1 2012-05-15 00:10:19 <rcorreia> In test/rpc_tests.cpp:
  2 2012-05-15 00:10:31 <rcorreia> 18     Array addresses;
  3 2012-05-15 00:10:46 <rcorreia> should the last line be this instead:
  4 2012-05-15 00:10:48 <rcorreia> 20     if (address2) addresses.push_back(address2);
  5 2012-05-15 00:10:50 <rcorreia> ?
  6 2012-05-15 00:27:22 <osmosis_> gmaxwell, i have a use case for my suggestion around locking an address to not be spent.
  7 2012-05-15 00:27:54 <osmosis_> gmaxwell, im holding some coins for someone, and I point them to the blockchain with the address showing that the coins exist. I sign a msg to prove that I am holding that address.
  8 2012-05-15 00:28:46 <osmosis_> as long as the coins stay in that address, my associate knows the coins are still there. In order to do this currently requires me having multiple wallet.dat files and moving files around manually. No way to do this within the client.
  9 2012-05-15 00:39:46 <luke-jr> do P2SH scriptCheck sigops count toward the limit of the redeeming block?
 10 2012-05-15 00:40:39 <luke-jr> they do. crap.
 11 2012-05-15 00:42:57 <luke-jr> guess I just found a p2pool vuln then
 12 2012-05-15 00:43:53 <luke-jr> a minor one, since it depends on txn spam
 13 2012-05-15 00:49:04 <forrestv> luke-jr, hurry up and get that getmemorypool revision finished <.<
 14 2012-05-15 00:49:53 <gmaxwell> osmosis_: thanks.  Thats a good way to put it.
 15 2012-05-15 00:50:05 <gmaxwell> I have varrious public adresses I've segregated for that reason.
 16 2012-05-15 00:50:17 <BlueMatt> is 0.6.2 mac-only, or is it everything now?
 17 2012-05-15 00:57:42 <Diablo-D3> NASA astronauts launching from Cosmodrome at 11pm EDT http://identi.ca/url/69744824
 18 2012-05-15 00:58:09 <BlueMatt> gmaxwell: whats up with 0.6.2?
 19 2012-05-15 00:58:58 <gmaxwell> BlueMatt: ... What do you mean whats up with it? It's out. It's being widely adopted. We shiped a critcial security advisory urging people to upgrade right away today.
 20 2012-05-15 00:59:14 <luke-jr> forrestv: it's mostly stable now I think?
 21 2012-05-15 00:59:24 <gmaxwell> BlueMatt: http://bitcoin.org/dos.html
 22 2012-05-15 00:59:26 <forrestv> ah
 23 2012-05-15 00:59:35 <BlueMatt> oh, sorry I was thinking 0.6.2.2...
 24 2012-05-15 00:59:43 <BlueMatt> nvm...
 25 2012-05-15 00:59:48 <BlueMatt> 0.6.2.2 is osx-only, though?
 26 2012-05-15 01:00:23 <gmaxwell> BlueMatt: yes 0.6.2.2 is OSX only.
 27 2012-05-15 01:00:35 <BlueMatt> ok, just wanted to make sure
 28 2012-05-15 01:00:44 <gmaxwell> And basically replaces 0.6.2 for OSX  (fixes the 0.6.1 busy wait regression)
 29 2012-05-15 01:00:49 <BlueMatt> yea
 30 2012-05-15 01:01:01 <BlueMatt> just wanted to make sure I wasnt missing a ppa upgrade
 31 2012-05-15 01:01:04 <gmaxwell> (apparently freebsd was also hit by that but we dont ship freebsd binaries)
 32 2012-05-15 01:01:07 <gmaxwell> ::nods::
 33 2012-05-15 01:01:56 <Diablo-D3> lift off
 34 2012-05-15 02:42:28 <gmaxwell> http://arstechnica.com/uncategorized/2012/05/bitcoins-worth-87000-plundered/  links to http://eprint.iacr.org/2012/248.pdf , claiming " although a recently published research paper reports limitations that allow the same bitcoins can be double spent"
 35 2012-05-15 02:42:54 <gmaxwell> and the 'paper', I guess a preprint bogusly claims that the bitcoin developers recomment people accept unconfirmed transactions.
 36 2012-05-15 02:43:15 <gmaxwell> s/recomment/recommend/
 37 2012-05-15 02:43:38 <gmaxwell> We should probably publish a statement correcting this factual inaccuracy.
 38 2012-05-15 02:43:58 <BlueMatt> and probably send ars a correction
 39 2012-05-15 02:44:42 <BlueMatt> that paper looks kinda cool, though, because they actually test the listening period stuff
 40 2012-05-15 02:44:46 <BlueMatt> (or so it looks)
 41 2012-05-15 02:44:54 <gmaxwell> Indeed, but they're exagerating the importance of it.
 42 2012-05-15 02:45:19 <BlueMatt> fair enough, I was just skimming
 43 2012-05-15 02:45:41 <gmaxwell> Bitcoin developers claim in [6] that
 44 2012-05-15 02:45:42 <gmaxwell> the cost of mounting double-spending attacks for
 45 2012-05-15 02:45:44 <gmaxwell> spending attacks.
 46 2012-05-15 02:45:52 <gmaxwell> [6] is just a generic link to the FAQ page on the wiki.
 47 2012-05-15 02:45:58 <BlueMatt> heh
 48 2012-05-15 02:46:04 <BlueMatt> so send the authors an email
 49 2012-05-15 02:46:53 <luke-jr> well, it's true that "accept zero confirms for cheap stuff" is the answer usually given to "but confirmations are slow"
 50 2012-05-15 02:47:50 <BlueMatt> do they do a listening period peered with miners?
 51 2012-05-15 02:48:16 <BlueMatt> ^ that is the usual response, not just lisening
 52 2012-05-15 02:48:25 <BlueMatt> (or, at least, the one I always give)
 53 2012-05-15 02:49:09 <gmaxwell> The software itself only shows transactions confirmed at six blocks.
 54 2012-05-15 02:49:40 <gmaxwell> luke-jr: I grepped my logs in #bitcoin and can find many people being told to wait 2-6 transactions. (Actually this just happed a few minutes ago before I saw that paper)
 55 2012-05-15 02:50:07 <BlueMatt> anyway, someone should send the authors of that paper an email, getting someone to write a paper on the security of peering with miners would be nice
 56 2012-05-15 02:50:22 <luke-jr> gmaxwell: often, 2-6 confirms is too long
 57 2012-05-15 02:50:29 <luke-jr> people want their coffee *now*
 58 2012-05-15 02:51:05 <BlueMatt> thats why you do a risk analysis and use a 3-rd party tx insurance provider who peers with miners
 59 2012-05-15 02:51:19 <BlueMatt> and/or has substantial mining capabilities themselves
 60 2012-05-15 02:51:38 <gmaxwell> The initial version of the response in the FAQ was rounded than the one authors of the paper would have seen: https://en.bitcoin.it/w/index.php?title=FAQ&diff=next&oldid=7094
 61 2012-05-15 02:52:11 <gmaxwell> and geesh someone needs to show them how to cite wikis.
 62 2012-05-15 02:52:50 <gmaxwell> luke-jr: we should _never_ advise people to accept zero confirms... gpumax has made finnyattacks cheap and easy.
 63 2012-05-15 02:53:05 <gmaxwell> (well accept from people they have normal recourse against, of course)
 64 2012-05-15 02:53:36 <luke-jr> gmaxwell: sometimes it's 1 BTC at 0 confirms, or no BTC *shrug*
 65 2012-05-15 02:53:56 <gmaxwell> Here is the version the authors of the paper would have read: https://en.bitcoin.it/w/index.php?title=FAQ&oldid=25163#Do_you_have_to_wait_10_minutes_in_order_to_buy_or_sell_things_with_Bitcoin.3F
 66 2012-05-15 02:54:01 <gmaxwell> which is indeed crap.
 67 2012-05-15 02:54:01 <luke-jr> that's one reason to broadcast double-spends
 68 2012-05-15 02:54:13 <gmaxwell> It is but that doesn't help you with finney attacks.
 69 2012-05-15 02:54:51 <luke-jr> well, even with GPUMAX, don't you still need to spend about 100 BTC on 2 blocks?
 70 2012-05-15 02:55:21 <gmaxwell> luke-jr: 50 BTC on one block and you keep it! thats the thing about a finny attack, no reversal is required.
 71 2012-05-15 02:55:54 <gmaxwell> I mine while sitting at your order page. I solve a block sending my payment back to me, quickly I finish the order sending a conflicting payment to you.. then I announce the block.
 72 2012-05-15 02:56:08 <gmaxwell> The only think I lose is the small increase in orphan risk from the delay.
 73 2012-05-15 02:56:19 <gmaxwell> Which I can make ~0 if you process my payment fast and I automate it.
 74 2012-05-15 02:56:53 <gmaxwell> and no amount of doublespend announcing helps because it's already too late by the time the doublespend is visible.
 75 2012-05-15 02:57:15 <TuxBlackEdo> i have a question
 76 2012-05-15 02:58:23 <TuxBlackEdo> there is more then one nounce that could be valid for the same block (which includes the same txns), right?
 77 2012-05-15 02:58:50 <TuxBlackEdo> nonce*
 78 2012-05-15 03:02:06 <gmaxwell> sometimes.
 79 2012-05-15 03:02:17 <gmaxwell> usually there are zero valid nonces.
 80 2012-05-15 03:03:20 <TuxBlackEdo> but it would never get stuck because there are incoming transactions
 81 2012-05-15 03:03:48 <TuxBlackEdo> what if there are no new transactions, does the timestamp help against getting stuck with no solution?
 82 2012-05-15 03:10:37 <gmaxwell> TuxBlackEdo: yes, but even more than that there is a second nonce inside the coinbase... the software increments that when the nonce space is exausted and the time isn't new yet.
 83 2012-05-15 03:11:03 <TuxBlackEdo> ah ok thanks
 84 2012-05-15 03:14:29 <gmaxwell> luke-jr: do you know of any sites that accept zero confirmation txn?  Perhaps I should perform a finney attack (with consent) to make a demonstration video .. because apparently people are not well informed.
 85 2012-05-15 04:45:23 <gribble> New news from bitcoinrss: jgarzik opened issue 1312 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1312>
 86 2012-05-15 06:24:43 <gmaxwell> FWIW, blockchain.info now appears to be evading blocks of their collector IPs.
 87 2012-05-15 06:24:57 <gmaxwell> (since it's reporting txn relayed through my nodes that block their /24)
 88 2012-05-15 06:32:12 <Graet> blockchain.info properly broken now
 89 2012-05-15 06:32:50 <t7> why ?
 90 2012-05-15 06:33:07 <Graet> dont know, nice cats tho :/
 91 2012-05-15 06:33:14 <t7> bitonica ?
 92 2012-05-15 06:33:35 <Graet> Error 500
 93 2012-05-15 06:33:43 <t7> when did they become master miners
 94 2012-05-15 06:33:55 <Graet> who?
 95 2012-05-15 06:34:01 <t7> bitonica
 96 2012-05-15 06:34:06 <Graet> <Graet> blockchain.info properly broken now
 97 2012-05-15 06:34:07 <Joric> so it has come to this
 98 2012-05-15 06:34:07 <t7> is that a split
 99 2012-05-15 06:34:17 <Graet> no the sites screwed
100 2012-05-15 06:34:39 <Graet> it happens regularly . and it isnt definitive
101 2012-05-15 06:34:52 <t7> CHAIN SPLIT!
102 2012-05-15 06:34:55 <t7> PANIC!
103 2012-05-15 06:35:01 <Joric> blockchain info should be checked for 'expect mass leak soon'
104 2012-05-15 06:35:05 <Graet> http://blockorigin.pfoe.be/top.php this is more accurate
105 2012-05-15 06:35:55 <t7> 'mining team reddit'.... kill me now
106 2012-05-15 06:36:17 <Joric> t7, u mostly mined there
107 2012-05-15 06:36:22 <Joric> *me
108 2012-05-15 06:36:37 <Joric> what's wrong with mining team reddi
109 2012-05-15 06:37:49 <t7> i imagine its full of overly politicly correct wankers
110 2012-05-15 06:56:35 <kinlo> can anyone tell me what the difference is betwen version 0.6.2 and 0.6.2.2?
111 2012-05-15 06:57:06 <Diablo-D3> some mac related fixes
112 2012-05-15 07:06:21 <kinlo> yeah, can't find a changelog
113 2012-05-15 07:08:01 <Diablo-D3> its build problems
114 2012-05-15 07:08:06 <Diablo-D3> nothing relevant
115 2012-05-15 07:14:01 <kinlo> ic, thanks
116 2012-05-15 07:26:35 <sipa> kinlo: a 100% cpu issue because we used a locking primitive that was implemented as a spinlock on OSX
117 2012-05-15 07:27:47 <kinlo> ic
118 2012-05-15 07:27:50 <kinlo> thanks
119 2012-05-15 07:41:55 <davout> hey
120 2012-05-15 07:42:29 <davout> is it just me or doesn't listtransactions allow to query for all transactions no matter the account when specifying a tx count ?
121 2012-05-15 07:42:45 <davout> listtransactions [account] [count=10] [from=0]
122 2012-05-15 07:42:52 <davout> Returns up to [count] most recent transactions skipping the first [from] transactions for account [account].
123 2012-05-15 07:43:18 <davout> i want all the txes in the wallet, so i want to set count to some really high value
124 2012-05-15 07:43:35 <davout> but then the account parameter demands a value, "" doesn't work :/
125 2012-05-15 07:46:32 <sipa> try "*"
126 2012-05-15 07:47:01 <davout> already tried, I get "type mismatch"
127 2012-05-15 07:47:35 <davout> same with just *
128 2012-05-15 07:47:56 <davout> also tried ""*"" no success either :/
129 2012-05-15 07:48:00 <sipa> hmm
130 2012-05-15 07:50:52 <gribble> New news from bitcoinrss: davout opened issue 1313 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1313>
131 2012-05-15 07:51:06 <davout> gribble: that's right bitch
132 2012-05-15 07:51:09 <davout> :D
133 2012-05-15 07:51:56 <davout> sipa: thank you very much for confirming i'm not crazy and i didn't overlook something obvious
134 2012-05-15 07:52:28 <sipa> i'll have a look at the code in a minjte
135 2012-05-15 07:52:31 <sipa> minute
136 2012-05-15 07:57:33 <davout> sipa: sweet!
137 2012-05-15 08:00:51 <sipa> davout: can't test here with a wallet, but ./bitcoind listtransactions "*" 10
138 2012-05-15 08:00:54 <sipa> works
139 2012-05-15 08:00:59 <sipa> (though the result is empty)
140 2012-05-15 08:01:04 <davout> oh yes
141 2012-05-15 08:01:19 <davout> that's not really what i'm expecting :)
142 2012-05-15 08:01:43 <sipa> but you said you get a type error?
143 2012-05-15 08:01:54 <davout> oh
144 2012-05-15 08:01:57 <davout> on some
145 2012-05-15 08:01:58 <davout> hang on
146 2012-05-15 08:02:28 <davout> * -> type mismatch
147 2012-05-15 08:02:43 <davout> '*' -> type mismatch
148 2012-05-15 08:02:56 <davout> "*" -> type mismatch
149 2012-05-15 08:03:07 <sipa> can you give the full command line?
150 2012-05-15 08:03:13 <davout> yeah
151 2012-05-15 08:03:23 <drizztbsd> davout: * is expanded by the shell, you have to quote it :P
152 2012-05-15 08:03:33 <davout> $ ./bitcoin listtransactions "" 100
153 2012-05-15 08:03:45 <davout> and this, returns transactions belonging to the 100 account
154 2012-05-15 08:04:06 <davout> this is getting more and more confusing
155 2012-05-15 08:04:26 <davout> i'm relaunching Instawallet's client, I'll confirm the exact client version
156 2012-05-15 08:05:02 <sipa> davout: it should return the 100 first results of the "" account
157 2012-05-15 08:05:09 <davout> yes
158 2012-05-15 08:05:39 <sipa> is that wrong?
159 2012-05-15 08:06:28 <davout> i tried that and it returned transactions belonging to the "100" account :D
160 2012-05-15 08:06:44 <davout> (that was on bitcoin-central's client)
161 2012-05-15 08:06:52 <davout> it's a slightly patched client
162 2012-05-15 08:07:00 <davout> adds monitortx calls
163 2012-05-15 08:07:10 <davout> and monitorblock
164 2012-05-15 08:07:11 <sipa> i see no way how that could happen, looking at the code
165 2012-05-15 08:07:29 <davout> i'll run a couple more tests against my instance
166 2012-05-15 08:07:47 <davout> so on instawallets client i get version 60099
167 2012-05-15 08:08:26 <davout> based on master from like a month ago
168 2012-05-15 08:08:42 <davout> $ ./bitcoin listtransactions "" 100
169 2012-05-15 08:08:53 <davout> this yields an empty JSON array
170 2012-05-15 08:10:50 <sipa> well, do you have transactions in the "" account?
171 2012-05-15 08:11:06 <davout> no
172 2012-05-15 08:11:12 <davout> i don't use accounts at all
173 2012-05-15 08:11:20 <davout> i assumed "" was the default account
174 2012-05-15 08:11:55 <sipa> yes, it is
175 2012-05-15 08:12:09 <davout> and it's instawallet's bitcoin client
176 2012-05-15 08:12:14 <sipa> and just listtransactions works?
177 2012-05-15 08:12:21 <davout> so it has like thousands of txes
178 2012-05-15 08:12:23 <davout> yes
179 2012-05-15 08:12:35 <davout> it does, but i can only fetch 10, the default
180 2012-05-15 08:12:44 <sipa> and ./bitcoind listtransactions "*" 100
181 2012-05-15 08:13:04 <davout> type mismatch
182 2012-05-15 08:13:10 <davout> oh wait a second
183 2012-05-15 08:13:37 <davout> the "bitcoin" in my command is a bash script
184 2012-05-15 08:13:49 <davout> which contains apart form the shebang a single line
185 2012-05-15 08:14:10 <davout> "/home/bitcoin/instawallet/bitcoind -conf=/home/bitcoin/instawallet/bitcoin.conf -datadir=/home/bitcoin/instawallet/data"
186 2012-05-15 08:14:19 <davout> without the quotes
187 2012-05-15 08:14:29 <sipa> and it doesn't pass the arguments down? :s
188 2012-05-15 08:14:46 <davout> "/home/bitcoin/instawallet/bitcoind -conf=/home/bitcoin/instawallet/bitcoin.conf -datadir=/home/bitcoin/instawallet/data $@"
189 2012-05-15 08:14:52 <davout> sorry missed the last part
190 2012-05-15 08:14:59 <sipa> you need quotes around $@
191 2012-05-15 08:15:05 <davout> do i?
192 2012-05-15 08:15:08 <sipa> otherwise it gets interpreted by the shell
193 2012-05-15 08:15:22 <sipa> and you'll get exactly what you observed
194 2012-05-15 08:15:39 <sipa> "" becomes no-argument-at-all
195 2012-05-15 08:15:46 <davout> oh
196 2012-05-15 08:15:50 <davout> dang
197 2012-05-15 08:15:50 <sipa> and * becomes a list of files in the current directory
198 2012-05-15 08:16:07 <davout> i feel so stupid now :/
199 2012-05-15 08:16:14 <sipa> also, that -conf is not necessary
200 2012-05-15 08:16:22 <davout> thank you so much
201 2012-05-15 08:16:23 <sipa> it picks the config file in the datadir by default
202 2012-05-15 08:17:53 <davout> thanks for the tips, and sorry for the loss of time
203 2012-05-15 09:26:03 <kinlo> does blockchain.info has a testnet version?
204 2012-05-15 09:35:36 <freewil> kinlo: not sure, but block explorer does http://blockexplorer.com/testnet
205 2012-05-15 09:39:18 <kinlo> I know about blockexplorer, but blockchain gives more info
206 2012-05-15 09:40:17 <davout> libbitcoin isn't used at all in the satoshi client right ?
207 2012-05-15 09:40:35 <davout> they're like orthogonal if i understand correctly
208 2012-05-15 09:40:39 <freewil> correct that is a separate project
209 2012-05-15 09:41:39 <freewil> i think the main idea with libbitcoin is to separate out and decouple the logic so it can be used more easily in custom projects
210 2012-05-15 09:41:52 <freewil> everything is so tightly coupled with the satoshi client
211 2012-05-15 09:42:13 <sipa> davout: libbitcoin is not a year old
212 2012-05-15 09:42:25 <davout> sipa: i know :)
213 2012-05-15 09:42:58 <davout> if i understand correctly the awesome thing with it is that it's fully asynchronous
214 2012-05-15 09:44:24 <sipa> i know genjix paid a lot of attention to that, yes
215 2012-05-15 09:44:46 <sipa> bluematt's cblockstore will also make the satoshi code more asynchronous, though
216 2012-05-15 09:59:03 <gribble> New news from bitcoinrss: fanquake opened pull request 36 on bitcoin/bitcoin.org <https://github.com/bitcoin/bitcoin.org/pull/36>
217 2012-05-15 12:01:41 <sipa> ;;bc,tslb
218 2012-05-15 12:01:44 <gribble> Time since last block: 5 minutes and 31 seconds
219 2012-05-15 12:06:53 <jgarzik_> wow, sure am seeing a lot of orphan TX's right now
220 2012-05-15 12:11:58 <sipa> i'm trying to construct a stuck blockchain
221 2012-05-15 12:12:28 <sipa> though i should probably have done so on testnet, as it may not require me to mine a stale block...
222 2012-05-15 12:12:34 <sipa> *now
223 2012-05-15 12:13:41 <helo> fuzzing?
224 2012-05-15 12:14:03 <sipa> nah, added a rule that made one particular block invalid at connecting
225 2012-05-15 12:14:13 <sipa> then redownloaded the chain
226 2012-05-15 12:27:30 <sipa> seems i've actually constructed a stuck chain
227 2012-05-15 12:27:42 <sipa> "blocks" : 178000,
228 2012-05-15 12:27:46 <sipa> "errors" : "WARNING: Displayed transactions may not be correct!  You may need to upgrade, or other nodes may need to upgrade."
229 2012-05-15 12:32:25 <sipa> ;;bc,tslb
230 2012-05-15 12:32:26 <gribble> Time since last block: 9 minutes and 13 seconds
231 2012-05-15 12:39:39 <gmaxwell> hm. I tried manually rejecting a block then redownloaded and it didn't stick for me.
232 2012-05-15 12:43:31 <sipa> gmaxwell: see my "forkmode" branch
233 2012-05-15 12:48:02 <sipa> let's see if it sticks when a new block is found
234 2012-05-15 12:51:43 <jgarzik> sipa: is there any problem with using boost::mutex directly in bitcoin code?  there is a case where we might need it but CRITICAL_SECTION isn't yet up, it sounds like (#917)
235 2012-05-15 12:52:31 <sipa> jgarzik: why isn't it up yet?
236 2012-05-15 12:52:55 <sipa> hmm, i thought util.h included sync.h
237 2012-05-15 12:54:51 <sipa> ;;bc,tblb 20m
238 2012-05-15 12:54:52 <gribble> 1 hour, 10 minutes, and 21 seconds
239 2012-05-15 12:56:12 <sipa> jgarzik: using boost::mutex shouldn't be a problem, in particular when you know for sure it will not cause deadlocks
240 2012-05-15 12:56:17 <gavinandresen> I just sent alerts on the dos issue and updated https://en.bitcoin.it/wiki/Alerts
241 2012-05-15 12:56:25 <gmaxwell> gavinandresen: thanks!
242 2012-05-15 12:56:33 <sipa> ah, great
243 2012-05-15 12:56:41 <sipa> c'mon you block
244 2012-05-15 12:57:12 <gavinandresen> (actually the 0.6.0 alert will go out in about three minutes....)
245 2012-05-15 13:00:55 <luke-jr> hmm, gavinandresen's alert is falsely bothering me
246 2012-05-15 13:01:06 <sipa> same
247 2012-05-15 13:01:07 <luke-jr> as in, flooding me
248 2012-05-15 13:01:17 <luke-jr> I have 5 alerts popping up
249 2012-05-15 13:01:23 <luke-jr> and they keep coming as fast as they go away
250 2012-05-15 13:01:28 <sipa> 05/15/12 15:00:30 accepted alert 1017, AppliesToMe()=1
251 2012-05-15 13:01:29 <gavinandresen> wha???
252 2012-05-15 13:01:39 <gavinandresen> I'll cancel that one now....
253 2012-05-15 13:01:40 <luke-jr> gavinandresen: better cancel that one
254 2012-05-15 13:01:53 <kinlo> why is the alert actually sent out to 0.6.2?
255 2012-05-15 13:02:06 <luke-jr> also, note 0.6.0.7+ isn't affected
256 2012-05-15 13:02:30 <luke-jr> (myself and a few other pools are using this)
257 2012-05-15 13:02:37 <luke-jr> s/myself/Eligius
258 2012-05-15 13:02:45 <gavinandresen> yeah, I'm not going to write special code to deal with 0.6.0.7
259 2012-05-15 13:03:09 <gribble> New news from bitcoinrss: Diapolo opened pull request 1314 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1314>
260 2012-05-15 13:03:28 <kinlo> any reason why you still use 0.6.0?
261 2012-05-15 13:03:29 <luke-jr> as long as it doesn't trigger safemode or crap, I don't care
262 2012-05-15 13:03:41 <luke-jr> kinlo: because I haven't taken the time to rebase Eligius's codebase on 0.6.2
263 2012-05-15 13:03:46 <kinlo> ic
264 2012-05-15 13:03:48 <luke-jr> easier to just pull 0.6.0.x fixes
265 2012-05-15 13:04:03 <sipa> ;;bc,tblb 40m
266 2012-05-15 13:04:05 <gribble> 7 hours, 50 minutes, and 13 seconds
267 2012-05-15 13:04:07 <luke-jr> plus, 0.6.2 just came out. I'd rather wait until it's confirmed working sanely elsewhere :p
268 2012-05-15 13:04:46 <kinlo> I've deliberately chosen not to make too many patches so I could upgrade easily
269 2012-05-15 13:05:11 <luke-jr> kinlo: well, I'm instead working hard to get my patches suitable for mainline
270 2012-05-15 13:05:24 <luke-jr> maybe 0.7 will actually be usable OOTB :p
271 2012-05-15 13:05:35 <sipa> out of the blue?
272 2012-05-15 13:05:38 <luke-jr> out of the bxo
273 2012-05-15 13:05:40 <kinlo> out of the box :p
274 2012-05-15 13:05:40 <luke-jr> box*
275 2012-05-15 13:05:40 <sipa> ah
276 2012-05-15 13:06:00 <luke-jr> (ok, not likely since I still need to patch in fee policies etc, but still can make my changes minimal)
277 2012-05-15 13:06:18 <kinlo> wasn't there a possibility to send the alerts to specific versions?
278 2012-05-15 13:06:38 <sipa> kinlo: yes, and i'm sure gavin used that
279 2012-05-15 13:06:39 <luke-jr> yes, gavin just doesn't want to do it for 0.6.0.7 it seems
280 2012-05-15 13:06:49 <gavinandresen> yes, but separating the PROTOCOL_VERSION from the CLIENT_VERSION seems to have broken something
281 2012-05-15 13:06:49 <kinlo> sipa: I'm getting the alert on my 0.6.2
282 2012-05-15 13:06:54 <luke-jr> (and why it's hitting 0.6.2 is unknown)
283 2012-05-15 13:07:27 <kinlo> in any case, I can override my code to ignore the message, no big deal
284 2012-05-15 13:07:30 <sipa> i get it on git master
285 2012-05-15 13:07:36 <sipa> kinlo: -nosafemode
286 2012-05-15 13:07:57 <kinlo> is safe mode enforced?
287 2012-05-15 13:08:23 <sipa> actually, not sure
288 2012-05-15 13:08:39 <kinlo> safe mode only disables wallet features, right?
289 2012-05-15 13:08:48 <sipa> yes
290 2012-05-15 13:08:55 <kinlo> as long as submitblock (getmemorypool with data) keeps working, I think I'm safe
291 2012-05-15 13:08:58 <kinlo> lemme check in code
292 2012-05-15 13:11:06 <drizztbsd> 0.6.2 too
293 2012-05-15 13:11:33 <drizztbsd> can you remove that alert? or should I start bitcoin-qt with -nosafemode?
294 2012-05-15 13:11:46 <gavinandresen> I'm working on it....
295 2012-05-15 13:12:58 <Joric> look
296 2012-05-15 13:13:07 <Joric> the url
297 2012-05-15 13:13:08 <Joric> http://www.pcworld.idg.com.au/article/424589/hackers_break_into_bitcoin_exchange_site_bitcoinica_steal_90_000_bitcoins/
298 2012-05-15 13:13:15 <drizztbsd> so?
299 2012-05-15 13:13:16 <Joric> steal 90000 bitcoins
300 2012-05-15 13:13:36 <jeremias> wrong number
301 2012-05-15 13:13:39 <Joric> "It's more serious than we thought," said Bitcoinica founder Zhou Tong
302 2012-05-15 13:13:40 <jeremias> over 9000
303 2012-05-15 13:13:49 <drizztbsd> do you trust pcworld :P
304 2012-05-15 13:13:49 <Joric> sure it is )
305 2012-05-15 13:14:06 <jeremias> "It's more serious than we thought", said 17-year-old Bitcoin CEO Zhou Tong
306 2012-05-15 13:14:17 <jeremias> "And now we are closing all bitcoin operations worldwide"
307 2012-05-15 13:14:36 <drizztbsd> 17 yo?
308 2012-05-15 13:14:37 <drizztbsd> :P
309 2012-05-15 13:14:48 <jeremias> well, now he is probably 19
310 2012-05-15 13:14:58 <jeremias> he was 17 year old when he started bitcoinica
311 2012-05-15 13:15:03 <drizztbsd> cool
312 2012-05-15 13:15:31 <Joric> "Bitcoin CEO (Zhou Tong Jr) to cease all bitcoin operations from this day to forever"
313 2012-05-15 13:15:44 <drizztbsd> LOL bitcoin ceo
314 2012-05-15 13:15:54 <Joric> the company just wasn't profitable, he said
315 2012-05-15 13:16:02 <jeremias> it would be cool to get that kind of news article to pcworld
316 2012-05-15 13:16:03 <sipa> gavinandresen: you'd need nVersion=60001 and setSubVer={"/Satoshi:0.6.0/", ... }
317 2012-05-15 13:16:09 <jeremias> maybe the onion?
318 2012-05-15 13:16:34 <luke-jr> getting alerts again
319 2012-05-15 13:16:38 <luke-jr> just 1 this time it looks like
320 2012-05-15 13:16:46 <gavinandresen> thanks sipa
321 2012-05-15 13:16:58 <luke-jr> ironically, it was from my git master client, NOT my 0.6.0.7 client
322 2012-05-15 13:17:05 <sipa> gavinandresen: or look at http://bitcoin.sipa.be/seeds.txt
323 2012-05-15 13:18:06 <gavinandresen> sipa: I was trying to figure out how to cancel a subset...
324 2012-05-15 13:18:25 <sipa> you can only cancel up to a previus alert number, no?
325 2012-05-15 13:18:55 <gavinandresen> there's a setCancel
326 2012-05-15 13:19:18 <sipa> ha, nice
327 2012-05-15 13:23:02 <Joric> bitcoinica's written entirely on ruby, right? could anyone tell is it an appropriate language for such things? i barely know what's ruby
328 2012-05-15 13:23:27 <luke-jr> Joric: I hate Ruby.
329 2012-05-15 13:23:44 <luke-jr> bitcoin.org is also Ruby.
330 2012-05-15 13:24:06 <sipa> i thought it was html ;(
331 2012-05-15 13:24:26 <Joric> yeah it's static maybe preprocessed a bit
332 2012-05-15 13:24:35 <sipa> Joric: yes, i know :)
333 2012-05-15 13:26:17 <sipa> ;;bc,tblb 1h2m
334 2012-05-15 13:26:18 <gribble> Error: '1h2m' is not a valid argument.
335 2012-05-15 13:26:20 <sipa> ;;bc,tblb 1h 2m
336 2012-05-15 13:26:22 <gribble> 2 days, 15 hours, 19 minutes, and 41 seconds
337 2012-05-15 13:26:23 <Joric> ruby looks like python but without forced indentation it may be autoformatted and shit
338 2012-05-15 13:27:47 <drizztbsd> https://gist.github.com/2702648 less lines to do anything
339 2012-05-15 13:30:53 <gavinandresen> 0.6.2 shouldn't be showing the alert any more, but I think I accidentally cancelled the alerts for older versions, too...
340 2012-05-15 13:31:09 <sipa> gmaxwell: my chain is definitely stuck
341 2012-05-15 13:31:39 <gavinandresen> sipa: where is that patch?  I've got a stuck chain, too (one I didn't try to make stuck)
342 2012-05-15 13:31:59 <sipa> gavinandresen: it's called "unstuck"
343 2012-05-15 13:32:48 <sipa> https://github.com/sipa/bitcoin/commit/c31b6427a67b20b1a52e0c998a3578e94e535370
344 2012-05-15 13:35:24 <luke-jr> confirmed no alerts visible here
345 2012-05-15 13:35:25 <sipa> seems to work!
346 2012-05-15 13:42:11 <etotheipi_> I assume it's been mentioned, but just in case:  I see  "Bitcoin version vCLIENT_VERSION_MAJOR.CLIENT_VERSION_MINOR.CL"
347 2012-05-15 13:42:20 <sipa> etotheipi_: fixed already
348 2012-05-15 13:44:15 <gribble> New news from bitcoinrss: sipa opened pull request 1315 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1315>
349 2012-05-15 13:52:33 <gmaxwell> sipa: whats with leaving in this commented out code? +//            pfrom->PushGetBlocks(pindexBest, uint256(0));
350 2012-05-15 13:52:45 <sipa> uh
351 2012-05-15 13:53:23 <sipa> right, that was for testing
352 2012-05-15 13:53:44 <gmaxwell> ... yea. :)
353 2012-05-15 13:58:24 <jgarzik> gmaxwell: I'm curious about your thoughts on DeathAndTaxes' analysis in the zero confirmation thread you started.  Are those figures vaguely accurate?
354 2012-05-15 13:59:03 <jgarzik> sipa: the comment message is a bit vague...  that one-line summary will be much less meaningful in 3-6 months
355 2012-05-15 13:59:18 <sipa> ok i'll try to elaborate
356 2012-05-15 14:02:03 <gmaxwell> jgarzik: his analysis was exactly what I was thinking, but the actual invalidation risk is a little harder to figure out just because propagation isn't instant.
357 2012-05-15 14:02:58 <gmaxwell> jgarzik: And of course, because we don't currently propagate double spends, or even notify the user if we happen to hear about them, that kind of delay doesn't help you against non-finney attacks.
358 2012-05-15 14:04:56 <gmaxwell> The finny attack can also be enhanced by DOS attacking pools or just by buying up more hash power (thus depriving the legit network of it) and using it to continue your finney chain. (Though I admit that starts looking like a majority attack)
359 2012-05-15 14:06:47 <jgarzik> gmaxwell: I'm interesting in the topics, because I've been designing a file storage site.  Among other things, random people on the internet will be able to send bitcoins to an address, to add "credits" to a file being hosted, thereby keeping that file alive longer.  once credits run out, the file is no longer available.
360 2012-05-15 14:07:02 <jgarzik> gmaxwell: 0-conf will be likely
361 2012-05-15 14:07:28 <jgarzik> gmaxwell: but since it's a digital download, the cost of a double-spend seems low, even if successful
362 2012-05-15 14:07:38 <gmaxwell> You can always price in your 'shrinkage'
363 2012-05-15 14:07:43 <jgarzik> yep
364 2012-05-15 14:08:17 <gmaxwell> But there is a difference between thinking its secure against reversal and knowing that the loss is bounded and you can easily eat it. :)
365 2012-05-15 14:08:19 <luke-jr> jgarzik: that's a genious idea btw
366 2012-05-15 14:08:30 <jgarzik> gmaxwell: true ;-)
367 2012-05-15 14:09:04 <gmaxwell> The finny attack actually becomes more expensie the more people are doing it: drives up the market price for on demand hashing power. There are only so many such attacks that can go on network wide.
368 2012-05-15 14:09:15 <gmaxwell> (every block could be a finny attack, thats the maximum)
369 2012-05-15 14:09:18 <jgarzik> Also a great point in the thread, that I had not thought of:  sites without instant delivery can accept 0-conf initially, and then check confirmation count > X at a later time (e.g. before shipping the product)
370 2012-05-15 14:09:35 <jgarzik> obvious to some, perhaps, but I hadn't thought of it
371 2012-05-15 14:09:52 <gmaxwell> Yea, I'e long pointed that out to people. ... and here I though I was just stating the obvious!
372 2012-05-15 14:09:59 <jgarzik> ;)
373 2012-05-15 14:10:33 <forrestv> gmaxwell, unless finny-pools arise that let multiple people make attacks in the same block..
374 2012-05-15 14:10:44 <gmaxwell> forrestv: oh, good point.
375 2012-05-15 14:12:02 <gmaxwell> forrestv: speaking of that have you given any consideration to an idea like "p2pool shares disclose the identity of w/ fee transactions they contain, and subsiquent shares are required to contain those transactions to be considered valid (until the next bitcoin block)"?  Would require a lot more bitcoin integration.
376 2012-05-15 14:12:06 <luke-jr> forrestv: which is automatic when fees can be customized&
377 2012-05-15 14:12:18 <gmaxwell> forrestv: but it would give you committments to include transactions very fast.
378 2012-05-15 14:12:57 <forrestv> mm.. it would also restrict the freedom of choosing whichever txns you want to mine
379 2012-05-15 14:13:13 <gmaxwell> forrestv: thats also why I limited it to "with fee" in that statement.
380 2012-05-15 14:13:29 <forrestv> it's still a restriction :p
381 2012-05-15 14:13:30 <gmaxwell> It could even be "with a sufficiently high fee as determined by a vote of the users"
382 2012-05-15 14:13:34 <luke-jr> gmaxwell: so more-centralized fee rules?
383 2012-05-15 14:15:38 <gmaxwell> forrestv: basically it would encourage people to include fees in order to get the 10-second ???confirmations.. not really valuable with only a few hundred GH/s on p2pool, but it would be if some p2pool like system were a hashpower majority.
384 2012-05-15 14:17:01 <gmaxwell> luke-jr: You'd be alone.
385 2012-05-15 14:17:06 <luke-jr> gmaxwell: I don't think so.
386 2012-05-15 14:18:17 <luke-jr> if Deepbit started using its hashpower to enable double spends, everyone would see it that way
387 2012-05-15 14:18:20 <luke-jr> why is p2pool any differnet?
388 2012-05-15 14:18:43 <sipa> because the transactions are decided by the individual miners
389 2012-05-15 14:18:52 <luke-jr> sipa: not in the case gmaxwell is suggesting
390 2012-05-15 14:18:55 <gmaxwell> enable double spends?!@#!
391 2012-05-15 14:18:57 <gmaxwell> huh?
392 2012-05-15 14:19:07 <gmaxwell> I think you misunderstood me.
393 2012-05-15 14:19:09 <n0n00dz4u> wow you guys keep booting me
394 2012-05-15 14:19:15 <n0n00dz4u> I didn't even say anything
395 2012-05-15 14:19:17 <n0n00dz4u> wrong
396 2012-05-15 14:19:19 <n0n00dz4u> but okay
397 2012-05-15 14:19:22 <n0n00dz4u> whatever
398 2012-05-15 14:19:30 <luke-jr> gmaxwell: I'm a p2pool miner. I send a 0-fee transaction to someone. Then I use p2pool to finney attack him with a dobule spend with fee.
399 2012-05-15 14:20:00 <gmaxwell> luke-jr: you can use use btc+deepbit for that already.
400 2012-05-15 14:20:09 <gmaxwell> just send your finny txn to both ... tada!
401 2012-05-15 14:20:26 <luke-jr> gmaxwell: Deepbit replaces transactions with ones with higher fees?
402 2012-05-15 14:20:44 <kinlo> what's a finney attack?
403 2012-05-15 14:20:47 <luke-jr> ooh good, I think I just confirmed my code changes are safe
404 2012-05-15 14:20:58 <luke-jr> kinlo: double-spend at 0 confirmations by finding a block
405 2012-05-15 14:21:05 <luke-jr> kinlo: ie, so the double spend is never seen before the block
406 2012-05-15 14:21:16 <sipa> https://en.bitcoin.it/wiki/Weaknesses#The_.22Finney.22_attack
407 2012-05-15 14:22:37 <sipa> https://bitcointalk.org/index.php?topic=3441.msg48384#msg48384
408 2012-05-15 14:22:44 <kinlo> luke-jr: as a p2pool miner you can't influence the transactions being mined by p2pool, now can you
409 2012-05-15 14:23:25 <kinlo> perhaps you can modify the transactions you mine yourself, but that means you only have your own hashing power, so that equals to solo mining
410 2012-05-15 14:23:43 <luke-jr> kinlo: gmaxwell was suggesting that forrestv make it so you can
411 2012-05-15 14:23:53 <kinlo> luke-jr: your own
412 2012-05-15 14:24:09 <luke-jr> kinlo: no, he wants everyone else ot be forced to include txns with some fee
413 2012-05-15 14:24:43 <luke-jr> [16:12:02] <gmaxwell> forrestv: speaking of that have you given any consideration to an idea like "p2pool shares disclose the identity of w/ fee transactions they contain, and subsiquent shares are required to contain those transactions to be considered valid (until the next bitcoin block)"?  Would require a lot more bitcoin integration.
414 2012-05-15 14:25:19 <kinlo> mmmz
415 2012-05-15 14:25:22 <jgarzik> does anybody have a pre-generated block chain dataset, that permits one to test Reorganize() ?
416 2012-05-15 14:25:49 <jgarzik> ideally there's block chain A, and then a longer branch that may be injected, causing a multi-block switch
417 2012-05-15 14:26:02 <luke-jr> jgarzik: I think someone (maybe Gavin or sipa?) was working on something like that
418 2012-05-15 14:28:01 <jgarzik> makes me want to write BitcoinTestScript language or macros, that could express this test case, and then generate it at low difficulty
419 2012-05-15 14:33:01 <sipa> jgarzik: i have a forkmode branch, which allows you to create a split chain
420 2012-05-15 14:33:16 <sipa> if you do it on testnet, and mine a bit yourself, you can get a side branch
421 2012-05-15 14:52:33 <gavinandresen> jgarzik: BitcoinTestScript is a good idea, the ParseScript() code I wrote for the test/script_tests.cpp unit tests might be a good start
422 2012-05-15 14:54:02 <luke-jr> 5% secure against CVE-2012-2459
423 2012-05-15 14:55:15 <luke-jr> (4% of which is 0.6.2)
424 2012-05-15 15:13:44 <helo> ouch
425 2012-05-15 15:41:50 <gribble> New news from bitcoinrss: jgarzik opened pull request 1316 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1316>
426 2012-05-15 15:44:03 <phantomcircuit> osmosis, "it's broken"
427 2012-05-15 15:46:02 <osmosis> phantomcircuit, there are munin plugins for bitcoind but they are broken?
428 2012-05-15 15:46:13 <phantomcircuit> was a joke about that state of bitcoind
429 2012-05-15 15:46:27 <osmosis> hm
430 2012-05-15 15:46:39 <osmosis> well, i could probably put together my own pretty easily
431 2012-05-15 15:46:43 <gmaxwell> phantomcircuit: I can't seem to find any open issues by you in the system, am I looking under the wrong name?
432 2012-05-15 15:46:52 <osmosis> the bitcoind stats cmd doesnt really give that much info to graph though
433 2012-05-15 15:47:15 <phantomcircuit> gmaxwell, i can never figure out exactly what's going on
434 2012-05-15 15:47:27 <phantomcircuit> but for example rpc will stall when trying to pull 10k transactions
435 2012-05-15 15:47:39 <phantomcircuit> which is reasonable :)
436 2012-05-15 15:47:45 <gmaxwell> I was going to say
437 2012-05-15 15:48:08 <gmaxwell> also, that sounds like an area that would be easy to improve.
438 2012-05-15 15:48:18 <osmosis> blocks, connections, difficulty are about all could graph. Not a lot of useful network information.
439 2012-05-15 15:48:29 <osmosis> via  ./bitcoind getinfo
440 2012-05-15 15:48:51 <phantomcircuit> gmaxwell, basically the issue is
441 2012-05-15 15:48:57 <phantomcircuit> a single rpc call will block
442 2012-05-15 15:49:06 <sipa> wallet rpc's shouldn't lock cs_main, and the other way around
443 2012-05-15 15:49:11 <phantomcircuit> but there isn't a sane queueing mechanism
444 2012-05-15 15:49:31 <phantomcircuit> i understand it's supposed to use the socket queue as an rpc queue
445 2012-05-15 15:49:39 <jgarzik> phantomcircuit: latest upstream changed all that
446 2012-05-15 15:49:52 <sipa> i hope that blockstore can be merged soon, so we can limit the use of cs_main further
447 2012-05-15 15:49:57 <phantomcircuit> but that means if something takes longer than the timeout for socket queue i get exceptions i have to catch :)
448 2012-05-15 15:50:14 <osmosis> gtk-gnutella has some amazing network packet analysis, displayed in the gui even. Id like to have stats like that available from bitcoind getinfo
449 2012-05-15 15:50:24 <gmaxwell> Well multithreaded rpc would help that too, even without all the locking changes.
450 2012-05-15 15:50:42 <gmaxwell> osmosis: start coding!
451 2012-05-15 15:51:11 <phantomcircuit> osmosis, i can pretty much guarantee you're not using a lot of bandwidth
452 2012-05-15 15:51:45 <sipa> unless you're downloading or distributing the blockchain
453 2012-05-15 15:52:01 <sipa> gmaxwell: anyway, if you're interested, i have a backup of my stuck chain
454 2012-05-15 15:52:08 <gmaxwell> sipa: weird timing.
455 2012-05-15 15:52:12 <gmaxwell> I was just about to ask.
456 2012-05-15 15:52:42 <gmaxwell> I'm testing your stuck fix some as we speak.. but just basic sanity testing.. syncing a new node with it, etc.
457 2012-05-15 15:53:11 <gmaxwell> commands to add and kill peers, and manipulate the blacklist would be good too.
458 2012-05-15 15:53:24 <osmosis> phantomcircuit, at 172 connections averaging about 151 Kb out
459 2012-05-15 15:53:41 <sipa> gmaxwell: i used -forkmode=1M3nGRvJHELWqt8iJXZbeBBTbTWCAebDCJ to create my chain (which is first used in block 178001)
460 2012-05-15 15:53:44 <gmaxwell> But yea, with ipv6 and the onion support peer management rpcs are more important than they used to be.
461 2012-05-15 15:53:45 <phantomcircuit> osmosis, impressive
462 2012-05-15 15:53:49 <phantomcircuit> i see like
463 2012-05-15 15:53:58 <phantomcircuit> maybe 1 KB/s
464 2012-05-15 15:54:13 <sipa> osmosis: probably one is downloading the chain
465 2012-05-15 15:54:15 <sipa> or more than one
466 2012-05-15 15:54:18 <gmaxwell> Also, wumpus console patch increased the audience for RPC commands considerably,  more value in new rpc commands now.
467 2012-05-15 15:54:31 <jgarzik> gmaxwell: yes, that was nice to see
468 2012-05-15 15:55:02 <sipa> gavinandresen: by the way, had time to test threadsync?
469 2012-05-15 15:55:12 <gavinandresen> sipa: nope
470 2012-05-15 15:55:44 <jgarzik> CBlockStore seems nice but... ambitious and freakin' huge :)
471 2012-05-15 15:56:02 <jgarzik> I would rather simplify CTransaction and CBlock first -- remove the file I/O stuff from those classes
472 2012-05-15 15:56:26 <sipa> jgarzik: i think matt rebased it more than enough by now
473 2012-05-15 15:56:30 <jgarzik> having an object open a database and read itself is... inside out IMO
474 2012-05-15 15:56:40 <diki> I see you guys took my suggestion and implemented it :)
475 2012-05-15 15:56:57 <diki> Where number of blocks left are displayed in the progress bar
476 2012-05-15 15:57:08 <sipa> agree on that, but blockstore brings real improvement, like parallel block chain checking
477 2012-05-15 15:57:12 <diki> though mine was num blocks dled/total left
478 2012-05-15 15:57:16 <diki> but still, an improvement
479 2012-05-15 15:57:29 <jgarzik> sipa: I think it still has design issues... it's huge and may not best choice for the future.  tries to do too much
480 2012-05-15 15:57:41 <sipa> huge?
481 2012-05-15 15:57:56 <sipa> the only large change is movement of some code from main to protocol
482 2012-05-15 15:58:06 <sipa> large in terms of lines changed
483 2012-05-15 15:58:07 <gmaxwell> I'd prever we avoid pulling both block database changes and the chain pull fix without at least a few days gap.
484 2012-05-15 15:58:39 <diki> love the changes
485 2012-05-15 15:58:49 <diki> bitcoin-qt > bitcoin 3
486 2012-05-15 15:59:01 <jgarzik> sipa: the class interface
487 2012-05-15 15:59:40 <sipa> hmm maybe, but the block chain management is a beast
488 2012-05-15 15:59:47 <jgarzik> sipa: RegisterFoo, RegisterBar, ...  an ugly hardcoded interface that would be better served with signal classes or something more modular
489 2012-05-15 16:00:07 <jgarzik> sipa: the class interface is just a long, unorganized, all-other-the-place laundry list
490 2012-05-15 16:00:31 <jgarzik> transactions, blocks, commit alerts, it's got it all in one class
491 2012-05-15 16:00:44 <sipa> well, currently it's all not in any class at all
492 2012-05-15 16:01:04 <jgarzik> sipa: true -- but that does not justify CKitchenSink :)
493 2012-05-15 16:02:18 <sipa> jgarzik: blockstore currently is two things: a message dispatcher and block chain management
494 2012-05-15 16:02:51 <sipa> the idea was to first split these from rpc/net/ui, and in a second step split block chain management from the message dispatcher too
495 2012-05-15 16:03:15 <sipa> so it becomes a regular signal handler like the wallet
496 2012-05-15 16:04:09 <jgarzik> sipa: it does: event dispatching, event _handling_, network alert relaying, block index services, ...   it is a kitchen sink
497 2012-05-15 16:04:23 <jgarzik> it is so large, comments divide up the class into sections: https://github.com/TheBlueMatt/bitcoin/blob/6bb36f2343ef5c00666520d3ae84eabf0204f28b/src/blockstore.h
498 2012-05-15 16:04:48 <jgarzik> what does "RelayAlerts" have to do with a "blockstore"?
499 2012-05-15 16:05:14 <sipa> well, maybe the name isn't well chosen, but as i said, it's a message dispatcher
500 2012-05-15 16:05:49 <jgarzik> and handler.  the mail carrier _and_ the mail.  plus random shite, because it wasn't already in a class.  :)
501 2012-05-15 16:06:03 <jgarzik> random bits of state like IsInitialBlockDownload
502 2012-05-15 16:06:28 <sipa> as the idea is that all network inputs passes the dispatcher to be processed in the right place, transactions, blocks, alerts, ... need their place in it
503 2012-05-15 16:06:51 <jgarzik> fixing the flow is a good thing.  mixing dispatcher and dispatched is not.
504 2012-05-15 16:07:17 <jgarzik> anybody, babysitting time
505 2012-05-15 16:07:25 <sipa> i'm sure you would have complained the code change was too large to do in one step then
506 2012-05-15 16:09:28 <jgarzik> no, it's a design issue as described, mixing multiple idioms
507 2012-05-15 16:10:53 <sipa> blockstore becomes the central communication between a) network b) blockchain c) wallets
508 2012-05-15 16:11:08 <sipa> for now, the blockchain handling is part of it, rather than be a separate part
509 2012-05-15 16:11:24 <sipa> so yes, it's a mix of a message handler and doing some of the handling
510 2012-05-15 16:12:12 <sipa> but it's an improvement towards a better design; right now those three are arbitrarily-linked together in a very strange net
511 2012-05-15 16:12:19 <jgarzik> a boost-ish solution would be a generic message dispatcher, plus some auto-registering event handlers that do the real work
512 2012-05-15 16:12:52 <sipa> it uses boost signals; not sure what more you want?
513 2012-05-15 16:17:49 <sipa> ?
514 2012-05-15 16:18:06 <gmaxwell> I'd think the goal of cleaner interfaces wouldn't be a C++ api as much as bit of C++ glue to hook up a socket api and move the wallet and gui into anothe rprocess.
515 2012-05-15 16:18:24 <luke-jr> gmaxwell: yes, that's pretty ideal
516 2012-05-15 16:19:10 <sipa> as soon as the interface between blockchain, wallets, ui and network happens through a message dispatcher, it's pretty easy to make the message dispatcher work via network
517 2012-05-15 16:19:27 <gmaxwell> (pretty crazy that we handle long term private keys worth potentially tens of thousands of dollars in a process directly exposed to the public internet, even though the code base has done well against classical exploitation)
518 2012-05-15 16:21:53 <sipa> luke-jr: which C++ api do you speak of, by the way?
519 2012-05-15 16:22:10 <luke-jr> sipa: there were some hooks for callbacks that got removed during the 0.5 or 0.6 window IIRC
520 2012-05-15 16:22:19 <luke-jr> and some more that was removed during 0.3
521 2012-05-15 16:22:27 <sipa> you mean the pubsub system?
522 2012-05-15 16:22:29 <luke-jr> yeah
523 2012-05-15 16:22:37 <sipa> that was for network communication
524 2012-05-15 16:22:40 <luke-jr> O.o
525 2012-05-15 16:23:02 <sipa> so merchants could publish what they were selling and stuff
526 2012-05-15 16:31:32 <neofutur> ;;title https://bitcointalk.org/index.php?topic=71542
527 2012-05-15 16:31:33 <gribble> [BOUNTY - 25 BTC] : read only blockchain patch for bitcoind
528 2012-05-15 16:33:00 <gmaxwell> neofutur: What you're suggesting is the wrong way to accomplish what you want.  You've been told this before.
529 2012-05-15 16:37:47 <gribble> New news from bitcoinrss: rebroad opened issue 1317 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1317>
530 2012-05-15 16:38:46 <weex> gmaxwell: i would also like to see this feature...is a bounty a decent way to go about it?
531 2012-05-15 16:39:18 <sipa> what we need is a nice split between blockchain+net and wallet+ui
532 2012-05-15 16:40:30 <weex> does that cut much bone to use a medical metaphor?
533 2012-05-15 16:41:07 <luke-jr> a bad metaphor; what do you mean? XD
534 2012-05-15 16:44:05 <weex> youre right, i can't go anywhere with that
535 2012-05-15 16:45:50 <weex> primarily what i would like is to be able to have multiple bitcoinds running that i can query to see if they've seen transactions for a set of addresses
536 2012-05-15 16:47:35 <weex> unconfirmed txns mainly but also to support thin clients who just have a few private keys
537 2012-05-15 16:47:53 <weex> they want to check their balance and broadcast txns
538 2012-05-15 16:49:19 <weex> i guess that is accomplished by sipa's split but i don't require the bitcoinds i'm querying not to have their own wallets/ui
539 2012-05-15 16:50:23 <weex> mike hearn on somethign close to the subject: https://bitcointalk.org/index.php?topic=57435.msg892884#msg892884
540 2012-05-15 16:51:38 <luke-jr> weex: you're describing what sipa said :p
541 2012-05-15 16:51:53 <luke-jr> weex: at least as I envision it
542 2012-05-15 16:52:34 <luke-jr> while Bitcoin-Qt would probably spawn bitcoind itself on a local socket (stdio?), you could just run bitcoind as a TCP server still
543 2012-05-15 16:53:16 <weex> so it's easy! :D
544 2012-05-15 16:54:50 <weex> lol....i'm well aware that it's not and that i'll not be the one to do it so the best i can do is provide my perspective
545 2012-05-15 16:55:14 <gribble> Order id 8063 created.
546 2012-05-15 16:55:14 <jine> ;;sell 1 Domain at 1000 BTC Bitcoins.LC - send bids to domains@jine.se
547 2012-05-15 16:55:20 <jine> Oh, crap.. wrong channgel :)
548 2012-05-15 16:56:45 <weex> with this kind of feature, it becomes quite easy to accept fast payments as you just query 10 bitcoinds
549 2012-05-15 16:57:38 <weex> until 80% accept a txn for you
550 2012-05-15 16:58:50 <graingert> weex: there is a site that does this
551 2012-05-15 16:59:14 <graingert> weex: http://transactionradar.com/
552 2012-05-15 16:59:33 <weex> hey graingert , how's it going?
553 2012-05-15 16:59:38 <graingert> weex: heya
554 2012-05-15 16:59:39 <weex> i'm aware of that but it's not fast
555 2012-05-15 16:59:57 <weex> 0-conf fast i mean
556 2012-05-15 17:00:50 <weex> plus i want not 1 server but tens to thousands
557 2012-05-15 17:02:39 <weex> instead of a transaction, i'd love to see it accept an address or list of addresses
558 2012-05-15 17:12:35 <phantomcircuit> gavinandresen, around?
559 2012-05-15 17:12:59 <phantomcircuit> is there someone here who can whitelist a forum account?
560 2012-05-15 17:13:22 <graingert> weex: what do you want to do?
561 2012-05-15 17:13:33 <graingert> what do you want to sell at 0 conf?
562 2012-05-15 17:14:29 <graingert> because if you want to sell a subscription then 0 conf is fine
563 2012-05-15 17:14:32 <graingert> you just cut them off
564 2012-05-15 17:14:33 <graingert> and sue
565 2012-05-15 17:15:49 <jgarzik> phantomcircuit: I can try
566 2012-05-15 17:15:58 <jgarzik> phantomcircuit: I have mod on -a- board, so maybe it will work
567 2012-05-15 17:18:14 <weex> graingert: i'm selling downloads at coindl
568 2012-05-15 17:19:29 <weex> sue for a subscription lol...bbl
569 2012-05-15 17:22:58 <gavinandresen> phantomcircuit: I'm back
570 2012-05-15 17:26:44 <phantomcircuit> gavinandresen, can you whitelist someone on the forum?
571 2012-05-15 17:27:18 <gavinandresen> phantomcircuit: yes
572 2012-05-15 17:27:45 <phantomcircuit> please whitelist Bitcoinica Consultancy
573 2012-05-15 17:28:49 <gavinandresen> ok... (looking for the whitelist button, always have a hard time finding it....)
574 2012-05-15 17:29:04 <gavinandresen> done, I think
575 2012-05-15 17:29:36 <phantomcircuit> thanks
576 2012-05-15 17:33:32 <luke-jr> weex: that check won't beat finney
577 2012-05-15 17:38:21 <eian> how many simultaneous connections does bitcoind support?  Has anyone hit an upper bound?
578 2012-05-15 17:38:51 <XMPPwocky> eian: it should grow to fill memory
579 2012-05-15 17:39:14 <XMPPwocky> in theory, you might hit MAX_INT or MAX_SHORt
580 2012-05-15 17:40:11 <eian> just wondering what type of peering  #'s the large exchanges reach with bitcoind
581 2012-05-15 17:41:39 <eian> MAX_INT doesn't seem likely - I don't think a kernel can handle billions of open socket descriptors haha
582 2012-05-15 17:43:19 <XMPPwocky> eian: well, I figured in bitcoind
583 2012-05-15 17:45:55 <luke-jr> eian: bitcoind 0.6.2 supports 1 simultaneous (JSON-RPC) connection
584 2012-05-15 17:46:17 <luke-jr> eian: bitcoind uses select(), so it can never have more than 1024 file descriptors
585 2012-05-15 17:46:37 <luke-jr> eian: exchanges have no particular need to peer much, but at least MtGox does not use bitcoind
586 2012-05-15 17:48:03 <terry> luke-jr: "PHP can do anything"
587 2012-05-15 17:48:06 <terry> ?
588 2012-05-15 17:48:13 <luke-jr> terry: what?
589 2012-05-15 17:48:31 <terry> MagicalTux's signature used to be: "PHP can do anything"
590 2012-05-15 17:48:36 <luke-jr> O.o
591 2012-05-15 17:49:02 <eian> what does MtGox use?
592 2012-05-15 17:49:40 <luke-jr> eian: presumably QBitcoin Core (which hasn't been released yet)
593 2012-05-15 17:50:10 <eian> gotcha, thanks
594 2012-05-15 17:50:26 <luke-jr> MagicalTux might be able to comment further
595 2012-05-15 17:57:04 <luke-jr> argh
596 2012-05-15 17:57:22 <luke-jr> just renewed my domains& made a brief google for coupon codes, none found
597 2012-05-15 17:57:37 <sipa> qbitcoin core?
598 2012-05-15 17:57:56 <luke-jr> later wanted to go back and see what their first attempted upsell was, but it didn't try to upsell again; click the X to close the window, and it offers me 10% off if I don't close it without ordering -.-
599 2012-05-15 17:58:04 <luke-jr> sipa: yes, MagicalTux's client
600 2012-05-15 17:58:13 <luke-jr> https://en.bitcoin.it/wiki/QBitcoin
601 2012-05-15 17:59:01 <luke-jr> MagicalTux: btw, that page still says 0.1.0 planned for Jan 2011 :P
602 2012-05-15 18:10:32 <upb> written in 100% pure PHP
603 2012-05-15 18:10:50 <luke-jr> Qt
604 2012-05-15 18:28:12 <neofutur> gmaxwell: thanks for posting on https://bitcointalk.org/index.php?topic=71542
605 2012-05-15 18:28:19 <neofutur> i updated the op with your comment
606 2012-05-15 18:30:26 <gmaxwell> neofutur: thanks!
607 2012-05-15 18:31:45 <neofutur> thanks to you ! no doubt you know better than me the best way to address those problems / needs ;)
608 2012-05-15 19:51:10 <gribble> New news from bitcoinrss: Diapolo opened pull request 1319 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1319>
609 2012-05-15 19:58:37 <Matt_von_Mises> Anyone know how to link OpenSSL to a program on OSX?
610 2012-05-15 20:01:18 <Matt_von_Mises> -lcrypto -lssl Linker flags worked
611 2012-05-15 20:01:51 <TD> damnit. i hate how people make it hard to find their email addresses these days
612 2012-05-15 20:02:07 <TD> what's with your project website linking your name to github which in turn only links to your twitter page
613 2012-05-15 20:02:28 <TD> which then in turn links to a google profile
614 2012-05-15 20:02:29 <TD> argh
615 2012-05-15 20:07:07 <luke-jr> TD: I just cheat and use git.
616 2012-05-15 20:07:19 <luke-jr> clone some repo of theirs and git log :D
617 2012-05-15 20:07:22 <gmaxwell> TD: clone their repo on github and look at the log.
618 2012-05-15 20:07:28 <TD> hah
619 2012-05-15 20:07:30 <TD> didn't think of that
620 2012-05-15 20:07:38 <gmaxwell> I had to do this just the other day had the same complaint as you. :)
621 2012-05-15 20:13:52 <eian> Is there a programatic way to get the geographical location of the peers I'm connected to? (Assuming I have the IP address)
622 2012-05-15 20:14:22 <eian> Is there some online database I can query?
623 2012-05-15 20:14:28 <luke-jr> http://www.maxmind.com/app/ip-location
624 2012-05-15 20:15:00 <eian> luke, thanks
625 2012-05-15 20:16:34 <eian> luke, any idea how much that costs?
626 2012-05-15 20:16:50 <luke-jr> eian: nothing, if you just want states
627 2012-05-15 20:17:05 <eian> oh, I didn't notice the free version
628 2012-05-15 20:17:15 <eian> I was hoping for something that did international lookups
629 2012-05-15 20:22:41 <luke-jr> ?
630 2012-05-15 20:22:44 <luke-jr> afaik it's international
631 2012-05-15 20:23:35 <gmaxwell> eian: the nothing is for state or lower resolution.
632 2012-05-15 20:23:59 <gmaxwell> The commercial DB's will get you down to street addresses (and they're accurate at least to zipcodes in many cases)
633 2012-05-15 20:25:49 <luke-jr> whois an IP is also handy, but I don't think you can hit it too hard
634 2012-05-15 20:32:39 <eian> thanks guys
635 2012-05-15 20:39:41 <weex> < luke-jr> weex: that check won't beat finney
636 2012-05-15 20:40:23 <weex> thankfully the economics of small transactions <$100 make finney unlikely to be tried
637 2012-05-15 21:05:40 <ashconnor> Anybody managed to compile bitcoin on osx using homebrew instead of macports?
638 2012-05-15 21:21:58 <gribble> New news from bitcoinrss: chickentech opened issue 1320 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1320>
639 2012-05-15 21:29:20 <gmaxwell> weex: finney costs ~nothing to perform.
640 2012-05-15 21:29:57 <gmaxwell> weex: I mean sure below some value you don't really care if a few jerks rip you off, but thats not the same as the attacks not happening.
641 2012-05-15 21:31:02 <gmaxwell> weex: also 'get lots of nodes to accept a txn' doesn't even require a finney attack to defeat.
642 2012-05-15 21:32:34 <gmaxwell> weex: I just near simultaneously release a txn paying you to your node. ... and a conflicting one directly to deepbit and btcguild's nodes. You'll get lots and lots of nodes to accept the one paying you, and you'll still likely not make it in the chain.
643 2012-05-15 21:51:43 <tosku> I updated to 0.6.2 for Windows.
644 2012-05-15 21:51:58 <tosku> There seems to be something wrong with the Swedish translation.
645 2012-05-15 21:52:14 <tosku> Every button text starts with "amp; "
646 2012-05-15 21:53:19 <tosku> Well, not every, but the home button, send bitcoins button and the receive bitcoins button.
647 2012-05-15 21:53:35 <Diablo-D3> [07:52:53] <MagicalTux> we have written consent from Dwolla, basically now Dwolla has "normal" merchant accounts, and "special" merchant accounts with extra requirements
648 2012-05-15 21:54:26 <luke-jr> tosku: fix it? :P
649 2012-05-15 21:55:10 <tosku> luke-jr: Sure, if you tell me how :)
650 2012-05-15 21:55:27 <luke-jr> some website
651 2012-05-15 21:55:54 <gavinandresen> transifex.com
652 2012-05-15 21:56:02 <luke-jr> ^ :D
653 2012-05-15 21:56:11 <gavinandresen> err, .net maybe
654 2012-05-15 21:56:22 <tosku> it redirects
655 2012-05-15 21:56:23 <luke-jr> (I actually suspect transifex might be to blame for it, since it keeps being reintroduced&)
656 2012-05-15 21:57:17 <gavinandresen> https://www.transifex.net/projects/p/bitcoin/
657 2012-05-15 21:58:46 <luke-jr> man, ton of fee txns tonight
658 2012-05-15 21:58:49 <luke-jr> 815 pending now
659 2012-05-15 22:03:45 <tosku> gavinandresen: From what I can see online, there's ampersands where there are supposed to be ampersands. "&Overview" is translated to "&??versikt" and "&Transactions" is translated to "&Transaktioner". On my computer, the Overview button is displayed as "amp& ??versikt", but the Transaction button is displayed correctly. It seems the problem is not with the translation itself.
660 2012-05-15 22:04:20 <sipa> wumpus: ^ any idea?
661 2012-05-15 22:29:08 <weex> gmaxwell: so the big miners should definitely be on the list of nodes to check
662 2012-05-15 23:21:46 <gmaxwell> weex: the attacker only needs one sufficiently big miner, and he'll be successful at least some of the time. You can't know all their addresses...