Time Nick Message 07:46 MTDiscord luatic: Should I look at it on apple arm HW? 11:12 luatic if you want to, sure, you can try tweaking the tolerance until you get one where it doesn't fail in many runs. 11:13 luatic (ultimately though it's not that important and i can eventually look at it in CI) 11:37 pgimeno I should probably take a look at that test 11:38 pgimeno IEEE float operations should be much more consistent among platforms than that, maybe there's a too aggressive math optimization performed on Apple ARM? 11:40 pgimeno also I'm not sure if that degree of imprecision is acceptable for regular use 11:40 luatic pgimeno: neither am i, i do think the tests hint at a problem 11:40 pgimeno I suspect the flakey cases are those near gimbal lock 11:41 pgimeno I'll take a look when I can, I'm currently busy 11:44 pgimeno is there an example of test case values that make the test fail? 12:32 luatic you should be able to reproduce it (on apple arm) using the catch seed 4072041937, see https://github.com/luanti-org/luanti/actions/runs/16634420051/job/47071789741 17:06 Krock sfan5: I am confused about the 'gather_git_credits.py' script. Isn't it supposed to list all major contributors under "previous contributors" ? 17:06 Krock For me, `git log --mailmap --pretty="format:%h %aN <%aE>" HEAD | wc -l` gives 2732 results, which is probably because I only have a shallow clone ..? 17:07 sfan5 yeah you need a full clone 17:08 sfan5 result on my pc https://0x0.st/s/Ep7S48_f5opF3M1_-ifzow/8RC-.txt 17:08 Krock alright thanks. I'll go through that. 17:08 sfan5 everything below the second --------- is no longer a major contributor, so Jeija is the last to appear in the credits 17:13 sfan5 Krock: feel free to also commit this https://0x0.st/8RCb.txt 17:15 Krock I was about to add a simple comment but this is even better. thanks. 17:20 Krock updated the PR 17:54 Krock I'd have time tomorrow to do the release (forum post already prepared) - unless someone would already like to help out with this right now