1 2012-11-05 00:00:17 <sipa> does that mean we have to implement our own TCP/IP stack?
2 2012-11-05 00:00:28 <edcba> ipx/spx
3 2012-11-05 00:00:28 <Luke-Jr> no, edcba will do that for us
4 2012-11-05 00:00:33 <sipa> ha, yes!
5 2012-11-05 00:00:50 <wizkid057> Luke-Jr: the fact that you dream about bitcoin means you need to get out more :)
6 2012-11-05 00:01:07 <Luke-Jr> wizkid057: probably. but my next scheduled "get out" also involves Bitcoin!
7 2012-11-05 00:01:13 <wizkid057> ...
8 2012-11-05 00:01:15 <Luke-Jr> lol
9 2012-11-05 00:01:19 <wizkid057> when/where is that again?
10 2012-11-05 00:01:24 <Luke-Jr> wizkid057: Nov 20 in Orlando
11 2012-11-05 00:01:27 <wizkid057> hmm
12 2012-11-05 00:01:30 <wizkid057> ACTION checks calander
13 2012-11-05 00:01:40 <wizkid057> ah yeah, that wont happen
14 2012-11-05 00:01:41 <wizkid057> lol
15 2012-11-05 00:01:45 <Luke-Jr> looking forward to meeting gmaxwell and forrestv ???
16 2012-11-05 00:01:54 <edcba> why ? end of world is 21 dec !
17 2012-11-05 00:01:55 <conman> luckily I wont be there
18 2012-11-05 00:02:49 <sipa> Luke-Jr: have a gpg key signing party!
19 2012-11-05 00:03:54 <edcba> or just exchange coins
20 2012-11-05 00:04:32 <wizkid057> so far my CPPSRB code is ~1200 lines of craziness
21 2012-11-05 00:04:44 <sipa> cppsrb?
22 2012-11-05 00:04:46 <conman> does it involve byte flipping?
23 2012-11-05 00:05:00 <wizkid057> conman: yeah, had to convert a solution->block hash
24 2012-11-05 00:05:01 <wizkid057> lol
25 2012-11-05 00:05:03 <sipa> conman: better, *double* byteflips
26 2012-11-05 00:05:23 <conman> in the pike position
27 2012-11-05 00:05:24 <Luke-Jr> sipa: Capped PPS with Recent Backpay :P
28 2012-11-05 00:05:26 <wizkid057> sipa: http://eligius.st/wiki/index.php/Capped_PPS_with_Recent_Backpay
29 2012-11-05 00:05:32 <conman> from the top of a 10m diving platform
30 2012-11-05 00:06:22 <edcba> wow i don't understand anything
31 2012-11-05 00:06:42 <conman> get in line
32 2012-11-05 00:06:46 <edcba> pools have really complex retribution schemes
33 2012-11-05 00:07:18 <edcba> CPPSRB storage rounds shares PPLNS
34 2012-11-05 00:09:29 <wizkid057> definitely need to clean this code up
35 2012-11-05 00:09:39 <wizkid057> before main() gets even more unmanagable
36 2012-11-05 00:10:04 <edcba> you are reading bitcoin main.cpp ?
37 2012-11-05 00:10:08 <conman> holy cow still no share?
38 2012-11-05 00:10:15 <conman> grr
39 2012-11-05 00:10:19 <edcba> https://en.bitcoin.it/wiki/Comparison_of_mining_pools...
40 2012-11-05 00:10:50 <edcba> what is wrong with proportional ?
41 2012-11-05 00:11:01 <wizkid057> read up on pool hopping
42 2012-11-05 00:11:50 <edcba> because you don't know how to calc prop efficiently
43 2012-11-05 00:12:20 <conman> guh here we go again
44 2012-11-05 00:12:38 <conman> edcba, please feel free to set up another prop pool... so I can hop it
45 2012-11-05 00:12:48 <conman> we need more hoppable pools
46 2012-11-05 00:12:57 <wizkid057> sounds good
47 2012-11-05 00:13:02 <edcba> the problem is i would need to setup my own interface
48 2012-11-05 00:13:22 <edcba> ie i'd need to rewrite mining code
49 2012-11-05 00:13:37 <conman> excuses excuses, hurry up so we can rip you off
50 2012-11-05 00:13:42 <wizkid057> ok, have to stop getting distracted if i'm going to get enough code done for some payouts
51 2012-11-05 00:13:52 <edcba> because i don't think miners would like to change their code
52 2012-11-05 00:14:40 <edcba> and i would also need to write some p2p thingy because i don't want to use a server that would be easily ddosed :)
53 2012-11-05 00:16:06 <sipa> edcba: main.cpp doesn't have a main() :P
54 2012-11-05 00:16:40 <edcba> that doesn't even surprise me :/
55 2012-11-05 00:18:30 <conman> hmm I still think I should just give up on gbt and accept only stratum
56 2012-11-05 00:18:40 <conman> this is the timewaster 2000
57 2012-11-05 00:19:55 <wizkid057> :9
58 2012-11-05 00:19:59 <wizkid057> * :(
59 2012-11-05 00:22:58 <conman> what a piece of horseshit
60 2012-11-05 00:23:00 <conman> I give up
61 2012-11-05 00:24:08 <conman> oops wrong tab
62 2012-11-05 00:34:04 <sipa> etotheipi_: the parallizability is a nice argument in favor of tries
63 2012-11-05 00:34:12 <conman> well flipping the bytes before it's given to the mining machinery produces no shares at all
64 2012-11-05 00:38:47 <etotheipi_> sipa: thanks
65 2012-11-05 00:39:47 <etotheipi_> I'm getting tired of this thread... but I still feel compelled to fight it out :)
66 2012-11-05 00:40:10 <etotheipi_> maybe I will switch my priorities to working on that after I release Armory Beta, instead of the new wallet :)
67 2012-11-05 00:40:31 <sipa> i wish i had time for this...
68 2012-11-05 00:44:08 <sipa> adding a merkle patricia tree structure on top of the existing utxo set now that's merged in mainline shouldn't be too difficult though
69 2012-11-05 00:47:27 <D34TH> i think i figured out my issue in compiling *coin*
70 2012-11-05 00:47:35 <D34TH> testing now
71 2012-11-05 00:54:37 <D34TH> sipa: mind if i suggest 2 things for bitcoin-qt.pro
72 2012-11-05 00:56:50 <D34TH> screw it, fork + commit + issue
73 2012-11-05 01:02:41 <D34TH> sipa: https://github.com/bitcoin/bitcoin/pull/1984
74 2012-11-05 01:21:23 <D34TH> Luke-Jr, any idea what would cause this? http://pastebin.com/hWaZKkDM
75 2012-11-05 01:21:44 <Luke-Jr> D34TH: I've not compiled a LevelDB client yet, sorry.
76 2012-11-05 01:21:56 <D34TH> thanks anyways
77 2012-11-05 01:22:39 <Luke-Jr> edcba: pretty sure I've open sourced enough code to setup a simple prop pool
78 2012-11-05 01:23:08 <Luke-Jr> edcba: also, wtf is so complicated about CPPSRB? O.o
79 2012-11-05 01:24:28 <D34TH> Error: Modules with different CPU types were found.
80 2012-11-05 01:24:33 <D34TH> i think thats my problem
81 2012-11-05 01:30:21 <SuckMyWub> why do you all suck so many wubs?
82 2012-11-05 01:31:11 <forsetifox> <.<
83 2012-11-05 01:31:30 <SuckMyWub> mother fuckin wub suckers
84 2012-11-05 01:31:45 <conman> you make a lot of sense SuckMyWub
85 2012-11-05 01:31:57 <conman> allow me to send you all my money
86 2012-11-05 01:32:03 <D34TH> yes, me too
87 2012-11-05 01:32:04 <SuckMyWub> ok
88 2012-11-05 01:32:05 <SuckMyWub> gimme
89 2012-11-05 01:32:41 <SuckMyWub> 13DRAm57deBUgUfPAoZfp3cqHsqWkVhWRg
90 2012-11-05 01:32:43 <SuckMyWub> gimme money
91 2012-11-05 01:33:18 <gmaxwell> hopefully that wasn't a bad idea.
92 2012-11-05 01:33:34 <D34TH> bad ideas have the best intentions
93 2012-11-05 01:44:48 <conman> aww I wanted to send him every cent
94 2012-11-05 01:49:41 <phantomcircuit> gmaxwell, send him 1 satoshi
95 2012-11-05 01:49:46 <phantomcircuit> laugh when he cant spend it
96 2012-11-05 01:49:50 <phantomcircuit> bwahaha
97 2012-11-05 01:49:51 <conman> hah
98 2012-11-05 01:49:55 <conman> I will send him one
99 2012-11-05 01:49:56 <conman> :D
100 2012-11-05 01:50:15 <phantomcircuit> i was kidding you can use large txout's to "vacuum" smaller txout's
101 2012-11-05 01:50:27 <sipa> better, send him 0 satoshi
102 2012-11-05 01:50:32 <phantomcircuit> lol
103 2012-11-05 01:50:37 <conman> yeah but that I've already don
104 2012-11-05 01:50:37 <gmaxwell> meh.
105 2012-11-05 01:50:38 <conman> e
106 2012-11-05 01:58:57 <conman> sigh I wonder if I should start up trying to debug this again
107 2012-11-05 01:59:15 <conman> wizkid057 was kind enough to put up the share log, maybe that will help
108 2012-11-05 01:59:30 <wizkid057> hope it helps
109 2012-11-05 01:59:39 <conman> me2 :s
110 2012-11-05 02:00:03 <conman> still dont understand those last 4 bytes being different to what I submitted...
111 2012-11-05 02:00:21 <wizkid057> dunno
112 2012-11-05 02:00:46 <Luke-Jr> if you can convince him to push the code, I can probably find the problem(s) ;p
113 2012-11-05 02:01:09 <wizkid057> conman: still have Luke-Jr on ignore? lol
114 2012-11-05 02:01:17 <conman> put him back on it
115 2012-11-05 02:01:24 <conman> doesn't take much sorry
116 2012-11-05 02:01:28 <wizkid057> <Luke-Jr> if you can convince him to push the code, I can probably find the problem(s) ;p
117 2012-11-05 02:01:31 <conman> guess I could undo ..
118 2012-11-05 02:01:53 <conman> might even leave him off ignore (for a while) if he ends up fixing it
119 2012-11-05 02:01:53 <forsetifox> Heh.
120 2012-11-05 02:02:07 <conman> but I feel so dirty
121 2012-11-05 02:02:20 <conman> ok might push then
122 2012-11-05 02:02:29 <wizkid057> ACTION hands conman some hand sanitizer
123 2012-11-05 02:02:30 <conman> but first fud you go back on ignore
124 2012-11-05 02:04:58 <conman> https://github.com/ckolivas/cgminer/blob/gbt/cgminer.c#L1362 onwards
125 2012-11-05 02:05:07 <conman> most of the gbt functions are one after the other there
126 2012-11-05 02:05:28 <D34TH> str.contains("_gbt_")?
127 2012-11-05 02:05:33 <D34TH> :3
128 2012-11-05 02:06:02 <forsetifox> You could always vent in the cgminer comments. =3
129 2012-11-05 02:06:20 <conman> and the submission code is https://github.com/ckolivas/cgminer/blob/gbt/cgminer.c#L2298
130 2012-11-05 02:07:17 <D34TH> https://github.com/bitcoin/bitcoin/pull/1984
131 2012-11-05 02:40:37 <Luke-Jr> conman: don't have it working yet, but 3 bugs found so far:
132 2012-11-05 02:40:45 <conman> nod
133 2012-11-05 02:40:50 <Luke-Jr> 1. you need to calc_midstate the GBT data, at least for some devices
134 2012-11-05 02:41:05 <conman> oh I thought I did
135 2012-11-05 02:41:08 <conman> must have forgotten it
136 2012-11-05 02:41:20 <Luke-Jr> 2. probably want to check for submit/coinbase support, and support sendign the whole block if you want to support solo mining
137 2012-11-05 02:41:33 <Luke-Jr> 3. not really a bug, but flip256 is a duplicate of swab256 in miner.h
138 2012-11-05 02:41:41 <conman> no it's not
139 2012-11-05 02:41:52 <conman> it only flips bytes within each uint32_t
140 2012-11-05 02:42:02 <conman> whereas swab256 flips bytes and reorders variables
141 2012-11-05 02:42:20 <Luke-Jr> ah, I see
142 2012-11-05 02:42:25 <Luke-Jr> you're right
143 2012-11-05 02:42:41 <conman> hmm so far the midstate lack is a big problem
144 2012-11-05 02:42:55 <conman> I just want to get the regular pool mining working first
145 2012-11-05 02:43:01 <conman> thanks so far
146 2012-11-05 02:47:01 <conman> come on share...
147 2012-11-05 02:47:25 <D34TH> ^story of my childhood
148 2012-11-05 02:47:34 <maaku> lol
149 2012-11-05 02:48:08 <Luke-Jr> conman: the last 32-bit of the prevblock header in submissions is something else
150 2012-11-05 02:48:24 <conman> yah but what...
151 2012-11-05 02:49:14 <conman> all the machinery for properly refreshing and queueing new work is not in place yet, so don't bother mentioning that
152 2012-11-05 02:49:46 <conman> there we go
153 2012-11-05 02:49:51 <conman> rejected
154 2012-11-05 02:50:32 <conman> ACTION checks wizkid057's page
155 2012-11-05 02:51:04 <Luke-Jr> conman: the only use of flip256 needs to flip all 20 sets of 32-bit ints, not only 256 bits worth
156 2012-11-05 02:52:17 <conman> hmm where'd all those zeroes suddenly come from?
157 2012-11-05 02:53:02 <conman> yeah that screwed something up
158 2012-11-05 02:53:25 <conman> for starters, there are only 80 bytes
159 2012-11-05 02:53:36 <conman> can't flip 128 bytes when there are only 80...
160 2012-11-05 02:54:54 <conman> heh but the function only flips 80 anyway
161 2012-11-05 02:55:05 <D34TH> seems like it needs a check
162 2012-11-05 02:55:23 <conman> nah, just rename it
163 2012-11-05 02:55:24 <Luke-Jr> 256 bits = 32 bytes
164 2012-11-05 02:56:29 <Luke-Jr> bad-txnmrklroot now that I'm flipping all 20 blocks of data
165 2012-11-05 02:57:59 <conman> well once again that may be my misunderstanding the docs on how to do it
166 2012-11-05 03:01:07 <Luke-Jr> conman: are you including the coinbase in the transactions? (it doesn't look like it, but I didn't dig deep to double-check)
167 2012-11-05 03:01:15 <conman> yes
168 2012-11-05 03:01:44 <Luke-Jr> I mean, in the merkle tree of them
169 2012-11-05 03:02:23 <Luke-Jr> ah, I see it
170 2012-11-05 03:02:29 <conman> list 1446
171 2012-11-05 03:02:32 <conman> line*
172 2012-11-05 03:10:16 <conman> wizkid057, well now the first 160 bytes are at least matching what your share counter is showing
173 2012-11-05 03:10:19 <conman> byteswapped
174 2012-11-05 03:10:28 <wizkid057> :D
175 2012-11-05 03:11:10 <conman> still nfi if it's supposed to be byteswapped
176 2012-11-05 03:11:18 <conman> but whatever...
177 2012-11-05 03:19:17 <Luke-Jr> conman: when you're copying the end of the coinbase transaction (after the inserted data), you use the new extended length for the source too
178 2012-11-05 03:19:34 <Luke-Jr> this is where the extra 4 octets is coming from
179 2012-11-05 03:19:41 <Luke-Jr> (still bad-txnmrklroot though???)
180 2012-11-05 03:19:50 <conman> am I not supposed to?
181 2012-11-05 03:20:14 <conman> am I supposed to use the original coinbase txn?
182 2012-11-05 03:20:52 <conman> there is nothing in the docs to make that clear
183 2012-11-05 03:21:52 <Luke-Jr> conman: I mean you're copying from 4 bytes late in the source coinbase txn
184 2012-11-05 03:21:57 <Luke-Jr> effectively skipping it
185 2012-11-05 03:23:25 <conman> strcat(gbt_block, work->gbt_coinbase);
186 2012-11-05 03:23:42 <conman> where is that 4 bytes late?
187 2012-11-05 03:24:07 <Luke-Jr> hex2bin(coinbase + 42 + *extra_len, pool->coinbasetxn + 84 + (*extra_len * 2), cbt_len - *extra_len - 42); <-- this is after *extra_len is incremented by 4
188 2012-11-05 03:24:59 <conman> the plus is ok, the minus is not, I see
189 2012-11-05 03:25:25 <Luke-Jr> out of the three *extra_len, only the first one should be incremented by 4
190 2012-11-05 03:25:49 <Luke-Jr> gen_hash(txn_bin, pool->txn_hashes + (32 * i), txn_len); <-- txn_len is the length in hex here, but we get the hash of the binary data, this hashes too much
191 2012-11-05 03:26:13 <Luke-Jr> now I am getting shares accepted, just a matter of interpreting the reply correctly
192 2012-11-05 03:26:49 <Luke-Jr> http://codepad.org/t84IMPz8 <-- all the changes I made to your git to get shares accepted, fwiw
193 2012-11-05 03:27:01 <conman> appreciate it, thanks
194 2012-11-05 03:27:04 <conman> honestly
195 2012-11-05 03:28:01 <Luke-Jr> np, it was a good challenge
196 2012-11-05 03:28:28 <Luke-Jr> regarding the response: null is the "accepted" response, and rejects are a string reason
197 2012-11-05 03:28:32 <conman> helps if you know what the code is actually supposed to do
198 2012-11-05 03:28:50 <conman> and you're just about the only person to have implemented the client side of gbt mining
199 2012-11-05 03:28:51 <wizkid057> good job guys :)
200 2012-11-05 03:28:54 <conman> since you invented it...
201 2012-11-05 03:29:49 <senseless> wtf
202 2012-11-05 03:29:57 <conman> wwf
203 2012-11-05 03:30:06 <senseless> I shut down bitcoind with bitcoind stop
204 2012-11-05 03:30:10 <senseless> ps ux make sure the executable is gone
205 2012-11-05 03:30:15 <senseless> copy the new executable
206 2012-11-05 03:30:27 <senseless> load bitcoind and i get "wallet.dat corrupt, salvage failed"
207 2012-11-05 03:30:28 <senseless> awesome
208 2012-11-05 03:31:41 <wizkid057> hopefully there were no coins there
209 2012-11-05 03:32:55 <senseless> 5
210 2012-11-05 03:33:13 <kjj_> what version to what version?
211 2012-11-05 03:33:21 <senseless> 0.7.0 -> 0.7.1
212 2012-11-05 03:33:24 <kjj_> and did you detach when you stopped?
213 2012-11-05 03:33:38 <senseless> i run it in daemon mode and access it from the console
214 2012-11-05 03:33:42 <senseless> thats why i issued bitcoind stop
215 2012-11-05 03:33:50 <senseless> then made sure the process was no longer there with ps ux
216 2012-11-05 03:34:48 <Luke-Jr> that doesn't detach
217 2012-11-05 03:34:50 <Luke-Jr> by default
218 2012-11-05 03:34:59 <kjj_> hmm, the wallet is always supposed to detach, from what I hear
219 2012-11-05 03:35:06 <kjj_> I always forget that too
220 2012-11-05 03:35:13 <Luke-Jr> well yes
221 2012-11-05 03:35:20 <kjj_> what do the logs show when you start it back up?
222 2012-11-05 03:35:29 <senseless> I ran a satoshi bot on this thing for awhile and my wallet.dat is like 500mbytes
223 2012-11-05 03:35:31 <senseless> does that matter?
224 2012-11-05 03:35:48 <kjj_> the last few lines of logs from the last time that you shut down 0.7.0 would have been useful too, but it might be difficult to get now
225 2012-11-05 03:37:56 <senseless> http://pastebin.com/e7mzyVyh
226 2012-11-05 03:38:19 <senseless> there is the successful shutdown of 0.7.0 and the boot up of 0.7.1
227 2012-11-05 03:39:06 <senseless> the user has full permissions to the .bitcoin and wallet.dat so why it couldnt make a copy is beyond me
228 2012-11-05 03:39:51 <D34TH> if i may inquire, openssl 0.9.8k
229 2012-11-05 03:39:59 <D34TH> i thought we were using 1.0.1c
230 2012-11-05 03:39:59 <senseless> centos
231 2012-11-05 03:40:05 <D34TH> oh
232 2012-11-05 03:42:23 <senseless> Is there anyway to recover anything from the wallet.dat ?
233 2012-11-05 03:42:55 <D34TH> did you back it up. pre-upgrade
234 2012-11-05 03:44:08 <senseless> ya
235 2012-11-05 03:44:19 <D34TH> restore the backup?
236 2012-11-05 03:44:28 <senseless> I could, but I want to know what happened :/
237 2012-11-05 03:44:29 <D34TH> nevermind, you probably tried that
238 2012-11-05 03:46:08 <kjj_> anything show up in the db.log?
239 2012-11-05 03:47:30 <senseless> http://pastebin.com/7a68zVX6
240 2012-11-05 03:47:56 <senseless> those are the last lines in the db.log
241 2012-11-05 03:49:03 <kjj_> have you tried clearing your .bitcoin directory? and by clearing, I mean renaming
242 2012-11-05 03:49:39 <senseless> To what end?
243 2012-11-05 03:49:53 <D34TH> mv .bitcoin .bitcoin.old
244 2012-11-05 03:49:54 <D34TH> ?
245 2012-11-05 03:50:04 <kjj_> well, those BDB errors worry me a bit. make me think the problem isn't the wallet.dat so much as something else
246 2012-11-05 03:50:47 <kjj_> and since I'm not sure what else you've tried, I'm thinking that a totally new directory, with no .dat files, and not .log files might be worth checking
247 2012-11-05 03:52:34 <conman> -\t\tunsigned char data[80];
248 2012-11-05 03:52:40 <senseless> So, use my backup wallet.dat with a freshly downloaded copy of the blockchain?
249 2012-11-05 03:52:48 <conman> if the header is only created from the first 80 bytes, why make it 128 Luke-Jr ?
250 2012-11-05 03:53:08 <Luke-Jr> conman: quick hack to flip more without redoing the flip function
251 2012-11-05 03:53:18 <conman> k
252 2012-11-05 03:53:19 <kjj_> senseless: with a newly created .bitcoin directory with nothing else in it, no blockchain at all, at least at first
253 2012-11-05 03:53:24 <conman> thought as much
254 2012-11-05 03:55:05 <senseless> Gives me the old, create the bitcoin.conf error with rpc user/pass settings in it
255 2012-11-05 03:55:54 <kjj_> doesn't it create that with a random password?
256 2012-11-05 03:55:54 <senseless> I creat bitcoin.conf and it loads and starts syncing right away
257 2012-11-05 03:56:07 <senseless> it just tells me to and gives me a random password
258 2012-11-05 03:56:44 <senseless> brand new wallet.dat, everything syncing
259 2012-11-05 03:56:49 <senseless> looks perfectly normal
260 2012-11-05 03:57:49 <senseless> oh screw me, i excluded the bitcoin directory for back up since it was so huge
261 2012-11-05 03:58:01 <kjj_> is the RPC system starting?
262 2012-11-05 03:58:24 <senseless> ya, rpc works fine
263 2012-11-05 03:58:44 <kjj_> do you happen to know one of your addresses? can you dumpprivkey on it?
264 2012-11-05 03:58:55 <senseless> well, im not using my wallet
265 2012-11-05 03:59:00 <senseless> should i copy the corrupted wallet.dat?
266 2012-11-05 03:59:12 <senseless> I don't know the privkeys unfortunately
267 2012-11-05 03:59:13 <kjj_> this is 0.7.1, right?
268 2012-11-05 03:59:16 <senseless> yes
269 2012-11-05 03:59:23 <senseless> i've already upgraded the bin to 0.7.1
270 2012-11-05 03:59:28 <kjj_> ok, use the "stop true" RPC command
271 2012-11-05 03:59:44 <senseless> it's stopped
272 2012-11-05 03:59:48 <senseless> no bitcoind running in ps ux
273 2012-11-05 03:59:51 <kjj_> when it stops, check your .log files
274 2012-11-05 04:00:13 <senseless> just a bunch of new blocks
275 2012-11-05 04:00:15 <senseless> what am i looking for
276 2012-11-05 04:00:48 <kjj_> er, sorry, I mean the log.xxxxxxxx files. how many and how big? and is there a "Bitcoin exited" line in the debug.log?
277 2012-11-05 04:01:28 <senseless> Yes, same cleanly exited line i saw before
278 2012-11-05 04:01:35 <senseless> (when i exited in 0.7.0)
279 2012-11-05 04:01:48 <senseless> 1 log file 10mbytes in size
280 2012-11-05 04:01:58 <senseless> i didnt leave bitcoin syncing for very long i shut it down pretty fast
281 2012-11-05 04:02:10 <kjj_> and you used "bitcoind stop true" when you shut down?
282 2012-11-05 04:02:17 <senseless> yep
283 2012-11-05 04:02:32 <kjj_> ok. now try copying your wallet.dat in
284 2012-11-05 04:04:14 <senseless> oh sweet it worked
285 2012-11-05 04:04:39 <senseless> Or not, it's not showing my original addresses
286 2012-11-05 04:04:41 <kjj_> now try using "bitcoin dumpprivkey 1SOMEADDRESS"
287 2012-11-05 04:05:10 <kjj_> hmm, not showing your addresses? check your logs again
288 2012-11-05 04:05:28 <senseless> I copied the wrong wallet.dat file er
289 2012-11-05 04:05:36 <kjj_> heh
290 2012-11-05 04:07:30 <senseless> it doesnt want anything to do with the wallet now heh
291 2012-11-05 04:07:40 <senseless> it just keeps renaming it to .bak and exiting
292 2012-11-05 04:08:37 <kjj_> log shows salvage found no records?
293 2012-11-05 04:10:59 <senseless> taking awhile to manually run salvage
294 2012-11-05 04:12:58 <kjj_> hopefully that is a good sign
295 2012-11-05 04:13:45 <kjj_> salvage still opens the wallet.dat in the context of the database. what I think happened is that your other DB files were trashed
296 2012-11-05 04:14:23 <senseless> What happens when you run a slavage, does it rename the file and then rebuild it from the original?
297 2012-11-05 04:14:27 <senseless> Or how does that work?
298 2012-11-05 04:14:46 <senseless> what is this .bak?
299 2012-11-05 04:15:00 <senseless> I'm trying to recover again from the same .bak and it's not wanting to process the file
300 2012-11-05 04:15:02 <kjj_> it does a rename in the context of the database environment (not merely a filesystem rename), and then scans it
301 2012-11-05 04:15:19 <senseless> how would i make the bitcoind rescan the .bak?
302 2012-11-05 04:16:08 <senseless> oh wait there it goes
303 2012-11-05 04:16:11 <senseless> took it awhile
304 2012-11-05 04:16:30 <kjj_> there doesn't seem to be any way to do it on a file other than wallet.dat, and the output always goes into wallet.dat, so you can't use it to merge
305 2012-11-05 04:16:50 <senseless> but isn't the .bak a wallet.dat file?
306 2012-11-05 04:17:28 <kjj_> yeah, I'm just saying you can't do it on two different versions and get the keys into one file
307 2012-11-05 04:18:06 <senseless> What do you mean by 2 different versions? (0.7.1 and 0.7.0) ?
308 2012-11-05 04:18:17 <senseless> Should i have reverted back to 0.7.0?
309 2012-11-05 04:18:31 <kjj_> I mean that if you have two broken wallet.dat files from different times, you can't try to get keys out of both of them into the one file
310 2012-11-05 04:18:53 <senseless> Ah, ya, I'm just trying to get it to re-read the same .bak and create another wallet.dat from the .bak
311 2012-11-05 04:19:18 <kjj_> you can do that, but you'll end up with two different wallet.dat files.
312 2012-11-05 04:19:32 <kjj_> just copy out the good copy, and rename one of the .bak files to .dat
313 2012-11-05 04:20:24 <senseless> It's weird because everytime I run it I get different resulting .dat file sizes, with different keys in them
314 2012-11-05 04:20:35 <senseless> Well, 1 key (I'm guessing some default created address)
315 2012-11-05 04:21:01 <kjj_> after the salvage finishes, I think you need to stop, and then start with -rescan
316 2012-11-05 04:22:28 <senseless> I'll let it finish syncing then rescan and see what happens
317 2012-11-05 04:22:35 <senseless> sucks, pool is basically out of business now :(
318 2012-11-05 04:22:43 <kjj_> which pool?
319 2012-11-05 04:22:54 <senseless> nothing public just my private pool, and by out of business i mean offline
320 2012-11-05 04:27:44 <Luke-Jr> conman: longpollid is almost certain to change every longpoll
321 2012-11-05 04:28:07 <conman> oh
322 2012-11-05 04:28:10 <conman> hrm
323 2012-11-05 04:28:21 <conman> gotta prevent a race on it is all..
324 2012-11-05 04:28:24 <conman> will fix
325 2012-11-05 04:28:41 <conman> at least it's accepting shares now
326 2012-11-05 04:34:49 <senseless> well weird thing kjj
327 2012-11-05 04:34:54 <senseless> i went back to using the .bitcoin2
328 2012-11-05 04:35:00 <senseless> renamed it back to .bitcoin
329 2012-11-05 04:35:07 <senseless> ran the salvage again and it successfully recovered everything
330 2012-11-05 04:35:14 <senseless> and wallet is back to operating normally
331 2012-11-05 04:35:25 <senseless> Dunno
332 2012-11-05 07:18:16 <conman> [2012-11-05 19:17:20] Rejected df390d86 Diff 1/1 GPU 1 pool 2 (high-hash)
333 2012-11-05 07:18:25 <conman> I get a lot of those
334 2012-11-05 07:18:27 <conman> I wonder ytf
335 2012-11-05 07:18:39 <abrkn> ;;ident abrkn
336 2012-11-05 07:18:40 <gribble> Nick 'abrkn', with hostmask 'abrkn!~pialur@195.159.164.228', is not identified.
337 2012-11-05 07:30:35 <conman> I'm guessing yet more byteswap fuckage
338 2012-11-05 08:05:47 <abrkn> can someone explain why listtransactions doesn't return "move" transactions in the proper order? my c is horrible, tried to work it out yesterday
339 2012-11-05 09:24:34 <rebroad> 0.6.3-beta is stuck at block 187352 now since a few days...
340 2012-11-05 09:25:40 <an3k> why don't you use the latest version?
341 2012-11-05 09:25:42 <sturles> 0.6.3-beta is an old beta version. Upgrade.
342 2012-11-05 09:26:07 <jouke> you should upgrade, 0.6.3 is an old version.
343 2012-11-05 09:26:08 <jouke> ;P
344 2012-11-05 09:26:18 <rebroad> if the latest version is needed, why isn't the client telling me?
345 2012-11-05 09:26:42 <rebroad> I thought it was supposed to alert when there was a needed upgrade..
346 2012-11-05 09:27:02 <rebroad> I'm uneasy about the apparent constant need to upgrade
347 2012-11-05 09:27:16 <jouke> No, you shouldn't need to upgrade.
348 2012-11-05 09:27:29 <jouke> I also use 0.6.3 still
349 2012-11-05 09:28:01 <rebroad> it saw a block 187353 that was invalid, and somehow it's got stuck
350 2012-11-05 09:28:48 <rebroad> i was wondering whether i should file a bug report for it, in case it's still in the latest version
351 2012-11-05 09:29:19 <rebroad> I've seen nothing online about getting stuck on block 187352, which I would have expected if it affected other people
352 2012-11-05 09:31:18 <rebroad> block 00000000000000b68ab3 was the invalid one
353 2012-11-05 09:31:52 <rebroad> no google results for it though!
354 2012-11-05 09:32:52 <rebroad> according to blockexplorer that block should be valid... http://blockexplorer.com/block/00000000000000b68ab313eab5014ab00ea45095e4c27296add1c322f8b1c1e1
355 2012-11-05 09:33:31 <rebroad> so,.. I'm confused how other people are still keeping up to date and using 0.6.3
356 2012-11-05 09:34:50 <rebroad> it's Bitcoin version v0.6.3-beta (2012-06-19 16:44:55 -0400) that I'm using
357 2012-11-05 09:35:52 <rebroad> hmmm. that shows up nothing in Google either... is this some really rare version I'm using,....?!
358 2012-11-05 10:05:25 <senseless> its kinda old
359 2012-11-05 10:05:30 <senseless> 0.7.1 is latest
360 2012-11-05 10:05:33 <senseless> 0.7.0 has been out for months
361 2012-11-05 10:11:57 <gmaxwell> rebroad: The reason you saw it invalid was probably due to database corruption, not due to running 0.6.3.
362 2012-11-05 10:52:56 <xenland> What is the RPC command to look up the transactions in the block index?
363 2012-11-05 10:53:25 <andyrossy> getblock
364 2012-11-05 10:53:28 <andyrossy> then decodehash
365 2012-11-05 10:53:34 <andyrossy> then decoderawtransaction
366 2012-11-05 10:53:59 <xenland> Excellent I'm going to try that out :D thanks
367 2012-11-05 10:55:20 <xenland> Is decodehash found in 7.1?
368 2012-11-05 10:56:40 <gmaxwell> there is no 'decodehash'
369 2012-11-05 10:57:18 <gmaxwell> just run decoderawtransaction <txid> 1
370 2012-11-05 10:58:36 <xenland> I'm under the impression that txid isn't the same as a block hash
371 2012-11-05 10:58:49 <xenland> or are they?
372 2012-11-05 11:02:41 <gmaxwell> xenland: the txid is not the block id, no.
373 2012-11-05 11:04:08 <xenland> Okay, when I do getblockindex 0 I get the hash of the genesis block, I copied that hash and do getblock *hash here* but i get JSON data... what do i do with this json data to get the TX list
374 2012-11-05 11:05:00 <xenland> I'm trying to figure out how to verify blocks
375 2012-11-05 11:06:17 <kjj_> the tx field is a list of all of the transactions in that block
376 2012-11-05 11:07:17 <kjj_> you can look them up with getrawtransaction
377 2012-11-05 11:07:56 <kjj_> except for 4a5e1e4baab89f3a32518a88c31bc87f618f76673e2cc77ab2127b7afdeda33b. You can't fetch that one
378 2012-11-05 11:10:05 <xenland> thanks kjj your my saviour!
379 2012-11-05 11:10:11 <xenland> More bitcents for you
380 2012-11-05 11:12:12 <kjj_> heh. you don't have to pay me every time I answer a question. :)
381 2012-11-05 11:12:26 <xenland> heh okay
382 2012-11-05 11:12:37 <xenland> but still... YTMD
383 2012-11-05 11:15:28 <xenland> and the geneisis block dosen't return any information when using getrawtransaction and thats because there is no inputs correct?
384 2012-11-05 11:18:16 <kjj_> no, just because that block is a quirk in the code
385 2012-11-05 11:18:48 <kjj_> the genesis transaction is unspendable because it never gets added to the index
386 2012-11-05 11:21:04 <kjj_> for more info on the oddness of block 0, see this thread: https://bitcointalk.org/index.php?topic=119530.0;all
387 2012-11-05 11:43:22 <rebroad> gmaxwell I've upgraded to a later version and it's working ok now.
388 2012-11-05 11:43:54 <rebroad> although I'm noticing a new "bug" in the latest version - the date format is wrong. it's showing MM/DD/YYYY rather than DD/MM/YYYY
389 2012-11-05 11:46:15 <abrkn> can someone explain why listtransactions doesn't return "move" transactions in the proper order? my c is horrible, tried to work it out yesterday
390 2012-11-05 11:55:44 <abrkn> is there a mechanism for suggesting features with a bounty? i'm specifically looking for an optional parameter to the move command so that it guarantees the source account not ending up with a minus
391 2012-11-05 12:11:10 <xenland> abrkn: Why do you need to use move? Whats wrong with a DB?
392 2012-11-05 12:16:07 <abrkn> why use a db when there is move?
393 2012-11-05 12:22:27 <xenland> abrkn: I never experience a "Possibiliity" of a random movment of money happen when I use a database. It sounds like you are in need of a guarantee with your financial information something "move" isn't exactly for from what it sounds like
394 2012-11-05 12:22:36 <xenland> Why not do the math before you do a "move" command?
395 2012-11-05 12:25:57 <kjj_> yeah, do the math in advance
396 2012-11-05 12:28:28 <kjj_> I've only found one situation where the move command is really useful. if I'm collecting multiple small transactions into one bigger transaction to the same address, I use move to reduce the destination account so that it accurately shows the real incoming amount
397 2012-11-05 12:30:10 <xenland> Is decoderawtransaction straight hex to ascii? or is there some other proccess involved? my work dosen't seem to be checking correctly
398 2012-11-05 12:30:47 <xenland> oh wait it seems its serialized aswell
399 2012-11-05 12:31:29 <abrkn> well, if i use a db then i need to keep updating it when it's credited with btc
400 2012-11-05 12:31:39 <abrkn> so need to monitor deposit addresses for every user
401 2012-11-05 12:31:51 <abrkn> and credit their db account when they receive money
402 2012-11-05 12:32:01 <abrkn> sounds like work...
403 2012-11-05 12:36:04 <xenland> Financial websites get hacked for lazyness and negligence your going to get hacked if you don't do your homework and do redundant checking of your data. Perhaps you could have everything sent to the correct addresses and then do "Getrecievedfrom"
404 2012-11-05 12:37:49 <xenland> abrkn: Make sure your aware of floating point precision in-accuracies as well other wise you'll lose people money/Bitcoins. Its best to do all calculations in satoshi and make the output/display to the user with decimals
405 2012-11-05 12:39:29 <abrkn> xenland: ok. and what method is good for knowing when a user's account has been credited?
406 2012-11-05 12:40:29 <xenland> getrecievedbyaddress <bitcoinaddress> [minimum confirmations]
407 2012-11-05 12:40:43 <abrkn> but that would require me to poll every user's receive address?
408 2012-11-05 12:41:02 <xenland> Yes i believe so
409 2012-11-05 12:41:09 <abrkn> hmm ok
410 2012-11-05 12:41:28 <xenland> Is polling an issue?
411 2012-11-05 12:41:44 <abrkn> not really
412 2012-11-05 12:41:59 <abrkn> just pretty psyched to get my game out there but all this accounting slowing me down
413 2012-11-05 12:42:08 <abrkn> but you're right, better safe than sorry
414 2012-11-05 12:42:44 <xenland> oooo a game you say?
415 2012-11-05 12:43:05 <abrkn> classic video poker
416 2012-11-05 12:43:07 <abrkn> http://i.imgur.com/pcxv7.jpg
417 2012-11-05 12:43:28 <xenland> nice
418 2012-11-05 12:43:34 <kjj_> abrkn: I just submitted a patch that lets you specify a script to get run any time a transaction hits your wallet
419 2012-11-05 12:43:57 <abrkn> kij: nice, ive been using blockchain.info to monitor addresses
420 2012-11-05 12:44:24 <xenland> kjj_: nice!
421 2012-11-05 12:44:29 <kjj_> it doesn't do all of the work for you, and it'll trigger even on unconfirmed transactions
422 2012-11-05 12:45:05 <kjj_> but then you'll be able to stash the TxIDs in your database and poll just those until they get enough confirmations, rather than having to poll each address or whatever
423 2012-11-05 12:46:34 <abrkn> xenland: i will use what's called a 9-6 payout table which pays 99.543904% back and 0.45% goes to a progressive jackpot (leaving house edge at 0.006096% which i think is quite fair)
424 2012-11-05 12:46:49 <abrkn> 0.0045% i mean
425 2012-11-05 12:47:02 <abrkn> err, im getting mixed up. nto used to %s :)
426 2012-11-05 12:47:20 <abrkn> original game edge is 0.99543904, jackpot contribution is 0.0045
427 2012-11-05 12:47:48 <xenland> nice! I've been looking for games to show my non-techie friends to play
428 2012-11-05 12:47:51 <xenland> good to see more games
429 2012-11-05 12:47:59 <xenland> for winning bitcoins that is
430 2012-11-05 12:48:01 <abrkn> and the nice thing is, because the jackpot is progressive
431 2012-11-05 12:48:06 <abrkn> you should write a bot that plays when jackpot is high
432 2012-11-05 12:48:11 <abrkn> positive edge very fast =)
433 2012-11-05 12:48:50 <abrkn> http://wizardofodds.com/games/video-poker/strategy/jacks-or-better/9-6/optimal/ <-- bot recipe
434 2012-11-05 12:51:52 <xenland> How do i deserialise the getrawtransaction output?
435 2012-11-05 12:52:05 <kjj_> deoderawtransaction
436 2012-11-05 12:52:19 <kjj_> or you can parse it yourself. In your thread, I gave you links
437 2012-11-05 12:52:33 <kjj_> https://en.bitcoin.it/wiki/File:TxBinaryMap.png
438 2012-11-05 12:52:34 <xenland> oh okay, I'm sure the answer is in those links thanks mate
439 2012-11-05 12:53:28 <Icoin> hi guys what is the timeframe for libdb5.2 and qt 5 integration? BTC aswell as DVC works simultanously on RPI https://bitcointalk.org/index.php?topic=101559.msg1316721#msg1316721
440 2012-11-05 12:53:34 <Icoin> hi xenland :)
441 2012-11-05 12:53:47 <Icoin> good to see you bro
442 2012-11-05 12:54:13 <abrkn> btc to satoshi is what? x * 1e8?
443 2012-11-05 12:54:36 <jurov> yes
444 2012-11-05 12:55:00 <xenland> 100000000 = 1 Bitcoin
445 2012-11-05 12:55:34 <xenland> Oh hey icoin
446 2012-11-05 12:55:38 <xenland> long time no talk
447 2012-11-05 13:06:00 <slush> https://bitcointalk.org/index.php?topic=122438.0
448 2012-11-05 13:08:17 <kinlo> slush: nice
449 2012-11-05 13:08:36 <xenland> excellent mate!
450 2012-11-05 13:08:41 <kinlo> slush: however, can it be made usefull for protecting servers/hot wallets too?
451 2012-11-05 13:09:06 <kinlo> ie, is the signing capable of for example, signing at the most 10 bitcoins per day or something?
452 2012-11-05 13:09:08 <slush> kinlo: unless you want to press the button everytime server will make outgoing transaction, then probably no
453 2012-11-05 13:09:32 <kinlo> it depends on how much logic is inside the chip...
454 2012-11-05 13:09:42 <slush> kinlo: but implementing cold storage for your backup will be very easy with this
455 2012-11-05 13:09:51 <kinlo> true
456 2012-11-05 13:09:51 <slush> for your server, of course
457 2012-11-05 13:11:28 <slush> currently I'm very affraid that I'll make some mistake during handling cold storage for the pool. One mistake and I can burn all that coins easily. Or I can forget keys somewhere on the computer where I handle it. With wallet like this, I don't need to load cold storage wallet to any machine at all..
458 2012-11-05 13:11:54 <abracadabra> what if you lose the device?
459 2012-11-05 13:11:55 <abracadabra> :p
460 2012-11-05 13:12:01 <abracadabra> hehe
461 2012-11-05 13:12:17 <slush> abracadabra: Device itself will ask you for doing paper backup of the initial seed
462 2012-11-05 13:12:28 <slush> abracadabra: you'll be able to load that seed to computer or to another device
463 2012-11-05 13:12:45 <slush> but displaying seed again will be improssible once initialization is done...
464 2012-11-05 13:12:53 <abracadabra> i'm sure you got it covered slush
465 2012-11-05 13:13:00 <abracadabra> j/k around.. humans have a tendency to lose things
466 2012-11-05 13:13:08 <abracadabra> ;)
467 2012-11-05 13:21:39 <kjj_> slush: how do you communicate back and forth?
468 2012-11-05 13:22:07 <slush> kjj_: It will act as a generic USB HID device
469 2012-11-05 13:22:24 <kjj_> don't take this the wrong way, but fail. :)
470 2012-11-05 13:23:11 <kjj_> unless you are bit banging the USB protocol yourself, your wallet should not be a USB device
471 2012-11-05 13:23:19 <slush> kjj_: why?
472 2012-11-05 13:24:51 <kjj_> safety
473 2012-11-05 13:25:37 <slush> why?
474 2012-11-05 13:26:04 <kjj_> do you know that whatever system you are using doesn't have any back doors or debugging modes?
475 2012-11-05 13:26:27 <kjj_> unless you wrote it yourself using GPIO pins, the answer is probably no
476 2012-11-05 13:27:13 <slush> kjj_: how can it be used for attack?
477 2012-11-05 13:27:25 <kjj_> which means that you don't really know what is going to happen when you plug it in. malware might know about a backdoor that lets it dump out the seed and send it to whoever owns that computer
478 2012-11-05 13:27:53 <slush> "dump out the seed" ? :-)
479 2012-11-05 13:28:17 <slush> kjj_: from the announcement: "* Impossibility to obtain private keys from the device"
480 2012-11-05 13:28:55 <kjj_> heh. and I'm telling you that if you are using a USB interface that you didn't write yourself on a device that complex, you don't have any idea if it is possible or not. you hope it isn't, you claim it isn't, but you don't KNOW.
481 2012-11-05 13:29:11 <slush> eh
482 2012-11-05 13:29:47 <slush> kjj_: I'm not sure if you catched the idea of that device
483 2012-11-05 13:30:54 <xenland> I think it depends on the specs some hardware is built with security in mind
484 2012-11-05 13:31:01 <slush> kjj_: desktop client will ask for master public key, then you can see your wallet on the desktop. When you want to send transaction, desktop client will knock up transaction template and send it over USB to the device. Device then check the transaction, display information (mostly target address and amount of transfer) on the display. And will ask user to confirm it by pressing button
485 2012-11-05 13:31:02 <kjj_> it is a linux box that you didn't write, running on hardware that you didn't design. I'm not sure if you understood the implications of those factors
486 2012-11-05 13:31:25 <slush> kjj_: it's not linux box, it is custom software burned directly to the chip
487 2012-11-05 13:31:43 <kjj_> oh, so not the Raspberry Pi in your announcement?
488 2012-11-05 13:32:10 <slush> and that DYI version will be just an application which you can install on clean (trusted) raspberry pi distro
489 2012-11-05 13:32:16 <slush> kjj_: that DYI version, yes
490 2012-11-05 13:33:04 <slush> or you can compile your own linux kernel, if you don't trust raspbian
491 2012-11-05 13:33:06 <kjj_> sec, gotta run, back in a few. but unless you wrote the USB stack yourself, you don't know what is in there.
492 2012-11-05 13:33:25 <slush> no, there's no possible attack vector over malicious USB drivers
493 2012-11-05 13:36:32 <xenland> Is the ability to process a Script required to verify Blocks in the block chain?
494 2012-11-05 13:38:17 <Luke-Jr> yes
495 2012-11-05 13:38:48 <kjj_> slush: ha!
496 2012-11-05 13:39:09 <slush> kjj_: well, of course you must install trusted distro to that RPi. But firstly it is "community version" of the product and also there's no reason to think that raspberian is less secure than other linux distributions or even Windows machines.
497 2012-11-05 13:39:50 <helo> other linux distros and windows machines very likely have USB stack exploits... they have in the past, at least
498 2012-11-05 13:39:57 <kjj_> also, there are almost certainly debug modes burned right into the USB chip, even if the code is clean
499 2012-11-05 13:40:35 <Luke-Jr> slush: hey, I'll get an answer to your question in a sec ;)
500 2012-11-05 13:40:53 <kjj_> for a truly secure wallet, you should be using something more transparent than USB. that's always been my opinion, but I hope your device works well. the world surely needs these
501 2012-11-05 13:43:19 <xenland> perhaps wifi module so you can specifically send out packets exactly how you want with little to none reverse penetration attack vectors
502 2012-11-05 13:45:00 <xenland> I have faith in slush though
503 2012-11-05 13:45:17 <xenland> just putting my two cents in with what i've been learning about hardware the past year or so
504 2012-11-05 13:47:10 <kjj_> heh. I always think back to this paper: http://cm.bell-labs.com/who/ken/trust.html and then I add "Stage IV" where the attack is embedded in a chip rather than a binary
505 2012-11-05 13:47:49 <xenland> i think i saw something like this on a one of those hardware hacking convention videos on youtube
506 2012-11-05 13:47:57 <slush> helo: "very likely have USB stack exploits" - then just buy that custom hardware, there won't be any malicious linux drivers (there won't be linux at all)
507 2012-11-05 13:50:54 <helo> tough to know if you're safe or not without auditing/authoring the code, as kjj_ said
508 2012-11-05 13:52:45 <xenland> slush: Is this machine code?
509 2012-11-05 13:52:46 <kjj_> when I was thinking of doing this, my plan was to do straight up RS-232 serial with a software implementation. impossible to sneak anything in, unless the chip designer was psychic and knew which GPIO pins I was going to use well before I did
510 2012-11-05 13:53:28 <helo> it's a little on the paranoid side of caution, but when there may be epic attacks mounted to steal money, if there is a hole it will probably be found
511 2012-11-05 13:53:36 <kjj_> you could even embed a USB to serial adapter in the device so that it plugged in as USB, but still safe because if had a physical layer that was too dumb for mischief
512 2012-11-05 13:56:37 <slush> kjj_: what's the difference between serial-over-usb and HID?
513 2012-11-05 13:56:47 <slush> kjj_: I really don't see your problem.
514 2012-11-05 13:57:48 <xenland> The unknowingness of the USB communication is the issue i believe
515 2012-11-05 13:58:22 <xenland> with out knowing the exact method, its difficult to trust i suppose, but who trusts more then a few Bitcoins on a stick its more for On-the-go or shopping style things like a "wallet"
516 2012-11-05 13:58:42 <xenland> I doubt someone be holding a million BTC on your stick thing (No offence)
517 2012-11-05 13:59:03 <xenland> so not a big issue indeed
518 2012-11-05 13:59:09 <slush> I don't see any problem with USB at all.
519 2012-11-05 13:59:17 <slush> Even if USB stack is hacked - so what?
520 2012-11-05 13:59:49 <xenland> Run commands that bypass your softwares security possibly
521 2012-11-05 13:59:53 <xenland> my guess
522 2012-11-05 14:00:27 <xenland> You said this wasn't linux is this like arduino where its its own machine code and you program the logic into it?
523 2012-11-05 14:00:47 <drizztbsd> firewire is worst
524 2012-11-05 14:00:55 <drizztbsd> with firewire you can read memory :P
525 2012-11-05 14:01:06 <xenland> drizztbsd: good to know
526 2012-11-05 14:02:01 <slush> what security you can bypass by running command? :)
527 2012-11-05 14:03:40 <xenland> Read privatekeys, steal BTC
528 2012-11-05 14:05:25 <slush> are we talkin about DYI version or about that custom design?
529 2012-11-05 14:05:49 <slush> as I said, DYI on Raspberry will be as safe as you can install trusted software on your RPi
530 2012-11-05 14:07:51 <slush> that custom device is lowlevel code (no operating system) and USB stack is done by underlying hardware
531 2012-11-05 14:14:25 <xenland> ah, well you have my trust in that case then, can't speak for kjj_ though
532 2012-11-05 15:10:16 <Luke-Jr> slush: 65416 bytes before the stratum connection goes idle
533 2012-11-05 15:11:18 <slush> Luke-Jr: it happen everytime to you?
534 2012-11-05 15:11:22 <Luke-Jr> slush: yes
535 2012-11-05 15:11:27 <slush> hm, weird
536 2012-11-05 15:12:18 <Luke-Jr> slush: Wireshark tells me that it stops when the TCP window is full
537 2012-11-05 15:13:26 <slush> Luke-Jr: I tested it again in telnet. another machine, another network, cannot simulate your problem...
538 2012-11-05 15:13:40 <slush> can you try it on another machine?
539 2012-11-05 15:15:22 <Luke-Jr> hmm, socat finishes fine
540 2012-11-05 15:24:25 <Luke-Jr> slush: seems maybe this problem is in Con's socket code, I'll have to debug it more on my end
541 2012-11-05 15:25:52 <slush> if it works in telnet for you, then this is likely
542 2012-11-05 15:32:04 <xenland> Where do i find the command list for requesting blocks to download from peers?
543 2012-11-05 15:38:36 <BlueMatt> anyone who doesnt have dnsseed.bluematt.me NS cached in their nameserver: does looking up dnsseed.bluematt.me work for you?
544 2012-11-05 15:41:57 <jgarzik> BlueMatt: yes, it works
545 2012-11-05 15:43:08 <BlueMatt> nice, now I can turn it off on jenkins
546 2012-11-05 16:48:18 <jgarzik> gmaxwell, sipa: ultraprune ready for mining? I would like to upgrade my testnet3 node to HEAD, and wondered if the "don't use for mining" proviso still stood?
547 2012-11-05 16:48:46 <jgarzik> and
548 2012-11-05 16:49:59 <jgarzik> random_cat: my inbox has an emailed notification about a MeetUp of "IBD enthusiasts"
549 2012-11-05 16:50:08 <jgarzik> ACTION kicks xchat autocorrect
550 2012-11-05 16:50:13 <jgarzik> Random : my inbox has an emailed notification about a MeetUp of "IBD enthusiasts"
551 2012-11-05 16:52:19 <helo> how can anyone be enthusiastic of either interpretation of IBD?
552 2012-11-05 17:07:48 <gmaxwell> jgarzik: for testnet? sure.
553 2012-11-05 17:09:10 <jgarzik> gmaxwell: Just wanted to double-check. A couple weeks ago, there was panic, making sure there were some non-ultraprune miners on testnet ;p
554 2012-11-05 17:09:22 <gmaxwell> jgarzik: yes, that bug is fixed.
555 2012-11-05 17:11:14 <sipa> jgarzik: there is one potential bug, which would cause switching to a later block at the same height sometimes
556 2012-11-05 17:11:25 <sipa> but that doesn't prohibit mining
557 2012-11-05 17:13:19 <sipa> jgarzik: nonetheless, i prefer having the warning in the code until we hit rc's or later... if a significant number of miners switch, and there is still a bug...
558 2012-11-05 17:13:32 <jgarzik> ok
559 2012-11-05 17:14:35 <sipa> being hypocritical myself, i've been mining on ultraprune for over a month now...
560 2012-11-05 17:38:28 <Tykling> is there a way to use the rpc interface to list the unconfirmed balance of an account ? like the gui client does ?
561 2012-11-05 17:49:46 <random_cat> jgarzik: what am i to make of this?
562 2012-11-05 17:49:55 <jgarzik> random_cat: ignore
563 2012-11-05 17:50:15 <jgarzik> random_cat: I typed "Random: " and xchat changed it automatically to "random_cat: "
564 2012-11-05 17:50:18 <random_cat> oic...
565 2012-11-05 17:52:17 <kjj_> another reason to hate automatic "correction" features
566 2012-11-05 18:13:10 <an3k> ;;seen tenakha
567 2012-11-05 18:13:11 <gribble> I have not seen tenakha.
568 2012-11-05 18:13:20 <an3k> oh for sure you have
569 2012-11-05 18:13:54 <an3k> oh, sry, wrong channel :)
570 2012-11-05 18:32:08 <BlueMatt> anyone wanna propose a better name for bloom filter seed than nSeedTwiddleFactor? :)
571 2012-11-05 18:32:22 <sipa> nSeedSeed
572 2012-11-05 18:32:31 <sipa> :p
573 2012-11-05 18:33:05 <sipa> nSalt ?
574 2012-11-05 18:33:06 <gmaxwell> nTweak
575 2012-11-05 18:33:23 <sipa> ^ that
576 2012-11-05 18:33:51 <BlueMatt> nSeedTweak
577 2012-11-05 18:33:52 <gmaxwell> (tweak seems to be the word used for large block ciphers for basically this)
578 2012-11-05 18:34:21 <BlueMatt> ok
579 2012-11-05 18:35:33 <jgarzik> is it possible for testnet3 bitcoin addresses to start with 'n', as well as 'm', yes?
580 2012-11-05 18:35:53 <jgarzik> *it is
581 2012-11-05 18:35:53 <sipa> correct
582 2012-11-05 18:35:55 <gmaxwell> Yes.
583 2012-11-05 18:37:22 <sipa> i once saw a failed -reindex some time ago, which i didn't understand at all
584 2012-11-05 18:37:50 <sipa> i now realize i may have compiled jgarzik's reindex instead of my reindex2...
585 2012-11-05 18:49:56 <Traffic> bitcoin-qt makes a lot of traffic up to 4 GB per day?
586 2012-11-05 18:50:36 <gmaxwell> Traffic: The traffic is normall quite low, especially if you're not listening??? of course there is a fair amount during the initial sync.
587 2012-11-05 18:50:36 <sipa> if someone downloads the entire block chain from you, yes
588 2012-11-05 18:51:53 <sipa> or when you dowbload the chain from somewhere, sure
589 2012-11-05 18:52:01 <Traffic> 18 GB traffic in 5 days - including normal surfing 5GB per month
590 2012-11-05 18:52:35 <Traffic> i already have the chain - - the wallet shows up to 44 conections
591 2012-11-05 18:53:03 <sipa> right, you're likely distributing the chain to many people
592 2012-11-05 18:53:04 <gmaxwell> Right, so you're listening and other new nodes are pulling it from you.
593 2012-11-05 18:53:35 <Traffic> ok then i petter stop listening - how can i do ?
594 2012-11-05 18:53:44 <Traffic> better
595 2012-11-05 18:57:30 <ThomasV_> bitcoin-qt is not a good name
596 2012-11-05 18:59:00 <Traffic> how can i stop my wallet from listening so that traffic decrease ?
597 2012-11-05 19:00:24 <sipa> Traffic: -nolisten or -maxconnections may help you
598 2012-11-05 19:00:48 <sipa> maybe those can be set in the config window, not sure
599 2012-11-05 19:00:56 <gavinandresen> Traffic: run with -listen=0 or put listen=0 in your bitcoin.conf file. See https://en.bitcoin.it/wiki/Running_Bitcoin
600 2012-11-05 19:02:10 <kjj_> does -maxconnections=0 mean no connections, 8 connections, or unlimited?
601 2012-11-05 19:02:25 <BlueMatt> Traffic: my seednode almost never uses any significant amount of traffic
602 2012-11-05 19:02:40 <BlueMatt> Traffic: a few GB in a few days - Ive never seen that much
603 2012-11-05 19:03:08 <BlueMatt> kjj_: 8 iirc, but you should use nolisten
604 2012-11-05 19:04:34 <sipa> kjj_: -maxconnections is the total
605 2012-11-05 19:04:40 <sipa> both incoming and outgoing
606 2012-11-05 19:04:42 <Traffic> thx
607 2012-11-05 19:05:06 <BlueMatt> sipa: iirc it will still make the 8 outgoing though
608 2012-11-05 19:05:15 <sipa> quite sure it won't :)
609 2012-11-05 19:05:32 <BlueMatt> sipa: oh, I wasnt looking at code, I just remember it ignoring max connections for outgoing, maybe Im wrong
610 2012-11-05 19:06:08 <sipa> i haven't looked at the code either
611 2012-11-05 19:06:15 <BlueMatt> Traffic: even listening, you should not see any really measurable bw usage (unless you are limited to like 2GB/mo, then it may matter)
612 2012-11-05 19:06:20 <sipa> but i remember seeing something "if maxconnections < 8 ..."
613 2012-11-05 19:06:31 <BlueMatt> sipa: oh, well then maybe Im wrong
614 2012-11-05 19:06:39 <BlueMatt> or...maybe someone should look at code ;)
615 2012-11-05 19:07:01 <sipa> int nMaxOutbound = min(MAX_OUTBOUND_CONNECTIONS, (int)GetArg("-maxconnections", 125));
616 2012-11-05 19:07:24 <BlueMatt> mmm, ok, I stand corrected
617 2012-11-05 19:07:34 <sipa> now, less funny: my mining node segfaulted
618 2012-11-05 19:07:50 <BlueMatt> :(
619 2012-11-05 19:08:14 <sipa> it was running git head of a few days ago
620 2012-11-05 19:09:04 <sipa> BlueMatt: does pulltester have a test which verifies that when receiving two blocks of equal work, the first one is retained?
621 2012-11-05 19:09:53 <BlueMatt> you mean connectedBlock -> A, connectedBlock -> B check that the new connectedBlock is the first one of A/B sent
622 2012-11-05 19:10:07 <sipa> indeed
623 2012-11-05 19:10:23 <BlueMatt> IIRC, yes, gimme a sec...
624 2012-11-05 19:11:03 <BlueMatt> yes, many, many times (its just a part of various tests as it tests new chain head after each block)
625 2012-11-05 19:11:12 <sipa> strange!
626 2012-11-05 19:12:03 <BlueMatt> if you find a bug in block acceptance, as always, ping me and Ill add a new test for it
627 2012-11-05 19:13:12 <sipa> because i can't find code in current head that enforces this (i've submitted a pullreq, and was waiting to get it pulltested)
628 2012-11-05 19:13:59 <BlueMatt> why would it need to be explicitly enforced? should it not not connect a block which has the same work as the current tip?
629 2012-11-05 19:15:16 <BlueMatt> (ie in the same check that checks for < work)
630 2012-11-05 19:15:27 <sipa> it may be debatable, but it mayb have very serious consequences
631 2012-11-05 19:15:56 <BlueMatt> yea...other question...why did test_bitcoin segfault when testing your pull...
632 2012-11-05 19:16:15 <BlueMatt> anyone else seen that on a pull tester log?
633 2012-11-05 19:16:16 <sipa> i've seen that bug before; maybe some old code leaked in
634 2012-11-05 19:16:24 <sipa> i'll try it locally
635 2012-11-05 19:16:32 <BlueMatt> we fixed that bug a very, very long time ago
636 2012-11-05 19:17:25 <gmaxwell> Being willing to switching on equal difficulty at a very least make it take longer to converge... and increases the risk of non-convergence.
637 2012-11-05 19:18:12 <jgarzik> what is current tip height, on testnet3?
638 2012-11-05 19:18:26 <jgarzik> 35138?
639 2012-11-05 19:18:58 <BlueMatt> gmaxwell: yes, it absolutely shouldn't be allowed, my comment was more "in what weird code design is an explicit check for that required?"
640 2012-11-05 19:19:35 <BlueMatt> (ie there must be a check to not connect a block who's work is < current tip work, but why cant that just be <=)
641 2012-11-05 19:19:57 <sipa> BlueMatt: current code has a function ConnectBestBlock, which finds the best block among all known, and connects it
642 2012-11-05 19:20:15 <BlueMatt> ahh, ok, still...seems strange
643 2012-11-05 19:20:36 <sipa> and if two have the same amount of work, any of them may be first
644 2012-11-05 19:21:16 <jgarzik> sipa: seems like the first one encountered should be first
645 2012-11-05 19:21:39 <BlueMatt> jgarzik: yep, thats what sipa's pull (theoretically) does
646 2012-11-05 19:21:59 <sipa> BlueMatt: think i know why it segfaults
647 2012-11-05 19:22:41 <sipa> pindexBest may be NULL, in case of the genesis block being connected
648 2012-11-05 19:22:45 <sipa> yup, fixed it
649 2012-11-05 19:22:46 <BlueMatt> ahh
650 2012-11-05 19:23:27 <BlueMatt> yea, grep only shows your pull with a "Test setup error"
651 2012-11-05 19:25:32 <sipa> BlueMatt: updated
652 2012-11-05 19:50:40 <jgarzik> w00t, picocoin is happily parsing the testnet3 block headers, and generating addresses. Just sent the first transaction /to/ a picocoin wallet.
653 2012-11-05 19:51:35 <ThomasV_> jgarzik: what is picocoin?
654 2012-11-05 19:52:07 <jgarzik> ThomasV: SPV client in C
655 2012-11-05 19:52:27 <ThomasV_> oh I see
656 2012-11-05 19:52:59 <jgarzik> ThomasV_: Under development, not ready for users yet. https://github.com/jgarzik/picocoin/
657 2012-11-05 20:08:54 <sipa> i wonder whether using a patricia tree for a UTXO merkle structure is even useful
658 2012-11-05 20:09:27 <edcba> who is patricia ?
659 2012-11-05 20:10:07 <sipa> it's an acronym, afaik
660 2012-11-05 20:10:30 <sipa> Practical Algorithm To Retrieve Information Coded In Alphanumeric
661 2012-11-05 20:10:36 <edcba> ok
662 2012-11-05 20:15:24 <kjj_> sipa: since hashes are binary, you end up with a radix tree in base 256
663 2012-11-05 20:15:52 <sipa> i'd certainly do it bitwise, so base 2
664 2012-11-05 20:16:17 <sipa> the advantage patricia trees have over normal tries is compact representation along common paths
665 2012-11-05 20:16:29 <sipa> but you don't expect many common paths if the keys are hashes
666 2012-11-05 20:17:35 <kjj_> for faster lookups, I'd chunk it out
667 2012-11-05 20:18:20 <sipa> but the representation of "how many bits are compressed in the following path" and those bits themself also require storage
668 2012-11-05 20:18:28 <kjj_> make a base32 tree and pull bits out 5 at a time. moving down the tree is than AND, shift and multiply
669 2012-11-05 20:18:53 <sipa> and i'm not convinced this is less than what you gain from the actual compression
670 2012-11-05 20:19:44 <kjj_> is everyone talking in the channel getting messages from some guy looking to buy or sell bitcoins?
671 2012-11-05 20:20:13 <maaku> kjj_: yes, set your client to ignore
672 2012-11-05 20:20:30 <maaku> and send your logs to the police :)
673 2012-11-05 20:20:32 <kjj_> yeah, did. was just wondering if I was special, or if he just messaged everyone that talks
674 2012-11-05 20:21:05 <kjj_> binary data is a bitch to search effectively
675 2012-11-05 20:21:42 <kjj_> particularly hashes, since by design they don't have any commonality in particular
676 2012-11-05 20:23:54 <kjj_> the best I was able to come up with was a tree/list hybrid. each tree node had 2^n downstream pointers, and the thing pointed to was either another tree node, or a list node. the idea was you'd follow n bits at a time down the tree until you hit a list, then just iterate the few entries in the list. if a list hit a high water mark, split into a tree node
677 2012-11-05 20:24:47 <sipa> anything with a branching factor higher than 2 will result in needless hash operations
678 2012-11-05 20:25:19 <kjj_> they are already hashed
679 2012-11-05 20:25:35 <maaku> sipa: prefix trie?
680 2012-11-05 20:25:52 <sipa> kjj_: i'm talking about an authenticated utxo set
681 2012-11-05 20:25:58 <sipa> so like a merkle tree
682 2012-11-05 20:26:30 <sipa> where every node stores pointers to its subnodes + their hashes
683 2012-11-05 20:26:46 <sipa> to be able to efficiently recompute the merkle root when doing updates
684 2012-11-05 20:26:59 <kjj_> ahhh, that's a whole different pain in the ass
685 2012-11-05 20:27:16 <edcba> is it that long to recompute merkle root ?
686 2012-11-05 20:27:53 <kjj_> either the tree is going to get unbalanced as hell, and suck for searching, or you'll have to redo tons of hashes each time you update it
687 2012-11-05 20:28:27 <sipa> well, that's the discussion... balanced tree or trie
688 2012-11-05 20:28:45 <edcba> just try !
689 2012-11-05 20:28:59 <sipa> balances tree certainly has better theoretical properties, like a guarantee on its height in function of log(elements)
690 2012-11-05 20:29:04 <kjj_> I think that the practical problems of maintaining such a tree will sink plans to distribute utxo. if I had to guess, I'd say that Satoshi knew that and that's why he didn't use them
691 2012-11-05 20:29:32 <maaku> kjj_: which is why you use a prefix trie--it avoids those practical problems
692 2012-11-05 20:29:46 <sipa> tries have the risk of becoming unbalanced
693 2012-11-05 20:29:52 <kjj_> maaku: seems unlikely to avoid all of them.
694 2012-11-05 20:30:06 <sipa> especially when the whole world can see the data structure, and try to unbalance it
695 2012-11-05 20:30:18 <sipa> but there are computational limits to how badly you can unbalance it
696 2012-11-05 20:30:20 <edcba> yeah better to stay in safe side
697 2012-11-05 20:30:25 <sipa> when the keys are hashes
698 2012-11-05 20:30:40 <maaku> sipa: and the consequences of one person's mischief would be????
699 2012-11-05 20:30:51 <edcba> anyway satoshi just relied on hardware advances so don't care about perf :)
700 2012-11-05 20:30:52 <sipa> maaku: slowing down block validation for every full node
701 2012-11-05 20:31:15 <sipa> there are currently around 1.25 M entries in the utxo set
702 2012-11-05 20:31:21 <maaku> you'd have to have a LOT of bit dust to do that
703 2012-11-05 20:31:59 <edcba> some ppl may have it and time on their hands now
704 2012-11-05 20:32:00 <kjj_> maaku: not really. I pull a block every few months from my mining rigs. I could totally do a MAX_BLOCK_SIZE block that was nothing but dust, pay no fees on it
705 2012-11-05 20:32:56 <maaku> kjj_: and that would cause, what, a couple dozen extra hashes?
706 2012-11-05 20:32:58 <kjj_> and since the each entry only costs me 1 satoshi, I don't even care if I have private keys to match the pubkeys in the scripts
707 2012-11-05 20:34:33 <kjj_> what it costs the network depends on the tree's details. I can make searches for legit transactions more costly, and/or I can make future additions/deletions more costly
708 2012-11-05 20:34:46 <maaku> if you don't have the private keys, then that structure never gets updated and the hashes can be cached
709 2012-11-05 20:35:30 <kjj_> maaku: only if each block's transactions end up in their own branch. which might be a good idea, I'm not sure
710 2012-11-05 20:35:50 <maaku> kjj_: with a prefix trie there's a maximum limit on how long a search can take
711 2012-11-05 20:36:05 <sipa> yes, 256 steps :)
712 2012-11-05 20:36:36 <maaku> if you use a binary trie, yes, but you could use a 1-byte radix
713 2012-11-05 20:36:38 <maaku> maximum 32 steps
714 2012-11-05 20:36:55 <kjj_> 64, not 32.
715 2012-11-05 20:37:00 <sipa> 32
716 2012-11-05 20:37:06 <sipa> 32 bytes = 256 bits
717 2012-11-05 20:37:07 <kjj_> er, no, 32. gah. I need coffee or something
718 2012-11-05 20:37:27 <edcba> if distribution is uniform i don't see what you gain
719 2012-11-05 20:37:35 <sipa> gain from what?
720 2012-11-05 20:38:21 <sipa> maaku: higher fan-out means more data to hash in each step
721 2012-11-05 20:38:52 <maaku> sipa: it also means less total hashes
722 2012-11-05 20:38:57 <sipa> so you only need 1/8 the number of steps, but you hash 128x as much data in each step
723 2012-11-05 20:39:20 <kjj_> and an attacker should be able to extend maybe 5,000 to 10,000 nodes per block, forcing all of them to need recalculation
724 2012-11-05 20:41:20 <kjj_> if we assume that hashing is less costly than comparing by at least a factor of 128, the binary tree looks best
725 2012-11-05 20:41:30 <kjj_> er, more costly. where's that damn coffee?
726 2012-11-05 20:46:06 <maaku> log2(1.5M) ~= 20.5, so balanced-binary vs prefix trie *worst-case* results in 56% more hashes for a targeted "attack"--although it's unclear to me what, if anything that attack would accomplish
727 2012-11-05 20:46:36 <sipa> 56% ?
728 2012-11-05 20:46:38 <maaku> what i'm saying is that's not orders of magnitude, so we should pay attention to non-performance considerations
729 2012-11-05 20:46:42 <maaku> 32/20.5
730 2012-11-05 20:47:32 <kjj_> also keep in mind that bitcoin is growing. it is safe to assume every day that the utxo set is smaller than it ever will be in the future
731 2012-11-05 20:47:34 <sipa> at least compare balanced-256-b-tree to 256-choice-prefix-trie or balances-binary to binary prefix trie
732 2012-11-05 20:47:50 <sipa> kjj_: it has actually shrunk in the past
733 2012-11-05 20:48:12 <kjj_> I'm sure it has. I don't expect that to be a common thing in the future
734 2012-11-05 20:48:21 <sipa> sure
735 2012-11-05 20:49:56 <maaku> sipa: are you talking about doing a balanced-256-b-tree?
736 2012-11-05 20:52:23 <sipa> no
737 2012-11-05 20:52:47 <sipa> but it's not a fair comparison if you compare a balanced binary tree against a 256-fanout-trie
738 2012-11-05 21:07:31 <jgarzik> Sigh. Compile with -O2, a new picocoin test passes, hashes match. Compile with just '-Wall -g', and the new test fails.
739 2012-11-05 21:28:48 <Luke-Jr> ACTION ponders
740 2012-11-05 21:50:24 <Luke-Jr> ahaha, I spent all that time debugging and it's a bug on slush's pool after all! :<
741 2012-11-05 21:50:29 <Luke-Jr> slush: ;p
742 2012-11-05 21:51:54 <edcba> lol
743 2012-11-05 21:52:19 <upb> ACTION stalks edcba
744 2012-11-05 21:52:32 <edcba> help !
745 2012-11-05 21:52:39 <edcba> upb wants to steal my bitcoinz !
746 2012-11-05 21:53:59 <upb> false i just want to stalk you