1 2013-02-15 00:01:58 rdymac has joined
2 2013-02-15 00:06:22 RazielZ has quit (Ping timeout: 256 seconds)
3 2013-02-15 00:11:55 Zarutian has joined
4 2013-02-15 00:12:11 ashams has quit (Read error: Connection reset by peer)
5 2013-02-15 00:12:18 JZavala has joined
6 2013-02-15 00:13:46 K1773R has joined
7 2013-02-15 00:14:21 * jaakkos created ext4 on zvol, fixed the bdb issue
8 2013-02-15 00:15:10 one_zero has joined
9 2013-02-15 00:15:51 rdymac has quit (Quit: This computer has gone to sleep)
10 2013-02-15 00:17:33 <sipa> Luke-Jr: smaller representation of Pi in base Pi: 3.0033332333111332313333122033213033233312132103331323221213121120332232113111122133221313032233331113
11 2013-02-15 00:17:56 * Luke-Jr hides from sipa
12 2013-02-15 00:18:28 <Luke-Jr> sipa: please tell me you didn't spend the last 2 hours on that
13 2013-02-15 00:18:34 <sipa> I didn't
14 2013-02-15 00:18:49 <sipa> now, my status bar at this point in irssi (and it's not intentional):
15 2013-02-15 00:18:57 <sipa> [01:00] [sipa(+i)] [17:freenode/#bitcoin-dev(+Ccnt)] [Act: 3,14,15,39]
16 2013-02-15 00:19:10 <sipa> i have no channel numbered 92, unfortunately
17 2013-02-15 00:21:17 * andytoshi starts looking for patterns
18 2013-02-15 00:22:11 Muis has quit (Quit: Connection reset by apple)
19 2013-02-15 00:22:59 <BlueMatt> /join #bitcoin-conspiracies
20 2013-02-15 00:24:25 <andytoshi> hey... it says i'm an admin, and i'm the only one in there
21 2013-02-15 00:24:33 <andytoshi> it's like....i stepped out of the matrix
22 2013-02-15 00:24:36 <andytoshi> it was all in my head
23 2013-02-15 00:24:45 <K1773R> u just joined a matrix in a matrix
24 2013-02-15 00:26:09 Ssateneth has joined
25 2013-02-15 00:26:19 <Ssateneth> webb why you scamming?
26 2013-02-15 00:27:07 dvide has quit ()
27 2013-02-15 00:27:08 BTC4amazon has joined
28 2013-02-15 00:27:12 BTC4amazon has left ()
29 2013-02-15 00:27:34 <K1773R> wrong channel Ssateneth
30 2013-02-15 00:28:59 <Ssateneth> he stole $400 and isn't in any other channel as far as I know.
31 2013-02-15 00:29:25 gavinandresen has quit (Quit: gavinandresen)
32 2013-02-15 00:29:47 <K1773R> still, this is the dev channel
33 2013-02-15 00:30:16 <ProfMac> can I configure bitcoin-qt so that it accepts connections, but does not initiate them?
34 2013-02-15 00:30:32 <sipa> ProfMac: -connect=0.0.0.0 -listen
35 2013-02-15 00:31:14 <Ssateneth> idc, he stole $400. by telling me to leave, you're defending his scammer ways
36 2013-02-15 00:31:22 <ProfMac> thanks sipa. Is the IPv6 version -connect=[::] -listen ?
37 2013-02-15 00:31:49 <K1773R> no i dont, u will reach totally nothing if you want to talk in this channel
38 2013-02-15 00:32:09 <K1773R> and he probably wont answer, use #bitcoin-otc or if hes on bitcointalk open a scam accusation thread
39 2013-02-15 00:32:38 <K1773R> ProfMac: 0.0.0.0 is fine, since it cant connect to it
40 2013-02-15 00:32:44 <Ssateneth> i'd rather stay here in case he comes back, kthx
41 2013-02-15 00:33:13 <K1773R> i didnt say u cant stay here, just said u shouldnt discuss such thing in this channel.
42 2013-02-15 00:33:32 PhantomSpark has quit (2!~kvirc@137.125.143.209|Read error: Connection reset by peer)
43 2013-02-15 00:33:59 PhantomSpark has joined
44 2013-02-15 00:35:17 BTC4amazon has joined
45 2013-02-15 00:35:40 dust-otc has quit (Remote host closed the connection)
46 2013-02-15 00:35:47 <BTC4amazon> ;;ident webb
47 2013-02-15 00:35:47 <gribble> Nick 'webb', with hostmask 'webb!webb@b1f-003.static.bnc4free.com', is not identified.
48 2013-02-15 00:36:19 one_zero has quit (Read error: Connection reset by peer)
49 2013-02-15 00:38:40 one_zero has joined
50 2013-02-15 00:38:40 one_zero has quit (Read error: Connection reset by peer)
51 2013-02-15 00:39:14 dparrish has quit (Ping timeout: 256 seconds)
52 2013-02-15 00:39:29 BTC4amazon has left ()
53 2013-02-15 00:40:29 word has quit (Quit: Konversation terminated!)
54 2013-02-15 00:42:15 dhill has joined
55 2013-02-15 00:42:41 <dhill> is it possible to run bitcoind without threads, or max 1 thread?
56 2013-02-15 00:43:27 <K1773R> why would u do that?
57 2013-02-15 00:43:44 <dhill> trying to see if i can debug this crash
58 2013-02-15 00:44:05 <dhill> http://gbpaste.org/dJryl
59 2013-02-15 00:44:13 word has joined
60 2013-02-15 00:44:40 <ProfMac> Can bitcoin-qt re-read the bitcoin.conf file without stopping and restarting?
61 2013-02-15 00:44:51 <sipa> dhill: there's always at least the network thread and the message handler thread
62 2013-02-15 00:45:27 dparrish has joined
63 2013-02-15 00:45:36 <dhill> sipa: what boost version are you using?
64 2013-02-15 00:46:10 one_zero has joined
65 2013-02-15 00:47:54 kiceek has quit (Ping timeout: 256 seconds)
66 2013-02-15 00:48:49 <K1773R> i remember one boost version is creating segfaults (similiar ones), cant remember which one
67 2013-02-15 00:48:59 meLon has quit (Quit: leaving)
68 2013-02-15 00:49:10 <dhill> i am on 1.42.0
69 2013-02-15 00:49:12 meLon has joined
70 2013-02-15 00:49:12 meLon has quit (Changing host)
71 2013-02-15 00:49:13 meLon has joined
72 2013-02-15 00:49:19 <dhill> wonder if i should give 1.52.0 a go
73 2013-02-15 00:49:44 meLon has quit (Client Quit)
74 2013-02-15 00:50:08 <K1773R> you could compile the libs and link them static into bitcoind, so you dont have to replace the system wide ones
75 2013-02-15 00:50:40 meLon has joined
76 2013-02-15 00:50:41 meLon has quit (Changing host)
77 2013-02-15 00:50:41 meLon has joined
78 2013-02-15 00:51:17 vessenes has joined
79 2013-02-15 00:57:08 rbecker is now known as RBecker
80 2013-02-15 01:00:16 kiceek has joined
81 2013-02-15 01:00:21 Michail1 is now known as Michail1_
82 2013-02-15 01:00:23 darkskiez has quit (Ping timeout: 245 seconds)
83 2013-02-15 01:00:58 PhantomSpark has quit (2!~kvirc@137.125.143.209|Read error: Connection reset by peer)
84 2013-02-15 01:01:22 webb has quit (Quit: I'm using a Free IRC Bouncer from BNC4FREE - http://bnc4free.com/)
85 2013-02-15 01:01:24 PhantomSpark has joined
86 2013-02-15 01:01:35 Toresh_ has joined
87 2013-02-15 01:02:45 darkskiez has joined
88 2013-02-15 01:03:58 Seeraber has joined
89 2013-02-15 01:04:21 Toresh has quit (Ping timeout: 246 seconds)
90 2013-02-15 01:04:21 Toresh_ is now known as Toresh
91 2013-02-15 01:07:28 comboy has quit (Ping timeout: 245 seconds)
92 2013-02-15 01:10:35 Hashdog has quit (Remote host closed the connection)
93 2013-02-15 01:15:24 Tatsuya has joined
94 2013-02-15 01:17:06 JWU42_ is now known as JWU42
95 2013-02-15 01:17:17 JWU42 has quit (Changing host)
96 2013-02-15 01:17:17 JWU42 has joined
97 2013-02-15 01:18:47 comboy has joined
98 2013-02-15 01:22:53 vicep has quit (Ping timeout: 245 seconds)
99 2013-02-15 01:23:52 Seeraber has quit (Quit: Bye!)
100 2013-02-15 01:23:56 toffoo has joined
101 2013-02-15 01:26:00 bitafterbit has quit (Remote host closed the connection)
102 2013-02-15 01:32:31 vessenes has quit (Remote host closed the connection)
103 2013-02-15 01:51:02 LargoG has quit (Remote host closed the connection)
104 2013-02-15 01:58:12 m00p has quit (Ping timeout: 248 seconds)
105 2013-02-15 01:59:19 denebole has joined
106 2013-02-15 01:59:37 dlb76 has quit ()
107 2013-02-15 02:01:46 sgornick has joined
108 2013-02-15 02:04:03 denebole has quit (Quit: Leaving)
109 2013-02-15 02:06:31 Toresh_ has joined
110 2013-02-15 02:06:41 PhantomSpark has quit (2!~kvirc@137.125.143.209|Ping timeout: 276 seconds)
111 2013-02-15 02:09:07 Toresh has quit (Ping timeout: 252 seconds)
112 2013-02-15 02:09:07 Toresh_ is now known as Toresh
113 2013-02-15 02:09:40 Michail1_ is now known as Michail1
114 2013-02-15 02:16:07 dhill has quit (Ping timeout: 240 seconds)
115 2013-02-15 02:19:05 ImpStout has joined
116 2013-02-15 02:19:28 rdymac has joined
117 2013-02-15 02:23:57 rdymac has quit (Read error: No route to host)
118 2013-02-15 02:28:08 meLon has quit (Quit: leaving)
119 2013-02-15 02:33:36 <jaakkos> how about making the bitcoin client first download block headers and asking merkle branches from a full neighbour to provide "full" functionality, then download merkle trees in the background?
120 2013-02-15 02:34:01 <jaakkos> so that there will always be a good chance that you have full neighbors
121 2013-02-15 02:40:02 rdymac has joined
122 2013-02-15 02:44:21 <Eliel_> jaakkos: That's the long term plan, as I understand. It requires someone to take up the task and actually do it to happen though.
123 2013-02-15 02:46:00 rdymac has quit (Read error: Connection reset by peer)
124 2013-02-15 02:50:42 agricocb has joined
125 2013-02-15 02:52:36 Tatsuya has quit (Ping timeout: 248 seconds)
126 2013-02-15 02:54:31 freakazoid_ has quit (Ping timeout: 252 seconds)
127 2013-02-15 02:57:20 <sipa> jaakkos: no need for merkle branches, just the headers in the best chain suffices
128 2013-02-15 02:57:24 <sipa> and indeed, that's the plan
129 2013-02-15 02:57:38 <jaakkos> sipa: how do you know your balance then
130 2013-02-15 02:57:52 <sipa> you don
131 2013-02-15 02:57:58 <sipa> 't
132 2013-02-15 02:58:44 <jaakkos> well you can't even pay because you don't know the outputs
133 2013-02-15 02:58:50 <jaakkos> that kind of sucks?
134 2013-02-15 02:59:10 <sipa> wait
135 2013-02-15 02:59:14 <sipa> there's two different ideas
136 2013-02-15 03:00:01 <sipa> 1) download headers first, and then sync block data in the background, only having wallet functionality after the second step
137 2013-02-15 03:00:35 <jaakkos> (... sowhy bother the first step at all?)
138 2013-02-15 03:01:09 <sipa> that's already extremely useful, as it means you know the best chain in advance, can do sigcheck skipping without hardcoded checkpoints, do load spreading of the actual block data downloa (as it's known what needs to be downloaded), ...
139 2013-02-15 03:01:28 <sipa> it's just an optimization for synchronizing a full node, like we have now
140 2013-02-15 03:02:11 <sipa> the other idea, "headers only mode", is starting as an SPV node, which downloads headers and transactions interesting to the user's wallet first; and in the background upgrade to a full node if resources allow it
141 2013-02-15 03:02:28 <sipa> that's probably what you were referring to
142 2013-02-15 03:02:55 PhantomSpark has joined
143 2013-02-15 03:03:54 Ssateneth has left ()
144 2013-02-15 03:05:17 <jaakkos> yes
145 2013-02-15 03:05:34 <sipa> imho, we should just disconnect both entirely
146 2013-02-15 03:05:40 <sipa> wallet == spv node
147 2013-02-15 03:05:49 <sipa> blockchain == full node
148 2013-02-15 03:06:00 <sipa> run both if you want the security of a full node + a wallet
149 2013-02-15 03:06:05 <Luke-Jr> :D
150 2013-02-15 03:06:42 <sipa> (it can still be the same codebase, or even the same binary)
151 2013-02-15 03:06:58 <Luke-Jr> sipa: or better, a shared library <.<
152 2013-02-15 03:07:07 <sipa> sure
153 2013-02-15 03:07:29 <sipa> there is currently nothing in bitcoin-qt that relies on having the full block data present
154 2013-02-15 03:07:41 <sipa> (afaik)
155 2013-02-15 03:08:18 * Luke-Jr kicks 1BmFHTUM7Dtfyo15SaXHs2EoMWLKMBbfTe
156 2013-02-15 03:09:11 CodeShark has quit (Remote host closed the connection)
157 2013-02-15 03:19:29 PhantomSpark has quit (2!~kvirc@pool-71-190-230-179.nycmny.fios.verizon.net|Ping timeout: 276 seconds)
158 2013-02-15 03:27:44 paybitcoin has joined
159 2013-02-15 03:36:32 Goonie has quit (Ping timeout: 260 seconds)
160 2013-02-15 03:38:15 Maged has joined
161 2013-02-15 03:38:37 D34TH has quit (Read error: Connection reset by peer)
162 2013-02-15 03:42:35 sgornick has quit (Ping timeout: 264 seconds)
163 2013-02-15 03:44:09 PhantomSpark has joined
164 2013-02-15 03:48:54 PhantomSpark has joined
165 2013-02-15 03:50:02 B0g4r7 has quit (Ping timeout: 276 seconds)
166 2013-02-15 03:51:42 fiesh has quit (Ping timeout: 260 seconds)
167 2013-02-15 03:54:07 fiesh has joined
168 2013-02-15 03:55:08 sgornick has joined
169 2013-02-15 03:56:40 MobGod is now known as Skav
170 2013-02-15 03:57:32 B0g4r7 has joined
171 2013-02-15 03:58:13 da2ce7_d is now known as da2ce7
172 2013-02-15 03:58:48 yellowhat has joined
173 2013-02-15 03:58:48 yellowhat has quit (Remote host closed the connection)
174 2013-02-15 04:03:01 yellowhat has joined
175 2013-02-15 04:03:13 yellowhat has quit (Remote host closed the connection)
176 2013-02-15 04:03:39 yellowhat has joined
177 2013-02-15 04:11:25 dlb76 has joined
178 2013-02-15 04:17:20 sgornick has quit (Quit: Ex-Chat)
179 2013-02-15 04:30:32 wumpus has quit (No Ping reply in 180 seconds.)
180 2013-02-15 04:30:47 wumpus has joined
181 2013-02-15 04:33:14 sgornick has joined
182 2013-02-15 04:34:56 Dyaheon has joined
183 2013-02-15 04:35:19 MobiusL has quit (Remote host closed the connection)
184 2013-02-15 04:36:29 MobiusL has joined
185 2013-02-15 04:41:34 BlueMattBot has joined
186 2013-02-15 04:41:34 BlueMattBot has quit (Changing host)
187 2013-02-15 04:41:34 BlueMattBot has joined
188 2013-02-15 04:47:53 paraipan has quit (Quit: Saliendo)
189 2013-02-15 04:52:04 AtashiCon has quit (Quit: AtashiCon)
190 2013-02-15 04:54:53 freakazoid_ has joined
191 2013-02-15 05:05:50 freewil has joined
192 2013-02-15 05:06:25 AtashiCon has joined
193 2013-02-15 05:09:22 [7] has quit (Disconnected by services)
194 2013-02-15 05:09:31 TheSeven has joined
195 2013-02-15 05:10:12 CodeShark has joined
196 2013-02-15 05:14:28 CodeShark has quit (Ping timeout: 248 seconds)
197 2013-02-15 05:15:37 ciphermonk has joined
198 2013-02-15 05:17:11 twobitcoins has joined
199 2013-02-15 05:17:17 freewil has quit (Remote host closed the connection)
200 2013-02-15 05:19:46 WolfAlex has joined
201 2013-02-15 05:20:01 SomeoneWeird has quit (Excess Flood)
202 2013-02-15 05:22:22 Guest92994 has joined
203 2013-02-15 05:22:29 WolfAlex_ has quit (Ping timeout: 245 seconds)
204 2013-02-15 05:23:12 Muis has joined
205 2013-02-15 05:23:49 HM has quit (Ping timeout: 252 seconds)
206 2013-02-15 05:27:05 moore_ has joined
207 2013-02-15 05:29:19 HM has joined
208 2013-02-15 05:31:16 snakie has quit (Remote host closed the connection)
209 2013-02-15 05:33:26 Anonymoussss has joined
210 2013-02-15 05:43:47 Guest92994 has quit (Changing host)
211 2013-02-15 05:43:47 Guest92994 has joined
212 2013-02-15 05:43:55 Guest92994 is now known as SomeoneWeird
213 2013-02-15 05:50:43 eoss has joined
214 2013-02-15 05:53:05 Zarutian has quit (Quit: Zarutian)
215 2013-02-15 05:54:55 techlife has quit (Ping timeout: 252 seconds)
216 2013-02-15 05:57:40 techlife has joined
217 2013-02-15 05:58:48 Anonymoussss has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
218 2013-02-15 06:02:01 JZavala has quit (Ping timeout: 252 seconds)
219 2013-02-15 06:03:08 Anonymoussss has joined
220 2013-02-15 06:12:55 root2_ has joined
221 2013-02-15 06:13:48 Anonymoussss has quit (Quit: Colloquy for iPad - http://colloquy.mobi)
222 2013-02-15 06:14:24 ShaTwo has joined
223 2013-02-15 06:15:36 ShaTwo has left ()
224 2013-02-15 06:15:51 root2 has quit (Ping timeout: 256 seconds)
225 2013-02-15 06:17:02 [\\\] has quit (Ping timeout: 252 seconds)
226 2013-02-15 06:19:13 sgornick has quit (Quit: Ex-Chat)
227 2013-02-15 06:21:14 [\\\] has joined
228 2013-02-15 06:21:38 rdponticelli_ has joined
229 2013-02-15 06:24:52 HM has quit (Ping timeout: 252 seconds)
230 2013-02-15 06:27:20 rdponticelli has quit (Ping timeout: 276 seconds)
231 2013-02-15 06:29:21 HM has joined
232 2013-02-15 06:30:17 brwyatt is now known as brwyatt|Away
233 2013-02-15 06:38:03 Tatsuya has joined
234 2013-02-15 06:43:19 MrTiggr has quit (Ping timeout: 256 seconds)
235 2013-02-15 06:55:11 moore_ has quit (Ping timeout: 256 seconds)
236 2013-02-15 07:04:05 lupine313 has joined
237 2013-02-15 07:05:43 nus- has joined
238 2013-02-15 07:06:00 vicep has joined
239 2013-02-15 07:09:57 nus has quit (Ping timeout: 276 seconds)
240 2013-02-15 07:10:40 CodeShark has joined
241 2013-02-15 07:13:39 Anonymoussss has joined
242 2013-02-15 07:14:58 CodeShark has quit (Ping timeout: 248 seconds)
243 2013-02-15 07:16:59 andre__ has joined
244 2013-02-15 07:17:22 andre__ is now known as Guest40413
245 2013-02-15 07:19:44 vicep has quit (Ping timeout: 252 seconds)
246 2013-02-15 07:21:00 FredEE has quit (Quit: FredEE)
247 2013-02-15 07:25:13 Cory has quit (Ping timeout: 260 seconds)
248 2013-02-15 07:26:41 Cory has joined
249 2013-02-15 07:28:06 vicep has joined
250 2013-02-15 07:29:48 swappermall_ has quit (Ping timeout: 256 seconds)
251 2013-02-15 07:30:29 Guest40413 has quit (Ping timeout: 248 seconds)
252 2013-02-15 07:37:43 freakazoid_ has quit (Ping timeout: 252 seconds)
253 2013-02-15 07:38:10 Michail1 has left ()
254 2013-02-15 07:39:25 vicep has quit (Read error: Operation timed out)
255 2013-02-15 07:40:36 vicep has joined
256 2013-02-15 07:45:08 RBecker is now known as rbecker
257 2013-02-15 07:45:45 Goonie has joined
258 2013-02-15 07:50:40 vicep has quit (Ping timeout: 252 seconds)
259 2013-02-15 07:51:00 vicep has joined
260 2013-02-15 07:57:02 nibcoin has joined
261 2013-02-15 08:00:01 vicep has quit (Ping timeout: 252 seconds)
262 2013-02-15 08:01:20 mapppum has joined
263 2013-02-15 08:01:49 mappum has quit (Ping timeout: 256 seconds)
264 2013-02-15 08:11:41 nibcoin has quit (Ping timeout: 245 seconds)
265 2013-02-15 08:19:23 ovidiusoft has joined
266 2013-02-15 08:26:26 mapppum has quit (Ping timeout: 256 seconds)
267 2013-02-15 08:27:40 Apexseals has quit (Ping timeout: 244 seconds)
268 2013-02-15 08:29:10 Apexseals has joined
269 2013-02-15 08:34:27 Apexseals has quit (Ping timeout: 276 seconds)
270 2013-02-15 08:34:57 Apexseals has joined
271 2013-02-15 08:39:42 MagicalTux has quit (Excess Flood)
272 2013-02-15 08:40:08 MT`AwAy has joined
273 2013-02-15 08:46:01 Tatsuya has quit (Ping timeout: 256 seconds)
274 2013-02-15 08:47:51 jdnavarro has joined
275 2013-02-15 08:49:59 ZephyrVoid has quit (Ping timeout: 256 seconds)
276 2013-02-15 08:50:28 RazielZ has joined
277 2013-02-15 08:55:32 ciphermonk has quit (Ping timeout: 276 seconds)
278 2013-02-15 08:55:46 MT`AwAy has quit (Changing host)
279 2013-02-15 08:55:46 MT`AwAy has joined
280 2013-02-15 08:55:50 MT`AwAy is now known as MagicalTux
281 2013-02-15 09:01:35 pooler has quit (Remote host closed the connection)
282 2013-02-15 09:04:06 ZephyrVoid has joined
283 2013-02-15 09:09:21 Insu has joined
284 2013-02-15 09:11:20 Insu has quit (Remote host closed the connection)
285 2013-02-15 09:15:48 Arnavion has quit (Quit: Arnavion)
286 2013-02-15 09:16:03 Arnavion has joined
287 2013-02-15 09:18:52 Tatsuya has joined
288 2013-02-15 09:20:51 [ken] has joined
289 2013-02-15 09:25:59 t7 has joined
290 2013-02-15 09:33:32 toffoo has quit ()
291 2013-02-15 09:35:47 Tatsuya has quit (Read error: Connection reset by peer)
292 2013-02-15 09:42:07 CodeShark has joined
293 2013-02-15 09:46:45 lupine3131 has joined
294 2013-02-15 09:46:45 lupine313 has quit (Read error: Connection reset by peer)
295 2013-02-15 09:46:49 lupine3131 has quit (Client Quit)
296 2013-02-15 09:55:02 Pasha has joined
297 2013-02-15 09:56:23 Cory has quit (Ping timeout: 264 seconds)
298 2013-02-15 10:00:06 Pasha is now known as Cory
299 2013-02-15 10:09:53 eoss has quit (Ping timeout: 256 seconds)
300 2013-02-15 10:11:28 CodesInChaos has joined
301 2013-02-15 10:13:36 kuzetsa has quit (Quit: Stop doing that it. You don't want it to fall off.)
302 2013-02-15 10:13:41 rdymac has joined
303 2013-02-15 10:16:05 one_zero has quit ()
304 2013-02-15 10:22:38 B0g4r7 has quit (Ping timeout: 276 seconds)
305 2013-02-15 10:25:53 m00p has joined
306 2013-02-15 10:27:11 B0g4r7 has joined
307 2013-02-15 10:35:43 m00p has quit (Ping timeout: 256 seconds)
308 2013-02-15 10:36:22 axhlf has joined
309 2013-02-15 10:38:50 dvide has joined
310 2013-02-15 10:47:08 veerboot has quit (Quit: Very funny Scottie.... Now beam down my clothes!)
311 2013-02-15 10:55:58 drizztbsd has joined
312 2013-02-15 11:02:00 Hashdog has joined
313 2013-02-15 11:04:30 root2_ has quit (Read error: Connection reset by peer)
314 2013-02-15 11:04:56 root2_ has joined
315 2013-02-15 11:20:35 root2_ is now known as root2
316 2013-02-15 11:27:04 Arnavion has quit (Quit: Arnavion)
317 2013-02-15 11:27:05 MrTiggr has joined
318 2013-02-15 11:27:16 root2_ has joined
319 2013-02-15 11:27:28 root2 has quit (Killed (cameron.freenode.net (Nickname regained by services)))
320 2013-02-15 11:27:28 root2_ is now known as root2
321 2013-02-15 11:29:30 talso has quit (Ping timeout: 264 seconds)
322 2013-02-15 11:32:23 Arnavion has joined
323 2013-02-15 11:32:57 kenbeans has joined
324 2013-02-15 11:35:09 kenbeans has left ()
325 2013-02-15 11:42:27 talso has joined
326 2013-02-15 11:42:40 drivelights has left ()
327 2013-02-15 11:43:05 i2pRelay has quit (Remote host closed the connection)
328 2013-02-15 11:43:29 i2pRelay has joined
329 2013-02-15 11:48:55 ashams has joined
330 2013-02-15 11:50:58 ciphermonk has joined
331 2013-02-15 11:52:23 nus has joined
332 2013-02-15 11:52:23 nus- has quit (Read error: Connection reset by peer)
333 2013-02-15 11:58:17 Tiggr has joined
334 2013-02-15 11:58:18 Tiggr has quit (Client Quit)
335 2013-02-15 11:59:03 MrTiggr has quit (Ping timeout: 256 seconds)
336 2013-02-15 12:00:04 kuzetsa has joined
337 2013-02-15 12:06:20 rdymac has quit (Ping timeout: 260 seconds)
338 2013-02-15 12:08:09 rdymac has joined
339 2013-02-15 12:11:46 ciphermonk has quit (Remote host closed the connection)
340 2013-02-15 12:11:51 ralphtheninja has joined
341 2013-02-15 12:15:19 MrTiggr has joined
342 2013-02-15 12:16:00 CodeShar_ has joined
343 2013-02-15 12:16:21 CodeShark has quit (Read error: Connection reset by peer)
344 2013-02-15 12:17:59 ralphtheninja has left ()
345 2013-02-15 12:25:57 skeledrew has quit (Ping timeout: 245 seconds)
346 2013-02-15 12:26:55 axhlf has quit (Remote host closed the connection)
347 2013-02-15 12:29:51 rdymac has quit (Ping timeout: 256 seconds)
348 2013-02-15 12:30:22 ciphermonk has joined
349 2013-02-15 12:33:21 axhlf has joined
350 2013-02-15 12:35:59 rdymac has joined
351 2013-02-15 12:40:38 <ciphermonk> How do you determine the transaction ordering used to compute the merke tree? Is it just the order in which you received the transactions in a block using "getdata" ?
352 2013-02-15 12:42:20 <ciphermonk> I'm confused because some dumps hint an the existence of a vMerkleTree ( ex: here: https://en.bitcoin.it/wiki/Dump_format#CBlock ) but I can't find any reference to such a structure in the protocol itself
353 2013-02-15 12:42:56 <Luke-Jr> ciphermonk: correct
354 2013-02-15 12:43:33 <Luke-Jr> ciphermonk: vMerkleTree is just a cache of all the hashes used in the calculation
355 2013-02-15 12:44:21 <ciphermonk> ah ok! thanks a lot :)
356 2013-02-15 12:44:48 copumpkin has quit (Ping timeout: 252 seconds)
357 2013-02-15 12:45:25 PsyKick1 has quit (Remote host closed the connection)
358 2013-02-15 12:45:29 copumpkin has joined
359 2013-02-15 12:53:52 <sipa> ciphermonk: the order of transactions in a block is fixed (and meaningful)
360 2013-02-15 12:57:02 <ciphermonk> ok, I wasn't sure if I missed an additional structure or protocol message that would specify the ordering. It's good to know that the order they are sent on the wire is fixed and this order is used to compute the merkle root
361 2013-02-15 12:59:15 t7 has quit (Read error: Connection reset by peer)
362 2013-02-15 12:59:26 LargoG has joined
363 2013-02-15 12:59:41 t7 has joined
364 2013-02-15 13:01:09 vicep has joined
365 2013-02-15 13:06:02 <ciphermonk> thanks for the clarification!
366 2013-02-15 13:07:33 skeledrew has joined
367 2013-02-15 13:12:31 rdymac has quit (Read error: Connection reset by peer)
368 2013-02-15 13:14:56 rdymac has joined
369 2013-02-15 13:25:49 MobiusL has quit (Quit: Ex-Chat)
370 2013-02-15 13:28:08 Pasha has joined
371 2013-02-15 13:29:25 Cory has quit (Ping timeout: 248 seconds)
372 2013-02-15 13:33:12 Pasha is now known as Cory
373 2013-02-15 13:33:17 MobiusL has joined
374 2013-02-15 13:34:14 rdymac has quit (Read error: Connection reset by peer)
375 2013-02-15 13:34:24 devrandom has quit (Ping timeout: 276 seconds)
376 2013-02-15 13:34:47 rdymac has joined
377 2013-02-15 13:34:52 devrandom has joined
378 2013-02-15 13:35:05 rdymac has quit (Read error: Connection reset by peer)
379 2013-02-15 13:35:28 rdymac has joined
380 2013-02-15 13:35:59 rdymac has quit (Read error: Connection reset by peer)
381 2013-02-15 13:37:04 rdymac has joined
382 2013-02-15 13:39:27 Jackneill has joined
383 2013-02-15 13:41:08 [ken] has quit (Quit: leaving)
384 2013-02-15 13:45:45 rdymac has quit (Read error: Connection reset by peer)
385 2013-02-15 13:46:16 rdymac has joined
386 2013-02-15 13:46:31 datagutt has joined
387 2013-02-15 13:50:15 rdymac has quit (Read error: Connection reset by peer)
388 2013-02-15 13:50:22 paraipan has joined
389 2013-02-15 13:51:10 rdymac has joined
390 2013-02-15 13:54:00 paybitcoin has quit (Read error: Connection reset by peer)
391 2013-02-15 13:54:59 t7` has joined
392 2013-02-15 13:55:12 Hashdog has quit (Remote host closed the connection)
393 2013-02-15 13:55:12 MobiusL has quit (Ping timeout: 276 seconds)
394 2013-02-15 13:57:04 t7 has quit (Ping timeout: 245 seconds)
395 2013-02-15 13:57:10 JWU42_ has joined
396 2013-02-15 13:57:45 t7`` has joined
397 2013-02-15 13:58:16 JWU42 has quit (Disconnected by services)
398 2013-02-15 13:58:20 JWU42_ is now known as JWU42
399 2013-02-15 13:58:24 JWU42 has quit (Changing host)
400 2013-02-15 13:58:24 JWU42 has joined
401 2013-02-15 13:58:34 TD has joined
402 2013-02-15 14:01:04 t7` has quit (Ping timeout: 252 seconds)
403 2013-02-15 14:01:51 rdymac has quit (Quit: This computer has gone to sleep)
404 2013-02-15 14:11:01 paybitcoin has joined
405 2013-02-15 14:11:29 MobiusL has joined
406 2013-02-15 14:16:47 t7`` has quit (Ping timeout: 264 seconds)
407 2013-02-15 14:31:35 Zarutian has joined
408 2013-02-15 14:32:47 techlife has quit (Ping timeout: 255 seconds)
409 2013-02-15 14:34:06 daybyter has joined
410 2013-02-15 14:35:28 techlife has joined
411 2013-02-15 14:35:29 techlife has quit (Max SendQ exceeded)
412 2013-02-15 14:35:59 techlife has joined
413 2013-02-15 14:36:01 techlife has quit (Max SendQ exceeded)
414 2013-02-15 14:36:48 ciphermonk has quit (Ping timeout: 276 seconds)
415 2013-02-15 14:37:10 techlife has joined
416 2013-02-15 14:37:11 techlife has quit (Max SendQ exceeded)
417 2013-02-15 14:38:16 techlife has joined
418 2013-02-15 14:38:17 techlife has quit (Max SendQ exceeded)
419 2013-02-15 14:39:09 techlife has joined
420 2013-02-15 14:39:12 techlife has quit (Max SendQ exceeded)
421 2013-02-15 14:39:29 rdponticelli_ has quit (Quit: No Ping reply in 180 seconds.)
422 2013-02-15 14:39:41 techlife has joined
423 2013-02-15 14:39:44 techlife has quit (Max SendQ exceeded)
424 2013-02-15 14:41:30 techlife has joined
425 2013-02-15 14:41:31 techlife has quit (Max SendQ exceeded)
426 2013-02-15 14:43:05 techlife has joined
427 2013-02-15 14:43:07 techlife has quit (Max SendQ exceeded)
428 2013-02-15 14:44:22 techlife has joined
429 2013-02-15 14:44:36 MobiusL has quit (Ping timeout: 276 seconds)
430 2013-02-15 14:44:37 t7 has joined
431 2013-02-15 14:45:39 MobiusL has joined
432 2013-02-15 14:54:30 t7 has quit (Read error: No route to host)
433 2013-02-15 14:56:00 agricocb has quit (Quit: Leaving.)
434 2013-02-15 14:57:23 gavinandresen has joined
435 2013-02-15 14:57:47 CodeShar_ has quit (Remote host closed the connection)
436 2013-02-15 15:01:21 techlife has quit (Ping timeout: 260 seconds)
437 2013-02-15 15:03:45 techlife has joined
438 2013-02-15 15:03:47 techlife has quit (Max SendQ exceeded)
439 2013-02-15 15:04:40 techlife has joined
440 2013-02-15 15:06:15 <Jouke> When I want to get the public hash of a standard transaction, can I just remove "76a914" from the beginning and "88ac" from the end?
441 2013-02-15 15:07:13 <jgarzik> Jouke: Your question does not make sense. Each transaction is constructed, and then a hash is performed over all its data.
442 2013-02-15 15:07:18 <sipa> what is a 'public hash' ?
443 2013-02-15 15:07:24 agath has quit (Read error: No route to host)
444 2013-02-15 15:07:53 agath has joined
445 2013-02-15 15:07:59 <sipa> the txid of a transaction is the double-sha256 hash of its serialized form
446 2013-02-15 15:08:02 <Jouke> Damn it, I am not comfortable with all the terminology
447 2013-02-15 15:08:41 <sipa> i still have no idea what you're asking :0
448 2013-02-15 15:08:47 <Jouke> I basically want to get a bitcoinaddress out of a standard transaction.
449 2013-02-15 15:09:08 <sipa> a transaction doesn't have an addres
450 2013-02-15 15:09:16 <sipa> transaction outputs have output scripts
451 2013-02-15 15:09:27 <sipa> and some output scripts correspond to an address
452 2013-02-15 15:09:29 <Jouke> Sorry, the output svript
453 2013-02-15 15:09:33 <sipa> (in practice, most)
454 2013-02-15 15:09:58 <sipa> ah
455 2013-02-15 15:12:18 ciphermonk has joined
456 2013-02-15 15:13:21 <sipa> Jouke: if the script starts with those 3 bytes, and ends with those 2, you can indeed drop them to obtain the hash160 of the pubkey (which, when encoded in base58, is the address)
457 2013-02-15 15:13:22 <gavinandresen> Jouke: if it is 25 bytes long, begins with 76a914, and ends with 88ac⦠then yes, the 20-byte hash is the other 20 bytes
458 2013-02-15 15:13:51 LargoG has quit (Ping timeout: 276 seconds)
459 2013-02-15 15:13:55 <Jouke> Thanks :)
460 2013-02-15 15:14:19 LargoG has joined
461 2013-02-15 15:16:27 <gavinandresen> just be careful, that is not the only way to encode a standard transaction (you can replace the 0x14 byte with OP_PUSHDATA opcode plus extra bytes)
462 2013-02-15 15:17:29 <Jouke> Does the bitcoinclient at this moment produce such transactions gavinandresen?
463 2013-02-15 15:17:45 <gavinandresen> no
464 2013-02-15 15:18:12 <Jouke> Ok, thanks for the info.
465 2013-02-15 15:18:22 <gavinandresen> When I say "be careful", I mean: think about if an attacker might be able to cause trouble if they encoded it differently
466 2013-02-15 15:18:29 moore_ has joined
467 2013-02-15 15:18:49 <sipa> gavinandresen: are you ok with the comments in #2310?
468 2013-02-15 15:19:34 ciphermonk has quit (Quit: Leaving)
469 2013-02-15 15:20:20 ciphermonk has joined
470 2013-02-15 15:20:45 <gavinandresen> sipa: next time we add a checkpoint, how do we get: total number of transactions between genesis and last checkpoint
471 2013-02-15 15:21:08 <sipa> gavinandresen: look at debug.log (the SetBestChain line tells you tx=...)
472 2013-02-15 15:21:30 <gavinandresen> ah-- that is non-obvious, would be nice if the comment said so
473 2013-02-15 15:21:34 <sipa> ack
474 2013-02-15 15:22:11 Grishnakh has quit (Read error: Connection reset by peer)
475 2013-02-15 15:22:11 <gavinandresen> other than that, looks great
476 2013-02-15 15:22:34 Grishnakh has joined
477 2013-02-15 15:23:50 B0g4r7_ has quit (Remote host closed the connection)
478 2013-02-15 15:23:56 <sipa> i'm afraid about people complaining "oh my, this is SO much slower! 0.8.0rc1 did like 50% in a few minutes, and it's taken an HOUR already here!"
479 2013-02-15 15:24:10 ashams has quit (Read error: Connection reset by peer)
480 2013-02-15 15:24:58 <gavinandresen> if we do our job right, we'll be in exponential growth, and we'll have a lot more new customers than old customers. So new customer experience is more important....
481 2013-02-15 15:30:10 agricocb has joined
482 2013-02-15 15:30:53 Hashdog has joined
483 2013-02-15 15:30:57 SugarPink has joined
484 2013-02-15 15:31:51 <sipa> gavinandresen: update pushed
485 2013-02-15 15:36:38 rdponticelli has joined
486 2013-02-15 15:38:32 meLon has joined
487 2013-02-15 15:38:33 meLon has quit (Changing host)
488 2013-02-15 15:38:33 meLon has joined
489 2013-02-15 15:41:03 Goonie has quit (Remote host closed the connection)
490 2013-02-15 15:54:33 <discrete> ;;genrate 600
491 2013-02-15 15:54:34 <gribble> The expected generation output, at 600.0 Mhps, given difficulty of 3275464.58657, is 0.0921224766392 BTC per day and 0.00383843652663 BTC per hour.
492 2013-02-15 15:54:42 <discrete> ;;genrate 700
493 2013-02-15 15:54:43 <gribble> The expected generation output, at 700.0 Mhps, given difficulty of 3275464.58657, is 0.107476222746 BTC per day and 0.00447817594774 BTC per hour.
494 2013-02-15 15:54:52 <discrete> ;;genrate 900
495 2013-02-15 15:54:53 <gribble> The expected generation output, at 900.0 Mhps, given difficulty of 3275464.58657, is 0.138183714959 BTC per day and 0.00575765478995 BTC per hour.
496 2013-02-15 16:00:41 <jgarzik> ;;genrate 11.2
497 2013-02-15 16:00:42 <gribble> The expected generation output, at 11.2 Mhps, given difficulty of 3275464.58657, is 0.00171961956393 BTC per day and 7.16508151638e-05 BTC per hour.
498 2013-02-15 16:01:26 <jgarzik> so CPU mining would gross US$16.75/year
499 2013-02-15 16:01:44 freakazoid_ has joined
500 2013-02-15 16:03:05 FredEE has joined
501 2013-02-15 16:03:20 * jgarzik compresses his coins down to a single coin
502 2013-02-15 16:03:44 B0g4r7_ has joined
503 2013-02-15 16:07:31 <helo> neutron coin
504 2013-02-15 16:07:48 ciphermonk has quit (Ping timeout: 276 seconds)
505 2013-02-15 16:10:11 moore_ has quit (Ping timeout: 264 seconds)
506 2013-02-15 16:13:39 LargoG has quit (Ping timeout: 276 seconds)
507 2013-02-15 16:15:21 <MC1984> eh
508 2013-02-15 16:18:19 darkee has joined
509 2013-02-15 16:23:01 CodesInChaos has quit (Ping timeout: 244 seconds)
510 2013-02-15 16:24:45 da2ce7 has quit (Read error: Connection reset by peer)
511 2013-02-15 16:26:19 da2ce7 has joined
512 2013-02-15 16:26:49 ThomasV has joined
513 2013-02-15 16:35:33 testnode9 has quit (Ping timeout: 248 seconds)
514 2013-02-15 16:42:00 <gavinandresen> sipa: I've got a reproduceable crash-at-shutown, running bitcoin against an intentionally-corrupt leveldb datadir
515 2013-02-15 16:42:18 <gavinandresen> sipa: crash is caused by deleting CCheckQueue after exitâ¦.
516 2013-02-15 16:43:30 <gavinandresen> sipa: https://gist.github.com/gavinandresen/4961485
517 2013-02-15 16:44:07 <gavinandresen> I wonder if interrupting all of our boost::threads in Shutdown would be an easy fix...
518 2013-02-15 16:50:03 <sipa> gavinandresen: deleting CCheckQueue? it's statically allocated?
519 2013-02-15 16:51:11 <gavinandresen> right, destructors are being called after exit() ...
520 2013-02-15 16:52:11 <sipa> hmm, and why is that a problem? sounds very expected
521 2013-02-15 16:52:14 <gavinandresen> ⦠in some random order, which might be the issue with random crashes at exit
522 2013-02-15 16:52:39 <gavinandresen> I'm seeing some weirdness, recompiling Bitcoin-Qt to rule out a bad build
523 2013-02-15 16:52:55 <TD> shutdown in a threaded program is known to be a surprisingly hard problem
524 2013-02-15 16:53:08 grau has joined
525 2013-02-15 16:53:36 <TD> sipa: fyi some core google libraries (like bigtable) don't actually support clean shutdown at all. they are written on the assumption that they'll only ever be cleanly killed by the kernel
526 2013-02-15 16:53:42 <TD> for exactly this reason
527 2013-02-15 16:53:49 <TD> we don't have that luxury, i guess
528 2013-02-15 16:54:08 <sipa> yeah, crash-only software is a nice idea
529 2013-02-15 16:54:37 <gavinandresen> does Windows have a "kill me right now" signal/whatever ?
530 2013-02-15 16:57:34 <TD> _exit
531 2013-02-15 16:57:38 <TD> it bypasses shutdown handlers
532 2013-02-15 16:57:39 <broomkorn> http://serverfault.com/questions/151196/how-is-the-windows-kill-process-works
533 2013-02-15 16:57:54 <TD> i believe _exit is a standard thing, actually
534 2013-02-15 16:57:55 <broomkorn> http://support.microsoft.com/kb/178893
535 2013-02-15 16:58:10 <TD> http://msdn.microsoft.com/en-us/library/6wdz5232(v=vs.80).aspx
536 2013-02-15 16:58:33 <TD> note that exit() also flushes stream buffers, which is probably useful for logging. so if you switch to _exit it'd be needed to do it manually
537 2013-02-15 16:58:41 <TD> though for software of the complexity of bitcoin clean shutdown should be achievable
538 2013-02-15 16:59:26 <TD> the other possibility of course is to not allocate CCheckQueue statically and then delete it yourself, which achieves a predictable ordering
539 2013-02-15 16:59:39 <TD> the problem is ThreadScriptCheck is still waiting on it
540 2013-02-15 17:00:28 <TD> so the final possibility is to deliberately leak CCheckQueue and suppress valgrind warnings
541 2013-02-15 17:01:54 <gavinandresen> Here's a radical idea: have Shutdown wait only on the wallet.dat-flushing thread, then _exit().
542 2013-02-15 17:05:19 <TD> could work
543 2013-02-15 17:06:00 <sipa> gavinandresen: i think we'll forget things, and cause more problems
544 2013-02-15 17:06:04 <TD> though i wonder what happens if you _exit in the middle of leveldb doing things
545 2013-02-15 17:06:10 <TD> clean shutdown does seem very much preferable
546 2013-02-15 17:06:26 <TD> i'd just leak the CCheckQueue. no biggie.
547 2013-02-15 17:08:21 <gavinandresen> TD: yes, that's probably the right answer for now. That's what we do for the logging mutex in our OutputDebugStringF(), which has to work after exit()
548 2013-02-15 17:08:23 <sipa> in theory, ConnectBlock cannot exit while the checkqueue is active still
549 2013-02-15 17:08:40 <TD> what about leveldb shutdown?
550 2013-02-15 17:08:44 <TD> i don't remember writing any code for that
551 2013-02-15 17:08:54 swappermall_ has joined
552 2013-02-15 17:10:27 <gmaxwell> oh hey, did someone else hit that CCheckQueue mutex deallocation hang?
553 2013-02-15 17:10:50 <sipa> TD: there's a sync flush to both leveldb databases, at least
554 2013-02-15 17:10:52 <gavinandresen> gmaxwell: crashes reproducibly for me on a corrupt datadir
555 2013-02-15 17:10:57 <TD> ok
556 2013-02-15 17:11:19 <sipa> TD: and a delete of the db, and the env
557 2013-02-15 17:12:12 <gmaxwell> gavinandresen: that sounds similar to my reproduction case tooâ re-coin-indexing a blktree that had been zzufed. Sipa and I spent a while banging our heads on it and.. It wasn't reliably reproducable for meâ would take about 100 tries or so.
558 2013-02-15 17:12:14 freakazoid_ has quit (Ping timeout: 252 seconds)
559 2013-02-15 17:12:26 <gavinandresen> I looked hard at the shutdown code a long time ago, came up with a plan to make it better, then got busy with other things.
560 2013-02-15 17:12:28 <TD> ok
561 2013-02-15 17:12:43 <gmaxwell> gavinandresen: interestingly, making the mutex inside the CCheckQueue a global made the problem 'go away'.
562 2013-02-15 17:13:25 <gmaxwell> still, it makes no sense. All of the validators should be stopped long before anything exits.
563 2013-02-15 17:14:09 <gavinandresen> gmaxwell: it's not a normal exit in this case, I think it is still somewhere in AppInit2
564 2013-02-15 17:14:12 <sipa> all validators should be stopped before ConnectBlock _can_ exit
565 2013-02-15 17:14:39 <sipa> though that doesn't mean they're necessary exited
566 2013-02-15 17:15:50 <sipa> i'll try to have a look at the shutdown cases for ccheckqueue this weekend
567 2013-02-15 17:16:55 esreskkesketske has joined
568 2013-02-15 17:17:00 <esreskkesketske> hi
569 2013-02-15 17:17:02 <esreskkesketske> need help
570 2013-02-15 17:17:54 esreskkesketske has quit (Client Quit)
571 2013-02-15 17:18:42 <gavinandresen> sipa: order of events for the crash I'm seeing is: ⦠Step 4 in init.cpp spawns threads. Step 7: load the block chain exits with InitError(). exit() called, crash in mutex destructor.
572 2013-02-15 17:18:56 grau_ has joined
573 2013-02-15 17:19:51 <gavinandresen> TD: you asked about paymentrequest work yesteday: yes, I've started plowing through the "client" side of the TODO list. I bit the bullet and read a couple books about Qt programming...
574 2013-02-15 17:20:01 <TD> cool. Qt is a pretty nice API.
575 2013-02-15 17:20:05 <TD> it shouldn't prove too hard.
576 2013-02-15 17:20:13 <gavinandresen> yes, I'm really liking Qt
577 2013-02-15 17:20:13 <TD> i have some java code for validating requests, but i never finished it
578 2013-02-15 17:20:36 rbecker is now known as RBecker
579 2013-02-15 17:20:43 <TD> there are still higher priorities around correctness and performance, but i guess i'll see how quickly you make progress
580 2013-02-15 17:21:35 grau has quit (Ping timeout: 252 seconds)
581 2013-02-15 17:21:39 <gavinandresen> TD: we never nailed down details of mime types, or if we're going to extend bitcoin: URIs for payment requests
582 2013-02-15 17:21:46 <TD> that is true
583 2013-02-15 17:21:56 <TD> those should be fairly simple and uncontroversial things though?
584 2013-02-15 17:22:02 <TD> application/x-bitcoin-payment-request
585 2013-02-15 17:22:32 <TD> bitcoin:address?value=1.23&req=https://merchant.com/req/1a2b3c
586 2013-02-15 17:22:35 <gavinandresen> I did: applications/x-btc-signed-payment-request and x-btc-payment-request
587 2013-02-15 17:23:01 <gavinandresen> (blah, application/.â¦)
588 2013-02-15 17:23:17 <TD> oh yeah. i remember us debating whether it was better to have two mime types or just make pki_data optional
589 2013-02-15 17:23:35 <gavinandresen> ah, right-- with a pki_type of "none" ?
590 2013-02-15 17:23:42 toffoo has joined
591 2013-02-15 17:23:48 axhlf has quit (Remote host closed the connection)
592 2013-02-15 17:24:15 <TD> sure, or missing also
593 2013-02-15 17:24:18 <gavinandresen> do you think the bitcoin clients will always pull the payment request from the server?
594 2013-02-15 17:24:42 <TD> no. i expect some to fetch it via bluetooth, NFC, load them from email attachments, receive them via Android intents and other things
595 2013-02-15 17:24:58 <TD> but for opening a wallet app from a web browser, it seems ideal. and if you have something that contains a URL you probably have a server also
596 2013-02-15 17:25:08 <TD> so we may as well start there
597 2013-02-15 17:25:13 <gavinandresen> and it fits in a QR code nicely.
598 2013-02-15 17:25:28 <TD> a URL does. a fully blown payment request might not.
599 2013-02-15 17:25:47 <TD> at least not one that scans quickly and reliably. but you want a server anyway, really, otherwise you end up re-using addresses
600 2013-02-15 17:25:53 <gavinandresen> right-- the alternative is : click on a link, and the server serves up the payment request data directly
601 2013-02-15 17:26:06 <TD> i'm sure there'll be cheap hosting services that generate payment requests for you
602 2013-02-15 17:26:07 <TD> yeah
603 2013-02-15 17:26:14 <gavinandresen> ⦠although I can't figure out how to make my Mac Do The Right Thing with a payment request sent directly
604 2013-02-15 17:26:20 <TD> that also works. extending a bitcoin: uri has the nice advantage of backwards compatibility
605 2013-02-15 17:27:09 <TD> otherwise you'll get websites giving uses complicated instructions
606 2013-02-15 17:27:11 <TD> which would kinda suck
607 2013-02-15 17:27:28 <TD> and more importantly, set low UI standards for the community.
608 2013-02-15 17:27:41 <TD> so i like extending the custom URI. it also works quite nicely with mobile platforms.
609 2013-02-15 17:34:51 <gavinandresen> TD: I'm still fuzzy on if or how a webwallet would handle bitcoin: URIs. And even fuzzier on if or how a webwallet could handle the click-and-get-paymentrequest-data â¦.
610 2013-02-15 17:35:09 <TD> HTML5 allows websites to register themselves as URL handlers and it's now fairly widely supported
611 2013-02-15 17:35:41 <gavinandresen> mime type handlers, too? or just new URL schemes?
612 2013-02-15 17:36:03 <TD> https://developer.mozilla.org/en-US/docs/DOM/navigator.registerProtocolHandler
613 2013-02-15 17:36:07 <TD> just new URL schemes AFAIK
614 2013-02-15 17:36:30 <gavinandresen> TD: mmm: http://dev.opera.com/articles/view/html5-custom-protocol-and-content-handlers/ â¦. says mime types, too....
615 2013-02-15 17:37:13 <TD> cool
616 2013-02-15 17:39:03 daybyter has quit (Quit: Konversation terminated!)
617 2013-02-15 17:42:03 <TD> gavinandresen: btw what's the current ETA for 0.8 final?
618 2013-02-15 17:42:37 <gavinandresen> darn good question. I'm waffling on whether or not we can live with the random-crash-on-exit bug
619 2013-02-15 17:43:24 <TD> *shrug* you can always follow up with a 0.8.1
620 2013-02-15 17:44:07 <gavinandresen> true. besides that bug, and some reports of db corruption (that might just be flaky hard drives), 0.8 feels very solid
621 2013-02-15 17:44:45 <TD> good. it's time to get it out there then.
622 2013-02-15 17:44:45 moore_ has joined
623 2013-02-15 17:45:10 <Arnavion> Make the application safe to crash! Then "crashing the application to quit it super fast" can be a feature!
624 2013-02-15 17:45:19 <TD> i want to get new MultiBit and android wallet releases out maybe a week after 0.8, so there are enough nodes to not get smacked by every client trying to use bloom filtering at once
625 2013-02-15 17:45:34 <TD> assuming people upgrade fairly fast of course
626 2013-02-15 17:46:26 <TD> which i am skeptical about â¦. i bet 90%+ of people running nodes have no way to find out about new ones
627 2013-02-15 17:46:28 <TD> an alert might be appropriate
628 2013-02-15 17:47:24 Cory has quit (Ping timeout: 264 seconds)
629 2013-02-15 17:48:39 <gmaxwell> An alert would be crying wolf here.
630 2013-02-15 17:49:03 <gmaxwell> Uptake of new versions has been reasonably quick in the past.
631 2013-02-15 17:49:11 <TD> yeah. i guess we can see how it goes.
632 2013-02-15 17:49:35 <gmaxwell> ... though not 'a week' quick, even with alerts.
633 2013-02-15 17:49:46 BTCOxygen has joined
634 2013-02-15 17:50:25 dhill has joined
635 2013-02-15 17:50:30 * TD -> home
636 2013-02-15 17:50:49 <dhill> latest boost didn't fix my issue.. now trying db 4.8.30 .. instead of 4.6.21
637 2013-02-15 17:51:24 TD has quit (Quit: TD)
638 2013-02-15 17:51:26 <gmaxwell> Causing fast uptake would vastly increase risk in any case. It will be good for 0.8 to roll slowly. We had previously discussed explicitly discouraging miners and merchants from running it. I don't think we'll need to do that now... but still, no need to rush in.
639 2013-02-15 17:51:54 BTCOxygen has quit (Ping timeout: 252 seconds)
640 2013-02-15 17:52:55 grau_ has quit (Remote host closed the connection)
641 2013-02-15 17:56:19 <sipa> ;;bc,tslb
642 2013-02-15 17:56:20 <gribble> Error: unexpected EOF while parsing (<string>, line 1)
643 2013-02-15 17:58:59 dust-otc has joined
644 2013-02-15 18:02:48 <jgarzik> gmaxwell: in the current state, I think encouraging miners to use 0.8 is nice
645 2013-02-15 18:02:55 <jgarzik> and full nodes
646 2013-02-15 18:05:58 <andytoshi> dumb question: is there any firewall setup needed to run a full node properly?
647 2013-02-15 18:06:31 <sipa> unless your firewall by default blocks all incoming and outgoing traffic, no
648 2013-02-15 18:07:35 <andytoshi> thx, i'm good then
649 2013-02-15 18:09:44 grau has joined
650 2013-02-15 18:10:06 t7 has joined
651 2013-02-15 18:11:38 grau has quit (Remote host closed the connection)
652 2013-02-15 18:13:54 PhantomSpark has quit (2!~kvirc@pool-71-251-16-105.nycmny.fios.verizon.net|Ping timeout: 276 seconds)
653 2013-02-15 18:16:26 Lolcust has quit (Ping timeout: 255 seconds)
654 2013-02-15 18:17:38 Lolcust has joined
655 2013-02-15 18:25:24 <dhill> um, i don't want to jinx myself, but i am thinking db 4.8 is working
656 2013-02-15 18:25:43 <dhill> only to block 132000 so far, but we shall see
657 2013-02-15 18:26:10 <andytoshi> congrats :)
658 2013-02-15 18:26:16 <dhill> DONT JINX IT!!
659 2013-02-15 18:26:20 <dhill> :)
660 2013-02-15 18:27:31 <dhill> will keep you posted
661 2013-02-15 18:27:36 * andytoshi holds his breath
662 2013-02-15 18:27:50 <gmaxwell> jgarzik: it's nice unless there is another lurking fork creating bug, and then it's not nice. â I'm not opposed to it now, but I do think that I'd be opposed to using alerts for it.
663 2013-02-15 18:34:51 freakazoid_ has joined
664 2013-02-15 18:36:21 moore_ has quit (Ping timeout: 276 seconds)
665 2013-02-15 18:36:28 drizztbsd has quit (Remote host closed the connection)
666 2013-02-15 18:36:56 ThomasV has quit (Ping timeout: 252 seconds)
667 2013-02-15 18:47:41 enolan has joined
668 2013-02-15 18:49:58 <sipa> ;;bc,nethash
669 2013-02-15 18:50:02 <gribble> 26960.757628215793
670 2013-02-15 18:50:54 TD has joined
671 2013-02-15 19:00:37 dust-otc has quit (Remote host closed the connection)
672 2013-02-15 19:04:32 twobitcoins_ has joined
673 2013-02-15 19:05:40 FredEE_ has joined
674 2013-02-15 19:07:31 twobitcoins has quit (Ping timeout: 248 seconds)
675 2013-02-15 19:10:14 FredEE has quit (Ping timeout: 248 seconds)
676 2013-02-15 19:10:14 FredEE_ is now known as FredEE
677 2013-02-15 19:15:41 CodesInChaos has joined
678 2013-02-15 19:15:47 Skav is now known as MobGod
679 2013-02-15 19:20:51 echius has quit (Quit: Leaving.)
680 2013-02-15 19:28:19 Pasha has joined
681 2013-02-15 19:35:30 SugarPink has quit (Read error: Connection reset by peer)
682 2013-02-15 19:36:21 agricocb has quit (Remote host closed the connection)
683 2013-02-15 19:40:43 <dhill> so, if test/wallet_tests.cpp fail .. is that something to be concerned about?
684 2013-02-15 19:41:04 RBecker is now known as rbecker
685 2013-02-15 19:41:15 <gmaxwell> Yes.
686 2013-02-15 19:41:48 eoss has joined
687 2013-02-15 19:41:48 HM has quit (Read error: Connection reset by peer)
688 2013-02-15 19:42:03 <dhill> test/wallet_tests.cpp(107): error in "coin_selection_tests": check setCoinsRet.size() == 3 failed [4 != 3]
689 2013-02-15 19:42:06 <dhill> test/wallet_tests.cpp(146): error in "coin_selection_tests": check nValueRet == 18 * CENT failed [19000000 != 18000000]
690 2013-02-15 19:42:09 <dhill> test/wallet_tests.cpp(158): error in "coin_selection_tests": check nValueRet == 11 * CENT failed [14000000 != 11000000]
691 2013-02-15 19:42:12 <dhill> those 3 lines just repeat over and over
692 2013-02-15 19:42:18 <Luke-Jr> BSD?
693 2013-02-15 19:42:21 <dhill> yup
694 2013-02-15 19:43:20 HM has joined
695 2013-02-15 19:44:53 <Luke-Jr> dhill: pretty sure it's known-broken on BSD. interested in becoming the BSD dev? :P
696 2013-02-15 19:45:43 <andytoshi> these are weird bugs.. setCoinsRet.size() should be algorithmically determined, no?
697 2013-02-15 19:46:27 BTCOxygen has quit (1!~BTCOxygen@199.193.119.32|Changing host)
698 2013-02-15 19:46:27 BTCOxygen has joined
699 2013-02-15 19:46:31 BTCOxygen is now known as 1!~BTCOxygen@unaffiliated/mroxy/bot/btcoxygen|BTCOxygen
700 2013-02-15 19:46:42 <andytoshi> it's the number of inputs needed to get some value, given a crappy set of availble inputs
701 2013-02-15 19:47:28 BTCOxygen is now known as BTCOxygen|1
702 2013-02-15 19:47:36 BTCOxygen is now known as 1!~BTCOxygen@unaffiliated/mroxy/bot/btcoxygen|BTCOxygen|2
703 2013-02-15 19:47:42 BTCOxygen is now known as 2!~BTCOxygen@unaffiliated/mroxy/bot/btcoxygen|BTCOxygen|3
704 2013-02-15 19:47:48 BTCOxygen is now known as 3!~BTCOxygen@unaffiliated/mroxy/bot/btcoxygen|BTCOxygen
705 2013-02-15 19:47:55 <dhill> there are warnings when compiling wallet_tests.cpp
706 2013-02-15 19:48:29 <dhill> comparison between signed and unsigned integer expressions
707 2013-02-15 19:49:32 <dhill> /usr/local/include/boost/test/test_tools.hpp:536: warning: comparison between signed and unsigned integer expressions
708 2013-02-15 19:50:08 <Luke-Jr> dhill: I'd start looking there
709 2013-02-15 19:50:13 <dhill> template <class Left, class Right>
710 2013-02-15 19:50:13 <dhill> predicate_result equal_impl( Left const& left, Right const& right )
711 2013-02-15 19:50:13 <dhill> { return left == right;
712 2013-02-15 19:50:25 <dhill> }
713 2013-02-15 19:50:35 <Luke-Jr> almost certainly the bug is higher-up
714 2013-02-15 19:51:36 <dhill> http://gbpaste.org/Lbj6g
715 2013-02-15 19:51:54 <andytoshi> can you check if sizeof (int64) is 8 on your BSD?
716 2013-02-15 19:52:00 <andytoshi> it is typedeff'd to long long iirc, so it should be
717 2013-02-15 19:52:34 <Luke-Jr> gavinandresen: have you run test_bitcoin on OS X btw?
718 2013-02-15 19:52:40 <Scrat> he was on openbsd/i386 iirc
719 2013-02-15 19:52:55 <Scrat> that combo gotta be 1 in a million :p
720 2013-02-15 19:53:20 <dhill> yes, it is 8
721 2013-02-15 19:53:30 BTCOxygen has quit ()
722 2013-02-15 19:53:37 * Luke-Jr thinks it's pretty reasonable to expect someone running obscure platforms to submit and support pullreqs to fix them <.<
723 2013-02-15 19:53:40 BTCOxygen has joined
724 2013-02-15 19:54:03 CodesInChaos has quit (Read error: Connection reset by peer)
725 2013-02-15 19:54:04 HM has quit (Read error: Connection reset by peer)
726 2013-02-15 19:54:50 <andytoshi> at the start of wallet_test.cpp, we have typedef set<pair<const CWalletTx*,unsigned int> > CoinSet;
727 2013-02-15 19:55:29 <dhill> unsigned int is of course 4
728 2013-02-15 19:55:34 <dhill> is that the 4 != 3 ?
729 2013-02-15 19:55:45 <Luke-Jr> ⦠no
730 2013-02-15 19:56:24 <andytoshi> no, the .size() is not a datatype size
731 2013-02-15 19:57:38 HM has joined
732 2013-02-15 19:57:55 <dhill> any debug statement i can throw in to see how it gets to 4?
733 2013-02-15 19:57:56 <gavinandresen> Luke-Jr: yes, I run test_bitcoin regularly on my mac
734 2013-02-15 19:57:59 <andytoshi> it is the number elements in the set returned from CWallet::SelectCoinsMinConf in setCoinsRet
735 2013-02-15 19:58:40 <Luke-Jr> dhill: FWIW, the bug is almost certainly in SelectCoinsMinConf
736 2013-02-15 19:58:55 <Luke-Jr> looking at wallet_tests is not really likely to find anything
737 2013-02-15 19:59:14 <dhill> right
738 2013-02-15 19:59:46 <andytoshi> dhill: one sec, i'll throw a patch up
739 2013-02-15 19:59:48 <andytoshi> on pastebin
740 2013-02-15 19:59:50 <dhill> ok
741 2013-02-15 20:00:51 sgornick has joined
742 2013-02-15 20:00:51 HM has quit (Read error: Connection reset by peer)
743 2013-02-15 20:01:17 HM has joined
744 2013-02-15 20:01:17 <dhill> make sure it is against 0.7.2
745 2013-02-15 20:01:28 <dhill> or i can figure it out if it isn't
746 2013-02-15 20:02:29 <andytoshi> dhill: yep, i'm doing so
747 2013-02-15 20:02:43 <andytoshi> your line numbers are wrong for git anyway :)
748 2013-02-15 20:03:05 MobGod has quit (Remote host closed the connection)
749 2013-02-15 20:03:06 paraipan has quit (Remote host closed the connection)
750 2013-02-15 20:03:20 <andytoshi> ..sorry, build is very slow..
751 2013-02-15 20:03:47 <andytoshi> i have new RAM backordered at NCIX
752 2013-02-15 20:04:13 paraipan has joined
753 2013-02-15 20:04:57 mappum has joined
754 2013-02-15 20:05:02 toffoo has quit ()
755 2013-02-15 20:05:08 MobGod has joined
756 2013-02-15 20:05:25 <andytoshi> okay, printf does not work? where is bitcoin_test magic documented?
757 2013-02-15 20:05:32 MobGod is now known as Guest74534
758 2013-02-15 20:05:35 <andytoshi> oh, i got it, stderr works..
759 2013-02-15 20:09:28 <andytoshi> OK, this is a start..
760 2013-02-15 20:10:02 <andytoshi> diff of wallet.cpp
761 2013-02-15 20:10:08 <andytoshi> http://gbpaste.org/6Ygp4
762 2013-02-15 20:10:24 <andytoshi> but that dumps several hundred lines
763 2013-02-15 20:11:15 <dhill> hmm
764 2013-02-15 20:12:04 <dhill> how do i apply that?
765 2013-02-15 20:12:08 <dhill> i am used to patch
766 2013-02-15 20:12:31 <dhill> git diff
767 2013-02-15 20:12:33 <andytoshi> patch doesn't work?
768 2013-02-15 20:12:59 <andytoshi> oh, right, i can use git..
769 2013-02-15 20:13:05 MrTiggr has quit (Ping timeout: 260 seconds)
770 2013-02-15 20:13:19 <dhill> thx
771 2013-02-15 20:13:57 <andytoshi> http://gbpaste.org/JRmfi
772 2013-02-15 20:14:30 <andytoshi> sorry, i was using the 0.72 download, forgot i could just checkout a git tag :}
773 2013-02-15 20:15:16 <dhill> ok, that applied
774 2013-02-15 20:16:27 <andytoshi> alright, i dunno how to make it run that specific line 107 test
775 2013-02-15 20:17:57 <andytoshi> but my RAM just came in, so i'm gonna bike over to NCIX, be back in an hour
776 2013-02-15 20:18:02 andytoshi is now known as andytoshi-away
777 2013-02-15 20:19:49 pooler has joined
778 2013-02-15 20:21:27 zebedee_ has quit (Quit: Leaving)
779 2013-02-15 20:22:29 BTCOxygen is now known as Guest4895472
780 2013-02-15 20:22:35 Guest4895472 has quit ()
781 2013-02-15 20:22:51 BTCOxygen has joined
782 2013-02-15 20:23:04 BTCOxygen has quit (Client Quit)
783 2013-02-15 20:23:19 BTCOxygen has joined
784 2013-02-15 20:23:59 BTCOxygen has quit (Client Quit)
785 2013-02-15 20:24:11 BTCOxygen has joined
786 2013-02-15 20:38:53 agricocb has joined
787 2013-02-15 20:47:03 t7 has quit (Quit: Konversation terminated!)
788 2013-02-15 20:49:06 Jackneill has quit (Ping timeout: 244 seconds)
789 2013-02-15 20:55:56 <dhill> 104 // try making 34 cents from 1,2,5,10,20 - we can't do it exactly
790 2013-02-15 20:55:59 <dhill> 105 BOOST_CHECK( wallet.SelectCoinsMinConf(34 * CENT, 1, 1, vCoins, setCoinsRet, nValueRet));
791 2013-02-15 20:56:02 <dhill> 106 BOOST_CHECK_GT(nValueRet, 34 * CENT); // but should get more than 34 cents
792 2013-02-15 20:56:06 <dhill> 107 BOOST_CHECK_EQUAL(setCoinsRet.size(), 3); // the best should be 20+10+5. it's incredibly unlikely the 1 or 2 got included (but possible)
793 2013-02-15 20:56:09 <dhill> Insert at line 1129 (coin value ``0.20'')
794 2013-02-15 20:56:09 <dhill> Insert at line 1129 (coin value ``0.10'')
795 2013-02-15 20:56:09 <dhill> Insert at line 1129 (coin value ``0.05'')
796 2013-02-15 20:56:09 <dhill> Insert at line 1129 (coin value ``0.01'')
797 2013-02-15 20:56:11 <dhill> SelectCoins() best subset: 0.20 0.10 0.05 0.01 total 0.36
798 2013-02-15 20:56:29 <dhill> wonder why the 0.01 is in there too
799 2013-02-15 21:01:13 Guest74534 is now known as MobGod
800 2013-02-15 21:01:17 owowo has joined
801 2013-02-15 21:01:22 MobGod has quit (Changing host)
802 2013-02-15 21:01:22 MobGod has joined
803 2013-02-15 21:01:22 MobGod has quit (Changing host)
804 2013-02-15 21:01:22 MobGod has joined
805 2013-02-15 21:06:35 HM has quit (Read error: Connection reset by peer)
806 2013-02-15 21:09:49 HM has joined
807 2013-02-15 21:10:42 PhantomSpark has quit (Ping timeout: 276 seconds)
808 2013-02-15 21:10:56 vigilyn has quit (Ping timeout: 255 seconds)
809 2013-02-15 21:14:09 TD has quit (Quit: TD)
810 2013-02-15 21:18:54 CodesInChaos has joined
811 2013-02-15 21:22:49 HM has quit (Read error: Connection reset by peer)
812 2013-02-15 21:23:17 HM has joined
813 2013-02-15 21:25:27 <i2pRelay> <code@i2p> is this like western union
814 2013-02-15 21:28:51 <MC1984> ...no
815 2013-02-15 21:29:45 toffoo has joined
816 2013-02-15 21:30:37 <i2pRelay> <KillYourTV@i2p> some of the visitors make me sad about humanity..
817 2013-02-15 21:30:44 TD has joined
818 2013-02-15 21:31:40 PhantomSpark has joined
819 2013-02-15 21:34:40 andytoshi-away is now known as andytoshi
820 2013-02-15 21:37:36 andytoshi has quit (Quit: gotta reboot)
821 2013-02-15 21:39:31 <i2pRelay> <K1773R@i2p> what a worthless life... waste of organic material
822 2013-02-15 21:39:51 <sipa> ?
823 2013-02-15 21:40:42 <i2pRelay> <K1773R@i2p> these trolls/idiots
824 2013-02-15 21:41:35 <i2pRelay> <KillYourTV@i2p> i was referring to "join <code> is this like western union part"
825 2013-02-15 21:41:52 <i2pRelay> <KillYourTV@i2p> sorry for the noise over there :/
826 2013-02-15 21:42:20 <K1773R> as i said, lets make all the relay chans moderated ;)
827 2013-02-15 21:42:22 <i2pRelay> <KillYourTV@i2p> </ot>
828 2013-02-15 21:43:21 <kytv> t'aint my channels, /me no haz that power
829 2013-02-15 21:43:34 <i2pRelay> <K1773R@i2p> ah well, i forgot we have no OP here
830 2013-02-15 21:43:45 <i2pRelay> <K1773R@i2p> qry
831 2013-02-15 21:43:50 * kytv just runs the bot (and hangs his head in shame when the morons come)
832 2013-02-15 21:44:26 Goonie has joined
833 2013-02-15 21:45:48 PhantomSpark has quit (2!~kvirc@pool-71-251-16-105.nycmny.fios.verizon.net|Ping timeout: 276 seconds)
834 2013-02-15 21:47:44 HM has quit (Read error: Connection reset by peer)
835 2013-02-15 21:49:11 HM has joined
836 2013-02-15 21:50:43 CodesInChaos has quit (Ping timeout: 252 seconds)
837 2013-02-15 21:51:57 [\\\] has quit (Read error: Connection reset by peer)
838 2013-02-15 21:53:52 andytoshi has joined
839 2013-02-15 21:54:18 PhantomSpark has joined
840 2013-02-15 21:54:36 <andytoshi> wow! i upgraded my laptop to 4gb RAM and it's twice as fast
841 2013-02-15 21:54:48 <dhill> riiight
842 2013-02-15 21:55:41 <andytoshi> firefox + clang were constantly fighting over the last 100mb of RAM before
843 2013-02-15 21:58:51 [\\\] has joined
844 2013-02-15 22:00:02 <dhill> andytoshi: SelectCoins() best subset: 0.20 0.10 0.05 0.01 total 0.36
845 2013-02-15 22:00:15 <dhill> when testing the 34cents
846 2013-02-15 22:00:24 <dhill> dunno why it adds in that extra 0.01
847 2013-02-15 22:00:30 <dhill> hunting still
848 2013-02-15 22:00:34 <andytoshi> hmm
849 2013-02-15 22:01:33 <K1773R> maybe it needs em to be used as fees?
850 2013-02-15 22:01:48 axhlf has joined
851 2013-02-15 22:01:56 <andytoshi> K1773R: no, this is in the test suite
852 2013-02-15 22:02:00 <andytoshi> and only dhill is seeing this behavior
853 2013-02-15 22:02:18 <K1773R> ok
854 2013-02-15 22:02:31 <sipa> is there perhaps something wrong with rand()?
855 2013-02-15 22:03:10 <dhill> is there any srand?
856 2013-02-15 22:03:29 <dhill> The rand_r() is a thread-safe version of rand().
857 2013-02-15 22:03:43 <dhill> DESCRIPTION These interfaces are obsoleted by random(3).
858 2013-02-15 22:03:48 <dhill> let me switch it
859 2013-02-15 22:04:20 <andytoshi> it would be nice if we could just seed the RNG in the test suite and that fixed it..
860 2013-02-15 22:04:21 <dhill> sipa: you are talking about rand() in ApproximateBestSubset, right?
861 2013-02-15 22:04:28 <sipa> dhill: yes
862 2013-02-15 22:04:32 <dhill> ok, sec
863 2013-02-15 22:06:44 <andytoshi> but i guess, even with a seed, BSD and linux might do different things
864 2013-02-15 22:09:21 <dhill> um
865 2013-02-15 22:09:23 <dhill> i think it worked
866 2013-02-15 22:09:27 <dhill> no way
867 2013-02-15 22:09:54 <dhill> Insert at line 1102 (coin value ``1.00'')
868 2013-02-15 22:09:54 <dhill> Insert at line 1102 (coin value ``1.00'')
869 2013-02-15 22:09:54 <dhill> Insert at line 1102 (coin value ``1.00'')
870 2013-02-15 22:09:54 <dhill> Insert at line 1102 (coin value ``1.00'')
871 2013-02-15 22:09:54 <dhill> Running 70 test cases...
872 2013-02-15 22:09:57 <dhill> *** No errors detected
873 2013-02-15 22:09:59 <dhill> HAHAHAHA
874 2013-02-15 22:10:02 <dhill> !!!!!!!!!!
875 2013-02-15 22:10:03 <gribble> Error: "!!!!!!!!!" is not a valid command.
876 2013-02-15 22:10:06 <sipa> LOL
877 2013-02-15 22:10:07 <andytoshi> loll
878 2013-02-15 22:10:26 <K1773R> gribble dosnt like u being happy dhill ;)
879 2013-02-15 22:11:11 <andytoshi> dhill: can you post a patch? it would be nice to check on linux and mac
880 2013-02-15 22:11:32 <sipa> you just changed rand() to random()?
881 2013-02-15 22:12:14 <dhill> no, i used openbsd's arc4random .. which is prolly on linux and freebsd and mac.
882 2013-02-15 22:12:19 <dhill> but, let me try with random
883 2013-02-15 22:12:27 <dhill> i think we need to seed
884 2013-02-15 22:12:42 <gmaxwell> I'm still somewhat concerned that the subsetsum is failing on such a trivial problem.
885 2013-02-15 22:12:43 <andytoshi> line 1020? also, don't have a manpage for arc4random here..
886 2013-02-15 22:13:14 <dhill> oh, arc4random on linux is in libbsd .. so we don't want another dependancy
887 2013-02-15 22:13:19 <dhill> let me try random()
888 2013-02-15 22:13:27 <gmaxwell> seems very improbable to me.
889 2013-02-15 22:13:58 <gmaxwell> moreover, changing the random function shouldn't change anything except stir around where the failure happens.. unless rand() is outright broken.
890 2013-02-15 22:14:19 <sipa> gmaxwell: that's my guess, that it returns a constant number or so
891 2013-02-15 22:14:30 <gmaxwell> Yea, that would do it.
892 2013-02-15 22:14:37 <sipa> gmaxwell: turning the "exceptionally unlikely" case into a "always or never"
893 2013-02-15 22:14:43 <dhill> is it seeded anywhere?
894 2013-02-15 22:14:49 <dhill> i don't see an srand() call
895 2013-02-15 22:14:54 <sipa> there is none, no
896 2013-02-15 22:15:11 <gmaxwell> dhill: shouldn't matter.
897 2013-02-15 22:15:23 <andytoshi> random() passes tests over here fwiw
898 2013-02-15 22:15:47 <sipa> on linux, rand and random have the same implementation, afaik
899 2013-02-15 22:15:49 <gmaxwell> can you instrument the test where it fails and print out rand() ?
900 2013-02-15 22:16:10 <gmaxwell> I'm going to guess that your rand() is some busted LCG that always gives even results or something totally insane like that.
901 2013-02-15 22:16:21 <gmaxwell> or maybe gives an even/odd sequence.
902 2013-02-15 22:16:50 <K1773R> random(0, 1) lol
903 2013-02-15 22:16:52 <dhill> so, without the seed, rand() _always_ returns 1103527590 on my machine
904 2013-02-15 22:17:05 <gmaxwell> thats throughly busted.
905 2013-02-15 22:17:26 <dhill> from the manpage
906 2013-02-15 22:17:27 <dhill> If no seed value is provided, the functions are automatically seeded with
907 2013-02-15 22:17:30 <dhill> a value of 1.
908 2013-02-15 22:17:40 <gmaxwell> But in any case, it's not unusual for random functions to be busted.
909 2013-02-15 22:17:41 <gavinandresen> broken randomness on some platform is a catastrophic, insidiuous, "might not notice it is broken until my coins are gone" problem for bitcoin
910 2013-02-15 22:17:59 <gmaxwell> IIRC thats the only place we use the system rand()
911 2013-02-15 22:18:14 <gmaxwell> we don't use the good random because the good random is slow and it makes a visible difference there.
912 2013-02-15 22:18:30 <sipa> dhill: does changing it to random() help?
913 2013-02-15 22:18:55 <dhill> no
914 2013-02-15 22:19:31 <sipa> in that case, i would suggest adding a srand(time()) somewhere in init
915 2013-02-15 22:19:38 <gmaxwell> I suggest we replace that with http://git.xiph.org/?p=opus.git;a=blob;f=tests/test_opus_common.h;h=d1e6b452918ed0e29c8208fcb7682bf174973b49;hb=HEAD#l55 or the KISS99 random, and seed it with the real RNG.
916 2013-02-15 22:19:40 <sipa> oh, and in test_bitcoin
917 2013-02-15 22:20:09 <gmaxwell> sipa: if rand() can return a constant without seeding that must mean it is broken and can go constant in other cases.
918 2013-02-15 22:20:20 <sipa> true
919 2013-02-15 22:20:29 <sipa> that random looks simple enough
920 2013-02-15 22:20:43 <sipa> what period does it have?
921 2013-02-15 22:21:03 <andytoshi> maybe we should write our own shit RNG just to make it consistent?
922 2013-02-15 22:21:16 <dhill> can you use openssl somehow instead?
923 2013-02-15 22:21:21 <andytoshi> we don't need much, and if speed is a concern..
924 2013-02-15 22:21:40 <andytoshi> we've already got an openssl dependency for pretty-much everything else, so that'd be my first response
925 2013-02-15 22:22:02 <sipa> dhill: we do use openssl's random, for almost everything
926 2013-02-15 22:22:05 <sipa> dhill: but it is slow
927 2013-02-15 22:22:22 <gmaxwell> sipa: about 2^60
928 2013-02-15 22:22:43 <sipa> better not initialize it to Rz=0, Rw=0
929 2013-02-15 22:22:45 jcv has joined
930 2013-02-15 22:24:41 <gmaxwell> Wikipedia article on that kind of generator: http://en.wikipedia.org/wiki/Multiply-with-carry Main attraction is that they're as fast as LCGs and generally much better.
931 2013-02-15 22:25:50 <dhill> also, not sure if it matters, but rand_r() is for threaded apps
932 2013-02-15 22:26:02 <gmaxwell> dhill: it doesn't matter.
933 2013-02-15 22:26:11 <dhill> linux manpage says
934 2013-02-15 22:26:13 <dhill> If no seed value is provided, the rand() function is automatically seeded with a value of 1.
935 2013-02-15 22:26:17 <dhill> same as openbsd
936 2013-02-15 22:26:31 <gmaxwell> dhill: can you find the source do your openbsd rand() I'm interested to see how they screwed this up. :)
937 2013-02-15 22:27:07 <sipa> the unit test runs (almost) single-threadedly
938 2013-02-15 22:27:24 <dhill> gmaxwell: http://www.openbsd.org/cgi-bin/cvsweb/~checkout~/src/lib/libc/stdlib/rand.c?rev=1.9;content-type=text%2Fplain
939 2013-02-15 22:28:06 <andytoshi> interesting numbers
940 2013-02-15 22:28:19 <dhill> truly random would be to use arc4random
941 2013-02-15 22:28:23 <dhill> but that is bsd
942 2013-02-15 22:28:35 <sipa> that is not truly random
943 2013-02-15 22:28:41 <sipa> just less predictable
944 2013-02-15 22:28:44 <gmaxwell> lol. It's also irrelevant.
945 2013-02-15 22:28:56 <gmaxwell> And I don't see how that code could give a constant result.
946 2013-02-15 22:29:12 <andytoshi> if RAND_MAX is zero somehow..
947 2013-02-15 22:29:23 <andytoshi> though, that should flip 0,1,0,1 then
948 2013-02-15 22:29:30 <dhill> return (*seed % ((u_int)RAND_MAX + 1));
949 2013-02-15 22:29:35 <dhill> seed is 1
950 2013-02-15 22:29:36 <K1773R> dhill there is no random in our universe...
951 2013-02-15 22:29:41 <gmaxwell> dhill: ...
952 2013-02-15 22:29:42 <dhill> yea yea yea yea
953 2013-02-15 22:29:54 <sipa> if RAND_MAX is 0, the output is always zero
954 2013-02-15 22:29:59 <gmaxwell> dhill: The function increments seed. With a seed of 1 it should return
955 2013-02-15 22:30:16 <dhill> ohhh
956 2013-02-15 22:31:58 <gmaxwell> 1103527590 2524885223 662824084 3295386429 4182499122 ... for seed 1.
957 2013-02-15 22:32:51 <jcv> K1773R: not really usable for any reasonable cases, but all you need is radioactive decay and you do have truely random numbers
958 2013-02-15 22:33:13 <andytoshi> dhill: can you use gdb to step into rand()?
959 2013-02-15 22:33:13 <jcv> just saying, there is random in our universe, just probably not something you want to be using
960 2013-02-15 22:33:20 <dhill> oh
961 2013-02-15 22:33:27 <dhill> 1 sec
962 2013-02-15 22:33:35 <K1773R> why should it be random? just because we dont know the rules yet?
963 2013-02-15 22:33:44 <sipa> jcv: whether or not that is random is a philosophical question
964 2013-02-15 22:33:49 <sipa> but it doesn't matter to us
965 2013-02-15 22:33:59 HM has quit (Ping timeout: 252 seconds)
966 2013-02-15 22:34:10 <Luke-Jr> K1773R: it's random as a simple attempt to improve privacy
967 2013-02-15 22:34:12 <gmaxwell> nor to the subset sum algorithim.
968 2013-02-15 22:34:21 rbecker is now known as RBecker
969 2013-02-15 22:34:30 <dhill> wait
970 2013-02-15 22:34:32 <Luke-Jr> K1773R: a predictable algorithm would make it more likely someone can deduce your wallet from your transactions
971 2013-02-15 22:34:35 <dhill> you are right
972 2013-02-15 22:34:36 <dhill> for (i = 0; i < 10; i++)
973 2013-02-15 22:34:36 <dhill> printf("%d\n", rand());
974 2013-02-15 22:34:41 <gmaxwell> Luke-Jr: actually noâ the subset sum algorithim actually only produces good results if the sequence is 'random'.
975 2013-02-15 22:35:16 <gmaxwell> (so independant of private, the rand there needs to be randomish)
976 2013-02-15 22:35:31 <Luke-Jr> gmaxwell: 1,2,3,4 is perfectly valid as a random output :P
977 2013-02-15 22:35:36 <Luke-Jr> or even 1,1,1,1,1
978 2013-02-15 22:35:55 <sipa> 1,1,1,1,1 would result in a very bad subset sum approximation
979 2013-02-15 22:36:04 <andytoshi> Luke-Jr: i think that -is- our output, and it causes a test failure :)
980 2013-02-15 22:36:11 <Luke-Jr> >_<
981 2013-02-15 22:36:13 <K1773R> yes, as i said. there is no random. i dont need a TRNG (hardware based RNG) for BTC anyway, still this wouldnt be random too
982 2013-02-15 22:36:26 <gmaxwell> not just a test failure, but an objectively bad subset sum solution.
983 2013-02-15 22:36:39 <Luke-Jr> K1773R: without some kind of TRNG, you can't practically have a Bitcoin wallet
984 2013-02-15 22:36:46 <Luke-Jr> K1773R: see that embedded wallet fiasco
985 2013-02-15 22:37:18 <K1773R> fiasco? well if you believe/use deprecated stuff its already programmed to happen :P
986 2013-02-15 22:37:32 * Luke-Jr ponders if HD wallets "solve" that dependency
987 2013-02-15 22:37:42 <gmaxwell> K1773R: you're confused.
988 2013-02-15 22:37:46 <Luke-Jr> K1773R: deprecated? the problem was solely a lack of TRNG IIRC
989 2013-02-15 22:38:04 <gmaxwell> Luke is talking about signing with constant K values.
990 2013-02-15 22:38:13 <K1773R> ah, well thats stupid
991 2013-02-15 22:38:19 <Luke-Jr> gmaxwell: or even predictable, I'd think
992 2013-02-15 22:38:36 <gmaxwell> K1773R: and openbsd doesn't get to "deprecate" C89.
993 2013-02-15 22:39:00 <sipa> technically, the ECDSA K value can be generated deterministically from Hash(message + privatekey)
994 2013-02-15 22:39:04 <K1773R> wasnt talking about openbsd
995 2013-02-15 22:39:27 <sipa> so combined with deterministic wallets, you can reduce the generation of randomness to just the wallet seed
996 2013-02-15 22:39:31 HM has joined
997 2013-02-15 22:39:47 <sipa> but IANAC, so don't trust me :)
998 2013-02-15 22:39:59 <Luke-Jr> sipa: well, you certainly know better than me :P
999 2013-02-15 22:40:12 <gmaxwell> In any case, we still have a mystery here. The openbsd source pretty clearly can't produce a constant output unless miscompiled.
1000 2013-02-15 22:40:20 <Luke-Jr> but either way, you always need a TRNG for the wallet seed
1001 2013-02-15 22:40:32 sgornick has quit (Quit: Ex-Chat)
1002 2013-02-15 22:40:34 <sipa> what's the T in TRNG?
1003 2013-02-15 22:40:38 <Luke-Jr> sipa: True?
1004 2013-02-15 22:40:48 <Luke-Jr> ie, non-pseudo
1005 2013-02-15 22:40:51 <K1773R> TRNG (hardware based RNG)
1006 2013-02-15 22:41:05 <sipa> ideally, sure
1007 2013-02-15 22:41:10 <sipa> but you can do without
1008 2013-02-15 22:41:25 <Luke-Jr> â¦
1009 2013-02-15 22:41:37 <K1773R> the T is somewhat stupid, since its not really "True Random"
1010 2013-02-15 22:41:47 <Luke-Jr> sipa: without a TRNG, you can't properly seed a PRNG
1011 2013-02-15 22:42:01 <sipa> Luke-Jr: apparently people do :)
1012 2013-02-15 22:42:05 <gmaxwell> can the random philosophy stuff go elsewhere? I'm going to miss it if dhill figures out why his random is constant.
1013 2013-02-15 22:42:06 <Luke-Jr> sipa: who?
1014 2013-02-15 22:42:18 <andytoshi> dhill: if you're still here, can you use gdb to step into rand()?
1015 2013-02-15 22:42:19 <dhill> no, i am wrong
1016 2013-02-15 22:42:23 <sipa> Luke-Jr: afaik, I have no hardware based random number generator
1017 2013-02-15 22:42:24 <dhill> for (i = 0; i < 10; i++)
1018 2013-02-15 22:42:24 <dhill> printf("%d\n", rand());
1019 2013-02-15 22:42:28 <andytoshi> oh?
1020 2013-02-15 22:42:28 <dhill> that prints a different number each time
1021 2013-02-15 22:42:33 <Luke-Jr> #bitcoin-mining is idle, move TRNG chat there?
1022 2013-02-15 22:42:47 <dhill> but i have no idea why switching rand() to arc4random() makes the tests pass
1023 2013-02-15 22:43:09 <andytoshi> dhill: what was that pastie site you used?
1024 2013-02-15 22:43:14 <andytoshi> pastebin is blocking my tor node
1025 2013-02-15 22:43:17 <dhill> http://gbpaste.org
1026 2013-02-15 22:43:33 <midnightmagic> I have a hardware-based random number generator..
1027 2013-02-15 22:43:50 <gmaxwell> dhill: how did you get the result that rand was the same every time?
1028 2013-02-15 22:44:55 <andytoshi> you can check your RNG is at least uniform with http://gbpaste.org/dKOeK
1029 2013-02-15 22:44:57 <sipa> dhill: can you just annotate the rand() in ApproximateSubSetSum, and print out the values it generates?
1030 2013-02-15 22:44:59 <andytoshi> well, crudely uniform anyway
1031 2013-02-15 22:45:09 ThomasV has joined
1032 2013-02-15 22:45:09 <sipa> dhill: for the failing case
1033 2013-02-15 22:45:14 <dhill> sec
1034 2013-02-15 22:47:16 FredEE has quit (Quit: FredEE)
1035 2013-02-15 22:47:19 freewil has joined
1036 2013-02-15 22:53:38 brwyatt is now known as Away!~brwyatt@brwyatt.net|brwyatt
1037 2013-02-15 22:54:07 <gmaxwell> andytoshi: uniform by that metric isn't relevant here.
1038 2013-02-15 22:54:44 <gmaxwell> andytoshi: because that rand is a plain LCG over 2^32-1 its outputs will be an even odd sequences.. which is not going to be good for us.
1039 2013-02-15 22:55:10 <gmaxwell> There is basically no entropy in the least significant bit, which is the only one we're using.
1040 2013-02-15 22:55:19 <andytoshi> ah, gotcha
1041 2013-02-15 22:55:26 <gmaxwell> So noâ his rand isn't constant, but its almost as bad.
1042 2013-02-15 22:56:07 <andytoshi> well, maybe RAND_MAX is some big prime
1043 2013-02-15 22:56:41 <andytoshi> though, the choice of numbers there suggests this RNG was supposed to be flagrantly bad
1044 2013-02-15 22:56:52 <andytoshi> to avoid anybody using it unwittingly for secure things
1045 2013-02-15 22:57:41 <dhill> #define RAND_MAX 0x7fffffff
1046 2013-02-15 22:57:48 <dhill> anyways, let me get the rand() values
1047 2013-02-15 22:57:55 <dhill> sorry, had to take a phone call
1048 2013-02-15 22:58:00 <gmaxwell> andytoshi: those aren't bad parameters. They'll produce a maximum period output for that kind of generator.
1049 2013-02-15 22:58:47 <andytoshi> yeah, i did factor them -- but having a default sequence of 1, 12346, ..., seems like a purposeful indicator, that's all
1050 2013-02-15 22:59:09 <gmaxwell> andytoshi: thats not the sequence with a seed of 1.
1051 2013-02-15 22:59:42 <gmaxwell> 1103527590 377401575 662824084 1147902781 2035015474 368800899 1508029952 486256185 1062517886 267834847 is...
1052 2013-02-15 22:59:55 <andytoshi> oh :P never mind, i'm just illiterate
1053 2013-02-15 23:00:15 <gmaxwell> In any case, that RNG will make our subsetsum make the same decisions on every iteration no matter what the seed is when the problem has an even number of coins.
1054 2013-02-15 23:00:28 <gmaxwell> so there is no mystery.
1055 2013-02-15 23:00:59 <andytoshi> maybe we should just use higher bits to extract entropy then
1056 2013-02-15 23:01:23 <gmaxwell> no, we should use the RNG that I linked to or one like it. And never again risk getting screwed by a system RNG.
1057 2013-02-15 23:01:42 Hashdog has quit (Read error: Connection reset by peer)
1058 2013-02-15 23:01:53 <andytoshi> yeah, i guess that's for the best
1059 2013-02-15 23:02:10 <andytoshi> it's sad that the C standard has those functions and allows things like this, seems like a waste
1060 2013-02-15 23:02:32 <gmaxwell> meh, randomness is hard.
1061 2013-02-15 23:02:59 <jrmithdobbs> the standard says that rand() isn't required to have the properties crypto systems think a rng should have
1062 2013-02-15 23:03:02 <gmaxwell> The mistake here is ours, a LCG is the classic implementation of rand().
1063 2013-02-15 23:03:02 <jrmithdobbs> it's pretty straight forward
1064 2013-02-15 23:03:12 <gmaxwell> jrmithdobbs: we're not using it in a cryptosystem.
1065 2013-02-15 23:04:54 sgornick has joined
1066 2013-02-15 23:05:08 <dhill> ok, so
1067 2013-02-15 23:05:11 <jrmithdobbs> gmaxwell: yes, but you were relying on bad assumptions about rand() that you would normally be able to rely on with a cryptographically strong prng
1068 2013-02-15 23:05:18 <dhill> rand:1099508847 rand%2:1
1069 2013-02-15 23:05:18 <dhill> rand:1070329212 rand%2:0
1070 2013-02-15 23:05:18 <dhill> rand:1981289989 rand%2:1
1071 2013-02-15 23:05:18 <dhill> rand:1809127002 rand%2:0
1072 2013-02-15 23:05:18 <dhill> rand:2049319051 rand%2:1
1073 2013-02-15 23:05:21 <dhill> rand:380134760 rand%2:0
1074 2013-02-15 23:05:39 <dhill> it is always 0,1,0,1,0,1
1075 2013-02-15 23:05:53 one_zero has joined
1076 2013-02-15 23:06:08 <gmaxwell> For some definition of you, :P at least in theory I knew that rand() would do this.
1077 2013-02-15 23:06:18 <jrmithdobbs> yes not you as in gmaxwell ;p
1078 2013-02-15 23:06:32 <gmaxwell> (well lot of good my hypothetical knoweldge did here)
1079 2013-02-15 23:06:54 <sipa> gmaxwell: ACK on putting that 64-bit MWC rand() in bitcoin :)
1080 2013-02-15 23:07:04 <gmaxwell> sipa: yea working on a patch.
1081 2013-02-15 23:07:47 * dhill just happy that is fixed
1082 2013-02-15 23:07:51 <dhill> now i can use it :)
1083 2013-02-15 23:07:59 <andytoshi> thx for your help dhill
1084 2013-02-15 23:08:06 <gmaxwell> sipa: think I should just stick it inside selectcoins?
1085 2013-02-15 23:08:17 <gmaxwell> (and seed it every run with the regular rand functions)
1086 2013-02-15 23:08:27 <Luke-Jr> why not just use OpenSSL random for selectcoins?
1087 2013-02-15 23:08:34 <gmaxwell> because its visibly slow
1088 2013-02-15 23:08:50 <Luke-Jr> even if you just use it once to seed rand()? <.<
1089 2013-02-15 23:08:54 <andytoshi> i think it should be a separate bitcoin_fast_shit_rand() function, for things like this
1090 2013-02-15 23:08:59 <andytoshi> and seeded from openssl on first run
1091 2013-02-15 23:09:01 <gmaxwell> Luke-Jr: !@#!@#!@#!@#!@#!
1092 2013-02-15 23:09:17 <gmaxwell> Luke-Jr: ahem. I mean. it doesn't matter how you seed it. on dhill's machine it will fail.
1093 2013-02-15 23:09:28 <Luke-Jr> gmaxwell: BSD sucks that bad? :P
1094 2013-02-15 23:09:36 <gmaxwell> andytoshi: insecure_rand()
1095 2013-02-15 23:09:46 <dhill> i'd steal arc4random from libbsd
1096 2013-02-15 23:09:49 <gmaxwell> Luke-Jr: the sequence that kind of LCG will produce is even-odd.
1097 2013-02-15 23:09:51 <gmaxwell> dhill: NO
1098 2013-02-15 23:09:54 <dhill> :)
1099 2013-02-15 23:09:58 <sipa> using rand() / (RAND_MAX/2) would also work
1100 2013-02-15 23:10:06 <sipa> instead of rand() % 2
1101 2013-02-15 23:10:08 <gmaxwell> jesus. it's actually important that this one is fast.
1102 2013-02-15 23:10:08 <andytoshi> yeah, insecure_rand() is better :P
1103 2013-02-15 23:10:31 jcv has left ()
1104 2013-02-15 23:10:34 <gmaxwell> it gets 1000 * the number of coins in your wallet random numbers per run.
1105 2013-02-15 23:10:48 <Luke-Jr> int insecure_rand() { return 6; /* chosen by dice roll, guaranteed to be random */ }
1106 2013-02-15 23:10:53 <Luke-Jr> kthx
1107 2013-02-15 23:11:00 <sipa> Luke-Jr: xkcd fail; it's 4
1108 2013-02-15 23:11:06 <Luke-Jr> sipa: my dice said 6
1109 2013-02-15 23:11:12 <dhill> arc4random is extremely fast
1110 2013-02-15 23:11:38 <sipa> dhill: still ridiculously slow compared to gmaxwell's MWC
1111 2013-02-15 23:11:38 <gmaxwell> dhill: it is three orders of of magnitude slowerâ at leastâ than the LCG.
1112 2013-02-15 23:12:19 <dhill> ok
1113 2013-02-15 23:12:23 <sipa> even though it's probably very fast compared to OpenSSL's
1114 2013-02-15 23:12:25 <dhill> well, i will look for it and test it out
1115 2013-02-15 23:13:17 <jrmithdobbs> gmaxwell: why not just use salsa8 with the privkey of the last used key in the last transaction or if there are none then the newest/last privkey in the memory pool as the key and it's compressed/truncated pubkey as the starting nonce, and it reseeds itself
1116 2013-02-15 23:13:45 <sipa> there's really no need for a cryptographic random function here
1117 2013-02-15 23:13:53 <sipa> if we need one, use openssl's random
1118 2013-02-15 23:14:10 <jrmithdobbs> or chacha8, it's almost as fast as mwc on sse4/neon hardware
1119 2013-02-15 23:14:15 <gmaxwell> because of performance reasons one is undesirable. If we are okay with subsetsum being slowerâ we should run more iterations, not use a 'better' rng.
1120 2013-02-15 23:15:03 jdnavarro has quit (Remote host closed the connection)
1121 2013-02-15 23:15:04 <andytoshi> maybe we should use the bits of the input hashes as a RNG :P
1122 2013-02-15 23:15:38 <sipa> andytoshi: do not invoke the wrath of Sergio
1123 2013-02-15 23:15:41 <jrmithdobbs> gmaxwell: which impl did you link I missed it
1124 2013-02-15 23:16:05 <Luke-Jr> sipa: lol
1125 2013-02-15 23:16:08 <andytoshi> http://git.xiph.org/?p=opus.git;a=blob;f=tests/test_opus_common.h;h=d1e6b452918ed0e29c8208fcb7682bf174973b49;hb=HEAD#l55
1126 2013-02-15 23:16:46 <Luke-Jr> sipa: as unimportant as some of his exploits might seem at times, I think Sergio is an important part of the team ;)
1127 2013-02-15 23:16:55 <sipa> agree
1128 2013-02-15 23:18:43 CodeShark has joined
1129 2013-02-15 23:23:38 TD has quit (Quit: TD)
1130 2013-02-15 23:24:54 <gmaxwell> andytoshi: I think I'm going to opt against insecure_rand() simply because the requirements for an insecure_rand() could be pretty application specific, and right now this is the only one I'm aware of in the codebase.
1131 2013-02-15 23:26:03 <andytoshi> yeah, that's a good point
1132 2013-02-15 23:27:25 ThomasV has quit (Quit: Quitte)
1133 2013-02-15 23:29:56 <gmaxwell> https://people.xiph.org/~greg/random.patch but I wonder if I shouldn't just make it constant instead of rand_bytes seeded.
1134 2013-02-15 23:32:37 eoss has quit (Remote host closed the connection)
1135 2013-02-15 23:33:38 <andytoshi> IMHO constant
1136 2013-02-15 23:33:43 <gmaxwell> sipa: your opinion on just making the random sequence constant for every run of ApproximateBestSubset? it would make the tests more determinstic.
1137 2013-02-15 23:34:40 <sipa> gmaxwell: for unit tests that makes sense; for actual usage i'm less sure
1138 2013-02-15 23:34:46 MagicalTux has quit (Excess Flood)
1139 2013-02-15 23:34:59 <sipa> gmaxwell: as Luke-Jr said, there is a chance that it reveals something about your wallet
1140 2013-02-15 23:35:16 <gmaxwell> IIRC the inputs to approximate best subset are already securely randomly ordered, no?
1141 2013-02-15 23:35:21 * gmaxwell checks
1142 2013-02-15 23:35:58 <sipa> they're sorted from low to high
1143 2013-02-15 23:36:06 <gmaxwell> yea, okay. right.
1144 2013-02-15 23:36:09 <gmaxwell> fine enough.
1145 2013-02-15 23:36:48 Guest63658 has joined
1146 2013-02-15 23:36:56 <sipa> before the ApproximateSubSetSum refactor/rewrite by dooglus, they were randonly ordered i think
1147 2013-02-15 23:37:08 <bitmarco> 16:51 < Luke-Jr> gmaxwell: BSD sucks that bad? :P
1148 2013-02-15 23:37:10 <bitmarco> haha
1149 2013-02-15 23:37:16 <bitmarco> linux doesnt suck at all!
1150 2013-02-15 23:37:55 <sipa> iirc, the algorithm used is a very traditional PRNG algorithm, and there's even a traditional advise not to rely on its low-order bits :)
1151 2013-02-15 23:39:43 <gmaxwell> yea, this behavior is well know.. hasn't stopped it from causing a lot of problems. (google for TCP ISN prediction)
1152 2013-02-15 23:40:54 D34TH has joined
1153 2013-02-15 23:53:09 freewil has quit (Remote host closed the connection)
1154 2013-02-15 23:55:58 Pasha is now known as Cory