Time Nick Message 00:28 paradust erle: This rendering bug.. do you have instructions for reproducing it? I can usually narrow things down without bisecting, if I have a reproducer. 00:30 paradust Especially on the web platform, there is a lot of good tooling. One can watch the render happen frame by frame, mesh by mesh. 00:33 erle paradust wait, which rendering bug? 00:35 paradust I thought you said you've been trying to bisect a rendering bug. "slime UV thing"? 00:35 erle x2048 sfan5 i summarized the nodebox rendering optimization discussion on IRC in the forum https://forum.minetest.net/viewtopic.php?f=7&t=28068&p=409857#p409857 00:36 erle paradust oh yeah! wait, i look for the issue 00:36 erle paradust https://github.com/minetest/minetest/issues/11852 00:36 erle paradust, https://github.com/minetest/minetest/issues/11852#issuecomment-992478961 00:37 paradust thanks 00:38 erle paradust, it would be nice if you crosspost any findings to https://git.minetest.land/Mineclonia/Mineclonia/issues/211 08:15 erle sfan5 x2048 it's “let's make every cube, stair & slab that is not ground content a lego brick” time https://git.minetest.land/erlehmann/studs/src/branch/master/screenshot.png 08:15 erle :D 08:16 erle oops, wrong channel 08:17 erle x2048 anyways, you can use this mod to stresstest your rendering improvements, it makes lots of faces. just remove the check for ground content to get a worst case scenario. 08:18 MTDiscord oh man, you even made studs on the stairs and slabs, that looks so good! 08:19 erle yes and i hope x2048 can turbocharge the face recognition so we i can turn the entire world into legos 08:20 erle also i can't think of a better stress test that is actually fun 08:20 erle i guess the holy grail for optimization problems are the GreenXenith hexagonal pillars 08:21 sfan5 Especially on the web platform, there is a lot of good tooling. One can watch the render happen frame by frame, mesh by mesh. 08:21 sfan5 renderdoc works for that on desktop, but only with the ogles2 driver 08:22 erle btw, for me some full nodeboxes with studs were entirely black without smooth lighting checked (ogl renderer). any idea if i missed some parameter? 08:22 erle (or could this be a rendering bug?) 08:24 sfan5 Zughy[m]: sneak animations are up to the game but the engine should provide a way to make it work with local animations 08:24 sfan5 so yes what Warr said 09:01 Zughy[m] sfan5: if it's alright with you, I'd open a different issue and close that one, because it's misleading imho 09:18 sfan5 rename it? 09:22 Zughy[m] I can't 09:33 celeron55 github's "triage" permission level doesn't allow some things like that, and it's silly 09:33 celeron55 the next up is "write", which allows push access and i think also renaming issues 09:34 erle that is silly indeed 09:34 celeron55 i'd be fine with just giving that to triagers, and trusting them not to mess with the code 09:34 celeron55 anyone else think the same? 09:36 erle could you disallow pushing to master for everyone except a list of users maybe? 09:36 erle i'm probably too spoiled by the feature set of gitea lol 09:44 sfan5 better not to give triagers write permission IMO 10:39 celeron55 well then, maybe just complain very loudly to github 13:01 josiah_wi Hi. I want to compile a list of known architectural or design limitations that have been encountered by people trying to refactor Minetest. If you've run into anything before, please let me know. 13:58 erle celeron55 if “complain to github” would do anything, they would by now be able to render the git log topologically ordered :P 17:18 MTDiscord @paradust I think this message was to you... 17:46 Zughy[m] is #4714 still a thing? 17:46 ShadowBot https://github.com/minetest/minetest/issues/4714 -- Visual weapon swing time doesn't correspond with full punch interval 18:08 MTDiscord I'm pretty sure it isn't 18:15 sfan5 http://sprunge.us/vhPYTg?diff pushing this soon 18:29 josiah_wi As the only outspoken proponent of ugly bracing in this channel, I am here obliged to mention that the '=' sign is unnecessary. xD 18:30 sfan5 uh huh so it's the same syntax 18:50 josiah_wi I'm not sure whether you meant that as a statement or a question. Semantically, I believe it's the same. Syntactically, I think it's different, but I can't even remember the term. C++ has innumerable types of initialization lol. 18:53 josiah_wi Ah, aggregate initialization is the right term for what you did in the patch. No idea what it even is. 18:53 josiah_wi Nope, nvmnd, that's a specific subset of the other kinds. 18:54 sfan5 https://en.cppreference.com/w/cpp/language/zero_initialization 18:54 sfan5 (2) 18:55 josiah_wi Ah, it's list initialization. 18:55 josiah_wi It's specifically zero initialization using list initialization. 18:55 josiah_wi Now I can forget all that because it's painful to try to remember. 18:57 sfan5 for extra confusion: zero initialization in C is = {0} while in C++ that only initializes the first member 18:57 rubenwardy merging https://github.com/rubenwardy/minetest/commit/f4a3c5a67d32229305386168776189105055ef60 in 5 18:57 sfan5 rubenwardy: is that backportable 18:57 rubenwardy yeah, trivially 18:57 rubenwardy I can add it to backports too 18:57 sfan5 no need