1 2012-12-09 00:00:04 <sipa> jgarzik: hmm, i wonder why no OpenSSL code allocates BIGNUMs on the stack
2 2012-12-09 00:00:59 <jgarzik> sipa: looking at ecdsa_do_verify(), it looks like allocs might be grouped
3 2012-12-09 00:01:07 <jgarzik> whereas Hal just does BN_new()
4 2012-12-09 00:01:25 <gmaxwell> well for BN_new it can get resized, no?
5 2012-12-09 00:02:34 <jgarzik> Confirmed. Internally, OpenSSL calls BN_CTX_get() which provides a pool of temporaries.
6 2012-12-09 00:02:35 <jgarzik> Confirmed. Internally, OpenSSL calls BN_CTX_get() which provides a pool of temporaries.
7 2012-12-09 00:02:51 <jgarzik> All freed with BN_CTX_end(ctx);
8 2012-12-09 00:03:03 <sipa> jgarzik: see "my" Hal code
9 2012-12-09 00:03:07 <sipa> that's the first thing i changed
10 2012-12-09 00:03:08 <sipa> that's the first thing i changed
11 2012-12-09 00:03:13 <jgarzik> ok :)
12 2012-12-09 00:03:41 <sipa> I also precalculate Glam, bnn, bnp, ... now
13 2012-12-09 00:03:42 <sipa> I also precalculate Glam, bnn, bnp, ... now
14 2012-12-09 00:04:37 <sipa> btw, benchmark of rebuilding 193k->211k with -par=6: 14:40 without halcode, 12:00 with halcode
15 2012-12-09 00:04:38 <sipa> btw, benchmark of rebuilding 193k->211k with -par=6: 14:40 without halcode, 12:00 with halcode
16 2012-12-09 00:06:00 <jgarzik> hmmm
17 2012-12-09 00:06:04 <slush> is there any graph displaying blocks v1 versus v2?
18 2012-12-09 00:06:11 <slush> well - there is one, but I lost URL
19 2012-12-09 00:06:17 <jgarzik> ACTION is tempted to turn on CPU mining for mainnet. Seriously. :)
20 2012-12-09 00:06:31 <jgarzik> Would be a useful ultraprune test
21 2012-12-09 00:07:12 <sipa> jgarzik: i've already mined a *block* with ultraprune :)
22 2012-12-09 00:07:15 <jgarzik> Exercise more code, even if the likelihood of solving a block is infinitesimal
23 2012-12-09 00:07:27 <sipa> (p2pool, but still)
24 2012-12-09 00:07:31 <jgarzik> heh
25 2012-12-09 00:07:32 <jgarzik> heh
26 2012-12-09 00:08:18 <gmaxwell> I've had cpu mining enabled w/ ultraprune on mainnet at one point, though not at the moment. Obviously I'm mining with it on testnet.
27 2012-12-09 00:08:19 <gmaxwell> I've had cpu mining enabled w/ ultraprune on mainnet at one point, though not at the moment. Obviously I'm mining with it on testnet.
28 2012-12-09 00:09:16 <jgarzik> stored orphan tx 9066f46261 (mapsz 10000)
29 2012-12-09 00:09:55 <jgarzik> lotsa orphans, compared to 0.7.1:
30 2012-12-09 00:09:57 <jgarzik> stored orphan tx 8b3cd52c7b (mapsz 103)
31 2012-12-09 00:10:02 <jgarzik> uptimes vaguely similar
32 2012-12-09 00:10:39 <slush> ok, here it is http://blockorigin.pfoe.be/top.php
33 2012-12-09 00:10:56 <sipa> jgarzik: yes, rebroadcasts of fully spent transactions end up as orphans in HEAD
34 2012-12-09 00:11:02 <gmaxwell> jgarzik: this is because recently mined and consumed txn that get rebroadcast (why?) are 'orphans'.
35 2012-12-09 00:11:03 <gmaxwell> jgarzik: this is because recently mined and consumed txn that get rebroadcast (why?) are 'orphans'.
36 2012-12-09 00:11:46 <jgarzik> sipa: boy it just shit itself
37 2012-12-09 00:11:48 <jgarzik> [jgarzik@us2 ~]$ grep -c 'ERROR: mempool transaction missing input' /spare/bitcoin/data/debug.log
38 2012-12-09 00:11:49 <jgarzik> 6216
39 2012-12-09 00:12:02 <jgarzik> in less than 60 seconds, since CPU mining enabled
40 2012-12-09 00:12:41 <gmaxwell> what version are you running?
41 2012-12-09 00:13:04 <gmaxwell> (bugs that could result in crud in the mempool were fixed recently)
42 2012-12-09 00:13:05 <gmaxwell> (bugs that could result in crud in the mempool were fixed recently)
43 2012-12-09 00:13:14 <jgarzik> gmaxwell: HEAD, as of a few minutes ago
44 2012-12-09 00:13:18 <jgarzik> CreateNewBlock(): total size 38755
45 2012-12-09 00:13:26 <jgarzik> love the compare/contrast :)
46 2012-12-09 00:13:27 <jgarzik> love the compare/contrast :)
47 2012-12-09 00:13:33 <jgarzik> 122 vs. 3814 :)
48 2012-12-09 00:13:37 <jgarzik> *112
49 2012-12-09 00:15:29 <sipa> jgarzik: sure those mempool tx missing input error are from recent HEAD?
50 2012-12-09 00:15:59 <jgarzik> sipa: hmmm
51 2012-12-09 00:16:21 <jgarzik> From git://github.com/bitcoin/bitcoin
52 2012-12-09 00:16:27 <jgarzik> perhaps it was older than I thought
53 2012-12-09 00:16:57 <jgarzik> sipa: circa Nov 27
54 2012-12-09 00:19:16 <sipa> 5c37be2 didn't include that fix
55 2012-12-09 00:22:45 <jgarzik> sipa: sorry for the noise. too many public nodes. the $other one is running HEAD. just updated that one to HEAD as well.
56 2012-12-09 00:24:05 <sipa> np
57 2012-12-09 00:26:58 <sipa> slush, kinlo: interesting site
58 2012-12-09 00:31:26 <kinlo> sipa: ?
59 2012-12-09 00:32:14 <kinlo> oh right
60 2012-12-09 00:32:19 <kinlo> :)
61 2012-12-09 00:40:33 <sipa> reddit down?
62 2012-12-09 00:41:19 <sipa> wait... my DNS is down
63 2012-12-09 01:40:49 <Luke-Jr> I guess we're teaching people English with the GitHub Issue tracker now???
64 2012-12-09 01:44:36 <sipa> Luke-Jr: post a list to translate.google.com ?
65 2012-12-09 01:44:41 <sipa> *link
66 2012-12-09 01:46:44 <Luke-Jr> sipa: tempting
67 2012-12-09 02:18:08 <SomeoneWeird> is there a good C++ btc lib?
68 2012-12-09 02:23:50 <vazakl> libbitcoin
69 2012-12-09 02:38:13 <SomeoneWeird> isn't libbitcoin py?
70 2012-12-09 02:38:18 <SomeoneWeird> http://bitcoinmedia.com/libbitcoin-first-steps/
71 2012-12-09 02:38:41 <SomeoneWeird> at least, it definitely LOOKS like python :p
72 2012-12-09 02:44:02 <jgarzik> If you want a python lib, you want /bitcoin/ of https://github.com/jgarzik/pynode/
73 2012-12-09 02:44:16 <jgarzik> (for the public record; I know SomeoneWeird asked about C++)
74 2012-12-09 02:44:44 <SomeoneWeird> :)
75 2012-12-09 02:44:45 <jgarzik> SomeoneWeird: for C++ I thought there was libcoin and libbitcoin
76 2012-12-09 02:44:53 <jgarzik> where libbitcoin is -not- python
77 2012-12-09 02:45:02 <SomeoneWeird> hrm
78 2012-12-09 02:45:11 <SomeoneWeird> it says it has a C++ api yet I can't find in docs for it
79 2012-12-09 02:45:49 <jgarzik> SomeoneWeird: https://github.com/libcoin/libcoin (NOTE: _not_ a recommendation, just a link; I think the author is MIA)
80 2012-12-09 02:46:08 <jgarzik> and you can always use my C lib </plug>
81 2012-12-09 02:46:52 <SomeoneWeird> link to that? :)
82 2012-12-09 02:47:40 <jgarzik> SomeoneWeird: https://github.com/jgarzik/picocoin/
83 2012-12-09 03:00:15 <vazakl> picocoin? looks cool
84 2012-12-09 03:00:21 <vazakl> haven't heard about that till now
85 2012-12-09 03:36:39 <etotheipi_> anyone else find this tx concerning? http://blockchain.info/tx/9106ac6859097079d39127aaac86208ac2a2b9bade92c3ae109192b7bc340872
86 2012-12-09 03:37:02 <etotheipi_> the guy is doing an excellent job spamming...
87 2012-12-09 03:37:38 <etotheipi_> and for some reason wants to link all those addresses to an inappropriate "Public Note"
88 2012-12-09 03:39:50 <gmaxwell> etotheipi_: sounds like something to take to #blockchain.info, not #bitcoin-dev. :P
89 2012-12-09 03:41:07 <etotheipi_> just making others aware, Bitcoin Foundation and Bitcointalk Forums are on that list, too
90 2012-12-09 03:43:44 <noagendamarket> lol
91 2012-12-09 03:49:51 <Luke-Jr> etotheipi_: so blockchain.info should just delete the note
92 2012-12-09 04:38:59 <bcb> how can I verify a message in a text file with a bitcoin address?
93 2012-12-09 04:40:14 <bcb> bitcoind <btcaddress> <sig> message.txt
94 2012-12-09 04:41:35 <gmaxwell> bitcoind <btcaddress> <sig> `cat message.txt`
95 2012-12-09 04:43:41 <bcb> gmaxwell negative
96 2012-12-09 04:44:50 <bcb> "message.txt"
97 2012-12-09 04:44:54 <bcb> 'message.txt'
98 2012-12-09 04:45:07 <bcb> * 'cat message.txt'
99 2012-12-09 04:45:13 <bcb> * "cat message.txt"
100 2012-12-09 04:47:41 <gmaxwell> ` is not '
101 2012-12-09 04:47:48 <gmaxwell> and are you on windows? that won't work on windows.
102 2012-12-09 04:47:58 <gmaxwell> on windows you're out of luck. put the content on the commandline.
103 2012-12-09 04:48:35 <Luke-Jr> or use the GUI???
104 2012-12-09 04:49:13 <bcb> linux
105 2012-12-09 04:49:29 <gmaxwell> Luke-Jr: the gui doesn't let you use a file does it?
106 2012-12-09 04:49:40 <gmaxwell> bcb: then what I said, exactly, will work.
107 2012-12-09 04:49:50 <Luke-Jr> gmaxwell: no, but it lets you reasonably put a lot of text content in the verification
108 2012-12-09 04:49:54 <gmaxwell> but what I said with random changes, will not.
109 2012-12-09 04:50:37 <Luke-Jr> gmaxwell: does POSIX `` properly quote? :o
110 2012-12-09 04:50:51 <bcb> `cat message.txt` returns error: {"code":-1,"message":"verifymessage <bitcoinaddress> <signature> <message>\\nVerify a signed message"}
111 2012-12-09 04:51:20 <Luke-Jr> bcb: try: bitcoind <bcaddress> "<sig>" "$(<message.txt)"
112 2012-12-09 04:51:58 <bcb> luke-jr worked - thx
113 2012-12-09 04:52:22 <Luke-Jr> ACTION ponders which part made the difference :P
114 2012-12-09 04:52:52 <Luke-Jr> gmaxwell: hah, we both missed the verifymessage command bit :P
115 2012-12-09 04:53:31 <bcb> no i used verifymessage
116 2012-12-09 04:53:39 <bcb> `message.txt`
117 2012-12-09 04:53:42 <bcb> didn't work ??
118 2012-12-09 04:56:05 <Luke-Jr> bcb: with `` you need the cat command
119 2012-12-09 04:56:26 <bcb> `cat message.txt` didn't work
120 2012-12-09 04:57:40 <gmaxwell> bcb: does your message have spaces in it?
121 2012-12-09 04:57:55 <bcb> yes
122 2012-12-09 04:58:18 <bcb> and line breaks
123 2012-12-09 04:59:18 <gmaxwell> may need to do something like
124 2012-12-09 04:59:28 <gmaxwell> \\"`cat message.txt`\\"
125 2012-12-09 05:01:42 <bcb> nope
126 2012-12-09 05:01:52 <bcb> np the other way worked - just curious
127 2012-12-09 05:02:00 <bcb> could I cat to an env?
128 2012-12-09 05:02:28 <SomeoneWeird> ?
129 2012-12-09 05:03:37 <bcb> someoneweird you get my PM
130 2012-12-09 05:33:17 <jgarzik> evorhees, on reddit: "Further, SD increased the minimum bets to .01 btc, so it's not like "dust" or spam transactions are even occurring at all." http://www.reddit.com/r/Bitcoin/comments/14i9e7/your_stance_toward_satoshidice/
131 2012-12-09 05:33:28 <jgarzik> That seems to conflict with the claim that losing bets get sent dust.
132 2012-12-09 05:43:46 <jgarzik> ACTION trolls sipa and gmaxwell for upvotes ;p http://www.reddit.com/r/Bitcoin/comments/14i9e7/your_stance_toward_satoshidice/c7dkrmh
133 2012-12-09 06:03:25 <bcb> how do I open a wallet.dat file on v0.7.1-beta on windows
134 2012-12-09 06:03:36 <bcb> a different wallet.dat
135 2012-12-09 06:03:52 <sipa> for now: put
136 2012-12-09 06:04:13 <sipa> it in the place of your existing one
137 2012-12-09 06:04:26 <bcb> yes
138 2012-12-09 06:04:31 <sipa> jgarzik: willdo; can you do the same?
139 2012-12-09 06:04:32 <sipa> jgarzik: willdo; can you do the same?
140 2012-12-09 06:06:22 <jgarzik> done, I just see the one from you
141 2012-12-09 06:06:57 <bcb> how did I locate wich file the client is currently accessing
142 2012-12-09 06:07:01 <bcb> on windoze
143 2012-12-09 06:07:56 <MC1984> process explorer
144 2012-12-09 06:10:21 <bcb> * which wallet.dat file the client is accessing
145 2012-12-09 06:10:49 <sipa> ;;google bitcoin data directory
146 2012-12-09 06:10:50 <gribble> Data directory - Bitcoin: <https://en.bitcoin.it/wiki/Data_directory>; Securing your wallet - Bitcoin: <https://en.bitcoin.it/wiki/Securing_your_wallet>; Running Bitcoin - Bitcoin: <https://en.bitcoin.it/wiki/Running_Bitcoin>
147 2012-12-09 06:11:07 <sipa> bcb: ^
148 2012-12-09 06:12:24 <bcb> sipa ty
149 2012-12-09 06:13:03 <jgarzik> sipa: hah. -benchmark is turning into an important psychological hack ;p "How fast does my hardware go, on this software?" is a question geeks love to answer. https://bitcointalk.org/index.php?topic=129861.0
150 2012-12-09 06:13:24 <jgarzik> myself included... I need to try that branch ;p
151 2012-12-09 06:19:41 <jgarzik> chain-verf: spend block @ 160000
152 2012-12-09 06:19:42 <jgarzik> chain-verf: spend block @ 160000
153 2012-12-09 06:20:03 <denisx> no freebsd or osx build for me ;(
154 2012-12-09 10:17:14 <denisx> jgarzik: you started cpu miner nov 2010, what was used before? what did satoshi use?
155 2012-12-09 10:24:35 <slush1> denisx: satoshi client had cpu miner inside
156 2012-12-09 10:25:42 <denisx> slush1: oh yeah, I forgot
157 2012-12-09 10:30:56 <denisx> snowing here in berlin
158 2012-12-09 10:30:57 <denisx> snowing here in berlin
159 2012-12-09 11:06:40 <sipa> denisx: that built-in miner even still exists, and it is far faster than the version in satoshi's original code
160 2012-12-09 11:07:26 <denisx> sipa: yes, I simply forgot it
161 2012-12-09 11:07:26 <sipa> TD: Hal confirmed he indeed already did the optimization
162 2012-12-09 11:07:33 <TD> ah
163 2012-12-09 11:08:06 <sipa> jgarzik: none of the reported speeds even include -benchmark output
164 2012-12-09 14:17:22 <zeks2> Can someone help me to setup bitcoin mining pool?:) I can pay if needed?
165 2012-12-09 15:00:24 <etotheipi_> gmaxwell, sipa: what's the chance that in a hard fork we could update the TxIn serialization/format to include the 8-byte input values?
166 2012-12-09 15:04:28 <etotheipi_> ever since I had that one person send me a 90kB tx that had 4 MB of supporting tx...and trying to help out the HW wallet guys, it's unnecessarily bloat-y to not have txIns standalone
167 2012-12-09 15:07:35 <sipa> etotheipi_: what's the point of having the value if you don't have the txout?
168 2012-12-09 15:07:43 <sipa> *txout script
169 2012-12-09 15:10:02 <etotheipi_> sipa: for offline verification
170 2012-12-09 15:10:19 <sipa> and more generally... the reason to have supporting transactions is so they can be relayed to be included in blocks
171 2012-12-09 15:10:28 <sipa> nothing except the full transactions will do
172 2012-12-09 15:10:41 <etotheipi_> if the value is part of the serialization, the offline system can confirm everything it needs to know that the tx is either exactly what it thinks it is, or invalid
173 2012-12-09 15:10:46 <Diablo-D3> >90kb tx
174 2012-12-09 15:10:47 <etotheipi_> but not malicious
175 2012-12-09 15:10:50 <Diablo-D3> >4mb supporting tx
176 2012-12-09 15:10:56 <Diablo-D3> it wasnt me was it?
177 2012-12-09 15:11:05 <etotheipi_> Diablo-D3: it was 483 inputs
178 2012-12-09 15:11:12 <Diablo-D3> miiiiight have been me
179 2012-12-09 15:11:21 <etotheipi_> I recommended the guy try to split it up, but it was still a pain
180 2012-12-09 15:11:21 <sipa> etotheipi_:but you can claim any value there
181 2012-12-09 15:11:22 <Diablo-D3> some asshole sent me 10btc in 0.001 increments
182 2012-12-09 15:11:24 <Diablo-D3> with fees off.
183 2012-12-09 15:11:38 <sipa> etotheipi_: if you're offline there is no way to verify those values match the true values of those txouts
184 2012-12-09 15:11:39 <Diablo-D3> it took the network 28 hours to include them all in blocks
185 2012-12-09 15:11:41 <etotheipi_> sipa: if it's part of the serialization
186 2012-12-09 15:12:00 <etotheipi_> then the tx will be invalid if the wrong value is signed
187 2012-12-09 15:12:01 <Diablo-D3> for months I had to pay fees out my ass
188 2012-12-09 15:12:19 <Diablo-D3> because it basically posioned my entire wallet with useless ammounts
189 2012-12-09 15:12:25 <sipa> etotheipi_: yes, so? you can only verify that if you have a UTXO set anyway
190 2012-12-09 15:12:33 <sipa> etotheipi_: with or without amounts in the txin
191 2012-12-09 15:13:01 <etotheipi_> sipa: in this situation the online computer makes the tx
192 2012-12-09 15:13:12 <etotheipi_> saves it to a USB key and takes it to the offline computer
193 2012-12-09 15:13:36 <etotheipi_> the offline computer parses it and says "this transaction is to send 33 BTC to address X. Is this correct?"
194 2012-12-09 15:13:46 <Diablo-D3> etotheipi_: I dont see the point
195 2012-12-09 15:14:03 <etotheipi_> the problem with sending just the tx itself, is there's no way for the offline system to know how much fee is part of the tx
196 2012-12-09 15:14:12 <etotheipi_> that's the only thing it can't confirm
197 2012-12-09 15:14:12 <sipa> etotheipi_: to sign it, you need the txout scripts too, no?
198 2012-12-09 15:14:16 <sipa> not just the amounts
199 2012-12-09 15:14:28 <etotheipi_> sipa: the txout scripts were already injected into the tx by the online computer
200 2012-12-09 15:14:38 <etotheipi_> the tx is exactly as it needs to be for signing
201 2012-12-09 15:14:43 <sipa> so why not the amounts?
202 2012-12-09 15:14:56 <etotheipi_> it's just a matter of the offline computer not being able to be tricked into signing something else
203 2012-12-09 15:15:05 <etotheipi_> becuase the txout scripts are part of the signature
204 2012-12-09 15:15:15 <etotheipi_> if the wrong txout scipt is supplied, the resulting tx is invalid
205 2012-12-09 15:15:20 <sipa> ah
206 2012-12-09 15:15:31 <sipa> so what you want is the signaturehash to include the txout value?