1 2015-05-02 00:01:41 <phantomcircuit> hulkhogan_, huh?
  2 2015-05-02 00:04:46 <hulkhogan_> er, i'm trying to test the effect of a single peer message on a node, there isn't a way to do that probably via RPC, is there? (i noticed there was a new p2p testing framework committed, but i wasn't sure if that would help me any)
  3 2015-05-02 00:05:14 <hulkhogan_> *its a custom inv message, specifically
  4 2015-05-02 00:20:20 <hulkhogan_> I suppose, since I'm just testing the behavior of sending large messages, I just want to understand if a thrown exception will cause the sending node to shutdown (or whatever the behavior might be)
  5 2015-05-02 00:21:09 <gmaxwell> We silently discard oversized messages.
  6 2015-05-02 00:21:21 <gmaxwell> (if you mean recieving an oversized message)
  7 2015-05-02 00:23:24 <hulkhogan_> well actually i noticed 5843 a second ago, which cleanly answered my question about the receiver, but actually i was inquiring about the effect of sending such messages (and throwing , say, in a check before PushMessage)
  8 2015-05-02 00:24:26 <hulkhogan_> (basically was trying to see if i could address sipa's nit above, ie 'what happens when we try sending an oversized message')
  9 2015-05-02 00:31:20 <hulkhogan_> no worries though, this was my first dig through the p2p code, i think ill be able to figure it out after a bit more exploration... wasn't as hairy as my imagination led me to believe it would be..
 10 2015-05-02 01:32:05 <warren> phantomcircuit: ping
 11 2015-05-02 01:54:34 <notplato> When I compile bitcoin-qt, I use make from the directory with the src subdirectory.  Is that the right place from which to run make check?
 12 2015-05-02 01:56:17 <hulkhogan_> most probably
 13 2015-05-02 01:56:37 <hulkhogan_> there ought to be some relevant output too
 14 2015-05-02 01:57:00 <notplato> When I run make bitcoin_test_clean, it only works from the src subdirectory.  (No such target from the parent folder)
 15 2015-05-02 01:57:36 <notplato> Ok.  I've been doing make check from src too, and it passes for bitcoind, but fails for bitcoin-qt.  I'll try it from the parent.  Thanks.
 16 2015-05-02 01:58:39 <hulkhogan_> i think -qt is built from /src/qt, so that might be why its failing
 17 2015-05-02 01:59:17 <hulkhogan_> and the `test clean` routine is in the src/ makefile so thats why it only works there
 18 2015-05-02 02:00:28 <hulkhogan_> *er, 'bitcoin_test_clean'
 19 2015-05-02 02:01:21 <hulkhogan_> notplato: try running `make check` from src/qt if you want to run that for -qt
 20 2015-05-02 02:06:47 <notplato> make check from src/qt: FAIL: qt/test/test_bitcoin-qt.exe ================================= 1 of 1 test failed Please report to info@bitcoin.org <-- From Nitrogenics build instructions, used against master (from 16 commits ago).
 21 2015-05-02 02:08:23 <notplato> And no src/qt/test/test_bitcoin-qt.log was produced.  No output at all from qt/test/test_bitcoin-qt.exe.
 22 2015-05-02 02:12:45 <hulkhogan_> hm, not sure how much help i'll be.. i did a windows build maybe once or twice, and it sadly was a clunky affair
 23 2015-05-02 02:13:07 <hulkhogan_> i can say that command works on linux, so there is some hope for your cause ;p
 24 2015-05-02 02:13:30 <hulkhogan_> (possibly check your dependencies?)
 25 2015-05-02 02:16:20 <notplato> Some people in here must compile bitcoin-qt for Windows.  Is it klunky for everyone?
 26 2015-05-02 02:16:32 <hulkhogan_> are you using mingw?
 27 2015-05-02 02:16:38 <notplato> yes
 28 2015-05-02 02:17:02 <notplato> I'm following Nitrogenics instructions at https://bitcointalk.org/index.php?topic=149479.
 29 2015-05-02 02:17:09 <hulkhogan_> looking
 30 2015-05-02 02:17:43 <hulkhogan_> i think ive followed these before.. why do you need to run `make check`?
 31 2015-05-02 02:18:20 <hulkhogan_> ah, i remember that i had trouble compiling qt, hmm
 32 2015-05-02 02:18:21 <notplato> I wanted to make a patch so that the Windows version as compiled with his inx could run the unit tests.
 33 2015-05-02 02:18:46 <notplato> It compiled qt and I can run it just fine (master minus 16 commits).
 34 2015-05-02 02:18:47 <hulkhogan_> inx ?
 35 2015-05-02 02:18:52 <notplato> instructions.
 36 2015-05-02 02:18:58 <hulkhogan_> ea
 37 2015-05-02 02:19:00 <hulkhogan_> yea*
 38 2015-05-02 02:19:48 <hulkhogan_> thats interesting, wish i knew more about the tests, i guess it wouldnt be related to boost probably
 39 2015-05-02 02:20:39 <hulkhogan_> well
 40 2015-05-02 02:20:42 <hulkhogan_> looking over these
 41 2015-05-02 02:20:49 <hulkhogan_> it looks like he adds a flag called '--disable-tests'
 42 2015-05-02 02:20:54 <hulkhogan_> you turned that off right
 43 2015-05-02 02:20:55 <hulkhogan_> ?
 44 2015-05-02 02:22:01 <hulkhogan_> ie 'Boost must be configured with --with-test option'
 45 2015-05-02 02:22:17 <hulkhogan_> might take that route-
 46 2015-05-02 02:22:32 <notplato> Well, nitro's instructions might only cover the unit tests for test_bitcoin.exe.  Maybe qt quits because there is other stuff missing.  It compiles it without error.
 47 2015-05-02 02:23:05 <notplato> Yes, I removed --disable-tests from the configure command, and recompiled boost with the --with-test option.
 48 2015-05-02 02:23:43 <notplato> Do you have a test_bitcoin-qt.exe, and if so, does it generate output when you run it?
 49 2015-05-02 02:23:55 <hulkhogan_> it would be wierd if the -qt tests were incompatible with win32..
 50 2015-05-02 02:25:26 <hulkhogan_> yea, and its passing.
 51 2015-05-02 02:25:43 <hulkhogan_> this is on x64 linux, though
 52 2015-05-02 02:25:43 <notplato> Well, I confirmed the behavior is the same with or without the changes I made to avoid Nitro's requirement that Win32 builders install an external hexdump and then alter Makefile.test.include.  I'll just let him know that.
 53 2015-05-02 05:14:24 <wumpus>  ================================= 1 of 1 test failed Please report to info@bitcoin.org <- where does info@bitcoin.org even go? we should change the instructions to point to the github issue tracker, probably :)
 54 2015-05-02 11:14:53 <harding> wumpus: I have no idea where info@bitcoin.org goes.  I'll ask the person who maintains the mail aliases.
 55 2015-05-02 15:12:25 <Luke-Jr> wumpus: I don't know what the .rc is used for - shoudl I just revert it?
 56 2015-05-02 15:12:55 <sipa> i think it's for the "properties" of the windows .exe file
 57 2015-05-02 15:13:32 <jgarzik> nod
 58 2015-05-02 15:14:50 <Luke-Jr> make sense, though I don't know if that means I should pull out its changes or leave them
 59 2015-05-02 15:19:32 <Luke-Jr> makes*
 60 2015-05-02 16:57:41 <dgenr8> has anyone ever promoted a minimum amount per output hard fork?
 61 2015-05-02 16:57:51 <sipa> that would be a soft fork
 62 2015-05-02 16:57:57 <sipa> and you'd be lynched :)
 63 2015-05-02 17:00:47 <dgenr8> why is it important to be able to make dust?
 64 2015-05-02 17:01:23 <sipa> it isn't
 65 2015-05-02 17:01:30 <sipa> but what 'dust' means may change over time
 66 2015-05-02 17:01:51 <dgenr8> mmm... well so can rules
 67 2015-05-02 17:02:07 <sipa> policy is much easier to change than consensus rules
 68 2015-05-02 17:02:42 <sipa> and consensus rules are intended to be pretty much not changes - that's what they are: consensus, what everyone agrees on
 69 2015-05-02 17:02:55 <sipa> people can individually enforce different rules on top
 70 2015-05-02 17:04:35 <dgenr8> the cost of hoping for frozen consensus rules appears to be significant
 71 2015-05-02 17:04:57 <sipa> it should be significant
 72 2015-05-02 17:05:10 <sipa> changes to the consensus rules can change bitcoin's economic properties very significantly
 73 2015-05-02 17:05:28 <sipa> they shouldn't be possible without very widespread agreement that the change is useful
 74 2015-05-02 17:05:43 <dgenr8> quite agreed
 75 2015-05-02 17:06:09 <sipa> and yes, this has a cost - the cost is that we can't experiment easily
 76 2015-05-02 17:10:30 <dgenr8> benefits of minimum output amount... network congenstion relief, block size relief (at least 2 ways -- incentive for fewer outputs, incentive for fewer non-economic transactions), one less reason for IsStandard, higher cost to anti-privacy dust painting
 77 2015-05-02 17:36:42 <dgenr8> ... UTXO set size or at least growth rate reduction, resulting in memory and disk usage relief and some validation speed increase, plus a follow-on effect as child spends would now have fewer inputs on average
 78 2015-05-02 17:55:08 <hrzn> WOW:O vid.me/vCPc
 79 2015-05-02 18:50:10 <jgarzik> jonasschnelli, was there a problem with the picocoin PRs/issues, or just no movement?
 80 2015-05-02 18:50:46 <jonasschnelli> hm.... they where open for over such a long time. I though they are not of a interest.
 81 2015-05-02 18:50:54 <jonasschnelli> but can reopen if you care
 82 2015-05-02 18:52:05 <jonasschnelli> jgarzik: maybe https://github.com/jgarzik/picocoin/pull/22 is interesting. But its from March 2014. I would have to check again if my ideas where correct
 83 2015-05-02 18:57:28 <jgarzik> jonasschnelli, they are definitely of interest
 84 2015-05-02 18:57:43 <jgarzik> jonasschnelli, that's the main thing with each PR - need to check each for relevancy.  if still useful, I'll merge rapidly.
 85 2015-05-02 18:58:20 <jonasschnelli> but let me first re-check https://github.com/jgarzik/picocoin/pull/22
 86 2015-05-02 18:59:07 <jgarzik> jonasschnelli, completing the SPV client to HD wallet status is still a goal, though much lower priority alas :)
 87 2015-05-02 18:59:21 <jgarzik> jonasschnelli, as is making libccoin a useful all-round library for C
 88 2015-05-02 18:59:37 <jonasschnelli> Yes. Some minutes ago i was thinking about trying to add HD to picocoin
 89 2015-05-02 18:59:45 <jgarzik> jonasschnelli, it is capable of fully validating the chain, though it is pre-BIP66
 90 2015-05-02 18:59:50 <jonasschnelli> it would be cool so s/openssl/libsecp256k
 91 2015-05-02 19:00:03 <jgarzik> jonasschnelli, I'd merge that in a heartbeat, yes :)
 92 2015-05-02 19:00:04 <jonasschnelli> s/so/to
 93 2015-05-02 19:01:12 <jgarzik> jonasschnelli, it doesn't do payment protocol, so it would be possible to remove openssl completely IIRC.  there was already some work to replace openssl with polarssl, but that is just a half-step.
 94 2015-05-02 19:02:12 <jonasschnelli> jgarzik: Right. If bip70 would be there once it could be opt-in compiled and so openssl could be opt-in also.
 95 2015-05-02 19:02:22 <jgarzik> jonasschnelli, ACK picocoin PR #21... ok to merge?
 96 2015-05-02 19:02:30 <jgarzik> fix mac compile
 97 2015-05-02 19:02:50 <jgarzik> I don't have a good way to test
 98 2015-05-02 19:03:07 <jgarzik> I tested that it doesn't break Linux
 99 2015-05-02 19:03:09 <jonasschnelli> it's from May 2013. :) let me first do some testing. Give me 30mins.
100 2015-05-02 19:03:22 <jonasschnelli> or maybe 5mins
101 2015-05-02 19:04:58 <jonasschnelli> it even has merge conflicts #21
102 2015-05-02 19:05:33 <jgarzik> minor ones
103 2015-05-02 19:05:47 <jgarzik> easy enough for me to simply re-create
104 2015-05-02 19:11:15 <jonasschnelli> jgarzik: force pushed but need to fix: "configure: error: Missing required libevent"
105 2015-05-02 19:11:26 <jonasschnelli> ah. that's probably a local issue
106 2015-05-02 19:12:34 <jonasschnelli> jgarzik: is there a rpc server or why does picocoin requires libevent and libjansson?
107 2015-05-02 19:16:23 <jgarzik> jonasschnelli, for the spv wallet in progress
108 2015-05-02 19:17:19 <jgarzik> jonasschnelli, bitcoind will likely want libevent soon, anyways.  https://github.com/bitcoin/bitcoin/pull/5677
109 2015-05-02 19:18:11 <jonasschnelli> jgarzik: yes. I hope wumpus will continue soon. :)
110 2015-05-02 19:18:49 <jonasschnelli> jgarzik: picocoin: get some compile error. I try to solve them an open a pr
111 2015-05-02 19:19:37 <jgarzik> jonasschnelli, ok!  make sure to pull master as of a few minutes ago
112 2015-05-02 19:19:45 <jgarzik> jonasschnelli, just re-tested here on Linux
113 2015-05-02 19:19:59 <jonasschnelli> yes. i saw. Just fetched and reseted to master.
114 2015-05-02 19:20:06 <jgarzik> jonasschnelli, your PR #22 looks ok, sans the trailing "{" in function definitions... ;p
115 2015-05-02 19:20:12 <jonasschnelli> autoreconf and  configure no works
116 2015-05-02 19:20:31 <jonasschnelli> yes. Let me overhaul #22. Don't even look at it now. :)
117 2015-05-02 19:21:02 <jonasschnelli> hmm... off64_t is non std
118 2015-05-02 19:21:10 <jonasschnelli> in brd.c
119 2015-05-02 19:21:59 <jgarzik> jonasschnelli, for the tests, set TEST_TESTNET3_VERF to a linearized blockchain (import format).  ditto TEST_MAINNET_VERF for mainnet.
120 2015-05-02 19:22:14 <jgarzik> jonasschnelli, the tests run a _much_ longer time, and validate the entire chain
121 2015-05-02 19:22:28 <jonasschnelli> ok. Thanks
122 2015-05-02 19:22:50 <jgarzik> jonasschnelli, standard gnu makefile: "make check" "make distcheck" etc.
123 2015-05-02 19:24:23 <jgarzik> jonasschnelli, any and all changes welcome - comment/doc additions etc.
124 2015-05-02 19:25:14 <jgarzik> jonasschnelli, the goal for libccoin is a bitcoin library for wallets & utilities & full nodes.  the goal for picocoin was an extremely minimal SPV wallet in C.
125 2015-05-02 19:25:29 <jgarzik> all portable to multiple platforms and embedded situations
126 2015-05-02 19:26:18 <jgarzik> libccoin is feature complete for a basic bitcoin library.  picocoin is much more of a WIP.
127 2015-05-02 19:26:33 <jonasschnelli> jgarzik: agreed. A tiny c based spv library would be very flexible.
128 2015-05-02 19:27:01 <jonasschnelli> does libccoin do netwoking, the whole SPV stuff?
129 2015-05-02 19:28:42 <jgarzik> jonasschnelli, no networking really (a few dns seed helper routines)  the idea was to layer those away from libccoin
130 2015-05-02 19:28:49 <jgarzik> jonasschnelli, networking is in src/
131 2015-05-02 19:29:18 <jonasschnelli> okay. will dive in soon.
132 2015-05-02 19:29:43 <jonasschnelli> first i need to solve the compile issues. off64_t and lseek64 does not compile on osx.
133 2015-05-02 19:29:45 <jgarzik> jonasschnelli, oh yeah, "brd" is "block relay daemon", a minimal wallet-free node.
134 2015-05-02 19:29:55 <jgarzik> jonasschnelli, forgot about that ;p
135 2015-05-02 19:30:03 <jonasschnelli> internally off_t and lseek is already 64bit
136 2015-05-02 19:30:09 <jgarzik> jonasschnelli, also like 30% complete ;p
137 2015-05-02 19:31:22 <jgarzik> jonasschnelli, maybe a lseek64 and off64_t wrapper is best approach then
138 2015-05-02 19:31:23 <jonasschnelli> Ah. Now i see and remember. brd is not part of libccoin or picocoin
139 2015-05-02 19:31:47 <jgarzik> jonasschnelli, I like to keep the code as free of ifdefs as possible
140 2015-05-02 19:31:48 <jonasschnelli> jgarzik: yes. Just added #ifndef's for the *64_t stuff
141 2015-05-02 19:31:59 <jgarzik> (in the main C code, I mean)
142 2015-05-02 19:37:19 <jgarzik> jonasschnelli, this guy did/found same on osx: https://github.com/jgarzik/picocoin/issues/29
143 2015-05-02 19:38:09 <jonasschnelli> jgarzik: right. Almost the same what i just added. Reminds me to first look at the issues page before starting coding. :)
144 2015-05-02 19:38:41 <joshc> hi :)
145 2015-05-02 19:51:47 <jonasschnelli> jgarzik: here we go: https://github.com/jgarzik/picocoin/pull/37
146 2015-05-02 19:52:23 <jonasschnelli> with #37 i can compile it without issues on my mac
147 2015-05-02 19:54:09 <joshc> unfortunately, still breaks linux
148 2015-05-02 19:55:04 <jonasschnelli> joshc: oh. Let me test.
149 2015-05-02 19:55:57 <joshc> needs some smarter autoconf magic, probably, to first check glibc, then check for libargp before failing
150 2015-05-02 19:56:26 <jonasschnelli> on my linux machine it even start before that with: "configure.ac:99: warning: macro `AM_PATH_GLIB_2_0' not found in library"
151 2015-05-02 19:57:26 <jgarzik> jonasschnelli, needs glib to autogen
152 2015-05-02 19:57:51 <jgarzik> jonasschnelli, welcome glib removal or embedding..
153 2015-05-02 19:58:20 <joshc> is glib removal a goal of the project?
154 2015-05-02 19:59:32 <jonasschnelli> the gnu stuff would make the library not useable within Apple app stores (iOS / Mac)
155 2015-05-02 19:59:41 <jonasschnelli> IIRC license issues
156 2015-05-02 19:59:59 <jgarzik> joshc, jonasschnelli: ideal is near zero dependencies for core libccoin, to enable embedded use
157 2015-05-02 20:01:06 <joshc> roll you're own ECDSA? ;)
158 2015-05-02 20:01:17 <jonasschnelli> i had no issue compiling on linux with PR #37 (after installing libglib2)
159 2015-05-02 20:04:08 <jgarzik> joshc, use libsecp256k1
160 2015-05-02 20:04:33 <jgarzik> libccoin seems a possible candidate for libconsensus use long term
161 2015-05-02 20:05:14 <joshc> jonasschnelli: http://vpaste.net/QUy8Q
162 2015-05-02 20:06:13 <jonasschnelli> okay... will try to check for libargb only if glib-2 is not there
163 2015-05-02 20:06:46 <joshc> it isn't clear to me the relationship between glib-2 and libargp?  argp_parse is provided by _glibc_, at least on my system
164 2015-05-02 20:08:40 <jgarzik> there is no relationship on linux
165 2015-05-02 20:14:01 <jonasschnelli> https://github.com/jgarzik/picocoin/pull/37
166 2015-05-02 20:14:01 <jonasschnelli> now it should work on linux and osx: