1 2016-06-13 07:13:50 <benj2> hello. I'm working on a wallet for exchanges
2 2016-06-13 07:14:15 <benj2> NBitcoin offers quite a lot very useful functions. are some of these available in other stacks?
3 2016-06-13 07:14:33 <benj2> I'm familiar with bitcoind API, bitcoinJ, python-bitcoin
4 2016-06-13 09:05:58 <rubensayshi> hmm `bitcoin-cli estimatefee 1` used to work right? when the fees were lower and it was still able to estimate a sane for for nblocks=1 ?
5 2016-06-13 09:23:56 <benj2> is there any work on some java stuff besides bitcoinj?
6 2016-06-13 10:43:01 <wumpus> estimatefee 1 doesn't work anymore with recent versions - it's not possible to estimate the fee necessary for confirmation in one block
7 2016-06-13 10:43:20 <wumpus> (it never was, but older versions returned a pretty useless guess in that case)
8 2016-06-13 20:47:37 <murch> Hey, I'm trying to understand ApproximateBestSubset in Wallet.cpp. in #1789 it calls insecure_rand(). Does insecure_rand() give back true or false in 50:50 chance?
9 2016-06-13 20:47:48 <murch> https://github.com/bitcoin/bitcoin/blob/master/src/wallet/wallet.cpp#L1789
10 2016-06-13 21:27:31 <instagibbs> murch, no, it gives the first bit as "1" with 50/50 chance, if you assume it's secure
11 2016-06-13 21:27:37 <instagibbs> err, random, or whatever
12 2016-06-13 21:27:49 <instagibbs> it then does &1 with that value
13 2016-06-13 21:27:52 <instagibbs> to only take that value
14 2016-06-13 22:17:54 <murch> instagibbs: Thanks, that's what I wanted to know.
15 2016-06-13 22:20:35 <murch> I.e. this would mean that each utxo is added to the inputs with a 50% chance, but they're sorted, so either the smaller or larger get picked much more often.
16 2016-06-13 22:22:28 <murch> Ah yes, descendig order, so large ones get picked more often. Might have something to do with global UTXO numbers growing.
17 2016-06-13 22:23:19 <murch> In my simulation I had been picking at random, but I'd bet that picking from the largest with a 50% chance would make a lot of difference.
18 2016-06-13 22:35:07 <luke-jr> anyone care to review? https://github.com/luke-jr/eloipool/pull/12
19 2016-06-13 22:40:12 <kanzure> what was openfec being used for recently? what about https://tahoe-lafs.org/trac/zfec