< bitcoin-git> [bitcoin] fanquake opened pull request #17567: gui: remove macOS start on login code (master...remove_macos_startup_item) https://github.com/bitcoin/bitcoin/pull/17567
< bitcoin-git> [bitcoin] fanquake closed pull request #16738: gui: Remove needless GCC diagnostic pragma (master...20190827-gcc-pragmas) https://github.com/bitcoin/bitcoin/pull/16738
< bitcoin-git> [bitcoin] achow101 opened pull request #17568: wallet: fix when sufficient preset inputs and subtractFeeFromOutputs (master...fix-bnb-preset-sub) https://github.com/bitcoin/bitcoin/pull/17568
< meshcollider> achow101: I'm not sure why my test passes but yours doesn't
< meshcollider> I'm sure I've built the right one and running the right test ?
< meshcollider> Can someone else test #17568 quickly pls
< gribble> https://github.com/bitcoin/bitcoin/issues/17568 | wallet: fix when sufficient preset inputs and subtractFeeFromOutputs by achow101 . Pull Request #17568 . bitcoin/bitcoin . GitHub
< wumpus> sure i'll test
< wumpus> meshcollider: I get "test_framework.authproxy.JSONRPCException: Fee exceeds maximum configured by -maxtxfee (-4)" on master with the new test
< meshcollider> Hmm ok it must be something on my end then
< wumpus> isn't that what you expected?
< meshcollider> It's what is supposed to happen yeah, but it passed when I ran it lol
< wumpus> strange!
< wumpus> now tried to run it a few times, it seems reproducible, rpc_fundrawtransaction.py fails every time
< wumpus> after recompiling with #17568 it passes
< gribble> https://github.com/bitcoin/bitcoin/issues/17568 | wallet: fix when sufficient preset inputs and subtractFeeFromOutputs by achow101 . Pull Request #17568 . bitcoin/bitcoin . GitHub
< bitcoin-git> [bitcoin] laanwj opened pull request #17569: build: Allow export of environ symbols (master...2019_11_environ) https://github.com/bitcoin/bitcoin/pull/17569
< wumpus> fanquake: hopefully #17569 is the last step
< gribble> https://github.com/bitcoin/bitcoin/issues/17569 | build: Allow export of environ symbols by laanwj . Pull Request #17569 . bitcoin/bitcoin . GitHub
< bitcoin-git> [bitcoin] KaanKC opened pull request #17570: test: add unit test for non-standard txs w/ too large tx size (master...test_unit_IsStandardTx_tx-size) https://github.com/bitcoin/bitcoin/pull/17570
< JeremyCrookshank> Morning, is there a handy guide for developing the GUI using QT? I swear i've seen one but can't find it
< sipsorcery> JeremyCrookshank: only one I've seen is readme at https://github.com/bitcoin/bitcoin/tree/master/src/qt
< sipsorcery> someone else asked the same question here last week.
< JeremyCrookshank> Perfect cheers
< emilengler> How can I execute multiple tests in one file?
< emilengler> Adding the second class to main does not work :/
< wumpus> JeremyCrookshank: there's a lot of documentation on the Qt site, I think you need to be more specific in your question; FWIW, you can open the .ui files in qt designer
< wumpus> emilengler: which kind of tests? unit or functional?
< emilengler> wumpus: Functional
< jonatack> emilengler: did you have a look at test/functional/example_test.py?
< emilengler> jonatack: Sure
< emilengler> In #17536 it was said that it is possible to add multiple *.main() calls to the main function and then execute them
< gribble> https://github.com/bitcoin/bitcoin/issues/17536 | test: Add test for rpc_whitelist by emilengler . Pull Request #17536 . bitcoin/bitcoin . GitHub
< wumpus> emilengler: you should only have one class per implementation file there; I'm not sure why you'd want multiple, some bigger tests do logical separateion of parts by having more methods and calling them in turn
< emilengler> But it does not work, also a "tail test/functional/*" does not show this aprroach for any other file (as far as I looked so far)
< emilengler> wumpus: I need to test two states of a cli arg
< emilengler> foo=0 and foo=1
< emilengler> As bitcoind needs a restart for it I thought about doing it in 2 files
< wumpus> you can restart bitcoind from your test, no problem
< emilengler> Or in one file with multiple classes which seem to not work
< emilengler> wumpus: Really? How?
< wumpus> there's plenty of tests that check different argument combinations, the easiest way is to have it spawn multiple bitcoind nodes initially, but if that's too much combinations, you can always start and stop nodes on request
< emilengler> Ok I will try, thanks :)
< wumpus> stop_node, restart_node,..
< jonatack> emilengler: test_framework.py:445: def restart_node(self, i, extra_args=None):
< jonatack> emilengler: (git grep "def restart")
< wumpus> AppVeyor is failing on "bench_verify_script.obj : error LNK2001: unresolved external symbol "struct CMutableTransaction __cdecl BuildSpendingTransaction(class CScript const &,struct CScriptWitness const &,class CTransaction const &)" (?BuildSpendingTransaction@@YA?AUCMutableTrans"
< wumpus> did anyone forget to update the MSVC build system?
< sipsorcery> wumpus: on master?
< wumpus> I think so, I'm seeing the same error in many PRs, even #17561 which only updates docs
< gribble> https://github.com/bitcoin/bitcoin/issues/17561 | doc: Changed MiniUPnPc link to https in dependencies.md by sandakersmann . Pull Request #17561 . bitcoin/bitcoin . GitHub
< sipsorcery> will check with a local msvc build (otherwise it's a 1hr turn around time for appveyor)
< wumpus> I think #17542 is most likely to have introduced it of recent changes
< gribble> https://github.com/bitcoin/bitcoin/issues/17542 | build: Create test utility library from src/test/util/ by brakmic . Pull Request #17542 . bitcoin/bitcoin . GitHub
< sipsorcery> wumpus: yes the msvc build of bench_bitcoin is failing on master. i will investigate.
< wumpus> sipsorcery: thanks!
< sipsorcery> wumpus: np. you were correct the new libtest_util lib from #17452 is not part of the msvc build config. i've added and will submit PR imminently.
< gribble> https://github.com/bitcoin/bitcoin/issues/17452 | lib: update fuzz directory in .gitignore by jonatack . Pull Request #17452 . bitcoin/bitcoin . GitHub
< darosior> meshcollider, wumpus, achow101: The test added in #17568 pass on master for me as well..
< gribble> https://github.com/bitcoin/bitcoin/issues/17568 | wallet: fix when sufficient preset inputs and subtractFeeFromOutputs by achow101 . Pull Request #17568 . bitcoin/bitcoin . GitHub
< bitcoin-git> [bitcoin] sipsorcery opened pull request #17571: WIP: Add libtest_util library to msvc build configuration (master...msvc_libtestutil) https://github.com/bitcoin/bitcoin/pull/17571
< JeremyCrookshank> Anyone know why Qt-creator says the Bitcoin form .h files aren't part of the project? Not able to add slots because of it
< JeremyCrookshank> Doesn't do it on the sample project
< wumpus> the .pro isn't a suitable build system for building the project, it simply collects the form files for editing in qt creator / designer, it might be that some advanced functionality doesn't work
< JeremyCrookshank> Hmm okay so i'll have to code in slots manually as the qt-creator option doesn't work
< wumpus> yes
< JeremyCrookshank> Thats annoying :p
< wumpus> yes it is, but it's also not really worth it to maintain another build system for it, and I think most devs prefer implementing the hooks and such manually to have more control over what happens
< wumpus> also .pro is very limited as build systems go, I don't think it'd even be possible to use it anymore to build the full project. ofc, your could simply add the .h files to it and see that helps.
< wumpus> if it's just a matter of the IDE needing to find the code
< JeremyCrookshank> Yeah tried that already unfortunately
< fanquake> wumpus hopefully ?
< wumpus> at least locally it seems to make the gitian linux build pass again
< wumpus> or not, I get a NX failure after "Checking binary security" for RISC-V now, wtf
< wumpus> crap, after a failed build it doesn't cache the dependencies, so it needs to build qt all over again next time...
< fanquake> ugh
< wumpus> I wonder what broke the rv64 build, it might well be another change entirely, or something is messed up here locally
< bitcoin-git> [bitcoin] fanquake pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/0b79caf65801...33c103e2fe67
< bitcoin-git> bitcoin/master 794fe91 Hennadii Stepanov: doc: Update and improve Developer Notes
< bitcoin-git> bitcoin/master 33c103e fanquake: Merge #17539: doc: Update and improve Developer Notes
< bitcoin-git> [bitcoin] fanquake merged pull request #17539: doc: Update and improve Developer Notes (master...20191120-developer-notes) https://github.com/bitcoin/bitcoin/pull/17539
< achow101> meshcollider: darosior: it consistently fails for me on master
< wumpus> same here
< cfields> jonasschnelli: Looks like your macos sig is invalid :(
< cfields> cory@Macbook:~/dev/bitcoin$ spctl --assess -vv /Volumes/Bitcoin-Core/Bitcoin-Qt.app
< cfields> /Volumes/Bitcoin-Core/Bitcoin-Qt.app: invalid Info.plist (plist or signature have been modified)
< cfields> Oh wait, might be my local problem.
< cfields> Yup, all good.
< fanquake> ?
< cfields> Gitian builders: detached sigs for v0.19.0.1 are up
< bitcoin-git> [bitcoin] TheBlueMatt opened pull request #17573: Seed RNG with precision timestamps on receipt of net messages. (master...2019-11-rng-netmsg) https://github.com/bitcoin/bitcoin/pull/17573
< wumpus> cfields: jonasschnelli: great, thanks!
< bitcoin-git> [bitcoin] fanquake pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/33c103e2fe67...2c98e2024d52
< bitcoin-git> bitcoin/master b509554 Aaron Clauson: Added libtest_util library to msvc build configuration.
< bitcoin-git> bitcoin/master 2c98e20 fanquake: Merge #17571: Add libtest_util library to msvc build configuration
< bitcoin-git> [bitcoin] fanquake merged pull request #17571: Add libtest_util library to msvc build configuration (master...msvc_libtestutil) https://github.com/bitcoin/bitcoin/pull/17571
< provoostenator> ^ rebase time!
< wumpus> the gitian RV64 check failure is not caused by #17270, it appears, but some later change, I'm bisecting
< gribble> https://github.com/bitcoin/bitcoin/issues/17270 | Feed environment data into RNG initializers by sipa . Pull Request #17270 . bitcoin/bitcoin . GitHub
< sipa> wumpus: strange
< wumpus> i suspect it's some build system change that makes it eat a required flag
< fanquake> wumpus maybe it was broken when we removed OpenSSL ?
< wumpus> definitely possible !
< fanquake> "PS this seems to be causing a noise inversely proportional to the complexity of the change" https://github.com/bitcoin/bitcoin/pull/17563#issuecomment-557829862
< fanquake> What else was he expecting when he PR'd an initially poorly explained change to the ifdef hackery inside random.cpp
< wumpus> we're pretty consistent in this, too, I've closed PRs of myself too when 'working around a warning' was just not worth the mess
< fanquake> The initial change just felt like a Debian OpenSSL Valgrind scenario
< wumpus> bisected it to b49b6b0f7090cc15860d815fb0ef306ddfc718ba, so indeed seems that the change has to do with openssl removal, very curious
< fanquake> interesting
< wumpus> it's not even a build system change, just the last change that causes no symbols from openssl to be used anymore
< sipa> heh
< wumpus> it looks like we never explicitly pass -Wl,--noexecstack except to openssl
< meshcollider> darosior: thanks for testing, good to know it's not just me haha. I'll try and investigate today
< wumpus> all the other architectures default to it, though, so might as well be an ubuntu compiler issue
< potatoe_face> %
< wumpus> it's good that we have the security check or we'd never have noticed; the workaround of adding -Wl,-z,noexecstack explicitly on RISC-V in the descriptor seems to work, but I'm still trying to figure out why it doesn't get set automatically there
< fanquake> wumpus good hunting
< wumpus> thanks! this one was very sneaky; hopefully this was the last problem and we can start using gitian again
< achow101> meshcollider: are you sure you have an updated master? The issue was only after #17290 was merged
< gribble> https://github.com/bitcoin/bitcoin/issues/17290 | Enable BnB coin selection for preset inputs and subtract fee from outputs by achow101 . Pull Request #17290 . bitcoin/bitcoin . GitHub
< meshcollider> well yeah I merged it haha
< achow101> could be an old branch/ different worktree, i dunno
< meshcollider> Yeah I'll double check later and recompile fresh master