1 2012-05-06 01:20:36 <jgarzik> (in rpcsrv, not bitcoin)
  2 2012-05-06 03:09:57 <gribble> New news from bitcoinrss: randomproof opened pull request 1208 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1208>
  3 2012-05-06 03:26:29 <BlueMatt> sipa: ping
  4 2012-05-06 03:26:55 <gmaxwell> BlueMatt: kinda early in sipatime.
  5 2012-05-06 03:27:18 <BlueMatt> gmaxwell: well his whois isnt disconnected...
  6 2012-05-06 03:27:35 <BlueMatt> but...meh, at least he'll ping when hes back
  7 2012-05-06 03:28:04 <gmaxwell> BlueMatt: can you get whatever nodes you're running up on the addrman fixes? I just pulled them into master.
  8 2012-05-06 03:29:09 <gmaxwell> I'd like us to cut a release fairly soon with them, but I think they need a pretty good number of total userhours baking. (in particular, I'm not eager to cut a release that replaces a rare crash with a somewhat less rare crash-due-to-overeager-asserts)
  9 2012-05-06 03:29:24 <BlueMatt> gmaxwell: any clue why this tiny snippet doesnt unlock properly (boost "Assertion `res == 0' failed.")
 10 2012-05-06 03:29:37 <BlueMatt> gmaxwell: sadly, I have no permanently up nodes...
 11 2012-05-06 03:29:42 <gmaxwell> tisk tisk
 12 2012-05-06 03:30:21 <BlueMatt> https://github.com/TheBlueMatt/bitcoin/commit/03410e0fd5cbdd10a28b13980982d10c56b45de5#diff-5
 13 2012-05-06 03:30:26 <BlueMatt> oops, forgot the link...
 14 2012-05-06 03:38:26 <luke-jr> gmaxwell: deploying it on Eligius now
 15 2012-05-06 03:43:26 <dub> error: {"code":-4,"message":"Error: Transaction creation failed  "}
 16 2012-05-06 03:43:51 <dub> whats the largest amount to be sent in one tx?
 17 2012-05-06 03:44:22 <gmaxwell> dub: there isn't a maximum amount of bitcoin, there is a maximum amount of _data_.
 18 2012-05-06 03:44:48 <gmaxwell> so if you have a wallet with a zillion tiny inputs you can bump into that data limit.
 19 2012-05-06 03:45:15 <dub> roger
 20 2012-05-06 03:45:54 <gmaxwell> Is that something you hit on testnet or real bitcoin?
 21 2012-05-06 03:46:04 <dub> i0coin :)
 22 2012-05-06 03:47:34 <gmaxwell> ltc carries a patch to control the minimum size inputs that get considered for use.
 23 2012-05-06 03:47:56 <Anon007> Santoshi Created the Bitcoin to help the Little Guy its not meant for use by the Corporations and it is Geared Towards People who are Honest and Trustworthy... Not the ELITES Stay Away from our Money You Print Your Own DEBT Backed Money... Don't Let BitCoins Become Centralized ANONYMOUS is Watching!
 24 2012-05-06 03:48:30 <dub> coolstorybro
 25 2012-05-06 03:48:36 <DiabloD3> csb
 26 2012-05-06 03:48:37 <Anon007> Its The Truth
 27 2012-05-06 03:49:38 <luke-jr> Anon007: how about instead of watching, Anonymous write code to address the scaling issues they're whining about?
 28 2012-05-06 03:49:59 <dub> ITT: Truth
 29 2012-05-06 03:50:28 <Joric> 'Don't Let BitCoins Become Centralized' is there a way to let them become centralized?
 30 2012-05-06 03:50:59 <paulo_> Anon007: are you also AnonMiner?
 31 2012-05-06 03:51:04 <Anon007> You have a Point there but you must remember anyone can say they are ANONYMOUS!
 32 2012-05-06 03:51:20 <Anon007> You Caught me Paulo :)
 33 2012-05-06 03:51:32 <DiabloD3> Anon007: stop trolling dude
 34 2012-05-06 03:52:11 <Anon007> I am not Anonymous I am just another guy that is pissed off with Elites Trying to Control the 99% and I won't Stand for it Anymore!
 35 2012-05-06 03:53:04 <Anon007> Sorry Devil I didn't mean to Burst a Blood Vessel I am just Voicing my Opinion am I allowed to have my own Thoughts and not be accused of Trolling???
 36 2012-05-06 03:53:15 <gmaxwell> Less talk, more code.
 37 2012-05-06 03:53:30 <DiabloD3> jarly
 38 2012-05-06 03:56:13 <paulo_> lol Devil
 39 2012-05-06 03:57:45 <wumpus> gmaxwell: right!
 40 2012-05-06 03:58:27 <paulo_> as Satoshi said, "I'm better with code than with words though."
 41 2012-05-06 03:59:34 <Anon007> Diablo Translates to Spanish for Devil I speak English so I simply Translated it :P
 42 2012-05-06 03:59:43 <Joric> he must be terrible with words
 43 2012-05-06 03:59:49 <Anon007> LOL
 44 2012-05-06 04:00:26 <gmaxwell> Joric: When you show that you can write gabling sites that I don't instantly see how to exploit then you can complain about Satoshi's code.
 45 2012-05-06 04:01:06 <paulo_> heh.
 46 2012-05-06 04:01:08 <Anon007> Santoshi is a Genius and he needs to take control of his invention and not allow it to fall into the wrong hands... Max Keiser is Awesome and he is On Board with the BitCoin and he says we can Crash JP Morgan by simply Mining BitCoin and Buying Physical Silver!
 47 2012-05-06 04:02:10 <gmaxwell> Anon007: taking control is sort of the opposite of what makes bitcoin worthwhile.
 48 2012-05-06 04:02:12 <wumpus> yea yea it needs to be more decentralized but a single person needs to take control of it...
 49 2012-05-06 04:02:56 <wumpus> did a clown escape from #bitcoin?
 50 2012-05-06 04:03:02 <luke-jr> paulo_: except it turns out Satoshi wasn't too good with code either! <.<
 51 2012-05-06 04:03:10 <Anon007> Yes you are correct gmaxwell but when Gavin said he has no interest in making the bitcoin system easier for anyone to mine he is saying that he is focused on the money and not helping the 99% :(
 52 2012-05-06 04:03:38 <gmaxwell> ...
 53 2012-05-06 04:03:54 <jine> luke-jr: He only said that he's better in code then text.
 54 2012-05-06 04:03:58 <Anon007> Please Tell me I am Misunderstanding something when I listened to Gavin he claimed that there were no plans on making the BitCoin Mining more Profitable for people with slow rigs :(
 55 2012-05-06 04:04:00 <luke-jr> jine: :p
 56 2012-05-06 04:04:22 <jine> That doesn't say to much tho
 57 2012-05-06 04:04:29 <gmaxwell> Anon007: It _can't_ be more profitable for people with slow rigs. The design of bitcoin precludes that possiblity.
 58 2012-05-06 04:04:45 <Anon007> Ok
 59 2012-05-06 04:04:46 <luke-jr> it's not even supposed to be profitable at all
 60 2012-05-06 04:04:51 <gmaxwell> As luke says.
 61 2012-05-06 04:05:42 <gmaxwell> Anon007: also its really just about as (un)profitable, in terms of return on investment, for everyone.
 62 2012-05-06 04:05:58 <Anon007> Ok I misunderstand him than... It seemed like he was more interested in the Money than Helping Average Individuals get setup with mining and he also said he has no plans in supporting the Mining Pools :(
 63 2012-05-06 04:06:20 <luke-jr> &&&
 64 2012-05-06 04:06:33 <gmaxwell> Mining pools create centeralization and are generally bad for bitcoin.
 65 2012-05-06 04:07:02 <gmaxwell> They transfer power/control and wealth into a smaller set of pool operators. This has good and bad sides.
 66 2012-05-06 04:07:07 <jine> gmaxwell:
 67 2012-05-06 04:07:10 <jine> Upgraded.
 68 2012-05-06 04:07:32 <Anon007> Ok
 69 2012-05-06 04:08:07 <Anon007> so how do I get bitcoins without a Pool... It would take ages to solve a block at 150 megahashes a second :(
 70 2012-05-06 04:08:33 <luke-jr> Anon007: can you make a youtube video of you doing this rant verbally? it would make good humour material.
 71 2012-05-06 04:08:39 <Anon007> I would be better off setting up a system that accepts bitcoins in exchange for services :)
 72 2012-05-06 04:09:11 <wumpus> there are many such sites, just look around a bit on the wiki instead of ranting in the developer channel
 73 2012-05-06 04:09:15 <Anon007> NAH just asking questions you never know unless you ask :)
 74 2012-05-06 04:09:28 <gmaxwell> Anon007: mining isn't cumulative. It doesn't take 'ages' as much as 'its unlikely to happen at any given instant' ... but the better way to obtain bitcoins is to sell goods and services that people want.
 75 2012-05-06 04:09:48 <BlueMatt> ummm...ok, so why does this dead-simple test fail: https://github.com/bitcoin/bitcoin/pull/771/files#diff-25
 76 2012-05-06 04:09:50 <Anon007> Ok Thank You GMaxWell
 77 2012-05-06 04:10:03 <Anon007> That is the Best Answer here :)
 78 2012-05-06 04:10:05 <BlueMatt> (gets "Didnt ever leave CS???")
 79 2012-05-06 04:10:38 <gmaxwell> If you're really interested in mining  You can also use decenteralized pools like p2pool, or Luke's decenteralized pool support.. though both these things are currently pretty technical. But mining just isn't a great way to get bitcoins.
 80 2012-05-06 04:11:21 <Anon007> I almost have 1 Bit Coin :P
 81 2012-05-06 04:11:38 <Anon007> After Weeks of Mining and starting to get Blue Screens I stopped
 82 2012-05-06 04:11:55 <luke-jr> BlueMatt: it's too timing-dependent?
 83 2012-05-06 04:12:28 <gmaxwell> as luke says, the sleeps are so short god knows what happens.
 84 2012-05-06 04:12:33 <gmaxwell> You can't count on sleeps to be accurate.
 85 2012-05-06 04:13:01 <wumpus> good that you stopped, getting blue screens while mining usually means there is a problem with cooling of your rig
 86 2012-05-06 04:13:06 <BlueMatt> luke-jr: doesnt seem to be, changing the sleep significantly doesnt help
 87 2012-05-06 04:13:32 <gmaxwell> BlueMatt: add a bunch of prints so you can see the sequence its really taking.
 88 2012-05-06 04:14:13 <BlueMatt> arg...I was hoping on focusing on rebasing and letting someone else find the problem there...
 89 2012-05-06 04:14:14 <BlueMatt> oh well
 90 2012-05-06 04:14:39 <jine> gmaxwell: Does the "walletversion" : 59900 matter?
 91 2012-05-06 04:15:08 <jine> I was planning to go grap a new wallet.dat anyway so, cause of the size... taking to much space and ram, and it's growing pretty fast.
 92 2012-05-06 04:15:15 <gmaxwell> jine: no, not really. You can tell it to upgrade it.. thats been made a manual process so that users retain the freedom to downgrade.
 93 2012-05-06 04:15:38 <jine> Ah, whats the diffrence between the two versions?
 94 2012-05-06 04:15:58 <gmaxwell> the 0.6.0 version adds the support for compressed public keys, IIRC.
 95 2012-05-06 04:16:31 <jine> I c, how do i upgrade?
 96 2012-05-06 04:16:44 <jine> nvm
 97 2012-05-06 04:16:49 <jine> found -upgradewallet
 98 2012-05-06 04:18:28 <gmaxwell> Backing up first is recommended of course.
 99 2012-05-06 04:18:34 <jine> ofc :)
100 2012-05-06 04:18:55 <jine> Can i add it to bitcoin.conf, restart the client and then later remove the line?
101 2012-05-06 04:19:07 <jine> Or do i need to run it separaretly?
102 2012-05-06 04:20:53 <jine> gmaxwell: ^
103 2012-05-06 04:21:13 <gmaxwell> It should work in the conf sorry, was checking the code. Never tested that.
104 2012-05-06 04:21:44 <jine> Hah, i c :)
105 2012-05-06 04:26:31 <jine> Fully upgraded, thanks :)
106 2012-05-06 04:28:03 <jine> Also updated keypoolsize to having to take backup less often :)
107 2012-05-06 04:32:47 <gribble> New news from bitcoinrss: laanwj opened pull request 1209 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1209>
108 2012-05-06 05:15:21 <splatster> back
109 2012-05-06 05:36:11 <paulo_> is there an exact date when SHA-3 gets chosen?
110 2012-05-06 05:37:48 <splatster> paulo_: I don't think so.
111 2012-05-06 10:36:16 <gribble> New news from bitcoinrss: Diapolo opened issue 1210 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1210>
112 2012-05-06 10:37:33 <sipa> ;;later tell BlueMatt test_bitcoin in cblockstore seems to work here; shouldn't it?
113 2012-05-06 10:37:34 <gribble> The operation succeeded.
114 2012-05-06 10:38:12 <sipa> BlueMatt: and my irc client runs on my vps, so it's online when i'm not
115 2012-05-06 12:28:41 <gribble> New news from bitcoinrss: Diapolo opened pull request 1211 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1211>
116 2012-05-06 13:05:00 <gribble> New news from bitcoinrss: Diapolo opened pull request 1212 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1212>
117 2012-05-06 16:28:29 <nanotube> possibly bug report: before 0.6.1, when creating a new address with the 'new address' button, the address was created and focus was set on it. now with 0.6.1 it is not focused on after creation, and one needs to find it in the list. (which may be problematic if it has the same label as some others). anyone else see this behavior? anyone else agree that it is a bug?
118 2012-05-06 16:29:03 <luke-jr> nanotube: I've seen that for a month or so at least now, and it's been annoying. I didn't realize it was a regression.
119 2012-05-06 16:29:10 <luke-jr> 0.6.0 didn't have that issue?
120 2012-05-06 16:29:39 <nanotube> no, i only started seeing this behavior as of a few days ago, which iirc coincides with my upgrade to .1
121 2012-05-06 16:30:09 <luke-jr> hmm
122 2012-05-06 16:30:24 <nanotube> not 100% sure though. i could double check by starting the .0 client.... there's no backward-incompatibilities between .1 and .0 is there?
123 2012-05-06 16:30:31 <nanotube> s/is/are/g
124 2012-05-06 16:31:10 <gmaxwell> nanotube: no, you should be able to start 0.6.0.
125 2012-05-06 16:31:31 <nanotube> ok i'll give it a try then.
126 2012-05-06 16:32:52 <nanotube> gmaxwell: luke-jr: confirmed, 0.6.0 properly sets focus.
127 2012-05-06 16:33:18 <luke-jr> k, bisecting..
128 2012-05-06 16:33:26 <nanotube> (this is on linux for the record, but y'all already know that :) )
129 2012-05-06 16:40:44 <wumpus> probably introduced in 98e61758744ed34e8b7f59b37edb6d09b33d5517
130 2012-05-06 16:41:33 <wumpus> and yes, it's a bug, please file an issue
131 2012-05-06 16:53:51 <luke-jr> 98e61758744ed34e8b7f59b37edb6d09b33d5517 is the first bad commit
132 2012-05-06 16:53:53 <luke-jr> yep
133 2012-05-06 17:13:39 <gmaxwell> "walletversion" : 10500,
134 2012-05-06 17:13:42 <gmaxwell> ^ wtf?
135 2012-05-06 17:14:26 <DiabloD3> ?
136 2012-05-06 17:14:40 <gmaxwell> log says
137 2012-05-06 17:14:41 <gmaxwell> 05/06/12 19:11:36 nFileVersion = 59900
138 2012-05-06 17:14:49 <gmaxwell> DiabloD3: thats a screwed up version.
139 2012-05-06 17:15:09 <DiabloD3> I didnt know wallets had versions
140 2012-05-06 17:15:23 <gmaxwell> Yes.
141 2012-05-06 17:17:13 <luke-jr> DiabloD3: that's why you can downgrade to 0.5 after running 0.6 only iff you started with 0.5
142 2012-05-06 17:18:13 <DiabloD3> can I make it upgrade a wallet?
143 2012-05-06 17:19:02 <gmaxwell> DiabloD3: yes.
144 2012-05-06 17:19:32 <DiabloD3> how
145 2012-05-06 17:19:42 <gmaxwell> Anyone have any ideas why I'd be getting 10500? from a 59900 wallet?
146 2012-05-06 17:35:18 <gmaxwell> wallet.h:    FEATURE_BASE = 10500, // the earliest version new wallets supports (only useful for getinfo's clientversion output)
147 2012-05-06 17:35:57 <gmaxwell> I guess thats okay then.. kinda odd.
148 2012-05-06 17:46:30 <nanotube> wumpus: ok i'll file
149 2012-05-06 17:46:38 <wumpus> I guess it's not needed
150 2012-05-06 17:46:46 <wumpus> already working on fix
151 2012-05-06 17:46:49 <nanotube> ok i won't then. :)
152 2012-05-06 17:46:54 <nanotube> thanks! :)
153 2012-05-06 18:00:16 <amiller> has anyone seen this proposal, it's pretty much the same as what i'm working on https://en.bitcoin.it/wiki/User:DiThi/MTUT
154 2012-05-06 18:02:03 <gmaxwell> amiller: you mean like this? https://bitcointalk.org/index.php?topic=21995.0
155 2012-05-06 18:03:30 <amiller> yeah that's exactly right.
156 2012-05-06 18:03:34 <amiller> so i've figured out how to do that
157 2012-05-06 18:04:02 <amiller> you should look at this paper Persistent Authenticated Dictionaries http://www.cs.brown.edu/cgc/stms/papers/isc2001.pdf
158 2012-05-06 18:04:12 <gmaxwell> Can you also produce efficient provable no-txn found results?
159 2012-05-06 18:04:27 <gmaxwell> The way that wikipage describs it can't do that.
160 2012-05-06 18:04:29 <amiller> yes you get non-membership verification with the same O(log N)
161 2012-05-06 18:04:36 <gmaxwell> (but I know it can be done)
162 2012-05-06 18:04:56 <gmaxwell> what I don't know how to do is make something that does non-membership secure against attack.
163 2012-05-06 18:05:22 <amiller> the trick is to used a balanced merkle tree
164 2012-05-06 18:05:24 <gmaxwell> Becuase the only way I see to do non-membership requrires txn to have a known connection point based on their hash. But then an attacker can make many txn that connect at the same point.
165 2012-05-06 18:05:29 <gmaxwell> To unbalance the tree.
166 2012-05-06 18:05:35 <sipa> gmaxwell: that nFileVersionis simply the last version that wrote to it, to correct bugs for example
167 2012-05-06 18:05:58 <amiller> this paper i linked describes self balancing merkle trees
168 2012-05-06 18:06:04 <amiller> so the worst case verification time is still log n
169 2012-05-06 18:06:04 <sipa> gmaxwell: the number reported by walletversion is different
170 2012-05-06 18:06:14 <amiller> you might also look at the code i'm playing with for this https://github.com/amiller/redblackmerkle/
171 2012-05-06 18:07:11 <gmaxwell> amiller: K. paper in my reading queue.
172 2012-05-06 18:09:21 <gmaxwell> amiller: in any case, I think this sort of thing is absolutely essential for something like namecoin... and potentially of value for bitcoin but not as important.
173 2012-05-06 18:11:10 <amiller> one thing missing from your proposal and from the wiki thing i suppose is that i would like to see the proof-of-work changed to being a proof-of-throughput for accessing this merkle tree
174 2012-05-06 18:13:04 <gmaxwell> Yea, and I don't think thats a grand idea, but I'm still open to being convinced.  (I think that proof that you've got it is great, proof of throughtput seems to me to be trading one kind of busywork for another because the maximum amount of throughput you could possibly need for validation is necessairly low, otherwise the system loses decenteralization)
175 2012-05-06 18:14:53 <amiller> if all you do is include the merkle tree root in your blocks, you haven't proved you have access to the tree
176 2012-05-06 18:16:06 <gmaxwell> amiller: Yes, and?
177 2012-05-06 18:16:09 <amiller> but if you include the result of accumulating k random elements from the tree
178 2012-05-06 18:16:23 <amiller> then someone can verify your work in O(k * log N)
179 2012-05-06 18:16:58 <amiller> but you would have had to access the tree on average O(T) times
180 2012-05-06 18:17:49 <gmaxwell> I'm not sure what you're telling me that I didn't know. I think proof of access is fine. I don't think it's a good thing to derive the difficulty of the block from.
181 2012-05-06 18:20:10 <gmaxwell> E.g. you can make a very cheap proof of access, e.g. hash of the second to last block is your random seed for your N lookups.  N could be as small as 1.  You could also include a proof based on the validation you had to do for the prior block since you had to do that work anyways if you're not cheating.
182 2012-05-06 18:21:24 <gmaxwell> Thats enough to remove the incentive to be forgetful, but without the gratitious change in the proof of work type.
183 2012-05-06 18:21:36 <amiller> let me use k for what you're calling N
184 2012-05-06 18:22:02 <amiller> if you only have to access k=1 value, then you could cheat at the work by only storing one value
185 2012-05-06 18:22:13 <amiller> and you keep generating random numbers until that's the lookup you have to do
186 2012-05-06 18:22:16 <gmaxwell> You don't control what value you have to access.
187 2012-05-06 18:22:25 <gmaxwell> Negatory ghostrider.
188 2012-05-06 18:22:48 <amiller> okay so it doesn't depend on your nonce, you don't have to do a lookup for each work?
189 2012-05-06 18:22:59 <gmaxwell> Right.
190 2012-05-06 18:23:11 <amiller> then only one person needs to have that data and they can broadcast
191 2012-05-06 18:23:17 <amiller> it doesn't really improve decentralization at all
192 2012-05-06 18:23:34 <gmaxwell> There is no reason to do lookups per each work. The total lookup throughput _must_ be bounded if the system is to be decenteralized.
193 2012-05-06 18:24:06 <gmaxwell> amiller: proving people have lots of lookup throughput doesn't help decenteralzation either.
194 2012-05-06 18:25:06 <gmaxwell> I'm late for lunch, but we should talk about this later. I'd like one of us to convince the other on this.
195 2012-05-06 18:25:15 <amiller> ok, i'll be around
196 2012-05-06 18:30:55 <Diapolo> hi all
197 2012-05-06 18:42:24 <Diapolo> Why is Monospace used or defined as "address" font?
198 2012-05-06 18:46:49 <wumpus> because it is consistent and makes addresses all the same width
199 2012-05-06 19:09:19 <Diapolo> wumpus: But is intended only for addresses?
200 2012-05-06 19:11:10 <wumpus> yes
201 2012-05-06 19:14:09 <wumpus> it's basically so that they line up in the address book (note that addresses look ugly in any font, so it doesn't matter so much)
202 2012-05-06 19:14:32 <gmaxwell> I think monospace makes them look nicer.
203 2012-05-06 19:14:37 <wumpus> better to not show them at all, where that's possible
204 2012-05-06 19:14:38 <wumpus> yes
205 2012-05-06 19:14:46 <wumpus> slightly :)
206 2012-05-06 19:16:22 <gmaxwell> hm. I wonder if some letter pairs look nicer, and if someone could define finite state machine for an encoding that produces nicer looking mixtures.
207 2012-05-06 19:17:36 <Diapolo> wumpus: someone used Monospace on the amount label, I was not sure if it was ok to remove it but now I'm convinced :D.
208 2012-05-06 19:18:39 <ali1234> monospace technically isn't a font, it's a font family that just means "any available monospace font"
209 2012-05-06 19:19:01 <Diapolo> thanks for clarification :)
210 2012-05-06 19:19:02 <ali1234> assuming you are talking about the Qt UI anyway, which uses (in theory) CSS properties
211 2012-05-06 19:19:12 <wumpus> that would be like creating better looking hex memory addresses
212 2012-05-06 19:19:15 <ali1234> not that i've looked at the code but i can only guess this is what you are refering to
213 2012-05-06 19:19:27 <Diapolo> you are right ali
214 2012-05-06 19:19:36 <gmaxwell> ali1234: generally addresses look better in any monospace face than any non-monospace face.
215 2012-05-06 19:19:46 <wumpus> yes it's a family, not a specific one
216 2012-05-06 19:19:49 <gmaxwell> (if for no other reason then they line up)
217 2012-05-06 19:20:04 <gmaxwell> I'm fond of Inconsolata.
218 2012-05-06 19:21:35 <ali1234> when it comes to monospace i prefer the output of the saa5243
219 2012-05-06 19:21:59 <Diapolo> Should monospace also be used in the transactionlist on the main tab and the transactiontable?
220 2012-05-06 19:22:54 <wumpus> ideally, yes, but only when it's not too much complexity to do so
221 2012-05-06 19:23:30 <Diapolo> wumpus: These are things I can do and don't take too much time the real big GUI changes are your part currently ^^.
222 2012-05-06 19:23:44 <wumpus> not all text outputs allow mixing different fonts, if you need to go as far as subclassing a renderer, don't bother
223 2012-05-06 19:25:16 <Diapolo> will look into it tomorrow, too late today :)
224 2012-05-06 19:25:42 <wumpus> for example in the transaction table it shows label (address). You can only have one font per cell, and making them all monospace looks weird
225 2012-05-06 19:26:49 <Diapolo> wumpus: that makes sense sure
226 2012-05-06 19:27:16 <gmaxwell> It might look nicer if all addresses displayed in the interface were display between brackets. e.g. [17Q5ZBy5N9ke2dU1X7gGGT2vWfKJEm3gW4]
227 2012-05-06 19:28:08 <wumpus> yea you could use brackets instead of parentheses
228 2012-05-06 19:28:39 <wumpus> though I don't think you should put brackets around the addresses in the address book
229 2012-05-06 19:28:41 <Diapolo> btw. is anyone here that can build a Windows executable that I can redistribute to test my experimental blockchain-writing code? Seems I'm not able to meet BlueMatt in here anymore.
230 2012-05-06 19:28:49 <sipa> Diapolo: i can
231 2012-05-06 19:28:54 <sipa> (gitian)
232 2012-05-06 19:28:59 <ali1234> couldn't you do something with columns? and then hide the addresses unless the user decides to show that column?
233 2012-05-06 19:29:17 <sipa> Diapolo: just point me to a commit
234 2012-05-06 19:29:22 <Diapolo> sipa: cool I made some progress https://github.com/Diapolo/bitcoin/tree/InitBlockDL-exp
235 2012-05-06 19:29:27 <wumpus> ali1234: that's already the case
236 2012-05-06 19:29:49 <wumpus> the address is hidden *unless* it has an empty or no label
237 2012-05-06 19:30:02 <Diapolo> sipa: I'm sure this should be not that hard for Linux-style OSes, too ... but I had no time to investigate the API calls
238 2012-05-06 19:30:03 <ali1234> if it's on a different column you should be able to set the font for only that column (again in theory)
239 2012-05-06 19:30:15 <wumpus> which is why it's not really an issue for 'good' users, which label all addresses they use for transactions
240 2012-05-06 19:30:49 <ali1234> i never label anything :)
241 2012-05-06 19:30:53 <Diapolo> wumpus: perhaps merge my labeling fix, I heard about a possible 0.6.2 and this fixes a medium issue
242 2012-05-06 19:31:08 <ali1234> but then again i don't use the GUI anyway
243 2012-05-06 19:31:17 <Diapolo> I only use the GUI ^^.
244 2012-05-06 19:31:20 <sipa> Diapolo: still no per-file pointers?
245 2012-05-06 19:31:41 <sipa> Diapolo: ah, i see
246 2012-05-06 19:31:43 <wumpus> most people label addresses because it's the only way to currently distinguish transactions/recipients
247 2012-05-06 19:31:44 <sipa> you just append anyway
248 2012-05-06 19:31:46 <sipa> Diapolo: nevermind
249 2012-05-06 19:31:48 <gmaxwell> sipa: I have got someone with a stuck chain who is willing to test your fixed fix, BTW.
250 2012-05-06 19:32:17 <gmaxwell> He's on 0.6.1 release and still stuck, so it should be a worthwhile test case.
251 2012-05-06 19:32:23 <Diapolo> sipa: I tried to do the things you suggested ^^ no globals, use the tx-db for settings
252 2012-05-06 19:32:53 <wumpus> Diapolo: right, I really already started merging for 0.7.0 and then it was decided to do 0.6.2
253 2012-05-06 19:33:03 <wumpus> (it's the same every time...)
254 2012-05-06 19:33:52 <Diapolo> wumpus: Are there meetings where such things are decided, sometimes it seems a bit random to me as Newb ^^.
255 2012-05-06 19:34:23 <sipa> gmaxwell: i can do a gitian build, if necessary?
256 2012-05-06 19:34:33 <wumpus> it would make more sense to have two branches, so we can start merging for 0.7.0 even while 0.6.x is still being maintained
257 2012-05-06 19:34:56 <luke-jr> wumpus: we have 4 branches now, I can add a 0.6.x branch easily
258 2012-05-06 19:35:14 <luke-jr> (in fact, I have one, just not pushed yet)
259 2012-05-06 19:35:22 <wumpus> I mean in the official repository... I know you maintain the older versions
260 2012-05-06 19:35:30 <wumpus> which makes sense
261 2012-05-06 19:35:34 <Diapolo> sipa: while you are at it can you do an experimental Windows exe for me :)? With that branch ...
262 2012-05-06 19:35:39 <sipa> Diapolo: sure
263 2012-05-06 19:35:49 <gmaxwell> That is not equivalent. We don't have a large enough testing base to justify dilluting it with a lot of compeating heads.
264 2012-05-06 19:35:58 <luke-jr> wumpus: the stable repository is just as official as master on github
265 2012-05-06 19:36:02 <wumpus> I don't want 'lots' of competing heads
266 2012-05-06 19:36:11 <wumpus> just the old and new one
267 2012-05-06 19:36:14 <Diapolo> sipa: great :) hope it compiles ^^ my local setup is sometimes a little different to official Gitian envs
268 2012-05-06 19:36:43 <gmaxwell> sipa: If you want to just go ahead and do all platform builds with your chain fix, I suppose that would be useful.
269 2012-05-06 19:36:56 <sipa> gmaxwell: ok
270 2012-05-06 19:37:10 <Diapolo> wumpus: I agree as it's sometimes hard to see good pulls that don't get into the master and can be tested. I don't want to setup remotes to every of your repos ^^.
271 2012-05-06 19:37:13 <sipa> Diapolo: i'll first do the stuck chain fix builds, then yours
272 2012-05-06 19:37:45 <wumpus> Diapolo: agreed... so we end up with lots of conflicting pull requests, which we already know we want to merge
273 2012-05-06 19:38:24 <gmaxwell> Diapolo: I try to be pretty agressive at pulling the obviously minor stuff quickly.
274 2012-05-06 19:38:26 <luke-jr> ('next' is just pullreqs with core ACKs)
275 2012-05-06 19:39:05 <Diapolo> sipa: no hurry, I guess I'm off in a few minutes, but can you post the link here: https://github.com/bitcoin/bitcoin/issues/776 or perhaps somewhere I can find it ^^
276 2012-05-06 19:39:39 <luke-jr> anyhow, I guess my key point is that every possible branch already exists, and any new one sounds like trying to reinvent something that's already there
277 2012-05-06 19:40:02 <Diapolo> gmaxwell: that was no offence to anyone, I just produce small pulls quite often for example ^^ and I would really like to test the RPC console or IPV6 :)
278 2012-05-06 19:40:22 <gmaxwell> Diapolo: you know you can get any pull request as a patch and git am it easily, right?
279 2012-05-06 19:40:26 <wumpus> luke-jr: right, but it'd be useful if it existed in one repository on github, so coordination is easier
280 2012-05-06 19:40:33 <gmaxwell> Diapolo: e.g. wget https://github.com/bitcoin/bitcoin/pull/1075.patch
281 2012-05-06 19:40:57 <gmaxwell> Thats how I test every pull before I pull it. Perhaps there is some more githubish way, but I think its really easy.
282 2012-05-06 19:41:13 <Diapolo> gmaxwell: I have heard of wget ... but never used this.
283 2012-05-06 19:41:28 <gmaxwell> Diapolo: wget is just a command to save a url to a file.
284 2012-05-06 19:41:41 <sipa> Diapolo: wget is just a program for downloading a file; the wget is not important, it's the url
285 2012-05-06 19:41:43 <luke-jr> wumpus: does GitHub let you have different commit-access per branch?
286 2012-05-06 19:41:52 <luke-jr> curl > wget
287 2012-05-06 19:41:56 <sipa> wumpus: not afaik
288 2012-05-06 19:42:01 <sipa> eh, luke-jr
289 2012-05-06 19:42:01 <wumpus> I generally just do   git remote add, git fetch, git checkout <user>/<branch>
290 2012-05-06 19:42:04 <Diapolo> I'm that lazy Windows Guy who hates command prompts ^^.
291 2012-05-06 19:42:10 <wumpus> no, it doesn't
292 2012-05-06 19:42:11 <luke-jr> Diapolo: tried next-test?
293 2012-05-06 19:42:24 <gmaxwell> e.g. you can wget https://github.com/bitcoin/bitcoin/pull/1075.patch ; git am 1075.patch ; make .. ; and when you're done git reset --hard origin
294 2012-05-06 19:42:38 <gmaxwell> Diapolo: ah well you can also use your browser to save that url. :)
295 2012-05-06 19:42:56 <luke-jr> wumpus: well, then moving 'next' to that repo implies either A) adding me as a pusher to it, or B) someone else takes the job of maintaining it ;)
296 2012-05-06 19:43:12 <Diapolo> ^^ I will have to look into all this, I'm the growing and learning one here ;).
297 2012-05-06 19:43:19 <luke-jr> wumpus: I presume (A) would only happen independent of this conversation, so is someone volunteering for (B)?
298 2012-05-06 19:43:38 <gmaxwell> Diapolo: you also create a lot of GUI patches which I don't, as a rule, pull unless they're very obviously okay.
299 2012-05-06 19:43:42 <luke-jr> Diapolo: https://bitcointalk.org/?topic=79884.0
300 2012-05-06 19:43:45 <luke-jr> Diapolo: https://bitcointalk.org/?topic=79884.0
301 2012-05-06 19:44:10 <Diapolo> luke-jr: I was not sure if that builds are trustworthy or could perhaps be modified to do bad things ...
302 2012-05-06 19:44:11 <sipa> indeed, i sometimes comment on GUI patches, but leave the pulling mostly to wumpus
303 2012-05-06 19:44:35 <gmaxwell> One thing that would probably be helpful is having a jenkins setup that drives gitian builds of _everything_ ... I'd be glad to do this, and my jenkins kungfu is reasonable, but I can't seem to get gitian working.
304 2012-05-06 19:44:39 <luke-jr> Diapolo: well, that depends on if you trust me. the branches also exist in my git repo, if you wanted to verify the builds yourself
305 2012-05-06 19:45:05 <sipa> gmaxwell: never had a problem with gitian (except some learning curve)
306 2012-05-06 19:45:10 <luke-jr> gmaxwell: that'd be neat
307 2012-05-06 19:45:34 <sipa> gmaxwell: but you don't run ubuntu, i suppose?
308 2012-05-06 19:45:36 <gmaxwell> sipa: I don't run ubuntu normally so I have it running in a VM.
309 2012-05-06 19:46:03 <gmaxwell> and the machine I want to run it in can't do the fast nested kvm, so I'm trying to get the LXC mode working.
310 2012-05-06 19:46:11 <Diapolo> gmaxwell: "as a rule, pull unless they're very obviously okay." I don't get what you mean, sorry ... can you explain?
311 2012-05-06 19:46:12 <wumpus> the only annoying thing about gitian is that I can't run it together with virtualbox
312 2012-05-06 19:46:13 <gmaxwell> (I want to run it on my 32 core boxes)
313 2012-05-06 19:46:51 <gmaxwell> Diapolo: I mean your GUI patches take longer to get pulled because mostly only wumpus pulls them.  I only pull gui related patches if they're obviously correct text changes or the like.
314 2012-05-06 19:47:11 <Diapolo> gmaxwell: now it's clear
315 2012-05-06 19:47:20 <gmaxwell> (and likewise for sipa)
316 2012-05-06 19:48:11 <Diapolo> luke-jr: I guess I should try your next-test to play around with pulls, so I don't need to rely on merging them into local repos and compiling them for now.
317 2012-05-06 19:48:15 <luke-jr> but I guess you'd want the LXC network jails for this kind of thing
318 2012-05-06 19:48:30 <gmaxwell> luke-jr: the gitian stuff is kind of horiffic. It requires sudoing.
319 2012-05-06 19:49:00 <Diapolo> damn, time is running to fast ...
320 2012-05-06 19:49:03 <gmaxwell> I looked into trying to get it running on a fedora host with just the ubuntu vms and gave up pretty quickly.
321 2012-05-06 19:49:28 <gmaxwell> luke-jr: sipa: What version of ubuntu is your gitian host systems?
322 2012-05-06 19:50:20 <gmaxwell> Mine is ubuntu-10.04.4-server-amd64.iso which is perhaps why I'm having problems where no one else is.
323 2012-05-06 19:50:44 <luke-jr> gmaxwell: how can you tell?
324 2012-05-06 19:52:56 <Diapolo> I have to say good night. sipa: It's very well okay if you do my build tomorrow, if you are out of time.
325 2012-05-06 19:54:50 <sipa> gmaxwell: gitian requires sudo?
326 2012-05-06 19:55:07 <luke-jr> sipa: yes
327 2012-05-06 19:55:49 <sipa> In file included from src/qt/messagepage.cpp:16:
328 2012-05-06 19:55:52 <sipa> build/ui_messagepage.h:165: error: class QLineEdit has no member named setPlaceholderText
329 2012-05-06 19:56:39 <ali1234> introduced in Qt 4.7
330 2012-05-06 19:56:46 <luke-jr> are we building Win32 with an old Qt? :/
331 2012-05-06 19:57:22 <wumpus> sipa: huh, that's in the generated code?
332 2012-05-06 19:58:02 <sipa> wumpus: that's the gitian linux build
333 2012-05-06 19:59:21 <wumpus> yes but that means uic generates code that doesn't compile with the same qt version it is packaged with... darn
334 2012-05-06 19:59:54 <sipa> this is problematic, as i can't do a gitian build anymore :)
335 2012-05-06 20:02:41 <luke-jr> I need a testnet address please
336 2012-05-06 20:04:22 <wumpus> miW96tTsuBt6fMoxfBH8JSfuCMfxy5SZv2
337 2012-05-06 20:05:28 <luke-jr> ty
338 2012-05-06 20:05:38 <luke-jr> [21:59:21] <wumpus> yes but that means uic generates code that doesn't compile with the same qt version it is packaged with& darn <-- this is why I asked before it was merged -.-
339 2012-05-06 20:05:54 <wumpus> I didn't think qt would be that stupid
340 2012-05-06 20:06:18 <wumpus> anyway, I'm moving it back to the code now
341 2012-05-06 20:06:48 <gmaxwell> darn users!
342 2012-05-06 20:07:06 <gmaxwell> My stuck chain test subject applied the proper patch but also deleted his chain at the same time.
343 2012-05-06 20:07:19 <sipa> haha
344 2012-05-06 20:10:13 <wumpus> luke-jr: and I was right on the xml validation! :p
345 2012-05-06 20:11:25 <luke-jr> I don't suppose anyone is mining testnet right now?
346 2012-05-06 20:11:49 <wumpus> I never use the real testnet, just the one in a box
347 2012-05-06 20:12:07 <gmaxwell> Testnet is mostly broken now.
348 2012-05-06 20:12:23 <sipa> we really need to reset it for 0.7.0
349 2012-05-06 20:14:52 <vision_> I had an idea yesterday to make transactions easy and fast in normal life, but i don't have any idea of programming so i would just like to share and if you see interesting, just take the idea and make it possible...
350 2012-05-06 20:15:15 <luke-jr> vision_: what can be faster than the instantaneous we have right now?
351 2012-05-06 20:15:24 <sipa> luke-jr: seems you need a better monitor
352 2012-05-06 20:15:26 <vision_> yes
353 2012-05-06 20:15:30 <luke-jr> sipa: ?
354 2012-05-06 20:15:40 <vision_> giving a coin would be faster
355 2012-05-06 20:15:42 <sipa> luke-jr: i do see alternating colors in your screenshot
356 2012-05-06 20:15:42 <vision_> :)
357 2012-05-06 20:15:45 <vision_> let me explain
358 2012-05-06 20:15:54 <sipa> luke-jr: though only softly
359 2012-05-06 20:15:54 <vision_> I was a bartender in a disco
360 2012-05-06 20:16:48 <vision_> and i was thinking that bitcoin would never work in a bar or a disco because the amount of time you need to set-up or do the transfer is to high
361 2012-05-06 20:17:06 <vision_> so in that time the bartender is NOT serving like 10 drinks
362 2012-05-06 20:17:10 <luke-jr> sipa: #F8F7F6 vs #FFFFFF isn't visible :p
363 2012-05-06 20:17:18 <vision_> which makes the bar lose money
364 2012-05-06 20:17:24 <vision_> well whatever
365 2012-05-06 20:17:35 <vision_> i thought it would be great this system
366 2012-05-06 20:17:38 <luke-jr> vision_: nonsense, bartender shows a QR Code, you scan, done
367 2012-05-06 20:17:38 <sipa> vision_: it could be made to work in seconds (NFC, for example)
368 2012-05-06 20:17:46 <sipa> or QR, for now
369 2012-05-06 20:17:48 <vision_> what is NFC?
370 2012-05-06 20:17:53 <sipa> near-field communication
371 2012-05-06 20:18:02 <wumpus> the bitcoin android app supports that, at least
372 2012-05-06 20:18:05 <vision_> which is like a RDIF work
373 2012-05-06 20:18:13 <vision_> isn't it?
374 2012-05-06 20:18:24 <vision_> wumpus really?
375 2012-05-06 20:18:29 <sipa> yes
376 2012-05-06 20:18:30 <vision_> how does it work?
377 2012-05-06 20:18:47 <sipa> shop generates a bitcoin url with address and amount in a qr code
378 2012-05-06 20:18:51 <sipa> *uri
379 2012-05-06 20:19:01 <sipa> you scan it with your phone that has a bitcoin app
380 2012-05-06 20:19:12 <sipa> asks you whether you want to pay
381 2012-05-06 20:19:24 <sipa> and then creates/sends the transaction
382 2012-05-06 20:19:40 <vision_> i think my system is better
383 2012-05-06 20:19:44 <sipa> which is?
384 2012-05-06 20:20:05 <vision_> people should need to buy a RDIF card of the system
385 2012-05-06 20:20:21 <sipa> that can work, but that's not bitcoin anymore
386 2012-05-06 20:20:21 <wumpus> sipa: latest commit should fix the gitian build problem
387 2012-05-06 20:20:24 <vision_> this card would have a serial number on it (or a QR that you would have to scan)
388 2012-05-06 20:20:31 <sipa> that's a payment processing system on top of it
389 2012-05-06 20:20:35 <vision_> let me just finish
390 2012-05-06 20:20:44 <sipa> ok!
391 2012-05-06 20:20:46 <vision_> :)
392 2012-05-06 20:20:48 <vision_> ok
393 2012-05-06 20:20:58 <vision_> because i don't like the middleman idea
394 2012-05-06 20:21:07 <luke-jr> can someone send me TNBTC? mjovDSJv8gCJ8fSz12xEH7LSuT278FXLJh
395 2012-05-06 20:21:26 <vision_> it is true that you would have to build a special app which would be a kind of wallet
396 2012-05-06 20:21:40 <vision_> then you would setup this wallet with the serial number of the card
397 2012-05-06 20:21:44 <vision_> and 4 number digit
398 2012-05-06 20:22:10 <vision_> (if someone tryes 3 number it would block the number and then you should change in your cellphone)
399 2012-05-06 20:22:52 <sipa> and the actual wallet is on your cell phone?
400 2012-05-06 20:22:56 <vision_> in the bar side: a special rdif scanner shows the amount of money you have to pay, if you agree you put your card next to the scanner and put the 4 digit
401 2012-05-06 20:23:20 <luke-jr> cool. I'm going to make my rdif scanner show 5 BTC then charge 50 BTC
402 2012-05-06 20:23:39 <sipa> what's the difference with holding your cell phone next to the scanner, and typing the digits on your cell phone?
403 2012-05-06 20:23:48 <ali1234> that's ok cos i'm going to make a rdif wrier that puts 1,000,000 btc on my card
404 2012-05-06 20:23:56 <wumpus> luke-jr: just make it charge TBC, no one will understand anyway :)
405 2012-05-06 20:24:02 <luke-jr> wumpus: TBC are smaller
406 2012-05-06 20:24:04 <vision_> sipa if the cellphone holds a rdif nothing at all
407 2012-05-06 20:24:09 <wumpus> darn
408 2012-05-06 20:24:29 <sipa> vision_: or a QR code scanner?
409 2012-05-06 20:24:32 <luke-jr> vision_: I like the QR Code system that works easier with current phones
410 2012-05-06 20:24:39 <vision_> but i have a galaxy s
411 2012-05-06 20:24:46 <vision_> and the qr scanning is slowww
412 2012-05-06 20:24:58 <sipa> that will certainly improve
413 2012-05-06 20:25:04 <sipa> or be replaced by other technology
414 2012-05-06 20:25:17 <sipa> i'm just saying there is no fundamental difference with your idea
415 2012-05-06 20:25:21 <luke-jr> vision_: at least it works. bet that galaxy s doesn't work with your idea ta tall :p
416 2012-05-06 20:25:21 <vision_> sipa think that when i was a bartender i used to put an average of 98 drinks per hour
417 2012-05-06 20:25:26 <luke-jr> at all*
418 2012-05-06 20:25:42 <vision_> luke it would
419 2012-05-06 20:25:45 <sipa> except all the hardware is under your control, and not the merchant
420 2012-05-06 20:25:56 <sipa> vision_: where is the wallet stored in your idea?
421 2012-05-06 20:26:11 <vision_> in the cell
422 2012-05-06 20:26:19 <vision_> (or other computer around the world)
423 2012-05-06 20:26:27 <sipa> how does the cardreader communicate with it?
424 2012-05-06 20:26:33 <sipa> internet?
425 2012-05-06 20:26:36 <vision_> yes
426 2012-05-06 20:26:44 <riush> luke-jr: there.. do we need another testnet faucet? ;)
427 2012-05-06 20:26:50 <luke-jr> riush: maybe, ty
428 2012-05-06 20:27:17 <sipa> vision_: so, if i scan your rfid, and remember its id, and the pin you typed, i can withdraw any amount from your wallet?
429 2012-05-06 20:27:57 <vision_> and if you copy my visa card and remember the pin i typed?
430 2012-05-06 20:28:25 <ali1234> then the nasty men from VISA show up at yor place of business with wrenches
431 2012-05-06 20:28:46 <vision_> you could make a lock in the cell aplication
432 2012-05-06 20:28:48 <vision_> for example
433 2012-05-06 20:29:46 <vision_> you don't like the idea, isn't it?
434 2012-05-06 20:30:22 <BlueMatt> sipa: wait, so why cant we send messages while holding cs_vNodes?
435 2012-05-06 20:30:44 <BlueMatt> sipa: and yea, found the mistake in the cblockstore tweaks of the testsuite
436 2012-05-06 20:31:22 <BlueMatt> also, its depressing when you find simple bugs in libc trying to test bitcoin...
437 2012-05-06 20:31:41 <gmaxwell> BlueMatt: What libc bug did you find?
438 2012-05-06 20:31:57 <BlueMatt> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=671789
439 2012-05-06 20:32:02 <gmaxwell> I don't know if I've ever found a libc bug before. I've found libm bugs, however.
440 2012-05-06 20:32:36 <BlueMatt> getaddrinfo returns different results looking up "127.0.0.1" and "localhost" if you are/arent connected to the internet
441 2012-05-06 20:32:40 <gmaxwell> (crap like exp() returns wrong results or takes 100000x longer for some inputs)
442 2012-05-06 20:32:52 <BlueMatt> ewwww
443 2012-05-06 20:50:34 <pgibson> I'm not sure how to configure my build to not use miniupnpc, would it hurt to just remove the lines that include it from net.cpp?
444 2012-05-06 20:50:47 <BlueMatt> build with USE_UPNP=
445 2012-05-06 20:51:15 <pgibson> awesome!
446 2012-05-06 20:51:26 <sipa> BlueMatt: because you don't need vnodes while processing messages
447 2012-05-06 20:51:42 <sipa> BlueMatt: so vnodes is locked lower
448 2012-05-06 20:51:52 <BlueMatt> well, isnt PushMessage essentially just a callback?
449 2012-05-06 20:51:59 <BlueMatt> oh, it is?
450 2012-05-06 20:54:01 <luke-jr> for qmake, you need USE_UPNP=-
451 2012-05-06 21:24:12 <gmaxwell> BlueMatt: https://launchpad.net/~bitcoin/+archive/bitcoin  < can you update these for 0.6.1?
452 2012-05-06 21:29:10 <sipa> hello gavin
453 2012-05-06 21:29:19 <gavinandresen> hey sipa
454 2012-05-06 21:29:41 <sipa> are we going for a 0.6.2 now, or 0.7.0?
455 2012-05-06 21:29:57 <gavinandresen> I'd like to get the addrman fix in ASAP
456 2012-05-06 21:30:15 <luke-jr> it's in I think
457 2012-05-06 21:30:19 <gavinandresen> ... so a quick 0.6.2 that is 0.6.1 plus addrman cherry-picked is what I think we should do
458 2012-05-06 21:30:32 <luke-jr> the problem now is, so are some other things that invalidate all the translations&
459 2012-05-06 21:30:44 <gavinandresen> that's why I say cherry-picked
460 2012-05-06 21:30:52 <luke-jr> gavinandresen: OK, want me to prepare that then?
461 2012-05-06 21:31:02 <luke-jr> also, what about that other fix?
462 2012-05-06 21:31:12 <sipa> gmaxwell had someone to test the unstuck fix, but it seemed to have removed his block db
463 2012-05-06 21:31:26 <luke-jr> wtf
464 2012-05-06 21:31:29 <gavinandresen> how risk is the unstuck fix?
465 2012-05-06 21:31:33 <gavinandresen> risky
466 2012-05-06 21:32:18 <gmaxwell> luke-jr: ::shrugs:: users!
467 2012-05-06 21:33:11 <sipa> risky: worst case it doesn't work
468 2012-05-06 21:33:14 <gavinandresen> ... or to ask another way:  are we sure the unstuck fix won't do something nasty like create an infinite request-block loop or something?
469 2012-05-06 21:34:02 <sipa> hmk
470 2012-05-06 21:34:09 <gmaxwell> I don't think the unstuck fix is risky all that risky but it should get some testing actually the addrman addition of a bunch of assertions is actually somewhat risky (may convert some silent bugs to hard crashes). I'd really like to see both get a couple days testing at least.
471 2012-05-06 21:34:26 <gavinandresen> assertions are compiled out unless -DDEBUG, aren't they?
472 2012-05-06 21:35:37 <sipa> no, compiled in unless -DNDEBUG
473 2012-05-06 21:35:59 <gavinandresen> right...  that's exactly the kind of thing I'm really good at forgetting....
474 2012-05-06 21:36:33 <sipa> 0.6.0rc4 failed on an assert failure
475 2012-05-06 21:36:36 <gmaxwell> For the future perhaps we could have some kind of soft assert that kicks nodes into safe mode and logs like crazy to use for newly introduced assertions.
476 2012-05-06 21:36:37 <sipa> iirc
477 2012-05-06 21:36:49 <luke-jr> http://codepad.org/6iCEju8s <-- proposed commits to cherry-pick into 0.6.2
478 2012-05-06 21:36:59 <luke-jr> (note, I omitted the assert additions)
479 2012-05-06 21:37:26 <gmaxwell> (perhaps also drops the walletpassphrase out of memory too)
480 2012-05-06 21:38:35 <gmaxwell> I'm happy about the additional asserts I'm just a little concerned that in the process of squashing a very rare bug, we'll introduce a less rare crash on an overeager assert.
481 2012-05-06 21:38:58 <luke-jr> gmaxwell: is there a risk to omitting the asserts?
482 2012-05-06 21:39:09 <DiabloD3> ha ha you said asshurts
483 2012-05-06 21:39:09 <gmaxwell> Though on the subject of very rare, I think the addrman bug is a bit of a timebomb: once nodes have made enough outbound connects they'll hit it much more often.
484 2012-05-06 21:39:36 <gavinandresen> yes, that's why I want it fixed ASAP
485 2012-05-06 21:39:38 <gmaxwell> luke-jr: without them there may be other bugs that silently corrupt memory which will leave nodes still running.
486 2012-05-06 21:40:34 <DiabloD3> gmaxwell: whats that bug do?
487 2012-05-06 21:40:41 <gmaxwell> DiabloD3: what bug?
488 2012-05-06 21:40:56 <DiabloD3> addrman bug
489 2012-05-06 21:41:41 <gmaxwell> Causes mostly boring memory corruption and usually a crash+addr.dat corruption to nodes which have made many successful outbound connections.
490 2012-05-06 21:41:45 <XMPPwocky> addrman, addrman, does whatever an addrman does
491 2012-05-06 21:41:49 <gavinandresen> luke-jr: pulling non-critical bugfixes into 0.6.2 would be a mistake, IMHO.
492 2012-05-06 21:42:27 <gavinandresen> So: how critical is the unstuck fix?  Doesn't seem to be affecting many people, is there any reason to believe that problem will get worse?
493 2012-05-06 21:42:34 <luke-jr> gavinandresen: does d41f22c count as critical? I'm not sure what effects it might have
494 2012-05-06 21:43:23 <gribble> New news from bitcoinrss: TheBlueMatt opened issue 1213 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/issues/1213>
495 2012-05-06 21:44:19 <BlueMatt> gmaxwell: yea, working on it...
496 2012-05-06 21:44:23 <luke-jr> ie, what happens if printf gets more arguments than it expects?
497 2012-05-06 21:45:06 <gavinandresen> luke-jr: what is d41f22c?  (and what's the git command to tell me...)
498 2012-05-06 21:45:25 <BlueMatt> git checkout d...; git log
499 2012-05-06 21:45:43 <luke-jr> 82e6b92  Bugfix: %-12I64d is not valid and causes the parameter to be skipped, use %12"PRI64d" instead
500 2012-05-06 21:45:49 <luke-jr> gavinandresen: git show 82e6b92
501 2012-05-06 21:45:57 <luke-jr> (sorry, d41f22c was the backport of it)
502 2012-05-06 21:46:12 <gmaxwell> luke-jr: that code doesn't get executed for most userrs though IIRC, it requires an argument.
503 2012-05-06 21:46:23 <luke-jr> gmaxwell: right, it requires 2 arguments
504 2012-05-06 21:46:25 <luke-jr> -debug -printpriorities
505 2012-05-06 21:46:54 <luke-jr> but if enabled, I'm not familiar with the side effects of passing extra arguments to printf
506 2012-05-06 21:46:58 <gmaxwell> I expect the consequence is just printing random memory meaning no harm to occasional crashes, plus minor information leaks into the log. Given that it needs -debug -printpriorities  I'm not worried.
507 2012-05-06 21:47:25 <luke-jr> gmaxwell: well, this is *extra* arguments, not too few
508 2012-05-06 21:47:26 <gmaxwell> luke-jr: it's not just extra, it shifts the arguments down one too, IIRC.
509 2012-05-06 21:47:27 <luke-jr> ok
510 2012-05-06 21:47:36 <gmaxwell> extra to vararg is itself completely safe.
511 2012-05-06 21:47:38 <luke-jr> how about bd1aabe  Bugfix: store source address in addrman ?
512 2012-05-06 21:48:01 <DiabloD3> gmaxwell: but I want unlimited sonnections :<
513 2012-05-06 21:50:54 <jine> bitcoin.org/bitcointalk/mtgox down?
514 2012-05-06 21:51:10 <jine> *sigh*
515 2012-05-06 21:51:36 <luke-jr> gavinandresen: 0.6.x branch pushed to stable repo; let me know what else to add before tagging v0.6.2 I guess
516 2012-05-06 21:51:49 <luke-jr> gavinandresen: right now, it's just 56f1e91 (Fix addrman crashes)
517 2012-05-06 21:52:05 <jine> Or my routing sucks.
518 2012-05-06 21:53:53 <luke-jr> devrandom: ping?
519 2012-05-06 21:57:22 <luke-jr> gavinandresen*
520 2012-05-06 21:57:28 <gmaxwell> I'm inclined to pull the reorg fix to master right now in the hopes of getting it more testing. Any objections?
521 2012-05-06 21:57:37 <gavinandresen> gmaxwell: ACK
522 2012-05-06 21:57:50 <gavinandresen> luke-jr: what?
523 2012-05-06 21:58:22 <luke-jr> gavinandresen: any ruling on 82e6b92 and/or bd1aabe? :p
524 2012-05-06 21:58:55 <sausage2> So. How is the dev going?
525 2012-05-06 22:04:53 <sipa> i'm quite sure the asserts are harmless (i didn't trigger any while reproducing the addrman crash, and i believe all code paths have been tested now
526 2012-05-06 22:05:02 <sausage2> gmaxwell: hi, how's it going? nice wather today, huh?
527 2012-05-06 22:05:20 <sipa> on the other hand, addrman failures are not that critical
528 2012-05-06 22:05:28 <sausage2> So, the dev is going pretty well, isnt it? :)
529 2012-05-06 22:06:16 <sausage2> Isnt it so gmaxwell? You're a BTC-dev too, I've heard. An eminent one.
530 2012-05-06 22:06:21 <sipa> the other bug fix was just not storing a value to base bucket decisions on
531 2012-05-06 22:06:40 <sipa> i don't see how it could influence much
532 2012-05-06 22:06:59 <sausage2> so, how do I use JSON for btc?
533 2012-05-06 22:06:59 <sipa> but that's certainly not critical
534 2012-05-06 22:07:03 <sausage2> the rpc service?
535 2012-05-06 22:07:25 <gmaxwell> sausage2: https://en.bitcoin.it/wiki/API_reference_%28JSON-RPC%29
536 2012-05-06 22:07:45 <sausage2> is it so gmaxwell? Can I do it in linux?
537 2012-05-06 22:08:30 <sausage2> Thanks gmaxwell, you are a very kind person :)
538 2012-05-06 22:08:44 <luke-jr> so just 56f1e91(addrman crash fix) and 6860133(stuck download fix) then?
539 2012-05-06 22:08:44 <sipa> ?
540 2012-05-06 22:08:58 <gmaxwell> (he was spamming me in PM with scatological nonsense)
541 2012-05-06 22:09:08 <sipa> luke-jr: fine by me
542 2012-05-06 22:10:02 <gavinandresen> luke-jr: I'll create a 0.6.2 branch on github, version numbers in READMEs/etc need to be bumped, too
543 2012-05-06 22:10:28 <luke-jr> gavinandresen: I already made a branch&
544 2012-05-06 22:10:42 <gavinandresen> okey doke
545 2012-05-06 22:12:31 <gavinandresen> https://github.com/bitcoin/bitcoin/tree/0.6.2
546 2012-05-06 22:13:10 <luke-jr> &
547 2012-05-06 22:13:33 <gavinandresen> So: how about we all compile/run that branch overnight, and assuming no issues by tomorrow night I'll tag and we'll gitian-build it as 0.6.2?
548 2012-05-06 22:14:24 <luke-jr> gavinandresen: so you're including the non-critical bugfix and risky assertions?
549 2012-05-06 22:14:27 <gmaxwell> We also need to test new nodes with this branch, and groups of nodes connected to each other with this branch (to check for inv loops)
550 2012-05-06 22:14:42 <gmaxwell> luke-jr: The assertions are not pure risk. There is risk either way.
551 2012-05-06 22:14:57 <luke-jr> gmaxwell: there is no *additional* risk to leaving them out
552 2012-05-06 22:15:50 <sipa> i wouldn't mind leaving them out
553 2012-05-06 22:16:11 <gmaxwell> I don't care about 'additional' risk, I care about the total resulting safty and stability. At least the risk from assertions is just DOS and not data corruption of security compromises.  I think the risk from them is adequately addressed by testing in any case.
554 2012-05-06 22:17:21 <gavinandresen> I want to fail early, so assertions stay in
555 2012-05-06 22:18:04 <gmaxwell> with how slowly people tend to deploy it's not that scarry.
556 2012-05-06 22:19:45 <luke-jr> sounds like we're good to go until tomorrow then
557 2012-05-06 22:22:59 <gmaxwell> luke-jr: actually we could use gitian builds of that branch, so we can nag more people to run them.
558 2012-05-06 22:23:35 <luke-jr> am I the only one with working gitian now? :P
559 2012-05-06 22:23:41 <sipa> mine works
560 2012-05-06 22:23:58 <BlueMatt> give me like 2 days
561 2012-05-06 22:24:15 <riush> why is there no address displayed for mined/coinbase tx?
562 2012-05-06 22:24:57 <luke-jr> riush: known bug. want to fix it?
563 2012-05-06 22:25:01 <freewil> because i think that part of the code was made with the assumption that a coinbase wouldnt be sent
564 2012-05-06 22:25:10 <luke-jr> bbiab (hopefully with gitina builds)
565 2012-05-06 22:25:18 <freewil> ... and you would be generating it for yourself
566 2012-05-06 22:25:34 <sipa> gavinandresen: no annotated tag for 0.6.2?
567 2012-05-06 22:25:50 <sipa> (git describe says v0.6.1-5-gf367e5d)
568 2012-05-06 22:26:35 <sipa> apology for offtopicness, but i find this hilarious: http://www.youtube.com/watch?v=w1gusAIRN8M
569 2012-05-06 22:27:00 <BlueMatt> thats very offtopic...
570 2012-05-06 22:27:40 <gavinandresen> sipa: I don't want to tag it until it gets a little testing
571 2012-05-06 22:28:04 <sipa> ok
572 2012-05-06 22:30:18 <sipa> i'll tag it myself as 0.6.2rc1, ok?
573 2012-05-06 22:30:32 <sipa> (just locally, to get it in the version info)
574 2012-05-06 22:30:41 <luke-jr> sipa: your build won't match mine if you do
575 2012-05-06 22:30:47 <sipa> ah, right
576 2012-05-06 22:30:52 <graingert> gavinandresen: do you think "bitcoin.org" could be a commercial sponsor for a University Group Design Project?  Every student on a four year course has to participate in a GDP.  Which will either be a standard issue exercise or a problem submitted by a commercial sponsor.  A commercial sponsor suggests a problem, eg "bitcoin multicast":  something that they want doing that should take 5 people 2/3 of a semester to do.
577 2012-05-06 22:31:02 <graingert> "(01:30:27) gavinandresen: try gmaxwell or sipa, maybe they have cool project ideas...."
578 2012-05-06 22:31:36 <luke-jr> O.o
579 2012-05-06 22:31:54 <luke-jr> graingert: bitcoin.org is tcatm or sirius anyway
580 2012-05-06 22:32:00 <theorbtwo> Bitcoin multicast actually sounds like an interesting idea.
581 2012-05-06 22:32:05 <graingert> gavinandresen: okay if I pastbin the entire chat log from 01:24:25
582 2012-05-06 22:32:12 <theorbtwo> Too bad multicast doesn't actually work on the general internet.
583 2012-05-06 22:32:20 <graingert> theorbtwo: it works over JANET
584 2012-05-06 22:32:27 <graingert> ie all of UK
585 2012-05-06 22:32:43 <luke-jr> theorbtwo: yet to be seen on IPv6
586 2012-05-06 22:32:49 <theorbtwo> graingert: .ac.uk.
587 2012-05-06 22:32:58 <gavinandresen> graingert: sure, I don't mind if everybody sees me being a curmudgeon....
588 2012-05-06 22:33:01 <graingert> sorry yes
589 2012-05-06 22:33:02 <theorbtwo> luke-jr: By my definitions, ipv6 isn't the general internet.
590 2012-05-06 22:33:10 <graingert> http://pastie.org/3871060
591 2012-05-06 22:33:16 <luke-jr> theorbtwo: you must be living in 1990
592 2012-05-06 22:33:31 <graingert> ipv6 is as general internet as you can get
593 2012-05-06 22:33:40 <graingert> you've got google
594 2012-05-06 22:33:42 <graingert> and facebook
595 2012-05-06 22:33:45 <graingert> what more do you want
596 2012-05-06 22:33:49 <theorbtwo> What I mean is that you can't expect J. Random user, connected to a commercial DSL modem, to have working multicast.
597 2012-05-06 22:34:04 <graingert> theorbtwo: no but it's not too far off
598 2012-05-06 22:34:47 <theorbtwo> graingert: widespread ipv6 is one of those things that's always a couple of years off.
599 2012-05-06 22:34:52 <Dagger2> graingert: I want Youtube and Yahoo too!
600 2012-05-06 22:35:01 <graingert> back on topic?
601 2012-05-06 22:35:06 <theorbtwo> Back on topic.
602 2012-05-06 22:35:17 <graingert> cool ideas for GDP, gmaxwell sipa bitinstant
603 2012-05-06 22:35:42 <graingert> mtgox is currently deliberating
604 2012-05-06 22:36:06 <theorbtwo> graingert: Considering sponsoring a gdp, you mean?
605 2012-05-06 22:36:13 <graingert> yes
606 2012-05-06 22:36:30 <theorbtwo> Interesting.
607 2012-05-06 22:36:33 <sipa> what does sponsoring imply?
608 2012-05-06 22:37:33 <graingert> sipa: you're telepresence is required at a few meetings: requirements and demo presentation
609 2012-05-06 22:37:37 <graingert> your*
610 2012-05-06 22:38:13 <theorbtwo> I do wonder if you / one could work out a way to make quick bitcoin for POS applications, which would need to transparently interoperate with normal bitcoin.
611 2012-05-06 22:38:47 <graingert> theorbtwo: that's been done
612 2012-05-06 22:38:59 <graingert> or do you mean a SPV?
613 2012-05-06 22:39:40 <theorbtwo> I've not been following things closely, apparenytly.
614 2012-05-06 22:40:00 <graingert> "Do an SPV: Go!" would be an interesting project request
615 2012-05-06 22:40:41 <sipa> Do a barrel roll!
616 2012-05-06 22:40:44 <XMPPwocky> solar powered vehicle?
617 2012-05-06 22:40:53 <graingert> !google SPV bitcoin
618 2012-05-06 22:40:53 <gribble> Thin Client Security - Bitcoin: <https://en.bitcoin.it/wiki/Thin_Client_Security>; Alternative Chains - Bitcoin: <https://en.bitcoin.it/wiki/Alternative_Chains>; Scalability - Bitcoin: <https://en.bitcoin.it/wiki/Scalability>
619 2012-05-06 22:40:54 <TiggrBot> [SPV bitcoin] http://webcache.googleusercontent.com/search?q=cache:9VZRyUeW_N4J:https://en.bitcoin.it/wiki/Thin_Client_Security+SPV+bitcoinhl=ct=clnk (Cached)
620 2012-05-06 22:41:04 <sipa> graingert: you start by using bitcoinj, then ...
621 2012-05-06 22:41:14 <BlueMatt> 1. use bitcoinj
622 2012-05-06 22:41:16 <BlueMatt> 2. ???
623 2012-05-06 22:41:18 <BlueMatt> 3. profit
624 2012-05-06 22:41:28 <graingert> 4. finney attack
625 2012-05-06 22:41:30 <graingert> 5. cry
626 2012-05-06 22:41:38 <sipa> graingert: ipv6 multicast would actually be very interesting, imho
627 2012-05-06 22:41:46 <graingert> truefax
628 2012-05-06 22:42:04 <graingert> and I think implementing it as a fake client would probably be the easiest
629 2012-05-06 22:42:14 <graingert> for a prototype
630 2012-05-06 22:42:30 <graingert> eg you add 127.0.0.1:port as a peer
631 2012-05-06 22:42:47 <graingert> and that peer pretends to be a bitcoin node interested in tx and giving you tx
632 2012-05-06 22:43:03 <sipa> so, a gateway
633 2012-05-06 22:43:06 <graingert> when in actual fact it's communicating with everyone else on JANET
634 2012-05-06 22:43:14 <graingert> yes
635 2012-05-06 22:43:16 <sipa> that's probably the easiest way to implement it
636 2012-05-06 22:43:19 <graingert> yes
637 2012-05-06 22:43:45 <graingert> super
638 2012-05-06 22:43:52 <graingert> I have a supervisor who <3 multicast
639 2012-05-06 22:44:57 <graingert> and <3 ipv6
640 2012-05-06 22:45:49 <sipa> i'll see if i can come up with some other ideas
641 2012-05-06 22:49:04 <da2ce7> I would love to have a feature where you enter a ipv6 address and a public key hash... and your client contacts the remote bitcoin client and that remote client retunrs a signed new public key to send the coins to.
642 2012-05-06 22:49:17 <da2ce7> *signed by their main public key.
643 2012-05-06 22:49:38 <da2ce7> kinda like the orignal send-to-ip... but secure.
644 2012-05-06 22:49:51 <sipa> why does it need to be ipv6?
645 2012-05-06 22:49:57 <luke-jr> da2ce7: a script hash would be better
646 2012-05-06 22:50:00 <luke-jr> sipa: because IPv4 is dead
647 2012-05-06 22:50:11 <luke-jr> da2ce7: so N-of-M can be used if desired
648 2012-05-06 22:50:20 <sipa> it should just return a script
649 2012-05-06 22:50:26 <sipa> not an address or a hash
650 2012-05-06 22:50:42 <luke-jr> sure
651 2012-05-06 22:50:49 <luke-jr> I mean to authenticate the connection
652 2012-05-06 22:50:58 <da2ce7> well it depends how we design it... however I think that scritps should be stored in a DHT.
653 2012-05-06 22:51:07 <da2ce7> *stored.
654 2012-05-06 22:51:10 <sipa> ooh, a DHT!!!
655 2012-05-06 22:51:27 <gmaxwell> graingert: I have a whole speculative features list page though I'm probably due to update it.... a few things on it are done or very near done, and of course I've had more thoughts since then.
656 2012-05-06 22:52:15 <gmaxwell> da2ce7: you do realize that you've just trigged the bayesian idiot detector here, right? or was that the point.
657 2012-05-06 22:52:35 <graingert> stabbing machine?
658 2012-05-06 22:52:53 <sipa> luke-jr: also, however much i'd like your statement about ipv4 to be true, i think you should stop dreaming
659 2012-05-06 22:53:20 <graingert> sipa: much the same way one can still slide use a dead horse as a sledge
660 2012-05-06 22:53:22 <gmaxwell> luke-jr: I really wish it were dead, instead carriers are deploying massive nat boxes (or gearing up to do so where they haven't yet)
661 2012-05-06 22:53:24 <graingert> ipv4 is dead
662 2012-05-06 22:53:31 <gmaxwell> graingert: https://en.bitcoin.it/wiki/User:Gmaxwell/features
663 2012-05-06 22:53:33 <da2ce7> well won't there be private scripts, that are only known on spend, and scripts that are public from the start?
664 2012-05-06 22:53:44 <sipa> da2ce7: read BIP16
665 2012-05-06 22:54:05 <sipa> even if ipv6 starts to take off now (which i assume it will), most services on the internet will remain on IPv4 for years to come, just to remain accessible to both
666 2012-05-06 22:55:11 <graingert> IPv6 at all would be nice :p
667 2012-05-06 22:55:20 <sipa> graingert: done
668 2012-05-06 22:55:26 <graingert> oh really
669 2012-05-06 22:55:36 <graingert> gmaxwell: you need to keep that up to date
670 2012-05-06 22:55:47 <sipa> graingert: https://github.com/bitcoin/bitcoin/pull/1021
671 2012-05-06 22:55:50 <BlueMatt> pulling bitcoin: uris into a full secure payment system with support for multisig on multiple devices passing info back and forth and all would be cool
672 2012-05-06 22:56:43 <gavinandresen> I still prefer a mime-type / file download to bitcoin: URIs...
673 2012-05-06 22:56:45 <gmaxwell> graingert: I just went and updated it a bit.
674 2012-05-06 22:57:59 <sipa> combinding it with BIP10 would be cool
675 2012-05-06 22:58:04 <sipa> *combing
676 2012-05-06 22:58:08 <sipa> **combining
677 2012-05-06 22:58:13 <graingert> why not just use https
678 2012-05-06 22:58:18 <graingert> or webid
679 2012-05-06 22:58:34 <BlueMatt> gavinandresen: except uris are handled more gracefully in many cases...
680 2012-05-06 22:59:00 <sipa> it can well be a URI that is required to point to a file with a particular mime-type
681 2012-05-06 22:59:09 <gmaxwell> graingert: look at all the bitcoin sites with invalid HTTPS certs, or the fact that 100% of them are downgrading attack vulnerable... I couldn't even talk any of them into using HSTS.
682 2012-05-06 22:59:30 <graingert> you'd demand https at the protocol
683 2012-05-06 22:59:34 <graingert> ie like webid
684 2012-05-06 22:59:49 <da2ce7> ok... I stil feel stupid... you need the serialized script to spend a output. if you loose the script the the output is unspendable. So it is as-important to know the scripts as the private keys, but they don't need to remain private.  Why is it such as awful idea to store them in DHT?
685 2012-05-06 22:59:57 <graingert> eg to pay to a domain
686 2012-05-06 23:00:06 <BlueMatt> sipa: that just makes it overcompliced, its ugly to add a ton of crap to a uri, but doable...
687 2012-05-06 23:00:19 <graingert> type the domain in the box, it looks up https://domain/.well-know/bitcoin
688 2012-05-06 23:00:23 <graingert> type the domain in the box, it looks up https://domain/.well-known/bitcoin
689 2012-05-06 23:00:38 <graingert> that file contains a json file with a new address each time it's called
690 2012-05-06 23:00:50 <BlueMatt> anyone have the link to gavinandresen's gist where we were discussing payment systems a while back?
691 2012-05-06 23:00:53 <gmaxwell> da2ce7: because if you know the private key then you know the script.  There is also no such thing as an usably attack resistant DHT, which is why DHT is usually an idiot filter trigger in the context of bitcoin.
692 2012-05-06 23:01:23 <gavinandresen> BlueMatt: I probably do :)
693 2012-05-06 23:01:31 <da2ce7> why don't reqire every script-has of the DHT to be already included in the block-chain?
694 2012-05-06 23:01:43 <da2ce7> *hash
695 2012-05-06 23:02:03 <graingert> I think https urls is the way to go for payments
696 2012-05-06 23:02:16 <graingert> ie if I want to add money to my mtgox account
697 2012-05-06 23:02:34 <graingert> I would enter https://mtgox.com/user/graingert in my client
698 2012-05-06 23:02:39 <sipa> graingert: if you trust the SSL PKI
699 2012-05-06 23:02:50 <gavinandresen> BlueMatt:  this one?  https://gist.github.com/2217885
700 2012-05-06 23:02:54 <BlueMatt> graingert: avoiding trusting _anyone_ would be nice
701 2012-05-06 23:02:54 <sipa> (not that i mean to say you shouldn't, but certainly not everyone will)
702 2012-05-06 23:02:58 <gmaxwell> The SSL PKI is demonstrably untrustworthy.
703 2012-05-06 23:03:03 <BlueMatt> gavinandresen: thats the one!
704 2012-05-06 23:03:15 <graingert> BlueMatt: that would be nice but you'd need to trust mtgox.com if you were sending to them
705 2012-05-06 23:03:23 <BlueMatt> well, yea
706 2012-05-06 23:03:29 <BlueMatt> but minimal trust is the goal
707 2012-05-06 23:03:36 <graingert> you can't do it any other way
708 2012-05-06 23:03:45 <graingert> if I was a shop I'd have to deliver my bitcoin address to you
709 2012-05-06 23:03:51 <graingert> and the best way at the moment is ssl in a page
710 2012-05-06 23:03:58 <gmaxwell> da2ce7: great you just encouraged blockchain spamming in order to 'register' addresses, address reuse which hurts anonymity in order to leverage existing registration, ... and for no real purpose because you didn't need to store anything more to begin with.
711 2012-05-06 23:04:24 <graingert> okay so the cool challenges are
712 2012-05-06 23:04:34 <graingert> nicer secure payment system
713 2012-05-06 23:04:40 <graingert> and IPv5 multicast
714 2012-05-06 23:04:47 <graingert> durp
715 2012-05-06 23:04:48 <graingert> 6*
716 2012-05-06 23:05:02 <sipa> da2ce7: if you own an address, it means you generated the script associated with it (to be able to give out the hash) using the private keys you have
717 2012-05-06 23:05:20 <gmaxwell> (and to whatever extent you did, even if you remove the DOS vulnerability by requring the chain you didn't really save yourself much because the distributed storage isn't guarenteed, so you still have to save it yourself)
718 2012-05-06 23:05:33 <sipa> da2ce7: assuming your client does that in a minially-controlled way, it can easily recreate the scripts
719 2012-05-06 23:05:37 <sipa> given the keys
720 2012-05-06 23:05:59 <da2ce7> sipa: but if we have  2-of-3 script, how can I recreate that determeistaly?
721 2012-05-06 23:06:32 <graingert> anyone want to write that up/be a sponsor? sipa?
722 2012-05-06 23:07:00 <gmaxwell> graingert: I don't think multicast is espeically interesting.
723 2012-05-06 23:07:00 <sipa> da2ce7: anyone of the 3 owners of that address have an interest in keeping the script around, as much as they have an interest in keeping their private keys around
724 2012-05-06 23:07:09 <graingert> gmaxwell: :(
725 2012-05-06 23:07:35 <gmaxwell> (I mean, it's generally interesting, but not for bitcoin the reason is that the inv process makes bitcoin very transmission efficient)
726 2012-05-06 23:08:07 <graingert> that's true
727 2012-05-06 23:08:55 <sipa> gavinandresen: just realized, my build script doesn't call gsign unless it is a build with a version tag :s
728 2012-05-06 23:09:14 <gmaxwell> da2ce7: You should just consider the script to be part of the private key.  In bitcoin we call the scripts scriptpubkey and such for that reason.
729 2012-05-06 23:09:36 <gmaxwell> da2ce7: the crazy proposal on bitcoin talk which actually removes ECDSA entirely from bitcoin (MAVE) is pretty enlightening in that regard.
730 2012-05-06 23:09:43 <amandapanda> hey
731 2012-05-06 23:13:38 <sipa> gmaxwell: in theory, you could replace the inv/tx/block system by just a block/tx transmit on multicast
732 2012-05-06 23:13:47 <graingert> gmaxwell: ?
733 2012-05-06 23:13:48 <sipa> but it's probably way too easy to exploit
734 2012-05-06 23:13:51 <graingert> gmaxwell: how?
735 2012-05-06 23:14:07 <graingert> amandapanda: hey
736 2012-05-06 23:14:59 <gribble> New news from bitcoinrss: TheBlueMatt opened pull request 1214 on bitcoin/bitcoin <https://github.com/bitcoin/bitcoin/pull/1214>
737 2012-05-06 23:15:14 <gmaxwell> graingert: the same way namecoin prevents claim jumping.
738 2012-05-06 23:15:44 <graingert> have you got a link to that?
739 2012-05-06 23:16:20 <gmaxwell> graingert: https://bitcointalk.org/index.php?topic=76073.20  (a lot of the rest of his paper is meh... he has a high spew to research ratio)
740 2012-05-06 23:16:38 <gmaxwell> oops page one of that thread, of course.
741 2012-05-06 23:19:07 <graingert> got an abstract on the "clever bit"
742 2012-05-06 23:23:39 <gmaxwell> At any moment, each address has a secret matched with it. The H(secret) is stored in the chain.  When that address whants to send it first announces H(secret+H(txn)+H(new secret)) + txn, and once that is good and burried it announces secret,H(new secret).
743 2012-05-06 23:24:12 <graingert> oh
744 2012-05-06 23:24:15 <graingert> that's clever
745 2012-05-06 23:24:19 <gmaxwell> The advantage of this system is that it's very very fast. The disadvantage is that a >50% attacker can steal everyone's money for as far back as he can cut the chain.
746 2012-05-06 23:24:50 <graingert> but not an advantage in bitcoin
747 2012-05-06 23:24:57 <graingert> as that's not the bottleneck
748 2012-05-06 23:24:59 <graingert> afaik
749 2012-05-06 23:25:01 <graingert> it's disk
750 2012-05-06 23:25:09 <gmaxwell> it's also resistant to attacker by quantum computers (except in so far as a QC enabled attacker can mount a >50% attack), but without a lot of space bloat that you'd get from a real QC resistant signature function.
751 2012-05-06 23:26:02 <gmaxwell> Oh I agree.. or rather, if you allow the txn volume to go high enough that signauture validation is a bottleneck, then you throughly have lost decenteralization due to storage requirements. but it's still a cute idea if not useful.
752 2012-05-06 23:27:42 <gmaxwell> The whole secret management part also discourages using fresh addresses for everything.. breaks privacy.
753 2012-05-06 23:29:45 <graingert> so what's the txn then
754 2012-05-06 23:29:58 <graingert> how is that generated from secret_b
755 2012-05-06 23:31:47 <graingert> gmaxwell: ^
756 2012-05-06 23:33:39 <gmaxwell> graingert: the txn is just like a regular bitcoin transaction. Basically the idea is that you commit to preimages in the blockchain, and then reveal the secrets to prove that your transactions are yours.  (I think his actual protocol had one more step than my toy example, in order to close a DOS attack, but thats the basic idea)
757 2012-05-06 23:34:02 <graingert> but how do you send money to secrets?
758 2012-05-06 23:34:36 <gmaxwell> You send to an address. which already has a preregistered secret in the chain. You don't give anyone your address until you've successfully registered it.
759 2012-05-06 23:34:49 <graingert> oh I see
760 2012-05-06 23:35:04 <graingert> and you can only get that by mining it yourself
761 2012-05-06 23:36:20 <gmaxwell> You dont have to mine it yourself no one has any incentive to claim jump your address because you won't use it unless you see that you're successful in getting it in.  I think he proposed some external POW system (e.g. not mining) to rate limiting the production of new addresses.
762 2012-05-06 23:36:33 <graingert> oh
763 2012-05-06 23:36:36 <graingert> that makes sense
764 2012-05-06 23:38:34 <luke-jr> gmaxwell: my carrier is IPv6 only with NAT64
765 2012-05-06 23:38:48 <gmaxwell> luke-jr: Are you using a cellphone in japan?
766 2012-05-06 23:38:52 <luke-jr> gmaxwell: USA cellphone
767 2012-05-06 23:39:03 <luke-jr> T-Mobile
768 2012-05-06 23:39:15 <gmaxwell> Ah, I knew that AT&T was going to do that on the LTE stuff, didn't know that anyone else had done it yet.
769 2012-05-06 23:39:20 <luke-jr> the only USA carrier compliant with international 3G frequencies
770 2012-05-06 23:40:18 <gmaxwell> the cell providers are a bit crazy most are giving each phone three IPs on their LTE networks.
771 2012-05-06 23:40:35 <gmaxwell> (one for internet, one for voice, one for management)
772 2012-05-06 23:40:40 <luke-jr> 03b23aea05a107448512a5fbf791529021b950b901fc10ef0712a36fde03da10  build/out/bitcoin-0.6.2-win32-setup.exe
773 2012-05-06 23:40:54 <luke-jr> hmm. was I supposed to save the actual file for that? <.<
774 2012-05-06 23:41:03 <gmaxwell> hahah
775 2012-05-06 23:41:27 <luke-jr> XD
776 2012-05-06 23:41:38 <luke-jr> I didn't actually delete it, but I did shutdown
777 2012-05-06 23:41:49 <luke-jr> in hopes of minimizing my nested-KVM-panic risk
778 2012-05-06 23:43:41 <luke-jr> http://luke.dashjr.org/programs/bitcoin/files/bitcoind/0.6.2/test/rc1/
779 2012-05-06 23:44:04 <luke-jr> (uploading in progress)
780 2012-05-06 23:55:16 <amiller> so gmaxwell, if you have any attention available for this proof-of-work discussion
781 2012-05-06 23:55:42 <amiller> i'm not really sure where the essence of our disagreement is
782 2012-05-06 23:56:25 <amiller> it's better for bitcoin the more people that can independently validate transactions and blocks
783 2012-05-06 23:56:49 <amiller> since you get reimbursed for investing in mining hardware, lots of people will have whatever is required to mine
784 2012-05-06 23:57:16 <amiller> so if the proof-of-work requires a transaction validation machine, then lots of people will be able to validate transactions
785 2012-05-06 23:59:36 <gmaxwell> If ~regular users without investements in special mining hardware can not fully validate transactions then bitcoin has lost its value???because it won't be very decenteralized anymore.
786 2012-05-06 23:59:58 <amiller> regular users (without a coin database) can still verify transactions if they're handed O(log N) proofs