1 2012-09-21 00:37:16 <eennaam> echo
2 2012-09-21 00:37:22 <jgarzik> cardiogram
3 2012-09-21 00:37:25 <eennaam> hello world
4 2012-09-21 00:38:08 <eennaam> can i report some minor annoyance with bitcoin client?
5 2012-09-21 00:44:51 <forrestv> eennaam, sure, why not
6 2012-09-21 00:47:15 <eennaam> everytime a new version is out, i install it, the shortcut to it is also replaced, it links to the executable, but i change it with some attributes behind the executable, i thought maybe in new version could try not to replace the old shortcut if it is already there, because it removes the attributes
7 2012-09-21 00:47:27 <eennaam> but it is just very small issue i guess
8 2012-09-21 01:25:56 <weex> eennaam: it's better to create an issue at github
9 2012-09-21 01:26:31 <eennaam> okay
10 2012-09-21 02:55:57 <randy-waterhouse> hi I'm getting a db error with bitcoin-0.7 (ubuntu 11.04) ... I've run the bitcoind -detachdb prior to launch ... any ideas?
11 2012-09-21 02:56:10 <randy-waterhouse> ************************
12 2012-09-21 04:19:59 <jgarzik> The ZeroAccess Botnet: Mining and Fraud for Massive Financial Gain - http://www.sophos.com/en-us/medialibrary/PDFs/technical%20papers/Sophos_ZeroAccess_Botnet.pdf?dl=true
13 2012-09-21 04:20:22 <jgarzik> botnet is P2P; even has super nodes and leaf nodes
14 2012-09-21 04:25:41 <Gladamas> wow
15 2012-09-21 04:29:48 <Joric> mining is probably the least profitable thing that can be done with a botnet
16 2012-09-21 04:31:29 <Joric> or maybe it's just me and other ppl got better gpus
17 2012-09-21 04:46:05 <wumpus> the botnet explicitly disables gpu mining, so the profit cannot be that large per machine, though ofcourse they have many, many machines
18 2012-09-21 04:58:18 <_dr> using like 30% or whatever of the gpu would seem more covert
19 2012-09-21 04:58:52 <_dr> even uninformed users will notice there's something wrong with their machine if all cpus work at full load
20 2012-09-21 04:59:29 <_dr> that is, if someone actually uses the infected machine and it's not just sitting somewhere
21 2012-09-21 05:00:29 <amiller> their estimate seems to be that the botnet makes 10% of its income from mining
22 2012-09-21 05:03:39 <_dr> i want a botnet
23 2012-09-21 05:27:02 <Diablo-D3> what the fuck
24 2012-09-21 06:15:30 <amiller> "regressive" describes the tendency of a lottery to be played by poor people
25 2012-09-21 06:15:53 <amiller> lotteries with high jackpots tend to be played by more affluent people
26 2012-09-21 06:16:44 <amiller> p2pool is a regressive lottery
27 2012-09-21 06:18:30 <amiller> solomining is progressive
28 2012-09-21 06:33:52 <amiller> lol
29 2012-09-21 06:42:23 <amiller> hah i think i figured out why satoshi dice is a problem
30 2012-09-21 06:43:03 <amiller> it's wasted risk-tolerance
31 2012-09-21 06:43:25 <amiller> the underlying cause is the fixed block difficulty
32 2012-09-21 07:01:07 <yellowhat> amiller can you expand on the satoshidice problem?
33 2012-09-21 07:02:45 <amiller> lets say that bitcoin is like a state lottery in that harnesses risky-greedy behavior and turns it into public utility
34 2012-09-21 07:03:18 <amiller> imagine that a state lottery has one game you can play, it pays out once a week and the jackpot is $1,000,000
35 2012-09-21 07:04:36 <Joric> and all that goes to eric woorhees
36 2012-09-21 07:04:38 <amiller> hopefully that's an interesting game and a lot of people play - the expected earning is negative, but the proceeds go to some useful cause
37 2012-09-21 07:05:37 <amiller> typically there are a lot of different games to choose from
38 2012-09-21 07:07:03 <amiller> an interesting thing about lotteries is that poor people tend to play the games that have smaller jackpots but a higher chance of winning
39 2012-09-21 07:07:24 <amiller> but affluent people tend to play the games with a higher payout
40 2012-09-21 07:07:38 <amiller> http://www.dartmouth.edu/~chance/chance_news/for_chance_news/ChanceNews13.02/OsterShort.pdf
41 2012-09-21 07:08:43 <epscy> affluent people are more greedy?
42 2012-09-21 07:09:43 <yellowhat> because for a rich person it might not make an impact on his life if he wins 10.000 $ for a poor person, it will
43 2012-09-21 07:10:58 <yellowhat> so you point is - from a socialist state perspective it is morally wrong that eric vorhees wins?
44 2012-09-21 07:11:55 <amiller> no - i'm saying it's wrong from a bitcoin scalability perspective
45 2012-09-21 07:12:01 <amiller> it's too bad bitcoin only lets you play at one game
46 2012-09-21 07:12:11 <epscy> err
47 2012-09-21 07:12:15 <epscy> i don't understand
48 2012-09-21 07:12:35 <epscy> you could run lots of different types of lotteries in bitcoin
49 2012-09-21 07:12:37 <amiller> a variation of bitcoin with dynamic block difficulty would be able to pull in more revenue, more mining work
50 2012-09-21 07:12:47 <yellowhat> well i expect more competitors with more interesting UIs to spring up competing with SD
51 2012-09-21 07:12:53 <epscy> oh you are talking about mining
52 2012-09-21 07:13:36 <epscy> but you don't want to it to vary too much
53 2012-09-21 07:13:48 <epscy> because that might make bitcoin insecure
54 2012-09-21 07:14:00 <yellowhat> it took me a while to get that too :)
55 2012-09-21 07:15:37 <amiller> ah i need to pick a phrase besides "dynamic block difficulty" because that makes it seem like i'm talking about varying the difficulty threshold over time
56 2012-09-21 07:15:56 <amiller> what i'm talking about is more like having a bunch of jackpots at different difficulties and you can pick your bet wherever you want
57 2012-09-21 07:17:56 <epscy> interesting idea
58 2012-09-21 07:20:32 <amiller> i need to come up with a proof of work scheme where it's easy to pay for the mining
59 2012-09-21 07:24:39 <amiller> bitcoin can already do part of what's required - you can ask someone to mine on a block for you without revealing what it is
60 2012-09-21 07:25:55 <amiller> the reason it's not safe to outsource mining right now is that the service provider can tell if you've won or not
61 2012-09-21 07:26:22 <amiller> you should be able to hire a miner to produce a high-work lottery ticket for you, but only you can scratch it off to see if you won
62 2012-09-21 07:27:14 <amiller> (tl;dr for the previous half of this rant: the problem with satoshi dice is that it's more exciting than bitcoin ming)
63 2012-09-21 07:34:59 <epscy> heh
64 2012-09-21 07:35:08 <epscy> not sure how this would work though
65 2012-09-21 07:35:30 <epscy> you could limit the lower jackpots to smaller blocks perhaps
66 2012-09-21 07:36:02 <epscy> not sure if there would be enough of an incentive to mine the higher jackpots though
67 2012-09-21 07:38:20 <amiller> well there's a natural incentive which is that you have less chance of being preempted at the higher jackpot
68 2012-09-21 07:38:28 <amiller> of having a stale fork
69 2012-09-21 07:38:50 <amiller> propagation/validation goes proportionally faster at the higher difficulties
70 2012-09-21 07:38:55 <epscy> preempted?
71 2012-09-21 07:39:23 <epscy> you mean less competition?
72 2012-09-21 07:40:51 <amiller> the stale block rate is smaller up there
73 2012-09-21 07:41:20 <amiller> someone who's mining on a very difficult block
74 2012-09-21 07:41:29 <amiller> wants damn sure to have the fastest network and to always be mining at the front
75 2012-09-21 07:41:30 <epscy> hmmm, not sure how much of a factor that is
76 2012-09-21 07:41:53 <epscy> the stale block rate i mean
77 2012-09-21 07:42:04 <amiller> because the risk of winning a block and being beaten because someone else made a bigger one faster is too unbearable
78 2012-09-21 07:46:34 <amiller> it's kind of a two dimensional lottery
79 2012-09-21 07:46:50 <amiller> and there's supply/demand for each
80 2012-09-21 07:49:49 <epscy> yeah, i wonder how that would work with a single blockchain
81 2012-09-21 07:50:09 <epscy> you could argue we already have what you suggest
82 2012-09-21 07:50:29 <epscy> there is litecoin for people who think mining bitcoin is too hard
83 2012-09-21 07:50:39 <epscy> though litecoin is flawed imho
84 2012-09-21 07:52:55 <amiller> divisiveness isn't the answer, the goal is to consolidate all the work
85 2012-09-21 07:53:26 <epscy> lower jackpot, means lower difficulty
86 2012-09-21 07:53:48 <amiller> also both bitcoin and litecoin are flawed in that their proof-of-work puzzles have nothing to do with propagation/validation
87 2012-09-21 07:53:59 <amiller> dhtcoin will kick ass though
88 2012-09-21 07:54:04 <epscy> dhtcoin?
89 2012-09-21 07:54:26 <epscy> i thought that was feature, the network and validation could be abstracted away
90 2012-09-21 07:54:55 <epscy> bitcoin could run over the telephone or snail mail, in theory
91 2012-09-21 07:55:19 <amiller> even if you have telephone and snail mail, you still need to validate transactions
92 2012-09-21 07:55:40 <phantomcircuit> epscy, well no not really since you would be epic vulnerable to a sybil attack
93 2012-09-21 07:56:34 <epscy> phantomcircuit: yeah, in practice there would be all sorts of issues
94 2012-09-21 07:57:42 <epscy> i worry about the security of a blockchain where some blocks were found with a lower difficulty than others
95 2012-09-21 07:58:26 <amiller> epscy, you'd always pick the "strongest" block, the one with the most total work
96 2012-09-21 07:58:40 <amiller> er, strongest chain*
97 2012-09-21 07:59:56 <epscy> sounds like that would make mining for lower jackpots very risky
98 2012-09-21 08:01:05 <amiller> well if you're willing to cooperate with others, you'd want to pool together on something like p2pool, mine at the higher difficulty, and split the rewards
99 2012-09-21 08:01:18 <amiller> you're right - mining for lower jackpots is risky and inefficient
100 2012-09-21 08:02:47 <amiller> high frequency trading is kind of like an advanced form of satoshidice
101 2012-09-21 08:04:23 <phantomcircuit> amiller, except on average you win
102 2012-09-21 08:04:24 <phantomcircuit> :)
103 2012-09-21 08:04:44 <phantomcircuit> also not sure if anybody noticed but intersango is now much faster
104 2012-09-21 08:04:46 <phantomcircuit> stupid PDO
105 2012-09-21 08:06:15 <amiller> phantomcircuit, well, it's easy to win when there's a trapdoor that could be coerced out somehow :/
106 2012-09-21 08:08:02 <epscy> phantomcircuit: PDO?
107 2012-09-21 08:10:00 <phantomcircuit> epscy, php data object
108 2012-09-21 08:10:22 <phantomcircuit> basically instead of using prepared statements just using escaped strings results in substantially better query planning from postgresql
109 2012-09-21 08:10:30 <phantomcircuit> and thus a massive speed up
110 2012-09-21 08:10:48 <Diablo-D3> that sounds like a load of bullshit
111 2012-09-21 08:11:02 <epscy> is that a framework bug or a postgres bug?
112 2012-09-21 08:11:08 <Diablo-D3> has to be a framework bug
113 2012-09-21 08:11:15 <Diablo-D3> prepared statements is how you make sql fast
114 2012-09-21 08:11:34 <phantomcircuit> Diablo-D3, hilariously not in this case
115 2012-09-21 08:11:57 <Diablo-D3> wtf is the query?
116 2012-09-21 08:11:59 <phantomcircuit> i dumped the query plan in the log and it turns out the plan is totally retarded when using prepared statements
117 2012-09-21 08:12:10 <Diablo-D3> that sounds extremely wrong
118 2012-09-21 08:12:12 <phantomcircuit> Diablo-D3, just selecting the orders linked to an account
119 2012-09-21 08:12:26 <Diablo-D3> whos db is this?
120 2012-09-21 08:12:29 <phantomcircuit> i agree
121 2012-09-21 08:12:30 <phantomcircuit> and yet
122 2012-09-21 08:12:43 <phantomcircuit> it's postgres
123 2012-09-21 08:12:55 <epscy> i thought postgres was supposed to be a big piece of awesomeness
124 2012-09-21 08:12:57 <Diablo-D3> because it sounds very much like the table has the wrong indexes set
125 2012-09-21 08:13:00 <Diablo-D3> epscy: it is
126 2012-09-21 08:13:04 <phantomcircuit> a small number of users have a much greater % of orders than other users
127 2012-09-21 08:13:06 <Diablo-D3> its basically the best traditional db I know of
128 2012-09-21 08:13:14 <phantomcircuit> for those users a sequential scan of the orders table makes sense
129 2012-09-21 08:13:18 <phantomcircuit> for everybody else it doesn't
130 2012-09-21 08:13:38 <phantomcircuit> if the planner doesn't know which account you're looking at when it creates the plan it goes with sequential
131 2012-09-21 08:13:44 <phantomcircuit> not sure why exactly but it does
132 2012-09-21 08:13:55 <Diablo-D3> phantomcircuit: erm
133 2012-09-21 08:13:57 <Diablo-D3> thats all tunable
134 2012-09-21 08:14:12 <Diablo-D3> and you can make indexes by user AND force the query to use the index
135 2012-09-21 08:14:15 <phantomcircuit> Diablo-D3, yes except this is the way i was told by grey beards to do it
136 2012-09-21 08:14:16 <phantomcircuit> heh
137 2012-09-21 08:14:29 <phantomcircuit> the query plan is totally different
138 2012-09-21 08:14:33 <epscy> php grey beards or postgres grey beards?
139 2012-09-21 08:14:33 <phantomcircuit> not just indexes
140 2012-09-21 08:14:39 <phantomcircuit> postgres
141 2012-09-21 08:14:42 <epscy> because not all grey beards are equal
142 2012-09-21 08:14:45 <phantomcircuit> there are no php grey beards
143 2012-09-21 08:15:07 <Diablo-D3> this is true
144 2012-09-21 10:42:05 <BlueMattBot> Yippie, build fixed!
145 2012-09-21 10:42:06 <BlueMattBot> Project Bitcoin build #68: FIXED in 6 hr 35 min: http://jenkins.bluematt.me/job/Bitcoin/68/
146 2012-09-21 12:35:03 <gmaxwell> sipa: I confirm IsCanonicalSignature has 100% branch coverage now. Hurray.
147 2012-09-21 12:35:31 <gmaxwell> IsCanonicalPubKey does not yet.
148 2012-09-21 12:35:42 <gmaxwell> (none of the failures are tested)
149 2012-09-21 13:20:22 <diki> Why does the coding style forbid using tabs? I find tabs to be much easier to use.
150 2012-09-21 13:20:28 <diki> and make code cleaner
151 2012-09-21 13:20:44 <helo> uniform appearance regardless of settings
152 2012-09-21 13:21:33 <sipa> it should look the same whether you use tabs or spaces (given the right tab width setting), and in a decent editor both are just as easy to use
153 2012-09-21 14:01:19 <_dr> helo: you can set up vim so it's identical
154 2012-09-21 14:01:27 <helo> yup
155 2012-09-21 14:02:25 <epscy> take the number of spaces you use to indent code, subract two, stab yourself that many times
156 2012-09-21 14:06:49 <helo> sounds like a good argument for using \\t
157 2012-09-21 14:08:08 <helo> depending on one's interpretation of -2 stabs
158 2012-09-21 14:08:21 <_dr> obviously: stab someone else two times :)
159 2012-09-21 14:55:36 <helo> would it be healthy for colored bitcoin to be used by large organizations (municipal governments, for example) to track property ownership?
160 2012-09-21 14:56:29 <helo> the additional fees would help support mining, but the block space competition could make it more expensive to send bitcoin as a currency
161 2012-09-21 14:57:35 <helo> it kind of seems like we could use some block space competition to push fees up, at least to help compensate for the reward halving
162 2012-09-21 14:57:56 <kjj_> I think that there are fundamental issues with coloring that make the idea unhealthy, but I don't think it would be VERY bad for the network
163 2012-09-21 14:58:15 <kjj_> helo: we don't want the fees to be unnaturally high
164 2012-09-21 14:59:29 <helo> right... 0.5btc might not be an unreasonable fee for someone to pay if they're transferring ownership of their vehicle to someone else, but it would be pretty prohibitive if someone wants to send 1btc to someone
165 2012-09-21 15:00:20 <helo> so colored bitcoin may have the potential to significantly harm bitcoin's viability as a currency
166 2012-09-21 15:00:29 <helo> if it were used on a large scale
167 2012-09-21 15:02:10 <kjj_> I don't know how significant the harm would be. and people could argue the other way, that it helps
168 2012-09-21 15:03:00 <helo> massive fees would make for a pretty astronomical hashrate... maybe an excessively high hashrate. there's no reason to have more hashing power than we need to be secure.
169 2012-09-21 15:03:18 <kjj_> the point is that we (as users) want fees to be at their natural level, just barely paying for network security
170 2012-09-21 15:04:18 <kjj_> but if we are competing with non-currency uses of the system, the fees will necessarily be higher with those other uses than they would be otherwise, which means that we (collectively) are paying too much for currency transactions
171 2012-09-21 15:04:23 <helo> a ~major weakness may be that the fees are based on block space scarity, rather than the need for an appropriate hashrate
172 2012-09-21 15:05:15 <copumpkin> http://bitbin.it/4FcNx3sI
173 2012-09-21 15:05:16 <helo> yeah, it seems colored bitcoin may be bad for currency bitcoin...
174 2012-09-21 15:05:17 <kjj_> over time, the fees will get to the level that pays for the real cost of the system. we can change the maximum block size if we need to
175 2012-09-21 15:10:01 <gmaxwell> 09:55 < helo> would it be healthy for colored bitcoin to be used by large organizations (municipal governments, for example) to track property ownership?
176 2012-09-21 15:10:15 <gmaxwell> I don't think that has any value, either to bitcoin or the colored-users.
177 2012-09-21 15:10:29 <helo> yeah, that
178 2012-09-21 15:10:50 <gmaxwell> You're still depending on a centeralized orginization to give the coloring real meaning. So you might as well have them track the ordering of property exchanges.
179 2012-09-21 15:11:31 <helo> right
180 2012-09-21 15:12:38 <helo> changing topic... "my BitCoin gobbler takes 0.1 from every transaction and shifts it off as transaction fee/lost in verification" "we got a worm inside the client and it spreads instantly through the entire network"
181 2012-09-21 15:16:20 <copumpkin> helo: yeah, good times
182 2012-09-21 15:17:43 <helo> claims to be making $1500/week :/
183 2012-09-21 15:19:44 <fiesh> no, per day
184 2012-09-21 15:19:45 <gmaxwell> helo: hah where is this from?
185 2012-09-21 15:19:55 <helo> gmaxwell: copumpkin's link above
186 2012-09-21 15:20:00 <helo> ahh right, per day
187 2012-09-21 15:20:07 <kjj_> gmaxwell: http://pastebin.com/EcNCwaEu
188 2012-09-21 15:20:25 <gmaxwell> ah I stopped by like the 10th line because the stupid was burning my eyes.
189 2012-09-21 15:20:33 <kjj_> from July 2011
190 2012-09-21 15:21:21 <Eliel> ACTION wonders if it'd make any sense to have the client set the minimum fees in a way that attempts to keep the average block reward at around 50BTC.
191 2012-09-21 15:21:47 <kjj_> no
192 2012-09-21 15:22:00 <gmaxwell> Eliel: jesus, no, :P
193 2012-09-21 15:22:36 <kjj_> maybe if the subsidy function was something other than >> that could have worked
194 2012-09-21 15:23:07 <gmaxwell> Eliel: there is a reason bitcoin doesn't work anything like that; whats the proper income for mining? whatever it is, it certantly shouldn't have anything to do with the bitcoin exchange rate.
195 2012-09-21 15:25:37 <fiesh> exchange rate?
196 2012-09-21 15:26:23 <kjj_> he's talking about mining income in non-bitcoin units of value
197 2012-09-21 15:26:54 <Eliel> gmaxwell: income for mining will hover around barely worth doing whatever we do :P
198 2012-09-21 15:27:15 <kjj_> as in, why should mining be worth $500 per block today, but maybe $5000 per block next year?
199 2012-09-21 15:27:24 <fiesh> kjj_: oh I see... but that will inevitably depend on the exchange rate, no?
200 2012-09-21 15:27:51 <Eliel> gmaxwell: the real question is, how high network hashrate is enough :)
201 2012-09-21 15:28:07 <helo> ^
202 2012-09-21 15:28:17 <kjj_> if the mining reward was fixed in BTC, the real-world value would float based on the exchange rate
203 2012-09-21 15:28:23 <gmaxwell> Eliel: No, it won't thats the point.
204 2012-09-21 15:28:46 <kjj_> by not having a fixed BTC reward, the real-world value can match the real-world costs, and the BTC amount can float
205 2012-09-21 15:28:50 <fiesh> kjj_: oh yes, I agree that's not a good idea
206 2012-09-21 15:28:57 <gmaxwell> Eliel: if you fix it like that then income depends on how much distortion it puts on the exchange rate.
207 2012-09-21 15:29:08 <helo> the bitcoin hashrate depends on free block competition, but the sufficient hashrate for security depends on whatever an attacker has access to
208 2012-09-21 15:29:28 <helo> "free block space competition"
209 2012-09-21 15:30:05 <gmaxwell> (feel free to replace exchange rate with "the price of goods in btc"
210 2012-09-21 15:30:06 <gmaxwell> )
211 2012-09-21 15:37:20 <eian> Are there desktop motherboards with TPMs built in?
212 2012-09-21 15:37:29 <eian> If so, where the hell do I buy one
213 2012-09-21 15:38:57 <kjj_> the NSA gift shop has them
214 2012-09-21 15:39:04 <eian> :)
215 2012-09-21 15:39:40 <eian> There's all this hoopla about trusted computing
216 2012-09-21 15:39:49 <eian> I guess it's meant for industry
217 2012-09-21 15:41:33 <kjj_> you mostly see it on server boards
218 2012-09-21 15:43:01 <eian> yeah
219 2012-09-21 15:43:08 <eian> pff
220 2012-09-21 15:43:16 <eian> I wanted to built a trusted cloud in my room
221 2012-09-21 15:43:20 <eian> build*
222 2012-09-21 15:43:26 <kjj_> trusted by who?
223 2012-09-21 15:43:28 <PK> my 6 years old notebook has a tpm chip, iirc.
224 2012-09-21 15:43:30 <eian> me
225 2012-09-21 15:43:33 <eian> lol
226 2012-09-21 15:43:39 <PK> make a cloud out of notebooks.
227 2012-09-21 15:43:41 <eian> I just wanted to figure some stuff out to be honest
228 2012-09-21 15:43:47 <eian> PK, who is the vendor?
229 2012-09-21 15:43:52 <PK> HP
230 2012-09-21 15:43:59 <eian> wow
231 2012-09-21 15:44:25 <PK> but I think pretty much every notebook should have it by now.
232 2012-09-21 15:45:13 <PK> eian, btw, gigabyte uses tpm's too, maybe you find a desktop mainboard there.
233 2012-09-21 15:54:24 <amiller> there's a strategy in lotteries by which you buy a "trump ticket"
234 2012-09-21 15:54:29 <amiller> and purchase every possible combination of ticket
235 2012-09-21 15:54:43 <amiller> 11% of lotteries have positive expected value if you're able to do that http://college.holycross.edu/RePEc/eej/Archive/Volume32/V32N4P673_684.pdf
236 2012-09-21 15:55:13 <amiller> because of the costs involved in doing that, it has never happened
237 2012-09-21 15:55:59 <amiller> an australian 'consortium' once grouped together to try to purchase all the tickets and 'corner' a $25 million jackpot, but they tried and were only able to obtain about 40% of a complete set
238 2012-09-21 16:02:41 <kjj_> anyone know git/github?
239 2012-09-21 16:11:13 <copumpkin> kjj_: pretty sure nobody's ever heard of either
240 2012-09-21 16:12:09 <kjj_> heh. I'm pretty sure plenty have, I'm just not sure if any are paying attention
241 2012-09-21 16:12:37 <jgarzik> amiller: https://github.com/jgarzik/pybond/blob/master/dht.py should be in good shape now
242 2012-09-21 16:13:00 <kjj_> I made a pull request, then updated it based on the comments and re-pushed. now, when I look at the diff, it shows the changes from the previous commit, not from the master. does it look like that for everyone, or just me? and if everyone, how do I fix it?
243 2012-09-21 16:16:38 <jgarzik> kjj_: It sounds like you created a new commit just to address the comments (total commits == 2) rather than 'git reset' or 'git rebase' to revise the original commit (total commits == 1).
244 2012-09-21 16:17:27 <kjj_> yeah, first time using git, and I don't know the terms well enough to use google to steer me to the right method
245 2012-09-21 16:17:37 <jgarzik> kjj_: The former method is what Linus Torvalds recommends and is more git-friendly. The latter method involves 'git push --force', rewriting history and other git unfriendly practices...... that sometimes lead to a cleaner net result.
246 2012-09-21 16:18:09 <jgarzik> It is common in bitcoin to rebase (latter method), as it produces a cleaner diff, and cleaner upstream history.
247 2012-09-21 16:18:29 <kjj_> ok, how do I do it the way that works best for bitcoin?
248 2012-09-21 16:19:10 <jgarzik> kjj_: do what suits your personal style and workflow best
249 2012-09-21 16:19:35 <kjj_> heh. I don't have a style or a workflow. I have a diff
250 2012-09-21 16:19:44 <jgarzik> kjj_: a commit is a logical, atomic change of the code. think how you want your code arranged, if it is presented as a stream of reviewable, successive changes.
251 2012-09-21 16:19:59 <kjj_> how can I get that diff to the rest of you in the way that works best for you?
252 2012-09-21 16:20:28 <kjj_> yeah, I get the stream and the tree models
253 2012-09-21 16:21:03 <jgarzik> kjj_: one important principle is that the code is "bisectable", which means there is never a commit that breaks the build/test workability of the software. Sometimes you wind up creating a bunch of cosmetic, algebraically-equivalent transformations, just to prepare the code for a big change.
254 2012-09-21 16:22:31 <jgarzik> kjj_: when working on projects inside branches of github/kjj/bitcoin.git, for submittal as bitcoin pull requests, you choose between a submission that looks like "1) my change, 2) fixes and updates based on community feedback" or "1) my change"
255 2012-09-21 16:22:42 <jgarzik> neither choice is "right" or "wrong"
256 2012-09-21 16:23:31 <gmaxwell> I think rewriting history is the right thing to do for personal repositories. Keeping the public history as linear as reasonably possible has value, and consolidating diffs makes review much easier.
257 2012-09-21 16:23:43 <jgarzik> I tend to rebase. If there is community feedback, I'll revise the original commit, and push back to github with "git push --force"
258 2012-09-21 16:23:48 <jgarzik> But that is just a personal choice.
259 2012-09-21 16:24:21 <jgarzik> Torvalds argues that if anybody is watching your tree (via 'git pull'), they are royally screwed if you rebase.
260 2012-09-21 16:24:37 <kjj_> so, your pulls are always master -> latest, not master -> old -> latest ?
261 2012-09-21 16:24:54 <amiller> can you rebase into a different branch and then pull-request that
262 2012-09-21 16:24:56 <jgarzik> kjj_: always work on a side branch
263 2012-09-21 16:25:07 <jgarzik> amiller: sure
264 2012-09-21 16:25:22 <kjj_> yeah, I'm in a branch inside my repo. kjj2/stopdetach
265 2012-09-21 16:25:29 <amiller> that might avoid screwing the followers then
266 2012-09-21 16:26:01 <kjj_> I assume I have no followers. how do I rebase so that the pull-req is clean?
267 2012-09-21 16:26:16 <jgarzik> kjj_: I only ever pull bitcoin/bitcoin.git -> local master
268 2012-09-21 16:26:25 <jgarzik> kjj_: then create a branch from that local master
269 2012-09-21 16:26:40 <jgarzik> kjj_: then, after work, push branch to jgarzik/bitcoin.git
270 2012-09-21 16:27:11 <jgarzik> Another nice thing about git is that it is not critically important that you always work on the tip of the tree.
271 2012-09-21 16:27:39 <jgarzik> It only matters insofar as Gavin needs to be able to merge the resultant pull request
272 2012-09-21 16:28:13 <jgarzik> thus, try to avoid the habit of "git checkout mybranch ; git pull github/bitcoin/bitcoin.git master"
273 2012-09-21 16:28:36 <jgarzik> (i.e. fetch latest git HEAD upstream) Do that only when you know you need to.
274 2012-09-21 16:29:08 <kjj_> I hate to say it, but most of this is going way over my head
275 2012-09-21 16:29:38 <kjj_> mostly because my total experience with git has been setting up my repo, making my changes, and making a pull-request
276 2012-09-21 16:30:04 <gavinandresen> kjj_: Run: git rebase -i origin/master
277 2012-09-21 16:30:49 <gavinandresen> kjj_: ... then squash the commits into one (put "f" for fixup in the first column of the interactive rebase edit session)
278 2012-09-21 16:31:04 <gavinandresen> kjj_: ... then git push -f to your repository
279 2012-09-21 16:31:30 <kjj_> fixup is a third command after the two picks?
280 2012-09-21 16:31:46 <gavinandresen> wait... no... rebase -i upstream/master, if you've got an 'upstream' remote that is the bitcoin tree
281 2012-09-21 16:31:46 <kjj_> (not interactive for me)
282 2012-09-21 16:32:42 <gavinandresen> kjj_: rebase -i should open up an editor window with instructions on what you can do
283 2012-09-21 16:34:11 <kjj_> ok, when rebasing, do I fixup one commit, or both?
284 2012-09-21 16:34:55 <gavinandresen> squash (fixup) the second onto the first.
285 2012-09-21 16:35:05 <gavinandresen> ... and edit the first's commit message, if appropriate
286 2012-09-21 16:38:20 <kjj_> ok, the commit is right now, except for the comment
287 2012-09-21 17:08:14 <jgarzik> Bitfloor reopens: https://plus.google.com/109620439233076225324/posts/bLJRDHApjSP
288 2012-09-21 17:08:24 <jgarzik> met him at the conf... seemed like he was serious about fixing problems
289 2012-09-21 17:08:59 <jgarzik> the interesting bit to me: he might have taken advice to heart, about running a testnet exchange!
290 2012-09-21 17:11:32 <lianj> can he cover the lost coins?
291 2012-09-21 17:12:45 <jgarzik> lianj: over time, from profits, apparently
292 2012-09-21 17:14:44 <lianj> yea, just read the full post. thanks
293 2012-09-21 17:15:02 <kjj_> ooh. so he went fractional publicly?
294 2012-09-21 17:16:27 <phantomcircuit> jgarzik, maths says ~8 years time...
295 2012-09-21 17:22:50 <gavinandresen> jgarzik: recompiled 0.7.0 osx binaries now up on sourceforge; the osx10.5 problem was miscompiled dependencies. Still not sure what the Lion crash is/was, but I'm hoping compiling -pthread with an updated version of Qt will fix the issue.
296 2012-09-21 17:41:24 <midnightmagic> .. fractional..?!
297 2012-09-21 17:53:38 <enolan> I'm getting exception: St9bad_alloc and sometimes SIGABRTs and sometimes segfaults
298 2012-09-21 17:53:58 <enolan> all of this is new since upgrading to 0.7.0 via the official ubuntu ppa
299 2012-09-21 17:54:05 <enolan> here's a log: http://pastebin.com/ki1iyjxd
300 2012-09-21 17:54:46 <gmaxwell> enolan: are you running it with tight ulimits or something?
301 2012-09-21 17:56:06 <phantomcircuit> midnightmagic, hmm?
302 2012-09-21 17:56:24 <enolan> oh, yes. ulimit -v 1500000 is in my bashrc
303 2012-09-21 17:56:39 <Belkaar> Hello. Since the update to 0.7 i have "CWalletTx::GetAmounts: Unknown transaction type found, txid ----" in the log when using listtransactions. any ideas?
304 2012-09-21 17:56:54 <enolan> I put it there to stop runaway memory allocations in my code eating all the system memory
305 2012-09-21 17:56:59 <Belkaar> Example: Unknown transaction type found, txid ce8a1b3ecc5400238e913579868e6c4f092f529c49f9c9041dece43b6fab1097
306 2012-09-21 17:57:04 <phantomcircuit> iirc you're about 500MB away from working
307 2012-09-21 17:57:10 <gmaxwell> enolan: well there you go. Why the heck would you limit the virtual memory? other than TLB entries virtual memory is free.
308 2012-09-21 17:57:37 <phantomcircuit> FREE MEMMORIES
309 2012-09-21 17:57:41 <phantomcircuit> YOU HEARD IT HERE FIRST
310 2012-09-21 17:57:45 <phantomcircuit> ACTION runs
311 2012-09-21 17:58:00 <enolan> gmaxwell: it was supposed to stop runaway memory allocation in my code from eating all the system memory and causing swap thrash
312 2012-09-21 17:58:21 <enolan> would doubling it to 3gb be sufficient? this machine has plenty of ram
313 2012-09-21 17:58:53 <phantomcircuit> yes but more so why limit virtual memory
314 2012-09-21 17:59:03 <phantomcircuit> limiting actual resident memory i understand
315 2012-09-21 17:59:06 <phantomcircuit> but why virtual?
316 2012-09-21 18:00:20 <enolan> wouldn't limiting resident memory just force processes that allocate lots of RAM to swap thrash earlier?
317 2012-09-21 18:01:14 <gmaxwell> you're limiting the address space, not memory usage.
318 2012-09-21 18:01:55 <enolan> maybe I don't understand how ulimit -v works as much as I thought I did...
319 2012-09-21 18:19:16 <gavinandresen> Belkaar: that's a p2pool transaction with a weird, non-standard last output. Don't know what changed with listtransactions between the last release and this, maybe Luke-Jr would remember (he added code related to displaying generation transactions)
320 2012-09-21 18:20:18 <Belkaar> I think I read something about making txs non-standard that contain 0-outputs. maybe it's related to that?
321 2012-09-21 18:20:32 <Belkaar> the blocks seem to be accepted tho
322 2012-09-21 18:21:10 <Luke-Jr> gavinandresen: he's talking about in the debug.log, not output from the JSON-RPC itself
323 2012-09-21 18:21:15 <gavinandresen> Belkaar: non-standard just affects whether or not transactions are relayed across the network... since this is generation txn, that shouldn't apply
324 2012-09-21 18:21:31 <gavinandresen> Luke-Jr: right, but from what he says he didn't get that in debug.log before this release
325 2012-09-21 18:21:50 <gavinandresen> ... which makes me think some new information reported in listtransactions is triggering it
326 2012-09-21 18:22:46 <kjj_> confirmed, the messages appear in the log when running listtransactions
327 2012-09-21 18:23:34 <Luke-Jr> gavinandresen: yes, looks like you're right
328 2012-09-21 18:24:10 <Luke-Jr> gavinandresen: this warning is coming from code sorting out the accounts, which had been special-cased out for generation before
329 2012-09-21 18:25:13 <gavinandresen> Belkaar: can you open an issue on github?
330 2012-09-21 18:25:37 <Luke-Jr> I wonder if the client should be ignoring outputs it doesn't recognize
331 2012-09-21 18:25:42 <Belkaar> I'll try. not registered yet. but theres a first time for everything :-)
332 2012-09-21 18:26:56 <gavinandresen> ACTION has a feeling fixing that tiny little wart will open up a can of worms...
333 2012-09-21 18:27:53 <Luke-Jr> this explains another bug I had reported too
334 2012-09-21 18:28:47 <Luke-Jr> maybe
335 2012-09-21 18:31:59 <Belkaar> ok. opened an issue at github. i hope i did it right
336 2012-09-21 18:32:13 <kjj_> wait, does this mean that the 0.0 BTC marker transactions in the p2pool coinbases are getting added to our wallets?
337 2012-09-21 18:34:33 <Belkaar> what is a marker transaction?
338 2012-09-21 18:35:05 <Luke-Jr> Belkaar: looks good
339 2012-09-21 18:35:17 <Luke-Jr> kjj_: I doubt it
340 2012-09-21 18:35:22 <gavinandresen> it is a marker output, and yes, the transaction that output is part of is added to your wallet
341 2012-09-21 18:35:33 <gavinandresen> ... because you own one of the other outputs
342 2012-09-21 18:35:49 <gavinandresen> (assuming you use p2pool)
343 2012-09-21 18:35:50 <kjj_> oh, right. the wallet holds the whole transaction, not just my txout
344 2012-09-21 18:38:13 <Luke-Jr> ACTION interpreted the question as "does the marker txout appear in listtransactions?"
345 2012-09-21 18:39:05 <kjj_> ahh, no it does not. I run that pretty often to see if my rigs have earned anything recently.
346 2012-09-21 18:39:34 <gmaxwell> pedantically, it's not a 'marker' output, it's the merging the p2pool does. and it's there instead of the coinbase because python sucks. :P (forrest might protest, but I believe that it would be done normally if recalculating it from python weren't slow)
347 2012-09-21 18:40:47 <kjj_> heh, yeah, I didn't know what to call it exactly.
348 2012-09-21 18:43:32 <kjj_> so... do we ask him to fix it, or do we ignore 0.00000000 txouts in that loop, or do we ignore it?
349 2012-09-21 18:43:59 <Luke-Jr> it seems to me it would be appropriate to ignore txouts that aren't ours in this case
350 2012-09-21 18:44:29 <Luke-Jr> doesn't fix my problem, since they ARE mine (and it could be argued it shouldn't be fixed in this case), but it would fix for p2pool
351 2012-09-21 18:58:18 <Belkaar> Is there a way to get the last 100 transactions no matter the account?
352 2012-09-21 18:58:22 <jgarzik> is luke-jr the only mainnet miner that accepts non-standard transactions?
353 2012-09-21 18:58:35 <jgarzik> good for luke-jr, sad for experimenters.
354 2012-09-21 18:59:06 <kjj_> bitcoind listtransactions "*" 100
355 2012-09-21 19:00:00 <Belkaar> thanks
356 2012-09-21 19:18:53 <Luke-Jr> jgarzik: I bet there would be more, if it didn't require recompiling bitcoind custom
357 2012-09-21 19:19:33 <Luke-Jr> jgarzik: let me know if you guys ever want to reconsider a hidden config option??? ;)
358 2012-09-21 19:20:00 <jgarzik> Luke-Jr: if you have sufficient hashpower on the network, I imagine one has sufficient skill to compile
359 2012-09-21 19:21:59 <Luke-Jr> jgarzik: I'm thinking about p2pool miners mostly
360 2012-09-21 20:38:32 <diki> I've noticed that the pager on the forum outputs ALL pages from 1 to X and am wondering why it's not fixed to display 1,2,3...NUMpages
361 2012-09-21 22:41:01 <Dragonsangel> Heya, I am looking for a comparison tool to compare hashes made from UTF-16 string? Anyone know any sites that do them? I only seem to find UTF-8 versions
362 2012-09-21 22:59:27 <Anonymous-man> Donate Bitcoins to the cause of Anonymous! http://pastebin.com/r56SVjBs