--- Log opened Mon Jul 05 00:00:57 2021
00:55 -!- reallll is now known as belcher
05:20 -!- kallewoof is now known as kalle
05:41 -!- kanzure [~kanzure@user/kanzure] has joined #secp256k1
06:07 -!- dr-orlovsky [~dr-orlovs@31.14.40.19] has quit [Quit: ZNC 1.8.0 - https://znc.in]
06:29 < nickler_> And yes, it sucks most of the testing happens in CI. However, there's this cirrus-cli command which reads cirrus.yml and runs the tasks in docker. 
06:29 < nickler_> Needs this fix: https://github.com/cirruslabs/cirrus-cli/issues/418 but should be easy to integrate into a review workflow.
07:02 < gmaxwell> probably a shell one-liner perhaps coupled with some comments in the cirrus config could just invoke the bulk of the tests locally-- at least the ones that don't need weirdo tooling setups.
07:03 < gmaxwell> or even just enabling -Werror and running make check would be pretty good, though it wouldn't hit all the configs.
07:30 -!- roconnor [~roconnor@host-45-78-208-216.dyn.295.ca] has joined #secp256k1
08:15 < nickler_> most issues have happened in the weirdo setups
08:15 < nickler_> but yeah, in this case, `git config --add githubmerge.testcmd "make clean && ./configure CFLAGS=-Werror && make check"` would have prevented this
08:32 < gmaxwell> https://github.com/bitcoin-core/secp256k1/pull/844#pullrequestreview-698937387  this person has left a couple of these
08:49 < real_or_random> yeah I reported it but  I can't do anything about it... apparently github lets me delete only comments but not PR reviews...
09:12 -!- luke-jr [~luke-jr@user/luke-jr] has quit [Ping timeout: 252 seconds]
09:13 -!- luke-jr [~luke-jr@user/luke-jr] has joined #secp256k1
11:35 -!- Netsplit *.net <-> *.split quits: gmaxwell, michaelfolkson, sipa
11:40 -!- Netsplit over, joins: michaelfolkson, sipa, gmaxwell
13:22 < gmaxwell> real_or_random: it worked! (#963)
13:47 < nickler_> gmaxwell: want to have another look? otherwise I'd merge
13:52 < gmaxwell> nickler_: ACK.
13:53 < gmaxwell> nickler_: changes all look good, plus ci is failed on master, so ... getting that fixed will be good.
13:55 < nickler_> cool, thanks
14:00 < gmaxwell> A follow on PR might want to see if the diff to remove the -fpermissive is too offensive.
14:15 < roconnor> wasn't SECP256K1_ECMULT_CONTEXT_PREALLOCATED_SIZE important?
14:15 < roconnor> oh it was static
14:16 < roconnor> I guess it wasn't important.
14:16 < roconnor> no wait, everything is static
14:23 < real_or_random> roconnor: it is important but it was declared twice, in the header and the _impl.h
14:25 < roconnor> oh right. I see now.
14:26 < real_or_random> this was just me not knowing how to write C
14:38 < real_or_random> ok, that's not true. it was valid C but not valid C++
14:38 < real_or_random> but it's a C library, so who cares ^^
14:46 -!- luke-jr [~luke-jr@user/luke-jr] has quit [Read error: Connection reset by peer]
14:47 -!- luke-jr [~luke-jr@user/luke-jr] has joined #secp256k1
14:49 < gmaxwell> C++ tried pretty hard to be backwards compatible with C, when you're not you're probably doing something at least somewhat weird or gross. :)
17:25 -!- belcher [~belcher@user/belcher] has quit [Ping timeout: 268 seconds]
17:38 -!- belcher [~belcher@user/belcher] has joined #secp256k1
--- Log closed Tue Jul 06 00:00:58 2021