMS
239 2011-06-03 03:20:06 <jgarzik> though I bet MT would lower price
240 2011-06-03 03:20:06 <Raccoon> jgarzik: in what regard
241 2011-06-03 03:20:09 <gjs278> rihahaha
242 2011-06-03 03:20:17 <cacheson> heh
243 2011-06-03 03:20:18 <Raccoon> jgarzik: are you claiming bitcoins you have mined?
244 2011-06-03 03:20:22 <kreal-> uh nice jgarzik
245 2011-06-03 03:20:24 <lfm> like there seems no reasonable way that sms messages should cost more that email messages but they do
246 2011-06-03 03:20:29 <cacheson> nice, $5 per SMS
247 2011-06-03 03:20:42 <jgarzik> <jgarzik> though I bet MT would lower price
248 2011-06-03 03:20:53 <cacheson> lfm: it's one of those "what the market will bear" situations
249 2011-06-03 03:20:55 <jgarzik> MT probably hasn't updated site for latest value
250 2011-06-03 03:21:11 <Raccoon> lol wow.  bitcoins for SMS
251 2011-06-03 03:21:19 <Raccoon> thats a pretty fair deal right there
252 2011-06-03 03:21:25 <Niedar> 5$
253 2011-06-03 03:21:29 <lfm> wow 0.5 btc per sms? so like $5 per sms!
254 2011-06-03 03:21:38 <Raccoon> oh wait nvm
255 2011-06-03 03:21:46 <Raccoon> i thought as a means of buying bitcoins :P
256 2011-06-03 03:21:47 <Raccoon> lol
257 2011-06-03 03:22:23 <Raccoon> "send an SMS to 99999 with your BTC Address, and we'll charge your phone provider $5 per 0.5 BTC"
258 2011-06-03 03:22:39 <lfm> now that would be nice
259 2011-06-03 03:22:48 <Raccoon> that would be pretty slick
260 2011-06-03 03:24:30 <lfm> gotta go ... see yall
261 2011-06-03 03:25:39 <zooko> jgarzik: did you declare your mining results to the IRS?
262 2011-06-03 03:28:53 <Niedar> http://www.daopay.com/
263 2011-06-03 03:29:42 <Niedar> dunno reputation but you could use something like that
264 2011-06-03 03:29:54 <Niedar> for what you are talking about
265 2011-06-03 03:34:28 <Niedar> actually here is a list of different service providers for that type of billing
266 2011-06-03 03:34:30 <Niedar> http://www.blackhatworld.com/blackhat-seo/making-money/287848-pay-per-call-pay-per-sms-website.html
267 2011-06-03 03:50:43 <jgarzik> Bitcoin now world's most powerful supercomputer: http://forum.bitcoin.org/index.php?topic=11544.0   (edit:  please check my math and proofread... ktnx)
268 2011-06-03 03:52:42 <gjs278> yeah folding@home got raped
269 2011-06-03 03:54:14 <tommygunner> stresstest applications or benchmarks should mine bitcoins too
270 2011-06-03 03:55:33 <jgarzik> gjs278: any chance you have a link or good google string for F@H numbers?
271 2011-06-03 03:55:41 <jgarzik> my goog-fu is weak, apparently
272 2011-06-03 04:09:09 <diki> so, i ripped the part where the hexstr get's converted to a valid block so it can be confirmed, however...it posts back some jibberish
273 2011-06-03 04:09:10 <diki> http://pastebin.com/0qWH2mbM
274 2011-06-03 04:09:27 <diki> it's basically a small program i am going to use
275 2011-06-03 04:09:58 <diki> dont mind the includes, i just wasn's sure which to use
276 2011-06-03 04:12:21 <sgstair> I have to seriously question any calculation which suggests that a double sha256 takes >10000 floating point operations. For one, it doesn't. (I count approximately 5000 individual integer operations, unoptimized) For two though, they're obviously not floating point, and there's no comparison.
277 2011-06-03 04:13:21 <cacheson> using floats in your hash algorithm would make it really... interesting
278 2011-06-03 04:13:44 <diki> <cacheson> usin<- to who were you referring?
279 2011-06-03 04:13:55 <cacheson> diki: sgstair
280 2011-06-03 04:17:29 <diki> so, any ideas why the program is not working?
281 2011-06-03 04:22:01 <kish> is the public channel log searchable?
282 2011-06-03 04:22:38 <diki> use the browser search function
283 2011-06-03 04:23:25 <jgarzik> sgstair: did you read the referenced thread?
284 2011-06-03 04:23:45 <sgstair> I did not see it, looking
285 2011-06-03 04:24:14 <gjs278> how many confirms do you need to send those coins
286 2011-06-03 04:24:26 <gjs278> that you've received
287 2011-06-03 04:26:11 <jgarzik> gjs278: at least one
288 2011-06-03 04:26:24 <jgarzik> gjs278: >1 confirms give ever-greater confidence of no double spend
289 2011-06-03 04:26:27 <gjs278> alright
290 2011-06-03 04:26:33 <gjs278> 3 should be good enough then
291 2011-06-03 04:26:42 <jgarzik> yes
292 2011-06-03 04:30:36 <jrmithdobbs> if things are built with -DwxDEBUG_LEVEL=0 why am I still getting wxWidgets Debug Alert dialogs pop up?!
293 2011-06-03 04:31:17 <jrmithdobbs> instead of it going to stdout
294 2011-06-03 04:31:25 <jrmithdobbs> (or stderr, whatever)
295 2011-06-03 04:32:20 <diki> hey jrmith give me suggestions
296 2011-06-03 04:32:30 <jrmithdobbs> diki: give me some money
297 2011-06-03 04:36:58 <jrmithdobbs> ah, makefile.osx is missing -D__WXDEBUG__
298 2011-06-03 04:38:06 <diki> you have a mac?
299 2011-06-03 04:40:28 <jrmithdobbs> oh nm i see the problem, has to actually be turned off in wxwidgets at build time. fucking annoying. remembering why i never build the damned gui
300 2011-06-03 04:40:33 <jrmithdobbs> diki: yes?
301 2011-06-03 04:41:22 <diki> so my program does seem to work in a way, i just want it to print a sha256 hash, not some mumbo jumbo
302 2011-06-03 04:44:22 <forexmasterja> can Everybody say $11 -> 1 !!!
303 2011-06-03 04:47:28 <noagendamarket> woot
304 2011-06-03 05:40:07 <devon_hillard> ;;bc,calc 65000
305 2011-06-03 05:40:07 <gribble> The average time to generate a block at 65000 Khps, given current difficulty of 434882.7217497 , is 47 weeks, 3 days, 14 hours, 4 minutes, and 53 seconds
306 2011-06-03 05:45:48 <Raccoon> any client devs currently present?
307 2011-06-03 06:04:27 <intelliot> i'm interested in building a site like mtgox, but with a different fee structure (x bitcoins per trade). contact me if you're able to do this
308 2011-06-03 06:05:15 <da2ce7> I'm looking for a professional c++ programer...
309 2011-06-03 06:09:23 <Xenland__> are there any non-free opensource lisences
310 2011-06-03 06:11:42 <davep> depends how you define free. http://www.opensource.org/licenses/category
311 2011-06-03 06:16:34 <Xenland__> Basicly i want to sell my already opensource software, but i have a client that wants to have the option to not have to be forced to give out his modified code
312 2011-06-03 06:17:07 <Xenland__> He will pay me to have the right to keep it opensource but he will have to pay to have the right to keep it private aswell?
313 2011-06-03 06:24:55 <vegard> Xenland__: if you wrote/own all the code, you just distribute it to your client under a private license that says they're not allowed to redistribute it
314 2011-06-03 06:25:21 <Xenland__> vegard: and a few words will hold up in court?
315 2011-06-03 06:25:46 <Xenland__> Too me if your vauge its hard to prove that the offending person did anything wrong.... i may be naive but that just what i think
316 2011-06-03 06:26:04 <vegard> just sign a contract, what's the problem?
317 2011-06-03 06:26:22 <Xenland__> Well a license is a contract... soo i guess no problem
318 2011-06-03 06:26:29 <Xenland__> are you a lawyer?
319 2011-06-03 06:26:42 <Xenland__> I dont plan on suing anybody I would just like to know that i have that proctection
320 2011-06-03 06:27:11 <vegard> IANAL, developer
321 2011-06-03 06:27:52 <molecular> what's the status with wx. Do I still need wx 2.9 to build bitcoin or will it work with 2.8 with newest git version?
322 2011-06-03 06:28:26 <sipa> bluematt has a patch to make it work with 2.8
323 2011-06-03 06:29:38 <Xenland__> vegard: ISNO, grammer
324 2011-06-03 06:30:33 <cacheson> Xenland__: look up "dual licensing"
325 2011-06-03 06:30:40 <cacheson> that sounds like what you're looking for
326 2011-06-03 06:30:57 <Xenland__> Cool, I'll check that out
327 2011-06-03 06:31:27 <cacheson> Xenland__: https://secure.wikimedia.org/wikipedia/en/wiki/Multi-licensing
328 2011-06-03 06:31:39 <Xenland__> Thanks mate this is exactly what i was talking about
329 2011-06-03 06:31:47 <cacheson> sure thing  :)
330 2011-06-03 06:31:47 <Xenland__> Non-free Opensource licenses
331 2011-06-03 06:31:50 <Xenland__> :D
332 2011-06-03 06:45:37 <da2ce7> Hiring a C++ programer.
333 2011-06-03 06:45:38 <da2ce7> https://forum.bitcoin.org/index.php?topic=11573.0
334 2011-06-03 06:49:11 <eps1> ;;bc,mtgox
335 2011-06-03 06:49:12 <gribble> {"ticker":{"high":13.6,"low":9.55,"vol":50289,"buy":12.9003,"sell":13.0997,"last":12.9003}}
336 2011-06-03 06:55:45 <vegard> what does "coinbase" refer to?
337 2011-06-03 06:56:42 <vegard> ah, the transaction with the coins from block generation?
338 2011-06-03 06:56:55 <gjs278> ;;bc,stats
339 2011-06-03 06:56:57 <gribble> Current Blocks: 128327 | Current Difficulty: 434882.7217497 | Next Difficulty At Block: 129023 | Next Difficulty In: 696 blocks | Next Difficulty In About: 4 days, 1 hour, 3 minutes, and 12 seconds | Next Difficulty Estimate: 525349.41062639
340 2011-06-03 06:58:30 <gjs278> ;;bc,mtgox
341 2011-06-03 06:58:31 <gribble> {"ticker":{"high":13.6,"low":9.55,"vol":50688,"buy":12.9,"sell":13.0998,"last":13}}
342 2011-06-03 07:09:39 <Xenland> round(high) = $14
343 2011-06-03 07:09:45 <Xenland> OMG MTGOX IS $14!!!!
344 2011-06-03 07:11:24 <cacheson> no rounding allowed in here
345 2011-06-03 07:44:37 <matija> Hi to all!
346 2011-06-03 07:45:01 <matija> How can I contribute with localization?
347 2011-06-03 07:46:29 <cacheson> hi matija  :)
348 2011-06-03 07:46:45 <cacheson> none of the devs seem to be awake at the moment, maybe try back later?
349 2011-06-03 07:47:01 <cacheson> alternatively, maybe do a search on the forums
350 2011-06-03 07:47:09 <matija> cahceson, I will. Thank you! :)
351 2011-06-03 07:47:16 <cacheson> sure thing  :)
352 2011-06-03 07:59:07 <eps1> ;;bc,mtgox
353 2011-06-03 07:59:08 <gribble> {"ticker":{"high":14,"low":9.55,"vol":53244,"buy":13.85,"sell":13.9,"last":13.9}}
354 2011-06-03 08:10:06 <ordex> hello
355 2011-06-03 08:10:34 <ordex> I have problems while running minerd with bitcoind. is this the right place to ask? I just want to understand what cause the error
356 2011-06-03 08:10:48 <PirateMarmalade> There is also #bitcoin-mining
357 2011-06-03 08:11:03 <PirateMarmalade> If the problem is mining specific
358 2011-06-03 08:11:08 <ordex> not really
359 2011-06-03 08:11:23 <ordex> I just got this message while running minerd: HTTP request failed: The requested URL returned error: 500
360 2011-06-03 08:11:27 <Phoebus> wtb new bitcoin version with less transaction fees sooner :P considering the new heights of exhange rates...
361 2011-06-03 08:11:56 <ordex> PirateMarmalade: I'm tring to understand what causes the 500 error
362 2011-06-03 08:12:43 <PirateMarmalade> 500 is internal server error
363 2011-06-03 08:12:59 <PirateMarmalade> Not sure under what conditions bitcoind returns that
364 2011-06-03 08:13:10 <ordex> mhhh
365 2011-06-03 08:13:11 <PirateMarmalade> (I don't user minerd either)
366 2011-06-03 08:13:13 <ordex> so it is bitcoind fault
367 2011-06-03 08:13:19 <ordex> not minerd
368 2011-06-03 08:13:22 <ordex> right?
369 2011-06-03 08:13:25 <jaromil> goodmorning
370 2011-06-03 08:13:29 <PirateMarmalade> Might just be the way minerd is configured
371 2011-06-03 08:13:45 <jaromil> FYI a good reason to rethink about strrpc pull req that genjix did
372 2011-06-03 08:13:46 <jaromil> http://blog.programmableweb.com/2010/10/19/the-twitter-id-shuffle-text-vs-numbers/
373 2011-06-03 08:13:51 <ordex> mh I'm only giving it the user and the pass
374 2011-06-03 08:15:49 <PirateMarmalade> Can you maybe post the configuration you are using (u/n and pass obscured of course)
375 2011-06-03 08:16:50 <ordex> I'm tring to directly connect to 127.0.0.1:8332 with my browser and after loggin in I'm getting this error: {"result":null,"error":{"code":-32700,"message":"Parse error"},"id":null}
376 2011-06-03 08:16:59 <ordex> PirateMarmalade: for bitcoind?
377 2011-06-03 08:17:05 <PirateMarmalade> For minerd
378 2011-06-03 08:17:13 <PirateMarmalade> whichever flags you are using
379 2011-06-03 08:17:44 <ordex> ah I just passed --userpass user:pass, nothing more
380 2011-06-03 08:18:37 <ordex> should I use a config?
381 2011-06-03 08:18:46 <PirateMarmalade> that should be fine
382 2011-06-03 08:18:59 <PirateMarmalade> try use --url maybe
383 2011-06-03 08:19:09 <PirateMarmalade> then again i dont know minerd
384 2011-06-03 08:19:55 <ordex> PirateMarmalade: ok thanks, but even with --url it doesn't work. is the error I pasted before normal?
385 2011-06-03 08:20:33 <PirateMarmalade> I would test it, but no bitcoin on the pc im on nao
386 2011-06-03 08:21:03 <ordex> ah ok
387 2011-06-03 08:22:21 <edcba> someone should email that adam cohen ask him if he still abide to what he said on quora and if yes make a 1 year longbet with him to make things factuals
388 2011-06-03 08:23:23 <edcba> oh wait there is a follow up from 20 hours ago...
389 2011-06-03 08:26:08 <edcba> http://news.ycombinator.com/item?id=2612237
390 2011-06-03 08:26:19 <edcba> seems some better conclusions
391 2011-06-03 08:34:44 <doublec> jgarzik getting called out in a hacker new comment for 'marketing'
392 2011-06-03 08:35:01 <doublec> "or example, in a Gawker story about drugs, Jeff Garzik, a core developer, inaccurately commented that Bitcoin could not be used to purchase anything anonymously. He surely knows that it wouldn't be difficult to do so; he appears simply to have been underplaying a regulatory risk that the currency faces in order to get part of the public on his side."
393 2011-06-03 08:35:11 <doublec> http://news.ycombinator.com/item?id=2614896
394 2011-06-03 08:35:33 <tommygunner> everybody and their mother is trying to make their journalistic bones with bitcoin
395 2011-06-03 08:35:51 <doublec> antiscam is very antibitcoin in that thread
396 2011-06-03 08:46:03 <rlifchitz> ;;bc,stats
397 2011-06-03 08:46:13 <gribble> Current Blocks: 128334 | Current Difficulty: 434882.7217497 | Next Difficulty At Block: 129023 | Next Difficulty In: 689 blocks | Next Difficulty In About: 4 days, 0 hours, 16 minutes, and 7 seconds | Next Difficulty Estimate: 523307.07372643
398 2011-06-03 08:46:37 <tommygunner> ;;bc,gen 292000
399 2011-06-03 08:46:46 <andrew12> jrmithdobbs: some people may not understand that right off the bat
400 2011-06-03 08:47:20 <tommygunner> ;;bc,gend 292000 523307.07372643
401 2011-06-03 08:47:42 <tommygunner> talk to me gribble :/
402 2011-06-03 08:47:52 <andrew12> lol
403 2011-06-03 08:48:01 <andrew12> it should pm you
404 2011-06-03 08:48:07 <BlueMatt> gribble down yet again?
405 2011-06-03 08:48:16 <tommygunner> if i triggered spam?
406 2011-06-03 08:48:25 <tommygunner> someone else just try it pls
407 2011-06-03 08:48:34 <BlueMatt> ;;bc,mtgox
408 2011-06-03 08:48:43 <tommygunner> zing
409 2011-06-03 08:48:48 <gribble> The expected generation output, at 292000 Khps, given current difficulty of 434882.7217497 , is 0.675358832259 BTC per day and 0.0281399513441 BTC per hour.
410 2011-06-03 08:48:49 <gribble> The expected generation output, at 292000 Khps, given the supplied difficulty of 523307.07372643, is 0.561241958836 BTC per day and 0.0233850816182 BTC per hour.
411 2011-06-03 08:48:56 <tommygunner> oh hai gribble
412 2011-06-03 08:49:07 <BlueMatt> laggedy lag
413 2011-06-03 08:49:42 <gribble> {"ticker":{"high":14,"low":9.55,"vol":57508,"buy":12.5799,"sell":12.58,"last":12.58}}
414 2011-06-03 09:00:36 <wumpus> jaromil: yes, it'd be better to support strings in the json, so that no conversion to double is needed when you want to represent coins as decimal or integer type
415 2011-06-03 09:01:22 <wumpus> using doubles for financial stuff is a big nono
416 2011-06-03 09:01:55 <andrew12> is there any sort of feed, i.e. telnet, somewhere for the block and tx data that #bitcoin-watch has?
417 2011-06-03 09:04:11 <xata> Hi, I have a problem - i am using ubuntu 11.04 with fluxbox. Bitcoin starts minimized in tray and i can not make it show by pressing left/right buttons/ How can i now find out number of my wallet? I am new to this.
418 2011-06-03 09:07:50 <BlueMatt> xata: use 0.3.22 rc
419 2011-06-03 09:07:52 <BlueMatt> bug in wx
420 2011-06-03 09:08:41 <xata> BlueMatt, it is in git/somewhere else?
421 2011-06-03 09:08:56 <BlueMatt> you can grab builds from sf/the forums, or checkout and built git head yourself
422 2011-06-03 09:09:12 <BlueMatt> well checkout and build any bitcoin version with wx-trunk that is
423 2011-06-03 09:10:01 <xata> BlueMatt, so i need trunk of wx or of bitcoin?
424 2011-06-03 09:10:31 <BlueMatt> no, its exclusively a wx issue
425 2011-06-03 09:10:34 <BlueMatt> any bitcoin and wx trunk
426 2011-06-03 09:10:55 <BlueMatt> or a 0.3.22 build which are built against wx 2.9.1 with the relevant patches
427 2011-06-03 09:11:17 <matija> hi again
428 2011-06-03 09:12:20 <xata> BlueMatt, oh, thanks. i better go build bitcoin - it is easier and will not versions in system.
429 2011-06-03 09:13:20 <BlueMatt> yea...you have to build all of wx first, no distros have 2.9 in their repos
430 2011-06-03 09:13:24 <BlueMatt> much easier to dl bitcoin
431 2011-06-03 09:13:32 <BlueMatt> (binary)
432 2011-06-03 09:14:00 <ordex> why do I get this on get work: error: {"code":-9,"message":"Bitcoin is not connected!"}
433 2011-06-03 09:14:04 <ordex> ?
434 2011-06-03 09:14:16 <BlueMatt> because bitcoin is not connected?
435 2011-06-03 09:14:29 <ordex> what should I do to connect it?
436 2011-06-03 09:14:39 <ordex> I didn't find any suitable guide for this
437 2011-06-03 09:14:55 <BlueMatt> its slow, give it time
438 2011-06-03 09:15:14 <ordex> ah so it doesn't depend on me or on my configuration?
439 2011-06-03 09:15:27 <ordex> oh, now I'm downloading blocks
440 2011-06-03 09:15:41 <BlueMatt> well it will speed up significantly if you open port 8333 to your bitcoin node
441 2011-06-03 09:16:01 <ordex> BlueMatt: another question, why does minerd get  HTTP request failed: The requested URL returned error 500  ?
442 2011-06-03 09:16:15 <ordex> BlueMatt: I'm behind a natted MAN..cannot do more
443 2011-06-03 09:16:26 <BlueMatt> because the requested url returned a 500???
444 2011-06-03 09:16:55 <ordex> BlueMatt: the requested url is http://127.0.0.1:8332 I don't know why it should return 500
445 2011-06-03 09:17:01 <ordex> this is why I'm asking :/
446 2011-06-03 09:17:51 <ordex> weird
447 2011-06-03 09:18:20 <BlueMatt> nope something else on your computer is doing that, not bitcoin
448 2011-06-03 09:18:27 <BlueMatt> bitcoin never returns 500, there is no 500 in its code
449 2011-06-03 09:18:51 <ordex> ah
450 2011-06-03 09:18:57 <BlueMatt> or wait...no 1 sec
451 2011-06-03 09:18:58 <ordex> double weird
452 2011-06-03 09:19:00 <ordex> ?
453 2011-06-03 09:20:13 <BlueMatt> there are a couple places bitcoin might return 500, but none of them should be triggered if you are using a sane client
454 2011-06-03 09:20:37 <vegard> I got 500 before the block chain was completely downloaded
455 2011-06-03 09:20:38 <ordex> I'm using cpuminer-1.0.1
456 2011-06-03 09:20:42 <ordex> ah
457 2011-06-03 09:20:46 <ordex> so this could be the reason
458 2011-06-03 09:20:51 <ordex> as I'm still downloading blocks
459 2011-06-03 09:21:03 <BlueMatt> maybe
460 2011-06-03 09:21:17 <_flow_> is there someone who can help my recovering my wiki password?
461 2011-06-03 09:21:28 <ordex> error: {"code":-10,"message":"Bitcoin is downloading blocks..."}
462 2011-06-03 09:21:34 <BlueMatt> yea that you should see
463 2011-06-03 09:21:46 <BlueMatt> though I suppose I didnt realize bitcoin returned 500 when it did that...
464 2011-06-03 09:21:51 <BlueMatt> learn something new every day
465 2011-06-03 09:22:04 <BlueMatt> oh, of course, Im blind...ErrorReply
466 2011-06-03 09:22:20 <vegard> ah, yeah, I think I remember cpuminer doesn't report either the http code or the error message
467 2011-06-03 09:22:37 <vegard> I actually used protocol analyser to see that it was a 500 and that it had "downloading blocks" error message
468 2011-06-03 09:22:47 <BlueMatt> bitcoin doesnt report http code, cpuminer doesnt report error message
469 2011-06-03 09:23:16 <vegard> maybe that's how it was, yeah
470 2011-06-03 09:27:01 <andrew12> i think it's kind of funny that when everyone was saying "i'm using x's wonderful kernel", i didn't realize they were talking about opencl kernels (because i didn't know what those were) until last night
471 2011-06-03 09:32:36 <edcba> it's funny that a site aimed at exchanging magic the gathering cards is now exchanging bitcoins :)
472 2011-06-03 09:33:17 <edcba> it's quite an opposite 'money' exchanged here :)
473 2011-06-03 09:34:54 <jeremias> lol, was that the original purpose of mtgox?
474 2011-06-03 09:35:53 <zamgo2> to make profit, I assume
475 2011-06-03 09:44:00 <edcba> jeremias: mtgox
476 2011-06-03 10:09:19 <gavinandresen> Good morning everybody.  Did I miss anything while I was gone :-)
477 2011-06-03 10:09:34 <UukGoblin> gavinandresen, hi! :-)
478 2011-06-03 10:09:43 <UukGoblin> gavinandresen, price is at ~$13.8
479 2011-06-03 10:09:51 <MartianW> Wow, it's gavin!
480 2011-06-03 10:10:09 <BlueMatt> hey gavin is back :)
481 2011-06-03 10:10:09 <MartianW> UukGoblin, wow, I'm feeling stupid for selling at $4 now.
482 2011-06-03 10:10:23 <gavinandresen> Yeah, I noticed that.  I completely slept through all of the $11's.  Which is too bad, because 11 is my favorite number.
483 2011-06-03 10:10:34 <UukGoblin> ;-(
484 2011-06-03 10:11:11 <BlueMatt> oh gavin, do you want win/linux builds for git head for rc7/0.3.22
485 2011-06-03 10:11:36 <uppesnuppe> how many rc's are you planning? ;>
486 2011-06-03 10:11:48 <gavinandresen> 11 I hope!
487 2011-06-03 10:11:52 <BlueMatt> well we were planning 6 until yesterday...
488 2011-06-03 10:11:58 <uppesnuppe> :D
489 2011-06-03 10:13:06 <sipa> welcome back, gavinandresen :)
490 2011-06-03 10:13:19 <gavinandresen> BlueMatt: I don't have any use for a win build, and built a git-head bitcoind myself this morning....
491 2011-06-03 10:13:38 <BlueMatt> ok, fair enough...just wondered if you wanted semi-official ones
492 2011-06-03 10:14:08 <roconnor> what is the cited number of visa tranactions per second?
493 2011-06-03 10:14:16 <gavinandresen> Are any high-traffic sites using .22 yet?
494 2011-06-03 10:14:32 <BlueMatt> well mtgox has been complaining that .21 crashes with segfaults occasionally
495 2011-06-03 10:14:38 <BlueMatt> whereas .19 doesnt :(
496 2011-06-03 10:14:47 <gavinandresen> roconnor: See https://en.bitcoin.it/wiki/Scalability
497 2011-06-03 10:14:57 <BlueMatt> dont think he wants to upgrade for that reason...
498 2011-06-03 10:15:42 <gavinandresen> BlueMatt: I don't blame him.   ClearCoin was running a .21 bitcoind, which chose exactly the wrong time to stop responding to RPC requests
499 2011-06-03 10:16:04 <BlueMatt> lol yea heard about that
500 2011-06-03 10:16:36 <gavinandresen> I want to try to get more emphasis on reliability and bug fixing, and less on new features...
501 2011-06-03 10:16:45 <roconnor> gavinandresen: I see.  At visa number of transaction per second, I couldn't possibly verify the transactions myself  on my laptop.
502 2011-06-03 10:16:52 <gavinandresen> roconnor: nope
503 2011-06-03 10:17:01 <BlueMatt> not a bad idea...plus general code cleanup all over the place
504 2011-06-03 10:17:16 <gavinandresen> roconnor: then again, a laptop in 10 years might be able to handle it...
505 2011-06-03 10:17:34 <gavinandresen> BlueMatt: and unit tests
506 2011-06-03 10:17:40 <BlueMatt> that too
507 2011-06-03 10:17:48 <sipa> it would be great to fix these reported rpc lockups and crashes
508 2011-06-03 10:17:50 <roconnor> openssl speed ecdsa gives my friend:
509 2011-06-03 10:17:55 <roconnor> Doing 256 bit verify ecdsa's for 10s: 3268 256 bit ECDSA verify in 9.97s
510 2011-06-03 10:18:08 <roconnor> that seems 10x slower than the 3msec per verification cited on the wiki
511 2011-06-03 10:18:15 <roconnor> or maybe my math is wrong
512 2011-06-03 10:18:16 <sipa> but has anyone been able to reproduce these reliably?
513 2011-06-03 10:18:25 <roconnor> or maybe that elliptic curve is slow
514 2011-06-03 10:18:34 <sipa> roconnor: secp256k1?
515 2011-06-03 10:18:37 <gavinandresen> sipa:  I think some re-architecting of bitcoin internals might be needed to avoid the RPC deadlock issues
516 2011-06-03 10:18:47 <BlueMatt> totally agree
517 2011-06-03 10:18:51 <roconnor> sipa: I don't know what curve openssl speed ecdsa is using.
518 2011-06-03 10:18:52 <BlueMatt> 0.4.5 == rearch
519 2011-06-03 10:18:54 <BlueMatt> z?
520 2011-06-03 10:20:10 <sipa> gavinandresen: what do you see making 0.3.23/0.4.0/later versions?
521 2011-06-03 10:20:57 <phantomcircuit> ;;bc,blocks
522 2011-06-03 10:20:58 <gribble> 128353
523 2011-06-03 10:21:38 <sipa> BlueMatt: that'd be great
524 2011-06-03 10:21:43 <ordex> how many blocks should I have before starting mining?
525 2011-06-03 10:21:55 <llandru> anyone running the test network?
526 2011-06-03 10:21:59 <sipa> rpc and gui code shouldn't be touching any internal datastructures at all
527 2011-06-03 10:22:08 <gavinandresen> sipa: good questions; I think I agree with jgarzik that we need a 0.3.23 soon to fully implement the new 5-mill default fee
528 2011-06-03 10:22:23 <BlueMatt> I would think 0.4.0 would be quick enough
529 2011-06-03 10:22:33 <phantomcircuit> yeah and the connect timeout fee
530 2011-06-03 10:22:42 <BlueMatt> now that I want to see
531 2011-06-03 10:22:44 <sipa> gavinandresen: depends how quickly 0.3.22 is adopted - if so, 0.3.23 with new default fee policy
532 2011-06-03 10:22:54 <gavinandresen> llandru:  darn good question-- anybody know what's up with the test network?  Somebody did a huge block chain re-org that broke the test faucet
533 2011-06-03 10:22:59 <sipa> and maybe BlueMatt's patch to make it configurable
534 2011-06-03 10:24:03 <phantomcircuit> so anybody? conenct timeout?
535 2011-06-03 10:24:16 <sipa> connect timeout fee?
536 2011-06-03 10:24:31 <phantomcircuit> sipa, lol no my patch that makes Connect timeout
537 2011-06-03 10:24:51 <phantomcircuit> actually it needs one change first
538 2011-06-03 10:25:32 <sipa> wow, we crossed $14 :o
539 2011-06-03 10:25:51 <BlueMatt> if we do do a 0.3.23 with more than just connect fixes and one-line fee fixes, can we get a 0.4.0 branch as I dont want to see that sit and sit with features that may be great tagged for it...
540 2011-06-03 10:27:02 <sipa> i don't think 0.3.23 should be more than the absolutely necessary
541 2011-06-03 10:27:19 <BlueMatt> hm, out of all the nodes out there, my dnsseed has only been able to find 2441 which accept incoming, and only 2413 >= 0.3.19 on port 8333
542 2011-06-03 10:27:24 <doublec> have you tried using helgrind to narrow down deadlock issues?
543 2011-06-03 10:27:40 <TomyBoy3G> where i can get rc verison for win?
544 2011-06-03 10:28:10 <BlueMatt> forum thread/sf
545 2011-06-03 10:28:18 <gmaxwell> BlueMatt: I wounder how many of those 2441 will still be there in 24 hours?
546 2011-06-03 10:28:33 <BlueMatt> gmaxwell: there is pretty large turnover
547 2011-06-03 10:28:41 <BlueMatt> but the total stays pretty constant
548 2011-06-03 10:29:51 <gavinandresen> BlueMatt:  I think our current bottleneck is a lack of people really testing and banging on the code, and I fear that a 0.4 branch would just sit, untested, for a very long time.  Then when we eventually got to it, there would be a bunch of old bugs in that new code that were harder to track down and fix.
549 2011-06-03 10:30:00 <roconnor> I wonder if it would have been better to use RSA signatures instead of elliptic curve signatures:  10x larger signatures for 10x faster verification.
550 2011-06-03 10:30:27 <davex__> gavinandresen, is there something in the new version that rejects blocks generated by bitcoin nodes that are "too old"
551 2011-06-03 10:30:30 <BlueMatt> gavinandresen: totally agree, but if we just let it sit, half the stuff there will all of a sudden not even apply cleanly and may need serious work
552 2011-06-03 10:30:45 <gmaxwell> roconnor: 10x more data stored and transfered... doesn't sound like a win. I don't think the difference is actually 10x either, with the curve that bitcoin uses.
553 2011-06-03 10:30:58 <BlueMatt> does anyone have the resources to create huge dummy load on a bitcoind?
554 2011-06-03 10:31:22 <gavinandresen> davex__: really old nodes might still be trying to include bad transactions (see https://en.bitcoin.it/wiki/Incidents )
555 2011-06-03 10:31:24 <BlueMatt> hm, maybe Ill go do that...
556 2011-06-03 10:31:52 <davex__> gavinandresen, hmm, so .3.20 should have been ok.
557 2011-06-03 10:31:54 <gavinandresen> BlueMatt: not applying cleaning and requiring some work / a second look  is a good thing....
558 2011-06-03 10:32:05 <gmaxwell> roconnor: there are major performance difference between different ecc curves, I think bitcoin uses one from a class of faster curves.
559 2011-06-03 10:32:53 <gavinandresen> BlueMatt: You can create a pretty good stress test on a multicore system running two bitcoinds that talk to each other....
560 2011-06-03 10:33:07 <sipa> i believe i read a post by Hal saying that he wasn't sure whether the speed advantage was actually exploited by openssl
561 2011-06-03 10:33:18 <gavinandresen> BlueMatt:  at least, that's how I tracked down the last set of RPC deadlock issues.
562 2011-06-03 10:33:19 <BlueMatt> gavinandresen: maybe so...I just dont like the idea of some cool stuff sitting around that could be *very* useful for some people
563 2011-06-03 10:33:38 <BlueMatt> sipa: I believe I read something like that as well
564 2011-06-03 10:33:46 <BlueMatt> because its a fairly less-used curve
565 2011-06-03 10:34:00 <BlueMatt> gavinandresen: alright...I suppose Ill go see if I can set something like that up
566 2011-06-03 10:34:26 <BlueMatt> MagicalTux: ping
567 2011-06-03 10:34:33 <MagicalTux> pong
568 2011-06-03 10:34:57 <BlueMatt> do either of you guys happen to have any clues as to which rpc commands might be causing the lockup
569 2011-06-03 10:35:03 <BlueMatt> ie what ones do you use on your sites
570 2011-06-03 10:35:15 <BlueMatt> just send/check balance/check received/getnewaddr?
571 2011-06-03 10:35:49 <ordex> when should my bitcoind stop downloading blocks?
572 2011-06-03 10:35:58 <BlueMatt> ;;bc,blocks
573 2011-06-03 10:35:58 <gribble> 128354
574 2011-06-03 10:36:01 <BlueMatt> when you have that many
575 2011-06-03 10:36:51 <phantomcircuit> gavinandresen, the current release is basically completely broken for normal users
576 2011-06-03 10:37:10 <gavinandresen> phantomcircuit: how so?
577 2011-06-03 10:37:37 <gavinandresen> phantomcircuit: (and I assume you mean people who download and run the bitcoin client?)
578 2011-06-03 10:37:43 <ordex> BlueMatt:thanks
579 2011-06-03 10:37:58 <gmaxwell> gavinandresen: New nodes being nat basically don't come up. Or sometimes take continual restarting to get any connections at all, or may take dozens of hours to get connections&
580 2011-06-03 10:38:07 <gmaxwell> s/being/behind/
581 2011-06-03 10:38:27 <phantomcircuit> gavinandresen, if you dont specify a node via -connect it can take a good hours if not days to get a connection to another node
582 2011-06-03 10:38:31 <gavinandresen> Ah, the connection problems.  Yup, that's a screw-up.
583 2011-06-03 10:38:42 <phantomcircuit> yeah well i fixed it
584 2011-06-03 10:38:53 <gmaxwell> gavinandresen: http://forum.bitcoin.org/index.php?topic=11126.msg158368#msg158368
585 2011-06-03 10:39:02 <gmaxwell> sipa: yes, reachablity would be great.
586 2011-06-03 10:39:35 <phantomcircuit> upnp dnsseed and a connect timeout are all completely necessary
587 2011-06-03 10:39:46 <gmaxwell> sipa: really there are several important things that all need to happen soonish. Connect, seeding improvements, reachablity, upnp....
588 2011-06-03 10:41:09 <gavinandresen> Those are exactly the sort of reliability improvements I want to concentrate on.  Also more DoS prevention measures (dropping ill-behaving peers) and connection rotation (should prevent all sorts of classes of Sybil attacks)
589 2011-06-03 10:41:10 <sipa> phantomcircuit: i'll have a look at your patch
590 2011-06-03 10:41:46 <BlueMatt> sipa: my current issue with that flag: you cant disable it once you have broad casted that you have support for it
591 2011-06-03 10:41:46 <gmaxwell> gavinandresen: I guess you missed the discussion about goofy timing attacks too?
592 2011-06-03 10:42:01 <BlueMatt> as nodes will not remove services from addr.dat, only add to
593 2011-06-03 10:42:05 <gavinandresen> gmaxwell: I skimmed it (I was SUPPOSED to be on vacation....)
594 2011-06-03 10:42:19 <gmaxwell> gavinandresen: if you simply have access to lots of IPs you can drive a listening target's network time ???70 minutes.
595 2011-06-03 10:42:26 <BlueMatt> gavinandresen: lol, like you can do that...
596 2011-06-03 10:43:13 <phantomcircuit> gavinandresen, for connection rotation you want to maintain half and half
597 2011-06-03 10:43:15 <gmaxwell> gavinandresen: since the gap can be >2h if one is a miner you can make another node ignore that miner's blocks, just creating a split. Though I don't think anyone has actually tried doing this to see if it works.
598 2011-06-03 10:43:39 <gavinandresen> gmaxwell: seems like defining "time way out of whack" as one category if 'ill-behaving peer' might be a good approach to prevent it.  Along with changing your peers regularly.
599 2011-06-03 10:43:46 <sipa> BlueMatt: i've been thinking about it as well - assume the whole network upgrade to v0.3.XX which has the extra flag, and you enter the network as new client - you don't advertize yourself as reachable, so no one will try to reach you, right?
600 2011-06-03 10:44:18 <phantomcircuit> a reachable flag is a bad idea
601 2011-06-03 10:44:19 <BlueMatt> sipa: yea pretty much...or you could flip the meaning of the flag, meaning you cant turn it off...
602 2011-06-03 10:44:27 <gmaxwell> gavinandresen: yea well the bigger issue there is that it doesn't appear to forget adjustments for peers that have gone away, so the rotation wouldn'tcurrently fixit.
603 2011-06-03 10:44:33 <BlueMatt> which means you can never turn on "connect to me"
604 2011-06-03 10:45:11 <gmaxwell> gavinandresen: and the adjustment it clamped, but only to 70 minutes, so the mutal gap can be wide enough to trigger the block check.
605 2011-06-03 10:45:30 <BlueMatt> to pull this off either way, you have to carefully consider how flags are handled, as its an easy sybil if you overwrite any given flag, but or'ing them together also doesnt really work...
606 2011-06-03 10:45:30 <sipa> BlueMatt: maybe the solution is simply to remove addresses from the database which have been confirmed not to be reachable?
607 2011-06-03 10:45:51 <gavinandresen> gmaxwell: I'm currently more worried about day-to-day reliability of bitcoin services than theoretical, difficult-to-pull-off timing attacks.
608 2011-06-03 10:45:57 <sipa> BlueMatt: so they won't be passed on either
609 2011-06-03 10:46:09 <BlueMatt> sipa: maybe...but Im not sure how effective that would be
610 2011-06-03 10:46:32 <phantomcircuit> gavinandresen, that timing attack would be trivial if combined with a sybil attack
611 2011-06-03 10:46:34 <llandru> if anyone is running on the test network, please msg
612 2011-06-03 10:46:36 <gmaxwell> gavinandresen: Quite fair. Well right. Basic reachabliity has problems now, at least for firewalled nodes.
613 2011-06-03 10:46:40 <phantomcircuit> and i already have the code for the sybil
614 2011-06-03 10:46:49 <gavinandresen> phantomcircuit: I didn't say I'm NOT worried about it....
615 2011-06-03 10:46:53 <sipa> llandru: i have a reachable node on testnet, but i'm not mining
616 2011-06-03 10:46:55 <BlueMatt> maybe, not as a solution, but as an addition to the solution, if your node has not connected in x days, it pulls from dnsseed and connects to one of those *first* then gets a more recent node list?
617 2011-06-03 10:46:59 <gmaxwell> (and the connectivity problems make sybil worse, of course)
618 2011-06-03 10:47:10 <BlueMatt> imo that + connect timeout would mostly fix it
619 2011-06-03 10:47:31 <gmaxwell> Makes control of dnsseed more attractive for an attacker, no?
620 2011-06-03 10:47:43 <BlueMatt> gmaxwell: thats why we need many dnsseeds :)
621 2011-06-03 10:47:56 <phantomcircuit> i should probably change from a connect() loop to connect + select, but that's slightly more complicated logic
622 2011-06-03 10:47:56 <sipa> phantomcircuit: hmm, where is your patch?
623 2011-06-03 10:48:05 <phantomcircuit> sipa, one sec
624 2011-06-03 10:48:12 <BlueMatt> but we only have one irc server now so dnsseed is a big improvement anyway
625 2011-06-03 10:48:30 <sipa> BlueMatt: i disagree - i want the bitcoin network to be as self-contained as possible, dnsseed shouldn't be more than an entry point
626 2011-06-03 10:48:36 <sipa> not a way for finding good peers
627 2011-06-03 10:48:49 <BlueMatt> also, yes we should remove nodes that we failed to connect to, my question, sipa, is not if we should do it but if it would work as they would still be passed by other nodes and probably end up right back in your db
628 2011-06-03 10:49:23 <BlueMatt> maybe we always try to reverse connect to nodes which just connected to us and then only put them in db if that worked?
629 2011-06-03 10:49:37 <sipa> maybe
630 2011-06-03 10:49:45 <sipa> and if that fails, let them know they are unreachable
631 2011-06-03 10:49:50 <BlueMatt> does gnutella regrab its http seed stuff if it is failing?
632 2011-06-03 10:49:51 <gavinandresen> BlueMatt:  seems like one category of "ill-behaving peers" is peers that send you lots of addresses of nodes that are not connectable
633 2011-06-03 10:49:56 <sipa> after which they stop announcing their own IP
634 2011-06-03 10:50:13 <phantomcircuit> sipa, http://bitcoinconsultancy.com/nonblocking.patch
635 2011-06-03 10:50:19 <BlueMatt> sipa: I dont think you can stop announcing your ip
636 2011-06-03 10:50:20 <gmaxwell> BlueMatt: if it's failing is different then if X days. If it's failing it should dnsseed again I think.
637 2011-06-03 10:50:28 <BlueMatt> nodes will still publish their list of "Im connected to"
638 2011-06-03 10:50:34 <BlueMatt> well old nodes, that is
639 2011-06-03 10:51:04 <sipa> BlueMatt: right, it does need some serious thinking
640 2011-06-03 10:51:16 <BlueMatt> gmaxwell: true, but they probably happen at the same time 90% of the time
641 2011-06-03 10:51:40 <BlueMatt> gavinandresen: yes, absolutely needs a ton of thought...
642 2011-06-03 10:51:45 <ordex> wa
643 2011-06-03 10:52:37 <gmaxwell> BlueMatt: yes, but  saves figuring out a good value for X, and reduces the value of attacks on dnsseed by at least the 10%. :)
644 2011-06-03 10:53:26 <BlueMatt> fair enough
645 2011-06-03 10:54:59 <gmaxwell> BlueMatt: counter argument is that if only one of the nodes you remember works, then that node can stick you on a fantasy network pretty easily.
646 2011-06-03 10:55:23 <gmaxwell> (just tells you about his own sybil nodes)
647 2011-06-03 10:56:02 <BlueMatt> can we do some kind of trust thing for addr.dat, ie rank nodes higher on the connect-to list if youve heard of it from more nodes?
648 2011-06-03 10:56:23 <anarchyx> ;;bc,stats
649 2011-06-03 10:56:26 <gribble> Current Blocks: 128357 | Current Difficulty: 434882.7217497 | Next Difficulty At Block: 129023 | Next Difficulty In: 666 blocks | Next Difficulty In About: 3 days, 19 hours, 12 minutes, and 18 seconds | Next Difficulty Estimate: 525411.41916702
650 2011-06-03 10:56:49 <gmaxwell> Subject to sybil attack itself. Might make you less likely to get even _one_ good node?  "The 10,000 clones didn't tell me about you!"
651 2011-06-03 10:56:59 <ordex> why is my block counter stuck even if I have 6 connesions open?
652 2011-06-03 10:57:07 <phantomcircuit> a sybil attack against the current network code would be absolutely trivial
653 2011-06-03 10:57:17 <phantomcircuit> seriously who wants to see one? ill do it right now
654 2011-06-03 10:58:03 <gmaxwell> BlueMatt: is there any authentication on the dnsseed? e.g. if there are N dnsseeders, can I actually tell that I'm getting data from N parties?
655 2011-06-03 10:58:25 <gmaxwell> (sorry, haven't really had a chance to look at the dnsseed stuff)
656 2011-06-03 10:58:26 <sipa> phantomcircuit: doesn't compile here
657 2011-06-03 10:58:31 <BlueMatt> you could, but currently it doesnt, but why is that necessary?
658 2011-06-03 10:58:39 <phantomcircuit> sipa, what's the error?
659 2011-06-03 10:58:47 <sipa> is your code windows-only?
660 2011-06-03 10:58:59 <phantomcircuit> no im running it on gentoo right now
661 2011-06-03 10:59:04 <gmaxwell> BlueMatt: I guess it doesn't. Someone who controls your network well enough to fake your dnsseed requests can just answer all your node connect attempts.
662 2011-06-03 10:59:05 <sipa> net.cpp: In function bool ConnectSocket(const CAddress&, SOCKET&, int):
663 2011-06-03 10:59:07 <phantomcircuit> lol gtting it to work on windwos took like 4 hours
664 2011-06-03 10:59:08 <sipa> net.cpp:133: error: WSAEALREADY was not declared in this scope
665 2011-06-03 10:59:10 <phantomcircuit> it was a huge pita
666 2011-06-03 10:59:10 <sipa> net.cpp:133: error: WSAEINVAL was not declared in this scope
667 2011-06-03 10:59:14 <BlueMatt> gmaxwell: yep
668 2011-06-03 10:59:17 <phantomcircuit> sipa, oh i forgot util.h modifications
669 2011-06-03 10:59:48 <BlueMatt> phantomcircuit: how big is this patch getting?
670 2011-06-03 10:59:57 <phantomcircuit> sipa, http://bitcoinconsultancy.com/nonblocking.patch
671 2011-06-03 11:00:05 <phantomcircuit> BlueMatt, it's that big
672 2011-06-03 11:01:00 <phantomcircuit> it could be much smaller
673 2011-06-03 11:01:32 <gmaxwell> -    devlist = upnpDiscover(2000, multicastif, minissdpdpath, 0);
674 2011-06-03 11:01:33 <gmaxwell> +    devlist = upnpDiscover(2000, multicastif, minissdpdpath, NULL);
675 2011-06-03 11:01:39 <gmaxwell> doesn't seem strictly minimal. :)
676 2011-06-03 11:01:42 <sipa> a separate function for MakeNonBlock() would be nice, i guess
677 2011-06-03 11:02:30 <BlueMatt> can you not add more crap to headers.h, if you need to include something only in that file, please put it in net.h or relevant
678 2011-06-03 11:02:49 <BlueMatt> and yea, why did you change the upnp code?
679 2011-06-03 11:02:59 <phantomcircuit> oh that's just a side effect
680 2011-06-03 11:03:24 <BlueMatt> and reapplied sipa's dnsseed patch
681 2011-06-03 11:03:30 <phantomcircuit> BlueMatt, also you're reading the patch wrong i dont modify headers.h at all
682 2011-06-03 11:03:45 <sipa> indeed
683 2011-06-03 11:03:54 <BlueMatt> oh sorry, thats my bad
684 2011-06-03 11:04:04 <BlueMatt> but why did you change util.h?
685 2011-06-03 11:04:34 <BlueMatt> if you need to define new constants cant you do it in net.h or net.cpp?
686 2011-06-03 11:04:34 <phantomcircuit> BlueMatt, definitions of WSAEINPROGRESS etc
687 2011-06-03 11:04:39 <sipa> i suppose for compatibility with windows
688 2011-06-03 11:04:46 <phantomcircuit> yeah i could but all the other ones are in util.h
689 2011-06-03 11:04:54 <phantomcircuit> so i figured it was best to put them in the same place
690 2011-06-03 11:04:59 <sipa> that looks reasonable to me
691 2011-06-03 11:05:11 <phantomcircuit> also this is a git diff so sipas patch is included jsut because
692 2011-06-03 11:05:30 <BlueMatt> but if you already have half of the connect stuff in ifdef's, why define constants, just use the original names?
693 2011-06-03 11:05:57 <BlueMatt> though I suppose that is better...
694 2011-06-03 11:06:07 <BlueMatt> nvm
695 2011-06-03 11:06:24 <phantomcircuit> this patch only has the non blocking parts in ifdefs
696 2011-06-03 11:06:28 <phantomcircuit> the rest of the code is shared
697 2011-06-03 11:07:22 <phantomcircuit> also i wouldn't put it in a function MakeNonBlocking because it sets it back to blocking for compatibility reasons
698 2011-06-03 11:08:10 <BlueMatt> yea...hm, well if there is no better way to do non-blocking than the ugly while loops...
699 2011-06-03 11:08:29 <phantomcircuit> and actually those headers are unnecessary now
700 2011-06-03 11:08:45 <phantomcircuit> BlueMatt, well there is but i didn't want to do it because it's slightly more obscure
701 2011-06-03 11:08:52 <phantomcircuit> and everybody kept saying to make it simpler
702 2011-06-03 11:08:55 <BlueMatt> and it is?
703 2011-06-03 11:09:01 <phantomcircuit> connect/select
704 2011-06-03 11:09:06 <BlueMatt> ah
705 2011-06-03 11:10:41 <phantomcircuit> BlueMatt, this was is simpler but i would only do it this way for a single socket
706 2011-06-03 11:10:56 <sipa> 500x Sleep(1) ?
707 2011-06-03 11:11:09 <phantomcircuit> the real solution here is to try multiple connections simultaneously and then drop ones to get back to 8
708 2011-06-03 11:11:25 <phantomcircuit> but well screw that
709 2011-06-03 11:11:33 <BlueMatt> my problem is just that it will end up calling sleep so many times...so many loop iterations
710 2011-06-03 11:11:56 <phantomcircuit> it iterates a maximum of 500 times
711 2011-06-03 11:13:35 <sipa> it does work
712 2011-06-03 11:13:51 <BlueMatt> its just such a hack...
713 2011-06-03 11:13:58 <phantomcircuit> http://bitcoinconsultancy.com/nonblocking.patch
714 2011-06-03 11:14:03 <phantomcircuit> there slightly better
715 2011-06-03 11:14:18 <phantomcircuit> BlueMatt, this is why im not super interested in refactoring mainline
716 2011-06-03 11:14:25 <phantomcircuit> to fix shit you need to use hacks everywhere
717 2011-06-03 11:14:40 <BlueMatt> then rewrite it proper
718 2011-06-03 11:14:47 <BlueMatt> ie this connection stuff
719 2011-06-03 11:14:52 <phantomcircuit> you dont get it
720 2011-06-03 11:15:00 <gmaxwell> I don't see why a hack is required... use select() instead of sleeping, for example.
721 2011-06-03 11:15:03 <phantomcircuit> a proper rewrite would be extensive
722 2011-06-03 11:15:06 <phantomcircuit> and a massive pita
723 2011-06-03 11:15:16 <gmaxwell> Though, yes, running them in parallel would make a lot more sense.
724 2011-06-03 11:15:21 <BlueMatt> way less than starting ground-up
725 2011-06-03 11:15:43 <phantomcircuit> what?
726 2011-06-03 11:16:07 <phantomcircuit> the entire networking stack could fit into a couple hundred loc with boost::asio
727 2011-06-03 11:16:21 <phantomcircuit> the serialization of structures is going to be the hardest part
728 2011-06-03 11:16:46 <BlueMatt> ok then rewrite it to do that
729 2011-06-03 11:16:57 <BlueMatt> reimplement network from ground up
730 2011-06-03 11:17:05 <BlueMatt> instead of a whole client
731 2011-06-03 11:17:21 <phantomcircuit> sigh
732 2011-06-03 11:17:22 <phantomcircuit> no
733 2011-06-03 11:17:23 <BlueMatt> you dont lose existing patches/work and you still get all the advantage
734 2011-06-03 11:17:24 <BlueMatt> s
735 2011-06-03 11:18:29 <matija> Hi, who can point me to the shortest way tolocalization of client and documentation?
736 2011-06-03 11:18:49 <BlueMatt> you want what now?
737 2011-06-03 11:18:55 <matija> I saw a thread in the forum ...
738 2011-06-03 11:19:14 <phantomcircuit> he wants to help translate the client i believe
739 2011-06-03 11:19:16 <matija> I can translate the client and the documentation in Slovene.
740 2011-06-03 11:19:31 <BlueMatt> for the client: http://forum.bitcoin.org/index.php?topic=8390.0
741 2011-06-03 11:19:32 <matija> phantomc., exactly
742 2011-06-03 11:20:24 <ezyang> I wrote an article on the crypto bitcoin uses, if any devs are around and what to help fact check before a publish...
743 2011-06-03 11:20:46 <phantomcircuit> ezyang, sure
744 2011-06-03 11:20:52 <sipa> sure
745 2011-06-03 11:20:54 <BlueMatt> linkidy link?
746 2011-06-03 11:20:59 <matija> BlueM., thank you. And for documentation, FAQ, website?
747 2011-06-03 11:21:16 <ezyang> http://blog.ezyang.com/2011/06/the-cryptography-of-bitcoin/
748 2011-06-03 11:21:36 <BlueMatt> mtrlt: the part on the client is out-of-date, but the stuff about the website should be roughly right: http://forum.bitcoin.org/index.php?topic=151.0;all
749 2011-06-03 11:22:28 <sipa> ezyang: cryptopp is only used for mining, i believe
750 2011-06-03 11:22:47 <BlueMatt> yep
751 2011-06-03 11:22:56 <phantomcircuit> ezyang, yeah i was about to say
752 2011-06-03 11:22:57 <forexmasterja> what the hell happened in the past few hours ???? I go to sleep for a few hours and wake up to see $14 ????? !!! Wow
753 2011-06-03 11:22:58 <BlueMatt> openssl is used for ecdsa and such
754 2011-06-03 11:23:11 <phantomcircuit> forexmasterja, smartmoney.com
755 2011-06-03 11:23:47 <sipa> ezyang: openssl is used for normal hashing
756 2011-06-03 11:23:55 <ezyang> Cool, updated that.
757 2011-06-03 11:24:16 <BlueMatt> ezyang: the use of secp256k1 is quite non-arbitrary, it was picked for speed as it has the potential for some large improvements in calculations
758 2011-06-03 11:24:28 <BlueMatt> which could be quite a big deal if bitcoin becomes big
759 2011-06-03 11:25:05 <phantomcircuit> BlueMatt, nobody ahs actually implemented those improvements though
760 2011-06-03 11:25:07 <ezyang> Oh, I do think I remember reading something about that.
761 2011-06-03 11:25:12 <phantomcircuit> so for the time being it's purely theoretical
762 2011-06-03 11:25:13 <ezyang> http://forum.bitcoin.org/index.php?topic=3238.0
763 2011-06-03 11:25:19 <BlueMatt> phantomcircuit: yes, but it could be done if necessary
764 2011-06-03 11:25:27 <BlueMatt> that is why it would have been chosen
765 2011-06-03 11:25:53 <phantomcircuit> ezyang, also you dont seem to make a difference between splitting the block chain and completely reconstructing it
766 2011-06-03 11:25:57 <phantomcircuit> which are very different things
767 2011-06-03 11:26:11 <BlueMatt> satoshi was no idiot, otherwise he would have chosen a standard one, he mustve had a reason and afaict that is the only legitimate reason
768 2011-06-03 11:26:52 <ezyang> phantomcircuit: Ah, is reconstructing it "resolving" the entire thing.
769 2011-06-03 11:27:01 <sipa> i do believe the choice for secp256k1 was a bit arbitrary
770 2011-06-03 11:27:23 <ezyang> I /think/ the attack I described was the chain splitting, not reconstruction attack.
771 2011-06-03 11:27:27 <BlueMatt> I dont think satoshi was near stupid enough for that
772 2011-06-03 11:27:45 <phantomcircuit> ezyang, reconstructing the entire block chain is requires the same fundamental break as splitting the block chain, but requires doing all of the blocks
773 2011-06-03 11:28:01 <ezyang> OK, fair enough.
774 2011-06-03 11:28:50 <ezyang> Wait wait, but that attack won't work
775 2011-06-03 11:28:56 <ezyang> because the clients have built in known good chains.
776 2011-06-03 11:29:29 <BlueMatt> no all they have is if (nHeight == xxxx) hash must be yyyyyyyyyyyy
777 2011-06-03 11:29:40 <ezyang> oh. Ha.
778 2011-06-03 11:30:00 <ezyang> OK, reworded.
779 2011-06-03 11:30:03 <Insti> !seen genjix
780 2011-06-03 11:30:35 <phantomcircuit> Insti, would you like so speek with him?
781 2011-06-03 11:30:36 <phantomcircuit> speak*
782 2011-06-03 11:30:55 <sipa> ezyang: also, if you're not attacking the block chain in any way, you will need to break both sha256+ripemd160 and ecdsa for single-use-addresses, as you only know ripemd160(sha256(ecdsa_pubkey(X)))==A, and you need to find an X
783 2011-06-03 11:30:59 <Insti> phantomcircuit yes
784 2011-06-03 11:31:36 <vegard> why do you have to break sha256+ripemd160? surely it's enough to find X?
785 2011-06-03 11:31:39 <ezyang> sipa: Not true, I can inject my input at any stage of the hash.
786 2011-06-03 11:31:46 <vegard> (the public key is part of the address, right?)
787 2011-06-03 11:32:02 <ezyang> This requires a pretty ridiculous break in the hashing algorith, though.
788 2011-06-03 11:32:25 <sipa> ezyang: you need to create a valid signature for your transaction, right?
789 2011-06-03 11:32:42 <ezyang> Hmm, what specific scenario are we talking about here?
790 2011-06-03 11:32:50 <phantomcircuit> Insti, ^
791 2011-06-03 11:32:56 <vegard> sipa: nvm. I was wrong.
792 2011-06-03 11:33:38 <sipa> ezyang: i admit i haven't read all details, so correct me if that's not what you're trying to do - if you want to spend a coin i have, addresses to a single-use address of mine
793 2011-06-03 11:33:52 <sipa> *addressed
794 2011-06-03 11:34:01 <ezyang> Oh oh, right, you can't use a replay attack on single use address, that's correct.
795 2011-06-03 11:35:10 <ezyang> (cuz there's no signature to replay)