12017-05-05T00:12:38 *** goksinen has joined #bitcoin-core-dev
22017-05-05T00:12:57 <luke-jr> it seems that my system univalue returns a temporary, and for some reason Core is including that instead of the embedded univalue :|
32017-05-05T00:13:34 <praxeology> Not sure how "garbage collection" works with std::string. Does that line make a string array on the stack or is it allocated in heap?
42017-05-05T00:14:14 *** marcoagner has quit IRC
52017-05-05T00:14:23 <sipa> praxeology: c++ has no garbage collection
62017-05-05T00:14:37 *** marcoagner has joined #bitcoin-core-dev
72017-05-05T00:14:41 <sipa> when the variable goes out of scope, it is destroyed
82017-05-05T00:14:46 <luke-jr> what is supposed to add src/univalue/include/ to the include path? it's not there..
92017-05-05T00:15:25 <sipa> praxeology: and strings are always allocated on the heap (in practice), but that's not relevant
102017-05-05T00:15:40 <sipa> praxeology: there is no refcounting for references
112017-05-05T00:16:30 <sipa> some c++ standard libraries implement refcounting for strings internally, but that still only works with different _string_ objects with the same content, not multiple references to the same one
122017-05-05T00:16:43 <luke-jr> oh, system univalue is used by default, as expected, ok
132017-05-05T00:17:04 <luke-jr> so I guess options are: 1) require newer univalue to use system, or 2) rewrite this code to work either way
142017-05-05T00:17:08 *** goksinen has quit IRC
152017-05-05T00:18:12 <luke-jr> it's probably slightly optimised to do num 2
162017-05-05T00:18:49 <luke-jr> is it valid to assign the return std::string temporary to a reference? or is G++ just extra tolerant?
172017-05-05T00:18:57 <gmaxwell> 'system univalue' dear lord, why would that exist?
182017-05-05T00:18:58 <luke-jr> const std::string& account_param = request.params[0].get_str();
192017-05-05T00:19:15 <luke-jr> gmaxwell: because that's the right way to do libraries
202017-05-05T00:19:16 <sipa> luke-jr: it is valid
212017-05-05T00:19:24 <sipa> (i commented on the PR)
222017-05-05T00:20:34 <praxeology> sipa: yea thought so... "it's taking a pointer to a temporary", not sure why you used the word "temporary" there because if allocated on the heap, its not temporary in any way.
232017-05-05T00:20:35 <luke-jr> which PR?
242017-05-05T00:20:51 <luke-jr> praxeology: it's not. if there's no "new", it's not on the heap
252017-05-05T00:20:57 <gmaxwell> luke-jr: if it were actually a library and not some unmaintained code dump that only we use...
262017-05-05T00:21:24 <praxeology> luke-jr: are you sure there is no "new" deeper in the library?
272017-05-05T00:21:25 <sipa> praxeology: 'heap' and 'stack' don't exist in the C++ spec, they're implementation details
282017-05-05T00:21:34 <sipa> praxeology: 'temporary' however exists in the spec
292017-05-05T00:22:05 <sipa> praxeology: a temporary is the result of an expression that returns something by-value, and isn't assigned to a variable
302017-05-05T00:22:30 <bitcoin-git> [bitcoin] luke-jr opened pull request #10341: rpc/wallet: Workaround older UniValue which returns a std::string temporary for get_str (master...rpcwallet_uv_workaround) https://github.com/bitcoin/bitcoin/pull/10341
312017-05-05T00:23:14 <sipa> praxeology: in practice, the std::string is allocated on the stack, but the string object itself allocates the string/length data on the heap
322017-05-05T00:24:43 <sipa> but again, that doesn't matter
332017-05-05T00:24:56 <sipa> the compiler may even optimize part out and not have such objects hit memory at all
342017-05-05T00:25:15 <sipa> what matters is that it's returning an object and not storing it in a variable
352017-05-05T00:26:16 <praxeology> Hm, yea I guess somehow with the automatic out-of scope clearing/freeing/deleting of the actual data... you need to assign the std::string to a stack variable for the duration you want it to exist
362017-05-05T00:26:38 *** tw2006 has joined #bitcoin-core-dev
372017-05-05T00:26:53 <sipa> no, you don't
382017-05-05T00:26:57 <sipa> that's what i just pointed out
392017-05-05T00:27:25 <sipa> you can take a reference to a temporary, and in that case, the lifetime of the temporary is automatically extended as long as the reference exists
402017-05-05T00:27:46 <sipa> but that's done by static analysis by the compiler, not by reference counting
412017-05-05T00:28:16 <praxeology> its returning an object that becomes part of stack, but in order for that to happen you need to assign the object to a variable that exists in the stack
422017-05-05T00:28:34 <sipa> stop talking about stacks, that doesn't matter
432017-05-05T00:30:00 <praxeology> ok, scope somehow. but that is less clear to me how such works... I guess we don't have to discuss it here, I can look into it myself
442017-05-05T00:31:11 *** tw2006 has quit IRC
452017-05-05T00:36:56 <praxeology> const std::string* account = nullptr;
462017-05-05T00:37:38 <praxeology> woops sorry I'll leave you guys alone now
472017-05-05T00:37:42 * luke-jr puts praxeology on the stack
482017-05-05T00:40:27 *** owowo has quit IRC
492017-05-05T00:46:09 *** owowo has joined #bitcoin-core-dev
502017-05-05T00:48:20 *** Ylbam has quit IRC
512017-05-05T00:56:05 *** tw2006 has joined #bitcoin-core-dev
522017-05-05T01:02:52 *** jtimon has quit IRC
532017-05-05T01:34:03 *** dcousens has joined #bitcoin-core-dev
542017-05-05T01:34:39 <dcousens> sipa: "its modifying rng_state, protected by the lock" --- I suppose, for an RNG, racing memcpy of random data isn't as much an issue? :D haha
552017-05-05T01:36:37 <gmaxwell> it's very much an issue, as it could cause arbritary corruption.
562017-05-05T01:36:54 <dcousens> gmaxwell: would it could corruption outside of rng_state though?
572017-05-05T01:36:58 <gmaxwell> Yes.
582017-05-05T01:37:04 <dcousens> OK, then yes, issue
592017-05-05T01:37:20 <gmaxwell> yea, I was just being stupid and thought that was the copy into the output too.
602017-05-05T01:39:00 <dcousens> gmaxwell: ooi, what could I lookup to understand the implications of that, I wasn't aware it could cause arbitrary corruption
612017-05-05T01:39:02 <gmaxwell> dcousens: in C/C++ the compiler is allowed to assume that it has exclusive access to any variable it writes to in a scope for the duration of it (the technical guarentee is more complicated and I don't fully understand it, but this approximation is enough). So the compiler could do optimizations like temporarily store the stack pointer into memory that it knows it's going to later overwrite. t
622017-05-05T01:39:08 <gmaxwell> hen copy it back out again before overwriting it.
632017-05-05T01:39:09 <dcousens> gmaxwell: answered as I asked
642017-05-05T01:40:03 <dcousens> cheers
652017-05-05T01:40:11 <gmaxwell> Its very rare for it to _in practice_ cause arbritary corruption, but it's not unheared of.. intel's compiler is a little more prone to crazy stunts that expose things like that than GCC is, but presumably future versions of GCC will keep getting smarter until nothing runs. :)
662017-05-05T01:40:20 *** d_t has joined #bitcoin-core-dev
672017-05-05T01:40:27 <dcousens> ha
682017-05-05T02:00:10 *** dermoth has quit IRC
692017-05-05T02:01:05 *** dermoth has joined #bitcoin-core-dev
702017-05-05T02:03:30 *** laurentmt has quit IRC
712017-05-05T02:13:24 *** goksinen has joined #bitcoin-core-dev
722017-05-05T02:17:08 *** d_t has quit IRC
732017-05-05T02:17:54 *** goksinen has quit IRC
742017-05-05T03:01:16 *** atroxes has quit IRC
752017-05-05T03:02:35 *** atroxes has joined #bitcoin-core-dev
762017-05-05T03:02:39 *** talmai has joined #bitcoin-core-dev
772017-05-05T03:12:28 *** tw2006 has quit IRC
782017-05-05T03:51:01 *** d9b4bef9 has quit IRC
792017-05-05T03:52:19 *** d9b4bef9 has joined #bitcoin-core-dev
802017-05-05T04:13:33 *** tw2006 has joined #bitcoin-core-dev
812017-05-05T04:14:09 *** goksinen has joined #bitcoin-core-dev
822017-05-05T04:17:56 *** tw2006 has quit IRC
832017-05-05T04:18:27 *** goksinen has quit IRC
842017-05-05T04:27:18 *** sw1f7 has joined #bitcoin-core-dev
852017-05-05T04:41:04 *** goksinen has joined #bitcoin-core-dev
862017-05-05T04:50:26 *** goksinen has quit IRC
872017-05-05T04:53:34 *** goksinen has joined #bitcoin-core-dev
882017-05-05T04:58:32 *** goksinen has quit IRC
892017-05-05T05:03:13 *** goksinen has joined #bitcoin-core-dev
902017-05-05T05:08:40 *** goksinen has quit IRC
912017-05-05T05:14:20 *** tw2006 has joined #bitcoin-core-dev
922017-05-05T05:17:01 *** kadoban has quit IRC
932017-05-05T05:19:12 *** tw2006 has quit IRC
942017-05-05T05:37:05 *** iprokg has joined #bitcoin-core-dev
952017-05-05T05:38:43 *** Ylbam has joined #bitcoin-core-dev
962017-05-05T05:44:55 *** iprokg has quit IRC
972017-05-05T05:46:24 *** RubenSomsen has joined #bitcoin-core-dev
982017-05-05T05:49:24 *** talmai has quit IRC
992017-05-05T06:00:07 *** dermoth has quit IRC
1002017-05-05T06:00:43 <jonasschnelli> <*highlight> [22:47:52] <BlueMatt:#bitcoin-core-dev> jonasschnelli: you might kill me for #10340, sorry :/
1012017-05-05T06:00:44 <gribble> https://github.com/bitcoin/bitcoin/issues/10340 | Add harmless missing cs_wallet lock in qt CoinControlDialog by TheBlueMatt · Pull Request #10340 · bitcoin/bitcoin · GitHub
1022017-05-05T06:00:52 *** dermoth has joined #bitcoin-core-dev
1032017-05-05T06:00:57 * jonasschnelli looking at 10340
1042017-05-05T06:02:02 <jonasschnelli> BlueMatt: seems reasonable to me (10340)
1052017-05-05T06:03:20 *** moli_ has quit IRC
1062017-05-05T06:03:44 *** moli_ has joined #bitcoin-core-dev
1072017-05-05T06:07:56 *** moli_ has quit IRC
1082017-05-05T06:08:19 *** moli_ has joined #bitcoin-core-dev
1092017-05-05T06:13:05 *** moli_ has quit IRC
1102017-05-05T06:13:43 *** moli_ has joined #bitcoin-core-dev
1112017-05-05T06:15:14 *** tw2006 has joined #bitcoin-core-dev
1122017-05-05T06:20:01 *** tw2006 has quit IRC
1132017-05-05T06:43:44 *** d_t has joined #bitcoin-core-dev
1142017-05-05T06:52:57 *** BashCo has quit IRC
1152017-05-05T07:01:20 *** To7 has quit IRC
1162017-05-05T07:01:48 *** RubenSomsen has quit IRC
1172017-05-05T07:05:17 *** goksinen has joined #bitcoin-core-dev
1182017-05-05T07:09:27 *** goksinen has quit IRC
1192017-05-05T07:11:43 *** d_t has quit IRC
1202017-05-05T07:11:58 *** tonikt has joined #bitcoin-core-dev
1212017-05-05T07:16:09 *** tw2006 has joined #bitcoin-core-dev
1222017-05-05T07:19:17 *** RubenSomsen has joined #bitcoin-core-dev
1232017-05-05T07:20:37 *** tw2006 has quit IRC
1242017-05-05T07:27:53 *** Dyaheon has quit IRC
1252017-05-05T07:30:26 *** Dyaheon has joined #bitcoin-core-dev
1262017-05-05T07:31:45 *** tonikt has quit IRC
1272017-05-05T07:40:09 *** BashCo has joined #bitcoin-core-dev
1282017-05-05T07:49:04 <jonasschnelli> BlueMatt: Ah.. know I now why you mentioned " you might kill me". Massive layer violation.
1292017-05-05T08:08:39 *** _anthony_ has joined #bitcoin-core-dev
1302017-05-05T08:08:45 *** anthonyjd has quit IRC
1312017-05-05T08:17:05 *** tw2006 has joined #bitcoin-core-dev
1322017-05-05T08:22:04 *** tw2006 has quit IRC
1332017-05-05T08:23:22 *** jannes has joined #bitcoin-core-dev
1342017-05-05T08:26:16 *** gielbier has joined #bitcoin-core-dev
1352017-05-05T08:30:03 *** Giszmo has quit IRC
1362017-05-05T08:30:04 *** gielbier has quit IRC
1372017-05-05T08:50:45 *** vicenteH has joined #bitcoin-core-dev
1382017-05-05T08:58:11 *** Ylbam has quit IRC
1392017-05-05T09:05:58 *** goksinen has joined #bitcoin-core-dev
1402017-05-05T09:11:10 *** goksinen has quit IRC
1412017-05-05T09:18:03 *** tw2006 has joined #bitcoin-core-dev
1422017-05-05T09:22:25 *** tw2006 has quit IRC
1432017-05-05T09:27:38 *** Guyver2 has joined #bitcoin-core-dev
1442017-05-05T09:37:44 *** marcoagn1 has joined #bitcoin-core-dev
1452017-05-05T09:37:48 *** marcoagner has quit IRC
1462017-05-05T10:19:12 *** tw2006 has joined #bitcoin-core-dev
1472017-05-05T10:23:33 *** tw2006 has quit IRC
1482017-05-05T10:31:51 *** praxeology has left #bitcoin-core-dev
1492017-05-05T10:36:24 *** tw2006 has joined #bitcoin-core-dev
1502017-05-05T10:53:21 *** jtimon has joined #bitcoin-core-dev
1512017-05-05T11:06:42 *** goksinen has joined #bitcoin-core-dev
1522017-05-05T11:11:07 *** goksinen has quit IRC
1532017-05-05T11:54:20 *** Ylbam has joined #bitcoin-core-dev
1542017-05-05T12:44:34 *** NewLiberty has joined #bitcoin-core-dev
1552017-05-05T12:56:01 *** [\\\] has joined #bitcoin-core-dev
1562017-05-05T12:57:02 *** vicenteH` has joined #bitcoin-core-dev
1572017-05-05T12:57:05 *** mol has joined #bitcoin-core-dev
1582017-05-05T13:01:38 *** atroxes_ has joined #bitcoin-core-dev
1592017-05-05T13:01:49 *** Dyaheon- has joined #bitcoin-core-dev
1602017-05-05T13:02:51 *** Ylbam has quit IRC
1612017-05-05T13:02:52 *** marcoagn1 has quit IRC
1622017-05-05T13:02:52 *** vicenteH has quit IRC
1632017-05-05T13:02:52 *** Dyaheon has quit IRC
1642017-05-05T13:02:52 *** moli_ has quit IRC
1652017-05-05T13:02:52 *** atroxes has quit IRC
1662017-05-05T13:02:52 *** tripleslash has quit IRC
1672017-05-05T13:02:53 *** mturquette has quit IRC
1682017-05-05T13:02:54 *** atroxes_ is now known as atroxes
1692017-05-05T13:07:28 *** goksinen has joined #bitcoin-core-dev
1702017-05-05T13:09:27 *** marcoagn1 has joined #bitcoin-core-dev
1712017-05-05T13:11:57 <bitcoin-git> [bitcoin] jtimon closed pull request #10119: Util: Remove ArgsManager wrappers: (master...0.14-args-wrappers) https://github.com/bitcoin/bitcoin/pull/10119
1722017-05-05T13:12:13 *** goksinen has quit IRC
1732017-05-05T13:14:28 *** Ylbam has joined #bitcoin-core-dev
1742017-05-05T13:14:28 *** mturquette has joined #bitcoin-core-dev
1752017-05-05T13:25:23 *** AaronvanW has joined #bitcoin-core-dev
1762017-05-05T13:26:28 <jonasschnelli> Collecting stats over CScheduler with a interval of 2000ms (TBD), would it still make sense to collect a timestamp per sample of is CScheduler in general precise enough so we can calculate (index-of-sample * 2000ms) == timeoffset?
1772017-05-05T13:26:31 *** Ylbam has quit IRC
1782017-05-05T13:26:31 *** mturquette has quit IRC
1792017-05-05T13:26:57 *** Ylbam has joined #bitcoin-core-dev
1802017-05-05T13:27:48 *** mturquette has joined #bitcoin-core-dev
1812017-05-05T13:28:34 *** laurentmt has joined #bitcoin-core-dev
1822017-05-05T13:29:02 *** goksinen has joined #bitcoin-core-dev
1832017-05-05T13:31:42 *** laurentmt has quit IRC
1842017-05-05T13:31:54 <rafalcpp> dcousens: with any UB in c++, anything can happen. In theory a race condition in some debug code, could make your node start mining dogecoin instead
1852017-05-05T13:32:49 <dcousens> rafalcpp: haha, I suppose I was simply affirming it was U/B, not just a product of something else I'd missed
1862017-05-05T13:33:22 *** goksinen has quit IRC
1872017-05-05T13:35:38 <bitcoin-git> [bitcoin] jnewbery opened pull request #10342: [tests] Improve mempool_persist test (master...improve_wait_until_test) https://github.com/bitcoin/bitcoin/pull/10342
1882017-05-05T13:36:14 <Chris_Stewart_5> Is there a way to have two local regtest node share the same chain, while having different wallets?
1892017-05-05T13:40:00 <instagibbs> Chris_Stewart_5, this is #bitcoin question, but yes
1902017-05-05T13:46:39 *** talmai has joined #bitcoin-core-dev
1912017-05-05T13:50:34 *** d_t has joined #bitcoin-core-dev
1922017-05-05T13:55:04 *** d_t has quit IRC
1932017-05-05T13:57:19 *** laurentmt has joined #bitcoin-core-dev
1942017-05-05T13:58:28 *** laurentmt has quit IRC
1952017-05-05T14:03:07 *** talmai has quit IRC
1962017-05-05T14:08:10 *** JackH has quit IRC
1972017-05-05T14:12:13 *** zrts_ has left #bitcoin-core-dev
1982017-05-05T14:19:23 *** RubenSomsen has quit IRC
1992017-05-05T14:25:26 <BlueMatt> jonasschnelli: ehh, I didnt feel like making needless copies of everything 3 times
2002017-05-05T14:25:36 <BlueMatt> but i can do that if you prefer
2012017-05-05T14:26:39 <jonasschnelli> BlueMatt: I don't like accessing the wallet's insides from outside of walletmodel. We can make an exception,... but I guess ryanofsky has to do more work in case we want process sep.
2022017-05-05T14:27:06 <BlueMatt> well we can merge ryanofsky's pr first and then maybe extend it even further and never expose COutput outside of wallet
2032017-05-05T14:27:12 <BlueMatt> ultimately we should move towards ^
2042017-05-05T14:27:32 <jonasschnelli> Yes. Thats a good point. For intermediate steps, I think it's fine.
2052017-05-05T14:28:48 <BlueMatt> I mean I'm more than happy to not take 10340 yet and do the real fix, as long as we get it in for 0.15
2062017-05-05T14:28:52 <BlueMatt> it shouldnt be /too/ hard
2072017-05-05T14:29:02 <ryanofsky> my pr that stops using coutput/cwallettx (#10244) is ready, just needs review
2082017-05-05T14:29:04 <gribble> https://github.com/bitcoin/bitcoin/issues/10244 | [qt] Add abstraction layer for accessing node and wallet functionality from gui by ryanofsky · Pull Request #10244 · bitcoin/bitcoin · GitHub
2092017-05-05T14:33:35 *** Alina-malina has quit IRC
2102017-05-05T14:35:10 *** QBcrusher__ has quit IRC
2112017-05-05T14:37:34 *** JackH has joined #bitcoin-core-dev
2122017-05-05T14:37:52 *** goksinen has joined #bitcoin-core-dev
2132017-05-05T14:45:04 <BlueMatt> oh, heh, missed that
2142017-05-05T14:45:06 <BlueMatt> k, will review
2152017-05-05T14:46:46 <bitcoin-git> [bitcoin] TheBlueMatt closed pull request #10340: Add harmless missing cs_wallet lock in qt CoinControlDialog (master...2017-05-fix-mapwallet-zap-runtime) https://github.com/bitcoin/bitcoin/pull/10340
2162017-05-05T14:47:37 *** jnewbery has quit IRC
2172017-05-05T14:48:26 *** jnewbery has joined #bitcoin-core-dev
2182017-05-05T14:51:43 *** Alina-malina has joined #bitcoin-core-dev
2192017-05-05T14:52:55 *** tonikt has joined #bitcoin-core-dev
2202017-05-05T14:59:09 *** Alina-malina has quit IRC
2212017-05-05T14:59:10 *** Alina-malina has joined #bitcoin-core-dev
2222017-05-05T15:02:51 *** tw2006 has quit IRC
2232017-05-05T15:05:44 *** Dyaheon- has quit IRC
2242017-05-05T15:06:32 *** Dyaheon has joined #bitcoin-core-dev
2252017-05-05T15:09:51 *** tonikt has quit IRC
2262017-05-05T15:10:19 *** RubenSomsen has joined #bitcoin-core-dev
2272017-05-05T15:18:23 *** RubenSomsen has quit IRC
2282017-05-05T15:33:11 *** Giszmo has joined #bitcoin-core-dev
2292017-05-05T15:35:45 *** harrymm1 has joined #bitcoin-core-dev
2302017-05-05T15:36:24 *** harrymm1 has joined #bitcoin-core-dev
2312017-05-05T15:38:27 *** harrymm has quit IRC
2322017-05-05T15:44:31 *** Giszmo1 has joined #bitcoin-core-dev
2332017-05-05T15:47:05 *** abpa has joined #bitcoin-core-dev
2342017-05-05T15:59:00 <bitcoin-git> [bitcoin] practicalswift opened pull request #10343: Remove redundant on-the-same-line-repetition of type names (DRY): RepeatedTypeName foo = static_cast<RepeatedTypeName>(bar) (master...auto) https://github.com/bitcoin/bitcoin/pull/10343
2352017-05-05T16:02:12 *** tw2006 has joined #bitcoin-core-dev
2362017-05-05T16:12:22 *** marcoagn1 has quit IRC
2372017-05-05T16:17:12 *** SopaXorzTaker has joined #bitcoin-core-dev
2382017-05-05T16:27:05 *** elkalamar has quit IRC
2392017-05-05T16:28:58 *** dcousens has quit IRC
2402017-05-05T16:48:20 *** n1ce has quit IRC
2412017-05-05T16:53:33 *** [\\\] is now known as tripleslash
2422017-05-05T17:06:42 <bitcoin-git> [bitcoin] jnewbery opened pull request #10344: [tests] Fix abandonconflict.py intermittency (master...fix_abandon_conflict) https://github.com/bitcoin/bitcoin/pull/10344
2432017-05-05T17:30:50 *** moli_ has joined #bitcoin-core-dev
2442017-05-05T17:32:22 <BlueMatt> sipa: really? per-txout ends with a smaller dbcache? that's (marginally) surprising to me
2452017-05-05T17:32:40 *** goksinen has quit IRC
2462017-05-05T17:33:21 <sipa> BlueMatt: likewise
2472017-05-05T17:33:23 *** vicenteH` has quit IRC
2482017-05-05T17:33:53 *** mol has quit IRC
2492017-05-05T17:34:00 <sipa> BlueMatt: in my memory the master dbcache usage for a full reindex was around 4GB or so, but apparently i haven't done one in a very long time...
2502017-05-05T17:34:17 <BlueMatt> heh
2512017-05-05T17:34:52 * BlueMatt would be curious to see the results of default-dbcache-on-tmpfs
2522017-05-05T17:39:22 <BlueMatt> jtimon: do you feel so inclined to redo #9494 with scripted-diff?
2532017-05-05T17:39:24 <gribble> https://github.com/bitcoin/bitcoin/issues/9494 | Introduce an ArgsManager class encapsulating cs_args, mapArgs and mapMultiArgs by jtimon · Pull Request #9494 · bitcoin/bitcoin · GitHub
2542017-05-05T17:39:33 <BlueMatt> it would be kinda nice, though I'd understand if you dont want to redo it at this point
2552017-05-05T17:51:04 <jtimon> BlueMatt: yeah, for some parts of it it would make a lot of sense. I think I may have broke something in the last edit from what travis said. I will fix that and will s/argsGlobal/gArgs/ and see if I can move some parts to scripte-diff without much effort
2562017-05-05T17:52:14 <BlueMatt> sounds good
2572017-05-05T17:52:16 <BlueMatt> cool
2582017-05-05T17:53:40 <jtimon> BlueMatt: in the meantime #8855 remains very easy to review :p
2592017-05-05T17:53:42 <gribble> https://github.com/bitcoin/bitcoin/issues/8855 | Use a proper factory for creating chainparams by jtimon · Pull Request #8855 · bitcoin/bitcoin · GitHub
2602017-05-05T18:05:08 *** alpalp has joined #bitcoin-core-dev
2612017-05-05T18:10:24 *** praxeology has joined #bitcoin-core-dev
2622017-05-05T18:20:52 *** goksinen has joined #bitcoin-core-dev
2632017-05-05T18:33:24 *** AaronvanW has quit IRC
2642017-05-05T18:37:44 <BlueMatt> heh, sipa, I was about to PR https://github.com/bitcoin/bitcoin/commit/64d45d0fc3b6f7669d0f2c24bca3beaa37cec0b3
2652017-05-05T18:37:54 <BlueMatt> (and https://github.com/bitcoin/bitcoin/commit/5e03232a5dcc4674c620187f67163dc462fbafd1)
2662017-05-05T18:39:17 <sipa> BlueMatt: ha
2672017-05-05T18:39:25 <sipa> RandAddSeedSleepSanity has no return statement
2682017-05-05T18:39:42 <luke-jr> sounds very random.
2692017-05-05T18:39:44 <BlueMatt> oh, well, i never said i had gotten a chance to test it yet
2702017-05-05T18:40:55 <BlueMatt> sipa: it seems a waste to call GetPerformanceCounter and *not* go ahead and add it to our random state
2712017-05-05T18:41:51 <sipa> BlueMatt: how about we add that after #10338 ?
2722017-05-05T18:41:52 <gribble> https://github.com/bitcoin/bitcoin/issues/10338 | Maintain state across GetStrongRandBytes calls by sipa · Pull Request #10338 · bitcoin/bitcoin · GitHub
2732017-05-05T18:42:13 <BlueMatt> sipa: initially it was built on both, but right now everything is pushed into openssl's prng state, so...meh?
2742017-05-05T18:44:48 *** mkwia has left #bitcoin-core-dev
2752017-05-05T18:45:26 *** elkalamar has joined #bitcoin-core-dev
2762017-05-05T18:46:04 <sipa> BlueMatt: like this? https://github.com/bitcoin/bitcoin/pull/10322/commits/999021ea0c6fce6dcef1ff05514c983e60ddc7eb
2772017-05-05T18:47:02 <BlueMatt> sipa: sure! RandAddSeed() also memory_cleanse()s, but I dont think thats actually useful
2782017-05-05T18:48:26 *** alpalp has quit IRC
2792017-05-05T18:50:18 *** alpalp has joined #bitcoin-core-dev
2802017-05-05T19:00:13 *** PaulCapestany has joined #bitcoin-core-dev
2812017-05-05T19:04:58 *** SopaXorzTaker has quit IRC
2822017-05-05T19:05:30 *** alpalp has quit IRC
2832017-05-05T19:15:40 <jonasschnelli> hmm... CScheduler seems to be innacurate. I've set up scheduler.scheduleEvery(fn,2000) (every 2000ms)... I often get a callback with a delta of 10 seconds between each other...
2842017-05-05T19:15:50 <BlueMatt> huh?
2852017-05-05T19:15:52 <BlueMatt> thats...not good
2862017-05-05T19:16:43 <jonasschnelli> I need to look into it
2872017-05-05T19:17:04 <BlueMatt> jonasschnelli: which version? cfields fixed something with sub-second precision not too long ago
2882017-05-05T19:17:10 <BlueMatt> but i believe the symptom was high cpu usage
2892017-05-05T19:17:41 <jonasschnelli> Hmm.. my Laptop runs pretty quite... i'll investigate
2902017-05-05T19:18:08 <jonasschnelli> running in LLDB shoudln't cause such high offsets (i'm not halting the process)¨
2912017-05-05T19:20:05 <jonasschnelli> BlueMatt: What if another Scheduler task requires a couple of seconds to complete? (not familiar with the internals of the Scheduler)
2922017-05-05T19:20:30 <BlueMatt> sure, that could cause it, but the only thing in scheduler is wallet flush, really
2932017-05-05T19:21:21 <jonasschnelli> hmm... I'm not sure if that is the right approach for collecting stats... well,.. at least I have to add a timestamp then to each sample
2942017-05-05T19:21:43 <BlueMatt> hmm?
2952017-05-05T19:22:07 <sipa> jonasschnelli: you'll have a timestamp on each sample regardless
2962017-05-05T19:22:34 <jonasschnelli> sipa: You mean I should have one per sample?
2972017-05-05T19:22:37 <sipa> yes
2982017-05-05T19:22:49 <sipa> you can't guarantee that you'll always be able to gather any statistic at any given point in time at that accuracy
2992017-05-05T19:22:52 * BlueMatt has no idea wtf is going on
3002017-05-05T19:22:54 <jonasschnelli> Yes. I just removed it when I switch to useing CScheduler.. but I guess your right
3012017-05-05T19:23:02 <sipa> BlueMatt: jonas is gather statistics
3022017-05-05T19:23:07 <BlueMatt> stats for what?
3032017-05-05T19:23:07 <sipa> every 2 seconds
3042017-05-05T19:23:22 <jonasschnelli> BlueMatt: my stats collector PR (== mempool graph)
3052017-05-05T19:23:24 <sipa> i assume things like memory usage, bandwidth, ...
3062017-05-05T19:23:28 <BlueMatt> ohoh, that one
3072017-05-05T19:23:35 <jonasschnelli> For now it does only collect mempool data..
3082017-05-05T19:23:47 <jonasschnelli> but it would be trivial to expand it for bandwidth, etc.
3092017-05-05T19:24:26 <jonasschnelli> sipa: you think it's worth keeping a int64 for the starttime and only uint32 deltas per sample?
3102017-05-05T19:24:29 <BlueMatt> well if people want to start seriously using the scheduler we can give it another thread
3112017-05-05T19:25:06 <sipa> jonasschnelli: i assume your resolution is never less than a second
3122017-05-05T19:25:16 <sipa> jonasschnelli: in that case you can use int32 for the starttime and int16 for the deltas :)
3132017-05-05T19:26:36 <jonasschnelli> sipa: I have now different precision levels and I wanted to keep it absolute per sample... but,... I can't remember why I have damned the idea of deltaes between samples..
3142017-05-05T19:27:13 <jonasschnelli> Yes. I guess this would work now that I run over the Schedular (rather then collect "on event")
3152017-05-05T19:27:47 <sipa> using absolute time and data is certainly easier... you can just delete entries to reduce resolution over time
3162017-05-05T19:27:56 <sipa> but uses more memory
3172017-05-05T19:27:56 <jonasschnelli> But still strange that sometimes the Scheduler fires the 2000ms interval after 10000ms (its an inaccuracy of x5)
3182017-05-05T19:28:17 <sipa> that does not sound normal and should be investigated regardless
3192017-05-05T19:28:51 <sipa> you're sure this is not due to lock contention on main or mempool, right?
3202017-05-05T19:29:21 <jonasschnelli> Some sample output of a 2000ms scheduler:
3212017-05-05T19:29:22 <jonasschnelli> https://0bin.net/paste/HXqOTrnFkRzwECHA#8D9wGGWTFJwNZ1vkcQqarVAawSDdZZtENrMBPKeQcYu
3222017-05-05T19:29:23 <gribble> https://github.com/bitcoin/bitcoin/issues/8 | RPC command to sign text with wallet private key · Issue #8 · bitcoin/bitcoin · GitHub
3232017-05-05T19:29:37 <jonasschnelli> No,... it does only lock a new cs_stats
3242017-05-05T19:29:57 <jonasschnelli> The mempool obtain it's own atomic caches in my current working branch
3252017-05-05T19:31:01 <jonasschnelli> Although I run with the GUI... but shoudln't make any different. It's a calm regtest peer anyways
3262017-05-05T19:38:43 *** carbonneau has joined #bitcoin-core-dev
3272017-05-05T19:46:31 *** vicenteH has joined #bitcoin-core-dev
3282017-05-05T19:53:16 *** carbonneau has quit IRC
3292017-05-05T19:55:23 <bitcoin-git> [bitcoin] sdaftuar opened pull request #10345: [P2P] Timeout for headers sync (master...2017-05-timeout-headers-sync) https://github.com/bitcoin/bitcoin/pull/10345
3302017-05-05T20:00:26 *** CubicEarth has joined #bitcoin-core-dev
3312017-05-05T20:03:05 <paveljanik> BTW mempool - "standard" nodes are now running with ~300MB mempools...
3322017-05-05T20:03:50 <gmaxwell> paveljanik: yep, "mempoolminfee": 0.00002822
3332017-05-05T20:04:06 <gmaxwell> hurray, system is back to a healthy state.
3342017-05-05T20:04:27 <paveljanik> not so often tested/run code parts are now being tested :-)
3352017-05-05T20:04:28 <BlueMatt> gmaxwell: +1
3362017-05-05T20:04:30 <instagibbs> what was happening before?
3372017-05-05T20:04:37 <instagibbs> I missed all the excitement it seems
3382017-05-05T20:04:58 <gmaxwell> instagibbs: nah just for the last couple months we were riding at the minimum mempoolminfee... everything was getting confirmed.
3392017-05-05T20:05:37 <instagibbs> RIP sub-sat/byte confirms
3402017-05-05T20:08:48 *** goksinen has quit IRC
3412017-05-05T20:13:26 * paveljanik going to test dump/load mempool
3422017-05-05T20:24:15 *** goksinen has joined #bitcoin-core-dev
3432017-05-05T20:28:57 *** goksinen has quit IRC
3442017-05-05T20:29:47 *** goksinen has joined #bitcoin-core-dev
3452017-05-05T20:42:49 *** Guyver2 has quit IRC
3462017-05-05T20:46:25 *** tw2006 has quit IRC
3472017-05-05T21:06:41 <bitcoin-git> [bitcoin] practicalswift opened pull request #10346: Use range-based for loops (C++11) when looping over vector elements (master...range-based-for-loops) https://github.com/bitcoin/bitcoin/pull/10346
3482017-05-05T21:22:12 <bitcoin-git> [bitcoin] practicalswift closed pull request #10346: Use range-based for loops (C++11) when looping over vector elements (master...range-based-for-loops) https://github.com/bitcoin/bitcoin/pull/10346
3492017-05-05T21:32:14 *** kadoban has joined #bitcoin-core-dev
3502017-05-05T21:37:31 <bitcoin-git> [bitcoin] practicalswift opened pull request #10347: Use range-based for loops (C++11) when looping over vector elements (master...range-based-for-loops) https://github.com/bitcoin/bitcoin/pull/10347
3512017-05-05T21:38:54 *** jannes has quit IRC
3522017-05-05T21:47:16 *** tw2006 has joined #bitcoin-core-dev
3532017-05-05T21:51:46 *** tw2006 has quit IRC
3542017-05-05T22:04:03 <phantomcircuit> gmaxwell, odd
3552017-05-05T22:04:15 <phantomcircuit> my bitcoind running patches to accept anything
3562017-05-05T22:04:19 <phantomcircuit> is still "mempoolminfee": 0.00000000
3572017-05-05T22:04:43 <sipa> "usage": 294825248,
3582017-05-05T22:04:46 <sipa> "mempoolminfee": 0.00001801
3592017-05-05T22:05:02 <phantomcircuit> "usage": 298733328,
3602017-05-05T22:05:07 <phantomcircuit> oh
3612017-05-05T22:05:16 <phantomcircuit> it's because i have maxmempool set to like infinity also
3622017-05-05T22:05:17 <phantomcircuit> derp
3632017-05-05T22:06:26 <luke-jr> lol
3642017-05-05T22:07:34 <phantomcircuit> so it's just on the edge
3652017-05-05T22:21:10 <BlueMatt> phantomcircuit: yea, well because most of your peers are limiting to 3MB its rare that your mempool ever gets much above 3mb
3662017-05-05T22:21:16 <BlueMatt> whether you limit or not
3672017-05-05T22:21:24 <sipa> 3MB ??
3682017-05-05T22:21:32 <sipa> 300?
3692017-05-05T22:21:53 <BlueMatt> oh, yes, 300
3702017-05-05T22:22:06 *** goksinen has quit IRC
3712017-05-05T22:22:32 <phantomcircuit> BlueMatt, i have peers that are before any kind of mempool limiting
3722017-05-05T22:22:34 <phantomcircuit> but yeah
3732017-05-05T22:22:50 <BlueMatt> well their peers :p
3742017-05-05T22:23:04 <BlueMatt> even my 200-connection bitcoind seednode has only 307MB of txn
3752017-05-05T22:27:04 *** goksinen has joined #bitcoin-core-dev
3762017-05-05T22:27:05 *** bsm1175321 has quit IRC
3772017-05-05T22:30:07 *** str4d has joined #bitcoin-core-dev
3782017-05-05T22:31:57 *** goksinen has quit IRC
3792017-05-05T22:45:22 *** goksinen has joined #bitcoin-core-dev
3802017-05-05T22:47:46 *** elkalamar has quit IRC
3812017-05-05T22:48:18 *** tw2006 has joined #bitcoin-core-dev
3822017-05-05T22:52:59 *** tw2006 has quit IRC
3832017-05-05T22:53:35 *** goksinen has quit IRC
3842017-05-05T23:10:21 *** goksinen has joined #bitcoin-core-dev
3852017-05-05T23:13:29 *** PaulCapestany has quit IRC
3862017-05-05T23:14:57 *** goksinen has quit IRC
3872017-05-05T23:16:18 *** PaulCapestany has joined #bitcoin-core-dev
3882017-05-05T23:22:44 *** PaulCapestany has quit IRC
3892017-05-05T23:22:57 *** PaulCapestany has joined #bitcoin-core-dev
3902017-05-05T23:30:35 *** moli_ has quit IRC
3912017-05-05T23:30:59 *** moli_ has joined #bitcoin-core-dev
3922017-05-05T23:41:20 *** tunafizz has joined #bitcoin-core-dev
3932017-05-05T23:49:17 *** tw2006 has joined #bitcoin-core-dev
3942017-05-05T23:53:39 *** tw2006 has quit IRC
3952017-05-05T23:53:55 *** goksinen has joined #bitcoin-core-dev
3962017-05-05T23:59:17 *** goksinen has quit IRC