1 2014-08-24 00:12:59 <michagogo> jgarzik: ping
  2 2014-08-24 00:13:02 <michagogo> 317000?
  3 2014-08-24 00:13:12 <michagogo> So that's a new file?
  4 2014-08-24 00:13:39 <jgarzik> michagogo, yes
  5 2014-08-24 00:13:45 <michagogo> How come?
  6 2014-08-24 00:13:57 <jgarzik> michagogo, Much time passed since beta
  7 2014-08-24 00:14:10 <michagogo> Also, how much size difference are we talking about?
  8 2014-08-24 00:14:30 <michagogo> (i.e. should I bother linearizing those last blocks, or just update the torrent?
  9 2014-08-24 00:14:31 <michagogo> )
 10 2014-08-24 00:15:13 <gribble> 317193
 11 2014-08-24 00:15:13 <michagogo> ;;blocks
 12 2014-08-24 00:18:06 <BlueMatt> https://projects.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD
 13 2014-08-24 00:18:11 <BlueMatt> fucking --with-incompatible-bdb
 14 2014-08-24 00:19:13 <gmaxwell> BlueMatt: no use 'cussin about it, we knew that they'd do that.
 15 2014-08-24 00:19:18 <Luke-Jr> BlueMatt: I don't see anything at that link. Are they pushing it to end users?
 16 2014-08-24 00:19:27 <BlueMatt> Luke-Jr: yes
 17 2014-08-24 00:19:30 <gmaxwell> it's crappy, but there you have it.
 18 2014-08-24 00:19:34 <Luke-Jr> sigh
 19 2014-08-24 00:19:58 <BlueMatt> are there any oses of note that dont have 5.1 yet?
 20 2014-08-24 00:20:17 <Luke-Jr> 
ACTION doesn't think Arch doing this warrants bumping it

 21 2014-08-24 00:20:30 <BlueMatt> yep, ubuntu
 22 2014-08-24 00:20:36 <Luke-Jr> BlueMatt: note if we bump it, we break downgradability too
 23 2014-08-24 00:20:38 <BlueMatt> Luke-Jr: no, but its always been the intention to bump it when we can
 24 2014-08-24 00:20:55 <BlueMatt> I'd bet arch is far from the only ones doing it
 25 2014-08-24 00:21:12 <Luke-Jr> … no that there's really any excuse at all to be running an old version on Windows
 26 2014-08-24 00:21:13 <Luke-Jr> not*
 27 2014-08-24 00:21:19 <gmaxwell> I am really not very interested in updating it... due to downgradability, and 5.1 seeming to generally be a regression from 4.8. (it appears to crash much more readily on corrupted files, though 4.8 is no star in that respect either)
 28 2014-08-24 00:21:26 <Luke-Jr> (which is frankly the only OS the bdb4.8 should matter for)
 29 2014-08-24 00:21:47 <gmaxwell> BlueMatt: we can't even stay current with bdb anymore due to license compatiblity.
 30 2014-08-24 00:21:49 <Luke-Jr> gmaxwell: true, plus we hope to get rid of it "someday"
 31 2014-08-24 00:22:06 <cfields> wow, codebase builds just fine against musl...
 32 2014-08-24 00:22:10 <BlueMatt> gmaxwell: yea, but most linuxes dont either, so we're safe there
 33 2014-08-24 00:22:15 <cfields> hash anyone worked towards that? or does it actually "just work" ?
 34 2014-08-24 00:22:24 <cfields> -h
 35 2014-08-24 00:22:34 <gmaxwell> BlueMatt: for the moment. But it also means oracle isn't maintaining any of these things anymore anyways.
 36 2014-08-24 00:22:43 <BlueMatt> gmaxwell: true
 37 2014-08-24 00:24:01 <gmaxwell> I assume wallet-ng-format-import-tool will need to link 5.1 anyways for compatiblity with both, upgrading wouldn't be the end of the world, but as you note, there are supported systems that don't have it.. also not eager to make two non-downgradable wallet changes.
 38 2014-08-24 00:24:06 <BlueMatt> but, if we had the option to upgrade (mostly ubuntu) then I would say yes
 39 2014-08-24 00:24:26 <BlueMatt> true
 40 2014-08-24 00:24:38 <BlueMatt> but is there even anyone working on planing wallet-ng-format?
 41 2014-08-24 00:25:30 <sipa> i had a working version ages ago
 42 2014-08-24 00:25:32 <gmaxwell> Not actively, sipa did some design a long time ago.   The other issue we have with upgrading is that we really need wallet tests...
 43 2014-08-24 00:25:37 <gmaxwell> oh more than design then. :P
 44 2014-08-24 00:25:46 <sipa> it missed some error handling etc
 45 2014-08-24 00:25:54 <sipa> and compaction
 46 2014-08-24 00:27:06 <gmaxwell> Did it also end up forcing you to always have the full wallet in memory?
 47 2014-08-24 00:27:36 <sipa> i believe so, yes
 48 2014-08-24 00:27:48 <sipa> but we're doing that already anyway at the application level
 49 2014-08-24 00:28:41 <BlueMatt> yea, but a wallet rewrite shouldnt need to
 50 2014-08-24 00:28:42 <sipa> cfields: i don't think anyone worked on that
 51 2014-08-24 00:29:04 <cfields> hmm, neat
 52 2014-08-24 00:29:22 <sipa> BlueMatt: i really just changed bdb for different (and trivial) append-only checksummed key-value system
 53 2014-08-24 00:29:26 <cfields> it creates a completely static binary, no reliance on runtime glibc
 54 2014-08-24 00:29:35 <cfields> so it's a definite contender for us in the future, imo
 55 2014-08-24 00:30:09 <sipa> BlueMatt: changing what is actually stored in the wallet is a much more invasive change
 56 2014-08-24 00:30:37 <jgarzik> michagogo, not much size difference
 57 2014-08-24 00:30:44 <jgarzik> michagogo, I would just update the torrent
 58 2014-08-24 00:32:20 <gmaxwell> cfields: it still leaves the libstdc++ dependency, so whats the improvement?  statically linking any kind of libc will break domainname resolution on some systems.
 59 2014-08-24 00:33:42 <BlueMatt> sipa: I havent looked at wallet format in forever, what would actually need changed in there?
 60 2014-08-24 00:33:50 <sipa> ?
 61 2014-08-24 00:35:12 <cfields> gmaxwell: from what i've read, libstdc++ can be safely linked statically (or libc++ looking forward). Works for my tests, at least. The dns resolution is a fair point, and i'm not qualified to argue that one.
 62 2014-08-24 00:35:38 <Luke-Jr> BlueMatt: as long as you want address reuse, I think you *have* to keep the wallet in memory..
 63 2014-08-24 00:35:45 <Luke-Jr> of course, that could be dropped :p
 64 2014-08-24 00:36:24 <gmaxwell> cfields: you're right on libstdc++. Name resolution isn't the only thing that potentially breaks w/ static libc though it's one of the most common things.
 65 2014-08-24 00:36:45 <gmaxwell> I'd have to do more research on it myself to figure out how I feel about it... we're not very dependant on name resolution....
 66 2014-08-24 00:37:03 <cfields> indeed
 67 2014-08-24 00:37:26 <Diablo-D3> gmaxwell: did anything get committed to git that improves download performance yet?
 68 2014-08-24 00:37:41 <BlueMatt> Luke-Jr: you might have to keep some parts of wallet around in memory, but that doesnt mean you have to keep everything around
 69 2014-08-24 00:37:49 <gmaxwell> Diablo-D3: nope, because you've failed to test anything and report back. :(
 70 2014-08-24 00:38:01 <gmaxwell> Diablo-D3: what are you doing, we're all waiting on you.
 71 2014-08-24 00:38:13 <Diablo-D3> gmaxwell: lol
 72 2014-08-24 00:38:44 <gmaxwell> Diablo-D3: feel free to test https://github.com/bitcoin/bitcoin/pull/4468 on top of git master and report how it goes... and if it fails be sure to keep the debug logs. :)
 73 2014-08-24 00:39:27 <cfields> gmaxwell: anyway, i certainly wasn't arguing for considering it now. I was just impressed that it actually builds/runs without any patching.
 74 2014-08-24 00:39:45 <Diablo-D3> gmaxwell: yeah I don't like running bitcoin git for some reason =/
 75 2014-08-24 00:39:55 <gmaxwell> Diablo-D3: you can run multiple copies. :)
 76 2014-08-24 00:40:45 <gmaxwell> failure to run bitcoin from git means that bitcoin release will be as bad as you feared bitcoin from git would be. I also agree that running on a serious mainnet wallet is a bad idea... but testnet or a second copy is safe.
 77 2014-08-24 00:41:15 <Diablo-D3> gmaxwell: yeah
 78 2014-08-24 00:41:20 <Diablo-D3> gmaxwell: I used to run git
 79 2014-08-24 00:41:26 <Diablo-D3> but then bitcoin started having actual value
 80 2014-08-24 00:44:11 <stacksmash> I use git with testnet, still using 0.8.6 for real money.
 81 2014-08-24 00:44:43 <Diablo-D3> yeah I should just try it out on testnet
 82 2014-08-24 00:49:35 <Diablo-D3> gmaxwell: btw, I ended up building another box to take over linux duties
 83 2014-08-24 00:49:52 <Diablo-D3> one of those 20th anniversary haswell pentiums
 84 2014-08-24 00:50:22 <Diablo-D3> 16gb of ram, dual crucial m550 128gb caching dual raid1 seagate 2tb enterprise drives I already had laying around
 85 2014-08-24 00:50:29 <Diablo-D3> its goddamned fast for such a tiny box
 86 2014-08-24 00:51:13 <Diablo-D3> idling with spindles off 23w, idling with spindles idle 31w, maximum power usage with gpu off 55w.
 87 2014-08-24 00:51:29 <Diablo-D3> well, maximum Ive seen thus far
 88 2014-08-24 00:51:36 <Diablo-D3> was rsyncing over the network when I did that
 89 2014-08-24 00:52:33 <SomeoneWeird> Diablo-D3: "i already had laying around"
 90 2014-08-24 00:52:35 <SomeoneWeird> damn you :p
 91 2014-08-24 00:53:17 <Diablo-D3> my last workstation was 2tb raid1
 92 2014-08-24 00:53:44 <Diablo-D3> my new one is crucial m500 960gb raid1
 93 2014-08-24 00:56:54 <SomeoneWeird> Diablo-D3: that is a decently sized ssd :)
 94 2014-08-24 00:57:55 <Diablo-D3> I was building a totally silent box
 95 2014-08-24 00:58:05 <Diablo-D3> the hard part was making the 7970 silent
 96 2014-08-24 00:58:30 <Diablo-D3> turning up the fans on everything I can just _barely_ hear it
 97 2014-08-24 00:59:34 <Diablo-D3> 2x noctua 120m, 2x noctua 140mm, noctua nh-d14, and artic accelero xtreme iv 280x, all in a corsair obsidian 550d
 98 2014-08-24 01:00:41 <Luke-Jr> hm
 99 2014-08-24 01:00:51 <Diablo-D3> the noctua fans are all plugged in with the undervolt cables, and then the pwmable ones (nh-d14 has 120mm and 140m of its own, but cant be pwmed) at 15% speed, and the accelero's 3 fans at 15%, and the psu's fan halts at low enough wattage
100 2014-08-24 01:00:58 <Luke-Jr> 
ACTION wonders what behaviour exactly BCCore manifests not having undo data for a reorg

101 2014-08-24 01:01:23 <stacksmash> Nice, I have Accelero fans on my cards but my one card with the stock blower fan is louder than all the others combined.
102 2014-08-24 01:01:32 <Diablo-D3> so thats 6 noctua fans, 3 gpu fans, and a psu fan
103 2014-08-24 01:01:36 <Diablo-D3> 10 fucking fans.
104 2014-08-24 01:01:48 <Diablo-D3> and its dead silent.
105 2014-08-24 01:02:04 <Diablo-D3> its more silent than any water cooling rig I've met
106 2014-08-24 01:02:15 <Diablo-D3> pumps are always too fucking loud
107 2014-08-24 01:03:30 <stacksmash> I think I had 15 fans in my old mining rig. It was definitely not silent.
108 2014-08-24 01:03:54 <Diablo-D3> 2 120mm in the front, 1 140mm in the floor blowing right into the gpu, 1 140mm behind the nh-d14 for exhaust
109 2014-08-24 01:04:08 <Diablo-D3> and the 550d is a specially silenced case
110 2014-08-24 01:04:53 <Diablo-D3> my new box is more conventional
111 2014-08-24 01:06:19 <stacksmash> At some point I'll convert my old equipment into a gaming rig or something. Right now it's doing nothing.
112 2014-08-24 01:06:21 <Diablo-D3> 53w tdp haswell pentium, 2 ssds, 2 hdds, fanless psu with extremely high efficiency at the 20-30w range (seasonic ss-400fl2, 400w, fanless, 80plus platinum, highest efficiency at 20-30w Ive seen of any psu)
113 2014-08-24 01:07:01 <Diablo-D3> noctua nh-u9b (think of it as a baby nh-d14 without the split block, uses dual 92mm non-pwm fans, also with the undervolt cables)
114 2014-08-24 01:07:17 <Diablo-D3> and then I left the corsair 140mm fan in the front of the case moved dowwn infront of the harddrives
115 2014-08-24 01:07:22 <Diablo-D3> and its a corsair 350d
116 2014-08-24 01:07:28 <Diablo-D3> same series as the 550d, but not specially silenced
117 2014-08-24 01:07:33 <BlueMatt> 
ACTION wonders if he's in ##hardware

118 2014-08-24 01:07:46 <Diablo-D3> BlueMatt: hardware fetishism is always on topic ;)
119 2014-08-24 01:10:21 <stacksmash> My 5870 is still secured to my tower with zip ties.
120 2014-08-24 02:07:51 <earlz> does bitcoin master support headers only now?
121 2014-08-24 02:08:01 <earlz> (ie, super fast blockchain downloads :) )
122 2014-08-24 02:09:38 <fanquake> earlz see https://github.com/bitcoin/bitcoin/pull/4468
123 2014-08-24 02:12:14 <earlz> so exciting :)
124 2014-08-24 02:12:26 <SomeoneWeird> (so no, not yet)
125 2014-08-24 02:13:02 <earlz> I saw some variables referencing headersOnly recently, but i guess that's just preperation
126 2014-08-24 02:13:52 <earlz> headers only means no orphan blocks? that's interesting
127 2014-08-24 02:14:29 <earlz> seems kind of like moving to a concensus form of tracking blocks
128 2014-08-24 02:17:06 <earlz> seems like it'd make a wallet getting stuck much more likely though
129 2014-08-24 02:20:30 <icanon> whois abrkn
130 2014-08-24 02:43:17 <jgarzik> PSA: Today's torrent will be the last bootstrap.dat torrent.  Future torrents will split the block chain into files no larger than 1GB (and often smaller).  https://bitcointalk.org/index.php?topic=145386.msg8507165#msg8507165
131 2014-08-24 02:43:44 <jgarzik> IMO, bootstrap.dat itself may be deprecated, but that's open for debate...
132 2014-08-24 02:48:31 <dmj`> where do I go to see the rise and rise of bitcoin documentary?
133 2014-08-24 02:49:49 <Luke-Jr> dmj`: off-topic and afaik does not exist. #bitcoin to continue this conversation
134 2014-08-24 02:50:19 <dmj`> Luke-Jr: I asked there, either I'm muted or no one replied
135 2014-08-24 02:50:46 <dmj`> Luke-Jr: of course it exists: http://theriseandriseofbitcoin.com/
136 2014-08-24 02:53:38 <Luke-Jr> dmj`: again, this is off-topic here. just because nobody wants to discuss it does not mean it's acceptable to bring it to the wrong venue.
137 2014-08-24 02:58:38 <dmj`> Luke-Jr: please stop telling me what to do
138 2014-08-24 03:00:51 <jgarzik> michagogo, Testing the new format torrent now, release in another 3000 blocks or so, perhaps
139 2014-08-24 03:01:25 <jgarzik> michagogo, https://github.com/bitcoin/bitcoin/pull/4757 generates old bootstrap.dat or new format
140 2014-08-24 03:04:56 <BlueMatt> jgarzik: I think headers first may deprecate it
141 2014-08-24 03:07:24 <BlueMatt> (hopefully
142 2014-08-24 03:07:25 <BlueMatt> )
143 2014-08-24 03:24:04 <jgarzik> BlueMatt, Ideally the torrent will be relegated to research purposes and infrequent uses once headers-first lands
144 2014-08-24 03:24:14 <jgarzik> but who knows, once archive nodes are deployed
145 2014-08-24 03:24:21 <BlueMatt> research purposes?
146 2014-08-24 03:27:37 <jgarzik> BlueMatt, if you need the dataset, easier to just download the block data directly, than run a full node and have it indirectly perform the same function
147 2014-08-24 03:27:42 <phantomcircuit> BlueMatt, it's well formatted blockchain data basically
148 2014-08-24 03:27:59 <phantomcircuit> jgarzik, not to mention it's not full of orphans and/or garbage
149 2014-08-24 03:28:04 <jgarzik> you know the blocks are in-order, with no orphans
150 2014-08-24 03:28:06 <BlueMatt> jgarzik: true
151 2014-08-24 03:28:14 <BlueMatt> mostly in-order w/o orphans
152 2014-08-24 04:56:00 <jgarzik> 2014-08-24 04:55:21 ERROR: SignatureHash() : nOut=1 out of range
153 2014-08-24 04:56:00 <jgarzik> 2014-08-24 04:55:21 ERROR: SignatureHash() : nOut=2 out of range
154 2014-08-24 04:56:00 <jgarzik> 2014-08-24 04:55:21 UpdateTip: new best=0000000000000000070337cf17903810498b47e9c7b077920214d17ea7f0ae7f  height=313937  log2_work=80.035541  tx=43818183  date=2014-08-04 11:01:22 progress=0.924860
155 2014-08-24 04:56:02 <jgarzik> 2014-08-24 04:55:21 ERROR: SignatureHash() : nOut=1 out of range
156 2014-08-24 04:56:03 <jgarzik> 2014-08-24 04:55:21 ERROR: SignatureHash() : nOut=1 out of range
157 2014-08-24 04:56:06 <jgarzik> 2014-08-24 04:55:22 UpdateTip: new best=00000000000000001004f5fdaa079dd8ca2c95c077fb2d87494ffb9de234cff1  height=313938  log2_work=80.035635  tx=43819630  date=2014-08-04 11:37:39 progress=0.924955
158 2014-08-24 04:56:28 <jgarzik> _hopefully_ those out-of-range are referring to garbage incoming from network, and not garbage in my blk*.dat
159 2014-08-24 05:25:25 <phantomcircuit> jgarzik, probably should change the code to specify....
160 2014-08-24 08:34:31 <dionyziz> Hello.
161 2014-08-24 08:34:57 <dionyziz> Can I get some bitcoin testnet coins? 0.5BTC would be fine. Thanks! mozULF1a2qPS1cWXYNk2tie8XGWo4r1PhV
162 2014-08-24 08:44:27 <justanotheruser> dionyziz: 5f1b14f02e6a2e02456a4f61f618a6a410df60908a228b71935361b20279ac30
163 2014-08-24 08:48:52 <dionyziz> justanotheruser: Thanks :)
164 2014-08-24 09:09:59 <BlueMatt> hmm, is there a reason syncing from a node running master should be slower than syncing from a node running 0.9.2.1?
165 2014-08-24 09:10:27 <gmaxwell> none that I'm aware of, seems unlikely to me.
166 2014-08-24 09:21:11 <midnightmagic> comments on whether pullreq based on: https://github.com/midnightmagic/bitcoin/commit/244915c31d7a563a64b8519fae93fc05e3e3e837#diff-d41d8cd98f00b204e9800998ecf8427e or a pullreq based on: https://github.com/midnightmagic/bitcoin/commit/f2c050781f8dbf497d988113d9740610de6150f2#diff-d41d8cd98f00b204e9800998ecf8427e would be more successful? the pkgsrc people are requesting a merge upstream to eliminate the dangers of late merge
167 2014-08-24 09:21:17 <midnightmagic> fails.
168 2014-08-24 09:24:35 <gmaxwell> midnightmagic: first hunk is incorrect— I think, that should be accomplished by adjusting autotools to detect and set the right include path.
169 2014-08-24 09:25:26 <gmaxwell> (which it does just fine on my fedora system— it's in /usr/include/libdb4/db_cxx.h )
170 2014-08-24 09:26:21 <gmaxwell> No opinion on the leveldb part, that should go upstream to leveldb ultimately so it should be what they'd want.
171 2014-08-24 09:26:27 <gmaxwell> Netbase change looks okay to me.
172 2014-08-24 09:27:16 <midnightmagic> case in point then. these were the straight pkgsrc patch files. I can correct that in the pkgsrc wip entry easy.
173 2014-08-24 09:27:18 <gmaxwell> Why is the walletdb change needed? A boost bug on netbsd or?
174 2014-08-24 09:27:53 <midnightmagic> there's a build issue, this one here: http://www.robertnitsch.de/notes/cpp/cpp11_boost_filesystem_undefined_reference_copy_file
175 2014-08-24 09:29:13 <midnightmagic> in exploring it, it appears it's not immediately easy to get boost homogeneously building with bitcoin (and others) with/without the -std=c++0x flag
176 2014-08-24 09:30:46 <midnightmagic> so, rather than risk a minor makefile failure messing it up in the future, i thought maybe easier to avoid it altogether, since it only affects copy_file() and not copy()..
177 2014-08-24 09:30:58 <midnightmagic> I'm not attached to any of those changes btw.
178 2014-08-24 09:33:11 <midnightmagic> tangential irrelevance: it's super irritating that ldd is hardwired to verbose mode when resolving symbol versions, while still retaining all the programmatic logic to support suppression of associated errors
179 2014-08-24 09:35:16 <BlueMatt> naa, found my bug
180 2014-08-24 09:36:23 <kdomanski__> midnightmagic: that issue is resolved with #define BOOST_NO_CXX11_SCOPED_ENUMS
181 2014-08-24 09:37:20 <kdomanski__> ah, right, the page you linked says so
182 2014-08-24 09:37:51 <midnightmagic> kdomanski__: one would think so. the error message of a missing copy_file() def with enums persists after patching even right down to the included boost headers.
183 2014-08-24 09:38:41 <Diablo-D3> hey guys
184 2014-08-24 09:39:03 <Diablo-D3> did any of you ever come across a pci-e flexible riser that was advertised as 3.0 compatible?
185 2014-08-24 09:39:34 <midnightmagic> buildlink3 dumps boost headers into the temp working (build) tree and setting even the superceding define does not eliminate the problem; is there a hard necessity for the use of copy_file rather than copy?
186 2014-08-24 09:40:01 <kdomanski__> midnightmagic: what boost version?
187 2014-08-24 09:40:15 <gmaxwell> midnightmagic: are you putting in the define ahead of including the boost headers?
188 2014-08-24 09:40:26 <midnightmagic> kdomanski__: exact version is 1.55.0nb1
189 2014-08-24 09:40:40 <midnightmagic> packages required for building are boost-headers-1.55.0nb1 and boost-libs-1.55.0nb1
190 2014-08-24 09:42:13 <midnightmagic> gmaxwell: I did, yes. And then, irritated, I manually modified the boost headers themselves, and then manually edited the proto to eliminate that line itself. The last attempt worked. I didn't trace with a cpp raw output yet or exhaust all possibilities in the interest of getting a functional build going faster.
191 2014-08-24 09:46:13 <midnightmagic> also tangentially, pkgsrc lists vulns for qt-4.8.x that have been open and unresolved for .. like two years or something, and a number of unresolved qt-5.2.x recent vulns which are also unresolved. So building full bitcoind-qt under the pkgsrc framework requires manual intervention to explicitly ignore those URLs, or shut off pkgsrc vuln checking altogether.
192 2014-08-24 09:56:14 <midnightmagic> gmaxwell: possibility exists I made a mistake of course.
193 2014-08-24 09:58:00 <sipa> jgarzik: that return 1 in case of out of range with sighash_one is significant and consensus critical
194 2014-08-24 10:01:34 <Diablo-D3> I think I answered my own question
195 2014-08-24 10:01:35 <Diablo-D3> http://solutions.3m.com/wps/portal/3M/en_US/Interconnect/Home/Products/ProductCatalog/Catalog/?PC_Z7_RJH9U5230O73D0ISNF9B3C3SI1000000_nid=TMHBQ9MBNLitJ249BH4P7KglS6SKK96VN3bl
196 2014-08-24 10:01:40 <Diablo-D3> I can get that for $59
197 2014-08-24 11:32:26 <elichai2> petertodd, here?
198 2014-08-24 14:06:38 <jgarzik> sipa, not sure what you're referring to
199 2014-08-24 16:15:51 <alfacent> hi