1 2013-10-28 00:00:42 <sipa> BlueMatt: is there nothing like memenv for that b2?
2 2013-10-28 00:00:42 <sipa> eh, h2
3 2013-10-28 00:01:39 <sipa> BlueMatt: so, what is holding that pullreq back? the fact that there is no tmpfs on the pulltester vm?
4 2013-10-28 00:02:00 <BlueMatt> probably, havent looked recently
5 2013-10-28 00:02:05 <BlueMatt> yep
6 2013-10-28 00:02:31 <BlueMatt> feel free to login and fix pull-tester.py
7 2013-10-28 00:03:24 <sipa> or /etc/fstab?
8 2013-10-28 00:03:26 <BlueMatt> not sure, try it
9 2013-10-28 00:03:33 <BlueMatt> may break some rm -r or so
10 2013-10-28 00:05:45 <sipa> AAAARGH
11 2013-10-28 00:05:57 <sipa> i've been testing with ./bitcoin-qt all the time
12 2013-10-28 00:06:03 <sipa> instead of ./src/qt/bitcoin-qt
13 2013-10-28 00:06:11 <BlueMatt> hah nice
14 2013-10-28 00:06:36 <BlueMatt> this is why you git clean -f -x -d when switching to/from 0.8
15 2013-10-28 00:06:48 <sipa> 0.8?
16 2013-10-28 00:06:54 <sipa> this is just git head
17 2013-10-28 00:07:07 <BlueMatt> and ./bitcoin-qt is not
18 2013-10-28 00:07:08 <sipa> and i have way too many non-checked in files that are useful to run a command like that
19 2013-10-28 00:07:10 <BlueMatt> ie its from 0.8
20 2013-10-28 00:07:22 <BlueMatt> ahh, well keep those up a level
21 2013-10-28 00:07:35 <sipa> it's from a headersfirst build, apparently
22 2013-10-28 00:07:45 <sipa> long after 0.8.99
23 2013-10-28 00:07:49 <sipa> but before autotools
24 2013-10-28 00:09:48 <BlueMatt> s#to/from 0.8#to/from autotools#
25 2013-10-28 00:09:49 <BlueMatt> happy?
26 2013-10-28 00:11:34 <sipa> yup!
27 2013-10-28 00:11:38 <sipa> also, your latest comparison script on tmpfs seems to work nicely
28 2013-10-28 00:12:10 <warren> ACTION has watchonly working on 0.8.5
29 2013-10-28 00:13:31 <sipa> warren: cool
30 2013-10-28 00:13:34 <BlueMatt> sipa: yea...finally
31 2013-10-28 00:13:59 <warren> sipa: I'm just making a bitcoin-0.8.5 tree with all the same backports as in litecoin-0.8.5.x
32 2013-10-28 00:14:33 <warren> I should stuff leveldb 1.13 in here too
33 2013-10-28 00:15:06 <warren> sipa: is there a syntax to update and resquash an existing subtree?
34 2013-10-28 00:16:04 <sipa> git subtree pull, with some extra arguments
35 2013-10-28 00:16:38 <sipa> -P src/leveldb in any case
36 2013-10-28 00:16:51 <sipa> and then the name of a branch in the leveldb repo
37 2013-10-28 00:17:46 <sipa> there's a leveldb 1.14 btw
38 2013-10-28 00:17:57 <sipa> but no changes that are relevant for us, i think
39 2013-10-28 00:18:16 <warren> I rather test what will be in 0.9
40 2013-10-28 00:18:21 <warren> unless you're putting 1.14 into 0.9
41 2013-10-28 00:18:32 <sipa> i'm not inclined to now
42 2013-10-28 00:18:51 <sipa> so do a git remote add of the git@github.com/bitcoin/leveldb repo
43 2013-10-28 00:19:13 <sipa> and then git subtree pull --squash -P src/leveldb leveldb/master
44 2013-10-28 00:19:21 <sipa> that should do it, iirc
45 2013-10-28 00:19:26 <warren> ah, thanks
46 2013-10-28 00:19:27 <sipa> but haven't tried myself now
47 2013-10-28 00:19:44 <warren> I'll ship both litecoin and bitcoin-0.8.5-awesomestuff with that
48 2013-10-28 02:22:10 <skinnkavaj> https://bitcointalk.org/index.php?topic=319688.0
49 2013-10-28 02:22:19 <skinnkavaj> Bitcoin's Killer App = High Speed Anonymous Internet (TOR like)
50 2013-10-28 02:39:07 <warren> sigh. my laptop can't survive syncing the bitcoin block chain
51 2013-10-28 02:39:10 <warren> overheat kills itself
52 2013-10-28 02:39:52 <gmaxwell> ACTION rages at the world
53 2013-10-28 02:40:11 <gmaxwell> It's really bullshit that we should be asked to deal with crap like that (I know you're not...)
54 2013-10-28 02:40:53 <propumpkin> crap like what?
55 2013-10-28 02:41:00 <warren> that
56 2013-10-28 02:41:06 <gmaxwell> propumpkin: Hardware that fails if you use it.
57 2013-10-28 02:41:08 <propumpkin> oh
58 2013-10-28 02:42:01 <gmaxwell> 99% of random desktops users spend 99% of their time idle. Their machines crash randomly anyways.. so it's commercially viable to sell hardware that can't actually take being used.
59 2013-10-28 02:42:51 <gmaxwell> then joe-random-user, understandably since it doesn't frequently otherwiseâ or when it does its often because some program broke and busylooped, blames bitcoin when his machine overheats.
60 2013-10-28 02:43:24 <warren> this is a Thinkpad T410s. It never could handle running both cores at full speed. Once to render a video I held the laptop up to a 10,000 BTU AC for an hour. The temperature hovered ~98C ... at 100C it kills itself.
61 2013-10-28 02:43:30 <gmaxwell> unlike the hardware or the OS vendor we can't reasonably add things like thermal throttling to bitcoin, because the mechenisms are too system specific.
62 2013-10-28 02:43:42 <lianj> i thought those machines are only 70% idle because the rest is taken for LTS game of malwares
63 2013-10-28 02:44:28 <gmaxwell> lianj: worse, the malware is also a not uncommon cause of overheating, so people will sometimes assume they got malware.
64 2013-10-28 02:44:45 <warren> when they are really just mining for bitcoins
65 2013-10-28 02:44:47 <warren> =)
66 2013-10-28 02:45:04 <lianj> warren: just sell some bitcoin and get a new machine
67 2013-10-28 02:45:29 <warren> lianj: my last three generations of Thinkpad are still working
68 2013-10-28 02:45:31 <gmaxwell> Thinkpads are usually pretty good. My x230 is fine with all four cores pegged and the lid closed.
69 2013-10-28 02:45:40 <warren> four cores?!
70 2013-10-28 02:45:44 <warren> I'd love a 4 core thinkpad
71 2013-10-28 02:46:11 <lianj> get one :P
72 2013-10-28 02:46:16 <gmaxwell> model name : Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz
73 2013-10-28 02:46:17 <warren> I didn't know they existed
74 2013-10-28 02:46:40 <warren> gmaxwell: how heavy?
75 2013-10-28 02:47:01 <sipa> 1.3 kilobytes
76 2013-10-28 02:47:09 <lianj> ^^
77 2013-10-28 02:47:24 <lianj> including the crappy display :(
78 2013-10-28 02:47:50 <gmaxwell> it's an x230, it's not an ultraportable netbook but it's smaller than $random_laptop yea, the bummer is that the display is nowhere near as good as my old x61t (SXGA+) was.
79 2013-10-28 02:48:15 <warren> I need some expenses to offset income. maybe it's time to buy a new laptop
80 2013-10-28 02:48:41 <gmaxwell> I understand the new haswell ones are nice... maybe even some models coming out with better displays, but I don't know if those are shipping yet.
81 2013-10-28 02:48:44 <lianj> i mean resolution. so sad no new xseries got a good one. staying with my x220 until the ship one
82 2013-10-28 02:49:05 <warren> haswell is out? I haven't been following hardware at all.
83 2013-10-28 02:49:19 <gmaxwell> warren: yes, the desktop and mobile ones are out.
84 2013-10-28 02:49:36 <lianj> warren: thats why you whine about bitcoin freezing your gameboy :P
85 2013-10-28 02:49:53 <gmaxwell> multisocket server parts are still stuck with IB-E
86 2013-10-28 02:50:26 <warren> my T410s has a weird battery that is very expensive to replace
87 2013-10-28 02:50:33 <warren> I guess that's true of all T*s
88 2013-10-28 02:51:18 <gmaxwell> warren: hm? usually there is an unending supply of vendors on ebay with mostly okay greymarket thinkpad batteries.
89 2013-10-28 02:51:28 <warren> gmaxwell: this is a weird one
90 2013-10-28 02:51:53 <warren> gmaxwell: only T400s and T410s, and both were unpopular models because of omgwtfbbq hardware defects
91 2013-10-28 02:52:03 <warren> I've had this laptop service 9 times under warranty
92 2013-10-28 02:53:41 <warren> grr,, the T431s has an entirely different power adapter and non-replacceable battery
93 2013-10-28 02:54:35 <gavinandresen> We should all get together and vow to only buy laptops that run on standard AAA batteries. Who's with me?
94 2013-10-28 02:54:52 <lianj> hell no
95 2013-10-28 02:55:03 <gavinandresen> what, you want D-cells?
96 2013-10-28 02:55:06 <gavinandresen> 9 volts?
97 2013-10-28 02:55:07 <warren> hahaa
98 2013-10-28 02:55:12 <sipa> gavinandresen: does it also apply to laptops you *get*, rather than buy?
99 2013-10-28 02:55:19 <warren> gavinandresen: potato batteries
100 2013-10-28 02:55:33 <gavinandresen> sipa: you getting sent lots of laptops these days?
101 2013-10-28 02:55:38 <gavinandresen> warren: Hand Crank FTW!
102 2013-10-28 02:55:42 <sipa> warren: you played too much portal
103 2013-10-28 02:55:47 <sipa> gavinandresen: one suffices
104 2013-10-28 02:55:50 <warren> actually I haven't played portal
105 2013-10-28 02:56:14 <sipa> actually, portal 2
106 2013-10-28 02:56:30 <Luke-Jr> gavinandresen: do any Loongson laptops run on AAA?
107 2013-10-28 02:56:51 <warren> the last game I played was FFXIII
108 2013-10-28 02:57:06 <sipa> in soviet russia, battery drains you
109 2013-10-28 02:57:57 <warren> damn, there's no thin laptops with quad core
110 2013-10-28 02:58:10 <lianj> define thin
111 2013-10-28 02:58:42 <warren> T4xxs or thinner
112 2013-10-28 03:00:14 <Luke-Jr> not even quad core ARM? :P
113 2013-10-28 03:00:24 <warren> the current laptops don't look much better than what I have
114 2013-10-28 03:00:49 <warren> I just need to strap on a 10,000 BTU A/C to make my laptop usable.
115 2013-10-28 03:01:17 <lianj> ACTION gives up
116 2013-10-28 03:01:24 <lianj> just don't blame bitcoin
117 2013-10-28 03:01:40 <Luke-Jr> sigh, I seem to have gotten into a debate with someone who insists GPU mining with DirectCompute (DirectX) makes sense
118 2013-10-28 03:05:23 <warren> laptop died again. time to limit it to 50% of one core.
119 2013-10-28 03:06:46 <phantomcircuit> warren, what are you doing
120 2013-10-28 03:07:01 <warren> phantomcircuit: syncing bitcoin blockchain
121 2013-10-28 03:09:01 <lianj> warren: you run thinkfan?
122 2013-10-28 03:09:33 <warren> lianj: this laptop design just sucks. at 100% fan it can't handle full power.
123 2013-10-28 03:09:46 <lianj> how do you know its 100%
124 2013-10-28 03:10:00 <lianj> controlling the can by yourself?
125 2013-10-28 03:10:05 <warren> because I owned this laptop since 2010 and I've fiddled with it before
126 2013-10-28 03:10:37 <lianj> so with thinfan and last level being 127 aka real full speed it still happens? find it hard to believe
127 2013-10-28 03:11:30 <warren> lianj: this laptop just sucks
128 2013-10-28 03:12:12 <warren> there, 80% of one core and temperature is hovering at 95C
129 2013-10-28 03:12:14 <warren> not pushing it harder
130 2013-10-28 03:13:58 <lianj> withs the fans rpm?
131 2013-10-28 03:14:02 <lianj> *whats
132 2013-10-28 03:14:25 <warren> dunno, I fiddled with it in windows and later /sys something
133 2013-10-28 03:14:31 <phantomcircuit> warren, which model?
134 2013-10-28 03:14:40 <warren> phantomcircuit: Thinkpad T410s
135 2013-10-28 03:15:12 <lianj> warren: if you dunno and still cry, you are bad at debugging
136 2013-10-28 03:15:16 <phantomcircuit> that's nearly identical to the T61p i have
137 2013-10-28 03:15:23 <phantomcircuit> which also has terrible thermal management
138 2013-10-28 03:15:28 <lianj> cat /proc/acpi/ibm/fan
139 2013-10-28 03:16:12 <warren> lianj: I've looked years ago and came to the conclusion this thing can't handle full power
140 2013-10-28 03:16:14 <phantomcircuit> lianj, nah ibm screwed up on this one
141 2013-10-28 03:16:37 <phantomcircuit> they rely really heavily on copper heatsinks to move the heat from the cpu to the fan
142 2013-10-28 03:16:51 <phantomcircuit> the problem is it doesn't move 100% of the hear
143 2013-10-28 03:16:53 <phantomcircuit> heat*
144 2013-10-28 03:17:10 <phantomcircuit> and the fans are hilariously underpowered
145 2013-10-28 03:17:49 <lianj> that might be, but i can't believe doing `echo "level disengaged" > /proc/acpi/ibm/fan` won't make it at least a bit better
146 2013-10-28 03:18:11 <phantomcircuit> lianj, the fan on this model will rip itself apart if you do that
147 2013-10-28 03:18:23 <phantomcircuit> i went through 4 fans on my T61p before i just gave up
148 2013-10-28 03:18:49 <warren> phantomcircuit: you know where to buy replacement hsf's cheap?
149 2013-10-28 03:19:11 <lianj> phantomcircuit: but you destroyed that piece of hardware then, right :)
150 2013-10-28 03:19:31 <phantomcircuit> warren, you can either buy the cheap chinese ones where you have to actually replace the fan unit in the heatsink
151 2013-10-28 03:19:45 <phantomcircuit> or buy the ones from lenovo that are like $80 each
152 2013-10-28 03:21:27 <warren> I feel like if I can replace the HSF, keyboard and mouse, this thing will last another 3 years.
153 2013-10-28 03:23:04 <phantomcircuit> warren, unlikely but go for it
154 2013-10-28 03:23:55 <warren> I'd upgrade if there were a quad core laptop of similar weight.
155 2013-10-28 03:24:30 <gmaxwell> warren: huh? the T410 is like 5lbs.
156 2013-10-28 03:24:36 <gmaxwell> The x230s are 3 lbs.
157 2013-10-28 03:24:58 <warren> gmaxwell: T410 yes. T410s with the drive bay removed (useless anyway) is lighter than an ultrabook
158 2013-10-28 03:25:20 <warren> less than 3lbs
159 2013-10-28 03:25:27 <gmaxwell> ah. yea, well, the really light notebook market has mostly been killed by crappy tablets.
160 2013-10-28 03:25:36 <lianj> gmaxwell: you like the keyboard? (never tried, but skeptik)
161 2013-10-28 03:25:49 <warren> I miss the old thinkpad keyboard.
162 2013-10-28 03:26:33 <MC1984> warren there is probably a nice blanket of navel lint on the inside of the heatsink fins
163 2013-10-28 03:26:40 <gmaxwell> lianj: I have mixed feelings about the new keyboards. They feel fine, pretty good actually. But they don't have as strong of a "positive confirmation" that you hit the button, so it makes me miss characters more often (esp r, e, t for some reason)
164 2013-10-28 03:26:53 <warren> gmaxwell: exactly!
165 2013-10-28 03:27:11 <warren> MC1984: I clean it thoroughly because it sucks horribly if I don't.
166 2013-10-28 03:27:28 <MC1984> thermal paste?
167 2013-10-28 03:28:06 <gmaxwell> They're easier to keep clean though.
168 2013-10-28 03:28:29 <gmaxwell> and I get the impression that it'll last longer. on the x61t I think I replaced my keyboard 5 times.
169 2013-10-28 03:28:35 <MC1984> this one i have here is an i7 quad + nvidia 650gt something chip serviced by one heatpipe and about 4 inches of heatsink
170 2013-10-28 03:28:40 <MC1984> it seems sufficient
171 2013-10-28 03:29:08 <gmaxwell> Hm. actually probably only three times, I replaced Kat's x61t keyboard twice.
172 2013-10-28 03:29:27 <lianj> gmaxwell: yea, likewise. but loved the x60t. never used the touch much but rotating the screen to another person was priceless
173 2013-10-28 03:30:03 <MC1984> im thinking of getting an old thinkpad for my mother for christmas actually
174 2013-10-28 03:30:12 <MC1984> theres a UK site that is selling them cheap w/ coreboot installed
175 2013-10-28 03:30:36 <gmaxwell> lianj: yea. it was also handy for reading things on airplanes... thats pretty much the only times I used it. But both our x61ts hinges eventually failed, and they were a pita to replace. Kat used the tablet mode enough that she has the x230t now.
176 2013-10-28 03:31:54 <lianj> yea, what was the downsid of x230t vs no t, i think no additional pci slot or something?
177 2013-10-28 03:33:08 <gmaxwell> lianj: yea, well, also t-type battery, and the doomed to be less reliable t-hinge. it's also a bit thicker.
178 2013-10-28 03:33:40 <warren> actually. one of the reasons why I quit bitcoin in 2010 was my laptop overheated and killed itself.
179 2013-10-28 03:33:53 <warren> same laptop
180 2013-10-28 03:33:56 <lianj> and now you want to keep it for another 3 years
181 2013-10-28 03:34:11 <warren> well, it's fine as long as I use one core at 50% speed
182 2013-10-28 03:34:12 <warren> =)
183 2013-10-28 03:34:17 <gmaxwell> really my favorite thinkpad was the x60 except for the screen. .. no stupid touchpad, super small and light. Even with the jumbo gives-it-12-hours slice battery it was thinner than most laptops.
184 2013-10-28 03:34:21 <lianj> warren: great
185 2013-10-28 03:35:12 <gmaxwell> my second favorite was the x61t. ... last 4:3 laptop I'll probably ever be able to own, awesome sxga+ screen. good original lenovo keyboard.
186 2013-10-28 03:35:42 <lianj> yea, slice battery was/is great. although i destroyed 2 when charging them in the us. i know it makes no sense but was reproduceable (thus the two instead of one)
187 2013-10-28 03:35:57 <MC1984> heh warren thats really not normal
188 2013-10-28 03:37:07 <warren> this model was a serious disaster. ALL of them had this defect: http://www.youtube.com/watch?v=rM9_ZM_8qkQ
189 2013-10-28 03:37:21 <warren> they refused to fix it until I posted this video
190 2013-10-28 03:37:39 <gmaxwell> the x61* sxga+ displays got bubbles.
191 2013-10-28 03:38:12 <lianj> mine didn't, a friend still uses it
192 2013-10-28 03:39:08 <gmaxwell> lianj: the two here did, it was commonly reported. But they tended to be on the edge and so you (or at least we) got used to them. they were worth it.
193 2013-10-28 03:39:16 <MC1984> that screen thing seems like it might be the LVDS bundle or its connector
194 2013-10-28 03:39:29 <warren> perhaps one day bitcoin will be discussed here again. I'm sorry.
195 2013-10-28 03:39:43 <lianj> you started it
196 2013-10-28 03:39:48 <warren> I know.
197 2013-10-28 03:49:34 <BlueMatt> gavinandresen: while I'm at it, anything else need done to pull-tester?
198 2013-10-28 03:49:46 <BlueMatt> or anyone else have tests they want added to BitcoindComparisonTool?
199 2013-10-28 03:49:50 <BlueMatt> gmaxwell, sipa: ^
200 2013-10-28 03:55:19 <gmaxwell> BlueMatt: the tester ought to have some bad serialization tests, scriptpubkey invalidity (we don't check scriptpubkeys for the most part, so putting invalid things it is permitted), and dead script branch invailidy tests cases that I think it lacks.
201 2013-10-28 03:56:03 <gmaxwell> apparently btcd got forked on testnet after passing pull tester (bummer, since it kinda suggests that they were just iterating to pass pull tester, alas), but they didn't open an issue on the specific trigger transaction so I dunno which one it was.
202 2013-10-28 03:56:29 <BlueMatt> davec added a script branch with invalid opcode test a few weeks ago
203 2013-10-28 03:56:48 <BlueMatt> Im assuming to address that specific oversight
204 2013-10-28 03:56:55 <BlueMatt> bad serialization sounds fun
205 2013-10-28 03:57:53 <gmaxwell> BlueMatt: yea, I can give an example of that, one sec.
206 2013-10-28 04:00:52 <gavinandresen> BlueMatt: pull-tester should be working? So if I manually re-run https://github.com/bitcoin/bitcoin/pull/3164 it should succeed?
207 2013-10-28 04:01:53 <BlueMatt> gavinandresen: working on the tmpfs thing now
208 2013-10-28 04:01:56 <BlueMatt> once I get that done, yes
209 2013-10-28 04:02:17 <gavinandresen> BlueMatt: awesome.
210 2013-10-28 04:13:22 <BlueMatt> gavinandresen: windows builds etc are all running?
211 2013-10-28 04:15:27 <gavinandresen> BlueMatt: yes, but windows builds were not being tested against the block-tester or unit tests run
212 2013-10-28 04:15:39 <BlueMatt> why no unit tests?
213 2013-10-28 04:16:28 <gavinandresen> BlueMatt: I was just cutting down number of variables to get things working-- wanted to cut out the wine dependency
214 2013-10-28 04:16:42 <BlueMatt> :(
215 2013-10-28 04:16:48 <BlueMatt> is wine installed in the chroot?
216 2013-10-28 04:17:16 <gavinandresen> BlueMatt: yes, I think so. Would probably work...
217 2013-10-28 04:17:26 <BlueMatt> also, the number of variable in pull-tester.py is fracking unreadable...
218 2013-10-28 04:17:47 <gavinandresen> ⦠but I'm not sure running unit tests twice, once under wine and once native, is worth it.
219 2013-10-28 04:17:56 <BlueMatt> it has caught issues in the past
220 2013-10-28 04:18:03 <BlueMatt> as has running the block tester twice
221 2013-10-28 04:18:11 <gavinandresen> There is definite benefit to having a quick commit-get-pull-tester-results-quickly cycle
222 2013-10-28 04:18:39 <BlueMatt> you can check the logs right after it starts
223 2013-10-28 04:18:49 <gavinandresen> Running git HEAD against the block tester and unit tests in wine / etc makes sense to me.
224 2013-10-28 04:20:32 <BlueMatt> well it looks like jenkins is stopped?
225 2013-10-28 04:20:38 <gavinandresen> in my humble opinion⦠ideal system is per-pull pull-tester should compile linux and windows binaries and make them available (and comment).
226 2013-10-28 04:20:54 <BlueMatt> no builds since sept :(
227 2013-10-28 04:20:56 <gavinandresen> ⦠and jenkins does the whole-shebang test-everything-up-the-wazoo
228 2013-10-28 04:21:15 <gavinandresen> Yes, I turned off jenkins because Everything Was Changing.
229 2013-10-28 04:21:23 <BlueMatt> ahh
230 2013-10-28 04:21:46 <gavinandresen> It was one of those "can't change just this little thing without it turning into a three day project"
231 2013-10-28 04:22:01 <gavinandresen> (I don't remember details)
232 2013-10-28 04:22:11 <BlueMatt> sounds like fun...
233 2013-10-28 04:22:20 <gavinandresen> yeah, not.
234 2013-10-28 04:22:34 <Tril> window 13
235 2013-10-28 04:22:59 <BlueMatt> hmm...so you're saying I should learn autotools and restructure the test scripts to make it either run pull-tester simple tests of jenkins full tests?
236 2013-10-28 04:23:06 <gavinandresen> I haven't heard from Cory or Andreas, I was hoping they'd make it work.
237 2013-10-28 04:23:07 <BlueMatt> that sounds like a grand ol time
238 2013-10-28 04:24:06 <gavinandresen> BlueMatt: no, what I'm really saying is we need a paid QA lead, but nobody has stepped up and proven they're competent to do that.
239 2013-10-28 04:24:41 <gavinandresen> well, competent and WILLING
240 2013-10-28 04:25:02 <gavinandresen> Plus, the autotools stuff still hasn't settled down 100% last I heard
241 2013-10-28 04:36:03 <gmaxwell> BlueMatt: what problem are you having with autotools?
242 2013-10-28 04:36:19 <BlueMatt> gmaxwell: have to add new conditionals and such, halfway done though
243 2013-10-28 04:37:19 <gmaxwell> K. (lots of people floating around have autotools expirence so make sure you ask if you get stuck, no one in the universe actually understands itâ mind youâ but some expirence is usually good enough)
244 2013-10-28 04:38:33 <gavinandresen> Where are we at with documentation and autotools, by the way? I haven't been paying very close attention....
245 2013-10-28 04:39:01 <Luke-Jr> autotools is getting there slowly but surely
246 2013-10-28 04:39:09 <Luke-Jr> (void*)(intptr_t)a <-- ugly :<
247 2013-10-28 04:41:24 <gmaxwell> I observed last night that doc/build-unix.md at a minimum needs updating as it refers to modifying makefiles.
248 2013-10-28 04:43:48 <gmaxwell> gavinandresen: do you see any problem with creating issues for test infrastructur that ought to exist but doesn't, perhaps tagged with label test?
249 2013-10-28 04:45:37 <gavinandresen> gmaxwell: no, good idea.
250 2013-10-28 04:45:57 <BlueMatt> gmaxwell: like?
251 2013-10-28 04:47:33 <gavinandresen> like: every commit should be unit tested so git bisect always works
252 2013-10-28 04:48:08 <BlueMatt> you mean by jenkins or pull-tester?
253 2013-10-28 04:48:24 <gavinandresen> pull-tester
254 2013-10-28 04:48:33 <gavinandresen> .. by the time it gets into git HEAD it is too late
255 2013-10-28 04:48:51 <BlueMatt> pull-tester needs an initial "looks ok" response and then a delayed full-test tests each commit + runs big reorg checks in block tester "LOOKS GOOD TO MERGE" response
256 2013-10-28 04:49:12 <BlueMatt> gavinandresen: can pull-tester.py be made open source yet?
257 2013-10-28 04:49:22 <BlueMatt> arent all the important things in env vars by now?
258 2013-10-28 04:49:29 <gavinandresen> BlueMatt: yes, it is in the tree under qa/pull-tester
259 2013-10-28 04:49:59 <gavinandresen> BlueMatt: confusing because there is a pull-tester.sh and a pull-tester.py thereâ¦. ah well
260 2013-10-28 04:50:04 <BlueMatt> oops, so when I edit it on the server its evil...
261 2013-10-28 04:50:16 <gmaxwell> BlueMatt: e.g. https://github.com/bitcoin/bitcoin/issues/3170
262 2013-10-28 04:50:37 <gavinandresen> ok to edit it on the server, just please submit changes through github
263 2013-10-28 04:50:54 <BlueMatt> will do
264 2013-10-28 04:51:07 <BlueMatt> gmaxwell: OH GOD that would take hours
265 2013-10-28 04:51:22 <gmaxwell> BlueMatt: yea so? doesn't have to run for every pull. :P
266 2013-10-28 04:51:36 <BlueMatt> we need a new test infrastructure...
267 2013-10-28 04:51:48 <BlueMatt> pull-tester+jenkins distinction aint gonna work no longer methinks
268 2013-10-28 04:52:26 <gavinandresen> BlueMatt: mmm. We've got three different environments for building stuff (pull-tester chroot, jenkins, and gitian)â¦. should be just gitian
269 2013-10-28 04:53:03 <gmaxwell> its a test we could run once prior to release (I've been doing a similar test prior to release: I always reindex the chain during RC ... but that wouldn't have caught a disconnect only bug), though I think we should just run it constantly and add computing power as required.
270 2013-10-28 04:53:08 <BlueMatt> yea, but you still need infrastructure to run gitian with different flags for all-tests, some-tests, tests-for-each-commit-on-master, test-run-daily, etc
271 2013-10-28 04:55:09 <davec> BlueMatt/gmaxwell: Yes that is the script I added - invalid opcode in a dead execution branch - I don't belive TD has added it though yet since it's against your branch
272 2013-10-28 04:55:31 <BlueMatt> davec: if its on my branch itl get run on pull-tester
273 2013-10-28 04:55:33 <Luke-Jr> miner.c:2434:2: warning: format â%xâ expects argument of type âunsigned intâ, but argument 4 has type âunsigned intâ [-Wformat=] <-- wtf GCC?
274 2013-10-28 04:55:40 <BlueMatt> davec: getting into master for block-tester doesnt really matter
275 2013-10-28 04:57:28 <davec> BlueMatt: aye, I don't think he's merged it over though last time I looked - checking now
276 2013-10-28 04:58:14 <BlueMatt> it is now
277 2013-10-28 04:58:25 <BlueMatt> I hadnt had a chance to get the whole shebang working till yesterday
278 2013-10-28 04:58:48 <davec> ah great
279 2013-10-28 04:59:11 <phantomcircuit> BlueMatt, could you setup jenkins to run on AWS EC2 instances
280 2013-10-28 04:59:22 <BlueMatt> phantomcircuit: if I get it running at all, maybe
281 2013-10-28 04:59:22 <phantomcircuit> like you spin up an EC2 instance just to run one test
282 2013-10-28 04:59:29 <warren> sigh. 50% speed on one core isn't enough. died again.
283 2013-10-28 04:59:47 <phantomcircuit> warren, pull the keyboard off and point a desktop fan at it
284 2013-10-28 04:59:48 <phantomcircuit> :/
285 2013-10-28 04:59:50 <phantomcircuit> (seriously)
286 2013-10-28 05:00:03 <warren> phantomcircuit: I did before, that 10,000 BTU A/C, to render a video
287 2013-10-28 05:00:49 <warren> I'm open to buying a new laptop, but nothing appears to be much better.
288 2013-10-28 05:00:58 <BlueMatt> warren: get a desktop...
289 2013-10-28 05:01:05 <warren> BlueMatt: moving soon...
290 2013-10-28 05:01:12 <BlueMatt> laptop for mobile stuff, move important things to desktop
291 2013-10-28 05:01:13 <warren> expensive to ship things from Hawaii
292 2013-10-28 05:01:27 <BlueMatt> warren: you dont have to take it out of the box, just keep it in the motherboard box as a case
293 2013-10-28 05:01:29 <BlueMatt> (its what I do)
294 2013-10-28 05:01:38 <warren> heh
295 2013-10-28 05:01:41 <BlueMatt> 1 motherboard box + psu is pretty small
296 2013-10-28 05:01:58 <warren> with six pcie cables and ards
297 2013-10-28 05:02:16 <BlueMatt> what are you doing, gpu mining?
298 2013-10-28 05:02:53 <phantomcircuit> BlueMatt, initial sync
299 2013-10-28 05:02:53 <phantomcircuit> lol
300 2013-10-28 05:03:19 <phantomcircuit> warren, just give up and buy two laptops, a desktop replacement and a mobile one
301 2013-10-28 05:03:34 <warren> what's funny, litecoind doesn't overheat this laptop
302 2013-10-28 05:03:43 <BlueMatt> ACTION just ran resync on his desktop in <4 hours
303 2013-10-28 05:03:46 <warren> something to do with "scrypt is too slow" and "no transactions to verify"
304 2013-10-28 05:03:50 <BlueMatt> and the disk is like 5 years old
305 2013-10-28 05:04:06 <Luke-Jr> warren: you've finally found a use case for litecoin!
306 2013-10-28 05:04:12 <Luke-Jr> "doesn't destroy crummy hardware"
307 2013-10-28 05:04:12 <warren> Luke-Jr: yeah!
308 2013-10-28 05:04:20 <Luke-Jr> lol
309 2013-10-28 05:04:27 <warren> Luke-Jr: I never claimed Litecoin is good.
310 2013-10-28 05:04:33 <warren> I just made it less bad.
311 2013-10-28 05:04:38 <Luke-Jr> XD
312 2013-10-28 05:05:08 <warren> I think all software is bad.
313 2013-10-28 05:06:12 <dobry-den> https://en.bitcoin.it/wiki/Protocol_specification#headers says that block_header[] items may be greater than 81bytes because they have a var_int (txn_count).
314 2013-10-28 05:06:30 <dobry-den> But every txn_count I get is 0 as it says on https://en.bitcoin.it/wiki/Protocol_specification#Block_Headers
315 2013-10-28 05:08:01 <CodeShark> that field is not used
316 2013-10-28 05:08:53 <dobry-den> Is the wiki entry wrong or am I misunderstanding?
317 2013-10-28 05:10:44 <dobry-den> It goes on to specify that miners do get block headers with a txn_count that does depict actual txn-count. But only miners get that version of the packet
318 2013-10-28 05:11:05 <dobry-den> Er, it says the other way around - that it's miners that receive 0 txn_count
319 2013-10-28 05:13:52 <CodeShark> the wiki entry is wrong
320 2013-10-28 05:14:27 <CodeShark> or rather, the tx count is not used at all in hashing
321 2013-10-28 05:16:12 <davec> the hash is the first 80 bytes (version 4, prevhash 32, merkle 32, ts 4, bits 4, nonce 4)
322 2013-10-28 05:18:03 <warren> hmm.... now I can't connect anything to RPC
323 2013-10-28 05:19:29 <BlueMatt> gavinandresen: ok, new version of pull tester all set up without the large reorg tests (thus no tmpfs needed)...however, now the pull must be merged before pull-tester can run on other pulls
324 2013-10-28 05:21:28 <gavinandresen> BlueMatt: ok. It will run on itself, though, right?
325 2013-10-28 05:21:36 <gavinandresen> (and "the pull" is 3164?)
326 2013-10-28 05:21:44 <BlueMatt> gavinandresen: its running on 3164 now, but aside from that its paused
327 2013-10-28 05:21:57 <gavinandresen> BlueMatt: cool. If it succeeds on itself, I will pull it.
328 2013-10-28 05:22:31 <gavinandresen> (and we'll fix it if it breaks on other things)
329 2013-10-28 05:23:45 <BlueMatt> ok, Ill look at re-enabling jenkins too
330 2013-10-28 05:28:39 <BlueMatt> aaaannnnddd...pull-tester success
331 2013-10-28 05:29:10 <warren> laptop died again. Screw this. I'm doing dev in a fast VM.
332 2013-10-28 05:29:19 <BlueMatt> warren: desktop.
333 2013-10-28 05:29:25 <BlueMatt> they're dirt cheap...
334 2013-10-28 05:29:38 <warren> BlueMatt: no time to deal with a desktop, and I want less stuff to move
335 2013-10-28 05:43:17 <BlueMatt> gavinandresen: thanks
336 2013-10-28 05:43:38 <warren> maybe I can change the threshold temperature where the computer kills itself...
337 2013-10-28 05:43:40 <BlueMatt> pull-tester running again, I likely broke it all, so ping me when you find the bugs :)
338 2013-10-28 05:44:03 <BlueMatt> warren: if you have software doing it, it usually just scales down, there is one thats in-hardware that you cant touch
339 2013-10-28 05:44:07 <gavinandresen> BlueMatt: thank YOU very much for all the
340 2013-10-28 05:44:22 <BlueMatt> dont thank me yet, wait till I get jenkins working :p
341 2013-10-28 05:45:48 <warren> um... I have no RPC port listening at all
342 2013-10-28 05:45:50 <warren> [warren@caprica bitcoin]$ netstat -ln |grep 833
343 2013-10-28 05:45:51 <warren> tcp 0 0 0.0.0.0:8333 0.0.0.0:* LISTEN
344 2013-10-28 05:45:51 <warren> tcp6 0 0 :::8333 :::* LISTEN
345 2013-10-28 05:45:58 <warren> daemon=1 server=1 should do it right?
346 2013-10-28 05:47:17 <Tril> warren: you need - rpcuser, rpcpassword, rpcallowip set
347 2013-10-28 05:47:48 <Tril> rpcport for good measure
348 2013-10-28 05:48:25 <warren> I have rpcuser and rpcpassword which should be good enough for localhost only
349 2013-10-28 05:48:38 <Tril> rpcallowip=127.0.0.1
350 2013-10-28 05:49:24 <Tril> it should listen when you do that
351 2013-10-28 05:49:52 <warren> "kill" is also failing to kill bitcoind
352 2013-10-28 05:49:56 <warren> only kill -9 works
353 2013-10-28 05:50:22 <Tril> it takes a long time to shutdown, especially if it's still starting up
354 2013-10-28 05:50:57 <warren> it's struggling to sync the block chain
355 2013-10-28 05:51:05 <warren> I'm pointing a AC unit at it right now
356 2013-10-28 06:38:13 <Phoebus> bitcoin-q left open for another 24 hours, data dir is 11.4GB, still not caught up. Displays no progress... out of sync wallet etc. Any ideas?
357 2013-10-28 06:48:07 <MC1984> warren im willing to bet there is somthing seriously wrong with the thermal interface of cpu and heatsink
358 2013-10-28 06:50:43 <warren> MC1984: this is my third CPU and HSF for this laptop, all three were just as bad.
359 2013-10-28 06:51:11 <BlueMatt> MC1984: it is a laptop....there's by far more often than not a design issue there
360 2013-10-28 06:51:39 <warren> It's perfectly fine! (If I have an A/C blowing into the intake.)
361 2013-10-28 06:55:39 <MC1984> ive never heard anything like it
362 2013-10-28 06:55:54 <MC1984> most intel cpus will turn themselves down before catching fire
363 2013-10-28 06:57:13 <BlueMatt> or, if they turn themselves down and still are gonna hit 100, they just poweroff
364 2013-10-28 06:58:07 <MC1984> have you considered there may be something wrong with the chassis to that the heatsink retention mechanism is bent and not touching the cpu
365 2013-10-28 06:58:51 <warren> MC1984: this laptop has been rebuilt 3 times by the depot due to different hardware defects
366 2013-10-28 06:58:56 <warren> I was just screwed in purchasing this model
367 2013-10-28 06:59:02 <warren> other owners of this model have similar issues
368 2013-10-28 06:59:40 <MC1984> what model is it?
369 2013-10-28 06:59:47 <MC1984> thinkpads are usually solid
370 2013-10-28 07:00:18 <BlueMatt> ACTION has a thinkpad that is literally always down-throttled because its overheated
371 2013-10-28 07:00:23 <BlueMatt> relatively recent t420 too
372 2013-10-28 07:00:39 <BlueMatt> any devs awake that wanna merge https://github.com/bitcoin/bitcoin/pull/3174
373 2013-10-28 07:01:30 <MC1984> ive got a y500 and its solid
374 2013-10-28 07:01:52 <warren> MC1984: Thinkpad T410s ... super thin and light, thin and light on the HSF too...
375 2013-10-28 07:01:59 <MC1984> i only managed to make the cpu thermal throttle by caning it with prime95 + furmark on a 40% gpu overclock
376 2013-10-28 07:02:07 <warren> it seems they fixed this problem with T420s
377 2013-10-28 07:02:08 <MC1984> thats an extreme workload
378 2013-10-28 07:02:22 <BlueMatt> and Ive gotten the motherboard replaced once and manually reapplied thermal paste and checked the thermal coupling several times
379 2013-10-28 07:03:03 <MC1984> are the x20x series ok? Looking at a cheapo one for mother
380 2013-10-28 07:04:03 <BlueMatt> if you shy away from one of the higher-end cpus (ie the cpu upgrade), you should be ok
381 2013-10-28 07:04:37 <MC1984> either core2duo or i5, somewhere over 2ghz
382 2013-10-28 07:09:25 <MC1984> http://www.notebookcheck.net/Review-Lenovo-Thinkpad-T410-Notebook-Optimus.42580.0.html
383 2013-10-28 07:38:11 <Luke-Jr> anyone checking 085 setup.exe on SF already?
384 2013-10-28 07:47:57 <BlueMatt> Luke-Jr: if nothing else, I get the right hash
385 2013-10-28 07:48:04 <BlueMatt> no idea if its valid, dont have wine on here (yet)
386 2013-10-28 07:48:20 <Luke-Jr> meh, if it matches the gitian hashes, it's fine I'm sure
387 2013-10-28 07:48:34 <BlueMatt> well, the signed hashes, not gitian
388 2013-10-28 07:51:02 <BlueMatt> wumpus: lol, I didnt think anyone was awake...that pull was (mostly) untested...lets see if it works
389 2013-10-28 07:52:04 <wumpus> BlueMatt: lol, yeah I just assumed it was ok the changes looked sane, well indeed let's see
390 2013-10-28 07:52:51 <BlueMatt> shit, forgot rpcport...
391 2013-10-28 07:53:32 <gmaxwell> sipa: Is what he's saying clear to you? https://bitcointalk.org/index.php?topic=320090.0
392 2013-10-28 07:53:54 <wumpus> I'm pretty easy with merging pulls that only involve tests, as you'd expect them to be tested :P
393 2013-10-28 07:56:50 <wumpus> gmaxwell: it's clear to me; for private keys we add a 01 byte to signify it's to be used with a compressed key
394 2013-10-28 07:57:28 <wumpus> gmaxwell: as we use a 'add to the end' convention for the private keys, and 'prefix' convention for the public keys, and the datas have the same size, there is a potential ambiguity if somehow one gets in the place of the other
395 2013-10-28 07:59:07 <wumpus> it would have been more consistent to always put the marker byte at the beginning and use a non-overlappng range for private and public key types
396 2013-10-28 07:59:43 <gmaxwell> wumpus: we don't have any base 58 encoded public key types at all.
397 2013-10-28 08:00:25 <wumpus> then again, it is usually quite clear whether some object is a private or public key (from the encoding, or otherwise) so it's not a very important issue
398 2013-10-28 08:00:30 <wumpus> right
399 2013-10-28 08:00:58 <warren> watchonly for 0.8.5. is working nice
400 2013-10-28 08:01:14 <wumpus> I doubt that this ever caused any problems for anyone
401 2013-10-28 08:01:15 <warren> I'm making this for bitcoin to help the Coinpunk guy have a stable client
402 2013-10-28 08:02:22 <gmaxwell> wumpus: we also have the type prefixs when in the base 58 encoding which makes them unambigious.
403 2013-10-28 08:02:44 <gmaxwell> it sounds like he wants an additional prefix on the non-base58 data?
404 2013-10-28 08:06:56 <wumpus> I don't like his 'trivial algorithm' at the end very much though, it still very much relies on the size and encoding of the data to determine the type
405 2013-10-28 08:07:08 <BlueMatt> wumpus: ok, noooww you can merge it
406 2013-10-28 08:07:09 <BlueMatt> https://github.com/TheBlueMatt/bitcoin/commit/b2b7bf4bf9b3a7b1c4bb8470289da9180da9f13b
407 2013-10-28 08:07:12 <BlueMatt> https://github.com/bitcoin/bitcoin/pull/3175
408 2013-10-28 08:07:13 <BlueMatt> sorry
409 2013-10-28 08:08:12 <wumpus> gmaxwell: well I thought he wanted to define some encoding-independent scheme, but I'm not entirely sure anymore :-)
410 2013-10-28 08:09:51 <BlueMatt> ok, Im taking a wild guess that jenkins wont fail now, so Im gonna enable irc notifications
411 2013-10-28 08:10:00 <BlueMatt> if it fails, have fun with those tonight...
412 2013-10-28 08:10:02 <BlueMatt> ACTION -> bed
413 2013-10-28 08:10:53 <gmaxwell> wumpus: I think the issue is that 'sx' has adopted a weird architecture where base58 decoding is like a seperate commandline tool, and then you're stuck with untyped hex that its passing around.
414 2013-10-28 08:12:02 <Luke-Jr> wtf for? :/
415 2013-10-28 08:12:11 <Luke-Jr> base58 is an obvious case for a simple shared library
416 2013-10-28 08:12:12 <wumpus> gmaxwell: that's a likely explanation
417 2013-10-28 08:12:33 <gmaxwell> Luke-Jr: because it's attempted to factor out all the parts of bitcoin in order to more expose it for power users.
418 2013-10-28 08:12:51 <gmaxwell> Which is certantly a notion I'm sympathic to... more unixy.
419 2013-10-28 08:13:05 <gmaxwell> Though that doesn't mean that any particular factorization is a good idea. :)
420 2013-10-28 08:13:14 <Luke-Jr> sigh
421 2013-10-28 08:13:40 <Luke-Jr> as if hex is any less an encoding than base58
422 2013-10-28 08:14:20 <wumpus> well, yes, conceptually it'd make sense to consistently label the raw hex data with a marker byte at the beginning instead of rely on the data length and encoding to convey meaning
423 2013-10-28 08:14:29 <gmaxwell> Luke-Jr: well, it might matter if we used base58 for anything which was interesting to manipulate directly, like transaction data.
424 2013-10-28 08:15:00 <wumpus> but I'm afraid that ship has kind of sailed, and introducing more types of encoding doesn't make it easier
425 2013-10-28 08:15:40 <wumpus> it just means that there are even more options...
426 2013-10-28 08:16:47 <gmaxwell> wumpus: yea, that was just the only thing we could do to support compressed keys in a way that was both backwards compatible _and_ didn't require allocating another base58 type value.
427 2013-10-28 08:16:59 <gmaxwell> I don't recall why we didn't want to allocate another base58 type value though.
428 2013-10-28 08:17:46 <gmaxwell> (sipa would remember, no doubt)
429 2013-10-28 08:21:59 <warren> sipa: not having luck with subtree
430 2013-10-28 08:33:15 <sipa> gmaxwell: a new base58 version for what?
431 2013-10-28 08:34:00 <warren> sipa: tried many variants of git subtree pull -P src/leveldb ... and nothing is achieving the goal
432 2013-10-28 08:34:35 <sipa> did you add and fetch the leveldb repo as a remote?
433 2013-10-28 08:34:42 <warren> yes
434 2013-10-28 08:36:07 <sipa> oh, my mistake
435 2013-10-28 08:36:17 <sipa> pull is for when you're pulling from another repo
436 2013-10-28 08:36:28 <sipa> if it is within one repo, merge should suffice
437 2013-10-28 08:36:58 <warren> I want the leveldb-bitcoin-1.13 tag?
438 2013-10-28 08:37:21 <warren> hmm
439 2013-10-28 08:38:15 <sipa> git subtree merge --squash -P src/leveldb leveldb-bitcoin-1.13
440 2013-10-28 08:38:20 <sipa> indeed
441 2013-10-28 08:39:05 <warren> that did it!
442 2013-10-28 08:39:09 <warren> thanks
443 2013-10-28 08:39:34 <warren> we'll have some widespread testing of 1.13 long before 0.9
444 2013-10-28 09:36:58 <michagogo> cloud|warren: have you considered liquid cooling?
445 2013-10-28 09:37:46 <michagogo> cloud|Hook up an external keyboard, mouse, and monitor, and dunk the laptop in a tank of mineral oil
446 2013-10-28 09:38:16 <michagogo> cloud|(Disclaimer: I've never done this myself. Research before attempting.)
447 2013-10-28 09:39:41 <michagogo> cloud|I think I saw something at some point about this guy who overclocked some old 100-something MHz processor to over 2 GHz, with a bathtub of glycerin and a disassembled fridge
448 2013-10-28 10:04:36 <MrDaneelOlivaw> morning!
449 2013-10-28 10:16:34 <SomeoneWeird> michagogo|cloud, don't stick hdds in oil
450 2013-10-28 10:25:04 <wumpus> I'm getting so friggin annoyed here... can someone slap pjotr_n on the forums for me please https://bitcointalk.org/index.php?topic=306923.msg3426364#msg3426364
451 2013-10-28 11:59:52 <monode> hi!
452 2013-10-28 12:00:05 <monode> is it still feasible to do CPU mining on the testnet?
453 2013-10-28 12:00:18 <killerstorm> yes
454 2013-10-28 12:00:37 <killerstorm> it is locked to difficulty of 1
455 2013-10-28 12:08:47 <killerstorm> Hi. We are looking for Python programmers who can help with colored coins client (NGCCC) development. Currently we offer 0.25 BTC/hr (limited time offer!). Easy and small tasks are available.
456 2013-10-28 12:09:25 <killerstorm> If interested please pm me or join #bitcoinx
457 2013-10-28 12:14:47 <sipa> killerstorm: eh, not at all
458 2013-10-28 12:14:52 <sipa> testnet difficulty is around 18000
459 2013-10-28 12:15:44 <killerstorm> wow. it is reset from time to time, right?
460 2013-10-28 12:16:29 <sipa> there have been 3 testnet instances
461 2013-10-28 12:16:45 <sipa> and there is a special rule that allows creating a difficulty 1 block if there hasn't been one in 20 minutes
462 2013-10-28 12:16:53 <Happzz> colored coins client?
463 2013-10-28 12:17:22 <sipa> and there is a bug that if this special-difficulty-1 block happens at retarger (every 2016 blocks), the difficulty resets permanently to 1
464 2013-10-28 12:17:30 <killerstorm> ah, ok.
465 2013-10-28 12:17:38 <sipa> well, not permanently, just acvtually resets
466 2013-10-28 12:17:44 <sipa> rather than allowing a 1-time exception
467 2013-10-28 12:18:08 <killerstorm> I've been mining for a couple of weeks, got 200 coins so far. I guess I started before the difficulty jump.
468 2013-10-28 12:19:09 <killerstorm> colored coins intro: https://github.com/bitcoinx/colored-coin-tools/wiki/colored_coins_intro
469 2013-10-28 12:54:07 <jgarzik> mornin'
470 2013-10-28 12:54:09 <jgarzik> yee gads
471 2013-10-28 12:54:33 <jgarzik> A discussion of error codes could not be more boring ;p
472 2013-10-28 12:57:14 <petertodd> killerstorm: didn't like nSequence-based ordering?
473 2013-10-28 12:57:37 <killerstorm> Sorry for the lack of reply. I like it.
474 2013-10-28 12:58:11 <petertodd> killerstorm: likewise. Cool, I think it's the best I've come up with, and it's safe to use nSequence like that.
475 2013-10-28 12:58:20 <petertodd> killerstorm: (by using it, you'll make it safe in the future too...)
476 2013-10-28 12:59:13 <petertodd> killerstorm: unfortunately nSequence is the *only* data stored in a CTxIn that we can both set arbitrarily, and is signed :(
477 2013-10-28 13:00:07 <killerstorm> Well, it's possible to use least-significant bits. But that's uglier, I guess.
478 2013-10-28 13:00:07 <petertodd> killerstorm: everything else doesn't meet that criteria, which means you can do a CoinJoin'd colored coin transaction - useful to obscure who actually bought it.
479 2013-10-28 13:00:28 <killerstorm> One dude found a way how to use lsb without problems with central registry.
480 2013-10-28 13:00:28 <petertodd> killerstorm: right, CTxOuts are signed, but they may be set to their values by someone else.
481 2013-10-28 13:00:35 <petertodd> killerstorm: how so?
482 2013-10-28 13:01:05 <killerstorm> each color has a set of tags, like (13, 37, 55, ...)
483 2013-10-28 13:01:26 <killerstorm> when we need to use several colors in one transaction, we find tags which are unambiguous.
484 2013-10-28 13:01:42 <killerstorm> basically, from a symmetric difference of tag sets of all involved colors.
485 2013-10-28 13:01:56 <petertodd> killerstorm: hmm... sounds kinda complex :)
486 2013-10-28 13:02:25 <killerstorm> I wouldn't call it very complex,but this would require largish "color definitions".
487 2013-10-28 13:03:20 <killerstorm> do you think clients should implement infinite-precision arithmetic, or that 2^64 is enough for everybody?
488 2013-10-28 13:03:21 <petertodd> killerstorm: I dunno, given that there isn't a version of colored coins that doesn't require tx's all the way back to the genesis txout, I'm not sure I see the advantage of using anything but the nSequence method. Maybe one day we'll be able to write scriptPubKeys that constrain their outputs, but that's a long way off.
489 2013-10-28 13:03:51 <petertodd> killerstorm: Ha, 2^64 is probably fine - use the lshift rshift by two trick I told you about.
490 2013-10-28 13:03:55 <killerstorm> Well, one advantage is that it is very easy to detect transactions with nSequence :)
491 2013-10-28 13:04:12 <petertodd> killerstorm: Sure, but it's a completely standard part of the protocol!
492 2013-10-28 13:05:06 <petertodd> killerstorm: Now if that is an issue, I guess you're stick with LSB ordering, maybe with a bit of crypto hiding to let you brute-force LSB's that others in a coinjoin aren't using.
493 2013-10-28 13:07:27 <petertodd> killerstorm: also, keep in mind that with OP_RETURN <data> you can store your txin->txout mapping there.
494 2013-10-28 13:08:10 <petertodd> killerstorm: heck, do it as either nSequence is used, or OP_RETURN <data> is used - that should cover everything
495 2013-10-28 13:09:18 <petertodd> if devs get nasty and try to block that, you use a pubkeyhash output, and even with P2SH^2 you can easily brute-force the first few bits to store your color mapping data
496 2013-10-28 13:09:43 <killerstorm> Eh, why not order-based coloring?
497 2013-10-28 13:10:07 <killerstorm> It isn't exactly undetectable, but detecting it isn't easy.
498 2013-10-28 13:10:35 <petertodd> killerstorm: that works fine too, but as I understand it, putting that in an arbitrary coinjoin tx gets a bit ugly
499 2013-10-28 13:13:38 <killerstorm> Yes, that's true. I'm not sure that 'ability to coinjon' is a major criterion.
500 2013-10-28 13:13:57 <killerstorm> Anyway, coloring per se is a small part of the client code.
501 2013-10-28 13:14:18 <killerstorm> So it isn't a problem to implement several versions, although eventually it is better to settle on one.
502 2013-10-28 13:15:02 <petertodd> Yeah, regardless there are a lot more UI issues. (taking user interface as broadly as possible)
503 2013-10-28 13:16:02 <petertodd> One thing that I was thinking, is in practice SPV w/ color-coins is probably not a big deal: it's easy to write an SPV client that follows the colors from the genesis outputs, just like you would with any other wallet transaction.
504 2013-10-28 13:17:00 <petertodd> Yes, there's a bunch of bandwidth, but do the math: it's really not so bad given there's that 1MB/block upper limit anyway.
505 2013-10-28 13:17:17 <petertodd> I don't see people "daytrading" colored coins... :)
506 2013-10-28 13:18:33 <jgarzik> petertodd, why not?
507 2013-10-28 13:18:42 <jgarzik> petertodd, if they represent a stock or bond, its inevitable
508 2013-10-28 13:19:35 <killerstorm> The problem with SPV client is that it isn't possible to prove lack of spend.
509 2013-10-28 13:20:08 <petertodd> killerstorm: but you don't have too! just make up a buy transaction in a trust free way, if it gets mined, that's all the evidence you need that the coin wasn't spent
510 2013-10-28 13:20:56 <petertodd> jgarzik: yes, but my point is there are sufficient limits on the day-trading due to fees that a ccoin will likely still be just a few hundred megs to download at most
511 2013-10-28 13:21:08 <killerstorm> OK, well, if done properly the only problem is DoS attack on thin clients.
512 2013-10-28 13:21:14 <petertodd> jgarzik: the real issue is even a bit of trading is more than enough to kill full nodes disk IO bandwidth...
513 2013-10-28 13:21:40 <killerstorm> Basically someone tells you that coins is of color X and you have to scan a lot of data to prove that it isn't. Maybe not a real issue.
514 2013-10-28 13:21:42 <petertodd> jgarzik: IE, picture each ccoin SPV client doing a full re-scan of the blockchain, just because someone added a coin
515 2013-10-28 13:22:09 <jgarzik> petertodd, SPV is kinda pointless with any layered scheme
516 2013-10-28 13:22:13 <petertodd> killerstorm: yup, and in that case I think the solution is to have a UI where it can have ccoin tx data added to it in a big dump
517 2013-10-28 13:22:34 <petertodd> jgarzik: what do you mean? SPV is great for layered schemes - saves having to actually do any of the work!
518 2013-10-28 13:22:44 <petertodd> jgarzik: just make sure your scheme is bloom filter comaptible
519 2013-10-28 13:23:02 <sipa> petertodd: huh? a colored coin fullnode would keep track of color in its UTXO set, i expect? why would it need to go rescan?
520 2013-10-28 13:23:28 <petertodd> sipa: I mean the rescan required when a user adds a new coin to their wallet
521 2013-10-28 13:24:05 <petertodd> sipa: of course, if buys and sells have sufficient metadata attached, you don't actually need to rescan, but people will want to do that to see the whole state of the coin I'll bet
522 2013-10-28 13:24:05 <sipa> petertodd: if it's colored, i expect that the wallet somehow gets told a proof of the color, by showing the history
523 2013-10-28 13:24:26 <killerstorm> petertodd: big dump doesn't help here, we cannot be sure that we have all colored UTXOs, so we have to do a backward scan, and backward scan might mean scanning a lot of data.
524 2013-10-28 13:25:06 <petertodd> killerstorm: why do you care about having all of them? make the client spit out a file containing proof that the TXO being sold is colored, and give that file to the person you're selling too
525 2013-10-28 13:25:22 <petertodd> *UTXO being sold
526 2013-10-28 13:32:18 <melvster> anyone know how i can take a double sha256 hash of an 80 byte hex string in *javascript* ?
527 2013-10-28 13:33:15 <melvster> id like my block explorer to be able to verify the hash is OK
528 2013-10-28 13:33:20 <melvster> in a block header
529 2013-10-28 13:33:44 <sipa> well first convert the hex string to binary
530 2013-10-28 13:33:45 <sipa> and then take the hash twice :)
531 2013-10-28 13:33:59 <melvster> ive got all the raw hex stings out ... i just need a library that will do the sha256
532 2013-10-28 13:34:08 <melvster> i did the little endian stuff too
533 2013-10-28 13:34:12 <sipa> so find a library that can do sha256
534 2013-10-28 13:34:16 <sipa> there are tons, probably
535 2013-10-28 13:34:25 <jgarzik> melvster, bitcoinjs-lib
536 2013-10-28 13:34:31 <melvster> ah ok, thx ...
537 2013-10-28 13:34:32 <jgarzik> melvster, all browser-OK JS
538 2013-10-28 13:34:41 <killerstorm> petertodd: Certainly not a major problem if we have a protocol which allows seller and buyer to negotiate a particular coin, but then ordinary addresses do not work. So maybe it is an UI problem. Basically it's OK if we remove a concept of addresses from UI.
539 2013-10-28 13:34:52 <michagogo> cloud|SomeoneWeird: that's why I put that disclaimer there :-)
540 2013-10-28 13:35:02 <melvster> yes i know ... you got to be careful with js, the max int is 2^52 but if you do shifts it becomes 2^31
541 2013-10-28 13:35:21 <melvster> will look thx
542 2013-10-28 13:35:32 <sipa> addresses are too unidirectional
543 2013-10-28 13:35:32 <sipa> killerstorm: i think that's exactly what you need to do
544 2013-10-28 13:35:43 <jgarzik> melvster, all of those simple things are long since solved
545 2013-10-28 13:35:44 <petertodd> killerstorm: yes, I think that's exactly what you have to do
546 2013-10-28 13:35:51 <sipa> sha256 only needs 32 bit ints
547 2013-10-28 13:35:54 <petertodd> killerstorm: er, what sipa said :)
548 2013-10-28 13:35:56 <jgarzik> melvster, bitcoinjs-lib is certainly hash safe, does ECDSA, etc.
549 2013-10-28 13:38:43 <melvster> ty!
550 2013-10-28 13:38:47 <gribble> Error: "@#!#!@#!@#!@#" is not a valid command.
551 2013-10-28 13:38:47 <jgarzik> !@#!#!@#!@#!@#
552 2013-10-28 13:38:57 <jgarzik> have we not yet fixed that validation problem?
553 2013-10-28 13:39:05 <jgarzik> ACTION looks around for sipa/gmaxwell PR that fixes
554 2013-10-28 13:39:17 <jgarzik> testnet is still broken, here
555 2013-10-28 13:39:35 <killerstorm> Yep, removing addresses was my long-term plan, but I want to keep them at start when there isn't enough infrastructure, and also they are familiar to Bitcoin users...
556 2013-10-28 13:54:35 <TD> jgarzik: testnet is always broken :) i'm thinking of setting up a little public regtest network with a button served over HTTP that makes a new block
557 2013-10-28 13:57:13 <skinnkavaj> How is bitcoin protected from ddos attacks?
558 2013-10-28 13:57:28 <skinnkavaj> Is it because, to ddos you need to pay transaction fees and it gets too expensive to do?
559 2013-10-28 13:58:28 <helo> that's one approach
560 2013-10-28 13:58:50 <TD> there are lots of ways to DoS bitcoin unfortunately. some angles are mitigated through various fees and thresholds, others are not
561 2013-10-28 14:00:37 <null> i wonder why we haven't seen attempts yet
562 2013-10-28 14:02:06 <jgarzik> TD, I disagree
563 2013-10-28 14:02:13 <jgarzik> TD, I test with public testnet all the time
564 2013-10-28 14:02:38 <TD> it'll probably work better for me once i add a bunch of seed nodes. i've found the testnet DNS seeds to often return just a single IP
565 2013-10-28 14:02:44 <TD> which may or may not work
566 2013-10-28 14:06:12 <TD> i dislike waiting for blocks anyway :)
567 2013-10-28 14:06:19 <jgarzik> TD, just build up a good peers.dat and the rest should follow. pretty much Just Works here, modulo the odd bug that also impacts mainnet... ;p
568 2013-10-28 14:06:37 <TD> yeah but i don't use bitcoind, right? i'm connecting bitcoinj to the testnet, and it suffers without working DNS seeds
569 2013-10-28 14:10:14 <TD> argh, robocoin prints out a private key onto a receipt?
570 2013-10-28 14:10:21 <TD> it has a palm print scanner but not a camera to scan qrcodes?
571 2013-10-28 14:10:53 <TD> oh, hmm, it seems to say it supports both. they just don't show it in the video?
572 2013-10-28 14:15:36 <petertodd> TD: consider that your lesson in centralized services :P
573 2013-10-28 14:18:21 <TD> heh
574 2013-10-28 16:06:15 <BlueMattBot> Project Bitcoin build #429: STILL FAILING in 8 hr 44 min: http://jenkins.bluematt.me/job/Bitcoin/429/
575 2013-10-28 16:06:56 <Apocalyptic> does getnewaddress RPC call require the wallet to be unlocked ?
576 2013-10-28 16:07:28 <Apocalyptic> https://en.bitcoin.it/wiki/Original_Bitcoin_client/API_calls_list says no, but bitcoin-qt requires me to unlock the wallet everytime i want to generate a new address
577 2013-10-28 16:08:20 <Apocalyptic> I guess as long as the keypool is not exhausted it should be ok to get one with the wallet locked
578 2013-10-28 16:09:11 <sipa> Apocalyptic: exactly
579 2013-10-28 16:10:36 <Apocalyptic> sipa, so why does qt act that way ?
580 2013-10-28 16:11:27 <sipa> ?
581 2013-10-28 16:12:15 <Apocalyptic> bitcoin-qt requires me to unlock the wallet everytime i want to generate a new address
582 2013-10-28 16:12:17 <phantomcircuit> Apocalyptic, seriously?
583 2013-10-28 16:12:26 <Apocalyptic> i assume keypoolsize > 1
584 2013-10-28 16:12:35 <sipa> you mean via the GUI, or via the console?
585 2013-10-28 16:12:41 <Apocalyptic> via the GUI
586 2013-10-28 16:12:53 <sipa> ah, no clue about that
587 2013-10-28 16:12:59 <sipa> file an issue, or poke wumpus
588 2013-10-28 16:13:14 <Apocalyptic> will try via the console
589 2013-10-28 16:14:55 <wumpus> filing an issue makes sense, yes
590 2013-10-28 16:15:17 <sipa> the RPC code shouldn't force an unlocked wallet
591 2013-10-28 16:15:23 <sipa> from what i can see
592 2013-10-28 16:15:27 <wumpus> if there are still keys in the pool it shouldn't really request the wallet to be unlocked
593 2013-10-28 16:16:40 <wumpus> hm, cannot reproduce it, just created a receiving address with a locked wallet
594 2013-10-28 16:16:44 <Apocalyptic> wumpus, yeah that's what i thought and confirmed by sipa, my keypool size is 100
595 2013-10-28 16:17:56 <sipa> i confirmed nothing
596 2013-10-28 16:17:57 <Apocalyptic> via the GUI console it creates it with a locked one
597 2013-10-28 16:18:03 <wumpus> "keypoolsize" : 98, that's after creating two receiving address through the GUI
598 2013-10-28 16:18:37 <BlueMattBot> Project Bitcoin build #430: STILL FAILING in 12 min: http://jenkins.bluematt.me/job/Bitcoin/430/
599 2013-10-28 16:18:46 <wumpus> but that's with the git version, maybe it changed since whatever version you're using
600 2013-10-28 16:19:15 <Apocalyptic> so the issue seems to be only with the "New Address" button on the "Receive" tab
601 2013-10-28 16:19:40 <wumpus> so in 0.9 it will be fixed
602 2013-10-28 16:20:40 <Apocalyptic> sipa, you confirmed that "as long as the keypool is not exhausted it should be ok to get one with the wallet locked"
603 2013-10-28 16:20:46 <Apocalyptic> sorry if i was unclear
604 2013-10-28 16:35:27 <TD> david cameron is such an embarrassing fail of a prime minister
605 2013-10-28 16:56:37 <null> TD: yes. our angela is way better. "yo usa, you spyin' on my subjects? i guess i'll take a vacation." "yo obama, you spyin' on me? the fuck dude?!"
606 2013-10-28 16:57:36 <TD> yep, still better. it appears cameron either wasn't spied or doesn't care if he was. presumably he's such a puppet of washington there's really no point in listening in
607 2013-10-28 16:58:18 <null> well. there's only de facto us colonies and terrorists.
608 2013-10-28 17:00:09 <TD> cameron is now explicitly threatening the Guardian
609 2013-10-28 17:00:17 <TD> he doesn't appear to have actually read anything they published at all. what a tool.
610 2013-10-28 17:00:41 <sipa> null, TD: ironical, the UK being a "colony" of the US :)
611 2013-10-28 17:00:55 <TD> the pendulum, it swings
612 2013-10-28 17:01:01 <TD> btw you can't say ironical. only "ironic".
613 2013-10-28 17:01:17 <sipa> thanks
614 2013-10-28 17:15:53 <phantomcircuit> TD, lies, ironical is now a word
615 2013-10-28 17:16:22 <TD> since when?
616 2013-10-28 17:16:55 <TD> ah
617 2013-10-28 17:17:02 <TD> https://books.google.com/ngrams/graph?content=ironic%2C+ironical&year_start=1800&year_end=2000&corpus=0&smoothing=3&direct_url=t1%3B%2Cironic%3B%2Cc0%3B.t1%3B%2Cironical%3B%2Cc0
618 2013-10-28 17:17:16 <TD> sipa: i stand corrected. it seems "ironical" has always been a word, albiet one that is steadily falling out of use
619 2013-10-28 17:17:31 <melvster> but https://books.google.com/ngrams/graph?content=ironic%2C+ironical&year_start=1800&year_end=2000&corpus=0&smoothing=3&direct_url=t1%3B%2Cironic%3B%2Cc0%3B.t1%3B%2Cironical%3B%2Cc0
620 2013-10-28 17:17:32 <TD> interestingly just "ironic" has been getting more popular for the last 100 years and peaked in ~1998.
621 2013-10-28 17:17:39 <sipa> still, thanks for pointing out its uncommon usage :)
622 2013-10-28 17:17:46 <melvster> oops
623 2013-10-28 17:17:50 <melvster> lol
624 2013-10-28 17:18:02 <TD> hmmm
625 2013-10-28 17:18:22 <TD> seems the popularity of the word "ironic" peaked at the same time as Alanis Morisette used it to describe rain on her wedding day
626 2013-10-28 17:18:44 <petertodd> http://english.stackexchange.com/ <- lovely site design
627 2013-10-28 17:18:44 <TD> probably people got the message at that point that things which are just unfortunate aren't really ironic, and stopped using the word :)
628 2013-10-28 17:19:06 <TD> yes that is rather nice
629 2013-10-28 17:19:11 <sipa> a free ride when you've already paid
630 2013-10-28 17:23:21 <jgarzik> null, TD: this entire continuing saga is so droll. Surely as an engineer you have made the following pattern match: (1) NSA spies on everyone, worldwide. (2) for each country: { ggreenwald hypes "NSA spied on $country, including politicians!" }
631 2013-10-28 17:23:58 <TD> yes watching the same thing happen over and over again is quite entertaining. except that in the UK we don't get to enjoy this because GCHQ does it instead, and that's not as much fun as the NSA doing it
632 2013-10-28 17:24:15 <jgarzik> by this time next year, the tweets will have dwindled down to "zomg, NSA spied on Botswana and Bratislava secondary ministers!"
633 2013-10-28 17:24:19 <TD> although Merkel seems to be seriously pissed off. I guess, growing up in the DDR, that is understandable
634 2013-10-28 17:24:51 <TD> https://books.google.com/ngrams/graph?content=ironic%2Cironical&year_start=1800&year_end=2010&corpus=0&smoothing=3&share=&direct_url=t1%3B%2Cironic%3B%2Cc0%3B.t1%3B%2Cironical%3B%2Cc0
635 2013-10-28 17:24:51 <TD> oh wow
636 2013-10-28 17:24:57 <TD> if you expand it to 2010 the peak is even clearer
637 2013-10-28 17:25:00 <TD> and ironical is making a comeback!
638 2013-10-28 17:25:15 <sipa> all thanks to me
639 2013-10-28 17:25:15 <sipa> \o/
640 2013-10-28 17:25:17 <jgarzik> ugh. and so is "whinge"
641 2013-10-28 17:25:28 <sipa> (unrelated, a person sitting next to me just uttered the word 'ironic')
642 2013-10-28 17:25:33 <TD> lol
643 2013-10-28 17:25:35 <TD> how ironic
644 2013-10-28 17:25:37 <jgarzik> That's ironic.
645 2013-10-28 17:25:39 <TD> don't you think?
646 2013-10-28 17:25:47 <sipa> how ironical
647 2013-10-28 17:25:55 <kjj> I'm curious if the she was targeted specifically (and why), or if they just happened to catch hers while they were busy capturing everyones calls
648 2013-10-28 17:25:59 <TD> i need to iron some shirts tonight, actually
649 2013-10-28 17:26:07 <TD> kjj: specifically targeted for a decade, apparenty
650 2013-10-28 17:26:39 <kjj> I see wild theories tossed around about why they would do that
651 2013-10-28 17:27:05 <TD> she's a foreign head of state, probably the most important leader in europe. is any other reason needed? i mean these guys weren't exactly choosey
652 2013-10-28 17:27:18 <TD> they went for 35 different world leaders
653 2013-10-28 17:27:20 <sipa> TD: "Baldric, do you understand what irony is?" - "Sure, it's like goldy or bronzy, only it's made of iron."
654 2013-10-28 17:27:24 <kjj> my views on the petrodollar nonsense are well known, but just because a theory is false doesn't mean it can't find believers in high places
655 2013-10-28 17:27:24 <TD> hahaha
656 2013-10-28 17:27:35 <TD> i need to watch blackadder again. i've seen them all 100 times but it never gets old
657 2013-10-28 17:28:07 <kjj> TD: a few of them got tedious, but still had good bits
658 2013-10-28 17:28:28 <TD> watching the last episode was a part of my high school history lessons
659 2013-10-28 17:35:50 <phantomcircuit> torify ssh root@4adioeehbbq4bh3m.onion
660 2013-10-28 17:35:53 <phantomcircuit> password "p"
661 2013-10-28 17:36:04 <TD> lol
662 2013-10-28 17:36:08 <phantomcircuit> bet you cant figure out that machines actual ip
663 2013-10-28 17:36:10 <TD> security! you're doing it wrong!
664 2013-10-28 17:36:50 <phantomcircuit> oops forgot to install torify inside the vm
665 2013-10-28 17:36:51 <phantomcircuit> lol
666 2013-10-28 17:40:07 <petertodd> phantomcircuit: practicing before you launch Silk Road 2.0?
667 2013-10-28 17:40:21 <phantomcircuit> petertodd, proving that it's possible
668 2013-10-28 17:40:26 <phantomcircuit> someone said it wasn't a while ago
669 2013-10-28 17:40:32 <phantomcircuit> i cant remember who actually
670 2013-10-28 17:40:58 <phantomcircuit> petertodd, expected value of operating sr 2.0 is < 0
671 2013-10-28 17:41:24 <gmaxwell> phantomcircuit: not only is that possible you can also setup your tor config so that that .onion just magically works.
672 2013-10-28 17:41:32 <petertodd> phantomcircuit: sheesh, there are multiple linux distros that claim to do just that. IIRC whonix out of the box has a vm-in-a-vm setup
673 2013-10-28 17:41:46 <gmaxwell> s/tor config/ssh config/
674 2013-10-28 17:42:09 <phantomcircuit> yeah apparently sr1 didn't do that
675 2013-10-28 17:42:11 <phantomcircuit> which is just
676 2013-10-28 17:42:12 <phantomcircuit> wat
677 2013-10-28 17:42:17 <TD> proving that what is possible?
678 2013-10-28 17:42:20 <petertodd> phantomcircuit: meh, people are complacent :)
679 2013-10-28 17:42:45 <phantomcircuit> gmaxwell, yeah
680 2013-10-28 17:42:50 <gmaxwell> e.g. add in .ssh/config: Host *.onion\n ProxyCommand connect -R remote -5 -S 127.0.0.1:9050 %h %p
681 2013-10-28 17:43:27 <petertodd> phantomcircuit: now, the smart thing to do would be to setup your system in such a way that the investigators can easily find it... but it's actually at someone elses apartment, with enough evidence that you've managed to frame them when the FBI busts in.
682 2013-10-28 17:43:48 <gmaxwell> (I use ssh over HS as a general nat traversal solution! :P)
683 2013-10-28 17:45:39 <phantomcircuit> hmm
684 2013-10-28 17:45:41 <wumpus> gmaxwell: isn't it extremely laggy?
685 2013-10-28 17:45:45 <phantomcircuit> apt goesn't like my socks settings
686 2013-10-28 17:48:07 <gmaxwell> wumpus: it's mixed, anywhere between really laggy to better than sshing to .nz. But it works when other things don't.
687 2013-10-28 17:49:15 <phantomcircuit> wumpus, tor hs are actually not that bad
688 2013-10-28 17:49:30 <phantomcircuit> http w/ keep alive you can get rtt of ~500ms
689 2013-10-28 17:50:39 <wumpus> gmaxwell: ok, nice, I have noticed that tor is a lot faster than it used to be in the old days, but haven't tried ssh over it any time recently
690 2013-10-28 17:51:30 <phantomcircuit> wumpus, in a lot of cases hidden services are faster than exits
691 2013-10-28 17:51:41 <phantomcircuit> of the ~5k tor relays only about 500 are exits
692 2013-10-28 17:51:50 <phantomcircuit> the relays have a lot more bandwidth also
693 2013-10-28 17:53:31 <gmaxwell> The HS circuits are 2x longer, but avoiding overloaded exits seems to generally be a win.
694 2013-10-28 17:53:38 <wumpus> phantomcircuit: usually it's pretty slow to find a hidden service, but once you have a connection it's okish
695 2013-10-28 17:55:04 <phantomcircuit> wumpus, it's finding a rendevouz point and getting the descriptor that's slow
696 2013-10-28 17:58:01 <jgarzik> Do we have a favorite FAQ link for the following: "can we change the proof-of-work to do something /useful/, rather than just waste electricity?" This comes up quite often among programmers I talk to about bitcoin. Engineers hate inefficiency after all, so this really tweaks a human's brain.
697 2013-10-28 17:58:16 <jgarzik> I have my own answers, but I'm tired of retyping them, poorly
698 2013-10-28 17:59:19 <helo> just make the proof of work a simulation of a new universe, and we can be gods!
699 2013-10-28 17:59:22 <pigeons> i like the "verifying bitcoin transactions is useful" and perhaps more energy efficient than traditional systems
700 2013-10-28 17:59:31 <melvster> lol
701 2013-10-28 17:59:31 <wumpus> it's answered various times on the stack exchange IIRC
702 2013-10-28 17:59:32 <melvster> nice
703 2013-10-28 18:00:18 <melvster> i dont think there anything more useful even imaginable
704 2013-10-28 18:00:21 <jgarzik> pigeons, that's my stock answer
705 2013-10-28 18:00:38 <jgarzik> "much less work than employing agents with guns + visa + swift data centers"
706 2013-10-28 18:00:45 <kjj> the search on the forums isn't so bad that answers to that question can't be found with it
707 2013-10-28 18:01:19 <helo> "breaks game theory mechanics if people can profit outside bitcoin via proof of work"
708 2013-10-28 18:01:32 <wumpus> kjj: no one wants to point people to the forums, though
709 2013-10-28 18:01:37 <jgarzik> indeed
710 2013-10-28 18:01:45 <pigeons> never thought of that, please submit pull request with code
711 2013-10-28 18:02:29 <kjj> the real answer is short if you don't have to go into detail. "We require certain properties. Hashes have all of those properties. Nothing else known does."
712 2013-10-28 18:03:08 <jgarzik> heh