1 2015-02-26 00:20:20 <Luke-Jr> wtf, users can create hardlinks to files owned by other users and which they can't read/write O.o
  2 2015-02-26 00:22:49 <sipa> of course
  3 2015-02-26 00:23:09 <sipa> all you need is access to the directory in which yiu are creating the link
  4 2015-02-26 00:23:30 <sipa> creating a hardlink does not access the file itself
  5 2015-02-26 00:23:49 <Luke-Jr> sipa: it creates security issues in many cases
  6 2015-02-26 00:24:58 <Luke-Jr> compromise somedaemon, do `ln /etc/passwd /var/log/somedaemon/error.log`, now when the system reboots and somedaemon's init script fixes the permissions for error.log, it's changing /etc/passwd
  7 2015-02-26 00:25:26 <sipa> hearn: the JNI bindings can use an update, but pull requests welcome
  8 2015-02-26 00:26:07 <sipa> hearn: also, yes, tested on ARM and wumpus even wrote asm for it (not merged yet); i'd consider arm support even more experimental than the rest, though
  9 2015-02-26 00:26:13 <hearn> nice
 10 2015-02-26 00:27:03 <sipa> hearn: the JNI code will need to change significantly when we introduce explicit contexts, and get rid of the globals (see some other pull)
 11 2015-02-26 00:27:10 <hearn> ok
 12 2015-02-26 00:27:13 <hearn> yes i saw
 13 2015-02-26 00:27:17 <cfields> sipa: i was working on updating the jni code, that's what sparked the discussion :)
 14 2015-02-26 00:27:45 <cfields> sipa: question was about the contexts and supporting just the flags and/or setter functions as part of the java api
 15 2015-02-26 00:28:08 <cfields> i think i'm just going to ignore the setters and require the flags during construction. that work for you?
 16 2015-02-26 00:29:15 <sipa> cfields: the JNi code currently only supports validation anyway
 17 2015-02-26 00:31:03 <sipa> so that's fine
 18 2015-02-26 00:31:12 <cfields> sipa: well i was creating a small raii context wrapper class to be a little future-proof. You want me to just hard-code it to verification for now?
 19 2015-02-26 00:31:51 <sipa> i think the api will need to change later anyway, but fell free
 20 2015-02-26 00:32:16 <cfields> ok, i'll just keep it simple then
 21 2015-02-26 00:34:21 <instagibbs> ok back with gitian q's again. running first build line and it's breaking. Log shows: http://pastebin.com/ZbbbAKfs
 22 2015-02-26 00:34:48 <instagibbs> wget appears to be able to find samba.org, at least, when I run it manually on the vm
 23 2015-02-26 00:36:02 <instagibbs> my bad, sipa asked about this few months ago.
 24 2015-02-26 00:39:16 <instagibbs> wumpus: can you explain what you meant by "make download": https://bitcoinstats.com/irc/bitcoin-dev/logs/2014/12/04#l1417709019
 25 2015-02-26 00:40:11 <sipa> instagibbs: type into your shell the command "make download" in the depends dir in the gitian host
 26 2015-02-26 00:43:31 <instagibbs> thanks! (I should probably write down all the issues I'm having to supplement tutorial)
 27 2015-02-26 02:59:02 <instagibbs> sipa: anything else I have to do? Downloads seem to have happened, yet same error.
 28 2015-02-26 03:18:35 <instagibbs> hold that thought
 29 2015-02-26 04:55:12 <fanquake> ;;blocks
 30 2015-02-26 04:55:13 <gribble> 345194
 31 2015-02-26 05:38:29 <earlz> So, sometime ago I got a link to some quick performance characteristics of libsecp256
 32 2015-02-26 05:39:03 <earlz> I'm curious to know how fast it can sign and verify on modern hardware
 33 2015-02-26 05:42:39 <sipa> let's say around 260k cpu cycles for an ecdsa verification
 34 2015-02-26 05:43:01 <sipa> it depends on which optimalizations are enabled
 35 2015-02-26 05:43:20 <sipa> and whether assembly routines are used
 36 2015-02-26 05:43:35 <earlz> hmm
 37 2015-02-26 05:43:56 <earlz> Assume fastest I guess.. 260K.. around how many ms would that be? <1?
 38 2015-02-26 05:43:57 <sipa> but it's that ballpark for modern i7 cpu's in 64bit mode
 39 2015-02-26 05:44:01 <Diablo-D3> https://www.battleforthenet.com/
 40 2015-02-26 05:44:15 <sipa> depends on the clock speed
 41 2015-02-26 05:44:41 <sipa> ;;calc 260000 / 3000000000 * 1000
 42 2015-02-26 05:44:42 <gribble> 0.0866666666667
 43 2015-02-26 05:44:51 <sipa> that many ms on a 3 GHz
 44 2015-02-26 05:44:53 <earlz> well of course heh. But I mean just a simple ballpark figure
 45 2015-02-26 05:45:05 <earlz> ah
 46 2015-02-26 05:45:14 <Diablo-D3> sipa: assuming the memory bus is fast enough if everything cant be prefetched
 47 2015-02-26 05:46:11 <sipa> signing is a bit faster, but libsecp256k1 hasn't been particularly optimized for signing speed, more for avoiding sidechannel attacks
 48 2015-02-26 05:46:40 <earlz> faster would make it more vulnerable to sidechannel attacks?
 49 2015-02-26 05:46:49 <sipa> no
 50 2015-02-26 05:47:24 <sipa> but protecting against sidechannel attacks typically means slower constant-time algorithms
 51 2015-02-26 05:47:31 <earlz> ah
 52 2015-02-26 05:52:11 <gmaxwell> signing speed is almost always unimportant.
 53 2015-02-26 05:53:23 <gmaxwell> where is tens of thousands of signatures per second not enough?
 54 2015-02-26 07:15:23 <kska> hello all
 55 2015-02-26 08:41:09 <wumpus> gmaxwell: agreed, I can only think of malicious reasons for fast signing
 56 2015-02-26 08:41:56 <wumpus> gmaxwell: oh maybe: extremely low-end hardware wallets
 57 2015-02-26 11:09:44 <harding> wumpus: you around?  I'd like to enable Travis CI test builds of the Bitcoin.org website and I seem to have the permissions, but cfields smartly suggested I do it when you're around in case I somehow mess up the queues for other github.com/bitcoin projects.
 58 2015-02-26 11:20:52 <wumpus> harding: sounds good to me
 59 2015-02-26 11:21:22 <sipa> i'm around too if there are problems
 60 2015-02-26 11:22:55 <harding> wumpus, sipa: thanks!  I'll start now.
 61 2015-02-26 11:40:31 <harding> wumpus, sipa: the first build job was successful.  Thanks again!
 62 2015-02-26 11:40:37 <wumpus> great!
 63 2015-02-26 12:05:33 <nbenoit> Hi, I submitted PR 5823 (Replaced current function names with __func__ in LogPrintf() calls, in main.cpp), it was merged into the trivial tree. Shall I proceed similarly with other files ? Maybe submit to the trivial-next branch directly ?
 64 2015-02-26 12:08:54 <wumpus> nbenoit: if it is a similar change (e.g. messages only), yes, please submit to trivial-next branch directly
 65 2015-02-26 12:13:21 <nbenoit> wumpus: all right, will do. Is it better to submit one PR per file or can I do a single PR for multiple files ?
 66 2015-02-26 12:15:10 <wumpus> combining similar changes is easiest for reviewing
 67 2015-02-26 12:17:23 <nbenoit> wumpus: ok, thanks
 68 2015-02-26 12:17:34 <gmaxwell> A commit should be whatever is thought of as a single 'logical' change. Doesn't matter how many files are touched.  Things should be organized to minimize review as the first considersation, and to simplify history management (Rebasing, bisecting, etc.) as a second consideration.
 69 2015-02-26 12:18:36 <sipa> In case you suspect that some part of the changes may require separate review (may be more controversial, for example), better split them up (though this shouldn't apply for trivial patches, which should always be completely uncontroversial)
 70 2015-02-26 12:21:54 <xabbix> gmaxwell, can you say if Evil-Kenivel actually found a bug that allowed easy DoS on bitcoin-core?
 71 2015-02-26 12:24:26 <nbenoit> gmaxwell, sipa: ok, have a nice day !
 72 2015-02-26 12:25:19 <fanquake> wumpus What’s your new GH avatar? An angry bee?
 73 2015-02-26 12:33:15 <wumpus> fanquake: yes, sort of :-) this was the "busy" mouse cursor on Atari ST
 74 2015-02-26 12:37:41 <wumpus> in modern UIs they use a hourglass or spinner for that, but a bee is of course much cooler
 75 2015-02-26 12:38:02 <fanquake> wumpus yes definitely. Very nice.
 76 2015-02-26 12:50:41 <wumpus> incidentally, if that computer crashed it'd show a row of bombs (depending on the error code) they should bring that back too. https://en.wikipedia.org/wiki/Bomb_%28symbol%29#Atari_ST_TOS
 77 2015-02-26 12:52:49 <sipa> unary encoding for error codes
 78 2015-02-26 12:52:50 <sipa> ha!
 79 2015-02-26 12:54:50 <instagibbs> BSoD is the only way to go *fond memories of Windows ME*
 80 2015-02-26 12:55:45 <wumpus> "The very first version of TOS used mushroom clouds; this was quickly changed, as it was considered politically incorrect." hehe, these bombs are much more politically incorrect these days
 81 2015-02-26 12:56:45 <wumpus> sipa: I must say I never saw more than ~5 at a time, had no idea
 82 2015-02-26 15:21:14 <coderwill> Looks like Circle is having some type of issue, homepage 504, android app in maintenance mode.
 83 2015-02-26 15:28:18 <coderwill> ...and it's back up.
 84 2015-02-26 15:28:37 <sipa> coderwill: off topic here :)
 85 2015-02-26 15:52:38 <coderwill> sipa: apologies!
 86 2015-02-26 19:15:43 <Luke-Jr> is it impossible to show addresses in the tx list in 0.10 :/
 87 2015-02-26 19:33:32 <Luke-Jr> wumpus: would you be opposed to reverting https://github.com/bitcoin/bitcoin/pull/4610 ?
 88 2015-02-26 19:33:59 <Luke-Jr> not sure what the reasoning was - the removed feature does do what it says :|
 89 2015-02-26 19:39:05 <hjkjh> Hello can anyone help me with something, i install bitcoin qt and have the blockchain synced. Transfered some bitcoin to an address but the transaction did not show up in the qt. I could do "dumpprivkey" command for the address without an issue but i don't see a balance. Localbitcoin shows 2 confirmations last i checked
 90 2015-02-26 19:39:10 <hjkjh> Transfered some bitcoin to an address in my bitcoin qt*
 91 2015-02-26 19:39:15 <hjkjh> from somewhere
 92 2015-02-26 19:39:20 <hjkjh>  am i missing something? i don't want to redownload the blockchain its several gb
 93 2015-02-26 19:40:01 <hjkjh> please help
 94 2015-02-26 19:40:36 <kadoban> hjkjh: Pick a channel, don't ask in both. Sounds more appropriate for #bitcoin, where you're already receiving responses anyway.
 95 2015-02-26 19:40:57 <hjkjh> kadoban, due to the lack of response i posted here.
 96 2015-02-26 20:08:44 <maaku> hjkjh: for future reference, you will not get a response here
 97 2015-02-26 20:16:58 <hjkjh> ok thanks
 98 2015-02-26 20:53:06 <Luke-Jr> can I get some testnet to muhSokhfvS2RiEFCiC1HjihhZ6FNuevKea ?
 99 2015-02-26 20:53:14 <belcher> yeah sec
100 2015-02-26 20:55:02 <Luke-Jr> thx
101 2015-02-26 20:55:49 <belcher> chance to show off the coinjoin tx
102 2015-02-26 21:12:06 <ajweiss> how weird.  last night i was explaining what the bulletin board world used to be like to a few friends and mentioned the atari st.  (in the back of my head i briefly thought about the similarity between getting numbers for boards and getting bitcoin)  then today, lo and behold, wladimir has turned into the gem bee.
103 2015-02-26 22:13:17 <gmaxwell> https://bitcointalk.org/index.php?topic=937058.msg10593623#msg10593623
104 2015-02-26 22:19:34 <Luke-Jr> odd
105 2015-02-26 22:23:35 <justanotheruser> too bad bitcoin uses c++, if we used a language without complex stuff like strong typing it would be easier to read and less buggy
106 2015-02-26 22:23:50 <cfields> gmaxwell: huh
107 2015-02-26 22:27:48 <helo> ACTION throws an old tomato at justanotheruser
108 2015-02-26 22:42:24 <ahmed_> hi all does anyone here know how i can use a python script to run bitcoind as another user?