1 2012-09-13 00:00:44 <MC1984> jgarzik thats what they said 10 years ago
2 2012-09-13 00:16:48 <denisx> Luke-Jr: I think I know why deepbit sometimes serves an unknown block
3 2012-09-13 00:27:42 <sipa> denisx: explain?
4 2012-09-13 00:28:15 <denisx> sipa: on my pool when I found a block alot of miner switch to with really big GH/s
5 2012-09-13 00:28:42 <denisx> so if you artificially send an unknown block out, they will think you got one and switch to you
6 2012-09-13 00:28:52 <Luke-Jr> lol
7 2012-09-13 00:29:18 <gmaxwell> ah hopperbaiting.
8 2012-09-13 00:29:23 <denisx> gmaxwell: yes
9 2012-09-13 00:29:42 <Diablo-D3> lol
10 2012-09-13 00:29:42 <gmaxwell> denisx: dude, you run a proportional pool? Don't do that. It's not fair to your honest but ignorant miners.
11 2012-09-13 00:29:44 <Diablo-D3> thats dirty
12 2012-09-13 00:29:51 <denisx> gmaxwell: no, I don't
13 2012-09-13 00:29:53 <gmaxwell> oh okay.
14 2012-09-13 00:30:10 <denisx> but some miners still hop
15 2012-09-13 00:30:14 <gmaxwell> right.
16 2012-09-13 00:30:49 <sipa> so deepbit is luring those in? Ha.
17 2012-09-13 00:30:55 <denisx> sipa: maybe
18 2012-09-13 00:31:15 <Luke-Jr> Tycho seems to think it's a load balancing issue
19 2012-09-13 00:31:17 <Luke-Jr> fwiw
20 2012-09-13 00:31:30 <Luke-Jr> also, the blocks were real mainnet blocks
21 2012-09-13 00:31:48 <denisx> Luke-Jr: I thought you said total unknown blocks
22 2012-09-13 00:31:53 <Luke-Jr> BFGMiner seems to be buggy when printing the error message - it's the first 32 bytes of the header, not the prevblock
23 2012-09-13 00:32:11 <gmaxwell> accidentally beneficial load balancing issue, too perhaps.
24 2012-09-13 00:34:12 <Luke-Jr> on another note, it sounds like Deepbit is freezing withdrawls of hoppers instead of(?) banning them
25 2012-09-13 00:34:55 <BlueMatt> do they still not use anti-hopper algorithms?
26 2012-09-13 00:35:05 <BlueMatt> s/algorithms/payout algorithms/
27 2012-09-13 00:35:14 <Luke-Jr> BlueMatt: same old prop or PPS
28 2012-09-13 00:35:17 <MC1984> whats an unknown block
29 2012-09-13 00:35:23 <BlueMatt> wow...
30 2012-09-13 00:36:48 <gmaxwell> BlueMatt: there are a fair number of miners who are totally confused by hopping.. they think that proportional is 'fair' and that anything else peanalizes them for having a lossy network connection or bad luck in part of a round.
31 2012-09-13 00:37:31 <gmaxwell> in #p2pool there is a slow but regular pace of people who are either upset about the initial payouts when they start, or convinced p2pool is flawed because they paid hours after stopping.
32 2012-09-13 00:37:40 <Luke-Jr> lol\\
33 2012-09-13 00:38:28 <gmaxwell> Esp since a lot of miners think that works accumulates to solve a block.
34 2012-09-13 00:38:42 <gmaxwell> so obviously proportional is the one true fair way.
35 2012-09-13 00:38:55 <Luke-Jr> >_<
36 2012-09-13 00:39:50 <BlueMatt> gmaxwell: wow...people's lack of understanding of basic math never fails to amaze me
37 2012-09-13 00:41:10 <gmaxwell> doesn't help that hard PPS pool often have high fees. (in no way compares to how you'll get screwed by hoppers however). I bet this actually works in pools that do both's favor, a lot of people who don't grok the math at least try to test.. perhaps they get less on proportional due to hopping and move to PPS?
38 2012-09-13 00:44:06 <sipa> what? deepbit is only 6% anymore?
39 2012-09-13 00:44:14 <sipa> (according to blockchain.info)
40 2012-09-13 00:44:28 <Luke-Jr> sipa: Deepbit blocks them to make detection harder
41 2012-09-13 00:44:59 <denisx> my pool is also hidden
42 2012-09-13 00:44:59 <sipa> ah
43 2012-09-13 00:45:56 <MC1984> is deepbit a tad sinister or what then
44 2012-09-13 00:46:13 <gmaxwell> Nah, not at all.
45 2012-09-13 00:46:30 <MC1984> oh
46 2012-09-13 00:46:40 <gmaxwell> the hidden stuff is fighting (probably ineffectively) against hoppers.
47 2012-09-13 00:47:10 <gmaxwell> but it's not like the hoppers cost a pool operator money, unless the swings are so intense that they overload when there and leave the pool stranded when gone.
48 2012-09-13 00:48:04 <Luke-Jr> ironically, a group of hoppers has been discussing setting up their own prop pool(s) to hop :p
49 2012-09-13 00:48:07 <MC1984> the whole pooled mining thing is kinda mega emergent property of bitcoin int it
50 2012-09-13 00:48:18 <denisx> Luke-Jr: lolz
51 2012-09-13 00:48:40 <sipa> Luke-Jr: ?
52 2012-09-13 00:48:43 <gmaxwell> MC1984: I suspect it wouldn't have happened if stuff like p2pool had been implemented first (even though there are some advantages from central pools)
53 2012-09-13 00:48:52 <gmaxwell> Luke-Jr: facepalm.
54 2012-09-13 00:49:01 <gmaxwell> Unless they're planning on tricking suckers to use them.
55 2012-09-13 00:49:13 <BlueMatt> hmm...has anyone spent any time with hal's faster sig checker? I implemented it in bitcoinj and must've missed something because Im getting false negatives...
56 2012-09-13 00:49:13 <Luke-Jr> gmaxwell: exactly
57 2012-09-13 00:49:16 <MC1984> yeah p2pool is taking a while to get going
58 2012-09-13 00:49:32 <sipa> BlueMatt: haven't tried it yet
59 2012-09-13 00:49:40 <MC1984> seems like the pools are gonna be around for a good while
60 2012-09-13 00:49:54 <gmaxwell> MC1984: it's harder to use. and there is a huge based of people that picked a pool once a year ago and never looked back.
61 2012-09-13 00:50:10 <MC1984> pretty much
62 2012-09-13 00:50:20 <Luke-Jr> MC1984: ASICs kill centralized pools mostly
63 2012-09-13 00:50:34 <denisx> Luke-Jr: why that?
64 2012-09-13 00:50:43 <gmaxwell> Plus its prone to misunderstanding, and it had a big chunk of poor performance though thats solidly fixed now.
65 2012-09-13 00:50:43 <MC1984> i wonder what would happen if p2pool just existed as part of the satoshi client, so would get good uptake as people upgrade
66 2012-09-13 00:50:57 <Luke-Jr> denisx: getwork can't handle the load, and the ASIC-compatible getblocktemplate only supports decentralized
67 2012-09-13 00:51:09 <Luke-Jr> MC1984: clients shouldn't favour pools
68 2012-09-13 00:51:17 <gmaxwell> Luke-Jr: I expect a lot of asic users to end up on centeralized pools; it's not like plunking down $30k on unicorn chips means you're a science wiz
69 2012-09-13 00:51:31 <MC1984> well its not really a pool
70 2012-09-13 00:51:38 <gmaxwell> MC1984: don't argue this with luke.
71 2012-09-13 00:51:49 <Luke-Jr> MC1984: yes it is
72 2012-09-13 00:51:51 <MC1984> okie dokie
73 2012-09-13 00:52:04 <Luke-Jr> MC1984: the only unique factors to p2pool are bad things :p
74 2012-09-13 00:52:12 <denisx> if the clients use rollntime and maybe diff-4 or diff-8 all should be fine
75 2012-09-13 00:52:22 <Luke-Jr> denisx: rollntime only gets about 1 TH of work
76 2012-09-13 00:52:57 <gmaxwell> denisx: doesn't matter, it's not had to make a protocol that allows local extranonce generation.
77 2012-09-13 00:53:02 <gmaxwell> s/had/hard/
78 2012-09-13 00:53:08 <BlueMatt> jgarzik: tell the unc redhat recruiters they should be doing a better job, btw
79 2012-09-13 00:54:00 <BlueMatt> ACTION would like to see jgarzik doing the recruiting
80 2012-09-13 00:54:26 <denisx> this should all be done before the asics arrive
81 2012-09-13 00:55:04 <Luke-Jr> denisx: it mostly is, I've been working on it almost constantly the past month
82 2012-09-13 00:55:33 <Luke-Jr> could use some better non-technical documentation tho
83 2012-09-13 00:55:37 <gmaxwell> denisx: getblocktemplate is partially made to facilitate that.
84 2012-09-13 00:55:39 <Luke-Jr> https://bitcointalk.org/index.php?topic=108854.0
85 2012-09-13 01:00:25 <denisx> I predict alot of pools will not implement this and maybe die
86 2012-09-13 01:00:56 <Luke-Jr> denisx: worse would be if they implement slush's centralized extranonce protocol :/
87 2012-09-13 01:02:36 <gmaxwell> it's really a question of poolserver software, I assume only the oldest pools are running straight up custom poolserver software.
88 2012-09-13 01:02:52 <sipa> well, for a traditional miner on a centralized pool, slush's protocol is probably easier and faster
89 2012-09-13 01:03:50 <Luke-Jr> sipa: for a miner, all they care about is running the software
90 2012-09-13 01:04:10 <denisx> I use a very modfified version of pushpoold
91 2012-09-13 01:04:12 <Luke-Jr> implementing StratumMP is at least more work to implement
92 2012-09-13 01:04:30 <gmaxwell> Indeed, what matters is whats in the software.. and its kinda lame if miners end up having to support four protocols.
93 2012-09-13 01:04:45 <gmaxwell> s/miners/miner software authors/
94 2012-09-13 01:05:07 <denisx> I like that right now the pool software is mostly a proxy (with stats) for bitcoind
95 2012-09-13 01:06:05 <gmaxwell> denisx: getblocktemplate can generally preserve that??? bitcoind offers getblocktemplate. (it's how many of the more modern poolservers talk to it)
96 2012-09-13 01:06:11 <Luke-Jr> denisx: yours is?
97 2012-09-13 01:06:25 <denisx> yes
98 2012-09-13 01:06:51 <Luke-Jr> gmaxwell: well, pools need to make their own coinbase at least
99 2012-09-13 01:07:10 <denisx> I always wanted to move the merkletree building to pushpoold, but never had the resources to do that
100 2012-09-13 01:07:39 <Luke-Jr> denisx: pretty sure somebody did
101 2012-09-13 01:07:41 <gmaxwell> Luke-Jr: yea, a little less of a proxy, the rest can be passthrough though.
102 2012-09-13 01:11:23 <denisx> and the asics clients will not support the old way?
103 2012-09-13 01:13:40 <sipa> who knows?
104 2012-09-13 01:15:09 <sipa> as far as i know, no ASIC miner exists :)
105 2012-09-13 01:15:13 <gmaxwell> I expect they will but it won't work well at least for the larger ones.
106 2012-09-13 01:16:49 <denisx> asics are overestimated and the diificulty will help me to prove it ;)
107 2012-09-13 01:17:15 <jgarzik> asics are hot air
108 2012-09-13 01:17:22 <Luke-Jr> ???
109 2012-09-13 01:17:46 <gmaxwell> jgarzik: Thought you bought some BFL unicorn boxes?
110 2012-09-13 01:17:58 <jgarzik> gmaxwell: yes, I've spent bitcoins on hot air
111 2012-09-13 01:18:05 <jgarzik> gmaxwell: two providers worth, even
112 2012-09-13 01:20:06 <gmaxwell> I was a bit too concerned that if their first part run is defective they'll be bankrupt. Not worth it. I mine for fun and because I like bitcoin. I can buy the next wave of parts, or this one at a deep discount once people realize the the difficulty spike screwed up their economic planning.
113 2012-09-13 01:20:48 <jgarzik> it seems doubtful that miners will make a profit, but who knows
114 2012-09-13 01:21:07 <jgarzik> ACTION just wants first run bitcoin ASICs, whoever produces them
115 2012-09-13 01:21:10 <Luke-Jr> gmaxwell: I figure as long I get mine for the first difficulty period, I'll be ok <.<
116 2012-09-13 01:21:21 <jgarzik> even [Tycho] is working on ASIC, and is selling ASIC bonds for future dev
117 2012-09-13 01:21:41 <jgarzik> icbit is a bit too daunting, so no purchases made there
118 2012-09-13 01:21:58 <jgarzik> ACTION seriously wonders if he could do a cubesat bond
119 2012-09-13 01:22:09 <sipa> cubesat?
120 2012-09-13 01:22:13 <gmaxwell> sure, though most everyone is doing cheap Taiwanese 130nm fab runs. (and charging prices that are ... very profitable against those costs)... doesn't seem super wise to me.
121 2012-09-13 01:22:21 <gmaxwell> [[Cubesat]]
122 2012-09-13 01:22:33 <gmaxwell> aw this channel doesn't have the wikipedia link deref bot.
123 2012-09-13 01:22:43 <gmaxwell> http://en.wikipedia.org/wiki/CubeSat
124 2012-09-13 01:22:53 <gmaxwell> jgarzik: it ought to have an asic miner in it of course.
125 2012-09-13 01:23:00 <copumpkin> here I was thinking it was a satisfiability algorithm
126 2012-09-13 01:23:16 <gmaxwell> and bitcoin txn relaying.
127 2012-09-13 01:23:58 <jgarzik> hehehe, I was thinking more along the lines of a boring comm sat, maybe with a camera or two
128 2012-09-13 01:24:18 <sipa> copumpkin: one that runs in O(n^3), and thereby proves P=NP ?
129 2012-09-13 01:24:22 <jgarzik> cheap off the shelf parts, space-wise, means stuff too weak to run bitcoin software
130 2012-09-13 01:24:27 <copumpkin> sipa: ZOMG
131 2012-09-13 01:24:29 <copumpkin> yesh
132 2012-09-13 01:24:38 <jgarzik> 100 Mhz processors, 128 MB RAM, etc.
133 2012-09-13 01:24:43 <copumpkin> anyway, SMT >> SAT
134 2012-09-13 01:24:48 <copumpkin> SAT is so last century
135 2012-09-13 01:24:52 <sipa> haha
136 2012-09-13 01:28:39 <denisx> hi [7]
137 2012-09-13 01:31:26 <gmaxwell> jgarzik: pshaw. thats not ambitious, a bitcoin relay would be ambitious, plus camera of course. It would have shade of "some crazy dictatorship blocks the internet, fuck you, bitcoin users launch satellite" :P (so people there can exchange coins using the computers and radios they don't have, :( )
138 2012-09-13 01:31:53 <jgarzik> hehehe, true
139 2012-09-13 01:32:05 <jgarzik> but since I really am spec'ing out a cubesat project, I have to be realistic ;p
140 2012-09-13 01:32:21 <jgarzik> was only joking about the bitcoin funding part
141 2012-09-13 01:35:38 <whitenissan> hey guys, my bitcoin client keeps crashing recently. I have uninstalled, reinstalled, deleted everything from roaming/bitwallet except for wallet.dat, i'm expecting bitcoins to come in. but it can't even stay loaded for long enough to confirm the money i have
142 2012-09-13 01:35:50 <copumpkin> what kind of a crash is it?
143 2012-09-13 01:36:13 <whitenissan> c++ runtime
144 2012-09-13 01:36:16 <jgarzik> whitenissan: can you pastebin your debug.log from the crash?
145 2012-09-13 01:36:28 <whitenissan> okay, one second
146 2012-09-13 01:36:55 <whitenissan> 000406f99c height=185043 work=358101141221929278492 ProcessBlock: ACCEPTED sending getdata: tx b8908052325825b04eaf askfor block 00000000000009f3129d 1347504130000000 askfor block 00000000000007993c0c 1347504130000000 askfor block 00000000000000f0aeb5 1347504130000000 askfor block 00000000000001d8f2cb 1347504130000000 askfor block 00000000000009aee66a 1347504130000000 askfor block 00000000000005eb087c 134750413000
147 2012-09-13 01:37:02 <denisx> whitenissan: did you try without your wallet.dat?
148 2012-09-13 01:37:34 <whitenissan> I have not done that. but would that not delete my current infomation causing me to lose my current address?
149 2012-09-13 01:38:27 <whitenissan> It says "this application has asked runtime to terminate it in an unusual way."
150 2012-09-13 01:40:16 <whitenissan> oh, there was 2 debug files
151 2012-09-13 01:40:24 <whitenissan> file wallet.dat has LSN 1228/1501276, past end of log at 332/7394824 Commonly caused by moving a database from one database environment to another without clearing the database LSNs, or by removing all of the log files from a database environment DB_ENV->log_flush: LSN of 1228/1501276 past current end-of-log of 332/7394824 Database environment corrupt; the wrong log files may have been removed or incompatible database files imp
152 2012-09-13 08:10:34 <epscy> gmaxwell: what is the point of the base64 encoded source code in this rfc? http://tools.ietf.org/html/rfc6716#appendix-A.3
153 2012-09-13 11:44:32 <sipa_> gavinandresen, wumpus: 0.7.0rc3 is a perfect match, it seems
154 2012-09-13 11:45:05 <gavinandresen> huzzah!
155 2012-09-13 11:46:24 <sipa> even with a self-built qt
156 2012-09-13 11:50:35 <gavinandresen> wumpus : did you gitian build Linux? I see your sigs in rc3-win32/laanwj but not in rc3/
157 2012-09-13 11:57:42 <BlueMatt> gavinandresen: been pretty busy lately, building rc3 now
158 2012-09-13 11:57:53 <gavinandresen> BlueMatt: thanks
159 2012-09-13 12:32:46 <helo> the "This transaction requires a transaction fee of at least..." dialog really needs to allow one to determine exactly what (and to what extent) transaction elements are at fault
160 2012-09-13 12:36:13 <helo> i guess without a way to manipulate the transaction, the info isn't very pertinent
161 2012-09-13 12:36:53 <Joric> did we really start blk0003.dat already? http://blockchain.info/charts/blocks-size shows just 2 gigs :-O
162 2012-09-13 12:39:07 <sipa> Joric: "we" certainly didn't, but if you'd for example remove blkindex.dat without removing the blk00* files, it'll redownload and append
163 2012-09-13 12:42:26 <Joric> i got blk00001 / blk00002 capped at 2000M and blk0003 (250M) are you telling me it shouldn't be so?
164 2012-09-13 12:45:11 <Joric> ok then i'll try to redownload it from sf
165 2012-09-13 12:46:40 <sipa> i have around 3GB of block data right now
166 2012-09-13 12:48:07 <sipa> Joric: and if you're using 0.7, you can just move the old blk files away, and import them back using -loadblock=filename
167 2012-09-13 12:54:33 <BlueMatt> can someone send me a copy of their qt-win32-4.7.4-gitian-r1.zip and bitcoin-qt.exe from 0.7.0rc3, mine doesnt match
168 2012-09-13 12:55:31 <sipa> uploading
169 2012-09-13 12:56:36 <sipa> none of our qt builds match, by the way
170 2012-09-13 12:56:45 <sipa> but bitcoin-qt does
171 2012-09-13 12:59:25 <sipa> BlueMatt: http://bitcoin.sipa.be/builds/0.7.0rc3/
172 2012-09-13 13:11:18 <devrandom> hi tonikt
173 2012-09-13 13:11:35 <tonikt> hi
174 2012-09-13 13:11:37 <sipa> devrandom: starting lxc-start within linux32 would help, i suppose
175 2012-09-13 13:11:46 <BlueMatt> sipa: thanks
176 2012-09-13 13:11:55 <tonikt> can I help you, devrandom?
177 2012-09-13 13:12:22 <devrandom> maybe, but I was going to help you with gitian ;)
178 2012-09-13 13:12:47 <devrandom> tonikt: on-target has a sudo for lxc-start
179 2012-09-13 13:13:01 <tonikt> :) I edited my last note in the issue - I maneged to make it work after all
180 2012-09-13 13:13:04 <devrandom> maybe that's the source of the password request?
181 2012-09-13 13:13:13 <devrandom> are you logged in as ubuntu on host?
182 2012-09-13 13:13:22 <tonikt> no - different username
183 2012-09-13 13:13:45 <tonikt> but, if you read my latest comment: I removed the folder, cloned it over again and redone - it it went fine
184 2012-09-13 13:13:53 <tonikt> so there had to be some trashy leftovers
185 2012-09-13 13:14:06 <tonikt> and I dont even have them anymore
186 2012-09-13 13:14:08 <devrandom> hm... doesn't look like github send email on edit
187 2012-09-13 13:14:26 <devrandom> okay, got it, glad it's working
188 2012-09-13 13:14:32 <tonikt> but the thing with the umount - I have the sleep 5 now...
189 2012-09-13 13:14:51 <tonikt> so if you want me to try some solution, after removing the sleep - I'd be happy to help
190 2012-09-13 13:17:20 <devrandom> ACTION is lagging
191 2012-09-13 13:17:44 <sipa> ACTION switches to devurandom
192 2012-09-13 13:17:52 <kinlo> :p
193 2012-09-13 13:28:12 <BlueMatt> hmm...so my bitcoin-qt.exe doesnt even get close to matching (and not for any obvious reason)
194 2012-09-13 13:28:32 <BlueMatt> same base image as last week, but new kvm
195 2012-09-13 13:28:49 <BlueMatt> if anyone wants a copy, Ill upload it, but I dont have time to beat on this for a while...
196 2012-09-13 13:51:08 <devrandom> tonikt: the thing to do is to find out what is the process that holds a reference there for a second or two...
197 2012-09-13 13:52:07 <devrandom> perhaps put a os.system('fuser %s/dev' % self.context.chroot_dir)
198 2012-09-13 13:52:24 <devrandom> instead of the sleep
199 2012-09-13 13:52:34 <devrandom> and os.system('ps auxww')
200 2012-09-13 13:52:38 <gmaxwell> Can someone cluestick me, wrt bitcoin-dev, I can't figure out what value matthew sees in being able to request whole levels of the hash tree from peers; and asking him is just getting me lectures on the values of hash trees.
201 2012-09-13 13:53:33 <devrandom> tonikt: actually fuser -v would do it
202 2012-09-13 13:53:37 <Diablo-D3> gmaxwell: Im not subscribed, but specific levels instead of the entire tree? I dont get it
203 2012-09-13 13:54:04 <tonikt> devrandom: I'm on it..
204 2012-09-13 13:54:15 <sipa> gmaxwell: i thought i understood, until i read your comment
205 2012-09-13 13:54:19 <devrandom> tonikt: I'm going to be away the rest of the day. will research it soon if you don't get around to it
206 2012-09-13 13:54:32 <tonikt> ok
207 2012-09-13 13:54:57 <phantomcircuit> gmaxwell, it's necessary to run a pruning client without having to process all the transactions in the tree
208 2012-09-13 13:54:57 <sipa> gmaxwell: i thought he wanted to do iterative deepening of the tree, skipping subtrees whose hash is already known
209 2012-09-13 13:55:00 <tonikt> devrandom: if I find something - how should I send you the message?
210 2012-09-13 13:55:06 <phantomcircuit> but really that's not the best way to do it i dont think
211 2012-09-13 13:55:15 <sipa> gmaxwell: but that doesn't make sense without having the txids in advance anyway
212 2012-09-13 13:55:16 <gmaxwell> If you're a full node you need all the txn, so you'd just work by grabbing the lowest level. If you're an spv node you need the branches connecting txn to the root. (vertically)
213 2012-09-13 13:55:59 <gmaxwell> phantomcircuit: no, it's not, you don't need whole levels, just the vertical trace between the txn of interest and the root.
214 2012-09-13 13:56:18 <gmaxwell> sipa: yea, and if you have all the txn you can compute the tree yourself.
215 2012-09-13 13:56:41 <sipa> gmaxwell: exactly, so no: i have no idea what it's useful for
216 2012-09-13 13:57:17 <devrandom> tonikt: on the github bug is probably most reliable
217 2012-09-13 13:57:28 <tonikt> ok
218 2012-09-13 14:00:28 <sipa> gmaxwell: i think he wants to be able to first request 8 hashes, be sure that those indeed constitute parts of the tx merkle tree, and then indivudally download the transactions leading up to those hashes
219 2012-09-13 14:01:02 <sipa> that is at least meaningful, but i don't see a use case
220 2012-09-13 14:03:13 <sipa> or maybe: in the case you want to download the different segments from different peers, make sure that the initial step (which cannot be parallellized) needs only to transfer the minimal amount of data transfer
221 2012-09-13 14:06:43 <Eliel> ACTION wonders why bitcoind and bitcoin-qt are not built on top of libcoin (or something similar) yet.
222 2012-09-13 14:07:34 <gmaxwell> sipa: hm, I didn't see how to limit the width to do that.. far cheaper to get the txids from each peers. Then build the tree. Then if it doesn't match do checking to figure out whos lying.
223 2012-09-13 14:08:20 <gmaxwell> But at least that makes sense, but it presumes you're going to fetch the whole tree... a lot more data than just fetching the leaves.
224 2012-09-13 14:09:05 <gmaxwell> not quite sure what N-way parallel fetching does for you if you first inflate the data to fetch by M. :-/
225 2012-09-13 14:09:25 <sipa> well, typical transactions are around 256 bytes something
226 2012-09-13 14:09:33 <sipa> a txid or intermediate hash is 32 bytes
227 2012-09-13 14:09:52 <sipa> so as soon as you have 8 levels, fetching the tree is as much data as fetching just the transactions
228 2012-09-13 14:10:17 <gmaxwell> a full block is something like 12 levels.
229 2012-09-13 14:15:27 <sipa> gmaxwell: not sure if his proposal requires traversing all levels though (haven't looked at it in detail)
230 2012-09-13 14:16:15 <sipa> gmaxwell: 1) ask 8 level-3 hashes 2) in 8 separate threads, ask the txids used to build the 8 level-3 hashes 3) in the same threads, fetch the transactions you don't have yet
231 2012-09-13 14:19:30 <gmaxwell> sipa: there we go, thats reasonable, and doesn't require fetching the whole tree... which is all that I was getting out of it.
232 2012-09-13 14:21:05 <sipa> that said, no idea whether that's possible with his proposal
233 2012-09-13 14:27:06 <gmaxwell> sipa: I'm pretty sure it wasn't his intent at least, as his protocol has a bunch of stuff to indicate e.g. the deepest level of the tree that a peer knows about. But who knows, there is also a bunch of text about how to handle duplicate transactions.
234 2012-09-13 14:37:03 <ciscoftw> rpc command 'getpeerinfo' returns a "startingheight" attribute... what is the significance of this? ...only one of my root seeds has a starting height of zero????
235 2012-09-13 14:38:16 <sipa> ciscoftw: how many blocks the node announced
236 2012-09-13 14:38:20 <sipa> to have
237 2012-09-13 14:38:50 <ciscoftw> why would any seed nodes NOT have the entire chain?
238 2012-09-13 14:39:23 <gmaxwell> ciscoftw: 'root seeds' ?
239 2012-09-13 14:39:38 <ciscoftw> yeah, i knew that wasnt right term :)
240 2012-09-13 14:40:14 <sipa> i don't know what you mean
241 2012-09-13 14:40:27 <sipa> seed nodes aren't special in any way
242 2012-09-13 14:40:30 <ciscoftw> one sec. tying my example oout now
243 2012-09-13 14:43:12 <ciscoftw> 'getinfo' says i have 95 'connections' (limited on my firewall to 100...) ...which are peers connected to me on port 8333 (i am seeding them with blockchain info), however i only have 8 peers via 'getpeerinfo' -from which i get my blockchain info from. question is why would any of the 8 peers that im connected to NOT have a starating height of 0???
244 2012-09-13 14:44:17 <sipa> getpeerinfo only returns 8 connections, while getinfo says you have 95?
245 2012-09-13 14:44:35 <ciscoftw> i believe this to be unreleated... but if someone could tie in an explanation of http://bitcoin.sipa.be/seeds.txt that would be helpful too
246 2012-09-13 14:44:45 <ciscoftw> yep
247 2012-09-13 14:45:00 <ciscoftw> of hte 95 connection, all but 9 are via 8333
248 2012-09-13 14:45:00 <sipa> and i'm not sure why you think getpeerinfo says you anything about which peers you download blockchain info from
249 2012-09-13 14:45:20 <sipa> it's a peer-to-peer gossip network: you get your information from whoever tells you
250 2012-09-13 14:45:49 <sipa> ciscoftw: some are inbound, others are outbound
251 2012-09-13 14:45:59 <sipa> outbound connections use a random port at your system
252 2012-09-13 14:46:21 <ciscoftw> random about 1024 :)
253 2012-09-13 14:46:50 <sipa> but inbound/outbound has no influence on how that node is dealt with
254 2012-09-13 14:46:58 <sipa> it's just who made the initial connection
255 2012-09-13 14:47:18 <ciscoftw> really!!!
256 2012-09-13 14:47:24 <ciscoftw> that just kinda blew my mind
257 2012-09-13 14:47:27 <sipa> are you being sarcastic?
258 2012-09-13 14:47:30 <ciscoftw> no
259 2012-09-13 14:47:36 <ciscoftw> that is why this was fucking with me so badly
260 2012-09-13 14:47:47 <ciscoftw> ...but why does getpeerinfo only show 8 connections?
261 2012-09-13 14:48:04 <ciscoftw> cause it just so happens that i only have 8 outbound connections to 8333
262 2012-09-13 14:48:18 <ciscoftw> which is what getpeerinfo shows
263 2012-09-13 14:48:20 <sipa> can you paste your getpeerinfo somewhere?
264 2012-09-13 14:48:27 <ciscoftw> yeah, one sec
265 2012-09-13 14:48:29 <gmaxwell> ciscoftw: are you sure it isn't just scrolling off?
266 2012-09-13 14:48:35 <ciscoftw> :(
267 2012-09-13 14:48:41 <ciscoftw> let me check that acutally
268 2012-09-13 14:48:54 <sipa> haha
269 2012-09-13 14:49:19 <ciscoftw> fuck.
270 2012-09-13 14:49:24 <gmaxwell> ciscoftw: why did you limit the connections on the firewall? Bitcoin itself will limit connections (to 124). Kinda surprising to hear you have so many in fact.
271 2012-09-13 14:49:29 <ciscoftw> was scrolling off hte buffer
272 2012-09-13 14:49:58 <ciscoftw> what is the significance of http://bitcoin.sipa.be/seeds.txt
273 2012-09-13 14:50:03 <ciscoftw> cause im ranked pretty high :)
274 2012-09-13 14:50:28 <gmaxwell> ciscoftw: that doesn't matter.
275 2012-09-13 14:50:44 <gmaxwell> (being listed and being good matters, sure)
276 2012-09-13 14:50:47 <ciscoftw> ok, i understant this, then why is sipa tracking it?
277 2012-09-13 14:51:02 <sipa> and that seeds.txt list is generated by my crawler, which is served via my dns seeder
278 2012-09-13 14:51:20 <sipa> being high on that list means my crawler thinks you're a good node
279 2012-09-13 14:51:26 <ciscoftw> dns hardcoded into how bitcoin resolved peers?
280 2012-09-13 14:51:26 <sipa> not more, not less :)
281 2012-09-13 14:51:40 <sipa> yes, there are 4 DNS seeds hardcoded
282 2012-09-13 14:51:51 <ciscoftw> isnt that a fail?
283 2012-09-13 14:52:00 <sipa> bitcoin can survive fine without them
284 2012-09-13 14:52:08 <ciscoftw> gov pwns dns all the time
285 2012-09-13 14:52:11 <sipa> but it helps finding initial nodes to connect with
286 2012-09-13 14:52:22 <ciscoftw> is that one attribute that gives NC advantage?
287 2012-09-13 14:52:30 <sipa> NC?
288 2012-09-13 14:52:34 <ciscoftw> NameCoin
289 2012-09-13 14:52:36 <ciscoftw> NMC
290 2012-09-13 14:52:42 <sipa> namecoin is no different, afaik
291 2012-09-13 14:54:11 <ciscoftw> thanx for the info sipa/gmaxwell
292 2012-09-13 14:54:38 <eian> Hypothetical quesiton about the Satoshi client: Client A is connected to Client B. A sends a getblocks/getdata but B does not respond. What is A's defined behavior? Does A sit there waiting or should it relay other messages to B while it waits?
293 2012-09-13 14:55:05 <sipa> eian: it continues to work as normal
294 2012-09-13 14:55:23 <eian> I see
295 2012-09-13 14:55:24 <eian> thanks
296 2012-09-13 14:56:04 <Joric> sipa, i downloaded blockchain from eu1, theres no blkindex.dat i don't have it either how i should rebuild it
297 2012-09-13 14:56:30 <sipa> Joric: bitcoind -loadblock=blk000file
298 2012-09-13 14:56:53 <sipa> (0.7 only)
299 2012-09-13 14:57:07 <Joric> there are two dat files will it continue on blk0002?
300 2012-09-13 14:57:22 <sipa> then specify -loadblock twice
301 2012-09-13 14:57:28 <sipa> (in the correct order)
302 2012-09-13 14:57:48 <Joric> bitcoind -loadblock=blk0001.dat -loadblock=blk0002.dat ? :D
303 2012-09-13 14:58:00 <sipa> yes, if the blk000* files are in the current directory
304 2012-09-13 14:58:07 <sipa> do NOT put them in ~/.bitcoin
305 2012-09-13 14:58:19 <Joric> ah
306 2012-09-13 15:07:10 <TD> god evening
307 2012-09-13 15:07:12 <TD> good
308 2012-09-13 15:08:43 <sipa> hi there
309 2012-09-13 15:15:44 <Joric> <Cory> Still 0.05 BTC hidden at http://bitcoinclock.com/
310 2012-09-13 15:16:37 <Joric> 'Cory was here.' hashes to an address with 0.1 BTC but they were redeemed 64 blocks ago... any clues? i know gmaxwell likes that
311 2012-09-13 15:22:31 <helo> you mean 'Cory was here.' can be used as the seed in electrum or multibit to generate the private key?
312 2012-09-13 15:24:23 <Joric> helo, sha256('Cory was here.') -> 145Sph2eiNGp5WVAkdJKg9Z2PMhTGSS9iT
313 2012-09-13 15:26:35 <helo> interesting sha256 you have there ;)
314 2012-09-13 15:27:01 <Diablo-D3> I assume thats after bitcoin address encoding
315 2012-09-13 15:27:01 <Joric> yeah i skipped a few steps
316 2012-09-13 15:27:23 <lianj> Bitcoin.pubkey_to_address(Bitcoin.regenerate_public_key(Digest::SHA256.hexdigest("Cory was here."))) #=> 145Sph2eiNGp5WVAkdJKg9Z2PMhTGSS9iT
317 2012-09-13 15:27:36 <Diablo-D3> lolperl
318 2012-09-13 15:27:51 <Diablo-D3> wait, thats not perl
319 2012-09-13 15:27:53 <Diablo-D3> wtf is that?
320 2012-09-13 15:27:59 <lianj> ruby
321 2012-09-13 15:28:04 <Diablo-D3> fucking ruby
322 2012-09-13 15:28:18 <Diablo-D3> the language written by hipsters for hipsters
323 2012-09-13 15:28:18 <Joric> zhou tong's favorite
324 2012-09-13 15:28:24 <helo> lianj: right, thanks :)
325 2012-09-13 15:28:39 <lianj> Diablo-D3: sure???
326 2012-09-13 15:28:41 <Diablo-D3> Joric: zhou tong? oh, you mean nefario, thats what hes going by now
327 2012-09-13 15:29:03 <helo> how many people are making rainbow tables using that?
328 2012-09-13 15:29:17 <helo> well i guess the right question is "is anyone of you..."
329 2012-09-13 15:29:19 <Joric> yeah he's quite nefarious
330 2012-09-13 15:29:27 <Diablo-D3> Joric: he was pirate too, I bet
331 2012-09-13 15:47:31 <OneEyed> lianj: the ruby library looks cool!
332 2012-09-13 15:52:25 <OneEyed> Does coinbase always have to be the first transaction in a block?
333 2012-09-13 15:54:13 <lianj> yes
334 2012-09-13 16:14:32 <eco_> 0.7.0rc1 means ? link ? is test?
335 2012-09-13 16:15:11 <sipa> rc = release candidate
336 2012-09-13 16:15:35 <sipa> it's a prerelease, and if no showstopper bugs are found, it gets released as 0.7.0 final
337 2012-09-13 16:15:46 <eco_> thanks
338 2012-09-13 16:15:59 <sipa> also, rc2 already exists, and rc3 is almost there
339 2012-09-13 16:17:34 <eco_> https://bitcointalk.org/index.php?topic=108782.0 open coin dot com ? related bitcoin?
340 2012-09-13 16:20:01 <sipa> read the thread; nobody knows more than what is said there
341 2012-09-13 16:29:42 <gmaxwell> I .. can .. not... believe that he responded to me explaining what merkle tress are as though I were a child.
342 2012-09-13 16:31:17 <gavinandresen> ... he even drew a nice picture...
343 2012-09-13 16:32:08 <gmaxwell> I know! ahhhh! it is nice. And the explination seems fine and good.
344 2012-09-13 16:32:16 <gmaxwell> And was entirely not what I needed. :'(
345 2012-09-13 16:32:41 <gavinandresen> I've been kinda horrified by his cbitcoin posts on the forums. If I believed in regulation, I'd make him get a Coding License before going any further....
346 2012-09-13 16:33:09 <jeremias> lol
347 2012-09-13 16:33:56 <gmaxwell> gavinandresen: I've tried to refrain from saying things like that in public. :) God knows what would be said about me if people weren't so polite. :) But I've had some real challenges communicating with him, not sure why.
348 2012-09-13 16:34:44 <gavinandresen> We can all be idiots SOMETIMES, but certain people are more consistent about it.
349 2012-09-13 16:35:37 <sipa> he sure is persistent, though
350 2012-09-13 16:35:40 <sipa> in his work
351 2012-09-13 16:37:09 <gmaxwell> Thats powerful, if usefully directed.
352 2012-09-13 16:37:20 <sipa> i'll answer
353 2012-09-13 16:46:04 <eian> who are you guys talking about?
354 2012-09-13 16:52:43 <gmaxwell> MatthewLM; wrt a thread on bitcoin-dev.
355 2012-09-13 16:53:12 <gmaxwell> He's proposed a bunch of protocol messages for doing block downloads; I think it's rather complicated, and I'm trying to figure out what its supposted to accomplish over something much simpler, and not having much luck.
356 2012-09-13 17:05:19 <eian> :P
357 2012-09-13 17:32:18 <BlueMatt> found the bug in my bitoinj impl of hal's faster secp256k1, tuns out it was a bug in bouncycastle's multiply+add method that I copied (luckily it wont effect ecdsa...)
358 2012-09-13 17:32:49 <BlueMatt> sadly, the speedup for me is essentially nothing, but it probably is the fault of the still-slow multiply-and-add method Im still using...guess that goes next
359 2012-09-13 17:37:12 <diki> BlueMatt:where is your implementation located at?
360 2012-09-13 17:37:37 <BlueMatt> its not pushed yet (and the code is about 50% comments of slower and test implementations...)
361 2012-09-13 17:49:39 <test___> hi, how can i get a bitcoin adress? using a windows client is the only way?
362 2012-09-13 17:50:43 <sipa> go to receive addresses, and click new
363 2012-09-13 17:52:44 <test___> sipa: im using diablominer, how do i get an address, not using windows client
364 2012-09-13 17:54:50 <sipa> depends what software you want to use to manage your wallet
365 2012-09-13 17:55:10 <sipa> see http://bitcoin.org/clients.html
366 2012-09-13 17:55:51 <test___> thanks a lot sipa
367 2012-09-13 18:04:26 <diki> I recently noticed that I didn't know why ever block hash had leading zeroes(big-endian). Would be nice if someone clarifies.
368 2012-09-13 18:04:53 <diki> s/ever/every
369 2012-09-13 18:05:12 <sipa> to satisfy the proof of work requirement
370 2012-09-13 18:05:27 <diki> Which means?
371 2012-09-13 18:06:06 <gmaxwell> to assemble a mighty army
372 2012-09-13 18:06:17 <sipa> it's just some criterion that is hard to satisfy, so the fact that found blocks do have leading zero bits, proves that work was spent on creating it
373 2012-09-13 18:06:47 <diki> Ok, that makes sense.
374 2012-09-13 18:08:50 <Joric> cbitcoin ftw! if it's written by linus torwalds
375 2012-09-13 18:09:02 <BlueMatt> do we currently allow using obscenely large pubkeys?
376 2012-09-13 18:09:24 <BlueMatt> sipa: as the person doing the canonical sig stuff ^
377 2012-09-13 18:09:49 <diki> Joric:I like the idea of cbitcoin, but the author seems to be inexperienced when it comes to windows, apparently. This does not mean he wont be able to add windows support.
378 2012-09-13 18:10:01 <Joric> eww camelcase in C
379 2012-09-13 18:10:05 <BlueMatt> eg ima use 10240-bit ecdsa for even better security!!!111one (and in the process DoS the network)
380 2012-09-13 18:10:06 <Joric> just checked it out
381 2012-09-13 18:10:31 <sipa> BlueMatt: i think any bytestring not (0x00, 0x02/0x03 + 32 bytes or 0x04/0x06/0x07 + 64 bytes) will be considered an invalid pubkey
382 2012-09-13 18:11:00 <BlueMatt> sipa: so we currently allow it, but its not canonical in the current impl of canonical sigchecking?
383 2012-09-13 18:11:11 <sipa> by invalid i mean openssl invalid
384 2012-09-13 18:11:18 <sipa> as CKey is hardcoded to be secp256k1
385 2012-09-13 18:11:19 <BlueMatt> oh..hmmm
386 2012-09-13 18:11:24 <BlueMatt> yea, yea
387 2012-09-13 18:11:40 <BlueMatt> dur, sorry
388 2012-09-13 18:12:34 <Joric> well not camelcase more like capitalized words but why even use C prefix if there are no classes
389 2012-09-13 18:13:20 <Joric> ah CB must be CBitcoin
390 2012-09-13 18:14:43 <Joric> 'Filenames should begin with CB. Functions, types and variables with linker visibility outside the library should begin with CB'
391 2012-09-13 18:22:12 <diki> Joric:Vanitygen for instance uses vg as a prefix for everything.
392 2012-09-13 18:23:20 <Joric> it's fine while it's lower case
393 2012-09-13 18:29:56 <sipa> i'm sure there are many reasonable criticisms on code, but if the best you've got is complaining about the aestethics of upper- vs lowercase ...
394 2012-09-13 18:43:38 <Luke-Jr> lol
395 2012-09-13 19:05:17 <MC-Eeepc> Hard drives traditionally run on air, but HGST claims that using helium, which is seven times less dense than air, will reduce power consumption by 23 percent while extending capacity by around 40 percent.
396 2012-09-13 19:05:29 <MC-Eeepc> so there is still r&d going on for spinning storage
397 2012-09-13 19:06:06 <sipa> of course, for many workloads spinning disks have far better performamce per price than SSD
398 2012-09-13 19:07:34 <fiesh> I kind of wonder about the applications of fast scsi hdds though
399 2012-09-13 19:07:57 <fiesh> who would but a spinning sas disk, and for what application?
400 2012-09-13 19:08:35 <gmaxwell> fiesh: sequential writes.
401 2012-09-13 19:09:14 <gmaxwell> DB transaction logs, for example. Archiving your spying on the entire internets, for another.
402 2012-09-13 19:10:10 <fiesh> I thought even for sequential writes, sdds were better now, but maybe I'm wrong
403 2012-09-13 19:10:29 <fiesh> and surely the write cycles aren't really an issue any more
404 2012-09-13 19:10:57 <gmaxwell> fiesh: note that sipa said price performance.
405 2012-09-13 19:11:06 <MC-Eeepc> i cant believe moores law caught up for the ability to monitor and filter and store the entire internet
406 2012-09-13 19:11:09 <gmaxwell> They're also very cheap per capacity.
407 2012-09-13 19:11:15 <fiesh> oh yes, that's why I refer to sas disks that are very expensive normally
408 2012-09-13 19:11:18 <MC-Eeepc> i suppose it was inevitable, once traffic growth slowed down
409 2012-09-13 19:11:35 <fiesh> cheapo sata 3.5" disks, obviously much better for many things
410 2012-09-13 19:11:57 <gmaxwell> MC-Eeepc: Communications speeds have always grown much slower than moores law. Plus internet spying is based on sampling and automatic targeting, so they're not constantly logging everything.
411 2012-09-13 19:12:14 <MC-Eeepc> not everything
412 2012-09-13 19:12:28 <MC-Eeepc> but that datacenter in the desert is awfuly large
413 2012-09-13 19:18:52 <freewil> i think they do log everything
414 2012-09-13 19:18:59 <freewil> then they sample from that
415 2012-09-13 23:14:22 <whitenissan> does anyone have any idea why bitcoins that i know have been sent have not been recieved in 8 hours?
416 2012-09-13 23:14:56 <gmaxwell> whitenissan: how are you defining recieved? and how many confirmations do you see for the tx?
417 2012-09-13 23:16:11 <whitenissan> well what happened is my bitcoin wallet kept having problems over and over and over. eventually the only way to fix it was to delete wallet.dat and get bitcoins sent to the new address.
418 2012-09-13 23:16:27 <whitenissan> I have ordered bitcoins off the site multiple times, with no problems
419 2012-09-13 23:17:48 <gmaxwell> whitenissan: er. if you're having to delete your wallet then something is ver wrong.
420 2012-09-13 23:17:55 <gmaxwell> er very
421 2012-09-13 23:19:01 <whitenissan> well it was crashing. then I deleted everything but wallet.dat because that's the only fix i could find on the forums, so after that it worked but before it even confirmed my bitcoins it would crash
422 2012-09-13 23:19:32 <whitenissan> so I came on here, and someone gave me the great idea to delete wallet.dat saying i wouldnt lose my bitcoins...
423 2012-09-13 23:19:41 <midnightmagic> lol troll?
424 2012-09-13 23:19:59 <whitenissan> already out 400 dollars there, now another 300 for the ones i just bought.
425 2012-09-13 23:23:29 <whitenissan> I feel like 90% of the people in this chat shouldnt be, and the ones that should be are AFK so they can't help anyone.
426 2012-09-13 23:23:47 <gmaxwell> I'm currently searching my logs to see what happened before.
427 2012-09-13 23:24:15 <midnightmagic> wauugh brutal denisx didn't clarify after spoonfeeding you
428 2012-09-13 23:25:25 <gmaxwell> whitenissan: jesus. okay. your previously deleted bitcoins may not be unrecoverable.
429 2012-09-13 23:25:27 <midnightmagic> whitenissan: Sorry for the troll comment. There are a lot of them that come in here causing trouble with problems similar to yours. Please accept my apologies.
430 2012-09-13 23:25:53 <midnightmagic> "here" being IRC in general.
431 2012-09-13 23:25:57 <whitenissan> i'm more worried about the ones i just ordered
432 2012-09-13 23:26:12 <gmaxwell> whitenissan: how many blocks does your client say it has?
433 2012-09-13 23:26:43 <whitenissan> currently says 98k remaining
434 2012-09-13 23:26:51 <gmaxwell> oh well thats why you don't see them. Expected.
435 2012-09-13 23:27:06 <whitenissan> but usually when it has low blocks the bitcoins would say unconfirmed, am i wrong?
436 2012-09-13 23:27:12 <gmaxwell> Now before you do anything else, shut down, and make a backup of wallet.dat to external media.
437 2012-09-13 23:27:31 <gmaxwell> whitenissan: no it won't show it at all if the blocks the coin is based on have not been accepted yet.
438 2012-09-13 23:27:47 <gmaxwell> whitenissan: and then you should run this tool: https://bitcointalk.org/index.php?topic=25091.0
439 2012-09-13 23:28:05 <BlueMatt> arg...my ecdsa impl may be 5x faster than bouncycastle, but its still 5x slower than hal's openssl-based ecdsa and 4x slower than stock openssl...
440 2012-09-13 23:28:19 <gmaxwell> which will make a wallet file with everything on your harddrive that looks like a bitcoin private key; which may recover what you lost.
441 2012-09-13 23:28:54 <BlueMatt> either I'm missing something obvious, or the openjdk jit isnt as good as it should be...
442 2012-09-13 23:29:49 <gmaxwell> whitenissan: what denisx _meant_ was that he was asking if you've tried moving the wallet.dat out of the way. E.g. not deleting it. I'm sorry I wasn't around at that second.. usually when anyone mentions doing anything with anything the advice is backup backup backup before anything.
443 2012-09-13 23:30:11 <midnightmagic> whitenissan: Your wallet.dat contains your keys: bitcoin uses that information to determine whether bitcoins are yours or not. Therefore, your wallet is the most important file you have. Don't delete it. And don't listen to anyone who seems to be telling you to delete it.
444 2012-09-13 23:31:03 <gmaxwell> the message you were getting was because your wallet.dat was corrupted, probably as a result of deleting the database/ directory while they wallet was either actively being written or immediatly after bitcoin crashed while writing to the wallet.
445 2012-09-13 23:31:08 <whitenissan> I have absolutely no idea why i never checked this before, but i found it in the recycle bin. so i made a backup of the one I got the bitcoins sent to and then restored it to check. but it crashed in the same way
446 2012-09-13 23:31:14 <gmaxwell> But that kind of corruption isn't irrepariable.
447 2012-09-13 23:31:42 <denisx> I never said deleting
448 2012-09-13 23:31:43 <gmaxwell> whitenissan: thats okay, the tool in that thread can extract the keys out of it and make a new wallet.dat which is not corrupted and has the same keys.
449 2012-09-13 23:32:06 <midnightmagic> denisx: I know you didn't. But you spoonfed him and didn't follow-up. :(
450 2012-09-13 23:32:20 <midnightmagic> denisx: Honest mistake IMO.
451 2012-09-13 23:32:48 <whitenissan> but yes, that must have been what happened. If i use that tool will it keeps the address i just recently got coins sent to?
452 2012-09-13 23:33:06 <denisx> midnightmagic: I said "did you try without wallet.dat" he did not answer that and then left the channel
453 2012-09-13 23:33:11 <gmaxwell> denisx: you didn't; indeed. But surely you can see how someone not knoweldgable might have done something daft... I'm not faulting you. Bad communication happens.
454 2012-09-13 23:33:52 <midnightmagic> denisx: 30 seconds after you said that, he asked, "I have not done that. but would that not delete my current infomation causing me to lose my current address?"
455 2012-09-13 23:34:21 <midnightmagic> it *appears* he inferred you meant him to delete his wallet.dat file.
456 2012-09-13 23:34:34 <denisx> ops, that was not my intention
457 2012-09-13 23:34:36 <whitenissan> let's not cause an argument here, I misunderstood.
458 2012-09-13 23:34:37 <gmaxwell> whitenissan: if you use that tool it will create a new wallet.dat file with the content of all wallets it finds.
459 2012-09-13 23:34:40 <denisx> I'm sorry
460 2012-09-13 23:34:50 <gmaxwell> (though it takes a long time to sweep your whole disk.
461 2012-09-13 23:34:51 <gmaxwell> )
462 2012-09-13 23:35:13 <gmaxwell> if you run it just on a single wallet file it will make a new wallet file with just the stuff in that single input.
463 2012-09-13 23:35:30 <gmaxwell> You can switch between multiple wallets by shutting down and swapping out the files. Just take care while doing it.
464 2012-09-13 23:35:34 <whitenissan> alright, thanks for all the help
465 2012-09-13 23:35:35 <gmaxwell> And as always backup backup backup.
466 2012-09-13 23:35:53 <midnightmagic> whitenissan: Good luck. We would like to hear whether you managed to successfully recover any keys.
467 2012-09-13 23:36:12 <gmaxwell> make backups on a usb key of all wallets before doing anything tricky. make periodic backups. as you go. Good luck. If you get stuck please feel free to ask more questions.
468 2012-09-13 23:36:53 <gmaxwell> never overwrite a backup copy just make more backups. :)
469 2012-09-13 23:37:57 <whitenissan> anyone know if that tool can be run on DSL (damn small linux)
470 2012-09-13 23:38:09 <midnightmagic> whitenissan: It's obvious it was just a simple misunderstanding, I don't mean to imply I blame anyone. This sort of thing happens all the time, and I sympathize a lot with you and denisx both.
471 2012-09-13 23:38:12 <gmaxwell> whitenissan: the binary is static, should work on any linux.
472 2012-09-13 23:38:33 <whitenissan> alright thanks for all the help
473 2012-09-13 23:40:21 <denisx> but maybe his wallet.dat is already corrupt, because it did not work with it yesterday before he deleted it
474 2012-09-13 23:41:01 <midnightmagic> gah poor guy
475 2012-09-13 23:41:11 <gmaxwell> denisx: what I expect happened was that he probably had addr.dat corruption after an unclean shutdown, which is a frequent problem in versions with addr.dat.
476 2012-09-13 23:41:34 <gmaxwell> And that made his client not start. Also his wallet was in an unclean state due to the unclean shutdown.
477 2012-09-13 23:41:44 <gmaxwell> Forum advice was to delete everything except the wallet.
478 2012-09-13 23:42:01 <gmaxwell> But then the software was unable to read the wallet because the db logs had been deleted and the wallet was unclean.
479 2012-09-13 23:42:23 <gmaxwell> Fortunately, that wallet recovery tool will rescue the keys from a wallet in that state (so long as they're not encrypted).
480 2012-09-13 23:42:40 <gmaxwell> There are some bdb tools that would probably fix it too.
481 2012-09-13 23:43:41 <midnightmagic> db_dump | db_recover might do it: what are the chances that it was failing on a keygen
482 2012-09-13 23:44:02 <gmaxwell> midnightmagic: db_dump will probably refuse to read it too.. though there might be some arguments to make it go.
483 2012-09-13 23:45:09 <gmaxwell> but the recovery tool is pretty good.. it just doesn't worry about reading the database.. it matches bytes looking for things that look like the seralized private keys.
484 2012-09-13 23:45:14 <MC-Eeepc> rc3 changelog?
485 2012-09-13 23:45:34 <midnightmagic> db_dump is way more robust than I thought it was. -R and -r are the beasties to pull data.
486 2012-09-13 23:45:40 <gmaxwell> MC-Eeepc: posted in the thread, kinda, but ... use the git luke. :)
487 2012-09-13 23:45:51 <midnightmagic> yeah the recovery tool is probably way mor thorough.
488 2012-09-13 23:46:09 <gmaxwell> midnightmagic: it actually found some coins on my system that I didn't know I'd lost. :-/
489 2012-09-13 23:46:23 <denisx> did he say something about the amount of bitcoins in his wallet?
490 2012-09-13 23:46:23 <MC-Eeepc> can i have a quick link to the git page please
491 2012-09-13 23:46:29 <gmaxwell> and also about 5k LTC.
492 2012-09-13 23:46:32 <midnightmagic> awesome. I'm always paranoid about that, which is why I so sorely miss dumpwallet and importwallet.
493 2012-09-13 23:46:49 <midnightmagic> denisx: A few hundred dollars, and then a few hundred more.
494 2012-09-13 23:47:01 <midnightmagic> whatever that means.
495 2012-09-13 23:47:06 <gmaxwell> denisx: well just now he said he lost $400. But I expect he'll recover them all.
496 2012-09-13 23:47:33 <gmaxwell> (he found the fine he 'deleted' in the recycle bin)
497 2012-09-13 23:50:25 <gmaxwell> midnightmagic: I think we just need to make a habbit of hitting every single person in here that asks a question with "Have you made a wallet.dat backup lately?" even if they're asking just how signmessage works.
498 2012-09-13 23:52:05 <midnightmagic> I hadn't made a backup in weeks, even post wallet upgrade, until last night. I should probably be automating it or something.
499 2012-09-13 23:52:27 <MC-Eeepc> why doesnt the client bput a backup away itself periodically
500 2012-09-13 23:52:41 <MC-Eeepc> i mean it took microsoft 10 years to learn that lesson with office.........
501 2012-09-13 23:52:43 <midnightmagic> MC-Eeepc: Because then it's doing something surprising with a sensitive file.
502 2012-09-13 23:52:59 <midnightmagic> and deleting wallet.dat might not be deleting it, secure wipe or not.
503 2012-09-13 23:53:08 <MC-Eeepc> more surprising than "surprise you just lost $400"
504 2012-09-13 23:54:05 <MC-Eeepc> couldnt it just dump wallet.dat.backup in the very same folder
505 2012-09-13 23:54:11 <MC-Eeepc> just incase the real one gets munged
506 2012-09-13 23:54:26 <midnightmagic> wallet.dat.1 wallet.dat.2 wallet.dat.3 wallet.dat.4
507 2012-09-13 23:54:39 <midnightmagic> MC-Eeepc: Create a pull request. :)
508 2012-09-13 23:55:21 <MC-Eeepc> a pull request is something im only used to doing in bars and clubs
509 2012-09-13 23:57:04 <MC-Eeepc> implying i regularly go to bars and clubs anymore, who am i kidding