< petertodd> bsm117532: pull-reqs accepted :) I'm gonna bother the guys who said they were doing a segwit pull later
< GitHub61> [bitcoin] jmcorgan opened pull request #8284: Backport remaining commits for out-of-tree builds from master to 0.12 branch (0.12...build-oot-0.12) https://github.com/bitcoin/bitcoin/pull/8284
< GitHub81> [bitcoin] jonasschnelli pushed 2 new commits to master: https://github.com/bitcoin/bitcoin/compare/3685e0c1da29...6a87eb0e4b47
< GitHub81> bitcoin/master a7897c0 Wladimir J. van der Laan: qt: Remove client name from debug window...
< GitHub81> bitcoin/master 6a87eb0 Jonas Schnelli: Merge #8281: qt: Remove client name from debug window...
< GitHub119> [bitcoin] jonasschnelli closed pull request #8281: qt: Remove client name from debug window (master...2016_06_qt_remove_client_name) https://github.com/bitcoin/bitcoin/pull/8281
< jonasschnelli> wumpus: merged #8281 after the freeze (trivial non-feature). I hope this is okay...
<@wumpus> sure :)
<@wumpus> removing things is never a problem after the freeze
<@wumpus> thanks for testing
< NicolasDorier> looooooong timeout
< NicolasDorier> it was blocked, and strangely, when, I hit "Enter" in the console windows, all the socket dropped
< NicolasDorier> it is not the first time where I see a windows app completely frozen and unlocked by hitting enter in the console windows
< NicolasDorier> I guess the long timeout and this problem are related. But not really sure it comes from bitcoind.
<@wumpus> I've heard of the problem before, seems only to happen with windows bitcoind.exe
<@wumpus> one guess would be that the network code somehow ends up reading from the console file descriptor, blocking it
<@wumpus> pretty rare and annoying to debug, I hope cfields's libevent networking will make it go away
< GitHub75> [bitcoin] laanwj opened pull request #8285: windows: Add testnet link to installer (master...2016_06_testnet_link_windows) https://github.com/bitcoin/bitcoin/pull/8285
< jonasschnelli> On my XEON E3 machine I get the deadlock assertions all the time...
<@wumpus> jonasschnelli: which deadlock assertion?
<@wumpus> in networking?
< jonasschnelli> Can't really run master with --enable-debug on that machine..
< MarcoFalke> Could the testnet windows link cause any problems/confusion when a user does not know the distinction between mainnet and testnet?
< sipa> It may result in people accidentally starting to use it, and testnet gaining exchange value again :p
< jonasschnelli> hehe...
< jonasschnelli> We could move it into a folder called "development/" or so in the windows startmenu?!
< jonasschnelli> No strong opinion though.
< GitHub67> [bitcoin] MarcoFalke opened pull request #8286: [doc] typos, READMEs, comments, headers (master...Mf1607-trivialPre13) https://github.com/bitcoin/bitcoin/pull/8286
<@wumpus> sigh
<@wumpus> couldn't we have that discussion *before* I started implementing it? I wouldn't have bothered
<@wumpus> I'd say a green icon and (testnet) would eb clear enough
<@wumpus> but if not, never mind
< MarcoFalke> It is helpful for 0.13
< MarcoFalke> but we may want to revert it for 0.14
<@wumpus> why?
<@wumpus> the issue was that people don't test on testnet enough
< sipa> i think it is fine
<@wumpus> I'm not sure why that'd be version dependent
< MarcoFalke> On the other hand. The people that can't figure out how to start testnet are not going to do fancy pentesting
< sipa> i think testnet gaining value would be a fantastic problem to have
< sipa> because it would imply it is being used
<@wumpus> so is it anywhere in the testnet GUI not clear that it is running testnet?
<@wumpus> the icon color is different, the splash is different, the window title is different
< sipa> it doesn't even accept the same addresses
<@wumpus> right!
<@wumpus> I'd say if people get confused by that, that's toobad for them
< sipa> nobody will accidentally try to pay from the testnet client
< sipa> if there are complaints about it causing confusion we can address them
<@wumpus> nodoby can have themselves paid out to a testnet client
< sipa> i think it's a good thing regardless to give testnet more exposure
<@wumpus> anyhow, we should have had this discussion sooner, I wouldn't have dived into all the windows bullshit
< sipa> i think it's all fine
< MarcoFalke> Sure, nothing fatal can happen.
<@wumpus> it's not like it is any fun
<@wumpus> I thought there was agreement to do it
< sipa> there was
< sipa> and there is, as far as i'm concerned
<@wumpus> let's not be over-worried, if you should be worried about something it'd be a fatal bug in any new feature of 0.13
< MarcoFalke> Agree, we should totally try that and see what happens.
< MarcoFalke> Maybe we can then close https://github.com/bitcoin/bitcoin/issues/3172
<@wumpus> that one was never entirely clear to me, it's not a request for bitcoin core feature
<@wumpus> (at least how I read it it could be implemented as an external script in terms of rpc)
<@wumpus> jonasschnelli: thanks for the new icon, going to try
< sipa> I think 3172 proposes a service which we run
< sipa> the code for which could be included in the reposiutory
<@wumpus> I don't think it's something we want to encourage though
<@wumpus> bouncing coins that way
<@wumpus> people may think it's a valid way to refund
<@wumpus> if someone wants to make a service like that, they can, anyhow
< paveljanik> the only confusion can be sample address in the Pay To in the Send tab. It is 1NS... which is a bad example in testnet.
< sipa> we could add an exampleaddress std::string to chainparams
<@wumpus> or heck, just remove the example address
< paveljanik> I'd prefer removal, yes :-)
<@wumpus> especially with the new address formats coming up
< sipa> also fine
< paveljanik> or we can take an address used in coinbase of the genesis block
< paveljanik> 8)
<@wumpus> lol, it doesn't have to be a valid address at al
<@wumpus> it's not like people are supposed to send to it
<@wumpus> (it's luckily not easy to do that, as it will disappear if you start typing)
< paveljanik> yes, but better to show them something they are supposed to enter in the field.
< paveljanik> you can't even copy&paste it... so this is good.
<@wumpus> although I've never even thought about people trying to do that - removing the example address is definitely safest
<@wumpus> indeed, you can't
<@wumpus> you could make a screenshot then type it over
<@wumpus> or OCR it :-)
< paveljanik> providing example address can be of some value though, for some people...
<@wumpus> or just memorize it
<@wumpus> nah, all in all, it's probably a liability
<@wumpus> I also used to have that opinion that it had some value, but that value may well be slightly negative :)
<@wumpus> can we generate an address that *looks* like an address for current network
<@wumpus> but is invalid?
<@wumpus> we can't change that message anymore for 0.13 as its part of a translation strnig
<@wumpus> but the address is parametrized
<@wumpus> maybe instead of a full address show the first few characters and ...
< paveljanik> then we can print the address for pubkey 0.
< paveljanik> privkey of course
< paveljanik> 1HZ or so it was...
<@wumpus> pubkey 0 would work too, probably easier to generate
< paveljanik> isn't pubkey 0 valid? ;-)
< sipa> the example address IS an invalid address i hope
< paveljanik> privkey 0 is valid when sending 8)
< sipa> like, its checksum fails
<@wumpus> IIRC it is valid
<@wumpus> src/test/data/base58_encode_decode.json:["00eb15231dfceb60925886b67d065299925915aeb172c06647", "1NS17iag9jJgTHD1VXjvLCEnZuQ3rJDE9L"],
<@wumpus> ok it's just valid base58, not more, you may be right about the checksum
<@wumpus> simple enough to use the same principle to generate an address on the fly with the correct network-specific prefix
< GitHub69> [bitcoin] MarcoFalke opened pull request #8287: [wallet] Set fLimitFree = true (master...Mf1607-walletLimitFree) https://github.com/bitcoin/bitcoin/pull/8287
< MarcoFalke> paveljanik: Off-by-one for boolean is fatal. I am pretty sure I got it right :)
< sipa> In other news: C++17 removes the ++ operator for booleans
< paveljanik> MarcoFalke, ;-) So in the sendrawtransaction you have just set it to false as it was before, only named the constant.
< paveljanik> so in the commit with =true, you applied +bool fLimitFree = false; ;-)
< paveljanik> a bit unexpected :-)
< MarcoFalke> The rpc is just refactoring
< paveljanik> but correct :-)
< paveljanik> yes
< MarcoFalke> Hopefully someone will remove that soon
< paveljanik> but there are usecases where you want to accept free tx into your mempool. Especially when it is submitted locally...
< MarcoFalke> Then do it via sendrawtx
< paveljanik> right now, yes. But after removal?
< MarcoFalke> 7533 would make it such that you could overwrite it
< MarcoFalke> Instead of an infinite list of params you just pass a set
< GitHub175> [bitcoin] laanwj opened pull request #8288: qt: Network-specific example address (master...2016_06_network_based_example_address) https://github.com/bitcoin/bitcoin/pull/8288
<@wumpus> voila
< paveljanik> FYI, github fixed the certificate on githubusercontent.com....
< GitHub15> [bitcoin] roques opened pull request #8289: bash-completion: Adapt for 0.12 and 0.13 (master...completion) https://github.com/bitcoin/bitcoin/pull/8289
< luke-jr> wumpus: what is your LC_COLLATE? :/
< GitHub73> [bitcoin] MarcoFalke opened pull request #8291: [util] CopyrightHolders: Check for untranslated substitution (master...Mf1607-utilCopy) https://github.com/bitcoin/bitcoin/pull/8291
< NicolasDorier> wumpus: I'm not sure the timeout comes from network code. It already happened to me before any connection were done. (ie, when I rebooted, during the block verification phase) I don't think we should bother though, I'm almost sure it happened to me on other projects, and I've not seen it happening without printtoconsole