1 2011-09-07 00:06:53 <cjdelisle> hrm
  2 2011-09-07 00:07:09 <cjdelisle> no nxstack and stack canaries flags
  3 2011-09-07 00:24:34 <copumpkin> jgarzik, nameless|: you around?
  4 2011-09-07 00:47:22 <diki> im around
  5 2011-09-07 00:48:11 <copumpkin> gmaxwell would work too
  6 2011-09-07 00:48:31 <gmaxwell> Wha?
  7 2011-09-07 00:49:03 <copumpkin> gmaxwell: think you could kill the annoying spambot in #bitcoin? it PMs people for petty reasons :)
  8 2011-09-07 02:00:39 <CIA-92> bitcoin: aldiyen * r7424dd005368 Phoenix-Miner-personal/WorkQueue.py: Log abandon work message as debug
  9 2011-09-07 02:10:38 <CIA-92> bitcoin: aldiyen * r1a28a109c070 Phoenix-Miner-personal/WorkQueue.py: Forget that we abandoned work next time a range is fetched, even if there was already work to be had
 10 2011-09-07 02:20:41 <CIA-92> bitcoin: aldiyen * rc9a9cd2e8068 Phoenix-Miner-personal/Miner.py: Set version to v0.2.4
 11 2011-09-07 02:39:18 <bittwist> SolidCoin: fixing the 51% attack flaw in bitcoin
 12 2011-09-07 02:39:21 <bittwist> details at 11
 13 2011-09-07 02:41:23 <ymirhotfoot> Yes.  By a subtle cryptographic analogue of the New US Senate Rules, a 60% majority, beyond the old 50%+ majority is required for certification of SOLIDCOIN WARRANTS.
 14 2011-09-07 02:51:58 <ymirhotfoot> Perhaps of interest, a gathering 10-12 September 2011 in new York City:
 15 2011-09-07 02:52:01 <ymirhotfoot> http://openvideoconference.org/2011/08/building-a-new-payment-standard-on-the-web/
 16 2011-09-07 03:17:49 <nameless> |copumpkin: what's up?
 17 2011-09-07 08:20:45 <CIA-92> bitcoin: Con Kolivas * r26d1264c005a cgminer/README: FAQ updates.
 18 2011-09-07 08:30:46 <CIA-92> bitcoin: Con Kolivas * r01a0d0a7452b cgminer/main.c: Add a --no-adl option to disable ADL monitoring and GPU settings.
 19 2011-09-07 09:30:41 <shadders> Question for the protocol experts:
 20 2011-09-07 09:31:33 <MrSam> ACK
 21 2011-09-07 09:32:12 <D0han> timeout, RST
 22 2011-09-07 09:32:37 <shadders> If I send a tx to bitcoind because I've received it but can't verify it myself (because I'm a dumb node).  I don't want to wait for bitcoind until it feels like repropogating the tx.. If I calculate the hash myself then I can send a getdata requesting the specific tx...
 23 2011-09-07 09:33:00 <shadders> if the tx is valid presumably bitcoind will return it...
 24 2011-09-07 09:33:11 <shadders> if it's not then will it simply not respond?
 25 2011-09-07 09:33:48 <shadders> or can I rely on bitcoind to immediately rebroadcast the tx once it's verified?
 26 2011-09-07 09:34:21 <bespike> hello, could anyone tell me how to make my own coins, like bitcoins and soldcoins?
 27 2011-09-07 09:34:42 <shadders> you need a blacksmith
 28 2011-09-07 09:35:01 <bespike> very funny
 29 2011-09-07 09:35:01 <tcatm> shadders: I don't think you should assume the bitcoind will respond to INV(TX) or broadcast it immediately.
 30 2011-09-07 09:35:40 <shadders> are new tx's not considered high priority for relaying?
 31 2011-09-07 09:35:40 <tcatm> bespike: change all "magic values", change the port, genesis block and remove the checkpoints
 32 2011-09-07 09:36:11 <bespike> thank you tcatm :)
 33 2011-09-07 09:36:41 <tcatm> oh and change the address version, too
 34 2011-09-07 09:36:57 <bespike> ok :)
 35 2011-09-07 09:39:03 <tcatm> shadders: not exactly. the code will do that but that might change. it would be better to ask it for everything you need to verfiy the transaction yourself
 36 2011-09-07 09:42:44 <shadders> verifying tx is not an attractive option for the scenario... the 'dumb node' is a proxy in front of bitcoind taking network load off it.  But in some case like 'receive new tx' you want to repropogate as quick as possible.  But don't want propogate until the tx is validated... which is bitcoind's job..
 37 2011-09-07 09:42:53 <shadders> it's a bit of a square peg, round hole problem...
 38 2011-09-07 09:46:03 <tcatm> shadders: is that the proxy mentioned on the mailing list?
 39 2011-09-07 09:46:37 <shadders> yep
 40 2011-09-07 09:48:20 <tcatm> in that case you should just send it to bitcoind and relay it once you get it back
 41 2011-09-07 09:52:01 <shadders> I suppose the delay is no more than you'd have from a normal bitcoind node in that case...
 42 2011-09-07 09:53:32 <shadders> I also suspect I'm going to have to maintain 2 connections to the bitcoind.  I'm guessing it won't bother to relay the tx back through the connection it came from... So the 2nd connection should fool it into thinking it's a different node and it should come back through that connection
 43 2011-09-07 09:54:03 <tcatm> the code in main.cpp does relay it just after verification within the same thread
 44 2011-09-07 09:55:57 <shadders> ahh... does it exclude the incoming connection from the relay?
 45 2011-09-07 09:56:30 <shadders> sorry, handicapped.. no c++ here...
 46 2011-09-07 09:57:00 <tcatm> I'm not sure but I can't find any code that prevents it from sending it back.
 47 2011-09-07 09:57:49 <shadders> ok... will test it... but probably shouldn't assume it always will send back on originating conn...
 48 2011-09-07 09:58:38 <tcatm> I think it would be okay to tweak the protocol a little to make proxying easier if needed.
 49 2011-09-07 10:06:01 <shadders> hmm... The trickiest bit is how to deal with request/response exchanges... For bitcoind the response route is obvious because it's the connection the request came from...
 50 2011-09-07 10:06:47 <shadders> introducing sequence id's would be a big pain the arse on the bitcoind side...
 51 2011-09-07 10:07:15 <tcatm> requests are stateless, aren't they?
 52 2011-09-07 10:09:26 <shadders> the only way I can think to attack it is to generate a unique key from the request that can regenerated from the expected response... e.g. getdate: request_type + hashes... this won't necessarily be unique to clients but that's ok because if two identical requests come in the same response can be fed back to both...
 53 2011-09-07 10:09:37 <shadders> getdate=getdata
 54 2011-09-07 10:10:13 <shadders> tcatm: stateless yes but a stateless request still implies you know who to send the response to...
 55 2011-09-07 10:10:32 <shadders> bitcoind won't because it see's all requests coming from one node
 56 2011-09-07 10:11:49 <tcatm> you could keep a small map in the proxy. i.e. "client 2 requested tx with hash XYZ". then you request that tx from bitcoind (or from your local cache) and forward it to client 2
 57 2011-09-07 10:13:15 <shadders> yeah that's what I was thinking with the unique key that can be generated off both the request and the response... butt ugly compared to a sequence id but it does allow for reuse of responses to identical requests from remote nodes
 58 2011-09-07 10:15:04 <shadders> really depends how intelligent you want to get... and where I'm having trouble working out where to draw the line... e.g. with getdata you could even cache all the blocks individually.. if a new getdata comes in check if you've got all the requested blocks in cache if not send bitcoind a getdata for the missing ones otherwise just build the response and send it...
 59 2011-09-07 10:15:52 <shadders> but if you take it to the nth degree you'll just end up building a new bitcoind node ;0
 60 2011-09-07 10:15:59 <tcatm> a solid cache framework is the most important part of such a proxy
 61 2011-09-07 10:16:32 <tcatm> or maybe just use memcached
 62 2011-09-07 10:16:42 <shadders> tell me about it... poolserverj's got caches coming out it's arse...
 63 2011-09-07 10:18:03 <TuxBlackEdo> i am selling dirt 3 keys for 1btc
 64 2011-09-07 10:21:45 <Eliel> MagicalTux: I was looking at this graph http://pi.uk.com/bitcoin/charts/n-transactions and started wondering how big portion of those are transfer initiated or received by mtgox. Is there any chance you might be willing to publish that data?
 65 2011-09-07 10:23:10 <MagicalTux> hm
 66 2011-09-07 10:23:12 <MagicalTux> not that much
 67 2011-09-07 10:23:22 <MagicalTux> most of the traffic is coming from money laundering
 68 2011-09-07 10:24:34 <ThomasV> huh?
 69 2011-09-07 10:24:53 <xelister> ThomasV: Im just wondering what if entire world would be more like usa
 70 2011-09-07 10:25:21 <ThomasV> xelister: np, I was reacting to MagicalTux's statement
 71 2011-09-07 10:27:45 <Eliel> MagicalTux: money laundering... as in, just moving bitcoins between different addresses a few times?
 72 2011-09-07 10:28:38 <ThomasV> moving coins is not sufficient to launder money
 73 2011-09-07 10:28:53 <ThomasV> you need to mix your coins with someone elses's coins
 74 2011-09-07 10:32:28 <MagicalTux> that's what's happening every day on the blockchain
 75 2011-09-07 10:36:52 <phantomcircuit> Eliel, people seem to think that obscures the origin
 76 2011-09-07 10:36:53 <phantomcircuit> but
 77 2011-09-07 10:36:56 <phantomcircuit> it really doesn't
 78 2011-09-07 11:06:09 <WildSoil> ;;bc,stats
 79 2011-09-07 11:06:11 <gribble> Current Blocks: 144353 | Current Difficulty: 1777774.4820015 | Next Difficulty At Block: 145151 | Next Difficulty In: 798 blocks | Next Difficulty In About: 5 days, 13 hours, 26 minutes, and 36 seconds | Next Difficulty Estimate: 1807327.39582545
 80 2011-09-07 11:06:12 <xelister> Eliel: you can make money more private by moving
 81 2011-09-07 11:06:14 <xelister> in the chain
 82 2011-09-07 11:06:30 <xelister> but yea using exchange it could be another protection
 83 2011-09-07 12:20:59 <ThomasV> gavinandresen: hi, did you find the deadlock ?
 84 2011-09-07 12:21:14 <gavinandresen> yup
 85 2011-09-07 12:21:24 <gavinandresen> wait, no, I lied-- sipa found the deadlock
 86 2011-09-07 12:21:56 <gavinandresen> I'll pull the patch as soon as I finish writing an email in that other window.
 87 2011-09-07 12:22:12 <ThomasV> I just compiled from git. I noticed that when I do "getinfo", the response time is highly variable. could it be that it is waiting for an event on the network before it answers rpc ?
 88 2011-09-07 12:23:45 <gavinandresen> if it is busy processing events from the network the getinfo will have to wait...
 89 2011-09-07 12:24:09 <ThomasV> it is downloading blocks
 90 2011-09-07 12:24:22 <gavinandresen> ... that'll do it
 91 2011-09-07 12:24:55 <gavinandresen> I kind of used a sledgehammer to fix a whole slew of deadlock issues with RPC calls--
 92 2011-09-07 12:25:24 <ThomasV> a sledgehammer ?
 93 2011-09-07 12:25:28 <gavinandresen> Each RPC call used to be responsible for obtaining the locks it needed to do its thing.  Now the RPC handler obtains the cs_main and cs_wallet locks before calling the RPC method
 94 2011-09-07 12:26:07 <ThomasV> oh, is that the reason why I see this delay ?
 95 2011-09-07 12:26:09 <gavinandresen> That way the RPC code can't screw up and obtain them in the wrong order.  It does make RPC calls that don't really do much (like getinfo) slower.
 96 2011-09-07 12:26:24 <ThomasV> ok
 97 2011-09-07 12:29:19 <gavinandresen> There's more reworking of locks to be done, but that's a job for another release/refactoring.
 98 2011-09-07 12:29:31 <UukGoblin> gavinandresen, ah, running a full day on -noirc now, it's still alive(tm)
 99 2011-09-07 12:29:47 <CIA-92> bitcoin: Gavin Andresen master * r9406696 / src/net.cpp : Fix AddAddress cs_mapaddresses/db transaction deadlock - http://git.io/2mGiHg
100 2011-09-07 12:31:20 <gavinandresen> I've been running with the deadlock fix all night no problems (with irc)
101 2011-09-07 12:32:10 <ThomasV> is this tx spam ? http://blockexplorer.com/tx/e9f5520572e4657417ae58abae736fb4d7b4dac23ed5d471c6ac2063551a9acd
102 2011-09-07 12:32:35 <ThomasV> it is the end of a long series of tx sending 0.0000001 btc
103 2011-09-07 12:32:49 <ThomasV> each hasa a 0.005 fee
104 2011-09-07 12:36:47 <gavinandresen> There have been a lot of tiny, spammy transactions added to the block-chain recently.  Somebody is being antisocial and trying to see if they can break bitcoin in a similar way to the way ArtForz (accidently) half-broke solidcoin
105 2011-09-07 12:37:48 <ThomasV> this is older though
106 2011-09-07 12:40:10 <gavinandresen> ... huh, right you are.  Do the addresses decode into ASCII?  Could be somebody copying Dan Kaminsky's hack to get content into the block-chain.
107 2011-09-07 12:41:03 <ThomasV> I did not check
108 2011-09-07 12:41:57 <ThomasV> I have been running a script to see where allinvain's stolen bitcoins have been going, and I found this very long chain of transactions
109 2011-09-07 12:42:48 <martind> I hope people don't mind that I post this here, #bitcoin seems to be more focused on mining/trading-related activities: https://bitcoin.org.uk/forums/topic/210-call-for-contributions-bitcoin-weekend-at-the-london-hackspace-2425-sept-2011/ we're planning a Bitcoin event at the London Hackspace on 24/25 Sept and are looking for people interested in giving talks, holding workshops, or contributing in
110 2011-09-07 12:44:47 <ThomasV> martind: is that the same as what genjix organises ?
111 2011-09-07 12:55:51 <BurtyB> martind, you're having 2 days of talks?
112 2011-09-07 13:04:23 <phantomcircuit> cjdelisle, lol
113 2011-09-07 13:04:36 <cjdelisle> really
114 2011-09-07 13:05:18 <cjdelisle> if you don't have a lot of processor demand (generally bitcoin doesn't) just use event programming and one thread.
115 2011-09-07 13:05:21 <cjdelisle> no more lockd
116 2011-09-07 13:05:24 <cjdelisle> *locks
117 2011-09-07 13:06:10 <cjdelisle> if you have a ton of processor demand and you really need to exploit multicore, spawn another process to do a single job and pipe the data to it.
118 2011-09-07 13:07:02 <cjdelisle> I can't see threads really exploiting the power of a multicore processor if they are constantly grabbing locks.
119 2011-09-07 13:07:23 <cjdelisle> let alone a multi socket machine
120 2011-09-07 13:07:39 <edcba> depends on how much time they hold the lock
121 2011-09-07 13:08:36 <cjdelisle> Every time you grab a lock, you need to use the buss between the sockets to make sure nobody else is holding that lock
122 2011-09-07 13:08:41 <cjdelisle> (that is the worst case)
123 2011-09-07 13:09:42 <cjdelisle> if there are cores that don't have the lock in their cache then they can be skipped....
124 2011-09-07 13:09:59 <cjdelisle> threads == endless PITA
125 2011-09-07 13:10:34 <log0s> i think the bitcoin code really just needs a total rewrite...obviously the developers disagree, or they'd have done it already
126 2011-09-07 13:10:43 <edcba> i agree
127 2011-09-07 13:10:50 <edcba> but i'm too lazy to do it
128 2011-09-07 13:11:26 <jgarzik> oh, obviously
129 2011-09-07 13:11:31 <edcba> ;;bc,mtgox
130 2011-09-07 13:11:31 <gribble> {"ticker":{"high":7.59916,"low":6.402,"avg":6.837939745,"vwap":6.898355101,"vol":50716,"last":6.86103,"buy":6.86103,"sell":6.86282}}
131 2011-09-07 13:11:36 <edcba> ouch
132 2011-09-07 13:11:51 <jgarzik> the world needs a rewrite.  but obviously people don't care, or they would have done it already.
133 2011-09-07 13:11:58 <BlueMatt> log0s: total rewrite is never the right solution, piece-by-piece rewrite is the way to go...
134 2011-09-07 13:11:58 <cjdelisle> ^ :)
135 2011-09-07 13:12:15 <cjdelisle> yup, modularize and swap
136 2011-09-07 13:12:17 <BlueMatt> log0s: also, rewriting is a shitton of work and should always be a last resort
137 2011-09-07 13:12:56 <helo> i feel like this is #python discussing threads :)
138 2011-09-07 13:13:03 <cjdelisle> has anyone tried building with nxstack and stack canaries?
139 2011-09-07 13:13:04 <edcba> log0s: you know there is bitcoinj ? :)
140 2011-09-07 13:13:30 <jrmithdobbs> also libbitcoin
141 2011-09-07 13:13:35 <log0s> BlueMatt: took me only a few weeks worth of work spread out over a month to write my own java bitcoin library (no json-rpc or gui)
142 2011-09-07 13:13:39 <gavinandresen> cjdelisle: will nxstack catch something that valgrind won't?
143 2011-09-07 13:13:59 <cjdelisle> nxstack is just important to preventing stack smashing attacks
144 2011-09-07 13:14:24 <cjdelisle> If there's a buffer overflow, setting the nx bit on the stack pages will make some blackhat very sad
145 2011-09-07 13:14:32 <edcba> or not
146 2011-09-07 13:14:34 <log0s> i'll finally get around to testing my networking code one of these days soon, as that's the most recent part i've written
147 2011-09-07 13:14:38 <BlueMatt> log0s: yes, and when your devs are all volunteers and have day-jobs that takes a long time and is a lot of work that could be put into doing something more useful...
148 2011-09-07 13:15:05 <cjdelisle> but there is some annoying stuff that requires executable stack like some trickery used in mplayer so sometimes it just doesn't work.
149 2011-09-07 13:15:34 <edcba> but bitcoin doesn't use trickeries like that
150 2011-09-07 13:17:20 <log0s> edcba: yes, i know of bitcoinj. i looked at bitcoinj's and openssl's code to see how expanding and compacting the target worked when i was working on my library
151 2011-09-07 13:18:09 <BlueMatt> log0s: also, why are you writing another java bitcoin library? bitcoinj is very well done and well commented and I know TD is always looking for more people to contribute
152 2011-09-07 13:18:25 <gavinandresen> cjdelisle: setting the nx bit on machines that support it sounds like a very good idea. Don't know if anybody has tried it.
153 2011-09-07 13:18:37 <cjdelisle> trying a build now
154 2011-09-07 13:19:06 <BlueMatt> (on ubuntu/debian aka bitcoin's releases)
155 2011-09-07 13:20:51 <cjdelisle> -pie -fPIE -fstack-protector-all -Wstack-protector -D_FORTIFY_SOURCE=2 -Wa,--noexecstack -Wl,-z,relro -Wl,-z,now -Wl,-z,noexecstack
156 2011-09-07 13:21:53 <cjdelisle> not likeing PIE, I think it's because I didn't build wx with fPIC
157 2011-09-07 13:21:59 <jrmithdobbs> gavinandresen: since you're around again i'd like to bring this one back up as well: https://github.com/bitcoin/bitcoin/pull/202
158 2011-09-07 13:23:35 <jrmithdobbs> if nxstack is worth considering to protect against possible future issues, so is that ;p
159 2011-09-07 13:24:01 <BlueMatt> nxstack doesnt require setting a suid bit on bitcoin
160 2011-09-07 13:24:07 <gavinandresen> jrmithdobbs: i don't know nuthin about linux capabilities bounding set support. Doesn't seem to be a lot of enthusiasm
161 2011-09-07 13:24:13 <jrmithdobbs> BlueMatt: neither does that
162 2011-09-07 13:24:25 <phantomcircuit> you dont need to suid bitcoin to use capabilities...
163 2011-09-07 13:24:28 <BlueMatt> jrmithdobbs: thats what you told me
164 2011-09-07 13:24:43 <BlueMatt> or...sorry that you have to add other similar bits...
165 2011-09-07 13:24:43 <jrmithdobbs> BlueMatt: no that's what you selectively remembered of the 3 ways to make that work
166 2011-09-07 13:24:56 <BlueMatt> jrmithdobbs: iirc the other ones werent much of an option?
167 2011-09-07 13:25:44 <jrmithdobbs> BlueMatt: one is starting it as root, one is setuid root ... both of those are horrible ideas, the correct way to do it is to use filesystem capabilities to give it *one* capability that only allows it to set the processes capabilities which it uses to drop all capabilities first thing in main()
168 2011-09-07 13:26:07 <jrmithdobbs> and set the bounding set so that they can't be renabled
169 2011-09-07 13:26:15 <BlueMatt> jrmithdobbs: problem is, you are setting odd bits on the bitcoin exe, which I see as a problem as we dont currently have an installer
170 2011-09-07 13:26:19 <cjdelisle> +#ifdef __cplusplus
171 2011-09-07 13:26:20 <cjdelisle> +extern "C" {
172 2011-09-07 13:26:22 <cjdelisle> sold
173 2011-09-07 13:26:53 <jrmithdobbs> BlueMatt: doesn't matter, it's a build time option and the code does nothing at all if it doesn't have the perms to do it
174 2011-09-07 13:28:02 <log0s> BlueMatt: bitcoinj, like bitcoind, does not provide the modularity and flexibility that i want...libbitcoin is much closer to what i want in those respects, but i have my reasons for not adopting that as well
175 2011-09-07 13:28:13 <BlueMatt> jrmithdobbs: meh, Im on the fence here, I dont honestly see it as adding much to bitcoin at all...it could protect the rest of the system from further exploitation, but thats not up to the program to do...I would prefer things like apparmour or selinux rules for bitcoin
176 2011-09-07 13:28:20 <jrmithdobbs> log0s: i argued with him so hard not to make that gplv3 ;p
177 2011-09-07 13:28:23 <phantomcircuit> log0s, be interesting to know what those are?
178 2011-09-07 13:28:40 <jrmithdobbs> at least, i'm guessing it's the gplv3 thing
179 2011-09-07 13:29:18 <jrmithdobbs> BlueMatt: apparmor/selinux actually don't provide this same protection
180 2011-09-07 13:29:23 <BlueMatt> log0s: how much modularity do you need
181 2011-09-07 13:29:34 <BlueMatt> jrmithdobbs: chroot jail then
182 2011-09-07 13:29:40 <jrmithdobbs> BlueMatt: that doesn't either
183 2011-09-07 13:30:29 <jrmithdobbs> BlueMatt: it's something different just like nxstack is
184 2011-09-07 13:30:34 <BlueMatt> if you have no way of getting root it does
185 2011-09-07 13:30:38 <jrmithdobbs> (though nx is at the hw level)
186 2011-09-07 13:33:48 <jrmithdobbs> BlueMatt: there are some obscure cases where that's not entirely true
187 2011-09-07 13:34:01 <jrmithdobbs> BlueMatt: but the same arguments you just made could be used against wallet crypto and you're fine with that. :)
188 2011-09-07 13:34:21 <luke-jr> BlueMatt: sorry you can't read English
189 2011-09-07 13:34:38 <BlueMatt> luke-jr: ?
190 2011-09-07 13:34:46 <BlueMatt> jrmithdobbs: I dont think wallet crypto is worth anything, the reason it gets merged (IMO) is that its good for pr
191 2011-09-07 13:34:58 <BlueMatt> setting bounding capability set...not so much
192 2011-09-07 13:35:11 <d33tah> i heard bitcoin devs work on switching from wx to qt. is that true?
193 2011-09-07 13:35:36 <BlueMatt> d33tah: someone wrote qt version, and it will probably be merged and transitioned to after 0.4, maybe as early as 0.4.1
194 2011-09-07 13:35:36 <gavinandresen> wallet crypto is good to prevent casual my-little-brother-spent-my-coins-on-porn attacks. and to protect backups.
195 2011-09-07 13:35:59 <jrmithdobbs> BlueMatt: you get what i'm saying, though, i obviously understand the pr aspects ;p
196 2011-09-07 13:36:08 <jrmithdobbs> it's just a precautionary measure
197 2011-09-07 13:36:18 <wardearia> BlueMatt, I would imagine a lot of modulously modular modulism including extra modulistic modulosity
198 2011-09-07 13:36:37 <BlueMatt> jrmithdobbs: the difference is wallet crypto protects bitcoin, bounding set protects the rest of the system from bitcoin failing
199 2011-09-07 13:36:38 <jrmithdobbs> i just thought I'd bring it up again because the reasons it was closed out were because the person who closed it out doesn't understand what the code does and didn't bother reading the documentation that was clearly referenced in the request
200 2011-09-07 13:36:42 <d33tah> i had an idea recently
201 2011-09-07 13:37:14 <BlueMatt> jrmithdobbs: Im pretty sure jgarzk understands how linux works
202 2011-09-07 13:37:21 <BlueMatt> pretty damn well Id say
203 2011-09-07 13:37:25 <d33tah> do you think it would be possible to create an opensource 'paranoid bitcoin token' thingy, that would store the private keys of user's money and would never send them to computer?
204 2011-09-07 13:37:35 <jrmithdobbs> BlueMatt: his statements in that thread of that pull request demonstrates he does not know how that portion works
205 2011-09-07 13:37:41 <d33tah> when a transaction is made, it would generate it itself and just pass the encrypted data
206 2011-09-07 13:37:48 <MacRohard> d33tah, it's called a smart card
207 2011-09-07 13:37:59 <d33tah> kind of, but open source
208 2011-09-07 13:38:02 <BlueMatt> d33tah: its been brought up many times, yes its possible, there are several projects which have never gone anywhere, but a ecdsa smart card could do it
209 2011-09-07 13:38:05 <jrmithdobbs> BlueMatt: or that it's been part of the kernel for a decade (sans bounding set which is a trivial extension to it)
210 2011-09-07 13:38:08 <manveru> d33tah: try yubikey
211 2011-09-07 13:38:10 <d33tah> some atmega thingy
212 2011-09-07 13:38:25 <jrmithdobbs> BlueMatt: or that lots of very well established software does in fact use it
213 2011-09-07 13:38:31 <d33tah> manveru: i'll have a look
214 2011-09-07 13:38:42 <jrmithdobbs> but sure other than not understanding it at all i guess he seems to understand it perfectly?!
215 2011-09-07 13:38:55 <wardearia> jrmithdobbs, of course!  don't you understand?
216 2011-09-07 13:39:05 <BlueMatt> jrmithdobbs: like?
217 2011-09-07 13:39:10 <jrmithdobbs> BlueMatt: bind
218 2011-09-07 13:39:19 <BlueMatt> mmm
219 2011-09-07 13:39:19 <d33tah> nah, i don't like the yubikey idea
220 2011-09-07 13:39:21 <wardearia> Prepare for the mothingness
221 2011-09-07 13:39:22 <jrmithdobbs> BlueMatt: the kernel internals itself
222 2011-09-07 13:39:36 <jrmithdobbs> BlueMatt: go read the thread in that pull request i already answered these questions :(
223 2011-09-07 13:39:48 <jrmithdobbs> and clarified the misconceptions
224 2011-09-07 13:40:08 <jrmithdobbs> and i have to clean my damned keyboard now because something is stuck under one of my shift keys
225 2011-09-07 13:40:13 <BlueMatt> jrmithdobbs: "this has nothing to do with guarding against running as root" oh really, what does it do then?
226 2011-09-07 13:40:14 <d33tah> d33tah@deetah-laptop:~/bitcoin-qt/src$ LC_ALL=C make -f makefile.unix
227 2011-09-07 13:40:15 <d33tah> make: *** No rule to make target `obj/crypter.o', needed by `bitcoin'.  Stop.
228 2011-09-07 13:40:17 <d33tah> wtf?
229 2011-09-07 13:40:19 <BlueMatt> prevents getting root access
230 2011-09-07 13:40:37 <d33tah> ah, qmake
231 2011-09-07 13:41:25 <gavinandresen> jrmithdobbs: if jgarzik doesn't get it, then there is little hope that lots of other people will.  So there's likely to be a lot of "what the heck is THIS doing in bitcoin???" questions.  Which wastes everybody's time, all for a "shutting the barn door after the cows have come home" feature.   In my opinion.
232 2011-09-07 13:41:55 <gavinandresen> (sorry for the mixed metaphor, I meant shutting the barn door after your chickens have come home to roost)
233 2011-09-07 13:44:22 <jrmithdobbs> gavinandresen: which is why I contributed it as a build-time option turned off by default that's completely segregated code, people who do "get it" can turn it on and people who don't will never see it
234 2011-09-07 13:44:51 <BlueMatt> then you get even more questions from people trying to build asking "should i turn this on or not and what does it do?"
235 2011-09-07 13:45:03 <phantomcircuit> lol
236 2011-09-07 13:45:14 <phantomcircuit> just turn it off by default
237 2011-09-07 13:45:18 <phantomcircuit> 99% of people wont notice
238 2011-09-07 13:45:24 <jrmithdobbs> BlueMatt: jgarzik was saying the point was to prevent people from running bitcoind as root at all, that's a side effect but not the intent, the intent is to prevent other bad things happening if bitcoin is compromised in some way
239 2011-09-07 13:45:45 <jrmithdobbs> BlueMatt: you mean like how every single person who builds it doesn't understand the broke use of USE_UPNP?
240 2011-09-07 13:45:57 <BlueMatt> jrmithdobbs: yep, exactly
241 2011-09-07 13:46:25 <luke-jr> s/blames/credits/
242 2011-09-07 13:46:34 <BlueMatt> call it as you like
243 2011-09-07 13:47:20 <luke-jr> the problem is the build system not being the standard ./configure etc people are used to
244 2011-09-07 13:47:28 <BlueMatt> true
245 2011-09-07 13:48:24 <gavinandresen> patches welcome
246 2011-09-07 13:48:36 <gavinandresen> (but I don't know nuthin about autotools)
247 2011-09-07 14:02:33 <martind> ThomasV: yes, the same thing
248 2011-09-07 14:02:41 <martind> BurtyB: one day talks, one day workshops
249 2011-09-07 14:02:55 <martind> (roughly)
250 2011-09-07 14:11:38 <jrmithdobbs> gavinandresen: patches welcome? jaromil had a fully functional autotools that was tested and working 90%+ cases and it never got merged so fell behind to the point where it'd almost have to be redone?!
251 2011-09-07 14:12:49 <gavinandresen> jrmithdobbs: i didn't follow the debate around jaromil's patch.  Like I said, I don't know nuthin about autotools.
252 2011-09-07 14:13:36 <BlueMatt> jrmithdobbs: jaromil had a functional for ubuntu autotools that was tested and known to not work on many platforms and it never got merged so fell behind when jaromil failed to update with patches that were sent to him
253 2011-09-07 14:14:02 <BlueMatt> jrmithdobbs: and from what Ive been told (by people who know better than me) his autotools was kinda hacked together
254 2011-09-07 14:15:44 <jrmithdobbs> BlueMatt: iirc they were patches to things he couldn't test, it should have been merged when it was working for most cases (which it was outside of win32) so that the issues could be sorted out by a wider audience
255 2011-09-07 14:16:52 <jrmithdobbs> gavinandresen: also, how do I need to change that >2G log patch on 32bit linux to get it merged? That's a bugfix. Tell me how to get it merged. The current patch has less race conditions than the original code did. (which was jgarzik's objection)
256 2011-09-07 14:17:41 <gavinandresen> jrmithdobbs: bottleneck is testing, but that should be getting better as alex gets up to speed
257 2011-09-07 14:18:10 <gavinandresen> jrmithdobbs: recuit some people to help q/a and that'd be a big help in getting patches accepted or rejected quicker
258 2011-09-07 14:18:37 <gavinandresen> (writing a test-plan for the >2G log would be good, too)
259 2011-09-07 14:18:46 <gavinandresen> (how WILL that get tested???)
260 2011-09-07 14:19:15 <jrmithdobbs> gavinandresen: what needs to be tested specifically on that >2GB log patch? I tested that it would continue appending to the log beyond 2GB and removed the auto-trucating behaviour completely so I'm not sure what else needs testing
261 2011-09-07 14:19:41 <gavinandresen> developers should never test their own code.  Me included.
262 2011-09-07 14:19:45 <jrmithdobbs> I think the build options might need to be slightly different on *bsd/solaris/etc vs linux
263 2011-09-07 14:19:55 <jrmithdobbs> gavinandresen: I agree, I'm just saying what needs to be tested?
264 2011-09-07 14:20:40 <gavinandresen> I dunno, haven't looked at that pull request in a while, there are much higher priorities
265 2011-09-07 14:20:44 <jrmithdobbs> I mean, the test plan for that is pretty straight forward: dd if=/dev/zero of=debug.log; tail -f debug.log & bitcoind -daemon; and make sure there's logs getting appended
266 2011-09-07 14:21:09 <gavinandresen> jrmithdobbs: cool, so put that in the pull request and add it to alex's queue.
267 2011-09-07 14:21:17 <jrmithdobbs> how do I do the latter?
268 2011-09-07 14:21:27 <gavinandresen> alex posted to the bitcoin-dev mailing list
269 2011-09-07 14:21:37 <alexwaters> :)
270 2011-09-07 14:21:43 <jrmithdobbs> oh, and it has to be done on a 32bit machine/build, of course
271 2011-09-07 14:24:10 <jrmithdobbs> gavinandresen: what was the subject? I don't recall that message
272 2011-09-07 14:24:35 <Matth1a3> http://bit.ly/qWosyB
273 2011-09-07 14:24:55 <Matth1a3> jrmithdobbs: that link
274 2011-09-07 14:27:46 <tcatm> http://eu1.bitcoincharts.com/blockchain/ daily signed blockchain snapshots
275 2011-09-07 14:27:52 <jrmithdobbs> alexwaters: there you go.
276 2011-09-07 14:28:15 <alexwaters> thank you, awesome
277 2011-09-07 14:30:12 <jrmithdobbs> alexwaters: it probably needs to be tested on *bsd/solaris/etc like i said, because the defines might be a bit different for different libcs
278 2011-09-07 14:31:40 <alexwaters> ok
279 2011-09-07 14:39:58 <phantomcircuit> tcatm, are you aware that bcc was down sporadically today?
280 2011-09-07 14:45:28 <tcatm> phantomcircuit: yes. the server is a little busy with all those new markets (15 mtgox currencies and 16 from ruxum)
281 2011-09-07 14:46:31 <phantomcircuit> they have almost no activity though
282 2011-09-07 14:47:43 <tcatm> The way I store the market depth is a little inefficient. I'm working on that
283 2011-09-07 14:51:56 <ymirhotfoot> Bloom filters
284 2011-09-07 14:52:12 <ymirhotfoot> More Bloom filters and more Merkle trees
285 2011-09-07 14:52:25 <phantomcircuit> tcatm, i can help you with that for a fee
286 2011-09-07 14:53:38 <tcatm> phantomcircuit: how much?
287 2011-09-07 14:53:58 <phantomcircuit> depends on how screwy your current setup is
288 2011-09-07 14:54:06 <tcatm> very screwy ;)
289 2011-09-07 14:54:16 <phantomcircuit> lol
290 2011-09-07 14:55:01 <phantomcircuit> well im going to add PLN to intersango.com tonight but after that my schedule is only completely filled
291 2011-09-07 14:55:04 <phantomcircuit> so
292 2011-09-07 14:55:09 <phantomcircuit> name a price
293 2011-09-07 14:56:24 <phantomcircuit> using mysql?
294 2011-09-07 14:58:09 <tcatm> yes, but I've changed the database layout already and wrote wrappers. Only migrating the old code to use the new data format is not done yet. I think I'm going to write that tomorrow.
295 2011-09-07 14:58:27 <phantomcircuit> ah
296 2011-09-07 14:58:31 <phantomcircuit> new database fast enough?
297 2011-09-07 14:58:51 <tcatm> yes
298 2011-09-07 15:00:57 <phantomcircuit> you'll probably be fine them
299 2011-09-07 15:32:24 <unclemanti> i have bitcoind running as daemon'
300 2011-09-07 15:32:35 <unclemanti> i am trying to get the current block and i am getting this error
301 2011-09-07 15:32:55 <unclemanti> unclemantis@li291-14:~/bitcoin/src$ ./bitcoind -dns
302 2011-09-07 15:36:50 <ymirhotfoot> ps aux | grep itco
303 2011-09-07 15:36:59 <ymirhotfoot> I think that will work.
304 2011-09-07 15:42:36 <unclemanti> do what?
305 2011-09-07 15:43:29 <ymirhotfoot> type command at shell prompt; look for line with name of program in it
306 2011-09-07 15:43:50 <ymirhotfoot> if there is such a line the thing is running; might not be running right though
307 2011-09-07 15:52:27 <unclemanti> all i want to do is to ask the deamon waht the current block i
308 2011-09-07 15:52:28 <unclemanti> is
309 2011-09-07 16:18:17 <bx_> lookign for skilled web devs exp with js (ajax), php, mysql, etc PM me ... pay and equity in large bitcoin products along with Contract-by-contract for small biz clients. PM me ! :)
310 2011-09-07 16:29:06 <Joric> i made this using ajax! http://ragecoins.appspot.com
311 2011-09-07 16:36:43 <BlueMatt> ;;seen gavinandresen
312 2011-09-07 16:36:43 <gribble> gavinandresen was last seen in #bitcoin-dev 2 hours, 15 minutes, and 15 seconds ago: <gavinandresen> alex posted to the bitcoin-dev mailing list
313 2011-09-07 16:40:29 <tcatm> BlueMatt: http://eu1.bitcoincharts.com/blockchain/ is now signed with http://pgp.mit.edu:11371/pks/lookup?op=vindex&search=0x0120DFFD0E10F0EE
314 2011-09-07 16:41:10 <BlueMatt> tcatm: just saw that, Ill deal with links in a couple minutes...
315 2011-09-07 16:44:51 <cjdelisle> there's a lotta [connection reset by peer] in the #bitcoin* channels on lfnet
316 2011-09-07 16:51:15 <Eliel> Joric: an idea, you could let people play it for free too.
317 2011-09-07 16:53:03 <Matth1a3> anyone able to answer some of my questions about how to test https://github.com/bitcoin/bitcoin/pull/505 ?
318 2011-09-07 16:53:44 <Matth1a3> like, how many blocks should I generate (currently have about 20), and is there an easy way to check them against eachother in order to verify that the txids are unique?
319 2011-09-07 16:55:50 <Joric> how would you rate the damage caused by that mybitcoin affair
320 2011-09-07 16:58:20 <Eliel> Joric: well, the exchange rate plunged quite a bit so... severe?
321 2011-09-07 16:59:47 <SoftCoin> Mmmm, no new version of Bitcoin for over 2 months
322 2011-09-07 16:59:51 <SoftCoin> stagnant.
323 2011-09-07 17:00:11 <BlueMatt> clogged pipe with so many new features more like
324 2011-09-07 17:00:15 <BlueMatt> plus 0.4 is in rc now
325 2011-09-07 17:01:57 <SoftCoin> lol
326 2011-09-07 17:02:03 <SoftCoin> so many new features?
327 2011-09-07 17:03:06 <diki> you know
328 2011-09-07 17:03:14 <diki> i hope these features are not just under the hood
329 2011-09-07 17:03:24 <diki> i want to feel them
330 2011-09-07 17:04:54 <SoftCoin> hehe
331 2011-09-07 17:05:01 <SoftCoin> wallet stealer trojan released
332 2011-09-07 17:05:10 <SoftCoin> 6 months later, wallet encryption!
333 2011-09-07 17:05:23 <BlueMatt> diki: sadly...not really
334 2011-09-07 17:05:40 <SoftCoin> ___b___i__t____c____o____i_____n___@"
335 2011-09-07 17:05:54 <BlueMatt> SoftCoin: you want buggy releases of financial software?
336 2011-09-07 17:06:07 <SoftCoin> no
337 2011-09-07 17:06:25 <BlueMatt> also, gavin was gone for a while and stuff moved even slower, but now hes paid (as of like two weeks ago) and stuff is moving quick
338 2011-09-07 17:06:25 <SoftCoin> just any realeases
339 2011-09-07 17:06:49 <ThomasV_> BlueMatt: paid by whom ?
340 2011-09-07 17:06:53 <noagendamarket> wart
341 2011-09-07 17:06:55 <BlueMatt> trucoin
342 2011-09-07 17:06:57 <b4epoche> the man
343 2011-09-07 17:07:01 <BlueMatt> the cia
344 2011-09-07 17:07:08 <BlueMatt> oh wait, I already answered...damn
345 2011-09-07 17:07:11 <ThomasV_> huh?
346 2011-09-07 17:07:13 <noagendamarket> lol
347 2011-09-07 17:07:16 <b4epoche> the clan?
348 2011-09-07 17:07:26 <BlueMatt> http://www.trucoin.com/
349 2011-09-07 17:07:27 <b4epoche> oh, that's an 'i'
350 2011-09-07 17:07:35 <noagendamarket> he hets [paid cheese pizza ?
351 2011-09-07 17:08:14 <alexwaters> bluematt, any chance you have 2 minutes to help me with something?
352 2011-09-07 17:08:19 <BlueMatt> not sure how to interpret that, but Im assuming hes probably paid for his lunch breaks and thats about it...
353 2011-09-07 17:08:20 <b4epoche> does he have 'another' job?
354 2011-09-07 17:08:25 <BlueMatt> b4epoche: no
355 2011-09-07 17:08:36 <BlueMatt> never did afaik, though his wife works
356 2011-09-07 17:08:41 <BlueMatt> some uni something or other iirc
357 2011-09-07 17:08:49 <BlueMatt> alexwaters: uhh...sure
358 2011-09-07 17:09:13 <BlueMatt> trucoin also pays (paid?) stefan to work on bitcoinjs
359 2011-09-07 17:09:31 <ThomasV_> who owns trucoin?
360 2011-09-07 17:09:42 <BlueMatt> http://www.trucoin.com/?page_id=6
361 2011-09-07 17:10:00 <imsaguy2> I find it interesting the marketing guy doesnt have an email address.
362 2011-09-07 17:10:07 <BlueMatt> heh, funny
363 2011-09-07 17:10:18 <b4epoche> imsaguy:  and a PhD
364 2011-09-07 17:10:41 <imsaguy2> that just means he spent more money on school than some.
365 2011-09-07 17:11:06 <ThomasV_> hmm, how old is gavin?
366 2011-09-07 17:11:58 <BlueMatt> now this is getting creepy...
367 2011-09-07 17:12:08 <b4epoche> not sure what gets you a PhD in marketing&  mastering BS?
368 2011-09-07 17:12:40 <imsaguy2> who said the phd was in marketing?
369 2011-09-07 17:12:45 <BlueMatt> hehe, probably phyc major or something
370 2011-09-07 17:12:54 <BlueMatt> psyc*
371 2011-09-07 17:13:03 <imsaguy2> PSYCH!!!!!
372 2011-09-07 17:13:08 <BlueMatt> fu
373 2011-09-07 17:13:12 <imsaguy2> ah, the flashbacks
374 2011-09-07 17:13:15 <b4epoche> I never did&  just wondering
375 2011-09-07 17:13:37 <imsaguy2> BlueMatt, that wasn't a dig at you, just an old school reference
376 2011-09-07 17:13:42 <BlueMatt> oh, ok
377 2011-09-07 17:13:49 <imsaguy2> derp
378 2011-09-07 17:13:53 <BlueMatt> my uni uses psyc, not psych
379 2011-09-07 17:14:14 <b4epoche> why do you even know that?
380 2011-09-07 17:14:29 <imsaguy2> back in the day, they used to say it, but I have no clue as to the 'proper' spelling of it
381 2011-09-07 17:14:37 <BlueMatt> b4epoche: too much time looking at class lists
382 2011-09-07 17:14:56 <imsaguy2> http://www.urbandictionary.com/define.php?term=psych
383 2011-09-07 17:15:19 <b4epoche> imsaguy2:  you must be my age
384 2011-09-07 17:15:24 <imsaguy2> or not.
385 2011-09-07 17:16:13 <imsaguy2> historical knowledge doesn't necessarily prove age
386 2011-09-07 17:16:25 <mabus> sike!
387 2011-09-07 17:16:38 <mabus> [sic]
388 2011-09-07 17:17:01 <b4epoche> imsaguy2:  you used to say it, I can tell
389 2011-09-07 17:30:20 <Matth1a3> http://bitcoin.stackexchange.com/ is now public, FYI