1 2012-10-15 02:47:50 <jgarzik> Diablo-D3: "ragequit" new phrase, I like it
2 2012-10-15 03:18:11 <Diablo-D3> jgarzik: dude, thats like a decade old phrase
3 2012-10-15 03:18:27 <Diablo-D3> and the action it describes is like a million years old
4 2012-10-15 03:20:08 <midnightmagic> ACTION wonders how you could live through multiplayer doom/quake/unreal tournament without hearing it at least once.
5 2012-10-15 03:23:02 <Diablo-D3> yeah
6 2012-10-15 03:23:12 <Diablo-D3> I think ut even has that as a built in quit reason or some shit
7 2012-10-15 04:04:50 <jgarzik> games? what is that?
8 2012-10-15 04:06:29 <Luke-Jr> now if only we could get Diablo-D3 to ragequit #bitcoin-dev ???
9 2012-10-15 04:27:40 <midnightmagic> Luke-Jr: you big meanie.
10 2012-10-15 04:28:55 <Luke-Jr> :p
11 2012-10-15 06:42:41 <Guest12345> i need some tester's for my pool
12 2012-10-15 06:42:58 <Guest12345> anyone wan't to join
13 2012-10-15 07:10:01 <t7> i dont trust you
14 2012-10-15 07:59:38 <ThomasV> while bitcoind is catching up with the blockchain, it seems to be checking mempool transactions (calling FetchInputs). isn't it a waste of time?
15 2012-10-15 09:12:29 <jgarzik> ThomasV: sometimes
16 2012-10-15 09:58:04 <ThomasV> jgarzik: I'm asking because it's been extremely slow
17 2012-10-15 09:58:22 <ThomasV> and the log is full of that
18 2012-10-15 09:59:52 <sipa> ThomasV: if you're interested, you can try my "ultraprune" branch; it should be much faster
19 2012-10-15 10:00:24 <ThomasV> sipa: it prunes spent coins?
20 2012-10-15 10:00:31 <sipa> yes and no
21 2012-10-15 10:00:45 <ThomasV> I don't want that, it's for my electrum server
22 2012-10-15 10:00:58 <sipa> what does it need?
23 2012-10-15 10:01:18 <ThomasV> it needs all the blockchain
24 2012-10-15 10:01:24 <sipa> it keeps the entire blockchain
25 2012-10-15 10:01:42 <ThomasV> so, in which way does it differ?
26 2012-10-15 10:01:45 <sipa> and it is fully RPC compatible with 0.7, except for getrawtransaction
27 2012-10-15 10:02:12 <sipa> it keeps an explicit database with just the unspent transaction outputs, in addition to the block chain
28 2012-10-15 10:02:22 <sipa> (it doesn't have blkindex.dat anymore though)
29 2012-10-15 10:02:41 <ThomasV> what is in blkindex.dat?
30 2012-10-15 10:02:47 <sipa> the txid -> diskpos map
31 2012-10-15 10:02:53 <sipa> and where each coin is spent
32 2012-10-15 10:03:33 <ThomasV> sipa: does it have identical blk*.dat files ?
33 2012-10-15 10:04:00 <sipa> yes
34 2012-10-15 10:04:18 <sipa> well, same format; they're named differently, in a subdirectory, and smaller
35 2012-10-15 10:04:27 <sipa> but the encoding format is exactly the same
36 2012-10-15 10:04:43 <ThomasV> oh, that means I'll need to mess with abe
37 2012-10-15 10:05:12 <Luke-Jr> ACTION wonders if xanatos is seriously making an issue out of the requirement to include a copy of the GPL text with the program, sigh
38 2012-10-15 10:05:34 <sipa> Luke-Jr: i'm really not convinced he's wrong
39 2012-10-15 10:05:41 <sipa> (though i certainly don't know for sure)
40 2012-10-15 10:06:01 <Luke-Jr> no, we do technically need to include a copy of the license itself, but it feels a bit pedantic in this case
41 2012-10-15 10:06:08 <sipa> apart from that
42 2012-10-15 10:07:05 <sipa> the GPL requires that anyone who receives the binary has the same rights as the author; if someone takes the source now, modifies it under the MIT license, and produces a binary from that, and distributes it without source, he is not violating the MIT license, but he is violating the GPL license of the icon (in the assumption that using the icon makes bitcoin a derived work)
43 2012-10-15 10:07:23 <Luke-Jr> sipa: yes, that is true
44 2012-10-15 10:07:25 <sipa> and our MIT license does not guarantee that someone can't do that
45 2012-10-15 10:07:41 <Luke-Jr> the GPL doesn't require other code to guarantee that.
46 2012-10-15 10:07:55 <Luke-Jr> it just requires actual compliance with that as long as the GPL'd code is there.
47 2012-10-15 10:08:16 <sipa> well, that is where it gets fuzzy
48 2012-10-15 10:08:31 <sipa> but i'd rather not leaving it up to lawyers to decide
49 2012-10-15 10:09:14 <ThomasV> I don't know what the limiting factor is for my bitcoind catching up with the blockchain, but it does not seem to be cpu: I am at 0% usage most of the time, and sometimes it jumps to 30% for a little while (this coincides with blocks being added according to the log)
50 2012-10-15 10:09:31 <ThomasV> but still, most of the time isz spent at 0% usage
51 2012-10-15 10:09:40 <ThomasV> or 1%
52 2012-10-15 10:10:05 <ThomasV> so it could be disk access, or download speed
53 2012-10-15 10:10:06 <sipa> ThomasV: the limiting factor is BDB\\
54 2012-10-15 10:10:08 <Luke-Jr> sipa: under the GPL, the copyright holder would be obliged to warn us and give us time to comply by releasing source before any legal action could be taken
55 2012-10-15 10:10:19 <ThomasV> sipa: care to explain?
56 2012-10-15 10:10:32 <Luke-Jr> worst case scenario, we rush an emergency "some idiot complained about a grey area on our icons" release
57 2012-10-15 10:10:32 <sipa> ThomasV: BDB is not really efficient
58 2012-10-15 10:10:53 <sipa> ThomasV: it continuously needs to rewrite large parts of the database, and does tons of synchronous writes
59 2012-10-15 10:10:57 <Luke-Jr> their*
60 2012-10-15 10:11:09 <ThomasV> sipa: why is it used then?
61 2012-10-15 10:11:15 <sipa> ThomasV: because satoshi chose it
62 2012-10-15 10:11:21 <sipa> ThomasV: in 0.8 we'll switch to leveldb
63 2012-10-15 10:11:39 <ThomasV> is there a branch where leveldb is used?
64 2012-10-15 10:12:09 <sipa> yes, ultraprune :)
65 2012-10-15 10:12:26 <ThomasV> :)
66 2012-10-15 10:12:45 <ThomasV> sipa: bu that means your blk*.dat will differ
67 2012-10-15 10:12:51 <sipa> ThomasV: *NO*
68 2012-10-15 10:12:56 <ThomasV> why not?
69 2012-10-15 10:13:10 <sipa> the blk* files are just a concatenation of blocks
70 2012-10-15 10:13:19 <sipa> raw, directly to disk, no index or database
71 2012-10-15 10:13:26 <ThomasV> oh ok
72 2012-10-15 10:13:34 <ThomasV> so it is database agnostic
73 2012-10-15 10:13:38 <sipa> (except for blkindex.dat, which is a BDB database file like wallet.dat and formerly addr.dat)
74 2012-10-15 10:14:53 <sipa> ThomasV: on my VPS, sync from scratch used to take over a day (it has very high I/O latency)
75 2012-10-15 10:15:01 <sipa> with leveldb/ultraprune it's more like 1-2 hours
76 2012-10-15 10:16:27 <ThomasV> sipa: nice. but why is leveldb tied to ultraprune? these seem to be two distinct developments
77 2012-10-15 10:16:49 <sipa> ThomasV: leveldb was implemented separatedly by mike
78 2012-10-15 10:17:09 <sipa> ThomasV: with a compatibility layer to make it behave like BDB
79 2012-10-15 10:17:39 <sipa> as i rewrote the validation engine anyway, i did so with leveldb's model in mind, so after rebasing it, no compatibility layer was needed anymore
80 2012-10-15 10:18:18 <jgarzik> sipa: maybe it's confusing that "blk*dat" refers to both BDB index and blockchain flat files
81 2012-10-15 10:19:16 <sipa> jgarzik: yeah, in ultraprune it's blocks/*, blktree/* and coins/*
82 2012-10-15 10:19:30 <jgarzik> ThomasV: leveldb and ultraprune (bitcoin db index change) are indeed two distinct changes, but it makes sense to merge them at the same time. Everybody has to upgrade... no sense in putting users through a BDB->leveldb upgrade, and _then_ another upgrade to radically change the bitcoin db index formats.
83 2012-10-15 10:19:48 <gmaxwell> Luke-Jr: speaking of licensing, we're also not conformant with the LGPL. We ship LGPLed things, so we must include a copy of the license. I'm also unclear on what we need to do wrt the static linking. Closed source apps static linkers technically have to ship a relinkable set of object files so the lgpl parts can be replaced and are also obligated to offer the source for the LGPL things we link. But since we're not a closed app ourselves I'm
84 2012-10-15 10:20:06 <sipa> gmaxwell: ourselves I'm [...]
85 2012-10-15 10:20:10 <jgarzik> ThomasV: BDB is a big part of currently-slow network sync'ing
86 2012-10-15 10:20:19 <gmaxwell> app ourselves I'm unsure of how much we should be doing there.
87 2012-10-15 10:20:35 <Luke-Jr> gmaxwell: yeah, I think we can just leave that problem for whoever wants to close it
88 2012-10-15 10:20:38 <jgarzik> On licenses: License conformance is the important bit
89 2012-10-15 10:20:44 <jgarzik> shipping the text is optional but useful
90 2012-10-15 10:20:59 <gmaxwell> jgarzik: well BDB made much less of a difference after ultraprune, but the both carry common bug text.
91 2012-10-15 10:21:04 <jgarzik> Most source code include "if you don't have a copy, go to this $url" language
92 2012-10-15 10:21:44 <jgarzik> gmaxwell: common bug text?
93 2012-10-15 10:21:55 <gmaxwell> jgarzik: indeed, I was just following along with the pedantry. In general we ourselves aren't likely to non-conform in a material way as what _we_ do would even work under the GPL.
94 2012-10-15 10:22:02 <gmaxwell> jgarzik: s/text/risk/ sorry, just woke up
95 2012-10-15 10:22:05 <gmaxwell> :P
96 2012-10-15 10:22:15 <sturles> In my experience slow nodes is the major factor holding back network sync. 2 GB going out through a home ADSL line isn't fast.
97 2012-10-15 10:23:02 <Luke-Jr> jgarzik: shipping the text is not optional for GPL/LGPL, technically
98 2012-10-15 10:23:16 <gmaxwell> sturles: it wasn't previously.
99 2012-10-15 10:23:23 <sipa> gmaxwell: on VPS's the difference between BDB and LevelDB (both on ultraprune) is massive
100 2012-10-15 10:23:40 <sipa> like an order of magnitude
101 2012-10-15 10:23:56 <ThomasV> sipa: ok, let me try your branch
102 2012-10-15 10:24:11 <jgarzik> LevelDB simply does not update as many files/blocks as BDB does, for the same amount of block processing work
103 2012-10-15 10:24:14 <gmaxwell> But if we're serious about the MIT license being actually MITish, we need to think about making it easy for downstream reusers to be license kosher when they do closed sourcy things we don't.
104 2012-10-15 10:24:19 <jgarzik> lower pagecache usage
105 2012-10-15 10:24:26 <jgarzik> + less disk thrashing on lower resource nodes
106 2012-10-15 10:24:31 <sipa> jgarzik: exactly
107 2012-10-15 10:24:52 <jgarzik> gmaxwell: agreed
108 2012-10-15 10:25:08 <jgarzik> license pedantry is pedantry... but it is important to get the legal stuff right
109 2012-10-15 10:25:22 <gmaxwell> sipa: thats interesting, I just remembered my own??? on ssd tests??? and some numbers you had (I guess on your laptop)
110 2012-10-15 10:26:03 <sipa> gmaxwell: even on ultraprune, iirc syncing the full chain took a day on my VPS
111 2012-10-15 10:26:43 <gmaxwell> jgarzik: yea, agreed. I only point out that it's pedantry to emphasize that it's not urgent, we can slowly grind in the right direction. (wumpus was stressing a bit about some of the pull requests arguments the other day)
112 2012-10-15 10:27:59 <jgarzik> my sanity level is high today, because I have been deleting all those github posts unread ;-)
113 2012-10-15 10:46:57 <jgarzik> my life would move much more slowly, without shell history
114 2012-10-15 10:47:32 <lianj> and zsh
115 2012-10-15 10:47:51 <Luke-Jr> jgarzik: I find it handy to have per-directory shell histories (with an eternal global-history sorted before it)
116 2012-10-15 10:51:41 <ThomasV> sipa: I am running your branch now. it's not only faster, it also seems to use much less memory
117 2012-10-15 10:56:38 <ThomasV> yes, that's an order of magnitude difference
118 2012-10-15 10:57:27 <sipa> glad to hear; let me know if you find any issues
119 2012-10-15 10:58:18 <ThomasV> it hasn't yet reached the Satoshi Dice part of the blockchain
120 2012-10-15 11:22:57 <ThomasV> sipa: but it does not have getrawmempool??
121 2012-10-15 11:24:32 <sipa> it should
122 2012-10-15 11:24:55 <sipa> here it does :)
123 2012-10-15 11:31:57 <ThomasV> oh ok
124 2012-10-15 11:34:41 <sipa> why wouldn't it?
125 2012-10-15 11:36:58 <ThomasV> no, you said above it does not have getrawtransaction, and I confused it fot getrawmempool
126 2012-10-15 11:37:07 <ThomasV> *for*
127 2012-10-15 11:38:42 <sipa> it does have getrawtransaction as well, but it's a lot slower, and only works for not-completely-spent transactions
128 2012-10-15 11:39:46 <ThomasV> will 0.8 have getrawtransaction?
129 2012-10-15 11:40:20 <sipa> i may add an optional txindex later on, to speed up getrawtransaction
130 2012-10-15 11:41:16 <jgarzik> ThomasV: 0.8 should have everything currently in sipa's ultraprune branch
131 2012-10-15 11:41:52 <jgarzik> ThomasV: 0.8 will have getrawtransaction, yes. it will not work for spent transactions.
132 2012-10-15 11:43:23 <jgarzik> we should add "require c++0x" to the todo list
133 2012-10-15 11:43:37 <jgarzik> seems doable, if we can upgrade mingw compile env
134 2012-10-15 11:44:32 <sipa> moving to c++0x would be nice, indeed
135 2012-10-15 11:45:22 <Luke-Jr> would it be much overhead to store a bloom filter for the txids in every N blocks, to enable a slower getrawtx on spent ones?
136 2012-10-15 11:46:39 <jgarzik> sipa: it is disappointing that ultraprune's blkNNNN.dat flushing is a lot less efficient than current HEAD
137 2012-10-15 11:47:00 <jgarzik> sipa: current HEAD is write+flush, ultraprune does a much more expensive open+flush+close
138 2012-10-15 11:47:16 <sipa> jgarzik: hmm?
139 2012-10-15 11:47:33 <jgarzik> sipa: reading https://github.com/sipa/bitcoin/commit/b0fe5885f2adb837f10555da24e2dbcc351d2c24
140 2012-10-15 11:47:35 <sipa> head doesn't keep the block file open either, afaik
141 2012-10-15 11:48:22 <jgarzik> sipa: HEAD syscall trace looks like: write(2), fdatasync(2). ultraprune appears to write(2), close(2), open(2), fdatasync(2), close(2)
142 2012-10-15 11:48:49 <jgarzik> well write(2), fdatasync(2), close(2)
143 2012-10-15 11:48:53 <jgarzik> but the point stands
144 2012-10-15 11:49:15 <jgarzik> unless ultraprune fixes that behavior in a later commit
145 2012-10-15 11:49:45 <sipa> no, i plan to redo the block syncing thing after a global block cache
146 2012-10-15 11:50:17 <sipa> the problem is that doing it in one call requires knowing in advance whether the block is to be flushed after a certain connect, passing that down into the block tree logic, ...
147 2012-10-15 11:51:08 <sipa> but yes, minor regression right now, i agree
148 2012-10-15 11:51:41 <sipa> ultraprune should be slightly safer than HEAD, though
149 2012-10-15 11:51:47 <sipa> in terms of disk consistency
150 2012-10-15 11:52:18 <jgarzik> sipa: how so?
151 2012-10-15 11:52:43 <jgarzik> sipa: HEAD flushes after writing block, then does an index update wrapped in a BDB transaction
152 2012-10-15 11:53:04 <sipa> not during IBD
153 2012-10-15 11:53:12 <jgarzik> true
154 2012-10-15 11:53:18 <sipa> ultraprune flushes blocks right before any DB write
155 2012-10-15 11:53:26 <sipa> but does so less frequently in IBD
156 2012-10-15 11:54:02 <jgarzik> sipa: what is leveldb flush behavior during IBD? after every index update (every block)?
157 2012-10-15 11:54:26 <jgarzik> do you batch multiple blocks into one leveldb batch, during IBD?
158 2012-10-15 11:54:38 <sipa> jgarzik: ultraprune has a globally cached pcoinsTip object, which maintains the coin status
159 2012-10-15 11:54:55 <sipa> during IBD that objects is flushed to disk only after 5000 modified transactions
160 2012-10-15 11:55:09 <sipa> and there are no DB writes apart from that
161 2012-10-15 11:55:29 <sipa> and before that write, it flushes blocks and blockdb
162 2012-10-15 11:58:16 <sipa> there is still a weakness though, as updates to the block index are not cached but written immediately (though a sync is issued right before writing coin data), so there is a chance that the block index update makes it to disk, but the blocks themselves don't
163 2012-10-15 11:58:30 <sipa> (during IBD)
164 2012-10-15 12:06:39 <jgarzik> sipa: cool
165 2012-10-15 12:07:06 <jgarzik> sipa: not so worried about IBD anyway; just wondering. We can blow away the db and restart, and things are still efficient.
166 2012-10-15 13:22:47 <jgarzik> http://bitcoinmagazine.net/interview-with-glbses-nefario/
167 2012-10-15 13:24:02 <D34TH> jgarzik, i totally just got an invite to lemon wallet's dev challenge we should get bitcoin on it
168 2012-10-15 13:37:47 <gmaxwell> jgarzik: nice bus tracks all over thymos' back there.
169 2012-10-15 13:38:39 <sipa> ?
170 2012-10-15 13:40:24 <gmaxwell> sipa: the article he linked to, nefario spends considerable length accusing Theymos of undermining GLBSE and saying that he is hold customer deposits in excess of what he's claimed he's holding.
171 2012-10-15 13:46:22 <jgarzik> yah
172 2012-10-15 13:46:55 <jgarzik> theymos v. Nefario gossip, plus perhaps UK FSA opinion of bitcoin changing
173 2012-10-15 13:47:18 <jgarzik> the "all bitcoin use requires AML" is sad but not unexpected
174 2012-10-15 13:49:14 <helo> so if you want to sell something for bitcoin, you have to get verifiable id?
175 2012-10-15 13:49:25 <helo> (from the customer obv)
176 2012-10-15 13:51:09 <gmaxwell> helo: sounds really unlikely??? as that doesn't exist for cash; I'd expect all bank like usage to end up like that.
177 2012-10-15 13:59:19 <helo> ACTION learns what "currency changing" means... ah, good
178 2012-10-15 14:00:54 <jgarzik> This is in the context of GLBSE-like assets. It seems unlikely that a merchant selling regular goods for bitcoins would have any AML requirements over and above what they already have for fiat cash.
179 2012-10-15 14:05:11 <devrandom> /win 21
180 2012-10-15 14:05:16 <devrandom> argh
181 2012-10-15 14:06:17 <Luke-Jr> /lose 22
182 2012-10-15 14:15:09 <helo> sipa: should i be alarmed to see "failed to connect best block" with ultraprune?
183 2012-10-15 14:21:31 <t7> asdasasdasddsa
184 2012-10-15 14:22:45 <sipa> helo: does it still work?
185 2012-10-15 14:24:23 <helo> nope, aborts
186 2012-10-15 14:31:48 <sipa> helo: what errors do you see in debug.log?
187 2012-10-15 14:33:12 <helo> REORGANIZE: Disconnect 1 blocks; 000000000000048d4dc3..00000000000000d94da9
188 2012-10-15 14:33:24 <helo> REORGANIZE: Connect 2 blocks; 000000000000048d4dc3..0000000000000176ec0e
189 2012-10-15 14:33:28 <helo> ERROR: DisconnectBlock() : added transaction mismatch? database corrupted
190 2012-10-15 14:33:36 <helo> ERROR: SetBestBlock() : DisconnectBlock 00000000000000d94da9 failed
191 2012-10-15 14:34:24 <helo> i likely pulled the rug out from under it when i rebooted just prior to this
192 2012-10-15 14:35:41 <sipa> helo: did you upgrade the code?
193 2012-10-15 14:35:57 <helo> nope, at 41f98b3c506de5a81e143fc7538aa85d325e9c66
194 2012-10-15 14:37:03 <helo> hmm... that's older than i thought
195 2012-10-15 14:37:32 <sipa> try latest
196 2012-10-15 14:40:11 <helo> i thought i was on latest... git clone https://github.com/sipa/bitcoin.git && cd bitcoin && git checkout -b ultraprune -t origin/ultraprune
197 2012-10-15 14:46:49 <sipa> helo: does git head fix the problem?
198 2012-10-15 14:46:58 <helo> building
199 2012-10-15 14:49:27 <helo> src/sethash.{cpp,h} missing...
200 2012-10-15 14:50:23 <sipa> helo: means i screwed up; those files shouldn't be referenced
201 2012-10-15 14:51:32 <helo> k, removed from the .pro
202 2012-10-15 14:54:20 <helo> sipa: same error
203 2012-10-15 14:54:53 <sipa> helo: the old code may have screwed up
204 2012-10-15 15:01:02 <helo> ACTION reprimands the boogey man
205 2012-10-15 15:02:07 <helo> sipa: ultraprune currenlty does no pruning by default, correct?
206 2012-10-15 15:04:02 <sipa> helo: none at all
207 2012-10-15 15:04:28 <helo> it's just using leveldb to organize things better for pruning?
208 2012-10-15 15:04:40 <helo> rather, "using leveldb and organizing..."
209 2012-10-15 15:11:19 <jgarzik> helo: Berkeley DB is used to index the blockchain. That software is replaced with leveldb, which is used to index the blockchain.
210 2012-10-15 15:15:34 <MC1984> ultraprune tunred out to be a massive misnomer
211 2012-10-15 15:16:50 <sipa> MC1984: it maintains an ultra-pruned version of the chain in addition to the blockchain
212 2012-10-15 15:17:04 <sipa> but yes, it's confusing
213 2012-10-15 15:17:08 <sipa> maybe i should still rename it
214 2012-10-15 15:17:30 <MC1984> ultrachain or something
215 2012-10-15 15:22:40 <alexmat> anyone interested in buying a 12month access to any porn site of your choosing!
216 2012-10-15 15:23:21 <kjj_> ooh. which op is going to ban the spammer first?
217 2012-10-15 15:23:31 <MC1984> >paying for porn
218 2012-10-15 15:23:36 <copumpkin> seems nicer to just ask him not to spam
219 2012-10-15 15:23:42 <copumpkin> rather than to ban immediately
220 2012-10-15 15:24:26 <kjj_> yes, because if there is anything that having email for the last 15 years has taught me, it is that the problem is lack of education.
221 2012-10-15 15:28:11 <jgarzik> sipa: just rename the branch to "zooooom"
222 2012-10-15 15:29:15 <kjj_> lol. he PMd me, called me a shithead and told me to mind my own business
223 2012-10-15 15:30:07 <alexmat> deserve it!
224 2012-10-15 15:30:59 <jgarzik> alexmat: behave, and do not spam this channel
225 2012-10-15 15:37:05 <jdnavarro> is there a maximum size for a transaction?
226 2012-10-15 15:39:12 <maaku> jdnavarro: 1MB
227 2012-10-15 15:39:22 <maaku> but in practice much lower
228 2012-10-15 15:40:09 <jdnavarro> maaku: alright
229 2012-10-15 15:40:26 <jdnavarro> I'll update the protocol specification in the wiki
230 2012-10-15 15:40:31 <maaku> jdnavarro: no
231 2012-10-15 15:40:43 <maaku> it's not part of the protocol
232 2012-10-15 15:40:54 <jdnavarro> hmm
233 2012-10-15 15:40:54 <maaku> it just that the block limit is 1mb
234 2012-10-15 15:41:07 <maaku> so a larger transaction could never make it into a block
235 2012-10-15 15:41:27 <jdnavarro> but you can still send it around
236 2012-10-15 15:41:57 <maaku> maybe, maybe not. but does that matter?
237 2012-10-15 15:42:00 <jdnavarro> unless there is a limit in the size of tx message
238 2012-10-15 15:42:32 <maaku> that's an implementation detail, not a protocol requirement
239 2012-10-15 15:44:26 <jdnavarro> maaku: but that would be an obvious DoS attack
240 2012-10-15 15:44:56 <jdnavarro> just sending huge transactions like crazy would overload many nodes, wouldn't it?
241 2012-10-15 15:46:14 <jgarzik> jdnavarro: large orphan transactions are not remembered, for similar reasons
242 2012-10-15 15:46:40 <jgarzik> normally bitcoin remembers orphans (though not immediately spendable, due to lack of dependent transactions) in case the parents appear later
243 2012-10-15 15:47:39 <maaku> jdnavarro: so? that doesn't mean it should be on the protocol page. the bitcoin protocol page should cover the protocol, and not be cluttered up with DoS protection trivialities
244 2012-10-15 15:48:36 <jdnavarro> maaku: well, there are other messages with size limits
245 2012-10-15 15:49:27 <jdnavarro> jgarzik: you mean by remember that it receives it, deletes them and doesn't relay them?
246 2012-10-15 15:49:51 <jgarzik> jdnavarro: correct, for the first two
247 2012-10-15 15:49:57 <jgarzik> I don't _think_ it relays, but I could be wrong
248 2012-10-15 15:50:57 <jdnavarro> ACTION is checking the source code of the official client
249 2012-10-15 16:00:01 <sipa> transactions must not be over 1000000 bytes
250 2012-10-15 16:00:06 <sipa> that is an actual network rule
251 2012-10-15 16:00:43 <sipa> (although it is redundant, as such a transaction cannot be in a block without that block going over its size limit)
252 2012-10-15 16:03:13 <jdnavarro> I just found there is a limit for the message size: https://github.com/bitcoin/bitcoin/blob/master/src/main.cpp#L3090
253 2012-10-15 16:03:31 <jdnavarro> sipa: it's that the limit or there is a specific one for transactions?
254 2012-10-15 16:11:56 <phantomcircuit> sipa, the rule is actually that no network message can be > MAX_SIZE so it's slightly smaller than the usually quoted number
255 2012-10-15 16:12:25 <jdnavarro> I just realized that the limit for the transaction is the same as the limit for the block
256 2012-10-15 16:12:56 <sipa> phantomcircuit: MAX_SIZE is 32 MiB, if my hex-foo is good
257 2012-10-15 16:13:14 <phantomcircuit> is it?
258 2012-10-15 16:13:19 <jdnavarro> sipa: I can confirm it
259 2012-10-15 16:13:20 <phantomcircuit> i seem to remember it being 10MiB
260 2012-10-15 16:13:23 <phantomcircuit> hmm
261 2012-10-15 16:13:23 <sipa> serialize.h:static const unsigned int MAX_SIZE = 0x02000000;
262 2012-10-15 16:13:25 <phantomcircuit> guess not
263 2012-10-15 16:13:33 <sipa> MAX_BLOCK_SIZE is 1000000, though
264 2012-10-15 16:14:39 <sipa> jdnavarro: anyway, both CheckBlock and CheckTransaction (basic context-independent validity checks) require the respective block/tx being at most MAX_BLOCK_SIZE
265 2012-10-15 16:15:15 <jdnavarro> sipa: right
266 2012-10-15 16:15:41 <jdnavarro> I'll go with it
267 2012-10-15 16:15:44 <jdnavarro> thanks guys
268 2012-10-15 16:30:47 <sipa> helo: did you restart from scratch? let me know if the same thing happens still with newer code
269 2012-10-15 16:49:54 <helo> sipa: yeah, i restarted from the beginning... i doubt i can exit at the same state it was at, but after ibd i'll reload and let you know
270 2012-10-15 18:16:37 <BCB> any idea why a transaction from my 0.7 bitcoind would not hit the blockchain
271 2012-10-15 18:18:46 <c_k> BCB: give it time
272 2012-10-15 18:18:51 <c_k> BCB: how long has it been?
273 2012-10-15 18:18:57 <BCB> 10 min
274 2012-10-15 18:19:07 <jgarzik> BCB: what do you mean "hit the blockchain"?
275 2012-10-15 18:19:11 <BCB> I usually have 1 confirm by now
276 2012-10-15 18:19:35 <jgarzik> BCB: 10 minutes is only an average. it could take an hour if you're unlucky.
277 2012-10-15 18:20:19 <BCB> jgarzik: to even show up in the blockchain??
278 2012-10-15 18:20:22 <BCB> I've never seen that
279 2012-10-15 18:20:44 <BCB> sent at 2012-15-10 15:43:52
280 2012-10-15 18:20:49 <kjj_> yes, that's what the blockchain is. or do you mean in the memory pool of the node that runs blockchain.info's website?
281 2012-10-15 18:21:10 <BCB> "48b0d02003a926d6db7fceb39b3acf5b43e9f65dbc03718883c05626dbfb75b6"
282 2012-10-15 18:21:30 <sipa> ;;bc,tslb
283 2012-10-15 18:21:31 <gribble> Time since last block: 2 minutes and 21 seconds
284 2012-10-15 18:21:48 <BCB> I have a transaction on my bitcoind
285 2012-10-15 18:21:55 <BCB> funds are gone
286 2012-10-15 18:22:06 <BCB> but when i check blockchain.info nothing
287 2012-10-15 18:22:17 <BCB> i've just not seen that before
288 2012-10-15 18:22:36 <sipa> wait half an hour, there can have been a network hiccup when the transaction was submitted
289 2012-10-15 18:22:43 <sipa> your client will eventually retransmit
290 2012-10-15 18:23:16 <BCB> sipa:will do
291 2012-10-15 18:23:36 <BCB> sipa: can I cancel that tranasction
292 2012-10-15 18:23:43 <jgarzik> BCB: there is a good place to watch: http://bitcoincharts.com/bitcoin/txlist/
293 2012-10-15 18:24:21 <sipa> BCB: no, because there is no knowing of who might have seen the transaction
294 2012-10-15 18:24:28 <sipa> they will keep considering it valid anyway
295 2012-10-15 18:25:36 <BCB> sipa:jarzik:thx
296 2012-10-15 18:27:07 <kjj_> I wish he would purge those bogus transactions, or at least move them to a different page
297 2012-10-15 18:29:16 <BCB> where does that data come from on bitcoincharts and is it available from my bitcoind
298 2012-10-15 18:30:17 <sipa> yes
299 2012-10-15 18:30:35 <sipa> but since you sent the transaction, you yourself will certainly have it
300 2012-10-15 18:30:38 <helo> jgarzik: why gevent with pynode?
301 2012-10-15 18:30:44 <sipa> the question is whether it propages through the network
302 2012-10-15 18:31:11 <helo> jgarzik: that is, why did you choose gevent for pynode
303 2012-10-15 18:31:39 <jgarzik> helo: it's better than asyncore, and somebody else did the work ;p I actually dislike the model gevent exposes quite a bit: these soft threads are nothing like what goes on "under the hood", leading to odd code occasionally.
304 2012-10-15 18:32:03 <Luke-Jr> jgarzik: so why not networkserver? :p
305 2012-10-15 18:32:28 <jgarzik> Luke-Jr: never heard of it
306 2012-10-15 18:32:48 <Luke-Jr> http://gitorious.org/bitcoin/eloipool/blobs/master/networkserver.py
307 2012-10-15 18:33:08 <jgarzik> Luke-Jr: I am tempted to use python 3.x for pynode though: 3.3 includes os.sendfile. pynode's current flat-file blockchain format precisely matches a network "block" message, to prepare for this.
308 2012-10-15 18:33:37 <Luke-Jr> no sendfile before 3.3? O.o
309 2012-10-15 18:33:50 <jgarzik> nope AFAICT
310 2012-10-15 18:33:55 <jgarzik> *not
311 2012-10-15 18:34:11 <Diablo-D3> heh
312 2012-10-15 18:34:15 <Diablo-D3> silly python
313 2012-10-15 18:34:36 <Diablo-D3> HAI LETS MAKE TWO VERSIONS OF PYTHON THAT BOTH GET ACTIVE DEVELOPMENT, ARE INCOMPATIBLE, AND INCLUDE FEATURES THE OTHER DOESNT HAVE THAT DEVELOPERS WANT
314 2012-10-15 18:34:38 <Diablo-D3> HURRRRRRRRRRR
315 2012-10-15 18:34:48 <Luke-Jr> looks like it was a non-included module or such: http://pypi.python.org/pypi/py-sendfile/
316 2012-10-15 18:34:51 <sipa> Diablo-D3: just consider them to be two different languages
317 2012-10-15 18:35:03 <Luke-Jr> Diablo-D3: Python 2 has no active development afaik
318 2012-10-15 18:35:18 <jgarzik> just 90% of the active users :)
319 2012-10-15 18:35:40 <Luke-Jr> dunno, I think I can uninstall Python 2 soon
320 2012-10-15 18:39:54 <Diablo-D3> sipa: I do
321 2012-10-15 18:39:58 <Diablo-D3> but its still annoying as fuck
322 2012-10-15 18:40:03 <Diablo-D3> its one of the major reasons why I wont use python
323 2012-10-15 18:43:46 <jgarzik> heh, transaction fees as messaging: https://bitcointalk.org/index.php?topic=117630.msg1274260#msg1274260
324 2012-10-15 19:02:52 <MC1984> whos gonna make a metro ui for bitcoin for windows 8 out next month?
325 2012-10-15 19:03:03 <MC1984> you guys gonna draw straws or what?
326 2012-10-15 19:04:03 <MC1984> http://www.youtube.com/watch?v=QZ1oPDtdhFo
327 2012-10-15 19:06:23 <sipa> MC1984: there are very few windows devs active
328 2012-10-15 19:10:45 <EPiSKiNG> why does my latest version of Bitcoin-QT require .0005 TX Fee for nearly every transaction?
329 2012-10-15 19:11:28 <copumpkin> probably depends on how fragmented your coins are?
330 2012-10-15 19:11:33 <sipa> probably because you have very young/small coins that would otherwise be considered spam by most nodes on the network
331 2012-10-15 19:13:20 <BlueMatt> TD: to work around https://github.com/netty/netty/issues/535 (triggered when bitcoinj sees high connection count sometimes), can we upgrade netty?
332 2012-10-15 19:13:24 <gmaxwell> EPiSKiNG: none of that code has changed in a very long time, so 'latest version' shouldn't be relevant.
333 2012-10-15 19:14:04 <EPiSKiNG> hmm
334 2012-10-15 19:14:13 <sturles> EPiSKiNG: Wait until you have at least 6 confirmations before sending again, and use sendmany when possible.
335 2012-10-15 19:14:22 <EPiSKiNG> ok
336 2012-10-15 19:14:36 <gmaxwell> and aviod getting paid lots of little payments when you can easily avoid it.
337 2012-10-15 19:14:51 <gmaxwell> (e.g. don't do daily payouts from mining pools)
338 2012-10-15 19:15:42 <gmaxwell> Also upgading your wallet format to the latest version and flush the keypool so all your addresses are for compressed keys (unless your wallet is already fairly newly created)
339 2012-10-15 19:15:43 <sipa> the relay minimum was changed to 0.0005 on june 5th 2011 :)
340 2012-10-15 19:22:07 <TD> BlueMatt: what version do we need?
341 2012-10-15 19:22:28 <BlueMatt> apparently 3.5.6...
342 2012-10-15 19:27:58 <gavinandresen> sipa jgarzik gmaxwell BlueMatt : I'm preparing for the foundation meeting later this week, and trying to come up with a hardware/software wish list for the dev team.
343 2012-10-15 19:28:34 <gavinandresen> So, what do we wish for?
344 2012-10-15 19:28:51 <Luke-Jr> ???
345 2012-10-15 19:29:18 <sipa> some infrastructure to run BlueMatt's pull tester on?
346 2012-10-15 19:29:22 <gavinandresen> Also, anybody know if Apple or Microsoft have inexpensive developer subscriptions for open source projects?
347 2012-10-15 19:29:40 <jgarzik> gavinandresen: DNS seed
348 2012-10-15 19:29:49 <gavinandresen> sipa: yes, I put two beefy dedicated servers on the wish list.
349 2012-10-15 19:29:50 <jgarzik> gavinandresen: some full nodes with SSDs, good relaying ones
350 2012-10-15 19:30:24 <Luke-Jr> if the Foundation is going to run a DNS seed, it would make sense to give it a name and add it in for 0.7.1 IMO
351 2012-10-15 19:30:45 <gavinandresen> too late for 0.7.1, I think.
352 2012-10-15 19:31:07 <Luke-Jr> oh well, gives whoever time to get it running I guess
353 2012-10-15 19:31:17 <gavinandresen> jgarzik: how much does a full node with SSD cost per month?
354 2012-10-15 19:31:29 <jgarzik> ACTION wonders about BF sponsoring an http bootstrap.dat download from Amazon S3
355 2012-10-15 19:31:40 <jgarzik> gavinandresen: dunno... and family dinner time. bbi ~1hr
356 2012-10-15 19:31:52 <gavinandresen> jgarzik: cool. I should get dinner going, too....
357 2012-10-15 19:32:16 <sipa> i'm not sure whether I'd rather have disk+more RAM or SDD+less RAM
358 2012-10-15 19:32:22 <Luke-Jr> gavinandresen: if you get a dedi with "unmetered" bandwidth, it should be a fixed cost
359 2012-10-15 19:32:46 <sipa> ACTION is afraid of the crowd's reaction to a seed.bitcoinfoundation.org in the source code...
360 2012-10-15 19:32:55 <Luke-Jr> worst case it can't handle the load and upgrading it can be visited in a future month
361 2012-10-15 19:33:23 <Luke-Jr> (in that case, it would probably still be beneficial to the network even if bogged down a bit)
362 2012-10-15 19:34:46 <gavinandresen> sipa: don't be frightened... and I vote for seed.btc.org
363 2012-10-15 19:35:00 <gavinandresen> (foundation owns both bitcoinfoundation.org and btc.org domains)
364 2012-10-15 19:35:11 <sipa> ha :)
365 2012-10-15 19:36:32 <Luke-Jr> is that "crowd" any bigger than "mostly just Atlas"? <.<
366 2012-10-15 19:38:25 <TD> BlueMatt: ok. i'll try and remember. if you file a bug i'll do it later this week
367 2012-10-15 19:38:43 <TD> gavinandresen: code signing certs
368 2012-10-15 19:38:58 <rdponticelli> Well, Atlas alone is not less than 5 guys :)
369 2012-10-15 19:39:02 <TD> haha
370 2012-10-15 19:39:06 <gmaxwell> I was wondering if perhaps we shouldn't start calling seeds 'bootstraps' .. but then realized thats what jeff called the blockchain preloads.
371 2012-10-15 19:39:17 <gavinandresen> TD: yes... just need to join the Microsoft/Apple dev programs to get code signing certs, I believe
372 2012-10-15 19:39:22 <sipa> entry nodes?
373 2012-10-15 19:39:37 <TD> for microsoft i believe it's something like verisign. apples costs $99 a year, or something like that
374 2012-10-15 19:39:38 <jgarzik> Walmart Greeters
375 2012-10-15 19:39:40 <TD> just involves some paperowrk
376 2012-10-15 19:40:07 <gavinandresen> TD: yes, I already asked Patrick to start the "get a DUNS number" that is the beginning of the process
377 2012-10-15 19:40:31 <TD> DUNS number?
378 2012-10-15 19:41:03 <TD> gavinandresen: more secure hosting in general. ideally a dedicated box in a cage in a datacenter near where you live
379 2012-10-15 19:41:15 <TD> no VPS. no remote root access.
380 2012-10-15 19:41:21 <TD> for hosting file downloads and the website
381 2012-10-15 19:41:47 <gavinandresen> TD: near where I live? I don't want to get the pager beep to drive to the datacenter and replace that failed hard drive....
382 2012-10-15 19:41:49 <jgarzik> ACTION whips out his phatty DUNS number, SAMS number, _and_ US Patent Office Customer Number
383 2012-10-15 19:42:34 <TD> or near where mark lives :)
384 2012-10-15 19:42:36 <Luke-Jr> gavinandresen: set it up to automatically update from git/uploads when the commit is signed by 3+ people?
385 2012-10-15 19:42:49 <TD> just generally, we know that free virtual hosting seems to be a vulnerability
386 2012-10-15 19:43:04 <TD> so it'd suck if the site hosting bitcoin downloads got hacked and the binaries replaced. it'd take a while for people to notice that
387 2012-10-15 19:43:05 <Luke-Jr> also, HTTPS would be good if we're going to secure bitcoin.org
388 2012-10-15 19:43:07 <jgarzik> gavinandresen: see this possible lock bug report? https://bitcointalk.org/index.php?topic=117874.msg1273313#msg1273313
389 2012-10-15 19:43:07 <sipa> linode is very trustworrthy, indeed
390 2012-10-15 19:43:10 <TD> ditto for defacement of the website
391 2012-10-15 19:45:34 <gavinandresen> I'm not sure I agree; I think I'd rather have an early warning system that checked downloads once an hour and reacted automatically (e.g. emailed a few of us and automatically shutdown downloads until problem was resolved)
392 2012-10-15 19:46:26 <TD> well that would be good too, but even if caught, a hacked server would be confidence shaking
393 2012-10-15 19:47:52 <gmaxwell> gavinandresen: that isn't exclusive.
394 2012-10-15 19:48:06 <gavinandresen> I agree, but I don't have faith that even if the Foundation spent bunches of money it's website would be secure.
395 2012-10-15 19:49:20 <gavinandresen> ... would be 100% secure, I should say.
396 2012-10-15 19:49:58 <sipa> if you want 100% security, make sure it's disconnected from the network and powered off
397 2012-10-15 19:50:06 <gavinandresen> exactly
398 2012-10-15 19:50:22 <sipa> doesn't mean certain steps can make it significantly harder to attack
399 2012-10-15 19:50:26 <sipa> *can't
400 2012-10-15 19:53:54 <gmaxwell> Besides if we follow best pratices and get hacked we at least advance the state of the art in knowing what doesn't work.
401 2012-10-15 19:54:08 <gmaxwell> If we don't follow best practices and get hacked nothing is learned. :P
402 2012-10-15 19:54:45 <gmaxwell> (of course, what is "best practices"? ??? a question the bitcoin ecosystem would benefit from knowing the answer to!)
403 2012-10-15 19:54:54 <gavinandresen> writing up best practices is one of the missions of the Foundation
404 2012-10-15 19:55:43 <Luke-Jr> I'd suggest hiring a security expert for that; unless someone here is one?
405 2012-10-15 19:55:54 <Luke-Jr> (or even if someone here is one)
406 2012-10-15 19:56:26 <gmaxwell> Securing a static webserver is not a unique and special activity.
407 2012-10-15 19:57:52 <c_k> I suspect there is quite a number of security experts in the bitcoin community
408 2012-10-15 19:58:24 <gmaxwell> Unfortunately there are also quite a number of security "experts"... :P
409 2012-10-15 19:58:29 <c_k> however surely most of us are rather well versed in what constitutes good security practices and how to learn them
410 2012-10-15 19:58:54 <sipa> c_k: judging by several events in bitcoin's history, i'm not too sure about that :)
411 2012-10-15 19:59:05 <sipa> well, events in its ecosystem
412 2012-10-15 19:59:19 <Luke-Jr> I think the community would like to hear Bitcoin.org is secured so that compromising it means getting at at least 3 people, 2 of whom must be in different jurisdictions
413 2012-10-15 19:59:23 <gmaxwell> securing dynamic stuff is hard.
414 2012-10-15 19:59:49 <gmaxwell> Luke-Jr: step 1) replace https with something new that accomidates that mode of operation.
415 2012-10-15 20:00:06 <Luke-Jr> gmaxwell: if we don't use https, nobody NEEDS to compromise the server in the first place
416 2012-10-15 20:00:14 <Luke-Jr> they can just replace it
417 2012-10-15 20:00:21 <c_k> sipa: haha you are right, there are certainly some developers of bitcoin related web apps with poor security design
418 2012-10-15 20:01:18 <gmaxwell> Luke-Jr: you misunderstood me, I wasn't arguing against https... I'm pointing out that http/https doesn't facilitate secure decenteralized operation. There must be at least one server that if compromised makes the site compromised.
419 2012-10-15 20:02:19 <Luke-Jr> gmaxwell: I'm thinking more like there's no remote access, local access requires someone to fly out from another country, and updates are made via git commits signed off by 3+ (including 2 in diff jurisdictions)
420 2012-10-15 20:02:27 <gmaxwell> Luke-Jr: and when you try to get smart you take risks.. say the site itself gitian builds and checks gitian signatures. ... okay great, now someone inserts an exploit in a makefile. :( (not that these things can't be done safely, but it takes careful work)
421 2012-10-15 20:02:53 <sipa> ACTION zZzZ
422 2012-10-15 20:02:59 <gmaxwell> and if you lock down access too much you can't apply urgent security patches (less criticial if its a locked down otherwise static server)
423 2012-10-15 20:03:13 <Luke-Jr> true
424 2012-10-15 20:03:57 <Luke-Jr> anyhow, it's probably a bit too complicated as the "next step tomorrow"
425 2012-10-15 20:04:38 <gmaxwell> but sure, some kind of highly locked down, yubikey and ACL mediated... thing that takes updates by periodically fetching a multisigned tar with a copy of the whole site .. or something.. might be possible to make secure without boiling the oceans.
426 2012-10-15 20:04:49 <TD> by the way
427 2012-10-15 20:05:04 <TD> at some point i think it'd be a good idea to move the FAQ to static HTML stored in git
428 2012-10-15 20:05:08 <TD> the current wiki based FAQ is a mess
429 2012-10-15 20:05:50 <TD> possibly some other wiki content could move as well
430 2012-10-15 20:06:03 <TD> storing it in git means changes have to go through peer review before they go live and you don't get stupid edit wars
431 2012-10-15 20:07:45 <gmaxwell> have we had edit war problems on the FAQ?
432 2012-10-15 20:08:15 <gmaxwell> not that I don't agree.. though I also think the FAQ is fairly imbalanced right now, e.g. misses some actual FAQs and spends perhaps too much time on some subjects.
433 2012-10-15 20:08:19 <TD> yeah
434 2012-10-15 20:08:49 <TD> i made some edits and then some random guy reverted all of them, but not completely so the answer is now an unintelligible mess
435 2012-10-15 20:08:57 <TD> he seems to think the word "recourse" means "reversible"
436 2012-10-15 20:09:09 <TD> so the answer talks about non-recourse transactions, which makes no sense in English
437 2012-10-15 20:09:49 <TD> also the answer to the deflation question is a rambling essay that doesn't say much of anything
438 2012-10-15 20:10:28 <gmaxwell> ugh. well, one FAQ is in fact about non-technical methods of attack resistance and deterrence... lots of people seem surprisingly blown away by the idea that you could use totally boring mechnisms to secure bitcoin just as you use to secure other things.
439 2012-10-15 20:11:11 <Luke-Jr> TD: nanotube could always lock the page or something for the same effect
440 2012-10-15 20:11:12 <TD> also i linked to the ETH paper from that faq entry
441 2012-10-15 20:11:43 <TD> and that edit was reversed on the grounds that it was "blatant advertising for a paper which presented an idea that was already known"
442 2012-10-15 20:11:55 <Luke-Jr> O.o
443 2012-10-15 20:12:08 <gmaxwell> I think for some things like deflation we should go towards minimal answers "here are where people have discussed it. The general ideas are this. But bitcoin is an expirement and time will be the ultimate test."
444 2012-10-15 20:12:18 <TD> so just generally i think an FAQ that uses real HTML and is hammered out via code review process, would work better
445 2012-10-15 20:12:23 <TD> yeah. agreed.
446 2012-10-15 20:12:57 <gmaxwell> E.g. lets avoid getting into fighting about economic stuff in the faq. Lay out the broad ideas and give pointers and try to avoid making any furhter promises. :)
447 2012-10-15 20:13:07 <gmaxwell> hah further.
448 2012-10-15 20:15:18 <TD> the hashpower thread is sad
449 2012-10-15 20:15:42 <gmaxwell> where is this?
450 2012-10-15 20:16:10 <TD> i trolled it a bit
451 2012-10-15 20:16:21 <gmaxwell> (I mostly only read /tech unless I get complaints in mining or someone points me to something)
452 2012-10-15 20:16:21 <TD> https://bitcointalk.org/index.php?topic=118069.40;topicseen
453 2012-10-15 20:16:48 <TD> "So far I'm happy with the service and it's on par with Gpumax. I don't really see the sensationalizing, alarmist talk about these services. I've pointed my BFL Single crunching away at 800mhash/s and honestly as a miner I do not in any way shape or form care where my hashes go. "
454 2012-10-15 20:16:53 <TD> "Don't get me wrong, I support Bitcoin but I also go where the money is."
455 2012-10-15 20:17:31 <TD> http://www.regretsy.com/2012/09/10/barack-obama-is-tired-of-your-shit/
456 2012-10-15 20:17:45 <gmaxwell> yea, these threads make me regret being honest. So many people just _begging_ to be ripped off and to enable random strangers to rip off others.
457 2012-10-15 20:18:00 <gmaxwell> Interest in that stuff _after_ clipse poofed with everyone's coin.
458 2012-10-15 20:18:21 <gmaxwell> (Clipse, goat, and GPUmax were the prior most notable parties paying suspiciously high PPS rates for mining)
459 2012-10-15 20:18:51 <TD> how did they do that? it was backed by the ponzi scheme?
460 2012-10-15 20:19:41 <gmaxwell> Clipse stopped paying at the same time as pirate shut down. My personal best guess is that he was fronting zeekrewards.
461 2012-10-15 20:20:03 <gmaxwell> Of course, they won't disclose their methods??? or they give confused and illogical answers.
462 2012-10-15 20:20:38 <TD> i guess it's possible that ASICs will make it harder to run these sorts of services
463 2012-10-15 20:20:41 <gmaxwell> When I pressed goat about it on the forums he put up a couple hundred BTC bounty to "find" evidence connecting me to unlawful activities.
464 2012-10-15 20:20:58 <TD> i suppose it's harder to proxy work if you have to serve up new work every few milliseconds
465 2012-10-15 20:21:02 <gmaxwell> (he pulled it down after I started trying to collect on it!)
466 2012-10-15 20:21:12 <TD> lol
467 2012-10-15 20:21:23 <gmaxwell> TD: nope, the pools are solving that problem in any case, with local work creation.
468 2012-10-15 20:22:45 <TD> oh, right. they give you a coinbase to modify. you can't see the other transactions.
469 2012-10-15 20:22:56 <gmaxwell> Right.
470 2012-10-15 20:23:34 <gmaxwell> Well, alternatively they could offer getblocktemplate??? luke's pool (and pool software) does??? and his miner talks to it; which creates transparency at least.
471 2012-10-15 20:28:58 <BCB> FYI this transaction still has 0 comfirmations "48b0d02003a926d6db7fceb39b3acf5b43e9f65dbc03718883c05626dbfb75b6"
472 2012-10-15 20:30:16 <BCB> sent 2012-15-10 15:43:52 EST or about 3 hours ago
473 2012-10-15 20:30:26 <BCB> any was to track where these coins migh tbe
474 2012-10-15 20:30:42 <Luke-Jr> blockchain.info's never heard of it
475 2012-10-15 20:31:40 <BCB> Sorry we could not find any blocks or transactions matching this hash
476 2012-10-15 20:36:36 <BCB> Luke-Jr: any way to trace it
477 2012-10-15 20:39:09 <vampireb> well what did you use to send it? your local client?
478 2012-10-15 20:41:36 <gmaxwell> BCB: did you use the reference client to generate it?
479 2012-10-15 20:42:07 <jgarzik> BCB: go to http://bitcoincharts.com/bitcoin/txlist/ and look for your TX id. if it is not listed, it probably did not make it to the outside world.
480 2012-10-15 20:42:19 <jgarzik> BCB: perhaps due to insufficient fees? were you sending a tiny amount?
481 2012-10-15 20:43:21 <gmaxwell> jgarzik: stock client can't do that (insufficient fees) except via raw transactions.
482 2012-10-15 20:43:38 <jgarzik> gmaxwell: -settxfee
483 2012-10-15 20:43:57 <gmaxwell> Again, it can't result in insufficient fees for relaying.
484 2012-10-15 20:44:19 <gmaxwell> you can set it to zero all you want, it'll still force a the relay fee if it itself wouldn't relay it.
485 2012-10-15 20:44:37 <gmaxwell> the GUI asks but if you say no??? no transaction.
486 2012-10-15 20:45:12 <BCB> I'm using "version" : 70003,
487 2012-10-15 20:45:32 <gmaxwell> From bitcoin.org?
488 2012-10-15 20:45:46 <BCB> yes
489 2012-10-15 20:45:53 <BCB> send 7.09
490 2012-10-15 20:45:57 <gmaxwell> In any case, can you open the debug console and type getrawtransaction 48b0d02003a926d6db7fceb39b3acf5b43e9f65dbc03718883c05626dbfb75b6 an then pastebin what it spits back?
491 2012-10-15 20:46:27 <MC1984> heh greg nearly said fuhrer, total freudian slip, atlas was right!
492 2012-10-15 20:49:13 <BCB> http://pastebin.com/g31tP0Dp
493 2012-10-15 20:51:21 <gmaxwell> BCB: is that really it? it ended with 5626?
494 2012-10-15 20:51:51 <BCB> no wait
495 2012-10-15 20:52:39 <BCB> http://pastebin.com/mR35eYPw
496 2012-10-15 20:52:54 <BCB> ends in 7ff888ac00000000
497 2012-10-15 20:53:26 <gmaxwell> it's not all making it into the pastebin
498 2012-10-15 20:54:15 <gmaxwell> oh nevermind
499 2012-10-15 20:57:46 <gmaxwell> well, it's spending inputs I know nothing about.
500 2012-10-15 20:58:01 <gmaxwell> BCB: do you place dice?
501 2012-10-15 20:58:30 <BCB> play dice?
502 2012-10-15 20:58:31 <BCB> no
503 2012-10-15 20:59:01 <BCB> i imported and exported a few private keys
504 2012-10-15 20:59:30 <gmaxwell> do you have multiple unconfirmed transactions in your wallet?
505 2012-10-15 21:02:22 <gmaxwell> the reason I ask is because you're spending transactions my node hasn't seen.. which should only happen if they're outputs you wrote yourself.
506 2012-10-15 21:02:39 <gmaxwell> (you meaning your client)
507 2012-10-15 21:12:18 <BCB> gmaxwell: this is the first one I've seen
508 2012-10-15 21:12:42 <BCB> gmaxwell: no, there are not others
509 2012-10-15 21:13:42 <BCB> gmaxwell: I'm using a php program to generate keypairs
510 2012-10-15 21:13:56 <BCB> and importing those private keys to bitcoind
511 2012-10-15 21:14:46 <gmaxwell> can you paste me the getrawtransaction 2b7a3a4d24e05d1b0a4017835fb66044025d8e9be9b3d56c7a26312384b0bbf0
512 2012-10-15 21:14:59 <BCB> gmaswell: I also did a few multisig transac
513 2012-10-15 21:15:04 <BCB> tions
514 2012-10-15 21:16:40 <BCB> gmaxwell:http://pastebin.com/X2wTs0sQ
515 2012-10-15 21:17:24 <gmaxwell> okay, that one spends cc2b8a0335b8af49c49ae94aea1a97c959922591848c94b0611d74a85b2b3f48 that my node knows nothiing about ...
516 2012-10-15 21:17:36 <gmaxwell> can you getrawtransaction cc2b8a0335b8af49c49ae94aea1a97c959922591848c94b0611d74a85b2b3f48 ?
517 2012-10-15 21:21:18 <lianj> ERROR: ConnectInputs() : 2b7a3a4d24e05d1b0a4017835fb66044025d8e9be9b3d56c7a26312384b0bbf0 prev tx already used at (nFile=1, nBlockPos=1422618605, nTxPos=1422620633)
518 2012-10-15 21:22:01 <gmaxwell> BCB: you ever get paid by coinbase during their free coin for signing up thing?
519 2012-10-15 21:24:09 <MC1984> i dont get this hashpower buying and selling stuff
520 2012-10-15 21:24:41 <MC1984> how can it ever be more profitable to sell hashes than to just use them to mine
521 2012-10-15 21:25:00 <BCB> gmaxwell: no
522 2012-10-15 21:25:17 <jgarzik> MC1984: (1) if you want to mine a non-standard transaction, (2) if you want to launder money, exchanging dirty coins for clean
523 2012-10-15 21:25:43 <jgarzik> I'd love to know how to enable the first without the second
524 2012-10-15 21:26:07 <MC1984> oh
525 2012-10-15 21:26:17 <amiller> Luke-Jr, wtf "networkserver" is just more crap piled on top of asyncore
526 2012-10-15 21:26:20 <MC1984> not more profitable, but has ancilliary uses
527 2012-10-15 21:26:21 <amiller> Luke-Jr, what about pynode doesn't work in python3?
528 2012-10-15 21:26:23 <amiller> also what features does eloipool have that pynode doesn't? maybe it won't be too hard to merge our efforts..
529 2012-10-15 21:26:23 <gmaxwell> MC1984: (3) you're just a scammer who is going to run with the funds or otherwise invest the float in questionable 'investments'
530 2012-10-15 21:26:47 <gmaxwell> amiller: eloipool doesn't implement a bitcoin node.
531 2012-10-15 21:27:42 <BCB> gmaxwell: http://pastebin.com/PvNBf7MX
532 2012-10-15 21:28:32 <amiller> nvm i misread the conversation above and thought someone was complaining about eloipool and pynode being two different versions duplicating effort
533 2012-10-15 21:28:36 <BCB> gmaxwell: that last transaction was probably made on a bitcoind 3 or four versions ago
534 2012-10-15 21:28:48 <jgarzik> pynode still lacks some of the heretofor unneeded block building validation checks
535 2012-10-15 21:28:51 <amiller> anyway that's still the case regarding networkserver
536 2012-10-15 21:29:10 <gmaxwell> BCB: and how many confirmations does it have?
537 2012-10-15 21:29:17 <gmaxwell> and what is your current block count?
538 2012-10-15 21:30:08 <BCB> gmaxwell: "confirmations" : 31797,
539 2012-10-15 21:30:23 <alexmat> wtb bitcoins