1 2012-02-05 00:00:20 <sipa> no standard client will do so
   2 2012-02-05 00:00:28 <sipa> but you can create one that does, indeed
   3 2012-02-05 00:00:43 <roconnor> technically I haven't tested it yet, but I have little doubt.
   4 2012-02-05 00:01:34 cdecker has quit (Ping timeout: 240 seconds)
   5 2012-02-05 00:01:36 <gmaxwell> Sipa's fix plus adding heights to the API would still address it and would remove the third party api user's hazard.
   6 2012-02-05 00:01:45 Guest89132 has joined
   7 2012-02-05 00:01:54 osmosis has quit (Ping timeout: 240 seconds)
   8 2012-02-05 00:02:02 <sipa> roconnor's idea is disallowing any coinbase to be reused
   9 2012-02-05 00:02:07 <sipa> and keeping them forever
  10 2012-02-05 00:02:13 welterde has joined
  11 2012-02-05 00:02:16 <gmaxwell> I know but it's unbounded storage (though realastically not bad)
  12 2012-02-05 00:02:20 <shargs> gmaxwell=bitcoin master
  13 2012-02-05 00:02:32 <roconnor> it effectively adds 256bits to the "blockheader"
  14 2012-02-05 00:03:20 dooglus has quit (Ping timeout: 252 seconds)
  15 2012-02-05 00:03:24 <gmaxwell> In terms of short term behavior.. how are gavin's discouraging patches doing? discouraging duplicate coinbases would pratically close this.
  16 2012-02-05 00:03:44 <luke-jr> http://www.csoonline.com/article/695680/cracking-md5-…-with-google- lol
  17 2012-02-05 00:03:48 Sze has joined
  18 2012-02-05 00:04:11 Cablesaurus has quit (Quit: If you think nobody cares, try missing a few payments)
  19 2012-02-05 00:05:08 <gmaxwell> (since with dircouraging the attacker would need to produce back to back blocks to get their dupe coinbase into the chain)
  20 2012-02-05 00:05:45 AAA_awright_ is now known as AAA_awright
  21 2012-02-05 00:05:47 <roconnor> gmaxwell: even that wouldn't work if the miners are willing to fork the chain at that point.
  22 2012-02-05 00:06:37 EPiSKiNG- has quit ()
  23 2012-02-05 00:07:20 <gmaxwell> roconnor: discouraging isn't a complete fix, but it would at least require miners go along with the attack. E.g. preventing people from using it for griefing while the real fix is worked out.
  24 2012-02-05 00:07:42 <roconnor> yes
  25 2012-02-05 00:08:36 <gmaxwell> though I don't know if it's needed... or if the normal barrier of mining a block is enough to stop casual trouble making.
  26 2012-02-05 00:09:07 Cablesaurus has joined
  27 2012-02-05 00:10:00 theorb has joined
  28 2012-02-05 00:10:18 <roconnor> gmaxwell: how big is the zombie mining network?
  29 2012-02-05 00:10:31 <roconnor> gmaxwell: if there is a real attack based on this, then there will be motivation.
  30 2012-02-05 00:10:35 theorbtwo has quit (Ping timeout: 252 seconds)
  31 2012-02-05 00:10:43 theorb is now known as theorbtwo
  32 2012-02-05 00:10:48 <roconnor> I mean; I have no idea what happens when the reorg corrupts the chain db.
  33 2012-02-05 00:11:02 <roconnor> maybe nothing bad
  34 2012-02-05 00:11:10 <roconnor> maybe something very bad.
  35 2012-02-05 00:11:14 Clipse has quit (Ping timeout: 240 seconds)
  36 2012-02-05 00:11:25 <sipa> you get a block chain split as soon as the coinbase is spent the second time
  37 2012-02-05 00:11:33 <sipa> those who saw the reorg won't accept it
  38 2012-02-05 00:11:39 <sipa> those who didn't will
  39 2012-02-05 00:12:38 Sze has quit (Read error: Connection timed out)
  40 2012-02-05 00:12:54 jercos_ is now known as jercos
  41 2012-02-05 00:13:00 RazielZ has quit (Quit: Leaving)
  42 2012-02-05 00:13:02 <roconnor> so you put your target in a part of the fork with low mining capabilities
  43 2012-02-05 00:13:12 <gmaxwell> yet another reason why big miners should all @#$@# fully mesh their p2p connections.
  44 2012-02-05 00:13:26 <roconnor> wait for a while for the difficulty to drop and start 50% attacking them.
  45 2012-02-05 00:13:46 <roconnor> ... a promising start for an attack ...
  46 2012-02-05 00:14:03 <gmaxwell> roconnor: the ones who didn't see the reorg can still reorg onto the chain of those who saw it.
  47 2012-02-05 00:14:07 <gmaxwell> (if it's longer)
  48 2012-02-05 00:14:27 * roconnor 's head hurts
  49 2012-02-05 00:14:31 <gmaxwell> roconnor: the site that won't ever move is the side thats is far more likely to be the bigger hashpower side. (thankfully!)
  50 2012-02-05 00:14:55 <gmaxwell> (simply because if you're not online all the time, thus will see the reorg, you're not likely to be doing much mining)
  51 2012-02-05 00:15:03 num1 has quit (Quit: Textual IRC Client: http://www.textualapp.com/)
  52 2012-02-05 00:15:12 osmosis has joined
  53 2012-02-05 00:15:26 <sipa> you need a coinbase with at least two txouts, to pull off this attack, imho
  54 2012-02-05 00:15:50 ahbritto has joined
  55 2012-02-05 00:15:50 ahbritto has quit (Changing host)
  56 2012-02-05 00:15:50 ahbritto has joined
  57 2012-02-05 00:17:03 paraipan has quit (Remote host closed the connection)
  58 2012-02-05 00:17:52 paraipan has joined
  59 2012-02-05 00:23:26 denisx has joined
  60 2012-02-05 00:26:41 datagutt has quit (Quit: Computer has gone to sleep.)
  61 2012-02-05 00:32:28 Zarutian has quit (Quit: Zarutian)
  62 2012-02-05 00:33:23 Rabbit67890 has quit (Quit: Rabbit67890)
  63 2012-02-05 00:34:35 <userhj> what is bip16+ver?
  64 2012-02-05 00:36:59 <shargs> bipalope
  65 2012-02-05 00:37:48 Nicksasa has quit (Quit: I'll be back nubs)
  66 2012-02-05 00:37:59 Nicksasa has joined
  67 2012-02-05 00:38:05 Rabbit67890 has joined
  68 2012-02-05 00:39:14 <luke-jr> userhj: BIP 16 modified to use the txn version instead of a magic script
  69 2012-02-05 00:39:30 <sipa> not "instead of"
  70 2012-02-05 00:39:32 <gmaxwell> s/instead of/in addition to/
  71 2012-02-05 00:39:51 <gmaxwell> but I guess that shows the fundimental nature of luke's objection. :)
  72 2012-02-05 00:40:08 <luke-jr> …
  73 2012-02-05 00:40:23 <gmaxwell> I mean that idea that changing the version makes the script less magic. :)
  74 2012-02-05 00:40:40 <luke-jr> it's not really much better if it's "in addition to" -.-
  75 2012-02-05 00:41:03 <sipa> but you need that script to be there, or you lose backward compatibility
  76 2012-02-05 00:41:15 <gmaxwell> luke-jr: there isn't any way to remove the magic script and #include <sipa>
  77 2012-02-05 00:41:40 <luke-jr> how so?
  78 2012-02-05 00:42:03 <luke-jr> I'm thinking the current rules, plus a if(version==2) Exec(scriptcopy.top) afterward
  79 2012-02-05 00:42:19 <luke-jr> and disallow non-push in scriptSig ofc
  80 2012-02-05 00:42:34 <sipa> right, the script would be non-magic if it isn't a version 2 transaction
  81 2012-02-05 00:42:56 <luke-jr> ?
  82 2012-02-05 00:42:57 * gmaxwell nods
  83 2012-02-05 00:43:19 <gmaxwell> the idea that _adding_ a requirement makes is less magic is weird to me.
  84 2012-02-05 00:43:31 <sipa> if you'd create a BIP16 scriptPubKey, but in a version=1 transaction, it would be a stupid pay to miner
  85 2012-02-05 00:43:35 <gmaxwell> But thats fine. it's not me that needs to be made happy.
  86 2012-02-05 00:43:56 <gmaxwell> pay to non-lazy miner. ;)
  87 2012-02-05 00:44:08 <luke-jr> sipa: that's why if(version==2)
  88 2012-02-05 00:44:08 theymos has joined
  89 2012-02-05 00:44:18 <sipa> gmaxwell: luke came up with a proposal that would *only* allow BIP16 transactions in version=2 transactions
  90 2012-02-05 00:44:20 <luke-jr> on the scriptPubKey txn
  91 2012-02-05 00:44:52 <gmaxwell> sipa: I know. I happy with that, I guess. I don't personally see it as any better (or worse) than BIP16 but I can see why he would.
  92 2012-02-05 00:45:10 <sipa> it does make the version logic more complicated, though
  93 2012-02-05 00:45:16 <sipa> as you need a third version number
  94 2012-02-05 00:45:49 <luke-jr> no rush on deploying versions IMO
  95 2012-02-05 00:46:03 <luke-jr> just P2SH
  96 2012-02-05 00:47:05 <luke-jr> but why not allow any scriptPubKey with version==2 to execute the top scriptSig item? … answer to my own question: because existing clients will be tricked into accepting txns they can't spend :/
  97 2012-02-05 00:48:32 <luke-jr> blarg
  98 2012-02-05 00:49:00 <luke-jr> I suppose anything but "in addition to" has the problem of mixing different output types
  99 2012-02-05 00:49:49 <luke-jr> well that sucks :/
 100 2012-02-05 00:51:09 marf_away has quit (Ping timeout: 245 seconds)
 101 2012-02-05 00:52:32 paraipan has quit (Remote host closed the connection)
 102 2012-02-05 00:54:34 Clipse has joined
 103 2012-02-05 00:54:36 paraipan has joined
 104 2012-02-05 00:54:59 paraipan has quit (Remote host closed the connection)
 105 2012-02-05 00:55:26 paraipan has joined
 106 2012-02-05 00:55:52 Cablesaurus has quit (Quit: Say What?)
 107 2012-02-05 01:00:37 <sipa> luke-jr: you bring up a very interesting point
 108 2012-02-05 01:00:50 <sipa> a client should not accept transactions of a version it doesn't know
 109 2012-02-05 01:01:12 <luke-jr> if they do now, it greatly limits us
 110 2012-02-05 01:01:21 <sipa> and i guess this makes introducing the version logic a bit more dangerous as well
 111 2012-02-05 01:01:57 <luke-jr> I didn't check if that's the case yet
 112 2012-02-05 01:01:57 Guest89132 has quit (Quit: Clever quit message!)
 113 2012-02-05 01:02:20 <AAA_awright> sipa: Why? One could spoof the version number right?
 114 2012-02-05 01:02:28 <sipa> how so?
 115 2012-02-05 01:02:38 <sipa> that changes the txid
 116 2012-02-05 01:02:42 <sipa> and the signatures
 117 2012-02-05 01:02:45 phantomcircuit has joined
 118 2012-02-05 01:02:47 <sipa> so that's not a problem
 119 2012-02-05 01:02:54 <sipa> but it implies that an address (or whatever is used as descriptor for creating txouts) must be associated with a particular version
 120 2012-02-05 01:02:57 ThomasV has quit (Ping timeout: 245 seconds)
 121 2012-02-05 01:03:15 <sipa> which is a problem
 122 2012-02-05 01:03:23 <AAA_awright> Not changing the database, when you're generating the transaction
 123 2012-02-05 01:03:24 <sipa> as it implies you can't create a transaction with mixed outputs
 124 2012-02-05 01:03:29 <gmaxwell> AAA_awright: because it might not actually be spendable by them, so they shouldn't reflect it in their balance.
 125 2012-02-05 01:03:53 <shargs> ok
 126 2012-02-05 01:04:04 <luke-jr> looks like IsMine ignores version :/
 127 2012-02-05 01:04:12 <sipa> everything ignores versions
 128 2012-02-05 01:04:45 * luke-jr gets the feeling the real reason Satoshi remains anonymous is to avoid developers <.<
 129 2012-02-05 01:05:27 <sipa> haha
 130 2012-02-05 01:06:20 <sipa> this is a serious set-back for my proposal, i guess
 131 2012-02-05 01:08:07 Rabbit67890 has quit (Quit: Rabbit67890)
 132 2012-02-05 01:10:05 iocor has quit (Quit: Computer has gone to sleep.)
 133 2012-02-05 01:12:15 Katniss_ has joined
 134 2012-02-05 01:14:22 Katniss has quit (Ping timeout: 256 seconds)
 135 2012-02-05 01:16:15 pingdrive has joined
 136 2012-02-05 01:16:19 denisx has quit (Quit: denisx)
 137 2012-02-05 01:16:19 MobiusL has joined
 138 2012-02-05 01:17:19 Katniss has joined
 139 2012-02-05 01:19:22 Katniss_ has quit (Ping timeout: 252 seconds)
 140 2012-02-05 01:19:37 ThomasV has joined
 141 2012-02-05 01:22:39 osmosis has quit (Quit: Leaving)
 142 2012-02-05 01:27:53 Turingi has quit (Read error: Connection reset by peer)
 143 2012-02-05 01:36:53 sacredch1o has joined
 144 2012-02-05 01:37:14 Cablesaurus has joined
 145 2012-02-05 01:37:34 ThomasV has quit (Ping timeout: 240 seconds)
 146 2012-02-05 01:38:05 sacredchao has quit (Ping timeout: 276 seconds)
 147 2012-02-05 01:38:56 erle- has quit (Quit: erle-)
 148 2012-02-05 01:40:21 marf_away has joined
 149 2012-02-05 01:44:02 userhj has quit (Quit: Leaving)
 150 2012-02-05 01:45:53 da2ce7 has quit (Ping timeout: 276 seconds)
 151 2012-02-05 01:46:48 splatster has quit (Quit: gtg)
 152 2012-02-05 01:47:54 osmosis has joined
 153 2012-02-05 01:51:23 da2ce7 has joined
 154 2012-02-05 01:56:50 d4de has joined
 155 2012-02-05 01:56:50 d4de has quit (Changing host)
 156 2012-02-05 01:56:50 d4de has joined
 157 2012-02-05 01:59:04 marf_away has quit (Ping timeout: 245 seconds)
 158 2012-02-05 01:59:35 eoss has quit (Ping timeout: 260 seconds)
 159 2012-02-05 02:06:30 shargs has quit (Ping timeout: 252 seconds)
 160 2012-02-05 02:07:48 sacarlson has quit (Read error: Connection reset by peer)
 161 2012-02-05 02:09:45 BTC_Bear is now known as hbrntng!~BTC_Bear@unaffiliated/btc-bear/x-5233302|BTC_Bear
 162 2012-02-05 02:09:49 JFK911_ has joined
 163 2012-02-05 02:10:04 JFK911 has quit (Read error: Connection reset by peer)
 164 2012-02-05 02:10:30 b4epoche_ has joined
 165 2012-02-05 02:10:56 SomeoneWeirdzzzz is now known as SomeoneWeird
 166 2012-02-05 02:11:11 occulta has quit (Quit: KVIrc 4.1.1 Equilibrium http://www.kvirc.net/)
 167 2012-02-05 02:11:36 b4epoche has quit (Ping timeout: 256 seconds)
 168 2012-02-05 02:11:36 b4epoche_ is now known as b4epoche
 169 2012-02-05 02:15:10 JFK911_ is now known as JFK911
 170 2012-02-05 02:15:15 eoss has joined
 171 2012-02-05 02:18:01 Rabbit67890 has joined
 172 2012-02-05 02:22:03 splatster has joined
 173 2012-02-05 02:22:08 copumpkin is now known as Ukdo
 174 2012-02-05 02:23:29 Ukdo is now known as copumpkin
 175 2012-02-05 02:26:11 sacarlson has joined
 176 2012-02-05 02:28:23 Mango-chan has quit (Ping timeout: 245 seconds)
 177 2012-02-05 02:32:41 da2ce7 has quit (Ping timeout: 276 seconds)
 178 2012-02-05 02:32:56 cryptoxchange has quit (Read error: No route to host)
 179 2012-02-05 02:33:16 cryptoxchange has joined
 180 2012-02-05 02:33:17 cryptoxchange has quit (Changing host)
 181 2012-02-05 02:33:17 cryptoxchange has joined
 182 2012-02-05 02:35:26 da2ce7 has joined
 183 2012-02-05 02:38:26 theymos has quit (Remote host closed the connection)
 184 2012-02-05 02:41:15 FROTUSCI has joined
 185 2012-02-05 02:46:26 num1 has joined
 186 2012-02-05 02:49:17 a_meteor has joined
 187 2012-02-05 02:50:14 paraipan has quit (Ping timeout: 276 seconds)
 188 2012-02-05 02:50:54 JRWR has quit (Ping timeout: 240 seconds)
 189 2012-02-05 02:51:03 a_meteorite has quit (Ping timeout: 248 seconds)
 190 2012-02-05 02:51:03 a_meteor is now known as a_meteorite
 191 2012-02-05 02:51:04 paraipan has joined
 192 2012-02-05 02:53:50 MrTiggr has joined
 193 2012-02-05 02:54:05 merde has quit (Ping timeout: 245 seconds)
 194 2012-02-05 03:00:32 merde has joined
 195 2012-02-05 03:03:54 MobiusL has quit (Remote host closed the connection)
 196 2012-02-05 03:05:08 MobiusL has joined
 197 2012-02-05 03:08:16 colin__ has joined
 198 2012-02-05 03:08:29 colin__ is now known as c00w
 199 2012-02-05 03:14:00 Cablesaurus has quit (Quit: I used to think I was indecisive, but now I'm not too sure.)
 200 2012-02-05 03:17:59 Cablesaurus has joined
 201 2012-02-05 03:18:00 Cablesaurus has quit (Changing host)
 202 2012-02-05 03:18:00 Cablesaurus has joined
 203 2012-02-05 03:23:08 ivan\ has quit (Quit: ERC Version 5.3 (IRC client for Emacs))
 204 2012-02-05 03:23:12 devrandom has quit (Remote host closed the connection)
 205 2012-02-05 03:24:04 [7] has quit (Read error: Operation timed out)
 206 2012-02-05 03:24:51 TheSeven has joined
 207 2012-02-05 03:26:50 ivan\ has joined
 208 2012-02-05 03:29:29 devrandom has joined
 209 2012-02-05 03:34:31 <sipa> hmm, currently 582644 addresses with non-zero balance
 210 2012-02-05 03:34:49 splatster has quit (Quit: Leaving...)
 211 2012-02-05 03:36:01 <pingdrive> one of them is MEEEEEEEE!
 212 2012-02-05 03:37:02 <FROTUSCI> cool
 213 2012-02-05 03:37:05 cryptoxchange has quit (Read error: Connection reset by peer)
 214 2012-02-05 03:37:39 <copumpkin> sipa: what's the largest balance at a single address?
 215 2012-02-05 03:38:32 cryptoxchange has joined
 216 2012-02-05 03:39:00 <gmaxwell> copumpkin: mtgox has like 450k btc at a single address.
 217 2012-02-05 03:39:27 <copumpkin> yeah, I know about that one
 218 2012-02-05 03:39:32 <copumpkin> I think it's a little less than that now isn't it?
 219 2012-02-05 03:39:39 <copumpkin> (and I'm still not 100% positive that it is mtgox :P)
 220 2012-02-05 03:40:09 <copumpkin> it'd be nice to get a real confirmation from MagicalTux that it is or isn't
 221 2012-02-05 03:42:00 <gmaxwell> copumpkin: we got confirmation before.
 222 2012-02-05 03:42:15 <copumpkin> oh? how?
 223 2012-02-05 03:42:26 <gmaxwell> (MagicalTux moved an amount of coin specified by someone on IRC)
 224 2012-02-05 03:42:59 <FROTUSCI> the amount had 1337 in it
 225 2012-02-05 03:43:04 <sipa> copumpkin: looking
 226 2012-02-05 03:43:16 <sipa> damn, bitcoind is fast when ran from a tmpfs
 227 2012-02-05 03:43:26 <sipa> from disk it would have taken hours to calculate this
 228 2012-02-05 03:46:17 <sipa> largest i find is 105k BTC
 229 2012-02-05 03:49:02 splatster has joined
 230 2012-02-05 03:49:04 splatster has quit (Changing host)
 231 2012-02-05 03:49:04 splatster has joined
 232 2012-02-05 03:50:35 phantomfake has quit (Ping timeout: 240 seconds)
 233 2012-02-05 03:52:37 phantomfake has joined
 234 2012-02-05 04:07:02 Moron__ has quit ()
 235 2012-02-05 04:11:50 c00w has quit (Ping timeout: 255 seconds)
 236 2012-02-05 04:13:23 <MagicalTux> [12:33:43] <copumpkin> it'd be nice to get a real confirmation from MagicalTux that it is or isn't <- we are not storing more than 5k btc per address anymore
 237 2012-02-05 04:13:34 <copumpkin> ah
 238 2012-02-05 04:13:47 <copumpkin> but those huge 500k movements a while back were you?
 239 2012-02-05 04:14:14 <BTC_Bear> MagicalTux: Did Theymos get ahold of you?  He couldn't ssh into the forum.
 240 2012-02-05 04:14:53 <MagicalTux> BTC_Bear: that's fixed
 241 2012-02-05 04:14:53 colin__ has joined
 242 2012-02-05 04:15:42 Katniss_ has joined
 243 2012-02-05 04:17:25 Rabbit67890 has quit (Quit: Rabbit67890)
 244 2012-02-05 04:17:37 Katniss has quit (Ping timeout: 244 seconds)
 245 2012-02-05 04:17:45 Rabbit67890 has joined
 246 2012-02-05 04:17:54 eoss has quit (Quit: Leaving)
 247 2012-02-05 04:18:29 booo has joined
 248 2012-02-05 04:27:17 Katniss has joined
 249 2012-02-05 04:27:39 RobinPKR_ has joined
 250 2012-02-05 04:28:15 RobinPKR has quit (Ping timeout: 240 seconds)
 251 2012-02-05 04:28:15 RobinPKR_ is now known as RobinPKR
 252 2012-02-05 04:29:10 Katniss_ has quit (Ping timeout: 252 seconds)
 253 2012-02-05 04:31:00 colin__ is now known as clr_
 254 2012-02-05 04:38:33 <luke-jr> what OS does gitian need? -.-
 255 2012-02-05 04:38:45 <luke-jr> I assumed Ubuntu, but it doesn't have python-vm-builder or apt-cacher
 256 2012-02-05 04:40:07 <sipa> it has here
 257 2012-02-05 04:40:23 <sipa> Ubuntu 11.10
 258 2012-02-05 04:41:39 <luke-jr> O.o
 259 2012-02-05 04:41:43 <luke-jr> same version too
 260 2012-02-05 04:42:04 <luke-jr> do I need some non-default repo?
 261 2012-02-05 04:42:06 clr_ has quit (Quit: Ex-Chat)
 262 2012-02-05 04:42:21 clr_ has joined
 263 2012-02-05 04:42:23 pingdrive has quit (Quit: Leaving)
 264 2012-02-05 04:42:48 clr_ is now known as Guest82933
 265 2012-02-05 04:43:58 <sipa> not that i know
 266 2012-02-05 04:44:09 <sipa> it may be in universe or multiverse, though
 267 2012-02-05 04:44:47 <sipa> universe, apparently
 268 2012-02-05 04:46:17 <FROTUSCI> cool
 269 2012-02-05 04:46:29 <luke-jr> looks like I need to upgrade my host-KVM first anyway… one I had installed doesn't support nesting
 270 2012-02-05 04:47:04 <sipa> it does now?
 271 2012-02-05 04:47:13 <luke-jr> it should O.o
 272 2012-02-05 04:47:15 <sipa> nice
 273 2012-02-05 04:47:36 <luke-jr> I know I saw it in kernel changelog :p
 274 2012-02-05 04:48:53 <gmaxwell> hm. inability to nest is one reason I haven't moved my entire system into kvm... pretty cool.
 275 2012-02-05 04:51:47 <luke-jr> it's not working <.<
 276 2012-02-05 04:52:01 <luke-jr> latest kvm got rid of --enable-nested and I can't figure out the cmdline option that replaced it yet
 277 2012-02-05 04:52:46 <sipa> i read this somewhere: modprobe kvm_amd nested=1
 278 2012-02-05 04:53:29 <graingert> ส็็็็็็็็็็็็็็็็็็็็็็็็็
 279 2012-02-05 04:53:47 <gmaxwell> graingert: you don't say?
 280 2012-02-05 04:54:06 <graingert> gmaxwell: it's very odd char
 281 2012-02-05 04:55:11 <luke-jr> sipa: I use Intel
 282 2012-02-05 04:55:12 <luke-jr> :p
 283 2012-02-05 04:55:22 <luke-jr> but it's the KVM cmd line that's the problme
 284 2012-02-05 04:55:25 <FROTUSCI> downton abbey
 285 2012-02-05 04:56:55 booo has quit (Ping timeout: 260 seconds)
 286 2012-02-05 04:57:45 <luke-jr> k, got it
 287 2012-02-05 04:57:57 <luke-jr> I had disabled nested=1 on the kernel module for some reason; now it just works
 288 2012-02-05 04:58:15 <luke-jr> let's see if 2 GB RAM is enough to do this all on LiveCD
 289 2012-02-05 04:58:44 <luke-jr> since qemu apparently can't do more
 290 2012-02-05 05:00:06 <luke-jr> I'm impressed with how Ubuntu boots fairly fast, and autodetects qemu :D
 291 2012-02-05 05:00:50 <luke-jr> hmm, gitian requires 64-bit? -.-
 292 2012-02-05 05:00:55 <luke-jr> Ubuntu recommended 32-bit
 293 2012-02-05 05:01:18 <sipa> don't think you can run a 64-bit guest on a 32-bit host
 294 2012-02-05 05:01:33 <sipa> and gitian uses a 64-bit guest to do the 64-bit builds
 295 2012-02-05 05:03:56 <luke-jr> what if I only want 32-bit builds? <.<
 296 2012-02-05 05:04:05 <sipa> hack the script, i guess
 297 2012-02-05 05:04:17 <luke-jr> didn't think we supported win64
 298 2012-02-05 05:04:43 Cablesaurus has quit (Quit: We be chillin - IceChat style)
 299 2012-02-05 05:04:44 <sipa> we don't
 300 2012-02-05 05:12:05 sacarlson has quit (Ping timeout: 260 seconds)
 301 2012-02-05 05:13:02 sacarlson has joined
 302 2012-02-05 05:19:40 Guest82933 has quit (Ping timeout: 260 seconds)
 303 2012-02-05 05:20:11 minimoose has joined
 304 2012-02-05 05:25:11 <graingert> sipa: you can in some cases
 305 2012-02-05 05:25:16 <graingert> 64 on 32
 306 2012-02-05 05:25:28 <sipa> ?
 307 2012-02-05 05:25:31 <sipa> ah
 308 2012-02-05 05:25:32 <graingert> Virtualbox
 309 2012-02-05 05:25:47 <sipa> without cpu emulation?
 310 2012-02-05 05:25:55 <graingert> yup
 311 2012-02-05 05:26:32 <sipa> even when the host cpu doesn't support 64 bit?
 312 2012-02-05 05:26:56 <graingert> not sure, only used it when the host cpu supports
 313 2012-02-05 05:27:00 <graingert> but os does not
 314 2012-02-05 05:27:41 <graingert> http://geekswithblogs.net/twickers/archive/2009/02/06/129243.aspx
 315 2012-02-05 05:27:48 <graingert> "requires additional overhead"
 316 2012-02-05 05:28:07 Clipse has quit (Quit: Clipse)
 317 2012-02-05 05:29:38 graingert has quit (Read error: Connection reset by peer)
 318 2012-02-05 05:30:04 merde has quit (Ping timeout: 245 seconds)
 319 2012-02-05 05:31:14 osmosis has quit (Quit: Leaving)
 320 2012-02-05 05:32:18 jamescarr has joined
 321 2012-02-05 05:33:11 merde has joined
 322 2012-02-05 05:37:15 Clipse has joined
 323 2012-02-05 05:50:46 helo_ has quit (Quit: leaving)
 324 2012-02-05 05:54:05 Rabbit67890 has quit (Quit: Rabbit67890)
 325 2012-02-05 06:11:56 da2ce7 has quit (Quit: KVIrc 4.1.3 Equilibrium http://www.kvirc.net/)
 326 2012-02-05 06:19:49 osmosis has joined
 327 2012-02-05 06:22:12 b4epoche_ has joined
 328 2012-02-05 06:22:12 da2ce7 has joined
 329 2012-02-05 06:23:04 b4epoche has quit (Ping timeout: 252 seconds)
 330 2012-02-05 06:23:31 b4epoche_ is now known as b4epoche
 331 2012-02-05 06:24:15 <devrandom> mmmmmm... nested kvm
 332 2012-02-05 06:24:36 <luke-jr> devrandom: is gitian intentionally Ubuntu-only? :/
 333 2012-02-05 06:24:50 <luke-jr> for host OS
 334 2012-02-05 06:28:13 <gmaxwell> sad: ozco.in (mining pool) was being robbed by that same RPC bruteforcer that there was a thread about on the forum recently.
 335 2012-02-05 06:28:33 <gmaxwell> Several hundred BTC stolen... actively siphoned as the blocks matured.
 336 2012-02-05 06:29:00 <sipa> maybe time to have a delay before disconnecting if an RPC client gives a wrong password
 337 2012-02-05 06:29:41 <gmaxwell> sipa: we have one— which does little more than tell you that the password is worth bruitforcing or not. :(
 338 2012-02-05 06:30:04 <luke-jr> wtf, another one?
 339 2012-02-05 06:30:12 <gmaxwell> (there is a delay which depends on the length of your rpcpassword)
 340 2012-02-05 06:30:19 <luke-jr> how are they even getting to the RPC port?
 341 2012-02-05 06:30:39 <gmaxwell> I don't know for sure if ozco.in was via rpc.. thats what the same attacker did to m3ta though.
 342 2012-02-05 06:30:58 <gmaxwell> so I think it's likely.
 343 2012-02-05 06:31:15 * luke-jr ponders building a bitcoind w/o send*
 344 2012-02-05 06:31:22 <gmaxwell> ozco.in appears to have left it running while they were getting robbed... kinda freaky.
 345 2012-02-05 06:32:13 Diablo-D3 has joined
 346 2012-02-05 06:32:22 <gmaxwell> sipa: in any case, while rpc is single threaded any delay means that it's a dos attack vector.
 347 2012-02-05 06:33:14 <sipa> well you shouldn't expose the RPC port in the first place
 348 2012-02-05 06:33:20 <sipa> but why do we have a password then
 349 2012-02-05 06:33:22 <gmaxwell> sipa: I think I prefer instead to take N attempts in M interval from an IP and then start returning 'nope' to every RPC call for a few minutes... this way it's only a DOS if they can attempt from your own systems.
 350 2012-02-05 06:33:22 <luke-jr> any bad effects if I enable wallet encryption on Eligius?
 351 2012-02-05 06:33:46 <gmaxwell> sipa: well yea, you shouldn't.. but people make mistakes.
 352 2012-02-05 06:34:02 <luke-jr> besides being permanently unable to unlock it?
 353 2012-02-05 06:34:13 <gmaxwell> sipa: one thing we should do is autogenerate a rpcuser/rpcpassword when we create a config file.. and hopefully people will leave the auto ones in.
 354 2012-02-05 06:34:30 <gmaxwell> luke-jr: it can't add to the keypool without the encryption key.
 355 2012-02-05 06:34:44 <gmaxwell> luke-jr: but that may not matter for your usage.
 356 2012-02-05 06:34:45 <luke-jr> gmaxwell: I use a static address for mining now
 357 2012-02-05 06:35:07 <gmaxwell> great. then there should be no harm. (a backup is always advised first, of course)
 358 2012-02-05 06:37:07 <gmaxwell> I kinda wonder if there shouldn't be code with wallet encryption to limit the unlock to the IP address that called the unlock rpc.
 359 2012-02-05 06:37:22 <sipa> gmaxwell: one other advantage of determinstic wallets... you can add keys while locked
 360 2012-02-05 06:37:37 <gmaxwell> because even if ozco.in had been using encrpytion, the attacker could just poll like crazy and then outrace the unlock.
 361 2012-02-05 06:38:02 <gmaxwell> I guess if your rpc is insecure you've already lost.
 362 2012-02-05 06:38:09 <gmaxwell> sipa: Indeed.
 363 2012-02-05 06:38:53 <sipa> this would be so much easier to implement if i could throw the old wallet's key handling stuff out
 364 2012-02-05 06:41:19 minimoose has quit (Quit: minimoose)
 365 2012-02-05 06:42:21 dissipate has joined
 366 2012-02-05 06:42:21 dissipate has quit (Changing host)
 367 2012-02-05 06:42:21 dissipate has joined
 368 2012-02-05 06:42:50 BTC_Bear is now known as BTC_Bear|hbrntng
 369 2012-02-05 06:56:53 Cablesaurus has joined
 370 2012-02-05 06:56:54 Cablesaurus has quit (Changing host)
 371 2012-02-05 06:56:54 Cablesaurus has joined
 372 2012-02-05 06:58:50 Rabbit67890 has joined
 373 2012-02-05 07:03:44 dissipate has quit (Remote host closed the connection)
 374 2012-02-05 07:08:13 dirus- has quit (Quit: WeeChat 0.3.5)
 375 2012-02-05 07:15:54 da2ce7 has quit (Remote host closed the connection)
 376 2012-02-05 07:17:05 dissipate has joined
 377 2012-02-05 07:19:14 da2ce7 has joined
 378 2012-02-05 07:20:47 Rabbit67890 has quit (Quit: Rabbit67890)
 379 2012-02-05 07:32:22 <Diablo-D3> gmaxwell, sipa: what happened to -server on gui?
 380 2012-02-05 07:32:30 <sipa> nothing?
 381 2012-02-05 07:32:35 <gmaxwell> it's still there.
 382 2012-02-05 07:32:43 <gmaxwell> Just not in help (was it ever?)
 383 2012-02-05 07:32:46 <Diablo-D3> so why does it not start it up?
 384 2012-02-05 07:32:54 <sipa> anything in debug.log?
 385 2012-02-05 07:32:58 <gmaxwell> you're not setting a user/password?
 386 2012-02-05 07:33:20 <Diablo-D3> gmaxwell: yes I am
 387 2012-02-05 07:33:27 <Diablo-D3> I see nothing useful in debug.log
 388 2012-02-05 07:33:46 <Diablo-D3> oh wait, wtf
 389 2012-02-05 07:33:48 <Diablo-D3> now its working
 390 2012-02-05 07:33:52 <Diablo-D3> I havent touched a thing
 391 2012-02-05 07:34:01 <sipa> was it still loading the block chain index?
 392 2012-02-05 07:34:56 <Diablo-D3> no, its been awake for days
 393 2012-02-05 07:41:22 dissipate has quit (Remote host closed the connection)
 394 2012-02-05 07:41:46 helo_ has joined
 395 2012-02-05 07:43:27 <helo_> txparam = "txdata=" *pchar
 396 2012-02-05 07:45:03 <helo_> ^ add to bip 21 okplzthx
 397 2012-02-05 07:46:30 <sipa> ?
 398 2012-02-05 07:46:49 <sipa> what would it mean?
 399 2012-02-05 07:47:46 ThomasV has joined
 400 2012-02-05 07:48:39 <helo_> probably bs, but it could allow people to embed entire transaction payloads in a URI to make it easy for others to send/verify a transaction
 401 2012-02-05 07:49:29 <sipa> the entire serialized transaction?
 402 2012-02-05 07:50:12 <helo_> this isn't a very practical application, but if you wanted to give someone bitcoin for their birthday, or allow them to withdraw bitcoin at their leisure, you could send them an email with the transaction for them to click and broadcast themselves
 403 2012-02-05 07:50:16 <helo_> yeah
 404 2012-02-05 07:50:23 splatster has quit (Quit: Na na na na na na na na na na na na na na na na... SPLATMAN!!!)
 405 2012-02-05 07:51:34 <sipa> it'd probably have to be req-txdata
 406 2012-02-05 07:52:52 <helo_> or i want to be sure i have access to enough money to fly me home should something bad happen, but dodn't want to have the balance on any address that i'm in possesion of... or something
 407 2012-02-05 07:53:22 <helo_> or my dad wants me to have access to money while i'm away at college should the need arise
 408 2012-02-05 07:54:14 <helo_> but doesn't want a virus to be able to steal it from my wallet (passphrase sniffing) when i decrypt it to pay rent each month
 409 2012-02-05 07:56:29 <helo_> i'm betting there is a better way to achieve those aims, but there are other applications too, and it would be kind of a different mechanism to be commonly accessible
 410 2012-02-05 07:57:04 <helo_> (if it was officially supported by the main client)
 411 2012-02-05 07:57:38 <helo_> could also have, instead of sending directly, a checkbox for "generate url containing transaction payload for sending later"
 412 2012-02-05 07:57:46 <gribble> New news from bitcoinrss: gmaxwell opened pull request 798 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/798>
 413 2012-02-05 07:58:21 <helo_> ahh good gmax is awake... i'm sure he'll take the opportunity to shoot this idea down properly :)
 414 2012-02-05 08:00:26 <helo_> could also allow other applications to send bitcoin transactions without implementing the networking side of actually sending a transaction into the network
 415 2012-02-05 08:01:01 <sipa> just sending a transaction is quite trivial, though
 416 2012-02-05 08:01:15 <helo_> but a web browser can't do it by itself
 417 2012-02-05 08:01:37 <sipa> true
 418 2012-02-05 08:01:38 <gmaxwell> helo_: I saw it, but thought it was kinda silly.
 419 2012-02-05 08:01:40 <helo_> javascript app or whatever
 420 2012-02-05 08:01:47 <helo_> it is...
 421 2012-02-05 08:02:06 <gmaxwell> But it's not my place to shoot down ideas that are merely silly.
 422 2012-02-05 08:02:25 <helo_> as long as it's not dangerous like most of my ideas :)
 423 2012-02-05 08:04:07 <helo_> hmm... you could generate a transaction, and completely destroy the wallet the funds belong to
 424 2012-02-05 08:04:28 <Graet> ^^ right now i like this idea
 425 2012-02-05 08:04:36 <Graet> qucik tell me how!!
 426 2012-02-05 08:05:09 <helo_> save transaction data transferring all of the funds (currently known) to exist at an address
 427 2012-02-05 08:05:16 <helo_> instead of sending it
 428 2012-02-05 08:05:54 <helo_> then destroy the wallet, and put the transaction into a safety deposit box
 429 2012-02-05 08:06:01 <helo_> if you want...
 430 2012-02-05 08:06:22 <helo_> kind of once-removed offline storage
 431 2012-02-05 08:06:56 <helo_> you could create a transaction moving an amount of bitcoin that doesn't currently reside at an address
 432 2012-02-05 08:07:09 <gmaxwell> Darnit how the hell do I update a pull request.
 433 2012-02-05 08:07:16 <helo_> to be able to used in the future when it does
 434 2012-02-05 08:07:27 <sipa> gmaxwell: just push to the branch the pull request uses
 435 2012-02-05 08:08:01 <bd_> helo_: why not just send to a special, one-time-use address and keep the private key in a secure location?
 436 2012-02-05 08:08:55 <helo_> that is the normal way to do things, yes
 437 2012-02-05 08:09:04 <gmaxwell> sipa: okay, I did that but the pull request didn't update.
 438 2012-02-05 08:09:17 <sipa> is it closed?
 439 2012-02-05 08:09:31 <gmaxwell> No.
 440 2012-02-05 08:09:55 <gmaxwell> https://github.com/gmaxwell/bitcoin/commit/b04f301c8edb0d062864af58e20a65079f9624b7  is the head of the branch
 441 2012-02-05 08:10:16 <gmaxwell> (I screwed up and missed commiting part of my change because I did --amend without -a and only updated the commit message)
 442 2012-02-05 08:10:28 <gmaxwell> pull request still shows https://github.com/gmaxwell/bitcoin/commit/9d33dc71cfbfc89e89284338c691d1e104c60665
 443 2012-02-05 08:10:37 <sipa> you updated rpcpassword, but the pullreq uses advertise0fixes
 444 2012-02-05 08:10:37 <gmaxwell> (the difference is that its missing the logging part of the change)
 445 2012-02-05 08:10:38 <helo_> maybe it is true that multiple weird newish kinds of arrangements/scenarios could be made accessible if transaction payload could be encoded into a URI
 446 2012-02-05 08:10:47 <gmaxwell> bleh.
 447 2012-02-05 08:10:54 <helo_> and maybe some of these would be useful in some situations :)
 448 2012-02-05 08:11:43 josephcp has quit (Ping timeout: 252 seconds)
 449 2012-02-05 08:12:38 <gmaxwell> sipa: thanks. I fail at github. (fixing)
 450 2012-02-05 08:12:53 [eval] has quit (Read error: Operation timed out)
 451 2012-02-05 08:13:18 <helo_> or maybe if you were to die, nobody would know your passphrase, and all of your money would be gone
 452 2012-02-05 08:13:42 josephcp has joined
 453 2012-02-05 08:14:32 molecular has quit (Ping timeout: 272 seconds)
 454 2012-02-05 08:15:14 <helo_> so you create a transaction sending your savings to your lawyer to be used in the event that you die
 455 2012-02-05 08:15:46 molecular has joined
 456 2012-02-05 08:16:17 <sipa> i think this is useful, but i think it's more appropriate to have a file format for it, and the ability to export/import it
 457 2012-02-05 08:17:06 <helo_> yeah, possibly
 458 2012-02-05 08:18:04 <sipa> or an RPC call
 459 2012-02-05 08:18:05 <gribble> New news from bitcoinrss: gmaxwell opened pull request 799 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/799>
 460 2012-02-05 08:19:20 <gmaxwell> helo_: if you were going to create such a recovery thing... why not just have the password in it?
 461 2012-02-05 08:20:53 <helo_> maybe your house burns down with you and your private key backups in it
 462 2012-02-05 08:22:22 <helo_> would URI support cover any ground that a file format would not?
 463 2012-02-05 08:26:41 marf_away has joined
 464 2012-02-05 08:28:22 <helo_> a bitcoin transaction is kind of like a check you have written
 465 2012-02-05 08:29:46 <sipa> how would you store the URL?
 466 2012-02-05 08:31:32 <helo_> in an html file with javascript to load the URI when it is opened ;)
 467 2012-02-05 08:33:12 FROTUSCI has quit (Ping timeout: 248 seconds)
 468 2012-02-05 08:33:56 <helo_> URI would apparently just be a method to store a transaction in a very awkward way Oo
 469 2012-02-05 08:51:18 merde has quit (Ping timeout: 255 seconds)
 470 2012-02-05 08:52:25 RazielZ has joined
 471 2012-02-05 08:56:22 danbri has joined
 472 2012-02-05 08:58:55 merde has joined
 473 2012-02-05 09:13:48 Katniss has quit (Ping timeout: 255 seconds)
 474 2012-02-05 09:15:47 helo_ has quit (Quit: leaving)
 475 2012-02-05 09:19:32 Katniss has joined
 476 2012-02-05 09:34:34 Cablesaurus has quit (Quit: Few women admit their age.  Few men act theirs.)
 477 2012-02-05 09:40:41 sje has joined
 478 2012-02-05 09:43:21 Katniss has quit (Ping timeout: 260 seconds)
 479 2012-02-05 09:47:38 Katniss has joined
 480 2012-02-05 10:02:32 ThomasV has quit (Ping timeout: 256 seconds)
 481 2012-02-05 10:03:50 <Eliel> helo: there's a few factors that make the transaction approach somewhat inconvenient. 1) the recipient has to have a wallet 2) If you want the transaction to include a whole wallet, the transaction would need to be remade every time money is added to the wallet.
 482 2012-02-05 10:04:13 <Eliel> helo: oh and 3) you have to know the recipient address well in advance.
 483 2012-02-05 10:04:40 <Eliel> ... and 4) you have to have the bitcoins in advance.
 484 2012-02-05 10:09:43 Guest13384 has quit (Ping timeout: 248 seconds)
 485 2012-02-05 10:13:54 ivan` has quit (Quit: ERC Version 5.3 (IRC client for Emacs))
 486 2012-02-05 10:14:26 jamescarr has quit (Ping timeout: 256 seconds)
 487 2012-02-05 10:16:24  has joined
 488 2012-02-05 10:17:04 pusle has joined
 489 2012-02-05 10:18:03 ivan` has joined
 490 2012-02-05 10:22:37 bitlad has joined
 491 2012-02-05 10:27:04 bitlad has quit (Remote host closed the connection)
 492 2012-02-05 10:33:35 bitlad has joined
 493 2012-02-05 10:33:41 copumpkin has quit (Ping timeout: 252 seconds)
 494 2012-02-05 10:34:18 copumpkin has joined
 495 2012-02-05 10:34:37 b4epoche_ has joined
 496 2012-02-05 10:35:51 b4epoche has quit (Ping timeout: 260 seconds)
 497 2012-02-05 10:35:51 b4epoche_ is now known as b4epoche
 498 2012-02-05 10:52:39 Joric has joined
 499 2012-02-05 10:52:39 Joric has quit (Changing host)
 500 2012-02-05 10:52:39 Joric has joined
 501 2012-02-05 10:53:32 diki has joined
 502 2012-02-05 10:54:17 <diki> A few months ago I wrote a...php proxy for bitcoin, the miner connected to the proxy but sadly, even though I was sending the correct json headers...nothing happened
 503 2012-02-05 10:54:25 <diki> my question is, was it Apache's fault?
 504 2012-02-05 10:55:14 <diki> Since a miner has only a few specified headers which curl reports, where apache has a few more
 505 2012-02-05 10:58:17 bitlad has quit (Remote host closed the connection)
 506 2012-02-05 11:04:07 copumpkin has quit (Ping timeout: 248 seconds)
 507 2012-02-05 11:04:35 copumpkin has joined
 508 2012-02-05 11:07:02 cdecker has joined
 509 2012-02-05 11:13:40 <gribble> New news from bitcoinrss: laanwj opened pull request 800 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/800>
 510 2012-02-05 11:15:13 <makomk> luke-jr: you can run gitian on Gentoo too - I do - but it's pretty much undocumented and a bit fiddly.
 511 2012-02-05 11:26:16 Nicksasa has quit (Ping timeout: 240 seconds)
 512 2012-02-05 11:26:59 erle- has joined
 513 2012-02-05 11:38:10 MrTiggr has quit (Ping timeout: 252 seconds)
 514 2012-02-05 11:46:01 Turingi has joined
 515 2012-02-05 11:46:01 Turingi has quit (Changing host)
 516 2012-02-05 11:46:01 Turingi has joined
 517 2012-02-05 11:46:55 ivan` has quit (Quit: ERC Version 5.3 (IRC client for Emacs))
 518 2012-02-05 11:47:27 MrTiggr has joined
 519 2012-02-05 11:48:09 iocor has joined
 520 2012-02-05 11:48:10 danbri has quit (Ping timeout: 244 seconds)
 521 2012-02-05 11:50:53 sacarlson has quit (Quit: Leaving.)
 522 2012-02-05 11:57:25 Nicksasa has joined
 523 2012-02-05 11:57:41 ivan` has joined
 524 2012-02-05 11:57:55 ivan` has quit (Excess Flood)
 525 2012-02-05 12:00:20 ivan` has joined
 526 2012-02-05 12:00:48 larsivi has quit (Read error: Connection reset by peer)
 527 2012-02-05 12:04:33 larsivi has joined
 528 2012-02-05 12:04:45 danbri has joined
 529 2012-02-05 12:25:34 gjs278 has quit (Remote host closed the connection)
 530 2012-02-05 12:26:04 gjs278 has joined
 531 2012-02-05 12:28:55 cryptoxchange has quit (Ping timeout: 248 seconds)
 532 2012-02-05 12:38:14 Zarutian has joined
 533 2012-02-05 12:50:31 paraipan has quit (Quit: Saliendo)
 534 2012-02-05 12:51:53 occulta has joined
 535 2012-02-05 12:58:30 phantomfake has quit (Read error: Connection reset by peer)
 536 2012-02-05 13:01:36 booo has joined
 537 2012-02-05 13:07:47 paraipan has joined
 538 2012-02-05 13:11:24 datagutt has joined
 539 2012-02-05 13:11:37 Nicksasa has quit (Ping timeout: 240 seconds)
 540 2012-02-05 13:23:01 <osmosis> it would be nice if bitcoin-qt showed and allowed a sort by creation datetime on the list of addresses
 541 2012-02-05 13:27:23 <osmosis> whenever I create a new address, i got back to look for it and its lost in alphabetical order
 542 2012-02-05 13:28:28 storrgie has joined
 543 2012-02-05 13:28:37 slush has joined
 544 2012-02-05 13:37:59 denisx has joined
 545 2012-02-05 13:38:36 booo has quit (Ping timeout: 240 seconds)
 546 2012-02-05 13:38:44 slush has quit (Quit: Leaving.)
 547 2012-02-05 13:40:20 gjs278 has quit (Read error: Connection reset by peer)
 548 2012-02-05 13:45:00 gjs278 has joined
 549 2012-02-05 13:47:29 paul0 has joined
 550 2012-02-05 13:51:47 baz has joined
 551 2012-02-05 13:53:41 <luke-jr> makomk: how? python-vm-builder looks liek an Ubuntu-specific thing
 552 2012-02-05 13:57:49 finway has joined
 553 2012-02-05 13:58:22 <finway> Is libcoin's initial downloading really that fast ? 3.5 times faster ?
 554 2012-02-05 13:58:34 <finway> we should import that feature.
 555 2012-02-05 13:59:32 <luke-jr> finway: slower IIRC
 556 2012-02-05 13:59:54 <finway> luke-jr: i guess you're right!   java can't be faster...
 557 2012-02-05 14:01:21 <finway> And what do you guys think about this: http://marc.info/?l=full-disclosure&m=132810929830371&w=2
 558 2012-02-05 14:01:34 <finway> Is this really dangerous ?
 559 2012-02-05 14:01:58 <finway> Skip verifying transaction signatures during initial block-chain down… …load
 560 2012-02-05 14:01:59 <Diablo-D3> if thats the url Im thinking of, go read the replies
 561 2012-02-05 14:02:06 <cjd> biggest risk is realsolid's goons spamming it around causing fud
 562 2012-02-05 14:02:37 denisx_ has joined
 563 2012-02-05 14:03:13 denisx_ has quit (Remote host closed the connection)
 564 2012-02-05 14:03:29 denisx_ has joined
 565 2012-02-05 14:03:58 <finway> Oh shit,  Dan Kaminsky replied to this.
 566 2012-02-05 14:04:10 <cjd> it's "unexploitable"
 567 2012-02-05 14:05:57 denisx has quit (Ping timeout: 240 seconds)
 568 2012-02-05 14:05:57 denisx_ is now known as denisx
 569 2012-02-05 14:06:07 osmosis has quit (Quit: Leaving)
 570 2012-02-05 14:06:45 <finway> I think _exploiting_bitcoin_network _get_all_money are all hackers' wet dream...
 571 2012-02-05 14:07:41 <cjd> I find that cleverness and morals come together
 572 2012-02-05 14:09:32 storrgie has quit (Ping timeout: 245 seconds)
 573 2012-02-05 14:09:41 baz has quit (Remote host closed the connection)
 574 2012-02-05 14:09:57 <finway> So we get Initial downloading faster by skiping ECDSA checking before the CHECKPOINT ?
 575 2012-02-05 14:13:27 <finway> What a bad idea!
 576 2012-02-05 14:13:38 <luke-jr> finway: libcoin isn't Java…
 577 2012-02-05 14:13:52 <finway> luke-jr: oh...
 578 2012-02-05 14:14:21 <cjd> actually it's a good idea
 579 2012-02-05 14:14:25 <justmoon> finway: the exploit didn't regard skipping verification below the checkpoints, but verification of quickly received blocks
 580 2012-02-05 14:14:31 <cjd> and a better idea is to simply not download the blocks at all
 581 2012-02-05 14:14:41 <luke-jr> finway: the ECDSA checking is redundant because there are *already other checks* in place
 582 2012-02-05 14:14:52 <cjd> just download the internal state of the program at that checkpoint
 583 2012-02-05 14:15:20 <cjd> but my understanding is it doesn't represent it's state in a way that it would work to do that
 584 2012-02-05 14:15:28 <cjd> so making that change would be expensive
 585 2012-02-05 14:16:46 <justmoon> cjd: it would also not be much faster
 586 2012-02-05 14:18:11 <cjd> that also depends
 587 2012-02-05 14:18:19 <justmoon> if all you're doing is checking sha hashes and creating indexes and your database handles bulk inserts well it's really not much of a difference performance wise between a data import and copying the database
 588 2012-02-05 14:18:49 * cjd realizes he's suggesting massive changes to the internals of the program so this is purely theoretical...
 589 2012-02-05 14:19:16 <justmoon> cjd: it's not, bitcoinjs has already created two custom database backends and I'm working on a third one
 590 2012-02-05 14:19:53 <cjd> don't store the unspent transactions themselves, just store bitfields representing the outputs where a set bit represents an unspent output and a cleared bit represents a spent/nonexistant one
 591 2012-02-05 14:20:45 <cjd> when someone wants to spend a transaction, use the existing protocol to make them fork over the transaction so you can check it
 592 2012-02-05 14:20:54 <finway> luke-jr,justmoon: Since you say so, i feel good now.
 593 2012-02-05 14:21:57 <justmoon> cjd: that would increase network load dramatically - arguably network load is more critical than storage or CPU
 594 2012-02-05 14:22:00 <luke-jr> finway: it was discussed on the 0.5.2 thread ;)
 595 2012-02-05 14:22:13 <finway> luke-jr: thanks.
 596 2012-02-05 14:22:15 <cjd> also I'm kind of talking out my ass because I don't entirely know how transactions are hashed into blocks
 597 2012-02-05 14:22:58 <cjd> the thing about network load is gmax can setup his satellite transponder and bcast these transacrtions out to the world
 598 2012-02-05 14:23:12 <cjd> with storage, there's no way around it, everyone has to store everything
 599 2012-02-05 14:23:38 <cjd> and everyone-stores-everything won't scale nomatter what kind of cleverness you use
 600 2012-02-05 14:25:18 <justmoon> cjd: I've been working on a merkle tree of unspent outputs proposal for a while, it's unfinished, but you can take a look here: https://en.bitcoin.it/wiki/User:Justmoon/IMTUO
 601 2012-02-05 14:26:23 <justmoon> under my proposal, nodes (even full mining, verifying nodes) only store recent blocks
 602 2012-02-05 14:27:31 <luke-jr> justmoon: well, that's only viable if users trust the miners of the past ;)
 603 2012-02-05 14:27:39 <luke-jr> so full verification still requires the complete block chain
 604 2012-02-05 14:27:41 <cjd> Bitfield indicating where the side branches attach (0 - prefix; 1 - suffix);  <-- can't you just do 2 memcmp() ?
 605 2012-02-05 14:27:43 <finway> justmoon: so new full nodes should trust other nodes ?
 606 2012-02-05 14:27:59 <justmoon> luke-jr: we already trust the miners of the past - hence no ecdsa verification
 607 2012-02-05 14:28:18 denisx_ has joined
 608 2012-02-05 14:28:41 <cjd> "Currently, Bitcoin sends all transactions again when a block is transmitted, even though the receiving node may know most of them already. We propose that blocks be transmitted including a list of transaction hashes and missing transactions be requested as needed." <-- changes to the actual protocol har tough
 609 2012-02-05 14:28:42 <justmoon> finway: no, new nodes don't any trust other than that blocks older than 1024 blocks ago are valid
 610 2012-02-05 14:29:53 <justmoon> cjd: protocol changes are the easy part - the hard part is getting a trustworthy MTUO hash in the coinbase - if only a minority of miners participate you can't enforce its validity
 611 2012-02-05 14:30:10 <cjd> RIPEMD160 <-- note that merkle trees are much easier to attack than bitcoin addresses so ripemd160 may not be enough
 612 2012-02-05 14:30:29 <justmoon> cjd: interesting, how so?
 613 2012-02-05 14:30:30 <cjd> with a bitcoin address, you need to generate a key for each guess
 614 2012-02-05 14:30:45 <luke-jr> justmoon: no, the ECDSA verification is only disabled when redundant
 615 2012-02-05 14:30:57 <justmoon> cjd: no, you can increment a single key
 616 2012-02-05 14:31:23 <cjd> with a merkle tree you can add phony transactions and use them as nonces to attack the hash directly
 617 2012-02-05 14:31:37 denisx has quit (Ping timeout: 240 seconds)
 618 2012-02-05 14:31:37 denisx_ is now known as denisx
 619 2012-02-05 14:31:50 <justmoon> luke-jr: I don't follow
 620 2012-02-05 14:31:53 <cjd> it's mostly academic because 160, 256, whocares
 621 2012-02-05 14:32:34 <luke-jr> justmoon: the ECDSA signatures are only skipped because the checkpoints ALREADY check they've validated
 622 2012-02-05 14:32:35 <cjd> A lightweight client stores:
 623 2012-02-05 14:32:36 <cjd> Block headers
 624 2012-02-05 14:32:46 <cjd> ^^recent block headers should be enough
 625 2012-02-05 14:33:08 <justmoon> luke-jr: yes, but ecdsa is also skipped beyond the last checkpoint for blocks that are received in quick succession
 626 2012-02-05 14:33:34 <justmoon> luke-jr: see that mailing list post finway just posted: http://marc.info/?l=full-disclosure&m=132810929830371&w=2
 627 2012-02-05 14:33:54 <justmoon> seems to be in current master as well: https://github.com/bitcoin/bitcoin/blob/master/src/main.cpp#L886
 628 2012-02-05 14:34:18 <justmoon> now, I have no idea if I understand this correct, since I haven't been here for the change or for the discussion of it
 629 2012-02-05 14:34:19 luke-jr has quit (otg!~luke-jr@2001:470:5:265:222:4dff:fe50:4c49|Read error: Connection reset by peer)
 630 2012-02-05 14:34:54 luke-jr has quit (Read error: Connection reset by peer)
 631 2012-02-05 14:34:55 <justmoon> cjd: you're right
 632 2012-02-05 14:35:47 <cjd> A full client stores:
 633 2012-02-05 14:35:47 <cjd> Block headers
 634 2012-02-05 14:35:47 <cjd> Coinbases
 635 2012-02-05 14:35:47 <cjd> Merkle branch connecting coinbase tx to merkle root
 636 2012-02-05 14:35:48 <cjd> Recent blocks (full blocks including vinfo)
 637 2012-02-05 14:35:59 <cjd> also needs to store all unspent transactions / outputs
 638 2012-02-05 14:36:13 <justmoon> cjd: no it doesn't, that's the whole point of the proposal
 639 2012-02-05 14:36:14 luke-jr has joined
 640 2012-02-05 14:36:49 <cjd> so you only store merkle branches to transactions which are your money?
 641 2012-02-05 14:37:28 <justmoon> yes, I imagine there would be a dht where you could look it up if you needed to also
 642 2012-02-05 14:37:47 <cjd> ok
 643 2012-02-05 14:37:53 luke-jr has joined
 644 2012-02-05 14:37:56 <cjd> this means you need a binary merkle tree
 645 2012-02-05 14:38:08 <luke-jr> it seems nested VMX isn't very stable :/
 646 2012-02-05 14:38:08 <justmoon> as opposed to?
 647 2012-02-05 14:38:57 <cjd> if you just made a list and created a merkle tree from it then when you added a transaction, it would cause the pairs which were hashed together to shift be one and break everything
 648 2012-02-05 14:39:20 <phantomcircuit> luke-jr, it's not at all
 649 2012-02-05 14:39:32 <justmoon> cjd: please actually read the section on how hashes are added and removed from the merkle tree
 650 2012-02-05 14:39:46 * cjd works from top down
 651 2012-02-05 14:39:51 <luke-jr> phantomcircuit: not even if I upgrade to 3.2.2?
 652 2012-02-05 14:39:59 <phantomcircuit> :)
 653 2012-02-05 14:40:07 <phantomcircuit> luke-jr, nested vmx is hard
 654 2012-02-05 14:40:11 <justmoon> cjd: kk, sorry :)
 655 2012-02-05 14:40:16 <phantomcircuit> yo dawg i heard you like virtualization...
 656 2012-02-05 14:40:27 <luke-jr> phantomcircuit: well Linux shouldn't kernel panic over it :P
 657 2012-02-05 14:40:33 <justmoon> phantomcircuit: rofl I was just typing that same joke
 658 2012-02-05 14:43:16 <cjd> looks like you're making something like a binary hash tree
 659 2012-02-05 14:43:29 <cjd> I don't fully understand the balancing algo though
 660 2012-02-05 14:43:33 baz has joined
 661 2012-02-05 14:44:13 <justmoon> cjd: not sure if balancing is misleading here, since it's not balancing like you would balance a binary merkle tree
 662 2012-02-05 14:44:17 <justmoon> i.e. where you move stuff around
 663 2012-02-05 14:44:22 datagutt has quit (Ping timeout: 244 seconds)
 664 2012-02-05 14:44:27 <justmoon> this is a totally unordered merkle tree
 665 2012-02-05 14:44:41 <justmoon> and the balancing happens simply because lower insertion points are preferred over higher ones
 666 2012-02-05 14:44:48 hexTech has joined
 667 2012-02-05 14:45:11 <cjd> my binary merkle tree design is not balanced at all but I can prove that no branch will run longer than 255 nodes
 668 2012-02-05 14:45:22 <cjd> because each node represents a bit in the hash...
 669 2012-02-05 14:45:40 b4epoche_ has joined
 670 2012-02-05 14:46:01 <justmoon> yes, in this design no branch will run over whatever the limit is - a sensible one would be 50% over average
 671 2012-02-05 14:46:07 <cjd> and that is a nice defense against someone who would send you infinite nodes which don't ever end at any root
 672 2012-02-05 14:46:22 b4epoche has quit (Ping timeout: 245 seconds)
 673 2012-02-05 14:46:23 b4epoche_ is now known as b4epoche
 674 2012-02-05 14:46:28 <cjd> hmm
 675 2012-02-05 14:46:29 <justmoon> so with a couple billion transactions the maximum merkle branch that is theoretically possible would be 50 or so
 676 2012-02-05 14:46:32 <phantomcircuit> cjd, the bitcoin merkle tree is a balanced tree
 677 2012-02-05 14:46:59 <phantomcircuit> the implementation isn't actually a tree
 678 2012-02-05 14:47:07 <justmoon> phantomcircuit: we're talking about something else: https://en.bitcoin.it/wiki/User:Justmoon/IMTUO
 679 2012-02-05 14:47:10 <phantomcircuit> it just could maybe be a tree
 680 2012-02-05 14:47:18 <cjd> phantomcircuit: like sort, take middle, first quarter, last quarter etc?
 681 2012-02-05 14:47:35 <cjd> I'm curious how it actually works
 682 2012-02-05 14:47:37 <phantomcircuit> cjd, yes
 683 2012-02-05 14:47:42 <phantomcircuit> the codes really simple
 684 2012-02-05 14:47:45 <cjd> and skeered of the .cpp files
 685 2012-02-05 14:50:09 <cjd> justmoon: http://btc.pastebay.org/144544 my proposal speaks to some of the problems you mentioned
 686 2012-02-05 14:50:13 Nicksasa has joined
 687 2012-02-05 14:50:27 <cjd> namely not having all miners on board
 688 2012-02-05 14:50:52 bitlad has joined
 689 2012-02-05 14:51:38 <justmoon> cjd: awesome, that's what I'm really struggling with
 690 2012-02-05 14:52:02 <roconnor> sipa, gmaxwell: I think I know how to turn 1000 confirmation coins back into 1 confirmation coins using duplicate transactions.
 691 2012-02-05 14:53:25 <luke-jr> oh well, since I lost my runtime, I'm going to shutdown and look at this stupid 5850
 692 2012-02-05 14:53:28 <luke-jr> bbiab
 693 2012-02-05 14:53:38 luke-jr has quit (Quit: Lost terminal)
 694 2012-02-05 14:53:41 luke-jr has quit (otg!~luke-jr@2001:470:5:265:222:4dff:fe50:4c49|Quit: ZNC - http://znc.sourceforge.net)
 695 2012-02-05 14:54:03 SomeoneWeird is now known as SomeoneWeirdzzz
 696 2012-02-05 14:54:17 <cjd> justmoon: I decided I don't like that solution that much though
 697 2012-02-05 14:54:41 <cjd> I'd rather store the headers and bitfields instead
 698 2012-02-05 14:54:48 <cjd> it's much easier for me
 699 2012-02-05 14:55:25 <cjd> also you can make the person who has the money store the proof of it
 700 2012-02-05 14:57:18 <cjd> I still need binary hash trees for dns and at the moment they're making a fool of me
 701 2012-02-05 14:57:49 da2ce7 has joined
 702 2012-02-05 14:58:37 bitlad has quit (Remote host closed the connection)
 703 2012-02-05 14:59:47 bitlad has joined
 704 2012-02-05 15:02:18 BTC_Bear is now known as hbrntng!~BTC_Bear@unaffiliated/btc-bear/x-5233302|BTC_Bear
 705 2012-02-05 15:04:13 finway has quit (Quit: Page closed)
 706 2012-02-05 15:06:23 <justmoon> cjd: I'm trying to understand your proposal - you talk about "unspent transaction" - there is no such thing, there are only unspent outputs
 707 2012-02-05 15:07:56 <justmoon> I think it would be possible to use the bitfield idea from your proposal and merge it with the output hashing (based on outpoint, value and script) from my proposal
 708 2012-02-05 15:12:06 eoss has joined
 709 2012-02-05 15:12:06 eoss has quit (Changing host)
 710 2012-02-05 15:12:06 eoss has joined
 711 2012-02-05 15:13:25 p0s has joined
 712 2012-02-05 15:14:42 luke-jr has joined
 713 2012-02-05 15:20:42 doublec has joined
 714 2012-02-05 15:21:04 doublec_ has quit (Ping timeout: 260 seconds)
 715 2012-02-05 15:24:47 Sedra has quit (Remote host closed the connection)
 716 2012-02-05 15:24:57 Sedra has joined
 717 2012-02-05 15:26:45 Sedra has quit (Remote host closed the connection)
 718 2012-02-05 15:27:32 da2ce7 has quit (2!~da2ce7@gateway/tor-sasl/da2ce7|Ping timeout: 276 seconds)
 719 2012-02-05 15:32:11 Sedra has joined
 720 2012-02-05 15:33:24 b4epoche has quit (Quit: Textual IRC Client: http://www.textualapp.com/)
 721 2012-02-05 15:34:13 datagutt has joined
 722 2012-02-05 15:34:41 devrandom has quit (Ping timeout: 276 seconds)
 723 2012-02-05 15:35:25 user_ has joined
 724 2012-02-05 15:38:31 Sedra- has joined
 725 2012-02-05 15:40:54 Moron__ has joined
 726 2012-02-05 15:41:18 BTC_Bear is now known as BTC_Bear|hbrntng
 727 2012-02-05 15:41:22 devrandom has joined
 728 2012-02-05 15:41:23 Sedra has quit (Ping timeout: 245 seconds)
 729 2012-02-05 15:42:59 <cjd> justmoon: yea, it should be "Unspent Output Trees", my fault there
 730 2012-02-05 15:45:37 pingdrive has joined
 731 2012-02-05 15:46:03 Guest82933 has joined
 732 2012-02-05 15:46:03 Guest82933 has quit (Read error: Connection reset by peer)
 733 2012-02-05 15:49:34 pingdrive has quit (Client Quit)
 734 2012-02-05 16:07:49 <gribble> New news from bitcoinrss: tcatm opened pull request 801 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/801>
 735 2012-02-05 16:09:03 <devrandom> luke-jr: it's just what was convenient.  I'm all for adding platforms.
 736 2012-02-05 16:11:02 splatster has joined
 737 2012-02-05 16:19:31 sje has quit (Remote host closed the connection)
 738 2012-02-05 16:48:00 dissipate has joined
 739 2012-02-05 16:56:33 TD has joined
 740 2012-02-05 16:56:41 <TD> evening
 741 2012-02-05 16:57:07 <justmoon> TD: guten abend
 742 2012-02-05 16:57:13 <TD> how's it going ?
 743 2012-02-05 16:57:33 <justmoon> just got a pull request adding actual features: https://github.com/zootreeves/bitcoinjs-lib/commit/b972e868c52b86af8dfbf7871d8b128f1e9f2360
 744 2012-02-05 16:57:41 <justmoon> exciting days
 745 2012-02-05 16:58:12 <TD> oh, cool!
 746 2012-02-05 16:58:17 <TD> i haven't done any multisig stuff yet
 747 2012-02-05 16:58:29 <justmoon> blockchain.info has it deployed already I believe
 748 2012-02-05 16:58:31 dissipate has quit (Remote host closed the connection)
 749 2012-02-05 16:59:06 <TD> yeah, i saw that today
 750 2012-02-05 16:59:21 <TD> they have some kind of checkmultisig ui, but you have to co-ordinate moving signatures and addresses around by hand
 751 2012-02-05 16:59:45 <justmoon> yes, but hey, if I get a pull request out of it, I'm happy :P
 752 2012-02-05 17:00:50 <TD> heh, yep, that's pretty cool
 753 2012-02-05 17:01:08 <TD> i see lots of clones of bitcoinj on google code, but not many people have contributed patches this cycle. oh well.
 754 2012-02-05 17:02:20 <TD> i guess there are a bunch of things i need to do as a higher priority before i get to play with multi-sigs
 755 2012-02-05 17:02:24 <TD> sigh
 756 2012-02-05 17:02:55 <TD> btw, description of how to do android+micropayments: https://groups.google.com/forum/#!topic/bitcoinj/ssTt-0FsPwQ
 757 2012-02-05 17:03:03 sausage2 has joined
 758 2012-02-05 17:03:13 <TD> funding apps via transparent bitcoin micropayments would be fun. and not that hard to implement, really
 759 2012-02-05 17:03:14 <sausage2> hi. Is that Satoshi guy here?
 760 2012-02-05 17:03:45 <TD> satoshi is gone for a long time now
 761 2012-02-05 17:03:47 <TD> you can't talk to him
 762 2012-02-05 17:03:58 <sausage2> why would he want to be anonymous?
 763 2012-02-05 17:04:07 <sausage2> unless YOU are him?
 764 2012-02-05 17:04:18 <TD> keeps things simple, i guess
 765 2012-02-05 17:04:18 <sausage2> do you think he comes here once in a while?
 766 2012-02-05 17:04:41 <TD> no. he left the project a year ago
 767 2012-02-05 17:05:16 <sausage2> maybe he died IRL by sickness or an accident?
 768 2012-02-05 17:05:30 <TD> no. he said specifically it was time for him to leave
 769 2012-02-05 17:05:47 <sausage2> I'm sure someone IRL knows who he is. He's closest friends or so.
 770 2012-02-05 17:06:11 <Moron__> my bet is he comes on here via tor all the time
 771 2012-02-05 17:06:26 <sausage2> did anyone of you chat with him when he was active?
 772 2012-02-05 17:06:54 <TD> i talked to him via email a few times
 773 2012-02-05 17:06:59 <TD> and he posted a lot on the forums, of course
 774 2012-02-05 17:07:16 <sausage2> he's a weird guy, I must say.
 775 2012-02-05 17:07:20 <sausage2> mysterious.
 776 2012-02-05 17:07:33 <TD> he never seemed that weird to me
 777 2012-02-05 17:09:28 <sausage2> if he gave an interview or anything.... it'd be a good thing
 778 2012-02-05 17:09:47 <sausage2> a lot of people are curious about him
 779 2012-02-05 17:10:10 <Moron__> i think thats part of his genius
 780 2012-02-05 17:10:17 <sausage2> I guess
 781 2012-02-05 17:10:24 <Moron__> if he started spouting reasons why he created bitcoin, people would use that against him, no matter what he said
 782 2012-02-05 17:10:36 <Moron__> in a way, to him, bitcoin was just an academic project
 783 2012-02-05 17:10:38 <sausage2> sure
 784 2012-02-05 17:11:05 <sausage2> Moron__: well, we can't tell that for sure. He might as well have made a fortune on this, without really caring about the project
 785 2012-02-05 17:11:24 <Moron__> i donno, he hasnt seemed to use any of his coins yet tho?
 786 2012-02-05 17:11:26 <sausage2> that's a possibility. But that doesnt mean it matters anymore.
 787 2012-02-05 17:11:43 <sausage2> Moron__: really?
 788 2012-02-05 17:11:54 <Moron__> doesnt the blockchain show that?
 789 2012-02-05 17:11:59 <sausage2> how much does he have?
 790 2012-02-05 17:12:02 <sausage2> Moron__: maybe
 791 2012-02-05 17:12:04 <Moron__> he mined loads
 792 2012-02-05 17:12:09 <sausage2> lol
 793 2012-02-05 17:12:11 <sausage2> surprising
 794 2012-02-05 17:12:30 <Moron__> that was back when the diifficulty was rediculously low
 795 2012-02-05 17:12:36 <Moron__> so all he needed was a cpu
 796 2012-02-05 17:12:57 <sausage2> but if all coins are logged in the blockchain... can't you see if he mined anything before he even booted up the project?
 797 2012-02-05 17:12:58 booo has joined
 798 2012-02-05 17:13:16 <marf_away> he didnt
 799 2012-02-05 17:13:19 <sausage2> or if he had "ready hashes" allready, just to "put that into the...  eh I don't know how all that works really
 800 2012-02-05 17:13:27 <Diablo-D3> he didnt mine much
 801 2012-02-05 17:13:38 <Diablo-D3> he mined the first few blocks to make the chain work
 802 2012-02-05 17:13:54 <Diablo-D3> you cant strictly tell what he mined because you dont know what addresses are his
 803 2012-02-05 17:14:19 <sausage2> couldnt he had mined coins offline, and then harvest them at some point?
 804 2012-02-05 17:14:19 <Diablo-D3> sausage2: and ready hashes makes no sense
 805 2012-02-05 17:14:23 <marf_away> i think, bitcoin was a prototype for satohi, not the real one...but it becamethe real one ;D
 806 2012-02-05 17:14:33 <Diablo-D3> sausage2: you cant mine offline
 807 2012-02-05 17:14:54 <Diablo-D3> sausage2: each new block has a field of the hash of the previous block.... and each block has a field for the time it was mined.
 808 2012-02-05 17:15:00 <Diablo-D3> you cant predict the future.
 809 2012-02-05 17:15:18 <marf_away> i think many of first million chips are gone for ever, lost in debugingprocess ;D
 810 2012-02-05 17:15:32 <lianj> until block 170, there were only coinbase blocks
 811 2012-02-05 17:15:57 <Diablo-D3> marf_away: in a sense.
 812 2012-02-05 17:16:12 <Diablo-D3> lianj: the first 170 were mined by satoshi or in coordination with him
 813 2012-02-05 17:16:19 <Diablo-D3> or something like that
 814 2012-02-05 17:16:24 <lianj> yea
 815 2012-02-05 17:16:54 <Diablo-D3> ;;bc,blocks
 816 2012-02-05 17:16:55 <gribble> 165498
 817 2012-02-05 17:16:58 <Diablo-D3> we're now at that many.
 818 2012-02-05 17:17:06 <sausage2> do you think his friends knows who he is? and arent there any IP-information left by him in e-mails headers, forum-http-logs or anything like that?
 819 2012-02-05 17:17:16 <Diablo-D3> sausage2: nope.
 820 2012-02-05 17:17:36 <sausage2> has any serious coordinated attempts to track him been made?
 821 2012-02-05 17:17:53 <lianj> some people have tried
 822 2012-02-05 17:18:05 <sausage2> no success I guess?
 823 2012-02-05 17:18:52 <lianj> are you researching for an article or what?
 824 2012-02-05 17:19:12 * pusle thinks Elvis came back to give us Bitcoin. Then he left again ^^
 825 2012-02-05 17:19:38 <TD> there haven't been any serious efforts to track him down
 826 2012-02-05 17:19:46 <TD> the man likes his privacy and why should we not respect that?
 827 2012-02-05 17:19:59 <riush> no public attempts at least...
 828 2012-02-05 17:20:00 <lianj> pusle: until otherwise confirmed thats how it took place
 829 2012-02-05 17:20:07 <riush> for all we know they might have found him :p
 830 2012-02-05 17:21:49 <Moron__> its thought if you rearrange the letters in the name satoshi nakamoto it gives you a clue as to who he is
 831 2012-02-05 17:21:59 <Moron__> but the community is still working on it
 832 2012-02-05 17:23:45 Clipse has quit (Read error: Connection reset by peer)
 833 2012-02-05 17:24:20 <sausage2> lianj: no. I'm just very curious who this man might be.
 834 2012-02-05 17:24:56 <sausage2> Satoshi Nakamoto .. yeah, I read somewhere about Motorola etc etc... but that seems outrageous. It's probably just one guy
 835 2012-02-05 17:25:03 <sausage2> or more. But not large organizations
 836 2012-02-05 17:25:11 <lianj> what TD said
 837 2012-02-05 17:26:25 <sausage2> =/
 838 2012-02-05 17:26:31 <sausage2> =(
 839 2012-02-05 17:27:16 BTC_Bear is now known as hbrntng!~BTC_Bear@unaffiliated/btc-bear/x-5233302|BTC_Bear
 840 2012-02-05 17:29:58 <sausage2> what do you think his regular job is?
 841 2012-02-05 17:30:00 booo has quit (Ping timeout: 272 seconds)
 842 2012-02-05 17:30:27 <sausage2> a host within NASA connects to the bitcoin network
 843 2012-02-05 17:30:51 <sausage2> but he wouldnt be that uncareful
 844 2012-02-05 17:31:52 <sausage2> so are you all developers?
 845 2012-02-05 17:34:34 <TD> satoshi is probably a 40+ year old british developer with an interest in macroeconomics, who taught himself cryptography in order to build bitcoin
 846 2012-02-05 17:34:46 <TD> and yes some of us here are indeed developers
 847 2012-02-05 17:35:40 <TD> at any rate, who he is does not matter
 848 2012-02-05 17:35:54 <TD> who remembers the inventor of packet switched networks? very few. the invention itself is what matters
 849 2012-02-05 17:37:09 <sausage2> or the C developer who recently died =(.......
 850 2012-02-05 17:37:16 <sausage2> that's sad
 851 2012-02-05 17:38:27 <sausage2> how's the developement going?
 852 2012-02-05 17:38:36 <sausage2> how enthusiastic are you guys?
 853 2012-02-05 17:38:41 <Moron__> meh
 854 2012-02-05 17:39:20 <TD> it's going ok
 855 2012-02-05 17:39:27 <TD> i wish i could spend more time on it
 856 2012-02-05 17:39:40 <TD> the possibilities and improvements that can be made outstrip the communities resources, at the moment
 857 2012-02-05 17:41:44 <BlueMatt> outstrip by far, sadly
 858 2012-02-05 17:43:12 <sausage2> how is it to be a developer in a project (like this)?
 859 2012-02-05 17:43:17 <sausage2> do you have regular jobs, or what?
 860 2012-02-05 17:43:35 <sausage2> of course you have. But, I mean.. how many hours/day do you spend with the project?
 861 2012-02-05 17:45:05 <Moron__> i dont really develop, but i am working on a nice new wallet-stealing trojan module for the client
 862 2012-02-05 17:45:40 * BlueMatt doesnt, Im just an undergrad
 863 2012-02-05 17:47:05 <TD> yeah, most of us have jobs i think
 864 2012-02-05 17:47:11 <TD> or are freelance/consulting
 865 2012-02-05 17:47:28 <justmoon> someone need consulting? I do consulting for food!!
 866 2012-02-05 17:47:34 <sausage2> but what are you guys doing here then?
 867 2012-02-05 17:47:44 <TD> justmoon: heh :)
 868 2012-02-05 17:47:44 <sausage2> well, what I am doing here?
 869 2012-02-05 17:47:49 <TD> very existential
 870 2012-02-05 17:47:50 <sausage2> but I thought you guys were the devs :)
 871 2012-02-05 17:47:55 <TD> yes
 872 2012-02-05 17:48:00 <sausage2> TD: you are?
 873 2012-02-05 17:48:00 <TD> most bitcoin development is volunteer driven
 874 2012-02-05 17:48:01 b4epoche has joined
 875 2012-02-05 17:48:08 <sausage2> ofc
 876 2012-02-05 17:48:12 <TD> i work on a reimplementation of the system in java
 877 2012-02-05 17:48:18 <TD> which is used by various apps and mobile wallets
 878 2012-02-05 17:48:26 <justmoon> sausage2: TD created bitcoinj, one of the most popular client libraries
 879 2012-02-05 17:48:26 <sausage2> TD: all by yourself?
 880 2012-02-05 17:48:38 <sausage2> gonna check it out :)
 881 2012-02-05 17:48:40 <TD> devrandom helps me a lot
 882 2012-02-05 17:48:41 <TD> and a few others
 883 2012-02-05 17:48:52 <TD> but it's mostly me yes
 884 2012-02-05 17:48:57 <TD> http://code.google.com/p/bitcoinj/
 885 2012-02-05 17:48:59 <sausage2> devrandom?
 886 2012-02-05 17:49:12 <TD> aka miron cuperman
 887 2012-02-05 17:49:16 <justmoon> a colleague of TD's at google
 888 2012-02-05 17:49:59 <sausage2> oh, so devrandom is a person
 889 2012-02-05 17:50:09 <justmoon> :D
 890 2012-02-05 17:50:16 <sausage2> bitcoin client should be java based instead of c++ based
 891 2012-02-05 17:50:16 <BlueMatt> cat /dev/random>bitcoinj source
 892 2012-02-05 17:50:19 <justmoon> he's in this channel right now actually
 893 2012-02-05 17:50:20 <sausage2> for various reasons
 894 2012-02-05 17:50:27 <BlueMatt> ;)
 895 2012-02-05 17:50:31 <sausage2> lol
 896 2012-02-05 17:50:40 <TD> well, feel free to send me patches for the java version :)
 897 2012-02-05 17:50:54 <justmoon> <TD> devrandom helps me a lot
 898 2012-02-05 17:51:04 <justmoon> that statement is indeed really funny out of context
 899 2012-02-05 17:51:23 <TD> it is occasionally helpful when implementing crypto :)
 900 2012-02-05 17:51:29 <justmoon> ^^
 901 2012-02-05 17:51:38 <TD> but mostly i just throw a few dice and then put the results i got into an array
 902 2012-02-05 17:51:45 <TD> with a counter to choose the next random number
 903 2012-02-05 17:51:57 <TD> i don't trust RNGs i can't understand :)
 904 2012-02-05 17:52:10 <BlueMatt> heh, whats the xkcd?
 905 2012-02-05 17:52:21 <BlueMatt> http://xkcd.com/221/
 906 2012-02-05 17:52:28 <TD> yeah :)
 907 2012-02-05 17:52:47 <josephcp> justmoon: btw: a while back i was going to implement bitcoinjs-gui on iphone but was too unfamiliar with the code so I gave up because it look like it'd take too long (working on other stuff right now). i think it's possible to dump the entire bitcoinjs-gui into a "Data: URI" with a pre-created hardcoded wallet
 908 2012-02-05 17:53:06 <josephcp> and have that as an icon on the home screen
 909 2012-02-05 17:53:38 <justmoon> josephcp: hmm interesting idea, would love to see a proof of concept
 910 2012-02-05 17:53:51 <justmoon> josephcp: with a simple javascript hello world program
 911 2012-02-05 17:53:54 <josephcp> it got too difficult because the current revision uses define[]
 912 2012-02-05 17:54:02 <josephcp> http://blog.clawpaws.net/post/2007/07/16/Storing-iPhone-apps-locally-with-data-URLs
 913 2012-02-05 17:54:15 <TD> wow
 914 2012-02-05 17:54:17 <TD> what a glorious hack
 915 2012-02-05 17:54:26 <justmoon> TD: careful what you say next
 916 2012-02-05 17:54:30 <TD> though now apple let in a wallet app, maybe it'd be easier to just have a webkit wrapper app
 917 2012-02-05 17:54:34 <josephcp> so yeah it's technically possible
 918 2012-02-05 17:54:35 Joric has quit ()
 919 2012-02-05 17:54:58 <justmoon> TD: yeah, but even if you don't stay in the 'hood, the 'hood stays in you
 920 2012-02-05 17:55:06 <justmoon> so we gotta keep it underground!
 921 2012-02-05 17:55:08 <josephcp> you just need a generator that can generate a custom data: uri that creates its own wallet
 922 2012-02-05 17:55:13 <TD> haha
 923 2012-02-05 17:55:47 <justmoon> josephcp: that's brilliant dude - I agree with TD that we can probably get a phonegap or similar app in the appstore now, but this is still really cool
 924 2012-02-05 17:56:12 <justmoon> and you never know, when the shit hits the fan we might get booted out of the appstore and market pretty quickly
 925 2012-02-05 17:56:39 <josephcp> yeah it's a good "just in case", and you dont need to rely on Apple (however the downside is you can't use the camera)
 926 2012-02-05 17:57:08 <TD> i think there is some effort make an html5 camera api
 927 2012-02-05 17:57:24 <justmoon> TD: indeed
 928 2012-02-05 17:57:27 <justmoon> <3 google
 929 2012-02-05 17:57:47 <TD> i'd like to see us move away from qrcodes though
 930 2012-02-05 17:57:53 <TD> convenient though they are at the moment
 931 2012-02-05 17:58:02 <josephcp> i was screwing around with it using just a pure bookmarklet with using HTML5 dbs, but it didn't work because Safari is very strict and denies access to dbs in blank namespaces
 932 2012-02-05 17:58:03 <BlueMatt> what is with your hate of qr codes? what do you want to move to?
 933 2012-02-05 17:58:11 <justmoon> TD: so get working on a NFC HTML5 api lol
 934 2012-02-05 17:58:15 <TD> heh
 935 2012-02-05 17:58:15 <TD> well
 936 2012-02-05 17:58:22 <TD> NFC just opens an arbitrary URL
 937 2012-02-05 17:58:23 <TD> at least on android
 938 2012-02-05 17:58:30 <TD> and there's already an html5 api to handle URL schemes
 939 2012-02-05 17:58:38 <BlueMatt> mmm, thats cool, but doesnt replace all the use-cases of a qr
 940 2012-02-05 17:58:54 <TD> sure
 941 2012-02-05 17:59:07 <TD> i'd like there to be an easy way to click a link on your laptop and have it open up the wallet on your phone
 942 2012-02-05 17:59:13 <TD> and i'd like to see social network integration
 943 2012-02-05 17:59:13 <josephcp> it worked in firefox though, it also works in safari if you use the bookmarklet while browsing ON a website (isntead of a blank page), but then that page could hypothetically have access to your wallet, which led me to the conclusion that a hardcoded custom generated wallet in data: is probably the most practical
 944 2012-02-05 17:59:27 <TD> so i can just pick a friend from a list pulled down via facebook or whatever, and then have a new address created in the background
 945 2012-02-05 17:59:28 <TD> all seamless
 946 2012-02-05 17:59:48 <BlueMatt> easier link -> android would be cooler, chrome send to phone is nice, but has a lot of cool potential that isnt being used
 947 2012-02-05 18:00:06 <justmoon> josephcp: you can still call the internet, so you could have the deterministic wallet seed stored and all other less critical user data in the cloud somewhere
 948 2012-02-05 18:00:13 <BlueMatt> yea, social network integration is good for just about any project ;)
 949 2012-02-05 18:00:21 <justmoon> seed stored in the data url*
 950 2012-02-05 18:00:31 <josephcp> BTW: using data: URIs makes bitcoinjs-gui viable even on the desktop because you don't need to trust the server to not screw with the HTML and compromise your wallet key
 951 2012-02-05 18:00:32 <BlueMatt> (after all when people spend hours a day on one site, having something available there is always nice
 952 2012-02-05 18:00:59 <josephcp> yeah, having a deterministic wallet (preferable) or just use one private key would work i think
 953 2012-02-05 18:01:03 <justmoon> josephcp: again, the better way is probably a qtwebkit wrapper or a chrome app, no?
 954 2012-02-05 18:01:30 <josephcp> yeah, for sure
 955 2012-02-05 18:02:23 <justmoon> I still marvel at the brilliance of the idea and that this actually works :)
 956 2012-02-05 18:02:33 <justmoon> do you know what kind of size limits are in effect there?
 957 2012-02-05 18:02:40 <josephcp> none as far as i know
 958 2012-02-05 18:02:51 <josephcp> i think i tested a 1MB junk file and put it on the home screen
 959 2012-02-05 18:02:57 <justmoon> O_o
 960 2012-02-05 18:03:10 <justmoon> that's kind of funny
 961 2012-02-05 18:04:32 <TD> wow
 962 2012-02-05 18:04:37 <josephcp> yeah haha, i gave up on it because I'm not a super expert on javascript :-(
 963 2012-02-05 18:04:42 <TD> yeah i guess every image also has to be encoded as a data uri
 964 2012-02-05 18:04:50 <TD> resulting in double base64 encoding
 965 2012-02-05 18:05:08 <justmoon> TD: you'd think there would be a limit on bookmark urls though (or home screen urls)
 966 2012-02-05 18:05:10 <josephcp> yeah you can include inline images in HTML though
 967 2012-02-05 18:05:21 <josephcp> there is in IE i think
 968 2012-02-05 18:05:36 <TD> well there's going to be a practical limit
 969 2012-02-05 18:05:41 <josephcp> none in firefox/safari that matter at a reasonable level (no idea about chrome)
 970 2012-02-05 18:05:44 <TD> yeah, IE limited URL length. very annoying.
 971 2012-02-05 18:05:53 <justmoon> TD: total persistent memory :P
 972 2012-02-05 18:06:38 <TD> divided by three or however many times it gets copied :)
 973 2012-02-05 18:06:41 booo has joined
 974 2012-02-05 18:06:50 <justmoon> hehe true
 975 2012-02-05 18:07:12 <TD> how is bitcoinjs coming along?
 976 2012-02-05 18:07:14 <TD> had much time to work on it lately?
 977 2012-02-05 18:07:25 <splatster> etotheipi_: 10.7 doesn't like Armory
 978 2012-02-05 18:07:42 <justmoon> josephcp: iirc there is a compiler for require.js that compiles out the defines and you end up with the plain compressed js - that should then be includable
 979 2012-02-05 18:07:57 <josephcp> i think require.js broke everything
 980 2012-02-05 18:08:02 <etotheipi_> splatster: I'm out of town so there's not much I can do at the moment... (no dev environment here), but even if I was home I'm not sure what I could do
 981 2012-02-05 18:08:07 gronager has joined
 982 2012-02-05 18:08:18 <etotheipi_> I was thinking maybe I should buy a crappy, used Mac, just for this stuff
 983 2012-02-05 18:08:34 <josephcp> but this was a couple months ago that i played around with it
 984 2012-02-05 18:08:44 <justmoon> TD: not too much time unfortunately, I'm in the process of moving and I had some other projects to take care of
 985 2012-02-05 18:08:56 Clipse has joined
 986 2012-02-05 18:08:57 <TD> :(
 987 2012-02-05 18:08:59 <TD> know the problem
 988 2012-02-05 18:09:13 <justmoon> TD: it's always cyclical for me, I think I'm at the beginning of a more active phase again now
 989 2012-02-05 18:09:21 <TD> excellent
 990 2012-02-05 18:09:29 <splatster> etotheipi_: Get a nice Mac Mini if you already have keyboards/monitor/etc
 991 2012-02-05 18:09:35 <TD> i feel like 90% of my time on bitcoinj gets sucked up with yak shaving exercises
 992 2012-02-05 18:09:41 <justmoon> josephcp: broke everything?
 993 2012-02-05 18:10:19 * justmoon had to look up yak shaving
 994 2012-02-05 18:10:23 <TD> oh, sorry
 995 2012-02-05 18:10:25 <etotheipi_> splatster: that's a good idea... I was trying to decide what would be the best, low-cost solution for doing it (and I do have extra keybaord&mouse)... is mac-mini the smallest-yet-usable-option?
 996 2012-02-05 18:10:41 <justmoon> don't be sorry, seems like a good expression to know
 997 2012-02-05 18:10:47 <etotheipi_> s/smallest/cheapest/g
 998 2012-02-05 18:10:50 <josephcp> i don't know, debugging javascript is not my forte :-P i just remember trying it with a really really old git revision before require.js was used and it was workable
 999 2012-02-05 18:10:51 <TD> it's some dumb silicon valley expression that means "performing an apparently pointless task that is actually required to complete yet another pointless task that is required in order to reach your actual goal"
1000 2012-02-05 18:11:08 <TD> so i spend a lot of time on resolving minor API issues, bug fixing edge cases etc.
1001 2012-02-05 18:11:15 <TD> not much time left to experiment with the core technology
1002 2012-02-05 18:11:21 <splatster> etotheipi_: smallest cheapest all of the above
1003 2012-02-05 18:11:34 <josephcp> but it could've also been a modificiation somewhere else, this was a while back, sorry
1004 2012-02-05 18:12:01 <splatster> The newer Mac minis come with some nice specs
1005 2012-02-05 18:12:05 <justmoon> TD: speaking of yak shaving, I recently implemented a blog for bitcoinjs: http://bitcoinjs.org/blog/2012/02/04/simplifying-our-architecture.html
1006 2012-02-05 18:12:34 <justmoon> first time I used jekyll's blog feature (the static website generator powering github pages)
1007 2012-02-05 18:12:44 <splatster> etotheipi_: the problem is that it ships with lion, which I haven't been able to get armory to work with
1008 2012-02-05 18:12:52 <TD> justmoon: great. i wondered why there were so many subprojects before
1009 2012-02-05 18:12:56 <josephcp> one thing i did learn then as a result is that mongodb is crazy fast for bitcoin, so im using mongodb to store bitcoin transactions for a blockexplorer (but an incompatible db format :-P)
1010 2012-02-05 18:13:07 <etotheipi_> splatster: that sounds like exactly the one I want, then
1011 2012-02-05 18:13:09 <splatster> I am going to get a mac mini with server specs and use it as my desktop comp
1012 2012-02-05 18:13:18 <josephcp> indexed mysql is so slow
1013 2012-02-05 18:13:24 <etotheipi_> I need to figure out how to get it to work on the systems that it doesn't work on yet :)
1014 2012-02-05 18:13:27 <splatster> Have a dual boot partition
1015 2012-02-05 18:13:30 <splatster> get ubuntu
1016 2012-02-05 18:13:31 <TD> why is mongodb so much faster?
1017 2012-02-05 18:13:45 <splatster> make an offline comp out of the ubuntu partition
1018 2012-02-05 18:13:57 <splatster> encrypt with trucrypt
1019 2012-02-05 18:14:00 <etotheipi_> splatster: wait, isn't there a way to install OSX on intel-CPU systems?
1020 2012-02-05 18:14:10 <justmoon> TD: mongodb is a document database, mysql is relational
1021 2012-02-05 18:14:31 <etotheipi_> splatster: what's the encryption for?  if it's offline and the wallet already has encryption...?
1022 2012-02-05 18:14:35 <splatster> etotheipi_: It's called Hackintosh, and it will only set you back farther as almost nothing works on it
1023 2012-02-05 18:14:38 <justmoon> people experience mongodb as being faster because it forces them to design their applications properly
1024 2012-02-05 18:14:47 <etotheipi_> (though I wouldn't be offended if folks didn't want to trust the encryption, yet)
1025 2012-02-05 18:14:59 <splatster> etotheipi_: It's just for my paranoia's sake
1026 2012-02-05 18:15:15 <justmoon> it also has lower overhead (not a scalability issue, but a performance issue)
1027 2012-02-05 18:15:36 <etotheipi_> splatster: well I just want to remind you :)  I found the 0.4.0 wallet-not-really-encrypted bug, and that was my inspiration for my wallet file format... to avoid exactly that
1028 2012-02-05 18:16:28 <splatster> etotheipi_: Wow
1029 2012-02-05 18:16:32 <splatster> didn't know that
1030 2012-02-05 18:16:48 <etotheipi_> it was because I was investigating the satoshi wallet format (BDB)
1031 2012-02-05 18:17:05 <etotheipi_> but I didn't feel like figuring out BSDDB, so I wrote a binary searcher
1032 2012-02-05 18:17:09 <splatster> Imma donate some of what you gave me back to Armory, just for that
1033 2012-02-05 18:17:27 <etotheipi_> to my surprise, I found 36 of 108 private keys when I ran it on my Satoshi wallet
1034 2012-02-05 18:17:59 <splatster> that's scary
1035 2012-02-05 18:18:16 <justmoon> etotheipi_: is your wallet file format documented anywhere? is it suitable as a standard to be adopted by other clients?
1036 2012-02-05 18:18:16 <etotheipi_> it turns out that BSDDB engine manages the memory itself... when you tell it to delete/overwrite, it doesnt' necessarily do it
1037 2012-02-05 18:18:34 <TD> talking of wallet formats, bitcoinj is moving to protocol buffer serialization soon
1038 2012-02-05 18:18:42 <etotheipi_> http://bitcoinarmory.com/index.php/armory-wallet-files
1039 2012-02-05 18:18:44 <TD> so those wallets will be readable by any language that has protobuf suppotr
1040 2012-02-05 18:19:23 <etotheipi_> justmoon: I have to update the format to enable forward-compatibility, but otherwise it should be easy to implement in any language
1041 2012-02-05 18:19:48 <justmoon> etotheipi_: is it binary, json, or something else?
1042 2012-02-05 18:19:56 <etotheipi_> plain binary
1043 2012-02-05 18:20:40 <etotheipi_> but everything is straightforward, has consistency checks, and the keys are easy to get to and don't require any special libraries... and they are guaranteed to be overwritten in place if you encrypt it
1044 2012-02-05 18:21:23 <justmoon> the main problem I'm worried about is extensibility - how does it handle deterministic wallets, multisig wallet or whatever the next thing is we come up with
1045 2012-02-05 18:21:45 <etotheipi_> well I already use it for deterministic wallets with the ability to import addresses
1046 2012-02-05 18:22:08 <etotheipi_> and it will support any potential protocol extensions just by adding new codes
1047 2012-02-05 18:22:44 <justmoon> hmhmhm
1048 2012-02-05 18:22:46 <etotheipi_> the bottom row is the existing codes... you'll notice I even put in an entry-type for OP_EVAL... before it was sacked
1049 2012-02-05 18:23:10 <justmoon> the bottom row? did you post a link I missed?
1050 2012-02-05 18:23:13 <TD> http://code.google.com/p/bitcoinj/source/browse/src/bitcoin.proto
1051 2012-02-05 18:23:16 <etotheipi_> however, the format probably needs to be substantiated with some kind of yes
1052 2012-02-05 18:23:16 <TD> here is the format we're using
1053 2012-02-05 18:23:20 <etotheipi_> er..
1054 2012-02-05 18:23:22 <TD> it is designed to be forwards/backwards compatible
1055 2012-02-05 18:23:26 <TD> so you might get some inspiration there
1056 2012-02-05 18:23:28 <etotheipi_> http://bitcoinarmory.com/index.php/armory-wallet-files
1057 2012-02-05 18:24:35 <TD> it is still kinda new though so there is room for improvement
1058 2012-02-05 18:24:41 <etotheipi_> justmoon: the root key and chaincode exists in the header to accommodate determistic wallets, then after the header is a sequence of appended address information, comments, and OP_EVAL scripts
1059 2012-02-05 18:24:48 <TD> it doesn't have any way to flag encrypted keys, for example
1060 2012-02-05 18:25:03 <etotheipi_> although I guess I'll be changing that to BIP16-scripts
1061 2012-02-05 18:25:48 PK has joined
1062 2012-02-05 18:26:33 BLZNGPNGN has quit (2!~kvirc@S0106602ad0726c1f.vf.shawcable.net|Read error: Connection reset by peer)
1063 2012-02-05 18:26:35 <etotheipi_> my plan is that all future entries will use a standard format:   [code | size | data] which should allow me to add something like BIP16-scripts, and older clients can just skip over entries that have codes they don't recognize
1064 2012-02-05 18:27:06 <justmoon> interesting: "Google makes heavy use of Protocol Buffers in JS (GMail, etc.) through their Closure Library, generating JS code with a (unfortunately non-open-sourced) modified protoc (it would probably have to be ported to a protoc extension before being open-sourced)."
1065 2012-02-05 18:27:41 <TD> etotheipi_: handing of new fields such that old clients ignore them is the sort of thing protobufs is designed to handle. serialization with (tag, length, data) tuples is really a solved problem.
1066 2012-02-05 18:27:43 <justmoon> so yeah, if protocol buffers in javascript ever happens, your format would look really good TD
1067 2012-02-05 18:27:54 <TD> justmoon: i didn't know the protobuf-to-js compiler wasn't open sourced
1068 2012-02-05 18:28:08 <TD> justmoon: i think mostly we convert them to json for interop with the js world
1069 2012-02-05 18:28:09 <justmoon> otherwise I'm gonna have to realistically use a json-based format just based on the environment I'm in most of the time
1070 2012-02-05 18:28:21 <justmoon> convert them to json?
1071 2012-02-05 18:28:49 <TD> there's some library internally that takes a protocol buffer and the C++/Java compiled code, then serializes the data to json automatically
1072 2012-02-05 18:29:05 <TD> i don't know if there's code that turns json back into type safe objects on the client side or anything. i never tried it.
1073 2012-02-05 18:29:16 <TD> i vaguely recall serializing protobufs to json a few years ago as part of some toy project i had
1074 2012-02-05 18:29:25 <TD> but i guess as your server is JS too, that wouldn't help you
1075 2012-02-05 18:29:28 <etotheipi_> having human-readable wallet files scares me a tiny bit... I want the user to at least have to know how to open a file in binary in order to mess with it
1076 2012-02-05 18:29:38 <TD> protobuf serializes to binary or text
1077 2012-02-05 18:29:45 <justmoon> if I google "protocol buffers json" all I find is a blog article that says json serialization/deserialization is faster than protocol buffers lol
1078 2012-02-05 18:29:48 <etotheipi_> ahh
1079 2012-02-05 18:29:50 <TD> binary is the "real" format. the text format is a bit bizarre. it was mostly intended for debugging.
1080 2012-02-05 18:30:03 <etotheipi_> maybe I should have both...
1081 2012-02-05 18:30:05 <TD> justmoon: well i'd hope so :) modern JS engines probably eliminate much perf difference though
1082 2012-02-05 18:30:29 <TD> justmoon: https://github.com/sirikata/protojs ?
1083 2012-02-05 18:30:38 <TD> seems like that guy went pretty deep
1084 2012-02-05 18:31:18 <TD> though i'm not sure it'd be a good idea in this case, i guess
1085 2012-02-05 18:31:21 <justmoon> TD: you'd hope so? I don't follow
1086 2012-02-05 18:31:27 josephcp has quit (Ping timeout: 252 seconds)
1087 2012-02-05 18:31:30 <justmoon> I would have expected protocol buffers to be faster
1088 2012-02-05 18:31:33 <TD> justmoon: json parsing is done natively in C++, right?
1089 2012-02-05 18:31:42 <TD> whereas protobuf parsing would be done purely in javascript, i guess
1090 2012-02-05 18:31:43 <justmoon> TD: no this was a java benchmark, not javascript
1091 2012-02-05 18:31:46 <TD> oh
1092 2012-02-05 18:31:55 <justmoon> i.e. a java json implementation versus java protobuf
1093 2012-02-05 18:31:59 <justmoon> http://blog.juma.me.uk/2009/03/25/json-serializationdeserialization-faster-than-protocol-buffers/
1094 2012-02-05 18:32:08 <TD> i don't know. it could be correct. protobufs has two versions
1095 2012-02-05 18:32:24 <TD> the first one is internal only. it was written by speed freaks and highly optimized, it was also deemed "too messy" to open source
1096 2012-02-05 18:32:26 eoss has quit (Remote host closed the connection)
1097 2012-02-05 18:32:37 <justmoon> ...
1098 2012-02-05 18:32:42 <TD> some guy rewrote it into "protobufs 2" which has totally new java APIs and implementations, which and he did it in the "modern java style"
1099 2012-02-05 18:32:50 <TD> which is a euphamism for academic and without any attention paid to performance
1100 2012-02-05 18:33:03 <TD> the C++ implementation was largely left alone, but even then, it got a bit slower and had to be re-optimized
1101 2012-02-05 18:33:12 josephcp has joined
1102 2012-02-05 18:33:30 <TD> more problematically the new Java APIs were totally different, resulting in absurd amounts of interop pain. the whole initiative was very poorly done and made the poor guy quite notorious inside the firm
1103 2012-02-05 18:33:39 <TD> proto2 has wasted more time than i thought possible for such a trivial thing
1104 2012-02-05 18:33:57 <justmoon> did it prevail in the end or did you go back?
1105 2012-02-05 18:34:04 <TD> so i wouldn't want to make any claims about the performance of protobufs these days, especially not the java version. the old c++ version that is most widely used was very tight
1106 2012-02-05 18:34:19 <TD> bit of both. i think most codebases still have  a mix of proto1 and proto2 more or less at random.
1107 2012-02-05 18:34:35 <TD> i've had to port code back to proto1 before, typically due to bizarre API interop issues rather than performance
1108 2012-02-05 18:35:00 <cjd> lol
1109 2012-02-05 18:35:34 <TD> i work mostly in the c++ codebase where the kind of crap you see in the java world is much less prevalent. proto2 for us was mostly a non issue. you ran into issues with things like SWIG where protobufs were being passed to/from interpreted environments
1110 2012-02-05 18:35:40 <TD> the java world suffered a lot harder
1111 2012-02-05 18:36:00 <cjd> looking at protobuf, I think it can be very fast, with java though, optimization is full of magic
1112 2012-02-05 18:36:22 <TD> the format was designed for the core search engine. so yeah, it can indeed be very performant. and i mean it's much better than xml or whatever.
1113 2012-02-05 18:36:26 <TD> if you have a good implementation
1114 2012-02-05 18:36:51 <TD> the java implementation of proto2 would make me want to kill things though, api wise. fortunately miron wrote all that code.
1115 2012-02-05 18:36:58 * cjd has concluded that trying to optimize java is never a good plan because there are too many unknowns about what the vm decides to do
1116 2012-02-05 18:37:03 <TD> yeah
1117 2012-02-05 18:37:10 <TD> we have a lot of internal folklore about optimizing java
1118 2012-02-05 18:37:58 <TD> even so it's probably more like 2x as expensive as the c++ equivalent, when everything is taken into account. java vs c++ flamewars and interop stupidity are a huge drag on the company. almost makes me wish we'd do a zuckerberg and standardize the company on PHP
1119 2012-02-05 18:38:20 <cjd> heh
1120 2012-02-05 18:38:44 * cjd likes C, it takes forever to write something but when you're done you have a real work of art
1121 2012-02-05 18:38:46 <sausage2> is facebook runned on PHP?
1122 2012-02-05 18:38:50 <sausage2> thought it used ajax as well?
1123 2012-02-05 18:39:24 <josephcp> yeah facebook uses PHP, but they use their own compiler
1124 2012-02-05 18:39:27 <TD> it is implemented mostly in PHP which they then compile to C++, then onto machine code
1125 2012-02-05 18:39:35 <TD> ajax isn't a language
1126 2012-02-05 18:39:38 <TD> so you can't compare them
1127 2012-02-05 18:39:42 <sausage2> oh
1128 2012-02-05 18:39:49 <justmoon> TD: everyone knows ajax is a cleaning product
1129 2012-02-05 18:39:55 <sausage2> why would facebook do that?
1130 2012-02-05 18:39:57 <TD> cjd: i think c++ is a good middle ground. you have the abstractions but you can also do things as precisely as you need
1131 2012-02-05 18:39:59 <Zarutian> bit like the sues crises nipping out for bun kind of thing?
1132 2012-02-05 18:40:15 <sausage2> how can they run a webserver on C++? llol?
1133 2012-02-05 18:40:24 <TD> sausage2: cost. interpreted php is, like any interpreted language, very slow and that means expensive.
1134 2012-02-05 18:40:34 <TD> google web search servers are all written in c++
1135 2012-02-05 18:40:37 <TD> it's easy
1136 2012-02-05 18:40:48 <TD> (same for quite a lot of other products web servers)
1137 2012-02-05 18:41:27 <sausage2> how does that work? Are the http-requests handeled directly by machine code?
1138 2012-02-05 18:41:32 <sausage2> but google uses ngix
1139 2012-02-05 18:41:34 <sausage2> so do youtube
1140 2012-02-05 18:41:41 <sausage2> nginx*
1141 2012-02-05 18:41:48 <Zarutian> TD: well PHP is actually slower than many other interpreted languages due to its semantics, afaik
1142 2012-02-05 18:41:51 <TD> google does not use nginx, i assure you. youtube might have done at some point but not any more.
1143 2012-02-05 18:42:05 <TD> Zarutian: yes. it pays the penalty of being both interpreted and loosely/dynamically typed
1144 2012-02-05 18:42:13 <justmoon> the php facebook compiles is a reduced syntax
1145 2012-02-05 18:42:14 <sausage2>  TD: I've got nginx error messages while surfing youtube
1146 2012-02-05 18:42:16 <TD> yeah
1147 2012-02-05 18:42:20 <TD> sausage2: how recently?
1148 2012-02-05 18:42:30 <sausage2> TD: months
1149 2012-02-05 18:42:39 <sausage2> I have such a bad memory, not sure when
1150 2012-02-05 18:42:42 <TD> perhaps it still lurks in some corners. the youtube guys do their own thing, to some extent
1151 2012-02-05 18:42:45 <justmoon> I mostly get that custom google "Oops" page
1152 2012-02-05 18:42:59 <sausage2> yes, that's now
1153 2012-02-05 18:43:00 <TD> "we screwed up. that's all we know"
1154 2012-02-05 18:43:05 <justmoon> yeah
1155 2012-02-05 18:43:14 <sausage2> but... how's that C++-webserverthing working?
1156 2012-02-05 18:43:23 <TD> c++ web servers "work" in the same way any web server works. why would language make any difference?
1157 2012-02-05 18:43:24 <Zarutian> TD: no, it is slow because the semantics is so ambigiously parsed from the syntax. I have seen dynamically typed languages that run faster than PHP.
1158 2012-02-05 18:43:33 <josephcp> sausage2: well if you wanted to do it yourself, just implement FastCGI
1159 2012-02-05 18:43:37 <TD> Zarutian: i'll take your word for it. i don't know much about PHP
1160 2012-02-05 18:43:42 <josephcp> you can probably get good results if you google C++ FastCGI
1161 2012-02-05 18:43:54 <TD> sausage2: same as any web server works. it receives a network connection, parses the headers, etc, calculates the result, writes it back
1162 2012-02-05 18:43:56 <sausage2> just curious how google does it
1163 2012-02-05 18:44:06 <TD> there's a C++ library that implements HTTP serving
1164 2012-02-05 18:44:12 <sausage2> TD and that is done by a binary program?
1165 2012-02-05 18:44:15 <TD> yes
1166 2012-02-05 18:44:20 <sausage2> neat
1167 2012-02-05 18:44:22 <TD> you compile it, run it. now you have a program that's listening on a port.
1168 2012-02-05 18:44:41 <sausage2> I've always wondered why you'd waste so much CPU on the same interpretation, over and over again
1169 2012-02-05 18:45:07 <sausage2> TD: so facebook recompile everytime them make a change on their site?
1170 2012-02-05 18:45:19 <josephcp> because human time/effort is a cost as well, often more expensive than machine cycles
1171 2012-02-05 18:45:30 <Zarutian> TD: PHP was first made as an specification language implemented in unoptmisable perl (that is, non strict) and it hasnt got any better. I recommend that people use Ruby, Python, Lua with decent HTTP frameworks.
1172 2012-02-05 18:45:33 <TD> i believe for dev/testing they use the interpreter, and every few days/once a week they compile it to C++ and push it out to the production site
1173 2012-02-05 18:45:35 <bitlad> http://en.wikipedia.org/wiki/Google_platform#Software <-- google servers
1174 2012-02-05 18:46:03 <TD> sausage2: but it is expensive. that's one reason java costs more to run than a c++ server doing the equivalent would
1175 2012-02-05 18:46:05 <justmoon> sausage2: also, modern frameworks don't recompile on every request, like node.js or mongrel2
1176 2012-02-05 18:46:43 <makomk> luke-jr: sorry, didn't notice your message. "Notes for Gentoo users: emerge qemu-kvm ruby debootstrap multipath-tools apt-cacher-ng dev-python/cheetah, manually download and install vm-builder-0.12.4+bzr469" according to my note to myself.
1177 2012-02-05 18:46:57 <TD> bitlad: i guess that's a reasonable overview
1178 2012-02-05 18:47:01 <TD> the backend of gmail is C++. the frontend is java
1179 2012-02-05 18:47:12 <makomk> It helps that Gentoo has ebuilds for apt, debootstrap and apt-cacher-ng
1180 2012-02-05 18:47:17 <justmoon> TD: java? o_O
1181 2012-02-05 18:47:20 <bitlad> no, backend is java for gmail
1182 2012-02-05 18:47:26 * justmoon tries to imagine the gmail java applet
1183 2012-02-05 18:47:29 <TD> it depends how you define "backend" i guess
1184 2012-02-05 18:47:31 <sausage2> TD: are there any standard solutions for this, or are they using their own stuff?
1185 2012-02-05 18:47:34 <Zarutian> TD: java? dont you mean emcascript which is better known as js or javascript?
1186 2012-02-05 18:47:35 <TD> but trust me
1187 2012-02-05 18:47:39 <josephcp> my understanding is google uses java a lot
1188 2012-02-05 18:47:42 <TD> i work on gmail.
1189 2012-02-05 18:48:05 <bitlad> google uses c/c++, java and python.
1190 2012-02-05 18:48:05 <TD> the "frontend server" which handles peoples HTTP requests and manages the UI is java. storage, spam, delivery, message routing, smtp etc is all c++
1191 2012-02-05 18:48:27 <TD> if you define "frontend" as "stuff that runs in the browser" then obviously that is javascript
1192 2012-02-05 18:48:48 <Zarutian> x out my last message as I thought this was the common misconception of javascript == java.
1193 2012-02-05 18:48:51 m00p has joined
1194 2012-02-05 18:48:53 <TD> bitlad: and javascript and sawzall and shell and a bunch of other random in house languages we never released
1195 2012-02-05 18:49:06 <TD> lots of languages
1196 2012-02-05 18:49:15 <makomk> luke-jr: also, don't forget to actually start apt-cacher-ng, I always do for some reason.
1197 2012-02-05 18:49:49 <bitlad> i need to play around with python, not a fan of php. i use mostly java for everythinbg
1198 2012-02-05 18:50:21 <Moron__> not a fan of java
1199 2012-02-05 18:50:24 <Moron__> i use basic for most things
1200 2012-02-05 18:50:24 <Moron__> :P
1201 2012-02-05 18:50:27 <Diablo-D3> td: now only if google would make a language I like
1202 2012-02-05 18:50:36 <bitlad> GO ?
1203 2012-02-05 18:50:46 <Diablo-D3> didnt care for go or dart
1204 2012-02-05 18:51:09 <Diablo-D3> basically, I want C as a scripting language with ONE feature from js: the obsession with callbacks and being able to inline define them
1205 2012-02-05 18:51:27 <Diablo-D3> and have it be able to also be compiled as native C, with a very minimal runtime lib
1206 2012-02-05 18:52:05 <TD> i think there is some kind of c-as-an-interpreter scripting engine out there somewhere
1207 2012-02-05 18:52:13 <TD> though why you'd want one is beyond me
1208 2012-02-05 18:52:21 <Diablo-D3> because pretty much I hate all languages.
1209 2012-02-05 18:52:30 <Diablo-D3> C is the only language Ive never really come to hate
1210 2012-02-05 18:52:53 <bitlad> I prefer c over c++ :-)
1211 2012-02-05 18:53:06 <josephcp> yeah C feels more solid/predictable than C++ :-P
1212 2012-02-05 18:54:04 <Diablo-D3> C is a very minimal language, which is why its so damned useful
1213 2012-02-05 18:54:26 <Zarutian> Diablo-D3: what is Forth then? super minimal?
1214 2012-02-05 18:54:49 <Diablo-D3> forth is a different way of doing coding
1215 2012-02-05 18:55:19 <Diablo-D3> depending on which forth variant you use, its usually less complex than C
1216 2012-02-05 18:55:27 <Diablo-D3> problem is that leads to it being TOO simple
1217 2012-02-05 18:55:55 sausage2 has quit (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111220165912])
1218 2012-02-05 18:56:09 <Zarutian> it is and I recommend it like Scheme and E. If you program in those for a while and then perhaps never again then your programming in other languages improves.
1219 2012-02-05 18:56:38 <Diablo-D3> Zarutian: eh
1220 2012-02-05 18:56:38 <Zarutian> (same for Prolog and Erlang)
1221 2012-02-05 18:56:42 <Diablo-D3> you describe the lisp effect.
1222 2012-02-05 18:56:52 <Diablo-D3> lisp profoundly changes how you code, more than any other language
1223 2012-02-05 18:58:56 <Zarutian> beleave me, writing forth words also changes how you code. (You have to keep an eye on the stacks' effects and side effects or you are just blown out. So you test each word after you write it to make sure it works as you intend it to)
1224 2012-02-05 18:59:04 <cjd> I'm always amazed at how much like C javascript really is
1225 2012-02-05 18:59:27 <cjd> for such a high level language, the code pattern is very similar to the bottom level
1226 2012-02-05 18:59:29 <Diablo-D3> Zarutian: okay, so Im writing a miner kernel in forth.
1227 2012-02-05 18:59:39 <Diablo-D3> cjd: javascript is closer to smalltalk than it is C
1228 2012-02-05 19:00:00 * cjd will have to lookup smalltalk
1229 2012-02-05 19:00:04 <justmoon> cjd: I recently answered a stackoverflow question and gave javascript code example - then I wanted to add a C version - I changed like "var" -> "int" and that was it
1230 2012-02-05 19:00:15 <cjd> ^^exactly
1231 2012-02-05 19:00:15 <Zarutian> cjd: well C got its syntax from algol60 iirc and javascript just bolted that onto a system that is akin to scheme and smalltalk.
1232 2012-02-05 19:00:24 <luke-jr> var foo = <xml><garbage id='g'/></xml>; foo.getElementById('g');
1233 2012-02-05 19:00:26 <luke-jr> *so* like C…
1234 2012-02-05 19:00:59 <Zarutian> Diablo-D3: yes, go on.
1235 2012-02-05 19:01:00 <Diablo-D3> luke-jr: erm, whats with the unquoted string there?
1236 2012-02-05 19:01:07 <cjd> you can do that? o_O
1237 2012-02-05 19:01:13 <luke-jr> justmoon: you can write C-like code in any language :P
1238 2012-02-05 19:01:19 <luke-jr> Diablo-D3: it's not a string
1239 2012-02-05 19:01:25 <luke-jr> cjd: yes, officially anyhow
1240 2012-02-05 19:01:26 <justmoon> there a little supported extension that allows xml in javascript in some interpreters
1241 2012-02-05 19:01:27 <Diablo-D3> you just set foo to a string of xml
1242 2012-02-05 19:01:41 <Diablo-D3> justmoon: its still improper js.
1243 2012-02-05 19:01:43 <luke-jr> Diablo-D3: no, I set it to an XML DOM object
1244 2012-02-05 19:01:51 <cjd> java resists efforts to write C like code
1245 2012-02-05 19:01:58 <Diablo-D3> luke-jr: you do realize javascript is used for things other than xml/sgml dom manipulation, right?
1246 2012-02-05 19:01:58 <Zarutian> nope, not on my watch. String literals in languages that rely on yacc kind of parsing is a big nono in my books.
1247 2012-02-05 19:02:11 <Diablo-D3> cjd: bullshit
1248 2012-02-05 19:02:12 <Diablo-D3> ask TD
1249 2012-02-05 19:02:14 <Diablo-D3> hes seen my java
1250 2012-02-05 19:02:17 <Diablo-D3> its VERY C-like
1251 2012-02-05 19:02:20 <luke-jr> http://www.ecma-international.org/publications/files/ECMA-ST/Ecma-357.pdf
1252 2012-02-05 19:02:29 <cjd> I didn't say it was impossible
1253 2012-02-05 19:02:38 <cjd> it's just that the language resists it
1254 2012-02-05 19:02:51 <Diablo-D3> heh, no it really doesnt
1255 2012-02-05 19:02:54 <TD> it's not so bad
1256 2012-02-05 19:03:03 <Diablo-D3> java is very easy to write correctly.
1257 2012-02-05 19:03:12 <Diablo-D3> I dont care what those javaee faggots in corporate land do
1258 2012-02-05 19:03:14 <Zarutian> the only thing I like about java is Minecraft and classnamer.com
1259 2012-02-05 19:03:15 <Diablo-D3> or notch, for that matter
1260 2012-02-05 19:03:17 <justmoon> "E4X is supported by Mozilla's Rhino, used in OpenOffice.org and several other projects, and SpiderMonkey, used in Firefox, Thunderbird, and other XUL-based applications. It is also supported by Tamarin, the JavaScript engine used in the Flash virtual machine. It is not supported by Nitro (Safari), V8 (Google Chrome), Carakan (Opera), nor Internet Explorer."
1261 2012-02-05 19:03:22 <luke-jr> quote from the spec: var person = <person><name>John</name><age>25</age></person>;
1262 2012-02-05 19:03:35 <Diablo-D3> thats disgusting.
1263 2012-02-05 19:03:45 <Diablo-D3> good luck making that do ANYTHING on something that isnt js manipulating a dom
1264 2012-02-05 19:03:47 <Diablo-D3> like node.js
1265 2012-02-05 19:03:55 Graet has quit (Ping timeout: 248 seconds)
1266 2012-02-05 19:03:59 <cjd> user@debo8:~$ node
1267 2012-02-05 19:03:59 <cjd> > var person = <person><name>John</name><age>25</age></person>;
1268 2012-02-05 19:03:59 <cjd> ...
1269 2012-02-05 19:04:07 denisx_ has joined
1270 2012-02-05 19:04:11 <cjd> it gave me the old …
1271 2012-02-05 19:04:22 <justmoon> cjd: v8 doesn't support it, it's a mozilla specific extension
1272 2012-02-05 19:04:32 <justmoon> and mozilla deprecated it as well
1273 2012-02-05 19:04:33 <Zarutian> so it is a xml literal. It still should be quoted.
1274 2012-02-05 19:04:38 <justmoon> https://developer.mozilla.org/en/E4X
1275 2012-02-05 19:04:40 <cjd> figures they would be the ones to come up with such a silly idea
1276 2012-02-05 19:04:41 <Diablo-D3> [01:54:30] <Zarutian> Diablo-D3: yes, go on.
1277 2012-02-05 19:04:41 <k9quaint> javascript is such an abortion
1278 2012-02-05 19:04:45 <Diablo-D3> btw, theres nothing to go on
1279 2012-02-05 19:04:50 <Diablo-D3> Im brutally raping AMD's compiler
1280 2012-02-05 19:04:54 <Diablo-D3> k9quaint: thats the thing, its not
1281 2012-02-05 19:04:57 <Diablo-D3> its how fucking faggots used it
1282 2012-02-05 19:04:58 <luke-jr> justmoon: it's not Mozilla-specific, it's an ECMA standard
1283 2012-02-05 19:05:04 <Diablo-D3> its another language ruined by the userbase
1284 2012-02-05 19:05:07 <Diablo-D3> same with java
1285 2012-02-05 19:05:13 <Zarutian> Diablo-D3: I meant the writing a miner in forth.
1286 2012-02-05 19:05:32 <Diablo-D3> Zarutian: no, but Im checking the stack and op layout every minor edit
1287 2012-02-05 19:05:39 <cjd> luke-jr: POSEX requires that 2100 is a leapyear, standards are to be taken with a grain of salt
1288 2012-02-05 19:05:40 <Diablo-D3> Im beating the compiler with a crowbar
1289 2012-02-05 19:05:41 denisx has quit (Read error: Operation timed out)
1290 2012-02-05 19:05:41 denisx_ is now known as denisx
1291 2012-02-05 19:05:54 <Diablo-D3> cjd: POSIX damnit
1292 2012-02-05 19:05:55 <luke-jr> cjd: fair enough ;)
1293 2012-02-05 19:05:55 <justmoon> luke-jr: no, it's not in the standard: http://www.ecma-international.org/publications/standards/Ecma-262.htm
1294 2012-02-05 19:06:06 * cjd can't spell
1295 2012-02-05 19:06:08 <luke-jr> justmoon: it's ECMA-357
1296 2012-02-05 19:06:19 <justmoon> luke-jr: javascript is ECMA-262
1297 2012-02-05 19:06:20 <Diablo-D3> s/EMCA-/Magnum /
1298 2012-02-05 19:07:01 <Diablo-D3> Colt Pythons... or the only thing nearly as awesome as a crowbar.
1299 2012-02-05 19:07:02 <Zarutian> saw this gag in an picture from an programming shop once: "We are looking for a suffiently smart compiler. Will pay 5 cents per statement compiled."
1300 2012-02-05 19:07:12 <Diablo-D3> Zarutian: :D
1301 2012-02-05 19:07:57 <Zarutian> and frankly that isnt such a bad idea for very time critical hard real time code.
1302 2012-02-05 19:08:43 <Diablo-D3> Im pretending Im a two stage compiler.
1303 2012-02-05 19:09:14 <Diablo-D3> its taking me now 4 days to do the first stage.
1304 2012-02-05 19:09:16 * cjd is a two stage compiler pretending to be human
1305 2012-02-05 19:09:16 <splatster> Selling $4.58 Amazon gift card
1306 2012-02-05 19:09:22 <splatster> oops
1307 2012-02-05 19:09:30 <splatster> thought this was -otc
1308 2012-02-05 19:09:38 <Diablo-D3> Ive shaved 1 cycle off 2.1 -v 2, and am quickly catching 2.6 up to 2.1
1309 2012-02-05 19:10:08 <Zarutian> Diablo-D3: what exactly are you handcompiling? a mining client?
1310 2012-02-05 19:10:11 denisx has quit (Remote host closed the connection)
1311 2012-02-05 19:10:27 denisx has joined
1312 2012-02-05 19:10:31 <Diablo-D3> Zarutian: the DM kernel
1313 2012-02-05 19:11:00 <Zarutian> Diablo-D3: I know too many acronyms. What does DM stand for in this context, please?
1314 2012-02-05 19:11:46 <Eliel> it's DiabloMiner
1315 2012-02-05 19:12:04 <Diablo-D3> DiabloMiner!
1316 2012-02-05 19:13:32 denisx_ has joined
1317 2012-02-05 19:14:18 <TD> justmoon: http://code.google.com/p/bitcoinj/wiki/ReleaseNotes?ts=1328465774&updated=ReleaseNotes
1318 2012-02-05 19:14:26 <TD> justmoon: stuff in the new release
1319 2012-02-05 19:14:50 <TD> should be out soon, hopefuly
1320 2012-02-05 19:16:12 <justmoon> TD: great stuff! offline transactions are cool!
1321 2012-02-05 19:16:18 denisx has quit (Ping timeout: 240 seconds)
1322 2012-02-05 19:16:18 denisx_ is now known as denisx
1323 2012-02-05 19:16:36 <TD> yeah. after 0.4 i want to change the wallet api so that you can give it some arbitrary incomplete transaction, and it finishes it off for you
1324 2012-02-05 19:16:42 <TD> it's not so hard to do actually. just a refactoring.
1325 2012-02-05 19:16:59 phantomfake has joined
1326 2012-02-05 19:17:04 <TD> then to create, say, a multi-send, you just create a fresh Transaction, add some outputs as needed, then tell the wallet "plz finish". it will add enough inputs, sign them, add a change output, etc
1327 2012-02-05 19:17:36 <justmoon> interesting, now I want to do it like that too :)
1328 2012-02-05 19:18:08 <justmoon> TD: https://bitcointalk.org/index.php?topic=62718.0
1329 2012-02-05 19:18:21 <TD> nice!
1330 2012-02-05 19:18:32 <TD> looking forward to it. let me know which train you're getting, i'll meet you on the platform
1331 2012-02-05 19:18:50 <justmoon> ok!
1332 2012-02-05 19:19:05 <justmoon> I have to get in touch with bart and find out when he'd like to meet up
1333 2012-02-05 19:19:08 <TD> cool
1334 2012-02-05 19:19:36 <justmoon> I'll bring a few books as reference material so I can prove to you once and for all why git is superior to svn
1335 2012-02-05 19:19:44 <justmoon> just a few bags full
1336 2012-02-05 19:19:46 <TD> hah
1337 2012-02-05 19:19:48 <TD> we already converted to git
1338 2012-02-05 19:19:52 <TD> i'm getting used to it :)
1339 2012-02-05 19:19:57 <justmoon> :D
1340 2012-02-05 19:20:01 onelineproof has joined
1341 2012-02-05 19:20:03 <TD> managed to screw up once already though. some kind soul helped me repair the damage
1342 2012-02-05 19:20:14 <TD> but the community hath spaketh
1343 2012-02-05 19:20:21 <TD> lots of clones appeared. so hopefully patches will follow
1344 2012-02-05 19:21:10 <Diablo-D3> heh git is nice
1345 2012-02-05 19:22:43 Graet has joined
1346 2012-02-05 19:22:56 iocor has quit (Quit: Computer has gone to sleep.)
1347 2012-02-05 19:22:57 <justmoon> TD: wow, the forum topic id of the first zurich meetup was 2716, this one is 62718
1348 2012-02-05 19:23:02 Graet is now known as Guest60357
1349 2012-02-05 19:23:03 <justmoon> seems like such a long time ago now
1350 2012-02-05 19:23:17 <TD> yeah. it's amazing how much things grew in just one year.
1351 2012-02-05 19:23:47 Guest60357 is now known as Graet
1352 2012-02-05 19:24:05 <luke-jr> TD: what's with Google's new ToS?
1353 2012-02-05 19:24:08 Graet has quit (Changing host)
1354 2012-02-05 19:24:08 Graet has joined
1355 2012-02-05 19:24:21 <TD> i don't think we changed the tos. you mean the privacy policy?
1356 2012-02-05 19:24:43 <luke-jr> TD: I mean ToS. It basically exempts Google from the requirements of GPL/AGPL if anyone uploads the code on Google's sites.
1357 2012-02-05 19:25:03 <Diablo-D3> no it doesnt.
1358 2012-02-05 19:25:18 <TD> i don't think google is going to violate the GPL anytime soon. we have an entire legal team and set of processes put in place to ensure we don't.
1359 2012-02-05 19:25:38 <TD> if the ToS can be read in such a way that it looks like we require that, it's probably a mistake
1360 2012-02-05 19:25:40 <luke-jr> TD: the point is the ToS gives them permission to do so legally.
1361 2012-02-05 19:25:44 * Diablo-D3 giggles like a schoolgirl as the op count goes down
1362 2012-02-05 19:25:53 <luke-jr> it grants them a license to basically do anything
1363 2012-02-05 19:26:10 <TD> i suspect it's the legal boilerplate that says we can create derivative works, etc.
1364 2012-02-05 19:26:17 <TD> that applies to any uploaded content
1365 2012-02-05 19:26:20 <luke-jr> right
1366 2012-02-05 19:26:30 <Diablo-D3> but creating a derivative of GPL is okay
1367 2012-02-05 19:26:33 <Diablo-D3> it just has to be GPL too
1368 2012-02-05 19:26:37 <Diablo-D3> so whats the issue?
1369 2012-02-05 19:26:48 <TD> i don't know. i don't deal with that. like i said, it sounds like an oversight/result of too-aggressive simplification, perhaps. you could raise the issue with cdibona, perhaps
1370 2012-02-05 19:26:55 <luke-jr> Diablo-D3: the ToS grants Google a license to create deriv *without* the GPL restrictions
1371 2012-02-05 19:27:07 <TD> or just send me a mail outlining the problem and i'll forward it to danny
1372 2012-02-05 19:27:07 <Diablo-D3> yes, and the moment they try I sue the fucking shit out of them
1373 2012-02-05 19:27:08 <luke-jr> no idea who that is
1374 2012-02-05 19:27:19 <luke-jr> Diablo-D3: and they win?
1375 2012-02-05 19:27:20 <Diablo-D3> their ToS does not preempt copyright law.
1376 2012-02-05 19:27:26 <TD> chris dibona runs the open source programs office. he deals with anything open source related.
1377 2012-02-05 19:27:53 <luke-jr> cdibona@google.com ?
1378 2012-02-05 19:28:45 <TD> actually
1379 2012-02-05 19:28:51 <TD> just email hearn@google.com and i'll find the right person
1380 2012-02-05 19:29:01 <luke-jr> k
1381 2012-02-05 19:31:38 marf_away has quit (Ping timeout: 245 seconds)
1382 2012-02-05 19:35:46 marf_away has joined
1383 2012-02-05 19:36:34 <justmoon> TD: do you have any contact details for cdecker? if so, can you notify him about the meetup in case he wants to come with us? if not, I'll send him a pm on the forums
1384 2012-02-05 19:36:47 <TD> sure
1385 2012-02-05 19:36:51 <justmoon> thx
1386 2012-02-05 19:36:53 <cdecker> Hi there
1387 2012-02-05 19:36:57 <justmoon> lol
1388 2012-02-05 19:36:57 <TD> ah ha
1389 2012-02-05 19:37:04 <TD> i was just about to forward the mail
1390 2012-02-05 19:37:18 <cdecker> Meetup?
1391 2012-02-05 19:37:25 <justmoon> cdecker: wanna go to geneva bitcoin meetup with us? saturday 11th
1392 2012-02-05 19:37:33 <TD> justmoon: maybe we should have a "swiss bitcoin" fb or +page that people can follow or something
1393 2012-02-05 19:37:38 <TD> rather than an ad hoc mailing list
1394 2012-02-05 19:37:43 <TD> though mail is probably more effective
1395 2012-02-05 19:37:44 <cdecker> Ah that would be great
1396 2012-02-05 19:37:55 <cdecker> But I have the deadline for my thesis this monoth
1397 2012-02-05 19:38:08 <justmoon> TD: feel free to set something up - me and social media are like gollum and social media
1398 2012-02-05 19:38:11 <cdecker> And my results aren't as conclusive as I would like them to be
1399 2012-02-05 19:38:40 <Moron__> screw your thesis!
1400 2012-02-05 19:38:51 <Moron__> ohps
1401 2012-02-05 19:38:52 <Moron__> i mean
1402 2012-02-05 19:38:58 <Moron__> to pieces
1403 2012-02-05 19:39:00 <cdecker> Hehe, I'm not that far away from actually dropping everything :D
1404 2012-02-05 19:39:27 <justmoon> the next meetup will be in zurich again, in 1-2 months hopefully, so no worries
1405 2012-02-05 19:39:42 <justmoon> good luck with the thesis - distributed computing?
1406 2012-02-05 19:40:00 <Moron__> cdecker: http://www.youtube.com/watch?v=gAYL5H46QnQ
1407 2012-02-05 19:40:02 <cdecker> Great, looking forward to meeting everybody again
1408 2012-02-05 19:40:26 <cdecker> Yep, distributed computing, crawling some poor unsuspecting Bittorrent users
1409 2012-02-05 19:40:52 <justmoon> cdecker: so you chickened out about doing a thesis on bitcoin? :P
1410 2012-02-05 19:41:03 <cdecker> They didn't allow me to
1411 2012-02-05 19:41:27 <justmoon> aww, well you'll have plenty of opportunity to research bitcoin later :)
1412 2012-02-05 19:41:27 <Moron__> you need to fight... for your right... to theorize (on bitcoin...)
1413 2012-02-05 19:41:31 <cdecker> But I'm not so far away from convincing the prof about opening some future thesis to BTC
1414 2012-02-05 19:41:43 <justmoon> Moron__: lol, love your cultural references
1415 2012-02-05 19:41:48 <TD> that's cool
1416 2012-02-05 19:42:03 <cdecker> Well if Google doesn't want me I'll start a PhD, and the prof is not that far away from allowing it :)
1417 2012-02-05 19:42:59 <justmoon> a PhD who knows what he's talking about - holy crap, if you do that you'll threaten my whole world view
1418 2012-02-05 19:43:11 <TD> lol
1419 2012-02-05 19:43:43 <justmoon> (that joke makes slightly more sense in the economics depts where I picked it up)
1420 2012-02-05 19:43:49 <cdecker> Well the Prof wants me exactly because I'm not a theory guy
1421 2012-02-05 19:44:28 <justmoon> I'd love to do a phd, if I had an extra life
1422 2012-02-05 19:44:42 <TD> i could never go back to university. hated it there.
1423 2012-02-05 19:44:49 <TD> 3 years and i was out :)
1424 2012-02-05 19:45:01 <justmoon> 6 months here, couldn't take it
1425 2012-02-05 19:45:04 <Moron__> im all into theoretics but somethings hypothosizing just gets tiring and you wanna write an antithesis
1426 2012-02-05 19:45:15 <TD> was so nice to have free weekends for the first time. though, now  i'm doing german homework at the moment :)
1427 2012-02-05 19:45:29 <TD> education is dragging me back in :)
1428 2012-02-05 19:45:32 <cdecker> Need help?
1429 2012-02-05 19:45:34 <cdecker> :-)
1430 2012-02-05 19:45:52 <TD> it's fine :) a bit boring
1431 2012-02-05 19:46:13 <TD> Warum sind Frau Schusters Präsentation effektiv? Sie spricht deutlich, und ihren Grafik sind logisch und klar.
1432 2012-02-05 19:46:31 <cdecker> Mostly korrekt :D
1433 2012-02-05 19:46:46 <TD> Vor einer Präsentation hat man oft <Lampenfieber>. Aber man muss <selbstsicher> sein, wenn man eine Präsentation hält
1434 2012-02-05 19:46:48 <onelineproof> I was in geneva last week, just returned. I was trying to find someone to sell my CHF for bitcoins.
1435 2012-02-05 19:46:48 <TD> etc etc
1436 2012-02-05 19:46:59 <onelineproof> Next time I go, I wanna know who to ask.
1437 2012-02-05 19:47:03 <TD> onelineproof: whereabouts are you now? i do chf/btc trades in zurich sometimes
1438 2012-02-05 19:47:14 <onelineproof> no im in Canada now
1439 2012-02-05 19:47:23 <TD> p2p exchange needs a lot of work
1440 2012-02-05 19:47:28 <TD> none of the map sites ever got critical mass
1441 2012-02-05 19:47:41 <onelineproof> i tried btcnearme and bitcoin-otc
1442 2012-02-05 19:48:12 <PK> td: happen to be in bern sometime? I would buy some btc
1443 2012-02-05 19:48:39 <justmoon> doesn't our train go through bern? lol
1444 2012-02-05 19:48:41 <cdecker> Anyway, got to go
1445 2012-02-05 19:48:53 <cdecker> Let me know about the next meetup :-)
1446 2012-02-05 19:48:57 <justmoon> cdecker: cya, send thesis when done!
1447 2012-02-05 19:49:00 <justmoon> will do
1448 2012-02-05 19:49:05 <cdecker> Will do ^^
1449 2012-02-05 19:49:22 <TD> yeah
1450 2012-02-05 19:49:25 <lianj> TD: in the above case, Präsentationen and ihre
1451 2012-02-05 19:49:41 <TD> PK: normally not, but indeed on saturday we'll pass through bern :) why not join the meetup?
1452 2012-02-05 19:49:45 <TD> lianj: thanks
1453 2012-02-05 19:50:00 * luke-jr ponders writing a script to launch an Ubuntu VM to run gitian in <.<
1454 2012-02-05 19:50:09 <PK> td: where's the meetup?
1455 2012-02-05 19:50:19 <TD> https://bitcointalk.org/index.php?topic=62718.0
1456 2012-02-05 19:50:25 <TD> in geneva. see link.
1457 2012-02-05 19:50:31 dwon has quit (Quit: Leaving)
1458 2012-02-05 19:50:31 <PK> thanks
1459 2012-02-05 19:50:46 <PK> but... they speak french and crazy stuff there..
1460 2012-02-05 19:50:47 <TD> PK: otherwise i'm happy to get off at that stop, do the trade and then get back on
1461 2012-02-05 19:50:55 <justmoon> PK: meetup is in english, don't worry
1462 2012-02-05 19:50:57 <TD> oui, c'est vrai
1463 2012-02-05 19:51:03 <PK> td: speed trade or next train? :)
1464 2012-02-05 19:51:08 <TD> haha
1465 2012-02-05 19:51:16 <TD> i'm not sure mobile wallets are fast enough to do it whilst the train is on the platform
1466 2012-02-05 19:51:24 <TD> next train
1467 2012-02-05 19:51:53 <justmoon> yeah we could go grab a coffee if you want
1468 2012-02-05 19:52:41 <PK> or I just join you guys in the train to geneva. Advantage of having a GA, doesn't cost me anything but a bit of time.
1469 2012-02-05 19:53:03 <justmoon> do it, bitcoin meetups are the best!
1470 2012-02-05 19:53:27 <TD> yep!
1471 2012-02-05 19:53:29 Nicksasa has quit (Read error: Connection reset by peer)
1472 2012-02-05 19:53:34 Nicksasa has joined
1473 2012-02-05 19:53:42 <TD> i wish there was a website that had an easy to remember address and converted bitcoin URIs into QRcodes for you
1474 2012-02-05 19:53:54 <TD> http://www.paymyaddress.com/
1475 2012-02-05 19:53:57 <TD> that's pretty cool
1476 2012-02-05 19:53:59 <lianj> like key signing parties, but for coins :P
1477 2012-02-05 19:54:11 <TD> except the qrcodes seem to be malformatted
1478 2012-02-05 19:54:48 a_meteorite is now known as a_meteor
1479 2012-02-05 19:54:51 <TD> PK: do you have an android wallet? if not then ensuring you have a qrcode i can scan would be useful
1480 2012-02-05 19:55:03 <justmoon> PK, TD: I'm looking at trains - how about the one that leaves zurich 10:32, bern 11:34, geneva 13:15 - then we have lunch with bart, then the meetup
1481 2012-02-05 19:55:10 a_meteor is now known as a_meteorite
1482 2012-02-05 19:55:12 <TD> sounds good
1483 2012-02-05 19:55:14 <PK> no, I'll let you type it all in manually >:P
1484 2012-02-05 19:55:50 <TD> well. i guess that's what the checksum is for :)
1485 2012-02-05 19:55:58 <TD> or email me an address, old school but it works
1486 2012-02-05 19:56:22 <PK> or post it here and hope on some stray coins ending up :)
1487 2012-02-05 19:56:41 <justmoon> if you post it here, people will be able to trace your coins!!!!111oneone
1488 2012-02-05 19:57:04 <PK> I'm fine with that. Got nothing to hide.
1489 2012-02-05 19:57:37 <justmoon> guess there goes that whole "bitcoin is the currency for druggies and arms dealers" theory
1490 2012-02-05 19:57:54 <TD> PK: out of interest, what kind of phone do you have ?
1491 2012-02-05 19:58:09 <PK> we swiss use trust old number bank accounts for that :P
1492 2012-02-05 19:58:18 <justmoon> lol
1493 2012-02-05 19:58:31 <PK> TD: x10
1494 2012-02-05 19:58:44 <TD> that's the sony ericsson android right
1495 2012-02-05 19:58:46 <PK> yes
1496 2012-02-05 19:59:04 <TD> ah ha
1497 2012-02-05 19:59:07 <TD> btc.to can do it
1498 2012-02-05 19:59:15 <TD> kind of unintuitive, you have to edit the url by hand.
1499 2012-02-05 19:59:37 <PK> is there no good, trusty android app?
1500 2012-02-05 20:00:03 <justmoon> TD: can't google charts api do it?
1501 2012-02-05 20:00:04 <TD> there are a couple that work ok
1502 2012-02-05 20:00:07 booo has quit (Ping timeout: 272 seconds)
1503 2012-02-05 20:00:11 npouillard has quit (Quit: leaving)
1504 2012-02-05 20:00:17 <TD> justmoon: i want something easy damnit :) and which formats the qrcode properly. so far i found at least two that were type: text
1505 2012-02-05 20:00:28 <TD> PK: there is "Bitcoin Wallet" by Andreas Schildbach. it connects directly to the p2p network
1506 2012-02-05 20:00:40 <TD> PK: the first time you install it, it'll take a few minutes to catch up with the chain head.
1507 2012-02-05 20:00:53 <TD> PK: there's also BitcoinSpinner that relies on a third party server to do all the heavy lifting. so it's faster.
1508 2012-02-05 20:00:57 <TD> but i think it has fewer features also
1509 2012-02-05 20:01:09 <PK> TD, nothing offline?
1510 2012-02-05 20:01:12 <TD> PK: with "Bitcoin Wallet" it puts a qrcode on the screen so it's pretty easy
1511 2012-02-05 20:01:30 <TD> PK: well, you don't need to be online. just me.
1512 2012-02-05 20:01:30 <PK> or that only connects to send trx without being a full wallet checking all transactions?
1513 2012-02-05 20:01:39 <TD> neither are full wallets
1514 2012-02-05 20:01:51 <TD> bitcoin wallet is much lighter weight than the regular satoshi client
1515 2012-02-05 20:02:21 <justmoon> I'd say just try bitcoin wallet and see if you like it
1516 2012-02-05 20:02:38 <TD> it still walks the chain but doesn't process most of it. so it can download blocks much faster. at any rate, you don't need to be online to receive coins, i can just grab the address from your screen and then you'll get the coins later, when you let it sync up
1517 2012-02-05 20:02:45 <TD> you can send them on to a desktop wallet, mtgox, whatever you want
1518 2012-02-05 20:02:47 <PK> what's the data traffic it eats?
1519 2012-02-05 20:03:26 <TD> good question. i'm not sure. more than bitcoinspinner. if you're very sensitive to data costs, that might be a better approach. or just write down an address from your desktop wallet :)
1520 2012-02-05 20:04:22 <k9quaint> war of the two letter nicknames!
1521 2012-02-05 20:05:25 * justmoon wonders what kind of person gets a flatrate for public transport, but not for his mobile phone ;P
1522 2012-02-05 20:05:46 <TD> sigh, got an ANR from bitcoin wallet. only brief.
1523 2012-02-05 20:05:56 <JFK911> pretty easy.  transport is necessary, telephone is not
1524 2012-02-05 20:05:57 <TD> seems i need to make reading the chain head lock free
1525 2012-02-05 20:06:07 * Eliel <-- has flatrate for public transport but not for mobile phone.
1526 2012-02-05 20:06:30 <JFK911> i own a car but no telephone
1527 2012-02-05 20:06:39 <justmoon> JFK911: but a GA costs like 100 times what a mobile flat costs here
1528 2012-02-05 20:06:55 Cablesaurus has joined
1529 2012-02-05 20:06:55 Cablesaurus has quit (Changing host)
1530 2012-02-05 20:06:55 Cablesaurus has joined
1531 2012-02-05 20:07:08 <JFK911> what's a GA ?
1532 2012-02-05 20:07:09 <TD> PK: how many BTC would you want to buy?
1533 2012-02-05 20:07:09 <Eliel> ... well, in practise my mobile phone contract tends to be very close to flatrate though.
1534 2012-02-05 20:07:17 * TD gets 1gig a month and it's plenty
1535 2012-02-05 20:07:22 splatster is now known as splatster|away
1536 2012-02-05 20:07:24 <TD> PK: if you sync it on wifi, no problem, right ?
1537 2012-02-05 20:07:39 <justmoon> JFK911: let's you ride any bus, train etc in switzerland
1538 2012-02-05 20:07:40 <PK> justmoon: someone who needs pubilc transport more than mobile internet. I usually have a wlan to connect to with my mobile. So 250MB is more than enough.
1539 2012-02-05 20:08:02 graingert has joined
1540 2012-02-05 20:08:09 <justmoon> PK: ahh, ic, 250MB is more than enough for bitcoin wallet too, if you sync it regularly from your wifi
1541 2012-02-05 20:08:36 <JFK911> Interesting, my experience in europe is mobiles cost even more than usa
1542 2012-02-05 20:08:36 <JFK911> the places where mobile services are worth their cost are places like venezuela, russia, somalia.
1543 2012-02-05 20:08:46 <JFK911> no way im paying T-mobile 4 euros for a mbyte
1544 2012-02-05 20:08:54 <PK> btw, mobile internet flat costs around 600-700 CHF a year, a GA only 3500 or so. Only 6 times more, not 100 times :P
1545 2012-02-05 20:08:55 <JFK911> because i am prepaid or roaming
1546 2012-02-05 20:08:59 <Eliel> ... oh wait, I do have a flatrate for data.
1547 2012-02-05 20:09:23 <Eliel> 10€/month for unlimited data (although bandwidth is limited to 1Mbps)
1548 2012-02-05 20:09:34 <JFK911> how many dollars has a franc?  like 2?
1549 2012-02-05 20:09:46 <JFK911> oh wait i have a 1000 chf note so probably not so much
1550 2012-02-05 20:09:54 <luke-jr> are there gitian builds of boost and qt published?
1551 2012-02-05 20:10:14 <justmoon> 1 Swiss franc = 1.09087 U.S. dollars
1552 2012-02-05 20:10:19 <PK> JFK911: some providers have roaming packages. With swisscom I pay 7 CHF a day for 10MB. Still a lot but it is improving, slowly.
1553 2012-02-05 20:10:44 <JFK911> I'd pay $7/day, if I could choose a day and pay $7
1554 2012-02-05 20:10:50 <JFK911> but most days I won't use it at all
1555 2012-02-05 20:11:24 <JFK911> the transport pass seems like a superb value compared to the cost and taxes of a car
1556 2012-02-05 20:11:31 <PK> yea, it's a day and max 10MB. You can choose the days. Actually it does it automatically. Charge you per kb until you reach 7 CHF and then it's flat until you hit the 10MB
1557 2012-02-05 20:11:40 <TD> PK: i pay around 25 CHF a month and get 1GB a month
1558 2012-02-05 20:11:45 <TD> so 300 CHF a year not 600
1559 2012-02-05 20:11:46 <JFK911> that's fair!
1560 2012-02-05 20:11:52 <PK> TD you're under 26?
1561 2012-02-05 20:11:55 <TD> no
1562 2012-02-05 20:11:59 <JFK911> So strange to hear about a telco that's not predatory
1563 2012-02-05 20:11:59 <TD> http://www1.orange.ch/en/mobile/options-and-services/data-and-internet/mobile-internet/
1564 2012-02-05 20:12:05 <TD> i have "mobile internet max"
1565 2012-02-05 20:12:21 <PK> ok, orange. I have a orange stick that's flat for 45 CHF a month. But that's orange.
1566 2012-02-05 20:12:36 * justmoon has an irrational fear that his skin will turn orange if he gets a contract with orange
1567 2012-02-05 20:12:40 <TD> no good coverage for you?
1568 2012-02-05 20:13:41 <PK> no
1569 2012-02-05 20:13:59 copumpkin has quit (Quit: Computer has gone to sleep.)
1570 2012-02-05 20:17:03 <Moron__> chf??
1571 2012-02-05 20:17:12 <TD> swiss franc
1572 2012-02-05 20:18:29 <onelineproof> For testing, should I use https://github.com/bitcoin/bitcoin/tarball/master
1573 2012-02-05 20:20:41 etotheipi_ has quit (Ping timeout: 245 seconds)
1574 2012-02-05 20:23:23 <gmaxwell> wumpus: someone in p2pool is saying they are literally getting "%n second ago" from a tooltip in bitcoin-qt.
1575 2012-02-05 20:23:56 <gmaxwell> wumpus: I looked at the localizations and I don't see why, unless %n is supposted to be replaced with %1 there (but if so, then there are a bunch that are incorrect)
1576 2012-02-05 20:24:38 RazielZ has quit (Ping timeout: 248 seconds)
1577 2012-02-05 20:25:06 RazielZ has joined
1578 2012-02-05 20:26:33 occulta has quit (Ping timeout: 276 seconds)
1579 2012-02-05 20:26:52 splatster has quit (away!~splatster@unaffiliated/splatster|Quit: Leaving...)
1580 2012-02-05 20:28:28 danbri has quit (Remote host closed the connection)
1581 2012-02-05 20:28:28 chmod755 has joined
1582 2012-02-05 20:30:30 <PK> ok, I go with bitcoin spinner then. The bitcoin wallet used 300MB just for 10 days of chain to download. :O
1583 2012-02-05 20:31:56 <TD> the full chain is only 893mb and that's two years worth
1584 2012-02-05 20:32:01 <TD> so i doubt 10 days was 300mb
1585 2012-02-05 20:32:18 <TD> but it's true that bitcoin wallet is more bandwidth intensive. i think andreas is going to implement a "nighttime catchup" mode
1586 2012-02-05 20:32:22 <chmod755> hi, i had some orphaned blocks and tried to move all my btc to a new wallet so i only have my real btcs
1587 2012-02-05 20:32:23 genjix has joined
1588 2012-02-05 20:32:31 <TD> so it can do the bulk of the sync when it's charging and has home wifi access
1589 2012-02-05 20:32:35 <PK> hey genjix
1590 2012-02-05 20:32:46 <genjix> hey PK
1591 2012-02-05 20:32:53 <genjix> PK: TD is from switzerland too
1592 2012-02-05 20:33:02 <genjix> there's a geneva meeting
1593 2012-02-05 20:33:05 <TD> i'm from the uk. i live in switzerland :)
1594 2012-02-05 20:33:07 <PK> genjix: we figured out :)
1595 2012-02-05 20:33:13 <chmod755> but it looks like *ALL* of my transactions were rejected
1596 2012-02-05 20:33:14 <genjix> oh cool :)
1597 2012-02-05 20:33:53 <TD> PK: you're on wifi though, right ?
1598 2012-02-05 20:34:14 <PK> TD: right now, yes. And it's full traffic, not just the download. There is protocol overhead and IRC stuff too.
1599 2012-02-05 20:35:50 <TD> it doesn't use IRC discovery. by default it uses DNS
1600 2012-02-05 20:35:53 <TD> that surprises me a lot
1601 2012-02-05 20:35:58 <TD> i can't think of any reason it'd use 300mb of bandwidth
1602 2012-02-05 20:36:03 <genjix> PK: https://bitcointalk.org/index.php?topic=30646.msg730898#msg730898
1603 2012-02-05 20:36:13 <TD> are you sure that was only the wallet app? ICS breaks down bandwidth usage for you
1604 2012-02-05 20:36:56 m00p has quit (Ping timeout: 255 seconds)
1605 2012-02-05 20:37:17 <PK> TD: It was not only the wallet, but mostly. Maybe it wasn't IRC but it said something about 6 connections. I assume it would have some overhead there.
1606 2012-02-05 20:37:46 Xunie has joined
1607 2012-02-05 20:38:04 <genjix> hey Xunie
1608 2012-02-05 20:38:19 <Xunie> Sup?
1609 2012-02-05 20:38:19 <TD> PK: the connections are fairly light. you see transaction announcements and stuff
1610 2012-02-05 20:38:29 <TD> PK: it may be a bug that caused it, i can guess
1611 2012-02-05 20:38:31 copumpkin has joined
1612 2012-02-05 20:38:33 <TD> there's a big upgrade coming along
1613 2012-02-05 20:38:38 <TD> it'll reduce bandwidth consumption for new users a lot
1614 2012-02-05 20:38:48 <genjix> which one?
1615 2012-02-05 20:38:48 <TD> need to fix the bugs and get testing first though
1616 2012-02-05 20:38:55 <genjix> ah bitcoinj
1617 2012-02-05 20:39:07 <Moron__> is it possible to flip a bitcoin?
1618 2012-02-05 20:39:18 <PK> TD: it would be nice to have a lightweight app that uses blockexplorer as backend.
1619 2012-02-05 20:39:42 <genjix> Moron__: no, the scripting system is completely deterministic
1620 2012-02-05 20:39:49 <TD> PK: that's pretty much the BitcoinSpinner model
1621 2012-02-05 20:39:52 <genjix> PK: that's stratum
1622 2012-02-05 20:40:04 <genjix> #stratum + #electrum
1623 2012-02-05 20:40:12 <TD> PK: you tell a remote server your public keys and what you want to do. it gives you back a transaction to sign
1624 2012-02-05 20:40:16 <PK> genjix: is that an android app too?
1625 2012-02-05 20:40:31 <TD> PK: BitcoinSpinner is an android app yes
1626 2012-02-05 20:40:34 <genjix> no, stratum is an api for a full node (like json rpc++)
1627 2012-02-05 20:40:37 <TD> PK: it has some advantages and disadvantages to the Wallet model
1628 2012-02-05 20:40:48 <genjix> electrum is simply one client using stratum
1629 2012-02-05 20:41:00 <PK> TD: what's the disadvantage?
1630 2012-02-05 20:41:21 <PK> TD: I know BitcoinSpinner is, I have that installed now :)
1631 2012-02-05 20:41:45 <TD> you rely on a semi-trusted server. if it goes away you have to find a new one. there's no privacy: the server knows all your transactions and your balance.
1632 2012-02-05 20:41:55 <TD> the server can make you believe you have received payments you haven't
1633 2012-02-05 20:42:00 ThomasV has joined
1634 2012-02-05 20:42:32 <PK> yea, that's why a client using blockexplorer (instead of its own private servers) would be more appealing. But that's just paranoid me.
1635 2012-02-05 20:42:41 <TD> well, the model is the same
1636 2012-02-05 20:42:49 <PK> almost
1637 2012-02-05 20:42:49 <TD> block explorer had an outage the other day. so your app  would have broken.
1638 2012-02-05 20:43:00 <TD> the full p2p model is almost always harder to get right
1639 2012-02-05 20:43:06 <genjix> PK: try this https://bitcointalk.org/index.php?topic=50721.0
1640 2012-02-05 20:43:22 <TD> i think they'll converge over time and the advantages of the super-lightweight model will erode away.
1641 2012-02-05 20:43:26 <genjix> you can make a wrapper using that and blockexplorer/blockchain.info
1642 2012-02-05 20:43:27 <TD> but we'll see
1643 2012-02-05 20:44:32 <PK> nice. I couild use that to make a webbased bitcoin wallet that I host on my own server.
1644 2012-02-05 20:44:36 <TD> bitcoinj isn't as efficient as it can be, for sure, and protocol extensions can eliminate most of the delta left between the BCCAPI/electrum model and SPV clients
1645 2012-02-05 20:44:42 <justmoon> TD: some MTUO ideas by me: http://dcao.org/wiki/index.php?title=User:Justmoon/IMTUO and one by cjd: http://btc.pastebay.org/144544
1646 2012-02-05 20:44:54 <justmoon> TD: to be read at your leisure
1647 2012-02-05 20:45:09 <TD> justmoon: the first link requires a login. what's it about?
1648 2012-02-05 20:45:15 <justmoon> ah crap
1649 2012-02-05 20:45:23 <TD> ah ha
1650 2012-02-05 20:45:23 <justmoon> https://en.bitcoin.it/wiki/User:Justmoon/IMTUO
1651 2012-02-05 20:45:55 booo has joined
1652 2012-02-05 20:45:58 splatster has joined
1653 2012-02-05 20:46:31 splatster has quit (Client Quit)
1654 2012-02-05 20:46:43 <TD> justmoon: yes this kind of thing is very good
1655 2012-02-05 20:47:36 <justmoon> TD: I've been struggling with a method to do this if/while it doesn't have majority miner support yet
1656 2012-02-05 20:47:43 <justmoon> maybe some sort of merged mining type thing would work
1657 2012-02-05 20:47:57 <justmoon> since you're the mm guru, maybe you could think about it for a bit?
1658 2012-02-05 20:48:24 <TD> i don't fully understand why the vinfo does not include output index
1659 2012-02-05 20:48:48 <TD> i'd like to see blocks be transmitted as lists of tx hashes indeed. that seems like a simple, independent improvement
1660 2012-02-05 20:49:11 <justmoon> because the outputs are indexed by a hash (defined under heading MTUO) and you have everything to figure out what that hash is
1661 2012-02-05 20:49:44 <justmoon> you need to know the outputs outpoint (which is in the input), the script and the value (which are in the vinfo)
1662 2012-02-05 20:49:56 <TD> oh, sorry
1663 2012-02-05 20:49:57 <TD> i see
1664 2012-02-05 20:49:58 datagutt has quit (Quit: kthxbai)
1665 2012-02-05 20:49:58 FROTUSCI has joined
1666 2012-02-05 20:50:44 splatster has joined
1667 2012-02-05 20:51:37 <justmoon> take a look, it's just an idea, I'm not sure the network tradeoff is worth it - maybe bandwidth will be a more precious resource than a couple tb worth of storage anyway
1668 2012-02-05 20:52:43 <TD> i'd like to see it broken down into less ambitious goals
1669 2012-02-05 20:52:59 <TD> like, first goal, optimize the protocol by allowing propagating blocks to be broadcast as header+hashes.
1670 2012-02-05 20:53:26 <TD> second goal, start calculating the MTUOs and putting the roots into the coinbase, get agreement on it, have nodes check each others MTUOs
1671 2012-02-05 20:53:28 <justmoon> I actually had a section like that, but removed it because there were too many problems/unknowns with it still
1672 2012-02-05 20:53:37 <TD> discourage blocks with incorrect MTUOs
1673 2012-02-05 20:53:45 <TD> it's useful for lightweight clients
1674 2012-02-05 20:53:53 <justmoon> "discourage"?
1675 2012-02-05 20:53:56 <TD> even without changing the whole things around for regular clients
1676 2012-02-05 20:54:29 <TD> miners calculate the tree independently. if they see a broadcast block that connects to the best chain with an incorrect MTUO root, don't build on it
1677 2012-02-05 20:54:33 <TD> but don't fork the chain
1678 2012-02-05 20:54:40 <TD> if another block appears that builds on it, go with that
1679 2012-02-05 20:55:25 <justmoon> that would punish the punisher as much as the punishee if the punishers aren't a majority yet
1680 2012-02-05 20:55:48 <TD> yes. i think reaching the point where lightweight clients can verify free standing/broadcast transactions without needing a full block index would be enough work to start with
1681 2012-02-05 20:56:18 barmstrong has quit (Remote host closed the connection)
1682 2012-02-05 20:56:39 <TD> though i suppose they could not generate the vinfos
1683 2012-02-05 20:56:40 <TD> hmm
1684 2012-02-05 20:56:48 copumpkin has quit (Quit: Computer has gone to sleep.)
1685 2012-02-05 20:56:51 <TD> yeah i need to think about it more
1686 2012-02-05 20:57:35 chmod755 has left ("Leaving.")
1687 2012-02-05 20:57:37 <justmoon> other proposals just regenerate the whole merkle tree for every block - intuitively that seems like it would be expensive
1688 2012-02-05 20:57:48 <justmoon> not for the hashing, but for the loading all the hashes
1689 2012-02-05 20:58:02 <TD> yeah
1690 2012-02-05 20:58:09 <TD> i'm not really worried about storage/bandwidth for full nodes
1691 2012-02-05 20:58:19 <gmaxwell> TD: I have really mixed feelings about the value there. I think unspent transaction committments aren't actually a helpful fit with the SPV security model.
1692 2012-02-05 20:58:36 baz has quit (Remote host closed the connection)
1693 2012-02-05 20:58:36 m00p has joined
1694 2012-02-05 20:58:37 <TD> allowing low-trust verification of a transaction without needing all the data seems kinda valuable, except i guess you can't produce such proofs yourself unless you know how to build the merkle branch
1695 2012-02-05 20:58:55 <TD> gmaxwell: how so ?
1696 2012-02-05 20:58:56 <gmaxwell> TD: because I could just mine one (gonna be orphaned) block on top of the chain with bogus commitments and make an SPV+MTUO node believe it.
1697 2012-02-05 20:59:04 <gmaxwell> (s/an/all/)
1698 2012-02-05 20:59:18 <TD> yeah, you'd have to do what SPV clients already do - only trust blocks once they reach a certain depth
1699 2012-02-05 20:59:38 <TD> as it happens bitcoinj today trusts any tx that appears in a valid connecting block. but i'm trying to make this more nuanced with the TransactionConfidence API
1700 2012-02-05 20:59:44 <gmaxwell> Right, thats good— but does it actually make things faster for the SPV case?
1701 2012-02-05 20:59:48 <TD> so users can decide what level of trust they want
1702 2012-02-05 20:59:54 <TD> well
1703 2012-02-05 21:00:05 <TD> it means somebody can pass you a tx offline, with the proof, and you can have more confidence in it
1704 2012-02-05 21:00:15 <TD> so for the vending machine use case or whatever the machine could verify your payment itself
1705 2012-02-05 21:00:28 <TD> if there's a network outage, it can still vend crisps or diamonds or whatever it's selling
1706 2012-02-05 21:00:47 <TD> because unless you're spending very young outputs, it can check the TX is valid without having to talk to a full node
1707 2012-02-05 21:00:53 <TD> but i admit that's only a minor advantage
1708 2012-02-05 21:00:53 <justmoon> TD: can't I just create a conflicting transaction online in that case?
1709 2012-02-05 21:01:13 <TD> yes, if you have network access and it doesn't
1710 2012-02-05 21:01:32 <TD> that was a bad example
1711 2012-02-05 21:01:53 <TD> maybe it's better to say it's cheaper to run because you don't need a full validating node to do the work for the machine. it can be self contained and just hold open some connections to the regular p2p network
1712 2012-02-05 21:02:38 <gmaxwell> TD: Okay. Fair enough— it's just important to not ascribe too much to it. It's an improvement on SPV, but a narrow one.
1713 2012-02-05 21:02:42 <TD> yeah
1714 2012-02-05 21:02:50 <TD> i'd work on other stuff first, myself
1715 2012-02-05 21:02:54 occulta has joined
1716 2012-02-05 21:02:57 <TD> like connection filtering
1717 2012-02-05 21:03:23 <TD> i want to be able to suppress or reduce needless traffic and download blocks with only the relevant transactions included (ie, you get a header+relevant transactions+merkle branches)
1718 2012-02-05 21:04:22 <justmoon> gmaxwell: is it true that bitcoin skips/used to skip ecdsa verification even for the latest blocks during the initial download?
1719 2012-02-05 21:04:34 <justmoon> i.e. what this guy claims: http://marc.info/?l=full-disclosure&m=132810929830371&w=2
1720 2012-02-05 21:05:19 <gmaxwell> justmoon: What you're saying isn't correct, what he's saying was.
1721 2012-02-05 21:05:38 <justmoon> how was my version incorrect?
1722 2012-02-05 21:07:08 barmstrong has joined
1723 2012-02-05 21:07:24 <gmaxwell> justmoon: for 0.5 only (not 0.5.2) it would also skip past the highest checkpoint if and only if the second to last block recieved in the current best chain was >24 hours old, and the block after it was recieved within 10 seconds of getting that one.
1724 2012-02-05 21:08:10 <justmoon> which is the case during the initial download isn't it?
1725 2012-02-05 21:08:19 h4ckm3 has joined
1726 2012-02-05 21:09:04 <gmaxwell> justmoon: it could in theory also be the case with a >>50% power attacker (or not high power if you were totally isolated), mining blocks with incorrect timestamps and then feeding them to you in a wad.
1727 2012-02-05 21:09:26 <gmaxwell> In either case you'd reorg onto the correct network eventually, but he could have tricked you in the meantime.
1728 2012-02-05 21:10:32 <gmaxwell> Of course, he also could have just double spent on you. But the disabled signature validations mean that he could have use bitcoin that wasn't even his to trick you.
1729 2012-02-05 21:10:55 jeewee has quit ()
1730 2012-02-05 21:11:09 <justmoon> sure, I just looked at the code earlier today and it seemed like it did what he said, just wanted to check I didn't miss anything
1731 2012-02-05 21:11:09 <gmaxwell> justmoon: In git/0.5.2 it just disables up to the highest checkpoint.
1732 2012-02-05 21:11:25 <justmoon> nBestHeight < Checkpoints::GetTotalBlocksEstimate(), yes
1733 2012-02-05 21:11:33 marf_away has quit (Quit: Nettalk6 - www.ntalk.de)
1734 2012-02-05 21:12:20 danbri has joined
1735 2012-02-05 21:16:12 pingdrive has joined
1736 2012-02-05 21:21:53 pusle has quit ()
1737 2012-02-05 21:26:07 Katniss has quit (Read error: Connection reset by peer)
1738 2012-02-05 21:26:28 iocor has joined
1739 2012-02-05 21:28:33 <makomk> gmaxwell: the scary thing is mostly that if enough people had upgraded, they wouldn't reorg onto the correct network because the version with the fake spend would become the correct network.
1740 2012-02-05 21:31:10 <FROTUSCI> cool
1741 2012-02-05 21:31:25 <gmaxwell> makomk: briely at least, there are plenty of nodes with versions going way back out there that it would be noticed. I'm now running a .19 node and if it disagrees with the git nodes, I get yelled at.
1742 2012-02-05 21:31:31 <gmaxwell> I guess I should make that output to IRC too.
1743 2012-02-05 21:33:13 <gmaxwell> hm. I guess it could trigger as soon as a block fails vaildation, since it's behind -connect, in fact.
1744 2012-02-05 21:34:59 Moron__ has quit ()
1745 2012-02-05 21:38:28 <luke-jr> hmm, something funny going on?
1746 2012-02-05 21:39:23 <gmaxwell> luke-jr: what are you seeing?
1747 2012-02-05 21:39:47 <gmaxwell> Graet pointed me to two recent height 1 forks.
1748 2012-02-05 21:40:19 <luke-jr> gmaxwell: ljrbot seem to be stuck on that fork
1749 2012-02-05 21:40:24 <gmaxwell> oh actually, a single height 2 fork.
1750 2012-02-05 21:40:43 <gmaxwell> luke-jr: according to blockchain.info eligius side won, with an extension by btcguild.
1751 2012-02-05 21:40:51 <luke-jr> gmaxwell: yes, but ljrbot is stuck
1752 2012-02-05 21:41:04 danbri has quit (Remote host closed the connection)
1753 2012-02-05 21:41:52 <luke-jr> 0.3.21-based still
1754 2012-02-05 21:41:54 <gmaxwell> luke-jr: what do its logs say?
1755 2012-02-05 21:41:55 <luke-jr> ;;bc,blocks
1756 2012-02-05 21:42:00 <gribble> 165523
1757 2012-02-05 21:42:04 <gmaxwell> http://blockchain.info/block-index/859800/00000000000000192af6a40b747a6a8c3d5ac41956c94bba2d0be0dc90fb6b94
1758 2012-02-05 21:42:09 <luke-jr> odd, ljrbot has the same height
1759 2012-02-05 21:42:11 <gmaxwell> http://blockchain.info/block-index/859803/0000000000000a50f1de5eb6182778660266764e919b573c1ea8697fef0974cb
1760 2012-02-05 21:42:36 D_H_ has joined
1761 2012-02-05 21:42:48 <luke-jr> debug.log is huge
1762 2012-02-05 21:43:20 <gmaxwell> yes, fine look at the bottom.
1763 2012-02-05 21:43:38 <luke-jr> yeah, less is slow on big file ;P
1764 2012-02-05 21:45:01 <luke-jr> looks like a long line with a ton of  +numbers
1765 2012-02-05 21:45:53 <cdecker> If less is slow use tail
1766 2012-02-05 21:46:06 ovidiusoft has quit (Ping timeout: 265 seconds)
1767 2012-02-05 21:46:30 <luke-jr> Added time data, samples 21499, offset -1 (+0 minutes)
1768 2012-02-05 21:46:37 <luke-jr> followed by a ton
1769 2012-02-05 21:46:38 D_H_ has quit (Quit: Page closed)
1770 2012-02-05 21:47:31 <luke-jr> furthermore, what is 0000000000000572c437edf738aa3cc1e733bbfd4bdd51f3a6bb985fae130d9a
1771 2012-02-05 21:48:49 <luke-jr> restarting ljrbot's bitcoind…
1772 2012-02-05 21:49:03 <luke-jr> looks like debug.log overflowed
1773 2012-02-05 21:51:17 <luke-jr> SetBestChain: new best=00000000000000fa663e  height=165524  work=229173008572693874424
1774 2012-02-05 21:51:45 <luke-jr> blockchain.info says this is 165520
1775 2012-02-05 21:52:09 ovidiusoft has joined
1776 2012-02-05 21:56:01 <gmaxwell> gah, these log entries need to be changed to give the other @#$@# side
1777 2012-02-05 21:56:12 <gmaxwell> a bunch of @#$#@ zeros is not helpful
1778 2012-02-05 21:57:23 <gmaxwell> blockchain.info appears to just be busted.
1779 2012-02-05 21:58:15 * Graet is happy he's on winning side :D
1780 2012-02-05 21:58:26 pingdrive has quit (Quit: Leaving)
1781 2012-02-05 21:58:56 <luke-jr> hmm
1782 2012-02-05 21:59:00 <gmaxwell> luke-jr: is lrjbot still stuck?
1783 2012-02-05 21:59:05 <luke-jr> gmaxwell: not sure
1784 2012-02-05 21:59:28 <gmaxwell> (or did the reorg just confuse it?)
1785 2012-02-05 21:59:31 <luke-jr> I think not, but it skipped announcing it
1786 2012-02-05 21:59:41 b4epoche_ has joined
1787 2012-02-05 22:00:08 <gmaxwell> My .19 node wasn't stuck, but to be sure I droped its chain and I'm resyncing it to make sure it also reaches the same chain endpoint.
1788 2012-02-05 22:00:09 <luke-jr> Graet: we should be on the winning side more often! wanna addnode each other? :P
1789 2012-02-05 22:00:19 <luke-jr> Graet: does that mean your pool is back online btw?
1790 2012-02-05 22:00:28 b4epoche has quit (Ping timeout: 244 seconds)
1791 2012-02-05 22:00:28 b4epoche_ is now known as b4epoche
1792 2012-02-05 22:01:07 <Graet> no, but eu node is ok, so i was going to solo there, but a few guys have donated some charity hashes, no stats - but any blocks go toward helping with losses :D
1793 2012-02-05 22:01:20 <luke-jr> ah
1794 2012-02-05 22:01:26 <Graet> i'm just generating o my wallet for the mo
1795 2012-02-05 22:01:46 <Graet> so it was surprise to find one, and relief not to be orphan :)
1796 2012-02-05 22:02:02 <luke-jr> gmaxwell: I successfully gitian-built boost in a VM btw
1797 2012-02-05 22:02:06 <luke-jr> ie, nested VMs
1798 2012-02-05 22:02:53 <gmaxwell> so.. whew. false alarm.
1799 2012-02-05 22:03:12 <luke-jr> gmaxwell: dunno, it didn't recover until I restarted it, so maybe something unrelated going on?
1800 2012-02-05 22:03:24 <gmaxwell> luke-jr: the bot or the node?
1801 2012-02-05 22:03:42 <Graet> dont know if it related i had to restart my wallet for it to show too
1802 2012-02-05 22:04:04 <luke-jr> I wonder if I hadn't, if it would have tracked the orphan chain :x
1803 2012-02-05 22:04:08 <gmaxwell> Graet: the wallet doesn't show blocks until they are burried by one.
1804 2012-02-05 22:04:13 splatster has quit (Quit: Linkinus - http://linkinus.com)
1805 2012-02-05 22:04:16 <luke-jr> we never actually tested what happened with that exploit
1806 2012-02-05 22:04:38 <luke-jr> gmaxwell: oh? I'm pretty sure my bitcoinds did O.o
1807 2012-02-05 22:04:54 <luke-jr> at confirm=1, which is the start
1808 2012-02-05 22:04:56 <gmaxwell> bitcoinds do.
1809 2012-02-05 22:05:00 <gmaxwell> bitcoin-qt doesn't.
1810 2012-02-05 22:05:10 <luke-jr> ah
1811 2012-02-05 22:05:23 <luke-jr> Graet: which one did you use in this case?
1812 2012-02-05 22:05:39 <gmaxwell> So, my .19 node wasn't stuck, but I don't know if it saw that fork at all— and stupidly I deleted the logs when I deleted the chain to make sure it would resync.
1813 2012-02-05 22:05:45 paul0 has quit (Quit: paul0)
1814 2012-02-05 22:05:53 <gmaxwell> lemme hit up the network to see.
1815 2012-02-05 22:07:32 <gmaxwell> I don't see any evidence of stuck nodes.
1816 2012-02-05 22:08:04 <Graet> i'm using bitcoin-qt 0.5.2-beta nad there were 4 blocks on top of ours before i restarted it
1817 2012-02-05 22:10:17 PK has quit ()
1818 2012-02-05 22:10:48 MrTiggr has quit (Ping timeout: 244 seconds)
1819 2012-02-05 22:11:19 <gmaxwell> blockchain.info shows the generation from the losing side of the fork in the balance of the related addresses.
1820 2012-02-05 22:11:59 <TD> blockchain.info doesn't handle re-orgs?
1821 2012-02-05 22:12:48 bitlad has quit (Remote host closed the connection)
1822 2012-02-05 22:13:24 <gmaxwell> Guess not, http://blockchain.info/address/416bd52d4c17be883253c0acf8cc87671972ca6d
1823 2012-02-05 22:13:34 <gmaxwell> thats generation output from a losing fork.
1824 2012-02-05 22:14:11 <TD> perhaps i should try doing a block explorer type app based on bitcoinj
1825 2012-02-05 22:15:07 <luke-jr> the problem is, sites like BE and BChain.info shouldn't *care* which chain prevails
1826 2012-02-05 22:15:25 <luke-jr> they should just display the data, and maintain a separate index
1827 2012-02-05 22:17:11 <lianj> http://bitcoin.interesthings.de/tx/843206a626bdec9cd4ef171af2827e4ff9865697e38043c8aabd4658c5242b63.json ha, didn't now value 0.00000000 was allowed
1828 2012-02-05 22:17:11 dwon has joined
1829 2012-02-05 22:17:18 <lianj> *know
1830 2012-02-05 22:19:19  has quit (Clown|!Clown@static-87-79-93-140.netcologne.de|Ping timeout: 240 seconds)
1831 2012-02-05 22:19:32 <luke-jr> lianj: Eligius charges extra for those :P
1832 2012-02-05 22:19:40 <riush> http://blockchain.info/orphaned-blocks
1833 2012-02-05 22:20:06 <Graet> :D
1834 2012-02-05 22:21:13 <luke-jr> a lot of orphaning going on
1835 2012-02-05 22:21:24 <luke-jr> 95.120.241.167 looks hostile
1836 2012-02-05 22:22:13 <luke-jr> LOL @ two "DeepBit" conflicting
1837 2012-02-05 22:22:45 pingdrive has joined
1838 2012-02-05 22:23:58 imsaguy is now known as n00bie
1839 2012-02-05 22:24:04 n00bie is now known as imsaguy
1840 2012-02-05 22:26:31 <gmaxwell> 02/02/12 12:29:35 version message: version -1125599027, blocks=-1
1841 2012-02-05 22:26:58 <luke-jr> O.o
1842 2012-02-05 22:28:53 pingdrive has quit (Quit: Leaving)
1843 2012-02-05 22:33:06 justmoon has quit (Quit: Leaving)
1844 2012-02-05 22:34:24 <gmaxwell> I made a simply python script that uses version messages from the logs to tell you about lagging nodes:
1845 2012-02-05 22:34:28 <gmaxwell> http://pastebin.com/EympQwyG
1846 2012-02-05 22:34:52 <gmaxwell> of course, it's noisy for versions with few active clients.
1847 2012-02-05 22:35:11 erle- has quit (Quit: erle-)
1848 2012-02-05 22:36:12 ThomasV has quit (Ping timeout: 255 seconds)
1849 2012-02-05 22:36:37 <gmaxwell> Here is what I get from it: http://pastebin.com/sb7A3jNT
1850 2012-02-05 22:36:41 <gmaxwell> looks mostly fine.
1851 2012-02-05 22:48:11 <gmaxwell> here are the transactions the loser vs ozcoin+eligius had, that ozcoin did not:
1852 2012-02-05 22:48:15 <gmaxwell> da9f68c1fb266bdddc529f21f250a3e3fbdf902af38123340f71045cb1483b08
1853 2012-02-05 22:48:23 <gmaxwell> ae3d4c2a619f007870bdd6af927083b7f38cfec9f78dca8eb57bc2c8eb4196a9
1854 2012-02-05 22:48:36 <gmaxwell> a98b734582b8ac0bfe63ba6455c117f316108ac037ddaa9abea4ddf78c5a6e65
1855 2012-02-05 22:48:48 <gmaxwell> 5ca7526ab2bae32df4c628ca2b4c29344864ea2ed75b4af59128fb69222dc203
1856 2012-02-05 22:49:19 <gmaxwell> ozcoin did not have any transactions the loser lacked (except the coinbases in each case, of course)
1857 2012-02-05 22:50:37 <gmaxwell> All of those orphan txn were eventually mined in 165521.
1858 2012-02-05 22:51:22 <gmaxwell> so nothing weird.
1859 2012-02-05 22:53:08 <gmaxwell> and my 3.19 node synced back up just fine.
1860 2012-02-05 22:54:35 <luke-jr> hmm
1861 2012-02-05 22:55:29 coblee_ has joined
1862 2012-02-05 22:56:23 <gmaxwell> Perhaps we can get the person operating blockchain.info to automatically report transaction differences on the orphans page?
1863 2012-02-05 22:56:37 <gmaxwell> That would be really good information to have while investigating a potential problem.
1864 2012-02-05 22:56:57 coblee has quit (Read error: Operation timed out)
1865 2012-02-05 22:56:57 coblee_ is now known as coblee
1866 2012-02-05 22:57:00 <TD> how does the satoshi client handle the case where a peer is sending the block chain and then just stops
1867 2012-02-05 22:57:06 <TD> i forgot .... been a while since i read that code
1868 2012-02-05 22:57:10 <TD> if anyone knows offhand?
1869 2012-02-05 22:57:16 <TD> does it notice and switch to a different peer?
1870 2012-02-05 22:57:24 <gmaxwell> No idea.
1871 2012-02-05 22:57:42 <tcatm> I think it'll "wait" until it receives a new block and then ask that peer for more.
1872 2012-02-05 22:58:02 <TD> yeah
1873 2012-02-05 22:58:04 <TD> so it just stops
1874 2012-02-05 22:58:18 <TD> i've noticed sometimes peers just stop delivering blocks during chain download, but don't actually disconnect
1875 2012-02-05 22:58:29 <gmaxwell> doh.
1876 2012-02-05 22:58:40 <gmaxwell> Do they actually have them?
1877 2012-02-05 22:59:06 <gmaxwell> Iirc it pulls first from the first connected peer... perhaps not the best metric either.
1878 2012-02-05 22:59:28 <TD> yeah that's what bitcoinj does
1879 2012-02-05 22:59:42 <TD> and i sometimes see instability in chain download for that reason. annoying.
1880 2012-02-05 22:59:43 <tcatm> It would be much better to download from multiple peers.
1881 2012-02-05 22:59:45 <TD> yeah the peers claim to have them
1882 2012-02-05 22:59:59 <TD> well, switching between peers if one doesn't deliver fast enough is possible
1883 2012-02-05 23:00:13 <gmaxwell> tcatm: we really need backwards fetching to make that a good thing.
1884 2012-02-05 23:00:43 <FROTUSCI> publish it via cdn and have the peers confirm the hashes
1885 2012-02-05 23:00:54 <TD> 184.154.125.210:8333
1886 2012-02-05 23:00:57 <gmaxwell> FROTUSCI: that doesn't address any issue that we actually have.
1887 2012-02-05 23:01:04 <FROTUSCI> yes it does
1888 2012-02-05 23:01:11 <TD> this node claimed to have all the blocks but just randomly stopped delivering. i wonder if we're hitting a DoS check
1889 2012-02-05 23:01:21 <TD> i should do some tests against a local node again to see if i can reproduce it
1890 2012-02-05 23:01:22 <FROTUSCI> cdn is faster and more reliable than requesting individual blocks
1891 2012-02-05 23:01:26 <tcatm> Which version does that node run?
1892 2012-02-05 23:01:39 <TD> 11:47:05 11 TCPNetworkConnection.<init>: Connected to peer: version=32400, subVer='', services=0x1, time=Sun Feb 05 23:46:46 CET 2012, blocks=165530
1893 2012-02-05 23:01:51 <josephcp> yeah might be the usual older nodes anti DDoS stuff getting hit?
1894 2012-02-05 23:01:59 <TD> ah, is that the version with the DoS bug?
1895 2012-02-05 23:02:06 <TD> i forgot which ones were affected. that could be it.
1896 2012-02-05 23:02:06 <gmaxwell> No. 24 should be okay.
1897 2012-02-05 23:02:15 <gmaxwell> Also, that behavior makes it hang up.
1898 2012-02-05 23:02:22 <TD> thats true
1899 2012-02-05 23:02:35 <tcatm> 0.3.24 is the version fixing the DoS bug
1900 2012-02-05 23:02:36 <gmaxwell> TD: what height does it stop at.
1901 2012-02-05 23:02:39 <FROTUSCI> encode it with an information dispersal algo, say a tornado code and its even more robust in the face of server disconnects
1902 2012-02-05 23:02:54 <gmaxwell> FROTUSCI: please be quiet until you become competent.
1903 2012-02-05 23:03:02 <TD> gmaxwell: i need to retry and see if it starts delivering again. just a sec.
1904 2012-02-05 23:03:22 <TD> it stopped at height 109994, but i didn't start the download from genesis
1905 2012-02-05 23:03:29 <FROTUSCI> gmaxwell: if you have nothing of substance to say don't speak at all
1906 2012-02-05 23:03:45 <TD> FROTUSCI: the CDN idea makes sense for some clients, i think, but the obvious problem we have ATM is "who pays for the CDN"
1907 2012-02-05 23:03:53 <TD> FROTUSCI: finding reliable, fast peers is just a matter of coding
1908 2012-02-05 23:03:55 <cjd> /topic 18:11 < FROTUSCI> gmaxwell: if you have nothing of substance to say don't speak at all
1909 2012-02-05 23:03:57 <TD> once it's done, it's done
1910 2012-02-05 23:04:04 <gmaxwell> FROTUSCI: This is a channel for bitcoin development, not for spouting random buzzwords.
1911 2012-02-05 23:04:11 booo has quit (Read error: Operation timed out)
1912 2012-02-05 23:04:21 cryptoxchange has joined
1913 2012-02-05 23:04:32 <FROTUSCI> true TD. good thing is that there are lots of ppl who are happy to provide free cdn service (in the samw wa linux distros are mirrored)
1914 2012-02-05 23:05:15 <TD> gmaxwell: yeah reconnecting to it makes it work again
1915 2012-02-05 23:05:28 <TD> i'd love to see its logs
1916 2012-02-05 23:05:28 <FROTUSCI> gmaxwell: you are very hostile
1917 2012-02-05 23:05:31 <TD> wish the protocol was more verbose
1918 2012-02-05 23:05:33 <gmaxwell> td: bleh. I can't think of why it would just stop.
1919 2012-02-05 23:05:47 <TD> gmaxwell: maybe it's a bug in my code. a race or something.
1920 2012-02-05 23:05:51 <FROTUSCI> http://en.wikipedia.org/wiki/Tornado_code  is an ideal way to distribute the blockchain over multiple cdn providers
1921 2012-02-05 23:06:03 <TD> it's possible i guess ..... it usually stops on a boundary (when doing a new getblocks/getheaders)
1922 2012-02-05 23:06:22 <gmaxwell> FROTUSCI: Yes, I'm familar with the subject. It is also agressively irrelevant to bitcoin today.
1923 2012-02-05 23:06:44 <FROTUSCI> except its a painful usability issue for new users
1924 2012-02-05 23:06:59 <gmaxwell> FROTUSCI: which is _completely_ unrelated to that.
1925 2012-02-05 23:07:08 eldentyrell has joined
1926 2012-02-05 23:07:20 <gmaxwell> FROTUSCI: the time it takes to sync the blockchain is pretty much entirely due to validation, not transfer.
1927 2012-02-05 23:07:28 <luke-jr> FROTUSCI: initial blockchain load is not merely downloading
1928 2012-02-05 23:07:28 <FROTUSCI> usability is irrelevant to "bitcoin today" ? cool story
1929 2012-02-05 23:07:40 <k9quaint> PATCHES ARE WELCOME!
1930 2012-02-05 23:07:42 * k9quaint flees
1931 2012-02-05 23:07:42 <luke-jr> FROTUSCI: it has to actually *process* the daa
1932 2012-02-05 23:07:47 <tcatm> Well, using such codes to spread transactions (for archival and later use) could be an idea.
1933 2012-02-05 23:08:00 <gmaxwell> FROTUSCI: my message was in response to your tornado link.
1934 2012-02-05 23:08:31 <FROTUSCI> validation would be reduced to verifying the hashes of long segments your peers give you
1935 2012-02-05 23:08:35 <cjd> ahh the good old distracting-developers-with-almost-plausible-jibberish attack, one of few vulnerabilities which bitcoin has yet to patch :)
1936 2012-02-05 23:08:46 <Graet> ^^
1937 2012-02-05 23:08:51 <gmaxwell> FROTUSCI: No. Validation in bitcoin means something more significant than simply validating hashes from peers.
1938 2012-02-05 23:09:11 <TD> FROTUSCI: you could distribute the databases as well via a CDN
1939 2012-02-05 23:09:17 <FROTUSCI> aye
1940 2012-02-05 23:09:24 <TD> and in fact i agree, i think we should be distributing finalized databases with the client download
1941 2012-02-05 23:09:32 <gmaxwell> FROTUSCI: Bitcoin doesn't trust peers to have followed the rules of the system, so each node validates all the bitcoin rules itself.
1942 2012-02-05 23:09:34 <k9quaint> cjd: I submitted a patch for that the other day, there seems to be some loopholes in my code tho :(
1943 2012-02-05 23:09:35 <TD> at least until/unless users are all on lightweight clients
1944 2012-02-05 23:09:46 <TD> given that you already trust the client download
1945 2012-02-05 23:09:52 <TD> it has no impact beyond making bitcoin start faster
1946 2012-02-05 23:09:58 <TD> and making the download a lot larger of course
1947 2012-02-05 23:10:11 <gmaxwell> TD: you can audit the software that you download. Opaque data is opaque.
1948 2012-02-05 23:10:11 <luke-jr> TD: well, I'd have the installer download it optionally
1949 2012-02-05 23:10:32 <TD> gmaxwell: you can audit the included data just by running it with -rescan -checkblocks
1950 2012-02-05 23:10:37 <k9quaint> the entire point of bitcoin is the network enforces the integrity of the data
1951 2012-02-05 23:10:51 <k9quaint> downloading all the data from a single source is nioctib
1952 2012-02-05 23:10:58 <tcatm> luke-jr: Or let the client fetch it itself after asking the user...
1953 2012-02-05 23:11:12 <TD> k9quaint: it's equivalent. you _already trust the client_ explicitly. if you haven't read the code and compiled it yourself, that program could do anything
1954 2012-02-05 23:11:19 <FROTUSCI> you dont have to do all that validation on the entire blockchain to start with since the proof-of-work is good enough assurance. in any case, you can validate signatures in the background
1955 2012-02-05 23:11:27 <tcatm> It could even do this by connecting to a trusted node first and downloading the chain from there.
1956 2012-02-05 23:11:29 <gmaxwell> TD: yes, if someone wrote that functionality then that would be good. It's not acceptable to only run it on demand.. but e.g. checking some random segments would be tine.
1957 2012-02-05 23:11:30 <TD> there's no pointing having untrusted software and then saying, i downloaded from the network and now it's trusted
1958 2012-02-05 23:11:43 <k9quaint> TD: so, because I trust chrome, I trust everything chrome can access?
1959 2012-02-05 23:11:47 <gmaxwell> FROTUSCI: It's only assurance _at all_ because other nodes are doing full validation.
1960 2012-02-05 23:12:24 <TD> k9quaint: they're totally different
1961 2012-02-05 23:12:44 <k9quaint> TD: garbage in, garbage out, client irrelevant
1962 2012-02-05 23:12:58 <TD> k9quaint: if the data comes with the code, then anyone who can modify the data can also modify the code
1963 2012-02-05 23:12:59 <TD> right?
1964 2012-02-05 23:13:04 <FROTUSCI> checking the hashes gives you assurance that the chain was difficult to create. you dont really need to check individual transactions until you rely on them for some payment
1965 2012-02-05 23:13:05 cdecker has quit (Ping timeout: 252 seconds)
1966 2012-02-05 23:13:17 <gmaxwell> FROTUSCI: That is not the bitcoin security model.
1967 2012-02-05 23:13:45 <k9quaint> TD: that is a great method for doing a 51% attack without needing the hashpower
1968 2012-02-05 23:14:08 <TD> no, it's not. i don't see why this is so hard to understand. the client already hardcodes checkpoints into it
1969 2012-02-05 23:14:21 <TD> when you trust the binary you download, you *by definition* trust any data it comes with
1970 2012-02-05 23:14:26 <gmaxwell> TD: Because the validation behaivor is transparent and auditable.
1971 2012-02-05 23:14:34 <TD> so is the included block chain
1972 2012-02-05 23:14:39 <gmaxwell> TD: I can't change the software without taking a non-trivial risk the people will see the change.
1973 2012-02-05 23:14:48 [Tycho] has joined
1974 2012-02-05 23:14:53 <gmaxwell> TD: only if the client actually audits it.
1975 2012-02-05 23:14:58 <gmaxwell> TD: if it audits it, then great.
1976 2012-02-05 23:15:23 <TD> you can easily change the software if you have write access to the server. how many people do you think check the GPG signatures?
1977 2012-02-05 23:15:24 <FROTUSCI> gmaxwell: its no less secure to delay the check until you need it or when its a new tx
1978 2012-02-05 23:15:28 <cjd> is that weird coinbase still kicking around?
1979 2012-02-05 23:15:37 <gmaxwell> FROTUSCI: Yes, it is. Sorry.
1980 2012-02-05 23:15:38 <TD> only a small number. same as the number who would re-audit the included chains
1981 2012-02-05 23:16:05 <gmaxwell> FROTUSCI: I don't have time to debate with every idiot who has a grand idea. Write some patches, in the process you'll realize why you're incorrect— and if not I'll help you out then.
1982 2012-02-05 23:16:09 <TD> you could make miners do it automatically if you wanted to. but as anyone who can modify the included data can modify the binaries, it doesn't buy you much.
1983 2012-02-05 23:16:10 <FROTUSCI> gmaxwell: no it plainly isn't. if you never need to know if a signature is valid then checking it is a waste of time
1984 2012-02-05 23:16:38 <FROTUSCI> CDN is a novel idea? lol
1985 2012-02-05 23:16:46 <[Tycho]> cjd: what coinbase ?
1986 2012-02-05 23:17:06 <gmaxwell> TD: this is what gittian solves pretty much completely.
1987 2012-02-05 23:17:09 <FROTUSCI> a perfectly obvious extension to bitcoin is not some major upheaval
1988 2012-02-05 23:17:23 <cjd> there was a strange forkey thing that just happened, gmax understands it better than me
1989 2012-02-05 23:17:29 <FROTUSCI> but adhoms and handwaving dont make a convincing counterargument
1990 2012-02-05 23:17:38 <gmaxwell> cjd: Nah, you have wires crossed.
1991 2012-02-05 23:17:44 <cjd> :D
1992 2012-02-05 23:17:46 <k9quaint> FROTUSCI: #clientservercoin is where you should be at :)
1993 2012-02-05 23:17:55 <gmaxwell> cjd: nothing weird happened, there was concern that something weird happened. But it seems not.
1994 2012-02-05 23:18:05 <cjd> ok
1995 2012-02-05 23:18:23 <cjd> because I was half paying attention and it looked like what rocnner described yesterday
1996 2012-02-05 23:18:25 <FROTUSCI> if you never rely on a previous tx for your own payments, why would you need to verify it immediately?
1997 2012-02-05 23:18:30 <FROTUSCI> that can obviously be pushed to the background
1998 2012-02-05 23:18:40 <gmaxwell> TD: better to just make it randomly audit 1:N at least at first startup which could be arbitrarily fast and would retain good security properties.
1999 2012-02-05 23:18:54 barmstrong has quit (Remote host closed the connection)
2000 2012-02-05 23:19:30 <TD> probably you could have a hash of the included files in the binary as well, if you like
2001 2012-02-05 23:19:43 <FROTUSCI> yeah TD that would be pretty simple to do
2002 2012-02-05 23:19:44 <gmaxwell> TD: Thats fine but not helpful. Again. Not auditable.
2003 2012-02-05 23:19:52 <k9quaint> the client is not where the security of bitcoin resides
2004 2012-02-05 23:20:18 <gmaxwell> TD: we already have that in the checkpoints effectively, except we don't have a rescan which checks the hashes (but we really should due to disk/memory corruption in any case)
2005 2012-02-05 23:20:54 <TD> i don't see why you are so hung up on checking the data. gitian doesn't "solve" anything for 99% of users, it just lets people who are interested ensure the binaries match the source. anyone who is willing to go through  that checking process can also just force a rescan/recheck of the included data files and ensure it passes. make it a part of the build, even
2006 2012-02-05 23:21:45 <k9quaint> TD: peers checking the data is the foundation upon which the security of bitcoin is built
2007 2012-02-05 23:21:54 <TD> and yeah, i guess what i'm saying is do checkpoints but not just for the chain head
2008 2012-02-05 23:21:55 <gmaxwell> TD: I don't see why you are so eager to abandon the zero trust properties of the bitcoin full node for zero gain.
2009 2012-02-05 23:21:58 <TD> but for included databases as well
2010 2012-02-05 23:22:04 <TD> i don't see it as affecting the trust properties of the system at all
2011 2012-02-05 23:22:13 <[Tycho]> Hmm, that reorg was a bit surprising :)
2012 2012-02-05 23:22:22 <TD> the vast majority of all users don't verify the binary does what we claim it does
2013 2012-02-05 23:22:25 <TD> they take it on faith
2014 2012-02-05 23:22:45 <gmaxwell> TD: yes, and thats a problem but _at least_ if we were lying it would be pretty straightforwardly discoverable.
2015 2012-02-05 23:22:47 <TD> that's why gitian is a good investment. otherwise whoever compiles the binary essentially controls all bitcoin users who run it
2016 2012-02-05 23:22:58 <TD> well, it's discoverable now by people who take the time to check it
2017 2012-02-05 23:23:17 <TD> now if you extend the download to include a chain and database, then the subterfuge is still discoverable .... by people who take the time to check it
2018 2012-02-05 23:23:19 <gmaxwell> TD: thats not good enough. We know for a fact that only a infinitesmal number of users check.
2019 2012-02-05 23:23:23 <TD> it's the same
2020 2012-02-05 23:23:26 <TD> sure. but we can't avoid that.
2021 2012-02-05 23:23:29 <TD> it only takes one or two anyway
2022 2012-02-05 23:23:48 <TD> once somebody has a binary that claims to be a deterministic output of the source but actually contains a backdoor, it will be in the press
2023 2012-02-05 23:23:58 <TD> people will read about it. spread the word. bitcoins credibility is destroyed.
2024 2012-02-05 23:24:11 <TD> same for a bogus database
2025 2012-02-05 23:24:15 <gmaxwell> It really isn't the same. I read all the diffs committed (for whatever thats worth) and match that the source matches. I have no way of telling that some 800 mb opaque binary crap is correct, and I'm never going to check it.
2026 2012-02-05 23:24:28 <TD> sure you do
2027 2012-02-05 23:24:33 <TD> just download the chain from scratch and then compare
2028 2012-02-05 23:24:49 <gmaxwell> I'm not going to download and check an 800 mb file with every release. No one is.
2029 2012-02-05 23:25:04 <gmaxwell> I do download the release tar files and check the source.
2030 2012-02-05 23:25:24 [Tycho] has quit (Remote host closed the connection)
2031 2012-02-05 23:25:28 <TD> i question that assumption. checking the chain is slow/expensive ..... for machines. whose time is cheap.
2032 2012-02-05 23:25:28 <gmaxwell> If the code still validates the blob then that checking carries through.
2033 2012-02-05 23:25:35 <tcatm> Looks like a tool to verify a blk0001.dat could be useful ;)
2034 2012-02-05 23:25:35 <TD> checking the source is slow and expensive for skilled professionals
2035 2012-02-05 23:25:53 <TD> rebuilding the chain/db "from the source" is something a shell script can do
2036 2012-02-05 23:26:28 <gmaxwell> TD: this is a stupid discussion. Do you agree that just having every system check 1:N automatically would be trivial and have basically no performance impact (for sutiable N)?
2037 2012-02-05 23:27:02 <da2ce7> good morning all :)
2038 2012-02-05 23:27:23 <FROTUSCI> there is no harm in distributing the chain with the binary. we already have checkpoints and that is equivalent to it
2039 2012-02-05 23:28:16 <TD> it's not a stupid discussion and I agree with FROTUSCI, there's no need to get vitriolic. i don't see any reason to have a complicated implementation - how do you pick N and how do you check the random transactions? you can't know if the spent flags are correct without checking all the transactions
2040 2012-02-05 23:28:54 <TD> if you worry about the distribution servers getting hacked or whatever, just have a cron job that downloads the user-facing binary
2041 2012-02-05 23:28:59 <TD> if the hash has changed, rerun the automated checks
2042 2012-02-05 23:29:05 <TD> it's no different to gitian in that respect.
2043 2012-02-05 23:29:06 <gmaxwell> TD: Sure you can. You pick N to give good performance but still making any modification trivially to detect. one in 1000 blocks would be fine.
2044 2012-02-05 23:29:58 <gmaxwell> TD: And you don't need to check the spent flags, you just check that the inputs exist. And assume if they were tampered with then that would be detected in their respective blocks (if not by you then by someone else like you).
2045 2012-02-05 23:30:19 <TD> alright. so i put in a few blocks that double spend the same input, multiplying my wealth. the chances of you picking all three blocks i chose out of a few hundred thousand isn't that high. but if you don't, you can't tell i am double spending
2046 2012-02-05 23:30:56 splatster has joined
2047 2012-02-05 23:31:33 <FROTUSCI> for the truly paranoid, you can have a /nocheckpoints flag to disable any pre-defined blocks
2048 2012-02-05 23:32:12 <gmaxwell> TD: the odds of some systems detecting it with, e.g. 40k users are _excellent_
2049 2012-02-05 23:32:49 <TD> no, the odds are tiny because when it's detected, what are you going to do? pop up an alert? so the user sees "WARNING: <some technobabble>" and then presses the "yeah whatever" button and ignores it. after all, the system still works, doesn't it.
2050 2012-02-05 23:32:50 <gmaxwell> It isn't acceptable to just have afew users manally check.
2051 2012-02-05 23:33:13 <TD> if you have a small number of technically skilled people, who understand what they're doing, checking the downloads, the chances of a successful detection and escalation are very good
2052 2012-02-05 23:33:16 <gmaxwell> TD: No. The software just refuses to work at that point... and throws up a message.
2053 2012-02-05 23:33:33 <TD> so then the user thinks "ugh, bitcoin is broken" and either gives up or switches to another client
2054 2012-02-05 23:33:48 <FROTUSCI> then you can ddos the network that way, killing every bitcoin node
2055 2012-02-05 23:33:51 <gmaxwell> Maybe. But thats a pretty significant risk.
2056 2012-02-05 23:34:01 <gmaxwell> (for the attacker)
2057 2012-02-05 23:34:49 <gmaxwell> TD: Why are you so eager to remove validation of the network rules from bitcoin?
2058 2012-02-05 23:35:01 <TD> a jury of randomly selected users will never be as reliable as some paranoid programmers with a cron job and a knowledge of who can make noise about a hack
2059 2012-02-05 23:35:14 <TD> because the current client is crap and puts users off with its 24 hour+ startup time
2060 2012-02-05 23:35:32 <gmaxwell> TD: Thats orthorgonal.
2061 2012-02-05 23:35:47 RazielZ has quit (Quit: Leaving)
2062 2012-02-05 23:35:48 <gmaxwell> TD: it has a 24 hour startup time due to implementation stupidity.
2063 2012-02-05 23:36:01 <gmaxwell> TD: throw it on tmpfs, you can sync up from zero in 30 minutes.
2064 2012-02-05 23:36:20 <gmaxwell> (want debug.logs demonstrating that?)
2065 2012-02-05 23:36:39 <TD> so if you throw a giant ramdisk at it, it only takes 30 minutes? that's great. i can't think of any programs i used lately that weren't ready within 30 seconds of me starting them
2066 2012-02-05 23:36:44 minimoose has joined
2067 2012-02-05 23:37:23 <gmaxwell> TD: it starts un under 5 seconds. Thats the time for the initial syncup and downloading the data.
2068 2012-02-05 23:37:26 <FROTUSCI> yeah not really scalable.. once the block is multi-GB in size...
2069 2012-02-05 23:37:36 <TD> yeah but first run time is the most important for leaving a good impression
2070 2012-02-05 23:37:43 <TD> i wouldn't build an app that downloads its own source code, a compiler and then sits and churns for an hour before it's usable
2071 2012-02-05 23:37:46 <gmaxwell> TD: I'm sad that you actually don't share any of the value of the bitcoin system.
2072 2012-02-05 23:38:00 <gmaxwell> TD: why aren't you instead recommending people use webclients and electrum then?
2073 2012-02-05 23:38:09 <FROTUSCI> the chain will be 2GB by the end of the year or more
2074 2012-02-05 23:38:37 <FROTUSCI> you can really assume everything can just sit in ram
2075 2012-02-05 23:38:39 <gmaxwell> TD: those things also don't provide the full security model of bitcoin, but their existance doesn't remove the security provided by all the full nodes that do exist.
2076 2012-02-05 23:38:44 <FROTUSCI> *can't
2077 2012-02-05 23:38:52 <luke-jr> I know I'd hate to download a blockchain snapshot every release, when I already have it
2078 2012-02-05 23:38:53 <TD> you haven't shown why distributing a database removes any security.
2079 2012-02-05 23:38:59 <TD> the attacks are exactly the same
2080 2012-02-05 23:39:36 <FROTUSCI> you wouldnt have to luke-jr, just pick the package without it
2081 2012-02-05 23:39:39 <gmaxwell> TD: distributing the database is fine, it's not something I've expressed any opposition it. Failing to validate it is what I opposed.
2082 2012-02-05 23:39:43 <TD> luke-jr: the app can just download the chain databases as a separate file if there's no chain already present. it doesn't physically have to be in the same download, as long as the app verifies the hash of the download.
2083 2012-02-05 23:39:48 <FROTUSCI> someone will eventually end up distributing the chain that way i msure
2084 2012-02-05 23:40:14 <cjd> suppose you stored the headers and a bitfield w/ each one, each bit in the bitfield representing a tx output boolean (valid && unspent), that should download pretty quickly and then when someone wants to spend some money, shallange them to show you the block.
2085 2012-02-05 23:40:29 <tcatm> FROTUSCI: already done. It's just not used widely...
2086 2012-02-05 23:40:40 <gmaxwell> TD: and I've argued why _not validating the bitcoin rules_ is a major negative change. I'm sorry that I haven't found a clear way of expressing that arguement to you.
2087 2012-02-05 23:40:50 <cjd> can store a hash of that structure along with the lockin
2088 2012-02-05 23:40:59 <FROTUSCI> theres a difference between validating up front and lazy validation when you actually need it
2089 2012-02-05 23:41:04 <gmaxwell> tcatm: it's also completely insecure as it stands, because there is no way to validate those files.
2090 2012-02-05 23:41:10 <TD> you've expressed your argument, i just think it makes no sense. it could be used to justify distributing bitcoin as gcc+source+app which compiles during install time
2091 2012-02-05 23:41:18 <tcatm> gmaxwell: they are GPG signed
2092 2012-02-05 23:41:37 <TD> there's no point making lots of computers redundantly recalculate the same thing over and over again
2093 2012-02-05 23:41:44 <gmaxwell> tcatm: anyone ever download the signatures? :)  But, regardless, if I compromise you.. it can't be told.
2094 2012-02-05 23:42:17 <luke-jr> gmaxwell: Bitcoin-Qt could bundle the sigs it accepts
2095 2012-02-05 23:42:18 <tcatm> I don't keep logs so I can't tell.
2096 2012-02-05 23:42:31 <FROTUSCI> cjd: makes sense. it would be a big change to the internals, but equivalent security with less cpu load
2097 2012-02-05 23:42:35 <gmaxwell> TD: It could be used to argue that.
2098 2012-02-05 23:42:41 <gmaxwell> TD: So?
2099 2012-02-05 23:42:46 <TD> nobody _does_ argue that though, because it would be absurd.
2100 2012-02-05 23:43:01 <luke-jr> it's not absurd
2101 2012-02-05 23:43:01 <TD> having all our users compile their own code wouldn't achieve anything unless they all read it and understand it, which obviously they never will
2102 2012-02-05 23:43:15 <gmaxwell> TD: I mean thats what downloading via gittan-installer will basically do. But sure- there are degrees and costs.
2103 2012-02-05 23:43:25 <Diablo-D3> I dont understand the bitcoin code... I mean, boost? really?!
2104 2012-02-05 23:43:57 <FROTUSCI> indeed TD. even freebsd's vpn code had some subtle bugs introduced deliberately that went undiscovered for years
2105 2012-02-05 23:43:59 <gmaxwell> TD: And I nodded my hat to the costs by pointing out that randomized auditing is almost as good, in agregate, for our trust of the system.
2106 2012-02-05 23:44:06 <TD> so as our users are already essentially trusting 3rd party auditors to ensure that the binary they download really comes from the source it claims to, it changes nothing at all to have them trust that the database it comes with matches the chain it claims to come from
2107 2012-02-05 23:44:13 <TD> both can be checked deterministically using tools
2108 2012-02-05 23:44:34 <TD> i wouldn't trust any system that relies on random, busy, non-expert users to raise the alarm
2109 2012-02-05 23:44:43 <gmaxwell> TD: one doesn't exclude the other.
2110 2012-02-05 23:44:45 <TD> i've seen how many users report serious bugs when they occur
2111 2012-02-05 23:44:51 <TD> (maybe 1 in 1000)
2112 2012-02-05 23:44:56 <gmaxwell> TD: you're advocating a reduction in security for _zero_ benefit.
2113 2012-02-05 23:45:13 <gmaxwell> 1:1000 is pretty good, I've seen rates worse than that for sure. :)
2114 2012-02-05 23:45:15 <TD> the benefit is, download gets a bit larger, then the first time a new user who has read about bitcoin starts it, it's ready to go in 5 seconds
2115 2012-02-05 23:45:21 <FROTUSCI> oh it was openbsd's ipsec that was tampered with
2116 2012-02-05 23:45:35 <gmaxwell> TD: Thats not a benefit.
2117 2012-02-05 23:45:43 <gmaxwell> TD: You can have that either way.
2118 2012-02-05 23:45:57 <luke-jr> better would be distributing an "unspent outputs" summary ;)
2119 2012-02-05 23:46:00 <gmaxwell> (a bit larger, hahaha. but okay)
2120 2012-02-05 23:46:16 <cjd> it's ready to go in 5 seconds <-- that is huge
2121 2012-02-05 23:46:30 <gmaxwell> luke-jr: Or just have people install thin client or SPV nodes! thats what they are there for!
2122 2012-02-05 23:46:49 <gmaxwell> and their security is boistered by the existance of full validating nodes.
2123 2012-02-05 23:47:10 <FROTUSCI> wait talk about reduction in security.. SPV offers very little security
2124 2012-02-05 23:47:16 <cjd> take away real security, add in gold padlocks and green stripes, people will come in droves.. like it or not, UI is all that people care about.
2125 2012-02-05 23:47:21 <luke-jr> gmaxwell: yes, but this way you validate the blcokchain eventually
2126 2012-02-05 23:47:36 <gmaxwell> luke-jr: You can't validate it eventually from unspent output summaries. :)
2127 2012-02-05 23:48:02 <luke-jr> gmaxwell: you can download the blockchain without breaking user experience
2128 2012-02-05 23:48:07 <gmaxwell> luke-jr: just fully validate it in the background. I don't see why TD is actually so opposed to the bitcoin trust model. Too much SPV interaction must have damaged him.
2129 2012-02-05 23:48:38 <FROTUSCI> SPV is great if you know the tradeoffs
2130 2012-02-05 23:48:46 genjix has quit (Quit: leaving)
2131 2012-02-05 23:48:50 <FROTUSCI> but i dont see what more recent checkpoints have to do w/spv
2132 2012-02-05 23:49:26 <FROTUSCI> in inbound payment can be marked with a yellow flag if you havent validated all the input signatures in the background yet
2133 2012-02-05 23:49:34 <gmaxwell> luke-jr: I'm somewhat frightened by people wanting to degrade full nodes to SPV like security in the name of UX, when its just a false argument.
2134 2012-02-05 23:50:04 <luke-jr> gmaxwell: I think the concern is user experience
2135 2012-02-05 23:50:49 <luke-jr> FROTUSCI has a decent idea actually…
2136 2012-02-05 23:50:50 <gmaxwell> luke-jr: sure, but it's a seperate matter. New installs should bootstrap in SPV mode and fetch the headers backwards, and connect to a preloaded copy.
2137 2012-02-05 23:51:04 <luke-jr> why not just accept blocks blindly, and only validate your own transactions?
2138 2012-02-05 23:51:08 <luke-jr> (not for miners ofc)
2139 2012-02-05 23:51:14 <luke-jr> wait, I know why
2140 2012-02-05 23:51:18 <gmaxwell> luke-jr: no he doesn't. because checking something inputs doesn't preclude double spends unless you've checked everything.
2141 2012-02-05 23:51:35 <luke-jr> gmaxwell: no, you can check for double-spends without doing script validation
2142 2012-02-05 23:51:53 <luke-jr> but unless you check everything, it's possible the block with the txn is invalid
2143 2012-02-05 23:51:57 <luke-jr> even if the txn is
2144 2012-02-05 23:51:59 <gmaxwell> luke-jr: sure but script validation is cheap, the io to find inputs is costly.
2145 2012-02-05 23:52:04 <gmaxwell> Yep that too.
2146 2012-02-05 23:52:26 <FROTUSCI> check all the ancestors that it depends on.. that is cheap compared to checking every tx in history
2147 2012-02-05 23:52:49 jamescarr has joined
2148 2012-02-05 23:53:16 <FROTUSCI> in any case it lets you start using the client immediately in caution mode
2149 2012-02-05 23:53:24 <tcatm> FROTUSCI: Until you hit the coinbase(s). At that point you have to verify all previous blocks completely.
2150 2012-02-05 23:54:08 <FROTUSCI> transactions outside of your input tree wouldnt matter to you
2151 2012-02-05 23:54:12 <gmaxwell> tcatm: even before that, following a chain of transactions doesn't protect you from double spends, as you wouldn't see them. You'd have to trust that someone else prevented the double spends.
2152 2012-02-05 23:54:27 Insti has quit (Ping timeout: 256 seconds)
2153 2012-02-05 23:54:31 SomeoneWeirdzzz is now known as SomeoneWeird
2154 2012-02-05 23:54:38 <gmaxwell> (not a good assumption if the reference client were changed to no longer do this!)
2155 2012-02-05 23:55:00 Insti has joined
2156 2012-02-05 23:55:13 <FROTUSCI> yes theres a whole tree of tx's you need to verify. but its a smaller % of the total tree
2157 2012-02-05 23:55:44 <tcatm> Yes, but Bitcoin is also about being able to trust the chain without having to verify every bit.
2158 2012-02-05 23:56:24 <FROTUSCI> someone could analyze the tree, but ill bet the set of transactions making up the dependency set would be less than 1%
2159 2012-02-05 23:56:25 <gmaxwell> tcatm: indeed, but thats only a sane idea when other people are actually validating the chain, and mining with the extreme risk of being orpahned if they do it wrong.
2160 2012-02-05 23:57:18 <tcatm> Miners must always verify every bit.
2161 2012-02-05 23:58:00 <tcatm> Client nodes shouldn't have to.
2162 2012-02-05 23:58:24 <gmaxwell> tcatm: but if _only_ miners verify anything but their own then there is a bit incentive to lie in a way which is jointly beneificial to miners and not everyone else.
2163 2012-02-05 23:58:45 <luke-jr> did sipa fix the addrman crash/corrupt?
2164 2012-02-05 23:59:13 <gmaxwell> Or at leat that how I've viewed it— the fact that every node is no longer a miner is still compensated for by that fact that there are a great many validators.
2165 2012-02-05 23:59:22 <tcatm> Well, clients should still verify new blocks.
2166 2012-02-05 23:59:53 <gmaxwell> (that the consolidation of mining is offset by the fact that there is still lots of checking)