1 2015-04-18 00:08:09 <StephenM347> http://imgur.com/3hLMbbT <-- Parsing blocks on disk, I can see the exact point in the blockchain I downloaded the headers-first client.
 2 2015-04-18 00:09:22 <StephenM347> The height of the peaks correspond to number of blocks that were downloaded without having all the previous blocks of that block.
 3 2015-04-18 01:21:12 <denisx> has anybody a bitcoin running in testnet mode with less than 262k difficulty?
 4 2015-04-18 01:34:28 <justanotheruser> denisx: could you rephrase?
 5 2015-04-18 01:38:45 <denisx> justanotheruser: I made a port of ckpool to freebsd, but before I use it I would to test it on testnet, but my old miners have a hard time finding a block on testnet because the difficulty is 262k
 6 2015-04-18 01:39:04 <gmaxwell> denisx: just start two isolated nodes yourself?
 7 2015-04-18 01:39:29 <gmaxwell> you can use invalidate block to pin it wherever you want in history, perhaps right by a point where it recently warped back to diff 1.
 8 2015-04-18 01:39:37 <justanotheruser> Is there a reason you can't do 1 node? Is it just the way it is?
 9 2015-04-18 01:39:48 <gmaxwell> er "invalidateblock" it's a hidden rpc.
10 2015-04-18 01:40:11 <gmaxwell> (it's just hidden because its kind of a footgun, and not normally useful except for testing)
11 2015-04-18 01:40:31 <justanotheruser> that's valid on mainnet?
12 2015-04-18 01:40:38 <gmaxwell> whats valid on mainnet?
13 2015-04-18 01:40:44 <justanotheruser> invalidateblock?
14 2015-04-18 01:40:47 <gmaxwell> sure
15 2015-04-18 01:41:04 <justanotheruser> neato, I can become a checkpointing authority with just a few lines of code then
16 2015-04-18 01:41:15 <gmaxwell> its matching partner is "reconsiderblock"
17 2015-04-18 01:41:36 <gmaxwell> justanotheruser: stupid use will make us take it out! :P
18 2015-04-18 02:57:31 <denisx> gmaxwell: I resetted my testnet and invalidatet the 5th block, but now bitcoind is reporting "Bitcoin is downloading blocks..." for getblocktemplate
19 2015-04-18 02:58:21 <denisx> ah, because it still sees other clients who say they have more than 5?
20 2015-04-18 02:59:16 <gmaxwell> you canot mine if you're below the highest checkpoint (which iirc is at block 500 or so for testnet)
21 2015-04-18 02:59:36 <denisx> ok, thanks, I will try that
22 2015-04-18 03:00:00 <gmaxwell> you could run with checkpoints=0  or move up to 500 or so. there is actually a block which includes a transaction with a 100k TNBC fee maybe around 20k or so that I put in specifically for people to fork there for testing.
23 2015-04-18 03:06:32 <denisx> I tried 600, nothing changed
24 2015-04-18 03:06:44 <denisx> gmaxwell: should I jump after that block at 20k?
25 2015-04-18 03:11:18 <gmaxwell> denisx: grr. last changes we made to IsInitialBlockDownload make your life hard. are you in a position to edit the code? If so add "return false;" to the first line inside IsInitialBlockDownload in main.cpp
26 2015-04-18 03:11:33 <gmaxwell> we should probably add an override hidden option for that.
27 2015-04-18 03:14:34 <Luke-Jr> see also https://github.com/bitcoin/bitcoin/pull/5987
28 2015-04-18 03:22:04 <denisx> hope this helps, if not my last idea would be to direct my pool to testnet for some seconds ;)
29 2015-04-18 03:57:02 <denisx> gmaxwell: ok, thanks, that did help
30 2015-04-18 04:29:43 <TYDIRocks> Can a normal private key be paired with multiple chain codes to create multiple extended keys? (BIP32)
31 2015-04-18 10:11:07 <ScarfFace> Hey everybody, I want to setup a Service, where you can post a TX to be relayed only under certain conditions (block height, output X spent, timestamp, ...) - this will be free to use for anybody, I will need it for precondition for another project though, and I want it to be on an dedicated server.. I would need a webserver, to add TXs, and probab
32 2015-04-18 10:11:07 <ScarfFace> ly a bitcoind to listen for blocks and incoming TXs.. I have used Java a lot for Android and Tomcat development, so I defo feel most comfortable using it for this. Since this isn't too problematic security wise, I will start off with a cheap VPS. Any tips for the integration of bitcoind into Java? I would need a callback for each incoming TX/Block,
33 2015-04-18 10:11:07 <ScarfFace>  together with a Database-Handler and a way to validate/post a TX (probably the easiest step)..
34 2015-04-18 10:11:35 <ScarfFace> Hm .. reading this, I think this might be better suited in the forum
35 2015-04-18 10:26:12 <ScarfFace> Ah I see, bitcoinj has a BlockChainListener, I guess this will come in handy
36 2015-04-18 16:03:41 <phantomcircuit> ScarfFace, rpc should work well enough
37 2015-04-18 16:03:57 <phantomcircuit> ScarfFace, can you expand on why you want the triggers in the first place?
38 2015-04-18 16:42:41 <ScarfFace> oh nice, someone responden :) .. phantomcircuit: I want to try to look into the lightning network/payment channels, try some things out ... many of the trust issues are resolved by transactions being sent in certain conditions. this usually requires all participants to be online all the time, so I thought of such a system
39 2015-04-18 16:43:35 <ScarfFace> and thats why I want it to be separate and autonomous aswell
40 2015-04-18 16:43:48 <ScarfFace> most likely open source aswell - pretty simple anyway
41 2015-04-18 16:46:34 <ScarfFace> rpc means having bitcoind in background and accessing it via the json-api, right? do you think this is easier/better, than implementing bitcoinj, given it already has the right listener in place
42 2015-04-18 17:22:56 <hulkhogan42o> have there been any announced efforts to unit test #5885? asking out of interest (and curiosity in understanding RNG test process)
43 2015-04-18 17:26:43 <gmaxwell> 5885 is just a discussion PR, it's not something thats on a path at the moment to actively go in.  What we'd want to do there is show that the fortuna implementation is faithful to the spec (or at least agress with independant fortuna implementations)-- e.g. use another one to build test vectors.  We'd want to estimate the entropy of our snakeoil randomness gatherer and make sure that it's not broken
44 2015-04-18 17:26:49 <gmaxwell> , and we'd want to write tests for the combiner (that isn't in there yet) that show that the compiler correctly depends on both inputs and still is random if either input is constant.
45 2015-04-18 17:27:05 <gmaxwell> It sounds like from discussion the direction will be to break this work into a seperate library.
46 2015-04-18 17:28:02 <hulkhogan42o> yeah, i saw the comment about maybe looking at the nist stats tests for the entropy testing
47 2015-04-18 17:28:46 <hulkhogan42o> i was just curious, i suppose in terms of if anyone had started writing a test harness for whats in place (or if there is much of a need for that, even)
48 2015-04-18 17:29:19 <gmaxwell> The nist tests/dieharder are not really of much value in something like this, it's fine to do... but unless something is grossly wired the fortuna output is just an AES stream.
49 2015-04-18 17:29:55 <hulkhogan42o> ah, hmm
50 2015-04-18 17:30:03 <gmaxwell> (and if the gross miswiring is on the wrong side of the aes, dieharder will pass even if the construction is totally insecure)
51 2015-04-18 17:30:49 <hulkhogan42o> i think i probably misunderstood, since my understanding of the stats tests were they tested the entropy
52 2015-04-18 17:31:09 <hulkhogan42o> ansd not the output, but yes it seems more reading req.
53 2015-04-18 17:32:45 <gmaxwell> Well the comment was output; but yea, not useful there. For input, thats another matter.  So a problem there is that many of the inputs are not very random. Thats fine. We feed them to sha512.  The output of sha512 is obviously going to pass any test of randomness.
54 2015-04-18 17:34:21 <hulkhogan42o> yeah, there was a page i located that discussed unit testing fortuna in depth, unfortunately via comparing implementations, the author of that was using pycrypto as comparison
55 2015-04-18 17:35:36 <gmaxwell> hulkhogan42o: so the whole of the inputs will not past the nist test; the output from the input compression will (even if there is almost no entropy in the inputs).  Probably if you wanted to do something with a randomness test suite you'd test things some of the indivigual input parts, to estimate their entropy. But I dunno if this really has value. our assumption is that in aggreage all that input
56 2015-04-18 17:35:42 <gmaxwell> is not very random and is only done as a last ditch defense against the stronger mechenisms failing.
57 2015-04-18 17:37:25 <hulkhogan42o> it doesnt seem like it would have mmuch value, i was mainly tipped off by dieharder and its role in statistical testing of rngs- but it was a good point you made re: comparing implementations testing, and seeing as how others already started (and wrote) about doing this, seems like a more likely area for `value-add`
58 2015-04-18 17:37:58 <hulkhogan42o> gmaxwell: er, forgot to prefix
59 2015-04-18 17:39:31 <hulkhogan42o> ok, that gives me a starting point for exploration, thanks for the tip !
60 2015-04-18 18:02:53 <Luke-Jr> erm
61 2015-04-18 18:03:01 <Luke-Jr> !ops