1 2012-10-21 00:16:42 <forrestv> you can accidentally start two bitcoind instances if you specify different -port and -rpcport options; is there any danger of database corruption if you do that?
2 2012-10-21 00:24:30 <gmaxwell> hm. we have a lock file.
3 2012-10-21 00:37:49 <forrestv> gmaxwell, hm. this is actually litecoin testnet... i definitely had four started. i'll look some into it
4 2012-10-21 00:38:09 <forrestv> actually, i can start multiple bitcoinds in that way
5 2012-10-21 00:39:54 <forrestv> or not
6 2012-10-21 00:40:02 <gmaxwell> there should be a .lock created.
7 2012-10-21 00:40:42 <echelon> is it possible to make the debug.log not as verbose?
8 2012-10-21 00:43:30 <MC1984> would be nice if it didnt print how happy it is about every block
9 2012-10-21 00:43:37 <jgarzik> gut feeling: ultraprune is so fast, it exacerbates some IBD races/problems
10 2012-10-21 00:43:40 <echelon> heh
11 2012-10-21 00:43:59 <jgarzik> I am seeing thousands of orphans and dups here
12 2012-10-21 00:44:10 <jgarzik> but progress continues, and it never gets truly stuck
13 2012-10-21 00:50:30 <gmaxwell> jgarzik: I'd really thought it was the other way: slower IBD makes it more likely to get tripped up. if its fast enough you'll pull it all before another block triggers you into orphan space.
14 2012-10-21 01:52:35 <jgarzik> kjj_: quoting fsck-up https://bitcointalk.org/index.php?topic=117982.msg1286972#msg1286972
15 2012-10-21 02:45:42 <MC1984> "ASIC's are not a natural technological generation leap like going from CPU to GPU was. ASIC's are simply specialized processing units made specifically for Bitcoin.ASIC's are not a natural technological generation leap like going from CPU to GPU was. ASIC's are simply specialized processing units made specifically for Bitcoin."
16 2012-10-21 02:46:11 <MC1984> i dont know why that came out twice, jokes on me
17 2012-10-21 03:18:05 <versakill00> It looks like Gmaxwell doesn't want a Bitcoin-Qt Wikipedia article.
18 2012-10-21 03:19:05 <versakill00> Why?
19 2012-10-21 03:21:20 <gmaxwell> versakill00: because there ought not be one.
20 2012-10-21 03:21:53 <gmaxwell> Wikipedia is an encyclopedia not a software directory. The little there is to say about the reference software can be said in the main bitcoin article
21 2012-10-21 03:22:19 <gmaxwell> As far as I can tell the only reason the article was created is because Atlas and his socks were removing all mention of it from the main Bitcoin article.
22 2012-10-21 03:22:35 <versakill00> :(
23 2012-10-21 03:22:49 <MC1984> edit warz
24 2012-10-21 03:23:24 <MC1984> i contributed to wikipedia once, it was a minor spelling or grammar correction
25 2012-10-21 03:23:26 <MC1984> it got reverted
26 2012-10-21 03:23:26 <versakill00> "Mining and node implementations include core software such as Bitcoind/Bitcoin-Qt, libbitcoin, cbitcoin[38] and BitCoinJ.[39][40] As of 2012, Bitcoind is the only one capable of fully verifying the Blockchain.[citation needed] It is still the most widely used implementation.[citation needed]"
27 2012-10-21 03:24:01 <versakill00> that needs to be cited hmm
28 2012-10-21 03:24:59 <gmaxwell> The JS implementation is close to complete, but I'm very skeptical if its correct. (simply because its creation didn't result in a pile of bug reports)
29 2012-10-21 03:26:01 <gmaxwell> The bitcoin article is basically crap now, it basically provides no distinguishing of node types, no technical details on how the system works??? says basically nothing about the consensus algorithim, etc.
30 2012-10-21 03:27:36 <versakill00> there's a bitcoin mining section
31 2012-10-21 03:39:25 <gmaxwell> versakill00 which gives minutia that would be better covered with a link to [[Hashcash]] but it never says how bitcoin achieves consensus??? which is the invention that makes bitcoin _possible_
32 2012-10-21 03:39:45 <gmaxwell> the article is embarassing now, it's just been warred over by people pushing stupid politics.
33 2012-10-21 03:47:00 <MC1984> shit is the SEC really looking into the pirate HYIP scheme?
34 2012-10-21 03:58:33 <lianj> Luke-Jr: your node down?
35 2012-10-21 04:16:23 <midnightmagic> AAAARGH it's The Graph again!
36 2012-10-21 04:16:28 <midnightmagic> http://en.wikipedia.org/wiki/File:Total_bitcoins_over_time.png
37 2012-10-21 04:17:47 <conman> heh 2am in jgarzik town; probably not a good time to start a chat
38 2012-10-21 05:01:10 <jgarzik> woah, wtf?
39 2012-10-21 05:01:24 <jgarzik> my balance disappeared, post ultraprune upgrade
40 2012-10-21 05:01:52 <jgarzik> 'listtransactions' accurately shows my last transaction, but 'listunspent' returns nothing, and 'getbalance' returns 0.0
41 2012-10-21 05:02:14 <jgarzik> restoring wallet from pre-ultraprune backup, doing -rescan under ultraprune, restores expected balance
42 2012-10-21 05:02:23 <jgarzik> not good
43 2012-10-21 05:03:23 <jgarzik> wallet.dat.bad and wallet.dat.good are same size, but cmp -l shows differences
44 2012-10-21 05:03:39 <conman> sounds like you have bigger problems to deal with
45 2012-10-21 05:05:03 <jgarzik> it's 3am here, gonna take some cough medicine and sleep
46 2012-10-21 05:05:14 <conman> sleep tight
47 2012-10-21 05:05:23 <jgarzik> conman: we need to figure out those GBT improvements
48 2012-10-21 05:05:29 <jgarzik> (some other night than tonight :))
49 2012-10-21 05:05:32 <conman> right
50 2012-10-21 05:05:39 <conman> maybe tomorrow earlier
51 2012-10-21 05:05:43 <conman> I'm home during the day
52 2012-10-21 05:08:59 <MC1984> is consensus on that most recent bitcoin research paper that its bunk?
53 2012-10-21 05:09:54 <copumpkin> the shamir one?
54 2012-10-21 05:22:57 <gmaxwell> jgarzik: reproduced here, interestingly on a node that was an older version of ultraprune previously.
55 2012-10-21 05:23:24 <gmaxwell> deleted the chain, resynced it under new ultraprune.. no balance.. but I see transactions
56 2012-10-21 05:26:29 <MC1984> yes
57 2012-10-21 06:52:31 <gmaxwell> what the fuck.
58 2012-10-21 06:52:38 <gmaxwell> atlas is running around rallying against ultraprune
59 2012-10-21 06:52:43 <Diablo-D3> gmaxwell: notice you took care of that fucking thread
60 2012-10-21 06:53:29 <Diablo-D3> gmaxwell: I wonder if all the BFL people can just be banned on the forum
61 2012-10-21 06:55:41 <gmaxwell> Diablo-D3: do you still have any room free in your mailbox?
62 2012-10-21 06:56:22 <Diablo-D3> ACTION checks
63 2012-10-21 06:56:31 <Diablo-D3> google says 0.4 of my 10.1 used.
64 2012-10-21 06:56:47 <Diablo-D3> thats gb, btw
65 2012-10-21 06:57:58 <gmaxwell> hm. I could have sworn I had a gigabytes of complaints on that thread already.
66 2012-10-21 06:58:25 <Diablo-D3> seriously though, I wonder if theymos could officially warn them or something
67 2012-10-21 07:00:42 <midnightmagic> Why bother? What are commercial companies even doing trying to carry on business on a forum filled with trolly mctrollersons?
68 2012-10-21 07:01:02 <Diablo-D3> midnightmagic: that isnt it
69 2012-10-21 07:01:06 <Diablo-D3> they HIRE people to troll the forums
70 2012-10-21 07:01:12 <Diablo-D3> as far as I can tall, thats all they hired inaba for
71 2012-10-21 07:01:14 <midnightmagic> And why do you expect businesses to operate at a higher standard than the endless socks and trolls?
72 2012-10-21 07:01:57 <midnightmagic> If you're going to hold BFL / CP to a standard, it's time to start slamming the banhammer down on the endless douchebaggeries that are far, far less civil.
73 2012-10-21 07:01:59 <gmaxwell> ACTION is looking for the hemlock
74 2012-10-21 07:02:24 <midnightmagic> I hear hemlock is a good way to go.
75 2012-10-21 07:02:35 <gmaxwell> Inaba and CP both believe they are in the right, for reasons that make sense to me when I consider their perspectives.
76 2012-10-21 07:03:13 <Diablo-D3> except
77 2012-10-21 07:03:16 <Diablo-D3> it violates one of the golden rules
78 2012-10-21 07:03:17 <gmaxwell> They've also been civil??? except for their persistance in their bickering.
79 2012-10-21 07:03:17 <midnightmagic> Hey man, your mod, your rules. They don't like it, they can retreat to their own forums that nobody ever visits.
80 2012-10-21 07:03:20 <midnightmagic> lol
81 2012-10-21 07:03:29 <Diablo-D3> dont troll other people's threads, especially when they're your competitor
82 2012-10-21 07:03:38 <midnightmagic> Diablo-D3: That's b-s, it's a forum.
83 2012-10-21 07:03:47 <Diablo-D3> bfl did it before with enterpoint, and I publicly shamed them for that
84 2012-10-21 07:04:00 <gmaxwell> yea, but it's self suicide, thats a social norm for sure??? and because of breaking it a lot of people have lost respect for Inaba.
85 2012-10-21 07:04:04 <Diablo-D3> midnightmagic: in the real world, it would be anticompetitive behavior and DoJ worthy.
86 2012-10-21 07:04:21 <Diablo-D3> but none of these are real companies, so /me shrugs
87 2012-10-21 07:04:23 <gmaxwell> I'm doing CP a favor by _not_ removing Inaba. :(
88 2012-10-21 07:04:25 <midnightmagic> I never hear anybody complaining about all the trolls when they attack BFL.
89 2012-10-21 07:04:37 <Diablo-D3> midnightmagic: I get reports of that too
90 2012-10-21 07:05:06 <gmaxwell> midnightmagic: People complain. BFL complained. And I did a similar stuff. Asked people to chill out... and nuked threats and severe incivility. (which there hasn't been in this case)
91 2012-10-21 07:05:14 <gmaxwell> s/ a //
92 2012-10-21 07:05:19 <midnightmagic> WHoah, threats?!
93 2012-10-21 07:05:32 <Diablo-D3> yeah, its nuts
94 2012-10-21 07:05:44 <Diablo-D3> I warned people not to invest in BFL, they didnt listen, so everyone can just shut the fuck up.
95 2012-10-21 07:05:51 <midnightmagic> It must be a sort of .. weird inbred dysfunctional social group dynamic going on..
96 2012-10-21 07:05:54 <Diablo-D3> you lost your money, deal with it
97 2012-10-21 07:06:01 <Diablo-D3> midnightmagic: basically.
98 2012-10-21 07:06:05 <midnightmagic> "lost your money"?
99 2012-10-21 07:07:15 <Diablo-D3> yes, I do not think bfl is going to deliver
100 2012-10-21 07:07:20 <midnightmagic> Like, ever?
101 2012-10-21 07:07:25 <Diablo-D3> pretty sure
102 2012-10-21 07:07:29 <Diablo-D3> even if they do, you still got screwed.
103 2012-10-21 07:07:29 <midnightmagic> lol
104 2012-10-21 07:07:42 <midnightmagic> You know something I don't?
105 2012-10-21 07:07:58 <midnightmagic> Good heavens, you're not reading PMs too are you?
106 2012-10-21 07:08:12 <Diablo-D3> if they dont deliver: your money is gone.
107 2012-10-21 07:08:21 <gmaxwell> MiningBuddy-: Wouldn't you like to know?
108 2012-10-21 07:08:24 <gmaxwell> oops midnightmagic
109 2012-10-21 07:08:29 <midnightmagic> lol
110 2012-10-21 07:08:38 <Diablo-D3> if they do deliver: difficulty is going to rise 100x in the next 3 months, 1000x by the next 2 years.
111 2012-10-21 07:08:42 <midnightmagic> That mental image is excellent.
112 2012-10-21 07:09:12 <Diablo-D3> bfl TCO is so high that you may not pay them off unless you have them mining for all 3 of those months
113 2012-10-21 07:09:16 <JyZyXEL> if jeri ellsworth was able to produce a working asic on her first try then so can anyone
114 2012-10-21 07:09:17 <Diablo-D3> which is dec/jan/feb
115 2012-10-21 07:09:30 <JyZyXEL> and she had to make a whole computers worth
116 2012-10-21 07:09:34 <Diablo-D3> bfl may miss the chinese new year deadline
117 2012-10-21 07:09:43 <midnightmagic> 100x? You're trying to say difficulty is going up to 2000TH?
118 2012-10-21 07:09:45 <Diablo-D3> if they do, they're fucked.
119 2012-10-21 07:09:57 <Diablo-D3> midnightmagic: yeah, somewhere around there
120 2012-10-21 07:10:00 <conman> it will be at least 50x
121 2012-10-21 07:10:02 <midnightmagic> LOL
122 2012-10-21 07:10:24 <midnightmagic> Since when did BFL say they were going to deliver all orders inside of 3 months?
123 2012-10-21 07:10:42 <midnightmagic> Last I heard, they were delivering order #3000 somewhere by next Feb.
124 2012-10-21 07:10:44 <Diablo-D3> remember, if bfl DOES ship but misses the chinese new year launch window, asicminer will beat them
125 2012-10-21 07:11:03 <Diablo-D3> so by the time anyone gets their BFL hardware, its worthless.
126 2012-10-21 07:11:04 <midnightmagic> right with all their 12TH and broken assets.
127 2012-10-21 07:11:07 <JyZyXEL> how long is the chinese new year?
128 2012-10-21 07:11:12 <midnightmagic> and competing against their own customers.
129 2012-10-21 07:11:14 <Diablo-D3> JyZyXEL: one month.
130 2012-10-21 07:11:37 <JyZyXEL> gotta love those 3 month dealextreme deliveries :p
131 2012-10-21 07:11:53 <Diablo-D3> everything is shut down by the end of the first week of jan
132 2012-10-21 07:12:23 <gmaxwell> midnightmagic: and undermining confidence in bitcoin by demonstrating the viability of a single party with a substantial fraction of the hashpower.
133 2012-10-21 07:12:35 <Diablo-D3> >broken assets
134 2012-10-21 07:12:36 <Diablo-D3> hrm?
135 2012-10-21 07:12:41 <midnightmagic> gmaxwell: Right! That too..
136 2012-10-21 07:12:56 <midnightmagic> Diablo-D3: What are asicminer shares worth right now and how can they be bought?
137 2012-10-21 07:13:10 <Diablo-D3> you cant buy any unless someone is willing to sell
138 2012-10-21 07:13:36 <Diablo-D3> friedcat is currently collecting claims of ownership for people who can sufficiently prove it ahead of glbse releasing data
139 2012-10-21 07:13:55 <midnightmagic> "sufficiently prove it"? What does that mean?
140 2012-10-21 07:13:59 <Diablo-D3> screenshots, etc
141 2012-10-21 07:14:05 <midnightmagic> photoshopshots
142 2012-10-21 07:14:12 <Diablo-D3> ACTION shrugs
143 2012-10-21 07:14:23 <Diablo-D3> DMC claimed we only have 1000, so its not like its a huge claim
144 2012-10-21 07:14:25 <midnightmagic> ACTION respects friedcat even more now.
145 2012-10-21 07:14:47 <midnightmagic> ah, that was a cruel thing to say.. I retract that, I didn't mean that.
146 2012-10-21 07:14:57 <Diablo-D3> midnightmagic: btw, not all asicminer shares were held on glbse
147 2012-10-21 07:15:02 <Diablo-D3> many people held them with friedcat directly
148 2012-10-21 07:15:13 <midnightmagic> Diablo-D3: I am aware of that. Lots were held directly with friends and family.
149 2012-10-21 07:15:37 <Diablo-D3> yeah, and friedcat's own company owns like what, 10% of the shit?
150 2012-10-21 07:15:44 <midnightmagic> Not a nepotistic little org of concentrated hashrate at all.. :-/
151 2012-10-21 07:15:56 <Diablo-D3> ACTION shrugs
152 2012-10-21 07:16:01 <Diablo-D3> if they can get shit going, its fine
153 2012-10-21 07:16:09 <gmaxwell> no worries, I can prove the transactions I want reversed were really mine; I've got screenshots.
154 2012-10-21 07:16:21 <Diablo-D3> the surviving companies will just have to work harder
155 2012-10-21 07:16:27 <Diablo-D3> and bfl wont survive
156 2012-10-21 07:18:27 <midnightmagic> There are a lot of people with very high estimates of "within 3 months" hashrate..
157 2012-10-21 07:18:39 <midnightmagic> .. but nobody ever says how they calculate it.
158 2012-10-21 07:18:40 <bladezor> Argh
159 2012-10-21 07:18:48 <bladezor> Testnet difficulty is up to 201
160 2012-10-21 07:18:55 <bladezor> and bitcoin testnet faucet is down
161 2012-10-21 07:19:15 <gmaxwell> bladezor: you can mine a block at difficulty 1 after a 20 minute gap...
162 2012-10-21 07:19:23 <gmaxwell> bladezor: or give me an address and I'll send you some coins.
163 2012-10-21 07:19:53 <Diablo-D3> midnightmagic: its based on estimates of hardware shipping
164 2012-10-21 07:21:01 <bladezor> gmaxwell: how do I set the difficulty in bitcoind?
165 2012-10-21 07:21:27 <midnightmagic> bladezor: You can't set it, it's a calculated, derived value built from incoming blocks.
166 2012-10-21 07:21:34 <gmaxwell> bladezor: you don't??? it automatically will go to one after there has been a 20 minute gap.
167 2012-10-21 07:21:39 <gmaxwell> (in testnet)
168 2012-10-21 07:21:41 <bladezor> oh okay
169 2012-10-21 07:22:31 <bladezor> gmaxwell: ms7d6bFfEeUG2wm9CCGn9nXFJHHKXgv2b9
170 2012-10-21 07:22:34 <bladezor> Thanks :)
171 2012-10-21 07:24:17 <gmaxwell> febb1285ac1b4e44add5a3d807efe10c3a87378425463bb9fd58f056bf000316
172 2012-10-21 07:37:03 <sipa> jgarzik: can you db5.1_dump both wallets, and tell/give/show the difference? (removing private keys, of course)
173 2012-10-21 07:37:44 <sipa> jgarzik: iirc with -p, private keys really show up as "...key....."
174 2012-10-21 09:36:30 <slush1> Can anybody send me few testnet bitcoins to miWqkZQ1nLaEKtahPcqXWjNn2GzYbruvqX ?
175 2012-10-21 09:36:43 <slush1> Faucet looks broken and I need something to test.
176 2012-10-21 09:37:00 <Diablo-D3> heh its a slush
177 2012-10-21 09:38:05 <slush1> Diablo-D3: heh, so? :)
178 2012-10-21 09:38:19 <Diablo-D3> dont see you on irc much anymore
179 2012-10-21 09:45:32 <slush1> Diablo-D3: I'm here most of the time, but not talking in -dev too much
180 2012-10-21 09:46:30 <slush1> nobody with at least few testnet coins here? :(
181 2012-10-21 09:46:50 <Diablo-D3> not me
182 2012-10-21 09:48:02 <senseless> i got a block on my test net wallet if testnet hasnt been regenisesed since then
183 2012-10-21 09:48:10 <senseless> give me a bit and ill load it up
184 2012-10-21 09:49:14 <slush1> oh, nevermind, I just found one lost wallet with 120700 tesnet coins O:-)
185 2012-10-21 09:49:27 <Diablo-D3> >lost
186 2012-10-21 09:49:28 <Diablo-D3> wat?
187 2012-10-21 09:49:34 <senseless> slush1: e770b26259d9a045ff05a6347f392b8eb6e9cb65bcb7164f06425868ad04d254
188 2012-10-21 09:49:40 <slush1> senseless: thanks :)
189 2012-10-21 09:49:49 <slush1> Diablo-D3: found on old development machine
190 2012-10-21 09:49:57 <Diablo-D3> >lost machine
191 2012-10-21 09:49:58 <Diablo-D3> wat?
192 2012-10-21 09:50:25 <Diablo-D3> what is this, the bash quote about losing a machine, but it still reponds to pings, you just dont know exactly where it is?
193 2012-10-21 09:51:08 <slush1> hehe, I don't understand you
194 2012-10-21 09:51:15 <slush1> I simply found one wallet.dat with few coins
195 2012-10-21 09:51:43 <MC1984> atlas is a twat
196 2012-10-21 09:51:48 <Diablo-D3> http://www.bash.org/?5273
197 2012-10-21 09:51:57 <MC1984> also that guy strata seems like a nob
198 2012-10-21 09:52:04 <slush1> Diablo-D3: lol
199 2012-10-21 09:52:15 <Diablo-D3> MC1984: Achievement Unlocked: MASTER OF THE OBVIOUS
200 2012-10-21 09:53:26 <MC1984> why cant they just start banning fools
201 2012-10-21 09:53:43 <Diablo-D3> because theymos wont let me :<
202 2012-10-21 09:53:44 <MC1984> forums are forums, someone runs it and its not a free speech zone
203 2012-10-21 09:54:19 <MC1984> i think those forums are a detriment to bitcoin right now
204 2012-10-21 09:54:22 <senseless> because they believe in liberty and free speech, which to them means they can come shit in your living room because "it's their right"
205 2012-10-21 09:54:38 <Diablo-D3> yes, and I can also shoot them because its my right.
206 2012-10-21 09:54:49 <Diablo-D3> also, my living room doubles as a private armory.
207 2012-10-21 09:54:57 <senseless> I just mean, if they get banned people will start pouting and whining like girls :/
208 2012-10-21 09:55:15 <Diablo-D3> you know what happens to girls that pout and whine?
209 2012-10-21 09:55:34 <Diablo-D3> hint: resistance only makes my dick harder.
210 2012-10-21 09:55:38 <senseless> they end up single with 50 cats?
211 2012-10-21 09:55:40 <senseless> im not sure
212 2012-10-21 09:56:11 <MC1984> the sad thing is it would be great to have people checking the dev team and shit, but that guy makes it look like an errand for asshats
213 2012-10-21 09:56:29 <Diablo-D3> ACTION goes back to working on Seaking
214 2012-10-21 09:57:22 <MC1984> diablo implies he frequently rapes intrasigent women into submission
215 2012-10-21 09:57:40 <MC1984> in the context of the forum bitches, i support this initiative
216 2012-10-21 09:58:03 <Diablo-D3> I wonder if I can buy condoms in bulk on amazon
217 2012-10-21 09:58:31 <MC1984> fucking atlas didactic asshole
218 2012-10-21 09:58:35 <Diablo-D3> http://www.sfgate.com/bayarea/article/FBI-must-pay-S-F-journalist-470-000-3965054.php
219 2012-10-21 09:58:36 <MC1984> fffffffffffffffffffffff
220 2012-10-21 09:58:55 <Diablo-D3> ACTION puts his stamp of approval on this
221 2012-10-21 09:59:15 <Diablo-D3> raping the fbi in court? sounds like my kind of party.
222 2012-10-21 09:59:43 <MC1984> keep laughing, thats your money
223 2012-10-21 10:00:07 <Diablo-D3> what money?
224 2012-10-21 10:00:18 <Diablo-D3> its all fake money that doesnt exist
225 2012-10-21 10:00:37 <MC1984> still your fake money
226 2012-10-21 10:00:42 <Diablo-D3> what money?
227 2012-10-21 10:00:49 <Diablo-D3> seriously, Im like the poorest person in here
228 2012-10-21 10:00:53 <MC1984> id pay any fine with someone elses magic money too
229 2012-10-21 10:01:19 <MC1984> also observe as the FBI receives a tip off about CP on that guys computer a couple of years from now
230 2012-10-21 10:03:06 <slush1> I still don't understand why testnet difficulty is different than nbits
231 2012-10-21 10:03:16 <sipa> slush1: it is not
232 2012-10-21 10:03:30 <sipa> the rules for calculating difficulty/nbits are just different
233 2012-10-21 10:03:33 <slush1> When I calculate target from nbits, it is much higher than 1 given in getinfo difficulty
234 2012-10-21 10:03:36 <slush1> oh
235 2012-10-21 10:03:45 <MC1984> oh wow is ultraprune stuff officially in the mainline now?
236 2012-10-21 10:03:56 <sipa> MC1984: yup
237 2012-10-21 10:04:08 <MC1984> you magnificent bastard
238 2012-10-21 10:04:16 <slush1> sipa: btw congratulation for this. I really like you development in this area
239 2012-10-21 10:04:24 <sipa> slush1: thanks :)
240 2012-10-21 10:05:16 <slush1> sipa: btw where are these different rules for nbits described?
241 2012-10-21 10:05:39 <sipa> slush1: good question - i??? afraid only in the source code
242 2012-10-21 10:05:52 <slush1> quite typical ;)
243 2012-10-21 10:07:17 <sipa> the informal rule is "when the time different with the parent block is more than 20 minutes, difficulty is 1; blocks afterwards resume the normal difficulty of the 2016-block-range they are in"
244 2012-10-21 10:07:50 <sipa> but there are weird (and unintended) interactions when the transition block has a 20 minute delay itself
245 2012-10-21 10:08:15 <slush1> oh, so nbits is still the same and these alternate rules apply only to difficulty?
246 2012-10-21 10:09:02 <sipa> yes
247 2012-10-21 10:09:31 <sipa> but nbits will reflect the difficulty 1
248 2012-10-21 10:09:38 <slush1> hm, skipped target checking, but I'm still unable to generate testnet block
249 2012-10-21 10:19:35 <MC1984> where can i get mainline builds
250 2012-10-21 10:19:48 <MC1984> is there such a thing
251 2012-10-21 10:20:13 <sipa> you mean releases?
252 2012-10-21 10:20:30 <MC1984> test builds
253 2012-10-21 10:20:35 <sipa> we don't have daily builds of mainline
254 2012-10-21 10:20:38 <MC1984> like yours
255 2012-10-21 10:20:52 <sipa> but BlueMatt's pull tester creates builds for every pull request
256 2012-10-21 10:21:18 <MC1984> whats the plan for testing ultraprune then
257 2012-10-21 10:21:59 <sipa> daily builds would be nice, i suppose
258 2012-10-21 10:22:20 <sipa> for now, testing has relied on people building themselves, until release candidates are there
259 2012-10-21 10:23:48 <sipa> and there is Luke-Jr's next-test builds
260 2012-10-21 10:24:56 <MC1984> i wanted to continue to test
261 2012-10-21 10:25:14 <MC1984> but im a pleb, i cant compile programs
262 2012-10-21 10:51:44 <Joric> someone made a front page! http://www.reddit.com/r/Bitcoin/comments/11u55q/jgarzik_my_bitcoin_balance_disappeared_post/
263 2012-10-21 11:14:11 <sipa> jgarzik, gmaxwell: bug found
264 2012-10-21 11:46:55 <Luke-Jr> jgarzik: IMO what might make GBT simpler to implement in miners would be to require coinbasetxn ;)
265 2012-10-21 11:48:22 <Diapolo> Seems I'm currently not able to compile with Ultraprune directly on Windows. I get a:
266 2012-10-21 11:54:41 <sipa> Diapolo: try building leveldb separately beforehand
267 2012-10-21 11:54:51 <sipa> Diapolo: just go into the subdirectory, and make there
268 2012-10-21 11:56:01 <Diapolo> sipa: just mingw32-make in leveldb folder?
269 2012-10-21 11:56:23 <sipa> yes
270 2012-10-21 11:56:31 <sipa> oh, no
271 2012-10-21 11:56:37 <sipa> make libleveldb.a libmemenv.a
272 2012-10-21 11:57:50 <Diapolo> I never used make, not even sure if this is part of mingw ;) will try, just a sec.
273 2012-10-21 11:58:33 <Diapolo> no, there is no make
274 2012-10-21 11:58:44 <sipa> oh, mingw32-make in your case
275 2012-10-21 11:59:01 <sipa> i just mean you need to build libleveldb.a and libmemenv.a, not the default targets
276 2012-10-21 12:00:04 <Diapolo> C:\\Users\\Diapolo\\bitcoin.Qt\\src\\leveldb>mingw32-make libleveldb.a libmemenv.a leads to:
277 2012-10-21 12:00:22 <Diapolo> process_begin: CreateProcess(NULL, sh C:\\Users\\Diapolo\\bitcoin.Qt\\src\\leveldb\\build_detect_platform build_config.mk, ...) failed.
278 2012-10-21 12:01:25 <sipa> no clue about that - sorry
279 2012-10-21 12:01:38 <D34TH> hold on ill try
280 2012-10-21 12:01:38 <sipa> the win32 leveldb port was contributed by justmoon
281 2012-10-21 12:01:54 <D34TH> link?
282 2012-10-21 12:01:58 <Diapolo> That's why I asked if anyone compiled it on windows :D.
283 2012-10-21 12:02:06 <D34TH> or is it in head
284 2012-10-21 12:02:10 <sipa> D34TH: HEAD
285 2012-10-21 12:02:23 <sipa> Diapolo: well i did builds for windows, not on windows
286 2012-10-21 12:03:05 <D34TH> made
287 2012-10-21 12:03:08 <D34TH> just fine
288 2012-10-21 12:03:09 <Diapolo> that's the difference it seems ... any idea how this can be run ob Windows?
289 2012-10-21 12:03:16 <D34TH> "Unknown" platform
290 2012-10-21 12:03:36 <Diapolo> D34TH: How did you proceed?
291 2012-10-21 12:04:14 <D34TH> v
292 2012-10-21 12:04:15 <D34TH> http://pastebin.com/s0YYNChh
293 2012-10-21 12:04:54 <Diapolo> build_detect_platform is a shell script I dunno how that would work on Windows
294 2012-10-21 12:05:03 <D34TH> i have sh
295 2012-10-21 12:05:07 <D34TH> notice my use of ls
296 2012-10-21 12:05:07 <sipa> well you need a shell obviously
297 2012-10-21 12:05:13 <D34TH> MSYS
298 2012-10-21 12:05:44 <Diapolo> I never use this ... is this the MinGW Shell link in start menu?
299 2012-10-21 12:05:59 <sipa> did bitcoin even build before, without a shell?
300 2012-10-21 12:06:13 <Diapolo> Bitcoin-Qt build just fine, yes I never compiled just bitcoind
301 2012-10-21 12:06:29 <sipa> how was build.h generated, for example?
302 2012-10-21 12:06:31 <Diapolo> I have a shell but it's the Windows command-prompt :D
303 2012-10-21 12:06:52 <sipa> ah, that has a !windows in the bitcoin-qt.pro file
304 2012-10-21 12:06:57 <Diapolo> build.h was not build
305 2012-10-21 12:07:00 <Diapolo> yes
306 2012-10-21 12:07:13 <D34TH> i did run into an issue though compiling bitcoind
307 2012-10-21 12:07:29 <D34TH> makefile.mingw:98: *** missing separator. Stop.
308 2012-10-21 12:08:02 <sipa> oh, that's an actual bug
309 2012-10-21 12:08:11 <sipa> just change those 4 spaces into a tab
310 2012-10-21 12:08:35 <D34TH> yeay boost-1.5
311 2012-10-21 12:09:50 <Diapolo> how can I supply the platform with make?
312 2012-10-21 12:10:06 <Diapolo> $ make
313 2012-10-21 12:10:07 <Diapolo> Unknown platform!
314 2012-10-21 12:10:13 <D34TH> the file is made
315 2012-10-21 12:11:13 <sipa> now type
316 2012-10-21 12:11:18 <sipa> make libmemenv.a
317 2012-10-21 12:12:12 <D34TH> i would assume TARGET_OS=OS_WINDOWS_CROSSCOMPILE make
318 2012-10-21 12:12:12 <Diapolo> I now have these 2 files, but they are just 8 Bytes in size?
319 2012-10-21 12:12:54 <Diapolo> that's it, it seems to build now
320 2012-10-21 12:13:07 <D34TH> :3
321 2012-10-21 12:13:33 <Diapolo> damn, now the paths to boost are missing -_-
322 2012-10-21 12:13:42 <D34TH> because it doesnt ask for a version
323 2012-10-21 12:13:45 <D34TH> specify your version
324 2012-10-21 12:14:16 <sipa> feel free to submit a pullreq to simplify this process
325 2012-10-21 12:14:53 <Diapolo> how can I specify my version?
326 2012-10-21 12:19:02 <D34TH> i prefer editing the makefile
327 2012-10-21 12:19:05 <D34TH> line 120
328 2012-10-21 12:19:13 <D34TH> **119
329 2012-10-21 12:20:41 <Diapolo> I see in build_detect_platform you mean
330 2012-10-21 12:20:55 <D34TH> yea
331 2012-10-21 12:20:57 <D34TH> whoops
332 2012-10-21 12:21:39 <D34TH> http://pastebin.com/SxnSpnzR
333 2012-10-21 12:21:43 <D34TH> im at that right now
334 2012-10-21 12:23:19 <sipa> you don't need to build the .so file
335 2012-10-21 12:23:38 <sipa> (still shouldn't fail, though0
336 2012-10-21 12:24:27 <D34TH> i make the .a's fine
337 2012-10-21 12:24:48 <sipa> libleveldb.a and libmemenv.a is all you need
338 2012-10-21 12:24:58 <D34TH> i dont have a leveldb/libleveldb.lib for bitcoind
339 2012-10-21 12:25:07 <Diapolo> BOOST chrono seems missing
340 2012-10-21 12:25:14 <D34TH> yea i added it
341 2012-10-21 12:25:16 <D34TH> want my changed makefile
342 2012-10-21 12:25:26 <sipa> D34TH: please submit a pullreq :)
343 2012-10-21 12:25:38 <D34TH> sipa after i specify its own case
344 2012-10-21 12:25:42 <D34TH> of NATIVE_WINDOWS
345 2012-10-21 12:25:55 <D34TH> instead of borking crosscompile
346 2012-10-21 12:25:58 <sipa> ok
347 2012-10-21 12:26:00 <Diapolo> good idea for a patch :)
348 2012-10-21 12:26:24 <Diapolo> yes now it's working
349 2012-10-21 12:26:26 <sipa> hmm, what is that .lib file?
350 2012-10-21 12:26:51 <D34TH> mingw32-make: *** No rule to make target 'leveldb/libleveldb.lib', needed by 'ob
351 2012-10-21 12:26:55 <Diapolo> build_detect_platform line 119:
352 2012-10-21 12:27:42 <D34TH> Diapolo i was going to specify the libs found in makefile.mingw
353 2012-10-21 12:28:14 <Diapolo> since boost 1.5 the thread lib needs chrono on Windows
354 2012-10-21 12:28:31 <Diapolo> you can create your pull and I'll be glad to comment and test it out :)
355 2012-10-21 12:28:54 <sipa> you may want to talk to justmoon, as he contributed it
356 2012-10-21 12:29:00 <sipa> (and maybe i screwed it up)
357 2012-10-21 12:30:11 <Diapolo> Well Qt-Creator still doesn't fine libleveldb.a but it's there ... hmm
358 2012-10-21 12:30:34 <D34TH> i could tell you why
359 2012-10-21 12:30:44 <D34TH> are you using your mingw or qt-creators mingw
360 2012-10-21 12:30:59 <D34TH> if you `dont know` you have 2 mingws
361 2012-10-21 12:32:10 <Diapolo> not the included one
362 2012-10-21 12:32:26 <Diapolo> upgraded to mingw 4.7.1 and compiled the Qt stuff for myself with that
363 2012-10-21 12:33:00 <D34TH> ahh
364 2012-10-21 12:35:10 <D34TH> proposing file change
365 2012-10-21 12:35:51 <D34TH> stupid litecoin screwed up my push
366 2012-10-21 12:36:22 <Diapolo> sipa: When I have the 2 .a files in leveldb, is there anything more to be done or do I just need to link to them / use them?
367 2012-10-21 12:37:06 <sipa> just that
368 2012-10-21 12:37:27 <sipa> though i wonder why the makefile refers to libleveldb.lib
369 2012-10-21 12:38:20 <Diapolo> .lib is not MinGW that's MS VC++ no?
370 2012-10-21 12:38:25 <D34TH> https://github.com/bitcoin/bitcoin/pull/1940
371 2012-10-21 12:39:02 <D34TH> line 99: makefile.mingw
372 2012-10-21 12:39:07 <Diapolo> it seems to build Bitcoin-Qt now, I removed all that magic except:
373 2012-10-21 12:39:08 <D34TH> thats probably why
374 2012-10-21 12:39:13 <sipa> D34TH: do you have to refer to specific boost versions?
375 2012-10-21 12:39:33 <D34TH> i dont believe windows boost compiles general versions
376 2012-10-21 12:39:36 <D34TH> but no you shouldnt
377 2012-10-21 12:41:19 <Diapolo> somehow the boost version should be supplyable via the shell or we need to supply the version we require for Windows anyway, which is MinGW 4.4 with Boost 1.50 currently AFAIK.
378 2012-10-21 12:41:57 <D34TH> ahh, in the makefile.mingw it says 4.5
379 2012-10-21 12:44:04 <Diapolo> https://github.com/bitcoin/bitcoin/blob/master/bitcoin-qt.pro#L319
380 2012-10-21 12:44:52 <D34TH> https://github.com/bitcoin/bitcoin/blob/master/src/makefile.mingw#L19
381 2012-10-21 12:45:26 <sipa> ah, we already do it seems
382 2012-10-21 12:45:27 <sipa> meh
383 2012-10-21 12:46:09 <Diapolo> D34TH: strange makes no sense to use 2 different hard-coded MinGW version references IMO.
384 2012-10-21 12:46:35 <D34TH> ACTION points slowly
385 2012-10-21 12:46:50 <D34TH> https://github.com/bitcoin/bitcoin/commit/a1755ec4264248323f9a38ac0c3f19200b77c9f8#src/makefile.mingw
386 2012-10-21 12:47:32 <Diapolo> I have no idea, why 4.5 is mentioned there ^^.
387 2012-10-21 12:48:24 <Diapolo> sipa: I have to warnings left when compiling Bitcoin-Qt now
388 2012-10-21 12:48:40 <Diapolo> C:\\Users\\Diapolo\\bitcoin.Qt\\src\\init.cpp:83: Warnung:deleting object of polymorphic class type 'CCoinsViewCache' which has non-virtual destructor might cause undefined behaviour [-Wdelete-non-virtual-dtor]
389 2012-10-21 12:48:49 <Diapolo> and C:\\Users\\Diapolo\\bitcoin.Qt\\src\\init.cpp:84: Warnung:deleting object of polymorphic class type 'CCoinsViewDB' which has non-virtual destructor might cause undefined behaviour [-Wdelete-non-virtual-dtor]
390 2012-10-21 12:48:51 <sipa> interesting
391 2012-10-21 12:49:33 <Diapolo> and Bitcoin-Qt is crashing with only wallet.dat and bitcoin.conf in the folder I need to investigate this a little
392 2012-10-21 12:49:40 <sipa> :o
393 2012-10-21 12:53:20 <D34TH> i hate libdb
394 2012-10-21 12:53:41 <sipa> haha
395 2012-10-21 12:55:48 <D34TH> http://pastebin.com/ktpt27mV
396 2012-10-21 12:56:15 <D34TH> if you go down past the txdb and leveldb youll see my hatred
397 2012-10-21 12:56:31 <D34TH> Diapolo, you should send me your libdb's
398 2012-10-21 12:57:50 <Diapolo> the .a's?
399 2012-10-21 12:57:57 <D34TH> yea
400 2012-10-21 12:58:53 <Diapolo> I'll zip and up them, just a sec.
401 2012-10-21 13:02:01 <Diapolo> http://wikisend.com/download/622246/leveldb.zip
402 2012-10-21 13:02:05 <Diapolo> there you go
403 2012-10-21 13:02:09 <D34TH> thnx
404 2012-10-21 13:02:35 <D34TH> nonono, libdb not libleveldb
405 2012-10-21 13:03:03 <D34TH> libdb.a, libdb_cxx.a
406 2012-10-21 13:05:21 <Diapolo> ah sorry ...
407 2012-10-21 13:05:32 <D34TH> ill just try compiling it
408 2012-10-21 13:06:49 <Diapolo> that was not easy I remember ^^
409 2012-10-21 13:07:01 <Diapolo> http://wikisend.com/download/287114/libdb.zip
410 2012-10-21 13:07:18 <D34TH> <3
411 2012-10-21 13:07:52 <sipa> ACTION hopes to get rid of the bdb dependency soon
412 2012-10-21 13:08:14 <D34TH> still getting that stupid db error
413 2012-10-21 13:08:15 <D34TH> D:
414 2012-10-21 13:08:57 <Diapolo> sipa: could the posted warnings above cause a hard crash with Bitcoin-Qt?
415 2012-10-21 13:09:21 <Diapolo> well I guess no as they are in Shutdown() ^^ sorry
416 2012-10-21 13:11:01 <D34TH> renaming libdb to libfubar
417 2012-10-21 13:11:51 <Diablo-D3> heh
418 2012-10-21 13:16:07 <Diapolo> sipa: still crashes ... will try with a clean datadir
419 2012-10-21 13:16:40 <D34TH> oh wow
420 2012-10-21 13:16:45 <D34TH> i feel stupid
421 2012-10-21 13:17:18 <D34TH> that fixed most of my issues
422 2012-10-21 13:17:21 <D34TH> LOL
423 2012-10-21 13:18:11 <Diapolo> D34TH: do you build bitcoind or Bitcoin-Qt currently?
424 2012-10-21 13:18:18 <D34TH> attempting to do qt
425 2012-10-21 13:19:57 <Diapolo> good will be interesting to see if it's working for you then
426 2012-10-21 13:23:22 <sipa> Diapolo: anything in debug.log when it crashes?
427 2012-10-21 13:35:37 <Diapolo> sipa: nothing unusal and that is what bothers me...
428 2012-10-21 13:35:56 <Diapolo> 10/21/12 15:17:16 ThreadDNSAddressSeed started
429 2012-10-21 13:35:56 <Diapolo> these are the last 4 lines before the crash:
430 2012-10-21 13:46:22 <D34TH> having some leveldb errors
431 2012-10-21 13:51:59 <D34TH> still getting libleveldb errors
432 2012-10-21 13:53:04 <Diapolo> which ones?
433 2012-10-21 13:53:39 <D34TH> oh wiat
434 2012-10-21 13:53:42 <D34TH> just got past them
435 2012-10-21 13:53:54 <D34TH> why are your leveldb.a's so much larger
436 2012-10-21 13:55:42 <Diapolo> I'm asking myself if it is intended that leveldb is compiled with -O3 -march=native -mtune=native or if that comes from my env.
437 2012-10-21 13:56:10 <Diapolo> oh I set this as env-flag ^^
438 2012-10-21 13:56:22 <D34TH> built
439 2012-10-21 13:56:45 <D34TH> and error
440 2012-10-21 13:56:48 <D34TH> whoot
441 2012-10-21 13:57:37 <D34TH> and now it insta crashes
442 2012-10-21 13:58:19 <fiesh> march implies mtune...
443 2012-10-21 13:59:54 <D34TH> still different
444 2012-10-21 14:00:41 <Diapolo> D34HT: it built bitcoin-qt.exe and crashes on startup?
445 2012-10-21 14:00:46 <D34TH> yep
446 2012-10-21 14:00:49 <Diapolo> same here
447 2012-10-21 14:01:07 <Diapolo> I see the splash screen and before the GUI shows up it crashes hard
448 2012-10-21 14:01:28 <D34TH> i dont even see splash
449 2012-10-21 14:01:40 <Diapolo> strange ... what's in debug.log
450 2012-10-21 14:02:09 <D34TH> doesnt get that far
451 2012-10-21 14:02:40 <D34TH> immediate werfault
452 2012-10-21 14:03:45 <D34TH> bitcoind compiled and is running just fine
453 2012-10-21 14:04:00 <D34TH> must be gay QT
454 2012-10-21 14:04:05 <Diapolo> now that is strange ...
455 2012-10-21 14:04:39 <Diapolo> so at least the leveldb libs seem to be okay then?
456 2012-10-21 14:04:52 <D34TH> yes
457 2012-10-21 14:06:09 <Diapolo> the last thing I see with my Bitcoin-Qt is "Loading wallet..."
458 2012-10-21 14:07:04 <Diapolo> crashes with QtGui4.dll
459 2012-10-21 14:07:22 <D34TH> put the 2 dll's in the folder
460 2012-10-21 14:07:43 <Diapolo> no no I was able to built Bitcoin-Qt just fine before the Ultraprune patch
461 2012-10-21 14:11:43 <D34TH> sipa: if leveldb can be used for transaction db why not wallet
462 2012-10-21 14:13:39 <Diapolo> hey wow now it's starting ...
463 2012-10-21 14:13:58 <sipa> D34TH: mainly the fact that it is overkill, and requires an entire directory for storage instead of just a file
464 2012-10-21 14:14:23 <D34TH> Diapolo, what id you do
465 2012-10-21 14:14:35 <Diapolo> sipa: I moved NewThread(ThreadImport, vPath); in init.cpp into the if-clause ... and it starts could there be some problem with the -loadblock Thread?
466 2012-10-21 14:15:31 <sipa> Diapolo: if you do that, you disable the bootstrap.dat ability
467 2012-10-21 14:15:48 <D34TH> add -bootstrap?
468 2012-10-21 14:15:49 <Diapolo> right. but if I don't use it why should it spawn a thread
469 2012-10-21 14:16:12 <sipa> D34TH: what's the point of that? then we could just use -loadblock
470 2012-10-21 14:16:17 <sipa> Diapolo: "meh"
471 2012-10-21 14:17:04 <Diapolo> sorry? I'm rather sure this causes my crash:
472 2012-10-21 14:17:18 <sipa> why would it?
473 2012-10-21 14:17:35 <sipa> show me the bug, and I'll believe you, but in that code alone is no problem
474 2012-10-21 14:17:36 <Luke-Jr> O.o
475 2012-10-21 14:17:44 <sipa> maybe there is a bug in ThreadImport, though
476 2012-10-21 14:17:50 <Luke-Jr> well, it doesn't make sense either :P
477 2012-10-21 14:18:55 <Diapolo> sipa: that was what I thought, too ... a bug in ThreadImport
478 2012-10-21 14:19:22 <D34TH> quick find and murder
479 2012-10-21 14:19:27 <D34TH> ill get the raid
480 2012-10-21 14:20:03 <Luke-Jr> (reading the full code does make sense tho)
481 2012-10-21 14:20:04 <D34TH> mine still faults immediatly
482 2012-10-21 14:20:13 <D34TH> its gotta be my QT
483 2012-10-21 14:20:47 <Diapolo> if vPath does not contain any file, can this cause a crash?
484 2012-10-21 14:21:25 <sipa> Luke-Jr: ThreadImport should move to init.cpp, with the import loop code in a function in main.cpp
485 2012-10-21 14:21:33 <sipa> that would make things clearer as well
486 2012-10-21 14:21:42 <sipa> Diapolo: it shouldn't
487 2012-10-21 14:23:23 <Diapolo> I verified it, when I comment-out NewThread(ThreadImport, vPath); Bitcoin-Qt does NOT crash.
488 2012-10-21 14:24:03 <sipa> ok, add some print statements inside of ThreadImport then, and localize where the problem is
489 2012-10-21 14:24:11 <sipa> or run it in a debugger
490 2012-10-21 14:24:17 <sipa> and break in that function
491 2012-10-21 14:24:41 <Diapolo> sipa: I'll for sure investigate further, but at least my Qt Creator doesn't allow me to debug currently :-/.
492 2012-10-21 14:24:54 <sipa> you don't have gdb?
493 2012-10-21 14:25:38 <sipa> anyway, adding debug output can function as a weak replacement of a debugger :)
494 2012-10-21 14:25:43 <Diapolo> I did not compile the recent Qt version as debug ^^
495 2012-10-21 14:26:47 <Luke-Jr> Diapolo: just ensure Bitcoin-Qt gets compiled with -ggdb -O0
496 2012-10-21 14:26:53 <Luke-Jr> maybe install DrMingw
497 2012-10-21 14:27:09 <Diapolo> never heard of that Luke
498 2012-10-21 14:27:15 <Luke-Jr> tho Qt Creator might do everything DrMingw does itself
499 2012-10-21 14:27:24 <Luke-Jr> DrMingw just displays crash info
500 2012-10-21 14:27:38 <Diapolo> even that -ggdb -O0 flag is new to me ^^
501 2012-10-21 14:28:12 <Luke-Jr> -ggdb = include extended gdb-compatible debugging information
502 2012-10-21 14:28:18 <Luke-Jr> -O0 = don't optimize anything, period
503 2012-10-21 14:28:24 <Luke-Jr> (optimizations can make debugging a pain)
504 2012-10-21 14:28:26 <Diapolo> -O0 was clear
505 2012-10-21 14:32:59 <Diapolo> now that fix was easy ... it seems
506 2012-10-21 14:33:37 <Diapolo> if (!vFiles->empty()) {
507 2012-10-21 14:33:38 <Diapolo> ....
508 2012-10-21 14:34:12 <Luke-Jr> ???
509 2012-10-21 14:34:37 <sipa> Diapolo: you don't get it; even if vFiles is empty, bootstrap.dat should still be processed
510 2012-10-21 14:34:42 <sipa> and that happens in that thread
511 2012-10-21 14:34:51 <Diapolo> right and it happens after
512 2012-10-21 14:35:14 <sipa> i don't get it
513 2012-10-21 14:35:16 <Diapolo> first part manual -loadblock (for which this check is)
514 2012-10-21 14:35:24 <sipa> and that helps? :S
515 2012-10-21 14:35:27 <Diapolo> and second part is bootstrap
516 2012-10-21 14:35:36 <Diapolo> which this check doesn't touch
517 2012-10-21 14:35:52 <sipa> does it help?
518 2012-10-21 14:36:14 <Diapolo> http://pastebin.de/30366
519 2012-10-21 14:36:19 <Diapolo> it fixes the crash for me
520 2012-10-21 14:36:32 <Luke-Jr> sipa: IMO there's definitely some fclose()s missing
521 2012-10-21 14:36:41 <jgarzik> wakeup email, pass #1: delete all SF admin spam. pass #2: delete all Windows-related github emails. pass #3: actually read email
522 2012-10-21 14:36:52 <sipa> even then, Diapolo's change doesn't make sense
523 2012-10-21 14:36:58 <sipa> as in his case, no file is opened
524 2012-10-21 14:37:05 <Luke-Jr> yes, unrelated to that I mean
525 2012-10-21 14:37:24 <sipa> also, LoadExternalBlockFile closes itself
526 2012-10-21 14:37:25 <Luke-Jr> Diapolo: what if you only comment out this line? uiInterface.InitMessage(_("Starting block import???"));
527 2012-10-21 14:37:34 <Luke-Jr> sipa: ew O.o
528 2012-10-21 14:37:35 <sipa> as it uses CAutoFile which closes when going out of scope
529 2012-10-21 14:37:43 <sipa> yeah, eww indeed
530 2012-10-21 14:37:49 <jgarzik> sipa: will test your zero-balance fix later tonight
531 2012-10-21 14:37:59 <jgarzik> sipa: I'm surprised you did not see this before :)
532 2012-10-21 14:38:17 <sipa> jgarzik: well it requires importing from scratch with a full wallet
533 2012-10-21 14:38:29 <jgarzik> sipa: which will be the normal upgrade case :)
534 2012-10-21 14:38:37 <sipa> but i should have noticed, indeed
535 2012-10-21 14:38:38 <Diapolo> Luke-Jr: I'll try
536 2012-10-21 14:39:08 <jgarzik> RE fclose()... yeah it's annoying, passing FILE to CAutoFile()
537 2012-10-21 14:39:22 <jgarzik> I would rather pass the FILE*, wait for the func to return, then close it
538 2012-10-21 14:39:26 <jgarzik> or simply pass CAutoFile
539 2012-10-21 14:39:29 <sipa> jgarzik: indeed
540 2012-10-21 14:39:44 <Diapolo> Luke that's it ... seems the initMessage causes an error here
541 2012-10-21 14:40:12 <sipa> Luke-Jr: why does your client turn ... into ??? ?
542 2012-10-21 14:40:41 <sipa> i was already wondering whether the source code containing unicode could be Diapolo's issue
543 2012-10-21 14:41:06 <Luke-Jr> sipa: ???
544 2012-10-21 14:41:35 <sipa> Diapolo: if your remove the comment-out, it still crashes reproducibly?
545 2012-10-21 14:42:12 <Luke-Jr> my theory is that InitMessage's QApplication::instance()->processEvents() can't be called from other threads
546 2012-10-21 14:42:14 <Diapolo> sipa: commenting out uiInterface.InitMessage(_("Starting block import...")); -> no crash, leaving this in -> crash
547 2012-10-21 14:43:04 <sipa> hmm, ok
548 2012-10-21 14:43:11 <sipa> makes it kind of useless, no?
549 2012-10-21 14:43:13 <Diapolo> Luke-Jr: perhaps because it's not locked?
550 2012-10-21 14:43:21 <Diapolo> multi-thread issue?
551 2012-10-21 14:43:31 <Luke-Jr> sipa: no, up till now, it was only used from the init thread
552 2012-10-21 14:43:45 <jgarzik> it might be nice to produce some unofficial ultraprune Windows test builds, for the testing crowd. maybe add an additional "this is really really really only a test version" additional popup messagebox at startup.
553 2012-10-21 14:43:47 <Luke-Jr> sipa: it doesn't make sense to do it in ThreadImport IMO since the GUI is starting in parallel
554 2012-10-21 14:44:28 <Luke-Jr> jgarzik: I'll probably make a next build with it when the other pullreqs are rebased (for next-test)
555 2012-10-21 14:44:34 <Diapolo> Luke-Jr: Could we just place a LOCK() for the initMessage?
556 2012-10-21 14:44:48 <Luke-Jr> Diapolo: doubt it
557 2012-10-21 14:44:50 <sipa> Diapolo: that doesn't make it run in another thread, does it?
558 2012-10-21 14:45:11 <sipa> Luke-Jr: ok, the message can be moved to AppInit2, i guess
559 2012-10-21 14:45:19 <Luke-Jr> isn't the point of ThreadImport so that it runs after the GUI is done loading basically?
560 2012-10-21 14:45:39 <Luke-Jr> InitMessage only affects the splash screen *before* the GUI is loaded
561 2012-10-21 14:45:40 <sipa> hmm, right - it already provides feedback anyway
562 2012-10-21 14:45:46 <sipa> through the progressbar
563 2012-10-21 14:45:49 <Luke-Jr> so it only makes sense to use it in the main init thread
564 2012-10-21 14:45:58 <sipa> ok, just delete the uiInterface things
565 2012-10-21 14:46:05 <sipa> in ThreadImport
566 2012-10-21 14:46:17 <Diapolo> I will check back later, thanks for all your leveldb help sipa and D34TH!
567 2012-10-21 14:47:05 <Diapolo> btw. src/txdb.h is included twice in the project file for Bitcoin-Qt
568 2012-10-21 14:47:55 <sipa> jgarzik: i think in general having daily builds of git HEAD would be nice
569 2012-10-21 14:56:44 <jgarzik> sipa: BlueMatt used to do nightly builds, in fact
570 2012-10-21 14:57:04 <sipa> yes, but his jenkins seems a bit overworked
571 2012-10-21 15:19:51 <darkip> Is there any way to subscribe to the bitcoin client for block find notifications?
572 2012-10-21 15:20:08 <sipa> -blocknotify
573 2012-10-21 15:20:40 <darkip> oh awesome, thanks
574 2012-10-21 15:20:52 <darkip> and thanks for the help with the hash rate stuff the other day
575 2012-10-21 15:21:09 <darkip> just written the script to calculate it all this weekend :)
576 2012-10-21 15:22:37 <sipa> what formula? just average in a block window?
577 2012-10-21 15:23:04 <darkip> Yeah, average over time period/blocks
578 2012-10-21 15:23:56 <darkip> What is regarded as a reasonable window size to avoid luck being too much of a factor?
579 2012-10-21 15:24:04 <sipa> it's a delicate balance
580 2012-10-21 15:24:18 <sipa> a longer window means you're no longer measuring the current speed, but the speed some time ago
581 2012-10-21 15:24:26 <sipa> but it also reduces variance
582 2012-10-21 15:24:47 <darkip> I'm currently looking at the last day
583 2012-10-21 15:37:11 <gmaxwell> < Joric> someone made a front page! < please don't feed atlas
584 2012-10-21 15:40:14 <sipa> well, at least he changed the thread title to something less dramatic
585 2012-10-21 15:45:20 <sipa> hmm, the thread is also locked; maybe some mod changed the title?
586 2012-10-21 15:46:35 <gmaxwell> sipa: under that name. He also started two other threads, one about pruning destroying the network, and one with a title like the bitcointalk one.
587 2012-10-21 15:47:29 <sipa> what are you talking about?
588 2012-10-21 15:47:58 <gmaxwell> sipa: my comment to joric was about reddit posts. (I realize that may not be clear now)
589 2012-10-21 15:48:25 <sipa> i realized you were talking about reddit, but i was talking about bitcointalk.org
590 2012-10-21 15:49:30 <BlueMatt> sipa: waiting on resolution of a jdk bug to fix jenkins now :(
591 2012-10-21 15:49:46 <gmaxwell> sipa: yea, I didn't follow your initial response??? I thought you were saying the reddit thread had a better name.
592 2012-10-21 15:50:21 <sipa> gmaxwell: no, i noticed Atlas' original FUD thread on bitcointalk.org had changed name, while also having become locked
593 2012-10-21 15:50:58 <BlueMatt> maybe ill just give up and write a third dnsseed based on pynode...
594 2012-10-21 15:52:16 <sipa> what's wrong with mine? (not that i have a problem with more variety in dns seeding software)
595 2012-10-21 15:57:09 <maaku> BlueMatt: sipa's works really well
596 2012-10-21 15:58:00 <sipa> well, there are some flaws - i want to reduce the cpu usage, make it schedule retries better, and make it check for actual blocks
597 2012-10-21 15:58:41 <sipa> still, that's less work than writing one from scratch, i think
598 2012-10-21 15:58:49 <sipa> but if you enjoy doing that, please do :)
599 2012-10-21 15:59:39 <maaku> sipa: btw, my alt chain branch has been working fine for a few weeks now
600 2012-10-21 16:01:05 <sipa> ok, good to hear
601 2012-10-21 16:01:58 <TD> good evening
602 2012-10-21 16:02:18 <BlueMatt> sipa: its either write my own (again) which provides variety in dns seed software or track down the 100% cpu bug in yours (which I dont think should be too hard, but Id rather go for the variety)
603 2012-10-21 16:02:29 <sipa> TD: hi there
604 2012-10-21 16:02:37 <BlueMatt> sipa: and writing one from scratch with bitcoinj was really easy...until I hit jdk bugs
605 2012-10-21 16:03:30 <sipa> ic
606 2012-10-21 16:12:48 <BlueMatt> TD: hey
607 2012-10-21 16:14:14 <BlueMatt> TD: I did a few minor refactors to my branch over the weekend (rebased onto your master, tweaked a few things, etc)
608 2012-10-21 16:15:14 <TD> cool
609 2012-10-21 16:15:22 <TD> i was busy last week but i plan to take another look this week
610 2012-10-21 16:15:29 <BlueMatt> np
611 2012-10-21 16:16:14 <BlueMatt> (working on reverse headers sync now...its really easy to get working at the application level, but moving it into PeerGroup is a pain, plus I havent thought much about how to deal with wallets there...)
612 2012-10-21 16:19:25 <BlueMatt> TD: will a wallet blow up, that you know of, if you tell it that you downloaded the chain twice (once without txes, once with)?
613 2012-10-21 16:19:39 <TD> i think it will get corrupted, yes. it's not designed to have blocks replayed into it
614 2012-10-21 16:20:04 <BlueMatt> alright
615 2012-10-21 16:30:00 <gmaxwell> hmph, after salvagewalleting that ultraprune mangled wallet, I'm now getting a nice generic "10/21/12 18:05:32 Bitcoin: Cannot initialize keypool" and then the node stops cleanly.
616 2012-10-21 16:34:38 <sipa> gmaxwell: nice catch
617 2012-10-21 16:34:47 <sipa> -salvagewallet should set fFirstRun to false
618 2012-10-21 16:36:58 <sipa> or at least the keypool init shouldn't be done
619 2012-10-21 16:50:36 <sipa> gmaxwell: does my patch fix the zero-balance problem?
620 2012-10-21 16:51:51 <gmaxwell> sipa: yes, it appears to... I'm not completely done resyncing the chain yet, but it's part way done and the balance is not zero.
621 2012-10-21 16:52:13 <gmaxwell> (I tried a number of other things first, e.g. the rescan and salvagewallet)
622 2012-10-21 16:52:43 <sipa> gmaxwell: well, the bug that was fixed (which may or may not be the entire issue) would have caused to go to zero immediately at startup, and never change
623 2012-10-21 16:53:15 <gmaxwell> sipa: ah, I wasn't sure if it was at startup or during resync since I reproduced it before with a chainless node.
624 2012-10-21 16:53:32 <gmaxwell> In any case, yes it's fixed.
625 2012-10-21 16:56:50 <sipa> seems i can keep a node on valgrind keep up with the network, without problems :)
626 2012-10-21 16:57:28 <gmaxwell> sipa: yep. I have a couple that run that way.
627 2012-10-21 17:09:44 <Diapolo> Re, any news on that InitMessage bug?
628 2012-10-21 17:10:24 <sipa> Diapolo: yes, i'll just remove those uiMessages
629 2012-10-21 17:11:33 <Diapolo> If you want I can create that patch, so you can work on more important stuff?
630 2012-10-21 17:11:56 <sipa> go ahead
631 2012-10-21 17:14:21 <sipa> jgarzik: i'm reworking your -reindex patch on top of ultraprune
632 2012-10-21 17:16:02 <Diapolo> sipa: shouldn't the ThreadImport() also reside in a try-catch like in other threads we use?
633 2012-10-21 17:17:23 <sipa> Diapolo: it would be cleaner, but it's safe now; LoadExternalBlockFile has a try-catch around the essential part
634 2012-10-21 17:26:20 <yellowhat> short quistion: where can i find the pgp signed announcment for 0.7.1 ?
635 2012-10-21 17:36:06 <Diapolo> sipa: as it's not worth further discussion, my "blockchain -> block chain" pull is now non-translation changing, are you fine with just merging it to easen the progress?
636 2012-10-21 17:36:20 <Diapolo> LOL thanks ^^
637 2012-10-21 17:36:24 <gmaxwell> yellowhat: there doesn't appear to be a signed _announcement_ there are the signed release files however.
638 2012-10-21 17:36:53 <yellowhat> so how do you check it practically
639 2012-10-21 17:37:35 <gmaxwell> yellowhat: you check the actual files.
640 2012-10-21 17:38:02 <gmaxwell> http://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.7.1/ < SHA256SUMS.asc is the signed hashes
641 2012-10-21 17:43:13 <midnightmagic> or you accept that the person who signs things is the same one that has commit access to the github repository and you ask politely for all the devs to start signing off on all their changes automatically. :-)
642 2012-10-21 17:44:59 <gmaxwell> midnightmagic: the tags are signed in the repository.
643 2012-10-21 17:45:09 <midnightmagic> that's pretty awesome. I didn't know that.
644 2012-10-21 17:45:31 <midnightmagic> my git is old enough it doesn't know how to git log --show-signature
645 2012-10-21 18:07:58 <Diapolo> gmaxwell: Thanks for the single Testnetnode which is reacheable via Tor HS ^^.
646 2012-10-21 18:08:21 <Diapolo> I had my testnet client running for hours and now it's getting blocks from you ^^.
647 2012-10-21 18:10:29 <Diapolo> sipa: Seems I just got an Assertian failure in main.cpp 1530 from assert(pindex->pprev == view.GetBestBlock()); while doing an initial chain-sync. Can't tell if it was the testnet oder mainnet client though.
648 2012-10-21 18:11:06 <Diapolo> alright, it was the testnet client :D as that one is gone now after clicking OK
649 2012-10-21 18:12:00 <sipa> Diapolo: that shouldn't happen anyway...
650 2012-10-21 18:13:29 <Diapolo> just wanted to mention it any idea?
651 2012-10-21 18:13:55 <sipa> you've given me very little context
652 2012-10-21 18:15:06 <Diapolo> init chain sync, Win 7 x64 own build with my core proxy patches applied, -onlynet="Tor" ... it was just sitting in my system tray
653 2012-10-21 18:15:18 <Diapolo> and afaik I had setgenerate true 4
654 2012-10-21 18:18:28 <root2> gmaxwell: Bitcoin uses ECDSA with 256-bit keys correct?
655 2012-10-21 18:18:48 <sipa> root2: yes, it uses the secp256k1 curve
656 2012-10-21 18:19:21 <root2> is ECDSA amenable to key lengthening? Or, is there a suitable 512 bit curve to use?
657 2012-10-21 18:19:40 <sipa> there are other curves, yes, with different sizes
658 2012-10-21 18:19:52 <sipa> not sure what you mean by key lengthening
659 2012-10-21 18:20:18 <root2> can the algorithm be easily adjusted for arbitrary (power of two) length keys?
660 2012-10-21 18:20:41 <sipa> you need a specific curve - there are some predefined ones, or you can generate your own
661 2012-10-21 18:20:51 <sipa> it's not as easy as just saying : use size X"
662 2012-10-21 18:20:53 <Titanium2> but what if I want a 1428.75 bit key?
663 2012-10-21 18:21:15 <sipa> also, changing the signature algorithm is not particularly easy in bitcoin
664 2012-10-21 18:21:26 <Luke-Jr> root2: the question is, is there a good reason to use a different curve ever?
665 2012-10-21 18:21:29 <sipa> it's possible, and even without a hrd fork, but that's about it
666 2012-10-21 18:21:41 <root2> wouldnt 512 be sufficent? if security is equal to half the key size, 256 should be sufficient for a looong time
667 2012-10-21 18:22:10 <gmaxwell> root2: 128 should also be sufficient, absent currently unknown attacks??? which may be just as effective for 512.
668 2012-10-21 18:22:14 <gmaxwell> er 256.
669 2012-10-21 18:22:20 <Luke-Jr> root2: I'm not aware of any known vulnerabilities in ECDSA, and when one is found, there is no reason to think merely using a different curve would help
670 2012-10-21 18:22:20 <root2> well ECDSA with 256 bit keys takes only 2^128 to brute force, yes?
671 2012-10-21 18:22:25 <Titanium2> when you ahve millions of dollars being protected, you want more bits
672 2012-10-21 18:22:26 <sipa> root2: correct
673 2012-10-21 18:22:53 <root2> im not speaking of vulnerabilities in the algorithm, just the difficulty of the brute force
674 2012-10-21 18:23:01 <sipa> root2: and the entire bitcoin network hasn't performed 2^70 sha iterations even
675 2012-10-21 18:23:03 <gmaxwell> root2: just incrementing a non-reversable binary counter 2^128 times (much less doing something with it) is 'start harnessing all the output of the sun' level computing.
676 2012-10-21 18:23:24 <Luke-Jr> root2: use multisig. 3-of-3 is 256^256^256 ;)
677 2012-10-21 18:23:27 <Titanium2> it costs about nothign to use more bits
678 2012-10-21 18:23:38 <Titanium2> a bit of storage space, and a bit of work when creating a block
679 2012-10-21 18:23:51 <gmaxwell> Generally 128 bit security is considered the a solid standard of brute force unreachability.
680 2012-10-21 18:23:58 <root2> I was aware that all of our sun's energy would be insufficient to count to 2^256, but i was unaware that 128 fell in the same range
681 2012-10-21 18:24:09 <jgarzik> sipa: sigh. what was Atlas's thread?
682 2012-10-21 18:24:39 <sipa> jgarzik: https://bitcointalk.org/index.php?topic=119566.0
683 2012-10-21 18:25:04 <Luke-Jr> jgarzik: what *wasn't* Atlas's thread?
684 2012-10-21 18:25:10 <gmaxwell> root2: all the sun's energy is enough to count to 2^128 under an ideally efficient non-reversable computer model. But not a computer buildable by all the current efforts of man.
685 2012-10-21 18:26:46 <gmaxwell> Titanium2: in any case, it's not cheap. If it had originally been larger we'd be talking about half the scalability / decenteralization tradeoff. And fitting it in later would require carrying both, which is ugly. And just upping the ECC level doesn't forclose QC oriented attacks or attacks from mathmatical breakthroughs that break ecc.
686 2012-10-21 18:27:15 <Diapolo> sipa: anything more I can give you for that assertion issue?
687 2012-10-21 18:28:02 <root2> now is that taking into account the speedups associated with quantum computing? My understanding of shors algorithm and grovers algorithm indicates that superposition of states is being used, and that these might represent more than a traditional bit with lower energy costs fordiscrete state transitions
688 2012-10-21 18:28:07 <gmaxwell> Diapolo: can you reproduce it?
689 2012-10-21 18:28:09 <root2> but, i could be confused
690 2012-10-21 18:28:56 <gmaxwell> root2: I have been blinded by your word salad. Can you try asking that another way? :P
691 2012-10-21 18:29:04 <Diapolo> gmaxwell: currently no, I had a guess that I perhaps generated a testnet block via setgenerate, while doing the chain sync, can this be the cause?
692 2012-10-21 18:29:19 <Luke-Jr> sipa: re the optional extra index, maybe it should maintain info needed to replace blockchain.info with a GUI page?
693 2012-10-21 18:29:37 <Luke-Jr> sipa: eg, balances of arbitrary scriptPubKeys
694 2012-10-21 18:30:09 <sipa> Luke-Jr: that's one step further even, and I don't like having things depend on the availability of "balance of an address"
695 2012-10-21 18:30:43 <gmaxwell> root2: er, I can try answering. QC is magical unicorn land right now, it may turn out that there are physical limits that make building the kind of very large QC's required for cracking asymetric crypto infeasable; they're certantly pratically infeasable now. But assuming that they become possible to build
696 2012-10-21 18:30:47 <Luke-Jr> sipa: users obviously want to be able to see that info though
697 2012-10-21 18:31:00 <sipa> Luke-Jr: unfortunately
698 2012-10-21 18:31:23 <Luke-Jr> root2: afaik, QC in theory breaks ECDSA entirely
699 2012-10-21 18:31:29 <sipa> but if a full txid-to-diskpos index is optionally available, while not go all the way and do it for address-to-txid as well
700 2012-10-21 18:32:06 <gmaxwell> root2: ecdsa??? with really??? any length is insecure against a sufficiently large QC, the slowdown from additional size would be too low to help. The same issue is not true for hash functions because grovers gives a tightly bounded sqrt() speedup on general non-linear search.
701 2012-10-21 18:32:33 <gmaxwell> root2: If very large QC's became a concern we wouldn't deploy larger ecdsa we'd deploy lamport signatures.
702 2012-10-21 18:32:44 <jgarzik> sipa: seeing as how people are jumping on ultraprune testing (great!), we should probably go ahead and merge the wallet fix
703 2012-10-21 18:33:05 <jgarzik> ACTION hasn't had a chance to verify it yet... today was family pumpkin patch day ;p
704 2012-10-21 18:33:40 <gmaxwell> Unforuntately lamport signatures are very large. The ultraprune plus UTXO set ideas show how bitcoin could be viable with signatures as large as lamport ones... but its not something that would make sense speculative. The loss of scalablity and decenteralization that comes with block bloat is a bigger risk than pixie dust computers.
705 2012-10-21 18:34:06 <root2> gmaxwell: Cool, Grovers just had me tweaked a bit, 128 bits can be searched in 2^64, thats not unreasonable for a government or large corporate entity to undertake in the next few years
706 2012-10-21 18:34:46 <root2> is NTRUsign viable post-quantum?
707 2012-10-21 18:34:58 <sipa> Luke-Jr: the motivation behind ultraprune is making resources scale with the size of the UTXO set, and not with the size of history
708 2012-10-21 18:35:04 <root2> i dont know many QC secure assymetric algos
709 2012-10-21 18:35:21 <root2> asymetric
710 2012-10-21 18:35:21 <sipa> Luke-Jr: an index for the full history breaks that
711 2012-10-21 18:35:32 <gmaxwell> root2: you wouldn't use grovers for ecdsa. On a sufficiently large QC (enormous) you can break 256 bit ecdsa in ~2^32 operations.
712 2012-10-21 18:35:40 <sipa> Luke-Jr: it's certainly useful for debugging, and that is why i'd consider adding it
713 2012-10-21 18:36:02 <sipa> Luke-Jr: but apart from that, i want to discourage any use that depends on the availability of such an index
714 2012-10-21 18:36:36 <Luke-Jr> sipa: a blockchain.info replacement (ie, no more web blockchain viewer required) needs it AFAIK
715 2012-10-21 18:36:46 <gmaxwell> root2: NTRUsign is fairly unattractive for multiple reasons??? for one, it's not obviously as _classically_ secure as ecdsa. My recommendation is lamport. It is intutively as strong as any other pratical signature scheme implementation.
716 2012-10-21 18:37:31 <sipa> Luke-Jr: of course it does - i consider that 'debug use'; using it for anything else (like fast rescanning for "the blockchain is your wallet!" purposes) is a problem imho
717 2012-10-21 18:37:56 <Diapolo> gmaxwell: a simple no your idea is stuipd or yes could be, would be nice ^^ even in between a Quantum computer discussion ^^
718 2012-10-21 18:38:15 <gmaxwell> Diapolo: I don't know. It sounds like a guess.
719 2012-10-21 18:38:28 <gmaxwell> Diapolo: I think we already know the internal miner is somewhat buggy.
720 2012-10-21 18:39:57 <gmaxwell> Our ecosystem dependency on blockchain.info (esp now that blockexplorer is pretty much dying) bugs me but I'm not sure what to do about it. Even with good integrated 'explorer' functionality in the client people will still use the website out of convience.
721 2012-10-21 18:40:50 <Luke-Jr> gmaxwell: I don't see how a website is more convenient
722 2012-10-21 18:41:06 <Luke-Jr> gmaxwell: a well-indexed local client can be much more responsible and user friendly
723 2012-10-21 18:41:19 <Diapolo> gmaxwell: I always loved to use it for a quick testnet-block generation ... if it's bugged it needs fixing or removal IMHO.
724 2012-10-21 18:41:21 <gmaxwell> Luke-Jr: because I can just give you a url to a website.
725 2012-10-21 18:41:30 <Luke-Jr> gmaxwell: bitcoin URIs version 2 :p
726 2012-10-21 18:41:34 <gmaxwell> Diapolo: patches accepted! :P
727 2012-10-21 18:42:38 <sipa> Luke-Jr: now THAT is exactly what I don't want to see
728 2012-10-21 18:42:54 <gmaxwell> Hopefully luke does the ui integration for BFG miner and we could just remove the internal miner and package bfgminer with bitcoin.
729 2012-10-21 18:42:57 <Diapolo> gmaxwell: I don't know what's broken + I really don't want to mess with mining internals :).
730 2012-10-21 18:42:59 <sipa> as it means a community that learns to depend on the availability of all historic transaction data
731 2012-10-21 18:43:13 <sipa> well... i suppose we already have that anyway
732 2012-10-21 18:43:35 <gmaxwell> sipa: we do??? though they're used to picking it up from a couple of centeralized websites.
733 2012-10-21 18:44:14 <gmaxwell> At a _minimum_ we should try to avoid making it hard to run an explorer sites... so there at least will be lots of choice.
734 2012-10-21 18:44:31 <gmaxwell> e.g. having to write your own node software to accomplish that kinda stinks.
735 2012-10-21 18:45:14 <jgarzik> gmaxwell: bleh
736 2012-10-21 18:45:18 <jgarzik> gmaxwell: I like the internal miner
737 2012-10-21 18:45:54 <jgarzik> We already package pyminer with the reference client ;-)
738 2012-10-21 18:45:56 <jgarzik> ACTION runs
739 2012-10-21 18:46:02 <sipa> we do? :o
740 2012-10-21 18:46:17 <Luke-Jr> sipa: in contrib, but it's trivial
741 2012-10-21 18:46:24 <jgarzik> sipa: all sorts of strange gadgetry in contrib/
742 2012-10-21 18:46:36 <sipa> yah
743 2012-10-21 18:47:15 <sipa> anyway, it's a balance; a community depending on historic data via a centralized site is worse than one that depends on such data being available locally, but both are significantly worse than not depending on either
744 2012-10-21 18:49:38 <Luke-Jr> O.o
745 2012-10-21 18:49:56 <Luke-Jr> Walmart photo CDs (from film) have the Qt DLLs in the root
746 2012-10-21 18:52:38 <sipa> jgarzik: making AcceptBlock deal with the genesis block is fun
747 2012-10-21 18:52:45 <sipa> it fails several tests :p
748 2012-10-21 18:52:57 <Luke-Jr> sipa: also unsafe ;)
749 2012-10-21 18:53:02 <sipa> ?
750 2012-10-21 18:53:22 <Luke-Jr> sipa: you might index the coinbase txn!
751 2012-10-21 18:53:26 <sipa> i just added an if (hash != hashGenesisBlock) { DO_TESTS }
752 2012-10-21 18:53:49 <sipa> Luke-Jr: thanks for reminding me to check that
753 2012-10-21 18:54:12 <jgarzik> sipa: eh?
754 2012-10-21 18:54:39 <sipa> jgarzik: to prevent the genesis block being added on every -reindex, you want to reuse the existing one on disk
755 2012-10-21 18:54:44 <Luke-Jr> hmm, we should probably document that explicit somewhere
756 2012-10-21 18:54:49 <jgarzik> sipa: In the context of -reindex? In the old configuration, -reindex always added a new genesis block as a consequence of LoadBlockIndex
757 2012-10-21 18:54:51 <jgarzik> sipa: ah ok
758 2012-10-21 18:55:25 <jgarzik> sipa: a new genesis block seemed like a small cost, compared to additional code just for one block
759 2012-10-21 18:55:30 <jgarzik> ACTION dislikes special case code
760 2012-10-21 18:55:46 <sipa> jgarzik: well, it's very hard to make the current block file code deal with it otherwise
761 2012-10-21 18:56:09 <sipa> as there is no simple "append to block file" logic anymore, it picks a file and position based on the metadata about the block files
762 2012-10-21 18:56:25 <sipa> without a database, that means at the beginning of the first file, resulting in a pre-allocation at that point
763 2012-10-21 18:56:32 <sipa> overwriting the first few thousand blocks
764 2012-10-21 18:56:59 <sipa> Luke-Jr: the block connection logic is already special cased to ignore connecting the genesis block
765 2012-10-21 18:57:33 <conman> yo jgarzik
766 2012-10-21 18:57:44 <jgarzik> sipa: what is preallocation, specifically? I saw that in log messages, and it raised my curiosity. Do you write zeroes?
767 2012-10-21 18:57:47 <jgarzik> conman: word
768 2012-10-21 18:57:52 <sipa> jgarzik: yes - for now
769 2012-10-21 18:58:12 <sipa> jgarzik: there is a function in util.cpp for that - it's not required to overwrite the specified range, but allowed to
770 2012-10-21 18:58:24 <jgarzik> wow
771 2012-10-21 18:58:24 <sipa> it can be overridden with OS-specific implementations
772 2012-10-21 18:58:30 <jgarzik> orphan map 100% full, here
773 2012-10-21 18:58:35 <jgarzik> block height is correct
774 2012-10-21 18:58:46 <jgarzik> 10/21/12 20:58:13 stored orphan tx 51d305bbd9 (mapsz 10001)
775 2012-10-21 18:58:49 <jgarzik> been like that for hours
776 2012-10-21 18:59:04 <sipa> are there still transactions being accepted?
777 2012-10-21 18:59:08 <jgarzik> yes
778 2012-10-21 18:59:20 <jgarzik> bitcoind is otherwise functioning normally, AFAICT
779 2012-10-21 18:59:32 <Diapolo> sipa: I'd like to look into OS specific code for the pre-alloc on Windows over the next weeks.
780 2012-10-21 18:59:41 <sipa> Diapolo: it's very easy :)
781 2012-10-21 19:00:13 <Diapolo> Because of your AllocateFileRange()?
782 2012-10-21 19:00:23 <jgarzik> sipa: later on, when the dust settles, I might write a pullreq for: #ifdef linux ... posix_fallocate()