1 2012-05-18 00:02:32 <jgarzik> I do understood it takes a step or nine to get there
  2 2012-05-18 00:03:40 <BlueMatt> agreed
  3 2012-05-18 00:04:29 <gmaxwell> Hm. I don't think that I really would have split further than wallets and blockchain/p2p... simply because you can define the blockchain<->network protocol as the current p2p protocol, and the null program is then your network program for the p2p network.
  4 2012-05-18 00:06:22 <BlueMatt> i prefer it mostly because a vuln in the p2p client only means you can block new blocks from getting added, not changing the chain itself
  5 2012-05-18 00:07:07 <BlueMatt> and, imho, security should be more closely watched with chroot/atleast separate processes for bitcoin
  6 2012-05-18 00:08:00 <BlueMatt> ok, I think the pull is sane, be careful when building if you have some objects not built with DEBUG_LOCKORDER/DEBUG_LOCKCONTENTION and some with
  7 2012-05-18 00:08:34 <BlueMatt> oh, wait, no need to make stuff before b39 work
  8 2012-05-18 00:10:34 <gmaxwell> I think debug lockorder is still broken.
  9 2012-05-18 00:10:56 <BlueMatt> it wasnt on the plane (after the commit titled Fix DEBUG_LOCKORDER)
 10 2012-05-18 00:11:05 <gmaxwell> oh. missed that one!
 11 2012-05-18 00:13:13 <jgarzik> gmaxwell: yeah, it is debatable whether network is a separate entity, or should be within the conceptual "blockchain server"
 12 2012-05-18 00:14:57 <gmaxwell> In any case, thats down the road the wallet blockchain step is obviously higher priority and a bigger win (due to running N:1 wallets to blockchains, and the privsep)
 13 2012-05-18 00:15:12 <luke-jr> yeah
 14 2012-05-18 00:15:24 <BlueMatt> yea, cleaner wallet/chain is next after cblockstore
 15 2012-05-18 00:15:37 <BlueMatt> (atleast not callbacks to wallet, but its pretty close)
 16 2012-05-18 00:15:44 <gmaxwell> luke-jr: it has risk too. Breaking network communications can leave you on forks and if the blockchain is the only one doing e.g. script validation a lot of harm could just tunnel through (and if they both do, again potential forks)
 17 2012-05-18 00:16:07 <luke-jr> hmm
 18 2012-05-18 00:30:23 <gribble> New news from bitcoinrss: gavinandresen opened pull request 1349 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1349>
 19 2012-05-18 00:40:29 <seco> wtf BIGipServerIngeus_ScanSafe_COL3_AppSrv_app~Ingeus_ScanSafe_COL3_AppSrv_pool' (attacker '80.254.146.132')
 20 2012-05-18 00:40:43 <seco> offtopic from apache^^
 21 2012-05-18 00:43:45 <seco> Scansafe Ops ...
 22 2012-05-18 01:35:40 <JohanBerg> Mr. Gmaxwell, we need to have a chat.
 23 2012-05-18 01:43:34 <Tril> trying to build out of source tarball download and it worked, but with warning about not a git repo and bitcoind -? shows version vCLIENT_VERSION_MAJOR.CLI...
 24 2012-05-18 01:43:48 <Tril> 0.6.2
 25 2012-05-18 01:44:14 <gmaxwell> Tril: this is fixed in 0.7.0 but not in the 0.6.2 release.
 26 2012-05-18 01:44:29 <gmaxwell> Tril: the patch is pretty simple if you want to apply it directly.
 27 2012-05-18 01:44:45 <gmaxwell> or you could just pull 0.6.2 from git
 28 2012-05-18 01:45:26 <gmaxwell> git clone git://github.com/bitcoin/bitcoin.git ; git checkout 0.6.2
 29 2012-05-18 01:46:46 <gribble> New news from bitcoinrss: jgarzik opened pull request 1350 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1350>
 30 2012-05-18 02:11:35 <gmaxwell> Hazards of coinbase triggered feature switchover, p2pool just suffered the blockchain equivilent of a boiling liquid expanding vapor explosion.
 31 2012-05-18 02:12:05 <luke-jr> gmaxwell: ?
 32 2012-05-18 02:12:43 <gmaxwell> luke-jr: bug in a corner case bit of latest protocol code which was triggered after p2pool crossed the 95% switchover threshold but fell below it.
 33 2012-05-18 02:12:54 <gmaxwell> Caused upgraded peers to reject their own shares.
 34 2012-05-18 02:13:17 <luke-jr> O.o
 35 2012-05-18 02:13:24 <gmaxwell> amusingly it shouldn't have harmed their ability to mine bitcoins, they all just became donor nodes to random unupgraded nodes. :)
 36 2012-05-18 02:13:39 <sipa> jgarzik: with DB_HASH, does BDB use its own hash function, or does it expect hashed keys to be passed in?
 37 2012-05-18 02:13:42 <Tril> allright what's a v0.6.2.2-dirty-beta
 38 2012-05-18 02:13:59 <gmaxwell> it should self correct once the fixed nodes produce enough shares.
 39 2012-05-18 02:14:12 <weex> when you say upgrade gmaxwell, you mean the latest security upgrade?
 40 2012-05-18 02:14:14 <gmaxwell> Tril: naughty you running dirty dirty code *wink*
 41 2012-05-18 02:14:26 <Tril> gmaxwell: which you told me to (what is it?)
 42 2012-05-18 02:14:37 <sipa> Tril: dirty = from a source tree that was not entirely clean (modifications not in commits)
 43 2012-05-18 02:14:38 <gmaxwell> weex: no, nothing to do with bitcoin.
 44 2012-05-18 02:15:00 <gmaxwell> sipa: I think the dirty detection is somewhat lossy, since I've seen that tag on code that was probably clean... didn't interest me enough to check.
 45 2012-05-18 02:15:01 <weex> oh the p2pool program?
 46 2012-05-18 02:15:05 <gmaxwell> weex: yes.
 47 2012-05-18 02:15:43 <gmaxwell> Tril: it just means that it thinks your code is not exactly equal to the v0.6.2.2 tag. e.g. you changed something else.
 48 2012-05-18 02:16:05 <sipa> gmaxwell: that may well be - once gitian was able to build without the dirty tag, it was fine to me
 49 2012-05-18 02:16:45 <Tril> sipa: shouldn't be I did git clone, git checkout 0.6.2, build it (64-bit), make clean, scp to 32-bit machine machine, then built it again. Make clean doesn't stop it from being dirty?
 50 2012-05-18 02:17:24 <gmaxwell> it shouldn't be dirty unless you actually edited some of the files after doing the checkout... but don't worry about it, the detection is probably just hyperactive.
 51 2012-05-18 02:17:37 <sipa> Tril: do both 32 and 64 bit builds have the dirty tag?
 52 2012-05-18 02:17:45 <gmaxwell> The reason for that tag is so when you ask for support we know to ask you what evil patches you've applied.
 53 2012-05-18 02:18:06 <luke-jr> network is 6.5% secure against CVE-2012-2459
 54 2012-05-18 02:18:16 <Tril> sipa: no just the second one does
 55 2012-05-18 02:18:29 <gmaxwell> luke-jr: is that a seeds.txt based count?
 56 2012-05-18 02:18:35 <luke-jr> yes
 57 2012-05-18 02:18:35 <sipa> Tril: scp may have messed up timestamps
 58 2012-05-18 02:19:14 <sipa> seeds is not very reliable, as it does not contain addresses of nodes that are frequently off
 59 2012-05-18 02:19:59 <luke-jr> Unfortunately, while my OSX zlib and minipnpc builds are deterministic, the crosscompiler, bdb, and openssl builds are not :<
 60 2012-05-18 02:20:03 <sipa> well, it is reliable, but for weighted data (counting active nodes more strongly)
 61 2012-05-18 02:20:12 <luke-jr> sipa: those don't matter much afaik
 62 2012-05-18 02:20:18 <sipa> luke-jr: ah, too bad :(
 63 2012-05-18 02:20:25 <gmaxwell> luke-jr: oh you can cross comple for OSX?
 64 2012-05-18 02:20:30 <luke-jr> gmaxwell: yes, working on it
 65 2012-05-18 02:20:36 <luke-jr> gmaxwell: https://gitorious.org/cross-osx
 66 2012-05-18 02:21:07 <gmaxwell> very cool
 67 2012-05-18 02:21:40 <sipa> doesn't apple use clang these days?
 68 2012-05-18 02:21:48 <luke-jr> shrug
 69 2012-05-18 02:23:08 <luke-jr> I'm aiming for max compatibility, not "what apple does"
 70 2012-05-18 02:23:10 <luke-jr> :p
 71 2012-05-18 02:24:24 <sipa> no, but it'll probably take a lot of maintanence work just to keep it working
 72 2012-05-18 02:24:40 <sipa> impressive that some parts already work
 73 2012-05-18 02:26:10 <gmaxwell> Might actually end up being the only way to compile w/ modern gcc for osx at some point in the future. :)
 74 2012-05-18 02:32:17 <sipa> jgarzik: seems bdb uses its own hash function; i wonder if we should use a custom one that uses a per-node secret, as you may expose yourself to a dos attack where one bucket is deliberately filled with transactions otherwise
 75 2012-05-18 03:02:02 <jgarzik> sipa: or a custom one that simply returns N bits of the key
 76 2012-05-18 03:02:10 <jgarzik> since it's already hashed...
 77 2012-05-18 03:03:10 <jgarzik> sipa: fyi there is a reference counting bug somewhere in my bdb commits.  should have it nailed down tomorrow
 78 2012-05-18 03:03:55 <jgarzik> data seems intact, but refcounts show crazy numbers upon shutdown
 79 2012-05-18 03:13:09 <jgarzik> would probably produce some new and useful warnings
 80 2012-05-18 04:57:42 <jgarzik> ok, fixed refcount crap
 81 2012-05-18 05:07:02 <sipa> jgarzik: that's even easier to attack
 82 2012-05-18 05:07:45 <sipa> jgarzik: creating tons of transactions in the same bucket will make lookup operations linear
 83 2012-05-18 05:08:27 <sipa> not sure how much of a problem that is, though
 84 2012-05-18 06:00:40 <Eliel> I remember seeing this bar graph somewhere that had a bar for each day counting the number of bitcoins that were last transfered (or created) during that day. Does anyone have the link?
 85 2012-05-18 06:06:32 <jarpiain> Eliel: http://ecdsa.org/stats.html
 86 2012-05-18 06:07:22 <Eliel> ah, thank you :)
 87 2012-05-18 06:08:09 <Eliel> ah, unfortunately, it hasn't been updated since the last time I looked at it :/
 88 2012-05-18 06:10:59 <gribble> New news from bitcoinrss: Diapolo opened pull request 1351 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1351>
 89 2012-05-18 06:49:15 <moa7> my client is stuck downloading blocks on 180380 ... anyone else reported this problem?
 90 2012-05-18 06:49:51 <moa7> ERROR: Reorganize() : ConnectBlock 00000000000009851668 failed InvalidChainFound: invalid block=00000000000005ff826d  height=180455  work=326201203339013528496 InvalidChainFound:  current best=00000000000002edc64e  height=180380  work=325642889600634942846 InvalidChainFound: WARNING: Displayed transactions may not be correct!  You may need to upgrade, or other nodes may need to upgrade.
 91 2012-05-18 06:50:12 <moa7> using local build of official 0.6.2-beta release
 92 2012-05-18 06:50:23 <moa7> on ubuntu 11.04
 93 2012-05-18 06:53:00 <gribble> New news from bitcoinrss: Diapolo opened pull request 1352 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1352>
 94 2012-05-18 06:54:11 <Eliel> moa7: those symptoms sound similar to the attack the latest release was supposed to fix. I wonder if your version has th fix since it's got -beta attached the end of the version.
 95 2012-05-18 06:54:15 <rebroad> moa7 that is interesting...
 96 2012-05-18 06:54:39 <moa7> all clients are beta ... as far as i'm aware
 97 2012-05-18 06:54:43 <rebroad> Eliel, what got fixed?
 98 2012-05-18 06:54:54 <rebroad> moa7, yes, I think they're all beta too
 99 2012-05-18 06:55:52 <moa7> anybody got a client that will get over the latest "hump" in the blockchain ... getting kinda old
100 2012-05-18 06:56:19 <moa7> maintaining a blockchain is only for devs now?
101 2012-05-18 06:59:20 <rebroad> hmmm. I'm having git troubles :-s I'm just doing a rebase -i upstream/master in my master, and I've got about 25 lines of picks. Many of them repeated 2 or 3 times each. Would anyone know what I need to do here please?
102 2012-05-18 07:02:14 <Eliel> rebroad: you didn't see the announcement?
103 2012-05-18 07:02:56 <Eliel> rebroad: https://bitcointalk.org/index.php?topic=81749.0
104 2012-05-18 07:08:12 <t7> oops
105 2012-05-18 07:08:16 <t7> better upgrade my node
106 2012-05-18 07:13:19 <t7> got i cant find what screen session bitcoind is running on :|
107 2012-05-18 07:13:25 <t7> im such a unix noob
108 2012-05-18 07:20:26 <t7> how do i give user access to a folder in unix?
109 2012-05-18 07:21:38 <Eliel> that depends on which users you want to have access to it. If just one user, you can change the owner. Alternatively you can give a group access to the folder. Third option is to make it public.
110 2012-05-18 07:22:03 <t7> ah chown
111 2012-05-18 07:22:05 <Eliel> you can use the chown command to do the first two and chmod command to alter the access rights
112 2012-05-18 07:22:11 <t7> ls
113 2012-05-18 07:34:32 <t7> its taking years to build
114 2012-05-18 07:34:35 <t7> bloody 512 ram
115 2012-05-18 07:35:36 <TuxBlackEdo> t7, boost?
116 2012-05-18 07:35:48 <TuxBlackEdo> boost takes few minute to compike
117 2012-05-18 07:35:52 <t7> bitcoinrpc.cpp took like 5 mins
118 2012-05-18 07:35:58 <t7> i think it ran outa ram
119 2012-05-18 07:36:04 <t7> i turned off my java webserver
120 2012-05-18 07:36:10 <t7> and its finished :)
121 2012-05-18 07:36:47 <t7> i need to redo my website
122 2012-05-18 10:15:11 <gribble> New news from bitcoinrss: Diapolo opened pull request 1353 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1353>
123 2012-05-18 10:44:00 <drizztbsd> hi, is there any (json?) api to have the last blocks found by someone? (p2pool)
124 2012-05-18 10:44:10 <drizztbsd> or website
125 2012-05-18 11:19:40 <TD> sipa: the stuck blockchain issue ..... it basically occurs if you're on a side chain > 500 blocks?
126 2012-05-18 11:21:28 <sipa> TD: actually if you already have 500+ blocks of the chain you're switching to
127 2012-05-18 11:21:43 <sipa> or should be switching to
128 2012-05-18 11:22:38 <sipa> it should only occur if there is a rule change on the network, or a bug
129 2012-05-18 11:23:13 <TD> sipa: ok
130 2012-05-18 11:23:35 <TD> sipa: what version of boost are you using?
131 2012-05-18 11:23:42 <TD> it seems the version i have is not new enough but there's no version check in the build
132 2012-05-18 11:24:00 <TD> i seem to have 1.49
133 2012-05-18 11:25:16 <sipa> that's certainly new enough (i believe the PPA builds are done using 1.37, i have 1.46, and i'm sure several other versions are used as well)
134 2012-05-18 11:25:21 <sipa> what error do you get?
135 2012-05-18 11:25:21 <TD> ok
136 2012-05-18 11:25:43 <TD> /Users/hearn/bitcoin/gitrepo/src/qt/qtipcserver.cpp:32: error: no matching function for call to 'boost::interprocess::message_queue_t<boost::interprocess::offset_ptr<void, int, long unsigned int, 0u> >::timed_receive(char (*)[257], long unsigned int, size_t&, unsigned int&, boost::posix_time::ptime&)'
137 2012-05-18 11:26:09 <TD> inside message_queue.hpp i get another error, so maybe my boost is broken somehow
138 2012-05-18 11:26:15 <TD> /opt/local/include/boost/interprocess/ipc/message_queue.hpp:325:0 /opt/local/include/boost/interprocess/ipc/message_queue.hpp:325: error: no matching function for call to 'get_rounded_size(long unsigned int, const unsigned int&)'
139 2012-05-18 11:26:17 <sipa> that's current git head?
140 2012-05-18 11:26:21 <TD> yeah
141 2012-05-18 11:26:39 <sipa> it's possible someone recently broke something for OSX
142 2012-05-18 11:27:24 <sipa> i believe the IPC code differs between platforms
143 2012-05-18 11:27:43 <TD> hmm
144 2012-05-18 11:27:51 <TD> seems like it may be a 32 vs 64 bit snafu again
145 2012-05-18 11:27:58 <TD> let's see if it makes any difference
146 2012-05-18 11:28:07 <sipa> afaik gavin does the OSX builds for 32-bit
147 2012-05-18 11:28:22 <TD> yeah. i think when i set things up i ended up with 64 bit builds of the dependencies
148 2012-05-18 11:28:33 <TD> so now i always have to remember to build as 64 bit. rerunnnig qmake reverts it to 32 bit
149 2012-05-18 11:28:35 <TD> sigh ...
150 2012-05-18 11:28:36 <gavinandresen> yes, I have macports build everything 32-bit i386-only
151 2012-05-18 11:28:54 <gavinandresen> The get_rounded_size problem sounds familiar....
152 2012-05-18 11:29:04 <TD> ah, there he is. i was about to compliment you on the new github profile pic, very professional looking :-)
153 2012-05-18 11:30:05 <TD> gavinandresen: i didn't see any obvious bugs with the cache, but yeah i think the eviction policy could use a bit more discussion. also i'm not really a fan of code like (if r.first) return r.second;
154 2012-05-18 11:30:08 <TD> otherwise lgtm
155 2012-05-18 11:30:17 <TD> oh
156 2012-05-18 11:30:34 <TD> the 200 bytes per cache entry
157 2012-05-18 11:30:35 <TD> presume tha
158 2012-05-18 11:30:43 <TD> does that include the overhead of std::map ?
159 2012-05-18 11:31:17 <gavinandresen> yes, I was generous, it is 32 byte hash, 75 byte signature, 65 byte public key
160 2012-05-18 11:31:23 <gavinandresen> ... so lots of space for map overhead
161 2012-05-18 11:31:27 <TD> cool
162 2012-05-18 11:31:47 <TD> maybe remark on how 200 bytes was arrived at in the comment
163 2012-05-18 11:32:13 <gavinandresen> RE: r.first/r.second:  what's your favorite C++ paradigm for returning both a result and a "have-result" code?
164 2012-05-18 11:32:52 <gavinandresen> I could change it to match the rest of the code and pass in a reference to the result....
165 2012-05-18 11:33:08 <TD> some libraries have a Maybe<T> or Optional<T> type, somewhat a la haskell. alternatively just struct FooResult { boolean success; boolean valid; } or something
166 2012-05-18 11:33:52 <gavinandresen> ah, Boost.Optional....
167 2012-05-18 11:33:55 <sipa> also, pointers have an implicit "no-data" value, NULL
168 2012-05-18 11:34:10 <TD> yeah, but they also add memory management problems. returning a value type feels right
169 2012-05-18 11:34:17 <TD> it's just a question of how to make the code as readable as possible
170 2012-05-18 11:34:26 <sipa> (i didn't check the actual code yet)
171 2012-05-18 11:34:45 <sipa> a Maybe type certainly sounds perfect if it exists
172 2012-05-18 11:34:55 <gavinandresen> http://www.boost.org/doc/libs/1_38_0/libs/optional/doc/html/boost_optional/examples.html#boost_optional.examples.optional_return_values
173 2012-05-18 11:35:16 <TD> yep
174 2012-05-18 11:35:19 <gavinandresen> in for a penny, in for a pound, right?
175 2012-05-18 11:36:38 <TD> looks like then you can do if ((fValid = GetSigCache(...)) return *fValid;
176 2012-05-18 11:36:40 <TD> or something like that
177 2012-05-18 11:36:57 <TD> my god the documentation for this thing is crazy
178 2012-05-18 11:36:59 <TD> they wrote an entire book on it
179 2012-05-18 11:37:55 <TD> sipa: got an IPv6 connection handy?
180 2012-05-18 11:38:09 <TD> sipa: i'm wondering if you can connect to my node or if i'm firewalled off (most likely)
181 2012-05-18 11:38:43 <gavinandresen> ok, I'm looking at optional.hpp and I'm getting scared....
182 2012-05-18 11:40:50 <sipa> TD: IP?
183 2012-05-18 11:41:09 <TD> 2620:0:105f:311:fa1e:dfff:fee9:c223
184 2012-05-18 11:41:15 <TD> i can try connecting to your node if not
185 2012-05-18 11:41:19 <TD> right now all my connections at v4
186 2012-05-18 11:41:57 <BlueMatt> 2001:470:9ff2:1:ee55:f9ff:fec6:e666 should work
187 2012-05-18 11:43:28 <sipa> sec, rebuilding master
188 2012-05-18 11:43:51 <BlueMatt> TD: getting connection timeout
189 2012-05-18 11:44:13 <TD> ok
190 2012-05-18 11:44:19 <gavinandresen> #include <boost/optional.h> pulls in almost 31,000 lines of code
191 2012-05-18 11:44:22 <TD> probably we drop inbound connections on 8333 at the border
192 2012-05-18 11:44:25 <TD> i'll try connecting outbound
193 2012-05-18 11:44:33 <TD> gavinandresen: er, wow. how did they manage that??
194 2012-05-18 11:44:48 <BlueMatt> oops, I might have that on too, lemme go check the fw
195 2012-05-18 11:44:50 <gavinandresen> dependencies on their generic template metaprogramming libraries
196 2012-05-18 11:44:58 <BlueMatt> been too long sice Ive been home
197 2012-05-18 11:45:10 <TD> ugh
198 2012-05-18 11:45:13 <sipa> mine is 2001:6a8:200:b::2 (and should be up in a few seconds)
199 2012-05-18 11:45:18 <TD> a simple equivalent might be useful
200 2012-05-18 11:45:22 <TD> or not. up to you.
201 2012-05-18 11:45:25 <gavinandresen> ok, I'm backing away slowly from optional.hpp ....
202 2012-05-18 11:45:42 <sipa> an Optional<T> should be maybe 20 lines to implement...
203 2012-05-18 11:46:18 <gavinandresen> yes, but them we forevermore have to answer people asking "why not just use boost::Optional ?
204 2012-05-18 11:46:51 <gavinandresen> life is too short, too many higher priorities
205 2012-05-18 11:46:52 <sipa> and to prevent that you put a comment on top of the implementation "Using ad-hoc implementation instead of boost::Optional, as that pulls in 32k lines of code"
206 2012-05-18 11:47:57 <Dagger2> talking of v6, what's up with 2a02:348:5e:5a29::1? it's the one and only v6 DNS seed, but it's refusing connections
207 2012-05-18 11:47:58 <TD> sipa: am i supposed to surround with braces or something? i tried with -connect=2001:6a8:200:b::2 but it doesn't seem to be working
208 2012-05-18 11:48:08 <sipa> TD: try braces around it
209 2012-05-18 11:48:13 <BlueMatt> Dagger2: ohh...ops my fault
210 2012-05-18 11:48:17 <BlueMatt> one sec
211 2012-05-18 11:48:25 <BlueMatt> anyone have any 24/7 ipv6 nodes?
212 2012-05-18 11:48:32 <TD> i might be able to bring one online
213 2012-05-18 11:48:39 <Dagger2> (I'm being awkward and trying to bring up a node on a v6-only machine, and it's getting nowhere as a result)
214 2012-05-18 11:48:47 <TD> sipa: right that works
215 2012-05-18 11:48:58 <sipa> accepted connection [2620:0:105f:311:fa1e:dfff:fee9:c223]:51382
216 2012-05-18 11:49:22 <kinlo> I might also be able to maintain an ip6 node
217 2012-05-18 11:49:40 <kinlo> should the need be there
218 2012-05-18 11:49:40 <sipa> there's a -onlynet=ipv6 option, if you like
219 2012-05-18 11:49:44 <TD> yay
220 2012-05-18 11:49:59 <sipa> receive version message: version 60001, blocks=179521, us=[2001:6a8:200:b::2]:8333, them=0.0.0.0:0, peer=[2620:0:105f:311:fa1e:dfff:fee9:c223]:51382
221 2012-05-18 11:50:02 <TD> sipa: though it doesn't seem to be receiving the block chain
222 2012-05-18 11:50:12 <sipa> TD: my node is busy catching up itself
223 2012-05-18 11:50:16 <TD> ah that'd be why
224 2012-05-18 11:50:36 <Dagger2> the other possible issue I noted is that none of the providers in GetMyExternalIP() work over IPv6
225 2012-05-18 11:50:41 <luke-jr> sipa: https://bitcointalk.org/?topic=79884.msg907347#msg907347 <-- IPv6 issues?
226 2012-05-18 11:50:48 <Dagger2> at least one of them has v6 but it's broken...
227 2012-05-18 11:51:34 <TD> sipa: is IPv6 only in master?
228 2012-05-18 11:51:42 <sipa> TD: i think so
229 2012-05-18 11:52:22 <TD> ok
230 2012-05-18 11:52:25 <sipa> luke-jr: could be, it's far less tested on windows
231 2012-05-18 11:52:58 <sipa> TD: it seems your bitcoind did not manage to find out its own IPv6 address
232 2012-05-18 11:53:00 <Dagger2> (specifically www.showip.com; 2a01:608:ffff:2872::10 is dead, despite still being published in DNS)
233 2012-05-18 11:53:10 <TD> the sync bar went away? :(
234 2012-05-18 11:53:27 <TD> sipa: it does that using IRC or something, no?
235 2012-05-18 11:53:35 <TD> sipa: so if you use -connect how does it find out its own address
236 2012-05-18 11:54:06 <sipa> TD: oh, if you use -connect it disables listening, so it will never announce its own address
237 2012-05-18 11:54:10 <TD> yeah
238 2012-05-18 11:54:13 <TD> mystery solved
239 2012-05-18 11:54:23 <TD> once IPv6 is released i'll throw up a node on an ipv6 colo box i have
240 2012-05-18 11:54:33 <TD> if i run random master builds there i'll just forget to update
241 2012-05-18 11:54:40 <TD> or fall behind when dependencies change, etc
242 2012-05-18 11:54:43 <kinlo> ah good, that was still an open question for me wether my hidden bitcoin node was announcing it's address or not
243 2012-05-18 11:55:30 <sipa> kinlo: try connecting to me, and i'll tell you what it announces
244 2012-05-18 11:55:39 <sipa> [2001:6a8:200:b::2]
245 2012-05-18 11:55:53 <kinlo> sipa: I'll do it tomorrow if you don't mind, I have to configure ipv6 first on this node
246 2012-05-18 11:55:58 <sipa> sure
247 2012-05-18 11:56:25 <gmaxwell> kinlo: nodes that have listening disabled never annouced.. that was changed to do that around 0.6.0 I think.
248 2012-05-18 11:57:14 <kinlo> ic
249 2012-05-18 11:58:00 <kinlo> sipa: do ask for it or I will forget
250 2012-05-18 11:58:03 <Diapolo> sipa: is [2001:6a8:200:b::2] a testnet node or on normal net?
251 2012-05-18 11:58:16 <sipa> normal
252 2012-05-18 11:58:33 <sipa> (but it's not up very often)
253 2012-05-18 11:58:44 <Diapolo> I can't connect it seems ...
254 2012-05-18 11:58:51 <sipa> command line?
255 2012-05-18 11:58:59 <Diapolo> long ^^
256 2012-05-18 11:59:18 <sipa> i didn't ask about its length ;)
257 2012-05-18 12:00:27 <Diapolo> -externalip="6to4 IP from my Router 2002:..." -bind="local machine IP 2002:..." -onlynet="IPv6" -discover=0 -connect="2001:6a8:2bin00:b::2"
258 2012-05-18 12:00:52 <BlueMatt> double check that address
259 2012-05-18 12:01:20 <sipa> Diapolo: put [] around my address
260 2012-05-18 12:01:43 <sipa> as it probably misinterprets it as [2001:6a8:200:b:]:2
261 2012-05-18 12:02:26 <BlueMatt> you also have a random bin thrown in the address, as you pasted it
262 2012-05-18 12:02:26 <Diapolo> Bound to [2002:5898:...]:8333
263 2012-05-18 12:02:47 <Diapolo> so I dont use "" for addresses but []
264 2012-05-18 12:02:50 <Diapolo> ?
265 2012-05-18 12:02:53 <sipa> what is that 'bin' doing inside my address?
266 2012-05-18 12:03:11 <Diapolo> you are right ... wait a sec
267 2012-05-18 12:03:53 <Diapolo> nice
268 2012-05-18 12:03:56 <Diapolo> working
269 2012-05-18 12:04:20 <sipa> ok, i see tons of getblocks
270 2012-05-18 12:04:31 <Diapolo> yes it is dl the chain
271 2012-05-18 12:04:58 <BlueMatt> ok, random non-accepting ipv6 should be removed from dnsseed after the next update
272 2012-05-18 12:04:58 <sipa> it's a sixxs tunnel over wifi, via 2 nats... don't expect it to be fast :)
273 2012-05-18 12:05:24 <Diapolo> sipa: it IS fast :) sorry
274 2012-05-18 12:05:52 <Diapolo> really seems rather quick to me
275 2012-05-18 12:06:11 <Diapolo> sipa: do I have to use  [IPV6_Addr] always to be safe?
276 2012-05-18 12:07:10 <sipa> Diapolo: yes
277 2012-05-18 12:07:23 <sipa> Diapolo: if it's a service; if it's a hostname, it doesn't matter
278 2012-05-18 12:07:32 <Diapolo> alright, thanks
279 2012-05-18 12:07:32 <sipa> let me double check that
280 2012-05-18 12:08:55 <Diapolo> sipa: are you able to -connect to my IP, too? I use server=1 but am unsure if my routers firewall is really IPv6 ready and passes the packets to me
281 2012-05-18 12:09:10 <sipa> address?
282 2012-05-18 12:09:34 <gribble> New news from bitcoinrss: fanquake opened pull request 1354 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1354>
283 2012-05-18 12:09:38 <Diapolo> sipa: [2002:5898:c338:8000:be05:43ff:febc:fc65]
284 2012-05-18 12:09:40 <moa7> so has anybody else reported the 180380 block problem?
285 2012-05-18 12:09:49 <sipa> moa7: still stuck?
286 2012-05-18 12:09:52 <Dagger2> it should be easy enough to remove the requirement for using [] in the case of not specifying a port
287 2012-05-18 12:10:00 <moa7> yep, tried three different builds now
288 2012-05-18 12:10:11 <moa7> bitcoind-0.5.99
289 2012-05-18 12:10:24 <moa7> bitcoin-qt 0.6.2-beta
290 2012-05-18 12:10:28 <Dagger2> i.e. if there's more than one :, treat the whole thing as an IP unless there are also [] brackets in it
291 2012-05-18 12:10:43 <moa7> bitcoin-qt-0.6.99-alpha
292 2012-05-18 12:11:09 <sipa> moa7: how long did you leave 0.6.2 running
293 2012-05-18 12:11:45 <moa7> it was trying to reorganize 30-40 blocks past 180380 ... or so
294 2012-05-18 12:12:01 <sipa> debug.log?
295 2012-05-18 12:12:03 <moa7> how long should I leave it running?
296 2012-05-18 12:12:18 <sipa> enough to see a new block coming in
297 2012-05-18 12:12:41 <moa7> 'new' as in how?
298 2012-05-18 12:12:51 <sipa> new as in freshly mined
299 2012-05-18 12:13:15 <moa7> nup didn't do that
300 2012-05-18 12:13:31 <luke-jr> hmm, bdb isn't deterministic because it has some LC_UUID in binaries
301 2012-05-18 12:13:58 <moa7> what is in the block that is causing the issue?
302 2012-05-18 12:14:51 <sipa> moa7: either you have a strange reorgnisation issue (i don't think forks of >4 blocks have occurred on mainnet), or you were just not patient enough
303 2012-05-18 12:15:21 <sipa> so a debug.log would be very useful, to rule out the first
304 2012-05-18 12:17:02 <moa7> is bitcoind.0.5.99 the latest stable?
305 2012-05-18 12:19:58 <luke-jr> moa7: 0.5.99 was never stable
306 2012-05-18 12:20:10 <moa7> so which one was?
307 2012-05-18 12:20:13 <luke-jr> moa7: latest stable: 0.6.2, 0.6.0.7, 0.5.5, 0.4.6
308 2012-05-18 12:20:20 <moa7> bitcoind?
309 2012-05-18 12:20:29 <luke-jr> yes
310 2012-05-18 12:21:04 <sipa> moa7: there was never even a 0.5.99 released - that's the internal version number used for builds done from code during 0.6.0's merge window
311 2012-05-18 12:21:12 <luke-jr> 0.6.2 is most likely to recover from stuck blockchain
312 2012-05-18 12:21:41 <sipa> any 0.5.99 you're running is most certainly broken in many ways
313 2012-05-18 12:21:49 <moa7> thnx
314 2012-05-18 12:22:28 <sipa> running unreleased code is always welcome to help testing, but don't use it for production environments, and please update frequently in that case
315 2012-05-18 12:23:11 <moa7> i'm just doing git pull and build and that is the version it runs ...
316 2012-05-18 12:23:18 <Diapolo> sipa: are you able to connect?
317 2012-05-18 12:23:34 <sipa> moa7: of course, no problem, but don't keep running that code for 4 months...
318 2012-05-18 12:23:37 <luke-jr> moa7: never run git master in production :p
319 2012-05-18 12:23:43 <Joric> any idea why crome extensions for bitcoind dont work anymore? i tried my own js it works fine
320 2012-05-18 12:24:38 <Joric> what does -debug do will it show requests?
321 2012-05-18 12:24:42 <sipa> trying connection [2002:5898:c338:8000:be05:43ff:febc:fc65] lastseen=0.0hrs
322 2012-05-18 12:24:47 <sipa> Diapolo: ^
323 2012-05-18 12:24:58 <Diapolo> you kicked me ^^
324 2012-05-18 12:25:07 <sipa> ?
325 2012-05-18 12:25:27 <Diapolo> no connection anymore
326 2012-05-18 12:25:43 <sipa> yes, i restarted
327 2012-05-18 12:25:53 <Diapolo> okay re-connected
328 2012-05-18 12:27:35 <Diapolo> seems I currently don't get any blocks
329 2012-05-18 12:27:51 <Joric> did you change authorisation or something since bitcoin-wx? it's really strange, can anyone confirm safebit app works with 0.6.2? https://chrome.google.com/webstore/detail/ablfbfgmgfoggdhhcnledndibdihggpb
330 2012-05-18 12:28:14 <Diapolo> sipa: ah didn't see this connect() failed after select(): Permission denied
331 2012-05-18 12:28:52 <Diapolo> my local system firewall should allow incoming IPv6 i used server=1 in the conf what more needs to be done?
332 2012-05-18 12:29:05 <sipa> server=1 means you want to run an rpc server
333 2012-05-18 12:29:09 <moa7> http://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.6.2/ so source from here is stable?
334 2012-05-18 12:29:14 <sipa> that's entirely independent from the p2p network
335 2012-05-18 12:29:19 <sipa> moa7: yes
336 2012-05-18 12:29:28 <Diapolo> sipa: you are right
337 2012-05-18 12:29:38 <sipa> moa7: but you can use git just find; just run "git checkout v0.6.2" and then build
338 2012-05-18 12:29:39 <moa7> why does it build with weurf version number?
339 2012-05-18 12:29:47 <sipa> weurf?
340 2012-05-18 12:29:51 <moa7> weird
341 2012-05-18 12:30:08 <sipa> you mean the vCLIENT_VERSION_MAJOR.... thing?
342 2012-05-18 12:30:23 <sipa> low priority bug for non-git builds of 0.6.2; fixed in 0.7.0
343 2012-05-18 12:30:24 <moa7> ya
344 2012-05-18 12:30:50 <moa7> kind of confusing if you want to know what version you are running ...
345 2012-05-18 12:30:56 <sipa> i know
346 2012-05-18 12:31:08 <sipa> Joric: don't think anything changed
347 2012-05-18 12:31:38 <Diapolo> ahh I need to add an IPv6 fw rule, didn't know my router allows this ^^
348 2012-05-18 12:32:43 <moa7> so source forge and git checkout v0.6.2 will be identical?
349 2012-05-18 12:33:55 <Joric> sipa, maybe chrome has changed
350 2012-05-18 12:34:48 <Joric> there are a few chrome rpc extensions neither of those works now
351 2012-05-18 12:36:26 <Diapolo> sipa: can you try once more to connect
352 2012-05-18 12:43:50 <sipa> moa7: yes
353 2012-05-18 12:44:16 <sipa> Diapolo: still can't connect
354 2012-05-18 12:44:37 <moa7> except for the "vCLIENT_VERSION_MAJOR.... thing"
355 2012-05-18 12:45:11 <sipa> moa7: that's not really the source that differs, but there's a script that determines at compile time what version you're running, and in 0.6.2 it fails for non-git builds
356 2012-05-18 12:45:14 <sipa> the code is identical
357 2012-05-18 12:46:38 <moa7> aside :anybody made progress on a man page?
358 2012-05-18 12:47:02 <sipa> there is one, but it's outdated and debian-only, afaik
359 2012-05-18 12:47:24 <gavinandresen> TD: reworked the signature cache code to only cache valid signatures, which finesses the whole Optional<T> thing
360 2012-05-18 12:48:23 <moa7> so what does anybody who can't remember all the CLI's ... use the web wiki? ... (curious)
361 2012-05-18 12:49:27 <sipa> moa7: use ./bitcoind help ?
362 2012-05-18 12:49:43 <moa7> is that what you do?
363 2012-05-18 12:49:49 <sipa> no, i check the source code :)
364 2012-05-18 12:50:15 <moa7> right so no the right person to ask ... i mean people who actually use the code not write it
365 2012-05-18 12:52:38 <gavinandresen> sipa: what's up with CSemaphore on the mac?  git HEAD I'm getting sync.h:140: error: Sleep was not declared in this scope
366 2012-05-18 12:55:30 <sipa> gavinandresen: there was a recent issue about that; solution is either move Sleep from util to sync, or pulling #1292 (but i'd like you to verify that doesn't reintroduce the 100% cpu problem on OSX)
367 2012-05-18 12:55:51 <gavinandresen> I'll test 1292
368 2012-05-18 13:03:29 <TD> gavinandresen: yeah i guess that makes sense
369 2012-05-18 13:08:24 <luke-jr> moa7: Debian has a man page
370 2012-05-18 13:09:04 <moa7> luke-jr : okay I'll go install Debian ...
371 2012-05-18 13:09:28 <luke-jr> moa7: Debian's package is also ancient and has known exploits.
372 2012-05-18 13:09:43 <moa7> maybe i won't
373 2012-05-18 13:10:15 <moa7> so i'm really curious people use ./bitcoind help and the web wiki ... ?
374 2012-05-18 13:10:24 <gavinandresen> sipa: 1292 looks good, running -listen=0 I've got 8 connections and minimal CPU usage
375 2012-05-18 13:10:32 <sipa> gavinandresen: ok, great
376 2012-05-18 13:10:46 <luke-jr> moa7: the builtin help is handy
377 2012-05-18 13:12:28 <luke-jr> anyone care to pull #936 ? I can back-out the longpoll commit if that's a cause of hesitation&
378 2012-05-18 13:13:13 <sipa> luke-jr: wasn't there very recent discussion about it?
379 2012-05-18 13:13:29 <luke-jr> sipa: yes, that's resolved
380 2012-05-18 13:14:53 <luke-jr> (also, this is needed for p2pool to fix an exploit I reported)
381 2012-05-18 13:15:08 <gmaxwell> luke-jr: for one thing I was waiting to hear from forrest on it, and get a chance to actually test it myself. (though with the issues last night it might be a few days before I can)
382 2012-05-18 13:17:11 <luke-jr> forrestv: ^ want to re-ACK it? :p
383 2012-05-18 13:18:24 <drizztbsd> luke-jr: do you know why p2pool made that upgrade this morning/night?
384 2012-05-18 13:18:56 <luke-jr> drizztbsd: AIUI, it was a "so many % upgraded" changeover
385 2012-05-18 13:19:20 <gmaxwell> I wonder if bitcointalk and other bitcoin sites should display active alerts.
386 2012-05-18 13:19:26 <drizztbsd> Pool: 380 GH/s. Stale rate: 29.41%. Share difficulty: 597.0673328023785.
387 2012-05-18 13:19:27 <drizztbsd> :P
388 2012-05-18 13:19:51 <gmaxwell> drizztbsd: the stats are goofy from when it was broken.
389 2012-05-18 13:19:56 <luke-jr> drizztbsd: if you're having issues, you're welcome to try out Eligius's decentralized mining :P
390 2012-05-18 13:20:18 <drizztbsd> lol, it's not MY pool. it's the p2pool global stats
391 2012-05-18 13:20:21 <gmaxwell> All the nodes did a concurrent automatic switchover (of a kind we've proposed for bitcoin but never done) and there was a bug.
392 2012-05-18 13:20:50 <luke-jr> gmaxwell: imo, all this means is the obvious "we should testnet it before deploying" :P
393 2012-05-18 13:20:59 <drizztbsd> is eligius decentralized mining made also on python?
394 2012-05-18 13:21:34 <kinlo> drizztbsd: eligius has a central server, luke really should explain the decentralized part of his pool coz I don't get it :)
395 2012-05-18 13:22:07 <gmaxwell> luke-jr: the corner case was only triggered if it went above the final trigger threshold and then back below it.
396 2012-05-18 13:22:11 <luke-jr> drizztbsd: yes, but real Python, not p2pool's Twisted crap
397 2012-05-18 13:22:17 <luke-jr> gmaxwell: testable
398 2012-05-18 13:22:23 <gmaxwell> (and, in fact, I did ask about that case)
399 2012-05-18 13:22:26 <drizztbsd> p2pool does not work on my atom machine :P
400 2012-05-18 13:22:43 <gmaxwell> luke-jr: almost everything is testable, but no testing is perfect.
401 2012-05-18 13:23:05 <luke-jr> gmaxwell: hence why we should learn from p2pool's problem for the future
402 2012-05-18 13:23:14 <luke-jr> drizztbsd: http://eligius.st/wiki/index.php/Getting_Started_with_Decentralized_Mining
403 2012-05-18 13:23:20 <gmaxwell> Yes, thats why I mentioned the p2pool blowup in here.
404 2012-05-18 13:24:00 <gmaxwell> I think we knew that chain coordinated autoswitchvers had more risk of spectacular failure, but it was interesting to see that proved out.
405 2012-05-18 13:24:35 <gmaxwell> if p2pool had more hashpower it would be more viable to test by splitting off a chunk of the pool.
406 2012-05-18 13:29:30 <Eliel> luke-jr: have you written a text anywhere that explains in what ways gmp-proxy improves things over regular pool mining?
407 2012-05-18 13:30:26 <moa7> number of connected nodes still dropping or going up again?
408 2012-05-18 13:32:34 <gmaxwell> moa7: there is no way to measure it
409 2012-05-18 13:32:58 <gmaxwell> moa7: don't believe anyone who claims to
410 2012-05-18 13:33:02 <moa7> i thought there was a guestimate chart floating around somewhere?
411 2012-05-18 13:33:28 <moa7> false memory?
412 2012-05-18 13:33:33 <gmaxwell> moa7: I removed the code that caused nodes who were not listening to announce themselves. So you can't even really guestimate now.
413 2012-05-18 13:33:57 <jgarzik> moa7: here ya go, http://bitcoinstatus.rowit.co.uk/
414 2012-05-18 13:34:25 <gmaxwell> Those estimates are totally worthless. (also, until a few days ago their collector node was broken)
415 2012-05-18 13:35:02 <moa7> seems like maintaining a blockchian is going beyond hobby status so interested to get a feel for it ...
416 2012-05-18 13:35:05 <gmaxwell> (the listening node / block/ version count is okay except for the collector node brokeness I mentioned)
417 2012-05-18 13:36:40 <moa7> sipa : stopped trying to reorganize after 81 failed reconnect length
418 2012-05-18 13:36:58 <gavinandresen> gmaxwell: RE: chain auto-switchovers:  what mechanism did p2pool use?
419 2012-05-18 13:36:59 <sipa> moa7: can you post a debug.log?
420 2012-05-18 13:37:38 <moa7> would it really help that much ... the are quite few other variables
421 2012-05-18 13:37:46 <forrestv> gavinandresen, it used a flag, and when the flag was in 95% of previous shares, it switched
422 2012-05-18 13:38:02 <sipa> moa7: ?
423 2012-05-18 13:38:04 <moa7> i'm going to accept my blockchin is goosed ... i think
424 2012-05-18 13:38:09 <gavinandresen> forrestv: ah.  Have you read my proposal?  https://gist.github.com/2355445
425 2012-05-18 13:38:12 <forrestv> (that part, at least, worked.. the problem was that it switched back when it shouldn't have)
426 2012-05-18 13:38:23 <gmaxwell> moa7: it is, but we'd like to fix the software to recover from goosed chains.
427 2012-05-18 13:38:37 <sipa> moa7: can you stop complaining and just give us the information to check what the problem is? maybe you're not the only one who has this problem?
428 2012-05-18 13:40:16 <moa7> well which debug.log would you like ... i have any number of attempts to recover the chain on different code versions and from different points in the past ...?
429 2012-05-18 13:40:51 <sipa> moa7: 0.6.2, one that has been running for a while
430 2012-05-18 13:41:06 <moa7> define a 'while'
431 2012-05-18 13:41:46 <sipa> an hour
432 2012-05-18 13:42:01 <gmaxwell> The peers.dat code doesn't appear to delete the old addr.dat
433 2012-05-18 13:42:03 <moa7> ok, i can do that, then i'm out.
434 2012-05-18 13:43:11 <forrestv> gavinandresen, looks good... p2pool used some of those ideas: the "flag" was a version field in shares, and clients started warning users when an unsupported version was in 50% of recent shares
435 2012-05-18 13:44:05 <gavinandresen> forrestv: so was the switch-back bug just a plain-old implementation bug, or a problem with the general idea of using count-of-previous-blocks-in-best-chain as the criteria to switch?
436 2012-05-18 13:44:58 <gavinandresen> Or, in other words, is there some block-reorg corner case that you tripped over that I haven't thought of....
437 2012-05-18 13:45:29 <forrestv> implementation bug- there was a rule that said that an old-share can't follow a new-share, but it was only partially enforced
438 2012-05-18 13:45:41 <gmaxwell> ...But instead of a couple of early adopters going 'ouch this hurts' the whole thing came to a screeching halt.
439 2012-05-18 13:45:44 <forrestv> so people mined old-shares when the version count went down, but none of them validated
440 2012-05-18 13:46:41 <gavinandresen> forrestv: ouch....
441 2012-05-18 13:47:03 <gmaxwell> There were also two thresholds in this system one that made the change admissable and the other that made it happen, allowing e.g. if we get stuck under the lower threshold we could manually trigger a switch.
442 2012-05-18 13:47:22 <gmaxwell> (er stuck under the higher one rather)
443 2012-05-18 13:48:03 <forrestv> with a system as large as bitcoin, someone would manually trigger the switch unless there were some other restriction, like a special signature or hashed secret
444 2012-05-18 13:48:26 <gmaxwell> I dunno about that. You couldn't assume they wouldn't at least.
445 2012-05-18 13:55:17 <gmaxwell> gavinandresen: here is forrestv's writeup: http://groups.google.com/group/p2pool-notifications/browse_thread/thread/9d7678823d9931e0
446 2012-05-18 13:55:29 <gavinandresen> gmaxwell: thanks
447 2012-05-18 13:55:43 <forrestv> s/writeup/rushed notice trying to fix it/
448 2012-05-18 13:56:24 <gmaxwell> Also note that it was partially a consequence of a incompletely understood high level programming construct being used in a protocol rule.
449 2012-05-18 14:03:59 <moa7> sipa : debug.log is over 8 Mb
450 2012-05-18 14:04:36 <gmaxwell> moa7: just zip it up and email it if you don't have a place to post it.
451 2012-05-18 14:04:52 <moa7> to who?
452 2012-05-18 14:06:19 <drizztbsd> python--
453 2012-05-18 14:07:53 <moa7> pieter.wuille@gmail.com is sipa?
454 2012-05-18 14:07:55 <gavinandresen> mmmmm, python.....
455 2012-05-18 14:08:05 <sipa> moa7: yes, is good
456 2012-05-18 14:08:22 <moa7> thnx ... i wasn't complaining btw
457 2012-05-18 14:08:36 <sipa> ok :)
458 2012-05-18 14:08:55 <moa7> you guys got a tiger by the tail ... s/ware wise
459 2012-05-18 14:09:03 <moa7> don't let go ...
460 2012-05-18 14:09:50 <drizztbsd> python3 is completely NOT compatibile with python2 :P
461 2012-05-18 14:13:41 <luke-jr> drizztbsd: Eloipool (and gmp-proxy) is 100% Python 3
462 2012-05-18 14:14:16 <drizztbsd> yes, I mean about p2pool and all the twisted shit
463 2012-05-18 14:15:03 <forrestv> https://github.com/forrestv/p2pool/commit/0dbb9da2851108bfe0c7439c769ddeb1bd85d02a would have fixed the problem if noticed earlier
464 2012-05-18 14:20:56 <jgarzik> gmaxwell: yeah, leaving around old addr.dat is intentional
465 2012-05-18 14:21:22 <jgarzik> gmaxwell: less code, it will eventually disappear as ppl upgrade, and it might be nice to have it around if you downgrade
466 2012-05-18 14:23:00 <eian> is libbitcoin still in beta?
467 2012-05-18 14:23:04 <eian> Is it stable yet?
468 2012-05-18 14:23:20 <jgarzik> gmaxwell gavinandresen: speaking of database files...  I am moving txhash.dat, blkhash.dat, blkmeta.dat and blk????.dat into blockchain/ subdirectory (sipa suggestion).  the code will re-index upon first upgrade.  it will pop up a "this will take a LONG time to upgrade db" dialog with InitWarning().
469 2012-05-18 14:24:59 <gmaxwell> jgarzik: are you going to split the db envrioment at the same time (or have you already done this?)
470 2012-05-18 14:26:27 <luke-jr> eian: genjix isn't here most of the time; try #bitcoinconsultancy
471 2012-05-18 14:26:37 <eian> luke, thanks
472 2012-05-18 14:27:54 <gavinandresen> jgarzik: ACK on blockchain/
473 2012-05-18 14:28:14 <sipa> jgarzik: will reindexing mean copying blk0001.dat?
474 2012-05-18 14:28:24 <sipa> or do you move the file, and reindex it on-the-fly
475 2012-05-18 14:29:42 <sipa> gavinandresen: any reason why ExtractAddresses doesn't use ExtractAddress for the single-sig case?
476 2012-05-18 14:30:18 <gavinandresen> sipa: let me look, see if it jogs my memory
477 2012-05-18 14:30:47 <jgarzik> sipa: unfortunately I think the existing code works best if you rewrite blk????.dat along with the rest
478 2012-05-18 14:31:07 <jgarzik> sipa: i.e. LoadExternalBlockFile() calls ProcessBlock(), which ... stores in blk????.dat as well as indices
479 2012-05-18 14:31:18 <jgarzik> sipa: I could reindex, sure, but that won't use the existing code
480 2012-05-18 14:31:37 <sipa> jgarzik: ok
481 2012-05-18 14:31:47 <sipa> jgarzik: fine for now, i guess
482 2012-05-18 14:31:57 <gmaxwell> Just means that we need to have a disk space free guard of 2x blk???.dat.
483 2012-05-18 14:32:01 <sipa> reindexing on-the-fly would certainly be harder
484 2012-05-18 14:32:14 <gmaxwell> what does LoadExternalBlockFile do with multiple block files?
485 2012-05-18 14:32:37 <sipa> gmaxwell: it just processes all blocks found in the file one by one, as if they were received by network
486 2012-05-18 14:32:58 <gavinandresen> sipa: going the other way, having ExtractAddress call ExtractAddresses would probably be the easier refactor
487 2012-05-18 14:33:37 <gavinandresen> sipa: ... and if I recall correctly I didn't do that so backporting would be easier.
488 2012-05-18 14:34:09 <sipa> gavinandresen: i'm splitting CBitcoinAddress into CKeyID, CScriptID, CTxDestination (variant of the former two), and swap the script->base58 dependency around
489 2012-05-18 14:34:39 <gavinandresen> gulp.  Ok....   (CBitcoinAddress is what made backporting a pain)
490 2012-05-18 14:34:47 <sipa> so we don't need base58-depending code to pass addresses around
491 2012-05-18 14:35:54 <luke-jr> sipa: is there a way for users to disable IPv6 binding?
492 2012-05-18 14:36:00 <luke-jr> eg, in the config file
493 2012-05-18 14:36:17 <sipa> luke-jr: -onlynet=ipv4
494 2012-05-18 14:36:55 <luke-jr> sipa: I think my next-test predates that option
495 2012-05-18 14:37:05 <luke-jr> guess I'll need to rebuild to debug this further
496 2012-05-18 14:37:34 <sipa> there used to be -blocknet=... instead of -onlynet, but i'm not sure whether it disables binding (don't think so)
497 2012-05-18 14:39:36 <luke-jr> sipa: yes, it looks like this version just binds IPv6 localhost always
498 2012-05-18 14:40:11 <luke-jr> forrestv: poke
499 2012-05-18 14:41:06 <sipa> luke-jr: oh, is it that old; yes, before -bind was introduced it always just listened on IPv6, which doesn't work on platforms where the v6 and v4 stacks are separated (like winxp)
500 2012-05-18 14:41:40 <forrestv> luke-jr, ?
501 2012-05-18 14:42:09 <luke-jr> forrestv: gmaxwell was hoping you'd comment on/ACK #936 again
502 2012-05-18 14:48:11 <jgarzik> gmaxwell: yes, that means 2x blk????.dat, at least temporarily
503 2012-05-18 14:49:01 <eian> Boy, the academic work around bitcoin is getting strange...
504 2012-05-18 14:49:33 <eian> "Another  idea is to reuse the energy instead of (or in addition to) the computation result. Every computation converts electricity into heat pretty efficiently. Therefore, PoW could very well be used to help heat buildings."
505 2012-05-18 14:49:44 <eian> .....
506 2012-05-18 14:50:08 <eian> I Ddon't think Satoshi saw that coming
507 2012-05-18 14:50:24 <jgarzik> eian: not strange, that is a common data center thought in general.  data centers generate a lot of heat energy
508 2012-05-18 14:51:08 <gavinandresen> I've been saying I want a mining space heater for my office for the last two winters
509 2012-05-18 14:51:18 <eian> hahaha :)
510 2012-05-18 14:51:24 <gavinandresen> maybe next year....
511 2012-05-18 14:51:39 <luke-jr> announcing Eligius space heaters
512 2012-05-18 14:51:48 <jgarzik> 10 years ago, when my main workstation was a DEC Alpha and I was single, I used that Alpha for heating in the winter
513 2012-05-18 14:51:52 <gavinandresen> how many bitcoins?  I'll buy three!
514 2012-05-18 14:51:53 <luke-jr> in collaboration with Sprint, these space heaters are subsidized
515 2012-05-18 14:51:54 <luke-jr> :p
516 2012-05-18 14:52:22 <sipa> boost::variant sure gives rise to nice compiler errors
517 2012-05-18 14:52:36 <gavinandresen> I can see the power companies:  Free Heat!  Just Let Us Install Your New Heating System....
518 2012-05-18 14:52:36 <sipa> /usr/include/boost/variant/variant.hpp:1776:13:   instantiated from static typename Visitor::result_type boost::variant<T0, T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12, T13, T14, T15, T16, T17, T18, T19>::internal_apply_visitor_impl(int, int, Visitor&, VoidPtrCV) [with Visitor = boost::detail::variant::invoke_visitor<const CBitcoinAddressVisitor>, VoidPtrCV = const void*, T0_ = CNoDestination, T1 = CKeyID, T2 = CScriptID, T3 =...
519 2012-05-18 14:52:43 <sipa> boost::detail::variant::void_, T4 = boost::detail::variant::void_, T5 = boost::detail::variant::void_, T6 = boost::detail::variant::void_, T7 = boost::detail::variant::void_, T8 = boost::detail::variant::void_, T9 = boost::detail::variant::void_, T10 = boost::detail::variant::void_, T11 = boost::detail::variant::void_, T12 = boost::detail::variant::void_, T13 = boost::detail::variant::void_, T14 = boost::detail::variant::void_, T15 =...
520 2012-05-18 14:52:47 <t7> what do you want?
521 2012-05-18 14:52:49 <sipa> boost::detail::variant::void_, T16 = boost::detail::variant::void_, T17 = boost::detail::variant::void_, T18 = boost::detail::variant::void_, T19 = boost::detail::variant::void_, typename Visitor::result_type = bool]
522 2012-05-18 14:53:19 <luke-jr> gavinandresen: well, I think it'd be easier to sell standalone heaters that use cellular radios to mine
523 2012-05-18 14:53:58 <gavinandresen> luke-jr: Good Idea.
524 2012-05-18 14:54:10 <luke-jr> the problem is, how do you stop people from opening them up and taking the radeons?
525 2012-05-18 14:54:21 <jgarzik> sipa: that is why C++ sucks ;-)
526 2012-05-18 14:54:42 <sipa> luke-jr: that's why you put a specialized ASIC in it
527 2012-05-18 14:54:53 <luke-jr> sipa: but ASICs don't generate as much heat?
528 2012-05-18 14:54:54 <jgarzik> a single boost backtrace line in gdb often fills half the lines in a terminal window
529 2012-05-18 14:55:01 <sipa> luke-jr: use more ASICs :p
530 2012-05-18 14:55:06 <jgarzik> metametametaprogramming makes logical sense, but...
531 2012-05-18 14:55:09 <luke-jr> sipa: >_<
532 2012-05-18 14:57:42 <forrestv> luke-jr, http://u.forre.st/u/drktbydk/gmp.txt ?
533 2012-05-18 14:57:43 <gavinandresen> lol:  "I spent almost an hour thinking about this code snippet."  https://github.com/gavinandresen/bitcoin-git/commit/4add41a2a6fa1229277e00e01a8b4111e7cb21d6#commitcomment-1348784
534 2012-05-18 14:58:37 <luke-jr> forrestv: doh, forgot to push the fix for that; refetch
535 2012-05-18 15:00:32 <luke-jr> gavinandresen: doing inexpensive checks before expensive ones seems useless to me, as long as it's not any more difficult to pass the inexpensive checks&
536 2012-05-18 15:01:01 <luke-jr> that is, an attacker would just make their CPU-consuming spam pass all the simple checks and then fail the last sig chck
537 2012-05-18 15:01:18 <luke-jr> and it doesn't cost them any more to make
538 2012-05-18 15:01:30 <gavinandresen> luke-jr: failing a signature check gets them banned
539 2012-05-18 15:02:02 <luke-jr> ah
540 2012-05-18 15:02:10 <luke-jr> ok, that makes sense then
541 2012-05-18 15:03:57 <forrestv> luke-jr, why are you using AnyTxToJSON? is any of that data actually useful..?
542 2012-05-18 15:04:30 <luke-jr> forrestv: doesn't hurt, and might be useful to someone?
543 2012-05-18 15:06:13 <forrestv> mm.. it complicates the output quite a bit and is redundant.. someone can just parse the data if they really care
544 2012-05-18 15:06:47 <luke-jr> parsing the data won't tell you about inputs or wallet info
545 2012-05-18 15:08:11 <luke-jr> I don't care strongly about whether it's in there though
546 2012-05-18 15:10:11 <gavinandresen> less is more
547 2012-05-18 15:12:37 <luke-jr> k
548 2012-05-18 15:13:07 <luke-jr> I guess I can submit longpoll later too
549 2012-05-18 15:17:39 <gribble> New news from bitcoinrss: luke-jr opened pull request 1355 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1355>
550 2012-05-18 15:25:09 <helo> mining efficiency increases with distance from equator :)
551 2012-05-18 15:25:22 <JohnSmith78> Can somebody unban me from #bitcoin? I was banned for no reason, period.
552 2012-05-18 15:26:35 <helo> the bans aren't permanent, particularly if you did nothing too extreme
553 2012-05-18 15:27:01 <JohnSmith78> I discussed Bitcoin and compared it to gold.
554 2012-05-18 15:27:13 <JohnSmith78> Then Gmaxwell decided to ban me.
555 2012-05-18 15:27:26 <JohnSmith78> Watch, he's going to come in ban me here too.
556 2012-05-18 15:27:47 <helo> just wait, and it will be lifted
557 2012-05-18 15:28:25 <JohnSmith78> Well thanks for your help.
558 2012-05-18 15:28:45 <helo> he is at least fair enough to give you the opportunity to benefit from learning from your mistakes, don't worry
559 2012-05-18 15:29:02 <JohnSmith78> >implying I made a mistake
560 2012-05-18 15:29:49 <luke-jr> denying mistakes usually precludes learning from them
561 2012-05-18 15:30:12 <jgarzik> gmaxwell sipa: should I delete data/blk0001.dat, after upgrading to data/blockchain/blk0001.dat?
562 2012-05-18 15:30:16 <JohnSmith78> and authority is infallible and divine by the grace of god
563 2012-05-18 15:30:26 <jgarzik> even at huge disk space cost
564 2012-05-18 15:32:31 <luke-jr> [Friday, May 18, 2012] [3:24:46 AM] <gmaxwell>  pacpac: he's been trolling in multiple channels today, earlier he was going on about child porn in -otc.
565 2012-05-18 15:33:27 <nanotube> he exhibited a classic troll pattern of dropping into chans, starting discussions on provocative topics, frequently switching pov's, identities, and topics
566 2012-05-18 15:34:35 <luke-jr> "frequently switching pov's" sounds like playing devil's advocate more than trolling to me
567 2012-05-18 15:34:44 <luke-jr> "&topics" sounds like ADHD :p
568 2012-05-18 15:39:31 <nanotube> "i'm with the fbi, we can put a cap on all bitcoin mining by just raiding everyone" "oh wait i'm not just kidding" "bitcoin is like gold" "oh wait actually it isn't" ... yes it's a fine line, but i'm not going to quibble over the ban. :)
569 2012-05-18 15:41:11 <luke-jr> O.o
570 2012-05-18 15:45:45 <jgarzik> ok, "blockchain/" sub-dir and database upgrade changes pushed out to #blockindex
571 2012-05-18 15:59:46 <jgarzik> sipa: have you tested -loadblocks?
572 2012-05-18 15:59:48 <jgarzik> I get
573 2012-05-18 15:59:50 <jgarzik> EXCEPTION: NSt8ios_base7failureE
574 2012-05-18 15:59:55 <jgarzik> ok a working blk0001.dat file
575 2012-05-18 15:59:57 <jgarzik> *in
576 2012-05-18 16:01:20 <luke-jr> jgarzik: worked for me, mostly
577 2012-05-18 16:01:39 <luke-jr> I had to move it after the wallet loading, due to some local patches
578 2012-05-18 17:17:15 <nadrimajstor_> Hello everyone,
579 2012-05-18 17:18:41 <nadrimajstor_> I tried google but no luck. I'we got bitcoin 0.6.2 win32 running at Win7 64 bit
580 2012-05-18 17:19:21 <nadrimajstor_> From time to time, bitcoin fill up my upload bandwith to the top.
581 2012-05-18 17:20:03 <sipa> jgarzik: i did all my bdb-stresstests using it, for a few months
582 2012-05-18 17:22:13 <sipa> jgarzik: it's -loadblock=, by the way
583 2012-05-18 17:22:56 <gmaxwell> nadrimajstor_: disabling listening.
584 2012-05-18 17:27:21 <nadrimajstor_> The issues is, as my upload rate is only 35kBytes/sec that bitcoin easily fills up, making eg. web browsing really slow.. And it stays filled up for hours.
585 2012-05-18 17:27:35 <sipa> nadrimajstor_: try -nolisten, indeed
586 2012-05-18 17:27:51 <sipa> nadrimajstor_: your problem is probably caused by someone trying to download the blockchain from you
587 2012-05-18 17:28:24 <sipa> as you can't do so at a reasonable speed anyway, it's probably better for everyone to not be reachable either
588 2012-05-18 17:28:56 <jgarzik> sipa: is there something special with the lowest level of scope, within a function or method, WRT LOCK()?  I see at least three examples of:  type MyFunc() { { LOCK(cs_foo); do something; } }
589 2012-05-18 17:29:01 <jgarzik> i.e. extra level of scope
590 2012-05-18 17:29:14 <jgarzik> LoadExternalBlockFile() does this, as do a few other places
591 2012-05-18 17:30:02 <sipa> jgarzik: that was simply to limit the code change necessary
592 2012-05-18 17:31:05 <sipa> when changing CRITICAL_BLOCK(cs) { ... } to { LOCK(cs); ... }
593 2012-05-18 17:31:27 <jgarzik> ok
594 2012-05-18 17:31:44 <sipa> if you modify code that involves such a double scope, feel free to improve it
595 2012-05-18 17:34:51 <gribble> New news from bitcoinrss: luke-jr opened issue 1356 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1356>
596 2012-05-18 17:39:20 <jgarzik> sipa: yeah
597 2012-05-18 17:39:33 <jgarzik> sipa: no justification to touch LoadExternalBlockFile() or the other locations, sadly...
598 2012-05-18 17:41:16 <jgarzik> (fresh d/l)
599 2012-05-18 17:41:27 <sipa> jgarzik: using loadexternalblockfile?
600 2012-05-18 17:45:59 <eian> Is there some place where I can get a history of attacks to peripheral bitcoin services (e.g., exchanges that have collapsed)/
601 2012-05-18 17:49:00 <nadrimajstor_> So far, so good. No internet hogging. For users with just few BTC in wallet, assume they are poor (with a terrible internet connection) and make -nolisten as default option.
602 2012-05-18 17:49:10 <nadrimajstor_> Just kidding. Thank you all for your time.
603 2012-05-18 17:49:40 <jgarzik> sipa: no, network
604 2012-05-18 17:49:54 <jgarzik> sipa: I zapped mine to make sure the new code works both ways
605 2012-05-18 17:50:10 <jgarzik> *zapped my blockchain/* data
606 2012-05-18 17:51:32 <sipa> ok
607 2012-05-18 18:51:01 <gribble> New news from bitcoinrss: luke-jr reopened pull request 1128 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1128>
608 2012-05-18 19:20:50 <jgarzik> interesting watching the initial block download
609 2012-05-18 19:21:02 <jgarzik> behavior is bursty
610 2012-05-18 19:21:29 <jgarzik> it never gets stuck -- but sometimes things get out of order, and we download a bunch of orphan blocks
611 2012-05-18 19:21:50 <jgarzik> the orphans are then correctly processed in a burst
612 2012-05-18 19:23:33 <gmaxwell> jgarzik: happens when a new block comes in for example.
613 2012-05-18 19:26:47 <jgarzik> yep
614 2012-05-18 19:28:17 <wumpus> hmm one of my bitcoin clients (running current git master) is stuck at block 171192, displays the "WARNING: Displayed transactions may not be correct!  You may need to upgrade, or other nodes may need to upgrade." warning.
615 2012-05-18 19:28:35 <Diapolo> sipa: strange thing with your IPv6 node, my Client reports 168000 as block count which is a checkpoint ... it should get the current number from your node, no?
616 2012-05-18 19:29:43 <wumpus> ERROR: Reorganize() : ConnectBlock 00000000000002aa05fa failed
617 2012-05-18 19:30:27 <Diapolo> wumpus: I think someone asked about that one earlier, perhaps its in the IRC logs.
618 2012-05-18 19:32:46 <sipa> wumpus: which version?
619 2012-05-18 19:33:00 <sipa> wumpus: seems like a corrupt blockchain
620 2012-05-18 19:33:17 <wumpus> latest master
621 2012-05-18 19:34:15 <wumpus> lots of errors in the log like: ERROR: FetchInputs() : 0756c2abba mempool Tx prev not found 24bfb27f09
622 2012-05-18 19:34:51 <Diapolo> I get that error very often ... never thought about what it means.
623 2012-05-18 19:35:26 <luke-jr> wumpus: immediatley preceding the Reorg fail?
624 2012-05-18 19:35:31 <jgarzik> you'll get that error, if your blockchain is not up-to-date
625 2012-05-18 19:35:55 <wumpus> I can simply throw away the block chain and start over, but it's still kind of weird
626 2012-05-18 19:36:17 <jgarzik> FetchInputs() is trying to connect today's transactions...  but with a stuck blockchain, you will not have received those TX's dependents (probably)
627 2012-05-18 19:37:24 <Diapolo> sipa: version message: version 60001, blocks=180662 but why is my client not using that number?
628 2012-05-18 19:37:29 <wumpus> luke-jr: nope, it starts with the reorg fail
629 2012-05-18 19:38:02 <Diapolo> wumpus: could tha be a Qt only problem?
630 2012-05-18 19:38:22 <wumpus> this is how it starts: http://www.hastebin.com/hasiqukepi.hs
631 2012-05-18 19:38:25 <wumpus> lol, a qt problem?!
632 2012-05-18 19:38:46 <Diapolo> wumpus: my client keeps DL'ing blocks from Sipas node, but shows no progressbar, because it reads the last checkpoint as block-count, although sipas node reports blocks=180662
633 2012-05-18 19:38:59 <Diapolo> wumpus: sorry was not talking about your stuck issue
634 2012-05-18 19:39:53 <wumpus> Diapolo: you need multiple outgoing connections before it trusts the total block count
635 2012-05-18 19:40:08 <Diapolo> wumpus: alright, didn't know that fact
636 2012-05-18 19:40:28 <Diapolo> wumpus: but it's kind of ugly not to see a progressbar, but the spinner
637 2012-05-18 19:40:31 <wumpus> this is a sanity measure, to prevent one node from lying and messing up the counts
638 2012-05-18 19:40:40 <wumpus> normally that never happens anyway
639 2012-05-18 19:40:48 <jgarzik> heh
640 2012-05-18 19:40:50 <jgarzik> 05/18/12 21:37:41 storing orphan tx 95a22bc215 (mapsz 4686)
641 2012-05-18 19:41:02 <jgarzik> tons of orphans (as expected) during initial block download
642 2012-05-18 19:41:19 <Diapolo> wumpus: playing aroung with IPv6, that's the cause ^^ any idea to get the progressbar back for such situations?
643 2012-05-18 19:41:27 <wumpus> Diapolo: connect to more nodes
644 2012-05-18 19:42:28 <Diapolo> wumpus: no more IPv6 ones are found :-P
645 2012-05-18 19:42:36 <sipa> Diapolo: no need to run IPv6-only
646 2012-05-18 19:42:50 <Diapolo> sipa: only for testing :)
647 2012-05-18 19:43:13 <Diapolo> but okay not a bug and that's fine in the end
648 2012-05-18 19:43:26 <wumpus> no, it's not a bug
649 2012-05-18 19:43:45 <Diapolo> as I said ;)
650 2012-05-18 19:46:03 <Diapolo> sipa: is it okay that I see a few duplicate AddLocal with IPv6 addresses but different weights
651 2012-05-18 19:46:04 <Diapolo> AddLocal([2002:5898:cc5b:0:8d15:c44e:90f3:45a5]:8333,2)
652 2012-05-18 19:46:14 <Diapolo> +same
653 2012-05-18 19:47:16 <wumpus> redownloading the block chain from scratch, I've saved the corrupted one in case someone someday wants to debug it
654 2012-05-18 19:47:35 <luke-jr> wumpus: perhaps it should trust a single node's estimate provided it's reasonable? :p
655 2012-05-18 19:47:38 <Diapolo> sipa: the first is a forced bound and the second one is "discovered" I guess speaking of [2002:5898:cc5b:0:8d15:c44e:90f3:45a5]
656 2012-05-18 19:47:39 <sipa> wumpus: you can try to -loadblock= the old blk0001.dat
657 2012-05-18 19:47:54 <sipa> Diapolo: yes, that's normal if it's discovered through different means
658 2012-05-18 19:47:55 <luke-jr> wumpus: or at least increase the estimate by 1 for every 90 mins we're behind?
659 2012-05-18 19:48:22 <gmaxwell> wumpus> hmm one of my bitcoin clients (running current git master) is stuck at block 171192,  < How long has it been stuck there. e.g. did it get stuck while running prior code?
660 2012-05-18 19:48:30 <wumpus> luke-jr: I suppose so, but that will make the judgement logic more complex, currently it's simply a median filter and that works well for all normal situations
661 2012-05-18 19:48:47 <Diapolo> sipa: this does not harm the networking code that we habe a duplicate address in?
662 2012-05-18 19:48:53 <sipa> Diapolo: no
663 2012-05-18 19:49:00 <Diapolo> okay
664 2012-05-18 19:49:15 <sipa> it's stored in a map from CNetAddr to weight, so it cannot result in a duplicate
665 2012-05-18 19:49:37 <jgarzik> grumble!  1) Fedora's rpm requires db4-utils, preventing use of libdb-utils to inspect db 5.x databases.     2) db_stat from libdb-utils, when force-installed, cannot see any data
666 2012-05-18 19:49:48 <jgarzik> number of keys: 1
667 2012-05-18 19:50:20 <wumpus> gmaxwell: I don't remember it was stuck last time, then again, I run this one very sparsely, which is why it was two months behind
668 2012-05-18 19:53:23 <gmaxwell> luke-jr: the values reported by peers  should probably be clamped to  HighestCheckpoint .. BestGuessBlock.Height+max(0,288*Day-BestGuessBlock.Day) ; where BestGuessBlock is either the highest checkpoint, or the best block if its over the highest checkpoint... then subject to the median filter, and a median of might then be reasonable to allow.
669 2012-05-18 19:54:24 <moa7> sipa : i had three more attempts, two failed at "Postponing 87 reconnects" and one failed at 260ish ...3 short of catching up :(
670 2012-05-18 19:54:44 <moa7> i feel like I'm chasing and can't catch up
671 2012-05-18 19:55:31 <sipa> moa7: your blockchain looks corrupted, from what i see in the debug.log file
672 2012-05-18 19:55:51 <sipa> moa7: it cannot find a transaction that was mined in a block 3 blocks earlier
673 2012-05-18 19:56:21 <moa7> what do you recommend I do at this point I have some btc I want to sell
674 2012-05-18 19:56:25 <moa7> ?
675 2012-05-18 19:56:30 <sipa> delete your blockchain and redownload
676 2012-05-18 19:56:42 <sipa> but since you were running 0.5.99, it's hard to say whether it was the result of buggy old code, a still-existing bug, or a harddisk corruption
677 2012-05-18 19:56:47 <moa7> not gonna happen on my internet connection i'm afraid
678 2012-05-18 19:57:29 <Eliel> moa7: then you might need to work on exporting the private keys to a light weight client and use that.
679 2012-05-18 19:57:31 <sipa> if you're on git head, you can try deleting blkindex.dat and move blk0001.dat to blk0001.dat.old, and then run -loadblock=.../blkindex.dat.old
680 2012-05-18 19:58:21 <Eliel> sipa: that functionality is not activated in the releases?
681 2012-05-18 19:58:34 <wumpus> it wasn't merged yet
682 2012-05-18 19:58:37 <sipa> Eliel: it will be, but there hasn't been a release since it was merged
683 2012-05-18 19:59:17 <Eliel> which release is planned to activate multisig addresses on mainnet?
684 2012-05-18 19:59:29 <sipa> 0.6.1, afaik
685 2012-05-18 19:59:46 <kinlo> but it still isn't useable right?
686 2012-05-18 20:00:09 <sipa> you can send to multisig addresses, and redeem if you have all necessary keys in one wallet
687 2012-05-18 20:00:22 <Eliel> so, 0.6.2 supports it to some extent?
688 2012-05-18 20:00:24 <kinlo> that defeats the purpose :)
689 2012-05-18 20:00:53 <sipa> there just is no consensus/idea/implementation of how negotiation of multisig spendings must happen
690 2012-05-18 20:01:01 <sipa> well, there is BIP10
691 2012-05-18 20:01:02 <moa7> i think "trust noone" is going to have to be dumped ...
692 2012-05-18 20:01:06 <kinlo> ic
693 2012-05-18 20:01:08 <sipa> moa7: ?
694 2012-05-18 20:01:17 <Guest30328> bah
695 2012-05-18 20:01:19 <kinlo> that was my followup question, how it was supposed to happen :p
696 2012-05-18 20:01:56 <moa7> i can get a copy of a good blockchain from someone manually I guess?
697 2012-05-18 20:02:11 <graingert_ecs> moa7: yes
698 2012-05-18 20:02:21 <graingert_ecs> it would be handy to maintain a list of nodes that you control
699 2012-05-18 20:02:26 <graingert_ecs> eg laptop desktop and phone
700 2012-05-18 20:02:38 <sipa> moa7: that takes exactly as much network traffic as downloading it from the bitcoin network...
701 2012-05-18 20:02:58 <graingert_ecs> sipa: but not storage space
702 2012-05-18 20:03:02 <moa7> my test was to build everything myself and do all my own blockchain maintenance but I was vigilante enough about which version I was running ,...
703 2012-05-18 20:03:05 <graingert_ecs> uses much less storage space
704 2012-05-18 20:03:14 <graingert_ecs> if you have one box which does all that
705 2012-05-18 20:03:14 <sipa> graingert_ecs: huh?
706 2012-05-18 20:03:24 <graingert_ecs> and then other clients that trust that box
707 2012-05-18 20:03:31 <graingert_ecs> I'd like to have that setup
708 2012-05-18 20:03:39 <moa7> sipa : i mean met them and get file on USB stick, or similar
709 2012-05-18 20:03:41 <graingert_ecs> eg I just have a box that sits under my desk
710 2012-05-18 20:03:52 <sipa> moa7: try git head and use -loadblocks
711 2012-05-18 20:03:57 <sipa> -loadblock i mean
712 2012-05-18 20:04:03 <graingert_ecs> and then I have my phone and other machines run all the important tx by it
713 2012-05-18 20:04:04 <moa7> sipa : more info
714 2012-05-18 20:04:05 <moa7> ?
715 2012-05-18 20:04:33 <sipa> moa7: -loadblock is a function to import a blockchain from a (potentially even corrupted) blk0001.dat file
716 2012-05-18 20:04:40 <sipa> it's not in 0.6.2 yet though
717 2012-05-18 20:04:49 <moa7> so it is CLI for bitcoind ?
718 2012-05-18 20:04:58 <sipa> it's a command line option yes
719 2012-05-18 20:05:05 <sipa> for bitcoind or bitcoin-qt
720 2012-05-18 20:05:08 <moa7> could you paste the actual command?
721 2012-05-18 20:05:26 <sipa> which OS?
722 2012-05-18 20:05:42 <moa7> "$./bitcoind -rescan -daemon -loadblock"  ?
723 2012-05-18 20:05:46 <moa7> linux
724 2012-05-18 20:05:48 <Eliel> sipa: personally, I'd like an implementation that allows the user interface to be built so the user doesn't need to coordinate the communication at all.
725 2012-05-18 20:06:05 <sipa> Eliel: not sure what you mean
726 2012-05-18 20:06:58 <sipa> moa7: rm ~/.bitcoin/blkindex.dat; mv ~/.bitcoin/blk0001.dat ~/.bitcoin/blk0001.dat.old; ./bitcoind -daemon -loadblock=$HOME/.bitcoin/blk0001.dat.old
727 2012-05-18 20:07:03 <wumpus> personally, I'd like a flying pony
728 2012-05-18 20:07:18 <sipa> wumpus: with a spoiler
729 2012-05-18 20:07:31 <sipa> don't forget about the spoiler
730 2012-05-18 20:07:35 <Eliel> sipa: I mean that one of the keyholders prepares the transaction and hits "send" and as a result, the other keyholders get a message about it asking if they agree. All automatic, no need to coordinate it.
731 2012-05-18 20:07:38 <wumpus> hehe
732 2012-05-18 20:08:02 <moa7> sipa : thnx
733 2012-05-18 20:08:22 <Eliel> if enough of the others agree, the transaction is then broadcast.
734 2012-05-18 20:08:22 <sipa> Eliel: very nice, but quite impossible without a third party doing coordination (who doesn't need any authority over the actual keys, though)
735 2012-05-18 20:08:40 <sipa> but yes, that'd be perfect
736 2012-05-18 20:08:50 <Eliel> sipa: the third party could be an auxiliary p2p network too
737 2012-05-18 20:09:51 <sipa> or a DHT! *ducks*
738 2012-05-18 20:10:41 <kinlo> sipa: I just read bip 10, what are the reasons for not implementing it?
739 2012-05-18 20:11:07 <moa7> sipa : so when you say git head you mean build the alpha stuff again and use that?
740 2012-05-18 20:11:14 <moa7> not git expert
741 2012-05-18 20:11:47 <jgarzik> moa7: the block chain may be downloaded manually from http://eu1.bitcoincharts.com/blockchain/
742 2012-05-18 20:11:52 <Eliel> sipa: good enough for this purpose I think :)
743 2012-05-18 20:12:02 <sipa> moa7: yes
744 2012-05-18 20:12:09 <sipa> kinlo: someone has to do it
745 2012-05-18 20:12:22 <moa7> jgarzik : thnx but my internet would not cope as it is too large now ... :(
746 2012-05-18 20:12:24 <Eliel> sipa: I mean, this auxiliary network could be called bitcoin wallet network :)
747 2012-05-18 20:12:41 <sipa> Eliel: i see no reason to use another P2P network for this
748 2012-05-18 20:12:47 <Eliel> sipa: while the current one is bitcoin node network
749 2012-05-18 20:12:59 <kinlo> sipa: it would require more flexibility on the input transactions I guess
750 2012-05-18 20:13:12 <moa7> jgarzik : also goes outside my trust noone test ...
751 2012-05-18 20:13:15 <kinlo> sipa: you really need to be able to spend one particular transaction
752 2012-05-18 20:13:37 <gmaxwell> sipa: The SpoilerPony should be made of Cake.
753 2012-05-18 20:14:04 <Eliel> sipa: the wallet network would be mostly for communications between users.
754 2012-05-18 20:14:19 <Eliel> not for transactions as such.
755 2012-05-18 20:14:21 <sipa> Eliel: and do all users need to hear about all transactions?
756 2012-05-18 20:14:26 <gmaxwell> kinlo: More flexiblity would be generally desirable but can you explain your use of the word _need_ there?
757 2012-05-18 20:14:29 <sipa> Eliel: that's just not viable
758 2012-05-18 20:14:45 <Eliel> of course not. All the net needs to do is help coordinate it in case both are not online
759 2012-05-18 20:14:54 <Eliel> and if both are online, help them find each other
760 2012-05-18 20:15:10 <sipa> Eliel: you mean TCP/IP?
761 2012-05-18 20:15:30 <gmaxwell> Eliel: ... why do you think that belongs in bitcoin? It sounds like you're describing .. well, IP .. or at worse a gnutella or chat client.
762 2012-05-18 20:15:37 <Eliel> sipa: IP addresses are not what I'd use to identify other wallets.
763 2012-05-18 20:16:26 <kinlo> gmaxwell: don't I need to be able to specify my input transaction here more specifically?  assume I want to use 5 BTC of our shared account I might want to use one input transaction, but if I want to spend 3 myself and 2 from the shared account... I'll need to select that, no?
764 2012-05-18 20:17:29 <Eliel> gmaxwell: simply because I'm not aware of any other method that would be simultaneously both secure and simple to use for this sort of communication and the need as well as use case is there.
765 2012-05-18 20:17:56 <Eliel> gmaxwell: but note, it'd be for wallets, not nodes.
766 2012-05-18 20:18:03 <gmaxwell> kinlo: ah, you're talking about multisig, sorry I was out of sync. I expect all multisig spending will likely be (partially) manual.
767 2012-05-18 20:18:36 <gmaxwell> Eliel: why are you imaging it would be secure by virtue of putting it inside the memory space of a process that handles high value private keys??
768 2012-05-18 20:19:12 <gmaxwell> our current network protocol has basically no security feature because the blockchain consensus service doesn't really need any.
769 2012-05-18 20:19:20 <gmaxwell> features*
770 2012-05-18 20:20:07 <Eliel> gmaxwell: because it could use those keys for addressing and identification as well as verification.
771 2012-05-18 20:20:21 <luke-jr> ugh,  how do you merge .ui files? -.-
772 2012-05-18 20:20:26 <gmaxwell> Thats really really super duper bad bad bad.
773 2012-05-18 20:20:55 <Eliel> gmaxwell: why?
774 2012-05-18 20:21:00 <gmaxwell> Eliel: You totally blow up bitcoin's anonymity by using addresses you spend to for general identification purposes.
775 2012-05-18 20:21:45 <Eliel> gmaxwell: ... you're not imagining making it a social network are you?
776 2012-05-18 20:21:56 <sipa> CoinBook!
777 2012-05-18 20:22:05 <gmaxwell> Eliel: I scan the internet and make a map of Eliel nodes and ask them for their identity for example.  Now I have a nice table of address ownership to IP addresses.
778 2012-05-18 20:22:09 <wumpus> haha, like facebook, but for bitcoins!
779 2012-05-18 20:22:28 <gmaxwell> FacePonyCakeSpoiler
780 2012-05-18 20:22:44 <gmaxwell> (deluxe!)
781 2012-05-18 20:23:33 <Eliel> gmaxwell: I was more thinking using the keys to identify source of messages and only to the recipient of the message.
782 2012-05-18 20:23:36 <gmaxwell> in any case, go build what you want externally to bitcoin and solve all the hard problems and prove it can be usefully done then argue to include it.
783 2012-05-18 20:24:02 <sipa> gmaxwell: Pro
784 2012-05-18 20:24:13 <gmaxwell> Eliel: I don't know of any protocol for doing that that lets you still route.
785 2012-05-18 20:24:40 <sipa> gmaxwell: obviously, broadcast-to-all
786 2012-05-18 20:25:15 <Eliel> sipa: broadcast-to-all is too heavy. there'll be light nodes taking part.
787 2012-05-18 20:25:24 <sipa> Eliel: and that's the problem :)
788 2012-05-18 20:26:05 <gmaxwell> Eliel: I don't think you can have routing and blinding at the same time (or at least I don't know how), and without routing you could only flood which is clearly problematic.
789 2012-05-18 20:26:38 <Eliel> ok, it's rather difficult :)
790 2012-05-18 20:27:04 <gmaxwell> Maybe you could build some kind of flooding network based on using bitcoin mining as a POW to mitigate traffic. Who knows. But all we have at the moment is armwaving.
791 2012-05-18 20:28:23 <gmaxwell> (e.g. nodes could agree to accept messages to flood if they come attached with a bitcoin blockshare of recent height of at least difficulty 0.1 which commits to the hash of the message in the coinbase)
792 2012-05-18 20:28:59 <Diapolo> luke-jr: via text-editor :D
793 2012-05-18 20:29:08 <jgarzik> moa7: there is no need to trust that blockchain
794 2012-05-18 20:29:12 <jgarzik> moa7: bitcoin will validate it
795 2012-05-18 20:29:14 <gmaxwell> But you can build all this with no direct bitcoin integration, just like p2pool requires no direct bitcoin integration. (and I wouldn't mind adding some extra rpc calls to extract any missing data it needed)
796 2012-05-18 20:29:23 <jgarzik> moa7: -loadblock=/PATH/TO/blk0001.dat
797 2012-05-18 20:29:34 <Eliel> it'd still not be scalable enough. This thing would pretty much need to be able to scale to billions of nodes
798 2012-05-18 20:29:39 <wumpus> Diapolo: the annoying thing to merge are tablelayouts, because they contain explicit row/column numbers
799 2012-05-18 20:29:54 <moa7> jgarzik : ok, i see thnx ... so i'd just need to get it delivered physically and then we're all good.
800 2012-05-18 20:30:07 <wumpus> billions, even...
801 2012-05-18 20:30:11 <gmaxwell> Eliel: well, then you need to go earn a fields medal first for your blinded but routable messaging ...
802 2012-05-18 20:30:13 <jgarzik> blkhash.dat/txhash.dat/blkmeta.dat verified via network
803 2012-05-18 20:30:18 <jgarzik> now running upgrade code
804 2012-05-18 20:30:33 <moa7> my problem is i have enough bandwidth, etc resources to maintain a blockchain but not a full node
805 2012-05-18 20:30:41 <jgarzik> LoadExternalBlockFile() works quite well
806 2012-05-18 20:30:56 <moa7> and I dropped the ball keeping the right version because I didn't put enough time into git, etc
807 2012-05-18 20:30:58 <Diapolo> wumpus: forget the last thing with the frame, but I feel really more comfortable with adding code via XML-edit, than via the Qt Designer sometimes.
808 2012-05-18 20:31:00 <gmaxwell> moa7: Can you restate you problem? That didn't make sense to me.
809 2012-05-18 20:31:18 <moa7> gmaxwell : which part?
810 2012-05-18 20:31:32 <gmaxwell> moa7: "i have enough bandwidth, etc resources to maintain a blockchain but not a full node"
811 2012-05-18 20:31:37 <wumpus> Diapolo: yes, especially when moving widgets it's sometimes easier to do by hand.. do check the result in the editor though :-)
812 2012-05-18 20:31:45 <moa7> i have a limited internet connection
813 2012-05-18 20:31:50 <gmaxwell> I don't know what you mean by 'a blockchain but not a full node'
814 2012-05-18 20:31:55 <gmaxwell> moa7: sure. and?
815 2012-05-18 20:32:08 <gmaxwell> The maximum long term rate for the blockchain at the moment is about 14kbit/sec.
816 2012-05-18 20:32:08 <moa7> so i can upload the latest blocks every 3-4 days or so ..
817 2012-05-18 20:32:20 <gmaxwell> the actual rate is more like 2kbit/sec.
818 2012-05-18 20:32:30 <moa7> but i can't leave a client running all the time broadcasting everyone elses transactions
819 2012-05-18 20:32:41 <gmaxwell> Just set -listen=0 and -maxconnections=1
820 2012-05-18 20:32:52 <moa7> but when my blckchain got too far behind it got corrupted
821 2012-05-18 20:33:13 <gmaxwell> moa7: thats not because it was too far behind.
822 2012-05-18 20:33:24 <Eliel> gmaxwell: hmm... something tor based might work... I'm just not too hot on having to bring tor in as a luggage for it to work :P
823 2012-05-18 20:33:25 <Diapolo> wumpus: I found out about a nice option in the Designer, dunno what's it called in english but it resides under Layout -> the last one (easen layout or sth.)
824 2012-05-18 20:33:33 <gmaxwell> Mostly likely it was due to running a pre-release version of 0.6 as a best guess.