1 2013-12-10 01:58:35 <amiller> okay so...
  2 2013-12-10 01:58:52 <amiller> what happens when you add a new key to a wallet managed by an spv bloom filtering node
  3 2013-12-10 01:59:27 <amiller> you tell your peers to update their bloom filter, so that means you are going to hear about any new transactions/blocks
  4 2013-12-10 02:00:10 <amiller> if you generated the key from scratch, then you can rely on it not being present in any old blocks
  5 2013-12-10 02:00:19 <amiller> what happens when you import a key? does that trigger a rescan?
  6 2013-12-10 02:00:43 <gmaxwell> amiller: you can import a key with a date, and it rescans from there.
  7 2013-12-10 02:00:56 <amiller> do keys normally come with dates?
  8 2013-12-10 02:01:02 <gmaxwell> if you don't have a date with it— its rescanning from day 1.
  9 2013-12-10 02:01:20 <amiller> so a filtering node will start asking its peers for every block forever?
 10 2013-12-10 02:01:24 <gmaxwell> amiller: the multibit wallet format does, WIF private keys often do not, mostly because we didn't think about it in the initial serialization.
 11 2013-12-10 02:01:54 <gmaxwell> amiller: it'll ask it for every past block, w/ filtering, once at import time and then its caught up for that key.
 12 2013-12-10 02:01:56 <amiller> that seems like a lot of burden on the peers...they have to scan through the whole blockchain including applying new filters.. if i understand that right
 13 2013-12-10 02:02:02 <amiller> okay.
 14 2013-12-10 02:02:46 <gmaxwell> In _general_ importing a key shouldn't be a normal case, as it totally busts the durability of backups in a way which may be surprising to users. (especially for clients with wallets which are normally determinstic)
 15 2013-12-10 03:41:59 <Blaster> hello... if I show a bitcoin address on my web app, how can I make the app aware when that address receives payment?   Do I use the JSON-RPC?
 16 2013-12-10 03:42:44 <weex> Blaster: yes, I'd use listtransactions
 17 2013-12-10 03:43:24 <Blaster> I want it to generate a new address when a user clicks a link, and then when they pay that address, I want to display a popup window
 18 2013-12-10 03:43:29 <weex> but be sure to use both transaction id and address to uniquely identify a payment
 19 2013-12-10 03:43:57 <Blaster> weex: why transaction id?
 20 2013-12-10 03:44:37 <weex> well in the past i only used txid but then some people made multioutput transactions
 21 2013-12-10 03:44:47 <kjj> https://bitcointalk.org/index.php?topic=269886.0
 22 2013-12-10 03:45:03 <kjj> should be in the FAQ or something
 23 2013-12-10 03:45:24 <Blaster> walletnotify?
 24 2013-12-10 03:45:32 <Blaster> is that an event that gets fired?
 25 2013-12-10 03:46:02 <kjj> it runs a process on the local (to bitcoind) machine
 26 2013-12-10 03:46:27 <Blaster> How quickly can I verify payment has been received after it gets sent?  I don't want to make my customers wait several minutes for a confirmation
 27 2013-12-10 03:46:53 <kjj> as soon as your node sees the transaction come in over the network
 28 2013-12-10 03:47:18 <kjj> but be aware that you may get notified more than once for a transaction.  read that whole thread
 29 2013-12-10 03:47:24 <Blaster> ok
 30 2013-12-10 03:48:50 <Blaster> it says to check the confirmation count of the transactions, not the balance of the account, but if you do that how can you verify the correct amount was sent?
 31 2013-12-10 03:49:18 <kjj> decode the transaction
 32 2013-12-10 03:49:40 <kjj> you should probably get accounts right out of your head now if you are working on a web service.
 33 2013-12-10 03:50:20 <Blaster> doesn't it mean wallet ID by account?
 34 2013-12-10 03:50:50 <kjj> I'm not sure what you mean by wallet ID
 35 2013-12-10 03:51:03 <kjj> but I can tell you that accounts do not do what you expect, and they will not help you at all
 36 2013-12-10 03:51:35 <weex> accounts are no good
 37 2013-12-10 03:51:44 <Blaster> isn't the wallet ID what someone would send a payment to?  I think I mean address.
 38 2013-12-10 03:51:44 <weex> better to call them crazy labels
 39 2013-12-10 03:52:20 <kjj> ok, when you are presenting a payment page, you generate a new address and tuck it into your database
 40 2013-12-10 03:52:37 <Blaster> ok
 41 2013-12-10 03:52:51 <kjj> then, when walletnotify triggers, you decode it, check the outputs.  compare the addresses to your database, along with the amounts
 42 2013-12-10 03:53:10 <andytoshi> Blaster: http://download.wpsoftware.net/bitcoin/bitcoin-faq.pdf might help untangle addresses and accounts for you
 43 2013-12-10 03:53:15 <Blaster> I read that you should pipe wallet notify to a file so it doesn't cause any hang ups
 44 2013-12-10 03:53:54 <kjj> that thread describes a client/server model
 45 2013-12-10 03:54:19 <kjj> if you are SURE that your site won't get a lot of traffic, you can get away with some sketchy shit
 46 2013-12-10 03:54:22 <Blaster> so is what I am trying to do though, isn't it?
 47 2013-12-10 03:54:40 <Blaster> well my site will be getting a lot of traffic so I do want to do at least a decent job
 48 2013-12-10 03:54:52 <Blaster> define: a lot of traffic
 49 2013-12-10 03:55:08 <kjj> it varies.
 50 2013-12-10 03:55:35 <Blaster> let's say < 5000 tx per month
 51 2013-12-10 03:55:47 <kjj> basically, walletnotify triggers at least once per transaction involving your wallet.  a big chunk of the bitcoind node will pause until that process ends
 52 2013-12-10 03:55:56 <Blaster> I consider that a lot, but it's probably a small amount if you're an exchange or something
 53 2013-12-10 03:56:33 <Blaster> oh yeah, well I think piping to a file is probably a good idea
 54 2013-12-10 03:56:34 <kjj> if that process does something slow, it can get backed up and crash your box, or at least your node
 55 2013-12-10 03:56:40 <Blaster> then ending immediately
 56 2013-12-10 03:56:52 <kjj> so, you want a very light, very fast process that does the bare minimum, then ends.
 57 2013-12-10 03:57:14 <Blaster> yeah
 58 2013-12-10 03:57:17 <kjj> then you have either a long running process or a cron job that does the slower processing in an orderly fashion
 59 2013-12-10 03:57:43 <Blaster> does it make sense to store all the TX from walletnofify in the DB?
 60 2013-12-10 03:58:13 <kjj> again, depends on what you'll need
 61 2013-12-10 03:58:25 <andytoshi> hey devs, i am doing a manual coinjoin, /msg me or email apoelstra@wpsoftware dot net to get in on it
 62 2013-12-10 03:58:43 <kjj> some TXs can get huge, so you may want to include only the address of yours that is getting paid, and the amount
 63 2013-12-10 03:58:59 <gmaxwell> andytoshi: you might want to mention that the optimal txout side for your txn is 1 BTC.
 64 2013-12-10 03:59:16 <Blaster> I want to display a modal immediately when payment is received, so I think my javascript will be sending an AJAX request every few seconds to check if the payment has been received, the script that the AJAX requests will probably be either parsing that file with the notifywallet output, or checking the DB
 65 2013-12-10 03:59:25 <kjj> I'll go in for 1 BTC
 66 2013-12-10 04:00:06 <andytoshi> cool kjj
 67 2013-12-10 04:00:14 <andytoshi> i don't understand "you may want to include only the address of yours that is getting paid, and the amount"
 68 2013-12-10 04:00:19 <kjj> how manual?
 69 2013-12-10 04:00:36 <andytoshi> well, i am asking people to msg me utxo's, then i'm createrawtransacting them together
 70 2013-12-10 04:00:46 <kjj> andytoshi: that was to blaster
 71 2013-12-10 04:01:01 <kjj> like in vi?  or do you have a script?
 72 2013-12-10 04:01:03 <Blaster> kjj: what do you mean by how manual?
 73 2013-12-10 04:01:13 <kjj> Blaster: that was to andy
 74 2013-12-10 04:01:20 <Blaster> Which was to me?
 75 2013-12-10 04:01:24 <andytoshi> kjj: createrawtransaction just needs txid's
 76 2013-12-10 04:01:45 <andytoshi> then the resultant hash, i send to all you guys and you each sign
 77 2013-12-10 04:01:49 <andytoshi> not hash
 78 2013-12-10 04:01:51 <Blaster> kjj: 1 BTC?
 79 2013-12-10 04:01:56 <andytoshi> resultant hex-encoded transaction
 80 2013-12-10 04:02:31 <kjj> Blaster: "some TXs can get huge..." was to you
 81 2013-12-10 04:02:57 <Blaster> oh ok
 82 2013-12-10 04:03:06 <Blaster> so does the process above sound the right way?
 83 2013-12-10 04:03:19 <Blaster> what would be better in that scenario, parsing the text file, or selecting from the DB?
 84 2013-12-10 04:04:27 <kjj> Blaster: if you have a DB, hit that.  the DB will at least pretend to try to be ACID
 85 2013-12-10 04:04:56 <andytoshi> gmaxwell: do you know roughly how many inputs will force us to add a fee
 86 2013-12-10 04:05:02 <andytoshi> one player wants to add over a dozen..
 87 2013-12-10 04:07:27 <gmaxwell> andytoshi: 10kb will need a fee. just assume everyone is using uncompressed keys and estimate the size.
 88 2013-12-10 04:07:37 <andytoshi> kk thx
 89 2013-12-10 04:08:35 <gmaxwell> and the fee should be at least 0.0001 btc/kb once there is a fee.. unless you go get wizkid or luke to play and they can prioritize. :)
 90 2013-12-10 04:08:47 <gmaxwell> might be an optimal strategy. :)
 91 2013-12-10 04:09:20 <kjj> should I be worried that I now know the syntax of createrawtransaction by muscle memory?
 92 2013-12-10 04:09:40 <andytoshi> kjj: oh yes, your bashisms will be more destructive than ever
 93 2013-12-10 04:09:57 <andytoshi> rm . * is peanuts :)
 94 2013-12-10 04:10:32 <kjj> DWIM: corrected "rm . *" to "rm /*"
 95 2013-12-10 04:12:12 <andytoshi> oh, i meant to say * .o
 96 2013-12-10 04:12:48 <gmaxwell> kjj: only worry when you can hexedit txn to change nlocktime and ad op_return txn without thinking about it.
 97 2013-12-10 04:13:00 <gmaxwell> ... and not get tripped out by byteorder on longer values. :P
 98 2013-12-10 04:14:12 <Blaster> how many GB is the block chain currently?
 99 2013-12-10 04:14:44 <kjj> 11.7 for blocks
100 2013-12-10 04:14:58 <kjj> wait, that's not right
101 2013-12-10 04:15:29 <kjj> 12.9
102 2013-12-10 04:15:36 <Blaster> I don't think I have the HDD space to store it on my development VM.  Damn SSD's.
103 2013-12-10 04:16:03 <kjj> didn't I hear that TB SSDs were coming out soon?
104 2013-12-10 04:17:11 <Blaster> probably, but only affordable for a trillionaire
105 2013-12-10 04:17:49 <kjj> if it makes you feel any better, my main p2pool node is running the blockchain in a RAMdisk
106 2013-12-10 04:18:29 <Blaster> cool, how much ram?
107 2013-12-10 04:18:47 <Blaster> price and capacity
108 2013-12-10 04:19:08 <amiller> it's weird to me the trust model with bloom filtering
109 2013-12-10 04:19:31 <kjj> the box has 32 gigs.  4 sticks of 8, probably like $160 total when I bought them
110 2013-12-10 04:19:55 <amiller> if you have to trust the node and it has to provide you some services, why stop at just the bloom filter on new tx and getblocks
111 2013-12-10 04:19:57 <andytoshi> hi amiller, do you want to get in on a big coinjoin?
112 2013-12-10 04:20:06 <andytoshi> msg me a txid/vout
113 2013-12-10 04:20:18 <amiller> what coin size
114 2013-12-10 04:20:21 <andytoshi> 1btc
115 2013-12-10 04:20:27 <amiller> pff
116 2013-12-10 04:20:40 <Blaster> what's a coinjoin?
117 2013-12-10 04:20:42 <andytoshi> well, you are welcome to give me several txids
118 2013-12-10 04:20:45 <andytoshi> :}
119 2013-12-10 04:21:13 <andytoshi> Blaster: when you spend bitcoins, you sign a transaction which passes your coins to another person
120 2013-12-10 04:21:30 <andytoshi> what we are doing is, we are combining all our transactions into one, then we each sign the big transaction
121 2013-12-10 04:21:47 <Blaster> andytoshi: what's the benefit of doing that?
122 2013-12-10 04:22:05 <andytoshi> Blaster: normally when you have multiple coins in the same transaction, it is because the same person owns all the coins
123 2013-12-10 04:22:23 <andytoshi> so people trying to track movement of money can use that to discover identities
124 2013-12-10 04:22:29 <kjj> this way, when the feds come knocking on my door, I can tell them that Greg did it
125 2013-12-10 04:22:42 <andytoshi> and greg will blame me, and i'll blame kjj
126 2013-12-10 04:22:45 <Blaster> why would the feds come knocking on your door for a BTC tx?
127 2013-12-10 04:23:02 <kjj> feds get bored
128 2013-12-10 04:23:51 <kjj> anyone else have a namecoin node get stuck on block 150021 ?
129 2013-12-10 04:24:20 <amiller> if i could easily express the logic of what to do when new data comes in, then one good thing to do would be to have filtered "sweeps"
130 2013-12-10 04:25:14 <amiller> in other words if i don't mind waiting and orderness problems, it would be efficient to have your trusty peer node do scans every so often and notify you if it collects any changes
131 2013-12-10 04:48:21 <kjj> one problem that I see with coinjoin is the output size thing.
132 2013-12-10 04:49:41 <kjj> to do things other than just mixing back into our own wallets, we'd need a pile of them at multiples of 2
133 2013-12-10 04:50:27 <kjj> and winding into and out of a stack of joins to get the error under the limit that you care about could take all day
134 2013-12-10 04:50:29 <andytoshi> yeah
135 2013-12-10 04:50:43 <andytoshi> a few people have already wrecked the size-1 rule
136 2013-12-10 04:51:00 <andytoshi> so if you wanna split yours up you are welcome to
137 2013-12-10 04:51:05 <kjj> I intend someday to script this and automate it, but still...
138 2013-12-10 04:51:17 <kjj> 1 BTC is fine by me.  I'm just in it for the notoriety
139 2013-12-10 04:51:21 <amiller> i like how it begins with a txoutput.
140 2013-12-10 04:51:32 <amiller> i'm working on a script to do iddo's lottery game.
141 2013-12-10 04:51:39 <amiller> i'm not relaly sure how it should go though
142 2013-12-10 04:51:54 <amiller> the main tricky thing is that commitments have to be secure ahead of time
143 2013-12-10 04:52:13 <kjj> I have a decent outline of a coinjoiner service, but every time I start writing it, real life pulls me away
144 2013-12-10 04:52:52 <andytoshi> i still haven't checked out iddo's lottery game
145 2013-12-10 04:53:06 <andytoshi> it'd be cool to have a coinjoin service..i'd have to think about how well we could keep people anonymous
146 2013-12-10 04:53:18 <andytoshi> i suppose the first few people to sign pretty-much have to reveal their hand
147 2013-12-10 04:53:22 <kjj> don't bother, at first
148 2013-12-10 04:53:27 <amiller> andytoshi, did you see the eprint technical paper that commemorates it with formal drawing and such
149 2013-12-10 04:53:38 <andytoshi> no, i didn't
150 2013-12-10 04:53:51 <kjj> accept that the service can break your privacy.  use layering
151 2013-12-10 04:54:09 <kjj> and later on, get into advanced stuff like chaum tokens to blind everything
152 2013-12-10 04:54:13 <amiller> andytoshi, http://eprint.iacr.org/2013/784 <- really great paper about lottery protocol in bitcoin
153 2013-12-10 04:54:31 <andytoshi> oh, thx amiller, i do have that one
154 2013-12-10 04:54:36 <andytoshi> but it'll be a while before i can read it
155 2013-12-10 04:54:42 <amiller> k
156 2013-12-10 04:54:56 <amiller> well, anyway, there are about three steps to the lottery protocol
157 2013-12-10 04:55:15 <amiller> first you have to make some security deposits with someone's public key attached.
158 2013-12-10 04:55:28 <amiller> if you know who you're going to want to play with, you can totally pre-compute all of these
159 2013-12-10 04:56:15 <amiller> to play the game, you propose and agree on a wager with the one or two other people you're playing against
160 2013-12-10 04:56:51 <amiller> then everyone exchanges timelock transactions that show that either you release your committed value by time T, or else the timelock transactions let everyone take the security deposits from you
161 2013-12-10 04:57:46 <amiller> then you sign the wager, wait for everyone to reveal their value and take their security deposits back, and one person also gets to take the pot
162 2013-12-10 04:58:12 <andytoshi> cool
163 2013-12-10 04:58:27 <amiller> so i need to make a gambling wallet
164 2013-12-10 04:58:28 <andytoshi> it's definitely not clear how this can be ordered safely
165 2013-12-10 04:58:36 <andytoshi> but i think if i study it i'll see
166 2013-12-10 05:01:05 <andytoshi> it'd be cool if we bet this coinjoin :P
167 2013-12-10 05:03:32 <amiller> one problem is that this is a lottery you can only play on eligius :p
168 2013-12-10 05:03:35 <amiller> it uses non-standard transactions
169 2013-12-10 05:03:50 <amiller> but... my gambling wallet can still make use of spv bloom filters and stuff to find them easily.
170 2013-12-10 05:09:55 <andytoshi> ok, thx everyone for participating in the coinjoin -- i'll start collecting signatures tomorrow afternoon/evening
171 2013-12-10 05:10:06 <andytoshi> until then, anyone who wants to join is welcome to
172 2013-12-10 05:10:13 <andytoshi> for now i'm going to bed
173 2013-12-10 05:10:46 <lechuga> coinjoin?
174 2013-12-10 05:11:13 <lechuga> nm ic
175 2013-12-10 05:14:01 <lechuga> is this like tor for the blockchain
176 2013-12-10 05:14:32 <amiller> pretty close.
177 2013-12-10 05:14:50 <amiller> though it's a little more like the old mixmailers
178 2013-12-10 05:14:58 <amiller> it's a little slower and less automated as tor.
179 2013-12-10 05:15:40 <gmaxwell> 20:53 < andytoshi> i suppose the first few people to sign pretty-much have to reveal their hand
180 2013-12-10 05:16:53 <gmaxwell> thats why I talked about chaum tokens. There is no reason the service needs to be able to break people's privacy.  Esp it's easy to achieve security in the "honest but curious" model, where the server might snoop if it was easy, but wouldn't outright behave maliciously
181 2013-12-10 05:17:17 <amiller> yeah, also there are flat-out straightforward to use implementation of chaum tokens
182 2013-12-10 05:17:21 <amiller> like Lucre
183 2013-12-10 05:17:25 <amiller> by ben laurie
184 2013-12-10 05:17:37 <gmaxwell> maaku implemented chaum blinding. it's trivial
185 2013-12-10 05:18:03 <kjj> what's the short version of implementing chaums for coinjoin?
186 2013-12-10 05:19:24 <gmaxwell> You take your output address and blind it.  You give me an input you want to spend and your blinded output address. I sign the blinded output address and return it. You unblind.. and come back later from a new tor connection and give me the address with my signature on it.
187 2013-12-10 05:19:31 <lechuga> hmm chaum tokens
188 2013-12-10 05:20:02 <gmaxwell> if there is more than one output token I need to use one public key per value, these could be well known so I couldn't substitute easily to deanonymize people
189 2013-12-10 05:20:12 <kjj> what does blinding look like, and what kind of signature survives the inverse operation?
190 2013-12-10 05:21:50 <gmaxwell> kjj: https://en.wikipedia.org/wiki/Blind_signature#Blind_RSA_signatures.5B2.5D:235
191 2013-12-10 05:23:24 <gmaxwell> the attack that exists in this system is either when the server can watermark via changing the keys... or if it makes you meet up with nothing but sybils.
192 2013-12-10 05:23:53 <gmaxwell> the former you can design out easily enough, the latter is probably harder to design out.
193 2013-12-10 05:24:18 <gmaxwell> but the blind signature thing is so simple to do, it kinda seems silly to not do it in a complete system.
194 2013-12-10 05:24:21 <kjj> ok, so the signature passes ordinary verification?
195 2013-12-10 05:24:28 <gmaxwell> kjj: yes.
196 2013-12-10 05:25:25 <kjj> can the signer publish a large list of pubkeys in advance and let each participant pick ones at random to use?
197 2013-12-10 05:26:47 <kjj> and when you say you need multiple pubkeys for multiple outputs, do you mean per batch, or ever?
198 2013-12-10 05:27:22 <gmaxwell> yes, but what if each participant picks a different one?   one possiblity would be for some cryptosystems (e.g. blind signing with ec-schnorr) to just derrive the pubkey with server_root+H(value)
199 2013-12-10 05:28:19 <gmaxwell> kjj: you want the server not to be able to tell two otherwise equal outputs apart. If those two outputs use two different keys, the server can.
200 2013-12-10 05:28:57 <gmaxwell> what you also don't want is someone giving you a 1 BTC input and then trying to give you a 5 btc output so there must be some binding on the values.
201 2013-12-10 05:29:43 <kjj> gmaxwell: the way I'm designing my system, all outputs will be equal, and all inputs must be at least that amount
202 2013-12-10 05:30:01 <gmaxwell> then you only need one key.
203 2013-12-10 05:30:05 <maaku> kjj: there's really no reason to mandate equal size outputs
204 2013-12-10 05:30:23 <gmaxwell> but yea, seems like a needless constrant, even if its preferred.
205 2013-12-10 05:30:46 <kjj> it makes it easier to assemble transactions
206 2013-12-10 05:30:49 <maaku> my protocol has a setup phase where someone suggests batches of output sizes (based on constraints provided by the participants)
207 2013-12-10 05:31:03 <maaku> the joiner then generates a unique key per batch
208 2013-12-10 05:31:52 <kjj> ok, so your join would have (n outputs of 1 BTC, m outputs of 0.5 BTC, o outputs of .25 BTC, etc?
209 2013-12-10 05:32:49 <kjj> and then people could pick the output(s) they need and provide at least that much input?
210 2013-12-10 05:33:43 <kjj> yeah, that can work.
211 2013-12-10 05:36:11 <kjj> is there a EC version of blind signing?
212 2013-12-10 05:38:22 <gmaxwell> kjj: sure, but not for ecdsa (except obnoxiously)
213 2013-12-10 05:39:08 <gmaxwell> schnorr is trivial to blind.
214 2013-12-10 05:39:44 <warren> gmaxwell: saw someone posted a patch to adjust bitcoin to use fedora's openssl?
215 2013-12-10 05:39:50 <gmaxwell> which might result in a tidy thing where you just sign with pubkey+H(amount)  and then the person gives you the unblinded value w/ amount.
216 2013-12-10 05:40:04 <gmaxwell> warren: yea, by copying the parts of openssl fedora has off into bitcoin. lol
217 2013-12-10 05:40:04 <kjj> are there common command line tools for this?
218 2013-12-10 05:40:10 <gmaxwell> kjj: nope.
219 2013-12-10 05:40:30 <kjj> hmm
220 2013-12-10 05:40:38 <gmaxwell> could be pretty easily. (just some minor modification to ed25519 code)
221 2013-12-10 05:41:09 <kjj> this ends up being a client/server thing.  the client needs to be able to do the blinding
222 2013-12-10 05:41:59 <warren> gmaxwell: so in other words, they have dumb reasons for not shipping the relevant parts of EC
223 2013-12-10 05:42:06 <gmaxwell> IIRC you don't have to change the validation process to do ed25519... just add a blinding function, and a signing function that injests a blinded value.
224 2013-12-10 05:43:06 <kjj> ok, well, I'm going to stick to unblinded for my first run
225 2013-12-10 05:43:24 <kjj> mostly so that people won't need to install anything special to use it
226 2013-12-10 05:44:13 <kjj> well, other than some scripts to convert the BIP10 request into something that bitcoind can sign, and then back into BIP10 submission form
227 2013-12-10 05:44:59 <gmaxwell> ed25519 in javscript has lots of hits. But sure, I don't think there is anything wrong in doing something first thats simple.
228 2013-12-10 05:45:30 <gmaxwell> I think a complete tool should do the blinding because its just some code... one time cost.
229 2013-12-10 05:46:17 <kjj> asking people to install strange new crap on their boxes seem like a bit much for a proof of concept
230 2013-12-10 05:47:24 <kjj> when there are dozens of coinjoiner nodes running, or when ever bitcoin node is one by default, we can put that stuff in the client so that it can churn your coins while you sleep
231 2013-12-10 05:56:52 <kjj> ooh.  "Legacy banking system".  just the right level of condescension
232 2013-12-10 06:06:34 <grau> gmaxwell: did the test case I sent you made sense? Could you please tell me what shatoshi dislikes or what digest it uses?
233 2013-12-10 06:39:29 <maaku> kjj: somewhere on my todo list is figuring out how to wrangle ed25519 into doing schnorr blind sigs as gmaxwell describes, for the purpose of coinjoin
234 2013-12-10 06:39:37 <maaku> in the mean time rsa works fine
235 2013-12-10 06:40:27 <maaku> and rsa you "don't have to install anything special"
236 2013-12-10 06:59:23 <gmaxwell> grau: So I get hash 11646b8f1ef3e60daa4e85a177e6d95c4f2d99d1e88cd15ad5c03c25cdd65ae1 on both your bad transaction and the one generated by bitcoind that passes.
237 2013-12-10 06:59:35 <gmaxwell> I can verify that the ECDSA validation is actually failing.
238 2013-12-10 07:00:30 <grau> gmaxwell: thanks a lot that is some new pointer. I check the hash.
239 2013-12-10 07:02:57 <Blaster> when using bitcoin-qt on ubuntu how do I find out where my wallet and blockchain data is stored?
240 2013-12-10 07:03:14 <Blaster> I want to use the torrent version of the blockchain so I don't have to wait a long time to get caught up
241 2013-12-10 07:04:22 <Blaster>  actually I think I may have found it, inside ~/.bitcoin
242 2013-12-10 07:04:24 <Plarkplark_> Hello. I am having consistent problems with bitcoind on Debian stable.
243 2013-12-10 07:04:32 <Plarkplark_> getinfo displays: Warning: Displayed transactions may not be correct! You may need to upgrade, or other nodes may need to upgrade.
244 2013-12-10 07:04:35 <Plarkplark_>  "version" : 80500,
245 2013-12-10 07:04:53 <Plarkplark_> is the node functional? Or do I need to take action?
246 2013-12-10 07:05:56 <gmaxwell> Plarkplark_: grep nvalid ~/.bitcoin/debug.log
247 2013-12-10 07:06:24 <Plarkplark_> a lot of these
248 2013-12-10 07:06:25 <Plarkplark_> InvalidChainFound:  current best=0000000000001e5aedcfabd7acfadc75fcd8fd38142cce66f60416b2ed56d19b  height=128969  log2_work=63.481279  date=2011-06-06 07:49:05
249 2013-12-10 07:06:28 <Plarkplark_> InvalidChainFound: Warning: Displayed transactions may not be correct! You may need to upgrade, or other nodes may need to upgrade.
250 2013-12-10 07:06:50 <Plarkplark_> mostly  2011/2012 dated.
251 2013-12-10 07:07:56 <gmaxwell> Plarkplark_: you are in a severly corrupted state. knowing the first error before it started producing those might be interesting.
252 2013-12-10 07:08:17 <Plarkplark_> ouch. I just took 24 hours to get synced.
253 2013-12-10 07:08:36 <Plarkplark_> it did bug out at one time because of memory issues, increased mem for it.
254 2013-12-10 07:08:41 <Plarkplark_> any way to rebuild?
255 2013-12-10 07:09:39 <gmaxwell> Plarkplark_: -reindex but something is screwy ... you also haven't made it very var in the validation at all.
256 2013-12-10 07:10:23 <Plarkplark_> its runnig
257 2013-12-10 07:10:28 <Plarkplark_> ProcessBlock: ACCEPTED
258 2013-12-10 07:10:28 <Plarkplark_> SetBestChain: new best=00000000712bcf48c796bd4fb485910002e165614a93e48bfb7608953f5c8c37  height=13175  log2_work=45.685647  tx=13288  date=2009-05-03 21:59:54 progress=0.000221
259 2013-12-10 07:10:35 <grau> gmaxwell: yes, I have a different hash as expected.
260 2013-12-10 07:10:38 <Plarkplark_> So maybe interesting where it goes wrong.
261 2013-12-10 07:11:05 <grau> gmaxwell: could you do one more favor and also dump what is hashed (in hex) by satoshi?
262 2013-12-10 07:11:10 <gmaxwell> grau: okay so now explain how you don't fail the txn signed by bitcoind? :P
263 2013-12-10 07:11:14 <benten> gmaxwell: whats the max number of blocks that can be reversed at a given time?
264 2013-12-10 07:11:45 <gmaxwell> benten: all of them?  .. its not clear to me what you're asking.
265 2013-12-10 07:11:46 <grau> gmaxwell: I know it is more than strange, but we are probably one step from solving the puzzle.
266 2013-12-10 07:12:59 <gmaxwell> I will grab.
267 2013-12-10 07:13:29 <gmaxwell> the only difference I saw was your signature (I think?) was one byte shorter in one of the signatures.
268 2013-12-10 07:14:03 <benten> gmaxwell:  could you go 100 blocks backwards, reproduce all the work and then push them all at once as a new block chain if you could outpace the current network?
269 2013-12-10 07:15:13 <benten> In short I'm just asking if there is a limit on the amount of work you could do in the past in regards to how reversing blocks works.
270 2013-12-10 07:15:28 <Plarkplark_> gmaxwell: getting hundreds of thse now
271 2013-12-10 07:15:29 <Plarkplark_> ProcessBlock: ORPHAN BLOCK, prev=000000000000043c19be2e9c55f83dcdb3fd966a70df2dc34785a71e54792cc8
272 2013-12-10 07:15:43 <benten> hm
273 2013-12-10 07:15:54 <gmaxwell> benten: absolutely, if you could enormously outpace the network.
274 2013-12-10 07:16:17 <gmaxwell> benten: any prohibition on that creates network destroying forever inconsistency risk.
275 2013-12-10 07:17:08 <benten> gmaxwell: thanks, that answers my curiosity about transaction reversale
276 2013-12-10 07:18:22 <Plarkplark_> gmaxwell: up to a few thousand of those ORPHAN blocks now.
277 2013-12-10 07:19:36 <benten> gmax, thats why you need about 25 confirmations before a transaction is considered to be 'fairly secure' from being reversed, right?
278 2013-12-10 07:20:59 <gmaxwell> it depends on the value in question... most people are looking for 6 confirmst most of the time.
279 2013-12-10 07:21:22 <gmaxwell> there is a formula in the bitcoin whitepaper, where you can plug in attacker hashpower and get back successrates.
280 2013-12-10 07:23:00 <Plarkplark_> benten, it's a balance of hashing power, value, trust
281 2013-12-10 07:23:17 <Plarkplark_> if you make million dollars worth transaction, wait 6 or maybe 10-ish confirms
282 2013-12-10 07:23:36 <Plarkplark_> if you pay 100 bucks, I would wait 1 confirm if I did not trust the person
283 2013-12-10 07:23:47 <Plarkplark_> if you pay for beer, 0 confirm would be considered "fine".
284 2013-12-10 07:25:51 <gmaxwell> Plarkplark_: 1 confirm is not terribly secure, hashpowerless attackers can reverse 1 confirm with fair chance just by multiannouncing and hoping for a natural fork.
285 2013-12-10 07:27:35 <benten> this is weird. ive been mining on 50btc for the last month and their pool fees appear to be 100%
286 2013-12-10 07:28:12 <phantomcircuit> benten, go to sleep fool
287 2013-12-10 07:28:27 <benten> ok…. :)
288 2013-12-10 07:29:12 <Plarkplark_> yeah but for 100 bucks?
289 2013-12-10 07:30:09 <phantomcircuit> lol
290 2013-12-10 07:30:20 <gmaxwell> benten: https://en.bitcoin.it/wiki/Comparison_of_mining_pools < as you can see, their fees are indeed 100%
291 2013-12-10 07:30:20 <grau> gmaxwell: NO the hashes are identical, you just printed in reverse order!
292 2013-12-10 07:33:42 <grau> gmaxwell: this must be some difference in the signature itself
293 2013-12-10 07:35:10 <Plarkplark_> gmaxwell: sigh. Im just going to redown the blockchain again
294 2013-12-10 07:35:15 <gmaxwell> grau: hm ... any chance you're just truncating a byte off it or something.
295 2013-12-10 07:35:26 <gmaxwell> Plarkplark_: it won't help you.
296 2013-12-10 07:35:33 <gmaxwell> Plarkplark_: there is something wrong with your setup.
297 2013-12-10 07:35:40 <gmaxwell> Plarkplark_: where did your bitcoin come from?
298 2013-12-10 07:35:59 <Plarkplark_> http://sourceforge.net/projects/bitcoin/files/Bitcoin/
299 2013-12-10 07:36:42 <gmaxwell> Plarkplark_: are you running the binary from there or did you compile it?
300 2013-12-10 07:36:52 <Plarkplark_> binary. copied to /usr/bin
301 2013-12-10 07:37:31 <gmaxwell> Were you able to go through the debug.log and find the first Invalid instance?  if you've since tried reindexing it's probably been blown out of the log.
302 2013-12-10 07:38:38 <Plarkplark_> Yes. Therefore i'm trying again. Will p ost back with results.
303 2013-12-10 07:38:59 <Plarkplark_> I have a lot of CPU power and memory at my disposal wanted to help btc network by running a few big nodes.
304 2013-12-10 07:40:01 <Plarkplark_> downloading the bootstrap.dat this time. -loadblock=./bootstrap.dat ?
305 2013-12-10 07:40:39 <gmaxwell> no. just put it in the directory, you don't need to loadblock it
306 2013-12-10 07:41:02 <Plarkplark_> it auto detects it? put in .bitcoin/ directly?
307 2013-12-10 07:41:14 <gmaxwell> though its likely you have hardware problems, I wanted to see the initial error to get an idea of memory or disk / disk bus were more likely the problem.
308 2013-12-10 07:41:18 <gmaxwell> Plarkplark_: yes.
309 2013-12-10 07:41:33 <Plarkplark_> i think it bugged out when the mem was only 512mb ;)
310 2013-12-10 07:41:38 <Plarkplark_> 4GB now
311 2013-12-10 07:44:25 <grau> gmaxwell: This drives me mad. You said one of the sigs would be shorter. I do not see that. It looks like that on brainwallet.org but it is only because the proportional font it uses.
312 2013-12-10 07:45:01 <gmaxwell> you can see it directly in the serialization
313 2013-12-10 07:45:03 <gmaxwell> 48 vs 47
314 2013-12-10 07:45:12 <gmaxwell> (hex)
315 2013-12-10 07:46:21 <grau> you are right the tx serialization is longer.
316 2013-12-10 08:42:16 <Krellan> Hmm - fresh compile of 0.8.6 failed for me, the old boost chrono unresolved reference problem
317 2013-12-10 08:42:21 <Krellan> but 0.8.5 and latest Git worked.
318 2013-12-10 08:51:44 <grau> gmaxwell: actually one of the satoshi signatures is shorter. BOP has two of same size.
319 2013-12-10 08:52:19 <grau> gmaxwell: I reparsed both transactions manually and evaluated both sets of signatures. They are all valid on the same hash.
320 2013-12-10 08:53:29 <grau> curramba: I am tired mixed up.
321 2013-12-10 08:53:42 <grau> gmaxwell: hold on I have to recheck what I did.
322 2013-12-10 08:58:05 <Krellan> Got it, was able to fix the boost chrono problem.
323 2013-12-10 08:59:02 <Krellan> Needed to add to makefile.unix, under the LIBS += block, -l boost_chrono$(BOOST_LIB_SUFFIX) and -l rt
324 2013-12-10 08:59:10 <Krellan> That worked.
325 2013-12-10 09:04:16 <Krellan> Same for the bitcoin-qt: had to add -lboost_chrono -lrt to the end of the big LIBS line in Makefile.  That worked.  Also did qmake BDB_INCLUDE_PATH=/usr/include/db4.8 otherwise it didn't find it.
326 2013-12-10 09:10:02 <Stephen> I'm looking to do some work similar in tech to what http://www.prrofofextistence.com. Anyone have a small task I could complete for them to earn 0.00005 BTC?
327 2013-12-10 09:11:33 <Stephen> Or, if you can spare more, I'll happily develop something comparable to the fee
328 2013-12-10 09:16:20 <Stephen> Note: 0.00005 BTC = $0.05 USD
329 2013-12-10 09:17:08 <wumpus> Krellan: yes, the crappyness of the build system is well-known, that's why 0.9.x switched to autotools which should auto-detect more
330 2013-12-10 09:17:27 <Stephen> Or, if anyone can direct me to something similar to Amazon Turk which pays in BTC, I'll gladly work it through
331 2013-12-10 09:20:39 <grau> gmaxwell: I rechecked. My previous parse is correct. One of the satoshi signatures is shorter.
332 2013-12-10 09:23:43 <grau> sipa: is there a guide of what canonical checks are in place? I am fighting with signature not accepted by satoshi.
333 2013-12-10 09:34:59 <deanclkclk>  1 BTC stored as 100,000,000 (one-hundred-million of the smallest possible bitcoin unit)
334 2013-12-10 09:35:10 <deanclkclk> folks question....I reading the wiki on bitcoin money handling and i"m not sure of this:
335 2013-12-10 09:35:21 <deanclkclk> " 1 BTC stored as 100,000,000 (one-hundred-million of the smallest possible bitcoin unit)"
336 2013-12-10 09:35:27 <deanclkclk> can someone explain?
337 2013-12-10 09:40:59 <shesek> the smallest unit of bitcoins is 0.00000001. there are 100,000,000 of those in 1 BTC
338 2013-12-10 09:41:25 <sipa> Blaster: use getreceivedbyaddress when walletnotify fires
339 2013-12-10 09:50:24 <sipa> grau: i mailed it to the mailinglist once, and there is a reference in the source code
340 2013-12-10 09:50:27 <sipa> let me find it
341 2013-12-10 09:50:50 <grau> could it be that checks are stricter for P2SH?
342 2013-12-10 09:51:10 <grau> I do not seem to run into any problem or difference but only for P2SH redemption.
343 2013-12-10 09:51:40 <Plarkplark_> gmaxwell: using the bootstrap did the trick. I think the blockchain got corrupted deu to low memory at initial sync over p2p
344 2013-12-10 09:51:42 <sipa> https://github.com/bitcoin/bitcoin/blob/master/src/script.cpp#L256
345 2013-12-10 09:52:41 <grau> I see, but strange is that my signatures were always accepted until I tried to use P2SH
346 2013-12-10 09:53:06 <grau> I wen through with gmawell that digest calculation for my signatures is correct
347 2013-12-10 09:53:15 <grau> also in P2SH case.
348 2013-12-10 09:55:58 <ThomasZ> Is it possible to make my bitcoin-qt flush old blocks? To lower the memory footprint.
349 2013-12-10 09:56:09 <ThomasZ> or, actually, on-disk footprint
350 2013-12-10 09:57:23 <ThomasZ> Stephen: want to earn money? Come to my place and do some vacuum cleaning ;)
351 2013-12-10 09:58:08 <Stephen> ThomasZ: LOL, If I could earn a few fractions of BTC in the DFW area vacuuming, I'd do so.
352 2013-12-10 09:58:26 <ThomasZ> :D
353 2013-12-10 09:59:26 <v3ry3l33te> ThomasZ, impossible as of now
354 2013-12-10 09:59:54 <ThomasZ> v3ry3l33te: ok, thanks for the answer.
355 2013-12-10 10:06:02 <deanclkclk> so folks...to run the bitcoin client as a server...I download the bitcoin-qt client?
356 2013-12-10 10:07:28 <e-dard> deanclkclk: If you mean run it on a server as a daemon, there is a headless version
357 2013-12-10 10:07:42 <deanclkclk> what's the url?
358 2013-12-10 10:07:47 <deanclkclk> yes e-dard
359 2013-12-10 10:08:16 <e-dard> It's on github with the QT client I believe
360 2013-12-10 10:11:52 <mrkent> does paytxfee=0.00 mean no fees paid on all transactions?
361 2013-12-10 10:30:23 <michagogo> cloud|mrkent: it means no fees paid on transactions that meet the rules for free transactions
362 2013-12-10 10:30:41 <michagogo> cloud|See ,,(bc,wiki transaction fees)
363 2013-12-10 10:30:42 <gribble> https://en.bitcoin.it/wiki/Transaction_fees | Nov 19, 2013 ... Transaction fees may be included with any transfer of bitcoins from one address to another. At the moment, many transactions are typically ...
364 2013-12-10 10:33:15 <mrkent> michagogo|cloud, okay thanks
365 2013-12-10 10:37:32 <Plarkplark_> Hello
366 2013-12-10 10:37:49 <Plarkplark_> What are the bandwith stats for a full bitcoin node? Are we talking 10's 100s or 1000s of GB per month?
367 2013-12-10 10:39:44 <Stephen> Wow, the sheer bombardment...
368 2013-12-10 10:40:00 <Plarkplark_> ?
369 2013-12-10 10:41:21 <Stephen> Sorry, offhand comment
370 2013-12-10 10:42:03 <ThomasZ> Plarkplark_: the full database of all blocks since the beginning of time is 15Gb.
371 2013-12-10 10:42:20 <ThomasZ> Plarkplark_: you can extrapolate from there ;)
372 2013-12-10 10:42:56 <Plarkplark_> ThomasZ: Yes, but once I have a node running with tweaked max_connections (to help network) and having 1Gb uplink to the internet, what can I expect?
373 2013-12-10 10:43:20 <Plarkplark_> 100 or 1000GB is not a problem (a month) but multiple TB a month would make this costly.
374 2013-12-10 10:43:20 <ThomasZ> Plarkplark_: that it runs just fine ?
375 2013-12-10 10:44:03 <ThomasZ> you only download the full blocks once in your lifetime. So I doubt you'll have to worry
376 2013-12-10 10:44:40 <Plarkplark_> yes, but the node would be uploading a lot..
377 2013-12-10 10:46:17 <ThomasZ> it would? Seems you know already :)
378 2013-12-10 10:48:41 <tablecoder> let's say i create a bali transaction (offline without sending it, such as like this: http://brainwallet.org/#tx)
379 2013-12-10 10:48:58 <grau> sipa: I implemented and went though all isCanonical checks and the BOP signatures are ok with them. I just can not figure why satoshi rejects the transaction.
380 2013-12-10 10:49:48 <tablecoder> how do i a) get the transaction ID from it, and b) can i create another offline transaction using the outputs from my previous offline transaction as inputs to a new transaction, and c) if i send the first transaction, then the second immediately after, will both transactions successfully go through?
381 2013-12-10 10:50:09 <grau> sipa: These transactions are needed for the BTC1K party ticket sale...
382 2013-12-10 10:50:44 <Plarkplark_> ThomasZ: Im asking the question :( - how much uploading does a node do?
383 2013-12-10 10:51:52 <grau> Any chance someone could figure why the transaction http://pastebin.com/ANeJEtS7 is rejected by satoshi?
384 2013-12-10 10:52:52 <grau> I wen through checking inputs digests signed canonical signature. Everything I could come up with. Signatures appear to be valid with Bouncy castle but somehow rejected by satoshi client.
385 2013-12-10 10:54:04 <grau> The BTC 1000$ ticket sale is pending on resolution of this issue. Please help.
386 2013-12-10 10:54:07 <ThomasZ> Plarkplark_: As said, I don't know. I expect it to be trivial amounts compared to your limits.
387 2013-12-10 10:55:32 <Plarkplark_> ok thanks
388 2013-12-10 10:55:56 <Plarkplark_> One more question: Opening ICMP, 8333 is enough?
389 2013-12-10 11:00:16 <ThomasZ> Plarkplark_: Not ICMP, just UDP.  Other than that, yes.
390 2013-12-10 11:00:41 <ThomasZ> Plarkplark_: actually its 8882 here...
391 2013-12-10 11:03:17 <Plarkplark_> ok. And udp what ports?
392 2013-12-10 11:12:19 <dansmith_btc>  HI, I'm struggling to setup gitian following https://github.com/bitcoin/bitcoin/tree/master/contrib/gitian-descriptors. I'm on Ubuntu 12.10 and I get weird errors when running "bin/make-base-vm --arch amd64". Could someone confirm that gitian instructions worked for him?
393 2013-12-10 11:15:00 <dansmith_btc> Or alternatively, I should probably launch a virtualbox 12.04 stock iso ubuntu and setup gitian that way.
394 2013-12-10 11:19:26 <stonecoldpat> grau: Are you signing a simplified version of the transaction?
395 2013-12-10 11:19:51 <grau> stonecoldpat: I do not know what you mean with simplified
396 2013-12-10 11:20:17 <grau> I am signing the same digest satoshi does.
397 2013-12-10 11:20:52 <stonecoldpat> they talk about it here https://bitcointalk.org/index.php?topic=244314.0
398 2013-12-10 11:20:53 <grau> I have a version of the same transaction signed by satoshi and can validate both set of signatures with the same transaction digest
399 2013-12-10 11:20:59 <stonecoldpat> but if its the same digest then it should be ok
400 2013-12-10 11:21:22 <stonecoldpat> well actually, what do you mean by digest
401 2013-12-10 11:21:24 <michagogo> cloud|dansmith_btc: I assume you cloned github:devrandom/gitian-builder?
402 2013-12-10 11:21:51 <dansmith_btc> michagogo|cloud, I did acc. to bitcoin github page
403 2013-12-10 11:21:52 <grau> The signatures are a function of a transaction digest (hash)
404 2013-12-10 11:22:09 <michagogo> cloud|dansmith_btc: follow the readme of that repo
405 2013-12-10 11:22:12 <lianj> grau: you mean the signature hash?
406 2013-12-10 11:22:17 <grau> It is not the transaction hash in common sense as transaction is modified before signature
407 2013-12-10 11:22:27 <grau> yes I mean signature hash
408 2013-12-10 11:23:02 <lianj> ok and you checked that you generate the same signature hash as bitcoind does for you testing/example tx?
409 2013-12-10 11:23:05 <michagogo> cloud|I've done it in precise and raring -- haven't tried it in quantal, but I may be able to help out
410 2013-12-10 11:23:57 <dansmith_btc> michagogo|cloud, I will setup a pristine 12.04  vbox VM and then try afresh
411 2013-12-10 11:23:57 <grau> Here the two sets of signatures: http://pastebin.com/inr9d3G1
412 2013-12-10 11:24:17 <lianj> for the resulting signature, did you put a signature hash_type on it when adding it as input finally
413 2013-12-10 11:24:25 <grau> sure
414 2013-12-10 11:24:38 <michagogo> cloud|dansmith_btc: note that if you're running in a VM, you won't be able to use qemu-kvm
415 2013-12-10 11:24:47 <michagogo> cloud|You need to use lxc
416 2013-12-10 11:25:06 <grau> wait...
417 2013-12-10 11:25:27 <grau> the signature hashType is added before computing the signature hash
418 2013-12-10 11:25:36 <grau> that is what you mean?
419 2013-12-10 11:26:21 <michagogo> cloud|dansmith_btc: what machine are you trying to run gitian on with quantal?
420 2013-12-10 11:26:24 <lianj> no afterwards, it doesn't matter because the signaturehash doesn't/can't include your input script (with your signrature and that hash_type) aynway
421 2013-12-10 11:26:55 <michagogo> cloud|And does the CPU have hardware virtualization support?
422 2013-12-10 11:27:11 <dansmith_btc> michagogo|cloud, yes VT support is there
423 2013-12-10 11:27:38 <michagogo> cloud|What error(s) are you getting?
424 2013-12-10 11:28:09 <michagogo> cloud|;;google --lucky inurl:devrandom gitian site:github.com
425 2013-12-10 11:28:10 <gribble> https://github.com/devrandom/gitian-builder
426 2013-12-10 11:29:09 <grau> yes, hasType is also appended to signature and it is SIGHASH_ALL in all cases
427 2013-12-10 11:30:42 <dansmith_btc> michagogo|cloud, the errorlog: http://pastie.org/8541793
428 2013-12-10 11:30:52 <michagogo> cloud|dansmith_btc: what errors are you getting? (And, I assume you installed all the prereqs from https://github.com/devrandom/gitian-builder/blob/master/README.md?)
429 2013-12-10 11:32:16 <michagogo> cloud|Hmm
430 2013-12-10 11:32:18 <lianj> grau: you are saying http://pastebin.com/inr9d3G1 the --BOP parts are invalid?
431 2013-12-10 11:32:29 <dansmith_btc> michagogo|cloud, yes dependencies are exact same ones in both readmes
432 2013-12-10 11:32:46 <michagogo> cloud|dansmith_btc: --suite quantal is one thing I've never seen used/tried
433 2013-12-10 11:32:57 <grau> lianj: yes they are rejected by satoshi
434 2013-12-10 11:33:24 <grau> they are from the transaction  http://pastebin.com/ANeJEtS7
435 2013-12-10 11:33:40 <michagogo> cloud|dansmith_btc: try it with lucid or precise
436 2013-12-10 11:33:51 <dansmith_btc> michagogo|cloud, I can rerun wit lucid but it will take a while
437 2013-12-10 11:34:26 <lianj> grau: you have a different error though. the signatures look fine: http://paste.mhanne.net/raw/f9c7dd533b3db9df50b18c1bb3e9436ebce5b3b5
438 2013-12-10 11:34:53 <grau> That is the puzzle
439 2013-12-10 11:35:11 <grau> If you try to send the raw transaction to satoshi it rejects it
440 2013-12-10 11:35:17 <lianj> yes but the signatures aren't wrong. something else is
441 2013-12-10 11:35:21 <grau> although here is the transaction parse:
442 2013-12-10 11:35:35 <lianj> so  http://pastebin.com/ANeJEtS7 is what you want to send and generated on your own?
443 2013-12-10 11:36:16 <michagogo> cloud|dansmith_btc: what are you looking to gbuild?
444 2013-12-10 11:36:24 <grau> yes
445 2013-12-10 11:36:35 <grau> This is the same manually reparsed http://pastebin.com/EDhKkpg0
446 2013-12-10 11:37:31 <lianj> that looks like a redeeming p2sh tx
447 2013-12-10 11:37:36 <grau> exactluy
448 2013-12-10 11:37:57 <grau> I never had a problem with my transaction sign/verify until I tried to send P2SH
449 2013-12-10 11:38:24 <grau> But since I sign the same hash as satoshi and sigs are valid I am puzzled
450 2013-12-10 11:39:10 <lianj> i doubt that you sign the same hash. the p2sh case is special for signature hash
451 2013-12-10 11:39:34 <grau> ERROR: CScriptCheck() : 3d7b6899c8519535f5418cb82e7caed661b613187912af18a98d996a3f4ef680 VerifySignature failed
452 2013-12-10 11:39:40 <grau> I do sign the same hash
453 2013-12-10 11:40:22 <michagogo> cloud|dansmith_btc: what do you plan to build in gitian?
454 2013-12-10 11:41:17 <grau> I know how p2sh is special and I am able to verify them if created by satoshi. The verification fails if I forge sigs
455 2013-12-10 11:41:20 <dansmith_btc> michagogo|cloud, bitcoin first, then my own code
456 2013-12-10 11:42:02 <grau> Please if you can debug satoshi
457 2013-12-10 11:42:52 <grau> run http://pastebin.com/H63Wd6fG
458 2013-12-10 11:43:02 <grau> and tell me why exactly it rejects
459 2013-12-10 11:44:23 <grau> you can confirm by the way that the correct hash to sign is e15ad6cd253cc0d55ad18ce8d1992d4f5cd9e677a1854eaa0de6f31e8f6b6411 (not that this is straight not reverse order as for blocks)
460 2013-12-10 11:46:50 <lianj> yes e15ad6cd253cc0d55ad18ce8d1992d4f5cd9e677a1854eaa0de6f31e8f6b6411
461 2013-12-10 11:47:03 <lianj> but the tx looks ok from my viewpoint.
462 2013-12-10 11:47:34 <grau> everithing looks ok except it is not accepted :(
463 2013-12-10 11:47:50 <michagogo> cloud|dansmith_btc: For current bitcoin (0.8.x and earlier) you do need lucid VMs
464 2013-12-10 11:48:05 <michagogo> cloud|One each i386 and amd64
465 2013-12-10 11:48:06 <lianj> grau: which bitcoind and which error message?
466 2013-12-10 11:48:55 <dansmith_btc> michagogo|cloud, why 2 architectures, I'm on 64, do i still need a 32 bit vm ?
467 2013-12-10 11:49:07 <michagogo> cloud|For bitcoin, yes
468 2013-12-10 11:49:12 <grau> Bitcoin version v0.8.5-beta (2013-09-12 13:35:18 +1000)
469 2013-12-10 11:49:29 <michagogo> cloud|The Linux build is actually 2 builds, one for each architecture
470 2013-12-10 11:49:43 <michagogo> cloud|I think current git head (and therefore 0.9) might have upgraded to precise, by the way
471 2013-12-10 11:50:00 <grau> ERROR: CScriptCheck() : 3d7b6899c8519535f5418cb82e7caed661b613187912af18a98d996a3f4ef680 VerifySignature failed
472 2013-12-10 11:50:03 <grau> is in the log
473 2013-12-10 11:50:12 <michagogo> cloud|(I'd check and tell you for sure, but I'm on my phone atm)
474 2013-12-10 11:50:14 <grau> this is an other similar tx
475 2013-12-10 11:51:26 <lianj> ERROR: CScriptCheck() : c36efe11c9f2d854c1111f01fc4ba644935e4dbc7fec7748921e64d0fdc18d73 VerifySignature failed
476 2013-12-10 11:51:30 <lianj> yes hm
477 2013-12-10 11:51:41 <grau> thanks for confirming
478 2013-12-10 11:52:25 <grau> It is really a show stopper for me. I plan to start ticket sales for the BTC1K party in Frankfurt with a P2SH vault and this breaks the schedule....
479 2013-12-10 11:53:25 <grau> I should have started yesterday actually ... https://bitcointalk.org/index.php?topic=285771.0;topicseen
480 2013-12-10 11:54:17 <lianj> grau: pm'd you
481 2013-12-10 11:54:57 <grau> lianj: where?
482 2013-12-10 11:55:51 <lianj> dunno, somewhere in your client another window/tab should have pop up or blinking
483 2013-12-10 11:56:15 <grau> nope.
484 2013-12-10 11:56:48 <lianj> then pm me
485 2013-12-10 11:57:34 <grau> pm'd you
486 2013-12-10 12:02:15 <Blaster> When I run bitcoin-qt from the terminal on Ubuntu and then go do something else, it gets killed
487 2013-12-10 12:02:29 <Blaster> $ bitcoin-qtKilled
488 2013-12-10 12:02:47 <Blaster> what the heck!
489 2013-12-10 12:16:50 <michagogo> cloud|dansmith_btc: wb
490 2013-12-10 12:17:09 <michagogo> cloud|dansmith_btc: Just checked, and git head does indeed use precise
491 2013-12-10 12:19:39 <Blaster> it bitcoin runs out of ram does it just crash?
492 2013-12-10 12:19:40 <Blaster> if*
493 2013-12-10 12:34:33 <dansmith_btc> 
494 2013-12-10 12:34:33 <dansmith_btc> ./bin/gbuild:21:in `system!': failed to run on-target true (RuntimeError)
495 2013-12-10 12:34:33 <dansmith_btc> michagogo|cloud, I'm getting ssh: connect to host localhost port 2223: Connection refused
496 2013-12-10 12:34:41 <dansmith_btc> do you know what is wrong?
497 2013-12-10 12:37:52 <michagogo> cloud|Doing what?
498 2013-12-10 12:42:39 <dansmith_btc> michagogo|cloud, running, ./bin/gbuild --commit bitcoin=HEAD ../bitcoin/contrib/gitian-descriptors/gitian.yml
499 2013-12-10 12:43:23 <michagogo> cloud|Can you pastebin the terminal scrollback?
500 2013-12-10 12:43:33 <michagogo> cloud|And perhaps a ls -a?
501 2013-12-10 12:44:32 <dansmith_btc> michagogo|cloud, http://pastie.org/8541964 and http://pastie.org/8541965
502 2013-12-10 12:45:10 <michagogo> cloud|What do you have checked out in ../bitcoin?
503 2013-12-10 12:45:12 <michagogo> cloud|master?
504 2013-12-10 12:45:57 <michagogo> cloud|Go to ../bitcoin, `git checkout v0.8.6`
505 2013-12-10 12:45:58 <dansmith_btc> michagogo|cloud, yes git branch says master
506 2013-12-10 12:46:04 <michagogo> cloud|then go back to gitian-builder
507 2013-12-10 12:46:21 <michagogo> cloud|`./bin/gbuild --commit bitcoin=v0.8.6 ../bitcoin/contrib/gitian-descriptors/gitian.yml`
508 2013-12-10 12:46:57 <michagogo> cloud|Building current git HEAD is done in a precise VM
509 2013-12-10 12:47:01 <michagogo> cloud|not lucid
510 2013-12-10 12:48:09 <dansmith_btc> michagogo|cloud, running as you suggested still gives ssh: connect to host localhost port 2223: Connection refused
511 2013-12-10 12:51:50 <michagogo> cloud|Hmm
512 2013-12-10 12:52:10 <michagogo> cloud|I'm not that familiar with gitian's internals
513 2013-12-10 12:52:13 <michagogo> cloud|;;seen devrandom
514 2013-12-10 12:52:14 <gribble> devrandom was last seen in #bitcoin-dev 5 days, 14 hours, 21 minutes, and 50 seconds ago: <devrandom> at least for now
515 2013-12-10 12:52:45 <michagogo> cloud|Hmm, whois says 23 hours idle
516 2013-12-10 12:53:21 <michagogo> cloud|dansmith_btc: I'd join #bitcoin-build, mention what you're doing and what's failing over there
517 2013-12-10 12:53:46 <michagogo> cloud|And wait a while -- eventually someone will see and help you out
518 2013-12-10 12:54:06 <michagogo> cloud|In the meantime, you could try in raring or precise
519 2013-12-10 13:05:16 <MoALTz> i have at least 31 orphan blocks in my blk*n.dat files. tempted to write a script to zap them (backing up the original blk files first of course)
520 2013-12-10 13:07:08 <Blaster> I copied the blockchain from one computer to another.  Then I ran bitcoin-qt -rescan, and it's been doing that for about 30 mins.  How long should it take?
521 2013-12-10 13:11:45 <michagogo> cloud|Blaster: Did you include blocks/index/ and chainstate/?
522 2013-12-10 13:12:10 <Blaster> only blocks/index, not chainstate
523 2013-12-10 13:12:25 <Blaster> should I copy that over and restart the rescanning? or can it work through it?
524 2013-12-10 13:12:26 <michagogo> cloud|So it's more than a rescan
525 2013-12-10 13:12:29 <michagogo> cloud|It's a reindex
526 2013-12-10 13:12:49 <Blaster> oh
527 2013-12-10 13:13:17 <Blaster> will killing it and copying chainstate make it quicker?
528 2013-12-10 13:13:42 <michagogo> cloud|probably
529 2013-12-10 13:13:58 <michagogo> cloud|(but copy over blocks/ again as well)
530 2013-12-10 13:18:57 <Blaster> do I need bitcoind for notifywallet?
531 2013-12-10 13:19:46 <Blaster> walletnotify I mean
532 2013-12-10 13:47:05 <wumpus> Blaster: in contrast to what? -walletnotify should work with the GUI as well AFAIK (never tested though)
533 2013-12-10 14:18:06 <gribble> Error: The command "info" is available in the Factoids, GPG, MessageParser, and RSS plugins.  Please specify the plugin whose command you wish to call by using its name as a command before "info".
534 2013-12-10 14:18:06 <Plarkplark_> ;;info
535 2013-12-10 14:18:08 <Plarkplark_> ;;blocks
536 2013-12-10 14:18:09 <gribble> 274177
537 2013-12-10 14:28:24 <Plarkplark_> ;;blocks
538 2013-12-10 14:28:25 <gribble> 274177
539 2013-12-10 14:36:43 <Plarkplark_> http://miki.it/articles/papers/#bitiodine