1 2013-09-03 00:24:56 <phantomcircuit> gmaxwell, i cant replicate the invalid block issue on debian sid
  2 2013-09-03 00:25:11 <phantomcircuit> there's more to the story than just using debian sid
  3 2013-09-03 00:25:29 <phantomcircuit> (this is with reindexing and also randomly killing between 5-10 minutes)
  4 2013-09-03 01:18:58 <ahmedbodi> is Luke-Jr around? or anyone thats used eloipool before?
  5 2013-09-03 02:23:20 <maaku> ahmedbodi: better luck on #bitcoin-mining or similar
  6 2013-09-03 02:23:31 <ahmedbodi> thnks maaku
  7 2013-09-03 02:28:16 <amiller> i'm trying to understand how conflicting transactions affect misbehavior
  8 2013-09-03 02:28:29 <amiller> if you send a node a transact that conflicts with a transaction in its mempool
  9 2013-09-03 02:28:39 <amiller> then does it assign you a misbehavior penalty?
 10 2013-09-03 02:29:11 <amiller> i think it does
 11 2013-09-03 02:36:20 <CodeShark> no, it does not - it would be a bad thing because it would make it easy for an attacker to split the network
 12 2013-09-03 02:41:00 <amiller> CodeShark, this line is the else branch for if a coin is rejected from mempool
 13 2013-09-03 02:41:01 <amiller> https://github.com/bitcoin/bitcoin/blob/master/src/main.cpp#L3768
 14 2013-09-03 02:41:28 <amiller> uh its really the check on state.IsInvalid after that
 15 2013-09-03 02:41:51 <CodeShark> that's if it's missing inputs - not if it conflicts
 16 2013-09-03 02:43:27 <CodeShark> line 820 is where the check for conflicts begins
 17 2013-09-03 02:45:05 <amiller> ahh okay so a conflict makes mempool.accept() return false but it doesn't mark validation state as invalid
 18 2013-09-03 02:46:21 <amiller> thanks :]
 19 2013-09-03 03:09:56 <warren> Luke-Jr: http://luke.dashjr.org/programs/bitcoin/files/charts/versions.txt  if this is up to date, this is somewhat surprising.
 20 2013-09-03 03:19:30 <gmaxwell> amiller: the _goal_ with the misbehavior code is that nodes which are correctly working but potentially surrounded by attackers won't make an honest peer disconnect them. Or, in other words, misbehavior should never be node-transitive across a correct node.
 21 2013-09-03 03:19:48 <amiller> ok that makes more sense
 22 2013-09-03 03:20:32 <gmaxwell> amiller: It's perfectly possible that we've screwed that up someplace so you should watch out for that, but every time we add one we're evaluating it against that criteria. (and if you find such a case let us know!)
 23 2013-09-03 03:25:14 <Luke-Jr> warren: it should be; why is it surprising?
 24 2013-09-03 03:58:51 <numismatics> is there a (python) script that will perform OP_CHECKSIG?
 25 2013-09-03 04:14:00 <maaku> numismatics: look at pynode
 26 2013-09-03 04:14:32 <numismatics> ty
 27 2013-09-03 04:49:08 <gmaxwell> https://bitcointalk.org/index.php?topic=172009.msg3069540#msg3069540  :-/ oh gesh what now
 28 2013-09-03 04:51:38 <k9quaint> just more grist for my booming tinfoil business
 29 2013-09-03 04:52:47 <gmaxwell> "Bitfoil"
 30 2013-09-03 04:52:59 <gmaxwell> someone should totally make a site that sells little foil hats for bitcoin.
 31 2013-09-03 04:53:08 <gmaxwell> "Protection against the state's mind control rays!"
 32 2013-09-03 04:56:25 <gmaxwell> "With enhanced bernanke blocking bimetal construction"
 33 2013-09-03 05:02:46 <porquilho> gmaxwell lol
 34 2013-09-03 05:38:15 <k9quaint> grrr, somebody already has bitfoil.com :(
 35 2013-09-03 05:40:43 <warren> Luke-Jr: surprising so many are still running 0.8.1
 36 2013-09-03 05:41:07 <gmaxwell> warren: we haven't done an alert induced upgrade beyond that.
 37 2013-09-03 05:41:22 <gmaxwell> and there have always been a lot of older nodes...
 38 2013-09-03 05:42:01 <warren> we alerted all previous clients and a large number of them are still running the year old version ...
 39 2013-09-03 05:44:01 <phantomcircuit> gmaxwell, "Note that it starts with 43 zero bits. Why? The block target difficulty was much lower (around 32 bits), so we can assume Satoshi did this on purpose."
 40 2013-09-03 05:44:02 <phantomcircuit> wat
 41 2013-09-03 05:44:04 <phantomcircuit> no we cant
 42 2013-09-03 05:44:38 <k9quaint> are you applying logic to that post?
 43 2013-09-03 05:45:01 <k9quaint> I hope you are at least wearing safety goggles
 44 2013-09-03 05:45:22 <gmaxwell> phantomcircuit: yea evidence suggests he didn't in fact, there wasn't enough extra nonce motion there to indicate substantially more than 2^32 work, and moreover there were no 2^32 sufficient solutions with smaller nonces.
 45 2013-09-03 05:45:50 <gmaxwell> Sergio does some good analysis generally, though I think he concludes a little too fast (sometimes a flaw I share too!)
 46 2013-09-03 05:47:45 <k9quaint> "There is a long history of using aluminum to both deflect mind control signals and to control them. Evidence suggests that the ancient Atlanteans used aluminum armor and psychotrons in their wars with the Phoenician Old Dynasty."
 47 2013-09-03 05:48:17 <k9quaint> sometimes, I think the url specification is too flexible when I land on sites with stuff like that on them
 48 2013-09-03 05:49:28 <warren> k9quaint: I thought ancient Atlanteans had powered force fields and hyperdrive tech, or at least the ones on TV did...
 49 2013-09-03 05:50:11 <k9quaint> I started looking up how to make top qualify hats out of tinfoil and almost immediately regretted it
 50 2013-09-03 05:50:22 <k9quaint> I will just stick to selling foil in bulk for BTC :(
 51 2013-09-03 05:50:57 <nsh> gmaxwell, what's being contended there? (re: satoshi/nonce motion)
 52 2013-09-03 05:51:30 <nsh> ;;google Note that it starts with 43 zero bits
 53 2013-09-03 05:51:31 <gribble> IPv6 address - Wikipedia, the free encyclopedia: <http://en.wikipedia.org/wiki/IPv6_address>; SparkNotes: Holes: Chapters 36–43: <http://www.sparknotes.com/lit/holes/section10.rhtml>; Oracle Certified System Configurations - Java SE 6: <http://www.oracle.com/technetwork/java/javase/system-configurations-135212.html>
 54 2013-09-03 05:51:34 <nsh> meh
 55 2013-09-03 05:52:13 <nsh> 
ACTION reads the thread

 56 2013-09-03 05:52:22 <warren> k9quaint: they also make lead foil.  It's quite fragile, but assuming you have no holes it might be better than aluminum at this goal...
 57 2013-09-03 05:53:35 <gmaxwell> it needs to be quad metal foil, a layer of tin (for durability), a layer of alumnium (for mind control rays), a layer of sliver (for RF immunity), and a layer of gold (found soundness).
 58 2013-09-03 05:53:41 <gmaxwell> s/found/for/
 59 2013-09-03 05:53:58 <k9quaint> yes, because the symptoms of heavy metal poisoning is just what we want to add on to the various neurosis of the bitcoin world ;)
 60 2013-09-03 05:54:19 <nsh> help! my cheap multifoil has clay feet!
 61 2013-09-03 05:55:00 <warren> k9quaint: I want to put that quote on a billboard somewhere.
 62 2013-09-03 05:55:10 <warren> context isn't needed
 63 2013-09-03 05:55:41 <k9quaint> "This too shall pass" <-- another statement that is always true, no matter the context :P
 64 2013-09-03 07:44:54 <arioBarzan> is bitcoin still using irc as one of its bootstrap methods?
 65 2013-09-03 07:45:57 <gmaxwell> No.
 66 2013-09-03 07:46:03 <gmaxwell> and hasn't in a long time.
 67 2013-09-03 07:47:19 <SomeoneWeird> a very long time
 68 2013-09-03 07:47:24 <arioBarzan> I see strMainNetDNSSeed and also pnSeed in net.cpp. are these two main sources of finding initial nodes, assuming peers.dat is empty?
 69 2013-09-03 07:49:01 <warren> generally yes.  IRC was disabled a while ago.
 70 2013-09-03 07:49:03 <gmaxwell> yep, unless you've provided nodes via addr.txt addnode or connect.
 71 2013-09-03 07:49:27 <gmaxwell> warren: not "generally" the code is gone.
 72 2013-09-03 07:49:47 <warren> yes to his question
 73 2013-09-03 07:49:53 <warren> nevermind...
 74 2013-09-03 07:50:27 <arioBarzan> gmaxwell: may I ask whether you are a contributor to litecoin as well?
 75 2013-09-03 07:50:44 <warren> 
ACTION facepalm

 76 2013-09-03 07:51:21 <gmaxwell> arioBarzan: only by virtue of litcoin copying code I've written.
 77 2013-09-03 07:52:40 <JyZyXEL> can the current blockchain be downloaded with the older 0.7 clients or has there been a fork yet?
 78 2013-09-03 07:53:14 <warren> JyZyXEL: fork has happened, use Luke-Jr's 0.7 backport which remains compatible
 79 2013-09-03 07:53:38 <gmaxwell> I made some recommendations about LTC's broken fee policy after it first started and as a result the litcoin community accused me of attacking it. Art argued against using scrypt for POW with me, then put it in an altcoin. I got called names by the LTC community when I pointed out that the state space of the parameters they used wouldn't stop gpus at all.
 80 2013-09-03 07:53:55 <Diablo-D3> gmaxwell: btw, its not that scrypt is bad
 81 2013-09-03 07:53:59 <Diablo-D3> its that they use scrypt wrtong
 82 2013-09-03 07:54:02 <gmaxwell> While there are some perfectly fine litecoin users, the ltc community as a whole strikes me as being filled with not nice people.
 83 2013-09-03 07:54:08 <Diablo-D3> if I started an alt coin right now
 84 2013-09-03 07:54:12 <Diablo-D3> I'd use scrypt
 85 2013-09-03 07:54:44 <gmaxwell> Diablo-D3: its a dumb goal in any case, go look at the logs from here in late 2010 / early 2011 with art and I arguing over it. He convinced me.
 86 2013-09-03 07:55:23 <warren> gmaxwell: the ltc community as defined as who?  Seems to be entirely different people from back then.
 87 2013-09-03 07:55:33 <Diablo-D3> gmaxwell: well, I think the goal of "cpu defined mining only" is a worthy goal
 88 2013-09-03 07:55:53 <warren> Yes, Art failed at that, on purpose or not doesn't matter today.
 89 2013-09-03 07:55:57 <gmaxwell> POW converts energy into proof, the goal is consuming a scarce resource so the ideal rational behavior is to only put all your effort on the one chain you think will dominate. Assuming all the details hold, all you can do is shuffle around some constant factors.
 90 2013-09-03 07:55:58 <warren> It is what it is.
 91 2013-09-03 07:56:39 <Diablo-D3> gmaxwell: yeah but if I were to write my own altchain
 92 2013-09-03 07:56:47 <Diablo-D3> gmaxwell: I would obviously be believing my chain will win
 93 2013-09-03 07:57:13 <gmaxwell> Diablo-D3: this wasn't a commentary between altcoins doof, it's a comment about actually achieving a consensus inside one.
 94 2013-09-03 07:57:28 <Diablo-D3> gmaxwell: yeah, but we dont have to do that, thats what the software is for
 95 2013-09-03 07:57:33 <gmaxwell> (Though I suppose it applies somewhat between too all other things equal)
 96 2013-09-03 07:57:39 <gmaxwell> ...
 97 2013-09-03 07:57:49 <Diablo-D3> what?
 98 2013-09-03 07:58:00 <Diablo-D3> if they dont think bitcoin is the future, why are they using it so much
 99 2013-09-03 07:58:21 <Diablo-D3> and dont give me that "well, they use ltc" shit
100 2013-09-03 07:58:29 <Diablo-D3> they cant use their ltc, exchange them for btc, and then sell the btc.
101 2013-09-03 07:58:51 <JyZyXEL> warren: know how long it took after 15th of May when the safeguard was lifted that a fork occured?
102 2013-09-03 07:58:59 <warren> What did folks expect GPU owners would do when ASIC's made it impossible to mine Bitcoin?   Just sell hardware to <gaming market> and quit?  (This is not arguing that scrypt GPU is a good thing.  It just is, and one day in the future it won't be.)
103 2013-09-03 07:59:14 <warren> JyZyXEL: hm, happened in recent weeks
104 2013-09-03 07:59:41 <gmaxwell> warren: well litecoin promised that gpus couldn't mine it. ::shrugs::
105 2013-09-03 07:59:53 <gmaxwell> JyZyXEL: "fork" is complicated.
106 2013-09-03 08:00:18 <warren> I mean, Litecoin exists and it is what it is.  I'm not saying it's good.  There are things about it that I don't like and can't be changed.
107 2013-09-03 08:00:39 <gmaxwell> JyZyXEL: versions prior to 0.8 couldn't follow along with new blocks in a somewhat non-determinstic way that depended on the fine structure of the node's local database.
108 2013-09-03 08:01:25 <gmaxwell> JyZyXEL: there are still unmodified 0.7 nodes that are trucking along, though the recent trigger block seems to have gotten most of them.
109 2013-09-03 08:01:52 <gmaxwell> JyZyXEL: the non-determinism of it was why it was absolutely essential to fix it via a hardfork.
110 2013-09-03 08:02:50 <gmaxwell> warren: profitability of bitcoin gpu mining was pretty low by many people's standards long before any asic stuff was available. And we had reasonably efficient fpga miners for a long time to bridge the gap.
111 2013-09-03 08:03:24 <gmaxwell> warren: most people with really big gpu farms (myself included) had wound them down and sold them off sometime before.
112 2013-09-03 08:03:24 <JyZyXEL> gmaxwell: are you saying its not certain a one trigger block exists that triggers all the 0.7 users?
113 2013-09-03 08:03:26 <warren> gmaxwell: quite clever, those who bought FPGA's then scared away more FPGA buyers with promises of ASIC "soon" =)
114 2013-09-03 08:03:47 <Diablo-D3> yes and now
115 2013-09-03 08:03:50 <Diablo-D3> asic buyers
116 2013-09-03 08:03:51 <gmaxwell> warren: hm, I don't think "those woh bought FPGAs" were doing that, BFL did that.
117 2013-09-03 08:03:53 <Diablo-D3> are scaring away potential buyers
118 2013-09-03 08:03:54 <Diablo-D3> with
119 2013-09-03 08:03:57 <arioBarzan> If one wants to mine some fake bitcoins on mainnet, I thought he would have to run at least two nodes locally. Apparently you don't need that. if you mine only one fake block, then you could put its hash in checkpoints and continue mining as a misbehaving node.
120 2013-09-03 08:03:57 <Diablo-D3> 28nm asics
121 2013-09-03 08:04:03 <Diablo-D3> and talk about some mystical end game
122 2013-09-03 08:04:26 <gmaxwell> Diablo-D3: no one with asic in hands seem to have even the slighest problem selling them— the prices the market is paying for miners are higher than they've ever bin.
123 2013-09-03 08:04:37 <Diablo-D3> gmaxwell: maybe
124 2013-09-03 08:04:45 <Diablo-D3> that reminds me, I need to pay dmc dividends
125 2013-09-03 08:04:49 <gmaxwell> arioBarzan: huh? it will not mine unless there is a connection up.
126 2013-09-03 08:05:22 <gmaxwell> getblocktemplate and getwork should both return nothing when you have no connections up.
127 2013-09-03 08:05:35 <warren> gmaxwell: I'm not sure about "most", when ASIC's finally pushed BTC difficulty well outside of any GPU profitability, we saw massive scrypt hashrate increases in Litecoin and the <too many> scrypt clones equivalent to maybe 15-20TH of sha256d GPU's.
128 2013-09-03 08:05:46 <gmaxwell> (of course you can go remove that test)
129 2013-09-03 08:06:03 <gmaxwell> warren: or someone finally has some reasonably efficient scrypt mining FPGA farms going.
130 2013-09-03 08:06:23 <gmaxwell> there was never 20TH of sha256d gpu power, not even close.
131 2013-09-03 08:06:32 <warren> gmaxwell: probably a mix of both
132 2013-09-03 08:06:49 <warren> there's also a lot of people buying GPU's *now* /me facepalm
133 2013-09-03 08:07:18 <JyZyXEL> i wonder how many GPU's have died from mining
134 2013-09-03 08:07:31 <gmaxwell> (I think people don't quite appreciate how fast the 28nm FPGAs were, since no one was productizing them)
135 2013-09-03 08:07:36 <arioBarzan> gmaxwell: After putting my fake block in checkpoints I connected to mainnet and because it connects to other nodes I could continue mining. but my mined blocks obviously did not get accepted by anybody rather than myself.
136 2013-09-03 08:07:55 <gmaxwell> arioBarzan: oh sure.
137 2013-09-03 08:08:17 <gmaxwell> kind of a really lame and inefficient way to replace a single extra command.
138 2013-09-03 08:08:44 <gmaxwell> arioBarzan: the only reason it works that way is so that you don't waste a lot of power mining on a dead end chain on a totally isolated node.
139 2013-09-03 08:11:39 <arioBarzan> I suspect my ip get blacklisted by other nodes very fast, because of lots of invalid blocks and transactions.
140 2013-09-03 08:13:55 <warren> arioBarzan: what is your goal in asking these seemingly random questions?
141 2013-09-03 08:14:18 <arioBarzan> I guess if one node pull a %51 attack, maybe other nodes could fork their own chain and put it in checkpoints.
142 2013-09-03 08:18:28 <gmaxwell> arioBarzan: they could, but why bother?
143 2013-09-03 08:18:58 <gmaxwell> the attacker would just spoil them again, and if you can establish consensus using checkpoints you don't need bitcoin at all— base your currency on that mechenism whatever it is.
144 2013-09-03 08:22:39 <warren> damn it.  I was going to respond to him.
145 2013-09-03 08:25:17 <gmaxwell> http://bitslog.wordpress.com/2013/09/03/new-mystery-about-satoshi/  here is sergio's latest post.
146 2013-09-03 08:26:35 <gmaxwell> though I don't think there is any great mystery (and endianness? lol) ... the nonce gets reset when there is a new block. if you seep nonces starting at zero at some particular rate which isn't a big multiple of 2^32/600 you'll end up with a non-uniform distribution.
147 2013-09-03 08:37:17 <JyZyXEL> a secret message embedded into the blockchain! :-O
148 2013-09-03 08:39:21 <gmaxwell> no, no secret message.
149 2013-09-03 08:39:33 <gmaxwell> Broken software, almost certantly.
150 2013-09-03 08:39:45 <gmaxwell> or just statistics...
151 2013-09-03 08:42:37 <JyZyXEL> i like how insane it gets with the list of values picked up from the statistics and then the ideas of what those values could mean :))
152 2013-09-03 10:40:58 <kha0Z> hello
153 2013-09-03 10:45:31 <_dr> gmaxwell: I think his point in the blog was that even with some nonce wrapping there is still some slight bias towards 0 in the msb
154 2013-09-03 10:45:37 <_dr> which, to me, seems about right
155 2013-09-03 10:45:54 <arioBarzan> Since Microsoft is buying Nokia mobile phone unit, is there a concern about possible patent claims on qt-framework, although it is currently licensed under LGPL 2.1?
156 2013-09-03 10:46:06 <_dr> but I have to admit that I've never been too much of a statistics guy
157 2013-09-03 10:46:19 <_dr> it just sounds right 'intuitively' ;)
158 2013-09-03 10:51:35 <_dr> but his interpretation is rather disappointing
159 2013-09-03 12:02:23 <Luke-Jr> Diablo-D3: if you used scrypt correctly, you'd make it worse in other ways - verification would require just as much memory as mining
160 2013-09-03 12:19:45 <TD> petertodd: is your testnet seed fully working at the moment? i can only get 1 IP out of it
161 2013-09-03 12:20:08 <TD> petertodd: and that peer is dead
162 2013-09-03 12:42:22 <phantomcircuit> TD, is it just me or is it impossible to build multibit?
163 2013-09-03 12:45:04 <TD> er. i doubt it's impossible.
164 2013-09-03 12:45:09 <TD> i don't work on it so it's been a while since i tried
165 2013-09-03 12:45:14 <TD> what is the issue?
166 2013-09-03 12:50:53 <TD> phantomcircuit: anyway i suggest you just email jim. for some reason the maven file refers to some "MB-ALICE" version of bitcoinj, that i've never heard of. it must be jims branch of it.
167 2013-09-03 12:50:58 <TD> 
ACTION has no clue what ALICE is short for

168 2013-09-03 12:51:30 <phantomcircuit> TD, it's not even that it's things like org.multibit.store not found
169 2013-09-03 12:51:38 <phantomcircuit> which makes sense because that doesn't exist
170 2013-09-03 12:51:50 <TD> which branch are you trying to build?
171 2013-09-03 12:52:05 <phantomcircuit> TD, i've tried both master and develop with the same issue
172 2013-09-03 12:52:20 <phantomcircuit> it feels like there's some code that's not in git
173 2013-09-03 12:52:32 <TD> ah
174 2013-09-03 12:52:36 <TD> it discusses this in the readme
175 2013-09-03 12:53:11 <TD> look at the section "A note on the bitcoinj dependency"
176 2013-09-03 12:53:28 <TD> it says you need to clone his fork and then do a "mvn clean install"
177 2013-09-03 12:53:37 <TD> after that it should be able to find the org.multibit.store classes and stuff
178 2013-09-03 12:54:28 <TD> jim should really unfork himself
179 2013-09-03 12:54:31 <TD> but i think he knows that
180 2013-09-03 12:54:35 <phantomcircuit> lets give it a try
181 2013-09-03 13:00:50 <mhanne> hi! since you're talking about bitcoinj.. i would like to run the blocktester thing against bitcoin-ruby, but i'm a bit lost.. is there any documentation about it i've missed?
182 2013-09-03 13:01:38 <mhanne> i think mvn clean package on master is working now, but not on bluematt/blocktester branch.. and even then i have no idea how to run it..
183 2013-09-03 13:01:46 <phantomcircuit> TD, lol it fails a test
184 2013-09-03 13:02:11 <TD> mhanne: email matt. the block tester uses bitcoinj but it's not maintained by me.
185 2013-09-03 13:02:21 <phantomcircuit> transactionConfidence
186 2013-09-03 13:02:32 <phantomcircuit> PeerGroupTest
187 2013-09-03 13:02:33 <TD> phantomcircuit: doh! you can do "mvn install -DskipTests" to get past that
188 2013-09-03 13:02:42 <TD> phantomcircuit: can you pastebin me the failure? it might be flaky
189 2013-09-03 13:03:01 <mhanne> TD: thanks, will do
190 2013-09-03 13:04:51 <ThomasV> with bip32 and public derivations, if someone knows the private key of m/0/, will they be able to infer the parent's private key m/ ?
191 2013-09-03 13:05:54 <phantomcircuit> ThomasV, yes
192 2013-09-03 13:05:56 <phantomcircuit> TD, http://pastebin.com/raw.php?i=k0VPXGv1
193 2013-09-03 13:05:58 <phantomcircuit> there's a ton more info but it's hard to tell where the test starts
194 2013-09-03 13:05:59 <phantomcircuit> (i think)
195 2013-09-03 13:06:45 <TD> that's a multi-line error. anyway, i'd just skip it for now. the tests don't fail in upstream bitcoinj. so it's either got some system dependency i'm not aware of, or it's flaky, or jim broke it :)
196 2013-09-03 13:06:56 <ThomasV> phantomcircuit: care to explain how?
197 2013-09-03 13:07:15 <phantomcircuit> ThomasV, i actually dont understand it i just remember that being on of the issues
198 2013-09-03 13:09:12 <phantomcircuit> TD, lol that fails in the examples... let me get you an error
199 2013-09-03 13:09:30 <TD> i guess i should try his branch at some point
200 2013-09-03 13:09:54 <phantomcircuit> TD, http://pastebin.com/raw.php?i=75vnde2U
201 2013-09-03 13:10:09 <phantomcircuit> he merged a bunch of conflicts on the 31st
202 2013-09-03 13:10:17 <phantomcircuit> im guessing he didn't get it exactly right
203 2013-09-03 13:10:31 <TD> ah
204 2013-09-03 13:10:32 <TD> yes
205 2013-09-03 13:10:47 <TD> i took that dependency verification stuff out of examples because it wasn't ready to go
206 2013-09-03 13:11:01 <TD> just delete that section of the examples/pom.xml - the stuff related to the dependency verifier
207 2013-09-03 13:11:13 <TD> it's telling you, you have a "hacked" copy of the library. which as it's a fork is sorta true but mostly not :)
208 2013-09-03 13:11:20 <TD> or you can just skip that crap and go into the core directory
209 2013-09-03 13:11:24 <TD> and run "mvn install" from there
210 2013-09-03 13:11:30 <TD> that'll skip examples and tools which you don't really need anyway
211 2013-09-03 13:12:36 <ThomasV> " Given a child extended private key (ki,ci) and the integer i, an attacker cannot find the parent private key kpar more efficiently than a 2256 brute force of HMAC-SHA512. "
212 2013-09-03 13:13:19 <phantomcircuit> ThomasV, iirc there is a huge difference between private derivation and public
213 2013-09-03 13:13:21 <ThomasV> I'm wondering if there is a difference between private and public derivations
214 2013-09-03 13:13:30 <ThomasV> yes that's my point
215 2013-09-03 13:13:54 <ThomasV> than sentence is from the wiki
216 2013-09-03 13:13:58 <ThomasV> *that*
217 2013-09-03 13:15:56 <phantomcircuit> TD, Failed tests:   testExportTransactions(org.multibit.file.ExportTransactionsSubmitActionTest): Row 1 incorrect expected:<29 Jul 2013 [10]:23,"Sent to ""unenc...> but was:<29 Jul 2013 [02]:23,"Sent to ""unenc...>
218 2013-09-03 13:16:09 <phantomcircuit> that looks like a timezone problem
219 2013-09-03 13:16:16 <ThomasV> phantomcircuit: oh, the wiki says it. thanks
220 2013-09-03 13:16:23 <phantomcircuit> i'll just skip the tests again...
221 2013-09-03 13:16:28 <TD> that's some test jim has added
222 2013-09-03 13:16:35 <TD> yeah looks like a dependency on british time
223 2013-09-03 13:17:05 <phantomcircuit> ONLY THE BRITISH MAY WORK ON MULTIBIT
224 2013-09-03 13:17:10 <TD> :)
225 2013-09-03 13:19:56 <phantomcircuit> TD, http://pastebin.com/raw.php?i=YyzDHWxC
226 2013-09-03 13:19:59 <phantomcircuit> wat
227 2013-09-03 13:21:05 <TD> lol
228 2013-09-03 13:21:25 <phantomcircuit> incase you're wondering
229 2013-09-03 13:21:31 <phantomcircuit> that file seems to actually exist
230 2013-09-03 13:21:35 <phantomcircuit> and be readable
231 2013-09-03 13:22:09 <TD> he checked in a complete copy of a mac jdk into git?
232 2013-09-03 13:22:23 <phantomcircuit> TD, it would appear so
233 2013-09-03 13:22:41 <TD> well, i doubt you care about the packaging and shrinking aspects.
234 2013-09-03 13:22:53 <TD> you've got a multibit.jar
235 2013-09-03 13:23:04 <TD> you should be able to run it now: cd target; java -jar multibit.jar
236 2013-09-03 13:23:34 <phantomcircuit> TD, actual directory src/main/skeleton/mac/MultiBit.app/Contents/PlugIns/
237 2013-09-03 13:23:42 <phantomcircuit> TD, directory it's looking for src/main/skeleton/mac/MultiBit.app/Contents/Plugins
238 2013-09-03 13:23:42 <TD> ahh
239 2013-09-03 13:23:47 <phantomcircuit> >.>
240 2013-09-03 13:23:50 <TD> HFS+ isn't case sensitive, is it?
241 2013-09-03 13:23:55 <TD> i think jim develops on a mac
242 2013-09-03 13:24:17 <TD> it's a single developer project, obscure dependencies on his setup probably sneak through because not many other people try to work on it
243 2013-09-03 13:24:59 <phantomcircuit> TD, i can understand why it happens
244 2013-09-03 13:25:16 <phantomcircuit> but i cant understand why people think this is production quality
245 2013-09-03 13:25:30 <phantomcircuit> especially since clearly nobody else is building it
246 2013-09-03 13:26:36 <TD> *shrug* it has users. they download and run it. it works. it'd be nice if multibit was a larger project, for sure.
247 2013-09-03 13:27:02 <TD> unfortunately the alternatives are bitcoin-qt which has lots of developers, but is unusable as it's only full mode, or blockchain.info which is also just a single developer
248 2013-09-03 13:27:49 <phantomcircuit> TD, as of 0.8.1 even the archive mode bitcoin-qt loads in about 48 hours on my netbook
249 2013-09-03 13:28:01 <phantomcircuit> which is super underpowered compared to even the cheapest laptop
250 2013-09-03 13:29:04 <TD> i'm pretty sure basically all wallets except the reference client are single developer, unfortunately. although in reality multibit is really "jim+me+others who work on bitcoinj"
251 2013-09-03 13:29:15 <TD> ditto for the android app
252 2013-09-03 13:31:04 <TD> i'm expecting that bitcoin-qt will become an SPV wallet as well at some point. then there'll be more competition
253 2013-09-03 13:31:35 <phantomcircuit> TD, lol i tried to remove the wallet code
254 2013-09-03 13:31:47 <phantomcircuit> it took me like 2 hours and i still had things randomly breaking
255 2013-09-03 13:31:53 <phantomcircuit> just to remove it
256 2013-09-03 13:32:05 <TD> sipa has done headers first sync, which is the bulk of the work.
257 2013-09-03 13:32:15 <Belxjander> SPV?
258 2013-09-03 13:32:20 <TD> although there are quite a lot of details to handle after that
259 2013-09-03 13:32:30 <TD> spv aka "fast mode"
260 2013-09-03 13:33:11 <Belxjander> Who works on the "Bitcoin Wallet" for Android? or the "Mycellum Wallet" ?
261 2013-09-03 13:34:10 <Belxjander> TD: I've been considering making a "port" to AmigaOS for actually dealing with BlockChain's and CryptoCurrency wallets in a somewhat abstracted sense with any Application able to make use of a "Wallet" class
262 2013-09-03 13:36:01 <TD> AmigaOS?
263 2013-09-03 13:36:06 <TD> the android app is andreas schildbach
264 2013-09-03 13:36:35 <Belxjander> yeah ...
265 2013-09-03 13:36:44 <Belxjander> www.hyperion-entertainment.biz for the current developers
266 2013-09-03 13:37:04 <Belxjander> www.acube-systems.biz + www.a-eon.com for the hardware
267 2013-09-03 13:37:22 <Belxjander> modified from "Embedded" systems
268 2013-09-03 14:28:57 <TD> goddamnit. what i wouldn't do for a functioning testnet block explorer
269 2013-09-03 14:32:30 <jgarzik> mapsz 1400 on testnet?  good grief.
270 2013-09-03 14:32:32 <jgarzik> TD, amen
271 2013-09-03 14:58:09 <jouke> TD: I'll see if I can setup an abe instance for testnet.
272 2013-09-03 14:58:24 <TD> jouke: that would be fantastic. does abe show unconfirmed transactions as well?
273 2013-09-03 15:00:50 <jgarzik> +1
274 2013-09-03 15:03:35 <jouke> I know it has mempool support for bitcoin, I guess it will work with testnet as well.
275 2013-09-03 15:12:25 <TD> jouke: that would be great
276 2013-09-03 15:13:03 <JyZyXEL> does it show unconfirmed txs?
277 2013-09-03 15:13:47 <AndyOfiesh> Is there a net gain in security running Bitcoin-Qt as root, given that the installation has not been tampered with?
278 2013-09-03 15:15:10 <Luke-Jr> no
279 2013-09-03 15:15:20 <Luke-Jr> there is never any security gain running things as root, AFAIK
280 2013-09-03 15:15:27 <AndyOfiesh> It seems that would prevent user-level applications from yanking your private keys out of RAM. Is the risk greater than the benefit?
281 2013-09-03 15:15:49 <Luke-Jr> AndyOfiesh: well, I'd suggest running it as a different user than your other software, but not as root ;)
282 2013-09-03 15:24:14 <jouke> Hmm, my new testnetnode is not getting any peers. Can I just copy a peers.dat from an other node?
283 2013-09-03 15:25:07 <michagogo> jouke: No connections at all?
284 2013-09-03 15:25:32 <michagogo> You could try a couple `addnode <address> onetry`
285 2013-09-03 15:25:33 <michagogo> s
286 2013-09-03 15:25:55 <jouke> I don't know any testnet nodes.
287 2013-09-03 15:26:25 <abrkn> does sendMany accept string as numbeR?
288 2013-09-03 15:26:49 <michagogo> no, it's {"string":decimal}
289 2013-09-03 15:26:59 <abrkn> ok, that makes it a bit difficult with javascript
290 2013-09-03 15:27:28 <abrkn> i JS, 0.00000001 evaluates to 1e-8
291 2013-09-03 15:27:34 <abrkn> (is expressed as)
292 2013-09-03 15:28:07 <abrkn> JSON.stringify({ a: 0.00000001 }) == "{"a":1e-8}"
293 2013-09-03 15:29:47 <michagogo> jouke: what OS?
294 2013-09-03 15:30:26 <michagogo> Use whatever tool is available on $os to run a DNS lookup on testnet-seed.bluematt.me and/or testnet-seed.bitcoin.petertodd.org
295 2013-09-03 15:31:46 <jouke> michagogo: I have copied an existing peers.dat and used it.
296 2013-09-03 15:31:51 <michagogo> Ah, okay
297 2013-09-03 15:32:21 <michagogo> (you could also `addnode 93.172.79.35 onetry`)
298 2013-09-03 15:32:34 <jouke> Oh. Hmmm. It could pretty well be my routers DNS-suckage indeed.
299 2013-09-03 15:33:54 <cz3141> have a "lost wallet issue". willing to offer a bounty, any suggestions/recommendations pls pm.
300 2013-09-03 15:34:15 <michagogo> cz3141: What does "lost" mean in this case?
301 2013-09-03 15:35:46 <michagogo> cz3141: Lost in what way?
302 2013-09-03 15:35:59 <cz3141> wallet on hd that has been reformatted and had new os installed.
303 2013-09-03 15:36:38 <cz3141> and has been used for some time now. just received payment to wallet address that i thought i had updated to new wallet.
304 2013-09-03 15:36:54 <michagogo> Ah.
305 2013-09-03 15:37:02 <_dr> hm. really noone wrote a tool to find wallets yet?
306 2013-09-03 15:37:33 <michagogo> First thing: if that hard drive is currently mounted read/write, unmount it right away before anything else
307 2013-09-03 15:37:42 <cz3141> yes.
308 2013-09-03 15:37:57 <Belxjander> cz3141: full or quick formatting?
309 2013-09-03 15:38:26 <cz3141> full ): on debian
310 2013-09-03 15:38:33 <Belxjander> then its gone
311 2013-09-03 15:38:38 <michagogo> Ah, a full format... That's likely to zap it
312 2013-09-03 15:38:43 <cz3141> that's what i guessed.
313 2013-09-03 15:38:46 <michagogo> How much bitcoin are we talking here?
314 2013-09-03 15:38:47 <Belxjander> michagogo: yeah its zapped
315 2013-09-03 15:38:48 <cz3141> 21
316 2013-09-03 15:38:54 <michagogo> Ouch. :-/
317 2013-09-03 15:39:03 <_dr> some forensics lab might be able to recover them
318 2013-09-03 15:39:06 <cz3141> its an address i forgot was out there to receive pmt tx
319 2013-09-03 15:39:07 <Belxjander> cz3141: that is a substantial loss
320 2013-09-03 15:39:12 <_dr> but i guess that will cost more than 21btc :(
321 2013-09-03 15:39:21 <Diablo-D3> ;;ticker
322 2013-09-03 15:39:21 <gribble> MtGox BTCUSD ticker | Best bid: 146.40000, Best ask: 146.40004, Bid-ask spread: 0.00004, Last trade: 146.40000, 24 hour volume: 12973.98319875, 24 hour low: 142.20000, 24 hour high: 148.90893, 24 hour vwap: 145.56612
323 2013-09-03 15:39:23 <cz3141> not insignificant unfortunately
324 2013-09-03 15:39:28 <Diablo-D3> ;;calc 21 * 146
325 2013-09-03 15:39:28 <gribble> 3066
326 2013-09-03 15:39:31 <Diablo-D3> actually
327 2013-09-03 15:39:33 <michagogo> cz3141: Yeah, addresses "out there to receive payment" are usually bad things
328 2013-09-03 15:39:35 <Diablo-D3> it wont cost you that much to do it
329 2013-09-03 15:39:39 <michagogo> Address reuse is bad.
330 2013-09-03 15:39:49 <Diablo-D3> about $500
331 2013-09-03 15:40:09 <cz3141> i'm listening Diablo-D3
332 2013-09-03 15:40:39 <Diablo-D3> theres a few companies that can do it, google for em
333 2013-09-03 15:41:27 <cz3141> m'ok. will shop around. thanks for suggestion.
334 2013-09-03 15:41:44 <michagogo> cz3141: There does exist software that will scan the HD for anything resembling bitcoin privkeys, such as http://www.btcnn.com/2011/09/recovering-lost-due-to-format-bitcoin.html
335 2013-09-03 15:41:59 <michagogo> But I don't know that any of that can help if a hard drive was full-formatted
336 2013-09-03 15:42:34 <cz3141> i think i saw that. anyone tried it?
337 2013-09-03 15:42:40 <_dr> michagogo: yes. downloading binaries and executing them with root privileges is always a good idea :)
338 2013-09-03 15:42:54 <cz3141> lol. why yes.
339 2013-09-03 15:43:13 <cz3141> which is why i hadnt tried it yet, thought i would ask around some first.
340 2013-09-03 15:43:16 <michagogo> Well, it *is* open source
341 2013-09-03 15:43:27 <michagogo> You can read the source and check that it's not malicious.
342 2013-09-03 15:43:28 <cz3141> true. will take a peek.
343 2013-09-03 15:43:44 <Diablo-D3> michagogo: well
344 2013-09-03 15:43:44 <_dr> just have a look at the source code of the client. maybe the wallet stores addresses (which include a hashsum)
345 2013-09-03 15:43:48 <Diablo-D3> full formatting still doesnt
346 2013-09-03 15:43:58 <Diablo-D3> but you wont be able to recover it with THAT software
347 2013-09-03 15:44:08 <Diablo-D3> but it might still be able to be recovered
348 2013-09-03 15:44:27 <Diablo-D3> (its why military spec /dev/random writing exists, just to prevent ANY recovery)
349 2013-09-03 15:44:30 <_dr> cz3141: you would have to get a forensics lab to recover the previous (i.e. pre-formating) contents of your disk
350 2013-09-03 15:44:32 <cz3141> some info out there dicusses opening hex editor and searching manually
351 2013-09-03 15:44:40 <_dr> and then you can run the wallet-finder on this raw image
352 2013-09-03 15:44:56 <cz3141> _dr ok.
353 2013-09-03 15:45:20 <pigeons> is /dev/random writing any better than writing zeros?
354 2013-09-03 15:46:20 <_dr> to really really safely delete you also need several iterations
355 2013-09-03 15:46:41 <michagogo> Diablo-D3: What *exactly* does this format operation do?
356 2013-09-03 15:47:09 <Diablo-D3> michagogo: full? overwrites it with zeros
357 2013-09-03 15:47:17 <pigeons> really, you can recover from writing zeros once, but not several times? how many times?
358 2013-09-03 15:47:28 <Diablo-D3> pigeons: depends on the drive
359 2013-09-03 15:47:40 <Diablo-D3> its the urandom that fucks you over
360 2013-09-03 15:48:05 <pigeons> oh i just use the java SecureRandom instead
361 2013-09-03 15:48:08 <pigeons> ;)
362 2013-09-03 15:48:09 <Diablo-D3> do that 3 times, and no one is recovering it, I dont care if the military standard is like 7 (which is bullshit anyhow, since they throw decomissioned drives into a furnace)
363 2013-09-03 15:48:32 <_dr> heh
364 2013-09-03 15:48:48 <_dr> talk about green it
365 2013-09-03 15:49:47 <maaku> pigeons: magnetic writes on disk don't perfectly overlap
366 2013-09-03 15:50:06 <maaku> and SSDs corrode differently based on their charge
367 2013-09-03 15:50:06 <phantomcircuit> Diablo-D3, afaik nobody has ever recovered data from a driver after even a single pass on any modern drive (manufactured within the last say 30 years)
368 2013-09-03 15:50:41 <phantomcircuit> maaku, the principle issue with doing an overwrite is more that drives silently remap sectors they think are failing
369 2013-09-03 15:50:44 <Diablo-D3> phantomcircuit: of urandom? fbi claims they can do 3
370 2013-09-03 15:50:51 <k9quaint> thats why I store all my sensitive information in /dev/null
371 2013-09-03 15:50:54 <Diablo-D3> phantomcircuit: zeros however do much less damage
372 2013-09-03 15:50:57 <maaku> phantomcircuit: you can buy atomic-force data recovery for a few thousands dollars
373 2013-09-03 15:51:00 <maaku> it's a commodity service
374 2013-09-03 15:51:21 <Diablo-D3> now if you're on SSD and secure erased it?
375 2013-09-03 15:51:22 <Diablo-D3> boned
376 2013-09-03 15:51:27 <phantomcircuit> maaku, that's really not for overwritten data
377 2013-09-03 15:51:38 <maaku> phantomcircuit: yes it is, and yes i have
378 2013-09-03 15:51:39 <phantomcircuit> it's for where the controller/head crashed
379 2013-09-03 15:51:58 <maaku> phantomcircuit: you can get atomic force recovery too
380 2013-09-03 15:52:04 <maaku> they offer more than one service
381 2013-09-03 15:52:22 <maaku> the only difference is how you filter the data
382 2013-09-03 15:52:27 <cz3141> so... i guess i have a new day tading target :D
383 2013-09-03 15:53:33 <cz3141> *trading
384 2013-09-03 15:57:50 <phantomcircuit> maaku, securely erasing a specific file on an ssd is more or less impossible
385 2013-09-03 15:58:43 <phantomcircuit> Diablo-D3, i doubt the FBI can actually recover much of anything after even a single pass of random data
386 2013-09-03 15:59:25 <phantomcircuit> im sure they have various means of finding copies of the files though
387 2013-09-03 15:59:48 <pigeons> they can use http://getprsm.com/ to get the backups
388 2013-09-03 15:59:49 <phantomcircuit> especially under windows where things like volume shadow copy are constantly silently copying things all over the place
389 2013-09-03 16:01:00 <gmaxwell> Luke-Jr: Had any time to think about https://github.com/bitcoin/bitcoin/pull/2945 ?
390 2013-09-03 16:02:06 <Luke-Jr> gmaxwell: not really. it makes things cheaper for spammers to do 1:1 stuff; maybe it should use a difference?
391 2013-09-03 16:02:35 <Luke-Jr> gmaxwell: eg, you only get the input "costless" when there isn't a matching output
392 2013-09-03 16:03:21 <gmaxwell> Luke-Jr: If you do that then you will incentivize people adding extra outputs when they can do so for free, as they can then redeem them later. I point this out.
393 2013-09-03 16:03:49 <Luke-Jr> hmm
394 2013-09-03 16:03:56 <gmaxwell> Luke-Jr: it won't make it cheaper for spammers to do 1:1 stuff, except in so far as that it effectiviely lowers the free priority threshold a bit.
395 2013-09-03 16:04:11 <gmaxwell> And I considered doing that sensible just because of how much bitcoin has increased in value.
396 2013-09-03 16:04:30 <gmaxwell> e.g. 1 BTC for 1 day costs a lot more now then it did a couple years ago.
397 2013-09-03 16:04:32 <Luke-Jr> gmaxwell: it doesn't affect the size used by fee-by-kB calc?
398 2013-09-03 16:04:39 <gmaxwell> No, it doesn't.
399 2013-09-03 16:05:06 <gmaxwell> Didn't do that simply because I'm still assuming miners are income-maximizing on those.
400 2013-09-03 16:06:16 <gmaxwell> though there probably should be some tradeoff, like "I'm willing to forgo 0.01 BTC/block in order to optimize for priority rather than income" but then it goes deep down the path of non-unique solutions again.
401 2013-09-03 16:08:06 <abrkn> gmaxwell: do you know how the json rpc handles sendmany { "1addr....": "0.00000001" }? javascript's JSON.stringify of the number 0.00000001 is "1e-8", which im not sure if bitcoind's json parser would handle
402 2013-09-03 16:08:59 <gmaxwell> abrkn: we handle that.
403 2013-09-03 16:09:29 <gmaxwell> wait "0.00000001" ?
404 2013-09-03 16:09:29 <TD> jouke: the DNS seeds are broken. here are two
405 2013-09-03 16:09:37 <TD> 199.26.85.40,54.243.211.176
406 2013-09-03 16:10:07 <abrkn> gmaxwell: "0.00000001" (one satoshi, as a string)
407 2013-09-03 16:10:19 <gmaxwell> abrkn: we handle { "1addr....": 0.00000001 }  and { "1addr....": 1e-8 } The quotes are right out.
408 2013-09-03 16:10:47 <abrkn> gmaxwell: as im sure you know JS handles small numbers as floats
409 2013-09-03 16:11:00 <abrkn> gmaxwell: not sure how small they need to be to lose accuracy
410 2013-09-03 16:11:30 <Luke-Jr> restarted DNS seed
411 2013-09-03 16:12:07 <gmaxwell> abrkn: all bitcoin values are preserved in an implementation using doubles and normal conversion. But you could avoid your concern just by constructing the request manually.
412 2013-09-03 16:12:29 <gmaxwell> abrkn: e.g. printing the whole thing, not just the value.
413 2013-09-03 16:12:38 <gmaxwell> abrkn: in any case, no you can't use a string there.
414 2013-09-03 16:13:01 <abrkn> gmaxwell: ok, thanks for clarifying
415 2013-09-03 16:14:26 <phantomcircuit> abrkn, if it's off it's going to be off by a trivially small amount
416 2013-09-03 16:14:41 <phantomcircuit> the question is how much effort is it worth to make sure it's never off
417 2013-09-03 16:18:25 <jouke> TD: It is running, but it is behind a NAT at the moment. I'll set up a reverse tunnel tomorrow (it will take some time to catch up anyway).
418 2013-09-03 16:24:05 <TD> jouke: yay!
419 2013-09-03 16:40:12 <numismatics> no dice on pynode
420 2013-09-03 16:40:26 <numismatics> any suggestions for script to compute UP_CHECKSIG?
421 2013-09-03 16:54:39 <maaku> numismatics: https://github.com/jgarzik/python-bitcoinlib/blob/master/bitcoin/scripteval.py#L368
422 2013-09-03 16:54:46 <maaku> (that's part of pynode)
423 2013-09-03 16:56:17 <numismatics> ty, i gues i'm just not smart enough to figure out what I'm trying to do
424 2013-09-03 16:57:57 <jgarzik> numismatics, define "compute"
425 2013-09-03 16:58:11 <jgarzik> numismatics, are you trying to execute a script, construct a script, or something else?
426 2013-09-03 16:58:44 <michagogo> Has anyone here been seeding the bootstrap.dat torrent for more than a week or so?
427 2013-09-03 16:59:06 <jgarzik> michagogo, has it been more than a week since I updated it?
428 2013-09-03 16:59:37 <michagogo> jgarzik: Just checked, it was on the 30th
429 2013-09-03 16:59:52 <michagogo> Specifically, I'm looking for people still seeding the old versions
430 2013-09-03 17:00:23 <jgarzik> woo!  more pynode users appear.
431 2013-09-03 17:00:28 <jgarzik> 
ACTION looks at petertodd 

432 2013-09-03 17:00:47 <numismatics> i want to know the hash of the output to be signed
433 2013-09-03 17:01:20 <numismatics> so if I fed the script the r & s values to compute z
434 2013-09-03 17:01:46 <jgarzik> numismatics, the hash of the output to be signed is specified as a transaction input, COutpoint
435 2013-09-03 17:05:33 <numismatics> i guess i'm confused as to how this guy computed the z values in http://www.nilsschneider.net/2013/01/28/recovering-bitcoin-private-keys.html
436 2013-09-03 17:06:37 <numismatics> and i dont understand how to use pynode to accomplish said task
437 2013-09-03 17:06:44 <gmaxwell> numismatics: If you rob fewer people you'll tend to have less of these problems.
438 2013-09-03 17:07:00 <numismatics> :) i'm not going to rob anyone
439 2013-09-03 17:07:19 <numismatics> i want to understand the underlying crypto function
440 2013-09-03 17:07:34 <numismatics> clearly, i lack the knowledge to be dangerous =P
441 2013-09-03 17:07:50 <gmaxwell> yea but you're working on it!
442 2013-09-03 17:08:16 <numismatics> hah :)
443 2013-09-03 17:08:21 <gmaxwell> numismatics: signature validation has two inputs— the message being signed and the public key. You're looking for the message being signed there.
444 2013-09-03 17:09:32 <numismatics> then what's referred to as z will be part of the message being signed?
445 2013-09-03 17:10:16 <gmaxwell> technically the hash of the data being signed.
446 2013-09-03 17:13:18 <numismatics> gotchya; and pynode lets me explore the underlying messages (unlike blockchain.info or blockexplorer?)
447 2013-09-03 17:23:38 <handle> pynode?
448 2013-09-03 17:23:53 <handle> 
ACTION looks it up

449 2013-09-03 17:24:28 <handle> ooh, this is pretty neat
450 2013-09-03 17:25:00 <handle> so is this just a reference bitcoin clienthalf- implementation written in python?
451 2013-09-03 17:25:55 <handle> jgarzik: ^
452 2013-09-03 17:26:10 <jgarzik> s/reference//
453 2013-09-03 17:26:20 <handle> that works
454 2013-09-03 17:26:22 <jgarzik> it's an implementation + library (python-bitcoinlib)
455 2013-09-03 17:26:53 <handle> interesting
456 2013-09-03 17:27:08 <handle> so I assume you're supposed to run your own bitcoind and connect this to it?
457 2013-09-03 17:27:40 <phantomcircuit> handle, it's a library
458 2013-09-03 17:27:45 <phantomcircuit> you do whatever you want with it
459 2013-09-03 17:28:12 <handle> well yeah, certainly
460 2013-09-03 17:28:22 <handle> but if it au
461 2013-09-03 17:28:28 <handle> yeah, suppose you're right :P
462 2013-09-03 17:31:43 <jgarzik> handle, for pynode, it is fully validating (and slow!), so you may connect it to any node.
463 2013-09-03 17:31:58 <jgarzik> handle, python-bitcoinlib is a full library, not requiring any other software
464 2013-09-03 17:33:22 <jgarzik> Luke-Jr, now that RPC is a part of python-bitcoinlib, we need to figure out a good disposition for python-bitcoinrpc
465 2013-09-03 17:35:55 <Luke-Jr> jgarzik: ?
466 2013-09-03 17:36:07 <numismatics> anyone have a node I can connect to?
467 2013-09-03 17:36:11 <numismatics> rather than my own =P
468 2013-09-03 17:36:17 <jgarzik> numismatics, us2.exmulti.net or eu3.exmulti.net
469 2013-09-03 17:36:29 <jgarzik> Luke-Jr, to fix the eloipool problem
470 2013-09-03 17:36:50 <Luke-Jr> jgarzik: revert the regression? :P
471 2013-09-03 17:36:57 <numismatics> credentials?
472 2013-09-03 17:37:05 <jgarzik> numismatics, ?
473 2013-09-03 17:37:05 <Luke-Jr> or merge the fix that guy made; it's not perfect, but it's better than nothing
474 2013-09-03 17:37:15 <numismatics> i thought i needed RPCusername/pass
475 2013-09-03 17:37:31 <jgarzik> Luke-Jr, inside python-bitcoinlib, the utility of faking authserviceproxy is quite low
476 2013-09-03 17:37:39 <jgarzik> versus python-bitcoinrpc
477 2013-09-03 17:37:59 <Luke-Jr> jgarzik: it's not related to faking ASP
478 2013-09-03 17:38:00 <jgarzik> numismatics, P2P does not require credentials
479 2013-09-03 17:38:25 <Luke-Jr> hmm, lemme try cooking up a real fix, 2 sec
480 2013-09-03 17:38:35 <michagogo> Does addnode work with dns addresses?
481 2013-09-03 17:39:23 <michagogo> ...apparently it does.
482 2013-09-03 17:39:51 <numismatics> forgive my ignorance, then what would my pynode config look like?
483 2013-09-03 17:45:53 <handle> https://github.com/bitcoin/bitcoin.org/blob/master/jgarzik-exmulti.asc do you have a stake in exmulti?
484 2013-09-03 17:48:37 <jgarzik> handle, 100% stake, yes
485 2013-09-03 17:48:50 <jgarzik> numismatics, key=value
486 2013-09-03 17:49:15 <handle> ah nice
487 2013-09-03 17:59:20 <Luke-Jr> petertodd: I'm gonna have to revert that patch.. my client keeps crashing -.-
488 2013-09-03 18:40:05 <jgarzik> Luke-Jr, thanks
489 2013-09-03 18:40:16 <jgarzik> Luke-Jr, now I need to see if that applies to python-bitcoinlib also
490 2013-09-03 18:40:26 <jgarzik> Luke-Jr, ideally, python-bitcoinrpc is entirely retired
491 2013-09-03 18:42:01 <Luke-Jr> jgarzik: I don't see a reason for python-bitcoinlib for simple stuff
492 2013-09-03 18:42:20 <Luke-Jr> sounds like it does a lot more than just RPC communications
493 2013-09-03 18:44:18 <jgarzik> Luke-Jr, yes.  If you need to create base58, build or parse or sign transactions or blocks or P2P messages, ...   the idea is that python-bitcoinlib will do it for you, with only openssl as a dep
494 2013-09-03 18:44:45 <jgarzik> Luke-Jr, most bitcoin-RPC apps need additional bitcoin lib code, besides just RPC.
495 2013-09-03 18:44:45 <Luke-Jr> jgarzik: right, my point is that if all someone wants to do is RPC, they don't want all that other stuff, or an additional dep
496 2013-09-03 18:44:53 <Luke-Jr> jgarzik: especially a GPL-incompatible dep >_<
497 2013-09-03 18:45:31 <jgarzik> Luke-Jr, what does eloipool use for P2P?  its own code?
498 2013-09-03 18:45:59 <jgarzik> P2P message parsing, etc.
499 2013-09-03 18:46:19 <jgarzik> gavinandresen, gmaxwell, any final 0.8.4 blockers?
500 2013-09-03 18:47:17 <Luke-Jr> jgarzik: yes
501 2013-09-03 18:47:38 <gmaxwell> Luke-Jr: huh?
502 2013-09-03 18:47:42 <Luke-Jr> jgarzik: as Eloipool is AGPL, it wouldn't work to use an OpenSSL-tied dep
503 2013-09-03 18:47:46 <Luke-Jr> gmaxwell: ?
504 2013-09-03 18:48:00 <gmaxwell> jgarzik: I have been scratching my head why 0.8.4 isn't shipped yet. That needs to get out.
505 2013-09-03 18:48:13 <gmaxwell> Luke-Jr: thought you were commenting on 0.8.4 blockers.
506 2013-09-03 18:48:53 <Luke-Jr> ah, no. just saying Eloipool has its own node code (which predates other Python node code..)
507 2013-09-03 18:49:01 <abrkn> gmaxwell: im still struggling with this scentific notation. sending the amount 1e-8 gives me an error that i dont have enough in the wallet, which i assume mean its not getting parsed right by the json rpc. ideas?
508 2013-09-03 18:49:14 <Luke-Jr> abrkn: you can't send amount 1e-8
509 2013-09-03 18:49:48 <abrkn> Luke-Jr: ok, im struggling because javascript insist on expressing one satoshi as 1e-8
510 2013-09-03 18:49:51 <gmaxwell> abrkn: outputs that small won't be relayed or mined by the network currently.
511 2013-09-03 18:49:56 <Luke-Jr> abrkn: 1 satoshi is too small to send
512 2013-09-03 18:50:00 <gmaxwell> abrkn: I think your problem is not what you think it is.
513 2013-09-03 18:50:13 <Luke-Jr> abrkn: try sending 1 BTC
514 2013-09-03 18:50:20 <Luke-Jr> abrkn: and do testing on testnet. that's what it's there for.
515 2013-09-03 18:50:42 <abrkn> but when doing sendmany, isnt a fee added anyway?
516 2013-09-03 18:51:15 <Luke-Jr> abrkn: you cannot send that amount, even with a fee
517 2013-09-03 18:51:25 <abrkn> i see. what it the smallest i can send?
518 2013-09-03 18:51:31 <Luke-Jr> something like 5000 satoshis I think
519 2013-09-03 18:51:38 <Luke-Jr> but seriously. TESTNET
520 2013-09-03 19:23:23 <abrkn> Luke-Jr: thanks for the heads up. how does blockchain.info send tiny transactions? i noticed they take quite a while to get included
521 2013-09-03 19:25:17 <Luke-Jr> abrkn: if they do, I presume they have some private deals with irresponsible miners
522 2013-09-03 19:26:05 <michagogo> (or just a node connection to a pool that does that)
523 2013-09-03 19:26:30 <abrkn> what's a reasonable minimum to set for withdraws to not be considered spamming the network?
524 2013-09-03 19:26:42 <michagogo> 5460 satoshis is the minimum to be relayed, IIRC
525 2013-09-03 19:26:55 <michagogo> (though a bunch of people have it wrong and think it's 5430)
526 2013-09-03 19:27:38 <gribble> 0.007834404
527 2013-09-03 19:27:38 <michagogo> But remember, that's still tiny: ,,(calc [ticker --last] * 0.0000543)
528 2013-09-03 19:27:47 <michagogo> less than a penny
529 2013-09-03 19:32:33 <Luke-Jr> michagogo: pools that do that are the irresponsible miners ;p
530 2013-09-03 19:32:49 <michagogo> Yeah, but no need for "private deals"
531 2013-09-03 19:32:58 <michagogo> (also, I thought eligius mined everything it sees?)
532 2013-09-03 19:34:42 <Luke-Jr> michagogo: Eligius will take non-standard, but not spammy looking garbage, and generally requires transaction fees if it seems possibly spammy.
533 2013-09-03 19:34:50 <michagogo> Ah, I see
534 2013-09-03 19:35:10 <michagogo> I guess you made all kinds of algorithms for that?
535 2013-09-03 19:35:16 <Luke-Jr> yeah
536 2013-09-03 19:35:39 <Luke-Jr> it also looks for fees on dependent transactions to pay for the ones they need
537 2013-09-03 19:35:59 <Luke-Jr> so a transaction might not get confirmed until Eligius sees it spent with fees
538 2013-09-03 19:36:10 <michagogo> BTW, did I mention that a couple weeks ago I made a script that completely automates the gitian-build?
539 2013-09-03 19:36:35 <michagogo> I double-click, type in the release number, and it will download, build, sign, and pullrequest
540 2013-09-03 19:36:54 <michagogo> (I'm sure there's a better way, but I did the pull request with octokit and a ruby oneliner)
541 2013-09-03 19:37:15 <Luke-Jr> sipa has had that for months :P
542 2013-09-03 19:37:17 <michagogo> (oh, and add, commit and push)
543 2013-09-03 19:37:30 <michagogo> Yeah, it's not that hard :-P
544 2013-09-03 19:38:06 <michagogo> I think I made the basic script after a build or two that I did manually to ~understand what it was doing (treating gitian as a black box)
545 2013-09-03 19:39:07 <michagogo> But I made it fully automated, with a prompt for the version number and the automated pullrequest, after 4rc2
546 2013-09-03 20:05:10 <BlueMatt> sipa/others who are interested in block-tester's stuff: the code should be fairly readable at https://code.google.com/p/bitcoinj/source/browse/core/src/test/java/com/google/bitcoin/core/BitcoindComparisonTool.java but essentially the algorithm is to send a single-block inv, wait for bitcoind to fetch the block with a getdata then respond with the block, ping, and getheaders to make sure bitcoind has accepted/rejected the block
547 2013-09-03 20:07:56 <gmaxwell> BlueMatt: can you make the tester rerun on https://github.com/bitcoin/bitcoin/pull/2945 ?
548 2013-09-03 20:09:43 <BlueMatt> gmaxwell: done
549 2013-09-03 20:09:52 <gmaxwell> BlueMatt: thanks!
550 2013-09-03 20:10:30 <BlueMatt> sipa: so does block-tester thinggy just need to be able to respond to getheaders in addition to getblocks?
551 2013-09-03 20:11:40 <gmaxwell> BlueMatt: oh does it not respond to getheaders?  It needs to. Assuming sipa's code is right then that should be it.
552 2013-09-03 20:11:55 <BlueMatt> ahh, ok, yea that shouldn't be hard
553 2013-09-03 20:12:25 <gmaxwell> headers first might also justify having some longer reorg tests.. e.g. chains where there is a long run of blocks with an invalid one in the middle.
554 2013-09-03 20:12:44 <gmaxwell> To test the case where if follows a line of headers and must reorg off of it because there is an invalid block.
555 2013-09-03 20:13:12 <gmaxwell> BlueMatt: oh while you're here! the fork alert stuff seems to be causing a fair amount of false positives for people. :(
556 2013-09-03 20:13:25 <BlueMatt> gmaxwell: bug link?
557 2013-09-03 20:14:02 <gmaxwell> There have been a number of forum posts and a couple people on IRC: https://bitcointalk.org/index.php?topic=286013.0
558 2013-09-03 20:14:26 <gmaxwell> I think one case that is triggering it is if you're syncing, and then you get a new block, causing you to start pulling in the other direction and get a bunch of orphan blocks.
559 2013-09-03 20:35:13 <michagogo> Cool, uploading ~200 kB/s bootstrap to the netherlands
560 2013-09-03 20:39:24 <michagogo> Are release tags done by gavin exclusively?
561 2013-09-03 20:40:28 <gavinandresen> good morning everyone.  I'll tag 0.8.4rc2 as final and start gitian building this morning
562 2013-09-03 20:41:02 <michagogo> 
ACTION greets gavin with his morning beverage of choice

563 2013-09-03 20:41:12 <BlueMatt> gmaxwell: yea, probably...you should totally write some so that it does that :)
564 2013-09-03 20:42:41 <gavinandresen>  * [new tag]         v0.8.4 -> v0.8.4
565 2013-09-03 20:43:33 <warren> 
ACTION starting gitian build

566 2013-09-03 20:44:06 <michagogo> Okay, it's quarter to midnight for me and I should get to sleep -- but I'll reboot into Ubuntu and build first.
567 2013-09-03 20:44:10 <michagogo> Goodnight everyone.
568 2013-09-03 20:44:19 <Luke-Jr> gavinandresen: excellent
569 2013-09-03 20:45:42 <michagogo> Expect my PR in a few minutes :-)
570 2013-09-03 20:45:49 <cfields> mm.. I think the chan was ignoring me before because I was unregistered. Can you guys hear me now?
571 2013-09-03 20:46:45 <Luke-Jr> cfields: yes
572 2013-09-03 20:46:52 <cfields> thanks
573 2013-09-03 20:47:01 <cfields> 
ACTION <-- TheUni

574 2013-09-03 20:47:10 <cfields> <cfields> i just got both :)
575 2013-09-03 20:47:10 <cfields> <cfields> warren: was there some bounty around for deterministic osx builds? or linux->osx builds?
576 2013-09-03 20:48:11 <Luke-Jr> hmmnet.cpp:1689: warning: dereferencing pointer ‘sockaddr.1335’ does break strict-aliasing rules
577 2013-09-03 20:48:17 <Luke-Jr> hmm net.cpp*
578 2013-09-03 20:48:37 <Luke-Jr> cfields: wait, really?
579 2013-09-03 20:48:48 <warren> cfields: Luke-Jr mentioned 25 BTC was donated to him for that work, but it wasn't indicated how much of that would be allocated to whoever picks it up.  The Litecoin team will contribute some if Bitcoin people contribute.
580 2013-09-03 20:49:12 <Luke-Jr> warren: will this development upset your friend? did he start on it?
581 2013-09-03 20:49:27 <warren> Luke-Jr: I'll ask, I think he's too busy
582 2013-09-03 20:50:54 <cfields> Luke-Jr: yep, bitcoind verified working. Checking on tests/qt now
583 2013-09-03 20:50:58 <warren> hmm, my 0.8.4 gitian build failed with: cp: cannot stat `/home/ubuntu/out/src/doc/README.md': No such file or directory
584 2013-09-03 20:51:03 <Luke-Jr> cfields: IMO deterministic OS X builds are good enough to say "too bad Windows folks" XD
585 2013-09-03 20:51:28 <cfields> hmm? win32 are deterministic too..?
586 2013-09-03 20:51:52 <warren> cfields: yes
587 2013-09-03 20:52:05 <Luke-Jr> cfields: I mean people trying to build on Windows ;p
588 2013-09-03 20:52:11 <Luke-Jr> aka Diapolo and nobody else
589 2013-09-03 20:52:21 <cfields> oh, heh
590 2013-09-03 20:53:38 <warren> supposedly the virtualbox gitian works, so it would be possible for windows users to do gitian builds
591 2013-09-03 20:53:54 <warren> would require cygwin and porting the entire debian stack to somehow work there...
592 2013-09-03 20:56:21 <Luke-Jr> warren: yes, but Diapolo wants to just click Compile on Qt Creator :p
593 2013-09-03 20:56:27 <michagogo> What's this error I'm getting, something about sha256sum and no such file or directory?
594 2013-09-03 20:57:28 <warren> my gitian builds are failing
595 2013-09-03 20:57:41 <BlueMatt> gmaxwell: hmm, well Ill see if I can get block-tester to respond to getheaders properly, but I may not have time to touch the fork notification stuff for a while yet
596 2013-09-03 20:57:50 <BlueMatt> 
ACTION -> out

597 2013-09-03 20:57:56 <michagogo> I think it had the string "protobuf" in there somewhere?