1 2012-07-01 00:00:25 <ali1234> i ask because i have an old smartphone with a half finished linux port. the screen, keyboard and usb works but that's about it. it's no use as a phone but would be perfect for this, and it's just sitting unused on my desk.
   2 2012-07-01 00:01:07 <jine> Well, there is bitcoin clients (both full clients and thiny ones) for that might work on your phone.
   3 2012-07-01 00:01:23 <jine> I have no idea about the status for such a project tho, haven't seen anything public about it (as a pressrelease or similar)
   4 2012-07-01 00:02:40 <ali1234> thin client would be a start, but i would want to integrate it with a PC side full client
   5 2012-07-01 00:03:33 <ali1234> the phone would only be needed when something needed to be signed with the private key. i would want to keep the USB protocol as simple as possible to avoid bugs, and have no other network connections at all (since they don't work anyway)
   6 2012-07-01 00:06:56 graingert has quit (Read error: Connection reset by peer)
   7 2012-07-01 00:07:22 <ali1234> actually bluetooth works as well, that could be interesting
   8 2012-07-01 00:13:25 DAN- has quit (Read error: Connection reset by peer)
   9 2012-07-01 00:19:09 minimoose has joined
  10 2012-07-01 00:20:06 <jrmithdobbs> ali1234: why even bother? do it over a tls encrypted/authenticated tcp connection from the get go and you get all the same functionality without the wire snooping and driver/hid wrapper writing
  11 2012-07-01 00:20:34 <jrmithdobbs> ali1234: usb is not a secure bus and can be sniffed by other devices on it
  12 2012-07-01 00:21:30 <ali1234> it doesn't matter
  13 2012-07-01 00:21:35 <ali1234> the key will never be sent over the usb bus
  14 2012-07-01 00:21:42 <ali1234> so there is nothing to snoop
  15 2012-07-01 00:22:13 rdponticelli has quit (Ping timeout: 246 seconds)
  16 2012-07-01 00:22:51 <ali1234> instead the message digest would be sent, and then the signature would be returned
  17 2012-07-01 00:23:01 <ali1234> and that's all public information that goes in the blockchain
  18 2012-07-01 00:23:24 <jgarzik> hrm
  19 2012-07-01 00:23:37 <jgarzik> has anyone yet written a python interpreter for bitcoin scripts?
  20 2012-07-01 00:26:28 <jrmithdobbs> jgarzik: i think either genjix or phantom have a partial one somewhere
  21 2012-07-01 00:26:43 <jgarzik> phantomcircuit: ^
  22 2012-07-01 00:27:14 <jrmithdobbs> don't know how complete, doesn't libbitcoin have py bindings though? that good enough?
  23 2012-07-01 00:27:17 <phantomcircuit> jgarzik, i have one but there's a ton of bugs i just dont care to fix
  24 2012-07-01 00:28:40 <jrmithdobbs> I have a weird question
  25 2012-07-01 00:29:06 <jgarzik> hrm
  26 2012-07-01 00:29:30 <jrmithdobbs> so i was going through the udacity.com crypto course thing to see what their stuff was like and the guy covers bitcoin (very briefly at very high level, in fact he gets scripts/coin xfer completely wrong ...)
  27 2012-07-01 00:29:42 <jrmithdobbs> and he made a couple statements that I don't agree with and wanted opinions ;p
  28 2012-07-01 00:30:57 <jrmithdobbs> eg, re: sha256(sha256()) he says the only justification was to make the POW "more" work specifically saying that preventing pre-image/collision issues is not the reasons for it. I know those two properties *are* a benefit we've seen vs certain possible attacks ....
  29 2012-07-01 00:31:40 <jrmithdobbs> but was that the original reason? It seems it's unnecessary for that specific reason since you can just change the diff adjust alg to account for 2x as many blocks per interval, for instance
  30 2012-07-01 00:32:20 TD has quit (Quit: TD)
  31 2012-07-01 00:32:44 <jrmithdobbs> thoughts/comments?
  32 2012-07-01 00:34:24 <ali1234> ah, it looks like armory has most of what i need
  33 2012-07-01 00:37:16 <phantomcircuit> jgarzik, https://github.com/phantomcircuit/bitcoin-alt
  34 2012-07-01 00:37:24 <phantomcircuit> there's a bug with the block organization
  35 2012-07-01 00:37:33 <phantomcircuit> but other than that it mostly works
  36 2012-07-01 00:38:19 <jrmithdobbs> noone? really? ;p
  37 2012-07-01 00:40:29 rdponticelli has joined
  38 2012-07-01 00:42:19 <jgarzik> phantomcircuit: building my own block reorg handling, just hoping to avoid reinventing the script bits
  39 2012-07-01 00:42:19 <ali1234> jrmithdobbs: wait, how does hashing twice prevent collisions?
  40 2012-07-01 00:42:45 Diablo-D3 has quit (Ping timeout: 244 seconds)
  41 2012-07-01 00:42:53 <phantomcircuit> jgarzik, someone else had done the scripting in python
  42 2012-07-01 00:43:03 <phantomcircuit> there are a ton of very subtle mistakes you can make though
  43 2012-07-01 00:43:13 <phantomcircuit> so reading through the mainline code is very important
  44 2012-07-01 00:43:24 <phantomcircuit> and of course always err on the side of rejecting
  45 2012-07-01 00:43:40 <jrmithdobbs> ali1234: because colliding the single hash is already difficult, having to go through 2 iterations is statistically harder
  46 2012-07-01 00:43:51 <jgarzik> phantomcircuit: bitcoin.script seems reasonably self-contained
  47 2012-07-01 00:44:27 <jrmithdobbs> ali1234: or rather, the probability of their being an issue which can be carried through multiple iterations is lower than one that one be able affect a single iteration
  48 2012-07-01 00:44:57 <ali1234> ok. but it doesn't prevent accidental collisions at all. maybe that's what he meant?
  49 2012-07-01 00:45:32 <jrmithdobbs> oh well sure, it doesn't prevent accidental collisions but the comment was in the context of security model / blockchain structure
  50 2012-07-01 00:46:39 <jrmithdobbs> ali1234: it does, on the other hand, prevent the length extension we know that sha2 is theorhetically weak against
  51 2012-07-01 00:48:23 brwyatt is now known as brwyatt|Away
  52 2012-07-01 00:48:23 <jrmithdobbs> (once anyone is actually able to show a collision, anyways)
  53 2012-07-01 00:48:40 <ali1234> what about hash chains?
  54 2012-07-01 00:48:57 <jrmithdobbs> what about them?
  55 2012-07-01 00:49:22 <ali1234> if you used that type of attack and found a collision for sha256(x), then the collision for sha256(sha256(x)) is just the previous number in the chain
  56 2012-07-01 00:54:52 graingert_ has quit (Ping timeout: 256 seconds)
  57 2012-07-01 00:55:56 <jrmithdobbs> ali1234: not exactly ... you do know that there are, for instance, reduced round attacks on aes, right? It's a similar situation, just because you can perform the attack through x number of iterations doesn't mean you can through y at the same cost/benefit
  58 2012-07-01 00:56:00 t7 has quit (Quit: ChatZilla 0.9.88.2 [Firefox 13.0.1/20120614114901])
  59 2012-07-01 00:59:09 <jrmithdobbs> ali1234: so say there is a sha2 attack that successfully reduces the search to 2^50-60ish (the pratical, but still hard range), that would leave sha256(sha256()) at about 2^90-100+ which is still quite infeasible
  60 2012-07-01 00:59:15 <jrmithdobbs> ali1234: https://bitcointalk.org/index.php?topic=45456.0;wap2 basically
  61 2012-07-01 01:07:22 rdponticelli has quit (Ping timeout: 246 seconds)
  62 2012-07-01 01:09:52 <luke-jr> jine: no u troll
  63 2012-07-01 01:11:31 <gmaxwell> luke-jr: figure out what happened to your bitcoin directory?
  64 2012-07-01 01:12:09 <gmaxwell> luke-jr: I assume you had a 7GB debug.log?
  65 2012-07-01 01:12:23 Xunie has joined
  66 2012-07-01 01:13:38 MysteryBanshee has quit ()
  67 2012-07-01 01:14:01 * jgarzik ponders logging addresses of orphan TX suppliers
  68 2012-07-01 01:15:33 brwyatt is now known as Away!~brwyatt@brwyatt.net|brwyatt
  69 2012-07-01 01:15:44 <ali1234> jrmithdobbs: if you have an attack that reduces upper bound for a single collision to 2^60 then calculations required for double-hash is 2^61, no?
  70 2012-07-01 01:15:56 theodore has joined
  71 2012-07-01 01:19:36 <jrmithdobbs> ali1234: important part of that is near the bottom
  72 2012-07-01 01:19:42 <jrmithdobbs> ali1234: "Finding collisions is irrelevant for bitcoin, you need to find partial preimage to extend the blockchain, or 2nd-preimage to replace a non-recent block in the blockchain (in the old thread Satoshi mentioned adding extra field with new hash function to blocks, in case 2nd-preimage attack on sha256 becomes practical)."
  73 2012-07-01 01:20:37 <ali1234> ok, well replace "collision" with "preimage" in what i said...
  74 2012-07-01 01:22:09 <galambo>  it doesnt really matter because after its on everyones block chain you cant substitute the merkle root as well
  75 2012-07-01 01:22:34 <galambo> to add extra transactions or whatever you're planning
  76 2012-07-01 01:23:57 <jrmithdobbs> ali1234: yes, you're right, i think
  77 2012-07-01 01:24:12 <ali1234> i think the guy talking about collisions really meant to say preimage anyway
  78 2012-07-01 01:24:17 <jrmithdobbs> ali1234: so the distinction for bitcoin matters since it's not the same kind of search
  79 2012-07-01 01:24:22 <jrmithdobbs> ya
  80 2012-07-01 01:24:41 rdponticelli has joined
  81 2012-07-01 01:24:47 <jrmithdobbs> so he was right re: collisions then
  82 2012-07-01 01:25:11 <jrmithdobbs> but definitely wrong re: pre-image resistance at the same time
  83 2012-07-01 01:25:19 <jrmithdobbs> or possibly, anyways
  84 2012-07-01 01:27:08 <jrmithdobbs> ok, i see the origins now
  85 2012-07-01 01:27:11 <ali1234> if you have z = sha256(y=sha256(x)) and you can compute y in n attacks, then you can compute y and then use it to compute z in another n attacks, ie 2n in total, not n^2.
  86 2012-07-01 01:27:15 <jrmithdobbs> http://crypto.stackexchange.com/questions/779/hashing-or-encrypting-twice-to-increase-security
  87 2012-07-01 01:28:13 <ali1234> and yes, hash(x) = hash(y) implies hash(hash(x) = hash(hash(y)) - which is why i said it doesn't prevent accidental collisions at all
  88 2012-07-01 01:28:37 <jrmithdobbs> so it protects against first (but not second) pre-image sometimes but mostly against length extension
  89 2012-07-01 01:30:20 <ali1234> it doesn't really protect against either very much afaict
  90 2012-07-01 01:31:03 <ali1234> if either is feasible for a single hashing then it's feasible for two as well
  91 2012-07-01 01:31:25 <jrmithdobbs> eventually, but not necessarily immediately
  92 2012-07-01 01:31:31 <ali1234> unless it's right on the cusp of feasibility
  93 2012-07-01 01:31:40 <ali1234> like you're willing to wait 10 years but not 20
  94 2012-07-01 01:31:51 <jrmithdobbs> which would theorhetically give you time to switch out primitives in the protocol
  95 2012-07-01 01:32:09 <jrmithdobbs> s/would/could/
  96 2012-07-01 01:33:04 <jrmithdobbs> (besides the fact that you need to break it well enough to fake merkle trees if you want to do anything interesting without having to create blocks)
  97 2012-07-01 01:34:27 <jrmithdobbs> or if you can collide in <10min 1% or more of the time
  98 2012-07-01 01:34:37 <jrmithdobbs> at that point it falls apart of course
  99 2012-07-01 01:35:02 <jrmithdobbs> not even 1%
 100 2012-07-01 01:35:37 <ali1234> doubling the complexity is not going to buy you much time
 101 2012-07-01 01:35:43 rdponticelli has quit (Ping timeout: 246 seconds)
 102 2012-07-01 01:36:01 <luke-jr> gmaxwell: yep
 103 2012-07-01 01:36:11 <galambo> i think the greatest threat is from integer overruns (the only successful exploit so far)
 104 2012-07-01 01:36:36 <luke-jr> gmaxwell: it's too bad Linux doesn't have some kind of special log file type that just saves the last n MB
 105 2012-07-01 01:39:47 random_cat__ has quit (Ping timeout: 276 seconds)
 106 2012-07-01 01:40:18 <JFK911> logs belong in databases anyway
 107 2012-07-01 01:40:26 MobiusL has quit (Ping timeout: 276 seconds)
 108 2012-07-01 01:40:29 <luke-jr> um, no thanks
 109 2012-07-01 01:40:44 <JFK911> youve never handled serious logs before.
 110 2012-07-01 01:40:56 random_cat__ has joined
 111 2012-07-01 01:41:00 <luke-jr> sure I have
 112 2012-07-01 01:41:03 <gmaxwell> Yea ... no kidding.  In any case, every other daemon manages to have rotatable logs without issue.
 113 2012-07-01 01:41:16 <luke-jr> gmaxwell: bitcoind has log rotation on Gentoo :P
 114 2012-07-01 01:41:24 MobiusL has joined
 115 2012-07-01 01:41:42 <jrmithdobbs> gmaxwell: svlogd
 116 2012-07-01 01:41:42 <luke-jr> I was just thinking, instead of having it cut at a random line
 117 2012-07-01 01:42:05 <jrmithdobbs> gmaxwell: because no, they don't, so much crap software re: log rotation
 118 2012-07-01 01:42:31 <jgarzik> bitcoind reopens on SIGHUP now.  easy log rotation.
 119 2012-07-01 01:42:36 <luke-jr> sadly, jrmithdobbs is right when it comes to using syslog
 120 2012-07-01 01:42:56 <luke-jr> syslog with debug-level detail can really kill performance
 121 2012-07-01 01:43:18 EricLombrozo has joined
 122 2012-07-01 01:43:29 <EricLombrozo> is there a fee anomaly in block 0000000000000600498a139cfff8940a434e2351c418f8fd566236535b487ddd ?
 123 2012-07-01 01:43:56 <EricLombrozo> the miner is claiming .1341 BTC in fees but I'm adding up a total of .1346 BTC
 124 2012-07-01 01:44:09 <luke-jr> ouch
 125 2012-07-01 01:44:09 <EricLombrozo> I have two questions:
 126 2012-07-01 01:44:16 <EricLombrozo> 1) is my calculation wrong?
 127 2012-07-01 01:44:27 <EricLombrozo> and 2) what happens if the miner claims less in fees than is entitled to?
 128 2012-07-01 01:44:31 <gmaxwell> EricLombrozo: nothing requires a miner to claim all the value.
 129 2012-07-01 01:44:39 <gmaxwell> And people have intentionally claimed lets in the past.
 130 2012-07-01 01:44:48 <EricLombrozo> so the bitcoins just vanish?
 131 2012-07-01 01:44:59 <gmaxwell> Yes.
 132 2012-07-01 01:45:18 <EricLombrozo> so in principle there could be less than 21 million bitcoins in the end
 133 2012-07-01 01:45:25 <luke-jr> there will be
 134 2012-07-01 01:45:30 <gmaxwell> EricLombrozo: There always would be less than 21 million.
 135 2012-07-01 01:46:07 <luke-jr> EricLombrozo: if it's really off, that suggests a bug in Eloipool :|
 136 2012-07-01 01:46:10 <gmaxwell> 21 million is the limit as time goes to infinity and it would even underrun that because the coin precision is finite.
 137 2012-07-01 01:46:14 <luke-jr> EricLombrozo: care to help find it?
 138 2012-07-01 01:46:17 <EricLombrozo> the geometric series converges on 21 million if we assume satoshis are infinitely divisible
 139 2012-07-01 01:46:32 <EricLombrozo> but since satoshis are indivisible, we get less than 21 million
 140 2012-07-01 01:46:35 <jgarzik> hrm
 141 2012-07-01 01:46:36 <luke-jr> EricLombrozo: http://gitorious.org/bitcoin/eloipool
 142 2012-07-01 01:46:39 <jgarzik> that's... odd
 143 2012-07-01 01:46:40 <jgarzik> receive version message: version 31800, blocks=186970, us=127.0.0.1:8333, them=127.0.0.1:8333, peer=81.106.112.56:54136
 144 2012-07-01 01:46:43 <jgarzik> 127.0.0.1?
 145 2012-07-01 01:47:12 <jgarzik> receive version message: version 60000, blocks=186975, us=0.0.0.0:0, them=0.0.0.0:0, peer=77.247.181.165:13805
 146 2012-07-01 01:47:22 <jrmithdobbs> gmaxwell: http://pastebin.com/kT7XPfVh
 147 2012-07-01 01:47:42 <gmaxwell> jgarzik: current nodes send 0:0 when they don't know a routable IP for themselves.
 148 2012-07-01 01:48:01 <jgarzik> gmaxwell: hopefully we check for non-routeable addresses and not just 0.0
 149 2012-07-01 01:48:01 <jrmithdobbs> jgarzik: this is much easier because svlogd lets you rotate on size/time/etc
 150 2012-07-01 01:48:14 <jrmithdobbs> jgarzik: nice that logrotate will work now tho
 151 2012-07-01 01:48:23 <EricLombrozo> so if someone wanted to destroy a bunch of bitcoins, they could mine a two transaction block where the second transaction spends everything on fees and the first transaction (generation transaction) claims nothing
 152 2012-07-01 01:48:37 <gmaxwell> jgarzik: We do everywhere we'd need to check, though those particular messages don't do anything important today.
 153 2012-07-01 01:49:16 <gmaxwell> EricLombrozo: or they could forward coins to 1BitcoinEaterAddressDontSendf59kuE and save all the mining trouble.
 154 2012-07-01 01:49:29 <EricLombrozo> or right - they could send it to a black hole address
 155 2012-07-01 01:49:51 <gmaxwell> Or write an unredeemable script.
 156 2012-07-01 01:49:54 <jrmithdobbs> there's already lots of lost coins
 157 2012-07-01 01:50:00 <jgarzik>         "version" : 31800,
 158 2012-07-01 01:50:00 <jgarzik>         "subver" : "/BitCoinJ:0.5.1/Bitcoin Wallet:2.17/",
 159 2012-07-01 01:50:04 <jgarzik> so the weirdos are bitcoinj
 160 2012-07-01 01:50:07 <jgarzik> thank you getpeerinfo
 161 2012-07-01 01:50:10 <EricLombrozo> but there's a difference - an unredeemable script still shows an amount in the block chain
 162 2012-07-01 01:50:25 <Karmaon> Does anyone know how I would determine which pool a block header submitted by a miner is valid for?
 163 2012-07-01 01:50:29 <jrmithdobbs> EricLombrozo: there's even some not-fully-redeemed coinbase transactions in some blocks
 164 2012-07-01 01:50:30 <Karmaon> for example, a miner doing round robin.
 165 2012-07-01 01:50:47 <jrmithdobbs> EricLombrozo: as in, the 50/25 you get for finding the block
 166 2012-07-01 01:50:48 <gmaxwell> EricLombrozo: Thats a distinction, for most purposes I don't know that there is a difference.
 167 2012-07-01 01:51:22 <gmaxwell> jrmithdobbs: well, one. I don't think there is more than mindnightmagic's.
 168 2012-07-01 01:51:32 <jrmithdobbs> gmaxwell: i thought there was 2
 169 2012-07-01 01:51:40 <EricLombrozo> is there a good reason why the protocol shouldn't enforce exact calculation of coinbase and fees?
 170 2012-07-01 01:51:54 <gmaxwell> There are two that destroy coin by virtue of having duplicate coinbases.
 171 2012-07-01 01:51:56 <jrmithdobbs> gmaxwell: thought luke or someone had a bug for like a couple hours
 172 2012-07-01 01:51:59 <jgarzik> hrm
 173 2012-07-01 01:52:05 <jrmithdobbs> oh ya, that was different
 174 2012-07-01 01:52:05 * jgarzik looks at BlueMatt...
 175 2012-07-01 01:52:07 <jgarzik> [jgarzik@eu3 ~]$ grep -c 'socket send buffer full warning' /spare/bitcoin/data/debug.log
 176 2012-07-01 01:52:07 <jgarzik> 384
 177 2012-07-01 01:52:11 Turingi has quit (Read error: Connection reset by peer)
 178 2012-07-01 01:52:20 <gmaxwell> EricLombrozo: because it would be pointless?
 179 2012-07-01 01:52:55 <jrmithdobbs> EricLombrozo: basically it's only going to happen because of software bugs or people being asshats (the midnightmagic block) ;p
 180 2012-07-01 01:53:12 <gmaxwell> jrmithdobbs: asshats is the wrong word, it doesn't hurt anyone.
 181 2012-07-01 01:53:29 <jgarzik> 524 / 5569 log lines in git HEAD are this send buffer warning
 182 2012-07-01 01:53:33 <EricLombrozo> also, I am of the opinion that block height should be stored in the block headers and amounts should be stored in transaction inputs. it is a minimal cost in space for a significant improvement in performance and simplicity in code logic
 183 2012-07-01 01:53:40 <jrmithdobbs> gmaxwell: no, he specifically did it because he wanted the total # of coins off by some .35149685207 random-ish number iirc
 184 2012-07-01 01:53:47 <jrmithdobbs> gmaxwell: he's not an asshat, but that is asshatery
 185 2012-07-01 01:53:50 <jrmithdobbs> heh
 186 2012-07-01 01:54:01 <jgarzik> 805 / 5949
 187 2012-07-01 01:54:12 <jrmithdobbs> midnightmagic: <3
 188 2012-07-01 01:54:19 <gmaxwell> jrmithdobbs: he wanted to toss 1 satoshi to signify the absense of satoshi from bitcoin.
 189 2012-07-01 01:54:40 <jrmithdobbs> close enough, still qualifies ;p
 190 2012-07-01 01:54:42 <gmaxwell> But he goofed up and discarded 0.02000001 instead.
 191 2012-07-01 01:55:41 <EricLombrozo> 8 additional bytes in the header, 8 additional bytes per input isn't going to add significantly to block sizes
 192 2012-07-01 01:55:56 <jgarzik> 1547 / 6834
 193 2012-07-01 01:55:59 <EricLombrozo> and it would make it far easier to reject malformed blocks
 194 2012-07-01 01:56:02 <BlueMatt> jgarzik: hmm...well should probably not print that line unless fDebug then
 195 2012-07-01 01:56:22 <jgarzik> BlueMatt: yeah, it's 99% of the log lines on this public node
 196 2012-07-01 01:56:32 <EricLombrozo> besides, blocks can be compressed when stored on disk - and the block frequency isn't so high as to make it a bandwidth issue
 197 2012-07-01 01:56:53 <gmaxwell> EricLombrozo: We will be putting the height in the coinbase txn. (where it will serve a useful purpose)
 198 2012-07-01 01:56:55 <BlueMatt> jgarzik: hmm...probably wanna decrease default send buffer even further and not print that line
 199 2012-07-01 01:57:14 <gmaxwell> EricLombrozo: putting values in the inputs doesn't help anything, because you'd still need to lookup the inputs to check the values.
 200 2012-07-01 01:58:02 <EricLombrozo> gmaxwell, you still need to look up the inputs to verify values and check the scripts - but at least you could tentatively compute fees and see if they match what the miner claims...or if the miner claims an excess
 201 2012-07-01 01:58:06 <gmaxwell> EricLombrozo: also, 8 bytes per input is a significant increase in transaction size.
 202 2012-07-01 01:58:15 <jgarzik> BlueMatt: surely it will print that warning whenever it needs to throttle -- a normal occurence on a network?
 203 2012-07-01 01:58:22 EricLombrozo has quit ()
 204 2012-07-01 01:58:48 EricLombrozo has joined
 205 2012-07-01 01:58:58 <BlueMatt> jgarzik: yep, or, more simply, whenever the remote node is downloading a bunch of blocks
 206 2012-07-01 01:59:01 <EricLombrozo> starting when will the coinbase be storing height?
 207 2012-07-01 01:59:11 <BlueMatt> s/simply/the most common one, Id think/
 208 2012-07-01 01:59:17 <gmaxwell> EricLombrozo: If you're going to mine an invalid block you can happily set those high enough so that test passes and you don't fail until you look up the inputs. It achieves _nothing_ against a malicious party except increasing transaction size by 5% or so.
 209 2012-07-01 02:00:07 <gmaxwell> EricLombrozo: most likely 0.7.0, though it won't be enforced as a rule until later.
 210 2012-07-01 02:01:32 paraipan has quit (Remote host closed the connection)
 211 2012-07-01 02:03:08 <jrmithdobbs> oh those run scripts remind me, need to rebuild and add onion transit
 212 2012-07-01 02:03:19 * jgarzik files an issue
 213 2012-07-01 02:04:30 <gmaxwell> jrmithdobbs: addnode=6hgmaxwellgpv2oe.onion
 214 2012-07-01 02:04:31 <gribble> New news from bitcoinrss: jgarzik opened issue 1544 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1544>
 215 2012-07-01 02:04:46 <jrmithdobbs> gmaxwell: gonna be a bit, i don't actually have tor running on that box atm
 216 2012-07-01 02:05:07 <jrmithdobbs> gmaxwell: but thanks, need a test node ;p
 217 2012-07-01 02:05:43 <jrmithdobbs> gmaxwell: did you seriously bruteforce a vanity onion addr? heh
 218 2012-07-01 02:05:53 <jrmithdobbs> or can those be done more easily?
 219 2012-07-01 02:06:08 <gmaxwell> It computationally easier than bitcoin addresses.
 220 2012-07-01 02:06:25 <jrmithdobbs> doable on cpu?
 221 2012-07-01 02:06:34 <gmaxwell> that one took something like 60 cpu days.
 222 2012-07-01 02:06:37 <gmaxwell> IIRC
 223 2012-07-01 02:06:50 <jrmithdobbs> got the code up somewhere? ;p
 224 2012-07-01 02:07:02 <gmaxwell> https://github.com/katmagic/Shallot
 225 2012-07-01 02:08:04 <gmaxwell> you might want to fix a write-past-the-end-of-a-buffer bug in it. It also has some heisenbug when there are more than 24 cores or so. I wasn't easily able to track it down but running it inside gdb made it hide.
 226 2012-07-01 02:08:14 hnz has quit (Ping timeout: 252 seconds)
 227 2012-07-01 02:08:46 <jrmithdobbs> gmaxwell: i'm only going for 4 chars, don't care position, and not gonna run it on that many cores ;p
 228 2012-07-01 02:10:01 TuxBlackEdo is now known as TuxTest
 229 2012-07-01 02:10:21 <gmaxwell> yea, thats about instant.. actuall 4 chars pinned to the beginning or end is about instant.
 230 2012-07-01 02:10:22 TuxTest is now known as TuxBlackEdo
 231 2012-07-01 02:10:28 Joric has joined
 232 2012-07-01 02:10:42 <gmaxwell> sipa: is your hidden service node down?
 233 2012-07-01 02:11:13 Joric has quit (Remote host closed the connection)
 234 2012-07-01 02:11:17 * luke-jr stabs Linux
 235 2012-07-01 02:11:33 <gmaxwell> luke-jr: leapsecond knock out one of your hosts?
 236 2012-07-01 02:11:43 <luke-jr> gmaxwell: is that what did it? O.o
 237 2012-07-01 02:12:06 <luke-jr> gmaxwell: one of my GPUs died, and now Linux is turning any process touchign the PCI bus into an unkillable zombie
 238 2012-07-01 02:12:10 hnz has joined
 239 2012-07-01 02:12:24 <gmaxwell> luke-jr: some kernels have a deadlock inserting the leapsecond.
 240 2012-07-01 02:12:31 <gmaxwell> Jun 30 23:59:59 carbide80 kernel: [1168941.330709] Clock: inserting leap second 23:59:60 UTC
 241 2012-07-01 02:12:35 <luke-jr> O.o
 242 2012-07-01 02:12:57 <luke-jr> Freezing of tasks failed after 20.00 seconds (4 tasks refusing to freeze, wq_busy=0):
 243 2012-07-01 02:13:34 <gmaxwell> http://serverfault.com/questions/403732/anyone-else-experiencing-high-rates-of-linux-server-crashes-today
 244 2012-07-01 02:14:04 EricLombrozo has quit ()
 245 2012-07-01 02:14:29 <luke-jr> gmaxwell: no sign of it in my system
 246 2012-07-01 02:16:29 <gmaxwell> jgarzik: is height in getpeerinfo the starting height or the current one?
 247 2012-07-01 02:16:31 <jrmithdobbs> gmaxwell: don't need any special build options to enable onion/tor/socks?
 248 2012-07-01 02:16:50 <gmaxwell> jrmithdobbs: nope.
 249 2012-07-01 02:16:59 <gmaxwell> jrmithdobbs: just currentish git.
 250 2012-07-01 02:18:48 setkeh` has joined
 251 2012-07-01 02:18:52 setkeh has quit (Ping timeout: 252 seconds)
 252 2012-07-01 02:23:00 <BlueMatt> jgarzik: https://github.com/bitcoin/bitcoin/pull/1545
 253 2012-07-01 02:23:34 <gmaxwell> jrmithdobbs: I'm glad you're setting up, I want to try getpeerinfo with it and the other nodes are down.
 254 2012-07-01 02:24:50 <gribble> New news from bitcoinrss: TheBlueMatt opened pull request 1545 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1545>
 255 2012-07-01 02:37:24 wasabi1 has quit (Ping timeout: 245 seconds)
 256 2012-07-01 02:38:54 Zarutian has quit (Quit: Zarutian)
 257 2012-07-01 02:41:53 <DBordello> so, the qt client doesn't react well to ungraceful gracedown huh :/
 258 2012-07-01 02:41:54 [7] has quit (Disconnected by services)
 259 2012-07-01 02:42:02 TheSeven has joined
 260 2012-07-01 02:42:40 <gmaxwell> DBordello: care to elaborate?
 261 2012-07-01 02:44:36 <DBordello> gmaxwell, http://i.imgur.com/tfkO4.png
 262 2012-07-01 02:45:49 <DBordello> (this stupid computer has a driver problem and won't suspend cleanly)
 263 2012-07-01 02:46:22 <DBordello> and this is after I deleted the entire block chain earlier today
 264 2012-07-01 02:48:02 <gmaxwell> DBordello: yea, I think that happens if the last block referenced in the blockindex is past the end of your blockfile.
 265 2012-07-01 02:48:49 <DBordello> delete 'er all and restart?
 266 2012-07-01 02:49:31 <gmaxwell> DBordello: if you're running 0.7 you can loadblock= instead, but if not... then yes.. delete the blocks file and the databases and restart. (presumably you have a recent wallet backup)
 267 2012-07-01 02:57:48 <DBordello> hmmm, we will try restarting the whole prces
 268 2012-07-01 02:58:44 sgornick has joined
 269 2012-07-01 03:01:04 e0s_ has quit (Read error: Connection reset by peer)
 270 2012-07-01 03:08:17 Z0rZ0rZ0r has quit (Ping timeout: 248 seconds)
 271 2012-07-01 03:09:14 Z0rZ0rZ0r has joined
 272 2012-07-01 03:17:40 twobitcoins has quit (Quit: Leaving)
 273 2012-07-01 03:18:12 <Tril> how do i access testnet on master, it tries to go to some testnet3 thing and finds no peers
 274 2012-07-01 03:19:50 <gmaxwell> Tril: Testnet has been replaced with testnet3. You should be finding peers however.
 275 2012-07-01 03:20:27 <gmaxwell> My testnet3 node has 11 peers right now.
 276 2012-07-01 03:24:58 <Tril> gmaxwell: yeah it's working now, not sure what happened. it was in a loop before, trying the same peer repeatedly.
 277 2012-07-01 03:27:58 ivan\ has quit (Quit: ERC Version 5.3 (IRC client for Emacs))
 278 2012-07-01 03:28:54 twobitcoins has joined
 279 2012-07-01 03:37:39 <jgarzik> gmaxwell: nStartingHeight
 280 2012-07-01 03:38:37 <gmaxwell> yea. I went and looked eventually. I suspect this is going to cause users to spaz out thinking they have peers which are behind on the chain.
 281 2012-07-01 03:39:52 <jgarzik> gmaxwell: I don't think we track that, so the options would be (a) track it, imperfectly or (b) ditch it
 282 2012-07-01 03:40:20 <jgarzik> we just know what they report in 'version', and what they request, AFAICS
 283 2012-07-01 03:40:56 <gmaxwell> or just call it "initialheight" or "startingheight" in the output.
 284 2012-07-01 03:41:07 <gmaxwell> Which is what I thought your patch did initially.
 285 2012-07-01 03:41:53 <jgarzik> true.  I can change it back to startingheight
 286 2012-07-01 03:42:14 <jgarzik> if that will not be just as confusing to users
 287 2012-07-01 03:43:04 <gmaxwell> I think that would have less potential for confusion, and the starting height is at least somewhat useful. (though its more useful in combination with the current height)
 288 2012-07-01 03:49:56 theodore has quit (Quit: theodore)
 289 2012-07-01 04:00:49 ivan\ has joined
 290 2012-07-01 04:08:36 <jgarzik> "services" : "00000000"
 291 2012-07-01 04:08:47 <jgarzik> so bitcoinj does not assert NODE_NETWORK... makes sense
 292 2012-07-01 04:16:10 Guest513 has quit (Changing host)
 293 2012-07-01 04:16:10 Guest513 has joined
 294 2012-07-01 04:16:20 Guest513 is now known as SomeoneWeird
 295 2012-07-01 04:17:35 luke-jr has quit (Remote host closed the connection)
 296 2012-07-01 04:23:19 luke-jr has joined
 297 2012-07-01 04:24:44 theodore has joined
 298 2012-07-01 04:26:34 nanook7 has joined
 299 2012-07-01 04:30:13 Joric has joined
 300 2012-07-01 04:31:15 nanook7 has quit (Client Quit)
 301 2012-07-01 04:52:40 <midnightmagic> I fail to see why people think underclaiming is so bad when there are thousands of bitcoins completely lost to the void. What, being able to see what happened offends you more than knowing it happened?
 302 2012-07-01 04:54:02 <midnightmagic> If I want to burn my money in my backyard, who gives a !$%@? Welcome to a deflationary currency.
 303 2012-07-01 04:54:04 <midnightmagic> Asshats.
 304 2012-07-01 04:55:05 <galambo> lol
 305 2012-07-01 04:55:13 <gmaxwell> midnightmagic: 'people' ?
 306 2012-07-01 04:55:37 <gmaxwell> midnightmagic: I think the asshat comment wasn't claiming you were doing wrong just that you were being silly.
 307 2012-07-01 04:56:20 <galambo> jrmithdobbs> midnightmagic: <3
 308 2012-07-01 04:56:21 <midnightmagic> gmaxwell: This isn't the first time someone has called me names for underclaiming both the block reward and the tx fees, this is just the first time it's annoyed me in a while. Apologies for ranting, it's not directed at you.
 309 2012-07-01 04:57:51 <galambo> what about the guys that send bitcoins to the address from the genesis block ...
 310 2012-07-01 04:57:57 <midnightmagic> I did see the <3 but the name-calling is unnecessary and just makes me want to actually go through with all the things I've been contemplating the last year or so.
 311 2012-07-01 04:58:47 <midnightmagic> galambo: There's a whole thread on bitcointalk dedicated to verified- and claimed-destroyed bitcoins that are permanently lost.
 312 2012-07-01 04:58:52 <midnightmagic> lemme see if I can dig it up..
 313 2012-07-01 05:00:35 <midnightmagic> https://bitcointalk.org/index.php?topic=7253
 314 2012-07-01 05:00:42 <galambo> i have a confession to make i lost about 10 bitcoins forever on the testnet when i got them from the faucet and accidentally wrote over my wallet.dat
 315 2012-07-01 05:02:15 <Joric> good guy galambo
 316 2012-07-01 05:02:41 <midnightmagic> testnet's being reset again anyway
 317 2012-07-01 05:03:22 <D34TH> whats this?
 318 2012-07-01 05:03:28 <D34TH> someone needs testnet coins?
 319 2012-07-01 05:03:30 <D34TH> addr
 320 2012-07-01 05:04:03 <Joric> i usually run testnet-in-a-box for all kinds of tests
 321 2012-07-01 05:04:11 <D34TH> i use -testnet
 322 2012-07-01 05:04:12 <D34TH> D:
 323 2012-07-01 05:04:47 <D34TH> Balance: 1185.58409643
 324 2012-07-01 05:05:46 <freewil> i made this with a Makefile to try to make life easy... https://github.com/freewil/bitcoin-testnet-box
 325 2012-07-01 05:06:41 <galambo> i switched to regular bitcoin for testing because testnet3 is too hard to bootstrap for me
 326 2012-07-01 05:07:03 <galambo> once i have actual tests ill switch back :)
 327 2012-07-01 05:07:35 <galambo> for now the test is (still accepting blocks? yep.)
 328 2012-07-01 05:10:18 <Joric> lol @ http://blockexplorer.com/address/1BitcoinEaterAddressDontSendf59kuE
 329 2012-07-01 05:10:32 <Joric> Received BTC: 0.159
 330 2012-07-01 05:11:32 <osmosis> ya
 331 2012-07-01 05:11:34 <osmosis> ha
 332 2012-07-01 05:12:26 <midnightmagic> lol it would amuse me to NO end if someone actually had a private key for that.. somehow..
 333 2012-07-01 05:12:39 <midnightmagic> although I guess generating a keypair like that would be pretty difficult
 334 2012-07-01 05:14:55 Joric has quit (Ping timeout: 245 seconds)
 335 2012-07-01 05:17:02 <gmaxwell> midnightmagic: certantly you can get a more improbable address if you're willing to expand your definition of improbable.. but in that one the full 160 bits is eaten up by ascii...
 336 2012-07-01 05:17:20 <gmaxwell> er. I mean english.
 337 2012-07-01 05:20:27 <midnightmagic> :-)
 338 2012-07-01 05:20:54 <midnightmagic> running pycryptopp and python-ed25519 benchmarks on my raspberry pi units. this should be fun!
 339 2012-07-01 05:21:13 RainbowDashh has joined
 340 2012-07-01 05:21:55 <galambo> it would amuse me more if satoshi had a private key to 00000000000000000 and a transaction list that collided with the merkle root in the genesis block
 341 2012-07-01 05:21:57 <midnightmagic> i'm pretty sure bitcoind won't run on a raspberry pi anymore.
 342 2012-07-01 05:22:13 <midnightmagic> lol
 343 2012-07-01 05:22:23 <midnightmagic> that would similarly awesome.
 344 2012-07-01 05:22:25 * galambo raises eyebrow
 345 2012-07-01 05:22:57 <midnightmagic> what?
 346 2012-07-01 05:23:54 <galambo> im sorry i said that wrong
 347 2012-07-01 05:24:29 <galambo> a block that hash ed to 00000000000000
 348 2012-07-01 05:24:53 <galambo> and although that probably wouldnt work for some reason i cant think of right now
 349 2012-07-01 05:25:12 <midnightmagic> that would tell me sha256 was broken and we should change hash algo
 350 2012-07-01 05:25:36 <gmaxwell> galambo: What you're saying makes no sense— it wouldn't do anything.
 351 2012-07-01 05:26:39 <galambo> i know but people make a big deal about "premined" coins and i thought it was a funny idea
 352 2012-07-01 05:29:00 <gmaxwell> People do? huh? in bitcoin?
 353 2012-07-01 05:29:21 RainbowDashh has quit (Quit: SLEEP MODE.  [NSFW!]  [WIERD] [WTF] http://pastebin.com/6L0WjKRk)
 354 2012-07-01 05:29:32 <galambo> whenever they talk about poorly conceived alt chains
 355 2012-07-01 05:34:23 <gmaxwell> ah.
 356 2012-07-01 05:35:23 <gmaxwell> I did at one point run into someone on IRC who thought there was some possibility that there existed premined coins in bitcoin, and I think I was unable to shake him from this position.  For a moment I thought you were saying it was more than one guy.
 357 2012-07-01 05:37:14 RainbowDashh has joined
 358 2012-07-01 05:37:16 Joric has joined
 359 2012-07-01 05:37:32 RainbowDashh has quit (Client Quit)
 360 2012-07-01 05:39:00 gfinn has quit (Ping timeout: 276 seconds)
 361 2012-07-01 05:39:05 word has quit (Ping timeout: 252 seconds)
 362 2012-07-01 05:39:37 theodore has quit (Quit: theodore)
 363 2012-07-01 05:41:32 word has joined
 364 2012-07-01 05:43:01 copumpkin is now known as eigenpumpkin
 365 2012-07-01 05:46:26 MrTiggr has quit (Quit: getting bouncy)
 366 2012-07-01 05:47:38 Guest64211 has quit (Changing host)
 367 2012-07-01 05:47:38 Guest64211 has joined
 368 2012-07-01 05:47:47 RainbowDashh has joined
 369 2012-07-01 05:48:06 Guest64211 is now known as MrTiggr
 370 2012-07-01 05:49:22 <midnightmagic> gmaxwell: some people incorrectly describe the early miners as "preminers", implication being they deliberately took advantage of the early difficulty and hoarded for the long-term
 371 2012-07-01 05:49:23 maaku has joined
 372 2012-07-01 05:51:19 <midnightmagic> gmaxwell: which is exactly what some of them did, but..  premining? ponzi scheme? unjust enrichment? sigh. I'm not surprised at all Art never came back, for all the times he had to explain the same thing over and over again.
 373 2012-07-01 05:51:45 RainbowDashh has quit (Client Quit)
 374 2012-07-01 05:54:18 <gmaxwell> midnightmagic: The person I was speaking about wasn't especially technical and just wouldn't accept that it isn't possible for there to exist an unbounded amount of unknown coin.
 375 2012-07-01 05:58:09 <midnightmagic> gmaxwell: I'm a little surprised we don't get so many of those, these days given the massively growing popularity of btc.
 376 2012-07-01 05:59:04 <midnightmagic> gmaxwell: for python-ed25519 speed benchmarks, my rpi is slower than a >10 yr old pentium III
 377 2012-07-01 05:59:08 <midnightmagic> lol
 378 2012-07-01 05:59:22 <midnightmagic> how the hell is it doing full-screen 1080p if that's the case.
 379 2012-07-01 06:02:49 gfinn has joined
 380 2012-07-01 06:03:00 maaku has quit (Quit: maaku)
 381 2012-07-01 06:04:25 RainbowDashh has joined
 382 2012-07-01 06:08:05 Motest031 has quit (Read error: Operation timed out)
 383 2012-07-01 06:09:04 Motest003 has joined
 384 2012-07-01 06:09:50 devrandom has joined
 385 2012-07-01 06:22:33 MysteryBanshee has joined
 386 2012-07-01 06:24:14 RazielZ has joined
 387 2012-07-01 06:30:24 D34TH has quit (Read error: Connection reset by peer)
 388 2012-07-01 06:31:01 Lexa has quit (Quit: Lexa)
 389 2012-07-01 06:31:14 minimoose has quit (Quit: minimoose)
 390 2012-07-01 06:40:24 theodore has joined
 391 2012-07-01 06:51:06 RainbowDashh has quit (Ping timeout: 246 seconds)
 392 2012-07-01 06:52:31 Joric has quit ()
 393 2012-07-01 06:54:11 RainbowDashh has joined
 394 2012-07-01 07:08:51 molecular has quit (Ping timeout: 264 seconds)
 395 2012-07-01 07:09:37 molecular has joined
 396 2012-07-01 07:14:29 RainbowDashh has quit (Disconnected by services)
 397 2012-07-01 07:14:29 RainbowD_ has joined
 398 2012-07-01 07:29:24 brwyatt is now known as brwyatt|Away
 399 2012-07-01 07:30:10 theodore has quit (Quit: theodore)
 400 2012-07-01 07:33:57 osmosis has quit (Quit: Leaving)
 401 2012-07-01 07:38:34 theodore has joined
 402 2012-07-01 07:43:46 PK has joined
 403 2012-07-01 07:49:55 RainbowD_ is now known as RainbowDashh
 404 2012-07-01 08:15:44 theodore has quit (Quit: theodore)
 405 2012-07-01 08:21:36 OneFixt_ has joined
 406 2012-07-01 08:23:13 Diablo-D3 has joined
 407 2012-07-01 08:25:13 OneFixt has quit (Ping timeout: 246 seconds)
 408 2012-07-01 08:42:42 OneFixt_ is now known as OneFixt
 409 2012-07-01 08:46:43 Z0rZ0rZ0r has quit (Quit: Leaving)
 410 2012-07-01 08:59:45 da2ce7 has joined
 411 2012-07-01 09:02:45 Joric has joined
 412 2012-07-01 09:22:04 da2ce7 has quit (Ping timeout: 240 seconds)
 413 2012-07-01 09:24:06 Prattler has joined
 414 2012-07-01 09:26:45 sirk390 has joined
 415 2012-07-01 09:28:59 imsaguy has quit (Remote host closed the connection)
 416 2012-07-01 09:44:32 imsaguy has joined
 417 2012-07-01 09:58:27 imsaguy has quit (Remote host closed the connection)
 418 2012-07-01 10:07:18 sytse has quit (Ping timeout: 245 seconds)
 419 2012-07-01 10:07:29 Joric has quit ()
 420 2012-07-01 10:08:07 sytse has joined
 421 2012-07-01 10:23:28 <PK> can someone explain me the logic of this? why do I have a negative account value? http://www.pastebin.ca/2165989
 422 2012-07-01 10:23:41 imsaguy has joined
 423 2012-07-01 10:31:21 mxmxmx has left ()
 424 2012-07-01 10:36:20 sacredchao has joined
 425 2012-07-01 10:41:22 datagutt has joined
 426 2012-07-01 10:43:26 Turingi has joined
 427 2012-07-01 10:43:26 Turingi has quit (Changing host)
 428 2012-07-01 10:43:26 Turingi has joined
 429 2012-07-01 10:45:56 t7 has joined
 430 2012-07-01 10:50:25 TD has joined
 431 2012-07-01 10:53:01 Tykling has quit (Excess Flood)
 432 2012-07-01 10:54:23 Tykling has joined
 433 2012-07-01 10:56:49 Tykling has quit (Excess Flood)
 434 2012-07-01 10:57:57 Tykling has joined
 435 2012-07-01 10:57:59 ageis has quit (Quit: http://ageispolis.net)
 436 2012-07-01 10:58:09 Tykling has quit (Excess Flood)
 437 2012-07-01 10:59:04 ageis has joined
 438 2012-07-01 10:59:58 <doublec> PK: you sent coins from the default account, "", which put it in negative
 439 2012-07-01 10:59:58 BeTep has quit (Read error: Connection reset by peer)
 440 2012-07-01 11:00:19 <doublec> PK: maybe you used "sendtoaddress" instead of "sendfrom"
 441 2012-07-01 11:00:46 tower has quit (Remote host closed the connection)
 442 2012-07-01 11:01:16 tower has joined
 443 2012-07-01 11:01:31 weather has joined
 444 2012-07-01 11:01:51 <TD> good day
 445 2012-07-01 11:04:53 Tykling has joined
 446 2012-07-01 11:05:12 Prattler has quit (Remote host closed the connection)
 447 2012-07-01 11:05:20 weather is now known as BeTep
 448 2012-07-01 11:06:30 <PK> doublec: yes, I did.
 449 2012-07-01 11:08:05 <PK> doublec: if I don't use "sendfrom" is there any way to find out where I sent it from later? I added address with the 50 BTC from "" to "Account 1" after sending the coins.
 450 2012-07-01 11:13:26 <PK> doublec: it stored only the fromaccount, not the from address, either I don't understand how bitcoin works or that seems rather flaky
 451 2012-07-01 11:13:49 <PK> and goodday TD, welcome to the club of twoletter nicks.
 452 2012-07-01 11:15:27 * TD has been in that club for a long time :)
 453 2012-07-01 11:21:05 paraipan has joined
 454 2012-07-01 11:41:22 ThomasV has joined
 455 2012-07-01 11:41:57 <doublec> PK: I don't generally use the accounts feature
 456 2012-07-01 11:42:16 <doublec> PK: it's mainly there for web sites that want to use bitcoind to manage user account balance I believe
 457 2012-07-01 11:45:36 paraipan has quit (Ping timeout: 276 seconds)
 458 2012-07-01 11:45:42 <sipa> PK: if you don't use sendfrom, it is always sent from ""; no need to find it out later
 459 2012-07-01 11:46:03 <sipa> but note that that has nothing to do with the actual coins selected for transfer
 460 2012-07-01 11:46:10 <PK> sipa: yes, but if I move the address into another account, then I have a negative value
 461 2012-07-01 11:46:26 paraipan has joined
 462 2012-07-01 11:46:58 <sipa> you'll need to do a move to change the account balance as well
 463 2012-07-01 11:48:57 sirk390 has quit (Quit: Leaving.)
 464 2012-07-01 11:50:47 sirk390 has joined
 465 2012-07-01 11:52:47 t7 has quit (Ping timeout: 246 seconds)
 466 2012-07-01 11:59:47 d4de has quit (Ping timeout: 246 seconds)
 467 2012-07-01 12:15:20 ThomasV has quit (Quit: Quitte)
 468 2012-07-01 12:21:53 RainbowDashh has quit (Quit: SLEEP MODE.  [NSFW!]  [WIERD] [WTF] http://pastebin.com/6L0WjKRk)
 469 2012-07-01 12:28:48 Tykling has quit (Ping timeout: 272 seconds)
 470 2012-07-01 12:35:46 Tykling has joined
 471 2012-07-01 12:38:36 wasabi1 has joined
 472 2012-07-01 12:41:32 Tykling has quit (Excess Flood)
 473 2012-07-01 12:41:34 Diablo-D3 has quit (Remote host closed the connection)
 474 2012-07-01 12:42:01 Diablo-D3 has joined
 475 2012-07-01 12:52:53 Tykling has joined
 476 2012-07-01 13:04:17 Diapolo has joined
 477 2012-07-01 13:07:56 Tykling has quit (Excess Flood)
 478 2012-07-01 13:16:36 setkeh` is now known as setkeh
 479 2012-07-01 13:16:57 sytse has quit (Ping timeout: 248 seconds)
 480 2012-07-01 13:17:04 sytse has joined
 481 2012-07-01 13:17:42 zebedee_ has joined
 482 2012-07-01 13:19:23 Tykling has joined
 483 2012-07-01 13:21:34 zebedee_ has quit (Client Quit)
 484 2012-07-01 13:21:43 zebedee_ has joined
 485 2012-07-01 13:47:21 guruvan_ has joined
 486 2012-07-01 13:49:17 Diapolo has left ()
 487 2012-07-01 13:52:09 darksk1ez has quit (Ping timeout: 248 seconds)
 488 2012-07-01 13:53:04 darksk1ez has joined
 489 2012-07-01 13:53:26 tsche has joined
 490 2012-07-01 13:54:12 <jrmithdobbs> wow, binary size significantly increased
 491 2012-07-01 13:54:31 <jrmithdobbs> compared to .3-ish
 492 2012-07-01 13:54:53 <jrmithdobbs> oh nm, forgot to strip and relink the debug info
 493 2012-07-01 13:57:05 leotreasure has joined
 494 2012-07-01 13:58:33 darksk1ez has quit (Ping timeout: 248 seconds)
 495 2012-07-01 14:06:11 dvide has joined
 496 2012-07-01 14:08:01 leotreasure has quit (Read error: Connection reset by peer)
 497 2012-07-01 14:08:21 leotreasure has joined
 498 2012-07-01 14:10:02 leotreasure has quit (Client Quit)
 499 2012-07-01 14:10:39 Clipse has quit (Ping timeout: 246 seconds)
 500 2012-07-01 14:12:26 Detritus has quit (Ping timeout: 246 seconds)
 501 2012-07-01 14:14:20 theorb has joined
 502 2012-07-01 14:14:37 theorbtwo has quit (Ping timeout: 252 seconds)
 503 2012-07-01 14:14:49 theorb is now known as theorbtwo
 504 2012-07-01 14:16:52 <jgarzik> jrmithdobbs: FWIW, if binary size really matters to you, you can carry the debug info in a separate file
 505 2012-07-01 14:18:24 sirk390 has quit (Quit: Leaving.)
 506 2012-07-01 14:20:43 <PK> if size matters to you, you should strip.... one of those sentences you should never use out of context.
 507 2012-07-01 14:26:17 Tykling has quit (Ping timeout: 248 seconds)
 508 2012-07-01 14:30:59 smtmnyz has joined
 509 2012-07-01 14:34:48 <jgarzik> interesting
 510 2012-07-01 14:35:16 <jgarzik> if two TX's are broadcast to a node at nearly the same time... is bitcoin smart enough to batch those together into a single 'inv', for offering to other nodes?
 511 2012-07-01 14:35:25 <jgarzik> I think it might be... checking code in a bit
 512 2012-07-01 14:42:47 TD has quit (Quit: TD)
 513 2012-07-01 14:44:57 rdponticelli has joined
 514 2012-07-01 14:45:22 Clipse has joined
 515 2012-07-01 14:46:53 Tykling has joined
 516 2012-07-01 14:49:00 Tykling has quit (Read error: Connection reset by peer)
 517 2012-07-01 14:51:20 eigenpumpkin is now known as copumpkin
 518 2012-07-01 14:52:16 Detritus has joined
 519 2012-07-01 14:53:54 Tykling has joined
 520 2012-07-01 14:54:56 graingert has joined
 521 2012-07-01 14:56:15 MC1984 has quit (Ping timeout: 264 seconds)
 522 2012-07-01 14:58:35 MC1984 has joined
 523 2012-07-01 14:59:39 copumpkin has quit (Ping timeout: 246 seconds)
 524 2012-07-01 15:00:11 copumpkin has joined
 525 2012-07-01 15:17:08 freewil has quit (Quit: Leaving)
 526 2012-07-01 15:21:32 TD has joined
 527 2012-07-01 15:23:25 PK has quit ()
 528 2012-07-01 15:28:37 sirk390 has joined
 529 2012-07-01 15:37:09 none_ has joined
 530 2012-07-01 15:37:37 <none_> getreceivedbyaddress only can access the amount received in a local wallet using the bitcoin api
 531 2012-07-01 15:38:02 <none_> how do I get the amount received by someone else's address
 532 2012-07-01 15:38:31 <none_> I am looking to do this in a command line way from a web service
 533 2012-07-01 15:41:54 <gmaxwell> You can not. The reference software doesn't index other people's addresses.
 534 2012-07-01 15:43:36 <luke-jr>  heh
 535 2012-07-01 15:43:41 <luke-jr> that use never even occurred to me XD
 536 2012-07-01 15:52:58 <jine> Jikes, the leap-second really messed things up.
 537 2012-07-01 15:53:16 <luke-jr> jine: ?
 538 2012-07-01 15:53:21 <jine> We got 80 duplicated deposits in our system.
 539 2012-07-01 15:53:34 <luke-jr> ouch
 540 2012-07-01 15:53:41 <luke-jr> hopefully on honest accounts?
 541 2012-07-01 15:54:01 <jine> First of all, java-applications hung - which cased poolserverj to crash/be really unstable
 542 2012-07-01 15:54:04 <jine> A reboot solved that
 543 2012-07-01 15:54:33 <jine> Secondly, mysql (or mariadb) hung and took 100% cpu after the leap-second (GMT) - which caused site and pool outages
 544 2012-07-01 15:54:37 <jine> A reboot solved that to
 545 2012-07-01 15:54:42 <luke-jr> weird
 546 2012-07-01 15:54:56 <luke-jr> you'd think software could handle stuff like this, considering NTP might set the clock back occasionally
 547 2012-07-01 15:55:16 <jine> Now - 17h later, i got a PM here on IRC with a question why he had such a high balance on his account - It was an old deposit that SOMEHOW got processed 3 times tonight.
 548 2012-07-01 15:55:22 <jine> Which left him with 6 BTC in his account.
 549 2012-07-01 15:56:05 <gmaxwell> luke-jr: other than at boot ntp will normally just slow the clock if its too fast... not step backwards.  (though the leap second isn't a jump backwards either)
 550 2012-07-01 15:56:12 JStoker has quit (Excess Flood)
 551 2012-07-01 15:56:15 <jine> I have about.. 70-80 deposits tonight, which have been duplicated. Strangly enough - no-one seems to have noticed it until now, cause noone has actually "stolen" anything.
 552 2012-07-01 15:56:18 <luke-jr> gmaxwell: it isn't?
 553 2012-07-01 15:56:34 <luke-jr> gmaxwell: I thought I read that Linux currently just sets the clock back a second at midnight
 554 2012-07-01 15:57:00 <TD> leap second handling is a PITA
 555 2012-07-01 15:57:13 <TD> it causes so many bugs that at google we simply lengthen seconds leading up to it
 556 2012-07-01 15:57:22 <TD> so most software doesn't actually see the leap second at all
 557 2012-07-01 15:57:24 <jine> The issue in linux (2.6 and 3.x) is that when ntp tells the kernel there is going to be a leap-second, it could result in a softlock
 558 2012-07-01 15:57:34 <gmaxwell> jine: thats not the only issue.
 559 2012-07-01 15:57:50 <luke-jr> TD: IIRC that was removed in 2.6.23 or something
 560 2012-07-01 15:57:59 <jine> But, that's not the issues I've seen or heard about (except a few cases) - it's mostly... strange issues, such as java and mysql (innodb only?) hangs.
 561 2012-07-01 15:58:09 <gmaxwell> luke-jr: yea, lots of debian systems out there with old kernels.
 562 2012-07-01 15:58:36 <luke-jr> gmaxwell: not that old!
 563 2012-07-01 15:58:38 <jine> Debian squeeze, latest stable, with latest amd64 kernel seems to be affected so
 564 2012-07-01 15:58:41 <jine> It's
 565 2012-07-01 15:58:42 <gmaxwell> luke-jr: hm, seems your right. In any case, ntp normally only steps the clock at bootup. During runtime if you get too fast it slews.
 566 2012-07-01 15:58:44 <jine> Linux pool 2.6.32-5-amd64
 567 2012-07-01 15:59:01 <gmaxwell> Debian stable = software from the mid 1920s.
 568 2012-07-01 15:59:18 <luke-jr> gmaxwell: Debian stable often has newer versions than Gentoo stable, these days
 569 2012-07-01 15:59:43 <jine> The diffrence with debian stable from other dists, is that it's actually IS stable.
 570 2012-07-01 15:59:45 <luke-jr> at least for a few months after releases <.<
 571 2012-07-01 15:59:53 <luke-jr> jine: yep
 572 2012-07-01 16:00:04 <luke-jr> I'm disappointed in the next-stable plans tho
 573 2012-07-01 16:00:25 <luke-jr> they're removing OpenVZ support, and apparently removing pure-amd64 (no multilib) support
 574 2012-07-01 16:00:56 <jine> Wiho, no more fake, buggy vz hosting!
 575 2012-07-01 16:01:02 <jine> Use kvm or xen ffs.
 576 2012-07-01 16:01:08 <jine> Or actual bsd jails.
 577 2012-07-01 16:01:09 <luke-jr> no, OpenVZ is far better
 578 2012-07-01 16:01:19 <jine> I'm not even going to respond to that.
 579 2012-07-01 16:01:20 <luke-jr> and not fake
 580 2012-07-01 16:01:42 <luke-jr> just because it implements a different (more common) use case doesn't make it fake
 581 2012-07-01 16:01:46 <jine> luke-jr: Fake as in companys sells it as "cloud hosting", while it is in fact just a jail.
 582 2012-07-01 16:01:59 <luke-jr> jine: complain to said companies
 583 2012-07-01 16:02:23 <luke-jr> besides, "cloud hosting" is just marketting speak with no real meaning anyway
 584 2012-07-01 16:03:33 none_ has quit (Ping timeout: 245 seconds)
 585 2012-07-01 16:03:55 darksk1ez has joined
 586 2012-07-01 16:06:37 brwyatt is now known as Away!~brwyatt@brwyatt.net|brwyatt
 587 2012-07-01 16:06:43 JStoker has joined
 588 2012-07-01 16:11:37 nanook7 has joined
 589 2012-07-01 16:13:17 <phungus> HyperV really whips the llama's ass!
 590 2012-07-01 16:14:10 <phungus> just kidding
 591 2012-07-01 16:14:35 rdponticelli has quit (Ping timeout: 246 seconds)
 592 2012-07-01 16:14:48 <sipa> I certainly hope hyperV doesn't require winamp
 593 2012-07-01 16:14:54 <luke-jr> lol
 594 2012-07-01 16:16:42 <MC1984> debian not so stable
 595 2012-07-01 16:16:49 <MC1984> oh well time for arch
 596 2012-07-01 16:17:15 Tykling has quit (Read error: Connection reset by peer)
 597 2012-07-01 16:17:45 darksk1ez has quit (Ping timeout: 248 seconds)
 598 2012-07-01 16:22:22 darksk1ez has joined
 599 2012-07-01 16:24:40 sirk390 has quit (Quit: Leaving.)
 600 2012-07-01 16:25:52 nanook7 has quit (Quit: Colloquy for iPhone - http://colloquy.mobi)
 601 2012-07-01 16:27:55 Tykling has joined
 602 2012-07-01 16:28:23 <phungus> hmm
 603 2012-07-01 16:28:27 <phungus> my debian squeeze was just fine
 604 2012-07-01 16:28:33 <phungus> typing on it now
 605 2012-07-01 16:28:45 <phungus> it's just a simple vps though
 606 2012-07-01 16:29:16 <brwyatt> My server didn't have issues either.
 607 2012-07-01 16:30:34 freewil has joined
 608 2012-07-01 16:35:21 Tykling has quit (Ping timeout: 248 seconds)
 609 2012-07-01 16:36:25 darksk1ez has quit (Ping timeout: 248 seconds)
 610 2012-07-01 16:39:08 darksk1ez has joined
 611 2012-07-01 16:39:46 <luke-jr> so what was the verdict on 0000000000000600498a139cfff8940a434e2351c418f8fd566236535b487ddd ?
 612 2012-07-01 16:39:55 <luke-jr> is it really destroying fees? :|
 613 2012-07-01 16:41:29 <brwyatt> ?
 614 2012-07-01 16:42:40 <luke-jr> EricLombrozo was in here saying the fees added up to more than was claimed
 615 2012-07-01 16:43:28 <gmaxwell> appears to.
 616 2012-07-01 16:43:41 <gmaxwell> My figuring has 0.1346 in fees.
 617 2012-07-01 16:43:47 <luke-jr> hrm
 618 2012-07-01 16:43:52 <gmaxwell> it pays out 50.1341.
 619 2012-07-01 16:43:53 Tykling has joined
 620 2012-07-01 16:44:03 <luke-jr> is there a way to ask bitcoind? :p
 621 2012-07-01 16:44:15 <luke-jr> blockchain.info seems to think 1341 is right :/
 622 2012-07-01 16:44:50 <sipa> i guess you can change the fee check from >= to ==, and emit a warning for the > case
 623 2012-07-01 16:46:47 <gmaxwell> gettransaction sadly doesn't easily let me see the fees anymore.
 624 2012-07-01 16:47:00 <gmaxwell> Otherwise I'd give you the shell oneliner to figure it out.
 625 2012-07-01 16:47:20 <luke-jr> Transaction Fees	-0.00000001 BTC
 626 2012-07-01 16:47:22 * luke-jr stabs bc.info
 627 2012-07-01 16:47:35 <luke-jr> http://blockchain.info/block-index/133504/0000000000004c78956f8643262f3622acf22486b120421f893c0553702ba7b5
 628 2012-07-01 16:48:01 <gmaxwell> Yea, I try not to use those sites they're all crap.
 629 2012-07-01 16:48:11 <luke-jr> sigh
 630 2012-07-01 16:48:35 eoss has joined
 631 2012-07-01 16:48:36 eoss has quit (Changing host)
 632 2012-07-01 16:48:36 eoss has joined
 633 2012-07-01 16:48:36 <gmaxwell> "Oh hai, I used "float" for all these bitcoin values"
 634 2012-07-01 16:49:09 <luke-jr> I wonder if Inaba has that Eloipool instance running still
 635 2012-07-01 16:49:13 <luke-jr> it should have the data to debug
 636 2012-07-01 16:49:33 <gmaxwell> in any case, it sounds like it was exactly one fee short.
 637 2012-07-01 16:50:08 sirk390 has joined
 638 2012-07-01 16:51:42 Apexseals has joined
 639 2012-07-01 16:55:05 darksk1ez has quit (Ping timeout: 248 seconds)
 640 2012-07-01 16:55:34 darksk1ez has joined
 641 2012-07-01 17:01:11 <sipa> why oh why are the hashNext values written to disk in CDiskBlockIndex entries?
 642 2012-07-01 17:01:58 <sipa> they can be derived from hashBestBlock, and removing it simplifies reorg code
 643 2012-07-01 17:02:48 * TD grimaces at the description from the coinbase CEO of "needing a PhD to transfer bitcoins"
 644 2012-07-01 17:02:56 * TD mentally marks them down a notch
 645 2012-07-01 17:03:49 <sipa> marketing talk :)
 646 2012-07-01 17:04:20 <ersi> s/marketing/idiotic bullshit/
 647 2012-07-01 17:04:27 <ersi> It's not the same, dear sir.
 648 2012-07-01 17:05:14 <luke-jr> ersi: what's the difference?
 649 2012-07-01 17:05:53 <ersi> It's a fine line in between, but it's mostly that it's exceptionally inane
 650 2012-07-01 17:06:17 darksk1ez has quit (Ping timeout: 248 seconds)
 651 2012-07-01 17:06:18 D34TH has joined
 652 2012-07-01 17:07:44 <lianj> TD: well its not that easy to securely use bitcoin for non geeks
 653 2012-07-01 17:08:40 <TD> securely, not yet, but they didn't say that :)
 654 2012-07-01 17:09:03 <TD> i dislike iPhone users saying bitcoin is hard to use because their chosen device maker forces them to screw about with a web browser
 655 2012-07-01 17:09:54 <TD> i did a trade btc for cash in person the other day
 656 2012-07-01 17:10:06 <TD> (with the android bitcoin wallet). it was trivial and there were no problems.
 657 2012-07-01 17:17:22 <galambo>  i think phd is a metaphor for "a lot of patience"
 658 2012-07-01 17:17:47 <galambo> pretending that theres no learning curve isnt helpful
 659 2012-07-01 17:17:59 meLon has joined
 660 2012-07-01 17:18:00 meLon has quit (Changing host)
 661 2012-07-01 17:18:00 meLon has joined
 662 2012-07-01 17:18:00 <ersi> PhD, also known as 'Pile it Higher and Deeper'
 663 2012-07-01 17:19:09 <galambo> ahhaa yeah education is totally shit mate
 664 2012-07-01 17:20:42 <TD> there is a learning curve indeed
 665 2012-07-01 17:22:35 <copumpkin> I quit mine!
 666 2012-07-01 17:29:09 TD has quit (Quit: TD)
 667 2012-07-01 17:29:46 Prattler has joined
 668 2012-07-01 17:29:59 <jgarzik> pooh, TD disappears two seconds before I try to write
 669 2012-07-01 17:32:33 sirk390 has quit (Quit: Leaving.)
 670 2012-07-01 17:42:21 [Tycho] has joined
 671 2012-07-01 17:46:16 coblee has quit (Ping timeout: 244 seconds)
 672 2012-07-01 17:52:43 <gmaxwell> 09:58  * TD grimaces at the description from the coinbase CEO of "needing a PhD to transfer bitcoins"
 673 2012-07-01 17:53:08 <gmaxwell> ^ this kind of half-truth is appealing to investors, so that may be a contributing factor.
 674 2012-07-01 17:53:56 <copumpkin> lol
 675 2012-07-01 17:54:01 <copumpkin> gmaxwell: do you have one?
 676 2012-07-01 17:54:06 <copumpkin> a Ph.D., that is
 677 2012-07-01 17:54:11 <sipa> i believe the fact that they got funding for a startup is more beneficial to the bitcoin economy than the fact that they are spreading a half-truth to get it
 678 2012-07-01 17:54:45 <gmaxwell> copumpkin: Nope. Though I often feel left out, most of my friends do! :)
 679 2012-07-01 17:54:56 <copumpkin> poor thing!
 680 2012-07-01 17:55:02 <sipa> gmaxwell: you did publish papers, didn't you?
 681 2012-07-01 17:55:19 <gmaxwell> Yes, I've published.
 682 2012-07-01 17:55:19 <copumpkin> sipa: do you?
 683 2012-07-01 17:55:35 <sipa> copumpkin: yes
 684 2012-07-01 17:55:39 <copumpkin> Dr. Sipa
 685 2012-07-01 17:57:09 <sipa> :)
 686 2012-07-01 17:57:48 <sipa> not sure it was the best thing i did with my time, though
 687 2012-07-01 17:58:02 <copumpkin> takes some patience! I wasn't able to :)
 688 2012-07-01 17:58:42 <gmaxwell> cool. getpeerinfo works as expected for outbound onions.
 689 2012-07-01 17:59:12 <sipa> ah, good
 690 2012-07-01 17:59:45 <gmaxwell> sipa: are you running new enough code to have getpeerinfo?  I'm connected to you currently via onion, so you should have an inbound one.
 691 2012-07-01 18:00:08 <gmaxwell> I expect it'll show 127.0.0.1 ... not really all that informative. :( perhaps it should also display the us/them fields from the initial handshake.
 692 2012-07-01 18:00:36 <sipa> gmaxwell: my vps is running older code now, but i'm quite sure it will show 127.0.0.1
 693 2012-07-01 18:01:13 <sipa> indeed stroing and showing the incoming them would be useful
 694 2012-07-01 18:02:43 <gmaxwell> I should probably nag tor for mutually authenticated onion<->onion connections... thats what torchat wants, and in order to get it torchat must have access to your onion private keys.
 695 2012-07-01 18:02:58 <gmaxwell> Which is kinda lame.
 696 2012-07-01 18:05:45 JStoker has quit (Excess Flood)
 697 2012-07-01 18:06:33 coblee has joined
 698 2012-07-01 18:07:59 minimoose has joined
 699 2012-07-01 18:08:34 Motest003 has quit (Ping timeout: 252 seconds)
 700 2012-07-01 18:08:43 Motest003 has joined
 701 2012-07-01 18:13:17 nanook7 has joined
 702 2012-07-01 18:16:46 nanook7 has quit (Client Quit)
 703 2012-07-01 18:21:04 bonks has joined
 704 2012-07-01 18:21:04 meLon has quit (Remote host closed the connection)
 705 2012-07-01 18:24:26 meLon has joined
 706 2012-07-01 18:24:26 meLon has quit (Changing host)
 707 2012-07-01 18:24:26 meLon has joined
 708 2012-07-01 18:26:14 JStoker has joined
 709 2012-07-01 18:31:12 Tykling has quit (Quit: quit)
 710 2012-07-01 18:34:05 bitcoinbulletin has quit (Remote host closed the connection)
 711 2012-07-01 18:35:09 meLon has quit (Remote host closed the connection)
 712 2012-07-01 18:39:55 theodore has joined
 713 2012-07-01 18:43:31 <jgarzik> does testnet3 mint blocks on a regular basis?
 714 2012-07-01 18:43:47 <luke-jr> jgarzik: I think only when someone is testing
 715 2012-07-01 18:43:48 <jgarzik> testnet2 was running around ~1-2 hours per confirmation
 716 2012-07-01 18:43:53 <luke-jr> I did a bunch recently to test refactor_times
 717 2012-07-01 18:44:17 bitcoinbulletin has joined
 718 2012-07-01 18:46:24 Tykling has joined
 719 2012-07-01 18:49:43 <gmaxwell> We should probably merge some only mine at 20minutes code.
 720 2012-07-01 18:50:15 <gmaxwell> I'm not currently mining on testnet3. If anyone needs blocks on it, lemme know...
 721 2012-07-01 18:51:05 <gmaxwell> (or if you just need coins lemme know)
 722 2012-07-01 18:52:08 <jgarzik> writing an app (pastecoin v2) which will need testing eventually.  trying to plan.
 723 2012-07-01 18:57:02 <gmaxwell> I've been trying to encourage people running bitcoin services to run testnet versions. I don't know if that will be successful.
 724 2012-07-01 19:04:13 <luke-jr> gmaxwell: I was thinking something along the lines of "consider the pool dead if target isn't diff 1" for BFGMiner
 725 2012-07-01 19:05:44 <ersi> jgarzik: pastecoin? that'd be a pastebin? or file host? ;o
 726 2012-07-01 19:05:50 <gmaxwell> luke-jr: hm. thats kinda hacky. Also, even if the diff gets down to 1 I'd prefer the background mining only run at the 20 minute mark.
 727 2012-07-01 19:06:04 <ersi> I got a testnet instance up, not mining it though.
 728 2012-07-01 19:06:11 <luke-jr> gmaxwell: well, I wouldn't want to mine, if someone else is keeping it at higher diff
 729 2012-07-01 19:07:00 <gmaxwell> luke-jr: right but if it gets to 1 because no one mined last cycle, you don't want to constantly mine it until 2016 blocks pass.
 730 2012-07-01 19:07:25 <luke-jr> true
 731 2012-07-01 19:07:32 <gmaxwell> it would be bonkers.. e.g. if I did that on my own farm everytime testnet hit diff 1 I'd spend several hours cranking it directly to diff 4.
 732 2012-07-01 19:07:36 <luke-jr> "minimum time between shares"? :p
 733 2012-07-01 19:08:30 <gmaxwell> luke-jr: or how about bitcoin just returning the "I'm downloading" error... and make bfgminer back off when it gets that?
 734 2012-07-01 19:09:00 <luke-jr> gmaxwell: seems the wrong place to put the logic, but that would work right now
 735 2012-07-01 19:09:04 <gmaxwell> e.g. a switch to bitcoind -genslow=1 which makes it only return work when >20 minutes have passed.
 736 2012-07-01 19:09:16 <gmaxwell> then make the miner respond gracefully to that case.
 737 2012-07-01 19:09:24 <gmaxwell> e.g. don't try to constantly get work.
 738 2012-07-01 19:09:33 <jgarzik> ersi: yep, pastecoin is a file host.  you can post files, and pay yourself to host/store them... or the community can pay to a bitcoin address and "keep the file alive"
 739 2012-07-01 19:10:26 khalahan has quit (Ping timeout: 248 seconds)
 740 2012-07-01 19:10:37 <luke-jr> jgarzik: what will you do about illegal files?
 741 2012-07-01 19:10:41 <gmaxwell> luke-jr: maybe even a parameter like -genslow=40 to make it wait 40 minutes before it tries mining.
 742 2012-07-01 19:11:17 <jgarzik> luke-jr: if legal takedown notices are sent, we will comply
 743 2012-07-01 19:11:48 <jgarzik> luke-jr: US company, complying with any applicable US laws
 744 2012-07-01 19:11:54 <gmaxwell> luke-jr: one thing such a site should do is encrypt everything with the key in the URL ... so you can at least make an argument that its the URL which is the thing which needs to be taken down. (though obviously any such site should comply with applicable laws)
 745 2012-07-01 19:12:18 <luke-jr> jgarzik: and hosting payments forfeit? how can someone respond to the takedown to get it back up?
 746 2012-07-01 19:12:29 Turing_i has joined
 747 2012-07-01 19:12:29 Turing_i has quit (Changing host)
 748 2012-07-01 19:12:29 Turing_i has joined
 749 2012-07-01 19:12:42 <jgarzik> luke-jr: there would need to be an appeals process, like a youtube etc.
 750 2012-07-01 19:13:03 <luke-jr> youtube doesn't take down illegal content :/
 751 2012-07-01 19:13:12 <gmaxwell> Should be pretty easy to pretty much automate notice and takedown.
 752 2012-07-01 19:13:37 Turingi has quit (Ping timeout: 244 seconds)
 753 2012-07-01 19:13:44 <jgarzik> Youtube takes down content, though they prefer to strike deals where they direct all ad revenue to the music publisher, regardless of uploading party.
 754 2012-07-01 19:13:59 Turing_i is now known as Turingi
 755 2012-07-01 19:14:07 <gmaxwell> Youtube's takedown process is bizarre, they are hyperoverreactive to big media companies— even when they make bogus claims— and pretty unresponsive to claims from small entities. (e.g. they'll only comply strictly with the law, requring them to mail their designated agent)
 756 2012-07-01 19:14:09 <jgarzik> That's the norm for "illegal" content on youtube: permit the upload, send money to content owner.
 757 2012-07-01 19:14:31 <luke-jr> jgarzik: I didn't mean copyright infringment, I meant criminal
 758 2012-07-01 19:14:35 <gmaxwell> Ah. The revenue stream. Interesting.
 759 2012-07-01 19:14:55 <luke-jr> some idiots posted instructions on setting up your own botnet using Eligius
 760 2012-07-01 19:14:59 <luke-jr> YouTube refuses to take it off
 761 2012-07-01 19:15:02 <gmaxwell> luke-jr: like the wallet stealing malware people have promoted on youtube?
 762 2012-07-01 19:15:10 <luke-jr> gmaxwell: I imagine there's that too
 763 2012-07-01 19:15:11 <sipa> luke-jr: why would that be illegal?
 764 2012-07-01 19:15:20 <luke-jr> sipa: computer cracking is illegal most places
 765 2012-07-01 19:15:27 <sipa> right
 766 2012-07-01 19:15:27 <jgarzik> Youtube has an internal system where (a) algorithms detect that a new upload is content from artist ABC, or (b) music publishers file claims, claiming that a new upload is their property
 767 2012-07-01 19:15:33 <luke-jr> sipa: it also violates Eligius's ToS
 768 2012-07-01 19:15:43 <jgarzik> they might have more, but that's what I knew from ~12 months ago
 769 2012-07-01 19:16:32 <jgarzik> luke-jr: if there is a criminal complaint (case number, whatever) from an authority, pastecoin will pursue it
 770 2012-07-01 19:16:38 <jgarzik> though I really hope it doesn't come to that...
 771 2012-07-01 19:16:52 <gmaxwell> luke-jr: for example, https://www.youtube.com/watch?v=8Hws-OruuqE
 772 2012-07-01 19:17:06 <luke-jr> jgarzik: seems pretty inevitable for something like that IMO
 773 2012-07-01 19:17:36 <jgarzik> hopefully limited in number
 774 2012-07-01 19:17:48 <gmaxwell> (I'm not sure if thats the one I looked at before— the one I looked at before dorked with the wallet file to give you a fake balance while simultaniously sending your wallet to some compromised hosts. Youtube wouldn't remove it, but reporting the compromised hosts got them taken offline)
 775 2012-07-01 19:17:52 <jgarzik> if too much time is wasted dealing with law, there is no time left to code :)
 776 2012-07-01 19:18:43 <gmaxwell> jgarzik: did you see the "Anonymous Publishing Is Dead" post on tor-talk? http://cryptome.org/2012/06/anon-pub-dead.htm
 777 2012-07-01 19:19:05 <luke-jr> jgarzik: yeah, there's a bunch of vigilantes out there demanding I spend time hunting down botnets abusing Eligius
 778 2012-07-01 19:19:19 <jgarzik> gmaxwell: nope. reading...
 779 2012-07-01 19:19:53 <gmaxwell> luke-jr: More than just TuxBlackEdo?  (which I can't help but laugh at... considering the source)
 780 2012-07-01 19:20:36 <luke-jr> gmaxwell: hosts-file.net is accusing Eligius of malware distribution unless I hunt the botnets down and somehow stop them
 781 2012-07-01 19:21:16 <luke-jr> gmaxwell: MyWOT.com as well
 782 2012-07-01 19:21:59 <gmaxwell> luke-jr: ... the operator of hosts-file.net is in the UK? man, thats ill advised.
 783 2012-07-01 19:22:14 <luke-jr> gmaxwell: not sure where he is
 784 2012-07-01 19:22:17 <jgarzik> I thought about trying to find a business model that permit the customer to pay once, and host some content forever.
 785 2012-07-01 19:22:19 <gmaxwell> (In the UK you can get a successful defamation claim against someone because they looked at you funny)
 786 2012-07-01 19:22:33 <luke-jr> gmaxwell: without paying a bunch to lawyers up front?
 787 2012-07-01 19:22:54 <gmaxwell> luke-jr: well, I wasn't suggesting you do that... but it seems to me that someone will eventually.
 788 2012-07-01 19:23:14 <luke-jr> if it wasn't expensive, I'd consider it :p
 789 2012-07-01 19:23:30 <luke-jr> it's basically blackmail
 790 2012-07-01 19:23:38 <gmaxwell> (In particular in the UK the truth of a defamatory claim isn't an automatic defense)
 791 2012-07-01 19:23:55 <luke-jr> hah, so you don't need to even argue the facts?
 792 2012-07-01 19:24:04 <luke-jr> wow
 793 2012-07-01 19:25:19 <jgarzik> also thought about doing a massively distributed storage service, which would use those el-cheap web site accounts as a final backing store.  Just need a cheap layer of caching and indexing servers on top.
 794 2012-07-01 19:25:41 <jgarzik> betcha could beat Amazon S3
 795 2012-07-01 19:25:52 <gmaxwell> jgarzik: You should also probably talk to zooko as he's spent a lot of time thinking about that.
 796 2012-07-01 19:26:09 <gmaxwell> S3 prices are pretty impressively high, but perhaps I don't have a real grasp of the true cost of that business.
 797 2012-07-01 19:26:40 RazielZ has quit (Quit: Leaving)
 798 2012-07-01 19:27:04 <jgarzik> Amazon does a lot of replication internally (so much so there is a cheaper pricing tier with lower reliability advertised), and they offer services such as "FedEx a hard drive, we load it into S3"
 799 2012-07-01 19:27:28 <jgarzik> even so, IMO they need competition :)
 800 2012-07-01 19:28:03 <freewil> google apparently just launched some cloud service that is supposed to be a competitor
 801 2012-07-01 19:28:04 <jgarzik> needs more APIs, too.  Why not a public service that offers iSCSI or NFS or, well, anything but HTTP REST?
 802 2012-07-01 19:28:06 <freewil> havnet looked at it yet
 803 2012-07-01 19:28:08 minimoose has quit (Ping timeout: 246 seconds)
 804 2012-07-01 19:28:31 <sipa> google drive?
 805 2012-07-01 19:28:56 <gmaxwell> iSCSI implementations have really awesome failure modes in the face of packet loss.. but perhaps those would get fixed if things like iSCSI-s3 existed.
 806 2012-07-01 19:28:57 minimoose has joined
 807 2012-07-01 19:29:17 <freewil> sipa, http://cloud.google.com/products/compute-engine.html
 808 2012-07-01 19:29:28 <jgarzik> Google's EC2 competitor
 809 2012-07-01 19:30:09 <gmaxwell> I filled out their request form to get info.. and not even a you're-not-good-enough-for-us autoresponder got back to me.
 810 2012-07-01 19:30:47 <gmaxwell> (Mostly I just wanted to know if they had subhour billing ... but I hear from someone they did accept that they don't, and their pricing is the same as te EC2 non-spot pricing. 'meh')
 811 2012-07-01 19:31:18 erska has quit (Ping timeout: 265 seconds)
 812 2012-07-01 19:31:35 hnz has quit ()
 813 2012-07-01 19:36:01 khalahan has joined
 814 2012-07-01 19:36:23 egecko has quit (Quit: ~ Trillian Astra - www.trillian.im ~)
 815 2012-07-01 19:43:56 membersonlyguy has joined
 816 2012-07-01 19:50:26 topi` has joined
 817 2012-07-01 19:51:07 meLon has joined
 818 2012-07-01 19:51:07 meLon has quit (Changing host)
 819 2012-07-01 19:51:07 meLon has joined
 820 2012-07-01 19:51:33 Zarutian has joined
 821 2012-07-01 20:08:11 t7 has joined
 822 2012-07-01 20:08:17 erska has joined
 823 2012-07-01 20:08:31 phantomcircuit is now known as MANLYMANMANMAN
 824 2012-07-01 20:08:42 MANLYMANMANMAN is now known as phantomcircuit
 825 2012-07-01 20:08:56 Diapolo has joined
 826 2012-07-01 20:13:10 <xorgate> is there any plan to make bitcoin-qt support multiple wallets?
 827 2012-07-01 20:13:32 membersonlyguy has quit (Read error: Connection reset by peer)
 828 2012-07-01 20:13:34 <sipa> yes, but not immediately
 829 2012-07-01 20:13:56 <sipa> (internally a lot of work for that is already done)
 830 2012-07-01 20:17:47 trickybear has joined
 831 2012-07-01 20:19:14 khalahan has quit (Quit: Bye)
 832 2012-07-01 20:20:43 trickybear has quit (Client Quit)
 833 2012-07-01 20:30:29 Diapolo has left ()
 834 2012-07-01 20:34:16 <xorgate> any multibit devs around perchance?
 835 2012-07-01 20:39:23 t7 has quit (Read error: Connection reset by peer)
 836 2012-07-01 20:40:01 jurov is now known as away!aktooj@84.245.71.31|jurov
 837 2012-07-01 20:41:10 Matt_von_Mises has joined
 838 2012-07-01 20:44:33 hnz has joined
 839 2012-07-01 20:54:54 <Matt_von_Mises> Another question: Why is the protocol version, timestamps and block height in the version message signed? Also where is the version message in the source code?
 840 2012-07-01 20:55:48 ThomasV has joined
 841 2012-07-01 20:56:28 <sipa> signed?
 842 2012-07-01 20:56:30 <gmaxwell> Matt_von_Mises: huh? nothing is _signed_. All messages have checksums, is that what you're refering to?
 843 2012-07-01 20:56:56 <Matt_von_Mises> They are signed integers
 844 2012-07-01 20:57:01 <sipa> ah :D
 845 2012-07-01 20:57:03 <Matt_von_Mises> You know. + or -
 846 2012-07-01 20:57:11 <gmaxwell> oh!
 847 2012-07-01 20:57:18 <gmaxwell> hah
 848 2012-07-01 20:57:23 <sipa> certainly timestamps shouldn't be
 849 2012-07-01 20:57:32 <kinlo> should be fairly easy to change them to unsigned ones?
 850 2012-07-01 20:57:47 <sipa> for block height and version it shouldn't matter too much
 851 2012-07-01 20:57:53 <kinlo> as long as everybody upgrades before 2^31 blocks
 852 2012-07-01 20:57:53 <Matt_von_Mises> I see no reason whatsoever why any of them should be signed.
 853 2012-07-01 20:58:08 <sipa> Legacy.
 854 2012-07-01 20:58:44 <Matt_von_Mises> I'm mainly wondering why it was chosen to be that way. I assume it was a mistake then.
 855 2012-07-01 20:58:47 <sipa> satoshi didn't pay too much attention to these things, unfortunately
 856 2012-07-01 20:58:58 <sipa> "int" is shorter to type than "unsigned int"
 857 2012-07-01 20:59:26 <sipa> in many place where it matters the signedness has since been corrected, but certainly not everywhere
 858 2012-07-01 21:00:05 <galambo> its a massive system and he was just one guy give  him a break ;)
 859 2012-07-01 21:00:19 <Matt_von_Mises> The timestamps will eventually need to be 64 bits. Signed 32 bit integers help but not forever.
 860 2012-07-01 21:01:08 <Matt_von_Mises> It seems Satoshi just released bitcoin a little bit early. Oh well. It still works.
 861 2012-07-01 21:01:09 <galambo> based on lines of code this project would take one person 70 months to code
 862 2012-07-01 21:01:36 <sipa> not really a problem; the serialized versions can remain 32 bits; it's easy enough to guess the upper 32 bits
 863 2012-07-01 21:02:03 <gmaxwell> fortunately the p2p protocol isn't that big a deal to change.
 864 2012-07-01 21:02:03 datagutt has quit (Quit: Computer has gone to sleep.)
 865 2012-07-01 21:02:23 <sipa> right, but timestamps are in block headers
 866 2012-07-01 21:02:45 <gmaxwell> sure but as you point out, it's easy to figure out the right upper bits.
 867 2012-07-01 21:03:13 <gmaxwell> we can worry about that in 2030 if bitcoin is still around. ;)
 868 2012-07-01 21:03:42 <Matt_von_Mises> When an overflow occurs the next block would have a lower time than the previous so when that occurs you could just add to the first 32 bits.
 869 2012-07-01 21:03:43 <sipa> and there hasn't been a hard fork in that time which Fixed Everything(R)
 870 2012-07-01 21:03:45 <galambo> you can change the block storage with etotheipi's proposal to map unspent CTxOut
 871 2012-07-01 21:04:24 <gmaxwell> galambo: changing the block header format is a harder change that most things.
 872 2012-07-01 21:04:25 <galambo> i think it provides a good method to say which storage you should check
 873 2012-07-01 21:04:41 <galambo> keep the old blk.dat and the new file
 874 2012-07-01 21:04:54 <sipa> galambo: we're working on changing block storage
 875 2012-07-01 21:05:06 hattorihanzo-rea has joined
 876 2012-07-01 21:05:42 hattorihanzo has quit (Ping timeout: 244 seconds)
 877 2012-07-01 21:05:52 <Matt_von_Mises> Do people agree that you could just detect overflows when going through the block chain?
 878 2012-07-01 21:06:21 <galambo> you can
 879 2012-07-01 21:06:36 <galambo> just like they do in regular accounting systems
 880 2012-07-01 21:06:45 <sipa> galambo: eto's proposal does several things: 1) changing storage 2) keeping an address-to-tx index 3) committing a hash of the open tx set tree to the block chain
 881 2012-07-01 21:07:01 <galambo> add everything up and check that the balance is zero
 882 2012-07-01 21:08:39 <sipa> galambo: i'm working on (1) in bitcoind; (3) will be useful one day for lighter nodes; (2) is useful for certain purposes but can nicely be done in an alt chain
 883 2012-07-01 21:10:34 Tykling has quit (Ping timeout: 248 seconds)
 884 2012-07-01 21:11:01 <galambo> well i dont think it changes storage but it would allow for changes in storage for people that don't want you refactoring their blockchain. you'd have a key-value pair for each unspent txn that identifies which store to pull the transaction from.
 885 2012-07-01 21:11:54 <sipa> what i'm doing right now is store the txout set in a very compact form that doesn't even keep the transactions themselves around
 886 2012-07-01 21:12:07 <sipa> which is sufficient for validating the blockchain
 887 2012-07-01 21:12:11 <jrmithdobbs> sipa: i'm beating up on ipv6 and here in a few on onion stuff on my weird network setup, not found any issues yet other than ipv6 being slower to sync when encap'ed as many times as that vpn is ;p
 888 2012-07-01 21:12:50 <sipa> galambo: but is not enough for rescanning, reorganising or serving blocks
 889 2012-07-01 21:13:22 <sipa> galambo: but those can be limited to only recent blocks, so you keep the full blocks around as well, but only recent ones
 890 2012-07-01 21:13:27 <sipa> jrmithdobbs: :)
 891 2012-07-01 21:13:30 <jrmithdobbs> sipa: completed a complete block sync via v6 on a node with no ip4
 892 2012-07-01 21:13:36 <jrmithdobbs> err blockchain
 893 2012-07-01 21:14:48 <sipa> nice!
 894 2012-07-01 21:15:08 <jrmithdobbs> sipa: btw, re: new wallet format you're working on, it's not going to have any extra files/env like bdb right so we'll be able to have it stored completely separate now?
 895 2012-07-01 21:15:54 Tykling has joined
 896 2012-07-01 21:16:10 <galambo> sipa its nice to see that someones actually working on it
 897 2012-07-01 21:16:29 <sipa> jrmithdobbs: that's the plan, but i'm working on other things first now
 898 2012-07-01 21:16:51 Tykling has quit (Excess Flood)
 899 2012-07-01 21:16:58 darkee has quit (Ping timeout: 276 seconds)
 900 2012-07-01 21:17:16 darkee has joined
 901 2012-07-01 21:17:19 <sipa> galambo: to get an idea: the full txout set serialized takes around 65 MiB in storage
 902 2012-07-01 21:17:38 <jrmithdobbs> sipa: damn, would fix how slow this box syncs :( (aes128xts fde -> blockchain+wallet in aes256ctr disk image)
 903 2012-07-01 21:17:41 <jrmithdobbs> ;p
 904 2012-07-01 21:17:54 <galambo> yeah i remember you talking about having it to that level before etotheipi even posted his thread
 905 2012-07-01 21:17:55 <sipa> in a database it will likely be close to twice as much
 906 2012-07-01 21:18:15 <galambo> so i shouldnt call it "his" proposal
 907 2012-07-01 21:18:16 MobiusL has quit (Ping timeout: 276 seconds)
 908 2012-07-01 21:18:36 <sipa> he wasn't aware of what i was working on at the time
 909 2012-07-01 21:18:48 <sipa> but the idea of keeping an open txout set is much much older
 910 2012-07-01 21:18:55 <galambo> yeah
 911 2012-07-01 21:19:09 <gmaxwell> galambo: all this stuff except for the the idea of just indexing by address (which was shown to be unworkable for validation) isn't original to him, it's been discussed in varrious forms for over a year. Evidence that it's a good idea: people keep independantly inventing it.
 912 2012-07-01 21:19:25 MobiusL has joined
 913 2012-07-01 21:19:38 <sipa> he doesn't claim to have invented it either :)
 914 2012-07-01 21:19:42 <jrmithdobbs> we talking pruning methods i'm assuming
 915 2012-07-01 21:19:50 <gmaxwell> jrmithdobbs: no, not .. quite.
 916 2012-07-01 21:19:53 <sipa> it's a combination of a few ideas
 917 2012-07-01 21:20:03 <gmaxwell> I mean, what sipa is doing is more like straight pruning.
 918 2012-07-01 21:20:14 <galambo> someone needs to figure out a safe way to update it over the network
 919 2012-07-01 21:20:21 <sipa> upgrade?
 920 2012-07-01 21:20:22 <gmaxwell> jrmithdobbs: there is now this elegant idea that it's possible to have ~all nodes be fully pruned without breaking the security model.
 921 2012-07-01 21:20:57 <galambo> update the txout file
 922 2012-07-01 21:21:41 <sipa> serialized blocks can be considered cryptograhpically secured patch files for the txout set
 923 2012-07-01 21:21:58 <gmaxwell> jrmithdobbs: the notion is that you have a tree of txouts which nodes update as they go along and process blocks... and miners commit to the identity of this tree.  So you can securely obtain the tree from other parties and know its the real tree. (or alternatively, not obtain it and just have remote nodes query it for you and provide tree fragements that prove their queries)
 924 2012-07-01 21:22:23 <gmaxwell> jrmithdobbs: and you can even bootstrap in this state without seeing the full history and be reasonably confident no cheating happened.
 925 2012-07-01 21:22:42 <jrmithdobbs> how reasonably?
 926 2012-07-01 21:22:52 <galambo> is somebody saying that all nodes can be fully pruned??? i havent heard anyone suggest that
 927 2012-07-01 21:22:59 <gmaxwell> jrmithdobbs: because anyone who has observed cheating  (e.g. spending against the rules) could prepare a message which _proves_ the cheating.. and broadcast it.
 928 2012-07-01 21:23:34 Prattler has quit (Quit: Ex-Chat)
 929 2012-07-01 21:23:40 <gmaxwell> (and because information is hard to stifle a farily low density of honest nodes is needed to preserve security in that model)
 930 2012-07-01 21:23:42 <jrmithdobbs> but you'd still want full nodes somewhere wouldn't you?
 931 2012-07-01 21:24:03 <sipa> i think we need to define full node first; it's used with different meanings
 932 2012-07-01 21:24:26 <jrmithdobbs> you you'd still want nodes with fully unpruned copies of the blockchain sowhere wouldn't you?
 933 2012-07-01 21:24:30 <gmaxwell> jrmithdobbs: These nodes would be 'full' in that they enforce all the rules.
 934 2012-07-01 21:24:58 <jrmithdobbs> well, maybe corrected copies with forks removed and such
 935 2012-07-01 21:25:03 <gmaxwell> jrmithdobbs: you might, but it's not strictly required for the security model to be largely upheld.
 936 2012-07-01 21:25:05 <jrmithdobbs> but you know what i mean
 937 2012-07-01 21:25:55 <jrmithdobbs> i remember seeing a thread on the mailling list but hadn't read through it
 938 2012-07-01 21:26:25 <gmaxwell> galambo: they could be all fully pruned (well, there are some risks around reorgs where you want to keep an undo log at least as long as the longest reorg you'll reasonable tolerate)
 939 2012-07-01 21:27:07 <jrmithdobbs> gmaxwell: so the idea is noone has to store the full thing
 940 2012-07-01 21:27:36 <jrmithdobbs> gmaxwell: there any way to leverage that to *prevent* anyone from storing the whole thing?
 941 2012-07-01 21:28:06 <gmaxwell> jrmithdobbs: just the current set of unspent txouts, and recent txn (either in the form of blocks or an undo log). Alas, I don't see any way to prevent that. :)
 942 2012-07-01 21:28:07 <jrmithdobbs> or discourage, at least
 943 2012-07-01 21:28:15 <galambo> the point is that its an accounting ledger, there are two sides to every transaction. you can do what you are saying in accounting too and only keep track of balances but its not as safe.
 944 2012-07-01 21:28:26 <jrmithdobbs> gmaxwell: ya, prevent was bad choice of words
 945 2012-07-01 21:29:27 <gmaxwell> galambo: Explain the threat model you're thinking of, and I'll tell you how its addressed.
 946 2012-07-01 21:29:41 <luke-jr> gmaxwell: any background with JTAG?
 947 2012-07-01 21:29:44 <sipa> jgarzik: it seems it is perfectly possible to remove the hashNext field from CDiskBlockIndex (and derive the best chain from hashBestBlock), which means that the reorg/connect logic doesn't need to touch any block index data at all (the block index entries are only written when saving the block, and are immutable after that), so storing the block index data in a separate file or database becomes trivial, as it doesn't need atomic updates together...
 948 2012-07-01 21:29:50 <sipa> with transaction index updates (except for hashBestBlock)
 949 2012-07-01 21:30:04 <galambo> you can't replay how you achieved the current state
 950 2012-07-01 21:30:11 <galambo> if you only store the current state
 951 2012-07-01 21:30:13 <jrmithdobbs> if it can be strongly discouraged it greatly strengethens the pseudo/anonymity claimns in several cases
 952 2012-07-01 21:30:18 MobiusL has quit (Remote host closed the connection)
 953 2012-07-01 21:30:22 <galambo> the replay is what makes the system of accounts valuable
 954 2012-07-01 21:30:43 <jrmithdobbs> galambo: maybe that's not so bad?
 955 2012-07-01 21:30:56 Mobius_ has joined
 956 2012-07-01 21:30:58 <sipa> bitcoin basically has two levels of security: 1) validate everything 2) trust what the majority of hashpower votes for
 957 2012-07-01 21:31:03 <galambo> what would you do if there was another integer overflow in the code
 958 2012-07-01 21:31:14 <jrmithdobbs> galambo: so long as you know you can have confidence in the current state at any given time (within params), it makes things much nicer really
 959 2012-07-01 21:31:15 <galambo> cant go back in time like you did before
 960 2012-07-01 21:31:19 <gmaxwell> galambo: because of reorgs you need to retain some undo data for a localized replay. But lets ignore that.  Can I instead change your criticism to "You can't independantly validate that the rules were followed by miners in the past, if no one gives you all of the past state"?
 961 2012-07-01 21:31:22 <sipa> once you prune blocks, you cannot serve data for others to validate everything
 962 2012-07-01 21:31:54 <jrmithdobbs> aye
 963 2012-07-01 21:32:20 <sipa> but you can still validate everything yourself (given that you trust your own old state)
 964 2012-07-01 21:32:33 <gmaxwell> galambo: I can weaken the I-personally-validated-the-whole-history security model very slightly and give you something almost as good which doesn't run into the problem of having to replay everything.
 965 2012-07-01 21:32:51 <jrmithdobbs> gmaxwell: please
 966 2012-07-01 21:32:55 <galambo> ok
 967 2012-07-01 21:33:13 khalahan has joined
 968 2012-07-01 21:33:31 <galambo> but thats not the same. thats why these two systems have to exist concurrently.
 969 2012-07-01 21:33:49 <gmaxwell> galambo: Instead I can give: nodes-commit-to-the-state-constantly-you-can-trust-the-commitment-because-if-there-is-only-_one_-honest-node-that-wittnessed-misbheavior-he-can-cheaply-prove-it-to-everyone.
 970 2012-07-01 21:34:22 <galambo> and how do you correct misbehavior if you cant play back
 971 2012-07-01 21:34:24 <jrmithdobbs> that that 'can' be strengthened any?
 972 2012-07-01 21:34:30 <jrmithdobbs> s/that that/can that/
 973 2012-07-01 21:34:34 <sipa> galambo: you always keep some undo information
 974 2012-07-01 21:35:00 <gmaxwell> galambo: and misbehavior doesn't happen in the first place because of POW burrying, and because it has no chance of success.
 975 2012-07-01 21:35:03 <jgarzik> sipa: makes sense
 976 2012-07-01 21:35:18 <jgarzik> sipa: hashNext is just a convenience, after all
 977 2012-07-01 21:35:19 <gmaxwell> (no chance, if you assume there exists at least one honest node which is not totally censored)
 978 2012-07-01 21:35:35 <gmaxwell> jrmithdobbs: Well, can would be 'will', really.
 979 2012-07-01 21:35:54 <galambo> how would your system deal with this https://en.bitcoin.it/wiki/Incidents#CVE-2010-5139
 980 2012-07-01 21:36:05 <jrmithdobbs> gmaxwell: so it's 'will' for anyone running uncompromised official satoshi clients
 981 2012-07-01 21:36:11 <galambo> how would you know what state to return to
 982 2012-07-01 21:36:12 <jrmithdobbs> at least initially
 983 2012-07-01 21:36:21 <gmaxwell> jrmithdobbs: exactly, at least if this is implemented.
 984 2012-07-01 21:36:32 <sipa> galambo: hoping that it gets fixed before everyone's undo information until that point is lost
 985 2012-07-01 21:36:35 <jrmithdobbs> in that case, that's a pretty damned strong assurance, ya
 986 2012-07-01 21:36:44 egecko has joined
 987 2012-07-01 21:36:51 <galambo> yeah and id say thats a pretty big problem
 988 2012-07-01 21:36:58 <gmaxwell> galambo: you couldn't roll back futher than your replay data goes. Realistically nodes should be keeping a fairly large amount of replay data. Sipa and I were talking about on the order of a few thousand blocks.
 989 2012-07-01 21:37:02 <jrmithdobbs> galambo: why?
 990 2012-07-01 21:37:22 <gmaxwell> Also, a reorg that big today would quite possibly be fatal for bitcoin in any case.
 991 2012-07-01 21:37:35 <galambo> but its at least possible
 992 2012-07-01 21:37:45 <sipa> indeed, and an incident like that CVE would be devastating right now, imho
 993 2012-07-01 21:37:45 <gmaxwell> also, if you _did_ want to do a big replay you'd simply need to provide the required data.
 994 2012-07-01 21:37:46 <jrmithdobbs> galambo: and there will always be nodes that keep the whole thing somewhere
 995 2012-07-01 21:38:01 <jrmithdobbs> even if they archive it off and don't keep it live
 996 2012-07-01 21:38:07 <sipa> galambo: just realised that one should commit the replay data to the blockchain as well
 997 2012-07-01 21:38:14 <galambo> jrmithdobbs: they're suggesting that there wont be any nodes that keep the whole thing
 998 2012-07-01 21:38:19 <sipa> s/reply/undo/
 999 2012-07-01 21:38:31 <sipa> gmaxwell: just realised that one should commit the undo data to the blockchain as well
1000 2012-07-01 21:38:33 <jrmithdobbs> galambo: they're suggesting that there wont *have* to be any nodes that keep the whole thing
1001 2012-07-01 21:38:33 <gmaxwell> galambo: I'm saying no one would have to for the system to work. People will of course.
1002 2012-07-01 21:38:41 <jrmithdobbs> you're naive if you think no nodes will
1003 2012-07-01 21:38:49 <galambo> oh ok sorry then were wasting our time
1004 2012-07-01 21:38:51 <galambo> sorry
1005 2012-07-01 21:39:23 <galambo> but you should be careful to think that the information your are pruning has zero value
1006 2012-07-01 21:39:26 <gmaxwell> Basically my intent with this thinking is making it so that it's very cheap to run nodes that fully enforce the rules, so we can avoid losing decenteralizaton as bitcoin grows.
1007 2012-07-01 21:39:43 t7 has joined
1008 2012-07-01 21:40:07 <jrmithdobbs> gmaxwell: i still wish p2pool was more widely adopted to address the actual validation portion
1009 2012-07-01 21:40:24 <gmaxwell> Certantly I'd prefer to keep things simple and store everything, but increasingly users (even technically sophicated ones) are moving to online wallets and SPV nodes.  And I'm concerned that if this increases you'll have an effectively centeralized system at the end.
1010 2012-07-01 21:41:33 <jeremias> what's the problem, people enjoy some centralized entity abusing them, you want to deny their enjoyment?
1011 2012-07-01 21:41:44 <jeremias> (just joking, sorry)
1012 2012-07-01 21:41:54 Tykling has joined
1013 2012-07-01 21:42:03 <gmaxwell> jrmithdobbs: becaues of the cost of the mining pow I don't think we can really depend on mining alone as the sole source of decenteralized validation. Miners have correlated interests which would help them achieve agreements to tell common lies.
1014 2012-07-01 21:42:26 <gmaxwell> E.g. to keep the miners honest there needs to be a large number of non-miners validating too.
1015 2012-07-01 21:43:18 <freewil> what is the incentive for an individual non-miner to validate?
1016 2012-07-01 21:43:42 <jrmithdobbs> there isn't any at this point
1017 2012-07-01 21:43:44 <jrmithdobbs> good will
1018 2012-07-01 21:43:45 Motest003 has quit (Ping timeout: 265 seconds)
1019 2012-07-01 21:43:46 <gmaxwell> freewil: to keep miners from inflating the currency, and to make themselves more secure against short fork attacks.
1020 2012-07-01 21:43:56 <freewil> hmm that seems to be a flaw
1021 2012-07-01 21:44:08 <gmaxwell> But those incentives are fairly weak, which is why its important to make it possible to validate very inexpensively.
1022 2012-07-01 21:44:41 <freewil> seems to be a form of the "tragedy of the commons" - everyone lets everyone else take the responsibility
1023 2012-07-01 21:44:52 <jrmithdobbs> gmaxwell: another approach is to add benefit to the validation
1024 2012-07-01 21:44:57 <gmaxwell> I called it that specifically in one of my bitcoin-dev posts.
1025 2012-07-01 21:45:14 <jrmithdobbs> gmaxwell: eg, a bitcoin killer app that requires a bitcoind (or similar) running/accessible at your house/etc
1026 2012-07-01 21:45:43 <jrmithdobbs> where having fresh validation info is actually useful to the end user
1027 2012-07-01 21:45:43 <gmaxwell> jrmithdobbs: we'll need everything of course. :) But lowering the validation costs is a strict win.
1028 2012-07-01 21:45:52 <jrmithdobbs> yes definitely
1029 2012-07-01 21:46:28 <gmaxwell> jrmithdobbs: well, it already is some.. for example, right now I could use gpumax to mine a six block fork where 1m bitcoins is given to me from nowhere and paid to you, and your SPV node would gladly accept it.
1030 2012-07-01 21:47:01 <jrmithdobbs> gmaxwell: ?
1031 2012-07-01 21:47:04 <gmaxwell> That attack doesn't work against full nodes. — I'd have to at least have the million bitcoins to begin with, and be doing so leave more of an evidence trail.
1032 2012-07-01 21:47:16 <jrmithdobbs> ah
1033 2012-07-01 21:47:26 <gmaxwell> jrmithdobbs: if you don't validate your only evidence that I'm not making up money is the number of blocks behind it.
1034 2012-07-01 21:47:56 <gmaxwell> Validating nodes are simply more secure and always will be.
1035 2012-07-01 21:48:18 <gmaxwell> But ... thats probably not enough incentive if it's a big burden to run one.
1036 2012-07-01 21:48:29 <jrmithdobbs> yes but there's no benefit to be validating current data unless you're about to use it
1037 2012-07-01 21:48:41 <gmaxwell> 'cause everyone picks convenience over security.
1038 2012-07-01 21:48:57 <jrmithdobbs> so there's large numbers of these types of nodes that get shut off frequently for very long periods and aren't actually helpful to the network
1039 2012-07-01 21:49:07 <TuxBlackEdo> wtf I got highlighted here?
1040 2012-07-01 21:49:23 <jrmithdobbs> good for you?
1041 2012-07-01 21:49:56 erska has quit (Ping timeout: 245 seconds)
1042 2012-07-01 21:50:01 <TuxBlackEdo> 19:15 	luke-jr 	jgarzik: yeah, there's a bunch of vigilantes out there demanding I spend time hunting down botnets abusing Eligius
1043 2012-07-01 21:50:01 <TuxBlackEdo> 19:15 	gmaxwell 	luke-jr: More than just TuxBlackEdo? (which I can't help but laugh at... considering the source)
1044 2012-07-01 21:50:15 <TuxBlackEdo> wtf is that supposed to mean?
1045 2012-07-01 21:51:02 <gmaxwell> TuxBlackEdo: AFAIK your vocation is search engine spamming and you're worried about mild mannered botnet miners. I find this amusing. Perhaps I'm weird.
1046 2012-07-01 21:51:06 <jrmithdobbs> TuxBlackEdo: let me translate to internet 12yo for you: "19:15 <gmaxwell> luke-jr: More than just that dum faggot TuxBlackEdo?"
1047 2012-07-01 21:51:54 <TuxBlackEdo> gmaxwell, there are people that go into #eligius (on two different occasions) that have complained a mining trojan has surfaced on their computer mining on eligius
1048 2012-07-01 21:52:12 <TuxBlackEdo> and search engine spamming and SEO are two different things
1049 2012-07-01 21:52:30 <gmaxwell> Not that I approve of botnet mining in the slighest, ... people people manipulating search results for commercial gains has a much more negative impact on my life, than someone running some miner on my systems without my permission would.
1050 2012-07-01 21:52:35 <TuxBlackEdo> Ad hominem.
1051 2012-07-01 21:52:43 <TuxBlackEdo> Ad hominem!!
1052 2012-07-01 21:52:52 <luke-jr> TuxBlackEdo: and I help them file a police report
1053 2012-07-01 21:52:59 <TuxBlackEdo> gmaxwell argues in logical fallacies
1054 2012-07-01 21:53:01 <jrmithdobbs> please save us the self-serving defense of your profession and be a grown up, just drop it.
1055 2012-07-01 21:53:21 <gmaxwell> TuxBlackEdo: Hey, I was clear enough that I was remarking about the source there. ::shrugs::
1056 2012-07-01 21:53:25 <TuxBlackEdo> plus I don't do just SEO, i write iphone apps as well
1057 2012-07-01 21:53:38 <jrmithdobbs> you know what 99.9999% of the world thinks of your profession, you know that the situation he pointed out is amusing to most of the world, get over it or find another job that doesn't make you hate yourself?
1058 2012-07-01 21:53:53 <TuxBlackEdo> i don't hate myself...?
1059 2012-07-01 21:53:58 makomk has joined
1060 2012-07-01 21:53:59 <luke-jr> TuxBlackEdo: so spamming is OK as long as you bring crappy food to the people in jail?
1061 2012-07-01 21:54:03 <gmaxwell> then fine, no need to argue.
1062 2012-07-01 21:54:09 <TuxBlackEdo> I actually like what I do
1063 2012-07-01 21:54:18 <jrmithdobbs> really? you just got all pissed off and defensive about some random comment that was harmless and kind of amusing several hours ago
1064 2012-07-01 21:54:26 <TuxBlackEdo> I help businesses that deserve to be #1 in google results become #1
1065 2012-07-01 21:54:27 <jrmithdobbs> sure seems like you're secure in your profession to me
1066 2012-07-01 21:54:38 <gmaxwell> Ahem. BITCOIN.
1067 2012-07-01 21:54:48 <TuxBlackEdo> jrmithdobbs stop trolling
1068 2012-07-01 21:55:02 <jrmithdobbs> no
1069 2012-07-01 21:55:20 <sipa> take this elsewhere, please
1070 2012-07-01 21:55:25 <TuxBlackEdo> me?
1071 2012-07-01 21:55:28 <sipa> everyone
1072 2012-07-01 21:55:29 <gmaxwell> Sorry.
1073 2012-07-01 21:56:03 <gmaxwell> jrmithdobbs: in any case, under the model where a single honest validating node can produce an alert/proof of misbehavior, even if many of the validators are offline at any given moment means that the system should still be secure.
1074 2012-07-01 21:57:48 <gmaxwell> and if the cost is low enough you can count on 'altruists' making sure that there are always some honest nodes online. Also, because the proofs can be relatively small, and the honest nodes are passive unless they have something to report, it would be very hard to find and shut them all down.
1075 2012-07-01 21:59:13 <jrmithdobbs> gmaxwell: i don't know if the altruist thing will hold, but it's a start, i mean, we're not talking a project like tor that's saving people's lives in some cases
1076 2012-07-01 21:59:18 <jrmithdobbs> money brings out the worst in everyone
1077 2012-07-01 22:00:32 <jrmithdobbs> (eventually)
1078 2012-07-01 22:00:55 <Zarutian> well it is quantified power so to speak
1079 2012-07-01 22:01:14 <freewil> i think that is true somewhat, some are more easily separated from their morals
1080 2012-07-01 22:01:17 <gmaxwell> jrmithdobbs: well, again, thats the point of making it sufficiently cheap. If the cost of running a validating node is low enough to be hobby level for people then it's likely many people will run them.
1081 2012-07-01 22:04:50 TuxBlackEdo has quit (Quit: Leaving)
1082 2012-07-01 22:05:06 rdponticelli has joined
1083 2012-07-01 22:05:27 <jrmithdobbs> gmaxwell: yes, it's a start, i don't like solutions that rely on altruism, they don't sit right
1084 2012-07-01 22:05:35 <jrmithdobbs> maybe i really have gotten too cynical
1085 2012-07-01 22:05:38 <jrmithdobbs> ha
1086 2012-07-01 22:07:14 <gmaxwell> jrmithdobbs: well, I don't generally. But depending on there existing a minimum of one semi-altruist (after all, keeping the currency honest is a selfish end too) worldwide at all times doesn't bother me.
1087 2012-07-01 22:07:16 <Zarutian> I hope I understand this correctly, validating nodes are full nodes that valdiate transaction for other more lightweight clients? (Such as wallet apps on smartphones). So what is stopping people for packaging such as a subscribable service?
1088 2012-07-01 22:07:30 <Eliel_> jrmithdobbs: I don't think it really relies on altruism. I mean, there's no personal benefit to most users to withhold information about alerts.
1089 2012-07-01 22:07:59 <gmaxwell> Zarutian: lightweight clients only check the headers.. they use the blockchain as evidence of validation.
1090 2012-07-01 22:08:44 da2ce7 has joined
1091 2012-07-01 22:08:49 <gmaxwell> Zarutian: but the 'problem' there is that if validation itself is costly there may be relatively few parties performing it. (costly validation means that its less expensive in total to consolidate it)
1092 2012-07-01 22:08:54 erska has joined
1093 2012-07-01 22:09:35 <gmaxwell> And if it's consolidated its a point of potential control. ("Mine this bogus transaction that gives us a million bitcoin from thin air or we do_bad_things")
1094 2012-07-01 22:09:42 <Eliel_> jrmithdobbs: in fact, not sharing alerts is rather suspicius in itself, so there's an incentive to share them (most of them anyway)
1095 2012-07-01 22:10:25 <gmaxwell> Eliel_: if you're heeding an alert it is VERY much in your interest to forward them as far and fast as possible.
1096 2012-07-01 22:10:33 <gmaxwell> (because you want to make sure everyone else does too)
1097 2012-07-01 22:10:40 TD has joined
1098 2012-07-01 22:10:56 <Eliel_> ah, true
1099 2012-07-01 22:11:42 <gmaxwell> But creating one? well. thats less clear. Its in everyone's interest to prevent inflation, unless they're the ones getting the inflated funds. But there is a tragedy of the commons risk. "someone else will watch out for this, so I don't have to"
1100 2012-07-01 22:11:48 <Zarutian> as I see it validation is costly mainly in keeping the full block chain and random accessing it. (Disk seek times are looong)
1101 2012-07-01 22:12:04 <jrmithdobbs> gmaxwell: yes, that bit's a problem.
1102 2012-07-01 22:12:37 <gmaxwell> Zarutian: Correct. ECDSA is also computationally costly, though not really an issue with the current maximum blocksize.
1103 2012-07-01 22:12:50 <jrmithdobbs> gmaxwell: you know off top of your head if you ignore evenything >=median block size what the average is?
1104 2012-07-01 22:13:12 <jrmithdobbs> as in, what's the average small block size
1105 2012-07-01 22:13:25 <jrmithdobbs> (including the frontloaded no txn ones)
1106 2012-07-01 22:15:02 <gmaxwell> jrmithdobbs: for i in {1..100} ; do blk=`bd getblockhash $i` ; echo $i `bd getblock $blk | grep size` ; done
1107 2012-07-01 22:15:10 <gmaxwell> Don't ask me, this is now a shell oneliner question.
1108 2012-07-01 22:15:42 <Zarutian> gmaxwell: ECDSA could be speeded up by providing faster implementations of the primitives it is built out of or faster implementation of the algorythm itself (AFAIK which isnt much in this case). It is the blockchain storage structuring that I am intrested in at this point it time.
1109 2012-07-01 22:15:46 <gmaxwell> (not the most efficient way to dump sizes.. but :shrugs:)
1110 2012-07-01 22:16:04 <Matt_von_Mises> Can the secp256k1 EC keys be used for asymmetric encryption?
1111 2012-07-01 22:16:06 * jgarzik wonders
1112 2012-07-01 22:16:12 <jgarzik> under what model is "getheaders" ever used?
1113 2012-07-01 22:16:27 <sipa> Matt_von_Mises: there is no EC system for encryption that i know of
1114 2012-07-01 22:16:29 * jgarzik would have thought it would return more transaction data, at least a full merkle or list of txid's
1115 2012-07-01 22:16:49 <Matt_von_Mises> I don't know much about it: http://en.wikipedia.org/wiki/Elliptic_curve_Diffie%E2%80%93Hellman
1116 2012-07-01 22:16:55 <gmaxwell> jgarzik: I thought the SPV nodes used getheaders for points in time prior to the earliest key in their wallet.
1117 2012-07-01 22:17:09 <jrmithdobbs> gmaxwell: i thought you might already have it in a csv or something, ha
1118 2012-07-01 22:17:10 <sipa> Matt_von_Mises: but there is EC-IEC; which uses EC diffie-hellman to establish a shared secret key, and then uses AES to encrypt/decrypt using it
1119 2012-07-01 22:17:10 <gmaxwell> Matt_von_Mises: ECDH is key argreement, not encryption.
1120 2012-07-01 22:17:15 minimoose has quit (Quit: minimoose)
1121 2012-07-01 22:17:20 <Matt_von_Mises> Ah
1122 2012-07-01 22:17:22 <gmaxwell> jrmithdobbs: only old copies.
1123 2012-07-01 22:17:27 <jgarzik> gmaxwell: in theory.  don't see it in practice ATM
1124 2012-07-01 22:17:47 <gmaxwell> jgarzik: hmph. I was pretty sure that multibit didn't scan past blocks.
1125 2012-07-01 22:17:54 <gmaxwell> TD: ?
1126 2012-07-01 22:18:12 <TD> ?
1127 2012-07-01 22:18:21 <jgarzik> my 3 public nodes don't seem to be seeing any 'getheaders' traffic
1128 2012-07-01 22:18:24 <sipa> TD: does BitcoinJ use getheaders?
1129 2012-07-01 22:18:27 <gmaxwell> TD: what if anything do you use getheaders for?
1130 2012-07-01 22:18:29 <TD> yes
1131 2012-07-01 22:18:40 <TD> fast catchup from the last checkpoint for new users
1132 2012-07-01 22:19:01 <TD> where by "checkpoint" i mean whatever the shipped block chain head is. it doesn't do real checkpointing yet
1133 2012-07-01 22:19:21 <TD> new users of multibit/android wallet are probably still a rare enough occurrence that you won't see that traffic on some random nodes
1134 2012-07-01 22:19:23 ThomasV has quit (Quit: Quitte)
1135 2012-07-01 22:19:48 <TD> especially as it just uses the DNS seeds each time. so i suspect those clients connect to the same nodes over and over. using addr broadcasts is another todo list item
1136 2012-07-01 22:21:46 <jgarzik> TD: it connects to nothing other than the DNS seed-returned nodes?
1137 2012-07-01 22:22:31 <sipa> hmm, we use over 160 bytes to store block index data
1138 2012-07-01 22:22:40 <sipa> should be able to do it with less than 50
1139 2012-07-01 22:23:11 <jgarzik> sipa: if it's truly just an index you just need a file number and file offset
1140 2012-07-01 22:23:28 <jgarzik> hash->file locator->data
1141 2012-07-01 22:23:39 <sipa> jgarzik: i mean eveything CDiskBlockIndex is storing
1142 2012-07-01 22:24:38 <jgarzik> oh yeah, height too
1143 2012-07-01 22:24:48 <jgarzik> personally I would ditch everything but file, offset and height
1144 2012-07-01 22:24:49 <sipa> height isn't stored afaik
1145 2012-07-01 22:25:04 <jgarzik> sipa: sure it is
1146 2012-07-01 22:25:12 <sipa> oh, right
1147 2012-07-01 22:25:43 <sipa> i thought i added that in my branch, as height is used in the filename, but it was already there
1148 2012-07-01 22:26:27 theodore has quit (Quit: theodore)
1149 2012-07-01 22:26:50 <sipa> jgarzik: not storing the block header at all is a bad idea, i think, as it would require reading the entire blockfile at startup to calculate difficulty, total work, ...
1150 2012-07-01 22:27:00 <luke-jr> hmm
1151 2012-07-01 22:28:30 <gmaxwell> sipa: you could just store the bits though.
1152 2012-07-01 22:29:01 X-Scale has joined
1153 2012-07-01 22:29:10 <sipa> right
1154 2012-07-01 22:29:26 <sipa> or even just the time
1155 2012-07-01 22:29:33 <sipa> that is enough to reconstruct the bits
1156 2012-07-01 22:33:05 ageis has quit (Ping timeout: 240 seconds)
1157 2012-07-01 22:33:16 phungus has quit (Ping timeout: 245 seconds)
1158 2012-07-01 22:33:42 theodore has joined
1159 2012-07-01 22:34:17 DBordello has quit (Ping timeout: 260 seconds)
1160 2012-07-01 22:34:58 Tril has quit (Ping timeout: 276 seconds)
1161 2012-07-01 22:35:00 <sipa> but even just the fact that a separate db entry is used for each index entry... if you concatenate a few hundred of them: you don't need to store hashPrev for each, time can be stored as difference with the previous, difficulty can be stored just once, height becomes trivial, ...
1162 2012-07-01 22:36:14 Tril has joined
1163 2012-07-01 22:36:18 DBordello has joined
1164 2012-07-01 22:36:37 Tril is now known as Guest84189
1165 2012-07-01 22:39:40 ageis has joined
1166 2012-07-01 22:40:51 phungus has joined
1167 2012-07-01 22:42:29 <sipa> plus that would reduce startup time in case the block db isn't available in OS cache
1168 2012-07-01 22:47:52 theodore has quit (Quit: theodore)
1169 2012-07-01 22:49:50 AlexWaters has joined
1170 2012-07-01 22:50:26 Matt_von_Mises has quit (Quit: Leaving.)
1171 2012-07-01 22:54:02 <jrmithdobbs> sipa: i don't see how to specify the onion address to advertise
1172 2012-07-01 22:54:30 <gmaxwell> jrmithdobbs: externalip
1173 2012-07-01 22:54:53 <gmaxwell> jrmithdobbs: RTFM, it has good guidance. (or if you did, FTFM (fix the fine manual) please)
1174 2012-07-01 22:55:31 <jrmithdobbs> gmaxwell: ya, i'm confused on how i'm supposed to pass it because i've got the manual and sipa's description via email jumbled in my head i think
1175 2012-07-01 22:55:54 <sipa> email...?
1176 2012-07-01 22:56:05 <jrmithdobbs> on the list you linked to something or otehr
1177 2012-07-01 22:56:23 <jrmithdobbs> with a description of how you were encoding the onion addresses
1178 2012-07-01 22:56:58 <sipa> and that same mail pointed to doc/Tor.txt for how to use it in practice
1179 2012-07-01 22:57:03 <sipa> see https://github.com/bitcoin/bitcoin/blob/817ee0d826087d418b5f0bffcdd92429574284e2/doc/Tor.txt
1180 2012-07-01 22:58:52 sirk390 has joined
1181 2012-07-01 22:59:05 <jrmithdobbs> i missed your example at the very bottom, ya
1182 2012-07-01 22:59:38 <sipa> hmm, that file should also mention -bind and -bind=127.0.0.1 in particular
1183 2012-07-01 23:00:37 <jrmithdobbs> the -? output for externalip could probably use a minor update too with examples of ipv4/6/onion addresses
1184 2012-07-01 23:01:05 <gmaxwell> jrmithdobbs: perhaps a "see doc/tor.txt" somewhere in the help output?
1185 2012-07-01 23:01:32 <jrmithdobbs> gmaxwell: no because it's not just onion it doesn't mention notation for ipv6 either
1186 2012-07-01 23:02:38 graingert_ has joined
1187 2012-07-01 23:02:41 <jrmithdobbs> there's several v4 examples (esp 127.0.0.1) through but none of the others ;p
1188 2012-07-01 23:02:52 <jrmithdobbs> s/through/through out/
1189 2012-07-01 23:03:34 <jrmithdobbs> maybe it should be mentioned at the top or bottom somewhere
1190 2012-07-01 23:03:43 <jrmithdobbs> of -? not on externalip
1191 2012-07-01 23:10:01 <jrmithdobbs> sipa: can tor nodes be seeded via dns using those packed addrs as aaaa records or a better way?
1192 2012-07-01 23:10:43 sirk390 has quit (Quit: Leaving.)
1193 2012-07-01 23:11:02 <jrmithdobbs> sipa: I don't understand what's meant by "can connect to seed as proxy" ?
1194 2012-07-01 23:11:18 <gmaxwell> jrmithdobbs: I don't think thats a good way to seed tor, simply because DNS queries for dnsseed don't really work over tor.
1195 2012-07-01 23:11:48 <jrmithdobbs> gmaxwell: aye, i'm trying to run a bridged node basically
1196 2012-07-01 23:12:20 <gmaxwell> I expect that eventually we'll hardcode a seed list of hidden service nodes and just let addr rumoring take care of the rest.
1197 2012-07-01 23:12:38 <jrmithdobbs> ya, why i was bringing it up
1198 2012-07-01 23:12:56 <gmaxwell> where is "can connect to seed as proxy" that you're quoting?
1199 2012-07-01 23:13:00 <jrmithdobbs> was wondering if y'all had come up with anything besides hardcoded seeds
1200 2012-07-01 23:13:24 <jrmithdobbs> gmaxwell: -dnsseed in Tor.txt
1201 2012-07-01 23:13:45 <gmaxwell> -dnsseed DNS seeds are not resolved directly when a SOCKS5 proxy server is set. Rather, a short-lived proxy connection to the dns seed hostname is attempted, and peer addresses are requested.
1202 2012-07-01 23:14:06 <jrmithdobbs> wait, is that describing socks?
1203 2012-07-01 23:14:24 <gmaxwell> jrmithdobbs: basically it just connects to the dnsseed domain names like nodes.
1204 2012-07-01 23:15:03 <gmaxwell> meaning it only effectively gets one peer per dnsseed.. though the addr rumoring takes care of the rest from there.
1205 2012-07-01 23:15:21 <jrmithdobbs> can I suggest the wording on that be changed to something like "dns seeded names will be resolved through the socks4a or socks5 name resolution methods."
1206 2012-07-01 23:15:28 <jrmithdobbs> when using a socks proxy*
1207 2012-07-01 23:15:52 <gmaxwell> jrmithdobbs: that isn't what it does. It doesn't resolve them, it just connects to them by name.
1208 2012-07-01 23:16:14 <gmaxwell> (stupid resolving via socks can only return one address, at least via tor, so it's not generally useful to us)
1209 2012-07-01 23:16:44 <jrmithdobbs> well, it's the same thing, can only resolve to one or connect to one, i get what it means now
1210 2012-07-01 23:16:52 <jrmithdobbs> just confusing
1211 2012-07-01 23:22:04 <jrmithdobbs> gmaxwell: just had an interesting idea on an extension of the altruism incintive
1212 2012-07-01 23:24:10 * gmaxwell waits
1213 2012-07-01 23:24:12 <jrmithdobbs> gmaxwell: if you no longer have to validate the whole chain just current state back a couple hundred blocks, because of how cheap you've made it (10s of MB 1000s of cycles of cpus so long as p2p is throttled in connections a bit) ... there's actually a not horrible argument for trying to get it included (wallet-less) in things like tails/torbundle/etc, as that *is* a good target audience for those kinds of expectations
1214 2012-07-01 23:24:43 <jrmithdobbs> eg, a stripped down verifier p2p node
1215 2012-07-01 23:24:50 <jrmithdobbs> we already know it uses no bandwidth
1216 2012-07-01 23:25:06 <TD> jgarzik: right
1217 2012-07-01 23:25:22 <jrmithdobbs> and it aids tor in adding legit traffic with an interesting traffic profile to the mix
1218 2012-07-01 23:25:24 silp has joined
1219 2012-07-01 23:25:41 <jrmithdobbs> everybody wins
1220 2012-07-01 23:25:56 <TD> jgarzik: it's not great, but recording addr broadcasts, using them, ensuring you still start fast, etc, it's a bunch of code that isn't written yet
1221 2012-07-01 23:26:20 <TD> btw, what's the reason for asking? somebody wants to remove getheaders?
1222 2012-07-01 23:26:48 <gmaxwell> jrmithdobbs: sure. Also, you can trade diskspace for bandwidth. E.g. you can do full validation with no disk usage at all, so long as you have untrusted peers who are willing to give you the required tree fragments.  Also, nodes doing that could choose to validate only one in n blocks. If the number of nodes is high enough or n is small enough then it would be vanishingly unlikely that none would validate any given block.
1223 2012-07-01 23:27:09 <gmaxwell> TD: jgarzik is currently adding per-peer message stats.
1224 2012-07-01 23:27:22 <TD> i see
1225 2012-07-01 23:29:06 Guest84189 has left ()
1226 2012-07-01 23:29:21 <TD> getheaders is a very useful optimization
1227 2012-07-01 23:29:43 <TD> it takes only a minute or two to fully catch up with the head of chain, usually, and that's with a bunch of missed optimization opportunities
1228 2012-07-01 23:29:46 <TD> so it's well worth having
1229 2012-07-01 23:31:59 minimoose has joined
1230 2012-07-01 23:36:23 membersonlyguy has joined
1231 2012-07-01 23:36:31 elkingrey has joined
1232 2012-07-01 23:37:09 graingert__ has joined
1233 2012-07-01 23:38:55 TD has quit (Quit: TD)
1234 2012-07-01 23:39:40 graingert_ has quit (Ping timeout: 244 seconds)
1235 2012-07-01 23:40:30 <jgarzik> sigh, TD gone again :)
1236 2012-07-01 23:40:36 <jgarzik> demands of a 2-year-old
1237 2012-07-01 23:40:58 <jgarzik> for the record, I'm not interested in removing getheaders...  just wondered about the usage, since my per-msg stats were showing very little usage
1238 2012-07-01 23:41:24 silp has left ()
1239 2012-07-01 23:41:26 <jgarzik> but anyway... ouch
1240 2012-07-01 23:41:36 <jgarzik> BitcoinJ only uses the nodes directly returned from DNS seeds
1241 2012-07-01 23:41:54 <jgarzik> that is quite a different picture from what I had assumed
1242 2012-07-01 23:42:16 <jgarzik> makes it more urgent that I find a non-static DNS seed setup, too
1243 2012-07-01 23:43:39 Apexseals has quit (Ping timeout: 246 seconds)
1244 2012-07-01 23:44:04 Apexseals has joined
1245 2012-07-01 23:45:42 <jgarzik> bleh
1246 2012-07-01 23:45:52 da2ce7 has quit (Ping timeout: 250 seconds)
1247 2012-07-01 23:46:01 <luke-jr> jgarzik: I like sipa's bitcoin-seeder (though it does use quite a bit of CPU time)
1248 2012-07-01 23:46:03 Geebus has quit (Ping timeout: 245 seconds)
1249 2012-07-01 23:46:07 <jgarzik> I need to find someone who can do HTML/CSS for BTC.  Graphic design work is just not my interest.
1250 2012-07-01 23:46:27 <jgarzik> oh that reminds me...  'getpeers' RPC!
1251 2012-07-01 23:46:30 <luke-jr> I'd ask in #bitcoin-otc
1252 2012-07-01 23:47:03 Zarutian has quit (Quit: Zarutian)
1253 2012-07-01 23:49:12 <gmaxwell> jgarzik: the fact that one of the dnsseeds is static isn't a big deal.  I think the diversity of methods is good.
1254 2012-07-01 23:52:43 <jgarzik> gmaxwell: two factors are amplified for any nodes listed via DNS seed:  more likely to serve IBDs, and far more likely to be used by BitcoinJ
1255 2012-07-01 23:53:41 <jgarzik> two fixable negative incentives
1256 2012-07-01 23:54:22 <jgarzik> and a static DNS seed inflicts more of such "punishment" than a more dynamic one
1257 2012-07-01 23:54:23 Apexseals has quit ()
1258 2012-07-01 23:56:02 <gmaxwell> sure. At the same time, I can rob bitcoinj nodes blind if I can convince DNS seeds to only give out the addresses of nodes I control.
1259 2012-07-01 23:57:57 <jgarzik> yep
1260 2012-07-01 23:58:58 * jgarzik 's view is admittedly colored by "running a public node" viewpoint
1261 2012-07-01 23:59:22 <jgarzik> does anybody keep stats on the number of public(listening) nodes, across months/years?