1 2015-05-09 03:24:57 <hulkhogan_> sdaftuar: here ?
 2 2015-05-09 07:09:03 <gmaxwell> oh realized I didn't mention it here; I attempted to reproduce matt's (well the ubuntu builder) arm crash in test_bitcoin with 0.10.1; and I cannot reproduce it on my novena.
 3 2015-05-09 07:09:15 <gmaxwell> (I told matt earlier)
 4 2015-05-09 07:38:21 <hulkhogan_> how are you running a full node on arm? just curious..
 5 2015-05-09 07:38:57 <gmaxwell> hulkhogan_: how? it's just a computer! there is nothing magic about it. What would you have expected to be the issue?
 6 2015-05-09 07:39:38 <wumpus> quadcore ARM boxes can run a full node fine
 7 2015-05-09 07:40:10 <gmaxwell> I have bitcoin core running on an odroid u2 (though its not reliable due to thermal issues) and a novena board; both run linux.  (I've also previously run it on a TI pandaboard; though those are a bit resource constrained.); Other people have it running on rpi though thats a bit nuts, and rpi2 which is more reasonable.
 8 2015-05-09 07:41:45 <hulkhogan_> hah, i just meant - at least for my arm device, i haven't thought about how to store the blockchain
 9 2015-05-09 07:42:01 <hulkhogan_> i suppose i could fork out some for a microsd..
10 2015-05-09 07:42:18 <wumpus> cheap USB hardisk  is best
11 2015-05-09 07:43:04 <hulkhogan_> i was just checking the buildlog from earlier but couldnt find what about the install was failing it
12 2015-05-09 07:43:13 <wumpus> (you can actually use USB stick too, but those things tend to be unreliable for the sustained throughput, expensive at that size, and expected to wear out more quickly than normal usage)
13 2015-05-09 07:43:29 <hulkhogan_> usb harddisk ?
14 2015-05-09 07:43:55 <hulkhogan_> i dont know what those are actually
15 2015-05-09 07:44:23 <hulkhogan_> oh i guess an enclosed drive
16 2015-05-09 07:44:28 <wumpus> an external harddisk (enclosure) that is connected to USB? those are bog-standard
17 2015-05-09 07:44:33 <hulkhogan_> yea
18 2015-05-09 07:45:04 <hulkhogan_> should be non-ssd for long term use eh
19 2015-05-09 07:45:19 <wumpus> there is a lot of price stunting with them as normal consumers buy them for backup purposes
20 2015-05-09 07:45:22 <hulkhogan_> [to prevent wear-outs]
21 2015-05-09 07:45:23 <hulkhogan_> [to prevent wear-outs]
22 2015-05-09 07:45:56 <hulkhogan_> very interesting purchase behavior
23 2015-05-09 07:46:39 <hulkhogan_> its easy to throw up buzzwords to pull off sales
24 2015-05-09 07:46:40 <hulkhogan_> its easy to throw up buzzwords to pull off sales
25 2015-05-09 07:46:50 <midnightmagic> gmaxwell: aw damn, the odroid overheats?!
26 2015-05-09 07:48:36 <wumpus> syncing the bitcoin block chain is a great burn-in test for thermal design of a board :)
27 2015-05-09 07:49:07 <midnightmagic> darnit. :(
28 2015-05-09 07:50:49 <hulkhogan_> seems like its just a segfault in one of the tests
29 2015-05-09 07:50:58 <hulkhogan_> instead of a huge fire
30 2015-05-09 07:51:18 <wumpus> (which is non-trivial for high-end ARM boards; iMX6 is hardly ever run at full 1.2GHz but tends to be downclocked to 1.0GHz)
31 2015-05-09 07:51:56 <wumpus> yes it's a figurative term, if a burn-in test results in a huge fire, there is something wrong with the thermal design of your working environment...
32 2015-05-09 07:52:06 <hulkhogan_> haha
33 2015-05-09 07:52:13 <hulkhogan_> probably true
34 2015-05-09 07:54:43 <gmaxwell> my odroid has a 64gb emmc on it; with the novena I use an external hdd (the novena has a SATA interface too.)
35 2015-05-09 07:55:00 <hulkhogan_> running a full node would be cool on a rpi or equivalent
36 2015-05-09 07:55:20 <wumpus> yes, SATA is even better if possible
37 2015-05-09 07:55:25 <hulkhogan_> a question, does it take up the entire cpu ?
38 2015-05-09 07:55:30 <midnightmagic> my earlier-version rpi overheat and spontaneously reboot themselves. or .. whatever is the cause. I'm guessing it's a heating issue.
39 2015-05-09 07:55:32 <gmaxwell> midnightmagic: yea, I think its fine if there is a fan on it; but without it reliably suffers memory corruption once signature validation cranks up. 100% repeatable (though it fails at different spots)
40 2015-05-09 07:55:51 <wumpus> hulkhogan_: that is really a #bitcoin question
41 2015-05-09 07:56:20 <gmaxwell> hulkhogan_: syncing takes as much as you'll give it-- of course, to finish asap.  Once its synced the usage is pretty modest.  Rpi is really on the edge of reasonable hardware for it; rpi is especially underpowered even for an arm board.
42 2015-05-09 07:56:26 <gmaxwell> oops didn't notice the channel! :)
43 2015-05-09 07:56:41 <wumpus> gmaxwell: me neither.
44 2015-05-09 07:58:31 <hulkhogan_> ah thanks gmaxwell, i might try that soon then !
45 2015-05-09 08:16:56 <hulkhogan_> ack, i just realized you said you /cannot/ reproduce the test_bitcoin error from that logfile earlier
46 2015-05-09 08:19:18 <gmaxwell> cannot. yea, doesn't crash for me; may be compiler/libc/etc. specific.
47 2015-05-09 09:44:03 <btcdrak> Does anyone have the link to the site which was measuring the rollout of BIP66?
48 2015-05-09 09:47:42 <gmaxwell> btcdrak: http://bitcoin.sipa.be/ver-10k.png
49 2015-05-09 09:48:47 <btcdrak> thanks
50 2015-05-09 10:40:10 <jgarzik> gmaxwell, sipa: gah, the nuttery of EU date formats
51 2015-05-09 10:41:56 <gmaxwell> gnuplot date settings are weird. and yea.. that bugs me about sipa's graphs.
52 2015-05-09 10:42:12 <gmaxwell> sipa: iso8601 plz.
53 2015-05-09 10:43:06 <jgarzik> any computer science nerd should recognize the superiority of year/month/day (abbrev. month/day).  The US and EU both get bits of that wrong ;p
54 2015-05-09 10:51:12 <gmaxwell> I think tonal dates would probably be an improvement over the dates on sipas graphs. :P
55 2015-05-09 11:15:25 <sturles> At least the numbers are in a sensible order.  Not mixed like US formats.  9.11.01 - how are we supposed to know which number is the most significant and the least significant?  Placing the day in the middle between month and year is even more nuts, imho.
56 2015-05-09 11:16:19 <sturles> Byt, yes, YYYYMMDD is much prefered and very often used.
57 2015-05-09 11:48:28 <Diablo-D3> sturles: most people use yyyymmdd just so it sorts correctly.
58 2015-05-09 11:50:38 <buZz> best reason
59 2015-05-09 12:14:33 <gavinandresen> I always say 11 Oct 2011 to be perfectly clear.
60 2015-05-09 13:51:43 <jtimon> please help libconsensus move forward by reviewing #6051
61 2015-05-09 13:51:43 <jtimon> please help the policy encapsulation move forward reviewing #6068
62 2015-05-09 17:58:19 <hulkhogan_> morn
63 2015-05-09 17:59:04 <hulkhogan_> will look at that a bit latter jtimon
64 2015-05-09 20:16:36 <michagogo> !j
65 2015-05-09 20:16:37 <michagogo> !j
66 2015-05-09 20:16:38 <gribble> Error: "j" is not a valid command.
67 2015-05-09 20:19:00 <michagogo> ...wrong channel
68 2015-05-09 20:19:09 <michagogo> 
ACTION slaps IRCCloud's iPhone app around a bit with a large cow shark