1 2012-05-21 00:03:27 <jgarzik> interesting
  2 2012-05-21 00:03:37 <jgarzik> db_tuner util suggests page size 16384 for blkindex.dat
  3 2012-05-21 00:03:46 <jgarzik> (default is 4096 generally)
  4 2012-05-21 00:23:56 <BlueMatt> jgarzik: instant gratification can come in the form of an on-site balance, not just in your wallet, but...
  5 2012-05-21 00:27:14 <jgarzik> BlueMatt: but then it's not cool and bitcoiny and distributed and anonymous
  6 2012-05-21 00:27:15 <jgarzik> :)
  7 2012-05-21 00:27:41 <jgarzik> (mind you _I_ am not arguing for this, just saying, this must be what is going through peoples' heads)
  8 2012-05-21 00:38:11 <BlueMatt> jgarzik: I understand, doesnt mean I dont think they are stupid, but, then...most of getting people to gamble is illogical anyway, so I can see the argument
  9 2012-05-21 01:13:18 <SmithJohansen> Can I speak to a mod?
 10 2012-05-21 01:26:47 <jgarzik> ok, so I'm not crazy.  lotsa orphan tx's right now, one every few seconds
 11 2012-05-21 01:26:52 <jgarzik> mapsz 300
 12 2012-05-21 01:31:39 <imsaguy> satoshidice tx?
 13 2012-05-21 01:36:20 <jgarzik> imsaguy: orphan tx's are unspendable by anybody
 14 2012-05-21 01:36:28 <jgarzik> (without the dependent tx)
 15 2012-05-21 01:36:49 <imsaguy> any idea what's causing them?
 16 2012-05-21 01:40:17 <Diablo-D3> so its burning money?
 17 2012-05-21 01:42:03 <jgarzik> or might simply be never-valid trash cluttering up node memory
 18 2012-05-21 01:42:19 <Diablo-D3> so its not burning money?
 19 2012-05-21 01:42:55 <jgarzik> orphan tx's are unspendable... until they are spendable.  which might be never.
 20 2012-05-21 01:43:46 <Diablo-D3> so it is burning money?
 21 2012-05-21 01:49:48 <jgarzik> Diablo-D3: no
 22 2012-05-21 01:50:08 <Diablo-D3> k.
 23 2012-05-21 02:29:55 <luke-jr> fyi, make doesn't like -j0
 24 2012-05-21 02:29:56 <luke-jr> <.<
 25 2012-05-21 02:33:16 <Diablo-D3> luke-jr: erm, wouldnt that be infinite gccs?
 26 2012-05-21 02:33:29 <luke-jr> Diablo-D3: or zero.
 27 2012-05-21 04:50:13 <dub> luke-jr: FYI nobody likes j00
 28 2012-05-21 05:11:59 <SomeoneWeird> Anyone use bitcoinjs?
 29 2012-05-21 06:10:38 <gmaxwell> 20:26 <@jgarzik> ok, so I'm not crazy.  lotsa orphan tx's right now, one every few seconds
 30 2012-05-21 06:12:52 <SomeoneWeird> Anyone use bitcoinjs?
 31 2012-05-21 06:14:09 <gmaxwell> I saw that too My WAG earlier was that someone repeated-input-txned the dice site resulting in a bunch of invalid transactions.
 32 2012-05-21 06:45:31 <gmaxwell> almost all the orphans form a linear chain.
 33 2012-05-21 06:54:24 <gmaxwell> http://people.xiph.org/~greg/orphans_straight.png (big image warning)
 34 2012-05-21 06:56:44 <Diablo-D3> I think your image broke firefox
 35 2012-05-21 06:56:50 <Guest1603> mines fine :D
 36 2012-05-21 06:57:08 <Diablo-D3> oh wait
 37 2012-05-21 06:57:11 <Diablo-D3> I had to zoom in
 38 2012-05-21 06:57:14 <ageis> it's not like it's the Hubble Ultra-Deep Field
 39 2012-05-21 06:57:32 <Diablo-D3> gmaxwell: what are those orphans of?
 40 2012-05-21 06:57:36 <Diablo-D3> because those giant forks are kind of scary
 41 2012-05-21 06:58:36 <gmaxwell> Diablo-D3: orphan transactions, not block Diablo-D3
 42 2012-05-21 06:59:46 <Diablo-D3> ahh
 43 2012-05-21 06:59:50 <Diablo-D3> but...
 44 2012-05-21 06:59:52 <Diablo-D3> why is it doing that?
 45 2012-05-21 08:23:48 <GMP> 55 min without a block.... what is the probability of that to happen?
 46 2012-05-21 08:24:49 <GMP> lol!
 47 2012-05-21 08:25:01 <GMP> someone read it and posted a block
 48 2012-05-21 08:28:26 <SomeoneWeird> Anyone use bitcoinjs?
 49 2012-05-21 08:30:35 <gmaxwell> GMP: 100%
 50 2012-05-21 08:31:05 <GMP> lol great answer!
 51 2012-05-21 08:31:26 <gmaxwell> ;;bc,tblb 55m
 52 2012-05-21 08:31:28 <gribble> 1 day, 12 hours, 34 minutes, and 35 seconds
 53 2012-05-21 08:32:23 <gmaxwell> They hapeen once every day and a half on average.
 54 2012-05-21 08:32:28 <kinlo> GMP: it happens often that it takes a while to find a block...
 55 2012-05-21 09:19:02 <SomeoneWeird> Anyone use bitcoinjs?
 56 2012-05-21 09:19:04 <SomeoneWeird> >.<
 57 2012-05-21 09:20:12 <moa7> i wonder if the blocks take longer when there are more tainted coins running thru the system? ;)
 58 2012-05-21 09:50:52 <sipa> ;;bc,tblb
 59 2012-05-21 09:50:54 <gribble> Error: There's really no reason why you should have underscores or brackets in your mathematical expression.  Please remove them.
 60 2012-05-21 10:02:47 <SomeoneWeird> lol sipa
 61 2012-05-21 10:08:39 <coingenuity> anyone know whether -discover is a valid argument for bitcoin?
 62 2012-05-21 10:09:32 <coingenuity> it appears to be via net.cpp but the wiki does not mention it
 63 2012-05-21 10:09:48 <sipa> coingenuity: it's in master, but not in any released version
 64 2012-05-21 10:10:50 <coingenuity> ah, ok. if i compile from master, am i to assume it'll do what i've inferred from the .cpp and search for nodes to peer with?
 65 2012-05-21 10:11:01 <sipa> sure
 66 2012-05-21 10:11:11 <coingenuity> :) cool, i should give it a try
 67 2012-05-21 10:11:24 <sipa> -discover is default on; it means bitcoin should try to find its own public ip
 68 2012-05-21 10:11:25 <coingenuity> thanks sipa, mucho gracias
 69 2012-05-21 10:11:49 <sipa> you can use -nodiscover in combination with -externalip=... to explicitly specify the address(es) you're reachable at
 70 2012-05-21 10:12:21 <coingenuity> oh, interesting
 71 2012-05-21 10:12:31 <coingenuity> that's quite nifty
 72 2012-05-21 10:12:46 <gmaxwell> Which will helpfully avoid it connecting to the bitcoin node tracing service^W^W^W^W^Wwhatismyip.
 73 2012-05-21 10:13:14 <coingenuity> lol
 74 2012-05-21 10:13:51 <coingenuity> i'm fairly curious about how i would force a client to peer more heavily than the default, is my main goal
 75 2012-05-21 10:14:30 <gmaxwell> listen for inbound connections.
 76 2012-05-21 10:14:34 <coingenuity> i know i can use -maxconnections but i'm wondering how i would get my client to handshake with more peers more quickly
 77 2012-05-21 10:14:45 <gmaxwell> Otherwise? Sorry we don't have a -dosattacknetwork=1 mode. :)
 78 2012-05-21 10:15:22 <coingenuity> oh, i'm far from trying to dos ;)
 79 2012-05-21 10:16:05 <coingenuity> i'd basically like to have a web of trusted nodes that each can specify on -connect, so that transactions can be verified by multiple segments of the bitcoin network that are highly seeded
 80 2012-05-21 10:20:28 <gribble> New news from bitcoinrss: Diapolo opened pull request 1368 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1368>
 81 2012-05-21 10:22:37 <gmaxwell> each what can?
 82 2012-05-21 10:26:38 <coingenuity> each node
 83 2012-05-21 10:26:50 <coingenuity> so that node1 -connects to node2, etc
 84 2012-05-21 10:30:10 <gmaxwell> coingenuity: okay,  the connects (/addnodes) go up more or less right away.
 85 2012-05-21 11:16:00 <dusty_> /m gribble help
 86 2012-05-21 11:19:45 <dusty_> !faq
 87 2012-05-21 11:19:46 <gribble> See http://www.bitcoin.org/faq and http://www.bitcoin.org/wiki/doku.php?id=more_faqs
 88 2012-05-21 11:31:13 <dusty_> sipa: the second link of faqs given by gribble is not working
 89 2012-05-21 11:31:28 <dusty_> I don't know who manages the config of the bot
 90 2012-05-21 11:41:06 <BlueMatt> nanotube: runs gribble
 91 2012-05-21 11:41:08 <BlueMatt> ^
 92 2012-05-21 11:48:51 <dusty_> BlueMatt: thanks
 93 2012-05-21 13:03:13 <luke-jr> ThomasV: any reason electrum/stratum aren't CIA'd as modules under bitcoin? :p
 94 2012-05-21 13:03:42 <ThomasV> they are under #bitcoinconsultancy and #electrum
 95 2012-05-21 13:04:43 <ThomasV> genjix did it; I do not know how to set these things up
 96 2012-05-21 13:05:08 <luke-jr> i c
 97 2012-05-21 13:17:00 <t7> has anyone found a sha256 fixed point?
 98 2012-05-21 13:17:17 <drizztbsd> ?
 99 2012-05-21 13:18:46 <sipa> drizztbsd: a 256-bit value that hashes to itself
100 2012-05-21 13:19:00 <sipa> but i'm quite sure no such point is known
101 2012-05-21 13:20:35 <gribble> New news from bitcoinrss: sipa opened pull request 1369 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1369>
102 2012-05-21 13:22:12 <copumpkin> no such point is known
103 2012-05-21 13:22:24 <copumpkin> it's easy to compute the probability of one existing though
104 2012-05-21 13:22:53 <sipa> 1-(1-2^(-256))^(2^256)
105 2012-05-21 13:23:38 <Joric> t7, noone has found md5 fixed point even
106 2012-05-21 13:24:05 <Joric> not sure about its probability
107 2012-05-21 13:24:15 <sipa> which is 1-1/e, afaik
108 2012-05-21 13:24:15 <t7> i read 60% online
109 2012-05-21 13:24:25 <luke-jr> lol 60%
110 2012-05-21 13:24:27 <t7> which seems too high :|
111 2012-05-21 13:24:31 <sipa> 63%
112 2012-05-21 13:24:48 <t7> i has to be the same length
113 2012-05-21 13:25:00 <t7> oh wait im thinking with strings
114 2012-05-21 13:25:04 <t7> carry on
115 2012-05-21 13:25:47 <Joric> is there any malicious use for that? can't find one
116 2012-05-21 13:26:11 <t7> you know if the source string is x
117 2012-05-21 13:26:18 <t7> where x = f x
118 2012-05-21 13:26:21 <t7> :)
119 2012-05-21 13:26:40 <t7> so then i know the plain text
120 2012-05-21 13:26:50 <t7> as long as its x
121 2012-05-21 13:27:46 <t7> unless its something else that hashes to x
122 2012-05-21 13:29:10 <Joric> if only blocks were starting from zeros... infinite money!
123 2012-05-21 13:36:21 <diki> would anyone perhaps have a DAWG implementation in C?
124 2012-05-21 13:36:28 <sipa> yo dawg
125 2012-05-21 13:37:28 <Joric> must be a recursive language
126 2012-05-21 13:38:08 <Joric> oh it's a data structure
127 2012-05-21 13:38:14 <Joric> Directed Acrylic Word Graph
128 2012-05-21 13:38:58 <Joric> *Acyclic
129 2012-05-21 13:42:51 <luke-jr> [15:39:43] <ljrbot> Blk 00000000000000a2974f462cab5a64e816cb9bfdb2fe1f9f904fe0b006bb50ff: 18tyJrqKKcHdJ46th2icwoKZBZVp6qUsTx 47.90167571 BTC, 1MgH8iNkdZrDW8rsfzifDzsqJnCjBVyRBV 2.50787429 BTC, (None) 0 TBC
130 2012-05-21 13:42:53 <luke-jr> hmm
131 2012-05-21 13:42:57 <luke-jr> what's going on there?
132 2012-05-21 13:44:35 <diki> wow...you paid someone a whole block?
133 2012-05-21 13:44:47 <diki> ...unintentionally
134 2012-05-21 13:44:52 <diki> I wish that was me..
135 2012-05-21 13:45:24 <diki> wow, you paid him twice
136 2012-05-21 13:46:22 <luke-jr> looks like maybe a forked p2pool
137 2012-05-21 14:34:44 <Joric> https://bitcointalk.org/index.php?topic=80815.0 will anyone else do that or you're leaving it to me
138 2012-05-21 14:35:03 <Joric> a whole 25 btc for 3 months
139 2012-05-21 14:37:56 <Joric> i will be able to feed my family and Thomas will devote more time to his fantasy footbal team jk
140 2012-05-21 14:39:01 <Joric> love this video http://www.theonion.com/content/video/more_american_workers_outsourcing
141 2012-05-21 14:41:42 <gribble> New news from bitcoinrss: laanwj reopened pull request 1365 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1365>
142 2012-05-21 14:41:44 <bitcoinhacker> Joric: you want some free bitcoin
143 2012-05-21 14:42:04 <Joric> bitcoinhacker, sure, drop here 1JoricCBkW8C5m7QUZMwoRz9rBCM6ZSy96
144 2012-05-21 14:42:08 <bitcoinhacker> Joric: k
145 2012-05-21 14:42:11 <luke-jr> graingert, are you confessing?
146 2012-05-21 14:42:27 <bitcoinhacker> luke-jr: If I confess to you is it legally protected speach?
147 2012-05-21 14:42:48 <luke-jr> IANAL
148 2012-05-21 14:43:17 <bitcoinhacker> !google IANAL
149 2012-05-21 14:43:18 <gribble> IANAL - Wikipedia, the free encyclopedia: <http://en.wikipedia.org/wiki/IANAL>; Urban Dictionary: IANAL: <http://www.urbandictionary.com/define.php?term=IANAL>; IANAL: <http://www.catb.org/jargon/html/I/IANAL.html>
150 2012-05-21 14:43:28 <bitcoinhacker> luke-jr: I see
151 2012-05-21 14:44:04 <graingert_ecs> sigh
152 2012-05-21 14:44:13 <graingert_ecs> jgarzik: >.>
153 2012-05-21 14:45:11 <user__> bitcoinhacker: please some coins here: 1DApNttZXeSCyggDPuJeerC7VUf4W3PYt6
154 2012-05-21 14:45:17 <Joric> you may confess freely, luke-jr is a local priest
155 2012-05-21 14:45:23 <gavinandresen> Joric: sweet
156 2012-05-21 14:45:23 <luke-jr> Joric: don't lie
157 2012-05-21 14:45:53 <graingert_ecs> sigh I just picked up all of gavin's !later messages
158 2012-05-21 14:46:00 <sipa> haha
159 2012-05-21 14:46:08 <sipa> please don't do that anymore
160 2012-05-21 14:46:12 <graingert_ecs> it was mine warning him to set ENFORCE ON
161 2012-05-21 14:46:28 <jgarzik> graingert_ecs: purpose is irrelevant.  stop impersonating.
162 2012-05-21 14:46:31 <graingert_ecs> !later tell gavinandresen Please turn ENFORCE ON on your nick in NickServ
163 2012-05-21 14:46:31 <gribble> The operation succeeded.
164 2012-05-21 14:46:37 <jgarzik> we don't need that crap in here.
165 2012-05-21 14:46:48 <graingert_ecs> jgarzik: bitcoinhacker is one of my official nicks
166 2012-05-21 14:47:04 <sipa> ...
167 2012-05-21 14:47:15 <sipa> you really want a reputation, no?
168 2012-05-21 14:47:54 <graingert_ecs> sipa: sorry I can't see what you are saying can you turn off OTR?
169 2012-05-21 14:48:27 <sipa> ?
170 2012-05-21 14:48:44 <graingert_ecs> sipa: it's a complicated counter-troll
171 2012-05-21 14:49:12 <sipa> i'm not following, but i don't really care
172 2012-05-21 14:49:23 <graingert_ecs> sipa: that's fair
173 2012-05-21 14:49:44 <graingert_ecs> it was mainly for the people over my shoulder
174 2012-05-21 14:49:59 <luke-jr> this is #bitcoin-dev, not #nickfloodlulz
175 2012-05-21 14:50:24 <graingert_ecs> luke-jr: the fact that chaning nicks spams in all channels is anoying
176 2012-05-21 14:50:32 <sipa> then don't do it
177 2012-05-21 14:50:38 <luke-jr> graingert_ecs: it's a nick, not a message
178 2012-05-21 14:50:45 <graingert_ecs> changing*
179 2012-05-21 14:50:47 <luke-jr> you're not SUPPOSED to change nick generally
180 2012-05-21 14:51:03 <graingert_ecs> is it even ratelimited?
181 2012-05-21 14:51:38 <Joric> i don't care about impersonating but at least stop promising free coins i didn't get anything
182 2012-05-21 14:52:02 <luke-jr> lol
183 2012-05-21 14:52:03 <graingert_ecs> Joric: you really don't want coin off of BitcoinHacker
184 2012-05-21 14:52:25 <graingert_ecs> because 'gox will probably blacklist them because they are 'gox
185 2012-05-21 14:52:39 <Xeranos> tainted coins are tainted
186 2012-05-21 14:53:22 <graingert_ecs> well untill they are well shuffled
187 2012-05-21 14:53:26 <graingert_ecs> eg into the forceit
188 2012-05-21 14:54:36 <user__> grainget-ecs: i want 1DApNttZXeSCyggDPuJeerC7VUf4W3PYt6
189 2012-05-21 14:55:09 <Joric> see what you've done
190 2012-05-21 14:59:52 <copumpkin> http://dl.dropbox.com/u/361503/weird.png seems a bit contradictory
191 2012-05-21 14:59:57 <Diapolo> @Core Devs, is there an expected release date or any plan available for 0.7 RC-phase? I want to make sure we have a current translation ready by that time.
192 2012-05-21 15:04:07 <Diapolo> luke-jr: This was not meant to push for a RC, I just want to ensure translations get the love they need :).
193 2012-05-21 15:04:25 <gmaxwell> Diapolo: Not real soon quite yet.
194 2012-05-21 15:04:39 <Diapolo> then everythin is alright
195 2012-05-21 15:04:59 <luke-jr> yay, SIGHUP merged :p
196 2012-05-21 15:05:03 <Diapolo> I want to suggest, that there is a relase shedule in the Wiki or somewhere else, if that's not the case.
197 2012-05-21 15:05:28 <luke-jr> Diapolo: IMO, it's not a time-based thing
198 2012-05-21 15:07:46 <jgarzik> hum
199 2012-05-21 15:08:07 <jgarzik> MAX_ORPHAN_TRANSACTIONS would seem to conflict with a reorg that involves a large number of transactions
200 2012-05-21 15:08:16 <Diapolo> But it should be planned somehow, to have deadlines ensures people work on parts that are important ;).
201 2012-05-21 15:08:40 <jgarzik> we start evicting random orphan tx's at 10,000
202 2012-05-21 15:08:47 <luke-jr> Diapolo: the problem is, merges are held up mostly by reviewers and pullers, not the developers and users who want it ;)
203 2012-05-21 15:09:54 <luke-jr> so IMO it should be more of "when the features ready at the opening of the window are all pulled"
204 2012-05-21 15:12:02 <luke-jr> in any case, "when it's ready" tends to work for most open source projects
205 2012-05-21 15:12:10 <luke-jr> don't see a need for something more formal
206 2012-05-21 15:12:59 <Diapolo> luke-jr: Let me re-phrase what I said, translations need time ... create a new bitcoinstrings.cpp, create a new english-master file, let translators on Transifex do their jobs and then fetch from Tx and create a pull with updated translations. I would say this needs at least 7, better 14 days to mature. With pre-0.7 releases there simply was no real translation-planning or time to do this, as I always got surprised with a RC-rela
207 2012-05-21 15:13:39 <gmaxwell> Diapolo: we should call that the RC window I think it's okay to update translations during RC... and by that time we won't be doing big code changes that disrupt translations.
208 2012-05-21 15:13:57 <Diapolo> gmaxwell: agreed
209 2012-05-21 15:14:19 <luke-jr> I suppose we could call the first one an "alpha" release since it's not really a candidate if we're intending to update translations anyway
210 2012-05-21 15:14:44 <gmaxwell> meh. that would really hurt testing.
211 2012-05-21 15:14:52 <gmaxwell> the strings are alpha, not the code.
212 2012-05-21 15:15:01 <luke-jr> for 0.7, it seems there is some bugs in the new stuff already merged that need cleanup, so it might make sense to declare feature-freeze before RCs get made
213 2012-05-21 15:15:36 <luke-jr> (thinking of the verify message thing mainly - but if coincontrol is merged before it's cleaned up, that too)
214 2012-05-21 15:16:03 <gmaxwell> also we still have sipa's redundant block puller than need to be fixed or come out.
215 2012-05-21 15:16:21 <luke-jr> gmaxwell: I thought it was fixed?
216 2012-05-21 15:16:48 <gmaxwell> Well it works now but it's over eager and pulls when it doesn't need to sometimes.
217 2012-05-21 15:16:50 <sipa> luke-jr: it's been improved already, but there are still redundant getblocks (no redundant getdata's though)
218 2012-05-21 15:17:06 <luke-jr> oh
219 2012-05-21 15:17:31 <gmaxwell> no biggie, I only brought it up because you were listing things that were must fix before release.
220 2012-05-21 15:17:38 <luke-jr> anyhow, we probably have a week worth of fix/cleanup we can do between freeze and rc1
221 2012-05-21 15:18:16 <gmaxwell> luke-jr: did you find it odd that coincontrol can't select single inputs?
222 2012-05-21 15:18:49 <luke-jr> gmaxwell: I don't particularly care for coincontrol as-is, but I don't know if showing individual inputs would make me care any more than I do now.
223 2012-05-21 15:19:25 <gmaxwell> Yea, it kinda sucks. :-/
224 2012-05-21 15:19:42 <luke-jr> in fact, multiple wallets seems to make current coincontrol worthless
225 2012-05-21 15:20:09 <luke-jr> BUT, dooglus's refactoring code probably makes sense either way
226 2012-05-21 15:21:20 <jgarzik> that's sorta my feeling -- multiple wallets preferred over coin control
227 2012-05-21 15:21:38 <jgarzik> easier for the user to grasp and manage
228 2012-05-21 15:27:24 <luke-jr> as long as we're talking about easy to grasp/manage, I have long felt it'd be useful to have a nice GUI with visual coins (amount-sized circles, representing one output each) that can be shown in "account" or address groups and manually hand-chosen for "melting" into a new transaction ;)
229 2012-05-21 15:30:23 <sipa> proposal for command-line argument interaction, when tor hidden service support is added: https://gist.github.com/2763381
230 2012-05-21 15:32:22 <luke-jr> sipa: it's unclear if -bind 127.0.0.1 causes it to bind :: and 0.0.0.0 or not
231 2012-05-21 15:32:57 <luke-jr> sipa: how does -upnp's default relate to USE_UPNP=0/1 ?
232 2012-05-21 15:33:11 <sipa> right, let me update
233 2012-05-21 15:34:04 <luke-jr> what does -proxydns do if you turn it on without -proxy||-onion? :p
234 2012-05-21 15:35:38 <luke-jr> sipa: shouldn't -proxy -externalip result in binding only localhost by default?
235 2012-05-21 15:37:01 <sipa> hmm, that becomes complex...
236 2012-05-21 15:37:46 <luke-jr> otoh, someone might have a normal proxy with port forwarding&
237 2012-05-21 15:38:01 <Diapolo> sipa: I remember I had some findings related to IPv6 that you intended to change, too ... can you remember?
238 2012-05-21 15:38:16 <sipa> Diapolo: sorry, no
239 2012-05-21 15:38:42 <luke-jr> sipa: is it possible to bind sockets such that they only get connections FROM a specific IP? <.<
240 2012-05-21 15:38:51 <sipa> luke-jr: don't think so
241 2012-05-21 15:39:45 <Diapolo> sipa: I ething related to -discover=0 but can't remember too ... and there are 3 compiler-warnings left in the IPv6 code
242 2012-05-21 15:39:45 <luke-jr> possible to detect if proxy server is Tor?
243 2012-05-21 15:39:56 <Diapolo> "I think something" was the first part ^^
244 2012-05-21 15:40:06 <sipa> luke-jr: not afaik
245 2012-05-21 15:40:21 <sipa> luke-jr: and judging based on whether it's port 9050 is sloppy
246 2012-05-21 15:40:26 <luke-jr> sipa: possible to specify -bind=lo ? :p
247 2012-05-21 15:40:37 <sipa> luke-jr: no
248 2012-05-21 15:40:41 <luke-jr> meh
249 2012-05-21 15:40:53 <sipa> maybe later
250 2012-05-21 15:41:25 <luke-jr> sounds like the best solution is to default to loopback binding if -proxy is specified, and treat -bind=1 as inaddr_any
251 2012-05-21 15:42:07 <sipa> i don't like listening by default in any case
252 2012-05-21 15:42:35 <sipa> well, any potential attempt at privacy
253 2012-05-21 15:43:34 <imsaguy2> While having the text in the status bar showing how many blocks remain until a client is 'caught up' is a good thing, its almost impossible to read.
254 2012-05-21 15:44:28 <imsaguy2> it needs a color shade behind it or something
255 2012-05-21 15:44:34 <luke-jr> sipa: I mean -proxy -externip
256 2012-05-21 15:44:51 <luke-jr> imsaguy2: that's an OS thing
257 2012-05-21 15:45:01 <Diapolo> imsaguy2: We use OS-default style for the progressbar, no chance to fix it for every possible theme setting.
258 2012-05-21 15:45:32 <imsaguy2> you can't put the text in a background color'd box?
259 2012-05-21 15:46:02 <imsaguy2> you might as well leave it out then, because it really does look like crap on a default windows install
260 2012-05-21 15:46:07 <wumpus> Diapolo: agreed, it's simply the OS theme, if it's not right for your theme fix your theme
261 2012-05-21 15:46:15 <imsaguy2> thats stupid
262 2012-05-21 15:46:18 <imsaguy2> I use the default theme
263 2012-05-21 15:46:27 <sipa> wumpus: i don't think i have any theme to set, or i wouldn't know where
264 2012-05-21 15:46:29 <imsaguy2> you're telling me I need to restyle my whole OS because of a single program? lol
265 2012-05-21 15:46:29 <luke-jr> imsaguy2: there is no default theme
266 2012-05-21 15:46:37 <imsaguy2> luke-jr: wrong.
267 2012-05-21 15:46:42 <wumpus> no, that's not stupid... embedding styles is stupid
268 2012-05-21 15:47:05 <sipa> wumpus: for normal things, but this combination bites so many people with typical themes already
269 2012-05-21 15:47:07 <wumpus> other qt programs will have the same issue if they use a progress bar with text on it
270 2012-05-21 15:47:15 <imsaguy2> that doesn't make it right wumpus
271 2012-05-21 15:47:16 <imsaguy2> derp
272 2012-05-21 15:47:23 <wumpus> file an issue with qt
273 2012-05-21 15:47:38 <sipa> i agree with the basic notion that you should use default styles as much as possible
274 2012-05-21 15:47:45 <sipa> but it's just unreasonable here
275 2012-05-21 15:47:50 <luke-jr> wumpus: we're not using Qt's "text on progressbar" feature though, are we?
276 2012-05-21 15:48:00 <wumpus> luke-jr: afaik we are
277 2012-05-21 15:48:16 <luke-jr> wumpus: how? QProgressBar by default won't let you override the %&
278 2012-05-21 15:48:19 <sipa> exactly; i don't think there is any guarantee or intention to make test-on-progressbar look good, judging from what i've seen
279 2012-05-21 15:48:43 <wumpus> if you change the progressbar then people will complain it looks ugly on their os... people will always complain
280 2012-05-21 15:48:44 <imsaguy2> so then put the remaining number above/next to the progressbar itself
281 2012-05-21 15:48:53 <sipa> furthermore, i wouldn't have any clue where to change a style
282 2012-05-21 15:49:04 <Diapolo> I took many approached with that bar and I remember we had I style I think looked good, but asap we got a complaint it would not match the themes colors...
283 2012-05-21 15:49:13 <wumpus> yeah I remember that too
284 2012-05-21 15:49:15 <luke-jr> outline on the text a good compromise?
285 2012-05-21 15:49:18 <wumpus> we can't keep fipflopping
286 2012-05-21 15:49:28 <imsaguy2> just put the text in a box
287 2012-05-21 15:49:35 <sipa> just fix it... that way it's already readable everywhere
288 2012-05-21 15:49:48 <Diapolo> imsaguy2: if that would be poossible it's an option
289 2012-05-21 15:49:52 <Diapolo> -o
290 2012-05-21 15:50:17 <luke-jr> wumpus: for testing changes, how can I force the progressbar to be visible?
291 2012-05-21 15:50:26 <Diapolo> sipa: how could we cover all possibilities a Got uses perhaps a black style we use black text ... next complaint ;).
292 2012-05-21 15:50:35 <Diapolo> goth ^^
293 2012-05-21 15:50:36 <wumpus> luke-jr: start bitcoin in a new datadir
294 2012-05-21 15:50:40 <luke-jr> >_<
295 2012-05-21 15:50:47 <imsaguy2> Diapolo: my use case isn't an outliar
296 2012-05-21 15:50:54 <sipa> Diapolo: therefore, fix the style - don't use a default
297 2012-05-21 15:50:59 <imsaguy2> its the basic windows theme
298 2012-05-21 15:51:08 <Diapolo> What OS?
299 2012-05-21 15:51:15 <sipa> Diapolo: that way some may find it ugly, but at least it will be readable
300 2012-05-21 15:51:39 <wumpus> progressBar->setFormat(tr("~%n block(s) remaining", "", nRemainingBlocks));   that's as close to qt default code as it gets
301 2012-05-21 15:51:43 <sipa> i figured my case is an outlier, but apparently i'm not the only one
302 2012-05-21 15:52:10 <wumpus> if it's not visible on some default themes, I suggest to file an upstream issue with qt
303 2012-05-21 15:52:36 <imsaguy2> lol
304 2012-05-21 15:52:48 <sipa> how can qt do anything about that... it uses the OS's default progressbar, and puts a default text on top of it, i guess
305 2012-05-21 15:52:57 <Diapolo> not lol if it would be easy to fix it would already be merged :D
306 2012-05-21 15:53:09 <wumpus> that's the same as asking what we can do about it... we simply use qt and trust it
307 2012-05-21 15:53:11 <luke-jr> we're not using progress bar text right now
308 2012-05-21 15:53:12 <imsaguy2> so then perhaps text on the progress bar itself isn't the righ thting
309 2012-05-21 15:53:15 <luke-jr> going to prototype a fix
310 2012-05-21 15:53:17 <imsaguy2> right*
311 2012-05-21 15:53:25 <sipa> i just don't understand why you give priority to "looking nice" while it's just unreadable to others
312 2012-05-21 15:53:27 <imsaguy2> even windows doesn't do that
313 2012-05-21 15:54:03 <wumpus> the percentage on the progress bar was also text
314 2012-05-21 15:54:19 <imsaguy2> back when the progress bars were full color, not blue boxes
315 2012-05-21 15:54:21 <gmaxwell> 10:19 < luke-jr> in fact, multiple wallets seems to make current coincontrol worthless
316 2012-05-21 15:54:25 <luke-jr> err
317 2012-05-21 15:54:38 <gmaxwell> Not at all because even if you have multiple wallets, your wallts should have multiple addresses!
318 2012-05-21 15:54:38 <wumpus> we simply use "%n blocks left" instead of "xxx%" as setFormat now
319 2012-05-21 15:55:08 <imsaguy2> when the progress bars were/are blue boxes, windows never puts text on it.  Microsoft only put the text on the progress bars when they were a solid color that filled in.
320 2012-05-21 15:55:15 <gmaxwell> jgarzik: the coincontrol patch also contains a collection of good unit tests for coinselection.
321 2012-05-21 15:55:20 <imsaguy2> There's a reason for that.
322 2012-05-21 15:55:26 <imsaguy2> text on the boxes isn't readible
323 2012-05-21 15:56:13 <imsaguy2> One could ask why you even need a progress bar
324 2012-05-21 15:56:23 <imsaguy2> I'm 2k blocks behind and the only thing missing is the last block
325 2012-05-21 15:56:24 <sipa> i prefer multiwallet as wallet management mechanism over coin control, but that doesn't make the latter useless
326 2012-05-21 15:56:27 <imsaguy2> a lot of good that does me
327 2012-05-21 15:56:37 <Diapolo> because if there is none people complain they don't see how fast block-sync is?
328 2012-05-21 15:56:38 <jgarzik> FWIW, speaking generally, there are very few things of consequence you will see my NAK on
329 2012-05-21 15:56:43 <jgarzik> bitching on IRC != NAK :)
330 2012-05-21 15:57:13 <jgarzik> i.e. you don't see a NAK on CBlockStore
331 2012-05-21 15:57:39 <imsaguy2> you can see how fast block sync is by the numbers counting down
332 2012-05-21 15:57:52 <imsaguy2> only right now, I can't see the numbers because the blocks that aren't changing make it impossible
333 2012-05-21 15:58:18 <luke-jr> wumpus: yeah, this looks like a Qt bug
334 2012-05-21 15:58:31 <wumpus> luke-jr: thanks for checking
335 2012-05-21 15:58:48 <imsaguy2> me as an end user complaining to QT will get referred to the original project
336 2012-05-21 15:59:05 <luke-jr> gmaxwell: choosing individual addresses is IMO not useful; individual coins might be, but the next useful level up is logical wallets
337 2012-05-21 15:59:20 <luke-jr> gmaxwell: coincontrol just tries to shove multiple logical wallets inside one real wallet
338 2012-05-21 15:59:29 <sipa> luke-jr, gmaxwell: update: https://gist.github.com/2763381
339 2012-05-21 15:59:47 <imsaguy2> but anyway, I'm done.
340 2012-05-21 16:00:53 <gmaxwell> luke-jr: Well, it _is_ useful, both for some real cases, and as an educational tool. But it's not a replacement for muliple wallets, not even a partial one and it's weak as an educational tool when you can't also select inputs.
341 2012-05-21 16:00:58 <wumpus> imsaguy2: you can still file a github issue with us, put some screenshots, we can file the qt issue with more details... I just don't agree with the solution of setting a fixed style on every os
342 2012-05-21 16:02:01 <gmaxwell> luke-jr: e.g. it allowed a lot of people who took the theieves coins to send exactly those coins on to your collection address... and let them do so without the forethought of establishing another wallet for that purpose.
343 2012-05-21 16:04:11 <luke-jr> gmaxwell: only if they made a new address
344 2012-05-21 16:04:43 <gmaxwell> luke-jr: indeed, there were some who didn't, and because it couldn't select inputs no joy.
345 2012-05-21 16:20:59 <sipa> luke-jr: updated again
346 2012-05-21 16:27:20 <jgarzik> kyotocabinet's HashDB sure is nice
347 2012-05-21 16:30:21 <jgarzik> several interesting in-memory data structures, including one that claims to be better than std::map
348 2012-05-21 16:34:11 <wumpus> imsaguy2: can you put up a screenshot how it looks for you?
349 2012-05-21 16:35:36 <sipa> this is what it looks like for me: http://bitcoin.sipa.be/bitcoinqt.png
350 2012-05-21 16:37:59 <wumpus> sipa: hm looks like the base windows style http://doc-snapshot.qt-project.org/4.8/gallery-windows.html
351 2012-05-21 16:38:41 <sipa> wumpus: it's XMonad on Ubuntu 11.10
352 2012-05-21 16:39:17 <wumpus> most windows users get the windowsvista or windowsxp style
353 2012-05-21 16:39:32 <wumpus> on linux it's usually plastique or gtk
354 2012-05-21 16:40:04 <sipa> where would one configure that?
355 2012-05-21 16:40:39 <wumpus> try passing -style plastique  (that should work, if it passes the arguments to qt as it's supposed to)
356 2012-05-21 16:41:15 <sipa> wow, you can do that? :D
357 2012-05-21 16:41:16 <gmaxwell> thats what it looks like for me too, though my colors are inverted.
358 2012-05-21 16:41:49 <gmaxwell> man, someday someone will shop sipa -geometry and it will blow his mind.
359 2012-05-21 16:42:00 <wumpus> typically one of "windows", "motif", "cde", "plastique", "windowsxp", or "macintosh" (they probably don't all work, depends on which ones are installed)
360 2012-05-21 16:42:06 <wumpus> hehe
361 2012-05-21 16:42:24 <gmaxwell> s/shop/show/
362 2012-05-21 16:42:40 <sipa> "Wow, you mean I can *change* which screen things show up on, using DISPLAY=...?" ;)
363 2012-05-21 16:43:04 <jgarzik> interesting pattern: "DB.transaction(self, proc, hard=False)"    Perform entire transaction by a functor ("proc").
364 2012-05-21 16:43:51 <sipa> jgarzik: you mean, instead of starting a transaction in acceptblock, and ending it in any of 3 different places?
365 2012-05-21 16:44:03 <sipa> ;)
366 2012-05-21 16:44:09 <jgarzik> sipa: something like that ;)
367 2012-05-21 16:45:35 <jgarzik> That pattern is from kyotocabinet's python interface (though it's C++ and other APIs support it also).  bitcoin's equivalent would be something like:  pdb->transaction(WriteBlockIndex) or pdb->transaction(AcceptBlock) etc.  The database calls your app's functor callback, wrapped inside a transaction
368 2012-05-21 16:45:44 <jgarzik> if the functor returns an error, the transaction is aborted
369 2012-05-21 16:45:49 <jgarzik> otherwise its committed
370 2012-05-21 16:46:31 <jgarzik> would be a pain to apply to bitcoin, so it's largely of academic interest
371 2012-05-21 16:46:53 <jgarzik> kyotocabinet also uses the visitor pattern that is starting to appear
372 2012-05-21 16:47:11 <sipa> jgarzik: that is not *that* different from having a RAII CTxDB we already have, though it's used horribly
373 2012-05-21 16:47:37 <jgarzik> true
374 2012-05-21 16:48:10 <jgarzik> though I really dislike RAII's side effects in bitcoin...  buried DEEP inside a call chain, you are opening an entire db environment -and- a database
375 2012-05-21 16:49:05 <jgarzik> and we do strange things like instantiate a CTxDB _inside_ a CTxDB method
376 2012-05-21 16:49:17 <wumpus> hehe
377 2012-05-21 16:50:33 <sipa> any comments on #1357, by the way?
378 2012-05-21 16:54:17 <jgarzik> <shrug> seems like an unobjectionable minor cleanup
379 2012-05-21 16:58:12 <jgarzik> It is disappointing that bsddb.db shipped with python 2.7 in Fedora 16 is locked to db 4.8
380 2012-05-21 16:58:44 <wumpus> hmm I think boost::apply_visitor(ValidateAddressVisitor(ret), dest);  is a bit confusing... you pass in a hash to return the return value in. Don't  boost visitors have proper return values?
381 2012-05-21 17:00:01 <sipa> wumpus: what would you make it return?
382 2012-05-21 17:00:15 <sipa> oh, that one
383 2012-05-21 17:00:24 <wumpus> well it's a validation visitor, so it should return a value whether it is validated or not?
384 2012-05-21 17:00:37 <sipa> i guess it could return a JSON object
385 2012-05-21 17:00:48 <sipa> but there are common fields that need to be added to it as well
386 2012-05-21 17:01:39 <wumpus> sounds better
387 2012-05-21 17:02:12 <wumpus> and maybe rename it to describeaddressvisitor, seems more like what it does :)
388 2012-05-21 17:02:29 <sipa> if i could, i'd change the rpc name as well
389 2012-05-21 17:02:58 <wumpus> hm yeah
390 2012-05-21 17:03:19 <sipa> and i still need a way to return failure as well
391 2012-05-21 17:03:27 <sipa> though that's not really used now
392 2012-05-21 17:04:20 <wumpus> maybe return a pair <bool, Object>
393 2012-05-21 17:04:32 <wumpus> or use an exception
394 2012-05-21 17:04:49 <sipa> bah :)
395 2012-05-21 17:05:03 <wumpus> first is better in this case, as it's not really exceptional
396 2012-05-21 17:06:04 <sipa> it shouldn't ever happen
397 2012-05-21 17:06:17 <wumpus> hehe, famous last words
398 2012-05-21 17:06:35 <sipa> CBitcoinAddress::IsValid() being true means no CNoDestination
399 2012-05-21 17:06:59 <sipa> unless there is a "merge" operation for JSON objects, i'd prefer to keep it this way
400 2012-05-21 17:07:06 <waldston> anyone knows a safe vps service for hosting bitcoind + website ?
401 2012-05-21 17:07:42 <jgarzik> waldston: depends on the site value.  if you want safe... control the hardware or face another hack like LInode or RackSpace
402 2012-05-21 17:08:02 <jgarzik> waldston: bitcoin hackers have already targeted large cloud providers' management systems
403 2012-05-21 17:08:10 <waldston> jgarzik: rackspace was hacked too ???
404 2012-05-21 17:08:13 <jgarzik> so I doubt a tiny VPS provider will be better secured
405 2012-05-21 17:08:56 <waldston> well, one could use my service like a wallet ... so I could not face a hacking event
406 2012-05-21 17:09:21 <waldston> * I can not
407 2012-05-21 17:09:34 <luke-jr> waldston: Bitcoinica was Rackspace
408 2012-05-21 17:09:34 <sipa> wumpus: the purpose of that visitor is to add whatever information there is about that destination to a given json object
409 2012-05-21 17:10:06 <waldston> luke-jr: but in the second hack, they said in their release that it was not rackspace's fault
410 2012-05-21 17:10:08 <sipa> wumpus: i could make it return that information in an object, but it would need to be merged with whatever we already have
411 2012-05-21 17:10:30 <waldston> not like in the first hack, with was entirely linode's fault
412 2012-05-21 17:10:34 <wumpus> sipa: I know, but that's not clear from that line
413 2012-05-21 17:10:36 <wumpus> sipa: hm let's see
414 2012-05-21 17:10:53 <luke-jr> waldston: you mean third?
415 2012-05-21 17:11:23 <waldston> luke-jr: wow, there was one before linode ? didn't knew
416 2012-05-21 17:11:55 <gavinandresen> waldston: if you are holding other people's bitcoins then my opinion is you insure accounts for losses up to X bitcoin. And when I say insure I really mean insure-- buy insurance, or if you can't buy insurance put aside X*nAccounts bitcoins to self-insure against hacking.
417 2012-05-21 17:11:57 <waldston> so, my service is ready for launch ... I just need a host
418 2012-05-21 17:12:34 <waldston> gavinandresen: I think no insurance company here in brazil will understand what a bitcoin is
419 2012-05-21 17:12:50 <gavinandresen> ok, so self-insure with an auditable off-line cold-storage wallet.
420 2012-05-21 17:13:55 <waldston> gavinandresen: I'd take some months to put that working ... if I could just pay a trustable vps that would be much better
421 2012-05-21 17:14:02 <c_k> waldston: use Amazon EC2
422 2012-05-21 17:14:15 <ThomasV> use Linode!
423 2012-05-21 17:14:16 <sipa> waldston: will the vps be able to send money?
424 2012-05-21 17:14:19 <sipa> or only receive?
425 2012-05-21 17:14:35 <waldston> ThomasV: linode ?? gotta be kidding me
426 2012-05-21 17:14:40 <ThomasV> yeah :)
427 2012-05-21 17:14:43 <waldston> c_k:  can one run bitcoind there ?
428 2012-05-21 17:14:48 <waldston> sipa: yeah
429 2012-05-21 17:15:01 <diki> c_k:Clark Kent I presume?
430 2012-05-21 17:15:11 <gavinandresen> waldston: how would sending 5,000 BTC to offline cold-storage and publishing a signed message proving you have the key to those funds take months to setup?
431 2012-05-21 17:15:26 <BlueMatt> waldston: there is no such thing as a trustable impenetrable vps
432 2012-05-21 17:15:27 <gavinandresen> (5,000 BTC being insuring 1,000 accounts for 5 BTC each, just an example)
433 2012-05-21 17:15:47 <wumpus> sipa: from what I see (json spirit is pretty hard to read), Object is simply a typedef of map, so should be possible to merge them simply with insert(obj.begin(), obj.end());
434 2012-05-21 17:15:51 <BlueMatt> any online machine is not credible as an insurance-wallet-holding machine
435 2012-05-21 17:16:45 <ThomasV> captain kent!
436 2012-05-21 17:17:20 <BlueMatt> Id think there is
437 2012-05-21 17:17:42 <luke-jr> whoever does it would need to at least hire a security expert to audit the systems being insured
438 2012-05-21 17:17:46 <ThomasV> hum, would 17-years old people have to  pay more?
439 2012-05-21 17:17:48 <luke-jr> to check that they're reasonably secure
440 2012-05-21 17:17:59 <BlueMatt> luke-jr: ofc
441 2012-05-21 17:18:08 <ThomasV> (like for car driving)
442 2012-05-21 17:18:13 <luke-jr> ThomasV: I see no reason for an age-bias, so long as you're insuring the service and not the users directly.
443 2012-05-21 17:18:23 <ThomasV> lol
444 2012-05-21 17:18:35 <luke-jr> (if you're insuring the users directly, then you need to consider the un-prosecutability of the legal minor running off with the money)
445 2012-05-21 17:19:06 <luke-jr> ThomasV: of course, someone could make a rule that anyone with access to the wallet needs to pass a security test
446 2012-05-21 17:19:21 <ThomasV> and a honesty test :)
447 2012-05-21 17:19:31 <BlueMatt> luke-jr: I would expect such a rule
448 2012-05-21 17:19:34 <luke-jr> a 17 year old is no likely to be less honest than a 45 year old
449 2012-05-21 17:20:10 <diki> luke-jr:are you referring to zhoutong?
450 2012-05-21 17:20:17 <ThomasV> I have discovered the true nature of pirateat40's business: bitcoinica insurer
451 2012-05-21 17:20:20 <luke-jr> diki: I'm not referring to anyone specifically.
452 2012-05-21 17:20:33 <gmaxwell> people still have diki not ignored?
453 2012-05-21 17:20:55 <ThomasV> gmaxwell: me
454 2012-05-21 17:21:09 <gmaxwell> weird.
455 2012-05-21 17:21:16 <diki> gmaxwell:jelly? If so, go away please.
456 2012-05-21 17:21:32 <ThomasV> I just ignore the /ignore command
457 2012-05-21 17:22:52 <sipa> wumpus: ret.insert(detail.begin(), detail.end());   ->   one screen long error message
458 2012-05-21 17:23:09 <luke-jr> gmaxwell: I generally ignore diki the old fashoned way, but this time he asked a slightly reasonable question
459 2012-05-21 17:23:10 <sipa> i never managed to decode json-spirits's inner workings
460 2012-05-21 17:25:17 <user__> waldston: are you from brazil? what service you will launch
461 2012-05-21 17:25:36 <wladston> user__: it's a paid qa site
462 2012-05-21 17:26:06 <wladston> user__: yup, from brazil. but will launch in english.
463 2012-05-21 17:26:07 <user__> wladston: o que ?? qa?
464 2012-05-21 17:26:38 <sipa> wumpus: nevermind, fixed
465 2012-05-21 17:28:20 <sipa> wumpus: pushed
466 2012-05-21 18:14:00 <wladston> thanks for all the help guys
467 2012-05-21 18:14:15 <gavinandresen> sipa jgarzik: valgrind and my mac are throwing fits with the commit that added #define BOOST_SPIRIT_THREADSAFE...
468 2012-05-21 18:14:32 <wladston> let's hope I can launch my service without hacking events
469 2012-05-21 18:14:39 <wladston> see ya!
470 2012-05-21 18:14:52 <gavinandresen> "let's hope" <-- that doesn't bode well....
471 2012-05-21 18:16:07 <sipa> hope... the quintessential of all human delusions
472 2012-05-21 18:17:20 <luke-jr> cuz you're certainly going to see them at some point if you're successful& better to have problems exposed earlier
473 2012-05-21 18:18:31 <sipa> gavinandresen: running valgrind
474 2012-05-21 18:18:45 <gavinandresen> sipa: valgrind against test_bitcoin
475 2012-05-21 18:19:03 <gavinandresen> ... that's where I'm running into the issue.  Invalid write of size for in _pthread_mutex_init
476 2012-05-21 18:19:08 <luke-jr> O.o
477 2012-05-21 18:19:23 <gavinandresen> ... called from boost::mutex::mutex
478 2012-05-21 18:19:30 <gavinandresen> (called from boost::spirit)
479 2012-05-21 18:33:26 <gavinandresen> frickin fracking cross-platform differences... valgrind and the unit tests are perfectly happy in my ubuntu VM, fail miserably on my Mac.
480 2012-05-21 18:34:44 <luke-jr> :/
481 2012-05-21 18:34:52 <luke-jr> I got a ton of OpenSSL things on Gentoo
482 2012-05-21 18:34:57 <luke-jr> but not pthread-related
483 2012-05-21 18:36:35 <luke-jr> gavinandresen: btw, what do Mac users do for bitcoind?
484 2012-05-21 18:36:53 <gavinandresen> luke-jr: compile it themselves if they want it.
485 2012-05-21 18:36:55 <luke-jr> i c
486 2012-05-21 18:47:34 <Diapolo> gavinandresen: What about the added startuptime thing in the RPC call getinfo? Could luke and I convince you?
487 2012-05-21 18:47:44 <gavinandresen> no'
488 2012-05-21 18:48:36 <gavinandresen> ... and if you start a long discussion on how it is a great little feature that I will never regret adding I'll just put you on ignore
489 2012-05-21 18:49:05 <sipa> Diapolo: i use -logtimestamps if i need to know the startuptime (and that is only for benchmarks, hardly worth an extra info field)
490 2012-05-21 18:49:55 <gavinandresen> first line of debug.log includes the startup time, even without -logtimestamps, I believe
491 2012-05-21 18:50:13 <sipa> orly?
492 2012-05-21 18:50:26 <gavinandresen> I think so, I get Bitcoin version v0.6.2.2-beta (2012-05-11 10:09:45 -0400)
493 2012-05-21 18:51:01 <Diapolo> should be the compile time ...
494 2012-05-21 18:51:17 <sipa> that's time of the last commit
495 2012-05-21 18:51:54 <luke-jr> maybe it should. sounds like a reasonable compromise.
496 2012-05-21 18:52:45 <gavinandresen> ah, right, I'm constantly compiling then running so it always looked like startup time to me....
497 2012-05-21 18:52:58 <gmaxwell> Anyone looked into these big orphan floods anymore?
498 2012-05-21 18:53:29 <sipa> gmaxwell: hmm. no
499 2012-05-21 18:53:31 <gmaxwell> it's really hard to figure out anything about them... and gettransacion doesn't return anything on stored orphans. :-/
500 2012-05-21 18:53:53 <sipa> shouldn't be hard to add
501 2012-05-21 18:54:09 <gmaxwell> sipa: see the crazy pngs I posted early this morning... there are enormous chains of orphanes being announced for the last day or so.
502 2012-05-21 18:54:39 <gmaxwell> 01:54 < gmaxwell> http://people.xiph.org/~greg/orphans_straight.png (big image warning)
503 2012-05-21 18:54:47 <[Tycho]> OMG, looks like my FPGA-farm is not the biggest anymore...
504 2012-05-21 18:56:13 <[Tycho]> gmaxwell: text labels are way too small
505 2012-05-21 18:56:29 <graingert_ecs> gmaxwell: I do love straight orphans
506 2012-05-21 18:56:45 <[Tycho]> http://bitfury.org - was this thing announced already ?
507 2012-05-21 18:57:45 <bonks> [Tycho]: what's the biggest now?
508 2012-05-21 18:58:15 <gmaxwell> http://people.xiph.org/~greg/orphans_flat.png (very big image warning, likely to OOM your browser)
509 2012-05-21 18:58:43 <sipa> gmaxwell: works fine in chromium
510 2012-05-21 18:58:58 <gavinandresen> hmmm.  we're being awfully nice handling orphan transactions as gently as we do, I wonder if it would be better to just drop orphans
511 2012-05-21 18:59:19 <gmaxwell> 300 Mh/s  on S6-lx150? wild.
512 2012-05-21 19:00:06 <sipa> the S6LX150 on my ztex boards don't want to go above 212MHz
513 2012-05-21 19:00:24 <sipa> ah, they have shorter paths it seems
514 2012-05-21 19:00:25 <gmaxwell> sipa: sure, because the ztex design fails timing above that.
515 2012-05-21 19:01:05 <gmaxwell> looks like their hand layout is 90degs off from eldentyrell's.
516 2012-05-21 19:01:18 <gavinandresen> gmaxwell sipa : seriously, what if we refused to store/relay orphan transactions, and instead relied on the sending node to re-send them when their parents are in the chain (or otherwise made it to us) ?
517 2012-05-21 19:01:53 <gavinandresen> It would make the code a fair bit simpler.
518 2012-05-21 19:02:26 <sipa> thinking
519 2012-05-21 19:03:46 <sipa> hmm, it seems there is no "receive orphan -> request parents" code for txn?
520 2012-05-21 19:03:54 <gmaxwell> some of the random forwarding means that you may just end up hearing transactions out of order.
521 2012-05-21 19:04:58 <sipa> orphans are not announced, so not forwarded
522 2012-05-21 19:05:16 <sipa> so you should only hear about a transaction when the sender also has the parents
523 2012-05-21 19:05:37 <jgarzik> does a reorg or any other chain event ever create orphans?
524 2012-05-21 19:05:47 <jgarzik> I wouldn't want a reorg to create orphans... then we drop the orphans
525 2012-05-21 19:05:58 <jgarzik> i.e. a tx goes from confirmed to unconfirmed to dropped?
526 2012-05-21 19:06:14 <sipa> jgarzik: reorging would always add the tx's parents to the pool as well
527 2012-05-21 19:06:20 <sipa> as they were connected to the chain
528 2012-05-21 19:06:34 <sipa> except if the parents would conflict with something in the new chain
529 2012-05-21 19:06:37 <luke-jr> sipa: what if there are more txns being pulled out in the reorg than the client remembers, and the parents drop out first?
530 2012-05-21 19:07:28 <sipa> i think they'll just be dropped; not go to orphanage
531 2012-05-21 19:10:38 <sipa> gavinandresen: i'd like to see some statistics on the reasons for orphans currently (chain not caught up, parent missing, parent already in block that hasn't been seen yet, ...)
532 2012-05-21 19:11:03 <sipa> chain not caught up... i don't think we should be storing orpans at all during initial block download
533 2012-05-21 19:12:01 <gavinandresen> agreed on not storing them during initial block download.
534 2012-05-21 19:12:47 <sipa> TD pointed out that there are cases where transactions can get into the memory pool, even though they conflict with things already in the chain
535 2012-05-21 19:13:06 <sipa> but that's the memory pool, not the orphan txs
536 2012-05-21 19:14:17 <gavinandresen> yup. I'm looking at ResendWalletTransactions, and as long as the code does the right thing when re-orgs happen then it seems to me dropping orphans wouldn't hurt anything much,
537 2012-05-21 19:14:48 <sipa> my worry is transactions not being propagated in the correct order
538 2012-05-21 19:15:07 <sipa> that could make propagation a lot harder
539 2012-05-21 19:15:47 <gribble> New news from bitcoinrss: Diapolo opened pull request 1370 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1370>
540 2012-05-21 19:15:48 <sipa> but it surprises me that the code doesn't ask for parent txs when it receives an orphan
541 2012-05-21 19:15:53 <gavinandresen> I don't see how that happens, assuming messages aren't dropped (and they typically aren't)
542 2012-05-21 19:16:04 <sipa> as the sender is guaranteed to have the parents
543 2012-05-21 19:17:51 <sipa> how do you mean?
544 2012-05-21 19:18:13 <luke-jr> I mean it keeps a small buffer of txns it is "currently sending" and once it falls out of that buffer, it won't serve the data for it
545 2012-05-21 19:18:24 <gavinandresen> that's anti-social
546 2012-05-21 19:18:54 <gavinandresen> ... and smells a lot like a node that might be trying to DoS you by sending you lots of orphan txns
547 2012-05-21 19:18:54 <luke-jr> I found it rather annoying when I was trying to get that BIP16-invalid txn that was messing up pre-BIP16 miners back in April
548 2012-05-21 19:19:09 <sipa> of course it could refuse to send transactions, but i'm wondering why satoshi didn't implement such an orhpan-parant-request, given that he assumed that all nodes were running his code
549 2012-05-21 19:19:18 <luke-jr> gavinandresen: I'm talking about current client behaviour
550 2012-05-21 19:19:57 <gavinandresen> luke-jr: oh, requesting an invalid transaction, that's not relevant to this discussion
551 2012-05-21 19:20:12 <luke-jr> no, the clients I was trying to get it from considered it valid
552 2012-05-21 19:20:13 <sipa> luke-jr is right i think
553 2012-05-21 19:20:51 <sipa> if you request a tx from a node, they serve it from mapRelay, not from the block database
554 2012-05-21 19:20:55 <luke-jr> mapRelay
555 2012-05-21 19:21:02 <luke-jr> yep
556 2012-05-21 19:21:03 <gavinandresen> ok, I sit corrected
557 2012-05-21 19:21:06 <sipa> i never noticed that
558 2012-05-21 19:21:24 <gavinandresen> makes sense, I suppose, otherwise you could disk-seek-DoS a node by constantly requesting random past txns
559 2012-05-21 19:21:52 <luke-jr> looks like 15 minutes
560 2012-05-21 19:22:05 <luke-jr> (net.h)
561 2012-05-21 19:22:19 <luke-jr> gavinandresen: aren't unconfirmed txns saved in memory too?
562 2012-05-21 19:22:54 <luke-jr> maybe if it's not in mapRelay, we can fallback to mapTransactions (or whatever jgarzik replaced that with)
563 2012-05-21 19:23:10 <luke-jr> unless someone wants to verify we can just remove mapRelay entirely? :p
564 2012-05-21 19:23:21 <sipa> i wouldn't remove mapRelay
565 2012-05-21 19:25:37 <sipa> gavinandresen: i *think* removing orphans would not be a problem, but this has to be thought through very well...
566 2012-05-21 19:26:26 <gavinandresen> sipa: ACK, I mostly ask because if what gmaxwell is seeing is a DoS attack a quick fix for anybody attacked might be "don't store orphan transactions"
567 2012-05-21 19:26:41 <luke-jr> sipa: how is mapRelay better than mapTransactions?
568 2012-05-21 19:26:58 <sipa> luke-jr: i'm not suggesting to drop either
569 2012-05-21 19:27:19 <luke-jr> sipa: I'm suggesting mapTransactions does everything mapRelay does, but without expiration
570 2012-05-21 19:27:54 <sipa> that sounds reasonable as well, actually, at first sight
571 2012-05-21 19:28:57 <RedEmerald> is there an ETA on getting tor hidden services working fully? i can definitely help with testing
572 2012-05-21 19:29:10 <luke-jr> RedEmerald: does the latest next-test not work?
573 2012-05-21 19:30:00 <sipa> RedEmerald: 0.7.0 :)
574 2012-05-21 19:30:20 <RedEmerald> luke-jr, didn't realize it was in next-test. i'll play with it tonight then
575 2012-05-21 19:31:04 <sipa> i plan to set up a permanent bitcoin-torhs soon
576 2012-05-21 19:31:29 <sipa> gavinandresen: any comment on https://gist.github.com/2763381 btw?
577 2012-05-21 19:31:48 <sipa> (suggested semantics and interaction between network parameters)
578 2012-05-21 19:33:36 <sipa> luke-jr: SPV nodes may ask for transactions maybe, even though they're already in a block?
579 2012-05-21 19:34:02 <RedEmerald> sipa: torhs?
580 2012-05-21 19:34:09 <sipa> RedEmerald: tor hidden service
581 2012-05-21 19:34:12 <RedEmerald> oh duh
582 2012-05-21 19:34:52 <luke-jr> sipa: if they're already in a block, chances are they're over 15 mins old?
583 2012-05-21 19:35:08 <sipa> luke-jr: a very recent block :)
584 2012-05-21 19:35:37 <RedEmerald> https://en.bitcoin.it/wiki/Fallback_Nodes#Tor_nodes got these already, will be nice when its all properly handled though
585 2012-05-21 19:36:21 <sipa> wow, didn't know there was such a list already
586 2012-05-21 19:37:30 <RedEmerald> you have to add "mapaddress" to your torrc, so its a little bit of setup
587 2012-05-21 19:37:56 <sipa> i know
588 2012-05-21 19:38:28 <sipa> HiddenServiceDir / HiddenServicePort
589 2012-05-21 19:40:42 <RedEmerald> namecoin
590 2012-05-21 19:40:51 <sipa> right
591 2012-05-21 19:41:12 <gavinandresen> sipa: no comment, network/tor configuration isn't something I know or care much about
592 2012-05-21 19:41:45 <sipa> gavinandresen: but maybe you have comments on least-surprise behaviour or the complexity of command-line parameter interactions
593 2012-05-21 19:43:59 <gavinandresen> oh, you're changing -listen default again based on other params....   yeah, I don't like that
594 2012-05-21 19:44:11 <Matt_von_Mises> What is OP_NOP for?
595 2012-05-21 19:44:19 <gavinandresen> nothing
596 2012-05-21 19:44:41 <Matt_von_Mises> Well it does nothing. But what would you want that?
597 2012-05-21 19:45:30 <gavinandresen> sipa:  I'd rather one or more meta-flags that SoftSetArg() stuff...
598 2012-05-21 19:46:09 <gavinandresen> Or is that what -proxy and -onion are meant to be?
599 2012-05-21 19:46:26 <gavinandresen> Matt_von_Mises: doing nothing is sometimes the right thing to do.
600 2012-05-21 19:46:33 <gavinandresen> actually, it is often the right thing to do....
601 2012-05-21 19:46:55 <sipa> gavinandresen: currently, -proxy does some thing, but changes the defaults for 5 other parameters ONLY when it is port 9050
602 2012-05-21 19:47:47 <sipa> this should be an improvement, and i think the defaults are reasonable (you don't need softset, by the way, just pass a default when doing Get[Bool]Arg)
603 2012-05-21 19:48:19 <sipa> -listen was disabled earlier when you used tor mode, or used -connect already
604 2012-05-21 19:48:38 <gavinandresen> jgarzik protested that -connect= would disable -listen, if I recall correctly
605 2012-05-21 19:49:05 <sipa> i don't mind changing that, i just thought it was the case already
606 2012-05-21 19:50:53 <sipa> there are several places in the code that prevent trying to obtain the current address if -connect is used
607 2012-05-21 19:51:02 <sipa> which seems silly if you still want to be connectable
608 2012-05-21 19:51:32 <Matt_von_Mises> gavinandresen: Is a NOP needed for empty scripts? Can't a script have zero operations?
609 2012-05-21 19:51:35 <gavinandresen> I like simplifying.  And I like nice, orthogonal command-line options
610 2012-05-21 19:51:46 <luke-jr> afaik, the purpose of -connect is to talk to only one specific node O.o
611 2012-05-21 19:51:51 <sipa> luke-jr: indeed
612 2012-05-21 19:52:31 <sipa> at least, that is what i expect from it
613 2012-05-21 19:52:35 <luke-jr> Matt_von_Mises: BIP 12 and 17 are examples of how NOPs are useful
614 2012-05-21 19:52:52 <Matt_von_Mises> THanks luke-jr I will look
615 2012-05-21 19:53:04 <sipa> Matt_von_Mises: a script is true if it ends with a non-zero value on the stack
616 2012-05-21 19:53:24 <sipa> so empty or OP_NOP makes no difference for that
617 2012-05-21 19:53:33 <gavinandresen> -connect works to multiple nodes: BOOST_FOREACH(string strAddr, mapMultiArgs["-connect"])
618 2012-05-21 19:53:53 <luke-jr> gavinandresen: yes, but the difference between -addnode and -connect is that -connect inhibits the usual peer finding
619 2012-05-21 19:54:18 <gavinandresen> misfeature, in my opinion. Should be -addnode then a separate arg for "don't look for other peers"
620 2012-05-21 19:54:56 <luke-jr> perhaps.
621 2012-05-21 19:54:58 <gavinandresen> But there are WAY WAY WAY higher things on the list to worry about!
622 2012-05-21 19:55:13 <gavinandresen> (so why are we talking about trivial features?)
623 2012-05-21 19:55:40 <Matt_von_Mises> luke-jr: BIP12 and 17 don't use OP_NOP.
624 2012-05-21 19:55:48 <luke-jr> Matt_von_Mises: yes, they do.
625 2012-05-21 19:55:54 <Matt_von_Mises> At least what it says on the wiki
626 2012-05-21 19:56:14 <sipa> gavinandresen: because i want to have onion support merged, but i don't like the mess of command-line options now, and onion support would make it worse
627 2012-05-21 19:56:15 <luke-jr> Matt_von_Mises: you're reading the actual BIPs?
628 2012-05-21 19:56:29 <Matt_von_Mises> https://en.bitcoin.it/wiki/BIP_0017 I search for OP_NOP and all it says is "proceed as if an OP_NOP" not, "and then use OP_NOP"
629 2012-05-21 19:57:04 <luke-jr> &
630 2012-05-21 19:57:21 <sipa> Matt_von_Mises: both BIP 12 and BIP 17 redefine OP_NOP1
631 2012-05-21 19:57:25 <luke-jr> Matt_von_Mises: put simply, the purpose of OP_NOP is so it can be redefined in the future without breaking compatibility
632 2012-05-21 19:57:40 <Matt_von_Mises> OP_NOP isn't the same as OP_NOP1, though is it?
633 2012-05-21 19:57:42 <luke-jr> Matt_von_Mises: (also, do note both 12 and 17 were withdrawn)
634 2012-05-21 19:57:47 <luke-jr> Matt_von_Mises: functionally, they are the same
635 2012-05-21 19:57:52 <sipa> Matt_von_Mises: same semantics
636 2012-05-21 19:59:30 <sipa> if (fUseProxy || mapArgs.count("-connect") || fNoListen)
637 2012-05-21 19:59:36 <sipa> (that's from 0.5.3)
638 2012-05-21 19:59:47 <Matt_von_Mises> So they are just reserved including OP_NOP?
639 2012-05-21 20:00:26 <sipa> gavinandresen: i don't get why -connect would prevent exposing your own address, if it still listens
640 2012-05-21 20:01:14 <sipa> so it does seem to have some magic behaviour associated with it
641 2012-05-21 20:06:25 <gribble> New news from bitcoinrss: gavinandresen opened issue 1371 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1371>
642 2012-05-21 20:08:52 <gavinandresen> sipa: if you need to cleanup the networking options, then I think we should make them the way they aught to be, renaming anything that would be incompatible and checking/warning/suggesting if the user's got old, obsolete options in their bitcoin.conf.
643 2012-05-21 20:09:17 <gavinandresen> e.g. "-connect is obsolete, please use -addnode with the -foobar option instead"
644 2012-05-21 20:10:03 <xenland> So in the beaglebone I typed in init 1 and it said going into safe mode or boot mode or somthing and now when I connect through the usb with linux ubuntu 11.10 terminal/shell it just gives me a blank screen .. what shall i do to get my angstrom login screen back?
645 2012-05-21 20:10:14 <xenland> whoops
646 2012-05-21 20:10:15 <xenland> wrong chat
647 2012-05-21 20:10:21 <gmaxwell> sipa: thats two of us suggesting renaming around old stupid behavior.
648 2012-05-21 20:10:53 <sipa> i didn't consider -addnode/-connect stupid actually
649 2012-05-21 20:11:25 <luke-jr> I considered -connect's name stupid, at first.
650 2012-05-21 20:11:35 <luke-jr> I used to think it was -addnode
651 2012-05-21 20:11:57 <luke-jr> xenland: try init 4 or 5
652 2012-05-21 20:12:16 <sipa> and i'm all for renaming if semantics change in a way that old users would want to reconsider
653 2012-05-21 20:12:50 <gmaxwell> I don't know what to name the rest of connect's implied behavior though.
654 2012-05-21 20:12:52 <gavinandresen> maybe -connect is just fine the way it is, as I said I don't really care... but I don't like things like -upnp=1 is ignored if -listen=0  (but would be fine if doing that generated a warning in debug.log)
655 2012-05-21 20:12:53 <luke-jr> how about renaming it, but leaving -connect=& working undocumented for N versions?
656 2012-05-21 20:13:09 <luke-jr> -p2p=0/1 ?
657 2012-05-21 20:13:43 <luke-jr> so -p2p=0 to turn off p2p
658 2012-05-21 20:13:51 <luke-jr> and only connect when explicitly told to
659 2012-05-21 20:14:04 <sipa> -upnp is never ignored; but if you're not accepting any connections, there is just no point in trying to open up ports, or finding out what your address is
660 2012-05-21 20:14:14 <gmaxwell> I think connect and addnode are very easily confused.  And the fact that I've seen users with configs hat look like connect=foo connect=bar addnode=baz supports this.
661 2012-05-21 20:14:18 <sipa> and nobody who uses -bind would expect listening to be disabled
662 2012-05-21 20:14:43 <gmaxwell> sipa: unless they think it's going to change the binding for the rpcport?
663 2012-05-21 20:15:12 <sipa> well, that should be clearly documented of course
664 2012-05-21 20:15:22 <luke-jr> so how about an undocumented -connect=& that implies -p2p=0 -listen=0 -addnode=&
665 2012-05-21 20:15:47 <sipa> there's always a risk confusing users who don't realize the difference between P2P and RPC traffic, but there's little to do about that
666 2012-05-21 20:15:53 <sipa> luke-jr: what does -p2p do?
667 2012-05-21 20:16:03 <luke-jr> sipa: turns off p2p? :p
668 2012-05-21 20:16:10 <luke-jr> ie, looking for and connecting to peers
669 2012-05-21 20:16:12 <sipa> so, it turns networking off?
670 2012-05-21 20:16:13 <sipa> ah
671 2012-05-21 20:16:42 <gmaxwell> well what -connect does, logically, is it implied nolisten and nooutboundp2p while also addnoding
672 2012-05-21 20:17:06 <sipa> ok, -onlynode= ?
673 2012-05-21 20:17:23 <sipa> gmaxwell: apparently it never implied nolisten
674 2012-05-21 20:17:43 <sipa> but it *did* prevent exposing your own public ip
675 2012-05-21 20:18:37 <gmaxwell> Hm. Why to I remember non-local p2pool users complaining about it breaking them?
676 2012-05-21 20:21:19 <sipa> gmaxwell: i can't find any evidence for that in 0.6.0, 0.5.3 or 0.4.0
677 2012-05-21 20:22:17 <sipa> but i don't get it - there are multiple places in the code where -connect is explicitly checked, to prevent exposing our public address
678 2012-05-21 20:22:28 <sipa> but it never prevents listening
679 2012-05-21 20:23:25 <gmaxwell> does it change the binding to be 127.0.0.1?
680 2012-05-21 20:23:31 <sipa> no
681 2012-05-21 20:23:38 <sipa> that wasn't ever done
682 2012-05-21 20:24:12 <gmaxwell> Well darn. Did you think it did because of me? Or was there some common source for this mistaken belief?
683 2012-05-21 20:27:24 <diki> sipa:there are smart people out there that would do it if it was neccessary
684 2012-05-21 20:27:40 <sipa> diki: would do what?
685 2012-05-21 20:37:55 <sipa> the only existing commands that are changed by my proposal are: -proxy always turns of listening (instead of only if it's port 9050), -connect turns of listening (though i'm willing to change that), -bind always enabled listening (duh...), no listening disables upnp by default (there is absolutely no point in using it otherwise)
686 2012-05-21 20:38:38 <sipa> are these changes so invasive that we should rename the options?
687 2012-05-21 20:39:56 <sipa> (i'm always talking about defaults, in the above sentence, never changing explicitly specified options)
688 2012-05-21 20:44:35 <jgarzik> drat, gavin is gone
689 2012-05-21 20:44:37 <jgarzik> re orphans
690 2012-05-21 20:44:47 <jgarzik> we already have MAX_ORPHAN_TRANSACTIONS limit
691 2012-05-21 20:45:09 <luke-jr> sipa: FWIW, Eligius's work-node has nolisten=1 in the config, which suggests connect=& alone didn't take care of it, maybe
692 2012-05-21 20:45:13 <jgarzik> orphan tx's are randomly evicted, as new orphans appear
693 2012-05-21 20:45:36 <luke-jr> gmaxwell: http://pastebin.com/uyHE2Mku <-- Deepbit not only spams, but processes the least for the network
694 2012-05-21 20:47:04 <sipa> jgarzik: want to have a look at https://gist.github.com/2763381 ?
695 2012-05-21 20:47:06 <luke-jr> [Tycho]: ^
696 2012-05-21 20:48:06 <diki> jgarzik:le orphans
697 2012-05-21 20:48:13 <diki> or better yet "le orfans"
698 2012-05-21 20:58:37 <sipa> jgarzik: want to have a look at https://gist.github.com/2763381 ?
699 2012-05-21 20:59:08 <jgarzik> sipa: can -bind be used multiple times?
700 2012-05-21 20:59:33 <sipa> yes
701 2012-05-21 20:59:39 <sipa> externalip too
702 2012-05-21 21:03:47 <luke-jr> [23:03:15] <ljrbot> Txn 53597ae26a65fd151900852f49828d6c75f850022e82c773b44cb556beb1a1da: 198bLhyREhk2u94F5TnD8E8edbAEqEhPjE 0.00949999 BTC, 1HdAGo2KdfS9PL2BJbe7fxqCZrafhp4t46 0.00000001 BTC
703 2012-05-21 21:03:48 <luke-jr> [23:03:17] <ljrbot> Txn 6144933ff8d0d836052c0b818cdb2b2ca91a9524e3e2413497971b8d48bf4d7d: 12K5SyY2Z3DNsqFtTCnyGC3J7jYTCjM54m 0.00949999 BTC, 1MeACfxWxEnRoYAgLg7Ex1rUtbJuRUVUiC 0.00000001 BTC
704 2012-05-21 21:03:50 <luke-jr> satoshispam
705 2012-05-21 21:05:25 <[Tycho]> gmaxwell: they are talking about 300 MH/s at 1.5v - serious overvolting, may be.
706 2012-05-21 21:06:20 <luke-jr> [Tycho]: &
707 2012-05-21 21:06:31 <luke-jr> [Tycho]: why does Deepbit process the least transactions out of all the big pools?
708 2012-05-21 21:07:47 <[Tycho]> luke-jr: may be because they are considered "spam"
709 2012-05-21 21:08:02 <luke-jr> [Tycho]: I don't mean just quantity, I mean quality too
710 2012-05-21 21:08:12 <[Tycho]> Quality ?
711 2012-05-21 21:08:22 <luke-jr> Deepbit only processed 13kBTC of transactions in the last 10 blocks
712 2012-05-21 21:08:56 <[Tycho]> Can you give me an example of TXes, refused by Deepbit ? Then I can tell you why they weren't accepted.
713 2012-05-21 21:09:04 <luke-jr> no, I don't know them
714 2012-05-21 21:09:16 <luke-jr> I just know on average Deepbit processes far less than other pools
715 2012-05-21 21:09:32 <[Tycho]> Mostly they are refused because of fees too low. Some - by size.
716 2012-05-21 21:10:16 <[Tycho]> Do you know that my minimal fee is not 0.00005, but 0.01 for non-free TXes ?
717 2012-05-21 21:12:26 <gribble> New news from bitcoinrss: jgarzik opened pull request 1372 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1372>
718 2012-05-21 21:21:07 <luke-jr> OK, confirmed deterministic bdb, openssl, zlib, and miniupnpc for OSX :D
719 2012-05-21 21:21:26 <luke-jr> next up: boost :x
720 2012-05-21 21:28:13 <sipa> how heavily do you need to patch gcc, luke-jr?
721 2012-05-21 21:48:20 <luke-jr> sipa: not too heavy, starting from Apple GCC
722 2012-05-21 21:48:33 <luke-jr> sipa: GCC isn't deterministic though
723 2012-05-21 21:53:36 <luke-jr> sipa: the one patch I had to write myself was supplying the mode to an open() call that was missing it
724 2012-05-21 21:53:47 <luke-jr> someone should review that to make sure I put the right value :P
725 2012-05-21 21:54:16 <luke-jr> not that it should matter, since nothing from the cross-compiler should make it into the final binaries itself
726 2012-05-21 21:54:26 <luke-jr> (i think it was in some stdlib)
727 2012-05-21 22:15:04 <sipa> how do you mean gcc isn't deterministic? the build of gcc itself?
728 2012-05-21 22:21:55 <luke-jr> sipa: right
729 2012-05-21 22:49:40 <gribble> New news from bitcoinrss: fanquake opened issue 1373 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1373>
730 2012-05-21 22:50:58 <diki> well I'll be
731 2012-05-21 22:51:01 <diki> WindowsError: [Error 8] Not enough storage is available to process this command
732 2012-05-21 22:51:09 <diki> odd...why is Abe doing this
733 2012-05-21 22:51:22 <luke-jr> because you use Windows
734 2012-05-21 22:51:29 <diki> har har
735 2012-05-21 22:51:37 <diki> Linux is pretty much dead as a DESKTOP os
736 2012-05-21 22:51:48 <luke-jr> no u
737 2012-05-21 22:51:51 <diki> http://www.pcworld.com/businesscenter/article/252516/why_linux_on_the_desktop_is_dead.html
738 2012-05-21 22:52:10 <gmaxwell> luke-jr: Are you okay?
739 2012-05-21 22:52:30 <luke-jr> gmaxwell: yeah, forgot it was diki for a second