12016-11-12T00:03:57 <BlueMatt> cfields: hum, why do we even need fSuccessfullyConnected?
22016-11-12T00:04:15 *** d9b4bef9 has joined #bitcoin-core-dev
32016-11-12T00:04:15 <BlueMatt> cfields: cant we just set nVersion at that location in ::VERSION processing and just keep using nVersion != 0
42016-11-12T00:04:33 <BlueMatt> cfields: alternatively, shouldnt fSuccessfullyConnected be set in ::VERACK, not ::VERSION?
52016-11-12T00:04:58 <cfields> BlueMatt: see the commit message. Sometimes we bail before fully finishing version
62016-11-12T00:05:12 <BlueMatt> cfields: yes, I'm saying move that setting down
72016-11-12T00:05:31 <cfields> BlueMatt: i think the confusion comes from trying to repurpose that var. What it really means is fCanSendToPeer
82016-11-12T00:05:37 <cfields> how about renaming to that?
92016-11-12T00:05:49 <BlueMatt> cfields: I'm asking if its redundant
102016-11-12T00:06:06 <BlueMatt> the confusion is that it seems redundant
112016-11-12T00:07:19 <cfields> BlueMatt: I suppose with the fDisconnect completely worked out, yes, it's redundant
122016-11-12T00:08:24 <cfields> BlueMatt: the issue before was that sometimes we'd bail halfway through processing VERSION, leaving nVersion set, but we wouldn't want to send any new outgoing messages. But because some places didn't check for fDisconnected, outgoing messages got through anyway
132016-11-12T00:08:58 <cfields> I believe that's fixed in that PR. So in that case, yes, they should be redundant. Will fix.
142016-11-12T00:09:12 <BlueMatt> cfields: yes, i see that, but how hard is it to move the pfrom->nVersion = ... thing down 50 lines and do PushWithVersion for everything in between (is there anything)
152016-11-12T00:09:34 <BlueMatt> ok
162016-11-12T00:09:35 <BlueMatt> thanks
172016-11-12T00:10:53 *** cdecker has quit IRC
182016-11-12T00:52:57 *** Chris_Stewart_5 has quit IRC
192016-11-12T01:11:15 *** testnet has quit IRC
202016-11-12T01:11:15 *** wasi has quit IRC
212016-11-12T01:11:46 *** wasi has joined #bitcoin-core-dev
222016-11-12T01:11:55 *** testnet has joined #bitcoin-core-dev
232016-11-12T01:13:47 *** testnet has left #bitcoin-core-dev
242016-11-12T01:25:03 *** windsok has quit IRC
252016-11-12T01:26:30 *** windsok has joined #bitcoin-core-dev
262016-11-12T01:44:47 *** btcdrak has joined #bitcoin-core-dev
272016-11-12T02:00:58 *** fanquake has joined #bitcoin-core-dev
282016-11-12T02:01:06 *** Alopex has quit IRC
292016-11-12T02:02:11 *** Alopex has joined #bitcoin-core-dev
302016-11-12T02:10:21 *** MarcoFalke has quit IRC
312016-11-12T02:30:07 *** Alopex has quit IRC
322016-11-12T02:31:12 *** Alopex has joined #bitcoin-core-dev
332016-11-12T02:40:03 <BlueMatt> cfields: if you want to be extra super duper awesome for me you could even add a skip_bytes method in your CVectorWriter (see https://github.com/bitcoinfibre/bitcoinfibre/commit/70673283326f0ab7b542dfb16da32dd81f70176d) but thats not really a comment, just a note since I have a similar class in FIBRE and it would be useful
342016-11-12T02:43:29 <sipa> BlueMatt: that just increments the write pointer?
352016-11-12T02:43:36 <BlueMatt> essentially, yes
362016-11-12T02:43:45 <sipa> the equivalent of writing zeroes, i guess?
372016-11-12T02:43:50 <BlueMatt> sipa: yes
382016-11-12T02:44:02 <BlueMatt> sipa: though for my use-case i couldnt care less if its 0s or garbage
392016-11-12T02:44:12 <sipa> what is it used for?
402016-11-12T02:44:21 <BlueMatt> though, i guess, sending un-init'd memory over the wire is generally frowned upon
412016-11-12T02:46:26 <BlueMatt> cfields: didnt really read in too much detail, but the concept looks like what i was asking for
422016-11-12T02:49:54 *** fanquake has quit IRC
432016-11-12T02:50:01 <sipa> BlueMatt: why do want to send (don't care) bytes over the wire?
442016-11-12T02:50:20 <BlueMatt> sipa: pad transactions out so that they (often) start on packet/fec-chunk boundaries
452016-11-12T02:50:26 <BlueMatt> you end up with null space
462016-11-12T02:50:43 <BlueMatt> in an earlier implementation it even sent shorter packets to not send the null space, but you still use it for fec-coding
472016-11-12T02:50:57 <BlueMatt> i think the current code sends the 0s just because it was annoying to try to code
482016-11-12T03:00:28 *** Ylbam has quit IRC
492016-11-12T03:02:25 *** ybit_ is now known as ybit
502016-11-12T03:02:41 *** Ylbam has joined #bitcoin-core-dev
512016-11-12T03:10:12 *** Ylbam has quit IRC
522016-11-12T03:52:05 *** amiller has quit IRC
532016-11-12T04:03:07 *** Alopex has quit IRC
542016-11-12T04:03:51 <cfields> BlueMatt: hmm, it may make more sense to do that at the point where we're actually writing to the net?
552016-11-12T04:04:09 <cfields> s/net/socket
562016-11-12T04:04:12 *** Alopex has joined #bitcoin-core-dev
572016-11-12T04:04:42 <cfields> no problem adding it to CVectorWriter though, if that's the approach that makes sense
582016-11-12T04:04:43 *** sanada has joined #bitcoin-core-dev
592016-11-12T04:06:28 <cfields> BlueMatt: it surprises me that you'd see the benefits of padding like that though, considering how many layers of caching there are to get through on the OS side
602016-11-12T04:08:34 <cfields> BlueMatt: and thanks for looking, btw
612016-11-12T04:44:07 *** Alopex has quit IRC
622016-11-12T04:45:12 *** Alopex has joined #bitcoin-core-dev
632016-11-12T04:57:35 *** Guest47022 has joined #bitcoin-core-dev
642016-11-12T04:58:34 <gmaxwell> matt's data neeeds to be padded in memory, not for the net.
652016-11-12T04:59:42 <gmaxwell> he seralizes the block into memory and then runs forward error correction over it. The padding is added so that transactions begin on FEC packet bundaries.
662016-11-12T05:01:07 <gmaxwell> The reason for this is that when using the mempool for reconstruction when a txn is missing the whole packet containing is missing. To prevent miss amplification there is some padding to get txn onto packet boundaries.
672016-11-12T05:01:11 <BlueMatt> cfields: note that if you've gotten to the point where you're sending tx data over the wire, you've failed horribly....at that point you've already sent a ton of fec data
682016-11-12T05:01:25 <BlueMatt> also what gmaxwell said
692016-11-12T05:01:40 <gmaxwell> if matt's FEC stuff was more optimized he'd probably never send the original packets at all ever. :P
702016-11-12T05:01:46 <cfields> ah, i see
712016-11-12T05:02:07 <BlueMatt> i mean i could just do that, but since i already have the data and all the plumbing for using it since the header stuff does...might as well send it :)
722016-11-12T05:02:18 *** Chris_Stewart_5 has joined #bitcoin-core-dev
732016-11-12T05:02:30 *** abpa has joined #bitcoin-core-dev
742016-11-12T05:02:45 <cfields> completely misunderstood what you were getting at, thanks for the explanation
752016-11-12T05:10:27 *** juscamarena has joined #bitcoin-core-dev
762016-11-12T05:13:49 *** juscamarena has quit IRC
772016-11-12T05:24:02 *** Giszmo has quit IRC
782016-11-12T05:24:32 *** Alopex has quit IRC
792016-11-12T05:25:37 *** Alopex has joined #bitcoin-core-dev
802016-11-12T05:36:09 *** Chris_Stewart_5 has quit IRC
812016-11-12T05:38:11 *** Alopex has quit IRC
822016-11-12T05:39:16 *** Alopex has joined #bitcoin-core-dev
832016-11-12T06:01:27 *** Alopex has quit IRC
842016-11-12T06:02:32 *** Alopex has joined #bitcoin-core-dev
852016-11-12T06:09:37 *** Naphex has joined #bitcoin-core-dev
862016-11-12T06:12:21 *** Alopex has quit IRC
872016-11-12T06:13:27 *** Alopex has joined #bitcoin-core-dev
882016-11-12T06:22:07 <luke-jr> I keep getting on master ./bench/data/block413567.raw.h:124989:40: error: redefinition of âconst unsigned char block_bench::block413567 []â
892016-11-12T06:22:15 <luke-jr> have to manually delete the file and try again
902016-11-12T06:23:13 <luke-jr> the rules to generate it only append, shouldn't the first create anew?
912016-11-12T06:23:42 <luke-jr> (and does make do the deletion of files when the rule fails, or must we?)
922016-11-12T06:24:45 <luke-jr> looks like we need to do it..
932016-11-12T06:56:28 *** DigiByteDev has joined #bitcoin-core-dev
942016-11-12T06:58:30 *** thokon00 has quit IRC
952016-11-12T07:03:01 *** d9b4bef9 has quit IRC
962016-11-12T07:03:02 <wumpus> luke-jr: strange, I just copied that rule from the tests, so I'd expect it would work as-is. But indeed it should create the file anew, not append to it, that's weird
972016-11-12T07:03:43 <luke-jr> is there one of these in the tests I should fix too?
982016-11-12T07:03:57 *** thokon00 has joined #bitcoin-core-dev
992016-11-12T07:03:57 *** fengling has joined #bitcoin-core-dev
1002016-11-12T07:04:05 <wumpus> no, I think it's my fault, I removed the namespace{} around it, the first line in the test is probably ok
1012016-11-12T07:04:08 *** d9b4bef9 has joined #bitcoin-core-dev
1022016-11-12T07:04:28 <wumpus> should just change the first >> $@ to > $@
1032016-11-12T07:04:34 <wumpus> are you going to do it or should I?
1042016-11-12T07:04:50 <luke-jr> well, if any of these steps fails, we need to delete the file too :x
1052016-11-12T07:05:05 <luke-jr> (or else make will think it's up to date next run)
1062016-11-12T07:07:27 <wumpus> the best option then would be to write to a temporary file
1072016-11-12T07:07:39 <wumpus> then at the end mv it over atomically
1082016-11-12T07:07:57 <wumpus> eg write to $@.tmp
1092016-11-12T07:08:38 <wumpus> in that case you need to do it in Makefile.test.include too
1102016-11-12T07:08:45 <luke-jr> hm, I was thinking http://codepad.org/SXhNyVzI
1112016-11-12T07:09:21 <wumpus> meh, with my solution you never actually need to delete anything
1122016-11-12T07:09:25 <wumpus> it's just not created if naything fails
1132016-11-12T07:09:41 <luke-jr> true
1142016-11-12T07:10:11 <wumpus> or ... isn't that the case for yours too?
1152016-11-12T07:10:14 <wumpus> you write the output to a pipe
1162016-11-12T07:10:24 <wumpus> oh wait, yes
1172016-11-12T07:10:54 <luke-jr> http://codepad.org/Glnudlpw ?
1182016-11-12T07:10:55 <wumpus> the pipe doesn't 'wait'. SO yes a temporary file then atomic move is probably better
1192016-11-12T07:11:06 <wumpus> yes
1202016-11-12T07:16:16 *** aalex has joined #bitcoin-core-dev
1212016-11-12T07:17:01 <bitcoin-git> [bitcoin] luke-jr opened pull request #9140: Bugfix: Correctly replace generated headers and fail cleanly (master...bugfix_genheaders) https://github.com/bitcoin/bitcoin/pull/9140
1222016-11-12T07:17:21 *** fengling has quit IRC
1232016-11-12T07:26:13 *** DigiByteDev has quit IRC
1242016-11-12T07:35:58 <bitcoin-git> [bitcoin] luke-jr closed pull request #7534: Minimal RPC & wallet support for CLTV-enabled multisig addresses (master...cltv_address) https://github.com/bitcoin/bitcoin/pull/7534
1252016-11-12T07:37:36 <bitcoin-git> [bitcoin] luke-jr closed pull request #8388: [0.13] mining: Optimise for typical mining use with blockmaxsize (0.13...blockmaxsize_opti-0.13) https://github.com/bitcoin/bitcoin/pull/8388
1262016-11-12T07:40:51 *** DigiByteDev has joined #bitcoin-core-dev
1272016-11-12T07:42:22 *** DigiByteDev has quit IRC
1282016-11-12T07:45:22 *** DigiByteDev has joined #bitcoin-core-dev
1292016-11-12T07:51:41 *** aalex has quit IRC
1302016-11-12T08:19:24 *** Eliel has quit IRC
1312016-11-12T08:20:00 *** Eliel has joined #bitcoin-core-dev
1322016-11-12T08:27:48 *** windsok has quit IRC
1332016-11-12T08:29:04 *** Guyver2 has joined #bitcoin-core-dev
1342016-11-12T08:32:22 *** windsok has joined #bitcoin-core-dev
1352016-11-12T08:35:06 <BlueMatt> cfields: ouch, lets avoid any copies? I dont see why we should ever need that...once you PushMessage, PushMessage should just take the message, not the send code
1362016-11-12T08:46:07 *** whphhg_ has joined #bitcoin-core-dev
1372016-11-12T08:46:33 *** whphhg has quit IRC
1382016-11-12T08:46:36 *** whphhg_ is now known as whphhg
1392016-11-12T09:02:04 *** windsok has quit IRC
1402016-11-12T09:06:53 *** Ylbam has joined #bitcoin-core-dev
1412016-11-12T09:10:49 <bitcoin-git> [bitcoin] jonasschnelli opened pull request #9141: Remove unnecessary calls to CheckFinalTx (master...2016/11/istrusted) https://github.com/bitcoin/bitcoin/pull/9141
1422016-11-12T09:26:18 <bitcoin-git> [bitcoin] jonasschnelli opened pull request #9142: Move -salvagewallet, -zap(wtx) to where they belong (master...2016/11/wallet_init) https://github.com/bitcoin/bitcoin/pull/9142
1432016-11-12T09:27:29 *** Guyver2 has quit IRC
1442016-11-12T09:41:44 *** moli has quit IRC
1452016-11-12T09:42:36 *** eljaycee has joined #bitcoin-core-dev
1462016-11-12T09:57:36 <Victorsueca> morning
1472016-11-12T09:59:20 <bitcoin-git> [bitcoin] jonasschnelli opened pull request #9143: Refactor ZapWalletTxes to avoid layer vialotions (master...2016/11/wallet_db_refactoring_1) https://github.com/bitcoin/bitcoin/pull/9143
1482016-11-12T09:59:32 *** windsok has joined #bitcoin-core-dev
1492016-11-12T10:30:26 *** Ylbam has quit IRC
1502016-11-12T10:30:41 *** AtashiCon has quit IRC
1512016-11-12T10:30:45 *** Ylbam has joined #bitcoin-core-dev
1522016-11-12T10:45:21 *** AtashiCon has joined #bitcoin-core-dev
1532016-11-12T10:45:59 *** AaronvanW has quit IRC
1542016-11-12T10:48:28 *** arowser has quit IRC
1552016-11-12T10:48:28 *** xiangfu has quit IRC
1562016-11-12T10:49:17 *** arowser has joined #bitcoin-core-dev
1572016-11-12T10:50:37 *** xiangfu has joined #bitcoin-core-dev
1582016-11-12T10:51:10 *** AaronvanW has joined #bitcoin-core-dev
1592016-11-12T10:51:10 *** AaronvanW has joined #bitcoin-core-dev
1602016-11-12T10:53:43 *** Naphex has quit IRC
1612016-11-12T11:02:27 *** cdecker has joined #bitcoin-core-dev
1622016-11-12T11:03:22 *** fanquake has joined #bitcoin-core-dev
1632016-11-12T11:06:05 *** MarcoFalke has joined #bitcoin-core-dev
1642016-11-12T11:25:13 *** DigiByteDev has quit IRC
1652016-11-12T11:29:08 *** Ylbam has quit IRC
1662016-11-12T11:30:35 *** Ylbam has joined #bitcoin-core-dev
1672016-11-12T11:46:31 *** John_Doe has joined #bitcoin-core-dev
1682016-11-12T12:19:00 *** moli has joined #bitcoin-core-dev
1692016-11-12T12:39:27 <bitcoin-git> [bitcoin] fanquake opened pull request #9144: [Trivial] Correct waitforblockheight example help text (master...rpc-commands) https://github.com/bitcoin/bitcoin/pull/9144
1702016-11-12T12:40:04 *** laurentmt has joined #bitcoin-core-dev
1712016-11-12T12:40:19 <fanquake> Excuse the PRs/issue closing, going to try and clean up a lot of issues.
1722016-11-12T12:42:48 <paveljanik> fanquake, no need to apologise, good work!
1732016-11-12T12:43:53 *** laurentmt has quit IRC
1742016-11-12T13:52:21 <bitcoin-git> [bitcoin] MarcoFalke opened pull request #9145: [qt] Make network disabled icon 50% opaque (master...Mf1611-qtNetworkIcon) https://github.com/bitcoin/bitcoin/pull/9145
1752016-11-12T14:06:51 *** eljaycee has quit IRC
1762016-11-12T14:18:47 <wumpus> fanquake: yes, thanks a lot
1772016-11-12T14:36:32 *** Giszmo has joined #bitcoin-core-dev
1782016-11-12T14:38:45 *** jtimon has joined #bitcoin-core-dev
1792016-11-12T14:42:19 *** John_Doe has left #bitcoin-core-dev
1802016-11-12T14:58:24 *** owowo has quit IRC
1812016-11-12T15:02:23 *** Giszmo has quit IRC
1822016-11-12T15:14:33 *** owowo has joined #bitcoin-core-dev
1832016-11-12T15:14:34 *** owowo has joined #bitcoin-core-dev
1842016-11-12T15:14:34 *** owowo has joined #bitcoin-core-dev
1852016-11-12T15:19:04 *** owowo has quit IRC
1862016-11-12T15:24:00 *** owowo has joined #bitcoin-core-dev
1872016-11-12T15:31:56 *** aalex has joined #bitcoin-core-dev
1882016-11-12T15:38:53 *** aalex has quit IRC
1892016-11-12T15:54:42 *** haya has joined #bitcoin-core-dev
1902016-11-12T15:56:00 *** haya has quit IRC
1912016-11-12T16:50:22 *** cdecker has quit IRC
1922016-11-12T16:50:36 *** cdecker has joined #bitcoin-core-dev
1932016-11-12T16:55:36 *** MarcoFalke has left #bitcoin-core-dev
1942016-11-12T17:20:56 *** nibor has quit IRC
1952016-11-12T17:21:48 *** nibor has joined #bitcoin-core-dev
1962016-11-12T17:31:01 *** d9b4bef9 has quit IRC
1972016-11-12T17:32:08 *** d9b4bef9 has joined #bitcoin-core-dev
1982016-11-12T17:40:04 *** nibor has quit IRC
1992016-11-12T17:41:48 *** nibor has joined #bitcoin-core-dev
2002016-11-12T17:57:09 *** justanotheruser has quit IRC
2012016-11-12T17:58:50 *** justanotheruser has joined #bitcoin-core-dev
2022016-11-12T18:40:37 *** belcher has quit IRC
2032016-11-12T18:51:50 *** belcher has joined #bitcoin-core-dev
2042016-11-12T19:05:12 *** achow101 has quit IRC
2052016-11-12T19:05:32 *** achow101 has joined #bitcoin-core-dev
2062016-11-12T19:38:45 *** AaronvanW has quit IRC
2072016-11-12T19:50:30 *** Guyver2 has joined #bitcoin-core-dev
2082016-11-12T20:50:59 *** Guyver2__ has joined #bitcoin-core-dev
2092016-11-12T20:53:26 *** Guyver2 has quit IRC
2102016-11-12T20:53:32 *** Guyver2__ is now known as Guyver2
2112016-11-12T20:58:13 *** achow101 has quit IRC
2122016-11-12T21:02:01 *** d9b4bef9 has quit IRC
2132016-11-12T21:03:08 *** d9b4bef9 has joined #bitcoin-core-dev
2142016-11-12T21:03:23 *** achow101 has joined #bitcoin-core-dev
2152016-11-12T21:03:40 *** achow101 has quit IRC
2162016-11-12T21:49:39 *** PaulCapestany has quit IRC
2172016-11-12T21:52:28 *** PaulCapestany has joined #bitcoin-core-dev
2182016-11-12T21:58:15 *** PaulCapestany has quit IRC
2192016-11-12T21:59:19 *** PaulCapestany has joined #bitcoin-core-dev
2202016-11-12T22:09:18 *** PaulCapestany has quit IRC
2212016-11-12T22:11:12 *** PaulCapestany has joined #bitcoin-core-dev
2222016-11-12T22:20:53 *** cdecker has quit IRC
2232016-11-12T22:21:27 <BlueMatt> someone wanna tag #9148 0.14?
2242016-11-12T22:21:28 <gribble> https://github.com/bitcoin/bitcoin/issues/9148 | Wallet RPCs can return stale info due to ProcessNewBlock Race · Issue #9148 · bitcoin/bitcoin · GitHub
2252016-11-12T22:23:02 *** PRab has joined #bitcoin-core-dev
2262016-11-12T22:31:28 *** achow101 has joined #bitcoin-core-dev
2272016-11-12T22:41:34 *** justanotheruser is now known as YolandiVisser
2282016-11-12T22:47:57 *** YolandiVisser is now known as justanotheruser
2292016-11-12T23:39:45 *** belcher has quit IRC
2302016-11-12T23:52:09 *** belcher has joined #bitcoin-core-dev
2312016-11-12T23:58:00 *** btcdrak has quit IRC