1 2013-10-27 00:00:04 <MrDaneelOlivaw> this is the full https://dpaste.de/Yqdu with boost dev installed
2 2013-10-27 00:00:32 <Apocalyptic> skinnkavaj, wrong channel
3 2013-10-27 00:06:39 <gmaxwell> MrDaneelOlivaw: It will build with --disable-tests I expect.
4 2013-10-27 00:07:42 <gmaxwell> MrDaneelOlivaw: can you try that?
5 2013-10-27 00:07:59 <Luke-Jr> skinnkavaj: designers and developers are usually two not-very-intersecting groups
6 2013-10-27 00:10:47 <gmaxwell> MrDaneelOlivaw: I'm taking a WAG that it's actually the next test failing and some dangling variable is causing it to print the wrong cause.
7 2013-10-27 00:13:11 <MrDaneelOlivaw> gmaxwell: trying
8 2013-10-27 00:13:44 <MrDaneelOlivaw> I may have copied the libs dependencies from the package in the repo which probably was built that way
9 2013-10-27 00:15:01 <gmaxwell> MrDaneelOlivaw: you should have, as the fine manual states, done an apt-get install libboost-all-dev
10 2013-10-27 00:24:50 <MrDaneelOlivaw> gmaxwell: somehow libboost-all-dev didn't fix it but installing the -dev as well as .0 package of test fixed it (likewise --disable-test worked) thanks a lot
11 2013-10-27 00:26:29 <MrDaneelOlivaw> gmaxwell: I'm interested in using the regtest functionality, I understand testnet in a box is not used much anymore
12 2013-10-27 00:33:55 <MrDaneelOlivaw> not sure if it's funny, need to finish it, but it is something (no woman, no drive) http://www.youtube.com/watch?v=aZMbTFNp4wI
13 2013-10-27 00:38:27 <MrDaneelOlivaw> gmaxwell: actually WARNING: unrecognized options: --disable-test, but test package fixed it
14 2013-10-27 00:38:59 <MrDaneelOlivaw> ouch, typo, sorry.
15 2013-10-27 00:40:29 <warren> are there any existing tools to read peers.dat?
16 2013-10-27 00:41:49 <sipa> gmaxwell: http://bitcoin.stackexchange.com/a/14092/208
17 2013-10-27 00:41:54 <sipa> warren: not that i know
18 2013-10-27 00:42:51 <sipa> hmm, -debug is painful:
19 2013-10-27 00:42:53 <sipa> $ ls -lh ~/.bitcoin/debug.log
20 2013-10-27 00:42:54 <sipa> -rw------- 1 pw pw 20G Oct 27 02:42 /home/pw/.bitcoin/debug.log
21 2013-10-27 00:45:15 <warren> :q
22 2013-10-27 00:45:18 <warren> oops
23 2013-10-27 00:46:36 <gmaxwell> sipa: yea... that reddit thread is full of people linking to tools that report "entropy" of hundreds of bits for stupid phrases that have been cracked as brainwallets.
24 2013-10-27 00:47:16 <sipa> warren: i should write a python tool for it
25 2013-10-27 00:47:35 <sipa> warren: if i have time somewhere the next 5 years :)
26 2013-10-27 00:50:22 <gmaxwell> MrDaneelOlivaw: ".0 package of test" which package is this?
27 2013-10-27 00:52:11 <MrDaneelOlivaw> gmaxwell: libboost-{system,filesystem,chrono,program-options,thread,test}1.54{-dev,.0}
28 2013-10-27 00:52:46 <sipa> the dev packages don't depend on their corresponding non-dev packages?
29 2013-10-27 00:52:58 <MrDaneelOlivaw> sipa: yeah but a different version
30 2013-10-27 00:53:04 <sipa> eww
31 2013-10-27 00:53:08 <MrDaneelOlivaw> 1.53
32 2013-10-27 00:53:20 <sipa> that's asking for trouble
33 2013-10-27 00:53:38 <MrDaneelOlivaw> no sorry i thought you were talking about the generic -dev version
34 2013-10-27 00:53:49 <sipa> no, what i mean
35 2013-10-27 00:54:02 <sipa> i infer that you already had some -dev package installed
36 2013-10-27 00:54:08 <MrDaneelOlivaw> i guess it does, maybe i've been too explicit
37 2013-10-27 00:54:21 <sipa> but only when you also installed the corresponding .o package, the problem was fixed?
38 2013-10-27 00:54:25 <sipa> .0
39 2013-10-27 00:54:27 <MrDaneelOlivaw> no
40 2013-10-27 00:54:30 <MrDaneelOlivaw> i needed test
41 2013-10-27 00:54:53 <MrDaneelOlivaw> just saying that the libboost-all-dev didn't do it for me
42 2013-10-27 00:55:03 <sipa> you didn't have libboost-test-dev?
43 2013-10-27 00:55:09 <MrDaneelOlivaw> i did
44 2013-10-27 00:55:12 <sipa> even with libboost-all-dev?
45 2013-10-27 00:55:13 <MrDaneelOlivaw> and it didn't work
46 2013-10-27 00:55:33 <sipa> oh, you didn't have libboost-test?
47 2013-10-27 00:56:57 <MrDaneelOlivaw> when i first started no, i just checked what bitcoin dependencies where on the bitcoin ubuntu package and installed the dev parts assuming dependencies had not changed too much, which is what worked for me on 0.8.5
48 2013-10-27 00:57:12 <sipa> ah, i see
49 2013-10-27 00:57:21 <sipa> but libboost-dev-all should have fixed it then?
50 2013-10-27 00:57:29 <sipa> or does -all not depend on -test?
51 2013-10-27 00:58:23 <MrDaneelOlivaw> I don't know, when i tried the above and it worked i was happier as it was an even later version anyway
52 2013-10-27 00:59:04 <gmaxwell> Matters to us because our instructions say to install -all.
53 2013-10-27 00:59:15 <gmaxwell> If that doesn't work we will need to fix them.
54 2013-10-27 00:59:16 <MrDaneelOlivaw> I can try again
55 2013-10-27 00:59:20 <MrDaneelOlivaw> one sec
56 2013-10-27 00:59:27 <gmaxwell> Might also be good to figure out how to make the error more informative.
57 2013-10-27 00:59:27 <MrDaneelOlivaw> just a small change in my docker file
58 2013-10-27 00:59:51 <MrDaneelOlivaw> disabled test or not =
59 2013-10-27 00:59:54 <gmaxwell> I was going on "well, the reported error cannot be true, so maybe its the _next_ thing that failed."
60 2013-10-27 01:00:05 <gmaxwell> MrDaneelOlivaw: don't disable.
61 2013-10-27 01:08:23 <MrDaneelOlivaw> gmaxwell: yes it fails, let me upload the file for you
62 2013-10-27 01:17:24 <MrDaneelOlivaw> gmaxwell: http://dpaste.de/Qjhh
63 2013-10-27 01:19:30 <MrDaneelOlivaw> libboost-all-dev is already the newest version.
64 2013-10-27 01:23:43 <gmaxwell> MrDaneelOlivaw: what message did that fail with on stdout?
65 2013-10-27 01:24:03 <midnightmagic> I have a /window goto 16
66 2013-10-27 01:25:59 <MrDaneelOlivaw> gmaxwell: configure: error: Could not find a version of the library!
67 2013-10-27 01:26:26 <gmaxwell> can you post the config that sets up your vm? (e.g. with the apt-get lines?)
68 2013-10-27 01:31:46 <MrDaneelOlivaw> gmaxwell: https://dpaste.de/7KyN
69 2013-10-27 01:47:35 <MrDaneelOlivaw> is there a lot of people thinking bitcoin was created by a group of conspirator with tremendous insight as opposed to one guy called Satoshi ? I was reading an article from schneider and I wonder if the weakness is technical or simply by design (mining)
70 2013-10-27 01:54:06 <Luke-Jr> MrDaneelOlivaw: what weakness?
71 2013-10-27 02:00:11 <MrDaneelOlivaw> Luke-Jr: eventually mining could be come even more centralized but we wouldn't even be able to know for sure and it is quite unaccountable
72 2013-10-27 02:00:36 <Luke-Jr> MrDaneelOlivaw: there isn't really a better alternative
73 2013-10-27 02:00:44 <Luke-Jr> thankfully, things aren't playing out that way
74 2013-10-27 02:02:49 <MrDaneelOlivaw> Luke-Jr: alternatively there could be some weakness in crypto which we are not aware off, just like DES and differential cryptoanalysis
75 2013-10-27 02:03:02 <MrDaneelOlivaw> Luke-Jr: I agree we don't have a better alternative
76 2013-10-27 02:04:11 <MrDaneelOlivaw> I've been following your work for a while Luke-Jr, yours and a few others. I feel i should congratulate with you
77 2013-10-27 02:04:51 <Luke-Jr> thanks
78 2013-10-27 02:05:51 <MrDaneelOlivaw> Luke-Jr: what are you working on these days if I may ask ?
79 2013-10-27 02:06:07 <Luke-Jr> MrDaneelOlivaw: BFGMiner's been keeping me busy lately; lots of new devices coming out still
80 2013-10-27 02:06:20 <Luke-Jr> and plenty of room to improve the basic program
81 2013-10-27 02:08:09 <MrDaneelOlivaw> I've stopped mining a month back, sold my devices and moved on. It's not my thing, moves too fast. I feel it will be less and less accessible and margins smaller and smaller (which is just natural)
82 2013-10-27 02:08:45 <MrDaneelOlivaw> more interested in the recent protocol changes like p2sh and this payment protocol, although i still have to study the CA side of things
83 2013-10-27 02:09:16 <MrDaneelOlivaw> Is there a roadmap as far as you know ?
84 2013-10-27 02:09:28 <Luke-Jr> in various peoples' heads âº
85 2013-10-27 02:09:45 <MrDaneelOlivaw> isn't it always :P
86 2013-10-27 02:10:08 <MrDaneelOlivaw> do you have a view on btcd ? tried it ?
87 2013-10-27 02:10:30 <MrDaneelOlivaw> and thoughts on electrum
88 2013-10-27 02:10:38 <MrDaneelOlivaw> I'm talking about the design-performance-etc
89 2013-10-27 02:12:36 <Luke-Jr> MrDaneelOlivaw: I haven't tried btcd yet. Electrum is still (last I checked) old Python2, but I hear it's a SPV client now.
90 2013-10-27 02:13:12 <Ascendion> grrrr
91 2013-10-27 02:13:56 <MrDaneelOlivaw> Luke-Jr: yeah i read that too but what do you mean by 'old' python2, python2 is still used far more than 3.x
92 2013-10-27 02:14:08 <Luke-Jr> MrDaneelOlivaw: it's still old/outdated ;)
93 2013-10-27 02:14:11 <MrDaneelOlivaw> Ascendion: writing an electrum competitr ?
94 2013-10-27 02:14:21 <Luke-Jr> Python 3.x has been around a while now, and I use it almost exclusively.
95 2013-10-27 02:14:26 <Luke-Jr> (well, for Python)
96 2013-10-27 02:14:39 <MrDaneelOlivaw> Luke-Jr: I need to check if everything is available
97 2013-10-27 02:14:52 <Luke-Jr> ?
98 2013-10-27 02:15:01 <MrDaneelOlivaw> for 3.x of what i use in python 2.x
99 2013-10-27 02:15:10 <MrDaneelOlivaw> anyhow, it'd be interesting to see pypy move to 3.x
100 2013-10-27 02:15:47 <Ascendion> R.Daneel -- dunno since I dunno what electrum is :)
101 2013-10-27 02:16:08 <Luke-Jr> heh, I didn't even notice his nick XD
102 2013-10-27 02:16:21 <Ascendion> LOL :)
103 2013-10-27 02:16:46 <MrDaneelOlivaw> I can never hide
104 2013-10-27 02:17:58 <MrDaneelOlivaw> I tried reporter Chetter Hummin but that didn't work out, so I tried MrDaneelOlivaw the russian programmer but ..
105 2013-10-27 02:18:43 <Ascendion> what I'm working on right now is scraping bitcoind via json-rpc to import the blockchain and all transactions into an sql database, then keep that database in sync in realtime as long as the "import" app is running
106 2013-10-27 02:19:20 <Ascendion> then I plan to do the same for litecoin and a few others
107 2013-10-27 02:20:37 <Ascendion> Daneel -- I prolly would have caught almost any name from that series considering I reread the entire series just a few months back :)
108 2013-10-27 02:22:51 <MrDaneelOlivaw> Ascendion: I've done so in a couple of languages at least 3 times, a great story full of details
109 2013-10-27 02:23:43 <MrDaneelOlivaw> Ascendion: have you implemented pruning ?
110 2013-10-27 02:23:55 <Ascendion> ok damn it -- why all of a sudden can I not get the raw transaction data for the genesis transaction
111 2013-10-27 02:24:14 <Ascendion> pruning as in detecting forks/reorgs ?? yes
112 2013-10-27 02:24:50 <Ascendion> I only keep what my local bitcoind considers to be the current best chain in the database -- I dont store orphans
113 2013-10-27 02:25:34 <MrDaneelOlivaw> Ascendion: no, pruning as in not keeping addresses without spendable outputs
114 2013-10-27 02:26:20 <Ascendion> I dont particularly store addresses separate from the transaction outputs where they are referenced so no -- I dont prune them
115 2013-10-27 02:27:44 <MrDaneelOlivaw> Ascendion: eta to sync?
116 2013-10-27 02:28:24 <Ascendion> 10k years if I dont figure out why I cannot fetch the transaction in the genesis block like I was able to do about an hour ago
117 2013-10-27 02:38:21 <Ascendion> it keeps coming back with an internal server error response code
118 2013-10-27 02:45:27 <Ascendion> hmmm it was able to get the transaction from the 2nd block
119 2013-10-27 02:47:38 <Ascendion> "no information available about transaction" -- HUH ??? this is the transaction in the genesis block -- I've got the right transaction id, validated it on blockchain.info -- why can I not fetch it from bitcoind ??????????
120 2013-10-27 02:51:17 <gmaxwell> Ascendion: didn't I strongly advise you before to not test with the genesis block?
121 2013-10-27 02:51:27 <gmaxwell> Ascendion: is there a reason you didn't take my advice? :P
122 2013-10-27 02:52:19 <gmaxwell> Maybe I didn't. Don't test with the genesis block, it's special.
123 2013-10-27 02:52:34 <gmaxwell> The coinbase transaction in the genesis block doesn't exist as far as the system is concerned.
124 2013-10-27 02:52:41 <gmaxwell> Though coins are forever unspendable.
125 2013-10-27 02:53:34 <Luke-Jr> Those*
126 2013-10-27 02:55:20 <Ascendion> then why return a transaction id at all in the block, why does blockchain.info show data for that transaction ?? I HATE SPECIAL CASES !!!
127 2013-10-27 02:57:32 <gmaxwell> Ascendion: because there is a transaction in the block, but that transaction was never inserted into the database by the bitcoin system, because the code that hardcodes the genesis block didn't do so.
128 2013-10-27 02:59:31 <Ascendion> personally I call a block with a transaction listed that cannot be accessed a BUG :) no matter if its the genesis block or not :) just breaks the rules that every other block follows
129 2013-10-27 03:01:45 <gmaxwell> Ascendion: The transaction was never introduced to the network. It's a hard forking requirement of the blockchain protocol now.
130 2013-10-27 03:02:20 <gmaxwell> It could be returned by the rpc anyways, e.g. by internally special casing it, but that would be inaccurate.. e.g. it would make you think that coin could be spent when it couldn't.
131 2013-10-27 03:02:22 <Ascendion> doesnt have to be on the network... just add the transaction to the db hardcode :)
132 2013-10-27 03:03:29 <Ascendion> thats no different than any other "lost my wallet/keys/coins" ... probably plenty of unspendable outputs in the database for just that reason
133 2013-10-27 03:04:39 <gmaxwell> Ascendion: but then you'll fetch it and add up the coins that could be possibly spent and get a wrong answer.
134 2013-10-27 03:04:48 <Ascendion> doesnt make a difference to my system -- someone has to provide the address and a signed message by that address before I will credit their "wallet" with those coins
135 2013-10-27 03:05:01 <gmaxwell> unless, again, you special case it. And if you're going to special case, you might as well just pretend the genesis block doesn't exist.
136 2013-10-27 03:05:21 <gmaxwell> uh. ... I probably don't want to know what you're doing.
137 2013-10-27 03:05:45 <Ascendion> I fetch the genesis block for no other reason than to validate that the next block points back to it :)
138 2013-10-27 03:06:25 <gmaxwell> Ascendion: so? stop at 1 and just consider it the first block.
139 2013-10-27 03:06:32 <Ascendion> importing the entire blockchain/transaction data into a SQL database, then maintaining it in realtime as blocks arrive and reorgs happen
140 2013-10-27 03:07:00 <Ascendion> I already special cased the code -- it ignores the transactions for height = 0
141 2013-10-27 03:08:35 <Ascendion> adding code to track transactions processed and time elapsed so I can see how fast its importing
142 2013-10-27 03:11:22 <Ascendion> and blowing my brains out on pain meds cause there is a tstorm moving through thats setting off my joint pain :)
143 2013-10-27 05:34:18 <gmaxwell> http://www.reddit.com/r/Bitcoin/comments/1p3zoz/gavin_andresen_core_development_update_5/ccyk7qe I appreciate the enthusiasm, but it would be helpful if he wasn't misinforming people that they will have to "rewrite their code".
144 2013-10-27 06:32:25 <BlueMatt> https://github.com/bitcoin/bitcoin/pull/3164 boom
145 2013-10-27 06:32:38 <BlueMatt> gavinandresen: ^ :)
146 2013-10-27 06:33:11 <gmaxwell> BlueMatt: thanks!
147 2013-10-27 06:33:31 <gmaxwell> BlueMatt: hey, if you're working on it right now can you add a test for OP_RETURN reorgs? :P
148 2013-10-27 06:34:10 <gmaxwell> Basically we need a block with a transaction that has a single OP_RETURN output, and that block needs to get reorged out of the chain. I believe this will fail on current master.
149 2013-10-27 06:34:33 <gmaxwell> (it would be nice to have some additional txn in the block: two OP_return, and one OP_RETURN one regular output, but less essential.)
150 2013-10-27 06:34:58 <BlueMatt> hmm...ok, lets see if I can break master then
151 2013-10-27 06:35:25 <BlueMatt> ACTION needs to buy a new disk for this new workstation...Ive gotten spoiled by my laptop's ssd and now everything is sooooo sloooowwww
152 2013-10-27 06:36:20 <gmaxwell> Sipa also has a pull which should fix that bug.
153 2013-10-27 07:46:07 <BlueMatt> gmaxwell: <gmaxwell> (it would be nice to have some additional txn in the block: two OP_return, and one OP_RETURN one regular output, but less essential.) <-- E_NOPARSE
154 2013-10-27 07:47:01 <gmaxwell> BlueMatt: a transaction with a single output that is nothing but the opcode OP_RETURN (0x6a I think?)
155 2013-10-27 07:47:19 <gmaxwell> another transaction with two outputs, each an OP_RETURN
156 2013-10-27 07:47:38 <gmaxwell> another transaction with a regular pay to address output, and a OP_RETURN
157 2013-10-27 07:47:45 <BlueMatt> in separate blocks, or the same?
158 2013-10-27 07:48:00 <gmaxwell> same block is fine. The block then needs to get reorged out of the chain.
159 2013-10-27 07:48:01 <flound1129> any devs around?
160 2013-10-27 07:48:22 <gmaxwell> Bounus points if you double up that last transaction to have both orders of outputs. OP_RETURN first and last.
161 2013-10-27 07:48:27 <flound1129> I'm trying to find out why a block my pool found didn't get into the chain
162 2013-10-27 07:48:54 <flound1129> it should have been block 226249, but that block was found by bitminter 10 minute slater
163 2013-10-27 07:48:58 <gmaxwell> flound1129: some roughly 1% of all blocks are orphaned naturally by chance. Do you have an reason to think it was anything else?
164 2013-10-27 07:49:05 <flound1129> http://pastebin.com/EiaRJaH6
165 2013-10-27 07:49:09 <BlueMatt> gmaxwell: should be easy enough...let me see if I can whip it up
166 2013-10-27 07:49:09 <gmaxwell> hm. 10 minutes, how are you judging that?
167 2013-10-27 07:49:24 <flound1129> timestamp in the db
168 2013-10-27 07:49:29 <flound1129> when the share was submitted
169 2013-10-27 07:49:44 <gmaxwell> flound1129: what is the hash of your block?
170 2013-10-27 07:49:58 <gmaxwell> can you email me your full debug log?
171 2013-10-27 07:50:42 <flound1129> yes
172 2013-10-27 07:50:47 <flound1129> hash is 00000000000000054dcf9fbb3263ca4abcc3cc5b1dd03a19169481bd24dd56aa
173 2013-10-27 07:50:53 <gmaxwell> gmaxwell@gmail.com works.
174 2013-10-27 07:51:39 <gmaxwell> neither of my nodes have seen that block.
175 2013-10-27 07:52:51 <feddy3> blockchain.info doesn't have it either
176 2013-10-27 07:53:13 <gmaxwell> feddy3: yea thats not super informative. :P
177 2013-10-27 07:53:27 <feddy3> my node doesn't have it either ;)
178 2013-10-27 07:54:21 <flound1129> just sent you mail
179 2013-10-27 07:54:38 <sipa> well then it sounds like the pool software never sent it out?
180 2013-10-27 07:54:51 <gmaxwell> thats why I wanted the debug log, downloading it as we speak.
181 2013-10-27 07:55:22 <flound1129> ProcessBlock: ACCEPTED
182 2013-10-27 07:55:22 <flound1129> ThreadRPCServer method=submitblock
183 2013-10-27 07:55:23 <gmaxwell> absolutely no mention of thash hash in your debug log.
184 2013-10-27 07:56:32 <sipa> check for the mined block's parent hash
185 2013-10-27 07:56:56 <sipa> this looks like it was processed as a (true) orphan
186 2013-10-27 07:57:10 <gmaxwell> flound1129: for the future you may want to run with logtimestamps=1
187 2013-10-27 07:57:15 <flound1129> ok will do
188 2013-10-27 07:57:49 <gmaxwell> ProcessBlock: ACCEPTED
189 2013-10-27 07:57:49 <gmaxwell> ThreadRPCServer method=submitblock
190 2013-10-27 07:58:11 <gmaxwell> ... and many lines later:
191 2013-10-27 07:58:14 <gmaxwell> received block 0000000000000009e45551733bf49e4f2517a0e907337e5e05c326a04f6b3d3e
192 2013-10-27 07:58:19 <gmaxwell> SetBestChain: new best=0000000000000009e45551733bf49e4f2517a0e907337e5e05c326a04f6b3d3e height=266249 log2_work=73.208971 tx=26064781 date=2013-10-26 22:25:11 progress=0.999998
193 2013-10-27 07:58:32 <flound1129> yeah that's the bitminter block
194 2013-10-27 07:58:34 <gmaxwell> flound1129: do you have an actual copy of this block?
195 2013-10-27 07:58:44 <flound1129> there's nothing in the wallet
196 2013-10-27 07:58:46 <flound1129> no transactions
197 2013-10-27 07:59:13 <gmaxwell> seems like your node didn't do anything with it at all, like it didn't think it was the best at the time it recieved it.
198 2013-10-27 07:59:26 <gmaxwell> flound1129: what bitcoin version is your node?
199 2013-10-27 07:59:51 <gmaxwell> Oh nevermind Bitcoin version v0.8.3-dirty-beta (2013-06-25 10:27:24 -0400)
200 2013-10-27 08:00:12 <flound1129> yeah I'm compiling 0.8.5 now
201 2013-10-27 08:01:08 <gmaxwell> wow, really slow host. is this running on an ec2 instance or something?
202 2013-10-27 08:01:11 <gmaxwell> Flushed 14916 addresses to peers.dat 7676ms
203 2013-10-27 08:01:23 <flound1129> hmmmm
204 2013-10-27 08:01:34 <flound1129> no, but my VPS was having issues last night
205 2013-10-27 08:01:34 <gmaxwell> though I don't see how that could be a problem here.
206 2013-10-27 08:02:01 <flound1129> disk issues on that physical host
207 2013-10-27 08:02:26 <flound1129> another VPS was using all the I/O bandwidth
208 2013-10-27 08:02:33 <flound1129> could that have factored into this?
209 2013-10-27 08:02:35 <gmaxwell> logs indicate really surprisingly slow IO, I'd look into that though I don't think it's the cause of your orphaning.
210 2013-10-27 08:02:52 <gmaxwell> though I'd like to get a copy of the actual block
211 2013-10-27 08:03:01 <flound1129> how can I get that?
212 2013-10-27 08:03:14 <gmaxwell> I take it your pool software doesn't log it?
213 2013-10-27 08:03:25 <flound1129> it logged the same hash
214 2013-10-27 08:03:37 <gmaxwell> yea, but I don't want the hash, I want the block.
215 2013-10-27 08:04:03 <gmaxwell> one possiblity is that you were lagged out when your pool last fetched work and this was actually a solution one block behind.
216 2013-10-27 08:04:13 <gmaxwell> that would explain your node not moving up to it.
217 2013-10-27 08:04:15 <flound1129> the bitcoind was lagged out you mean?
218 2013-10-27 08:04:33 <flound1129> hmm
219 2013-10-27 08:04:39 <flound1129> wouldn't it be rejected though
220 2013-10-27 08:05:07 <gmaxwell> flound1129: or your pool software. No, it just wouldn't be best so it wouldn't move to it. I don't /think/ we log anything special in that case.
221 2013-10-27 08:05:25 <flound1129> ok
222 2013-10-27 08:05:27 <gmaxwell> it would pass the initial sanity checks and then just not be attractive unless some other block extended it and made it the longer chain.
223 2013-10-27 08:05:43 <flound1129> make sense
224 2013-10-27 08:07:56 <BlueMatt> gmaxwell: yep, definitely is broken on master
225 2013-10-27 08:08:14 <BlueMatt> gmaxwell: remind me to push the latest version to pull-tester when the pull that fixes it is merged
226 2013-10-27 08:08:23 <gmaxwell> BlueMatt: yep. thought so. Can you test sipa's fix?
227 2013-10-27 08:08:31 <BlueMatt> branchname?
228 2013-10-27 08:08:37 <gmaxwell> BlueMatt: https://github.com/bitcoin/bitcoin/pull/3163
229 2013-10-27 08:08:45 <BlueMatt> lazy
230 2013-10-27 08:08:57 <gmaxwell> (I was holding off acking it because I didn't have a reorg reproduction, only the startup time test reproduction)
231 2013-10-27 08:09:12 <gmaxwell> sipa:allunspendable
232 2013-10-27 08:09:22 <BlueMatt> yea, building now
233 2013-10-27 08:10:37 <gmaxwell> sipa: congrats, your anal startup tests found a forking bug.
234 2013-10-27 08:11:24 <sipa> gmaxwell: ... which I introduced
235 2013-10-27 08:11:39 <gmaxwell> pft. and everyone else acked. :P
236 2013-10-27 08:12:24 <sipa> true
237 2013-10-27 08:12:55 <gmaxwell> sipa: I tested by reindexing the chain... as I knew the chain already had a number of these things.
238 2013-10-27 08:13:13 <sipa> same
239 2013-10-27 08:13:27 <sipa> well, it seemed my laptop's node hadn't run im two weeks
240 2013-10-27 08:13:28 <gmaxwell> we don't currently really have a way to do the disconnect test to the whole chain, due to the memory limit.
241 2013-10-27 08:14:01 <sipa> so i just rannit until i crossed the all-unspent case you hot, and restarted
242 2013-10-27 08:14:11 <sipa> you GOT
243 2013-10-27 08:14:42 <BlueMatt> sipa: it didnt get released, so who cares, right? :)
244 2013-10-27 08:14:53 <sipa> sure
245 2013-10-27 08:14:55 <gmaxwell> ah, thats how you got a reproduction? :P that works. you could have used your blacklist patch to trigger a reorg to once above, no? I think I commented on that pull basically saying we could use it to try to disconnect the whole chain.
246 2013-10-27 08:15:13 <gmaxwell> BlueMatt: it did almost get missed, so I think its worth thinking about how we need to improve the process.
247 2013-10-27 08:15:16 <sipa> hmm, i missed thar
248 2013-10-27 08:16:09 <sipa> gmaxwell: for testing purposes a unblacklist woukd be useful too
249 2013-10-27 08:16:13 <gmaxwell> BlueMatt: basically I shut down my node at just the right time while moving between rooms at linuxcon and got "chainstate corruption", sipa's startup test failing at start, which I then bit into and didn't release until we figured out what it was.
250 2013-10-27 08:16:20 <BlueMatt> gmaxwell: everyone should be required to write a shitton of bitcoindcomparisontool tests :)
251 2013-10-27 08:16:43 <sipa> that would be extremely helpful
252 2013-10-27 08:16:45 <gmaxwell> BlueMatt: Yea, I need to figure out how to run that locally and build it.
253 2013-10-27 08:16:58 <sipa> if only for forcing us to understand the code
254 2013-10-27 08:17:00 <BlueMatt> gmaxwell: intellij -> import bitcoinj -> run :)
255 2013-10-27 08:17:07 <gmaxwell> I think we should have required an opreturn test for the comparison tool.
256 2013-10-27 08:17:18 <BlueMatt> (just dont try to build the jar in intellij..)
257 2013-10-27 08:17:31 <gmaxwell> We slacked because these things were already in the chain, if there were none in the chain I think it's likely I would have demanded a test.
258 2013-10-27 08:17:46 <sipa> i haven't built any java in years i think
259 2013-10-27 08:18:17 <gmaxwell> sipa: https://github.com/bitcoin/bitcoin/pull/2839#issuecomment-21302247
260 2013-10-27 08:19:20 <gmaxwell> BlueMatt: anyways when you confirm its fixed please comment on that pull that you confirmed the bug is forking and that the patch fixes it.
261 2013-10-27 08:20:16 <sipa> oh there, right
262 2013-10-27 08:20:59 <gmaxwell> (well, some consolation that at least a kind of test that would have caught it had been mentioned.)
263 2013-10-27 08:21:20 <gmaxwell> sipa: unblacklist would be neat, you could thrash the heck out of a node blacklisting block 1 and unblacklisting it. :P
264 2013-10-27 08:22:11 <sipa> haha
265 2013-10-27 08:22:21 <BlueMatt> ok, reorg off+on a block with: {outputs: {OP_RETURN, OP_TRUEx4}}, {outputs: {OP_RETURNx2}, inputs: {tx1_1}}, {outputs: {OP_RETURN, OP_TRUE}, inputs: {tx1_2}}, {outputs: {OP_TRUE, OP_RETURN}, inputs: {tx1_3}}, {outputs: {OP_RETURN}, inputs: {tx1_4}}
266 2013-10-27 08:22:25 <BlueMatt> success with sipa's patch
267 2013-10-27 08:22:41 <sipa> \o/
268 2013-10-27 08:22:48 <BlueMatt> gmaxwell: how's that for bonus points?
269 2013-10-27 08:23:14 <sipa> ACTION does the BitcoindComparisonTool.jardance
270 2013-10-27 08:23:41 <gmaxwell> BlueMatt: sweet. Thats good!
271 2013-10-27 08:25:58 <BlueMatt> sipa: https://github.com/TheBlueMatt/test-scripts (and yes, you now need all the jars in the folder...blame intellij and my unwillingness to crapify my new workstation with 30 java ides)
272 2013-10-27 08:26:14 <BlueMatt> that will fail on master, Ill throw it on the pull-tester when sipa's fix gets merged
273 2013-10-27 08:27:15 <sipa> can't you just unzip + rezip them together?
274 2013-10-27 08:27:29 <BlueMatt> gmaxwell: and if you want to see how easy that test was to create, look at https://code.google.com/r/bluemattme-bitcoinj/source/detail?r=cb616b6f4ec9430551f0ae7767944aa46d2f606e&name=blocktester
275 2013-10-27 08:27:55 <BlueMatt> sipa: nfc, but you cant just throw them all in the jar normally 'cause some of the deps are signed classes
276 2013-10-27 08:28:03 <BlueMatt> I think there is magic that makes it work (eclipse knows how)
277 2013-10-27 08:28:11 <BlueMatt> but I dunno it, and Im lazy
278 2013-10-27 08:28:27 <BlueMatt> gmaxwell: its literally all copy/paste/change-1-char
279 2013-10-27 08:29:20 <sipa> where is b86 (in the comment) ?
280 2013-10-27 08:29:40 <BlueMatt> I forgot to write it :)
281 2013-10-27 08:31:06 <gmaxwell> BlueMatt: yea, I believe it's easy, but installing a java ide ... yuck. :P
282 2013-10-27 08:37:36 <BlueMatt> gmaxwell: generally, I might agree, but intellij is surprisingly nice
283 2013-10-27 08:37:43 <BlueMatt> for an ide, ofc
284 2013-10-27 08:38:57 <sipa> why am i awake?
285 2013-10-27 08:39:10 <BlueMatt> sipa: I was just asking myself that very question...
286 2013-10-27 08:39:15 <BlueMatt> why is gmaxwell awake?
287 2013-10-27 08:39:16 <sipa> it's before 10am on a sunday
288 2013-10-27 08:42:38 <MC1984> we starting a nite owls club?
289 2013-10-27 08:42:55 <BlueMatt> ACTION 's been in that one forever
290 2013-10-27 08:43:25 <BlueMatt> 3rd night in a row up >4am working on things that should've been done months ago...
291 2013-10-27 08:43:44 <MC1984> you dont know who youre dealing with
292 2013-10-27 08:43:58 <BlueMatt> ACTION was up till 7 on thursday...
293 2013-10-27 08:44:15 <BlueMatt> (and a wee bit drunk on irc, sorry 'bout that one y'all)
294 2013-10-27 08:44:30 <BlueMatt> GOD pull-tester is slow as fuck now
295 2013-10-27 08:44:31 <MC1984> but drunk irc is best irc
296 2013-10-27 08:44:34 <BlueMatt> need more tmpfs
297 2013-10-27 08:47:11 <sipa> BlueMatt: tmpfs on the devserver?
298 2013-10-27 08:47:19 <BlueMatt> yea
299 2013-10-27 08:47:19 <sipa> or at least the pulltester vm
300 2013-10-27 08:47:24 <BlueMatt> pulltester vm
301 2013-10-27 08:47:32 <sipa> why did that change?
302 2013-10-27 08:47:41 <BlueMatt> I think it used to have one
303 2013-10-27 08:47:49 <sipa> or does it just need more memory now
304 2013-10-27 08:47:50 <BlueMatt> but now it needs to be in a different place
305 2013-10-27 08:47:54 <sipa> ic
306 2013-10-27 08:48:02 <BlueMatt> the tester now needs disk for its db since it does huge-reorg-tests
307 2013-10-27 08:48:18 <BlueMatt> so it needs tmpfs for that to run reasonably on huge reorgs for every pull
308 2013-10-27 08:48:31 <BlueMatt> not sure if that version ever got pushed or if autotools came first and broke it
309 2013-10-27 08:49:19 <BlueMatt> sipa: ok, you got your b89
310 2013-10-27 08:49:22 <BlueMatt> now Im off to bed
311 2013-10-27 08:49:56 <sipa> ... 86
312 2013-10-27 08:50:13 <sipa> nite!
313 2013-10-27 08:56:57 <sipa> i'm not sure i understand the current state
314 2013-10-27 08:57:12 <sipa> is the comparison tool active in pulltester?
315 2013-10-27 09:05:40 <warren> wow. bitcoin-qt.exe works in wine.
316 2013-10-27 09:07:07 <sipa> doesn't surprise me that qt works well on wine
317 2013-10-27 09:08:24 <warren> I'm playing with bitcoin-qt.exe and litecoin-qt.exe built with the new mingw
318 2013-10-27 09:10:11 <Ascendion> hmmmm taking forever to verify last 288 blocks on startup of bitcoind
319 2013-10-27 09:10:56 <Ascendion> finally :)
320 2013-10-27 09:14:49 <Ascendion> OUCH
321 2013-10-27 09:18:45 <Ascendion> getting the db schema right is giving me one heck of a headache :)
322 2013-10-27 09:24:25 <warren> hmm... litecoin-qt.exe is using less RAM than the gitian linux build
323 2013-10-27 10:23:21 <feddy3> On the standard scriptSig with 2 pushdatas, what are the size requirements of each of the pushes (so that the tx is standard) ?
324 2013-10-27 10:23:29 <feddy3> Is the 2nd push always 33 bytes?
325 2013-10-27 10:23:46 <feddy3> i.e., a compressed pubkey? or can it be uncompressed?
326 2013-10-27 10:24:24 <sipa> for standardness, there are no limitations
327 2013-10-27 10:24:33 <sipa> the scriptSig just has to be push-only
328 2013-10-27 10:24:47 <sipa> if you don't give a valid public key though, the transaction won't be valid
329 2013-10-27 10:24:57 <feddy3> sure
330 2013-10-27 10:25:18 <sipa> in general, you cannot know what the meaning of the data in the pushes is
331 2013-10-27 10:25:30 <feddy3> of course
332 2013-10-27 10:25:33 <sipa> though i guess you could do something like "this looks like a public key, show it as one"
333 2013-10-27 10:25:41 <sipa> or "this looks like a signature, show it as one"
334 2013-10-27 10:26:00 <feddy3> I want to do something simpler than that: just give a basic interpretation of the script
335 2013-10-27 10:26:06 <feddy3> "This is a standard spend"
336 2013-10-27 10:26:21 <sipa> unless you also look at the output being consumed, you can't do that (in general)
337 2013-10-27 10:26:41 <feddy3> So giving an interpretation of scriptSig isn't useful?
338 2013-10-27 10:26:55 <feddy3> by itself
339 2013-10-27 10:27:14 <sipa> in general, no
340 2013-10-27 10:27:18 <sipa> in practice, probably yes
341 2013-10-27 10:27:39 <sipa> brb, i'll explain
342 2013-10-27 10:27:54 <feddy3> what approach would you think is best? If all(opcode == pushdata) then interpretation = "standard spend" ?
343 2013-10-27 10:27:57 <feddy3> ok
344 2013-10-27 10:33:00 <sipa> feddy3: i don't think you need to worry about standardness or not
345 2013-10-27 10:33:22 <sipa> feddy3: in particular, i think such a transaction disassembler is exactly most useful for non-standard spends
346 2013-10-27 10:33:30 <feddy3> sure
347 2013-10-27 10:34:08 <sipa> feddy3: but it may make sense to just decode the script operations, and when you see something that looks like a pubkey, show it as a pubkey and the corresponding address for example
348 2013-10-27 10:34:18 <feddy3> I'm interested in providing an "interpretation" of standard scripts, not necessarily qualifying validity
349 2013-10-27 10:35:01 <feddy3> for instance, standard pay to pubkeyhash can show the destination address, right?
350 2013-10-27 10:35:22 <sipa> that's outputs; that's different
351 2013-10-27 10:35:31 <feddy3> yup
352 2013-10-27 10:35:39 <sipa> there you can definitely perform pattern matching to see if it fits an address template
353 2013-10-27 10:35:45 <feddy3> right
354 2013-10-27 10:35:46 <sipa> and then show the corresponding address
355 2013-10-27 10:35:58 <feddy3> and I have implemented basic matching already
356 2013-10-27 10:36:12 <feddy3> I'm just curious what a decent way of interpreting input scripts should be, or not at all
357 2013-10-27 10:36:16 <sipa> you only really need pay-to-pubkeyhash and pay-to-scripthash
358 2013-10-27 10:36:21 <sipa> and probably pay-to-pubkey too
359 2013-10-27 10:36:24 <feddy3> and multisig
360 2013-10-27 10:37:01 <sipa> but for example, in the case of a P2SH input, it would be really nice to disassemble the subscript
361 2013-10-27 10:37:17 <sipa> (though detecting one is certainly harder)
362 2013-10-27 10:37:23 <feddy3> yeah
363 2013-10-27 10:37:33 <sipa> for inputs, i
364 2013-10-27 10:37:43 <sipa> for inputs, i'd just try to give meaning to the individual pushes, if any
365 2013-10-27 10:37:53 <feddy3> you have to know that the referenced output was a p2sh to begin with..
366 2013-10-27 10:38:09 <feddy3> ah, interesting
367 2013-10-27 10:38:32 <sipa> well, you can try decoding it as a script, and if it succeeds, and contains an OP_CHECKSIG or any of a few other very common ones, assume it is a subscript
368 2013-10-27 10:38:50 <sipa> as at least public keys and signatures are very recognizable
369 2013-10-27 10:39:06 <feddy3> ahh, yeah
370 2013-10-27 10:39:18 <feddy3> for every pushdata, decode it as a script and see if it matches a regular pattern?
371 2013-10-27 10:39:25 <feddy3> hmm
372 2013-10-27 10:39:32 <sipa> i'd just do pubkey and signatures first
373 2013-10-27 10:39:38 <feddy3> yeah
374 2013-10-27 10:40:12 <sipa> for a pubkey you can say: whether it's compressed or not, whether it's valid or not (but that requires some ec crypto), what the corresponding address is
375 2013-10-27 10:41:05 <sipa> for a signature you can say whether it's canonical or not
376 2013-10-27 10:41:06 <gmaxwell> [OT] the sha1 bounty ( 37k7toV1Nv4DfmQbmZ8KuZDQCYK9x5KpzP ) is up to 2.47 btc now.
377 2013-10-27 10:41:33 <sipa> (though detecting non-canonical signatures again becomes more fuzzy)
378 2013-10-27 10:41:40 <Anduck> the tx still isn't relayed.....
379 2013-10-27 10:41:44 <Anduck> 3 days now
380 2013-10-27 10:41:49 <sipa> Anduck: it has a fee of 0.00001
381 2013-10-27 10:41:58 <sipa> i don't think any node will relay that
382 2013-10-27 10:41:59 <Anduck> well even 0-fee's get relayed to other nodes
383 2013-10-27 10:42:07 <sipa> if they are high priority
384 2013-10-27 10:42:10 <sipa> yours isn't
385 2013-10-27 10:42:32 <Anduck> well bitcoind made it...
386 2013-10-27 10:42:36 <sipa> that worries me
387 2013-10-27 10:42:42 <sipa> 0.8.5, you say?
388 2013-10-27 10:42:44 <Anduck> yes
389 2013-10-27 10:42:46 <gmaxwell> hm.
390 2013-10-27 10:42:52 <gmaxwell> Anduck: with the raw transaction interface?
391 2013-10-27 10:42:54 <Anduck> 0.8.5 is what i am using
392 2013-10-27 10:42:55 <Anduck> no, sendmany
393 2013-10-27 10:43:20 <feddy3> sipa: Ok, thanks.
394 2013-10-27 10:43:28 <gmaxwell> Anduck: can you pastebin the raw transaction to me?
395 2013-10-27 10:43:31 <Anduck> sure
396 2013-10-27 10:45:38 <sipa> feddy3: how do i paste a raw transaction?
397 2013-10-27 10:45:51 <feddy3> sipa: I've updated fetchtx.info again with some error handling.. it'd be very cool if you had feedback on the site
398 2013-10-27 10:46:06 <sipa> well, i pasted a raw transaction and get an empty page
399 2013-10-27 10:46:28 <feddy3> really? can I get the tx?
400 2013-10-27 10:46:42 <gmaxwell> Anduck: I handed it manually to eligius and it appears to have accepted it.
401 2013-10-27 10:46:58 <feddy3> sipa: hit random tx, then click 'Show' and overwrite the tx in there
402 2013-10-27 10:47:01 <gmaxwell> it's not yet in their blocktemplates, however.
403 2013-10-27 10:47:17 <Anduck> thx
404 2013-10-27 10:47:39 <gmaxwell> (presumably because its at the bottom of the priority list)
405 2013-10-27 10:48:01 <Anduck> but are 0-fees higher than my tx?
406 2013-10-27 10:48:24 <sipa> Anduck: anything below 0.0001 (iirc) is considered as 0-fee
407 2013-10-27 10:48:50 <Anduck> btw i am not 100% sure the tx is made with 0.8.5 because i did change client vers there couple times
408 2013-10-27 10:49:06 <Anduck> but i am pretty confident it's 0.8.5 because i didnt send txs from the other vers...
409 2013-10-27 10:49:19 <sipa> is it possible you created it with a recent git head?
410 2013-10-27 10:49:27 <Anduck> yes
411 2013-10-27 10:49:47 <Anduck> it's the version i used for a while but i am pretty sure i didn't use it for the sendmany tx
412 2013-10-27 10:49:53 <Anduck> but it's possible.......
413 2013-10-27 10:50:19 <gmaxwell> well that would certantly explain itâ¦
414 2013-10-27 10:51:13 <sipa> gmaxwell: how so?
415 2013-10-27 10:51:33 <sipa> ooh!
416 2013-10-27 10:51:36 <gmaxwell> right.
417 2013-10-27 10:51:59 <sipa> removing the subcent rule caused his node to create a transaction with 0 fee
418 2013-10-27 10:52:08 <sipa> and the change was below dust, so was moved to fee
419 2013-10-27 10:53:50 <gmaxwell> I'm kinda wondering if maybe we shouldn't ship 0.9 with the fee defaulted to 0.0001 BTC/kb to avoid that gotcha.
420 2013-10-27 10:54:28 <gmaxwell> probably not worth the political noise.
421 2013-10-27 10:54:42 <gmaxwell> I say we abandon bitcoin-qt and create a new project bitcoin2-qt. :P
422 2013-10-27 10:54:59 <sipa> how about ditching the toolkit
423 2013-10-27 10:55:07 <gmaxwell> And then we can change all the defaults without anyone feeling like we killed their dog.
424 2013-10-27 10:55:08 <sipa> i hear a lot of good things about wx
425 2013-10-27 10:55:24 <gmaxwell> I vote for ncurses.
426 2013-10-27 10:55:28 <sipa> +1 !
427 2013-10-27 10:55:32 <swulf--> In the long run, separating the projects entirely would be a good move, no?
428 2013-10-27 10:55:49 <sipa> first step: separating the processes
429 2013-10-27 10:55:55 <sipa> second step: separating the binaries
430 2013-10-27 10:56:01 <sipa> third step: separating the codebase
431 2013-10-27 10:56:07 <sipa> fourth step: separating the projects
432 2013-10-27 10:56:26 <swulf--> yeah
433 2013-10-27 10:56:29 <gmaxwell> fifth step: gavin and the log splitter.
434 2013-10-27 10:56:57 <sipa> running java code always makes me somehow feel... infected
435 2013-10-27 10:57:07 <gmaxwell> "What do you mean that binary fission doesn't work on developers??"
436 2013-10-27 10:57:14 <CodeShark> gmaxwell: https://github.com/CodeShark/BitShell
437 2013-10-27 10:57:31 <gmaxwell> "I thought you said they didn't have sex?" "No! I said they don't get laid!"
438 2013-10-27 10:58:17 <gmaxwell> CodeShark: interesting!
439 2013-10-27 10:58:17 <sipa> gmaxwell: now you los me
440 2013-10-27 10:58:23 <sipa> *lost
441 2013-10-27 10:59:43 <gmaxwell> sipa: a log splitter is a mechnical device that splits logs (the wooden kind) in half. Binary fission is the process bacteria use to reproduce, as opposed to sexual reproduction.
442 2013-10-27 10:59:51 <HM2> damn CodeShark
443 2013-10-27 10:59:55 <HM2> this is something i've though about
444 2013-10-27 11:00:16 <sipa> gmaxwell: oh, i wasn't familiar with that biological term
445 2013-10-27 11:03:32 <feddy3> Is it possible to get a bypass of the newbies-only limitation on bitcointalk here?
446 2013-10-27 11:05:09 <sipa> hmm, it seems that matt's pullreq enables both the comparisontool, and fixes an issue with it
447 2013-10-27 11:05:33 <gmaxwell> feddy3: whats your username?
448 2013-10-27 11:05:41 <feddy3> gmaxwell: fetchtx
449 2013-10-27 11:06:51 <gmaxwell> feddy3: if you go spamming around a bunch of places I'll personally kill you. You're whitelisted now. :)
450 2013-10-27 11:07:09 <HM2> lol
451 2013-10-27 11:07:10 <feddy3> gmaxwell: hah, thanks ;) I just want to start a single thread for my site
452 2013-10-27 11:07:25 <sipa> I heard that gmaxwell executions are especially slow and painful.
453 2013-10-27 11:07:52 <sipa> involving opious amounts of capitalization
454 2013-10-27 11:07:57 <sipa> *copious
455 2013-10-27 11:08:05 <wumpus> hehe
456 2013-10-27 11:08:53 <sipa> here is a nice example: https://bitcointalk.org/index.php?topic=311000.msg3345309#msg3345309
457 2013-10-27 11:10:33 <feddy3> ah yes, I read that earlier
458 2013-10-27 11:11:14 <feddy3> I'd like to see him expand on the "it's not hard to memorize 128 bits" part, so that people who want to use brainwallets can do it right
459 2013-10-27 11:11:33 <gmaxwell> feddy3: what electrum does, for example.
460 2013-10-27 11:12:08 <gmaxwell> It has a bijection between 128 bit uniform numbers and sets of 12 not-uncommon english words.
461 2013-10-27 11:12:20 <feddy3> Ah, interesting
462 2013-10-27 11:12:36 <feddy3> Is it the same recovery "peom" system that blockchain.info wallets have?
463 2013-10-27 11:14:01 <gmaxwell> Now. _memorizing_ that and making that your only storage mechnism is totally possible, but its risky... because memory is a bit brittle. E.g. you can remember something fine for years.. and the one day, some magic cell dies, or you run a fever and poof. can't remember it... so I still don't recommend it, but its way better than letting the human come up with the key.
464 2013-10-27 11:14:11 <sipa> feddy3: two nits: left-align the txid input box, and allowing pasting a raw transaction in hex there
465 2013-10-27 11:14:25 <gmaxwell> feddy3: No idea what bc.i has.
466 2013-10-27 11:14:58 <gmaxwell> So for an electrum wallet, by all means, memorize it, but also write it down. If you're worried about people finding your written copy, split it in half or whatever.
467 2013-10-27 11:15:58 <feddy3> sipa: the txid input box on the main page? the one on the txview page is left aligned.
468 2013-10-27 11:17:00 <feddy3> hmm, I need to do some magic to allow pasting the the raw tx on the main page
469 2013-10-27 11:17:55 <CodeShark> embed the even bits into a youtube video, write the odd bits down and lock them in a safe
470 2013-10-27 11:17:57 <CodeShark> :p
471 2013-10-27 11:18:34 <HM2> write it on every sheet of toilet paper in the house, then reroll it
472 2013-10-27 11:18:37 <HM2> you'll never forget
473 2013-10-27 11:19:07 <CodeShark> or embed the key into several youtube videos such that the xor of all the vids combined produces the key
474 2013-10-27 11:19:18 <CodeShark> and then store the list of vids in a safe :)
475 2013-10-27 11:19:31 <feddy3> and then youtube closes down
476 2013-10-27 11:19:44 <feddy3> better to use the blockchain itself as your xor source
477 2013-10-27 11:19:46 <gmaxwell> THIS VIDEO IS UNAVAILABLE IN YOUR COUNTRY (on one of them)
478 2013-10-27 11:19:49 <CodeShark> lol
479 2013-10-27 11:19:50 <HM2> I heard the NSA has backdoor deals with safe manufacturers
480 2013-10-27 11:20:14 <gmaxwell> obviously you should first apply forward error correction.
481 2013-10-27 11:20:59 <gmaxwell> so then you only need a simple majority of the steganographic cat pictures to recover your coins.
482 2013-10-27 11:21:38 <CodeShark> u haz bitcoin?
483 2013-10-27 11:21:46 <gmaxwell> but really, "the attacker is finding bits of paper in my home/office" isn't a primary threat model.
484 2013-10-27 11:21:54 <gmaxwell> at least not for most people.
485 2013-10-27 11:22:13 <gmaxwell> and not relative to "oops, I forgot my crazy hiding scheme"
486 2013-10-27 11:28:05 <HM2> steganography is a cool concept but kind of useless generally
487 2013-10-27 11:28:36 <CodeShark> I've found a number of practical applications for it
488 2013-10-27 11:28:43 <sipa> well, it IS security through obscurity in a way
489 2013-10-27 11:28:47 <HM2> if you want to hide something at home you're better off burying it in your neighbours garden. if you want to smuggle digital content over a border you can just push it over the net
490 2013-10-27 11:28:57 <sipa> as the actual secret becomes the location
491 2013-10-27 11:29:06 <sipa> or even just the knowledge that the data exists
492 2013-10-27 11:29:19 <CodeShark> the point of steganography is not simply to hide information - but to not arouse suspicion you're hiding it in the first place
493 2013-10-27 11:29:28 <sipa> right
494 2013-10-27 11:30:05 <HM2> except all the steganography software on C: and google searches for "how to hide bitcoin wallets in jpegs"
495 2013-10-27 11:31:24 <sipa> it took me a while to realize you were talking about a filesystem and not a programming language :D
496 2013-10-27 11:31:37 <HM2> that's a good sign
497 2013-10-27 11:31:45 <HM2> (I don't have a C: drive either)
498 2013-10-27 11:31:48 <CodeShark> steganography can be as simple as "if we're meeting, post a picture of a puppy on your blog. otherwise, post a picture of a kitten"
499 2013-10-27 11:36:14 <melvster> when i see "Branch
500 2013-10-27 11:36:42 <sipa> 'coinbase block header' makes no sense
501 2013-10-27 11:36:55 <sipa> the coinbase is the first transaction in a block
502 2013-10-27 11:37:25 <melvster> sorry i meant coinbase.com
503 2013-10-27 11:37:34 <melvster> https://coinbase.com/network/blocks/0000000006d10c918748d91be4b27b1b2d31edbcfbc29dd40ce056ec297f8a6d
504 2013-10-27 11:37:52 <melvster> it has a block chain viewer
505 2013-10-27 11:37:55 <sipa> heh do they also have a b
506 2013-10-27 11:37:59 <sipa> lock explorer
507 2013-10-27 11:38:08 <melvster> but that field doesnt appear in other explorers
508 2013-10-27 11:38:18 <melvster> Branch
509 2013-10-27 11:38:25 <sipa> my guess in that case that it means that block is part of the active/main/best chain
510 2013-10-27 11:38:44 <sipa> and not an orphan/inactive/extinguished/reorganized/stale one
511 2013-10-27 11:38:52 <melvster> got it, thanks ... so branch "Testnet" may be another?
512 2013-10-27 11:38:57 <sipa> no
513 2013-10-27 11:38:58 <melvster> ohhh
514 2013-10-27 11:39:00 <melvster> i see
515 2013-10-27 11:39:04 <sipa> that is something else entirely
516 2013-10-27 11:39:06 <melvster> so the longest winning chain
517 2013-10-27 11:39:18 <sipa> longest valid chain, indeed
518 2013-10-27 11:39:23 <HM2> that's a lot of adjectives
519 2013-10-27 11:39:33 <melvster> got it, thanks sipa
520 2013-10-27 11:40:37 <melvster> im also working on a block explorer now
521 2013-10-27 11:40:51 <melvster> but i realized there's different nets, the main net, and the test net etc.
522 2013-10-27 11:40:59 <melvster> so i was thinking of marking which is which
523 2013-10-27 11:41:19 <melvster> but was unsure of the terminology ... I guess I could put in the genesis block, that should work
524 2013-10-27 11:41:31 <melvster> because main net and test net have different genesis
525 2013-10-27 11:41:56 <sipa> different networks
526 2013-10-27 11:42:24 <melvster> also yes
527 2013-10-27 11:42:34 <sipa> no, i mean, that is the terminology
528 2013-10-27 11:42:56 <melvster> oic thanks
529 2013-10-27 11:43:41 <melvster> i see yes, that's consistent with the white paper
530 2013-10-27 11:45:50 <melvster> ill put the genesis block in there too, as there's only ever going to be one of those ...
531 2013-10-27 11:46:06 <melvster> then i can do an explorer for both test and main
532 2013-10-27 13:04:35 <MrDaneelOlivaw> hello folks
533 2013-10-27 13:06:09 <MrDaneelOlivaw> does anyone know if electrum server supports the regtest regression test mode ?
534 2013-10-27 13:06:49 <MrDaneelOlivaw> also, seems like the electrum patch is not needed for latest bitcoin
535 2013-10-27 13:10:29 <sipa> regtest mode is pretty new
536 2013-10-27 13:11:12 <sipa> no idea about any patches
537 2013-10-27 13:12:18 <MrDaneelOlivaw> yeah i noticed, i'm a bit confused by it, is testnet returns no, i understand it is done for dev/testing so prod net rules, isolated, with minimal difficulty ?
538 2013-10-27 13:13:59 <sipa> testnet has special rule to keep its difficulty low, but it still has difficulty checks
539 2013-10-27 13:14:33 <MrDaneelOlivaw> ok so far i tried : prod, testnet and testnet in a box. Next I'm trying regtest and want to understand how it differs from the last.
540 2013-10-27 13:16:40 <MrDaneelOlivaw> i also confirm that building tip of bitcoin on latest ubuntu (saucy) is broken with libboost-dev-all (1.53 all packgs), i had to install manually the 1.54 dev packages (test, system, program_option, thread, chrono) to get it to build.
541 2013-10-27 13:16:51 <sipa> ok, different network magic (fabfb5da), subsidy halving every 150 blocks, initial difficulty = .0000000002328 (any block is good), and the genesis block is different
542 2013-10-27 13:17:48 <sipa> so creating blocks for regtest is trivial
543 2013-10-27 13:17:56 <sipa> and you get to see the effects of subsidy halving very quickly
544 2013-10-27 13:18:35 <sipa> MrDaneelOlivaw: good to know about saucy
545 2013-10-27 13:19:06 <MrDaneelOlivaw> fabfb5da is testnet and F9BEB4D9 is prod, so there are only these two in bitcoind without patching right ?
546 2013-10-27 13:19:48 <sipa> testnet is 0b110907
547 2013-10-27 13:20:04 <MrDaneelOlivaw> aha', i got confused by https://github.com/bitcoin/netspec/blob/master/protocol_specification.txt
548 2013-10-27 13:20:06 <sipa> mainnet is f9beb4d9
549 2013-10-27 13:20:48 <sipa> oh wow
550 2013-10-27 13:20:53 <sipa> i haven't seen that document in years :S
551 2013-10-27 13:21:05 <sipa> that may have been the magic of testnet1 or testnet2
552 2013-10-27 13:21:10 <MrDaneelOlivaw> lol bitcoin just has a few years
553 2013-10-27 13:21:28 <sipa> i've been aroun for a while
554 2013-10-27 13:21:42 <MrDaneelOlivaw> the more i read comments around git hub the more i notice ;P
555 2013-10-27 13:21:52 <sipa> ?
556 2013-10-27 13:21:58 <MrDaneelOlivaw> i've been around for a couple of years but mostly in the shadow
557 2013-10-27 13:22:02 <sipa> ic
558 2013-10-27 13:25:00 <MrDaneelOlivaw> there's a lot to lurk .. and apparently the changes in getblock to add the verbose flag means that electrum can work without patching bitcoin
559 2013-10-27 13:32:55 <MrDaneelOlivaw> I'm really surprised the btcd channel is empty
560 2013-10-27 13:34:06 <Belxjander> btcd?
561 2013-10-27 13:34:43 <swulf--> That's the go implementation, correct?
562 2013-10-27 13:34:57 <sipa> one of them
563 2013-10-27 13:35:06 <swulf--> there's more than one?:)
564 2013-10-27 13:36:08 <MrDaneelOlivaw> yeah the go one, as far as I know is the only complete one ? is bitcoinj just something that goes in front of bitcoin ?
565 2013-10-27 13:37:03 <sipa> bitcoinj is a completely separate implementation, mostly intended for implementing SPV clients
566 2013-10-27 13:37:04 <MrDaneelOlivaw> Well it is a client only by default but it does have an experimental full client mode
567 2013-10-27 13:37:12 <sipa> though it has an incomplete full validation mode
568 2013-10-27 13:37:25 <MrDaneelOlivaw> SPV clients means that electrum server are not necessary ?
569 2013-10-27 13:37:34 <sipa> the model is different
570 2013-10-27 13:37:55 <sipa> SPV clients participate directly in the P2P network, and validate PoW but don't validate transactions
571 2013-10-27 13:38:13 <MrDaneelOlivaw> and don't store the full blockchain, just headers right ?
572 2013-10-27 13:38:17 <sipa> indeed
573 2013-10-27 13:38:22 <sipa> not even all headers, just recent ones
574 2013-10-27 13:38:31 <sipa> electrum relies on a central server to index transactions clients are interested in
575 2013-10-27 13:38:46 <MrDaneelOlivaw> yes i know i've been playing with it
576 2013-10-27 13:38:53 <sipa> they do use the same crypto to prove that the transactions are actually part of the block chain
577 2013-10-27 13:39:00 <sipa> but electrum clients don't connect to the p2p network
578 2013-10-27 13:39:27 <MrDaneelOlivaw> is it commonly foreseen that SPV will take over electrum style client server ?
579 2013-10-27 13:39:38 <sipa> no idea
580 2013-10-27 13:39:57 <MrDaneelOlivaw> i'm trying to understand, what's the point of electrum old model when we have SPV
581 2013-10-27 13:40:16 <sipa> SPV has been around for much longer, it's even described in satoshi's whitepaper
582 2013-10-27 13:40:18 <MrDaneelOlivaw> much more reliable, the only thing I wonder is how long it takes for an SPV client to sync
583 2013-10-27 13:40:27 <sipa> for a new wallet? seconds
584 2013-10-27 13:40:39 <sipa> resyncing from scratch takes longer (minutes)
585 2013-10-27 13:40:46 <MrDaneelOlivaw> even old addresses ?
586 2013-10-27 13:40:47 <MrDaneelOlivaw> right
587 2013-10-27 13:40:51 <sipa> no, not old addresses
588 2013-10-27 13:41:00 <sipa> if you import things, you have to request history
589 2013-10-27 13:41:12 <MrDaneelOlivaw> this is the difference, electrum server caches each address with spendable outputs
590 2013-10-27 13:41:20 <sipa> yes, it relies on an index
591 2013-10-27 13:41:22 <sipa> SPV doesn't
592 2013-10-27 13:42:10 <MrDaneelOlivaw> I guess there could be an altorithm on top of SPV which caches the information close to where it is requested to improve speed sync
593 2013-10-27 13:42:12 <sipa> there is a difference in privacy too: iirc the electrum server knows exactly which addresses are yours
594 2013-10-27 13:42:19 <sipa> with SPV you can hide it
595 2013-10-27 13:42:23 <sipa> at least to an extent
596 2013-10-27 13:42:31 <MrDaneelOlivaw> sipa: the electrum server knows only when I spend which addresses are mine
597 2013-10-27 13:42:42 <sipa> MrDaneelOlivaw: your client asks for the balance of its addresses
598 2013-10-27 13:42:53 <MrDaneelOlivaw> i can ask for any address sipa
599 2013-10-27 13:42:57 <MrDaneelOlivaw> doesn't mean its mine
600 2013-10-27 13:43:01 <sipa> well, sure
601 2013-10-27 13:43:02 <MrDaneelOlivaw> until i sign something
602 2013-10-27 13:43:08 <sipa> but it can see which ones you are interested in
603 2013-10-27 13:43:14 <sipa> that is a (mild) privacy leak
604 2013-10-27 13:43:17 <MrDaneelOlivaw> but then again the SPV peers will see this too
605 2013-10-27 13:43:20 <sipa> no
606 2013-10-27 13:43:28 <sipa> you don't give your addresses, you give a filter
607 2013-10-27 13:43:37 <sipa> and the filter has false positives
608 2013-10-27 13:43:44 <MrDaneelOlivaw> no but at some point i need to send out my transaction
609 2013-10-27 13:43:52 <MrDaneelOlivaw> i guess it could be just relying
610 2013-10-27 13:44:08 <MrDaneelOlivaw> i.e. a bit like tor/freenet
611 2013-10-27 13:44:09 <sipa> but SPV doesn't connect to a single server
612 2013-10-27 13:44:13 <MrDaneelOlivaw> no, to N
613 2013-10-27 13:44:16 <sipa> it just broadcasts transactions on the network
614 2013-10-27 13:44:19 <sipa> just like everyone else
615 2013-10-27 13:44:27 <sipa> yes, that also allows for some analysis, indeed
616 2013-10-27 13:44:30 <sipa> but it's a lot weaker
617 2013-10-27 13:44:36 <MrDaneelOlivaw> right, so credible deniability
618 2013-10-27 13:44:48 <sipa> at least in theory
619 2013-10-27 13:45:01 <sipa> because of a node identifies itself as BitcoinJ, you know it won't be relaying any transactions
620 2013-10-27 13:45:10 <sipa> so anything it broadcasts can be assumed to be its own
621 2013-10-27 13:45:18 <MrDaneelOlivaw> makes sense
622 2013-10-27 13:45:30 <sipa> then again, you can switch peers all the time
623 2013-10-27 13:45:57 <sipa> anyway, currently the SPV model does put rather high load on the nodes it requests transactions from
624 2013-10-27 13:46:52 <MrDaneelOlivaw> i can imagine but then so does a full client synching
625 2013-10-27 13:47:17 <sipa> yes, but a bit differently
626 2013-10-27 13:50:40 <MrDaneelOlivaw> less cpu intensive and more network intensive perhaps
627 2013-10-27 13:50:45 <MrDaneelOlivaw> or IO in general
628 2013-10-27 13:51:02 <MrDaneelOlivaw> I guess the blockchain doesn't compress well at all
629 2013-10-27 13:51:05 <sipa> indeed, disk/cpu vs network is different
630 2013-10-27 13:51:18 <MrDaneelOlivaw> being full of entropy
631 2013-10-27 13:51:33 <sipa> no attempt is even made to compress it
632 2013-10-27 13:51:41 <sipa> you could get 20-40% off
633 2013-10-27 13:52:17 <MrDaneelOlivaw> not too bad but depends on at what cost of cpu/time and what algorithm
634 2013-10-27 13:58:29 <sipa> gzip -9 removes some 25%
635 2013-10-27 13:58:34 <sipa> xz -9 removes 40%
636 2013-10-27 13:59:07 <sipa> (and that's on a bunch of blocks together, as you still need to access them individually, you can only compress them individually, with likely worse compression)
637 2013-10-27 14:23:34 <sipa> BlueMatt, gmaxwell: just ran comparison tool on #3163 + #3164... it took hours and resulted in:
638 2013-10-27 14:23:37 <sipa> Blocks which were not handled the same between bitcoind/bitcoinj: 361
639 2013-10-27 17:01:52 <sipa> wth testnet difficulty is 18800? :o
640 2013-10-27 17:02:14 <sipa> even with a bunch of asics that requires hours to mine...
641 2013-10-27 17:05:30 <MrDaneelOlivaw> sipa: I hope people won't starting selling them again
642 2013-10-27 17:07:48 <sipa> MrDaneelOlivaw: we can always reset it again
643 2013-10-27 17:09:07 <phantomcircuit> sipa, that's probably my fault
644 2013-10-27 17:09:08 <phantomcircuit> sorry
645 2013-10-27 17:09:17 <MrDaneelOlivaw> sipa: sure but the difficulty adjusts more often on testnet so if it is so hard is because people are furious on it so even if you reset won't take much at all before things are back ?
646 2013-10-27 17:10:10 <sipa> MrDaneelOlivaw: no, it adjusts just as frequently
647 2013-10-27 17:10:28 <sipa> ;;genrate 10200
648 2013-10-27 17:10:29 <gribble> The expected generation output, at 10200.0 Mhps, given difficulty of 390928787.638, is 0.0131216902664 BTC per day and 0.000546737094432 BTC per hour.
649 2013-10-27 17:10:33 <gribble> The expected generation output, at 10200.0 Mhps, given difficulty of 18800.0, is 272.85353551 BTC per day and 11.3688973129 BTC per hour.
650 2013-10-27 17:10:33 <sipa> ;;genrate 10200 18800
651 2013-10-27 17:11:03 <MrDaneelOlivaw> sipa: oh, for some reason i thought otherwise. maybe just on test net in a box or maybe i must have been looking at some patches
652 2013-10-27 17:11:34 <sipa> MrDaneelOlivaw: testnet in a box is just testnet on a separate chain
653 2013-10-27 17:11:38 <sipa> it has all the same rules
654 2013-10-27 17:12:12 <sipa> this is crazy... testnet difficulty is higher than it was on mainnet when i first mined :p
655 2013-10-27 17:12:13 <MrDaneelOlivaw> fair enough
656 2013-10-27 17:58:47 <dobry-den> sipa: I don't know how personal this question is, but I like how I see the big bitcoin contributors like you in here all time. Is this something you do as a full-time hobby or do you do it alongside $dayjob?
657 2013-10-27 18:08:21 <BlueMatt> sipa: seems like there is an issue when running with the "expensive tests"
658 2013-10-27 18:08:26 <BlueMatt> sipa: if you set that flag to 0 it works
659 2013-10-27 18:17:27 <melvster> anyone know if there's any PHP code than can verify any of block headers / merkle trees / tx sigs / scripts ?
660 2013-10-27 18:18:11 <BlueMatt> melvster: if it does exist, its yet another weekend reimplementation in $MY_FAVORITE_LANGUAGE