1 2011-10-26 03:17:39 <osmosis> http://bitcoin.org/   says   Full announcement (including signatures)  , but when I click the link I dont see any signatures.
  2 2011-10-26 03:26:35 <gmaxwell> osmosis: the mailing list archive appears to be hiding them.
  3 2011-10-26 03:27:00 <gmaxwell> osmosis: if you go get the identical message posted to the forums you'll see the signatures on the bottom and you can attempt to validate them.
  4 2011-10-26 03:27:15 <gmaxwell> Though they're signed by gavin's key and I'm unable to find a link to his key from mine.
  5 2011-10-26 03:27:18 <Diablo-D3> "Seriously, five ancient sages of bitchdom all gathered together one day on the peaks of Mount Bitch to proclaim your birth, and a hundred years later, when all the bitch stars had aligned, you were born and made everybody's life around you a living hell because you are such a bitch!"
  6 2011-10-26 03:27:25 <gmaxwell> (much less a reasonably short one)
  7 2011-10-26 03:29:39 <osmosis> gmaxwell, i see some keys included in the archive in the gitian dir
  8 2011-10-26 03:39:04 <gmaxwell> osmosis: but the guy replacing the binary as you download it is replacing those too!
  9 2011-10-26 03:39:31 <osmosis> gmaxwell, right
 10 2011-10-26 03:40:00 <Diablo-D3> I should make that my forum sig
 11 2011-10-26 04:06:55 <CIA-101> poolserverj: shadders * 629f0f222c1d r169 /poolserverj-main/src/main/java/com/shadworld/poolserver/ (6 files in 3 dirs):
 12 2011-10-26 04:59:26 <CIA-101> poolserverj: shadders * 8e167caa2acd r170 /poolserverj-main/src/main/java/com/shadworld/poolserver/ (4 files in 2 dirs): (log message trimmed)
 13 2011-10-26 04:59:27 <CIA-101> poolserverj: longpoll passthru. To address the double longpoll issue.
 14 2011-10-26 04:59:28 <CIA-101> poolserverj: What happens then is that one chain advances to the next block. PSJ checks the
 15 2011-10-26 07:11:20 <shadders> Is there any particular reason the diff target needs to be in coinbase?  Or the 04 prefix... or the length prefix for extraNonce?
 16 2011-10-26 07:30:43 <gribble> The operation succeeded.
 17 2011-10-26 07:30:43 <nanotube> ;;later tell graingert ok ! works now in -dev. don't remember why i changed that for -dev in the first place. heh.
 18 2011-10-26 09:00:47 <jav__> why do exchange APIs, like Tradehill's or the new API version from Mt.Gox, return price information and such as strings? .. is this accepted best practice? what's wrong with the JSON number type?
 19 2011-10-26 09:04:59 <jav__> I see, the JSON specification doesn't specify the exact type of a JSON number... I guess that could be considered problematic
 20 2011-10-26 09:15:51 <CIA-101> poolserverj: shadders * 44d705f11bc2 r171 /poolserverj-main/src/main/java/com/shadworld/poolserver/LongpollHandler.java: added trace message to indicate longpoll passthru has started/stopped
 21 2011-10-26 09:25:29 <edcba> jav__: it's a javascript number afaik
 22 2011-10-26 09:27:10 <edcba> http://en.wikipedia.org/wiki/JSON#Data_types.2C_syntax_and_example
 23 2011-10-26 09:28:08 <edcba> and since it's floating point it's very bad for money handling :)
 24 2011-10-26 09:34:08 <jav__> edcba: If the objection is to floating point, then I'm not sure how a floating point number pretty-printed as a String changes much
 25 2011-10-26 09:41:14 <upb> it changes because the client can choose how to interpret it
 26 2011-10-26 09:59:13 <edcba> jav__: how do you know it's a floating point pretty printed ? :)
 27 2011-10-26 10:10:59 <Eliel> I think we should identify these for bitcoin http://falkvinge.net/2011/03/22/the-swarms-activation-ladder/
 28 2011-10-26 10:16:32 <CIA-101> bitcoinj: hearn@google.com * r254 /trunk/target: Remove javadocs from repo, they are available at javadoc.bitcoinj.org instead.
 29 2011-10-26 12:34:09 <graingert> !later tell BlueMatt it might be better to have
 30 2011-10-26 12:34:10 <gribble> The operation succeeded.
 31 2011-10-26 12:34:14 <graingert> arse
 32 2011-10-26 12:39:07 <graingert> !later tell BlueMatt it might be better to have a dev and release channel on https://launchpad.net/~bitcoin/+archive/bitcoin a little like the chrome PPA
 33 2011-10-26 12:39:07 <gribble> The operation succeeded.
 34 2011-10-26 12:40:02 <imsaguy2> I think !later needs an undo
 35 2011-10-26 14:03:11 <Diablo-D3> gmaxwell: you know, I think I figured out what eligius problem is
 36 2011-10-26 14:03:18 <Diablo-D3> its deleting active works
 37 2011-10-26 14:03:33 <Diablo-D3> instead of doing it 120 seconds after being issued, it does it when the client gets another one
 38 2011-10-26 14:03:48 <Diablo-D3> which is insane when clients frequently hold multiple works
 39 2011-10-26 14:04:21 <Diablo-D3> so, I think I can now conclude luke is a lying motherfucker and will rot in hell according to his own pedophiliac beliefs.
 40 2011-10-26 14:07:00 <nejon> :D
 41 2011-10-26 14:08:40 <UukGoblin> what's this Scrypt and why is it faster on CPU than GPU?
 42 2011-10-26 14:11:47 <necrodearia> https://bitcointalk.org/index.php?topic=41456.msg594335#msg594335
 43 2011-10-26 14:12:21 <gmaxwell> Diablo-D3: No, it's not doing that if it did that I'd have enormous reject rates on my cgminers with 12 threads.
 44 2011-10-26 14:12:23 <jrmithdobbs> UukGoblin: it's a key deriv function that uses several (varying number of) rounds of salsa20/8 and some other fun stuff
 45 2011-10-26 14:12:48 <tcatm> necrodearia: see about page...
 46 2011-10-26 14:13:01 <jrmithdobbs> UukGoblin: it uses a lot of ram and the work isn't parallizable, hence, gpus not very useful
 47 2011-10-26 14:13:17 <UukGoblin> jrmithdobbs, interesting, got a link or something?
 48 2011-10-26 14:13:37 <necrodearia> mm
 49 2011-10-26 14:13:38 <jrmithdobbs> http://www.tarsnap.com/scrypt.html
 50 2011-10-26 14:13:41 <UukGoblin> thanks
 51 2011-10-26 14:13:48 <gmaxwell> Diablo-D3: it certantly does it when clients get a long poll and have to switch work. It would be insane to award credit against old blocks.
 52 2011-10-26 14:13:51 <jrmithdobbs> UukGoblin: basically it's bcrypt for the 21st century
 53 2011-10-26 14:14:21 <gmaxwell> jrmithdobbs: though it's GPU hardness depends on it actually being configured to use enough memory to be hard on a gpu.
 54 2011-10-26 14:14:46 <jrmithdobbs> gmaxwell: in part, salsa20 also isn't optimized on them
 55 2011-10-26 14:15:27 <gmaxwell> jrmithdobbs: yes, but thats just a lack of engineering, not a practical limit AFAIK. You could have said the same thing about SHA256 a year ago. :)
 56 2011-10-26 14:15:42 <imsaguy2> gmaxwell, why are you running 12 threads?
 57 2011-10-26 14:15:47 <jrmithdobbs> gmaxwell: fair enough ;p
 58 2011-10-26 14:16:07 <gmaxwell> imsaguy2: six gpus, two threads each.
 59 2011-10-26 14:16:13 <imsaguy2> ah, gotcha
 60 2011-10-26 14:16:29 <jrmithdobbs> gmaxwell: but there's also no ready-made salsa20 circuits for asics either which there have been for sha2 for a while
 61 2011-10-26 14:16:39 <jrmithdobbs> but ya, still a matter of engineering
 62 2011-10-26 14:16:44 <tcatm> necrodearia: we have no easy way to credit those who help with translations since we are now using transifex. maybe you could figure out how to do that?
 63 2011-10-26 14:17:31 <jrmithdobbs> gmaxwell: i'd have to go read the salsa20 paper again but i think there's some other things that make it unparallizable
 64 2011-10-26 14:17:35 <gmaxwell> jrmithdobbs: indeed, but none of the ones for sha256 are all that exciting for mining.
 65 2011-10-26 14:17:39 <jrmithdobbs> or at least, hard to do
 66 2011-10-26 14:17:45 <gmaxwell> jrmithdobbs: sha256 is also unparallizable.
 67 2011-10-26 14:18:15 <UukGoblin> yeah, it's just the case of needing to compute a lot of them
 68 2011-10-26 14:18:16 <gmaxwell> But for mining (or cracking) you create a pipeline and do 128 at once.
 69 2011-10-26 14:18:30 <gmaxwell> But each instance is completely serial.
 70 2011-10-26 14:18:35 <jrmithdobbs> right
 71 2011-10-26 14:18:46 <jrmithdobbs> and for mining you're only doing two rounds
 72 2011-10-26 14:19:16 <gmaxwell> You can't do that for scrypt because each instance needs a lot of fast memory attached... well, it does if you choose to use scrypt with a lot of memory.
 73 2011-10-26 14:19:41 <jrmithdobbs> even using the defaults in his provided utilities it slows it down to some extant
 74 2011-10-26 14:20:02 <jrmithdobbs> it uses between 256K-20M
 75 2011-10-26 14:20:32 <jrmithdobbs> (depending on availability in the generating system)
 76 2011-10-26 14:20:45 <gmaxwell> (I'm pointing out the memory thing because that claimed-gpu-hostile cryptocurency uses a rather small amount of memory in their scrypt IIRC way less than the scrypt defaults... perhaps it doesn't work well on current GPUs, but I'm skeptical that you couldn't still get nice speedups with e.g. fpgas)
 77 2011-10-26 14:21:09 <jrmithdobbs> there's a claimed gpu-hostile cryptocurrency?
 78 2011-10-26 14:21:35 <tcatm> necrodearia: oh and you can create pull requests for bitcoin.org. I think a page that explains how people can contribute (translations, code, help users, improve website) to the project with a section for credits might be useful.
 79 2011-10-26 14:21:42 <necrodearia> hmm?
 80 2011-10-26 14:22:03 <jrmithdobbs> gmaxwell: hey if it leads to some nice fast salsa20 hw implementations that sounds great to me whether their tech choices re: mem size/etc are valid or not ;p
 81 2011-10-26 14:22:37 <necrodearia> tcatm, thanks for pointing out the about page.  I didn't notice it until after you pointed it out.  clumsy me...
 82 2011-10-26 14:23:15 <necrodearia> although, that is a good idea. ^_^
 83 2011-10-26 14:23:26 <tcatm> necrodearia: discussing "developers  vs. contributors" isn't going to get us anywhere, while pull requests that can easily be reviewed and merged if there is consensus will
 84 2011-10-26 14:23:32 <jrmithdobbs> gmaxwell: what "other" currency are you referring to so I can go see how they're using scrypt exactly ;p
 85 2011-10-26 14:23:48 <epscy> tenebrix?
 86 2011-10-26 14:24:10 <gmaxwell> That.
 87 2011-10-26 14:24:42 <gmaxwell> https://bitcointalk.org/index.php?topic=49888.0  < someone posted on the forum complaining that they're losing inbound connections when their IP changes... another argument for UPNP renewals.
 88 2011-10-26 14:25:08 <jrmithdobbs> gmaxwell: the upnp code doesn't renew?
 89 2011-10-26 14:25:17 <da2ce7> !topci
 90 2011-10-26 14:25:18 <gribble> Error: "topci" is not a valid command.
 91 2011-10-26 14:25:23 <da2ce7> !topic
 92 2011-10-26 14:25:24 <gribble> Main: http://bitcoin.org/ | Wiki: https://en.bitcoin.it/wiki/ | Latest version: 0.4 *OLD VERSIONS HARM THE NETWORK AND YOUR SECURITY* | Bitcoin Development - We're here to help develop the Bitcoin system. All related discussions are welcome. | If you have a question, simply ask and wait for a reply. | Main support/discussion chan #bitcoin | Public channel logs: bit.ly/iPFi3X
 93 2011-10-26 14:25:29 <gmaxwell> jrmithdobbs: no. not ever.
 94 2011-10-26 14:25:49 <jrmithdobbs> interesting
 95 2011-10-26 14:25:55 <jrmithdobbs> never looked at the upnp code
 96 2011-10-26 14:26:08 <tcatm> gmaxwell: we debugged that problem last year and found it is caused by faulty routers; though reducing some timeouts within bitcoin might help
 97 2011-10-26 14:31:14 <jrmithdobbs> gmaxwell: so did art just s/SHA256/scrypt/g basically?
 98 2011-10-26 14:35:10 <jrmithdobbs> gmaxwell: ah, interesting, only in the POW stuff
 99 2011-10-26 15:10:58 <BlueMatt> can I get a few testnet coins: mkzNHRa1eUjWXkJzzo7gvc2jexprFsSbzU ?
100 2011-10-26 15:11:46 <rjk2> BlueMatt: https://testnet.freebitcoins.appspot.com/ will give you 500
101 2011-10-26 15:12:05 <BlueMatt> yea, I just did that 1 sec after I wrote that...
102 2011-10-26 15:12:09 <BlueMatt> damn forgetful me
103 2011-10-26 15:13:00 <BlueMatt> although the faucet lags...
104 2011-10-26 15:14:04 <rjk2> am I right when i say that theoretically you could run your own copy of blockexplorer (say on a local machine) just by having a good copy of the blockchain?
105 2011-10-26 15:14:32 <BlueMatt> yes
106 2011-10-26 15:14:51 <rjk2> are there any guides or any such somewhere?
107 2011-10-26 15:14:53 <BlueMatt> in fact, iirc a few such programs exist
108 2011-10-26 15:15:04 <rjk2> haven't been able to find much
109 2011-10-26 15:15:12 <Eliel> I think we should identify these for bitcoin http://falkvinge.net/2011/03/22/the-swarms-activation-ladder/
110 2011-10-26 15:15:18 <BlueMatt> bitcoin-js has one irrc
111 2011-10-26 15:15:34 <BlueMatt> node-bitcoin-explorer
112 2011-10-26 15:16:03 <Eliel> that is, to analyse what are the steps for becoming involved and what could be done to make them easier
113 2011-10-26 15:16:28 <BlueMatt> arg testnet coins arent coming in, someone want to testnet mine for me?
114 2011-10-26 15:16:43 <BlueMatt> (and maybe send again)
115 2011-10-26 15:17:04 <nathan7> Hi, earthlings!
116 2011-10-26 15:34:08 <diki> Tbh, if bitcoin did not have the possibility of some attacks such as 51% etc...maybe it would have gotten more attention(ok, my true words would be succeeded)
117 2011-10-26 15:34:28 <diki> I mean, once i know that weaknesses exist, it gets harder for me to accept something
118 2011-10-26 15:34:45 <nathan7> That weakness can not be prevented.
119 2011-10-26 15:35:03 <nathan7> But it is near-impossible to abuse.
120 2011-10-26 15:35:05 <diki> Its bad that it can
121 2011-10-26 15:35:12 <diki> but the point here is its possible
122 2011-10-26 15:35:36 <[Tycho]> Everything has weaknesses.
123 2011-10-26 15:35:53 <diki> Sure, but bitcoin is money
124 2011-10-26 15:35:56 <[Tycho]> Paper money can be forged.
125 2011-10-26 15:36:09 <diki> paper money can be forged, but there is no possible 51% attack
126 2011-10-26 15:36:31 <diki> where suddently my paper money is just that...paper..wothless paper
127 2011-10-26 15:36:42 <rjk2> oh really?
128 2011-10-26 15:36:45 <[Tycho]> Actually it's possible.
129 2011-10-26 15:36:50 <rjk2> see zimbabwew for an example
130 2011-10-26 15:36:54 <[Tycho]> Happened many times.
131 2011-10-26 15:36:55 <rjk2> zimbabwe*
132 2011-10-26 15:37:03 <nathan7> diki: The 51% is the bank here.
133 2011-10-26 15:37:15 <nathan7> And it's way over 51%.
134 2011-10-26 15:37:23 <rjk2> 100%
135 2011-10-26 15:37:43 <diki> while its possible paper money to go worthless, i dont like the idea of looking someday at my client and seeing everything invalidated
136 2011-10-26 15:38:23 <[Tycho]> diki, you didn't read the wiki page about attacks.
137 2011-10-26 15:38:47 <[Tycho]> No one can invalidate YOUR money. 51% attack mostly allows the attacker to reverse HIS transactions.
138 2011-10-26 15:38:57 <[Tycho]> He can't control other people's wallets.
139 2011-10-26 15:39:52 <gavinandresen> BlueMatt: I threw a little hashing power at the testnet (70mhash), but will probably take an hour or two to find a block...
140 2011-10-26 15:40:36 <[Tycho]> Oh, I'll need testcoins soon :))
141 2011-10-26 15:41:02 <[Tycho]> But it will be easier to mine myself, I think.
142 2011-10-26 15:42:34 <diki> why do you need testnet coins?
143 2011-10-26 15:42:43 <diki> you've got a pool to get the real stuff from :D
144 2011-10-26 15:43:20 <[Tycho]> Testcoins are needed for experiments and testing new funny stuff.
145 2011-10-26 15:43:28 <rjk2> because he doesn't want to test with real money
146 2011-10-26 15:44:08 <[Tycho]> Otherwise I can either lose some money or make lots of people go "what the hell it was in the blockchain right now ?!"
147 2011-10-26 15:44:15 <rjk2> makes me wonder though, how hard could it be to temporarily merged-mine on the testnet chain whenever you need a few confirmations
148 2011-10-26 15:44:28 <diki> rjk2:bad
149 2011-10-26 15:44:40 <rjk2> wat
150 2011-10-26 15:44:48 <rjk2> just for like 10 minutes
151 2011-10-26 15:46:55 <BlueMatt> gavinandresen: thanks, though for some reason I still dont see the faucet tx...
152 2011-10-26 15:47:14 <BlueMatt> (though I was synced and on network at send time...)
153 2011-10-26 15:47:26 <BlueMatt> (or does bitcoin-qt now not show unconfirmed txes?)
154 2011-10-26 15:48:16 <gavinandresen> It should show unconfirmed transactions if you're caught up with the block-chain
155 2011-10-26 15:48:26 <BlueMatt> odd...I am
156 2011-10-26 15:48:51 <BlueMatt> ah, bitcoin-qt doesnt think its caught up because the last block is hours ago
157 2011-10-26 15:49:04 <gavinandresen> If somebody was messing with the testnet block chain it could've broken the faucet's wallet... I'll check
158 2011-10-26 15:49:06 <BlueMatt> (still showing the syncing icon)
159 2011-10-26 15:51:58 <gavinandresen> The testnet faucet has a bunch of 0-confirmation transactions: probably your transaction is part of the change for another 0-conf txn...
160 2011-10-26 15:52:17 <gavinandresen> ... which your bitcoind didn't see, so it is an orphan (and orphans aren't shown)
161 2011-10-26 15:52:29 <Eliel> rjk2: if testnet mode was changed to running with the namecoin modifications, it ought to be possible to mergemine it just the same as namecoin... at least that's how I see it.
162 2011-10-26 15:53:36 <BlueMatt> gavinandresen: m, makes sense
163 2011-10-26 15:53:39 <rjk2> Eliel: yeah that's what I was thinking
164 2011-10-26 15:53:53 <rjk2> so a pool oper could have a tickbox when he wanted a few confirms
165 2011-10-26 15:53:58 <rjk2> but not leave it on too long
166 2011-10-26 15:54:06 <rjk2> so it screws the difficulty
167 2011-10-26 15:54:45 <Eliel> if you used the type of merged mining setup used at eligius, you could easily start/stop merged mining without affecting regular mining at all.
168 2011-10-26 15:56:49 <ThomasV> is the wiki still down ?
169 2011-10-26 16:00:27 <gavinandresen> ThomasV: is for me
170 2011-10-26 16:00:33 <ThomasV> hi gavin
171 2011-10-26 16:00:36 <diki> very interesting
172 2011-10-26 16:00:46 <diki> i launched bitcoin a few minutes ago and it is still not downloading blocks
173 2011-10-26 16:01:06 <diki> ...speak of the devil, i hate it when i take my time to write this and it starts downloading blocks
174 2011-10-26 16:01:39 <diki> no wait, its not..i confused myself with the ixcoin client
175 2011-10-26 16:02:27 <ThomasV> gavinandresen: I have another question. I figured out my keypairs are not working. if I import keypairs from my wallet, my transaction signing script works. if I use keypairs generated by myself, it does not work, even though the pubkeys seem to generate valid bitcoin addresses. any idea ?
176 2011-10-26 16:02:51 <gavinandresen> ThomasV: no idea.
177 2011-10-26 16:03:42 <ThomasV> gavinandresen: are there special requirements on private keys, other than using secp256k1 ?
178 2011-10-26 16:04:26 <gavinandresen> ThomasV: none that I know of.
179 2011-10-26 16:04:35 <ThomasV> :-(
180 2011-10-26 16:04:45 <gavinandresen> ... but I'm not an ECDSA expert by any stretch of the imagination
181 2011-10-26 16:05:07 <ThomasV> is there someone else I should ask ?
182 2011-10-26 16:05:25 <jrmithdobbs> ThomasV: how are you genn'ing the keys?
183 2011-10-26 16:05:37 <jrmithdobbs> ThomasV: i'd bet you've got an endian issue
184 2011-10-26 16:05:52 <jrmithdobbs> eg, flipping them when you shouldn't
185 2011-10-26 16:05:58 <ThomasV> jrmithdobbs: I use python-ecdsa
186 2011-10-26 16:06:06 <jrmithdobbs> ThomasV: link to code?
187 2011-10-26 16:06:42 <ThomasV> jrmithdobbs: my script works well with keypairs that I import from an existing wallet, so I don't think it's an endian issue
188 2011-10-26 16:07:15 <jrmithdobbs> ThomasV: it could be on the generation step since that code wont get hit on import
189 2011-10-26 16:07:20 <ThomasV> I can paste my code in pastebin if you are interested, one sec
190 2011-10-26 16:07:36 <jrmithdobbs> ThomasV: are you importing both public and private keys? what happens if you rearrange your code so it only imports the private key and re-gens the pubkey?
191 2011-10-26 16:07:37 <ThomasV> jrmithdobbs: yes that's what I suspect
192 2011-10-26 16:08:13 <jrmithdobbs> ThomasV: i bet you're flipping the pubkey after generation or flipping the privkey before or similar
193 2011-10-26 16:08:26 <jrmithdobbs> i had headaches with that shit and openssl, i know
194 2011-10-26 16:08:26 <ThomasV> how so ?
195 2011-10-26 16:08:42 <ThomasV> moment..
196 2011-10-26 16:08:48 <jrmithdobbs> ThomasV: paste the code somewhere ;p
197 2011-10-26 16:08:58 <jrmithdobbs> ThomasV: i suck re: python though
198 2011-10-26 16:11:02 <ThomasV> you want the whole file, or just the part the generates the keys ?
199 2011-10-26 16:12:36 <ThomasV> jrmithdobbs: http://pastebin.com/gmyEhyqL
200 2011-10-26 16:12:38 <jrmithdobbs> ThomasV: the import and generating code should be enough
201 2011-10-26 16:13:02 <ThomasV> oh the import was me copy pasting keys from pywallet --dump
202 2011-10-26 16:14:00 <jrmithdobbs> ThomasV: what are the use or whatever statements for the libraries that uses?
203 2011-10-26 16:14:19 <ThomasV> it uses python-ecdsa
204 2011-10-26 16:15:12 <ThomasV> and pywallet too, byt that is mostly for debugging; I don't really need it
205 2011-10-26 16:15:24 <ThomasV> *but*
206 2011-10-26 16:15:35 <jrmithdobbs> ThomasV: what's SecretToASecret?
207 2011-10-26 16:16:44 <ThomasV> it's in pywallet, it converts to base58
208 2011-10-26 16:17:39 <ThomasV> in that piece of code I generate the bitcoin address from secexp, and I save sec and the address
209 2011-10-26 16:17:47 <jrmithdobbs> ThomasV: my first guess is that pywallet.public_key_to_bc_address is the culprit
210 2011-10-26 16:18:15 <ThomasV> huh?
211 2011-10-26 16:18:22 <jrmithdobbs> ThomasV: can you check whether that's swapping endianness? I bet it's expecting a LE value when it's already BE
212 2011-10-26 16:18:59 <jrmithdobbs> as in, you're genning a valid address, it just doesn't relate to the key you think it does
213 2011-10-26 16:20:24 <ThomasV> yes I understand what you mean, but I don't think it's trhe case
214 2011-10-26 16:20:58 <jrmithdobbs> ThomasV: see if it works if you reverse the string you generate in that call
215 2011-10-26 16:21:16 <jrmithdobbs> easy enough to test ;p
216 2011-10-26 16:22:04 <jrmithdobbs> because your usage of python-ecdsa looks correct afaict
217 2011-10-26 16:22:04 <ThomasV> I have several instances of     assert addr == pywallet.public_key_to_bc_address( chr(4) + public_key.to_string() )
218 2011-10-26 16:22:18 <ThomasV> and they all pass
219 2011-10-26 16:22:40 <ThomasV> but let me add one at the generation step...
220 2011-10-26 16:23:12 <jrmithdobbs> ThomasV: that code's too simple for it to really be anything else but something reversing order you're not expecting
221 2011-10-26 16:27:43 <jrmithdobbs> ThomasV: i see the problem
222 2011-10-26 16:27:49 <ThomasV> yes?
223 2011-10-26 16:27:53 <jrmithdobbs> ThomasV: you know pywallet isn't expecting a hex string, right?
224 2011-10-26 16:28:02 <ThomasV> for what?
225 2011-10-26 16:28:05 <jrmithdobbs> it's actually expecting an unsigned char array
226 2011-10-26 16:28:15 <jrmithdobbs> for public_key_to_bc_address()
227 2011-10-26 16:28:30 <jrmithdobbs> and hash_160() and b58encode()
228 2011-10-26 16:28:57 <ThomasV> I didn't know
229 2011-10-26 16:29:08 <jrmithdobbs> ya that's going to be the problem
230 2011-10-26 16:29:21 <jrmithdobbs> you're giving it an ascii hex string when it's expecting binary values
231 2011-10-26 16:29:27 <ThomasV> hmm, thanks
232 2011-10-26 16:29:41 <jrmithdobbs> look at line 338ish of https://github.com/jackjack-jj/pywallet/blob/master/pywallet.py
233 2011-10-26 16:29:50 <jrmithdobbs> the pywallet code pretty clearly says what it's expecting
234 2011-10-26 16:30:09 <jrmithdobbs> """ encode v, which is a string of bytes, to base58.
235 2011-10-26 16:30:30 <jrmithdobbs> implying that things up the call stack are probably expecting the same, the docs are a bit lacking there, heh
236 2011-10-26 16:31:12 <ThomasV> indeed :-)
237 2011-10-26 16:32:09 <ThomasV> heh, I lost 1 bitcent that I sent to that address :-)
238 2011-10-26 16:35:46 <Diablo-D3> [12:13:48] <gmaxwell> Diablo-D3: it certantly does it when clients get a long poll and have to switch work. It would be insane to award credit against old blocks.
239 2011-10-26 16:35:48 <Diablo-D3> nope, its not that
240 2011-10-26 16:36:05 <Diablo-D3> DM dosent send works in on dead works, infact LP explicitly empties the queue
241 2011-10-26 16:36:11 <Diablo-D3> only ones already in process would slip through
242 2011-10-26 16:36:30 <Diablo-D3> gmaxwell: but thus far 6322 good, 63 bad
243 2011-10-26 16:36:48 <Diablo-D3> or 0.67%
244 2011-10-26 16:36:51 <Diablo-D3> which is insane
245 2011-10-26 16:38:00 <Diablo-D3> gmaxwell: hell, its rejected shares from works less than a second old and no LP has happened
246 2011-10-26 16:38:28 <ThomasV> jrmithdobbs: no, it's not that
247 2011-10-26 16:38:32 <Diablo-D3> if its not doing what I said, its just forgetting to record works
248 2011-10-26 16:38:34 <Diablo-D3> which is just as bad
249 2011-10-26 16:38:40 <Diablo-D3> Im going to bed, night all
250 2011-10-26 16:38:51 <jrmithdobbs> ThomasV: oh ya?
251 2011-10-26 16:38:56 <ThomasV> jrmithdobbs: the assert statement would fail
252 2011-10-26 16:39:10 <ThomasV> I check that I get the same address back from the priv key
253 2011-10-26 16:39:40 <jrmithdobbs> what's the assert?
254 2011-10-26 16:40:13 <ThomasV> assert original_address == pywallet.public_key_to_bc_address( chr(4) + public_key.to_string() )
255 2011-10-26 16:40:37 <ThomasV> where public_key is derived from private key
256 2011-10-26 16:40:59 <gmaxwell> Diablo-D3: does a LP happen shortly after?
257 2011-10-26 16:41:41 <jrmithdobbs> ThomasV: what's original_address derived from?
258 2011-10-26 16:42:33 <ThomasV> in the case of imported keys, it is derived from the bitcoin wallet with pywallet --dump
259 2011-10-26 16:42:51 <jrmithdobbs> ThomasV: because if you're setting original_address the same way as in the generation code they'll still match
260 2011-10-26 16:42:53 <ThomasV> the assert passes with both my keys and the imported keys
261 2011-10-26 16:43:08 <ThomasV> yes, but it passes both
262 2011-10-26 16:43:12 <jrmithdobbs> ah
263 2011-10-26 16:44:03 <jrmithdobbs> ThomasV: debugger time tbqh
264 2011-10-26 16:44:13 <ThomasV> heh
265 2011-10-26 16:44:17 <ThomasV> thanks anyway
266 2011-10-26 16:44:57 <jrmithdobbs> ThomasV: i'd still see what happens if you plug one of the imported private keys into the generation code and just skip the random number step and see if it generates the expected public key
267 2011-10-26 16:45:32 <ThomasV> that's exactly what I did with the assert
268 2011-10-26 16:47:02 <ThomasV> jrmithdobbs: http://pastebin.com/DcgrMA6E
269 2011-10-26 16:48:10 <jrmithdobbs> ThomasV: crazy
270 2011-10-26 16:48:59 <ThomasV> jrmithdobbs: I thought that there was some other constraint on the secret number, that I missed
271 2011-10-26 16:49:26 <jrmithdobbs> ThomasV: no all the weirdness is in the encoding for storage/display
272 2011-10-26 16:49:42 <ThomasV> ok
273 2011-10-26 16:50:02 <jrmithdobbs> i'm intriguied but can't be arsed to install python 2.7.2
274 2011-10-26 16:50:03 <jrmithdobbs> heh
275 2011-10-26 16:53:22 <diki> w00t bitcoin's got 80 connections
276 2011-10-26 16:53:24 <diki> me like
277 2011-10-26 16:55:20 <jrmithdobbs> pansy
278 2011-10-26 16:55:21 <jrmithdobbs> "connections" : 512,
279 2011-10-26 16:55:27 <ThomasV> anyway, time to eat
280 2011-10-26 16:55:29 <ThomasV> bbl
281 2011-10-26 16:58:19 <BlueMatt> gavinandresen: I need rc2 so that I can properly name my debian packages...
282 2011-10-26 16:58:45 <BlueMatt> gavinandresen: also, it looks like Ive got pretty solid packages, will you link to the ppa when 0.5 comes out (and maybe link in 0.5rc2 release announcement for testing?)
283 2011-10-26 16:59:30 <gavinandresen> BlueMatt: grumble grumble... I really wanted the how-to-build-on-windows-and-mac instructions to be correct for rc2...
284 2011-10-26 16:59:47 <BlueMatt> heh, I was just complaining
285 2011-10-26 16:59:56 <BlueMatt> you dont have to, it really can wait
286 2011-10-26 17:00:47 <gavinandresen> RE: linking to the ppa:  I don't know nuthin about ppa's, but if you write up something and it's ok with jgarzik I have no problem including it in release notes.
287 2011-10-26 17:01:12 <BlueMatt> all that would need be done is (maybe) link to https://launchpad.net/~bitcoin/+archive/bitcoin and tell people to execute: sudo apt-add-repository ppa:bitcoin/bitcoin
288 2011-10-26 17:01:36 <BlueMatt> (though bitcoin package is currently broken until I can reupload...)
289 2011-10-26 17:01:54 <gavinandresen> What is the trust model for that?  Have to trust the ppa package maintainers and whoever controls launchpad.net?
290 2011-10-26 17:02:03 <BlueMatt> yes and yes
291 2011-10-26 17:02:13 <BlueMatt> 1st being me (and others if they are added to the team)
292 2011-10-26 17:02:16 <gavinandresen> Ok, thanks.
293 2011-10-26 17:02:25 <BlueMatt> launchpad being canonical (the makers of ubuntu)
294 2011-10-26 17:02:25 <gavinandresen> (I trust you, by the way...)
295 2011-10-26 17:03:03 <BlueMatt> so Id say theve got a good level of trust... (not that it wouldnt be better to do gitian-style stuff, but os-standard is the way to go IMO)
296 2011-10-26 17:03:24 <BlueMatt> (that said, Id of course like to see bitcoin released on win32 via gitian, and updated the same)
297 2011-10-26 17:03:43 <BlueMatt> (osx, maybe you could submit to app store, not like it would get accepted, but its worth a shot...)
298 2011-10-26 17:04:03 <gavinandresen> agreed.  By the way:  I was very close to successfully building in a Windows AWS instance, but ran into a pthread issue compiling bitcoin that stumped me
299 2011-10-26 17:04:18 <diki> gavinandresen:amazon ec2?
300 2011-10-26 17:05:14 <BlueMatt> gavinandresen: and, for the record, I would prefer to have you at least control the group and launchpad account, even if you dont have the time/effort to do the actual package uploads
301 2011-10-26 17:05:33 <gavinandresen> diki: yes
302 2011-10-26 17:56:10 <diki> the windows instance is expensive
303 2011-10-26 17:56:23 <diki> and considering you want to setup a build environment
304 2011-10-26 17:58:34 <diki> My VS Studio environment is somehow screwed
305 2011-10-26 17:58:43 <diki> I was able to compile curl once, now i get tons of errors
306 2011-10-26 17:58:46 <rjk2> get a linux instance, install virtualbox, install windows ??? profit :P
307 2011-10-26 17:58:54 <rjk2> although slow as hell
308 2011-10-26 17:59:00 <diki> isnt that...a hassle>
309 2011-10-26 17:59:10 <diki> cause initially you need to ssh to the machine
310 2011-10-26 17:59:10 <rjk2> </sarcasm>
311 2011-10-26 18:01:03 <jgarzik> gavinandresen, BlueMatt: that sounds like a good ppa setup
312 2011-10-26 18:05:17 <diki> jgarzik
313 2011-10-26 18:05:49 <diki> Can you tell me in your 4way algo in minerd, where exactly is the nonce increased..i saw no ++ variable or +=X or something
314 2011-10-26 19:37:14 <ThomasV> jrmithdobbs: the problem is not caused by my key generation...
315 2011-10-26 20:10:40 <CIA-101> libbitcoin: genjix * r5a5abb36f86b / (7 files in 5 dirs): base58_encode/decode(...). Borried heavily from bitcointools and Satoshi bitcoin (thanks)
316 2011-10-26 21:11:43 <Davincij15> Hello? Is anyone here I have a question about the bitcoin.conf file...
317 2011-10-26 21:11:58 <Davincij15> Is there away to use connect=dnsname.com
318 2011-10-26 21:12:16 <Davincij15> Instead of an IP?
319 2011-10-26 21:15:18 <tcatm> Davincij15: yes, there's a -dns switch
320 2011-10-26 21:18:10 <diki> the bitcoin wiki is down
321 2011-10-26 21:18:25 <rjk2> been down all day
322 2011-10-26 21:18:29 <rjk2> wonder whats up
323 2011-10-26 21:24:08 <diki> hope it gets fixed soon
324 2011-10-26 22:25:58 <graingert> !later tell BlueMatt that's true but I thought ~bitcoin was a group? Perhaps there should be informal sub groups focusing on release/dev/daily build streams
325 2011-10-26 22:25:58 <gribble> The operation succeeded.
326 2011-10-26 23:06:59 <gjs278> ;;bc,stats
327 2011-10-26 23:07:02 <gribble> Current Blocks: 150756 | Current Difficulty: 1468195.4272208 | Next Difficulty At Block: 151199 | Next Difficulty In: 443 blocks | Next Difficulty In About: 3 days, 18 hours, 11 minutes, and 59 seconds | Next Difficulty Estimate: 1236664.43277946 | Estimated Percent Change: -15.7697667592
328 2011-10-26 23:23:22 <NullMoogleCable> hi
329 2011-10-26 23:26:47 <cjdelisle> ho
330 2011-10-26 23:27:02 <NullMoogleCable> so im trying to figure out how i can compile cpuminer to run better on my arm system
331 2011-10-26 23:27:21 <cjdelisle> how many systems do you have?
332 2011-10-26 23:27:47 <NullMoogleCable> about 4 diferent arm based systems i want to play with
333 2011-10-26 23:27:57 <cjdelisle> hmm
334 2011-10-26 23:28:28 <cjdelisle> just so you know, it's not going to be profitable unless you had like a botnet of a million and a half arm systems
335 2011-10-26 23:28:48 <cjdelisle> but for fun, that's cool..
336 2011-10-26 23:29:03 <NullMoogleCable> I was thinking more of using it as a benchmark
337 2011-10-26 23:29:37 <cjdelisle> hmm
338 2011-10-26 23:30:04 <cjdelisle> it's not a good indicator of system performance since it requires lots of bit twiddling and very little state is kept
339 2011-10-26 23:30:29 <cjdelisle> *not a good indicator of overall system perfornamce