1 2013-06-11 01:55:35 <warren> hmm, git blame is good to know when a line was added, but how do you easily determine when a line was removed? Sorry, I can't remember.
2 2013-06-11 01:57:10 <toffoo> hey anyone here running the bitcoin-qt 0.8.2 mac client?
3 2013-06-11 02:17:59 <etotheipi_> sipa: question about LevelDB... I was under the impression that if I have an iterator, and then change the data in its path (single-threaded), it will read the updated data when it reaches it
4 2013-06-11 02:18:19 <etotheipi_> I assumed that I needed to create a snapshot if I wanted a consistent read-view of the DB
5 2013-06-11 02:18:39 <etotheipi_> but I just read something that suggests otherwise
6 2013-06-11 02:20:23 <etotheipi_> (basically suggesting that iterators do essentially operate on snapshots, perhaps making/maintaining one while the iterator is active)
7 2013-06-11 02:29:11 <nospinzy> hey
8 2013-06-11 02:29:35 <nospinzy> http://blockchain.info/tx/f2ba5bf8523437761000bb691bc0a7356c27312678c19d3b05b5d7a7bebe2368
9 2013-06-11 02:29:35 <nospinzy> what way do you think this person sent this transaction
10 2013-06-11 02:32:21 <gmaxwell> nospinzy: thats not really ontopic for this channel. you might want to try another one.
11 2013-06-11 02:39:39 <warren> https://en.bitcoin.it/wiki/Transaction_fees is that Settings table still up to date?
12 2013-06-11 02:40:28 <sipa> etotheipi_: i've never done concurrent updates while iteratin
13 2013-06-11 02:40:53 <etotheipi_> sipa: okay
14 2013-06-11 02:41:26 <etotheipi_> I had made an assumption that I suddenly realized may not be true
15 2013-06-11 02:41:48 <etotheipi_> I guess I'll just have to play with it
16 2013-06-11 02:59:36 <phantomcircuit> etotheipi_, in general modifying the underlying data should be assumed to invalidate an iterator... even if they claim otherwise ;)
17 2013-06-11 03:44:05 <TheUni> warren: if you didn't find it yet, git log -S
18 2013-06-11 03:44:53 <warren> TheUni: thanks
19 2013-06-11 03:47:06 <TheUni> np
20 2013-06-11 03:47:22 <warren> did you push a fix?
21 2013-06-11 03:47:47 <TheUni> no, realized it didn't work completely
22 2013-06-11 03:48:22 <warren> TheUni: my solution sucks, but works!
23 2013-06-11 03:48:42 <TheUni> heh
24 2013-06-11 04:20:55 <warren> TheUni: good news is with my hack the build is deterministic
25 2013-06-11 04:21:14 <warren> TheUni: please add $MAKEOPTS and commit it under my name?
26 2013-06-11 04:21:24 <warren> TheUni: I can't send you pull requests for reasons I don't understand.
27 2013-06-11 04:21:52 <TheUni> warren: only thing i can come up with is that i'm a fork of the original..
28 2013-06-11 04:22:00 <TheUni> there's nothing different about my repo otherwise
29 2013-06-11 04:22:29 <warren> I pulled your branch into my tree, branched from it, uploaded to my repo and was unable to create a pull request
30 2013-06-11 04:23:32 <TheUni> warren: what's your github user?
31 2013-06-11 04:24:09 <TheUni> nm, it's at the top of my mailbox
32 2013-06-11 04:26:27 <TheUni> warren: yea, that's it, you'd have to clone mine
33 2013-06-11 04:26:41 <TheUni> github sees that it's 'forked from bitcoin', so that's the upstream choice it gives you
34 2013-06-11 04:26:41 <warren> damn
35 2013-06-11 04:26:53 <TheUni> that's a bit.. odd
36 2013-06-11 04:27:03 <warren> TheUni: It gives me 100 other person's bitcoin forks as options to send a pull request to, but not you.
37 2013-06-11 04:27:17 <TheUni> hmm, sec
38 2013-06-11 04:27:53 <TheUni> now?
39 2013-06-11 04:28:00 <warren> errr
40 2013-06-11 04:28:00 <warren> hold
41 2013-06-11 04:30:35 <warren> nope
42 2013-06-11 04:30:59 <warren> "theuni" right?
43 2013-06-11 04:31:03 <TheUni> yea
44 2013-06-11 04:31:10 <warren> it isn't an option for base repo
45 2013-06-11 04:31:22 <warren> forked from bitcoin/bitcoin
46 2013-06-11 04:31:22 <warren> wtogami / bitcoin
47 2013-06-11 04:31:44 <warren> TheUni: can you just add me as a remote and cherry-pick it into your tree?
48 2013-06-11 04:31:50 <TheUni> sure
49 2013-06-11 04:31:58 <warren> I assume you know the syntax
50 2013-06-11 04:32:16 <warren> TheUni: pushed here: https://github.com/wtogami/bitcoin/tree/win32buildfix
51 2013-06-11 04:32:19 <TheUni> yep
52 2013-06-11 04:33:10 <TheUni> warren: i'm only going to c-p some of it though, ok with you? i'm solving the configure problem in another way
53 2013-06-11 04:33:31 <TheUni> but the commit will be yours
54 2013-06-11 04:35:12 <warren> ok
55 2013-06-11 04:35:20 <warren> curious to see how you do it
56 2013-06-11 04:35:35 <TheUni> warren: still not quite sure. i've struck out 3 ways now...
57 2013-06-11 04:36:09 <TheUni> seems the libfaketime devs made it hard to use systematically on purpose
58 2013-06-11 04:36:54 <warren> I narrowly attacked that one thing because it is the only thing stopping it.
59 2013-06-11 04:41:15 <TheUni> warren: seems it still wants to run sometimes after changing that, though?
60 2013-06-11 04:41:39 <warren> TheUni: what wants to run?
61 2013-06-11 04:41:43 <sipa> warren: you cannot do pull requests to two different repositories that are related
62 2013-06-11 04:41:45 <warren> TheUni: it builds just fine every time for me
63 2013-06-11 04:41:54 <sipa> warren: github limitation
64 2013-06-11 04:42:08 <TheUni> warren: it builds, but you don't see rerunning configure?
65 2013-06-11 04:42:08 <warren> sipa: ok, what alternative do I do?
66 2013-06-11 04:42:16 <sipa> warren: send a patch?
67 2013-06-11 04:42:28 <warren> TheUni: looking at log
68 2013-06-11 04:43:00 <warren> TheUni: you mean make runs configure within it
69 2013-06-11 04:43:01 <TheUni> warren: no worries, i've got something that works now. was just curious to see if yours was still rerunning it
70 2013-06-11 04:43:01 <warren> ?
71 2013-06-11 04:43:05 <TheUni> yea
72 2013-06-11 04:44:06 <warren> TheUni: nope, not re-running during "make"
73 2013-06-11 04:44:33 <warren> TheUni: hmm, $MAKEOPTS didn't add any -j value, quite slow build
74 2013-06-11 04:46:24 <TheUni> warren: i'm not sure where else it'd be coming from
75 2013-06-11 05:11:41 <TheUni> warren: pushed
76 2013-06-11 05:45:34 <TheUni> warren: i'm headed to bed, will check backlog to see if you had a chance to give it a try
77 2013-06-11 06:13:02 <Anduckk> if you do sha256(A+B) = C and you know C + A, can you guess B somehow?
78 2013-06-11 06:13:11 <Anduckk> does it lower the security to know B
79 2013-06-11 06:13:23 <tgs3> dave4925 needs a kick in the connection
80 2013-06-11 06:14:11 <sipa> Anduckk: if that is possible, SHA256 would be broken
81 2013-06-11 06:14:38 <tgs3> Anduckk: by trying every possible value of it
82 2013-06-11 06:15:27 <Anduckk> yeh, thanks for answers
83 2013-06-11 06:19:35 <mrkent> what's the best way to check if a X btc has been paid to address Y if the wallet is not on the server?
84 2013-06-11 06:19:54 <Jc_Dev> what is the rough size of the bitcoin test blockchain?
85 2013-06-11 06:21:32 <Anduckk> mrkent: you can check it from blockchain.info or blockexplorer.com
86 2013-06-11 06:21:41 <Anduckk> or with bitcoin-qt but its complicated
87 2013-06-11 06:22:11 <matjeh> Jc_Dev: 188MB here, 84118 blocks
88 2013-06-11 06:22:30 <Anduckk> of course when checking from blockchain.info or blockexplorer.com you have to trust them to show correct info
89 2013-06-11 06:22:40 <Anduckk> the hard way is to check it straight from the real blockchan via bitcoind
90 2013-06-11 06:22:44 <mrkent> anduckk, can't I do it with bitcoind?
91 2013-06-11 06:22:46 <Jc_Dev> matjeh: excellent thanks - and testnet3 genesis block is from Feb 2011 right?
92 2013-06-11 06:22:47 <Anduckk> you can
93 2013-06-11 06:22:51 <sipa> you can't
94 2013-06-11 06:23:03 <sipa> well, not without iterating over all blocks and transactions
95 2013-06-11 06:23:05 <Anduckk> yeah
96 2013-06-11 06:23:07 <matjeh> testnet on blockexplorer.com is out of date
97 2013-06-11 06:23:08 <sipa> which would take ~hours
98 2013-06-11 06:23:09 <Anduckk> but its very complicated
99 2013-06-11 06:23:16 <sipa> it's trivial, but slow
100 2013-06-11 06:23:19 <matjeh> Jc_Dev: yeah
101 2013-06-11 06:23:24 <Jc_Dev> k
102 2013-06-11 06:23:25 <mrkent> How so Anduckk?
103 2013-06-11 06:23:32 <Anduckk> mrkent: what sipa said
104 2013-06-11 06:23:44 <sipa> mrkent: you really just don't want to do that
105 2013-06-11 06:23:56 <Anduckk> ACTION thinks its complicated task to check for a tx with bitcoind or local blockchain
106 2013-06-11 06:24:07 <sipa> once we add support for watch-only addresses it will be easy
107 2013-06-11 06:24:53 <mrkent> hum... How does blockexplorer/blockchain do it?
108 2013-06-11 06:25:07 <sipa> by keeping a huge index
109 2013-06-11 06:25:16 <sipa> of all addresses and all transactions affecting it
110 2013-06-11 06:26:27 <mrkent> i see. Is watch only address high in the queue?
111 2013-06-11 06:27:05 <sipa> it may end up in 0.9
112 2013-06-11 06:27:43 <Anduckk> also might be possible with armory, dunno
113 2013-06-11 06:27:47 <tgs3> https://fbcdn-sphotos-h-a.akamaihd.net/hphotos-ak-frc1/580405_10200937018724134_1262619752_n.jpg
114 2013-06-11 06:28:17 <matjeh> on a related note, how big is an abe/mysql db thesedays? last time i did an import a few months ago it was about 40GB
115 2013-06-11 06:34:45 <Jc_Dev> mrkent: fyi android app BitCare does watch-only wallets, it gets its info from block explorer
116 2013-06-11 06:38:02 <mrkent> blockchain no longer gives free api access?
117 2013-06-11 06:38:54 <SomeoneWeird> mrkent, ?
118 2013-06-11 06:40:07 <mrkent> Does blockchain.info have api for developers to use?
119 2013-06-11 06:40:38 <SomeoneWeird> yes?
120 2013-06-11 06:44:49 <sipa> it's lovely to see how bitcoin allows trust-free management of money, and everyone trusts a single site (which has been down and even plain wrong in the past...)
121 2013-06-11 06:46:25 <petertodd> sipa: incidentally, I was talking to the guy running the #bitcoin-otc tip bot, and turns out he's using blockchain.info and completely misunderstood my query about auditability, kinda scary
122 2013-06-11 06:52:04 <TD> sipa: lol
123 2013-06-11 06:52:10 <mrkent> my purpose is just for testing only
124 2013-06-11 06:52:11 <TD> sipa: yes
125 2013-06-11 06:52:15 <TD> sipa: APIs!!
126 2013-06-11 06:52:25 <TD> sipa: we haz none
127 2013-06-11 06:52:29 <mrkent> sipa, although watch addresses would be nice
128 2013-06-11 06:53:51 <mrkent> any python-wrapped bitcoin managers you recommend?
129 2013-06-11 06:55:32 <petertodd> mrkent: the Armory wallet of all things might be the closest to what you want that you can actually get right now
130 2013-06-11 06:55:55 <petertodd> mrkent: IIRC the contrib/ directory in the repo has some scripts to work with it, although it's really ram intensive
131 2013-06-11 06:57:02 <mrkent> petertodd, you mean for watch addresses?
132 2013-06-11 06:57:22 <petertodd> mrkent: yeah, my understanding is it maintains a UTXO database
133 2013-06-11 06:58:22 <sipa> petertodd: i thought it loaded the entire transaction history
134 2013-06-11 06:58:30 <sipa> (but my info may be very old)
135 2013-06-11 06:58:56 <petertodd> sipa: yeah, and I thought that as it does that it winds up with a UTXO database, but I've never actually looked
136 2013-06-11 06:59:16 <petertodd> Anyway, point is it does have the code for watch-only in there... somewhere.
137 2013-06-11 07:02:31 <mrkent> Any US based devs wanna beta test my service? I can really use feedback. Will add 0.1 btc to your account to buy stuff from amazon.
138 2013-06-11 07:03:16 <TD> http://www.ieee-security.org/TC/SP2013/papers/4977a080.pdf
139 2013-06-11 07:04:28 <petertodd> sipa: Incidentally, still working on a mempool rewrite, and I realized that a simple greedy memoizing implementation of child-pays-for-parent is doable with O(log n) work to add a tx, and O(n) work to process a block - is this well know already?
140 2013-06-11 07:05:00 <sipa> petertodd: that sounds very low
141 2013-06-11 07:05:37 <petertodd> Basically, just store total unconfirmed fees and size for every tx, and as new tx's come in inherit that unconfirmed fees/size from the parent and put the tx into a sorted list of profit.
142 2013-06-11 07:06:01 <petertodd> It's greedy, because it fails on certain topologies, like one tx with a lot of children, but it works well in most cases and is really cheap.
143 2013-06-11 07:06:29 <petertodd> Adding non-greedy incremental updating should be doable too.
144 2013-06-11 07:08:32 <mrkent> TD interesting article
145 2013-06-11 07:10:12 <mrkent> seems kinda impossible to prevent ddos attacks on hidden services
146 2013-06-11 07:10:56 <petertodd> mrkent: it's quite possible if you add micropayments for the resources you consume
147 2013-06-11 07:13:30 <mrkent> and how to track who paid?
148 2013-06-11 07:14:42 <Jc_Dev> i have a public address and i'm trying to construct an ECKey in bitcoinj, but the constructor wants a byte[] - any idea what best way to convert the string to a byte[] is? no idea what encoding, etc...
149 2013-06-11 07:14:55 <petertodd> you don't have too, just pay with a spendable token embedded in each layer of the onion packet
150 2013-06-11 07:15:20 <sipa> Jc_Dev: there's no way to convert an address into a public key
151 2013-06-11 07:15:36 <sipa> (assuming that's what ECKey wants, i don't know BitcoinJ's API)
152 2013-06-11 07:18:52 <Jc_Dev> yeah that makes sense, address is a hash of public key i think? i have an address and i'm trying to get bitcoinj to grab the transactions for it, it seems to want me to make an ECKey, add it to a wallet, then have the wallet watch the blockchain - maybe it's just not possible.
153 2013-06-11 07:20:03 <TD> mrkent: the real interesting result is that anyone can de-anonymize the silk road for about $11k
154 2013-06-11 07:20:14 <TD> i would assume law enforcement would be very interested in that result
155 2013-06-11 07:21:24 <petertodd> if the silk road is smart that result will just uselessly lead law enforcement to virtual servers, not people
156 2013-06-11 07:21:41 <petertodd> upps the game however
157 2013-06-11 07:21:50 <mrkent> TD, can you explain how its de-anonymized?
158 2013-06-11 07:21:56 <TD> read the paper
159 2013-06-11 07:22:34 <TD> basically they manipulate the consensus and then do traffic analysis attacks
160 2013-06-11 07:23:18 <TD> petertodd: if it were useless, they wouldn't use tor. if they could access the silk road servers they could certainly obtain a lot of information about the operators.
161 2013-06-11 07:23:44 <TD> they also measured the popularity of the silk road
162 2013-06-11 07:23:51 <TD> looks like they get around 15,000 users per day
163 2013-06-11 07:24:38 <petertodd> TD: virtual servers can be shutdown trivially, but they don't have to lead to actual operators, tor is what keeps them from being shut down
164 2013-06-11 07:25:37 <TD> probably not all of the users (or even most of them) are using pgp, so i'd imagine you can bust a lot of dealers even if you couldn't find DPR
165 2013-06-11 07:25:43 <petertodd> TD: but as I say, it upps the game into the realm of the sophisticated rather than making a silkroad site something anyone with a bit of tech knowledge can run, which makes it more likely that you'll need employees, and thus one will talk
166 2013-06-11 07:25:51 <mrkent> where did u get $11k number?
167 2013-06-11 07:25:57 <TD> mrkent: the paper! :)
168 2013-06-11 07:26:28 <petertodd> TD: I've spoken to people who use silk road, amazingly PGP is actually very widely used there
169 2013-06-11 07:26:54 <TD> yeah right. i don't believe that for one moment. i'm sure it *looks* like it's widely used. but this is like saying 2 factor auth is widely used on mt gox.
170 2013-06-11 07:26:56 <petertodd> TD: It and buying VPN services are the only times I've ever heard of anyone outside the Bitcoin community using Bitcoins...
171 2013-06-11 07:27:00 <TD> the only way you can really know, is by being the operator
172 2013-06-11 07:27:33 <TD> mrkent: the interesting thing is the conclusion, really. you can just read the last paragraph. they propose some ways to reduce the efficiency of their attacks. but ultimately the current design of Tor hidden services is not able to reliably prevent de-anonymization.
173 2013-06-11 07:27:40 <TD> it would require a significant redesign.
174 2013-06-11 07:28:27 <TD> to me that says a sufficiently technically advanced and well resourced operation could probably find the underlying server of the silk road. the dutch police have hacked hidden services before (directly, not using tor protocol attacks), so they are willing to do that kind of thing
175 2013-06-11 07:28:28 <petertodd> Meh, I don't have much sympathy for some guy buying pot who is too stupid to use PGP - point is the sellers nearly all advertise PGP keys according to what those people have said.
176 2013-06-11 07:28:34 <mrkent> what is a guard node? paper doesn't seem to define it before talking about it
177 2013-06-11 07:28:41 <mrkent> and im not very familiar with tor
178 2013-06-11 07:28:43 <TD> it assumes familiarity with the architecture of hidden services
179 2013-06-11 07:28:44 <petertodd> mrkent: read the Tor documentation
180 2013-06-11 07:32:23 <petertodd> TD: Yeah, the overall design of Tor is scary given traffic analysis.
181 2013-06-11 07:33:06 <TD> i would assume NSA wiretaps simplify traffic analysis a lot, though i doubt they give a shit about drug dealers
182 2013-06-11 07:34:06 <petertodd> Yes, although at least for passive analysis Tor nodes are so bandwidth saturated that it gives defacto resistance to traffic analysis, usually.
183 2013-06-11 07:34:34 <mrkent> hum.. guess just need more nodes
184 2013-06-11 07:34:57 <petertodd> When I was running a Tor node on a fast server it was doing ~10MiB/second symetric with a few hundred connections, 24/7
185 2013-06-11 07:35:05 <petertodd> (relay node)
186 2013-06-11 07:36:24 <mrkent> is that a lot?
187 2013-06-11 07:36:50 <TD> well, it adds up fast if you are paying for it
188 2013-06-11 07:37:11 <petertodd> mrkent: *statisticly* it is for the guy trying to map incoming to outgoing packets
189 2013-06-11 07:38:39 <mrkent> hum
190 2013-06-11 07:38:58 <petertodd> TD: it says something that the VPS company went out of business... I was practically exploiting a loophole in their pricing structure
191 2013-06-11 07:39:05 <TD> oh dear
192 2013-06-11 07:39:43 <petertodd> TD: been a few years, but they were really underpricing a particular set of features
193 2013-06-11 07:39:45 <mrkent> If there is a way of allowing people to annonomously buy nodes, I bet you can make some moeny
194 2013-06-11 07:40:07 <mrkent> Just so silkroad would pay for it
195 2013-06-11 07:40:25 <petertodd> mrkent: silkroad has a lot of listings for VPS servers payable in bitcoins last time I looked