1 2013-01-31 00:06:30 <doublec> does it mine as many bitcoins as your compute rfrom 1996?
  2 2013-01-31 00:07:04 <BlueMatt> jgarzik: why no p2pool?
  3 2013-01-31 00:09:09 <Luke-Jr> BlueMatt:  he'd need to setup a node for that? plus, why p2pool? :p
  4 2013-01-31 00:09:26 <MC1984> haha its a pc power suply
  5 2013-01-31 00:09:38 <MC1984> at least it looks decent quality
  6 2013-01-31 00:09:42 <BlueMatt> jgarzik: ok, fine, why a centralized pool?
  7 2013-01-31 00:09:50 <MC1984> black pc components = decent right?
  8 2013-01-31 00:10:20 <Luke-Jr> BlueMatt: Eligius isn't centralized
  9 2013-01-31 00:10:28 <BlueMatt> Luke-Jr: yes, I realize that
 10 2013-01-31 00:10:42 <Luke-Jr> BlueMatt: did I miss something then? O.o
 11 2013-01-31 00:10:52 <BlueMatt> umm...when?
 12 2013-01-31 00:12:04 <Luke-Jr> BlueMatt: what makes you think jgarzik is mining on a centralized pool?
 13 2013-01-31 00:12:21 <BlueMatt> didnt he say slush's pool?
 14 2013-01-31 00:12:34 <BlueMatt> yes, he did
 15 2013-01-31 00:12:41 <BlueMatt> http://garzikrants.blogspot.ca/
 16 2013-01-31 00:12:51 <BlueMatt> "Got things going on slush's pool, for a little third party confirmation:"
 17 2013-01-31 00:12:55 <Luke-Jr> hum
 18 2013-01-31 00:13:01 <sipa> i guess he's trying several pools/tools
 19 2013-01-31 00:13:04 <Luke-Jr> well, he was asking about Eligius setup a few mins ago
 20 2013-01-31 00:13:25 <BlueMatt> well, ok...we'll see what he settles for
 21 2013-01-31 00:15:17 <sipa> as long as he's the only asic miner, i guess he can solo mine
 22 2013-01-31 00:15:43 <Luke-Jr> sipa: not really, they used cgminer
 23 2013-01-31 00:15:52 <Luke-Jr> which doesn't support solo mining
 24 2013-01-31 00:16:20 <sipa> i'm sure there's a way
 25 2013-01-31 00:16:30 <slush> yes, setup local pool :)
 26 2013-01-31 00:16:39 <owowo> why not go solo?! ;o)
 27 2013-01-31 00:16:40 <BlueMatt> Im pretty sure (not 100% sure, but I think so) that jgarzik knows how to program
 28 2013-01-31 00:16:52 <sipa> BlueMatt: any evidence to support that?
 29 2013-01-31 00:16:57 <Luke-Jr> slush: yeah, that'd work as well as p2pool
 30 2013-01-31 00:16:59 <BlueMatt> sipa: none, just a wild guess
 31 2013-01-31 00:17:06 <Luke-Jr> BlueMatt: no evidence he has source code yet
 32 2013-01-31 00:17:15 <Luke-Jr> or that Avalon is complying with the GPL
 33 2013-01-31 00:17:43 <BlueMatt> meh...Im pretty sure he can figure out a fun way to get in :)
 34 2013-01-31 00:17:49 <Luke-Jr> ???
 35 2013-01-31 00:17:55 <MC1984> p2pool or bust bro
 36 2013-01-31 00:18:06 <Luke-Jr> MC1984: pool fanboy :P
 37 2013-01-31 00:18:12 <BlueMatt> p2pool ftw
 38 2013-01-31 00:19:52 <MC1984> wouldnt he be like 1/5th of p2pool if he sets the asic on it?
 39 2013-01-31 00:20:50 <BlueMatt> yes
 40 2013-01-31 00:21:18 <BlueMatt> but he could decrease his difficulty to not effect variance of other miner's shares while still increasing block count
 41 2013-01-31 00:21:42 <HM2> what order of magnitude are we talking about with these ASICs?
 42 2013-01-31 00:21:46 <BlueMatt> s/ decrease his difficulty/increase his own personal difficulty/
 43 2013-01-31 00:21:47 <HM2> cpu:gpu:asic
 44 2013-01-31 00:22:06 <Luke-Jr> HM2: you forgot FPGAs
 45 2013-01-31 00:22:19 <MC1984> i wnder if that would piss p2pool ppl off enough that they leave :/
 46 2013-01-31 00:22:36 <BlueMatt> MC1984: read my last statement
 47 2013-01-31 00:22:54 <Luke-Jr> someone should 51% p2pool
 48 2013-01-31 00:22:59 <Luke-Jr> that'd be fun to watch
 49 2013-01-31 00:23:18 <HM2> FPGA and ASIC can't be that far apart, surely
 50 2013-01-31 00:23:18 <sipa> HM2: the device jeff got does 68 GH/s, and is claimed to use 400W
 51 2013-01-31 00:23:28 <sipa> they are miles apart
 52 2013-01-31 00:23:29 <HM2> the ASIC process isn't top of the range nanometer stuff is it
 53 2013-01-31 00:23:33 <BlueMatt> HM2: nooooo
 54 2013-01-31 00:23:34 <MC1984> its a magnitude i think
 55 2013-01-31 00:24:32 <HM2> well a decent GPU can spit out a few GH/s now can't it
 56 2013-01-31 00:24:44 <sipa> HM2: i have 0.8 GH/s ASIC device that consumes 40W
 57 2013-01-31 00:24:52 <Luke-Jr> HM2: BFL's ASICs do 1500 Gh/s at 1500 W
 58 2013-01-31 00:25:05 <MC1984> ALLEGEDLY
 59 2013-01-31 00:25:17 <sipa> HM2: eh, FPGA device
 60 2013-01-31 00:25:18 <HM2> nice
 61 2013-01-31 00:25:23 <sipa> 
ACTION haz no asic

 62 2013-01-31 00:25:35 <HM2> 
ACTION looks shiftily at sipa

 63 2013-01-31 00:25:53 <MC1984> i wonder how far artfors got
 64 2013-01-31 00:26:04 <MC1984> he had asics 2 years ago
 65 2013-01-31 00:26:08 <sipa> HM2: GPU's do up to a few hundred MH/s using a few hundred W
 66 2013-01-31 00:26:11 <sipa> MC1984: S-ASICs
 67 2013-01-31 00:26:17 <BlueMatt> MC1984: Id think if he had kept going he would have stuck around here
 68 2013-01-31 00:26:30 <MC1984> i dont know what that is
 69 2013-01-31 00:26:41 <Luke-Jr> BlueMatt: I heard he left IRC to comply with a NDA
 70 2013-01-31 00:26:49 <HM2> what hash are we talking about here?
 71 2013-01-31 00:26:50 <BlueMatt> Luke-Jr: really...?
 72 2013-01-31 00:26:50 <MC1984> maybe he just retired to the backgorund and printed moeny this whole time
 73 2013-01-31 00:26:51 <sipa> MC1984: the metal part is custom, but the silicon is shared
 74 2013-01-31 00:26:53 <Luke-Jr> MC1984: S-ASIC is basically mass-produced FPGA
 75 2013-01-31 00:26:55 <sipa> HM2: what hash what?
 76 2013-01-31 00:26:59 <Luke-Jr> BlueMatt: that's what I heard
 77 2013-01-31 00:27:12 <MC1984> ok
 78 2013-01-31 00:27:17 <BlueMatt> Luke-Jr: ahhh
 79 2013-01-31 00:27:19 <HM2> as a measure
 80 2013-01-31 00:27:29 <sipa> HM2: i have no idea what you're asking
 81 2013-01-31 00:27:35 <HM2> SHA-256 x 1?
 82 2013-01-31 00:27:43 <sipa> ah, double SHA256
 83 2013-01-31 00:28:10 <sipa> like everything in Bitcoin
 84 2013-01-31 00:28:21 <sipa> except addresses
 85 2013-01-31 00:30:09 <HM2> few hundred MH/s sounds low for a GPU to me
 86 2013-01-31 00:30:28 <sipa> HM2: trust me, people have spent ages on optimizing that :)
 87 2013-01-31 00:30:37 <HM2> I'm sure i read about a GPU doing 4 billion SHA-1/s a while back, and SHA-1 is only about twice as fast as SHA-256?
 88 2013-01-31 00:30:50 <Luke-Jr> sipa: http://codepad.org/LbNjWNJG - slightly ugly; thoughts?
 89 2013-01-31 00:30:59 <HM2> although I guess that puts double SHA-256 in the 500 MH/s range
 90 2013-01-31 00:32:25 <sipa> HM2: https://en.bitcoin.it/wiki/Mining_hardware_comparison
 91 2013-01-31 00:33:44 <sipa> Luke-Jr: hmm, why needed?
 92 2013-01-31 00:33:59 <sipa> Luke-Jr: also, NULL is a valid uint256>
 93 2013-01-31 00:34:01 <sipa> ?
 94 2013-01-31 00:34:03 <Luke-Jr> sipa: that's the one where the results are lost
 95 2013-01-31 00:34:13 <Luke-Jr> I don't know if NULL is valid there.. didn't want to take any chances
 96 2013-01-31 00:34:37 <HM2> there we go, some of those GPUs are up at 700 MH/s
 97 2013-01-31 00:34:49 <sipa> HM2: dual GPU cards
 98 2013-01-31 00:35:06 <Ymgve> http://golubev.com/gpuest.htm
 99 2013-01-31 00:35:14 <Ymgve> sure you're not thinking about MD5 speed?
100 2013-01-31 00:35:33 <sipa> Luke-Jr: why not just use the passed state for everything?
101 2013-01-31 00:35:57 <Luke-Jr> sipa: does the caller want to know about the blocks it isn't processing?
102 2013-01-31 00:36:09 <Luke-Jr> sipa: seems like a way someone could trick another node into being DoS'd
103 2013-01-31 00:36:34 <sipa> Luke-Jr: that would require someone to build upon an invalid block, which itself violates DoS rules
104 2013-01-31 00:36:44 <sipa> Luke-Jr: though i agree it's somewhat ughy
105 2013-01-31 00:37:04 <Luke-Jr> sipa: wouldn't this be processing orphans that are waiting for the block?
106 2013-01-31 00:37:08 <HM2> Ymgve: it was probably a pair of top end radeons
107 2013-01-31 00:37:29 <Luke-Jr> so eg, someone makes an invalid orphan on top of your block, then when your block is received it DoSs the user sending your valid block..
108 2013-01-31 00:37:39 <HM2> it wasn't a bitcoin related thread, it was a discussion on cracking leaked password hashes
109 2013-01-31 00:38:26 <sipa> Luke-Jr: orphans are processed by ProcessBlock or AcceptBlock, in (not in the current code, but I assume you were going to change to) a separate state
110 2013-01-31 00:38:55 <sipa> so those would be calling SetBestChain themself, using a validationstate that is lost
111 2013-01-31 00:39:05 <Luke-Jr> sipa: so you're *certain* it's safe to just use the passed state here then? <.<
112 2013-01-31 00:39:17 <Luke-Jr> I don't follow this code, so I'll just take your word for it if you're sure
113 2013-01-31 00:39:50 <sipa> i'm quite sure - i'll have another look tomorrow
114 2013-01-31 00:39:53 <Luke-Jr> k
115 2013-01-31 00:43:10 <sipa> however, in ProcessBlock, the processing of orphans must certainly be done using a separate state
116 2013-01-31 00:48:02 <Luke-Jr> yep, will get to that next..
117 2013-01-31 00:48:22 <sipa> for ConnectBestBlock, it depends on whether you want to punish a node that happens to send you a block that depends on an invalid block, and causes a reorganisation
118 2013-01-31 00:48:49 <sipa> i must admit i didn't consider this, but i think the answer is yes
119 2013-01-31 00:49:17 <HM2> uint256s can be NULL?
120 2013-01-31 00:49:34 <sipa> HM2: yes, that i wondered about too
121 2013-01-31 00:49:45 <HM2> why does line 10 use 0 but line 23 use NULL
122 2013-01-31 00:50:27 <CodeShark> NULL = 0 in C++, no?
123 2013-01-31 00:50:57 <HM2> it's defined in a standard C header
124 2013-01-31 00:51:01 <HM2> C++ has "nullptr"
125 2013-01-31 00:51:03 <jrmithdobbs> error: undefined usage of macro as lvalue
126 2013-01-31 00:51:04 <jrmithdobbs> ;p
127 2013-01-31 00:51:23 <CodeShark> I mean NULL == 0 :p
128 2013-01-31 00:52:00 <jrmithdobbs> I don't remember in c++ but i know in c89/99 NULL is defined as 0 and says the platform should provide a header that does this
129 2013-01-31 00:52:19 <jrmithdobbs> s/does/defines/
130 2013-01-31 00:52:47 <HM2> well 0 is a valid address
131 2013-01-31 00:52:55 <HM2> it's a worthless construct
132 2013-01-31 00:53:00 <jrmithdobbs> not on any real platform it's not
133 2013-01-31 00:53:04 <jrmithdobbs> at least, not an accesible one
134 2013-01-31 00:53:17 <HM2> unless you live in kernel land :P
135 2013-01-31 00:53:42 <HM2> this is a digression
136 2013-01-31 00:54:12 <HM2> also, in GCC/ GNU libc NULL is actually __null
137 2013-01-31 00:54:21 <HM2> which is a compiler intrinsic, apparently
138 2013-01-31 00:54:23 <HM2> hurray
139 2013-01-31 00:54:46 <HM2> http://gcc.gnu.org/onlinedocs/libstdc++/manual/bk01pt02ch04s03.html
140 2013-01-31 00:55:26 <CodeShark> so it matches pointers first - but failing that, it will match integers
141 2013-01-31 00:56:16 <HM2> something like that
142 2013-01-31 00:56:30 <CodeShark> so it's automatically typecast to an integer
143 2013-01-31 00:57:01 <HM2> Are you saying that makes the code OK? :)
144 2013-01-31 00:58:37 <CodeShark> hmm - perhaps I'm wrong
145 2013-01-31 00:58:52 <CodeShark> well, it will typecast to int - but g++ gives a warning
146 2013-01-31 00:59:03 <CodeShark> warning: converting to non-pointer type ???int??? from NULL [-Wconversion-null]
147 2013-01-31 00:59:53 <CodeShark> and no, it doesn't make the code OK
148 2013-01-31 01:00:09 <CodeShark> in what source file did you see this, HM2?
149 2013-01-31 01:00:30 <HM2> the codepad link earlier
150 2013-01-31 01:00:41 <HM2> that Luke-Jr is working on
151 2013-01-31 01:00:56 <Luke-Jr> HM2: I trashed that on sipa's advice :P
152 2013-01-31 01:01:17 <HM2> lol ok
153 2013-01-31 01:02:50 <HM2> CodeShark: tbh i probably wouldn't have an implicit constructor on an uint256 data type, but everyone has a different style and there's nothing wrong with doing so
154 2013-01-31 01:03:30 <Luke-Jr> HM2: I'd want it to behave exactly like a uint64
155 2013-01-31 01:04:05 <CodeShark> I would also like it to behave like a uint64
156 2013-01-31 01:04:08 <CodeShark> just wider
157 2013-01-31 01:04:58 <CodeShark> once we have 256-bit processors, it will be a primitive datatype :)
158 2013-01-31 01:05:33 <CodeShark> it actually should be a primitive datatype even on 64-bit processors
159 2013-01-31 01:05:33 <HM2> AVX2 introduces 256b int ops
160 2013-01-31 01:05:36 <CodeShark> yeah
161 2013-01-31 01:05:43 <CodeShark> haswell
162 2013-01-31 01:06:03 <HM2> i think GCC has compiler primitives for it
163 2013-01-31 01:06:54 <HM2> http://gcc.gnu.org/onlinedocs/gcc/Vector-Extensions.html
164 2013-01-31 01:06:55 <jrmithdobbs> how exactly can you have a primitive that wont fit in a register?
165 2013-01-31 01:07:02 <jrmithdobbs> (serious question)
166 2013-01-31 01:07:07 <jgarzik> well I'll say this
167 2013-01-31 01:07:12 <Luke-Jr> jrmithdobbs: the ops use multiple registers
168 2013-01-31 01:07:14 <HM2> it's emulated in software but the compiler chooses the best instructions as it sees fit
169 2013-01-31 01:07:14 <jgarzik> punishing the miner via power, and it's surviving
170 2013-01-31 01:07:17 <jgarzik> go team
171 2013-01-31 01:07:34 <Luke-Jr> jrmithdobbs: at least, that's how the 128-bit ops work
172 2013-01-31 01:07:36 <HM2> I'm fairly sure MSVC++ has vecexts as well
173 2013-01-31 01:07:36 <jgarzik> er, oops, wrong channel
174 2013-01-31 01:07:54 <HM2> you might benefit but it'd make your code a bit #ifdefy
175 2013-01-31 01:08:21 <jrmithdobbs> Luke-Jr: but uint128_t isn't a primitive there's just ops that work on 128 bits of data at a time out of two distinct registers
176 2013-01-31 01:08:43 <jrmithdobbs> but maybe I'm being too pedantic on the meaning of "primitive"
177 2013-01-31 01:09:06 <HM2> compiler intrinsic then
178 2013-01-31 01:09:16 <CodeShark> yeah, two levels of "primitive" here
179 2013-01-31 01:09:19 <Luke-Jr> jrmithdobbs: GCC does have __int128
180 2013-01-31 01:09:34 <jrmithdobbs> HM2: gotcha.
181 2013-01-31 01:09:50 <jrmithdobbs> you mean primitive to the compiler not the cpu the code gets compiled for ;p
182 2013-01-31 01:09:58 <HM2> pointers aren't exactly primitives either
183 2013-01-31 01:10:19 <HM2> although i guess you have instructions that dereference
184 2013-01-31 01:10:19 <jrmithdobbs> sure they are, they're just ints
185 2013-01-31 01:10:41 <bsdunx> ya, register spanning IMO
186 2013-01-31 01:11:14 <jrmithdobbs> anyways, I'm going to go back to playing with blake2
187 2013-01-31 01:12:04 <HM2> i miss the days when hashes had cute names like whirlpool and tiger
188 2013-01-31 01:12:38 <jrmithdobbs> HM2: well, the name started off cute
189 2013-01-31 01:12:47 <jrmithdobbs> HM2: LAKE->BLAKE->BLAKE2
190 2013-01-31 01:14:07 <HM2> LAKE -> FLAKE -> FLAKEY
191 2013-01-31 01:14:28 <jrmithdobbs> the b meant something, heh, i forget what
192 2013-01-31 01:14:36 <HM2> Botched?
193 2013-01-31 01:15:01 <jrmithdobbs> eh? there's nothing wrong with any of the iterations of blake afaict
194 2013-01-31 01:15:16 <jrmithdobbs> sha-3 just liked the novel construction of keccak better
195 2013-01-31 01:16:37 <HM2> I'm just poking fun, i know nothing about designer hash functions :P except a few novice things like length extension attack and meet in the middle
196 2013-01-31 01:16:46 <HM2> *designing
197 2013-01-31 01:18:40 <jgarzik> Updated with machine info (cpuinfo, meminfo, dmesg): http://garzikrants.blogspot.com/2013/01/avalon-tonight.html
198 2013-01-31 01:20:49 <HM2> using an xbox controller for size comparison is a touch of class there
199 2013-01-31 01:23:27 <eckey> My Bitcoin-Qt process hung around for 15 minutes after Quit.  OSX 10.8.2 with encrypted disk.  Does Bitcoin write the block chain to disk during termination?
200 2013-01-31 01:24:35 <eckey> Can someone else with OSX open Activity Monitor and quit Bitcoin?  Thanks.
201 2013-01-31 01:25:14 <eckey> The Apple Store guy said my HDD is ok.
202 2013-01-31 01:25:54 <eckey> Bitcoin-Qt 0.7.2
203 2013-01-31 01:28:30 <MC1984> jgarzik threatens to mine on testnet
204 2013-01-31 01:28:32 <MC1984> luls
205 2013-01-31 01:35:45 <Diablo-D3> hey jgarzik
206 2013-01-31 01:35:55 <Diablo-D3> I want proof your thing works
207 2013-01-31 01:36:37 <Diablo-D3> have it connect to p2pool =P
208 2013-01-31 01:38:34 <sipa> eckey: do you have -detachdb enabled?
209 2013-01-31 01:38:37 <BlueMatt> he said he was gonna rotate some pools and see how it works on a number of them
210 2013-01-31 01:38:43 <Diablo-D3> ahh
211 2013-01-31 01:38:49 <Diablo-D3> I want him to perm mine on p2pool though
212 2013-01-31 01:38:56 <eckey> sipa: no
213 2013-01-31 01:38:59 <BlueMatt> Diablo-D3: as do i :)
214 2013-01-31 01:39:05 <sipa> eckey: large -dbcache ?
215 2013-01-31 01:39:06 <eckey> but is this normal?
216 2013-01-31 01:39:09 <sipa> no
217 2013-01-31 01:39:13 <eckey> 15 min?
218 2013-01-31 01:39:52 <sipa> eckey: can you run with -logtimestamps, and then paste the part of debug.log corresponding to your 15 min shutdown?
219 2013-01-31 01:39:53 <mappum> if he mines on p2pool we should see a jump from 340gh/s to 400
220 2013-01-31 01:40:13 <Diablo-D3> we're at 311 atm
221 2013-01-31 01:41:01 <mappum> i'm just looking at http://p2pool.info/
222 2013-01-31 01:43:28 <eckey> this is what i find in db.log--this is after starting Bitcoin-Qt-0.7.2
223 2013-01-31 01:43:32 <eckey> http://pastebin.com/5mHzZFP8
224 2013-01-31 01:44:31 <sipa> hmm, database troubles perhaps
225 2013-01-31 01:44:54 <eckey> should I erase and re-fetch?
226 2013-01-31 01:45:44 <HM2> https://github.com/bitcoin/bitcoin/blob/master/src/uint256.h#L193
227 2013-01-31 01:45:49 <HM2> there's something fishy here
228 2013-01-31 01:46:03 <HM2> pn[] is an array of unsigned ints
229 2013-01-31 01:46:19 <sipa> HM2: hardcoded fallback peers
230 2013-01-31 01:46:37 <HM2> say what?
231 2013-01-31 01:46:42 <sipa> oh
232 2013-01-31 01:46:47 <sipa> never mind
233 2013-01-31 01:46:53 <sipa> i didn't check the link
234 2013-01-31 01:47:49 <sipa> HM2: the conversion is well-defined, but it's strange that we don't see compiler warnings about comparison between signed and unsigned
235 2013-01-31 01:47:57 <HM2> i do
236 2013-01-31 01:47:59 <HM2> :|
237 2013-01-31 01:48:24 <sipa> ?
238 2013-01-31 01:48:34 <HM2> that's how i spotted it, a compiler warning
239 2013-01-31 01:48:43 <sipa> which compiler?
240 2013-01-31 01:48:48 <HM2> gcc 4.7.2
241 2013-01-31 01:49:12 <sipa> ah, i'm on 4.6.3
242 2013-01-31 01:49:55 <HM2> it might be -Wextra
243 2013-01-31 01:49:55 <mappum> i thought if you assign -1 to an unsigned variable it is just 0xffffffff
244 2013-01-31 01:49:57 <sipa> but i know why there's no warning
245 2013-01-31 01:50:07 <mappum> or however many bits
246 2013-01-31 01:50:10 <sipa> that function is never used :)
247 2013-01-31 01:51:51 <HM2> mappum: depends
248 2013-01-31 01:52:09 <HM2> C++ doesn't mandate that the machine you're using be twos compliment
249 2013-01-31 01:52:38 <mappum> i see
250 2013-01-31 01:52:57 <sipa> all supported machines do, however :)
251 2013-01-31 01:53:07 <sipa> and the code is certainly not platform-neutral
252 2013-01-31 01:53:15 <eckey> sipa: can I remove blk????.dat and blkindex.dat, restart, and let it recover the block chain?
253 2013-01-31 01:53:23 <sipa> eckey: yes
254 2013-01-31 01:53:26 <BlueMatt> if you're gonna write your code with the assumption that not everything is twos-complement.............
255 2013-01-31 01:53:40 <mappum> yeah, that could be fixed to just be the max value instead of -1
256 2013-01-31 01:55:40 <Luke-Jr> BlueMatt: then you'd better be prepared to also handle platforms where the endian changes at runtime too
257 2013-01-31 01:55:50 <HM2> actualyl
258 2013-01-31 01:55:53 <BlueMatt> Luke-Jr: lol...yep
259 2013-01-31 01:55:55 <HM2> blergh
260 2013-01-31 01:56:03 <BlueMatt> also...what platform does that?
261 2013-01-31 01:56:23 <sipa> Luke-Jr: does the number base also change? like from 2's complement to 10's complement?
262 2013-01-31 01:56:27 <sipa> and then to base phi
263 2013-01-31 01:56:47 <BlueMatt> and then we go with a decimal binary representation
264 2013-01-31 01:56:55 <sipa> check
265 2013-01-31 01:57:01 <Luke-Jr> sipa: my point is that runtime-endian machines actually exist
266 2013-01-31 01:57:16 <HM2> well C++ will cast that -1 to unsigned int -> UINT_MAX
267 2013-01-31 01:57:29 <sipa> Luke-Jr: well, runtime-chosen endian
268 2013-01-31 01:57:37 <sipa> Luke-Jr: it's not out of your control, right?
269 2013-01-31 01:58:16 <Luke-Jr> sipa: afaik the OS chooses
270 2013-01-31 01:59:00 <HM2> i prefer the term byte sex to endianness
271 2013-01-31 01:59:37 <sipa> Luke-Jr: then let's hope nobody creates an OS that changes your process' endianness unexpectedly :)
272 2013-01-31 02:00:39 <mappum> if your ram shakes too much all the bytes switch
273 2013-01-31 02:00:41 <BlueMatt> 
ACTION gets to kernel hacking so that he can write the most obnoxious bug reports known to man and file them against every project in existence...

274 2013-01-31 02:00:43 <mappum> true story
275 2013-01-31 02:01:37 <sipa> mappum: well if the bytes & the byteorder change at the same time, there's no problem :P
276 2013-01-31 02:01:58 <sipa> ... in theory
277 2013-01-31 02:02:09 <HM2> shoulda used a language without ints, problem solved
278 2013-01-31 02:02:28 <BlueMatt> must only use bytes
279 2013-01-31 02:02:32 <phantomcircuit> sipa, lol
280 2013-01-31 02:02:37 <mappum> sometimes some bits fall out though and it all gets out of sync
281 2013-01-31 02:02:45 <eckey> it's all Intel's fault
282 2013-01-31 02:02:50 <Luke-Jr> sipa: well, I mean where endian is unknown at compile time, not that it changes while it's running :p
283 2013-01-31 02:02:52 <eckey> IBM had it right
284 2013-01-31 02:03:42 <phantomcircuit> the headache of runtime endian code
285 2013-01-31 02:03:48 <phantomcircuit> that would be ridiculousness
286 2013-01-31 02:04:51 <BlueMatt> and here I was thinking google engineers fixed their mistakes quickly....reader has been fucked up all day
287 2013-01-31 02:06:34 <sipa> BlueMatt: the endianness of the pages swapped?
288 2013-01-31 02:07:10 <HM2> Intel also made a mistake ditching segmentation
289 2013-01-31 02:07:36 <BlueMatt> sipa: ohhhh...lemme dig out my big endian system and try
290 2013-01-31 02:07:36 <sipa> flat address spaces are so horrible to program for!
291 2013-01-31 02:07:47 <nibcoin> And not adopting a hardware BCD instruction. Oh wait, they just added that ^_^
292 2013-01-31 02:07:52 <HM2> sipa: sarcasm? :P
293 2013-01-31 02:07:59 <sipa> HM2: no, not at all!
294 2013-01-31 02:08:03 <nibcoin> Way to arrive 40 years late to party
295 2013-01-31 02:08:46 <HM2> lol
296 2013-01-31 02:09:18 <HM2> segmentation had features Intel has yet to reintroduce but have proven useful
297 2013-01-31 02:09:51 <nibcoin> We are still not to the vision of a single globally unified memory space though ;)
298 2013-01-31 02:10:16 <nibcoin> VMoIP
299 2013-01-31 02:10:30 <HM2> Googles NaCl uses it to sandbox their native code. PaX team use it to introduce loads of security features, and Intel are only just adding protection for userland code from kernel space with SMAPS
300 2013-01-31 02:12:34 <HM2> x86 is a bit of mad hatters tea party
301 2013-01-31 02:17:21 <bsdunx> alpha was the greatest cpu arch ever
302 2013-01-31 02:18:21 <jgarzik> bsdunx: damn straight
303 2013-01-31 02:18:36 <HM2> I'm fond of MIPS
304 2013-01-31 02:18:46 <BlueMatt> shit, I think pull-tester has been having some false negatives
305 2013-01-31 02:19:06 <nibcoin> why did you have to die, DEC =(
306 2013-01-31 02:19:22 <BlueMatt> in any case, Im adding some stuff over the next few days, anyone with a pull which got rejected over the past...I dont even know how long, please ping me
307 2013-01-31 02:19:26 <BlueMatt> sorry for the disturbance
308 2013-01-31 02:19:33 <BlueMatt> Luke-Jr: Im pretty sure that includes you
309 2013-01-31 02:20:17 <Luke-Jr> BlueMatt: it does. *ping*
310 2013-01-31 02:20:23 <BlueMatt> which #?
311 2013-01-31 02:20:37 <BlueMatt> (s)
312 2013-01-31 02:20:57 <Luke-Jr> 2243 2241 1816
313 2013-01-31 02:21:03 <BlueMatt> shit....
314 2013-01-31 02:21:08 <Luke-Jr> ?
315 2013-01-31 02:21:18 <BlueMatt> thats a lot of pulls...
316 2013-01-31 02:21:27 <BlueMatt> just means i probably got lots of them
317 2013-01-31 02:25:17 <MobGod> Luke-Jr do you have a min
318 2013-01-31 02:25:37 <Luke-Jr> MobGod: maybe.
319 2013-01-31 02:54:31 <Luke-Jr> sipa: https://github.com/bitcoin/bitcoin/pull/2253
320 2013-01-31 03:35:22 <vessenes> Hi all, we're working on setting up an in-memory blockchain API. I keep beating my head against the wall implementing it, though. I've done two rounds with redis, and two with mongo, plus one with sqlite, and one with mysql. Anyone out there worked on this?
321 2013-01-31 03:36:16 <BlueMatt> use bitcoinj
322 2013-01-31 03:36:21 <BlueMatt> its already half-implemented....
323 2013-01-31 03:36:36 <BlueMatt> as long as you dont mind a custom java db thats not really designed to be incredibly performant
324 2013-01-31 03:36:51 <CodeShark> I've done sql dbs
325 2013-01-31 03:37:00 <vessenes> Yeah, exactly.
326 2013-01-31 03:37:12 <BlueMatt> or cheat and use the h2 sql db and put it on a tmpfs
327 2013-01-31 03:37:14 <BlueMatt> lol
328 2013-01-31 03:37:19 <vessenes> This is a possibility
329 2013-01-31 03:37:31 <BlueMatt> no really, dont do that
330 2013-01-31 03:37:34 <vessenes> An example query from redis is appealing right now
331 2013-01-31 03:38:13 <Luke-Jr> the most ideal solution would be to extend the new optional "extra info" indexing in bitcoind IMO - but that's a bit of work :P
332 2013-01-31 03:38:44 <CodeShark> sipa's already gone some ways on that front
333 2013-01-31 03:38:56 <CodeShark> and I'm not sure that's the most ideal solution
334 2013-01-31 03:39:14 <Luke-Jr> CodeShark: ideal because it's a step toward a builtin blockchain.info replacement
335 2013-01-31 03:39:23 <vessenes> lrange a.1dice6YgEVBf88erBFra9BHf6ZMoyvG88.tx 0 -1 returns all txouts in 1.5 seconds from redis.
336 2013-01-31 03:39:29 <vessenes> That's 146307 or so
337 2013-01-31 03:40:05 <CodeShark> Luke-Jr: I would prefer to see bitcoind be a streamlined verification/relay agent and have separate apps for historical data
338 2013-01-31 03:40:22 <vessenes> I'm with Codeshark; I'd like to use bitcoind for just that
339 2013-01-31 03:40:37 <BlueMatt> lol, thats not what bitcoind is designed for at all
340 2013-01-31 03:40:42 <CodeShark> no?
341 2013-01-31 03:40:52 <Luke-Jr> ???
342 2013-01-31 03:40:56 <CodeShark> I thought that was its main function
343 2013-01-31 03:41:02 <vessenes> de-facto for sure
344 2013-01-31 03:41:05 <CodeShark> except for the "Streamlined" part :P
345 2013-01-31 03:41:34 <BlueMatt> yes, its main function is to verify and relay blocks, but its really not streamlined for that, it kinda has a wallet and rpc server hanging off the side.....
346 2013-01-31 03:41:58 <CodeShark> right - but the wallet and RPC stuff could be implemented by other projects without risk to the core network integrity
347 2013-01-31 03:42:05 <vessenes> The issue we face, say at the Foundation is: giant set of public addresses need to get turned into a weekly balance sheet. This is not easy to do automatically with the wallet.
348 2013-01-31 03:42:14 <CodeShark> however, there should be a reference implementation of transaction/block validation and relay
349 2013-01-31 03:43:15 <CodeShark> bitcoind was originally designed to be a full-featured node - to perform all functions necessary to interface the bitcoin network and make transactions
350 2013-01-31 03:43:26 <CodeShark> but I don't really like the monolithic approach
351 2013-01-31 03:43:32 <BlueMatt> CodeShark: bitcoind is the reference implementation of all of that
352 2013-01-31 03:43:54 <CodeShark> right, but certain functions have already been moved out of it - for instance, mining
353 2013-01-31 03:43:55 <BlueMatt> but Im completely with you, re-factoring bitcoind into its clear parts would be nice
354 2013-01-31 03:44:00 <BlueMatt> but Im not gonna get into that...
355 2013-01-31 03:44:11 <CodeShark> the mining code in bitcoind is only good for testnet
356 2013-01-31 03:44:15 <BlueMatt> no, Im pretty sure the mining code is still there
357 2013-01-31 03:44:24 <BlueMatt> that means it hasnt been moved out, its only rarely used....
358 2013-01-31 03:44:28 <CodeShark> yes, it's still there but nobody uses it on the main network
359 2013-01-31 03:44:59 <BlueMatt> hence, not "moved out"
360 2013-01-31 03:45:25 <BlueMatt> vessenes: can you be more specific on your problem here?
361 2013-01-31 03:45:43 <CodeShark> the code hasn't been moved out of it - but the function has
362 2013-01-31 03:45:57 <BlueMatt> ok, then we have different definitions of "moved out"
363 2013-01-31 03:46:01 <CodeShark> and the same thing has been happening with wallets
364 2013-01-31 03:46:20 <BlueMatt> as a developer, I dont care if its used or not, the code is there...in fact unused code is worse than used code that maybe should be removed....
365 2013-01-31 03:46:20 <vessenes> Here's a simple use-case: Given 1,000 public addresses, generate a second-by-second history of balances across all of them. Desired time, less than 250ms
366 2013-01-31 03:46:35 <jgarzik> CodeShark: as it happens, picocoin.git includes "brd" (Block Relay Daemon) which is intended to do nothing but relay TXs and blocks [when finished].
367 2013-01-31 03:46:42 <BlueMatt> CodeShark: huh? bitcoin-qt is kinda used a lot as a wallet....
368 2013-01-31 03:47:02 <CodeShark> I'm not saying it isn't used - but there are alternative wallets that have gained some market share
369 2013-01-31 03:47:16 <BlueMatt> like?
370 2013-01-31 03:47:22 <CodeShark> multibit, armory
371 2013-01-31 03:47:29 <BlueMatt> ahh, ok you mean alt desktop clients
372 2013-01-31 03:47:50 <Luke-Jr> CodeShark: bitcoind was originally designed to be a SPV node just as well as a full one, FWIW
373 2013-01-31 03:48:40 <BlueMatt> vessenes: not sure such an index is already built in any software.  Not to push one project over others, Id find a decent bitcoin library (bitcoinj!) and use that to make a db that gets dumped into which can later be queried
374 2013-01-31 03:48:47 <Luke-Jr> Satoshi just didn't finish that, nor has anyone else - yet
375 2013-01-31 03:49:05 <BlueMatt> vessenes: bitcoinj (should) make this easy by just hooking a listener to the block chain and dumping block/tx data into a db
376 2013-01-31 03:49:35 <CodeShark> I did that just using the p2p protocol
377 2013-01-31 03:49:42 <CodeShark> hooked up a listener to a node
378 2013-01-31 03:49:46 <CodeShark> and dump the data into a db
379 2013-01-31 03:49:49 <BlueMatt> that works too, ofc
380 2013-01-31 03:49:49 <Luke-Jr> vessenes: is that the only use case? you could index (time, scriptPubkey, balance)
381 2013-01-31 03:51:11 <jgarzik> hum
382 2013-01-31 03:51:13 <Luke-Jr> BlueMatt: woo, your pull tester thing passed :P
383 2013-01-31 03:51:26 <jgarzik> is it immediately known, when p2pool finds a mainnet block?
384 2013-01-31 03:51:41 <Luke-Jr> jgarzik: it's announced on #p2pool pretty quick
385 2013-01-31 03:51:43 <jgarzik> I guess you see it in your wallet...
386 2013-01-31 03:51:51 <Luke-Jr> jgarzik: but if it was immediate, we wouldn't need a blockchain ;)
387 2013-01-31 03:51:57 <BlueMatt> jgarzik: wallet gens are only shown after 1 confirm, I believe
388 2013-01-31 03:52:01 <Luke-Jr> jgarzik: Bitcoin-Qt only shows after a 2nd confirm
389 2013-01-31 03:52:18 <Luke-Jr> (my count includes the one builtin to the block itself)
390 2013-01-31 03:52:24 <BlueMatt> Luke-Jr: yes, Ill turn it loose on the list again
391 2013-01-31 03:52:40 <BlueMatt> finally update pull-tester to fix bugs and use some tests I wrote like a month ago.....
392 2013-01-31 03:52:47 <jgarzik> is there a history of p2pool blocks shown anywhere?  say the last 10 p2pool blocks?
393 2013-01-31 03:52:55 <BlueMatt> p2pool.info
394 2013-01-31 03:53:35 <jgarzik> so 1-2 blocks per day seems normal for p2pool
395 2013-01-31 03:56:03 <Luke-Jr> BlueMatt: pull tester can automatically fix our bugs now? :D
396 2013-01-31 03:56:40 <CodeShark> pull tester automatically implements new features now :)
397 2013-01-31 03:57:20 <Luke-Jr> yay, we're all obsolete as humans
398 2013-01-31 03:57:47 <CodeShark> jgarzik: autotools is now checking libraries - but it's still not adding the compiler options to link to them
399 2013-01-31 03:58:07 <CodeShark> I could just manually add them - but I figure there must be a better way
400 2013-01-31 03:58:40 <jgarzik> CodeShark: take a look at config.* to see what gets set, and what not
401 2013-01-31 03:58:59 <jgarzik> CodeShark: linking properly is pretty basic
402 2013-01-31 03:59:12 <jgarzik> CodeShark: Did you import the boost/bdb macros?
403 2013-01-31 03:59:20 <CodeShark> yes
404 2013-01-31 04:00:00 <CodeShark> do I need to add AC_SUBST() in configure.ac?
405 2013-01-31 04:00:55 <jgarzik> CodeShark: typically AC_SUBST is only for AC_CHECK_LIB env vars, e.g.
406 2013-01-31 04:00:57 <jgarzik> AC_CHECK_LIB(jansson, json_loads, JANSSON_LIBS=-ljansson,
407 2013-01-31 04:00:57 <jgarzik> [AC_MSG_ERROR([Missing required libjansson])])
408 2013-01-31 04:01:03 <jgarzik> AC_SUBST(JANSSON_LIBS)
409 2013-01-31 04:01:05 <CodeShark> yeah, that's what I was thinking
410 2013-01-31 04:01:56 <CodeShark> should the AX macros automatically take care of this for me?
411 2013-01-31 04:02:04 <CodeShark> I still have no idea how it all really works
412 2013-01-31 04:02:42 <vessenes> Thanks for the messages guys. I'm comfortable I'm not reinventing the wheel at least. :)
413 2013-01-31 04:02:45 <jgarzik> CodeShark: presuming you call the macros, they should...
414 2013-01-31 04:02:48 <jgarzik> http://www.gnu.org/software/autoconf-archive/ax_boost_filesystem.html#ax_boost_filesystem
415 2013-01-31 04:02:49 <jgarzik> says
416 2013-01-31 04:02:55 <jgarzik> AC_SUBST(BOOST_FILESYSTEM_LIB)
417 2013-01-31 04:02:55 <jgarzik> This macro calls:
418 2013-01-31 04:03:15 <jgarzik> so  AX_BOOST_FILESYSTEM goes into configure.ac
419 2013-01-31 04:03:27 <jgarzik> then @BOOST_FILESYSTEM_LIB@ goes into bitcoind_LDADD
420 2013-01-31 04:03:56 <BlueMatt> Luke-Jr: no, it doesnt bother with that, it just reads descriptions of pulls and writes them all from scratch
421 2013-01-31 04:04:34 <Luke-Jr> jgarzik: wouldn't he want it in just LDADD, since it's needed for bitcoin-qt and test_bitcoin as well?
422 2013-01-31 04:04:56 <CodeShark> jgarzik: so I need to explicitly add the line bitcoind_LDADD=@BOOST_FILESYSTEM_LIB@ to Makefile.am?
423 2013-01-31 04:05:25 <jgarzik> CodeShark: yes
424 2013-01-31 04:05:48 <CodeShark> and that will include the -L and the -l?
425 2013-01-31 04:05:53 <jgarzik> CodeShark: yes
426 2013-01-31 04:06:33 <jgarzik> CodeShark: if your libraries are in weird places, like /usr/local/lib on freebsd, the user is expected to tell this to the configure script, like
427 2013-01-31 04:06:40 <jgarzik> LDFLAGS="/usr/local/include" ./configure
428 2013-01-31 04:06:49 <CodeShark> right - I'm just doing run-of-the-mill ubuntu for now
429 2013-01-31 04:06:57 <CodeShark> it's all in /usr/include
430 2013-01-31 04:07:16 <CodeShark> and /usr/lib
431 2013-01-31 04:07:45 <Luke-Jr> BlueMatt: what is "b64"?
432 2013-01-31 04:08:08 <BlueMatt> did it fail?
433 2013-01-31 04:08:41 <BlueMatt> ahh, they're all failing
434 2013-01-31 04:08:42 <BlueMatt> shit
435 2013-01-31 04:09:16 <CodeShark> jgarzik: making progress
436 2013-01-31 04:10:15 <BlueMatt> oh shit, that one needs investigation
437 2013-01-31 04:10:36 <Luke-Jr> is it not my fault? :/
438 2013-01-31 04:10:38 <BlueMatt> gavinandresen: please give me day before you spin rc1, i think its probably a test bug, but Im gonna go sleep and it may not be
439 2013-01-31 04:10:46 <BlueMatt> Luke-Jr: dont think so, two pulls failed back-to-back
440 2013-01-31 04:18:11 <CodeShark> jgarzik: I'm getting a syntax error in the configure file itself:
441 2013-01-31 04:18:13 <CodeShark> checking for exit in -lboost_program_options-mt... yes
442 2013-01-31 04:18:20 <CodeShark> ./configure: line 5427: syntax error near unexpected token `;;'
443 2013-01-31 04:18:38 <CodeShark> I don't think it should have that -mt
444 2013-01-31 04:19:09 <jgarzik> CodeShark: the -mt suffix is definitely used on some platforms.  maybe not yours.
445 2013-01-31 04:19:29 <CodeShark> yeah, it's used on my OS X
446 2013-01-31 04:19:45 <CodeShark> but not on ubuntu precise, boost v 1.48
447 2013-01-31 04:20:44 <CodeShark> the more interesting question is why autoreconf is botching up that case statement
448 2013-01-31 04:30:25 <CodeShark> jgarzik: you can see my Makefile.am and configure.ac here https://github.com/CodeShark/bitcoin/tree/autotools/src
449 2013-01-31 04:32:19 <jgarzik> CodeShark: yes
450 2013-01-31 05:08:51 <jgarzik> ASIC is now p2pool'ing
451 2013-01-31 05:09:31 <petertodd> nice!
452 2013-01-31 05:10:04 <SomeoneWeird> can i have it
453 2013-01-31 05:10:05 <SomeoneWeird> >.>
454 2013-01-31 05:11:47 <petertodd> I still think you should mine solo until you get a block, and set the coinbase to "jgarzik pw0nz"
455 2013-01-31 05:12:02 <petertodd> should take what, 5 days average?
456 2013-01-31 05:13:09 <SomeoneWeird> roughly
457 2013-01-31 05:13:44 <jgarzik> p2pool epic fail
458 2013-01-31 05:13:44 <Luke-Jr> petertodd: or he can just use a GBT pool :P
459 2013-01-31 05:13:53 <Luke-Jr> well, except that it's cgminer :/
460 2013-01-31 05:14:08 <petertodd> jgarzik: ?
461 2013-01-31 05:14:24 <Luke-Jr> bfgminer you can just do --coinbase 'my message' ;)
462 2013-01-31 05:14:54 <BCB> ;;ident blitz-
463 2013-01-31 05:14:55 <gribble> Nick 'blitz-', with hostmask 'blitz-!blitz@is.super.duper.very.extremely.c00l.info', is not identified.
464 2013-01-31 05:15:11 <petertodd> Luke-Jr: ha, nice, so bfgminer has built-in solo mode?
465 2013-01-31 05:15:32 <Luke-Jr> petertodd: yes, but --coinbase works with any GBT pool too :p
466 2013-01-31 05:15:55 <Luke-Jr> for solo, you need --coinbase-addr <address> too
467 2013-01-31 05:16:17 <petertodd> Luke-Jr: ah, cool, I take it the GBT pool can then reject your shares if you would make the coinbase too large?
468 2013-01-31 05:16:41 <Luke-Jr> petertodd: well, you have to stay within the coinbase size limits of course
469 2013-01-31 05:16:52 <Luke-Jr> BFGMiner will truncate if necessary
470 2013-01-31 05:17:21 <Luke-Jr> 
ACTION ponders if BitMinter would accept a coinbase-too-large share <.<

471 2013-01-31 05:17:39 <petertodd> Luke-Jr: Easy thing to screw up...
472 2013-01-31 07:38:30 <porquilho> SomeoneWeird
473 2013-01-31 07:38:34 <porquilho> can you remove ban ?
474 2013-01-31 07:38:42 <porquilho> from bitcoin and bitcoin-otc
475 2013-01-31 07:38:52 <porquilho> becase i am not 'share'
476 2013-01-31 07:39:26 <porquilho> its going to break now the 20$ barrier
477 2013-01-31 07:39:47 <BTCOxygen> <@jgarzik> p2pool epic fail
478 2013-01-31 07:39:50 <BTCOxygen> why?
479 2013-01-31 07:44:14 <BTCOxygen> 
ACTION think jgarzik is busy playing with his ASIC

480 2013-01-31 07:44:24 <Luke-Jr> he went to bed
481 2013-01-31 07:44:27 <porquilho> the guy was here on irc ?
482 2013-01-31 07:44:37 <porquilho> the guy who get the isac ?
483 2013-01-31 07:44:50 <mappum> whats the update on that, does it actually perform?
484 2013-01-31 07:44:54 <porquilho> yes
485 2013-01-31 07:45:01 <porquilho> he is going to make $220 per day
486 2013-01-31 07:45:10 <porquilho> he is already making $220 per day
487 2013-01-31 07:45:14 <mappum> whoa
488 2013-01-31 07:45:20 <mappum> not on p2pool?
489 2013-01-31 07:46:41 <porquilho> http://bitcoinmagazine.com/working-avalon-asic-confirmed/
490 2013-01-31 07:47:16 <porquilho> its $240 per day
491 2013-01-31 07:49:55 <mappum> Is there really going to be an "asic revolution"? It seems like there is no supply
492 2013-01-31 07:50:28 <porquilho> ;;asks 20
493 2013-01-31 07:50:30 <gribble> There are currently 3842.9437 bitcoins offered at or under 20.0 USD, worth 76858.8651021 USD in total.
494 2013-01-31 07:51:03 <porquilho> only 3,8k BTC
495 2013-01-31 07:51:27 <porquilho> ;;ticker
496 2013-01-31 07:51:28 <gribble> BTCUSD ticker | Best bid: 19.99990, Best ask: 19.99999, Bid-ask spread: 0.00009, Last trade: 19.99990, 24 hour volume: 48530.01689193, 24 hour low: 19.15000, 24 hour high: 19.99999, 24 hour vwap: 19.65852
497 2013-01-31 07:51:32 <mappum> ;;bids 20
498 2013-01-31 07:51:33 <porquilho> ;;asks 20
499 2013-01-31 07:51:34 <gribble> There are currently 0 bitcoins demanded at or over 20.0 USD, worth 0.0 USD in total.
500 2013-01-31 07:51:36 <gribble> There are currently 3842.9437 bitcoins offered at or under 20.0 USD, worth 76858.8651021 USD in total.
501 2013-01-31 07:51:43 <mappum> ;;bids 19.9
502 2013-01-31 07:51:46 <gribble> There are currently 338.1959 bitcoins demanded at or over 19.9 USD, worth 6744.91064095 USD in total.
503 2013-01-31 07:51:53 <mappum> so close
504 2013-01-31 07:52:28 <mappum> ;;bids 19.95
505 2013-01-31 07:52:31 <gribble> There are currently 162.37553 bitcoins demanded at or over 19.95 USD, worth 3244.49479111 USD in total.
506 2013-01-31 07:52:34 <porquilho> ;;ticker
507 2013-01-31 07:52:36 <gribble> BTCUSD ticker | Best bid: 19.99990, Best ask: 20.00000, Bid-ask spread: 0.00010, Last trade: 20.00000, 24 hour volume: 49712.92651993, 24 hour low: 19.15000, 24 hour high: 20.00000, 24 hour vwap: 19.66688
508 2013-01-31 07:52:44 <porquilho> we are on 20$
509 2013-01-31 07:52:52 <mappum> i want to be the first bid at 20 :P
510 2013-01-31 07:52:52 <porquilho> its gone
511 2013-01-31 07:52:54 <porquilho> LOL
512 2013-01-31 07:52:56 <porquilho> mappum ahha
513 2013-01-31 07:52:58 <da2ce7> trade MtGox:          0.50   BTC @   20.00       MTGUSD
514 2013-01-31 07:53:18 <porquilho> NOW LETS SELL!
515 2013-01-31 07:53:19 <porquilho> :p
516 2013-01-31 07:53:21 <porquilho> im kidding
517 2013-01-31 07:54:21 <mappum> fuckin campbx decided to wait a few days to add my dwolla transfer to my account
518 2013-01-31 07:54:56 <mappum> i could have bought at 17
519 2013-01-31 07:57:39 <jeremias> seems like you have to develop better working exchange yourself
520 2013-01-31 07:58:54 <porquilho> mappum well maybe you still can
521 2013-01-31 07:59:02 <porquilho> if it goes lower
522 2013-01-31 07:59:19 <porquilho> it went to 20.4$
523 2013-01-31 08:01:37 <porquilho> SomeoneWeird
524 2013-01-31 08:01:38 <porquilho> SomeoneWeird
525 2013-01-31 08:01:42 <porquilho> SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird
526 2013-01-31 08:01:44 <porquilho> SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird SomeoneWeird
527 2013-01-31 08:01:54 <porquilho> 
ACTION slaps SomeoneWeird around a bit with a large trout

528 2013-01-31 08:01:56 <porquilho> 
ACTION slaps SomeoneWeird around a bit with a large trout

529 2013-01-31 08:01:57 <porquilho> 
ACTION slaps SomeoneWeird around a bit with a large trout

530 2013-01-31 08:02:33 <porquilho> remove ban
531 2013-01-31 08:02:40 <porquilho> why is he not listening to me
532 2013-01-31 08:02:47 <porquilho> he put on ignore
533 2013-01-31 08:02:58 <porquilho> i have to go join with a proxy
534 2013-01-31 08:03:01 <porquilho> to talk to him :\\
535 2013-01-31 08:03:19 <porquilho> remove ban Someguy123
536 2013-01-31 08:03:20 <porquilho> SomeoneWeird
537 2013-01-31 08:04:02 <Scrat> shut the fuck up
538 2013-01-31 08:04:04 <Scrat> mongoloid
539 2013-01-31 08:04:28 <SomeoneWeird> thankyou.
540 2013-01-31 08:05:08 <Ferroh> lol
541 2013-01-31 08:43:51 <Marina> gmaxwell
542 2013-01-31 08:43:54 <Marina> are you there ?
543 2013-01-31 08:45:43 <kinlo> how many blocks do the undo files go back on 0.8?
544 2013-01-31 08:46:16 <Marina> this fucking idiot SomeoneWeird is banning me becase he thinks i am some other people
545 2013-01-31 08:46:35 <Marina> you must be fucking retarded in head SomeoneWeird
546 2013-01-31 08:46:40 <Marina> you must be
547 2013-01-31 09:01:39 <gmaxwell> kinlo: all of them.
548 2013-01-31 09:02:46 <Marina> gmaxwell
549 2013-01-31 09:03:00 <Marina> someoneweird thinks i am someone else and is banning me
550 2013-01-31 09:03:06 <Marina> on bitcoin-otc
551 2013-01-31 09:03:08 <Marina> and bitcoin
552 2013-01-31 09:03:12 <gmaxwell> Marina: who are you and where did you get the idea that _I_ am a person to appeal bans to, or that #bitcoin-dev is a place to do it?
553 2013-01-31 09:03:45 <Marina> you have op
554 2013-01-31 09:03:56 <Marina> i tried to talk to him, but he puts me on ignore
555 2013-01-31 09:04:08 <Marina> i am not 'share' or 'sure' or whatever he thinks i am
556 2013-01-31 09:04:15 <Marina> gmaxwell i am porquilho
557 2013-01-31 09:04:29 <Marina> can you say to him to remove ban
558 2013-01-31 09:04:32 <Marina> maybe he listen to you
559 2013-01-31 09:04:52 <gmaxwell> whatever your issues are, #bitcoin-dev is not the place for them. I'll tell him you are complaining.
560 2013-01-31 09:05:02 <Marina> yes thank you
561 2013-01-31 09:05:18 <Marina> i know its no
562 2013-01-31 09:05:19 <Marina> not
563 2013-01-31 09:05:46 <Marina> he thinks i am someone else, and keeps banning me, like an idiot
564 2013-01-31 09:05:49 <Marina> makes me angry
565 2013-01-31 09:06:30 <Marina> and then puts me on ignore, like he is right and absolute sure that i am someone else. fuckign ignorant
566 2013-01-31 09:06:43 <CodeShark> can you please take the psychotherapy elsewhere?
567 2013-01-31 09:07:04 <Marina> its all SomeoneWeird fault
568 2013-01-31 09:07:19 <Marina> he creates this situation
569 2013-01-31 09:09:28 <kinlo> gmaxwell: hmmz, am I correct to state that in the level db there is a copy of each active transaction?  so no second seek must happen when the transaction must be validated?
570 2013-01-31 09:10:41 <gmaxwell> kinlo: correct.
571 2013-01-31 09:10:56 <gmaxwell> well lets be more specific.
572 2013-01-31 09:11:19 <gmaxwell> The coins database (there are several leveldb databases) has all that is needed to validate a new block.
573 2013-01-31 09:11:47 <gmaxwell> It does not have the data required to reorgnize (by itself at least), or serve older blocks or the transactions inside them.
574 2013-01-31 09:12:15 <kinlo> just trying to understand here: if the undo files must go back to block 1, shouldn't the undo tables be containing all data from the transaction/blocks, hence should the undo database not be bigger then the blockchain itself?
575 2013-01-31 09:12:45 <gmaxwell> the undo data is much smaller.
576 2013-01-31 09:13:23 <gmaxwell> kinlo: the undo data just contains the information required to undo the application of a new block to the coins database.
577 2013-01-31 09:14:05 <gmaxwell> (basically it contains the utxos which were spent by that block, the utxos of a block are fairly small??? most of the bulk of txdata is in scriptsigs, not scriptpubkeys)
578 2013-01-31 09:15:21 <gmaxwell> kinlo: are you enlightened now?
579 2013-01-31 09:15:47 <kinlo> for now, I'll take the information to the next round of testing, I guess more questions will pop up then :)
580 2013-01-31 09:20:27 <gladoscc> How can I get the balance of any address?
581 2013-01-31 09:21:24 <da2ce7> gladoscc: blockchain.info
582 2013-01-31 09:21:40 <gladoscc> without blockchain.info :P
583 2013-01-31 09:23:58 <erska> gladoscc: https://bitcointalk.org/index.php?topic=88584.0
584 2013-01-31 09:24:40 <erska> or Abe
585 2013-01-31 12:18:40 <davout> ohai
586 2013-01-31 12:18:58 <davout> does anyone know of a quick way to launch bitcoind without actually connecting to the network ?
587 2013-01-31 12:19:20 <davout> i want to import a massive load of private keys, having zero blocks in the DB would speed that up A LOT
588 2013-01-31 12:20:35 <sipa> davout: -connect=0.0.0.0 ?
589 2013-01-31 12:27:52 <HM> so Jeffs ASIC box is based around a TP-LINK  TL-WR703N
590 2013-01-31 12:28:23 <HM> a 3G travel router ^_^
591 2013-01-31 12:40:50 <davout> is it just me or is the forum extremely slow?
592 2013-01-31 12:45:31 <Luke-Jr> davout: not just you. maybe it's for the better.
593 2013-01-31 12:46:16 <davout> word
594 2013-01-31 12:47:51 <HM> economics is weird
595 2013-01-31 12:48:51 <HM> lots of people saying ASICs will push bitcoin prices up....but shouldn't ASICs increase supply as people recoup their investment?
596 2013-01-31 12:48:58 <HM> whoop wrong channel for this sorry
597 2013-01-31 12:49:10 <HM> back to lazy coding
598 2013-01-31 12:50:06 <Luke-Jr> HM: no reason to think ASICs will push bitcoin prices up, correct
599 2013-01-31 12:50:23 <Luke-Jr> good news: I have master running on Eligius's server successfully; now I just need to patch Eloipool to use it
600 2013-01-31 12:50:34 <sipa> it's enough that some people think it will push the prices, in order for it to actually push prices up
601 2013-01-31 12:52:55 <kinlo> Luke-Jr: you need to patch eloi?   did 0.8 change it's api's again?
602 2013-01-31 12:53:51 <Luke-Jr> kinlo: no, I need to patch Eloi to use both versions at the same time
603 2013-01-31 12:53:59 <kinlo> ah right
604 2013-01-31 12:53:59 <Luke-Jr> sipa: true
605 2013-01-31 12:54:16 <davout> is there any way to import a key but prevent the client to do a rescan ?
606 2013-01-31 12:54:44 <Scrat> davout: I've said a million times that they should run xcache+nginx+varnish (these are all drop in replacements)
607 2013-01-31 12:54:45 <davout> could be very useful for me if I want to import a large amount of keys and trigger a rescan only after
608 2013-01-31 12:54:45 <HM> from a dev perspective i wouldn't mind owning one of those ASICs if it could be repurposed for generic hashing after it had served its useful mining life
609 2013-01-31 12:54:47 <Luke-Jr> davout: latest code has it optional
610 2013-01-31 12:55:02 <davout> i wish i was using latest code
611 2013-01-31 12:55:16 <davout> it's missing the monitortx/monitorblocks AFAIK
612 2013-01-31 12:55:22 <davout> is it ?
613 2013-01-31 12:55:24 <Luke-Jr> davout: it can't be a very complicated patch to backport
614 2013-01-31 12:55:33 <Luke-Jr> davout: 0.6+ has -blocknotify
615 2013-01-31 12:55:39 <davout> what do you mean by backport ?
616 2013-01-31 12:55:58 <davout> i use gavin's patch that i ported to 0.6
617 2013-01-31 12:55:59 <Luke-Jr> davout: look at what the commit to master changed to make it optional, then make the same changes to your code
618 2013-01-31 12:56:13 <Luke-Jr> davout: monitor* are ugly :P
619 2013-01-31 12:56:18 <davout> why ?
620 2013-01-31 12:56:30 <davout> having http callbacks rocks in quite a few use cases
621 2013-01-31 12:56:33 <Luke-Jr> davout: they require you to run a webserver in your other process
622 2013-01-31 12:56:46 <davout> what would the other options be ?
623 2013-01-31 12:56:46 <Luke-Jr> davout: -blocknotify can run any command, like curl ;)
624 2013-01-31 12:56:52 <davout> oh i see
625 2013-01-31 12:56:56 <davout> that works too
626 2013-01-31 12:57:00 <davout> even better i guess
627 2013-01-31 12:57:10 <davout> so latest code has blocknotify
628 2013-01-31 12:57:12 <davout> right ?
629 2013-01-31 12:57:15 <Luke-Jr> yes
630 2013-01-31 12:57:18 <davout> what about txnotify ?
631 2013-01-31 12:57:26 <Luke-Jr> there's a -walletnotify pullreq, but it isn't merged yet
632 2013-01-31 12:57:33 <Luke-Jr> probably won't be until after 0.8 I guess
633 2013-01-31 12:57:35 <davout> what would it do exactly ?
634 2013-01-31 12:57:44 <Luke-Jr> runs a command when something happens involving your wallet
635 2013-01-31 12:57:45 <davout> notify *all* txes ? or only your own ?
636 2013-01-31 12:57:52 <davout> ok, only your own then
637 2013-01-31 12:58:00 <Luke-Jr> if you want to watch all txs, just connect to the p2p port..
638 2013-01-31 12:58:05 <davout> guess it would work like blocknotify
639 2013-01-31 12:58:12 <davout> ya, i'm not interested in all txes
640 2013-01-31 12:58:14 <kjj> transactions that hit your wallet
641 2013-01-31 12:58:20 <davout> that's awesome
642 2013-01-31 12:58:21 <kinlo> davout: there are so many tx's, you could just poll every few seconds, new ones arrive at any time
643 2013-01-31 12:58:33 <davout> polling is ugly
644 2013-01-31 12:58:46 <kinlo> true, but it depends on what you are interested
645 2013-01-31 12:59:04 <davout> i'm interested in making instawallet and bitcoin-central awesome
646 2013-01-31 12:59:12 <davout> or more awesome than what they already are :)
647 2013-01-31 12:59:23 <davout> part of this requires real-time tx notification
648 2013-01-31 12:59:35 <davout> that gets handled by the web backend
649 2013-01-31 12:59:47 <davout> and subsequently notified through websocket to the client
650 2013-01-31 12:59:49 <Scrat> you can do selective polling
651 2013-01-31 12:59:55 <davout> forget about polling
652 2013-01-31 13:00:05 <Scrat> ie. poll logged in user's addresses every minute, otherwise every hour
653 2013-01-31 13:00:18 <Scrat> yeah it is ugly
654 2013-01-31 13:00:24 <HM> you can hide polling server side
655 2013-01-31 13:00:36 <davout> every minute doesn't cut it, real notification also puts less strain on bitcoind
656 2013-01-31 13:00:40 <sipa> meh, shouldn't be necessary
657 2013-01-31 13:00:43 <Luke-Jr> davout: I'd suggest taking master and merging -walletnotify; then you can easily merge 0.8.0 final, and the 0.8.x branch to get bugfixes
658 2013-01-31 13:00:59 <Scrat> if you use blocknotify then you need a way to get older blocks in case your app server is down
659 2013-01-31 13:01:01 <davout> yea well, the problem with that is maintaining my own fork
660 2013-01-31 13:01:13 <davout> Scrat: yea, so what ?
661 2013-01-31 13:01:22 <Luke-Jr> davout: that's what the stable branches are for
662 2013-01-31 13:01:31 <davout> polling can be a fallback to repair the DB when stuff goes wrong
663 2013-01-31 13:01:41 <davout> not the default way of doing things
664 2013-01-31 13:01:56 <davout> brb
665 2013-01-31 13:02:01 <Luke-Jr> davout: and -walletnotify is almost certain to be merged for 0.9 imo
666 2013-01-31 13:02:59 <kjj> as-is?  or do I still need to rewrite it?
667 2013-01-31 13:03:35 <sipa> something needs to be done about dos potential, i think
668 2013-01-31 13:04:11 <kjj> I think that potential is much more potential than actual, but that has been the longstanding objection
669 2013-01-31 13:05:20 <Scrat> Luke-Jr: walletnotify should be modified or at least there should be a version that is always atomic and only triggers on incoming transactions
670 2013-01-31 13:05:36 <sipa> always atomic?
671 2013-01-31 13:07:13 <Scrat> sipa: developers shouldn't have to store txid's just to verify that bitcoind isn't sending them twice
672 2013-01-31 13:07:34 <Scrat> (for a given amount of confirmations)
673 2013-01-31 13:08:33 <kjj> walletnotify triggers when the wallet changes.  it tells you which one you need to look at
674 2013-01-31 13:09:05 <Scrat> all I'm saying is: incomingnotify
675 2013-01-31 13:09:08 <Scrat> something like that :p
676 2013-01-31 13:09:18 <HM> related to DoS, has anyone heard of anyone blocking the bitcoin protocol?
677 2013-01-31 13:09:29 <kjj> it isn't a payment notification, for two reasons.  the first is that it was super easy to do it the way it is, and the second is that things like payment notification are murky in bitcoin
678 2013-01-31 13:10:11 <Scrat> but does it have to be murky? is it the consensus of the devs that stuff like that should be 3rd party?
679 2013-01-31 13:10:32 <kjj> no, it isn't up to the devs.  it is murky because bitcoin doesn't act like anything you are used to
680 2013-01-31 13:11:02 <kjj> if you want an incomingnotify, you also need a incomingUNnotify for when that payment is pulled back by a reorg
681 2013-01-31 13:11:09 <Scrat> sure
682 2013-01-31 13:12:07 <kjj> getting a notification that just tells you to look at the transaction seems like the right thing to do.  you are going to have to make sense of it anyway, and this way we aren't stuffing a whole detailed callback system into bitcoin
683 2013-01-31 13:13:36 <davout> kjj: +1
684 2013-01-31 13:14:36 <sipa> kjj: regarding DoS potential, something that doesn't sound too hard imho, is having a set of txids to notify for, run the notifier in a separate thread, and never do more than one notify at one time
685 2013-01-31 13:14:59 <sipa> and loop over the set to notify for, and go to sleep when it is empty
686 2013-01-31 13:16:13 <kjj> hmm.  I'll have to learn threads for that, and boost queues
687 2013-01-31 13:16:38 <kjj> sounds like I have some time though.  I'll look into it
688 2013-01-31 13:17:18 <Scrat> kjj: I didn't say that a notify should just return an address/value pair
689 2013-01-31 13:18:11 <Scrat> but it definitely needs to be smarter than the current walletnotify
690 2013-01-31 13:19:27 <kjj> I agree that smartness is needed, but I disagree on which part should be smart.  I think it should be up to the consumer of the information to make sense of it
691 2013-01-31 13:19:48 <sipa> i think there are various ways of abstraction to look at a wallet
692 2013-01-31 13:19:55 <sipa> and we do mix them in confusing ways
693 2013-01-31 13:23:53 <kjj> some of that is unavoidable.  some system has to deal with all of the messy stuff at the bottom of the stack.
694 2013-01-31 13:24:14 <sipa> yes, i'm not saying we should only expose one
695 2013-01-31 13:24:23 <sipa> but more that they should be clearly separated
696 2013-01-31 13:24:31 <kjj> people want nice accounting layers, and they are right to want them.  those just need to be added on top, not wedged into the "physical" layer
697 2013-01-31 13:30:22 <Scrat> kjj: someone here said that a modular design is inevitable
698 2013-01-31 13:30:31 <Scrat> ie. core, UI, notifier, whathaveyou
699 2013-01-31 13:30:38 <kjj> I've said that myself a bunch of times
700 2013-01-31 13:31:17 <kjj> at the very least, I think the stock client should be broken into two parts.
701 2013-01-31 14:37:51 <rdponticelli> Is there a reason for relaying (or spamming) repeatedly all wallet transactions during reindex?
702 2013-01-31 14:41:33 <tomboy64> how secure is bitcoin against forgery if someone was capable of intercepting all traffic happening between one bitcoind/-qt client and the rest of the bitcoin-network?
703 2013-01-31 14:42:00 <HM> tomboy64: completely
704 2013-01-31 14:42:10 <HM> unless they're paying to IP, i think.
705 2013-01-31 14:42:21 <sipa> that's been disabled for a while
706 2013-01-31 14:42:26 <tomboy64> the 6-confirmations-thingy?
707 2013-01-31 14:42:33 <HM> sipa: cool
708 2013-01-31 14:42:46 <tomboy64> HM: ip?
709 2013-01-31 14:42:58 <sipa> tomboy64: a passive attacker can only hurt your privacy
710 2013-01-31 14:43:17 <HM> unless they can spy on your rpc session and you're not using TLS/SSL :P
711 2013-01-31 14:43:24 <sipa> tomboy64: an active attacker could try to get you on a forked chain, but he still needs the hash rate to produce blocks
712 2013-01-31 14:43:35 <sipa> they can't make you accept invalid transactions, for example
713 2013-01-31 14:44:38 <tomboy64> what will the blocks contain when new bitcoin blocks get scarcer?
714 2013-01-31 14:44:41 <gavinandresen> sipa: I'm mostly in meetings today??? but what's the status of the Validation stuff?
715 2013-01-31 14:45:02 <sipa> gavinandresen: luke-jr's pullreq looks good, but haven't tested yet
716 2013-01-31 14:45:03 <HM> they could also steal any SIGHASH_NONE funds?
717 2013-01-31 14:45:16 <gavinandresen> sipa: Ok, I'll try to review between meetings
718 2013-01-31 14:45:18 <tomboy64> i can understand new chunks are easily verifiable, but how will verification occur when those chunks don't get produced anymore?
719 2013-01-31 14:45:30 <sipa> tomboy64: no blocks == no confirmations
720 2013-01-31 14:45:40 <sipa> tomboy64: and producing blocks (valid or invalid) is hard
721 2013-01-31 14:46:11 <tomboy64> sipa: do you understand what i mean by "new chunks"?
722 2013-01-31 14:46:16 <sipa> tomboy64: no
723 2013-01-31 14:46:21 <sipa> tomboy64: i assume you mean blocks
724 2013-01-31 14:46:25 <tomboy64> the stuff with the 25BTC reward
725 2013-01-31 14:46:31 <sipa> those are called blocks
726 2013-01-31 14:46:47 <Luke-Jr> gavinandresen: I'm working on setting up latest master code on Eligius for side-by-side comparison with 0.6.0 with a variety of real-world possible-blocks, FWIW
727 2013-01-31 14:46:54 <tomboy64> as i understood it it's limitied at 21million - and production rate will decline logarithmically
728 2013-01-31 14:47:03 <sipa> tomboy64: no, exponentially
729 2013-01-31 14:47:35 <tomboy64> so what happens when that production declines to - let's say, 1 block a day, worldwide?
730 2013-01-31 14:47:41 <sipa> tomboy64: it can't
731 2013-01-31 14:47:48 <tomboy64> then new transactions can't get worked in anymore.
732 2013-01-31 14:47:50 <Luke-Jr> tomboy64: it declines by block value, not block count
733 2013-01-31 14:48:00 <sipa> tomboy64: difficulty will drop correspondingly to keep the average block rate at one per 10 minutes
734 2013-01-31 14:48:01 <Luke-Jr> tomboy64: so in 4 years, blocks will be worth 12.50 BTC each
735 2013-01-31 14:48:08 <tomboy64> 
ACTION scratches his head

736 2013-01-31 14:48:12 <Luke-Jr> tomboy64: but the same number of blocks per day
737 2013-01-31 14:48:15 <tomboy64> ah!
738 2013-01-31 14:48:20 <sipa> gavinandresen: one unintended (but perhaps wanted) effect of CValidationResult is that if an invalid DoS-triggering block in a side chain is produced, and someone builds upon it to make you reorg to it, he will get punished for the block he built upon
739 2013-01-31 14:48:46 <tomboy64> i see. i got something wrong then.
740 2013-01-31 14:49:09 <tomboy64> thanks for that clarification :)
741 2013-01-31 14:59:20 <rdponticelli> sipa: ResendWalletTransactions is needed during a reindex?
742 2013-01-31 14:59:46 <sipa> rdponticelli: hmm, not really i guess
743 2013-01-31 15:00:11 <BlueMatt> sipa: hmmm...Im getting different block acceptance on windows vs linux
744 2013-01-31 15:00:24 <rdponticelli> Yeah, I thought so
745 2013-01-31 15:00:32 <sipa> BlueMatt: woah?
746 2013-01-31 15:00:38 <rdponticelli> It's spamming my old transactions
747 2013-01-31 15:01:01 <BlueMatt> sipa: its a pretty strange block, but I dont really have time to debug it right now
748 2013-01-31 15:01:08 <BlueMatt> sipa: http://jenkins.bluematt.me/pull-tester/files/FullBlockTestGenerator.java find b64
749 2013-01-31 15:09:29 <BlueMatt> maybe I can test later
750 2013-01-31 15:09:29 <BlueMatt> that test is only failing when running in wine, needs testing to see if it also happens when compiled with w64
751 2013-01-31 15:14:39 <Luke-Jr> sipa: I seem to have corrupted a node :o
752 2013-01-31 15:14:53 <Luke-Jr> bitcoind: main.cpp:1605: bool CBlock::ConnectBlock(CValidationState&, CBlockIndex*, CCoinsViewCache&, bool): Assertion `pindex->pprev == view.GetBestBlock()' failed.
753 2013-01-31 15:15:51 <sipa> Luke-Jr: what code, what did you do?
754 2013-01-31 15:17:05 <Luke-Jr> sipa: some kind of race - I was flooding it with getblocktemplates and immediately returning a (valid*) submitblock for each one
755 2013-01-31 15:17:54 <sipa> Luke-Jr: git head, or some modifications?
756 2013-01-31 15:18:01 <Luke-Jr> sipa: in one case, the submitblock was an orphan, and it crashed corrupt
757 2013-01-31 15:18:34 <Luke-Jr> sipa: modifications; I doubt it can be reproduced easily without proposals
758 2013-01-31 15:18:58 <Luke-Jr> since then you'd need to find a proof-of-work constantly
759 2013-01-31 15:19:28 <sipa> Luke-Jr: do you know where ConnectBlock was called?
760 2013-01-31 15:20:08 <sipa> that message means the view passed in does not correspond to the parent block of the one being connected
761 2013-01-31 15:20:52 <Luke-Jr> sipa: afraid not in the first case - I get the same error whenever I try to start that node now, though
762 2013-01-31 15:20:56 <kjj> hey luke, I have logs now of a box not failing back to pool 0
763 2013-01-31 15:21:36 <sipa> Luke-Jr: ewww :S
764 2013-01-31 15:21:51 <Luke-Jr> sipa: likely -reindex will fix?
765 2013-01-31 15:21:55 <sipa> yes
766 2013-01-31 15:22:05 <Luke-Jr> ok, I'll see if I can reproduce it easily
767 2013-01-31 15:22:10 <sipa> even a delete of chaindata/ will fix
768 2013-01-31 15:22:26 <sipa> chainstate
769 2013-01-31 15:22:27 <Luke-Jr> anything I should backup?
770 2013-01-31 15:22:31 <TD> sipa: iirc some google base libraries are open sourced, things like the crash handlers and stuff
771 2013-01-31 15:22:48 <TD> sipa: it might be worth trying to merge them in at some point so we get nice google3 style crash logs
772 2013-01-31 15:23:14 <sipa> Luke-Jr: yes, chaindata/
773 2013-01-31 15:23:31 <reeep> i have an idea i want to bounce off some developers
774 2013-01-31 15:23:32 <sipa> getting that error at startup should be outright impossible, even with a corrupted database
775 2013-01-31 15:23:52 <reeep> image explanation: http://i.imgur.com/wOiMn6M.png
776 2013-01-31 15:23:58 <rdponticelli> kjj: Your problem was with p2pool and bfgminer, right?
777 2013-01-31 15:24:02 <reeep> crappy image nonetheless
778 2013-01-31 15:24:21 <sipa> reeep: talk to amiller_
779 2013-01-31 15:24:22 <HM> reeep: that's elftor level art right there ;)
780 2013-01-31 15:24:28 <kjj> rdponticelli: yes
781 2013-01-31 15:24:28 <reeep> right now bitcoin miners build on top of the first block they see, which means if blocks occurred every 10 seconds, we'd have a ton of orphaned blocks
782 2013-01-31 15:24:30 <rdponticelli> kjj: I've seen the same behavior with p2pool and cgminer
783 2013-01-31 15:24:47 <kjj> ooh.  I think I've found a candidate for the problem
784 2013-01-31 15:24:51 <reeep> but if we made honest miners choose the block with the most difficulty, everybody would be competing to orphan the top of the chain