1 2014-12-06 00:50:16 <sipa> lechuga_: sure, it's a bitcoin specific thing
 2 2014-12-06 00:52:10 <lechuga_> ok, so if you write a wallet and you use openssl and for whatever reason it's impractical for you to use libsecp256k1 then don't upgrade your txs to v3
 3 2014-12-06 00:52:44 <lechuga_> unless you're manually re-encoding with a low S value
 4 2014-12-06 00:57:04 <lechuga_> maybe it's worth mentioning openssl's behavior in this regard explicitly in the bip as is done wrt to its permissiability of noncanonical sigs for verification
 5 2014-12-06 00:57:20 <lechuga_> if u think so i can make a PR
 6 2014-12-06 01:43:47 <sipa> lechuga_: go ahead
 7 2014-12-06 01:43:58 <sipa> some example code about how to do the reencoding would be useful too
 8 2014-12-06 01:47:01 <lechuga_> k
 9 2014-12-06 01:48:28 <sipa> bitcoin has had such code until recently, before switching to libsecp256k1
10 2014-12-06 01:48:35 <cidon> hi how do i install bitcoin-qt client for linux?
11 2014-12-06 01:50:01 <sipa> #bitcoin please
12 2014-12-06 02:19:51 <Belxjander> is the torrented bootstrap.dat in any particular format?
13 2014-12-06 04:39:48 <BlueMatt> how strong are our guarantees that wallet transactions will always be in memory pool?
14 2014-12-06 04:40:28 <BlueMatt> actually, nvm
15 2014-12-06 06:57:10 <wumpus> Belxjander: the standard serialized block format as used everywhere
16 2014-12-06 06:58:20 <wumpus> Belxjander: the bootstrap file has been generated with the linarize.py script, which writes them head-to-tail and make sure there are no gaps in between, which sometimes can happen in the blkXXXX.dat files
17 2014-12-06 09:11:48 <aschildbach> saivann: Hi! What does "custodial" mean? I tried translating that to German but it doesn't make any sense in the context of your guideline proposal.
18 2014-12-06 09:29:39 <paveljanik> How can I read src/util.cpp:instance_of_cinit;?
19 2014-12-06 09:30:00 <paveljanik> I understand what it should do. But how it is done?
20 2014-12-06 09:30:20 <paveljanik> Is it to me unknown compiler or language feature?
21 2014-12-06 09:31:07 <paveljanik> looks like it is boostism 8)
22 2014-12-06 09:35:28 <average> paveljanik: boost heh.. was it that knowing modern C++ equates to knowing Boost ? :)
23 2014-12-06 09:36:59 <paveljanik> hmm, but boost has only is_instance_of_2...
24 2014-12-06 09:38:01 <paveljanik> but instance_of is in the preprocessor output, so it has to be language/compiler feature not boost one.
25 2014-12-06 09:38:40 <average> paveljanik: maybe dig deep into the boost docs and source code
26 2014-12-06 09:49:30 <paveljanik> and even more interesting, the class itself is class CInit ;-)
27 2014-12-06 09:55:15 <wumpus> paveljanik: it's normal c++ syntax - a class is defined, and an instance of that class is defined in the same statement
28 2014-12-06 09:55:49 <wumpus> normally you'd do class CInit {};   CInit instance_of_cinit; this is just shorthand
29 2014-12-06 09:56:04 <paveljanik> wumpus: I'm now reading the C++ books I have and there is no word about it 8)
30 2014-12-06 09:56:29 <paveljanik> ah - there is no ; there 8)))
31 2014-12-06 09:56:31 <wumpus> this existed in C!
32 2014-12-06 09:56:35 <paveljanik> I'm blind.
33 2014-12-06 09:57:08 <paveljanik> the coding staly is a bit new to me. newline between class decl and instance...;
34 2014-12-06 09:57:11 <paveljanik> style
35 2014-12-06 09:57:11 <wumpus> the motivation for this construction is to run some code at startup (before main) and at shutdown (after main)
36 2014-12-06 09:57:20 <paveljanik> yup, this is clear.
37 2014-12-06 09:57:40 <paveljanik> the newline there makes this unreadable - at least for me ;-)
38 2014-12-06 09:57:42 <wumpus> just have patience while reading code, and if you don't at first understand it keep starting at it until you do
39 2014-12-06 09:58:06 <paveljanik> this is why I run the preprocessor 8)
40 2014-12-06 09:58:08 <wumpus> don't start off complaining or being annoyed, it gets in the way of reviewing what the code *does*
41 2014-12-06 09:59:09 <paveljanik> in this case, it was clear what it does, but how it does it was not clear to me - I never write code like
42 2014-12-06 09:59:11 <paveljanik> int
43 2014-12-06 09:59:12 <paveljanik> a;
44 2014-12-06 09:59:13 <paveljanik> 8)
45 2014-12-06 10:00:21 <paveljanik> Thanks for a free lesson :-)
46 2014-12-06 10:10:14 <paveljanik> I'd call C blindness ;-)
47 2014-12-06 11:31:59 <average> meow
48 2014-12-06 14:09:29 <wumpus> don't let your cat on the keyboard :p
49 2014-12-06 16:19:08 <op_null> can anybody confirm or deny what this fool is saying? https://bitcointalk.org/index.php?topic=883675.0
50 2014-12-06 16:22:41 <op_null> as much as I can find, yes that's a problem if people have been relying on that function and leaving their RPC port exposed.
51 2014-12-06 16:23:09 <sipa> well there may be other local users
52 2014-12-06 16:23:28 <sipa> if we weren't concerned about security of the RPC even if non-exposed, it shouldn't a need a password at all
53 2014-12-06 16:23:37 <sipa> if there is one, i guess it needs some security
54 2014-12-06 16:25:30 <op_null> in other places like the OSX instructions it's bytes being read from /dev/urandom, so that's fine.
55 2014-12-06 16:28:41 <op_null> (I called him a fool because this is the guy who claimed to have broken ECDSA before and wouldn't shut up about it)
56 2014-12-06 16:53:38 <gmaxwell> op_null: sure, that was known and fixed two years ago... debian submitted packaging files were brain damaged.
57 2014-12-06 16:53:58 <gmaxwell> Indpendantly the Fedora package did the same thing, and I caught it before they got anyone to use it.
58 2014-12-06 16:55:09 <gmaxwell> (the original stupidity there was someone in the debian project, not Matt... he just bulk imported their existing packaging stuff, I believe)
59 2014-12-06 16:59:31 <HM2> you could always default to unix sockets for the RPC and then do peer verification
60 2014-12-06 16:59:42 <HM2> not sure if that's portable to other unixes
61 2014-12-06 17:00:06 <HM2> that's what databases like postgres default to
62 2014-12-06 17:00:09 <op_null> gmaxwell: right, I didn't realised he linked to the 0.9.1 tag. well that's a whole load more annoying from the guy.
63 2014-12-06 17:01:50 <gmaxwell> 0.8.1
64 2014-12-06 17:02:31 <op_null> typo.
65 2014-12-06 17:02:40 <gmaxwell> HM2: yea, I've mentioned domain sockets before... but it's not that portable (e.g. not windows). We already bind only to localhost unless told otherwise.
66 2014-12-06 17:02:59 <gmaxwell> and you must know the rpcpassword... so..
67 2014-12-06 17:06:25 <HM2> i think once someone has the ability to connect to loopback you're toast anyway
68 2014-12-06 17:06:49 <HM2> i think Linux has peer id for TCP
69 2014-12-06 17:07:01 <HM2> you could probably ifdef that in if you felt like it
70 2014-12-06 17:09:58 <HM2> hmm, perhaps not
71 2014-12-06 21:41:49 <English-> hey folks, just messing around with the code. Using reindex to go through the blockchain, to look at inputs/outputs for non coinbase transctions. I seem to be having an issue with the client crashing on IsPayToScriptHash. This occurs when an output from one transaction is used as the input for another, within the same block
72 2014-12-06 21:42:17 <English-> ExtractDestination is called first, on the input, which crashes when IsPayToScriptHash is called
73 2014-12-06 21:42:45 <English-> anyone know why this might be....? :)
74 2014-12-06 23:26:52 <BlueMatt> testnet coins? mxbQiyEaekyFH858KcKzV4NyAeMDYiRT4v