01:07:26 <UkoeHB> Well I'm thoroughly confused. If I make a std::thread and immediately join on it in the perf test main function, then my perf tests 'kind of' work (they spew sloppy result messages as if the test is actually running on multiple threads or something). If I make that std::thread in a { } block instead of in-line (so the thread destroys itself before the perf test is run), then my perf tests don't work.
01:41:41 <UkoeHB> whoops guess I had some UB from failing to initialize part of the test params
03:51:06 <jeffro256[m]> Damn UB
15:42:12 <luigi1111w> .merges
15:42:12 -xmr-pr- 8720 8721 8736 8737 8738 8739
15:47:44 <luigi1111w> am I tagging 8739?
15:47:46 <luigi1111w> selsta 
16:07:27 <selsta> luigi1111w: tagging?
16:07:34 <selsta> i'd wait a bit before tag
16:08:02 <selsta> see if everything builds etc
16:08:36 <luigi1111w> good day, sir
20:34:24 <selsta> hyc: any idea here? https://github.com/monero-project/monero/issues/8742
20:35:13 <hyc> never heard of such a thing
20:35:32 <hyc> maybe the full disk encryption is temporarily using space?
20:35:50 <hyc> or they have a virus :P
20:36:43 <hyc> ohhh. 15GB/hour - maybe that's just logfile output
20:37:17 <selsta> I'll ask for their config
20:37:36 <hyc> if logfile rotation is happening but somehow the filehandles didn't get closed, then deleting the old logfiles wouldn't release their space
20:38:45 <hyc> kinda makes sense with the "everything looks ok after restart"
20:41:30 <hyc> he should check if it occurs if using a separate drive, unencrypted
23:22:45 <selsta> luigi1111: you can tag now, everything good