Time Nick Message 07:44 TenPlus1 hi folks 07:44 TenPlus1 is anything up with the forums today, cannot sign in ?!?! 12:44 TenPlus1 hi folks 12:44 TenPlus1 is anyone able to sign onto forums ? 12:47 MTDiscord TenPlus1: I am able to sign in. 12:47 TenPlus1 hey luatic, so it seems to be just me then :) lol 12:49 TenPlus1 any forum devs/admin handy ?? 13:10 Krock TenPlus1: have you yet tried to clear the cookies for forum.minetest.net and forum.luanti.org selectively? perhaps something's corrupt. 13:11 TenPlus1 Hi Krock, yeah did all that, cleared cookies and cache and disabled adblock incase it was that, nothing lets me in and keeps referring to the new reCAPTCHA v3 13:13 TenPlus1 ah nice, someone has fiddled and fixed it for me :) yay! 13:13 * TenPlus1 wonders what the issue was ?! 13:21 TenPlus1 thanks to whoever fixed it (looks in ruben/celeron's) direction 15:15 celeron55 i don't think anyone has done anything. supposedly something was cached somewhere and then wasn't anymore 15:41 MTDiscord merging #15476, #15233, #15542, #15541, #15540 in 15m 15:41 ShadowBot https://github.com/minetest/minetest/issues/15476 -- Change `NS` symbol for `core.translate_n` to `PS` to avoid clash with old code by Wuzzy2 15:41 ShadowBot https://github.com/minetest/minetest/issues/15233 -- Performance boost for Raycast by sfence 15:41 ShadowBot https://github.com/minetest/minetest/issues/15542 -- Test & document conventions used by matrix4::setRotation* by appgurueu 15:41 ShadowBot https://github.com/minetest/minetest/issues/15541 -- Irrlicht: Get rid of obsolete `setDebugName` by appgurueu 15:41 ShadowBot https://github.com/minetest/minetest/issues/15540 -- Get rid of now obsolete workaround for collision detection bugs by appgurueu 16:04 MTDiscord done 16:57 sfan5 opinions welcome on 15552 16:57 sfan5 #15552 16:57 ShadowBot https://github.com/minetest/minetest/issues/15552 -- [no sq] Make OpenGL 3 the preferred (default) video driver by sfan5 17:14 Krock LGTM 17:34 sfan5 we should really split clientmap.cpp rendering and non-rendering parts 17:35 sfan5 the file has >1000 loc 17:35 sfan5 even the culler could get its own file 17:37 sfan5 @luatic I can no longer point nodes 17:37 sfan5 unless I'm inside of them 17:37 sfan5 this suggests we don't have a single unit test 17:38 sfan5 (on raycasts) 17:40 Krock https://github.com/minetest/minetest/pull/15233#issuecomment-2543192980 17:40 Krock I'll revert this in 15 minutes unless there are objections. 17:40 Krock I dug though the code for a while and making shortcuts like these is not trivial 17:41 sfan5 sounds ok 17:41 sfan5 unless luatic says he can fix it within 1h 17:48 MTDiscord There is no release soon, so I don't see a reason to revert it so soon? 17:49 sfan5 dev build are supposed to be usable 17:56 MTDiscord sfan5, Krock: I propose reverting. See #15553. 17:56 ShadowBot https://github.com/minetest/minetest/issues/15553 -- Revert "Optimize raycast performance" by appgurueu 17:57 MTDiscord I see no easy way to salvage this. 17:57 sfan5 rip 18:03 MTDiscord as a speck of optimism, i think there could be more elaborate ways to do basically the same optimization (for example visiting voxels on the ray and everything surrounding the ray in two entirely separate loops and then merging the results) but that will require re-doing everything properly including unittests, nothing i would want to rush in any case 18:05 MTDiscord merging #15553 in 10 minutes then. F 18:05 ShadowBot https://github.com/minetest/minetest/issues/15553 -- Revert "Optimize raycast performance" by appgurueu 18:12 Krock Should "core.get_connected_players" only return players where the on_joinplayer callback was already run? 18:13 Krock asking because it apparently may return pre-callback players too 18:14 sfan5 we will need a team of scientists to answer that one 18:14 Krock addActiveObject is called in ServerEnvironment::loadPlayer, which happens before Server::handleCommand_ClientReady (where the join callback is run) 18:15 Krock oh no. I am doomed. No-one can ever solve this issue. 18:16 Krock PR incoming I guess 18:16 sfan5 https://github.com/minetest/minetest/actions/runs/12331955766/job/34419184900?pr=15531#step:6:639 mysterious test failure appeared again 18:17 Krock sfan5: I've seen it occur in some other PRs too 18:17 [MatrxMT] #12189 ? 18:17 ShadowBot https://github.com/minetest/minetest/issues/12189 -- Reorder client initialization by SmallJoker 18:17 Krock maybe at a 10% chance? 18:18 Krock grorp: oh that explains why it seemed to familiar 18:18 MTDiscord i was about to wonder :p 18:19 MTDiscord sfan5: so we revert now, add the unit test later? 18:28 MTDiscord will revert for now so our master becomes usable again and add a unit test to devtest later 19:08 MTDiscord #15555 (dang, nice PR number) 19:08 ShadowBot https://github.com/minetest/minetest/issues/15555 -- Add unit test for raycasts falsely skipping nodes by appgurueu 19:20 pgimeno hm, are there non-deterministic unit tests, like checking random cases? 19:22 MTDiscord Could the node-pointing issue have to do with the dead node-pointing code I found earlier? 19:24 MTDiscord Ah, no, I had figured that out already and it wasn't dead code. 19:26 MTDiscord A unit test intended to be deterministic could also be undeterministic due to undefined behavior. 19:27 MTDiscord e.g. an uninitialized variable somewhere could cause an issue if it has a particular value, essentially introducing randomness into the test run. 19:28 MTDiscord off to watch a movie for now, i'll have to do a bit more work on the unit test later 19:33 MTDiscord pgimeno, I think the test is actually checking random cases: it doesn't set a specific seed when it does the map generation. 19:34 MTDiscord I didn't finish reading the mapgen code to see where the seed gets set, but my hypothesis is it picks a random one if one isn't configured. 19:34 MTDiscord One of the seeds must result in a map that violates the expected condition. 19:37 pgimeno thanks, I was referring to intentionally random unit tests; I was thinking about displaying the random seed used for the test, but if it comes from the mapgen, well... in my experience there isn't a consistent unique seed (not sure if that was ever fixed, I suppose not) 19:39 MTDiscord We probably could add something that just prints out the seed being used, and let that run in CI until the test fails again. 19:39 MTDiscord Then we look at what seed it had when it failed, and we can reproduce it. But if it fails 10% of the time like someone said, it can probably be reproduced with some runs locally as well. 20:22 MTDiscord I just noticed the new dev wiki is live 🎉 https://dev.luanti.org/ 20:35 MTDiscord The engine structure diagram is quite blurry. But this is awesome. 20:36 MTDiscord yeah wsor seems to have just saved the thumbnail of it. I have a full sized copy saved I can replace it with 20:36 MTDiscord I think the formatting for the thread responsibilities might be damaged as well. 20:36 MTDiscord just going to encourage discussion over in #minetest-docs-irc as to not spam here 20:37 MTDiscord iirc #mintest-docs for the irc folks 21:02 MTDiscord Nice work with the wiki, looks really good! 21:02 MTDiscord thanks, but also real thanks to https://themes.gohugo.io/themes/hugo-book/ for the template 21:06 MTDiscord Ah, nice. Hugo looks nice, personally never tried it though. 21:40 MTDiscord very nice