1 2012-08-28 00:02:54 logger_ has joined
   2 2012-08-28 00:06:09 Z0rZ0rZ0r has quit (Quit: Leaving)
   3 2012-08-28 00:06:34 Z0rZ0rZ0r has joined
   4 2012-08-28 00:07:43 <gmaxwell> weex: CRC? 0_o
   5 2012-08-28 00:07:55 <BlueMatt> wait, we have a crc?
   6 2012-08-28 00:08:04 <sipa> sure
   7 2012-08-28 00:08:11 <sipa> inside TCP packets...
   8 2012-08-28 00:08:14 <BlueMatt> heh
   9 2012-08-28 00:10:14 egecko has joined
  10 2012-08-28 00:14:33 <gmaxwell> sipa: It occured to me today that if we do the UTXO commitment stuff we should also commit to the sum difficulty. This way when a new node starts up it polls its peers for the highest sum diff commitment, and then only pulls the best. So it only ever has to check one chain unless the best that it checks is engaging in an expensive lie.
  11 2012-08-28 00:14:51 copumpkin has quit (Ping timeout: 244 seconds)
  12 2012-08-28 00:16:00 <sipa> isn't that already handled by pulling the headers first?
  13 2012-08-28 00:16:46 copumpkin has joined
  14 2012-08-28 00:16:47 brwyatt has left (Away!~brwyatt@brwyatt.net|)
  15 2012-08-28 00:17:12 <gmaxwell> sipa: this saves you from even having to potentially pull muliple sets of headers.
  16 2012-08-28 00:17:33 brwyatt has joined
  17 2012-08-28 00:17:44 <gmaxwell> (otherwise you may have to pull a complete header set from every peer you consider.)
  18 2012-08-28 00:18:05 [\\\] has quit (Ping timeout: 272 seconds)
  19 2012-08-28 00:24:38 <sipa> gmaxwell: ok, so only connect to one peer :p
  20 2012-08-28 00:25:20 toffoo has joined
  21 2012-08-28 00:25:32 Obsi has quit (Ping timeout: 245 seconds)
  22 2012-08-28 00:26:01 Obsi has joined
  23 2012-08-28 00:27:13 <Diablo-D3> Obsi: any good news?
  24 2012-08-28 00:27:35 <Obsi> re: what?
  25 2012-08-28 00:27:40 <Diablo-D3> abmo
  26 2012-08-28 00:27:59 <Obsi> got 2300 shares of ASICMINER traded in
  27 2012-08-28 00:28:47 <Diablo-D3> hrm
  28 2012-08-28 00:28:48 <eian> anyone know of desktop motherboards that support > 64 gb of ram?
  29 2012-08-28 00:29:32 <Diablo-D3> eian: no
  30 2012-08-28 00:29:32 <kjj_> dare we ask why?
  31 2012-08-28 00:29:32 <eian> heh
  32 2012-08-28 00:29:32 <gmaxwell> eian: are you trying to make a software cosmic ray detector?
  33 2012-08-28 00:29:32 <Diablo-D3> eian: you're not going to fine non-buffered ddr3 sticks big enough
  34 2012-08-28 00:29:32 <eian> gmaxwell, lmao
  35 2012-08-28 00:29:34 <Diablo-D3> actually wait
  36 2012-08-28 00:29:35 <Diablo-D3> lemme look
  37 2012-08-28 00:29:37 <gmaxwell> (>64GB of non-ecc memory??? 0_o)
  38 2012-08-28 00:29:49 <Diablo-D3> gmaxwell: its not ecc thats problematic
  39 2012-08-28 00:29:56 <Diablo-D3> he can get ecc on amd fine
  40 2012-08-28 00:30:00 <eian> I was planning to run operations in triplicate
  41 2012-08-28 00:30:01 PhantomSpark has quit (2!~kvirc@pool-71-251-16-25.nycmny.fios.verizon.net|Read error: Connection reset by peer)
  42 2012-08-28 00:30:02 <Diablo-D3> its getting non-buffered sticks
  43 2012-08-28 00:30:04 <eian> on 3 machines
  44 2012-08-28 00:30:06 <Diablo-D3> because hes limited to 8 slots
  45 2012-08-28 00:30:13 <Diablo-D3> both intel and amd boards only go up to 8
  46 2012-08-28 00:30:19 <eian> so f*ck comsic ray one bit flips
  47 2012-08-28 00:30:28 <Diablo-D3> so he'd need 8gb ddr3 sticks
  48 2012-08-28 00:30:28 PhantomSpark has joined
  49 2012-08-28 00:30:37 <gmaxwell> "News flash: Eian to win noble price for his ground breaking work in neutrino flavor oscillation and PC software based neutrino detector."
  50 2012-08-28 00:30:54 <Diablo-D3> gmaxwell: surprised google hasnt done that
  51 2012-08-28 00:30:55 <kjj_> newegg has a category for 16 GB DIMMs, but nothing in it yet
  52 2012-08-28 00:31:10 <Diablo-D3> kjj_: yeah didnt think so
  53 2012-08-28 00:31:20 <eian> I'm trying to run analysis on one month of collected bitcoin data
  54 2012-08-28 00:31:34 <kjj_> wait.  hang on.  your salvation
  55 2012-08-28 00:31:39 <kjj_> http://www.newegg.com/Product/Product.aspx?Item=N82E16820326202
  56 2012-08-28 00:31:39 <eian> I have an insatiable appetite for memory and limited funds
  57 2012-08-28 00:31:41 <Diablo-D3> er bullshit
  58 2012-08-28 00:31:45 <Diablo-D3> kjj_: the category is full
  59 2012-08-28 00:32:01 <eian> kjj, bahahahah 2k for a module
  60 2012-08-28 00:32:05 user_ has joined
  61 2012-08-28 00:32:07 <Diablo-D3> http://www.newegg.com/Product/ProductList.aspx?Submit=Property&Subcategory=147&Description=&Type=&N=100007611&IsNodeId=1&IsPowerSearch=1&srchInDesc=&MinPrice=&MaxPrice=&PropertyCodeValue=523%3A40080
  62 2012-08-28 00:32:10 <kjj_> seriously?  I did the power search and checked the box for 16GB, got nothing
  63 2012-08-28 00:32:23 <Diablo-D3> oh wait
  64 2012-08-28 00:32:24 <Diablo-D3> 16
  65 2012-08-28 00:32:25 <Diablo-D3> durrr
  66 2012-08-28 00:32:27 <kjj_> ha!
  67 2012-08-28 00:32:32 <Diablo-D3> oh goddamnit
  68 2012-08-28 00:32:34 <gmaxwell> eian: go find some few years old server boards with lots of sockets, you'll save money by using lower density parts.
  69 2012-08-28 00:32:35 <Diablo-D3> fucking kjj
  70 2012-08-28 00:32:40 <Diablo-D3> [08:24:39] <Diablo-D3> so he'd need 8gb ddr3 sticks
  71 2012-08-28 00:32:41 <Diablo-D3> [08:24:39] <Diablo-D3> so he'd need 8gb ddr3 sticks
  72 2012-08-28 00:32:41 <Diablo-D3> [08:24:39] <Diablo-D3> so he'd need 8gb ddr3 sticks
  73 2012-08-28 00:32:41 <Diablo-D3> [08:24:39] <Diablo-D3> so he'd need 8gb ddr3 sticks
  74 2012-08-28 00:32:41 <Diablo-D3> [08:24:39] <Diablo-D3> so he'd need 8gb ddr3 sticks
  75 2012-08-28 00:32:44 <Diablo-D3> WHAT DOES THAT SAY
  76 2012-08-28 00:32:49 <Diablo-D3> IT DOESNT SAY 16
  77 2012-08-28 00:32:55 <Diablo-D3> SAY WHAT ONE MORE TIME
  78 2012-08-28 00:32:56 <eian> gmaxwell, hah - hadn't though of older components
  79 2012-08-28 00:33:02 lunchtime has joined
  80 2012-08-28 00:33:04 <eian> thought*
  81 2012-08-28 00:33:08 <kjj_> meh.  he said > not =
  82 2012-08-28 00:33:23 <Diablo-D3> hrm
  83 2012-08-28 00:33:28 <Diablo-D3> yeah if he wants more than 64
  84 2012-08-28 00:33:29 <sipa> ;;bc,blocks
  85 2012-08-28 00:33:29 <gribble> 196002
  86 2012-08-28 00:33:31 <Diablo-D3> he needs 16gb sticks
  87 2012-08-28 00:33:50 <eian> This is the cheapest machine I could find: http://www.avadirect.com/1u-rack-server-configurator.asp?PRID=16914
  88 2012-08-28 00:33:53 <gmaxwell> eian: I have a bunch of boards with 32 sockets. :)
  89 2012-08-28 00:34:01 <Diablo-D3> eian: oh, you're looking at dedis?
  90 2012-08-28 00:34:03 <Diablo-D3> or what?
  91 2012-08-28 00:34:04 <kjj_> actually, with triple channel, maybe he can find a 12 slot board
  92 2012-08-28 00:34:14 <eian> Diablo, i'm looking for the cheapest system I can build
  93 2012-08-28 00:34:22 lunchtime has left ()
  94 2012-08-28 00:34:23 <Diablo-D3> didnt intel abandon triple channel with sandy/ivy bridge?
  95 2012-08-28 00:34:29 <gmaxwell> (tyan S2945 with M4985 (or whatever, models from memory))
  96 2012-08-28 00:34:37 user_ has quit (Client Quit)
  97 2012-08-28 00:34:48 <Diablo-D3> http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007954%20600082361&IsNodeId=1&name=16GB
  98 2012-08-28 00:34:49 <Diablo-D3> yeah
  99 2012-08-28 00:34:52 <eian> gmaxwell, how do you have all these boards sitting around
 100 2012-08-28 00:34:52 <Diablo-D3> all of those are registered
 101 2012-08-28 00:35:00 <kjj_> there are 12 slot boards, starting at just $479.99 + 11.13 shipping
 102 2012-08-28 00:35:03 luncht1me has joined
 103 2012-08-28 00:35:04 <gmaxwell> eian: they're not sitting around, I have a 160 core cluster.
 104 2012-08-28 00:35:06 <Diablo-D3> eian: if you go with a board that needs registered memory
 105 2012-08-28 00:35:12 <Diablo-D3> you can go with twice as many slots
 106 2012-08-28 00:35:18 <Diablo-D3> 16 and 32 slot boards arent uncommon
 107 2012-08-28 00:35:21 <eian> gmaxwell, what in the ever living hell are you doing :P
 108 2012-08-28 00:35:26 <gmaxwell> SCIENCE.
 109 2012-08-28 00:35:30 <eian> :)
 110 2012-08-28 00:35:42 * Diablo-D3 puts on his goggles
 111 2012-08-28 00:35:53 <gmaxwell> eian: I don't need to explain myself to someone who shows up asking for >64 gb in desktop boards!
 112 2012-08-28 00:35:56 <kjj_> in the server board category, 12 slots for $369.99 on up
 113 2012-08-28 00:36:03 <eian> :P
 114 2012-08-28 00:36:06 <eian> point taken
 115 2012-08-28 00:36:10 <Diablo-D3> kjj_: yeah but registered ram costs more
 116 2012-08-28 00:36:18 <Diablo-D3> eian: how many months do you need this?
 117 2012-08-28 00:36:19 <Diablo-D3> forever?
 118 2012-08-28 00:36:26 <Diablo-D3> because you can get 12gb+ dedis for $110/mo
 119 2012-08-28 00:36:30 <eian> yeah, probably
 120 2012-08-28 00:36:33 <eian> diablo, where?
 121 2012-08-28 00:36:53 <eian> I'm paying 300/mo with gigenet for a single f'ing server
 122 2012-08-28 00:37:31 <Diablo-D3> this is where I say "from me, but Im not open for business yet, so pm me"
 123 2012-08-28 00:37:37 <jgarzik> he blinded me with science.
 124 2012-08-28 00:37:38 <eian> :)
 125 2012-08-28 00:37:42 <Diablo-D3> so, pm me. ;)
 126 2012-08-28 00:37:52 gavinandresen has quit (Quit: gavinandresen)
 127 2012-08-28 00:38:58 * Luke-Jr has 16 GB RAM dedi for under $70/mo <.<
 128 2012-08-28 00:39:58 <kjj_> what kind of data are you talking about, and what kind of analysis?
 129 2012-08-28 00:40:02 user_ has joined
 130 2012-08-28 00:40:28 <jgarzik> My new tattoo, just 1.5 hours ago:  https://plus.google.com/105424721218711536033/posts/CCSBKw3BXtJ
 131 2012-08-28 00:41:05 <kjj_> how do we know that's your name, and not someone else's?
 132 2012-08-28 00:41:38 <eian> Luke, how many?
 133 2012-08-28 00:42:11 <eian> jgarzik, nice
 134 2012-08-28 00:42:12 Obsi has quit (Ping timeout: 245 seconds)
 135 2012-08-28 00:42:56 <jgarzik> kjj_: good question!
 136 2012-08-28 00:43:22 <sipa> kjj_: both options are not mutually exclusive
 137 2012-08-28 00:43:45 pnicholson has quit (Quit: pnicholson)
 138 2012-08-28 00:48:12 one_zero has joined
 139 2012-08-28 00:48:29 <kjj_> too bad fusion-io stuff is still so ungodly expensive.  it'd probably be perfect for your analysis
 140 2012-08-28 00:49:03 <jgarzik> (it's a henna tattoo, which washes off in 3 weeks)
 141 2012-08-28 00:49:14 rbkillea has quit (Ping timeout: 246 seconds)
 142 2012-08-28 00:50:14 * jgarzik was tempted to get an incorrectly spelled word, or maybe the name of an old girlfriend on there
 143 2012-08-28 00:50:57 <gmaxwell> hahahahah
 144 2012-08-28 00:52:57 <eian> kjj, haven't heard of it
 145 2012-08-28 00:53:11 <kjj_> http://www.fusionio.com/products/iodriveduo/
 146 2012-08-28 00:53:28 [\\\] has joined
 147 2012-08-28 00:53:30 <jgarzik> fusion-io is full of open source goodness...  they hired a lot of good kernel hackers
 148 2012-08-28 00:53:44 <jgarzik> (but sadly is closed source itself)
 149 2012-08-28 00:55:49 <eian> what do these things run for?
 150 2012-08-28 00:56:05 <eian> I don't see pricing information
 151 2012-08-28 00:56:06 <kjj_> not too bad.  one arm and a kidney
 152 2012-08-28 00:56:52 <eian> :(
 153 2012-08-28 00:57:18 <kjj_> like $8k each for the single speed 320 GB cards.  the ones I linked are twice as fast, not sure how much they cost
 154 2012-08-28 01:00:04 minimoose has joined
 155 2012-08-28 01:05:07 D34TH has quit (Read error: Connection reset by peer)
 156 2012-08-28 01:08:14 pnicholson has joined
 157 2012-08-28 01:16:38 <kreal> http://www.ocztechnology.com/ocz-revodrive-pci-express-ssd.html
 158 2012-08-28 01:16:55 <kreal> http://www.ocztechnology.com/products/solid-state-drives/pci-express/enthusiast.html
 159 2012-08-28 01:19:57 <sipa> anyone have a link to that one block that didn't claim all fee?
 160 2012-08-28 01:20:07 <sipa> midnightmagic: i seem to remember you knew
 161 2012-08-28 01:20:27 MC-Eeepc has quit (Ping timeout: 244 seconds)
 162 2012-08-28 01:20:45 MC-Eeepc has joined
 163 2012-08-28 01:20:54 <Ferroh> Are there any #bitcoin ops around?
 164 2012-08-28 01:21:03 <Ferroh> please go ban that guy in #bitcoin, it's getting pretty annoying
 165 2012-08-28 01:21:46 <Ferroh> What does it mean when vanitygen says "found delta"?
 166 2012-08-28 01:22:37 Arnavion has quit (Ping timeout: 245 seconds)
 167 2012-08-28 01:23:18 Arnavion has joined
 168 2012-08-28 01:25:54 <BlueMatt> Ferroh: who?
 169 2012-08-28 01:26:21 <BlueMatt> heh, oh, too late
 170 2012-08-28 01:26:49 kreal has quit ()
 171 2012-08-28 01:26:52 paraipan_ has joined
 172 2012-08-28 01:27:10 paraipan has quit (Ping timeout: 276 seconds)
 173 2012-08-28 01:27:19 Obsi has joined
 174 2012-08-28 01:32:43 pnicholson has quit (Ping timeout: 260 seconds)
 175 2012-08-28 01:40:08 eoss has quit (Remote host closed the connection)
 176 2012-08-28 01:40:53 ChuangTzu_ has left ()
 177 2012-08-28 01:41:21 DrNoboto has joined
 178 2012-08-28 01:41:33 <DrNoboto> i (chuangtzu) was banned?
 179 2012-08-28 01:42:04 <DrNoboto> que paso? :P
 180 2012-08-28 01:42:18 <BlueMatt> ?
 181 2012-08-28 01:42:27 <BlueMatt> you mean in #bitcoin?
 182 2012-08-28 01:42:31 <BlueMatt> try again
 183 2012-08-28 01:42:56 <BlueMatt> (it wasnt on purpose)
 184 2012-08-28 01:43:08 oneb has quit (Changing host)
 185 2012-08-28 01:43:08 oneb has joined
 186 2012-08-28 01:43:12 DBordello has joined
 187 2012-08-28 01:44:03 user_ has left ()
 188 2012-08-28 01:52:17 da2ce7_d has joined
 189 2012-08-28 01:54:20 phantomcircuit_ has joined
 190 2012-08-28 01:56:35 Joric has joined
 191 2012-08-28 01:56:35 Joric has quit (Changing host)
 192 2012-08-28 01:56:35 Joric has joined
 193 2012-08-28 01:57:03 <DrNoboto> k
 194 2012-08-28 01:57:04 <DrNoboto> thanks
 195 2012-08-28 02:01:37 brwyatt has quit (Ping timeout: 276 seconds)
 196 2012-08-28 02:02:00 maaku has joined
 197 2012-08-28 02:08:16 brwyatt has joined
 198 2012-08-28 02:09:18 EPiSKiNG- has quit ()
 199 2012-08-28 02:13:10 Luke-Jr has quit (Excess Flood)
 200 2012-08-28 02:13:31 Luke-Jr has joined
 201 2012-08-28 02:16:54 has_many has quit (Ping timeout: 265 seconds)
 202 2012-08-28 02:17:35 buttpirate has joined
 203 2012-08-28 02:18:12 has_many has joined
 204 2012-08-28 02:22:28 maaku has quit (Quit: maaku)
 205 2012-08-28 02:23:35 maaku has joined
 206 2012-08-28 02:26:03 BTCTrader_ has joined
 207 2012-08-28 02:28:03 buttpirate has quit (Remote host closed the connection)
 208 2012-08-28 02:28:08 BTCTrader has quit (Ping timeout: 244 seconds)
 209 2012-08-28 02:28:08 BTCTrader_ is now known as BTCTrader
 210 2012-08-28 02:28:24 Karmaon1 has joined
 211 2012-08-28 02:31:03 Karmaon_ has joined
 212 2012-08-28 02:31:03 Karmaon_ has quit (Changing host)
 213 2012-08-28 02:31:03 Karmaon_ has joined
 214 2012-08-28 02:33:58 Karmaon1 has quit (Ping timeout: 260 seconds)
 215 2012-08-28 02:34:44 Motest031 has quit ()
 216 2012-08-28 02:34:58 Cablesaurus has joined
 217 2012-08-28 02:39:42 fiesh has quit (Ping timeout: 245 seconds)
 218 2012-08-28 02:41:52 fiesh has joined
 219 2012-08-28 02:45:10 da2ce7_d has quit (Ping timeout: 276 seconds)
 220 2012-08-28 02:49:44 <osmosis> is there any python code to convert a raw transcation into json?
 221 2012-08-28 02:51:00 <BlueMatt> pynode could do it, but you'd have to write the json stuff, or 0.7 of the satoshi client has it
 222 2012-08-28 02:51:31 <osmosis> i could probably use getrawtransaction in bitcoind 0.7
 223 2012-08-28 02:51:42 <osmosis> yah
 224 2012-08-28 02:52:38 rdponticelli has quit (Ping timeout: 260 seconds)
 225 2012-08-28 02:54:58 slush has quit (Ping timeout: 260 seconds)
 226 2012-08-28 02:55:31 <Evilmax> hi
 227 2012-08-28 02:55:33 <Evilmax> someone can tell me if exists some software (platfomr) for a btc trade site costruction?
 228 2012-08-28 02:55:40 <Evilmax> what kind of software they use?
 229 2012-08-28 02:57:56 <osmosis> Evilmax, maybe  https://gitorious.org/bitcoin-markets/bitcoin-markets
 230 2012-08-28 02:58:34 <Evilmax> ok tnx
 231 2012-08-28 02:58:36 <osmosis> Evilmax, or  http://gitorious.org/intersango
 232 2012-08-28 02:58:38 <Evilmax> now i see
 233 2012-08-28 02:58:39 da2ce7_d has joined
 234 2012-08-28 03:04:04 cheako has quit (Remote host closed the connection)
 235 2012-08-28 03:07:41 Z0rZ0rZ0r_ has joined
 236 2012-08-28 03:09:50 Z0rZ0rZ0r has quit (Ping timeout: 246 seconds)
 237 2012-08-28 03:12:31 minimoose has quit (Quit: minimoose)
 238 2012-08-28 03:15:29 <Luke-Jr> someone should test IBD with 0.7rc1 IMO
 239 2012-08-28 03:19:53 <gmaxwell> Luke-Jr: Good call, I did one a while back. I'll try it now.
 240 2012-08-28 03:20:20 <Luke-Jr> someone reported problems with next-test IBD, but I see the same buffer changes in 0.7
 241 2012-08-28 03:21:02 <eian> IBD?
 242 2012-08-28 03:21:18 <jgarzik> Initial Block Download (aka network sync)
 243 2012-08-28 03:21:22 <eian> thanks
 244 2012-08-28 03:24:42 <Luke-Jr> on another note, I'm thinking I may need to just say stable backport support is limited to English :/
 245 2012-08-28 03:24:50 <Luke-Jr> Qt Linguist files are not merge-happy
 246 2012-08-28 03:29:15 root2 has quit ()
 247 2012-08-28 03:41:47 <freewil> this seems misleading...
 248 2012-08-28 03:41:49 <freewil> Add -onlynet to prevent connections to a given network
 249 2012-08-28 03:41:59 <freewil> https://github.com/bitcoin/bitcoin/blob/master/doc/release-notes.txt#L63
 250 2012-08-28 03:42:46 phantomcircuit_ has quit (Remote host closed the connection)
 251 2012-08-28 03:43:11 <gmaxwell> freewil: er, onlynet does what it says
 252 2012-08-28 03:43:39 <gmaxwell> freewil: good spotting.
 253 2012-08-28 03:43:43 <freewil> so i can give it a network like 192.168.1.0 and it WONT connect to that?
 254 2012-08-28 03:44:09 <gmaxwell> freewil: no, no. e.g. -onlynet=ipv6  use ipv6 only
 255 2012-08-28 03:44:47 <freewil> ok, thats even less clear then
 256 2012-08-28 03:45:26 <gmaxwell> the commandline help is reasonably clear
 257 2012-08-28 03:45:26 <gmaxwell> init.cpp:        "  -onlynet=<net>         " + _("Only connect to nodes in network <net> (IPv4, IPv6 or Tor)") + "\n" +
 258 2012-08-28 03:45:50 <freewil> yeah that works
 259 2012-08-28 03:45:56 <gmaxwell> The release notes just need to be fixed.
 260 2012-08-28 03:50:45 <gmaxwell> okay, what happened here? http://pastebin.com/sA5NSLQ3
 261 2012-08-28 03:50:58 <gmaxwell> ;;bc,blocks
 262 2012-08-28 03:50:58 <gribble> 196024
 263 2012-08-28 03:51:22 <gmaxwell> it was looking normal then a big run of SetBestChain in a row
 264 2012-08-28 03:51:29 <gmaxwell> it's now stuck at     "blocks" : 180458,
 265 2012-08-28 03:52:03 eb3kk has quit (Read error: Connection reset by peer)
 266 2012-08-28 03:52:37 <gmaxwell> Huston, we have a problem.
 267 2012-08-28 03:53:52 TheSeven has quit (Disconnected by services)
 268 2012-08-28 03:54:01 [7] has joined
 269 2012-08-28 03:55:34 wizkid057 has quit (Quit: be back around 3pm ET tomorrow...)
 270 2012-08-28 04:01:00 paraipan_ has quit (Quit: Saliendo)
 271 2012-08-28 04:02:29 maaku has quit (Quit: maaku)
 272 2012-08-28 04:14:13 <midnightmagic> Could it be possible that the seeming dust transactions which appear to do nothing are actually txn relay delay measurements in an effort to try to map out the network?
 273 2012-08-28 04:14:40 <gmaxwell> perhaps.
 274 2012-08-28 04:14:40 Karmaon_ has quit (Ping timeout: 256 seconds)
 275 2012-08-28 04:14:41 <midnightmagic> So everybody assumes "yeah laundry" but in reality they're serving a purpose?
 276 2012-08-28 04:15:07 <gmaxwell> "laundry" that creates more than a trivial amount of blockchain traffic is wrongheaded in any case.
 277 2012-08-28 04:15:35 <Luke-Jr> midnightmagic: you can measure it without sending
 278 2012-08-28 04:15:54 <midnightmagic> @i2p account was tweeting about a possible attack going on there too
 279 2012-08-28 04:16:21 <midnightmagic> Luke-Jr: If you generate blocks..?
 280 2012-08-28 04:16:30 <gmaxwell> So... yea.. on my IBD fetch test, it trucked along fine to about 180450 and then got stuck until a random peer gave me a block. Amusingly I think it was telling me about an old block. Now I'm fetching very slowly over the network instead of from a fast peer.
 281 2012-08-28 04:16:39 <Luke-Jr> midnightmagic: no, anyone can
 282 2012-08-28 04:16:51 <midnightmagic> Luke-Jr: Time insertions?
 283 2012-08-28 04:16:59 <gmaxwell> I can't tell why it stopped pulling from the fast local peer, nothing in my logs (on either the IBD node or the peer it was pulling from) makes that clear.
 284 2012-08-28 04:17:28 <Luke-Jr> gmaxwell: fast connection = buffer overflow faster?
 285 2012-08-28 04:17:37 <Luke-Jr> slow connection = takes time to fill buffer, which you empty faster
 286 2012-08-28 04:17:45 <gmaxwell> there was no buffer error messages. And it made it all the way to 180xxx.
 287 2012-08-28 04:18:04 <gmaxwell> On either the peer or the node itself.
 288 2012-08-28 04:18:16 <gmaxwell> I'm thinking that perhaps it got disrupted by a new block on the network.
 289 2012-08-28 04:20:55 <Luke-Jr> perhaps
 290 2012-08-28 04:21:25 <midnightmagic> Luke-Jr: Okay, I give up. I have two nodes connecting to radically different peers. How do I get a message with full fanout from one node to eventually hit another without generating blocks and without spending bitcoins in a txn (and potentially paying fees)?
 291 2012-08-28 04:21:53 <gmaxwell> Unrelated, this doesn't look good: http://pastebin.com/2qZYTY0z
 292 2012-08-28 04:22:23 <gmaxwell> midnightmagic: I know a way to do that, but I'd rather not mention it in public due to beans-in-nose factor.
 293 2012-08-28 04:22:32 <midnightmagic> ok
 294 2012-08-28 04:22:37 <Luke-Jr> lol
 295 2012-08-28 04:30:11 maaku has joined
 296 2012-08-28 04:32:30 <osmosis> how can I get libboost1.37-dev  for ubuntu 12.04
 297 2012-08-28 04:34:07 <jgarzik> gmaxwell: the IBD download code is so hacky
 298 2012-08-28 04:34:20 <gmaxwell> jgarzik: I believe you've misspelled broken.
 299 2012-08-28 04:34:30 <jgarzik> gmaxwell: a magic 'inv' prompts you to continue, IFF everything goes peachy
 300 2012-08-28 04:34:33 <MagicalTux> could anyone kill satoshidice or get miners to put more txs in blocks ?
 301 2012-08-28 04:34:43 <MagicalTux> (just saying)
 302 2012-08-28 04:34:53 <jgarzik> gmaxwell: if you don't get to the point where the magic 'inv' is sent, you're stuck
 303 2012-08-28 04:35:09 <gmaxwell> jgarzik: in any case, it started again, so it's not the end of the world. but god knows what happened. :(
 304 2012-08-28 04:35:38 <jgarzik> I think it needs a sanity timeout -- if nothing's happened within 30 seconds, try asking another peer
 305 2012-08-28 04:35:46 <gmaxwell> (and also, I eventually started the node with connect/listen=0 so I could pull the @#$@ thing locall and not wait a week to get this node back up)
 306 2012-08-28 04:38:17 <gmaxwell> midnightmagic: did you see this? http://arxiv.org/abs/1208.2534
 307 2012-08-28 04:41:39 <Luke-Jr> MagicalTux: PM ;p
 308 2012-08-28 04:43:46 eb3kk has joined
 309 2012-08-28 04:44:12 <midnightmagic> gmaxwell: I did not, thanks for the link.
 310 2012-08-28 04:51:00 pnicholson has joined
 311 2012-08-28 04:52:16 <gmaxwell> I .. think, in fact, the ibd logic may end up making it pull from the slowest peer. because it looks like block announcements from delayed peers interrupt the pull and switch the puller to the late announcer.
 312 2012-08-28 04:52:26 maaku has quit (Quit: maaku)
 313 2012-08-28 04:54:12 <jgarzik> gmaxwell: the magic inv is quite similar to non-magic inv's...
 314 2012-08-28 04:55:17 <gmaxwell> OH M YGOD!
 315 2012-08-28 04:55:24 <gmaxwell> https://bitcointalk.org/index.php?topic=103172.msg1138884#msg1138884
 316 2012-08-28 04:55:48 <jgarzik> wow
 317 2012-08-28 04:57:58 <amiller> hm
 318 2012-08-28 05:00:12 <midnightmagic> gmaxwell: It looks like that paper is assuming the observers know their location on the graph, and that the graph edges are known or mapped out (as in the case of the KwaZulu-Natal province (SA) cholera outbreak they used as an example.)
 319 2012-08-28 05:00:21 <Luke-Jr> gmaxwell: ?
 320 2012-08-28 05:00:53 <midnightmagic> holy crap he posted
 321 2012-08-28 05:00:56 <gmaxwell> midnightmagic: Correct, yea I mentioned that when I linked to it previously as a further justification for peer rotation. If the network is constantly changing it should be harder to map.
 322 2012-08-28 05:01:16 <gmaxwell> Luke-Jr: I was oh my goding at there being a post from Hal.
 323 2012-08-28 05:01:23 <Luke-Jr> ah
 324 2012-08-28 05:03:15 pnicholson has quit (Ping timeout: 252 seconds)
 325 2012-08-28 05:08:20 <midnightmagic> gmaxwell: So that paper would be step #2 in my theoretical douchebag mapping the network toolbox..
 326 2012-08-28 05:08:40 <gmaxwell> midnightmagic: right.
 327 2012-08-28 05:09:02 <gmaxwell> first extract graph, then use timings from sparse observers to locate sources on the graph.
 328 2012-08-28 05:16:19 ovidiusoft has joined
 329 2012-08-28 05:20:19 da2ce7 has joined
 330 2012-08-28 05:26:49 maaku has joined
 331 2012-08-28 05:27:14 Joric_ has joined
 332 2012-08-28 05:27:14 Joric_ has quit (Changing host)
 333 2012-08-28 05:27:14 Joric_ has joined
 334 2012-08-28 05:27:48 Joric has quit (Ping timeout: 272 seconds)
 335 2012-08-28 05:38:52 <wumpus> sipa: pushed new sigs, they match now
 336 2012-08-28 05:45:22 da2ce7 has quit (Ping timeout: 248 seconds)
 337 2012-08-28 05:49:49 maaku has quit (Quit: maaku)
 338 2012-08-28 05:50:03 Joric_ is now known as Joric
 339 2012-08-28 05:52:26 skeledrew has quit (Ping timeout: 260 seconds)
 340 2012-08-28 05:54:15 skeledrew has joined
 341 2012-08-28 06:02:29 dvide has quit ()
 342 2012-08-28 06:08:05 da2ce7_d is now known as da2ce7
 343 2012-08-28 06:16:24 jouke has joined
 344 2012-08-28 06:25:39 B0g4r7_ has joined
 345 2012-08-28 06:29:25 extor has quit (Ping timeout: 276 seconds)
 346 2012-08-28 06:31:02 ThomasV has joined
 347 2012-08-28 06:31:09 copperwa has joined
 348 2012-08-28 06:33:54 davout has joined
 349 2012-08-28 06:33:54 davout has quit (Changing host)
 350 2012-08-28 06:33:54 davout has joined
 351 2012-08-28 06:37:10 gjs278 has quit (Ping timeout: 244 seconds)
 352 2012-08-28 06:37:19 gjs278 has joined
 353 2012-08-28 06:41:00 <bitfoo> quick q: how do I compute difficulty from nBits?
 354 2012-08-28 06:41:38 doublec has quit (Read error: Operation timed out)
 355 2012-08-28 06:41:46 <ersi> Hm, does the 'old 0.6.3' client not time out connections after a while? Or was suspending the 'puter just a bad idea, with bitcoin-qt runnin'?
 356 2012-08-28 06:42:17 <amiller> bitfoo, some good resources for that are here: https://en.bitcoin.it/wiki/Difficulty
 357 2012-08-28 06:42:51 <bitfoo> amiller, oh, thanks! searched for "nbits" on the wiki but didn't find this one.
 358 2012-08-28 06:44:34 doublec has joined
 359 2012-08-28 06:56:26 doublec has quit (Ping timeout: 252 seconds)
 360 2012-08-28 07:02:04 iddo has quit (Ping timeout: 252 seconds)
 361 2012-08-28 07:03:08 davout has quit (Remote host closed the connection)
 362 2012-08-28 07:03:38 PK has joined
 363 2012-08-28 07:05:24 _dr has joined
 364 2012-08-28 07:08:13 doublec has joined
 365 2012-08-28 07:08:15 vigilyn has quit (Read error: Connection reset by peer)
 366 2012-08-28 07:11:15 iddo has joined
 367 2012-08-28 07:22:15 RainbowDashh has joined
 368 2012-08-28 07:23:59 Maccer has quit (Excess Flood)
 369 2012-08-28 07:40:13 _dr has quit (Remote host closed the connection)
 370 2012-08-28 07:41:10 Maccer has joined
 371 2012-08-28 07:53:01 BitcoinBaltar has joined
 372 2012-08-28 07:53:05 random_cat_ has joined
 373 2012-08-28 07:53:24 _dr has joined
 374 2012-08-28 07:53:55 random_cat has quit (Ping timeout: 276 seconds)
 375 2012-08-28 07:57:18 forrestv has quit (Ping timeout: 246 seconds)
 376 2012-08-28 07:57:56 TD[gone] has quit (Ping timeout: 245 seconds)
 377 2012-08-28 08:02:42 forrestv has joined
 378 2012-08-28 08:02:42 forrestv has quit (Changing host)
 379 2012-08-28 08:02:42 forrestv has joined
 380 2012-08-28 08:07:31 forrestv has quit (Ping timeout: 245 seconds)
 381 2012-08-28 08:13:44 just4dos_ has quit (Ping timeout: 272 seconds)
 382 2012-08-28 08:19:09 forrestv has joined
 383 2012-08-28 08:19:09 forrestv has quit (Changing host)
 384 2012-08-28 08:19:09 forrestv has joined
 385 2012-08-28 08:20:40 TD[gone] has joined
 386 2012-08-28 08:24:39 davout has joined
 387 2012-08-28 08:25:33 davout_ has joined
 388 2012-08-28 08:28:46 davout has quit (Ping timeout: 244 seconds)
 389 2012-08-28 08:32:31 ThomasV has quit (Ping timeout: 252 seconds)
 390 2012-08-28 08:32:33 just4dos has joined
 391 2012-08-28 08:37:42 sirk390 has joined
 392 2012-08-28 08:40:18 doublec has quit (Ping timeout: 248 seconds)
 393 2012-08-28 08:40:58 doublec has joined
 394 2012-08-28 08:46:34 tucenaber has quit (Ping timeout: 276 seconds)
 395 2012-08-28 08:56:58 CodesInChaos has joined
 396 2012-08-28 08:59:52 tucenaber has joined
 397 2012-08-28 08:59:52 tucenaber has quit (Changing host)
 398 2012-08-28 08:59:52 tucenaber has joined
 399 2012-08-28 09:00:50 doublec has quit (Changing host)
 400 2012-08-28 09:00:50 doublec has joined
 401 2012-08-28 09:04:50 Obsi has quit (Ping timeout: 248 seconds)
 402 2012-08-28 09:05:17 davout_ has quit (Remote host closed the connection)
 403 2012-08-28 09:05:34 davout has joined
 404 2012-08-28 09:05:34 davout has quit (Changing host)
 405 2012-08-28 09:05:35 davout has joined
 406 2012-08-28 09:08:47 sirk390 has quit (Quit: Leaving.)
 407 2012-08-28 09:11:50 luncht1me has quit (Quit: —I-n-v-i-s-i-o-n— 3.3 (November '11) , You know what time it is)
 408 2012-08-28 09:17:43 skeledrew1 has joined
 409 2012-08-28 09:24:59 pjorrit_ has quit (Ping timeout: 246 seconds)
 410 2012-08-28 09:26:11 pjorrit has joined
 411 2012-08-28 09:26:48 tucenaber has joined
 412 2012-08-28 09:27:15 bit2012 has joined
 413 2012-08-28 09:27:23 tucenaber has joined
 414 2012-08-28 09:27:23 Detritus has joined
 415 2012-08-28 09:28:20 toffoo has quit ()
 416 2012-08-28 09:28:35 mmoya has joined
 417 2012-08-28 09:28:51 mndrix has joined
 418 2012-08-28 09:28:51 mndrix has quit (Changing host)
 419 2012-08-28 09:28:51 mndrix has joined
 420 2012-08-28 09:29:21 Z0rZ0rZ0r_ has joined
 421 2012-08-28 09:29:21 mndrix has joined
 422 2012-08-28 09:30:37 forrestv has joined
 423 2012-08-28 09:30:49 pjorrit has quit (Ping timeout: 252 seconds)
 424 2012-08-28 09:32:29 pjorrit has joined
 425 2012-08-28 09:37:05 `2Fast2BCn has joined
 426 2012-08-28 09:44:22 Detritus has quit (Read error: Operation timed out)
 427 2012-08-28 09:45:01 bit2012 has quit (Quit: Computer has gone to sleep.)
 428 2012-08-28 09:46:28 Wack0 has joined
 429 2012-08-28 09:47:40 Detritus has joined
 430 2012-08-28 09:48:06 pjorrit has quit (Ping timeout: 260 seconds)
 431 2012-08-28 09:49:10 sirk390 has joined
 432 2012-08-28 09:52:50 t7 has joined
 433 2012-08-28 09:55:21 slush has joined
 434 2012-08-28 09:58:15 aq83 has quit (Ping timeout: 246 seconds)
 435 2012-08-28 10:00:21 pjorrit has joined
 436 2012-08-28 10:02:10 aq83 has joined
 437 2012-08-28 10:04:53 pjorrit has quit (Ping timeout: 256 seconds)
 438 2012-08-28 10:05:45 pjorrit has joined
 439 2012-08-28 10:06:28 Detritus has quit (Ping timeout: 272 seconds)
 440 2012-08-28 10:07:30 Detritus has joined
 441 2012-08-28 10:10:08 pjorrit has quit (Ping timeout: 245 seconds)
 442 2012-08-28 10:10:56 pjorrit has joined
 443 2012-08-28 10:13:26 iocor has joined
 444 2012-08-28 10:16:31 LuaKT has joined
 445 2012-08-28 10:16:31 LuaKT has quit (Changing host)
 446 2012-08-28 10:16:31 LuaKT has joined
 447 2012-08-28 10:22:19 ThomasV has joined
 448 2012-08-28 10:40:13 PK has quit (Quit: Leaving)
 449 2012-08-28 10:56:38 RainbowDashh has quit (Quit: Mithay (<HOMENAME>) left IRC. (Quit: Hi, I'm a sleep message virus. Please replace your old line with this line and help me take over the world of IRC.))
 450 2012-08-28 10:57:34 has_many has quit (Ping timeout: 244 seconds)
 451 2012-08-28 10:57:51 RainbowDashh has joined
 452 2012-08-28 10:59:55 copumpkin has quit (Ping timeout: 252 seconds)
 453 2012-08-28 11:00:57 has_many has joined
 454 2012-08-28 11:02:24 copumpkin has joined
 455 2012-08-28 11:07:41 ThomasV has quit (Ping timeout: 240 seconds)
 456 2012-08-28 11:19:13 <sirk390> Hi, does anyone know a working testnet3 peer ?
 457 2012-08-28 11:19:53 <sirk390> I can't connect to any peer of the peers on the bootstrap channel
 458 2012-08-28 11:27:47 d4de has quit (Ping timeout: 276 seconds)
 459 2012-08-28 11:30:16 copperwa has quit (Quit: Leaving)
 460 2012-08-28 11:36:29 rdponticelli has joined
 461 2012-08-28 11:41:49 sirk390 has quit (Quit: Leaving.)
 462 2012-08-28 11:50:05 tonikt has quit (Read error: Connection reset by peer)
 463 2012-08-28 11:51:52 datagutt has joined
 464 2012-08-28 11:52:02 davout_ has joined
 465 2012-08-28 11:54:36 davout has quit (Ping timeout: 245 seconds)
 466 2012-08-28 12:01:07 sirk390 has joined
 467 2012-08-28 12:03:11 agricocb has quit (Quit: Leaving.)
 468 2012-08-28 12:05:28 paraipan has joined
 469 2012-08-28 12:06:31 ErnestoJuarell has joined
 470 2012-08-28 12:10:16 minimoose has joined
 471 2012-08-28 12:10:35 Arnavion has quit (Quit: ==kimi ni nante okureba ii no==)
 472 2012-08-28 12:12:27 Arnavion has joined
 473 2012-08-28 12:15:15 d4de has joined
 474 2012-08-28 12:15:16 d4de has quit (Changing host)
 475 2012-08-28 12:15:16 d4de has joined
 476 2012-08-28 12:21:19 sirk390 has quit (Quit: Leaving.)
 477 2012-08-28 12:30:56 agricocb has joined
 478 2012-08-28 12:39:33 vigilyn has joined
 479 2012-08-28 12:42:35 smiddi has quit (Ping timeout: 252 seconds)
 480 2012-08-28 12:43:31 smiddi has joined
 481 2012-08-28 12:53:01 maaku has joined
 482 2012-08-28 13:00:14 bonks has quit (Remote host closed the connection)
 483 2012-08-28 13:01:12 Evilmax has quit ()
 484 2012-08-28 13:06:51 ThomasV has joined
 485 2012-08-28 13:15:32 sirk390 has joined
 486 2012-08-28 13:17:28 jouke has quit (Changing host)
 487 2012-08-28 13:17:28 jouke has joined
 488 2012-08-28 13:20:43 datagutt has quit (Quit: kthxbai)
 489 2012-08-28 13:21:08 datagutt has joined
 490 2012-08-28 13:30:58 <freewil> enum Network ParseNetwork(std::string net) {
 491 2012-08-28 13:30:58 <freewil>     boost::to_lower(net);
 492 2012-08-28 13:30:58 <freewil>     if (net == "ipv4") return NET_IPV4;
 493 2012-08-28 13:30:58 <freewil>     if (net == "ipv6") return NET_IPV6;
 494 2012-08-28 13:30:58 <freewil>     if (net == "tor")  return NET_TOR;
 495 2012-08-28 13:31:01 <freewil>     if (net == "i2p")  return NET_I2P;
 496 2012-08-28 13:31:03 <freewil> what is NET_I2P
 497 2012-08-28 13:31:23 <copumpkin> http://en.wikipedia.org/wiki/I2P
 498 2012-08-28 13:32:06 <gmaxwell> 08/28/12 13:25:11 askfor tx c9e97c7bc7a5518ae39e   1346160311000000 (13:25:11)
 499 2012-08-28 13:32:09 <gmaxwell> 08/28/12 13:25:11 askfor tx c9e97c7bc7a5518ae39e   1346160431000000 (13:27:11)
 500 2012-08-28 13:32:09 * freewil slaps self for not using the google
 501 2012-08-28 13:32:13 <gmaxwell> 08/28/12 13:25:11 askfor tx c9e97c7bc7a5518ae39e   1346160551000000 (13:29:11)
 502 2012-08-28 13:32:16 <gmaxwell> 08/28/12 13:25:11 askfor tx c9e97c7bc7a5518ae39e   1346160671000000 (13:31:11)
 503 2012-08-28 13:32:25 <gmaxwell> hm. I wonder why it's asking for txn more than once?
 504 2012-08-28 13:32:49 <freewil> it's not getting sent?
 505 2012-08-28 13:33:03 <freewil> or it's not being flagged as received properly?
 506 2012-08-28 13:33:05 <gmaxwell> freewil: they're happening so fast it doesn't have a chance.
 507 2012-08-28 13:33:16 <gmaxwell> once it finally does get one it stops.
 508 2012-08-28 13:33:29 <freewil> oh
 509 2012-08-28 13:33:53 <freewil> do those happen in separate threads?
 510 2012-08-28 13:34:10 <gmaxwell> though I boggle about the future request times.
 511 2012-08-28 13:35:23 one_zero has quit ()
 512 2012-08-28 13:37:00 djoot has quit (Read error: Operation timed out)
 513 2012-08-28 13:40:07 <freewil> where are these enum values defined?
 514 2012-08-28 13:40:08 <freewil> https://github.com/bitcoin/bitcoin/blob/master/src/netbase.h#L20
 515 2012-08-28 13:41:02 djoot has joined
 516 2012-08-28 13:41:16 <kjj_> freewil: right there
 517 2012-08-28 13:41:34 <freewil> NET_MAX what does that equal
 518 2012-08-28 13:41:36 <freewil> that must an int
 519 2012-08-28 13:42:15 <kjj_> uh, read this.  http://www.cprogramming.com/tutorial/enum.html
 520 2012-08-28 13:43:49 <freewil> oh got it
 521 2012-08-28 13:44:10 <freewil> so NET_MAX implicitly is 5
 522 2012-08-28 13:45:18 <kjj_> well, probably.  if you don't set it explicitly, I think the value is undefined (meaning the compiler sets whatever it wants)
 523 2012-08-28 13:45:57 RazielZ has joined
 524 2012-08-28 13:46:12 RazielZ has quit (Read error: Connection reset by peer)
 525 2012-08-28 13:46:27 RazielZ has joined
 526 2012-08-28 13:53:49 sirk390 has quit (Quit: Leaving.)
 527 2012-08-28 13:58:12 RainbowDashh has quit (Remote host closed the connection)
 528 2012-08-28 14:00:12 pickett_ has quit (Remote host closed the connection)
 529 2012-08-28 14:07:51 <topace> is there a way to trigger a -rescan from the json api? or do i have to shutdown the daemon and restart it ?
 530 2012-08-28 14:11:11 <kjj_> well...  importing a private key might do it.
 531 2012-08-28 14:11:32 <kjj_> I'm not a pro, but it looks like when you do that, it rescans everything instead of just looking for stuff related to the one new key
 532 2012-08-28 14:12:14 Bachfischer has joined
 533 2012-08-28 14:12:24 <sipa> kjj_: it has no way of knowing which blocks are related to the new key
 534 2012-08-28 14:13:03 <sipa> topace: but why do you want that, the only meaningful point to do it, is when you've copied an old wallet into a new installation
 535 2012-08-28 14:13:52 <kjj_> right, but pwalletMain->ScanForWalletTransactions() doesn't take a key for an argument, so it should be looking at every block for every key, unless I'm missing something
 536 2012-08-28 14:14:53 <kjj_> chances of me missing something is pretty good, of course.
 537 2012-08-28 14:15:03 <topace> just tryhing to figure out if im missing transactions
 538 2012-08-28 14:15:09 <topace> i sent myself a test
 539 2012-08-28 14:15:12 <topace> and it hasnt shown up
 540 2012-08-28 14:15:28 <kjj_> does it show up in listtransactions?
 541 2012-08-28 14:15:46 <topace> on the sender, not the receiver
 542 2012-08-28 14:15:53 <kjj_> has it confirmed yet?
 543 2012-08-28 14:16:06 <topace> still 0 confirmations for my test, but i had a few overnight that were suppsed to come in but didnt
 544 2012-08-28 14:16:59 <topace> if my keypool gets empty can it cause invalid addresses to be given out on getnewaddress ?
 545 2012-08-28 14:17:01 Diablo-D3 has quit (Ping timeout: 244 seconds)
 546 2012-08-28 14:17:03 <sipa> topace: if it shows up with 0 confirmations, it was found alright
 547 2012-08-28 14:17:13 <kjj_> topace: no on the key pool thing
 548 2012-08-28 14:17:16 <sipa> ah, not on the receiver
 549 2012-08-28 14:17:33 <sipa> and rescan only finds transactions missing that are in the blockchain
 550 2012-08-28 14:17:38 <sipa> so have >=1 confirmations
 551 2012-08-28 14:17:48 <kjj_> also, is this you sending to the same wallet?  or sending from one wallet to another?
 552 2012-08-28 14:18:01 <topace> different computers
 553 2012-08-28 14:18:21 <topace> just checking the addresses given out overnight to see if blockchain.info shows anything sent to them
 554 2012-08-28 14:18:23 <sipa> topace: also, since 0.3.20 or os, in almost all cases it rescans automatically
 555 2012-08-28 14:18:25 <topace> maybe the customer just didnt send
 556 2012-08-28 14:18:33 <kjj_> well, basic questions.  is the receiver fully caught up?
 557 2012-08-28 14:18:37 <topace> yes
 558 2012-08-28 14:18:46 <topace>     "blocks" : 196086,
 559 2012-08-28 14:19:21 <topace> bitcoind getreceivedbyaddress 1CZ5P6vfPGaBuXdG19sJ8az4fvyJVRTK71
 560 2012-08-28 14:19:21 <topace> 0.00000000
 561 2012-08-28 14:19:37 <topace> blockchain.info shows the transaction
 562 2012-08-28 14:19:45 <topace> (as unconfirmed still, but at least it sees it)
 563 2012-08-28 14:21:12 <topace> bitcoind getreceivedbyaddress 1CZ5P6vfPGaBuXdG19sJ8az4fvyJVRTK71 0
 564 2012-08-28 14:21:12 <topace> 0.00000000
 565 2012-08-28 14:21:23 sirk390 has joined
 566 2012-08-28 14:21:39 <topace> (its not in listtransactions either)
 567 2012-08-28 14:22:25 <sipa> topace: it's entirely possible the mempool transaction never made it to your receiver node
 568 2012-08-28 14:22:29 <sipa> that is not a problem
 569 2012-08-28 14:22:38 <sipa> the sender will keep broadcasting it until it is in a block
 570 2012-08-28 14:22:42 Arnavion3 has joined
 571 2012-08-28 14:22:42 Arnavion has quit (Disconnected by services)
 572 2012-08-28 14:23:17 tonikt has joined
 573 2012-08-28 14:23:51 Arnavion has joined
 574 2012-08-28 14:25:26 <kjj_> holy crap, there are a lot of pending transactions right now
 575 2012-08-28 14:25:46 <topace> heh maybe thats my issue
 576 2012-08-28 14:26:20 <kjj_> that payment to you doesn't look particularly low priority, so it should show up soon-ish.
 577 2012-08-28 14:26:45 <topace> where can you see pending transactions and their priority?
 578 2012-08-28 14:26:57 <kjj_> http://bitcoincharts.com/bitcoin/txlist/
 579 2012-08-28 14:27:05 Arnavion3 has quit (Ping timeout: 252 seconds)
 580 2012-08-28 14:28:44 Arnavion has quit (Ping timeout: 264 seconds)
 581 2012-08-28 14:28:56 <kjj_> looks like it just hit in the last block.  check for it now
 582 2012-08-28 14:30:09 sirk390 has quit (Quit: Leaving.)
 583 2012-08-28 14:30:44 RazielZ has quit (Read error: Connection reset by peer)
 584 2012-08-28 14:32:25 RazielZ has joined
 585 2012-08-28 14:33:54 sirk390 has joined
 586 2012-08-28 14:35:08 Evilmax has joined
 587 2012-08-28 14:36:34 djoot has quit (Quit: leaving)
 588 2012-08-28 14:38:27 maaku has quit (Quit: maaku)
 589 2012-08-28 14:39:27 agricocb has quit (Quit: Leaving.)
 590 2012-08-28 14:41:34 gavinandresen has joined
 591 2012-08-28 14:43:37 sirk390 has quit (Quit: Leaving.)
 592 2012-08-28 14:45:16 Arnavion has joined
 593 2012-08-28 14:45:39 djoot has joined
 594 2012-08-28 14:45:39 djoot has quit (Changing host)
 595 2012-08-28 14:45:39 djoot has joined
 596 2012-08-28 14:48:03 Internet13 has quit (Ping timeout: 246 seconds)
 597 2012-08-28 14:52:51 agricocb has joined
 598 2012-08-28 14:55:21 <topace> yea its there now, thanks
 599 2012-08-28 14:55:29 <topace> i guess the other 3 orders overnight just didnt send the coins
 600 2012-08-28 14:56:04 Internet13 has joined
 601 2012-08-28 14:56:29 iocor has quit (Quit: Computer has gone to sleep.)
 602 2012-08-28 15:01:01 ThomasV has quit (Ping timeout: 240 seconds)
 603 2012-08-28 15:02:43 <kjj_> Hey Gavin, regarding our other chain discussion the other day, it's even worse than I thought.
 604 2012-08-28 15:04:26 <kjj_> since the second chain has no difficulty, and the first chain can only handle ordering in a very rough sense, reorg attacks are super easy
 605 2012-08-28 15:04:34 safra has joined
 606 2012-08-28 15:04:40 <safra> hello
 607 2012-08-28 15:04:49 <safra> anyone know paysius?
 608 2012-08-28 15:04:57 <safra> i have a problem with it
 609 2012-08-28 15:05:54 maaku has joined
 610 2012-08-28 15:06:04 Arnavion3 has joined
 611 2012-08-28 15:06:04 Arnavion has quit (Disconnected by services)
 612 2012-08-28 15:07:23 <jgarzik> safra: offtopic, try #bitcoin-otc
 613 2012-08-28 15:07:39 <gavinandresen> kjj_: How so?  the Bitcoin blockchain defines the canonical ordering, so unless you can cause a re-org on the bitcoin blockchain how does the attack work?
 614 2012-08-28 15:08:13 <kjj_> because bitcoin doesn't understand the magic cookies that we are embedding into it.  as far as bitcoin is concerned, they are just regular transactions
 615 2012-08-28 15:08:17 <jgarzik> "second chain has no difficulty" -- /me wonders if this is some other bitcoin, other than the one I'm familiar with
 616 2012-08-28 15:08:40 <gavinandresen> kjj wants a chain of stock exchange transactions.
 617 2012-08-28 15:08:55 <gavinandresen> (initial offering and then trading of shares)
 618 2012-08-28 15:08:58 <sipa> gavinandresen: wumpus' and my build for 0.7.0rc1 match exactly
 619 2012-08-28 15:09:17 <gavinandresen> sipa: nice, do you have permission to upload to sourceforge?
 620 2012-08-28 15:09:21 <sipa> gavinandresen: no
 621 2012-08-28 15:09:44 <sipa> luke told me his build matches wumpus', but that was before wumpus rebuilt with a more recent base image
 622 2012-08-28 15:10:23 <gavinandresen> I fixed my gitian machine (had to reinstall ubuntu), I should have builds in a couple of hours
 623 2012-08-28 15:11:45 spitteler has quit (Ping timeout: 272 seconds)
 624 2012-08-28 15:17:22 <gavinandresen> kjj : I think you misunderstood what I was suggesting.  My suggestion is a separate transaction chain; periodically all those transactions are hashed together and a summary hash put into the bitcoin blockchain, either as a plain-old transaction (with a fee) or in a miner's coinbase.
 625 2012-08-28 15:17:49 <gmaxwell> I'm contemplating making CreateTransaction take an nOverPay parameter specifying the maximum amount its allowed to overpay the output in order to eliminate change or randomly transfer change jaggyness to the output side.
 626 2012-08-28 15:18:01 <gavinandresen> kjj_: if two people put two different transaction history summaries that disagreed, then whichever one ended up in the bitcoin blockchain first would be considered the valid one.
 627 2012-08-28 15:18:51 <gmaxwell> Miner's coinbase is obviously much more scalable.  What gavin's suggesting is unlike namecoin merged mining in that the second 'chain' depends on the bitcoin chain.
 628 2012-08-28 15:18:53 <kjj_> gavinadresen: how would one be first?  wouldn't the double spender transmit his summary during the same BTC block period?
 629 2012-08-28 15:19:17 <gavinandresen> Blocks and transactions have a well-defined ordering.
 630 2012-08-28 15:19:46 <gavinandresen> "first" transaction would be the one in the earliest block, in the earliest position in that block's transaction list.
 631 2012-08-28 15:20:01 <jgarzik> ugh.  if we want to encourage people to store non-bitcoin-related spam, let's add a rule that requires 0.01 BTC fee for any TX with OP_DROP
 632 2012-08-28 15:20:25 <edcba> jgarzik chasing fees :)
 633 2012-08-28 15:20:27 <kjj_> jgarzik: that's a very good idea
 634 2012-08-28 15:20:27 <gavinandresen> jgarzik: meh.  the right transaction fee policy will make the right things happen.
 635 2012-08-28 15:20:42 <jgarzik> ...for data publishers
 636 2012-08-28 15:20:45 <jgarzik> not necessarily for bitcoin holders
 637 2012-08-28 15:20:46 <gmaxwell> gavinandresen: meh.
 638 2012-08-28 15:20:50 <gmaxwell> what jgarzik said. :)
 639 2012-08-28 15:21:10 <gmaxwell> the right thing is to use coinbase merge mining to bind this data. It has perfect O(1) scaling for infinite different crazy uses.
 640 2012-08-28 15:21:15 <gavinandresen> "let's add a fee rule" should be up to the miners, in my humble opinion.
 641 2012-08-28 15:21:21 <gmaxwell> But it requires marginally more startup costs.
 642 2012-08-28 15:21:24 <kjj_> over time it would adjust, but .01 BTC isn't a bad default for now.  The next ASCII bernake would cost a buck
 643 2012-08-28 15:21:45 <gavinandresen> ASCII bernanke didn't use OP_DROP
 644 2012-08-28 15:22:01 <gavinandresen> (which is why I'm "meh" -- if we try to impose rules, people will try to get around them)
 645 2012-08-28 15:22:02 <topace> kjj_: so the customer _swears_ they sent coins to address: 1F2WLosYc3ifVs78UPffYrTT1xUQgqN5LX last night
 646 2012-08-28 15:22:03 <edcba> gavinandresen: miners would want to add a 99% fee for each tx :)
 647 2012-08-28 15:22:08 <gmaxwell> kjj_: And the reason to add friction there is to encourage people to use more scalable solutions as much as anything else.
 648 2012-08-28 15:22:10 <topace> but its not on blockchain.info
 649 2012-08-28 15:22:17 eian has quit (Read error: Connection reset by peer)
 650 2012-08-28 15:22:20 <kjj_> gmaxwell: merged mining was my first idea, but it has the problem that this chain isn't monetary, and shouldn't act like it is.  the namecoin problem
 651 2012-08-28 15:22:29 eian has joined
 652 2012-08-28 15:22:32 <jgarzik> kjj_: that's OK
 653 2012-08-28 15:22:47 <gmaxwell> kjj_: I .. don't see what that has to do with anything 0_o
 654 2012-08-28 15:23:04 <jgarzik> a non-monetary data chain should work just fine
 655 2012-08-28 15:23:23 <gmaxwell> Data commitment in bitcoin works just dandy.
 656 2012-08-28 15:23:31 <gavinandresen> yes-- has anybody written up a design doc for non-monetary data chains?
 657 2012-08-28 15:23:32 <kjj_> namecoin has about 40% of the difficulty of bitcoin, despite being super easy to set up and mine for free
 658 2012-08-28 15:23:55 <kjj_> people don't bother doing even that tiny bit of work because namecoinds are worthless (nearly so)
 659 2012-08-28 15:24:07 <gavinandresen> (ooh, sounds like a good task for while SOMEBODY is lounging by the beach working on their tan....)
 660 2012-08-28 15:24:32 slush1 has joined
 661 2012-08-28 15:24:44 <gmaxwell> kjj_: It's not quite super easy, and people think it diminishes their bitcoin returns, but regardless.... I still don't see what your point is. A data commitment doesn't have a difficulty other than the bitcoin one.
 662 2012-08-28 15:24:59 <gmaxwell> kjj_: there are many ways to use merged mining. The namecoin way where the chains are independant is only one.
 663 2012-08-28 15:25:24 <kjj_> topace: I don't see that in the pending transaction list
 664 2012-08-28 15:25:25 <gmaxwell> It's equally possible to use bitcoin for O(1) order/timestamping as UukGoblin's chronobit does.
 665 2012-08-28 15:25:40 <topace> kjj_: yea i checked there too.
 666 2012-08-28 15:25:52 <kjj_> topace: have them double check
 667 2012-08-28 15:25:54 <topace> would there be naother explaination, besides the customer is lying?
 668 2012-08-28 15:26:42 eian has quit (Ping timeout: 245 seconds)
 669 2012-08-28 15:26:42 kreal has joined
 670 2012-08-28 15:27:02 <kjj_> topace: could also be user error.  have them double check
 671 2012-08-28 15:27:36 eian has joined
 672 2012-08-28 15:27:37 <topace> yea, they have "triple checked" supposedly.
 673 2012-08-28 15:27:54 <kjj_> gmaxwell: it isn't just data commitment.  I don't want merely a log of stock transactions, I want my shares to be "owned" by my private keys, just like my bitcoins are owned by my bitcoin keys
 674 2012-08-28 15:28:12 <jgarzik> gavinandresen: :)
 675 2012-08-28 15:28:14 <kjj_> topace: can they get you a transaction ID?
 676 2012-08-28 15:28:38 <gmaxwell> kjj_: And?
 677 2012-08-28 15:28:48 <gmaxwell> kjj_: you only need ordering for that. Bitcoin can provide the ordering.
 678 2012-08-28 15:29:31 <gmaxwell> kjj_: so long as your application does not need lite nodes it doesn't matter that bitcoin isn't enforcing the rules.
 679 2012-08-28 15:29:38 <kjj_> gmaxwell: to an extent.  it isn't like the order of transactions in a block is meaningful.  you can pick one at random
 680 2012-08-28 15:29:59 <gmaxwell> I have no idea what you're talking about there.
 681 2012-08-28 15:30:15 davout_ has quit (Remote host closed the connection)
 682 2012-08-28 15:30:31 davout has joined
 683 2012-08-28 15:30:32 davout has quit (Changing host)
 684 2012-08-28 15:30:32 davout has joined
 685 2012-08-28 15:30:34 <kjj_> I mean that if I send a bitcoin transaction, wait a minute and send a second spend, the network will reject the second one, most of the time
 686 2012-08-28 15:30:42 <gmaxwell> You commit to kjjblocks in bitcoin coinbases. KJJ nodes know the real KJJblocks based on what got committed into bitcoin. kjjblocks can't be reordered without reordering bitcoin.
 687 2012-08-28 15:30:49 <kjj_> but only one of them will end up in a block
 688 2012-08-28 15:30:54 <gmaxwell> kjj_: yes, thats true and totally irrelevant.
 689 2012-08-28 15:31:45 <kjj_> ok, if you do it in the coinbase, it means that miners are involved.  which means that miners need an incentive
 690 2012-08-28 15:32:37 <gmaxwell> The incentive is trivial, first: the miners don't need to run your service, they only need to run a generic timestamping service.
 691 2012-08-28 15:32:59 <gmaxwell> (for namecoin you actually have to run namecoin, and if you're a pool you have to handle payouts and this is not at all trivial)
 692 2012-08-28 15:33:36 <gmaxwell> The primary incentive is to keep your BS data storage out of the bitcoin blockchain where it detracts from people using bitcoin as a currency, as intended.
 693 2012-08-28 15:34:07 <gmaxwell> And if we get a good implemention of generic merged timestamping, I expect we'll integrate it in the reference software; so there shouldn't be any setup either.
 694 2012-08-28 15:35:12 <kjj_> I'm missing the step where a timestamping service replaces the merged-chain node software
 695 2012-08-28 15:36:12 <gmaxwell> people interested in your service run your software. They can all trivially agree if a sequence of blocks is valid, but they can't agree on the ordering without some help.
 696 2012-08-28 15:36:48 <gmaxwell> They send block identities to the generic timestamping service, and they watch the bitcoin chain to validate the timestamps and get their orders.
 697 2012-08-28 15:36:56 eian has quit (Read error: Connection reset by peer)
 698 2012-08-28 15:37:00 <kjj_> I would phrase it differently.  they can't tell which of several valid sequences is correct, but go on
 699 2012-08-28 15:37:07 eian has joined
 700 2012-08-28 15:39:01 <kjj_> so, the rule would be "block with the oldest timestamp that made it into the main bitcoin chain" ?
 701 2012-08-28 15:39:32 djoot has quit (Quit: leaving)
 702 2012-08-28 15:39:33 <gmaxwell> Lowest hash rather than oldest timestamp most likely.
 703 2012-08-28 15:40:26 <gmaxwell> (you want a tiebreaking rule thats hard to influence)
 704 2012-08-28 15:40:27 <kjj_> ok, so everyone churns hashes over and over again until they get one lower than the lowest one they know about, and send it off for stamping
 705 2012-08-28 15:41:10 <gmaxwell> well they don't even need to, if there is no direct monetary incentive for being the node that made the block they just send what they have.
 706 2012-08-28 15:41:45 <gmaxwell> I guess to prevent flooding the timestamper you probably would have a small threshold to make it stiocastic. But it's only for ratelimiting, not security.
 707 2012-08-28 15:41:52 <kjj_> what about someone that wants to sell the same stocks twice?
 708 2012-08-28 15:42:57 <gmaxwell> kjj_: Your rules prevent that. A sequence that violates the rules is invalid no matter how much its timestampped.
 709 2012-08-28 15:44:31 <kjj_> but we have two blocks, each sending the same stock to two different recipients.  the attacker has until the next bitcoin block to keep trying to find a lower hash for the second one
 710 2012-08-28 15:45:15 <kjj_> and it won't be hard to do that, since no one else is looking for lower hashes because they have no incentive to, unless they are the recipient of the first transaction
 711 2012-08-28 15:45:17 <sipa> by the way: i uploaded 0.7.0rc1
 712 2012-08-28 15:45:19 DBordello has quit (Ping timeout: 272 seconds)
 713 2012-08-28 15:45:26 djoot has joined
 714 2012-08-28 15:45:29 djoot has quit (Changing host)
 715 2012-08-28 15:45:30 djoot has joined
 716 2012-08-28 15:45:37 <da2ce7> Any Windows experts, please consider having a look at: https://bitcointalk.org/index.php?topic=104075.0  <- I'm having great troubles.
 717 2012-08-28 15:45:39 <gmaxwell> And? You can't trust a transaction is complete until its settled in the chain, this isn't any different that bitcoin other than some small factor on how long you would have to wait.
 718 2012-08-28 15:46:14 <gmaxwell> kjj_: you can also show people the double spend and prove the partipants cheating, which might have a lot more incentive than we generically have with bitcoin.
 719 2012-08-28 15:46:30 <sipa> da2ce7: i know nothing about .net, sorry
 720 2012-08-28 15:46:32 <kjj_> good points
 721 2012-08-28 15:47:13 <sipa> gmaxwell: you had troubles with 0.7.0rc1 syncing from scratch?
 722 2012-08-28 15:47:52 DBordello has joined
 723 2012-08-28 15:48:03 <gmaxwell> sipa: kinda. It did complete.  See the backscroll. It made it fairly high and then just stopped. It isn't obvious to me why it stopped. It eventually continued when some really slow peer announced the current best block to me.
 724 2012-08-28 15:48:21 <sipa> we need reverse-headers fetch:)
 725 2012-08-28 15:48:21 <gmaxwell> (and then of course started pulling from that slow peer instead of my local node)
 726 2012-08-28 15:48:49 maaku has quit (Quit: maaku)
 727 2012-08-28 15:48:51 <da2ce7> sipa: no prob.
 728 2012-08-28 15:51:17 Arnavion3 has left ()
 729 2012-08-28 15:51:29 Arnavion has joined
 730 2012-08-28 15:54:54 Evilmax has quit ()
 731 2012-08-28 15:57:22 <kjj_> gmaxwell: I still sorta don't like that the "real" next block is more or less random.  And that the only way to have a chance of being included is to send it off for timestamping yourself
 732 2012-08-28 15:58:02 <kjj_> and possibly having to may multiple timestamping fees if a block that doesn't include your transaction gets luckier than yours
 733 2012-08-28 15:58:03 <devrandom> sipa: any luck with gitian?
 734 2012-08-28 15:59:08 <sipa> devrandom: yup, did a perfect build that matches byte-for-byte with wumpus'
 735 2012-08-28 15:59:13 <jgarzik> sipa, gmaxwell, gavinandresen: cool beans.  0.7rc SF upload seen here.
 736 2012-08-28 15:59:29 <devrandom> sipa: yay for perfection
 737 2012-08-28 15:59:46 <jgarzik> naming of the stuff is a bit vague "changes after foo.txt" and "contrib after foo.txt"
 738 2012-08-28 16:00:07 <devrandom> sipa: I should add checks for common issues (e.g. apt-cacher-ng not running)
 739 2012-08-28 16:00:20 <jgarzik> maybe changelog.txt and changelog-summary.txt
 740 2012-08-28 16:02:22 <sipa> jgarzik: i'll update my script before 0.7.0 final ;)
 741 2012-08-28 16:07:52 <gmaxwell> kjj_: there should be no timestamping fees at all. bitcoin timestamping has O(1) scaling. An infinite amount of data can be timetsamped in a block with constant cost.
 742 2012-08-28 16:08:09 ErnestoJuarell has quit (Quit: Leaving)
 743 2012-08-28 16:08:42 <gmaxwell> kjj_: I expect that timestamp aggregators will just use POW (mining a diff .1 bitcoin share for me) to rate-limit submission.
 744 2012-08-28 16:08:47 <kjj_> do you have a link I can read for how that would work?
 745 2012-08-28 16:09:35 <kjj_> or something I can search for?
 746 2012-08-28 16:11:43 <gavinandresen> devrandom: I'm having trouble with make_base_vm:  Process (['umount', '/tmp/tmpTaKWjw/dev']) returned 1. stdout: , stderr: umount: /tmp/tmpTaKWjw/dev: device is busy.
 747 2012-08-28 16:12:08 <jgarzik> bah.  why does every project (including my own) written in python have to begin with "py" prefix?
 748 2012-08-28 16:12:14 <jgarzik> makes command line completion annoying
 749 2012-08-28 16:12:30 <kjj_> jgarzik: marketing
 750 2012-08-28 16:15:06 <gmaxwell> kjj_: the most I can point you to easily is chronobit, which goes a step further and makes the commitments transitive through p2pool shares as well. (so all p2pool miners can participate). But the basic idea is that the miner only puts in a single merged mining hash (size 0 or 32 bytes depending on if they have one), which connects a tree.. which may connect other trees and so on. So your timestamp is the sequences of hashes that connect you
 751 2012-08-28 16:15:19 <gmaxwell> you can have them commit a hash which is the root of an aggregation tree run by someone else.
 752 2012-08-28 16:15:31 <mrtn> jgarzik: would jsomething be better ?
 753 2012-08-28 16:15:32 <gmaxwell> Hm. I did actually write up a long pastebin about this at some point.
 754 2012-08-28 16:15:37 <mrtn> jgarzik: I'd blame the iCompany ;)
 755 2012-08-28 16:16:05 <mrtn> jgarzik: how many attendees at the london event alread have registered?
 756 2012-08-28 16:16:19 <mrtn> wrong.
 757 2012-08-28 16:16:29 <midnightmagic> kjj_: namecoins aren't worthless if you are making use of the name resolve purpose for which namecoin exists
 758 2012-08-28 16:16:59 <midnightmagic> kjj_: The price is being depressed by people dumping namecoins to maximize bitcoin "profit"
 759 2012-08-28 16:17:20 soaring_eagle has joined
 760 2012-08-28 16:17:39 sebicas has joined
 761 2012-08-28 16:17:56 soaring_eagle has quit (Client Quit)
 762 2012-08-28 16:18:15 davout has quit (Remote host closed the connection)
 763 2012-08-28 16:18:17 soaring_eagle has joined
 764 2012-08-28 16:18:53 soaring_eagle is now known as pnicholson
 765 2012-08-28 16:23:37 pnicholson has quit (Quit: cya later!)
 766 2012-08-28 16:23:55 soaring_eagle has joined
 767 2012-08-28 16:26:07 soaring_eagle has quit (Client Quit)
 768 2012-08-28 16:26:36 soaring_eagle has joined
 769 2012-08-28 16:26:58 <BlueMatt> gavinandresen: did you modify the addpkg list to include wine or otherwise are pre-installing binfmt-support?
 770 2012-08-28 16:27:20 soaring_eagle has quit (Client Quit)
 771 2012-08-28 16:27:25 <BlueMatt> (which, btw I recommend - makes building for win32 much faster, but you have to install all deps except wine, or it will break)
 772 2012-08-28 16:27:28 <gavinandresen> BlueMatt: huh what?
 773 2012-08-28 16:27:42 pnicholson has joined
 774 2012-08-28 16:27:49 <BlueMatt> nevermind, I suppose you didnt modify the make-base-vm script...
 775 2012-08-28 16:28:15 <sipa> gavinandresen: that's a known problem
 776 2012-08-28 16:28:30 <gavinandresen> BlueMatt: nope.  If you have a spiffy way of speeding things up, please modify contrib/gitian-descriptors/README ....
 777 2012-08-28 16:28:30 <sipa> gavinandresen: but did you checkout a recent version of gitian-builder?
 778 2012-08-28 16:28:43 <BlueMatt> gavinandresen: try rebooting/reinstalling (or, atleast rebooting when doing that using livecd sometimes wfm)
 779 2012-08-28 16:28:47 <gavinandresen> sipa: yes, rebuilt the entire filesystem from scratch last night
 780 2012-08-28 16:29:01 <sipa> gavinandresen: hmm, maybe try LXC?
 781 2012-08-28 16:29:29 <gavinandresen> BlueMatt: I'll try rebooting.  After two tries the 64-bit image created successfully
 782 2012-08-28 16:29:42 <sipa> gavinandresen: otherwise, look here: https://github.com/devrandom/gitian-builder/issues/12
 783 2012-08-28 16:30:47 <gavinandresen> sipa: cool, thanks
 784 2012-08-28 16:30:58 aq83 has quit (Ping timeout: 246 seconds)
 785 2012-08-28 16:31:09 * gavinandresen goes back to figuring out why macdeployqtplus stopped working.....
 786 2012-08-28 16:31:27 <BlueMatt> gavinandresen: re: "spiffy" way of speeding things up...meh dont feel like committing it anywhere because itl probably break if any package names/versions/etc change, change the addpkg list in make-base-vm to include all wine deps except wine itself and binfmt-support, and add mingw32 then build a second i386 image, and use that as base when building for win32
 787 2012-08-28 16:32:18 <sipa> ;;bc,blocks
 788 2012-08-28 16:32:19 <gribble> 196101
 789 2012-08-28 16:32:21 TD has joined
 790 2012-08-28 16:32:59 <BlueMatt> (but thats helped pretty significantly by using the latest version (as of a few months ago) wen it uses writeback for the tmp disk
 791 2012-08-28 16:33:47 <TD> even with leveldb syncing the block chain is a pain in the arse
 792 2012-08-28 16:33:55 <TD> not sure where the time is going in the last parts of the chain, i guess it's all cpu
 793 2012-08-28 16:34:02 * TD is building ultraprune now to try it out
 794 2012-08-28 16:34:42 <sipa> seems that in my ultraprune branch, after the last checkpoint, most time is still spent in BDB committing
 795 2012-08-28 16:34:53 * sipa contemplates ultraprune+leveldb
 796 2012-08-28 16:35:00 <Joric> may i fetch ultraprune somewhere?
 797 2012-08-28 16:35:09 p0s has joined
 798 2012-08-28 16:35:11 <TD> that shoiuld be an easy rebase
 799 2012-08-28 16:35:37 <sipa> TD: i did move to two separate databases, is that a problem?
 800 2012-08-28 16:35:51 <sipa> jgarzik: sure, sipa/bitcoin on github, branch ultraprune
 801 2012-08-28 16:35:54 <sipa> eh, Joric
 802 2012-08-28 16:35:57 <TD> ok, maybe not so easy :-)
 803 2012-08-28 16:36:06 <TD> ctxdb changed of course
 804 2012-08-28 16:36:45 <Joric> bitcoin-0.7.0rc1-win32.zip < 2 hours ago hurray
 805 2012-08-28 16:37:16 <sipa> TD: but i don't think I even need the entire cdb stuff anymore
 806 2012-08-28 16:37:27 <TD> sipa: basically CTxDB moved out of db.h and became its own object (not CDB related)
 807 2012-08-28 16:37:40 <TD> and there's an #ifdef in a new <txdb.h> to select bdb or leveldb
 808 2012-08-28 16:37:42 <jgarzik> whee.  Prediction markets lawyer guy got back to me, with encouraging news about running a legal prediction market in the US.
 809 2012-08-28 16:37:55 <sipa> TD: as CCoinsDB is just opened once and kept open the entire time, and CChainDB only gets non-atomic single writes
 810 2012-08-28 16:37:58 <BlueMatt> jgarzik: you are gonna run a market
 811 2012-08-28 16:38:19 <jgarzik> BlueMatt: if I can do so legally... hopefully yes
 812 2012-08-28 16:38:20 <TD> sipa: yeah the whole CTxDB abstraction makes no sense but i kept it to keep the changeset minimal
 813 2012-08-28 16:38:26 <TD> jgarzik: why not make it a pure p2p market? :)
 814 2012-08-28 16:38:34 <TD> jgarzik: then you just write the code and nobody operates it, exactly
 815 2012-08-28 16:38:37 <sipa> TD: so I think i'll just import the leveldb code, and then manually patch the code layer above CDB to use that directly
 816 2012-08-28 16:38:40 <jgarzik> TD: I don't want to get arrested :)
 817 2012-08-28 16:38:54 <jgarzik> US gambling laws are very annoying
 818 2012-08-28 16:39:18 <BlueMatt> ?
 819 2012-08-28 16:39:28 <TD> sipa: the leveldb api is really straightforward
 820 2012-08-28 16:39:35 <Joric> may you please replace the splash image? this picture and colors are godawful
 821 2012-08-28 16:39:37 <TD> sipa: but yeah i guess you'd have to do some merge work
 822 2012-08-28 16:39:39 <sipa> yes, I looked at it already, didn't look hard at all
 823 2012-08-28 16:39:41 <TD> it's either i go first, or you go first
 824 2012-08-28 16:39:45 <TD> then i rebase on top of your code
 825 2012-08-28 16:39:46 <jgarzik> BlueMatt: grey area, where I might get nabbed for "aiding and abetting" by writing software.
 826 2012-08-28 16:39:48 <TD> either works for me
 827 2012-08-28 16:39:53 <jgarzik> plus, less compensation for my time :)
 828 2012-08-28 16:39:55 <TD> my migration code is kind of crap
 829 2012-08-28 16:40:01 <sipa> TD: meh, I'll just cherry-pick the leveldb commits themselves, and write my own glue
 830 2012-08-28 16:40:02 <TD> i think you mentioned you were going to redo it in a better way?
 831 2012-08-28 16:40:04 aq83 has joined
 832 2012-08-28 16:40:06 <TD> ok
 833 2012-08-28 16:40:34 <BlueMatt> jgarzik: nice
 834 2012-08-28 16:41:05 <jgarzik> TD: I was thinking about that too (sans legal annoyances).  Still requires humans somewhere, AFAICS... though maybe you could (a) set up two bitcoin addresses for each outcome (yes + no), (b) require payouts to require multiple signatures, (c) have an independent set of experts to verify outcomes, and each sign
 835 2012-08-28 16:41:25 <jgarzik> but collusion possibilities seem insurmountable
 836 2012-08-28 16:43:04 <jgarzik> something/someone must verify that an event reached, or did not reach, the given criteria
 837 2012-08-28 16:43:35 rdponticelli has quit (Ping timeout: 272 seconds)
 838 2012-08-28 16:43:54 <has_many> fwiw: bitcoin-qt and bitcoind are unusable on os x: not because of hd trashing or cpu usage or downloading 2 gb data initially. no, it's because of time machine. every hour multi GB backup sucks. any chance to move the huge "always dirty" dbs in to a subdir and keep the wallet above? (for T M exclusion?) or should i recommend/use other "lightweight" clients like MultiBit?
 839 2012-08-28 16:44:32 <sipa> has_many: will happen soon
 840 2012-08-28 16:44:38 <sipa> 0.8 probably
 841 2012-08-28 16:44:44 <has_many> thanks :)
 842 2012-08-28 16:45:04 <sipa> as we're moving to new databases, with each in a separate directory
 843 2012-08-28 16:45:15 <TD> jgarzik: yeah i think it's interesting to consider p2p designs.
 844 2012-08-28 16:45:34 <TD> jgarzik: look at it this way. if it's considered "aiding and abetting" to write such software then arguably you're already hosed for working on bitcoin anyway ....
 845 2012-08-28 16:45:43 Turingi has joined
 846 2012-08-28 16:45:48 <TD> jgarzik: did you read the wiki section about oracleS?
 847 2012-08-28 16:46:03 <devrandom> gavinandresen: that issue should have been fixed (I posted an update to issue 12)
 848 2012-08-28 16:46:11 <jgarzik> TD: levels of separation.  specifically working on "gambling software" is potentially a danger zone
 849 2012-08-28 16:46:27 <jgarzik> TD: no, will search wiki for oracles...
 850 2012-08-28 16:46:28 <TD> jgarzik: if the thing you are "predicting" can be measured by software, you can potentially do it that way. there are examples of predictions for things like currency prices, sports results, etc in there
 851 2012-08-28 16:46:32 <TD> jgarzik:  here it is
 852 2012-08-28 16:46:46 <TD> https://en.bitcoin.it/wiki/Contracts#Example_4:_Using_external_state
 853 2012-08-28 16:47:09 <devrandom> gavinandresen: perhaps try to see what's holding that mount point?  (fuser, lsof).  The culprit in my experience was crond, but I excluded it in that commit.
 854 2012-08-28 16:47:13 <TD> jgarzik: the basic concept is you embed a hash of a program which yields an address/output-script, into an output. the oracle runs the script you provide and signs the given tx if it matches the outputs
 855 2012-08-28 16:47:43 <TD> jgarzik: so you make the decision of whether the prediction came true automatic, in many cases. you can have a quorum of oracles and an oracle can be challenged with garbage requests, but it doesn't know which requests are real vs fake
 856 2012-08-28 16:47:48 <TD> so you can increase your trust in it that way
 857 2012-08-28 16:47:50 <jgarzik> TD: typically a prediction market needs a quantifiable answer.  i.e. not "who was elected President?" but "Barack Obama received more electoral college votes than Mitt Romney"   That is something that a computer may transmit, but somewhere you need humans feeding info into the system.
 858 2012-08-28 16:48:03 <TD> well you need to find a web page that will contain that information
 859 2012-08-28 16:48:10 <TD> and then have a program that scrapes it
 860 2012-08-28 16:48:24 t7_ has joined
 861 2012-08-28 16:48:26 <TD> often there will be such pages
 862 2012-08-28 16:48:31 <jgarzik> yes
 863 2012-08-28 16:48:37 <gavinandresen> devrandom: "see what's holding the mount point" means... doing what?  On the virtual machine that's being created?  How do I do that?
 864 2012-08-28 16:48:40 <TD> if the API exposed to the oracle program contains download_url or something like that, you can get quite far
 865 2012-08-28 16:48:51 <TD> gavinandresen: lsof
 866 2012-08-28 16:48:56 <TD> "list open files
 867 2012-08-28 16:48:58 <TD> "
 868 2012-08-28 16:49:13 <gavinandresen> TD: yes, but do I run lsof on the VM or on the host?  if on the VM, how?
 869 2012-08-28 16:49:30 <BlueMatt> host
 870 2012-08-28 16:49:33 * TD has no context, that's just how he finds what's holding mount points :-)
 871 2012-08-28 16:49:42 <BlueMatt> (Im assuming you are still talking about umounting /tmp)
 872 2012-08-28 16:49:43 <sipa> TD: it's inside the gitian VM
 873 2012-08-28 16:50:00 <BlueMatt> (in which case - its a ghost, lsof provides nothing of use, I have yet to figure out wtf is going on with it, I always umount -f and try again)
 874 2012-08-28 16:50:02 t7 has quit (Ping timeout: 245 seconds)
 875 2012-08-28 16:50:06 <BlueMatt> (it usually works then)
 876 2012-08-28 16:50:34 <gavinandresen> devrandom: I need to finish debugging the rc1 mac build, then I'll go back to the gitian build....
 877 2012-08-28 16:50:37 <devrandom> gavinandresen: I thought you are having an issue during vmbuilder
 878 2012-08-28 16:50:43 <devrandom> ok
 879 2012-08-28 16:50:44 <BlueMatt> I believe he is
 880 2012-08-28 16:51:02 ThomasV has joined
 881 2012-08-28 16:51:02 <gavinandresen> devrandom: issue is/was with make_base_vm
 882 2012-08-28 16:51:22 <devrandom> gavinandresen: for later - vmbuilder doesn't actually create a vm, it just does chroot
 883 2012-08-28 16:51:27 t7_ has left ()
 884 2012-08-28 16:52:04 <devrandom> the issue was that debbootstrap inside the chroot was starting crond on install, and that was preventing the loopback device from being unmounted
 885 2012-08-28 16:52:15 <Joric> still no coin control in rc1? joric's sad
 886 2012-08-28 16:52:26 <TD> jgarzik: i think we'll find more and more questions that can be answered via third party websites. google is gonna get a lot better at question answering in future
 887 2012-08-28 16:52:37 <TD> so to me automated oracles for prediction markets sound good
 888 2012-08-28 16:52:40 <devrandom> gavinandresen: so I excluded crond, but you still seem to be having an issue.  Either crond is still installed, or some other package is starting a daemon.
 889 2012-08-28 16:52:55 <TD> but yes it's basically automating something humans can also do
 890 2012-08-28 16:53:07 <sipa> Joric: there's the raw transaction interface, which is a lot more flexible
 891 2012-08-28 16:53:12 <devrandom> gavinandresen: in any case, this is all within the host
 892 2012-08-28 16:53:12 <sipa> Joric: RPC only, though
 893 2012-08-28 16:53:20 <gavinandresen> devrandom: ACK
 894 2012-08-28 16:53:23 <sipa> coincontrol was unmaintained
 895 2012-08-28 16:53:31 rdponticelli has joined
 896 2012-08-28 16:54:26 <gmaxwell> Joric: Perhaps you should have not ignored the many times muliple people said that if someone didn't step up to maintain it, it wouldn't go in.
 897 2012-08-28 16:54:44 <TD> jgarzik: i think oracles have a lot more uses than gambling, btw. if you implemented that infrastructure, i doubt there'd be any concerns
 898 2012-08-28 16:54:58 <TD> but if you want to be a market operator for the fees, then sure :)
 899 2012-08-28 16:54:58 <gmaxwell> Joric: I merged the small part of it which I'm willing to maintain; which along with the raw txn interface covers most of the usecases, if not in the most user friendly way.
 900 2012-08-28 16:56:36 <jgarzik> TD: Just read it...  the problem of collusion and cheating is largely hand-waved away, I think.  "People can, and should, regularly challenge the oracle in an automated fashion to ensure it always outputs what is expected."  In practice, nobody is really motivated to do such grunt-work.  They just want to use the service...  until rumor goes around it's not trusted, at which point they just stop using it.
 901 2012-08-28 16:56:59 <jgarzik> people might have some incentive to constantly test the Oracle, but in practice they won't
 902 2012-08-28 16:57:05 <TD> disagree. you can just have it a part of whatever software you're using
 903 2012-08-28 16:57:06 <sipa> gavinandresen: any place/way in unit tests to get a nice corpus of valid and invalid signatures? i'm writing a unit test for strict DER-encoding tests
 904 2012-08-28 16:57:15 <TD> at any rate, at least you _can_ test an oracle
 905 2012-08-28 16:57:17 <TD> unlike a human
 906 2012-08-28 16:57:24 <jgarzik> and some conditions - - scraping a presidential election website -- cannot be tested in practice until the very moment that the Oracle's answers must be correct.
 907 2012-08-28 16:57:48 <TD> bear in mind, you provide the program to run and the transaction to sign
 908 2012-08-28 16:57:51 <BlueMatt> sipa: please put them in data-driven transactions
 909 2012-08-28 16:57:53 <BlueMatt> form
 910 2012-08-28 16:57:53 <TD> but neither has to be actually a real bet
 911 2012-08-28 16:58:02 <TD> (or inheritance locked tx or whatever)
 912 2012-08-28 16:58:19 <BlueMatt> sipa: (obv would have to be a new one, as the current ones are valid/invalid)
 913 2012-08-28 16:58:21 pnicholson has quit (Quit: pnicholson)
 914 2012-08-28 16:58:23 <TD> so when setting up the bet, you can craft two or more programs and use them in the fake transactions. you can check that the behaviour of the oracles are what you expect
 915 2012-08-28 16:58:32 <TD> ie, the output is as they should be given the inputs
 916 2012-08-28 16:58:58 ErnestoJuarell has joined
 917 2012-08-28 16:59:29 <sipa> BlueMatt: my question was more whether I could reuse existing ones :)
 918 2012-08-28 16:59:52 <gmaxwell> (in fact, the most obvious construction of an oracle has private keys based on the hash of the program and an oracle's secret, so the oracle need not store anything per user)
 919 2012-08-28 16:59:57 <BlueMatt> sipa: oh, well then grab them from the existing data-driven tx tests
 920 2012-08-28 17:00:11 <BlueMatt> (I think there is one just normal valid sig-checking tx)
 921 2012-08-28 17:00:53 <TD> gmaxwell: what does mixing in the program hash to the private key give you
 922 2012-08-28 17:01:24 <jgarzik> TD: ...but testing is impossible until the event itself
 923 2012-08-28 17:01:36 <sipa> BlueMatt: i guess i'll make new ones; i need signatures, not scripts
 924 2012-08-28 17:01:46 <BlueMatt> those are full txes
 925 2012-08-28 17:01:46 <jgarzik> TD: at which point, the oracle -must- produce valid data
 926 2012-08-28 17:01:48 <BlueMatt> incl sigs
 927 2012-08-28 17:01:56 <TD> jgarzik: no. all you're testing is program yields the correct output given its inputs
 928 2012-08-28 17:02:06 <BlueMatt> sipa: the tx_valid.json/tx_invalid.json not script_*.json
 929 2012-08-28 17:02:06 <sipa> BlueMatt: not really useful, unless you have a guaranteed way of extracting signatures from them
 930 2012-08-28 17:02:15 <TD> jgarzik: you can test it with any website, any data
 931 2012-08-28 17:02:16 <BlueMatt> they are just serialized txes
 932 2012-08-28 17:02:18 <jgarzik> TD: and the input is unpredictable until the event, in many cases
 933 2012-08-28 17:02:33 <Joric> what if i only want one address, -keypool=0 ?
 934 2012-08-28 17:02:39 <jgarzik> it just does not seem realistically doable
 935 2012-08-28 17:02:44 <BlueMatt> sipa: they are outputs + serialized txn...
 936 2012-08-28 17:03:03 <gmaxwell> TD: Because it lets you have a oracle with ~no storage. E.g. say the oracle is implemented in secure hardware like an IBM 4764. There is no real room for persistant secure storage. The oracle needs some way of remembering funds sent to 1program1 are encoumbered by the rules embodied in 1program1.
 937 2012-08-28 17:03:11 <BlueMatt> sipa: and I suggest putting it in data-driven form because other clients will want to use the same tests
 938 2012-08-28 17:03:18 <sipa> BlueMatt: of course
 939 2012-08-28 17:03:35 <BlueMatt> and its a bigger pain if we have 100 different data-driven formats
 940 2012-08-28 17:03:49 <BlueMatt> sipa: whats your hash on boost-win32-1.50.0-gitian2.zip ?
 941 2012-08-28 17:04:19 <gmaxwell> TD: the simplest way to get that is for the private key for 1program1  to be derrived from H(H(program1)||oraclesecret) prior to the execution of program1. So you send your script with no input and it returns the public key. Later you send your script again with the stuff to sign and evidence required, and it does so if the program approves.
 942 2012-08-28 17:04:22 <TD> gmaxwell: oh, i see. so it's just instead of putting the program hash into the transaction to sign
 943 2012-08-28 17:04:32 <TD> gmaxwell: it seems more explicit and simpler to put the hash into the transaction output and drop it
 944 2012-08-28 17:04:49 <sipa> BlueMatt: what kind of hash do you want? :)
 945 2012-08-28 17:05:03 <BlueMatt> sipa: any? whats gitian's default, sha256?
 946 2012-08-28 17:05:20 <gmaxwell> TD: hm. that seems ugly to me. It bloats the transaction and you're not even sure if the oracle is willing to run the program.
 947 2012-08-28 17:05:34 <BlueMatt> sipa: 393cdf5779c556ce09689bacaa2a1083a47b6d7bc247ed8fb2cd3aa94cb0fa80  boost-win32-1.50.0-gitian2.zip look farmiliar?
 948 2012-08-28 17:05:36 <sipa> $ sha256sum boost-win32-1.50.0-gitian2.zip
 949 2012-08-28 17:05:37 <sipa> d25ddb68061c11ee0a3fc40d3e6dec2b7b5216ffc3ba0b4de947a7fb04a3900e  boost-win32-1.50.0-gitian2.zip
 950 2012-08-28 17:05:43 <BlueMatt> damn
 951 2012-08-28 17:05:48 <sipa> but that's not a problem
 952 2012-08-28 17:06:01 <BlueMatt> is it not in this case?
 953 2012-08-28 17:06:15 <sipa> wumpus and I have matching hashes for the builds
 954 2012-08-28 17:06:17 <sipa> completely
 955 2012-08-28 17:06:24 <BlueMatt> but not boost?
 956 2012-08-28 17:06:55 <sipa> don't think so
 957 2012-08-28 17:07:03 <BlueMatt> ok, Ill try the build then
 958 2012-08-28 17:07:18 <sipa> we both used a VM base image that was created yesterday
 959 2012-08-28 17:07:24 <BlueMatt> same
 960 2012-08-28 17:07:28 <BlueMatt> well, today
 961 2012-08-28 17:07:32 <sipa> before, he used an older base, and the hashes for the linux build were different
 962 2012-08-28 17:07:32 <TD> gmaxwell: it bloats it by one hash, hardly a big deal. the transaction is only broadcast if the oracle signs it, so if it refuses to run the program the block chain isn't touched
 963 2012-08-28 17:07:51 <BlueMatt> sipa: hmm..odd, it updates all packages at start...
 964 2012-08-28 17:08:04 <gmaxwell> TD: er yes, but your coins are now stuck forever.
 965 2012-08-28 17:08:40 <gmaxwell> TD: whereas if you consult the oracle to get the applicable address it can give you a promise to perform; or at least tell you it certantly won't perform if it knows.
 966 2012-08-28 17:09:07 balrog has quit (Quit: Bye)
 967 2012-08-28 17:09:08 <TD> oh, sure. but you could just as easily challenge the oracle with the same program
 968 2012-08-28 17:09:18 <TD> then you don't even have to ask if it "would" run it, you can make it actually run it and check it works
 969 2012-08-28 17:10:31 <gmaxwell> TD: sure sure, but if you've done that you're still talking to it in advance. So getting a key that way costs nothing. I can only see upsides to this: smaller transactions, improved privacy (the oracle locked transactions aren't distinctive in the chain). Funny though, I thought doing it that way was just the most obvious thing to do.
 970 2012-08-28 17:10:45 <eian> I want to implement IP prefix matching (CIDR lookups).  Is there a library that I can use to create a trie structure? I don't see one in boost...
 971 2012-08-28 17:12:58 <BlueMatt> sipa: sigs match, gavinandresen we have 3 sigs, can we get an upload?
 972 2012-08-28 17:12:58 <BlueMatt> or tcatm
 973 2012-08-28 17:12:58 <sipa> BlueMatt: already uploaded
 974 2012-08-28 17:12:59 <BlueMatt> oh...who uploaded before we had 3 sigs?
 975 2012-08-28 17:13:05 <sipa> gavin is still busy with the OSX build
 976 2012-08-28 17:13:11 <sipa> BlueMatt: i did
 977 2012-08-28 17:13:14 <BlueMatt> :(
 978 2012-08-28 17:13:27 <gmaxwell> BlueMatt: it's an RC at least.
 979 2012-08-28 17:13:40 <BlueMatt> still...
 980 2012-08-28 17:14:03 <TD> i think the most obvious way is usually to make your data structure members explicit :) merging them together would be an optimization that increase the complexity of the protocol, though yes, the privacy gain may be useful
 981 2012-08-28 17:14:16 <TD> you'd still need to connect these transactions to 2-of-3 multisigs in the case of a bet though
 982 2012-08-28 17:14:30 <TD> so, the transactions would end up being large-ish anyway
 983 2012-08-28 17:14:36 <eian> I have 25,000 ips associated with the bitcoin network - I want to look up country codes. I don't want to hammer an online service with potentially bazillions of lookups so I'm directly grabbing BGP info from ARIN/RIPE/etc.
 984 2012-08-28 17:14:52 <sipa> eian: interesting
 985 2012-08-28 17:14:54 <TD> eian: there are free ipgeo databases
 986 2012-08-28 17:14:58 <TD> eian: debian has one you can apt-get
 987 2012-08-28 17:15:00 <gmaxwell> yea, maxmind, etc.
 988 2012-08-28 17:15:15 <eian> TD: where are they getting their information from? Can I trust the results?
 989 2012-08-28 17:15:34 <gmaxwell> eian: You certantly can't trust results from mapping prefixes to ASNs.
 990 2012-08-28 17:15:51 <gmaxwell> Since networks routinely have ASNs spanning geographies.
 991 2012-08-28 17:16:21 <eian> I guess it is the best I have - I don't know where these online databases are getting this info.
 992 2012-08-28 17:16:29 <gmaxwell> eian: see http://www.maxmind.com/ they have basic geodata available for free.
 993 2012-08-28 17:16:40 <TD> i don't know where they get the data from
 994 2012-08-28 17:16:41 <gmaxwell> (and they document how they get their onfo)
 995 2012-08-28 17:16:47 <kjj_> eian: it is better than random, but don't use them for missile targeting
 996 2012-08-28 17:16:49 <gmaxwell> (or at least they used to)
 997 2012-08-28 17:17:19 <eian> kjj, fire ze missiles
 998 2012-08-28 17:17:33 <TD> eian: justmoon has a webgl globe that does this
 999 2012-08-28 17:17:41 <BlueMatt> eian: maxmind is pretty much an industry standard for geoip afaik
1000 2012-08-28 17:17:49 <BlueMatt> not /the/ but /an/
1001 2012-08-28 17:17:51 <gmaxwell> eian: I mean, I can tell you with absolute certanty that trying to grok out node location from arin ASN registrations is a waste of time. It will have very low accuracy.
1002 2012-08-28 17:18:20 <eian> gmaxwell: :(
1003 2012-08-28 17:18:31 <eian> gmaxwell: well that was a waste of my energy
1004 2012-08-28 17:18:55 <eian> I just needed a trie structure and I would have been finished lol
1005 2012-08-28 17:18:59 <gmaxwell> eian: well presumably you learned something useful in the process? :)
1006 2012-08-28 17:19:38 maaku has joined
1007 2012-08-28 17:20:49 <eian> "We arrive at the accuracy figure by checking known web user IP address and location pairs against the GeoLite City database."
1008 2012-08-28 17:20:56 <eian> i.e., trust us - we know what we are doing
1009 2012-08-28 17:21:21 <jgarzik> eian: free ip-to-country database: http://software77.net/geo-ip/
1010 2012-08-28 17:21:26 <jgarzik> eian: snarfed from ARIN etc.
1011 2012-08-28 17:21:45 <jgarzik> maxmind has a free db in most linux distros, though
1012 2012-08-28 17:22:07 <gmaxwell> eian: http://arxiv.org/pdf/1005.5674
1013 2012-08-28 17:22:46 <eian> gmaxwell: interesting, thanks - I'll take a look
1014 2012-08-28 17:23:20 RedEmerald has quit (Ping timeout: 276 seconds)
1015 2012-08-28 17:24:19 <eian> gmaxwell: maybe as an interesting test, I can finish up this tool and compare my lookups with these other databases
1016 2012-08-28 17:24:49 <eian> gmaxwell: I could tell the world exactly how stupid my approach is :)
1017 2012-08-28 17:25:10 <gmaxwell> eian: sounds good, if you just want longest matches, don't bother with the trie. Just split the prefixes by match length — 32 tables. Then do 1-32 exact match lookups and stop when you get a hit.
1018 2012-08-28 17:25:30 <gmaxwell> It's less memory efficient than a prefix trie for longest match lookups, but its ~as fast.
1019 2012-08-28 17:25:32 <TD> devrandom: think you could quickly review the latest autosave code today?
1020 2012-08-28 17:25:46 <jgarzik> Smart property is cute.  I am tempted to buy one of those $100 houses in Detroit, MI and sell it for bitcoins.
1021 2012-08-28 17:25:48 <gmaxwell> (thats also how TCAM based L3 switches, which can't do trie lookups, do routing)
1022 2012-08-28 17:25:57 <TD> lol
1023 2012-08-28 17:26:07 <TD> jgarzik: start with ssh keys ;)
1024 2012-08-28 17:26:24 <jgarzik> TD: ?
1025 2012-08-28 17:26:42 <TD> i mean, if you want to do smart property, shell accounts seem like an easier place to start than houses :-)
1026 2012-08-28 17:27:22 <jgarzik> heh
1027 2012-08-28 17:27:27 <jgarzik> but not as much fun
1028 2012-08-28 17:27:48 <eian> gmaxwell: if I stop at the first hit, it may not be the longest? (Maybe I am misunderstanding)
1029 2012-08-28 17:27:53 <jgarzik> and it would be amusing if bitcoiners suddenly owned large swaths of shitty houses in Detroit
1030 2012-08-28 17:28:05 <eian> oh wait, you mean to traverse from longest to shortest - duh
1031 2012-08-28 17:28:08 iocor has joined
1032 2012-08-28 17:28:14 <kjj_> too bad the armory closed.  now we can never visit our houses
1033 2012-08-28 17:28:25 <gmaxwell> eian: you start at /32. You have 32 tables. One with all the /32 routes, one with all the /31 routes, one with all the /30...
1034 2012-08-28 17:28:37 <eian> right, I'm slow
1035 2012-08-28 17:28:39 <eian> thanks
1036 2012-08-28 17:28:55 RedEmerald has joined
1037 2012-08-28 17:30:31 <jgarzik> anyway, on oracles, wolfram alpha and google are getting close
1038 2012-08-28 17:31:41 <TD> yeah though i suspect the most common kinds of questions can be phrased as queries to various APIs. eg, if you want to bet on the price of a commodity or currency
1039 2012-08-28 17:32:03 <TD> but yes. google will be answering more and more questions directly in future, i think
1040 2012-08-28 17:32:04 <gmaxwell> but then the security reduces to that of the api.. kinda meh.
1041 2012-08-28 17:32:16 <TD> sure. but would you rather depend on that, or a human
1042 2012-08-28 17:32:16 <kjj_> if you want to do oracle based transactions, make sure they are M of N, and that the oracles don't update their databases to match each other
1043 2012-08-28 17:32:44 <gmaxwell> TD: a single human? well. Right, but a multisign?  Or even just an oracle enforcing a massive N of M vote?
1044 2012-08-28 17:33:07 <TD> you have a multi-sign of oracles of course :-)
1045 2012-08-28 17:33:20 <TD> for huge multi-signatures we'd need a real implementation of threshold signatures
1046 2012-08-28 17:33:26 <TD> unless you're ok with paying big fees
1047 2012-08-28 17:33:29 <gmaxwell> yea, but if they're all calling google, thats bleh.
1048 2012-08-28 17:33:48 <gmaxwell> TD: you can do huge multisign using e.g. three oracles doing a the tallying.
1049 2012-08-28 17:34:10 <TD> gmaxwell: well, if all 3 check 3 different sources?
1050 2012-08-28 17:34:27 <TD> yes, but there are cryptographic techniques that let you have a single ECDSA signature that can only be calculated by N participants
1051 2012-08-28 17:34:29 Cablesaurus has quit (Quit: Give a man a fish and he will eat for a day. Teach him how to fish, and he will sit in a boat and drink beer all day)
1052 2012-08-28 17:34:35 <gmaxwell> are the 'sources' just http sites? boring.
1053 2012-08-28 17:34:37 <TD> or rather an n-of-m threshold
1054 2012-08-28 17:34:49 <TD> it's basically CHECKMULTISIG but with fancy maths instead of just trying each combination ...
1055 2012-08-28 17:35:18 <gmaxwell> TD: I'm aware, but I've still not seen how it's possible with ECDSA (I know how it's possible with other signature schemes)
1056 2012-08-28 17:35:38 balrog has joined
1057 2012-08-28 17:36:00 <TD> there are papers that claim it can be done
1058 2012-08-28 17:36:01 <eian> Stupid question: Is there a fast way to plot a simple line graph with a billion points?
1059 2012-08-28 17:36:06 <TD> i never understood exactly how :-)
1060 2012-08-28 17:36:37 <gmaxwell> eian: shuf -n 20000 points | sort > input ; gnuplot ; plot 'input'
1061 2012-08-28 17:36:43 <TD> oh, interesting
1062 2012-08-28 17:36:44 <jgarzik> yeah, for distributed, I was thinking a multi-sign of human oracles at the first step.  Just don't trust automated programs to answer direct queries, because even distributed software oracles would wind up reducing trust to that of the common API they query (goog, wolf, ...)
1063 2012-08-28 17:36:47 <TD> speaking of which: https://bitcointalk.org/index.php?topic=104086.0
1064 2012-08-28 17:36:48 <jgarzik> as gmaxwell just noted
1065 2012-08-28 17:37:11 <gmaxwell> TD: "meh"
1066 2012-08-28 17:37:15 <eian> gmaxwell: you are a wellspring of information
1067 2012-08-28 17:37:24 <eian> gmaxwell: I will call you "The Google"
1068 2012-08-28 17:37:29 toffoo has joined
1069 2012-08-28 17:37:55 <jgarzik> I rather like the N distributed human oracles approach, but collusion is a problem.
1070 2012-08-28 17:38:11 <kjj_> jgarzik: double blind can help that
1071 2012-08-28 17:38:57 <kjj_> the humans don't know who is asking or why, and half (or more) of the questions are fake.  the askers don't know who they are asking.
1072 2012-08-28 17:38:59 <TD> jgarzik: the point of automating it is humans can have all kinds of other pressures or desires on them, whereas a program (if it can't tell which challenges are real/fake) will just do what it's told each time
1073 2012-08-28 17:39:15 <TD> and yeah it means you have to trust the data provider. but so what? the human will eventually consult a data provider to make its decision too
1074 2012-08-28 17:39:29 <TD> unless it's something like "who won the election" where you can assume they check multiple sources, maybe. but a program can do that too
1075 2012-08-28 17:39:39 <devrandom> TD: yes
1076 2012-08-28 17:42:13 asa has quit (Quit:  HydraIRC -> http://www.hydrairc.com <- Now with extra fish!)
1077 2012-08-28 17:42:25 <gmaxwell> jgarzik: not just that, but the oracle case is just a superset. AskAHuman()  can be a function in the program too.
1078 2012-08-28 17:43:02 dvide has joined
1079 2012-08-28 17:44:27 <jgarzik> well you still have to trust the human running the oracle
1080 2012-08-28 17:44:37 <jgarzik> in addition to the data provider
1081 2012-08-28 17:44:58 <TD> the point of the challenge protocol is to reduce that trust
1082 2012-08-28 17:45:14 <TD> as the wiki points out, you could use trusted computing as well
1083 2012-08-28 17:45:17 <jgarzik> maybe oracle software would make for a great StorJ-like agent ;-)
1084 2012-08-28 17:45:22 <TD> if the oracle runs on a platform that supports it :-)
1085 2012-08-28 17:45:26 <TD> right. time to go.
1086 2012-08-28 17:45:35 TD has quit (Quit: TD)
1087 2012-08-28 17:45:46 <devrandom> TD: I already +1'ed commit 108c23 ... was there something else?
1088 2012-08-28 17:45:49 cdecker has joined
1089 2012-08-28 17:46:37 <devrandom> TD[gone]: ^^
1090 2012-08-28 17:47:01 <gmaxwell> This is the obvious base hardware for an oracle. http://www.ebay.com/itm/IBM-PCI-X-CRYPTOGRAPHIC-COPROCESSOR-FC-4764-12R8566-12R8565-12R8561-/120922986008?pt=LH_DefaultDomain_0&hash=item1c27925618
1091 2012-08-28 17:48:03 <gmaxwell> It's is a highly tamper resistant PPC with memory and a crypto accelerator on a card, and a operating system that lets it securely do remote attestation to prove the software its running (upto the tamper resistance), tractable back to IBM's certificates.
1092 2012-08-28 17:48:59 <gmaxwell> so an oracle constructed using that is secure unless the operator is evil _and_ conspires with IBM _or_ has figured out how to tamper with the hardware.
1093 2012-08-28 17:51:23 Cablesaurus has joined
1094 2012-08-28 17:51:32 Cablesaurus has quit (Changing host)
1095 2012-08-28 17:51:32 Cablesaurus has joined
1096 2012-08-28 17:51:44 <kjj_> gmaxwell: that coprocessor is still vulnerable to what I'll call the "Dewey Defeats Truman" attack.  All oracle schemes are
1097 2012-08-28 17:51:54 <gavinandresen> mac build uploaded....  macdeployqtplus is still broken, though, any applescript experts here?
1098 2012-08-28 17:52:22 <gmaxwell> kjj_: Only for oracles depending on daft services, but there are lots of interesting things where the evaluation uses only provable inputs.
1099 2012-08-28 17:52:31 <gmaxwell> (e.g. tallys for massive multisig votes)
1100 2012-08-28 17:53:31 <gmaxwell> kjj_: you could code around "Dewey Defeats Truman", e.g. require evidence from N newspapers from N distinct days. Though good luck writing a safe parser for that. :)
1101 2012-08-28 17:53:58 <gmaxwell> Doesn't help that SSL's proof isn't extensible to third parties. :(
1102 2012-08-28 17:54:29 <kjj_> gmaxwell: exactly.  and what if all of the papers ran with the story and it took >N days to sort out?
1103 2012-08-28 17:56:57 <gmaxwell> kjj_: You can always set your N high enough to get acceptable security. Bitcoin itself is only secure to episilon risk of failure. Nothing in life is certan, you might win but get hit by a falling satellite too. :)
1104 2012-08-28 18:00:52 paraipan has quit (Remote host closed the connection)
1105 2012-08-28 18:01:05 da2ce7 has quit (Remote host closed the connection)
1106 2012-08-28 18:02:15 iocor has quit (Quit: Textual IRC Client: http://www.textualapp.com/)
1107 2012-08-28 18:04:10 Obsi has joined
1108 2012-08-28 18:04:55 da2ce7 has joined
1109 2012-08-28 18:05:51 B0g4r7__ has joined
1110 2012-08-28 18:07:30 paraipan has joined
1111 2012-08-28 18:09:29 B0g4r7_ has quit (Ping timeout: 276 seconds)
1112 2012-08-28 18:28:35 MC-Eeepc has quit (Ping timeout: 248 seconds)
1113 2012-08-28 18:32:54 sirk390 has joined
1114 2012-08-28 18:33:47 rdponticelli has quit (Ping timeout: 272 seconds)
1115 2012-08-28 18:39:29 pnicholson has joined
1116 2012-08-28 18:53:17 D34TH has joined
1117 2012-08-28 18:53:18 D34TH has quit (Changing host)
1118 2012-08-28 18:53:18 D34TH has joined
1119 2012-08-28 18:56:29 [7] has quit (Disconnected by services)
1120 2012-08-28 18:56:29 TheSeven has joined
1121 2012-08-28 18:58:52 pusle has joined
1122 2012-08-28 19:00:59 cablepair has joined
1123 2012-08-28 19:01:40 sebicas has left ()
1124 2012-08-28 19:03:35 <gmaxwell> We have a long way to go: https://people.xiph.org/~greg/bitcoin_coverage/coverage/home/gmaxwell/src/bcm/bax/src/index.html  (this is just from test_bitcoin)
1125 2012-08-28 19:07:35 <gmaxwell> Luke-Jr: seems your block tests don't test the sigops count limit for p2sh.
1126 2012-08-28 19:08:04 <gmaxwell> ( https://people.xiph.org/~greg/bitcoin_coverage/coverage/home/gmaxwell/src/bcm/bax/src/main.cpp.gcov.html line 3667 )
1127 2012-08-28 19:08:06 <Luke-Jr> :|
1128 2012-08-28 19:09:29 <gmaxwell> hey, well, it's still one of the better tested functions in the codebase.
1129 2012-08-28 19:12:51 slush1 has quit (Ping timeout: 248 seconds)
1130 2012-08-28 19:15:09 <gmaxwell> gavinandresen: er, you let the doomsday alrert out on testnet
1131 2012-08-28 19:15:22 <gmaxwell> s/alrert/alert/
1132 2012-08-28 19:15:24 <gavinandresen> gmaxwell: what?
1133 2012-08-28 19:15:24 <gmaxwell>     "errors" : "Warning: This version is obsolete, upgrade required!"
1134 2012-08-28 19:15:28 PhantomSpark has quit (2!~kvirc@pool-71-251-16-25.nycmny.fios.verizon.net|Ping timeout: 246 seconds)
1135 2012-08-28 19:15:45 <gavinandresen> gmaxwell: grr....
1136 2012-08-28 19:16:41 <gavinandresen> Wait, that's not the doomsday alert
1137 2012-08-28 19:17:10 <gmaxwell> what the heck is that then?
1138 2012-08-28 19:17:46 <gmaxwell>     // Check the version of the last 100 blocks to see if we need to upgrade:
1139 2012-08-28 19:17:54 <gavinandresen> yes, that's what it is
1140 2012-08-28 19:18:03 datagutt has quit (Quit: kthxbai)
1141 2012-08-28 19:18:04 minimoose has quit (Quit: minimoose)
1142 2012-08-28 19:18:08 <gmaxwell> hurray for testing.
1143 2012-08-28 19:18:12 <gavinandresen> Somebody mined a bunch of upversion blocks
1144 2012-08-28 19:18:49 gravypod has joined
1145 2012-08-28 19:18:55 <gmaxwell> should I just reorg them out of the chain?
1146 2012-08-28 19:18:59 gravypod has quit (Read error: Connection reset by peer)
1147 2012-08-28 19:19:14 <gavinandresen> gmaxwell: probably the right thing to do, yes
1148 2012-08-28 19:19:57 <gmaxwell> hm. how the heck can I get the block versions via the rpc? (I need to know where to cut)
1149 2012-08-28 19:19:59 <gavinandresen> Luke-Jr : you weren't using the high-nibble of the block version number on testnet, were you?
1150 2012-08-28 19:20:11 <gavinandresen> gmaxwell: getblock doesn't tell you?
1151 2012-08-28 19:20:36 <Luke-Jr> gavinandresen: no
1152 2012-08-28 19:21:01 davout has joined
1153 2012-08-28 19:21:01 davout has quit (Changing host)
1154 2012-08-28 19:21:02 davout has joined
1155 2012-08-28 19:21:30 pnicholson has quit (Quit: pnicholson)
1156 2012-08-28 19:21:33 <Luke-Jr> although I think someone mentioned doing it on mainnet a few weeks ago
1157 2012-08-28 19:21:58 <gmaxwell> gavinandresen: they're all just version 2.
1158 2012-08-28 19:22:02 <Luke-Jr> maybe instead of reorging testnet, we should make sure the error goes away after 100 blocks?
1159 2012-08-28 19:22:32 <gmaxwell> main.h:    static const int CURRENT_VERSION=1;
1160 2012-08-28 19:22:37 <Luke-Jr> (and then reorg it :P)
1161 2012-08-28 19:23:15 <gavinandresen> gmaxwell: you running git HEAD?
1162 2012-08-28 19:23:15 wizkid057 has joined
1163 2012-08-28 19:23:29 <gmaxwell> oh darn right this node might not be, sorry for the goose chase.
1164 2012-08-28 19:23:43 <gmaxwell> I have so many nodes, I lose track. :(
1165 2012-08-28 19:24:37 <gmaxwell> yea, gone with head.
1166 2012-08-28 19:24:45 <gmaxwell> good. It did what it was supposted to do.
1167 2012-08-28 19:25:19 <gavinandresen> phew.  releasing the doomsday alert on testnet would be bad... (I think somebody could copy it to main net....)
1168 2012-08-28 19:25:57 <gmaxwell> gavinandresen: lets change the testnet alert key.
1169 2012-08-28 19:26:10 <gmaxwell> And end that problem forever.
1170 2012-08-28 19:26:22 <gavinandresen> gmaxwell: yeah, I generated one months ago and promptly forgot to commit that change
1171 2012-08-28 19:26:31 <gmaxwell> Now is an excellent time.
1172 2012-08-28 19:27:28 <gavinandresen> Right-after-0.7-is-final is a more excellent time, but I'll create a PULL to pull after 0.7 is final.
1173 2012-08-28 19:27:35 Joric has quit ()
1174 2012-08-28 19:28:06 <gmaxwell> hm? why not put it in 0.7? that way ~all testnet3 nodes will end up with the same key.
1175 2012-08-28 19:28:59 <gavinandresen> gmaxwell: you're assuming we'll have a rc2
1176 2012-08-28 19:29:09 <gmaxwell> oh, fair point.
1177 2012-08-28 19:29:09 rdponticelli has joined
1178 2012-08-28 19:29:42 <gavinandresen> but yeah, if we do have a rc2 then I think it'd be OK to pull.
1179 2012-08-28 19:32:57 <gmaxwell> Coverage reports are a bit better if I include a testnet sync as well as test_bitcoin: file:///home/gmaxwell/src/bcm/bax/src/coverage2/home/gmaxwell/src/bcm/bax/src/index.html
1180 2012-08-28 19:33:40 ErnestoJuarell has quit (Quit: Page closed)
1181 2012-08-28 19:35:39 <gavinandresen> oooh, you tricked me into clicking on that file:/// link....
1182 2012-08-28 19:36:01 <gmaxwell> hahah
1183 2012-08-28 19:36:02 <gmaxwell> sorry
1184 2012-08-28 19:36:16 <gmaxwell> https://people.xiph.org/~greg/bitcoin_coverage/coverage2/home/gmaxwell/src/bcm/bax/src/index.html
1185 2012-08-28 19:36:20 <ersi> Score++ for gmaxwell
1186 2012-08-28 19:36:50 <gmaxwell> I OWNING YOUR HOST. Er.
1187 2012-08-28 19:37:39 <gavinandresen> gmaxwell: I'm looking at (for example) version.cpp coverage info; what does [+ - ]  mean?
1188 2012-08-28 19:38:13 <gavinandresen> (and why doesn't version.cpp get 100% coverage, it's nothing but preprocessor stuff and constants)
1189 2012-08-28 19:38:39 <gavinandresen> (err, 100% branch coverage....)
1190 2012-08-28 19:38:54 <gmaxwell> So, if you have a branch, e.g.  if (a)  {} there are two possible outcomes. True and false. So thats + and -.   if you ahve if(a&&b) there are two branches each with their own +/- states.
1191 2012-08-28 19:39:20 <gmaxwell> We'll never get close to 100% branch in our codebase because there are 'hidden' branches in much of the C++ muck.
1192 2012-08-28 19:39:23 * gmaxwell looks at version.cpp
1193 2012-08-28 19:39:47 <gavinandresen> something weird like std::string can be either narrow or wide maybe?
1194 2012-08-28 19:40:35 <gmaxwell> my WAG would be std::string allocates heap and tests for failure, throws an exception if it fails, and the branch is that tes.t
1195 2012-08-28 19:40:53 <gmaxwell> but I'm a little C++ dumb, as I like to point out.
1196 2012-08-28 19:41:07 <Luke-Jr> would you guys want something like this for contrib? http://codepad.org/qIXQVu2U
1197 2012-08-28 19:42:05 <gavinandresen> ooh!  Luke tricked me into clicking on a mess of impenetrable code....
1198 2012-08-28 19:42:32 <BlueMatt> gmaxwell: any chance you could check coverage when running the bitcoinj block tester too?
1199 2012-08-28 19:42:58 <BlueMatt> gmaxwell: apply patch at http://jenkins.bluematt.me/pull-tester/files/ and run .jar
1200 2012-08-28 19:43:51 <sipa> Luke-Jr: hm, what is that?
1201 2012-08-28 19:44:42 <Luke-Jr> sipa: aids is adjusting translations to changes in the English inputs
1202 2012-08-28 19:44:45 <Luke-Jr> in*
1203 2012-08-28 19:45:06 <Luke-Jr> (and merging translations between branches)
1204 2012-08-28 19:45:16 <sipa> ic; know noting about that
1205 2012-08-28 19:45:59 <sipa> gavinandresen: no 0.7.0rc1 announcement yet?
1206 2012-08-28 19:46:00 <gmaxwell> Luke-Jr: is that ... encrypted?
1207 2012-08-28 19:46:04 <Luke-Jr> gmaxwell: no :/
1208 2012-08-28 19:46:40 <gavinandresen> sipa: right, I got distracted from that....
1209 2012-08-28 19:47:15 <gavinandresen> sipa: any reason you put extra stuff in SHASUMS.asc ?  OK if I rewrite/re-sign it with the osx shasum
1210 2012-08-28 19:47:19 <gmaxwell> I shouldn't kid, my pyhton looks horrible compared to that. "Oh cool, I can use lambdas to implement dijkstra in one line!" and never do I think "but wait, I shouldn't"
1211 2012-08-28 19:48:36 <sipa> gavinandresen: i just uploaded the output of my build script, which also creates a list of commits and such
1212 2012-08-28 19:48:44 <sipa> gavinandresen: i can exclude that if you prefer
1213 2012-08-28 19:48:50 <gmaxwell> gavinandresen: hopefully the SNR isn't too high on the coverage reports; there are many useful things there. E.g. we appear to not have testcases for the checkmultisig limits.
1214 2012-08-28 19:49:11 <sipa> gavinandresen: but obviously you can cut out anything you want, and add the OSX hashes
1215 2012-08-28 19:49:23 <gavinandresen> sipa: ACK, doing that now
1216 2012-08-28 19:50:08 <sipa> the idea is that my signature means "I guarantee that the binaries here match this source", so I add+sign information about that source
1217 2012-08-28 19:52:54 tcatm has quit (Quit: No Ping reply in 180 seconds.)
1218 2012-08-28 19:53:12 tcatm has joined
1219 2012-08-28 19:53:12 tcatm has quit (Changing host)
1220 2012-08-28 19:53:12 tcatm has joined
1221 2012-08-28 19:53:18 davout has quit (Remote host closed the connection)
1222 2012-08-28 19:55:20 * Luke-Jr ponders if it might be handy to keep a database of (git commit)-to-(binary hash) somewhere
1223 2012-08-28 19:56:11 p0s has quit (Read error: Connection reset by peer)
1224 2012-08-28 19:56:20 <BlueMatt> jenkins
1225 2012-08-28 19:56:38 <BlueMatt> or...actually no I think it deletes old builds (and I dont see the option to change that :( )
1226 2012-08-28 19:57:21 sirk390 has quit (Quit: Leaving.)
1227 2012-08-28 19:59:51 <gmaxwell> BlueMatt: when your blockthingy is 'done' we can give it nonce values for a real solution so it doesn't need modification.
1228 2012-08-28 19:59:59 <gmaxwell> But yes, I can test that too.
1229 2012-08-28 20:00:34 sirk390 has joined
1230 2012-08-28 20:00:50 <BlueMatt> gmaxwell: yea, give me coverage reports and let me see what else needs checked and then it can be 'done' ;)
1231 2012-08-28 20:01:53 Diapolo has joined
1232 2012-08-28 20:02:06 <gmaxwell> lcov isn't hard to use fwiw, add CFLAGS='--coverage -ftest-coverage -fprofile-arcs -O0' LDFLAGS='--coverage'  when building, run the program then lcov -c -d `pwd` -b `pwd` -t mytest -o mytest.info ; genhtml mytest.info -o coverage/
1233 2012-08-28 20:02:16 <gmaxwell> then to clear the coverage state lcov -d `pwd` -b `pwd` -z
1234 2012-08-28 20:02:19 ovidiusoft has quit (Read error: No route to host)
1235 2012-08-28 20:02:46 <gmaxwell> you can run several tests into seperate info files, then combine them lcov -a one.info -a two.info -a three.info -o all.info  and genhtml on all to composite the tests.
1236 2012-08-28 20:03:30 ovidiusoft has joined
1237 2012-08-28 20:03:44 <gmaxwell> For other projects I work on, I have jenkins run it for every commit; e.g. https://mf4.xiph.org/jenkins/job/opus-coverage/ws/coverage/index.html
1238 2012-08-28 20:04:58 ovidiusoft has quit (Client Quit)
1239 2012-08-28 20:05:02 <BlueMatt> nice, Ill throw it on the never-ending todo list ;)
1240 2012-08-28 20:05:19 ovidiusoft has joined
1241 2012-08-28 20:05:43 <Diapolo> BlueMatt: Is the PullTester currently DoSed by commit amount?
1242 2012-08-28 20:05:53 <BlueMatt> quite possible, give me a sec
1243 2012-08-28 20:06:37 <BlueMatt> nope, it died, sorry, running again
1244 2012-08-28 20:08:38 <gmaxwell> BlueMatt: is the tester suport to only print one block then say disconnected?
1245 2012-08-28 20:08:41 <gavinandresen> 0.7.0rc1 announced: https://bitcointalk.org/index.php?topic=104173  (and on bitcoin-development mailing list)
1246 2012-08-28 20:09:21 <BlueMatt> gmaxwell: huh? it should send ~50 blocks assuming its working
1247 2012-08-28 20:10:04 <BlueMatt> does your bitcoind already have blocks?
1248 2012-08-28 20:10:56 <gmaxwell> BlueMatt: same behavior with no .bitcoin directory at all.
1249 2012-08-28 20:11:09 <Luke-Jr> BlueMatt: well, I figure just a mapping database means you can google the binary hash to find someone with it :p
1250 2012-08-28 20:11:10 BTCTrader has quit (Quit: BTCTrader)
1251 2012-08-28 20:11:30 <BlueMatt> Luke-Jr: I dont see a use-case where git bisect doesnt do it
1252 2012-08-28 20:11:36 <BlueMatt> gmaxwell: what is the behavior?
1253 2012-08-28 20:11:39 <Luke-Jr> gavinandresen: shouldn't the BIP list include BIP 34? O.o
1254 2012-08-28 20:12:22 c_k has quit (Changing host)
1255 2012-08-28 20:12:22 c_k has joined
1256 2012-08-28 20:12:29 <Diapolo> gavinandresen: nice, the first time I got some insight in that whole process :) guess the coolest thing now is reading feedback ^^
1257 2012-08-28 20:12:55 <gmaxwell> BlueMatt: http://pastebin.com/bEd7uiMh
1258 2012-08-28 20:13:51 <BlueMatt> what does debug.log show?
1259 2012-08-28 20:14:00 <BlueMatt> or, is bitcoind not at localhost:8333?
1260 2012-08-28 20:14:13 <Diapolo> BlueMatt: It's running again means over the next days it will generate again :)? Thanks!
1261 2012-08-28 20:14:37 BTCTrader has joined
1262 2012-08-28 20:14:48 <BlueMatt> Diapolo: yea, sorry about the mess, guess I forgot to restart it after adding the chain check stuff last week
1263 2012-08-28 20:16:46 <Diapolo> BlueMatt: Well I love it and have to say thanks for such a tool, I don't expect a 100% uptime.
1264 2012-08-28 20:16:49 <gmaxwell> BlueMatt: ah, it wasn't listening.
1265 2012-08-28 20:17:22 <BlueMatt> Diapolo: thanks, now go write unit-tests ;)
1266 2012-08-28 20:17:46 <gavinandresen> Luke-Jr: yep, should mention BIP34.
1267 2012-08-28 20:17:58 <BlueMatt> Diapolo: esp bitcoin-qt unit tests, IIRC there is only one, but the code base is a bit bigger than one unit-test...
1268 2012-08-28 20:19:29 <Diapolo> BlueMatt: To be sure I and I guess wumpus too, should take that more serious ... but I've not looked into how those tests really are done. I know Gavin wrote a short how-to a few weeks ago... it's all just a matter of time in the end.
1269 2012-08-28 20:20:19 <BlueMatt> Diapolo: the qt stuff is a bit separate, I cant claim to know just about anything about the qt unit-test framework, but apparently its pretty good
1270 2012-08-28 20:20:50 <BlueMatt> (can apparently test by creating a gui and controlling it in the test, but I dunno how much effort is required)
1271 2012-08-28 20:21:48 <Diapolo> Diapolo: before 0.8 I will have looked into unit-tests I promise
1272 2012-08-28 20:21:58 <Diapolo> lol I send a message to me ^^
1273 2012-08-28 20:23:21 rdponticelli has quit (Ping timeout: 272 seconds)
1274 2012-08-28 20:24:57 <gmaxwell> BlueMatt: https://people.xiph.org/~greg/bitcoin_coverage/blocktest/home/gmaxwell/src/bcm/bax/src/index.html
1275 2012-08-28 20:25:25 <BlueMatt> thanks
1276 2012-08-28 20:25:54 <gmaxwell> BlueMatt: this is just your tests alone.
1277 2012-08-28 20:26:00 <BlueMatt> thanks
1278 2012-08-28 20:26:34 <Diapolo> In short, what is the difference between makefile.linux-mingw and makefile.mingw?
1279 2012-08-28 20:26:44 <BlueMatt> one is for xcompiling, the other is for native
1280 2012-08-28 20:27:10 <BlueMatt> Id give you about 50/50 odds makefile.mingw is broken at any given moment (dont think anyone uses it)
1281 2012-08-28 20:27:12 bitcoinz has joined
1282 2012-08-28 20:28:29 <Diapolo> I want to create the ASLR / DEP linker patch, I guess it needs to be in makefile.linux-mingw then or shall I put the flags into both?
1283 2012-08-28 20:28:40 <sipa> Diapolo: i suppose in both
1284 2012-08-28 20:28:44 <BlueMatt> both, preferably
1285 2012-08-28 20:28:55 <Diapolo> okay
1286 2012-08-28 20:28:55 overdone has joined
1287 2012-08-28 20:28:56 <gmaxwell> BlueMatt: on your blueskys todo list, one point would be pulltester running test_bitcoin and noting if the pull creates a net increase or decrease in coverage. Branch coverage is really the better target, but c++ cruft makes a lot of branches unreachable, so it would have to be line coverage.
1288 2012-08-28 20:29:53 <BlueMatt> gmaxwell: ack, also, more generally and acceptably now, we should more carefully enforce the idea that all pulls which do anything meaningful should include some form of reasonable unit tests
1289 2012-08-28 20:30:30 <BlueMatt> gmaxwell: what is +/-/#?
1290 2012-08-28 20:30:49 sirk390 has quit (Quit: Leaving.)
1291 2012-08-28 20:30:52 wizkid057 has quit ()
1292 2012-08-28 20:30:58 ovidiusoft has quit (Ping timeout: 245 seconds)
1293 2012-08-28 20:31:08 copumpkin is now known as stickle
1294 2012-08-28 20:31:31 <BlueMatt> gmaxwell: nevermind Ill google
1295 2012-08-28 20:31:31 dbe has joined
1296 2012-08-28 20:32:46 <gmaxwell> BlueMatt: In the long run I'd like to move in the direction that core code does not decrease coverage as a rule, subject to reason. I think we're too far off from enforcing something like that yet though.  UI code is often harder to test, so I don't have an opinion there.
1297 2012-08-28 20:33:36 <BlueMatt> gmaxwell: ui: meh, but I dont think its unreasonable or unenforceable to require most pulls to include unit tests
1298 2012-08-28 20:34:18 stickle is now known as copumpkin
1299 2012-08-28 20:34:54 <BlueMatt> (and wasnt that policy at one point that has now fallen by the wayside?)
1300 2012-08-28 20:38:17 <sipa> ow, the release notes mention -proxydns... which doesn't exist anymore
1301 2012-08-28 20:38:24 <sipa> also, maybe -loadblock should be mentioned?
1302 2012-08-28 20:38:54 <gavinandresen> sipa: can you fix the test/README.txt ?  Then ping me and I'll edit the version on the forums...
1303 2012-08-28 20:39:12 <gavinandresen> (ACK on both changes)
1304 2012-08-28 20:42:03 <gmaxwell> and yes, loadblock is a superduper feature.
1305 2012-08-28 20:44:58 <Diapolo> can someone take a look if that is sufficient, as I never touched non bitcoin-qt makefile stuff before ^^
1306 2012-08-28 20:45:01 <Diapolo> https://github.com/Diapolo/bitcoin/commit/216dd2501dc5a093f8e8d9569f72716651c8fb83
1307 2012-08-28 20:45:19 <kjj_> any chance someone could make a standard source tarball and stick it up on sf when you do releases?
1308 2012-08-28 20:45:25 <gmaxwell> BlueMatt: demanding tests isn't the same as demanding coverage. E.g. we have lots of tests that leave things uncovered, or they only test the fail or only test the pass cases.
1309 2012-08-28 20:45:32 <gmaxwell> kjj_: thats what the linux tarfile is.
1310 2012-08-28 20:45:41 copumpkin is now known as orifice
1311 2012-08-28 20:45:56 <kjj_> oh, I thought it was a pair of binaries the last time I grabbed it
1312 2012-08-28 20:48:14 <sipa> it contains source + binaries
1313 2012-08-28 20:48:36 <Luke-Jr> kjj_: GitHub also makes source-only tarballs
1314 2012-08-28 20:49:33 orifice is now known as copumpkin
1315 2012-08-28 20:50:02 <gmaxwell> Anyone else notice testnet3 not bootstrapping on its own?
1316 2012-08-28 20:50:23 <sipa> gmaxwell: noticed some time ago
1317 2012-08-28 20:50:34 <kjj_> Luke-Jr: that's what I grabbed, but they always have the git fuzz in the directory name
1318 2012-08-28 20:50:39 <gmaxwell> I just had to addnode it, slush mentioned that a few days ago.
1319 2012-08-28 20:50:50 OneFixt has quit ()
1320 2012-08-28 20:50:57 <gmaxwell> But I checked and my node is on irc, and I bootstrapped off it after I addnoded it.
1321 2012-08-28 20:51:55 <Diapolo> gmaxwell: thanks for your great explanation on Github about canonical :)
1322 2012-08-28 20:53:05 overdone has left ()
1323 2012-08-28 20:56:18 meLon has quit (Read error: Connection reset by peer)
1324 2012-08-28 20:56:42 rdponticelli has joined
1325 2012-08-28 20:58:29 <Diapolo> Thunderbird now includes an IRC chat... weird
1326 2012-08-28 21:01:44 Diapolo has left ()
1327 2012-08-28 21:01:54 Diapolo has joined
1328 2012-08-28 21:04:08 phantomcircuit_ has joined
1329 2012-08-28 21:05:04 Obsi has quit (Ping timeout: 268 seconds)
1330 2012-08-28 21:09:34 <Luke-Jr> Diapolo: is it Chatzilla?
1331 2012-08-28 21:10:12 <Diapolo> Luke-Jr: not sure, only saw that after updating
1332 2012-08-28 21:10:30 OneFixt has joined
1333 2012-08-28 21:11:34 <BlueMatt> gmaxwell: requiring the tests to reasonable would include covering new code reasonably well ;)
1334 2012-08-28 21:18:21 darkee has quit (Remote host closed the connection)
1335 2012-08-28 21:18:39 usagi has quit (Ping timeout: 244 seconds)
1336 2012-08-28 21:18:41 Diapolo has left ()
1337 2012-08-28 21:18:43 <gmaxwell> sipa: seems we've missed the irc default for testnet.
1338 2012-08-28 21:19:17 <gmaxwell> This is why testnet doesn't work on its own.
1339 2012-08-28 21:19:29 darkee has joined
1340 2012-08-28 21:24:08 <gavinandresen> gmaxwell: -testnet doesn't SetDefault("-irc", 1) ?
1341 2012-08-28 21:24:31 <gavinandresen> gmaxwell: I thought it did once upon a time, but the option-handling code has been reworked so many times.....
1342 2012-08-28 21:25:34 <gmaxwell> yea, seems not, I'll track it down and fix as required.
1343 2012-08-28 21:26:11 <gavinandresen> Make sure your bitcoin.conf isn't overriding default behavior....
1344 2012-08-28 21:26:24 pusle has quit ()
1345 2012-08-28 21:28:08 <gmaxwell> gavinandresen: clean node; actually I didn't catch this previously because my public nodes have irc=1 in their conf. :(
1346 2012-08-28 21:28:23 usagi has joined
1347 2012-08-28 21:29:20 edcba has quit (Remote host closed the connection)
1348 2012-08-28 21:29:27 Diablo-D3 has joined
1349 2012-08-28 21:29:33 <Luke-Jr> Give testnet it's own alert key. <-- FYI, "its" without an apostrophe
1350 2012-08-28 21:30:52 usagi has quit (Read error: Connection reset by peer)
1351 2012-08-28 21:30:58 <gavinandresen> it's not important.
1352 2012-08-28 21:31:02 usagi has joined
1353 2012-08-28 21:46:12 <sipa> gavinandresen: wait, test/README.txt ?
1354 2012-08-28 21:46:48 <gavinandresen> sipa: sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.7.0/test/README.txt
1355 2012-08-28 21:46:49 Turingi has quit (Read error: Connection reset by peer)
1356 2012-08-28 21:47:32 phantomcircuit_ has quit (Remote host closed the connection)
1357 2012-08-28 21:47:38 <sipa> gavinandresen: lol, right... i thought you were referring to src/test/README.txt in the source tree
1358 2012-08-28 21:48:13 rdponticelli has quit (Ping timeout: 272 seconds)
1359 2012-08-28 21:50:31 <justmoon> sipa: some more ultraprune feedback:  CScriptCompressor::IsToScriptID() is redundant, there is already a CScript::IsPayToScriptHash() - maybe all of the IsTo* methods should be moved to CScript?
1360 2012-08-28 21:50:53 cdecker has quit (Remote host closed the connection)
1361 2012-08-28 21:52:02 <Luke-Jr> the 'sv' translation replaces the period with a comma in the example; I thought period was always the decimal point for BTC?
1362 2012-08-28 21:55:44 RazielZ has quit (Ping timeout: 264 seconds)
1363 2012-08-28 21:57:55 <sipa> justmoon: I know, but it was done intentionally
1364 2012-08-28 21:58:03 <sipa> justmoon: i believe I added a comment about it, even
1365 2012-08-28 21:58:59 <justmoon> sipa: don't see one in script.h or script.cpp near those functions
1366 2012-08-28 21:59:14 <justmoon> sipa: what's the reason for the having the functions separate then?
1367 2012-08-28 21:59:42 <sipa> justmoon: they check for specific byte sequences
1368 2012-08-28 21:59:59 nhodges has quit (Quit: Connection closed for inactivity)
1369 2012-08-28 22:00:20 <justmoon> so does IsPayToScriptHash?
1370 2012-08-28 22:00:28 <sipa> right now, sure
1371 2012-08-28 22:01:07 <sipa> and probably it will stay that way
1372 2012-08-28 22:01:40 <sipa> i felt it was just cleaner to keep it separate, as I'm not testing for a script with certain semantics, but whether it matches a predefined case with a shortened notation
1373 2012-08-28 22:02:21 Marf has quit (Ping timeout: 240 seconds)
1374 2012-08-28 22:02:26 agricocb has quit (Remote host closed the connection)
1375 2012-08-28 22:02:38 <sipa> gavinandresen: updated docs/release-notes.txt
1376 2012-08-28 22:03:07 agricocb has joined
1377 2012-08-28 22:03:15 <justmoon> I see - well I still feel like they should be methods on CScript, but it's your call, just trying to be helpful :)
1378 2012-08-28 22:03:30 <sipa> sure, appreciate it!
1379 2012-08-28 22:04:48 <sipa> justmoon: but for example, the test for 0x04 pubkey outputs also tests whether it's a valid pubkey, as an invalid one cannot be compressed
1380 2012-08-28 22:04:58 <safra> hi, i am willing to pay someone 1 BTC to fix magento with paysius payment gateway
1381 2012-08-28 22:05:06 gavinandresen has quit (Quit: gavinandresen)
1382 2012-08-28 22:05:10 <safra> it is 99% working
1383 2012-08-28 22:05:53 <justmoon> sipa: ah, I see, by calling the CPubKey constructor it's verifying the key?
1384 2012-08-28 22:05:56 <justmoon> that might be worth a comment
1385 2012-08-28 22:06:02 cheako has joined
1386 2012-08-28 22:06:07 <sipa> agree
1387 2012-08-28 22:07:34 <justmoon> sipa: did you ever count how many times each of the special types currently appears on the block chain?
1388 2012-08-28 22:08:19 <sipa> actually, no
1389 2012-08-28 22:09:11 Luke-Jr has quit (Excess Flood)
1390 2012-08-28 22:09:14 Cryo has quit (Read error: Connection reset by peer)
1391 2012-08-28 22:09:31 Luke-Jr has joined
1392 2012-08-28 22:10:11 <justmoon> k, let me know if you do, would be quite interesting I think
1393 2012-08-28 22:10:20 Cryo has joined
1394 2012-08-28 22:10:20 Cryo has quit (Changing host)
1395 2012-08-28 22:10:20 Cryo has joined
1396 2012-08-28 22:10:42 AlexWaters has joined
1397 2012-08-28 22:11:03 ThomasV has quit (Quit: Quitte)
1398 2012-08-28 22:11:17 <sipa> my guess: 0x01 almost never, 0x00 almost always, 0x04 and 0x05 each around 100k times, 0x02 and 0x03 a few thousand times maybe
1399 2012-08-28 22:12:27 setkeh` has joined
1400 2012-08-28 22:13:08 setkeh has quit (Ping timeout: 264 seconds)
1401 2012-08-28 22:13:08 <justmoon> seems plausible
1402 2012-08-28 22:13:36 <sipa> justmoon: pushed new version with some comments
1403 2012-08-28 22:14:00 <justmoon> except wouldn't there be at least 190k to-pubkeys? because of the coinbases?
1404 2012-08-28 22:14:36 <gmaxwell> Hm. So, there are two issues with my testnet IRC tests. For some tests I had nolisten set, because I was trying to avoid inbound connections spoiling my sync up tests and I'd forgotten that this disables IRC. (about to submit a pull request to change that drainbramage)
1405 2012-08-28 22:15:05 <sipa> justmoon: 0x02-0x05 are all to-pubkeys, but indeed
1406 2012-08-28 22:15:12 <gmaxwell> The other is that we seem to have some issue where connect to unreachable addresses is causing 100% cpu long spins in connecting.. so 1 in N initial testnet startups takes a long time to get going.
1407 2012-08-28 22:16:08 <sipa> gmaxwell: hmm, still :(
1408 2012-08-28 22:18:37 <sipa> justmoon: if a signature previously had an R value less than 2^248, how did you encode it (before the DER fixes) ?
1409 2012-08-28 22:19:07 agricocb has quit (Remote host closed the connection)
1410 2012-08-28 22:19:12 <gmaxwell> justmoon: a large number of coinbases don't pay to pubkeys now.
1411 2012-08-28 22:21:52 <justmoon> sipa: it would have been encoded as a 31-byte number, see https://github.com/bitcoinjs/bitcoinjs-lib/blob/c0d740d2d4cd35ab18dc35e23936db42f17ad2ee/ecdsa.js#L210
1412 2012-08-28 22:22:49 <sipa> hmm, ok, good
1413 2012-08-28 22:23:08 <sipa> i still wonder where the excessively-padded signatures come from
1414 2012-08-28 22:23:24 <justmoon> sipa: CIA probably
1415 2012-08-28 22:23:27 <justmoon> :P
1416 2012-08-28 22:24:03 * justmoon dusts of tinfoil hat
1417 2012-08-28 22:24:08 <justmoon> off*
1418 2012-08-28 22:25:51 maaku has left ()
1419 2012-08-28 22:28:52 LuaKT has quit ()
1420 2012-08-28 22:32:26 vampireb has joined
1421 2012-08-28 22:33:28 <sipa> gmaxwell: another advantage of reverse-headers fetch: you know the number of total transactions in the chain before downloading it
1422 2012-08-28 22:33:29 <gmaxwell> sipa: http://pastebin.com/eUjwyMj2 This ring any bells for you? It gets a bunch of addresses but keeps trying a failing one.
1423 2012-08-28 22:33:50 <sipa> which can probably give a nicer progressbar than just estimated number of blocks
1424 2012-08-28 22:34:14 <sipa> gmaxwell: yes, looks familiar
1425 2012-08-28 22:38:12 <jgarzik> no gavin?
1426 2012-08-28 22:38:30 <jgarzik> gmaxwell sipa: did gavin mention anything about the release notes?  they are truncated, in his bitcoin-devel posting.
1427 2012-08-28 22:38:34 Matt_von_Mises has joined
1428 2012-08-28 22:38:44 <jgarzik> terribly odd
1429 2012-08-28 22:39:13 <jgarzik> "core bitcoin" section totally gone, BIPs 34 & 35 not mentioned... but BIP 22 is
1430 2012-08-28 22:39:23 <jgarzik> feels like an editing error
1431 2012-08-28 22:39:30 <gmaxwell> erp perhaps he posted an earlier version
1432 2012-08-28 22:39:47 <jgarzik> gmaxwell: nope...  it contains the Qt UI stuff, moved to the top
1433 2012-08-28 22:43:30 <jgarzik> Gavin's forum announce is similar
1434 2012-08-28 22:43:38 <jgarzik> except it mentions BIP 34
1435 2012-08-28 22:43:51 <sipa> weird
1436 2012-08-28 22:44:12 <jgarzik> he link-ified BIP 22 and BIP 34, so I bet he accidentally edited out the entire ("core bitcoin") section
1437 2012-08-28 22:45:07 sgornick has quit (Ping timeout: 260 seconds)
1438 2012-08-28 22:45:24 iocor has joined
1439 2012-08-28 22:45:49 <sipa> README.txt on sf also mentions 22 and 34
1440 2012-08-28 22:45:54 <sipa> and has no core section
1441 2012-08-28 22:46:51 <Matt_von_Mises> jgarzik: I responded to your concern here: https://bitcointalk.org/index.php?topic=103295.msg1140506#msg1140506 I really need to better know some bitcoin network stats.
1442 2012-08-28 22:46:57 <gmaxwell> huh, I didn't previously realize that IRC signaled the port numbers.
1443 2012-08-28 22:47:06 <sipa> gmaxwell: it always did
1444 2012-08-28 22:47:30 <gmaxwell> right I just didn't realize it. So one thing that is screwing us is that our default port preferences don't interact so great with upnp.
1445 2012-08-28 22:51:19 <kjj_> heh.  the change to addr.dat should save a couple hundred forum posts a month
1446 2012-08-28 22:51:35 <gmaxwell> Didn't we do that in 0.6.3?
1447 2012-08-28 22:52:19 <sipa> apparently not
1448 2012-08-28 22:52:24 <jgarzik> you seem addr.dat -> peers.dat?
1449 2012-08-28 22:52:26 <jgarzik> *mean
1450 2012-08-28 22:52:35 <jgarzik> I don't think that is in 0.6.3
1451 2012-08-28 22:52:41 Zarutian has joined
1452 2012-08-28 22:52:50 <gmaxwell> Sweet, thats a nice improvement.
1453 2012-08-28 22:52:54 <kjj_> yeah.  it is in the 070 release notes
1454 2012-08-28 22:53:09 <gmaxwell> It's been so long since it was written I thought it was out already. :(
1455 2012-08-28 22:53:19 <jgarzik> indeed
1456 2012-08-28 22:53:33 <jgarzik> this is why I've been pushing for an interim 0.7 release, before ultraprune and leveldb land
1457 2012-08-28 22:53:36 <kjj_> I'll have to update my mining nodes to stop nuking that file when restarting
1458 2012-08-28 22:53:43 <jgarzik> it's just a lot of small, but good, incremental changes
1459 2012-08-28 22:53:59 <kjj_> looks like a ton of good changes.  well done guys
1460 2012-08-28 22:54:07 <sipa> thanks :)
1461 2012-08-28 22:54:40 <sipa> 0.6-0.7 will be more than half a year :S
1462 2012-08-28 22:54:47 <gmaxwell>             if (!addr.IsValid() || setConnected.count(addr.GetGroup()) || IsLocal(addr))
1463 2012-08-28 22:54:50 <gmaxwell>                 break;
1464 2012-08-28 22:54:51 <gmaxwell> shouldn't that be a continue?
1465 2012-08-28 22:55:40 <kjj_> I've been thinking about using the transaction API to sweep my many small transactions into piles.  but first, beer
1466 2012-08-28 22:55:42 <gmaxwell> and shouldn't the ntries increment be above the addrman call, not that it really matters much.
1467 2012-08-28 22:55:54 <Luke-Jr> 0.6.1 had addrman…
1468 2012-08-28 22:56:27 eoss has joined
1469 2012-08-28 22:56:30 eoss has quit (Changing host)
1470 2012-08-28 22:56:30 eoss has joined
1471 2012-08-28 22:56:53 <sipa> gmaxwell: i'm trying to remember why i made it break instead of continue
1472 2012-08-28 22:57:03 Evilmax has joined
1473 2012-08-28 22:57:05 vampireb has quit (Quit: Lost terminal)
1474 2012-08-28 22:57:12 <sipa> iirc, it was a conscious decision, but i can't remember
1475 2012-08-28 22:57:35 <gmaxwell> Well, it contributes to not ever making its way to accepting non-default ports.
1476 2012-08-28 22:57:44 <gmaxwell> Most of the testnet3 nodes are on non-default ports.
1477 2012-08-28 22:58:03 <sipa> oh really? that explains a lot
1478 2012-08-28 22:59:04 agricocb has joined
1479 2012-08-28 22:59:20 <gmaxwell> Yep. There is only mine and that unreachable one it keeps trying.
1480 2012-08-28 22:59:36 <gmaxwell> And it never gets to allowing non-default ports because that innerloop doesn't run enough.
1481 2012-08-28 22:59:43 <gmaxwell> So the ntries break you added is a regression.
1482 2012-08-28 23:00:04 <sipa> why do people run them on non-standard ports...?
1483 2012-08-28 23:00:05 sgornick has joined
1484 2012-08-28 23:00:17 <gmaxwell> sipa: upnp will almost always result in a non-standard port.
1485 2012-08-28 23:00:24 <sipa> heh?
1486 2012-08-28 23:00:33 <sipa> what does upnp have to do with it?
1487 2012-08-28 23:00:38 <gmaxwell> people also intentionally run on other ports e.g. because they have mutiple nodes.
1488 2012-08-28 23:00:41 <jgarzik> us4.exmulti.net:18333 is still on testnet3
1489 2012-08-28 23:00:45 <jgarzik> if you need another node to test
1490 2012-08-28 23:01:07 <gmaxwell> jgarzik: why isn't it on IRC?
1491 2012-08-28 23:01:20 <jgarzik> ./bitcoind -testnet -daemon
1492 2012-08-28 23:01:25 <gmaxwell> sipa: the nat gateway assigns the external port.
1493 2012-08-28 23:01:26 <jgarzik> gmaxwell: because that is default behavior?
1494 2012-08-28 23:01:32 <sipa> gmaxwell: wut?
1495 2012-08-28 23:01:35 <gmaxwell> jgarzik: no, IRC is default for testnet.
1496 2012-08-28 23:02:03 <jgarzik> gmaxwell: I never set noirc
1497 2012-08-28 23:02:06 <gmaxwell> jgarzik: can you look in your logs and see if its using IRC?
1498 2012-08-28 23:02:06 <sipa> gmaxwell: if that's the case, bitcoin cannot deal with it at all
1499 2012-08-28 23:02:13 <jgarzik> gmaxwell: maybe irc server died, and bitcoin did not reconnect?
1500 2012-08-28 23:02:17 <jgarzik> gmaxwell: sure
1501 2012-08-28 23:02:23 <gmaxwell> jgarzik: it reconnects.
1502 2012-08-28 23:02:34 <sipa> we ask to open a mapping of external port GetListenPort() to us, and advertize GetListenPort()
1503 2012-08-28 23:02:34 <gmaxwell> sipa: perhaps I'm on drugs, but I'm pretty sure thats how it works.
1504 2012-08-28 23:02:48 <gmaxwell> hm.
1505 2012-08-28 23:02:57 <sipa> do you drink coffee?
1506 2012-08-28 23:03:04 <jgarzik> one sec
1507 2012-08-28 23:03:35 <gmaxwell> I'm fasting today because I pulled out a broken tooth. That must be the cause of my density. Thats my story and I'm sticking to it.
1508 2012-08-28 23:05:29 * jgarzik taps his foot, wondering why a 77MB log file takes so long to bzip2
1509 2012-08-28 23:06:40 <jgarzik> gmaxwell: http://gtf.org/garzik/bitcoin/debug.log.testnet3.gmaxwell.gz
1510 2012-08-28 23:07:01 <gmaxwell> ThreadIRCSeed started
1511 2012-08-28 23:07:01 <gmaxwell> ThreadIRCSeed exited
1512 2012-08-28 23:07:12 <gmaxwell> can I see your relevant bitcoin.conf?
1513 2012-08-28 23:07:27 <gmaxwell> oh it's connecting later down.
1514 2012-08-28 23:07:39 <jgarzik> gmaxwell: bitcoin.conf:
1515 2012-08-28 23:07:40 <jgarzik> rpcuser=XX
1516 2012-08-28 23:07:40 <jgarzik> rpcpassword=YY
1517 2012-08-28 23:07:40 <jgarzik> rpcport=18332
1518 2012-08-28 23:07:56 * sipa l33t h4xx0rz jgarzik's node
1519 2012-08-28 23:08:13 <jgarzik> good luck trying :)
1520 2012-08-28 23:08:32 <jgarzik> random tangent:  we should change rpc port, if testnet
1521 2012-08-28 23:08:40 <sipa> why?
1522 2012-08-28 23:08:48 <jgarzik> it stays at 8332 even for testnet, which obviously conflicts if a regular node is running
1523 2012-08-28 23:08:56 <jgarzik> and for safety, so you don't confuse the two
1524 2012-08-28 23:09:05 <sipa> good point
1525 2012-08-28 23:09:20 <gmaxwell> yea, most irritating thing about getting testnet running, IMO
1526 2012-08-28 23:09:30 * jgarzik files an issue as a reminder
1527 2012-08-28 23:09:59 Dagger2 has quit (Ping timeout: 250 seconds)
1528 2012-08-28 23:10:28 bitcoinz has quit (Remote host closed the connection)
1529 2012-08-28 23:10:30 freewil has quit (Quit: Leaving)
1530 2012-08-28 23:11:51 bitcoinz has joined
1531 2012-08-28 23:12:03 <gmaxwell> IRC got who
1532 2012-08-28 23:12:03 <gmaxwell> Added 69.64.46.74:18334 from 193.107.204.81: 0 tried, 6 new
1533 2012-08-28 23:12:07 <gmaxwell> 18334?
1534 2012-08-28 23:12:16 <gmaxwell> 0_o
1535 2012-08-28 23:12:34 <gmaxwell> u4tf5gfU7Pqx8fp
1536 2012-08-28 23:12:35 <gmaxwell> AddLocal(69.64.46.74:18333,4)
1537 2012-08-28 23:12:35 <gmaxwell> IRC SENDING: NICK u4tf5gfU7Pqx8fp
1538 2012-08-28 23:13:37 <gmaxwell> 19:07 -!- u4tf5gfU7Pqx8fp [u4tf5gfU7J@us4.exmulti.net]
1539 2012-08-28 23:13:37 <gmaxwell> 19:07 -!-  ircname  :  u4tf5gfU7JNQN2U
1540 2012-08-28 23:17:58 eb3kk has quit (Ping timeout: 244 seconds)
1541 2012-08-28 23:18:30 Gladamas_ has joined
1542 2012-08-28 23:18:41 ZephyrVoid has quit (Ping timeout: 240 seconds)
1543 2012-08-28 23:18:46 <Raccoon> jesus christ, it's taking me 2 weeks to download the last 65 days of blocks.
1544 2012-08-28 23:20:03 Gladamas has quit (Ping timeout: 272 seconds)
1545 2012-08-28 23:20:09 <Raccoon> i can torrent the entire 200 gig library of Doctor Who in that time
1546 2012-08-28 23:20:52 <sipa> it is also not the downloading that is the problem
1547 2012-08-28 23:20:56 <gmaxwell> Raccoon: it certantly shouldn't take two weeks unless something is wrong or weird with your computer.
1548 2012-08-28 23:21:05 <gmaxwell> But we've gone over this before, IIRC.
1549 2012-08-28 23:21:27 <Raccoon> someone said something about the dice game clogging up bitcoin?
1550 2012-08-28 23:21:46 freewil has joined
1551 2012-08-28 23:21:48 <sipa> they are responsible for a significant part of the recent blocks
1552 2012-08-28 23:23:34 rdponticelli has joined
1553 2012-08-28 23:24:04 <sipa> Raccoon: are you running on a USB stick, over network, on an encrypted filesystem, or on a very slow CPU?
1554 2012-08-28 23:24:44 <Raccoon> encrypted file system on a 1.7ghz single core laptop, pATA laptop
1555 2012-08-28 23:24:50 <gmaxwell> Truecrypt?
1556 2012-08-28 23:24:51 <Raccoon> er harddrive
1557 2012-08-28 23:24:54 <Raccoon> right.
1558 2012-08-28 23:25:31 <sipa> that's probably it; truecrypt seems to have a detrimental effect on disk write latency
1559 2012-08-28 23:25:58 <Raccoon> but i dunno, shouldn't bitcoin at least use some ram for caching?
1560 2012-08-28 23:26:03 <sipa> it does
1561 2012-08-28 23:26:15 <sipa> but caching doesn't mean less writes
1562 2012-08-28 23:26:22 <sipa> it means less reads
1563 2012-08-28 23:26:23 <gmaxwell> yea, well, http://www.tomshardware.com/reviews/truecrypt-aes-ni-encryption,2899-6.html < we can't do anything about truecrypt being insanely slow
1564 2012-08-28 23:26:26 <Raccoon> i have 4 gigs.  it could fit the entire database
1565 2012-08-28 23:26:42 <Raccoon> ram drive it
1566 2012-08-28 23:27:02 <Raccoon> the database is scarcly over 1 gig in size
1567 2012-08-28 23:27:12 <Raccoon> i can copy a gig in minutes
1568 2012-08-28 23:27:23 <Raccoon> not weeks
1569 2012-08-28 23:27:28 cheako has quit (Remote host closed the connection)
1570 2012-08-28 23:27:32 <sipa> it is not copying that is the problem
1571 2012-08-28 23:27:36 <sipa> it is keeping it consistent
1572 2012-08-28 23:27:41 <Raccoon> right
1573 2012-08-28 23:27:45 <Raccoon> but ram drive it
1574 2012-08-28 23:27:50 <gmaxwell> You may do that, yes. But bitcoin itself can't do that. It's not welcome to use gigabytes of ram, especially since normal operating systems do their own caching with free ram.
1575 2012-08-28 23:27:54 <Raccoon> why is it reading the hdd at all?
1576 2012-08-28 23:28:04 <gmaxwell> reading isn't the slow part.
1577 2012-08-28 23:28:06 <sipa> it's probably not reading much
1578 2012-08-28 23:28:14 <sipa> but hopefully leveldb will soon improve BDB's lack of I/O efficiency
1579 2012-08-28 23:28:16 RainbowDashh has joined
1580 2012-08-28 23:28:27 <Raccoon> firefox welcomes itself to a gig
1581 2012-08-28 23:28:40 <Luke-Jr> Raccoon: which is why I don't use firefox, and I doubt I'm alone
1582 2012-08-28 23:29:00 <safra> is there a PHP dev here?
1583 2012-08-28 23:29:01 <Raccoon> but at least for db sync, bitcoin could just use ram
1584 2012-08-28 23:29:09 slush1 has joined
1585 2012-08-28 23:29:15 <Raccoon> a few hours isn't bad compared to a few weeks
1586 2012-08-28 23:29:29 <justmoon> I have 12 gigs, so firefox is a good fit for me :P
1587 2012-08-28 23:30:06 <gmaxwell> Raccoon: so run bitcoin in a ramdisk, you can do this easily.
1588 2012-08-28 23:30:06 <jgarzik> gmaxwell: no :18334 here
1589 2012-08-28 23:30:16 <Raccoon> gmaxwell: how?
1590 2012-08-28 23:30:22 <Raccoon> what's a good ramdisk app?
1591 2012-08-28 23:30:24 <gmaxwell> jgarzik: right, but my node decodes your irc as :18334.
1592 2012-08-28 23:30:25 <jgarzik> gmaxwell: netstat sees 8332, 8333, 18332, 18333
1593 2012-08-28 23:30:33 <jgarzik> gmaxwell: huh
1594 2012-08-28 23:30:35 <Raccoon> and do I need to copy stuff back and forth?
1595 2012-08-28 23:30:47 <jgarzik> gmaxwell: git HEAD as of ~4 weeks ago
1596 2012-08-28 23:30:56 <jgarzik> vanilla
1597 2012-08-28 23:31:31 <Luke-Jr> justmoon: I have 16 GB, but I still don't have GBs to spare for a browser
1598 2012-08-28 23:31:45 <sipa> Raccoon: here's how it works: we do transactions, BDB caches recent data in RAM, but every write is written immediately to a .log file; when the .log files become too large together, they are flushed to the actual DB files
1599 2012-08-28 23:31:46 <gmaxwell> (and come on, why the @#$@ do we just @#$#@ seralize the memory of a @#$@# c-struct for this. fuck. it's not like this is performance critcial; though if thats the source of the problem I don't see it)
1600 2012-08-28 23:31:47 ZephyrVoid has joined
1601 2012-08-28 23:32:05 Gladamas_ is now known as Gladamas
1602 2012-08-28 23:32:09 <Raccoon> sipa >_<
1603 2012-08-28 23:32:11 <sipa> gmaxwell: what are you talking about?
1604 2012-08-28 23:32:29 <sipa> gmaxwell: ah, the IRC thing
1605 2012-08-28 23:32:59 logger_ has quit (Ping timeout: 256 seconds)
1606 2012-08-28 23:33:07 <sipa> Raccoon: other databases are not different; you can't avoid writing things immediately to disk if you don't want to lose them
1607 2012-08-28 23:33:19 <Raccoon> i have a feeling bitcoin is going to grind the polish off my harddisk platters
1608 2012-08-28 23:33:21 slush has quit (Ping timeout: 272 seconds)
1609 2012-08-28 23:33:36 <sipa> Raccoon: BDB has an option not to do that, but it also means risk corrupting the database in case of a crash, without detecting it corrupted
1610 2012-08-28 23:33:44 <jgarzik> Raccoon: if your hard drive spins in _just_ the right way, it explodes like a grenade
1611 2012-08-28 23:33:50 <Raccoon> sipa: i don't care if i lose them in a power failure.  i can always download it again.
1612 2012-08-28 23:34:08 <sipa> Raccoon: that's the point; indeed we don't care about losing it, but we do care about corrupting it
1613 2012-08-28 23:34:30 <sipa> and BDB does not provide a way to get one without the other
1614 2012-08-28 23:34:31 <justmoon> I find there is a certain irony - people were hating on client developers who just put the whole database in ram, now "put it on a ramdisk" is our standard response in case of performance issues :)
1615 2012-08-28 23:34:46 <Raccoon> sipa: so you're saying that rubbing off the magnetic media from my harddisk doesn't risk corruption?
1616 2012-08-28 23:35:01 <sipa> Raccoon: ?
1617 2012-08-28 23:35:06 <Luke-Jr> justmoon: …
1618 2012-08-28 23:35:37 <Raccoon> justmoon: it's OK to utilize ram on a temporary operation.
1619 2012-08-28 23:35:37 <justmoon> Luke-Jr: was that a silence of agreement or a punitive silence of disagreement?
1620 2012-08-28 23:35:47 <gmaxwell> justmoon: it's not a standard response, Raccoon is running an encryption later that gives something like 1/100th the random write performance; and he's asking bitcoin to use ram. I suggested that if he has enough ram and wants to use it like that, he can use a ramdisk.
1621 2012-08-28 23:36:08 <Raccoon> justmoon: eg, for 2 hours while the DB syncs, same as when someone is encoding a DVD/XviD
1622 2012-08-28 23:37:01 <gmaxwell> My answer would be, don't use an encryption layer that slays your random write performance for the bitcoin node.. or do so but understand that it's going to be very slow to sync that way.
1623 2012-08-28 23:37:04 <Raccoon> gmaxwell: bitcoin Wants to be encrypted.
1624 2012-08-28 23:37:07 <justmoon> Raccoon: future version of bitcoin will do just that - leveldb uses much larger RAM caches by default and sipa's ultraprune also goes in the direction of reducing the number of reads from disk significantly
1625 2012-08-28 23:37:21 <gmaxwell> justmoon: its NOT ABOUT READS>
1626 2012-08-28 23:37:30 <justmoon> Raccoon: not the block chain which is public
1627 2012-08-28 23:37:44 <gmaxwell> the performance is bounded by sync writes.
1628 2012-08-28 23:37:45 <Raccoon> jurov|away: if you want to hide that you even use bitcoin, yes
1629 2012-08-28 23:37:50 <Raccoon> justmoon:
1630 2012-08-28 23:38:32 <sipa> gmaxwell: sure reads matters
1631 2012-08-28 23:38:33 <Luke-Jr> Raccoon: this is PUBLIC RECORD DATA
1632 2012-08-28 23:38:34 <Raccoon> because bitcoin is legally defined as "just cause"
1633 2012-08-28 23:38:38 <Luke-Jr> there is NO BENEFIT to encrypting it
1634 2012-08-28 23:38:39 bobke has quit (Read error: No route to host)
1635 2012-08-28 23:38:42 <justmoon> gmaxwell: ultraprune is an additional index (for now) - so it increases the number of writes
1636 2012-08-28 23:38:43 bobke_ has joined
1637 2012-08-28 23:38:57 <sipa> justmoon: ultraprune certainly decreases the amount of writes
1638 2012-08-28 23:38:59 <Luke-Jr> Raccoon: nonsense
1639 2012-08-28 23:39:01 <gmaxwell> justmoon: no, it massively decreases them.
1640 2012-08-28 23:39:14 <sipa> but it also reduces the size of the working set
1641 2012-08-28 23:39:20 <sipa> which means that cache has more effect
1642 2012-08-28 23:39:29 <gmaxwell> @#$@@#$@ high level programming language users who haven't a clue what your computer is doing under the hood.
1643 2012-08-28 23:39:44 <justmoon> sipa: that part I get - it reduces writes how? because we don't update vSpent anymore?
1644 2012-08-28 23:39:56 bobke_ has left ()
1645 2012-08-28 23:39:59 <copumpkin> gmaxwell: omg how could they!
1646 2012-08-28 23:40:11 <sipa> justmoon: the database is 10x smaller
1647 2012-08-28 23:40:20 <Raccoon> Luke-Jr: sorry.  "probable cause"
1648 2012-08-28 23:40:38 <jgarzik> gmaxwell: not exclusively high level, c.f. cbitcoin
1649 2012-08-28 23:40:57 <gmaxwell> justmoon: the number of writes required to update the index depends on the index size. It's not an O(1) operation just because you make one call to the database. :)
1650 2012-08-28 23:40:59 <justmoon> sipa: if you run ultraprune instead of the traditional index, but I thought we kept that around?
1651 2012-08-28 23:41:14 <sipa> justmoon: no, blkindex.dat is gone
1652 2012-08-28 23:41:17 <justmoon> gmaxwell: I know that thanks
1653 2012-08-28 23:41:31 <sipa> i may add that again, but it's only necessary for getrawtransaction, basically
1654 2012-08-28 23:41:32 <justmoon> gmaxwell: your condescension is starting to irritate me a bit
1655 2012-08-28 23:42:04 <justmoon> sipa: oh ok, yeah then it reduces writes massively obv
1656 2012-08-28 23:42:23 <sipa> still, even if i'd re-add it, i believe the writes would still be a lot lower
1657 2012-08-28 23:42:27 <sipa> indeed because of vSpent
1658 2012-08-28 23:43:02 RainbowDashh has quit (Quit: Mithay (<HOMENAME>) left IRC. (Quit: Hi, I'm a sleep message virus. Please replace your old line with this line and help me take over the world of IRC.))
1659 2012-08-28 23:43:46 <sipa> it's really quite pathetic: the old index's format requires more storage to just save which outputs are spent + where, than ultraprune needs to know spentness PLUS the actual txout data
1660 2012-08-28 23:45:02 <justmoon> hehe yes, I was pretty blown away when libbzing indexed the whole block chain in two minutes
1661 2012-08-28 23:45:23 iocor has quit (Quit: Textual IRC Client: http://www.textualapp.com/)
1662 2012-08-28 23:45:36 <sipa> hmm?
1663 2012-08-28 23:46:13 <justmoon> libbzing was my blockchain database implementation in ANSI C that I used to compare the performance of different backends: https://github.com/bitcoinjs/bzing/blob/master/src/bzing.c#L657
1664 2012-08-28 23:46:36 <sipa> right, you showed me that when I was at your place?
1665 2012-08-28 23:46:39 <justmoon> yep
1666 2012-08-28 23:46:48 <Luke-Jr> Raccoon: what jurisdiction has defined Bitcoin to be probable cause, and probable cause for what?
1667 2012-08-28 23:46:49 <sipa> it inspired me to write ultraprune :)
1668 2012-08-28 23:46:56 * justmoon blushes
1669 2012-08-28 23:47:17 <justmoon> well good night all!
1670 2012-08-28 23:47:33 <sipa> cya
1671 2012-08-28 23:47:34 logger_ has joined
1672 2012-08-28 23:47:35 <justmoon> no hard feelings gmaxwell, I know us noobs can get frustrating ^^
1673 2012-08-28 23:47:39 <justmoon> cu
1674 2012-08-28 23:47:43 justmoon has quit (Quit: Leaving)
1675 2012-08-28 23:52:32 logger_ has quit (Ping timeout: 245 seconds)
1676 2012-08-28 23:52:57 <gmaxwell> sipa: As far as reads mattering, I don't really think they do for Raccoon's case, though it's hard to tell. But the benchmarks I found don't seem to suggest that truecrypt breaks read caching.
1677 2012-08-28 23:53:28 <gmaxwell> So I'm making the reasonable guess that it's all the sync writes killing it; it's speculation though.
1678 2012-08-28 23:54:02 <Matt_von_Mises> Here's how you can prune the block chain, easily: Block 193000 is hardcoded… so why not take all the unspent previous outputs at this block, hard code those into the software and then get the software to start at block 193000, forgetting the rest?
1679 2012-08-28 23:54:50 <sipa> Matt_von_Mises: that means effectively making security depend on a correct txout set being encoded in the client
1680 2012-08-28 23:55:13 <Matt_von_Mises> sipa: Security already depends on the client...
1681 2012-08-28 23:55:20 <Matt_von_Mises> In so many ways
1682 2012-08-28 23:55:22 <gmaxwell> Matt_von_Mises: Because doing so would substantially degrade the bitcoin security model from "the software you run verifies the rules for itself" to opaque trust.
1683 2012-08-28 23:55:43 <Matt_von_Mises> You already need to trust the software.
1684 2012-08-28 23:55:52 <sipa> of course it depends on that, but we can't have a better security model for the source code
1685 2012-08-28 23:56:00 <gmaxwell> Matt_von_Mises: you can audit the software, I did.
1686 2012-08-28 23:56:10 <sipa> bitcoin is unique exactly because it does not require you to trust the blockchain data
1687 2012-08-28 23:56:24 <Matt_von_Mises> Why can't you audit the unspent prev outputs? :P
1688 2012-08-28 23:56:29 <sipa> you can
1689 2012-08-28 23:56:35 <gmaxwell> Thats what it does. You're suggesting removing that.
1690 2012-08-28 23:56:55 <sipa> but there's a difference between having it being verified automatically, and just allowing people to verify it but not doing it by default
1691 2012-08-28 23:57:26 <gmaxwell> If that were the only way to get good performance; ... well, first I'd be selling all my bitcoins; but after that... well, ...  but it's not the only way.
1692 2012-08-28 23:57:33 logger_ has joined
1693 2012-08-28 23:57:46 <gmaxwell> It's an _easy_ way, but it comes at the expense of part of what makes bitcoin worth having.
1694 2012-08-28 23:57:47 <Matt_von_Mises> Yes but you could still ruin the software anyway. People need to look regardless. Not everyone wants to look therefore they trust the software.
1695 2012-08-28 23:58:17 <sipa> Matt_von_Mises: for the source code we use a model of "no automatical verification, but we assume you'll hear about mistakes"
1696 2012-08-28 23:58:25 <sipa> you can use that same model for the txout data, yes
1697 2012-08-28 23:58:30 <sipa> but we don't have to do that
1698 2012-08-28 23:58:37 <sipa> we CAN do better for that part
1699 2012-08-28 23:58:59 <gmaxwell> Well, moreover, if the reference software does not do this, then nothing will.
1700 2012-08-28 23:59:17 Zarutian has quit (Quit: Zarutian)
1701 2012-08-28 23:59:24 Obsi has joined
1702 2012-08-28 23:59:58 <gmaxwell> And if we don't solve the challenging problems behind making it acceptably fast with the motivation of it being on the critcial path, then few to none one will check manually becuase no one will bother making some exceptional thing that only weirdos do fast.